A total of 14 pigtail macaques (male/female, aged 4-11 years, 6-13 kg) were used. Supplementary Table 1 details animal characteristics, including MHC haplotypes and experimental vaccination history. Prior to enrollment, all animals were pre-screened and confirmed seronegative for the presence of antibodies to West Nile, dengue, and Zika viruses. At least 2 months prior to enrollment into the study, eight animals were previously enrolled in studies in which they received an experimental hepatitis B virus (HBV) vaccine consisting of a combination of CD180 targeted DNA and recombinant protein vaccines comprised of HBV core and surface antigens (38) and/or a replicating RNA COVID-19 vaccine (39) ( Supplementary Table 1 ) and were evenly distributed between the control and experimental groups. Seven pigtail macaques were infected intravenously with 10,000 infectious units (I.U.) of SIVmac239M (40) (gift from Dr. Brandon Keele, AIDS and Cancer Virus Program, Frederick National Laboratory for Cancer Research) and then co-infected with ZIKV 9 weeks later. All animals were inoculated subcutaneously with 5 x 10⁵ PFU of the Brazil_2015_MG strain of ZIKV (GenBank: KX811222.1), as previously described (30). All animals were subject to a Simian AIDS monitoring protocol as defined by the WaNPRC guidelines (41). ZIKV RNA was not detected in any specimen tested at any timepoint in one animal in the SIV-infected group (Z14109); therefore, this animal was excluded from all post-ZIKV analysis. Animals were observed daily and full physical exams were conducted at each experimental timepoint, as previously described (30). All animals were euthanized at the study endpoint at 4 weeks post-ZIKV infection under deep anesthesia, in accordance with the 2007 American Veterinary Medical Association Guidelines on Euthanasia, by administration of Euthasol^® (Virbac Corp., Houston, TX).

SIV plasma viremia was evaluated by quantitative real time reverse transcription polymerase chain reaction (RT-PCR) by the Virology and Immunology Core at the WaNPRC, as previously described (41), and by the NIAID DAIDS Nonhuman Primate Core Virology Laboratory (NHPCVL) for AIDS Vaccine Research and Development Contract. Complete blood counts (CBC) and serum chemistries were performed by the Research Testing Service (RTS) at the University of Washington Department of Laboratory Medicine. Peripheral CD4 counts were determined from CBC using flow cytometry-based methods by the Virology and Immunology Core at the WaNPRC, as previously described (42).

Peripheral blood mononuclear cells (PBMC) were isolated from NHP whole blood collected pre-SIV inoculation (Wk-3) and at weeks 2 and 6 post-SIV inoculation, as previously described (30). PBMC were maintained in RPMI medium supplemented with 10% fetal bovine serum (FBS; HyClone), 2 mM L-glutamine, 5 mM sodium pyruvate, 1x Antibiotic Antimycotic Solution, and 10 mM HEPES (cRPMI; complete RPMI). RPMI medium used for the ZIKV inoculation was supplemented with 1% FBS, 2 mM L-glutamine, 5 mM sodium pyruvate, 1x Antibiotic Antimycotic Solution, and 10 mM HEPES (iRPMI; infection RPMI). Vero cells (WHO, Geneva, Switzerland) were cultured in complete Dulbecco’s modified Eagle medium (cDMEM) supplemented with 10% FBS, 2 mM L-glutamine, 5 mM sodium pyruvate, 1x Antibiotic Antimycotic Solution, 10 mM HEPES and 1X non-essential amino acids. Vero cells tested negative for mycoplasma contamination. All cells were maintained in a 37°C incubator with 5% CO₂. Brazil Zika virus stock (GenBank: KX811222.1) was used for the PBMC inoculation.

