There is a clear need to identify targetable drivers of resistance and potential biomarkers for salvage therapy for patients with melanoma refractory to the combination of BRAF and MEK inhibition. In this study, we performed whole-exome sequencing on BRAF-V600E–mutant melanoma patient tumors refractory to the combination of BRAF/MEK inhibition and identified acquired oncogenic mutations in NRAS and loss of the tumor suppressor gene CDKN2A. We hypothesized the acquired resistance mechanisms to BRAF/MEK inhibition were reactivation of the MAPK pathway and activation of the cell-cycle pathway, which can both be targeted pharmacologically with the combination of a MEK inhibitor (trametinib) and a CDK4/6 inhibitor (palbociclib). In vivo, we found that combination of CDK4/6 and MEK inhibition significantly decreased tumor growth in two BRAF/MEK inhibitor–resistant patient-derived xenograft models. In vitro, we observed that the combination of CDK4/6 and MEK inhibition resulted in synergy and significantly reduced cellular growth, promoted cell-cycle arrest, and effectively inhibited downstream signaling of MAPK and cell-cycle pathways in BRAF inhibitor–resistant cell lines. Knockdown of CDKN2A in BRAF inhibitor–resistant cells increased sensitivity to CDK4/6 inhibition alone and in combination with MEK inhibition. A key implication of our study is that the combination of CDK4/6 and MEK inhibitors overcomes acquired resistance to BRAF/MEK inhibitors, and loss of CDKN2A may represent a biomarker of response to the combination. Inhibition of the cell-cycle and MAPK pathway represents a promising strategy for patients with metastatic melanoma who are refractory to BRAF/MEK inhibitor therapy.

In early 2015, the combination of BRAF inhibitor (BRAFi; dabrafenib) and MEK inhibitor (MEKi; trametinib) treatment for patients with BRAF-mutant (V600E/K) metastatic melanoma was FDA approved. While the combination of BRAF/MEK inhibitors (BRAFi/MEKi) increased progression-free survival (PFS) as compared with BRAFi monotherapy long-term durable responses remain a challenge (1). The combination of BRAFi/MEKi has a low five-year PFS rate (19%), indicating that acquired resistance and tumor regrowth is inevitable for a majority of patients (2). The current understanding of resistance mechanisms in oncogene-addicted tumors, such as BRAF-V600E–mutant melanomas, can occur through genetic and epigenetic mechanisms. Such resistance mechanisms include gatekeeper mutations in the drug-binding site, activation of bypass signaling pathways, such as the PI3K pathway, and/or reactivation of the MAPK pathway (3). MAPK pathway reactivation in response to a BRAF and/or MEK inhibitor is a major mechanism of resistance and occurs through multiple mechanisms including genetic mutations in NRAS, KRAS, MEK1, and MEK2, in addition to amplifications, fusions or alternative splicing of BRAF (4, 5). Upon clinical resistance to BRAFi, approximately 20% of patients harbored activating mutations in NRAS or KRAS (4, 6). Mutant NRAS mediates MAPK signaling even in the context of BRAF inhibition through activation of CRAF, inducing dimerization with either BRAF or CRAF, resulting in trans-activation of the MAPK pathway (7). However, acquired resistance mechanisms following the combination of BRAFi/MEKi remain poorly defined.

The cell-cycle pathway is commonly dysregulated in melanoma, promoting aberrant cell proliferation, and may drive tumorigenesis in addition to mutations in the MAPK pathway (8). Genomic mutations affecting the cell-cycle pathway are found in up to 90% of patients with melanoma, including mutations or deletions of CDKN2A and amplification of the genes encoding CDK4/6 and Cyclin D1 (CCDN1; ref. 9). Mutations or deletions in CDKN2A resulting in decreased p16INK4A protein expression results in loss of negative feedback of CDK4 and increased cyclin D1-CDK4/6 complex activity, promoting aberrant cell-cycle progression and tumorigenesis (10). As melanoma progresses from benign nevi to metastatic melanoma, p16INK4A expression is dramatically decreased, highlighting a critical role for the cell-cycle pathway in melanoma development (11). Furthermore, the cell cycle is positively regulated by the MAPK pathway through transcriptional regulation of cyclin D1 by ERK (12). Preclinical studies have identified roles for oncogenic NRAS mutations mediating cell-cycle progression. A genetically engineered mouse model of melanoma driven by NRAS-Q61K, in a CDKN2A−/− background, demonstrated that MEK inhibition alone was not enough to fully inhibit NRAS-mediated downstream signaling through CDK4, and the combination of MEK and CDK4/6 inhibition led to superior tumor reduction in vivo (13). Clinically, alterations in CDK2NA correlate with decreased PFS in response to BRAFi monotherapy in melanoma (14). Similarly, in vitro studies have identified that overexpression of cyclin D1 and CDK4 mediate BRAFi resistance (15). Small-molecule inhibitors targeting CDK4/6, such as palbociclib, abemaciclib, and ribociclib, have shown efficacy in a variety of cancers, including melanoma, in preclinical and clinical studies (16). Loss of CDKN2A correlated with sensitivity to the CDK4/6i palbociclib in vitro, over a panel of 47 melanoma cell lines (17). However, in the context of acquired BRAFi/MEKi resistance, it remains unclear whether CDKN2A mutational status and p16INK4A protein expression correlates with sensitivity to CDK4/6i alone or in combination with MEKi.

In this study, our approach integrated computational and experimental methods to identify genetic mechanisms of resistance and validate rational combination therapies to treat patients with BRAF-V600E–mutant melanoma who have relapsed on BRAFi/MEKi combination therapy. First, we identified oncogenic mutations in NRAS and loss of CDKN2A at the time of BRAFi/MEKi resistance. We hypothesized acquired resistance to BRAFi/MEKi therapy are through both MAPK pathway reactivation via NRAS mutations and activation of the cell cycle pathway through CDKN2A loss. We further hypothesized these acquired mutations could be targeted pharmacologically with the combination of MEKi (trametinib) and CDK4/6i (palbociclib). Using two in vivo patient-derived xenograft (PDX) models, we validated that BRAFi/MEKi resistance could be abrogated by inhibiting both MEK and CDK4/6. Next, we recapitulated our in vivo PDX findings with BRAF-V600E–mutant melanoma cell line models of acquired resistance to BRAFi, and observed superior growth inhibition to the combination of CDK4/6i and MEKi compared with single-agent treatment. Mechanistic investigation demonstrated that CDKN2A loss in acquired resistant BRAFi cells enhanced the sensitivity to growth inhibition after treatment with single-agent CDK4/6i, and resulted in a synergistic response to the combination of CDK4/6i and MEKi. Overall, these results indicate that inhibition of the cell-cycle and MAPK pathway represents a promising strategy for patients with BRAFi/MEKi–resistant melanoma. Our results also identify that CDKN2A loss may represent a potential biomarker of response to the combination of CDK4/6i and MEKi in patients who have relapsed and developed resistance to BRAFi/MEKi.

