Mutations in Ras family proteins are implicated in 33% of human cancers, but direct pharmacological inhibition of Ras mutants remains challenging. As an alternative to direct inhibition, we screened for sensitivities in Ras-mutant cells and discovered 249C as a Ras-mutant selective cytotoxic agent with nanomolar potency against a spectrum of Ras-mutant cancers. 249C binds to vacuolar (V)-ATPase with nanomolar affinity and inhibits its activity, preventing lysosomal acidification and inhibiting autophagy and macropinocytosis pathways that several Ras-driven cancers rely on for survival. Unexpectedly, potency of 249C varies with the identity of the Ras driver mutation, with the highest potency for KRASG13D and G12V both in vitro and in vivo, highlighting a mutant-specific dependence on macropinocytosis and lysosomal pH. Indeed, 249C potently inhibits tumor growth without adverse side effects in mouse xenografts of KRAS-driven lung and colon cancers. A comparison of isogenic SW48 xenografts with different KRAS mutations confirmed that KRASG13D/+ (followed by G12V/+) mutations are especially sensitive to 249C treatment. These data establish proof-of-concept for targeting V-ATPase in cancers driven by specific KRAS mutations such as KRASG13D and G12V.
Mutations that constitutively activate members of the Ras family of oncogenes (H-Ras, N-Ras, and K-Ras) are collectively responsible for about one third of all human cancers; KRAS mutations in particular are implicated in the most fatal malignancies: pancreatic (91%), colon (42%) and lung (33%)1,2,3,4. Despite decades-long efforts, we still lack clinically-approved drugs for the majority of oncogenic KRAS variants (apart from Amgen’s KRASG12C inhibitor), in large part because of substantial challenges in directly inhibiting mutant KRAS. As an alternative strategy for pharmacological intervention, efforts to map potentially druggable neomorphic dependencies of KRAS-mutant cancers have identified, among others, metabolic adaptations concomitant with activating KRAS mutations, including high levels of basal autophagy, a phenomenon termed ‘autophagy addiction’5,6,7,8,9, and constitutive activation of macropinocytosis (MP)10. Both autophagy and MP provide nutrients to promote tumor growth by degrading macromolecules in acidic lysosomes produced by vacuolar H+ ATPases (V-ATPases)11,12, and V-ATPase has been reported as an essential regulator of Ras-induced MP for nutrient supply, suggesting that targeting V-ATPase could be exploited to curtail the metabolic adaptation of Ras-mutant cells10.
V-ATPase is a multi-subunit transmembrane complex that operates as an ATP-driven rotary proton pump, coupling ATP hydrolysis in its peripheral V1 domain to proton translocation from the cytoplasm to the lumen of organelles through its membrane integral Vo domain. Through this activity, V-ATPase is essential for lysosome acidification and thus lysosomal degradation13,14 and more broadly regulates a wide array of membrane trafficking and intracellular transport processes15. Although specific V-ATPase inhibitors such as Bafilomycin A1 (BafA1) have been developed and serve as important tool compounds, their clinical use is limited by toxicity16. Additional autophagy-modulating agents have been described, such as the repurposed antimalarial compound chloroquine (CQ) and its less toxic derivative hydroxychloroquine (HCQ), which impair lysosomal acidification by targeting palmitoyl-protein thioesterase 1 (PPT1), but these compounds have severe side effects by interfering with human ether-à-go-go-related gene (hERG) responsible for electrical activity of the heart17,18 and they can also cause gastrointestinal side effects, impact visual and auditory function, and lead to hypoglycemia. In over 60 clinical trials (completed or ongoing), CQ or HCQ alone as a monotherapy do not demonstrate substantial therapeutic efficacy19. Although second-generation analogs of HCQ such as ROC325, Lys05, DC661, and DQ661 show enhanced preclinical lysosomal-autophagic inhibition and antitumor activity as single agents in vitro and in vivo, it remains to be determined whether any of them will show efficacy in clinical investigations while avoiding unacceptable toxicities17,20. These limitations motivate an urgent need to identify and develop additional autophagy/MP inhibitors with clean safety profiles.
Here we report the discovery and characterization of a class of dihydro-pyrazole-5-carboxamide compounds with exquisite potency in killing KRAS-mutant cells and excellent pharmacological properties. By combining chemoproteomics, comparative profiling, and chemo-genetic clustered regularly interspaced short palindromic repeats (CRISPR) screening, we determined that our lead compound 249C binds to and inhibits V-ATPase, resulting in inhibition of autophagy, MP, and lysosomal acidification. Through this activity, 249C selectively kills KRAS/BRAF-mutant cells, with the highest activity against KRASG13D and G12V mutant cells, pointing to excessive reliance on V-ATPase as a unique vulnerability resulting from these KRAS mutations. For over 1 year, the molecule has been in a multi-center Phase Ia/b dose escalation and dose expansion clinical trial for the following indications: Ras mutations, lung, colon, and pancreatic cancers across five different sites in the U.S. (NCT04678648; 249C = RSC-1255).
249C is a potent inhibitor of Ras- and Raf-mutant cells
We had recently synthesized ~300 dihydro-pyrazole-5-carboxamide small molecules to target a developmental pathway (Supplementary Fig. 1 and Supplementary Table 1). Here to evaluate the ability of these molecules to also inhibit growth of Ras- and Raf-mutant cells, we screened these molecules for anti-proliferative activity against three human cancer cell lines with mutations in Ras or Raf: A549 (KRASG12S); LOX IMVI (BRAFV600E); and MelJuso (HRASG13D and NRASQ61L). Using structure-activity relationships, we observed that several compounds potently reduced viability of all three cell lines (Fig. 1a and Supplementary Tables 1–3). Simultaneous evaluation of drug-like properties and half-maximal inhibitory concentration (IC50) values revealed 249C as the lead compound on the basis of physico-chemical properties (Fig. 1b,c). Compound 249C was chosen over 226C for further studies owing to the presence of a tertiary amine, imparting higher water solubility, and a slightly lower cLogP (3.07 versus 3.33). To further evaluate determinants of sensitivity to 249C, we measured its potency against 53 cell lines (lung, pancreatic, breast, ovarian, colorectal/gastric, prostate, liver/hematological, glioma, and melanoma) (Fig. 1d). Comparing IC50 values to gene mutation data from the Cancer Cell Line Encyclopedia (CCLE) revealed that the presence of at least one mutation in KRAS, HRAS, NRAS, and/or BRAF was significantly associated with lower IC50 values (P = 1.24 × 10−6), with median IC50 values of 65 nM in KRAS/HRAS/NRAS/BRAF-mutant cells versus 365 nM in cells with no mutations in these genes (Fig. 1e). KRAS showed the strongest single gene association with IC50 values (P = 2.34 × 10−3) (Supplementary Fig. 1b,c). We next mapped the response of A549 cells to these compounds using liquid chromatography–tandem mass spectrometry (LC–MS/MS)-based protein profiling coupled with quantitative tandem mass tags labeling21. Upon treatment of cells with vehicle or molecules 68, 226C and 249C, we observed substantially elevated levels of the autophagy receptor sequestosome 1 (SQSTM1/p62) common to this family of compounds (Fig. 1f) in a time-dependent fashion (Fig. 1g), suggesting that these compounds impact autophagy.
249C inhibits autophagy
To characterize the mode of action of our lead compound 249C, we used comparative profiling in a panel of primary-cell-based systems designed to model disease biology22. In comparison to 4,000 other drugs, the phenotypes elicited by 249C most closely resembled those elicited by the V-ATPase inhibitor BafA1 (Pearson’s r = 0.95 across all measurements; Fig. 2a,b, Supplementary Figs. 2 and 3, and Supplementary Table 4). 249C also exhibited similarity, albeit to a lesser degree, to the mTOR inhibitors rapamycin and Torin1 and the antimalarials CQ and quinacrine. 249C and BafA1 elicited progressive accumulation of SQSTM1/p62 and microtubule-associated protein 1A/1B-light chain 3 (LC3-II) in A549 cells, consistent with a block in autophagic flux; by contrast, the autophagy inducer rapamycin did not (Fig. 2c). Using transmission electron microscopy, we further observed late-stage autophagy inhibition and increased autophagic vesicle accumulation in A549 cells treated with 249C and BafA1 (Fig. 2d–i). Finally, treatment with both 249C and BafA1 caused a reduction in the presence of acidic organelles (Fig. 2j) and an increase in pH (Fig. 2j) of A549 cells relative to dimethyl sulfoxide (DMSO) and Torin1, confirming that 249C blocks lysosomal acidification and autophagy progression.
Chemical–genetic screen reveals V-ATPase as the target of 249C
To further pinpoint the molecular target of 249C, we turned to CRISPR interference (CRISPRi)-mediated chemical–genetic screening, which has emerged as a robust strategy for identification of small molecule targets23. To allow us to work with a previously validated CRISPRi cell line, we evaluated a panel of available CRISPRi cell lines for sensitivity to 249C and found that MDA-MB-231 cells (breast adenocarcinoma, KRASG13D and BRAFG464V) were highly sensitive. Thus, we conducted these screens in MDA-MB-231 cells. In brief, we infected MDA-MB-231 cells stably expressing the dCas9-KRAB CRISPRi effector24 with the genome-scale hCRISPRi-v2 single-guide RNA (sgRNA) library (targeting 18,905 genes)25, harvested a subpopulation at the outset of the experiment (t0), and then cultured the remaining cells without (DMSO) or with 249C treatment (Methods). Finally, we used next-generation sequencing to determine how each sgRNA impacted sensitivity to 249C (ρ) and growth in the absence of 249C (γ)26 (Fig. 3a). The resulting untreated growth phenotypes were well correlated with those from other cell types, suggesting that our screen faithfully captured phenotypes resulting from gene knockdown (Supplementary Fig. 4). Among the 147 genes for which knockdown strongly affected 249C sensitivity (Fig. 3b and Supplementary Tables 5 and 6), we were intrigued to find five V-ATPase subunits (ATP6V1H, ATP6V1A, ATP6VoA2, ATP6VoA3/TCIRG1, and ATP6VoE1), because 249C essentially phenocopied the V-ATPase inhibitor BafA1 in our previous assays (Fig. 2c). Indeed, knockdown of several V-ATPase subunits sensitized cells to 249C; the regulatory H subunit ATP6V1H in particular was one of the most strongly sensitizing hits. By contrast, knockdown of the poorly characterized subunits ATP6VoE1 and ATP6VoA3 (TCIRG1) protected cells against 249C. Individual re-tests using internally controlled drug-sensitivity assays further confirmed that knockdown of multiple V-ATPase subunits sensitized MDA-MB-231 cells to 249C and BafA1 (target: ATP6VoC)27 (Fig. 3c). Thus, genetic depletion of V-ATPase sensitizes cells to 249C, further suggesting that the cytotoxicity of 249C arises from inhibition of V-ATPase activity. Although we cannot rule out polypharmacology, the V-ATPase complex stood out on the basis of multiple subunit hits.
