Expression of HIV-1 Intron-Containing RNA in Microglia Induces Inflammatory Responses

Abstract

Chronic neuroinflammation is observed in HIV⁺ individuals on suppressive combination antiretroviral therapy (cART) and is thought to cause HIV-associated neurocognitive disorders. We have recently reported that expression of HIV intron-containing RNA (icRNA) in productively infected monocyte-derived macrophages induces pro-inflammatory responses. Microglia, yolk sac-derived brain-resident tissue macrophages, are the primary HIV-1 infected cell type in the central nervous system (CNS). In this study, we tested the hypothesis that persistent expression of HIV icRNA in primary human microglia induces innate immune activation. We established multiple orthogonal primary human microglia-like cell cultures including peripheral blood monocyte-derived microglia (MDMG) and induced pluripotent stem cell (iPSC)-derived microglia. Unlike MDMG, human iPSC-derived microglia (hiMG), which phenotypically mimic primary CNS microglia, were robustly infected with replication competent HIV-1, and establishment of productive HIV-1 infection and de novo viral gene expression led to pro-inflammatory cytokine production. Blocking of HIV-1 icRNA expression, but not multiply spliced viral RNA, either via infection with virus expressing a Rev-mutant deficient for HIV icRNA nuclear export or infection in the presence of small molecule inhibitor of CRM1-mediated viral icRNA nuclear export pathway, attenuated induction of innate immune responses. These studies suggest that Rev-CRM1-dependent nuclear export and cytosolic sensing of HIV-1 icRNA induces pro-inflammatory responses in productively infected microglia. Novel strategies targeting HIV icRNA expression specifically are needed to suppress HIV-induced neuroinflammation.

FIG 1

Monocyte-derived microglia (MDMG) are susceptible to HIV-1 infection. (A) Schematic of MDMG differentiation protocol. (B) Representative image of MDMs or MDMGs differentiated from the same donor. Bars = 20 µm. (C and D) Expression of (C) P2RY12 and (D) GAS6 mRNA in MDMGs was quantified by reverse transcription-quantitative PCR (qRT-PCR) and normalized to that of MDM generated from the same donor. (E) Representative immunofluorescence images of MDMGs stained for nucleus (DAPI, blue) and P2RY12 or IBA-1 (red). Bar = 20 µm. (F) MDMGs were infected with Lai/YU-2env (replication-competent CCR5-tropic HIV-1, MOI = 1), and production of p24^Gag in the culture supernatant was quantified by ELISA (3 days postinfection [dpi]). (G) Western blot analysis for total SAMHD1, phosphorylated SAMHD1 expression in MDMGs, MDMs and THP-1 cells. Actin was probed as a loading control. +, PMA-treated THP-1; –, unstimulated THP-1. (H) MDMGs and MDMs were infected with HIV-1 (LaiΔenvGFP/VSV-G, MOI = 2, in the absence or presence of SIV_mac239 Vpx VLPs), and production of p24^Gag in the culture supernatant was quantified by ELISA (3 dpi). NT, no treatment (DMSO); EFV, efavirenz (1 µM); Ral, raltegravir (30 µM). The means ± standard error of the mean (SEM) are shown, and each symbol represents an independent experiment. P values: one-sample t test (panel C, two-tailed), the Wilcoxon matched-pairs signed rank test (panel D, two-tailed), or one-way ANOVA followed by the Tukey-Kramer posttest (panel F) or Dunnett’s posttest comparing to mock (panel H). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

FIG 2

HIV-1 infection induces innate immune activation in MDMGs. (A) mRNA expression profiles in MDMGs infected with HIV-1 (LaiΔenvGFP/VSV-G, MOI = 2, in the presence of SIV_mac239 Vpx VLPs) were analyzed using the human neuroinflammation panel (NanoString). (A to C) Expression of mRNA in HIV-1-infected MDMGs was normalized to that in mock-infected MDMGs (A), in infected MDMGs in the presence of (B) efavirenz or (C) raltegravir, and genes which were expressed more than the mean + 2 × standard deviation (SD) are shown. (D) Production of IP-10 in HIV-1-infected MDMGs (MOI = 2, 3 dpi) measured by ELISA. (E to G) Effects of CRM1 inhibitor (KPT-330) on HIV-1-infected MDMGs or infection of MDMGs with a Rev mutant deficient for icRNA nuclear export (Rev*, M10) on (E) viral infection (multiply spliced viral RNA expression, Rev-independent, shown as ΔCT to GAPDH), (F) p24^Gag production (Rev-dependent) measured by ELISA, or (G) IP-10 mRNA expression (shown as ΔCT to GAPDH). The means ± SEM are shown, and each symbol represents an independent experiment. P values: one-way ANOVA followed by Dunnett’s posttest comparing to mock (panels D to G). *, P < 0.05; **, P < 0.01; ****, P < 0.0001. NT, no treatment (DMSO); EFV, efavirenz (1 µM); Ral, raltegravir (30 µM); KPT, KPT-330 (Selinexor, 1 µM); Rev*, M10.