Patient samples

Tissue and peripheral blood samples were collected from two patients with melanoma at the University of Colorado Cancer Center (UCCC) between 2010 and 2014. All sample collections were made with the patient's informed written consent and in accordance with the Colorado Institutional Review Board (COMIRB-05–0309) and the Declaration of Helsinki stored in the International Melanoma Biorepository and Research Laboratory at the University of Colorado Anschutz Campus.

Next-generation whole-exome sequencing and identification of somatic mutations

Whole-exome sequencing was performed on patient tumor tissue and peripheral blood samples as described previously (18). Briefly, genomic DNA was isolated using the DNeasy Blood and Tissue kit (Qiagen). Exome DNA libraries were constructed with Agilent SureSelect XT Target Enrichment System (Agilent Technologies) for Illumina Paired-End Multiplexed Sequencing Library (catalog no. G9641B; Illumina). Libraries were sequenced on the Illumina HiSeq 2000 with 125-bp paired-end reads. Data were analyzed for variants and copy number alterations using the IMPACT WES analysis pipeline (19).

Reagents

For in vivo studies, trametinib and palbociclib were purchased from LC Laboratories. For in vitro studies, trametinib, palbociclib, and vemurafenib were purchased from SelleckChem, and were dissolved in dimethyl sulfoxide at 10 mmol/L stocks.

Patient derived xenograft models and animal studies

All animal studies were conducted at the University of Colorado Anschutz Medical Campus in accordance with the NIH guidelines for the care and use of laboratory animals, and animals were housed in a facility accredited by the American Association for Accreditation of Laboratory Animal Care, and with the approval of the University of Colorado Institutional Animal Care and Use Committee (IACUC protocol number: 00021) before initiation. Patient derived tumors were collected from consented melanoma patients at the UCCC with approval by the COMIRB. The BRAFi/MEKi–resistant melanoma patient-derived xenografts were generated as previously described (20, 21). Female athymic nude (nu/nu) mice were purchased from Envigo at age 4–8 weeks. When tumors reached 100–300 mm3, tumor-bearing mice randomized and treated with the following for 30 days (n = 10 tumors per group): (i) vehicle, (ii) palbociclib 75 mg/kg oral gavage daily, (iii) ulixertinib 75 mg/kg oral gavage twice daily, (iv) trametinib 0.5 mg/kg oral gavage daily, (v) both palbociclib and ulixertinib, or (vi) palbociclib and trametinib. Mice were monitored daily for signs of toxicity and were weighed once weekly. Tumor size was evaluated twice per week by caliper measurements using the following formula: tumor volume = length × width2 × 0.52. Tumor volume was monitored and expressed as a percent of volume on day 1 for each treatment group at the time points evaluated. The mice were euthanized at 30 days after the initial treatment and the tumors were excised for histopathologic and molecular studies 1 hour after dosing. A Student t test was conducted on the tumor volumes of each arm at the end of study.

IHC staining and pathologist scores

Tumors were collected and fixed in 10% buffered formalin. Samples were embedded in paraffin and sections were cut. IHC staining for phospho-ERK (Cell Signaling Technology, 1:200; RRID: AB_2315112) and Ki67 (Thermo Fisher Scientific, 1:100; RRID: AB_2341197) was performed and scored as the percentage of positive cells by a pathologist. For each tumor, 100–200 cells in four distinct regions were scored for the number of positively stained cells to yield a percent (N positive cells/N total number of cells evaluated) and averaged for a final total positive percent staining per tumor sample.

Immunoblotting and quantification

Tumors or cell lysates were harvested in RIPA lysis and extraction buffer (Thermo Fisher Scientific) with 1X HALT protease and phosphatase inhibitor cocktail (Thermo). Protein lysates were resolved on 4%–12% gradient SDS-PAGE gels and were transferred to nitrocellulose membranes, and incubated at 4°C overnight with the indicated antibodies diluted in 1:3 Odyssey Blocking Buffer in TBST (LI-COR). Antibodies were purchased from Cell Signaling Technology: phospho-ERK1/2 (RRID: AB_2315112), ERK1/2 (RRID: AB10695739), Cyclin D1 (RRID: AB_2259616), phospho-RB (S780, RRID: AB_330015; S807, RRID: AB_11178658), RB (RRID: AB_823629), phospho-MEK1/2 (RRID: AB_330810) MEK1/2 (RRID: AB_10695868), β-actin (RRID: AB_2242334), β-Tubulin (RRID: AB_823664). The antibody for p16-INK4A (RRID: AB_2078303) was purchased from ProteinTech. Blots were incubated with secondary goat anti-mouse or anti-rabbit IRDye-conjugated antibodies (LI-COR Biosciences), and proteins were imaged with the Odyssey CLx imager (LI-COR Biosciences). Protein expression was quantified with the Odyssey Image studio software (Version 5.2.5), and normalized to the corresponding loading control.

Cell culture and generation of BRAF inhibitor–resistant cell lines

Human melanoma cell lines A375 (RRID: CVCL_0132) and SKMEL28 (RRID: CVLC_0526) were obtained from Dr. Mayumi Fujita in 2016 (University of Colorado, Anschutz Medical Campus; Aurora, CO) and grown in RPMI (Invitrogen) supplemented with 10% FBS (HyClone Laboratories). All cell lines were maintained at 37°C in 5% CO2. Cell lines were validated using STR profiling using the Applied Biosystems Identifier kit (#4322288) in the Barbara Davis Center Core Molecular Biology Core Facility, at the University of Colorado, as previously described (22). All cell lines were routinely monitored for Mycoplasma contamination using the Lonza Mycoalert system according to the manufacturer's directions. BRAF inhibitor–resistant cell lines were generated in vitro by continuously culturing cells in 5 μmol/L vemurafenib over 3–6 months. Resistance was defined once consistent growth occurred in 5 μmol/L vemurafenib and measured routinely by CellTiterGlo assays, as described below.

Viral transfections and generation of stable cell lines

Cell lines (A375 and A375-VR) were transduced with nuclear RFP lentivirus (Essen Biosciences) and were selected with puromycin for 7 days, and sorted by flow cytometry for RFP positive expression. shRNA targeting human CDKN2A (Sigma mission TRC2#: TRCN0000281415 (sh15) and TRCN0000255849 (sh49) and a scrambled control (shScr, Sigma mission SHC202) were purchased from the UCCC Functional Genomics Shared Resource. All shRNA constructs were packaged for lentiviral delivery via HEK293T cells using the Lenti-X lentiviral expression system (Clontech) and cells were transduced and selected with puromycin for 7 days.

