Eleven (11) male MCM (Macaca fascicularis, 4–10 years old) were procured from Bioculture US (Immokalee, FL). Major histocompatibility complex (MHC) haplotypes, which were determined by MiSeq sequencing of PBMC at the Genomic Servies Core of the Wisconsin National Primate Research Center, and 8 the 11 MCM shared at least one copy of the M1 MHC haplotype. The MCM were screened for Mtb, as well as herpes B virus, SIV, SRV, and STLV, and housed at the University of Pittsburgh. All experimental procedures were conducted according to the ethical regulations of the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC), and complied with the standards outlined in the Animal Welfare Act, the “Guide for the Care and Use of Laboratory Animals (8^th Edition)”, and “The Weatherall Report”. The University is fully accredited by AAALAC (accreditation number 000496), and its OLAW animal welfare assurance number is D16–00118. The IACUC reviewed and approved the study protocols 21048970 and 24044961. All veterinary procedures were performed under ketamine sedation. The MCM were monitored twice daily for any health issues.

All MCM were infected intravenously with a 3×10³ tissue culture infectious dose (TCID₅₀) of SIVmac239 which was grown in rhesus PBMC (Primate Assay Laboratory of the California National Primate Research Center, University of California, Davis CA). Plasma virus level was serially monitored by qPCR, as detailed below.

Sixteen weeks after SIV infection, the MCM were randomly allocated into two vaccination groups: IV BCG (n=5), and ID BCG (n=6) (Figure 1a). Animals were vaccinated with BCG Danish Strain 1331, which was expanded at Colorado State University following the Aeras protocol (29). The BCG was prepared for injection by suspending in phosphate-buffered saline (PBS) containing 0.05% Tween 80 (Sigma-Aldrich) to prevent the inoculum from clumping. For ID inoculations, the upper arm was shaved and disinfected with chlorhexidine followed by isopropanol, and the standard human dose of 5×10⁵ CFU was injected into the dermis in a volume of 1 mL. For IV inoculation, the saphenous vein was used to inject 5×10⁷ CFU in 1 mL total volume, the dose we showed previously to confer high-order protection in SIV-infected MCM (17). To reduce the risk of BCG dissemination in these SIV-infected animals, the IV BCG-vaccinated MCM were treated with an anti-BCG drug regimen (isoniazid, 15mg/kg; rifampicin, 20mg/kg; ethambutol, 55mg/kg (HRE)) given by mouth daily for 8 weeks starting 3 weeks after vaccination, as done previously (17).

Whole blood was collected by venipuncture. PBMC were isolated using Ficoll-Paque PLUS gradient separation (GE Healthcare Biosciences) and cryopreserved in RPMI 1640/fetal bovine serum (FBS) containing 10% DMSO in liquid nitrogen. Bronchoalveolar lavage (BAL) was done by instilling and recovering warmed PBS (4 × 10 mL washes) under direct visualization using a pediatric bronchoscope. The recovered BAL fluid (BALF) was pelleted to collect the cells and a 15 mL aliquot of BALF was cryopreserved. The cells were resuspended into ELISpot media (RPMI 1640, 10% heat-inactivated human albumin, 1% L-glutamine, and 1% HEPES), counted, and apportioned for flow cytometric staining.

PET-CT imaging using radiolabeled 2-deoxy-2-¹⁸Fluoro-D-glucose (FDG) was conducted within 3 days of necropsy. A MultiScan LFER-150 PET-CT scanner (Mediso Medical Imaging Systems, Budapest, Hungary) was used, as detailed previously (30, 31). Co-registered PET-CT images were analyzed using OsiriX MD software (v 12.5.2, Pixmeo, Geneva, Switzerland) to measure the total FDG avidity of the lungs, thoracic lymph nodes (32). Total FDG avidity correlates with overall inflammation.

