Impact of Influenza and Pneumococcal Vaccines on HIV Persistence and Immune Dynamics during suppressive Antiretroviral Therapy

Abstract

Objective:

We sought to determine if standard influenza and pneumococcal vaccines can be used to stimulate HIV reservoirs during antiretroviral therapy (ART).

Design:

Prospective, randomized, double-blinded, placebo-controlled, crossover trial of two clinically recommended vaccines (influenza and pneumococcal).

Methods:

Persons with HIV on ART (N=54) were enrolled in the clinical trial. Blood was collected at baseline and days 2,4,7,14 and 30 post-immunizations. Levels of cellular HIV RNA and HIV DNA were measured by ddPCR. Expression of immunological markers on T cell subsets were measured by flow cytometry. Changes in unspliced cellular HIV RNA from baseline to day 7 post-injection between each vaccine and placebo was the primary outcome.

Results:

Forty-seven participants completed at least one cycle and there were no serious adverse events related to the intervention. We observed no significant differences in the change in cellular HIV RNA after either vaccine compared to placebo at any timepoint. In secondary analyses we observed a transient increase in total HIV DNA levels after influenza vaccine, as well as increased T cell activation and exhaustion on CD4⁺ T cells after pneumococcal vaccine.

Conclusions:

Clinically recommended vaccines were safe but did not appear to stimulate the immune system strongly enough to elicit significantly noticeable HIV RNA transcription during ART.

INTRODUCTION

A problem with curing HIV is that the virus in its latent state evades immune surveillance during antiretroviral therapy (ART) that suppresses HIV RNA to undetectable levels.¹ If ART is stopped, plasma viremia returns¹. Many curative strategies have focused on developing methods to induce the virus from latently infected cells, so that viral proteins are revealed and allow cellular reservoirs to be cleared by the human immune response, while ART prevents new cells from being infected.^2,3 Unfortunately, interventions to shock the reservoir into HIV transcription, like histone deacetylase inhibitors, IL-7, and disulfiram, have demonstrated modest activity and did not reduce the size of the circulating latent reservoir.^2,4–8 Also, safety profiles of these interventions preclude their widespread use.⁹

Prophylactic vaccination represents a safe and effective mean of activating the immune system and has been associated with transient increases of cell-free HIV RNA after vaccination for influenza,^10–19 pneumococcus,^20–22 tetanus,^18,23 hepatitis B²⁴ and cholera,²⁵ and recently SARS-CoV-2.^26,27 This stimulation of the immune system and the HIV reservoir probably occurs because standard vaccines activate not only the small subset of antigen-specific cells within the lymphocyte population but also induce a bystander activation of non-specific cells. This hypothesis is supported by the measurable increases in various markers of systemic inflammation (e.g. pro-inflammatory cytokines [interleukin (IL)-6, IL-1, C reactive protein, tumor necrosis factor (TNF)], T-cell activation [CD38⁺HLA-DR⁺] and proliferation [Ki67⁺])^28–33 after influenza and pneumococcal vaccination. Preliminary data using retrospective samples from a clinical trial³⁴ showed a significant increase in HIV transcription following vaccine administration.^35,36 Limitations in the previous study included its retrospective design, the small sample size, and suboptimal single sampling time points after each vaccine. Also, multiple vaccines were administered at certain time points, making it difficult to discern effect of single vaccines. Most notably samples from study baseline were not available, which prevented comparisons of pre-and post-vaccination time points.

Here, we performed a placebo controlled randomized crossover trial to evaluate two clinically indicated vaccines (influenza and pneumococcal) as immune stimulatory interventions for perturbing the HIV reservoir and inducing release of virus and subsequent potential for viral clearance by the activated host immune system.

METHODS

Study Design and Sampling.

