MK-6482 as a potential treatment for von Hippel-Lindau disease-associated clear cell renal cell carcinoma.

MK-6482 targets one of the truncal downstream consequences of VHL functional loss, thereby impacting many of the effects seen with this mutation. Due to the efficacy seen in VEGF therapy refractory patients and in patients with VHL disease, its excellent safety and tolerability profile, and its oral formulation, MK-6482 is highly likely going to become a therapy of choice in patients with VHL disease, and may become a therapeutic option for individuals with treatment refractory, advanced sporadic RCC. Due to its appealing efficacy and safety profile, and convenient oral formulation, MK-6482 has a high likelihood of being prescribed by physicians who manage these diseases. One of the common challenges with anti-tumor agents is the development of resistance. Currently, the mechanisms of resistance to MK-6482 is under investigation. One of the proposed mechanisms is a G323E mutation in HIF2α protein that prevents MK-6482 binding 72, [ 76 ]. Another mechanism is F446L mutation in ARNT that increases the affinity of protein to HIF2α even in the presence of MK-6482. Further studies are required to delineate primary and acquired mechanisms of resistance to MK-6482 which will help in developing biomarkers predicting response and address resistance mechanisms. Another major challenge will be determing how MK-6482 interacts with other cell types and other therapies. Which cell populations outside of the tumor cell itself will be impacted by HIF-2α blockade? Is HIF-2α blockade synergistic, additive or antagonistic to antiangiogenic agents or checkpoint blocking antibodies? Appropriately designed, tissue rich studies are needed to fully understand the impact of HIF-2α blockade on the broader tumor microenvironment. In five years, this agent will likely be approved as monotherapy for the treatment of VHL disease, and for treatment refractory RCC. A key unmet need in VHL disease is prevention of lesion development, and a trial demonstrating that MK-6482 treatment stops new lesion formation in patients with VHL disease would be a significant advance. Beyond monotherapy, leveraging the effect of MK-6482 by testing rational combinations with agents targeting angiogenesis, cell cycle regulation and the immune microenvironment are needed to better understand how to further enhance the effectiveness of this agent. Additionally, the MTD was never found for this agent, and it is possible that dose escalation may further enhance the effectiveness of MK-6482. Lastly, there are few data on the effectiveness of MK-6482 in other cancers, and potential indications beyond RCC and VHL disease need to be explored. As our understanding of exactly how to use MK-6482 to modulate tumor, immune and endothelial cell populations and how combine this agent with others in our therapeutic armamentarium, its use in oncologic practice will undoubtedly continue to grow.

HIFα and its counterpart ARNT contain one basic helix–loop–helix (bHLH) and two Per-ARNT-Sim (PAS) domains, PAS A and B[ 42 ]. While the bHLH domain serves as a DNA-binding site[ 64 ], PAS domains are important for protein-protein interaction to form stable heterodimer of HIFα/ARNT [ 65 , 66 ]. The detailed structure of HIFα PAS domain was not well known until the studies of Bruick and Gardner laboratory were published. They showed that the C terminal PAS domain, PAS-B of HIF2α, directly binds to the ARNT PAS-B domain through the interaction of residues in the central strand of the beta-sheet. Supporting this finding, they demonstrated that point mutations of those residues in HIF2α abolished its binding to ARNT and abrogated hypoxic response [ 67 ]. Subsequent studies on understanding the HIF2α/ARNT PAS-B domain interaction with NMR and X-ray crystallography revealed a 290-Å 3 internal cavity buried within the HIF2α PAS-B domain [ 68 ]. This was a critical observation because the cavities of this large size are infrequent findings from the known high-resolution structures. This discovery further triggered the interest in searching for potential drug candidates disrupting the interaction of HIF2α/ARNT. Using a NMR-based ligand binding assay, a small-molecule library was screened, and THS-044 was the first compound that was identified which had the partial ability to disrupt HIF2α/ARNT heterodimer formation[ 68 ]. This molecule was not suitable for further characterization in cultured cells and led the investigators to use functional high throughput screening to search a larger library of molecules that could disrupt HIF2α/ARNT heterodimer formation. Subsequently, they discovered compounds that demonstrated the ability both in vitro and in cell culture to inhibit HIF2α/ARNT heterodimer formation, and resulted in a modest decrease in transcriptional activity [ 69 – 71 ]. Due to the unfavorable physical properties and modest cellular activity, these