Cytotoxic CD4 T Cells: Dual Agents in HIV-1 Pathogenesis and Persistence

Abstract

Purpose of Review

Cytotoxic CD4 T cells (CD4 CTL) have long been recognized for their potentially protective role in people with HIV, but only recently have their contributions in HIV-1 pathogenesis and viral persistence been appreciated. This review summarizes evidence highlighting their critical role in HIV-1-mediated CD4 T cell depletion and in maintaining the viral reservoir, with special consideration for their abundance and development in the gut.

Recent Findings

CD4 CTL are increased in frequency in people with HIV (PWH). Granzyme B activity in CD4 CTL promotes HIV-1 mediated death of gut CD4 T cells. CD4 CTL that express survival markers such as BCL-2, TNFR2/CD120b, and OX40 appear to resist HIV-1 mediated cell death, potentially contributing to the viral reservoir. Multiple cytokine (IL-2, IL-15) and transcriptional (BACH2, EOMES) pathways were implicated in CD4 CTL differentiation and maintenance. In the gut, CD4 CTL activity appears to be intricately linked to the microbiome, as cytotoxic protein expression can develop in response to bacterial exposure. CD4 CTL from the gut, induced by the presence of bacteria in vitro, are highly infectable by HIV-1.

Summary

CD4 CTL may act as dual agents in HIV-1 infection, amplifying tissue damage and serving as resilient cellular reservoirs. Understanding the mechanisms regulating their differentiation, cytotoxicity, and survival could inform new therapeutic strategies aimed at restoring gut mucosal integrity, resolving chronic inflammation, and targeting the persistent HIV-1 reservoir.

Introduction

Cytotoxic CD4 T cells (CD4 CTL) represent an intriguing and understudied component of HIV-1 immunology. While CD4 T cells are traditionally viewed as helper cells that coordinate adaptive immunity through cytokine secretion and provision of help to B cells and CD8 T cells, a growing body of evidence highlighted that a subset could acquire direct cytolytic function. These CD4 CTL are loosely defined based on the expression of at least one cytotoxicity marker such as Granzyme A, B, H and K; perforin, Fas/Fas-L, TRAIL, GNLY, 4-1BB, SLAMF7, NKG7, NKG2D, or ICAM-1, all of which were associated with CD8 T cell and natural killer (NK) cell cytotoxicity. Originally dismissed as artifacts of in vitro activation, CD4 CTL have now been identified in both health and disease in vivo [1, 2], including several viral infections, autoimmune disorders, and malignancies (Table 1). Diverse mechanisms of cytotoxicity were reported, ranging from specific to nonspecific, which have been reviewed elsewhere [1–8].

Table 1.

Observations of presence of CD4 CTL in humans during infection, cancer, and autoimmune conditions

Presence of CD4 CTL
Condition	Definition	Reference
HIV-1 (acute)	CD38 + + CD57- BCL2lo IFNγ + Ki67 TIA-1+	[24]
HIV-1 (chronic)	GZMA + perforin + TIA-1/GMP-17+	[46, 127]
Cytomegalovirus (acute)	CD27- CD28 + IFNγ + GZMB + TNFα	[128]
Cytomegalovirus (latent)	CD107a + GZMA + GZMB + IFNγ + MIP1B + Perforin + TNF+	[129]
Cytomegalovirus (post-transplant)	CD28- CX3CR1 + NKG2D + Perforin+	[130]
Epstein-Barr virus	Eomes + T-bet+	[131]
Influenza	perforin + GZMB+	[132]
Vaccinia	GZMB GZMA GZMK CD161 Rab27a granulysin TIA-1 Perforin	[133]
Herpes Simplex Virus	Fas-L Perforin GZMB	[134]
Hepatitis B	Perforin+	[135]
Hepatitis C	Perforin+	[135]
Hepatitis D	Perforin+	[135]
Dengue	Perforin+	[136]
Parvovirus	CD57+, GZMB+, Perforin+, IL-17+	[137]
Hantavirus	GZMB + Perforin+, CD107a+	[138]
Human Papillomavirus	CD28- NKG2D+	[139]
Lung Cancer	GZMB GZMA PRF1 GZMH KLRK1 CTSW CST7	[140]
Colorectal Cancer	NKG7 PRF1 GZMH GNLY FGFBP2 CX3CR1	[141]
Hepatocellular Carcinoma	GZMA GZMB NKG7 GNLY	[142]
Breast Cancer	ABCB1 APBA2 SLAMF7 GPR18 PEG10	[143]
Head And Neck Cancer	GZMA GZMB	[144]
Osteosarcoma	GZMA NKG7	[145]
Melanoma	GZMA GZMB EOMES SLAMF7	[144]
Bladder Cancer	GZMA GZMB NKG7 PRF1 GNLY GZMK INFG TNF	[146]
Crohn’s Disease	Eomes RUNX3 IL-17 NK receptors IFNγ GZMB Perforin	[147, 148]
Intestinal Colitis	RUNX3 Eomes CD8a GZMB IFNγ T-bet CRTAM	[36, 147, 149]
Multiple Sclerosis	Eomes + GZMB + Perforin + IFNγ + IL-17 + T-bet + NK receptors + CD8 + CX3CR1 + CD28-	[150, 151]
Rheumatoid Arthritis	NK receptors + IL-17 + IFNγ + CD8 + CXCR1 + CD28-	[151]
Acute Coronary Syndrome	CD4 + CD28- Perforin+	[152]
Atherosclerosis	CD4 + CD28- IFNG + PRF1 + KIR2DS2+	[153]

