HIF-2-Dependent Regulation of PTHrP and Paraneoplastic Hypercalcemia in Aggressive Clear-Cell Renal Cell Carcinoma

Abstract

Patients with renal cell carcinoma (RCC) and hypercalcemia (HC) have worse outcomes. HC often involves parathyroid hormone-related protein (PTHrP), and the role of hypoxia-inducible factor 2 (HIF-2) is incompletely understood. Leveraging RCC tumorgraft (TG) models of HC, which were characterized by tumor cell-autonomous inflammatory/immune signatures, we show that HIF-2 inhibition with PT2399 frequently normalized calcium, downregulated circulating PTHrP, and reduced HIF-2 binding to the PTHLH (PTHrP) promoter. Likely contributing to the selective induction of PTHrP in a subset of HIF-2-dependent tumors, the PTHLH locus was generally more accessible in TG(HC). However, PTHLH chromatin accessibility was grossly unaffected by PT2399, unlike elsewhere (including the EPO locus in a TG with paraneoplastic polycythemia). As in TGs, paraneoplastic HC in patients was associated with clear-cell (cc)RCC (and sarcomatoid/rhabdoid differentiation) and was rapidly corrected by PT2977/belzutifan, which unlike bisphosphonates downregulated PTHrP. Our data support evaluating HIF-2 antagonists for ccRCC patients with paraneoplastic HC, which may serve as a predictive biomarker.

Significance: This study uncovers a direct role for HIF-2 in driving humoral HC of malignancy in ccRCC through transcriptional activation of PTHLH (encoding PTHrP), identifies HC (and PTHrP) as potentially predictive biomarkers of HIF-2 engagement, and sets a foundation for the evaluation of HIF-2 antagonists for HC management in patients.

Figure 1-. Hypercalcemia management by PT2399 in RCC tumorgraft-bearing mice

a- Schematic with study design. b- Line graph showing serum calcium levels in hypercalemic TG lines at baseline and at the end of the drug trial (dotted line represents ULN of serum calcium in mice). TG lines labelled in green (but not brown) normalized calcium levels upon PT2399 treatment. (Two XP1169d mice, one assigned to vehicle and another to PT2399, had normal calcium levels at treatment onset.) c- Overall fold change in serum calcium level in mice treated with PT2399 (n=27) compared to vehicle (n=34). d- Bar graph showing circulating PTHrP protein levels in vehicle and PT2399-treated TG-bearing mice. e- Mean percent change in PTHrP protein level in mice treated with PT2399 compared to vehicle. f- Bar graph showing relative PTHrP mRNA expression levels in vehicle and PT2399-treated TGs. g- Bar graph showing relative mRNA expression levels for CCND1 and VEGFA in vehicle and PT2399-treated TGs. h- Line graph showing tumor volume in TG(HC) lines treated with vehicle (black) or PT2399 (pink). (a, Created with BioRender.com). Each point in the bar graphs represents an individual mouse. Data are presented as mean ± SEM. P values as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns, non-significant.

Figure 2-. PT2399 alleviates hypercalcemia-related paraneoplastic symptoms

a- Line graph showing mouse body weight of vehicle (black) and PT2399-treated (pink) mice bearing hypercalcemic TGs. b- Bar graph showing overall percent change in mouse weight of TG(HC) mice treated with PT2399 with respect to vehicle. c- Representative photograph showing distance travelled in an open field test with corresponding bar graph of vehicle and PT2399-treated mice. d- Bar graph showing overall percent change in distance travelled by hypercalcemic mice treated with PT2399 in comparison to vehicle. e- Representative kidney H&E photomicrographs from vehicle and PT2399-treated hypercalcemic mice. Red arrow points to calcium deposition. f- Semi-quantitative evaluation of dystrophic calcium deposition on kidney H&E sections from tumorgraft-bearing mice treated as shown in panel e. g- Bar graph with percent change in dystrophic calcium deposition in kidneys from mice treated with PT2399 or vehicle. h- Bar graph showing semi-quantitative evaluation of renal injury on H&E sections of kidneys from TG-bearing mice treated as shown. i- Bar graph with percent change in renal injury scores from mice treated with PT2399 or vehicle. Each point in bar graphs represents an individual mouse. Data are presented as mean ± SEM. P values as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns, non-significant

Figure 3-. PT2399 promptly corrects hypercalcemia

a- b- Bar graph showing PTHrP tumor mRNA expression (2^−ΔCT) (a) and serum protein (b) levels in TG-bearing mice after 4 days of treatment as indicated. c- Line graph showing serum calcium levels in TG(HC) mice at baseline and after 4 days of vehicle (black) or PT2399 treatment (pink). d- Bar graph showing percent change in serum calcium levels with PT2399 treatment for 4 days. Each point in bar graphs represents an individual mouse. Data are presented as mean ± SEM. P values as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ***P < 0.0001 and ns, non-significant

