henhaobio
Sign in

bioRxivNeurosciencePreprint

SPP1+ Microglia Are Associated with Neuroimmune Rewiring and Glutamatergic Neuronal Injury in ART-Suppressed People Living with HIV

Ciniso Sylvester ShabanguiD, Hongjie Chen, Manickam AshokkumariD, Nikesh Katuwal, Jacob B Lovins, Edward P. BrowneiD, Sara GianellaiD, Antoine ChailloniD, David M. MargolisiD, Yuyang Tang, Guochun JiangiD

PDFDOI

Abstract

Abstract Antiretroviral therapy (ART) effectively suppresses systemic HIV replication but does not eradicate viral reservoirs in the brain, where their identity and contribution to neurological injury remain poorly defined. Using bulk, single-cell, and single-nucleus transcriptomics of postmortem human brain tissue, we identify an activated SPP1 microglial population that expands 5.8-fold during ART and serves as the primary central nervous system (CNS) reservoir, preferentially harboring HIV transcripts. These reservoir microglia adopt a distinct, immune-evasive reprogramming state marked by chronic type I interferon signaling and inflammasome activation, which we recapitulate in primary human microglia via prolonged interferon-{beta} exposure. We show that viral transcription in ART-suppressed brains is dominated by nef, which may sustain the viral reservoir by disrupting host HLA-A and HLA-E presentation machinery. This persistent, immune-evasive state is associated with TREM2-C1QC complement-mediated synaptic pruning and severe DNA damage response dysregulation, culminating in a profound loss of VGLUT1 and GAD67 glutamatergic neurons that persists despite viral suppression. Our findings establish SPP1 microglia as an active, pathogenic CNS reservoir, identifying the SPP1, MHC-I, and complement pathways as therapeutic targets to eliminate viral persistence and reverse HIV-associated neurocognitive dysfunction.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 7 Nov 2026 · from the full text of version 2 (10 Sept 2026) · CC BY 4.0

In ART-suppressed human brain, an SPP1+ activated microglial subset holds most detected HIV RNA and shows chronic interferon-I and MHC-I signatures, alongside complement signaling and neuronal loss.

ART does not eradicate brain viral reservoirs, and their identity and contribution to neurological injury remain poorly defined. The paper assigns the reservoir to an SPP1+ microglial subset and links it to interferon, MHC-I and complement pathways.

Key results

  1. ShownActivated and interferon-stimulated microglia expand with infection and stay expanded on ART. SPP1 signaling in the CellChat analysis rises in viremia and persists under ART.

    Activated MG: 738 (uninfected), 2,605 (viremic), 4,330 (ART-suppressed; 5.8-fold). ISG-activated MG: 4, 374, 539. CellChat interactions: 4,162, 4,574, 7,070. SPP1 pathway strength: 0.46, 4.18, 2.91. SPP1+ cells were 51.9%-98.8% of activated MG across five ART donors. · Aberrant SPP1 Signaling Marks Incomplete CNS…; SPP1-Driven MG-to-Neuron Signaling Persists…

  2. SuggestedAlmost all HIV RNA+ nuclei fall in activated microglia, mostly the SPP1+ subset, so this subset is proposed as the main brain reservoir. The counts are very small.

    6 of 7 HIV RNA+ cells were in SPP1+ MG (6/3299 vs 1/3178 in SPP1-). In scRNA-seq, HIV RNA+ MG were about 0.1% (1 of 1,324 in donor 1; 2 of 3,332 in donor 2). Zero HIV reads in activated MG of uninfected controls. · Aberrant SPP1 Signaling Marks Incomplete CNS…; SPP1+ Activated MG Subset Marks a Persistent…; HIV RNA+/SPP1+ Activated MG Subset Are…; Figure legends

  3. SuggestedHIV+ microglia on ART carry an interferon-I and MHC-I program that was absent in viremic donors. Chronic IFNβ on primary human microglia reproduces much of it, including SPP1 induction.

    snRNA-seq: 232 HIV+-specific vs 410 HIV--specific genes, with HIV+ MG enriched for MHC-I and IFN pathways. Primary MG from donors without HIV, IFNβ for five days (n=3 vs 3): 377 DEGs (329 up, 48 down); SPP1 rose 10-fold by qPCR; HLA-A and HLA-E up; IHC showed higher HLA-A and HLA-E in ART-suppressed brain. · HIV+ MG Show Persistent IFN-I/MHC-I…; IFN Signaling Directly Amplifies MHC-I…; IFN-I Signaling Broadly Activates MHC-I within…

  4. SuggestedViral transcription on ART is dominated by 3' nef reads. AlphaFold-Multimer modeling predicts Nef contacts with HLA-A and HLA-E, offered as a route to MHC-I evasion; this mechanism was not tested experimentally.

