Bi-steric mTORC1 inhibitors induce apoptotic cell death in tumor models with hyperactivated mTORC1

The PI3K/AKT/mTOR pathway is commonly dysregulated in cancer. Rapalogs exhibit modest clinical benefit, likely owing to their lack of effects on 4EBP1. We hypothesized that bi-steric mTORC1-selective inhibitors would have greater potential for clinical benefit than rapalogs in tumors with mTORC1 dysfunction. We assessed this hypothesis in tumor models with high mTORC1 activity both in vitro and in vivo. Bi-steric inhibitors had strong growth inhibition, eliminated phosphorylated 4EBP1, and induced more apoptosis than rapamycin or MLN0128. Multiomics analysis showed extensive effects of the bi-steric inhibitors in comparison with rapamycin. De novo purine synthesis was selectively inhibited by bi-sterics through reduction in JUN and its downstream target PRPS1 and appeared to be the cause of apoptosis. Hence, bi-steric mTORC1-selective inhibitors are a therapeutic strategy to treat tumors driven by mTORC1 hyperactivation.


Introduction
The PI3K/AKT/mTOR network is a critical intracellular signaling pathway directing cell growth and metabolism in physiological and pathological conditions (1).The evolutionarily conserved mammalian target of rapamycin (mTOR) is a serine/threonine kinase ubiquitously expressed in mammalian cells (2).mTOR is the key protein in mTORC1 and mTORC2 protein complexes (3), and mTORC1 regulates a signaling cascade involved in protein synthesis, gene expression, glucose and lipid metabolism, and nucleotide biosynthesis (4).mTORC1 phosphorylates 4E-binding protein 1 (4EBP1) and S6 kinase (S6K), downstream targets involved in cap-dependent translation initiation and elongation (5).Complete loss of TSC1 or TSC2, occurring in both tuberous sclerosis complex (TSC) tumors and a variety of cancer types (6)(7)(8), leads to constitutive unregulated activation of mTORC1 (9).
Given the frequent activation of PI3K/AKT/mTOR signaling in human tumors, several generations of mTOR inhibitors have been developed (1).Rapalogs have been approved by the FDA for the treatment of several TSC-associated tumors as well as renal cell carcinoma (RCC) (10).However, rapalogs have shown limited benefit in treating patients with RCC, bladder cancer (BLCA), and cancers of any origin with biallelic mutations in either TSC1 or TSC2 (11)(12)(13), which may be due to the incomplete suppression of mTORC1 kinase activity by rapalogs, including a lack of effect on p-4EBP1 T37,46 , p-PRAS40 S183 , and p-Grb10 S150 (14).MLN0128 (sapanisertib), a second-generation ATP-competitive inhibitor, inhibits all the mTORC1 substrates, but also inhibits mTORC2, likely contributing to reported toxicities necessitating lowered dosage that may limit its clinical efficacy (15).
Third-generation "bi-steric" mTORC1-selective inhibitors have been developed, consisting of a rapamycin derivative and an mTOR active-site inhibitor connected by a linker region (16).These bi-steric inhibitors overcome the limitations of the firstand second-generation inhibitors by virtue of being mTORC1 selective and able to inhibit p-4EBP1.We hypothesized that these potent mTORC1-selective inhibitors have the potential for greater benefit than rapalogs in the treatment of tumors with mTORC1 hyperactivation.We show that bi-steric inhibitors, RMC-4627, RMC-6272, and RMC-5552 (a clinical candidate) (17), drive rapid and durable suppression of p-S6K T389 , p-4EBP1 T37, 46 , p-mTOR S2481,S2448 , and p-PRAS40 S183,T246 .Bi-steric inhibitors are tolerated and have consistent antitumor activity both in vitro and in vivo.Furthermore, these bi-steric mTORC1-selective inhibitors induce global reprogramming of anabolism and catabolism as compared with rapamycin, indicating broad effects of more effective mTORC1 inhibition.Notably, RMC-5552 is currently undergoing clinical evaluation in patients with cancer (ClinicalTrials.gov NCT04774952).Our work provides evidence that bi-steric mTORC1 inhibitors provide improved therapeutic benefit as compared with the currently used rapalogs in tumors with high mTORC1 activity, due to important differences in the effects of these agents on purine metabolism.These studies provide strong The PI3K/AKT/mTOR pathway is commonly dysregulated in cancer.Rapalogs exhibit modest clinical benefit, likely owing to their lack of effects on 4EBP1.We hypothesized that bi-steric mTORC1-selective inhibitors would have greater potential for clinical benefit than rapalogs in tumors with mTORC1 dysfunction.We assessed this hypothesis in tumor models with high mTORC1 activity both in vitro and in vivo.Bi-steric inhibitors had strong growth inhibition, eliminated phosphorylated 4EBP1, and induced more apoptosis than rapamycin or MLN0128.Multiomics analysis showed extensive effects of the bi-steric inhibitors in comparison with rapamycin.De novo purine synthesis was selectively inhibited by bi-sterics through reduction in JUN and its downstream target PRPS1 and appeared to be the cause of apoptosis.Hence, bi-steric mTORC1-selective inhibitors are a therapeutic strategy to treat tumors driven by mTORC1 hyperactivation.cells, consistent with a selective effect.RMC-6272 caused a more dramatic reduction in clone formation than RMC-4627 for all cell lines.RMC-6272 (1 nM, 24 hours) also led to G 0 /G 1 cell cycle arrest in contrast to rapamycin (1 nM, 24 hours) (Figure 1D).
