Multi-functional plasmacytoid dendritic cells redistribute to gut tissues during simian immunodeficiency virus infection

Authors


Correspondence: R. Keith Reeves, Division of Immunology, New England Primate Research Center, Harvard Medical School, One Pine Hill Drive, Southborough, MA 01772-9102, USA. E-mail: roger_reeves@hms.harvard.edu

Senior author: R. Keith Reeves

Summary

The objective of this study was to determine the systemic effects of chronic simian immunodeficiency virus (SIV) infection on plasmacytoid dendritic cells (pDCs). pDCs play a critical role in antiviral immunity, but current data are conflicting on whether pDCs inhibit HIV/SIV replication, or, alternatively, contribute to chronic immune activation and disease. Furthermore, previous pDC studies have been complicated by incomplete descriptions of generalized depletion during HIV/SIV infection, and the effects of infection on pDCs outside peripheral blood remain unclear. In scheduled-sacrifice studies of naive and chronically SIV-infected rhesus macaques we evaluated the distribution and functionality of pDCs in multiple tissues using surface and intracellular polychromatic flow cytometry. As previously observed, pDCs were reduced in peripheral blood and spleens, but were also depleted in non-lymphoid organs such as the liver. Interestingly, pDCs accumulated up to fourfold in jejunum, colon and gut-draining lymph nodes, but not in peripheral lymph nodes. Most unexpectedly, SIV infection induced a multi-functional interferon-α, tumour necrosis factor-α, and macrophage inflammatory protein-1β cytokine secretion phenotype, whereas in normal animals these were generally distinct and separate functions. Herein we show a systemic redistribution of pDCs to gut tissues and gut-draining lymph nodes during chronic SIV infection, coupled to a novel multi-functional cytokine-producing phenotype. While pDC accumulation in the mucosa could aid in virus control, over-production of cytokines from these cells could also contribute to the increased immune activation in the gut mucosa commonly associated with progressive lentivirus infections.

Introduction

Plasmacytoid dendritic cells (pDCs), also referred to as natural-interferon-producing cells, respond to microbial pathogens by rapidly producing interferon-α (IFN-α), as well as other pro-inflammatory cytokines such as tumour necrosis factor-α (TNF-α) and macrophage inflammatory protein-1β (MIP-1β) initiating a cascade of both innate and adaptive immune responses. In normal infections, the pDC response rapidly resolves, but recent evidence suggests that during pathogenic HIV/SIV infections pDCs are continually stimulated,[1, 2] resulting in over-production of cytokines. However, despite their potent antimicrobial role, pDCs are also directly implicated in suppression of T helper type 17 cells and innate lymphoid cells, and promoting generalized apoptosis and chronic inflammation.[3-9]

As first reported over a decade ago, multiple groups have verified that pDC numbers are severely reduced in blood and lymph nodes during HIV infection.[2, 10-12] Furthermore, loss of pDCs begins during primary infection, is associated with increasing viral loads, and is only partially reversible by highly active anti-retroviral therapy. A comparable loss of pDCs has been demonstrated during acute and chronic simian immunodeficiency virus (SIV) infection of rhesus, pig-tailed and cynomolgus macaques,[13-15] but not non-pathogenic host species of SIV such as sooty mangabeys and African green monkeys.[7, 16, 17] Studies of acute pathogenic SIV infection in macaques have demonstrated a transient increase of pDCs in peripheral blood because of rapid egress from the bone marrow, followed by rapid depletion of circulating pDCs and accumulation of apoptotic pDCs in lymph nodes.[13, 18] These data, combined with evidence that pDCs are targets of infection for SIV and HIV,[13, 19-22] led to a model where pDCs were generally depleted during progressive lentivirus infection. However, recently we entertained an alternative hypothesis – that pDCs are not depleted but are trafficking elsewhere. Indeed we and others have now shown that in fact pDCs are not depleted by HIV/SIV infection, but rather accumulate in the colorectum through an α4β7-dependent homing mechanism.[6, 17] This finding has redefined the view of pDCs in lentivirus disease, but was incomplete, only focusing on pDCs in blood and colorectal biopsies. Furthermore, the functionality of pDCs in the mucosae is still unclear. Herein, we address this deficit by investigating the systemic effects of SIV infection on pDC redistribution.

