Targeting the Activating Mutations of NOTCH1 in T-Cell Lymphoblastic Leukemia with a New SERCA Inhibitor CAD204520

Given its oncogenic role in human cancers, Notch1 signaling has garnered increased attention as a therapeutic target. Several Notch modulators, including γ-secretase inhibitors, have shown therapeutic efficacy in preclinical tumor models. However, despite this promise, few of these candidates have shown clinical benefit in patients, in part due to tissue-dependent on-target toxicities from the repression of both mutant and wild type Notch proteins.

The discovery of the P-type ATPase Sarco/Endoplasmic Reticulum Ca²⁺-ATPase (SERCA) as a modulator of oncogenic NOTCH1 suggested a new approach to treat T-cell Acute Lymphoblastic Leukemia (T-ALL). In fact, thapsigargin mediated SERCA inhibition had a stronger effect on the most common type of activating NOTCH1 mutants compared to wild type. However, the consequence of thapsigargin binding to SERCA is a rapid increase in cytosolic Ca²⁺ and a depletion of Ca²⁺ stored in the endoplasmic reticulum (ER). This shift of Ca²⁺ efflux might cause cardiac toxicity in human, suggesting the need to identify inhibitors with better drug-like properties and reduced off-target toxicity.

From a small molecule screening of 191,000 P-type ATPase modulators CAD204520 displayed ~25 and ~79-fold greater selectivity toward human SERCA compared to Na⁺/K⁺ and H+-ATPase respectively and emerged as a candidate for the development of novel anti-NOTCH1 agents.

In a series of in vitro studies, we showed that CAD204520 impairs the proliferation of a panel of lymphoid malignancies carrying activating mutations in the NOTCH1 heterodimerization domain (HD, such as T-ALL) and/or deletions in the degradation domain (PEST, such as mantle cell lymphoma (MCL)). Importantly, T-ALL lymphoblasts from patients carrying NOTCH1 mutations were more sensitive to CAD204520 suppression compared with normal lymphocytes or ALL wild type NOTCH1. Similar to thapsigargin, CAD204520 causes a defect in NOTCH1 trafficking that results in an accumulation of full-length NOTCH1 and in a depletion of the activated form: ICN1. Consequently, the NOTCH1 targets, MYC and DTX1, were repressed as measured by qRT-PCR. In contrast, we observed no effect at protein level in T-ALL wild type NOTCH1 suggesting that clinically relevant NOTCH1 mutations including PEST deletions are more sensitive to CAD204250 treatment than wild type proteins.

The next question is whether CAD204520 activity has limitations in vivo due to Ca²⁺ shifts. Thus we analysed cardiac mechanics in isolated rat cardiomyocytes and showed that compared to thapsigargin, CAD204520 minimally alters myocardial performance (~25%) suggesting that the heart will likely tolerate CAD204520 modulation in vivo. This effect may be caused to differences between Ca²⁺ mediated ER stress upon thapsigargin or CAD204520 treatment. To test this hypothesis, we treated T-ALL cell lines at equimolar doses of CAD204520 and thapsigargin. We showed that, despite a similar effect on NOTCH1 protein levels, CAD204520 does not induce the unfolded protein response (UPR) pathway as measured by the phosphorylation of eIF2α or the induction of the ER chaperone BIP. This result suggests that CAD204520 retains ATPase activity, responsible for Notch inhibition, decoupled from the Ca²⁺ transport activity responsible for the off-target effects.

A tempting hypothesis to explain these differences is that thapsigargin and CAD204520 inhibit SERCA in distinct pockets. Consistently, docking poses showed that CAD204520 binds SERCA in the trans-membrane helices M1, M3 and M4 while thapsigargin in the M3, M5 and M7 groove. Consequently, thapsigargin and CAD204520 co-treatment resulted in synergistic anti-proliferative effect in T-ALL.

To test CAD204520 in vivo we studied its effect in CD1 mice and established a maximum tolerated dose. Mice exposed to CAD204520 for 21 days at 30 mg/Kg BID showed no decrease in body weight and, importantly no effect on cardiac hemodynamic. Finally, to assess the in vivo efficacy of CAD204520 we xenografted SKW-3/KE-37 T-ALL cell line in L‐2Rγ null mice and demonstrated that drug's treatment reduced circulating and tissue infiltrating human leukemia T-ALL cells without causing heart-related off target effects.

In conclusion, this study supports the development of tolerated SERCA inhibitors for Notch dependent cancers and extends its application to cases with mutations in the PEST degradation domain such as MCL or chronic lymphocytic leukemia.