RMC-6272 treatment led to changes in mRNA expression compared with rapamycin or DMSO (Supplemental Figure 5, A  and D).Pathway enrichment analysis showed that genes downregulated by RMC-6272 versus rapamycin were enriched for those involved in cell cycle/cell phase transition and DNA replication in both HCV29 TSC1-null and 705 Tsc2-null cell lines (Figure 2, A and B, Supplemental Figure 5, B and C, and Supplemental Tables 2 and 3).These findings were consistent with the cell cycle arrest induced by RMC-6272 (Figure 1D), and suggest the importance of inhibition of 4EBP1 phosphorylation to prevent cell proliferation induced by mTORC1 (26).
Three hundred polar metabolites were analyzed using mass spectrometry.Metabolite set enrichment analysis (MSEA) identified purine metabolism as the metabolite set that was most enriched for metabolites decreased by RMC-6272 in comparison with rapamycin for both human HCV29 TSC1-null (Figure 2C) and mouse 705 Tsc2-null (Supplemental Figure 7A) cell lines.This differential effect of RMC-6272 on purine metabolites in comparison with rapamycin was further verified using another four TSC1/ TSC2-null cell lines (Supplemental Figure 6), for which differential effects on purine metabolite levels was among the top-ranked metabolite sets in each cell line.
MSEA showed no major change in glycolysis (Supplemental Figure 11A).Both rapamycin and RMC-6272 caused a significant support for human clinical trials of selective bi-steric mTORC1 inhibitors in patients with TSC1/TSC2-deficient cancers.
Low-dilution clone formation decreased up to 50% with rapamycin and showed little effect with MLN0128, while the bi-steric inhibitors caused at least 70% reduction in multiple TSC-null cell lines, including several with more than 95% reduction (Figure 1, E and F, and Supplemental Figure 3).Both bi-steric inhibitors caused a milder decrease in growth of the TSC1-add-back/control lipids detected.As shown by lipid maps, RMC-6272 led to an apparent conversion from phosphatidic acid to phosphatidylinositol in both cell lines (Figure 2, D and E, and Supplemental Figure 5, F and G), indicating a differential effect on lipid biogenesis in comparison with rapamycin.Lipid subtypes (ceramides reduction of glucose uptake in comparison with DMSO-treated cells, while RMC-6272 showed a greater effect than rapamycin in 4 of 4 cell lines (Supplemental Figure 11B).
Lipidomic analysis showed that both rapamycin and RMC-6272 caused perturbations among the 43 different subtypes of step of purine de novo synthesis, was reduced at both mRNA (Figure 3B and Supplemental Figure 7C) and protein levels after 24 hours of either RMC-4627 or RMC-6272 at 20 nM (Supplemental Figure 4, P, U, and V).Quantitative reverse transcriptase PCR analysis confirmed that PRPS1, but not PRPS2, which performs a similar catalytic function, was decreased up to 90% by RMC-6272 treatment compared with rapamycin, which caused a much smaller decrease (~20%) (Figure 3, C and D, and Supplemental Figure 7, D and E).Purine nucleotides are needed for many intracellular processes, including DNA replication and cell proliferation, and a lack of purine nucleotides would be expected to cause cell cycle arrest and effects on expression of genes regulating cell cycle progression, as noted in our RNA-Seq analysis (Figure 2A and Supplemental Figure 5B).To examine the hypothesis that purine nucleotide deficiency explained the growth-inhibitory effects of bi-steric inhibitors, we performed an add-back experiment with exogenous purines.The addition of IMP, AMP, and GMP at 100 nM each to the culture medium increased the IC 50 of bi-steric inhibitors 4-to 10-fold on the TSC1-null HCV29 cell line and the Tsc2-null 705 cell line (Figure 3, E-G, and Supplemental Figure 7, H-J; compare with Figure 1A, Supplemental Figure 2I, and Supplemental Table 1).There was a minor effect of the addition of these purines on the IC 50 of rapamycin and no effect on the IC 50 of MLN0128.Furthermore, PRPS1 knockdown (KD) by siRNA had a significant effect on HCV29 cell proliferation in comparison with control siRNA-treated cells (Figure 3, H-J).In addition, Prps1 knockout (KO) by CRISPR/Cas9 induced cell apoptosis with no surviving cells when applied to each of 3 different murine Tsc1/ Tsc2-null cell lines (Supplemental Figure 7G), confirming the critical role of Prps1 in purine synthesis for cell growth and survival.Interestingly, PRPS1 mRNA expression is associated with reduced survival and positively correlated with DNA replication, tumor proliferation, and the PI3K/AKT/mTOR signature in the TCGA data set (Figure 3, M-O), confirming its broad essential role in cell growth and proliferation.
