Long COVID's biological blueprint: new study offers promising insights

By Neha MathurReviewed by Danielle Ellis, B.Sc.Sep 27 2023

In a recent article published in Nature, researchers designed a cross-sectional study to investigate the biological underpinnings of “Long COVID” (LC), the post-acute infection syndrome (PAIS) triggered by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).

Background

Many microbial and viral infections lead to the development of PAISs, implying they are ubiquitous. Some have been extensively studied, like myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), yet the basic biology underlying PAIS development remains unclear.

As recovery from coronavirus disease 2019 (COVID-19) varies from person to person (heterogeneous), in some convalescent individuals, its symptoms might linger on for a prolonged duration, manifesting as LC.

LC presents a constellation of debilitating symptoms, such as fatigue and cognitive deficits; additionally, its prevalence is high. Some prospective studies have suggested that one in eight people with COVID-19 experience persistent symptoms, likely because viral remnants in tissues activate autoimmunity or cause microbial dysbiosis. In some cases, chronic inflammation causes tissue damage.

About the study

In the present study, researchers created five groups using the 273 participants of the Mount Sinai Yale Long COVID (MY-LC) cohort study in New York, United States of America (USA), as follows:

i) healthcare workers with SARS-CoV-2-infection before vaccination (HCW); 

ii) SARS-CoV-2-uninfected (healthy), demographically matched, vaccinated controls (HC); 

iii) previously infected, vaccinated people with no persistent symptoms, i.e., convalescent controls (CC); 

iv) people with persistent symptoms after acute SARS-CoV-2-infection (Long COVID, LC); and 

v) another group of individuals with persistent symptoms after acute infection from another study (External LC). 

They used multi-dimensional immunophenotyping and unbiased machine learning (ML) to identify potential LC biomarkers in data aggregated from all five cohorts. The study follow-up lasted over 12 months.

Results

The study findings pointed to marked immunological differences between the LC and control populations; for instance, LC-affected people had altered immune cell populations in peripheral blood circulation.  

While the numbers of monocytes, double-negative B cells, and interleukin (IL)-4/6 secreting CD4 T cells increased in LC-affected people, those of dendritic cells (DCs)1 and central memory CD4 T cells had decreased, which are responsible for antigen presentation and cytotoxic T-cell priming. In addition, cerebral spinal fluid (CSF) derived from LC-affected individuals had more T-cell immunoglobulin and ITIM domain (TIGIT)+ CD8+ T cells, indicating possible immune exhaustion. 

Conversely, autoantibodies to human exoproteome did not significantly differ between LC and controls. Thus, whether autoreactive T cells play a role in LC pathogenesis requires future investigation. 

Further, individuals with LC had more antibodies to SARS-CoV-2, Varicella Zoster virus (VZV), and particularly Epstein-Barr virus (EBV) antigens. Other studies have shown that EBV viremia occurs in patients hospitalized due to COVID-19; thus, it appears to be an independent predictor of the development of persistent symptoms later. Elevated IgG against EBV lytic antigens suggests that reactivation of latent herpesviruses might be a common feature of LC.  

Furthermore, the authors noted substantial differences in levels of circulating cytokines and hormones, particularly cortisol, in participants from both MY-LC and External LC cohorts. Persistently lower cortisol levels in patients with LC more than a year after acute infection warrants further investigation, albeit some studies have associated decreased cortisol levels during the early phases of COVID-19 with the development of respiratory LC symptoms. 

Participants with LC from two sites had markedly lowered systemic cortisol levels; however, this was not associated with a compensatory exacerbation in adrenocorticotropic hormone (ACTH) levels, suggesting LC likely blunted the hypothalamic-pituitary axis response to regulate cortisol. As ACTH has an extremely short half-life in plasma, future studies should confirm these preliminary findings.

Unbiased ML models detected potential targets for future LC biomarker development after matching LC and controls to minimize confounding due to gender, age, and vaccination status. Furthermore, the authors showed that patient-reported outcomes and immunological analyses were highly concordant in LC diagnoses. 

Conclusions

To summarize, the current study remarkably distinguished biological differences between participants with LC and matched controls and HCWs. Thus, future investigations into the immunological underpinnings of LC must continue.