Is Rejuvenating Research Akin to the Fountain of Youth?
By Barry Nguyen, Biochemistry & Molecular Biology
Authors note: I have always been interested in the aging research field. So much so, I watched ALL 8 podcasts episodes of Dr. David Sinclair’s aging podcast during the summer (which can be found on Spotify–highly recommend). A lot of the discussion is centered around developments in rejuvenating research and the various biological pathways associated with aging that can be activated depending on one’s lifestyle.
As we age, not only does our outward appearance change, but the biological clock hidden within our cells does too. The biological clock, an intrinsic feature shared among cells, allows for partial genetic reprogramming, creating an opportunity to defy the concept of time and aging [2]. This recent development of gene therapy is our closest bet to finding the Fountain of Youth.
About a decade ago, Shinya Yamanaka had shared the Nobel Prize for discovering a cocktail of proteins with the potential to revert somatic cells back into stem cells. These transcription factors are Oct 4, Sox2, Klf4, and cMYC and are now known as Yamanaka Factors [1]. Typically referred to as OSKM genes, the Yamanaka Factors play a significant role in regulating the developmental signaling network necessary for stem cell pluripotency (defined as the capacity to differentiate to virtually all types of cells) and therefore can revert the identity of virtually any cells in the body.
Recent advancements in the study of aging at the molecular level have been significant according to Dr. Diljeet Gill, a postdoctoral researcher at the Salk Institute’s Reik Lab, which conducts research on rejuvenation.“These developments have led to techniques that enable researchers to measure age-related biological changes in human cells,” says Dr. Gill [3].
Scientists have identified two defining phenomena of the aging process to assist in characterizing signs of aging. The first is the epigenetic clock, which describes the chemical tags present throughout the genome. The second hallmark is the transcriptome, which encapsulates all the gene readouts produced by the cells.
As an organism ages, the epigenetic markers become widely different. Epigenetic modifications are an intrinsic biological feature of aging, with older organisms showing a significantly different epigenetic profile than younger organisms [1]. Because Yamanaka Factors are able to alter the epigenetic landscape of somatic cells, reprogramming-induced rejuvenation strategies using the OSKM genes are made possible. Furthermore, an animal’s epigenome can be entirely reset by chemically modifying DNA and proteins that help regulate gene activity. Essentially, this form of gene editing allows scientists to revert the aging of cells.
Cells that have undergone cellular reprogramming not only appear younger, but also function like young cells. In a new study conducted in a collaboration between Dr. In Izpisua and the Altos Lab at the Salk institute have found that mice receiving long-term treatments of Yamanaka factors expressed a gene expression and metabolism profile that resembled that of much younger mice [2].
Results of the study may open up a future of therapeutic possibilities. Researchers observed notable effects in the APBA2 gene, a gene associated with Alzheimer’s Disease and the MAF gene, a gene associated with cataract development, in their transcriptional profile; both displayed a more youthful, more abundant level of transcription, meeting one of the criteria of reverse aging. The results were promising and, according to Dr. Gill, “proved that cells can be rejuvenated without losing their function and that rejuvenation looks to restore some function to old cells.” Moreover, Professor Reik, the group leader, stresses that future work can move towards targeting rejuvenating genes to reduce effects of aging.
The prospects of this new facet of aging research are extraordinary. However, it should be noted that yamanaka factors have the capacity to induce Teratomas, a germ cell tumor. Despite a limit in studies investigating the extent to which Yamanaka Factors can induce cell tumors, the ability for Yamanaka Factors to induce pluripotency and stem cell-like properties allow cells to reach a cancer-like state. Cancers are typically characterized as uncontrolled cell division. Furthermore, the differentiated cell’s ability to revert to pluripotency significantly increases the possibility for cells to take on cancer-like states.
Nevertheless, studies within this field are exciting, and researchers are united by a common goal of identifying methods to slow or even reverse the processes that lead to disease. As research continues, society is at a rapid pace in reaching a point where predicting, preventing, and even treating diseases through cellular rejuvenation becomes a reality.
References:
- Cellular rejuvenation therapy safely reverses signs of aging in mice. Salk Institute for Biological Studies. (2023, January 5). Retrieved February 5, 2023, from https://www.salk.edu/news-release/cellular-rejuvenation-therapy-safely-reverses-signs-of -aging-in-mice/
- Fan, S. (2022, April 4). Scientists used cellular rejuvenation therapy to rewind aging in mice. Singularity Hub. Retrieved February 5, 2023, from https://singularityhub.com/2022/04/06/scientists-used-cellular-rejuvenation-therapy-to-re wind-aging-in-mice/
- Garth, E. (2022, May 12). Research reverses aging in human skin cells by 30 years. Longevity.Technology – Latest News, Opinions, Analysis and Research. Retrieved February 5, 2023, from https://longevity.technology/news/research-reverses-aging-in-human-skin-cells-by-30-ye ars/
- Two research teams reverse signs of aging in mice | science | AAAS. (n.d.). Retrieved February 5, 2023, from https://www.science.org/content/article/two-research-teams-reverse-signs-aging-mice
Review of Literature: Use of Deep Learning for Cancer Detection in Endoscopy Procedures
By Nitya Lorber, Biology and Human Physiology ’23
Author’s Note: I think now more than ever, the reality of artificial intelligence is knocking on our doors. We are already seeing how the use of AI programs are becoming more and more normalized for our daily use. AI is now driving our cars, talking to us through chatbots, and opening our phones with facial recognition. Frankly, I find it both incredible and intimidating having an artificial and computerized program making decisions with the intent of modeling the reasoning capabilities of the human mind. As an aspiring oncologist, I was really interested to see how AI is being used in the healthcare system, specifically in the field of oncology. So when my biological sciences writing class asked me to write a literature review on a topic of my choice, it was a no brainer – no AI needed. I hope that readers of this review can come away with a sense of comfort that AI is being used for improving cancer detection to potentially save lives.
ABSTRACT
Deep learning is a new technological science programmed to emulate and broaden human intellect [1]. With technological improvements and the development of state-of-the-art machine learning algorithms, the applications are endless for deep learning in medicine, specifically in the field of oncology. Several facilities worldwide train deep learning to recognize lesions, polyps, neoplasms, and other irregularities that may suggest the potential presence of various cancers. For colorectal cancers, deep learning can help with the early detection during colonoscopies, increasing adenoma detection rate (ADR) and decreasing adenoma miss rate (AMR), both essential indicators of colonoscopy quality. For gastrointestinal cancers, deep learning systems, such as ENDOANGEL, GRAIDS, and A-CNN, can help in early detection, giving patients a higher chance of survival. Further research is required to evaluate how these programs will perform in a clinical setting as a potential secondary tool for diagnosis and treatment.
INTRODUCTION
Artificial intelligence is the ability of a computer to execute functions generally linked to human intelligence, such as the ability to reason, find meaning, summarize information, or learn from experience [2]. Over the years, computer computing power has significantly improved, and its progress has provided several opportunities for machine learning applications in medicine [1]. Generally, deep learning in medicine utilizes machine learning models to search medical data and highlight pathways to improve the health and well-being of the patient, most commonly through physician decision support and medical imaging analysis [3]. Machine intelligence collects data and identifies pixel-level features from microimaging structures, which are easily overlooked or invisible to the naked eye [1, 4]. Deep learning is a subfield of machine learning that uses artificial neural networks to learn patterns and relationships in data. Its basic structure involves trained interconnected nodes or “neurons” organized into layers [1]. What sets deep learning apart from other types of machine learning is the depth of the neural network, which allows it to learn increasingly complex features and relationships in the data. The field of oncology has begun to incorporate deep learning in their screenings for cancers by training deep learning to recognize lesions, polyps, neoplasms, and other irregularities that may suggest the potential presence of various cancers, including lung, breast, and skin cancers. In an experimental trial setting, deep learning has shown its ability to aid in early cancer detection for a variety of cancers, specifically colorectal and gastrointestinal cancers, and although few studies show its performance in clinical settings, preliminary studies illustrate promising results for future deep learning applications in revolutionizing oncology today. The traditional approach to detecting colorectal and gastrointestinal cancers is through screening endoscopy procedures, which allow physicians to view internal structures [5-8]. Colonoscopies are a type of endoscopy that inserts a long flexible tube called the colonoscope into the rectum and large intestine to detect abnormalities, such as precancerous and cancerous lesions [7-9]. Advancing diagnostic sensitivity and accuracy of cancer detection through deep learning helps save lives by catching the disease before it progresses too far [1, 4].
DETECTION OF COLORECTAL CANCERS
Colorectal cancers (CRC), cancers of the colon and rectum, have the second highest cancer death rate for men and women worldwide [5]. Frequent colonoscopy and polypectomy screening can reduce the occurrence and mortality from CRC by up to 68% [5, 7]. However, several significant factors determine colonoscopy quality: the number of polyps and adenomas found during colonoscopy, procedural factors such as bowel preparation, morphological characteristics of the lesion, and most importantly, the endoscopist [5-8]. The performance of the endoscopist can vary for several factors, including the level of training, technical and cognitive skills, knowledge, and years of experience inspecting the colorectal mucosa to recognize polypoid (elevated) and non-polypoid (non-elevated) lesions [6, 7].
The most essential and reliable performance indicator for individual endoscopists is their adenoma detection rate (ADR) [5, 6]. ADR is the percentage of average-risk screening colonoscopies in which one or more adenomatous colorectal lesions are found, quantifying the endoscopists’ sensitivity for detecting CRC neoplasia [5, 7]. ADR is inversely related to incidence and mortality of CRC after routine colonoscopies [5-7]. Another performance indicator commonly used to investigate differences between endoscopists or technologies is the adenoma miss rate (AMR), calculated in sets of two repeated colonoscopies on the same subject and by finding the number of lesions missed in the first trial but found in the second [7]. The issue with the current approach to detecting CRC is the variability in performance, leading to widely diverse ADRs and AMRs amongst endoscopists. This variability often results in missed polyps and overlooked adenomatous lesions in patients, which can have serious consequences [5-8].
DEEP LEARNING IN COLONOSCOPIES
Deep learning provides a possible solution to the endoscopist performance variability problem. Deep learning could provide a standardized approach to colonoscopy imaging that would help eliminate inaccuracies generated by endoscopists who may have been distracted, exhausted, or less experienced [6, 8]. Over the past few years, several studies have analyzed deep learning’s impact on endoscopy quality (i.e. ADR, AMR) and how it plays a role in reducing the rate of CRCs. Convolutional neural networks (CNNs) succeed in image analysis tasks, including finding and categorizing lesions [5]. In addition, another experimental approach involves developing a computer-aided detection (CADe) system using an original CNN-based algorithm for assisting endoscopists in detecting colorectal lesions during colonoscopy [7]. Overall, deep learning systems can improve endoscopy quality and possibly reduce the CRC death rate by increasing ADR and polyp detection rates in the general population [5-8].
The known fact that deep learning can increase ADR has led to several subsequent studies on how this technology may impact our current system. For instance, it was not previously known how the increase of ADR by deep learning relates to physician experience. In trying to determine this relationship, Repici A, et al. (2022) discovered that both experienced and non-experienced endoscopists displayed a similar ADR increase during routine colonoscopies with CADe assistance compared to those without CADe assistance [6]. Surprisingly, this study concluded that deep learning was a significant factor for the ADR score, while also finding that the level of experience of the endoscopist was not [6]. Along with increasing ADR, Kamba et al. 2021 explored how deep learning would impact AMR and found a reduced AMR in colonoscopies conducted with CADe assistance compared to standard colonoscopies [7]. This study further confirmed conclusions made by Repici A, et al., saying endoscopists of all experiences using CADe will benefit from the reduced AMR and increased ADR [6, 7].
Moreover, deep learning is exceptionally well-trained in detecting flat lesions, which are often overlooked by endoscopists [6-8]. In evaluating deep learning use for detecting Lynch Syndrome (LS), the most common hereditary CRC syndrome, Hüneburg R, et al. found a higher detection rate of flat adenomas using deep learning compared to the High-Definition White-Light Endoscopy (HD-WLE), a standard protocol commonly used to examine polyps [8]. However, unlike other studies, the overall ADR was not significantly different between deep learning and HD-WLE groups, most likely from the study’s small sample size and exploratory nature [8]. This study was not the only one to observe a lack of significant increase in ADR. Zippelius C, et al. (2022) sought to assess the accuracy and diagnostic performance of a commercially available deep learning system named the GI Genius system in real-time colonoscopy [5]. Although the GI Genius system performs well in daily clinical practice and could very well reduce performance variability and increase overall ADR in less experienced endoscopists [8], it performed no better than that of expert endoscopists [5]. Overall, deep learning demonstrated to be superior or equal to standard colonoscopy performance, but never worse [5-8].
DETECTION OF UPPER GASTROINTESTINAL CANCERS
Upper gastrointestinal cancers, including esophageal and gastric cancer, are among the highest-ranked malignancies and causes of cancer-related deaths worldwide [4, 10, 11]. Of these, gastric cancer is the fifth most common form of cancer and the third leading cause of cancer-related deaths worldwide, with approximately 730,000 deaths each year [10,11]. Most upper gastrointestinal cancers are diagnosed at late stages in cancer because their signs and symptoms go unnoticed or are too general to produce a correct prognosis [10]. On the other hand, if these cancers are detected early, the 5-year survival rate of patients can exceed 90% [10, 11]. To diagnose gastrointestinal cancers, endoscopists must first conduct esophagogastroduodenoscopy (EGD) procedures examining upper gastrointestinal lesions to first find the early gastric cancer (EGC) [4, 11]. However, similar to colonoscopies, endoscopists require long-term specialized training and experience to accurately detect the difficult-to-see EGC lesions with EGD [4, 11]. EGD quality varies significantly by the endoscopist performance, and consequently impacts patient health [4, 10-11]. Because of the subjective, operator-dependent nature of endoscopy diagnosis, many patients are at risk of leaving their endoscopy examinations with undetected suspicious upper gastrointestinal cancers, especially if they are in less developed remote regions [10]. The rates of undetected upper gastrointestinal cancers go as high as 25.8%, and 73% of these cases resulted from endoscopists’ mistakes, such as the inability to detect a specific lesion or by mischaracterizing the lesion as benign during a biopsy [11]. There is a dire need for improved endoscopy quality and reliability as current tests rely too greatly on endoscopist knowledge and experience, creating too great of a variable for EGC detection [10, 11].
DEEP LEARNING IN ENDOSCOPIES
Deep learning systems may effectively monitor blind spots during EGDs, but very little research on deep learning applications in upper gastrointestinal cancers was conducted before 2019 [4, 11]. Previously, deep learning had been mainly used to distinguish between neoplastic, or monoclonal, and non-neoplastic, or polyclonal, lesions [10, 11]. However, CNNs were not among the researched algorithms, and the then-examined systems could not sufficiently distinguish between malignant and benign lesions [10, 11]. The first functional deep learning system to specifically detect gastric cancer was the 2019 “original convolutional neural network” (O-CNN), but this system had a low statistical precision, rendering it unviable for clinical practice [11]. This prior lack of research led to the development of three deep learning systems that could be used to detect and diagnose upper gastrointestinal cancers in hopes of catching the disease in its early stages to help the patient best: GRAIDS, ENDOANGEL, and A-CNN.
The first deep learning system developed and validated was the Gastrointestinal Artificial Intelligence Diagnostic System (GRAIDS), a deep learning semantic segmentation model capable of providing the first real-time automated detection of upper gastrointestinal cancers [10]. Luo H, et al. (2019) trained GRAIDS to detect suspicious lesions during endoscopic examination using over one million endoscopy images from six hospitals of different experiences across China [10]. GRAIDS is designed to provide real-time assistance for diagnosing upper gastrointestinal cancers during endoscopies as well as for retrospectively assessing the images [10]. In the study, Luo H, et al. (2019) found that GRAIDS could detect upper gastrointestinal cancers retrospectively and in a prospective observational setting with high accuracy and specificity [10]. GRAIDS’s high sensitivity is similar to that of expert endoscopists. However, GRAIDS cannot recognize some gastric contours delineated by experts leading to an increased risk of false positives, suggesting that this system is most effective as a secondary tool [10]. GRAIDS is seen as a cost-effective method for early cancer detection that can help endoscopists of every experience level [10].
The second deep learning diagnostic system is called Advanced Convolutional Neural Network (A-CNN), an upgraded version of O-CNN developed by Namikawa K, et al. (2020) [11]. Improving upon its predecessor, A-CNNs were able to successfully distinguish gastric cancers from gastric ulcers with high accuracy, sensitivity, and specificity [11]. This upgraded system is an essential improvement because gastric ulcers are often mistaken for cancer, leading to unnecessary cancer treatments for the patient. A-CNN can now help endoscopists in early diagnosis, improving survival rates of gastric cancers [11]. In addition, this program also helps to standardize the endoscopy approach to assuage some of the endoscopist performance variability [11].
The third deep learning system is ENDOANGEL, developed by Wu L, et al. (2021). Like A-CNN, ENDOANGEL is an upgrade of an older algorithm derived from CNNs called WISENSE [4]. Before the update, WISENSE illustrated the ability to monitor blind spots and create phosphodocumentation in real time during EGD [4]. Compared to WISENSE, ENDOANGEL achieved real-time monitoring during EGD with fewer endoscopic blind spots, a longer inspection time, and EGC detection with high accuracy, sensitivity, and specificity [4]. The deep learning program shows potential for detecting EGC in real clinical settings [4].
FUTURE IMPROVEMENT IN DEEP LEARNING DEVELOPMENT
Because deep learning is a moderately new technology, much of the available research is prospective. These studies attempt to determine if deep learning is a possible approach to reducing endoscopist performance variability. However, most require further research to illustrate how this technology will be used in a clinical setting. For example, most studies involving deep learning systems that were not commercially available and were conducted in highly specialized centers cannot indicate deep learning’s performance for lesion detection in daily clinical practice on different populations around the world [4, 6-7, 10-11]. Additionally, studies need to incorporate a greater patient sample size before they can be generalized to a larger population [7, 8]. Lastly, researchers should still consider endoscopist performance in their trials to explore every option and ensure each patient will get the same treatment no matter who their physician may be or their personal views and acceptance of deep learning technology [4, 5, 8]. These preliminary studies show potential, but the systems need improvement and research before they can be used as standalone options [4, 10-11].
