The immune system touches every aspect of human health, yet most of what we know about it comes from studying mice. While illuminating, that’s also been problematic: Mouse and human biology differ in some fairly fundamental ways, and only a handful of immunology findings in mice have led to tangible improvements for human health.

This has spurred many scientists, including Stanford immunologist Mark Davis, to develop ways to use human blood and tissues, instead of mouse ones, to study the intricacies of human immune responses. Davis’s group is among the first to apply these new methods — things like mass cytometry, which allows the scanning of dozens of proteins attached to thousands of immune cells at a time — to ask basic questions about our defenses against disease that couldn’t be addressed before.

Davis recently coauthored a review describing how new approaches and technologies are transforming understanding of human immunology. These developments have fostered the new field of systems immunology — in which multiple aspects of our immune response are measured simultaneously. Such approaches could help scientists figure out the causes of autoimmune disease, better understand how the immune system battles tumors, and finally test what’s called the hygiene hypothesis — the proposal that increasing rates of allergy and immune-related disease are caused by a lack of exposure to disease-causing agents and an overly clean environment early in life.

This conversation has been edited for length and clarity.

Scientists have studied the ins and outs of the human immune system for decades. Is it largely figured out?

Oh, not at all. I think we’re just kind of rebooting human immunology research. It’s been difficult to analyze the human immune system in depth. We needed the technology before we could measure dozens of different immune-activating molecules and hundreds of cell types — just like the field of genomics wasn’t that feasible until DNA-sequencing machines could work out entire genomes.

We don’t have detailed metrics of immunological health — ways to measure and track the health of the immune system in the clinic. Here’s this field that’s been booming for the last 60 or so years, and why isn’t there more of it in medicine? Doctors can’t make measurements unless the results affect treatment decisions. That’s a challenge to immunology in general: Make yourself relevant and develop metrics of immune health that people can use.

What are some of the pressing questions in human immunology that you’d like to see answered?

There is so much potential in human work in the sense that humans have thousands of diseases. For example, we have this whole hygiene hypothesis, which postulates that the reason kids are getting increasingly early onset of autoimmunity and allergies — especially food allergies — has to do with the way modern life restricts their exposure to pathogens. Because of that, their immune systems don’t get a chance to tune themselves properly. This idea has been around for a while but there hasn’t been any real critical test of it in humans. 

In a lot of autoimmune diseases — ones in which a person’s immune system attacks her or his own body, such as multiple sclerosis or type 1 diabetes — the dominant paradigm is that they are triggered by an infectious disease. The idea is that, from the immune system’s perspective, some proteins or protein fragments on bacteria or viruses look similar to the body’s own proteins. These proteins and fragments — called antigens — can convince the body to make an immune response against its own tissue.

But we’re completely ignorant about what the antigens are in any of the major autoimmune diseases. We know what the antigen is for celiac disease. Other than that, we know nada.

Why aren’t animal models good enough? Are humans that different from animals?

The mouse model has been good for understanding basic immunology. Immunology has many moving parts and many bizarre mechanisms that are the same in humans and mice. One example is gene rearrangement, where DNA inside immune cells called B and T cells is shuffled around to create myriad protein receptors for recognizing a wide array of pathogens.

But then, there are so many differences. After all, the mouse is a different species that has developed an immune system meant to deal with completely different types of infections. Also, in the lab we have done away with environmental pathogens and genetic variety in the mice, and both of those things are key features of human beings.

Graphic showing how T cells “see” their targets, protein fragments called antigens. T cells have various receptors that recognize and bind to various antigens to set off an immune response. Antigens m

T cells can initiate different immune responses, depending on what a T cell “sees.” Often, that’s a protein fragment called an antigen from a disease-causing bacteria or virus. A protein on the T cell called a receptor binds antigen, stimulating the body’s attack. Some T cells may mistakenly recognize antigens from a person’s own normal cells, as in autoimmune disease. Other T cells recognize antigens from cancer cells, which assists in the body’s fight against tumors. Identifying the exact T cell receptors and antigens active in people could help researchers make better vaccines, halt autoimmune diseases and improve immunotherapies for cancer.

We don’t want the mice we work with to get diseases. We put them in the basement covered with HEPA filters. I’ve never seen an immunologist so angry as when someone tells them there’s a disease in the mouse house. It will ruin your experiments. But here we humans are, coughing and sneezing and getting all sorts of afflictions from birth.

The mouse is a very imperfect representation of the human immune system, and we’re never going to know how similar or different it is unless we have human data. Once we know how it is in humans, we can then go back to the mice and say, “Can I model this somehow?” And most of the time, you can.

Why haven’t scientists historically been able to ask these basic questions using human beings?

Most of what we do in the mouse world we can’t do in humans because it’s illegal or immoral. To get human data, we’ve had to develop a whole new suite of technologies, like mass cytometry and the ability to sequence genes from individual cells one at a time, which has helped us study T cell receptors. We’ve known about T cell receptors — proteins on the surface of T cells that recognize antigens — in mice and humans for decades now. And we know that there are a lot of diseases that involve T cells. So, a key question is: What are the T cell receptors recognizing in these cases to cause disease?

In the case of autoimmune disease, it’s not at all obvious what the T cell receptors are recognizing. Each T cell’s receptor is unique and made from two different genes. Twenty-five years ago, there was a paper describing the gene sequence for one-half of a T cell receptor commonly found in damaged brain tissue from multiple sclerosis patients. Absolutely nothing could be done with that information because it only described one of the two T cell receptor genes. It was impossible at the time to determine both gene sequences for an individual T cell’s receptor at once.

