Research ArticleCancer therapy

HIF-independent synthetic lethality between CDK4/6 inhibition and VHL loss across species

See allHide authors and affiliations

Science Signaling  01 Oct 2019:
Vol. 12, Issue 601, eaay0482
DOI: 10.1126/scisignal.aay0482
  • Fig. 1 RNAi screen for genes that are synthetically lethal with vhl inactivation in Drosophila S2R+ cells.

    (A) Relative mRNA expression for sima, the D. melanogaster ortholog of the human gene encoding HIFα, and the indicated sima-responsive genes in vhl-null S2R+ cells as compared to wild-type (WT) S2R+ cells. Data are means ± SD of n = 2 independent experiments. (B) Z scores for change in viable cell number, as determined by CellTiter-Glo assays, after a 5-day incubation with dsRNAs (three dsRNAs per gene on average, 448 genes) in vhl-null S2R+ (x axis) and WT S2R+ (y axis) cells. Each dot represents the median Z score (n = 3 biological replicates) for one dsRNA. dsRNAs targeting the pan-essential Drosophila gene thread are indicated in red; those targeting cdk4 are indicated in blue. (C) Quantification of select data in (B). LacZ and GFP dsRNAs are negative controls that do not affect cell viability. Data are means ± SD of n = 3 independent experiments. (D) Top hits from (B), based on the top-scoring dsRNAs for each gene. A hit was defined as a gene targeted by a dsRNA with Z < −1.5 in at least two-thirds of replicates in EV cells and not more than one-third of VHL cells.

  • Fig. 2 Small-molecule screen for chemicals that are synthetically lethal with VHL inactivation in ccRCC cell lines.

    (A) Immunoblot of HIF-2α, VHL, and vinculin (loading control) abundance in parental VHL−/− 786-O and UMRC-2 cells and those stably infected with lentivirus expressing GFP and VHL (VHL) or GFP alone (EV), as indicated. Blots are representative of three biological replicates. (B) Z scores assessing the change in viable cell number, as determined by GFP fluorescence, after a 48-hour incubation with DMSO, epothilone B (410 nM), AT7519 (1.23 μM), or flavopiridol (137 nM) in the indicated cell lines. Data are means ± SD of n = 2 independent experiments. (C) Top-scoring drugs based on differential Z scores (VHL − EV) in 786-O and UMRC-2 cells and their putative protein targets. Yellow highlighting indicates targets that were interrogated in the Drosophila dsRNA screen (Fig. 1) but were not hits in that screen. Green highlighting indicates the targets human CDK4/6 (ortholog cdk4) and, to a lesser extent, CDK2 (ortholog cdk2) that were hits in the Drosophila dsRNA screen. (D) Venn diagram showing overlap of the human orthologs of the Drosophila dsRNA screening library and the genes encoding the protein targets of the ccRCC drug screen library. Shaded regions indicate screen hits. The data behind this diagram are in data files S1 to S3.

  • Fig. 3 The CDK4/6 inhibitor palbociclib preferentially inhibits pVHL-deficient ccRCC cells in an on-target manner.

    (A) Ratio of 786-O cells stably infected with a bicistronic lentivirus expressing VHL and Tdtomato (VHL-Tdtomato) to 786-O cells infected with GFP alone (EV-GFP) that had been mixed (1:1) and then treated with 0, 200, or 400 nM palbociclib for 3 to 10 days. Data are means ± SD of n = 4 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by two-way ANOVA. (B) Immunoblot of total and Ser780-, Ser608-, Ser795-, and Ser807/811-phosphorylated pRb in 786-O cells expressing VHL-Tdtomato or EV-GFP and treated with 100, 200, 400, or 800 nM palbociclib, as indicated by the triangle, for 24 hours. Blots are representative of three biological replicates. (C) 786-O cells that underwent CRISPR/Cas9 editing with an RB1 sgRNA and then were infected, mixed, and treated as in (A). The ratio of RB1-null VHL-Tdtomato cells to RB1-null EV-GFP cells after treatment is shown. Data are means ± SD of n = 3 independent experiments. (D) Immunoblot of Rb, VHL, and actin (loading control) abundance in 786-O cells edited with an RB1 sgRNA (as indicated, +) and infected as described in (C), but not otherwise treated. Blots are representative of three biological replicates. (E) Ratio of 786-O cells stably expressing CDK6(D104S) and either VHL-Tdtomato or EV-GFP that had been mixed and treated as described in (A). Data are means ± SD of n = 3 experiments. (F) Immunoblot of 786-O cells stably infected with lentivirus expressing either WT or mutant (D104S) CDK6 and then treated with 50, 100, 200, 400, 800, or 1600 nM palbociclib, as indicated by the triangle, for 24 hours. Blots are representative of three biological replicates.

