Early immune dysregulation in Mtb/SIV co-infection resists cART treatment at the single-cell level

Riti Sharan; Yi Zou; Zhao Lai; Korri Weldon; Todd Bradley; Xi Luo; Hulin Wu; Shabaana A Khader; Deepak Kaushal

doi:10.1016/j.isci.2026.116944

. 2026 Jul 27;29(8):116944. doi: 10.1016/j.isci.2026.116944

Early immune dysregulation in Mtb/SIV co-infection resists cART treatment at the single-cell level

Riti Sharan ^1,^∗, Yi Zou ², Zhao Lai ², Korri Weldon ², Todd Bradley ^3,^4,⁵, Xi Luo ⁶, Hulin Wu ⁶, Shabaana A Khader ⁷, Deepak Kaushal ^1,^8,^∗∗

PMCID: PMC13446293 PMID: 42565131

Summary

Using single-cell transcriptomics of bronchoalveolar lavage cells from Mtb/SIV co-infected rhesus macaques on cART, we reveal profound immune dysregulation during early SIV co-infection of latent tuberculosis. SIV induces a sharp decline in CD4⁺ T cells, NK, and NKT cells, with incomplete recovery of Mtb-specific T_H1 effector responses despite viral suppression. Instead, a persistent T_H17-skewed environment emerges, alongside sustained myeloid inflammation driven by Type I interferon signaling and pro-inflammatory regulators such as KLF6 and NFKB1. Ligand-receptor network analyses demonstrate expanded CD4⁺ T cell-macrophage crosstalk and loss of immune homeostasis that cART fails to fully restore. These findings expose how SIV remodels the pulmonary immune landscape to impair protective immunity against Mtb, providing a transcriptomic framework to explain TB reactivation in HIV infection. Our work highlights the urgent need for adjunctive immunotherapies to complement cART, aiming to rebalance immune responses and improve TB control in co-infected individuals.

Keywords: scRNA-seq, LTBI, HIV, NHP, Type I IFN, TB/SIV

Graphical abstract

Highlights

•
SIV co-infection disrupts T_H1 immunity during latent TB
•
cART fails to fully restore Mtb-specific CD4⁺ T cell function
•
Type I IFN signaling drives persistent myeloid activation
•
Paracrine networks remain altered despite viral suppression

Health sciences; Medicine; Immunology

Introduction

Mycobacterium tuberculosis¹/Human Immunodeficiency Virus (HIV) co-infected individuals are at a higher risk of progression to tuberculosis (TB) disease.²^,³^,⁴ HIV co-infection impairs Mtb-specific immunity, resulting in the potential reactivation of a latent TB infection (LTBI).⁵^,⁶^,⁷ TB is characterized by the development of specialized multicellular pathological structures—granulomas—in the lung and other extra-thoracic organs.⁸^,⁹ HIV co-infection causes reactivation of TB with the disruption of granulomas and their remodeling due to the impact of HIV on CD4⁺ T cells as well as chronic immune activation.⁹^,¹⁰^,¹¹ However, key knowledge gaps remain in our understanding of the underlying mechanisms involved in the pathogenesis of Mtb/HIV co-infection. A deeper insight into the impact of HIV on LTBI immune responses would likely assist in the development of preventive and therapeutic targets for Mtb/HIV co-infection. A focused study of the early cellular immune responses at the single-cell level during the first few weeks of SIV co-infection of LTBI in the rhesus macaque (RM) lung would allow a better understanding of how to protect and treat Mtb/HIV co-infected individuals. Single-cell RNA sequencing (scRNAseq) provides an excellent tool to examine immune changes at the transcript level.⁸^,¹²^,¹³^,¹⁴ Our group has pioneered the development of an RM model of Mtb/SIV co-infection, using a true aerosol-based Mtb infection modality, and using SIV as a surrogate for HIV. Additionally, we have successfully implemented and studied the impact of combinatorial antiretroviral therapy (cART) administration in Mtb/SIV co-infected RMs⁵^,⁶^,⁷^,¹⁵^,¹⁶^,¹⁷ and scRNAseq in the macaque model of TB.⁸^,¹³^,¹⁸ RMs develop asymptomatic Mtb infection akin to human LTBI and exhibit reactivation upon SIV co-infection.

We have previously shown that reactivation of LTBI in the RM model (1) is governed by very early events after SIV coinfection;⁶ (2) correlates with chronic immune activation caused by SIV rather than the mere depletion of CD4⁺ T cells;⁵^,¹⁰ (3) and that timing of cART is critical in mitigating virus-driven chronic immune activation.⁶ Although initiating cART during the acute phase of SIV co-infection modestly reduced chronic immune activation, it did not reduce the risk of LTBI reactivation as evidenced by persistence of defective functional Mtb-specific response.⁵ We have previously demonstrated that inhibiting immunosuppressive pathways such as indoleamine 2,3-dioxygenase (IDO) in Mtb/SIV co-infected, cART-treated RMs significantly reduced markers of chronic immune activation and enhanced the quality of immune responses.¹⁹ These findings suggest that cART alone is insufficient to fully resolve chronic immune activation, and consequently, the risk of TB reactivation or re-infection may persist despite treatment. The objective of this study was to characterize the early transcriptomic events underlying immune activation and TB reactivation during Mtb/SIV co-infection, and to determine whether cART alone is sufficient to mitigate these deficits. Toward this goal, we performed scRNA-seq on bronchoalveolar lavage (BAL) samples from Mtb/SIV co-infected and cART-treated RMs from our published cohort.⁶ BAL serves as a valuable surrogate for lung tissue in nonhuman primate (NHP) studies, as it allows for minimally invasive, longitudinal sampling of immune cells and molecular signatures from the lower respiratory tract. This approach is particularly important given the ethical and logistical challenges of euthanizing animals at multiple time points, as well as the high cost associated with increasing group sizes in macaque studies.

Overall, our findings reveal that SIV co-infection disrupts both innate and adaptive immune response-associated gene expression in the airways of Mtb-infected RMs. While cART partially restores CD4⁺ T cells, it fails to reconstitute the T_H1 response and does not fully control SIV-induced immune activation. Persistent myeloid-driven Type I IFN signaling and chronic inflammation highlight ongoing immune dysfunction despite viral suppression. This study defines critical immune subsets and gene expression changes impacted by HIV co-infection in latent TB, offering valuable insights into Mtb/HIV pathogenesis and potential targets for therapeutic intervention.

Results

Incomplete immune reconstitution by cART in Mtb/SIV co-infected RMs

In the previously published study, to assess the impact of SIV on LTBI infection, we infected four Indian-origin RMs with a low dose CDC1551 (∼10 CFU deposited in the lungs) via aerosol, to assess the impact of SIV on LTBI infection (Figure 1A)⁵^,⁶^,⁷ (Table S1). Asymptomatic LTBI was characterized by a bacterial burden of less than 2 log₁₀ CFU of Mtb in BAL, no significant changes in body temperature or weight, and normal C-reactive protein levels post-Mtb-infection.⁶ Latently infected macaques were subsequently co-infected with SIV (300 TCID₅₀ SIVmac₂₃₉) via the intravenous route at week 9 post Mtb-infection. We used a physiologically higher dose of SIV through the IV route to facilitate LTBI reactivation, shorten the duration of the study in BSL3 containment, and reduce the number of animals required, while still ensuring adequate statistical power. Co-infected RMs were treated with cART starting at 2 weeks post-SIV co-infection for a total of 9 weeks. BAL cells were collected at wk5 (represents the asymptomatic phase of Mtb infection), wk11 (represents 2 weeks post-SIV co-infection), wk15 (represents post-SIV co-infection and 4 weeks of cART treatment) and necropsy (study endpoint after 9 weeks of cART treatment) (Figure 1A). Co-infected RMs received a drug regimen consisting of 20 mg/kg of nucleoside reverse transcriptase inhibitors (NRTIs) (R)-9-(2-phosphonylmethoxypropyl) adenine (tenofovir disoproxil fumarate (TDF), Gilead Sciences), 30 mg/kg of 2′,3′-dideoxy-5-fluoro-3′-thiacytidine (FTC, emtricitabine, Gilead Sciences), and 2.5 mg/mL of the integrase inhibitor dolutegravir (DTG, ViiV Healthcare). The combination of TDF, FTC, and DTG used in NHPs corresponds to the clinically used regimen of Truvada plus Tivicay.

We previously demonstrated that the pathological examination of lungs from Mtb/SIV co-infected, cART-treated RMs still harbored active, caseous granulomas at necropsy (Figure 1B). Certain lung sections also exhibited increased lymphoid tissue, occasionally accompanied by basophils and eosinophils. Additionally, the liver presented multifocal nodular aggregates of lymphocytes and macrophages, along with a marginal increase in the size of hilar and bronchial lymph nodes.⁶ As previously reported, three of the four cART treated RMs continued to harbor Mtb (1 × 10² CFU/g) at necropsy.⁶ Immunophenotyping of BAL samples revealed a marked decrease in CD4⁺ T cells two weeks after SIV co-infection (wk11), with only partial restoration observed following cART by week 20 (Figure 1C). Reanalysis of earlier data revealed that this partial recovery was associated with an increase in the percentage of CD4⁺ central memory (T_CM) T cells following cART (Figure 1D). However, SIV co-infection led to a marked reduction in CD4⁺ effector memory (T_EM) T cells at wks 11 and 15, a deficit that cART failed to reverse (Figure 1E), relative to levels seen during LTBI.⁶

Ex vivo stimulation of BAL cells with Mtb-specific ESAT-6/CFP-10 peptides showed a drastic decline in CD4⁺IFN-γ⁺ (Figure 1F) and CD4⁺TNF-α⁺ T cells (Figure 1G) at wk11, which cART could not restore to LTBI levels. In contrast, the proportion of Mtb-specific CD4⁺IL-17⁺ T cells increased at wks 11, 15, and 20 post-cART initiation (Figure 1H). These results indicate that although cART partially restores CD4⁺ T cell populations in the BAL, it fails to normalize the Mtb-specific T_H1 response or rebalance the T_H1/T_H17 axis in Mtb/SIV co-infected RMs. The resulting cytokine profile suggests a T_H17-skewed immune response, which may contribute to either protective immunity or immunopathology, depending on the local microenvironment. These findings highlight the limitations of cART in fully restoring protective immune responses during Mtb/SIV co-infection and underscore the need for adjunctive therapies that can rebalance the T_H1/T_H17 axis and enhance immune control of Mtb.

Single-cell transcriptomic profiling identifies distinct immune and epithelial cell populations

To characterize the pulmonary transcriptomic landscape in Mtb/SIV co-infected, cART-treated macaques, we performed scRNA-seq using the 10× Genomics platform on BAL cells collected from four macaques at four time points: wk5 (asymptomatic phase of Mtb infection), wk11 (2 weeks post-SIV co-infection), wk15 (4 weeks of cART treatment following SIV co-infection), and necropsy (study endpoint after 9 weeks of cART treatment). In total, 16 BAL samples were analyzed. The use of four macaques in this scRNA-seq study is justified by the high intra-animal data richness inherent to single-cell approaches, where each animal yields tens of thousands of individual cell profiles, making the cell, not the animal, the primary unit of analysis. This allows for robust statistical power despite a small cohort size. Furthermore, scRNA-seq enables detailed resolution of cellular heterogeneity, capturing diverse cell types, states, and rare populations that bulk sequencing would obscure. Importantly, the goal of this study is to identify shared immunological features and conserved mechanisms of early immune dysregulation during Mtb/SIV co-infection, which can be effectively achieved by analyzing common patterns across a small number of animals. All samples passed quality control in terms of cell quality (fraction reads in cells) and sequencing (Table S2).

To characterize the cellular composition of the analyzed samples, we performed unsupervised clustering and UMAP visualization of scRNA-seq data (Figure 2A). A total of 13 transcriptionally distinct clusters were identified, corresponding to major immune and epithelial cell populations. Cluster identification was performed using unbiased automated software, Azimuth_ALV1L1, Azimuth_ALV1L2, Azimuth_ALV2L4, and Clustermole. The genes were matched to expression clusters using the Human Protein Atlas (https://www.proteinatlas.org/about) and GeneCards (The Human Genome Database: https://www.genecards.org). We further confirmed the cluster annotation using the published signature gene list⁸ (Figure 2B and Table S3). The expression profiles of key marker genes across clusters are shown in the dot plot (Figure 2B). Monocyte and macrophage clusters (C1–C3) expressed high levels of CD14 and CD68, while dendritic cell clusters (C4–C5) showed enrichment for CD1C and CLEC9A. T cell populations (C7–C9) were characterized by strong expression of CD3D, CD7, and IFNG, with CD8A expression distinguishing cytotoxic T cells (C8). NK and NKT cell clusters (C10–C11) exhibited elevated KLRD1 expression, consistent with their cytotoxic phenotype. B cells (C12) were marked by CD19 expression, and ciliated epithelial cells (C13) displayed epithelial-specific markers such as FOXJ1. Feature plots further validated cluster-specific marker expression patterns (Figure 2C). For example, CD3D, CD8A, and IFNG expression localized to T cell clusters, KLRD1 to NK/NKT clusters, and CD14 and CD1C to myeloid lineages. MS4A2 expression confirmed the identity of mast cells (C6), while CDK1 expression was enriched in proliferating macrophages (C3), consistent with their cycling state. The overall distribution and identity of these major immune and epithelial populations were conserved across all time points, suggesting stable immune cell representation (Figure S1A). Quality control metrics indicated consistent sequencing depth and transcript capture across cell clusters (Figure S1B). Violin plots of nFeature_RNA, nCount_RNA, and percent.mt demonstrated comparable gene and UMI counts among clusters, with low mitochondrial content, confirming high-quality single-cell transcriptomes suitable for downstream analyses. The overall UMAP structure was conserved among RMs (Figure S2). While the relative abundance of specific clusters varied slightly between RMs, the overall transcriptional architecture remained stable. Together, these results demonstrate the successful resolution of major immune and epithelial subsets within the dataset and highlight distinct transcriptional programs associated with each cell type.