Figure 4-. PT2399 disrupts HIF-2 mediated PTHrP (PTHLH) transcription

4a- Bar graph showing PTHrP mRNA levels in vehicle and PT2399 treated (5μM for 24-hours) 786-O cells (biological replicates, n=2). b- Bar graph showing the PTHrP Gluc reporter activity in 786-O cells from samples treated as in (a) (biological replicates, n=3). c-d DNA electropherogram of PTHrP promoter region (P2 promoter (11)) following HIF-2α chromatin immunoprecipitation in 786-O cell line (c) and XP771b and XP1087 TGs from mice treated with either vehicle or PT2399 (d). Each dot in the bar graph represents an individual mouse. Data are presented as mean ± SEM. P values as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns, non-significant

Figure 5-. Paraneoplastic hypercalcemia and HIF-2α as predictive biomarkers of response to HIF-2 antagonists

a- Bar graph with percentage of mice with tumor growth response to HIF-2 antagonist for each individual TG line segregated according to their ability to induce HC. Odds ratios with confidence intervals shown at the top. b- Representative photomicrograph illustrating HIF-2α induction by immunohistochemistry in TG(HC). c- Violin plot of HIF-2α scores (product of HIF-2α staining intensity and percentage of positive cells) in normocalcemic and hypercalcemic TG tissue microarray (19). d- Violin plot showing HIF-2α IHC scores in normocalcemic and hypercalcemic TG slides. (The hypercalcemia cohort is from vehicle treated mice in this study and the normocalcemia cohort comprises vehicle treated mice from this study as well as other studies and other untreated TGs.) e- Illustrative photomicrographs of H&E-stained sarcomatoid TGs along with HIF-2α immunohistochemistry. f- Violin plot of HIF-2α IHC scores in sarcomatoid/rhabdoid normocalcemic and hypercalcemic TGs. g- Representative photomicrograph of HIF-2α IHC in XP490 (normocalcemic) and XP771b (hypercalcemic) TGs. h- [¹⁸F]PT2385 PET scan in a mouse simultaneously implanted with XP771b (hypercalcemic TG with high HIF-2α) and XP490 (normocalcemic TG with low HIF-2α).

Figure 6-. Transcriptomic and ATAC-seq profiles of hypercalcemic and normocalcemic tumorgrafts

a- Volcano plot of deregulated genes in hypercalcemic TGs (cutoffs |LogFC|≥1 & FDR < 0.0001). b- Heatmap of genes differentially regulated between hypercalcemic and normocalcemic TGs. c- Venn diagram of upregulated genes in TG(HC) (HyperCaUP), and PT2399 downregulated genes in in TG(HC) (HyperCa PT Down). d- Venn diagram of upregulated genes in TG(HC) (HyperCaUP), and increased ATAC-seq peaks in promoter regions (HyperCa ATAC Up). e- Chromatin accessibility at PTHLH (PTHrP), CCND1 and VEGFA genes assessed by ATAC-seq in hypercalcemic (green) and normocalcemic (violet) TGs. Red highlighted area corresponds to sequences for luciferase reporter assay with highlighted line indicating the HIF-2 binding site evaluated by ChIP. The most significant promoter peak for each gene, based on the lowest FDR value in HC compared to normocalcemic TGs, is labeled in gray on the plot. f- Venn diagram of genes with reduced ATAC-seq peaks (±3kb of TSS) and gene expression by PT2399 in TG(HC). g- Chromatin accessibility at PTHLH (PTHrP), CCND1 and VEGFA genes assessed by ATAC-seq after vehicle (black) and PT2399 (pink) treatment. Blue highlighted area corresponds to sequences used in luciferase reporter assay and the highlighted line represents the HIF-2 binding site evaluated by ChIP. The most significant promoter peak for each gene, based on the lowest FDR value in TG(HC) downregulated by PT2399, is labeled gray on the plot. N.A., not applicable.

Figure 7-. Patient analyses and translational implications

a- Flow chart of RCC patients with paraneoplastic hypercalcemia and nephrectomy. b- Line graph showing corrected calcium levels pre and post nephrectomy (n= 26 patients). c- Pathological characteristics of normocalcemic (n=2215) and hypercalcemic (n=26) tumors. d- Schematic of case study patient with ccRCC metastasis to lung and adrenal. e- Computed tomography (CT) and Magnetic Resonance Imaging (MRI) of pulmonary and adrenal metastases over treatment course. Longest diameter (LD) of target lesions marked by white bar. f- Graph illustrating LD changes, corrected calcium and PTHrP levels over time. (d, Created with BioRender.com)