Precise Genome Editing by a Single Stranded Break
By Saloni Dhopte, Genetics and Genomics, ’23
Author’s note: Do you think CRISPR-Cas9 genome editing is amazing? Well, let me tell you about another technique that has been proven to be more accurate and efficient than CRISPR systems. It’s prime editing – a method of genome editing that utilizes a single stranded nick to edit DNA! I first learnt about prime editing from my then graduate student mentor, Dr. Peter Lynagh, while I was working in the Comai Lab at UC Davis. I remember scouring the internet for papers and looking up applications of this revolutionary technology, as it fascinated me beyond measure. So naturally, when we were tasked with writing a scientific review for my UWP 102B class, I had to write about this! Prime editing is an up-and-coming tool in gene editing, and I hope through my review, readers with basic to intermediate understanding of molecular genetics are able to share my amazement and admiration for prime editing!
Abstract
Current methods of genome editing involve a double stranded break on the DNA molecule, thus involving high frequency of unwanted base pair insertion/deletions (indels) and off-target effects in the repair process. Some are relatively limited in scope in terms of the length of the edit and specific type of base pair substitution. Recently, a new technique called prime editing has been developed which creates a single stranded break and shows high accuracy and editing proficiency along with minimal off-target effects in the genome. The molecular machinery of prime editing is very accurate; it recognizes the region of interest in the genome and is precisely able to insert an edited DNA sequence. In this review we discuss the mechanics of prime editing, compare it to other methods of genome editing, and investigate its applications as well as limitations in the health field. Numerous papers from the PubMed database point towards the potential of prime editing in repairing disease-causing mutations in humans. However, researchers are not using it for in vivo experiments (experiments that take place in a whole living organism) just yet, as they believe we need to learn a lot more about safe methods of delivery to human cell lines, side effects of the treatment, and overall efficiency of editing. Regardless, a lot of progress is consistently being made, including optimization of the prime editing machinery, online search databases, and ex-vivo applications. With the current pace of scientific discovery, the goal of using prime editing to its full potential will become reality soon enough.
Keywords: single-stranded break, genome editing, CRISPR-Cas9, guideRNA, disease-causing mutation repair.
Introduction
In 2019, Dr. David Liu published a paper introducing a new method of genome editing called prime editing (PE). The method is novel in its approach since it surpasses a majority of the drawbacks of pre-existing methods of genome engineering such as base editing (BE) and Clusters of Regularly Interspaced Palindromic Repeats (CRISPR)-CRISPR associated protein (Cas9), a method of genome editing adapted from the bacterial immune system’s defense mechanism. The main reason why PE is able to overcome these drawbacks is because it involves a single stranded break on the DNA molecule, as opposed to CRISPR-Cas9 which involves a double-stranded break (DSB), thus significantly reducing unwanted indels in the genome [1]. So far, a lot of studies have compared the efficiencies of PE with BE and CRISPR-Cas9. PE has been put to use in modeling diseases in organoids (tissue cultures that mimic in vivo organs) and researchers are developing ways of correcting the mutations that cause these diseases. In theory, PE could correct 90% of all disease-causing mutations in humans [2]. However, the unanimous opinion remains: a lot of work needs to be done in the field before PE can be used to correct mutations in a safe manner in vivo [1], [3], [4].
In this review I will investigate the mechanism and scope of prime editing and see how it can be used to study disease-causing mutations. By analyzing the technique and comparing it against the existing and well researched types of genome-editing, I will investigate PE’s effectiveness in repairing these mutations and see how the mechanism can be optimized.
How does Prime Editing work?
Mechanism
The prime editing system is based off of the CRISPR-Cas9 genome editing system. In CRISPR-Cas9, there is a single-stranded, guide RNA molecule (sgRNA) that is complementary to a specific region of the DNA and it is associated with a DNA endonuclease enzyme, Cas9. The sgRNA searches for and binds to the sequence homologies within the genome. Directed by the sgRNA to this complementary region of interest, the Cas9 protein then cuts the DNA, creating a DSB[2]. This DSB can then be repaired by double-strand break repair processes of varying levels of efficiency. In the case of prime editing, we still have the guide RNA, but it is modified such that it includes the sequence of our desired edit (referred to as prime editing guide RNA or pegRNA). Further, the Cas9 protein is partially inactivated (referred to as Cas9-nickase) such that it only cuts the 3’ strand of the complementary region on the DNA- creating a single stranded break [5]. And this makes all the difference. Since PE involves a single stranded break, the machinery doesn’t rely on the inefficient and unpredictable DSB repair mechanisms.
The single stranded break caused by Cas9 nickase leads to the formation of a 3’ DNA flap. This flap binds to a sequence on the pegRNA called the primer binding site (PBS). The reverse transcriptase enzyme, which is fused to Cas9 nickase, elongates the DNA using the 3’ flap as a primer, thus synthesizing a new “edited” DNA strand. The newly edited strand can then be incorporated into the original DNA molecule through a number of ways. The various approaches through which this resolution happens makes up the different kinds of prime editing.
Reproducible Graphic : The mechanics of prime editing
https://www.synthego.com/guide/crispr-methods/prime-editing
Types of Prime Editing
The mechanism PE1 relies on an endogenous endonuclease – FEN1, which cuts the 5’ end of our unedited DNA so that it doesn’t come in the way of our edited strand. With this cut, our edited DNA is thermodynamically favored to bind to the original DNA molecule. For PE2, the Liu group modified the sequence of RT to increase its thermostability, binding to DNA template and enzyme processivity. To take the mechanism to the next level, Anzalone et al. developed PE3 where they introduced a second guide RNA (gRNA) molecule that creates a 5’ nick on the non-edited strand, which allows for the edited DNA strand to be used as a template to complete the process. Finally, we have PE3b in which the gRNA is programmed such that it creates the 5’ nick only after the edited strand has been completely formed [5]. This approach led to reduced indel formation and improved editing. According to Anzalone et al., PE3 is 1.5 to 4.2-fold more efficient than PE2. PE3b, while not showing a significantly higher editing efficiency, has a 13-fold reduction in indel formation [2].
To summarize, the way PE works is by utilizing a special RNA molecule called pegRNA that encodes our desired DNA edit and targets the exact region of DNA where the edit has to be incorporated. Along with the Cas-9 nickase, it accurately creates a single stranded break, synthesizes the edited DNA, and finally incorporates the edit through various mechanisms – PE1, PE2, PE3 and PE3b.
Why is Prime Editing the best genome editing method we know?
Prime Editing versus CRISPR-Cas9
There are multiple characteristics of PE that make it better than our pre-existing methods of genome editing.
First, PE involves a single stranded break in the DNA as opposed to CRISPR-Cas9 editing which involves a double stranded break (DSB). A DSB can be repaired in one of two ways – via homology directed repair (HDR) or non-homologous end joining (NHEJ). The latter leads to a lot of indels. Thus, the editing efficiency of genome editing methods employing DSB’s is low. In contrast, PE doesn’t involve a DSB to begin with, thus the process of repairing the break doesn’t rely on the generation of random indels. Chemello et. al showed how the mutations causing Duchenne Muscular Dystrophy (DMD) are corrected with less unwanted effects via PE as opposed to CRISPR-Cas9 [4].They used prime editing to correct one of the most common mutations of DMD – the deletion of exon 51. Chemello et al. first attempted to restore the correct open reading frame (ORF) by inducing exon skipping. They used CRISPR Cas9 to systematically make two cuts such that exon 52 would be skipped, and the ORF would be restored. However they were unsuccessful due to the high rate of indels and off-target effects of CRISPR-Cas9. On the other hand, with PE, they were able to reframe the exon and precisely inserted two nucleotides into exon 52, thus bypassing the need for exon skipping entirely. This demonstrates the ability of PE to specifically target and edit DNA sequences in order to correct disease-causing mutations without the unwanted effects of double-stranded break repair pathways.
Further, since PE requires 3 separate hybridization events (pegRNA spacer to target DNA for Cas9 binding, pegRNA PBS to target DNA, and target DNA 3’ flap to RT product) to occur, it has significantly less off-target effects in the genome. In CRISPR-Cas9, the guide RNA, as it is searching DNA for complementarity, can bind to other regions of the genome with similar sequences, leading to DSB’s in places that were not targeted [2]. Kim et al. were unsuccessful in trying to correct DMD using CRISPR-Cas9, due to these off-target effects. However, they found no significant unwanted indels in the genome of mice hepatocytes that were prime edited to correct for HT1 [6]. Similarly, Geurts et al. performed whole genome sequence analysis on prime edited five colon organoids and reported no mutational differences among the edited organoid sequences [3]. These findings establish the safety of PE as compared to CRISPR-Cas9.
Prime Editing versus base editing
PE’s battle with base editing is not as one-sided as it is with CRISPR Cas9. Base editing performs rather well in most experiments.
There are two kinds of base editing – adenine base editing that can change a nucleotide from A to G or G to A, and cytosine base editing, which changes C to T or T to C. The mechanism involves an inactive Cas9 protein (dCas9) fused to a deaminase molecule which makes the respective base change possible [1]. Base editing doesn’t involve a break on the DNA molecule at all. It also doesn’t rely on the generation of random indels for editing. Hence, it is not surprising that the efficiency of this method surpasses that of PE. Geurts et al. reported that base editing induced correct mutations in 50% of the colonic organoids whereas prime editing was only able to reach 22% [3]. Similarly, Schene et al. found that using PE for the correction of mutations in liver organoids was less effective than using base editing. In this way, both papers report the same finding – when working with a mutation that can be corrected by base editing, it outperforms PE.
But here is the catch. First, base editing can make only four of the twelve possible base pair changes. If the disease of interest requires an adenine to be corrected into a cytosine, base editing doesn’t even come into the picture– that substitution is beyond its capability. This severely limits the scope of genome editing. The second drawback deals with the size of the edit. Some diseases require a stretch of nucleotides to be corrected– not just a single nucleotide. If the editing window for BE is increased to more than one nucleotide, especially if the edit includes more A or C bases, a lot of by-stander edits are observed. This is because the Cas9-deaminase complex makes all base substitutions in its range, including those we don’t want [2]. This problem doesn’t arise in PE because the pegRNA encodes highly specific DNA insertions up to 80 base pairs in length. Because 98-99% of insertions, deletions and duplications in the pathogenic human genetic variants are smaller than 30 base pairs, researchers have claimed that with PE, we will be able to correct 90% of disease-causing mutations in humans [1].
What has been accomplished so far using Prime Editing?
One of the most sought-after goals of genome editing is to be able to correct diseases-causing mutations. While PE is being used increasingly for its precision in editing DNA, it is a relatively new technique and so all of the research takes place in vitro ( in an artificial environment simulated to mimic the human body). In the hopes of eventually overcoming this gap and moving on to in vivo studies, scientists are also working on optimizing PE to have even less off-target effects. They have varied the molecular machinery, model organoids that mimic in vivo organ systems, and target mutations in different combinations, to find an approach with the best results. Optimization of the prime editing machinery is a well-established path to achieving the goal of using this technique in therapeutic applications.
Repair and modeling of disease-causing variants
Prime editing has been successfully used to model several diseases in human organoids. Broadly speaking, there are two main goals of these particular studies. First, to study the efficiency of PE, and working on ways of improving the mechanism. Geurts et al. utilized PE to model the mutation causing cystic fibrosis in human adult derived colonic organoids and then used PE to correct the mutation. Scientists employed both PE and BE for these steps and found that BE was more efficient in inducing intended mutations as compared to PE, but again, it remains limited to 4 of the 12 base pair substitutions. They acknowledged that if edits need to be made outside of this window, PE is the best approach [3].
The second kind of PE editing studies investigate the ways in which disease-causing mutations can be corrected. Schene et al. modeled mutations in liver organoids to mimic the development of liver cancer, then used PE [1]. Schene et al. and Guerts et al. both confirmed that PE is the better choice only for the subset of mutations not applicable for correction using BE [3].
With that, it is quite clear that we need to work a lot more on PE to further increase its efficiency. To do this, multiple researchers are focusing their efforts on optimization of PE.
Optimization of pegRNA’s
The prime editing machinery is highly advanced in structure. Compared to CRISPR-Cas9, there are fewer elements involved and thus less unwanted indels. The key player making this possible is the prime editing guide RNA (pegRNA). Researchers have worked extensively on optimizing the performance of the pegRNA through a variety of approaches. Lin et al. identified two main factors that have shown increased efficiency of editing: first, designing primer binding sites (PBS) on the pegRNA with melting temperature less than 30 degrees Celsius, and second: using not one, but twopegRNA’s encoding the same edit [7]. Together, these boost the editing efficiency 17.4-fold.
Moreover, there is an increase in resources and tools for pegRNA optimization. PegFinder is an online software that allows scientists to program the specific pegRNA to fit their experiment [3]. More recently, Lin et al describe the construction of their own web application called PlantPegDesigner. They claim that their tool is more user-friendly than PegFinder, as the latter necessitates experimental testing of pegRNA’s. PlantPegDesigner only requires a single DNA sequence as an input and provides a variety of parameters to be optimized by the user – an ideal candidate pegRNA [7]. This technology has the potential to greatly simplify prime editing experiments with plants, which in turn might lead to quickly reducing the knowledge gap in the field. Another similar web application is PrimeDesign – a tool that not only provides the user with an ideal pegRNA but visualizes the entire prime editing event. It allows users to rank pegRNA’s based on efficiency and includes extensive annotations. Additionally, Hsu et al. created a database called PrimeVar using all of these results, which can be used to search for pegRNA’s correcting ~70,000 pathogenic human genetic variants [8].
Conclusion
Prime editing is a novel breakthrough in the field of genome editing. It has been only three years since the publishing of Dr. David Liu’s original paper introducing the world to PE. PE is able to target and edit any region of the genome while avoiding drawbacks of current gene-editing methods, made possible by the induction of a single stranded break. Scientists have demonstrated the superiority of PE when compared to base editing and CRISPR-Cas9 editing. BE, although more accurate and known for less off-target effects than PE, can only correct a subset of base-pair substitutions. CRISPR-Cas9 involves a double stranded break on the DNA molecule, leading to high rates of unwanted insertions/deletions in the genome as compared to PE. Within a span of two years, four distinct types of PE have been developed – PE1, PE2, PE3 and PE4 – each more efficient than the last. The development of online tools such as PrimeDesign and PlantPegDesigner show the rate at which scientists are making progress with PE. However, we are far from the finish line. Most researchers still remain skeptical about the use of PE for in-vivo applications. While some say it is imperative to develop safe methods of delivery to human cell lines, others question the consequences of off-target effects in the genome. We don’t fully understand how PE might affect other cells of the subject [2]. Additionally, researchers aren’t certain about the longevity of prime edited disease corrections [4]. Most agree that in theory, prime editing will be revolutionary in terms of advancing human health, but given the relative recentness of the technology, there is still a lot of work to be done. Despite the gray area, PE certainly has a lot of potential and will be one of our strongest tools in improving human health in the future.
References:
- I. F. Schene et al., “Prime editing for functional repair in patient-derived disease models,” Nat Commun, vol. 11, no. 1, p. 5352, Dec. 2020, doi: 10.1038/s41467-020-19136-7.
- A. V. Anzalone et al., “Search-and-replace genome editing without double-strand breaks or donor DNA,” Nature, vol. 576, no. 7785, pp. 149–157, Dec. 2019, doi: 10.1038/s41586-019-1711-4.
- M. H. Geurts et al., “Evaluating CRISPR-based prime editing for cancer modeling and CFTR repair in organoids,” Life Sci. Alliance, vol. 4, no. 10, p. e202000940, Oct. 2021, doi: 10.26508/lsa.202000940.
- F. Chemello et al., “Precise correction of Duchenne muscular dystrophy exon deletion mutations by base and prime editing,” Sci. Adv., vol. 7, no. 18, p. eabg4910, Apr. 2021, doi: 10.1126/sciadv.abg4910.
- H. Jang et al., “Application of prime editing to the correction of mutations and phenotypes in adult mice with liver and eye diseases,” Nat Biomed Eng, vol. 6, no. 2, pp. 181–194, Feb. 2022, doi: 10.1038/s41551-021-00788-9.
- Y. Kim et al., “Adenine base editing and prime editing of chemically derived hepatic progenitors rescue genetic liver disease,” Cell Stem Cell, vol. 28, no. 9, pp. 1614-1624.e5, Sep. 2021, doi: 10.1016/j.stem.2021.04.010.
- Q. Lin et al., “High-efficiency prime editing with optimized, paired pegRNAs in plants,” Nat Biotechnol, vol. 39, no. 8, pp. 923–927, Aug. 2021, doi: 10.1038/s41587-021-00868-w.
- J. Y. Hsu et al., “PrimeDesign software for rapid and simplified design of prime editing guide RNAs,” Nat Commun, vol. 12, no. 1, p. 1034, Dec. 2021, doi: 10.1038/s41467-021-21337-7.
Elizabethkingia anophelis: an Emerging, Opportunistic Pathogen
By Nelly Escalante, Molecular and Medical Microbiology, ’23
Overview
Elizabethkingia is a family of gram-positive, aerobic bacteria that includes the species Elizabethkingia meningoseptica, Elizabethkingia miricola, Elizabethkingia anophelis, Elizabethkingia bruuniana, Elizabethkingia ursingii, and Elizabethkingia occulta [1]. E. meningoseptica and E. anophelis are the only species within the genus that have been observed to cause disease in humans. While previous research has characterized E. meningoseptica’s predominant role in infection, emerging research has revealed that E. anophelis has been responsible for most of the recent Elizabethkingia case reports.
Elizabethkingia anophelis is an emerging pathogen first discovered in 2011. It is a symbiotic bacterium that resides in the midgut of the mosquito Anopheles gambiae, which resides in the Gambia River region in central Africa [2]. While A. gambiae is endemic to that region, outbreaks have been observed in several Asian and African countries, with the biggest outbreak so far occurring in the United States. Most cases of E. anophelis are not due to direct contact with its host, A. gambiae, but rather are community-acquired in hospitals through a yet undescribed method of transmission.
Diagnosis
Clinical Presentation
Typical symptoms of E. anophelis infection include bacteremia and meningitis. Pyrexia, chills, and dyspnea have also been observed across several case reports. E. anophelis presents the greatest bacterial burden in the blood, causing bacteremia that can lead to further complications such as sepsis and septic shock. Removal of catheters or central lines may be a necessary approach to relieve bacteremia when E. anophelis is suspected [3].
Most of the information known about E. anophelis has come from case reports, as an animal model has not been developed yet to examine its pathogenesis in vivo. The first identified human case of E. anophelis infection was a case of neonatal meningitis in Africa. In this case, the 8-day-old patient experienced pyrexia, seizures, and apnea. Cerebrospinal fluid (CSF) analysis revealed hypoglycorrhachia [4]. In another case, a 7-month-old patient suffered from pyrexia, ecchymotic spots on the body, respiratory failure, and hemorrhaging [5]. These symptoms, however, are not considered to be within the standard clinical presentation of an E. anophelis infection and would only be seen with an especially acute bacterial burden. In both cases, the final diagnosis of E. anophelis infection was made after positive bacterial cultures were observed.
