Targeting the mevalonate or Wnt pathways to overcome CAR T-cell resistance in TP53-mutant AML cells

TP53-mutant acute myeloid leukemia (AML) and myelodysplastic neoplasms (MDS) are characterized by chemotherapy resistance and represent an unmet clinical need. Chimeric antigen receptor (CAR) T-cells might be a promising therapeutic option for TP53-mutant AML/MDS. However, the impact of TP53 deficiency in AML cells on the efficacy of CAR T-cells is unknown. We here show that CAR T-cells engaging TP53-deficient leukemia cells exhibit a prolonged interaction time, upregulate exhaustion markers, and are inefficient to control AML cell outgrowth in vitro and in vivo compared to TP53 wild-type cells. Transcriptional profiling revealed that the mevalonate pathway is upregulated in TP53-deficient AML cells under CAR T-cell attack, while CAR T-cells engaging TP53-deficient AML cells downregulate the Wnt pathway. In vitro rational targeting of either of these pathways rescues AML cell sensitivity to CAR T-cell-mediated killing. We thus demonstrate that TP53 deficiency confers resistance to CAR T-cell therapy and identify the mevalonate pathway as a therapeutic vulnerability of TP53-deficient AML cells engaged by CAR T-cells, and the Wnt pathway as a promising CAR T-cell therapy-enhancing approach for TP53-deficient AML/MDS.

p53-encoded by its gene TP53 on chromosome 17-is a transcription factor acting as a tumor suppressor by mediating the response to a variety of cellular stress stimuli such as DNA damage and other pro-apoptotic signals (Levine and Oren, 2009).Current treatment options for AML/MDS, including conventional induction chemotherapy and HMAs +/− Ven, largely exert their efficacy via the induction of p53-mediated apoptosis (Schimmer et al, 2022;Thijssen et al, 2021).Therefore, novel therapeutic strategies-in particular immunotherapeutic approaches that possibly circumvent the need for p53-mediated apoptosis-appear particularly promising for this hard-to-treat patient population (Sallman et al, 2020).Chimeric antigen receptor (CAR) T-cell therapy is a novel and highly efficacious immunotherapeutic modality, that generated hitherto unseen remission rates in relapsed/refractory B-cell malignancies and multiple myeloma, and that could be an attractive option for TP53-mutant AML/MDS as well (Berdeja et al, 2021;Davila et al, 2014;Kochenderfer et al, 2015;Lee et al, 2015a;Maude et al, 2014;Munshi et al, 2021;Neelapu et al, 2017;Schuster et al, 2019;Turtle et al, 2016a).A fundamental prerequisite for the development of an effective CAR T-cell therapy in AML/MDS would be determining the most suitable AML/MDS-associated cell surface target antigen (Borot et al, 2019;Haubner et al, 2019;Perna et al, 2017).Current CAR T-cell therapeutic efforts in AML/MDS are, therefore, primarily devoted to testing such target antigens including the surface antigens CD33, CD117, CD123, and CD371 (Arai et al, 2018;Laborda et al, 2017;Mardiana and Gill, 2020;Myburgh et al, 2020;Tashiro et al, 2017;Wang et al, 2015).Equally important for therapeutic efficacy but much less well-studied are intrinsic and acquired resistance mechanism to CAR T-cellmediated killing.While target antigen loss is a well-known acquired resistance mechanism observed in B-cell lymphoma and B-lymphoblastic leukemia (Fraietta et al, 2018;Shah and Fry, 2019;Sotillo et al, 2015), cancer cell-intrinsic determinants of the efficacy of CAR T-cell therapies are less well understood, particularly in AML (Singh et al, 2020).Loss of TP53 in target cells has been shown to confer resistance to cytotoxic T-cells and NK-cells via transcriptional and non-transcriptional mechanisms but how this might affect CAR T-cell-mediated killing is largely unknown (Ben Safta et al, 2015;Chollat-Namy et al, 2019;Thiery et al, 2005;Thiery et al, 2015).Furthermore, TP53 mutations in AML/MDS induce an immunosuppressive phenotype in the immune microenvironment with a potential for negatively impacting CAR T-cell efficacy (Sallman et al, 2020).Recently, in the clinical setting of diffuse large B-cell lymphoma, TP53 deficiency has been shown to be a negative prognostic marker of CD19-directed CAR T-cell therapy (Shouval et al, 2022), whereas another recent study could not replicate this finding (Jain et al, 2022).It is, therefore, currently unclear, whether the TP53 mutational status might impact the efficacy of CAR T-cells in hematological malignancies.
In light of the above-mentioned, we addressed the biologically and clinically highly relevant question as to whether, and, if so, how TP53 deficiency in AML cells might confer resistance to CAR T-cell therapy.To this end, we took advantage of a recently developed, CRISPR/Cas9-engineered MOLM13-TP53 isogenic human AML cell line model (Boettcher et al, 2019) as well as a newly generated isogenic AML cell line model (MV4-11), and tested the efficacy of CAR T-cells targeting common (CD33, CD117, CD123) AML cell surface antigens using flow cytometry-based in vitro killing assays, live-cell imaging, transcriptional profiling (mRNA-seq), CRISPR/ Cas9-based gene editing in CAR T-cells, and a xenogeneic mouse model of AML.

TP53-mutant AML cells are resistant to CAR T-cell-mediated killing in vitro
In order to test whether the TP53 mutational status of AML blasts might impact the efficacy of CAR T-cell therapies, we took advantage of a recently developed isogenic human MOLM13 AML cell line model that was CRISPR/Cas9-engineered to express an allelic series of the six most common TP53 missense mutations, null as well as wild-type alleles (Boettcher et al, 2019).Secondgeneration CAR T-cells with the 4-1BB costimulatory domain and the CD3ζ activating domain were generated by lentiviral transduction of T-cells from at least three different healthy donors (Fig. EV1A).The single-chain variable fragment (scFv) portion of the CAR was directed against three different target antigens present on normal myeloid precursors and most AML: CD33, CD123, and CD117 (Figs. 1A and EV1A).CAR T-cells were expanded and purified by MACS to yield high purity (>95% by flow cytometry) products (Fig. EV1B).MOLM13-TP53 isogenic cells showed indistinguishable surface antigen expression and growth kinetics in the unperturbed state (Figs. 1B and EV1C).
Altogether, we revealed that TP53 deficiency in AML cells confers resistance against CAR T-cell-mediated killing in vitro.

