The authors have declared that no competing interests exist.
Conceived and designed the experiments: BAH DEK MCC. Performed the experiments: BAH ML PAV FAS. Analyzed the data: BAH. Contributed reagents/materials/analysis tools: MA BDW DEK. Wrote the paper: BAH MCC. Collected and registered patients for the study: EPS.
Despite the success of antiretroviral therapy (ART), it does not cure Human Immunodeficiency Virus (HIV) and discontinuation results in viral rebound. Follicular dendritic cells (FDC) are in direct contact with CD4+ T cells and they retain intact antigen for prolonged periods. We found that human FDC isolated from patients on ART retain infectious HIV within a non-degradative cycling compartment and transmit infectious virus to uninfected CD4 T cells in vitro. Importantly, treatment of the HIV+ FDC with a soluble complement receptor 2 purges the FDC of HIV virions and prevents viral transmission in vitro. Our results provide an explanation for how FDC can retain infectious HIV for extended periods and suggest a therapeutic strategy to achieve cure in HIV-infected humans.
Human immunodeficiency virus (HIV) can lead to acquired immunodeficiency syndrome, or AIDS. Before the introduction of anti retroviral therapy (ART) in the mid-1990s, people with HIV could progress to AIDS in just a few years. Today patients with HIV have a close to normal life expectancy. Worldwide, there are about 2 million new cases of HIV per year. Currently about 35 million people are living with HIV of which around 13 million receive ART. Still an estimated 1.5 million people die from the consequences of HIV each year. Despite the success of ART, it does not cure HIV and discontinuation results in viral rebound. Follicular dendritic cells (FDC), located central to the B cell follicle, are also in direct contact with T cells. FDCs retain intact antigen for prolonged periods. We found that human FDCs isolated from patients on ART retain infectious HIV and can transmit virus to uninfected T cells in vitro. Treatment of the HIV+ FDC with a soluble complement receptor 2 purges the FDC of HIV virions and prevents viral transmission to T cells in vitro. Our results can explain how FDCs retain infectious HIV and suggest a therapeutic strategy to come closer to a cure.
Anti-retroviral therapy (ART) is capable of suppressing plasma viral load to undetectable levels and in many cases results in restoration of circulating CD4 T cell counts to near normal values. Despite the success of ART, when treatment is halted the virus rebounds suggesting the presence of a long lived reservoir [
While latency in CD4 T cells represents one possible source of persisting virus [
HIV is capable of independently fixing complement through complement factor I and, paradoxically, this enhances HIV infectivity in vitro [
To determine if human FDCs take-up and cycle antigen similar to murine FDCs, complement C3 opsonized phycoerythrin immune complexes (PEIC) were used as a model antigen [
(A) Human FDCs from lymph node biopsies of HIV-negative subjects were loaded with complement C3-opsonized phycoerythrin immune complexes (PEICs), then either fixed (left), acid washed and fixed (middle), or acid washed, incubated in media for 30 min (recovery), and then fixed (right). Fixed FDCs were surface stained with an antibody against PE. PEICs detected on the surface of the FDC after acid wash and recovery must have come from the inside of the cell since no PEIC was detected on the surface after the acid wash. (B) Quantification of the images shows efficient stripping of PEICs by acid wash treatment and a robust recovery of PEICs on the surface after 30 min. ****P<0.0001 (Two-way ANOVA, multiple comparisons) n = 3 (3 subjects, 4 replicates each). (C) On human FDCs PEIC reside mainly in Tf positive compartments, as is the case in mice. The recycling Tf compartment was visualized by incubation of live cells with fluorescent Tf for 8 minutes. (D) The percentage of Tf or Lamp-1 positive vesicles within the PEIC positive vesicles was quantified. This shows ~45% of PEICs in Tf positive compartments and only ~12% in Lamp-1 positive compartments. ****P<0.0001 (Two-way ANOVA) n = 3 (3 subjects, 6 replicates each).