Following overnight incubation, PBMC cell suspensions were prepared, and the cell concentration and viability measured using the Countess 3 Automated Cell Counter (ThermoFisher Scientific). Approximately 6 x 10⁶ PBMC were inoculated with ZIKV at an MOI of 2 in a total volume of 200 µL RPMI infection medium (iRPMI) at 37°C for 2 hours (h). Cells were gently mixed by pipetting at 20 minute (min) intervals during incubation. After 2 h, the cells were spun at 300 relative centrifugal force (rcf) for 3 min at room temperature and the inoculum carefully removed without disturbing the cell pellet. The cells were washed with 300 µL iRPMI and then resuspended in pre-warmed complete RPMI (cRPMI). A total of 5 x 10⁵ PBMC were added to each well of a 24-well plate containing 1 mL of cRPMI. The plates were returned to 37°C and incubated until the designated timepoint for sample collection. At 4, 24, and 48 h post-ZIKV inoculation, supernatants were collected and spun at 300 rcf for 3 min at 4°C. 100 µL Versene solution (ThermoFisher Scientific) was added to each well to dislodge adherent cells from the TC plate. The clarified supernatant was transferred to a new tube and banked at -80°C until further processing by plaque assay or qRT-PCR assay. The Versene solution containing PBMC was added to the PMBC cell pellet from the supernatant and then spun at 300 rcf for 3 min at 4°C. The supernatant was carefully removed from the cell pellet and discarded. The cell pellet was then resuspended in 700 µL QIAzol for RNA analysis.

Vero cells (WHO, Geneva, Switzerland) were seeded at a density of 5 x 10⁵ cells per well in 6-well plates. The next day, the medium was removed from the monolayers and 200 µL of 10-fold serial dilutions of virus-containing supernatant in DMEM containing 2% FBS was added to respective wells in duplicate. Vero cell monolayers were incubated at 37°C for 2 h, with rocking at 15 min intervals. Monolayers were overlaid with 1% low-melting point SeaPlaque^® agarose (Lonza), set at 4°C for at least 20 min, and then returned to the 37°C incubator. Plaques were visualized and counted 4 days later by crystal violet staining.

The NanoString nCounter platform (NanoString, Seattle, WA, USA) was used to quantify mRNA counts in PBMC processed from whole blood at pre- and post-SIV and ZIKV infection timepoints. RNA was isolated from cryopreserved PBMC samples collected at pre-SIV inoculation (Wk -3), 2 and 6 weeks post-SIV using the miRNeasy Mini Kit (QIAGEN). RNA was isolated from 1-2 x 10⁶ PBMC resuspended in 700 µl QIAzol collected at week 7/Day -14 pre-ZIKV inoculation, and at days 2, 4, 7, 10, 14 and 21 post-ZIKV challenge using the miRNeasy Micro Kit (QIAGEN). From each sample, 100 ng RNA was loaded in accordance with manufacturer’s instructions for targeted expression with 2 custom-built curated NanoString Human Panels of 44 and 60 genes of interest which both represent gene biomarkers of innate immune activation and response, interferon response, and inflammatory response. Due to the small number of genes represented on the Code Set, nCounter data normalization was performed using a method which calculates a ratio between genes of interest to housekeeping genes (43). The ratio is calculated by dividing the counts of the genes of interest by the geometric mean of 4 housekeeping genes which have the lowest coefficient of variance across all samples. This is done for each sample independently, which generates normalized expression for the genes of interest. Significant differences (nominal P-val <0.01) were determined between baseline and post-infection time points for each group and between groups at each time point.

At necropsy representative samples of all tissues and organs were collected in formalin and after fixation were paraffin embedded and sectioned at 3-5 μm. For basic histology, sections were stained with hematoxylin and eosin. All histological findings are summarized in Supplementary Table 2 .

Isolated PBMC, rectal and peripheral lymph node biopsy cells were assessed for viability with a live/dead stain (Life Technologies) and stained with a panel of antibodies, details described in Supplementary Table 3 , in brilliant stain buffer (BD Biosciences) to identify innate immune cells as previously described (30) and were gated according to Supplementary Figure 1 . Paraformaldehyde fixed cells were acquired on a LSRII (BD Biosciences) using FACS Diva software (version 8). Samples were analyzed using FlowJo software version 10.8.1 (FlowJo, LLC). All events were first gated on FSC singlets, CD45⁺ leukocytes, live, and then cells according to FSC-A and SSC-A profiles. Immune cells were identified as follows: plasmacytoid dendritic cells (DCs) (CD20^-CD3^-HLA-DR⁺CD14^-CD123⁺CD11c^-), myeloid DCs (CD20^-CD3^-HLA-DR⁺CD14^- CD123^-CD11c⁺), monocytes (CD20^-CD3^-HLA-DR⁺CD14⁺CD16^+/-), and neutrophils (CD3^-CD11b⁺CD14⁺SSC-A^Hi). AXL receptor tyrosine kinase (AXL) positive cells were identified after FMO subtraction and meeting a cellular threshold (>100 cells/gate).