249C binds and inhibits V-ATPase
With multiple independent assays pointing toward V-ATPase as the target of 249C, we proceeded to validate the biochemical and biophysical effects of 249C on V-ATPase activity. V-ATPase activity depends upon reversible assembly of the V1 and Vo domains and is nutrient dependent; increased V1–Vo assembly increases catalytic activity15. To test whether 249C alters localization of V-ATPase subunits, we evaluated the partitioning of V-ATPase subunits into the cytosolic (containing the V1 domain composed of eight subunits (A–H) responsible for ATP hydrolysis) and membrane fractions (membrane-embedded Vo domain comprised of six subunits (a, c, c′ (absent in mammals), c′′, d and e) responsible for proton translocation)15 of 249C-treated HEK293T cells. Treatment with 249C but not BafA1 led to an increase in membrane-associated subunit B2 (cytosolic V1 domain), possibly indicating increased V-ATPase assembly (Fig. 4a,b). To independently corroborate the effect of 249C on V-ATPase and to probe whether proton pumping is impacted in the complexes on lysosomes, we measured lysosome acidification using a pH-dependent fluorescent probe, FITC-Dextran. Both 249C and BafA1 inhibited V-ATPase-dependent proton transport in mammalian cells, as measured by ATP-dependent fluorescence quenching of FITC-loaded lysosomes28 (Fig. 3c). Proton translocation through the Vo region is powered by ATP hydrolysis in the V1 region. In addition to blocking acidification, both 249C and BafA1 blocked ATP hydrolysis of intact mammalian V-ATPase purified from pig kidneys (but not yeast; Supplementary Fig. 5) at concentrations as low as 1 µM, as measured by a standard biochemical V-ATPase activity assay (Fig. 3d). To directly measure binding of 249C to V-ATPase, we performed bio-layer inferometry (BLI) with both the purified intact mammalian V-ATPase complex and the recombinant ATP6V1H subunit using biotinylated 249C loaded onto streptavidin sensors (Fig. 5e,f), which revealed binding with nanomolar affinities (Kd: complex = 23 nM; H sub = 501 nM) (Fig. 3e,f and Supplementary Fig. 6). Altogether, these results establish that 249C directly binds to and inhibits mammalian V-ATPase.
249C preferentially kills mouse embryonic fibroblasts bearing specific KRAS mutations
The most frequent KRAS mutations across all human cancers occur at codons 12 or 13 with replacement of glycine to other amino acids: G12D (35%); G12V (24%); and G13D (13%)29. To test the impact of 249C on cells carrying individual Ras/Raf mutations, we measured the potency of 249C against immortalized mouse embryonic fibroblasts (MEFs) genetically engineered to be ‘Ras-less’ with subsequent stable transduction of human wild-type (WT) or mutant KRAS/BRAF isoforms30 obtained from the National Cancer Institute. The MEF system: (1) permits comparison of individual KRAS mutations in the absence of other oncogenic drivers found in cancer cells; (2) is routinely and widely used to examine the selectivity of KRAS-directed therapies31,32,33; and (3) is used to study autophagy34. Strikingly, 249C exhibited increased potency against cells expressing KRAS/BRAF mutants rather than the WT genes (Fig. 5a). Moreover, potency depended on the identity of the mutant (IC50 values: WT (1.25 µM); G13D (0.07 µM); G12V (0.15 µM); G12S (0.23 µM); G12D (0.3 µM); Q61L (0.31 µM); G12C (0.44 µM); Q61R (0.55 µM); and BRAF V600E (0.11 µM)), suggesting that susceptibility to V-ATPase inhibition differs across KRAS/BRAF mutants (Fig. 5a and Supplementary Fig. 1c; G13D, P < 0.05). Treatment with BafA1 showed similar profiles except for G12C, G12D, and Q61L, whereas the HCQ analog (DC661) and the autophagy inducer (CMA) showed substantially different profiles relative to 249C, and the KRASG12C inhibitor (AMG510) specifically killed cells expressing KRASG12C. Of note, human and mouse ATP6V1H share 98.55% sequence identity; of the remaining, five amino acids are similar.
249C inhibits macropinocytosis in a KRAS-mutant-dependent manner
To investigate the mechanism of mutant selectivity of 249C, we explored how 249C impacts four main phenotypes in MEFs expressing KRAS/BRAF mutants: (1) lysosomal pH; (2) dependence on V-ATPase activity; (3) apoptosis; and (4) levels of MP.
Cancer cells are reported to have lower lysosomal pHs than untransformed cells, and stably expressing oncogenic KRASG12V in untransformed cells is sufficient to decrease lysosomal pH35. To test this in our system, we stained MEFs bearing KRAS WT and KRAS/BRAF mutants with a pH-sensitive dye and discovered that MEFs bearing KRASG13D exhibited the most acidic lysosomes (basal levels) compared with WT or other mutants (Fig. 5b). Thus the mere expression of KRASG13D versus WT (or some of the other mutants) renders the lysosomes in these cells more acidic. Since V-ATPases are responsible for acidifying lysosomes, these data indicate that KRASG13D cells could bear more biochemically active V-ATPase. As expected, treatment with 249C or BafA1 increased the pH of lysosomes of KRASG13D MEFs (Fig. 5b, Supplementary Fig. 7). To next test if these differences in lysosomal pH were caused by differential dependence on V-ATPase activity, we measured lysosome acidification in MEFs bearing WT KRAS or KRAS/BRAF mutants using FITC-Dextran as in Fig. 3f. Untreated KRASG13D MEFs accumulated the most FITC in their lysosomes, indicative of more active V-ATPase and thus more acidic lysosomes. As expected, 249C inhibited V-ATPase-dependent proton transport in most MEFs, as measured by ATP-dependent fluorescence quenching of FITC-loaded lysosomes (Fig. 5c and Supplementary Fig. 8). Strikingly, 249C inhibited fluorescence quenching most strongly in KRASG13D MEFs. V-ATPase inhibition and increased pH might subsequently halt the degradative process in lysosome-autophagic vesicle (AV) fusion. Indeed, when we analyzed these MEFs by transmission electron microscopy, we found considerably higher AV diameters in MEFs bearing KRASG13D, KRASG12V, or BRAFV600E after 249C treatment compared with KRAS WT and other mutants, probably owing to autophagic flux blockade (Fig. 5d,e). Next, as an orthogonal measure of cell death, when we assessed these MEFs by flow cytometry, we found that 249C induced apoptosis in MEFs bearing KRASG13D (P = 0.0009), KRASG12V (P = 0.002), and BRAFV600E (P = 0.001), but less so in MEFs expressing KRAS WT (Fig. 5f). Similarly, 249C significantly increased apoptosis in KRAS-mutant cancer cells as compared to KRAS WT cells (Supplementary Fig. 8g).
V-ATPase-dependent lysosome acidification is essential for autophagy and MP. We next tested if MP activity varies across different Ras/Raf-mutant cells and whether such variations correlated with sensitivity to 249C. KRASG13D cells (followed by KRASG12V cells) cells had the highest levels of basal MP (absolute values) measured by uptake of fluorescently labeled dextran (Supplementary Fig. 8a) suggestive of higher V-ATPase activity attributable to the identity of the KRAS mutant in these cells. Reciprocally, treatment with 249C caused the strongest inhibition of MP in MEFs bearing KRASG13D (94%) and BRAFV600E (90%), in agreement with the low IC50 of cells bearing these mutants (Fig. 5g,h; KRASG12V (73%)). Even though KRASG13D is not associated with the highest proliferation rate, we speculate that of all available metabolic pathways, KRASG13D cells derive a particularly high fraction of their energy from MP/autophagy and thus are particularly reliant on V-ATPase for survival.
Sensitivity to 249C varies with MAPK, IGF1R, and EGFR signaling
We next asked whether differential dependence on other signal transduction pathways in different KRAS-mutant cells contributes to the mutant specificity of 249C and whether such dependencies could be exploited for combination therapies. We first evaluated synergy between inhibitors of the MAPK signaling pathway and 249C, since concurrent inhibition of ERK1/2 and autophagy has been reported to synergistically suppress the growth of KRAS-mutant cells36. Using MEFs and cancer cell lines, we performed co-treatment with 249C and the following MAPK pathway inhibitors: AMG510 (KRASG12C); Dabrafenib (BRAFV600); Trametinib (MEK1/2); and SCH772984 (ERK1/2) (Supplementary Fig. 9a). As anticipated, KRASG12C-mutant lines were sensitive to KRASG12C-specific inhibition via AMG510, which was heightened when combined with 249C (Supplementary Fig. 9b). Also, as expected, the BRAFV600E-mutant lines were sensitive to BRAFV600 inhibition via Dabrafenib, which was heightened when combined with 249C (Supplementary Fig. 9c). Finally, Trametinib (MEKi) (Supplementary Fig. 9d) and SCH772984 (ERKi) (Supplementary Fig. 9e) alone modestly decreased proliferation in MEFs expressing KRAS mutants and selected cancer cell lines with KRAS mutations. Co-treatment with 249C led to further decreased viability, especially at increased concentrations, with effects ranging from additive to mildly synergistic in different cell lines, suggesting that 249C-treated cells may transduce varying levels of signaling via the MAPK pathway. Thus, the addition of 249C could sensitize cells bearing less sensitive KRAS mutants to MAPK pathway inhibitors, expanding the potential therapeutic relevance of 249C. For example, Panc1 cells are not particularly sensitive to Trametinib individually, but respond well when combined with 249C (Supplementary Fig. 9d).
