Human epidemiology shows that CD8+ cytotoxic T lymphocyte (CTL) responses targeting epitopes in structurally constrained, topologically important ('highly networked') regions of HIV proteins are strongly associated with spontaneous viral control, but this association has never been tested causally by vaccination. The open problem is to determine, prospectively and in a controlled in vivo setting, whether an immunogen that focuses vaccine-elicited CTL responses on highly networked epitopes protects against acquisition, lowers the post-infection viral set point, or delays viral rebound. Further questions about the reach of the structure-based epitope network score — the equal weighting of its residue contributions, its dependence on high-resolution structures, global HLA coverage, and compensatory viral escape — bear on how generally and how durably such a result would hold, but the unperformed test is the causal one. As of August 2026, all follow-up work remains observational, ex vivo, or mechanistic; no prospective efficacy test of a network-epitope-focused immunogen exists in any species.

Background

CD8+ cytotoxic T lymphocytes (CTLs, the cytotoxic arm of T cell immunity) kill virus-infected cells after recognizing short viral peptides (epitopes, typically 8–13 amino acids) displayed on the cell surface bound to human leukocyte antigen (HLA) class I molecules. In untreated HIV infection the immune system eventually loses this battle: CTL responses select viral escape mutants, and the infection settles at a quasi-steady 'viral set point' — the blood viral load reached weeks after infection — which predicts long-term disease course. A small fraction of infected people, called elite controllers, instead maintain suppressed viremia for years without drugs, and genome-wide studies show this is partly linked to particular HLA class I alleles such as HLA-B57 (termed 'protective' alleles, in contrast to 'risk' alleles such as HLA-B35). Why some epitope-targeted CTL responses help control the virus and others do not has been a central question in HIV immunology.

Gaiha, Rossin, Urbach, et al. (Science, 2019) proposed a structural answer. They represented each HIV protein of known three-dimensional structure as a network graph whose nodes are amino-acid residues and whose edges encode physical interactions (contact distance, metal-coordinated bonds, local hydrophobic packing), then computed network centrality metrics — standard graph-theoretic measures of a node's relative importance in the topology — to assign every residue a quantitative 'network score'. Mutating residues with high network scores disproportionately impaired viral replication, whereas sequence conservation alone did not predict this mutational fragility. Aggregating residue scores at the positions that determine CTL recognition — the HLA-anchor residues that bind the peptide into the HLA groove, the central residues that contact the T cell receptor (TCR), and the flanking residues that influence proteasomal processing — yielded epitope-level scores. Epitopes presented by protective HLA alleles were enriched in high-scoring (highly networked) residues, and across a large cohort of controllers and progressors, CTL targeting of highly networked epitopes distinguished people who naturally controlled HIV, even among those lacking protective HLA alleles. The authors concluded that prophylactic and therapeutic vaccination 'should aim, in part, to induce functional CTL responses against topologically important epitopes', and noted that with the growing supply of high-resolution structures the approach could extend to other pathogens.

Three properties make this an actionable but untested vaccine-design principle rather than a settled one. First, the evidence is observational: association between network-epitope targeting and control does not establish that inducing such targeting by vaccination would protect. Second, the reach of the method has bounds that later work has had to confront: the epitope-level score is defined only for residues with high-quality structural data (the source computed epitope scores for the ~89.2% of optimally defined CTL epitopes with such data, leaving dynamic, disordered, or poorly resolved regions uncovered), and the composite score sums the anchor, TCR-contact, and flanking contributions equally — a deliberate design choice in the source, made 'to evenly capture the described mechanisms of viral epitope escape' — whose consequences for epitope rankings under alternative weightings or centrality metrics have not been characterized. Third, the CTL response a vaccine induces is shaped by immunodominance — the tendency of the immune system to respond to certain epitopes over others — so an immunogen must deliberately 'focus' responses onto subdominant, topologically constrained epitopes, which is exactly what natural infection often fails to do; and the virus may find compensatory (epistatic) mutations that restore fitness after escaping focused pressure. Follow-up work between 2021 and 2026 has built the mechanistic case (dendritic-cell-based ex vivo focusing of CTL responses onto network epitopes; crystal-structure explanations of how protective HLA alleles present a networked epitope and resist its escape variant; transfer of the network approach to SARS-CoV-2 and sarbecoviruses) but none of it constitutes a prospective test of protection.

Problem Statement

The problem, as posed by the source study and its unresolved follow-up literature, has one objective: to determine causally whether CD8+ T cell immunity deliberately focused on highly networked HIV epitopes is protective. Concretely: design and deliver an immunogen (peptide, vector, nucleic-acid, or cell-based) that focuses vaccine-elicited CTL responses onto epitopes ranking high in the structure-based epitope network score, and test it prospectively in a controlled in vivo setting — an animal challenge model (e.g., macaques challenged with simian immunodeficiency virus (SIV), the macaque lentivirus analogue of HIV, or with the chimeric simian-human immunodeficiency virus (SHIV), which carries the HIV envelope on an SIV backbone) or a human vaccination study with infection endpoints or analytical treatment interruption (a pre-planned pause of antiretroviral therapy to observe whether and how quickly virus rebounds) — asking whether such focusing prevents acquisition, lowers the viral set point, or delays rebound relative to appropriate unfocused or sham controls. Two terms fix the independent variable: 'epitopes ranking high in the structure-based epitope network score' means an epitope set fixed in advance by an explicitly stated threshold or score bin of that published score, chosen before any in vivo outcome data are collected; and 'focuses' means that vaccine-elicited CTL responses are directed toward those pre-selected epitopes to a greater extent than in the unfocused or sham control arm, as established by a quantitative immunogenicity criterion (for example the proportion, breadth, or magnitude of CTL responses to the pre-selected epitopes) that is specified in advance and evaluated independently of the efficacy endpoints. The required deliverable is a pre-registered prospective-test design — an immunogen; an epitope set fixed by an explicitly stated threshold or score bin of the published network score; unfocused or sham controls; and pre-specified immunogenicity and efficacy endpoints with an analysis plan — that, if executed, would settle the causal question; producing the executed in vivo determination is not required. Observational associations in natural controllers, ex vivo focusing demonstrations, and structural-mechanistic studies alone do not answer the question.