Cytokine and chemokine levels in plasma and cerebrospinal fluid (CSF) were analyzed using a custom nonhuman primate ProcartaPlex 24-plex immunoassay (ThermoFisher Scientific), per the manufacturer’s protocol. The levels of the analytes were assessed on a Bio-Plex 200 system (Bio-Rad) and analyzed per the manufacturer’s protocol.

Viral RNA load was assessed in plasma, CSF, rectal cytobrush supernatant, and tissues using a ZIKV-specific RT-qPCR assay, as previously described (30). RNA was isolated from plasma and rectal cytobrush supernatant collected pre-challenge and at 1, 2, 3, 4, 7, 10, 14 and 21 days post-infection (dpi) and at necropsy (24-28 dpi). RNA was isolated from CSF collected pre-challenge and at days 4, 7 and 21 post-challenge and at necropsy (24-28 dpi). Rectal and PLN biopsy tissues were collected pre-challenge and at days 7 and 21 post-challenge and lymphoid and gut tissues collected at necropsy (24-28 dpi). The iScript Select cDNA Synthesis Kit (Bio-Rad) was used for gene-specific cDNA synthesis and cDNAs were quantified on a QuantStudio Real-Time PCR System (ThermoFisher Scientific). Ct values <39 in at least 2 of the triplicates and falling within the standard curve determined from diluted known quantities of ZIKV genome were considered positive.

Plasma and/or CSF quantification by ELISA of human soluble CD14 (sCD14), human fatty acid binding protein 2 (FABP2) (Fisher Scientific, Waltham, MA) or human LPS binding protein (LBP) (Biometec, Germany) was performed per the manufacturer’s instruction. Plasma was diluted as follows: 1:200 (sCD14), 1:2 (FABP2), or 1:3 (LBP). CSF was diluted 1:5 (sCD14). Results were analyzed using Prism version 10.4.1 (GraphPad) and using a four- or five- parameter logistic (4- or 5-PL) function for fitting standard curves.

NHP sera and/or plasma samples were assessed for anti-ZIKV envelope (E) IgG binding titers by an Enzyme-Linked Immunosorbent Assay (ELISA). Purified NHP IgG (MyBioSource MBS539659) was serially diluted to establish a range of IgG standards. ZIKV E protein (Fitzgerald Industries International, 30-1932) was diluted to 0.5 μg/mL and was used as the capture antigen. Capture antigen and IgG standards were coated overnight on high-binding 96-well plates (Costar 3590) to produce test and standard wells, respectively. All wells were subsequently blocked in blocking buffer (5% w/v nonfat dried milk (Bio-Rad Laboratories 1706404) and 0.05% v/v Tween-20 in PBS). Samples were diluted 1:100, 1:200, and/or 1:400 and tested in triplicate. NHP IgG standard and test wells were probed by a goat anti-monkey IgG antibody conjugated to Horseradish peroxidase (HRP) (Abcam ab112767). SureBlue Reserve TMB substrate (KPL) was added to all wells to initiate a color change reaction catalyzed by HRP. Reaction was stopped after 30 minutes with 1N HCl (VWR) and absorbance at 450nm (Abs₄₅₀) was measured on an EMax plate reader (Molecular Devices). Abs₄₅₀ of standard wells were used to produce a 5PL logistic fit (GraphPad Prism). Abs₄₅₀ of test wells were converted to μg/mL of anti-ZIKV E IgG binding titers via the 5PL logistic fit.

NHP sera collected pre-challenge (Day -14) and at necropsy (24-28 dpi) were tested in PRNT assay for neutralizing antibody production, as previously described (44). The PRNT assay was performed using serial two-fold dilutions of the serum samples. The highest serum dilution reducing plaque numbers by 50% (PRNT₅₀) were determined with a limit of detection (LOD) of 1:50. The assay was repeated twice in triplicate using the ZIKV Brazil 2015 virus.

