One of the most striking features of SARS-CoV-2, the virus that causes COVID-19, is the exceptional breadth of symptoms it causes in people. Of the nearly 30 million recorded infections to date, the vast majority of people experienced mild or moderate disease—which itself can range from no symptoms at all to pneumonia or long-term, debilitating neurological symptoms. A minority ended up with severe respiratory symptoms but eventually recovered. And some—nearly 940,000 worldwide, of which 196,000 are in the US—took a turn for the worse and died.

For our comprehensive coverage and latest updates on COVID-19 click here.

---

Why some people die while others recover is thought to depend in large part on the human immune response, which spirals out of control in severe disease. Over the past few months, researchers have developed a better understanding of this dysfunctional immune response. By comparing patients with varying degrees of disease severity, they’ve catalogued a number of dramatic changes across the human immune arsenal that are often apparent when patients first come into the hospital—from signaling cytokine proteins and first-responder cells of the innate immune system, to the B cells and T cells that confer pathogen-specific adaptive immunity.

The factors that trigger this immune dysregulation have so far remained elusive due to the complexity of the immune system, which consists of seemingly endless biological pathways that twist and turn and feed back on one another like a ball of spaghetti. But researchers—drawing on knowledge from other conditions such as sepsis, cancer, and autoimmune disease—are gradually building coherent theories of what puts patients en route to severe disease. Along the way, they’re also uncovering signals that clinicians could use to predict disease prognosis and identify potential new treatment avenues.

“We don’t have the clearest picture yet. Nor do we know why there’s variability in this immune response,” says Nuala Meyer, a critical care physician at the Hospital of the University of Pennsylvania who researches sepsis. While it’s well-established that underlying conditions increase the risk for developing severe COVID-19, “I definitely see patients with diabetes, obesity, and high lipids that did not become severe [cases],” she says. “I think we have a lot of work to do to understand precisely what accounts for this differential response.”

Cytokine mayhem in severe COVID-19

Earlier this year, researchers learned that some of the body’s very first defenses against SARS-CoV-2 seem to be perturbed in patients who develop severe disease. When SARS-CoV-2 begins to multiply in cells lining the respiratory tract, both infected and some bystander immune cells release interferons, cytokines that, in general, act to curtail viral replication. But in May, scientists at the Icahn School of Medicine at Mount Sinai reported strikingly low levels of interferons in infected human cells in culture, and in live ferrets infected with the virus. Instead of responding with interferons, the cells had boosted their inflammatory pathways. Then, in a July study of COVID-19 patients’ blood, researchers in France found that the interferon response seemed to be blunted in patients with severe and critical disease, whereas interferon release was robust in those with mild and moderate symptoms.

Studies on mice infected with the coronaviruses SARS-CoV and MERS-CoV have shown that an early, strong interferon response is critical to resolving the infections, but if the response is delayed, the mice develop inflammatory immune reactions instead. Some data suggest that SARS-CoV-2 is capable of blocking interferon production in the cells it infects, and it appears to be much more effective in doing so than its cousin SARS-CoV. But some patients also appear to be less capable of mounting an interferon response even in uninfected immune cells, notes Miriam Merad, who directs the Precision Medicine Institute at Mount Sinai. Either way, without a solid interferon response, the virus will persist, causing damage that activates inflammatory pathways. “The higher the damage is, the more the immune system is trying to get rid of the damage,” says Merad, “so it gets activated and at some point … it goes completely crazy.” 

This over activation is clearly evident in the form of high concentrations of pro-inflammatory cytokines in patients’ blood—the “cytokine storm” that COVID-19 has become known for. In a recent analysis of nearly 1,500 COVID-19 patients, Merad and her colleagues found that concentrations of IL-6, IL-8 and TNF-α in their serum upon admission correlated strongly with disease severity and death. Each cytokine could by itself predict whether a patient would survive. Interestingly, the nature of COVID-19’s cytokine response is markedly different from the cytokine storm side effect observed in some cancer patients who receive cellular immunotherapies and certain other hyperinflammatory conditions. In COVID-19 patients, concentrations of certain cytokines tend to be much lower—King’s College London immunologist Manu Shankar-Hari describes their increase as a “cytokine breeze” rather than a storm. But the increased cytokine levels are sustained over days and weeks, Merad says.

