R5 HIV-1 envelope attracts dendritic cells to cross the human intestinal epithelium and sample luminal virions via engagement of the CCR5

Abstract

The gastrointestinal tract is a principal route of entry and site of persistence of human immunodeficiency virus type 1 (HIV-1). The intestinal mucosa, being rich of cells that are the main target of the virus, represents a primary site of viral replication and CD4(+) T-cell depletion. Here, we show both in vitro and ex vivo that HIV-1 of R5 but not X4 phenotype is capable of selectively triggering dendritic cells (DCs) to migrate within 30 min between intestinal epithelial cells to sample virions and transfer infection to target cells. The engagement of the chemokine receptor 5 on DCs and the viral envelope, regardless of the genetic subtype, drive DC migration. Viruses penetrating through transient opening of the tight junctions likely create a paracellular gradient to attract DCs. The formation of junctions with epithelial cells may initiate a haptotactic process of DCs and at the same time favour cell-to-cell viral transmission. Our findings indicate that HIV-1 translocation across the intestinal mucosa occurs through the selective engagement of DCs by R5 viruses, and may guide the design of new prevention strategies.

Figure 1

R5 but not X4 HIV-1 induces DCs to migrate through a monolayer of epithelial cells. Caco-2 cells were grown on transwell filter to form a confluent monolayer then DCs were let to adhere to the bottom of the filter. Cell-free HIV-1, LPS (1 µg/ml) or medium (DMEM 10% FCS) were incubated on the apical side of the Caco-2 monolayer for 1.5 h. Filters were processed for CM. A–F. Three-dimensional rendering of representative fields obtained with Volocity 5.0 software. Mouse anti-human E-Cadherin + Alexafluor488 goat anti-mouse IgG2a (green) depicts interepithelial adherent junctions. Mouse anti-human DC-SIGN-PE (red) labels the DCs, which creep through epithelial cells in response to incubation with R5 HIV-1^J6363 (A, at 1 ng of p24 Ag) and HIV-1^AD8 (B, at 24 ng of p24 Ag) but do not in response to X4 HIV-1^IIIB (D, at 10 ng of p24 Ag) and HIV-1^pNL4.3 (E, at 40 ng of p24 Ag). LPS (C) and medium (F) were positive and negative controls. The experiment was repeated three times. G,H. Quantitative analysis of DCs migration across the Caco-2 cell monolayer at the apical (G) and medial (H) level of the cell layer is shown. Results are expressed as percentage of area occupied by DCs compared to that of the whole field. Bars represent mean ± SD of three or four fields of three different experiments. Statistic analysis was performed as described in Materials and Methods Section. *p < 0.05.

Figure 2

DCs migration does not alter junctional protein expression by Caco-2 cells and preserves the integrity of the HIV-1 treated monolayer. The Caco-2/DCs co-culture system was incubated with medium or R5 HIV-1^AD8 at 24 ng of p24 Ag for 1.5 h. A–D. CM cross-sectional images of specimens stained for DCSIGN-PE (red) and for the epithelial junctions (green) JAM (A), E-Cadherin (B) Occludin (C) ZO-1 (D) showed that HIV-1 as well as migrated DCs did not affect the intraepithelial junctions expression in the Caco-2/DC system. DAPI stained the nuclei. Results are from one representative experiment out of three. E. Ultrastructure of Caco-2/DCs culture treated with R5 HIV-1^J6363 (at 1 ng of p24 Ag) for 1.5 h. TEM images show DCs adhering to the filter (*), DCs protrusions inside the filter membrane pore (**) and a DC interposed between adjacent Caco-2 cells (***) Scale bar: 2 µm. Numbers identify the corresponding magnified images displayed in panel (E), and arrows point to interepithelial TJs (1, 2), TJs-like structures between DCs and Caco-2 cells (3–5), contiguity among cells inside the pore (6), and junction like-structures between DCs (7) Scale bar: 500 nm. Results from a representative experiment out of three are shown. F. Semi-thin sections for TEM labelled with Toluidine blue and the corresponding explicative colour mask below (Caco-2 cells red, DCs blue, filter grey) show the morphology and the spatial organization of cells. Caco-2 cells are columnar and polarized, displaying microvilli, dense cytoplasmic granules and vacuoles characteristic of epithelial cells. DCs are disposed along the lower face of the filter, inside the membrane pores and, in HIV-1 treated sample, intercalated in between Caco-2 cells, without destroying epithelial monolayer continuity. G. The permeability of the Caco-2/DCs culture to FD4 (4 kDa, 250 µg/ml) in the presence of medium, R5 HIV-1^AD8 (at 24 ng of p24 Ag) and CN54 gp140 protein (100 ng/ml) is shown as percentage of the positive control (i.e. FD4 added in the upper chamber of the transwell without Caco-2 cells). Triton X100 was included as control of barrier disruption. Results are mean values ± SD of triplicates from a representative experiment out of three.