Having discovered that PRPS1 mRNA levels are decreased after mTORC1 suppression, we sought to determine the mechanism of this effect.JASPAR (28) was used to predict transcription factors (TFs) that bind to the PRPS1 promoter region, and several candidate TFs that may regulate PRPS1 expression were identified, including JUN, E2F4, JUND, and ATF4.mTORC1 inhibitors had little effect on ATF4 expression at 24 hours (Supplemental Figure 4, P, U, and V), and no effect on E2F4 and JUND (data not shown).In contrast, JUN protein levels decreased dramatically starting at 3 hours of bi-steric inhibitor treatment, whereas rapamycin had a milder effect on JUN expression beginning at 24 hours (Supplemental Figure 4P).RNA-Seq data, in contrast, showed an approximately 50% increase in JUN mRNA in both HCV29 and 705 cells in response to RMC-6272 treatment (Figure 3B and Supplemental Figure 7B).This is consistent with prior reports that JUN autoregulates its expression through an inhibitory binding interaction with its own promoter region, such that a decrease in JUN protein triggers upregulation of its mRNA (29).The decreased JUN expression upon mTORC1 inhibition suggests that JUN may be activated in some manner downstream of mTORC1.TCGA pan-cancer reverse-phase protein assay data showed that JUN mRNA level is positively correlated with [Cer], phosphatidylinositol [PI], and triglycerides [TG]), fatty acid length, and number of unsaturated bonds did not show a significant difference between rapamycin and RMC-6272 treatment (Supplemental Figure 12).
Global proteomics analysis also showed major differences in the effects of the bi-steric mTORC1-selective inhibitor RMC-6272 in comparison with rapamycin.Proteins downregulated by RMC-6272 treatment relative to rapamycin were enriched for translation and ribosome-related proteins according to pathway analysis (adjusted P value < 0.05, |log 2 (fold change)| > 1.5; Figure 2F and Supplemental Figure 5E), consistent with effects of non-phosphorylated 4EBP1 on synthesis of those proteins induced by RMC-6272.In addition, there was little overlap between the most downregulated mRNAs and the most downregulated proteins, consistent with effects of 4EBP1 on translation initiation.
Since mTORC1 phosphorylates multiple targets and initiates a downstream phosphorylation cascade, global phosphoproteomics analysis was performed on Tsc2-null 705 cells treated with rapamycin, RMC-6272, and DMSO as control.Both rapamycin and RMC-6272 caused dephosphorylation at multiple sites on S6 (Supplemental Figure 5J); in contrast, only RMC-6272 eliminated phosphorylation at multiple sites of 4EBP1 (Supplemental Figure 5I).These findings are consistent with immunoblot findings (Figure 1G and Supplemental Figure 4).Differential analysis of the effects of RMC-6272 versus rapamycin on the phosphoproteome using Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways indicated that proteins involved in ribonucleoprotein biogenesis, RNA splicing, and ribosome biogenesis were the proteins whose phosphorylation was most commonly downregulated (Figure 2G).Genes undergoing alternative splicing (AS) identified by RNA sequencing (RNA-Seq) were then analyzed.There were 400 genes that had a difference in AS induced by RMC-6272 versus rapamycin and were involved in the DNA metabolic process and other pathways (Supplemental Figure 5H).
Although both rapamycin and RMC-6272 decreased protein synthesis, RMC-6272 had a much stronger effect with 80% reduction in protein synthesis in comparison with 50% for rapamycin (Supplemental Figure 11, D and E).
Bi-steric mTORC1-selective inhibition suppresses de novo purine synthesis through the mTORC1/JUN/PRPS1 axis.Multiple purine metabolites were decreased to a greater extent in RMC-6272treated HCV29 cells in comparison with rapamycin treatment, including adenosine, AICAR, AMP, IMP, and GMP (Figure 3A), with similar findings in 4 other TSC1/TSC2-null cell lines (Supplemental Figure 6).Notably, there was no significant change in ribose 5-phosphate after RMC-6272 treatment (Figure 3A and Supplemental Figure 6, B, D, F, and H), suggesting that downregulation of the purine de novo synthesis pathway was the cause of these changes.PRPS1, which catalyzes the first and rate-limiting mTORC1 activity in 27 of 32 cancer types (data not shown).To further explore the mechanism of how mTORC1 regulates JUN, cells were treated with actinomycin D (ActD).ActD reduced JUN protein levels at 12 hours, independent of DMSO/rapamycin/ RMC-6272 treatment.In contrast, cycloheximide (CHX) treatment caused a rapid and near-complete loss of JUN protein only with concurrent bi-steric treatment (Supplemental Figure 8, E and  F).The ActD effect suggests that JUN protein expression requires continued RNA synthesis, while the CHX effect only with bi-steric treatment suggests that complete mTORC1 inhibition caused a reduction in JUN protein stability.
