Six-month immune responses to mRNA-1273 vaccine in combination antiretroviral therapy treated late presenter people with HIV according to previous SARS-CoV-2 infection

Objective: Immune responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) mRNA vaccines in people with HIV (PWH) with a history of late presentation (LP) and their durability have not been fully characterized. Design: In this prospective, longitudinal study, we sought to assess T-cell and humoral responses to SARS-CoV-2 mRNA vaccination up to 6 months in LP-PWH on effective combination antiretroviral therapy (cART) as compared to HIV-negative healthcare workers (HCWs), and to evaluate whether previous SARS-CoV-2 infection modulates immune responses to vaccine. Methods: SARS-CoV-2 spike (S)-specific T-cell responses were determined by two complementary flow cytometry methodologies, namely activation-induced marker (AIM) assay and intracellular cytokine staining (ICS), whereas humoral responses were measured by ELISA [anti-receptor binding domain (RBD) antibodies) and receptor-binding inhibition assay (spike-ACE2 binding inhibition activity), before vaccination (T0), 1 month (T1) and 5 months (T2) after the second dose. Results: LP-PWH showed at T1 and T2 significant increase of: S-specific memory and circulating T follicular helper (cTfh) CD4+ T cells; polyfunctional Th1-cytokine (IFN-γ, TNF-α, IL-2)- and Th2-cytokine (IL-4)-producing S-specific CD4+ T cells; anti-RBD antibodies and spike-ACE2 binding inhibition activity. Immune responses to vaccine in LP-PWH were not inferior to HCWs overall, yet S-specific CD8+ T cells and spike-ACE2 binding inhibition activity correlated negatively with markers of immune recovery on cART. Interestingly, natural SARS-CoV-2 infection, while able to sustain S-specific antibody response, seems less efficacious in inducing a T-cell memory and in boosting immune responses to vaccine, possibly reflecting an enduring partial immunodeficiency. Conclusions: Altogether, these findings support the need for additional vaccine doses in PWH with a history of advanced immune depression and poor immune recovery on effective cART.

Data on the efficiency of immune response to SARS-CoV-2 in PWH are controversial [14]. T-cell and humoral responses after SARS-CoV-2 infection have been reported not inferior to those in HIV-negative peers in two previous studies [15,16], and yet related to CD4 þ /CD8 þ ratio [15]. Conversely, others found lower and more exhausted SARS-CoV-2-specific T cells [17][18][19], or reduced humoral responses [6] in PWH recovered from COVID-19, especially in those with detectable HIV viremia and low CD4 þ T-cell counts. Furthermore, in a murine model of acute SARS-CoV-2 infection, the depletion of CD4 þ T cells led to diminished antibodies response and delayed viral clearance [20], highlighting the paramount role of CD4 þ T cells in controlling viral replication and regulating SARS-CoV-2-specific immune responses.
The observation that PWH have lower responses to hepatitis B virus vaccination [21,22], show more rapid wane of neutralizing antibodies after yellow fever vaccination [23], and mount variable humoral responses to other vaccines depending on CD4 þ T-cell counts [24], raise the concern that PWH, especially those with incomplete immune reconstitution despite virally effective combination antiretroviral therapy (cART), may not adequately respond to SARS-CoV-2 vaccines.
Despite such concerns, several studies have recently shown that immune responses to mRNA vaccines in PWH with optimal immunological response are comparable to those in healthy individuals but may be reduced in PWH with poor CD4 þ T-cell recovery [25][26][27][28][29][30][31]. Likewise, vaccine effectiveness, namely the ability of vaccine to protect against symptomatic illness and hospitalization/death, has been proven among a large population-based cohort of PWH with well controlled HIV infection after two doses of SARS-CoV-2 vaccine [32], yet it remains to be determined in PWH who have moderate to severe immunodeficiency.
Additionally, while COVID-19-recovered individuals have been proven to mount stronger immune responses to SARS-CoV-2 vaccines compared to virus-naive ones in the general population [57][58][59], whether previous SARS-CoV-2 infection may modulate T-cell and humoral responses to such vaccines in PWH is still a matter of debate.
Hence, in this study, we sought to longitudinally investigate the magnitude and the quality of T-cell and humoral responses to mRNA-1273 vaccine up to 6 months apart from the primary cycle in successfully cART-treated PWH who had started therapy as late presenters (LP-PWH), compared to HIV-negative peers. We also aimed to evaluate the role of previous SARS-CoV-2 infection in conditioning immune responses to vaccine in this population.

