CXCR3+ Th1 CD4 Memory Under Latent CMV: Clonal Imprinting Plus Chemokine-Axis Retention

Question: How does latent cytomegalovirus infection reshape pre-existing CD4 memory T cell subsets, particularly CXCR3+ Th1 populations, and what mechanisms maintain their long-term persistence in seropositive individuals?

Answer

Across the retrieved papers, the position this survey takes is that latent CMV does not inflate CD4 memory uniformly but instead selectively amplifies Th1-programmed clones while simultaneously widening the chemokine-receptive (CXCR3/CCR5) target pool that those clones and their bystanders occupy — and that persistence is best explained by a combination of clone-intrinsic phenotypic fixation laid down at priming and slow homeostatic self-renewal within a compartment that is barely replenished by new thymic output. Human single-cell and TCR evidence supports genuine CMV-specific clonal expansion of CD4 Th1 cells alongside CD4 TEMRA [1], and CMV seropositivity is uniquely associated with highly differentiated cytotoxic CD27−CD28− CD4 cells that EBV or HSV do not produce [13]. Yet the strongest direct evidence for a CXCR3 axis comes from a non-human primate model in which RhCMV expands CXCR3-high, CCR5+ memory CD4 cells in gut and lymph node with systemically elevated CXCL9 — and in which the majority of those expanded cells did not react to RhCMV lysate [17]. Taken together, these papers support a two-layer model: an antigen-specific Th1 core maintained by clonal stability, embedded in a larger chemotactically recruited, largely non-CMV-reactive CXCR3+ memory CD4 halo.

Latent CMV expands CD4 Th1 clones specifically, not the memory compartment as a whole

Integrated scRNA-seq, TCR-seq and flow cytometry across six human cohorts identifies CD4 Th1 cells as a newly recognized element of the CMV response, showing clonal expansion of Th1 cells alongside CD4 TEMRA, with the same CMV-associated clonal expansions reproduced in two independent cohorts and rapid shifts in circulating cell frequencies detectable in a single seroconverting individual [1]. The differentiation signature CMV imposes on CD4 cells is unusually specific: CMV-seropositive adults carry a substantial CD27−CD28− CD4 population expressing CD57 and cytotoxic granules, whereas equivalent cells are rare in EBV- or HSV-seropositive individuals and, when present in rheumatoid arthritis patients, are non-cytotoxic with higher TCF1, CD127 and Ki67 [13]. In melanoma cohorts, CMV drove expansion of both CD8+ and CD4+ TEMRA with enhanced effector function without reducing naive populations [10]. Antigen-level work is consistent: responses to the HCMV immediate-early 2 protein are CD4-dominant and dominated by polyfunctional Th1 cytokine production (IFN-γ+/TNF+/IL-2+) [12].

The CXCR3/CXCL9 axis appears to set the size of the recruitable CD4 pool, but human evidence for this is indirect

RhCMV-seropositive macaques show expansion of CCR5+ target CD4 T cells in gut and lymph node, cells that expressed high levels of CXCR3, with CXCL9 systemically elevated, and the expanded CCR5+ memory CD4 pool in gut was maintained during acute SIV with greater SIV DNA seeding in the intestine [17]. This positions CXCR3 not merely as a Th1 marker but as part of a virally regulated trafficking axis that determines where memory CD4 cells accumulate. In humans, CXCR3+ Th1 identity is the phenotype within which CMV-specific clones are found enriched [1], but the retrieved papers contain no human study that manipulates or longitudinally tracks CXCR3 expression itself on CMV-specific CD4 cells. The Th1 program is heritable rather than transient: repertoire-based mapping of sorted subsets onto scRNA-seq shows high intrinsic program sustainability of circulating Th clones across a four-year interval, and that cytotoxic CD4 cells arise clonally from Th1 clones [15].

What persists is a clone-intrinsic identity fixed early, not a continuously re-differentiated state

The clearest longitudinal evidence comes from CD8 cells but is directly relevant to the persistence question: tracking pp65- and IE1-specific clones over 4.5–12.5 years in latently infected donors shows individual clones with distinct specificity-dependent transcriptional profiles that arose during acute infection and were preserved with striking fidelity across decades of latency, tracking with relative TCR avidity and TCR sequence [3]. A CMC1-expressing state sharing features with tumor exhaustion remained phenotypically stable and clonally intact for over a decade of viral latency in that study [3]. The same principle of priming-imprinted durability operates for human CD4: mode of priming (infection versus vaccination) leaves CD4 transcriptional and clonal differences detectable 3–4 years later, with vaccine-primed cells preferentially adopting Tfh- and Th1-associated phenotypes [6]. Bulk repertoire analysis likewise shows greater TCR repertoire dispersal in CMV-positive individuals and progressive age-associated V-gene changes [14].

