Serial Cervicovaginal exposures with Replication-deficient SIVsm induce higher Dendritic Cell (pDC) and CD4+ T-Cell Infiltrates not associated with prevention but a More Severe SIVmac251 Infection of Rhesus Macaques

Abstract

Objective

Intravaginal exposure to SIV acutely recruits IFN-α producing plasmacytoid dendritic cells (pDC) and CD4+ T-lymphocyte targets to the endocervix of nonhuman primates. We tested the impact of repeated cervicovaginal exposures to noninfectious, defective SIV particles over 72 hrs on a subsequent cervicovaginal challenge with replication-competent SIV.

Methods

34 Female Indian Rhesus macaques were given a three-day, twice-daily vaginal exposures to either SIVsmB7, a replication-deficient derivative of SIVsmH3 produced by a CEMX174 cell clone (n=16), or to CEM supernatant controls (n=18). On the fourth day, animals were either euthanized to assess cervicovaginal immune cell infiltration or intravaginally challenged with SIVmac251. Challenged animals were tracked for plasma viral load and CD4 counts and euthanized at 42 days post infection.

Results

At the time of challenge, macaques exposed to SIVsmB7, had higher levels of cervical CD123 pDCs (p=0.032) and CD4+ T-Cells (p=0.036) than those exposed to CEM control. Vaginal tissues showed a significant increase in CD4+ T-Cell infiltrates (p=0.048), and a trend towards increased CD68+ cellular infiltrates. After challenge, 12 SIVsmB7-treated macaques showed 2.5-fold greater daily rate of CD4 decline (p=0.0408), and viral load rise (p=0.0036) as compared to 12 control animals.

Conclusions

Repeated non-productive exposure to viral particles within a short daily timeframe did not protect against infection in spite of pDC recruitment, resulting instead in an accelerated CD4+ T-Cell loss with an increased rate of viral replication

Introduction

Women comprise over half of all individuals living with HIV-1 worldwide with intravaginal sex as the predominant route of transmission [1]. However, the rate of infection amongst exposed females is statistically low, indicating that only a fraction of viral exposures result in productive infection [2–5]. At each exposure, HIV can interact with the cervical epithelium [6, 7] mediating cellular infiltration [6, 8] and potentially engaging TLR-7 on dendritic cells [8–10]. The effects of non-productive HIV-1 or SIV cervicovaginal exposures (absent infection) on subsequent infection outcomes remain unclear.

The CAPRISA004 trial found a positive association between a pre-existing inflammatory cytokine profile within cervicovaginal fluids and subsequent HIV infection [11], supporting previously held views that vaginal immune activation increases HIV-1 infectivity. In parallel, a study examining SIV-infectivity in rhesus macaques after direct topical application of TLR-7/9 ligands to the cervicovaginal mucosa showed increased localized inflammation, and increased viral load after challenge [12]. However, there are well-documented cases of women with a history of chronic multi-year HIV-1 exposure without infection indicating that exposure-initiated mechanisms may reduce future infectivity upon re-exposure [13–15]. Of interest, studies of resistance in these women have largely failed to identify HIV-specific responses as a determinant of protection [15] but a loss of HIV-1 “resistance” in a subset of these women after reduced sex work (i.e., HIV-1 exposure) has suggested that resistance is sustained by innate rather than adaptive responses [16]

Acute cellular changes following cervicovaginal mucosal infection have been studied. Infection is established within CD4+ T-Cells and other permissive cells of the lamina propria [17–20] following a robust cellular infiltrate of pDCs and CD4 T-Cells [6], and increased mRNA expression of pro-inflammatory cytokines/chemokines by day 3 post infection (DPI) [21]. Of interest, Li et al. demonstrated that infiltrating cervical pDCs produce interferon-alpha (IFN) within hours of SIV exposure, preceding CD4+ T-Cells infiltrates by 72 hours [6].