In this prospective clinical trial (NCT02707692), we used a randomized, double blinded, crossover design to administer pneumococcal vaccine (Pneumovax-23^®), influenza vaccine (Fluarix^®), and placebo vaccine (saline) to participants, who were scheduled to attend a vaccination visit during each of 3 cycles followed by 5 visits at 2, 4, 7, 14, and 30 days after vaccination for collection of blood and optional genital secretion (Figure 1). All participants provided written informed consent for the study and testing of all collected biological samples. Main inclusion and exclusion criteria are listed in Supplementary Table 1. Randomization was performed by the study pharmacist who assigned participants to one of six sequences from a schedule generated prior to recruitment by the study statistician.

Figure 1. Study Design Figure.

Randomized, double blinded, crossover study design to administer pneumococcal vaccine (Pneumovax-23^®), influenza vaccine (Fluarix^®), and placebo vaccine (saline) to participants, who were scheduled to attend one vaccination visit during each of 3 cycles followed by 5 visits at 2, 4, 7, 14, and 30 days after vaccination for collection of blood and optional genital secretion. The washout period between cycles was at least 6 weeks long.

Sample Storage.

At each timepoint, peripheral blood mononuclear cells (PBMCs) were isolated from whole blood using a density gradient medium (Lymphoprep; Stemcell Technologies) and cryopreserved at −150°C in 95% fetal bovine serum (FBS) plus 5% dimethyl sulfoxide within 24 h of blood collection. Genital secretion was collected and stored³⁷.

Cellular HIV DNA and HIV RNA in PBMC.

For each longitudinal timepoint, cellular DNA and RNA were extracted simultaneously from 5 million PBMC for each time-point using AllPrep DNA/RNA Mini Kit (Qiagen, CA)³⁸. Total HIV DNA and 2-LTR circles were quantified by droplet digital (dd)PCR from extracted DNA³⁹. Copy numbers were calculated as the mean of replicate PCR measurements and normalized to one million cells as determined by housekeeping gene RPP30 (RPP30 ribonuclease P/MRP subunit p30, total cell count)⁴⁰. Unspliced HIV RNA encoding (HIV RNA_Gag) and multiply spliced HIV RNA (HIV RNA_TatRev) PCRs were performed as a duplex with Hexachlorofluorescein (HEX) (HIV RNA_Gag) and 6-carboxyfluorescein (FAM) (HIV RNA_TatRev) probes respectively^41,42. Copy numbers were calculated as the mean of replicate PCR measurements, and normalized to total RNA as determined by A260/A280 absorptivity ratio using a NanoDrop 2000 spectrophotometer (Thermo Scientific)⁴³.

Human Herpes Viruses in Genital Secretion.

Since active shedding of Human Herpes Viruses is associated with immune activation and possibly activating HIV transcriptions, we also measured levels of cytomegalovirus (CMV) and Epstein-Barr Virus (EBV) in DNA extracted from seminal plasma by real-time PCR⁴⁴.

Flow Cytometry.

Frozen cells were thawed in a 37°C water bath, washed, resuspended in phosphate-buffered saline (PBS; Corning), and counted using a BD Accuri C6 Plus (BD Biosciences). Live cells were stained using LIVE/DEAD aqua (Invitrogen) in PBS, per the manufacturer’s instructions. Next, the cells were washed and stained for extracellular markers for 30 min at 4°C in the dark in PBS plus 2% FBS plus 0.09% sodium azide. All antibodies and fluorochromes used for flow cytometry are listed in Supplementary Table 2⁴⁵. After extracellular staining, cells were fixed and permeabilized (BD fixation/permeabilization kit; BD Biosciences) for staining of the intracellular marker⁴⁵. Cells were analyzed on a BD FACS Canto analyzer (BD Biosciences). Compensation and gating were performed using FlowJo (version 10). Using the gating strategy summarized in Supplementary Figure 1, proportion of TGIT⁺PD1⁺, CD38⁺HLADR⁺, CD107a⁺, and Ki67⁺ expressions on CD4⁺ and CD8⁺ T cells were obtained for visits that occurred on Days 0 (prior to vaccination), 4, and 7.

HIV integration Site Sequencing and Analyses.

