A total of 40 RhMs (Macaca mulatta) of both sexes between 4.2 and 14.1 years of age were used. At the beginning of this study, all animals were antibody and virus-negative for HIV-2, SIV, STLV-1, simian retrovirus type D, and Macacine herpesvirus 1. All animals were housed at the Tulane National Primate Research Center (TNPRC, full accreditation by AAALAC) and were under the care of TNPRC veterinarians for all sample collection and surgical procedures. The study was approved by the TNPRC Institutional Animal Care and Use Committee (IACUC) and was conducted within the guidelines of the United States Public Health Service Policy and the Guide for the Care and Use of Laboratory Animals (23). All RhMs were grouped into uninfected controls (n=17), acute SIV infection (8–21 days after SIV infection, n=10), and chronic SIV infection (150–422 days after SIV infection, n=13) based on days after SIV_MAC251 infection ( Table 1 ). All acute and chronic SIV-infected RhMs were inoculated with 100-1000 TCID₅₀ pathogenic SIV_MAC251 using either intravenous (IV) or intravaginal (IVAG) routes ( Table 1 ). In our previous studies, we have not detected any association between viral dosage, CD4 depletion, and viral loads in RhMs (7, 24, 25). We have also used IV and IVAG routes of SIV inoculation to mimic the major routes of HIV transmission in humans. At necropsy, peripheral blood and jejunum tissues were collected. Samples were processed for subsequent cellular and molecular assays.

Macaques were housed indoors in Animal Biosafety Level 2 housing with 1:1 light/dark hour cycle, relative humidity of 30-70%, and climate-controlled conditions. Veterinarians monitored animals on a regular basis to ensure their welfare, and animals were fed commercially available monkey chow twice daily. The animals participated in the TNPRC environmental enrichment program, which is reviewed and approved by the IACUC semiannually. All clinical procedures, including administration of anesthesia and analgesics to minimize pain and distress during sampling procedures, were carried out under the supervision of a laboratory animal veterinarian. Animal surgeries and tissue collections were performed by pre-anesthetizing animals with ketamine hydrochloride, acepromazine, and glycopyrrolate, followed by intubation and maintenance of anesthesia with a mixture of isoflurane and oxygen. Buprenorphine was given intraoperatively and postoperatively for analgesia following the veterinarian’s recommendation.

Plasma viral RNA was quantified by either a bDNA signal amplification assay (Siemens Diagnostics, USA) or quantitative reverse transcription-PCR (RT-PCR) with a lower detection limit of 125 and 60 SIV RNA copies/mL of plasma, respectively (7).

Crypts from jejunum were isolated using a low-temperature method standardized in our laboratory and described previously (26). In brief, 3–5 cm of jejunum was collected in a 50 mL tube and thoroughly cleaned with sterile PBS by shaking the tube gently, discarding the fluid and repeating the washes until the fluid was clear. The tissue was chopped into pieces of 2–3 mm length in a Petri dish. The tissue pieces were transferred into a sterile 50 mL tube with 25 mL of 1 x HBSS (Thermo Fisher, USA) containing 1 mM Dithiothreitol (DTT, Thermo Fisher Scientific, USA) and 5 mM EDTA (Thermo Fisher) to remove jejunum villi. The tube was kept in ice and constantly stirred at 200 rpm for 5 min. The undigested tissues were settled down for a minute and the supernatant (which mostly contains villi) was discarded. The undigested tissues were treated with chelating buffer (containing 27 mM Na₃C₆H₅O₇ (G-Biosciences, USA), 5 mM Na₂HPO₄ (USB corporation, USA), 96 mM NaCl (Sigma-Aldrich, USA), 8 mM KH₂PO₄ (VWR, USA), 1.5 mM KCl (Thermo Fisher), 0.5 mM DTT, 55mM D-sorbitol (VWR), 44 mM sucrose (VWR) with pH 7.3) which predominantly removes villi enterocytes. The tube was inverted slowly by hand at least 60 times. The tissues were allowed to settle for a minute and the supernatant was discarded. The treatment of tissues with EDTA and chelating buffer were repeated alternately four times. Finally, the crypts were allowed to separate from the tissues by tapping or shaking the tube vigorously in chelating buffer. The detached crypts were filtered through 100 µm cell strainer and 1% BSA (Sigma-Aldrich) in DMEM (Thermo Fisher Scientific) was added to reach 40 mL volume. The filtrate containing crypts was centrifuged at 4°C and 200 × g for 10 min. The isolated crypts were resuspended in DMEM + 1% BSA and the number of viable crypts was determined using trypan blue in a Glasstic™ Slide 10 with Grids (KOVA International, USA). A total of 20,000 crypts were aliquoted for the enteroid culture as described below. When possible, the remaining crypts were digested for flow cytometry staining by shaking in 10 mL of 1 x HBSS containing 6 U of Dispase (Gibco, USA) and 0.1 mg/mL of DNAse I (Sigma-Aldrich) at 37°C and 210 rpm using an orbital shaking incubator for 20 min. The isolated single cell suspension was stained with Viastain AOPI staining solution and counted using Cellometer (Nexcelom, Lawrence, MA, USA), and used for flow cytometry analysis when cells were >90% viable.

