Within this comprehensive household study conducted in a peri-urban setting in The Gambia, we have established the high seroprevalence present early in the SARS-CoV-2 pandemic, and neutralising antibody, mucosal and T-cell responses to SARS-CoV-2 in adults and children. In prepandemic samples, there was a higher seroprevalence of nucleocapsid specific SARS-CoV-2 antibody responses in adults compared to young children. Whilst strong correlation was noted within each immune parameter, the correlations between the different immune parameters were weak suggesting differences in the cellular, systemic and mucosal immune responses elicited against SARS-CoV-2.
Following the first two waves of SARS-CoV-2 and prior to the emergence of the Delta variant, we reported high seroprevalence among unvaccinated and mostly asymptomatic adult individuals living in a peri-urban setting in The Gambia7. These findings are similar to those reported in an earlier study, which was conducted in Farafenni, a partly rural and partly urban settlement in The Gambia around the same time 10, but the previous study was limited to pregnant women and only reported SARS-CoV-2 total receptor binding domain (RBD) immunoglobulin (Ig) M/IgG.
Our study revealed noticeable age-specific differences in seroprevalence rates 11. When adjusted for spike protein only, we found that just 25% of children under 5 years old were seropositive, compared to over 70% of participants in the 18–49-year age group. Higher SARS-CoV-2 seroprevalence has been observed among younger and middle-aged adults in various African settings, albeit with significant variations, 12–15 and in high income settings such as the UK 16,17. The differences in SARS-CoV-2 seroprevalence across Africa, may be influenced by country-specific and regional factors, along with variations in behaviour and pandemic measures, which likely impacted transmission dynamics and may explain the observed regional disparities 18.
Neutralizing serum antibody responses were higher in 5-17-year age group compared to adults aged 18–49 years for the ancestral variant. Increased neutralizing antibody titres in young children compared to adults has been previously described19. The high neutralizing antibodies in these young children may arise from repeated viral infections especially to seasonal coronaviruses which are cross-reactive against SARS-CoV-220. Given the wide circulation of coronaviruses, especially amongst children, we accessed pre-pandemic samples to measure potential cross-reactive antibody to spike and nucleocapsid protein. The nucleocapsid specific responses to human coronaviruses are conserved and highly cross- reactive whilst the spike specific responses are more diverse. Our findings of high nucleocapsid specific responses to SARS-CoV-2, in prepandemic samples has also been reported in other African countries 21 and has been attributed to exposure to other seasonal coronavirus but also with the high burden of malaria 22 and dengue fever 23 in this region.
Mucosal IgA responses have been shown to protect against viral infections. An induction of broad and persistent mucosal antibody response after primary SARS-CoV-2 infection against the spike and receptor binding domain of the virus has already been described 24. In these SARS-CoV-2 infected individuals, spike and receptor binding domain mucosal antibodies increased within 7–9 days after the start of symptoms and correlated with a lower viral load and faster resolution of systemic symptoms. We showed that participants infected with SARS-CoV-2 (S + N+) had higher mucosal responses compared to those that were seronegative for all variants tested. Whilst we did not see any significant differences in mucosal antibody responses with age in our cohort, previous studies have reported an earlier and more robust induction of IgA responses in asymptomatic children compared to symptomatic children and adults 25. The early induction of mucosal IgA in asymptomatic may possibly control viral replication leading to mild infections. A significant correlation in mucosal responses for Ancestral, Alpha, Delta and Omicron variants was noted in our cohort, suggesting that mucosal responses could be more cross-reactive compared to serum antibody responses. Although we have not assessed durability of the mucosal antibody responses, detection of mucosal antibodies up to 9 months post infection has been reported 24. However, results from a human challenge model of SARS-CoV-2, showed that in adults, whilst both mucosal and systemic antibodies are detected within 10 days after infection, mucosal antibodies plateau at day 14 post infection whilst systemic antibodies continue to rise26.
We were able to assess T-cell responses to SARS-CoV-2 antigens in a subset of participants and found that among participants who were seropositive for both spike and nucleocapsid, 67.5% had a T-cell response to at least one antigen whilst 49.6% responded to two or more antigens. Interestingly, we also observed that 50% of (S-N+) and 34.3% of (S-N-) individuals had a T-cell response to at least one SARS-CoV-2 antigen. SARS-CoV-2 specific T cell responses in spike seronegative individuals has also been reported in Kenya with 70% of these individuals responding to multiple SARS-CoV-2 specific peptide pool as opposed to responses in single peptide pools when T cell responses were assessed in prepandemic samples 27. The difference in magnitude and breadth of SARS-CoV-2 specific T cells between prepandemic and exposed seronegative healthcare workers in the UK has also been reported 28. Whilst the observed T-cell responses in seronegative participants in our study cohort may warrant further investigations, the observation that these responses were not limited to a single peptide similar to the Kenyan study 27 suggests that they are genuine responses.
Cross-reactivity to other human coronaviruses, malaria 22, and viral infections such as dengue fever 23 and CMV 29 has been implicated in the detection of SARS-CoV-2 specific T cells in both spike and nucleocapsid seronegative individuals. Higher frequencies of SARS-CoV-2-specific T cells were found in CMV seropositive donors compared to CMV seronegative individuals using PBMCs collected pre-pandemic. Further analysis of these cross-reactive T cells showed that HLA-B:35:01 which is the most common HLA-B epitope in The Gambian population30 presents both the SARS-CoV-2 spike peptide FVSNGTHWF (FVS) and the unrelated CMV pp65 peptide IPSINVHHY (IPS)29. CMV seropositivity is very high in The Gambia with over 85% of individuals being positive by the first year of life 31. It will therefore be difficult to ascertain the exact contribution of CMV infection to the SARS-CoV-2 cross-reactive T cells in seronegative donors in our setting.
Abortive infections may be another explanation for the presence of SARS-CoV-2 T cells in seronegative participants. In seronegative participants, Swadling et al showed that SARS-CoV-2 T cells that respond to replication–transcription complex (RTC) epitopes could also recognize human coronaviruses and these were associated with abortive infections in SARS-CoV-2 exposed seronegative donors 28. This suggest that SARS-CoV-2 viruses that infect these seronegative individuals were most likely rapidly cleared by the highly conserved RTC specific T cells from prior infection with human coronaviruses. Swadling et al showed presence of T cell responses to both the structural and RTC region of SARS- CoV-2 in seronegative health care workers but a higher frequency of these T cells was directed against the RTC in seronegative healthcare workers compared to unexposed individuals from a prepandemic cohort and those infected with SARS-CoV-2. NSP12 and NSP13 of the RTC region were among the most conserved across SARS-CoV-2 clades and NSP-12 was shown to be the region that elicited the highest average magnitude and frequency of responders without triggering humoral immune responses 28.
The presence of SARS-CoV-2-specific CD4 + and CD8 + T cells is linked to reduced severity of COVID-19 during active infection 32. Emergence of SARS-CoV-2 variants is of great concern as it can lead to immune escape but majority of SARS-CoV-2 specific CD4 + T cells (93%) and of CD8 + T cell (97%) in naturally infected or Pfizer/BioNTech and Moderna COVID-19 mRNA vaccinated individuals are conserved across variants including Alpha, Betta and Gamma strains 33. Slower waning of SARS-CoV-2-specific T cells compared to antibodies had been previously described34. Whilst sustained CD4 + T and CD8 + T cell responses up to 8 months post infection were reported in a study by Dan et al, a more recent study reported a decrease in the magnitude of these SARS-CoV-2 specific T cell 10 months post infection with the ancestral SARS-CoV-2 strain35.
The findings that induction of CD4 + and CD8 + T cells in a cohort of seronegative SARS-CoV-2 exposed individuals was associated with time since exposure 36, suggest that in our cohort, viral infection with SARS-CoV-2 may have occurred at a lower threshold without seroconversion in these seronegative individuals.
The assessment of systemic, mucosal, and T-cell responses in the same individuals is a strength of our study as our data has shown that systemic antibody responses to infection differ to that at the site of pathogen entry, the mucosa. Analysis of samples from both blood and mucosal sites should be incorporated in future studies to increase our understanding of compartmentalised immune responses to infection. Our study also strengthens the evidence that seroprevalence data based on antibody responses alone may underestimate the true burden of SARS-CoV-2 in populations. We reported that whilst 43.1% of the individuals in our study had detectable antibody responses to both spike and nucleocapsid, 67.5% had a T response to at least one antigen within these seropositive individuals, suggesting that T-cell responses may be more suitable in identifying seropositive individuals.
Our study has some clear limitations. We could not assess SARS-CoV-2 specific T-cell responses in the pre-pandemic samples due to sample unavailability, but presence of SARS-CoV-2 spike, membrane, ORF3a, or ORF7 responses were detected in 31% of individuals tested using prepandemic samples 27. We report the existence of pre-existing antibody and T-cell responses to coronaviruses, but the specific strain of coronavirus inducing this response cannot be ascertain within our study. To assess the role of pre-existing responses from other human coronavirus infections, future studies should assess antibody and T cell responses against the four major human coronaviruses species (a-hCCC-NL63, a-hCCC-229E, and b-hCCCHKU1, b-hCCC-OC43) in pandemic samples. The lack of flow cytometry data to assess polyfunctional T-cells which was shown to increase with repeated exposure to SARS-CoV-2 37,38 is also a limitation of our study. Finally, we were limited by the fact that we did not have a rural population with pandemic data to compare with our data. However, this has been addressed by the study done by Janha et al 2023, where data from both rural and urban settings in The Gambia have been reported10.
In summary, our data shows high seroprevalence of SARS-CoV-2 in The Gambia after two waves of the virus, especially in adults. Infection with SARS-CoV-2 induced systemic, mucosal and T-cell responses, with significant correlation observed with earlier variants as opposed to the later variants. Variant-specific responses were highly correlated at the mucosal level but not the serum antibody responses.