Figure 3

DCs migration across a Caco-2 monolayer is reversible. Caco-2 cells were grown on transwell filter to form a confluent monolayer then DCs were let to adhere to the bottom of the filter. Cell-free R5 HIV-1^AD8 or medium were incubated on the apical side of the Caco-2 monolayer for 2 h (Apical stimulus). After extensive wash, samples were either immediately fixed with 2% PFA, or the transwell moved in a new plate containing either medium or fractalkine (FRK, 100 ng/ml) in the basal chamber and further incubated for 3 h before fixation (Basal stimulus). Quantitative analysis of DCs migration across the Caco-2 cell monolayer at the apical (A) and medial (B) level of the cell layer is shown. Results are the mean of three replicates each and are expressed as percentage of the positive control (i.e. area occupied by DCs following apical treatment with HIV-1). n.d., not done.

Figure 4

R5 HIV-1 induces migration of DCs through the colonic epithelium. Colonic tissue was either left untreated (A and B) or incubated with X4 HIV-1^pNL4.3 (C), R5 HIV-1^AD8 (D) or R5 HIV-1^J6363 (E–J) (at 50 ng of p24 Ag) for 30 min. CD11c⁺ cells were detected in the colonic lamina propria of untreated (B) and HIV-1^pNL4.3 treated (C) tissues but not in between epithelial cells. Following R5 HIV-1 incubation protruded dendrites (D) or whole DCs (E–H) were observed inside the epithelium. Moreover DCs (DCSIGN⁺/CD68⁻ cells) but not Mϕ (DCSIGN⁺/CD68⁺ and DCSIGN^-CD68⁺ cells) migrated through the epithelium upon HIV-1 stimulation (I and J). Cryosections were fixed with 4% PFA and immunostained with Hematoxylin–eosin (A), or for human epithelial antigen (mouse anti-HEA-FITC, green), DAPI (nuclei; blue) and either mouse anti-human CD11c + Alexafluor594 goat anti-mouse IgG (DCs; red) (B–H) or mouse anti-human DC-SIGN-PE (DCs and Mϕ; red) and mouse anti-human CD68⁺ Alexafluor488 goat anti-mouse IgG1k (monocytes/Mϕ; green) (I and J). The inset in panel (D) and (F) is magnified 3×. Panels (F), (H) and (J) are magnification (zoom 3×) of the area indicated by arrow in (E), (G) and (I), respectively. The inset of panel (F) evidences DCs after hiding of HEA channel. Scale bar: 50 µm in B–E, G and I. Each Figure is representative of results from five donor tissues.

Figure 5

DCs capture HIV-1 in colonic explant and in vitro Caco-2/DC model. Human colonic tissue was either left untreated (A) or incubated with R5 HIV-1^J6363 (B) or R5 HIV-1^AD8 (C and D) (at 50 ng of p24 Ag) for 30 min. HIV-1 virions co-localized with CD11c⁺ DCs that migrate inside the epithelium as well as with DCs closely underlying the epithelium (see white arrows in B and C). Moreover virions were detected at both the apical and basal side as well as penetrating the epithelium. Yellow arrows point to virions entrapped in the mucus. Absence of p24 Ag in the basal medium, determined with ELISA, confirmed the seal integrity of the tissue culture system without any viral leakage (as described in Materials and Methods Section). Panel (D) is a magnification (zoom 3×) of the boxed area in C. Cryosections were fixed with 4% PFA and stained with mouse anti-human CD11c-PE (DCs; red), mouse anti-p24 + Alexafluor488 goat anti-mouse IgG (HIV-1; green), and DAPI (nuclei; blue). Scale bar indicate the magnifications. Each Figure is representative of results from three donor tissues. Caco-2/DCs co-culture was incubated with R5 HIV-1^AD8-GFP (at 20 ng of p24 Ag) for 30 min (E), 1.5 h (F) or 4.5 h (G). (A and B) Three-dimensional renderings from CM z-series stacks of representative images (acquired with objective 40×, zoom 6.5×) showed GFP-expressing virions (green, indicated by arrows) either on migrated (E) as well as non-migrated DCs at the basal side (F, the bottom side of the filter is shown). Cells were stained for DAPI (all cells; blue) and mouse anti-human CD11c + AlexaFluor594 goat anti-mouse IgG (DCs; red). (G) Three-dimensional rendering from a CM z-series stack of images showed a cluster of DCs (visualized with mouse anti-human DC-SIGN-PE; red) migrated to the Caco-2 side of the culture (Caco-2 cells not shown in the Figure) after incubation with HIV-1 ^AD8–GFP. GFP-expressing virions (yellow) were associated to migrated DCs.

