Combined Blockade of MEK and CDK4/6 Pathways Induces Senescence to Improve Survival in Pancreatic Ductal Adenocarcinoma

Abstract

Activating KRAS mutations, a defining feature of pancreatic ductal adenocarcinoma (PDAC), promote tumor growth in part through the activation of cyclin-dependent kinases (CDKs) that induce cell cycle progression. p16^INK4a (p16), encoded by the gene CDKN2A, is a potent inhibitor of CDK4/6 and serves as a critical checkpoint of cell proliferation. Mutations in and subsequent loss of the p16 gene occur in PDAC at a rate higher than that reported in any other tumor type and results in Rb inactivation and unrestricted cellular growth. Therefore, strategies targeting downstream RAS pathway effectors combined with CDK4/6 inhibition may have the potential to improve outcomes in this disease. Herein, we show that expression of p16 is markedly reduced in PDAC tumors compared to normal pancreatic or pre-neoplastic tissues. Combined MEK inhibition (MEKi) and CDK4/6 inhibition (CDK4/6i) results in sustained downregulation of both ERK and Rb phosphorylation and a significant reduction in cell proliferation compared to monotherapy in human PDAC cells. MEKi with CDK4/6i reduces tumor cell proliferation by promoting senescence-mediated growth arrest, independent of apoptosis in vitro. We show that combined MEKi and CDK4/6i treatment attenuates tumor growth in xenograft models of PDAC and improves overall survival over 200% compared to treatment with vehicle or individual agents alone in Ptf1a^cre/+;LSL-KRAS^G12D/+;Tgfbr2^flox/flox (PKT) mice. Histologic analysis of PKT tumor lysates reveal a significant decrease in markers of cell proliferation and an increase in senescence-associated markers without any significant change in apoptosis. These results demonstrate that combined targeting of both MEK and CDK4/6 represents a novel therapeutic strategy to synergistically reduce tumor growth through induction of cellular senescence in PDAC.

Introduction

Treatment of pancreatic ductal adenocarcinoma (PDAC) remains a major therapeutic challenge due to resistance to standard therapies. The high frequency of heterogeneous mutations and the prevalence of redundant signaling pathways in PDAC are challenges to new drug development. These suggest that targeting a single pathway is unlikely to be successful.

Activating KRAS mutations are present in over 75-95% of PDAC cases(1). In numerous studies, KRAS mutations have been shown to promote cell cycle progression, in part, by overactivation of the cyclin D1 complex, which phosphorylates the retinoblastoma (Rb) tumor suppressor protein resulting in its inactivation and subsequent progression from G₁ to S phase(2,3). Due to the high prevalence of KRAS mutations, inhibiting the downstream mitogen-activated protein kinase (MEK) pathway using small molecular inhibitors has emerged as a promising therapeutic strategy in PDAC. Early studies of these inhibitors showed a durable clinical response in treating resistant melanoma and non-small cell lung carcinoma(4,5). However, MEK inhibitor monotherapy in PDAC has produced suboptimal results in both preclinical models and clinical studies due to the development of resistance and dose-limiting toxicities(6).

Alterations in the CDKN2A gene expression occur in 40-90% of PDAC tumors(7–9). These alterations occur most frequently due to homozygous deletion or truncation of the CDKN2A gene, while other causes (promoter methylation, insertion/deletion, etc.) make up less than 20% of cases(9). The CDKN2A gene encodes for the p16^INK4a (p16) tumor repressor protein, a potent inhibitor of cyclin dependant kinases 4 and 6 (CDK4/6), are activated in response to KRAS pathway signaling. Reduced p16 expression leads to enhanced activity of CDK4/6, resulting in phosphorylation of Rb (pRb) tumor suppressor protein and uninhibited G₁ to S phase cell cycle progression(10). Furthermore, p16 accumulation is crucial to RAS-induced senescence in pre-malignant cells, indicating that loss of p16 may be a potential mechanism by which transformed malignant cells escape senescence programming(11). This provides a strong rationale for combining downstream KRAS pathway inhibitors with CDK4/6 inhibitors in the treatment of PDAC.

Further supporting this approach is recent work in an NRAS mutant mouse model of melanoma showing that monotherapy with a MEK inhibitor was ineffective due to the upregulation of pathways governing cell-cycle progression following oncogenic RAS inhibition(12). Specifically, CDK4 was identified as the key driver of this therapeutic resistance, and combined inhibition of MEK and CDK4/6 resulted in an enhanced therapeutic effect. Interestingly, a similar phenomenon was seen in a preclinical model of KRAS mutant non-small cell lung carcinoma, whereby inhibition of CDK 4 signaling increased susceptibility to MEK inhibition through induced cellular senescence, preventing tumor progression(13). Taken together, these studies indicate that a potential mechanism by which PDAC may escape the targeted blockade of oncogenic KRAS signaling is through the upregulation of CDK4/6 signaling.