Single-cell transcriptomic profiling identifies distinct immune and epithelial cell populations

(A) UMAP visualization of integrated scRNA-seq profiles from BAL cells collected from rhesus macaques (n = 4) at week 5 post-*Mtb* infection (LTBI), week 11 following *Mtb*/SIV co-infection, week 15 after 4 weeks of cART, and at necropsy after 9 weeks of cART. Distinct transcriptional clusters are annotated and colored by major immune cell identities, including macrophages (C1), monocytes (C2), proliferating macrophages (C3), mDC (C4) pDC subsets (C5), mast cells (C6), CD4 T cells (C7), CD8 T cells (C8), T helper cells (C9), NKT (C10), NK cells (C11), B cells (C12), and ciliated epithelial cells (C13).

(B) Dot plot displays canonical marker gene expression across all identified clusters. Dot size indicates the proportion of cells expressing each gene, and dot color denotes scaled average expression, supporting the assignment of myeloid, lymphoid, and epithelial cell identities.

(C) Feature plots show UMAP expression patterns of representative lineage-defining genes, including markers for T cells (CD3D, IFNG, CD8A), NK cells (CD7, KLRD1), B cells (CD19), myeloid subsets (CD14, CD1C), mast cells (MS4A2), and proliferating cells (CDK1). These data reveal the cellular heterogeneity of BAL during *Mtb*/SIV co-infection and its immunologic remodeling following cART.

SIV co-infection leads to the depletion of key lymphocyte populations that are not fully restored by cART

To evaluate how SIV co-infection and subsequent cART affect immune cell composition during Mtb infection, we analyzed changes in major lymphocyte clusters identified by scRNA-seq across the course of infection (Figures 3A–3E). Cell counts within each cluster were compared among RMs at wk5, wk11, wk15, and at necropsy. At the onset of SIV co-infection (wk11), there was a pronounced decrease in several lymphocyte populations, including CD4⁺ T cells (Figure 3A), NK cells (Figure 3C), and NKT cells (Figure 3D), relative to wk5. While not significant, this depletion was most notable in CD4⁺ T cells and NK cells, consistent with SIV-driven CD4⁺ T cell suppression. Following the initiation of cART (wk11), partial recovery of these cell subsets was observed; however, frequencies did not return to the levels seen in Mtb-only infection (wk5) at necropsy. In contrast, CD8⁺ T cell (Figure 3B) and B cell cluster (Figure 3E) abundances remained relatively stable throughout infection. These findings indicate that SIV co-infection induces a transient but significant loss of critical effector cell subsets within the Mtb-infected lung microenvironment. Despite cART treatment, immune reconstitution remained incomplete, suggesting persistent immune dysregulation and impaired cellular recovery in the context of Mtb/SIV co-infection.

SIV co-infection leads to depletion of key lymphocyte populations that are not fully restored by cART

Bar plots display cell counts for five major lymphocyte clusters identified by scRNA-seq: CD4⁺ T cells (A), CD8⁺ T cells (B), NK cells (C), NKT cells (D), and B cells (E). Counts are shown at weeks 5, 11, and 15 post-*Mtb* infection and at necropsy (N) (n = 4). All pairwise analyses in 3A-E have been performed using paired Wilcoxon matched-pairs signed-rank tests. Data are shown as median with interquartile range (IQR). ∗p < 0.05; ∗∗p < 0.01; and ∗∗∗p < 0.001.

(F) To assess temporal changes in immune transcriptional programs during infection, we evaluated the expression of a predefined T cell-associated gene module across immune cell clusters and time points. The key genes are organized into the following module: T cell module -TBX21, IFNG, TNF, LTA, IL18RAP, BHLHE40, IL2, CCR6, RORA, RORC, IRF4, STAT3, IL23R, IL22, IL21, IL21R, IL17A, IL4, IL5, IL6, IL10, IL13, KLF4, TRAC, KLRD1, CCL5, GZMB, GZMH, CTLA4, ICOS, LAG3, NCAM1, KLRK1, NCR1, PRF1, GZMA, GZMB. For each cluster and time point, average normalized expression values were calculated, and differences between time points were assessed using paired Wilcoxon signed-rank tests with FDR correction. Heatmap shows the FDR-adjusted p values, visualized as −log10(FDR-adjusted p values).

(G and H) Gene Ontology (GO) enrichment analysis for down-regulated (wk 11 vs. wk 15) (G) and up-regulated (wk 11 vs. wk 5) (H) genes in CD4⁺ T cells (C7) using the org.Mmu.e.g.,.db annotation and clusterProfiler. The top 20 enriched biological process terms are shown as dot plots. Significance thresholds were FDR-adjusted p < 0.05 and Storey’s q < 0.2. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

To assess temporal changes in immune transcriptional programs during infection, we evaluated the expression of a predefined T cell-associated gene module across immune cell clusters and time points (Figure 3F). This module included genes involved in T cell activation, signaling, and effector function (IFNG, IL2, IL17A, CCR6, RORA, RORC, IRF4, and STAT3). For each cluster and time point, average normalized expression values were calculated, and differences between time points were assessed using paired Wilcoxon signed-rank tests with FDR correction. Heatmap in Figure 3F illustrates the FDR-adjusted p values, visualized as −log10(FDR-adjusted p values). The heatmap is thresholded at FDR <0.05. Several immune cell clusters exhibited significant temporal changes in module expression during the course of infection. Significant differences were detected across myeloid populations (clusters 1–4), as well as lymphoid immune clusters (clusters 6–12), indicating dynamic transcriptional responses over time. Although this module is enriched for genes commonly associated with T cell function, many of these genes participate in broader immune activation pathways and can also be expressed by myeloid populations during inflammatory responses. Thus, the detection of this module across multiple immune cell clusters likely reflects shared immune activation and cytokine-driven transcriptional programs within the lung microenvironment during SIV co-infection and subsequent cART treatment. Although the module was derived from genes typically associated with T cell function, we intentionally evaluated its expression across all immune clusters to determine whether T cell-associated transcriptional programs were altered more broadly in the lung microenvironment. This approach allows the detection of coordinated immune signaling changes that may also appear in other cell populations through intercellular signaling, activation states, or shared transcriptional programs. Among the three modules, only the T cell module retained significant differences after FDR correction, indicating dynamic regulation of T cell-associated transcriptional programs over time. In contrast, the myeloid and immune activation modules did not maintain significance following adjustment, suggesting relatively stable expression profiles in these compartments. Collectively, these findings highlight T cell-driven transcriptional remodeling as a key feature of the temporal immune landscape from early to late time points and necropsy.

To explore the functional pathways underlying transcriptional changes associated with SIV co-infection during LTBI, we performed Gene Ontology (GO) enrichment analysis using the org.Mmu.e.g.,.db RM annotation database (Bioconductor release 3.21) and the clusterProfiler R package. GO enrichment analyses were performed on differentially expressed genes identified within the CD4 T cell cluster (cluster 7). Significance thresholds were set at FDR-adjusted p < 0.05 and Storey’s q < 0.2. The top 20 significantly enriched biological process terms were visualized for both up- and down-regulated genes using dot plots (Figures 3G and 3H). Figure 3G shows GO enrichment of downregulated genes in CD4 T cells between wk11 and wk15, corresponding to the period following the initiation of cART. Among the down-regulated genes (Figure 3G), enriched GO terms were largely associated with the positive regulation of immune responses, including cell activation, leukocyte activation, lymphocyte activation, and B cell activation. The suppression of these pathways suggests that SIV co-infection dampens adaptive immune signaling during LTBI, potentially reflecting viral-mediated immune modulation or exhaustion of critical immune subsets. Figure 3H shows GO enrichment of upregulated genes in CD4 T cells between wk5 and wk11, representing transcriptional changes associated with the onset of SIV co-infection. In contrast, upregulated genes (Figure 3H) were strongly enriched for pathways linked to immune cell activation, differentiation, and signaling. Prominent terms included T cell activation, B cell activation, leukocyte differentiation, immune response-activating signaling pathway, and actin cytoskeleton organization. These analyses identify biological pathways underlying the transcriptional remodeling observed in CD4 T cells during SIV co-infection and subsequent cART treatment. Together, these data suggest that SIV co-infection of LTBI drives a shift in immune transcriptional profiles characterized by decreased antiviral and innate immune signaling alongside enhanced activation and differentiation of lymphocyte subsets. This pattern reflects an overall increase in the activation status of lymphoid cells during SIV and LTBI co-infection, consistent with heightened immune stimulation and remodeling of the host response.

Quantitative profiling reveals differential distribution of inflammatory and effector transcripts across lymphoid and myeloid compartments

To examine cell type-specific transcriptional features, we quantified the expression of key lineage-associated and effector genes across lymphoid, myeloid, and epithelial clusters (Figure S3). Transcripts associated with type 1 immune responses, including TBX21, IFNG, TNF, LTA, and IL2, were detected at relatively low levels across lymphoid populations, including CD4⁺ T (C7), CD8⁺ T (C8), T helper (C9), and NKT (C10) clusters (Figure S3A). Fewer than ∼20% of lymphoid cells expressed TBX21 or IFNG, with mean expression levels near baseline (average log fold change ≈ 0–0.3). Although TBX21 is classically associated with TH1 differentiation in CD4⁺ T cells, expression was also observed in other lymphoid populations, consistent with its broader role in inflammatory transcriptional programs across immune cell types. In contrast, myeloid populations, particularly macrophages (C1), monocytes (C2), and mDCs (C4), exhibited stronger expression of inflammatory mediators. TNF and IL18RAP transcripts were detected in approximately 40–60% of these cells with average expression intensities ∼1–1.5 log fold higher than those observed in lymphoid subsets, indicating that inflammatory signaling signatures were more prominent within myeloid compartments. Epithelial cells (C13) displayed minimal expression across these genes. Figure S3B summarizes the distribution of genes associated with T_H17-related transcriptional pathways, including RORC, STAT3, IL17A–F, IL21, and IL22. Expression of the transcriptional regulators RORC and STAT3 was most evident in T helper (C9) and CD4⁺ T (C7) clusters, where approximately 50–80% of cells expressed these transcripts (mean ≈1.2–1.8 log fold). In contrast, cytokine genes such as IL17A, IL17F, and IL22 were detected in a smaller fraction of cells and were largely restricted to lymphoid populations, with minimal expression in myeloid or epithelial compartments. Expression of CCR6 and IRF4, which are associated with T_H17-related transcriptional programs in CD4⁺ T cells, was also observed within these lymphoid subsets. Effector and activation-associated transcripts (GZMB, GZMH, KLRD1, CCL5, ICOS, CTLA4, LAG3, and CD38) are shown in Figure S3C. Cytotoxic gene signatures were most prominent in CD8⁺ T (C8), NK (C11), and NKT (C10) clusters, where a high proportion of cells expressed GZMB and CCL5. In contrast, CD4⁺ T (C7) and T helper (C9) clusters displayed the expression of activation and immune regulatory markers including ICOS, CTLA4, and LAG3 in a subset of cells. Collectively, these analyses highlight distinct transcriptional features across immune cell populations, with stronger inflammatory signatures in myeloid compartments and cytotoxic programs enriched within NK and CD8⁺ T cell populations.

To assess how SIV co-infection alters CD4⁺ T cell function during Mtb infection, we analyzed transcriptional changes within the CD4⁺ T cell cluster isolated from BAL across disease stages (Figure S4). Comparison of wk5 (Mtb infection alone) and wk11 (Mtb/SIV co-infection) revealed a pronounced suppression of T_H1-associated gene expression, including IFNG, STAT1, TNFAIP3, and IL2RA (average log₂FC range: −0.7 to −2.5; Figures S4A and S4D). This reduction corresponded with diminished expression of key T_H1 transcriptional regulators (TBX21, STAT1) and effector cytokines (IFNG, IL2), consistent with impaired classical T_H1 differentiation following SIV infection. In contrast, while T_H17-related genes (IL21, IL22RA1, RORC, IL23R) were also downregulated during acute SIV co-infection (wk11) (average log₂FC range: −0.5 to −1.3), several of these transcripts showed partial recovery at wk15 following the initiation of cART (log₂FC range: +0.5 to +1.5; Figures S4B and S4D). This suggests a shift in the T_H1/T_H17 balance, characterized by persistent T_H1 suppression alongside partial T_H17 reactivation during treatment. Despite this transcriptional rebound, markers of immune activation and exhaustion remained elevated at the necropsy time point, including CD38, LAG3, ICOS, CTLA4, and chemokine genes such as CXCL10 and CXCL13 (log₂FC range: +1.0 to +2.5; Figures S4C and S4D). Together, these data reveal that SIV co-infection disrupts protective T_H1 immunity and drives a persistent immunological imbalance characterized by incomplete T_H17 recovery and sustained immune activation, even under cART.