FIG 3

iPSC-derived microglia are highly susceptible to HIV-1 infection. (A) Representative phase-contrast images of iCell-MGs (Fujifilm Cellular Dynamics). Bar = 20 µm. (B) Representative immunofluorescence image of iCell-MGs stained for nucleus (DAPI, blue) and IBA-1 (red). Bar = 20 µm. (C) Representative flow cytometry profile of iCell-MGs stained for intracellular and surface P2RY12. (D) Western blot analysis for total SAMHD1, phosphorylated SAMHD1 expression in iCell-MGs, MDMGs, and MDMs. Actin was probed as a loading control. (E) Replication kinetics of HIV-1 in iCell-MGs. Cells were infected with HIV-1 (Lai/YU-2env, replication competent CCR5 tropic HIV-1, MOI = 1), and production of p24^Gag in the culture supernatant was quantified by ELISA. (F to H) iCell-MGs were infected with HIV-1 (Lai/YU-2env, MOI = 1), and (F) HIV-1 infection (intracellular p24^Gag expression) and (G and H) CD169 expression were analyzed by flow cytometry. (I and J) Production of the proinflammatory cytokines (I) IP-10 and (J) CCL2 in the culture supernatants was measured by ELISA (6 dpi). The means ± SEM are shown, and each symbol represents an independent experiment. P values: one-way ANOVA followed by Dunnett’s posttest comparing to mock (panels F and H to J), *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. NT,: no treatment (DMSO); EFV, efavirenz (1 µM); Ral, raltegravir (30 µM); Ver, verdinexor (KPT-335, 0.1 µM).

FIG 4

Establishment of iPSC-derived microglia/neuron coculture system. (A) Schematic of hiMG generation by coculturing hiMAC (yolk-sac-derived primitive macrophages) and neurons from human iPSCs. (B) Representative phase-contrast image of hiMGs and hiNeurons cocultured for 11 days. Bar = 50 µm. (C to E) Expression of (C) TMEM119, (D) CX3CR1, and (E) P2RY12 mRNA in hiMACs or hiMG-hiNeuron cocultures was quantified by qRT-PCR and normalized to that of hiNeuron solo culture. P values from one-way ANOVA test for panels C, D, and E were 0.0972, 0.0829, and 0.0814, respectively. (F) Representative immunofluorescence images of hiMG-hiNeuron cocultures stained for nucleus (DAPI, blue), neuron (tubulin beta 3: TUBB3, green), and MG markers IBA-1 or TMEM119 (red). Bars = 50 µm. (G) Representative flow cytometry profile of hiMG-hiNeuron coculture stained for neurons (TUBB3) and hiMGs (P2RY12). The means ± SEM are shown, and each symbol represents an independent experiment.

FIG 5

HIV-1 infection of hiMGs in hiMG-hiNeuron cocultures induces proinflammatory responses. hiMG-hiNeuron cocultures were infected with HIV-1 (Lai/YU-2env: replication-competent CCR5 tropic HIV-1, MOI = 1). (A and B) HIV-1 infection (intracellular p24^Gag expression) was analyzed by flow cytometry. (A) Representative flow cytometry profile is shown, and microglia (P2RY12⁺) and neuron (P2RY12⁻) populations are highlighted with pink and blue, respectively. (B) HIV-infected (p24^Gag+) cells in microglia (pink in panel A) were calculated. (C) Replication kinetics of HIV-1 in hiMG-hiNeuron coculture. Cocultures were infected with HIV-1 (Lai/YU-2env, replication-competent CCR5 tropic HIV-1, MOI = 1), and production of p24^Gag in the culture supernatant was quantified by ELISA. (D and E) Production of proinflammatory cytokines (D) IP-10 and (E) CCL2 was measured by ELISA (6 dpi). (F and G) The proportion of live cells in (F) microglia (pink in panel A) and (G) neurons (blue in panel A) was calculated. The means ± SEM are shown, and each symbol represents an independent experiment. P values: one-way ANOVA followed by Dunnett’s posttest comparing to mock (panels B and D to F); **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. The P value from one-way ANOVA test was 0.9662 for panel G. NT, no-treatment (DMSO); EFV, efavirenz (1 µM); Ral, raltegravir (30 µM); KPT, KPT-330 (selinexor, 1 µM).