Single cell suspension obtained from crypts were stained for flow cytometry following the protocol described previously (7, 27). Briefly, 1–1.5 X 10⁶ cells resuspended in 100 µL of flow wash buffer (0.5% BSA in PBS) were first stained with live/dead dye (Invitrogen, USA), and subsequently with fluorochrome-conjugated monoclonal antibody (MAb) cocktail for 25 min at room temperature. The MAbs used in the studies were anti-CD24 FITC, anti-LGR5 PE, anti-CD166 BV421, anti-CD45 BV605 and anti-CD44 BV711 ( Supplementary Table 1 ). CD166 (ALCAM) plays an important role in ISC maintenance and proliferative capacity (28). CD44 is a prominent Wnt signaling target in the intestine and is highly expressed in the regenerating ISCs (29). CD24 and CD44 are differentially expressed in LGR5+ stem cells (30). Anti-LGR5 antibody was custom conjugated with PE fluorochrome using lightning link R-PE conjugation kit (Novus Biologicals, USA) following the manufacturer’s protocol. The cells were washed with wash buffer and fixed in BD stabilizing and fixative buffer (BD Biosciences, USA). At least 50,000 cells were acquired either using BD LSRII or Fortessa instruments (BD Biosciences), and the data were analyzed using FlowJo software (version 10.7.1., BD Biosciences). Singlets and live cell populations were subjected to downstream analysis for quantifying all the cell surface markers ( Figure 1 ).

To quantify LGR5+ cells in jejunal tissues, formalin-fixed tissue sections were stained with anti- LGR5 antibodies using the Mach3 Rabbit AP-Polymer Detection Kit (Biocare Medical, USA) as performed earlier (31). The paraffin-embedded tissue sections were deparaffinized and epitope retrieved by heating the tissue sections in low pH citrate buffer (Vector Laboratories, USA) using a microwave. After blocking with a serum-free protein blocker (Agilent Dako, USA) for 30 min, the tissue sections were incubated with anti-LGR5 primary antibodies ( Supplementary Table 1 ). The negative control slide consisted of rabbit Ig fractions (R&D Systems, USA) used at the same isotype and concentration as LGR5 antibodies to determine the background intensity and staining. The tissue sections were subsequently treated with the kit’s probe and polymer and finally developed using BCIP/NBT (Agilent Dako) chromogen system. The slides were then mounted with Vecta Mount AQ (Vector Laboratories). An average of 19–20 view fields (0.116mm²/view field under a 400 x magnification) were enumerated in each of the slides to quantify LGR5+ cells manually. The tissue sites for this evaluation were selected randomly. Tissues were analyzed by two different blinded individuals to avoid bias and averaged prior to the statistical analysis.

Paraffin-embedded jejunum specimens from uninfected control RhMs before and after the crypt isolation process were stained with hematoxylin and eosin (H&E) and evaluated according to standard pathology scores.

Formalin-fixed tissue sections were processed for immunofluorescence with one or a combination of primary antibodies as described earlier (7, 8, 24). In brief, deparaffinized tissue sections were subjected to antigen retrieval and successively incubated with one or two different fluorescence antibodies including anti-cytokeratin, anti-E-cadherin, anti-HD5 [human α defensin 5 for PCs (6)], anti-Ki67, and/or anti-LGR5 primary antibodies ( Supplementary Table 1 ). Alexa Fluor 488 or Alexa Fluor 568 conjugated secondary antibodies (1:1000 dilution, Invitrogen, USA) were used for staining. Anti-nuclear ToPro-3 antibodies (1 µM, Life Technologies, USA) or DAPI (1:5000 dilution, EMD Millipore, USA) were used for nuclear counterstaining. Labeled tissue sections were mounted using Prolong Gold antifade medium (Invitrogen). Images were captured using a TCS SP2 confocal laser scanning microscope (Leica, Germany) equipped with an argon-krypton laser at 488 nm (green), a krypton laser at 568 nm (red), and a helium-neon laser at 633 nm (blue). Negative controls consisted of either omitting the primary antibody or using isotype IgG1 and IgG (H+L) controls. ImageJ (version 1.53a; National Institutes of Health, USA) and Adobe Photoshop (USA) were used to assign pseudocolors to the channels collected. Quantitative fluorescence densitometry was performed using ImageJ software to quantify E-cadherin and Ki67 expression in the jejunum. An average of 19–20 view fields chosen at random (0.232mm²/view field under a 200 x magnification) were quantified. For determining Ki67 expression in epithelial cells, region of interest (ROI) was drawn to limit analysis to epithelium only and fluorescence intensity was quantified within that region. The intensity of E-cadherin and Ki67 protein expression is represented as fluorescence pixel values. For quantification of HD5+ cells, positive cells were counted in a minimum of 20 fields using a Nuance FX multispectral imaging system with colors assigned using Nuance version 2.10 software (Cri, USA). The data is presented as positive cells/mm² of the tissue. For immunofluorescence staining of enteroids (see below), enteroids grown in 8-chambered polystyrene vessel tissue culture-treated glass slides (Falcon, USA) for 7 days were fixed with 2% paraformaldehyde for 30 minutes at room temperature and washed before staining with primary antibodies.

