CAR T cells for solid tumors – developments to watch in 2023

KEYWORDS:

• CAR T cells
• cellular Immunotherapy
• solid tumors
• clinical trials
• CAR Targets
• microbiome
• T cell migration

1. Introduction

Chimeric antigen receptor (CAR) T cell therapy has become the standard of care for multiple hematological malignancies. In contrast, treatment success in solid tumors with CAR T cells has been very modest so far. To tackle this issue, a multitude of new CAR targets, receptor designs, and treatment modalities are being developed and tested in clinical and preclinical studies. We will summarize and discuss our top picks of 2023 in the field of solid tumor CAR T cell therapy, including exciting early clinical results for anti-EpCAM, CLDN6, and GD2 CARs, the correlation between microbiome and response, as well as creative new ways to design or enhance CARs (Figure 1).

Figure 1. Schematic overview of major developments in 2023 in the field of CAR T cells for solid tumors. Created with BioRender.com.

Based on encouraging results in hematological cancers, CAR T cells are increasingly being tested for solid tumors, which account for the majority of all cancers. Initial enthusiasm to rapidly develop CAR-T cells also for solid tumors was quickly dampened by very modest clinical results and toxicities [1]. Therefore, improving the efficacy of CAR T cell therapy in solid tumors is one of the most important areas of research in immunotherapy. The main hurdles are considered to be the difficult target antigen selection to avoid toxicities and enhance efficacy, the immunosuppressive tumor microenvironment, T cell exhaustion, and the insufficient migration of CAR-T cells into tumors [2]. The last year has brought a number of important advances in preclinical studies, especially with novel CAR designs and methods on how migration into tumors could be improved. Importantly, several early phase clinical trials using solid tumor CAR-T cells report favorable safety profiles and first signs of efficacy in multiple entities. In this article, we would like to highlight some of the hottest developments in the field in 2023 from our perspective.

2. CAR T cells for solid tumors – our top picks in 2023

2.1. First clinical data on RNA vaccine-based anti-CLDN6 CAR T cell boosting

Given the exciting preclinical results [3], the community eagerly awaited the publication by Mackensen and Haanen, on the phase 1/2 BNT211–01 trial [4]. This is the first demonstration of safety for CAR T cells against the fetal-oncogene Claudin-6 (CLDN6) and the first use of RNA-based vaccines to boost CAR T cells in humans. Importantly, the CAR treatment and vaccination regimen showed a manageable toxicity profile, mostly limited to treatable cytokine release syndrome and only one case of grade 1 neurotoxicity. In this cohort, mainly composed of patients with germ cell or epithelial ovarian tumors, an overall response rate of 33% was observed with and without vaccine boosting (5 partial responses and 1 complete response among 22 treated patients). Future randomized trials will be essential to draw any conclusions on efficacy and the potential benefits of vaccine boosting.

2.2. EpCAM-targeting CAR T cell shows favorable toxicity profile in patients

In a publication spanning from CAR optimization over preclinical experiments to an early phase clinical study, Li et al. developed and demonstrated the feasibility of anti-EpCAM CAR T cells [5]. First, they compared anti-EpCAM CARs containing either 4-1BB, CD28, or Dectin-1 transmembrane and costimulatory domains in vitro and in vivo, settling for Decin-1 given its reduced tendency to induce T cell exhaustion. This CAR was efficacious in treating an intra-peritoneally injected xenograft model system and did not induce overt toxicities in a newly developed hEpCAM humanized mouse. The authors then report the application of these CARs in humans, showing a very favorable toxicity profile and not detecting any occurrence of cytokine release syndrome (CRS) in the 12 treated patients. Two partial responses were observed, the longest being observed with the highest CAR dose injected intraperitoneally. We will need to see how this result will translate into more advanced clinical phases and settings.