    Recovered HIV reads in donors 1-5 on ART showed the 3' LTR, antisense transcript ASP and the nef tail, with other genes silent. Nef protein detected by IHC in ART-suppressed and ART-interrupted brains. Predicted Nef-HLA-A contacts at 2.6-3.9 Å; HLA-E contacts at 3.50-3.84 Å. · Viral Sequence Recovery Identifies Nef-Mediated…; Sc/snRNA-Seq Analysis

  5. SuggestedSPP1+ microglia carry stronger complement and synapse-pruning signatures. Excitatory and inhibitory neuron abundance stays below baseline on ART, with DNA double-strand break marker staining.

    C3 5.22 in SPP1+ vs 1.62 in SPP1- MG; TREM2 and C1QC up in SPP1+ MG. Neuron abundance under ART: GABAergic 10-45%, glutamatergic 28-42% of uninfected. γH2AX rose after ART interruption and persisted on ART, alongside VGLUT1+ and GAD67+ neuron loss by IHC. · Inflamed SPP1+ MG and TREM2–C1QC/ “Eat Me”…; SPP1⁺ MG Associated with Aberrant DNA Damage…

Methods, data and reproducibility

Postmortem frontal-cortex tissue and isolated microglia from ART-suppressed, viremic and uninfected donors were profiled by scRNA-seq, snRNA-seq with a combined human plus HIV-1 reference, and IHC. Analyses used CellChat, differential expression, pathway enrichment and structural modeling; chronic IFNβ was tested in primary microglia by bulk RNA-seq and RT-qPCR.

  • snRNA-seq of postmortem human brain across ART-suppressed, viremic and uninfected donors · human · n = Results: 4 ART, 3 viremic, 3 uninfected; Methods: 5, 2, 2
  • scRNA-seq of isolated brain microglia from two ART-suppressed donors (donor 1, donor 2) · human · n = 2 donors
  • Total proviral HIV DNA in isolated brain microglia on suppressive ART · human · n = 5 donors
  • Bulk RNA-seq and RT-qPCR of primary microglia from HIV-negative brain, IFNβ for 5 days vs untreated · human · n = 3 treated vs 3 untreated
  • IHC of brain tissue (SPP1/CD68, HLA-A, HLA-E, Nef, γH2AX, VGLUT1, GAD67) · human
Methods & tools
Cell Ranger v9.0.1, Seurat v4.3.0, Harmony, CellChat v1.6.1, clusterProfiler v4.14.6, Ingenuity Pathway Analysis, STRING v11.5, AlphaFold-Multimer v2.3, DoubletFinder, ScType, QuPath v0.6.0, 10x Genomics Chromium 3'
Reagents & environment
Software versions stated (Cell Ranger v9.0.1, Seurat v4.3.0, CellChat v1.6.1, R v4.3.1); no container or requirements file.
Reproducibility gaps
Raw and processed snRNA-seq and bulk RNA-seq data are available only on request; no code or repository links given.
Limitations (4)
  • Age differences between postmortem donors cannot be controlled, and postmortem interval and comorbidities are possible confounders. · stated by the authors
  • ART is treated as a binary exposure (suppressed vs viremic); regimen-level pharmacokinetic analysis was not done and is not possible in postmortem tissue. · stated by the authors
  • Reservoir claims rest on very few HIV RNA+ cells (7 in the SPP1 comparison, about 0.1% of microglia) from a handful of donors; donor counts differ between Results (4/3/3) and Methods (5/2/2). · noted in this summary
  • The Nef-mediated MHC-I evasion rests on predicted structural contacts and transcript data; no functional test of Nef action in microglia is reported. · noted in this summary

Read this if you study HIV persistence in the brain and want a snRNA-seq and CellChat view of an SPP1+ microglial reservoir and its links to interferon, MHC-I and neuronal loss.

Citations
0

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.