In the CD4⁺ T cell cluster, differential expression analysis between wk5 (Mtb infection alone) and wk11 (Mtb/SIV co-infection) revealed extensive transcriptional reprogramming, with 412 genes upregulated (Figure S5A and Table S4) and 368 genes downregulated at wk11 (adjusted p < 0.05, log₂FC ≥ 0.25) (Figure S5B and Table S4). GO enrichment of upregulated transcripts highlighted strong activation of adaptive immune processes, including T cell activation (GO:0042110; FDR = 1.2 × 10⁻⁵), lymphocyte differentiation (GO:0030098; FDR = 3.4 × 10⁻⁴), and positive regulation of immune signaling (GO:0002684; FDR = 5.7 × 10⁻⁴), alongside cytoskeletal remodeling pathways such as actin filament organization (GO:0030036; FDR = 2.1 × 10⁻³) (Table S4). Upregulated genes driving these enrichments included TGFB1, CD28, TNFAIP3, and CD74, consistent with enhanced activation and structural polarization of CD4⁺ T cells during co-infection. Conversely, downregulated genes were significantly enriched for interferon- and virus-response pathways, including response to type I interferon (GO:0034340; FDR = 8.6 × 10⁻⁶) and defense response to virus (GO:0051607; FDR = 4.3 × 10⁻⁵), as well as for processes associated with protein folding (GO:0006457; FDR = 1.9 × 10⁻³) and mRNA splicing (GO:0000398; FDR = 3.2 × 10⁻³) (Table S4). Representative downregulated genes included MX1, ISG15, IFIT1, IRF7, and HSPA5. Collectively, these results indicate that during Mtb/SIV co-infection, CD4⁺ T cells undergo a quantitative shift from an interferon-driven antiviral program toward an adaptive immune activation state, accompanied by cytoskeletal reorganization.

Table S4. Gene Ontology (GO) enrichment analysis of differentially expressed genes from BAL scRNA-seq in macrophage cluster (C1) comparing wk5 LTBI to wk11 post-SIV co-infection (n = 4)

Enriched biological process, cellular component, and molecular function terms are shown with associated statistics. GO enrichment was performed using over-representation analysis with hypergeometric testing, and p-values were adjusted using the Benjamini-Hochberg false discovery rate (FDR). The results indicate a shift from antiviral/interferon programs at wk 5 toward lymphocyte activation and cytoskeletal remodeling at wk11.

mmc2.xlsx^{(36.7KB, xlsx)}

SIV co-infection alters transcriptional programs associated with myeloid activation and interferon signaling

scRNA-seq of BAL samples from Mtb and SIV co-infected RMs revealed time-dependent shifts in myeloid and innate immune cell populations across infection and cART (Figures 4A–4E). Macrophages and monocytes play a crucial role in controlling Mtb infection and serve as key drivers of pathogenesis in HIV infection.²⁰^,²¹ Prior to SIV co-infection (wk5), macrophage (C1) (Figure 4A) and monocyte (C2) (Figure 4B) clusters were readily detectable. Following SIV co-infection (wk11), macrophage and monocyte cluster frequencies increased, indicating a robust myeloid expansion during acute viral co-infection. These populations declined by wk15, after initiation of cART, and remained lower at necropsy. Myeloid dendritic cell (mDC) cluster (Figure 4C) frequencies remained stable throughout infection and treatment, whereas plasmacytoid dendritic cells (pDCs) (Figure 4D) showed a reduction from wk 5 to wk 11 and further decrease at necropsy. Mast cell numbers (Figure 4E) were unchanged between wks 5–15 but declined notably by necropsy. These data indicate that Mtb/SIV co-infection transiently amplifies macrophage and monocyte populations in the BAL, while pDC and mast cell populations progressively diminish during cART, reflecting a reorganization of the local innate immune landscape over the course of co-infection and therapy.

To define the transcriptional dynamics underlying myeloid activation during Mtb/SIV co-infection, we analyzed the expression of 14 key transcription factors across six major BAL cell clusters (C1–C6) over the study period (Figure 4F). The M1 macrophage-associated transcription factor NFKB2 showed consistently low expression (average expression <0.2, detected in <25% of cells) in macrophage (C1) and monocyte (C2) clusters throughout infection and treatment. In contrast, NFKB1 and KLF6 were among the most broadly expressed regulators (detected in >70% of cells in C2 (monocytes), C4 (mDCs), and C6 (mast cells)), with sustained high expression levels (average expression >1.0) across all time points, supporting persistent inflammatory programming in monocytes, mDCs, and mast cells. Transcription factors associated with immune regulation and M2 macrophage polarization—including GATA3 and IRF4—displayed minimal expression across clusters (average expression <10% of cells), except for modest IRF4 induction in the pDC cluster (C5; average expression ≈0.5, ∼40% of cells). Macrophage (C1) and mast cell (C6) clusters exhibited the lowest IRF4 expression (average expression < −1.0), consistent with a reduced M2-like signature under co-infection conditions. Following SIV co-infection (wk11), we observed pronounced induction of IRF9 and STAT1 across all myeloid clusters (average expression +1.5 to +2.0, >75% of cells), relative to the Mtb-only phase (wk5). This increase was most evident in macrophages (C1), where STAT1, STAT3, and IRF9 were co-expressed in >80% of cells, indicating a robust activation of Type I interferon signaling. After cART initiation (wk15 and necropsy), expression of these interferon-associated transcription factors decreased toward baseline (average expression <0.5), suggesting partial restoration of transcriptional homeostasis. Together, these data indicate that SIV co-infection transiently amplifies STAT1/STAT3/IRF9-driven Type I IFN responses across myeloid lineages while suppressing alternative (IRF4/GATA3-mediated) polarization programs. The resolution of this signature with cART underscores the dynamic interplay between viral replication, immune activation, and therapeutic immune modulation in the Mtb/SIV co-infected lung.

GO enrichment of downregulated genes in macrophages (C1) at wk11 (Mtb/SIV co-infection) versus wk5 (Mtb infection alone) revealed significant suppression of immune and antiviral processes (Figure 4G). Key pathways included defense response to other organisms, innate immune response, and regulation of cytokine production (adjusted p < 2 × 10⁻⁴; GeneRatio 0.06–0.08). Processes related to response to virus and viral genome replication were also reduced (GeneRatio ≈0.03–0.04; adjusted p = 4–5 × 10⁻⁴), with 40–80 genes contributing per term. Downregulation of ubiquitin- and proteasome-dependent protein catabolic processes (GeneRatio ≈0.05; adjusted p < 3 × 10⁻⁴) suggested impaired protein turnover. Collectively, these data indicate that SIV co-infection dampens macrophage-driven antiviral and cytokine signaling pathways within the lung, consistent with early immune suppression during co-infection. GO enrichment analysis revealed significant upregulation of immune-related biological processes between wk15 (Mtb/SIV coinfection with 4 weeks of cART) and necropsy (end of cART) (Figure 4H). The top enriched terms included positive regulation of immune system processes (48 genes, p.adjust = 0.0018), cell activation (43 genes, p.adjust = 0.0021), and leukocyte activation (36 genes, p.adjust = 0.0035). Additional pathways, such as T cell activation (33 genes, p.adjust = 0.0037) and adaptive immune response (24 genes, p.adjust = 0.0048), were also significantly upregulated. Processes linked to immune regulation and hematopoietic recovery, regulation of immune response, and response to oxygen-containing compound (46 genes, p.adjust = 0.0060) were enriched as well. Broader immune activation was indicated by the enrichment of defense response to symbiont (GO:0140546), innate immune response (GO:0045087), and response to virus (GO:0009615), involving interferon-stimulated and inflammatory genes (IFIT1, OAS1, IRF7, TNFAIP3, TLR2) at wk5 and necropsy (Figure S6A and Table S5). Metabolic pathways, including hexose biosynthetic process (GO:0019319) and gluconeogenesis (GO:0006094), were also enriched (Figure S6A and Table S5). GO enrichment analysis of monocytes (C2) at wk5 and necropsy revealed strong upregulation of antigen processing and presentation pathways, including “antigen processing and presentation of peptide antigen via MHC class I/Ib” (GO:0002428, GO:0002476; adjusted p = 0.0018) (Figure S6B and Table S6). These processes were driven primarily by MAMU-A, MAMU-E, and MAMU-A3, indicating increased antigen-presenting activity. Additional enriched biological processes involved T cell-mediated cytotoxicity and leukocyte activation, such as “positive regulation of T cell-mediated immunity” (GO:0002711, adjusted p = 0.0045) and “cell killing” (GO:0001906, adjusted p = 0.0106). Enriched cellular component terms included “endocytic vesicle membrane” and “lumenal side of endoplasmic reticulum membrane,” consistent with increased vesicular trafficking and peptide loading (Figure S6B and Table S6). Together, these data indicate that monocytes at wk5 up to necropsy exhibit heightened antigen-presenting capacity. Overall, these results indicate sustained activation of immune signaling and adaptive immune pathways from wk15 to necropsy despite ongoing cART.

Table S5. Gene Ontology (GO) enrichment analysis of differentially expressed genes in the macrophage cluster (C1), comparing wk 5 LTBI to necropsy following cART (n = 4)

Enriched biological process, cellular component, and molecular function terms are shown with corresponding statistics. Analysis was performed using hypergeometric testing with Benjamini-Hochberg FDR correction. The enrichment profile highlights pathways that are upregulated post-SIV co-infection and cART, including antigen processing and presentation, innate immune activation, cytokine-responsive programs, metabolic signaling, and cytoskeletal remodeling in macrophages at necropsy relative to LTBI.

mmc3.xlsx^{(19.8KB, xlsx)}

Table S6. Gene Ontology (GO) enrichment analysis of differentially expressed genes in the monocyte cluster comparing wk 5 LTBI to necropsy following cART treatment (n = 4)

Enriched biological process, cellular component, and molecular function terms are shown with corresponding statistics. Analysis was performed using hypergeometric testing with Benjamini-Hochberg FDR correction. The enrichment profile highlights pathways upregulated post-SIV co-infection and cART, including antigen processing and presentation, T cell-mediated immunity and cytotoxicity, leukocyte-mediated immune regulation, vesicle trafficking, cytoskeletal remodeling, ribosomal function, and structural molecule activity.

mmc4.xlsx^{(17.9KB, xlsx)}

We performed GO GSEA and visualized the top statistically significant up- and downregulated (up to six each) gene sets across Biological Process, Molecular Function, and Cellular Component categories, with significance defined as FDR-adjusted p < 0.05. Among all clusters analyzed, the mast cell cluster (C6) exhibited the most significant transcriptional changes following SIV co-infection at wk15 and necropsy (Figure 4I). GSEA revealed strong enrichment of platelet-associated pathways, including platelet aggregation (NES = 2.4, adjusted p < 0.01), platelet activation (NES = 1.9, adjusted p < 0.05), and regulation of platelet activation (NES = 1.8, adjusted p < 0.05) (Figure 4I). In contrast, immune-regulatory processes such as regulation of immune response (NES = −2.1, adjusted p < 0.05) and positive regulation of immune system process (NES = −2.0, adjusted p < 0.05) were negatively enriched. Mast cells play a pivotal role in early Mtb responses through cytokine secretion, granuloma organization, and recruitment of innate immune cells²²; thus, their shift toward a pro-activation and platelet-interactive phenotype suggests altered inflammatory dynamics during SIV co-infection. Collectively, these findings indicate that mast cells are potential players in persistent immune activation in the lung, amplifying local inflammatory signaling despite effective viral suppression by cART.

SIV co-infection induces persistent remodeling of paracrine signaling networks during Mtb infection

To characterize how paracrine signaling between immune cell subsets changes during low-dose Mtb infection, Mtb/SIV co-infection, and following cART, we analyzed ligand-receptor (LR) interaction networks between B cells, CD4⁺ T cells, CD8⁺ T cells, and macrophages across three timepoints: wk5 (LTBI phase), wk11 (Mtb/SIV co-infection), and at necropsy following cART. At wk5 post-Mtb infection, a total of 38 LR interactions were detected (Figures 5A and S7). The majority of these occurred between CD4⁺ T cells and macrophages (48%) and between macrophages and CD4⁺ T cells (32%). Prominent interactions included LTB-CD40, TGFβ1-ACVRL1, TNFSF13B-TNFRSF13C, and HSP90B1-TLR7, suggesting a balanced paracrine environment characterized by immune regulation and low-level inflammation consistent with controlled latent-like infection. At wk11 (Mtb/SIV co-infection), the total number of LR interactions increased by 1.7-fold (to 65 pairs), indicating broad remodeling of the intercellular signaling landscape (Figures 5B and S8). The strongest expansion was observed in the CD4⁺ T cell → macrophage axis (accounting for 55% of all interactions), coinciding with a reduction in B cell-mediated signaling. Several new LR pairs emerged, including HSP90AA1-CFTR, GNA2-ADORA1, SEMA4A-ITGA1, and PTMA-VIPR1, reflecting enhanced inflammatory and metabolic crosstalk characteristic of immune activation and tissue remodeling under SIV co-infection. The overall mean LR score also increased by ∼35% compared with wk5, suggesting heightened paracrine activity. Following cART treatment (necropsy), the total number of detected interactions declined to 47 pairs, indicating a partial reduction in paracrine signaling intensity (Figures 5C and S9). However, the signaling landscape did not revert to the LTBI-like pattern. Persistent LR pairs such as TGFβ1-TGFBR3, HSP90B1-TLR7, and TNFSF4-TNFRSF4 remained enriched, while ANXA1-FPR1/2/3 interactions became dominant, suggesting sustained anti-inflammatory and tissue repair signaling. CD4⁺ T cell-macrophage interactions continued to represent over half of all paracrine events, indicating sustained immune activation despite viral suppression. These data demonstrate that SIV co-infection substantially remodels paracrine signaling networks during Mtb infection, and that these alterations persist following cART, failing to return to the balanced LR interaction profile observed during LTBI.