To examine whether model IC are taken up and transported into the cycling or the lysosomal compartment, transferrin (Tf) and Lamp-1, respectively were used as endosomal markers. Lamp-1 primarily localizes with the degradative compartment, whereas Tf localizes mainly with the neutral pH recycling endosomal compartment [
A hallmark of FDCs is their ability to retain antigen for extensive periods in the form of an IC [
To assess whether human FDC retain HIV and localize it to a similar compartment as observed with PEIC, inguinal LN tissue was harvested from seven HIV+ subjects classified as chronic progressors (CP) and undergoing therapy with ART and three healthy controls (
Phenotype negative represents healthy subjects. CP phenotype is chronic progressor. ARV (anti-retroviral). CD4 and CD8 count are per ml of blood. Serum viral load was determined one week before surgery using the hospital standard method, which has a detection limit of 20 virions/ml.
1 | Negative | --- | --- | <48 | n/a | n/a |
2 | Negative | --- | --- | <20 | n/a | n/a |
3 | Negative | --- | --- | <48 | n/a | n/a |
4 | CP | 635 (32) | 926 (47) | 61 | 1 | 1 |
5 | CP | 1371 (33) | 1814 (44) | <20 | 5 | 10 |
6 | CP | 712 (35) | 717 (37) | <20 | 7 | 24 |
7 | CP | 458 (23) | 838 (41) | <20 | 0 | 8 |
8 | CP | 667 (24) | --- (39) | <20 | 5 | 7 |
9 | CP | 1083 (48) | --- (26) | <20 | 1 | 6 |
10 | CP | 419 (27) | 687 (44) | <20 | 1 | 13 |
CP = Chronic Progressor, n/a = not applicable, ARV = Anti RetroViral
Enriched FDC were cultured for 5 days to regain their dendritic morphology before imaging. HIV was detected in FDC stromal cells by confocal microscopy analysis of permeabilized cells after staining with KC-57 antibody specific for the HIV core protein p24 (
(A) Cultured human FDCs from HIV+ subjects on ART were stained for Tf (green), HIV (p24, red), Lamp-1 (blue) and Hoechst (cyan). (inset) A single vesicle was enlarged and only the red (HIV) channel was shown. The yellow line was used for line profile analysis, which measures the fluorescent intensity over that line in all channels. Line profiles were also made in the z direction to ensure co-localization of HIV+ vesicle in the X, Y and Z direction with Tf or Lamp-1. (B) Quantification of line profile measurements of HIV positive vesicles. The vast majority of HIV containing vesicles were positive for Tf (~80%), while only ~5% was positive for Lamp-1 ****P<0.0001 (Two-way ANOVA) n = 2 (2 subjects, 4 replicates each). (C) FDCs and FRCs from HIV positive individuals on ART treatment and HIV negative subjects were cultured and viral RNA levels were quantified. RNA levels are represented per well. FRC and FDC were isolated from the same subjects to control for contaminating cells. Each data point presents an individual subject. ****P<0.0001 (Two-way ANOVA, multiple comparisons) n = 4 (4 subjects, 3 replicates).
The number of FDCs per well was calculated by counting FDCs in multiple fields of view (FOV) within a known area. Based on this area and the total surface area of the well, an estimate of total FDCs per well was calculated. Staining FDC with anti-p24 and counting HIV positive vesicles determined positive cells. The table compares the estimated HIV positive vesicles per well based on imaging with the average RNA copy number found per well in the corresponding subjects.