We next also evaluated signaling through IGF-1R using an IGF-1R-specific inhibitor, AG-1024. MEFs bearing KRASG12V or KRASG13D were insensitive to AG-1024, which is likely due to a lack of signaling through this pathway (Supplementary Fig. 10a). By comparison, MEFs expressing WT KRAS and other mutants (G12C, G12D, G12S, Q61L, Q61R, and V600E) were sensitive to inhibition, indicating that they can readily signal through the insulin signaling pathway, which can promote cell proliferation (Supplementary Fig. 10b). Thus, we speculate that upon V-ATPase inhibition, cells bearing WT KRAS and other mutants could signal through an alternative pathway for survival, but cells bearing KRASG12V and KRASG13D cannot rely on the insulin pathway.
Finally, we explored the effect of EGFR inhibition using Erlotinib. Consistent with previous reports37,38,39,40, we found that EGFRi was strongest for cells expressing KRASG13D and WT in both MEFs and SW48 cells, indicating their dependence on EGFR signaling, whereas other mutants were not as responsive (Supplementary Fig. 10c,d). Since KRASG13D has been suggested to maintain dependence on EGFR signaling when compared with other KRAS alleles, we investigated the effects of 249C treatment on EGFR localization via immunofluorescence. We found that in 249C-treated KRASG13D cells, EGFR localized predominantly in a punctate manner, likely in vesicles, but in cells expressing WT KRAS and other mutants the staining pattern was diffuse (Supplementary Fig. 10).
Cells bearing KRAS WT and mutants other than KRASG12V, KRASG13D, and BRAFV600E appear to have alternative signaling capabilities that allow these cells to survive in the presence of 249C. By contrast, for the sensitive mutants, 249C appears to target a composite signaling ‘Achilles’s heel’ for the sensitive mutants.
249C inhibits tumor growth in vivo with minimal toxicity
Finally, to test whether 249C could inhibit autophagy and attenuate the growth of mutant KRAS-dependent cancers as a single agent in vivo, we evaluated 249C in a mouse xenograft model of lung cancer (NSCLC, A549). In brief, we injected mutant KRAS A549 (G12S) cells into the flanks of athymic mice to establish xenografts, randomly assigned mice to vehicle control (n = 7) or 249C-recipient groups (n = 7; 10 mg kg−1) with twice daily intraperitoneal (i.p.) injections, and recorded tumor volume and body weight over time. During the course of the study, all mice survived but tumor volume was significantly lower in 249C-treated mice (Fig. 6a; P = 2.9 × 10−7). We subsequently killed all mice and excised their tumors for further investigation. Tumor mass was lower in the 249C-treated group (Fig. 6b; P = 5.3 × 10−6), and tumors from mice treated with 249C showed upregulation of LC3-I/II, as assessed by immunoblotting, indicative of inhibition of autophagic flux (Supplementary Fig. 11a). Pharmacokinetic studies revealed that after 4 h, a safe maximum concentration of 20 µM of 249C was detected in blood (or plasma) of mice, confirming metabolic stability (Supplementary Fig. 11). Furthermore, 150 mg kg−1 of 249C was determined to be the maximum tolerated dose in mice and none of the following changed in either group: (1) body weight (Fig. 6c); (2) internal organ tissue weights; and (3) blood cell counts (Supplementary Fig. 11). This suggests a lack of toxicity from treatment; yet even a dose as low as 10 mg kg−1 is sufficient for antitumor activity.
To explore the relationship between individual KRAS mutants and 249C sensitivity in vivo, we used a colon cancer (SW48) xenograft model with a set of colon cancer cell lines that differ in KRAS mutation status but are otherwise genetically matched. Stable derivatives of the SW48 cell line were derived by introducing a KRASG12D, KRASG12V, or KRASG13D point mutation via recombinant adeno-associated viral (rAAV) gene-editing technology previously reported41. To compare KRAS mutants in a cancer relevant system, four cell lines: (SW48, SW48 KRASG12D/+, SW48 KRASG12V/+, and SW48 KRASG13D/+) were injected into athymic mice and 14 days of treatment (10 mg kg−1 249C or vehicle control, i.p.; n = 5 for all arms) was started when the average tumor volume reached approximately 100–200 mm3 following random segregation into groups. As before, KRASG13D tumors were most sensitive to 249C both in vitro and in vivo (Fig. 6d,e). On day 21, the antitumor activity of 249C, presented as treatment/control, was 18% for SW48 KRASG13D/+ (P = 9.6 × 10−6), 55% for SW48 KRASG12V/+ (P = 0.005), 75% for SW48 KRASG12D/+ (P = 0.79), and ~100% for parental SW48 (P = 0.97), and the corresponding tumor growth inhibition was 82%, 45%, 25%, and 0%, respectively (Fig. 6e). Changes in tumor volume are presented for animals bearing isogenic SW48 cells: parental/KRAS WT (Fig. 6f), KRASG12D/+ (Fig. 6g), KRASG12V/+ (Fig. 6h), KRASG13D/+ (Fig. 6i). Note that SW48 tumors with different KRAS variants show different basal growth rates, as reported previously (https://tiny.cc/horizondiscovery). Nonetheless, these data indicate that replacement of a single WT allele with a mono-allelic KRASG13D/+ mutation is sufficient to significantly and drastically alter sensitivity to 249C and result in tumor regression in a xenograft model. Further, no significant changes in body weight of the mice were observed during the study (Fig. 6j).
To further evaluate the potential for secondary effects and adverse events in human cells, which is key to preventing phase I failures, we assessed the effect of 249C on hERG. The autophagy inhibitors CQ and HCQ are known to induce a high risk of cardiac electrocardiogram long QT syndrome (LQTS) through inhibition of hERG. In a side-by-side comparison, 249C showed no inhibition of hERG at 30 µM (0% inhibition) (Fig. 6k), whereas HCQ (~60% inhibition), BafA1 (35% inhibition), and DC661 (100% inhibition) all showed substantial inhibition (Supplementary Fig. 11), indicating less risk and a sizeable therapeutic window for 249C18,42. Finally, we used BioMAP Phenotypic Safety and Toxicology profiling43 to screen for adverse effects. In human primary cells relevant to physiology, BafA1 showed toxicity in two categories but 249C was determined to have no adverse events across over 100 biomarker readouts, an endorsement of its candidacy to progress into the clinic (Fig. 6l). Collectively, these data indicate that 249C is efficacious in inhibiting tumor growth with minimal side effects (Supplementary Fig. 12) in our investigations thus far, establishing in vivo proof-of-concept for targeting autophagy/MP/lysosomal acidification with 249C in mutant KRAS cancers. Interestingly, while profiling for off-target kinase hits (Supplementary Fig. 12), we found that 249C bound to a tyrosine kinase called DDR1, which has been suggested as a therapeutic target for pancreatic ductal adenocarcinoma44,45. Very recently DDR1 was reported to get activated to compensate for loss of KRAS signaling46 implying that part of 249C’s activity could result from synergistic inhibition of DDR1 and the Ras pathway.
In this study, we describe the discovery, target identification, and mechanism of action of 249C, a lead compound for treatment of KRASG13D and KRASG12V mutant cancers. 249C kills cancer cells by targeting a druggable hotspot within the V-ATPase subunit ATP6V1H and thereby inhibiting biochemical activity, lysosomal acidification, and MP. Although 249C inhibits global cellular functions that depend on V-ATPase, we uncover that cells expressing KRASG13D and KRASG12V are particularly dependent on V-ATPase, creating a synthetic vulnerability exploited by 249C. With promising in vivo activity, low toxicity, and favorable pharmacokinetics, 249C is an ideal candidate for the development of patient-tailored drugs to treat cancers driven by KRASG13D and KRASG12V. In addition, the synergy of 249C with inhibitors of MAPK signaling in some KRAS-mutant backgrounds provides an opportunity to explore new combination therapies.
Identification of the relevant mechanisms of action and cellular targets has been a critical barrier to exploiting small molecules with therapeutic potential from being developed into approved drugs. Here three independent assays—chemo-proteomics, comparative profiling, and CRISPR-based chemical–genetic screens—pointed to V-ATPase as the molecular target of 249C, which we then confirmed using cell-based, biochemical and biophysical techniques. Compellingly, 249C directly bound the V-ATPase subunit ATP6V1H and inhibited proton pumping and ATP hydrolysis. ATP6V1H is part of the stator complex47 and is suggested to regulate bridging of rotor (Vo) and stator (V1) complexes and physically prevent ATP-driven rotation. 249C may alter V1–Vo bridging by binding to ATP6V1H. The physical coupling model proposed here is in alignment with previous reports of blocking V-ATPase using BafA1 and/or ConcanamycinA (ConA)48,49.
A striking feature of 249C is the selectivity for KRAS mutants, which we established using three separate systems: (1) a panel of 53 established cancer cell lines (Fig. 1d,c); (2) MEFs expressing specific KRAS alleles (Fig. 5a) and; (3) four SW48 colon cancer cells/xenografts that differ in KRAS mutation status but are otherwise genetically matched (isogenic) (in vitro (Fig. 6d,e) and in vivo (Fig. 6f–i)). Together, these three systems provide consistent and complementary evidence for the selectivity of 249C for KRASG13D/G12V and highlight a fundamental vulnerability of KRASG13D/G12V mutant cells that has not been reported in this context before.
This selectivity was not explained by differences in V-ATPase localization with different KRAS mutants (Supplementary Fig. 8), as might have been expected from previous reports10. Instead, this selectivity appears to arise from multiple converging factors. First, KRASG13D and KRASG12V cells have particularly high basal levels of MP and low lysosomal pH, suggesting that these cells derive a larger fraction of their energy from degradative processes such as autophagy and MP and thus are particularly sensitive to V-ATPase inhibition. Second, cells expressing KRAS alleles other than KRASG13D or KRASG12V appear to derive proliferative signaling from additional pathways including the IGF-1R pathway, which may allow these cells to better withstand the presence of 249C. Finally, KRASG13D cells rely more strongly on EGFR signaling for growth, and 249C treatment disrupts EGFR localization, removing this pro-proliferative pathway.