compounds were not translated in further clinical studies. However, this seminal works build the foundation of Peloton Therapeutics to discover more potent HIF2α inhibitors, PT2385 and PT2399, through a structure-based drug design (SBDD) approach, using previously discovered benzoxadiazole and tetrazole structured HIF2α PAS-B antagonists ( Figure 2A ) [ 72 , 73 ]. These compounds were tested in various in vivo RCC models and demonstrated an significant decrease in HIF2α target gene expression and had potent anti-tumor activity [ 72 , 74 , 75 ]. PT2385 was further tested in a phase I study [ 76 ]. In the dose escalation cohort, 26 heavily pretreated advanced RCC patients were enrolled. No dose-limiting toxicity was identified. Based on the safety, pharmacokinetic, and pharmacodynamics analyses, the RP2D was defined as 800mg by mouth twice per day. 25 mRCC patients were treated in the dose expansion cohort. The most common toxicities were anemia, peripheral edema and fatigue, and overall the treatment was well tolerated. Based on the PK/PD studies, there was high variability in drug exposure between patients. Not surprisingly, the objective response rate was only 14%. However, the patients with higher drug exposure had better efficacy, evidenced by improved progression-free survival. To investigate this inter-patient variability, further pharmacokinetic studies demonstrated that the efficacy of PT2385 might be limited in some patients due to the extensive glucuronidation in the enterocytes with the UGT2B17 enzyme and conversion to the PT2639 (PT2385-glucuronide) metabolite. Subsequently, the investigators modified PT2385 and changed its germinal difluoro group to the vicinal difluoro group, which decreased its glucuronidation and improved its pharmacokinetics ( Figure 2B ) [ 77 ]. This new compound was called PT2977, later renamed MK-6482 after Merck acquired Peloton Therapeutics. Based on the VEGFA secretion assay, with this new modification, the potency of the compound also improved 2- to 3- fold. Additionally, it had decreased lipophilicity, which changed the plasma protein binding from 82% to 52%. Eventually, improving both potency and increasing the free fraction significantly improved the free fraction adjusted EC 50 from 95ng/mL to 13ng/mL in the VEGFA secretion assay [ 77 ]. In phase I dose-escalation study ( NCT02974738 ), 34 patients with advanced solid tumors were treated with 20 – 240 mg daily doses of MK-6482 [ 77 ]. Plasma concentrations were measured on day 1 and day 15 for pharmacokinetic (PK) and weekly for pharmacodynamic (PD) studies. At all doses above 40 mg daily, the steady-state plasma concentration was above the free fraction adjusted EC 85 of 75ng/mL from the VEGFA secretion essay. Doses above 120 mg daily did not provide additional exposure benefit ( Figure 3A ). On Day 15, 120 mg MK-6482 had a T 1/2 of 20.9 hours, C max of 1.8 ug/mL, an AUC last of 25.9 ug/mL and a C 24h /C min of 0.71 ug/mL. In other words, at 24 hours, MK-6482 concentrations were 10-fold higher than the free fraction adjusted EC 85 of 75ng/mL from the VEGFA secretion essay. Additionally, on Day 15, MK-6482 glucuronide metabolite to parent AUC ratio was 0.32. Compared to the first generation HIF2a inhibitor, PT2385, this was a significant improvement in metabolism, when compared to the PT2385 metabolite/parent AUC ratio at Day 15 of 6. To evaluate PD response, HIF2α target erythropoietin (EPO) protein expression was measured in plasma. With a 40 mg daily dose, on day 8 there was an approximate 50% reduction in the EPO expression. However, with a 120 mg daily dose, day 8 EPO expression was reduced by nearly 80%. Doses above 120 mg did not provide a significant change in the PD effect ( Figure 3B ). In terms of safety, the most common adverse events (AEs) included anemia (35%), fatigue (24%), edema (17%), headache (17%), and nausea (17%) [ 78 ]. Altogether, with a feasible PK/PD profile and a favorable safety profile, 120 mg PO daily was determined to be the RP2D for MK-6482. In the dose-escalation part of the study, 6 mRCC patients were treated with 120 mg PO daily and above. At the data cutoff of March 31, 2018, among those 6 mRCC patients, 1 patient had partial response and 4 had stable disease. There are two phase II studies that have been reported to date with MK-6482. One enrolled patients with advanced sporadic RCC, and the other enrolled patients with VHL disease- associated RCC. In the NCT02974738 trial, the dose-expansion cohort enrolled 55 patients with previously treated advanced ccRCC patients and treated them with 120 mg PO daily MK-6482 [ 79 – 81 ]. The primary end-point of the study was safety, and key secondary end-points included objective response rate (ORR), duration of response (DOR) and progression-free survival (PFS). The patients enrolled in this study were heavily pretreated: 15% had one, 24% had two, and 62% had ≥ 3 prior systemic treatments. 