CONCLUSION
Overall, deep learning has demonstrated impressive ability in detecting colorectal and gastrointestinal cancers in experimental trial settings. Deep learning provides a more standardized approach to conducting colonoscopies and endoscopies that may help to homogenize efficient screenings for every patient, regardless of their endoscopist. In colorectal cancers, studies have illustrated increased ADR and decreased AMR using machine learning. In gastrointestinal studies, deep learning has shown its ability to detect cancer just as well as expert endoscopists. Despite these advances, neural networks can only partially improve the cancer detection problem at hand. Even if neural networks improve the overall accuracy and sensitivity of cancer screenings, it will be useless if patients do not get their recommended cancer screenings at the recommended time. At the moment, human intervention is still required, in conjunction with deep learning support, to give the patient their most accurate results. It still needs to be fully understood how deep learning will perform in clinical settings as a secondary tool locally and globally. However, the preliminary studies discussed in this review illustrate promising results for future deep learning applications in revolutionizing oncology today.
First steps in the development of small-scale 3D printed hydrogel bioreactors for protein production in space travel
By Maya Mysore, Laura Ballou, Anna Rita Moukarzel, Alex Cherry, David Duronslet, Lisette Werba, Nathan Tran, Hannah Mosheim, Stephen Curry, Simon Coelho
Advisors: Kantharakorn Macharoen, Matthew McNulty, Andrew Yao, and Dr. McDonald, Dr. Nandi, and Dr. Facciotti
Author’s Note: My name is Maya Mysore, and I am a team lead on the BioInnovation Group’s Plant Bioprinter project. The BioInnovation Group is a student organization that creates research and leadership opportunities for undergraduates. The Bioprinter project is one of these opportunities.
I joined the BioInnovation group (BIG) in the winter quarter of 2019, as a freshman looking for ways to get involved on campus. I knew I liked research; I had been working in another lab. However, I was looking to explore different aspects of research. I heard about BIG through some friends in my major and went to an information session. There, I tried to join the more tech-based microfluidics project; however, my previous lab experience with cell culture convinced the lead for the Bioprinter project to get me involved in their work. I spent the next couple quarters investigating how to trap viruses in hydrogel. In Fall 2019, I was offered the role of lead. I was shocked, surprised, and a little out of my depth– after all, I had practically joined the project by accident! But I took on the role, excited about the leadership opportunity and the freedom. Now over a year into being project lead, I am planning to transition into the organization’s leadership. However, as a swan song to my time in charge, I wanted to compile all the hard work those involved with the project have accomplished. This paper is a celebration of the work of tens of student researchers over a period of several years. Hopefully, this paper will be the first of many for the Bioprinter project and the BioInnovation Group.
Abstract
As human space exploration expands to include potential settlement on the Moon and Mars, the ability to build shelter, manufacture food, produce medicine, and create other necessities in space will become increasingly important. Currently, the high cost and size constraints of sending payloads into space challenges us to think beyond the traditional manufacturing and agricultural tool-kit. Engineers have proposed that additive manufacturing, particularly 3D printing, is a solution to lower the payload costs and to enable the manufacturing of a variety of products in situ. This study focuses on 3D printing engineered biological cells for the production of biologics (e.g. pharmaceuticals that are living or derived from a biological source). We describe in-progress work to design, build, and test a small and affordable 3D bioprinter capable of printing 3D structured hydrogels that can carry living cells. We provide a general overview of the project, our progress in converting a low-cost and compact 3D printer from printing plastics to printing hydrogels, and preliminary work testing the compatibility of bioink formulations with genetically engineered rice cells that produce and secrete the enzyme butyrylcholinesterase.
Background
As humans continue to explore space and potentially settle in distant locations such as the Moon, Mars, and beyond, it will become increasingly necessary to build shelter, create food, and develop medicine while in space. However, the major costs (roughly $20,000/kg) and size constraints of sending payloads into space create challenges for such long-duration space travel beyond low Earth orbit [1-4]. Challenges include the manufacturing of food, shelter, and even medicine. 3D printing has been proposed as a cost-effective method for addressing some of these challenges, as it might allow the opportunity to ship only the printer to remote sites and to source the majority of the printing materials from the settlement location [5].
Biological systems may also play a large role in this approach. Microorganisms have been envisioned to help construct habitats through biocementation, a process that uses microorganisms to solidify inorganic matter into 3D structures [6-8]. Plants and microbes together are proposed as possible tools for the creation of sustainable ecosystems that recycle and detoxify waste and produce food [9-11]. A purported advantage of biological systems is that they can self-replicate, as each organism carries the full set of genetic instructions to create copies of itself. This means that biological systems could be delivered as light-weight “seeds”, i.e. self-replicating units that can be shipped in small and light quantities and grown to larger quantities upon permanent settlement at remote bases.
We and others envision that the 3D printing of engineered living systems (e.g bioprinting) may prove useful for the manufacturing of biologicals; this includes pharmaceuticals of or derived from a biological source [12]. In this context, the engineered living system serves as an on-demand expandable factory for the production of the biological while the 3D printer serves to produce custom-made culturing and purification hardware that can be produced in the geometries required for specific cells and production sizes. We were interested in exploring this concept and better understanding the challenges associated with the proposed process of drug production through bioprinting. In order to do this, we needed a bioprinter. Depending on their feature sets, commercial bioprinters can cost anywhere between $10,000 and $200,000, which was well outside our budget. Therefore, as a first step, we sought to design, build, and test a low-cost and compact bioprinter that we could later customize and use to explore novel design ideas.
FExisting modalities of bioprinting were considered and four main existing modalities of 3D bioprinting were considered: inkjet, pressure-assisted, laser-assisted, and stereolithography. For a detailed review on this subject, see Li et. al [13]. The major factors that were considered in the selection of a printer were types of usable bioinks, potential for good cell viability, cost, and complexity of the system (e.g. ease with which it can be modified). Inkjet-based bioprinting uses computer controls to drop small drops of bioink onto a surface. This type of printing maintains high short-tem cell viability and is widely available at low cost. However, it is limited in printing materials and creates high thermal and mechanical stress on cells which risks damage to cells and may affect long-term viability. Pressure-assisted bioprinters extrude bioink continuously onto a surface. While the extrusion process is slower and can lower cell viabilities immediately after printing (ranging from 40-80%, compared to 90% for inkjet printing), it allows use of a greater variety of materials and incorporates cells directly into the bioink. Laser-assisted bioprinters use a laser to irradiate a bioink such that the droplets adhere to the desired surface. This method of bioprinting is very precise and results in the highest cell viability; however, it is the most expensive, time-consuming, and has the highest risk of metal contamination. Finally, stereolithography printing uses illumination of a light-sensitive polymer to solidify 3D shapes. This method is fast, cost-effective, and has high final cell viabilities, but it is primarily limited by the need for a light-sensitive bioink, many of which are not biocompatible.
We chose to build a pressure-assisted bioprinter primarily due to practical factors: (a) the availability of low cost and compact fused deposition modeling (FDM) printers that could be used as chassis, theoretically enabling a “simple” swap of printing nozzles and pumps while taking advantage of the existing build platforms and 3D control systems; (b) the easy access to safe and low cost of compatible bioinks, and (c) the ability to incorporate cells directly into the bioink for prototyping.
This paper describes the progress of our project in developing a functional bioprinter. In addition, we describe the chemical assays used to evaluate engineered rice cell viability within hydrogels and these cells’ cell’s ability in gels to produce the pharmacologically-relevant enzyme Butrylcholinesterase (BChE), which is a complex human serine hydrolase enzyme that provides protection against organophosphorus poisoning from toxic agents such as sarin.
Figure 1. This diagram demonstrates the model methodology for seeding the cells into the hydrogel, printing out the cell-gel complex, and extracting the protein of interest from this system.
Methods
Printer selection, modification and testing
Selection of chassis
We sought to find a low-cost and compact FDM printer system that could be reasonably modified to extrude bioink rather than plastic filament. We ultimately selected the Monoprice MP Select Mini 3D Printer V2 because of its high availability, low cost ($250), and relative ease of modification. An accurate open source 3D computer-aided design (CAD) model (https://www.thingiverse.com/thing:2681912) of this printer was already available, making it easier to design new features for this specific unit.
Construction of an bioink extruder
To start converting the 3D plastic printer into a bioprinter, the printer’s original extrusion mechanism was replaced with a standard syringe/syringe-pump mechanism typical of bioprinters [14].
Incorporating the syringe-based bioink extruder required the design and construction of the entire extrusion system. An interchangeable mount was designed to hold the 10 mL syringe on the printer access, as seen in Figures 2b and 2c. In Figure 2b, the interchangeable mount design is shown with a trapezoidal connection piece, allowing the mount to swap between holding the 3D printer plastic extruder and the bioprinter syringe extruder system.
Figure 2. a) Inside of the 3D printer after all electrical components and panels were removed b) 3D printed interchangeable mount used to exchange the plastic extruder and the syringe extruder. c) The hydraulic extrusion system as connected to the bioprinter d) The hydraulic extrusion tubing system
The 10 mL syringe was connected to a hydraulic pumping system through a plastic tube. The hydraulic system is controlled using a Nema 17 Bipolar 40 mm Stepper Motor connected to an 8 mm threaded rod, forming a linear actuation mechanism. Connected to the rod is a 60 mL syringe plunger which is pushed through a 60 mL syringe. A liquid is placed in the 60 mL syringe and the bioink is placed in the 10 mL syringe also with a plunger sitting on top of the syringe. When the motor turns on, this liquid is pushed from the 60 mL syringe through the tubing and into the 10 mL syringe. This system pushes the plunger through the 10 mL syringe and extrudes the bioink onto the printing surface.
A T fitting made from 6 mm brass tubes was attached to the middle of the tubing system in order to remove air bubbles from the tube, as shown in Figure 2d.
Integration of hydraulic motors with chassis
To power the motor for the syringe extruder, the electrical components needed to be rebuilt. With this in mind, an Arduino Mega 2560 was connected with the HiLetGo RAMPS 1.4 control panel and the A4988 stepper motor driver boards using the wiring setup diagrammed in Figure 3.
Figure 3. This diagram shows the wiring for the 3D printer using the Arduino.
The Z-axis switch was then repositioned and mounted to the printer chassis directly under the print head, as seen in Figure 2c.
Firmware
For the firmware, Marlin was selected because it is open sourced and easily modified with the Arduino IDE. After the firmware and electronics were set up, a G code file was needed to determine the print pattern. Cura was used to develop the file due to its compatibility with the Monoprice 3D printer. The Cura profile used with the bioprinter tests followed a cylindrical shape with a square-shaped infill grid. With this information established, the Cura profile was exported as G Code. In the printer design, an SD card is required to flash the firmware and upload the G code to the bioprinter. With the firmware and G code loaded onto the SD card, the bioprinter could be set up to run test prints with the bioink. The final cost spent to make the bioprinter came out to $375. Further information on the process of building the bioprinter can be found at https://www.instructables.com/Low-Cost-Bioprinter/.
Hydrogels
Hydrogels are porous water-based polymers that have many valuable uses, especially in fields such as drug delivery and tissue engineering. Here, we use hydrogels for their ability to selectively trap materials on a size basis, as this is what allows us to trap cells and release the protein of interest. Our hydrogel protocol was adapted from Seidel et al., 2017. Briefly, the hydrogel mixture contained agarose (0.2275% w/v), alginate (2.52% w/v), methyl cellulose (3% w/v), and sucrose (3% w/v). Agarose, alginate, and sucrose were mixed into deionized water at room temperature until dissolved. This mixture and the methyl cellulose powder were then autoclaved in separate containers for 20 minutes at 121 C. Upon completion of the autoclave cycle, methyl cellulose was mixed into the gel. The mixture was then left for 12 to 24 hours in a 2-8 C fridge to allow swelling to occur [15]. After this, the gel was ready to be seeded.
Seeding and Crosslinking the Gels
Transgenic rice cells were supplied by the McDonald-Nandi lab. The cells were genetically modified with the addition of a human BChE gene optimized for rice cell compatibility and cloned into the RAmy3D expression system for transformation into A. tumefaciens to allow insertion into rice cells [16]. This allowed the engineered cells to produce the pharmacologically-relevant BChE protein. The provided cell suspensions were mixed thoroughly via pipetting to obtain even distribution of cells. This suspension was then added directly to the hydrogel in a 50% volume split of cell suspension and gel and gently mixed to distribute cells evenly. To crosslink the gels and create solid structures for later use, a 0.1 M calcium chloride solution was prepared. The hydrogel was loaded into a syringe and deposited into weight boats containing enough CaCl2 solution to half-cover the extruded hydrogel. The hydrogel would then cure in the solution for at least 5 minutes or until the shape solidified. Upon completion of curing, the hydrogel could be removed and used for experiments.
Tetrazolium Chloride Viability Assay on Hydrogels
The TTC (2,3,5-triphenyltetrazolium chloride) assay is a method for testing cell viability. TTC is turned red from a colorless solution in the presence of metabolizing cells, allowing for quantification of cell viability. When used with defined standards and run on a spectrometer, it can be used to monitor cell survival over time.
Preparation of the TTC solution involved mixing 0.4% w/v TTC in 0.05 M sodium phosphate buffer, pH 7.5. Once the TTC solution was prepared, the TTC assay was performed.
5-6 mL of 0.05 M sodium phosphate buffer was added to a 15 mL Falcon tube with cured gel to submerge the cured gel entirely. The gel remained in the solution for 15 minutes. Then the Ellman buffer was removed from the tube and 500 μL of TTC were added to the tube with gel while mixing slightly. This tube was stored in a dark area for 24 hours.
If the gel was not cured, roughly 5 mL of gelled cells were first centrifuged in a 15 mL conical tube at 4500 g for 20 minutes. The supernatant was removed and 1 mL of Ellman buffer was added and mixed. The sample was centrifuged again at 4500 g for 15 minutes, the supernatant was removed, and 500 μL of TTC solution were added to the gel-cell mix. This sample was stored for 24 hours in a dark area.
After the 24 hours period ended, the sample-TTC mix was centrifuged at 4500 g for 15min. The supernatant was removed and the gel-cell mix was washed with 1 mL deionized water. The mixture was re-centrifuged at 4500 g for 10 minutes. The supernatant was removed again and 1 mL of 95% ethanol was added to the gel-cell mix. The sample was transferred to a microcentrifuge tube and placed in a 60C water bath for approximately 10 minutes. The sample is then centrifuged at 21.1 g for 15 minutes to recover the final supernatant. The supernatant was then run on a colorimeter or Tecan and the absorbance value was read at 485 nm. Beer’s law was then used to determine concentration from this value.
Seeded Cell-Ellman BChE Concentration Assay
The Ellman assay was used to measure BChE concentration for a sample at a given time point. This assay uses the kinetics of a color changing reaction to quantify the amount of BChE in solution. When in the presence of specific substrates, BChE turns a colorless solution yellow; the peak rate of this reaction can be determined and used to calculate BChE mass in a sample.
After cells were seeded into a hydrogel complex with a disc shape approximately 7 cm in diameter and 1 cm thick, the complex was suspended in 40 mL sucrose-free nutrient broth (NB-S).
The flask was then covered with a cloth filter and placed in the shaking incubator (37C, 5% CO2, 80 rpm). 50 μL media samples were collected from the flask daily over 14 days and the Ellman assay was run directly following collection of each of these samples.
The Ellman assay protocol was based on the Cerasoli lab protocol, which was adapted from Ellman et al., 1961 [17]. To perform the Ellman assay, a 20 mm stock solution of 5, 5’ – dithiobis-(2–nitrobenzoic acid) (DTNB) was prepared. A 75mM stock solution of S-Butyrylthiocholine (BTCh) iodide was also prepared.
Immediately prior to performing the Ellman assay, the Ellman substrate was prepared. 60 μL of DTNB and 30 μL of BTCh were added to the phosphate buffer in the falcon tube. The tube was temporarily stored in ice with light protection.
Then the Ellman assay was performed. In a 96-well plate, 50 μL of sample containing BChE was were diluted into 0.1 M phosphate buffer, pH 7.4, to ensure the generated? outputted slope readings (mOD/min) would fall in the range of 200-1000 when read for 3-5 minutes at 25 C. This dilution was done by estimating the approximate BChE concentration and estimating the mOD/min based on the expected value. 150 μL of Ellman substrate was added to each sample containing well. The optical density of the sample was immediately read at a wavelength of 405 nm for a total of 300 s (5 min) after the measurement was started.
After collecting data from the assay, Beer’s law was used to determine the concentration of product formed. From that value, we could estimate the mass of functional BChE in the total volume of the sample collected [18].
Results
TTC-Gel compatibility
To measure in-gel cell viability, we evaluated the use of the tetrazolium chloride (TTC) assay. This assay measures metabolic activity in live cells by reducing tetrazolium chloride to red formazan through the process of cell metabolism. Effectively, it provides an indication of how well the cells survive over time. Our team modified the assay for use in gels by including extra Ellman buffer and centrifugation steps to provide more opportunity for cells in the gel to be washed.
Figure 4. This figure shows the results of the TTC assay run on the transgenic rice cells in suspension. The leftmost tube is a positive control showing the TTC assay done on cell aggregates in suspension (i.e. without gel) that have been centrifuged into a pellet after the assay was performed. The middle and rightmost tubes are cells suspended in a hydrogel; the TTC assay was performed on this combination of cells in gels. In each tube, the cells have been stained red from the assay, indicating the presence of metabolic activity. These samples can go on to be washed and suspended in ethanol to obtain a viability data value.
To qualitatively assess how different factors like cell distribution and crosslinking might influence the results of the TTC assay, we performed additional variations of the assay. We first visually examined whether cell homogeneity was impacted by the gel. Then, we performed the TTC assay on E. coli cells alone as a positive control. After that, we tested the effects of non-crosslinked and crosslinked gel to ensure neither condition would prevent the use of the assay. E. coli was used for these tests due to our group’s ability to access it more regularly and grow it more easily than the genetically modified rice cells from Dr. McDonald’s lab. All of these tests together allowed us to determine that cell survival could indeed be measured within the gel, allowing us to monitor culture health over time. This will be critical in future use of the model, allowing us to determine ways to improve cell health and protein output by providing a metric for us to test against.
Homogeneous mixing of biological sample
To determine later TTC accuracy, the first key issue to address was homogeneity of cells in a hydrogel. This would determine whether sectioned off samples of cell-gel complexes would be representative of a whole sample. To ensure that the gel mixing protocol yielded a homogeneous suspension of the cells, we first tested our procedure by mixing E. coli expressing a transgenic green fluorescent protein (GFP) and imaged the suspension under UV light. We expect E. coli to distribute homogeneously in the gel similarly to the transgenic rice cells. This mixture was observed (Figure 5a) and confirmed by visual inspection of a homogeneous mix.
Test of TTC assay with bacterial suspension
To ensure that the TTC assay in later tests would be effective with E. coli, we first tested the TTC assay on an E. coli suspension as a positive control for later tests. We ran the modified TTC assay protocol described in Methods, and observed a color change in the solution. The resultant red solution (Figure 5b) matches the literature expectations for the output of this assay on living cells and indicates the assay is effective for E. coli.