To differentiate T_RM cells from intravascular cells, IV anti-CD45 antibody staining was performed as previously described (16, 33, 34). Briefly, 5 minutes before euthanasia, animals were maximally bled and then slowly infused intravenously with 5 mL of purified Alexa Fluor 488-conjugated αCD45 antibody to achieve a dose of 100 μg/kg. For the ID BCG vaccination group, 5 of 6 MCM were injected IV with anti-CD45 antibodies. When harvested tissues were analyzed by flow cytometry, CD45 immune cells present in parenchyma were unstained, while those in the vascular were positively stained with anti-CD45 antibody.

Five months after vaccination, animals were humanely euthanized and necropsied to collect lung tissue, thoracic and peripheral lymph nodes, liver, and spleen. Each tissue sample was divided and a portion was fixed in neutral-buffered formalin for histopathology, with the remainder homogenized to yield a single cell suspension, as previously described (24). To detect any viable BCG, serial dilutions of this homogenate were plated onto 7H11 agar and incubated at 37°C, 5% CO₂ for three weeks. Single cell suspensions were portioned immediately for flow cytometric staining with the remainder cryopreserved in RPMI 1640/FBS containing 10% DMSO in liquid nitrogen. Formalin-fixed tissue was embedded in paraffin, sectioned, and stained with hematoxylin and eosin for microscopic examination.

The number of SIV RNA copies per milliliter was measured using a two-step real-time qPCR assay. Viral RNA was extracted from 500 μL of plasma using an automated sample preparation platform QIAsymphony SP (Qiagen, Hilden, Germany), along with a Virus/Pathogen DSP midi kit and the cellfree500 protocol (Qiagen). The extracted RNA was annealed to a reverse primer specific to the gag gene of SIVmac239 (5’- CAC TAG GTG TCT CTG CAC TAT CTG TTT TG −3’), then reverse transcribed to obtain cDNA using SuperScript^™ III Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA) along with RNAse Out. The resulting cDNA was treated with RNAse H and combined with a custom 4x TaqMan^™ Gene Expression Master Mix (Thermo Fisher Scientific) containing primers and a fluorescently labeled hydrolysis probe specific for the gag gene of SIVmac239 (forward primer 5’- GTC TGC GTC ATC TGG TGC ATT C −3’, reverse primer 5’- CAC TAG GTG TCT CTG CAC TAT CTG TTT TG −3’, probe 5’- /56-FAM/CTT CCT CAG TGT GTT TCA CTT TCT CTT CTG CG/3BHQ_1/ −3’). The qPCR was performed using a StepOnePlus^™ or QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). To determine the SIV gag RNA copies per reaction, quantification cycle data and a serial dilution of a well-characterized custom RNA transcript containing a 730 bp sequence of the SIV gag gene were used. The mean RNA copies per mL were calculated by adjusting for the dilution factor. The limit of quantification (LOQ) was determined based on the exact binomial confidence boundaries and the percent detection of the target RNA concentration, which is approximately 62 copies per mL.

Cells from BAL, PBMC, and tissue collected at necropsy were stained for flow cytometric analysis. BAL and necropsy tissue were stained promptly upon collection, while cryopreserved PBMC were stained as a batch. Cells were seeded at 1 × 10⁶ cells/well in a 96-well plate in ELISpot media, then stimulated with either Mtb H37Rv whole cell lysate (WCL, 20 μL/mL; BEI Resources) or PDBu and ionomycin (P&I) for 6 hours at 37°C. After 2 hours of incubation, Golgi Plug (BD Biosciences, Cat no. 555029) was added at 8 μg/mL, incubated for an additional 4 hours, and washed twice with PBS. Live/Dead stain was added and incubated for 10 minutes at room temperature in the dark, then washed twice with PBS. A cocktail of the surface stain (Supplemental Table 1) supplemented with BD Horizon^™ Brilliant Stain buffer (BD Biosciences, Cat no. 566349) was added to each well (50 μL/well). After 20 minutes of incubation at 4°C, the cells were washed twice and permeabilized for intracellular staining (ICS) with BD Cytofix/Cytoperm^™ (BD Biosciences, Cat no. 554722) for 10 minutes at room temperature. For intranuclear staining (INS) of transcription factors (necropsy samples only), 1x True-Nuclear^™ (BD Biosciences, Cat. No. 424401) was added and incubated for 45 minutes. The cells were then washed twice with 1x perm wash for ICS or True-Nuclear^™ 1x Fix Perm wash buffer for INS. Staining was done with 50 μL/well of the ICS and INS antibody cocktails (Supplemental Table 1) for 20 and 30 minutes, respectively. This was followed by washing twice and resuspending the stained cells in 150 μL of FACs buffer. Samples were run on a BD Cytek Aurora and analyzed using FlowJo Software for Mac OS (v10.8.1). Detailed gating strategies are shown in Supplemental Figure 3a–d.