Our current study aims to simultaneously target RAS pathway signaling and cell-cycle progression through the use of combined MEK (MEKi) and CDK4/6 inhibition (CDK4/6i) to improve outcomes in PDAC. We show that combined MEKi with CDK4/6i reduces tumor cell proliferation by promoting senescence-mediated growth arrest, independent of apoptosis resulting in a significant survival benefit in an aggressive genetic mouse model of PDAC. These results demonstrate the efficacy and provide a mechanistic rationale for the use of combined MEKi and CDK4/6i that may be explored in future clinical studies in PDAC.

Materials and Methods

Cell lines, reagents, and primers

Authenticated human PDAC cell lines BxPC-3, MiaPaCa-2, Panc-1, CFPAC, Panc 10.05, HPNE-KRAS, and HPNE were purchased from the American Type Culture Collection (ATCC) and were maintained according to ATCC guidelines. Additionally, only low passage cells (<10) were utilized in experiments. Cell authentication was performed by using STR DNA profiling, and cell lines tested negative for Mycoplasma via Genetica cell line testing using eMYCO plus kit (iNtRON Biotechnology).

Primary antibodies used for Western blot analysis, immunohistochemistry, and immunofluorescence are listed in the Supplementary Table S1, gene-specific primers used for quantitative reverse transcription PCR (qRT-PCR) are listed in Supplementary Table S2, and chemicals and reagents used are listed in Supplementary Table S3.

Mouse studies

Athymic nude mice – Foxn1^−nu/nu (4-5 weeks old) – were purchased from Harlan Sprague Dawley, Inc. Ptf1a^cre/+;KRAS^G12D/+;Tgfbr2^flox/flox (PKT) mice were used in survival studies (Provided by Dr. Hal Moses, Vanderbilt University Medical Center, Nashville, TN)(14). Genotyping of alleles was performed using oligonucleotide primers as described previously(15,16).

Quantitative reverse transcriptase PCR (qPCR) assay

RNA was harvested using RNeasy + Mini Kit (Qiagen) and quantified using Nanodrop 2000C (Thermo). cDNA was then synthesized using High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems) and PCR was performed using RT² qPCR primer assays (Qiagen) for CDKN1A, CDKN1B, and CDKN2A with SYBR Green qPCR SuperMix (Bio-Rad) (Supplementary Table S2). Values were normalized to GAPDH and reported as fold-change relative to HPNE cell line using ddCT method.

Immunohistochemistry, immunofluorescence, and immunocytochemistry

Formalin-fixed, paraffin-embedded pancreatic tissue sections and ATCC cell lines were used for immunohistochemical and immunofluorescence-based analysis using either mouse or rabbit monoclonal antibodies (Supplementary Table S1). In brief, sections were deparaffinized using xylene and rehydrated in decreasing concentrations of alcohol. Antigen retrieval was performed by incubating samples with Tris-EDTA (pH 9.0) buffer and applying gentle heating. Detection of the primary antibody was performed by incubation with species-specific biotinylated secondary antibody (Jackson ImmunoResearch Laboratories) followed by VECTASTAIN® Elite® ABC peroxidase kit (Vector Labs) as per the manufacturer’s protocol using diaminobenzidine (DAB) as the chromogen. Similarly, detection of the primary antibody used in immunofluoresence was performed with Alexa Flour 594 nm tagged goat anti-mouse or anti-rabbit IgG (H+L) secondary antibody (Invitrogen) and nuclei were counterstained with Hoechst 33342 (Thermo). At least four representative images from individual tissue sections were captured using DMi8 and DM750M imaging platforms (Leica) and ImageJ v1.52k software with Fiji add-on were used for analysis(17,18).

Western blot analysis

Human PDAC cell lines were plated in a 6 well plate at a density of 1 x10⁶ cell per well. Cells were treated with varying concentrations of LEE011 (Array BioPharma) and MEK162 (Novartis Oncology) suspended in DMSO. Following treatment, cells were washed twice with ice-cold PBS and lysed in radioimmunoprecipitation (RIPA) assay buffer (Cell Signaling). The lysates were centrifuged at 12,000 g for 10 min at 4° C to remove debris. Aliquots of supernatants containing an equal amount of total protein (30 μg) were resolved on 4-20% polyacrylamide gels (BioRad), transferred onto polyvinylidene difluoride (PVDF) membranes (BioRad), and probed with relevant antibodies using standard protocols described previously prior to imaging using a chemiluminescent substrate(19).