The anti-HIV properties of pDCs and Type I IFN have been documented and involve intrinsic and secondary antiviral mechanisms, such as the induction of tetherin or APOPBEC family members, and the modulation of innate and adaptive effectors [22–27]. In a recent clinical study in individuals with chronic HIV-infection, IFN immunotherapy sustained viral suppression during anti-retroviral therapy interruption in a subset of subjects supporting the interpretation that IFN can inhibit HIV-1 production/replication [28]. However, the role of infiltrating pDCs as a source of type I IFNs and its impact on acute SIV infection if present before infection as a result of repeated non-infectious viral exposures remains unclear. Taking advantage of a replication-deficient SIVsmB7 clone derived from SIVsmH3, able to bind CD4 and fuse with its target without subsequent viral replication [29–31], we tested the hypothesis that repeated vaginal exposures to SIV in the absence of productive infection would decrease viral infection potential and/or decrease replication kinetics as a consequence of sustained cervical pDC infiltration despite the increase in local CD4+ T-Cell infiltrates.

Methods

Ethics Statement and Animal Procedures

Healthy Indian Rhesus monkeys (Macaca mulata) were acquired from the Caribbean Primate Research Center (CPRC) of the University of Puerto Rico (UPR) - Medical Sciences Campus (MSC). Animals were quarantined for six months and maintained at the AAALAC-accredited facilities of the Animal Resources Center, UPR-MSC.

All animal studies were approved by the UPR-MSC, Institutional Animal Care and Use Committee (IACUC), and comply with the Guide for the Care and Use of Laboratory Animals. Animal Welfare Assurance Number: A3421, Protocol number: 3380308.

In addition, steps were taken to reduce suffering in accordance with the recommendations of the Weatherall report, “The Use of Non-human Primates in Research.” For instance, all procedures were conducted under anesthesia by using ketamine 10–20 mg/kg, delivered IM.

Phase I

Ten (6 and 4) macaques were given either intravaginal SIVsmB7 or CEM mock control inoculations. Inoculations were given twice daily for 3 days in either the follicular or luteal phase, as shown in Figure 1 and Supplemental Digital Content 2. Macaques were euthanized on the fourth day.

Figure 1. Phase I/II criteria.

The study was separated into two separate phases. Ten macaques were used in Phase I: Four CEM mock control intravaginally inoculated animals and six SIVsmB7 intravaginally inoculated animals. These animals were used to assess cellular infiltration in the cervical and vaginal epithelium. For Phase II, 24 animals were divided equally either into CEM mock-control- or SIVsmB7-inoculated animals. The inoculation sequence for Phase I was repeated and SIVmac251 challenge.

Phase II

Female macaques were housed together prior to use in the study to allow the animals to reach menstrual synchrony. Twenty-four macaques, 12 macaques per group, were given either intravaginal SIVsmB7 or CEM mock control inoculations (Table 1 and Figure 1). Inoculations were given twice daily for three days during the luteal phase. On the fourth day post inoculation, each macaque was given a single 1ml challenge dose SIVmac251. Blood was taken prior to the challenge and on days 7, 14, 21, and 42 to determine plasma viral load (PVL) and CD4 count. Macaques were euthanized on days 42–46 based on analysis focused on early viral kinetics after a single exposure and infection.

Euthanasia was performed only on fully anesthetized animals by injection of Pentobarbital Sodium at 390 mg/ml; 1cc/10lbs IV. Vaginal and cervical tissues were taken for analysis.

Viruses

SIVmac251 was diluted 1:2 from a 20,000 TCID₅₀/ml viral stock grown in SPF rhesus macaque PBMC produced by Dr. Ron Desrosiers (New England National Primate Research Center, Harvard Medical School) and kindly provided by Dr. Nancy Miller (NIAID) through contract #N01-AI-30018. Animals were challenged with 1ml of the diluted viral stock. In vivo, titration was done as described in Supplemental Digital Content 1, Methods.