91% of the patients were previously treated with anti-VEGF/VEGFR agents, 80% previously received immune checkpoint inhibitors, and 71% of the patients received both anti-VEGF/VEGFR and immune checkpoint inhibitor therapy. The majority of the patients in this cohort, 76%, were intermediate- or poor-risk mRCC patients per International Metastatic RCC Database Consortium (IMDC) criteria. The preliminary results of the study were presented in the ESMO Annual meeting in 2019 and at the ASCO Genitourinary Malignancies meeting in 2020 and 2021. At the time of data cutoff, on June 1, 2020, all patients had treatment-related AEs, and the most common AEs at all severity were anemia (76%), fatigue (71%), dyspnea (49%), and nausea (36%). The most common grade 3 AEs were anemia (27%), hypoxia (16%), fatigue (5%), and dyspnea (5%). Two patients (4%) had treatment discontinuation due to treatment-related AEs. Four patients (7%) experienced grade 5 AEs, due to disease progression, acute kidney injury, cardiac arrest and died. None of the patients had died due to treatment-related AEs. In terms of efficacy, 24% of the patients had a partial response, and 56% had stable disease. The clinical activity was observed across IMDC risk categories. At the median follow-up of 27.7 months, 80% of patients discontinued the treatment, and 20% had ongoing treatment. The median duration of response was not reached, and the median PFS was 14.5 months. Overall, these efficacy results are quite remarkable for the heavily pretreated mRCC patient population represented in this study, and meets an unmet need for effective treatment after the progression on multiple lines of treatment including, anti-VEGF/VEGFR and immune checkpoint blockade. Another phase II study with MK-6482 tested the efficacy of this treatment in VHL disease ( NCT03401788 ) [ 82 – 84 ]. The primary end-point of the study was ORR in VHL disease-associated RCC tumors and key secondary end-points were ORR in non-RCC lesions, DOR in RCC and non-RCC lesions and safety. Sixty-one patients with VHL disease were enrolled on this study and treated with 120 mg MK-6482 PO daily. Eighty two percent of the patients had central nervous system (CNS) hemangioblastoma, 100% had pancreatic lesions and 26% had retinal lesions. Fifty-three percent of the patients had undergone prior partial nephrectomy for RCC. The most recent preliminary results of the study was presented at the ASCO Genitourinary Cancers Symposium 2021. At the data cutoff of June 1, 2020, 92% of the patients were receiving ongoing treatment in the study. Three patients had treatment discontinuations for various reasons, including personal decision, adverse event, and death. Per RCC lesion evaluation, 36% of the patients had a confirmed partial response, an additional 11% had unconfirmed PRs, and 62% had stable disease. The median time to response was 31 weeks and the median duration of response was not reached with a 69 week median follow up. As the secondary end-point of the study, ORR in non-RCC lesions was evaluated. The ORR was 64% in pancreatic lesions, with a 6.6% CR and a 57.4% PR. Among 43 patients with CNS hemangioblastoma, 30% had an ORR with a 11.6% CR and a 18.6% PR. Among 16 patients with retinal lesions, 68.8% showed improvement, and 25% had stable disease. In terms of safety, similar to the other studies with MK-6482, the most common all-grade AEs were anemia (90%), fatigue (60%), headache (38%), dizziness (36%), nausea (31%) and dyspnea (20%). Eight patients (13%) had grade 3 or higher treatment-related AEs. One patient discontinued treatment due to treatment-related grade 1 dizziness. Two patients observed grade 4/5 adverse events that were not treatment-related: one developed retinal detachment and the other died from a narcotic overdose. Overall, this study concluded that MK-6482 is a well tolerated and efficacious treatment for VHL disease related RCC and non-RCC tumors. There are currently five ongoing clinical trials testing the safety and efficacy of MK-6482 with different combinations, doses and treatment settings on metastatic RCC ( Table 3 .). One of these is a phase I study, evaluating the safety of abemaciclib, a cyclin dependent kinase 4/6 inhibitor, alone and in combination with MK-6482. Two of these are phase 2 studies, one evaluating the efficacy of MK-6482 at two different doses and the other testing the combination of MK-6482 with cabozantinib in the frontline setting and after immunotherapy in metastatic RCC patients. There are two phase 3 studies: one testing the efficacy of MK-6482 in comparison with everolimus in the salvage treatment setting, and the other comparing MK-6482 plus lenvatinib with cabozantinib. On July 29, 2020, Federal Drug Administration has granted a breakthrough therapy designation to MK-6482 for the treatment of patients with von Hippel-Lindau disease–associated RCC who have nonmetastatic tumors of less than 3 centimeters, unless immediate surgery is necessitated.