1 | 10k-100k | 112 | 0 | 0 | 0 | 0 | 2.4 |
2 | 4k-40k | 48 | 0 | 0 | 0 | 0 | 5 |
3 | 7-70k | 79 | 0 | 0 | 0 | 0 | 2 |
4 | 8k-80k | 86 | 19.8 | 24 | 1.4 | 2k-20k | 1620 |
5 | 5k-50k | 56 | 12.5 | 12 | 1.7 | 1k-10k | 609 |
6 | 5k-50k | 68 | 16.2 | 17 | 1.5 | 1k-10k | 1204 |
7 | 2k-20k | - | - | - | - | - | 910 |
8 | 4k-40k | - | - | - | - | - | 709 |
9 | 3k-30k | 36 | 22.2 | 23 | 2.9 | 2k-20k | 1422 |
10 | 15k-150k | 187 | 46.0 | 206 | 2.4 | 15k-150k | 1596 |
In order to determine if human FDC retain viral RNA in addition to viral protein, cellular RNA was prepared from enriched FDC harvested from HIV+ and HIV- subject LNs and analyzed by droplet digital RT-PCR (ddPCR). The viral copy number was determined per well. As expected, no viral RNA was observed in FDCs from HIV- subjects. By contrast, FDC isolated from HIV+ subjects retained significant levels of viral RNA, i.e. mean±SD, 1007±314.3 viral copies per well of FDC (
To determine if HIV retained by FDCs is infectious, FDCs isolated from ART treated HIV+ or healthy HIV- subjects were co-cultured with purified activated CD4 T cells prepared from PBMC of healthy volunteers. After 5 days of co-culture, CD4+ T cells were separated from the adherent FDC and RNA was extracted from both populations of cells (
FDC cultures (103 to 105 cells) of HIV positive subjects on ART were co-cultured with activated CD4 T cells (105 cells) from PBMC of healthy subjects for 5 days. Non-adherent CD4 T cells were removed from the co-cultures by washing and analyzed for viral RNA and DNA content. (A) RNA from CD4 T cells was collected for analysis. T cells co-cultured with FDCs from HIV+ subjects contained viral RNA, in contrast to the T cells co-cultured with FDCs from healthy subjects. This indicates transfer of HIV from FDC to T cell. Each data point presents an individual subject. **P<0.01 (Two-way ANOVA) n = 6 (6 subjects, 2 replicates each). (B) DNA from CD4 T cells was collected for analysis. Integrated viral DNA was detected in T cells co-cultured with FDCs from a HIV+ subject, in contrast to the T cells co-cultured with FDCs from a healthy subject. This indicates productive infection of T cell. DNA integrations normalized to 105 T cells. **P<0.01 (Student’s
Complement receptor CD21 is expressed primarily on B cells and FDCs in both human and murine tissues. Earlier studies identified an important role for CD21 in HIV retention in murine lymphoid tissues and in transmission of virus to human CD4 T cells [
To determine if human FDC retain HIV via CD21, FDCs positive for virus were cultured overnight with either a CD21 decoy receptor, i.e. fusion protein of the C3d binding domain of human CD21 and murine immunoglobulin, (sCD21-Ig), or an isotype control [
FDCs were isolated from LNs of HIV positive subjects on ART and cultured for 4 days. CD4 T cells from HIV negative subjects were activated in parallel. On day 4 FDC cultures were treated overnight with an isotype control antibody, soluble CD21-Ig (sCD21-Ig) or a cocktail of 3 broadly neutralizing antibodies (bNab; VRC01, PG16 and PGT121). Then cultures were washed and co-cultured for 5 additional days with 105 activated CD4 T cells. On day 5 samples were split into 3 groups: FDCs for imaging (A, B), FDCs for RNA quantification (C) and T cells for RNA quantification (D). (A) FDCs were fixed, stained for p24 (HIV, red), CD35 (FDC, green) and Hoechst (blue) and imaged by confocal microscopy. Virions were detected in the isotype and bNab cocktail treated cultures, but not in the sCD21-Ig treated culture. (B) Percentage of HIV containing FDCs. 10 random field of views were collected and quantified per sample. ns: not significant, **P<0.01 (Two-way ANOVA, multiple comparisons) n = 3. (C) HIV RNA quantification of FDCs after sCD21-Ig treatment. Each data point represents an individual subject. Total HIV RNA per well is depicted. ****P<0.0001 (Two-way ANOVA, multiple comparisons) n = 3 (3 subjects, 3 replicates each). (D) HIV RNA quantification of T cells incubated with FDCs treated with sCD21-Ig (left panel) or bNab cocktail (right panel). HIV RNA is depicted per 105 T cells, which is the total per well. *P<0.05, ***P<0.001 (Two-way ANOVA) n = 3 (3 subjects, up to 3 replicates).