Cell-cycle analysis

Human PDAC cell lines were treated with either LEE011, MEK162, or combination of LEE011 and MEK162. They were then harvested and washed with PBS cells were then fixed with ice-cold 70% ethanol for 30 minutes at 4° C. Cells were then washed with PBS twice and treated with RNAse (100μg/ml) followed by staining with DNA intercalating dye propidium iodide (50μg/ml) as discussed previously(20). Cell fluorescence signals were acquired using a flow cytometer (CytoFLEX Flow Cytometer, Beckman Coulter, Inc.) and analyzed using FlowJO.

EdU incorporation assay

Human PDAC cell lines were harvested, stained, washed and analyzed as discussed previously. Briefly, cells were plated on standard 24 well plates at initial concentrations of 5-10x10⁵ cells per well in full media. After 24 hours, cells culture media substituted with 2.5% fetal bovine serum (FBS) for 18 hours. They were then treated with DMSO, LEE011, MEK162, or a combination in culture media supplemented with 2.5% FBS for 48 hours. EdU was then added and cells were incubated for an additional 6 hours. Cells were then washed with PBS, trypsinized, filtered, and fixed overnight with 3% formalin. Cells were then re-suspended in a labeling solution (150mM Tris pH 8.5, 1.5mM CuSO4, 2μM fluor-Azide dye). Following labeling, cells were then washed in PBS, and additional labeling using propidium iodide was performed. Cells were then analyzed by flow cytometry (CytoFLEX Flow Cytometer, Beckman Coulter, Inc.) and the percentage of EdU positive cells reported(21).

Senescence associated β-galactosidase assay

Human PDAC cell lines were plated on standard 6-well plates at an initial density of 0.3 x 10⁶ cells per well. After allowing 24 hours for attachment, cells were treated with DMSO, LEE011, MEK162, or combination of LEE011 and MEK162 for 48 hours. Following treatment, cells were washed with cold PBS and fixed using 0.2% glutaraldehyde (Sigma Aldrich) at room temperature. Cells were then incubated overnight in β-galactosidase staining solution, protected from light(22). At least 4 representative images were taken from each well using the DM750 M imaging platform (Leica) and ImageJ v1.52k software with Fiji add-on was used for analysis(17,18).

Treatment of Ptf1a^cre/+;KRAS^G12D/+;Tgfbr2^flox/flox (PKT) mice

PKT mice were treated with vehicle (n=10, HPMC/Tween 80), LEE011 (n=9) (75 mg/kg/day), MEK162 (n=9) (3.5 mg/kg/day), or the combination of MEK162 and LEE011 (n=10) by oral gavage 5 days/week, starting at 4 weeks of age. Mice were euthanized and pancreatic tumor tissues obtained after 3 weeks of treatment were used for further analysis. The tumor sections were fixed in 10% neutral buffered formalin solution for 24 hours, embedded in paraffin and tissue sections were prepared. Immunohistochemical and immunofluorescence analysis was performed using specific antibodies listed in the Supplementary Table S1.

Isolation of pancreatic tumor cells from PKT mice

Pancreatic tumors harvested from PKT mice were minced using sterile scalpels and then enzymatically digested with 0.6mg/ml of collagenase P (Roche, #11213857001), 0.8 mg/ml Collagenase V (Sigma Aldrich, # C9263-1G), 0.6 mg/ml soybean trypsin inhibitor (Sigma Aldrich, #T6522-1G), and 1800 U/ml DNase I (Thermo Scientific, #EN0521) in RPMI medium for 20-30 minutes at 37°C. Samples were then washed and resuspended in cold PEB solution to quench the enzymatic reaction and prevent over-digestion of tissues. The dissociated tissue was strained through 40 μm mesh filter to obtain a single cell suspension. These single-cell suspensions were plated on 10cm² dish in complete media for 2-3 days. EpCAM positive, live PKT tumor cell populations were FACS-sorted using EpCAM/CD326 antibody (BioLegend, #118213) and used for subsequent downstream experiments.

Xenograft models

Subcutaneous tumors were established by injecting 2×10⁶ cells (Panc-1, MiaPaCa-2, or BxPC-3) into the flanks of Foxn1^nu/nu mouse (6 weeks old). Mice were treated in the same manner as PKT mice (described above). Treatment was initiated when the subcutaneous tumors reached 75 – 100 mm³ size. The subcutaneous tumor volume (V) was determined by caliper measurements obtained twice weekly and calculated using equation V = L × W² × 0.5, where L is length and W is the tumor’s width. Growth curves for tumors were plotted as the mean volume ± standard error of the mean (SEM) of mice’s tumors from each group. Animals were treated for thirty-one (31) days or until tumors became greater than 2,000 mm³ in size. All mice were sacrificed at the end of the study and primary tumors were removed for further analysis.

