The role of immune activation and antigen persistence in acute and long COVID

Abstract

In late 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) triggered the global coronavirus disease 2019 (COVID-19) pandemic. Although most infections cause a self-limited syndrome comparable to other upper respiratory viral pathogens, a portion of individuals develop severe illness leading to substantial morbidity and mortality. Furthermore, an estimated 10%–20% of SARS-CoV-2 infections are followed by post-acute sequelae of COVID-19 (PASC), or long COVID. Long COVID is associated with a wide variety of clinical manifestations including cardiopulmonary complications, persistent fatigue, and neurocognitive dysfunction. Severe acute COVID-19 is associated with hyperactivation and increased inflammation, which may be an underlying cause of long COVID in a subset of individuals. However, the immunologic mechanisms driving long COVID development are still under investigation. Early in the pandemic, our group and others observed immune dysregulation persisted into convalescence after acute COVID-19. We subsequently observed persistent immune dysregulation in a cohort of individuals experiencing long COVID. We demonstrated increased SARS-CoV-2-specific CD4⁺ and CD8⁺ T-cell responses and antibody affinity in patients experiencing long COVID symptoms. These data suggest a portion of long COVID symptoms may be due to chronic immune activation and the presence of persistent SARS-CoV-2 antigen. This review summarizes the COVID-19 literature to date detailing acute COVID-19 and convalescence and how these observations relate to the development of long COVID. In addition, we discuss recent findings in support of persistent antigen and the evidence that this phenomenon contributes to local and systemic inflammation and the heterogeneous nature of clinical manifestations seen in long COVID.

Summary

WHAT IS ALREADY KNOWN ON THIS TOPIC:

• Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes severe clinical impact and death.

• Many patients with severe SARS-CoV-2 infection exhibit extreme immune perturbations.

WHAT ARE THE NEW FINDINGS:

• Immune dysregulation persists beyond the acute stages of coronavirus disease 2019 (COVID-19).

• Many patients with long COVID exhibit a immunological signature that includes sustained S-protein specific T-cell activation.

HOW MIGHT THESE FINDINGS IMPACT CLINICAL PRACTICE?

• Further investigation into the mechanisms behind long COVID may provide information about immunological targets for symptom resolution or long COVID prevention.

Introduction

The coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has killed over 6 million people globally to date. ¹ While most people infected with SARS-CoV-2 have self-limited disease and recover within 1–2 weeks, approximately 20% developed more severe illness in the initial waves of COVID-19, with respiratory involvement leading to hospitalization, respiratory failure, and increased mortality.^2–9 While immunity acquired through natural infection and vaccination has decreased the frequency and presentation of severe disease, COVID-19 continues to be associated with hospitalizations and death. In addition to the acute disease process, a portion of people develop post-acute sequelae of COVID-19 (PASC), also referred to as long COVID.^10,11 The phenomenon of post-acute sequelae is not unique to SARS-CoV-2, as post-viral syndromes have been previously recognized in other viruses such as SARS-CoV-1, H1N1 influenza, Epstein-Barr virus (EBV), and others. ¹² Many post-viral syndromes are associated with neurocognitive dysfunction and rheumatologic symptoms, which could suggest a common underlying etiology driving chronic symptoms. However, the extraordinary number of global infections over a short period of time seen during the COVID-19 pandemic has led to a high burden of individuals experiencing persistent and sometimes debilitating syndromes.

SARS-CoV-2 is an enveloped, positive-sense RNA virus belonging to the Coronaviridae family. Notably, coronaviruses have caused two recent outbreaks: SARS-CoV-1 in 2003 and Middle East respiratory syndrome (MERS-CoV) in 2012. SARS-CoV-1 binds host cells via the angiotensin-converting enzyme 2 (ACE2) receptor, whereas MERS-CoV utilizes dipeptidylpeptidase 4 (DPP4).^13,14 Similar to SARS-CoV-1, SARS-CoV-2 utilizes host cell receptor ACE2.^15,16 ACE2 plays a homeostatic role in the renin–angiotensin–aldosterone system by regulating angiotensin II levels. ¹⁷ Notably, this receptor is found on various cell types throughout the body, contributing to the heterogeneous nature of clinical manifestations seen in both acute COVID-19 and long COVID.^18–25 The SARS-CoV-1 epidemic infected approximately 8000 people with a reported case fatality rate of ~10% across 30 different countries before it was ultimately contained within 1 year. ²⁶ MERS-CoV has infected approximately 2500 people across 28 countries and appears to be the most lethal of the three coronaviruses with an estimated case-fatality rate of ∼34%. ²⁷ SARS-CoV-2, while having a lower case fatality rate than SARS-CoV-1 and MERS-CoV, is much more transmissible, contributing to its lack of containment and subsequent global spread. ²⁸ Notably, both SARS-CoV-1 and MERS-CoV were associated with prolonged symptom duration and post-viral syndromes.^29,30