Plasma and 10–12 fold concentrated BALF were used to measure IgG and IgM titers to Mtb H37Rv WCL and lipoarabinomannan (LAM), as previously described (16). Briefly, 1 ug of WCL and LAM were coated onto 96-well Costar^® Assay plate (Nunc). Coated plates were blocked with PBS/10% FBS for two hours at room temperature and washed with PBS/0.05% Tween 20. Concentrated BALF or plasma 1:5 serially diluted (8 dilutions/sample) were then incubated for two hours at 37 °C and washed. Then 100 uL/well of HRP-conjugated goat anti-monkey IgG h+l (50 ng/ml; Bethyl Laboratories, Inc.), or IgM chain (0.1 ug/ml; Rockland Immunochemicals Inc.) was added and the plates incubated for 2 hours at room temperature. Plates were developed with 100 uL/well Ultra TMB^® substrate (Invitrogen) and fixed with 100 uL/well 2N sulfuric acid. Plates were read at 450 nm using a Spectramax190 microplate reader (Molecular Devices) and presented as endpoint titers for IgG or midpoint titers for IgM when samples did not titer to a cut-off. Endpoint titers are reported as the reciprocal of the last dilution with an optical density (OD) above the detection limit or twice the OD of an empty well.

All the data were transformed as log₁₀(x + 1). Longitudinal BAL, PBMC, and SIV viremia data were analyzed using a mixed model fit using the Restricted Maximum Likelihood (REML) with the Geisser-Greenhouse correction. Individual animals were considered a random variable. Fixed effect tests were used to evaluate the differences between different time points, vaccine route, or vaccine route versus time points. All results from pairwise comparisons of interest (comparisons among time points within vaccine group and comparison between IV BCG vs ID BCG within time points) were reported as uncorrected Fisher’s Least Significant Difference (LSD) p-values. Shapiro-Wilk test was used to test data for normality. Comparisons between IV BCG and ID BCG were made using the unpaired t-test for normally distributed data or the Mann-Whitney test for nonparametric data. Bivariate relationships were assessed using Pearson’s correlation coefficient (r). All tests conducted in the study were two-sided, and statistical significance was determined at a significance level of P < 0.05. P-values ranging from 0.05 to 0.10 were deemed indicative of a trend. GraphPad Prism 10 for MacOS (Version 10.0.2) was used for statistical analyses.

The study was designed to compare immune responses to IV or ID BCG in SIV-infected MCM without the confounding variable of Mtb challenge. Animals were infected with 3,000 TCID₅₀ SIVmac239 for 16 weeks and then randomized into two groups to be vaccinated with BCG by either the IV route (n=5) or the ID (n=6) route (Figure 1a). Animals in the IV BCG group were treated for two months with HRE, an anti-BCG drug combination, starting 3 weeks after vaccination to reduce the risk of BCGosis in these immunocompromised animals (19). We cultured blood samples at 1, 2, and 4 weeks, and BALF at 4 weeks after vaccination. Viable BCG was detected only in the blood of IV-vaccinated MCM one and two weeks post-BCG, and no culturable BCG was found in any BAL sample. Plasma viremia peaked within 2 weeks of SIV infection, then spontaneously fell to the viral set-point unique to each animal (Figure 1b). Animals with plasma viremia set points above 10⁵ copies/ml were considered to be “non-controllers”, indicating an inability to naturally control SIV replication (35–37). Two animals in each vaccination group were non-controllers by 14 weeks post-SIV infection (pre-BCG) (Figure 1b, non-controllers marked with star symbols). The median viremia at 14 weeks (pre-vaccination) was higher in the IV BCG group (Figure 1b, bottom panel). ID BCG vaccination did not significantly affect plasma viremia (Figure 1b, top panel), although SIV replication transiently increased in one animal 3 weeks after ID BCG. In contrast, IV BCG elicited a significant increase in plasma viremia in all animals within 2 weeks, which spontaneously returned to the approximate pre-BCG setpoints in most animals over the next 10 weeks (Figure 1b, bottom panel), as seen previously (19).