Cellular growth assays

Incucyte

100–500 cells per well were plated in triplicate in 96-well plates. Cells were treated with increasing concentrations of the indicated drugs at time zero for 7 days and 4 images per well were acquired every 6 hours on the Incucyte ZOOM (Essen Bioscience) in the UCCC Cell Technologies Shared Resource. The Incucyte S3 software (v2018A) was used to analyze the images over time and count the number of red cells over time (nuclear count/mm2). IC50 values were calculated using the area under the curve (AUC) for each concentration over time, using GraphPad Prism software (version 8.4.2).

BrdU

100–500 cells per well were plated in clear 96-well plates, in 100 μL RPMI media with 10% FBS. Cells were treated with the indicated drug concentrations for 5 days and cell proliferation was measured using the Cell Proliferation ELISA, BrdU (colorimetric; Sigma-Aldrich) assay and Synergy H1 plate reader (BioTek) following manufacturer's protocol.

CellTiter-Glo 2.0 assay

Cells were plated in opaque-walled 96-well pates, in 100 μL of RPMI media with 10% FBS. Cells were treated with the indicated drug concentrations for 3–7 days and cell proliferation was measured by luminescence using CellTiterGlo 2.0 (Promega) assay and Synergy H1 plate reader (BioTek) following manufacturer's protocol.

Calculation of drug synergy

Drug synergy was calculated by Combination Index using CalcuSyn (Version 2) software that performs multiple drug dose-effect calculations using the Median Effect methods as described previously (23).

Cell-cycle analysis

Cells were treated for 24 hours as indicated. Adherent and floating cells were pooled, pelleted, washed twice with PBS, and fixed by dropwise addition of ice-cold 70% ethanol for at least 4 hours. Fixed cells were washed twice and resuspended in PBS, treated with RNase A solution at 100 mg/mL, and stained with propidium iodide (Sigma) at 10 mg/mL overnight. Cell-cycle analysis was performed (MoFlo XDP 100) and data were analyzed using FlowJo software at the UCCC Flow Cytometry Shared Resource.

Statistical analysis

Data show the mean of at least three independent experiments ±SD or SEM, as indicated. GraphPad Prism statistical software (version 8.4.2) was used to perform the two-tailed Student t test and two-way ANOVA analysis. Fold changes are calculated from the mean values of each treatment group. For all statistical asterisks (*) indicate *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001; and n.s., not significant.

Patients with BRAF-V600E–mutant melanoma refractory to BRAFi/MEKi therapy harbor mutations in NRAS and CDKN2A

To identify actionable acquired mutations upon resistance to BRAFi/MEKi therapy, we analyzed tumors from two patients with superficial spreading melanoma harboring BRAF-V600E mutations. Upon progressive disease, refractory metastatic lesions were surgically resected and analyzed by whole exome sequencing (WES) and established as patient-derived xenograft (PDX) models for both patient 1 (PDX: MB1998) and patient 2 (PDX: MB2132; Fig. 1A). We used the IMPACT pipeline for WES analysis to identify somatic mutations and to predict drugs against deleterious variants (19); however, all identified mutations are available in Supplementary File S1. For Patient 1, we identified BRAF-V600E, NRAS-G12V mutation and a CDKN2A copy number loss at time of progression (Fig. 1B; Supplementary Table S1; Supplementary Fig. S1). For patient 2, WES was performed on tumors: (i) at baseline prior to MEKi, (ii) upon disease progression after BRAFi, and (iii) after disease progression following BRAFi/MEKi (Fig. 1C; Supplementary Table S2). Patient 2 developed an NRAS-Q61K mutation and lost one copy of CDKN2A, as these aberrations were not present in the analysis performed on the tumor prior to BRAFi/MEKi (Fig. 1C; Supplementary Fig. S1B). The BRAF-V600E mutation remained present in all resistant tumor samples for both patients (Supplementary Fig. S1A and S1B). These results suggest that the potential resistance mechanisms were through reactivation of the MAPK pathway and activation of the cell-cycle pathway.

Dual inhibition of MEK and CDK4/6 overcomes BRAFi/MEKi acquired resistance in PDX models

Because we identified mutations in NRAS and CDKN2A in patients refractory to BRAFi/MEKi therapy (Fig. 1B and C), we hypothesized that dual inhibition of the MAPK and the cell-cycle pathways would overcome BRAFi/MEKi acquired resistance. To test this hypothesis, we generated two PDX models from BRAFi/MEKi–resistant patient tumors (Fig. 1B and C) and focused on inhibition of CDK4/6, with palbociclib (CDK4/6i) in combination with either MEK inhibition or ERK inhibition, with trametinib (MEKi) or ulixertinib (ERKi), respectively.

We observed significant inhibition of tumor growth in the MB1998 PDX model (BRAFV600E/NRASG12V/CDKN2A-/+) at the end of the study (30 days) following treatment with CDK4/6i in combination with either MEKi (P ≤ 0.001, t test) or ERKi (P ≤ 0.01, t test), as compared to the vehicle-treated control. The combination of CDK4/6i and MEKi resulted in an average final tumor volume of 241.8 ± 43.4 mm3, representing an 80.4% average reduction in tumor volume compared with vehicle control tumors, with an average final tumor volume of 1243.1 ± 206.5 mm3 (Fig. 1D). Whereas the combination of CDK4/6i and ERKi showed a 64.8% average tumor reduction, with an average final tumor volume of 434.4± 116.2 mm3 (Supplementary Fig. S2A). We observed a similar trend in the second PDX model, MB2132 (BRAFV600E/NRASG12V/CDKN2A−/−), the combination of CDK4/6i and MEKi resulted in an average final tumor volume of 414.12 ± 60.3 mm3, representing a 42% average reduction in tumor volume compared with vehicle control tumors, with an average final tumor volume of 714.5 ± 123.5 mm3 (P = 0.06, n.s., t test; Fig. 1E). Furthermore, in the MB2132 model, we observed that the combination of CDK4/6i and ERKi only resulted in a modest 32.7% average reduction in tumor volume, with an average final tumor volume of 480.6 ± 55.6 mm3, which was not significant (P = 0.1, n.s., t test; Supplementary Fig. S2B). No toxicities were observed in the treatment arms as measured by net body weight (Supplementary Fig. S3A and S3B). At the end of the study (30 days), phospho-RB and phospho-ERK expression was analyzed by Western blot to validate the drug's targets and we observed enhanced inhibition of phospho-ERK and phospho-RB in the combination of CDK4/6i and MEKi for both MB1998 and MB2132 (Fig. 1F; Supplementary Fig. S2C and S2D). In both PDX models, the combination of CDK4/6i with either MEKi or ERKi resulted in decreased phospho-ERK expression compared with vehicle control tumors at the end of the study, measured by IHC (Fig. 1H, J; Supplementary Fig. S2E–S2J). Interestingly, we observed a significant decrease in phospho-ERK only in the MB1998 PDX tumors with single agent inhibition of CDK4/6i by Western blot and IHC (Fig. 1F, G, and H). Tumor proliferation, measured by Ki-67 staining of PDX tumors, revealed a significant decrease in Ki-67–positive cells in all treatment groups (MEKi, CDK4/6i, MEKi+CDK4/6i, and ERKi+CDK4/6i) except single-agent ERKi, as compared with vehicle control (Fig. 1I, K; Supplementary Fig. S2E-S2J). Going forward, our studies will focus on the combination of CDK4/6i and MEKi, given that MEK inhibitors are currently FDA approved for melanoma treatment, and our observation that this specific combination more significantly reduced PDX tumor growth.