Test of TTC assay with bacteria seeded in hydrogel
After confirming the TTC assay was effective with E. coli, it became important to determine how the presence of gel would affect the assay. We suspended the E. coli cells in the hydrogel and ran the modified TTC assay. The results seen in Figure 4c show the suspension turning red, which visually indicates the presence of cell metabolic activity and the effectiveness of the TTC assay.
Test of TTC assay with bacteria seeded in a crosslinked hydrogel
Upon determining the gel did not qualitatively affect the output of the TTC assay, it became necessary to determine whether crosslinking the gel had any effect on the effectiveness of the TTC assay. We reran the same experiment as the non-crosslinking hydrogel experiment, with the only change being the crosslinking process and the different first wash step. We found that the result of the TTC assay appears to be unaffected by the presence of the crosslinked out layer, as the solution turns red in the same way it does for the positive control and the non-crosslinked gel (Figure 5d).
These experiments allowed us to qualitatively determine whether the TTC assay could be an effective measure of cell viability. They also demonstrated that the introduction of a crosslinked hydrogel will not have visible impacts on measuring cell viability.
Figure 5. Qualitative TTC assays were run on E. coli with the pMax plasmid to test homogeneity within the gel and the effectiveness of the TTC assay in different hydrogel conditions. 5a shows the bacteria mixed homogeneously within the hydrogel, which is visible in the fluorescence that is present homogeneously through the sample. 5b shows the ethanol suspension output for a TTC assay run on a pMax E. coli culture, providing a control for later experiments and showing that the TTC assay is effective for E. coli. The left image is the control and the right image is the test condition. The control is run in the same conditions as the test, except the cells are placed in a 60C water bath for ten minutes prior to adding TTC in order to kill them. 5c shows the output prior to ethanol suspension for a TTC assay on E. coli pMax cells that were suspended in an uncured hydrogel. The left tube is the control and the right tube is the experimental condition. The red color visible in the right tube shows that the presence of the hydrogel does not prevent use of the TTC assay. 5d shows the ethanol suspension output for a TTC assay run on E. coli pMax cells that had been suspended in a cured hydrogel. The left image is the control and the right image is the experimental condition. The red color of the suspension indicates the TTC assay remained effective even with the addition of the crosslinked outer layer of the gel. Throughout this figure, variation in intensity of the redness of the samples is related to variations in time spent in suspension of the TTC solution, with redder samples correlating to longer time.
Initial Attempts at Measuring BChE Production
Our second major goal was to determine whether BChE could be collected from our model system (as seen in figure 1). This would allow us to determine if our model system was an effective way to collect our protein of interest for future space travel applications, as well as confirm that our test for BChE quantity would be effective in this system. To test this, our team ran the seeded cell-Ellman assay as described in methods to assess the amount of BChE that was escaping into the media. We first prepared a hydrogel, mixed the transgenic rice cells in, and cured it into a disc shape roughly 7 cm in diameter and 1 cm in height. We then suspended this cured cell-gel complex in NB-S media to stimulate BChE production, and we kept this mix in a spinning incubator to ensure aeration and adequate diffusion of materials in and out of the gel. Media samples were collected over the course of 14 days and were run with the Ellman assay for BChE detection on a spectrometer. The Ellman assay uses the enzyme kinetics of a color-changing reaction between BChE and a substrate to quantify the amount of BChE present in a sample at a given time point.
It is important to note that this test was intended as a trial run of the system in order to ensure that the assay works and that useful data is being collected. In addition, we sought to assess if BChE could escape from the gel at all. Therefore, no negative control was run and only one run of data was collected (shown in the figure below.) As a result, we cannot conclusively state anything about the data. However, the data does show a trend worth noting for future experimentation. The This is that active BChE concentration in the media increased for the first roughly 100 hours, after which the values dropped off. At the time point marked in figure 5, 96 hours, we see the maximal BChE present. If the unusually low value seen at the roughly 120 hour time point is considered erroneous (which we suspect), the data suggests increased production of BChE over the first 4 days of culture followed by a slow decay thereafter with production ending at around day 8. This provides an early quantitative estimate of the time-dependence of BChE production in this model system. This experiment is a first attempt and will be repeated with various parameter variations in the future.
Figure 6. This plot shows the approximate active concentration of BChE released into the media for various time points over 14 days. Each sample was a 50 μL amount of media pulled from the small scale system model. This figure shows a burst in production of active BChE until the 96 hour time point (denoted with a dashed red line), after which the values drop off. The data point at t=120 hours is most likely an outlier resulting from this data being for one set of samples from one test condition.
Preliminary Bioprinter Testing
The process of building and testing the bioprinter was done in parallel with the TTC and Ellman assay testing. Detailed bioprinter testing has not been performed; however, initial testing of the printer showed its ability to print hydrogel into pre-programmed patterns. The grid pattern seen in Figure 7 was printed into a petri dish containing CaCl2 curing solution. The print shows excess hydrogel accumulation near the edges where the printhead briefly paused and reversed direction. In the center of the print, the lines in the grid averaged 1.25mm +/- 0.4 in width. Further testing and refinement is currently in process.
Figure 7. This shows a test print from our modified 3D printer using the hydrogel described in the methods section and cured in standard CaCl2 curing solution. This structure is described as a lattice shape and will be the primary pattern for future prints.
Discussion
In these experiments we determined that the TTC assay was effective in hydrogels, the Ellman assay showed the ability of protein to be detected from solution, and the bioprinter was able to create the desired lattice shape for later use.
Printer Performance
Our experiments to date have demonstrated our ability to convert a low-cost and compact FDM printer into a preliminarily functional bioprinter. The conversion of the original chassis required the modification of the printhead support, the development of a syringe-based hydraulic pump, and the modification of electronic and software control systems. Preliminary prints indicate that the printer can successfully deposit a programmed pattern with feature sizes in the range of 1.5mm. Existing conventional commercial bioprinters can achieve resolutions of 100-200µm, (some even claim filament diameters as low as 3µm), suggesting that we have room to improve the resolution of our system [19]. In addition to improving the resolution of the prints, we want to explore alternate methods for delivering the CaCl2 curing solution during alginate filament deposition to minimize user interaction and allow complete processing inside a biosafety cabinet; this should allow us to increase sterility during printing and print quality.
Cell Viability
Since it is known that pressure-assisted printing may negatively impact cell viability during printing, a key concern was the resulting cell viability of the system. As a result, our general goal for this phase of the project was to test whether a pressure-assisted bioprinter system could maintain cell viability after extrusion. We adapted the TTC assay for this purpose and tested our protocol to determine the effect of bioink and extrusion on cell viability under conditions mimicking those experienced during bioprinting.
Generally, the TTC assay demonstrated the ability of the assay to cellular viability in the crosslinked hydrogel, despite the unknown nature of how crosslinking affects pore size. Despite this success, the TTC assay remains largely qualitative as it is challenging to get quantitative measurements of cell viability when cells are embedded in a gel. This is further complicated by factors like the heterogeneous distribution of cells (or cellular aggregates) in the gel (see figure 4, rightmost sample). If homogeneity is not maintained, we need to design assays that take into account heterogeneous distribution of cells in the gel. In future experiments, we seek to determine whether samples from a large complex of cells in a hydrogel will provide a representative sample.
In later experiments, additional key variables that may potentially affect viability will be tested. These variables include media composition, culture duration, environmental conditions such as temperature, gel architecture, and the additional variables associated with the printer extrusion process (e.g. pressure, needle pore diameter, etc.). Determining how these specific factors affect viability will allow us to modify the printer design to minimize the drop in cell viability upon extrusion.
Protein production
Having confirmed the effectiveness of the TTC assay in the hydrogel, we moved forward to analyzing BChE production and its diffusion into the media. The assay we adopted allowed us to develop a standard method for data collection that can be used to analyze how various factors impact the cells’ ability to produce BChE. Figure 6, for example, shows that we can measure BChE production and diffusion out of the gel, and that under our preliminary experimental conditions, production peaks at 96 hours and then falls over the next 150 hours. While encouraging, this experiment needs to be repeated with many more samples and replicates to obtain a more reliable assessment of measurement error associated with the assay. Despite needing to replicate the experiment, we are confident that this preliminary experiment answered the core question of whether such a large protein – 85 kDa monomers and 4 units in quaternary form, with a total size of 574 monomers [20] – can effectively diffuse out of the hydrogel and avoid denaturation long enough to be collected and purified.
In addition to replication, future experiments should be explored to further improve protein escape from the hydrogel. These tests could increase the mixing speed to use centrifugal force to free proteins from the gel, increase pore size to create more physical space for protein escape, or print the 3-dimensional lattice structure to increase surface area and allow greater escape. Other relevant variables whose impact on BChE production should be tested include media composition and media changing schedules, culture duration, environmental conditions, gel architecture, and growth temperature. In our initial experiments, plant cells were grown in a shaking incubator at 37C to mimic the environment of protein production in mammalian hosts. However, this growth condition may have stressed the plant cells for which growth at 27C is more typical [16, 21]. This may explain the trend shown in figure 6, where die-off occurs after 96 hours.
Finally, in our current studies, the presence of sucrose in gel formulation (which inhibits BChE production) may have adversely impacted the amount of protein produced. While we expected that overlaying a relatively large volume of sucrose-free media would effectively dilute the sucrose to low levels, the presence of sucrose in the initial formulation could have nevertheless impacted the cells’ initial states and therefore protein production. A followup experiment that more stringently controls for the presence of sucrose in the gel than in the studies described above seems warranted.
Conclusion
In this work, we successfully modified an off-the-shelf pressure-assisted 3D printer into a working bioprinter. In addition, we established that BChE producing rice cells are biocompatible with the different bioink gel formulations and that our assays for testing cell viability and protein production are effective when analyzing the cells within the gel. Having shown that we can print gel, assess cell survival, produce BChE, and quantify its abundance, we next seek to optimize both printer function and the measurement assays for cell viability and protein concentration in ways that provide more quantitative data and more refined control over printed structures. Eventually, we expect that such advances will allow us to optimize protein production itself and ultimately develop a bioprinter suitable for protein production during space travel or in other remote locations.
Acknowledgements
Thank you to the Molecular Prototyping and BioInnovation Lab for the lab space, the BioInnovation Group for the administrative, scientific, safety, and monetary support, the McDonald-Nandi lab for materials and mentorship, and all past, present and future members of the Bioprinter team for contributing to these experiments.
References:
- Joshi SC, Sheikh AA. 2015. 3D printing in aerospace and its long-term sustainability. Virtual Phy Prototy [Internet]. 10(4):175–185. doi.org/10.1080/17452759.2015.1111519
- Hodkinson PD, Anderton RA., Posselt BN, Fong KJ. 2017. An overview of space medicine. Br J Anaesth [Internet]. 119(suppl_1):i143–i153. doi.org/10.1093/bja/aex336
- Aglietti, GS. 2020. Current Challenges and Opportunities for Space Technologies. Front Space Tech [Internet]. 1:1. doi.org/10.3389/frspt.2020.00001
- Coopersmith J. 2011. The cost of reaching orbit: Ground-based launch systems. Space Policy [Internet]. 27(2):77–80. doi.org/10.1016/j.spacepol.2011.03.001
- Ghidini T. 2018. Regenerative medicine and 3D bioprinting for human space exploration and planet colonisation. J Thorac Dis [Internet]. 10(Suppl 20):S2363–S2375. doi.org/10.21037/jtd.2018.03.19
- Gleaton J, Lai Z, Xiao R, Chen Q, Zheng Y. 2019. Microalga-induced biocementation of martian regolith simulant: Effects of biogrouting methods and calcium sources. Constr Build Mater [Internet]. 229:116885. doi.org/10.1016/j.conbuildmat.2019.116885
- Pilehvar S, Arnho M, Pamies R, Valentini L, Kjøniksen AL. 2020. Utilization of urea as an accessible superplasticizer on the moon for lunar geopolymer mixtures. J Clean Prod [Internet]. 247:119177. doi.org/10.1016/j.jclepro.2019.119177
- Kumar A, Dikshit R, Gupta N, Jain A, Dey A, Nandi A., … Rajendra A. 2020. Bacterial Growth Induced Biocementation Technology, ‘Space-Brick’ – A Proposal for Experiment at Microgravity and Planetary Environments. BioRxiv [Internet]. 2020.01.22.914853. doi.org/10.1101/2020.01.22.914853
- Lopez JV, Peixoto RS, Rosado AS. 2019. Inevitable future: space colonization beyond Earth with microbes first. FEMS Microbiol Ecology [Internet]. 95(10). doi.org/10.1093/femsec/fiz127
- Bornemann G, Waßer K, Tonat T, Moeller R, Bohmeier M, Hauslage J. 2015. Natural microbial populations in a water-based biowaste management system for space life support. Life Sci Space Res [Internet]. 7:39–52. doi.org/10.1016/j.lssr.2015.09.002
- Bokulich NA, Lewis ZT., Boundy-Mills K, Mills DA. 2016. A new perspective on microbial landscapes within food production. Curr Opin Biotechnol [Internet]. 37:182–189. doi.org/10.1016/j.copbio.2015.12.008
- Varma A, Gemeda HB, McNulty MJ, McDonald KA, Nandi S, Knipe JM. 2021. Bioprinting transgenic plant cells for production of a recombinant biodefense agent. BioRxiv [Internet]. 2021.02.01.429263. doi.org/10.1101/2021.02.01.429263
- Li J, Chen M, Fan X, Zhou H. 2016. Recent advances in bioprinting techniques: approaches, applications and future prospects. J Transl Med [Internet]. 14:271. doi.org/10.1186/s12967-016-1028-0
- Pusch K, Hinton TJ, Feinberg AW. 2018. Large volume syringe pump extruder for desktop 3D printers. HardwareX [Internet]. 3:49–61. doi.org/10.1016/j.ohx.2018.02.001
- Seidel J, Ahlfeld T, Adolph M, Kümmritz S, Steingroewer J Krujatz F, … Lode A. 2017. Green bioprinting: extrusion-based fabrication of plant cell-laden biopolymer hydrogel scaffolds. Biofabrication [Internet]. 9(4):045011. doi.org/10.1088/1758-5090/aa8854
- Corbin JM, Hashimoto BI, Karuppanan K, Kyser ZR, Wu L, Roberts BA, … Nandi S. 2016. Semicontinuous Bioreactor Production of Recombinant Butyrylcholinesterase in Transgenic Rice Cell Suspension Cultures. Front Plant Sci [Internet]. 7:412. doi.org/10.3389/fpls.2016.00412
- Ellman GL, Courtney KD, Andres VJ, Feather-Stone RM. 1961. A new and rapid colorimetric determination of acetylcholinesterase activity. Biochem Pharmacol [Internet]. 7:88–95. doi.org/10.1016/0006-2952(61)90145-9
- Alkanaimsh S, Karuppanan K, Guerrero A, Tu AM, Hashimoto B, Hwang MS, … McDonald KA. 2016. Transient Expression of Tetrameric Recombinant Human Butyrylcholinesterase in Nicotiana benthamiana. Front Plant Sci [Internet]. 7:743. doi.org/10.3389/fpls.2016.00743
- Gu Z, Fu K, Lin H, He Y. 2020. Development of 3D bioprinting: From printing methods to biomedical applications. Asian J Pharm Sci [Internet]. 15(5):529–557. doi.org/10.1016/j.ajps.2019.11.003
- Ngamelue MN, Homma K, Lockridge O, Asojo OA. 2007. Crystallization and X-ray structure of full-length recombinant human butyrylcholinesterase. Acta Crystallogr [Internet]. 63(9):723-727. doi.org/10.1107/S1744309107037335
- Macharoen K, McDonald KA, Nandi S. 2020. Simplified bioreactor processes for recombinant butyrylcholinesterase production in transgenic rice cell suspension cultures. Biochem Eng J [Internet]. 163:107751. doi.org/10.1016/j.bej.2020.107751
COVID-19 survivors can retrain their smell to enjoy food and wine again
By Daniel Erenstein, Neurobiology, Physiology & Behavior ‘21
Author’s Note: Last spring, I enrolled in the inaugural offering of the University Writing Program’s wine writing course. Our instructor, Dr. Alison Bright, encouraged us to report on topics of personal interest through our news stories on the wine industry, viticulture, enology, and more. In this article, which was prepared for an audience of general science enthusiasts, I examine how biologists are making sense of a puzzling COVID-19 symptom — anosmia, or loss of smell — and what COVID-19 patients with this condition can do to overcome it. Eighteen months into this pandemic, scientists continue to study cases of COVID-19-related anosmia with dreams of a treatment on the horizon. I hope that readers feel inspired by this article to follow this in-progress scientific story. I extend my appreciation to Dr. Bright, who throughout the quarter shared approaches to rhetorical awareness that elevated my grasp of effective writing.
Image caption: Anton Ego, the “Grim Eater” from PIXAR’s Ratatouille, is reminded of his childhood by Remy’s rendition of ratatouille, a Provençal dish of stewed vegetables.
With a single bite of Remy’s latest culinary creation, the eyes of Anton Ego, a notoriously harsh food critic, dilate, and Ratatouille’s viewers are transported back in time with Monsieur Ego. The meal — a simple yet elegant serving of ratatouille, accompanied by a glass of 1947 Château Cheval Blanc — has triggered a flashback to one singular moment, a home-cooked meal during his childhood. The universal charm of this enduring scene resonates; in Ego’s eyes, many recognize how our senses of smell and taste can impact a culinary experience.
Imagine how a real-life version of this scene might change for the millions of COVID-19 patients who have lost their sense of smell [1]. Anosmia, the phenomenon of smell loss, has become one of the more perplexing COVID-19 symptoms since first observed in patients during the earliest months of the pandemic [2].
What happens when we lose our sense of smell? During the pandemic, scientists have studied smell loss, which affects more than 85 percent of COVID-19 patients according to research published this year in the Journal of Internal Medicine [3]. In fact, anosmia is so common in COVID-19 patients that physicians were offered guidance for testing olfactory function as an indicator of infection last year [4].
To simplify studies of these complicated senses, taste and smell are often examined independently of one another, even though these senses are usually experienced simultaneously.
“Smell is just — it’s so crucial to taste, which means it’s really crucial to everything that I do,” said Tejal Rao, a New York Times food critic, in a March episode of The Daily [5]. “And it’s really difficult to cook without a sense of smell if you’re not used to it. You don’t know what’s going on. It’s almost like wearing a blindfold.”
Rao, who lost her sense of smell in mid-January after contracting COVID-19, began to search for answers to this mystery from scientists. Rao’s journey started with TikTok “miracle cures” and other aromatherapies — unfortunately, they were too good to be true — but she eventually discovered the work of Dr. Pamela Dalton, a scientist at the Monell Chemical Senses Center in Philadelphia [6]. At the center, Dalton studies the emotions that are triggered by our sense of smell [7].