Isolated jejunum crypts were seeded and cultured in vitro using the protocols described for human ISCs with modifications (12, 32). In brief, 1000 crypts suspended in 25 μL DMEM + 1% BSA media were mixed with 25 μL of BD Matrigel basement membrane matrix so that 50 μL of the crypt suspension could be cultured per well in a pre-warmed 24-well cell culture plate (Corning, USA). After carefully dispensing the crypt and matrigel mixture on the plate well, the plate was kept at 37°C for 10 min to allow full gel polymerization. After polymerization, 750 μL of prewarmed complete seeding media was added to each well and the plate was placed into a 37°C incubator under 5% CO₂. Complete seeding media consisted of advanced DMEM/F12 medium (Life Technologies) with 2 mM glutamine (Life Technologies), 10 mM HEPES (Life Technologies), 100 U/mL penicillin (Life Technologies), 100 μg/mL streptomycin (Life Technologies), 1 x N2 supplement (Life Technologies), 1 x B27 supplement (Life Technologies), 1% BSA (Sigma-Aldrich), 50% Wnt-3A-conditioned medium (prepared in-house using L Wnt-3A cell line (ATCC, USA), 1 µg/mL R-Spondin 1 (R&D Systems), 1 mM N-acetylcysteine (Sigma-Aldrich), 500nM A-83-01 (Peprotech, USA), 10 nM [Leu]15-Gastrin (Sigma-Aldrich), 10 mM Nicotinamide (Peprotech), 50 ng/mL EGF (Sigma-Aldrich), 100 ng/mL Noggin 1 (R&D Systems), and 10 μM SB202190 (Sigma-Aldrich). During the first two days of culture, the culture medium was also supplemented with 2.5 µM Thiazovivin (Stemgent, USA) and 2.5 µM CHIR99021 (Stemgent). The culture was monitored regularly under an inverted brightfield microscope for cell division and development, and the medium was replaced with 750 μL of fresh prewarmed complete seeding medium every 2 days.

Fully grown enteroids with multiple budding and thickened and darkened epithelial membranes, typically 10-12 days after primary culture, were used for subculturing. Dissociation of enteroids was performed using gentle cell dissociation agent (Stemcell Technologies Inc., USA) and trypsin-EDTA (Gibco) treatment. In brief, enteroids were harvested after 11 min treatment in gentle cell dissociation reagent at room temperature. Harvested cells were washed with DMEM/F-12 media. The cells were further treated with trypsin-EDTA for 10 minutes at 37°C. Cells were disrupted by gently pipetting with wide orifice pipet tips followed by washing with DMEM/F-12 medium. Cells were examined under an inverted bright field microscope for sufficient disruption and achievement of a single cell suspension. Cells were then counted, and 1000 cells were used for subculturing after mixing with BD Matrigel basement membrane matrix as discussed initially to determine its efficiency to redevelop enteroids and maintenance of phenotypic characteristics by immunohistochemistry assays.

For gene expression analysis, jejunal crypts and enteroids prepared as described above from uninfected normal and chronically SIV-infected RhMs were resuspended in RLT buffer (Qiagen, Germany) and vortexed at high speed for 1–2 min. The lysate was loaded onto a QIAshredder spin column (Qiagen) which was then spun for 2 min at 11,000 rpm. The flow-through was carefully transferred into a new microcentrifuge tube and mixed with one volume of 70% ethanol. The cell lysate preparation was subsequently used for RNA isolation including DNAse treatment for 15 min at room temperature as per the manufacturer’s instructions (RNAeasy mini kit, Qiagen). The quantity and quality of the total RNA was assessed using RNA 6000 Pico kit of the Bioanalyzer system (Agilent Technologies, USA). cDNA from the isolated RNA was synthesized using the Superscript IV first-strand synthesis system (Thermo Scientific) using the manufacturer’s protocol. To remove residual RNA from the final product, 1 μL of E. coli RNAse H was added to 20 μL of the reaction mixture and incubated at 37°C for 20 min. The cDNA products were stored at -20°C for further use. PCR was performed using the DreamTaq green PCR master mix (Thermo Fisher Scientific). Each 25 µL PCR reaction mixture contained 1 x DreamTaq green PCR master mix, 10 µM forward primer, 10 µM reverse primer, and 1 µL of synthesized cDNA. The thermal cycling comprised an initial denaturation at 95°C for 5 min; 36 cycles of denaturation at 95°C for 30 s, annealing at 56–60°C ( Supplementary Table 2 ) for 40 s and extension at 72°C for 45 s; and a final extension at 72°C for 10 min (T1000 Thermal Cycler, Bio-Rad, USA). The primers in the study were designed using an online primer designing tool (Integrated DNA Technologies, USA) and analyzed with OligoAnalyzer (Integrated DNA Technologies). Proper negative control with no template and a positive control targeting GAPDH housekeeping gene were included for each reaction. The amplified products were separated in 1.5% low electroendosmosis agarose prepared in 1 x TAE buffer and visualized under UV gel documentation system.

To assess SIV-mediated changes in the enteroid transcriptome, enteroids derived from 4 SIV-uninfected normal and 3 chronically SIV-infected RhMs were subjected to next-generation sequencing. One µg of the extracted RNA from primary enteroids grown for 12–13 days was used for cDNA library construction at Novogene using an NEBNext^® Ultra RNA Library Prep Kit for Illumina^® (cat# E7420S, New England Biolabs, USA) according to the manufacturer’s protocol. Briefly, mRNA was first enriched by removing ribosomal RNA using the Ribo-Zero kit. The mRNA was fragmented randomly using divalent cation under elevated temperature in the NEBNext first-strand synthesis reaction buffer, then the first-strand cDNA was synthesized using the fragmented mRNA as template and random hexamers primer. Second-strand cDNA was subsequently generated using dNTPs, RNase H and DNA polymerase I. After a series of terminal repair, a ligation and sequencing adaptor ligation, the double-stranded cDNA library was prepared. cDNA libraries of 250–350 bp were preferentially selected and enriched with PCR using Phusion High-Fidelity DNA polymerase. The quantity and quality of the resulting cDNA was assessed by using a Qubit 2.0 fluorometer (Thermo Fisher Scientific) and an Agilent 2100 Bioanalyzer (Agilent Technologies), respectively. Qualified libraries were sequenced on an Illumina Nova Seq 6000 Platform (Illumina, USA) using a paired-end 150 run (2 x 150 bases). Twenty million raw reads were generated from each library. These raw reads were stored in the FASTQ format using (bcl2fastq2) conversion software (v2.17).