In vitro cell line models of acquired resistance to vemurafenib are more sensitive to CDK4/6i alone or in combination with MEKi

We next examined the efficacy of the combination of CDK4/6i and MEKi in two BRAF-V600E–mutant cutaneous melanoma cell lines, A375 and SKMEL28, with acquired resistance to the BRAFi vemurafenib (Supplementary Fig. S4A). Resistant cell lines were developed by culturing cells in a high concentration of the BRAFi vemurafenib (5 μmol/L). Short tandem repeat (STR) profiling was performed upon acquisition of vemurafenib resistance to confirm cell line genetic identity (Supplementary Tables S3 and S4). The A375-vemurafenib-resistant (A375-VR) and SKMEL28-vemurafinib resistant (SKMEL28-VR) cells maintained cross resistance to MEKi (Supplementary Fig. S4B). Baseline Western blot analysis demonstrated a trend toward increased phospho-ERK and phospho-RBS807 in VR cell lines, although not significant, and confirms that p16 protein expression trends with CDKN2A mutational status (Supplementary Fig. S4C).

To accurately monitor cell proliferation over time, A375 and A375-VR cells were labeled with nuclear RFP and imaged over time using high-content live-cell imaging after treatment with either a single agent or the combination of CDK4/6i and MEKi (Supplementary Table S3; Supplementary Fig. S4D). As expected, A375 cells are sensitive to low nanomolar concentrations of MEKi, and A375-VR cells are more resistant to MEKi, and maintain the ability to grow in concentrations >5 nmol/L (Fig. 2A). The A375-VR cells are more sensitive to single-agent CDK4/6i (Fig. 2B), represented by a 7.9-fold decrease in IC50 (0.024 μmol/L), as compared with A375 cells (IC50: 0.189 μmol/L; Fig. 2C). The combination of CDK4/6i and MEKi resulted in effective growth inhibition of both A375 parental and A375-VR cells (Fig. 2D and E). We observed that the A375 cells required a lower concentration of the MEKi (0.5 nmol/L) and a higher concentration of the CDK4/6i (0.125 μmol/L; Fig. 2D). Conversely, the A375-VR cells that required a higher concentration of MEKi (5 nmol/L) and a lower concentration of CDK4/6i (0.0625 μmol/L; Fig. 2E), to achieve a similar growth inhibition response to the combination.

We analyzed cell-cycle distribution after 24 hours of treatment with the single agent or combination of CDK4/6i and MEKi. Here, significant cell-cycle inhibition here is defined as a significant increase in the G1 population (P ≤ 0.05) cooccurring with a significant decrease in the S-phase population (P ≤ 0.05). We observed that MEKi significantly inhibited the cell cycle in the A375 cells; however, CDK4/6i did not significantly inhibit the cell cycle (Fig. 2F). Combination treatment with CDK4/6i and MEKi significantly inhibited the cell cycle, although to a similar extent as single agent MEKi, suggesting that cell-cycle progression is highly regulated by MEK inhibition in A375 cells. In contrast, the A375-VR cells are most sensitive to the combination of CDK4/6i and MEKi, resulting in significant alterations in G1 (** P ≤ 0.05) and S-phase (* P ≤ 0.05), as compared with either single agent (Fig. 2F, bottom).

To assess the activity of the MAPK and cell-cycle pathways, we monitored the expression of phosphorylated and total ERK and RB, and cyclin D1 by Western blot analysis after 24 hours of treatment with either single agent or the combination of CDK4/6i and MEKi. Single-agent MEKi reduced levels of phospho-ERK, phospho-RB and cyclin D1 in both A375 and A375-VR cells; however, the combination was required for the efficient reduction of phospho-RB in A375-VR cells (Fig. 2G; Supplementary Fig. S5A-S5C). Overall, our data suggests that A375-VR cells are more sensitive to dual inhibition of the MAPK and cell-cycle pathways, and are thus more reliant on both pathways for cellular proliferation compared with A375 parental cells.

BRAFi resistant CDKN2A-WT SKMEL28-VR cells are sensitive to CDK4/6i alone and in combination with MEKi

Next, we examined the effects of wildtype CDKN2A/p16INK4A expression in response to single-agent CDK4/6i or in combination with MEKi using the CDKN2A+/+ SKMEL28 and SKMEL28-VR cells. As expected, the SKMEL28 cells demonstrate a dose-dependent reduction in growth in response to MEKi, as compared with SKMEL28-VR cells that exhibit full resistance to MEKi (Fig. 3A, left). The SKMEL28-VR cells are more sensitive to single-agent CDK4/6i (Fig. 3A), represented by a 3.5-fold decrease in IC50 (0.344 μmol/L), as compared with parental SKMEL28 cells (IC50: 1.2 μmol/L; Fig. 3B).

We next tested the efficacy of the combination of CDK4/6i and MEKi in both SKMEL28 and SKMEL28-VR (Fig. 3C and D, top). Drug synergy was calculated using the Combination Index (CI) values (Fig. 3C and D, bottom). While both parental and VR cell lines exhibited a combination effect and synergy (CI values, 0–0.3), the SKMEL28-VR cells were more synergistic and demonstrated overall lower CI values compared with SKMEL28 (Fig. 3C and D). Compared with SKMEL28, the SKMLE28-VR cells required higher concentrations of MEKi to achieve similar results (Fig. 3D).

The SKMEL28 cells were equally sensitive to cell -cycle inhibition with either single-agent MEKi or CDK4/6i demonstrating significant cell-cycle inhibition (Fig. 3E). The combination of CDK4/6i and MEKi resulted in significant cell-cycle inhibition comparable with either single-agent MEKi or CDK4/6i. In contrast, CDK4/6i alone resulted in significant cell-cycle inhibition, in the SKMEL28-VR cells and the combination of CDK4/6i and MEKi produced the most significant alterations in G1 (**, P ≤ 0.01) and S-phase (***, P ≤ 0.001), compared with either single agent (Fig. 3E).

