While impaired SARS-CoV-2-reactive functional T-cell responses have been linked to progression from mild to severe forms of COVID-19 [4,5,7,8,10–15], it remains to be elucidated whether these are associated with clinical outcomes among critically ill patients. Here, we prospectively monitored SARS-CoV-2 S1/M-reactive IFN-γ T-cell responses using an in-house-developed flow cytometry assay in a cohort of 71 patients admitted to ICU, of whom most were mechanically ventilated (88%) and 28 died. Our antigen choice was based upon previously published data showing that a wide array of highly immunogenic T-cell epitopes map within S1 and M proteins that elicit immunodominant responses [2, 4–8]. In addition to further characterizing the dynamics of these T-cell subsets in this population group, which currently remains poorly defined, we aimed to establish whether SARS-CoV-2 S1/M-reactive IFN-γ T-cell counts were related to serum levels of biomarkers predicting poor outcomes and mortality across critically ill patients, and could thus be used as a surrogate prognostic marker. Our data allowed us to draw the following conclusions. First, virtually all patients developed SARS-CoV-2 S1/M-reactive IFN-γ T-cell responses (either CD8+, CD4+ or both) during ICU stay, yet CD4+ T-cell responses were detected more frequently and at higher levels than their CD8+ counterparts. Furthermore, overall, CD4+ T-cell responses appeared to fluctuate over time, while those involving CD8+ T cells tended to wane. Despite this general landscape, we noted wide variations at the individual level. In fact, fluctuating responses were observed more frequently than consistent (either detectable or undetectable) ones over time. Second, data obtained in the rhesus macaque experimental model clearly underscore the crucial role of SARS-CoV-2-reactive T-cell responses in contributing to virus clearance from the lower respiratory tract [27,28]. To determine whether this could be the case in our patient population, we compared the dynamics of SARS-CoV-2 load in TA to that of SARS-CoV-2 S1/M-reactive IFN-γ T cells in paired whole-blood specimens. Our data indicated that although the rate of detection and magnitude of SARS-CoV-2 T-cell responses appeared directly related to the level of virus replication in the lower respiratory tract, as inferred by viral RNA load in TA, the dynamics of virus clearance from this compartment was not consistently associated with that of peripheral blood SARS-CoV-2 S1/M-reactive IFN-γ T cells. This suggested that enumeration of these T-cell subset specificities in whole blood provides no reliable information on the course of virus infection in the lungs. Naturally, our findings do not detract from the role of T cells in affording protection against severe forms of COVID-19, but rather suggest that examination of SARS-CoV-2-driven immune responses at the lower respiratory tract could offer a better perspective of the interplay between virus replication and host immune responses during severe COVID-19. Indeed, different cellular immune profiles in the airways and blood have been documented in critically ill COVID-19 patients [30,31]. Fourth, sustained high serum levels of several biomarkers of inflammation (IL-6, ferritin, CRP), coagulation and fibrinolysis (D-dimer) and tissue damage (LDH) are associated with poor COVID-19 prognosis across critically ill patients [32,33]. Moreover, hyperinflammatory states may down-regulate ongoing T-cell responses [27]. In this context, we investigated whether (qualitative and quantitative) SARS-CoV-2 S1/M-reactive IFN-γ T-cell responses in our patients were somehow related to levels of the aforementioned biomarkers. This was found not to be the case, as serum levels of all biomarkers were similar regardless of detected or absent SARS-CoV-2 CD8+ or CD4+ T-cell responses; moreover, no correlation was found between SARS-CoV-2 T-cell counts and biomarker levels in paired specimens. Fifth, in our cohort, mortality was not consistently associated with either detection rate or the magnitude of SARS-CoV-2 S1/M-reactive IFN-γ T-cell responses. This is in line with data reported by Thieme and colleagues [16], who found that development of robust T-cell responses toward spike, membrane, and nucleocapsid SARS-CoV-2 proteins was not associated with survival in a small cohort of critical COVID-19 patients. In a more comprehensive study, Saris et al. [30] found high levels of TNF-α-producing S-reactive CD8+ T cells to be associated with increased mortality, while mono-functional CD4+ T- cell subsets could not be related to survival; nevertheless, survivors appeared to display broader and stronger virus-reactive poly-functional CD4+ T-cell responses than those who died; yet, as stated by the authors, no obvious combination of effector functions of CD4+ T cells could be linked to prognosis. The key finding of the study was that mucosal-associated invariant T (MAIT) cell activation is an independent and significant predictor of mortality. Likewise, in a very small study critically ill patients with hypertension who died exhibited prolonged low peripheral blood counts of SARS-CoV-2-S-reactive CD8+ and CD4+ T cells [20].
The current study has several limitations deserving comment. First, the limited sample size, particularly regarding the number of deceased patients, clearly undermine the robustness of the analyses. Sufficiently powered studies are needed to elucidate whether monitoring SARS-CoV-2 T-cell responses in peripheral blood may have prognostic value in critically ill COVID-19 patients. Second, although blood specimens were scheduled to be collected weekly, this was unfortunately not achieved in a number of patients. Third, like other flow cytometry-based immunoassays used for measuring SARS-CoV-2 T-cell responses, ours lacks appropriate standardization. Fourth, SARS-CoV-2-reactive T cells were examined only for IFN-γ production, thus we cannot rule out the possibility that other functional T-cell specificities are associated with survival. Also, no data on the state of differentiation of reactive T cells are provided. Fifth, only SARS-CoV-2 S1 and M-reactive T cells were measured; whether enumeration of SARS-CoV-2 T cells targeting other viral proteins may help to individualize mortality risk in critical COVID-19 patients remain to be defined. Sixth, SARS-CoV-2 T-cell responses in the lower respiratory compartment were not assessed. Seventh, most patients were under corticosteroid treatment within sampling times. Eight, the impact of tocilizumab use on serum levels of inflammatory biomarkers was not apparent in our series (not shown), although it cannot be completely dismissed.
In summary, we found no association between peripheral blood levels of SARS-CoV-2-S1/M-reactive IFN-γ CD4+ and CD8+ T cells and viral clearance from the lower respiratory tract, serum levels of biomarkers of poor prognosis and mortality. Further, larger studies centered on resolving these issues are warranted.