Why CD4+ T cells matter in adoptive cell therapy for solid cancers

Abstract

Cytotoxic CD8+ T cells have been the focus of research efforts to improve the efficacy of anticancer adoptive cell therapies. However, emerging evidence is highlighting the critical role of CD4+ T cells in these treatments, including chimeric antigen receptor-T cells. Although the presence of CD4+ T cells is known to be important for efficacy of cell therapies, the mechanisms by which CD4+ chimeric antigen receptor-T cells exert their beneficial effects remain incompletely understood. In this article, we review the roles of CD4+ T cells in adoptive cell therapy, highlighting understudied areas that are yet to be explored, and propose new directions to advance our understanding of their antitumour potential.

Highlights

• •CD8+ T cells have been the main focus to date of efforts to improve anticancer adoptive cell therapy.
• •CD4+ T cells play both a key helper role in enhancing CD8+ T cell function and mediate direct anticancer activity.
• •Autologous or engineered (TCR/CAR-T) CD4+ T cells have demonstrated tumour control in preclinical and clinical studies.
• •Further understanding of both the indirect and direct mechanisms of CD4+ T cell anticancer activity will be critical for improving outcomes of adoptive cell therapies in the future.

Introduction

Adoptive cell therapy (ACT), or transfer of tumour-specific T cells that have been expanded ex vivo and reinfused into patients, has improved outcomes for patients with advanced solid cancers, including melanoma and carcinoma.1, 2, 3 Several forms of ACT have been established, including harvesting of tumour infiltrating lymphocytes (TILs) from primary tumours, followed by development of T cell receptor (TCR) and chimeric antigen receptor (CAR)-T cells.⁴ Development of CAR-T cell therapy has especially transformed the treatment of haematological cancers. In people with relapsed/refractory diffuse large B cell lymphoma (DLBCL), treatment with anti-CD19 CAR-T cell therapy can provide long-term remission in 40% of all treated patients, whereas previously this population had a dismal prognosis, with an expected survival of only 4 months.^5,6 However, there remains scope for optimisation and development of more potent CAR-T cells that demonstrate long-term persistence. Recent data indicates characteristics of the tumour microenvironment (TME) in patients with B cell lymphoma predicts response to CD19 CAR-T therapy, with an inflamed phenotype promoting CD8+ T cell exhaustion, associated with worse response.⁷ Similarly, ACT for solid tumours has faced several hurdles, including contending with an immunosuppressive TME, greater tumour antigen heterogeneity and difficulties with T cell trafficking.⁸ In attempts to improve ACT efficacy and persistence in a solid tumour context, much attention has focussed on CD8+ CAR-T cells. However, there is emerging evidence for the vital role of CD4+ T cells in ACT, both directly and indirectly. It has been demonstrated both in vitro and in vivo that in the absence of CD4+ T cell support, CD8+ T cells fail to expand and sustain tumour control, likely due to reduced proliferation, greater T cell exhaustion and failure to generate functional memory.9, 10, 11 There is also increasing recognition for the ability of CD4+ T cells to exert effector and cytotoxic activity in the context of cancer. This review discusses the current understanding of the role of CD4+ T cells in ACT.

CD4+ T cells in cancer

CD4+ T cells play an essential role in promoting antitumour activity through several mechanisms, the most widely recognised of which is their helper role, supporting cytotoxic CD8+ T cell function. However, the direct cytotoxic potential of CD4+ T cells is being increasingly recognised, with reports of MHC class II-dependent tumour cell killing by CD4+ T cells, and more recently, activation of effector macrophages.12, 13, 14 In contrast, subsets of CD4+ T cells are known to promote tumour growth, including CD25+FOXP3+ T regulatory cells (T_REG), and more recently, cytotoxic T regulatory 1 (Tr1) cells characterised by high IL-10 production,¹⁵ highlighting the diverse functionality of the total CD4+ T cell population.^16,17 Although this review focuses on CD4+ T cells in the context of ACT, several excellent reviews are cited that explore their broader role in cancer. The reader is encouraged to refer to these sources for a deeper understanding of CD4+ T cell biology.^13,18, 19, 20, 21

The role of CD4+ T cells in adoptive immunotherapy for solid tumours

The majority of preclinical and clinical trials of adoptive cell transfer using transgenic T cells or human TILs, respectively, have been carried out using CD8+ T cells alone and have demonstrated antitumour activity in advanced/treatment-resistant cancers. Murine models of aggressive prostate carcinoma, renal carcinoma, metastatic sarcoma and melanoma have all shown impressive tumour regression after treatment with variably generated and administered CD8+ ACT.22, 23, 24, 25 Given these findings, subsequent clinical studies utilised a similar approach for treatment of advanced, treatment-refractory solid cancers. A phase I study adoptively transferred melan-A specific cytotoxic CD8+ T cells to patients with refractory stage IV melanoma, reporting three objective responses, including one complete response lasting >2 years. No responses in visceral lesions were reported and persistence of infused cells was not detected several weeks after cell transfer.²⁶ Subsequent studies have pointed to various factors influencing CD8+ ACT persistence, including development of a terminally differentiated phenotype and upregulation of checkpoint molecules of these cells, which impair CD8+ cytotoxic function.^27,28

