illustration of a coronavirus and antibodies
illustration of a coronavirus and antibodies

When the Immune Response Makes COVID-19 Worse

If the immune system makes mistakes—reacting late or getting the target wrong—it can amplify the damage wrought by SARS-CoV-2.

alejandra manjarrez
Alejandra Manjarrez

Alejandra Manjarrez is a freelance science journalist who contributes to The Scientist. She has a PhD in systems biology from ETH Zurich and a master’s in molecular biology from...

View full profile.


Learn about our editorial policies.

Sep 27, 2021

ABOVE: © ISTOCK.COM, ELENA FOMINA

From the early days of the pandemic, it has been evident that patients respond differently to SARS-CoV-2 infection, ranging from having no symptoms at all to needing hospitalization. So far, at least 4.7 million people have died from COVID-19 worldwide.  

Finding the explanation for this remarkable variability has been a main focus of COVID-19 research. How the immune system responds to the viral infection—depending on age, sex, viral load, genetics, and other known and unknown variables—largely defines the course of the disease. Two elements have emerged as essential: reacting on time and toward the right targets. Not controlling the infection early enough or confounding the self with the enemy can cost the body dearly.  

Interferons: crucial early defenders 

We can think of COVID-19 as a two-step disease, says immunologist Darragh Duffy of the Institut Pasteur in Paris. The first step is when “the immune system is trying to respond to the virus and there, for me, it’s all about the interferon response,” he says. Interferons are proteins that bind to specific surface receptors on cells in danger, warning them about the viral invaders and orchestrating a signaling pathway that will keep the viruses from multiplying.  

The disease moves into the second stage when interferons don’t do their job, leading to viral growth and spread. “If you don’t control the viral infection,” continues Duffy, “then you get the secondary hyperinflammation that goes systemic, and that’s where the comorbidities come into play.” 

Cumulative evidence suggests that patients who are defective in interferon production are especially vulnerable to a SARS-CoV-2 infection. Using ferrets and human cell lines as models, a team observed in March 2020 that the virus induced lower levels of interferons than viral respiratory infections typically do. In a study published in August 2020, Duffy and others reported that between 8 and 12 days after the onset of symptoms, the blood samples of patients with severe or critical COVID-19 showed low production and activity of a specific group of interferons known as type I. Many more studies have followed those initial findings, confirming the relevance of these signaling proteins to clinical outcomes. Gradually, scientists are also teasing out why some patients do not have enough of them to effectively fight the infection.  

Genetic risk factors have appeared as one culprit in deficient interferon production and activity, at least for a fraction of hospitalized patients. A team led by Jean-Laurent Casanova, an immunologist and geneticist at the Rockefeller University and the University of Paris, sequenced the whole genomes or exomes of patients with life-threatening COVID-19 pneumonia and compared them to patients with asymptomatic or mild SARS-CoV-2 infections, looking for rare gene variants that in previous studies had been associated with the type I interferon immune response to influenza virus. The researchers found that the cohort with critical COVID-19 was enriched with these variants compared to the controls.  

More recently, a team also led by Casanova found X-linked deleterious versions of another gene involved in type I interferon production in 16 of 1,202 male patients who developed critical COVID-19 pneumonia (15 of whom were younger than 60 years old). In contrast, none of the study cohort’s 331 infected male patients with no symptoms or mild cases carried the variants.   

Casanova and colleagues have further shown that interferon activity can be blunted by the body’s own antibodies. In a cohort of 3,595 patients with critical COVID-19, 13.6 percent carried autoantibodies to type I interferons called -α2 or -ω in their blood. Among critical patients above 80 years old, the percentage increased to 21 percent, as these particular autoantibodies become more frequent with age. Only 1 percent of the patients with mild or no symptoms carried them. The team then looked at a cohort of 10,778 uninfected individuals, and found that 2.3 percent of them carried these autoantibodies—for those older than 80, the percentage climbed to 6.3. Along with other studies, these findings suggest that autoantibodies targeting interferons -α2 and -ω predate infection, increase with age, and may be more common in men than in women.  

