BCG revaccination qualitatively and quantitatively enhances SARS-CoV-2 spike-specific neutralizing antibody and T cell responses induced by the COVISHIELD™ vaccine in SARS-CoV-2 seronegative young Indian adults

This study tested if prior BCG revaccination can further boost immune responses subsequently induced by a widely distributed and otherwise efficacious Oxford/AstraZeneca ChAdOx1nCoV-19 vaccine, referred to as COVISHIELD™, in India. We compared COVISHIELD™ induced longitudinal immune responses in 21 BCG re-vaccinees (BCG-RV) and 13 BCG-non-revaccinees (BCG-NRV), all of whom were BCG vaccinated at birth and latent tuberculosis negative, after COVISHIELD™ prime and boost with baseline samples that were collected pre-pandemic and pre-BCG revaccination. Compared to BCG-NRV, BCG-RV displayed significantly higher magnitude of spike-specific Ab and T cell responses, including a greater proportion of high responders; better quality polyfunctional CD4 and CD8 T cells that persisted and a more robust Ab and T cell response to the Delta mutant of SARS-CoV-2 highlighting greater breadth. Mechanistically, BCG adjuvant effects on COVISHIELD™ induced adaptive responses was associated with more robust innate responses to pathogen-associated-molecular-patterns through TNF-α and IL-1β secretion. This study provides first in-depth analysis of immune responses induced by COVISHIELD™ in India and highlights the potential of using a cheap and globally available vaccine, BCG, as an adjuvant to enhance heterologous adaptive immune responses induced by COVIDSHIELD™ and other emerging vaccines.


INTRODUCTION
Bacille Calmette-Guérin (BCG) vaccine is part of vaccination policy at birth for tuberculosis (TB) prevention in several TB endemic countries 1 . Beyond TB, heterologous benefits of BCG include reducing all-cause mortality and morbidity in infants and children 2,3 , with these non-specific effects providing cross-protection against other pathogens, particularly against viral respiratory infections [4][5][6][7] . The heterologous benefits of BCG vaccination are likely mediated via the combined induction of Th1/Th17, humoral and innate immune responses. Increasing evidence emphasize that BCG vaccination enhances innate immune memory or 'trained immunity' (TI) 8,9 , a biological process by which innate immune response to pathogen-associated molecular patterns (PAMP) is significantly amplified by prior BCG exposure 10 . BCG induced trained immunity is mediated through epigenetic reprogramming and cellular metabolism rewiring 11,12 , an imprint that can be retained, such that subsequent PAMP exposure induces more pronounced innate effector function in monocytes 13,14 , This, in-turn contributes to an immune response that can be more protective than the one generated in the absence of prior BCG exposure 10,[15][16][17] .
The COVID-19 outbreak has refocused interest in the cross-protective benefits of BCG. Despite effective vaccines 18 , COVID-19 remains a global concern due to emergence of newer variants and global vaccine inequity. Consequently, more widely available vaccines such as the Influenza, OPV, MMR, Varicella Zoster and BCG vaccines known to boost heterologous cross-protective immunity are currently being tested for their efficacy against SARS-CoV-2 in various clinical trials [19][20][21][22][23] , with the most compelling evidence for BCG emerging from murine studies 24,25 . The intravenous administration of BCG in human-ACE2 transgenic mice followed by lethal SARS-CoV-2 challenge reduced viral load, resulted in milder disease and increased survival 24 . However, a similar study in rhesus macaques found no evidence of aerosol BCG vaccination protecting against SARS-CoV-2 challenge 26 . Whilst comprehensive data of BCG impact on COVID-19 infection and /or disease severity is awaited from global human trials 27,28 , several lines of evidence highlight potential benefit. Mathematical modelling highlighted administration of vaccines with heterologous benefits including BCG even with 5-15% efficacy could reduce COVID-19 cases, hospitalization, and mortality in the United States 29 . Epidemiological studies show countries with mandatory BCG vaccination recorded a lower incidence of SARS-CoV-2 infection and COVID-19 related deaths compared to countries where routine immunization was absent or terminated [30][31][32][33][34][35][36] . Individuals vaccinated with BCG in the past five years reported lower incidence of illness and fatigue compared to BCG non-vaccinated subjects 37 and anti-SARS-CoV-2 IgG seroprevalence and selfreported COVID-19 symptoms were significantly less in healthcare workers with BCG vaccination history 38 . In contrast, other studies do not report similar benefits 39,40 perhaps due to the timing of BCG vaccination prior to SARS-CoV-2 exposure or the BCG strain used 41-43 .