Helper CD4+ T cells play a key role in promoting antitumour activity and their presence may be essential for optimal efficacy of ACT. One preclinical study demonstrated that transfer of antigen-specific tyrosinase-related protein 1 (TRP-1) CD4+ T cells with CD8+ T cells was essential for control of solid tumours in vivo and that CD8+ cells from treated mice were less exhausted²⁹ (Table 1).^12,30, 31, 32, 33, 34, 35, 36, 37, 38 Another study reported that presence of T follicular helper (T_FH) cells was essential for CD8+ mediated antitumour responses with anti-PD-L1 therapy.³⁰

Table 1.

Relevant preclinical studies investigating the role of CD4+ T cells in ACT for solid cancers

ACT	Model system	Summary of findings relating to role of CD4+ T cell in ACT	Reference
Adoptive transfer of donor-derived tumour-nonspecific or TRP-1-specific TFH	C57BL/6; melanoma (B16), breast cancer (4T1), colorectal cancer (CT26)	•Presence of CD4+ TFH cells reduced CD8+ T cell exhaustion and enhanced PD-L1 checkpoint inhibitor therapy in various murine solid tumour models •scRNA sequencing obtained from databases + primary samples showed an association between improved progression/survival and high TFH	Niogret et al.30
Adoptive transfer of TRP-1-specific CD4+ T cells	C57BL/6 mice with mutated TRP1; spontaneous melanoma (Grm-1 mice), melanoma (B16)	•Control of established tumours expressing MHC II following ACT in lymphodepleted mice, independent of other endogenous immune cells •CTLA-4 blockade enhanced antitumour effects and reduced levels of suppressive TREG cells	Quezada et al.31
Adoptive transfer of spleen and lymph node-derived CD4+ T cells	C57BL/6 MHC II−/− mice and TRP-1 transgenic; melanoma (B16), colorectal cancer (MC38)	•Adoptive transfer of CD4+ T cells + OX40 costimulation/CTLA-4 blockade released proinflammatory cytokines •Cytokines promoted neutrophil recruitment, increasing nitric oxide synthase, mediating inflammatory-cell death in antigen loss variants •Effects were observed even in MHC II absence on host or tumour cells	Hirschhorn et al.32
Adoptive transfer of CD4+ TRP-1-specific T cells	C57BL/6 and TRP-1 transgenic; melanoma (B16, HCmel12)	•CD4+ T cells induced inflammatory tumour cell death mediated via IFN-γ and TNF-α •Cytotoxicity was observed even in MHC I negative tumours not controlled by CD8+ T cells	Kruse et al.12
Adoptive transfer of TRP-1-specific TH9 T cells or mesothelin-targeting TH9 CAR-T cells	C57BL/6; melanoma (B16), ovarian cancer (Ovcar)	•Antigen loss variants could be eradicated by tumour-specific TH9 T cells via direct killing and bystander effects mediated by type I IFN-γ and recruitment of endogenous monocytes	Xue et al.33
HER-2 targeting CAR-T	BALB/c scid mice; breast cancer (MDA-MB-435 HER-2 xenograft)	•Presence of CD4+ HER-2 CAR-T cells in combination with CD8+ CAR-T cells at a 1 : 1 ratio resulted in superior tumour control compared with CD8+ CAR-T cells alone	Moeller et al.34
HER-2 targeting CAR-T	BALB/c scid mice; breast cancer (MDA-MB-435 HER-2 xenograft)	•Polarised CD4+ TH1 HER-2 CAR-T cells in combination with CD8+ CAR-T showed greater resistance to tumour rechallenge compared with TH2	Moeller et al.35
Mesothelin-targeting CAR-T delivered intrapleurally	NSG; pleural mesothelioma (MSTO-211H)	•Potent cytokine secretion and proliferation was observed in early-activated and stimulated CD4+ but not CD8+ CAR-T cells •CD8+ CAR-T effector function required presence of CD4+ CAR-T •Direct CD4+ CAR-T cytotoxicity was also observed	Adusumilli et al.36
IL13RAα2-targeting CAR-T	NSG; glioblastoma (PBT103-2)	•Infusion with mixed CD4/8 CAR-T cells: CD4+ but not CD8+ CAR+ component persisted following tumour challenge •CD8+ only CAR-T cells were unable to control glioblastoma, whereas presence of CD4+ CAR-T cells resulted in tumour control >200 days	Wang et al.37
LeY-targeting CAR-T	In vitro culture of CD8+ CAR-T cells with/without CD4+ CAR-T cells during production/stimulation of CAR	•Absence of CD4+ CAR-T cells during production/stimulation resulted in impaired CD8+ CAR-T proliferation and cytokine production	Meyran et al.38