To examine the relationship between JUN and PRPS1 in greater detail, JUN CRISPR gene deletion was performed in 6 different TSC1/TSC2-null cell lines.Both JUN and PRPS1 protein levels were reduced (Figure 3, K and L).To determine whether JUN was regulating PRPS1 expression directly, JUN and H3K27ac CUT&RUN was performed.JUN bound to PRPS1 in a representative cell line with mTORC1 hyperactivation (621-101), indicating that PRPS1 was a direct transcriptional target (Figure 3P).Our results were further validated by publicly available JUN ChIP-Seq data on 2 cancer cell lines (MDA-MB-231 and A549) (30, 31) (Supplemental Figure 8M).Furthermore, MSEA showed that the purine synthesis pathway was the most decreased pathway in c-JUN KD cells in comparison with c-JUN WT cell lines (Supplemental Figure 8, A and B).
Twenty-eight of 32 TCGA cancer types had increased expression of both JUN and PRPS1 in comparison with their respective normal tissues, indicating the importance of JUN and PRPS1 in tumorigenesis and tumor progression.In summary, these findings indicate that mTORC1 regulates purine de novo synthesis through modulation of JUN expression and downstream PRPS1.
mTORC1 bi-steric inhibitors are more effective than rapamycin in multiple TSC-deficient tumor models with mTORC1 hyperactivation in vivo.Consistent with our mechanistic observations, RMC-6272 drove deeper tumor regression in 4 weeks of treatment, and significantly delayed tumor regrowth upon treatment cessa-tion, compared with rapamycin, both at translatable doses, in a human TSC1-null BLCA patient-derived xenograft (PDX) model (n = 8 per group; Figure 4, A and B).MLN0128 at maximum tolerated dose showed little activity in this model.Modest weight loss occurred after each dose of RMC-6272, followed by recovery (Supplemental Figure 9A).To examine the in vivo mechanism of tumor response, a separate cohort of mice with PDX received a single dose of RMC-6272, followed by sacrifice at different time points (Figure 5A).Both immunoblot and IHC showed that S6K and 4EBP1 phosphorylation sites were eliminated by RMC-6272 for a week.In contrast, these markers showed a more limited and transient response to rapamycin and MLN0128 (Figure 4E and Supplemental Figure 10B).RMC-6272 treatment also led to dephosphorylation of multiple mTORC1 components, including mTOR, Raptor, and PRAS40 (Supplemental Figure 10B).Strikingly, RMC-6272-treated PDX tumors showed near-complete ablation of Ki-67 staining and evidence of apoptosis as indicated by c-PARP and cleaved caspase-3 (CCS3) staining.These were not seen with either rapamycin or MLN0128 treatment (Figure 4E and Supplemental Figure 9K).Metabolite analysis of the tumors at 24 hours (Figure 4, C and D) but not 4 hours (Supplemental Figure 9, B and C) showed a profound decrease in purine metabolites, in comparison with rapamycin, with adenosine showing a 90% reduction.JUN IHC also showed that total and nuclear JUN levels were reduced by RMC-6272 at all times, while this was more limited and transient with rapamycin and not seen at all with MLN0128 (Supplemental Figure 9, L and M).PRPS1, but not PRPS2, was decreased at the mRNA level (Supplemental Figure 9, I and J).
We evaluated further the in vivo antitumor activity of bi-steric mTORC1-selective inhibitors by testing RMC-5552 in 3 human BLCA PDX models with TSC1/TSC2 deficiency, including the same model in which RMC-6272 was assessed.RMC-5552 is a clinical candidate representing the bi-steric mTORC1-selective inhibitor class, and currently being tested in a phase I clinical trial (NCT04774952).All 3 BLCA PDX models with TSC1/TSC2 deficiency were sensitive to RMC-5552, which caused significant tumor volume reduction in 2 of 3 PDX models, and delayed tumor growth in the third (Supplemental Figure 9, D-H).
We also examined the effects of the bi-steric inhibitors on a genetically engineered mouse model of Tsc2 RCC (Tsc2 +/-A/J mice; Figure 6A) and a mouse lung tumor model (Figure 7).Both rapamycin and bi-steric inhibitors were tolerated as assessed by body weight (Figure 6C and Figure 7A) and significantly decreased tumor burden (Figure 6, B, D, and E).RMC-6272 showed less tumor regrowth than rapamycin or MLN0128 after treatment cessation (Figure 6B), particularly in the mouse lung tumor model, in which no tumor regrowth was observed (Figure 7,  B-D).Single-dose experiments confirmed that RMC-6272 caused more tumor apoptosis than rapamycin (Figure 5 and Supplemental Figure 10A).RMC-6272 showed a more prolonged suppressive effect on mTORC1 activity as compared with rapamycin (Figure 5, B and I, and Figure 6F), which occurred with reduced tumor cell proliferation and enhanced apoptosis (Figure 5, C-F).Differences in T cell numbers or macrophages were not observed, suggesting that apoptosis was due to intrinsic effects of mTORC1 inhibition on the tumor cells (Figure 5, G and H, and Figure 6F).pressor genes (PTEN, TSC1, TSC2, PIK3R1), and amplification of growth-promoting oncogenes (18).All of these events lead to enhanced mTORC1 signaling, which through diverse downstream effectors leads to cell anabolic processes for both critical biochem-

Discussion
The PI3K/AKT/mTOR pathway is commonly dysregulated in human cancer, due to mutation of dominantly acting oncogenes (PIK3CA, AKT isoforms, MTOR), inactivation of tumor sup- p-4EBP1 and p-S6K.Both compounds demonstrate selectivity for mTORC1 over mTORC2, with approximately 27-fold and 40-fold selectivity, respectively.This selectivity is afforded by modulation of both the active site inhibitor and the rapamycin core.For both RMC-6272 and RMC-5552 the carbonyl at the C32 position has been reduced to a hydroxyl group, which serves to modulate binding affinity to FKBP12 and improves chemical stability of the macrocyclic core.