Beyond improving immunity to natural infections, BCG can enhance adaptive responses to heterologous vaccines in adults and infants [44][45][46][47][48] . Antibody levels against Influenza A (H1N1) were boosted and seroconversion accelerated in individuals who received BCG prior to a trivalent influenza vaccine 44 . Humoral responses to vaccination against pneumococcus, tetanus toxoid, measles, mumps, diphtheria, polio were higher in BCG immunized compared to non-BCG immunized infants in several trials [46][47][48] . Importantly, data from a randomized study in Mexico showed neutralizing antibody (NAb) titers induced by the Pfizer-BioNTech SARS-CoV-2 vaccine to be higher in the group that received BCG first and then the SARS-CoV-2 vaccine relative to the group that received placebo and SARS-CoV-2 vaccine 49 . In an animal study, human-ACE2 transgenic mice given BCG coupled with a trimeric-Spike vaccine generated a higher titer of NAb and a greater Th1 response than controls receiving trimeric-Spike vaccine alone and cleared infection with minimal immunopathology following SARS-CoV-2 challenge 25 . However, prior exposure to BCG does not guarantee an enhanced adaptive response to all vaccines: Responses to the Vi polysaccharide typhoid fever vaccine (TFV) and the Haemophilus influenzae type B (anti-Hib) vaccine was not boosted upon prior BCG exposure 45, 47,48 . Individual vaccines, whether they are live attenuated, whole inactivated or subunit vaccines, can impact the immune system in various ways, and this may or may not necessarily synergize with the non-specific effects of BCG.
These considerations prompted us to explore the impact of prior BCG vaccination on immune responses induced by the Oxford-AstraZeneca ChAdOx1-S, the first SARS-CoV-2 vaccine to be rolled out in India, locally referred to as COVISHIELD TM manufactured by the Serum Institute of 5 India. Our study was exceptionally well positioned to address the above objective: we had initiated a BCG revaccination study of young healthcare workers in October 2019 residing at St. John's Medical College Hospital, Bangalore with baseline samples (prior to BCG revaccination) and samples at 1 day and 8-10 weeks post BCG revaccination collected before the first COVID-19 outbreak in India. We were therefore in a unique position to track these subjects' longitudinal responses to the COVISHIELD TM vaccine rolled out in January 2021 relative to their pre-pandemic baseline samples. Our study is the first to provide novel insights into the magnitude and quality of antibody and T cell responses elicited by the COVISHIELD TM vaccine in young Indian adults, who were seronegative against all SARS-CoV-2 proteins screened. This entitles unequivocal analysis of the primary immune response induced by COVISHIELD TM prime vaccination to be placed in context of the COVISHIELD TM booster dose in subjects who did and did not receive prior BCG revaccination. (Figure 1). Funded in 2019, this study, was initially designed to probe the efficacy of BCG revaccination on enhancing Mycobacterium tuberculosis (Mtb)-specific immune responses, following our previous work 50  Samples were collected at baseline (T0), 1 day (T1) and 8-10 weeks -post BCG revaccination (T2); these time points were between September 2019 and January 2020 before COVID-19 pandemic spread to India. Subsequent sample collection (T3) between March 2020 and August 2020 was abandoned. Following roll out of the COVISHIELD TM vaccine in January 2021, we collected samples post COVISHIELD TM vaccination from both BCG re-vaccinees (BCG-RV) and BCG nonvaccinees (BCG-NRV). The median interval between BCG revaccination and the first dose of COVISHIELD TM vaccine was 63 weeks. We collected samples at 2-, 3-or 4-weeks after COVISHIELD TM prime (collectively referred to as T4 prime time point), followed by 5-6 weeks after the booster vaccine (T5 time point), or at 20-23 weeks post-booster (T6 time point). Due to COVID-19 restrictions, subjects were sampled only at one time point post prime and one time point post boost. For COVISHIELD TM vaccine-induced immune data analysis, the following samples were excluded: (i) samples not matched to the above time points, (ii) all subjects who tested SARS-CoV-2 RT-PCR positive, (iii) subjects identified to be potentially exposed to SARS-CoV-2 based on a positive in-house IgG antibody binding assay to the SARS-CoV-2 nucleocapsid (N) protein (see Extended Data Fig. 1), and (iv) all dropout subjects where we were unable to collect samples at one of the prime and one of the boost time points. In total, we successfully collected time matched samples from 34 subjects, of which 21 were BCG Re-vaccinees (referred to as BCG-RV) and 13