ACT, adoptive cell therapy; CAR-T cell, chimeric antigen receptor-T cell; HER-2, human epidermal growth factor receptor 2; IL13RAα2, interleukin-13 receptor subunit α2; IFN, interferon; LeY, Lewis Y antigen; NSG, non-obese diabetic (NOD)/severe combined immunodeficiency (SCID) gamma; PD-L1, programmed death-ligand 1; scRNA, single-cell RNA; T_FH, follicular T helper cell; TIL, tumour infiltrating lymphocyte; TNF, tumour necrosis factor; T_REG, regulatory T cell; TRP-1, tyrosinase-related protein 1.

Although antitumour responses have been reported when infusing isolated CD8+ T cells for ACT in murine models, including prostate carcinoma and fibrosarcoma, clinical studies report less impressive responses.^25,26,39, 40, 41 A phase I study using MART-1 and gp100-specific CD8+ peripheral T cells that were adoptively transferred into patients with metastatic melanoma without preconditioning reported only minor responses (no RECIST [Response Evaluation Criteria in Solid Tumor] criteria was applied) in two of ten treated patients, with remaining patients experiencing stable or progressive disease.³⁹ Another related study failed to show clinical benefit in patients treated with antigen-specific peripheral T cells or TILs that were largely CD8+, and noted poor peripheral persistence as one causative factor.⁴² After lymphodepletion and adoptive transfer of g209-specific CD8+ T cells into patients with refractory metastatic melanoma, clinical responses remained limited.⁴⁰ Although there are multiple possible causes for the lack of clinical response to these therapies, such as the potential for central tolerance to the differentiation antigens MART-1 and gp100, as well as lack of standardised lymphodepletion before cell transfer, the authors postulated that without CD4+ T cell support, CD8+ T cells failed to traffic to tumour sites or underwent apoptosis.³⁹ Additionally, there are several clinical reports that demonstrate tumour control using adoptive transfer of autologous, circulating CD4+ T cells alone.^3,43

In one report, the authors harvested, expanded and reinfused peripheral CD4+ T cells specific for the tumour-associated antigen (TAA) NY-ESO-1 from a patient with recurrent, metastatic melanoma.⁴³ It was postulated that infusion with CD4+ T cells would enable stimulation of endogenous CD8+ T cells and that the CD4+ T cells themselves would persist in the patient. Complete resolution of measurable disease was achieved within 2 months of infusion, which persisted for over 22 months, including controlling NY-ESO-1-negative cells.⁴³ Even more encouraging was the demonstration of epitope spreading, whereby T cells were activated by TAAs such as MART-1 or MAGE-3, but not by antigens absent from the melanoma.⁴³ Although this case report is impressive, little progress has been made in understanding the mechanisms behind the effects of CD4+ T cells in this form of ACT and in establishing robust protocols for selecting and expanding the appropriate T cell subsets for clinical ACT. The question therefore remains whether the addition of CD4+ T cells into ACT may have altered outcomes in earlier trials where CD8+ ACT was administered alone. Other forms of ACT including CAR-T cell therapy have standardised production protocols in a clinical setting and the role of engineered CD4+ T cells in these treatments has already shown promise in preclinical studies.

Act of tumour-infiltrating CD4+ lymphocytes

One of the earliest forms of ACT involved ex vivo expansion and infusion of TILs from solid tumours, which showed moderate clinically efficacy, particularly against melanoma, with 50%-70% response rates and up to 40% complete responses observed in refractory cases.^2,44 Most recently, a phase 3 multicentre clinical trial compared treatment of patients with refractory melanoma with TILs compared with the anti-CTLA-4 antibody, ipilimumab, and reported and significantly longer overall survival in the TIL group (median overall survival 25.8 months compared with 18.9 months).² Within these studies, the infused TILs consisted of a heterogenous mix of CD4:CD8 T cells. Although early data suggested TILs enriched for CD8+ T cells may have been associated with more favourable responses than general TIL therapy, this was not shown in a randomised controlled trial in patients with melanoma.⁴⁵ Furthermore, when only CD8+ T cell clones were utilised, their clinical efficacy was limited.⁴⁵ It is possible that when bulk TILs are expanded and infused, the total population of CD4+ T cells includes a subset of immune suppressive cells, so-called bystander cells or exhausted cells, which may explain why some studies show increased presence of CD4+ T cells as detrimental to outcome.⁴⁶ This was demonstrated by a recent study that utilised RNA sequencing to define T cell subsets in TILs administered to melanoma patients, showing greater populations of T_H2, T_H17 and T_REG transcripts in patients with lower numbers of TILs, which was associated with shorter overall survival.⁴⁷ In contrast, another group used a similar analysis to evaluate CD4+ TILs from a patient cohort with metastatic melanoma with varying response to treatment and identified that the predominant neoantigen-specific T cell clones were of CD4+ origin. Furthermore, these clones induced an effector response against autologous tumour cells that was MHC II-dependent.⁴⁸