As expected, single-agent MEKi significantly reduced the levels of phospho-ERK, phospho-RB, and cyclin D1 in SKMEL28 cells, but only reduced phospho-ERK levels in SKMEL28-VR cells, as measured by Western blot analysis (Fig. 3F; Supplementary Fig. S6A-S6C). In the SKMEL28-VR cells CDK4/6i, either as a single agent or in combination with MEKi, is required to fully inhibit phospho-RB (Fig. 3F; Supplementary Fig. S6A-S6C). These data suggest that single-agent MEKi in parental SKMEL28 cells is sufficient to fully shut down both the MAPK pathway and the cell-cycle pathway, as evident by cyclin D1 and phospho-RB levels trending with decreased phospho-ERK levels. However, in SKMEL28-VR cells, MEKi alone cannot sufficiently decrease phospho-RB and cyclin D1 levels, resulting in uninhibited cell-cycle pathway signaling. Thus, the SKMEL28-VR cells require CDK4/6i to fully inhibit phospho-RB and cell growth. These data further support that BRAFi-resistant cells have increased sensitivity to CDK4/6i (Figs. 2B and 3B), suggesting that the cell-cycle pathway may be a key target to overcome BRAFi resistance.

Loss of CDKN2A/p16INK4A expression increases the sensitivity of BRAFi resistant cells to single-agent CDK4/6i

Next, we tested the hypothesis that loss of CDKN2A/p16INK4A after the development of acquired resistance to BRAFi promotes increased sensitivity to CDK4/6i. We genetically knocked down CDKN2A in the CDKN2A+/+ SKMEL28-VR cells using two shRNAs (shCDKN2A_15 and shCDKN2A_49), alongside a negative control nontargeting shRNA (shScr) in the SKMEL28-VR cells. Almost complete knockdown of CDKN2A was achieved, as measured at the mRNA level by qRT-PCR and at the protein level by assessing p16INK4A expression by western blot (Fig. 4A and B). Knockdown of CDKN2A significantly increased baseline levels of phospho-RB without significant alterations in phospho-ERK and cyclin D1, as compared with shScr cells (Fig. 4B).

Notably, knockdown of CDKN2A further sensitized SKMEL28-VR cells to single-agent CDK4/6i, as compared to SKMEL28-VR-shScr (Fig. 4C). In response to CDK4/6i, we observed an average 2.6-fold decrease in IC50 values for SKMEL28-VR-shCDKN2A cells (range: 0.17–0.21 μmol/L), as compared to SKMEL28-VR-shScr cells (IC50: 0.49; Fig. 4D). Although we observed increased sensitivity in SKMEL28-VR-shCDKN2A cells to single-agent MEKi (Fig. 4E), with a corresponding 3-fold decrease in IC50 values to SKMEL28-VR-shScr, these IC50 values are in the high nanomolar concentrations (range: 88.7–92.9 nmol/L) are considered to be resistant and are not clinically achievable (Fig. 4F). This data suggests that loss of CDKN2A expression after acquired resistance to BRAFi further increases the sensitivity of SKMEL28-VR cells to CDK4/6i.

Loss of CDKN2A/p16INK4A expression results in a synergistic growth response to the combination of CDK4/6i and MEKi in BRAFi resistant cells

Because knockdown of CDKN2A sensitized SKMEL28-VR cells to single-agent CDK4/6i, we hypothesized that the combination of CDK4/6i and MEKi would result in a synergistic growth response in SKMEL28-VR-shCDKN2A cells. We found that SKMEL28-VR-shCDKN2A cells are more sensitive to the combination of CDK4/6i and MEKi (Fig. 5A–C, top), and display increased synergy represented by CI values >0.3 (Fig. 5A–C, bottom), as compared with SKMEL28-VR-shScr cells.

Knockdown of CDKN2A in the SKMEL28-VR cells increased sensitivity to cell-cycle arrest with single-agent CDK4/6i, showing a significant increase in G1 and decrease in S-phase (Fig. 5D, middle and right), as compared with SKMEL28-VR-shScr cells (Fig. 5D, left panel). As expected, the SKMEL28-VR-shScr and shCDKN2A cells were sensitive to the combination of CDK4/6i and MEKi, whereas single-agent MEKi did not have a significant effect on the cell cycle (Fig. 5D). While single-agent CDK4/6 inhibition reduced phospho-RB levels in SKMEL28-VR-shScr cells, only the combination of CDK4/6i and MEKi resulted in reduced phospho-RB levels in SKMEL28-VR-shCDKN2A cells (Fig. 5E; Supplementary Fig. S7A–S7C). We observed that single agent MEKi failed to reduce phospho-RB and phospho-ERK levels in shCDKN2A cells to the same degree as observed in shScr cells (Fig. 5E; Supplementary Fig. S7A-S7C). Overall, our data indicate that knockdown of CDKN2A promotes sensitivity to single-agent CDK4/6i and increased synergy to the combination of CDK4/6i and MEKi in resistant SKMEL28-VR cells.

Although targeted inhibition of BRAF and MEK has significantly improved survival for patients with BRAF-mutant melanoma, acquired resistance to targeted therapeutics remains a significant clinical challenge. There is a clear need to identify targetable drivers of resistance and potential biomarkers for salvage therapy for patients who are refractory to the combination of BRAFi/MEKi, and to ultimately provide a meaningful improvement in the overall survival for patients with melanoma. In this study, we performed WES and found that two patients acquired oncogenic mutations in NRAS and the loss of the tumor suppressor gene CDKN2A at the time of resistance to BRAFi/MEKi. We hypothesized that acquired resistance to BRAFi/MEKi is associated with reactivation of the MAPK pathway and activation of the cell-cycle pathway. We demonstrated that simultaneous inhibition of CDK4/6 and MEK significantly decreased PDX tumor growth in vivo in PDX models harboring BRAF-V600E and NRAS mutations, and CDKN2A loss (Fig. 1); and significantly reduced cellular growth, promoted cell-cycle arrest and effectively inhibited downstream signaling of MAPK and cell-cycle pathways in BRAFi-resistant cell lines in vitro (Figs. 2 and 3). Furthermore, we found that knockdown of CDKN2A after acquired resistance to BRAFi increased sensitivity to single-agent CDK4/6 inhibition and increased the synergistic response to CDK4/6i and MEKi (Fig. 5). Thus, a key implication of our study is that the combination of CDK4/6i and MEKi overcomes acquired resistance to BRAFi/MEKi, and loss of CDKN2A may represent a biomarker of response to the combination that should be further explored in future studies.