Study Overview
were not re-vaccinated with BCG (referred to as BCG-NRV) (Fig. 1). of Ab and T cell responses was observed at 3 and 4 weeks post priming (T4:3 and T4:4) and remained significantly higher than baseline after the booster (T5 and T6). Ab binding and NAb peaked at 3 weeks, dipping at 4 weeks; lowest Ab responses were detected at 2 weeks postprime. CD4+ T cell response also peaked at 3 weeks and sustained up to 4 weeks post prime, whilst peak CD8+ T cell responses were noted at 4 weeks, with some individuals responding exceptionally well as early as 2 weeks post prime. Analysis of the booster vaccine effect revealed a clear increase in both Ab binding and NAb concentrations at 5-6 weeks post booster relative to the 4 week post prime levels; however, a similar increase was not noted for CD4+ or CD8+ T cell frequencies in majority of donors tested. We report >87% vaccine potency; thus, subjects who did not respond to both prime and boost was 0% for Ab binding ( Fig. 2A, 2D, 4/31) as far as we could measure. Of the responders, the percentage who required a booster to respond was variable: this was 20.6% for Ab binding ( Fig. 2A, 7/34) using 33.8BAU/ml cut-off, which may be high as only 3 of the same subjects were detected as needed a boost by NAB assay Taken together, these data highlight the COVISHIELD TM vaccine to be highly efficient in inducing immune responses in young Indian adults, with priming alone inducing significant Ab and T cell responses. The booster vaccine effect is particularly noticeable in Ab responses, consistent with other studies [51][52][53][54][55] . In contrast, the highest responses noted for CD4+ and CD8+ T cells was at 3-4 weeks post prime, also consistent with other reports [51][52][53][54][55] . Both Ab and T cell responses waned by 20-23 weeks; importantly, in vast majority of subjects the induced responses did not fall to baseline. vaccine. Instead, a contraction of spike-specific CD4+ and CD8+ T-cell frequencies was noted at both the T5 and T6 time points relative to peak values, with some noticeable differences: median CD4+ and CD8+ T cell frequencies were higher in BCG-RV versus BCG-NRV, with a trend to remain higher in BCG-RV at 20-23 weeks post boost ( Fig. 3C and 3D). Furthermore, BCG-NRV CD4+ and CD8+ T cell responses contracted to near baseline levels at T6, whereas in BCG-RV, the median remained 10-fold higher than the baseline. (iii) The proportion of COVISHIELD TM nonresponders (subjects who did not respond to either prime and boost) did not differ between BCG-RV and BCG-NRV.

Overview of COVISHIELD
Comparative analysis of COVISHIELD TM induced responses in BCG-RV and BCG-NRV was conducted by stratifying responders in time matched samples ( Fig. 3 scatter graphs) using a contingency Fischer's Exact Test comparing fold induction of responses, relative to T0. Fold changes were stratified into high and low, dependent on median fold change and spread recorded in each assay: for Ab binding (Fig. 3A) high responses (HR) were set at minimum 100-fold increase over baseline and low responses (LR) set at minimum 10-fold increase over baseline. For NAb (Fig. 3B), HR was 10-fold minimum and LR 3-fold minimum. For CD4+ (Fig. 3C) and CD8+ ( Fig. 3D) T cell responses, HR was 10-fold minimum and LR 4-fold minimum. In each immune assay, BCG-RV had a significantly higher proportion of HR and/or LR responders than BCG-NRV (between group comparison minimum p value p=0.0373; maximum p<0.0001). For NAb (Fig. 3B, scatter graph), the biggest difference was noted after the booster at 20-23 weeks. For both CD4+ ( Fig. 3C) and CD8+ T cell responses (Fig. 3D), BCG-RV had a greater proportion of high responders after COVISHIELD TM priming (T4) and the proportion of high responders persisted up to 20-23 weeks after the booster (T6). Taken together, the data highlights BCG-RV to be significantly better and more often high responders to the COVISHIELD TM vaccine than BCG-NRV subjects.
