By Christina Hernandez Sherwood
You can imagine the scene as an older gentleman lifts a thick, creamy envelope from his mailbox, seeing his own name written in richly scripted lettering. He beams with pride and gratitude at the sight of his granddaughter’s wedding invitation. Yet his next thought is a sober and serious one. Would he be taking his life in his hands by attending the ceremony?
This grandfather lives with primary progressive multiple sclerosis (MS), an autoimmune disorder that he controls with a medicine that depletes his body of the type of immune cells that make antibodies. So while he has completed his COVID-19 vaccine course, his immune system function isn’t very strong—and the invitation has arrived at a time when COVID-19 is still spreading rapidly.
“In the past, all we could do was [measure] the antibody response,” said Amit Bar-Or, MD, the Melissa and Paul Anderson President's Distinguished Professor in Neurology at the Perelman School of Medicine, and chief of the Multiple Sclerosis division. “If that person didn’t have a good antibody response, which is likely because of the treatment they’re on, we’d shrug our shoulders and say, ‘Maybe you shouldn’t go because we don't know if you’re protected.’”
Today, though, Bar-Or can take a deeper dive into his patients’ individual immune systems to give them far more nuanced recommendations. A clinical test for immune cells produced in response to the COVID-19 vaccine or to the SARS-CoV-2 virus itself—not just antibodies—was one of the first applied clinical initiatives of a major new Immune Health® project at Penn Medicine. Doctors were able to order this test and receive actionable answers through the Penn Medicine electronic health record for patients like the grandfather with MS.
“With a simple test and an algorithm we can have a very different discussion,” Bar-Or said. A test result showing low T cells, for instance, would tell Bar-Or his patient may get a meaningful jolt in immunity from a vaccine booster, while low antibody levels would suggest passive antibody therapy is more helpful. Or, the test might show his body is already well primed to protect him, making it reasonably safe to attend the wedding.
This COVID-19 immunity test is only the beginning.
Physicians and scientists at Penn Medicine are imagining a future where patients can get a precise picture of their immune systems’ activity to guide treatment decisions. They are working to bring the idea of Immune Health to life as a new area of medicine. In labs, in complex data models, and in the clinic, they are beginning to make sense of the depth and breadth of the immune system’s millions of as-yet-undeciphered signals to improve health and treat illnesses of all types.
Penn Medicine registered the trademark for the term “Immune Health” in recognition of the potential impact of this research area and its likelihood to draw non-academic partners as collaborators in its growth. Today, at the south end of Penn’s medical campus, seven stories of research space are being added atop an office building at 3600 Civic Center Boulevard, including three floors dedicated to Immune Health, autoimmunity, and immunology research.
The concept behind the whole project, said E. John Wherry, PhD, director of Penn Medicine’s Institute for Immunology and Immune Health (I3H), “is to listen to the immune system, to profile the immune system, and use those individual patient immune fingerprints to diagnose and treat diseases as diverse as immune-related diseases, cancer, cardiovascular disease, Alzheimer’s, and many others.”
The challenge is vast. Each person’s immune system is far more complex than antibodies and T cells alone. The immune system is made of multiple interwoven layers of complex defenders—from our skin and mucous membranes to microscopic memory B cells that never forget a childhood infection—meant to fortify our bodies from germs and disease. It is a sophisticated system that learns and adapts over our lifetimes in numerous ways, and it also falters and fails in some ways we understand and others that remain mysterious. And each person’s intricate internal battlefield is in some way unique.
The immune system is not just a set of defensive barricades, either. It’s also a potential source of deep insight about a person’s physiological functioning and responses to medical treatments.
“The immune system is sensing and keeping track of basically all tissues and all cells in our body all the time,” Wherry said. “It is surveying the body, trying to clean up any invaders and restore homeostasis by maintaining good health.”
“Our goal is to essentially break the code of the immune system,” said Jonathan Epstein, MD, executive vice dean of the Perelman School of Medicine and chief scientific officer at Penn Medicine. “By doing so, we believe we will be able to determine your state of health and your response to therapies in essentially every human disease.”
Untangling Millions of Messages in the Immune System
Measuring and making sense of the immune system is a crucial step in Penn Medicine’s Immune Health platform.