Materials and methods

Animals

Twenty-nine Indian rhesus macaques (Macaca mulatta) were analysed; 19 SIV-naive macaques and 10 macaques infected chronically with SIVmac239. All animals were free of simian retrovirus type D, simian T-lymphotrophic virus type 1 and herpes B virus, were housed at the New England Primate Research Center and were maintained in accordance with the guidelines of the Committee on Animals of the Harvard Medical School and the Guide for the Care and Use of Laboratory Animals.

Cell processing

Macaque peripheral blood mononuclear cells were isolated from EDTA-treated blood by density gradient centrifugation over lymphocyte separation medium (MP Biomedicals, Solon, OH) and contaminating red blood cells were lysed using a hypotonic ammonium chloride solution. Mononuclear cells were isolated from various tissue sections by both enzymatic and mechanical disruption as described previously for our laboratory.[5, 23, 24]

Antibodies and flow cytometric analyses

Flow cytometry staining of mucosal dendritic cells was performed as previously described.[6] LIVE/DEAD Aqua dye (Invitrogen, Carlsbad, CA) and isotype-matched controls and/or fluorescence-minus-one controls were included for all assays. Except where noted, all antibodies were obtained from BD Biosciences (La Jolla, CA) and included fluorochrome-conjugated monoclonal antibodies to the following molecules: Caspase-3 (Alexa647 conjugate, clone C92-605), CD3 [allophycocyanin (APC)-Cy7 conjugate, clone SP34.2], CD4 [Alexa700 and peridinin chlorophyll protein (PerCp)-Cy5.5 conjugates, clone L-200], CD8 (APC-H7 and PerCp-Cy5.5 conjugates, clone SK1), CD11c (APC conjugate, clone S-HCL3), CD14 [Alexa700 and phycoerythrin (PE)-Cy7 conjugates, clone M5E2], CD20 (PerCp-Cy5.5 conjugate, clone L27), CD45 (Pacific Blue conjugate, clone D058-1283), CD123 (PE and PE-Cy7 conjugates, clone 7G3), HLA-DR (PE-Texas Red conjugate, clone Immu-357, Beckman-Coulter), and Ki67 (FITC conjugate, clone B56). Acquisitions were made on an LSR II (BD Biosciences) and analysed using flowjo (version 9.5) software (Tree Star Inc., Ashland, OR).

Intracellular cytokine staining

After isolation, fresh mucosal mononuclear cells were resuspended in RPMI-1640 (Sigma-Aldrich, St Louis, MO) containing 10% fetal bovine serum alone (R10) or with imiquimod (Sigma-Aldrich) at a final concentration of 10 μm. Golgiplug (brefeldin A) and Golgistop (monensin) were added to all samples at final concentrations of 6 μg/ml and cells were then cultured for 12 hr at 37° in 5% CO2. After culture, pDCs were surface-stained using markers as shown in Fig. 1, and then cells were permeabilized using Caltag Fix & Perm. Intracellular cytokine staining was performed for MIP-1β (FITC conjugate, clone 24006, R&D Systems, Minneapolis, MN), IFN-α (PE conjugate, clone 225.C, Chromaprobe, Maryland Heights, MO) and TNF-α (Alexa 700 conjugate, Mab11). Raw data analysis was performed using FlowJo software and multi-parametric analyses were done using spice (version 5.22).[25]

Figure 1.

Distribution of plasmacytoid dendritic cells (pDCs) in lymphoid and mucosal tissues from naive and simian immunodeficiency virus (SIV) -infected macaques. (a) Representative gating strategy for CD123+ pDCs and CD11c+ myeloid DCs (mDCs) among live tissue mononuclear cells. Frequencies of (b) pDCs and (c) mDCs in multiple tissue sites. Box and whisker plots show medians and ranges of between 6 and 19 animals per tissue group. Mann–Whitney U-tests were used for naive versus SIV comparisons; *P < 0·05; **P < 0·01; ***P < 0·001. Only statistically significant differences are shown. PBMC, peripheral blood mononuclear cells; PaLN, MLN and PLN, pararectal/paracolonic, mesenteric and peripheral (axillary, inguinal) lymph nodes; ND, not done.

Plasma virus load quantification

RNA copy equivalents were determined in EDTA-treated plasma using a quantitative real-time RT-PCR assay based on amplification of conserved sequences in gag.[26] The limit of detection for this assay was 30 viral RNA copy equivalents/ml plasma.

Statistical analyses

All statistical analyses were performed using graphpad Prism 6.0 software (GraphPad Software, Inc., La Jolla, CA). Non-parametric Mann–Whitney U-tests and Spearman correlation tests were used where indicated and values of P < 0·05 were assumed to be significant in all analyses.