CAR T-cells engaging MOLM13-TP53 −/− AML cells proliferate less, exhibit sustained activation marker expression, an exhausted immunophenotype, and enhanced trogocytosis
To elucidate the underlying mechanisms for decreased killing efficacy of CAR T-cells attacking MOLM13-TP53 −/− cells, we first assessed CAR T-cell proliferation during 6 days of co-incubation.While CAR T-cells expanded 1-1.5 log-fold in the presence of MOLM13-TP53 +/+ cells, CAR T-cells co-incubated with MOLM13-TP53 −/− or MOLM13-TP53 missense/− cells showed a significantly lower expansion (Figs. 2A and EV1H).Next, we measured CAR T-cell activation as determined by the surface expression of the early T-cell activation marker CD25.The fraction of CAR T-cells with sustained CD25 expression at day 6 of co-incubation was Specific killing  significantly higher in CAR T-cells engaging MOLM13-TP53 −/− AML cells than those engaging MOLM13-TP53 +/+ AML cells (Fig. 2B,C; Appendix Fig. S1).
As sustained stimulatory signaling via the CAR is associated with CAR T-cell exhaustion, we next analyzed the expression of canonical T-cell markers that are associated with insufficient clearance of antigen-expressing cells, i.e., PD-1 (Programmed cell death protein 1), LAG3 (Lymphocyte activating 3), and TIM3 (Tcell immunoglobulin and mucin-domain containing-3).We observed TP53-dependent differences in exhaustion profiles: CAR T-cells co-incubated with MOLM13-TP53 −/− AML cells exhibited a more exhausted T-cell phenotype than CAR T-cells co-incubated with MOLM13-TP53 +/+ AML cells as demonstrated by a higher percentage of CAR T-cells being double-or triple-positive for the above exhaustion markers (Fig. 2D,E; Appendix Fig. S2).Untransduced T-cells did not show upregulated surface expression of exhaustion markers upon co-incubation with either AML cell line (Fig. 2D,E; Appendix Fig. S2).
Trogocytosis is the active process, described in CAR T-cells, of acquiring target antigens while in contact with their target cells, followed by displaying them incorporated into their own membrane surface (Hamieh et al, 2019).We found a significantly higher percentage of acquired CD33 expression on CAR T-cells co-incubated with MOLM13-TP53 −/− AML cells than on CAR T-cells co-incubated with MOLM13-TP53 +/+ AML cells or the untransduced T-cells in control conditions, hinting at a longer and/or more intimate CAR Tcell:AML cell interaction in the context of TP53 deficiency in AML cells (Fig. 2F,G; Appendix Fig. S3).Although trogocytosis has been recognized as a means of acquired resistance to CAR T-cell therapies (Hamieh et al, 2019) via reversible antigen loss on tumor cells, we did not observe a decrease in target antigen density on MOLM13-TP53 -/- AML cells (Fig. EV1I; Appendix Fig. S4A).We did neither observe evidence of antigen-negative escape by outgrowth of CD33 − MOLM13-TP53 isogenic AML cells, nor differences in surface expression of the canonical immune checkpoint marker PD-L1 (Programmed cell death 1 ligand 1) at steady state or upon coincubation with CAR T-cells (Fig. EV1J; Appendix Fig. S4A,B).
In summary, the inability of CAR T-cells to eradicate MOLM13-TP53 −/− AML cells in vitro likely involves sustained activating CAR signaling, exacerbated CAR T-cell exhaustion, enhanced trogocytosis, and decreased proliferation.

Live-cell imaging shows a longer duration of interaction between CAR T-cells and MOLM13-TP53 −/− AML cells
Increased trogocytosis in CAR T-cells attacking MOLM13-TP53 −/− cells suggests a longer duration of engagement between CAR T-cells and TP53-deficient AML cells.To further investigate CAR T-cell:AML cell interaction dynamics, we employed fluorescence live-cell microscopy.CAR T-cells were co-incubated with isogenic MOLM13-TP53 AML cells with either TP53 wild-type or TP53 null status at an E:T ratio of 1:1 on imaging slides containing microgrids to avoid cell clustering and allow for tracking of individual killing events (Fig. 2H; Movies EV1 and EV2).The time from engagement between CAR T-cells and AML cells, and subsequent AML cell lysis, as judged by propidium iodide (PI) influx, was significantly longer, i.e., increased from roughly 10 h to ~16 h, for CAR T-cells co-incubated with MOLM13-TP53 −/− as compared to MOLM13-TP53 +/+ AML cells (Fig. 2H,I; Movies EV1 and EV2).Under these specific experimental conditions of a high E:T ratio with cells confined to microgrids, the TP53 genotype-specific difference between MOLM13-TP53 −/− and MOLM13-TP53 +/+ AML cells killed by CAR T-cells was less pronounced than in the regular co-culture assays.However, there were still significantly fewer killing events upon CAR T-cell co-incubation with MOLM13-TP53 −/− than with MOLM13-TP53 +/+ AML cells (Fig. 2J).
Taken together, these data suggest that the delayed CAR T-cellmediated killing of MOLM13-TP53 −/− AML cells, and therefore prolonged antigen exposure to CAR T-cells likely contributes to the observed sustained activation, increased exhaustion, and consequently, lower therapeutic efficacy against TP53-mutant AML cell lines in vitro.

MOLM13-TP53 −/− AML cells are resistant to CAR T-cells in a therapeutic xenograft in vivo mouse model
Having established that loss of TP53 in AML cells promotes increased resistance to CAR T-cell therapies in vitro, we next employed a therapeutic AML mouse model to test the in vivo relevance of our findings.Isogenic MOLM13-TP53 AML cells with either TP53 −/− or TP53 +/+ status were transduced with a luciferase lentiviral construct (Myburgh et al, 2020), allowing us to track cells in vivo by bioluminescence.One hundred thousand   Altogether, these data confirm our above-described results from in vitro co-culture experiments corroborating our finding that the TP53 mutational status is a critical tumor-intrinsic genetic determinant of CAR T-cell therapeutic efficacy in AML in vitro and in vivo.
These findings suggest that during engagement between CAR T-cells and AML cells both CAR T-cell-intrinsic as well as AML cell-intrinsic pathways are differentially activated or repressed, which in turn, may subsequently mediate CAR T-cell efficacy or lack thereof, respectively.Notably, the precise nature of these responses is dependent on the TP53 mutational status of the AML cells.This might provide an opportunity for a therapeutic intervention to enhance the efficacy of CAR T-cells against TP53mutant AML.