During simple colds or viral infections, smell is normally affected when the molecules in food and other aromas are physically blocked off from chemoreceptors in our nose by congestion. Scientists have also cited Alzheimer’s and Parkinson’s diseases, head trauma, and chemotherapy as triggers for anosmia [8]. But a separate phenomenon was occurring in the case of COVID-19.
“COVID is different in that way, because most people who lost their sense of smell did so without having any nasal congestion whatsoever,” Dalton told Rao during an interview.
One study published in October of last year by Dr. Nicolas Meunier, a French neuroscientist, aimed to investigate how the SARS-CoV-2 virus, which causes COVID-19, may disrupt sustentacular cells [9]. These structural cells express the ACE2 receptor, which the virus hijacks to gain entry into our cells, at higher levels [10]. Sustentacular cells support the specialized neurons that transmit signals from the nose to the brain.
When Meunier and his team at Paris-Saclay University in France infected hamsters with the virus, tiny hair-like projections known as cilia on the surfaces of olfactory neurons began to peel back from sustentacular cells. This disruption is a possible explanation for the difficulties with smell that COVID-19 patients experience.
If it is true that damage to sustentacular cells causes anosmia, loss of smell is not an irreversible brain condition. In this case, the poor connection between incoming odors and brain networks that typically process these stimuli is at fault, not direct damage to the brain itself. The sudden onset of smell loss in COVID-19 patients supports this thinking.
“It was just like a light bulb got turned off or a switch got flicked to off,” Dalton said. “And one moment they could smell. And the next moment, nothing smelled.”
But because olfactory support cells regularly regenerate, this loss of smell is only temporary, which allows for retraining of our senses. Two months of smell training, also known as olfactory training, allowed Rao to regain her sense of smell.
Olfactory training gradually exposes patients to particularly strong smells. Spices such as cinnamon or cumin, for example, were perfect for Rao’s first smell training session [5], and AbScent, a British charity offering support to patients with anosmia, sells kits with rose, lemon, and eucalyptus [8]. Scientists have found that recurring exposure to these strong scents gives the brain time to recalibrate its networks, a feature known as neuroplasticity [11].
But “you don’t just go from hurt to healed overnight,” Rao said. “It’s more like adjusting and learning how to live in a new space. That’s really just the beginning.”
Our chemical senses have the power to satisfy, to inspire, even to cause our memory to reveal itself, as 20th-century French author Marcel Proust observed in his seven-volume novel À la recherche du temps perdu, or In Search of Lost Time. Researchers have even speculated that our sense of smell can facilitate learning in other sensory domains, including vision [12].
While scientists further investigate how coronavirus causes loss of smell, olfactory training can provide an avenue in the meantime for COVID-19 patients to recover this crucial sense. Indeed, many patients are “in search of lost time,” and smell training can help them to once again experience food and wine in its sensory entirety.
References:
- Allen J, Almukhtar S, Aufrichtig A, Barnard A, Bloch M, Cahalan S, Cai W, Calderone J, Collins K, Conlen M, et al. 2021. Coronavirus in the U.S.: Latest Map and Case Count. New York (NY): New York Times; [accessed 28 July 2021]. https://www.nytimes.com/interactive/2021/us/covid-cases.html.
- Symptoms of COVID-19. 2021. Atlanta (GA): Centers for Disease Control and Prevention, National Center for Immunization and Respiratory Diseases, Division of Viral Diseases; [accessed 28 July 2021]. https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html.
- Lechien JR, Chiesa-Estomba CM, Beckers E, Mustin V, Ducarme M, Journe F, Marchant A, Jouffe L, Barillari MR, Cammaroto G, et al. 2021. Prevalence and 6-month recovery of olfactory dysfunction: a multicentre study of 1363 COVID-19 patients. J Intern Med. 290(2):451–461. https://doi.org/10.1111/joim.13209.
- Whitcroft KL, Hummel T. 2020. Olfactory Dysfunction in COVID-19: Diagnosis and Management. JAMA. 323(24):2512–2514. https://doi.org/10.1001/jama.2020.8391.
- Antolini T, Dorr W, Powell D, Schreppel C. 2021. A Food Critic Loses Her Sense of Smell. New York (NY): New York Times; [accessed 28 July 2021]. https://www.nytimes.com/2021/03/23/podcasts/the-daily/coronavirus-smell-food.html.
- Rao T. 2021. Will Fish Sauce and Charred Oranges Return the World Covid Took From Me? New York (NY): New York Times; [accessed 28 July 2021]. https://www.nytimes.com/2021/03/02/dining/covid-loss-of-smell.html.
- What COVID 19 is teaching us about the importance of smell, with Pamela Dalton, PhD. 17 Mar 2021, 33:31 minutes. American Psychological Association; [accessed 28 July 2021]. https://youtu.be/0pG_U13XDog.
- Schoch D. 2021. Distorted, Bizarre Food Smells Haunt Covid Survivors. New York (NY): New York Times; [accessed 28 July 2021]. https://www.nytimes.com/2021/06/15/health/covid-smells-food.html
- Bryche B, St Albin A, Murri S, Lacôte S, Pulido C, Ar Gouilh M, Lesellier S, Servat A, Wasniewski M, Picard-Meyer E, et al. 2020. Massive transient damage of the olfactory epithelium associated with infection of sustentacular cells by SARS-CoV-2 in golden Syrian hamsters. Brain Behav Immun. 89(2):579–586. https://doi.org/10.1016/j.bbi.2020.06.032.
- Brann DH, Tsukahara T, Weinreb C, Lipovsek M, Van den Berge K, Gong B, Chance R, Macaulay IC, Chou HJ, Fletcher RB, et al. 2020. Non-neuronal expression of SARS-CoV-2 entry genes in the olfactory system suggests mechanisms underlying COVID-19-associated anosmia. Sci Adv. 6(31): eabc5801.
- Kollndorfer K, Kowalczyk K, Hoche E, Mueller CA, Pollak M, Trattnig S, Schöpf V. 2014. Recovery of Olfactory Function Induces Neuroplasticity Effects in Patients with Smell Loss. Neural Plast. 1–7. https://doi.org/10.1155/2014/140419.
- Olofsson JK, Ekström I, Lindström J, Syrjänen E, Stigsdotter-Neely A, Nyberg L, Jonsson S, Larsson M. 2020. Smell-Based Memory Training: Evidence of Olfactory Learning and Transfer to the Visual Domain. Chem Senses. 45(7):593–600. https://doi.org/10.1093/chemse/bjaa049.
Surviving COVID-19: Variables of Immune Response
By La Rissa Vasquez, Neurobiology, Physiology & Behavior ‘23
Author’s Note: In this paper, I analyze autopsy reports conducted on deceased COVID-19 patients and supply a breakdown of the body’s immune response. The purpose of this paper is to provide a more generalized synopsis of how the body is affected by the virus from the onset of infection to the escalating factors that contribute to cause of death. COVID-19 and SARS-CoV-2 are referenced countless times throughout this paper, but they should not be used interchangeably. The name of the pathogenic virus is “Severe Acute Respiratory Syndrome Coronavirus 2” (SARS-CoV-2), and the name of the illness is called COVID-19 and is the common usage in forms of discussion. This paper only scratches the surface of the virus’s complexity and its effects upon the body and societies around the world.
Introduction
On December 31, 2019, the first case of the novel coronavirus was reported in Wuhan, China [1]. The first case of the virus reported in the United States was on January 22, 2020 [2]. Within 22 days, the Coronavirus had traveled across the Pacific to wreak havoc upon countries woefully unprepared. Within a year, COVID-19 has managed to bring some of the most powerful countries in the world to heel. Economies and healthcare systems across the world continue to be devastated by an adversary only 60 to 140 nanometers in diameter [3]. On February 11, 2020, the International Committee on Taxonomy of Viruses (ICTV) formally identified the virus as Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). On March 11, 2020, the World Health Organization classified COVID-19 as a worldwide pandemic and global health crisis [4]. As of May 2021, the CDC has confirmed that the U.S. has over 32 million cases. Healthcare systems across the nation and around the world are overwhelmed by the number of infected patients. Many of them perish due to either a lack of resources or accurate and efficient testing.
SARS-CoV-2 Viral Pathogenesis
Humans have two levels of immunity. Innate immunity is the body’s first line of contact and defense against invading pathogens. Adaptive immunity learns and remembers how to effectively target and eliminate these pathogens.
Innate Immunity
Our innate immune system is composed of barrier tissues and cells specialized for defense against pathogens [5]. Barrier tissues are the first line of defense, and inside barrier tissues reside sentinel cells, which are capable of consistently recognizing repeated exposure to pathogen associated molecular patterns (PAMPs). The sentinel cells release proinflammatory mediators like cytokines, chemokines, or histamines and circulate within the blood vessels inviting more immune cells from the surrounding tissue into the bloodstream [5]. Cells such as neutrophils or monocytes differentiate into macrophages and migrate from the bloodstream and phagocytose (eat) the pathogens. Neutrophils will undergo programmed cell death, referred to as apoptosis. Macrophages will continue to phagocytose the rest of the pathogens and restore homeostasis by consuming the dead neutrophils [5].
Infection occurs when these viral pathogens in respiratory droplets from a sneeze or a cough enter a person’s mouth, nose, or eyes and attach to the ACE-2 receptors in the nose, throat, and especially the lungs. Like any virus, SARS-CoV-2 cannot replicate on its own and instead hijacks the body’s own cellular machinery. The virus inserts its own genetic information into the host cell to produce more copies of itself until the cell bursts and dies, spreading more of the virus around the body to infect more cells [6]. Infection of the host cell consists of the following five steps: attachment, penetration, biosynthesis, maturation, and release. Once a virus binds to host receptors (attachment), it enters host cells via endocytosis or membrane fusion (penetration). Once the viral contents are released inside the host cells, viral RNA are transported by protein molecules in the host cell’s cytoplasm and travel into the nucleus for replication via the nuclear pore complex (NPC). Viral mRNA then makes viral proteins (biosynthesis). Lastly, novel viral particles are made (maturation) and released [7]. This innate immune response is not as effective against SARS-CoV-2 due to the strength of the various proteins displayed in Figure 1, an ultrastructural morphology rendering, provided by the Centers for Disease Control and Prevention (CDC) Image Library on February 10 [8].
Figure 1
The SARS-CoV-2 virus contains “M (membrane), S (spike), E (envelope), and N (nucleocapsid)” proteins, which envelop the virion and act as a defensive shield [9]. The S or Spike viral surface protein, which consists of two subunits, S1 and S2, binds to the angiotensin converting enzyme 2 (ACE2) receptors of the host cells [7]. The primary role of ACE2 is the breakdown of the angiotensin II (ANG II) protein into molecules that neutralize its harmful effects. ANG II is responsible for increased inflammation and death of alveolar cells in the lungs, which reduces oxygen uptake. When the S (spike) protein of SARS-CoV-2 binds to the ACE2 receptors, they inhibit ACE2 from doing its job of regulating ANG II, allowing ANG II to freely damage tissue in the lungs. These ACE2 receptors are naturally present on the surface of the lung’s epithelial cells and other organs throughout the body, but the virus’ S protein uses these receptors to penetrate the cell membrane and replicate inside host cells. The N (nucleocapsid) protein is another viral surface protein of SARS-CoV-2, which inhibits interferons (IFN1 and IFN-β) responsible for cytokine production [10]. But if the signals for regulating proinflammatory response are disrupted by the pathogen’s surface proteins, the innate immune response becomes hyperactive and self-destructive. A malfunctioning innate immune response also compromises an adequate adaptive immune response [9].
Adaptive Immunity
Adaptive immunity consists of B-cell and T-cell responses. B-cells produce antibodies to trigger an immune response, while T-cells actively target and eliminate infected cells.
B-Cell Response
The innate immune response is not particularly equipped to combat pathogens that are especially complex and vicious because the innate immune response is non-specific and will attack anything it identifies as an invader. The adaptive immune response can target pathogens more precisely and powerfully by using proteins called antibodies, which are produced by B-cell lymphocytes that bind to antigens on the surface of pathogens [5]. Adaptive immunity can more efficiently handle foreign pathogens, like a virus, because antibodies can see through the debris of proteins and dead cells left by the cytokine storm. Antibodies uniquely bind to antigens, acting as a beacon for the adaptive immune response to converge on the invading pathogen [5]. More importantly, adaptive immunity has memory and learns how to become more effective by retaining its response to pathogens so that it can be even quicker at eliminating them after repeated exposure [5]. Widespread pandemics like COVID-19 occur because of a lack of protective antibodies in populations that have never been exposed to or vaccinated against the specificity of SARS-CoV-2 [5]. Figure 2 depicts the four ways in which antibodies attack pathogens: neutralization, complement fixation, opsonization, and antibody dependent cellular cytotoxicity.
Figure 2
Figure 2 – “Immunopathogenesis of Coronavirus Disease 2019 (COVID-19)” [3].
Neutralization is the process by which antibodies immediately bind to the surface antigens of a pathogen and block their S protein from attaching to the receptors of healthy cells, thereby neutralizing the virus’ ability to attach and insert its genetic information. Complement fixation occurs when antibodies are responsible for inviting complement proteins to bind to the antigens of the pathogen. This process coats the pathogen in attack proteins that can either initiate the complement cascade leading to cell lysis, the breakdown of the cell, or it can induce the third stage, opsonization. During opsonization, proteins called opsonins bind to the invading pathogen, acting as markers for phagocytotic cells like macrophages to identify and consume the pathogen. Lastly, antibody dependent cellular cytotoxicity (ADCC) is the process by which antibodies recognize the antigen of a pathogen and signal for natural-killer cells (NK cells) to release cytotoxic molecules which kill off the virally infected cell [5].
T-Cell Response
T-cell lymphocytes are produced by the bone marrow and mature in the thymus. They form the basis of cellular immunity because they directly attack foreign pathogens. Consequently, they are more effective than innate immune or B-cell responses at targeting intracellular pathogens like viruses [5]. Antibodies can get distracted by viral particles and proteins, so they rely on the blind T-cell lymphocytes to ignore the surrounding virus particles and eliminate the infected host cell at the source. As naive T-cells circulate the lymph nodes and spleen, they express T-cell receptors (TCR) that recognize cell surface peptides (antigens) attached to major histocompatibility complex (MHC) molecules on the surface of a specific pathogen. These surface MHC proteins tell the T-cells where to attack [5]. The dendritic cells work to activate the adaptive immune response by ingesting viral proteins and turning them into cell surface peptides that bind to MHC molecules, forming peptide-MHC complexes. The TCR of naive T-cells recognize the peptide-MHC complexes and activate the T-cell. For T-cells to become active, they also need to bind to proteins from the dendritic cell via co-simulation. They then undergo clonal expansion and differentiate into effector T-cells [5]. Effector T-cells are also referred to as cytotoxic T lymphocytes (CTLs). They travel through the body to hunt down peptide-MHC presenting pathogens and kill the infected cells by releasing cytotoxic molecules [5].
The adaptive immune response is stimulated by the recognition of pathogen-associated molecular patterns (PAMPs). Within 1-2 weeks after infection, the B-cells produce antibodies while T-cells simultaneously increase proinflammatory cytotoxic molecules in a forceful attempt to eliminate the virus [7]. The uptick in Interleukin cytokines abbreviated as IL-1, IL-6, IL-8, and so on, flood the body with proinflammatory substances, which “chronically increase the stimulation of T-cells, resulting in a cytokine storm and T-cell exhaustion” [9]. T-cell exhaustion not only means that the virus is overwhelming the body’s antibodies but also draining the strength of the T-cell’s ability to eliminate the virus at the source of infected host cells. SARS-CoV-2 is a “high-grade chronic viral infection because it decreases the responsiveness of T-cells leading to a decreased effector function and lower proliferative capacity” [9]. T-cell exhaustion is also linked to an increase in inhibitory receptors that can initiate apoptosis in T-cells. This results in the destruction of T-cells and their co-receptors, further suppressing the T-cells, as well as B-cells and NK cells, all of which are white blood cells (lymphocytes). Thus, explaining the general lymphopenia (the lack of lymphocytes) observed in severe COVID-19 cases and the increased number of cytokines [9]. Viral entry and attachment to ACE2 receptors trigger a vicious cycle of both innate and adaptive immune responses, mounting an intense attack by secreting proinflammatory substances that invite more lymphocytes to try and kill the virus. This releases more cytokines and chemokines [11]. The downregulation of the ACE2 enzyme results in a cascade of chemical reactions that lead to further inflammation and destruction of cells, weakening and damaging the body’s own immune response.
pathologies of a pandemic:
COVID-19 Autopsies
Once the SARS-CoV-2 attaches to alveolar type II cells, it propagates within the cells. Most viral particles cause apoptosis, releasing more self-replicating pulmonary toxins. Figure 3 displays normal ACE2 receptors located in the type II pneumocytes. Healthy alveoli are unobstructed to allow efficient diffusion of oxygen and carbon dioxide with red blood cells.
Figure 3
Figure 3 – “Type I pneumocytes are very thin in order to mediate gas exchange with the bloodstream (via diffusion). Type II pneumocytes secrete a pulmonary surfactant in order to reduce the surface tension within the alveoli” [12].
In contrast to Figure 3, Figure 4 shows the histopathology of alveolar damage (A) and inflammation (B) of the epithelial cells. As the epithelial cells detach from the alveolar wall the alveoli structures collapse and no longer inflate making it hard for patients with severe cases of COVID to breathe [13]. This results in diffuse alveolar damage with fibrin rich hyaline membranes and hemorrhages in the lungs [13]. The histopathology also detected multinucleated cells that lead to pulmonary fibrosis (scarring in the lungs). Infected cells are “abnormally large and often polynucleated cells showing a large cytoplasm with intense staining for the COVID-19 RNA probe” [13]. The viral Spike protein is also largely detected in the histopathology of COVID cases (C). The nuclei of Spike-positive cells appear an intense red stain and have abnormally enlarged cytoplasts (panel h) [13].
Figure 4
Figure 4 – “Histopathological evidence of alveolar damage, inflammation and SARS-CoV-2 infection in COVID-19 lungs” [13].
The cellular destruction detected in the histopathology is macroscopically reflected in the physical damage of lung tissue displayed in Figure 5.
Figure 5
Figure 5 – “Macroscopic appearance of COVID lungs” [13].
