The Role of Inherited Pathogenic CDKN2A Variants in Susceptibility to Pancreatic Cancer

Abstract

CDKN2A is cell cycle negative regulator and the role of CDKN2A in the development of pancreatic ductal adenocarcinoma, which is continues to be a lethal cancer, is well-established. Somatic loss of CDKN2A is considered one of the major drivers of pancreatic tumorigenesis. CDKN2A gene is one of the pancreatic cancer susceptibility gene and in addition to melanoma, pathogenic germline CDKN2A variants have been identified in up to 3.3% patients with pancreatic ductal adenocarcinoma depending on family history of disease. Carriers of a known pathogenic germline CDKN2A variant have up to a 12.3-fold increased risk of developing pancreatic cancer. Recently several studies have demonstrated the benefit of clinical surveillance in patients with pathogenic germline CDKN2A variants. Therefore, identification of patients with a pathogenic germline CDKN2A variant is important for screening of at-risk relatives for pancreatic cancer. It has the potential to leads detection of early, potentially curable, pancreatic cancer and precursor neoplasm and reduce mortality. Furthermore, patients with a germline pathogenic CDKN2A variant and somatic loss of CDKN2A may benefit in future from treatment with targeted therapies such as a CDK4/6 inhibitor.

Introduction

Advances in cancer therapy and screening in the last 30 years have led to significant improvement in survival for most types of cancer.¹ However, pancreatic ductal adenocarcinoma (PDAC) continues to be a lethal cancer, with a median survival rate is less than 6 months and a 5-year survival rate of only 10%.² In the United States, PDAC is estimated to become the second leading cause of cancer-related mortality by the year 2030.³ Novel approaches to risk assessment, early detection, and treatment for PDAC are critical to improving patient care.

Family history of PDAC is an important risk factor, with first-degree relatives of a patient with PDAC having a 2-to-3-fold increased risk of developing the disease.⁴ Tragically, some families have more than one member with PDAC. These families, with at least two affected first-degree relatives, represent approximately 5–10% of patients with PDAC and are designated as having “familial pancreatic cancer” (FPC).^5,6 Risk of PDAC increases with an increasing number of affected relatives; individuals with one, two, or three affected first-degree relatives have a 4.6-, 6.4-, or 32.0-fold increased risk of developing PDAC, respectively.⁷

Pathogenic germline variants in the ATM, BRCA1, BRCA2, CDKN2A, CPA1, CPB1, MLH1, MSH2, MSH6, PALB2, PMS2, PRSS1, and STK1 genes are associated with an increased risk of PDAC and are found in 8.0–11.0% of patients with FPC and 4.0 – 5.2% of patients without a family history of the disease.^8–12 Understanding inherited risk of PDAC is essential to identify patients with PDAC that may benefit from targeted therapies. For example, patients with PDAC and a pathogenic variant in BRCA1, BRCA2, or PALB2 may have tumors deficient in homology-directed DNA repair and consequently be exquisitely sensitive to poly(ADP)-ribose polymerase inhibitors.¹³ Similarly, patients with PDAC and a pathogenic variant in a Lynch syndrome gene (MLH1, MSH2, MSH6, and PMS2) may have tumors deficient in mismatch DNA repair that respond to immunotherapy.¹⁴ Furthermore, knowing the genes responsible for inherited risk allows identification of at-risk relatives eligible for screening for pancreatic and extra-pancreatic neoplasms.¹⁵

CDKN2A is a gene located at chromosome 9p2¹⁶ that encodes two proteins, p16^INK4a and p14^ARF, that control of cell proliferation.^17,18 p16^INK4a induces G1cell-cycle arrest by inhibiting CDK4 and preventing phosphorylation of the retinoblastoma protein.¹⁷ p14^ARF, however, inhibits the function of HDM2, resulting in increased p53 levels.¹⁹ CDKN2A has a prominent role in inherited cancer, in particular, in familial atypical multiple-mole melanoma (FAMMM), an autosomal dominant syndrome characterized by malignant melanomas in association with multiple atypical nevi.²⁰ Pathogenic germline CDKN2A variants have been linked to up to 40% of familial melanoma cases²¹ and is associated with higher prevalence of atypical nevi.²² Pathogenic germline CDKN2A variants also increase the risk of pancreatic cancer.^8,10,23,24 Carriers of pathogenic germline CDKN2A variants also have an increased risk of lung, head and neck, and gastro-esophageal cancers.^25,26 Somatic loss-of-function of the CDKN2A gene products by mutation, deletion, or promotor methylation has been observed in various cancers, such as melanoma, PDAC, gastric cancer, colon cancer, non-small cell lung cancer.^16,27–30 The role of CDKN2A in the development of PDAC is well-established and somatic mutations in CDKN2A are considered one of the major drivers of pancreatic tumorigenesis, along with somatic mutations in KRAS, TP53, and SMAD4.^31–34

In this review, we detail the current understanding of the role of inherited pathogenic variants in CDKN2A in development of PDAC, the implications for PDAC screening in carriers of pathogenic germline CDKN2A variants, and potential therapeutic implications in CDKN2A deficiency.

