Quantification and prognostic significance of interferon-γ secreting SARS-CoV-2 responsive T cells in hospitalized patients with acute COVID-19.

Dear Editor,

We read with interest the article by Martin-Vicente and colleagues, who found no differences in SARS-CoV-2-specific immune humoral responses between patients with well controlled HIV and healthy controls. However, little remains known about the T-cell response during acute COVID-19. We therefore investigated the systemic T-cell response during acute SARS-CoV-2 infection in a hospitalised cohort of patients with COVID-19, using a functional T-cell assay developed to measure T-cell responses to four antigenic domains of SARS-CoV-2.

We conducted a prospective observational cohort study of hospitalised and nosocomially infected adult patients at University Hospitals of Leicester NHS Trust. Patients were eligible if they were 16 years or over, tested positive for SARS-COV-2 on nasopharyngeal RT-PCR using the hospital assay, with no previous history or record of infection and no existing conditions or treatments associated with T-cell immunodeficiency. One 10 ml lithium heparin anticoagulated blood for T-cell functional assay and one 6 ml blood anticoagulated using EDTA to measure antibody response was taken from each study participant within 24 h of a positive routine PCR test. Serology was performed using the commercially available SARS-CoV-2 Total Assay manufactured by Siemens, which detects IgG and IgM to the S1 RBD antigen and gave a qualitative result. To measure T-cell responses, we used the T-SPOT® Discovery SARS-CoV-2 kit (T-SPOT), which uses an ELISpot technology to detect IFN-γ release from T-cells after exposure to four SARS-CoV-2 peptides antigens: Spike protein S1 and S2 domains, Membrane and Nucleoprotein peptides. Routine clinical, radiological, laboratory and demographic data at the time of sampling was collected and prospective outcomes during admission including requirement for CPAP, invasive ventilation and 28 day mortality, were recorded.

Between 8th February and 8th March 2021, 114 participants were recruited into our study. Table 1 shows participant demographic data. The median age was 64 (IQR 52–78). Most participants were (91%) were symptomatic (fever, cough, breathlessness, anosmia) at time of sampling. 31% of patients had received one dose of either the Pfizer BioNTech or Oxford AstraZeneca vaccine (n = 36, 31%) in the weeks prior to acute infection; 29 had received their vaccine 2 weeks or longer prior to admission. The median duration of symptoms prior to sampling was 10 days (IQR 7 to 15). Almost all patients were antibody positive at time of sampling (n = 95, 93%). 84 (73%) participants received oxygen during hospitalisation; a fifth required continuous positive airway pressure (CPAP) in the days following blood sampling (n = 24, 21%) for progressive respiratory failure. None required mechanical ventilation. 7 (6%) study participants died within 28 days of hospital admission.

Table 1

Participant demographic, clinical, laboratory and radiological data and clinical outcomes.

Variables	Patients	Missing data
Demographic data
Age – median years (IQR)	64 (52 to 78)	0
Male – n (%)	69 (61%)	0
White ethnicity – n (%)	89 (79%)
Asian ethnicity – n (%)	23 (20%)	0
Black ethnicity - n (%)	2 (1%)
Autoimmune disease – n (%)	20 (18%)	0
Hypertension – n (%)	42 (37%)
Diabetes – n (%)	30 (26%)
Ischaemic heart disease – n (%)	35 (31%)
Chronic kidney disease – n (%)	9 (8%)
Cancer – n (%)	5 (4%)
Chronic lung disease – n (%)	23 (20%)
Neurological disease – n (%)	13 (11%)
Gastroenterological/liver disease – n (%)	10 (9%)
Haematological- n (%)	5 (4%)
Number of comorbidities – median (IQR)	1 (1-2)
Clinical data
Admission oxygen saturations – median % (IQR)	96% (94 to 97)	0
Any oxygen in hospital – n (%)	84 (73%)
White cell count – median x109 cells/L (IQR)	8.0 (6.2-11.0)	13 (not performed on sampling)
Lymphocyte - median x109 cells/L (IQR)	1.27 (0.87-1.68)	13
Urea – median mmol/L (IQR)	7.1 (5.1-9.6)	13
Creatinine – median μmol/L (IQR)	68 (57-89)	13
CRP – median mg/L (IQR)	21 (5-60)	13
Haemoglobin – median g/L (IQR)	129 (113-141)	13
IL-6 – median pg/ml	23 (10-68)	86 (never performed)
Nosocomial acquired infection – n(%)	10 (8%)	0
Findings of COVID-19 pneumonia on CXR – n (%)	88 (79%)	2 (CXR never performed)
Duration of symptoms – median days (IQR)	10 (7 to 15)	0
Treatment with dexamethasone – n (%)	73 (64%)	0
Vaccinated - n (%)	36 (31%)	0
Pfizer – n (%)	24 (21%)
Astra Zeneca – n (%)	12 (11%)
Combined IgG/IgM positive– n (%)	95 (93%)	12 inconclusive
T cell responses – median spots (IQR)
Panel 1/S1 protein	5 (2 to 54)	27 inconclusive
Panel 2/S2 protein	5 (2 to 22)
Panel 3/Nucleocapsid protein	3 (0 to 7)
Panel 4/Membrane protein	3 (1 to 10)
Clinical outcomes
Received CPAP following sampling – n (%)	24 (21%)	0
28 day mortality – n (%)	7 (6%)	0

