Role of tunneling nanotube (TNT)-like structures during the early events of HIV-infection: novel features of tissue compartmentalization and mechanism of HIV spread.

Abstract

HIV has become a chronic disease despite the effective use of anti-retroviral therapy (ART). However, the mechanisms of tissue colonization, viral evolution, generation of viral reservoirs, and compartmentalization are still a matter of debate, and most of the underlying mechanisms behind these are still not well described due to the challenges involved in examining early events of infection at the cellular and molecular level. Thus, there is still an urgent need to explore these areas to develop effective HIV-cure strategies.

Here, we describe the early events of tissue colonization and compartmentalization as well as the role of tunneling nanotubes (TNTs)-like structures (TLS) during viral spread in the presence and absence of effective anti-retroviral treatment. To examine these mechanisms, NOD/SCID IL2 RG−/− (NSG) humanized mice were either directly infected with HIV_ADA or with low numbers of HIV_ADA-infected leukocytes to limit tissue colonization in the presence and absence of TAK779, an effective CCR5 blocker of HIV-entry. We identify that viral seeding in tissues occurs early in a tissue and cell type-specific manner (24 to 72 h). Reduction in systemic HIV-replication by TAK779 treatment did not affect tissue seeding or spreading, despite reduced systemic viral replication. Tissue associated HIV-infected cells had different properties than cells in the circulation because the virus continues to spread in tissues in a TLS dependent manner, despite ART. Thus, understanding these mechanisms can provide new approaches to enhance the efficacy of existing ART and HIV-infection cure strategy.

Introduction

The pathogenesis of HIV-infection involves a series of dynamic interactions between HIV and several host proteins to support effective HIV-infection, generation of viral reservoirs, replication, latency, reactivation, and associated inflammation (1–3). Despite the extensive characterization of host proteins associated with viral replication, there are almost no data about the early mechanisms of tissue viral seeding, compartmentalization, viral reservoir generation, and bystander inflammation generated by early events of tissue infection. Thus, understanding all these points is essential to prevent the formation of viral reservoirs that later complicate treatment and, subsequently, to cure HIV.

In humans, it is accepted that during the acute phase of HIV infection, the virus disseminates into different tissues (4, 5). It has also been proposed that early tissue dissemination may facilitate the establishment of viral reservoirs, differential viral evolution, and viral tissue compartmentalization. The term tissue compartmentalization has been examined mostly by examining viral evolution; however, this area is controversial, and currently, there are data against (6–8) and supporting this concept (9–12). For example, studies in genital and brain compartments indicate that HIV colonizes these tissues early during infection and may serve as distinct nucleation sites for differential viral evolution (13–15). Studies of pulmonary lymphocytes have also described the potential for an HIV reservoir in the lung and genetic compartmentalization due to higher homogeneity of virus compared to a virus found in circulating blood (16–18). Equally important is the presence of macrophages that can be infected in the lungs, which differentiates it from the blood and serves as selective pressures for HIV evolution. Similar data were obtained in monkeys subjected to acute SIV infection but with poor compartmentalization after anal inoculum (9). In contrast, other groups have shown that continuous migration of HIV-1 between blood and tissue compartments results in lower localization of evolution or clonality (19). The usage of ART may also have a role in establishing viral compartmentalization. The exposure of HIV-infected tissues to therapies lead to varying selective pressures against viral compartments that naïve compartments of HIV-1 would not otherwise experience. Further, two recent publications examining viral reactivation and early HIV seeding in humans indicate that early reactivation occurs from several different tissues and anatomical sites, and a single tissue compartment alone cannot function as a primary source of viral rebound, suggesting that multiple tissues are involved. Additionally, viral seeding occurs early after HIV-exposure in humans suggesting that viral seeding and evolution in tissue are imperative to understand the time course of the disease (20, 21).

Despite the complexities of the different routes of viral transmission (rectal, oral, intravenous, or in utero), barriers, and immune response, tissue colonization and heterogeneity of viral isolates specific to tissue compartments are observed. It is then not surprising that several cell types are cumulatively involved in the transmission, amplification, compartmentalization, and generation of viral reservoirs in different tissues. Overall, regardless of the transmission route and the initial type of HIV-targeted cells, within a few weeks, the virus reaches several tissues including the lymph nodes, bone marrow, brain, mucosal barriers, and gastrointestinal tract (22–24) to establish local areas of HIV-infection and generation of viral reservoirs. The spread of HIV infection within these tissues has been suggested to be via cell-to-cell transmission, which is different from the cell-free basis transfer in vitro (25, 26).

Cell-to-cell infection allows HIV to infect neighboring uninfected cells via cellular apparatus that promotes cell to cell communication. Recently, tunneling nanotubes (TNTs) have been shown to be one of such cellular apparatus that enables long-distance cell-to-cell communication, typical over 30-500 μm distance both in vitro and in vivo (27–29). TNTs have been shown to allow for the transfer of some cellular components, organelles, and viruses between connected cells over this long-distance using in vitro studies in primary cells (30–33). However, it still is highly controversial whether TNTs are present in vivo.

Here in this study, we investigate the early events of HIV tissue colonization and compartmentalization as well as whether tunneling nanotubes (TNTs)-like structures (TLS) exist in vivo and their potential role in the viral spread in multiple tissues. We identify that viral seeding in tissues occurs early in a tissue and cell type-specific manner in correlation with TNT expression and viral component transport. These findings are clinically important because the establishment of viral infected cells and generation of latency are essential targets for the HIV-cure strategies.

Materials and Methods

HIV infection of humanized mice.

NSG mice were purchased from the Jackson Laboratory (Bar Harbor, ME; www.jax.org/strain/005557) and housed under pathogen-free conditions following ethical guidelines for the care of laboratory animals at the National Institutes of Health and the University of Nebraska Medical Center. Briefly, NSG mice are immunodeficient. This bred of mice carries two mutations on the NOD/ShiLtJ genetic background leading to deficiencies in functional NK cells and associated inflammation. All experimental protocols were approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee and the University of Texas Medical Branch (IACUC).

To generate human CD34⁺ mice, the newborn NSG mice were irradiated with a RS 2,000 biological irradiator (Rad Source Technologies Inc.) followed with intrahepatic engraftment of human CD34⁺ HSCs that were isolated from human cord blood. The humanization of the animals was monitored monthly from peripheral blood using flow cytometry analysis on human cell markers. At 20–22 weeks of age, a total of 64 animals with replicate levels of human cell reconstitution were selected, then divided evenly into uninfected and HIV-1 infected mouse groups in the presence and absence of TAK779 treatment. The animals were infected intraperitoneally with HIV-1_ADA and then randomly distributed into groups that were sacrificed at days 3, 7, and 14 days post-viral challenge for further immune and viral analysis.

Adult human PBMCs mice were generated by intraperitoneal injection of adult human peripheral blood lymphocytes purified from HIV seronegative donor leukopaks into 8-week old NSG mice at 10×10⁶ PBMCs/mouse. Ten days after engraftment, animal humanization was confirmed by flow cytometry. In total, 42 animals for pilot and 128 animals for the TAK779 experiments were engrafted with human cells and divided into uninfected and HIV infected groups in the presence and absence of TAK779. HIV_ADA challenge was given intraperitoneally at 10⁴ TCID₅₀. Infected animals were then randomly distributed into groups that were sacrificed at days 3, 7, and 14 after viral infection for further immune and viral evaluations. We selected these time points based on a pilot experiment (n=3 independent experiments, data not shown) to evaluate the time course of cell infiltration and viral seeding in the spleen and brain. Overall, we observe that reliable detection of the virus was obtained after 2-3 days. After 14 days, HIV-replication was high and became difficult to evaluate local viral replication and to quantify clusters positive for HIV components, probably due to the high systemic replication.

Flow Cytometry

Peripheral blood was collected at designated time points into EDTA-coated tubes by cardiocentesis at each time point examined. Cellular phenotypes were analyzed for human antigens CD45, CD3, CD4, and CD8 (BD Pharmingen, San Diego, CA) using the fluorescence-activated cell sorting (FACS) system BD LSR2 (BD Immunocytometry Systems, Mountain View, CA) system. CD45⁺ human cells were gated from total lymphocytes. The percentages of CD4⁺ and CD8⁺ cells were obtained from the gate set for human CD3⁺ T cells (BD Pharmingen, San Diego, CA, catalog number 562722). Results were analyzed using FlowJo software (BD Pharmingen, San Diego, CA).

Viral Load Analyses

Plasma samples were isolated from peripheral animal blood by centrifugation. Plasma HIV-1 RNA levels were measured using an automated COBAS Amplicor V2.0/Taqman-48 system (Roche Molecular Diagnostics, Basel, Switzerland) as per the manufacturer’s instructions.

ATP Assay.