In a more comprehensive but smaller immunological study of moderate and severe COVID-19 patients, researchers documented a number of immune features that are relatively common in chronic viral infections, such as T cells bearing markers of exhaustion, which could mean they’re less able to fight off pathogens. But in severe cases, three cytokines stood out and strongly correlated with the severity of the disease: IL-6, IL-10, and particularly IP-10. “Some of the changes are very similar to what people have reported in sepsis previously,” notes Shankar-Hari, who is a coauthor on the study. But severely ill patients have the added complication of an impaired interferon response, which he says he sees as the “core mechanism” driving these changes in other kinds of cytokines.

Scientists at Yale University tracking the progression of COVID-19 patients found that the cytokine increase was followed by haphazard-seeming immune responses in severely ill patients. While people with moderate disease appeared to activate immune machinery designed to fight off viruses, those with severe disease seemed to recruit cells and proteins that are typically associated with combating parasitic worms as well as immune responses designed to go after fungi and bacteria that live outside of cells—an unusual response the team describes as “immunological misfiring,” as if the immune system is failing to activate the right program. And while the immune responses of those who recovered faded gradually over time, the heightened activity was maintained in patients with severe disease. Ultimately, their frenzied cytokine response doesn’t do much to stem the virus—based on swabs from the nose and throat, severe and moderate patients began with similar viral loads that only dropped off in the moderate group.

Trouble in the innate immune system

Severe COVID-19 is also marked by dysfunction in the immune cells that are first at the scene of a viral infection, including myeloid cells such as neutrophils and monocytes. For instance, researchers from Germany recently analyzed the properties of these cells in the blood of 109 individuals with mild, moderate, and severe COVID-19. Although patients with severe disease seemed to be manufacturing larger quantities of such cells, the cells themselves seemed to be only partially activated and dysfunctional. The neutrophils were largely immature, a feature that is thought to have a suppressive effect on the immune system, while the monocytes tended to have an inflammation-promoting phenotype and often lacked a critical surface protein (HLA-DR) needed for presenting viral material to T cells. The researchers didn’t spot these dysfunctional cells in mild or moderate cases.

Perhaps severe COVID-19 isn’t a purely inflammatory disease, but rather a dangerous loop of ineffective human immune responses and continuous tissue inflammation, says coauthor Leif Erik Sander, an immunologist and infectious disease specialist at the Charité hospital in Berlin.

Some of these myeloid cells in the blood of severely ill COVID-19 patients are indeed “functionally sluggish,” explains Stanford University immunologist Bali Pulendran, who recently found that the cells didn’t release cytokines when exposed to viral or bacterial debris in vitro. That suggested that the vast quantities of blood-borne cytokines weren’t coming from these first-responders in the blood, but from the lungs, where researchers have spotted high abundances of inflammatory macrophages and other myeloid-derived cells, he says. “The immune response in the lung was excessive inflammation, lots of cytokines. [But the] immune response in the blood was the opposite—it was suppression.” In contrast, moderately ill patients had significantly fewer suppressive cells in their blood. Sander notes that the precise sources of the cytokines, and their exact role in driving the severity of the disease, are yet to be elucidated.

But Merad suggests that the lung-dwelling inflammatory macrophages, the first immune cells to detect viral debris, could play an important role in driving the dysregulated cytokine response. 

These cells have been implicated strongly in the deleterious effects of aging, obesity, and diabetes—the same comorbidities linked to risk for severe COVID-19. Perhaps they’re “giving the wrong tone of inflammation in these patients” by over-activating T cells and other parts of the immune system, she says. “It is, I think, really a myeloid-triggered disease.” As for the sluggish, suppressive cells in the blood, maybe the body is somehow shutting down other parts of the immune response to prevent it from becoming overwhelmed, Pulendran says. The simultaneous immune suppression and over-activation has also been observed in sepsis, where it’s called “immunoparalysis.” 