Previous studies have demonstrated the efficacy of the combination of CDK4/6i and MEKi in melanoma, prior to the development of drug resistance. Upfront combination of CDK4/6i and MEKi was efficacious in both BRAF- and NRAS-mutant in vivo models (13, 24). In addition, upfront treatment with CDK4/6i in addition to BRAFi/MEKi delayed the onset of drug resistance, which was associated with the prevention of a reversible drug-tolerant phase observed with MAPK inhibition; however, the combination was not efficacious after acquired resistance to BRAFi (25). While upfront combination therapy of CDK4/6i and BRAFi/MEKi is efficacious, it raises the issue of new resistance mechanisms as a result of upfront CDK4/6i treatment. Interestingly, resistance to single-agent CDK4/6i has been associated with increased MAPK pathway signaling and dependence in breast cancer and KRAS-mutant lung cancer, further highlighting the cooperative signaling between the MAPK and cell-cycle pathways (26, 27). Furthermore, resistance mechanisms to upfront combination of CDK4/6i and MEKi were associated with acquired NRAS mutations and mTOR/mTORC1 pathway activation, where the addition of mTOR inhibitors overcame resistance (28). In agreement with previous studies, we also observed upfront sensitivity to the combination prior to resistance, however we found increased sensitivity to the combination after acquired resistance in vitro (Figs. 2 and 3). Although we did not observe the same genomic alterations in NRAS and CDKN2A in vitro (Supplementary Fig. S4) as observed in BRAFi/MEKi–resistant patient tumors, we nonetheless observed an enhanced response to the combination in vitro, indicating the importance of these pathways (Figs. 2 and 3). While the PI3K/AKT pathway has been extensively shown to mediate intrinsic or acquired resistance to MAPK pathway inhibition (29, 30), future studies determining the response to the combination of CDK4/6i and MEKi should also focus on resistant tumors with PI3K/AKT pathway activation.

Previous studies in melanoma have identified that mutation, loss or reduced expression of CDKN2A predicts sensitivity to single agent CDK4/6i in vitro (17, 31). To further assess the effects of CDKN2A expression on acquired resistance mechanisms to BRAFi, we used three in vitro models representing multiple scenarios of acquired resistance; (i) CDKN2A loss at baseline and prior to resistance (A375-VR) (Fig. 2), (ii) wild-type CDKN2A expression is maintained after resistance (SKMEL28-VR; Fig. 3) and (iii) loss of CDKN2A upon resistance (SKMEL28-VR-shCDKN2A; Figs. 4 and 5). In all in vitro models of acquired resistance, we found that VR cells are more sensitive to single agent CDK4/6i and demonstrate increased synergy to the combination of CDK4/6i and MEKi, as compared with parental cells. We demonstrate that inhibition of the MAPK pathway with MEKi is no longer sufficient to decrease phospho-ERK and also phospho-RB levels in VR cells, which warrants the use of a CDK4/6i in addition to MEKi to achieve a similar reduction in phospho-RB as seen in parental cells (Fig. 2G and 3F). Importantly, knockdown of CDKN2A in SKMEL28-VR cells resulted in increased synergistic inhibition of growth to the combination of CDK4/6i and MEKi (Fig. 5). While we specifically studied loss of p16INK4A encoded by CDKN2A, the CDKN2A gene locus (Chr9p21) also encodes for p14ARF protein, which prevents p53 degradation and regulates cell cycle, apoptosis and DNA repair (32). Thus, loss of the CDKN2A locus could also result in loss of p14ARF, which has been demonstrated to enhance melanoma development in the background of BRAF-V600E mutations in vivo (33), and should be considered in future studies. Overall our data suggests that loss of CDKN2A either prior to or after acquired BRAFi/MEKi resistance, may represent a potential biomarker for a synergistic response to the combination of CDK4/6i and MEKi. In support of this, thyroid cancer cells harboring BRAF-V600E mutations and CDKN2A−/− were sensitive to the combination of CDK4/6i and BRAFi prior to and upon BRAFi resistance (34). We found that either single-agent CDK4/6i or in combination with MEKi reduced phospho-RB levels in both PDX models, we observed a more robust decrease of phospho-RB in the MB2132 PDX model, which may be attributed to full loss of CDKN2A−/− (Fig. 1F and G). However, in the MB1998 PDX model with partial loss of CDKN2A–/+, we observed more robust tumor inhibition with the combination of CDK4/6i and MEKi. Interestingly, we also observed significant downregulation of phospho-ERK following single-agent CDK4/6i in MB1998, not observed in MB2132, and while the mechanism of CDK4/6i reducing phospho-ERK levels remains unclear, it should be explored in future studies as it may confer increased sensitivity to CDK4/6i. To gain a better mechanistic understanding CDKN2A loss after acquired resistance mediating sensitivity to CDK4/6i, future studies in additional cell lines and PDX models are warranted, which will aid in fully defining CDKN2A loss as a biomarker of response to CDK4/6i upon BRAFi/MEKi resistance. Furthermore, these findings may extend to other BRAF-V600E–mutant cancers with intrinsic or acquired resistance to MAPK pathway inhibitors. Currently, there are ongoing clinical trials for patients with melanoma and other solid tumors that are refractory to front line therapy of BRAFi/MEKi, or standard of care, assessing the addition of CDK4/6i (35). Importantly, the LOGIC-2 (NCT02159066), MATCH (NCT0246506) and the MatchMel (NCT02645149) trials are incorporating genetic mutational data with a focus on cell-cycle pathway alterations, including CDKN2A, with clinical responses to CDK4/6i, which will further define the critical role of CDKN2A loss as a biomarker of response to CDK4/6i.

In summary, the data presented here demonstrates that after acquired resistance, the combination of CDK4/6i and MEKi reduces cellular and tumor growth in vitro and in vivo. Thus, upon BRAFi/MEKi acquired resistance, the cell-cycle pathway represents a key bypass signaling pathway that is a targetable resistance mechanism. Clinically, it remains unclear as to which upfront combination patients would have the most durable long-term response from, a MEKi in combination with either a BRAFi or a CDK4/6i. The combination of CDK4/6i and MEKi may have the ability to globally overcome BRAFi/MEKi resistance, but raises an important clinical question of when, and how, to accurately define resistance to the combination of BRAFi/MEKi and switch therapies to MEKi and CDK4/6i to achieve the most durable long-term response. As continuous dosing of palbociclib elicits significant toxicities in patients (36), a key consideration in combining CDK4/6i with targeted therapy will be limiting toxicity-related events using strategic dosing strategies to maximize clinical benefit. In support of this, continuous dosing of MEKi with intermittent CDK4/6i led to more complete reduction in tumor burden compared with other dosing schedules in a melanoma in vivo model (28). Thus, further defining mechanisms of resistance and potential biomarkers of response to this combination will be critical to define the best upfront combination therapies, and/or salvage therapies to overcome acquired resistance mechanisms.