Diagnostic Criteria
Cultures are a powerful tool in the diagnosis of bacterial infection and are grown by sampling many bodily fluids, although blood and CSF are the most common. Once cultures are grown, they can be analyzed to identify the specific bacterium or bacteria causing the infection. Common methods used to identify the different Elizabethkingia species have been unable to differentiate between E. meningoseptica and E. anophelis with great accuracy. Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry, one of these common methods, utilizes a laser to ionize the bacterial sample and records the time it takes the ions to travel the length of a tube. Larger ions take a longer time, thus producing a mass spectrum that is known as a peptide mass fingerprint (PMF). The PMF of the sample is compared to the over 2,000 PMF of known bacteria species in the database [6].
MALDI-TOF mostly produces accurate identification of bacteria to the genus level. In combination with the lack of PMF samples of E. anophelis, this method has caused many cases of E. anophelis to be misidentified as E. meningoseptica. However, 16S ribosomal RNA gene sequencing has been successful in identifying E. anophelis as it directly uses genomic DNA to produce the 16s rRNA gene sequence and compare it to the more then 60,000 bacterial type strains in the database [7]. This analysis has shown that E. anophelis accounts for far more infections than E. meningoseptica.
Elizabethkingia anophelis culture taken from a patient and grown on 5% sheep blood agar
Treatment
Treatment regimens for E. anophelis infections have not yet been established because the range of antibiotic resistance of the bacterium has not been completely characterized. However, studies have shown that minocycline and levofloxacin are the most effective in treatment. Minocycline belongs to the class of tetracycline antibiotics that inhibit protein synthesis in both gram-positive and gram-negative bacteria and can be given orally. This medicine, however, cannot be given safely to children under the age of 8 [8]. Given that neonates are one of the most affected, other treatments are still being explored.
Levofloxacin, on the other hand, is part of a new group of fluoroquinolones that inhibits DNA gyrase and topoisomerase IV, enzymes that are essential to bacterial DNA replication. Levofloxacin is usually not administered to children except in life-threatening infections such as one by E. anophelis [9]. Moxifloxacin, a drug in the same class of antibiotics, was successful in the treatment of the first human case of E. anophelis infection.
E. anophelis has been classified as multi-drug resistant because it is not susceptible to common antibiotics such as β-lactams and β-lactam/lactamase inhibitors. Additionally, although many cases of E. anophelis have been misidentified as E. meningoseptica, they have distinct antimicrobial susceptibilities and require different treatments [10]. Many E. anophelis strains contain variants in the catB gene that confers antibiotic resistance to phenicol drugs and antibiotic inactivation enzymes [11].
The fatality of E. anophelis infection varies greatly across case reports, but in general has been estimated to be close to 30% [12]. Incorrect antimicrobial therapy regimens are a risk factor in the mortality of patients, which means deciding on the correct antibiotics is essential to ensuring a patient’s recovery and survival [13]. For example, antibiotics that are used to treat neonatal meningitis are ineffective against E. anophelis infection, further highlighting the importance of accurate diagnosis and treatment regimens.
Bacteriophages, also known as phages, are currently being investigated as an alternative to antibiotic treatment, given the multi-drug resistant nature of E. anophelis. In Taiwan, a phage named TCUEAP1 was isolated from the wastewater of a hospital. While there is no bacteriophage specific to E. anophelis, TCEUAP1 was able to infect three strains of the bacteria and reduce the number of colony-forming units (CFU). In a mouse model, the phage was able to decrease the bacterial load in their blood from 5×105 CFU/mL to 1×105 CFU/mL. In doing so, they were able to rescue 80% of the mice that would have otherwise died due to bacteremia [14]. Phages are a promising new therapy for treating multidrug resistant bacteria because they only attack their bacterial hosts and do so with a mechanism that is distinct from drugs.
Prevention and Future Research
As a nosocomial infection, the best prevention is good hygiene practices. Frequent hand washing by medical personnel as well as routine, thorough disinfection of surfaces may help in reducing the spread. Person-to-person transmission, either through direct or close contact with an infected individual, remains a possible infection mechanism that has yet to be confirmed by in vitro models. It has been proposed that mothers are able to vertically transmit the infection to their child during birth. The exact mechanism of how a person becomes infected by E. anophelis is unknown, but many research efforts are underway to describe its pathogenesis and route of transmission.
Recent research has shown that the bacterium has been able to evade the immune system’s defenses. Macrophages are among the first cells of the immune system to respond to an infection. They have an antibacterial polarization state known as classically activated (M1) macrophages and are activated when a pathogen is detected. In this state, they change their morphology to engulf pathogens through phagocytosis to reduce the bacterial burden. E. anophelis evades this detection and prevents M1 macrophages from activating through a yet unknown mechanism. If activated M1 macrophages are present, the bacteria are also able to avoid being engulfed, which may be due to the bacterial capsule surrounding the bacterium. This type of phagocytosis evasion using a bacterial capsule has been observed by other bacteria such as Salmonella and Mycobacterium [15]. Considering that most patients who contracted an E. anophelis infection were elderly, newborn, or immunocompromised, this type of immune system evasion may be a contributing factor to the high mortality of the infection [5].
Image taken of E. anophelis using phase contrast microscopy. Bacteria are stained with Maneval’s solution with empty space around the bacteria showing the bacterial capsule.
Overall, there are many mechanistic mysteries to Elizabethkingia anophelis that have yet to be investigated, but are nonetheless pertinent to the prevention of further outbreaks and improved patient outcomes.
References:
- Nicholson AC, Gulvik CA, Whitney AM, Humrighouse BW, Graziano J, Emery B, Bell M, Loparev V, Juieng P, Gartin J, Bizet C, Clermont D, Criscuolo A, Brisse S, Mcquiston JR. 2018. Revisiting the taxonomy of the genus Elizabethkingia using whole-genome sequencing optical mapping, and MALDI-TOF, along with proposal of three novel Elizabethkingia species: Elizabethkingia bruuniana sp. nov., Elizabethkingia ursingii sp. nov., and Elizabethkingia occulta sp. nov. A Van Leeuw J Microb [Internet]. 111(1):55-72. doi:10.1007/s10482-017-0926-3.
- Kampfer P, Matthews H, Glaeser SP, Martin K, Lodders N, Faye I. 2011. Elizabethkingia anophelis sp. Nov., isolated from the midgut of the mosquito Anopheles gambiae. Int J Syst Evol Micr [Internet]. 61:2670-2675. doi:10.1099/ijs.0.026393-0.
- Bush L. 2020. Bacteremia. Merck manual professional version. Merck Sharpe and Dohme. https://www.merckmanuals.com/professional/infectious-diseases/biology-of-infectious-disease/bacteremia?query=blood%20cultures.
- Frank T, Gody JC, Nguyen LBL, Berthet N, Le Fleche-Mateos A, Bata P, Rafai C, Kazanji M, Breurec S. 2013. First case of Elizabethkingia anophelis meningitis in the Central African Republic. Lancet [Internet]. doi:10.1016/S0140-6736(13)60318-9. 381(9880):1876.
- Mantoo MR, Ghimire JJ, Mohapatra S, Sankar J. 2021. Elizabethkingia anophelis infection in an infant: an unusual presentation. BMJ case reports [Internet]. 14(5):e240378. doi:10.1136/bcr-2021-243078.
- Singhal N, Kumar M, Kanaujia PK, Virdi JS. 2015. MALDI-TOF mass spectrometry: an emerging technology for microbial identification and diagnosis. Front Microbiol [Internet]. 6:791 doi:10.3389/fmicb.2015.00791
- Strejcek M, Smrhova T, Junkova P, Uhlik O. 2018. Whole-Cell MALDI-TOF MS versus 16s rRNA gene analysis for identification and dereplication of recurrent bacterial isolates. Front Microbiol [Internet]. 9:1294. doi:10.3389/fmicb.2018.01294
- Werth B. 2020. Tetracyclines. Merck manual professional version. Merck Sharpe and Dohme. https://www.merckmanuals.com/professional/infectious-diseases/bacteria-and-antibacterial-drugs/tetracyclines?query=minocycline.
- Werth B. 2020. Fluoroquinolones. Merck manual professional version. Merck Sharpe and Dohme. https://www.merckmanuals.com/professional/infectious-diseases/bacteria-and-antibacterial-drugs/fluoroquinolones?query=fluroquinolones
- Lin JN, Lai CH, Yang CH, Huang YH. 2018. Comparison of clinical manifestations, antimicrobial susceptibility patters, and mutations of fluoroquinolone target genes between Elizabethkingia meningoseptica and Elizabethkingia anophelis isolated in Taiwan. J Clin Med [Internet]. 7(12):538. doi:10.3390/jcm7120538. .
- Wang M, Gao H, Lin N, Zhang Y, Huang N, Walker ED, Ming D, Chen S, Hu S. 2019. The antibiotic resistance and pathogenicity of a multidrug-resistant Elizabethkingia anophelis isolate. Microbiol Open [Internet]. 8(11):e804. doi:10.1002/mbo3.804.
- Yang C, Liu Z, Yu S, Ye K, Li X, Shen D. 2021. Comparison of three species of Elizabethkingia genus by whole-genome sequence analysis. FEMS Microbiol Letters [Internet]. 368(5):fnab018. doi:10.1093/femsle/fnab018.
- Lin JN, Lai CH, Yang CH, Huang YH, Lin HH. 2018. Clinical manifestations, molecular characteristics, antimicrobial susceptibility patterns and contributions of target gene mutation to fluoroquinolone resistance in Elizabethkingia anophelis. J Antimicrob Chemoth [Internet]. 73(9):2497-2502. doi:10.1093/jac/dky197.
- Peng SY, Chen LK, Wu WJ, Paramita P, Yang PW, Li YZ, Lai MJ, Chang KC. 2020. Isolation and characterization of a new phage infecting Elizabethkingia anophelis and evaluation of its therapeutic efficacy in vitro and in vivo. Front Microbiol [Internet]. 11:728. doi:10.3389/fmicb.2020.00728.
- Mayura IPB, Gotoh K, Nishimura H, Nakai E, Mima T, Yamamoto Y, Yokota K, Matsushita O. 2021. Elizabethkingia anophelis, an emerging pathogen, inhibits RAW 264.7 macrophage function. Microbiol and Immunol [Internet]. 65:317-324. doi: 10.1111/1348-0421.12888.
An Evaluation of eDNA Sampling for Aquatic Species
By Isoline Donohue, Biological Sciences, ’23
Author’s Note: I wrote this literature review for UWP 102B during the spring quarter of 2022, and learned about the Aggie Transcript from that course. I chose to write about this topic because I am very interested in conservation biology and work as an undergraduate researcher in this field. I had heard of environmental DNA before at my lab, but wanted to learn more by doing my own research. From this piece, I hope readers learn about the new and exciting ways species monitoring is being done to preserve ecosystems.
Introduction
Marine species are experiencing higher population declines than many terrestrial species due to anthropogenic causes [1], such as increased water exports or runoff impacting habitats and behavioral patterns. Aquatic systems require a greater focus on species preservation, but keeping track of different species can be difficult. A first step in conservation involves genetic monitoring to track population decline. Genetic monitoring uses DNA to study variation within a species, as well as to discriminate between different species types in an environment. Monitoring a species may include the use of environmental DNA (eDNA) in lieu of DNA collected off of the organism itself. eDNA is a sample collected from a habitat, such as water from a stream, that contains the DNA of the species present. The skin, mucus, or hair of an organism when shed contains DNA that can be eluted and amplified to detect species [2]. Samples are then used for metabarcoding, or non-species specific detection, as well as for targeting a certain species of interest through assay development of different genetic markers [2]. Overall, this method of sample collection is both non-invasive and applicable for smaller populations when capture methods tend to lose reliability [1]. Traditional capture methods are not always successful in areas where a species may currently or recently inhabit, which is where eDNA can be used to discover new territories and monitor current ones.
Studies have been conducted with environmental samples to monitor endangered species, as well as non-native species that endanger other populations [1, 3]. eDNA detections often need validation through replicates or prior knowledge of inhabitants, as using eDNA is still a relatively new form of sample collection. Schmelzle et al. (2016) highlights how current eDNA research centers on standardizing DNA capture methods to ensure repeatability [4]. Therefore, studies are being conducted to compare eDNA to traditional capture methods to determine if positive detections can be made the same as if a tissue sample was taken from an organism.
Currently, there is concern over whether eDNA can be reliably used for instances where seasonality impacts species presence in an area [5], as well as instances where DNA may degrade before proper evaluation [6]. In order to understand how eDNA can be utilized to find the limitations of aquatic species detection, various studies have been compiled. In this review, we evaluate the current methodology, validity, effectiveness, and concerns of eDNA through studies centered on endangered or invasive aquatic species.
Methodology
While collection, extraction and analysis methods and materials are not always the same, the general process of going from an environmental sample to an identifiable DNA sequence of a species is. First, the water collected for any aquatic system sampling must be filtered. There are several methods to do this that involve different materials and containers to collect and filter water. Ratcliffe et al. (2020), for example, took water samples from the Irish and Celtic seas to detect key taxa in the area. The researchers strained the water using a syringe and filters, where the water was stored in the filter holder containers at -20°C until DNA extraction [7]. Alternatively, Boothroyd et al. (2016) filtered their water samples with a funnel and vacuum right before DNA extraction to prevent degradation. These samples are generally stored with ethanol at -20°C before DNA extraction and -80°C afterwards [1, 3, 4, 5, 6, 7, 8]. Franklin et al. (2018) and Dubreuil et al. (2021) used the Qiagen DNeasy Blood and Tissue Kit in a two day extraction process of lysis for DNA release, and subsequent washing/elution to isolate the DNA. Afterwards, qPCRs in each study were run against an assay containing the genetic marker for identification of a specific species of interest, or in some cases a range of taxa. Species and their relative abundances have the potential to be identified via DNA sequencing starting with an environmental sample.
Researchers take different approaches in validating positive detections, such as through the use of controls and assessment of assays. For example, Mauvisseau et al. (2019) detected pearl mussels in Lake Windermere. The researchers validated their assays targeting the COI and 16S genes, regions optimal for species identification, with statistical analysis on their level of detection and quantification before using them against the samples collected in the study. The eDNA detections were also tested against tissue samples and positive controls to ensure accuracy. This was a double-blind experiment during the water sample collection and filtering process, where sample site information was not revealed to the researchers until after analysis was done [2]. Dubreuil et al. (2021) cited Mauvisseau et al. (2019) as the researchers followed the same statistical guidelines, and their assays proved to be specific to only their species of interest via the positive detections observed. Lastly, Franklin et al. (2018) also took steps to evaluate an appropriate assay for their species. They found that the COI genetic marker also distinguished their species of interest, smallmouth bass, from other non-target species. They examined COI sequences of bass from different regions to obtain sequence specific primers. The developed assay was tested against database sequences, tissue samples, and finally eDNA itself [9]. These preliminary tests aid in ensuring accurate detections are being made when the assay is used against eDNA samples.
Additionally, repeatability in sample analysis can be used to rule out contamination and strengthen confidence in positive detections. Schmelzle et al. (2016) demonstrated this by testing each water sample with 6 replicate qPCRs alongside positive controls to detect tidewater gobies on the California coast. qPCRs are done to identify and quantify the DNA in the samples collected, while positive controls confirm the accuracy of target detections. Overall, research tends to center on validating eDNA detections, as variability can be high even within the same sampling region.
Genetic information from numerous species may be present in one water sample. Researchers may choose to target a specific species or to identify all of the species in the sample using DNA sequencing methods. National Park Service
Comparison to traditional methods
A non-invasive method
A benefit of eDNA sample collection is that it is noninvasive to an ecosystem. This is most beneficial for detecting low density species when capture methods tend to fail. Boothroyd et al. (2016) evaluated the effectiveness of eDNA in relation to traditional capture methods, specifically regarding spotted gar fish that were collected via netting. The downsides of netting include sample size limitations and disruption of habitats. Boothroyd et al. (2016) placed fyke nets (cylindrical fish traps) at the eDNA sample sites for 24 hours during the spotted gar spawning season to gather a representative sample. Fin clips from 12 reference samples of spotted gar captured were used for DNA extraction and downstream analysis to compare to the eDNA samples [1]. eDNA detections, made from 1 liter water samples, were consistent with areas spotted gar were previously known to inhabit, and were even made at sites where capture methods failed to pick up the species of interest when placed [1]. The researchers in this study also captured dozens of other fish species at the different bodies of water sampled while attempting to collect spotted gar. The variability of netting shows that it can cause disruption for more than just the species of interest, as well as limit the sample size.
Similarly, Schmelzle et al. (2016) used occupancy modeling to compare traditional capture methods against eDNA detection. This study noted the limitations of traditional capture methods in their low capture yield versus high cost and time requirement. A seine haul (vertical net placed in the water) of tidewater goby fish was taken at each sample location to compare to the water samples collected. It was found that eDNA detections from the water samples, validated against known goby presence and replicate qPCRs, were more effective than seining detections in accurately representing goby occupancy [4]. Schmelzle et al. (2016) concluded that eDNA has the potential to be the dominant method for tidewater goby tracking [4]. In another study, Dubreuil et al. (2021) set up baited fish traps for armored catfish at sample sites three times each for 16 hours. eDNA of the species of interest was detected in the water sampled at 18 sites. Alternatively, the fish traps detected catfish at 14 of the 18 sites where positive eDNA detections were made [8]. There were no captures made at sites eDNA did not detect. This study highlights that trapping is variable, as it can depend upon predation, breeding, and food sources [8]. eDNA may be subject to similar variability, however, due to its sensitivity even a species low in numbers can be detected. While trapping non-native species is not as large of a concern in terms of invasiveness, the possibility of capturing at risk species in the area can be avoided with eDNA.
Reaching new limits with low density species detection
eDNA has proven to detect species in a wide range of sampling locations, even where inhabitantance has not previously been verified with traditional capture methods [1, 4]. This is in part because traditional capture methods become all the more difficult as a species population size declines. de Souza et al. (2016) monitored the black warrior waterdog salamander and flattened musk turtle, two at risk species from the upper Black Warrior River basin of Alabama that are impacted by habitat degradation, using eDNA sampling. The low payoff of laborious methods such as dip netting, trapping, and electrofishing for these species makes eDNA a better alternative. The researchers highlighted how sampling with eDNA is most effective when used alongside knowledge of a species’s territorial range and migration patterns, which is also true for trapping/fishing except that the latter takes more time to gather the necessary sample size [5]. eDNA also has the sensitivity to detect low density species, colonization events, and target species against similar ones [1, 3, 4, 9]. For instance, Mauvisseau et al. (2019) found that eDNA was able to differentiate between freshwater pearl mussels and non-target species with two different assays used against other mussel species. Meanwhile, Franklin et al. (2018) had success in identifying smallmouth bass, even when simulations predicted the genetic marker of choice would amplify additional species from the eDNA sample. eDNA allows researchers to determine the presence of a species with low population numbers in order to increase the regions being targeted for conservation.