In people with HIV (PWH), CD4 CTL occupy a unique and paradoxical position, as both potential mediators of immune control and contributors to tissue damage and reservoir persistence. Strategies to counteract HIV-1 immunopathogenesis may thus require a deep mechanistic understanding of CD4 CTL biology.

CD4 CTL During HIV-1 Infection

CD4 CTL employ mechanisms analogous to classical cytotoxic CD8 T cells (Fig. 1) [9]. Upon antigen recognition via the major histocompatibility complex II (MHC-II), they form immunological synapses with target cells, releasing cytotoxic granules that contain perforin, a protein that forms pores in target cells [10], and granzymes. In particular, Granzyme B (GZMB) cleaves over 300 intracellular substrates via the IEPD motif [11] that include caspase-3, 6, 7, 9 and Bid, which can trigger cell death via extrinsic apoptosis [12]. In the context of CD8 CTL, this process leads to the killing of HIV-1-infected cells expressing viral peptides loaded on to MHC-I, temporally correlating with the initial drop in viremia among PWH [13–16].

Fig. 1.

Mechanisms of CD4 CTL target cell identification and GZMB entry. (A) GZMB + CD4 T cell containing GZMB releases active GZMB intracellularly; without the anti-GZMB protein SerpinB9 to counteract GZMB effects, this leads to apoptosis in the effector CD4 CTL. (B-C) TCR recognition of peptide-MHC-II (pMHC-II) results in classical immune synapse formation and granule release from the effector cell with perforin mediated entry into the target cell. (C-E) Non-specific cytotoxic granule release without pMHC-II recognition could lead to target cell death through perforin mediated entry or non-perforin mediated entry into the target cell. Granzyme B could cleave inactive, pro-forms of Caspase-3, Caspase-7, and Caspase-8 into their biologically active form, leading to apoptosis. GZMB can also cleave Bid which leads to cytochrome c release from mitochondria and apoptosis. Cytotoxic granules consist of cytotoxic proteins and perforin, a pore forming protein which oligomerizes and leads to the entry into target cells of the cytotoxic proteins. Caspase independent pathways leading to cell death include: ICAD cleavage to CAD causing apoptosis, Bid cleavage leading to SMAC/Diablo activation and apoptosis, ROCK2 cleavage inducing membrane blebbing, and Lamin A, lamin C, lamin B1, or NUMA1 cleavage leading to lamina disruption and loss of nuclear integrity to cause cell death

HIV-1-specific (MHC-II-restricted) CD4 CTL with cytolytic capabilities were isolated from peripheral blood mononuclear cells (PBMC) of PWH during acute and chronic stages (Table 2). These CD4 CTL had the ability to in vitro lyse viral protein pulsed cells [17], cells expressing viral proteins [18, 19], or HIV-infected target cells [20]. CD4 CTL viral specificity most commonly mapped to Gag [17, 19, 21–27] but some cells could instead target Env [19] or Nef [20, 22]. The cytotoxic mechanism included perforin mediated cytotoxicity (defined through blockade of perforin activity), granzyme B activity, degranulation by surface CD107a expression, and non-calcium dependent cytotoxicity (granule independent mechanism). Co-incubation of HIV-1-specific CD4 CTLs with HIV-1-specific CD8 CTLs and infected CD4 T cells in vitro resulted in better inhibition of viral replication and lower infected CD4 counts [26]. This led to the conclusion that HIV-1 specific CD4 CTL work synergistically with CD8 CTL to help control HIV-1 infection [26]. When present in vivo, HIV-1 specific CD4 CTL have been associated with lower viral load set point [27]. Elite controllers of HIV-1, e.g., individuals with a lack of disease progression and low viral load despite no ART, show higher levels of HIV-1-specific CD4 CTLs than chronic progressors [26]. Together, these data indicate a possibly protective role for HIV-1 specific CD4 CTLs during HIV-1 infection. However, potential pathogenic consequences were raised due to increased susceptibility to infection and dysregulated/non-specific cytotoxic activity.