Targeting the mevalonate or the Wnt pathway fully rescues CAR T-cell killing of TP53-deficient AML cells
In order to test whether the mevalonate and/or Wnt pathways, prominent members of which we found to be differentially expressed in a TP53-dependent manner, are involved in determining the efficacy of CAR T-cell-mediated killing of TP53-deficient AML cells, we sought to target these pathways pharmacologically in co-culture killing assays in vitro (Fig. 5A).
In terms of the mevalonate pathway in TP53-mutant AML cells, we investigated the effect of the competitive HMG-CoA reductase (HMGCR; 3-hydroxy-3-methylglutaryl coenzyme A reductase) inhibitor simvastatin, which blocks the rate-limiting step of the mevalonate pathway.Furthermore, we supplemented the simvastatin-treated cultures with the HMGCR metabolite mevalonate to reverse the effect of simvastatin.Through a series of experiments measuring low-density lipoprotein receptor (LDLR) surface expression by FACS and the expression of the sterolregulatory element (SRE) in an established luciferase reporter assay, we confirmed the expected on-target activity of simvastatin (Fig. EV4A-D; Appendix Fig. S5C).For all future experiments, we used 1 μM and 15 μM simvastatin for experiments with MOLM13 and MV4-11 cells, respectively, as determined in dose-titration experiments (Fig. EV4A,B).
We first performed a series of in vitro co-incubation experiments: Adding simvastatin at the previously established dose to our killing assays completely rescued the sensitivity of MOLM13-TP53 −/− or MV4-11-TP53 −/− AML cells co-incubated with CAR T-cells over several log2 fold E:T range with CD33-as well as CD123-targeting CAR T-cells (Figs. 5B,C and EV4E,F).Of note, the addition of   In terms of the Wnt pathway in CAR T-cells attacking TP53mutant AML cells, we sought to pharmacologically increase the activation of the canonical Wnt pathway in CAR T-cells.To this end, we added 0.5 μM of the GSK-3 inhibitor and thus activator of the canonical Wnt pathway, BIO-acetoxime to our killing assays (Fig. 5A).We confirmed on-target activity of BIO-acetoxime by a luciferase reporter system in T-cells (Fig. EV5C).
BIO-acetoxime treatment of the co-culture assays resulted in a complete rescue of the deficiency of CAR T-cells in killing MOLM13-TP53 −/− AML cells over several log2-fold range of E:T ratios as well as a reversal of most of the CAR T-cell dysfunction features: We observed decreased exhaustion marker expression in CAR T-cells attacking MOLM-TP53 −/− AML cells, normalized interaction times, and increased T-cell proliferation compared to control conditions (Figs. 5B-E, EV4E-G and EV5A,B).Additional killing assays using anti-CD33 and anti-CD123-targeting CAR T-cells together with isogenic MOLM13-TP53 and MV4-11-TP53 AML cells confirmed the complete reversal of the resistance of TP53-mutant AML cells against CAR T-cell killing (Fig. EV4E).

Pretreatment with simvastatin but not with BIO-acetoxime rescues CAR T-cell killing of TP53-deficient AML cells
Our transcriptional profiling results suggest that the effect of the mevalonate and Wnt pathways in mediating resistance of TP53-mutant AML cells to CAR T-cells are restricted to TP53mutant AML and CAR T-cells attacking TP53-mutant AML cells, respectively.In order to test this functionally, we performed a second set of co-incubation experiments, this time with simvastatin-or BIO-acetoxime-pretreated AML or CAR T-cells (Fig. 6A).
As a first step, we analyzed the kinetics of the on-target activity of both drugs at the established concentrations for several days after washout: we observed sustained upregulation of LDLR over the course of 6 days after removal of simvastatin in both analyzed AML cell lines and genotypes (Figs. 6B and EV4D).By contrast, washout of BIO-acetoxime led to a rapid normalization of the TCF-reporter signal with indistinguishable levels by day 4 after drug removal (Figs.6C and EV5C).
Consistent with the prolonged simvastatin effect as shown in the washout experiments, simvastatin pretreatment of AML cells before co-incubation with CAR T-cells led to a complete rescue of the specific killing of both TP53-mutant MOLM13 and MV4-11 AML cells (Fig. 6D,E).Simvastatin pretreatment of AML cell lines with wild-type TP53 +/+ or CAR T-cells did not show any effect on CAR T-cell efficacy, indicating that the effect is specific for TP53deficient AML cells (Figs. 6D,E and EV5D).These results were consistently observed with CD33-and CD123-directed CAR T-cells (Figs. 6D,E and EV5D).In line with our prior washout experiments and the rescued killing rate, we observed sustained LDLR upregulation in simvastatin pretreated AML cells, equivalent to AML cells incubated in the continuous presence of simvastatin, upon 6 days of co-incubation in our killing assays (Fig. EV5E).This observation held true for MOLM13-TP53 and MV4-11-TP53 cells and with CD33-as well as CD123-targeting CAR T-cells (Fig. EV5E).
However, BIO-acetoxime pretreatment of both AML cells and CAR T-cells failed to rescue the specific killing rate of TP53-mutant MOLM13 or MV4-11 AML cells by CAR T-cells (Fig. 6A,D,E and  EV5D)-consistent with above washout experiments, which showed rapid normalization of Wnt signaling upon drug removal (Figs.6C and EV5C).
Collectively, we here show that rational drug targeting of specific resistance mechanisms-i.e., the mevalonate pathway in TP53mutant AML cells or the Wnt pathway in CAR T-cells engaging them-can completely rescue TP53-dependent resistance of AML cells to CAR T-cell killing in vitro.

CRISPR/Cas9 knockout of Regnase-1 results in Wnt pathway upregulation and rescue of TP53 deficiency-associated CAR T-cell resistance
Despite these encouraging results suggesting that pharmacological co-interventions with simvastatin and/or BIO-acetoxime may improve the efficacy of CAR T-cells against TP53-mutant AML, toxicity associated with such high doses of simvastatin as well as the short-lived effect of BIO-acetoxime would raise concerns over their clinical applicability.We therefore tested a genetic strategy to achieve sustained Wnt pathway activation in CAR T-cells.
To this end, we targeted Regnase-1 (also termed MCPIP1 and encoded by ZC3H12A), a rapid-response ribonuclease primarily involved in immune regulation, with TCF7 as one of its primary targets (Matsushita et al, 2009;Uehata et al, 2013;Wei et al, 2019).Genetic manipulation of Regnase-1 has been investigated as a means of increasing TCF7 signaling for improved antitumor    effector function (Mai et al, 2023;Wei et al, 2019;Zheng et al, 2021).To test whether Regnase-1 depletion in CAR T-cells could overcome the resistance of TP53-mutant AML cells against CAR Tcells, we knocked out Regnase-1 in CD33-directed as well as CD123-directed CAR T-cells using CRISPR/Cas9 genome editing (Fig. 6F).Using a previously published Regnase-1-targeting small guide RNA (sgRNA) (Mai et al, 2023), we generated CAR T-cells with high levels of Regnase-1 gene editing (>95% indels, Fig. EV5F).As expected, gene-edited Reg −/− CAR T-cells showed increased TCF7 mRNA expression in comparison to untreated and CAR T-cells electroporated with non-targeting sgRNA (mock) (Fig. EV5G).Reg −/− CAR T-cells, co-incubated with isogenic MOLM13-TP53 as well as MV4-11-TP53 AML cells in our established killing assays, dramatically outperformed the untreated and mock CAR T-cell controls both against TP53 +/+ as well as TP53 −/− AML cells over several E:T ratios and in both MOLM13-and MV4-11-TP53 cells (Fig. 6G,H).The increased killing efficacy of Reg −/− CAR T-cells was even more pronounced in TP53 −/− cells, leading to comparable specific killing rates, and effectively overcoming the resistance of TP53-mutant AML cells against CAR T-cells at the investigated E:T range (Fig. 6G,H).In summary, genetic targeting of the negative regulator of the Wnt pathway Regnase-1 in CAR T-cells results in highly improved killing of TP53-deficient AML cells.