Detecting non-native species
Non-native species are found either at a low density due to recent colonization, or a high density due to successful adaptation. Both instances need to be monitored to assess plans of action for restoring ecosystem balance [8]. In a notable study, Franklin et al. (2018) detected smallmouth bass to maintain the population of pacific salmon and other native species of the Pacific Northwest. While the spread of smallmouth bass in the US has had economic benefits, close population management is required for habitat stability [9]. Smallmouth bass have been documented to consume 35% of a salmon run, which has negative consequences on salmon migration and local predator/prey interactions. eDNA, with its high sensitivity, was able to detect smallmouth bass successfully against non-target species to find sample locations that require conservation targeting [9]. Similarly, Dubreuil et al. (2021) also tested eDNA by tracking armored catfish who recently started to inhabit rivers in Martinique and compete with gobies for food sources. 22% of sample sites detected the aquatic invasive species (AIS), armored catfish, using water collections. This AIS did not appear to have habitat preferences, such as pH, oxygen level, or temperature. Therefore, this species is more likely to successfully adapt to new territories [8], hence the need for species monitoring.
Current limitations
Despite the promise that eDNA shows, there are potential concerns over its consistency and longevity. For example, de Souza et al. (2016) evaluated the effect of species’ seasonality on eDNA detection probabilities. Seasonality differences may limit the time frame eDNA can effectively be utilized for and introduce a sampling bias. The study discovered that warm and cool seasons played a significant role in the detection of their species of interest. Investigating a species’s migratory and spawning behavior can improve the reliability of eDNA in the same way that traditional sampling methods can be improved [5].
Detection results can also be impacted by eDNA degradation, as validating a species’s presence depends on whether there is recent DNA in the surrounding area [6]. Barnes et al. (2014) found that freshwater amphibian eDNA lasted for over two weeks, while marine fish eDNA only lasted for seven days. There are many possible influences for eDNA degradation, such as sunlight or pH [6], so it is important to consider how genetic degradation may impact detections, or lack of. Brys et al. (2020) also reported eDNA degradation rates of close to one week for the 7 fish and 2 amphibian species the researchers were sampling as their control, and cited Barnes et al. (2014) in how high temperatures, UV radiation etc. may be to blame.
eDNA degradation is also accompanied with dispersion, where detections can only be deemed as reliable when they are repeatable in the area. Boothroyd et al. (2016) observed replicate water samples and determined there was variation among the number of positive detections, despite eDNA being highly sensitive to amplification. The study mentions that a downside of eDNA is that actual organisms are not being detected, so DNA could possibly be due to run-off or the remains of a fish [1]. This implies that DNA detections can vary, which may make eDNA collections a better supplement rather than replacement to traditional capture methods.
The type of aquatic ecosystem and species present can influence detection probabilities in many cases as well. Brys et al. (2020) observed a lentic (standing water) system with high eDNA decay rates and distance limitations for DNA retrieval. Metabarcoding was used to detect various species in a pond, where natural water samples were compared to those taken in proximity to a known variety of locally caged species. The 12S genetic marker detected the known fish and amphibians in the area via DNA sequencing in metabarcoding. Positive detections depend on if a lentic or lotic (rushing water) system is being observed, where the flow of water in a lotic system increases the range of eDNA [3]. Species type and density were shown to impact eDNA dispersal rates as well in this case [3], meaning samples taken in proximity to one another could potentially show different detection results due to spatial limitations.
Conclusion
The goal of this review was to demonstrate how eDNA collection and analysis works, and how it can be useful for aquatic species conservation. Various factors, such as the behavior of a species, population density, and sample regions determine whether eDNA can outperform traditional sampling methods. Methodology is often altered to correspond with these factors, as the different species being monitored and types of aquatic systems being observed impact everything from sampling to analyzing. This implies that whether eDNA can replace traditional capture methods will depend on the study itself and what information researchers are seeking. Dubreuil et al. (2021) had greater success in their study using eDNA compared to trapping, noting that eDNA has the potential to detect species at a greater distance than a trap does. Future research should be aimed towards identifying the susceptibility of eDNA degradation and false detections in different bodies of water. The more eDNA is understood, the simpler it becomes to validate findings and lessen reliance on traditional capture methods.
References:
- Boothroyd M, Mandrak NE, Fox M, Wilson CC. 2016. Environmental DNA (eDNA) detection and habitat occupancy of threatened spotted gar (Lepisosteus oculatus). Aquatic Conserv: Mar. Freshw. Ecosyst. 26: 1107–1119.
- Mauvisseau Q, Burian A, Gibson C, Byrs R, Ramsey A, Sweet M. 2019. Influence of accuracy, repeatability, and detection probability in the reliability of species-specific eDNA based approaches. Scientific Reports. 9:580.
- Brys R, Haegeman A, Halfmaerten D, Neyrinck S, Staelens A, Auwerx J, Ruttink T. 2020. Monitoring of spatiotemporal occupancy patterns of fish and amphibian species in a lentic aquatic system using environmental DNA. Molecular Ecology. 30:3097–3110.
- Schmelzle MC, Kinziger AP. 2016. Using occupancy modeling to compare environmental DNA to traditional field methods for regional-scale monitoring of an endangered aquatic species. Molecular Ecology Resources. 16, 895-908.
- de Souza LS, Godwin JC, Renshaw MA, Larson E. 2016. Environmental DNA (eDNA) detection probability is influenced by seasonal activity of organisms. PLOS ONE.11(10): e0165273.
- Barnes MA, Turner CR, Jerde CL, Renshaw MA, Chadderton WL, Lodge DM. 2014. Environmental Conditions Influence eDNA Persistence in Aquatic Systems. Environmental Science & Technology. 48, 1819−1827.
- Ratcliffe FC, Uren Webster TM, de Leaniz CG. 2020. A drop in the ocean: Monitoring fish communities in spawning areas using environmental DNA. Environmental DNA. 3:43-54.
- Dubreuil T, Baudry T, Mauvisseau Q, Arqué A, Courty C, Delaunay C, Sweet M, Grandjean F. 2021. The development of early monitoring tools to detect aquatic invasive species: eDNA assay development and the case of the armored catfish Hypostomus robinii. Environmental DNA. 4:349–362.
- Franklin TW, Dysthe JC, Rubenson ES, Carim KJ, Olden JD. 2018. A non-invasive sampling method for detecting nonnative smallmouth bass (micropterus dolomieu). Northwest Science. 92(2):149-157.
A New Titan Among Bacteria
By Ethan Feild, Molecular and Medical Microbiology
Author’s Note: I have always been interested in “huge” single celled organisms, like the amoebas living at the bottom of the ocean or even slime molds. When I heard about a single bacteria cell reaching 2 centimeters, I almost couldn’t believe it. This article was originally written for my upper division writing course, to translate specialized knowledge for those who do not have the same experience with a topic that we might. I wrote about T. magnifica because I genuinely believe it to be a deeply interesting and amazing discovery, and I want to convey what makes this new species so interesting to people who don’t have a deep understanding of microbiology. I just had to know how it could grow that big. And now that I’ve learned more about it, I want to teach others that to reach this size, the cell has to break a lot of preconceptions we have about cellular limitations.
What are some of the largest living things you can think of? An elephant? A blue whale? What about the giant sequoias or towering redwoods? There are a lot of “biggest” organisms in the world, but would you ever expect a bacterium to claim that title? Just recently, a new king of giant bacteria was crowned.
Candidatus Thiomargarita magnifica, a bacteria whose cells can stretch up to an impressive two centimeters in length, was just discovered this February through joint research at Lawrence Berkeley National Laboratory and France’s Muséum National d’Histoire Naturelle [1].
“Only two centimeters? That’s as big as a penny!” You may think to yourself. And you’d be right, but this defies all expectations about how big bacteria can grow. Take a more common bacteria like E. coli, for example. They are the “gold standard” for many scientists, especially for bacterial size and morphology. A single cell averages only one micrometer in length (one thousand micrometers fit in a millimeter). By comparison, T. magnifica looks massive, reaching sizes over 20,000 times larger with just a single cell. If E. coli were as big as the average person, T. magnifica would tower over us at 100,000 feet (19 miles) tall, which is high enough to reach supersonic plane altitudes in the stratosphere. But if most bacteria are microscopic, what advantage is there to growing so massive?
Relative to a human-sized E. coli cell, a single Thiomargarita magnifica cell would be over twice as tall as Mount Everest.
Their massive size is a survival strategy for life in the outside environment. Bacteria like E. coli and Salmonella, are small because they are pathogens. When you mostly live inside another organism, staying small means you can reproduce faster, spread easier, avoid capture, and gather all the rich nutrients in your host. Yet outside our bodies live entire microbial ecosystems with members that have to depend on themselves rather than a gracious host for everything, like T. magnifica.
The Thiomargarita genus, who despite the name have nothing to do with margaritas, are part of a larger group called “sulfur bacteria” that “eat” sulfur to get their energy. These bacteria like to live in the mud of streams, ponds, and estuaries where sulfur levels are high and oxygen levels are just right for survival. Unlike other more active bacteria, Thiomargarita do not move around very much. Once they stick somewhere, they live their entire lives on that spot. By growing bigger, they can better handle any potential changes that affect the concentrations of their energy source without needing to relocate from a good spot [2]. This is a sound strategy used by many members of Thiomargarita, including the former size champion Thiomargarita namibiensis which only grows to 750 micrometers in diameter.
But what does T. magnifica do differently to grow so big? Think of a simple bacterial cell like a balloon filled with water, glitter, and a piece of string. The rubber of the balloon is the membrane, the water is the cytoplasm that fills the cell, the glitter floating around inside is the different materials the cell needs to stay alive, and the string is the DNA, the blueprint of the cell. For our balloon cell to stay alive, it needs to move its parts around fast enough to fix damage, take in food, and excrete waste. When the balloon is small those parts can move around easily enough on their own; just by random chance, the pieces of glitter will get where they need to be, and our stringy DNA is more than enough to call all the shots.
As the balloon gets bigger, it takes longer for everything to cross the distance, and the DNA can’t make it to where it is needed. The surface area of our balloon has increased, but the volume inside grows even faster. This is the major problem with diffusion, which many bacteria rely on for their survival [3]. At microscopic sizes it takes less than a second for a molecule to diffuse its way through the cytoplasm. However, it takes oxygen over an hour to diffuse across even a millimeter [4]. And larger, heavier molecules, like life-sustaining sugars, move even slower. Growing too big means there’s too much space to cover, and diffusion alone cannot meet the bacterial cell’s needs to survive.
But diffusion is all that bacteria really have. Bacteria belong to a group called the prokaryotes, whose cells appear relatively simple when compared to our own. The very definition of a prokaryote is a lack of membrane-bound organelles (tiny organs within the cell). This is why most bacteria stick to such small sizes; they physically can’t afford to grow too big without starving themselves.
Yet, some bacteria, like T. magnifica, are determined to prove both physics and biology wrong. Along with other members of the genus, T. magnifica fills most of its body with a large vacuole in order to effectively decrease the volume that needs diffusion. It’s like filling the water balloon with another large balloon, so that everything is pressed into the thin layer between. The cytoplasm is actually only around 3 micrometers thick the whole way around the cell, with the vacuole taking up 73% of its total volume [1]. So even though the cell may be very large, a lot of the important processes take place in the very periphery of the outer membrane in that very thin layer.
That thin layer may hold all the cellular machinery, but it wouldn’t make sense to waste all that extra space, either. So the vacuole also doubles as a storehouse for the various chemicals that the bacteria might need to keep away from their proteins or are simply harder to come across. Important compounds like nitrates can be stored inside and kept at concentrations much higher than the surrounding environment whilst protecting other parts of the cell from harmful side reactions [2]. Recent investigations even indicate that they could use this extra membrane as a way to produce energy, much like how our own cells use mitochondrial membranes to produce most of our energy. This multipurpose structure is one of the reasons why T. magnifica has been able to reach such staggering lengths. It also proves that bacteria are more than the simple balloons of chemicals that we might make them out to be.
Another difficulty with large cell size relates to the cell’s DNA. With smaller cells, it’s easy enough to read off of a single copy for every protein it needs. But in a bigger cell, even with a thin cytoplasm, proteins would take too long to reach their destinations. And this is where T. magnifica really stands out from the rest. Where most bacteria have one copy of their DNA, T. magnifica has many, a situation known as polyploidy.
Polyploidy itself is hardly new. Even we have two copies of each of our 23 chromosomes, and some plants can have dozens of them. The other Thiomargaritas also have multiple copies of their chromosome spread around the cell, but T. magnifica stands out for having over 737,000 copies in a fully grown cell [1]. With one of the highest known numbers of DNA copies, T. magnifica can spread them throughout the body and maintain its impressive size.
Having the most DNA copies is not enough, however. One more unusual trait about T. magnifica is how it stores its DNA. Bacteria usually let their chromosome float around in the cytoplasm, accessed whenever it’s needed. But this giant actually keeps its DNA stored within specialized compartments that its discoverers have adorably named a “Pepin” in analogy with the pips (small seeds) in a watermelon [1]. And the analogy isn’t too far off.
A closeup of T. magnifica’s insides. See how the cytoplasm is only a small part of the volume of the cell, with pepins scattered throughout to organize the DNA and protein factories.
These little pepins keep the DNA more compact and organized, preventing messy diffusion throughout the whole of the cell body. They even discovered that ribosomes, the cellular machinery that turns genetic code into proteins, are also housed within these little organelles. Some eukaryotes have multiple nuclei in their cells, but in bacteria, such a compartment is almost unheard of. The defining characteristic of prokaryotes is the very lack of a nucleus, a membrane compartment where DNA is stored, so what about these pepins? While not exactly the same in structure, it raises questions about just how complex bacteria can be, and where our own nuclei came from.
It is clear that without specialization, it is impossible for a single cell to grow so large. Our own cells, and the cells of every animal, plant, and fungus, house DNA within a nucleus. These pepins within T. magnifica play a similar role in holding DNA all around the cell’s body, which no other bacteria are known to do. It even uses a vacuole to lower its volume and potentially store nutrients like an organelle. If prokaryotes are not supposed to have such structures, then what does that mean for bacteria like T. magnifica? Do all prokaryotes have to be simple?
The more we learn from T. magnifica the more it defies our conventional definition of prokaryotic life. Further study of this amazing oddity could shed light on where eukaryotes like us came from, what the real limits of a cell’s size are, and if bacteria are as simple as we think. Or maybe this is just the tip of the giant bacteria iceberg. Maybe an even larger giant is still waiting to be found, ready to defy our expectations all over again.
References:
- Volland, J.-M., et al. (2022). A centimeter-long bacterium with DNA compartmentalized in membrane-bound organelles. BioRxiv.
- Salman, V., Bailey, J. V., & Teske, A. (2013). Phylogenetic and morphologic complexity of giant sulphur bacteria. A Van Leeuw J Microb. 104(2):169–186.
- Marshall, W. F., et al. (2012). What determines cell size? BMC Biol. 10(1).
- Levin, P. A., & Angert, E. R. (2015). Small but mighty: Cell size and bacteria. CSH Perspect Biol. 7(7).
A War of Multiple Fronts: How to Fight Duchenne
By Alex Neupauer, Genetics and Genomics, ’23
Author’s Note: As a Genetics and Genomics major and a person with Duchenne muscular dystrophy (DMD), I was compelled to write a review on how to alleviate the suffering imposed by this devastating genetic disease. I consulted various scholarly articles and interviewed five experts on DMD. Approaches to improve patient outcomes will involve a mixture of researching cures and improving the care of patients’ symptoms. While we may be far away from truly curing DMD, we are not currently powerless against it. With this article, I hope to promote awareness of DMD among both scientists and the public, as the DMD community relies on outside support to improve patient outcomes.
Abstract
Duchenne muscular dystrophy (DMD) is a devastating disease resulting in muscle degradation. DMD results from mutations to the dystrophin gene that impair the function of the dystrophin protein, an important structural component of muscle tissues. Individuals with the disease suffer from the impacts of weakened muscles on skeletal, circulatory, and pulmonary systems, often shortening their lifespan. The main solution consists of restoring dystrophin expression in cells. CRISPR/cas9 and gene replacement therapy can in theory restore dystrophin expression, while other methods can restore some of its function. To implement CRISPR or gene therapy as cures, researchers must figure out how to deliver them safely and effectively to cells. They must also improve and expand treatments already on the market or on trial to target a wider variety of DMD patients. Exon-skipping therapies only exist for skipping the exons most often implicated in DMD. And nonsense mutation readthrough appears to work better in theory than practice. Implementing and improving methods of dystrophin restoration will also rely on better modeling of DMD and more basic research in molecular and cellular biology. Improved care for the symptoms of DMD is also vital while researchers find and develop cures. To better care for patients’ DMD-related complications, we must create better standards of comprehensive care and facilitate their access to it. Finally, education about DMD is key. Public education will lead to increased funding, allowing for more necessary research. Education will also promote parents’ awareness of screening for DMD and clarify complex care considerations for patients and their families, making quality care easier to secure.
Background
DMD results from the lack of a functional gene encoding dystrophin. Muscle tissues must constantly bear the force of contraction and need proteins like dystrophin to reinforce their strength. Dystrophin ensures muscle fibers and their membranes are not damaged during contraction [1]. Specifically, it links actin filaments inside muscle cells to membrane proteins, which in turn link the membrane to the extracellular matrix [2]. Without dystrophin, actin filaments lack a stable connection to the extracellular matrix, resulting in muscle cell damage.
Dystrophin attaches f-actin to membrane-bound proteins on the muscle cell membrane (sarcolemma). These membrane proteins attach to the extracellular matrix. Without dystrophin, this membrane-bound complex is abolished and f-actin is not anchored to the extracellular matrix. Image from Mbakam et al [1].
The cell damage induces cellular stress, inflammation, and eventual cell death [3], and improper muscle function ensues. As a consequence, individuals with this disease typically become wheelchair-bound in their mid-teens. The lessened use of skeletal muscles results in increased bone fragility and the development of contractures [4] – permanent stiffening of the joints. People with DMD also develop complications in cardiac and respiratory muscles, the most serious complications as a result of the disease [1]. As such, DMD has a negative impact on lifespan and quality of life. And there is no present cure for the disease.