Non-parametric statistical methods were employed for all comparisons, unless otherwise noted. Specifically, Kruskal-Wallis tests were used for comparisons across timepoints in in vitro experiments, paired Wilcoxon tests were used to evaluate cell fraction differences to baseline at each timepoint, and Mann-Whitney tests were used to compare continuous values across groups. All analyses were conducted using two-sided tests at the 0.05 level. Analyses were conducted in Prism version 10.4.1 (GraphPad). Significant differences (nominal P-val <0.01) in gene expression were determined using a t test that compared baseline and infection timepoints for each group and between groups at each timepoints.

Pigtail macaques (PTM) (n=7) were infected with SIVmac239M and blood was collected prior to and at 2 and 6 weeks post-SIV infection. Both post-infection timepoints are associated with significant declines in peripheral CD4 counts and correspond with SIV peak and viral setpoint, respectively ( Supplementary Figure 2 ). To evaluate whether SIV infection alters the permissivity of peripheral blood mononuclear cells (PBMC) to Zika virus (ZIKV) infection, we isolated fresh PBMC from SIV+ PTM (n=4) at weeks 2 and 6 post-SIV infection and inoculated cells with ZIKV ex vivo. PBMC isolated prior to SIV infection (pre-SIV, SIV-) served as a control. At 4, 24 and 48 hours post-ZIKV infection (hpi), cells and culture supernatants were collected to measure ZIKV RNA and viral titer. As expected, control pre-SIV PBMC were permissive to ZIKV infection, as measured by qRT-PCR and plaque assay, with peak viral replication at 24 hpi ( Figures 1A, B , Supplementary Figure 3 ). ZIKV RNA levels in SIV+ PBMC collected at week 2 post-SIV were significantly decreased at 24 hpi compared to pre-SIV PBMC while SIV+ PBMC collected at week 6 post-SIV had similar ZIKV RNA levels to that in pre-SIV PBMC ( Figure 1A ). The kinetics of ZIKV were similar in pre-SIV and SIV+ PBMC; however, significantly lower levels of infectious virus were observed at 24 and 48 hpi in SIV+ PBMC collected at week 6 post-SIV ( Figure 1B ; Supplementary Figure 3 ). Supernatants from ZIKV-infected PBMC cultures were subjected to multiplex immunoassay to measure cytokine and chemokine concentration changes at 4, 24, and 48 hpi. All cultures accumulated the pro-inflammatory cytokines MCP-1 and VEGF-A during the 48 hr post-infection period; however, there was only a trend for an increase of IL-5 (p = 0.061) in Week 2 SIV-infected cultures relative to pre-SIV at 24 hr ( Supplementary Figure 4 ). These data suggest that cells from acutely SIV-infected animals are less permissive to ZIKV infection.

Monocyte frequencies increase in the blood during HIV and SIV infection and are the primary targets of ZIKV infection (28–30, 32). Since ZIKV replication was lower in PBMC derived from SIV-infected animals, we next determined whether this was because of fewer ZIKV cellular targets. To test this, we evaluated innate cells in fresh blood from SIV+ and SIV- PTM by flow cytometry. During acute SIV infection, the frequency of CD16+ monocytes and dendritic cells (DCs) significantly increased at Week 2, with levels returning to pre-infection levels at Week 6 ( Figure 1C ). We next examined AXL expression on CD16+ monocytes and DCs in response to SIV infection, as the TAM receptor tyrosine kinase is a known ZIKV viral entry receptor (45–48). Ex vivo AXL expression on CD16+ monocytes and DCs was unchanged at Weeks 2 or 6 post SIV-infection ( Supplementary Figure 5 ). Thus, PBMC from SIV-infected PTM have similar/greater levels of ZIKV cellular targets in comparison to naïve PTM and these cells express similar levels of surface AXL, but they are less vulnerable to ZIKV infection ex vivo.