Discussion

We investigated the impact of SIV co-infection on LTBI in a biologically relevant animal model at the single-cell level. Our work identifies specific cell subsets and cell-to-cell interactions impacted during the very early phase of SIV co-infection of LTBI in the lung airways. Alveolar cells from the lung airways have long been used as a surrogate to study lung immune responses in Mtb infection, particularly via the aerosol route, and in Mtb/SIV or Mtb/HIV co-infection, and recently have been used to study correlates of vaccine-induced protection against TB. The data from this study are crucial for developing immune-based interventions, combined with cART and anti-TB therapy, to control dysregulated immune responses during the early events of HIV co-infection of LTBI and provide long-term immune reconstitution. BAL cells were collected from the same RMs at four key time points: wk 5 (asymptomatic phase of Mtb infection; LTBI), wk 11 (2 weeks post-SIV co-infection of LTBI), wk 15 (following 4 weeks of cART), and at necropsy (study endpoint after 9 weeks of cART). This longitudinal sampling allowed us to track early transcriptomic changes within defined cell populations across distinct stages of Mtb/SIV co-infection. As a result, separate LTBI- or cART-naïve control groups were not required, since the wk5 and wk11 samples effectively served as internal controls. We identified a total of 13 clusters including macrophages, monocytes, proliferating macrophages, mDCs, pDCs, mast cells, CD4⁺ T cells, CD8⁺ T cells, T helper cells, natural killer T cells (NKT), natural killer cells (NK), B cells and ciliated cells.

Further analysis of a defined signature gene list revealed a marked reduction in the expression of T_H1-associated transcription factors in Mtb/SIV co-infected, cART-treated RMs. Notably, cART failed to alleviate the reduced T_H1 response at both the transcript and protein levels, despite effective viral suppression and partial CD4⁺ T cell reconstitution. In contrast, the expression of T_H17-associated transcription factors, STAT3 and RORA, remained stable following SIV co-infection and throughout the cART treatment phase. These findings suggest that Mtb/SIV co-infection induces a selective and persistent impairment in T_H1-associated immune responses that is not reversed by cART, highlighting a potential limitation of antiretroviral therapy in fully restoring protective T cell immunity, whereas T_H17-related pathways appear to be more resilient to SIV-induced dysregulation. Strikingly, SIV co-infection in the context of LTBI led to a substantial upregulation of genes associated with T cell activation, which were only partially controlled by cART. Importantly, SIV co-infection also triggered a distinct increase in the expression of key Type 1 IFN signaling transcription factors IRF9, STAT1, and STAT2 within myeloid cell clusters.

Our previous work demonstrated that initiating cART at peak viremia effectively limited SIV-driven immune activation.⁶ However, despite virologic suppression, Mtb-specific immune dysfunction persisted, resulting in continued granuloma formation and incomplete immune recovery. In this study, we extend these findings by showing that SIV co-infection causes a profound and sustained perturbation of the pulmonary immune landscape at the single-cell level that is only partially corrected by cART. Immunophenotypic and single-cell transcriptomic analyses of BAL samples revealed that SIV co-infection led to a sharp decline in CD4⁺ T cells, NK cells, and NKT cells at wk11, with only partial restoration following cART. While CD4⁺ T_CM cells expanded post-cART, CD4⁺ T_EM populations remained depleted relative to levels observed during latent infection,⁶ indicating an incomplete reconstitution of effector function. Functionally, this numerical recovery did not translate into restored Mtb-specific T_H1 responses. Ex vivo stimulation of BAL cells revealed persistent reductions in CD4⁺IFN-γ⁺ and CD4⁺TNF-α⁺ T cells, accompanied by a compensatory rise in IL-17-producing CD4⁺ T cells, suggesting a T_H17-skewed cytokine milieu. This imbalance was mirrored at the transcriptional level, where scRNA-seq revealed decreased expression of T_H1-associated transcription factors and selective modulation of T cell-related gene modules between wks 5 and 15. In contrast, myeloid and epithelial cell programs remained relatively stable, underscoring the dominant contribution of lymphoid dysregulation to the altered immune trajectory. These findings highlight that cART, while capable of suppressing viral replication and partially restoring CD4⁺ T cell numbers, fails to fully reconstitute T_H1 effector function or re-establish a balanced T_H1/T_H17 axis within the Mtb-infected lung. The persistence of a T_H17-biased environment may reflect an adaptive attempt to maintain mucosal integrity but could also contribute to ongoing immunopathology and impaired control of Mtb. These data underscore the limitations of cART as a sole intervention in the setting of Mtb/SIV co-infection and suggest that adjunctive immunomodulatory strategies targeting T cell differentiation and cytokine balance may be required to achieve durable immune restoration and optimal pathogen control.

Our findings reveal that SIV co-infection profoundly reshapes the myeloid landscape and transcriptional circuitry in the Mtb-infected lung, characterized by a transient amplification of Type I interferon-driven inflammatory programs and suppression of immune-regulatory pathways. scRNA-seq analysis demonstrated marked expansion of macrophage and monocyte populations during acute SIV co-infection (wk11), coinciding with strong induction of STAT1, STAT3, and IRF9 across myeloid clusters. This pattern indicates robust activation of Type I IFN signaling, consistent with previous reports linking viral co-infection to IFN-dependent macrophage activation and tissue pathology in HIV/TB comorbidity.²³ Notably, expression of NFKB2, a regulator of noncanonical NF-κB signaling, remained low throughout infection, whereas NFKB1 and KLF6 were persistently upregulated, suggesting a sustained pro-inflammatory transcriptional state. Reduced NFKB2 expression has been associated with heightened macrophage inflammation and defective immune regulation, potentially facilitating dissemination of Mtb when IFN-γ-mediated CD4⁺ T cell responses are impaired.²⁴

Dendritic cell subsets also exhibited distinct transcriptional remodeling under co-infection. While mDC frequencies remained stable, pDCs progressively declined following SIV co-infection and during cART, paralleling increased expression of IRF9 and NFKB1. These findings align with prior observations that HIV-infected dendritic cells exhibit exaggerated IFN responses and pro-inflammatory activity, which may perpetuate immune activation despite viral suppression.²⁵ Sustained KLF6 expression in both mDCs and macrophages further underscores its potential role as a key regulator of inflammatory programming during viral-bacterial co-infection,²⁶ though its specific contribution to dendritic cell function in Mtb/SIV pathology remains to be defined. GO analyses supported these transcriptional signatures, revealing the suppression of antiviral and cytokine response pathways during acute SIV co-infection, followed by reactivation of immune and antigen presentation programs after cART initiation. Monocytes displayed enhanced MHC class I-mediated antigen processing and T cell cytotoxicity pathways at necropsy, indicating persistent immune activation even under effective viral control. Interestingly, mast cells exhibited the strongest transcriptional shifts, including the enrichment of platelet activation and aggregation pathways and loss of immune-regulatory signatures, features suggestive of sustained inflammatory crosstalk within the pulmonary microenvironment. In addition to their immunomodulatory functions, mast cells have been reported to be susceptible to HIV and SIV infection, as they can express CD4 and CCR5 and support viral replication in vitro. Previous studies have demonstrated that infected mast cells may enter a quiescent state with reduced viral production that can be reactivated upon inflammatory stimulation (via NF-κB signaling), suggesting that mast cells may serve as a potential reservoir capable of contributing to residual viral persistence or rebound during infection. Collectively, these findings indicate that mast cells are potential players in persistent immune activation in the lung, amplifying local inflammatory signaling despite effective viral suppression by cART.²⁷ These results highlight that SIV co-infection disrupts the balance between myeloid activation and immune regulation during latent Mtb infection, promoting a Type I IFN-dominant, pro-inflammatory milieu that is only partially resolved by cART. The persistence of this inflammatory gene signature suggests that cART may normalize viral replication but fails to fully restore myeloid homeostasis, potentially impairing T cell priming. Future studies dissecting how transcriptional regulators such as KLF6, NFKB1/2, and IRF9 coordinate myeloid and dendritic cell responses during Mtb/SIV co-infection could elucidate mechanisms underlying immune dysfunction and inform adjunctive therapies aimed at improving immune reconstitution in TB/HIV co-infection.

Understanding the impact of SIV co-infection on pulmonary immune responses during TB is essential for elucidating mechanisms of long-term immune dysregulation and tissue damage. Our scRNA-seq analysis revealed early and coordinated transcriptional changes in CD4⁺ T cells, NK cells, macrophages, and dendritic cells within the first two weeks of SIV co-infection, coinciding with the establishment of acute viral replication. Previous studies have shown that SIV infection elicits robust innate activation and induction of IFN-regulated genes in NHPs,²⁸ contributing to a chronic state of immune activation largely driven by sustained interferon signaling.²⁹^,³⁰ In our model, intravenous infection with 300 TCID of the pathogenic SIVmac₂₃₉ clone leads to peak viral replication and rapid CD4⁺ T cell depletion within 2–3 weeks, aligning with the transcriptional evidence of heightened IFN responses observed at wk11 post-Mtb infection. These findings suggest that early IFN-driven inflammation not only promotes myeloid activation but may also disrupt CD4⁺ T cell homeostasis and antigen presentation, thereby impairing the initiation of effective T_H1 responses. We propose that the outcome of Mtb/SIV co-infection is determined by the balance between IFN-mediated antiviral signaling and the ability of antigen-presenting cells to sustain functional CD4⁺ T cell engagement during this critical early phase of viral replication.

Our analysis of LR networks across distinct stages of Mtb and SIV co-infection reveals that viral co-pathogenesis profoundly remodels paracrine signaling among major immune cell subsets and that these perturbations persist despite effective antiretroviral suppression. The dynamic changes in intercellular communication identified here highlight the complexity of host immune regulation in the setting of dual infection and underscore the long-term imprint of SIV on mycobacterial immunity. During the latent-like phase of low-dose Mtb infection (wk5), the paracrine landscape was relatively constrained and characterized by regulatory and homeostatic signaling. The dominance of CD4⁺ T cell ↔ macrophage interactions, particularly through axes such as LTB-CD40 and TGFβ1-ACVRL1, reflects balanced immune activation consistent with controlled infection and granuloma maintenance. These interactions suggest that effective immune containment during latent TB relies on a tightly regulated dialogue between T cells and macrophages, promoting macrophage activation while limiting excessive inflammation.

SIV co-infection at wk11 induced a marked reorganization of these networks, expanding the total number and diversity of LR pairs by nearly 2-fold. This expansion was most pronounced in the CD4⁺ T cell → macrophage signaling direction, coinciding with a contraction of B cell-derived signals. Newly emergent interactions, including HSP90AA1-CFTR and SEMA4A-ITGA1, are associated with stress responses, metabolic reprogramming, and tissue remodeling, collectively reflecting the inflammatory milieu characteristic of SIV-driven immune activation. The observed 35% increase in mean LR score further indicates intensified paracrine crosstalk, consistent with prior reports that viral co-infection amplifies cytokine and chemokine signaling within granulomatous lesions. Importantly, the loss of regulatory balance and emergence of inflammatory and metabolic pathways suggest that SIV infection disrupts the finely tuned communication necessary for maintaining Mtb containment. Although cART administration partially restored systemic viral control, the paracrine signaling network did not return to its pre-SIV configuration. Despite a reduction in the total number of LR pairs at necropsy, persistent enrichment of TGFβ1-TGFBR3 and HSP90B1-TLR7 interactions indicates ongoing immune modulation and stress signaling. The emergence of dominant ANXA1-FPR1/2/3 interactions further suggests a shift toward anti-inflammatory and tissue repair processes, potentially reflecting attempts to resolve chronic immune activation. However, the continued predominance of CD4⁺ T cell-macrophage signaling, even under cART, implies that immune homeostasis is not fully re-established. These findings are consistent with studies in HIV-infected individuals showing incomplete immune normalization and sustained inflammation despite viral suppression. Thus, SIV co-infection induces persistent remodeling of the paracrine signaling architecture during Mtb infection, characterized by both heightened inflammatory activity and compensatory regulatory pathways. The failure of cART to fully restore the LTBI-like LR landscape underscores the durability of SIV-induced immune reprogramming. This persistent dysregulation may contribute to the impaired control of Mtb observed in TB/HIV co-infection and highlights paracrine signaling networks as potential therapeutic targets to improve immune restoration and TB outcomes in the context of HIV infection.