In addition to the lack of a cure, patients often suffer from inadequate care of their symptoms. As the disease impacts patients beginning from a young age, critical care decisions are those of their parents. Despite their best intentions, patients’ families sometimes make choices that negatively impact their outcomes. For example, families refuse steroid treatments, which do not cure DMD but significantly improve patient outcomes. Jessica A. Guzman, an RN for the pediatric neuromuscular clinic at Lucile Packard Children’s Hospital, laments that parents may hold off on steroids out of fear of relatively minor side effects such as shortened stature or in anticipation of a ‘miracle drug’ [5]. Parents may also fail to work through the labyrinth of caring for the health of their child, as the disease impacts many body systems, requiring the attention of many specialist doctors. Scheduling appointments with cardiologists, pulmonologists, orthopedic doctors, and more can be overwhelming. Parents’ uninformed decisions and difficulty navigating the needs of DMD demonstrate the need for increased education on DMD and more support for patients from the medical community.
Stopping the suffering inflicted by DMD has medical, scientific, and social barriers. And it leaves behind it a path of destruction that humanity should not accept. The best approach to end the devastation of DMD consists of researching and implementing methods to restore dystrophin expression to all muscle tissues while improving the care of patients’ symptoms and overall health, employing public education, better models, and basic research to overcome implementation challenges.
Restoring Dystrophin Expression
One of the best solutions is to tackle the problem at the root: the lack of dystrophin protein. There are several approaches to restore the expression of this essential protein: CRISPR/cas9, gene replacement therapy, exon-skipping, and post-transcriptional methods.
Gene Replacement Therapy
One of the most noteworthy challenges in treating or curing DMD relates to the fact that many types of mutations can lead to dysfunction of the dystrophin gene. Patients can have deleted or duplicated exons or point mutations of a single nucleotide [1]. And there is much variability with respect to which exons are affected or which nucleotides are mutated. Gene replacement therapy sidesteps the need for mutation-specific cures by providing all cells with a brand-new copy of the dystrophin gene. Adeno-associated viruses (AAVs) can deliver the replacement dystrophin gene. However, the dystrophin gene is over 2 million bases while the capacity of AAVs is only around 4,700 bases [2]. Micro-dystrophin overcomes this capacity issue, encoding a 1,001 amino acid long protein compared to the 3,684 amino acid long wild-type protein [2]. Carly Siskind, MS, CGC from Stanford Health Care, believes the approach could be helpful, so long as all necessary parts of the protein are encoded by the miniature gene [6]. However, micro-dystrophin does not replace the original protein, which is much longer. In other words, micro-dystrophin appears to solve the problem as a mediocre substitute, which poses the question of whether this mechanism can wholly cure DMD or not.
Furthermore, Claudia Senesac, PT, PhD, PCS from the University of Florida, laments that gene replacement therapy does not appear to last forever, requiring eventual redosing [7]. Unless gene therapy targets satellite cells, the therapy can be lost as muscle cells are replaced. Muscle tissues naturally have turnover rates, which are elevated in the context of gene therapy, as it does not tend to remedy 100 percent of all muscle cells [2]. Satellite cells divide to replace damaged muscle cells, so they must also contain functional dystrophin genes to ensure replaced muscle cells also have a functional copy of the gene. Unfortunately, AAVs do not target satellite cells very effectively [2]. Gene replacement therapy avoids having to consider a patient’s particular mutation, and thus sounds like an excellent solution. However, gene replacement therapy in its current form does not constitute a cure. To make gene therapy a cure, we must develop delivery systems of higher capacity that target a wide array of muscle cell types and create longer-lasting therapies.
CRISPR/cas9
The mechanism of CRISPR/cas9 originates as a bacterial defense against viruses. The CRISPR array locus contains a catalog of sequence fragments from previously infecting viruses. These fragments are known as spacer sequences, which alternate with invariable repeats: spacer, repeat, spacer, repeat, etc. When a novel virus infects bacteria, they sequester fragments of viral DNA and insert them into their CRISPR array locus. When the virus infects again, bacteria transcribe their entire CRISPR array. The mRNA of each spacer sequence is cleaved from the array and loaded onto a cas9 protein. The cas9-RNA complexes containing RNA complementary to segments of the invading viral DNA bind to the viral DNA. Then cas9 cuts at the binding site. Thus, the CRISPR system has high specificity with regard to where it cuts. In the lab, researchers can create their own short RNA sequences called small guide RNA (sgRNA) to guide cas9 to specific sites in DNA.
Mechanism of CRISPR in a bacterium. Viral DNA sequences are saved in the CRISPR array locus and transcribed during future infection to target cas9 to cut viral DNA. Image from “A Tool for Genome Editing: CRISPR-Cas9,” BioRender in collaboration with the Doudna Lab at UC Berkeley.
Researchers have adapted the system to accomplish a variety of functions that may be applicable to curing/treating DMD. Three useful applications are exon knock-in, base editing, and prime editing. Exon knock-in consists of CRISPR cutting the DNA between exons and inserting the missing exon, both restoring lost information and the reading frame [2]. Base editing uses enzymes that chemically convert one base into another. These enzymes are attached to a CRISPR system to guide the modifications to specific nucleotides [1]. Prime editing uses reverse transcriptase, guided by CRISPR, to replace a short segment of DNA with new DNA synthesized from an RNA template [1]. As these templates can contain different, extra, or missing nucleotides relative to the original DNA sequence, prime editing can change, insert, or delete particular nucleotides at particular sites. Taken together, prime and base editing can resolve point mutations. Thus, the diversity of CRISPR editing methods fits well with the diversity of mutations causing DMD. One big issue remains: targeting every single muscle cell with this treatment.
AAVs can deliver the materials necessary for the cell to complete CRISPR editing [2]. However, this delivery mechanism would succumb to the same issues of targeting satellite cells. Lipid nanoparticles may be another possibility for delivery. Together, cas9 (cationic) and sgRNA (anionic) form an anionic ribonucleoprotein (RNP) complex, which is placed inside a cationic lipid nanoparticle [2]. The RNP enters the cell via endocytosis [2], the fusion of the lipid nanoparticle envelope with the phospholipid membrane. It could be a promising avenue given the success of COVID-19 vaccines designed with lipid nanoparticles. Unfortunately, CRISPR also must overcome the challenge of off-target cuts. Permanent changes to DNA can be dangerous if done incorrectly, as the cell’s genome is forever changed. Jacinda Sampson, MD, PhD, adds that CRISPR can now be modified to target mRNA instead [4]. Thus, permanently altering the genome is no longer a concern. Collectively, all experts interviewed express hope of CRISPR as a future treatment option but recognize the aforementioned challenges in achieving that goal [4, 5, 6, 7, 8].
CRISPR gene editing is an exciting possibility for a future cure. To use CRISPR as a cure, we must overcome its delivery challenges and investigate its safety in humans. Additionally, researchers should further investigate RNA CRISPR and lipid nanoparticle delivery as potential cures.
Small Molecules and Exon Skipping
There are other possible methods to restore dystrophin expression that do not involve changing the gene or delivering new copies. Rather, they work with the cell to use the mutated code in a way that allows for the production of dystrophin. One such approach uses small molecules to induce ribosomes to read through a nonsense mutation in dystrophin mRNA. Govardhanagiri et al define small molecules as molecules less than 900 Da, including drugs and biological molecules not including proteins, nucleic acids, or polysaccharides [9]. The fact that cells selectively degrade transcripts with premature stops suggests that premature and normal translation termination may have different mechanisms. Aminoglycosides have been able to promote nonsense mutation readthrough but are toxic after continued use [10]. Pharmaceutical company PTC Therapeutics used high throughput screens to identify nontoxic compounds that promote readthrough of exclusively premature stops in mammalian cells and found ataluren as their top candidate [10].
Chemical structure of ataluren.
While ataluren restored some fully functional dystrophin expression in mouse models, only a small proportion of patients in their study showed increases in dystrophin expression [10]. In theory, the right small molecule drug can allow ribosomes to skip over premature stops to make full length, fully functional dystrophin, halting disease progression. It seems that in practice, this reality does not yet exist. Dr. Sampson hopes that small molecule treatments will improve in the future [4]. However, point mutations like premature stops only represent a fraction of all patients [1].
Another method uses antisense oligonucleotides (AONs) to skip over problematic exons [11]. During splicing, exons must contain an exonic splicing enhancer (ESE) to be retained in the mRNA transcript. ESEs recruit splicing factors that ensure exon inclusion. AONs complementary to an ESE bind to the ESE, abolishing the binding of splicing factors and causing the exon to be spliced out [11]. Exon-skipping may partially restore dystrophin expression in a variety of contexts. Exons could have problematic point mutations that need to be skipped over. Or, an exon deletion could disrupt the reading frame, such that skipping neighboring exons would restore the reading frame. However, exon skipping produces shortened dystrophin, which significantly improves but cannot eliminate the disease phenotype entirely. Furthermore, because each exon is unique, a different treatment must be developed for each specific exon to be skipped. Some already exist on the market, but only for the most commonly problematic exons.
However, AON exon-skipping and nonsense mutation readthrough have already emerged in clinical trials or on the market, making them a more present solution than gene therapy or CRISPR. Nonsense mutation readthrough must improve and exon-skipping treatments must come to include patients with a larger variety of mutations. Exon-skipping remains a good way to mitigate the effects of DMD while more effective treatments are developed. Additionally, if nonsense mutation readthrough therapies become effective, they will be complete cures in their own right.
CRISPR/cas9, gene replacement therapy, and nonsense mutation readthrough can restore fully functional dystrophin, but require much development before patients can use them. Exon-skipping therapies are more readily available but do not wholly cure DMD. Restoring dystrophin attacks the problem at the source, making it an integral part of the fight against DMD.
Improved Care
Even without a present cure, we still have power against the disease. Patients can fight back with quality health care targeting the damage inflicted by DMD. Seeing DMD patients daily as a nurse, Guzman stresses the importance of comprehensive health care:
While research is crucial, we must not forget that while studies are ongoing, [patients] will get weaker. We need to make sure that the patient is well cared for in [their] cardiac function […] There is quite a difference between our patients that have parents championing them with good care and follow up, vs. patients that are not being compliant with care recommendations [5].
Guzman urges that while a cure remains on the minds of most patients and their families, they must not forget to care for the symptoms of DMD in the present. Furthermore, the fact that patients with good care fare much better than those without it demonstrates the importance of quality care. Similarly, Dr. Sampson notes the importance of monitoring patients’ bone and pulmonary health and preventing contractures [4]. Unfortunately, scheduling all these appointments with various specialists is difficult and overwhelming for patients and their families. Additionally, many DMD clinics do not have all specialists present at once [5]. DMD patients need more standardized care programs, which integrate multiple specialists into a single check-up. Patients would benefit from a healthcare liaison who helps them comply with care standards and schedule appointments, especially if clinics cannot schedule all specialists into one appointment.
But care encompasses much more than tending to physical health. If our culture becomes more supportive of folks with disabilities, the outcome for DMD patients will improve. While this article focuses on DMD, we must remember that social justice for folks with disabilities at large remains an important goal to improve the lives of countless people, not just those with DMD. The motivation for cures of serious diseases must originate from a desire to reduce suffering, rather than a desire to remedy resulting inabilities out of the belief that they reduce the value of the people who have them. Neither should we accept the personal ‘shortcomings’ of folks with disabilities as the source of their barriers; social attitudes and infrastructure are the real shortcomings we must resolve. Siskind argues that we “need to have a better society for people who are differently-abled,” filled with improved infrastructure, more considerations, and better mindsets for those using wheelchairs [6]. Alleviating the challenges of DMD in the present includes lessening the burden of wheelchair use. If we ease the daily lives of those with disabilities and provide infrastructure that expands their freedom, their quality of life will improve. Even if supporting the disability-related needs of DMD patients does not directly benefit their physical health, it improves their emotional state, which in turn improves physical well-being. If we look after the whole person – mind, body, and spirit – we will reduce much suffering from DMD right now.
Additional Considerations
Addressing Scientific Gaps in Knowledge
Improved modeling of the disease will help tackle the challenges of implementing new biological discoveries as treatments and better care standards. Human induced pluripotent stem cells (iPSCs) provide a more accurate cellular model in which to study DMD than mouse models. iPSCs are formed by extracting a patient’s cells and converting them to stem cells [2]. From there, they could be turned into muscle cells to study DMD. Vera et al describe one study, which found an herbal compound that reduced oxidative stress in DMD heart cells derived from iPSCs [3]. Thus, iPSCs are already clarifying disease processes and elucidating treatments to alleviate the damage caused by DMD. Artificial intelligence can improve conceptual models of DMD. In particular, natural language processing (NLP), can comb through existing literature to find possible treatment combinations or characterize processes of DMD that humans cannot [3]. Thus, AI could find new information about DMD within existing studies and published work. Even modeling at the stage of clinical trials is important. According to a doctorate of neurology from Stanford University who prefers to remain unnamed, “improved genotype:phenotype information before and after various treatments will provide real world data that will clarify the disease process, response to treatment and cause of unmet needs [8].” More modeling of the disease will reveal more about the processes of and players implicated in DMD. Thus we will be able to better predict what we need to remedy and how diseased muscle cells will respond to a given treatment. When improved models fill gaps in knowledge, solutions to the implementation of treatments become more clear. In the case of iPSCs, patients can even have personalized models, allowing them to receive more targeted treatments. Furthermore, these increases in knowledge will inspire brand new treatment approaches in addition to perfecting current ones for clinical implementation.
Basic research improves the general understanding of biology, eventually filling gaps in knowledge necessary to implement CRISPR, gene replacement therapy, and nonsense mutation read-through as cures. The anonymous researcher expressed that studying molecular and cellular biology, particularly in diseased cells, shows promise in the fight to cure DMD [8]. For example, improved knowledge of cell biology can better illuminate how cells import materials from their surroundings and respond to exogenous DNA. Such knowledge could lead to more effective deliveries to muscle cells and reveal why replacement genes in gene therapy lose effectiveness over time. We must also increase our knowledge specifically on the cell types that will be the target of cures. Even basic research in seemingly distant fields could bring forth unforeseen possibilities. For example, CRISPR/cas9 would not exist without having studied bacterial immunity. And computer science led to AI and NLP, which may one day lead the fight against DMD. Basic research provides the raw materials from which new cures are built.
Public Education and Outreach
Finally, public education will be necessary to both secure proper funding for research and promote screening to limit the disease’s toll. Dr. Senesac calls for public education to increase awareness of the disease [7]. Although the disease is devastating, the public will stand still until they become aware of its existence. As a small community, the DMD community can only make progress with the support of the public. Without knowledge, the public cannot provide support for movements, especially those securing more funding from corporations and governments. Three of five expert respondents directly stated that more money was needed for research [5, 7, 8]. Given that these respondents have direct ties to research, their personal opinion that more funding is needed speaks volumes. A better-educated public also results in parents who are more aware of screening options. Nurse Guzman explains that early screening can “not only identify the child that is affected, and [allow families to] start making interventions early, but can… have an impact on family planning [5].” She recalls that some families discovered their status as carriers too late, having multiple children with DMD. Had they known of their first kid’s diagnosis earlier, they would have not had more kids or screened embryos [5]. Educating parents or prospective parents on screening expands the ways they can look out for the health of their kids. An old adage suggests knowledge is power. In the case of DMD, education is an integral weapon in the fight.
Conclusion
Duchenne muscular dystrophy is responsible for immense suffering, but we have tools with which to fight the disease: research to restore dystrophin expression, improved care, improved modeling, and public education. There are various promising approaches to treatments requiring varying amounts of development to become a reality. For example, patients are already using exon-skipping therapies, while CRISPR requires much more research to be safely applied to humans. In addition to attacking the root cause of DMD, caring for the whole person results in better health and improved quality of life in the face of devastation. Basic research and modeling will close the necessary gaps in biology to develop improved treatments and care regimens. Finally, public education brings awareness to the issue to instill action against DMD, potentially overcoming the issue of funding. We have many tools to fight DMD, and winning the battle requires the use of all of them. Likewise, genetic counselor Siskind remarks that “we are a long way from truly curing DMD, and people with DMD need resources from many different avenues [6].”
In reality, the problem is not DMD alone. The approaches outlined here could apply to other monogenic diseases. But the issue of disease in general extends far beyond science. In addition to being a biological issue, disease is a social, economic, and political issue. We cannot act on any disease without the support of entire societies worldwide. Fighting disease requires many people working together to not only find a cure but to care for those ailing from them in the present. Science is a necessary component of fighting disease, but it is not sufficient. When we fight disease, we fight with our heads and our hearts.
References:
- Mbakam CH, Lamothe G, Tremblay G, Tremblay JP. 2022. CRISPR-Cas9 Gene Therapy for Duchenne Muscular Dystrophy. Neurotherapeutics [Internet]. 19(3):931-941. doi:10.1007/s13311-022-01197-9
- Min Y, Bassel-Duby R, Olson EN. 2019. CRISPR Correction of Duchenne Muscular Dystrophy. Annu Rev Med [Internet]. 70(1):239-255. doi:10.1146/annurev-med-081117-010451
- Vera CD, Zhang A, Pang PD, Wu JC. 2022. Treating Duchenne Muscular Dystrophy: The Promise of Stem Cells, Artificial Intelligence, and Multi-Omics. Front Cardiovasc Med [Internet]. 9:851491. doi:10.3389/fcvm.2022.851491
- Sampson, Jacinda, MD, PhD. Interview conducted by Alex Neupauer. May 13, 2022.
- Guzman, Jessica A., RN. Interview conducted by Alex Neupauer. May 13, 2022.
- Siskind, Carly, MS, CGC. Interview conducted by Alex Neupauer. May 13, 2022.
- Senesac, Claudia, PT, PhD, PCS. Interview conducted by Alex Neupauer. May 13, 2022.
- Anonymous. Interview conducted by Alex Neupauer. May 13, 2022.
- Govardhanagiri S, Bethi S, Nagaraju PG. 2019. “Small Molecules and Pancreatic Cancer Trials and Troubles.” In Breaking Tolerance to Pancreatic Cancer Unresponsiveness to Chemotherapy, edited by Ganji P. Nagaraju, 117-131. Elsevier. doi:10.1016/C2018-0-02682-1
- Peltz SW, Morsy M, Welch EM, Jacobson A. 2013. Ataluren as an agent for therapeutic nonsense suppression. Annu Rev Med [Internet]. 64:407-425. doi:10.1146/annurev-med-120611-144851
- Aartsma-Rus A, van Ommen GJ. 2007. Antisense-mediated exon skipping: a versatile tool with therapeutic and research applications. RNA [Internet]. 13(10): 1609-1624. doi:10.1261/rna.653607
The Gut Microbiome and Obesity
By Lazer Introlegator, Neurobiology, Physiology, and Behavior, ’23
Author’s Note: Ever since I learned about the existence of the microbiome, I have been fascinated. When Dr. Brenda Rinard assigned my UWP102B class (Writing in the Biological Sciences) the task of writing a formal scientific literature review on a topic of our choosing, I knew that the assignment would be the perfect opportunity to delve deeper into understanding the role of the gut microbiome, specifically the gut microbiome’s role in influencing obesity. I chose the topic of obesity specifically because in previous social science classes, I learned about some of the socio-economic reasons for widespread obesity, and I wanted to learn the perspective of the medical community and how science is progressing in its understanding of the gut microbiome’s influence on obesity. My hope is that after reading this review, the reader will be intrigued to learn more about the amazing role the gut microbiome plays in our body.