Figure 6

DC migration across the epithelial monolayer in response to gp140 protein of different HIV-1 subtypes. A,B. The monolayer of the Caco-2/DCs system was incubated for 1.5 h with decreasing dose of R5 CN54 gp140 trimeric protein. Medium and R5 HIV-1^AD8 were used as negative and positive control, respectively. Bar charts represent quantitative analysis of DCs migration across the Caco-2 cell at the apical (A) and medial (B) level of the cell monolayer (as in Fig 1). Results are expressed as percentage of area occupied by DCs compared to that of the whole field. Bars represent mean ± SD of three different fields of three different experiments. *p < 0.05. C–F. DC migration was induced with gp140 protein (100 ng/ml) of 92UG037 (C), 93BR029 (D), and TV1 (E) but not with that of SF162 (F). Representative three-dimensional renderings from CM z-series stacks of the Caco-2 and DCs on filter stained for DAPI (epithelial cells and DCs; blue) and mouse anti-human DC-SIGN-PE (DCs; red) are shown. Results are those from one representative experiment out of three.

Figure 7

DCs migration is dependent from the viral envelope. In the absence of the HIV-1 env as well as of the V1V3 env region, DCs did not migrate across the Caco-2 monolayer (A and B). The V1V3 env region of an R5 virus completely restored the migratory ability of DCs (C). A–C. The Caco-2/DCs co-culture was incubated apically for 1.5 h with HIV-1 deleted of the env (HIV-1^Δenv, 6 ng of p24 Ag) (A), the HIV-1^43ΔV deleted of the V1V3 region (40 ng of p24 Ag) as negative control (B) and a recombinant HIV-1 with the V1V3 env region of the R5 HIV-1^J6363 primary virus recombined with the HIV-1^43ΔV backbone (HIV-1^J6363-43ΔV, 20 ng of p24 Ag) (C). Three-dimensional reconstructions from CM z-series image stacks of the Caco-2/DCs culture stained for DAPI (epithelial cells and DCs; blue) and mouse anti-human DC-SIGN-PE (DCs; red) are shown. D,E. The monolayer of the Caco-2/DCs system was incubated for 1.5 h with medium (negative control), HIV-1^YU2 (positive control) or YU2 gp120 protein (100 ng/ml). Bar charts represent quantitative analysis of DCs migration across the Caco-2 cell at the apical (D) and medial (E) level of the cell monolayer (as in Fig 1). Results are expressed as percentage of area occupied by DCs compared to that of the whole field. Bars represent mean ± SD of three different fields of three different experiments. Statistic analysis was performed as described in Materials and Methods Section. *p < 0.05.

Figure 8

Engagement of CCR5 is necessary to trigger migration of DCs across the epithelium. A–C. DCs were left untreated (A) or were incubated for 4 h during attachment to the filter with CCL5 (200 ng/ml) (B) or Maraviroc (100 ng/ml) (C), and thereafter R5 HIV-1^J6363 was incubated apically on the Caco-2 cells monolayer for 1.5 h. Representative three-dimensional rendering from CM z-series stacks of the Caco-2/DCs culture stained with mouse anti-human JAM + Alexafluor 488 goat anti-mouse IgG (epithelial cells; green) and mouse anti-human DC-SIGN-PE (DCs; red) is shown. Results are of one representative experiment out of three. D,E. Quantitative analysis of DCs migration across the Caco-2 cell monolayer at the apical (D) and medial (E) level of the cell layer revealed complete inhibition of DCs migration, induced by the virus, following CCL5 or Maraviroc treatment. Results are expressed as percentage of area occupied by DCs compared to that of the whole field. Bars represent mean ± SD of three different fields of three different experiments. *p < 0.05. In the negative control the Caco-2 cells monolayer was stimulated with medium only and DCs were left untreated (bar on the left in D and E).