Additionally, we confirm the COVISHIELD TM vaccine induced T cell response to be specific. induced spike-specific CD4+ T cell subsets expressing 3, 2 and 1 effectors after prime, which were sustained until 20-23 weeks after the booster (Fig. 4A). In BCG-NRV, TNF-α+ and IL-2+ single positive (SP) expressing CD4+ T cells were not induced significantly after prime (Fig 4B), whilst cellular subsets expressing 3, some 2 and some 1 combinations of effectors were. Additionally, in BCG-NRV, only two of the seven subsets analysed (namely: IL-2/TNF-α double positive {DP} and IFN-γ/IL-2 DP) were sustained after the booster: in particular, cells expressing all 3 cytokines and single effectors were not sustained (Fig. 4B, Table). Unpaired t-test with Welch correction of matched BCG-RV and BCG-NRV samples after the prime-(Extended Data Fig. 6A) and boostervaccines (Extended Data Fig. 6B) showed significantly higher responses in BCG-RV for IFN-γ/IL-2 DP cells and IFN-γ SP after the prime dose and significantly higher IFN-γ/TNF-α DP, TNF-α SP and IFN-γ SP after the booster dose. There was also a trend (p=0.06-0.08) for most other subsets to be expressed higher in BCG-RV after the booster. A similar analysis of CD8+ T cell subsets ( Fig. 4C and 4D) highlighted striking differences: whereas five of seven subsets analysed were induced significantly in BCG-RV group at prime and six of seven subsets including 3+ effectors sustained after the booster dose (Fig. 4C, Table), in BCG-NRV five of the seven CD8+ subsets tested and four of seven were not significantly induced after the prime or booster doses, respectively (Fig. 4D). Unpaired t-test with Welch correction of matched BCG-RV and BCG-NRV samples shows no significant differences between the two groups after prime (Extended Data Fig.   6C) or booster (Extended Data Fig. 6D), although, four subsets, namely 3+ IFN-γ+TNF-α+IL-2+ effectors, IFN-γ+/TNF-α+ DP, TNF-α+ SP and IFN-γ+ SP effectors showed a trend (P=0.08) for higher expression in BCG-RV group.
Taken together, these data highlight that BCG revaccination significantly enhances the quality of the COVISHIELD TM induced T cell response, with the most pronounced effect being on spikespecific CD4+ rather than CD8+ T cell effectors. To understand if this difference is in part explained by inherent differences in the robustness by which BCG revaccination regulates adaptive CD4+ versus CD8+ T cells, we enumerated BCG-specific CD4+ and CD8+ T cell frequencies following BCG revaccination at 2 time points (Extended Data Fig. 7): first, at 8-10 weeks post BCG revaccination (T2), where a significant increase in the frequencies of BCG-specific IFN-γ or IL-2 CD4+ but not CD8+ T cells were noted in the BCG-RV but not BCG-NRV group, relative to paired baseline samples (T0) (Extended Data Fig. 7A and 7B). M. tuberculosis specific recombinant ESAT-6/CFP10 fusion protein served as a negative control 56 ; as BCG lacks this region. No change between time points in all samples tested was noted to ESAT-6/CFP10 stimulation (Extended Data Fig. 7B). Second, we demonstrate that higher BCG-specific CD4+ T cell responses noted in BCG-RV at 8-10 weeks post BCG vaccination persist until the T6 time point of the study, which is 78-94 weeks post BCG revaccination (Extended Data Fig. 7C and see Figure   1), with no change in MTb-specific CD8 T cell frequencies between BCG-RV and BCG-NRV, highlighting BCG vaccination to more robustly induce CD4+ rather than CD8+ T cell effectors at the time points studied.