Additional factors may contribute to the selectivity of 249C. For example, V-ATPase interacts with the Ragulator, which regulates mTORC1 activation and autophagy50. Thus, in addition to inhibiting autophagy and MP, 249C may also disrupt compensatory mTORC1 signaling that normally occurs in these situations. Cells expressing 249C-insensitive mutants like KRASG12D and KRASG12C may also have alternative metabolic signaling pathways that promote survival, whereas cells expressing KRASG13D, which derive most of their energy from lysosomal degradative processes, do not. In addition, differential changes in cellular pH from 249C across different KRAS mutants may affect downstream transcription factors differentially, which may further contribute to the inability of cells expressing KRASG13D or KRASG12V to engage other survival-promoting signal transduction pathways. It may be interesting to deconvolve how cells expressing KRASG12V or BRAFV600E, which are sensitive to the changes in pH caused by 249C, differ in their signaling requirements from cells expressing KRASG13D.
249C differs from known V-ATPase/autophagy inhibitors in both mechanism and pharmacology. Antifungal V-ATPase inhibitors, such as BafA1 and ConA, target ATP6VoC, and other autophagy inhibitors, such as the antimalarials HCQ and CQ, act through a different target, PPT1. While both antifungals and antimalarials have potential therapeutic activity, their clinical utility is hampered by toxicity. BafA1 and ConA have been relegated as laboratory compounds, and HCQ and CQ show substantial off-target activity on hERG. By contrast, 249C shows minimal cytotoxicity when compared with BafA1 and minimal inhibition of hERG while maintaining nanomolar potency. Further, BafA1 was first isolated in 1983 and to date no clinical trials have been launched. By contrast, 249C has already entered phase I clinical trials.
Available evidence, such as differences in signaling in cells expressing different Ras mutants36,51,52,53, increasingly suggests that there may not be a single anti-Ras therapeutic approach for all Ras-mutant cancers; instead, mutation-specific therapeutic strategies must be deployed for different Ras mutations. While the emergence of the covalent KRASG12C inhibitor being developed by Amgen has garnered enthusiasm, there remains a pressing need for therapeutic strategies targeted at cancers driven by other KRAS mutations, especially given the dismal outcomes for patients with cancers driven by mutations in KRAS codon 1354,55. The G13D and G12V mutations account for 13% and 24% of KRAS mutations across all human cancers, respectively. Thus, 249C could become a therapeutic for ~37% of all KRAS malignancies. Beyond KRAS-driven cancers, autophagy dysregulation has been linked to a wide range of diseases including metabolic disorders, aging, and neurodegenerative diseases56, motivating investigation of 249C as a drug candidate for other treatment applications.
Cell lines, reagents and proliferation assays
Most cells were obtained from the American Type Culture Collection (ATCC) and were cultured in ATCC-recommended medium supplemented with 10% fetal bovine serum and 2% antibiotics (penicillin–streptomycin). MDA-MB-231 CRISPRi cells were a gift from H. Goodarzi at UCSF (parent cells from ATCC). MEFs expressing human KRAS/BRAF were from the Frederick National Laboratory for Cancer Research, the National Cancer Institute. SW48 cells were obtained from Horizon Disc.
ATCC cell lines: H1944 (CRL-5907), A427 (HTB-53), A549 (CRM-CCL-185), SW1271 (CRL-2177), H292 (CRL-1848), H358 (CRL-5807), H1299 (CRL-5803), H460 (HTB-177), H1703 (CRL-5889), H2170 (CRL-5928), H2452 (CRL-5946), NCI H82 (HTB-175), Calu-3 (HTB-55), H838 (CRL-5844), H1975 (CRL-5908), H1650 (CRL-5883), H28 (CRL-5820), CFPAC-1 (CRL-1918), Capan-1 (HTB-79), Panc 02-13 (CRL-2554), ASPC1 (CRL-1682), Panc1 (CRL-1469), BxPC3 (CRL-1687), MDA-MB-231 (HTB-26), Sk-Br-3 (HTB-30), MCF7 (HTB-22), MDA-MB-436 (HTB-130), HCT116 (CRL-247EMT), Colo201 (CCL-224), N87 (CRL-5822), RKO (CRL-2577), SW48 (CCL-231), CaCO2 (HTB-37), SkMel5 (HTB-70), SKMEL30 (HTB-73), A375 (CRL-1619), A2058 (CRL-11147), DUI145 (HTB-81), PC3 (CRL-3471), U251 (HTB-17), U87 (HTB-14), HL-60 (CCL-240), and HepG2 (HB-8065). LOX IMVI, MelJuso, BT474 and HCC1937 cells were a gift from the Oritz–Urda lab at UCSF originally sourced from ATCC or as gifts. ES2, OVCAR8, HEYA8, and HEY cells were a gift from the Smith–McCune lab at UCSF originally sourced from ATCC.
BafA1, chloroquine, hydroxychloroquine, and DC661 were from Medchemexpress. Rapamycin was from Selleckchem and Torin was from Fisher. Erlotinib, AG-1024, and MAPK inhibitors were from Selleckchem. DNA transfection into cells was performed with TransIT-LT1 Transfection Reagent (Mirus Bio) by following the manufacturer’s recommendations.
Logarithmically growing cells were plated in antibiotic-free medium supplemented with 2% fetal bovine serum at a density of 5,000 cells per well. The next day, cells were treated in triplicates with increasing doses of in-house small molecule inhibitor drugs and a DMSO vehicle control for 3 days and subsequently assessed for cell viability by measurement of ATP with CellTiter-Glo Luminescent Cell Viability Assay (Promega). Signal intensity was read on a Glomax 96 Microplate Luminometer (Promega) with GLOMAX software (v.1.9.3) and percent cell survival was calculated on the basis of the reading of vehicle control cells set as 100% using Excel (2012) and GraphPad Prism (v.6). Each compound was tested a minimum of 2 times with n = 3. All cell proliferation assays (cancer cell lines or MEFs) were done in a similar way.
Chemical synthesis of 249C
249C and other small molecules (~300) were synthesized at UCSF and are covered by patent number PCT/US2017/039806 filed with the United States Patent and Trademark Office. In brief, small molecules to represent diversity in the compound library were synthesized by the synthesis method described in Supplementary Fig. 1. The synthetic route is designed to investigate functional group tolerability around the dihydro-pyrazole pharmacophore; simultaneous optimization was carried out for potency and physico-chemical properties. Analytical evaluation of compounds synthesized is presented in Supplementary Table 3.
Structure–activity relationship investigation
Compounds synthesized in this series are potent in multiple cancer cell lines. Most calculated properties meet drug-likeness criteria. Initial compounds are hydrophobic in nature, with relatively high logP and have low polar surface are. The lipophilicity of the compounds can be reduced via the introduction of polar side chains resulting in analogs with more favorable physico-chemical properties, that is drug likeness.
Mutation analysis in CCLE cell lines
The CCLE (https://sites.broadinstitute.org/ccle) was used to download somatic mutation data for over 1,000 cancer cell lines ((CCLE_DepMap_18q3_maf_20180718.txt) from CCLE at https://depmap.org/portal/download/ (2018))57. We filtered the data to retain only mutations (single nucleotide variants (SNVs) and indels) detected in 53 cell lines that were included in our viability assays (Fig. 1). Non-coding variants, silent mutations and in-frame insertions were removed, as these were not considered functional. We retained in-frame deletions as these may remove essential binding sites. The Combined Annotation Dependent Database (https://cadd.gs.washington.edu/) (v.1.6) was used to assess the predicted functional importance of SNVs, and we retained only variants with CADD Phred scores ≥15 (ref.58). Including these predicted damaging SNVs as well as indels (minus in-frame insertions), we determined the number of times each gene was mutated across the 53 cell lines. Genes with multiple mutations in the same cell line were only counted once. Multiple Wilcoxon rank sum tests were carried out, using the ‘lapply’ looping function in the R statistical environment, to test for association between presence/absence of mutations in each gene and IC50 values in each cell line. A Wilcoxon rank sum test was also carried out for Ras/Raf mutations, which included combined mutation data for genes KRAS, NRAS, HRAS, NF1, and BRAF.
Mass spectrometric whole proteome analysis
After treatment with small molecules, A549 cells were trypsinized, lysed in 8 M Urea, 1% SDS, 50 mM Tris pH 8.5, protease (Complete) and phosphatase (PhosStop) inhibitors, and samples were sent to the Thermo Fisher Center for Multiplexed Proteomics at Harvard to be processed as described3. In brief, extracts were purified by trichloroacetic acid precipitation, followed by labeling with tandem mass tags (TMT) reagents (Thermo Fisher), and subsequently desalted by StageTips before LC–MS/MS analysis. Data were collected using an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific) coupled with a Proxeon EASY-nLC 1200 LC pump (Thermo Fisher Scientific). For LC–MS/MS analysis, an MS3-based TMT method was used as previously described4. For mass spectrometry data processing and spectra assignment, an MS2 spectra assignment false discovery rate of less than 1% was implemented by applying the target–decoy database search strategy. To quantify TMT reporter ion intensity, a 0.003 m/z window centered on the theoretical m/z value of each reporter ion was monitored for ions, and the maximum intensity of the signal to the theoretical m/z value was recorded. Reporter ion intensities were normalized by multiplication with the ion accumulation time for each MS2 or MS3 spectrum and adjusted on the basis of the overlap of isotopic envelopes of all reporter ions. After the reporter ion signal was extracted, the isotopic impurities of the TMT reagent were corrected using the values specified by the manufacturer. Total signal-to-noise values for all peptides were summed for each TMT channel and all values were adjusted to account for variance and a total minimum signal-to-noise value of 200 was implemented59,60.
GSEA software (version 4.2.3) and Molecular Signature Database (MSigDB, v.7.4) from https://www.gsea-msigdb.org/gsea/index.jsp was used to query the top eight proteins from the Whole Proteome Analysis for Gene Ontology gene sets in NCI-60 cell lines61,62.
