LONDON – The largest study of its kind to date has identified new biomarkers of inflammation that are both indicators of severe COVID-19 infection and distinguish it from severe influenza.

In particular, the cytokine GM-CSF occurred at much higher levels in people who later developed life-threatening respiratory disease, and it was very pronounced in fatal COVID-19. GM-CSF levels increased in relation to severity of COVID-19 and were elevated early on in patients who progressed to severe disease, typically within four days of symptoms emerging.

“We looked at 33 cytokines. The one that stood out was GM-CSF, which was at levels 10 times higher in patients who died, compared to healthy controls,” said Ryan Thwaites, research associate at the National Heart and Lung Institute, Imperial College London and co-author of a paper reporting the research published in Science Immunology on March 11.

The researchers recruited 471 hospitalized COVID-19 patients, who were stratified by disease severity, along with 39 people with mild disease and healthy controls. From those samples, they generated an immune response profile that forms the basis of an inflammation network, illustrating how levels of the 33 cytokines relate to and feed back on each other.

IL-6 already has been fingered as a key culprit in the damaging host immune response, and the inflammation network shows that in common with GM-CSF, IL-6, too, is a network hub.

“We can see the way in which inflammatory processes are connected. It gives us a beautiful picture of disease processes,” said Peter Openshaw, professor of experimental medicine at Imperial College London and co-lead of the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) that carried out the research.

The finding of elevated GM-CSF would appear to augur well for Glaxosmithkline plc’s inhibitor, otilimab, which is in phase III development in rheumatoid arthritis, but last month turned in middling results in the first cut of data from a phase II trial in the treatment of respiratory disease caused by COVID-19 infection.

There was no significant difference between patients who received a single infusion of otilimab and standard of care. However, the ISARIC data indicate GM-CSF levels are higher in patients over 70 years of age and a preplanned efficacy analysis showed there was a significant effect in that age group, leading Glaxosmithkline to expand the over 70s cohort and try and confirm that finding.

Some published studies testing GM-CSF inhibitors are favorable, some are not, Openshaw noted. “It could be adding GM-CSF inhibitors to standard of care might actually be detrimental, or it could be it only helps some people,” he said. The cytokine data can give a steer, but it is always necessary to do clinical trials, he added.

It was not evident at the start of the pandemic that the human immune system, rather than the virus per se, causes the most severe respiratory disease, said Kenneth Baillie, academic consultant in critical care medicine at Edinburgh University and ISARIC lead.

While there are “lots of plausible ideas” on how to repress harmful immune responses, this analysis now gives pointers to “the best and most effective way to stop immune damage to the lungs, whilst leaving the immune system active against disease,” Baillie told attendees of a press briefing.

“The importance of this is that there are literally hundreds of anti-inflammatory drugs, but little evidence on how to select between them. Now we have some signposts,” Baillie said.

Unique to COVID-19

The research published this week builds on more than a decade of collaborative research by ISARIC since it was set up in response to the 2009 H1N1 swine flu outbreak. The consortium has set out a standardized clinical characterization protocol for reporting patient responses to respiratory virus infections that has since been used in a number of epidemics.

The existence of ISARIC meant the research was fast getting off the ground in obtaining and analyzing biological samples to profile the immune response.

ISARIC also has stored patient samples from earlier outbreaks, enabling the researchers to cross-reference the immune responses to patients who died from swine flu in 2009, and those seen in COVID-19 in detail.

“We could compare samples collected 10 years apart,” Thwaites said. Levels of IL-6 were equivalent in both, whilst GM-CSF was far higher in patients with fatal COVID-19 than in flu. That suggests this arm of the inflammatory response is different in the two. It is “intriguing” that GM-CSF seems to be unique to COVID-19 and is not found in influenza, another common respiratory infection, said Thwaites.

“For IL-6, we already know that inhibitors are effective. We don’t have trial evidence on that scale for anti-GM-CSF, and that should now be a higher priority,” he said.

In addition, further research is needed to see if GM-CSF levels can be used as a marker in early disease to identify those at risk of progressing to serious respiratory infection.

While the study shows many inflammatory markers are raised in COVID-19, there is no evidence of fast onset cytokine storms, as seen in septic shock. Rather, it is a cytokine “wind,” or possibly a “gale,” Thwaites said.

“The evidence is pointing in the direction that rather than an all-guns blazing immune response it is more nuanced. Some elements of the immune response are still being helpful in trying to clear the virus. We need to understand which bits are helpful, and which are unhelpful,” said Thwaites.