We next hypothesized that anti-viral responses induced early within SIV infection could influence ZIKV permissibility. To test this, we used a targeted custom-built NanoString nCounter gene expression assay to investigate a panel of 84 immune-related genes in PBMC collected at pre-SIV and at Weeks 2 and 6 post-SIV infection. Differential gene expression analysis was performed for each timepoint post-SIV infection relative to pre-SIV PBMC and identified several significantly upregulated innate immune and interferon stimulated genes (ISGs) ( Figure 1D ; Supplementary Table 4 ). With SIV infection, several genes related to innate immunity were upregulated in expression compared to pre-SIV PBMC. Notably, IFIT1, MX2, and OAS1 ISGs were increased in expression at both Weeks 2 and 6 post-SIV. ISG20 and ISG15 were significantly upregulated in expression at Week 6, while CXCL8 and CCL2 encoding monocyte chemoattractant proteins were significantly upregulated at Week 2 post-SIV. Retinoic acid-inducible gene-I (RIG-I) signaling activated by viral infection induces the expression of these antiviral genes that, in turn, are known to restrict ZIKV replication (49, 50). These data indicate that despite the presence and expansion of ZIKV cellular targets during acute SIV infection, increased innate immune gene expression in PBMC could render monocytes refractory to ZIKV infection.

SIV-infected PTM were co-infected with ZIKV at 63 days (9 weeks) post-SIV infection (SIV+ZIKV+) and compared to SIV-naïve PTM infected with ZIKV (SIV-ZIKV+) ( Figure 2A ). This timepoint from the post-acute phase of SIV infection was selected as it corresponds with the establishment of viral setpoint (median SIV viremia 5.41 (1.63-6.18) log₁₀ copies/mL of plasma) and there is evidence of immunosuppression including lowered, yet stable peripheral CD4 counts (median 399 (333–901) cells/µL of blood) and decreased frequencies of CD4 T-cells in the gut mucosa relative to SIV-naïve controls ( Supplementary Figure 2 , Supplementary Table 1 ). Blood, peripheral lymph nodes (PLN), cerebrospinal fluid (CSF), and rectal samples (biopsy, cytobrush) were longitudinally collected according to Figure 2A for 4 weeks until the time of necropsy. In the SIV+ZIKV+ cohort, SIV viremia and peripheral CD4 counts remained stable post-ZIKV coinfection and there was no evidence of enhanced gut barrier dysfunction ( Supplementary Figures 2 , 6 ). The most notable histologic findings were in Z16297, who had proliferative-occlusive pulmonary arteriopathy with thrombosis and infarction, which is a retroviral-strain-associated disease that was likely secondary to the SIV infection ( Supplementary Table 2 ). Overall, these findings suggest that acute ZIKV co-infection does not have a significant effect on peripheral SIV viral replication or disease progression during the 4-week period examined.

To assess the impact of the post-acute phase of SIV infection on ZIKV pathogenesis and tissue tropism, ZIKV burden was evaluated in longitudinal specimens and in tissues at necropsy by qRT-PCR. ZIKV RNA was not detected in one animal in the SIV+ZIKV+ cohort (Z14109) in any longitudinal sample tested nor in any necropsy tissue ( Supplementary Tables 7 , 8 ); therefore, there was no evidence of productive ZIKV infection, and the animal was excluded from all subsequent post-ZIKV analysis. In the SIV-ZIKV+ cohort, plasma viremia peaked 2-4 days post infection (dpi) (median 3 dpi) and ZIKV was cleared in the plasma in most animals (6/7) by 7 dpi ( Figure 2B ). The mean viral kinetics in this cohort was consistent with our previous findings in ZIKV-infected PTM (30). In contrast, in SIV+ZIKV+ PTM, peak ZIKV viremia was variable (1-10 dpi; median 4 dpi) and virus was still present in most animals (4/7) by 7 dpi ( Figure 2B ). Accordingly, median plasma viral loads at 2 and 3 dpi trended to be higher or significantly higher (2 dpi, p = 0.108; 3 dpi, p = 0.050) in the naïve animals compared to SIV+ZIKV+ PTM and at 7 dpi trended higher (p = 0.054) in SIV+ZIKV+ PTM ( Supplementary Figure 7 ). Together, these data indicate that peak ZIKV viremia in co-infected animals is heterogeneous, owing to inter-animal variability; however, there is a delay in initial ZIKV viremia and clearance in the periphery in SIV-infected pigtail macaques that is consistent with previous findings in rhesus macaques (34).

