Organs On Chips Will Make Drugs Safer—And More Precise
By Deborah Borfitz
November 21, 2019 | The origins of the organs-on-chips industry can be traced back more than three decades to when Cornell University biomedical engineering professor Michael Shuler set out to make a pharmacokinetic model of the human body in cell culture to expedite drug discovery and toxicology studies. In 2001, a small paper was published about the microfabrication of an animal-on-a-chip device for this process, Shuler says, followed by a series of papers on multiorgan microscale systems (aka “body-on-a-chip”) that effectively marked the birth of a new industry.
But it wasn’t until 2012 that the field started getting serious attention, says Vanderbilt University professor John Wikswo, a biological physicist and founding director of the Vanderbilt Institute for Integrative Biosystems Research and Education (VIIBRE). That’s the year the Defense Advanced Research Projects Agency (DARPA), the U.S. Food and Drug Administration (FDA), and the National Institutes of Health (NIH) partnered to form a collaboration to focus on advancing the development of organ-on-a-chip technology and launch the national tissue chips programs.
The main NIH tissue chips program is run by the National Center for Advancing Translational Sciences (NCATS). The projects of Shuler and Wikswo were among the many funded during the initial NIH tissue chips program, in parallel with the DARPA-funded work at Harvard University and the Massachusetts Institute of Technology (MIT), says Wikswo. The NCATS focus includes the development of tissue chip models for rare diseases that are difficult if not impossible to study in humans and may have no realistic animal model.
The most important catalyst for federal-level attention was a breathing lung-on-a-chip system developed by Donald Ingber, an accomplished scientist at Harvard University, together with his then post-doctoral trainee Dan Huh, as an alternative to existing drug-testing models, says Wikswo. It was heralded as one of the most important research advances in biological and medical science of 2010.
The lung model was the size of a rubber eraser and fabricated out of soft and stretchy material to effectively model the alveolus, says Wikswo. “They showed that you could put epithelial cells that line the inside of an alveolus on one side of a porous membrane, and vascular endothelial cells on the other side, stretch the membrane to mimic breathing, and record beautiful movies of an immune cell crawling down the vascular side, crossing the membrane through the holes in the barrier, and tracking down a bacterium.”
A significant influx of money from the NIH and DARPA gave microphysiological systems a sense of legitimacy, says Wikswo, and the rate of publishing on the subject skyrocketed from a handful of scientific papers annually to about 1,000 last year. Wikswo has been a thought leader on microphysiological systems from the beginning, and his special interest is in functional coupling of organ-on-a-chip devices involving the sequential transfer of biofluid from one to the next.
Launch of the NCATS tissue chip program was exciting news to scientists at the University of Washington’s (UW’s) Kidney Research Institute, who were all too familiar with the difficulties in making new therapies for kidney diseases, says Director Jonathan Himmelfarb. “The kidney is a very complicated organ with more cell types [40-plus] and complex architecture than the brain.”
Every cell in the fist-sized organ is governed to some extent by fluidics—700 liters of blood flow through the kidneys every day, making it the most vascularized organ in the human body, he adds. Its complex filtering apparatus processes about 140 liters of water daily. The tubular epithelium reabsorb 99% of the water and the rest gets excreted as urine. Sensors, receptors, and channels within the kidneys help maintain the blood in a steady state, Himmelfarb says, “so when they’re not functioning well every organ in the body does not function well.”
Two-dimensional in vitro assays are poorly representative of in vivo physiology for the kidneys, Himmelfarb continues, because they lack fluidics and the ability to induce important flow and shear stresses. Blood vessels account for 25% of the kidneys by volume and their vasculature system is highly complicated.
UW researchers were competitively selected to be part of the first two rounds of the NCATS tissue chip program and, earlier this year, also sent tissue chips to the International Space Station, Himmelfarb says.
Commercializing the Technology
“DARPA has been legendary for getting people past the Valley of Death [between academia and commerce],” says Wikswo, including developers of the internet, stealth aircraft, drones, and autonomous vehicles. Often, one of the DARPA program requirements is that investigators have a plan for commercializing their technology.
To that end, a Harvard group created a company called Emulate and an MIT group licensed their technology to UK-based CN Bio Innovations. Wikswo has licensed his MultiWell MicroFormulator to CN Bio Innovations, and a commercial product is expected in early 2020.
MIT succeeded in getting 10 coupled organs working together, which was DARPA’s initial plan, Wikswo says, and a paper was published describing the absorption, distribution, metabolism, and excretion toxicity of pharmaceutical compounds in the different organs. Anyone interested in toxicology should be interested in the multi-organ effects of a drug that organs-on-chips make possible. “It’s going to be hard to do that with organoids because they are so small.”