Introduction
Roughly 30% of people worldwide are overweight or obese [1]. In the past four decades, obesity rates have nearly doubled in over seventy countries [1]. Overweight and obese individuals are at higher risk for various illnesses such as type 2 diabetes and heart disease. A healthy diet and exercise are known methods to combat and prevent obesity [2]. Unfortunately, many people with low income in the United States do not have access to a healthy diet [3].
Another major factor known to play a role in a person’s weight is their gut microbiome. The gut microbiome refers to the trillions of bacteria, fungi, and viruses that live within the guts of animals. These microscopic organisms help their hosts digest food, modify gene expression, fight off foreign infections, and more. Understanding how bacteria within the gut microbiome plays a role in obesity will allow us to manipulate the gut microbiome and combat obesity.
In the growing field of gut microbiome research, scientists have been looking into how the metabolite butyrate, a gut bacteria byproduct, plays a role in obesity. There is conflicting research regarding whether butyrate plays a beneficial or harmful role within the gut microbiome. This review examines how butyrate affects the gut microbiome, the role of the bacterium Lactobacillus sakei (L. sakei) ADM14 in combating obesity, how diet and exercise play an integral role in combating or causing obesity, and whether fecal microbiota transplantation may be used as a therapy to combat obesity.
Altering the gut microbiome
Time Frame for Change
Understanding how long altering the gut microbiome takes is the first step in understanding how to manipulate the gut microbiome. Studies have shown that the gut microbiome can change rapidly even after one day of a changed diet [4]. In one study conducted by David et al., people were fed either a plant- or animal-based diet, resulting in the rapid modification of their gut microbiome. The different diets resulted in different species of bacteria becoming more prominent in the participants’ gut microbiome. David et al. were surprised by how quickly the gut microbiome of each of the participants began to resemble the known gut microbiomes of either herbivores or carnivores [4]. “David et al. theorized that the reason the human gut microbiome can change rapidly based on diet is due to human evolution requiring our ancestors’ bodies to quickly adapt to any food source available [4].” This leads to the possibility that the gut microbiome can be rapidly manipulated. Dietary changes, medications, and probiotics can potentially be used to quickly alter the gut microbiome to combat obesity and improve gut health.
Composition
The gut microbiome has various bacteria that play a role in combating or causing obesity. A study by Gao et al. used 192 people of varying weights to show a correlation between certain gut bacteria and obesity [5]. The study found that in all the participants, Bacteroidetes and Firmicutes were the largest phyla by number of bacteria within the microbiome. All the participants were also found to have a similar total number of bacteria present in their gut. However, the ratio of the different types of bacteria in each group differed. Obese participants were found to have the least diverse microbiomes; they had a greater amount of the phylum Proteobacteria present than non-obese participants [5]. At the genus level, the study found proportional differences amongst the groups of bacteria present in their gut microbiomes. The Pseudomonas and Fusobacterium were two genera found in small proportion amongst participants at a healthy weight and in much larger abundance amongst the obese participants. The absence of these genera was theorized to be associated with a healthy weight [5].
A large study by Peters et al. identified which bacteria need more evidence to show their connection to obesity [6]. This study researched whether the diversity and composition of the human gut microbiome in American adults could be associated with obesity. The study analyzed the microbiome composition of fecal samples from 599 individuals split into two separate populations. 451 people were in Minnesota and were between the ages of 50 – 75, and 176 people were in New York and were between the ages of 29 – 86. The study verified that the gut microbiome of obese people compared to people at a healthy weight differed in composition. Similar to the study conducted by Gao et al., the study by Peters et al. verified that the phyla of Bacteroidetes and Firmicutes were the largest groups of bacteria within the microbiome, but they did not find a correlation between obesity and the Firmicutes to Bacteroidetes phyla ratio [6].
However, Peters et al. did find a correlation between certain classes of bacteria in the Firmicutes phylum and obesity [6]. The classes of bacteria associated with obesity were the Bacilli class and the Clostridia class. Obese individuals were found to have higher levels of the Bacilli class and lower levels of the Clostridia class compared to the healthy weight individuals. Overweight people showed similar trends in their gut microbiome as the obese individuals. Additionally, the bacteria that were found in lower abundance in obese people and higher abundance in healthy weight individuals were found to be associated with the beneficial butyrate metabolism (decreasing the amount of butyrate excreted in fecal matter) [6]. This means obese people in the study were less likely to benefit from butyrate metabolism. Butyrate metabolism promotes increased gut health and potentially plays a role in preventing and combating obesity. While Peters et al. identified bacteria that may play a role in causing or preventing obesity in humans, the way these bacteria prevent or cause obesity is not yet understood.
The Role of Butyrate
The byproducts of the bacteria in the gut microbiome can play a role in obesity as well. One specific type of bacteria product, or metabolite, is the short-chain fatty acids (SCFAs). Many obesity studies have focused on an SCFA called butyrate. Nandy et al. researched whether there was a correlation between the metabolites of the various bacteria in the gut microbiome and obesity [7]. This study analyzed stool samples from 170 two-year-old children. To best determine which metabolites were involved in weight determination, Nandy et al. tracked 20 independent variables that could influence the outcome (such as maternal smoking and diet). The only metabolite that showed a significant influence on weight was the SCFA butyrate. The children who were overweight or obese showed much higher levels of butyrate in their stool samples compared to the children that were of healthy weight or underweight. This study is important because it provides insight into how butyrate may play a role in child weight gain [7]. This research provides a foundation for future studies investigating the influence other metabolites may have on human health. Additionally, this study can help the scientific and medical communities combat obesity by predicting which children may become obese later in life and implementing preventive measures.
Another study by De la Cuesta-Zuluaga et al. investigated whether there was any correlation between the amount of fecal SCFAs to either gut structure or obesity [8]. De la Cuesta-Zulanga et al. used fecal samples from 441 volunteers to determine their microbiome composition and SCFA concentration. Additional data was collected from each volunteer detailing their daily diet and exercise regimen. This research showed that those with a higher butyrate concentration in their fecal samples were more likely to be male, have a lot of fiber in their diet, and be obese. High butyrate concentrations, along with other SCFAs, also showed a correlation with having a less diverse gut microbiome. The study also found that high butyrate concentrations were shown to have a correlation with poor gut structure [8]. Poor gut structure in this context refers to gut permeability, also known as leaky gut. “Leaky gut occurs when pathogens are not successfully prevented from entering the gut microbiome where they can then cause an imbalance of bacteria in the gut microbiome.” De la Cuesta-Zuluaga et al. found that volunteers with low fecal SCFA concentrations were more likely to be female and at a healthy weight. The study also showed that lower butyrate levels were associated with a higher level of Bacteroides in the human gut microbiome [8]. The mechanisms of butyrate within the gut are not yet understood and require further research.
Therapies
Probiotics
Ingesting live bacteria (probiotics) is a potential therapy to alter the gut microbiome. Won et al. used 24 mice to test the potential of the probiotic Lactobacillus sakei (L. sakei) ADM14 in preventing obesity [9]. Won et al. isolated the L. sakei ADM14 from kimchi, a Korean dish made from fermented vegetables. The bacterium was then grown in a lab to be fed to certain mice over a ten-week period. Four groups of six mice each were made at random. Two groups fed a normal diet were labeled ND and NDA and two groups fed a high-fat diet were labeled HD and HDA. Both the NDA group and the HDA group had the L. sakei ADM14 included in their diet. Around five weeks into the experiment, the HDA group began to weigh significantly less than the HD group. By the end of the experiment, Won et al. calculated that the HD group had gained over 25% more weight than the HDA group. Additionally, the HD group had a significant increase in blood cholesterol and blood glucose levels compared to the ND and HDA groups [9].
Won et al. also found that the ratio of Bacteroidetes (a group of bacteria that assist in the breakdown of carbohydrates and certain sugars) was much lower in the HD group than the other groups [9]. Bacteroidetes produce short-chain fatty acids (SCFAs) which greatly help maintain gut health and are commonly found in the microbiome of lean individuals. The HD group was found to be lacking in SCFAs. The NDA group showed similar findings to the ND group. Won et al. took these findings to suggest that L. sakei ADM14 only plays a role when there is a high-fat intake occurring in an individual [9]. The implications of this study are that L. sakei ADM14 has an impact on the microbiome and obesity. This study suggests that L. sakei ADM14 can encourage healthy bacterial growth, and prevent weight gain from a high-fat diet in mice. Won et al. believe further studies on L. sakei ADM14 can provide the medical field with probiotics that can help combat obesity [9]. More studies with larger data sets must be done to further understand how L. sakei ADM14 affects the gut microbiome in humans.
Diet and Exercise
A healthy diet and physical activity may be enough to maintain a healthy gut microbiome and prevent obesity [2,10]. A study conducted by Wang et al. with mice tested whether moderate treadmill exercise had an effect on the gut microbiome and barrier [2]. The gut barrier prevents pathogens from entering the gut microbiome and causing an increase in detrimental bacteria and a decrease in beneficial bacteria. Wang et al. randomly divided 24 mice into four groups of six. The first group (SD + Sed) was fed a standard diet with no exercise. The second group (SD + Exe) was fed a standard diet combined with exercise. The third group (HFD + Sed) was fed a high-fat diet with no exercise. The fourth group (HFD + Exe) was fed a high-fat diet combined with exercise. The exercise groups ran on a treadmill for 45 minutes, five days a week, for twelve consecutive weeks. Researchers collected blood, fecal matter, and bodyweight levels each week. The two groups fed a standard diet had relatively low weights during the twelve weeks. The HFD + Sed saw the mice double their body weight. Wang et al. found that the high-fat diet with exercise group had a more diverse gut microbiome than the high-fat diet group with no exercise [2].
Both the exercise groups were found to have altered their microbial structure compared to the non-exercise groups. Exercise was also found to increase the amount of Verrucomicrobia in the gut microbiome which a high-fat diet typically reduces [2]. This research also found that exercise helped bring bacteria to a healthy concentration even though the high-fat diet caused an imbalance in the microbiome. An additional finding was that a high-fat diet caused the gut to have less structural stability and an increased chance of foreign bacteria and pathogens invading the gut microbiome. The study also found that the high-fat diet with exercise group showed much more order and structural integrity of their guts compared to the high-fat diet group with no exercise [2]. This study suggests that those with unhealthy diets can counteract some of the negative influences caused by a high-fat diet by beneficially altering their gut microbiome through exercise
Fecal Microbiota Transplantation
A study by Lai et al. gave mice fecal microbiota transplantations (FMT) [10]. FMT is a process where fecal matter with the desired gut bacteria is collected from one individual and put into another, typically via colonoscopy or enema. Lai et al. wanted to see if and how FMT from mice with a healthy gut microbiome to obese mice would affect the gut microbiome of the recipient obese mice [10]. The study took 47 five-week-old mice and randomly split them up into seven groups that differed in diet and exercise regimen. Antibiotics were given to the groups undergoing FMT for several days before receiving FMT to prepare them for the gut microbiome recolonization process. The recipients were given the FMT 5 days a week from week 12 until week 24 [10]. After two of the non-exercise groups received FMT from groups that did exercise, their gut microbiomes became similar to the group which had a high-fat diet and exercised regularly. Lai et al. understood this finding to imply that the recipient’s diet affects which microbes can colonize their gut microbiome after FMT [10]. This study is important because it shows that FMT can transfer the benefits of diet and exercise to mice that have poor diet and no exercise. This study complements previous studies that show that diet and exercise can beneficially alter gut microbiome composition [10]. Using the results of this study for future studies and eventually human trials can potentially help combat and prevent obesity in humans. However, the methods of FMT performed in this mouse trial are not practical in human trials yet, so future human trials will need to make adjustments to the FMT procedure.
Conclusion
Obesity is a serious health concern for people worldwide. Certain bacteria are associated with obesity and gut health. The role of bacteria and their products are yet to be fully understood, but methods of altering the gut microbiome have come a long way. Therapies are being developed such as exercise, improved diet, probiotics, and fecal microbiota transplantation to combat obesity. “Although a healthy diet is shown to have the most impact on maintaining a healthy gut microbiome, exercise is currently the most accessible and beneficial therapy in maintaining gut health and combating obesity.” The mechanism by which bacteria and their products affect obesity and the gut must be better understood, and therapy trials must move from mice to humans.
References:
- Afshin A, Forouzanfar MH, Reitsma MB, Sur P, Estep K, Lee A, Marczak L, Mokdad AH, Moradi-Lakeh M, Naghavi M, et al. 2017. Health Effects of Overweight and Obesity in 195 Countries over 25 Years. New England Journal of Medicine. 377(1):13–27.
- Wang J, Zhang Q, Xia J, Sun H. 2022. Moderate Treadmill Exercise Modulates Gut Microbiota and Improves Intestinal Barrier in High-Fat-Diet-Induced Obese Mice via the AMPK/CDX2 Signaling Pathway. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy. 15:209–223.
- Edin K, H Luke Shaefer. 2016. $2.00 a day: living on almost nothing in America. Boston: Mariner Books.
- David LA, Maurice CF, Carmody RN, Gootenberg DB, Button JE, Wolfe BE, Ling AV, Devlin AS, Varma Y, Fischbach MA, et al. 2013. Diet rapidly and reproducibly alters the human gut microbiome. Nature. 505(7484):559–563.
- Gao R, Zhu C, Li H, Yin M, Pan C, Huang L, Kong C, Wang X, Zhang Y, Qu S, et al. 2017. Dysbiosis Signatures of Gut Microbiota Along the Sequence from Healthy, Young Patients to Those with Overweight and Obesity. Obesity. 26(2):351–361.
- Peters BA, Shapiro JA, Church TR, Miller G, Trinh-Shevrin C, Yuen E, Friedlander C, Hayes RB, Ahn J. 2018. A taxonomic signature of obesity in a large study of American adults. Scientific Reports. 8(1):9749.
- Nandy D, Craig SJC, Cai J, Tian Y, Paul IM, Savage JS, Marini ME, Hohman EE, Reimherr ML, Patterson AD, et al. 2021. Metabolomic profiling of stool of two‐year old children from the INSIGHT study reveals links between butyrate and child weight outcomes. Pediatric Obesity. 17(1).
- De la Cuesta-Zuluaga J, Mueller N, Álvarez-Quintero R, Velásquez-Mejía E, Sierra J, Corrales-Agudelo V, Carmona J, Abad J, Escobar J. 2018. Higher Fecal Short-Chain Fatty Acid Levels Are Associated with Gut Microbiome Dysbiosis, Obesity, Hypertension and Cardiometabolic Disease Risk Factors. Nutrients. 11(1):51.
- Won S-M, Chen S, Lee SY, Lee KE, Park KW, Yoon J-H. 2020. Lactobacillus sakei ADM14 Induces Anti-Obesity Effects and Changes in Gut Microbiome in High-Fat Diet-Induced Obese Mice. Nutrients. 12(12):3703.
- Lai Z-L, Tseng C-H, Ho HJ, Cheung CKY, Lin J-Y, Chen Y-J, Cheng F-C, Hsu Y-C, Lin J-T, El-Omar EM, et al. 2018. Fecal microbiota transplantation confers beneficial metabolic effects of diet and exercise on diet-induced obese mice. Scientific Reports. 8(1).
The Effects of Ozone on Plant-Pollinator Interactions
By Hanna Francis, Biological Sciences ’22
Author’s Note: I grew interested in plants through botany and plant biochemistry courses at UC Davis and learned about insects while volunteering at the Bohart Museum of Entomology on campus. After taking a course about the toxicology of air pollutants, which focused primarily on human health outcomes, I began to wonder how air pollution affects organisms other than ourselves. In investigating current research at the intersection of these three things– plants, insects, and air pollution– I found that plant-pollinator interactions are affected by ozone pollution in a variety of ways, some of which may be harmful to agriculture and biodiversity. I hope that readers will become aware of just how complex the effects of pollution are on the natural world. The paper may at first seem bleak, however, there must always be hope. I truly believe that research like this will help inform regulatory standards for air pollution that will protect both humans and our environment from damage in the future.
Introduction:
The concentration of ozone (O3) in the earth’s atmosphere has increased dramatically due to industrialization and is predicted to increase by two- to four-fold over the next twenty years [1,2]. Human activity, especially in the form of transportation and manufacturing, emits pollutants such as nitrogen oxides (NOx), which react with oxygen and UV light in the atmosphere to form ozone [3,4]. Ozone is a highly reactive gas known to lead to human lung diseases such as asthma [5], but its effects on other organisms are not as well understood.
Pollination– a process vital to agriculture– is reliant upon insects and plants, so it is crucial to understand the effects of ozone pollution on this process [6]. Insect pollination is necessary for over 80% of global crops [6]. In addition to agricultural plants, wild plant species have also been shown to be sensitive to ozone and studies have found ozone to be associated with a loss in wild species diversity [7]. Recent research has focused on the ways in which ozone pollution may interrupt plant-insect interactions during pollination. This review aims to explain and categorize current knowledge about the various ways in which increased levels of ozone in the atmosphere may impact the process of pollination. I focused on studies which investigated the effects of ozone exposure on flowering plants or insects during pollination. It is clear that exposure to ozone in insect pollinators can lead to changes in mobility [2], behavior [1,2], and perception of volatile organic compounds, which plants release for signaling and communication [8]. There is also evidence that when plants are exposed to ozone, their reproductive performance [4], visual features [3], and signaling compounds are affected [9,10]. However, it is unclear if ozone pollution will have a significant enough impact on pollination to negatively impact agricultural yields and the survival of wild plants.
Ozone Exposure to Plants:
Effects on Plant Visual Traits
There are a variety of effects of ozone on the visual features of plants that may impact pollination, including anatomical and chemical changes [3]. One recent study exposed
an endangered species of alpine geranium called Erodium paularense to ozone. The scientists measured the petal area and the spectral reflectance, which is the ratio of the amount of light reflected off of a surface to the amount of light that hit the surface [3]. Using the spectral reflectance, an anthocyanin reflectance index was calculated, which estimates the content of anthocyanins (blue, purple and red pigments) in each petal. Changes in anthocyanin content would change the perceived color of the flower to pollinators [3]. The study found no significant difference in petal area after ozone fumigation, but ozone exposure did increase petal reflectance. Plants that were exposed to ambient levels of O3 had a significantly higher anthocyanin reflectance index than other treatments. The differing reflectances mean that some insects, such as flies, would see the O3-treated plants as more blue than before, though the effect was not large enough for there to be any difference in the perceived color for butterflies and bees, based on models [3]. This study demonstrates that ozone has a measurable effect on the visual attraction of flowers, which has implications for pollination efficiency in agriculture as well as for endangered wild plants like Erodium paularense.