Table 2.

Observed HIV-1 antigen specific CD4 CTL during HIV-1 infection

HIV-1 specific cytotoxic/lytic killing of target cells
Phenotype/CD4 CTL markers	Effector Mechanism	Specificity	Reference
CD4+	non-calcium dependent cytotoxicity	Gag, Env	[19]
GZMB perforin IL-4 IL-4 IFNγ MIP1a MIP1B TNFα	perforin-mediated cytotoxicity, Fas dependent cytotoxicity	Gag	[21]
Perforin	perforin-mediated cytotoxicity, IFNγ, TNFα, β-chemokines secretion	Gag	[17]
IFNγ IL-4 MIP1B CCL5	IFNγ secretion	Gag, Nef	[22]
CD45RO + CCR7- CCR5 + CD161 + CD244 + GZMA GZMB Perforin + IFNγ BCL2 CD127	Granzyme B/Perforin-mediated cytotoxicity IFNγ secretion	Gag	[23]
GZMB GZMA perforin CD107a IFNγ	MHC-II restricted degranulation, Granzyme B/Perforin-mediated cytotoxicity, IFNγ secretion	Nef	[20]
CD38+++ KI67 + TIA + BCL2lo IL-2 + IL7R + CCR5 + GZMB + Perforin+	Granzyme B/Perforin-mediated cytotoxicity	Gag	[24]
CD107a IFNγ	Degranulation (CD107a expression), IL-2, IFNγ secretion	Gag	[25]
CD57 + KLRG1 + IL-7R- GZMB + GZMA + Perforin + GZMK + IFNγ+	Degranulation (CD107a expression), IL-2, IFNγ secretion, Granzyme B/Perforin-mediated cytotoxicity IFNγ secretion	Gag	[27]
CD107a IFNγ GZMB Perforin KLRG1 CD57 NKG2D Granulysin EOMES	Granzyme B/Perforin-mediated cytotoxicity IFNγ secretion	Gag	[26]
GZMB NKG7 IL2RA CCL3 CCL4 IFNG CD40L TNFRSF9	Granzyme B/Perforin-mediated cytotoxicity IFNγ secretion	Gag	[64]

The gastrointestinal (GI) tract harbors a large fraction of CD4 T cells in the human body that are susceptible to HIV-1 due to the high expression of CCR5 and α4β7 [28, 29]. Consequently, the gut is the site of early, profound and irreversible CD4 T cell loss among PWH [30, 31]. We reported that, despite overall CD4 T cell depletion, GZMB + CD4 T cells are enriched in colon biopsies from chronic PWH relative to age- and sex-matched controls [24]. It remains unknown if these gut CD4 CTL are HIV-1-specific and contribute to virologic control. However, given that high levels of HIV-1 RNA can still be detected in this tissue compartment when substantial CD4 T cell depletion has occurred, these surviving CD4 CTL likely harbor persistent HIV-1.

CD4 CTL Ontogeny

Given strong evidence for CD4 CTL involvement in HIV-1 infection, defining the signals and transcriptional programs that drive CD4 CTL differentiation can provide insights on how these cells balance cytotoxicity, survival and susceptibility. Multiple mechanisms have been described for CD4 CTL differentiation (Fig. 2). Transcription factors T-Bet, EOMES, and RUNX3 are central regulators of cytotoxic programming in CD8 CTL [32] and may also be relevant for CD4 CTL. EOMES RNA are elevated in HIV-1-infected (HIV+) CD4 CTL from PWH [33]. In murine CD4 T cells, ThPOK suppresses RUNX3-dependent cytotoxic gene expression, and ThPOK downregulation permits the development of MHC-II-restricted CD4 CTL [34]. Type I interferon signaling also supports CD4 CTL differentiation: CD4 T cells from IFNAR1-/- mice produce lower GZMB and perforin during influenza A virus (IAV) infection [35]. In mice, CRTAM (class I-restricted T-cell associated molecule) signaling induces cytotoxic activity and CD4 CTL gene expression including GZMB, perforin, Eomes, and IFN-γ [36]. The transcription factor BLIMP-1 enhances GZMB expression in conjunction with IL-2 and IFNα in cancer and IAV models [37, 38]. In addition, TCR stimulation alone can induce GZMB expression and cytotoxicity in CD4 T cells directed against regulatory T cells [39]. Thus, CD4 CTL differentiation can arise through multiple transcriptional and cytokine-driven pathways.

Fig. 2.