To test whether blockade of C3d binding by CD21 could prevent transmission of infectious virus, sCD21-Ig was added to FDC cultures that were then co-cultured with uninfected CD4 T cells as described above (
Schematic model. HIV is captured and subsequently cycled by complement receptor 2 (CD21) on the follicular dendritic cell (FDC). The HIV virion resides in the protective recycling endosome of the FDC. Upon emerging from the endosome at the cell surface, HIV can infect surrounding T follicular helper (Tfh) cells that have been attracted to the FDC by a CXCL13 gradient. Infection occurs through binding of CD4 and either CXC-chemokine receptor 4 (CXCR4) or CC-chemokine receptor 5 (CCR5) as a co-receptor by gp120. The soluble CD21 receptor fusion protein (sCD21-Ig) competes with the CD21 on the FDC for binding of complement C3d on the virion. This facilitates the release of the virion and makes it available for degradation by other cells. To prevent infection of T cells at this stage the treatment should be combined with broadly neutralizing antibodies (bNab).
As further evidence that FDCs retain infectious HIV and as a control for viral transmission to healthy CD4 T cells, a cocktail of bNab was added to HIV+ FDC cultures prior to co-culturing with CD4 T cells. The bNabs used (VRC01, PG16 and PGT121) block CD4 binding to gp120 (VRC01) and gp120 glycans (PG16 and PGT121) [
Despite the success of ART in reducing HIV viral loads and partial restoration of circulating CD4 T cells, cessation of the drugs results in acute viral rebound, suggesting the importance of a viral reservoir [
It was important to first test if human FDC retain model antigen IC similar to that observed in mice. Using PEIC opsonized with complement, human FDC were loaded with the complexes in the
The number of viral antigen+ vesicles per FDC is in a similar range as that identified for viral RNA copies, i.e. 1.4 to 2.9 HIV+ vesicles compared to approximately 0.2–2.1 viral copies of RNA. Earlier studies of LNs of HIV+ subjects following initial treatment with ART identified a dramatic reduction in viral load in both CD4 T cells and FDC within the first a month [
Trapping of HIV on the FDC by immune complexes includes incorporation of the infectious virus into periodically cycling endosomes that retain the virus within the FDC. This not only further identifies the B cell follicle as an important reservoir of infectious virus, these data imply that the follicle can be a source of infectious virus for, potentially, several years after the individual is put on ART and has undetectable viremia in the blood.
A prediction about an HIV reservoir is that it will include multiple variants of HIV collected over an extended period. Therefore, one would expect FDC to act as an archive for multiple genotypes of virus. Earlier studies by Burton and colleagues compared the RNA sequence of
In the current study, our finding that FDC take up virus via CD21 and cycle the complex within the transferrin Tf compartment provides a mechanism how FDC might serve as an archive. Tf marks a mildly acidic compartment in contrast to the lysosomal compartment and would provide a “non-labile” environment in which the virus can remain infectious for an extended period yet maintain periodic contact with TFH within the LN follicles.
There is a growing literature to support ongoing replication in lymphatic tissues of patients on ART with suppression in blood [
Since the FDC cultures could potentially be contaminated with an alternative cellular source of HIV, it was important to rule this out. It was especially important to test for the presence of contaminating HIV+ CD4 T cells. Several lines of evidence make this possibility unlikely. First, LN stromal cell isolation includes a CD45 negative sort that eliminates most bone marrow derived cells, including T cells. The second step, which is positive selection with anti-CD35, further limits the possibility of T cell contamination. As an aside, we did observe a small population of CD35 low CD4 T cells among the lymphocytes isolated from an HIV infected subject. Immuno-sorting of the cells under similar conditions used for FDC determined that the CD35 positive selection step failed to enrich for the T cells probably due to the very low level of surface antigen expression (
Previous studies support ongoing replication in lymphatic tissues of patients on ART with viral load suppression in blood [
Earlier studies in mice identified the importance of CD21 receptor in uptake and retention of foreign antigen by FDC [
In summary, we hypothesize that C3-opsonized HIV particles are transported to the FDC by B cells, as has been shown with immune complexes, where it is taken-up and cycled in a neutral pH compartment until it encounters a CD4+ TFH (
C57BL/B6 background mice were purchased from Jackson Laboratories and maintained in specific- pathogen-free facilities at Boston Children’s Hospital Program in Cellular and Molecular Medicine (PCMM), Harvard Medical School.