An individual’s immune system—constantly adapting and responding to its environment—is sending millions of messages, such as a spiked fever during an infection. Most of these messages are still confounding to researchers. The challenge is to find ways to untangle those numerous signals in ways that broaden and deepen physicians’ understanding of patients’ health and response to disease.
Researchers across Penn Medicine, with the backing of I3H, are endeavoring to do so by tracking patients’ immune responses across the disease spectrum and, in some cases, partnering with informatics experts to use advanced artificial intelligence algorithms and machine-learning models to predict outcomes. Among the efforts: studying whether dietary interventions could enhance the efficiency of some cancer treatments, using immune signals to help determine MS treatments, and even testing a cancer prevention vaccine.
“The immune system operates very much like the nervous system in monitoring just about everything that goes on physiologically in our body,” Wherry said. “Unlike the nervous system, the immune system is mobile. Cells move around, survey different tissues, interpret their environment and then respond or, importantly, choose not to respond. In some ways, this cell movement is our opportunity. The blood system is the highway of the immune system, but also allows easy sampling, of at least a subset, of the cells in the immune system. If we know how to listen to the language of the immune system, we can use it to tell us about physiological changes that may not be obvious otherwise.”
Leading the Way in Cancer Immunotherapy and COVID-19
Penn Medicine arrived at this moment due to a combination of leadership in immune-based discoveries in cancer and recent advances using immune health insights to treat patients who were severely ill with COVID-19.
Much of Wherry’s own research for years had emphasized understanding patients’ immune responses to cancer and to cancer treatments that work by activating the immune system. Other Penn Medicine researchers—notably Carl June, MD, the Richard W. Vague Professor in Immunotherapy, along with many other collaborators—were pioneers of chimeric antigen receptor T cell (CAR T) cancer therapy, in which a patient’s own immune cells are reprogrammed to fight cancer cells. Once the first CAR T therapy was approved by the Food and Drug Administration in 2017, Robert H. Vonderheide, MD, DPhil, director of Penn Medicine’s Abramson Cancer Center and an immunotherapy researcher himself, said the Penn immunology community felt the moment had truly arrived to look for the clinical impacts they could have with the immune system beyond cancer.
“We realized there is this huge discrepancy between what we were measuring routinely in a tube of blood from a patient versus the billions of parameters that we can measure with the same tube of blood in a research lab 500 yards away,” Vonderheide said. “That was the start of immune health.”
The field truly began to explode at Penn three years later when COVID-19 struck. As doctors around the world were scrambling to find the best ways to treat severely ill patients, Wherry, who is also chair of Systems Pharmacology and Translational Therapeutics, thought his research approach for cancer patients could be applicable to combatting the new virus. His study of cancer patients’ immune signals to predict their response to certain treatments accelerated care for patients who were racing the clock.
So when, early in the pandemic, critical care physicians were struggling to effectively treat hospitalized COVID-19 patients, Wherry jumped into action. Leveraging the work already being done on a smaller and slower scale in cancer, Wherry and Michael Betts, PhD, a Penn microbiologist studying immunology in human infection and diseases, established the COVID-19 Processing Unit to use patients’ individual immune responses to the virus to help inform their treatment. In the clinic, they partnered with Nuala Meyer, MD, MS, a critical care physician treating patients with COVID-19 in the intensive care unit. Meyer headed a laboratory with experience quickly enrolling critically ill patients into a clinical trial to study sepsis, and she was well versed in how immunology could fill in gaps in clinical knowledge.
For Wherry and others in Penn’s immunology community, the pandemic presented a once-in-a-lifetime opportunity to show how their work could extend far beyond cancer. “We’ve been saying for a number of years that the immune system matters, and that it should be a key to helping to diagnose [and treat] diseases,” he said. “If there was ever an opportunity to put our money where our mouth is and test whether what we’ve been saying is true, this is when we have to do it.”
The Immune System’s All-Stars
The immune system is made up of dozens of types of cells that surveil for threats, communicate with one another, and defend and protect the body in a variety of ways.
Broadly Defensive Innate Immune Cells
This hardwired part of the immune system is quick to respond to injuries, viruses, bacteria, and more.
Neutrophils and macrophages are like the patrol officers responding first to an immediate threat. Both types of cells destroy viruses and bacteria soon after they are detected. Macrophages also release molecules called cytokines that cause inflammation and make it easier for more immune cells to reach the area.