In all pathological examinations of patients that died of COVID, their lungs sustained macroscopic damage [13]. Severe cases of COVID reveal congested and firm lungs (A) with “hemorrhagic areas and loss of air spaces (a’, c’)” [13]. As the virus ravages the body, some patients rapidly deteriorate and develop severe inflammation and clotting in the lungs (B) which shows “the thrombosis of large pulmonary vessels, often with multiple thrombi and in one case determining an extensive infarction in the right lobe (Fig. 5B panels a and b)” [13]. The lung’s architecture crumbles as cells lose their integrity and continue to die, thus resulting in the development of Acute Respiratory Distress (ARDS). ARDS develops in about 5% of COVID-19 patients, and of all the infected, the mortality rate remains around 1 to 2% [14]. Autopsies are beginning to reveal that rather than a singular cause of death, many factors seem to be responsible for higher mortality rates in patients that develop critical cases of COVID-19.
The fallout from the hyperactive immune response disrupts regular oxygen diffusion from the alveoli into the capillaries and consequently to the rest of the body. This commonly leaves fluid and dead cells, resulting in pneumonia, a condition in which patients experience symptoms such as coughing, fever, and rapid or shallow breathing [14]. If oxygen levels in the blood continue to drop, patients rely on breathing assistance by a ventilator to forcefully push oxygen into damaged lungs “riddled with white opacities where black space—air—should be” [14]. The presence of opacities in the lungs indicate the development of pneumonia into ARDS, which was found in the autopsy of a 77-year-old man with a history of comorbidities, including hypertension and the removal of his spleen (splenectomy) [15]. The decedent exhibited chills and an intermittent fever but no cough for 6 days. On March 20, 2020, emergency medical services responded to a call, stating that the deceased was experiencing weakness, fever, and shortness of breath. In route to the hospital, the decedent went into cardiac arrest and died shortly after reaching the hospital [15]. A postmortem nasopharyngeal swab was administered and came back positive for SARS-CoV-2.
|
Figure 6 |
Figure 7 |
 |
 |
| Figure 6 – Normal chest X-Ray of healthy lungs [16]. | Figure 7 – “Lesion segmentation results of three COVID-19 cases displayed using the software post-processing platform” [17]. |
Figure 7 shows opacities in the CT “of typical COVID-19 infection cases at three different infection stages: the early stage, progressive stage, and severe stage” [17]. Figure 7 highlights these opacities in red, which appear to intensify and cover more of the lung CT as the virus increases in severity (a-c). Patient 4 (c) exhibits what medical examiners refer to as a “complete whiteout” of the lungs. Indicating reduced air flow, whereas the normal scan of healthy lungs (Figure 6) has a black background, representing the transparency of free and unrestricted airflow.
The postmortem radiography of the deceased 77-year-old man showed “Diffuse, dense bilateral airspace consolidations (complete “whiteout”)” [15]. In most cases of severe COVID-19 “the greatest severity of CT findings is visible around day 10 after the symptom onset. Acute respiratory distress syndrome is the most common indication for transferring patients with COVID-19 to the ICU” [18].
ARDS in connection to SARS-CoV-2 was first documented in Wuhan, Hubei, China in December 2019 with over 90,000 deaths associated with organ dysfunction, particularly progressive respiratory failure and the formation of blood clots resulting in the highest mortality rates [19]. Another autopsy from Hamburg, Germany conducted on the first 12 documented consecutive cases of COVID-19 related deaths revealed that there was not only profuse alveolar damage in 8 out of the 12 patients but also a high rate of clotting resulting in death. 75% of patients that died were males within an age range of 52 to 87 years and 7 out of 12 patients autopsied (58%) presented venous thromboembolism, as displayed in Figure 7. A pulmonary embolism was the direct cause of death in 4 of the deceased [20].
Figure 8
Figure 8 – “Macroscopic autopsy findings: A. Patchy aspect of the lung surface (case 1). B. Cutting surface of the lung in case 4. C. Pulmonary embolism (case 3). D. Deep venous thrombosis (case 5)” [20].
The formation of clots results in pulmonary vasoconstriction, or the constriction of arteries and halting of blood delivery to the arteries and capillaries in the lungs. Blood cannot travel to the lungs, so oxygen levels drop. As a result, a cytokine storm from our hyperactive immune system occurs, destroying the alveolus and the endothelium and causing clots to form. Smaller clots come together and form a fatal giant blood clot, or the clots can break apart and travel to other parts of the body, causing a blockage and inadequate blood supply to organs or other parts of the body [19]. If the blood supply to fingers, toes, and other extremities is cut off by a clot, it is referred to as ischemia and often results in the amputation of digits and appendages once the flesh begins to die [19].
When SARS-CoV-2 enters the alveolar cells in the lungs via the ACE2 receptors, it can directly attack organs and indirectly cause damage to other organs by triggering a hyperactive immune response (cytokine storm). When the viral particles trigger a cytokine storm, they cause further inflammation of the lungs resulting in plummeting oxygen levels and the formation of blood clots in the arteries (arterial thrombosis).
Conclusion
SARS-CoV-2 is a multi-organ scourge, but it primarily attacks the lung by first attaching its spike protein to the host cell’s ACE2 receptors. This prevents the lungs from regulating their function because it inhibits ANG II protein breakdown, causing increased alveolar damage and inflammation of the lungs. The virion proteins create proinflammatory responses in the innate immune response and compromise an effective adaptive immune response. As the virus progresses the number of neutrophils from the innate immune response increase while the number of helpful lymphocytes (T-cells and B-cells) decrease. The ACE2 receptors overstimulate the innate and adaptive immune response to produce more proinflammatory molecules to eliminate the virus, thus causing more destruction to the body and its immune response. Autopsies of COVID-19 victims show ongoing cellular death and collapse of the respiratory system caused by inflammation and alveolar damage that eventually develop into ARDS. Extreme inflammation induced by the immune response causes difficulties in breathing and clotting in the lungs. Radiography of progressive stages of COVID identify opacities in lung CTs indicating obstructed airways and alveolar deterioration. Postmortem examinations reveal gross destruction of the lung tissue, such as pulmonary artery thrombosis, vasoconstriction, lung infarction, or pulmonary embolism. Progressive organ and respiratory failure and abnormal clotting are all contributing factors to the cause of death in the most severe cases of COVID-19.
SARS-CoV-2 efficiently exploits weaknesses not only within our innate and adaptive immune systems across sex, age, race, and ethnicity, but it also exploits weaknesses within our societies. The etymological origins of Pandemic are rooted in pandēmos , which is Greek for ‘all’ (pan)+ ‘people’ (demos). When simplified, pandemic literally means “all people” but the priorities of leadership across the world reveal that not all people suffer the burden of this pandemic equally. Regarding the United States’ approach to the pandemic, this quote from the Atlantic’s article “Why Some People Get Sicker Than Others” is sufficient; “the damage of disease and a global pandemic is not a mystery. Often, it’s a matter of what societies choose to tolerate. America has empty hotels while people sleep in parking lots. Food is destroyed every day while people go hungry. Americans are forced to endure the physiological stresses of financial catastrophe while corporations are bailed out. With the coronavirus, we do not have vulnerable populations so much as we have vulnerabilities as a population. Our immune system is not strong” [21].
References:
1. Fan, Jingchun, Xiaodong Liu, Weimin Pan, Mark W. Douglas, and Shisan Bao. “Epidemiology of Coronavirus Disease in Gansu Province, China, 2020.” Emerging Infectious Diseases 26, no. 6 (2020): 1257-265. doi:10.3201/eid2606.200251.
2. Stokes, Erin K., Laura D. Zambrano, Kayla N. Anderson, Ellyn P. Marder, Kala M. Raz, Suad El Burai Felix, Yunfeng Tie, and Kathleen E. Fullerton. “Coronavirus Disease 2019 Case Surveillance — United States, January 22–May 30, 2020.” MMWR. Morbidity and Mortality Weekly Report 69, no. 24 (2020): 759-65. doi:10.15585/mmwr.mm6924e2.
3. Wiersinga, W. Joost, Andrew Rhodes, Allen C. Cheng, Sharon J. Peacock, and Hallie C. Prescott. “Pathophysiology, Transmission, Diagnosis, and Treatment of Coronavirus Disease 2019 (COVID-19).” Jama 324, no. 8 (2020): 782. doi:10.1001/jama.2020.12839.
4. “Naming the Coronavirus Disease (COVID-19) and the Virus That Causes It.” World Health Organization. Accessed May 31, 2021. https://www.who.int/emergencies/diseases/novel-coronavirus-2019/technical-guidance/naming-the-coronavirus-disease-(covid-2019)-and-the-virus-that-causes-it.
5. “How The Body Reacts To Viruses – HMX: Harvard Medical School.” HMX | Harvard Medical School. February 19, 2021. Accessed May 31, 2021. https://onlinelearning.hms.harvard.edu/hmx/immunity/.
6. Drexler, Madeline. What You Need to Know about Infectious Disease. Washington, D.C., WA: National Academies Press, 2010.
7. Yuki, Koichi, Miho Fujiogi, and Sophia Koutsogiannaki. “COVID-19 Pathophysiology: A Review.” Clinical Immunology 215 (2020): 108427. doi:10.1016/j.clim.2020.108427.
8. “Image Library.” Centers for Disease Control and Prevention. February 10, 2020. Accessed May 31, 2021. https://www.cdc.gov/media/subtopic/images.htm.
9. Yazdanpanah, Fereshteh, Michael R. Hamblin, and Nima Rezaei. “The Immune System and COVID-19: Friend or Foe?” Life Sciences 256 (2020): 117900. doi:10.1016/j.lfs.2020.117900.
10. Lee, Amanda J., and Ali A. Ashkar. “The Dual Nature of Type I and Type II Interferons.” Frontiers in Immunology 9 (2018). doi:10.3389/fimmu.2018.02061.
11. Scully, Eileen P., Jenna Haverfield, Rebecca L. Ursin, Cara Tannenbaum, and Sabra L. Klein. “Considering How Biological Sex Impacts Immune Responses and COVID-19 Outcomes.” Nature Reviews Immunology 20, no. 7 (2020): 442-47. doi:10.1038/s41577-020-0348-8.
12. Cornell, Brent. “Lung Tissue.” BioNinja. 2016. Accessed May 31, 2021. https://ib.bioninja.com.au/options/option-d-human-physiology/d6-transport-of-respiratory/lung-tissue.html.
13. Bussani, Rossana, Edoardo Schneider, Lorena Zentilin, Chiara Collesi, Hashim Ali, Luca Braga, Maria Concetta Volpe, Andrea Colliva, Fabrizio Zanconati, Giorgio Berlot, Furio Silvestri, Serena Zacchigna, and Mauro Giacca. “Persistence of Viral RNA, Pneumocyte Syncytia and Thrombosis Are Hallmarks of Advanced COVID-19 Pathology.” EBioMedicine 61 (2020): 103104. doi:10.1016/j.ebiom.2020.103104.
14. Wadman, Meredith. “How Does Coronavirus Kill? Clinicians Trace a Ferocious Rampage through the Body, from Brain to Toes.” Science, 2020. doi:10.1126/science.abc3208.
15. Barton, Lisa M., Eric J. Duval, Edana Stroberg, Subha Ghosh, and Sanjay Mukhopadhyay. “COVID-19 Autopsies, Oklahoma, USA.” American Journal of Clinical Pathology 153, no. 6 (2020): 725-33. doi:10.1093/ajcp/aqaa062.
16. Gaillard, Frank. “Normal Chest X-ray: Radiology Case.” Radiopaedia Blog RSS. Accessed May 31, 2021. https://radiopaedia.org/cases/normal-chest-x-ray?lang=us.Case courtesy of Assoc Prof Frank Gaillard, Radiopaedia.org, rID: 8304
17. Wang, Xingrui, Qinglin Che, Xiaoxiao Ji, Xinyi Meng, Lang Zhang, Rongrong Jia, Hairong Lyu, Weixian Bai, Lingjie Tan, and Yanjun Gao. “Correlation between Lung Infection Severity and Clinical Laboratory Indicators in Patients with COVID-19: A Cross-sectional Study Based on Machine Learning.” BMC Infectious Diseases 21, no. 1 (2021). doi:10.1186/s12879-021-05839-9.
18. Salehi, Sana, Aidin Abedi, Sudheer Balakrishnan, and Ali Gholamrezanezhad. “Coronavirus Disease 2019 (COVID-19): A Systematic Review of Imaging Findings in 919 Patients.” American Journal of Roentgenology 215, no. 1 (2020): 87-93. doi:10.2214/ajr.20.23034.
19. Magro, Cynthia, J. Justin Mulvey, David Berlin, Gerard Nuovo, Steven Salvatore, Joanna Harp, Amelia Baxter-Stoltzfus, and Jeffrey Laurence. “Complement Associated Microvascular Injury and Thrombosis in the Pathogenesis of Severe COVID-19 Infection: A Report of Five Cases.” Translational Research 220 (2020): 1-13. doi:10.1016/j.trsl.2020.04.007.
20. Wichmann, Dominic, Jan-Peter Sperhake, Marc Lütgehetmann, Stefan Steurer, Carolin Edler, Axel Heinemann, Fabian Heinrich, Herbert Mushumba, Inga Kniep, Ann Sophie Schröder, Christoph Burdelski, Geraldine De Heer, Axel Nierhaus, Daniel Frings, Susanne Pfefferle, Heinrich Becker, Hanns Bredereke-Wiedling, Andreas De Weerth, Hans-Richard Paschen, Sara Sheikhzadeh-Eggers, Axel Stang, Stefan Schmiedel, Carsten Bokemeyer, Marylyn M. Addo, Martin Aepfelbacher, Klaus Püschel, and Stefan Kluge. “Autopsy Findings and Venous Thromboembolism in Patients With COVID-19.” Annals of Internal Medicine 173, no. 4 (2020): 268-77. doi:10.7326/m20-2003.
21. Hamblin, James. “Why Some People Get Sicker Than Others.” The Atlantic. August 19, 2020. Accessed May 31, 2021. https://www.theatlantic.com/health/archive/2020/04/coronavirus-immune-response/610228/.
Sox10 as a Focal Point for Understanding Schwann Cell Differentiation
By Carly Adamson, Neurobiology, Physiology & Behavior ‘21
Author’s Note: I wrote this literature review for a UWP 104E assignment for which we could pick any science topic that interested us. I chose neural crest cells (NCCs) because they are the research focus of Dr. Crystal Rogers’ developmental biology lab, which I intern for on campus, and have such diverse fates. When I wrote this piece, I had just started my internship, and I wanted to connect the lab’s research to my own interests in the peripheral nervous system. This review explains and connects five key discoveries within the history of NCC development research. My intention was to split my audience into both neurodevelopmental specialists and a broader group of biologists with little background in NCCs. I used the terminology necessary for this specialized analysis while also drawing main conclusions in simpler language. I relate to the reader with metaphorical and illustrative language, as I yearned for such explanations in my own exploration of this complex research topic.
Abstract
Schwann cell (SC) development through neural crest cell (NCC) migration and differentiation is a fascinating and important topic since these cells are critical for nervous system function. On the journey to becoming SCs, some Schwann cell precursors (SCPs) stay in their partially-differentiated state to guide other developing cells and to provide a ready supply of a variety of NCC derivatives whenever needed in development. There is a lot left to understand in this intricate process, including how the timeline of SCP development aligns with other neurodevelopmental processes. This research review focuses on key studies about the network of transcription factors, regulators, and enzymes that take multipotent cells from a central region to the final fate of SC maturity. This review also highlights Sox10, a key transcription factor, as a central point to ground the reader in all other discoveries surrounding SC differentiation.
Keywords: Schwann cells, Schwann cell precursors, Sox10, neural crest cells, neurodevelopment, glial growth factor
Introduction
Neural crest cells (NCCs) are the foundation for a variety of key structures, from pigment cells of the skin to neurons and glia of the periphery. NCCs are multipotent, meaning that they can continue to divide in an undifferentiated state as well as differentiate into a wide range of mature cell types. However, they are also the starting point for many pathologies in abnormal development, including digestive tract abnormalities, motor disabilities, and cancer [1]. These fascinating cells have long been shown to differentiate based on networks of environmental signals, as shown by extensive transplantation experiments [2]. All NCCs originate from a structure aptly named the neural crest (NC), which distinguishes vertebrates from other chordates [3]. There are many gene networks to pattern the differentiation, migration, and maintenance of pluripotency of these cells. NCCs must delaminate from the neural tube and migrate to their target tissues after a process known as epithelial-mesenchymal transition (EMT) to become diverse derivatives such as craniofacial bone, pigment, or neurons in the developing organism. Some NCCs even stay multipotent after migration to establish local stem cell populations [4]. NCCs were first described by Dr. Wilhelm His over 150 years ago, and so much about the cells’ diverse functions has been uncovered since then.
One of these diverse functions is the formation of Schwann cell precursors (SCPs). These NCC derivatives migrate along embryonic nerve fibers, supplying stem cells wherever needed in the embryo [5]. After a whirlwind of developmental signaling pathways and context-dependent regulation, mature Schwann cells (SCs) are created. SCs are a key player in the peripheral nervous system, providing insulation and structural support to nerves [6]. SCs create multilayered fatty structures called myelin sheaths that allow action potentials to quickly conduct along a nerve fiber. While a developing embryo must regulate many cell proliferation and differentiation processes at once, it is important to specifically balance SC myelination and differentiation to efficiently develop peripheral nerves that can relay information to their neighbors and listen to external signals. A failure to balance these processes can lead to motor and sensory disabilities in an individual. In addition, the number of SCs that proliferate must match the number of axons in a one-to-one relationship to properly sort cells for subsequent myelination [1]. This balance requires a hefty molecular team, each with key roles in guiding SCs to maturity.
A main player in SC development is Sox10, a protein that induces NCCs to differentiate into components of the peripheral nervous system. Sox10 is a transcription factor, meaning that it binds a specific DNA sequence to regulate the expression of other genes. Sox10 has been linked to a multitude of developmental processes, including three activities relevant to this review: the formation of the neural crest itself, the formation of the peripheral nervous system, and the complete differentiation process of SCs. Without the expression of the Sox10 gene, no glial cells can form in vivo or in vitro [6]. Since Sox10 plays such key roles in multiple stages of vertebrate development, its expression must be tightly regulated. In order to understand the path from multipotent NCCs to fully-developed SCs that keep the organism alive, one must dive into the complicated web of Sox10 control.
Identification of potential SC development factors
The significance of early work in SC development studies can be sorted into two categories: 1) setting the foundation for NCC isolation techniques and 2) identifying genes for further inquiry into their potential roles in SC development. A publication by Buchstaller et al. in the Journal of Neuroscience describes the genetic methods of expressing fluorescent proteins in mice to identify and isolate NCCs and developing SCs. It is important to obtain pure populations of NCCs to study their development and differentiation, as a clean starting point gives the most accurate results once specific induction factors are applied. These broad methods, along with the genetic protocols of RNA amplification and in situ hybridization, allowed later researchers to study many different NCC lineages. A notable candidate selected from this study is Oct6, a transcription factor that will be further discussed below for its role in SC development [7]. The work of Buchstaller et al. provided the foundation for further investigations into the roles of individual transcription factors and signaling proteins in SC differentiation.