Persistence is sustained by self-renewal and early-seeded precursors, not by ongoing thymic recruitment

Fate-mapping in MCMV shows that recent thymic emigrants present at the time of infection expand and are retained in the long-lived memory compartment, generating more effector-memory cells, whereas CD8 cells made during the latent phase contribute little to the memory pool [2]. This argues that the inflating pool is seeded early and then self-sustained. Direct quantitative support for self-renewal comes from aged virtual memory CD8 cells, which derive predominantly from cells generated early in life and are maintained through continuous self-renewal, progressively outcompeting newly generated cells [8], and from CD4 TRM in skin, which self-renew with clonal half-lives of roughly five weeks and ~2% daily replacement from precursors [16]. A molecular handle on inflation exists: MK2 deficiency alters MCMV-specific CD8 dynamics, reducing non-inflationary cells acutely and enhancing expansion of inflationary subsets during persistence while impairing KLRG1 expression [7]. Tissue residence also contributes to durability, with CMV-responsive T cells detectable in myocardium of seropositive MI patients and virus-specific memory CD8 residency established in the heart after latent MCMV [4], and virus-specific airway CD4 and CD8 T cells persisting as TRM over 18+ months without decline [9].

A substantial fraction of "CXCR3+ Th1 inflation" may be bystander recruitment, and may not be functionally intact

The expanded CXCR3-high CCR5+ CD4 pool in RhCMV-seropositive macaques was largely non-reactive to RhCMV lysate, implicating chemokine-driven recruitment of non-CMV-specific cells [17]. Assay artifacts compound this: activation-induced marker assays activate CD39+ Treg-like and CD26hi Th22-like cells in a TCR-independent, cytokine-driven manner, producing an average 12-fold overestimation of the antigen-specific population, and CXCR4 downregulation rather than standard AIM readouts identifies genuinely clonally expanded, TCR-dependent memory CD4 cells [11]. Expanded CMV-associated CD4 cells are not necessarily protective: in CMV-seropositive rheumatoid arthritis patients, senescent CD4 cells showed reduced degranulation and lower cytokine production [13], and CMV-driven CD4 TEMRA expansion with enhanced effector function had no impact on clinical or immunological response to immune checkpoint blockade [10].

Conflicts and gaps

The central conflict is causal attribution. One body of work treats CD4 Th1 expansion as antigen-driven clonal selection, evidenced by CMV-specific TCR enrichment within Th1 cells and reproducibility across cohorts [1]; the primate work shows the expanded CXCR3-high CD4 pool is mostly not CMV-reactive and attributes it to a virus-regulated CXCL9/CXCR3 trafficking axis [17]. These are not reconciled by any retrieved paper. Second, the mechanistic inflation literature is overwhelmingly CD8-based and murine [2] [7], so it is unknown whether CD4 Th1 cells use the same MK2/KLRG1 or RTE-recruitment machinery. Third, the RTE finding that cells generated during latency contribute little to the memory pool [2] sits awkwardly with models requiring continuous antigen-driven restimulation under lifelong latency. Fourth, no retrieved study measures CXCR3 expression on tetramer-defined CMV-specific CD4 cells longitudinally, so CXCR3 remains a correlative marker with no demonstrated requirement. Fifth, whether the expanded compartment is protective or dysfunctional is unresolved, with enhanced effector function reported in one cohort [10] and blunted function in another [13]. Finally, MCMV downregulates MHC class II presentation machinery in dendritic cells, including Ciita and H2-DM components [5], which raises the untested possibility that CD4 persistence is maintained with minimal antigen presentation rather than by it.

What would change the answer

A longitudinal paired scRNA-seq/TCR-seq and ATAC-seq study of human CXCR3+ CD4 memory cells spanning seroconversion to a decade of latency — the design already executed for CD8 by Klebanoff and colleagues [3] — would discriminate clone-intrinsic fixation from ongoing chemokine recruitment. Combining peptide-MHC class II tetramers for IE2 and pp65 [12] with surface CXCR3 and the CXCR4−CD137+ gating correction [11] would directly quantify what fraction of CXCR3+ Th1 cells is CMV-specific versus bystander-recruited. Causality requires CXCR3- or CXCL9-blockade, or CXCR3-Cre fate mapping, during MCMV latency with parallel CD4 and CD8 readouts. Deuterium or BrdU labelling in seropositive humans, analysed with the fate-mapping-plus-mathematical-model framework applied to CD4 TRM [16], would supply the self-renewal and replacement rates that the current literature infers only indirectly.