While the clinical features and underlying immunology for severe acute COVID-19 have been extensively studied since its recognition in late 2019, the mechanisms and immunology driving long COVID have yet to be fully elucidated. Earlier in the pandemic, our group and others found that immune dysregulation persisted into convalescence after acute SARS-CoV-2 infection. ³¹ In a subsequent study, we observed similar findings in individuals with long COVID. ³² We proposed that a portion of this persistent dysregulation was likely driven by viral reservoirs leading to chronic immune activation and inflammation. Recently, other groups have found evidence supporting the presence of persistent antigen after acute COVID-19, and some have tied this finding to the development of long COVID.^33–42 In this review, we provide a summary of findings during acute COVID-19 infection and convalescence and how these observations relate to the potential role of persistent antigen contributing to long COVID.

Acute COVID-19

Clinical presentation and risk factors

COVID-19 symptoms typically begin around 3- to 6-day post-infection and can involve most organ systems (e.g., respiratory, gastrointestinal, cardiovascular, neurocognitive).^4,5,7,8,43 Most commonly reported symptoms include cough, myalgias, headache, and nasal congestion.^44–47 Self-limited COVID-19 illness often resolves in 7–10 days^5,48,49; however, lower respiratory tract involvement can progress to acute respiratory distress syndrome, leading to the need for mechanical ventilation and dramatic increases in morbidity and mortality.^5,48,50 Early in the pandemic, reports emerged that over half of cases progressed to severe disease.^5,49,51 As the true prevalence became clearer, severity incidence settled around 15%–20%.^2–4,6–9 Risk factors for severe disease include older age, male sex, and a variety of comorbid conditions (e.g., immunocompromised states, metabolic disorders).^{2–5,8,9,52,53} Since its origins in late 2019, SARS-CoV-2 has continued to mutate, leading to various strains with enhanced transmission. Each wave is associated with sequential dominant variants, but timing of surges varied by geographic location. To date, there is limited data comparing the immunologic differences in COVID-19 based on the dominant variant at the time, especially with more recent variants such as Omicron.^54–63 Over time, SARS-CoV-2 has appeared to become less virulent while transmissibility has increased, although this is confounded by accumulated population immunity.^{44,58,59,61,63–68} Even though SARS-CoV-2 can have a longer incubation period than some other respiratory viruses, viremia appears to peak at time of symptom onset. Furthermore, viral shedding has been noted in asymptomatic COVID-19 patients, contributing to increased transmission.^36,69–73

T-cell immune responses during acute illness

The immunopathology behind severe COVID-19 has been intensively studied since the beginning of the pandemic (Figure 1). Investigations have shown that severe illness is associated with delays in mounting an effective adaptive immune response, caused by prominent lymphopenia with concurrent hyperactivation and inflammation promoting tissue damage that leads to hospitalization and potential mortality.^74–82 Initial reports on the differences in clinical outcomes with SARS-CoV-2 showed elevations in inflammatory cytokine levels such as interleukin (IL)-6, inducible protein 10, IL-10, and tumor necrosis factor alpha in patients with severe illness. In addition, prominent lymphopenia was associated with increased severity.^49,83–93 One of the first larger-scale analyses, including over 500 hospitalized patients, observed that severe disease was associated with depletion of total, CD4⁺, and CD8⁺ T cells, as well as increases in various inflammatory cytokines compared to milder illness. ⁹⁴ They also found that more severe disease was associated with increased expression of programmed cell death protein 1 (PD1), a regulatory molecule that can be a sign of cellular activation and exhaustion. Mathew et al. ⁹⁵ analyzed similar metrics in acute COVID-19 patients and compared them to patients who recovered from COVID-19 and uninfected controls. While they noted similar findings of lymphopenia and increased inflammation, they also analyzed various T-cell populations and found a decrease in naïve CD4⁺ T cells, an increase in activated and terminally differentiated CD4⁺ and CD8⁺ T cells, and an increase in circulating T-follicular helper (cTfh) cells. Subsequent studies analyzed more extensive panels of cell surface markers and noted functional exhaustion in CD4⁺ and CD8⁺ T cells, increased expression of regulatory markers (e.g., PD1, TIGIT, TIM3), and increased expression of activation markers (e.g., HLA-DR) in severe disease.^48,94–101

Figure 1.