Plasma/serum was collected before PBMC separation, and ATP concentration was determined using the ATPlite luminescence assay system (PerkinElmer, MA) by combining 100 μL of the sample with 100 μL of ATPlite reagent. Luminescence was measured using a PerkinElmer EnVision Multilabel Plate Reader. The extracellular concentration of ATP was determined by comparing sample luminescence to a standard curve generated using ATP standards provided by the manufacturer, and as we described (34).

Basic Immunohistochemistry

Tissue samples were collected at the time of animal autopsy, fixed with 4% paraformaldehyde, and embedded in paraffin. Tissue sections of 5 μm thickness were cut and immuno-stained with HLA-DR (clone CR3/43, 1:100, DAKO, Carpinteria, CA) and HIV-1 p24 (1:10, DAKO) antibodies. The DAKO EnVision polymer-based system was used for staining development, and all the sections were counterstained with Mayer’s hematoxylin. Images were obtained with a Nuance EX camera fixed to a Nikon Eclipse E800 microscope using Nuance software (Cambridge Research & Instrumentation, Woburn, MA). Human HLA-DR images were taken at 20x magnifications, and HIV-1 p24 photos were captured at 40x objective magnifications.

Determination of HIV unintegrated and integrated DNA, viral mRNA, and proteins using imaging.

Paraformaldehyde-fixed, paraffin-embedded lung, spleen, lymph nodes, and brain were sectioned (4–12 μm), deparaffinized in xylene and graded alcohol solutions, microwave heated in 10 mM citrate buffer (pH 6.0) for antigen retrieval, cooled down in the same buffer and washed in double-distilled water. HIV-1 RNA detection was modified from (35, 36). Sections were pre-treated at room temperature (unless mentioned otherwise) with 1x sodium chloride/sodium citrate (SSC; 0.15 M NaCl and 0.015 M sodium citrate tribasic dehydrate; pH 7.4; 2 min), 0.2 N HCL (10 min), 1x SSC (2 min), 10 μg/ml proteinase K in 10 mM Tris pH 7.4 and 0.2 mM CaCl₂ (10 min, 37 °C), 0.1 M glycine in Tris-buffered saline (TBS; 5 min), TBS (5 min), 0.1 M triethanolamine/0.25% acetic anhydride (10 min) and 2x SSC (2 min). Slides were then pre-hybridized with hybridization buffer (50% formamide, 4x SSC, 1x Denhardt’s solution, 250 μg/ml yeast transfer RNA, and 250 μg/ml salmon sperm DNA, Thermo-Fisher) in diethylpyrocarbonate water in a humidity chamber (30 min, 48 °C). Hybridization was performed using negative control sense or antisense digoxigenin-labeled riboprobes (Lofstrand Labs), which span the entire HIV-1 genome (8.9-kilobase fragment). The probes were diluted in hybridization buffer (2 ng/μl), denatured (5 min, 65 °C), cooled on ice, and incubated with the sections (overnight, 48 °C). Post-hybridization washes were performed with 50% formamide, 2x SSC (5 and 20 min, 48 °C), 2x SSC (1 min, 37 °C) and 62.5 ng/ml RNase (Roche) in 2x SSC (30 min, 37 °C). Sections were then washed in TBS and incubated with blocking buffer (2% sheep serum (Jackson ImmunoResearch), 0.1% tween, 0.1 M NaCl and 0.1 M Tris pH 7.4; 30 min, room temperature). For viral mRNA detection, RNAscope was used. For RNA staining coupled to protein detection by immunofluorescence after ISH (FISH), sections were also incubated with 20 μg/ml of goat-anti-human-Iba-1 or rabbit-IgG-anti-human-CD3 antibodies (DakoCytomation), washed with TBS, and further incubated with 5 μg/ml of 5-(4,6-dichlorotriazinyl) amino fluorescein (DTAF)-conjugated donkey-anti-sheep-IgG1, cyanin (Cy)5-conjugated goat-anti-mouse-IgG1 and Cy3-conjugated donkey-anti-rabbit-IgG secondary antibodies (Jackson ImmunoResearch; 1 h, room temperature). Finally, sections were washed, counterstained with DAPI, and mounted with Mowiol (Calbiochem Merck) containing 2.5% 1,4-diazobicyclo-(1.2.2)-octane.

For HIV-1 DNA detection, as we recently reported (35, 36), the sections were pretreated with proteinase K (DAKO) diluted 1:10 in TBS (25 min, 37 °C), washed in double-distilled water (twice, 5 min) and 96% ethanol (10 s), and air-dried. Hybridization was performed using peptide nucleic acid probes to HIV-1 nef (biotin-GCAGCTTCCTCATTGATGG) and Alu (Cy5-GCCTCCCCAAGTGCTGGGATTACAG). The probes were prepared in formamide and diluted in TBS (10 μM), denatured (5 min, 93 °C), and incubated with the sections in a humidity chamber (30 min at 45 °C followed by 1 h at 55 °C). Post-hybridization washes were performed with stringent buffer solution (DAKO; 55 °C), and the HIV-1 nef probe was detected by incubation with streptavidin-conjugated to Texas Red (30 min). The sections were then incubated with 5 μg/ml goat-IgG-anti-human-Iba-1 (Abcam, a marker of macrophage/microglia) or CD3 antibody (a T cell marker), washed with TBS, and further incubated with Cy3-conjugated anti-goat-IgG secondary antibody (Sigma–Aldrich). Sections were then washed, counterstained with DAPI, and mounted with diamond antifade (Thermo-Fisher).

Image acquisition and analysis

For DNA FISH, images were taken with a Nikon A1R confocal microscope equipped with spectral detection (Nikon, Japan), as described recently (35, 36), with a resolution corresponding to 20 nm per pixel indicating that co-localization of DAPI, HIV-1 nef DNA, and Alu DNA corresponds to integrated HIV-1 DNA. Co-localization levels were determined using the Menders’ coefficient, as described recently (37). Images were acquired using MetaMorph (Molecular Devices) or NIS-Elements (Advance Research, Nikon, Kanagawa, Japan) software. Image analysis was performed using ImageJ (NIH) or Imaris (Bitplane AG) software. Matched isotype controls were used to determine signal specificity and parametric settings for negative background staining.

Background non-specific signals were processed by the contrast/histogram stretch method (38) in entire images to present only the specific signal in the final figure as follows. After image acquisition of experimental and control samples, a thresholding analysis was performed to retrieve specific signal during image processing. This fluorescence signal thresholding is based on (1) the acquisition of signal from technical and biological controls (for example, isotype controls for antibody staining and non-infected tissue); and (2) these control images allowing for the establishment of a fluorescence signal threshold above which the signals detected in our experimental samples are considered specific. Experimental and control samples were acquired under identical microscope settings and processed the same way during post-acquisition image analysis, as we recently described (35, 36, 39). For quantification of infection, cell types, and TNT like structures, several serial tissue sections (20-30 sections per tissue and from all the animals per group) were analyzed. The high numbers of serial sections and the large size of the tissue analyzed were mainly due to obtaining significant numbers, but more important to identify and reconstruct the TNTs in the tissue.

Our definition of a cluster of HIV-infected cells corresponds to a group of 2 or more cells with HIV-integrated DNA, as described above, that colocalize with DAPI and Alu-repeats as well as human HLA-DR staining. As described in the Results sections, some of these clusters expressed viral mRNA or viral proteins, but this was not a condition to be considered a cluster. To compare and maintain the consistency among different tissues, a similar area was analyzed (~2.2±0.9 cm², except for lymph nodes that the entire tissue was analyzed). As a definition of TNTs, we have strict criteria; first, TNTs communicate two or more cells at a minimal distance of 30 μm. Second, the TNT process can branch and reach distances up to 350 μm. Third, TNTs can transport organelles, vesicular structures, and small molecules between TNT connected cells. Fourth, TNTs are positive for actin and negative or poorly positive for tubulin, a key difference with filopodia, which is positive for both. Lastly, TNTs are positive for several TNT markers not present in filopodia, including TTHY1, GAP43, and protein 14-3-3γ as described. For these experiments, a total of 21 animals per infection route were used. Nine uninfected and 12 HIV infected (3 per time point for controls and 4 per time point for HIV, 3, 7, and 4 days), 20-30 sections per tissue, and from all the animals per group). This pilot experiment allows us to set up a time course of tissue colonization. After the initial pilot experiment, a large experiment with 128 animals to perform the showed experiments at 7 and 14 days in control conditions (28 animals), control+TAK779 (28 animals), HIV (36 animals), and HIV+TAK779 (36 animals) as described for Fig. 2 to Fig. 6. The differences in the animal numbers were due to the potential toxicity of the virus and/or TAK779 treatment. We did not lose any animals during the time course examined.