Immune responses in the airway are associated with IV BCG-elicited protection against Mtb infection (16, 34). Analysis of BAL samples taken 4 and 15 weeks after SIV infection (before vaccination) revealed no significant difference in total leukocyte and T cell counts from pre-SIV levels (Figure 2a, b). However, CD4+ T cells were significantly reduced and CD4+CD8α+ T cells were modestly reduced 4 weeks after SIV infection in both groups, but Vγ9+γδ were significantly reduced in ID BCG group (Figure 2b). All three cell populations returned to pre-SIV levels prior to BCG vaccination, except in animals with high viremia (Figure 2b). B cells and CD11b-CD11c+ (likely dendritic cells) were also significantly depleted in the ID BCG groups following SIV infection (Supplemental Figure 1a).

Following IV BCG vaccination, animals exhibited modest increases in leukocyte count and significant increases in T, B, and NK cells in the airways. These cell counts remained unchanged in the ID-vaccinated animals (Figure 2a, Supplemental Figure 1a). Notably, IV BCG, but not ID BCG, induced significant increases in CD8αβ+, CD8αα+, Vγ9+γδ, and Vγ9-γδ T cells and modest increases in CD4+, and CD4+CD8α+ T cells 4 weeks after vaccination (Figure 2b, and Table 1). These cell types returned to pre-BCG levels by 18 weeks post-vaccination, except Vγ9-γδ T cells which remained elevated (Figure 2a, 2b, and Table 1). There was no correlation between airway CD4+ T cell levels with SIV viremia in either group (Data not shown). Thus, we show IV BCG but not ID BCG elicits an influx of T cells into the airway of SIV-infected MCM.

We evaluated the ability of CD4+ and CD8αβ+ T cells in the airways to produce cytokines, focusing on those involved in Th1 (IFNγ, IL-2, TNF) and Th17 (IFNγ, IL-2, IL-17, TNF) responses due to their important roles in TB protection (20, 38, 39). Cytokine responses were assessed after ex vivo restimulation with Mtb WCL to measure mycobacterial-specific responses. The number of antigen-specific CD4+ T cells in the airways producing Th1- and Th17-related cytokines significantly increased 4 weeks after BCG, regardless of vaccine route (Figure 3a). Th1 CD4+ T cell responses remained significantly elevated up to 18 weeks in only the IV BCG group, while Th17 responses remained elevated in both IV- and ID-vaccinated groups (Figure 3a). In contrast, only ID BCG elicited significant Th1 and Th17 responses in airway CD8αβ+ T cells by 4 weeks post-vaccination. However, by 18 weeks after vaccination, Th1 and Th17 CD8αβ+ T cells were significantly increased in IV-vaccinated animals, while those responses were waning in the ID-vaccinated animals (Figure 3b). In summary, BCG vaccination by both the ID and IV routes elicited cytokine responses in airway CD4+ T cells by 4 weeks, but CD8αβ+ T cell responses were delayed following IV BCG. IV BCG resulted in more sustained CD4+ and CD8αβ+ T cell responses than ID BCG.