Immune dysregulation of non-hospitalized vs hospitalized individuals during acute SARS-CoV-2 infection.

(Blue) Mild illness is associated with a timely immune response, indicated by appropriate activation of the adaptive immune response. This is evident by an increase in cytokine production and upregulation of activation markers on the surface of T cells (non-hospitalized visit 1). While mild illness can lead to a return to baseline function (blue dashed line), we observed that non-hospitalized patients had increased expression of activation and exhaustion markers 1-month post-infection (non-hospitalized visit 2, blue dotted line), indicating persistent immune dysregulation. (Orange/Red) Delayed viral clearance contributes to hyperactivation and worsening of symptoms, leading to hospitalization and mortality. This hyperactivation is associated with upregulation of several activation and exhaustion markers on the surface of T cells, as well as an increase in inflammatory cytokine production (hospitalized/mortality). Furthermore, persistent hyperactivation can lead to eventual exhaustion (hypoactivation) and either recovery (orange dashed line) or mortality (red dashed line).

SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

While these early reports focused on severe presentations of COVID-19, we sought to explore differences in the immune response between hospitalized and non-hospitalized patients during acute infection. To do this, we obtained peripheral blood samples from 85 COVID-19 infected donors (46 hospitalized and 39 non-hospitalized) within the early months of the pandemic. For detailed protocol, please see Files et al. ³¹ In brief, peripheral blood mononuclear cells (PBMCs) were isolated and washed from each sample. 100K PBMCs from each sample were then stained with immunological markers, including CD3, CD4, CD8, OX40, and CD69, and analyzed on a BD FACS Symphony flow cytometer. All data collected were in accordance with the University of Alabama at Birmingham’s Institutional Review Board. We observed that hospitalized patients had increased expression of activation markers OX40 and CD69 on CD4⁺ and CD8⁺ T cells, respectively, compared to both non-hospitalized patients and uninfected controls (Figure 2(a) and (b)). Hospitalized patients had increased expression of several other activation markers on T cells, including HLA-DR, CD38, and CD-154. We also found increased expression of classical exhaustion markers including PD1 and TIM3 in hospitalized patients, while TIGIT was increased on T cells from COVID-19 patients regardless of hospitalization status. ³¹ Around the same time, other groups observed that non-hospitalized patients had increased expression of activation markers compared to uninfected controls albeit lower than hospitalized patients.^102–105 Carsetti et al. ¹⁰² observed increases in HLA-DR expression on CD4⁺ T cells from both mild and severe COVID-19 patients. Oja et al. ¹⁰⁴ noted similar increases in PD1 in severe COVID-19 patients.

Figure 2.

Expression of activation markers on CD4+ and CD8+ T cells during acute COVID-19 infection. (a) OX40 expression on CD4+ T cells and (b) CD69 expression on CD8+ T cells from healthy, pre-pandemic controls (“CoV-”), hospitalized COVID-19-infected samples (“H”), and non-hospitalized COVID-19-infected samples (“NH”). All samples collected during acute infection and early convalescent time points. (c) OX40 expression on CD4+ T cells and (d) CD69 expression on CD8+ T cells from longitudinal non-hospitalized patient samples.

Visit number is shown on the x-axis and is represented by “1” and “2” for “Visit 1” and “Visit 2,” respectively. All data modified from Files et al. ³¹

***p < 0.001; significance determined by unpaired Wilcoxon tests (a and b) or paired Wilcoxon tests (c and d).

COVID-19, corona virus 2019.