SIVsmB7 is a virus like particle (VLP) derived from a clone of a CEMx174 cell line stably infected with SIVsmH3. SIVsmB7 is non-infectious due to a 1.6 kbp deletion including integrase, vif, vpr and vpx genes. Cell-free SIVsmB7 and CEMx174 supernatant (CEM mock control) were isolated by standard 20% (w/v) sucrose gradient ultracentrifugation. P27 ELISA was used to determine 500 μg P27 SIVsmB7 dose. CEMx174 dose was established by equal protein quantification with SIVsmB7 dose.

CD4 and CD8 Counts

CD4 and CD8 counts where monitored by TruCount Absolute Count Kits (BD Bioscience) used according to manufacture protocol.

SIV Viral Loads

Plasma Viral Load was assessed by qRT-PCR as previously described [32]. Full method can be found in Supplemental Digital Content 1, Methods.

Immuno-histochemistry

Samples from the vagina and cervix were harvested post mortem from each animal and fixed in 4% paraformaldehyde and embedded in paraffin for sectioning, SafeFix II (Fisher Scientific), or frozen. Immunohistochemical staining for CD123, CD68, CD4 and Mx1 was conducted as previously reported [6]. (See Supplemental Digital Content 1 for complete method). CD123+ cells were considered plasmacytoid dendritic cells (pDC) as a previous report by Li, et al [6] has shown that CD123+ cells infiltrating into the endocervical subepithelium 3 days post SIV exposure stain positive for both HLA-DR and IFNa confirming these cells are pDC.

Statistical Analysis

Cross-sectional Phase 1 and Phase 2 two group comparisons between SIVsmB7-exposed and CEM mock-control-inoculated animals were made using Wilcoxon rank sum tests, Fisher’s Test or Student’s T-Tests. Two tailed p-values less than 0.05 were considered statistically significant.

In order to evaluate the rate of change for LogVL and CD4, a change point model with random intercept terms for each monkey and fixed effects for baseline CD4, days-post-infection (DPI) (overall and after 14 DPI), a group indicator for SIVsmB7 exposure, and two-way interaction terms between DPI and group, was fitted. This model accounts for the within individual correlation in repeated measurements and allows for a different slope before and after 14 DPI. The models are listed below, with model variables estimates described in the results. The function (DPI -14)⁺ = 0 when DPI≤14 and (DPI -14)⁺ = DPI -14 when DPI>14

Model Equations for SIVsmB7 or CEM Innoculated Animals

$$\begin{array}{c}\mathit{LogV}{\widehat{L}}_{\mathit{CEM}}/CD{\widehat{4}}_{\mathit{CEM}}={\mathrm{\beta }}_0+\left({\mathrm{\beta }}_1\times \mathit{age}\right)+\left({\mathrm{\beta }}_2\times \mathit{Baseline}\_CD4\right)+\left({\mathrm{\beta }}_3\times \mathit{DPI}\right)+\left({\mathrm{\beta }}_4\times {(\mathit{DPI}-14)}^{+}\right)\\ \mathit{LogV}{\widehat{L}}_{\mathit{SIVsmB}7}/CD{\widehat{4}}_{\mathit{SIVsmB}7}=\left({\mathrm{\beta }}_0+{\mathrm{\beta }}_0^{\prime }\right)+\left({\mathrm{\beta }}_1\times \mathit{age}\right)+\left({\mathrm{\beta }}_2\times \mathit{Baseline}\_CD4\right)+\left({\mathrm{\beta }}_3+{\mathrm{\beta }}_3^{\prime }\right)\times \mathit{DPI}+\left(\left({\mathrm{\beta }}_4+{\mathrm{\beta }}_4^{\prime }\right)\times {(\mathit{DPI}-14)}^{+}\right)\end{array}$$

Statistical analysis was done using R 2.14.1 and Prism.

Results

Phase I: Acute cellular infiltrates after 72 h of SIVsmB7/CEM mock control exposure

Ten female macaques followed during their natural luteal or follicular phases were used in order to initiate three-day, twice-daily SIVsmB7/CEM mock control dosing within a three days period after the start of a menstrual phase (as exemplified for two luteal-staged animals in Supplemental Digital Content 2). Animals were inoculated with CEM mock or SIVsmB7 as described in Figure 1 and in the methods.