BioMAP and benchmarking
BioMAP assays comprising 12 human primary-cell-based systems (co-cultured cells simulating 12 different disease contexts) were treated in the presence or absence of test agents (known drugs and our 249C compound) as described63. All cells were from a pool of multiple donors (n = 2–6) and collected in accordance with appropriate regulatory protocols. Direct enzyme-linked immunosorbent assay was used to measure selected biomarker level readouts (adhesion receptors, cytokines, enzymes etc.) and activity profiles (normalized datasets) were generated for each test agent. The resulting profiles from biomarker readouts were compared or ‘benchmarked’ to a known 4,000-molecule database by statistical methods to identify similarities and mechanistic insights described previously64 to provide Pearson’s and Z-scores.
Protein extracts were prepared from drug-treated cells by lysis on ice for 20 min in M-PER Mammalian Protein Extraction Reagent (ThermoScientific) supplemented with protease and phosphatase inhibitors (Roche). Lysates were centrifuged at 16,000 r.p.m. for 20 min at 4 °C and the supernatants were assessed for concentration via a Bradford assay. Protein extracts (25–50 µg) were used for gel electrophoresis followed by immunoblotting onto PVDF membranes. Blots were probed with the following primary antibodies: GAPDH (Thermo Fisher, 1:20,000), Ras (1:1000), SQSTM1 (1:1000) and LC3 (1: 1000) (Cell Signaling).
A549 cells attached to ACLAR film65 were treated for 24 h with 249C (250 nM) or BafA1 (10 nM) or DMSO as a control. The cells on film were then fixed by immersion in 2% glutaradehyde in 0.08 M Na-cacodylate buffer, pH 7.3 containing 2 mM CaCl2 that had been pre-warmed to 37 °C for 1 h with gentle agitation during which the fixative cooled to room temperature. The samples were then rinsed with 0.1 M Na-cacodylate buffer at room temperature, post-fixed with 1% OsO4 containing 1.5% potassium ferrocyanide in 0.1 M Na-cacodylate buffer for 45 min on ice, rinsed with water, en bloc stained with 3.5% uranyl acetate in water for 1 h at room temperature, dehydrated in ethanol followed by propylene oxide, and embedded in Eponate 12 resin (Ted Pell). Thin sections were cut with a Leica UCT ultramicrotome using a Diatome diamond knife and picked up on Pioloform films on slot grids. The grids were then post-stained with 1% uranyl acetate followed by Sato’s lead66. Sections were imaged with an FEI T12 TEM equipped with a Gatan U895 4k × 4k camera at 120 kV (software: DigitalMicrograph (v.3.4.3) and SerialEM (v.3.8.6)). Quantification was performed with custom Python image analysis software. MEFs were processes in a similar way.
Cells were treated with the respective molecules in two-well chambered coverglass slides at a final concentration of 1 µM for 18 h followed by addition of lysosome-specific dyes. LysoTracker Red DND-99 and pH-sensitive LysoSensor Yellow/Blue DND-160 purchased from Thermo Fisher were added to the cells for 1 h at 37 °C along with Hoescht to stain DNA according to the manufacturer’s recommendations. Live cell images were captured at 20× with a Zeiss spinning disc confocal and TIRF Fura microscope at UCSF’s Laboratory for Cell Analysis.
Genome-scale screens using MDA-MB-231 CRISPRi cells (a gift from H. Goodarzi at UCSF) were performed in a similar manner as previously described25,67. For genome-wide knockdown, CRISPRi-v2 sgRNA libraries25 (targeting 18,905 genes and marked with blue fluorescent protein (BFP)) were transduced into MDA-MB-231 CRISPRi cells at a multiplicity of infection <1 (percentage of transduced cells 2 days after transduction: ~30%). Cells were maintained in DMEM in 20× T-182.5 flasks for the course of the screen. After transduction, the cells were selected with puromycin for 2 days, at which point transduced cells accounted for 90% of the population assessed as the fraction of BFP-positive cells by flow cytometry. After 1 day of recovery without puromycin, samples at time-point t0 with a minimum established coverage of >1,000 cells per sgRNA were harvested and the remaining cells were split into two populations for untreated growth (DMSO control) and 249C-treated growth. The cells were maintained in T-182.5 flasks at an average coverage of greater than 1,000 cells per sgRNA for the duration of the screen. For 249C treatment, 525 nM 249C was added to the cells at day 0 and day 5 and removed the following day. Cells were harvested on days 19 (DMSO) and 21 (249C) (6.1 doubling differences between treated and untreated populations). Genomic DNA was isolated and the sgRNA-encoding region was amplified and processed for next-generation sequencing on an Illumina HiSeq 4000 (HiSeq Software Suite v.3.4.0) as described previously68. Sequencing reads were aligned to the CRISPRi library sequences, counted, and quantified using the Python-based ScreenProcessing pipeline available at https://github.com/mhorlbeck/ScreenProcessing25. Generation of negative control genes and calculation of phenotypes and Mann–Whitney P values was performed as described previously67. Sensitivity phenotypes (ρ) were calculated by calculating the log2 change in enrichment of an sgRNA in the treated and untreated samples, subtracting the equivalent median value for all non-targeting sgRNAs, and dividing by the number of population doubling differences between the treated and untreated populations67,68,69. Similarly, untreated growth phenotypes (γ) were calculated from the untreated and t0 samples, dividing by the total number of doublings of the untreated population. Phenotypes from sgRNAs targeting the same gene were collapsed into a single sensitivity phenotype for each gene using the average of the top three scoring sgRNAs (by absolute value) and assigned a P value using the Mann–Whitney test of all sgRNAs targeting the same gene in comparison to the non-targeting controls. For genes with multiple independent transcription start sites (TSSs) targeted by the sgRNA libraries, phenotypes and P values were calculated independently for each TSS and then collapsed to a single score by selecting the TSS with the lowest Mann–Whitney P value. Read counts and phenotypes for individual sgRNAs are available in Supplementary Table 3. Gene-level phenotypes are available in Supplementary Table 4. All additional analyses were performed using Python 3.7 using a combination of Numpy (v.1.15.4), Pandas (v.0.25.3), and Scipy (v.1.4.1).
Individual CRISPR re-tests
Oligonucleotides encoding sgRNAs targeting the selected genes for individual re-tests were acquired from IDT and cloned into the pCRISPRia-v2 vector (Addgene #84832) backbone followed by DNA transformation, extraction, and sequencing to verify accuracy. The sgRNA expression plasmids were packaged into lentivirus and transduced into MDA-MB231 CRISPRi cells. 5 days after transduction, cells were left untreated (DMSO) or treated with 525 nM of 249C or 30 nM BafA1 for 24 h and assessed for BFP using an LSR II flow cytometer at UCSF’s Flow Cytometry Core facility. Enrichment of sgRNA-expressing cells was measured as the enrichment of BFP-positive cells [e = fraction(BFP+) / (1 − fraction(BFP+)], calculated relative to the DMSO-treated control cells. Experiments were performed in triplicates.
Cytoplasmic and membrane extracts from A549 cells left untreated (DMSO) or treated with 249C (20 µM) or BafA1 (1.5 µM) for 2 h were prepared using a Subcellular Protein Fractionation Kit by following the manufacturer’s recommendations (Fisher). Fractions were subjected to Western blotting as described above and blotted for ATP6V1B2 (a gift from D. Brown at Harvard University, 1:1,000) and ATP6VoD (Abcam, 1:1,000) antibodies.
Mammalian V-ATPase complex: the entire V-ATPase complex (800 kDa) was purified as described previously from porcine kidneys23 using the detergent glycol-diosgenin.
ATP6V1H (H subunit): Addgene plasmid #14658 was expressed in Rosetta2(DE3)pLysS cells and induced with isopropyl β-d-1-thiogalactopyranoside. Harvested lysates were loaded onto a GSTrap column and bound protein eluted with reduced glutathione. Thrombin was added to cleave the GST tag followed by loading onto a BioRad SEC650 size-exclusion column (Supplementary Fig. 5).
Bio-layer interferometry studies for binding affinity determination
Biotinylated 249C was diluted in phosphate-buffered saline (PBS) + 0.05% Tween + 0.2% bovine serum albumin (BSA), pH 7.4 and loaded onto Octet Streptavidin (SA) Biosensors (ForteBio) by following the manufacturer’s recommendations on an Octet RED384 machine (ForteBio, PALL Octet System). Biotin–249C immobilization was checked via Octet Software before introduction of the entire V-ATPase complex and the individual H subunit protein diluted in blocking buffer (1× PBS + 0.05% Tween 20 + 0.2 % BSA + 10 µM Biotin) to eliminate non-specific binding. A reference sensor was subtracted from the signal to blank the system. The ForteBio Octet Software (v.12) on the BLI system was used to calculate the Kd values. Graphical output from the Octet Software of representative data from two independent experiments is presented.
Proton pumping assay in mammalian cells
HEK293T cells were either left untreated (DMSO control) or treated with 10.5 µM 249C or 0.5 µM BafA1 for 1 h at 37 °C. This assay was performed as described previously70.
HEK293T cells (4 × 106) were plated in 10-cm plates. The next day, the medium was replaced with medium containing 2.2 mg ml−1 FITC-Dextran (Millipore Sigma FD40), to allow uptake of the dye by endocytosis. The following day, FITC-Dextran-containing medium was replaced with regular unlabeled medium that permits all dye to progress to the lysosomal compartment71. Cells were either left untreated (DMSO control) or treated with 15 µM 249C or 0.5 µM BafA1 for 1 h at 37 °C. After treatment, cells were placed on ice, rinsed with ice-cold PBS and collected by scraping into fractionation buffer (125 mM KCl, 1 mM ethylenediaminetetraacetic acid, 50 mM sucrose, 20 mM HEPES, 1 mM PMSF, and Halt protease inhibitors). Cells were collected by centrifugation at 1,200g for 5 min, resuspended in fractionation buffer and lysed by passing through a 27-gauge needle 10 times. Cell lysates were cleared of nuclei and intact cells by centrifugation at 2,000g for 10 min and the resulting supernatant was then centrifuged again for 15 min at 16,100g to sediment the FITC-Dextran-containing lysosomes72,73. The resulting pellets were resuspended in 100 µl of fractionation buffer and subjected to a Bradford assay to assess protein concentration74. To measure molecule-dependent proton pumping, 20 µg protein was added to fractionation buffer pre-warmed to 37 °C. Sample fluorescence was excited at 490 nm, and emission fluorescence at 520 nm was measured continuously in BioTek plate reader (Synergy) using Gen5 software (v.3.05). After initial fluorescence stabilization, 1 mM ATP and 2 mM MgCl2 were added to initiate ATPase activity and proton pumping into the lysosomes, which causes quenching of FITC in the lysosomal lumen. The V-ATPase dependence of quenching was verified by adding 249C and BafA1, which inhibit fluorescence quenching. Representative of five individual experiments with n = 2 for each is presented. MEFs were processes in a similar way.
