RapaCaspase-9-based suicide gene applied to the safety of IL-1RAP CAR-T cells

The safety of ACT is very important because of the potential for on-target/off-tumor, off-target and CRS toxicities; indeed, the on-target/off-tumor toxicities caused by CAR-T cells directed against HER2 [27] or a transgenic TCR directed against MAGE-13 protein led to respiratory failure and neurotoxicity [28]. Moreover, off-target toxicities, such as nonspecific recognition by the TCR, lead to off-target/off-tumor toxicities, such as a MAGE-A3-specific transgenic TCR recognizing the Titin protein and causing cardiac death [29, 30]. Other toxicities such as CRS, which induces more or less dangerous symptoms, and the neurotoxicities notably reported with CD19-specific CAR-T cells [31] have to be managed, and suicide genes are a supplementary safety measure in certain cases.

Our CAR-T-cell targeting the cell-surface protein IL-1RAP was functionally validated in chronic myeloid leukemia and AML and will be evaluated in a clinical trial; and its suicide gene functionality, that of the iCaspase9 suicide gene, was already validated. However, the inducer of iCaspase9 suicide gene is not yet available as an approved drug which limits the use of this suicide gene system in the clinic. To make the use of the suicide gene in the clinic easier, the RapaCaspase9 suicide gene would be a great alternative: it is induced by a well-known and already use drug (rapamycin), available as off-the-shelf pharmaceutical licensed for commercial sale worldwide with a good biodistribution. Moreover, it is known to have an anti-cancer activity, even in the case of activation of the safety switch, the dose will be very low. This medication was approved in 1999 for prevention of renal graft rejection [32] and exerts immunosuppressive activity: this activity could create a real interest because of the high T-cell activity in ACT but will not be harmful given the needed dose for suicide gene activation in comparison to that for prevention of renal graft rejection [12].

The RapaCaspase9 suicide gene is a fusion of a mutated FKBP12 protein with the rapamycin FRB of mammalian target rapamycin. FKBP12-FRB is fused to the catalytic domain of Caspase 9, with a short linker between the FRB and FKBP12. Physiologically, caspase 9 is activated by oligodimerization of its caspase activation and recruitment domain with the apoptosome. For the RapaCaspase9 suicide gene, rapamycin induces heterodimerization of FRB and FKBP12, allowing the activation of the caspase 9 catalytic domain.

Caspase 9 plays an important role from the beginning to the end of an individual’s life; indeed, it has been demonstrated that decreased apoptosis, caused by mutations in the caspase 9 gene, results in embryonic lethality and defective brain development [20, 33,34,35]. The importance caspase 9 was also confirmed by the existence of different SNPs in endogenous caspase 9 that are described to have functional impacts on its efficiency and important consequences. This is the case for the SNP Ex5+32G>A (rs1052576), which shows a protective effect for the AA alleles in different cancers, such as glioma brain tumors [20], non-small cell lung cancer [21], and B-cell lymphoma [22], and a negative impact for the GG allele indicated by a survival reduction [23].

In this study, the SNP Ex5+32G>A (rs1052576) was incorporated into the iCaspase9 and RapaCaspase9 suicide gene improving the safety of IL-1RAP-targeting CAR constructs to assess its impact and optimize the efficiency of the suicide gene.

Our first works on suicide genes were developed with retroviral constructs containing the iCaspase9 suicide gene with one or the other SNP: the SNPs showed an important impact on suicide gene efficiency, but unlike endogenous caspase 9, the G nucleotide is responsible for better induction of cell death. With biotechnological advances, retroviral constructs, allowing bad and more difficult transductions, are progressively being replaced by lentiviral constructs; indeed, the ability of retroviral particles to transduce only cells undergoing division considerably reduces the efficiency of primary T-cell genetic modification. In the context of improving the safety of our CAR-T cells, RapaCaspase9 suicide genes with the SNP Ex5+32G>A (rs1052576) were introduced into lentiviral constructs: unlike for the retroviral constructs, the SNP did not show an impact on induced cell death with RapaCaspase9, but the –G construct presented a better efficiency with the minimal effective dose determined previously [12]. Similar results were also observed with a range of doses of rapamycin, confirming, first, the weak impact of the SNP on RapaCaspase9 suicide gene induction and, second, the use of the RapaCasp9-G suicide gene to improve the safety of our construct. In regard to the difference observed between lentiviral and retroviral constructs, we hypothesized that the different promoters in the constructs explain this variation, but this hypothesis needs to be validated with supplementary investigations.