In conclusion, our study reveals that SIV co-infection of LTBI drives profound transcriptomic remodeling across both innate and adaptive immune cell subsets. These findings underscore the critical importance of interrogating early time points in Mtb/SIV co-infection to better understand the mechanisms underpinning progression to chronic disease. We identify novel LR interactions between cell subsets recruited following SIV co-infection, providing new insights into host-pathogen crosstalk. Using longitudinal BAL samples from Mtb/SIV co-infected, cART-treated RMs, we show that while BAL serves as a practical surrogate for the lung, future spatial transcriptomic studies will be essential to validate these findings in the tissue context. Importantly, although cART resolves certain inflammatory parameters, it fails to fully restore the immune landscape disrupted by viral co-infection, highlighting persistent gaps in immune reconstitution. The relatively short duration of cART in this macaque study (9 weeks) contrasts with the much longer treatment periods typically seen in people living with HIV (PLHIV), where CD4⁺ T cell restoration often spans several months to years. Consequently, the immune recovery observed here may not fully reflect the prolonged reconstitution dynamics characteristic of human infection and treatment. Extended cART studies will be crucial to assess long-term immune reconstitution in the context of Mtb/SIV co-infection. Collectively, our results provide a transcriptomic explanation for the heightened susceptibility of PLHIV with LTBI to TB reactivation or reinfection despite effective cART and suggest that cART alone is insufficient to fully counteract SIV-induced immune dysregulation. These insights pave the way for host-directed therapeutic strategies, such as IL-21 administration to enhance innate-adaptive immune interface³¹ or IDO inhibition¹⁹^,³² to prevent unproductive responses, which warrant rigorous evaluation in the Mtb/SIV/cART preclinical model prior to clinical translation.

Limitations of the study

This study has some limitations. First, the sample size was limited to four RMs, which, although appropriate for high-dimensional single-cell transcriptomic analyses, may restrict generalizability and statistical power at the animal level. Second, BAL was used as a surrogate for lung tissue; while minimally invasive and suitable for longitudinal sampling, BAL may not fully capture the spatial complexity and microenvironmental heterogeneity of granulomatous lesions. Third, the duration of cART in this model (9 weeks) is considerably shorter than treatment courses in PLHIV, potentially limiting the assessment of long-term immune reconstitution. Additionally, the use of intravenous high-dose SIV to ensure timely reactivation may not completely recapitulate natural HIV transmission dynamics. Finally, transcriptomic findings were not comprehensively validated by functional or spatial assays, and future studies integrating proteomics, spatial transcriptomics, and mechanistic perturbation approaches will be important to confirm causality and therapeutic relevance.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact: Dr. Deepak Kaushal (dkaushal@txbiomed.org).

Materials availability

This study did not generate new unique reagents.

Data and code availability

Data availability

The single-cell RNA-seq raw and processed files are available at NCBI Gene Expression Omnibus, and the accession number is GSE254038. All data were generated based on published methods and the study pipeline. Main tools involved: 10X cellranger: gene counts from raw sequence data: https://www.10xgenomics.com/support/software/cell-ranger/latest.

Code availability

R and Seurat 5: single cell analysis: https://satijalab.org/seurat/authors#citation.

GSEA: https://docs.gsea-msigdb.org/#GSEA/GSEA_User_Guide/.

All other items

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request, Deepak Kaushal (dkaushal@txbiomed.org).

Acknowledgments

This work was supported by National Institutes of Health awards K01OD031898, R21AI170148, and R56AI184089-01A1 to R.S; and R01AI111943 and R01AI123047 to D.K. We also acknowledge the role of institutional grant no. U42OD10442 (to D.K.), P30AI168439 (Texas D-CFAR, co-director D.K., and P30AI161943 (IN-TRAC [(Inter Disciplinary Tuberculosis Research Advancement Center)]; both R.S. and D. K. are members). SIVmac₂₃₉ was graciously provided by Drs. Preston Marx and Nick Manness, Tulane National Primate Research Center. SIV viral load assays were performed by the Nonhuman Primate Core Virology Laboratory for AIDS Research and Development, Division of AIDS, NIAID. PMPA and FTC were provided by Gilead Sciences and DTG was provided by ViiV Healthcare. Data were generated in the Genome Sequencing Facility, which is supported by UT Health San Antonio, NIH-NCI P30 CA054174 (Cancer Center at UT Health San Antonio), and NIH Shared Instrument grant no. S10OD030311 (S10 grant to NovaSeq 6000 System) S10OD028732 (to SNPRC), and CPRIT Core Facility Award (RP220662).

Author contributions

R.S. and D.K. designed the study. R.S. performed sample processing. R.S., Z.L., Y.Z., X.L., H.W., and D.K. performed the data analysis. Z.L. and K.W. performed quality control of BAL cells, 10× scRNA-seq, and NGS workflow. S.A.K. and T.B. helped R.S. and D.K. in writing the manuscript.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies

LIVE/DEAD fixable Near-IR stain	Thermo Fisher Scientific	L10119
anti-CD4–PerCP-Cy5.5	BD Biosciences, clone L200	catalog 552838; ; RRID:AB_394488
anti-CD8–APC	BD Biosciences, clone RPA-T8	catalog 555369; RRID:AB_398595
anti-CD3–Alexa Flour 700	BD Biosciences, clone SP34-2	catalog 557917; RRID:AB_396938
anti-CD95–BV421	BD Biosciences, clone DX2	catalog 562616; RRID:AB_2737679
anti-CD28–PECy7	BD Biosciences, clone CD28.2	catalog 560684; RRID:AB_1727459
anti-CD45–BUV395	BD Biosciences, clone D058-1283	catalog 564099; RRID:AB_2738591
anti–IFN-γ–APC-Cy7	BioLegend, clone B27	catalog 506524; RRID:AB_2566136
anti–IL-17–BV605	BioLegend, clone BL168	catalog 512326; RRID:AB_2563887
anti–TNF-α–BV650	BioLegend, clone Mab11	catalog 502938; RRID:AB_2562741

Bacterialandvirusstrains

Mtb CDC1551	BEI Resources	NR-13649
SIVmac₂₃₉	Preston Marx’s Laboratory, Tulane National Primate Research Center	N/A

Biologicalsamples

Bronchoalveolar lavage cells	From Rhesus Macaque at SNPRC	33343, 33994, 34741, 35974

Chemicals,peptides, andrecombinantproteins

ESAT-6	BEI Resources	NR- 34824
CFP-10	BEI Resources	NR- 49425
Brefeldin A	Sigma-Aldrich	B- 7651
ACK lysis buffer	Gibco^TM	A1049201
RPMI 1640	Gibco^TM	11875093
Cryostor	Sigma-Aldrich	C2874-100 ML

Criticalcommercialassays

Single Cell 3′ Gel bead and library kit version 3.1	10× Genomics	PN-1000128

Depositeddata

The single cell RNAseq raw and processed data have been deposited at NCBI Gene Expression Omnibus	This paper	GSE254038

Experimentalmodels:Organisms/strains

Specific Pathogen Free Indian-origin rhesus macaques	Southwest National Primate Research Center	33343, 33994, 34741, 35974

Software andalgorithms

FlowJo	FlowJo LLC, BD	v10.6.1
HALO	Indica Labs	v 4.0
Cell ranger Single Cell Software suite	10× Genomics	v7.0.1
GraphPad Prism	GraphPad Software	V9.4.1
R	The Comprehensive R Archive Network	V4.4.0 https://cran.r-project.org/
Reference genome mmul10	Genebank	https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_003339765.1/
SingleCellSignalR	Bioconductor	https://bioconductor.org/packages//release/bioc/html/SingleCellSignalR.html

Other

(R)-9-(2-phosphonylmethoxypropyl) adenine (PMPA, tenofovir)	Gilead Sciences	Research grade
30 mg/kg of 2′,3′-dideoxy-5-fluoro-3′-thiacytidine (FTC, emtricitabine)	Gilead Sciences	Research grade
integrase inhibitor Dolutegravir DTG	ViiV Healthcare	Research grade

Open in a new tab

Experimental model and study participant details

Study approval

All infected macaques were housed under Animal Biosafety Level 3 facilities at the Southwest National Primate Research Center, where they were treated according the standards recommended by AAALAC International and the NIH Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011). The study procedures were approved by the Animal Care and Use Committee of the Texas Biomedical Research Institute and performed as outlined in Protocol number 1676 MM.

Animal infection

Data were included from completed studies, wherein, a total of four specific pathogen–free Indian-origin rhesus macaques were enrolled from the Southwest National Primate Research Center colony⁶ (Table S1). All macaques (3 females, 1 male) were infected with a low dose of approximately 10 CFU Mtb CDC1551 (BEI Resources, catalog NR13649) via aerosol as described previously.⁵^,⁶^,¹⁵^,¹⁸ A tuberculin skin test was performed at weeks 3 and 5 after Mtb infection to confirm infection. All the macaques were monitored for CRP, percentage body weight, and body temperature weekly throughout the study period. The four RMs with LTBI were then coinfected with 300 TCID₅₀ SIVmac₂₃₉ via the intravenous route 9 weeks after Mtb infection⁵^,⁷^,³³ (provided by Preston Marx’s Laboratory, Tulane National Primate Research Center). All the procedures were conducted by a board-certified veterinary clinician. The viral infection was confirmed through plasma viral loads via RT-qPCR. Upon confirmation of SIV infection, the four macaques were started on cART at 2 weeks after SIV coinfection or 11 weeks after Mtb infection (cART at peak viremia). The macaques were euthanized after 9 weeks of cART treatment. We did not observe an association of sex with the outcomes.

Method details

cART regimen

Coinfected NHPs received a drug regimen consisting of 20 mg/kg of (R)-9-(2-phosphonylmethoxypropyl) adenine (PMPA, tenofovir, Gilead Sciences), 30 mg/kg of 2′,3′-dideoxy-5-fluoro-3′-thiacytidine (FTC, emtricitabine, Gilead Sciences), and 2.5 mg/mL of the integrase inhibitor DTG (ViiV Healthcare). The drugs were administered daily via subcutaneous injection of a cocktail of these 3 drugs in the vehicle KLEPTOSE (Roquette, parenteral grade 346111) at previously published doses.⁵

BAL cell isolation

BAL was performed with 80 mL of sterile saline in BSL3 on NHPs at weeks 5, 11, 15 and necropsy after Mtb infection. The BAL was filtered, centrifuged (1,200 rpm, 10 min at 4°C), lysed for red blood cells (RBCs) using ACK lysis buffer (Gibco) and resuspended in 2 mL final volume of RPMI 1640 media (Gibco). The cells were counted and frozen in Cryostor in internally threaded cryovials in liquid nitrogen for downstream scRNA-seq processing using 10× Genomics platform. The BAL cells were frozen in Cryostor (Sigma Aldrich). Cryostor is formulated to contain 2% dimethyl sulfoxide (DMSO). Cryostor is designed to preserve the cell viability and functionality through modulation of cellular biochemical response to the cryopreservation process. Our lab routinely utilizes this reagent to preserve freshly isolated BAL cells.

High-parameter flow cytometry

High-parameter flow cytometry was performed on BAL cells before Mtb, before SIV (weeks 3 and 9), after SIV, before cART (week 11), and after cART (week 20 or necropsy). The prepared single cells were then stained for surface and intracellular markers to study various cell phenotypes. The freshly collected BAL cells were stimulated ex vivo with Mtb-specific antigen, ESAT-6/CFP-10 (BEI Resources, 10 μg/mL), for a total of 16 h. Brefeldin A (0.5 μg/mL, Sigma-Aldrich) was added 2 h after the onset of stimulation. After stimulation, the cells were stained with LIVE/DEAD fixable Near-IR stain (Thermo Fisher Scientific) and stained subsequently with the following antibodies against cell-surface proteins: anti-CD4–PerCP-Cy5.5 (BD Biosciences, clone L200, catalog 552838), anti-CD8–APC (BD Biosciences, clone RPA-T8, catalog 555369), anti-CD3–Alexa Flour 700 (BD Biosciences, clone SP34-2, catalog 557917), anti-CD95–BV421 (BD Biosciences, clone DX2, catalog 562616), anti-CD28–PECy7 (BD Biosciences, clone CD28.2, catalog 560684), and anti-CD45–BUV395 (BD Biosciences, clone D058-1283, catalog 564099). Cells were then fixed, permeabilized, and stained with the following antibodies against intracellular proteins: anti–IFN-γ–APC-Cy7 (BioLegend, clone B27, catalog 506524), anti–IL-17–BV605 (BioLegend, clone BL168, catalog 512326), and anti–TNF-α–BV650 (BioLegend, clone Mab11, catalog 502938). Cells were washed, suspended in BD stabilizing fixative buffer, and acquired on a BD FACSymphony flow cytometer. Analysis was performed using FlowJo (v10.6.1) software and a previously published gating strategy.⁶

Gross pathology

The animals were euthanized for necropsy and lung lobes, spleen, liver, bronchial lymph nodes were collected. All the tissues were weighed at the time of collection. Tissues were fixed in 10% neutral-buffered formalin, paraffin embedded, sectioned at 5 μm thickness and stained with hematoxylin and eosin (H&E) using standard methods. Lung tissues were collected stereologically at necropsy by a board-certified veterinary pathologist. The H&E stained slides were scanned in Zeiss Axio Scan Z1 and the images were analyzed using HALO 4.0 version software. HALO scores served as an indication for the true percentage of lung affected (primary and secondary lesions from TB). The lesions in each lung lobe were scored for pleural thickening, intralobular septae inflammation, perivasculitis, pneumocyte hyperplasia and lymphadenitis. The score was then utilized to anntotate the different disease types: active non necriotizing, active suppurative, active caseous, latent sclerotic, latent fibrocalcific.