Proton pumping assay in yeast
Spheroplasts from wild-type yeast (a gift from the Walter lab at UCSF) were left untreated (DMSO control) or treated with 29 µM 249C, or 1 µM BafA1 for 2.5 h with shaking at 30 °C and vacuolar membrane vesicles were isolated as previously described75. Proton transport was measured using ATP-dependent quenching of 9-amino-6-chloro-2-methoxy-acridine (Acridine Orange, Thermo Fisher) fluorescence quenching for isolated vacuoles as previously described76.
Biochemical V-ATPase assay
V-ATPase was purified as described previously from yeast77 using the detergent dodecylmaltoside and porcine kidney78 using the detergent glycol-diosgenin. ATPase assays with purified V-ATPase were performed as described previously. 249C (mammalian: 0.1, 1, 10, 100 µM; and yeast: 5, 10 and 100 µM) or BafA1 (1 µM) were added to the reaction and compared with a negative control to which only DMSO was added. Enzyme-coupled ATPase biochemical activity assays were performed in a 96-well plate as described previously79.
Three-dimensional molecular modeling, binding pocket detection and docking
A preliminary sequence alignment between the yeast (Uniprot P41807) and human (Uniprot Q9UI12) sequences, which share 27% homology/45% similarity, was generated with EMBOSS Stretcher on EMBL-EBI webserver. The alignment was then manually adjusted to avoid alignment gaps within secondary structural elements of the human protein, as predicted by the DISSPRED webserver (https://comp.chem.nottingham.ac.uk/disspred/). Next the human V-type proton ATPase, encoded by ATP6V1H was modeled on the basis of the yeast homolog structure available in the PDB database (PDB ID 1HO8) using the MODELLER80 software v.9.19. The best model on the basis of the DOPE81 score was used for subsequent virtual screening.
The Small Molecule Drug Discovery Suite 2019-182 (Schrödinger) was used for binding pocket detection and virtual screening. The protein was prepared in a ready-to-dock-format with the Protein Preparation Wizard workflow. The hydroxyl group orientations and protonation states were assigned at pH 7 using PROPKA in Epik v.4.7. The SiteMap algorithm was used to screen for potential binding pockets within the entire surface of the protein pocket, using standard parameters. Small molecules were drawn with Maestro v.11.9, and their protonation states and tautomers assigned at pH 7.0 ± 2 in Epik v.4.7. Conformers were assigned using LigPrep v.4.9 with the OPLS3e force field. Next, Glide v.8.2 was used to generate a grid box for docking the small molecules into each detected binding site. Before docking, a 37 × 37 × 37 Å (ref. 82) docking grid was erected. The Glide XP83 scoring function was employed, and the strain energy was included in the final docking score, a maximum of ten docking poses was set for each ligand. Planarity of aromatic groups was enforced. Final binding poses were picked on the basis of a combination of visual inspection, chemical intuition, and score.
Macropinosome visualization and quantification
Macropinosome visualization was essentially done as previously described84. Cells were plated in black, clear bottom CellCarrier-96 Ultra Microplates (Perkin Elmer 6055302) and serum-starved for 24 h. The next day, they were treated with 2 µM 249C (1255) for 18 h followed by addition of 70 kDa tetramethylrhodamine (TMR)–Dextran (Invitrogen D1818) and Hoescht for 2 h at 37 °C before 3× cold PBS washes and fixation in 4% paraformaldehyde for 30 min. Following 3× cold PBS washes, mounting medium (Vector Laboratories) was added and images were captured using an Operetta CLS fluorescent microscope for High-Content Analysis (Perkin Elmer). Image analysis and quantification was performed with a spot segmentation algorithm using the Perkin Elmer Harmony PhenoLOGIC 3.2 software. In brief, the images were segmented and individual nuclei (DAPI stained) and cell bodies (by digital phase contrast imaging) were identified. The TMR signal was used for the spot segmentation analysis block. Number of spots, area, and TMR fluorescence intensity were calculated for each cell region, and displayed as mean values per cell. Data are presented as spots − number of objects per nucleus. For each cell line, five DMSO control and five 249C-treated wells were imaged and a minimum of three images were captured per well. Data representing three independent experiments are presented.
Cells were seeded onto glass coverslips and subsequently serum-starved for 24 h. After serum starvation, cells were treated with 500 nM 249C for ~15 h and fixed with 4% formaldehyde for 30 min at room temperature. The following sequential steps took place at room temperature: cells were washed twice with PBS, permeabilized (0.1% Tween in PBS) for 10 min, and blocked (5% bovine serum in PBS) for 30 min. The following primary antibodies were used: ATP6V1A (Abnova, H00000523-M02 1:250 dilution), ATP6V1B2 (Thermo Fisher, PA552518, 1:50 dilution), ATP6V1H (PA552518, Fisher, 1:50 dilution), EGFR (4267P, Cell Signaling, 1:100 dilution). Corresponding AlexaFluor-488 secondary antibodies were used at 1:500 dilution (Thermo Fisher, mouse: A11001, rabbit: A11008). Cells were mounted onto slides using Mounting Media (Vector Laboratories) containing DAPI. A minimum of three images per well were captured using a spinning disk confocal microscope (Zeiss with ZEN blue software (2012)). Data representing two independent experiments are presented.
Plasma membrane fractionation
The plasma membrane protein extraction kit (BioVision, K268-50) was used to separate the plasma membrane fraction from other cellular membranes according to the manufacturer’s recommended protocol. Data representing two independent experiments are presented.
Cells were treated with 100 nM 249C for 48 h and following two washes with ice-cold PBS, they were stained using a FITC Annexin V/PI Apoptosis Detection kit (BD Pharmingen) following the manufacturer’s recommendations. Samples were analyzed by flow cytometry on a BD LSR II flow cytometer with FACSDiva software (v.8.0). Gating strategy: preliminary forward scatter/side scatter (FSC/SSC) gates were applied followed by removal of doublets (FSC-H versus FSC-A and also SSC-H versus SSC-A). Next, single-color controls (FITC only or PI only) and double-positive controls (FITC+ and PI+) were run to set gates for the double-positive population. Finally, untreated and 249C-treated samples were run keeping these gates constant for all.
Animal studies, plasma concentration, body weight
A549 cells (5 × 106) were subcutaneously injected into the lower left flank of 6–8-week old athymic nude female mice (Charles River strain 088, homozygous NU/NU). To assess establishment of tumors, mice were examined 5 days after inoculation and were randomly segregated into two treatment groups (vehicle control, 10 mg kg−1 249C i.p., n = 7 per group) and were treated twice daily with the drug in N-methyl-2-pyrrolidone (NPM) or vehicle control (5% weight aqueous solution of carboxymethyl cellulose). Tumor volumes were calculated using the formula (length × width × height)/2 with recorded caliper measurements. Survival curves and statistical analyses were performed using Excel. Animals were maintained on a standard chow diet; 12 h/12 h dark/light cycle and in group housing in HEPA-filtered cages. All animal procedures were performed under IACUC-approved protocols and guidelines. Animals from both groups were weighed every 3 days. Plasma concentration was measured from blood collected using standard and established protocols.
For cell line xenografts, 6 × 106 SW48 cells were subcutaneously injected into the flanks of 6–8-week old athymic nude female mice (Taconic model NCRNU-F, CrTac:NCR-Foxn1 sp/sp) and tumor growth was monitored with caliper measurement. Once the mean tumor reached a certain volume (between 100–200 mm3), animals were randomized into different treatment groups of five animals per group. 249C for in vivo use was formulated in 0.5% methylcellulose and dosed i.p. (10 mg kg−1, for 2 weeks). Antitumor activity was determined by calculating the treatment over control tumor volume ratio at the end of the study. Either at the end of each study, of if any animal reached a humane end point, animals were killed with CO2 followed by cervical dislocation or another secondary method of killing approved by the IACUC protocol. Mouse manipulations were performed in accordance with UCSF’s IACUC protocol AN179973. Mice were housed in the UCSF Animal Care Facility LARC at the Helen Diller Family Cancer Center at UCSF Mission Bay. Mice were housed in a specific-pathogen-free individual suite. They are housed up to five per cage in ventilator cages, with ad libitum food and water on a 12-h light cycle and controlled temperature and humidity conditions (67–74 °F and 30–70%).
Ex vivo analysis
At the end point, tumors were harvested, weighed and minced up. Total protein from frozen tissues was prepared with T-PER (Thermo Fisher Scientific) supplemented with protease and phophatase inhibitors (Roche). Equal quantities of proteins were combined with 5× protein loading buffer and separated by SDS-PAGE followed by PVDF membrane transfer. Membranes were blocked with 5% milk followed by incubation with LC3 (Cell Signaling) and GAPDH (Thermo Fisher) antibodies. Blots were developed with ECL Reagents (Pierce).
Maximum tolerated dose Studies
Maximum tolerated dose studies were outsourced and performed at a local contract research organization (CRO) on the basis of their established protocols.
The fluorescence polarization-based assay to characterize the affinity of 249C for hERG responsible for electrical activity of the heart was performed as previously described85.
The BioMAP Toxicity Screening Profile was performed as described previously86.
The KINOMEscan Profiling Service consisting of 468 human kinases was used to assess 249C inhibitor activity to check for off-target effects as previously described87.