Study design
In this prospective, longitudinal study, we consecutively enrolled LP-PWH (CD4 þ T-cell nadir < 350 cells/ml and/or history of AIDS-defining events) on virologically effective cART (HIV-RNA < 20 copies/ml), who received mRNA-1273 vaccine (two doses 28 days apart) at the Clinic of Infectious Diseases and Tropical Medicine, ASST Santi Paolo e Carlo, Department of Health Sciences, University of Milan, Milan, Italy. HIVnegative healthcare workers (HCWs), who received BNT162b2 (two doses 21 days apart) were also consecutively enrolled as controls.
Peripheral blood samples were collected in EDTA tubes from all study participants before vaccination (T0), 1 month and 5 months after the second dose (T1 and T2, respectively) ( Figure S1, Supplemental Digital Content, http://links.lww.com/QAD/C879). Plasma was separated by centrifugation and stored at -808C. Peripheral blood mononuclear cells (PBMCs) were obtained by Ficoll density gradient centrifugation and stored at -808C and then in liquid nitrogen.
Demographic and clinical characteristics of the study population as well as HIV-related features of LP-PWH were also collected.
The study was approved by the Institutional Ethics Committee and written informed consent was obtained from each participant. All research was performed in accordance with the Declaration of Helsinki.
Total anti-receptor binding domain antibodies Total anti-receptor binding domain (anti-RBD) antibodies were determined in plasma samples by a homemade ELISA as previously described [62]. Briefly, high-binding 96-well plates (Greiner Bio-One) were coated with 3 mg/ml of recombinant wild-type SARS-CoV-2 receptor binding domain (RBD) (ACROBiosystems) diluted in 0.5 mmol/l of carbonate-bicarbonate buffer pH 9.6 (Sigma-Aldrich) and incubated overnight at 48C. Plates were washed with PBS-0.05% Tween-20 and blocked for 1 h with PBS-2% BSA at 378C. Plasma samples were serially diluted in PBS-1% BSA in triplicates (1:40, 1:240 and 1:1440), added to plates, and incubated for 2 h at 378C. A mix of biotinylated goat antihuman k and l light chain were used at 1:2500 (Bethyl Laboratories, Inc., A80-115B and A80-116B) for detection, followed by avidin-HRP diluted at 1:2000 (ThermoFischer Scientific), for 30 min at room temperature in the dark and mild agitation. The detection was carried out with 3,3',5,5'-tetramethylbenzidine (TMB) (Invitrogen) and quenched with 1 mol/l H 2 SO 4 . Two plasma samples collected before the SARS-CoV-2 pandemic were included as negative controls, whereas an RBD-specific monoclonal antibody (Human Anti-SARS-CoV-2 Spike RBD Monoclonal Antibody, clone BIB116, Creative Diagnostics) was included as positive control. The optical density (OD) was measured by using EnSight (Multimode Plate Reader, PerkinElmer) at 450 and 620 nm, and the area under the curve (AUC) was determined with GraphPad Prism 9.2.

Receptor-binding inhibition assay
A receptor-binding inhibition assay, based on antibodymediated blockage of ACE2-Spike RBD interaction, was employed to measure plasma spike-ACE2 binding inhibition (spike-blocking) activity, which is a surrogate of neutralization activity, as previously described [63,64]. Briefly, high-binding 96-well plates (Corning) were coated with 2 mg/ml of recombinant human ACE2-Fc (InvivoGen) diluted in 100 mmol/l carbonate-bicarbonate buffer pH 9.6 (Sigma-Aldrich) and incubated overnight at 48C. Plates were washed with PBS-0.05% Tween-20 and blocked with PBS-2% BSA for 1 h at room temperature. Plasma samples were diluted 1:80 in triplicates in PBS-1% BSA and incubated with 12 ng of recombinant wild-type SARS-CoV-2 RBD-HRP (ACROBiosystems) for 1 h at 378C. Plates were washed and incubated with the preincubated plasma and RBD-HRP for 1 h at room temperature, then detected with TMB and 1 mol/l H 2 SO 4 . RBD-HRP alone and plasma with no RBD-HRP were used as controls. The OD was measured by using EnSight (Multimode Plate Reader, PerkinElmer) at 450 and 570 nm. The results were expressed as percentage (%) of inhibition, calculated as [(1 -sample OD)/average negative control OD)] Â 100.