Together, these findings suggest that “the general mechanism of disease is a disruption of type I interferon immunity,” says Casanova: “whether genetic or autoimmune, the endpoint is insufficient type I interferon.” 

Timing matters 

Interferons appear to only play a positive role in the immune response to SARS-CoV-2 early in the course of infection, however. When it comes to the interferon response, timing is critical, says Duffy. “It could be detrimental later on in that it could be inhibiting some of the adaptive immune response [or] it could be driving certain inflammation,” he says. In line with this, administration of interferon to hospitalized patients in a clinical trial led by the World Health Organization Solidarity Trial Consortium did not reduce mortality. 

When it comes to the interferon response, timing is critical.

Interferons are not the only immune components that have time-dependent effects on patients. A recent study led by Akiko Iwasaki, an immunologist at Yale University, measured antibodies to SARS-CoV-2’s spike protein in the blood of 229 patients at different stages of viral infection, and found that there is a critical window in which these antibodies have to be developed for a good clinical outcome. If antispike antibodies were induced more than 14 days after symptom onset, they were no longer protective: patients who would eventually die from the disease, for instance, reached higher antispike antibody concentrations in later stages of the disease than patients who survived. 

“The inability to mount [a] proper sequence of events is likely leading to this kind of delay,” says Iwasaki. The immune systems of patients with the delayed response eventually “do develop antibodies, but it’s too late at that point because the virus infection has already spread.” In those patients, “it’s no longer the virus that is causing the disease, but [rather] the hyperimmune response . . . and at that point, antibodies might not be functional or even relevant.” 

Autoimmunity: when the body turns against itself 

If the immune system acting on time is crucial, so is responding to the right target: the virus. The body, however, sometimes confuses itself with the adversary, and ends up launching an immune response toward its own molecules and tissues—a phenomenon known as autoimmunity. While evidence suggests this issue predates COVID-19 in the case of antibodies to type I interferons, in other cases, the body appears to target itself as a result of the infection. 

“It is not anything strange to have an autoimmune response after infection,” says Ana Rodriguez Fernandez, a microbiologist at NYU Grossman School of Medicine. Rodriguez has studied autoimmune responses in malaria for several years. Other infections, such as tuberculosis and AIDS, are also known to induce an autoimmune response in parallel to the “good, antigen-specific response,” she says. 

In addition to interferons, autoantibodies can go after other molecules associated with the immune response itself. Using a high-throughput approach, a team led by Iwasaki and Aaron Ring of Yale University screened plasma samples of both healthy and SARS-CoV-2 infected individuals for autoantibodies against a library of 2,770 extracellular human proteins. They detected a greater number of reactivities against these antigens in COVID-19 patients compared to uninfected people, and many of the targets were immune-related proteins.  

For instance, the team found autoantibodies directed to white blood cells, says Iwasaki. Such antibodies could deplete patients’ B cells or T cells, which are needed to confront the virus. “In a way, these autoantibodies are causing a state of immunodeficiency, because they are attacking the very cells that are needed to fight off the infection,” she says.  

A study published September 14 also found a wide diversity of antigens—including targets related to immune response, such as cytokines—neutralized by autoantibodies in the blood of SARS-CoV-2-infected individuals. The research team found that around half of a cohort of 147 hospitalized patients had autoantibodies against one or more of the human antigens tested, while this was true for less than 15 percent of the healthy participants.  

If the immune system acting on time is crucial, so is responding to the right target: the virus.

A subset of the autoantibodies observed in COVID-19 patients appeared to be triggered by the infection. These autoantibodies were not present on the day the patients arrived at the hospital, explains Paul Utz, an immunologist at Stanford University who led the study, but were detected one week or more later, “suggesting that the virus was directly activating the immune system to make [them].” 