Here we show that several bi-steric mTORC1-selective inhibitors (a) show high selectivity for mTORC1 over mTORC2; (b) completely inhibit mTORC1 activity on all substrates tested; (c) have a higher potency for mTORC1 inhibition in TSC1/TSC2mutant cancer cell lines as compared with rapamycin, and show modest selectivity (5-to 10-fold shift in IC 50 ) for these cells in comparison with parallel or (isogenic) add-back cell lines; (d) show effects in inhibiting the growth of such cell lines in clonal dilution growth assays; and (e) show improvement over rapamycin in the treatment of a variety of in vivo models with mTORC1 hyperactivation due to TSC1/TSC2 loss, including syngeneic xenografts, native mouse genetic models, and human bladder cancer patientderived xenografts.
Bi-steric mTORC1-selective inhibitors caused global rewiring of cellular metabolism, with significant effects on RNA expression, metabolites, lipids, protein expression, and phosphoprotein levels in comparison with rapamycin.Most importantly, bi-steric inhibitors caused tumor cell apoptosis, in contrast to rapamycin, for several of the tumor models for which superiority to rapamycin was demonstrated.
Our assessment of the omics effects of bi-steric treatment led us to identify a profound reduction in purine metabolites in the bi-steric-treated cells.Subsequent studies identified an axis of transcriptional regulation of purine synthesis downstream of mTORC1, in which JUN drives expression of PRPS1, the critical first enzyme in the de novo purine synthesis pathway.We found that both JUN and PRPS1 are reduced in expression by immunoblotting following the bi-steric inhibitor treatment, both in vitro and in vivo, consistent with the reduction in purine metabolites and constituting potential substrates downstream of mTORC1 inhibition.In addition, several different purines individually rescued the growth-inhibitory effects of the bi-steric inhibitors in vitro, while PRPS1 was shown to be absolutely required for viability of all cell lines tested.Previously an mTORC1/ATF4/ MTHFD2 pathway was identified as critical for purine synthesis in cells with mTORC1 activation due to TSC1/TSC2 loss, and sensitive to rapalog inhibition (40).However, we saw no effect on ATF4 expression, and a modest approximately 50% reduction in MTHFD2 expression in response to bi-steric treatment.While it is not completely clear how mTORC1 regulates JUN, we observed that JUN protein levels were decreased by 3 hours after bi-steric inhibitor treatment but not rapamycin treatment, and complete mTORC1 suppression by bi-steric mTORC1-selective inhibitors caused a rapid drop in JUN levels when cotreated with CHX (Supplemental Figure 8E).Thus, mTORC1 appears to regulate JUN translation in a 4EBP1-dependent manner.Further studies are required, including investigation of the possibility that mTORC1 or a downstream kinase phosphorylates JUN, which might regulate its expression level.ical intermediates and macromolecules (e.g., ribosomes) that enable cell growth (32).Arguably, the strongest mTORC1-activating genetic events are those in which there is biallelic loss of TSC1 or TSC2.TSC1, TSC2, and TBC1D7 form a protein complex that converts RHEB.GTP to RHEB.GDP, and loss of this TSC protein complex function leads to high levels of RHEB.GTP and constitutive activation of mTORC1 (2).
Several generations of mTORC1 inhibitors have been created.Rapamycin and its analogs, rapalogs (33), cause partial responses in about 10% of RCC patients, and have been reported to have major benefit in occasional cases of bladder cancer and other cancer types with TSC1/TSC2 mutations (13).Rapalogs have shown more consistent benefit for several tumors (renal angiomyolipoma, cortical subependymal giant cell astrocytoma, pulmonary lymphangioleiomyomatosis) that develop as part of the TSC syndrome, as well as sporadically (34,35).However, rapalogs produce a modest but persistent response in many TSC patients, with an average 50% reduction in tumor volume, and requirement for lifelong treatment since tumors will recur when rapalog therapy is stopped (11).