M. Fujita reports grants from NIH/NCI and grants from VA merit award grant during the conduct of the study. No disclosures were reported by the other authors.

The contents of this study are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.

K.W. Nassar: Conceptualization, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. J.D. Hintzsche: Conceptualization, data curation, investigation, visualization, writing–original draft. S.M. Bagby: Formal analysis, investigation, visualization, writing–review and editing. V. Espinoza: Investigation, writing–review and editing. C. Langouët-Astrié: Formal analysis, investigation, visualization, writing–review and editing. C.M. Amato: Resources, project administration, writing–review and editing. T.-S. Chimed: Investigation. M. Fujita: Investigation, writing–review and editing. W. Robinson: Conceptualization, resources, funding acquisition, methodology, writing–review and editing. A.C. Tan: Conceptualization, resources, supervision, funding acquisition, methodology, writing–review and editing. R.E. Schweppe: Conceptualization, resources, supervision, methodology, writing–review and editing.

We would like to acknowledge the following core facilities at the University of Colorado, Anschutz Medical Campus- Aurora, CO: The Barbara Davis Center Molecular Biology Core, the Flow Cytometry Shared Resource and the Cell Technologies Shared Resource. We would also like to acknowledge the Amy Davis Foundation, the Moore Family Foundation and the Margaret T. Grohne Family Foundation. This work is partly supported by the Cancer League of Colorado (to A.C. Tan, J.D. Hintzsche), the David F. and Margaret T. Grohne Family Foundation (to W.A. Robinson), the Rifkin Endowed Chair (to W.A. Robinson), the Amy Davis Foundation and the Moore Family Foundation (to W.A. Robinson), NIH/NCI T32CA190216 (to K.W. Nassar), RNA Biosciences Initiative RNA Scholar Award (to K.W. Nassar). The UCCC DNA Sequencing, the Flow Cytometry Shared resource and the Cell Technologies shared resource is supported by NCI Cancer Center, grant P30 CA046934.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Long
GV
,
Stroyakovskiy
D
,
Gogas
H
,
Levchenko
E
,
de Braud
F
,
Larkin
J
, et al
Dabrafenib and trametinib versus dabrafenib and placebo for Val600 BRAF-mutant melanoma: a multicentre, double-blind, phase 3 randomised controlled trial
.
Lancet
2015
;
386
:
444
51
.
2.
Robert
C
,
Grob
JJ
,
Stroyakovskiy
D
,
Karaszewska
B
,
Hauschild
A
,
Levchenko
E
, et al
Five-year outcomes with Dabrafenib plus Trametinib in metastatic melanoma
.
N Engl J Med
2019
;
381
:
626
36
.
3.
Kakadia
S
,
Yarlagadda
N
,
Awad
R
,
Kundranda
M
,
Niu
J
,
Naraev
B
, et al
Mechanisms of resistance to BRAF and MEK inhibitors and clinical update of US Food and Drug Administration-approved targeted therapy in advanced melanoma
.
Onco Targets Ther
2018
;
11
:
7095
107
.
4.
Johnson
DB
,
Menzies
AM
,
Zimmer
L
,
Eroglu
Z
,
Ye
F
,
Zhao
S
, et al
Acquired BRAF inhibitor resistance: a multicenter meta-analysis of the spectrum and frequencies, clinical behaviour, and phenotypic associations of resistance mechanisms
.
Eur J Cancer
2015
;
51
:
2792
9
.
5.
Nazarian
R
,
Shi
H
,
Wang
Q
,
Kong
X
,
Koya
RC
,
Lee
H
, et al
Melanomas acquire resistance to B-RAF(V600E) inhibition by RTK or N-RAS upregulation
.
Nature
2010
;
468
:
973
7
.
6.
Van Allen
EM
,
Wagle
N
,
Sucker
A
,
Treacy
DJ
,
Johannessen
CM
,
Goetz
EM
, et al
The genetic landscape of clinical resistance to RAF inhibition in metastatic melanoma
.
Cancer Discov
2014
;
4
:
94
109
.
7.
Kaplan
FM
,
Kugel
CH
,
Dadpey
N
,
Shao
Y
,
Abel
EV
,
Aplin
AE
. 
SHOC2 and CRAF mediate ERK1/2 reactivation in mutant NRAS-mediated resistance to RAF inhibitor
.
J Biol Chem
2012
;
287
:
41797
807
.
8.
Luebker
SA
,
Koepsell
SA
. 
Diverse mechanisms of BRAF inhibitor resistance in melanoma identified in clinical and preclinical studies
.
Front Oncol
2019
;
9
:
268
.
9.
Sheppard
KE
,
McArthur
GA
. 
The cell-cycle regulator CDK4: An emerging therapeutic target in melanoma
.
Clin Cancer Res
2013
;
19
:
5320
8
.
10.
Otto
T
,
Sicinski
P
. 
Cell cycle proteins as promising targets in cancer therapy
.
Nat Rev Cancer
2017
;
17
:
93
115
.
11.
Sanki
A
,
Li
W
,
Colman
M
,
Karim
RZ
,
Thompson
JF
,
Scolyer
RA
. 
Reduced expression of p16 and p27 is correlated with tumour progression in cutaneous melanoma
.
Pathology
2007
;
39
:
551
7
.
12.
Lavoie
JN
,
L'Allemain
G
,
Brunei
A
,
Müller
R
,
Pouysségur
J
. 
Cyclin D1 expression is regulated positively by the p42/p44(MAPK) and negatively by the p38/HOG(MAPK) pathway
.
J Biol Chem
1996
;
271
:
20608
16
.
13.
Kwong
LN
,
Costello
JC
,
Liu
H
,
Jiang
S
,
Helms
TL
,
Langsdorf
AE
, et al
Oncogenic NRAS signaling differentially regulates survival and proliferation in melanoma
.
Nat Med
2012
;
18
:
1503
10
.
14.
Nathanson
KL
,
Martin
AM
,
Wubbenhorst
B
,
Greshock
J
,