Dendritic cells act like crime scene investigators. They break invader cells apart and bring their uniquely detectable component molecules and pieces, or antigens, back to the adaptive immune system to learn more about the enemy.
Specialized Learning Adaptive Immune Cells
The adaptive immune system is a trained against specific threats, and also learns and retains a memory of antigens it has encountered before.
T cells are the body’s specialized immune soldiers. They are a type of white blood cell that will destroy the body’s own cells when necessary if it detects that cell is infected or damaged. They can also direct the activities of many other parts of the immune system.
B cells multiply in high numbers when fighting off an infection or invader, and produce antibodies, which are specialized protein molecules that are custom-made by the body to latch onto and neutralize or destroy viruses, toxins, and any foreign molecules that are perceived as a threat.
Studying Immune System Function in COVID-19
The team of more than two dozen highly trained researchers who made up the COVID-19 Processing Unit first processed peripheral blood and plasma samples from hospitalized patients with COVID-19 to extract immune cells. Then, they ran an assay called flow cytometry to measure the activation of the 30 or so immune cell types in the blood, more or less evenly divided into innate—or hardwired—cells, and adaptive cells, such as T cells and B cells (the cells that make antibodies). Because each immune cell type can exist in various forms of activation and anywhere throughout the body—for instance, a single B cell from the lungs of a person who recently received the COVID-19 vaccine might be very active, while a B cell from an unvaccinated person’s lymph nodes might be in a resting state—the team produced a data set of thousands of features of each patient’s immune system.
Each patient’s immune response was charted on an immune map with the responses of other hospitalized COVID-19 patients, which is when the team noticed some surprising, and important, patterns. Unlike in most other viruses, the COVID-19 patients’ immune systems were not responding to the virus in a uniformly predictable way. Instead, the patients’ immune activity patterns seem to cluster in a few distinct groups. They published these findings in Science in July 2020.
One group of patients was characterized by their overactive immune systems—with severe inflammation, high levels of activated T cells, and a large number of plasmablasts (a type of B cell actively producing antibodies). Another group had what doctors deemed an appropriate viral response to COVID-19: Their immune systems activated to fight the virus, but not to the extent that the immune system was causing harm. The third group had a low, almost negligible, immune response.
The first group—those with overactive immune responses—were likely to see the best results from steroid treatment, while the group with little to no immune response might not see such benefits from steroids, which would further suppress their immune systems. It was a finding consistent with subsequent clinical trials that helped the overwhelmed front-line physicians make informed treatment decisions for their patients amid a global pandemic.
In one memorable instance, Meyer called the COVID-19 Processing Unit with an urgent request on a Friday afternoon: Could the team map the immune state of a patient whose case was confounding physicians? By Sunday morning, the COVID-19 Processing Unit told Meyer that her patient mapped to the hyperactive immune response group.
“It gave us a lot of insight about that patient’s immune status,” Meyer said. “It was convening the right minds… to give us a sense for which features of this patient’s immune system seemed out of balance. It shows the potential for this type of work.”
For a single patient, the COVID-19 Processing Unit team’s analysis—along with a review of the patient’s clinical data—took 12 to 24 hours. “We had a team of about 6 to 10 people working in shifts 24 hours a day,” Wherry said. “To make immune health functional from a clinical perspective, this had to be real-time.”
In just three months, the COVID-19 Processing Unit analyzed the immune responses of some 750 patients. “It was team science done in a new way,” said Allie Greenplate, PhD, director of Strategic Alliances and Operations for I3H and an adjunct assistant professor of Systems Pharmacology and Translational Therapeutics, who was then part of the COVID-19 Processing Unit as a post-doctoral researcher in Wherry’s lab. “Looking in detail at the immune system and discovering something about a person’s biology [has been done before]. What was unique was the ability to return the results to a physician in real time. To do it at the scale we did, I think, is still something that hasn’t been done elsewhere.”
Immune Health Fingerprints
The methods the Penn Medicine teams put in place to analyze individual patients’ blood and plasma samples for their immune cells’ activity and map those patterns into groupings had clear implications for patients beyond COVID-19, and even beyond cancer, where immune-based treatments are already most advanced. Once the pace of “emergency response science” slowed down, Wherry said the COVID-19 Processing Unit team saw potential in scaling up the systems they had built.
“The core infrastructure of immune health is disease agnostic,” he said. “The immune landscape analysis that we applied in COVID, we can apply to cancer, to autoimmunity or allergy. We get to look at all of those fingerprints across all patients.”