The rapalog-FKBP12 complex reduces mTORC1 kinase activity in a variable manner on its different substrates, with S6K among the most sensitive, and 4EBP1 among the most resistant, to rapalog inhibition (Figure 1, G and H).There are at least 10 mTORC1 substrates that are insensitive to rapalog inhibition (36).This likely contributes to the limited clinical benefit of rapalog therapy (37).
Bi-steric mTORC1-selective inhibitors were designed to combine the high selectivity of the rapamycin-FKBP12 complex for the FRB domain and an mTOR ATP binding domain moiety, to give high specificity for mTORC1 over mTORC2, along with near-complete inhibition of mTORC1 activity on all of its substrates (38,39).
RMC-4627 is a bi-steric inhibitor that contains a rapamycin core and PP242-derived active site inhibitor, which are linked through an ether bond at the C40 position on rapamycin and a Peg8 linker.Based on cellular assays using MDA-MB-468 cells, RMC-4627 displays potent inhibition of p-4EBP1 (IC 50 : 1.4 nM) and p-S6K (IC 50 : 0.28 nM) and demonstrates selectivity for mTORC1 over mTORC2 (calculated as p-AKT IC 50 /p-4EBP1 IC 50 ), with approximately 13-fold selectivity.RMC-6272 and RMC-5552 are 2 bi-steric inhibitors in which the C40-linked ether chemical handle was exchanged for a C40 carbamate to enable synthetic tractability and incorporated the XL388-derived (RMC-6272) and MLN0128-derived (RMC-5552) active site inhibitors, respectively.RMC-6272 and RMC-5552 display very potent inhibition of ment is also well tolerated in vivo in mouse models at doses that lead to therapeutic benefit.The bi-steric mTORC1-selective inhibitor RMC-5552 is now in human clinical trials and, based on the results presented herein, has the potential to benefit patients with TSC syndrome tumors, and patients with the common cancers in which TSC1/TSC2 mutations drive hyperactivated mTORC1 activity.
In summary, bi-steric mTORC1-selective inhibitors with moieties that bind to both FKBP12 and the ATP binding pocket on mTOR show high specificity for mTORC1, require lower concentrations for mTORC1 inhibition than rapalogs, induce apoptosis in in vivo models, and appear to act (at least in part) by suppressing de novo purine synthesis through reduction in PRPS1 expression.Bi-steric mTORC1-selective inhibitor treat-
Cell proliferation assay (including washout).The effect of the compounds on cell proliferation was measured by crystal violet staining.Cells were plated in 96-well plates at a density of 500-2,000 cells (sextuples per condition).The next day, cells were treated with concentrations (0-100 nM) prepared by serial dilution.After 3 days of treatment, cells were fixed by 10% formalin for 10 minutes at room temperature.0.05% crystal violet solution was added and incubated for 15 minutes at room temperature.Absorbance was read using Synergy HT (BioTek) at 540 nm.The inhibition curves and IC 50 values were calculated by GraphPad Prism 9.0.
14-Day low-dilution clone formation assay.Long-term effects on cell proliferation were assessed by counting of the number of clones after 200-1,000 cells were seeded on 10 cm plates (triplicates per condition).After 3 days, cells were treated at concentrations selected based on IC 50 values (with a maximum of 10 nM).Media were changed every 3 days.Cell clones were fixed with 10% formalin and stained with 0.05% crystal violet after 14 days of treatment.The plates were washed extensively and scanned with a flatbed scanner.For quantification, clone numbers with size ≥ 1 mm were counted manually and data analyzed using GraphPad Prism 9.0.
Cell cycle analysis.Cell cycle analysis was performed following the manufacturer's instructions (Abcam, ab-139148).Briefly, cells were harvested and prepared in single-cell suspension after being treated with DMSO, rapamycin, or RMC-6272 for 24 hours.Cell pellets were washed and fixed in 66% ethanol on ice for 2 hours.Cells were then resuspended in propidium iodide plus RNase staining solution at 37°C for 30 minutes.Cell populations were analyzed using FACS (BD LSRFortessa).
Western blotting.Cells were washed with cold PBS twice and lysed on ice for 30 minutes in 1× RIPA buffer (Cell Signaling Technology) supplemented with protease and phosphatase inhibitor cocktails (Sigma-Aldrich), followed by centrifugation at 18,000g at 4°C.Proteins incorporated.The protein samples were analyzed by positive-ion mode LC-MS/MS using a high-resolution hybrid QExactive HF Orbitrap Mass Spectrometer (Thermo Fisher Scientific) via higher-energy collisional dissociation with data-dependent analysis with 1 MS1 scan followed by 8 MS2 scans per cycle for the top 8 ions detected in the MS1 scan.MS/MS spectra were analyzed for the peptide samples with a parent ion tolerance of 18 ppm and fragment ion tolerance of 0.05 Da.Carbamidomethylation of cysteine (+57.0293Da) was specified as a fixed modification, and oxidation of methionine (+15.9949) and phosphorylation of serine/threonine/tyrosine (+79.97) as variable modifications.Results were imported into Scaffold Q+S 4.6 software (Proteome Software Inc.) with a peptide threshold of about 85% and a protein threshold of 95%, resulting in a peptide false discovery rate of approximately 1%.Further statistical analysis was performed using PANTHER (http://www.pantherdb.org/)after removal of contaminants such as keratins, caseins, trypsin, and BSA (47).Omics data were normalized to the median of each sample and then subjected to differential analysis (DESeq2) and pathway enrichment analysis (GSEA).Lipidomic data were used to generate a lipid map using Lipid Maps (49) (https://www.lipidmaps.org).