Discussion
In this study, we addressed the emerging topic of genetic determinants of resistance to CAR T-cell therapies in hematological malignancies.We focused on TP53-mutant AML/MDS because current therapeutic approaches, i.e., intensive induction chemotherapy or HMAs+/− Ven have failed to improve patient outcomes (Pollyea et al, 2022), and thus, immunotherapeutic strategies are currently being tested in this clinically challenging patient cohort (Sallman et al, 2022).We took advantage of isogenic human AML cell lines harboring TP53 null, missense, or wild-type alleles to exclude other potential genetic confounders.Such experimental approaches have successfully been employed to elucidate important aspects of the pathobiology of TP53-mutant AML/MDS (Boettcher et al, 2019;Sallman et al, 2020;Schimmer et al, 2022;Sellar et al, 2022;Thijssen et al, 2021).Our findings support a model, in which the intrinsic apoptotic defect in TP53deficient AML/MDS cells results in a longer duration of the cellular interaction between CAR T-cells and TP53-deficient AML/MDS cells as shown by indirect as well as direct evidence of enhanced trogocytosis and live-cell imaging, respectively.Longer temporal interaction between CAR T-cells and TP53-deficient AML/MDS leads to enhanced CAR T-cell exhaustion, reduced CAR T-cell proliferation, and eventually an overall decrease in AML cell killing and leukemia cell outgrowth (Fig. 7).
Although this is, to our knowledge, the first in-depth exploration of the impact of TP53 deficiency in cancer cells on CAR T-cellmediated killing, our data are supported by previous studies in physiologic cytotoxic T-cells (CTLs) and natural killer (NK)-cells.CTL-and NK-cell-mediated target cell apoptosis via perforin, granzyme B, or Fas-ligand was shown to be dependent on intact p53 in target cells (Ben Safta et al, 2015;Chollat-Namy et al, 2019;Thiery et al, 2005;Thiery et al, 2015).
Our results are also consistent with a recently published CRISPR-based genetic screen for antigen-independent CAR T-cell resistance mechanisms in acute B-lymphoblastic leukemia by Singh and colleagues, where the most-enriched sgRNAs belonged to the apoptosis pathway, with "p53 signaling" being in the top six mostenriched pathways (Singh et al, 2020).
The clinical evidence supporting a negative predictive role for TP53 mutations in lymphoma in the context of anti-CD19 CAR T-cell therapies is, however, controversial.Shouval et al reported that TP53 deficiency in large B-cell lymphoma (LBCL) led to inferior response and overall survival rates in patients treated with each of the clinically approved anti-CD19 CAR T-cell products (Shouval et al, 2022).By contrast, Jain et al could not confirm that TP53 mutations are negative predictive biomarkers in LBCL patients treated with anti-CD19 CAR T-cells (Jain et al, 2022).This discrepancy might be explained by various confounding factors including co-mutations beyond TP53.
One limitation of our study may be the exclusive use of isogenic cell lines rather than primary cells.However, pre-clinical studies using well-defined isogenic cell lines model solely differing in the TP53 mutational status appear ideally suited to rigorously addressing the question of the impact of individual gene mutations on therapy resistance.Of note, we observed highly consistent effects in two different isogenic AML cell line models.Our results, therefore, may help improve risk stratification and inform the design of future clinical trials of cellular therapies in TP53-mutant AML/MDS.Nevertheless, further studies are needed to confirm the clinical translatability of our findings-particularly in other cancer types.
Importantly, our data revealed both cancer cell-intrinsic as well as CAR T-cell-intrinsic pathways that could be exploited to overcome therapy resistance of TP53-mutant AML/MDS or to enhance therapeutic potency of CAR T-cells against TP53-mutant AML/MDS, respectively.Through transcriptional profiling of TP53-mutant AML cells under CAR T-cell attack we uncovered the mevalonate pathway to be most upregulated in comparison to  PreTx -

CART preTx Simvastatin
CART vs TP53 +/+ CART vs TP53 -/- AML preTx BIO-acetoxime CART vs TP53 TP53 wild-type AML cells.Accordingly, pharmacological inhibition of HMG-CoA reductase-the rate-limiting enzyme of the mevalonate pathway-using simvastatin could fully rescue CAR T-cell-mediated killing of TP53-mutant AML cells.These results corroborate and extend the existent literature on the impact of mevalonate pathway in cancer (Mullen et al, 2016).p53 is a known negative regulator of the mevalonate pathway with increased cholesterol levels shown to promote tumorigenesis in a pre-clinical cancer models of breast carcinoma and hepatocellular carcinoma (HCC) (Freed-Pastor et al, 2012;Moon et al, 2019).On the molecular level, decreasing the cholesterol content of a target cell was shown to increase the stiffness of its cell membrane, which in turn increased the perforin-dependent cell death and killing by cytotoxic T-cells (Basu et al, 2016;Lei et al, 2021).Moreover, several cohort studies and metaanalyses, comprising more than 600,000 cancer patients, have consistently found a small but statistically significant prolongation of overall survival and/or disease-specific survival associated with statin treatment (Bansal et al, 2012;Cardwell et al, 2015;Gray et al, 2016;Huang et al, 2017;Manthravadi et al, 2016;Nayan et al, 2017;Nielsen et al, 2012;Peng et al, 2015).One caveat is, however, that the typical plasma concentrations achieved with oral simvastatin therapy are about 2-3-log-fold lower than the one we used in the here reported in vitro killing assays (Bjorkhem-Bergman et al, 2011;Keskitalo et al, 2009).Therefore, prohibitive dose-limiting toxicities will prevent immediate clinical translation of our pre-clinical findings.Whether more specific inhibition of the mevalonate pathway in TP53-deficient AML/MDS in the context of CAR T-cell therapies might provide a therapeutic window remains to be determined and will be the focus of future studies.
With regards to CAR T-cell-intrinsic resistance pathways, transcriptional profiling of CAR T-cells attacking TP53-deficient AML cells revealed a significant downregulation of the transcription factors TCF7 and EOMES, both of which are signaling downstream of the canonical Wnt pathway, and are essential in T-cell development, differentiation, and survival (Staal et al, 2004;Zhou et al, 2010).These data strongly suggest that insufficient activation of the Wnt pathway in CAR T-cells may contribute to their reduced activity against TP53-deficient AML/MDS.Accordingly, our data demonstrate that sustained activation of the Wnt pathway by pharmacological intervention or by overexpression of TCF7 via knocking out its negative regulator Regnase-1, effectively overcomes TP53 deficiency-associated CAR T-cell resistance, and further enhances CAR T-cell efficacy.
Last, our study underscores the importance of achieving favorable E:T ratios in a therapeutic setting, as has been observed in clinical trials (Ali et al, 2016;Enblad et al, 2018;Milone and Bhoj, 2018;Raje et al, 2019;Turtle et al, 2016a;Turtle et al, 2016b).By increasing the in vitro E:T ratio we could eventually eradicate TP53-deficient AML cells and observed similar exhaustion and differentiation profiles as in CAR T-cells engaging isogenic, TP53proficient AML cells at lower E:T ratios.Our data suggest that treatment of patients with TP53-deficient hematological neoplasms will require a relatively higher dose of infused CAR T-cells at a given tumor load (e.g., minimal residual disease) to achieve complete remissions-an approach, which has been pursued with other malignancies stratified as a high-risk disease (Brudno et al, 2018).However, this may come at the cost of increased toxicity (Dasyam et al, 2020;Milone and Bhoj, 2018).Therefore, the combination of CAR T-cell therapies with pharmacological cointerventions or the use of genetically enhanced CAR T-cell products-as exemplified in this study-may be a preferable strategy towards more efficacious and tolerable cellular therapies for patients with TP53-mutant myeloid neoplasms.

