Enhanced Thrombotic Susceptibility, Endotheliopathy, and Abnormal Cognitive Behavior for up to 5 Months after Sars-Cov-2 Infection in a Mouse Model of Long COVID

The sheer number of individuals infected with SARS-CoV-2 during the COVID-19 pandemic revealed the development of a debilitating post-acute condition named post-acute sequelae of COVID-19 (PASC), also known as Long COVID. Symptoms of PASC can last for months to years after the initial infection and are multifaceted, including fatigue, exertional malaise, and a collection of cognitive impairments (e.g., brain fog). Conservative estimates suggest PASC occurs in ~10% of individuals infected with SARS-CoV-2. PASC is associated with hypercoagulability and persistent endotheliopathy, and several studies reported increased risks for cardiovascular disease and thrombosis up to 1 year after infection. The mechanisms responsible for PASC remain unclear. Numerous hypotheses are proposed, including chronic inflammation due to persistent viral reservoirs, development of autoimmunity, and immune-mediated tissue damage resulting from continuing complement activation and/or non-resolving anti-viral responses. This indicates the urgent need for new models to test these hypotheses and to understand the pathogenesis of PASC and PASC-associated thrombosis.

We developed a mouse model of Long-COVID based on the widely used K18-hACE2 (K18) transgenic mouse. Infection with a low viral dose of 2.5 x 10³ f.f.u. of SARS-CoV-2 (USA WA1/2020) prolonged the acute symptomatic phase recapitulating many of the COVID-19-Associated Coagulopathy (CAC) findings observed in human patients, including hypercoagulability, increased plasma concentration of D-Dimers, soluble thrombomodulin (sTM), and VWF. Surviving mice recovered their initial body weight two weeks post-infection and did not display signs of overt disease for their remaining lifetime.

Histological analysis of surviving mice revealed the presence of long-lasting inflammatory foci scattered around the lungs’ parenchyma, persisting up to 14 months post-infection, suggesting the existence of an enduring inflammatory stimulus of unknown origin leading to a state of chronic inflammation. Persistent endotheliopathy was evident from elevated levels of circulating sTM, sEPCR, sP-Selectin, and elevated levels of VWF in plasma lasting for at least 5 months post-infection.

At 70 days post-infection K18 survivors developed significantly larger thrombi 24 hours after stenosis of the vena cava compared to infected wild-type (WT) littermate controls who developed minimal disease during the acute infection. The average thrombus size in K18 survivors was in between that of uninfected factor V Leiden heterozygous and homozygous mice, suggesting the thrombotic risk of K18 survivors might be somewhere in between as well. At 5 months post-infection, K18 survivors showed a complete and sustained occlusion in a carotid artery thrombosis model induced by a low dose of ferric chloride, whereas no stable occlusion was observed in infected WT littermate controls, suggesting a persistently increased susceptibility for thrombosis.

To evaluate the occurrence of cognitive abnormalities comparable to those described in individuals with PASC, K18 survivors and infected WT littermate controls were challenged using a battery of behavioral tests 5 months post-infection. In an open field test, K18 survivors showed reduced exploratory drive and total distance traveled compared to controls, suggesting impaired locomotive activity and increased anxiety. Cognitive impairments of K18 survivors were exacerbated after physical or emotional stress. In a novel object recognition test preceded by 20 minutes treadmill running, K18 survivors spent significantly less time exploring the newly introduced object compared to infected WT littermate controls. Similarly, in a light/dark transfer test, a 30 minutes physical restraint, eliciting a corticosterone-mediated stress response, significantly reduced transitions from darkness to light in K18 survivors when compared to infected WT littermate controls. Thus, K18 survivors displayed a significant increase in stress and anxiety responses that are indicative of cognitive impairments that were not observed in infected WT littermate controls.

In summary, our Long COVID mouse model replicates key aspects of PASC in humans and provides a unique opportunity to study the molecular mechanisms contributing to PASC and to evaluate potential new therapeutic strategies.