In situ cell death detection (TUNEL staining).TUNEL staining was performed on FFPE sections.After dewaxing and hydrating, slides were permeabilized for 8 minutes at room temperature with 0.1% Triton X-100 prepared in 0.1% sodium citrate in PBS.Slides were incubated with TUNEL reaction mixture (prepared according to In Situ Cell Death Detection Kit Fluorescein, 11684795910) in a humidified atmosphere for 60 minutes at 37°C in the dark.After washing of the slides twice in PBS, DAPI (Sigma-Aldrich, D9542-10MG) was applied, and slides were mounted with Fluoromount-G.
Measurement of protein synthesis rate.Protein synthesis rate was tested using Click-iT HPG Alexa Fluor 594 Protein Synthesis Assay Kit (Thermo Fisher Scientific, C10429).Cells were treated with rapamycin (10 nM) or RMC-6272 (10 nM) for 24 hours.Protein synthesis rate was detected by fluorescent microscope or measured by plate reader.
RNA sequencing.RNA-Seq was performed at Novogene.Cells were treated with DMSO, rapamycin (10 nM), or RMC-6272 (3 nM) for 24 hours (triplicates per condition).Total RNA was extracted using an RNeasy Mini Kit (Qiagen, 74104).One microgram RNA per sample was used for cDNA library preparation using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs).Paired-end reads (~20 M/ sample) were aligned to the hg38 human or mm10 genome using Spliced Transcripts Alignment to a Reference (STAR; v2.5) software.HTSeq v0.6.1 (45) was used to count the read numbers mapped to each gene, followed by determination of fragments per kilobase of exon model per million mapped reads (FPKM).Differential expression analysis between 2 conditions was performed using the DESeq2 R package (2_1.6.3).The resulting P values were adjusted using the Benjamini-Hochberg approach for controlling the false discovery rate.Genes with adjusted P value less than 0.05 were considered differentially expressed.Gene set enrichment analysis (GSEA) for analysis of gene expression differences was performed using the Gene Ontology and KEGG gene sets.
Alternative splicing analysis.Alternative splicing (AS) analysis was performed using rMATS on RNA-Seq data (46).Exon skipping, alternative 5′ splice sites, alternative 3′ splice sites, mutually exclusive exons, and retained introns were identified and compared between RMC-6272-and rapamycin-treated cells.P values and FDRs were calculated for differential splicing events.
Multiomics.Triomics analysis of lipids, metabolites, proteins, and phosphoproteins was performed as previously described (47).Pellets from 10 million cells were obtained by spinning down at 500g at 4°C after 24 hours of DMSO, rapamycin (10 nM), or RMC-6272 (3 nM) treatment (triplicates per condition).Two hundred microliters of 1× PBS and 1.5 mL HPLC-grade methanol were added, followed by a vigorous vortex for 1 minutes at room temperature.The samples were shaken for 1 hour at room temperature after addition of 5 mL of HPLC-grade MTBE, anhydrous 99.8% (Sigma-Aldrich, 34875-2L).Then 1.2 mL HPLC-grade water was added, vortexed for 1 minute, and spun for 10 minutes.The resulting upper (lipids) and middle (metabolites) phases were collected separately in 1.5 mL glass vials and dried out using SpeedVac Concentrator (Thermo Fisher Scientific, SC110A).A protein pellet at the bottom was used for both proteomics and phosphoproteomics.High-Select TiO 2 Phosphopeptide Enrichment Kit (Thermo Fisher Scientific, PIA32993) was used to enrich phosphopeptides.The metabolite samples were resuspended in 20 μL liquid chromatography-mass spectrometrygrade (LC-MS-grade) water and run as previously described (48).The data were analyzed using Elements for Metabolomics (Proteome Software) with the NIST database incorporated (http://chemdata.nist.gov/mass-spc/msms-search/) followed by statistical analysis with Metabo-Analyst 5.0 (http://www.metaboanalyst.ca/).The lipid samples were resuspended in 30 μL of LC-MS-grade isopropanol/methanol (1:1), and 5 μL was injected for LC-MS/MS analysis.Lipidomic data were analyzed using LipidSearch 4.1.9software (Thermo Fisher Scientific) and Elements for Metabolomics (Proteome Software) with NIST database J Clin Invest.2023;133(21):e167861 https://doi.org/10.1172/JCI167861Tumors were measured twice weekly using calipers, and tumor volume was calculated using the formula 0.5 × length × width 2 .Mouse body weight was monitored twice per week.