In pigtail macaques, we previously detected ZIKV viral RNA in the lymphoid and gastrointestinal tissues (30); therefore, we next examined if SIV infection altered ZIKV viral burden or persistence in tissues. In PLN, ZIKV RNA was detected in most animals (SIV- 5/7; SIV+ 4/5) at 7 dpi and in a proportion of animals (SIV- 2/7; SIV+ 2/6) at 21 dpi ( Figure 2B ). In rectal biopsy tissue, ZIKV RNA was detected in 4/7 of SIV- and 2/5 of SIV+ animals at 7 dpi, and at the time of necropsy (24-18 DPI) was detected in 1/7 of SIV- and half (3/6) of SIV+ animals at necropsy ( Figure 2B ; Supplementary Figure 7 ), suggesting ZIKV persistence in the rectum during SIV infection. In rectal cytobrushes, ZIKV RNA was predominantly detected in SIV- animals at 7 dpi (4/7 PTM) and was sporadically, but more consistently detected in SIV+ animals at 10 dpi (4/6 PTM), and the total viral burden in rectal cytobrushes trended (p = 0.171) to be higher during SIV+ZIKV+ co-infection ( Figure 2B ; Supplementary Figure 7 ). ZIKV RNA was not detected in CSF of any animal at any timepoint throughout the study ( Supplementary Table 7 ) and corroborated our previous findings in PTM (30). Overall, these data suggest that ZIKV infectivity of the gut mucosa is more common and may be more persistent during post-acute SIV infection.

To evaluate the impact of SIV infection on ZIKV tropism and tissue persistence, we measured ZIKV viral burden in lymphoid, gastrointestinal (GI) mucosal, and neuronal tissues at necropsy (24-28 dpi). Consistent with our previous findings (30), ZIKV RNA was commonly detected in lymphoid tissues, variably in GI tissues, but not in brain tissue (including the brainstem, hippocampus, frontal lobe, parietal lobe, and occipital lobe), supporting the absence of neurotropism, as suggested by CSF analysis ( Figure 2C , Supplementary Table 8 ). At necropsy, the total number of ZIKV positive tissues (p = 0.108) and overall viral burden (p = 0.171) trended higher in SIV+ZIKV+ animals compared to SIV-ZIKV+ animals ( Supplementary Figure 8 ). Upon further investigation, we focused on ZIKV viral burden in gastrointestinal-associated tissues (e.g. GI mucosa and GI draining lymph nodes) and other lymphoid tissues. In both cohorts, ZIKV RNA was detected in at least one lymphoid tissue in most animals (5/6 SIV+ZIKV+; 4/7 SIV-ZIKV+) ( Figure 2C ) and ZIKV burden in individual lymphoid tissues or across all lymphoid tissues did not significantly differ between groups ( Figure 2D ; Supplementary Figure 8 ). In the GI-draining lymph nodes (mesenteric and iliac), ZIKV RNA was detected in both cohorts, but the individual tissue burden trended to be higher (p = 0.109) and total GI viral burden was significantly greater (p = 0.023) in the SIV+ZIKV+ group when compared to the SIV-ZIKV+ group ( Figures 2C, D ; Supplementary Figure 8 ). ZIKV RNA was detected in at least one GI tissue in most SIV+ PTMs (5/6) but only in a few SIV- PTM (2/7) ( Figure 2C ), reinforcing our previous observation that persistent ZIKV infection in the gut is uncommon in healthy PTM (30). Consistent with this, the median number of ZIKV-positive gastrointestinal tissues trended higher (p = 0.057) in SIV-infected animals and the total GI viral load also trended higher (p = 0.125) during SIV co-infection ( Supplementary Figure 8 ). Moreover, a strong positive correlation (r = 0.7370, p = 0.0063) was observed between ZIKV burden in GI tissues and GI-draining lymph nodes ( Figure 2E ). In summary, while ZIKV distribution and persistence are similar across lymphoid tissues, ZIKV detection and viral burden are more common and persistent in the gut mucosa and GI-associated lymphoid tissues during post-acute SIV infection.