Consistent with previously published data showing that infectious SIV exposure resulted in acute endocervical pDC and CD4+ T-Cell infiltration within 3 days of SIV exposure/infection [6], replication-deficient SIVsmB7 treated animals had a significant higher number of endocervical CD123+ pDC (p=0.0317, mean: 1328 cells/mm² vs 435 cells/mm²) and CD4 T-Cells (p=0.0357, CD4 mean: 141.2 cells/mm² vs 17.92 cells/mm²) as compared to CEM mock-control-inoculated animals at 72 hrs (Figure 2A) irrespective of hormonal phase. Moreover in animals exposed to SIVsmB7 we detected a trend in increased expression of Mx1 (an interferon stimulated gene, Figure 2A) localized in areas enriched for CD123+ cells. This is, consistent with recruitment of CD123+ pDCs to the endocervix and local production of Type I interferons. Although Exposure to SIVsmB7 did not result in increased vaginal pDC infiltration (Figure 2B), it did result in higher levels of CD4-positive cells in vaginal tissues (Figure 2B. p=0.0476, mean: 164.8 cells/mm² vs 37.95 cells/mm²). Similar to pDCs, SIVsmB7-treated animals had higher (although not statistically significant) levels of CD68 macrophages than control animals in the cervix, but not in the vaginal tissue. Altogether, results confirm that exposure to replication-deficient SIV can modulate the tissue microenvironment, allowing us to directly test how these local viral-induced changes (i.e., pDC and CD4 T-Cell infiltrates) would impact a challenge with infectious SIVmac251.

Figure 2. Infiltrate staining and quantification of cervico-vaginal tissue taken from SIVsmB7 and CEM mock-control-treated animals.

(A – Endocervix) [Top Panel: CD123 Staining 20x Image] SIVsmB7 treatment (left) induced large-scale pDC infiltration with cells directly underneath or near the columnar epithelium as compared to mock control (p=0.0317) [2nd panel: Mx1 Staining 20x Image] Corresponding with increased pDC infiltration in SIVsmB7 treated macaques there was increased levels of Interferon Responsive Gene product (ISG) Mx1 in treated macaques compared to controls [3rd panel: CD4 Staining 20x Image] SIVsmB7 treatment (left) induced moderate infiltration of CD4+ T-Cells as compared to CEM mock treatment (p=0.0357) [Bottom panel: CD68 Staining 20x Image] SIVsmB7 treatment (left) did not induce a significant increase in macrophage infiltration as compared to CEM mock control (right), but visually, macrophages appeared closer to and, in some cases, permeated the columnar epithelium, suggesting a trend of increased infiltration (p=0.0952) (B – Vagina) [Top panel: CD123 Staining 20x Image] SIVsmB7 treatment (left) induced no noticeable level of pDC infiltration as compared to CEM mock-control-treated animals (bottom) (p=0.2000) [Middle panel: CD4 Staining 20x Image] SIVsmB7 treatment (left) induced a sizable increase in CD4 T-Cell staining, with many positive staining cells found directly underneath the striated vaginal epithelium and comparably fewer cells found in the vaginal mucosa of CEM mock-control-treated animals (right) (p=0.0476) [Bottom panel: CD68 Staining 20x Image] Macaques had moderate levels of macrophage staining within the vaginal mucosa regardless of treatment. This level of CD68 staining was not impacted by treatment (p=0.3524). Pair-wise comparisons were done with Wilcoxon Rank-Sum tests.

Phase II: Outcome of SIVmac 251 infection following 72-h pre-conditioning with SIVsmB7 or CEM Control

Twenty-four female macaques were challenged with SIVmac251 after undergoing 72-h exposure (as described above) to either SIVsmB7 (n=12) or CEM control (n=12). The groups were similar in weight and inoculation date relative to menstrual cycle (Table 1; Supplemental Digital Content 3) and there was no significant difference in major histocompatibility complex (MHC) Class I allele distribution (Fisher’s exact test, p>0.05) or for those associated with spontaneous control between groups [33–35] (Supplemental Digital Content 4 “Table”).