Neuregulin-1/ErbB signaling
A key discovery earlier in neural crest experimentation was a particular environmental factor, glial growth factor (GGF), also known as neuregulin-1, that prevents rat NCCs from differentiating into neuronal cells and instructs them to instead differentiate into glial components, such as SCs. This environmental factor works to inhibit Mash1, a very early and essential marker for neuronal differentiation. By blocking Mash1, neuregulin-1 prohibits developing cells from ever starting down the path to neuronal maturity, suppressing the fate of neurons entirely. This result, along with rigorous experiments designed to replicate this finding, identified GGF as the first factor shown to both promote one NCC fate and suppress another [2]. This study changed the theory of SC differentiation by confirming neuregulin-1 as a key SC regulator and proposing the mechanism behind it.
Later research further investigated the role of neuregulin-1 in SC development by focusing on the protein’s interplay with Sox10. A 2020 study from Yang et al. found that the signaling pathway involving neuregulin-1 and its family of epidermal growth factor receptors, ErbB type, maintains expression of Sox10 in differentiating SCPs [6]. By uncoupling ErbB signaling from SC differentiation via two experimental groups, this study was able to show that the combination of ErbB2 and neuregulin are required to produce SC phenotypes and that neuregulin works by affecting Sox10 expression.
Oct6’s synergy with Sox10
Oct6 is a transcription factor that works synergistically with Sox10 to promote the myelination of SCs [6]. Jagalur et al. used cell culture and cloning methods in rat models to elucidate the role of Oct6 in connecting the regulation networks of the promyelinating SC, which has been paired with a neuron but lacks a complete myelin sheath, to the SC that actively ensheaths axons. The 2011 paper concluded through comparative genome studies that Sox10 proteins pair up to form structures called dimers and bind the Oct6 gene. This interaction creates a greater regulatory complex and allows developing SC populations to respond to environmental cues [8]. This study defines a key regulatory mechanism for timing the onset of SC myelination, which is highly important to neurodevelopment in an individual and acutely affects their prognosis. This understanding of Oct6 provides another piece to the puzzle of SC development: differentiating NCCs must be able to understand cues from pre-existing, mature cells.
Histone deacetylases modify multiple SC factors
Figure 1: Bottom left image shows that the inhibition of HDAC1/2 yields lower Pax3 expression in JoMa1 cells. Image credit: Jacob et al., doi:10.1523/jneurosci.5212-13.2014.
Histone deacetylases are transcriptional regulators that remove acetyl groups from DNA histones to condense chromatin and to decrease DNA interactions with transcription factors. These enzymes can also remove acetyl groups from transcription factors themselves to modulate their activity. These enzymatic activities tie into SC development as demonstrated by Jacob et al.’s discovery in 2014 on the functions of histone deacetylase 1 and 2 (HDAC1/2). This study explains that HDAC1/2 are necessary for the myelination and complete maturation of SCs. These enzymes can be placed into the web of protein interactions that together regulate SC differentiation and myelination. This study used a combination of mouse neural crest explants and colonies, which differentiate into glial components in the presence of neuregulin, from the NCC-derived lineage called JoMa1 to get a more complete picture of transcriptional regulation. Jacob et al. show that HDAC1/2 unwind the tightly-packed DNA of the Pax3 gene to facilitate the expression of Pax3, an important transcription factor that maintains Sox10 [9]. This study produced a major change in theory due to the recognition of HDAC1/2 as induction factors for peripheral glia, including SCs, through their control of lineage-specific transcription factors, like Sox10.
Hippo/YAP/TAZ signaling
Figure 2: Control numbers and myelination of Schwann cells contrasts sharply with TAZ/YAP double knockouts (dcKO) on the far right. Image credit: Deng et al., doi: 10.1038/ncomms15161.
Hippo signaling refers to a pathway that affects cell proliferation and a process of controlled cell death known as apoptosis.It is primarily moderated by three factors: Hippo, TAZ, and YAP. TAZ and YAP proteins activate cell cycle regulators to promote proliferation of SC. The two also work with Sox10 to direct differentiation regulators for myelination [3]. TAZ and YAP were found to regulate SC proliferation via the control of cell cycle regulators and regulate SC myelination through interactions with Sox10. The direct targets of these two proteins are still not fully understood, but immunolabeling results published from Deng et al. in 2017 revealed TAZ and YAP to be necessary for SC proliferation and myelin induction. Single knockouts , which disable the target gene from functioning in an organism, for TAZ or YAP in mice showed how these two proteins can compensate for one another’s expression to produce normal SC phenotypes if only one factor is present. In contrast, double knockout mice showed a dramatic reduction in mature SCs due to decreased Sox10 expression [1]. The research that led to understanding the interplay of TAZ and YAP in this signaling network is key to SC developmental theory because the prognosis of mice without these genes is so bleak. Individuals without the proper number of mature SCs are not viable for long after birth due to severe motor and sensory deficits. These unfortunate phenotypes show the importance of TAZ and YAP in SC development as well as the importance of SCs to vertebrate life.
Further neuregulin-1/ErbB signaling analysis
Research on the neuregulin-1/ErbB signaling mechanism is key to understanding SC differentiation because this signaling makes important decisions early in differentiation that completely change the fate of a young NCC.
A closer look at Shah et al.’s 1994 Cell publication reveals advanced techniques that set an impressive foundation for future NCC studies. In order to understand neuregulin’s role in SC development, the paper aimed to answer the following question: When do NC-derived cells first start responding to neuregulins? Scientists used antibody staining, or immunocytochemistry, to fluorescently label NCCs with key proteins to track SC development as well as neuronal development for contrast. They stained for three proteins: 1) glial fibrillary acidic protein (GFAP) to tag immature SCs, 2) a mature Schwann cell-associated tyrosine kinase, c-Neu, and 3) peripherin to track cells developing into neurons. By analyzing NCC colonies from their early, undifferentiated state, scientists were able to study neuregulin’s instructional role in SC development.
The study found colonies of NCCs grown in neuregulin-1 to have no peripherin staining but intense GFAP staining, indicative of high levels of developing SCs and no developing neurons. This result suggests that the presence of neuregulin guides NCCs towards SC development and away from neuronal development. To confirm this result, scientists tested neuregulin-positive cell colonies for two additional Schwann cell-specific markers: P0 and O4. The presence of these protein markers in the neuregulin-treated colonies precisely identifies the cells as SCs, thus confirming neuregulin’s role in exclusively instructing SC differentiation. To confirm neuregulin’s unique role in promoting SC differentiation while suppressing neuronal differentiation, Shah et al. conducted careful colony analysis to confirm that the control and neuregulin-treated colonies had similar percentages of colony survival. This result shows that neurons were inhibited from ever forming in the first place rather than later degrading in neuregulin-positive media [2]. This experiment used rigorous colony analysis and previously validated protein markers to confirm their hypothesis of neuregulin’s two-pronged effects in directing NCC fate.
Yang et al.’s 2020 publication work builds upon Shah’s 1994 publication to examine the interplay between neuregulin and Sox10. This research is important because the study of individual protein factors is not enough to understand the complex control of a developmental process such as SC differentiation. The scientists isolated three groups of bone marrow mesenchymal stem cells (BMSCs) from mice: a control cell group with standard induction factors, a group in which neuregulin’s main receptor was blocked, and one group that never received any neuregulin. RT-PCR was used to quantify the amount of key transcriptional regulator proteins as a measure of the degree of stem cell differentiation into Schwann-like cells. The most significant results come from the ErbB2-inhibited cells treated with immunofluorescence staining, which exhibited significant decreases in multiple SC markers, large reductions in SC proliferation, and a significant decrease in Sox10 expression. These results show that ErbB2 and neuregulin are required to work together to induce stem cell differentiation into SCs. Further analysis of these results uncovered a positive feedback loop between neuregulin and Sox10, meaning the two factors’ reaction yields high amplification of the signals required to quickly create mature SCs in the developing embryo [6].
Conclusion
Current state of theory
As of this year, the entire path from NCCs to SCs is still not completely understood. It is difficult to create conditions that allow both NCC induction and the maintenance of their undifferentiated state. A second difficulty arises because new proteins are frequently added to a puzzle of cell signaling that is also not yet fully understood. There are only a few transcriptional regulators of NCCs that have been studied in detail, and little is known about the products of effector genes for the migration of NCCs [3]. As established in this review, at least four major cell signaling pathways are described to affect SC development. However, a comprehensive map of how these different pathways communicate and combine to deliver the product of mature SCs is not yet defined. Sox10 continues to be a key factor in SC differentiation, and its regulation proves to be more and more complicated with each new protein factor discovery. Nonetheless, the collective of rigorous science allows clarity to be found bit by bit, and there is great potential in the future of NCC-based therapeutics.
Figure 3: Summary of the molecular interactions between Sox10 and the four protein factors described in this review, guiding NCCs through the process of SC differentiation.
Important questions for future research
A key area of inquiry examines how this research on animal models might translate to regenerative medicine and stem cell-based therapeutics in humans. To be therapeutic for regenerative medicine, induced NCCs need to have as close to normal differentiation and population patterns as possible. This involves future research on functional comparisons between NC-derived stem cells in postnatal organisms and embryonic stem cells [5]. SCs hold a high potential for regenerative medicine due to their natural role in axonal regrowth following peripheral nerve damage [10]. The applications of this research are exciting, but there is still a long way to go in understanding the wide range of applicable protein activities.
Understanding normal patterns of SC development will help develop treatments for abnormal patterns, like SC tumors. Uncontrollable SC differentiation is a known characteristic of some cancers [1]. Recent research has identified a tumor suppressor, Nf2, that leads to Schwannomas and hyperplasia in mouse models when inactivated [3]. In addition, future manipulation of the Hippo signaling described in this article could compensate for myelin insufficiency without risking an overproduction of myelin that may lead to tumors [1]. The wide range of proteins described thus far as regulators ofSox10’s activity demonstrates the importance of continued funding for basic SC research. Finally, the modular content of this review supports the importance of further studies that focus on the interplay between cell signaling pathways to one day obtain a highly detailed, web-like recipe for SC differentiation.
References:
- Deng, Y., Wu, L. M. N., Bai, S., Zhao, C., Wang, H., Wang, J., et al. 2017. “A reciprocal regulatory loop between TAZ/YAP and G-protein Gas regulates Schwann cell proliferation and myelination.” Nat. Commun. 8, 1–15. doi: 10.1038/ncomms15161.
- Shah, N. M., Marchionni, M. A., Isaacs, I., Stroobant, P., & Anderson, D. J. 1994. “Glial Growth Factor Restricts Mammalian Neural Crest Stem Cells to a Glial Fate.” Cell, 77, 349-360.
- Méndez-Maldonado, K., Vega-López, G. A., Aybar, M. J., & Velasco, I. 2020. “Neurogenesis From Neural Crest Cells: Molecular Mechanisms in the Formation of Cranial Nerves and Ganglia.” Frontiers in Cell and Developmental Biology, 8: 1-15. doi:10.3389/fcell.2020.00635.
- Kunisada, T., Tezulka, K., Aoki, H., & Motohashi, T. 2014. “The stemness of neural crest cells and their derivatives.” Birth Defects Research Part C: Embryo Today: Reviews, 102(3), 251-262. doi:10.1002/bdrc.21079
- Perera, S. N., & Kerosuo, L. 2020. “On the road again – establishment and maintenance of stemness in the neural crest from embryo to adulthood.” Stem Cells Journals. doi:https://doi.org/10.1002/stem.3283
- Yang, X., Ji, C., Liu, X., Zheng, C., Zhang, Y., Shen, R., & Zhou, Z. 2020. “The significance of the neuregulin-1/ErbB signaling pathway and its effect on Sox10 expression in the development of terminally differentiated Schwann cells in vitro.” International Journal of Neuroscience, 1-10. doi:10.1080/00207454.2020.1806266
- Buchstaller J, Sommer L, Bodmer M, et al. 2004. “Efficient isolation and gene expression profiling of small numbers of neural crest stem cells and developing Schwann cells.” Journal of Neuroscience. 24: 2357-2365.
- Jagalur NB, Ghazvini M, Mandemakers W, et al. 2011. “Functional dissection of the Oct6 Schwann cell enhancer reveals an essential role for dimeric Sox10 binding.” Journal of Neuroscience.;31(23):8585–8594.
- Jacob, C., Lotscher, P., Engler, S., Baggiolini, A., Tavares, S. V., Brugger, V., Suter, U. 2014. “HDAC1 and HDAC2 Control the Specification of Neural Crest Cells into Peripheral Glia.” Journal of Neuroscience, 34(17), 6112-6122. doi:10.1523/jneurosci.5212-13.2014.
- Nishio, Y., Nishihira, J., Ishibashi, T., Kato, H., & Minami, A. 2002. “Role of Macrophage Migration Inhibitory Factor (MIF) in Peripheral Nerve Regeneration: Anti-MIF Antibody Induces Delay of Nerve Regeneration and the Apoptosis of Schwann Cells.” Molecular Medicine, 8(9), 509-520. doi:10.1007/bf03402160.
After Eureka Comes Death
As insulin prices skyrocket, diabetics turn to increasingly dangerous solutions to manage their illnesses
By Jesse Kireyev, History ‘21
Author’s Note: There’s an indescribable type of heartbreak that comes from hearing a close diabetic family member or friend tell you they cannot afford their next dose and won’t be able to for weeks. A day or two of missed insulin shots could easily end in death as it did for at least one individual mentioned in this article. It’s an especially American experience to be gripped in fear for your loved one’s life because the barebones that they need to survive is out of reach, and despite the relatively high prevalence of diabetes in the population, it’s an issue that’s starkly ignored. Mothers, fathers, siblings, and children wilting away in hospital rooms don’t grab headlines as easily as the latest political trauma, and so too often they get entirely ignored. While I chose to conclude the article on a hopeful note, it’s vital to emphasize that while we wait far too patiently for that hope to materialize, more diabetics die. Many of those who outwaited their hope will never have it again. Americans can’t wait much longer.
Oakland, California, has always been a hub of counterculture. Set in the heart of the Bay Area, Oakland has hosted dozens of America’s most famous hip hop musicians, visual artists, social justice advocates, and tech pioneers. This cultural backdrop has facilitated the creation of the Open Insulin Project: a communal space where diabetics and biohackers meet twice a week to try to create an open source guide that would let type 1 diabetics produce insulin at home.
Diagnosis of this condition was once a death sentence. Prior to the discovery of artificially produced insulin, a child diagnosed with type 1 diabetes had only about a year or two to live after diagnosis [1]. Type 1 diabetes is an autoimmune disease that develops when the immune system starts treating beta cells in the pancreas as a threat and attacks them. The beta cells are responsible for insulin production, so as they are attacked, the pancreas becomes unable to produce insulin. The reasons for why this happens are unknown — current theories suggest environmental and hereditary factors may play a big role, but lifestyle factors do not seem to affect it significantly. This differs from type 2 diabetes, which causes your body to resist the effects of insulin rather than making it unable to produce it [2]. Insulin is a hormone that your body naturally produces to regulate the amount of glucose, or blood sugar, in your system. As type 1 diabetes develops and insulin production stops altogether, the body loses its ability to regulate blood sugar unless insulin is supplemented through injection.
The grim sites of hospital wards full of dying children pushed three scientists at the University of Toronto—Sir Frederick Banting, Charles Best, and JJR Macleod—to discover insulin in 1921. The theory of injecting insulin to regulate blood sugar wasn’t developed until a year later by James Collip, but the first patients treated with insulin injections suffered from severe allergic reactions due to the insulin’s impurity. Collip discovered a way to purify it, putting insulin to use to save children dying from diabetic ketoacidosis—a dangerous diabetic complication that can often end in a coma or death that develops when the body, lacking glucose, begins to produce ketones that acidify the blood. Shortly after discovering and producing insulin, the researchers refused any compensation for their discovery and gave exclusive production rights to a chemical manufacturing firm in Indianapolis, which was to sell insulin at three cents per unit. As a New York Herald article from 1923 put it, Sir Frederick Banting set out to make insulin “available for even the poorest sufferers from diabetes” [3]. Despite Banting’s intentions, the price of insulin has since soared.
The current cost to produce a single vial of human insulin is somewhere in the range of two-and-a-half to three-and-a-half dollars, and a year’s supply of the medicine could be sold to type 1 diabetes patients for as little as seventy-two dollars, a cost that is not only affordable to consumers but still produces a large profit margin for the pharmaceutical companies creating it [4]. And yet insulin prices remain far higher than they theoretically should be. A survey of medical prices conducted by the Health Care Cost Institute showed that in 2016 type 1 diabetics paid an average of $5,705 for insulin, over seventy-nine times the cost of what medical researchers believe they should be paying [5]. American diabetes patients paid an average of $300 for a vial of Humalog, a specific type of insulin, in 2019. That same vial costs just thirty-two dollars in Canada [6]. As a result, more people are trying to fight back against these skyrocketing costs.
Drug development can drain millions of dollars from a pharmaceutical company’s budget, meaning companies have to sell their drugs at high prices to recoup the costs of development. New developments to certain medications can temporarily raise their prices, which is a reason often used by pharmaceutical companies to defend their price increases, but that’s not the case with insulin. Dr. Nicholas Argento said in an interview with Business Insider, “The products that are out are not really new. They may have tweaked the manufacturing process and[…]they have better delivery pens and the like, but the increase in price has been astronomical” [7].
While the base drug itself hasn’t seen significant development in decades, the patent on it still remains, and keeps getting extended. “Drugs are kept on patent by making somewhat fairly small fluctuations or modifications to the particular thing, like insulin,” said Dr. Huising. This effectively prevents the production of a cheaper, generic version of the drug, leaving diabetics to rely on insulin produced by only three suppliers in the United States. Canadian patients are able to buy insulin at lower costs due to increased market price regulation, as well as the fact that Canada allows generic insulin into the market, side-stepping the patent issues that generic insulin would face in the United States. And the patents still have a while to go before expiring: just last year, French pharmaceutical company Sanofi, one of the three insulin suppliers in the United States, got its patent extended to 2031. The medical advocacy non-profit, Initiative for Medicine, Access, and Knowledge, points to this as one of the main culprits behind the price increases, stating in its 2018 report, “The U.S. cannot fix the drug pricing crisis until it solves the drug patent problem” [8].