Letrero
R
,
D'Andrea
K
, et al
Tumor genetic analyses of patients with metastatic melanoma treated with the BRAF inhibitor dabrafenib (GSK2118436)
.
Clin Cancer Res
2013
;
19
:
4868
78
.
15.
Smalley
KSM
,
Lioni
M
,
Palma
MD
,
Xiao
M
,
Desai
B
,
Egyhazi
S
, et al
Increased cyclin D1 expression can mediate BRAF inhibitor resistance in BRAF V600E-mutated melanomas
.
Mol Cancer Ther
2008
;
7
:
2876
83
.
16.
Guo
L
,
Qi
J
,
Wang
H
,
Jiang
X
,
Liu
Y
. 
Getting under the skin: The role of CDK4/6 in melanomas
.
Eur J Med Chem
2020
;
204
:
112531
.
17.
Young
RJ
,
Waldeck
K
,
Martin
C
,
Foo
JH
,
Cameron
DP
,
Kirby
L
, et al
Loss of CDKN2A expression is a frequent event in primary invasive melanoma and correlates with sensitivity to the CDK4/6 inhibitor PD0332991 in melanoma cell lines
.
Pigment Cell Melanoma Res
2014
;
27
:
590
600
.
18.
Hintzsche
JD
,
Gorden
NT
,
Amato
CM
,
Kim
J
,
Wuensch
KE
,
Robinson
SE
, et al
Whole-exome sequencing identifies recurrent SF3B1 R625 mutation and comutation of NF1 and KIT in mucosal melanoma
.
Melanoma Res
2017
;
27
:
189
99
.
19.
Hintzsche
J
,
Kim
J
,
Yadav
V
,
Amato
C
,
Robinson
SE
,
Seelenfreund
E
, et al
IMPACT: a whole-exome sequencing analysis pipeline for integrating molecular profiles with actionable therapeutics in clinical samples
.
J Am Med Informatics Assoc
2016
;
23
:
721
30
.
20.
Bagby
S
,
Messersmith
WA
,
Pitts
TM
,
Capasso
A
,
Varella-Garcia
M
,
Klauck
PJ
, et al
Development and maintenance of a preclinical patient derived tumor xenograft model for the investigation of novel anticancer therapies
.
J Vis Exp
2016
;
115
:
54393
.
21.
Tentler
JJ
,
Tan
AC
,
Weekes
CD
,
Jimeno
A
,
Leong
S
,
Pitts
TM
, et al
Patient-derived tumour xenografts as models for oncology drug development
.
Nat Rev Clin Oncol
2012
;
9
:
338
50
.
22.
Landa
I
,
Pozdeyev
N
,
Korch
C
,
Marlow
LA
,
Smallridge
RC
,
Copland
JA
, et al
Comprehensive genetic characterization of human thyroid cancer cell lines: A validated panel for preclinical studies
.
Clin Cancer Res
2019
;
25
:
3141
51
.
23.
Chou
TC
,
Talalay
P
. 
Analysis of combined drug effects: a new look at a very old problem
.
Trends Pharmacol Sci
1983
;
4
:
450
4
.
24.
Teh
JLF
,
Purwin
TJ
,
Greenawalt
EJ
,
Chervoneva
I
,
Goldberg
A
,
Davies
MA
, et al
An in vivo reporter to quantitatively and temporally analyze the effects of CDK4/6 inhibitor-based therapies in melanoma
.
Cancer Res
2016
;
76
:
5455
66
.
25.
Martin
CA
,
Cullinane
C
,
Kirby
L
,
Abuhammad
S
,
Lelliott
EJ
,
Waldeck
K
, et al
Palbociclib synergizes with BRAF and MEK inhibitors in treatment naïve melanoma but not after the development of BRAF inhibitor resistance
.
Int J Cancer
2018
;
142
:
2139
52
.
26.
De Leeuw
R
,
McNair
C
,
Schiewer
MJ
,
Neupane
NP
,
Brand
LJ
,
Augello
MA
, et al
MAPK Reliance via acquired CDK4/6 inhibitor resistance in cancer
.
Clin Cancer Res
2018
;
24
:
4201
14
.
27.
Haines
E
,
Chen
T
,
Kommajosyula
N
,
Chen
Z
,
Herter-Sprie
GS
,
Cornell
L
, et al
Palbociclib resistance confers dependence on an FGFR-MAP kinase-mTOR-driven pathway in KRAS-mutant non-small cell lung cancer
.
Oncotarget
2018
;
9
:
31572
89
.
28.
Teh
JLF
,
Cheng
PF
,
Purwin
TJ
,
Nikbakht
N
,
Patel
P
,
Chervoneva
I
, et al
In vivo E2F reporting reveals efficacious schedules of MEK1/2-CDK4/6 targeting and mTOR-S6 resistance mechanisms
.
Cancer Discov
2018
;
8
:
568
81
.
29.
Zuo
Q
,
Liu
J
,
Huang
L
,
Qin
Y
,
Hawley
T
,
Seo
C
, et al
AXL/AKT axis mediated-resistance to BRAF inhibitor depends on PTEN status in melanoma
.
Oncogene
2018
;
37
:
3275
89
.
30.
Paraiso
KHT
,
Xiang
Y
,
Rebecca
VW
,
Abel
EV
,
Chen
YA
,
Munko
AC
, et al
PTEN loss confers BRAF inhibitor resistance to melanoma cells through the suppression of BIM expression
.
Cancer Res
2011
;
71
:
2750
60
.
31.
McArthur
GA
,
Young
RJ
,
Sheppard
KE
,
Mar
V
,
Waldeck
K
,
Fox
SB
, et al
Clinical significance of genomic alterations of the CDK4-pathway and sensitivity to the CDK4 inhibitor PD 0332991 in melanoma
.
J Clin Oncol
2012
;
30
:
8520
.
32.
Meyle
KD
,
Guldberg
P
. 
Genetic risk factors for melanoma
.
Hum Genet
2009
;
126
:
499
510
.
33.
Goel
VK
,
Ibrahim
N
,
Jiang
G
,
Singhal
M
,
Fee
S
,
Flotte
T
, et al
Melanocytic nevus-like hyperplasia and melanoma in transgenic BRAFV600E mice
.
Oncogene
2009
;
28
:
2289
98
.
34.
Antonello
ZA
,
Hsu
N
,
Bhasin
M
,
Roti
G
,
Joshi
M
,
Van Hummelen
P
, et al
Vemurafenib-resistance via de novo RBM genes mutations and chromosome 5 aberrations is overcome by combined therapy with palbociclib in thyroid carcinoma with BRAFV600E
.
Oncotarget
2017
;
8
:
84743
60
.
35.
Garutti
M
,
Targato
G
,
Buriolla
S
,
Palmero
L
,
Minisini
AM
,
Puglisi
F
. 
CDK4/6 inhibitors in melanoma: a comprehensive review
.
Cells
2021
;
10
:
1334
.
36.
Klein
ME
,
Kovatcheva
M
,
Davis
LE
,
Tap
WD
,
Koff
A
. 
CDK4/6 inhibitors: the mechanism of action may not be as simple as once thought
.
Cancer Cell
2018
;
34
:
9
20
.