Statistical analysis

Descriptive statistics were calculated using Microsoft Excel and Prism software v8 (GraphPad Inc.). Results are shown as values of mean ± SEM unless otherwise indicated. Statistical analyses were performed using the ANOVA followed by Tukey’s multiple comparisons test or two-sided Student’s t-test where appropriate. Kaplan-Meier survival analysis was performed for survival analysis and differences between groups were assessed using the log-rank test. P-values less than 0.05 were deemed significant, except when indicated otherwise.

Study approval

All experimental animal protocols were approved by the IACUC of The University of Miami (#15-099 and #18-040) and Vanderbilt University and conducted according to the Association for the Assessment and Accreditation of Laboratory Animal Care guidelines.

Results

CDKN2A is downregulated in human PDAC cells and tissues

Previous studies have demonstrated that the downregulation of p16 expression is an often acquired trait in the progression of PDAC(23–25). We began by querying the Catalogue of Somatic Mutations in Cancer (COSMIC) database to verify both the extent and cause of CDKN2A loss in human pancreatic cancer (26). An alteration in the CDKN2A gene sequence was present in 84% (n=26) of pancreatic tumor samples contained in the database. The most common alterations were due to homozygous deletion (61%) and truncation (19%), while point mutations, promoter methylation, and intronic mutations occurred much less frequently (Supplementary Fig. S1).

To determine the extent of this downregulation in vitro, qPCR was performed on one normal human pancreatic ductal cell line (HPNE), HPNE with KRAS G12D Mutation (HPNE-KRAS), and five human PDAC cell lines (BxPC-3, Panc-1, MiaPaCa-2, CFPAC, Panc 10.05) to evaluate for mRNA levels of the naturally-occurring, cyclin-dependent kinase inhibitors CDKN2A (p16), CDKN1A (p21), and CDKN1B (p27). We observed minimal to null transcription of CDKN2A in any of the five human PDAC cell lines, while HPNE-KRAS displayed reduced CDKN2A transcription levels when compared to HPNE cells (Fig. 1A). Furthermore, transcription of other endogenous inhibitors of cyclin-dependent kinases, p21 and p27, was variable but robust in all tested cell lines indicating that transcriptional disruption of the tumor suppressor p16 is common in PDAC despite retained transcription of other endogenous cyclin-dependent kinase inhibitors. To evaluate p16 protein expression in human PDAC cell lines, we stained HPNE, HPNE-KRAS, and five human PDAC cell lines for p16 protein. Limited p16 staining was observed in any of the tested PDAC cell lines, while both PanIN and HPNE cell lines displayed moderate to high levels of p16 (Fig. 1B).

Figure 1.

CDKN2A is downregulated in human PDAC cells and tissues. (A) Results of qPCR assay performed on HPNE, HPNE with KRAS G12D Mutation (HPNE-KRAS), and five human PDAC cell lines (BxPC-3, Panc-1, MiaPaCa-2, CFPAC, Panc 10.05) which demonstrates loss of transcriptional expression of CDKN2A in all PDAC cell lines and minimal expression in HPNE-KRAS despite retained expression of other cell cycle inhibitors, p21 and p27. All outputs normalized to GAPDH and fold change were calculated with respect to HPNE. (B) HPNE, HPNE-KRAS, and five human PDAC cell lines were stained using immunofluorescence markers for CDKN2A gene product, p16 (Alexa Flour 594), in red and nuclear stain in blue. Expression of p16 was observed in HPNE and HPNE-KRAS cells only and there is a complete loss of p16 expression in the five PDAC cell lines. (C) The left panel shows representative images from histologic sections of human PDAC tumors and normal adjacent pancreatic tissue (NAT) stained for p16 with quantification of microarray shown in the right panel. Case matched controls were only available for 23 out of 30 pancreatic cancer tissue samples and one set was excluded due to lack of cellular material resulting in a total of 22 evaluated cases with case-matched controls. Loss of p16 expression is observed in tumors with retained expression in normal tissue. Images shown at x20 and x40. P value: ****p<0.0001.