Quality control for frozen BAL cells

Prior to running the BAL cells on 10× Genomics platform, the cells were analyzed for viability using (i) automated cell countess, (ii) manual counts using Trypan Blue and (iii) microscopic evaluation. Briefly, cells were thawed on ice. 100 μL of cells was washed once in 1 mL warmed 1× phosphate buffered saline (PBS) (Gibco), centrifuged, and resuspended in 1 mL of 1× PBS. Cells were mixed in 1:1 ratio with Trypan blue and counted in automated countess as well by hemocytometer. Cellular morphology, including shape and size was determined using a standard bright field light microscope. Institutional approved protocols were applied when removing samples from BSL3.

Single cell RNA library generation and sequencing

BAL cell suspensions were loaded onto Chromium instrument (10× Genomics) to generate single-cell beads in emulsion. Single-cell RNA-seq libraries were then prepared using Single Cell 3′ Gel bead and library kit version 3.1 (10× Genomics). Single cell barcoded cDNA libraries were quantified and sequenced on an Illumina NovaSeq 6000. Read lengths were 28bd for read 1, 10bp for index 1, 10bp for index 2, and 100bp for read 2. Cells were sequenced to about 50,000 reads per cell. A total of 16 libraries were generated from four animals (33343, 33994, 34741, and 35974). Sequencing yielded between approximately 185 million and 680 million total reads per sample (mean ≈460 million). The estimated number of cells captured (Est.Nu.Cells) ranged from ∼2,300 to 22,000, with most samples yielding between 8,000 and 10,000 cells suitable for downstream analyses. Across all libraries, the mean reads per cell ranged from 27,000 to 111,000, and the median number of genes detected per cell varied from ∼800 to 3,200, consistent with high-quality single-cell transcriptomic data. The fraction of reads mapped to cells was consistently high, averaging >93%, with the majority of samples exceeding 95%, indicating efficient capture and sequencing performance. Following standard quality control filtering (retaining cells with <8,000 detected features and mitochondrial content <5%), ∼60% of captured cells were retained per library for downstream analysis. Samples collected at wk11 and wk15 demonstrated higher mean reads and gene counts per cell compared with wk5 and necropsy samples, suggesting improved library complexity and transcriptional depth at these infection stages. Among RMs, 33994 consistently showed the strongest sequencing metrics across time points, whereas 34741 yielded fewer cells overall, particularly at wk11. The necropsy samples generally exhibited lower gene detection and read depth per cell, consistent with reduced transcriptional activity and cellular integrity at terminal stages of infection. Together, these results confirm that high-quality scRNA-seq data were generated across all time points of Mtb/SIV co-infection, providing a robust foundation for downstream analyses aimed at identifying conserved immune features and mechanisms of early immune dysregulation.

All samples were processed using the same experimental workflow and were loaded onto the 10× Chromium platform with a consistent target of 10,000 cells per library. Libraries were prepared in parallel using identical protocols to minimize technical batch effects. Cell viability prior to library preparation was within an acceptable range across samples (typically ∼60–80% by user assessment and ∼60–76% by GSF assessment), and sequencing quality metrics were comparable across datasets, with a high fraction of reads mapping to cells (∼87–97%). Although the number of recovered cells after filtering differed between libraries (estimated recovered cells ranging from ∼2,300 to ∼19,700), downstream analyses were not based solely on absolute cell counts. Instead, cell subset abundances were normalized relative to the LTBI baseline time point for each animal prior to cross-timepoint comparisons. This approach accounts for differences in overall cell recovery and allows evaluation of longitudinal changes within each animal. Figures 3A–3E and 4A–4E represent the proportion of cells within each cluster, including the CD4 T cell cluster, at each timepoint. This provides a more controlled representation of cell subset dynamics independent of total cell numbers recovered in each library and how each animal changes its cell count per cluster from its LTBI stage. The scRNA-seq analysis reflects the transcriptionally defined CD4 T cell cluster within the captured BAL cell population after QC filtering, whereas flow cytometry quantifies CD4 T cells directly in the bulk BAL sample. Differences between these measurements may therefore arise from sampling depth, capture efficiency, and filtering steps inherent to single-cell workflows.

Single cell data analysis

Cell ranger Single Cell Software suite (V7.0.1) from 10× was used to perform sample demultiplexing and generate fastq files. Resulting fastq files were aligned against reference genome mmul10 (Genebank, https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_003339765.1/) with cellranger count. The targeted cell recovery per sample was set to 10,000 cells. The cellranger counting results for 16 samples were further integrated and analyzed by R software with package Seurat (V4.4.0). The data matrix for each sample was read by Read10X and filtered by removing cells which have more than 8000 detected genes in each sample. All 16 samples data were merged, normalized with method “LogNormalize”, and most variable genes were detected by the FindVariableFeatures function with nfeatures 2000. Anchor genes were selected by SelectIntegrationFeatures and FindIntegrationAnchors, and further applied to integrated dataset by IntegrateData. The integrated data were scaled by ScaleData and principal component analysis³⁴ was performed by RunPCA with npcs = 30. To visualize the data, the TSNE dimensionality reduction was performed using the first 20 PCA. Data clustering was run by FindNeighbors (pca 20) and FindClusters (resolution 0.2). Basic marker genes for each cluster were firstly identified using FindAllMarkers function in Seurat R package by (logFC.threshold >0.25, minPct >0.1), then the marker genes with different cut-off were further studied and evaluated. Heatmaps were created by Seurat Package using the mean expression of markers in each cluster per time point. Differential gene expression analysis in the cell clusters across wk5 (LTBI), wk11 (Mtb/SIV co-infection), wk15 (cART), and necropsy was performed using average log₂ fold change, p-values, and Benjamini–Hochberg adjusted p-values, with the percentage of cells expressing each gene also calculated. Results are summarized in Table S7.

SingleCellSignalR analysis

We used the SingleCellSignalR package in R to perform Ligand-Receptor (LR) interaction interface and visualization. Individual cell transcriptomes were normalized to their 99^th read count percentile.³⁵ Cells that had the 99^th percentile equal to zero were not included. Paracrine interactions were infered using published methods.³⁵ Briefly, a paracrine interaction is defined as a crosstalk between two cell types, A (ligand) and B (receptor), with A not expressing B’s receptor (and vice versa). An LR score is considered significant if it is above the threshold.

Quantification and statistical analysis

Overall statistical analysis strategy

All statistical analyses were conducted using GraphPad Prism (version 9.4.1) and R (version 4.4.0). For single-cell-level analyses, the specific R packages and analytical methods employed are described in detail f. Multiple-testing correction was performed using the false discovery rate (FDR), with statistical significance defined at an adjusted p value <0.05. For sample-level analyses, Given the small sample size (n = 4 animals) and the longitudinal study design, non-parametric paired statistical methods were used in Figures 1C–1H, 3A–3E and 4A–4E. Specifically, paired Wilcoxon signed-rank tests were applied for comparisons between timepoints (wk5, wk11, wk15, and wk20 or Necropsy) were performed. No subsets of comparisons were pre-selected. Due to limited statistical power inherent to non-human primate studies, p-values from these pairwise comparisons were not adjusted for multiple testing, and are therefore reported as unadjusted p-values. All pairwise p-values are provided in Table S9. For all sample-level data, results are presented as median with interquartile range (IQR), consistent with non-parametric analysis.

Statistical anlysis of scRNAseq data

We used the genome-wide Rhesus annotation database org.Mmu.e.g., db (Bioconductor release 3.21) for enrichment analysis with the R package clusterProfiler. Significance was defined as FDR-adjusted p < 0.05 and Storey’s q < 0.2. Gene Ontology (GO) enrichment results for biological process, molecular function, and cellular component respectively were visualized using dot plots, bar plots, term similarity networks, and category–gene networks. For both up- and down-regulated genes, we displayed the up to top 20 enriched GO terms. We performed GO GSEA and visualized the top statistically significant up- and down-regulated (up to six respectively) gene sets for Biological Process, Molecular Function, and Cellular Component. Statistical significance was defined as FDR-adjusted p < 0.05. The full enrichment analyses is provided in Table S8. We plotted LR interaction scores for the following pairs by time point: CD4⁺ T cells–macrophages, CD4⁺ T cells–mDCs, CD4⁺ T cells–B cells, and CD8⁺ T cells–macrophages. We plotted violin plots of key genes using columns such as C1.wk5 (cluster 1 at week 5) for each cluster. Differences between time points were assessed using the Wilcoxon signed-rank test for paired comparisons. The key genes are organized into the following 3 modules: T cell: TBX21, IFNG, TNF, LTA, IL18RAP, BHLHE40, IL2, CCR6, RORA, RORC, IRF4, STAT3, IL23R, IL22, IL21, IL21R, IL17A, IL4, IL5, IL6, IL10, IL13, KLF4, TRAC, KLRD1, CCL5, GZMB, GZMH, CTLA4, ICOS, LAG3, NCAM1, KLRK1, NCR1, PRF1, GZMA, GZMB. Myeloid: STAT1, CEBPA, CEBPD, IRF9, KLF6, NFKB1, NFKB2, RELA, RELB, REL, PPARA, PPARD, PPARG, STAT3, STAT6, CEBPB, IRF4, KLF4, GATA3, IL2, IFNG, TNF, CXCL8, IL6, CXCR3, CCR6, CD209. Immune activation: HLA-DRA, CD14, CRP, CXCL8, CXCL10, CD38, LAG3.

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.116944.

Contributor Information

Riti Sharan, Email: rsharan@txbiomed.org.

Deepak Kaushal, Email: dkaushal@txbiomed.org.

Supplemental information

Document S1. Figures S1–S9, and Tables S1–S3

mmc1.pdf^{(5.7MB, pdf)}

Table S7. The table summarizes gene-wise average log2 fold changes (avg_log2FC), p-values (p_val), and adjusted p-values (p_val_adj) for comparisons between wk 5 (LTBI), wk 11 (Mtb/SIV co-infection), wk 15 (cART), and necropsy (N)

Percent of cells expressing each gene at each time point is shown (pct), along with metrics summarizing the maximum fold change (tcMaxFC) and maximum percent expression (tcMaxPct) across time points. This table highlights dynamic changes in gene expression across infection, SIV co-infection, cART treatment, and necropsy in 13 clusters.

mmc5.xlsx^{(265.4MB, xlsx)}

Table S8. Gene Ontology (GO) enrichment analysis was performed using the genome-wide Rhesus annotation database org.Mmu.e

.g.,.db (Bioconductor release 3.21) in conjunction with the clusterProfiler R package

Statistical significance was defined as a false discovery rate (FDR)-adjusted p value <0.05 and Storey’s q value <0.2. Enrichment results for biological process, molecular function, and cellular component categories were visualized using dot plots, bar plots, term similarity networks, and category–gene networks. For both upregulated and downregulated genes, the top 20 enriched GO terms are shown.

mmc6.xlsx^{(863.5KB, xlsx)}

Table S9. Complete statistical results for all pairwise comparisons between timepoints (wk 5, wk 11, wk 15, and necropsy) for Figures 1C-H, 3A-E and 4A-E (n = 4)

P-values were calculated using paired Wilcoxon signed-rank tests. All reported p-values are unadjusted. Due to the small sample size (n = 4 animals), the paired Wilcoxon signed-rank test has limited resolution, resulting in a minimum attainable two-sided p-value of 0.125. Consequently, many comparisons yield identical non-significant p-values despite consistent directional trends across animals.