To investigate possible clonal expansion of HIV-infected CD4⁺ T cells post vaccine administration, HIV integration site sequencing was performed at two consecutive timepoints (day 4 or 7, and day 30) post influenza vaccination for 3 selected participants with the greatest HIV DNA increase during follow up. We used an adapted version of the protocol developed by Wells et al.⁴⁶ Methodological details are included in supplementary material.

Statistical Approach.

This trial tested the primary hypothesis that participants would have a higher absolute increase in levels of cellular unspliced HIV RNA_Gag 7 days after receiving either pneumococcal or influenza vaccination, when compared to receiving placebo. We also tested the effect of vaccines on the secondary virologic outcomes total HIV DNA and 2-LTR circular HIV DNA and the exploratory outcome of cellular multiply spliced HIV RNA_TatRev. Power calculations were based on pilot data³⁵. With 56 enrolled participants, 6 individuals anticipated to be lost to follow-up, a paired t-test with alpha of 0.05 for each of two comparisons and an estimate of the standard deviation of unpliced HIV RNA_Gag copies/10⁶ CD4 T-cells of 625 the trial had 80% power to detect an effect size of 253 HIV RNA_Gag copies/10⁶ CD4 T-cells absolute increase or decrease (two-sided test) in cell associated HIV RNA_Gag levels following active vaccination versus placebo. Sample size calculations were performed using Stata⁴⁷.

Cellular HIV RNA and HIV DNA measurements were generated for each visit. Since as many as 32% HIV RNA and DNA measurements fell below the limits of detection and therefore were not normally distributed, we log-transformed cellular HIV RNA and total HIV DNA levels for secondary and exploratory analyses. Undetectable levels were coded as zero, and “1” was added before taking the logarithm.

To test the difference in outcomes by vaccines on the Day-7 visit (primary outcome), we conducted pre-specified paired t-tests. The window for the Day-7 visits ranged from 5 to 9 days from vaccination, but most visits (66%) occurred on the 7^th day from vaccination. The window for days 2,4,14, and 30, shown consecutively, are [2,3], [4,4], [11,17], and [27, 33].

Through secondary analysis, we tested the effect of vaccination on each listed outcome for all visits during the 30-day follow-up. We used mixed-effects linear regression models to test the effect of vaccinations on the change in each outcome measure. The models contained random intercepts for each participant, a crossed random effect⁴⁸ for vaccination type by participant,⁴⁸, and fixed effects for vaccination type, continuous time since vaccine administration, and the level of the reservoir outcome at the time of vaccination (day zero). As post-hoc sensitivity analyses, we investigated non-linear relationships between time and reservoir outcomes.

To explore the effects of vaccination on immunological markers, we constructed 24 separate models using a similar mixed-effects linear regression approach. For 4 immunological markers (TIGIT⁺PD-1⁺, CD38⁺HLA-DR⁺, CD107a⁺, and Ki67⁺) in CD4⁺ T cells and CD8⁺ T cells and stratified by 3 cell-memory types (Central, Effector, and Terminally Differentiated memory cells), we modelled the change in the proportion of immunological marker 4 and 7 days after vaccination. In these data, the relationship between the outcomes and their corresponding baseline measures was not linear, so we categorized baseline measures using quartiles within immunological marker, T cell type, and cell-memory type.

Due to the exploratory nature of this trial, no adjustments were made for multiple comparisons and alpha of 0.05 was used for all secondary statistical tests.

RESULTS

Cohort Description.

We screened 118 individuals and enrolled 54 with HIV and on suppressive ART between December 2016 and December 2020. Most participants (63%) reported taking an integrase inhibitor-based regimen, while 18% were on a non-nucleoside reverse transcriptase inhibitor (NNRTI)-based regimen, 12% were taking NNRTI plus integrase inhibitor and 5% were on a protease inhibitor-based regimen. The 6 randomized sequences each contained 9 participants. The trial was composed mostly of Hispanic participants (50%) and non-Hispanic White participants (35%), who ranged in age from 24 to 65 and averaged 45 years old (Table 1). Most were men (83%) and reported sex with other men (80%) as their HIV risk factor. At enrollment, participants had a mean CD4⁺ T cell count of 753 per μl (SD=249).