Statistical analyses
Wilcoxon signed-rank test was used for longitudinal analyses to assess immune responses to vaccine at T1 and T2 vs. T0. Mann-Whitney U test was used for crosssectional analyses to compare study groups. Spearman's correlation test was used to correlate immune status and immune responses to vaccine in LP-PWH. Data were analyzed and graphed with GraphPad Prism 9.2.0. Permutation test in SPICE 6.0 was employed to compare polyfunctionality patterns of SARS-CoV-2-specific cytokine-producing Th1 cells in the two groups. P values <0.05 were considered statistically significant.
Demographic and clinical characteristics of the study population as well as HIV-related features of LP-PWH are summarized in Table 1.
Next, we assessed polyfunctionality of vaccine-induced SARS-CoV-2-specific CD4 þ T cells, as multifunctional Th1 cells (IFN-gþTNF-aþIL-2þ) have been described to provide a better correlate of immune protection against infection after vaccination [65]. LP-PWH and HCWs displayed comparable median distribution of polyfunctionality profiles within SARS-CoV-2-specific cytokine-producing Th1 cells at both time points, with a nonsignificant trend towards higher relative frequencies of tri-functional S-specific Th1 cells in HCWs (Fig. 1d). Frequencies of mono-functional, bi-functional and trifunctional S-specific Th1 cells within total CD4 T cells Immunogenicity of mRNA-1273 vaccine in people with HIV Augello et al. 1507  + T cell  T  s  EM CD4 + T cell  T  s  EMRA R R CD4  A  + T cell  T  s  cTfh CD4 + T cell  T  s  Total CD4  T T  + T  Continued were comparable between LP-PWH and HCWs, except for higher percentages of mono-functional TNF-aþ IFN-g-IL-2-CD4 þ T cells in LP-PWH at T1 (P ¼ 0.0023) (Fig. 1e).
No correlations were found between T-cell and humoral responses to vaccine in both study groups (data not shown).

SARS-CoV-2-specific T-cell and humoral responses in SARS-CoV-2-naive vs. -experienced vaccinees
When assessing vaccine-elicited immune responses according to previous COVID-19, we found that SARS-CoV-2-experienced HCWs showed higher anti-RBD antibodies as compared to naïve individuals only prior to vaccine administration (P ¼ 0.0355); after vaccination, they displayed greater Spike-blocking activity at T1 (P ¼ 0.0041), but not at T2 (Fig. 4a). SARS-CoV-2-recovered LP-PWH showed higher anti-RBD antibodies compared to naive at baseline (P ¼ 0.022), but this effect was not retained at postvaccination timepoints; no differences were detected in Spike-ACE2 inhibition activity between experienced and naive in LP-PWH (Fig. 4a).

As for T-cell responses, SARS-CoV-2-experienced
HCWs displayed higher S-specific IFN-gþ CD4 þ T cells at both T0 (P ¼ 0.0251) and T1 (P ¼ 0.0325), but not at T2. On the contrary, LP-PWH with prior SARS-CoV-2 infection did not show greater frequencies of Sspecific T cells neither before nor after vaccination as compared to naive (Fig. 4b). No differences were found in other SARS-CoV-2-specific T-cell populations between naive and experienced vaccinees (data not shown).

Discussion
In this prospective, longitudinal study, we sought to comparatively assess the magnitude and the quality of Tcell and humoral responses to COVID-19 mRNA   + T cell  T  s  EM CD8 + T cell  T  s  EMRA R R CD8  A  + T cell  T  s  cTfc CD8 + T cell  T  s  Total CD8  T T  + T    vaccination up to 6 months apart from the primary cycle in cART-treated late presenter PWH as compared to HIV-negative people. We also evaluated whether previous SARS-CoV-2 infection modulates vaccine-elicited immune responses in these populations.