Production of lentiviral vectors
Replication-defective, third-generation lentiviral vectors were produced as previously described (Myburgh et al, 2020).In brief, HEK293T were transfected with the transfer plasmid containing the DNA sequence of interest as well as with psPAX2 packaging and pCAG-VSVG envelope plasmids (both kindly provided by Dr. Patrick Salmon, University of Geneva, Switzerland).JetPRIME ® transfection reagent (Polyplus transfection, Illkirch, France) was used to enhance transfection efficacy.Viral particles were harvested 2 days later and concentrated with Peg-it™ (System Biosciences, Palo Alto, CA, USA).Concentrated viral particles were stored at −80 °C until further use.

Generation of MOLM13-TP53 cells expressing GFP and luciferase
Lentiviral transduction of MOLM13-TP53 with Luciferase pCLX-UBI-GFP-Luc (kindly provided by Dr. Patrick Salmon, University of Geneva, Switzerland) was performed as described before (Myburgh et al, 2020) in the presence of hexadimethrine bromide (8 µg/ml, Sigma-Aldrich, St. Louis, MO, USA) by spin infection.Subsequent cell sorting of GFP+ populations on a FACS Aria III 4L (Becton Dickinson, Franklin Lakes, NJ, USA) of oligoclonal populations yielded pure MOLM13-TP53 +/+ Luc + and MOLM13-TP53 −/− Luc + populations.The resulting clones were kept in culture and monitored for growth kinetics.Stable clones were selected and cryopreserved for future experiments.

Cloning of lentiviral CAR expression vectors
The published sequences of the scFv domains targeting CD33 (clone SGN-33, Lintuzumab;Feldman et al, 2005) EV1A).The resulting plasmids were packaged into lentiviral vectors to produce viral particles for transduction as described above.

S1B. Expanded untransduced T-cells and CAR T-cells of different
healthy donors were cryopreserved and stored at liquid nitrogen for further use.

In vitro co-culture assays
For in vitro killing assays, purified CAR T-cells were co-cultured with isogenic MOLM13-TP53 or MV4-11-TP53 cells in round-bottom 96well plates at effector-to-target (E:T) ratios ranging from 2 1 :1 to 2 −8 :1 at a total cell number of 60,000 cells per well in R10 cell culture medium at 37 °C and 5% CO 2 .Cells were split every 2 days, supplemented with new R10 medium and kept for up to 10 days, aliquots of the cell culture were used for intermittent flow cytometric analyses as described below.If not stated otherwise, experiments were generally performed with three different healthy donors in technical duplicates and at least two biological replicates.Co-culture assays for mRNA-seq were set up in 96well plates using three full plates for every condition and using the same healthy donor untransduced T-cells and CAR T-cells for the three technical replicates started at different time points.For drug-treated in vitro co-culture assays, the addition of simvastatin, BIO-acetoxime and/or DL-Mevalonolactone was achieved using an HP® D300e Digital Dispenser (HP Inc.) at designated concentrations.Cells were incubated for a total of 6 days, at which flow cytometric analysis was performed.
For pretreatment co-culture assays, targets and effectors were seeded in six-well and 12-well plates, respectively, at a concentration of 1 × 10 6 cells/ml and exposed to simvastatin and/or BIOacetoxime at indicated concentrations for 24 h.In addition, IL-2 (100 U/ml, Gibco™, #CTP0021) was added to effectors during pretreatment period.Cells were then collected, washed with PBS and then seeded for co-culture assays as described above.

In vivo xenograft model
All experiments and procedures involving experimental animals were performed according to protocols approved by the Cantonal Veterinary Office Zurich (Kantonale Veterinäramt Zürich) under the license 194/ 2018 as well as to the ARRIVE guidelines.Animals were maintained under specific pathogen-free (SPF) conditions at the Laboratory Animal Services Center (LASC, Schlieren, Zurich, Switzerland) and allowed to acclimatize for 1 week before use in experiments.Mice were kept in individually ventilated cages (IVC type 2 long; IVC T2L) housing up to five mice per cage with standard housing enrichment (bedding, red mouse house, tissues, and crinklets).The animal facility is equipped with an automated watering system with a connected drinking valve providing fresh water ad libitum.A schematic outline of the murine xenograft leukemia model is presented in Fig. 4A.Female 8-12 weeks old NOD-Prkdc scid -IL2rg Tm1 /Rj mice (Janvier Labs®, Le Genest-Saint-Isle, France) were sublethally irradiated with 100 cGy using an RS-2000 irradiator (Rad Source, Buford, GA, USA).After 6 h mice were injected with 10 5 MOLM13-TP53 +/+ Luc + or MOLM13-TP53 −/− Luc + cells via tail vein injection.A sufficient dose of CD33-directed T-cells to control leukemic disease for at least 3 weeks was established by pilot experiments for three different healthy donor CAR T-cells.Accordingly, 2-8 × 10 6 anti-CD33 CAR T-cells of the respective healthy donors were injected 7 days after MOLM13-TP53 AML cell engraftment via tail vein injection.Bioluminescence measurements were conducted on the indicated days by intraperitoneal injection of D-Luciferin (250 mg/kg bodyweight, Xenolight, PerkinElmer®, Waltham, MA, USA, #122799).Time points for the summary analyzes shown in Fig. 4C were chosen at day 7 and the latest day with data from all mice (before occurrence of termination criteria in any mouse, see below), corresponding to day 15 or day 18 after MOLM13-TP53 AML injection.Mice were monitored for signs of disease and sacrificed by CO 2 euthanasia if any of the following termination criteria were observed: paralysis of any leg, moribund appearance (unresponsive to stimuli, labored respiration, limited ambulation), body weight loss >15-20% of starting weight or dehydration unresponsive to Glucose 5% injections.Part of the mice were sacrificed at the day 35 without showing signs of disease for terminal analysis.Pooled survival data of three biological replicates with three different healthy donors are shown in Fig. 4D.Animal counts per group: T-cells vs. MOLM13-TP53 +/+ Luc + =12, T-cells vs. MOLM13-