Drug toxicity was the initial emphasis of DARPA, which appreciated the public and scientific concern over drugs like GlaxoSmithKline’s antidiabetic drug Avandia, which had been implicated in over 100,000 cardiac fatalities after 10 years of clinical use, Wikswo says. As was extensively discussed in a 2017 perspective piece in Experimental Biology and Medicine (DOI: 10.1177/1535370217732765) he authored with his colleagues at Eli Lilly and the NIH, microphysiological devices can help drug developers answer a long list of questions related to pharmacology as well as efficacy.
Flurry of Activity
The emergence of organ-on-a-chip systems parallels a shift away from the study, in a plastic dish, of primary human cells or immortalized cell lines—including the cells famously taken from the cervical tumor of Henrietta Lacks nearly 70 years ago—to a “hopefully much more realistic recapitulation of 3-dimensional biology,” Wikswo says. Some of the more advanced platforms incorporate blood vessel networks that ensure survival of the surrounding tissue.
An organ-on-a-chip system can be constructed in about six hours by a skilled technician, backed by a lot of infrastructure, says Wikswo. A drug company can expect to spend almost $500 per device for one experiment, but that compares to an easy $5,000 to $10,000 for each statistically sound test involving a rat and scientist.
A 2017 review (DOI: c6lc01554a) of microphysiological systems identified over 20 companies working in this general space, says Shuler, a figure that has “probably increased to well over 50 by now.” Shuler is president and CEO of Hesperos, one of few companies that concentrates on multiorgan systems and the only one that operates much like a contract research organization. “Companies send us their drugs and we do the test for them.”
The single-organ systems under development couldn’t be more different. Among the latest in the news are an organ-on-an-electronic-chip platform that shrink-wraps sensors around heart cells to extract electrophysiological information from tissue, a cartilage-on-a-chip system the size of a coin that mimics osteoarthritis and a pancreas-on-a-chip that delivers synchronized glucose pulses to help scientists study and develop effective treatments for diabetes.
A heart-on-a-chip system engineered by TARA Biosystems has also demonstrated its ability to replicate drug responses found in humans. Another heart-on-a-chip, developed by Wikswo’s group, allows application of controlled forces to an engineered cardiac tissue and fitting of a numerical model of cardiac contraction. And Penn Engineering researchers came up with a blinking eye-on-a-chip that is expected to improve understanding and treatment of dry eye disease.
The ability of an organ-on-a-chip system to study the blood-brain barrier (BBB) is a very big deal, says Wikswo. For example, anti-epilepsy drugs and chemotherapeutics targeting cancer metastasized to the brain get met by transport proteins in the BBB that haul the drug out of the epithelial cells as fast as it enters, making the body resistant to treatment, Wikswo says.
VIIBRE researchers have had “very good luck” building brain-on-a-chip barriers, having made substantial progress over the past seven years, says Wikswo. VIIBRE’s NeuroVascular Unit (NVU) seeds two chambers with four different cell types, which are grown in opposition to one another to form the BBB within the NVU. The advantage of their NVU is that it has enough cells for untargeted metabolic activity—resulting from inflammatory changes that affect the BBB—to be measured with a mass spectrometer.
A number of papers about gut-on-a-chip models, most notably one by Dan Huh and Don Ingber of Harvard, have shown that intestinal epithelial cells could serve as a barrier to microbes living in the gut. The models are becoming more realistic and will ultimately model different regions of the intestinal track, says Wikswo.
A more holistic view of drug effects is being made possible by connected organ-on-a-chip systems.
Wikswo and his collaborators at the University of Pittsburgh, the University of Washington, Johns Hopkins University and the Baylor College of Medicine have looked at the way organs interact when they are functionally coupled by transporting media from a gut-chip (Hopkins-Baylor) to a liver chip (Pittsburgh), a kidney proximal-tubule chip (Washington) and the Vanderbilt NVU.
Their study (DOI: 10.1038/srep42296) showed that the organ-specific processing of vitamin C and a withdrawn antihistamine (terfenadine) was consistent with clinical data, demonstrating the potential use of coupled organs as toxicity models, says Wikswo.
To simulate the metabolic activity of microbes in the intestinal microbiome, which convert the amino acids choline and carnitine in food to the chemical tri-methyl amine (TMA), the researchers added TMA to the lumen of the intestine chip, he explains. The media from the vascular side of the intestine chip, which contained TMA transported across the epithelial barrier, was delivered to the liver chip where the TMA was metabolized into trimethylamine N-oxide (TMAO), a predictive marker for chronic renal disease, atherosclerosis and heart failure.