B-cell and antibody immune responses during acute illness

We also explored differences in the B-cell and humoral responses between hospitalized and non-hospitalized patients. Notably, we found increased expression of activation markers CD69 and CD95 on B cells from hospitalized patients compared to both non-hospitalized patients and uninfected controls. ³¹ Other groups noted similar findings of increased expression of activation markers, as well as decreases in total B cells.^{95,102,106–108} Mathew et al. ⁹⁵ observed an increase in CD27+CD38+ plasmablasts, comprising >30% of circulating B cells in some patients. Similarly, Carsetti et al. ¹⁰² observed an increase in plasmablasts in relation to more severe illness. They also noted that symptomatic patients, regardless of severity, had increased antibody production, and that this production seemed to persist into early convalescence. ¹⁰² This increase in antibody production and expansion of plasmablasts aligns with the hyperactivation seen in severe illness. This may be in part due to an initial delay in effective adaptive immune responses, leading to immune hyperactivation and perhaps contributing to viral persistence.

While these data evaluated the humoral responses of COVID-19-infected donors obtained prior to COVID-19 vaccination efforts, more recent studies have investigated SARS-CoV-2-specific humoral responses in vaccinated and unvaccinated individuals. Vaccination leads to an overall lower rate of infection,^109–112 although breakthrough infection is associated with a weaker humoral response against variants such as Delta and Omicron.^113–117 In addition, symptomatic breakthrough infection was associated with higher antibody levels than asymptomatic infection with mainly the Delta variant. ¹¹⁸ However, increased antibody responses were also found in vaccinated individuals with no infection vs unvaccinated individuals with Delta variant infection. ¹¹⁹ These data indicate that humoral immunity should continue to be assessed against the various SARS-CoV-2 variants, as well as how newer vaccination booster products impact the immune response to circulating and future strains.

Prolonged immune responses during convalescence

Another area of active inquiry has been the longitudinal impact of SARS-CoV-2 infection on T-cell activation following initial infection. Using the same protocol as described above and in Files et al., ³¹ we observed that many non-hospitalized patients exhibited sustained activation into the convalescent period. In longitudinal samples collected from two visit timepoints in non-hospitalized patients, we measured activation and exhaustion markers into early convalescence. Notably, all samples were collected within the first 40-day post-symptom onset. We found evidence that immune dysregulation persisted regardless of initial symptom severity as shown by the increased OX40 expression on CD4⁺ T cells and increased CD69 expression on CD8⁺ T cells at the later visit 2 timepoint (Figure 2(c) and (d)). ³¹ We observed a similar phenomenon with the activation marker HLA-DR and exhaustion markers TIM3 and TIGIT expression on T cells. Other groups have reported persistent adaptive immune system dysregulation into convalescence, notably changes in cell populations toward a more inflammatory, reactive state with increases in pro-inflammatory cytokines from various cells, and decreases in naïve T-cell and B-cell populations.^120–127 Multiple hypotheses could explain these findings including the inadequate dampening of the acute inflammatory response to SARS-CoV-2 infection, or a failure to eliminate viral antigen. This continual shift of T-cell populations toward activation states could suggest continual exposure to antigen due to failure to clear the virus. Persistent exposure to antigen leads to the immune system continuously activating various immune cells in response to active infection, thus leading to an increase in activated cell populations. Other viruses associated with persistent antigen such as cytomegalovirus (CMV) and HIV also demonstrate a shift toward effector and terminally differentiated cell populations. CMV is associated with a long-term increase in both memory and effector-memory cells which has been linked to immunosenescence.^128–130 HIV is associated with an increase in effector cells during acute infection, and an increase in effector and exhausted cells in chronic infection, with and without antiretroviral therapy.^131–138