  • Fig. 4 Increased HIF-2α is neither necessary nor sufficient for the synthetic lethal relationship between CDK4/6 and VHL in ccRCC.

    (A) Immunoblot of HIF-2α, VHL, cyclin D1, and vinculin (loading control) abundance in 786-O cells that underwent CRISPR/Cas9 editing with a HIF-2α sgRNA (as indicated, +) and were then infected to express VHL and Tdtomato (VHL-Tdtomato) or GFP alone (EV-GFP). Blots are representative of three biological replicates. (B) Ratio of HIF-2α–null VHL-Tdtomato cells to HIF-2α–null EV-GFP cells that were mixed (1:1) and then treated with 0, 200, or 400 nM palbociclib for 4 to 10 days. Data are means ± SD of n = 3 independent experiments. **P < 0.01 and ****P < 0.0001 by two-way ANOVA. (C) Immunoblot of HIF-2α and actin (loading control) abundance in 786-O cells stably expressing VHL and Tdtomato (VHL-Tdtomato) or GFP alone (EV-GFP) and treated with vehicle (0) or 25, 50, 100, or 200 μM FG4592 for 36 hours. Blots are representative of three biological replicates. (D) Ratio of VHL-Tdtomato cells to EV-GFP 786-O cells that were mixed (1:1) and then treated with 0, 200, or 400 nM palbociclib with or without 100 μM FG4592 for 10 days. Data are means ± SD of n = 3 independent experiments.

  • Fig. 5 Palbociclib and PT2399 synergistically suppress cell viability of VHL-null cells in PT2399-sensitive, but not PT2399-insensitive, ccRCC cell lines.

    (A) Ratio of VHL-Tdtomato–expressing to EV-GFP–expressing 786-O cells that were mixed (1:1) and then treated with 0, 200, or 400 nM palbociclib with or without 2 μM PT2399 for 10 days. Data are means ± SD of n = 3 independent experiments. (B to D) As described for (A) in A498 (B), UMRC-2 (C), and 769-P (D) cells. (E) Relative mRNA expression for CCND1 in EV-GFP–expressing 786-O cells treated with 2 μM PT2399, 400 nM palbociclib, or the combination (as indicated) for 48 hours. Data are means ± SD of n = 2 independent experiments. (F to H) As described for (E) in A498 (F), UMRC-2 (G), and 769-P (H) cells. **P < 0.01, ***P < 0.001, and ****P < 0.0001 by two-way ANOVA.

  • Fig. 6 In vivo antitumor activity of palbociclib in VHL-null ccRCC.

    (A) Representative BLIs of orthotopic tumors formed by firefly luciferase–expressing 786-O cells before and after mice were treated with vehicle or palbociclib (65 mg/kg), dosed daily for 28 days by oral gavage. Images are representative of n = 10 or 9 mice, respectively. (B) Quantification of BLI at day 28 in mice described in (A) and in those treated daily by oral gavage with PT2399 (20 mg/kg) or both PT2399 (20 mg/kg) and palbociclib (65 mg/kg) (combo). Data are means ± SD overlaying the individual data points from n = 10, 9, 11, and 11, respectively. Photon counts on day 28 were normalized to those on day 0 for each mouse individually. ****P < 0.0001 by one-way ANOVA. (C) Kaplan-Meier survival curves for mice described in (B). “Rx” bar indicates duration of treatment. Log-rank (Mantel-Cox) test, P < 0.0001; log-rank test for trend, P = 0.0001. (D to F) As described in (A) to (C) using UMRC-2 cells. n = 11 mice (vehicle), 8 mice (palbociclib), and 6 mice (combo). *P < 0.05 (P = 0.0112) by one-way ANOVA (F); P = 0.0012 by log-rank Mantel-Cox test (F), in which log-rank test for trend showed P > 0.05.