Research conformed to ethical guidelines established by the ethics committee of the Massachusetts General Hospital, University of Montreal Health Center. Inguinal lymph nodes were excised under anesthesia by a surgeon at Massachusetts General Hospital according to normal surgical procedures. Tissue was then transported on ice to the BL2+ facility at the Ragon Institute. The committee on microbiological procedures at Harvard Medical School approved the study. Protocol number 03097.
Studies in mice were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Institutional Animal Care and Use Committees (IACUC) at Harvard Medical School and the Program in Cellular and Molecular Medicine (PCMM) at Boston Children’s Hospital. Approval numbers 889 and 04156.
Written, informed consent, approved by the Partners Human Research Committee of the Massachusetts General Hospital, was provided and signed by study participants before enrolment in the study. Protocol number: 2012P001806.
B-Phycoerythrin (PE) (Anaspec) was used as a model Ag. ICs were generated by mixing 5 μg of PE, 5 μg of rabbit anti-PE IgG (Rockland), and 10% freshly isolated C57BL/B6 or human serum (as a source of complement) in 100 μl GVB++ buffer (Complement Tech) for 30 min at 37°C. Splenocytes from a C57BL/B6 mouse were then incubated with the immune complex mix for 30 min at 37°C to generate IC-bound B cells. Alternatively, the human B cell line Raji was used.
FDC isolation and culture procedures were modified from that described to be compatible with human [
HIV-1 and CD4 RNA was quantified using the QX100 Droplet Digital PCR system (Bio-Rad, Pleasanton, CA). The HIV-1 ddPCR mix consisted of: 10 μl 2x ddPCR super mix for probes (Bio-Rad); 500 nM of Forward (5’-CATGTTTTCAGCATTATCAGAAGGA-3’) and Reverse (5’-TGCTTGATGTCCCCCCACT-3’) primers; 250 nM probe mix (5’-FAM-CCACCCCACAAGATTTAAACACCATGCTAA-TAMRA-3’) and 3 μl of the cDNA into a final volume of 20 μl. The CD4 ddPCR mix consisted of: 10 μl 2x ddPCR super mix for probes (Bio-Rad); 500 nM of Forward (5’-AGTCCTCACACAGATACGCC-3’) and Reverse (5’-ACTCACATCCGAACACTAGCAA-3’) primers; 250 nM probe mix (5’-FAM-TGAAGTGGAGGACCAGAAGGAGGA-TAMRA-3’) and 3 μl of the cDNA into a final volume of 20 μl. The total mix was placed into the 8-channel cartridge, 70 μl of droplet generating oil was added and droplets were formed in the QX100 droplet generator (Bio-Rad). Droplet in oil suspensions were transferred to a 96 well plate and placed into the T100 Thermal Cycler (Bio-Rad). Cycling conditions were as follows: 95°C for 10 min, followed by 45 cycles of 94°C for 30 seconds and 60°C for 60 seconds. Subsequently, the droplets were automatically read by the QX100 droplet reader (Bio-Rad) and the data was analyzed with the QuantaSoft analysis software 1.3.2.0 (Bio-Rad).
FDCs were fixed in 1% paraformaldehyde solution. Fixed samples were pre-incubated with anti- FcR (2.4G2, house produced) and 2% bovine serum albumin for blocking nonspecific and Fc-mediated binding. Anti-PE (Rockland), donkey anti-rabbit DyLight 405, anti-CD35, anti-Lamp-1 (Biolegend), fluorescent transferrin and anti-p24 (KC57-RD1, Beckman-Coulter) were used for staining ex vivo cultures of FDCs. Images were acquired with a FluoView FV1000 confocal microscope (Olympus) with a 20x lens (NA: 0.7) or a 60x water immersion lens (NA: 1.2) and processed with FluoView software (Olympus). Data were analyzed with Fiji software.