Researchers and clinicians see potential to better understand connections across conditions by creating large-scale immune landscape maps, like the one used to understand how different patients responded to COVID-19, by categorizing individual patients’ “immune fingerprints” into immune subgroups across diseases. For instance, Wherry said, the weakened immune system of a cancer patient is, in many ways, actually the inverse of the overactivated immune system of a person with an autoimmune disorder. “There’s this subgroup of cancer patients that didn’t respond to this immune-stimulating drug,” he said. “Well, there are some autoimmune patients who fall in that same category. Maybe the drugs that didn’t work in cancer will now work in autoimmunity.”
Among the first steps to realizing this ambitious vision was to streamline the process used during the pandemic into a more scalable and sustainable system. The I3H team simplified assays, standardized workflows, added support, organized teams, and embedded quality control into the process. Another change: Without the urgency of COVID-19, researchers could relax their timelines. “We could return results on the scale of days to weeks,” Wherry said, “and still be real-time for the patients to use that information for treatment choices.”
That’s how it worked for Bar-Or’s work in COVID-19 immunity, where he led a key study published in Nature Medicine in 2021 showing that MS patients taking drugs that suppressed their immune systems’ production of antibodies still gained robust protective T cell responses to the COVID vaccines. His next I3H partnership will focus on getting the patients the right MS treatments for their own body’s immune type. There are a number of FDA-approved therapies available for MS, but doctors have little guidance about which work best for individual patients. Mapping out how different patients respond to the various treatments could help doctors better target their therapies—just as they did with extra COVID-19 protection.
Autoimmune diseases like MS are among I3H’s first targets because they are fundamentally diseases of the immune system itself. Autoimmune diseases, including rheumatoid arthritis, lupus, and Graves' disease, affect almost three times the number of people with cancer but receive seven times less funding, Greenplate said. At Penn, however, generous gifts totaling $60 million from Judy and Stewart Colton in 2021 and 2022 established and accelerated the Colton Center for Autoimmunity, which will partner with I3H. Penn researchers are already making advances in multiple approaches that arm the immune system to fight autoimmune disorders, such as a modified version of CAR T therapy for the autoimmune skin condition pemphigus vulgaris and for myasthenia gravis, a neuromuscular condition.
Another researcher is working to understand long COVID—perhaps the world’s newest autoimmune disease. With blood samples offering few clues about why some people develop long COVID, researchers in the laboratory of Michela Locci, PhD, an assistant professor of Microbiology at the Perelman School of Medicine, are adapting a tool from basic science to a question with clinical implications. Locci’s lab uses a fine needle aspirate (FNA) method to collect samples from the cervical lymph nodes of COVID-19 survivors and study their ongoing immune responses in the lymphoid tissue. A closer look at these immune cells might help clarify what’s causing some survivors to develop long COVID.
Locci said she believes the study is the first attempt to use FNAs to study immune responses to long COVID infection in humans. “I could not think of a better place to conduct immune health–related studies than Penn,” she said. “This increased focus on immune health will act as a propeller to facilitate human immunology studies with potential to guide clinical decisions.”
Creating Tools for Immune Health Research
Building new tools and adapting existing ones to the challenges of immune health are among the most crucial aspects of the work.
Penn Medicine patients have an important role to contribute to immune health research. Greenplate is hoping to make it easy for many patients to get involved by expanding on the model of the Penn Medicine BioBank, which already aggregates a wide range of clinical data from nearly a quarter of a million Penn patients along with tens of thousands of biological samples from those patients, for use in observational research. The Penn Medicine BioBank was established in 2012 but has grown rapidly in recent years since the option to consent to participate was built into the electronic patient portal for every Penn Medicine patient at every location, during the COVID-19 pandemic. Patients who opt into participating in the Penn Medicine BioBank have extra blood collected the next time they are scheduled for a blood draw in the course of their care, as well as any leftover tissue from biopsies saved for research. To date, about 44,000 of these patients have genomic data associated with their (anonymized) clinical histories, including diagnoses, visits, and clinical test results, available for researchers to study using the BioBank.
“I would love to see not only your genetic information, but your immune profile as part of your medical record,” Greenplate said.