Figure 2:

Quantification of HIV-replication, immune activation and compromise, and associated inflammation in blood and serum. (A) Determinations of systemic viral replication using serum isolated from control uninfected (C), uninfected plus TAK779 (C+TAK), HIV, and HIV+TAK779 treatment after 7- and 14-days post-treatment. Viral mRNA was determined by the COBAS Amplicor V2.0/Taqman-48 system. HIV infection of PBMCs or direct viral infection resulted in increased replication. Treatment of the animals with TAK779 reduced viral replication. *p≤0.0042 as compared to uninfected conditions, #p≤0.051 as compared to HIV infected conditions, n=170 different animals for the analysis. The numbers of animals involve animals from figure 1 (42 total) and 128 from the rest of the experiment, as described in the method section. (B) Determination of HIV-p24 protein synthesis as described for HIV mRNA. Viral replication was determined by HIV-p24 ELISA (Perkin Elmer, Akron, Ohio). Treatment of the animals with TAK779 reduced viral replication, HIV-p24. *p≤0.0032 as compared to uninfected conditions, #p≤0.0028 as compared to HIV infected conditions. (C) Quantification of CD45 positive cells on the animal blood using FACS. In uninfected conditions, a ~20% of human CD45⁺ cells circulate in the blood of the animals. Upon HIV infection treatment, the percentage of CD45⁺ cells increased to ~50%, and treatment of the HIV-infected animals with TAK779 increased further the percentage of CD45⁺ cells up to 70-80%. *p≤0.0043 as compared to uninfected conditions, #p≤0.0041 as compared to HIV infected conditions. (D) Quantification of CD3⁺ cells from the gated CD45⁺ population indicates that no changes in CD3⁺ cells in pre-infection (Pre), control, control+TAK, HIV, and HIV+TAK779 for 7 days. However, HIV-infection after 14 days decreased the percentage of CD3⁺ cells that were prevented by TAK779 treatment. *p≤0.0044 as compared to uninfected conditions, #p≤0.0041 as compared to HIV infected conditions after 14 days post-infection. (E) Quantification of CD4⁺ cells gated from the CD3⁺ gate indicates that HIV-infection increased CD4 numbers after 7 days post-infection and dramatically decrease at 14 days. TAK779 treatment of HIV infected animals increased further the numbers of CD4⁺ cells at 7 days and protected CD4⁺ cells at 14 days. *p≤0.0044 as compared to uninfected conditions, #p≤0.0041 as compared to HIV infected conditions. (F) Quantification of CD8⁺ cells from the CD3⁺ gate indicates that HIV infection-induced their proliferation only after 14 days post-infection and treatment with TAK779 prevented the increase in CD8⁺ cells. *p≤0.0012 as compared to uninfected conditions, #p≤0.005 as compared to HIV infected conditions after 14 days post-infection. (G) Quantification of the ratio between CD4/CD8 cells indicates that HIV infection increases the ratio early during infection (7 days post-infection) to later (14 days post-infection) collapse. *p≤0.0009 as compared to uninfected conditions. (H) To examine inflammation in the animal circulation, we determined ATP. ATP is a unique inflammatory and danger signal associated with cell death and inflammation. Our data indicate that HIV-infection increased circulating levels of ATP, and treatment with TAK779 prevented the increase in ATP. *p≤0.0013 as compared to uninfected conditions, #p≤0.0038 as compared to HIV infected conditions.

Figure 6:

Quantification of the diffusion of viral components in vivo using different tissues. Due to tissues and cells within tissues did not follow the systemic replication of the virus, we decided to quantify the diffusion of unintegrated HIV-DNA (a signal that is mostly cytoplasmic that do not colocalize with DAPI or Alu-repeats), HIV-nef mRNA, and HIV-p24 protein in a spatial manner using 3D reconstructions and tracing the area analyzed by moving around the HIV infected clusters with HIV integrated DNA up to 80 μm from the cluster. No staining was detected in uninfected tissues, lung, spleen, lymph nodes, and brain. We decided based on the size of TLS and their expression to analyze viral content every 10 μm. (A, B, and C) Spleen and lung data combined, where 0 distance corresponds to the cluster with HIV-integrated DNA. Curves represent the numbers of positive pixels for each viral marker, DNA, mRNA, and HIV-p24. (D-F) Data in lymph nodes. (G-I) Data in the brain. Curves with TAK779 in C, D, E, G, H and I, are significantly different than HIV infection alone, p≤0.0059. A clear separation of diffusion can be observed, and the pressure established by TAK779 to adapt these HIV infected cells or clusters to become highly flexible to the microenvironment and systemic replication. Here, we used 128 animals to perform the showed experiments at 7 and 14 days in control conditions (28 animals), control+TAK779 (28 animals), HIV (36 animals), and HIV+TAK779 (36 animals). The differences in the animal numbers were due to the potential toxicity of the virus and/or TAK779 treatment. We did not lose any animals during the time course examined. 20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS).

Statistical analysis.

Information on the statistical tests used, and the exact values of n (number of experiments) can be found in Figure Legends. All statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software Inc.). The statistical tests were chosen according to the following: two-tailed paired or unpaired t-test was applied on datasets with a normal distribution (Kolmogorov-Smirnov test), whereas two-tailed Mann-Whitney (unpaired test) or Wilcoxon matched-paired signed-rank tests were used otherwise for microscopy analysis in the same group. To compare the different animal group data sets, we used a two-way analysis of variance (ANOVA) to compare the responses among groups. p < 0.05 was considered as the level of statistical significance (* p=0.005; ** p=0.0005; *** p=0.00005; **** p=0.0001)

Results

HIV-seeding into tissues is an early event, and it is not affected by the subsequent anti-retroviral intervention.

To determine the time course of HIV-seeding, spreading, and compartmentalization, we infected humanized animal models in two different ways. As indicated in Fig. 1A, NOD/SCID IL-2 RG−/− (NSG) humanized mice were either injected with low numbers of HIV_ADA-infected leukocytes 10×10⁶ (instead of 25×10⁶) peripheral blood mononuclear cells (PBMCs) injection/mice (40) (Fig. 1A) or directly injected with the soluble virus at low viral titers (HIV_ADA infection, 0.001 MOI Fig. 1A) to limit systemic replication and reduce the numbers of HIV-infected cells early during infection in order to analyze the time course of colonization in the spleen, brain, and lymph nodes (isolated from the neck, inguinal, and popliteal) by staining for HIV-p24 and viral mRNA (Fig. 1B or C, by HIV mRNA quantification) as described in the method section. Lung tissue behaves similarly to the spleen. Thus, spleen data is presented.

Figure 1:

Time course of viral seeding in spleen, lung, brain, and lymph nodes. (A) Time course of sample collection and TAK779 treatment. (B) Time course of tissue colonization by HIV-infected cells in spleen, lung, brain, and lymph nodes using the PBMCs injection model. Viral mRNA was quantified by numbers of positive clusters in the different tissues analyzed. Similar data were found for HIV-p24 positive clusters (data not shown). Overall, after 3 days post-infection, no significant changes in the numbers of positive clusters for HIV mRNA or HIV-p24 were detected. All numbers, clusters of HIV-infected cells, are significant as compared to uninfected conditions, p=0.005. (C) Time course of tissue colonization by HIV-infected cells in spleen, lung, brain, and lymph nodes using the direct injection of the HIV_ADA injection model. As described above, graphs show the numbers of positive clusters in the different tissues analyzed for viral mRNA and HIV-p24. Overall, after 3 days, post-infection no significant changes in the numbers of positive clusters for HIV mRNA or HIV-p24 was detected. All numbers, clusters of HIV-infected cells, are significant as compared to uninfected conditions, p=0.0039. (D-E) A representative example of uninfected spleens stained for HIV-nef mRNA. No staining or background was detected. (F-G) A representative example of HIV-nef mRNA staining after 3 days post-infection at low and high magnification. (H-I) A representative example of HIV-nef mRNA staining after 7 days post-infection at low and high magnification. (J-K) A representative example of HIV-nef mRNA staining after 14 days post-infection at low and high magnification. As observed, the numbers of the cluster do not increase, but the intensity does. (L-S) A representative example of HIV-p24 expression as described by viral mRNA: HIV-p24 staining of uninfected tissues (L-M), 3 (N-O), 7 (P-Q), and 14 days (R-S) post-infection. Clusters are denoted by arrows. Bar: 2000 μm and for the higher amplification was 380 μm. For these experiments, a total of 21 animals per infection route were used. Nine uninfected and 12 HIV infected (3 per time point for controls and 4 per time point for HIV, 3, 7, and 4 days). 20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS)

A pre-infection sample was taken from a group of uninfected mice (Pre-FACS, as indicated in all figures). Designated mice were injected with HIV-infected or uninfected (control) PBMCs, or directly infected with soluble HIV_ADA virus on day 10. Starting from two days post-HIV infection, some of the mice received a daily injection of CCR5 antagonist TAK779 (150 μg/animal) (41) until 14 days post-infection, the last assay time point (Fig. 1A). Blood and tissue samples were collected at every time point to quantify the degree of tissue infiltration by HIV-infected cells, populations of immune cells, viral replication, and to quantify the numbers of human cell clusters infected with HIV within the tissues.