Table 1.

Demographics and study characteristics (N=54 participants)

Baseline characteristics	In % unless stated otherwise
Race/Ethnicity
Black	7 (13.0)
Hispanic	27 (50.0)
White	19 (35.2)
Other	1 (1.9)
Mean Age, years (range)	45 (24 – 65)
Gender
Female	6 (11.1)
Male	45 (83.3)
Non-binary/transgender	3 (5.6)
Sexual Orientation
Gay	38 (70.4)
Heterosexual	11 (20.4)
Other	5 (9.3)
Mean CD4+ T cell count, cells/μl (SD)*	752.8 (248.6)
Lowest CD4+ T cell count, cells/μl	583.4 (177.6)
Mean CD4+/CD8+ ratio (SD)*	1.0 (0.4)
HIV RNA suppressed (<50 copies/ml)*	52 (98.1)

SD=standard deviation,

^*CD4 and HIV RNA were not assessed for one participant at baseline.

Not all participants completed each vaccine cycle (Consort Diagram, Supplementary Figure 2). Participants completed between 2 and 18 visits. Thirty-six participants (63%) completed the trial; all participants completed an average of 5.6 visits per cycle and 2.3 cycles; and 9 participants completed part or all visits of only 1 cycle. These cycles produced 138 Day-0 timepoints and 650 post-vaccination time points. Forty-five participants received the pneumococcal vaccine, 45 the influenza vaccine, and 48 received saline. The washout period between cycles ranged from 6 weeks to 34 weeks (average 11.5 weeks).

During follow-up, staff recorded Grade 3 and 4 adverse events. One participant had a very severe, or Grade 4, adverse event (AE) for elevated AST/ALT, and one had a severe AE (Grade 3) for elevated AST/ALT. A physician reviewed the documentation for both AEs and concluded that neither was related to study vaccinations.

Impact of Vaccine Administration on Measures of Viral Persistence.

Raw values for the HIV RNA and HIV DNA endpoints were not normally distributed due to the number of results that fell below the limit of detection. Specifically, 11.6% of time points had undetectable levels of cellular HIV RNA_Gag; 29.6% had undetectable cellular HIV RNA_TatRev; 4.6% had undetectable total HIV DNA; and 31.7% had undetectable HIV DNA 2-LTR.

For our primary analysis, we evaluated changes in reservoir measures for participants who had paired results seven days after active vaccine administration (Figure 2 and Table 2). Thirty-five participants had 7-day visits both after receiving the pneumococcal vaccine and the placebo, and 40 participants had visits for Influenza vaccine and placebo. Participants did not experience any significant differences in levels of cellular HIV RNA_Gag after pneumococcal vaccine or influenza vaccine compared to placebo (p=0.33 and p=0.30 for a non-significant decrease, respectively). Similarly, individuals did not experience statistically significant differences in cellular HIV RNA_TatRev or total HIV DNA or 2-LTR circles outcomes seven days after vaccination compared to placebo (p>0.20).

Figure 2. Distribution of cellular HIV RNA (Primary Analysis) and Total HIV DNA Outcomes by Vaccine or Placebo Injection at Day 7.

Change in levels of cellular unspliced HIV RNA_Gag, spliced HIV RNA_TatRev, total HIV DNA, and HIV DNA 2-LTR (normalized for one million cells and log transformed) from day of vaccination to seven days later by vaccine and placebo administration. Outliers are indicated with arrows.

Table 2.

