Inhibition of glutaminase elicits senolysis in therapy-induced senescent melanoma cells

Abstract

The cyclin D1-Cyclin-Dependent Kinases 4 and 6 (CDK4/6) complex is crucial for the development of melanoma. We previously demonstrated that targeting CDK4/6 using small molecule inhibitors (CDK4/6i) suppresses Braf^V600E melanoma growth in vitro and in vivo through induction of cellular senescence. However, clinical trials investigating CDK4/6i in melanoma have not yielded successful outcomes, underscoring the necessity to enhance the therapeutic efficacy of CDK4/6i. Accumulated research has shown that while senescence initially suppresses cell proliferation, a prolonged state of senescence eventually leads to tumor relapse by altering the tumor microenvironment, suggesting that removal of those senescent cells (in a process referred to as senolysis) is of clinical necessity to facilitate clinical response. We demonstrate that glutaminase 1 (GLS1) expression is specifically upregulated in CDK4/6i-induced senescent Braf^V600E melanoma cells. Upregulated GLS1 expression renders Braf^V600E melanoma senescent cells vulnerable to GLS1 inhibitor (GLS1i). Furthermore, we demonstrate that this senolytic approach targeting upregulated GLS1 expression is applicable even though those cells developed resistance to the Braf^V600E inhibitor vemurafenib, a frequently encountered substantial clinical challenge to treating patients. Thus, this novel senolytic approach may revolutionize current CDK4/6i mediated melanoma treatment if melanoma cells undergo senescence prior to developing resistance to CDK4/6i. Given that we demonstrate that a low dose of vemurafenib induced senescence, which renders Braf^V600E melanoma cells susceptible to GLS1i and recent accumulated research shows many cancer cells undergo senescence in response to chemotherapy, radiation, and immunotherapy, this senolytic therapy approach may prove applicable to a wide range of cancer types once senescence and GLS1 expression are induced.

Fig. 1. KGA expression is increased in palbociclib-induced senescent cells.

A–D WM3918 (Braf^WTNras^WT), WM3912 (Braf^WTNras^WT), A375 (Braf^V600ENras^WT), 1205Lu (Braf^V600ENras^WT), Skmel28 (Braf^V600ENras^WT), Skmel2 (Braf^WTNras^Q61R), WM3451 (Braf^WTNras^Q61K), and normal human melanocytes (HEMa) were treated with DMSO or palbociclib (1 μM) for 1 day or 8 days. Treated cells were subjected to SA-βgal assay (8 days treatment only) and SA-βgal positive cells were quantified. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). A and RT-qPCR analysis for KGA (B), GAC (C) and GLS2 (D). RT-qPCR data (B–D) were normalized by GAPDH and represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). E 1205Lu and Skmel28 cells treated with DMSO (Control) or palbociclib (1 μM) for 8 days were subjected to western blot for GLS1 and βactin. An arrowhead indicates KGA expression. F 1205Lu cells treated with DMSO (Control) or palbociclib (1 μM) for 1 day or 8 days and 4 clones of 1205Lu palbociclib resistant cells (PR_1, PR_2, PR_3 and PR_4) maintained in the palbociclib-containing medium were harvested and subjected to RT-qPCR analysis for KGA expression. Data were normalized by GAPDH and represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). G 1205Lu, Skmel28 and Skmel2 cells were treated with DMSO or palbociclib for 8 days, and then treated with actinomycin D (0.5 μg/ml) for 2 h. Harvested cells were subjected to RT-qPCR analysis for KGA expression. Data were normalized by GAPDH and represent mean ± SD, p < 0.05 (two-tailed Student’s t-test; n = 3). NS indicates no significance (n = 3).

Fig. 2. CB-839 induces senolysis in palbociclib-induced senescent cells.

A Scheme for experimental schedule. B WM3918 (Braf^WTNras^WT), WM3912 (Braf^WTNras^WT), 1205Lu (Braf^V600ENras^WT), Skmel28 (Braf^V600ENras^WT), Skmel2 (Braf^WTNras^Q61R), WM3451 (Braf^WTNras^Q61K), HEMa_1 (normal human melanocytes) and HEMa_2 (normal human melanocytes) were treated with DMSO (Control) or palbociclib (1 μM) for 1 day or 8 days followed by CB-839 (50 μM) for an additional 3 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometry. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3), NS indicates non-significant (n = 3). C 1205Lu cells treated with DMSO (Control) or palbociclib (1 μM) for 1 day or 8 days followed by CB-839 (50 μM) for an additional 3 days were subjected to SA-βgal assay. D Quantification of SA-βgal positive cells from (C) was shown; data represent mean ± SD. E Skmel28 cells were treated with DMSO (Control), palbociclib (0.25 μM) for 8 days, CB-839 (50 μM) for 3 days, or palbociclib (0.25 μM) for 8 days followed by CB-839 (50 μM) for an additional 3 days. Colony outgrowth was stained with crystal violet after culturing cells for 35 days, including the treatment period. The numbers represent mean ± SD (n = 4).