mmc7.xlsx^{(13.3KB, xlsx)}

Data S1. Full data related to Table S8

mmc8.zip^{(317.9KB, zip)}

References

1.Adekambi T., Ibegbu C.C., Kalokhe A.S., Yu T., Ray S.M., Rengarajan J. Distinct effector memory CD4+ T cell signatures in latent Mycobacterium tuberculosis infection, BCG vaccination and clinically resolved tuberculosis. PLoS One. 2012;7 doi: 10.1371/journal.pone.0036046. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Mollel E.W., Todd J., Mahande M.J., Msuya S.E. Effect of tuberculosis infection on mortality of HIV-infected patients in Northern Tanzania. Trop. Med. Health. 2020;48:26. doi: 10.1186/s41182-020-00212-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.MacPherson P., Lebina L., Motsomi K., Bosch Z., Milovanovic M., Ratsela A., Lala S., Variava E., Golub J.E., Webb E.L., Martinson N.A. Prevalence and risk factors for latent tuberculosis infection among household contacts of index cases in two South African provinces: Analysis of baseline data from a cluster-randomised trial. PLoS One. 2020;15:e0230376. doi: 10.1371/journal.pone.0230376. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Badje A., Moh R., Gabillard D., Guéhi C., Kabran M., Ntakpé J.B., Carrou J.L., Kouame G.M., Ouattara E., Messou E., et al. Effect of isoniazid preventive therapy on risk of death in west African, HIV-infected adults with high CD4 cell counts: long-term follow-up of the Temprano ANRS 12136 trial. Lancet. Glob. Health. 2017;5:e1080–e1089. doi: 10.1016/S2214-109X(17)30372-8. [DOI] [PubMed] [Google Scholar]
5.Ganatra S.R., Bucşan A.N., Alvarez X., Kumar S., Chatterjee A., Quezada M., Fish A., Singh D.K., Singh B., Sharan R., et al. Antiretroviral therapy does not reduce tuberculosis reactivation in a tuberculosis-HIV coinfection model. J. Clin. Investig. 2020;130:5171–5179. doi: 10.1172/JCI136502. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Sharan R., Ganatra S.R., Bucsan A.N., Cole J., Singh D.K., Alvarez X., Gough M., Alvarez C., Blakley A., Ferdin J., et al. Antiretroviral therapy timing impacts latent tuberculosis infection reactivation in a Mycobacterium tuberculosis/SIV coinfection model. J. Clin. Investig. 2022;132 doi: 10.1172/JCI153090. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Bucşan A.N., Chatterjee A., Singh D.K., Foreman T.W., Lee T.H., Threeton B., Kirkpatrick M.G., Ahmed M., Golden N., Alvarez X., et al. Mechanisms of reactivation of latent tuberculosis infection due to SIV coinfection. J. Clin. Investig. 2019;129:5254–5260. doi: 10.1172/JCI125810. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Esaulova E., Das S., Singh D.K., Choreño-Parra J.A., Swain A., Arthur L., Rangel-Moreno J., Ahmed M., Singh B., Gupta A., et al. The immune landscape in tuberculosis reveals populations linked to disease and latency. Cell Host Microbe. 2021;29:165–178.e8. doi: 10.1016/j.chom.2020.11.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.McCaffrey E.F., Donato M., Keren L., Chen Z., Delmastro A., Fitzpatrick M.B., Gupta S., Greenwald N.F., Baranski A., Graf W., et al. The immunoregulatory landscape of human tuberculosis granulomas. Nat. Immunol. 2022;23:318–329. doi: 10.1038/s41590-021-01121-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Sharan R., Bucşan A.N., Ganatra S., Paiardini M., Mohan M., Mehra S., Khader S.A., Kaushal D. Chronic Immune Activation in TB/HIV Co-infection. Trends Microbiol. 2020;28:619–632. doi: 10.1016/j.tim.2020.03.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Foreman T.W., Nelson C.E., Kauffman K.D., Lora N.E., Vinhaes C.L., Dorosky D.E., Sakai S., Gomez F., Fleegle J.D., Parham M., et al. CD4 T cells are rapidly depleted from tuberculosis granulomas following acute SIV co-infection. Cell Rep. 2022;39 doi: 10.1016/j.celrep.2022.110896. [DOI] [PubMed] [Google Scholar]
12.Cai Y., Dai Y., Wang Y., Yang Q., Guo J., Wei C., Chen W., Huang H., Zhu J., Zhang C., et al. Single-cell transcriptomics of blood reveals a natural killer cell subset depletion in tuberculosis. EBioMedicine. 2020;53 doi: 10.1016/j.ebiom.2020.102686. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Akter S., Chauhan K.S., Dunlap M.D., Choreño-Parra J.A., Lu L., Esaulova E., Zúñiga J., Artyomov M.N., Kaushal D., Khader S.A. Mycobacterium tuberculosis infection drives a type I IFN signature in lung lymphocytes. Cell Rep. 2022;39 doi: 10.1016/j.celrep.2022.110983. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Xu Y., Tan Y., Zhang X., Cheng M., Hu J., Liu J., Chen X., Zhu J. Comprehensive identification of immuno-related transcriptional signature for active pulmonary tuberculosis by integrated analysis of array and single cell RNA-seq. J. Infect. 2022;85:534–544. doi: 10.1016/j.jinf.2022.08.017. [DOI] [PubMed] [Google Scholar]
15.Sharan R., Ganatra S.R., Singh D.K., Cole J., Foreman T.W., Thippeshappa R., Peloquin C.A., Shivanna V., Gonzalez O., Day C.L., et al. Isoniazid and rifapentine treatment effectively reduces persistent M. tuberculosis infection in macaque lungs. J. Clin. Investig. 2022;132 doi: 10.1172/JCI161564. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Kaushal D., Singh D.K., Mehra S. Immune Responses in Lung Granulomas during Mtb/HIV Co-Infection: Implications for Pathogenesis and Therapy. Pathogens. 2023;12 doi: 10.3390/pathogens12091120. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Kulkarni S., Endsley J.J., Lai Z., Bradley T., Sharan R. Single-Cell Transcriptomics of Mtb/HIV Co-Infection. Cells. 2023;12 doi: 10.3390/cells12182295. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Sharan R., Zou Y., Singh B., Shivanna V., Dick E.J., Jr., Hall-Ursone S., Luo X., Guo G., Khader S.A., Alvarez X., et al. Concurrent TB and HIV therapies control TB reactivation during co-infection but not chronic immune activation. Nat. Commun. 2025;17:499. doi: 10.1038/s41467-025-67188-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Singh B., Sharan R., Ravichandran G., Escobedo R., Shivanna V., Dick E.J., Jr., Hall-Ursone S., Arora G., Alvarez X., Singh D.K., et al. Indoleamine-2,3-dioxygenase inhibition improves immunity and is safe for concurrent use with cART during Mtb/SIV coinfection. JCI Insight. 2024;9 doi: 10.1172/jci.insight.179317. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Wong M.E., Jaworowski A., Hearps A.C. The HIV Reservoir in Monocytes and Macrophages. Front. Immunol. 2019;10:1435. doi: 10.3389/fimmu.2019.01435. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Cobeña-Reyes J., Wanjalla C., Feria M., Simmons J., Temu T., Nochowicz C., Arafat S., Kityo C., Erem G., Longenecker C., et al. Characterization of Distinct Monocyte Subtypes and Immune Features Associated with HIV, Tuberculosis, and Coronary Artery Disease in a Ugandan Cohort Using Mass Cytometry. Pathog. Immun. 2026;11:14–38. doi: 10.20411/pai.v11i1.945. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Villareal-Rivota B., Meneses-Preza Y.G., Campillo-Navarro M., Ruiz-Sánchez B.P., Soria-Castro R., Barrios-Payán J., Mata-Espinosa D., Donis-Maturano L., Pérez-Tapia S.M., Chávez-Blanco A.D., et al. Impaired control of Mycobacterium tuberculosis infection in mast cell-deficient Kit(W-sh/W-sh) mice. Tuberculosis. 2025;150 doi: 10.1016/j.tube.2024.102587. [DOI] [PubMed] [Google Scholar]
23.Zhang L., Jiang X., Pfau D., Ling Y., Nathan C.F. Type I interferon signaling mediates Mycobacterium tuberculosis-induced macrophage death. J. Exp. Med. 2021;218 doi: 10.1084/jem.20200887. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Jiang J., Cao Z., Li B., Ma X., Deng X., Yang B., Liu Y., Zhai F., Cheng X. Disseminated tuberculosis is associated with impaired T cell immunity mediated by non-canonical NF-κB pathway. J. Infect. 2024;89 doi: 10.1016/j.jinf.2024.106231. [DOI] [PubMed] [Google Scholar]
25.Kim H., Shin S.J. Pathological and protective roles of dendritic cells in Mycobacterium tuberculosis infection: Interaction between host immune responses and pathogen evasion. Front. Cell. Infect. Microbiol. 2022;12 doi: 10.3389/fcimb.2022.891878. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Svanberg C., Nyström S., Govender M., Bhattacharya P., Che K.F., Ellegård R., Shankar E.M., Larsson M. HIV-1 induction of tolerogenic dendritic cells is mediated by cellular interaction with suppressive T cells. Front. Immunol. 2022;13 doi: 10.3389/fimmu.2022.790276. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Jiang A.P., Jiang J.F., Wei J.F., Guo M.G., Qin Y., Guo Q.Q., Ma L., Liu B.C., Wang X., Veazey R.S., et al. Human Mucosal Mast Cells Capture HIV-1 and Mediate Viral trans-Infection of CD4+ T Cells. J. Virol. 2015;90:2928–2937. doi: 10.1128/JVI.03008-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Swainson L.A., Sharma A.A., Ghneim K., Ribeiro S.P., Wilkinson P., Dunham R.M., Albright R.G., Wong S., Estes J.D., Piatak M., et al. IFN-α blockade during ART-treated SIV infection lowers tissue vDNA, rescues immune function, and improves overall health. JCI Insight. 2022;7 doi: 10.1172/jci.insight.153046. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Wang B., Kang W., Zuo J., Kang W., Sun Y. The Significance of Type-I Interferons in the Pathogenesis and Therapy of Human Immunodeficiency Virus 1 Infection. Front. Immunol. 2017;8:1431. doi: 10.3389/fimmu.2017.01431. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Bayón-Gil Á., Martinez-Picado J., Puertas M.C. Viremic non-progression in HIV/SIV infection: A tied game between virus and host. Cell Rep. Med. 2025;6 doi: 10.1016/j.xcrm.2024.101921. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Prasanna P., Herrera B., Schlesinger L.S., Paiardini M., Sharan R. Advances in host-directed therapy for tuberculosis and HIV coinfection: enhancing immune responses. Trends Microbiol. 2025;33:961–975. doi: 10.1016/j.tim.2025.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Singh B., Moodley C., Singh D.K., Escobedo R.A., Sharan R., Arora G., Ganatra S.R., Shivanna V., Gonzalez O., Hall-Ursone S., et al. Inhibition of indoleamine dioxygenase leads to better control of tuberculosis adjunctive to chemotherapy. JCI Insight. 2023;8 doi: 10.1172/jci.insight.163101. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Foreman T.W., Mehra S., LoBato D.N., Malek A., Alvarez X., Golden N.A., Bucşan A.N., Didier P.J., Doyle-Meyers L.A., Russell-Lodrigue K.E., et al. CD4+ T-cell-independent mechanisms suppress reactivation of latent tuberculosis in a macaque model of HIV coinfection. Proc. Natl. Acad. Sci. USA. 2016;113:E5636–E5644. doi: 10.1073/pnas.1611987113. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Alexandrov L.B., Kim J., Haradhvala N.J., Huang M.N., Tian Ng A.W., Wu Y., Boot A., Covington K.R., Gordenin D.A., Bergstrom E.N., et al. The repertoire of mutational signatures in human cancer. Nature. 2020;578:94–101. doi: 10.1038/s41586-020-1943-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Cabello-Aguilar S., Alame M., Kon-Sun-Tack F., Fau C., Lacroix M., Colinge J. SingleCellSignalR: inference of intercellular networks from single-cell transcriptomics. Nucleic Acids Res. 2020;48 doi: 10.1093/nar/gkaa183. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S4. Gene Ontology (GO) enrichment analysis of differentially expressed genes from BAL scRNA-seq in macrophage cluster (C1) comparing wk5 LTBI to wk11 post-SIV co-infection (n = 4)

mmc2.xlsx^{(36.7KB, xlsx)}

Table S5. Gene Ontology (GO) enrichment analysis of differentially expressed genes in the macrophage cluster (C1), comparing wk 5 LTBI to necropsy following cART (n = 4)

mmc3.xlsx^{(19.8KB, xlsx)}

Table S6. Gene Ontology (GO) enrichment analysis of differentially expressed genes in the monocyte cluster comparing wk 5 LTBI to necropsy following cART treatment (n = 4)

Enriched biological process, cellular component, and molecular function terms are shown with corresponding statistics. Analysis was performed using hypergeometric testing with Benjamini-Hochberg FDR correction. The enrichment profile highlights pathways upregulated post-SIV co-infection and cART, including antigen processing and presentation, T cell-mediated immunity and cytotoxicity, leukocyte-mediated immune regulation, vesicle trafficking, cytoskeletal remodeling, ribosomal function, and structural molecule activity.

mmc4.xlsx^{(17.9KB, xlsx)}

Document S1. Figures S1–S9, and Tables S1–S3

mmc1.pdf^{(5.7MB, pdf)}

mmc5.xlsx^{(265.4MB, xlsx)}

Table S8. Gene Ontology (GO) enrichment analysis was performed using the genome-wide Rhesus annotation database org.Mmu.e

.g.,.db (Bioconductor release 3.21) in conjunction with the clusterProfiler R package

mmc6.xlsx^{(863.5KB, xlsx)}

Table S9. Complete statistical results for all pairwise comparisons between timepoints (wk 5, wk 11, wk 15, and necropsy) for Figures 1C-H, 3A-E and 4A-E (n = 4)

mmc7.xlsx^{(13.3KB, xlsx)}

Data S1. Full data related to Table S8

mmc8.zip^{(317.9KB, zip)}

Data Availability Statement

Data availability

Code availability

R and Seurat 5: single cell analysis: https://satijalab.org/seurat/authors#citation.

GSEA: https://docs.gsea-msigdb.org/#GSEA/GSEA_User_Guide/.

All other items

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request, Deepak Kaushal (dkaushal@txbiomed.org).