Pathways of CD4 CTL development. Cytotoxic CD4 T Lymphocytes are defined in multiple ways throughout the literature. The broad classifications include the expression of various granzyme genes (A, B, H, K, M) and/or expression of other factors including perforin and Fas/Fas-L. The transcription factors involved in the expression of these factors vary. The pathways known to induce cytotoxic molecule expression in CD4 T cells include: IL-15, IL-2, TCR stimulation, CRTAM signaling (sometimes in conjunction with IL-2 signaling), type I interferon signaling, and retinoic acid with TGFβ. The main transcription factors that have been linked to CD4 CTL development are EOMES, T-Bet, BLIMP-1, and RUNX3

The pathways that trigger CD4 CTL development may be tissue-specific [23]. HIV-1 infection is associated with dysregulation of the gut epithelial barrier, resulting in the translocation of bacteria and/or bacterial products from the lumen to the underlying lamina propria (LP) [40–43]. In the rhesus macaque-SIV model, microbial translocation was observed as early as 8 days post-infection, prior to extensive CD4 T cell depletion [43, 44]. Co-incubation of lamina propria mononuclear cells (LPMC) with bacterial lysates in vitro led to a dramatic increase in GZMB expression on CD4 T cells [45]. However, GZMB expression did not occur in similarly stimulated CD4 T cells from disaggregated tonsil tissue [46]. Blockade of IL-2 signaling led to an almost complete cessation of bacteria induced GZMB, while co-incubation of LPMC with IL-15 promoted GZMB induction in gut CD4 T cells [46], consistent with a critical role for IL-2 and/or IL-15 in CD4 CTL development in other contexts [35, 37, 46, 47]. Blockade of HLA-DR led to a 32% reduction in bacteria-driven gut CD4 CTL frequency [46], suggesting the involvement of antigen-presenting cells. In particular, gut dendritic cells (DC) are critical for bacteria-mediated enhancement of gut CD4 T cell proliferation [51]. Of note, the majority of CD4 T cells in the gut are tissue-resident (non-circulating TRM) [48–50], and may be primed to specifically respond to the microbiota. The specific bacterial antigens driving gut CD4 CTL development remain unknown, but multiple species belonging to the Prevotella, Escherichia, Ruminococcus, Bacteroides and Acinetobacter genera could also induce GZMB in CD4 T cells to various extents [24], suggesting that there may be shared bacterial antigens and/or stimulatory pathways involved. In addition, HIV-1 infection of LPMC leads to enhanced frequency of GZMB expression in gut CD4 T cells [45, 52]. It would be of interest to determine if viral proteins such as Vpr and Nef that are known to alter the cellular transcriptome and proteome [53–55] may contribute to overall CD4 CTL induction in the gut.

The development of CD4 CTL observed in the peripheral blood and CD4 CTL that develop from the gut appear to be different. The HIV-1 specific CD4 CTL observed in peripheral blood of PWH clearly required HIV-1 antigen stimulation to develop [22–24]. However, acute CD4 CTL development in the gut does not require viral infection or viral antigens to be present and can occur through bacterial exposure alone [46]. PBMC do not show this same ability to produce cytotoxic proteins in response to bacteria or bacterial products [46]. CD4 CTL development in the gut from bacterial exposure may thus be an innate rather than antigen specific process, as it is unlikely that the TCR of majority of CD4 T cells from the gut would be specific for the same bacteria (such as E. coli). This innate process in the gut may be exploited by HIV-1 as CD4 CTL are more active, more proliferative, and more frequently infected than non-cytotoxic counterparts in the same conditions [46, 52]. In other words, CD4 CTL appear to be a ‘desirable’ cell population for HIV-1 survivability.

GZMB Mechanism of Killing

The profound bacteria-mediated upregulation of GZMB in gut CD4 T cells may help explain why the gut is such a vulnerable site for HIV-1-mediated CD4 T cell depletion. HIV-1-mediated CD4 T cell death was extensively studied in the Human Lymphoid Aggregate Culture (HLAC) model, which involve infecting disaggregated tonsil mononuclear cells with X4-tropic HIV-1 [56, 57]. These studies revealed a critical role for caspase-1-mediated pyroptosis and IFI16-mediated innate sensing [57, 58]. Building on this framework, we established the Lamina Propria Aggregate Culture (LPAC) model [59], which utilizes LPMC and transmitted/founder HIV-1, which are primarily R5-tropic [60, 61]. Using this model, we discovered that in the presence of bacteria HIV-1-mediated cell death mechanism shifted from pyroptosis to caspase-3-mediated apoptosis [62]. Importantly, inhibition of GZMB activity rescued gut CD4 T cells from HIV-1-mediated cell death [23].