Viral seeding in the spleen, brain, and lymph nodes was detected reliably three days post-HIV infection irrespective of the infection carrier, PBMC infected cells (Fig. 1B), or soluble virus (Fig. 1C). Most HIV-infected cells within tissues were present as clusters; however, after three days post-infection, no new clusters of HIV-infected cells were detected by staining for HIV mRNA (Fig. 1, D–K) or HIV-p24 protein (Fig. 1L–S). These data indicate that viral seeding in spleen, lung, brain, and lymph nodes is an early event, and colonization of different tissues does not have significant differences.

Circulating levels of CD4 and CD8 cells, as well as ATP, indicates an early and late immune response to the virus.

To determine the degree of systemic infection, immune response, inflammation, and immune compromise, we determined the cellular content of the blood, viral mRNA, and HIV-p24 to assess viral replication and ATP secretion due to its high inflammatory capacity (Fig. 2). We determined these parameters during pre-infection (Pre), control uninfected (C), control plus TAK779 (C+TAK), HIV infection alone (HIV), and HIV plus TAK779 (HIV+TAK) after 7- and 14-days post-infection (Fig. 2). Since both infection models give similar results, we decided to combine data from both models for all subsequent experiments.

Systemic viral replication was determined in serum by quantifying HIV mRNA (Fig. 2A) and HIV-p24 protein (Fig. 2B). No HIV mRNA or HIV-p24 were detected in control or control +TAK779 animals (Fig. 2A). HIV-mRNA was detected after 7- and 14-days post-infection with a significant reduction in viral replication due to the early treatment of the animals with TAK-779 at both time points assayed. In agreement with the viral mRNA data, the determination of HIV-p24 protein also indicates that treatment with TAK779 reduced systemic viral replication at 7- and 14-days post-infection (Fig. 2B). Thus, in both models, HIV-replication was robust and decreased in response to the TAK779 treatment.

To evaluate the immune response, we quantified the total number of CD45⁺ cells in the blood of the animals. HIV-infection increased the percentage of CD45⁺ circulating cells, as analyzed by FACS (Fig. 2C). TAK779 increased further the % of CD45⁺ cells but only after 7 days post-infection (Fig. 2C). We believe that our experiments dissected two early stages of early HIV infection observed in humans, an early immune response that increases the CD45⁺ cells in response to the virus with a CCR5 component of the TAK779 treatment protecting the cells from apoptosis induced from viral exposure (7 days). At 14 days, the CD45⁺ cells remain high, with a significant CD4 compromise (see below), suggesting that at 14 days, apoptosis is high recapitulating the later events of early infection where T cells apoptose in response to viral infection. If CD3⁺ cells were gated from the CD45⁺ cell population, no changes in the percentage of CD3⁺ cells were observed in HIV or HIV+TAK779 infected conditions as compared to control uninfected conditions (Fig. 2D, C). However, 14 days post-HIV-infection, a significant decrease in the CD3⁺ cell population was observed (Fig. 2D, HIV). Treatment with TAK779 prevented the decay in CD3⁺ after HIV-infection (Fig. 2D, 14 days, HIV+TAK).

Gating into the CD4⁺ cells from the CD3⁺ cell gated population indicates that at the early time points of HIV-infection, 7 days, a significant increase in CD4⁺ cells was detected. Treatment of HIV infected animals with TAK779 increased further the percentage of CD4⁺ cells (Fig. 2E), suggesting some degree of early protection of CD4⁺ cells by blocking CCR5 against the virus (Fig. 2E, compare to the CD45⁺ cells). However, after 14 days post-infection, CD4⁺ cells decay to almost undetectable levels (Fig. 2E, yellow squares). In the day 14 HIV-infected mice treated with TAK779, we observed there was no significant CD4⁺ decay, as seen in the day 14 HIV infected mice group (Fig. 2E brown vs. yellow squares).

Gating CD8⁺ cells from the CD3⁺ cell gated population indicates that CD8⁺ cells proliferate at about 14 days post-infection (Fig. 2F), and treatment with TAK779 prevented the increase in CD8⁺ cells when compared to HIV infected at day 14 (Fig. 2F). Quantification of the ratio between CD4⁺/CD8⁺ indicates a robust immune response at 7 days post-infection and probably a compromised immune system after 14 days post-infection, due to the HIV-infection induced decay in CD4⁺ T cells and increase in CD8⁺ T cells (Fig. 2G). Thus, our model represents an early immune response to the virus at 7 days post-infection and, potentially, a late perturbed immune system in the animal model at 14 days post-infection.

Large increases in extracellular levels of ATP are associated with cell death and serve as a proinflammatory signal that leads to the recruitment and activation of different cell populations (42, 43). Also, our recent publication indicates that circulating levels of ATP are high in HIV-infected individuals despite effective ART (34). Thus, ATP denotes a specific type of inflammation observed in HIV and also is a biomarker of cognitive impairment as well as a direct contributor to blood-brain barrier disruption that probably helps HIV-infected cells to infiltrate into the CNS (34). In control and TAK779 treated conditions, at 7 and 14 days, the circulating levels of ATP were low as expected. However, upon HIV-infection, ATP levels increased significantly (~4-6 times) as compared to uninfected conditions (Fig. 2H, 7, and 14 days). Treatment of the HIV-infected animals with TAK779 prevented this significant increase in circulating levels of ATP. However, when compared to the uninfected controls with or without TAK779, the ATP levels were still higher (Fig. 2H), suggesting that in addition to changes in cell populations, a significant unspecific inflammation is ongoing in a replication-dependent manner (compare Fig. 2H to Fig. 2A and B).

Overall, our HIV replication, cellular response, and inflammation marker data indicate that our acute HIV-infection model shows a strong immune response against the virus early on (7 days post-infection). However, at 14 days post-infection, the reconstituted human immune system was compromised with high viral replication, strong CD4 decay, CD8 response, and high levels of circuiting ATP as a measure of systemic inflammation. Thus, using our model, we can study both stages of HIV infection (7- and 14-days post-HIV infection).

Quantification of HIV integrated DNA, viral RNA, and HIV-p24 positive clusters in the spleen, lymph nodes, and brain to examine tissue colonization and local replication.

To quantify the cluster numbers of human cells in the different tissues, we stained for human HLA-DR, a human cell marker. Overall, as described in Fig. 1, 10.3±3.65 to 32.8±19.75 cluster per tissue were observed depending on the tissue analyzed. For all the animals and tissue sections analyzed, the numbers of clusters per unit of the area were consistent and similar (~2.2±0.9 cm², except for lymph nodes that the entire tissue was analyzed, 20-30 sections per tissue and from all the animals per group). The main reason for these large analyses was the reconstruction of TNTs in particular areas of the tissues. TNT range from 30 to 500 μm; thus, to assure proper identification, a large 3D reconstruction is required. From all the human clusters identified, only 62.7±8.79% become infected with HIV, as described below (Fig. 3, black bars).

Figure 3:

Quantification of human clusters in the spleen, lymph nodes, and the brain to determine the ratio of HIV-integrated DNA, viral mRNA, and HIV-p24 protein expression using imaging. We calculate the % of clusters versus HIV expression. The first observation is that not all human clusters become infected with HIV, only 60-65%. (A) Quantification of HIV replication in the spleen by detecting HIV-integrated DNA (red bars), viral mRNA (blue bars), and HIV-p24 (green bars). We decided to combine the data from spleen and lung due that no significant differences can be observed. No HIV-integrated DNA, viral mRNA, and HIV-p24 were detected in control conditions (C or C+TAK779). HIV-infection for 7 and 14 days resulted in an even expression of HIV-integrated DNA (probes colocalized with DAPI and Alu-repeats, see methods), viral mRNA, and HIV-p24 protein. Thus, clusters with HIV-DNA were producing viral mRNA and proteins. TAK779 treatment of HIV-infected animals reduced viral mRNA and HIV-p24 protein expression in a similar manner and ratio as described for systemic replication in Fig. 2. *p≤0.0021 as compared to HIV infected conditions. (B) Quantification of HIV replication in lymph nodes (cervical and inguinal). The approach was like the one described above. Similar to spleen, HIV-integrated DNA correlates with viral mRNA and HIV-p24 protein expression at 7- and 14-days post-infection. However, TAK779 treatment of HIV-infected animals did not alter the expression of viral mRNA (blue bars). Still, it reduced the expression of HIV-p24 clusters in a significant manner as compared to HIV-infection alone, *p≤0.0012 as compared to HIV infected conditions. Thus, lymph nodes have different viral behavior as compared to spleen and lung in the presence of TAK779 immune and viral pressure. (C) Quantification of HIV replication in brain tissues. The approach was like the one described above. Similar to spleen, and lymph nodes, HIV-integrated DNA correlates with viral mRNA and HIV-p24 protein expression at 7- and 14-days post-infection. However, TAK779 treatment of HIV-infected animals further increased HIV-DNA, viral mRNA, and HIV-p24 at both time points analyzed. Thus, TAK779 increased infection within the brain. #p≤0.05 as compared to HIV infected conditions. Thus, lymph nodes and brain have a different viral behavior as compared to spleen and lung in the presence of TAK779 immune and viral pressure. Here, we used 128 animals to perform the showed experiments at 7 and 14 days in control conditions (28 animals), control+TAK779 (28 animals), HIV (36 animals), and HIV+TAK779 (36 animals). The differences in the animal numbers were due to the potential toxicity of the virus and/or TAK779 treatment. We did not lose any animals during the time course examined. 20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS)