BCG revaccination enhances the breadth of the COVISHIELD TM induced immune response.
We next determined the efficiency with which COVISHIELD TM vaccine induced NAb and T cells show that while both BCG-RV and BCG-NRV induce significant NAb response after prime, there is a sharp decline of Delta-specific NAb to near baseline across several donors at 20-23 weeks postbooster, implying that Delta-specific Abs were not as well sustained as NAb to the wild-type (WT) SARS-CoV-2 Wuhan strain (see Fig 2). The differences between groups comparing fold induction of responses over matched baseline values ( Fig. 5A scatter graphs) were striking: the BCG-RV group had a significantly higher proportion of high and low responders after prime, but these differences were not sustained after the booster with few high responders detected at 20-23 weeks post booster ( Fig. 5A scatter graphs). This was also confirmed using a Wilcoxon paired t-test analysis of the NAb response to WT versus Delta strains in each subject (Fig. 5A), which showed that after prime the BCG-RV group had equally efficient NAb to both strains, whereas BCG-NRV had significantly lower NAb to Delta. After the booster, there was a trend for subjects in both groups to have higher NAb to WT, but this difference did not reach significance. These data highlight priming alone induces a more efficient NAb to both WT and Delta in BCG-RV compared to BCG-NRV.
Matched analysis of the breadth of CD4+ and CD8+ T cell responses is shown in Fig. 5B and 5C respectively to peptides spanning the Delta mutation relative to matched epitopes in the Wuhan strain in a subset of 6 subjects. Extended Data Fig. 8B and 8C (line graphs) show that frequencies of IFN-γ or IL-2 expressing CD4+ and CD8+ T cells are enhanced in response to delta strain in both BCG-RV and BCG-NRV after prime, although the extent of induced response was considerably lower in BCG-NRV and waned substantially by 20-23 weeks post-booster in both groups. Fig. 5B and 5C show BCG-RV comprised significantly higher proportion of both CD4 and CD8 high and low responders respectively, highlighting BCG-NRV to be overall weaker T-cell responders to Delta SARS-CoV-2. This was also confirmed using Wilcoxon paired t-test analysis, which showed equally efficient induction of CD4 and CD8 T cell responses to both strains within a given donor at prime and a decline of Delta responses at 20-23 weeks post boost ( Fig. 5B and 5C). Figure 5D and 5E confirm that Delta-specific NAb and CD4+ or CD8+ T cell responses correlated significantly.
Collectively, above data highlight that BCG-RV mount more robust spike-specific NAb and CD4+ and CD8+ T cell responses to the parent Wuhan and Delta strains compared to BCG-NRV.

BCG revaccination boosts monocyte and PBMC PAMP-induced effector cytokines that are
associated with trained immunity. One acknowledged mechanism through which BCG potentially boosts immune responses to a heterologous vaccine is by augmenting monocyte and PBMC PAMP-stimulated responses including expression of TNF-α, IL-1β and IL-6 implicated in trained immunity. We therefore tested monocyte and PBMC responses to PAMP stimulation in the same subjects probed for COVISHIELD TM induced vaccine responses. Samples archived prior to the COVID-19 pandemic were tested to mitigate against the potential of either the COVISHIELD TM vaccine or exposure to SARS-CoV-2 to induce PAMP responses. We compared pre-pandemic baseline (T0) with samples harvested 8-10 weeks post BCG revaccination (T2, see The two principal immunological mechanisms by which COVISHIELD TM mediates protection against COVID-19 disease severity is through induction of SARS-CoV-2 spike-specific NAbs and T cells 57 . We demonstrate that BCG does not significantly alter the published kinetics of spikespecific Ab and T cell responses induced by the COVISHIELD TM vaccine in a seronegative population, with Ab binding and NAb levels peaking 3 weeks post prime (as recorded in Fig. 2) and further enhanced at 2 weeks post boost 51-55, 58-60 , a time point that we were unable to collect.
However, we show a clear boost to Ab binding and NAb levels at 5-6 weeks, implying the booster dose effect is sustained up to 5-6 weeks before waning. We show frequencies of spike-specific IFN-γ or IL-2 CD4+ and CD8+ T cells rising significantly over baseline at 2 weeks and continuing to rise over 3 and 4 weeks post prime (Fig. 2). However, unlike Ab responses, our data, as reported [51][52][53][54][55] , show no major booster effect of the COVISHIELD TM vaccine induced T cell responses (Fig. 2).