There are reports of persistent SARS-CoV-2-specific humoral and cellular responses over 8-month post-infection, regardless of initial severity, demonstrating that the body sustains longitudinal T- and B-cell immunity to SARS-CoV-2.^120,121,124 Specifically, one group observed elevated SARS-CoV-2-specific CD4⁺ T-cell responses in COVID-19 patients, regardless of acute disease severity. ¹⁰⁴ Our group and others found SARS-CoV-2-specific cTfh responses correlate with antibody neutralization.^139–142 cTfh are a specific subset of CD4⁺ T cells that support humoral responses such as the development of neutralizing antibodies.^143–150 Previous work on memory responses to seasonal and pandemic coronaviruses demonstrated durable memory T-cell responses to SARS-CoV-1 lasting as long as 17-year post-infection and to SARS-CoV-2 over 1-year post-infection. ¹⁵¹ In addition, other studies have shown cross-reactive antibodies to SARS-CoV-1 and seasonal coronaviruses in SARS-CoV-2 convalescent patients, as well as in SARS-CoV-2-uninfected controls, demonstrating conserved epitopes between coronaviruses and further suggesting long-lasting immunity.^152,153 One potential mechanism for antigen persistence is cross-reactive responses from prior coronavirus infections due to a lack of naïve lymphocytes, resulting in suboptimal adaptive responses to SARS-CoV-2. Based on these observations, we hypothesized that viral reservoirs may contribute to the persistent immune activation and dysregulation observed in early convalescence and could provide the basis for a subset of long COVID presentations.

An early observation from the COVID-19 pandemic was the phenomenon of persistently positive PCR-based testing, particularly when sampling the lower airway. Although this was often a consequence residual viral proteins and high-sensitivity PCR-based testing platforms, recovery of replication competent virus was observed, particularly in immunosuppressed hosts. ¹⁵⁴ Several groups have observed prolonged viral shedding in patients who recovered from SARS-CoV-2 infection with resolution of symptoms.^155–159 Persistent SARS-CoV-2 RNAemia has been associated with severe illness.^160,161 Similarly, Cai et al. ⁸³ noted that slower viral clearance was associated with progression to critical illness, suggesting persistent virus contributes to prolonged immune activation, inflammation, and ultimately tissue damage. The recognition of post-acute sequelae syndromes raises the possibility that delays in eliminating viral proteins may contribute to persistent symptoms.

Long COVID

Clinical presentation and risk factors

While many individuals recover fully from acute SARS-CoV-2 infection, a portion of individuals develop persistent symptoms that can last for weeks or months now referred to as long COVID or PASC. Risk factors for long COVID development include severe acute illness, older age, female sex, lack of vaccination, and various comorbidities.^162–167 Long COVID has a wide variety of manifestations including cardiopulmonary complications, persistent fatigue, and neurocognitive dysfunction, among a diverse array of persistent syndromes.^{164–166,168–171} Many early characterizations of long COVID reported varying symptom durations following initial diagnosis, ranging from 4 weeks to 4 months. The heterogeneous nature of long COVID and inconsistencies in defining symptom duration for diagnosis have complicated initial efforts to estimate overall prevalence. Prior to a broadly accepted definition, early reports stated long COVID prevalence as low as 9% and as high as 81%. ¹⁷² The World Health Organization has since defined long COVID as new or persistent symptoms lasting for at least 2 months and occurring at least 3-month post-infection. ¹⁰ Recent studies have reported long COVID prevalence to be 10%–20%.^10,11 These data were largely published using cohorts from early variants. Thus, the frequency of severe disease, a long COVID risk factor, was greater than what is being observed in subsequent waves.^{44,63–68,173} Similarly, population immunity was minimal during early phases of COVID-19, but through a patchwork of vaccination and repeated natural infection, that is no longer the case. More recent reports observed that, despite SARS-CoV-2 overall causing less severe disease, presentations consistent with long COVID continue to arise, albeit at lower frequencies than prior variants.^63,64,174 Data regarding long COVID risk after vaccination are underexplored, though a recent report observed that fully vaccinated patients with breakthrough infection had only a slightly reduced risk compared to unvaccinated individuals (HR = 0.85 (95% CI 0.82, 0.89)). ¹⁶² Although the massive burden of long COVID is now appreciated, the mechanisms and associated immune responses driving the clinical sequelae of long COVID have been less characterized.