Upon isolation of the several tissues, including spleen/lung (plotted together, Fig. 3A), lymph nodes (Fig. 3B), and brain (Fig. 3C), quantification of HIV DNA positive, mRNA, and p24 positive clusters was performed using in situ DNA and mRNA hybridization combined with HIV-p24 staining and subsequent confocal microscopy (Fig. 3). Quantification of cell clusters containing HIV-DNA (red bars), HIV-mRNA (blue bars), or HIV-p24 protein (green bars) positive cells in the spleen (Fig. 3A), lymph nodes (Fig. 3B), and brain (Fig. 3C) from uninfected control, uninfected control plus TAK779 treatment (C+TAK), HIV alone (HIV), and HIV plus TAK779 treatment (HIV+TAK) are represented in Fig. 3.

Analysis of uninfected control tissues in the presence or absence of TAK779 treatment indicates no non-specific staining for HIV-integrated DNA, viral RNA, or HIV-p24 protein as expected (Fig. 3A–C, C or C+TAK). HIV-infection for 7 days indicates that at least half of the human clusters in spleen/lung become infected with HIV, as seen in the even expression of HIV-integrated DNA, viral mRNA, and HIV-p24 protein (Fig. 3A, HIV). Treatment of the HIV-infected animals with TAK779 did not reduce the numbers or intensity of HIV DNA staining at both time points, 7 and 14 days (Fig. 3A, HIV+TAK). However, HIV-infected animals treated with TAK779 show a significant decrease (a reduction of 48.2±8.8 % as compared to HIV alone) in HIV-mRNA expression (Fig. 3A, HIV+TAK, blue bars). In agreement, HIV-p24 positive clusters were reduced to almost undetectable levels in the spleen by TAK779 treatment in HIV infected conditions (Fig. 3A, HIV+TAK, green bars). This suggests that in spleen/lungs, TAK779 treatment of HIV-infected animals silences some of the clusters to become negative for replication products such as viral RNA and proteins, but TAK779 did not alter the numbers of clusters with HIV-integrated DNA (Fig. 3A), suggesting that seeding of the infection is unperturbed by TAK779.

In a similar manner, analysis of lymph nodes (cervical and inguinal) indicates that there was no unspecific staining for HIV-integrated DNA, HIV-mRNA, or viral proteins in control conditions both in the presence or absence of TAK779 treatment, on the 7- and 14-days (Fig. 3B, C, and C+TAK). The first difference observed in lymph nodes is the lower number of clusters with HIV-integrated DNA when compared to the spleen. In these lymph nodes, HIV-infection resulted in at least 50% of the human cell clusters having HIV-integrated DNA at both time points, 7 and 14 days, without any observed changes when treated with TAK779 (Fig. 3B, red bars). Also, like the results in spleen and lung (data not shown), expression of HIV-DNA, mRNA, and p24 in a control condition, with and without TAK, was even at 7- and 14-days post-infection (Fig. 3B, HIV). However, in lymph nodes obtained from HIV infected animals treated with TAK779 showed that viral mRNA was not altered, but HIV-p24 expression was reduced in 39.75±10.5 % as compared to HIV-infection alone (Fig. 3B, HIV+TAK, green bars). These are significant differences with the data obtained in the spleen, where TAK779 treatment reduced both viral mRNA and HIV-p24 clusters (compare Fig. 3A and 3B).

Furthermore, the analysis of brains obtained from uninfected animals indicated that no unspecific staining was detected in the presence or absence of TAK779 treatment for HIV DNA, mRNA, and HIV-p24 (Fig. 3C). In brains obtained from HIV-infected animals, an even number of clusters containing HIV-integrated DNA, viral RNA, and HIV-p24 were detected at 7 and 14 days (Fig. 3C), suggesting active replication. However, treatment of the HIV-infected animals with TAK779 increased the numbers of clusters with HIV-integrated DNA producing mRNA and HIV-p24 at 7 and 14 days (Fig. 3C, HIV+TAK, increased 28.71±7.98 %). Also, TAK779 treatment of HIV-infected animals did not reduce the expression or intensity of HIV mRNA and HIV-p24 at both time points analyzed (Fig. 3C, HIV+TAK779). Thus, TAK779, despite being effective in spleen/lung and lymph nodes, is ineffective and might enhance HIV spread within the brain. In conclusion, depending on the tissue analyzed, infection and response to ART may be different.

HIV-infection and TAK779 treatment induce tunneling nanotube (TNT)-like structures (TLS) in T cells and macrophages in a tissue-dependent manner.

Recently, our laboratory demonstrates that TNTs are induced during the early stages of HIV spread as well as during events of viral reactivation (31, 44–48); however, there are few demonstrations of TNTs in vivo. Only recently, a demonstration of TLS in vivo has been described during glioblastoma tissue colonization and tumor growth (29, 49–51). Overall, in HIV infection, TNTs formation and associated communication are expected to help the virus spread between communicated (TNT-linked) cells; however, the cargo and mechanisms are unknown. Analysis of the tissues described above revealed several TLS observed under the microscope in all the tissues examined. Thus, we decided to investigate further TLS formation in tissues with a focus on CD3⁺ and macrophages using tissue staining and confocal microscopy.

We and others have defined TNTs recently (52). Overall, our definition involves several key events, such as: First, TNTs connect two or more cells at a minimal distance of 30 μm. Second, in general, they are linear; Third, the TNT process can branch and reach distances up to 500 μm. Fourth, TNTs can transport organelles, vesicular structures, and small molecules between TNT connected cells. Fifth, TNTs are positive for actin and negative or poorly positive for tubulin (32, 53, 54). Lastly, TNTs are positive for several TNT markers not present in filopodia, including TTHY1, GAP43, and protein 14-3-3γ (55). Thus, to identify TNTs in vivo, we used protein TTHY1, 14-3-3γ, (both developmental) and GAP43 (except in the brain, due to the general observation that it is localized in neuronal growth cones). In addition, DAPI (nucleus) and CD3⁺ (T lymphocytes) or Iba-1 (macrophages) as well as HIV-DNA, RNA, and viral protein staining were used (Fig. 4).

Figure 4:

TLS proliferate in response to HIV-infection in CD3⁺ and macrophage cells in a tissue and replication specific manner. Critical questions until now is how HIV and TAK779 regulate the size of the HIV-infected cells in different tissues and what is the role in viral spread and adaptation to treatment. TLS were often observed in all tissues analyzed. (A-B) Quantification of TLS in lung and spleen, as defined in the results section, indicates that TLS in control tissues were low in CD3⁺ cells (A) and macrophages (B). HIV-infection increased the numbers of TLS in both cell types, and treatment with TAK779 prevented TLS formation and associated communication in both cell types. *p≤0.0058 as compared to uninfected conditions, #p≤0.0048 as compared to HIV infected conditions. (C-D) Quantification of TLS in lymph nodes. In control, tissues in the presence and absence of TAK779 low levels of TLS were observed. HIV-infection increased TLS formation in CD3⁺ cells (C) and macrophages (D). However, TAK779 treatment of HIV-infected animals did not alter the numbers of TLS in T cells, CD3⁺ cells, but increased TLS formation in macrophage cells within the lymph nodes. *p≤0.005 as compared to uninfected conditions, #p≤0.0031 as compared to HIV infected conditions. (E-F) Quantification of TLS in brain tissue sections. The numbers of CD3⁺ cells are minimal within the brain, and a few found to have low levels of TLS formation in all the conditions analyzed. In contrast, HIV-infection increased TLS formation in macrophage/microglia, and TAK779 promote further the formation of TLS between myeloid cells. *p≤0.005 as compared to uninfected conditions, #p≤0.003 as compared to HIV infected conditions. Thus, TLS formation depends on the tissue and cell type analyzed. Here, we used 128 animals to perform the showed experiments at 7 and 14 days in control conditions (28 animals), control+TAK779 (28 animals), HIV (36 animals), and HIV+TAK779 (36 animals). The differences in the animal numbers were due to the potential toxicity of the virus and/or TAK779 treatment. We did not lose any animals during the time course examined. 20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS)

In order to quantify TLS from the human cells in our model, we focused on cells that are positive for HLA-DR, in combination with cells that either positive for CD3 (a T cell marker), or Iba-1 (a macrophage marker) (Fig. 4A and B) in addition to viral mRNA for nef and HIV-p24 (Fig. 4). Analysis of spleen from uninfected animals in the presence and absence of TAK779 indicates that TLS are minimally present in CD3⁺ cells (Fig. 4A) or Iba-1 positive cells (Fig. 4B). However, upon HIV-infection, TLS proliferate at 7- and 14-days post-infection in CD3⁺ and Iba-1 positive cells within the spleen/lung (Fig. 4A and B, HIV). The use of TAK779 to reduce viral replication in HIV-infected animals also reduced the numbers of TLS in the spleen in CD3⁺ cells at both times examined (Fig. 4A and B, HIV+TAK). Interestingly, macrophages present in the spleen and lung of HIV-infected animals did not show any response to TAK779 treatment, and the TLS remained at similar levels compared to HIV infection alone condition (Fig. 4B, HIV+TAK). Thus, TNTs formation in tissue is cell-type-specific, but it is dependent on HIV infection and, at least in CD3⁺ cells, on active replication of the virus.