mRNA sequencing
Samples for gene expression profiling by mRNA-seq were set up as described above in technical triplicates in three 96-well plates per condition with CD33-directed CAR T-cells of the same healthy donor and MOLM13-TP53 +/+ or MOLM13-TP53 −/− at an E:T of 2 −4 :1 for 6 days.After co-incubation, CAR T-cells and MOLM13-TP53 AML cells were sorted using a FACS Aria III 4L (Becton Dickinson, Franklin Lakes, NJ, USA).Details of the gating strategy and sorting results are provided in Fig. EV3.For mRNA extraction and isolation were performed using the PicoPure™ RNA isolation kit (ThermoFisher, #KIT0204) according to the manufacturer's instructions.Cells were directly sorted into extraction buffer and stored at −80 °C until further processing.High-throughput nextgeneration sequencing was performed at the Functional Genomics Center Zurich on a NovaSeq 6000 Sequencing System (Illumina®, San Diego, CA, USA).Quality control, mapping and differential gene expression analysis was performed using the SUSHI bioinformatics pipeline.Gene set enrichment analysis (GSEA) was performed using the GSEA software jointly developed by UC San Diego and the Broad Institute (Subramanian et al, 2005) and the Enrichr tool (Chen et al, 2013;Kuleshov et al, 2016;Xie et al, 2021).The curated gene set WP_CHOLESTEROL_BIOSYNTHESIS_-PATHWAY was used for the analysis of expression of genes involved in cholesterol biosynthesis.

Chemical compounds
The GSK-3 inhibitor BIO-acetoxime (Selleckchem, Cat.No. #S7915) and simvastatin (Selleckchem, Cat.No. #S1792) were ordered in powder formulation.DL-Mevalonolactone (Medchemexpress, Cat.No. HY-107855) was ordered in oil formulation.BIOacetoxime and DL-Mevalonolactone were dissolved or diluted in Dimethyl Sulfoxide (DMSO) respectively, aliquoted and stored at −80 °C.Simvastatin was activated by alkaline hydrolysis according to the manufacturer's protocol (Sadeghi et al, 2000).Briefly, the lactone pro-drug was dissolved in 95% ethanol at a pH of 1.0 at 50 °C for 2 h before being equilibrated to a pH of 7.2, upon which the drug was aliquoted and stored for further use at −20 °C.Drugs were always freshly thawed and diluted appropriately either in cell culture medium (for in vitro studies), DMSO (for digital dispensing of BIO-acetoxime and DL-Mevalonolactone) or PBS-Tween 0.3%

EMBO Molecular Medicine
Jan Mueller et al (for digital dispensing of simvastatin) to obtain desired concentrations.

CRISPR sequencing
Genome editing was assessed by ultra-deep amplicon nextgeneration sequencing performed at the Massachusetts General Hospital Center for Computational and Integrative Biology DNA core facility (MGH CCIB DNA core, Cambridge, MA).For this, DNA was extracted from cells using the QIAamp DNA Mini Blood Kit (QIAgen, Cat.No. 51304) and subjected to PCR amplification of genomic knockout regions (Phusion ® High-Fidelity PCR Master Mix with GC Buffer, New England BioLabs ® Inc, Cat.No. M0532S).PCR-amplified DNA was then purified (QIAquick PCR Purification Kit, QIAgen, Cat.No. 28104) and sent for sequencing.Purification was visualized by agarose gel electrophoresis (data not shown).

LDLR expression experiments
MOLM13-TP53 and MV4-11-TP53 cells were seeded in 96-well flat-bottom cell culture plates at a density of 1 × 10 5 cells/well.Cells were treated with (titrated) simvastatin and/or DLmevalonolactone at designated concentrations and staining with mouse anti-human LDLR PE (BD Pharmingen™, Cat.No. 565653) was performed to determine mean fluorescence intensity using BD LSRFortessa™ Flow Cytometer with the high-throughput sampler (HTS; Becton Dickinson, Franklin Lakes, NJ, USA) at different time points.Drug application was performed using an HP® D300e Digital Dispenser (HP Inc.).

Luciferase gene reporter assays
For gene reporter assays, plasmids containing SRE-and Wntresponsive promoters coupled to luciferase were purchased from Addgene (Addgene plasmid #90371: pLminP_Luc2P_RE29 for SRE reporter assays and Addgene plasmid #24307: 7TFC for Wnt reporter assays).Plasmids were propagated, and lentiviral vectors produced as described above.For SRE reporter experiments, MOLM13-TP53 and MV4-11-TP53 were transduced with lentiviral supernatants of pLminP_Luc2P_RE29 vectors, sorted for fluorochrome-expressing cells using a FACS Aria 4L (Becton Dickinson, Franklin Lakes, NJ, USA) and cryopreserved in aliquots for further experiments (MOLM13-TP53-SRE and MV4-11-TP53-SRE).Similarly, for Wnt reporter experiments, T-cells isolated from two different healthy donors were thawed, stimulated overnight with CD3/CD28 beads as described above and then transduced with lentiviral supernatants of 7TFC vectors.After bead removal on day 3, T-cells were then propagated in the presence of IL-2 (100 U/ml, Gibco™, Cat.No. #CTP0021), sorted for fluorochromepositive cells and expanded until day 10 after thawing.7TFC expressing T-cells were then cryopreserved in batches for further experiments (7TFC-T-cells).

Statistical analysis
Statistical significance testing was determined by the indicated tests using GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA).For in vivo experiments, no blinding or randomization occurred.All animals were included for the final analysis (no exclusion).The paper explained Problem TP53-mutant acute myeloid leukemia/myelodysplastic neoplasms (AML/MDS) are distinct clinicogenomic entities characterized by chemotherapy resistance, high relapse rates, and poor survival.Chimeric antigen receptor (CAR) T-cell therapy is a successful novel cellbased therapy in hemato-oncology, and it might also be a promising therapeutic option for TP53-mutant AML/MDS.However, the AMLintrinsic determinants of the efficacy of CAR T-cell-based therapies are largely unknown.

