(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).

(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).

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 t test) n = 1 (1 subject, 3 replicates). (C) DNA from FDCs was collected for analysis. FDCs from a HIV+ subject had no viral DNA integrated in their genome, just as FDCs from a healthy subject. DNA integration represented as total per well. ns = not significant (Student’s t test) n = 1 (1 subject, 3 replicates).

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).

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).