When comparing the price of insulin produced by each of these three companies, it is clear that each company sells its product at almost the exact same price, and price increases between the three companies have remained almost identical over the past few decades. While the companies deny any collusion or price fixing, by law, pharmaceutical corporations are not required to provide the reasons behind price increases, and they can raise them without limit [9]. The issue has become so dire that in 2018, the Congressional Diabetes Caucus released a report stating that the current system is “unfairly putting insulin out of reach, placing millions of lives at risk” [10].
Placing millions of lives at risk is not an over exaggeration: without enough insulin, blood sugar rapidly increases. A diabetic with high blood sugar runs the immediate risk of developing diabetic ketoacidosis, alongside long-term cardiovascular and nervous system problems, which can significantly shorten lifespans. According to Dr. Kasia Lipska, an assistant professor at the Yale School of Medicine, “About 1 in 5 people with type 2 need insulin to prevent short-term and long-term complications like blindness, kidney failure, and dialysis and heart disease” [11]. For other types of non-gestational diabetes, the risk is more immediate: “Insulin is a life-saving drug, people need it,” says Dr. Huising. “If you have type 1 diabetes and no insulin, you die.”
These problems can emerge from missing just a few doses of insulin, something that many diabetics increasingly have to resort to due to its cost. Lipska et al. published a study in the journal JAMA Internal Medicine that found that over a fourth of diabetes patients have had to cut back on insulin dosage due to the high price of the medicine. The human costs of insulin prices are very real. In a CBS interview, the mother of a man who died because he couldn’t afford his medicine spoke out against the costs. Alec, a young man with diabetes, began to ration his insulin when faced with a $1,300 price tag. Unfortunately, after struggling under a diabetic coma, the lack of insulin cost him his life. “I wanted to be there with him, to hold his hand, or to call for help. And then I think about if he had never moved out, if he had lived at home, somebody would have seen the signs,” she said. “I’ll probably feel guilty every day for the rest of my life” [12].
It is within this context that a number of rogue diabetics in Oakland have begun to try to synthesize their own insulin supply. The Open Insulin Project was started in 2015 by Anthony Di Franco, a type 1 diabetic who struggled for years with being able to buy his insulin. The project utilizes biohacking—a movement that applies do-it-yourself, rule-averse hacker practices to the exploitation of genetic material—to create a homebrew form of insulin for type 1 diabetics. The project’s motives are exceptionally ambitious, given that almost none of the people involved are trained biochemists. Di Franco and other project leaders such as David Anderson lack experience in biological or chemical sciences, and neither work in relevant fields—Di Franco is a computer scientist, and David Anderson is pursuing a degree in business economics [13]. The Project aims to one day create a fully safe and functional form of insulin, and recent developments in the chemical process have shown some promise. Currently, the Project is hoping to convert proinsulin—insulins’ chemical precursor—to full insulin, after which they will attempt to scale their process up [14]. But some have their doubts. Hank Greely, a professor at the Center for Law and Biosciences at Stanford University, warns that “manufacturing pharmaceuticals is difficult, painstaking, and dangerous. If you get the dosing or the strength on the insulin wrong, it’s death. If you let contaminants into the insulin, it’s possible death. If your insulin breaks down too quickly in storage, it’s death” [13].
Dosing is a difficult challenge in a homebrew environment, where biohackers might not be able to access the proper equipment to create safe and stable insulin with consistent doses and without contaminants. Immunogenic reactions—wherein a foreign molecule entering the body provokes an immune system response—predominates the list of concerns over impurities. “You’re injecting something. If there is an impurity there that is a foreign molecule, then your immune system might start to respond,” Dr. Huising said. “Doing it at scale with a quality that is consistent is extremely challenging to do. It’s not that hard to make it if you’re a trained biochemist, but making it with a quality and consistency that is compatible with injecting it as a drug over multiple batches is hard to achieve.” Nor does the Project have access to most of the biochemistry equipment Dr. Huising insists is necessary to create insulin safe for injection. “The motivation behind wanting to make insulin is clear,” he said. “But doing it homebrew style is just dangerous and irresponsible.”
Due to this, the Open Insulin Project may face legal challenges from the FDA and other regulatory agencies, challenges that the Project may not have the money or resources to address. This is not a problem that has escaped the minds of those running the Project, as they currently are trying to figure out the legal issues surrounding human testing and safety of human consumption. That is why the project has been focusing on creating a do-it-yourself guide to synthesizing insulin at home—while the FDA can regulate distribution of medicine; the first amendment stops them from regulating the distribution of a guide on how to make the medicine.
But the situation may soon change enough that the need for the Open Insulin Project will fade away entirely. Over the past few years, the FDA has been pursuing paths to change insulin regulatory procedures, introduce generic insulin to the market, and lower the costs of the drug—policies that the last three presidential administrations have publicly advocated for. Dr. Huising has hope that this public pressure might help insulin prices fall within the next few years—“Even in the past couple of years, there has been talk in Washington about how big pharma does overcharge. I don’t think that’s necessarily a left or a right talking point,” he said. “There is a recipe there for improvement, where we demand that insulin is made available at prices that don’t force people to self censor or self limit how much insulin they dose themselves with.”
References
- Editor. Diabetes history. Diabetes.co.uk. 2019 Jan 15. https://www.diabetes.co.uk/diabetes-history.html
- Causes of type 1 diabetes – JDRF. Jdrf.org. 2017 Oct 17. https://www.jdrf.org/t1d-resources/about/causes/
- Moulton Weekly Tribune. Newspaperarchive.com. https://moultonpl.newspaperarchive.com/moulton-weekly-tribune/1923-12-07/page-8/
- Gotham D, Barber MJ, Hill A. Production costs and potential prices for biosimilars of human insulin and insulin analogues. BMJ global health. 2018;3(5):e000850.
- U.S. insulin costs per patient nearly doubled from 2012 to 2016: study. Reuters. 2019 Jan 22. https://www.reuters.com/article/us-usa-healthcare-diabetes-cost-idUSKCN1PG136
- Goldman B. The soaring cost of insulin. CBC News. 2019 Jan 28. https://www.cbc.ca/radio/whitecoat/blog/the-soaring-cost-of-insulin-1.4995290
- Business Insider. Why insulin is so expensive. 2019 Feb 12. https://www.youtube.com/watch?v=7Ycd8zEdoVk
- I-mak.org. 2018. https://www.i-mak.org/wp-content/uploads/2018/08/I-MAK-Overpatented-Overpriced-Report.pdf
- Thomas K. Drug makers accused of fixing prices on insulin. The New York times. 2017 Jan 30. https://www.nytimes.com/2017/01/30/health/drugmakers-lawsuit-insulin-drugs.html
- Skyrocketing insulin cost: Congressional Diabetes Caucus highlights need and ways to bring prices down. House.gov. 2018 Nov 1. https://diabetescaucus-degette.house.gov/media-center/press-releases/skyrocketing-insulin-cost-congressional-diabetes-caucus-highlights-need
- Adam.com. http://pennstatehershey.adam.com/content.aspx?productId=35&gid=4470
- CBS This Morning. Mother says son died “because he could not afford his insulin.” 2019 Jan 4. https://www.youtube.com/watch?v=Zp_1ohad0Tg
- Osterath B. Deutsche Welle (www. dw.com). 2019. Do-it-yourself insulin: Biohackers aim to counteract skyrocketing prices. https://www.dw.com/en/do-it-yourself-insulin-biohackers-aim-to-counteract-skyrocketing-prices/a-48861257
- Di Franco A. New frontiers for the New Year. Openinsulin.org. 2018 Dec 31. https://openinsulin.org/our-blog/new/
The Scientific Cost of Progression: CAR-T Cell Therapy
By Picasso Vasquez, Genetics and Genomics ‘20
Author’s Note: One of the main goals for my upper division UWP class was to write about a recent scientific discovery. I decided to write about CAR-T cell therapy because this summer I interned at a pharmaceutical company and worked on a project that involved using machine learning to optimize the CAR-T manufacturing process. I think readers would benefit from this article because it talks about a recent development in cancer therapy.
“There’s no precedent for this in cancer medicine.” Dr. Carl June is the director of the Center for Cellular Immunotherapies and the director of the Parker Institute for Cancer Immunotherapy at the University of Pennsylvania. June and his colleagues were the first to use CAR-T, which has since revolutionized personal cancer immunotherapy [1]. “They were like modern-day Lazarus cases,” said Dr. June, referencing the resurrection of Saint Lazarus in the Gospel of John and how it parallels the first two patients to receive CAR-T. CAR-T, or chimeric antigen receptor T-cell, is a novel cancer immunotherapy that uses a person’s own immune system to fight off cancerous cells existing within their body [1].
Last summer, I had the opportunity to venture across the country from Davis, California, to Springhouse, Pennsylvania, where I worked for 12 weeks as a computational biologist. One of the projects I worked on was using machine learning models to improve upon the manufacturing process of CAR-T, with the goal of reducing the cost of the therapy. The manufacturing process begins when T-cells are collected from the hospitalized patient through a process called leukapheresis. In this process, the T-cells are frozen and shipped to the manufacturing facility, such as the one I worked at this summer, where they are then grown up in large bioreactors. On day three, the T-cells are genetically engineered to be selective towards the patient’s cancer by the addition of the chimeric antigen receptor; this process turns the T-cells into CAR-T cells [2]. For the next seven days, the bioengineered T-cells continue to grow and multiply in the bioreactor. On day 10, the T-cells are frozen and shipped back to the hospital where they are injected back into the patient. Over the 10 days prior to receiving the CAR-T cells, the patient is given chemotherapy to prepare their body for inoculation of the immunotherapy [2]. This whole process is very expensive and as Dr. June put it in his TedMed talk, “it can cost up to 150,000 dollars to make the CAR-T cells for each patient.” But the cost does not stop there; when you include the cost of treating other complications, the cost “can reach one million dollars per patient” [1].
The biggest problem with fighting cancer is that cancer cells are the result of normal cells in your body gone wrong. Because cancer cells look so similar to the normal cells, the human body’s natural immune system, which consists of B and T-cells, is unable to discern the difference between them and will be unable to fight off the cancer. The concept underlying CAR-T is to isolate a patient’s T-cells and genetically engineer them to express a protein, called a receptor, that can directly recognize and target the cancer cells [2]. The inclusion of the genetically modified receptor allows the newly created CAR-T cells to bind cancer cells by finding the conjugate antigen to the newly added receptor. Once the bond between receptor and antigen has been formed, the CAR-T cells become cytotoxic and release small molecules that signal the cancer cell to begin apoptosis [3]. Although there has always been drugs that help your body’s T-cells fight cancer, CAR-T breaks the mold by showing great efficacy and selectivity. Dr. June stated “27 out of 30 patients, the first 30 we treated, or 90 percent, had a complete remission after CAR-T cells.” He then goes on to say, “companies often declare success in a cancer trial if 15 percent of the patients had a complete response rate” [1].
As amazing as the results of CAR-T have been, this wonderful success did not happen overnight. According to Dr. June, “CAR T-cell therapies came to us after a 30-year journey, along with a road full of setbacks and surprises.” One of these setbacks is the side effects that result from the delivery of CAR-T cells. When T-cells find their corresponding antigen, in this case the receptor on the cancer cells, they begin to multiply and proliferate at very high levels. For patients who have received the therapy, this is a good sign because the increase in T-cells indicates that the therapy is working. When T-cells rapidly proliferate, they produce molecules called cytokines. Cytokines are small signaling proteins that guide other cells around them on what to do. During CAR-T, the T cells rapidly produce a cytokine called IL-6, or interleukin-6, which induces inflammation, fever, and even organ failure when produced in high amounts [3].
According to Dr. June, the first patient to receive CAR-T had “weeks to live and … already paid for his funeral.” When he was infused with CAR-T, the patient had a high fever and fell comatose for 28 days [1]. When he awoke from his coma, he was examined by doctors and they found that his leukemia had been completely eliminated from his body, meaning that CAR-T had worked. Dr. June reported that “the CAR-T cells had attacked the leukemia … and had dissolved between 2.9 and 7.7 pounds of tumor” [1].
Although the first patients had outstanding success, the doctors still did not know what caused the fevers and organ failures. It was not until the first child to receive CAR-T went through the treatment did they discover the cause of the adverse reaction. Emily Whitehead, at six years old, was the first child to be enrolled in the CAR-T clinical trial [1]. Emily was diagnosed with acute lymphoblastic leukemia (ALL), an advanced, incurable form of leukemia. After she received the infusion of CAR-T, she experienced the same symptoms of the prior patient. “By day three, she was comatose and on life support for kidney failure, lung failure, and coma. Her fever was as high as 106 degrees Fahrenheit for three days. And we didn’t know what was causing those fevers” [1]. While running tests on Emily, the doctors found that there was an upregulation of IL-6 in her blood. Dr. June suggested that they administer Tocilizumab to combat increased IL-6 levels. After contacting Emily’s parents and the review board, Emily was given Tocilizumab and “Within hours after treatment with Tocilizumab, Emily began to improve very rapidly. Twenty-three days after her treatment, she was declared cancer-free. And today, she’s 12 years old and still in remission” [1]. Currently, two versions of CAR-T have been approved by the FDA, Yescarta and Kymriah, which treat diffuse large B-cell lymphoma (DLBCL) and acute lymphoblastic leukemia (ALL) respectively [1].
The whole process is very stressful and time sensitive. This long manufacturing task results in the million-dollar price tag on CAR-T and is why only patients in the worst medical states can receive CAR-T [1]. However, as Dr. June states, “the cost of failure is even worse.” Despite the financial cost and difficult manufacturing process, CAR-T has elevated cancer therapy to a new level and set a new standard of care. However, there is still much work to be done. The current CAR-T drugs have only been shown to be effective against liquid based cancers such as lymphomas and non-effective against solid tumor cancers [4]. Regardless, research into improving the process of CAR-T continues to be done both at the academic level and the industrial level.
References:
- June, Carl. “A ‘living drug’ that could change the way we treat cancer.” TEDMED, Nov. 2018, ted.com/talks/carl_june_a_living_drug_that_could_change_the_way_we_treat_cancer.
- Tyagarajan S, Spencer T, Smith J. 2019. Optimizing CAR-T Cell Manufacturing Processes during Pivotal Clinical Trials. Mol Ther. 16: 136-144.
- Maude SL, Laetch TW, Buechner J, et al. 2018. Tisagenlecleucel in Children and Young Adults with B-Cell Lymphoblastic Leukemia. N Engl J Med. 378: 439-448.
- O’Rourke DM, Nasrallah MP, Desai A, et al. 2017. A single dose of peripherally infused EGFRvIII-directed CAR T cells mediates antigen loss and induces adaptive resistance in patients with recurrent glioblastoma. Sci Transl Med. 9: 399.
CD47-SIRPα Pathway as a Target for Cancer Therapeutics
By: Nicholas Garaffo, Biochemistry and Molecular Biology, 20’
Authors’ Note: I originally wrote this piece for my UWP 104E class Writing in the Science’s, but I have since expanded my topic and complicated my original analysis. Ultimately, I submitted this piece to the Norman J. Lang Prize, was awarded second place, and presented my research to the UC Davis college deans. I chose to focus my literary review on cell signaling pathways because I hope to study such topics in my PhD. This topic has impacted my life personally because my grandmother was diagnosed with non-hodgkin’s lymphoma my freshman year of college. In fact, during this review the drugs she was treated with were mentioned, and the CD47-SIRPa pathway may actually be used to treat such a disease.
Abstract
According to the American Cancer Institute, in 2018, cancer had an estimated 1,735,350 new cases and 609,640 people died in the United States alone1. Like many deadly diseases, cancer has found ways to evade the immune system. Many cancers overexpress CD47—a widely expressed “don’t eat me” signal—which interacts with the immune cells’ signal receptor protein alpha (SIRPα), to prevent programmed cell removal (PrCR) 2. ‘Don’t eat me’ signals are a class of cell surface proteins that tell the immune system the corresponding cell is healthy and performing properly. Recent advances have been made to target the CD47-SIRPα pathway to prevent the antiphagocytic activity seen in many cancers. The scope of this review is limited to two new methods used to inhibit the CD47-SIRPα pathway: anti-CD47 and SIRPα antibodies, and small peptide inhibitors. The antibodies for CD47 have shown effectiveness in clinical trials. Antibody inhibition for CD47 and SIRPα were compared, and SIRPα produced better cell type specific inhibition, but similar on-target healthy cell phagocytosis caused anemia in both trials. Several factors, including degradation and inability to penetrate dense tumors, hinder antibody treatment in all cancer patients; therefore, small peptide inhibitors offer an alternate route for inhibition to occur.
Introduction
PrCR is an efficient and accurate process that clears dead, dying, or infectious cells. Phagocytic macrophages—neutrophils, dendritic cells and monocyte derivatives—perform PrCR, and acts independently of apoptosis—programmed cell rupture. Without such processes apoptosis would release cellular contents, such as proinflammatory signals, into the extracellular space3, 4. Such signals can activate inflammatory responses leading to organ and tissue damage. Cells that are under oxidative stress release chemotactic factors that attract immune cells4. Once the macrophage locates the infected cell, it recognizes the cell through “don’t eat me” or “eat me” ligands to prevent or induce cell engulfment, respectively. The scope of this review is limited to a single signal—receptor interaction between CD47—a widely expressed transmembrane protein5—and signal receptor protein alpha (SIRPα)—a receptor expressed on phagocytic immune cells. CD47 links to SIRPα and acts as a “don’t eat me” signal to prevent cellular phagocytosis.
Macrophages activate specific transcription factors in response to environmental cues. Notch signaling describes the macrophages’ internal protein cascade upon receptor-ligand interactions. The macrophage responds by adjusting its polarization into either phagocytic, categorized as the M1 polarization, or non-phagocytic (M2)6. This is important because a macrophages’ phenotype is environmentally dependent on surrounding cell signals, and plays a critical role in PrCR. Upon binding CD47, SIRPα initiates a signal transduction via src homology-2 domain recruitment, a large protein complex. This complex importantly contains two tyrosine phosphatases: SHP-1 and SHP-2, which both interact with various proteins for signaling. Once activated, SHP-1 propagates a downstream antiphagocytic signal (M2) through an unknown mechanism2, 7, 8. Naturally, this ensures macrophages do not engulf healthy cells. In fact, a single CD47-SIRPα interaction is capable of preventing phagocytosis9.
One mechanism cancer uses to evade the immune system is through the CD47-SIRPα pathway. For cancer to propagate it must: prevent apoptosis, divide rapidly, and evade the immune system11. Many cancers overexpress CD47 and it is hypothesized that CD47 accumulation acts as a camouflage. Since CD47 is sufficient to prevent PrCR of healthy cells, when cancers overexpress this signal they can effectively prevent phagocytic clearance. Therefore, inhibiting the CD47-SIRPα pathway is a favorable route for therapeutics2. Efforts have been made to target CD47 and SIRPα individually through monoclonal antibodies (mAb) and high-affinity small peptides. These methods, coupled with known cancer therapeutics like Rituximab, have been shown to decrease tumor cell density in vitro, in vivo, and in clinical trials14. The main goal here is to assess the potential adverse effects presented in each therapeutic. Major hurdles include the potential for other phagocytic inhibitors, off-target effects, and the lack of long-term effects.