Impact
We demonstrate that TP53 deficiency in AML cells confers resistance to CAR T-cell therapy by inducing CAR T-cell dysfunction.We propose a model in which difficult-to-kill TP53-mutant AML cells promote CAR T-cell exhaustion, eventually leading to uncontrolled AML cell outgrowth.We further identify inhibition of the mevalonate pathway as a potential therapeutic vulnerability of TP53-mutant AML cells, and stimulation of the Wnt pathway as a promising avenue to enhance the efficacy of CAR T-cell therapy.Pharmacological co-interventions or genetic engineering of CAR T-cell products may thus be a promising strategy for more efficacious and tolerable cellular therapies for patients with TP53-mutant AML/MDS.Competitive co-culture killing assays
luciferase-expressing MOLM13-TP53 +/+ or MOLM13-TP53 −/− AML cells were injected into the tail vein of sublethally irradiated immunodeficient NOD-scid IL2Rγ null (NSG) mice.After 7 days, 2-8 × 10 6 anti-CD33 CAR T-cells or T-cell controls were injected via the tail vein (Fig. 3A).Mice injected with control T-cells showed rapid AML progression as determined by bioluminescence imaging and had to be sacrificed within 3 weeks because of symptomatic leukemic disease irrespective of the TP53 genotype of the injected MOLM13 cells (Figs. 3B-D and EV2).Mice injected with MOLM13-TP53 −/− AML cells and treated with anti-CD33 CAR T-cells started to succumb to leukemic disease within 3 weeks as well.By contrast, mice injected with MOLM13-TP53 +/+ AML cells
TP53 +/+ (CART)' (negatively correlated) 'MOLM-TP53 -/-(CART)' (positively correlated) -cultures of CAR T-cells and AML cells with wildtype TP53 did not lead to increased killing of AML cells (Fig.5B,C and EV4E,F).The addition of mevalonate to simvastatin-containing coincubation assays completely restored the TP53 deficiency-associated resistance of AML cells to CAR T-cell killing with no effect on killing of MOLM13-TP53 +/+ AML cells (Fig.5B,C and EV4E).Assaying the established hallmarks of CAR T-cell dysfunction in the context of TP53-mutant AML (Fig.2), we observed a rescue of CAR T-cell expansion upon simvastatin addition (Fig.EV4G).CAR T-cells co-incubated with MOLM13-TP53 −/− AML cells in the presence of simvastatin showed reduced exhaustion markers (PD-1, LAG3, TIM3) compared to the DMSO controls (Figs.5D and EV5A), and fluorescence live-cell imaging revealed normalization of CAR T-cell:AML cell interaction times upon addition of simvastatin (Figs.5E and EV5B).

Figure 6 .
Figure 6.Pretreatment with simvastatin but not with BIO-acetoxime rescues CAR T-cell killing of TP53-deficient AML cells.(A) Graphical representation outlining pretreatment co-incubation assays.(B) LDLR expression on the surface of MOLM13-TP53 cells upon simvastatin 1 μM treatment for 24 h.(biological replicates, n = 3; two technical replicates per biological replicate; symbols indicate individual replicates; thickened line indicates mean and error bars indicate SD; two-way ANOVA).(C) Luminescence signal from T-cells transduced with a Wnt-responsive luciferase gene reporter system and incubated with or without BIO-acetoxime at 0.5 μM for 1, 4, or 6 days.(data shown from one healthy T-cell donor; biological replicates, n = 2; two technical replicates per biological replicate; symbols indicate individual replicates; thickened lines indicate mean and error bars indicate SD; unpaired Student's t test).(D) Results of pretreatment co-incubation assays with specific anti-CD33 CAR T-cell-mediated killing of MOLM13-TP53 +/+ or MOLM13-TP53 −/− AML cells at various E:T ratios.(biological replicates, n = 3; two technical replicates per biological replicate; symbols represent means; error bars indicate SD; two-way ANOVA).(E) Summary data highlighting results from (D) at an E:T ratio of 1:8 (biological replicates, n = 2; two technical replicates per biological replicate; symbols represent individual replicates; thickened lines indicate mean and error bars indicate SD; two-way ANOVA).(F) Graphical representation outlining co-incubation assays with CRISPR/Cas9-edited CD33-and CD123-directed CAR T-cells against TP53 +/+ or TP53 −/− AML cells.nc: unperturbed CAR T-cells; mock: CAR T-cells electroporated with nonsense gRNA and Cas9; RegKO: CAR T-cells gene-edited by CRISPR/Cas9 to induce Regnase-1 deficiency.(G) Results of co-incubation assays showing CRISPR/Cas9-edited CD33-and CD123-directed CAR T-cell-mediated killing (nc, mock and RegKO) against MOLM13-TP53 and MV4-11-TP53 AML cells.(biological replicates, n = 2; two technical replicates per biological replicate; symbols represent means; error bars indicate SD; two-way ANOVA).(H) Summary data highlighting results from (G) of gene-edited anti-CD33 CAR T-cell mediated killing against MOLM13-TP53 at an E:T of 1:32 (biological replicates, n = 2; two technical replicates per biological replicate; symbols represent individual replicates; thickened lines indicate mean and error bars indicate SD; two-way ANOVA).Source data are available online for this figure.

Figure 7 .
Figure 7. TP53 deficiency in AML leads to resistance against CAR T-cell-mediated killing that can be overcome by targeting the mevalonate and Wnt pathways.Graphical representation of our proposed model.The intrinsic apoptotic defect in TP53-deficient AML cells leads to a longer duration of cellular interaction between CAR T-cells and TP53-deficient AML cells.Longer temporal interaction between CAR T-cells and TP53-deficient AML eventually leads to reduced CAR T-cell proliferation and enhanced CAR T-cell exhaustion resulting in the outgrowth of TP53-deficient AML cells.Inhibition of cholesterol biosynthesis in TP53-deficient AML cells or activation of the Wnt pathway in CAR T-cells can overcome resistance of TP53-deficient AML cells to CAR T-cell-mediated killing.
, with CAR T-cell to AML cell attachment until AML cell or CAR T-cell death as judged by propidium iodide influx or detachment of CAR T-cell from the AML cell scored for each individual attachment event.Three different healthy donor CAR T-cells were used as three biological replicates.Events per group: CAR T-cells vs. MOLM13-TP53 +/+ = 102, CAR T-cells vs. MOLM13-TP53 −/− = 88.
Boettcher S, Miller PG, Sharma R, McConkey M, Leventhal M, Krivtsov AV, Giacomelli AO, Wong W, Kim J, Chao S et al (2019) A dominant-negative effect drives selection of TP53 missense mutations in myeloid malignancies.Science 365:599-604 Borot F, Wang H, Ma Y, Jafarov T, Raza A, Ali AM, Mukherjee S (2019) Geneedited stem cells enable CD33-directed immune therapy for myeloid malignancies.Proc Natl Acad Sci USA 116:11978-11987 study shows that TP53 mutations in AML cells lead to increased resistance to CAR T-cells in vitro.CAR T-cells co-incubated with target TP53-mutant AML blasts exhibited decreased proliferation and increased exhaustion compared to wild-type TP53 AML blasts.Live-cell imaging revealed longer time-to-killing of TP53-mutant than wild-type TP53 AML cells upon attack by CAR T-cells, which ultimately led to an inability of CAR T-cells to control TP53-mutant cells.Furthermore, immunodeficient mice xenografted with TP53-mutant AML and treated with CAR T-cells exhibit shorter survival compared to mice engrafted with wild-type TP53 AML.Transcriptional profiling of AML cells with either wild-type or mutant TP53 under CAR T-cell attack revealed upregulation of the mevalonate synthesis pathway in TP53-mutant AML cells.Simultaneously, CAR T-cells engaging TP53-mutant AML demonstrated a downregulated Wnt pathway.Rational pharmacological targeting of either of these pathways rescued TP53-mutant AML cell sensitivity to CAR T-cell-mediated killing.Similarly, CRISPR/Cas9engineering of CAR T-cells to upregulate Wnt pathway signaling rescued TP53-mutant AML cell sensitivity and led to improved killing efficacy.