Whilst BCG did not alter the kinetics of the COVISHIELD TM response, there was significant impact on its quality in three major ways: First, Ab and T cell responses were significantly higher in BCG-RV, including a greater proportion of high responders based on fold induction of the measured immune response over baseline. Some BCG-RV individuals had exceptionally high T cell responses (>10-fold change) that persisted till 20-23 weeks post-boost; such high responses were not detected or declined sharply in BCG-NRV. This heterogeneity may be intrinsic to the COVISHIELD TM vaccine 61 and potentially amplified by BCG. BCG revaccination therefore synergizes with COVISHIELD TM to amplify vaccine-specific Ab and T cell responses as well as enhance the durability of the induced immune response. Second, BCG helped induce a more polyfunctional T cell response, a characteristic that ascertains vaccine efficacy against chronic viral infections 62 , including SARS-CoV-2 where vaccine-induced multifunctional T cells correlate with enhanced protection from emerging variants 63 . Interestingly, one study showed polyfunctional T cells to be enhanced following the ChAdOx1 nCoV-19 vaccine booster indicating the booster may serve to enhance the quality and not just magnitude of a vaccine-induced response 63 . We contend that significant induction of polyfunctional spike-specific T cells and their persistence after the booster can potentially contribute to the heterologous benefit of BCG. Third, BCG-RV produced a more robust response to the Delta mutant of SARS-CoV-2 highlighting greater breadth of immune responses, a function that was globally, including India, associated with milder disease in AstraZeneca vaccinees during second wave of the SARS-CoV-2 pandemic [64][65][66] . With a strong correlation noted between NAb and T cells specific for both the Wuhan and Delta strains, we contend that BCG vaccination has the potential to expand the breadth of the Ab and T cell response against SARS-CoV-2 variants of concern.
Our data is consistent with previous work highlighting the benefit of prior or synchronous BCG vaccination in boosting heterologous vaccine responses 67 , including to a trivalent influenza vaccine 44 and to vaccination against pneumococcus, tetanus toxoid, measles, mumps, diphtheria and polio in BCG immunized infants 46-48 . Our data is also consistent with the results of the Mexican study which demonstrated prior BCG vaccination to enhance the Pfizer/Biotech induced NAb response,4 weeks after BCG vaccination 49 . In the context of COVID-19, BCG may not be unique and is consistent with emerging acceptance of the benefits of heterologous vaccination strategies.
It's been noted that immunization with existing vaccines such as the Influenza, OPV, MMR, Varicella Zoster in the recent past (≤ 5 years) can confer protection against SARS-CoV-2 by reducing infection rates, improving clinical outcomes and /or boosting NAbs induced during infection 68 . Indeed, heterologous prime-boost immunization regimens per se maybe more beneficial, i.e., ChAdOx1 nCoV-19 and mRNA-1273 or ChAdOx1 nCoV- 19 and BioNTech have been shown to augment COVID-19 vaccine efficacy by enhancing spike-specific IgG, NAbs as well as CD4+ and CD8+ T cells including robust recognition of variants of concern above levels induced by homologous vaccination [69][70][71][72][73] .