Considerable preclinical data have shown a role for cytotoxic CD4+ T cells in mediating antitumour immunity against solid tumours.^31,49,50 Early in vivo work suggested CD4+ T cells may play a direct role in killing MHC II-positive tumours, or indirectly controlling MHC II-negative tumours, mediated by IFN-γ.^49,51,52 Subsequently, more recent studies have demonstrated direct CD4+ cytotoxicity in the absence of MHC II expression on tumour cells. Kruse et al.¹² showed that although indirect CD4+ killing via antigen-presenting cell (APC) interactions at the tumour margins, along with IFN-γ-activated macrophages, adoptively transferred CD4+ T cells can directly initiate an inflammatory tumour cell death in models that lack MHC expression and are unresponsive to IFN-γ. Hischhorn et al.³² reported that adoptively transferred CD4+ T cells with OX40 costimulation or CTLA-4 blockade killed MHC II-deficient tumours through an inflammatory process that is dependent on presence of endogenous neutrophils. This combination treatment resulted in elimination of antigen-negative tumour cells. Finally, Draghi et al.⁴⁹ used clustered regularly interspaced short palindromic repeats (CRISPR) to knock out B2M, CIITA and JAK in ex vivo patient-derived melanoma cells to resemble immune checkpoint-resistant melanoma. Using CD4+ TILs cocultured with these edited autologous melanoma cells, they demonstrated that CD4+ TIL cytotoxicity persisted, despite B2M knock out or impaired IFN-γ signalling; however, it was abrogated by CIITA knockout, suggesting MHC II-dependence (Table 1).

Clinical reports for CD4+ TIL therapy are lacking. One group infused autologous tumour-reactive infiltrating CD4+ T lymphocytes into a patient with metastatic cholangiocarcinoma, resulting in partial regression of metastatic lesions for 7 months (Table 2). The infused CD4+ T cells were characterised as polyfunctional T_H1 cells, due to robust expression of IFN-γ, tumor necrosis factor (TNF)-α and IL-2, and phenotyped as largely effector memory CD4+ T cells with cytotoxic potential.³ Pulmonary metastatic lesions resected >1 year later contained the adoptively transferred CD4+ T cells, which the authors report as evidence that these cells were contributing to disease stabilisation. However, applying these results at a population level is likely to be less rewarding. One study screened 44 TIL samples for CD4+ tumour reactivity in vitro and reported that only 20.5% of screened metastatic melanoma samples contained such cells, illustrating that 80% of TILs did not contain antigen-specific CD4+ T cells, thus would not have harboured antitumour activity.⁵¹ Numerous approaches for improving TIL therapy through enrichment of particular T cell subsets are being investigated, however a clear understanding for the role of CD4+ T cell subsets in TIL therapy remains incomplete.⁴⁶ Furthermore, the ability to identify and select antigen-specific clones is currently still lacking.⁴⁶ Given these variable findings, there is rationale to move toward infusion of engineered antigen-specific, and defined subsets of T cells to promote greater uniformity of treatment between patients and reduce interpatient variability.

Table 2.

Relevant clinical studies investigating the role of CD4+ T cells in ACT for solid cancers