2.3. Next-generation CAR-T cells show impressive signs of efficacy for neuroblastoma and glioblastoma in first trials

Following the first clinical use of anti-GD2 CAR T cells for gliomas in 2022 [6], the results from two phase I studies with next-generation anti-GD2 CARs were reported in 2023 [7,8]. Both employed an anti-GD2 CAR containing CD28 and 4-1BB costimulatory domains and showed a safety profile expected for lymphodepletion and CAR T cell infusion, mainly consisting of severe hematotoxicity. In the neuroblastoma study, 9/27 patients (33%) experienced a complete response with a remarkable overall response rate of 63%. The response rate in the glioblastoma study appears to be similar (4/8); however, no complete responses were observed. Thus, both approaches show seemingly unprecedented response rates for solid tumor CAR T cell therapy and larger randomized clinical trials to investigate their anti-tumor efficacy are clearly warranted. Interestingly, the CAR products also co-expressed the ‘suicide’ switch gene iCasp9 to enable in vivo eradication in case of toxicity. Indeed, Del Bufalo et al. made use of the iCasp9 induction with rimiducid in one patient with suspected neurotoxicity. Levels of CAR-T cells decreased rapidly but re-expanded 6 weeks after the last rimiducid infusion, with the patient continuing to show a complete response. This showcases the feasibility of suicide switch genes for clinical CAR management, although this is one of the very first reports of successful use of the system for actual toxicity management. A potential concern is that although Caspase 9 is a potent apoptosis inducer, the strategy (in this case thankfully) failed to eliminate all CAR T cells, which may be a concern in more severe cases of toxicity.

2.4. Antibiotics use and microbiome composition correlate with CAR T cell therapy response

The gut microbiome’s effects on immune checkpoint blockade (ICB) are already well described and remain an exciting field of research [9]. In 2023, an international consortium published their findings on the association between microbiome, antibiotics treatment, and anti-CD19 CAR T cell responses [10]. In the five participating centers, the use of high-risk broad-spectrum antibiotics is negatively correlated with response to CAR T cells. However, the authors concluded that this association is mainly driven by a correlation of antibiotics use with poor pre-treatment disease status and systemic inflammation. The exclusion of antibiotics pre-treated patients significantly improved the predictive potential of microbiome composition for CAR T cell response, highlighting that patient selection is key to evaluate microbiome composition as a diagnostic procedure in the future. The Akkermansia bacterial taxa that correlates with response is particularly interesting as it also correlates with elevated pre-treatment T cell levels and was previously correlated with ICB response. Lifestyle or therapeutic interventions to enhance microbiome composition and thereby improve CAR T cell effectiveness will certainly receive great attention in the following years. This could be especially important for solid tumors, which in many cases are exposed to the environment (e.g. skin, colon, or lung cancer) and have even been shown to harbor their own tumor microbiome [11].

2.5. Engineered bacteria boost CAR T cells in pre-clinical models

Independent of the tumors’ own microbiome, Vincent et al. designed an exciting combination strategy using engineered bacteria with CAR T cells [12]. They express a secreted extracellular matrix-binding superfolder-GFP antigen in a probiotic E. coli strain that selectively colonizes solid tumors in mice. This antigen could then be targeted with anti-GFP CAR T cells and lead to tumor-antigen-independent treatment efficacy in several murine model systems. The bacterial product was further modified to additionally secrete CXCL16 and thereby enhance CAR T cell migration into tumors. Such exciting new treatment combinations certainly warrant further development and clinical testing.

2.6. CRISPR screen reveals how lung sequestration of T cell is regulated by a ST3GAL1-ICAM1-LFA1 axis

Solid tumors often show low inflammatory activity, which is associated with low immune cell infiltration and poor response to immunotherapies. Additionally, CAR T cells injected intravenously are most prominently retained in the lung and liver, sequestering them away from the tumor site [13]. Hong et al. performed a CRISPR Cas9 knockout screen in a syngeneic mouse model to discover genes upregulated in activated T cells that cause this lung and liver sequestration. They discovered that the protein ST3GAL1 glycosylates LFA1 component CD18 that leads to sequestration of T cells in the lung. Furthermore, they found that overexpression of βII-spectrin could normalize ICAM1 adhesion and improve CAR T cells homing away from the lung into the tumor and draining lymph nodes, thereby improving CAR effectivity. This discovery highlights how T cell homing is an essential hurdle for adoptive solid tumor therapy, which must be further exploited to tune CAR T cells.