The liver effluent, with both TMA and TMAO, was delivered to the vascular side of both the kidney and NVU chips, Wikswo continues. Studies of media from the organ side of these two chips showed that the kidney eliminated both TMA and TMAO, as expected from the medical literature. Whether TMAO penetrated the BBB was unknown.
“The most exciting result from this study was that the NVU predicted that TMAO would easily cross the blood-brain barrier,” Wikswo says. Researchers in Italy subsequently found that TMAO was detectable in the human cerebrospinal fluid of more than 50 human subjects. This might be the first prediction made using tissue chips that was subsequently confirmed in clinical measurements on humans, he adds.
For its part of the experiment, UW’s Kidney Research Institute used a microphysiology platform commercially available from Nortis that spun out of UW engineering a decade ago, says Himmelfarb. Nortis manufactures a credit card-sized chip ideal for bioengineering kidney components, including the renal tubule and microvasculature, in a space of 70 nanoliters—one-thousandths the size of a water droplet.
The tubules being created are about 120 microns across, relatively close to the 60- to 80-micron size of a normal human tubule, and flow rates mimic tubules in vivo at about half a microliter per minute, Himmelfarb says. Shear stresses are similarly well matched. “We’ve shown in a number of papers that we can recapitulate the function and diseases of the proximal tubule... the most metabolically active part of the kidney.”
A functional rather than a physical coupling experiment was pursued as a matter of convenience, says Himmelfarb. Rather than try to assemble all the organ chips in one location, the biofluid was shuttled from university to university as it would be sequentially processed by organs in the body.
The Kidney Research Institute at Washington has on its own physically coupled a kidney and liver on the same chip to provide a mechanistic understanding of nephrotoxins and drugs that are injurious to the kidney, says Himmelfarb, noting that animal models often don’t mimic results seen in humans. UW researchers were also the first to successfully isolate endothelial cells from human kidneys and propagate them in vitro to better understand their structure and function and show that their phenotypic characteristics are different from endothelial cells elsewhere in the body.
To continue the functional coupling experiment, VIIBRE researchers plan to create coupled tissue chips, linked via surrogate blood, to study the direct transfer of chemical information from one organ to another, Wikswo says. “We believe we can build a coupled gut-liver-brain system to look at the transfer and modification of opioids from the gut through the liver to the brain, as well as the under-appreciated effect of inflammatory agents generated by the gut in response to opioids on both the liver and brain. We have already shown that cytokines released by cells can open up the blood-brain barrier, which means stuff can get in that we want to keep out and vice versa.”
What the Vanderbilt team wants to do, Wikswo says, is “test in vitro a hypothesis that some of the chronic pain that leads to opioid abuse is in fact due to inflammation of the gastrointestinal tract triggered by the opioids themselves.” For unknown reasons, opioids more effectively treat chronic pain and cause fewer problems when given with nonsteroidal anti-inflammatory drugs.
Literature on the microbiome is 10 times more voluminous than that on microphysiological systems and doubling every 3.3 years, surpassing environmental toxicology, says Wikswo, and “organ chips are ideal for studying the microbiome.” Donald Ingber has also been a leading voice in this arena.
“I think organs-on-chips will be extremely important in studying the microbiome, and not just in the gut,” Wikswo says. The vaginal microbiome is thought to play a role in premature delivery, for instance. And the skin microbiome varies from the right hand to the left—and can be passed around the office with a few handshakes, he notes.
Medicines tailored to individuals will not be possible until science untangles more of the complexities and mysteries surrounding the human microbiome, Wikswo says, which will be difficult. “The microbiome readily detected in the stool is different from the microbiome that is harder to measure in the mucus of the colon, which is different from the microbiome in the three regions of your small intestine,” he cites as an example.
“The human genome has around 21,000 genes and about 3,000 of them have been deemed druggable,” says Wikswo. “Of those, only 700 have FDA-approved drugs targeting them. And the microbiome, depending on how you count proteins, has at least three million genes and possibly 100 million.”
Pharmaceutical companies are probably both excited by the prospects and alarmed by the complexity and expense, he adds. But given the microbial mismatch between mice and humans, organ-on-a-chip technology appears to be a potentially important tool in the moonshot goal of precision medicine.
Wikswo says he is convinced that tissue and organ chips will be keeping scientists and engineers busy for decades unraveling some of the hardest problems in medicine—contributing to a “renaissance in human physiology and systems biology.”