Endogenous estrogen levels at the time of challenge between SIVsmB7 and CEM mock control arms were not significantly different (p=0.84) (Table 1), indicating the absence of differences in estrogen-related structural epithelial factors known to impact viral infectivity [36–38].

As it has been shown that cervicovaginal inflammation is common amongst captive rhesus macaques [39], a 28-plex non-human primate luminex assay was done within 1 week of challenge to assess levels of pro-inflammatory mediators in cervicovaginal lavages (CVL) in 13 of the 24 macaques (6 SIVsmB7/7 CEM mock). Twelve of the cytokines/chemokines tested were below the limit of detection and no cytokine/chemokine tested showed any significant difference between groups. Results for selected cytokines and chemokines are summarized in Supplemental Digital Content 5.

Following a single, intravaginal SIVmac251 challenge using an infectious dose previously determined via an intravaginal titration (see Supplemental Digital Content 1, “Methods”) all 24 macaques became infected, with a detectable viral load within 7 or 14 days post infection (DPI)(Figure 3B). Viral load peaked in all animals by 14 to 21 DPI, with a mean viral load of 6.63 and 8.02 for CEM mock control- and SIVsmB7-treated, respectively (Student’s T-Test: p=0.0720).

Figure 3. Change point model of log plasma viral load (LogVL) and CD4 for Phase II macaques.

(A) ΔCD4 for SIVsmB7 treated macaques was significantly higher than in CEM mock control. By 14 DPI, there was mean loss of −355.8 CD4/μl in SIVsmB7 treated animals compared to −120.4 CD4/μl for controls (Student’s T-Test: p=0.0268 [inset]). (B) Twenty-two of 24 Macaques in both treatment arms reached peak viral load on days 14 or 21. Four of the CEM macaques had significantly lower log viral loads than the remaining eight CEM mock macaques [t-test on Area Under Curve (AUC) with p<0.0001], which was not explained by known correlates of spontaneous SIV control or protection suggesting low viral load was normal variation. At no time point was difference in Log viral load significant, however starting at 14 DPI and ending at 42 DPI there was a trend towards increased viremia in SIVsmB7 inoculated macaques (Student’s T-Test: p=0.118). (C) As we detected a significant difference in ΔCD4 between groups, we generated a linear mixed effect spline model to investigate CD4 change between treatment groups over time (in days, as “DPI”). This model predicts an increased loss of CD4+ T-Cells per day of 14.5 cells/μl over top of the CEM mock control animals culminating at peak viremia (day 14). (D) To investigate the behavior of log viral viral load a linear mixed effect spline model for LogVL was also generated. The LogVL model predicts SIVsmB7-treated animals to have an increased log viral load rise per day of 0.106 over top of CEM mock treated animals up until peak viremia on day 14.

Although we didn’t detect a significant difference in peak viral load, there was a trend towards increased viral load in SIVsmB7 treated macaques as compared to CEM mock treated animals (Figure 3B). We did detect a statistically significant decline in CD4+ T-Cells at 14 DPI in SIVsmB7 treated macaques as compared to control animals (Student’s T-Test: p=0.0268, mean: −355.8 CD4/μl vs CEM −120.4 CD4/μl) (Figure 3D) consistent with a more severe CD4 depletion after infection in the SIVsmB7 treated females. Notably, four of the macaques in the control group (animals 8D4, M847, 31R, and 530) had lower plasma viral loads. These animals were similar to the other animals in the control and experimental groups in terms of MHC genotypes, batch of doses used, day of manipulation, estrogen levels, weight, age and history of parity (Data not shown). Analysis of baseline CVL samples from two of these macaques, 53O and 31R, showed no difference to other infected control macaques.