A second group of questions concerns the reach of the score rather than the causal test: how epitope rankings and control associations depend on the composite score's equal weighting of HLA-anchor, TCR-contact, and processing-flanking contributions and on the choice of network centrality metric; whether the approach extends to epitopes, regions, or conformations lacking high-resolution structures (including disordered and highly dynamic regions, e.g., via predicted or ensemble structures); how global HLA diversity constrains achievable population coverage; and what compensatory (epistatic) escape routes remain available to the virus when CTL pressure is concentrated on topologically constrained residues. They bear on how broadly and how durably any protection would generalize, but they are not required components of an answer to the causal question.

References

  1. Gaiha GD, Rossin EJ, Urbach J, Landeros C, Collins DR, Nwonu C, Muzhingi I, Anahtar MN, Waring OM, Piechocka-Trocha A, Waring M, Worrall DP, Ghebremichael MS, Newman RM, Power KA, Allen TM, Chodosh J, Walker BD. Structural topology defines protective CD8+ T cell epitopes in the HIV proteome. Science 364(6439):480-484 (2019). DOI: 10.1126/science.aav5095. https://doi.org/10.1126/science.aav5095
  2. Garcia-Bates TM, Palma ML, Anderko RR, Hsu DC, Ananworanich J, Korber BT, Gaiha GD, Phanuphak N, Thomas R, Tovanabutra S, Walker BD, Mellors JW, Piazza PA, Kroon E, Riddler SA, Michael NL, Rinaldo CR, Mailliard RB; I4C and RV254 Study Groups. Dendritic cells focus CTL responses toward highly conserved and topologically important HIV-1 epitopes. EBioMedicine 63:103175 (2021). DOI: 10.1016/j.ebiom.2020.103175. https://doi.org/10.1016/j.ebiom.2020.103175
  3. Nathan A, Rossin EJ, Kaseke C, Park RJ, Khatri A, Koundakjian D, Urbach JM, Singh NK, Bashirova A, Tano-Menka R, Senjobe F, Waring MT, Piechocka-Trocha A, Garcia-Beltran WF, Iafrate AJ, Naranbhai V, Carrington M, Walker BD, Gaiha GD. Structure-guided T cell vaccine design for SARS-CoV-2 variants and sarbecoviruses. Cell 184(17):4401-4413.e10 (2021). DOI: 10.1016/j.cell.2021.06.029. https://doi.org/10.1016/j.cell.2021.06.029
  4. Li X, Singh NK, Collins DR, Ng R, Zhang A, Lamothe-Molina PA, Shahinian P, Xu S, Tan K, Piechocka-Trocha A, Urbach JM, Weber JK, Gaiha GD, Takou Mbah OC, Huynh T, Cheever S, Chen J, Birnbaum M, Zhou R, Walker BD, Wang JH. Molecular basis of differential HLA class I-restricted T cell recognition of a highly networked HIV peptide. Nature Communications 14:2929 (2023). DOI: 10.1038/s41467-023-38573-8. https://doi.org/10.1038/s41467-023-38573-8
  5. Duette G, Lee E, Martins Costa Gomes G, Tungatt K, Doyle C, Stylianou VV, Lee A, Maddocks S, Taylor J, Khanna R, Bull RA, Martinello M, Sandgren KJ, Cunningham AL, Palmer S. Highly Networked SARS-CoV-2 Peptides Elicit T Cell Responses with Enhanced Specificity. ImmunoHorizons 7(6):508-527 (2023). DOI: 10.4049/immunohorizons.2300034. https://doi.org/10.4049/immunohorizons.2300034
  6. Bhattacharyya S, Crain CR, Goldberg B, Gaiha GD. Features of functional and dysfunctional CD8+ T cells to guide HIV vaccine development. Current Opinion in HIV and AIDS 18(5):257-263 (2023). DOI: 10.1097/COH.0000000000000812. https://doi.org/10.1097/COH.0000000000000812
  7. Borgo GM, Rutishauser RL. Generating and measuring effective vaccine-elicited HIV-specific CD8+ T cell responses. Current Opinion in HIV and AIDS 18(6):331-341 (2023). DOI: 10.1097/COH.0000000000000824. https://doi.org/10.1097/COH.0000000000000824
  8. Alrubayyi A, Traunbauer AK, Crain CR, Bhattacharyya S, Collins DR, Huang H, Kaseke C, Arshad U, Lian X, Vijayakumar A, Getz MA, Lichterfeld M, Yu XG, Walker BD, Lauer GM, Gaiha GD. NK-like and networked CD8+ T cell immunity mediates exceptional HIV control. bioRxiv preprint 2026.07.02.735504 (2026). DOI: 10.64898/2026.07.02.735504. https://doi.org/10.64898/2026.07.02.735504