[bib1] 1.Adekambi T., Ibegbu C.C., Kalokhe A.S., Yu T., Ray S.M., Rengarajan J. Distinct effector memory CD4+ T cell signatures in latent Mycobacterium tuberculosis infection, BCG vaccination and clinically resolved tuberculosis. PLoS One. 2012;7 doi: 10.1371/journal.pone.0036046. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Mollel E.W., Todd J., Mahande M.J., Msuya S.E. Effect of tuberculosis infection on mortality of HIV-infected patients in Northern Tanzania. Trop. Med. Health. 2020;48:26. doi: 10.1186/s41182-020-00212-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.MacPherson P., Lebina L., Motsomi K., Bosch Z., Milovanovic M., Ratsela A., Lala S., Variava E., Golub J.E., Webb E.L., Martinson N.A. Prevalence and risk factors for latent tuberculosis infection among household contacts of index cases in two South African provinces: Analysis of baseline data from a cluster-randomised trial. PLoS One. 2020;15:e0230376. doi: 10.1371/journal.pone.0230376. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Badje A., Moh R., Gabillard D., Guéhi C., Kabran M., Ntakpé J.B., Carrou J.L., Kouame G.M., Ouattara E., Messou E., et al. Effect of isoniazid preventive therapy on risk of death in west African, HIV-infected adults with high CD4 cell counts: long-term follow-up of the Temprano ANRS 12136 trial. Lancet. Glob. Health. 2017;5:e1080–e1089. doi: 10.1016/S2214-109X(17)30372-8. [DOI] [PubMed] [Google Scholar]

[bib5] 5.Ganatra S.R., Bucşan A.N., Alvarez X., Kumar S., Chatterjee A., Quezada M., Fish A., Singh D.K., Singh B., Sharan R., et al. Antiretroviral therapy does not reduce tuberculosis reactivation in a tuberculosis-HIV coinfection model. J. Clin. Investig. 2020;130:5171–5179. doi: 10.1172/JCI136502. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] 6.Sharan R., Ganatra S.R., Bucsan A.N., Cole J., Singh D.K., Alvarez X., Gough M., Alvarez C., Blakley A., Ferdin J., et al. Antiretroviral therapy timing impacts latent tuberculosis infection reactivation in a Mycobacterium tuberculosis/SIV coinfection model. J. Clin. Investig. 2022;132 doi: 10.1172/JCI153090. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Bucşan A.N., Chatterjee A., Singh D.K., Foreman T.W., Lee T.H., Threeton B., Kirkpatrick M.G., Ahmed M., Golden N., Alvarez X., et al. Mechanisms of reactivation of latent tuberculosis infection due to SIV coinfection. J. Clin. Investig. 2019;129:5254–5260. doi: 10.1172/JCI125810. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Esaulova E., Das S., Singh D.K., Choreño-Parra J.A., Swain A., Arthur L., Rangel-Moreno J., Ahmed M., Singh B., Gupta A., et al. The immune landscape in tuberculosis reveals populations linked to disease and latency. Cell Host Microbe. 2021;29:165–178.e8. doi: 10.1016/j.chom.2020.11.013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.McCaffrey E.F., Donato M., Keren L., Chen Z., Delmastro A., Fitzpatrick M.B., Gupta S., Greenwald N.F., Baranski A., Graf W., et al. The immunoregulatory landscape of human tuberculosis granulomas. Nat. Immunol. 2022;23:318–329. doi: 10.1038/s41590-021-01121-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Sharan R., Bucşan A.N., Ganatra S., Paiardini M., Mohan M., Mehra S., Khader S.A., Kaushal D. Chronic Immune Activation in TB/HIV Co-infection. Trends Microbiol. 2020;28:619–632. doi: 10.1016/j.tim.2020.03.015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Foreman T.W., Nelson C.E., Kauffman K.D., Lora N.E., Vinhaes C.L., Dorosky D.E., Sakai S., Gomez F., Fleegle J.D., Parham M., et al. CD4 T cells are rapidly depleted from tuberculosis granulomas following acute SIV co-infection. Cell Rep. 2022;39 doi: 10.1016/j.celrep.2022.110896. [DOI] [PubMed] [Google Scholar]

[bib12] 12.Cai Y., Dai Y., Wang Y., Yang Q., Guo J., Wei C., Chen W., Huang H., Zhu J., Zhang C., et al. Single-cell transcriptomics of blood reveals a natural killer cell subset depletion in tuberculosis. EBioMedicine. 2020;53 doi: 10.1016/j.ebiom.2020.102686. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Akter S., Chauhan K.S., Dunlap M.D., Choreño-Parra J.A., Lu L., Esaulova E., Zúñiga J., Artyomov M.N., Kaushal D., Khader S.A. Mycobacterium tuberculosis infection drives a type I IFN signature in lung lymphocytes. Cell Rep. 2022;39 doi: 10.1016/j.celrep.2022.110983. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Xu Y., Tan Y., Zhang X., Cheng M., Hu J., Liu J., Chen X., Zhu J. Comprehensive identification of immuno-related transcriptional signature for active pulmonary tuberculosis by integrated analysis of array and single cell RNA-seq. J. Infect. 2022;85:534–544. doi: 10.1016/j.jinf.2022.08.017. [DOI] [PubMed] [Google Scholar]

[bib15] 15.Sharan R., Ganatra S.R., Singh D.K., Cole J., Foreman T.W., Thippeshappa R., Peloquin C.A., Shivanna V., Gonzalez O., Day C.L., et al. Isoniazid and rifapentine treatment effectively reduces persistent M. tuberculosis infection in macaque lungs. J. Clin. Investig. 2022;132 doi: 10.1172/JCI161564. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] 16.Kaushal D., Singh D.K., Mehra S. Immune Responses in Lung Granulomas during Mtb/HIV Co-Infection: Implications for Pathogenesis and Therapy. Pathogens. 2023;12 doi: 10.3390/pathogens12091120. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Kulkarni S., Endsley J.J., Lai Z., Bradley T., Sharan R. Single-Cell Transcriptomics of Mtb/HIV Co-Infection. Cells. 2023;12 doi: 10.3390/cells12182295. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Sharan R., Zou Y., Singh B., Shivanna V., Dick E.J., Jr., Hall-Ursone S., Luo X., Guo G., Khader S.A., Alvarez X., et al. Concurrent TB and HIV therapies control TB reactivation during co-infection but not chronic immune activation. Nat. Commun. 2025;17:499. doi: 10.1038/s41467-025-67188-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Singh B., Sharan R., Ravichandran G., Escobedo R., Shivanna V., Dick E.J., Jr., Hall-Ursone S., Arora G., Alvarez X., Singh D.K., et al. Indoleamine-2,3-dioxygenase inhibition improves immunity and is safe for concurrent use with cART during Mtb/SIV coinfection. JCI Insight. 2024;9 doi: 10.1172/jci.insight.179317. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib20] 20.Wong M.E., Jaworowski A., Hearps A.C. The HIV Reservoir in Monocytes and Macrophages. Front. Immunol. 2019;10:1435. doi: 10.3389/fimmu.2019.01435. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Cobeña-Reyes J., Wanjalla C., Feria M., Simmons J., Temu T., Nochowicz C., Arafat S., Kityo C., Erem G., Longenecker C., et al. Characterization of Distinct Monocyte Subtypes and Immune Features Associated with HIV, Tuberculosis, and Coronary Artery Disease in a Ugandan Cohort Using Mass Cytometry. Pathog. Immun. 2026;11:14–38. doi: 10.20411/pai.v11i1.945. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Villareal-Rivota B., Meneses-Preza Y.G., Campillo-Navarro M., Ruiz-Sánchez B.P., Soria-Castro R., Barrios-Payán J., Mata-Espinosa D., Donis-Maturano L., Pérez-Tapia S.M., Chávez-Blanco A.D., et al. Impaired control of Mycobacterium tuberculosis infection in mast cell-deficient Kit(W-sh/W-sh) mice. Tuberculosis. 2025;150 doi: 10.1016/j.tube.2024.102587. [DOI] [PubMed] [Google Scholar]

[bib23] 23.Zhang L., Jiang X., Pfau D., Ling Y., Nathan C.F. Type I interferon signaling mediates Mycobacterium tuberculosis-induced macrophage death. J. Exp. Med. 2021;218 doi: 10.1084/jem.20200887. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Jiang J., Cao Z., Li B., Ma X., Deng X., Yang B., Liu Y., Zhai F., Cheng X. Disseminated tuberculosis is associated with impaired T cell immunity mediated by non-canonical NF-κB pathway. J. Infect. 2024;89 doi: 10.1016/j.jinf.2024.106231. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Kim H., Shin S.J. Pathological and protective roles of dendritic cells in Mycobacterium tuberculosis infection: Interaction between host immune responses and pathogen evasion. Front. Cell. Infect. Microbiol. 2022;12 doi: 10.3389/fcimb.2022.891878. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Svanberg C., Nyström S., Govender M., Bhattacharya P., Che K.F., Ellegård R., Shankar E.M., Larsson M. HIV-1 induction of tolerogenic dendritic cells is mediated by cellular interaction with suppressive T cells. Front. Immunol. 2022;13 doi: 10.3389/fimmu.2022.790276. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Jiang A.P., Jiang J.F., Wei J.F., Guo M.G., Qin Y., Guo Q.Q., Ma L., Liu B.C., Wang X., Veazey R.S., et al. Human Mucosal Mast Cells Capture HIV-1 and Mediate Viral trans-Infection of CD4+ T Cells. J. Virol. 2015;90:2928–2937. doi: 10.1128/JVI.03008-15. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Swainson L.A., Sharma A.A., Ghneim K., Ribeiro S.P., Wilkinson P., Dunham R.M., Albright R.G., Wong S., Estes J.D., Piatak M., et al. IFN-α blockade during ART-treated SIV infection lowers tissue vDNA, rescues immune function, and improves overall health. JCI Insight. 2022;7 doi: 10.1172/jci.insight.153046. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Wang B., Kang W., Zuo J., Kang W., Sun Y. The Significance of Type-I Interferons in the Pathogenesis and Therapy of Human Immunodeficiency Virus 1 Infection. Front. Immunol. 2017;8:1431. doi: 10.3389/fimmu.2017.01431. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Bayón-Gil Á., Martinez-Picado J., Puertas M.C. Viremic non-progression in HIV/SIV infection: A tied game between virus and host. Cell Rep. Med. 2025;6 doi: 10.1016/j.xcrm.2024.101921. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Prasanna P., Herrera B., Schlesinger L.S., Paiardini M., Sharan R. Advances in host-directed therapy for tuberculosis and HIV coinfection: enhancing immune responses. Trends Microbiol. 2025;33:961–975. doi: 10.1016/j.tim.2025.04.006. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] 32.Singh B., Moodley C., Singh D.K., Escobedo R.A., Sharan R., Arora G., Ganatra S.R., Shivanna V., Gonzalez O., Hall-Ursone S., et al. Inhibition of indoleamine dioxygenase leads to better control of tuberculosis adjunctive to chemotherapy. JCI Insight. 2023;8 doi: 10.1172/jci.insight.163101. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Foreman T.W., Mehra S., LoBato D.N., Malek A., Alvarez X., Golden N.A., Bucşan A.N., Didier P.J., Doyle-Meyers L.A., Russell-Lodrigue K.E., et al. CD4+ T-cell-independent mechanisms suppress reactivation of latent tuberculosis in a macaque model of HIV coinfection. Proc. Natl. Acad. Sci. USA. 2016;113:E5636–E5644. doi: 10.1073/pnas.1611987113. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Alexandrov L.B., Kim J., Haradhvala N.J., Huang M.N., Tian Ng A.W., Wu Y., Boot A., Covington K.R., Gordenin D.A., Bergstrom E.N., et al. The repertoire of mutational signatures in human cancer. Nature. 2020;578:94–101. doi: 10.1038/s41586-020-1943-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Cabello-Aguilar S., Alame M., Kon-Sun-Tack F., Fau C., Lacroix M., Colinge J. SingleCellSignalR: inference of intercellular networks from single-cell transcriptomics. Nucleic Acids Res. 2020;48 doi: 10.1093/nar/gkaa183. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Early immune dysregulation in Mtb/SIV co-infection resists cART treatment at the single-cell level

Riti Sharan

Yi Zou

Zhao Lai

Korri Weldon

Todd Bradley

Xi Luo

Hulin Wu

Shabaana A Khader

Deepak Kaushal

Summary

Graphical abstract

Highlights

Introduction

Results

Incomplete immune reconstitution by cART in Mtb/SIV co-infected RMs

Figure 1.

Single-cell transcriptomic profiling identifies distinct immune and epithelial cell populations

Figure 2.

SIV co-infection leads to the depletion of key lymphocyte populations that are not fully restored by cART

Figure 3.

Quantitative profiling reveals differential distribution of inflammatory and effector transcripts across lymphoid and myeloid compartments

SIV co-infection alters transcriptional programs associated with myeloid activation and interferon signaling

Figure 4.

SIV co-infection induces persistent remodeling of paracrine signaling networks during Mtb infection

Figure 5.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Data availability

Code availability

All other items

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Study approval

Animal infection

Method details

cART regimen

BAL cell isolation

High-parameter flow cytometry

Gross pathology

Quality control for frozen BAL cells

Single cell RNA library generation and sequencing

Single cell data analysis

SingleCellSignalR analysis

Quantification and statistical analysis

Overall statistical analysis strategy

Statistical anlysis of scRNAseq data

Footnotes

Contributor Information

Supplemental information

References

Associated Data

Supplementary Materials

Data Availability Statement

Data availability

Code availability

All other items

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases