Pre-pandemic SARS-CoV-2-specific IFN-γ and antibody responses were low in Ugandan samples and significantly reduced in HIV-positive specimens

Abstract

Introduction: We investigated whether prior SARS-CoV-2-specific IFN-γ and antibody responses in Ugandan COVID-19 pre-pandemic specimens aligned to this population's low disease severity.

Methods: We used nucleoprotein (N), spike (S), NTD, RBD, envelope, membrane, SD1/2-directed IFN-γ ELISpots, and an S- and N-IgG antibody ELISA to screen for SARS-CoV-2-specific cross-reactivity.

Results: HCoV-OC43-, HCoV-229E-, and SARS-CoV-2-specific IFN-γ occurred in 23, 15, and 17 of 104 specimens, respectively. Cross-reactive IgG was more common against the nucleoprotein (7/110, 15.5%; p = 0.0016, Fishers' Exact) than the spike (3/110, 2.72%). Specimens lacking anti-HuCoV antibodies had higher rates of pre-epidemic SARS-CoV-2-specific IFN-γ cross-reactivity (p-value = 0.00001, Fishers' exact test), suggesting that exposure to additional factors not examined here might play a role. SARS-CoV-2-specific cross-reactive antibodies were significantly less common in HIV-positive specimens (p=0.017; Fishers' Exact test). Correlations between SARS-CoV-2- and HuCoV-specific IFN-γ responses were consistently weak in both HIV negative and positive specimens.

Discussion: These findings support the existence of pre-epidemic SARS-CoV-2-specific cellular and humoral cross-reactivity in this population. The data do not establish that these virus-specific IFN-γ and antibody responses are entirely specific to SARS-CoV-2. Inability of the antibodies to neutralise SARS-CoV-2 implies that prior exposure did not result in immunity. Correlations between SARS-CoV-2 and HuCoV-specific responses were consistently weak, suggesting that additional variables likely contributed to the pre-epidemic cross-reactivity patterns. The data suggests that surveillance efforts based on the nucleoprotein might overestimate the exposure to SARS-CoV-2 compared to inclusion of additional targets, like the spike protein. This study, while limited in scope, suggests that HIV-positive people are less likely than HIV-negative people to produce protective antibodies against SARS-CoV-2.

Keywords: Common-cold coronaviruses; Cross protection; Disease severity; Pre-existing IFN-γ; SARS-CoV-2; Sub-Saharan Africa; Ugandan population; sero-crossreactivity.

Figure 1

Pre-epidemic IFN-γ responses to SAR-CoV-2, HCoV-229E and HCoV-OC43. Illustrates the virus-specific IFN-γ responses to 18 SARS-CoV-2 and two human coronavirus peptide pools (HuCoVOC43 and HuCoV229E). IFN-γ response magnitudes are depicted and quantified as spot-forming units per million PBMC cells. The horizontal dotted red line represents the cut-off threshold for a positive response of 55 SFU/million PBMCs (A). The data represents 104 pre-COVID specimens classified as positive in red and negative in blue based on HIV serostatus (B).

Figure 2

The SARS-CoV-2 and HuCoV peptide pools targeted by HIV-1 status. Summarizes the individual responses to each peptide pool for 39 of the participants who responded positively to at least one peptide pool. The data is stratified by HIV serostatus, with HIV negatives in blue (A) and HIV positives in red (B). The larger the size of the circle, the greater the response.

Figure 3

Pairwise correlation between SARS-CoV-2 and HuCoV-directed IFN-γ. Depicts correlation plots of all examined SARS-CoV-2 and human coronavirus peptide pools for 68 HIV-negative (A) and 36 HIV-positive (B) specimens using Spearman's rank correlation test. Blue dots represent a positive correlation, while red dots represent a negative correlation. The darker the circle, the stronger the correlation, and the lighter the circle, the weaker the correlation. The correlation matrices show correlation coefficients between peptides in HIV-positive (C) and HIV-negative (D) specimens using Spearman's rank correlation. P-values of 0.05 or less are considered significant. Non-significant correlations are indicated by blank cells.

Figure 4

Frequencies and magnitudes of SARS-CoV-2 -specific cross-reactive IgG binding antibodies by HIV status. The frequency of pre-epidemic cross-reactive anti-SARS-CoV-2 binding IgG antibodies in Ugandan pre-pandemic sera is depicted in Figure 4 . 110 specimens (A) were stratified by HIV-1 negative (B) and positive serostatus (C), and the proportions of subjects with cross-reactive IgG binding antibodies against the SARS-CoV-2 spike and nucleoprotein illustrated. (D, E) summarises the medians of means of duplicate IgG antibody OD values in nm and concentration levels in ng/ml, respectively. The cut-off OD values are shown by the dashed lines.

Figure 5

IFN-γ cross-reactivity categorised HIV status and spike (A) and nucleoprotein (B) IgG response. Depicts the simultaneous induction of SARS-CoV-2-specific IgG antibodies and IFN-γ cross-reactive responses in 104 specimens with antibody and T-cell data against the spike (A) and the nucleoprotein (B). The data is presented according to HIV serostatus, with HIV-positive individuals in red and HIV-negative individuals in blue. Open triangles represent participants with cross-reactive antibodies and IFN-γ responses.