As there was evidence for altered ZIKV pathogenesis in SIV-infected animals, we next evaluated whether this was associated with changes to ZIKV cellular targets or the immune response. Humoral responses are important for the control of ZIKV infection (51) and we measured anti-ZIKV envelope IgG responses in longitudinal plasma samples. SIV-ZIKV+ animals generated robust binding IgG antibodies that were detected at 7 dpi and peaked at 14 dpi, however these responses were lower overall in SIV+ZIKV+ animals (AUC, p = 0.014) ( Figure 3A ). Despite the difference in anti-ZIKV IgG between the two cohorts, similar levels of neutralizing antibodies (NAb) were generated against ZIKV at the time of necropsy (7/7 SIV-ZIKV+, 6/7 SIV+ZIKV+) ( Figure 3B ). Thus, neutralizing antibody production was equivalent in naïve and SIV+ZIKV+ animals, despite overall lower anti-ZIKV envelope IgG binding antibodies in the co-infected group.

Consistent with our previous findings (30), there is rapid and robust expansion of inflammatory CD16+ (non-classical and intermediate), CD16- (classical) monocytes, and DCs in the blood in the first few days after ZIKV infection (median peak 2 dpi), which corresponds to peak ZIKV viremia ( Figure 4A ; Supplementary Figure 9 ). In contrast, the expansion of CD16+ and CD16- monocytes was severely dampened and delayed during SIV infection (median peak 8.5 days) and corresponded with the delayed peak ZIKV viremia observed in these animals ( Figure 2B ). Cellular analysis in the tissues revealed that CD16- monocytes/macrophages were robustly and significantly recruited to the rectum and PLN in SIV- negative animals, whereas in contrast, inflammatory CD16+ monocytes/macrophages were recruited to the tissues during SIV+ZIKV+ co-infection ( Figure 4A ; Supplementary Figure 9 ). DC frequencies in the tissues were similar between the two groups ( Supplementary Figure 9 ). AXL expression was not significantly changed on ZIKV cellular targets in whole blood, rectum and PLN in either group post-ZIKV infection ( Supplementary Figure 10 ). These findings suggest that increased recruitment of inflammatory monocytes and macrophages to lymphoid and gastrointestinal tissues during SIV infection, which are also cellular targets of Zika virus infection, may contribute to ZIKV viral persistence at these sites.

Neutrophils are important for ZIKV dissemination and pathogenesis (30, 52, 53). Blood neutrophils declined during SIV+ZIKV+ co-infection starting at 1 dpi and were significantly lower in frequency in comparison to SIV-ZIKV+ animals through 21 dpi ( Figure 4B ). The frequencies of neutrophils in the rectum at 21 dpi were also lower in SIV+ZIKV+ compared to SIV-ZIKV+ animals, with no differences observed within the PLN ( Supplementary Figure 11 ). To assess neutrophil function, we evaluated plasma concentrations of myeloperoxidase (MPO), a neutrophil granule and marker of inflammation, during the first 4 days of ZIKV infection. Interestingly, prior to ZIKV infection, elevated levels of MPO were detected in 5/6 SIV+ animals while lower concentrations of MPO were detected in 7/7 SIV- animals ( Figure 4C ). Post-ZIKV infection, concentrations of plasma MPO continued to be higher in SIV+ZIKV+ animals in comparison to SIV-ZIKV+ animals ( Figure 4C ). These data suggest that neutrophils are more inflammatory during SIV infection and impaired in their recruitment to the blood and to gastrointestinal tissues post-ZIKV.

The inflammatory response to ZIKV in the plasma was evaluated using a multiplex immunoassay. In both groups, ZIKV induced pro-inflammatory responses, characterized by transient increases in interleukin-1 receptor agonist (IL-1RA), monocyte chemoattractant protein-1 (MCP-1), and vascular endothelial growth factor A (VEGF-A) ( Supplementary Figure 12 ), with no major differences between groups. Although we found no evidence of ZIKV infection in neuronal tissue, markers of neuroinflammation were evaluated in longitudinal CSF specimens. IL-1RA, IL-6, IL-8, and VEGF-A were elevated in the CSF of several animals across groups at varying timepoints ( Supplementary Figure 12 ). Transient increases in IL-8 were detected in the CSF of a few SIV-ZIKV+ animals after ZIKV infection, which is evidence of neuroinflammation. At necropsy, two animals in the SIV-ZIKV+ group (L07201 and Z08062) had mild, multifocal demyelination and axonal loss in the brainstem, which may be a result of neuroinflammation despite no evidence of ZIKV infection at this site ( Supplementary Tables 2 , 8 ). CSF concentrations of sCD14, another marker of neuroinflammation, did not change with ZIKV infection in either group ( Supplementary Figure 6 ). Overall, no significant differences between SIV+ZIKV+ vs SIV-ZIKV+ groups were found for any analytes detected in plasma or CSF during ZIKV infection, suggestive that SIV infection does not enhance ZIKV-induced systemic inflammation or neuroinflammation.