Effects on Volatile Organic Compounds
Ozone has a strong potential to react with and disrupt VOCs, which plants release for signaling and communication [8]. Plants use various VOCs in order to attract insect pollinators, so any reactions these compounds may undergo in the atmosphere could affect the ability of insects to locate and pollinate plants [8]. A study tested how different concentrations of ozone affect the floral scent of black mustard plants, Brassica nigra, and how the scent changes affect bumblebee behavior [9]. Flowers were placed into a system of tubes in which the floral compounds were exposed to three different concentrations of O3. The floral emissions were then collected at four different distances from the flower. Bees were given the choice between two tubes that visually mimicked mustard inflorescences, each containing either clean air or the VOCs of varying O3 exposure and collection distance [9]. The study found that the ozone concentration negatively affected the concentration of VOCs in the glass tube. In the behavioral experiments, bees showed a preference for ozone-exposed floral scents at shorter distances rather than longer distances. The results indicate that ozone concentration, ozone exposure, and distance from the scent source all affected bumblebee behavior [9]. Another study conducted research on the degradation of VOCs in which rapeseed (Brassica napus) flowers were exposed to ozone. Measurements of VOCs were taken at several timepoints after O3 fumigation [10]. The study found that ozone exposure led to a decrease in monoterpenes and sesquiterpenes (VOCs that contain alkenes) and an increase in many oxygenated species, which are formed when VOCs react with ozone. It was concluded that because monoterpenes and sesquiterpenes are very important in plant-insect signaling, there could be a significant negative impact on this signaling during pollution episodes [10]. Together, these studies demonstrate how ozone negatively affects chemical compounds released from plants and that these changes affect the behavior of pollinators.
Figure 1. Floral compounds were exposed to varying concentrations of ozone and then collected at one of four different distances from the ozone source (ozone generator): Distance 0, 1, 2, and 3.
Figure 2. Bees were given the choice between two tubes that visually mimicked mustard inflorescences (shown in green and yellow). The tubes contained either clean air or VOCs of varying ozone exposure and collection distance.
Effects on Plant Reproduction
The impacts of O3 exposure on the visual and chemical traits of plants may lead to differences in plant reproductive performance. One study investigates whether plant age at the time of exposure impacts reproduction [4]. After wild mustard plants (Sinapis arvensis) were exposed to heightened ozone concentrations at various stages of life, plant reproduction was quantified by measuring the number of seeds produced, seed weight, number of total fruits, and number of visitations by pollinator insects. The study found that the plants were affected differently by O3 depending on the age at which they were exposed. Younger plants (3 and 4 weeks old) tended to have increased reproductive performance (more fruits, more seeds, and higher seed weight than control). In contrast, plants that were exposed later in life had reduced reproductive performance, though this effect was only significant in the total seed weight in plants exposed at 5 weeks old. Overall, the study concluded that younger plants respond to ozone treatment by initiating higher investment in flower production and reproduction, whereas older plants are not as flexible and have decreased reproductive capabilities due to stress response [4]. Understanding how pollution episodes at different life stages could change crop yield is of great interest to agriculture, and may impact crop management strategies in the future.
Ozone Exposure to Insects:
Effects on Insect Motility and Behavior
Research has demonstrated that insects that have been exposed to ozone show negative health effects and changes in behavior, which may impact their ability to successfully pollinate host plants [1,2]. In one recent study, fig wasps were exposed to various O3 concentrations in fumigation chambers and their behavior was recorded at different time intervals afterward [2]. There was a significant deviation in motility behavior after exposure to O3. Specifically, there was a significant decrease in activity from control in wasps exposed to high levels of ozone (120 and 200 ppb) [2]. In a similar study, fig wasps were exposed to differing concentrations of O3 and given a choice between two stems of a Y-tube: one with a fresh air control and one with a mixture of VOCs that mimicked those of the wasps’ host plant [1]. Fig wasps that had been exposed to lower concentrations (80 ppb) of ozone significantly preferred the VOCs, those that had been exposed to an intermediate concentration (120 ppb) showed no preference, and the wasps that had been exposed to the highest concentration (200 ppb) of ozone had a significant preference for the clean air over the VOCs [1]. These results suggest that the higher levels of ozone impaired the wasps’ ability to distinguish VOCs from clean air, or made the VOCs no longer attractive to the insects. Decreased mobility and ability to locate host plants could dramatically reduce the effectiveness of pollinators.
Effects on Insect Perception of VOCs
There is potential for biochemical changes to occur within insects that could alter their sensing abilities. One study exposed fig wasps and bumblebees to varying O3 concentrations and lengths of time in fumigation chambers [1]. After exposure, electroantennograms (EAGs), a method of measuring the electrical response of an antenna to a scent [1], were measured in order to evaluate the insects’ sensitivity to different VOCs. There was a clear negative relationship between O3 exposure and the ability of pollinators to respond to and detect volatile organic compounds. However, many results were not significant and the responses depended on factors including species, the specific VOC, ozone concentration, and exposure time [1]. Further effects of ozone on insect perception of plant compounds were examined in a study which used western honeybees, a species that is important to agriculture worldwide [11]. The honeybee antennae were mounted into an airflow system in which they were stimulated with different sequences of control paraffin, O3, and three VOCs [8]. Responses were measured using EAG recording devices. One of the three compounds, (Z)-3-hexenyl acetate, showed a significant decrease in antennal responses in the O3 treatment group compared to the control group [8]. These results show that ozone can cause significant effects on antennal response to some VOCs. Therefore it may be worthwhile to test the effects of various ozone concentrations on other VOCs beyond those that were tested in this study. Impaired responses to plant compounds could potentially be detrimental to pollination, which could have far-reaching impacts for wild plant populations as well as for agriculture.
Conclusion:
Understanding how and to what extent ozone impacts plant-insect pollination interactions are active areas of research. Further experimentation will help to solidify understanding of the severity of the threat of ozone to agricultural systems and wild plants. Future directions of research include the difference between short-term high-concentration ozone pollution episodes versus long-term exposure to ambient ozone levels. The mechanisms of action of ozone interference with plant and insect biochemistry are also largely unknown [12]. Despite remaining questions, it is clear that there are many ways in which pollination is negatively affected by ozone exposure. Some of these effects are nearly imperceptible even at unrealistically high experimental concentrations of ozone [4,8], meaning that they may not be relevant to real-world scenarios. For example, one study used concentrations of 1000ppb, despite current ozone concentrations not usually exceeding 100ppb [4]. Nevertheless, ozone is a threat to both agriculture and to the biodiversity of wild plants [3].
In addition to further research, there are ways in which current knowledge can already be applied. Data from studies such as the ones included in this review can inform air pollution standards [3]. Reducing NOx emissions, which are primarily anthropogenic in source (about half come from fossil fuel combustion, and a further 20% from other combustion done by humans) would be beneficial in reducing or at least maintaining ozone levels [13].
References:
- Vanderplanck M, Lapeyre B, Brondani M, Opsommer M, Dufay M, Hossaert-McKey M, Proffit M. 2021. Ozone Pollution Alters Olfaction and Behavior of Pollinators. Antioxidants [Internet]. 10(5):636. doi: 10.3390/antiox10050636
- Vanderplanck M, Lapeyre B, Lucas S, Proffit M. 2021. Ozone Induces Distress Behaviors in Fig Wasps with a Reduced Chance of Recovery. Insects [Internet]. 12(11):995. doi: 10.3390/insects12110995
- Prieto-Benítez S, Ruiz-Checa R, Bermejo-Bermejo V, Gonzalez-Fernandez I. 2021. The Effects of Ozone on Visual Attraction Traits of Erodium paularense (Geraniaceae) Flowers: Modelled Perception by Insect Pollinators. Plants [Internet]. 10(12):2750. doi: 10.3390/plants10122750
- Duque L, Poelman EH, Steffan-Dewenter I. 2021. Plant age at the time of ozone exposure affects flowering patterns, biotic interactions and reproduction of wild mustard. Sci Rep-UK [Internet]. 11(1). Article number: 23448 (2021)
- Uysal N, Schapira RM. 2003. Effects of ozone on lung function and lung diseases. Curr Opin Pulm Med [Internet]. 9(2):144–150. doi: 10.4046/trd.2020.0154
- Aizen MA, Aguiar S, Biesmeijer JC, Garibaldi LA, Inouye DW, Jung C, Martins DJ, Medel R, Morales CL, Ngo H. 2019. Global agricultural productivity is threatened by increasing pollinator dependence without a parallel increase in crop diversification. Glob Change Biol [Internet]. 25(10):3516–3527. doi: https://doi.org/10.1111/gcb.14736
- Davison, A. W., & Barnes, J. D. (1998). Effects of Ozone on Wild Plants. New Phytol [Internet], 139(1), 135–151. doi: https://doi.org/10.1046/j.1469-8137.1998.00177.x
- Dötterl S, Vater M, Rupp T, Held A. 2016. Ozone Differentially Affects Perception of Plant Volatiles in Western Honey Bees. J Chem Ecol [Internet]. 42(6):486–489. doi: 10.1007/s10886-016-0717-8
- Farré-Armengol G, Peñuelas J, Li T, Yli-Pirilä P, Filella I, Llusia J, Blande JD. 2015. Ozone degrades floral scent and reduces pollinator attraction to flowers. New Phytol [Internet]. 209(1):152–160. doi: https://doi.org/10.1111/nph.13620
- Acton WJF, Jud W, Ghirardo A, Wohlfahrt G, Hewitt CN, Taylor JE, Hansel A. 2018. The effect of ozone fumigation on the biogenic volatile organic compounds (BVOCs) emitted from Brassica napus above- and below-ground. Plos One [Internet]. 13(12):e0208825. doi: https://doi.org/10.1371/journal.pone.0208825
- Toni CH, Djossa BA, Yedomonhan H, Zannou ET, Mensah GA. 2018. Western honey bee management for crop pollination. African Crop Science Journal [Internet]. 26(1):1. doi: 10.4314/acsj.v26i1.1
- Pinto DM, Blande JD, Souza SR, Nerg A-M, Holopainen JK. 2010. Plant Volatile Organic Compounds (VOCs) in Ozone (O3) Polluted Atmospheres: The Ecological Effects. J Chem Ecol [Internet]. 36(1):22–34. doi: 10.1007/s10886-009-9732-3
- Olivier JGJ, Bouwman AF, Van der Hoek KW, Berdowski JJM. 1998. Global air emission inventories for anthropogenic sources of NOx, NH3 and N2O in 1990. Environ Pollut [Internet]. 102(1):135–148. doi: https://doi.org/10.1016/S0269-7491(98)80026-2
A Review of Recent Research into Remote Control of Stem Cell Differentiation through Light
By Jacob Pawlak, Biochemistry and Molecular Biology ’23
Author’s Note: I wrote this piece to bring attention to the exciting new field of research being conducted primarily in China that aims to control the differentiation of stem cells by irradiating them with different wavelengths of light. This non-invasive method is potentially of great value to those working in regenerative medicine and has a strong foundation of research for future exploration. I hope to introduce this fascinating concept to future researchers to pique their interest in the field.
Introduction
Regenerative medicine is the use of human bodily mechanisms to restore functionality, cure diseases, and heal injuries. In particular, this field’s research is primarily concerned with healing previously untreatable injuries to the nervous system and regrowing musculoskeletal tissue from the aftereffects of traumatic injury or disease. The body is incapable of repairing traumatic injury to the brain or spinal cord completely, and regenerative medicine aims to use stem cells as a source of new tissue. This primarily consists of manipulating stem cells into specific cell types at the injury sites. These injuries are not particularly uncommon; in just the US, an average of 17,000 people a year are hospitalized for spinal cord damage [1].
Within the field of regenerative medicine, remote control of stem cell differentiation is one of its most promising areas of research, as it avoids the major issue of complications arising from invasive procedures, such as a risk for infection or an autoimmune response. Non-invasiveness is typically achieved through the use of near-infrared light, which is capable of penetrating tissue layers to reach target sites without harming normal cells in its path [2]. Invasive procedures require large surgical teams and expose the patient to potential infection. Instead, non-invasive procedures can be done as part of outpatient care, not requiring lengthy hospital visits. Photobiomodulation has recently emerged as one of the most promising candidates for remote control of stem cell differentiation. Photobiomodulation is the act of exposing cells to specific wavelengths of light to influence gene expression and shift cellular processes towards a specific target, such as increased differentiation into a target cell type or increased proliferation of the stem cells [2]. This often takes the form of the exposed light influencing crucial biochemical pathways in the cells, “pushing” the cells towards the researcher’s target cell type or towards increased replication. Upon reaching the site, the photons can influence the cell’s genetic expression on its own, or be converted into other influential wavelengths by novel optical devices we will go on to describe in this review [2].
There are two primary cell types targeted by researchers in regenerative medicine. The first are neural stem cells, which are directed to differentiate into astrocytes, cells that act to regulate blood flow and repair the nervous system following infections and injuries [3]. The second target for control are mesenchymal stem cells, which are found throughout the body and are capable of differentiating into a wide variety of musculoskeletal cell types such as bone, muscle, and cartilage [4]. These two cell types form the backbone of photobiomodulation research due to their potential use in regenerating complex, irreparably damaged organs such as the spinal cord.
This review presents the findings of recent research into the remote control of human stem cell differentiation. While these researchers have not worked with in vivo cells, they have laid the groundwork for a wide range of potential exploration routes for further research into stem cell differentiation control via photobiomodulation and novel optical methods.
Types of Stem Cells Researched
The two primary types of stem cells being researched for remote control are mesenchymal stem cells and neural stem cells. Mesenchymal stem cells are typically selected to create bone cell cultures, while neural stem cells are directed towards forming glial cells, which support the nervous system by forming sheaths around neural pathways and regulating blood flow [3,4].
Mesenchymal stem cells have been the primary focus of novel research, due to the relative ease of acquiring human-adipose derived stem cells (hADSC). These cells are extracted from human fat tissue and are capable of differentiation into multiple cell types. Most papers have focused on increasing the proliferation of these cells alongside increasing their differentiation into osteoblasts [7-9]. These cells are primarily useful in regenerative medicine for application to traumatic injuries to the musculoskeletal system.
In both types of cells, the light triggers photoreceptor complexes that are sensitive to the upper and lower bound of visible light wavelengths, or red and blue light [5-9]. These complexes induce the cellular modifications that lead to the changes in the stem cell’s rates of proliferation and differentiation. Thus, research focuses primarily on only red, blue, and occasionally green light, as stem cells are not uniquely reactive to more moderate colors on the visible light spectrum due to lack of sufficient sensitivity.
Novel Non-Invasive Methods – Photobiomodulation
Typically in photobiomodulation research, LED diodes are placed over cell cultures to irradiate them at specific light wavelengths for approximately 60 minutes daily over the course of 5-10 days [5-8]. Once this period is complete, researchers examine the cell cultures for signs that the cells have differentiated, such as the release of signature proteins into the culture medium or through visual inspection with a microscope [5-8].
In 2019 Wang et al. was successful in multiplying the proliferation of neural stem cells by 4.3x, and their differentiation rate into astrocytes by 2.7x through treatment with low-power blue light irradiation [5]. Proliferation measures the rate of population increase of cells, while differentiation measures how many of the stem cells develop into specialized cells. A newer study in 2021 by Yoon et al. found an increase in astrocyte proliferation through red-light treatment [6]. Notably, Yoon was able to find that red light could influence astrocyte proliferation without affecting other cells in the area, demonstrating the light’s effects in an environment closer to the human body, unlike Wang’s work on isolated astrocyte cultures. These papers making use of different wavelengths of light for different situations speaks to the versatility of photobiomodulation as a method of controlling astrocyte populations.
Comparing Light Wavelengths
While some innovative work has been done with novel optical devices, most research deals with the cheaper and simpler direct application of visible light. Visible light has been applied to both neural and mesenchymal stem cells to observe their reactions and to find wavelengths that can control differentiation and proliferation of these cells.
The application of visible light can be broadly split into two categories: red and blue. Exclusively red-light wavelengths have been found to increase the proliferation of both mesenchymal and neural stem cells [6, 8-9]. Meanwhile, blue light has mixed effects. On neural stem cells, it increases proliferation and differentiation into astrocytes, whereas on mesenchymal stem cells it has been found to lower proliferation while raising the rate of differentiation into osteoblasts [5, 7-8]. Some preliminary work has been done on green light, which has been found to cause the same effects as blue light on mesenchymal stem cells, due to the two colors’ close proximity on the visible light spectrum [7].
The most promising results come from the work of Crous et al., whose team found that they could increase both differentiation and proliferation in mesenchymal stem cells by alternating red and green light irradiation, synthesizing the effects of the two wavelengths [9]. In combination with Wang Y et al.’s work on mesenchymal stem cells with single wavelength application, this suggests that the inhibitory effects of green light on proliferation are less potent than the enhancing effects of red light, and may even be overwritten completely. This alternation between lower energy red light and higher energy green light poses the clearest path forward for future research into direct light application to stem cells, as increasing both effects is synergistic for tissue regeneration and injury repair.
Novel Non-Invasive Methods – Upconversion Nanoparticles (UCNPs)
While most research is conducted exclusively with the application of light, novel optical devices have been developed to work in conjunction with light application for finer control of stem cell differentiation. Upconversion nanoparticles (UCNPs) are artificial, nano-scaled lattices of various metal ions that exhibit the capacity to upconvert photons, a process that involves absorbing two lower-energy photons and releasing them as a single higher-energy photon [1, 13]. These are especially useful in regenerative medicine research as they are easily taken in by cells [13]. Wang K. et al and Zhang Y. et al. both used UCNPs in conjunction with near-infrared (NIR) light to control stem cell differentiation [11-12]. Both teams irradiated their cell cultures with NIR light, which can penetrate deeper than higher-energy light. This NIR light then activated UCNPs within the cultures to release UV light that activated stem cell differentiation factors from where they were loaded onto the UCNPs. Wang’s team was able to increase mesenchymal stem cell osteogenic differentiation, while Zhang’s team successfully observed increased differentiation into glial cells, a broad category which includes astrocytes and other nervous system support cells.
Additionally, UCNPs are not exclusively used to release differentiation factors, as Wang M et al.’s team was able to use UCNPs to upconvert NIR light into visible blue light to achieve deeply penetrating visible light exposure, which has high potential for use at injury sites [5]. The use of UCNPs can circumvent the problems of direct application of UV light to cell cultures, such as genetic damage and low tissue penetration. These UCNPs could provide a tool by which stem cells can be influenced in more selective regions, as their area of effect is limited to tissue sites where they have been directly implanted.