Among the autoantibodies detected by Utz and his colleagues were some that are often found in rare connective tissue diseases such as systemic sclerosis, systemic lupus erythematosus, and myositis. Previous studies have suggested that once diagnosed with these rare diseases, patients “pretty much have [the autoantibodies] forever,” says Utz. But it is not clear whether that is also true for COVID-19 patients. If these autoantibodies stick around, he says, that would suggest patients who start making them are at risk for developing the associated autoimmune disease, or that these proteins could be playing a role in long COVID

Rodriguez and her colleagues similarly found in a recent study that autoantibodies—specifically, those against cell-free host DNA and the fatty molecule phosphatidylserine—are associated with severe COVID-19 infections. Rodriguez had previously found that phosphatidylserine is targeted by autoantibodies in malaria, and another research team at the University of Michigan has also detected those autoantibodies in hospitalized COVID-19 patients. Rodriguez’s team found that within a cohort of 155 hospitalized SARS-CoV-2-infected subjects with different degrees of disease severity, 14 were positive for autoantibodies against phosphatidylserine and seven for autoantibodies against DNA. All but one of the patients in each group developed severe COVID-19. According to Rodriguez, early detection of these autoantibodies could help to predict which patients will potentially get very sick, and thus alert healthcare providers to give them special care. 

In an email to The Scientist, Utz writes that the small number of patients with anti-DNA antibodies makes him unsure whether this could be defined as a risk factor for severe COVID-19. But overall, he says that Rodriguez’s results are interesting and support the importance of studying autoantibodies in SARS-CoV-2 infection. 

If the virus can induce the immune system to develop autoantibodies, could the vaccine against it do the same? Together with colleagues, Utz explored this question, and their findings are reassuring. “We don’t see any new autoantibodies or any worsening of autoantibodies—if patients have them—following vaccination, but we do see it with infection,” says Utz. These results confirm the importance and safety of getting vaccinated, rather than taking “the risk of developing autoimmune disease later,” he adds.   

Mucosal Immunity: The First Line of Defense  

Most studies of the human immune response to SARS-CoV-2 to date have been based on blood samples. But what happens there might not always reflect what’s going on in other body compartments. The nasal mucosa, for instance, is the first site to come into contact with the virus, so details on how the immune response works there could lend insight into the eventual course of the disease.  

The “mucosal site is often overlooked compared to the blood,” says Darragh Duffy of the Institut Pasteur. In collaboration with colleagues, he recently analyzed antibodies, cytokines, and viral load both in plasma samples and nasopharyngeal swabs of 42 COVID-19 patients 8 to 12 days after their first symptoms. The team found that the local immune response in the nasal mucosa was often distinct from that in the blood of the same patient, suggesting that the immune response might be regulated differently there. A caveat, says Duffy, is that they only looked at one time point, so obtaining samples over the course of the disease could offer a better picture of this immune regulation. 

Strikingly, Duffy and colleagues found that the composition of the nasal microbiome of patients at the time they were swabbed correlated with local and systemic cytokine expression. For instance, the presence of opportunistic pathogenic bacterial genera correlated with elevated expression of inflammatory cytokines associated with disease severity, while beneficial commensals correlated with the cytokine profile of a good immune response—high levels of interferons, for example. Overall, those patients with severe or critical COVID-19 had less microbial diversity and more pathobionts in their mucosa. It’s not clear whether this different microbial composition was already present before infection or was a consequence of it, however. “I suspect it is probably a combination of both,” says Duffy, as there is already a lot of variability in the microbial profile of healthy subjects.  

Other studies that have zoomed in on the mucosal immune response reveal correlations with what has been observed in blood. For instance, one study found low levels of nasal interferons in a subset of critically ill infected patients who were also positive for autoantibodies against type I interferons in their blood and nasopharyngeal mucosa. This suggests that autoantibodies could affect immunity not only systemically, but also at the local mucosal level.