One important mechanism by which BCG vaccination can boost heterologous vaccine responses is its intrinsic PAMP characteristics and ability to regulate innate immunity. Individuals in our study who showed boosted adaptive responses to COVISHIELD TM also exhibited evidence of trained Our observation that the heterologous benefit of BCG was more evident on COVISHIELD TM induced spike-specific CD4+ rather than CD8+ T cells is consistent with BCG as a recognised inducer of Th1 CD4+ T cell effectors through three potential mechanisms: firstly, trained monocytes have higher expression of MHC-II and costimulatory molecules CD80 and CD86; thus making them better antigen presenting cells for CD4+ T cell activation 75,76 ; secondly, trained monocytes have a higher expression of PRRs like CD14, TLR4 and mannose receptor and produce more pro-inflammatory cytokines such as TNF-α and IL-1β which can enhance T cell responses 8 and thirdly, cytokines secreted by trained monocytes e.g., IL-1β and IL-6 are key drivers of Th differentiation to Th1, Th17 or ex-Th17-sub-sets that have been shown to be correlates of protection against viral and bacterial infections [77][78][79] . Apart from these suggested mechanisms, it has been speculated that BCG vaccination might lower thresholds for T cell activation on account of the cytokine milieu that exists due to primed/trained monocytes 80  John's Medical College-Hospital, Bangalore, India were invited to participate in this prospective observational study from October 2019 to June 2021. All potential participants underwent a screening criteria and subjects with chronic illness such as hypertension, diabetes mellitus, heart disease, cancer, kidney / thyroid illness, asthma, epilepsy, jaundice or with a history of clinical tuberculosis disease and on medication were excluded from the study. All included subjects were assigned a unique serial number and baseline information such as age, gender, medical history, occupation, vaccination status and family history pertaining to tuberculosis was obtained. All participants were BCG-vaccinated at birth. Basic anthropometry measurements such as height (cm), weight (kg) using standard validated and calibrated instruments were used, and body mass index (kg/m 2 ) was computed. Relevant clinical information of study participants was documented in a proforma and is summarized in Table 1 and detailed follow-up questionnaire is provided in Supplemental File 1. Blood from study participants was screened for Mtb infection by the standard QFT TB Gold In-tube test (Qiagen) performed at Department of Microbiology, SJMCH, India, and were classified as either IGRA+ or IGRA-of which 66 IGRA-subjects were enrolled for the study (Fig. 1A).

ICS assay for whole blood stimulation assay and multiparameter flow cytometry.
Heparinized whole blood was collected from participants and processed within 30-45 min of phlebotomy, as previously described 85 . Briefly, 400 µl of blood was pipetted into 5ml polypropylene tubes (Sarstedt, Germany) and stimulated with Ag85A peptide pools (1 µg/ml per peptide), BCG (0.2 x 10 6 CFU/ml), or purified recombinant protein ESAT-6/CFP10 (10µg/ml) together with anti-CD28/CD49d costimulatory mAbs at 0.5 µg/ml. Culture medium with anti-CD28/CD49d was used as unstimulated negative control. Blood was incubated at 37˚C for a total of 12 hr, and Brefeldin A 20 + Monensin (Biolegend) diluted to a final concentration of 1X from a 1000X stock was added in the final 10 hr of stimulation. After stimulation, blood was treated with 2 µM EDTA, RBCs were lysed with 4.5 ml 1X FACS Lysing solution (BD), and fixed white blood cells were transferred to liquid nitrogen in freezing medium containing 10% DMSO, 40% FCS, and 50% RPMI 1640. On the day of staining, cryopreserved whole-blood samples were thawed in a water bath at 37°C for 2 min. Please refer Table 2 for details of antibodies used for staining.

ICS assay with SARS-CoV-2 specific peptides. Flow cytometry was used to examine SARS-
CoV-2-specific CD4+ and CD8+ T cell responses using a validated ICS assay 50  Absolute concentration of cytokine was calculated using a standard curve. The linear range of detection for TNF-α was 40 -2500 pg/ml, for IL-1β was 10 -1250 pg/ml and for IL-6 it was 40 -5000 pg/ml. For each ELISA, assay background was subtracted from absorbance values. Also, the spontaneous cytokine release in cells cultured with medium alone was subtracted from all PAMP stimulation conditions.

Recombinant full-length Spike, RBD and Nucleocapsid proteins. Recombinant Spike and RBD
for ELISA were expressed and purified as previously described 87  In-house ELISA for anti-Spike, RBD or Nucleocapsid IgG. Plasma was isolated from heparinized whole blood by centrifugation at 400 g for 10 minutes at room temperature. The supernatant was further centrifuged at 800-1200 g for 10 minutes to obtain clear plasma and stored at -80°C until further use. Plasma was heat-inactivated at 56°C for 30 minutes prior to use. SARS-CoV-2 Spike, RBD or Nucleocapsid specific IgG was measured using an in-house ELISA as previously described 88 . Briefly, Spike/RBD purified as previously described 87

Conflict of Interest Statement
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.