Of 87 participants who had a valid T-SPOT assay reading, the responses to the spike protein antigens S1 and S2 were most sensitive, being positive in the highest proportion of participants and at the greatest amplitude. The median T-SPOT for S1 was 5 spots (IQR 2 to 54); S2: 5 spots (IQR 2 to 22); Nucleocapsid: 3 spots (IQR 0 to 7); Membrane: 3 spots (IQR 1 to 10). Strong correlation was observed between the response to S1 and responses to the other three antigens (Pearson's correlation coefficient to S2: 0.54, p < 0.001; Nucleocapsid: 0.32, p = 0.003; Membrane: 0.51, p < 0.001.) However, there was little concordance between T-SPOT responses and the antibody assay (Pearson's correlation coefficient between S1 and Antibody: 0.06, p = 0.27). We observed no association of T-cell responses with either prior vaccination status or interval after symptoms onset. The T-SPOT assay was also positive in the 12 asymptomatic patients and could be detected within 3 days of a positive PCR test. Finally, we noted that patients with higher T-cell responses to S1 protein were more likely to receive CPAP prospectively during hospitalization, following sampling (Fig. 1 ).

Fig. 1

Box plots illustrating T-SPOT values, stratified by whether they prospectively.

Our study is the first to evaluate T-cell responses using the T-SPOT assay in hospitalized patients with acute COVID-19. We found that T-cell responses appeared as early as two days after symptom onset in early COVID-19, and can be positive in the context of a negative combined antibody assay. T-cell responses did not differ according to vaccination status and appeared to be related to a more severe disease phenotype. Previous studies have assessed the utility of the T-SPOT assay in convalescent patients and found that T-SPOT responses to the same proteins in this study were seen in the absence of anti-Spike IgG on long-term follow-up. We found similar levels of discordance in patients with acute COVID-19, highlighting the potential of the T-SPOT assay to pick up immunological responses in COVID-19 positive patients where antibody responses are negative. T-SPOT responses were also much higher in studies in convalescent individuals, highlighting clear differences in the kinetics of the T-cell response compared with antibody response over time after infection and vaccination. ,

We provide real-world data on the T-cell response of patients who had received one dose of the Pfizer BioNTech or Oxford AstraZeneca vaccine prior to developing COVID-19 requiring hospitalization. We did not observe amplification of the T-cell response in this group, compared with the unvaccinated population. This could reflect the design of current vaccines, which focus on generating a neutralizing antibody response rather than T-cell response, measurement of T-cell responses in a primarily older cohort where there is immunosenescence or may reflect insufficient time for the vaccines to induce robust T-cell immunity, despite the fact that the majority of our patients had their first dose 2 weeks or longer prior to admission.

Finally, we found an association between higher T-SPOT responses (especially S1) and increasing disease severity at the time of sampling, as evidenced prospective need for CPAP. It is not clear whether these T-cell responses are protective or deleterious. A protective role for the higher T-cell responses in severe disease is supported by the low 28-day mortality rate in this cohort. However, it is possible that the elevated T-cell responses are a marker of immune hyperstimulation generating cytokine over-production and cell death. Studies comparing T cell phenotype and cytokine levels are needed to resolve this question.