In lymph nodes, low levels of TLS structures are observed in uninfected conditions in CD3⁺ and macrophages in the presence and absence of TAK779 treatment (Fig. 4C and D, C and C+TAK). HIV-infection for 7- and 14-days increased TLS structures in both CD3⁺ and Iba-1 positive cells (Fig. 4C and D, respectively) in lymph nodes, as seen in the spleen (Fig. 4A and B). Treatment of the HIV-infected animals with TAK779 did not change the numbers of TLS in CD3⁺ cells after 7- or 14-days post-infection (Fig. 4C, compare HIV versus HIV+TAK). However, the analysis of macrophages (Iba-1 positive) in lymph nodes obtained from HIV-infected animals treated with TAK779 shows an increase in TLS at −7 and −14 days (Fig. 4D, HIV+TAK). Our data indicate that lymphocytes and macrophages in spleen and lymph nodes have a different TLS behavior that is independent of systemic replication, but tissue and cell-type dependent.

Further analysis of brain tissues (subcortical areas and cortex) obtained from uninfected and HIV-infected animals in the presence and absence of TAK779 treatment indicates overall that, there is a low number of CD3⁺ cells in the brain parenchyma, representing only 4.52±2.35 % of the total cell number (Fig. 4E). In comparison, in the spleen, most cells in the human cluster were CD3⁺ cells, and only 18.98±4.62 % of Iba-1 positive cells, macrophages. These CD3⁺ cells in the brain parenchyma had low levels of TLS under all conditions: control, HIV, and TAK treated conditions (Fig. 4E, C, C+TAK, HIV, and HIV+TAK). In contrast, when we examined the Iba-1 positive cells, HIV-infection for 7 and 14 days increased the numbers of TLS (Fig. 4F). Treatment of HIV-infected animals with TAK779 further increased the number of TLS as compared to HIV-infection alone (Fig. 4F), suggesting a response to therapy like lymph nodes (see Fig 4D and F).

Overall, under control conditions, low levels of TLS are detected in both cell types analyzed: CD3, and macrophages. However, our data under HIV-infected conditions indicates that CD3⁺ (lymphocytes) and Iba-1 (macrophages) had a different cell-to-cell communication behavior depending on the tissue where they reside. Furthermore, the formation of TLS was dependent on HIV-infection and the cell type analyzed. TAK779 treatment of HIV-infected animals resulted in the systemic reduction of HIV replication (see Fig. 2), but in lymph nodes and brain, this treatment increased cell-to-cell communication.

HIV-infection induces the formation of TLS that helps to spread HIV components between connected cells.

To perform these experiments, serial and thicker tissue sections (20-50 μm) were analyzed by confocal microscopy to reconstruct the clusters of HIV-infected cells in the tissue examined to enable the identification and reconstruction of human clusters as well as TNT generated from them into neighboring clusters. Staining for nuclei (DAPI, blue staining), HIV-p24 (green staining), viral RNA (nef mRNA, red), and HIV-DNA (nef sequences, white staining) was performed. Later, imaging analysis was used to identify the co-localization of HIV-DNA, viral RNA, and proteins in the spleen, lymph nodes, and brain tissue sections. The main objective of this approach is to identify, localize and reconstruct all viral components and the presence of TLS within the tissues analyzed with a better resolution and quantification as compared to Fig. 1, where positive and negative was the main readout.

No unspecific staining was detected in tissue sections obtained from spleen, lymph nodes, or brain for HIV-DNA, viral RNA, or proteins (Fig. 5A–E, brain tissue 3D reconstructions are shown). Early HIV-infection, 3 days post-infection, indicates an early infiltration into all the tissues examined, as described in Fig. 1. In all these tissues, a clear association between HIV-DNA and viral RNA was detected with low HIV-p24 expression (Fig. 5F–J). Also, most of the HIV-DNA and viral RNA expression colocalized, and HIV-p24 was not associated with the cells (64.5±9.85 % of the staining was not associated with cells containing HIV-DNA) (Fig. 5F–J).

Figure 5:

TLS containing HIV components (HIV-DNA, mRNA, and proteins) increase in response to HIV infection and the immune/viral pressure induced by TAK779 treatment. To evaluate in detail the mechanism of tissue and cell-specific viral replication, we examine serial sections of each tissue to evaluate TLS formation and viral component movement. (A-E) Lymph nodes tissue, stained for DAPI (nuclear staining, blue, A), HIV-p24 (green staining, B), HIV-nef mRNA (red staining, C), HIV-DNA (white staining, D), and merge of all colors indicated in the last row in control conditions in the presence and absence of TAK779 treatment. No unspecific staining was detected. (F-J) HIV infected conditions after 3 days post-exposure, as described in Fig. 1. Most viral components colocalized, except HIV-p24, that was secreted. (K-O) Representative image of HIV-infected clusters after 7 days post-infection where it is possible to observe TLS formation (indicated by arrows) and the first evidence of viral components such as viral proteins (L), mRNA (M), and unintegrated DNA (N) to diffuse into neighboring cells. (P-T) Clusters of human cells infected with HIV and treated with TAK779 as observed TLS become larger and contained several viral components inside of the TLS, such as HIV-p24, mRNA, and unintegrated HIV DNA (see arrows) despite that significant decrease in systemic viral replication. Thus, tissues and blood are different compartments with different immune and viral behaviors. (U-Y) Correspond to a representative example after 14 days post-infection and (Z-A4) correspond to 14 days of infection in the presence of TAK779 with a similar behavior than 7 days post-infection in the absence and presence of TAK779. (A5) Quantification of the co-localization between HIV-mRNA and HIV-p24 protein in the cluster and the TLS in different tissues. In HIV conditions, 7- and 14-days post-infection, HIV-DNA, and production of viral mRNA were similar; however, in the presence of TAK779, the co-localization decreased mainly due to the spread of these viral components via TLS. *p≤0.0058 as compared to HIV infection alone. (A6) Quantification of the co-localization between HIV-p24 and HIV mRNA in the cluster and the TLS in different tissues. In HIV conditions, 7 days, most HIV-p24 did not localize with the viral mRNA in spleen/lung and lymph nodes, but no in the brain. At 14 days post-infection, both viral products highly colocalized. However, the treatment of HIV-infected animals with TAK779 increased co-localization in spleen/lungs but not in the other tissues. *p≤0.001 as compared to HIV infection alone. Thus, TLS proliferate under HIV conditions and containing viral components that can reach distances up to 500 μm in vivo in all the tissues analyzed. Here, we used 128 animals to perform the showed experiments at 7 and 14 days in control conditions (28 animals), control+TAK779 (28 animals), HIV (36 animals), and HIV+TAK779 (36 animals). The differences in the animal numbers were due to the potential toxicity of the virus and/or TAK779 treatment. We did not lose any animals during the time course examined. 20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS). Bar: 20 μm.

At 7- day post-HIV-infection, the human cell cluster was better defined, HIV-DNA and viral RNA colocalized, and HIV-p24 protein was mostly surrounding the clusters containing HIV-components with 54.1±5.42 % of the p24 staining colocalizing with HIV-DNA or viral mRNA positive cells (Fig. 5K–O). In addition, small cellular projections from the clusters containing HIV-p24 (green staining), viral RNA (red staining), and viral DNA (white staining) no integrated (this HIV-DNA staining did not colocalize with DAPI or Alu-repeats as we recently described, (35, 36, 39)) was also observed (Fig. 5K–O, see white arrows in N). Treatment of HIV-infected animals with TAK779 promoted the formation of long TLS processes containing HIV-p24 (Fig. 5Q), viral mRNA (Fig. 5R), and unintegrated HIV-DNA (Fig. 5S, see arrows). Interestingly, these TLS in some cases, reach distances up to 500 μm to reach other human clusters, and importantly, no interaction with mouse cells was observed (Fig. 5P–T). The transfer of HIV-DNA via TNTs that did no colocalized with DAPI (nucleus) opens the possibility that infectious DNA or circles can be shared via TNTs. Thus, TNTs are present in tissues during the active spread of the virus in tissues as well as contain viral components to help to spread infection or inflammation into neighboring cells.