The precise mechanism by which CD4 T cell-intrinsic GZMB activity participates in HIV-1-mediated gut CD4 T cell death remains uncertain. Two main possibilities can be envisioned: (1) a “suicide” mechanism, in which GZMB released intracellularly by the producing CD4 CTL triggers its own apoptosis; or (2) a “bystander” mechanism, in which GZMB released from one CD4 T cell enters and kills neighboring, often uninfected, target cells (Fig. 1). We evaluate these possibilities below:

1. Self-directed (suicide) death appears unlikely. If GZMB escaped from cytotoxic granules into the cytoplasm of the effector CD4 T cell, it could in theory induce apoptosis through caspase activation. However, in the LPAC model, GZMB + and GZMB- CD4 T cell subsets show similar levels of depletion [52, 62], and GZMB + CD4 CTL are enriched in the gut tissue of PWH [33, 46, 63–65]. If self-inflicted death were common, GZMB- cells would be expected to preferentially survive, but this has not been observed. Thus, available evidence argues against GZMB-mediated suicide as a major contributor to gut CD4 T cell loss.

2. Bystander killing is a more plausible explanation. In the LPAC model, GZMB-dependent killing in the context of HIV-1 infection occurs even when CD4 T cells were derived from HIV-uninfected donors [52], implying that the cytotoxic response does not require HIV-1-specific T cell receptor (TCR) recognition. Classical CD8 CTL release cytotoxic granules only after a precise TCR-peptide-MHC-I (pMHCI) interaction, but whether gut CD4 CTL require analogous TCR-pMHC-II engagement for degranulation is unclear [1]. Given that most HIV-1-induced CD4 T cell death affects uninfected bystander cells [66, 67], GZMB released nonspecifically by CD4 CTL could amplify this bystander apoptosis. Supporting this idea, CD4 CTL are more prone than CD8 CTL to non-specific or “bystander” degranulation [68, 69], suggesting that GZMB may sometimes be discharged without strict TCR-pMHCII signaling.

A key unresolved question is how GZMB produced by CD4 CTL enters target cells when perforin is absent or at very low levels. In the LPAC model, only about one-third of GZMB + CD4 T cells co-express perforin [46], leaving the majority without this conventional means of granzyme delivery. Further, single-cell transcriptomic analyses of CD4 CTL from PWH show heterogenous and often low perforin expression [63, 64, 70]. We speculate that alternative pathways for GZMB delivery may exist. One possibility is that non-perforin pore-forming proteins produced either by host cells or commensal enteric bacteria could create transient membrane disruptions that allow GZMB uptake. Bacterial toxins such as ClyA (in Salmonella), YaxAB (Yersinia enterocolitica), or suilysin (Streptococcus suis) can form such pores [71]. In addition, endogenous human antimicrobial peptides found in the intestinal milieu that include Reg3 lectins, defensins, and cathelicidins can also permeabilize membranes [72]. These naturally occurring pores could provide a conduit for extracellular GZMB to enter nearby cells, promoting apoptosis even in the absence of perforin.

Granzyme Activity Beyond Direct Cytotoxicity

Beyond its role in inducing apoptosis, GZMB may indirectly amplify CD4 T cell loss by promoting inflammation and disrupting epithelial integrity. GZMB activity can enhance the production and release of inflammatory cytokines such as TNFα, IL-6, and IL-8 [73–75], and it can cleave inactive cytokine precursors including IL-1α and IL-18 into their active forms [76, 77]. Induction of these cytokines would heighten immune activation and render CD4 T cells more permissive to HIV-1 infection. GZMB’s protease activity can also degrade tight junction proteins within the intestinal epithelium [78, 79]. The resulting barrier dysfunction facilitates continued microbial translocation, further driving immune activation and inflammation. Together, these data suggest that GZMB may contribute to HIV-associated immunopathology not only through direct cytotoxicity, but also by fueling inflammation and compromising gut mucosal barrier function, amplifying the cycle of immune activation and CD4 T cell loss.

Other Granzymes may also contribute to HIV-1 associated immunopathology. Similar to GZMB, Granzyme A (GZMA) can cleave IL-1β into the active form [80]. GZMA can also induce other proinflammatory cytokines such as TNFα and IL-6 [81] that could contribute to chronic immune activation in PWH. Granzyme K (GZMK) may also contribute to immune cell activation, through induction of IL-6 and MCP-1 (monocyte chemotactic protein-1) and activation of endothelial cells through cleavage of PAR-1 (protease-activated receptor-1) [82]. GZMK is also known to activate the complement cascade [83] which could contribute to both barrier disruption and immune cell activation in PWH. Granzyme M (GZMM) is associated with protective or deleterious effects in different inflammatory contexts. GZMM played a role in increasing TLR4 induced inflammation [84], but also had a protective effect in mouse models of inflammatory bowel disease [85]. Overall, multiple granzymes beyond GZMB could be contributing to HIV-1 associated inflammation through non-cytotoxic mechanisms in underappreciated ways.