With a strong biological rationale, HIF2α inhibitors demonstrated anti-tumor response of various in vitro and in vivo RCC models. Through a rigorous process of lead compound identification, iteration and optimization, MK-6482 was developed. MK-6482 has been tested in phase I studies on advanced solid tumors and shown to be safe to use in further studies. In a phase II study of heavily pretreated sporadic RCC patients with refractory mestastatic disease, it demonstrated significant clinical activity. Lastly, in the phase II study patients with VHL disease related RCC, MK-6482 achieved robust clinical activity in RCC, hemangioblastomas, pancreatic lesions and retinal hemangioblastomas. MK-6482 appears to be a safe and effective therapy for the treatment of VHL patients. Ongoing studies will define the role of MK-6482 in the treatment of sporadic, advanced, sporadic RCC either as monotherapy or in combination with agents that target angiogenesis, or the cell cycle.

Von Hippel-Lindau (VHL) disease is an inherited autosomal dominant syndrome caused by a germline mutation and/or deletion of the VHL gene. Affected individuals are at risk of developing tumors and cysts in multiple organs, such as hemangioblastomas of the brain and spinal cord, inner ear endolymphatic sac tumors, retinal angiomas, renal cysts and renal cell carcinomas (RCC), pheochromocytomas, pancreatic cysts and neuroendocrine tumors, and epididymal and broad ligament cystadenomas ( Table 1 ) [ 1 ]. Based on the phenotypic heterogeneity of the syndrome, VHL disease is classified into type 1 and type 2 according to the presence and absence of pheochromocytoma. Type 2 disease is further classified to 2A, 2B and 2C depending on the presence of RCC and hemangioblastoma ( Table 2 ) [ 2 – 5 ]. Among all VHL disease associated tumors, RCC has the highest incidence, up to 70%, and is a leading cause of death in VHL patients [ 6 ]. VHL associated RCCs are usually clear cell subtype and have a higher risk of metastasis once they reach 3 cm [ 7 – 9 ]. Patients with tumors ≤3 cm are usually monitored with active surveillance. Once the tumor reaches ≥3 cm, nephron-sparing partial nephrectomy is recommended. Ablative techniques, such as radiofrequency ablation and cryoablation, are also used in patients with smaller tumors who have high operative risk [ 10 ]. Complete nephrectomy is reserved for patients who have single or multiple kidney tumors that are not amenable for a partial nephrectomy or local ablation due to the location, size and number of tumors. Surgical management does cure VHL disease associated RCC but does not prevent new tumors from forming in the remaining or contralateral kidney. In fact, in a retrospective study, 5 of 21 patients (24%) developed local recurrence, and 2 developed metastatic disease with a median follow up of 29 months [ 11 ]. For patients who develop local recurrence, repeat surgeries with ablative techniques or partial/salvage nephrectomy are performed. A number of patients eventually develop renal insufficiency after multiple surgeries and require dialysis or undergo renal transplantation [ 12 ]. Overall, these manifestations limit the life expectancy of VHL disease patients with RCC either from developing end-stage renal disease or developing metastatic disease [ 13 ]. Once the patients develop metastatic disease from renal tumors, the management strategy is similar to that for sporadic metastatic RCC with systemic treatments targeting the vasculare endothelial growth factor (VEGF) pathway, and more recently, with immune checkpoint inhibitors [ 14 , 15 ]. Two VEGF receptor inhibitors were tested in prospective clinical trials specifically enrolling VHL disease patients. In a pilot study with sunitinib, 15 VHL patients were treated [ 16 ]. The primary end-point of the study was safety, and the key secondary end-point of the study was objective response rate. 12 (80%) of the patients had RCC, and a total 18 RCC lesions were evaluated. Fatigue, diarrhea, mucositis, anemia, hand-foot syndrome, and rash were the most common adverse effects, and 10/15 (67%) patients had a dose reduction due to toxicity. Six out of 18 (33%) lesions showed a partial response, 10/18 (67% stable disease and 2/18 (10%) developed progressive disease. In this study, no response was observed in other VHL disease- associated lesions. In another phase II study, 31 patients with VHL disease were treated with pazopanib [ 17 ]. The co-primary end-points of the study were objective response rate and safety. 