Article Text

Download PDFPDF

In the shadow of antibodies: how T cells defend against COVID-19
  1. David S Pisetsky1,
  2. Kevin L Winthrop2
  1. 1 Department of Medicine and Immunology and Medical Research Service, Duke University Medical Center and Veterans Administration Medical Center, Durham, North Carolina, USA
  2. 2 School of Public Health, Oregon Health & Science University, Portland, Oregon, USA
  1. Correspondence to Dr David S Pisetsky, Department of Medicine and Immunology and Medical Research Service, Duke University Medical Center and Veterans Administration Medical Center, Durham, North Carolina, USA; david.pisetsky{at}

Statistics from

The coronaviruses are a diverse family of single-stranded RNA viruses that underlie conditions that are variably endemic, epidemic or pandemic. These conditions also differ markedly in severity. At one extreme, coronaviruses cause the common cold, an upper respiratory infection that, while symptomatic, is more bothersome than concerning. At the other extreme, SARS-CoV-2 causes a lower respiratory infection that has led to millions of deaths from devastating complications such as adult respiratory distress syndrome, cytokine storm and immunothrombosis.1 2 In addition to its impact on individual patients, the COVID-19 pandemic has had major economic, societal and political repercussions that may persist well into the future.

The variability of coronavirus infections extends beyond the pattern and epidemiology of disease to the effects on individual patients. For COVID-19, approximately 80% or more of infected individuals will have mild-to-moderate symptoms; they can even be asymptomatic. The other patients will have severe respiratory involvement requiring hospitalisation and often have dire complications that necessitate intensive care, intubation and a wide variety of interventions that have included biologics and targeted immunosuppressives familiar to rheumatologists. Fortunately, the armamentarium of therapies continuously grows, with new antiviral agents able to attenuate infection and allow home treatment.3 4

The range of illnesses experienced by patients with COVID-19 is remarkable, raising important questions about the determinants of outcome. Certain facts are clear. Disease is worse in older individuals, men especially, and the presence of comorbidities such as hypertension and obesity; among other factors influencing outcomes, autoantibodies against type 1 interferons have been associated with more severe disease.5 Unlike other infections that display a U-shaped pattern of risk by age (worse in children and older individuals), COVID-19 relatively spares children, with increasing age the dominant factor in outcome. Interestingly, in the influenza pandemic of 1918, young adults were the most seriously affected, suggesting that viruses differ in their interaction with host factors.6 7

As is the case in other viral infections, defence against SARS-CoV-2 involves the innate as well as adaptive immune system, B cells as well as T cells. While the cellular elements of the immune system work in concert, their roles can be distinct and also may change during the course of disease.8–12 Elucidating the dynamics of these cell populations is important for understanding the determinants of outcome as well as for the design of strategies for treatment and prevention. Hundreds if not thousands of scientific papers have provided an extraordinary picture of the orchestration of cellular responses before, during and after infection with SARS-CoV-2.

The state of the immune system prior to infection with SARS-CoV-2 is a topic of great interest as it may have an impact on future disease course.13–15 Indeed, as shown in a fascinating study published in Nature by Swadling et al, a population of pre-existing virus-specific T cells may be critical in the early stages of the infection, acting in some people to abort viral replication; the rapidity of the response may thereby prevent full-blown infection and even short-circuit the development of an antibody production.16 As a result, seronegativity may represent a successful host response to infecting viruses as well as a lack of viral infection.

In their study, Swadling et al focused on healthcare workers (HCW) who had been intensely monitored during the first wave of the infection and were found to be negative for the virus by PCR, as well as negative for antibody by anti-spike-1 IgG, IgG and IgM antibodies to the nucleoprotein (NP), and neutralisation assays. These individuals were designated as seronegative or seronegative HCW (SN-HCW). As an explanation for seronegativity in the SN-HCW as well as their lack of virus by PCR, the study explored the possibility that these individuals had pre-existent memory T cells that, prior to the onset of antibody production, could rapidly terminate infection because of cross-reactivity to SARS-CoV-2 protein(s).

To delineate the contribution of T cells in the response of SN-HCW to the virus, the investigators analysed the responses of T cells in peripheral blood by in vitro enzyme-linked immune absorbent spot (ELISPOT) assays; for antigens, the investigators used overlapping peptides for structural proteins (spike, membrane, NP and ORF3a) as well as non-structural proteins, most importantly proteins in the replication transcription complex (RTC) that is transcribed early after infection of cells. These proteins include NSP7, SP12 and NSP13. Control populations included cohorts prior to the pandemic as well as a concurrent population that had acquired infection.