Lactate dehydrogenase assay
Cells were left untreated (DMSO) or treated with indicated concentrations of 249C, 68, 148D, and 226 C for 48 h. The lactate dehydrogenase (LDH) cytotoxicity assays were performed using a kit (Pierce 88953) and following the manufacturer’s recommendations.
All experiments were repeated a minimum of two times with at least duplicate or triplicate samples and measurements were taken from distinct samples. Values including controls are expressed as the mean ±s.d. or mean ± s.e.m. Two-tailed paired Student’s t-test was used to test for differences between two groups unless indicated otherwise. Differences with a P ≤ 0.05 were considered as statistically significant.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Datasets that support the findings of this study are available in the Source Data, Supplementary Figs 1–19 and Supplementary Tables 1–6. Mutation analysis in CCLE cell lines: the Combined Annotation Dependent Database (https://cadd.gs.washington.edu/) was used to assess the predicted functional importance of SNVs (v.1.6). Cell line mutation data (CCLE_DepMap_18q3_maf_20180718.txt) were downloaded from the CCLE at https://depmap.org/portal/download/ (2018). Mass spectrometry Gene Ontology gene sets: GSEA (v.4.2.3) software and Molecular Signature Database (MSigDB v.7.4) from https://www.gsea-msigdb.org/gsea/index.jsp was used to query the top proteins from the Whole Proteome Analysis for Gene Ontology gene sets in NCI-60 cell lines. CRISPR screen: sgRNA read counts and phenotypes for all pooled screens are provided as supplementary tables. All other data will be made available by the corresponding author upon request. Source data are provided with this paper.
Simanshu, D. K., Nissley, D. V. & McCormick, F. RAS proteins and their regulators in human disease. Cell 170, 17–33 (2017).
Cox, A. D., Fesik, S. W., Kimmelman, A. C., Luo, J. & Der, C. J. Drugging the undruggable Ras: mission possible? Nat. Rev. Drug Discov. 11, 828–851 (2014).
Waters, A. M. & Der, C. J. KRAS: the critical driver and therapeutic target for pancreatic cancer. Cold Spring Harb. Perspect. Med. 8, a031435 (2018).
Hobbs, A. G., Der, C. J. & Rossman, K. L. RAS isoforms and mutations in cancer at a glance. J. Cell Sci. 129, 1287–1292 (2016).
Choi, A. M., Ryter, S. W. & Levine, B. Autophagy in human health and disease. N. Engl. J. Med. 368, 651–662 (2013).
Guo, J. Y. et al. Autophagy provides metabolic substrates to maintain energy charge and nucleotide pools in Ras-driven lung cancer cells. Genes Dev. 30, 1704–1717 (2016).
Efeyan, A., Comb, W. C. & Sabatini, D. M. Nutrient-sensing mechanisms and pathways. Nature 517, 302–310 (2015).
Mizushima, N., Levine, B., Cuervo, A. M. & Klionsky, D. J. Autophagy fights disease through cellular self-digestion. Nature 451, 1069–1075 (2008).
Guo, J. Y. & White, E. Autophagy is required for mitochondrial function, lipid metabolism, growth, and fate of KRAS(G12D)-driven lung tumors. Autophagy 9, 1636–1638 (2013).
Ramirez, C., Hauser, A. D., Vucic, E. A. & Bar-Sagi, D. Plasma membrane V-ATPase controls oncogenic RAS-induced macropinocytosis. Nature 576, 477–481 (2019).
Rabinowitz, J. D. & White, E. Autophagy and metabolism. Science 330, 1344–1348 (2010).
Commisso, C. et al. Macropinocytosis of protein is an amino acid supply route in Ras-transformed cells. Nature 497, 633–637 (2013).
Forgac, M. Vacuolar ATPases: rotary proton pumps in physiology and pathophysiology. Nat. Rev. Mol. Cell Biol. 8, 917–929 (2007).
Zhao, J., Benlekbir, S. & Rubinstein, J. L. Electron cryomicroscopy observation of rotational states in a eukaryotic V-ATPase. Nature 521, 241–245 (2015). May 14.
Stransky, L., Cotter, K. & Forgac, M. The function of v-ATPases in cancer. Physiol. Rev. 96, 1071–1091 (2016).
Li, Z. et al. Complete elucidation of the late steps of bafilomycin biosynthesis in Streptomyces lohii. J. Biol. Chem. 292, 7095–7104 (2017).
Rebecca, V. W. et al. PPT1 promotes tumor growth and is the molecular target of chloroquine derivatives in cancer. Cancer Discov. 9, 220–229 (2019).
Michaud, V. et al. Risk assessment of drug-induced long QT syndrome for some COVID-19 repurposed drugs. Clin. Transl. Sci. 14, 20–28 (2021).
Rebecca, V. & Amaravadi, R. Emerging strategies to effectively target autophagy in cancer. Oncogene 35, 1–11 (2016).
Carew, J. S. & Nawrocki, S. T. Drain the lysosome: development of the novel orally available autophagy inhibitor ROC-325. Autophagy 13, 765–766 (2017).
McAlister, G. C. et al. MultiNotch MS3 enables accurate, sensitive, and multiplexed detection of differential expression across cancer cell line proteomes. Anal. Chem. 86, 7150–7158 (2014).
Berg, E. L., Kunkel, E. J., Hytopoulos, E. & Plavec, I. Characterization of compound mechanisms and secondary activities by BioMAP analysis. J. Pharmacol. Toxicol. Methods 53, 67–74 (2006).
Jost, M. & Weissman, J. S. CRISPR approaches to small molecule target identification. ACS Chem. Biol. 13, 366–375 (2018).
Gilbert, L. A. et al. CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell. 154, 442–451 (2013).
Horlbeck, M. A. et al. Compact and highly active next-generation libraries for CRISPR-mediated gene repression and activation. eLife 5, e19760 (2016).
Gilbert, L. A. et al. Genome-scale CRISPR-mediated control of gene repression and activation. Cell 159, 647–661 (2014).
Bowman, B. J. & Bowman, E. J. Mutations in subunit C of the vacuolar ATPase confer resistance to bafilomycin and identify a conserved antibiotic binding site. J. Biol. Chem. 277, 3965–3972 (2002).
Liberman, R., Bond, S., Shainheit, M. G., Stadecker, M. J. & Forgac, M. Regulated assembly of the V-ATPase is increased during cluster disruption-induced maturation of dendritic cells through a PI-3 kinase/ mTOR-dependent pathway. J. Biol. Chem. 289, 1355–1363 (2014).
Stolze, B., Reinhart, S., Bulllinger, L., Fröhling, S. & Scholl, C. Comparative analysis of KRAS codon 12, 13, 18, 61, and 117 mutations using human MCF10A isogenic cell lines. Sci. Rep. 5, 8535 (2015).
Drosten, M. et al. Genetic analysis of Ras signaling pathways in cell proliferation, migration and survival. EMBO J. 29, 1091–1104 (2010).
Teng, K. W. et al. Selective and noncovalent targeting of RAS mutants for inhibition and degradation. Nat. Commun. 12, 2656 (2021).
Janes, M. R. et al. Targeting KRAS mutant cancers with a covalent G12C-specific inhibitor. Cell. 172, 578–589 (2018).
Haza, K. Z. et al. RAS-inhibiting biologics identify and probe druggable pockets including an SII-α3 allosteric site. Nat. Commun. 12, 4045 (2021).
Lock, R. et al. Autophagy facilitates glycolysis during Ras-mediated oncogenic transformation. Mol. Biol. Cell. 22, 165–178 (2010).
Webb, B. A. et al. pHLARE: a new biosensor reveals decreased lysosome pH in cancer cells. Mol. Biol. Cell. 32, 131–142 (2021).
Hobbs, G. A. et al. Atypical KRASG12R mutant is impaired in PI3K signaling and macropinocytosis in pancreatic cancer. Cancer Discov. 10, 104–123 (2020).
McFall, T. & Stites, E. C. A mechanism for the response of KRASG13D expressing colorectal cancers to EGFR inhibitors. Mol. Cell Oncol. 7, 1701914 (2020).
McFall, T., Schomburg, N. K., Rossman, K. L. & Stites, E. C. Discernment between candidate mechanisms for KRAS G13D colorectal cancer sensitivity to EGFR inhibitors. Cell Commun. Signal. 18, 179 (2020). Nov 5.
McFall, T. et al. A systems mechanism for KRAS mutant allele-specific responses to targeted therapy. Sci. Signal. 12, eaaw8288 (2019).
McFall, T. & Stites, E. C. Identification of RAS mutant biomarkers for EGFR inhibitor sensitivity using a systems biochemical approach. Cell Rep. 37, 110096 (2021).
Hammond, D. E. et al. Differential reprogramming of isogenic colorectal cancer cells by distinct activating KRAS mutations. J Proteome Res. 14, 1535–1546 (2015).
Trudeau, M. C., Warmke, J. W., Ganetzky, B. & Robertson, G. A. H. E. R. G. A human inward rectifier in the voltage-gated potassium channel family. Science 269, 92–95 (1995).
Kleinstreuer, N. C. et al. Phenotypic screening of the ToxCast chemical library to classify toxic and therapeutic mechanisms. Nat. Biotechnol. 32, 583–591 (2014).
Aguilera, K. Y. et al. Inhibition of discoidin domain receptor 1 reduces collagen-mediated tumorigenicity in pancreatic ductal adenocarcinoma. Mol. Cancer Ther. 16, 2473–2485 (2017).
Ruggeri, J. M. et al. Discoidin domain receptor 1 (DDR1) is necessary for tissue homeostasis in pancreatic injury and pathogenesis of pancreatic ductal adenocarcinoma. Am. J. Pathol. 190, 1735–1751 (2020).
Diehl, J. N. et al. The KRAS-regulated kinome identifies WEE1 and ERK coinhibition as a potential therapeutic strategy in KRAS-mutant pancreatic cancer. J. Biol. Chem. 297, 101335 (2021).
Wilkens, S., Inoue, T. & Forgac, M. Three-dimensional structure of the vacuolar ATPase. Localization of subunit H by difference imaging and chemical cross-linking. J. Biol. Chem. 279, 41942–41949 (2004).