2.7. Novel AND-gate CARs using T cell signaling protein fusions demonstrate improved specificity

2023 saw a number of creative CAR designs being developed. Tousley et al. screened replacements for the classical CD3ζ stimulatory domain using various other T cell activation cascade proteins [14]. Thereby, they developed a ZAP70 kinase domain and interdomain B (Zap70-KIDB) CAR that demonstrated enhanced antitumor activity in preclinical models. They then ingeniously designed a novel AND-gate CAR approach combining LAT and SLP76-based receptors. The leakiness of this design could further be abrogated by rational deletion of the GADS adaptor sites and cysteine residues in the CD28 transmembrane domain. The final design showed substantially improved specificity and abrogated toxicity while maintaining anti-tumor effectivity in vivo. Such AND-gate CARs have tremendous potential for clinical development as they could enable targeting of solid tumor antigens that are only specific in combination. This AND-gate CAR relies solely on two human protein domains independent of transcriptional activation, unlike the previously published SynNotch system [15].

2.8. TET2 disruption causes somatic mutation and genomic instability in CAR T cells

Lastly, an important preclinical study demonstrated how disruption of the TET2 gene in CAR T cells – previously shown to enhance T cell memory formation – induced antigen-independent T cell expansion [16]. This BATF3 and Myc pathway-dependent process led to the acquisition of secondary somatic mutations posing a risk for CAR T cell leukemia development. This highlights how the maintenance of genomic stability is an essential factor to control in future genetic CAR T cell enhancement strategies.

3. Expert opinion

We believe that the expected and manageable safety profiles of anti-EpCAM, GD2, and CLDN6 CAR T cells in these early clinical trials are very encouraging (Table 1). These trials do not allow any conclusion on efficacy yet, but seeing complete and partial responses in these difficult-to-treat patient populations is certainly exciting. While the field has progressed significantly with encouraging clinical trials using these well-known antigens, there is also a great interest in discovering new solid tumor antigens. The ever-growing amount of single-cell RNA sequencing data of tumor cells as well as new engineering approaches such as AND-gate CARs could enable the design of safer and more widely applicable CARs [17,18].

Table 1. Summary of highlighted clinical studies and cohorts.

Study	Trial and cohort	CAR and treatment	Main findings
[4]	NCT04503278 Phase I/II trial: Germ-cell tumors, epithelial ovarian cancer, endometrial carcinoma, rare solid malignancies	Anti-hCLDN6-CAR (CD8α hinge and transmembrane region, 4-1BB and CD3ζ signaling domains) Lipoplexed RNA vaccination encoding hCLDN6.	Manageable toxicities, tolerated vaccine-boosting, 33% overall response rate (ORR)
[5]	NCT02915445 Phase I trial: nasopharyngeal, parotid gland, rectal, signet-ring and gastric cancer	Anti-hEpCAM-CAR (CD8α hinge, Dectin-1 transmembrane and signaling domain, CD3ζ signaling domains)	New Dectin-1-based CAR design. Manageable toxicity and no treatment related cytokine release syndrome (CRS) observed. Partial response in 2/12 patients.
[7]	NCT03373097 Phase I/II trial: pretreated children with neuroblastoma	Anti-hGD2-CAR (CD28, 4-1BB and CD3ζ signaling domains) also encoding iCasp9 suicide switch gene.	Manageable toxicity profile. 63% ORR, 9 complete responses (CR).
[8]	NCT03170141 Phase I trial: progressive glioblastoma (GBM)	Anti-hGD2-CAR (CD28 transmembrane and signaling domain, 4-1BB TRAF and CD3ζ signaling domains) also encoding iCasp9 suicide switch gene.	Manageable toxicity profile. 4/8 patients showed partial response, 1/8 stable disease.
[10]	Two cohorts of relapsing or refractory B cell lymphoma patients in Germany and the U.S.A. (n = 172 total).	Anti-hCD19 CARs: axicabtagene ciloleucel (CD28 hinge/TM/signaling and CD3ζ signaling), tisagenlecleucel (CD8α hinge and TM, 4-1BB and CD3ζ signaling), lisocabtagene maraleucel (IgG4 hinge, CD28 TM, 4-1BB and CD3ζ signaling)	High risk broad-spectrum antibiotics use correlated with adverse outcomes, but also increased pretreatment tumor burden and systemic inflammation. Several bacterial species correlate with CAR T cell responsiveness.