Analysis of tissues obtained from HIV-infected animals for 14 days indicates that the cluster of HIV-infected cells has a similar size and distribution to HIV-infected clusters at 7 days post-infection, but with a higher expression and expanded distribution of HIV-p24 (Fig. 5U–Y). Similar TLS were observed at 14 days, as indicated by arrows (Fig. 5X). Treatment of HIV-infected animals with TAK779 for 14 days shows that HIV clusters become bigger with TLS containing HIV-p24, viral RNA, and unintegrated HIV-DNA as described above (Fig. 5 Z–A4).

Quantification of the staining and 3D reconstructions described above using imaging software NIS Elements (Nikon), indicates that HIV-RNA and HIV-p24 co-localization in the cluster decreased with TAK779 treatment in a tissue-specific manner (Fig. 5A5). The inverted analysis of HIV-p24 and HIV-mRNA indicates that TAK-779 also dysregulates both co-localizations in a tissue and TAK779 dependent manner (Fig. 5A6). Overall, as indicated in our staining, HIV-p24 could diffuse further into areas no containing HIV-DNA. Thus, to determine whether HIV-p24 colocalizes with its mRNA, the reverse analysis was performed in Fig. 5A6. In spleen and lymph nodes, HIV-p24 minimally colocalized with HIV-RNA; however, most of this protein is surrounding the cluster of HIV infected cells (Fig. 5A6).

In contrast, in the brain, most of the HIV-p24 protein colocalized with HIV-RNA at 7 days post-infection (Fig. 5A6). Treatment of HIV-infected animals with TAK779 for 7 days increased the co-localization of HIV-p24 with HIV-RNA but did not change the baseline of co-localization in spleen and lymph nodes (Fig.5A6). HIV-infection for 14 days increased the colocation of both HIV markers in spleen and lymph nodes and maintained the co-localization values observed at 7 days in the brain. Furthermore, TAK779 treatment did not change the co-localization of HIV-p24 with HIV RNA (Fig. 5A6). No co-localization or significant background was found in uninfected tissues (Fig. 5A5 or A6, t=0 time). These data indicate that HIV replication in various tissues is different and the response to TAK779 treatment varies from tissue to tissue. It also shows that some antiretrovirals may increase the spread of HIV within the tissue area through the increasing of viral components inside, probably as an escape mechanism of the HIV-infected cells to the effective TAK779 systemic treatment.

Examination of the spread of HIV-components into neighboring cells using imaging analysis.

To address the potential lack of correlation between HIV-DNA, HIV-RNA, and HIV-p24 in tissues versus the effective systemic decrease in HIV-replication in the whole animal, we took advantage of the highly concentrated clusters of human cells containing HIV-DNA in the tissues analyzed. Thus, serial tissue sections (20-30 sections per tissue with the objectives to maintain a constant number of clusters and to identify and reconstruct TLS) were analyzed to generate a map of the clusters of HIV-infected cells and the “diffusion” of HIV-DNA, viral mRNA, and HIV-p24 as a function of the distance from cells containing HIV-integrated DNA. Human cell clusters (HLA-DR positive) with HIV-DNA were considered the zero-distance, as described in Fig. 6, due to integrated HIV-DNA was not altered during the time course of HIV-infection as well as the TAK779 treatment. Data were expressed as the percentage of pixels versus distance from HIV-DNA positive clusters to either side of the cluster.

Analysis of spleen indicates that clusters containing HIV-DNA expand from the center of the cluster into ~40 μm to either side (Fig. 6A, blue line). Treatment of HIV-infected animals with TAK779 increased the size of the cluster up to ~60 μm (Fig. 6A, green line). At 14 days post-infection, the cluster size reached up to ~60 μm (Fig. 6A, pink line), and TAK779 treatment increased the radius of the HIV-DNA containing cluster to ~80 μm (Fig. 6A, ochre line) (all curves are significant p≤0.05 as analyzed by ANOVA to compare the different groups). These data indicate that TAK779 promoted the growth of the HIV-infected cluster despite that TAK779 effectively reduced systemic replication. Surprisingly, the analysis of viral RNA was only significant when HIV was compared to HIV plus TAK779, and the radius of cells expressing viral RNA was smaller than the cells containing HIV-DNA reaching up to ~50 μm (Fig. 6B).

In contrast, HIV-p24 expression and spread were limited, but the degree of expression and spread was responsible for the systemic reduction in replication induced by TAK779 treatment (Fig. 6C). Overall, our data indicate that TAK779 negatively regulates the size of the HIV-infected clusters, but strongly reduced HIV-p24 expression. The regulation of the viral ratio among DNA/mRNA/protein in the spleen is novel and indicates that viral reservoirs can respond to effective ART.

Analysis of lymph nodes indicates that the HIV-DNA cluster size was around ~30 μm at 7 days post-infection and then increased at 14 days to ~50 μm. However, upon the treatment of HIV-infected animals with TAK779, the radius of the cells containing HIV-DNA increased up to ~80 μm (Fig. 6D). In contrast to the spleen, viral RNA expression was highly localized at the center of the cluster of HIV-infected cells, reaching only ~30 μm from the center (Fig. 6E) in the lymph nodes. However, upon treatment with TAK779, at both time points (7- and 14-days post-infection), viral RNA radius was up to ~80 μm in a manner similar to HIV-DNA (Fig. 6E). The HIV-p24 expression only spread up to ~40 μm with no changes in levels or radius of expression in response to TAK779 treatment (Fig. 6F). Thus, the response of the cluster of HIV-infected cells in the lymph nodes was different from the response observed in the spleen and to TAK779 treatment.

Analysis of brain samples indicates HIV-infected clusters are smaller than clusters in spleen and lymph nodes reaching ~30-40 μm from the center of the cluster (Fig. 6G). Treatment with TAK779 also increased the radius of cells containing HIV-DNA, indicating that HIV-infected cells have an adaptive behavior in response to decreased systemic HIV-replication by TAK779 treatment (Fig. 6G). However, the increased radius, despite been significant, was not so noticeable like the radius observed in both the spleen and lymph nodes (Fig. 6G). HIV RNA expression and distribution were similar to HIV-DNA distribution, and we observed an increased radius of HIV RNA expression at 7- and 14-days post-infection in the presence of TAK779 treatment (Fig. 6H). In contrast, HIV-p24 expression was high in HIV only condition, and the radius of HIV-p24 expression, was higher in the HIV-infected animals treated with TAK779 as compared with HIV-infection alone (Fig. 6I), indicating that despite the reduction in systemic HIV-replication, the release of HIV-p24 is still ongoing in the brain irrespective of the presence of effective systemic ART. Furthermore, our data indicate that each organ has a different escape mechanism to TAK779, suggesting that each tissue responds in its independent manner. Finally, HIV-infected cells adapt to treatment, increasing the size and extension of infection, probably in response to the low systemic replication achieved by TAK779 treatment.

Discussion

In the present study, using a humanized animal model, we identified that HIV-tissue colonization is an early event in the spleen, lungs, lymph nodes, and brain. Interestingly, the treatment of HIV-infected mice with TAK779 reduced systemic viral replication and prevented immune system compromise; however, it also modulates viral replication in tissues. Overall, we identified 3 different types of tissues based on viral seeding, local viral replication, TLS formation, and viral spread: First, a systemic kind of tissue that behaves similar to the viral replication in the blood; Second, tissues with mixed behavior such as lymph nodes (that partially follow blood replication patterns) and Lastly, tissues as the brain that behave in a completely different manner as compared to the blood. Thus, tissue-associated HIV-infected cells are cell type and tissue-specific.

Often, we monitor viral replication and examined viral reservoirs by a systemic methodology or based derived products (56–59). However, it is accepted that HIV can establish latent viral reservoirs in anatomically sequestered sites and it has been proposed that there is the persistence of replication-competent cells in the different cell types, including T cells and myeloid cells (60, 61). However, the role of resident viral reservoirs is currently still under investigation. In addition, most data are obtained from postmortem tissues or small samples from biopsies or tissue donations, which provides us a small insufficient window to observe HIV behavior in tissues (36, 39, 62, 63). Several studies showed that lymphoid tissues such as the spleen, thymus, and gut-associated lymphoid tissues, as well as compartmentalized tissues such as the brain, have viral reservoirs, and viral DNA can be recovered (63–65). Currently, the most examined HIV-reservoirs are different populations of lymphocytes that circulate in the blood, such as CD4⁺ T lymphocytes (9, 66, 67). However, it has become evident that blood is a poor representation of the actual presence of viral reservoirs in tissues. The circulation of blood and lymph, as well as their frequent contact with many tissues, gives a broad landscape of viral infection but does not distinguish between different populations of the virus being released from isolated tissue compartments then coming into the blood (8). The differences between HIV-replication in the circulation and the behavior of the HIV-infected cells within different tissues only reinforce our knowledge that T cells and myeloid cells can be HIV infected and become viral reservoirs. Equally important, their cellular and viral behavior is flexible, depending on the local and systemic environment. Thus, it is likely that depending on the tissue and cellular barriers present in every tissue as well as blood supply, the behavior of the viral reservoir can be different.