Difference in mean levels of Cellular HIV RNA and total HIV DNA outcomes, 7 days after vaccination

Outcome	Pneumococcal Vaccine					Influenza Vaccine
	N	Mean (95% CI)			p-value*	N	Mean (95% CI)			p-value*
		Vaccine	Placebo	Difference°			Vaccine	Placebo	Difference°
HIV RNAGag	35	−563.9 (−1691.3, 563.4)	423.8 (−458.9 to 1306.5)	−987.7 (−2996.1, 1020.6)	0.33	40	−38.1 (−103.5, 27.4)	373.5 (−395.2 to 1142.1)	−411.5 (−1205.3, 382.3)	0.30
Total HIV DNA	35	−8.2 (−40.3, 23.9)	−5.4 (−95.8 to 85.0)	−2.8 (−94.7, 89.1)	0.95	40	−8.5 (−51.4, 34.5)	−1.2 (−80.6 to 78.3)	−7.3 (−67.7, 53.1)	0.81
HIV DNA 2-LTR	35	−1.7 (−6.9, 3.4)	1.4 (−6.7 to 9.4)	−3.1 (−8, 1.8)	0.20	40	−1.1 (−5.6, 3.5)	1.3 (−5.8 to 8.4)	−2.4 (−6.5, 1.7)	0.25
HIV RNATatRev	35	2.4 (−8.2, 13.1)	10.7 (−0.3 to 21.7)	−8.3 (−24.3, 7.7)	0.30	40	0.5 (−8.7, 9.8)	8.6 (−1.3 to 18.5)	−8.1 (−22.1, 6)	0.25

^*Paired t-test,

^°Vaccine minus placebo, CI = confidence interval

Although changes in HIV RNA_Gag levels were not significantly different than zero, levels declined after vaccination for pneumococcal and influenza (−563.9 and −38.1) but increased after placebo. These changes were attributable to one participant whose HIV RNA_Gag levels extremely declined greatly after pneumococcal vaccination (−19,418) and increased after placebo (+15,142). Excluding this participant yielded a mean non-significant difference in the change after pneumococcal and influenza vaccination compared to placebo of only −0.6 and −15.0 (p=0.99 and 0.76; 95%Cis: −85.8 to 84.6 and −113.2 to 83.3, respectively).

In our secondary analysis, we modelled the change in cellular HIV RNA_Gag after vaccination using all the follow-up timepoints (n=650, Table 3 and Supplementary Figure 3), we found no statistically significant difference in the change in cellular HIV RNA_Gag during vaccine cycles compared to the placebo cycle (pneumococcal vaccine: p=0.80 and influenza vaccine: p=0.12). Similarly, there was no statistically significant difference in the change in cellular HIV RNA_TatRev after pneumococcal vaccine (p=1.00) or influenza vaccine (p=0.55), and no statistically significant difference in HIV DNA 2-LTR circles after pneumococcal vaccine (p=0.13) or influenza vaccine (p=0.41) compared to saline. While there was no statistically significant difference in the change in total HIV DNA after pneumococcal vaccine (p=0.13), we found a significant increase in total HIV DNA after influenza vaccine administration (p=0.04). Stratifying by post-vaccine visit, we found that the increase in HIV DNA was significant only for the 14^th day after the pneumococcal vaccination (p=0.01). At the end of each cycle, changes in the levels of reservoir outcomes were not significantly different than zero (p>0.53). No significant non-linear relationships were found between time and reservoir measures during post-hoc sensitivity analysis.

Table 3.

Change in Reservoir Measures since vaccination, N=54 participants and 650 timepoints

	HIV RNAGag		Total HIV DNA		HIV DNA 2LTR		HIV RNATat-Rev
Fixed Effect	Model Coefficient* (95% CI)	p-value	Model Coefficient* (95% CI)	p-value	Model Coefficient* (95% CI)	p-value	Model Coefficient* (95% CI)	p-value
Days from vaccine to visit	−0.002 (−0.006 to 0.002)	0.26	0.001 (−0.002 to 0.003)	0.51	0.001 (−0.002 to 0.004)	0.65	0.002 (−0.002 to 0.006)	0.39
Vaccines
Saline	Reference		Reference		Reference		Reference
Influenza	0.10 (−0.02 to 0.22)	0.12	0.11 (0.01 to 0.21)	0.04	−0.04 (−0.13 to 0.05)	0.41	0.04 (−0.10 to 0.18)	0.55
Pneumococcal	−0.02 (−0.14 to 0.11)	0.80	0.08 (−0.03 to 0.19)	0.13	−0.07 (−0.16 to 0.02)	0.13	<0.001 (−0.140 to 0.141)	1.00
Level of RNA/DNA on day of vaccination°	−0.76 (−0.86 to −0.66)	<.01	−0.32 (−0.41 to −0.24)	<.01	−0.70 (−0.80 to −0.60)	<.01	−0.92 (−1.03 to −0.82)	<.01