69% of patients had RCC with a total of 59 renal tumors. 2/59 (3%) RCC tumors had a complete response, and 29/59 (49%) had a partial response. Regarding other VHL associated lesions, 2/49 (4%) of hemangioblastomas and 9/17 (53%) pancreatic tumors achieved partial response with pazopanib treatment. Overall, in this study, the response rate was 42%. Similar to the sunitinib study, diarrhea, fatigue, HTN, nausea were the most common adverse events. Additionally, elevated transaminases were commonly observed, which is known as a typical pazopanib related toxicity. 21/31 (67%) of the patients had a dose reduction due to toxicity. In summary, VEGF targeted agents have demonstrated clinical activity in VHL disease-associated RCC. However, due to the therapy-related adverse events, long-term treatment with these agents is challenging for patients and providers. In addition to that, non-RCC VHL disease-related tumors such as hemangioblastomas had no or limited response in these studies. Thus, there is an urgent need to develop more tolerable treatments for VHL disease patients that could demonstrated activity both in RCC and other lesions associated with this syndrome. The VHL gene is located at the chromosome 3p25–26 locus [ 18 ], and similar to other tumor suppressor genes, loss of both alleles is required for tumor formation in VHL disease [ 19 ]. The VHL gene encodes 2 isoforms of VHL protein (pVHL), pVHL30 and pVHL19 [ 20 – 22 ]. pVHL interacts with elongin B and elongin C and forms the VCB ubiquitin E3 ligase complex [ 23 – 25 ] along with cullin 2 and RING finger protein RBX1 [ 26 – 28 ]. Forming this complex stabilizes pVHL and all other subunits of the VCB complex [ 29 ]. In contrast, point mutations of VHL can alter the inherent stability of the pVHL protein, and decrease its interaction with Elongin C, enhancing its proteasomal degradation [ 23 – 26 ]. The first known substrate of the VCB complex is hypoxia-inducible factor 1 alpha (HIF1α). Polyubiquitination of HIF1α by the VCB complex results in its proteasome-mediated degradation [ 30 ]. Crystal structure analysis of the VCB and HIF1α complex demonstrated that interaction of VHL and HIF1α is dependent on hydroxylation of proline residues on the oxygen-dependent degradation domain (ODDD) of HIF1α [ 31 – 33 ]. Proline hydroxylation is performed by a group of dioxygenase enzymes called prolyl hydroxylases, which include PHD1, PHD2, and PHD3 [ 33 – 37 ]. The obligate co-factor of the hydroxylation reaction is oxygen, and therefore, the reaction happens only in normoxic conditions and causes the degradation of available HIFs. However, during hypoxia, the absence of hydroxylation prevents interaction with VHL, decreases ubiquitination and results in HIF protein accumulation. Similarly, in VHL disease or in sporadic RCC, defective pVHL is unable to form a VCB complex or recognize its ubiquitin-ligase targets, and therefore, HIFs accumulate in the cytoplasm ( Figure 1 ) [ 32 , 33 , 38 – 40 ]. Accumulated HIF1α forms a heterodimer with aryl hydrocarbon receptor nuclear translocator (ARNT), also known as HIF1β, and translocates to the nucleus where it binds hypoxia-response elements (HREs) to induce gene expression [ 41 , 42 ]. Currently, there are more than 800 recognized HIF target genes. Most of these are essential for the physiological response to hypoxic conditions [ 43 , 44 ], and may also drive tumorigenesis through various mechanisms, such as inducing angiogenesis through VEGF or platelet derived growth factor (PDGF) overexpression [ 45 , 46 ], increased glucose uptake and metabolism by upregulating the expression of glucose transporters (GLUT1 and GLUT3), hexokinase 2, phosphoglycerate kinase 1, lactate dehydrogenase A, phosphofructokinase 1 and pyruvate dehydrogenase [ 47 ], and inducing cell cycling with cyclin D1 overexpression [ 48 ]. There are 3 different types of HIFs, and each has some common and subtype-specific targets [ 49 ]. This makes it more difficult to conclude which one is more important in RCC pathogenesis. Nevertheless, accumulating evidence suggests HIF2α functions as a tumor oncogene [ 50 – 57 ] and HIF-1α as a tumor suppressor [ 5 , 39 , 55 – 61 ]. Upregulation of HIF2α is seen both in RCC and also in VHL associated pheochromocytomas and paragangliomas [ 62 , 63 ]. Thus, targeting HIF-2α and inhibiting its transcription factor activity was explored as a novel treatment strategy for RCC and VHL.