ACT	Patient cohort and treatment description	Summary of findings relating to role of CD4+ T cell in ACT	Reference
Ex vivo activated CD4+ T cells with NY-ESO-1 specificity	Single case •Patient with refractory metastatic melanoma •Treatment with a single infusion of clonal CD4+ T cells without preconditioning or cytokine support.	•Complete resolution of measurable disease within 2 months of infusion •Persistence over 22 months •Disease control included regression of NY-ESO-1 negative tumour cells	Hunder et al.43
TIL	Single case •Patient with metastatic cholangiocarcinoma •Treatment with TILs containing ∼25% tumour-specific polyfunctional TH1 cells with IL-2 cytokine support	•Initial partial regression observed for >1 year •Retreatment with tumour-reactive TH1 TILs at disease progression resulted in regression of metastatic lesions for >6 months	Tran et al.3
TIL	Single case •Patient with advanced refractory metastatic melanoma •Treatment with bulk TIL resulted in disease stabilisation (per RECIST) and clinical improvement	•Lack of tumour reactivity by CD8+ TILs, inferring that the initial tumour response was likely mediated by CD4+ TILs	Friedman et al.51
TIL	•Evaluation of TIL products from n = 6 patients with metastatic melanoma	•Neoantigen-reactive TIL clones were largely CD4+ (not associated with response to treatment). •CD4+ TIL exerted an effector response against autologous tumour that was MHC II dependent	Hall et al.48
TIL	•Evaluation of TIL subsets from n = 24 patients with metastatic melanoma	•Proportion of CD4+ TILs negatively correlated with number of infused TILs •These CD4+ subsets appeared to be largely regulatory T cells, TH2 and TH17 •A subset of patients with long overall survival had high frequencies of CD4+ TILs	Thompson et al.47
TIL	•In vitro coculture using patient-derived TILs and matched autologous melanoma cells	•CD4+ TILs mediated melanoma cytotoxicity independent from tumour ianus activated kinase/signal transducer and activator of transcription signalling •CD4+ TILs killed melanoma cells independently from IFN-γ-induced MHC II upregulation	Draghi et al.49
TCR therapy using CD4+ T cells engineered with MAGE-A3 TCR	•Clinical trial n = 17 patients with metastatic/locally advanced refractory cancer •Patients received cyclophosphamide/fludarabine, followed by low or high-dose CD4+ TCR cells	•One complete remission (low-dose cohort) and three partial responses observed, lasting 4-29 months •No association between response and TCR peripheral persistence •Moderate side effects (fever, hepatic and renal toxicities)	Lu et al.53

ACT, adoptive cell therapy; JAK, Janus kinase; MAGE-A3, MAGE family member A3; MHC, major histocompatibility complex; STAT, signal transducer and activator of transcription; TCR, T cell receptor; TIL, tumour infiltrating lymphocyte.

TCR therapy

TCR therapy involves the introduction of specific TCR genes (typically α and β) into T cells, enabling production of tumour antigen-specific T cells. Compared with CAR-T, TCR-T cells are able to identify intracellular antigens expressed on MHC class I molecules, potentially increasing the repertoire of TAAs able to be targeted.⁵³ One preclinical study adoptively transferred a small number of TCR-modified CD4+ T cells, and in combination with immune checkpoint blockade, this resulted in eradication of established melanomas in vivo, mediated by MHC II expression on tumours (Table 1).³¹ It was shown that the inclusion of CTLA-4 blockade was necessary to improve T cell expansion, as well as hinder accumulation of T_REG cells.³¹ Adoptive transfer of a subset of circulating T_H9 CD4+ T cells, containing a modified TCR against TPR-1 was able to induce regression of established B16 melanomas in mouse models, even controlling antigen loss variants through IFN-γ signalling and recruitment of endogenous monocytes (Table 1).³³

Despite the promising preclinical data, limited clinical studies have solely focussed on CD4+ T cells in TCR therapy. One study treated 17 patients with metastatic or locally advanced refractory/recurrent solid tumours expressing the cancer germline antigen MAGE-3 with either a low or high dose of autologous CD4+ T cells with modified TCR against MAGE-A3 after lymphodepleting chemotherapy (Table 2).⁵³ Among both dosing cohorts, one patient with metastatic cervical cancer experienced a complete remission for >29 months and there were three partial responses observed that lasted between 4 and >18 months. Grade 2 fever and grade 2, 3 and 4 hepatic and renal toxicities were observed. Levels of infused CD4+ TCR-T cells were higher 1-month after infusion in the high-dose group; however, there was no association between response and TCR-T cell persistence. In two of the patients that achieved a partial response, CD4+ TCR cells were identified in the circulation at time of progression, suggesting antigen escape or dysfunctional antigen presentation. Although there is clear evidence that CD4+ T cells are critical for successful TIL and TCR therapy, their ability to directly recognise and eliminate tumour cells may be limited by the frequent absence of MHC class II expression on tumour cells and reliance on functional APC pathways.⁵⁴ In contrast, CAR-T cells are able to target tumour cells independently of MHC class II or APCs, allowing the potential for cytotoxic CD4+ CAR-T cells to directly target and eliminate tumour cells as well as assist CD8+ CAR-T cells.