The results of the analysis were striking. Thus, despite the absence of measurable antibodies, the SN-HCW had strong responses from multispecific memory T cells; frequencies of these cells were greater than those of the unexposed, prepandemic cohort. Furthermore, the T-cell responses of the SN-HCW showed greater reactivity against the RTC antigens than the responses of matched concurrent infected individuals. Among the SN-HCW population, those with the strong T-cell responses to RTC proteins showed an increase in levels of the interferon-inducible transcript IFI27 in blood as demonstrated by PCR. As this transcript is increased with infection,17 these results provide evidence that the SN-SWC had been infected with the virus but somehow mounted a successful antiviral defence that did not entail detectable B-cell response. Figure 1 depicts the pattern of responses of SN-HCW.

Figure 1

Immunological findings in seronegative healthcare workers (SN-HCW) following infection with SARS-CoV-2. The figure highlights the immune responses of individuals (SN-HCW) who remain seronegative following viral infection. As the figure indicates, infection leads to an increase in T cells specific for the replication transcription complex (RTC) as well as increased expression of the IFI27 gene. Other markers of infection are lacking.

An immediate question, therefore, concerns the origin of the pre-existing T cells that can abort or terminate infection. As the authors suggest, the most likely explanation for the RTC-reactive T cells is prior infection by a coronavirus; in this scenario, the non-structural proteins involved in early stages of viral replication likely show greater conservation and homology than structural proteins such as the spike protein; these proteins may, therefore, induce cross-reactive T cells more readily. As seasonal infection by coronaviruses is very common, there is abundant opportunity for the induction of a cross-reactive T-cell population that can act in infection with SARS-CoV-2 and potentially other coronaviruses. Given the frequency of coronavirus infection, it will be important to determine why memory T cells are not present more commonly in the general population although age may be a factor.18

Another potential setting to elucidate the interplay of B cells and T cells in defence against the virus concerns vaccination of patients who have received rituximab (RTX) to deplete CD20-positive B cells for the treatment with autoimmune and inflammatory disease. These studies have indicated that, in the absence of B cells induced by RTX treatment, antibody responses are diminished as would be expected.19–21 The effects on T-cell responses are more complex, however, and may depend on the vaccine administered and the manner in which T-cell responses are assessed. In addition, the generation of T cells may be affected by the effects of comedication as well as underlying immunological disturbances of this patient population. Nonetheless, these studies suggest that the majority of patients who have been treated with RTX and then are vaccinated can mount measurable T-cell responses; in some patients, T-cell responses can occur even in the absence of measurable humoral responses. Studies of this kind are very relevant for rheumatologists who are concerned about the timing of vaccine administration and, ultimately, utilisation of RTX as an immunomodulatory agent.

This paper provides a rich source of data relevant to many aspects of the COVID-19 pandemic. Certainly, from the perspective of epidemiology, the results indicate that PCR and antibody assays may not invariably detect infection since, as the paper shows, infected individuals can lack these biomarkers of infection. Determining infection on the basis of T-cell reactivity, however, is inherently more complicated than assays for antibodies, requiring large peptide arrays to get adequate coverage of potential antigenic sites on any given protein. Similarly, assay for IFI27 is based on PCR determinations of peripheral blood cells, which involve additional technology.17 Further studies will also be needed to determine the range of viral infections for which IFI27 transcripts are elevated to avoid false positive results.

The findings in this paper are also highly relevant for the design of vaccines. Current vaccines focus on structural proteins such as the spike protein and, thus, can induce antibodies to prevent viral attachment to cells for entry; the duration of antibody responses induced vaccination (as well as infection) is a potential limitation of this approach that can at least be addressed by booster injections.22–24 Although current vaccines also induce T cells,25 26 a vaccine targeting the RTC would involve a fundamentally different strategy; rather than blocking viral entry, the induced T cells would target infected cells, killing them before viral replication advances. As the RTC proteins are likely to be conserved, vaccines based on the induction of T cells to these proteins may have broad applicability, capable of preventing illnesses such as SARS and MERS as well as COVID-19 and whatever variants that may emerge as the pandemic evolves. Furthermore, a vaccine designed to induce T cells to the RTC could be a valuable option for patients treated with RTX, a consideration important for rheumatology as well as other subspecialties using B-cell depletion therapeutically.

The exploration of new targets of vaccines always comes with safety concerns but there are also practical aspects. Given the effectiveness of current vaccines, vaccines that target T cells would presumably be tested and ultimately deployed together with vaccines targeting the spike or other non-spike proteins. A recent study of household contacts exposed to SARS-CoV-2 showed that pre-existing cross-reactive T cells from exposure to other coronavirus were highly correlated with protection from becoming PCR positive for infection. Interestingly, in this study, non-spike protein elicited T-cell responses, further suggesting the importance of including non-spike antigens in future vaccines.15 Demonstrating the efficacy of such vaccines may become challenging, however, as time passes and the risk of infection in the population dwindles. Investigating these issues will be an important goal for the future and can build on the seminal findings by Swadling et al which show why, despite the potential of SARS-CoV-2 for devastation, some infected people simply do not get sick.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.



  • Handling editor Josef S Smolen

  • DSP and KLW contributed equally.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.