Adaptive immunity during long COVID

To explore potential immunologic signatures of long COVID, our group evaluated immune responses in a cohort of 50 patients infected with COVID-19 during the year 2020, meaning all samples were obtained prior to vaccination. A detailed overview of the protocols and the cohort used is given in Files et al. ³¹ In brief, longitudinal samples were collected from patients at three differing timepoints: (1) early convalescence (75 or less days from initial symptom onset), (2) intermediate convalescence (76–150 days from symptom onset), and (3) late convalescence (151+ days from symptom onset). Of note, 20 of these patients had persistent symptoms for at least 4-week post-infection, thus meeting the criteria for long COVID. The remaining 30 individuals experienced symptom resolution within 14 days of initial symptom onset. ³² All data collected were in accordance with the University of Alabama at Birmingham’s Institutional Review Board. We hypothesized that individuals with long COVID would exhibit persistent immune dysregulation that our group and others had observed in convalescent COVID-19 samples. However, our findings showed minimal systemic differences between recovered and long COVID groups in overall immune dysregulation between early, intermediate, or late convalescent timepoints. Using assays similar to our previous findings, isolated PMBCs from each timepoint were stained with various immune markers and analyzed on a BD FACS Symphony flow cytometer. Representative data showing OX40 expression on CD4⁺ T cells and CD69 expression on CD8⁺ T cells between the two groups are shown in Figure 3(a) and (b), respectively. No significant differences were identified between patients with prolonged symptoms (long COVID) or recovered infection. Additional T-cell activation and exhaustion markers were also investigated, including HLA-DR, CD38, CD154, PD1, TIGIT, and TIM3, as well as other immune cell populations including B cells and monocytes, but no statistically significant differences were observed.

Figure 3.

Expression of activation markers on CD4+ and CD8+ T cells during long COVID. (a) OX40 expression on CD4+ T cells and (b) CD69 expression on CD8+ T cells from patients with prolonged symptoms (long COVID; red) and recovered COVID-19 samples (dark blue).

Samples collected at early convalescent (“EC”), intermediate convalescent (“IC”), or late convalescent (“LC”) timepoints. All data modified from Files et al. ³¹

***p < 0.001; significance determined by unpaired Wilcoxon tests.

In this study, our group also investigated SARS-CoV-2-specific T-cell and antibody responses to the spike protein. More detailed protocol can be obtained in Files et al. ³¹ In brief, prepared peptides covering the whole length of the SARS-CoV-2 spike protein were prepared. PBMCs from each sample at each timepoint were stimulated with spike protein peptides for 18 h. Samples were then stained with various immunological markers including OX40, PDL1, CD69, and CD137.^175,176 Samples were analyzed on a BD FACS Symphony flow cytometer. In agreement with numerous other studies at the time, we observed persistent SARS-CoV-2-specific T-cell activation throughout our cohort as far as 6 month post-infection.^{41,121,177–179} Many of these studies, including Dan et al., ¹²¹ demonstrated a decay in the T-cell response magnitude over time. We observed a similar decrease in CD4⁺ T-cell response magnitude in individuals from our recovered group (Figure 4(a) and (b); shown in blue). In contrast, those experiencing long COVID demonstrated a sustained SARS-CoV-2-specific CD4⁺ T-cell response magnitude over time (Figure 4(a) and (b); shown in red). The differences in antigen-specific T-cell responses were statistically significant for response magnitudes at the late convalescent timepoint. Overall, these observations suggest patients with long COVID have increased SARS-CoV-2-specific CD4⁺ and CD8⁺ T-cell magnitude responses during late convalescence as compared to those who recovered quickly and without persistent symptoms. Several studies have subsequently shown similar findings in patients with long COVID at various timepoints.^{35,41,121,177,179–188} Galan et al. ¹⁸⁰ observed increased levels of functional memory CD8⁺ T cells, increased CD4⁺ regulatory T cells, and increased expression of PD1 on T cells in long COVID patients 7-month post-infection. Similarly, Peluso et al. ¹⁷⁷ noted increased SARS-CoV-2-specific adaptive immune responses in long COVID patients, regardless of their initial disease severity, as far as 9-month post-infection. In addition, Phetsouphanh et al. ¹⁸³ noted similar findings of SARS-CoV-2-specific adaptive immune responses as well as a lack a naïve T and B cells in long COVID patients at 8-month post-infection. Taken together, these data demonstrate that antigen-specific responses are sustained in long COVID patients at levels significantly higher than individuals with prompt resolution of symptoms. This observation raises the possibility that persistent viral reservoirs are maintained and drive the persistence of SARS-CoV-2 specific responses seen in long COVID patients.

Figure 4.

Antigen-specific CD4+ T-cell responses are sustained into late convalescence in long COVID-19 patients. (a) OX40+PDL1+ expression on CD4+ T cells following stimulation with S-protein peptide pool from patients with prolonged symptoms (long COVID; red) and recovered COVID-19 samples (dark blue). Samples collected at early convalescent (“EC”), intermediate convalescent (“IC”), or late convalescent (“LC”) timepoints. (b) OX40+PDL1+ expression on CD4+ T cells following stimulation with S-protein peptide pool shown as days post-symptom onset.