But first there are regulatory considerations to address before bringing research data to patients in a clinical setting. Research laboratories, where immune health breakthroughs are made, don’t have the clinical laboratory certification needed to return their data into patient’s medical records, Greenplate said. One option currently being studied by Angela Bradbury, MD, a Perelman School of Medicine associate professor of Hematology /Oncology and Medical Ethics and Health Policy, is asking patients if they are willing to receive research data.
So, while Greenplate envisions someday giving patients access to their immune health testing data inside their electronic medical records, “We’ll probably start on a smaller scale with research data,” she said. “Then, depending on the response, consider whether it’s worth the cost of making it an insurance reimbursable test.”
Making sense of that research data to arrive at useful insights for patients is itself another large area still scaling up. The informatics team at I3H is building an infrastructure of data management, sharing, and analysis that will make their discoveries not only possible, but also actionable. Dokyoon Kim, PhD, an assistant professor of Informatics in Biostatistics and Epidemiology, does work that integrates electronic health record information with various biomedical data sources such as biobanks, medical imaging, and different types of genomics and other “multi-omics” data, to predict disease outcomes. Now, as I3H associate director of informatics, Kim is leveraging artificial intelligence to further enhance his clinical risk prediction models. “If we have well-trained multimodal AI models,” Kim said, “we could bridge the gap between clinicians and data scientists, opening doors to broad applications including personalized medicine.”
Building those immune health data and prediction models into a standardized and scalable database is a multidisciplinary collaboration—encompassing data scientists and software engineers, immunologists, and phlebotomists—led by Joost Wagenaar, PhD, also an assistant professor of Informatics. The goal is a comprehensive data ecosystem combining multimodal clinical and research data that is easily accessible. Imagine the power of a database, Wagenaar said, that enables a physician to pull a quick analysis of patients on a specific medication or a scientist to build a cohort of information for drug discovery.
“What if all of the data is at your fingertips, and all you have to do is ask the right question?” Wagenaar said. “Would we be able to accelerate research? Would we be able to get to cures for patients faster? I think the answer is yes. Immune health is an extremely good use case where we can demonstrate that.”
Using Immune Health to Guide Medical Treatment
The ultimate goal of the work scientists and informaticists are doing in the lab is to untangle the thousands of immune signals into clear clinical messages. It’s only then that immune health data will be truly useful for doctors and patients.
Greenplate imagines “immune boards,” modeled on cancer’s tumor boards, that would bring together physicians and scientists to examine a patient’s immune health data and make decisions on how to move their care forward.
The I3H informatics team is also working to develop a streamlined Immune Health dashboard that can integrate with the electronic medical record. They have already created an Immune Health tab in Penn Medicine’s electronic health record that provides a home for test orders and test results like the COVID-19 immunity test. Their goal is to give clinicians easy access to immune health insights that offer meaningful guidance for patient care, as they work to identify and validate more of these measures.
“We don’t want to put those 100,000 features of your immune system in the electronic medical record,” Wherry said. “We want to find out which two or so features can tell whether you’re going to respond to a new MS drug better than one of the other drugs that could be used. That’s the actionable choice.”
In cancer—where immune health has a long history—some groundbreaking developments are perhaps close at hand. The field as a whole has made rapid advancements in recent years thanks to the explosion of immunotherapy research. Those developments, combined with mRNA and other gene therapy technology, bring the possibility of a cancer prevention vaccine within reach, Vonderheide said. One in the works at Penn is for individuals at high risk for breast cancer because of their genetic mutations. “There’s an active clinical trial using DNA to treat those individuals,” he said, “and boost their immune systems to intercept and prevent cancer.”
Vonderheide’s own research is showing the potential of treatments customized to an individual patient’s immune health. His team published a paper in Nature Medicine last year showing that certain patients with newly diagnosed metastatic pancreatic cancer responded extremely well to different combinations of chemotherapy and immunotherapy treatment. Depending on their immune health baseline, some patients responded well to combination A, while others found success with combination B.
Vonderheide’s team is following up on this finding with a forthcoming prospective study—selecting each trial participant’s treatment according to the predicted outcome. “That’s precision oncology,” he said. “We do that all the time, but mostly with the genetic sequence of the tumor.” This time, though, it’s entirely based on the patient’s immune system.
“This is really where the rubber meets the road,” he said. “We meet a patient and we say, ‘Based on your immune health, we think this therapy is best for you.’”