Next, we sought to determine the expression of p16 protein in human pancreatic cancer tissue samples. A tissue microarray (TMA) containing 30 PDAC tumors with case-matched normal adjacent tissue (NAT) was stained for p16, revealing high levels of p16 in all normal pancreatic tissue. In contrast, its expression was dramatically reduced or absent in all case-matched PDAC tissue sections (Fig. 1C). Taken together, these results demonstrate that dysregulation of CDKN2A is a common occurrence in PDAC and results in a loss of p16 protein expression. Thus, an exogenous analog for p16 may have promise in the treatment of PDAC.

Pharmacologic inhibition of MEK and CDK4/6 is concentration and time-dependent

To examine the effects of CDK4/6 inhibition in vitro, we selected three human PDAC cell lines (BxPC-3, Panc-1, and MiaPaCa-2) that recapitulate the most common genetic alterations seen in PDAC. The BxPC-3 cell line is KRAS wild-type, whereas the Panc-1 and MiaPaCa-2 both contain activating KRAS mutations.(27) Each of these cell lines exhibits inactivation of TP53 and CDKN2A, and only BxPC-3 demonstrates inactivated SMAD4. The selection of these cell lines was also influenced by previous work that established the IC₅₀ values of a highly selective MEK inhibitor (MEK162) in these PDAC cell lines(28). All cell lines were then exposed to increasing dosages of ribociclib (LEE011), a highly selective small molecule inhibitor of CDK4/6, which demonstrated significantly decreased pRb in a concentration and time-dependent fashion (Figs. 2A and B). After treating all cell lines with escalating doses of MEK162, we observed reduced phosphorylated ERK (pERK) levels at all doses and in a time-dependent manner (Figs. 2C and D). Finally, ctreatment with MEK162 and LEE011 alone and in combination resulted in suppressing both pERK and pRb levels in all cell lines (Fig. 2E). Although the most significant reduction in the levels of both pRb and total Rb (tRb) was observed with LEE011 treatment, we observed that MEK162 monotherapy resulted in a slight reduction in both pRb and tRb in the MiaPaCa-2 cell line, a finding that had been previously reported in other studies(28).

Figure 2.

Pharmacologic inhibition of MEK and CDK4/6 is concentration and time-dependent. (A) CDK4/Cyclin D1 induces constitutive inactivating Rb phosphorylation (pRb) in all cell lines. The CDK4/6 inhibitor, LEE011, has no direct effect on CDK4 or cyclin D1 levels but rather restricts its kinase activity, reducing Rb’s phosphorylation in a dose-dependent manner in three different PDAC cell lines after 24hrs of treatment in culture. (B) Reversal of Rb inactivating phosphorylation by LEE011 treatment (1μM) is time-dependent and requires 24hrs for maximal effect in all cell lines. (C) ERK is constitutively activated in KRAS-wildtype (BxPC-3) and KRAS-mutant (Panc-1 and MiaPaCa-2) cell lines. MEK162 inhibits MEK’s kinase activity, reducing the activating phosphorylation of ERK (pERK) at all tested doses in all cell lines after 24 hours of treatment in culture. (D) MEK162 (5μM) exhibits rapid onset activity within 1hr in Panc-1 and MiaPaCa-2, but requires 24hrs for maximal effect in BxPC-3. (E) Combined LEE011 (1μM) and MEK162 (5μM) reduces p Rb and pERK, respectively at 24hrs.

Combined inhibition of CDK4/6 and MEK reduces PDAC cell proliferation

To characterize the effect of combined MEKi and CDK4/6i in vitro, BxPC-3, Panc-1, and MiaPaCa-2 cells were treated with LEE011, MEK162, or the combination of both drugs and proliferating cell fractions in response to treatment were measured using EdU assay. In BxPC-3 cells, a significant decrease in proliferation was seen with individual treatment with LEE011 and MEK162 as well as the combination treatment (Fig. 3A). Compared to single-agent treatment, the combination of LEE011 and MEK162 resulted in a significant decrease in proliferation in Panc-1 and MiaPaCa-2 cells (Figs. 3B and C).

Figure 3.

Combined inhibition of CDK4/6 and MEK reduces PDAC cell proliferation. (A-C) Shown are results from the EdU cell proliferation assay performed on BxPC-3, Panc-1, and MiaPaCa-2 cell lines following 48 hours of treatment with DMSO, LEE011 (1μM), MEK162 (5μM), or a combination. Combined LEE011 and MEK162 reduce cell proliferation at 24 hrs in all tested cell lines. (D-F) Each cell line was then plated in standard 12 well plates at an initial density of 1 x 10⁵ cells per well and treatment was initiated after 24 hours in culture with DMSO, LEE011 (1 μM), MEK16 (5 μM), and combination. Cell density was evaluated at two-day intervals until treatment was terminated on day 9. Long term combination shows sustained inhibition of cell proliferation by 2D colony formation assay in all tested cell lines. Right panels show results of relative fold-change calculation after 9 day treatment. P value: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ^nsnot significant.