However, while UCNPs have utility, they still pose a challenge to future regenerative medicine research in that they must be somehow applied directly to the target stem cells, requiring an invasive method such as injections. This is a common issue that different research teams have run into. A novel alternative to UCNPs has been developed by Zhang S. et al that applies NIR light to copper sulfide nanostructures [10]. These nanostructures produce electromagnetic oscillations upon stimulation with NIR light that have been found to increase the differentiation of hADSCs into neuron-like cells. While innovative, this method still requires the invasive placement of copper-sulfide nanomaterials at the site of target stem cells to impact their differentiation. Although it provides a potential alternative to UCNPs, this method has not been tested by any other research teams on stem cell cultures and requires a great deal of further research before it can be implemented as a regenerative medicine procedure. Regardless of the nanobiology tools selected, researchers still must identify a way to place their developed structures near target cells.
Future Possibilities
A potential future for photobiomodulation research lies in the combination of Wang M et al.’s work with UCNPs in combination with visible light application and the alternating light method of Crous et al. to achieve increased proliferation and differentiation with relatively non-invasive deep tissue injury site access [5, 9].
Additionally, Crous’ work with earlier neural photobiomodulation studies gives researchers a way to potentially further increase astrocyte proliferation and differentiation by combining Wang M. et al.’s blue light method with Yoon SR et al.’s red light method, thereby activating two different gene expression pathways simultaneously [5-6].
The alternating light method requires further inquiry, but Crous et al. have delivered promising results in the form of increased cell movement towards a specific direction in their alternating light research [9]. This directionality is important for regenerative medicine, as cells need to be directed to grow and differentiate at specific points in injury sites to prevent undesired cell growth that could interfere with normal tissue function. Directional application of green and red light could be used in specific patterns to direct mesenchymal stem cells to grow towards a target site. Specific control over the shape and structure of stem cell differentiation is the next step for regenerative medicine research, as it allows for the construction of more complicated tissues and structures for larger injuries and for potential use in organ regeneration.
Ultimately, this research is based on the application of light; the specifics of potential applications can still be tweaked. As previously mentioned, LED diodes are placed to irradiate samples for roughly an hour a day for approximately a week [5-8]. Scientists have prioritized similar methods to achieve comparable results with each other, but the use of such regular conditions leaves photobiomodulation open for a great deal of further experimentation, as optimized application of visible light has not yet been determined. Longer or shorter exposure times, alongside lowering or raising the power of the light sources has not yet been attempted on stem cell cultures. Furthermore, work with direct light has not been performed on cells that are heavily obscured from the light source, as would be expected at the site of a deeply placed traumatic injury in a clinical setting.
Once photobiomodulation has been optimized and readied for clinical use, it is likely that further work will be performed on not just cell cultures, but on in vivo stem cells and 3D structures of stem cells. In vivo refers to cells experimented on in live organisms, rather than in isolated cultures. In vivo stem cells come with the problem of difficult to control environmental conditions, but they more accurately simulate tissue conditions. Given additional advancements in bioprinting technology, stem cells in 3D structures like hydrogels have the potential for applications of light at different angles. However, a wide array of new environmental variables would require many years of preparatory work to make up for a lack of current research. Theoretically, it may be possible for a 3D printed structure of stem cells in hydrogels to be selectively irradiated with light to grow more complex tissue formations. As such, photobiomodulation has several avenues for future research to explore as biotechnology advances.
Conclusion
The remote control of stem cells has seen great advancements in the past five years, with research on novel optical devices, a variety of wavelengths, cutting-edge manipulation methods, and the production of two different categories of human stem cells. The field has even more potential in the future, especially through the combination of different research team’s work, through synthesizing the use of different wavelengths of light, UCNPs, and currently unrealized advances in biotechnology and nanotechnology. The concept of applying light to cells to control them is an appealing one, and the field will likely expand as methods become more standardized and easier to implement in molecular biology labs. The past five years of research have laid a solid foundation for future research into remote control of stem cell differentiation, and future advances may grant regenerative medicine immensely useful tools for treating traumatic injuries to the nervous system and musculoskeletal system.
References:
- National Spinal Cord Injury Statistical Center. Spinal Cord Injury Model Systems 2020. Accessed October 30, 2021.
- Yamada M et al. 2020. Neurosci Res. 152:66-77
- Sofroniew MV & Vinters HV. 2010. Acta Neuropathol. 119(1):7-35.
- Mahla RS. 2016. Int J Cell Biol. 2016:6940283.
- Wang M et al. 2019. Biomaterials. 225:119539.
- Yoon SR et al. 2021. 10(7):1664.
- Wang Y et al. 2016. Sci Rep. 2016;6:33719.
- Wang Y et al. 2017. Sci Rep. 2017;7(1):7781.
- Crous A et al. 2021. Biochimie. 2021;S0300- 9084(21)00183-8.
- Zhang S et al. 2020. Nanoscale. 2020;12(17):9833-9841.
- Wang K et al. 2020. Nanoscale. 2020;12(18):10106-10116.
- Zhang Y et al. 2020. ACS Appl Mater Interfaces. 2020;12(36):40031-40041.
- Loo JF-C et al. 2019. Coordination Chemistry Reviews. 2019;400:213042.
Inference on the Dynamics of COVID-19 in Kerala, India
By Darya Petrov
Author’s Note: I worked on this research project at the peak of the COVID-19 pandemic, while we were fully remote and on lockdown. I chose this topic because it was extremely relevant given the circumstances. I hope this report conveys the importance and value of the union of statistical modeling and public health in pandemic response efforts.
1 Introduction:
The coronavirus pandemic has been ongoing since the beginning of 2020. As of April 11, 2022, there have been 497 million confirmed cases of which 6 million resulted in death worldwide [1]; the recent World Health Organization (WHO) report on the pandemic indicates that this massive number itself is probably a significant underestimate [2]. Our understanding of the evolution of the coronavirus has dictated many large-scale social distancing measures including mask mandates, lock-downs, and travel restrictions that have had major impacts on society, economy, and public health. Conventional epidemiological models of infectious diseases, such as the SIR (Susceptible, Infected, Recovered) model which measures the spread of a disease through the change of the population in each of the three compartments listed, do not readily apply to COVID-19 dynamics; they do not utilize information on the count of asymptomatic individuals, an unobservable variable. It is well-known that asymptomatic but infected individuals have been the major spreaders of the COVID-19 pandemic, and therefore, it is imperative to obtain an estimate of such individuals in the population from available data. India is an excellent candidate for the analysis of disease dynamics because at one point during the pandemic, it had the worst COVID-19 crisis in the world. On May 6th, 2021, India had the largest worldwide single-day spike of over 400,000 new infections with shortages of hospital beds and ventilators [3]. We analyze publically available data from the state of Kerala in India to gain a better understanding of COVID-19 dynamics using a previously proposed methodology. The model is expressed through a system of difference equations, and incorporates information on social distancing measures and diagnostic testing rates to characterize the dynamics of the pandemic. The model’s key feature is its ability to estimate the unobservable count of asymptomatic individuals mentioned previously. This methodology has already been used to analyze COVID-19 dynamics in the United States [4].
2 Methods:
2.1 The Model
A graphical representation of the disease propagation model is depicted in Figure 1. The color of each box represents the observability of the compartment: red indicates unobserved, blue indicates observed, and purple indicates partially observed, meaning the compartments are observed together. Suppose at time t, Ct , Dt , Tt respectively represent the number of confirmed COVID-19 cases, number of deaths due to the disease, and number of tests performed up to time t. Let At denote the number of asymptomatic individuals, of which Ht denotes the number of persons hospitalized and Qt denotes those quarantined due to COVID-19 at time t. St denotes the number of susceptible individuals in the population at time t. Rt denotes the number of recovered individuals, of which RtQ, RtH, RtA respectively are those recovered from quarantine, hospitalization, or from being asymptomatic but never quarantined or hospitalized.
Figure 1: Graphical representation of the disease propagation model
2.2 Data Preprocessing
We consider the dynamics of the spread of COVID-19 in Kerala, India for a time window of April 1, 2020 to December 31, 2020. The proposed model is based on the observed state-wise daily counts of confirmed infections, deaths, hospitalizations and reported recoveries. Daily counts of the confirmed COVID-19 cases, recoveries, deaths, tests, and hospitalizations were obtained from an application programming interface (API) [5]. Daily hospitalizations were obtained from a state dashboard owned by the Kerala government [6]. Unfortunately, obtaining hospitalization data for India was arduous. We extracted the data manually for each district in Kerala for each day, and then combined the data into one data frame. The social mobility data was obtained from Google [7]. The data was preprocessed and cleaned, removing any irregularities present such as abnormally large counts for a given day that are clearly due to a mistake in the data reporting. Getting rid of such outliers is crucial because they can have a significant impact on the model and distort the actual relationships and patterns in the data. These irregularities or any missing observations were replaced using K-Nearest Neighbor (KNN) imputation with k=6 nearest neighbors, i.e. a missing observation for a given day was replaced by the average of the six closest observations by date. Inherent noise present in the daily counts, visually represented by frequent vertical spikes on a graph, were removed by pre-smoothing the trajectories using the Locally Estimated Weighted Scatterplot Smoothing (LOWESS) method with bandwidth 1/16. This fits smooth curves to the data points to capture general patterns in the data, with the bandwidth indicating how much of the data to use when smoothing at each point. The smaller the bandwidth, the rougher the smoothed curve will be, i.e. the graph will have more bumps.
3 Results:
3.1 Case Study for Kerala, India
We present our analysis based on the data from Kerala for the time window between April 1st, 2020 to December 31st, 2020. Figure 2 plots the daily number of people in hospitals. Note that no vaccine was available during this time period, and so any immunity from the virus could only be obtained through exposure. An important assumption our model makes is that once somebody is infected (either showing symptoms or otherwise), the person remains immune to reinfection. The black curve plots the observed values, and the red curve plots the fitted values from the model. It can be seen that the fitted values obtained from the model closely follow the observed values. This validates our proposed model and the estimation procedure. From the data and the fit, it is visible that a wave started early July 2020. The number of hospitalizations peaked early October 2020 and started decreasing afterwards.
Figure 2: Daily hospitalizations, fitted by the model (red ⎯) and observed (black ⎯)
3.2 Estimation of Latent Compartment
The estimated number of infected asymptomatic individuals (Figure 3) shows a similar pattern with a high point around the beginning of October, and dipping afterwards. There is also a local peak around the end of August. Estimation of this latent, i.e. unobservable, compartment across time is a key feature of our proposed methodology, since this information cannot naturally be obtained from the conventional epi-models.
Figure 3: Estimated number of asymptomatic individuals
3.3 Analogue of Basic Reproduction Rate
One large wave, i.e. a surge in new infections, can be observed from the plot of the proposed analogue of the basic reproduction rate (Figure 4). It measures the transmissibility of COVID-19 at time t and is influenced by spread mitigation efforts. The basic reproduction rate less than 1 indicates a decrease, while greater than 1 indicates a growth in the number of asymptomatic-infected individuals. Its estimate was mostly larger than 1 in the sub-interval, namely from the end of April to the beginning of October, indicating the singular large wave.
Figure 4: The basic reproduction rate ( R0 ) is the rate of growth of asymptomatic-infected individuals.
The plot of the number of daily new and daily reported infections (Figure 5) shows a maximum near October. The black curve plots Ct, the number of observed confirmed cases at time t+1. The red curve plots NI(t), the daily number of new infections at time t, which is calculated as the estimated number of susceptible individuals that become asymptomatic-infected at time t.
Figure 5: Daily new infections observed by the change in confirmed cases, Ct (black ⎯), versus the estimated number of new infections, NI(red ⎯) .
3.4 Transmission Rates
Figure 6 shows the plots of the crude infection rate (CIR) and net infection rate (NIR) . The red curve represents the CIR(t), the ratio of the daily change in the number of confirmed cases relative to the number of confirmed cases at time t+1. The CIR under-represents the infection rate, so the model estimates the infection rate with the NIR. This explains why the black curve represented by NIR(t) tends to be larger than the CIR(t) curve. The NIR(t) is the ratio of the daily change in the number of asymptomatic-infected individuals relative to the number of asymptomatic-infected individuals at time t.
Figure 6: The observed crude infection rate , CIR (red ⎯), and the estimated net infection rate, NIR (black ⎯)
The observed doubling rate obtained from the observed number of confirmed cases (Ct) and its estimate from the cumulative number of new infections (CNI) appear to be very close after mid July (see Figure 7). This implies reporting kept pace with the spread of the disease starting mid July. The doubling rate obtained from Ct is represented as the black curve, and the estimate obtained from CNI is represented by the red curve. It is the inverse of the doubling time at time t. The doubling time is the amount of time it takes to double the amount of infected individuals at time t. The higher the doubling rate, the faster the spread of the infection. The doubling rate reflects the effect of spread mitigation efforts, including social distancing campaigns, improved hygiene, and case tracking.
Figure 7: Doubling rate, Ct (black ⎯) and CNI(red ⎯)
Figure 8 shows the crude and net case fatality rates, CFR and NFR respectively. The black curve represents the CFR(t) and the red curve represents the NFR(t). CFR(t) is given by the percent of total deaths to the total confirmed cases up to time t. NFR(t) is given by the percent of total deaths to the cumulative number of infections up to time t estimated by the model. It is important to note that the formulas of the CFR and NFR are the same, except the denominator of the NFR is the CNI(t) while the denominator of the CFR is Ct . The observed number of confirmed cases Ct will be strictly less than or equal to the estimated cumulative number of new infections CNI(t), and likely much less, therefore the CFR is naturally much larger than the NFR.
Figure 8: The crude case fatality rate, CFR (black ⎯), and the net case fatality rate, NFR (red ⎯).
3.5 Testing and Hospitalization
The daily number of tests and its effect in quarantining asymptomatic but infected people can be judged from Figures 9 and 10. Figure 9 plots the number of tests performed per hospitalization. Tt represents the number of COVID-19 tests at time t+1. Ht represents the number of hospitalized persons for COVID-19 up to time t. This measure is an approximation of the contact tracing intensity. Figure 10 plots the RCCF, the relative change in confirmed fraction. The RCCF measures the change in the rate of currently asymptotic-infected individuals with COVID-19 that are detected through testing and quarantined relative to the rate of detection of currently infected individuals. This measure shows the dynamics of the effectiveness of detecting and isolating asymptomatic-infected individuals from the population through testing. Empirical comparison of Figures 2 and 9 reveals that although the number of daily tests could keep pace with daily number of hospitalized patients up to early July, the growing number of hospitalized people from July to October ultimately outpaced the number of daily tests. The daily number of hospitalizations beginning to decrease in early October was accompanied by the daily testing beginning to increase.
Figure 9: Tt / Ht
Figure 10: The relative change in confirmed fraction, RCCF
Discussion:
In comparison to conventional SIR models which model disease dynamics from the number of susceptible, infected, and recovered individuals, the proposed model also incorporates information about testing and quarantine. It is important to note the following assumptions the proposed model is based on:
- Only an asymptomatic individual who is not either in quarantine or in hospital can transmit the disease to a susceptible individual.
- People who recover from the disease are immune from subsequent infection.
- False positive rate for the test is negligible, so that if somebody is confirmed to be positive, then he/she is assumed to be infected.
- Anybody who shows significant symptoms, whether being in quarantine or not, is immediately hospitalized, and is tested to be positive.
- There is no effective treatment regime for the asymptomatic individuals, and so they recover or turn symptomatic at the same rate regardless of whether they are tested positive (and hence quarantined) or not.
These assumptions are quite general, however, the model could be modified if necessary to adjust for assumptions not met. For example, assumption 2 and 3 can be generalized by adding a fraction of recovered individuals to the susceptible population. Additionally, violations of some assumptions, such as assumption 1, are unlikely to have a significant impact on the disease dynamics. However, the current model does not incorporate impact of vaccination on the disease dynamics, which renders it applicable to the data being studied. Clearly, analyzing more recent data would require using a more enhanced version of the model that includes vaccination effects.
Smoothing was a crucial technique in this model because counts are rough. It was used in the data preprocessing to reduce the impact of anomalies, such as abnormally high counts likely due to incorrect data reporting. Additionally, it was used in the estimation of time varying parameters, which is intrinsic because of the locally weighted time window.
The goal of this study was to analyze how well the proposed model, which has already been used to model data from the United States, models disease dynamics of COVID-19 in Kerala, India. The performance of the model is validated by its ability to capture the large wave Kerala experienced between August and December of 2020, which is visible in the number of hospitalizations, estimated number of asymptomatic individuals, and the basic reproduction rate. The number of new infections estimated by the model appears reliable compared to the reported number of new confirmed cases. This reported number is an underestimate of the number of new infections since not all infections are reported. For example, an individual may be infected with COVID-19, but not reported as a confirmed case of COVID-19 if they are asymptomatic and did not get tested. This underestimate of the number of new infections worsens as the number of asymptomatic cases increases. The plots of testing per hospitalization and RCCF give us an idea of contact tracing intensity in Kerala, and how well it was coping with the pandemic. This model can help evaluate the effectiveness of measures used in hopes of reducing disease spread, such as social distancing, curfews, and mask mandate. The proper response to a pandemic is a controversial topic, and this model can help make informed actions in future pandemics. Just as this model was originally used on data from the United States and applied here to data from India, this model can also be applied to other regions, as long as the necessary data is available, preprocessed, and cleaned.
It is important to note that variants can have a strong influence on disease dynamics. For example, the omicron variant of the original SARS-CoV-2 strain is more infectious and spreads faster [8]. Additionally, the current state of this model is most applicable to a pandemic in which a vaccine has not yet been developed, which can be a big chunk of a pandemic since vaccine development takes time. The first cases of COVID-19 were detected in December of 2019, and a vaccine wasn’t approved until a year later in December of 2020 [9, 10], but even then the vaccine supply was limited and was distributed in phases, prioritizing those most at risk [11]. Because vaccinations impact disease dynamics, a potential next step is incorporating data about vaccinations into the proposed model [12].
Acknowledgements:
Thank you to Sruthi Rayasam for scraping the data from online, Satarupa Bhattacharjee for helping with the R code, and Dr. Debashis Paul for supervising this project.