Figure EV1 .
Figure EV1.Details of TP53-associated resistance to in vitro CAR T-cell killing, relates to Figs. 1 and 2. (A) Vector design for second-generation CAR expression under an EF1 promotor with a CD8α hinge, a 4-1BB costimulatory domain, a CD3ζ activating domain and an RQR8 identification and selection peptide.(B) Lentiviral T-cell transduction and MACS purification yielded a > 95% pure CAR T-cell population for further experiments.(C) In vitro growth kinetics of unperturbed MOLM13-TP53 +/+ and MOLM13-TP53 -/-leukemia cells.(D) Graphical representation of in vitro competitive co-incubation assay.(E) Representative FACS plots of in vitro killing assays showing ratio of MOLM13-TP53 −/− GFP + over MOLM13-TP53 +/+ RFP + upon 10 days of co-incubation with untransduced T-cell controls and or CAR T-cells.Percentages of parental populations are shown.(F) −Log of MOLM13-TP53 +/+ /MOLM13-TP53 −/− ratios normalized to the calculated initial MOLM13-TP53 +/+ /MOLM13-TP53 -/-ratios plotted against time of co-incubation.Pooled results of all different E:T and KO:WT ratios are shown (biological replicates, n = 2; three technical replicates per biological replicate; symbols represent means and error bars indicate SD; two-way ANOVA).(G) Calculated specific killing from in vitro competitive co-incubation assays of anti-CD33 CAR T-cells against a mixture of MOLM13-TP53 +/+ and MOLM13-TP53 −/− leukemia cells.(H) Absolute CD3 + cell numbers at an E:T ratio of 1:16 for untransduced T-cell controls and anti-CD33 CAR T-cells co-incubated with MOLM13-TP53 +/+ (black), MOLM13-TP53 -/-(red) or MOLM13-TP53 missense/-(blue) AML cells on day 6 (biological replicates, n = 2; two technical replicates per biological replicate; symbols represent individual replicates; thickened lines indicate means and error bars indicate SD; two-way ANOVA).(I) CD33 target antigen density on MOLM13-TP53 +/+ and TP53 −/− AML cells in coincubation at an E:T of 1:16 with untransduced T-cell controls and anti-CD33-directed CAR T-cells on days 1 and 6 (biological replicates, n = 4; three technical replicates for each biological replicates; symbols represent individual replicates; thickened lines indicate means and error bars indicate SD; two-way ANOVA).(J) PD-L1 surface expression on MOLM13-TP53 +/+ and TP53 -/-AML cells in co-incubation with untransduced T-cell controls and anti-CD33-directed CAR T-cells on days 1 and 6 (biological replicates, n = 3; two technical replicates for each biological replicate; symbols represent individual replicates; thickened lines indicate means and error bars indicate SD; two-way ANOVA).

Figure EV5 .
Figure EV5.Details of rescue and pretreatment co-incubation assays as well as Regnase-1-deficient CAR T-cells, relates to Figs. 5 and 6. (A) Fraction of T-cells negative for all three investigated exhaustion markers (right) and negative for all three markers (left) treated with the indicated compounds (one biological replicate, n = 2 technical replicates; symbols represent means; error bars indicate SD; two-way ANOVA).(B) Representative stills of fluorescence live-cell timelapse imaging data of anti-CD33 CAR T-cell engaging MOLM13-TP53 +/+ AML or MOLM13-TP53 -/-AML cells in the presence of simvastatin 1 μM or BIO-acetoxime 0.5 μM (scale bars, 10μm).(C) Luminescence signal of T-cells transduced with a Wnt-responsive luciferase gene reporter system and incubated with or without BIO-acetoxime at 0.5 μM for 1, 4 or 6 days as well as after washout of BIO-acetoxime.(Data shown from a second healthy T-cell donor than in Fig. 6C; biological replicates, n = 2; 2 technical replicates per biological replicate; symbols indicate individual replicates; thickened lines denote means and error bars indicate SD; unpaired Student's t test).(D) Results from 3 different pretreatment co-incubation assays with anti-CD33 and anti-CD123 CAR T-cells against TP53 +/+ or TP53 -/-AML (MOLM13 and MV4-11) cells at various E:T ratios.Co-incubations include: anti-CD123 CAR vs MOLM13-TP53, anti-CD33 CAR vs MV4-11-TP53 and anti-CD33 CAR vs MV4-11-TP53.(biological replicates, n = 2-3; 2 technical replicates per biological replicate; symbols represent means; error bars indicate SD; two-way ANOVA).(E) Heatmaps depicting LDLR expression (mean fluorescence intensity) within treatment as well as pretreatment co-incubation assays with anti-CD33 and anti-CD123 CAR T-cells against MOLM13-TP53 and MV4-11-TP53 AML cells at various E:T ratios.(biological replicates, n = 3-4; 2 technical replicates per biological replicate; pseudocolors indicate signal of LDLR expression).(F) Sequencing results graphed as %indels of Regnase-1 CRISPRed anti-CD33 as well as anti-CD123 CAR T-cells from one healthy donor.(G) RT-qPCR measuring expression of TCF7 transcripts normalized to ACTB in nc, mock and RegKO anti-CD33 and anti-CD123 CAR T-cells (biological replicates, n = 2; 2 technical replicates per biological replicate; symbols indicate individual replicates; thickened lines indicate means and error bars indicate SD; paired Student's t test).