CD4 CTL in HIV-1 Persistence

Not all CD4 T cells die following HIV-1 infection. Following antiretroviral therapy (ART), the CD4 T cells that survive long-term serve as reservoirs of replication-competent HIV-1, re-initiating plasma viremia if ART is stopped. Due to this barrier to cure, major efforts have been made to unravel the molecular nature of the HIV-1 reservoir [87–89]. Identifying and targeting of the viral reservoir is exceedingly difficult, as reservoir cells are rare (~ 1/10⁶ in blood of PWH on ART) [86–89]. Most HIV-1 DNA/RNA + cells harbor defective provirus [90–92]. The heterogenous nature of HIV-1 infected cells results in a lack of a defining HIV-1 infected cell marker [93] (reviewed in [94–96]). Despite these difficulties, multiple studies to molecularly define the HIV-1 reservoir are ongoing. Due to easier availability of samples and cells, most of these studies initially focused on peripheral blood. Central memory CD4 T cells (TCM) have particularly gained attention given the long-lived nature of these cells [97–99]. Signaling proteins linked to cell survival such as CD30 [100, 101], OX40, the transcription factors BACH2 and Aiolos, and mitochondrial associated protein BCL-2 are protective for HIV-1 + CD4 T cells (Fig. 3). Treatment of PBMC from PWH ex vivo with an antibody-drug conjugate that targets CD30, brentuximab vedotin, reduced total HIV-1 DNA [100]. A single case of a PWH taking brentuximab vedotin for lymphoma treatment had no detectable virus in PBMC [102]. OX40 on HIV-1 + CD4 T cells, seen initially in screens of PBMC infected with HIV-1 in vitro, induces Survivin (BIRC5) expression [45, 103]. Survivin was hypothesized to protect HIV-infected CD4 T cells from apoptotic cell death and small molecule inhibitors were suggested as potential treatment agents [103, 104].

Fig. 3.

HIV-1 + CD4 T cell survival pathways/signaling. Cell surface receptors associated with HIV-1 + CD4 T cell survival include CD30, CD120b/TNFR2, and OX40 which signal through TRAF signal transducers to activate several transcription factors associated cell survival. CD30 activates AP-1 leading to IL-6 expression, cell activation and survival, and NFκB signaling to stimulate IL-2 production enhancing cell activation, proliferation and activation. CD120b signals through various secondary signal transducers to activate transcription factors associated with activation, cell survival, and proliferation including AP-1, STAT5, and NFκB. OX40 signaling results in NFκB mediated induction of antiapoptotic proteins Survivin (BIRC5), BCL-2 and BCL-XL. Transcription factor Aiolos in HIV-1 + CD4 T activates NFκB signaling, resulting in enhanced cell proliferation, cell survival and BCL-2 expression. BACH2 also increases infected CD4 T cell survival, activation, and proliferation, but the upstream mechanisms are poorly understood. BCL-2, CD120b, and OX40 have been specifically detected in CD4 CTL

Increased BCL-2 expression in HIV-1 + versus HIV-1 negative CD4 T cells were also observed in multiple contexts [64, 70, 105, 106]. BCL-2 is involved in the sequestration of proteins that cause apoptosis by permeabilizing the outer mitochondrial membrane, thus suppressing apoptotic cell death [107]. BCL-2 was preferentially upregulated in the HIV-1 inducible reservoir in PWH [108]. BCL-2^hi cells occur at a higher frequency in PBMC from PWH versus people without HIV-1 [109]. The SIV reservoir is also enriched in BCL-2^hi cells compared to other subsets [110]. Ex vivo, BCL-2^hi cells from PWH were more resistant to HIV-1 specific CD8 T cell responses [111]. In these in vitro studies, CD8 T cell-resistant cells were CD4 CTL that co-expressed BCL-2 [112]. In humanized mice, rhesus macaques, and in vitro PBMC models of infection, the BCL-2 inhibitor venetoclax reduced the size of the HIV-1 reservoir [113–115]. A phase I clinical trial investigating the safety of venetoclax among PWH is ongoing [116].