Antibody targeting of CD47 and SIRPα shows inhibition of anti-phagocytic signaling
Antibody targeting of CD47 is an effective therapeutic for specific cancers. Acute myelogenous leukemia (AML) is maintained by self-renewing leukemia stem cells (LSC) which evade phagocytosis through increased CD47 expression2, 4. By targeting CD47, researchers hope to activate a focused immune response against tumor cells. Both, in vitro and in vivo analysis of an anti-CD47 antibody (B6H12.2) in an AML LSC model reported a 3-5 fold increase in phagocytosis compared to macrophages and tumor cells alone12. In contrast, an anti-SIRPα antibody reported an increased phagocytosis only when coupled with trastuzumab—a known breast cancer therapeutic10. This contradiction is important, firstly, because it shows antibodies alone are insufficient to increase phagocytosis. Secondly, it hypothesizes other “don’t eat me” signals continue to inhibit phagocytosis after the CD47-SIRPα has been blocked. Lastly, it shows two alternate ways to inhibit the CD47-SIRPα pathway. The anti-SIRPα antibody is argued as a favored cancer therapeutic because CD47 is widely expressed across cell types. Targeting CD47 may cause unwanted on-target CD47 phagocytosis. Despite this possibility, an in vivo analysis of B6H12.2 reported no additional phagocytic activity even with equivalently coated cells4. However, therapeutic exposure only lasted 14 days and animal models were sacrificed afterwards; therefore, long term effects have not been assessed.
Anti-CD47 antibody development towards human variant
A limitation to antibody therapeutics is inter-species variation. B6H12.2s’ affinity decreased from mice to humans due to CD47 variation. Therefore, a human anti-CD47 antibody (5F9) was produced and grafted to immunoglobulin G4 scaffold (IgG4)13. The resulting antibody (Hu5F9-G4) was tested in vitro for its affinity towards human CD47 and
revealed strong attraction, illustrated by the incredibly small amount of dissociation betweenCD47-SIRPa (KD=1×10-12). Hu5F9-G4 was further tested in cynomolgus monkeys to assess potential toxicity in a human-like model. No serious adverse events were characterized except dose dependent anemia which was expected due to the high CD47 expression on red blood cells and reverted naturally after antibody treatment2. However, using healthy monkeys was a limitation to this study; tumor cell phagocytosis was not assessed in vivo. Furthermore, the toxic effects were only tested in a three week period and no long-term effects were characterized.
Clinical trials for the human CD47 antibody variant
Clinical trials of Hu5F9-G4 antibody coupled with rituximab are currently being conducted. Toxicity and effectiveness were assessed in 22 patients with aggressive and indolent lymphoma (this can be thought of as metastatic and benign cancer, respectively)14. From this sample, 50% had an objective response and 36% had a complete response. Furthermore, by day 28, white and red blood cells had approximately 100% of their CD47 receptors occupied. This is important because blocking all CD47-SIRPα interactions is needed for effective results and, since all cells are not degraded, other signals must be preventing phagocytosis in healthy cells9. As seen in other animal models, dose-dependent anemia was the most common side-effect but normal levels of red blood cells reverted at lower dosages or after the treatment period2. This coupled treatment showed promising results for patients with aggressive and indolent lymphoma.
High-affinity small peptides as an alternate CD47-SIRPα inhibitor
Another issue with antibody therapeutics is their poor permeability into dense tumors15. Given this hurdle, an alternate route is small peptide inhibitors against the CD47-SIRPα pathway. By antagonizing CD47 or SIRPα, the small peptides should block any anti-phagocytic signaling and allow PrCR to occur. Small peptides are highly specific antagonists modeled after invariable regions of their target. By analyzing the human SIRPαs’ binding domain, a competitive antagonist for human CD47 was produced16. The high-affinity SIRPα monomer (CV1) was tested in vitro to assess its affinity towards human and mouse CD47. CV1 presented the same inhibition between human and mouse CD47 variants (50,000-fold affinity increase and KD=34.0 pm). Since small peptides are modeled after invariable regions, their affinities are similar between species. This is important because affinity testing for humans can now be estimated through animal models; thereby, eliminating toxic and costly human trials. Furthermore, ex vivo co-treatment of CV1 with anti-Her2/neu—a well studied breast cancer antibody—increased phagocytosis of human breast cancer cells compared to anti-Her2/neu alone. This coupled treatment was tested in vivo and revealed increased anti-tumor responses in a mouse breast cancer model. Co-treatment illustrates the possibility for more “don’t-eat-me” signals present on cancer cells. Despite CV1s’ efficacy, its high affinity caused on-target CD47 binding across all cell types. Although this high-affinity is wanted in therapeutics, unwanted red blood cell phagocytosis occurred and resulted in anemia. This side-effect, however, is common between all CD47 inhibitors and naturally reverted after treatment16,17.
A solution to CD47 on-target side-effects is antagonizing SIRPα instead. CD47 is expressed widely across cell lines, while SIRPα is present on a subset of macrophages; therefore, SIRPα is arguably the favored target for cancer therapeutics10. One potential SIRPα antagonist, which showed similar potency as CV1, is Velcro-CD47– a high-affinity CD47 variant synthesized through a novel protein “velcro” technique9. Through in vitro analysis, Velcro-CD47 enhanced mAb-mediated phagocytosis by inhibiting anti-phagocytic signals. It is important to note that the small peptide inhibitors do not, by themselves, promote phagocytosis. While antibodies illicit a targeted immune response, small peptides rely on the immune systems’ natural clearance or other cancer therapeutics to clear cancer cells.
Other small peptide therapeutics for CD47-SIRPα inhibition include 4N1K and its derivative PKHB1. There has been substantial evidence that 4N1K increases PrCR in vivo18,19,20,21. Several papers highlight a difference between CD47 +/+ and CD47 -/- tumor cells removal upon 4N1K treatment22. Unlike B6H12/Hu5F9, 4N1K is able to potentiate PrCR of chronic lymphocytic leukemia (CLL) in soluble conditions; however, in human serum, 4N1K is degraded by proteases faster than antibodies-more than 90% was degraded in an 1-hour incubation18. This therapy, therefore, requires more injections for an accurate response. Furthermore, 4N1K has conflicting evidence for its CD47 specificity, and may cause off-target effects23. In order to combat these issues, two terminal residues were replaced on 4N1K with their D analogues. This new therapeutic, PKHB1, lasted longer in human serum, maintained its solubility, and continued to bind CD47. PKHB1 was then tested in vivo and showed higher rates of CLL PrCR 18. PKHB1 is currently in pre-clinical trials for CLL treatment.
Conclusion
Cancer therapeutics continue to progress towards more accurate and less toxic forms. In turn, this eliminates the need for deleterious options like chemotherapy. CD47-SIRPα presents a target for future immunological therapeutics. Although anemia and off-target effects must be further assessed, CD47-SIRPα inhibitors present a feasible and effective option.
Anti-CD47 antibodies increase tumor cell phagocytosis in a coupled therapy with Rituximab and show accurate responses in Phase I clinical trials14. In response to phagocytosis of healthy CD47-expressing cells, anti-SIRPα antibodies have been developed which illustrated similar phagocytic responses in vivo with higher cell type specificity10. Regardless of the target, blocking the CD47-SIRPα pathways still cause anemia in patients. This is an expected and treatable side-effect that naturally reverts after a short-term treatment. No long-term effects of antibody treatments have been assessed and remains a limitation to these studies. To combat the limitations seen in antibody treatments, small peptide inhibitors are being developed for the CD47-SIRPα pathway. Velcro-CD47 presented a novel protein manufacturing technique and provided a high-affinity peptide to prevent inhibitor signals16. 4N1K has been shown to increase tumor cell phagocytosis between CD47 +/+ and CD47-/- but small peptide inhibitors are hindered by their short half-life in blood serum due to protease activity.
By blocking the CD47-SIRPα pathway and other inhibitor signals, researchers can trigger a natural immune clearance of cancer cells. Although differences between CD47 and SIRPα therapeutics, long-term effects, and 4N1K off-target effects must be further assessed, preliminary research indicates this pathway as a potential target for future therapeutics.
Reference:
- “Cancer Statistics”. National Cancer Institute, 2019, https://www.cancer.gov/about-cancer/understanding/statistics.
- Oldenborg PA, Zheleznyak A, Fang Y, Lagenaur CF, Gresham HD, Lindberg FP. “Role of CD47 as a Marker of Self on Red Blood Cells.” Science. 2001;288: 2051- 54
- Lagasse E, and Weissman IL. “bcl-2 inhibits apoptosis of neutrophils but not their engulfment by macrophages.” J Exp Med. 1994 Mar 1;179(3):1047-52.
- Chao MP, Alizadeh AA, Tang C, Jan M, Weissman- Tsukamoto R, Zhao F, Park CY, Weissman IL, Majeti R. “Therapeutic antibody targeting of CD47 eliminates human acute lymphoblastic leukemia.” Cancer Res. 2011 Feb 15; 71(4): 1374-84.
- Brown EJ, and Frazier WA. “Integrin-associated protein (CD47) and its ligands.” Trends Cell Biol. 2001 Mar;11(3): 130-5.
- Alvey C, and Discher DE. “Engineering macrophages to eat cancer: from “marker of self” CD47 and phagocytosis to differentiation.” J Leukoc Biol. 2017;102: 31–40.
- Barclay AN, and Brown MH. “The SIRP family of receptors and immune regulation.” Nat Rev Immunol. 2006 Jun;6(6): 457-64
- Lin Y, Zhao JL, Zheng QJ, Jiang X, Tian J, Liang SQ, Guo HW, Qin HY, Liang YM, Han H. “Notch Signaling Modulates Macrophage Polarization and Phagocytosis Through Direct Suppression of Signal Regulatory Protein α Expression.” Front Immunol. 2018 July 30.
- Ho CC, Guo N, Sockolosky JT, Ring AM, Weiskopf K, Ozkan E, Mori Y, Weissman IL, Garcia KC. “‘Velcro’ engineering of high affinity CD47 ectodomain as signal regulatory protein α (SIRPα) antagonists that enhance antibody-dependent cellular phagocytosis.” J Biol Chem. 2015 May 15;290(20): 12650-63.
- Zhao XW et al. “CD47-signal regulatory protein-α (SIRPα) interactions form a barrier for antibody-mediated tumor cell destruction.” Proc Natl Acad Sci USA. 2011 Nov 8;108(45): 18342-7.
- Ottaviano M, De Placido S, Ascierto PA. “Recent success and limitations of immune checkpoint inhibitors for cancer: a lesson from melanoma.” Virchows Arch. 2019 Feb 12.
- Majeti R, Chao MP, Alizadeh AA, Pang WW, Siddhartha J, Gibbs Jr. KD, Rooijen N, and Weissman IL. “CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells.” Cell. 2009 July 13;138(2): 286-299.
- Liu J et al. “Pre-Clinical Development of a Humanized Anti-CD47 Antibody with Anti-Cancer Therapeutic Potential.” PLoS ONE. 2015.
- Advani R et al. “CD47 Blockade by Hu5F9-G4 and Rituximab in Non-Hodgkin’s Lymphoma.” N Engl J Med. 2018 Nov 1;378(18):1711-21.
- Chames P, Van Regenmortel M, Weiss E, Baty D. “Therapeutic antibodies: successes, limitations and hopes for the future.” Br J Pharmacol. 2009 May;157(2): 220-33.
- Weiskopf K et al. “Engineered SIRPα variants as immunotherapeutic adjuvants to anti-cancer antibodies.” Science. 2013 July 5;341(6141).
- Willingham SB, et al. “The CD47-signal regulatory protein alpha (SIRPα) interaction is a therapeutic target for human solid tumors.” Proc Natl Acad Sci. 2012;109:6662.
- Martinez-Torres AC, Quiney C, Attout T, et al. “CD47 agonist peptides induce programmed cell death in refractory chronic lymphocytic leukemia B cells via PLCγ1 activation: evidence from mice and humans.” PLoS Med. 2015;12(3): e1001796.
- Soto-Pantoja DR, et al. “Therapeutic opportunities for targeting the ubiquitous cell surface receptor CD47.” Expert Opin Ther Targets. 2013 Jan;17(1):89-103.
- Kanda S, Shono T, Tomasini-Johansson B, Klint P, Saito Y. “Role of Thrombospondin-1-Derived Peptide, 4N1K, in FGF-2-Induced Angiogenesis.” Exp Cell Res. 1999;252, 262–272.
- Kalas W et al. “Thrombospondin-1 receptor mediates autophagy of RAS-expressing cancer cells and triggers tumour growth inhibition.” Anticancer Res. 2013;33(4):1429-38.
- Fujimoto T.-T., Katsutani S., Shimomura T., Fujimura K. “Thrombospondin-bound integrin-associated protein (CD47) physically and functionally modifies integrin alphaIIbbeta3 by its extracellular domain.” J Biol Chem. 2013;278 26655–26665
- Jeanne A, Schneider C, Martiny L, Dedieu S. “Original insights on thrombospondin-1-related antireceptor strategies in cancer.” Front Pharmacol. 2015;6: 252.
Finding a Solution in the Source: Exploring the Potential for Early Beta Cell Proliferation to Disrupt Autoreactive Tendencies in a Type 1 Diabetes Model
By Reshma Kolala, Biochemistry & Molecular Biology ‘22
Residing in the pancreas are clusters of specialized cells, namely alpha, beta (), and delta cells. cells, more specifically, are insulin-secreting cells that are instrumental in the body’s glucose regulation mechanism. An elevation of the extracellular glucose concentrations allows glucose to enter cells via GLUT2 transporters, where it is subsequently metabolized. The resultant increase in ATP catalyzes the opening of voltage-gated Ca2+ channels, triggering the depolarization of the plasma membrane which in turn stimulates insulin release by cells (1). In individuals with Type 1 Diabetes (T1D), however, pancreatic islet beta cells are damaged by pro-inflammatory cytokines that are released by the body’s own immune cells. The loss of functional beta cell mass induces a dangerous dysregulation of glucose levels, resulting in hyperglycemia along with other harmful side effects. The absence of a regulatory factor in the bloodstream forces those with T1D to take insulin intravenously to remedy the consequences.
A new study led by Dr. Ercument Dirice, a Harvard Medical School (HMS) instructor and research associate at the Joslin Diabetes Center. has suggested that an increase in cell mass early in life diminishes the autoreactive behavior of immune cells towards cells, therefore halting the development of T1D (2). In a typical T1D model, the secretion of antigens from cells induces a response from the body’s immune cells. These immune cells bind to the epitopes (the recognizable portion of an antigen) that are displayed on the surface of professional antigen presenting cells (APC’s) which are littered throughout the pancreatic islets (3). This binding action induces a destructive autoimmune response to antigens secreted by cells, resulting in loss of functional beta-cell mass. It was found however that by increasing cell mass at an early age where the organs of the immune system are still developing, the immune cells stopped attacking cells.
The novel approach presented by Dirice et al. departs from the traditional method of targeting various other components involved in the destructive autoimmune response, namely APC’s or the pro-inflammatory cytokines associated with T1D progression. This method instead focuses on the source of the autoimmune marked “pathogenic” antigens, the cells themselves.
The studies were completed using two models of female non-obese diabetic (NOD) mice. One was a genetically engineered model of female mice (NOD-LIRKO) that showed increased cell growth soon after birth while the second model was done using a live mouse that was injected at an early age with an agent known to increase cell proliferation. While maintaining more than 99.5% isogenicity (4), it was found that the mice with increased cell mass had a significantly lower predisposition to develop diabetes when compared against the NOD control mice, which developed severe diabetes between 20-35 weeks of age. The study also observed the interaction between the modified cells and immune cells by monitoring the concentration of these immune cells in the spleen. In doing this, researchers were able to conclude which mice had a greater risk of developing T1D based on if mice had an abnormal increase in the concentration of these cells.
At first glance, this method appears counterintuitive as an increase in cell mass may lead one to naively assume that this would result in increased autoantigen production. This precise hypothesis illustrates the beauty of this approach. Although the specific details of this mechanism have yet to be made clear, it is believed that the rapid turnover of cells “confuses” the autoimmune reaction. The proliferated cells present unusual autoantibodies that are not observed in typical T1D progression. Dr. Rohit Kulkarni, a fellow HMS professor and researcher at Joslin noted that there is thought to be some alteration in the new cells where the autoantigens typically produced are reduced or dilated (2). As a result of the slow presentation of antigens, there is a lower proportion of autoreactive immune cells. This essentially results in a “reshapen immune profile that specifically protects cells from being targeted.” Some degree of autoimmunity would continue to exist in the body, so further immunosuppressive treatment would be required.
Early cell proliferation has been previously speculated to have a protective effect in those with reduced functional cell mass as in a Type 1 Diabetes model. Once this preventative quality is better understood, applications of this research may be further explored. Despite still being in the beginning stages, this novel approach holds tremendous potential for application to T1D if this method is able to be translated to a human model. The massive prevalence of a T1D diagnosis is illustrated by 2014 census data that states that T1D affects roughly 4.7% of the world’s adult population. Although extensive research continues to be done on several aspects of the disease, the introduction of new data by Dirice et al. may push us a small step closer to solving one of the body’s greatest metabolic mysteries.
References
- Komatsu, M., Takei, M., Ishii, H., & Sato, Y. (2013, November 27). Glucose-stimulated insulin secretion: A newer perspective. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4020243/
- Yoon, J., & Jun, H. (2005). Autoimmune destruction of pancreatic beta cells. Retrieved from https://www.ncbi.nlm.nih.gov/pubmed/16280652
- Pushing early beta-cell proliferation can halt autoimmune attack in type 1 diabetes model. (2019, May 06). Retrieved from https://www.sciencedaily.com/releases/2019/05/190506124102.htm
- Burrack, A. L., Martinov, T., & Fife, B. T. (2017, December 05). T Cell-Mediated Beta Cell Destruction: Autoimmunity and Alloimmunity in the Context of Type 1 Diabetes. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5723426/
- Dirice, E., Kahraman, S., Jesus, D. F., Ouaamari, A. E., Basile, G., Baker, R. L., . . . Kulkarni, R. N. (2019, May 06). Increased β-cell proliferation before immune cell invasion prevents progression of type 1 diabetes. Retrieved from https://www.nature.com/articles/s42255-019-0061-8?_ga=2.76180373.1669397493.1557550910-1092251988.1557550910