^*Coefficients estimated from multivariable, mixed-effects linear regression. Coefficients with p<0.05 imply that effects were different than zero,

^°Corresponds to the day of vaccination for each cycle

For further post-hoc analyses, we also measured the level of shedding for CMV and EBV on 34 participants (63%) and 20% of the 159 timepoints. Twelve of these participants (35%) exhibited detectable levels of CMV during at least one visit, and 21 (62%) had at least one positive EBV result. Since more than 50% of measurements, across all timepoints, for both viruses were undetectable, we dichotomized both variables and added them separately to the final models and found that the presence of CMV or EBV DNA was not significantly associated with cellular HIV RNA or HIV DNA independent of vaccine administration (CMV: p>0.37, EBV: p>0.46).

Sustained Clonal Expanded HIV Infected Cells are Detected Over Time.

Based on the observed transient increase in HIV DNA levels post influenza vaccine administration, we sequenced integration sites of HIV DNA to track clonally expanded HIV-infected T cell for 3 selected participants with the largest increase in HIV DNA after Influenza vaccine administration. By sampling participants at two consecutive timepoints after influenza vaccination, we wanted to assess whether this increase in HIV DNA was the results of homeostatic survival over time of identical clones. As shown in Supplementary Figure 4, we detected clones contributing to ~18% of all HIV-infected T cells sampled at individual time points, which is consistent with physiological T cell dynamics combined with sampling biases⁴⁹. When looking at the longitudinal trend, clonal diversity remained stable over time in all three participants despite increasing reservoir. Longitudinally, we found clonally expanded HIV infected cells accounting for up to 43% of all HIV infected cells and large population of clones that persisted over multiple timepoints. While limited to only 3 participants, longitudinal analyses did not support a significant clonal expansion of HIV-infected T cells or change in diversity of the integration site landscape after influenza vaccine administration.

Effect of Vaccines on Cellular Immunological Markers.

In additional secondary analyses, we assessed the effect of vaccine administration on cellular immune markers of T cell activation, cycling, exhaustion and degranulation/cytotoxicity (Supplementary Table 2). After receiving the pneumococcal vaccine and compared to placebo injection, participants had higher proportions of central memory CD4⁺ T cells expressing TIGIT⁺PD-1⁺ (b=0.88, p=0.01), higher proportions of central memory CD8⁺ T cells expressing CD38⁺HLA-DR⁺ (b=1.03, p=0.048); higher proportions of effector memory CD8⁺ T cells expressing TIGIT⁺PD-1⁺ and CD38⁺HLA-DR⁺ (b=1.67, p=0.02); and a higher proportion of terminally differentiated CD4⁺ T Cells expressing CD38⁺HLA-DR⁺ (b=5.89, p<0.01). No immunological cellular markers exhibited significantly different proportions after Influenza vaccine administration compared to receiving a saline injection. Models which produced these results included variables to control for the number of days since vaccination and the baseline proportion of each activated CD4⁺ T cells or CD8⁺ T cell.

DISCUSSION

The ‘shock and kill’ method is a possible strategy to unmask and eradicate cells latently infected with HIV². This strategy requires a reversal of latency sufficient to produce a measurable decrease in the HIV reservoir. No agent that can be used safely in vivo has thus far been able to achieve this benchmark.

Standard vaccines may be an attractive option, as vaccines are an important part of routine clinical care for people with HIV, are by their nature immune stimulatory and may be associated with transient increases in plasma HIV RNA^15,27.