We next evaluated production of the cytotoxic effector granzyme B (GzmB) and granzyme K (GzmK) by CD8αβ+ T cells and NK cells recovered from the airways. Cytotoxic effectors may contribute to Mtb protection by lysing infected cells and/or intracellular killing of bacteria (40–42). SIV infection significantly increased both GzmB+ and GzmK+ CD8αβ+ T cells in both groups (Figure 4a), but did not affect the number of NK cells producing granzymes (Figure 4b). Following vaccination, IV BCG, but not ID BCG, elicited an additional and significant boost in CD8αβ+ T cells producing GzmB or GzmK, with GzmB+ cells remaining modestly elevated 18 weeks after vaccination. IV BCG also significantly increased the number of NK cells producing either GzmB or GzmK, but returned to pre-vaccination levels 18 weeks after BCG (Figure 4b and Table 1). These data demonstrate that IV BCG induces an early influx of cytotoxic effectors from CD8αβ+ T cells and NK in the airways.

PBMC were analyzed by flow cytometry to assess effects of BCG vaccination on circulating cell populations. Following SIV infection, the number of CD4+ and Vγ9+γδ T cells were significantly depleted in the IV BCG and ID BCG groups, respectively, prior to vaccination (Supplemental Figure 1c). Surprisingly, IV BCG vaccination induced a significant depletion in the number of NK, B cells, CD4+ and CD4+CD8α+ T cells by 4 weeks post-vaccination, while Vγ9+γδ T cells were depleted 4 weeks after ID BCG vaccination. These cell types returned to their pre-BCG levels by 18 weeks post-vaccination, although a significant decrease in NK cells remained at this late time point in ID-vaccinated animals (Supplemental Figure 1b, 1c).

We used ELISAs to measure WCL- and LAM-specific antibodies in plasma and BALF after BCG vaccination. IgG and IgM levels were associated with Mtb protection in IV BCG-vaccinated macaques (16, 19, 21, 43). Here, we observed a significant increase in plasma IgG specific for WCL and LAM following both ID and IV BCG by 4 weeks (Figure 5a, top panels). However, LAM-specific IgG in plasma returned to pre-BCG levels in both IV and ID BCG-vaccinated MCM by 18 weeks post-vaccination, while WCL-specific IgG returned to pre-BCG levels only in the ID BCG-vaccinated group (Figure 5a). Both IV and ID BCG elicited significant WCL-specific IgM in plasma, but only IV BCG induced significant LAM-specific IgM 4 weeks post-vaccination (Figure 5a, bottom panels). By 18 weeks after BCG, the IgM responses remained significant in the IV- compared to the ID BCG-vaccinated group (Figure 5a, Table 2).

In airways, IV BCG, but not ID BCG, elicited a significant increase in WCL-specific IgG and IgM while LAM-specific IgG and IgM responses were significant in both IV and ID BCG 4 weeks after vaccination (Figure 5b). However, both WCL- and LAM-specific antibodies declined to pre-BCG levels by 18 weeks after vaccination. (Figure 5b). Comparing each time point between IV and ID BCG showed that IV BCG elicited higher levels of LAM- and WCL-specific IgG in the airways at 4 weeks post-BCG than ID BCG. However, by 18 weeks after BCG vaccination, only LAM-specific IgG and IgM were significant in the airway of IV compared to ID BCG-vaccinated group (Table 2).

We interrogated immune responses across various tissues in BCG-vaccinated SIV+ MCM without the confounding variable of Mtb challenge, something that was not possible in our previous study in which all SIV+, BCG-vaccinated animals underwent Mtb challenge (19). Here, we conducted necropsies 5 months after vaccination, without Mtb challenge. Samples from lung, lymph nodes, and spleen were plated on agar to detect any viable BCG. We did observe one bacterial colony in a lung sample from a single IV-vaccinated animal in the HRE treatment group, which was confirmed by PCR to be BCG. Since no BCG was cultured from any other HRE-treated animal, and since this animal yielded just a single colony, the HRE regimen was largely effective.