Raw reads (in FASTQ format) were subjected to sequence quality control using FastQC (v0.11.9 released: http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (33). FastQC performed a series of analysis modules on raw reads and created a report with statistics for the data analyzed. For each library, FastQC showed high per base/tile sequence quality, exceeding 34 on Phred scale (less than 1/2000 chance of a base being wrong) and few overrepresented sequences. We used TopHat2 (34) to map the raw reads to the reference sequences and annotations of RhM from iGenomes (https://support.illumina.com/sequencing/sequencing_software/igenome.html), which is a collection of reference sequences and annotation files for commonly analyzed organisms. Reads with multiple alignments were discarded. The gene expression counts were calculated using Htseq-count in Galaxy platform (35) (https://usegalaxy.org/).

Transcriptome profiling and data analysis were performed using DESeq2 (36) in R/Bioconductor (https://bioconductor.org/packages/release/bioc/html/DESeq2.html). At first, genes with a read count smaller than 10 were excluded from further analysis. In order to remove the dependence of the variance on the mean, we used regularized-logarithm transformation (rlog) to transform the read counts per gene to log2 scale. Then, principal component analysis (PCA) was applied to obtain an impression on the similarity of SIV-infected and uninfected controls using “DESeq2” so we could identify subgroups or outliers. Differentially expressed genes (DEGs) were identified by comparisons of the counts of SIV-infected and uninfected controls with false discovery rate (FDR) < 0.05 and the absolute value of log2 fold change > 1, where the FDR is the adjusted P value obtained by applying Benjamini and Hochberg’s (BH) FDR correction on the original P value, and fold change is a measure describing how much a quantity changes between different group of RhMs.

We ranked the DEGs based on the FDR from smallest to largest and created hierarchical clustering trees based on the Pearson correlation coefficient for each animal with the top 40 DEGs and the following top 60 DEGs, respectively. The heat maps were generated using the “pHeatmap” R package (https://cran.r-project.org/web/packages/pheatmap/index.html). The heat maps also directly reflected the clustering of groups of samples, enabling the detection of the differential expression patterns of the genes in the SIV-infected and uninfected RhMs.

In order to better understand the biological meaning behind the variant DEGs, we identified the molecular functions or pathways in which DEGs were involved. All DEGs were mapped to the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways using a functional annotation tool named Database for Annotation, Visualization, and Integrated Discovery Bioinformatics Resource (DAVID: https://david.ncifcrf.gov/) (37, 38) for pathway enrichment analysis. KEGG pathways predominantly harvest pathway clusters covering our knowledge on the molecular interaction and reaction networks in DEGs, which are more focused on the metabolic pathways (39). In this study, significantly enriched pathways were identified by DEGs if FDR < 0.05. Moreover, genes weighted by length and categories with FDR < 0.05 were identified as being significantly enriched in the corresponding pathways.

The GO functional analysis mainly covers domains of biological processes, molecular function, and cellular components. We performed GO functional analysis of DEGs that showed significant transcriptional changes in enteroids between SIV-infected and uninfected RhMs by the Gene Ontology Resource (http://geneontology.org/). We used the same criteria as for the pathway enrichment analysis; we claimed significantly enriched GO terms by DEGs if FDR < 0.05, and significantly enriched genes in the corresponding GO terms with FDR < 0.05.

Statistical analysis and graphical representation of the data were performed using GraphPad Prism version 9.1.1 (GraphPad Software, USA). One-way ANOVA was applied to determine any statistically significant differences between the group means. Scatter plots are presented as a graphical method for comparing the frequency of different subsets of LGR5+, LGR5-, and HD5+ cells as well as fluorescence intensity of E-cadherin and Ki67 expression in SIV-uninfected and infected RhMs. To compare the expression of CD24, CD44, and CD166 between LGR5+ and LGR5- cells, Mann-Whitney T-test was applied. Analysis of Variance (ANOVA) followed by Tukey-Kramer multiple comparison as post hoc analysis was applied for LGR5, HD5, E-cadherin, and Ki67 expression analysis to observe statistical significance between different groups. Correlation analyses were performed with a two-tailed Pearson correlation method. A p value of < 0.05 was considered statistically significant.

All SIV-infected RhMs had detectable plasma viral loads ranging from 4.2 x 10³ to 1.6 x 10⁸ copies of RNA/mL of plasma. Acute SIV-infected RhMs had significantly higher levels of plasma viral load (ranging from 3.5 x 10⁶ to 1.6 x 10⁸ copies of RNA/mL of plasma) compared to chronically infected macaques (ranging from 4.2 x 10³ to 2.4 x 10⁷copies of RNA/mL of plasma) ( Figure 1A , p = 0.001). No significant difference in plasma viral load was detected between IV and IVAG inoculated RhMs during either acute or chronic SIV infection ( Supplementary Figures 1A, B ).