Through single-cell multi-omic sequencing technology, it was found that among PWH on ART, persistent HIV-1 DNA or RNA is found in significant numbers of CD4 CTL (Table 3). Multiple persistence and survival pathways have been proposed. Collora et al. hypothesized that the trending increase in BCL-XL transcripts and SerpinB9 anti-apoptotic proteins may facilitate persistence and survival of HIV-1 RNA + cells CD4 CTL [64]. CST7 was proposed by Weymar et al. as a possible mechanism for evading elimination, as it reduces the effect of NK cell cytotoxicity [70]. Wei et al. found that Aiolos/IKZF3 led to HIV-1 RNA + cell persistence and survival by promoting proliferation [33]. Of note, although multiple groups found CD4 CTL in the HIV reservoir from PBMC, not all multi-omics studies have identified this cell type [117]. This was likely due to the complex heterogeneity of the reservoir, the low number of HIV-1 infected cells (due to ART control of infection), or differences in how an infected cell was defined (i.e. defective or intact provirus, HIV-1 RNA + versus HIV-1 DNA+).

Table 3.

Observation of CD4 CTL presence in PWH

CD4 CTL Presence in PWH
Phenotype/CD4 CTL markers	Cell Source	Viremia Status of Participants	HIV Infection Defined by	Reference
CD3 + CD8- GZMB+	Colon Tissue	Viremic	Intracellular infection not defined	[46]
GZMB GZMA GZMM NKG7 NPM1	PBMC	Viremic	RNA	[63]
GZMB GZMK Perforin IFNγ TNFα IL-2	PBMC	Viremic and ART	RNA	[64]
GZMB GZMH NKG7 IFNG ZEB2	PBMC	Viremic and ART	DNA RNA	[33]
GZMA GZMK CST7 LYAR DUSP2 CRTAM HLA-DR	PBMC	ART	DNA RNA	[70]
CD45RO + CD11a + CCR7- Perforin+	PBMC	Viremic and ART	Intracellular infection not defined	[127]
GZMA GZMB GZMH GZMM PRF1 GNLY NKG7 IFNG TBX21	PBMC	Viremic (infected CD4 CTL absent in ART)	RNA	[65]

In contrast to the blood, less is known about the HIV-1 reservoir in tissues. This is a major knowledge gap in the field, as it was estimated that the gut could contain up to 80–95% of the total infected HIV-1 reservoir [118, 119]. Emerging data from the transplantation field revealed that gut TRM cells are long-lived [49] and, therefore, could potentially sustain the viral reservoir. Among PWH not on ART, CD4 CTL are readily detected in both peripheral blood and colon biopsies [24, 46]. In the LPAC model, a significant fraction of gut CD4 CTL that survive HIV-1-mediated CD4 T cell death were HIV-1-infected and expressed OX40 and TNFR2 [23]. TNFR2 inhibits apoptosis by recruiting TRAF2 and cIAP2-containing complexes to activate survival-promoting NF-kB and PI3K/Akt pathways [120, 121] and we demonstrated that blockade of TNFR2 promoted the death of surviving CD4 CTL in vitro [52]. Viral factors, such as HIV-1 Nef, were also reported to have anti-apoptotic properties [122, 123] (reviewed in [124]).

Several survival mechanisms observed in PBMCs have also been observed in gut tissue. In situ hybridization identified co-localization of HIV-1 RNA and CD30 in ileal and rectal tissue from both viremic and ART suppressed PWH [100]. OX40 + CD4 T cells detected by immunohistochemistry in gut tissue of PWH on long-term ART correlated with total ileum HIV-1 DNA copies per million cells [125]. A recent single-cell transcriptomic study using CD4 T cells from colon biopsies of 10 PWH on ART recovered 99 HIV-1 DNA/RNA + cells [126]. These cells were primarily TRM cells with elevated BACH2 activity and a survival-oriented transcriptional program linked to IL7R, OX40, GITR and TNFR2, consistent with our LPAC results [52]. Notably, GZMB was not detected in these persistently-infected gut CD4 T cells in vivo [126]. One possibility is that during ART, GZMB may be downregulated in these long-lived TRM cells, while survival factors remain elevated. Modifying the LPAC model to include relevant antiretroviral drugs (an ‘LPAC-ART’ model) may facilitate deeper mechanistic understanding of the phenotypic changes that occur among surviving HIV-1-infected CD4 CTL during long-term ART.

Collectively, the studies outlined above reveal potential survival mechanisms that may sustain the HIV-1 reservoir. Targeting these interconnected survival pathways, either individually or in combination, may offer new opportunities to selectively deplete the circulating and tissue HIV-1 reservoir and move closer to durable remission or cure.