Results

Chronic SIV infection induces redistribution of pDCs from lymphoid organs to gut mucosae and associated draining lymph nodes

In a planned serial-sacrifice study we sought to comprehensively evaluate pDC distribution in both naive and chronically SIV-infected macaques. We first defined pDCs as CD45+ live mononuclear cells expressing HLA-DR but negative for lineage markers (CD3, CD14, CD20). Plasmacytoid DCs were also CD123bright, but negative for CD11c, a common myeloid dendritic cell marker (Fig. 1a). Similar to previous reports in HIV-infected persons and SIV-infected macaques,[10, 11, 13, 18] we observed a three-fold reduction of pDCs in peripheral blood (Fig. 1b). pDCs were also reduced in bone marrow of SIV-infected macaques, albeit not significantly, and previous reports have suggested that this reflects an efflux of precursor cells into the circulation.[18] Interestingly, we also found pDCs to be reduced in spleen and liver, suggesting a potential loss of pDCs from both lymphoid and non-lymphoid organs. Finally, we sought to enumerate pDCs in mucosae and associated lymph nodes. We and others recently reported that pDC loss in blood is actually reflective of migration to the gut mucosa,[6, 17] but these initial studies only demonstrated this phenomenon in colorectal biopsies. In this current study we found that pDCs are not only increased in colorectal biopsies, but in total colonic and jejunum tissues with fourfold and threefold increases, respectively. Furthermore, we found, perhaps unexpectedly, that pDCs also accumulated in pararectal/paracolonic and mesenteric lymph nodes, but not in inguinal and axillary lymph nodes. These data indicate that pDCs not only migrate and/or expand in the gut mucosa but also in the gut-associated draining lymph nodes. Interestingly, tissue numbers of pDCs did not correlate with viral loads in these animals (data not shown). By comparison, myeloid DC frequencies in each of these tissues were largely unchanged (Fig. 1c), suggesting that the redistribution of pDCs is cell-specific.

Gut pDCs from SIV-infected macaques develop a novel multi-functional phenotype

Some studies have suggested that during chronic HIV/SIV infection pDCs are dysfunctional[2, 11, 27, 28] whereas it has also been argued that pDCs are continually activated during infection and contribute to chronic immune activation.[7] However, previous studies have focused on pDCs in peripheral blood and the results have been inconclusive. In a smaller animal cohort we previously found that pDCs in the gastrointestinal tract produced IFN-α, MIP-1β and TNF-α.[6] Now using a simultaneous three-function assay in a large cohort of normal and SIV-infected macaques, we verified that colonic pDCs produce high levels of IFN-α and MIP-1β but very little TNF-α in normal animals (Fig. 2a). Consistent with previous findings in SIV-infected macaques pDC production of IFN-α was slightly, but not significantly, reduced while TNF-α production was up-regulated. Interestingly, IFN-α mean fluorescence intensity was significantly increased in SIV-infected macaques (Fig. 2b), suggesting a bulk increase in IFN-α due to numerical expansion of pDCs. However, the most unexpected finding was that although changes in overall cytokine production during infection were minimal, the per cell cytokine production was significantly altered during infection. In normal animals IFN-α, MIP-1β and TNF-α were generally produced by discrete subpopulations of pDCs, whereas in infected animals there was a significant increase in both bi-functional and tri-functional cells (Fig. 2c).

Figure 2.

Plasmacytoid dendritic cell (pDC) function in colonic tissue. (a) Percentages and (b) mean fluorescence intensities (MFIs) of colonic pDCs from naive and simian immunodeficiency virus (SIV)-infected macaques expressing intracellular interferon-α (IFN-α), macrophage inflammatory protein-1β (MIP-1β) and tumour necrosis factor-α (TNF-α) following imiquimod stimulation. Bars represent means ± SEM of six to eight animals per group. (c) Multi-parametric SPICE (version 5.22) analyses of the data shown in (a). Bars represent poly-functional subpopulations of stimulated pDCs and correspond to pie slices as shown. Arcs summarize overlap of each of the three functions. For multi-parametric analyses, non-responsive cells are not included in the analyses. Mann–Whitney U-tests were used for naive versus SIV comparisons; *P < 0·05. Only statistically significant differences are shown.