Model of Log Viral Load Rise and CD4 Decline in Phase II Macaques

As our data detected that SIVsmB7 exposures prior to an infectious challenge caused a significant decline in CD4 at peak viral load after infection, we generated linear models for CD4 or log viral load kinetics over days (DPI) and included a spline (or change-point) at peak viral load day 14 representing the end of the acute change in variables after infection (illustrated in both viral load and CD4 decline data in Figure 3C/D). As expected, both models showed that (1) DPI is a significant determinant of both CD4 decline and viral load rise for animals in either group as exemplified by its coefficient β₃ in control (p=0.0436 for CD4 and p<0.0001 for logVL; Methods and Table 2) and β₃′ in SIVsmB7 treated (p=0.0408 for CD4 and p=0.0036 for logVL), and (2) day 14 is a valid change-point as illustrated by the reverse of direction of estimates for β₃ and β₄ for the control and β₃′ and β₄′ for the SIVsmB7 group.

The model for CD4 T-Cell kinetics closely approximated our CD4 data and indicated a significantly different change rate between groups before day 14 (β₃′, p=0.0408; Figure 3 and Table 2). The effect of infection in the CEM mock control animals showed a loss of 10.06 CD4+ T-Cells/μl each day until peak viremia (140.8 loss at peak viral load). SIVsmB7-treated macaques had a 2.5-fold higher daily loss of CD4+ T-Cells totaling 25.0 CD4+ T-Cells/μl per day (342.8 loss at peak viral load).

As for viral load, the model showed a significantly higher rate of change in viral load for the SIVsmB7 over CEM control before day 14 (β₃′, p=0.0036; Figure 3 and Table 2) in support of greater CD4 T cell loss. Specifically, log viral load rise each day after infection was predicted to be 0.553 log per day in SIVsmB7-treated macaques as compared to 0.446 log per day for CEM mock control animals. Taking into account baseline CD4 count, CD8 count, CD4/CD8 ratio, age, or weight, as independent factors did not change the model’s output for viral load rise.

Altogether our results suggest that noninfectious SIV exposures and associated conditioning of the local microenvironment result in a greater CD4+ T-Cell decline with higher viral load.

Discussion

We show for the first time that an early, local, innate and CD4 T-Cell response to replication-deficient SIV particles can enhance the rate of CD4+ T-Cell decline and viral load rise upon a subsequent productive infection. Our results were contrary to our original hypothesis that the observed increase in SIV-induced cervical pDC infiltrates and local interferon-mediated mechanisms would inhibit productive infection. Instead, our data are consistent with viral exposures acutely modulating levels of vaginal and cervical CD4+ T-Cell infiltrates, providing a greater pre-existing “substrate” for viral infection, potentially overwhelming any inhibition provided by pDC and resulting in an accelerated decrease of CD4+ T-Cells in association with viral replication. Importantly, our data assessing the impact of recruited pDC was collected during natural progesterone high periods independent of pharmacological treatments commonly used to synchronize animals as these have been shown to impair pDC function [40, 41].

Our results parallel a report by Wang et al. where female macaques were pre-exposed to TLR-9 or TLR-7 ligands, CpG, or imiquimod, respectively, before and during SIVmac251 challenge [12]. Wang et al. did not quantify differences in cellular infiltrates before challenge nor model viral kinetics to peak viral load, but did observe increased inflammatory cytokines and infiltrate after dosing and a higher viral load eight weeks post infection. Their reported data is consistent with our observed accelarated CD4 loss and viral load rise in SIVsmB7 pre-exposed animals. We speculate the Wang et al. study showed a more marked increase in viral replication than our study due to use to two challenges using a SIVmac251 one log higher (greater viral inoculum [42]) and use of optimized TLR-7/9 ligands which may have induced a greater inflammatory environment (greater local immune activation [43]).