To examine the response to long-term treatment with these agents, longitudinal growth in Panc-1, MiaPaCa-2, and BxPC-3 cell lines were assessed using clonogenic assays over 10 days. All cell lines demonstrated a significant reduction in cell density with combined MEK162 and LEE011 treatment compared with either vehicle or monotherapy (Figs. 3D, E and F). These results demonstrate that combined MEKi and CDK4/6i reduce PDAC proliferation.

Combined MEK and CDK4/6 inhibition reduces PDAC growth by inducing cellular senescence

To determine the mechanism of PDAC tumor growth inhibition, we performed cell cycle analysis by flow cytometry on the three cell lines following treatment with LEE011 and MEK162 individually and in combination. In cell cycle analysis, BxPC-3, Panc-1 and MiaPaCa-2 cell lines displayed reduced S phase fraction and increased G₀/G₁ phase arrest, confirming inhibition of cell proliferation (Fig. 4A). Interestingly, we observed a significant increase in G₀/G₁ phase arrest with combined MEK162 and LEE011 treatment compared with all other treatment groups in BxPC-3 and Panc-1 cell lines. The early and late apoptotic fractions in treated cell lines were examined by observing the Sub G₀ percentage and Annexin-V⁺ staining. Interestingly, we did not observe any consistent increase in early or late apoptosis in any of the tested cell lines following treatment with combined MEK162 and LEE011 (Supplementary Figs. S2A – S2F), suggesting that the mechanism of growth inhibition with this combined treatment is not due to the induction of apoptosis in PDAC cells.

Figure 4.

Combined MEK and CDK4/6 inhibition reduce PDAC growth by inducing cellular senescence. (A) Cell cycle analysis of BxPC-3, Panc-1, and MiaPaCa-2 cells following 48 hours of treatment with DMSO, LEE011 (1μM), MEK162 (5μM), or combination shows a reduction in S-phase fraction following combination treatment with a concomitant increase in G0/G1 fraction, suggesting cell cycle arrest at G1/S checkpoint in all cell lines. (B-D) Each cell line was then plated in standard 6 well plates at a density of 3 x 10⁵ cell cells per well and treated for 48 hours with DMSO, LEE011 (1μM), MEK162 (5μM), or a combination. Treated cells were stained for β-galactosidase, a surrogate marker for cell senescence, and analyzed using brightfield microscopy. The right panel shows representative images and the left panel shows quantification of staining, which reveals that senescent cell fraction increased with LEE011 and MEK162 treatment in all cell lines, with the greatest increase noted in combination-treated cells. Increased senescence was also associated with smaller colonies, enlarged cells, and decreased cell number per high powered field. P value: *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

To further characterize this effect, the β-galactosidase assay was used to determine senescence following treatment with LEE011 and MEK162. Combined inhibition resulted in a significant increase in β-galactosidase staining when compared to monotherapy in all cell lines, suggesting that combined inhibition of CDK4/6 and MEK with LEE011 and MEK162, respectively, reduces proliferation through the induction of cellular senescence rather than through pro-apoptotic mechanisms in PDAC cell lines (Figs. 4B, C and D).

Combined MEK and CDK4/6 inhibition decreases tumor growth in PDAC xenografts

To determine the effect of combined CDK4/6i with MEKi in vivo, we utilized a flank tumor xenograft model in athymic nude mice with MiaPaCa-2, Panc-1, or BxPC-3 cell lines. Treatment using the combination of LEE011 and MEK162 significantly attenuated tumor growth compared with vehicle treatment and monotherapy alone in BxPC-3 and Panc-1 xenografts, with minimal tumor growth occurring following the start of the combination therapy (Fig. 5A and B). While combination therapy in MiaPaCa-2 xenografts was also able to significantly suppress tumor growth compared with vehicle and MEK162 monotherapy, tumor growth inhibition in this cell line was comparable to treatment with LEE011 alone (Fig. 5C). The combination therapy was well tolerated and did not result in any in-vivo toxicity (Supplementary Figs. S3A–S3C).

Figure 5.

Combined MEK and CDK4/6 inhibition decrease tumor growth in PDAC xenografts. Flank tumor xenografts were established in athymic, Foxn1^nu−/nu− mice using BxPC-3, Panc-1, or MiaPaCa-2 cells. Mice were then assigned to receive treatment with vehicle, LEE011 (75 mg/kg/day), MEK162 (3.5 mg/kg/day) or combination by oral gavage after tumors reached 75 mm³ in size. (A) In BxPC-3 xenografts, all treatments significantly inhibited growth, but the greatest reduction was seen in combination therapy (B) Panc-1 xenograft growth was significantly reduced by combination therapy alone. (C) MiaPaCa-2 xenograft growth was significantly suppressed by LE011 as well as combination therapy. P value: *p<0.05, ****p<0.0001.