The role of CD4+ CAR-T cells

A major barrier to durable tumour control after CAR-T cell infusion is long-term persistence of the circulating engineered T cells. A landmark report described circulating CAR-T cells collected from two patients with chronic lymphocytic leukaemia (CLL) who remained in clinical remission for >10 years after the initial infusion.⁵⁵ The CAR-T cell populations from both patients were largely composed of CD4+ T cells, which surpassed CD8+ T cells as the dominant population ∼1 year after infusion. These CD4+ T cells expressed high levels of Ki67 and showed a distinct cell surface protein expression profile, including activation markers CD38, HLA-DR and CD95, transcription factors eomesdermin and thymocyte selection-associated high mobility group box protein, and memory markers CCR7 and CD27.⁵⁵ The cells also demonstrated cytotoxic features, which were characterised via gene expression analysis, with upregulation of GZMK and GZMA in the CD4+ CAR-T cells compared with normal CD4+ T cells. These observations support a growing body of evidence highlighting the pivotal role of CD4+ T cells in CAR-T cell therapy.⁵⁶ The fitness of CD4+ CAR-T cells appears to be essential for maintenance of the antitumour response. Single-cell sequencing of circulating anti-B-cell maturation antigen (BCMA) CAR-T cells in multiple myeloma patients identified significantly higher numbers of so-called CD4+ cytotoxic effector CAR-T cells in patients with early relapse. These cells expressed exhaustion markers TIM3 and T-cell immunoreceptor with Ig and ITIM domains, which may have been driven by high levels of TGF-β in the bone marrow, leading to reduced antitumour function.⁵⁷

Although direct tumour cell killing is one mechanism of disease control, the enhancement of CD8+ CAR-T cell efficacy through CD4+ CAR-T cell support has been shown to be critical for control of solid cancers. An 1 : 1 infusion ratio of CD8+ and CD4+ HER2-directed CAR-T cells in immunodeficient mice resulted in durable regression of established mammary tumours, and importantly, protection from rechallenge, whereas infusion of CD8+ CAR-T cells alone resulted in reduced tumour control.³⁴ Additionally, in a lung metastasis model of HER-2 breast cancer, adoptive transfer with IFN-γ-secreting T_H1 CD4+ T cells was associated with increased resistance to tumour rechallenge, as opposed to T_H2 cells. Conversely, infusion of either CD8+ or CD4+ CAR-T cells alone did not result in an antitumour effect.^34,35 Locoregional infusion of CAR-T cells resulted in effective control of mesothelioma cells in vivo when CD4+ CAR-T cells were exposed to tumour antigen before infusion with CD8+ CAR-T cells and resulted in enhanced CD8+ CAR-T efficacy.³⁶ Furthermore, the authors demonstrated a direct cytotoxic function of the CD4+ CAR-T cells, which was granzyme- and perforin-dependent, illustrating the dual ‘helper’ T cell and cytotoxic role of CD4+ CAR-T cells.³⁶ A preclinical study by Wang et al.³⁷ developed T central memory (T_CM)-enriched CAR-T cells targeting IL13Rα2 for glioblastoma (GBM) treatment and identified that CD4+-depleted CAR-T cells failed to control the tumour long-term. Interestingly, CD4+-only CAR-T cells formed immune synapses at the tumour interface, similar to what is typically observed in CD8+ CAR-T cells, and expressed IFN-γ TNF-α and granzyme B, consistent with a cytotoxic phenotype. These cells were able to kill GBM tumour cells as effectively as CD8+ CAR-T cells in short-term cocultures. However, in long-term assays, the CD4+ CAR-T cells showed superior tumour control over CD8+ CAR-T cells, including after tumour rechallenge, echoing the clinical findings observed in patients with CLL with persistent circulating CAR-T cells.^37,55 To better understand the mechanisms of CD4+ CAR-T killing, a recent study identified that the ability of CD4+ CD19 CAR-T cells to produce IFN-γ was a key determinant of tumour clearance, as well as the intrinsic sensitivity of tumour cells to this cytokine, consistent with previous preclinical results.⁵⁸ Importantly, the IFN-γ produced by CD4+ CAR-T cells were also able to disseminate through the TME to induce apoptosis of distant, CAR antigen-negative cells, resulting in bystander killing. This study also reported that for patients with DLBCL, a high CD4:CD8 CAR-T ratio and elevated serum IFN-γ levels 1 week after anti-CD19 CAR-T cell infusion were associated with prolonged progression-free and overall survival.⁵⁸