Correlations between current immune status and
Although several recent studies have addressed the question examined in this work, most of them showing that PWH with optimal immunologic response to cART mount immune responses to SARS-CoV-2 vaccines comparable to those of the HIV-negative counterparts [25][26][27][28][29][30][31], our study specifically focused on PWH identified late in infection, a condition which has been reported to associate with enduring immune disfunction despite cART [43][44][45][46], thus raising the concern of suboptimal vaccine-elicited responses. Moreover, given that late presenters are at increased risk of poor immunological recovery [49,51,52,56,61,66], a state which has been described to be potentially linked with worse COVID-19 outcomes [8,14], it is of paramount importance to define the efficiency of immune responses to SARS-CoV-2 vaccines in such vulnerable population.
In our research, mRNA vaccine was able to expand SARS-CoV-2-specific T cells in a similar fashion in LP-PWH and HIV-uninfected people, with some minor   phenotypic differences. CM, EM, and cTfh CD4 þ T cells increased in both LP-PWH and HCWs, whereas Sspecific EMRA CD4 þ T cells only expanded in LP-PWH. Antigen-specific CD4 EMRA T cells have been reported to produce more IFN-g compared to other memory subsets and to be endowed with potent cytotoxic effector functions, therefore being implicated in protective immunity against viral pathogens such as dengue virus (DENV) [67][68][69]. The observation that DENVspecific CD4 EMRA T cells expand with repeated infections [69] suggests that a high-dose/repeated antigen exposure may be critical for the development of such cells. Therefore, the induction of potentially protective SARS-CoV-2-specific CD4 þ EMRA T cells in LP-PWH but not in HCWs may be explained with the different mRNA vaccine administered to the two groups, as mRNA-1273 has been proven to elicit stronger immune responses and to confer higher clinical protection compared to BNT162b2, possibly as a reflection of the 3.3-fold higher dose of mRNA [25,28,[70][71][72][73][74][75][76].
Unexpectedly, frequencies of S-specific Th1 cells appeared to be higher in LP-PWH. However, iMFI for such cytokine-producing T cells were comparable between the two study groups, suggesting no significant differences in the total functional T-cell response to vaccine. Indeed, while the frequency of antigen-specific T cells only evaluates the magnitude of T-cell responses, iMFI is a metric which incorporates both the magnitude and quality of the immune response, thus reflecting the total functional response of a population of cytokineproducing T cells, so that it can be used to better estimate vaccine-induced protection [65,91,92]. Furthermore, polyfunctionality patterns of S-specific Th1 cells were similar in LP-PWH and HCWs, reaffirming that vaccineinduced T-cell responses are qualitatively comparable in the two study groups. In accordance with published data on SARS-CoV-2-specific T cells in both convalescent and vaccinated individuals [15,33,58,70,71,[77][78][79][80], vaccine-induced CD4 þ T-cell responses in our cohort outnumbered CD8 þ T-cell responses.
We also found that mRNA vaccines induced spikeblocking anti-RBD antibodies in all vaccinees, irrespective of HIV status. Surprisingly, anti-RBD antibodies and spike-ACE2 binding inhibition activity, while comparable in the two groups at T1, were lower in HCWs at T2, possibly reflecting a faster decline of humoral responses to vaccine. As for the higher frequencies of vaccine-elicited CD4 þ EMRA T cells in LP-PWH, the different mRNA vaccine administered to the two groups may explain such finding.
In agreement with previous studies which described antibody levels to wane more rapidly than T cells after both natural infection [93] and vaccination [59,70,94], we observed different dynamics in T-cell and humoral responses to mRNA vaccines, with a tendency to be stable or increase at 6 months in SARS-CoV-2-specific T cells and a tendency to decline in Spike-blocking anti-RBD antibodies in both LP-PWH and HCWs.
In accordance with previous studies in PWH demonstrating the association between immune response to SARS-CoV-2 mRNA vaccines and current CD4 þ T-cell count [25][26][27][28][29][30][31][95][96][97], we hereby found a positive correlation between markers of immune recovery on cART and both SARS-CoV-2-specific CD8 þ T-cells and antibody spike-ACE2 binding inhibition activity. Interestingly, a positive correlation between CD4 þ T-cell nadir and spike-blocking function was also observed. These data suggest that PWH with low CD4 þ T-cell nadir and/or poor immune reconstitution may have suboptimal responses to a two-dose vaccine cycle, reinforcing the importance of additional mRNA vaccine doses in these subgroups of PWH.
It should also be noted that LP-PWH at baseline showed lower percentages of naïve and higher frequencies of effector memory CD4 þ and CD8 þ T cells, as well as higher circulating T follicular cytotoxic cells. Depletion of naive T cells -which are critical for effective immune responses to pathogens and vaccines -is a well known feature of HIV infection [98]. Likewise, cTfc cells have been previously reported to be enriched in PWH and proposed as an indicator of disease progression because, among PWH, those with lower frequencies of cTfc cells also have lower CD4 þ T-cell counts [99,100]. Nonetheless, in our cohort of LP-PWH, only naïve CD4 þ T-cell pools out of the different subsets showed a positive correlation with vaccine-elicited immune responses (i.e., spike-blocking activity).