To confirm that the low-temperature crypt isolation technique provides well-defined crypts, we performed H&E staining on both pre- and post-processing tissues ( Figures 1B, C ). Preprocessing jejunum tissue showed normal architecture with abundant crypts and villi which were lined by enterocytes. However, after crypt isolation using a low-temperature crypt isolation procedure, the overall intestinal architecture was maintained but a loss of enterocytes and crypts was noticeable, suggesting that this protocol was able to remove a majority of crypts from tissues without introducing major alterations of the remaining tissue ( Figures 1B, C ). Our isolation protocol yielded an average 0.6–0.8 x 10⁶ crypts/3 cm² of jejunum tissue from normal RhMs.

The phenotype of LGR5+ cells in jejunum was determined by flow cytometry assay using enzymatically digested crypts. Cells were analyzed by gating on singlets followed by all cells, live cells, and finally gated on CD45− cells. All CD45− cells were further gated on LGR5+ and LGR5− cells. Both LGR5+ and LGR5− cells were further analyzed for their CD24, CD44, and CD166 surface expression ( Figure 1D ). We detected significantly increased expression of CD24 (mean ± the standard error, 9.1 ± 2.0%), CD44 (mean ± the standard error, 60.0 ± 5.8%), and CD166 (mean ± the standard error, 74.7 ± 4.5%) in LGR5+ cells compared to LGR5− cells (mean ± the standard error, 0.2 ± 0.1%, 8.4 ± 2.0%, and 12.3 ± 3.2%, for CD24, CD44, and CD166 markers, respectively, p < 0.0001) ( Figure 1E ). Using combination gating strategy, we further analyzed those markers and found that the largest proportion of LGR5+ cells were CD24-CD44+CD166+ (mean ± the standard error, 45.6 ± 5.7%) followed by CD24-CD44-CD166- (mean ± the standard error, 22.9 ± 4.0%), CD24-CD44-CD166+ (mean ± the standard error, 17.2 ± 3.6%), and CD24+CD44+CD166+ cells (mean ± the standard error, 7.4 ± 1.8%) ( Figure 1F ). In contrast, the majority of the LGR5− cells were negative for CD24, CD44, and CD166 surface expression (mean ± the standard error, 84.9 ± 3.3%), followed by CD24-CD44-CD166+ (mean ± the standard error, 6.6 ± 1.6%) and CD24-CD44+CD166+ cells (mean ± the standard error, 5.2 ± 1.5%) ( Figure 1F ). CD24-CD44-CD166- cells were significantly higher compared to any other subpopulation in the LGR5− cell population ( Figure 1F ). We also measured the purity of the epithelial cells by performing flow cytometry on the digesting crypts. Digested crypts had little to no leukocyte contamination as detected by CD45 marker (pan leukocyte markers with mean ± the standard deviation, 4.3 ± 3.7%).

Enriched suspensions of intact crypts with classical tubular shape were obtained ( Figure 2A ). On average, more than 90% of the isolated crypts were viable as detected by trypan blue dye exclusion staining method (data not shown). Isolated crypts were cultured in basement membrane matrix, and from day 1, the crypts started rounding up to form enterospheres, with some of them starting their first budding at day 2. The majority of the crypt budding occurred by day 7, and then from day 9 onwards full enteroids were formed as indicated by the appearance of a lumen and epithelial lining ( Figure 2B , see also red lines at day 13). Enteroid cultures were maintained for a total of 12–13 days. Thereafter, enteroids began to disintegrate (data not shown). Most of the proliferating cells were detected in the bud of the tips along with the epithelial cells (Cytokeratin+, Figure 2C ). For subculturing, 10–12-day enteroids were used, as they are at their peak of differentiation and budding at this time. Enteroids were characterized based on their morphology and their ability to differentiate into intestinal epithelial lineages by measuring the expression of cytokeratin, E-cadherin [adherens junction (AJ) protein/differentiation marker for epithelial cells], HD5 (for Paneth cells), mucin-2 (for goblet cells), and Ki67 (for proliferating cells) in subcultured enteroids ( Figures 2D–H ). We detected expression of cytokeratin and Ki67 throughout the subcultured enteroids, suggesting ongoing epithelial cell proliferation in those cultures. LGR5+ and HD5+ cells were also detected in subcultured enteroids, suggesting their continued regeneration and differentiation ( Figures 2E–H ). The presence of cytokeratin and mucin-2 expression in the original and subcultured enteroids indicated presence of absorptive enterocytes and goblet cells, respectively ( Figures 2D, E, H ), further supporting differentiation of enteroid ISC into functional cells.

To confirm the stemness properties of crypts and enteroids, RNA isolated from freshly isolated crypts as well as fully grown primary enteroids (between 10–12 days of culture) of healthy normal RhMs were used to determine the expression of LGR5, OLFM4, VIL1, MUC2, LGR4, FOXA2, and SPON1 genes by RT-PCR ( Supplementary Figure 2 ). The reference gene IDs for these genes are given in Supplementary Table 2 . The amplicon size (base pairs) of our target genes LGR5, OLFM4, VIL1, MUC2, LGR4, FOXA2, SPON1, and GAPDH were 232, 143, 120, 113, 202, 270, 231, and 120, respectively ( Supplementary Figure 2 and Supplementary Table 2 ).

LGR5+ cells were localized to the crypt base columnar cells in normal formalin-fixed paraffin-embedded jejunum tissue ( Figures 3A and Supplementary Figures 3A–C ). Both LGR5+ and HD5+ cells were localized side by side where HD5 cells intersperse with LGR5 cells at the base of the crypts ( Supplementary Figure 3A ). Colocalization of LGR5+ cells with cytokeratin (cytokeratin+LGR5+ cells in yellow) was also detected as well as cell-to-cell adhesion molecule (E-cadherin+LGR5+ cells in yellow) ( Supplementary Figures 3B, C ), consistent with the epithelial cell lineage of ISCs.