As previously discussed, presence of CD4+ CAR-T cells is critical for CD8+ CAR-T cell function, which was recently observed for CD8+ CAR-T cells derived from naïve T cells (termed T_STEM-like CAR-T).^38,59 It was shown that the presence of early phenotype CD4+ CAR-T cells was required for CD8+ CAR-T cell proliferation and cytokine production. Depletion of CD4+ T cells during the production and/or expansion phase resulted in inferior proliferation and production of TNF-α and IFN-γ. Bulk RNA sequencing of CD8+ T_STEM-like CAR-T cells in the presence of CD4+ T_STEM-like CAR-T cells showed greater expression of Ki67 and ICOS than separately cultured cells, suggesting a role for CD4+ T cells in the activation and proliferation of CD8+ T_STEM-like CAR-T cells. In contrast, a study by Wang et al.³⁷ compared the transcriptional profiles of in vivo-stimulated CD8+ and CD4+ CAR-T cells, noting minimal alterations in gene expression for CD8+ CAR-T cells with or without the presence of CD4+ T cells. However, gene expression of CD4+ CAR-T cells was altered depending on the presence of CD8+ CAR-T cells, with reduced expression of the early-differentiation markers LEF1, KLF7 and FOXP1 and reduced IFNG expression when CD8+ T cells were present. In addition, CD4+ CAR-T cells alone were able to durably control aggressive GBM in an orthotopic in vivo model, compared with CD8+ CAR-T cells, or both subsets in a 1 : 1 ratio, which is in contrast to other preclinical solid tumour studies.35, 36, 37, 38 It should be noted that both previously mentioned studies used CD4+ CAR-T cells intratumorally, which may yield differential effects compared with systemic administration.^60,61 This study failed to demonstrate a beneficial effect of CD4+ CAR-T cells on CD8+ CAR-T cell function, in contrast to other preclinical solid tumour studies.^35,37,38 One explanation is that the initial CD4:CD8 ratio changed dramatically in this study, in some cases with >90% of the CAR-T cell product composed of CD8+ CAR-T cells, resulting in a very small pool of ‘helper’ T cells.³⁷ Thus, different tumour models and methods of CAR-T cell administration might influence T cell subsets over time, affecting the role of CD4+ CAR-T cells.

Clinical data supports the importance of the 1 : 1 CD4+ to CD8+ CAR-T cell ratio in haematological cancers. One clinical trial infused CD19 CAR-T cells with a 1 : 1 ratio of CD4+ and CD8+ CAR-T cells, reporting a 93% complete response rate in patients with relapsed/refractory B cell malignancies.⁶⁰ This study was pivotal in demonstrating the feasibility of manufacturing CAR-T products with predefined ratios. Subsequently, the US Food and Drug Administration approved the CD19-targeting CAR product, lisocabtagene maraleucel (liso-cel), to be manufactured at a 1 : 1 ratio of CD4+ and CD8+ T cells.⁶¹ Compared with the other approved CD19 CAR-T product, axicabtagene ciloleucel, which does not require predefined CD4+:CD8+ ratios, overall survival and progression-free survival were equivocal; however, liso-cel showed superior best overall response and greater safety. Patients treated with liso-cel showed significantly lower rates of cytokine release syndrome, neurotoxicity and intensive care unit admissions.⁶²

To understand the effect of CD4+ CAR-T cells on CD8+ CAR-T cell function, one study separated conventional CD4+ and CD8+ CD19 CAR-T cells via a Transwell membrane, which prevented T cell migration between compartments but enabled diffusion of soluble factors between cells. This experiment identified a significant reduction in CD8+ CAR-T cell proliferation when physically separated from CD4+ CAR-T cells, suggesting cell-cell contact between these populations is crucial for optimal CAR-T cell function. Furthermore, it was demonstrated that blocking CD40L–CD40 and CD70–CD27 interactions between CD4+ and CD8+ CAR-T cells resulted in impaired CD8+ CAR-T cell expansion, suggesting these molecules might be involved in the beneficial effects of CD4+ CAR-T cells. Beyond these recent data, there remains a gap in our knowledge of the molecular mechanisms behind the beneficial effects of CD4+ CAR-T cells on CD8+ CAR-T cells. Finally, accounting for the specific CD4+ T cell subtype is essential for understanding CAR-T cell functionality. One study depleted CD25+FOXP3+ T_REG cells from CD19 CAR-T cells and demonstrated enhanced CD8+ T cell function through increased secretion of IL-2, IFN-γ and TNFα, and promoting perforin and granzyme secretion, which was diminished when T_REG cells were reintroduced.⁶³ The use of naïve CD4+ T cells for CAR-T cell production could theoretically avoid enriching CAR-T cells with T_REG cells, with these recent data highlighting the importance of maintaining CAR-T cell stemness in vivo, which is discussed further below.