Other innate immune cells, such as natural killer (NK) cells, are also altered in severe cases of COVID-19 compared to non-severe cases, research by scientists in Sweden has shown. NK cells sense stressed cells and kill those infected with pathogens, while also releasing pro-inflammatory cytokines and influencing T cell responses. NK cells come in different flavors, from less-differentiated ones fresh out of the bone marrow—which are good at proliferating and secreting cytokines—to highly differentiated super killers specialized for taking down virally infected cells. NK cells were more skewed toward the former group in moderate disease. But “a subpopulation of NK cells that are the most terminally differentiated, that are the most skewed towards killing, they were specifically expanded in the severely sick patients,” explains senior author Niklas Björkström, an immunologist at the Karolinska Institute. Such killer cells have also been spotted in hantavirus and yellow fever virus infections, but it’s not clear if they play a deleterious role in driving further immune dysfunction or a protective role. “I’m a positive person. So I would still be hoping that the fact that we see them is because we need them for something at that severe stage, and that they are rather trying to do something good,” Björkström adds.

An adaptive immune system out of kilter

The components that make up our adaptive immune system also undergo drastic transformations in severe COVID-19. Some of these changes are expected. Like some other infectious disease patients, those with COVID-19 almost always have unusually low numbers of lymphocytes such as B cells and T cells in their blood—perhaps because those cells are dying off for some reason, or because they’re rushing into tissues to combat the infection, says Shankar-Hari. However, despite this shortage of lymphocytes, some patients’ T cells seemed to be highly activated, says Meyer, speaking of her own recent investigation of 125 hospitalized patients in Pennsylvania.

The combination of depleted yet activated T cells “was not uniformly seen in all patients, but that was one of the features that did associate with the more severe course and more organ failure,” Meyer adds. Other work in critically ill COVID-19 patients finds—somewhat counter-intuitively—that T cells specifically targeted to SARS-CoV-2 proteins are not associated with recovery. “Maybe it’s not a numbers game with the T cells but rather has to do with the type of response that the ones present are exhibiting,” Björkström says.

COVID-19 patients also have vast quantities of antibody-secreting plasmablasts—another unusual feature of the disease. While other viral infections can provoke such a response, the increase tends to be short-lived, whereas in COVID-19 it seems to persist, Meyer adds. The blood of COVID-19 patients is also flooded with antibodies. But interestingly, new data from a study of 22 hospitalized patients suggest that although the quantity of antibodies didn’t differ between survivors and those who died, their function, and which viral proteins they targeted, correlated with severity.

Another peculiarity of the antibody response in severe COVID-19 patients is the apparent lack of a critical antibody creation process that takes place in the so-called germinal center of the lymph nodes and spleen. Ordinarily in viral infections, following a wave of initial virus-targeting antibodies that begins in the first few days of an infection, germinal centers form in the lymph nodes and spleen, where specialized B cells and T helper cells gather to produce a highly refined batch of antibody-producing cells that are crucial for lasting antibody immunity. But a study published last month demonstrated that those structures were absent in 11 COVID-19 patients who died from the disease. The germinal center formation may be stunted due to high levels of certain cytokines, the authors posit, or perhaps due to the defects of antigen-presenting cells that help drive that response, Shankar-Hari notes.

In line with those findings, data reported in a preprint by a team of Emory University researchers show that the abundance of two subtypes of B cells that create short-lived antibodies outside the germinal center correlated strongly with disease severity in 17 hospitalized patients, while those cells were barely present in healthy controls. The expansion of these B cell types has been primarily associated with flare-ups of autoimmune diseases such as systemic lupus erythematosus, where they’re reflective of an overly inflammatory state. It’s possible that this abnormal B cell response is the body’s attempt to generate antibodies after being somehow hindered from making them in germinal centers. Yet findings from autoimmune diseases also hint that “it may be that some of these cells are actually worsening the inflammatory cascade,” says Richard Ramonell, one of the study’s coauthors and a fellow in pulmonary and critical care medicine at Emory.

Merad says she thinks these abnormalities of the adaptive immune system are driven by the dysregulated cytokine response, which may be ultimately rooted in myeloid defects. But the intimate dialog between the innate and adaptive immune systems makes it hard to tease out cause and effect—an issue that comes up in autoimmune diseases all the time, notes Matthew Woodruff, an immunologist at Emory. “You see an autoimmune patient in the clinic, and you’ve got this entire system that has already been thrown out of balance in some way. . . . You’ve already established a new homeostatic process, and you can describe it, you can intervene, you can do all of those things. But the chicken and egg questions continue to be very, very difficult to answer.”   