Combined MEK and CDK4/6 inhibition significantly increases survival in Ptf1a^cre/+;KRAS^G12D/+;Tgfbr2^flox/flox (PKT) mice

Having established that combined CDK4/6i and MEKi reduces tumor growth both in vitro and in vivo, the effect of this therapy on overall survival was assessed using PKT mice, a highly aggressive, genetically engineered mouse model (GEMM) of PDAC(14). PKT mice were treated with vehicle, MEK162, LEE011, or a combination of MEK162 and LEE011 starting at 4 weeks of age (Supplementary Fig. S4A). The median survival of the vehicle treatment group was 59 days. Monotherapy with MEK162 or LEE011 showed no significant difference in median survival (63 days and 68 days, respectively) compared with vehicle treatment. However, combined treatment with MEK162 and LEE011 resulted in significantly improved median survival of 154.5 days (p = 0.0001, log-rank test) (Fig. 6A). The combination therapy was well tolerated and did not result in any in vivo toxicity (Supplementary Fig. S4B).

Figure 6.

Combined MEK and CDK4/6 inhibition significantly increases survival in Ptf1a^cre/+;KRAS^G12D/+;Tgfbr2^flox/flox (PKT) mice. PKT mice were treated with vehicle, LEE011 (75 mg/kg/day), MEK162 (3.5 mg/kg/day), or combined LEE011/MEK162 by oral gavage 5 days/week, starting at 4 weeks of age. (A) Kaplan-Meier survival analysis shows significantly improved overall survival with LEE011 and MEK162 (median 154.5 days) compared with vehicle control (median 59 days). Neither monotherapy alone produced a significant effect with a median survival of 59-68 days. However, mice who received both drugs in combination had dramatically increased median survival to over 154 days. P value: ***p=0.0001, log-rank test. (B) Histologic sections of PDAC tumors were obtained from a separate cohort of mice similarly treated for three weeks as above. The left panel shows representative images from stained sections for H&E, cleaved caspase-3 (a marker for apoptosis), Ki67 (a marker for proliferation), and pRb (an inverse marker of senescence) and the right panel shows results of staining quantification. As measured by cleaved caspase-3 staining, there was no increase in apoptosis in any treatment group, and the combination-treated tumors had a significant decrease in proliferative fraction, as measured by Ki67 staining. pRb is the deactivated form of Rb and is an inverse marker of senescent cells with higher levels seen in dividing cells and lower levels seen in senescent cells. We found that combination-treated tumors had a significant increase in cellular senescence, as measured by an absence of pRB staining compared to control tumors. (C) PKT tumor cells were treated with LEE011, MEK162, or combination of LEE011/MEK162 and the senescent fraction was analyzed by staining for the presence of β-galactosidase. Cells treated with LEE011 alone or combination with MEK162 showed increase in β-galactosidase staining. P value (B and C): *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ^nsnot significant.

Histologic analysis demonstrated no change in cleaved caspase 3 (CC-3) levels across all treatment groups, suggesting no significant increase in apoptosis produced by the combined treatment, which is consistent with our in vitro observations (Fig. 6B). To examine if treatment results in a senescent phenotype in vivo, we performed immunofluorescence for pRb in histologic specimens from treated mice. pRb staining was highest in vehicle-treated mice while mice treated with LEE011, alone and in combination with MEK162, showed a significant reduction in pRb staining, suggesting induction of senescence in these tumors. Combined MEKi and CDK4/6i therapy additionally showed a significant decrease in Ki67 levels compared to vehicle or individual treatment alone, mirroring ours in vitro findings (Fig. 6B).

To further explore the effect of combined MEKi and CDK4/6i therapy on senescence in this tumor model, PKT tumor cells were isolated from untreated, tumor-bearing mice and treated with MEKi and CDK4/6i alone and in combination. The senescent fraction was analyzed by staining for the presence of β-galactosidase. Cells treated with LEE011, alone and in combination with MEK162, had a significant increase in β-galactosidase staining, indicating the induction of cellular senescence in response to MEKi and CDK4/6i therapy in PKT tumor cells (Fig. 6C).