For example, macrophages are capable of self-renewal during homeostatic conditions, and the persistence of HIV in them can be compared to viral persistence in the long-lived memory T cells (36, 68–70). The ability of HIV to infect both T cells and macrophages allows the virus the potential to establish reservoirs in both cell types. This provides benefits for HIV to establish reservoirs in macrophages with long half-lives that can sustain viremia for several months just like in T cells, as was explained recently in mice and SIV infected macaques study (62).

In addition, myeloid cells are relatively more resistant to apoptosis induced by HIV infection, and viruses produced by macrophages are mainly CCR5-tropic viruses, which is generally accepted as the initial virus strain that establishes infection and the most abundant during the early phase of HIV pathogenesis (71). Monocytes and macrophages are chemotactic cells and are capable of infiltrating into most tissues in the body; therefore, they have the widest capacity for mediating HIV spread. Only recently using similar techniques to the ones used in this report, identification of viral reservoirs in tissues such as urethral and bone marrow support our findings of heterogeneous viral reservoirs in human tissues even in the ART era (72–74). As a result, these anatomical sites can be challenging to access with conventional ART, and it is possible that local replication or secretion of viral proteins still occurs. It is, therefore, highly important to consider the contribution of non CD4⁺ T lymphocytes as reservoirs and sources of HIV viremia. Our data indicate that each organ has a different response to HIV-infection and ART. These new variables must be considered in early interventions as well as future curative approaches.

Several groups have suggested that early ART intervention can limit the formation of viral reservoirs and prevent viral spread (75–77); however, our experiments in the humanized mice indicate that the underlying mechanism for this initial seeding and establishment of viral pools are early and complex due to the cell and tissue dependency. In agreement with the complexities in early infection processes, multiple examples of early ART interventions like in the case of the “Mississippi baby”; the infant who began receiving ART ~30 h after birth. Specifically, the mother tested positive for HIV during labor, but therapies to prevent transmission of the virus could not be administered before the child was born. While the administration of early ART could not prevent HIV infection, it was successful in inhibiting viremia. However, upon ART cessation after 12 months of therapy, the authors observed viral rebound in the Mississippi baby, suggesting that ART intervention may not be sufficient to prevent seeding as well as the establishment of viral reservoirs. Hence there is a need to understand the complexities involved in seeding of HIV, the establishment of viral reservoirs, as well as the associated biology and virology (78).

Despite the complexities mentioned above, viral tissue colonization appears, and it is accepted to be uniform. Our data demonstrate that there are significant differences in cellular and tissue responses to HIV-infection as well as to successful ART. We propose that each anatomical compartment may establish its distinct viral reservoir that undergoes a different degree of isolation, viral evolution, and subsequent genetic diversity (79–81). These differences can emerge either at the onset of tissue compartment infection as a result of the following factors: differential viral entry, cell tropism, blood versus lymph, and the presence of physical barriers such as the blood-brain barrier (BBB), which can results in viral genetic drift or selection during viral evolution (11, 12, 82). These factors result in genetically and functionally different subpopulations of HIV in unconventional reservoirs such as macrophages and other CNS cells. This heterogeneity of HIV supports the idea that diverse variants of HIV can cumulatively affect virulence and response to antiviral therapy (83–85). Our data support this idea and indicates that ART results in differential responses depending on the tissue and cell type analyzed. Thus, local mechanisms of HIV spread, and enhanced tissue colonization is in place to perpetuate the virus in tissues when systemic replication is either low or undetectable. This data is novel and indicates that there needs to understand better the establishment and stabilization of viral reservoirs in multiple tissues to generate an effective cure.

We demonstrated that TLS proliferate in vivo in response to HIV, and antiretroviral treatment may further enhance their formation and communication. Our in vitro experiments demonstrate that, after CD4 and CCR5/CXCR4 mediated HIV entry, TLS contributes to cell-to-cell communication between HIV-infected cells and their surrounding uninfected cells (31, 47, 48). TLS transport host and viral components between communicated cells, but their role in vivo is still a matter of debate. Currently, the only known mechanism of direct cell-to-cell transfer of HIV is via specialized virological synapses (86, 87). HIV virologic synapses and TLS established between T cells and other cell types mediate direct translocation of large amounts of HIV antigens and proteins into the recipient target cell to facilitate infection (88, 89). Several groups have calculated that such cell-to-cell transmission is up to 18,000-fold more efficient than for cell-free viruses (88, 90). This highly efficient mode of transmission of HIV virulent factors shields the infectious agent from neutralizing antibodies and phagocytosis, allowing the infected cells to concentrate and transmit multiple viral genomes, reaching 10²-10³ copies in a single synapse or TNT (88, 91). This is becoming an even more important issue in our data where the systemic viral infection is extremely low, yet local transmission in tissues continue to occur. Thus, we present in this study that TLS could increase HIV concentration and circumvent anti-retroviral treatments, thus providing a sanctuary for local replication, mutations, and resistance. Based on the results presented here, we propose that TLS can provide an additional route of rapid transfer of HIV proteins or particles from donor HIV infected macrophages to uninfected recipient cells, efficiently amplifying infection. Similar mechanisms of TNT-mediated cell-to-cell amplification of disease have been observed in cancer and neurodegenerative disorders (e.g., Alzheimer’s and Parkinson’s disease where aggregated proteins are transferred) (48, 92–97).

The identification of TNTs in vivo is significant not only because we observe a novel long-range communication system used for HIV to spread in tissues, but we also identify, in vivo, several types of TNTs. We identify TNTs that are involved in a variety of functions such as the transport of only HIV-p24, others with unintegrated HIV-DNA, and others with only viral RNA. These different observations of variable classes of TNTs suggest that, as described in vitro (31, 47, 48), they all have different permeabilities for different cargo to be trafficked. However, a limitation of our study is the lack of specific blockers to prevent TNT formation in animal models. Currently, most TNT blockers target actin or proteins required for TNT formation but have never been used in animal models that involve experiments for days. Thus, more data is required to explore whether TNT formation induced by HIV infection help to spread infection and/or inflammation. Our data indicate that like tissue HIV replication colonization, TNTs are heterogeneous in structure but also in cargo transported. In our analysis, TNTs were selective to transport HIV-DNA, mRNA, and viral proteins suggesting a cutoff or specialization poorly understood in vitro and now in vivo. Clearly, TNT formation is induced by HIV infected cells, but surprisingly, HIV-systemic replication is poorly associated with TNT formation. More important, ART treatment, TAK779, resulted in decreased systemic replication, increased TNT formation, suggesting that viral reservoirs have a mechanism of preservation in different tissues.

Only recently, we reported TNTs in breast and human glioblastoma tissues (98). Here we identify that human TNTs are rich in GAP43 and 14-3-3γ protein and that upon glioblastoma treatment with temozolomide and radiation, tumor cells form TNTs with neighboring cells to spread a critical DNA repair enzyme for the survival of tumors, MGMT (O[6]-methylguanine-DNA methyltransferase), to prevent tumor apoptosis (98). Thus, based on our recent publication and the current report, we propose that pathogenic cell types such as viral reservoirs or glioblastoma stem cells, use TNTs to spread toxic signals into neighboring cells but also use TNTs to survive and perpetuate disease. TNTs are minimally expressed in healthy conditions; thus, targeting TNTs can provide an adjuvant treatment to prevent viral reservoir escape mechanism in different tissues.

In conclusion, our data demonstrate that HIV-infection, tissue colonization, local tissue replication, and generation of viral reservoirs are tissue-specific. Effective systemic ART contributes to the plasticity of HIV-infected cells depending on the cell type and the tissue analyzed. HIV infected cells in tissues are extremely flexible to adapt to systemic viral and immune changes not reported before. Thus, a better understanding of these tissue compartments is required to potentially blocking TNT-mediated cell-to-cell communications, to reduces viral tissue and viral reservoir escape in different tissues, and to create new therapeutic interventions aimed at achieving HIV cure.

Highlights.

• HIV seeding occurs early in different tissues and anatomical sites.

• HIV tissue spread in a cell type, tissue, and ART-dependent

• HIV to infect neighboring cells via tunneling nanotubes-like structures.