Supplementary Materials

  • stke.sciencemag.org/cgi/content/full/12/601/eaay0482/DC1

    Fig. S1. Control experiments for competition experiments done with isogenic ccRCC cell lines treated with CDK4/6 inhibitors.

    Fig. S2. The CDK4/6 inhibitor palbociclib preferentially inhibits pVHL-deficient cells in various ccRCC cell lines.

    Fig. S3. Changes in proliferation of ccRCC cells after pVHL reconstitution does not account for differential sensitivity to CDK4/6 inhibition.

    Fig. S4. Individual knockdown of CDK4 or CDK6 does not differentially affect the viability of ccRCC cells based on VHL status.

    Fig. S5. PT2399 attenuates palbociclib-induced up-regulation of cyclin D1 abundance in HIF-2α–dependent, but not HIF-2α–independent, cell lines.

    Fig. S6. Effect of palbociclib, PT2399, and their combination on cyclin D1 and phospho-pRb abundance in vivo.

    Fig. S7. Growth of ccRCC orthotopic xenografts during treatment with vehicle, palbociclib, PT2399, or their combination.

    Fig. S8. Growth of HIF-2α inhibition–resistant VHL-null ccRCC orthotopic xenografts during treatment with vehicle, palbociclib, or the combination of palbociclib with PT2399.

    Fig. S9. Antitumor activity of abemaciclib in VHL-null ccRCC orthotopic xenografts.

    Fig. S10. Antitumor activity of palbociclib and abemaciclib in a ccRCC PDX model.

    Fig. S11. Schematic of analogous signaling mechanisms in breast cancer and ccRCC.

    Table S1. sgRNA oligonucleotides.

    Table S2. PCR primers.

    Data file S1. Results of screen for synthetic lethality with vhl inactivation using dsRNA library in Drosophila cells.

    Data file S2. Results of screen for synthetic lethality with VHL inactivation using chemical library in human 786-O and UMRC-2 ccRCC cells.

    Data file S3. Overlap between genes encoding targets of chemicals that scored in chemical screen and human orthologs of Drosophila genes that scored in dsRNA screen.

  • The PDF file includes:

    • Fig. S1. Control experiments for competition experiments done with isogenic ccRCC cell lines treated with CDK4/6 inhibitors.
    • Fig. S2. The CDK4/6 inhibitor palbociclib preferentially inhibits pVHL-deficient cells in various ccRCC cell lines.
    • Fig. S3. Changes in proliferation of ccRCC cells after pVHL reconstitution does not account for differential sensitivity to CDK4/6 inhibition.
    • Fig. S4. Individual knockdown of CDK4 or CDK6 does not differentially affect the viability of ccRCC cells based on VHL status.
    • Fig. S5. PT2399 attenuates palbociclib-induced up-regulation of cyclin D1 abundance in HIF-2α–dependent, but not HIF-2α–independent, cell lines.
    • Fig. S6. Effect of palbociclib, PT2399, and their combination on cyclin D1 and phospho-pRb abundance in vivo.
    • Fig. S7. Growth of ccRCC orthotopic xenografts during treatment with vehicle, palbociclib, PT2399, or their combination.
    • Fig. S8. Growth of HIF-2α inhibition–resistant VHL-null ccRCC orthotopic xenografts during treatment with vehicle, palbociclib, or the combination of palbociclib with PT2399.
    • Fig. S9. Antitumor activity of abemaciclib in VHL-null ccRCC orthotopic xenografts.
    • Fig. S10. Antitumor activity of palbociclib and abemaciclib in a ccRCC PDX model.
    • Fig. S11. Schematic of analogous signaling mechanisms in breast cancer and ccRCC.
    • Table S1. sgRNA oligonucleotides.
    • Table S2. PCR primers.
    • Legends for data files S1 to S3

    [Download PDF]

    Other Supplementary Material for this manuscript includes the following:

    • Data file S1 (Microsoft Excel format). Results of screen for synthetic lethality with vhl inactivation using dsRNA library in Drosophila cells.
    • Data file S2 (Microsoft Excel format). Results of screen for synthetic lethality with VHL inactivation using chemical library in human 786-O and UMRC-2 ccRCC cells.
    • Data file S3 (Microsoft Excel format). Overlap between genes encoding targets of chemicals that scored in chemical screen and human orthologs of Drosophila genes that scored in dsRNA screen.

Stay Connected to Science Signaling

Navigate This Article