References

[1] Grabauskas, T., Trinity, L., Verschoor, C. P., Nehar-Belaid, D., Marches, R., Thibodeau, A., Eryilmaz, G., Mahajan, A. S., Ravichandran, S., Picard, E., Kuo, C.-L., Schmader, K. E., Colon-Emeric, C., Whitson, H., Paust, S., Garcia-Sastre, A., Yaari, G., Banchereau, J., Kuchel, G., Ucar, D.. "CMV-specific clonal expansion of Th1, GZMK⁺ CD8⁺, and TEMRA T cells revealed by human PBMC single cell profiling". bioRxiv. 2025-10-18T00:00:00.000Z. URL: https://doi.org/10.1101/2025.06.24.661167

[2] Hilt, Z. T., Reynaldi, A., Steinhilber, M., Zhang, S., Wesnak, S. P., Smith, N. L., Davenport, M. P., Rudd, B. D.. "Recent thymic emigrants are preferentially recruited into the memory pool during persistent infection". bioRxiv. 2025-02-08T00:00:00.000Z. URL: https://doi.org/10.1101/2025.02.06.636722

[3] Klebanoff, S. D., Lawrance, M., O'Brien, K., Smithmyer, M. E., Speak, C., Codd, A. S., Ford, E. S., Koelle, D. M., Buckner, J. H., Newell, E. W., Campbell, D. J.. "Heterogeneity of HCMV-Specific CD8+ T Cells is Established Early in Infection, Clonally Encoded, and Stable Over Time". bioRxiv. 2026-10-16T00:00:00.000Z. URL: https://doi.org/10.64898/2026.10.09.758110

[4] Dapergola, E., Sustic, M., Ashour, D. E., Cokaric Brdovcak, M., Materljan Franki, J., Golemac, M., Lisnic, B., Boivin-Jahns, V., Krammer, T., Le Gouge, K., Rech, L., Stangl, V. M., Kashofer, K., Rainer, P. P., Jahns, R., Mariotti-Ferrandiz, E., Terekhov, M., Marwan, H., Keshtkar, M., Saliba, A.-E., Cochain, C., Jonjic, S., Hofmann, U., Frantz, S., Brizic, I., Gasteiger, G., Campos Ramos, G.. "Cytomegalovirus latency exacerbates cardiac inflammation and tissue remodeling after myocardial infarction". medRxiv. 2025-10-29T00:00:00.000Z. URL: https://doi.org/10.1101/2025.10.27.25338932

[5] Parry, R. H., McMillan, C. L. D., Bruce, K. L., Farrell, H. E.. "Global transcriptional reprogramming by cytomegalovirus infection suppresses MHC class II antigen presentation while enhancing migration machinery in murine dendritic cells.". bioRxiv. 2025-10-03T00:00:00.000Z. URL: https://doi.org/10.1101/2025.03.31.646504

[6] Antoun, E., Liu, G., Jayathilaka, D., Yao, X., Rostron, T., Waugh, C., Clark, K., Sopp, P., Fry, J., Xia, T., Mentzer, A., Knight, J., Peng, Y., Dong, T.. "Mode of T cell priming durably shapes the TCR repertoire, effector function and α4β1 integrin expression of human virus-specific CD4+ T cells". bioRxiv. 2026-08-19T00:00:00.000Z. URL: https://doi.org/10.64898/2026.08.18.745226

[7] Panagioti, E., Yu, X., Kong, Y. W., Makakova, K., Mercado, N. B., Lawler, S. E., Yaffe, M. B., Cook, C. H.. "MAPKAP Kinase 2 Orchestrates Memory T Cell Inflation in Cytomegalovirus Infection". bioRxiv. 2026-01-20T00:00:00.000Z. URL: https://doi.org/10.64898/2026.01.16.699974

[8] Chiu, B. C.. "Lifelong self-renewal and competition shape the virtual memory CD8 T-cell compartment during aging". bioRxiv. 2026-06-11T00:00:00.000Z. URL: https://doi.org/10.64898/2026.06.07.730686

[9] Ramirez, S. I., Faraji, F., Lopez, P. G., Garin-Ortega, L., Hern, N. A., Eisenman, O., Azhan, A., Frazier, A., Phillips, E. J., Mallal, S., Sette, A., Dileepan, T., Jenkins, M. K., Crotty, S.. "Durability and diversity of human upper airway T cell memory". bioRxiv. 2026-09-18T00:00:00.000Z. URL: https://doi.org/10.64898/2026.09.15.751278