To study this medication in conditions of a clinical trial, we have to be closer to clinical conditions and treatments given to future patients: for that, the next doses studied were calculated to be similar to taking a pill and we could confirm the efficiency of clinical doses for suicide gene induction. In addition, we showed, in vitro, that rapamycin (Rapamune®) pills are as efficient as rapamycin powder in treating RapaCasp9-G-expressing GMTCs. This supports that Rapamune® pills can be used in conditions of a clinical trial, even if additional studies need to be perform. Moreover, in the context of improving the safety of this therapy, the speed of suicide gene induction must be as fast as possible to quickly eliminate GMTCs. Even though induced cell death reached a plateau at 48 h after rapamycin exposure, most of the genetically modified cells were already eliminated after 24 h and no longer represented a danger to the patient. In comparison to other suicide gene systems for which induction is efficient between 4 and 6 days [36], the RapaCasp9-G suicide gene exhibited faster induction. In case of 24 h exposure to rapamycin is not enough to correctly induce suicide gene induction, we could plan to expose patients to a second consecutive dose, as it showed a better suicide gene induction than 24 h exposure (Supplementary Fig. S6). For the purpose of a clinical trial, the IL-1RAP CAR-T-cell production process must be GMP compliant. To ensure the efficiency of the RapaCasp9-G suicide gene in these conditions, the freezing and activation process must not have an impact on it: notably, the RapaCasp9-G suicide gene still showed the same functionality regardless of the activation process in HDs. Moreover, even without a rest day to restart GMTCs metabolism after thawing and in the presence or absence of target stimulation, high rapamycin-induced killing percentages confirmed the RapaCasp9-G suicide gene functionality.

More importantly, as therapy with IL-1RAP CAR-T cells is developed for AML patients, the functionality of the RapaCasp9-G suicide gene in RapaCasp9-G-expressing GMTCs produced from AML patients is an important milestone. Our experiments strongly confirmed this functionality and support the interest in suicide gene use in our CAR-T-cell therapy targeting the IL-1RAP protein in different activation processes. Unfortunately, the scarcity of samples from AML patients prevented us from performing further study.

Finally, an important characteristic of the CID is acceptable or limited side effects. As a side effect, rapamycin showed only transitory transgene expression changes, confirming the specific elimination of RapaCasp9-G-expressing GMTCs, and a slowdown in the proliferation of untransduced T cells, which could be turned into a potential advantage. Indeed, we can speculate that this effect could produce an observable impact on RapaCasp9-G-expressing GMTCs: in the case of suicide gene use, there could be a synergistic effect between the slowdown of T-cell metabolism, regardless of whether the T cells are genetically modified, and the specific elimination of RapaCasp9-G-expressing GMTCs. Moreover, rapamycin induced only transitory transgene expression changes and specifically eliminated GMTCs.

Taking into account the efficiency of rapamycin, its limited side effects, its “off-the-shelf” status and its specificity in eliminating target cells, this CID seems to be a great candidate as a suicide gene inducer. Furthermore, in order to better eliminate GMTCs in case of toxicities, a second suicide gene system could be added to enhance GMTCs elimination if necessary [37]. Moreover, the whole point of improving safety with a suicide gene is the association with a cell therapy in order to be as close as possible to the real conditions of use: this model thus perfectly meets the conditions for validated in vitro functionality.

留言 (0)

沒有登入
gif