Importantly, our results are directly relevant to both low-dose SIV repeated challenge study designs commonly used in vaccine studies and to settings of repeated HIV exposures in women. Regarding SIV study designs, our data suggests that early SIV challenges may encounter a different microenvironment than later challenges as a result of immune cell infiltration. Infections after repeated exposures may increase local cell targets and result in accelerated infection kinetics and/or greater GALT depletion when compared to infections occurring at the start of exposures. Per human cohorts, our data suggests that a sex worker in an area with high HIV-1 prevalence may have different levels of intravaginal immune cell infiltration, independent of co-infections, as compared to women with low levels of HIV-1 exposure. The presence of increased CD4+ T-Cells and CD68 cells in women at higher risk of HIV-1 infection or in non-human primates following multiple SIV exposures may decrease the efficiency of prophylactic interventions by increasing infection efficiency. However, it would be of interest to determine if vaccine-elicited anti-viral antibodies in the presence of an increase in cellular infiltration could also provide a greater effector population to harness potentially protective ADCC mechanisms in high-risk uninfected women. Taken together our data suggests that SIV/HIV exposures can contribute to local immune modulation affecting the host tissue response to future viral infection as already established for other STD co-infections [4, 16, 44, 45]. Future experiments using a lower infectious dose or repeated low to high-dose escalation of virus would be needed to directly show if SIVsmB7 pre-exposures would increase infection rate over controls as suggested by our data.

Our study had limitations that we addressed. First, our replication-deficient virus, SIVsmB7, was derived from the human cell line, CEMx174. Previous studies have shown that xeno/allogeneic adaptive immune responses to MHC alone can block infection[46]. We found that both SIVsmB7 and CEMx174 supernatants had MHC Class I and Class II proteins (Supplemental Digital Content 6). The 72-hour time frame used for this experiment would limit development of adaptive xenoantigenic responses; moreover, we used a challenge virus derived from a rhesus macaque cell line to avoid this possible confounder response (Methods). It has been shown that innate allogenic responses can be rapid in response to live tissue grafts [47], yet little is known about the speed and strength of such responses through mucosal administration of MHC proteins. Mucosal studies to date have only examined such responses after the induction of the adaptive response [48, 49]. We interpret that since MHC proteins are present in both SIVsmB7 and control it is unlikely there would be a differential response yet to be detected based on MHC alone. Importantly, CD4+ T-Cell and pDC infiltration was significantly greater in SIVsmB7 treated animals indicating the control proteins did not induce change in absence of SIV proteins. As for other proteins shown in Supplemental Digital Content 6 & 7, proteomic analysis of both CEM mock and SIVsmB7 doses had comparable protein content to SIVsmB7 outside of the presence viral proteins or Mov10. Interestingly, the presence of Mov10 (Supplemental Digital Content 7) in SIVsmB7 proteomic analysis is consistent with enriched particle generation as previous reports have shown that Mov10 incorporates into viral particles [50]. It has also been shown that ectopic expression of Mov10 is an inhibitor of HIV-1/SIV infectivity [50–53]; however, our data together with a recent report showing that basal levels of Mov10 had no impact on HIV-1 replication or infectivity argues against its antiviral potential [51].

Second, previous studies have shown that captive rhesus macaques can have divergent pre-existing cervicovaginal inflammation [39, 54]. To control for this, we did random testing of cervicovaginal lavages from rhesus macaques and found no difference in inflammatory cytokine levels between treatment groups (Supplemental Digital Content 5) in spite of observing differential CD4 changes between groups as described. The only difference we did detect was a 1-year difference in age between groups (Table 1). We do not expect this minor difference impacts our data as previous research has evidenced no significant difference in immune parameters, such as immune activation or serum cytokine level, amongst young to middle-age macaques [55].

Third, our conclusions address the cervicovaginal mucosa and not other mucosal routes of exposure. Future studies will need to establish whether repeated replication-deficient rectal exposures induce CD4+ T-Cell infiltration and affect the severity of a productive infection. Intriguingly, sustained SIV-resistance has been reported following repeated low-dose intra-rectal SIV challenges, highlighting a potential difference in compartments or difference between 72 hour versus chronic exposure effects or both [56].

In conclusion, our data support a model where viral exposures in absence of productive infection induce changes in cervicovaginal cellular infiltrates that do not prevent infection. Furthermore, our data may indicate an added risk to repeated short-term exposures not previously identified, and introduces an unaccounted tissue change variable that may impact the outcome of SIV challenge in studies evaluating protection by low-dose repeated challenge formats using female macaques.