Conclusion

CD4 CTL represent a double-edged sword of HIV-1 pathogenesis, capable of both amplifying tissue damage and sustaining viral persistence. Their Granzyme-mediated activity may exacerbate bystander CD4 T cell loss in the gut, while survival pathways driven by TNFR2, BCL2, OX40, Aiolos, CD30 and BACH2 may enable long-term maintenance of the viral reservoir. Defining how CD4 CTL arise, persist, and contribute to immune dysfunction remains a major research frontier in HIV-1 immunology. Emerging studies suggest that CD4 CTL may exist along a continuum of effector and survival states, suggesting the possibility of reprogramming their fate by redirecting cytotoxic and survival pathways toward viral clearance and gut mucosal repair. Reshaping CD4 CTL biology could thus transform these cells from persistent hosts into instruments of HIV-1 eradication.

Key References

• Collora JA, Liu R, Pinto-Santini D, Ravindra N, Ganoza C, Lama JR, et al. Single-cell multiomics reveals persistence of HIV-1 in expanded cytotoxic T cell clones. Immunity. 2022;55(6):1013–31 e7.

  • A study that established HIV-1 RNA+ CD4 T cells from the PBMC of PWH on ART are enriched for cytotoxic CD4 T cells. Also found that HIV-1 RNA+ cells had large clone size, established during viremia, and persisted after viral suppression with ART.
• Johnson S, Eller M, Teigler JE, Maloveste SM, Schultz BT, Soghoian DZ, et al. Cooperativity of HIV-Specific Cytolytic CD4 T Cells and CD8 T Cells in Control of HIV Viremia. Journal of Virology. 2015;89(15):7494–505.

  • HIV-1 specific cytotoxic CD4 T cells from people with HIV defined with transcriptional, phenotypic, and functional analyses. Indicated CD4 T cytotoxic ability is similar to that of CD8 T cell cytotoxicity. Showed HIV-1 specific CD4 CTL develop early in acute infection and had an association with an earlier viral set point.
• Dillon SM, Mickens KL, Thompson TA, Cooper EH, Nesladek S, Christians AJ, et al. Granzyme B(+) CD4 T cells accumulate in the colon during chronic HIV-1 infection. Gut Microbes. 2022;14(1):2045852.

  • Tissue histology establishes granzyme B+ CD4 T cells accumulate and are enriched in the colon of PWH relative to people without HIV in vivo.In vitro commensal bacteria alone can upregulate granzyme B expression in gut-derived CD4 T cells, at higher frequency than occurs in peripheral blood or tonsil-derived CD4 T cells.
• Herrera A, Leyre L, Weiler J, Linden NL, Huynh TT, Wang F, et al. Multi-Omic Atlas reveals cytotoxic phenotype and ROS-linked metabolic quiescence as key features of CTL-resistant HIV-infected CD4 (+) T-cells. bioRxiv. 2024.

  • Study indicating that CD4 CTL are overrepresented in reservoir cells harboring HIV-1. Also, that reservoir cells (enriched in CD4 CTL) are resistant to CD8 CTL elimination.
• Soghoian DZ, Jessen H, Flanders M, Sierra-Davidson K, Cutler S, Pertel T, et al. HIV-Specific Cytolytic CD4 T Cell Responses During Acute HIV Infection Predict Disease Outcome. Science Translational Medicine. 2012;4(123).

  • Longitudinal study of cohort of PWH, in elite controllers from this cohort HIV specific CD4 CTL are higher than in progressors. The development of GZMA+ CD4 CTL was predictive of clinical outcome and slower progression.
• Geretz A, Ehrenberg PK, Clifford RJ, Laliberte A, Prelli Bozzo C, Eiser D, et al. Single-cell transcriptomics identifies prothymosin alpha restriction of HIV-1 in vivo. Sci Transl Med. 2023;15(707):eadg0873.

  • Multi-omics sequencing of PBMC from PWH during acute infection and 48 weeks after ART initiation. The highest abundance of HIV-1 RNA transcripts were found in memory CD4 T cells expressing cytotoxic markers.

Author Contributions

Original draft preparation by K.M. and M.S. K.M. prepared figures and tables. Review, editing, and additional concepts to be included S.D. K.G. C.W. All authors reviewed and approved the final manuscript.

Funding

This work was supported by the NIH R01 AI145428, AI108404 and AI134220 (CCW and MLS), the Colorado Clinical and Translational Sciences Institute NIH UM1 TR004399 and T32 TR004367 (KLM), the NIH Molecular Pathogenesis of Infectious Disease T32 AI052066 (KLM), the University of Colorado RNA Bioscience Institute (MLS, CCW, KG) and the University of Colorado ASPIRE Program (MLS, CCW, SMD, KG).

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Human and Animal Rights

All the reported studies/experiments with human or animal subjects performed by the authors have been previously published and complied with all applicable ethical standards (including Helsinki declaration and its amendments, institutional/national research committee standards, and international/national/institutional guidelines).

Competing interests

The authors declare no competing interests.