Discussion

Although there is significant genetic heterogeneity among tumors, certain cancers display “oncogene addiction”, whereby they become critically dependent on mutations in specific genes for sustained growth(29,30). Activating KRAS mutations are believed to be an indispensable driver of PDAC tumor initiation and progression, as evidenced by the nearly ubiquitous presence of oncogenic RAS pathway alterations in human PDAC tumors(31–33). However, the development of small-molecule inhibitors to successfully inhibit mutant KRAS has been largely unsuccessful, highlighting the need to further explore combination strategies that target the multiple downstream mediators of RAS signaling to overcome resistance.

In the present study, we demonstrate that combined MEKi and CDK4/6i is a novel combinatorial strategy to target the RAS-dependent pathways in PDAC. We show the efficacy of MEKi therapy is significantly enhanced by the addition of a CDK4/6 inhibitor in PDAC in vitro and in vivo. Combined inhibition of CDK4/6 and MEK simultaneously reduces phosphorylation of the Rb (pRb) and decreases pERK levels, thereby inhibiting cell cycle progression in human PDAC cell lines. We propose that the efficacy of this regimen is not due to the induction of apoptosis, but rather a decrease in cell proliferation and the induction of cellular senescence in both human PDAC cell lines and murine tumors, as we observed a significant increase in senescence-associated markers with combined therapy. Treatment of PKT mice with combined MEKi and CDK4/6i significantly increases the median overall survival in the highly aggressive PKT GEMM of PDAC.

Although PKT mice do not have a mutation or deletion in the CDKN2A gene, the TGF-β/SMAD signaling axis has been shown to play a critical role in regulating cell cycle progression(34). TGF-β signaling via SMAD proteins promotes transcription and activation of p21, p27, and p15(Ink4b), which act on cyclin-D, cyclin-E and CDK4/6 to prevent cell cycle progression(35,36). Although individually, these are less potent inhibitors of the cyclin-dependent kinases than p16, their combined downregulation due to the loss of TGF-β signaling may be responsible for CDK4/6 inhibitor treatment efficacy in this model. However, uncovering genomic alterations that may serve as predictive biomarkers of response to MEKi/CDK4/6i will be critical to the translational applicability of this regimen and should be explored in future studies.

Recent preclinical studies in non-pancreatic, solid organ malignancies have demonstrated an increase in apoptosis via both caspase and Bcl-2 mediated mechanisms following MEKi(37–39). Our combined MEKi and CDK4/6i results did not show any effect on apoptosis but instead produced senescence of tumor cells. It is known that inhibition of CDK4/6 can result in different cell fates such as senescence (40) and our results demonstrate that this pathway can be therapeutically exploited to arrest tumor growth. This finding is also particularly interesting as senescent cell survival is uniquely dependent upon the upregulation of numerous anti-apoptotic and immunomodulatory proteins that allow them to thrive in an otherwise toxic, inflammatory, or nutrient-poor condition in the PDAC tumor microenvironment(41,42). Senolytic drugs that selectively target the pathways necessary for the survival of senescent cells have already been shown to restore tissue homeostasis and reverse the effects of doxorubicin-induced chemotoxicity models of premature aging(43,44). Additionally, they have also been shown to induce xenograft regression in multiple tumor cell lines following Rb-dependent senescence induction (45). Thus, sequential treatment with drugs that induce cell death in senescent cells might be an additional strategy to potentiate the antitumor effects in PDAC.

The efficacy of CDK4/6i in PDAC has been previously explored. In a study by Franco et al., combination of CDK4/6 inhibitors with other pathway selective agents demonstrated a differential response to CDK4/6 inhibitors in several PDAC cell lines. Although treatment with a CDK4/6 inhibitor results in significant suppression of pRb, there may be residual low levels of persistent pRb that could be abrogated by the addition of selective antagonists such as PI3K/mTOR inhibitors that result in durable growth suppression of tumors(46). More recent work by Maust et al. has highlighted the cyclooxygenase-2 (COX-2) enzyme as an important mediator of response to MEKi and CDK4/6i in PDAC(47). Their study identified that certain PDAC cell lines responded more favorably to the combination of MEK and CDK4/6 inhibition (47). This suggests the potential applicability of MEK and CDK4/6 inhibitors in this class of tumors as part of a combinatorial regimen with anti-inflammatory agents.

We have demonstrated that the co-targeting of MEK and CDK4/6 induces a senescent phenotype in PDAC, resulting in a significant reduction in tumor growth. In addition, our results suggest the effectiveness of this combination is not dependent solely on CDKN2A deletion or mutation, as demonstrated by the dramatic response observed in the PKT mouse model. Further investigation into the molecular subtypes that predict response to this treatment combination is warranted and would significantly support the clinical use of the combined MEK and CDK4/6 inhibition in the treatment of PDAC.