Zoncu, R. et al. mTORC1 senses lysosomal amino acids through an inside-out mechanism that requires the vacuolar H+-ATPase. Science 334, 678–683 (2011).
Bar-Peled, L., Schweitzer, L. D., Zoncu, R. & Sabatini, D. M. Ragulator is a GEF for the rag GTPases that signal amino acid levels to mTORC1. Cell 150, 1196–1208 (2012).
Chung, C. Y. et al. Covalent targeting of the vacuolar H+-ATPase activates autophagy via mTORC1 inhibition. Nat. Chem. Biol. 15, 776–785 (2019).
Rabara, D. et al. KRAS G13D sensitivity to neurofibromin-mediated GTP hydrolysis. Proc. Natl Acad. Sci. USA 116, 22122–22131 (2019).
Yang, J. et al. Elevated expression of the V-ATPase D2 subunit triggers increased energy metabolite levels in KrasG12D-driven cancer cells. J. Cell Biochem. 120, 11690–11701 (2019).
Zhang, X., Cao, J., Miller, S. P., Jing, H. & Lin, H. Comparative nucleotide-dependent interactome analysis reveals shared and differential properties of KRas4a and KRas4b. ACS Cent. Sci. 4, 71–80 (2018).
Er, Tze-Kiong, Chen, Chih-Chieh, Bujanda, Luis & Herreros-Villanueva, Marta Clinical relevance of KRAS mutations in codon 13: where are we? Cancer Letters 343, 1–5 (2014).
Hobbs, G. A. & Der, C. J. Binge drinking: macropinocytosis promotes tumorigenic growth of RAS-mutant cancers. Trends Biochem. Sci. 45, 459–461 (2020).
Rubinsztein, D. C., Codogno, P. & Levine, B. Autophagy modulation as a potential therapeutic target for diverse diseases. Nat. Rev. Drug Discov. 11, 709–730 (2012).
Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603–607 (2012).
Kircher, M. et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nat. Genet. 46, 310–315 (2014).
Ting, L., Rad, R., Gygi, S. P. & Haas, W. MS3 eliminates ratio distortion in isobaric multiplexed quantitative proteomics. Nat. Methods 8, 937–940 (2011). Oct 2.
McAlister, G. C. et al. MultiNotch MS3 enables accurate, sensitive, and multiplexed detection of differential expression across cancer cell line proteomes. Anal. Chem. 86, 7150–7158 (2014).
Subramanian, T. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).
Liberzon, A. et al. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. 1, 417–425 (2015).
Kunkel, E. J. et al. Rapid structure–activity and selectivity analysis of kinase inhibitors by BioMAP analysis in complex human primary cell-based models. Assay Drug Dev. Technol. 2, 431–441 (2004).
Berg, E. L., Yang, J. & Polokoff, M. A. Building predictive models for mechanism-of-action classification from phenotypic assay data sets. J. Biomol. Screen. 18, 1260–1269 (2013).
Kingsley, R. E. & Cole, N. L. Preparation of cultured mammalian cells for transmission and scanning electron microscopy using ACLAR® film. J. Electron Microsc. Tech. 10, 77–85 (1988).
Sato, T. A modified method for lead staining of thin sections. J. Electron Microsc. 17, 158–159 (1968).
Gilbert, L. A. et al. Genome-scale CRISPR-mediated control of gene repression and activation. Cell 159, 647–661 (2014).
Jost, M. et al. Titrating gene expression using libraries of systematically attenuated CRISPR guide RNAs. Nat. Biotechnol. 38, 355–364 (2020).
Kampmann, M., Bassik, M. C. & Weissman, J. S. Integrated platform for genome-wide screening and construction of high-density genetic interaction maps in mammalian cells. Proc. Natl Acad. Sci. USA 110, E2317–E2326 (2013).
Stransky, L. A. & Forgac, M. Amino acid availability modulates vacuolar H+-ATPase assembly. J. Biol. Chem. 290, 27360–27369 (2015).
Galloway, C. J., Dean, G. E., Marsh, M., Rudnick, G. & Mellman, I. Acidification of macrophage and fibroblast endocytic vesicles in vitro. Proc. Natl Acad. Sci. USA 80, 3334–3338 (1983).
Steinberg, B. E. et al. A cation counterflux supports lysosomal acidification. J. Cell Biol. 189, 1171–1186 (2010).
Zoncu, R. et al. mTORC1 senses lysosomal amino acids through an insideout mechanism that requires the vacuolar H+-ATPase. Science 334, 678–683 (2011).
Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265–275 (1951).
Roberts, C. J., Raymond, C. K., Yamashiro, C. T. & Stevens, T. H. Methods for studying the yeast vacuole. Methods Enzymol. 194, 644–661 (1991).
Feng, Y. & Forgac, M. A novel mechanism for regulation of vacuolar acidification. J. Biol. Chem. 267, 5817–5822 (1992).
Benlekbir, S., Bueler, S. A. & Rubinstein, J. L. Structure of the vacuolar-type ATPase from Saccharomyces cerevisiae at 11-Å resolution. Nat. Struct. Mol. Biol. 19, 1356–1362 (2012).
Abbas, Y. M., Wu, D., Bueler, S. A., Robinson, C. V. & Rubinstein, J. L. Structure of V-ATPase from the mammalian brain. Science 367, 1240–1246 (2020).
Vasanthakumar, T. et al. Structural comparison of the vacuolar and Golgi V-ATPases from Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 116, 7272–7277 (2019).
& Eswar, N. et al. Comparative protein structure modeling using Modeller. Curr. Protoc. Bioinform. 15, 5.6.1–5.6.30 (2006).
Shen, M. Y. & Sali, A. Statistical potential for assessment and prediction of protein structures. Protein Sci. 15, 2507–2524 (2006).
Schrödinger Release 2019-1 (Schrödinger LLC, 2019).
Friesner, R. A. et al. Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein–ligand complexes. J. Med. Chem. 49, 6177–6196 (2006).
Commisso, C., Flinn, R. J. & Bar-Sagi, D. Determining the macropinocytic index of cells through a quantitative image-based assay. Nat. Protoc. 9, 182–192 (2014).
Piper, D. R. et al. Development of the predictor HERG fluorescence polarization assay using a membrane protein enrichment approach. Assay Drug Dev Technol. 6, 213–223 (2008).
Kleinstreuer, N. C. et al. Phenotypic screening of the ToxCast chemical library to classify toxic and therapeutic mechanisms. Nat. Biotechnol. 32, 583–591 (2014).
Davis, M. et al. Comprehensive analysis of kinase inhibitor selectivity. Nat. Biotechnol. 29, 1046–1051 (2011).
This study was supported in part by the UCSF Helen Diller Family Comprehensive Cancer Center Laboratory for Cell Analysis Shared Resource Facility through a grant from NIH (P30CA082103). All microscopy data was collected at the Lab for Cell Analysis. Flow Cytometry was performed at the UCSF Parnassus Flow Core facility (RRID: SCR_018206), which is supported in part by the NIH DRC Center Grant (P30 DK063720). The whole proteome mass spectrometry was performed at the Thermo Fisher Scientific Center for Multiplexed Proteomics. We thank several CROs for screening our molecule in targeted assays for a fee. The authors thank the following individuals for their generous contributions: H. Goodarzi at UCSF (MDA-MB-231 CRISPRi cells) and D. Brown at Harvard University (ATP6V1A and ATP6V1B2 antibodies used for Fig. 3). The MEF cells expressing human KRAS or BRAF were from the Frederick National Laboratory for Cancer Research, National Cancer Institute. We thank K. Yamamoto and J. Debnath from UCSF for valuable discussions. We thank M. Forgac and S. Wilkens for technical discussions. We thank L. Acevedo and N. Hoang for assistance with Western blotting. We thank M. Boone from the Walter lab at UCSF for yeast and related reagents. We thank the Wells lab at UCSF for use of the Octet RED384 for binding studies. We thank the Chief of the UCSF Thoracic Oncology laboratory for lab space and staff chemists for assistance with chemical synthesis/reagents. We thank P. Derish, UCSF Department of Surgery, for editorial assistance. Fig. 4 was created with BioRender (https://biorender.com). This work was supported by a UCSF Pancreas Center Pilot Project Grant Award funded by the Bern Schwartz Family Foundation (MMEPHMT603 to B.T.), the UCSF Hellman Family Award for Early Career Faculty (12RBP0002 to B.T.), the UCSF Thoracic Oncology Program laboratory (B.T.), the Canadian Institutes of Health Research postdoctoral fellowship (CIHR-IRSC:0525009424 to Y.Z.T.), an Ontario Graduate Scholarship (T.V.), the American Heart Association Postdoctoral Fellowship (18POST33960587 to P.B.), the Canada Research Chairs program and Canadian Institutes of Health Research grant (PJT166152 to J.L.R.), the National Institutes of Health (K99 GM130964 and R00 GM130964 to M.J. and 1RM1 HG009490-01 to J.S.W.). J.S.W. is a Howard Hughes Medical Institute Investigator.
B.T. is an inventor on the patent filed by UCSF covering these molecules. UCSF has licensed this patent to a start-up biotechnology company. The startup calls 249C ‘RSC-1255’ instead and is running a Phase Ia/1b dose escalation and dose expansion clinical trial for the following indications: Ras mutations; lung cancer; colon cancer; pancreatic cancer; and glioblastoma. It is being run as a multi-center trial across five different sites in the U.S. Recruitment is ongoing (NCT04678648). J.S.W. and M.J. have submitted patent applications related to CRISPR screening. J.S.W. declares outside interest in 5 AM Venture, Amgen, Chroma Medicine, DEM Biosciences, KSQ Therapeutics, Maze Therapeutics, Tenaya Therapeutics, Tessera Therapeutics and Velia Therapeutics. M.J. consults for Maze Therapeutics and Gate Biosciences. All other authors declare no competing interests.
Peer review information
Nature Biotechnology thanks Alessandra d’Azzo, Kirsten Bryant, Philippe Cassier and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Tolani, B., Celli, A., Yao, Y. et al. Ras-mutant cancers are sensitive to small molecule inhibition of V-type ATPases in mice. Nat Biotechnol (2022). https://doi.org/10.1038/s41587-022-01386-z