[10] Edwards, J. M., Senthi, S., Smith, R., Burridge, H., Owens, C., Shackleton, M., Andrews, M. C., van Zelm, M. C.. "Age and CMV are not associated with clinical or immunological response to immune checkpoint blockade". medRxiv. 2026-08-06T00:00:00.000Z. URL: https://doi.org/10.64898/2026.08.04.26359748

[11] Zheng, M. Z. M., Burmas, L., Tan, H.-X., Trieu, M.-C., Lee, H. J., Rawlinson, D., Haque, A. P., Kent, S. J., Wheatley, A. K., Juno, J. A.. "Deconvoluting TCR-dependent & -independent activation is vital for reliable Ag-specific CD4+ T cell characterization by AIM assay". bioRxiv. 2024-12-13T00:00:00.000Z. URL: https://doi.org/10.1101/2024.12.10.627643

[12] Panagioti, E., Welters, M. J. P., Arens, R., van der Burg, S.. "High-throughput epitope screening of the Human Cytomegalovirus immediate early protein 2 identifies promising antigenic T cell targets". bioRxiv. 2026-01-26T00:00:00.000Z. URL: https://doi.org/10.64898/2026.01.24.701419

[13] Williams, L., Saber, A. O., Xu, R., Jung, H., Awad, S., Shahane, A., Baker, J. F., Su, L. F.. "Divergent effects of cytomegalovirus and rheumatoid arthritis on senescent CD4+ T cells". bioRxiv. 2024-12-16T00:00:00.000Z. URL: https://doi.org/10.1101/2024.12.11.627966

[14] Pan, M., Hu, J., Wang, Y. T., Tan, Y., Hu, H., Yang, Z., Fleming, J., Pang, Y., Li, B.. "Antigen-aware TCR embeddings reveal conserved immune patterns for personalized immune monitoring". bioRxiv. 2026-10-16T00:00:00.000Z. URL: https://doi.org/10.64898/2026.10.09.757444

[15] Lukyanov, D. K., Kriukova, V. V., Ladell, K., Shagina, I. A., Staroverov, D. B., Minasian, B., Fedosova, A. S., Shelyakin, P., Suchalko, O. N., Komkov, A., Blagodatskikh, K. A., Miners, K. L., Britanova, O. V., Franke, A., Price, D. A., Chudakov, D. M.. "Repertoire-based mapping and time-tracking of helper T cell subsets in scRNA-Seq". bioRxiv. 2024-12-09T00:00:00.000Z. URL: https://doi.org/10.1101/2023.10.24.563704

[16] Chandler, J., Bullock, M. E., Swain, A., Williams, C., van Dorp, C., Seddon, B., Yates, A. J.. "Tissue resident memory CD4+ T cells are sustained by site-specific levels of self-renewal and continuous replacement". bioRxiv. 2025-03-17T00:00:00.000Z. URL: https://doi.org/10.1101/2024.09.26.615039

[17] Perdios, C., Suresh Babu, N., Coleman, C. D., Brown, A. T., Mostrom, M. J., Allers, C., Doyle-Meyers, L., Fennessey, C. M., Keele, B. F., Kaur, A., Freeman, M. L., Mudd, J. C.. "RhCMV Expands CCR5 Memory T Cells and promotes SIV reservoir seeding in the Gut Mucosa". bioRxiv. 2025-01-08T00:00:00.000Z. URL: https://doi.org/10.1101/2025.01.06.631163

How this survey was made

This survey answers one question: “How does latent cytomegalovirus infection reshape pre-existing CD4 memory T cell subsets, particularly CXCR3+ Th1 populations, and what mechanisms maintain their long-term persistence in seropositive individuals?”

We keep our own collection of recent preprints and refresh it every day from bioRxiv, medRxiv, arXiv, PsyArXiv, SocArXiv and ChemRxiv. For each survey, we search that collection for recent work and keep the strongest matches: papers that use the same terms as the question, and papers on the same topic in different words.

The text was written from the titles and abstracts of those papers; the full texts were not read. Statements that draw on a particular paper are marked with a number in brackets. The reference list just above is the 17 papers whose work this text draws on, chosen from the 100 papers those searches returned; the rest of the shortlist is not printed. It is not a complete review of everything published on the question.

Treat this as a starting point and a reading list, not a peer-reviewed review. The searches behind this survey ran on October 22, 2026 and covered work from the previous two years.