Graphic showing how scientists identify specific T cell receptors and thus their antigens using gene sequencing and other techniques.

Immunologists can figure out what antigen a T cell responds to by sequencing the genes that code for the T cell receptor. A T cell receptor is made up of two proteins, each encoded by a separate gene. In the past, gene sequencing could only be done using thousands or millions of cells at a time, so it was impossible to match a T cell’s two receptor genes back together. Using newly developed techniques, researchers can now sequence the genes from many individual T cells simultaneously and keep track of T cell receptor gene pairs. This has allowed them to match T cell receptors to their antigens.

Several years ago, we developed a high-throughput technique to get both gene sequences from single T cells with very high efficiency. We can do this for hundreds of individual T cells at a time, and narrow in on particular types of T cells activated in people with multiple sclerosis.

So suddenly we’re now in the position to find out what the antigen is — what the T cell receptor is binding to for a given disease. We don’t have that answer yet in multiple sclerosis. But we have made progress in one type of cancer.

We recently published a paper on cancer with Stanford molecular biologist Chris Garcia describing how to figure out which antigens the T cells recognize inside colon tumors as the T cells attempt to mount a defense. We used technologies like single-cell sequencing to find the most common T cell receptors in colon cancer biopsies from patients and defined several antigens that the immune cells recognized. Lots and lots of people are acutely interested in what T cells inside tumors “see” and how they can use that information to enhance immunotherapy.

How else are scientists working to address human questions in their studies of the immune system?

I think what people have to be creative about is taking advantage of the things that are unique to humans. What things are readily available that have immunological information? Blood banks are great. Vaccines are widely available. If you want some genetic input, go after twins. Another thing is that people live in different places with exposure to different diseases. We have a project where we’re comparing kids in Bangladesh with kids in Palo Alto, and [the data is not published yet, but] I can tell you that they’re really different in terms of their immune systems. Turning to such tools will allow us to address things like the hygiene hypothesis.

What other kinds of questions can researchers now ask that they couldn’t ask before by using these more creative strategies?

What in the immune system is determined genetically, and what isn’t? What role does the environment play? We did a big twin study here at Stanford. We looked at over 100 pairs of twins. We found, using systems immunology approaches, that something like 75 percent of the 200 traits that we could measure — things like the frequencies of various immune cell types such as B or T cells, and responses to flu vaccines — were barely affected by genetics. We could tell that younger twins were more influenced by genetics than older twins, not surprisingly. Genetics makes a contribution, but genetics is what you start with — and then life happens. Exposure to different pathogens over time likely has a bigger influence on how the immune system works.

How are new technologies changing your approach to studying human immunology?

From what I understand of history, in the 1930s and 1940s, biology was all about descriptive work — you know, like counting the hairs on a beetle. Later, labs would focus on one molecule in one kind of cell, trying to understand how it functions. And that works. I’ve done that with T cells. But the immune system is a system, and T cells are just one part of that, and all the components are working together to create layers of defense.

Until ten or so years ago, we didn’t have the technology that allowed us to measure those layers in a comprehensive way. Now we do. And I say, let’s look at everything and see what the whole immune system tells us is important. The systems immunology approaches, using things like mass cytometry, are a way to cast a broad net, to look at big chunks of the immune system just in a blood sample. It’s letting the system guide you in terms of what might be important. And then you dig deeper.


So you now have the technology for a systems approach. How do you actually apply that in human studies?

Right now, we’re working on data from a thousand different people. We’re pursuing metrics of immunological health such as each person’s particular mixture of immune cells present in the blood, and developing a cheat sheet to tell us if someone’s at risk for heart disease, cancer or flu. One of my colleagues at Stanford calls it “the human immumonitor.”

What findings might affect clinical practices in the near future?

We’re on the path to revolution in terms of understanding, in the broad spectrum of autoimmune diseases, just what the antigen is that T cells are recognizing. Once we know the antigen that triggers the T cells to attack the body in a given disease, we can try to figure out whether we can block that antigen or train immune cells to ignore it. In allergy responses, we know that if you give patients specific injections of the allergens, in a fraction of the patients you can desensitize the patients to the allergens. Maybe we can do the same for autoimmune diseases.

I think cardiology is destined to boom in terms of immunology. A lot of heart disease involves inflammation. We’ve been working with cardiologists here at Stanford to do more immune assays, and we published a paper last year where we found an immunological and metabolic basis for hypertension. We isolated some novel compounds that we saw elevated in people with persistent hypertension and when we put those compounds in mice, we gave the mice hypertension.

We also saw activation of genes that carry instructions for proteins involved in a certain immunological pathway. That pathway is typically associated with infectious diseases, and causes inflammation. But in this case, we saw the genes for that pathway active in older people that had persistent hypertension. So perhaps combating inflammation could help fight hypertension. It turns out that one way you can ameliorate the inflammatory pathway is by consuming coffee and dark chocolate. So that’s the good news.

How mature is this human approach to immunology and where do you think it will go?

Every field starts with description. In a way, human immunology is a new field, and so we have to start with description.

I think there’s going to be an explosion of discoveries in human immunology. That’s why I also think this convergence is sort of a sweet spot where scientists who are interested in discovering basic principles and people who are interested in understanding what’s going on in their patients can both benefit enormously.