References:
- World Health Organization. WHO coronavirus (COVID-19) dashboard. Accessed April 11, 2022. Available from: https://covid19.who.int/
- World Health Organization. 2022. Global excess deaths associated with covid-19, January 2020 – December 2021. World Health Organization. Accessed May 13, 2022. Available from: https://www.who.int/data/stories/global-excess-deaths-associated-with-covid-19-january-2020-december-2021
- Runwal, P. 2021. How India’s covid-19 crisis became the worst in the world. Science News. Accessed April 16, 2022. Available from: https://www.sciencenews.org/article/coronavirus-covid-india-crisis-social-distancing-masks-variant
- Bhattacharjee, S., Liao, S., Paul, D. Chaudhuri, S. 2022. Inference on the dynamics of COVID-19 in the United States. Sci Rep. 12(1): 2253. https://doi.org/10.1038/s41598-021-04494-z
- Babu, J., Shukla, A., & Bharath. Covid19-India API. Accessed April 21, 2021. Available from: https://data.covid19india.org/
- C-DIT. Kerala : COVID-19 Battle. GoK Dashboard. Accessed June 4, 2021. Available from: https://dashboard.kerala.gov.in/covid/index.php
- Google. Covid-19 Community Mobility Reports. Accessed August 9, 2021. Available from: https://www.google.com/covid19/mobility/
- Centers for Disease Control and Prevention. 2022. What you need to know about variants. COVID-19. Accessed April 16, 2022. Available from: https://www.cdc.gov/coronavirus/2019-ncov/variants/about-variants.html
- U.S. Food and Drug Administration.. FDA Approves First COVID-19 Vaccine. Accessed June 16, 2022. Available from: https://www.fda.gov/news-events/press-announcements/fda-approves-first-covid-19-vaccine
- Centers for Disease Control and Prevention. 2022. CDC Museum Covid-19 Timeline. Accessed June 16, 2022. Available from: https://www.cdc.gov/museum/timeline/covid19.html
- California Department of Public Health. Covid-19 vaccine prioritization recommendations for moving through vaccine phases and tiers. Accessed June 16, 2022. Available from: https://www.cdph.ca.gov/Programs/CID/DCDC/Pages/COVID-19/Vaccine-Prioritization.aspx?TSPD_101_R0=087ed344cfab2000b4d56b612d8caafb7743db82d9ff978793d54888b3c213b891734b5b4e6868b0087aa92767143000f5f590c2e053b6fe510cb7cedbfa962f4b18412db1954061d6f01003ae226244940e863d42ef53db7a5f6abb6d403af1
- Chen, X., Huang, H., Ju, J. Sun, R., Zhang, J. 2022. Impact of vaccination on the COVID-19 pandemic in U.S. states. Sci Rep. 12(1): 1554. https://doi.org/10.1038/s41598-022-05498-z
The Long-term Effects and Implications of Testicular Cancer Treatment
By Michael Guo, Molecular and Medical Microbiology ‘25
Author’s Note: After I returned home for winter break last year, I learned that a friend from my high school was diagnosed with testicular cancer. While I have limited experience with cancer and pathology, I hoped to educate myself about a topic that impacts and will continue to impact the life of one of my friends, and improve my medical literacy as well. This review is primarily based on discussing increased risk factors from testicular cancer and treatment, focusing on resulting secondary malignant neoplasms and cardiovascular disease.
Introduction:
In the past century, the two most prominent causes of death have been heart disease and cancer [1]. Heart disease disproportionately affects older adults, and cancer typically follows a similar pattern. One exception to this is testicular cancer, which in contrast to most types of cancer, occurs most often in 25-45 year old males [2, 3]. Another defining feature of testicular cancer is the extremely high survival rate in most patients, with most cases hovering around 95%. While this high survival rate is admirable, testicular cancer survivors have an increased risk for long-term effects and cancer recurrence from treatment compared to other cancers. Testicular cancer treatment greatly increases long-term effects from treatment compared to other cancer types treatments because survivors are often younger; the testes themselves are relatively exposed organs compared to the heart, lungs, etc, which makes cancer lumps more easily detectable, but also suggests younger survivors could live with long-term effects for multiple decades [4].
Caption: A study [10] depicts data collected on ages of TC patients, with a majority of survivors comprising the 25-45 year old age group.
Testicular cancer (TC), in its most simplified definition, is the uncontrolled division of cells (cancerous) in testicle tissues of males. Most treatments of TC consist of surgical removal of cancerous tissues, radiation therapy and chemotherapy, with the latter two encompassing more adverse effects. One such effect is the development of secondary malignant neoplasms (SMN), which are new cancerous cells that occur because of previous radiation therapy and/or chemotherapy [5].
The chemical drugs used in chemotherapy can also lead to various issues, including infertility, low testosterone, and various heart complications/diseases [6]. By deepening our understanding of the exact connections between TC treatments and their long-term effects, healthcare workers can greatly decrease mortality and improve the quality of life of testicular cancer survivors.
Testicular Cancer:
In order to distinguish when various TC treatments are utilized, there must be an understanding of the many forms and stages of TC. Testicular cancer almost always consists of germ cell tumors, which are cancerous cells that form from germ cells, or the sex cells (sperm in males). Non-germ cell tumor TCs are called stromal tumors; these are cancerous cells formed from supportive tissue in the testes, are relatively rare and can almost always be locally removed with surgery. There are two main types of TC germ cell tumors: seminomas and nonseminomas. Seminomas are localized in the testes and can be treated with surgery and subsequent radiation therapy and/or chemotherapy. Non-seminomas usually have spread throughout the body and require more intense chemotherapy, with subsequent surgery when needed depending on the stage/observations [7, 8]. These different forms of TC require different treatments; oftentimes patients receive a combination of multiple kinds. Most TC cases consist of germ cell tumors, have to be treated with some sort of radiotherapy/chemotherapy, and enclose many adverse, long-term side effects.
One of the most important designations in a cancer diagnosis is the stage. Cancerous tumors are categorized into five stages, with increasing severity from 0-4. Both Stage 0 and Stage 1indicate cancerous cells in one specific area, and are considered early stage cancers. Stage II and III are used when cancerous cells have expanded to surrounding tissues or the lymph nodes. Cancer cells that have spread are most commonly first found in lymph nodes, where developed cancerous cells can congregate from tissue circulation of the lymphatic system and spread to other organs. When cancerous cells spread past the original organ and lymph nodes, it is then classified as Stage IV, and is considered the most advanced stage of cancer [9].
Caption: Graphic showing the progression of cancer cell growth within a group of cells in the five cancer stages [7].
Statistically, most prognoses for TC patients are encouraging. A committee of German cancer statisticians observed that TCs only make up around 1.6% of cancers in men, and 90% of TC cases are diagnosed Stage I or II. On the other hand, TC also can also be categorized by various forms. Stromal or unknown non-germ cell tumors make up around 7% of all TC, while seminomas make up around 62%, and various non-seminomas around 31% [10]. Combined, the high diagnosis rate within early stages and the fact that the majority of cancer cases are localized result in close to a 95% 5-year survival rate, meaning a high number of patients survive for at least 5 years after diagnosis. Not to be overlooked however, is the impact of newly developed treatments of TC in the last 30 years.
Treatment and Side Effects:
Most of the adverse effects experienced from testicular cancer survivors are derived from the harsh treatment options available, rather than the cancer itself. So far, attempting to balance long-term effects from medication with sufficient treatment to remove the cancer cells has been proven to be the most successful course of medical care. All forms of TC can utilize surgical removal of tumors as a treatment option. When diagnosed early, surgery is almost always used to remove a testicle to prevent cancerous tissues from spreading. Surgical removal has little to no long-term side effects outside of normal surgical recovery standards.
More commonly seen with non-seminomas or more advanced stage TCs (stage III/stage IV) is the use of chemotherapy and radiation therapy. While chemotherapy uses specific drugs or drug combinations to target and kill rapidly dividing cells, radiation therapy uses focused radiation to break apart cancer cell DNA, which prevents division [8]. However, chemotherapy and radiation therapy come with different side effects. In chemotherapy of TC specifically, almost all treatments contain bleomycin, etoposide and cisplatin. This combination causes especially harsh long term side effects, including infertility, low testosterone, heart diseases and development of secondary cancers [2, 11]. In recent statistical studies tracking TC survivors who have undergone chemotherapy or radiation therapy (or a combination of both), researchers have noticed chemotherapy can increase the risk for SMNs and cardiovascular disease (CVD), while radiation therapy greatly increases the risk for SMNs but not necessarily CVD.
Caption: Graph visualization of cancer risk – blue line is depicting the risk of all cancers after a seminomatous TC diagnosis, compared to the green line representing general population risk of a seminomatous cancer. The red line depicts risk of all cancers after a non-seminomatous TC diagnosis, compared to the general population risk of non-seminomatous TC.
Secondary Malignant Neoplasms:
Healthcare professionals and researchers have long known that radiation and certain chemicals are able to cause cancer. Secondary malignant neoplasms (SMNs), as they have been termed, are when cancerous cells form outside of the original cancerous organ tissues, because of chemotherapy and/or radiation therapy. In a study published by Bokemeyer and Schmoll, research has suggested, “Radiotherapy is associated with a two- to threefold increased risk for secondary solid tumors” [4]. More recent studies have shown that patients exposed to higher amounts of radiation during treatment are more likely to develop SMNs than patients with lower amounts of exposure, since healthy tissues near cancer cells are exposed to high amounts of radiation at the time of treatment as well [4, 12]. Incidences of SMNs in TC survivors are especially noticeable and impactful; the younger age elicits more time for secondary cancers to occur post-treatment. While it seems that little can be done to combat the development of SMNs caused by radiation therapy during treatment, changes have been made in recent years. When doctors administer treatment of TC, if requiring radiation therapy, they aim to use minimal doses of radiation, and have completely stopped radiation therapy concentrated in the chest area in the past several years [13, 14]. While follow up appointments had been established before the long term effects of TC treatment have been quantified, follow up appointments are now taken more seriously and are being continued for a longer period of time following treatment, with a larger emphasis on secondary cancers.
Chemotherapy has also seen correlation with development of SMNs after initial treatment. Two of the drugs used in TC treatment, etoposide and cisplatin, have caused secondary malignancies to arise even in treatment of cancers other than TC. Researchers have come to agree that the impact of chemotherapy is less than radiation therapy in terms of development of SMNs. Decades old research has confirmed the correlation between complications following treatment and >4 cycles of cisplatin-based therapy [15]. Cisplatin in TC treatment has specifically been known to lead to increased risk of leukemia and myelodysplastic syndrome, both of which are related to complications to blood-forming cells in bone marrow. In another study, researchers equated the effects of chemotherapy and radiation therapy on SMNs and CVD to be similar to smoking, a well known carcinogen (cancer causing agent) [13, 14]. While the individual effects of both chemotherapy and radiation therapy regarding the development of SMNs have been documented, there have been few studies that differentiate the effects of each when both treatment options are combined. Future research surveying older survivors with more long-term effects could be the key to optimizing TC treatment when decreasing SMNs.
Cardiovascular Disease:
While both chemotherapy and radiation therapy have documented effects of SMNs, radiation therapy has not been connected to greater risk of CVD. However, among the various negative effects of chemotherapy, CVD has been one of the most important causes of premature death in TC survivors. In chemotherapy, cisplatin and bleomycin are heavy metals that with repetitive use, can build up in and weaken heart muscles, as well as cause hearing impairment and infertility [15, 16]. While the effects of bleomycin are similar to other heavy metals used in chemotherapy, cisplatin specifically damages mitochondrial or nuclear DNA of certain cells. This causes mammalian cells using ATP respiration in the mitochondria to create reactive oxygen species (ROS). ROS are unstable molecules formed from O2 that can then cause damage or cell death when reacted [17]. With repeated chemotherapy treatment, cisplatin can build up in certain areas, and is often the cause of side effects such as hearing loss (cochlea), hair loss (hair follicles), and CVD (inner linings of arteries). However, specifically why cisplatin builds up in certain tissues is currently unknown [11].
Cases of CVD from cancer treatment can especially be seen in TC because of the younger age of TC patients. Surviving TC at a mean age of 37 could mean living with an increased risk of CVD for upwards of 40 years. Researchers have attempted to substitute cisplatin with a similar compound called carboplatin in hopes of decreasing the known long-term effects. However, carboplatin has never produced efficient cure rates even while decreasing nerve and hearing damage [18, 15]. As stated before, it is unclear whether the long-term effects of cisplatin-based chemotherapy outweigh the monumental success rate of treatment of TC, simply because of the lack of data available from long-term survivors. Eventual data of TC survivors could help determine longer-term impacts of cisplatin in chemotherapy, and help with discussions regarding a need for finding a substitution for cisplatin.
Conclusion:
In regards to the post-diagnosis 5 year survival rate of testicular cancer patients, testicular cancer is one of the most survivable cancer types. However, the abnormally young age of TC patients allows us to more easily see the long-term effects of more advanced stage cancer treatment. Both radiation therapy and chemotherapy have been documented to increase risk for secondary malignant neoplasms, with chemotherapy also leading to a large variety of complications, including infertility, low testosterone, and cardiovascular diseases. While certain studies have shown such adverse effects in TC treatment, not enough data has been gathered from treated TC patients who have lived through a larger period of time since treatment. With future studies, researchers could discern the need for alternative treatments to testicular cancer, or methods to prevent the harmful effects of current testicular cancer treatment.
References:
- Leading Causes of Death. (2022, January 13). Centers for Disease Control and Prevention. https://www.cdc.gov/nchs/fastats/leading-causes-of-death
- Khan, O., & Protheroe, A. (2007). Testis Cancer. Postgraduate Medical Journal. Volume 83 (Issue: 984). https://doi.org/10.1136/pgmj.2007.057992
- Toni, L. D., ŠAbovic, I., Cosci, I., Ghezzi, M., Foresta, C., & Garolla, A. (2019). Testicular Cancer: Genes, Environment, Hormones. Frontiers in Endocrinology. https://doi.org/10.3389/fendo.2019.00408
- Bokemeyer, C., & Schmoll, H. J. (1995). Treatment of testicular cancer and the development of secondary malignancies. Journal of Clinical Oncology. Volume 13 (Issue: 1) https://doi.org/10.1200/jco.1995.13.1.283
- Virginia Cancer Institute. (n.d.). Secondary Malignancies. Retrieved March 7, 2022, from https://www.vacancer.com/diagnosis-and-treatment/side-effects-of-cancer/secondary-malignancies/
- Travis, L. B., Beard, C., Allan, J. M., Dahl, A. A., Feldman, D. R., Oldenburg, J., Daugaard, G., Kelly, J. L., Dolan, M. E., Hannigan, R., Constine, L. S., Oeffinger, K. C., Okunieff, P., Armstrong, G., Wiljer, D., Miller, R. C., Gietema, J. A., Leeuwen, F. E., Williams, J. P., . . . Fossa, S. D. (2010). Testicular Cancer Survivorship: Research Strategies and Recommendations. Journal of the National Cancer Institute. Volume 102 (Issue: 15). https://doi.org/10.1093/jnci/djq216
- Johns Hopkins Medicine. (n.d.). Types of Testicular Cancer. Retrieved March 7, 2022, from https://www.hopkinsmedicine.org/health/conditions-and-diseases/testicular-cancer/types-of-testicular-cancer
- The American Cancer Society medical and editorial content team. (2019, September 4). Treatment Options for Testicular Cancer, by Type and Stage. American Cancer Society. Retrieved March 7, 2022, from https://www.cancer.org/cancer/testicular-cancer/treating/by-stage.html
- Langmaid, S. (2016, November 28). Stages of Cancer. WebMD. Retrieved March 7, 2022, from https://www.webmd.com/cancer/cancer-stages
- Association of Population-based Cancer Registries in Germany & German Centre for Cancer Registry Data at the Robert Koch Institute. (2021, April 26). Testicular cancer. Zentrum Für Krebsregisterdaten. Retrieved March 7, 2022, from https://www.krebsdaten.de/Krebs/EN/Content/Cancer_sites/Testicular_cancer/testicular_cancer_node.html
- Murphy, M. P. (2009, January 1). How mitochondria produce reactive oxygen species. Biochemical Journal. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2605959/
- Curreri, S. A., Fung, C., & Beard, C. J. (2015). Secondary malignant neoplasms in testicular cancer survivors. Urologic Oncology: Seminars and Original Investigations. https://doi.org/10.1016/j.urolonc.2015.05.002
- The American Cancer Society medical and editorial content team. (2020, June 9). Second Cancers After Testicular Cancer. American Cancer Society. Retrieved March 7, 2022, from https://www.cancer.org/cancer/testicular-cancer/after-treatment/second-cancers.html
- van den Belt-Dusebout, A. W., de Wit, R., Gietema, J. A., Horenblas, S., Louwman, M. W. J., Ribot, J. G., Hoekstra, H. J., Ouwens, G. M., Aleman, B. M. P., & van Leeuwen, F. E. (2006). Treatment-specific risks of second malignancies and cardiovascular disease in 5-year survivors of testicular cancer. Journal of Clinical Oncology. Volume 25 (Issue: 28) https://doi.org/10.1200/jco.2006.10.5296
- Moul, J. W., Robertson, J. E., George, S. L., Paulson, D. F., & Walther, P. J. (1989). Complications of Therapy for Testicular Cancer. The Journal of Urology. Volume 142 (Issue: 6) https://doi.org/10.1016/S0022-5347(17)39135-8
- Travis, L. B., Fosså, S. D., Schonfeld, S. J., McMaster, M. L., Lynch, C. F., Storm, H., Hall, P., Holowaty, E., Andersen, A., Pukkala, E., Andersson, M., Kaijser, M., Gospodarowicz, M., Joensuu, T., Cohen, R. J., Boice, J. D., Dores, G. M., & Gilbert, E. S. (2005). Article Navigation Second Cancers Among 40 576 Testicular Cancer Patients: Focus on Long-term Survivors. Journal of the National Cancer Institute. https://doi.org/10.1093/jnci/dji278
- Breglio, A. M, et al. (2017, November 21). Cisplatin is retained in the cochlea indefinitely following chemotherapy. Nature Communications. https://www.nature.com/articles/s41467-017-01837-1?error=cookies_not_supported&code=95f5244f-996e-42ef-845e-9ca9f624b375
- Haugnes, H. S., Bosl, G. J., Boer, H., Gietema, J. A., Brydøy, M., Oldenburg, J., Dahl, A. A., Bremnes, R. M., & Fosså, S. D. (2012). Long-Term and Late Effects of Germ Cell Testicular Cancer Treatment and Implications for Follow-Up. Journal of Clinical Oncology. Volume 97 (Issue: 18) https://doi.org/10.1200/JCO.2012.43.4431
- Baird, D. C., Meyers, G. J., Darnall, C. R., & Hu, J. S. (2018). Testicular Cancer: Diagnosis and Treatment. American Family Physician. https://www.aafp.org/afp/2018/0215/p261.html
- Smith, A. (2019, December). The Long Haul: Facing the Long-Term Side Effects of Testicular Cancer. Cure Media. Retrieved March 7, 2022, from https://www.curetoday.com/view/the-long-haul-facing-the-long-term-side-effects-of-testicular-cancer