To further examine SIV effects on immune responses to ZIKV, we performed targeted gene expression analysis on longitudinal PBMC specimens collected post-ZIKV infection using a custom-built NanoString Code Set for interrogating 64 genes marking innate immune activation, inflammatory and antiviral responses. Changes in gene expression were determined at each timepoint post-challenge in comparison to uninfected timepoints. Gene expression was not significantly changed 7 weeks post-SIV infection relative to naïve PBMC. This indicated that while SIV infection stimulated gene expression, the changes just prior to ZIKV challenge were not significantly different from baseline levels. Following ZIKV challenge, a total of 23 genes were significantly differentially expressed (22 upregulated and 1 downregulated), 14 genes in SIV-ZIKV+ (13 up and 1 down), 19 genes in SIV+ZIKV+ (19 up), and 10 genes in both groups (10 up) ( Figure 5B ; Supplementary Table 5 ). ZIKV induced robust innate immune gene activation in SIV-ZIKV+ animals during acute infection (2-4 dpi), with the expression of most genes returning to baseline by 7 dpi ( Figure 5A ). The kinetics of the gene expression mirrors peak ZIKV viremia (median 3 dpi) and time to viral clearance (median 7 dpi). While these innate immune genes were also strongly upregulated in SIV+ZIKV+ co-infected animals during acute infection (2-4 dpi), in contrast to the SIV-ZIKV+ group, the gene signature was maintained in SIV+ZIKV+ animals throughout infection and remained highly expressed 7-21 dpi. The kinetics of gene expression during SIV+ZIKV+ co-infection corresponds to the shift in peak ZIKV viremia (median 4 dpi) and viral clearance (median 10 dpi). We also found that the expression of 10 genes were significantly different between SIV+ZIKV+ and SIV-ZIKV+ animals, primarily at 14 and 21 dpi, and predominantly consisted of genes associated with type I IFN signaling and ZIKV viral control (ISG15, IFIT1, MX1, ISG20, IRF7) ( Figure 5C , Supplementary Table 6 ). These data demonstrate that in the blood, SIV-ZIKV co-infection leads to persistent upregulation of genes associated with inflammation and innate immune activation. This persistent and hyperactivated innate immune state in response to ZIKV co-infection likely contributes to the impaired peripheral ZIKV clearance and its persistence in tissues.

Studies in NHP and human have revealed that ZIKV can persist in bodily fluid and tissues, including semen, CSF, lymphoid and gastrointestinal tissues for several months after infection; however, the mechanisms contributing to persistence are poorly understood (14, 33). In the pigtail macaque model, we previously found ZIKV persistence in lymphoid, but not in gastrointestinal, tissues (30). As several of our animals had ZIKV RNA present in gastrointestinal tissues at necropsy and was more common in SIV+ (5/6) versus SIV- (2/7) animals, we next examined the immune responses associated with ZIKV gut persistence, irrespective of SIV status. We performed correlational analysis between immune responses in the blood or rectum to levels of GI ZIKV burden, a summation of the gut viral load in each tissue at necropsy ( Supplementary Table 9 ). Among these, higher concentrations of plasma MPO (r = 0.8091, p = 0.0015), indicative of an inflammatory neutrophil response, higher concentrations of FABP2 (r = 0.6559, p = 0.0181), a marker of gut barrier dysfunction, and greater infiltration of inflammatory CD16+ monocytes and macrophages into the rectum (r = 0.5866 p = 0.0385) all positively correlated with higher ZIKV gut burden at necropsy ( Figures 6A, B ).