Gut pDC functions are associated with changes in T-cell activation and turnover

As increased cytokine production and alterations in the gut mucosa are considered to be prime mechanisms of inducing chronic immune activation we next evaluated relationships between ex vivo gut pDC cytokine production (as measured in Fig. 2a) and systemic markers of T-cell activation and turnover. To do so we quantified intracellular Ki67 as a marker of proliferation/activation and intracellular caspase-3 as a marker of apoptosis/turnover in bulk circulating CD4+ CD3+ and CD8+ CD3+ T cells. Despite its role as an inflammatory cytokine, we found no correlations between MIP-1β production and either of these T-cell markers (data not shown). However, TNF-α production by gut pDCs was positively associated with caspase-3 expression in both CD4+ (R = 0·886, P = 0·033) and CD8+ T cells (R = 0·600, P = 0·042), as well as with Ki67 expression in CD8 T cells (R = 0·429, P = 0·042). Interferon-α production also significantly correlated with caspase-3 expression in CD4+ T cells (R = 0·371, P = 0·049). Collectively, these data indicate that increased gut-homing and cytokine production by pDCs during SIV infection could contribute to the net increase in systemic immune activation observed in progressive disease.

Discussion

Until recently, our understanding of the role of pDCs in lentivirus disease has been somewhat incomplete and complicated by the perception of generalized pDC depletion. This deficit has made addressing the question of whether or not pDCs are ‘helpful’ or ‘hurtful’ to controlling disease difficult. To help better refine the model of lentivirus-induced modulation of pDC dynamics we present two new bodies of data: (i) pDCs are reduced in blood and non-lymphoid organs, but accumulate in the mucosae and associated lymph nodes; and (ii) chronic SIV infection alters mucosal pDC cytokine secretion by inducing a multi-functional phenotype.

In the post-highly active anti-retroviral therapy era one of the greatest causes of ongoing disease in HIV patients is chronic immune activation, contributing to cardiovascular diseases, non-AIDS-related cancers, and liver and kidney failure.[29-31] Although the breakdown of the gut microenvironment leads to activation via microbial translocation,[8, 9, 32] it is unclear why activation persists in the absence of obvious HIV replication. As pDCs are the primary source of IFN-α in the SIV-infected gut, our new data demonstrating the overall redistribution to gastrointestinal tissues raises the question – could pDCs be contributing to apoptosis in the gut, so fuelling microbial translocation and activation? Although the frequency of TNF-α+ and IFN-α+ pDCs did not significantly increase, the relative increase in pDCs in mucosal sites may contribute to a net increase in cytokines in the gut, which we have observed previously.[5] Indeed a significant shift in mean fluorescence intensity of IFN-α in pDCs from SIV-infected macaques (Fig. 2) supports this notion. Furthermore, HIV stimulates pDCs to secrete IFN-α and indoleamine 2,3-dioxygenase in vitro that inhibits T-cell proliferation and induces apoptosis.[33, 34] This observation is in line with our observed correlation between pDC-produced cytokines and T-cell activation and apoptosis in the circulation.

The alteration in the cytokine production profile of gut pDCs during SIV infection is puzzling. pDC cytokine production has long been shown to occur in sequential waves[35] – IFN-α followed by pro-inflammatory cytokines. Our data suggest that through some unknown mechanism SIV may induce these functions simultaneously in the gut. Indeed, seminal work from O'Brien et al.[1] also indicated that HIV induces a persistent IFN-α-producing phenotype in pDCs that may persist with production of other cytokines, perhaps as a result of a change in transcriptional regulation. Hence, what we are demonstrating as multi-functional pDCs may be the first in vivo observation of pDCs that have become static in cytokine production and partial maturity. Although pDCs are highly responsive to HIV RNA and DNA, this functional phenotype could also be a result of alternative stimuli such as CpG-DNA or microparticles.[32, 36, 37] Regardless, regulation of gut-trafficking and cytokine production in pDCs is probably highly complex, but delineation of such mechanisms will undoubtedly lead to a better understanding of how pDCs balance antimicrobial properties and activation/apoptosis in the gut.

Acknowledgements

The authors thank Angela Carville, Elaine Roberts and Joshua Kramer for animal care, and Tristan Evans and Michelle Connole for expert technical assistance. This work was supported by a CHAVI/HVTN Early Career Investigator award, grant number U19 AI067854, a Harvard University CFAR grant, number P30 AI060354 (both to RKR), and NIH NEPRC base grant P51 OD011103.

Disclosures

All authors report no conflicts of interest.

Ancillary