In this randomized blinded crossover trial, we measured cellular HIV RNA (spliced and unspliced) as a measure of HIV transcription and HIV DNA (total and 2-LTR) in participants receiving standard influenza and pneumococcal vaccines or placebo injections in random order. We measured markers of T cell activation, degranulation, exhaustion and cycling at baseline as well as 4- and 7-days post injection. In our study administration of standard vaccines against influenza and pneumococcus was safe for people with HIV, but our trial did not support a significant increase in HIV transcription during suppressive ART as compared to placebo injection despite an increase in activated and exhausted T cells post-pneumococcal vaccine. This latter finding may suggest that the immune activation induced by the pneumococcal vaccine was strong enough to be detected in our immunologic assays, while the stimulus from the influenza vaccine did not leave a discernable mark at least in the assays performed in this study. All study participants received the seasonal influenza vaccine at least 6 weeks prior to enrollment, which might have impacted our ability to observe a bystander activation post vaccination in this study. Alternatively, our flow cytometry panel might be limited by the number of markers to capture vaccine-specific immune response and didn’t include markers of earlier immune activation (e.g., CD69 or CD25). We hypothesize that lack of measurable HIV transcription could be because the presence of ART blocked HIV propagation and the viral production stimulated by the vaccines was too small and short-lived to result in detectable viremia. We did observe a transient increase in HIV DNA post influenza vaccine administration without evidence of clonal expansion of HIV infected cells, which is unlikely to be clinically significant and may not have been significant if adjusted for multiple comparisons.

Our data are in line with previous studies not showing significant effects of vaccine administration on various virologic parameters^50–54. Other studies demonstrated transient increases in plasma viremia^12,15 or cell-associated HIV RNA^36,55. Several factors may have contributed to these conflicting observations, including differences in plasma viremia, CD4 T cell counts, and ART status of the study participants at the time of study, sampling intervals, and sample size.

The study has several limitations. First, the limited sample size for the primary analysis due to larger than expected loss of follow up. While the crossover design was initially selected to increase our power, the long follow up became problematic and many participants dropped off due to competing obligations. While there was no discernable effect on cellular HIV transcription (primary outcome) after vaccine administration as compared to placebo, this negative result is unlikely to change even with a larger sample size because the observed difference was in the opposite direction of what was hypothesized.

Third, we selected two inactive vaccines for our trial (Pneumovax-23^® and Fluarix^®) and it is possible that other products that were live-attenuated or with different antigens or adjuvants might provide different results, possibly in combination with other latency-reversing agents or immunotherapies. Our study predominantly enrolled men: sex-based differences in immune responses post vaccine administration are well-documented^56–58, and it would be important to determine if vaccine-mediated activation of the viral reservoir differs between sexes and genders.

Finally, total HIV DNA was used as a surrogate marker of the viral reservoir. Although these measures are often used to evaluate curative interventions, these may not be the most sensitive marker of reservoir size given the large proportion of defective virus. Also, for this study, we decided against isolating CD4 from PBMC since CD4 T cells are often lost during the sorting process, especially activated T Cells, which have been shown to downregulate CD4 receptor and are often the cells releasing HIV RNA.

The quest continues for strategies to provoke HIV expression from quiescent cellular reservoirs and then eradicate the exposed, infected cells. We found in this trial that while pneumococcal vaccine caused an estimated increase ranging from 1.03 times (95%CI: 0.03–2.04) to 5.89 times (95% CI: 2.19 9.58) in the change in percent of T cell activation in secondary analyses, it did not appreciably increase HIV RNA transcription. Similarly, influenza vaccine administration did not increase HIV RNA_Gag, but we did observe a potential transient increase in HIV DNA compared to placebo after influenza vaccination. While the outcome of our clinical trial was mostly negative, our study provides insights about the effects of standard vaccine administration on viral and immune markers, which will be administered in people with HIV and will be useful for the design of future clinical trials. For example, future clinical trials should consider a different study design to maximize retention and perform targeted outreach to enroll a more diverse study population. Future studies should apply more sensitive and specific markers for assessing the reservoir size, such as measuring replication-competent or intact provirus on sorted CD4 T Cells. We also recommend selection of more immunogenic vaccines to maximize immune stimulation, and more use a broader panel of immunological markers, and markers specific to the tested antigens.