To determine the impact of SIV infection on ISCs, the frequency of LGR5+ cells in the jejunum of SIV-infected and uninfected RhMs was determined ( Figure 3B ). No significant differences were detected in the number of LGR5+ stem cells due to the effect of SIV infection. A slight decrease in LGR5+ stem cell population was detected during acute infection (mean ± the standard deviation, 54.4 ± 36.5 cells/mm² and 31.6 ± 12.4 cells/mm² for uninfected-normal and acute SIV RhMs, respectively; p = 0.244). However, the frequency of LGR5+ cells was similar to normal levels during chronic SIV infection (mean ± the standard deviation, 53.7 ± 42.9 cells/mm²). There was no significant increase in LGR5+ cells during chronic infection compared to acute infection (p = 0.343), and there was no significant difference in the frequency of LGR5+ cells between normal SIV-uninfected and chronically SIV-infected RhMs (p = 0.962). HD5+ cells (PCs) were detected in the crypt base in normal as well as SIV-infected jejunum ( Figures 3A and Supplementary Figure 3D ). To determine the impact of SIV on PCs, HD5+ cells were quantified in both SIV-infected and uninfected RhMs. A significant increase in HD5+ cells was detected during acute SIV infection (mean ± the standard deviation, 130.3 ± 20.5 cells/mm² and 156.3 ± 22.0 cells/mm² for uninfected-normal and acute SIV RhMs, respectively; p = 0.026). In addition, HD5+ cells were also increased in chronic SIV infection (mean ± the standard deviation, 157.9 ± 25.5 cells/mm²; p = 0.018) when compared to the values from uninfected normal jejunum tissues ( Figure 3C ). There were no significant differences in the HD5+ cells between acute and chronically SIV-infected RhMs. No significant positive or negative correlation between LGR5+ cells and HD5+ cells was detected (p = 0.96). More male animals were represented in the uninfected group (10/17) than in the acute (1/10) and chronic (1/13) SIV-infected group, and we compared LGR5+ and HD5+ cells on jejunum tissues from male and female SIV-uninfected control animals to determine whether there are any differences in LGR5 and HD5 expression between male and female RhMs ( Table 1 ). However, we did not observe any significant differences in cell population/mm² for either LGR5 (p = 0.354) or HD5+ (p = 0.251) cells ( Supplementary Figures 4A, B ).

E-cadherin, the core membrane protein of the AJ protein family in jejunum tissues, is responsible for regulating intestinal homeostasis and barrier function. To understand the impact of SIV infection on E-cadherin expression in epithelial region, we analyzed the expression of E-cadherin in jejunum tissues and found it was significantly reduced in both acutely (p = 0.032, mean of MFI ± standard deviation, 38.7 ± 5.8) and chronically (p = 0.012; mean of MFI ± standard deviation, 37.7 ± 5.8) SIV-infected RhMs compared to uninfected controls (mean of MFI ± standard deviation, 44.4 ± 3.9; Figures 4A, B ). Ki67, a marker of cell proliferation, was significantly increased in jejunum from both acute (mean of MFI ± the standard deviation, 33.4 ± 6.3) and chronic (mean of MFI ± standard deviation, 31.6 ± 5.8) SIV infection compared to the uninfected controls (mean of MFI ± the standard deviation, 23.9 ± 8.4; p = 0.012 and p = 0.047, respectively) ( Figures 4A, C ). Since no epithelial marker was used along with Ki67 to measure epithelial cell induced cell proliferation, ROIs were carefully drawn in the epithelial region in order to avoid the proliferation normally occurring in Peyer’s patches or other gut associated lymphatic tissue ( Supplementary Figure 5 ). Pearson correlation coefficient analysis of E-cadherin and Ki67 MFI values demonstrated that no significant correlation existed between the expression levels of E-cadherin and Ki67+ proliferating cells in jejunum tissue (p = 0.965), indicating that despite enhanced cell proliferation the disrupted barrier function could not be repaired in SIV infection.

Despite there was significantly decreased E-cadherin expression, we were unable to detect any statistically significant difference in LGR5+ stem cell population in jejunum between uninfected and chronically SIV infected RhMs. To determine whether LGR5+ cells during chronic SIV infection was able to maintain intestinal homeostasis and iEC regeneration as observed in SIV-uninfected ISCs, we compared enteroids from chronically SIV-infected RhMs with enteroids derived from uninfected animals and studied the transcriptomic profiling of ISCs (SRA accession number PRJNA765696). On average about 15-20% of the crypts were able to generate well and developed mature enteroids. We monitored the development of enteroids isolated from both uninfected and chronically SIV infected RhMs microscopically on a regular basis and did observe any differences in budding stages in enteroids between those two groups under the same culture condition ( Supplementary Figure 6 ). RNA was isolated from enteroid cultures grown from four SIV uninfected and three chronically SIV-infected RhMs and subjected to RNA-seq. In all our subsequent analyses we have compared primary culture enteroids from the uninfected control and chronically SIV-infected RhMs. PCA of the variable transcripts after DESeq2 regularized log transformation revealed a clear distinction between SIV-infected RhMs and uninfected controls along the first principal component (PC1) with 91% and the second principal component (PC2) with 4% of the total variance ( Figure 5A ).