All data modified from Files et al. ³¹

**p < 0.01 and *p < 0.05; significance determined by unpaired Wilcoxon tests (a) and by linear mixed effects model (b).

Localized immune responses during long COVID

One limitation of our study was that it exclusively investigated immune signatures from peripheral blood samples. Other groups have defined tissue-specific immunopathology and shown similar findings of persistent immune activation and inflammation.^{105,178,182,183,189–193} Several have now observed persistent immune activation and elevation of proteins associated with apoptosis, tissue repair, and epithelial injury in the lungs,^{105,193–202} while others have observed persistent immune activation and inflammation in the central nervous system (CNS) of patients with persistent neurocognitive symptoms.^203–206 Apple et al. ²⁰⁶ proposed that persistent immune activation and inflammation in the CNS may contribute to neurocognitive manifestations of long COVID after observing elevated inflammatory markers in the CSF from patients with long COVID neurocognitive symptoms. In line with these organ-specific findings, SARS-CoV-2 antigen has been recovered from multiple body sites including the gastrointestinal tract, hepatic tissues, the brain, lymphoid tissues, and in circulation up to 7-month post-infection^{34,40,207,208} Some groups proposed that, much like other chronic infections, persistent antigen exposure leads to chronic immune activation and subsequent systemic and compartmentalized inflammation, contributing to the wide array of clinical manifestations seen in long COVID.^{105,165,177,178,181,183,189–193,209–212}

Post-acute sequalae due to antigen persistence

Antigen persistence and subsequent clinical sequelae post-infection are not novel to SARS-CoV-2 as infectious causes of chronic autoimmune and inflammatory disorders have been explored for decades.^12,213–217 Chronic infections such as HIV, CMV, and EBV are associated with viral latency and persistence, leading to chronic exposure to viral antigens and subsequent immune activation and inflammation, contributing to the various sequelae of these infections. For example, HIV is known to cause accelerated aging due to chronic immune activation leading to early senescence and exhaustion of various immune cell populations.^{136–138,218–220} This persistent inflammation contributes to earlier development of a variety of comorbidities such as cardiometabolic disorders and neurocognitive dysfunction.^{136,219,221,222} EBV and CMV are two other latent viruses known to become reactivated in immunocompromised hosts and contribute to continued damage and inflammation throughout the body. ²²³ Latent viral infections are associated with continual, low-level viremia that can lead to reactivation and subsequent symptoms. This low-level, often compartmentalized viremia can continuously activate the immune system, contributing to negative effects later in life such as increased risk of comorbidities and early senescence and exhaustion of various immune cell populations. Other viruses such as influenza, Ebola, Coxsackie, and now SARS-CoV-2 have been associated with post-acute sequelae, and some have demonstrated antigen persistence.^12,224–232

For example, prolonged viral shedding has been observed after Ebola and West Nile virus infection for several years after acute infection.^224–227 Similarly, several papers have provided support for the presence of viral reservoirs for SARS-CoV-2.^{33,34,36–42,207,208} Recent reports have observed persistent RNAemia and circulating SARS-CoV-2 spike in individuals with long COVID.^38,40 Others have observed tissue- and organ-specific SARS-CoV-2 RNA and antigen persistence. Chertow et al. ²⁰⁷ performed autopsies on COVID-19 patients and reported persistent SARS-CoV-2 RNA in a wide array of anatomical locations including the respiratory tract, gastrointestinal tract, cardiovascular tissue, reproductive tissue, peripheral nerves, and several brain regions, almost 8-month post-infection, although this study is biased toward more severe acute disease. In addition, others have demonstrated associations between persistent antigenemia and long COVID manifestations, including in the olfactory bulb perhaps explaining persistent anosmia.^233,234 Zollner et al. ²³⁵ reported SARS-CoV-2 antigen in the gastrointestinal tract of patients who were later diagnosed with irritable bowel disease post-SARS-CoV-2 infection.