PDX 1-dose treatment.The same human BLCA PDX-2211 described above was used.Mice were treated once with vehicle, rapamycin (3 mg/kg), or RMC-6272 (8 mg/kg) when tumor size reached a range of 200-600 mm 3 .Mice were sacrificed 4, 24, 72, or 168 hours after the single dose.Each tumor sample was cut into 2 pieces, with one being snap-frozen in liquid nitrogen and the other being fixed in 10% formalin for further analysis.
Statistics.Data were analyzed using GraphPad Prism 9.0.Results are presented as mean ± SD.All cell and Western blot experiments were repeated at least twice with separately prepared samples.The log-rank (Mantel-Cox) test was used to determine significance for survival curves.For 2-group comparisons, a 2-tailed unpaired Student's t test was applied.One-way ANOVA was used for multiple comparisons in the experiments with more than 2 groups.P less than 0.05 was considered statistically significant.P values are denoted with asterisks: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.
Study approval.All animal experiments were conducted using protocols approved by the Brigham and Women's Hospital Institutional Animal Care and Use Committee.
Treatment of mouse lymphangioleiomyomatosis models.TTJ cells were resuspended in 1× sterile PBS and injected into 6-week-old male C57BL/6J mice (The Jackson Laboratory, 000664) through tail vein (1 million cells per mouse).Three days after injection, mice were randomly divided into 5 groups (n = 10 mice per group) and treated as described above.Weights were monitored daily.After 4 weeks of treatment, 5 mice from each group were sacrificed.Mouse lungs were inflated with 4% PFA and fixed in 4% PFA for FFPE blocks.The other 5 mice were monitored for up to 2 months, without further treatment, to evaluate tumor regrowth.

Figure 2 .
Figure 2. Multiomics analysis of effects of RMC-6272 versus rapamycin.(A) Gene set enrichment analysis (GSEA) of RNA data comparing RMC-6272-and rapamycin-treated TSC1-null HCV29 cells.The most downregulated pathways after RMC-6272 treatment are shown.(B) Seventy-three cell cycle genes were decreased in mRNA expression (FPKM) in RMC-6272-treated HCV29 cells compared with DMSO-or rapamycin-treated cells.(C) Purine metabolism was the metabolic pathway showing the most significant decrease in metabolite levels following RMC-6272 compared with rapamycin treatment in HCV29 TSC1-null cells, by MSEA.(D and E) Lipid map indicating lipid changes comparing RMC-6272-with rapamycin-treated HCV29 TSC1-null cells.Green nodes correspond to active lipids and arrows to conversion pathways.Z scores were used to assess significance.Reactions with a positive z score have green arrows, while negative z scores are colored purple.(F) Global proteomics analysis of RMC-6272-versus rapamycin-treated HCV29 TSC1-null cells led to identification of proteins involved in protein localization, targeting, and folding as being most downregulated with RMC-6272 treatment compared with rapamycin.(G) Global phosphoproteomics analysis of RMC-6272-versus rapamycin-treated 705 Tsc2-null cells identified ribonucleoprotein biogenesis and RNA splicing pathways enriched for phosphoproteins that were downregulated with RMC-6272 treatment compared with rapamycin using GSEA.

Figure 5 .
Figure 5. Single-dose assessment of bi-steric inhibitors in the Tsc2 +/-A/J mouse kidney cancer model.(A) Schematic diagram of the single-dose treatment strategy.(B) Immunoblotting of mTORC1 signaling pathway in whole kidneys after single dose, harvested at different time points.(C) Quantification of tumor cell apoptosis marker TUNEL.(D) Plasma concentrations of different compounds after 1 dose.(E) Quantification of proliferation marker Ki-67.(F) Measurement of kidney tumor volume in each mouse kidney.Different tumor sizes reflected the treatment effect and individual variation.(G and H) T cell (CD3) and macrophage (F4/80) infiltration.CD3 was used as a pan-T cell marker.F4/80 was used as a macrophage marker.(I) H&E and IHC staining of kidney tumors.For C-H, each bar and error bar represent mean ± SD (n = 6)

Figure 6 .
Figure 6.Bi-steric compounds showed more dramatic tumor suppression and less tumor regrowth in the Tsc2 +/-A/J mouse kidney cancer model.(A) Schematic diagram of treatment strategy.(B) Tumor volume from semiquantitative analysis of H&E slides immediately after treatment or 2 months after treatment.Dots are individual values (n = 6); a median line is shown.One-way ANOVA was used.*P < 0.05, **P < 0.01, ***P < 0.001.(C) Toxicity evaluation of compounds as judged by body weight.Each dot represents median (n = 6).(D and E) Images of whole mouse kidneys (D) and H&E-stained kidney sections (E) after 4 weeks of treatment and another 2-month tumor regrowth from mice as in A, both immediately after treatment and after 2-month regrowth.Scale bar: Mm ruler (D) and 5 mm (E).(F) H&E and IHC staining of kidney sections after 4 weeks of treatment and after 2 months of tumor regrowth.Scale bar: 30 μm.