The resulting number of significant DEGs with the official gene symbols (n=5293; Supplementary Table 3 ) was used for further downstream analysis. A majority of these DEGs were downregulated in response to SIV infection (2900, compared to 2393 that were upregulated). The volcano plot shows the upregulated (red) and downregulated (blue) DEGs of SIV-infected RhMs and uninfected controls ( Figure 5B ). We performed hierarchical clustering analysis of the regularized log transformed gene expression data of the top 40 DEGs ( Figure 5C ) and the next 60 DEGs ( Figure 5D ). Our top 100 DEGs denoted upregulation of some important genes associated with inflammatory diseases of the gut, like Crohn’s disease and ulcerative colitis, as well as carcinogenesis ( Supplementary Table 3 ). These included ARG2, DHFR, ANGPTL2, SCUBE3, AVIL, EMILIN1 and DUSP11.

There were 1789 DEGs enriched in 284 pathways in the KEGG database which contained at least two DEGs ( Supplementary Table 4 ). Among all the pathways, 38 were significantly enriched (p < 0.05) without the correction of multiple comparisons, and 9 pathways (red) with the p-value adjusted by the BH FDR (p < 0.05) ( Figure 5E and Supplementary Table 4 ). The latter included the metabolic pathway (mcc01100; p = 7.69E-08), mitochondrial pathway (namely, oxidative phosphorylation (OXPHOS)) (mcc00190; p = 2.79E-04), tri-carboxylic acid/TCA cycles (mcc00020; p = 6.68E-04), apoptosis (mcc04210; p = 0.004), Notch signaling pathway (mcc04330, p = 0.012), AMPK signaling pathway (mcc04152, p=0.033) and mTOR signaling pathway (mcc04150, p = 0.037) ( Supplementary Table 4 ).

The balance between proliferation and differentiation of ISCs to maintain homeostasis is controlled by several external and internal factors. Availability of cytokines, growth factors, and different nutrients are considered as important external factors in regulating intestinal homeostasis. Hence, impairment of nutrient metabolism pathways could impact intestinal stemness. KEGG pathway analysis revealed the metabolic pathway in enteroids from SIV-infected RhMs as the highest significantly enriched pathway compared to enteroids from SIV-uninfected RhMs. A total of 398 DEGs were associated with metabolic pathways, out of which 298 and 100 DEGs were downregulated and upregulated, respectively ( Supplementary Tables 4 – 6 ). In the glycolysis/gluconeogenesis pathway, we identified 29 DEGs, of which 23 genes were downregulated (ACSS2, PDHA1, ADH1A, PGAM1, PDHB, ENO2, HK1, ALDH3A2, MINPP1, ALDH1A3, LDHA, PFKL, PGK1, ALDOC, ALDOB, DLAT, ACSS1, ALDOA, DLD, GAPDH, PGM1, PCK2, ALDH9A1) and only 6 genes were upregulated (ADH4, GAPDHS, ADH7, PFKM, GCK, LDHAL6B). Sixteen DEGs in the pyruvate metabolism pathway were detected, where 15 genes (FH, ACSS2, PDHA1, MDH1, GLO1, PDHB, ALDH3A2, LDHA, LDHD, HAGH, DLAT, ACSS1, DLD, PCK2, ALDH9A1) were downregulated and only 1 gene (LDHAL6B) was upregulated. We noted 18 DEGs in fatty acid metabolism, of which 13 genes (PECR, ACAA2, OXSM, ACSL4, MCAT, ACADSB, HADHB, HADHA, PPT1, ACOX3, ACADM, HADH, ACADS) were downregulated and 5 genes (ELOVL5, ACSBG1, PPT2, CPT1B, ACSBG2) were upregulated. All DEGs (BDH2, HMGCS1, BDH1, OXCT1) involved in ketone body metabolism were downregulated. The AMPK signaling pathway, which regulates intestinal barrier function (40) and various energy metabolism pathways (41), was also significantly dysregulated in SIV-infected enteroids compared to uninfected enteroids. A significant downregulation of PRKAB1 and PRKAG1 genes, which encode β and γ subunits of master regulator enzyme AMPK, were detected in enteroids from SIV-infected RhMs ( Supplementary Table 4 ).

The metabolism of different nutrients is the sum of catabolic and anabolic reactions resulting in the production of byproducts that provide energy to drive various cellular processes, including ATP. Glycolysis, TCA cycle, and OXPHOS are among the ATP-generating pathways. KEGG analysis showed that the majority of DEGs associated with TCA and OXPHOS pathways were downregulated. Fifty-five out of 68 DEGs (nearly 81%) of OXPHOS were downregulated ( Figure 6A and Supplementary Table 4 ). Notable downregulated genes included COX (-7B, 4I1, 8A, 6B1, 7A2L, etc.), NDUF (-A10, A12, C2, S8 etc.), ATP (-5C1, 5G3, 5G1, 5F1, etc.), ATP6V (-0B, 0E1, 0D1, 1G1, etc.), SDH (-A, B, C), UQCR, UQCR (-C1, C2, H etc.), LHPP, and PPA1. Upregulated genes in enteroids from SIV-infected RhMs included ND (-1 to 6), ATP (-4A, 4B, 6, 8), ATP6V1B1, and COX11. Interestingly, all DEGs (100%) related to TCA cycles were downregulated. Downregulated genes were FH, PDHA1, MDH1, IDH (-1, 2), DLST, PDHB, CS, ACLY, SUCL (-A2, G1, G2), ACO1, DLAT, DLD, PCK2, and IDH3A. ( Figure 6B and Supplementary Table 4 ). Since mitochondrial metabolisms and signaling pathways play pivotal roles in maintaining the homeostasis, proliferation, differentiation and stemness of IECs, our observation may indicate loss of ISC functions due to mitochondrial metabolic stress in enteroids from chronically SIV-infected RhMs.