The broadly neutralizing antibodies used in this study are available through
(A) PBS (vehicle) or live GFP-HIV (green) was injected in the footpad of a mouse. After 48h the LN was collected, fixed and cryosectioned. Sections were stained with the FDC marker 8C12 (red). Note the dendritic pattern of FDC (red) within the follicles and the co-localization of HIV particles (green) with dendrites in HIV treated mice. Representative image. Each symbol represents a quantified follicle. (B) Quantification of the correlation of HIV-GFP signal with FDC signal was performed using the Cell Profiler software. Each symbol represents a different follicle; full z-stacks were analyzed per follicle. ***P<0.001 (unpaired Student’s
(TIF)
Line profiles are fluorescent intensity measurements over a line in an image, creating a graph. (A) Lines over the HIV vesicle in XY and YZ used to calculate the line profiles of all 3 stains. (B) Line profile graphs from the lines shown in A in XY and YZ. Results confirm co-localization of transferrin signal with p24 staining in both XY and YZ optical planes.
(TIF)
(A) Human FDCs were loaded with PEIC using Raji B cells in vitro, then FDCs were either mock treated with isotype control IgG (left panel) or sCD21-Ig treated (right panel) overnight and analyzed the following day. After sCD21-Ig treatment no PEIC+ vesicles were detected. By contrast, PEIC+ vesicles were identified in the isotype treated controls. (B) The mean fluorescent intensity (MFI) of 16 random fields of views (FOV) at lower magnification was compared. ****p<0.0001 (unpaired Student’s
(TIF)
LNs were harvested and FDCs isolated, first by depletion of CD45+ B and T cells, followed by positive selection for CD35+ cells. Cultures were either treated with sCD21-Ig, isotype control or RNA was directly isolated. The remaining samples were washed and HIV negative CD4 T cells were added to the FDC wells and co-cultured for 5 days. Afterwards FDCs and T cells were separated and RNA isolated. HIV RNA was quantified by ddPCR. Numbers in the graph represent total virions per well. The average cell number per well for this subject was between 15k and 150k (subject #10).
(TIF)
(A) CD35 expression on T and B cells shows expression on >96% of B cells as expected, interestingly 1% of T cells also express CD35, although the mean fluorescent intensity levels were lower than on B cells. (B) Before the positive CD35 bead selection for FDCs >79% of cells were CD3 positive. In contrast, after CD35 positive selection for FDC, only 0.3% (estimate 1 cell) was CD3 positive. Thus, the contribution of T cells to the CD35 positively selected FDC cultures, is negligible based on immune staining and flow cytometry.
(TIF)
To correlate mRNA levels with cellular contamination, we prepared a standard curve by sorting B and T cells separately and then spiking the B cells with known numbers of T cells (0 to 106 T cells in 10-fold increments, 7 data points). RNA was extracted from the B cell samples bearing a known number of contaminating CD4 T cells and the amount of CD4 mRNA quantitated using ddPCR and a standard curve prepared. (A) Standard curve of CD4 mRNA versus CD4 T cell number. (B) Extrapolation of the number of CD4 T cells in samples based on CD4 mRNA levels. Extrapolation of the number of CD4 T cells in a pure CD4 sample correlates with the number of T cells analyzed (approximately 105 cells). CD4 mRNA levels in the FDC samples indicate limited CD4 T cell contamination of the historical FDC samples. ***p<0.001 (unpaired Student’s
(TIF)
The authors thank all members of the M.C.C. lab and of the Ragon Institute for suggestions and help with experiments. We acknowledge the valuable assistance of Elisabeth M. Carroll, Galit Alter, Maria Buzón, Thomas T. Murooka, Thorsten R. Mempel, Stephanie Jost, Amy Baxter and Sara Miller.