Long COVID mechanisms and current evidence

Collectively, these findings strongly suggest that persistent SARS-CoV-2 antigen is a common feature in individuals recovering from COVID-19 infection and plays a role in the pathogenesis of long COVID. As suggested in Figure 5, this persistent viral antigen could explain the ongoing symptoms, persistent inflammation, and the sustained T-cell response magnitude observed in long COVID patients. However, this observation does not exclude a variety of other mechanisms that may also contribute to long COVID symptomatology. While our study focused on immune cell populations, other groups measured various plasma inflammatory markers in patients experiencing long COVID and found marked elevations that could contribute to systemic symptoms such as fever and fatigue.^{105,182,183,190,209,210,212,236} As discussed earlier, acute disease severity strongly correlates with likelihood and burden of long COVID symptoms.^162–167 Symptoms such as dyspnea, anxiety, and frailty commonly occur during severe disease and are at least partially due to end organ damage and fibrosis that commonly occurs during various critical illnesses.^237–244 A few groups have shown evidence of EBV reactivation in patients with long COVID^35,39,245; this could explain persistent fatigue and other long COVID symptoms. Others have shown support for gut dysbiosis, showing perturbations in the microbiome over 6-month post-infection, which may also explain a variety of gastrointestinal-related symptoms.^246–249 Also, a variety of reports have identified signatures of autoimmune responses from patients experiencing long COVID including the evidence of autoantibodies in long COVID patients, whereas others have not.^{35,37,39,250,251} The above processes are not mutually exclusive, resulting in a variety of complex overlapping symptomatic presentations all falling under the umbrella of long COVID. In summary, these findings lend credence to the numerous proposed mechanisms behind long COVID including persistent viral reservoirs, systemic and tissue-specific inflammation, latent virus reactivation, gut dysbiosis, and autoimmunity.^{33,35,37,164,189,250–253} While our data and others have shown support for persistent viral antigen and sustained immune activation and inflammation, there is also support for other proposed mechanisms. The wide array of clinical manifestations of long COVID suggests that there are likely multiple mechanisms driving pathogenesis, which contributes to the complexities of determining the biological basis of the condition. Future investigation is needed to further identify and describe these pathologic mechanisms of long COVID to identify novel therapeutics that may be beneficial in treating patients with prolonged symptoms.

Figure 5.

Proposed mechanism for how viral reservoirs and dysregulated immune responses contribute to the development of long COVID.

Persistent antigen contributes to cell population changes such as a decrease in naïve T and B cells and an increase in effector cells, as well as increased expression of activation and exhaustion markers as the immune system attempts to regulate this chronic activation state. We propose that viral reservoirs contribute to this dysregulation and lead to both localized and systemic inflammation depending on where antigen is located. Several groups have demonstrated antigen persistence in a variety of tissues (e.g., brain, lungs, heart, adipose, gastrointestinal tract) in SARS-CoV-2-infected patients both acutely and chronically. In addition, several groups have demonstrated tissue-specific immune responses and inflammation in relation to corresponding organ-specific symptoms.

COVID, coronavirus disease; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Conclusions

Since the beginning of the COVID-19 pandemic, a large numbers of studies have been performed investigating the immunopathology following infection with SARS-CoV-2. During acute SARS-CoV-2 infection, delayed viral clearance contributes to hyperactivation and worsening of symptoms, leading to hospitalization and mortality. Indeed, our group and others observed increased expression of activation and exhaustion markers on various immune cell populations and increased systemic inflammatory cytokines that persists into convalescence, suggesting sustained immune activation (Figure 1). We also observed sustained SARS-CoV-2-specific T-cell response magnitudes in long COVID patients into late convalescence. Other groups showed similar findings of SARS-CoV-2-specific lymphocytes over 1-year post-infection. We proposed that persistent viral reservoirs could contribute to the immune dysregulation seen in early convalescence and explain the persistent SARS-CoV-2 specific immune responses observed in long COVID patients during late convalescence. This persistent activation could contribute to the localized and systemic inflammation observed in long COVID which could manifest as various sequelae depending on where the persistent antigen is located. Indeed, several groups have demonstrated antigen persistence in a variety of tissues in SARS-CoV-2-infected patients (Figure 5). Further research on determining if certain immunologic or inflammatory signatures correlate with specific long COVID manifestations should be performed, as well as how various prevention and treatment options such as vaccination and monoclonal antibodies influence the development of long COVID.