Figure 9

HIV-1 penetrates within and in between epithelial cells. A. Virions (stained with mouse anti-p24 + Alexafluor488 goat anti-mouse IgG, green) were mainly localized at the apical surface of Caco-2 cells at 30 min but also inside the cytoplasm at 90 min of incubation with R5 HIV-1^AD8 (20 ng of p24 Ag). Shown are CM single plane cross sectional images of a Caco-2 monolayer (rabbit anti-human Occludin + Alexafluor594 goat anti-rabbit IgG; red) taken at apical and medial level of the cell layer. B. R5 HIV-1^AD8 (red; visualized with human anti-gp120 monoclonal antibody 2G12 + Alexafluor594 goat anti-human IgG) localized inside the cytoplasm of epithelial cells and in intrajunctional spaces (indicated by arrows). Shown are four single plane cross sectional images, taken from the apical to the medial plane along the z-axis of the Caco-2 monolayer (mouse anti-human E-Cadherin + Alexafluor488 rabbit anti-mouse IgG2a; green) incubated with R5 HIV-1^J6363 (at 5 ng of p24 Ag) for 90 min. Results are of one representative experiment out of three.

Figure 10

Transcytosis of cell-free HIV-1 through a tight monolayer of Caco-2 cells. A–E. R5 HIV-1^J6363 (A, at 1 ng of p24 Ag), X4 HIV-1^IIIB (B, at 10 ng of p24 Ag), R5 93BR029 gp140 protein (C, at 100 ng/ml), R5 HIV-1^SF162 (D, at 2 ng of p24 Ag), or medium (E), were added to the Caco-2/DCs cultures for 90 min and immunostained for CM analysis. HIV-1 (red; visualized with human anti-gp120 monoclonal antibody 2G12 + Alexafluor594 goat-anti-human IgG) was dispersed throughout the cytoplasm but mainly concentrated in the upper part of the Caco-2 cells (visualized with mouse anti-human JAM + Alexafluor 488 goat-anti-mouse IgG; green). DCs are not shown in the Figure. Transversal xy- and xz-plane visualization from representative fields of the monolayer were obtained with Volocity 5.0 software. 1 unit = 10.3 µm. Results show one representative experiment out of three. F. Transcytosis of HIV-1 R5 and X4 is comparable. The amount of transcytosed virus was evaluated measuring with ELISA the p24 Ag released in the basal chamber after 2.5 h of incubation with cell-free HIV-1 (at 20 ng of p24 Ag), either R5 HIV-1^AD8 or X4 HIV-1^pNL4.3, on the apical side of the Caco-2 monolayer cultured with or without DCs adherent to the filter. Results are expressed as mean ± SD of triplicates of three different experiments.

Figure 11

DCs capture virus and transfer infection to target cells. A,B. Caco-2/DCs system was apically incubated for 1.5 h with R5 HIV-1^AD8 at 10 or 40 ng (A) or 20 ng of p24 Ag (B). PBMCs (5 × 10⁵ cells) directly incubated with the same amount of input virus served as positive control (data not shown). (A) R5 virus does efficiently replicated when DCs were cultured with PBMCs (1 × 10⁶). Low levels of p24 Ag were observed when DCs were cultured alone (inset). (B) DCs transmit infectious virus in the long-term. PBMCs were added to detached DCs immediately or after 3 or 4 days. Results from a representative experiment out of three are expressed as mean of p24 Ag values of triplicate cultures ± SD. C. DCs transferred both R5 and X4 viruses to PBMCs. Caco-2 cells were treated with R5 HIV-1^AD8 (24 ng/ml) and X4 HIV-1^pNL4.3 (40 ng/ml) as described above. DCs collected from transwell were co-cultured with PHA-activated PBMCs. Mean ± SD of three different experiments performed in triplicate is shown. *p < 0.05.

Figure 12

Schematized proposed mechanism of HIV-1 penetration across the intestinal epithelium and DCs recruitment. HIV virions/gp120 env protein that interact with the apical surface of intestinal epithelial cells (ECs) can transiently open the TJs of EC (1) or enter into the EC and be released from the cytoplasm to the intercellular space (2) to create a gradient between adjacent EC and reach subepithelial DCs. Chemokines secreted by EC might possibly accumulate locally. The interaction between the env and the CCR5 receptor expressed on DCs induces within 30 min the elongation of DC's dendrites and migration of DCs across the epithelium to further capture luminal R5 virions (3). The process is inhibited by CCR5 inhibitors and anti-JAM antibodies (Abs) (4). Alternatively, virions can be transcytosed through ECs (5) and be captured by lamina propria cells. Migrated DCs can shuttle across the barrier and return to the lamina propria when appropriately stimulated. The captured virus is transmitted from DC to CD4⁺ T cells (6) or directly transported to lymph nodes to further spread the infection.