Future directions and challenges

CD4+ CAR-T cells play an important role in tumour control and increasing evidence has demonstrated the direct tumour-killing potential of these cells. The ability to manipulate and direct CD4+ CAR-T cells toward a cytotoxic phenotype is an appealing strategy that has not been thoroughly explored so far. This strategy could include altering cell culture conditions with or without gene editing. However, substantial challenges remain in pursuing this strategy, including how to induce a potent effector phenotype while preserving in vivo persistence and avoiding T cell exhaustion. Current data points to in vitro treatment of naïve CD4+ T cells with IL-2 to develop a cytotoxic phenotype, with several studies identifying upregulation of BLIMP-1 and CRTAM after CD4+ T cell culture under these conditions.64, 65, 66 Addtionally, genes uniquely overexpressed in cytotoxic CD4+ T cells isolated from bladder cancer after treatment with checkpoint immunotherapy included APBA2, SLAMF7, GPR18 and PEG10. However, ongoing treatment of T cells with IL-2 is likely to drive them into an exhausted phenotype and may even promote differentiation into a T_REG phenotype.⁶⁷ Additionally, characterisation of this cell subset in a CAR-T context is currently limited. Further research is needed to understand how these cells develop from naïve T cells, determine the proportion of CD4+ T cells required to induce an effective tumour response, and identify transcription factors driving their differentiation and expansion. Identifying drivers of CD4+ cytotoxic T cell differentiation could allow manipulation of CAR-T products that could promote this subtype within the TME (Figure 1). The current contrasting preclinical and clinical data highlight that it is unclear whether cytotoxic CD4+ CAR-T cells alone will be sufficient for tumour control, particularly in a solid tumour context. It is therefore essential to also understand CD4+ CAR-T helper mechanisms to improve CD8+ CAR-T cell function. By obtaining a deeper understanding of these mechanisms, it may be possible to exploit this knowledge to generate a more consistent and potentially more efficacious CAR-T product for the treatment of solid cancers. It could also enable identification of targets or pathways that enhance the activity and longevity of CAR-T cells and could reduce functional variation of CD4+ T cells between patients. Such pathways could be targeted via exogenous drugs; however, more recent studies suggest CRISPR/Cas9 engineering of patient cells to optimise functionality could be feasible, including knockdown of immune checkpoint molecules or other immunosuppressive targets such as the adenosine A2A receptor or TGF-β (Figure 1).68, 69, 70, 71, 72, 73 Conversely, overexpression of key ‘fitness’ genes could be considered as well. For example, one study overexpressed BCL6 in CD4+ T cells, resulting in transdifferentiation of the cells toward a T_FH phenotype.⁷⁴ The cells were transduced with anti-mesothelin or anti-ephrin receptor 2A CAR and infused in a 1 : 1 combination with CD8+ CAR-T cells into non-obese diabetic scid gamma mice harbouring pancreatic adenocarcinoma, with the T_FH-CAR-T treatment group experiencing improved survival over controls. Edited CD4+ CAR-T cells showed greater tumour infiltration and cytotoxicity, showing the potential for targeted gene expression to enhance CAR-T efficacy against solid cancers.⁷⁴

Figure 1.

Strategies to enhance and harness CD4+ CAR-T cell therapy. There are multiple ways to boost CD4+ CAR-T cell efficacy such as improving their stem-like phenotype, increasing their cytotoxic potential, enhancing their effect on CD8+ CAR-T cells and interfering with their immunosuppressive character. ABCB1, ATP-binding cassette subfamily B member 1; CAR, chimeric antigen receptor; FOXO1, forkhead box protein O1; GZMB, granzyme B; IL, interleukin; LAG-3, lymphocyte activation gene 3; TGF-β, transforming growth factor beta.

Another promising approach to improve CAR-T function is altering T cells to improve CAR-T cell stemness. Modification of T cell culture conditions through the use of IL-7 and IL-15 has been shown to promote a naïve or memory phenotype, resulting in improved CAR-T cell proliferation, cytokine production and in vivo persistence, including for CD4+ CAR-T cells.^38,75 Preconditioning CAR-T cells with the CDK4/6 inhibitor palbociclib led to more CD8+ T memory cells and fewer T_REG cells. Antitumour efficacy was improved in palbociclib pretreated cells and a prolonged in vivo persistence of both CD4+ and CD8+ CAR-T cells was observed.⁷⁶ Another group carried out CRISPR/Cas9 screening and identified inhibition of the mitogen activated protein kinase p38, as an effective strategy to enhance CAR-T cell memory, proliferation and antitumour efficacy.⁷⁷ However, this effect was limited to CD8+ T cells, with minimal impact on CD4+ CAR-T phenotype, which was not explored further. Additionally, treatment of CAR-T cells with small molecules in culture has the disadvantage of being transient, with the effects not being maintained once infused into the patient. Recent studies reported by our group, and others, identified that overexpression of the transcription factor FOXO1 in IL-15-conditioned CAR-T cells resulted in a stem-like phenotype, increased metabolic fitness and expression of oxidative phosphorylation genes and improved cytokine production and antitumour efficacy of CD8+ CAR-T cells.^78,79 Although the approach of improving CAR-T stemness shows great promise, to date there is little information on the specific effects on CD4+ CAR-T cells and how these cells can be exploited to improve CD8+ CAR-T cell function. By improving our understanding of CD4+ CAR-T cells at the molecular level and optimising their direct and indirect antitumour activity through approaches such as gene editing to enhance stemness, we may overcome some of the key barriers to CAR-T cell efficacy in solid tumours.