Turning immunology findings into biomarkers and treatments

At the moment, clinicians use several biomarkers to assess COVID-19 patients’ general inflammatory state, including D-dimer, which measures protein fragments that arise from blood clots, and the inflammatory indicator C-reactive protein (CRP), both of which correlate with disease severity. However, COVID-19–specific and earlier biomarkers would be needed to reliably identify patients on hospital admission or earlier who are destined to develop critical disease. “If we can somehow identify them before they come into the hospital, or while they’re in the emergency room before they get intubated, we could potentially change their disease course,” notes immunologist and pulmonologist F. Eun-Hyung Lee of Emory University.

Over the past few months, studies investigating immune features that correlate with COVID-19 severity have yielded a number of candidate biomarkers based on immune cells, specific antibody features, and cytokines. They include the IL-6, IL-8, and TNF-α cytokines that Miriam Merad and her colleagues found elevated in severely ill patients in a cohort of 1,500 individuals treated at Mount Sinai Health System. Merad’s team validated those cytokines in a second cohort of 231 patients, and demonstrated that they were indeed predictive of disease severity. Other research that has flagged IL-6, IL-10, and IP-10 as potentially predictive markers seems promising but would need to be validated in a larger group of patients, notes Nuala Meyer, a critical care physician at the Hospital of the University of Pennsylvania.

Some of Meyer’s research has found that the ratio of neutrophils to leukocytes—which is easy to measure in the clinic—can predict dysregulated immune responses in COVID-19 patients. Other work has flagged other cytokines and serum bicarbonate, which reflects electrolyte imbalances and is a readily available lab test. But in that study, “it did not look like all COVID patients fit into [the] hyper-inflammatory group. So in a way I think we need to do more work to understand what are better plasma biomarkers for severe COVID,” she says.

Findings from such studies are also providing insights into new treatment approaches for COVID-19. Several trials are underway that aim to administer interferon—which is already used to treat hepatitis C—to patients. However, timing is key when it comes to interferon treatment—if it’s given too late, it could worsen the disease. 

For later in the course of disease, immune-dampening steroid drugs such as dexamethasone and other corticosteroids—which can curtail the activity of multiple cytokines in tandem—appear to be effective in reducing the risk of death. But corticosteroid treatments often have side effects, especially for elderly patients, and more targeted therapies would be preferable. 

So far, researchers haven’t had any luck in tamping down levels of individual cytokines. Doing so for IL-6, for instance, is effective in dampening the cytokine storm that occurs in some cancer patients receiving cellular immunotherapy. The pharma giant Roche recently reported that its CONVACTA trial testing its IL-6–targeting drug Actemra did not improve the clinical status of hospitalized COVID-19 patients with severe pneumonia. And a smaller study testing the IL-6 blocking agent sarilumab in severe disease also proved disappointing to Merad, she says. A combination treatment that blocks multiple cytokines might be the way to go, she suggests. But COVID-19 is also a very heterogeneous disease, making it difficult to interpret clinical signals, she notes. “We need to tailor the treatment to the molecular effect.”

Alternatively, it may not be the cytokine increase that’s killing patients. Some researchers have recently expressed doubt that cytokines play a major role in driving the severity of the disease based on small studies suggesting that levels of pro-inflammatory cytokines are much lower in severe COVID-19 than in patients without COVID-19 who die of acute respiratory distress syndrome. 

Perhaps the ultimate culprit is a coagulation problem driven by the raging inflammation of blood vessels, says immunologist Niklas Björkström of the Karolinska Institute in Sweden, noting that infusions of the anticoagulant heparin reduced the fatality rate in some patient groups. “The fact that a lot of our patients have problems with clotting raises interest perhaps in the complement pathway or novel anticoagulant strategies,” Meyer says. In addition, “I do think we’re interested to know if we can perhaps intelligently intervene on this T cell activation without tipping the balance too much towards immune suppression,” she adds. 

Half a year into a pandemic, “I understand that the public is frustrated. Even my family’s like, ‘Can’t you figure it out?’” Merad says. However, answers are on their way, and at breakneck speed, in large part thanks to technological advances over the past decades. “We are [learning] much faster than any time in history where we’ve had a big disease [outbreak] like that. So I’m confident that the few months to come are going to be extremely informative.”

—Katarina Zimmer