We compared T cell responses in lungs, thoracic lymph nodes, and spleen five months after vaccination. Since T_RM cells were associated with IV BCG-elicited protection (16), we included anti-CD69 antibody, a canonical marker of T_RM cells and activation (33, 44) in our flow panels (Supplemental Table 1). To differentiate T_RM immune cells from those in circulation, fluorescently-labeled anti-CD45 antibody was administered just prior to euthanasia (16, 34). Thus, tissue-resident cells (anti-CD45 negative) can be differentiated from bloodborne cells (anti-CD45 positive). We designated the CD69+ivCD45- population to be T_RM cells (Fig. 6a), although this strategy may also mischaracterize some activated T cells as well. This analysis revealed significantly more T_RM cells in lungs of animals that received IV BCG compared to ID BCG (Figure 6a and 6b). Of the other T_RM cell subsets identified, there were significantly more CD8αβ+, CD8αα+, and γδ+ T cells in lung tissue of IV BCG-vaccinated animals (Figure 6c). CD4+ and CD4+CD8α+ T cell numbers did not differ significantly between the groups, most likely due to the low CD4+ T cell numbers in the MCM with higher SIV viral levels (Figure 6c). Indeed, CD4+ T cells in lung were modestly inversely correlated with SIV plasma viremia in both ID and IV BCG MCM (Data not shown). T_RM cell subset numbers in thoracic lymph node were similar in animals vaccinated with either ID or IV BCG (Supplemental Figure 2a), and the number of CD4+ and CD4+CD8α+ T cells in thoracic lymph nodes appeared independent of SIV viremia level (Supplemental Figure 2a).

We stimulated lung, thoracic lymph nodes, and spleen with WCL to assess production of cytokines associated with Th1 (IFNγ, IL-2, TNF) and Th17 (IFNγ, IL-2, IL-17, TNF) responses from CD4+ and CD8αβ+ T cells. The number of CD4+ T cells producing a combination of Th1- and Th17-associated cytokines in the lung were higher in animals vaccinated via IV BCG than ID BCG, but not in the thoracic lymph node (Figure 7a, 7b, and Supplemental Figure 2b, 2c). MCM with high SIV viremia had lower numbers of responding CD4+ T cells. CD8αβ+ T cells from all three sites produced some combination of Th1 and Th17 cytokines at similar levels in both IV- and ID-vaccinated animals (Figure 7a, 7b, and Supplemental Figure 2b, 2c).

We measured the number of CD8αβ+ T cells and NK cells producing granulysin, GzmB, GzmK, or perforin in lung, thoracic lymph nodes, and spleen. Our data revealed similar numbers of CD8αβ+ T cells and NK cells producing these cytotoxic effectors in lung and thoracic lymph nodes regardless of vaccination group (Supplemental Figure 2d–g). In the spleen, there were significantly more CD8αβ+ T cells producing GzmK and a trend towards more CD8αβ+ T cells producing GzmB in IV- than ID-vaccinated animals (Data not shown), but NK cells producing cytotoxic effector show no significant difference in number in all the tissues (Supplemental Figure 2d–g).

There was no apparent effect of plasma SIV viremia on granulysin-, GzmB-, GzmK-, or perforin-producing CD8αβ+ T cells in the lung. However, spleens of IV BCG animals with higher plasma SIV viremia had the lowest GzmB-, GzmK-, or perforin-producing NK cells (Data not shown).

The transcription factors Foxp3, RORγT, and T-bet are associated with regulatory T cells, Th17 cells, and Th1 cells, respectively (45, 46). The number of CD4+ and CD8αβ+ T cells expressing Foxp3, RORγT, or T-bet did not differ in lung, thoracic lymph node, or spleen between animals vaccinated by the IV or ID route (Data not shown).

In summary, our data show that IV BCG induces robust and sustained T helper and cytotoxic T cell response in the airways of SIV+ MCM. Notably, even 5 months after vaccination, IV BCG-vaccinated animals had higher numbers of T_RM cells in the lungs compared to ID BCG-vaccinated animals.