Fig. 3. Knockdown of GLS1 induces senolysis in palbociclib-induced senescent Braf^V600E melanoma cells.

A 1205Lu and Skmel28 cell lines, transduced to stably express inducible short hairpin RNAs targeting KGA (shKGA_1 and shKGA_2), were treated with or without doxycycline (100 ng/ml) for 5 days and subjected to western blot for GLS1 and βactin. Arrowheads indicate KGA expression. B 1205Lu and Skmel28 cells from (A) were treated with DMSO (Control), palbociclib (1 μM) for 13 days, or palbociclib (1 μM) for 8 days followed by doxycyclin for an additional 5 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometry. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). C 1205Lu and Skmel28 cells introduced with overexpression of KGA were subjected to western blot for GLS1 and βactin. An arrowhead indicates KGA expression. D 1205Lu and Skmel28 cells introduced with or without overexpression of KGA were treated with DMSO (Control) or palbociclib (1 μM) for 8 days. Treated cells were subjected to SA-βgal assay and SA-βgal positive cells were quantified. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). E 1205Lu and Skmel28 cells introduced with overexpression of KGA were treated with DMSO (Control) or palbociclib (1 μM) for 8 days followed by CB-839 (50 μM) for an additional 3 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometry. Data represent mean ± SD, NS indicates non-significant (n = 3).

Fig. 4. CDK4/6i and GLS1i induces senolysis in vemurafenib-resistant melanoma cells.

A Vemurafenib-resistant cells derived from 1205Lu and 983B (1205Lu_VR and 983B_VR) treated with DMSO or palbociclib (1 μM) for 1 day or 8 days were subjected to RT-qPCR analysis for KGA expression. Data were normalized by GAPDH and represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). B 1205Lu_VR and 983_VR cells treated with DMSO (Control) or palbociclib (1 μM) for 8 days were subjected to western blot for GLS1 and βactin. C 1205Lu_VR and 983B_VR cells were treated with DMSO (Control) or palbociclib (1 μM) for 1 day or 8 days followed by CB-839 (50 μM) for an additional 3 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometry. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). D 1205Lu_VR and 983B_VR cells were treated with palbociclib (1 μM) for 8 days and then cultured in a medium with or without glutamine for an additional 3 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometry. Data represent mean ± SD, NS indicates non-significant (n = 3).

Fig. 5. CDK4/6i and GLS1i induces senolysis in vivo.

A A total of 1 × 10⁶ cells of 1205Lu_VR cells were subcutaneously injected into 6-week-old SCID mice. Tumor-inoculated mice were treated with palbociclib (90 mg/kg) by oral gavage or CB-839 (200 mg/kg) by oral gavage after tumors reached 5 mm in diameter. After 8 days, a combination of palbociclib (90 mg/kg) and CB-839 (200 mg/kg) was given by oral gavage for an additional 9 days. In each group, comprising 10 tumors, tumor volumes were measured every 2 days. Data represent means ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 10). B Representative images of H&E staining for harvested tumors from (A). The scale bar indicates 50 μm. C The number of mitotic cells was enumerated within a representative 1 mm² area in each image. Data represent median and interquartile range (IQR), *p < 0.05 (two-tailed Student’s t-test; n = 10). D The proportion of necrotic cells was quantified in each H&E-stained image. Data represent median and interquartile range (IQR), *p < 0.05 (two-tailed Student’s t-test; n = 10). NS indicates non-significant (n = 10). E Representative images of SA-βgal staining and IHC staining for GLS1 and cleaved Caspase-3 in the tumors from (A). The numbers indicate H scores by IHC scoring (see Materials and Methods) from 10 tumors. The scale bar indicates 50 μm. The red arrow indicates Cleaved Caspase-3 positive cells.

Fig. 6. GLS1i induces senolysis in vemurafenib-induced senescent cells.

1205Lu cells treated with DMSO (Control) or vemurafenib (0.5 μM) for 8 days were subjected to SA-βgal assay (A) and RT-qPCR analysis for KGA expression (B). RT-qPCR data (B) were normalized by GAPDH and represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3). C 1205Lu cells treated with DMSO (Control) or vemurafenib (0.5 μM) for 8 days were subjected to western blot for p-ERK, total ERK, p-S6, total S6 and βactin. D 1205Lu cells were treated with DMSO (Control) or vemurafenib (0.5 μM) for 8 days followed by CB-839 (50 μM) for an additional 3 days. Harvested cells were subjected to Annexin V staining and analyzed by flow cytometer. Data represent mean ± SD, *p < 0.05 (two-tailed Student’s t-test; n = 3).