Human iPSC-derived neurons reveal NMDAR-independent dysfunction following HIV-associated insults

Abstract

The central nervous system encounters a number of challenges following HIV infection, leading to increased risk for a collection of neurocognitive symptoms clinically classified as HIV-associated neurocognitive disorders (HAND). Studies attempting to identify causal mechanisms and potential therapeutic interventions have historically relied on primary rodent neurons, but a number of recent reports take advantage of iPSC-derived neurons in order to study these mechanisms in a readily reproducible, human model. We found that iPSC-derived neurons differentiated via an inducible neurogenin-2 transcription factor were resistant to gross toxicity from a number of HIV-associated insults previously reported to be toxic in rodent models, including HIV-infected myeloid cell supernatants and the integrase inhibitor antiretroviral drug, elvitegravir. Further examination of these cultures revealed robust resistance to NMDA receptor-mediated toxicity. We then performed a comparative analysis of iPSC neurons exposed to integrase inhibitors and activated microglial supernatants to study sub-cytotoxic alterations in micro electrode array (MEA)-measured neuronal activity and gene expression, identifying extracellular matrix interaction/morphogenesis as the most consistently altered pathways across HIV-associated insults. These findings illustrate that HIV-associated insults dysregulate human neuronal activity and organization even in the absence of gross NMDA-mediated neurotoxicity, which has important implications on the effects of these insults in neurodevelopment and on the interpretation of primary vs. iPSC in vitro neuronal studies.

Keywords: HIV; NMDA; antiretroviral therapy; iPSC; microglia; neuron.

FIGURE 1

Human i³ cortical neurons are resistant to cytotoxicity from HIV-ADA-infected myeloid cell supernatants. (A) Representative images of i³ neurons following 96-h exposure to HIV-iMg supernatants (n = 3 donors, mean ± s.e.m.) and (D) HIV-MDM supernatants (n = 5 donors, mean ± s.e.m.) (B,E) Quantification of average number of MAP2⁺ cells and (C,F) average MAP2⁺ area per cell.

FIGURE 2

LPS + ATP-treated myeloid cell supernatants reduce neuronal area in an NMDAR-independent manor. (A) Representative images of i³ neurons following 96-h exposure to LPS-MDM supernatants (n = 4 donors, mean ± s.e.m., One-way ANOVA, *p < 0.05, **p < 0.01). (B) Quantification of average number of MAP2⁺ cells and (C) average MAP2⁺ area per cell.

FIGURE 3

Human i³ neurons are resistant to NMDA-mediated cytotoxicity. (A) Representative images of i³ neurons following 24-h exposure to NMDA (n = 3 neuronal differentiations, mean ± s.e.m.). (B) Quantification of average number of MAP2⁺ cells and (C) average MAP2⁺ area per cell. (D) RNA-Seq read counts per million of glutamate receptor subunit genes expressed by i³ neurons with and without iMg supernatant supplement (crossed out box = not detected, minimum CPM cutoff = 5).

FIGURE 4

Integrase inhibitors elvitegravir and raltegravir are not cytotoxic to i³ neurons. (A) Representative images of i³ neurons following 9 days of exposure to 3 doses of elvitegravir (EVG) or (B) raltegravir (RAL, n = 3 neuronal differentiations mean ± s.e.m). (C,D) Quantification of average number of MAP2⁺ cells. Staining for Beta III tubulin (Red) is present in merged images but is not part of the analysis.

FIGURE 5

Elvitegravir and raltegravir differentially alter i³ neuron spontaneous activity. MEA analysis of i³ neurons throughout 9 days of exposure to 3 doses of elvitegravir [EVG, (A–C)] or raltegravir [RAL, (D–F), n = 3 neuronal differentiations, mean ± s.e.m, one-way ANOVA, *p < 0.05]). (A,D) Viability and cell area as measured by conductance. (B,E) Spontaneous firing rate and rate normalized to total cellular material. (C,F) Synchronous activity as measured by number of local and network burst events.

FIGURE 6

Integrase inhibitors similarly alter i³ neuron gene expression. Analysis of significantly differentially expressed genes (DEGs) from RNA-seq on neurons post-MEA analysis. (A) Clustered heatmap of log₂ normalized counts per million for top 30 most significant DEGs between vehicle and 1 μM elvitegravir and (B) 1 μM raltegravir. (C) Top 10 enriched KEGG pathways containing significant DEGs induced by EVG and (E) RAL. (D) Venn diagram showing commonalities among significant DEGs induced by integrase inhibitors.

FIGURE 7

LPS-iMg supernatants alter i³ neuron activity and transcription. MEA analysis of i³ neurons throughout 4 days of exposure to supernatants from LPS + ATP treated iMg (n = 3 neuronal differentiations, mean ± s.e.m, one-way ANOVA with Dunnett’s correction for multiple comparisons, *p < 0.05). (A) Viability and cell area as measured by conductance. (B) Spontaneous firing rate and rate normalized to total cellular material. (C) Synchronous activity as measured by number of local and network burst events. (D) Clustered heatmap of log₂ normalized counts per million for top 30 most significant DEGs between neurons exposed to iMg and LPS-iMg supernatants. (E) Top 10 enriched KEGG pathways containing significant DEGs induced by LPS-iMg supernatants.