Finally, when we assessed the immune response according to previous COVID-19 diagnosis, we found a unique pattern of S-specific T-lymphocyte response in LP-PWH, that was not captured in HIV-uninfected people. Indeed, SARS-CoV-2-experienced LP-PWH, while showing higher anti-RBD antibodies at baseline similar to HIV-negative individuals, did not feature raised Sspecific T cells, suggesting similar humoral yet diminished T-cell responses to natural infection. Furthermore, although previously exposed to SARS-CoV-2, LP-PWH failed to raise T-cell response at the end of the first vaccine cycle compared to SARS-CoV-2-naive ones, as instead observed in HCWs, pointing to an inability of prior SARS-CoV-2 infection to boost immune responses to vaccine in LP-PWH. These observations may appear in partial contrast with two previous studies, which described comparable SARS-CoV-2-specific T-cell and humoral responses between HIV-positive and negative individuals after SARS-CoV-2 infection [15,16]; however, PWH of our cohort have a history of late HIV presentation and feature lower median CD4 þ and CD4 þ /CD8 þ ratio than those included in the abovementioned studies, which may be responsible for reduced T-cell responses to SARS-CoV-2 natural infection. However, in accordance with our data, other recent studies showed that, compared to HIV-negative peers, convalescent PWH while developing similar humoral response, show lower and more exhausted SARS-CoV-2specific T cells [17][18][19].
Some limitations need to be acknowledged in our study. Firstly, the lack of a CD3 marker in the flow cytometry antibody panels for both AIM and ICS assay, and the definition of the cTfh subset by means of the solely CXCR5 expression. Secondly, the small sample size, which may hinder the generalizability of the observations. Thirdly, the lack of data regarding SARS-CoV-2 variants of concern (VOCs), especially Omicron, not yet emerged at the time the study was conceptualized, and which is now the most widespread all over the world [101]. Fourthly, the relatively short follow-up and the lack of data on responses to a third/fourth dose of vaccine. Furthermore, the time between previous COVID-19 diagnosis and study enrollment was unknown in both LP-PWH and HCWs, so timing differences between the two groups might have influenced the dissimilarity found in infection-induced immune responses. Additionally, the mRNA vaccines administered were different in the two study groups, as per the initial indication of the Italian Ministry of Health (1273-mRNA in LP-PWH and BNT162b2 in HCWs), and this may account for a trend towards higher immune responses in LP-PWH.
In conclusion, in our cohort of PWH with pre-cART advanced immunodeficiency and current full virologic control on long-term cART, we herein showed that a two-dose mRNA-1273 vaccine cycle is able to induce both polyfunctional SARS-CoV-2-specific memory T cells and anti-RBD Spike-blocking antibodies, which are still above prevaccine baseline levels at 6 months. Of note, immune responses to vaccine do not appear inferior to those in HIV-negative peers overall, albeit a scarce immune recovery may hinder both T-cell and humoral vaccine-elicited responses. Furthermore, unlike HIVnegative individuals, in LP-PWH natural SARS-CoV-2 infection seems inefficient in inducing specific T-cell memory and in boosting T-cell and humoral responses to vaccine, reflecting an enduring partial immune dysfunction. Altogether, these findings support the need for additional vaccine doses in PWH with a history of advanced immune depression and poor immune recovery on effective cART.

Acknowledgements
We are grateful to all the people enrolled in this study who agreed to participate to this research. Our special thanks also go to all the physicians and nurses at the Clinic of Infectious Diseases and Tropical Medicine at San Paolo Hospital in Milan who continuously help in patients' care and vaccination management, as well as laboratory personnel whose role in the conduct of the study was crucial.
Author contributions: M.A. contributed to study design, enrolled participants, collected clinical data, designed and performed the experiments, analyzed and interpreted the data, designed the figures, and wrote the manuscript; V.B. and R.R. designed and performed the experiments and interpreted the data; C.T. contributed to study design, data interpretation and critical revision of the manuscript; A.d'A.M. contributed to critical revision of the manuscript; G.M. conceived and designed the study, interpreted the data, and wrote the manuscript; all authors contributed to the article and approved the submitted version.
Funding: This work was supported by grants from Fondazione Cariplo in collaboration with Regione Lombardia and Fondazione Umberto Veronesi (CAR_-RIC20GMARC_01 and CAR_RIC20GMARC_02) and by Fondazione di Comunità Milano (FON_NA-Z20ADARM_01).

Conflicts of interest
There are no conflicts of interest.
This study was presented in part at HIV Glasgow 2022 (#P248).