Apoptosis of iECs and loss of TJ proteins are characteristic of SIV infection (7). Transcriptomic analysis revealed 29 DEGs involved in apoptosis, among which 26 DEGs were downregulated and 6 DEGs were upregulated. Notable downregulated genes include XIAP, IKBKB, AKT2, TRAF2, NFKBIA, BIRC2, and BCL2L1 ( Figure 6C and Supplementary Table 4 ). These genes encode proteins that negatively regulate the expression of apoptosis genes. Thus, downregulation of these genes suggests uninhibited apoptosis processes. Further, we also detected upregulation of the ENDOG gene, which induces DNA fragmentation and cell apoptosis. However, we also observed downregulation of genes encoding different caspase proteins like CASP9, CASP7, CASP8, CASP10, and CASP3 that might play a role in compensating excess cellular apoptosis and overall increased cell proliferation.

The Notch signaling pathway regulates the proliferation and differentiation of ISCs. We detected 20 DEGs relevant to the Notch signaling pathway in enteroids from SIV-infected RhMs. Important downregulated genes were NCSTN, RFNG, JAG2, APH1A, ADAM10, PSENEN, and LFNG, while upregulated genes were ADAM17, NOTCH1, DTX3L, DTX (1 and 4), and DLL3 ( Figure 6D and Supplementary Table 4 ). The data suggest that SIV pathogenesis during chronic infection leads to dysregulation of the Notch pathway, which might have a significant impact on ISC regeneration, proliferation, and function.

The mTOR signaling pathway regulates a number of cellular processes like autophagy, protein biosynthesis, lipid biosynthesis, cell survival, and growth factor signaling. We identified several upregulated genes including TSC1, PIK3R1, and PIK3R3 that were involved in the mTOR pathway. Several downregulated genes were also detected including PIK3CB, TSC2, IKBKB, RPS6KA3, RPS6KB2, RHEB, EIF4E, and EIF4E2 ( Figure 6E and Supplementary Table 4 ).

Some of the DEGs encoding key proteins of the Wnt signaling pathway were significantly downregulated or upregulated. Downregulation of some positive regulator encoding genes (e.g., CTNNB1, DVL (-2, 3), LGR4, ASCL2) and upregulation of negative regulator encoding genes (e.g., APC2 and SOX9) were detected ( Figure 6F and Supplementary Table 4 ). We also noted upregulation of WNT4, VANGL2, and PRICKLE1 genes, which encode proteins that activate the Wnt signaling pathway, and downregulation of AXIN2 and DKK1 genes, which encode pathway inhibitory proteins. Upregulation of genes encoding activator proteins and downregulation of inhibitory proteins suggest ISCs are Wnt dependent and showed an effort to restore normal Wnt signaling pathway function during chronic SIV infection.

Intestinal inflammatory disease-related DEGs were either upregulated (IL12R (B1, B2), IL23A, ARG2, DHFR and RORA) or downregulated (IFNGR2 and IL18) in enteroids derived from chronic SIV-infected RhMs compared to enteroids derived from SIV-uninfected control RhMs ( Figure 6G ). Several DEGs (TICAM1, TIRAP, CXCL10, CCR4, CCR7, CD80, and CXCR6) associated with immune regulation were also upregulated ( Figure 6H ). We noted 18 out of 25 DEGs encoding TJ proteins and regulatory proteins were downregulated ( Figure 6I and Supplementary Table 4 ). Similarly, 15 of 22 DEGs encoding proteins of AJ were downregulated ( Supplementary Table 4 ). Notable downregulated genes encoding epithelial barrier regulatory proteins CLDN (-1, 3, 4, 7, 23), OCLN, PARD3, CTNNB1, CTNND1, CTNNA1, F11R, RHOA, DSG2, PPP2CB, PTPN1, CSNK2A1, CSNK2B, CDX2, and RAC1 ( Figure 6I and Supplementary Table 3 ).

Molecular function (MF), cellular components (CC), and biological processes (BP) of DEGs showed significant transcriptional changes in enteroids from SIV-infected RhMs compared to enteroids from uninfected RhMs. Under the MF category, 5293 DEGs were annotated, of which 2393 DEGs were upregulated and 2900 DEGs were downregulated. The results of molecular function GO analysis revealed significant upregulation of calcium ion binding GO terms and 11 significantly downregulated GO terms, namely, poly(A) RNA binding, GDP binding, GTP binding, translation initiation factor activity, structural constituent of ribosome, GTPase activity, mRNA binding, electro carrier activity, and fatty-acyl-CoA binding ( Table 2 and Supplementary Tables 7 , 8 ). The CC category also provided two significantly upregulated GO terms: integral component of plasma membrane and voltage-gated calcium channel complex ( Table 2 ). Twenty-five significantly downregulated GO terms were detected under the CC category including extracellular exosome, mitochondrion, membrane, focal adhesion, mitochondrial inner membrane, lysosome, endoplasmic reticulum, lysosomal membrane, nucleolus, and intracellular ribonucleoprotein complex ( Table 2 and Supplementary Tables 9 , 10 ). Similarly, one significantly downregulated GO term “proteolysis involved in cellular protein catabolic process” in the BP category was detected ( Table 2 and Supplementary Tables 11 , 12 ).