The correlation of the microbiome and antibiotics use with therapeutic response is striking, yet it remains unclear how this knowledge will be applied in the clinics. Avoiding antibiotics treatment to enhance CAR T cell therapy success is a risky strategy as infections are one of the main toxicities [19]. Therefore, better diagnostic procedures coupled to a more selective use of antibiotics, or the application of probiotics and metabolites should be investigated as alternatives.

The discovery of genes and pathways responsible for lung sequestration of CAR T cells could be a pivotal development to enhance therapy success. With billions of T cells infused into the patient, likely only a small percentage ever reaches the tumor and can fulfill their purpose. The findings by Hong et al. agree with previous imaging studies by Skovgard et al., who also showed that CAR T cells given intravenously mainly home to the lungs, liver, and bone marrow in mice and that regional administration could improve efficacy [20]. This is in line with recent preclinical experiments where delivery of CAR T cells into the surgical cavity inside a fibrin gel improved efficacy [21]. While local administration is a promising strategy currently tested clinically mainly for brain tumors, intravenous application will likely continue to be the method of choice for most metastatic cancers. We believe that the migration of CAR T cells applied intravenously to the tumor could be further boosted by coupling the abrogation of sequestration with the overexpression of selected chemokine receptors [22,23]. Moreover, new methods to recruit T cells into tumors are highly warranted, exemplified by chemokine-secreting tumor-homing bacteria.

Understanding how TET2 gene disruption leads to genetic instability and antigen-independent T cell expansion is especially important given the recent FDA investigation into T cell malignancies after CAR T cell application [24]. It remains key to avoid such potentially malignant transformations during CAR construct transduction, CRISPR-mediated optimization, or simply during ex vivo expansion to enable widespread use and long-term success of CAR T cell therapies.

Declaration of interests

S Kobold has received honoraria from TCR2 Inc., Miltenyi, Galapagos, Novartis, BMS, and GSK. S Kobold is an inventor of several patents in the field of immuno-oncology. S Kobold received license fees from TCR2 Inc and Carina Biotech. S Kobold received research support from TCR2 Inc., Tabby Therapeutics, Catalym GmBH, Plectonic GmBH and Arcus Bioscience for work unrelated to the manuscript. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

A reviewer on this manuscript has disclosed that they are a member of the senior leadership team at Leucid Bio, a CAR T-cell company focused on solid tumor immunotherapy. Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was funded by the Swiss National Science Foundation (SNSF Postdoc. Mobility grant to MP Trefny) and the Monika-Kutzner Stiftung (to MP Trefny). This work was supported by the Bavarian Cancer Research Center (BZKF) (TANGO to S Kobold), the Deutsche Forschungsgemeinschaft (DFG, grant number: KO5055-2-1 and KO5055/3-1 to S Kobold), the international doctoral program ‘i-Target: immunotargeting of cancer’ (funded by the Elite Network of Bavaria; to S Kobold), the Melanoma Research Alliance (grant number 409510 to S Kobold), Marie Sklodowska-Curie Training Network for Optimizing Adoptive T Cell Therapy of Cancer (funded by the Horizon 2020 programme of the European Union; grant 955575 to S Kobold), Else Krner-Fresenius-Stiftung (IOLIN to S Kobold), German Cancer Aid (AvantCAR.de to S Kobold), the Wilhelm-Sander-Stiftung (to S Kobold), Ernst Jung Stiftung (to S Kobold), Institutional Strategy LMUexcellent of LMU Munich (within the framework of the German Excellence Initiative; to S Kobold), the Go-Bio-Initiative (to S Kobold), the m4-Award of the Bavarian Ministry for Economic Affairs (to S Kobold), Bundesministerium für Bildung und Forschung (to S Kobold), European Research Council (Starting Grant 756017, PoC Grant 101100460 and CoG 101124203 to S Kobold), by the SFB-TRR 338/1 2021–452881907 (to S Kobold), Fritz-Bender Foundation (to S Kobold), Deutsche José Carreras Leukämie Stiftung (to S Kobold), Hector Foundation (to S Kobold), Bavarian Research Foundation (BAYCELLATOR to S Kobold), the Bruno and Helene Jöster Foundation (360° CAR to S Kobold).