Pathogenic Germline CDKN2A Variants in Patients With PDAC

In 1968, Lynch and Krush reported that some family members in FAMMM kindreds developed pancreatic cancer.³⁵ Additional kindreds with FAMMM in which a family member developed PDAC have been reported. Bergmann and colleagues reported that the number of PDAC cases observed in FAMMM family members was increased compared to the expected incidence of PDAC.³⁶ Furthermore, Goldstein and colleagues estimated that melanoma-prone families with a pathogenic germline CDKN2A variant have a 22-fold risk of PDAC.³⁷ These data, provide clinical evidence for the association between PDAC and malignant melanoma.^36,38 Importantly, as noted by Moskaluk and colleagues, some individuals with PDAC and a pathogenic germline CDKN2A variant do not have a family history of melanoma.³⁹

Subsequent studies have confirmed the role of pathogenic germline CDKN2A variants in susceptibility to PDAC and found that between 1.5–3.3% of patients with FPC carry a pathogenic germline CDKN2A variant (Table 1).^{8,10,23,40,41} Importantly, pathogenic germline CDKN2A variants identified in patients with PDAC have diverse functional consequences, and they include nonsense variants, frameshift insertions, frameshift deletions, splicing variants, missense variants.^{8,10,23,40,41} Most studies reporting pathogenic germline CDKN2A variants in patients with PDAC have used targeted sequencing of the CDKN2A coding region to identify deleterious variants. The spectrum of germline variants seen in PDAC patients in FPC kindreds is similar to that seen in individuals with familial melanoma. A recent study analyzing the whole genomes of 638 PDAC patients in FPC kindreds reported a similar number and spectrum of pathogenic variants as previous studies.²³ Specifically, 16 of 638 patients had a pathogenic germline CDKN2A variant and these included 1 nonsense variant, 3 frameshift deletions, and 12 missense variants.²³ The paucity of pathogenic CDKN2A variants in non-coding regions, when considering that whole genome sequence was analyzed, is likely due to the difficulty of interpreting these variants.²³

TABLE 1.

Prevalence of Germline CDKN2A Variants in Patients With FPC

Total No. Patients	Pathogenic Variants, n (%)	VUS, n (%)	Study, Year
120	4 (3.3)	—	McWilliams et al, 201140
515	13 (2.5)	10 (1.9)	Zhen et al, 201510
638	16 (2.5)	17 (2.5)	Roberts et al, 201623
185	4 (2.2)	8 (4.3)	Chaffee et al, 201841
343	5 (1.5)	—	Hu et al, 20188

FPC indicates familial pancreatic cancer; VUS, variants of unknown significance.

Most patients with PDAC do not meet diagnostic criteria for FPC and the role of pathogenic germline CDKN2A variants in this patient population is less well studied. However, some patients with PDAC who do not meet diagnostic criteria for FPC have pathogenic germline CDKN2A variants. Shindo and colleagues analyzed the prevalence of pathogenic germline variants in 32 genes, including CDKN2A and other established pancreatic cancer susceptibility genes, in 854 patients with PDAC unselected for family history and identified one patient (0.1%) with a pathogenic frameshift insertion variant in CDKN2A gene.¹² Similarly, Hu and colleagues analyzed the prevalence of pathogenic germline variants in 21 genes in 3030 patients with PDAC unselected for family history. In this study, a statistically significant association was observed between PDAC and pathogenic germline CDKN2A variants (0.3% of cases vs 0.02% of controls). Furthermore, the odds ratio (OR) for CDKN2A was 12.33 (95% confidence interval, 5.43–25.61).⁸ Subsequent studies of patients with PDAC either without a family history of PDAC or unselected for family history have identified pathogenic germline CDKN2A variants in up to 2.6 % patients, with a similar spectrum of mutations as observed in patients with FPC (Table 2).^42–46

TABLE 2.

Prevalence of Germline CDKN2A Variants in Patients With Nonfamilial PDAC and Unselected for Family History

Patient Family History	Total No. Patients	Pathogenic Variants, n (%)	VUS, n (%)	Study, Year
Nonfamilial PDAC	1410	5 (0.4)	NR	McWilliams et al, 201140
	201	0	3 (1.5)	Zhen et al, 201510
	117	0	1 (0.9)	Chaffee et al, 201841
Unselected PDAC	854	1 (0.1)	1 (0.1)	Shindo et al, 201712
	615	6 (1)	–	Lowery et al, 201843
	3594	40 (1.1)	–	Singhi et al, 201944
	298	1 (0.3)	–	Brand et al, 201845
	289	2 (0.7)	–	Yurgelun et al, 201942
	350	9 (2.6)	10 (2.9)	McWilliams et al, 201846
	3030	9 (0.3)	–	Hu et al, 20188

Nonfamilial PDAC are patients with pancreatic adenocarcinoma (PDAC) and no family history of PDAC or a family history that does not meet criteria for familial pancreatic cancer. Unselected PDAC are patients with PDAC with or without a family history of PDAC.

NR indicates not reported; VUS, variants of unknown significance.

Germline CDKN2A Variants of Unknown Significance in Patients With PDAC

Germline variants in CDKN2A can be classified as benign, of uncertain significance, or pathogenic based on American College of Medical Genetics guidelines.⁴⁷ While benign and pathogenic variants are clinically actionable, CDKN2A variants of unknown significance (VUS) are the cause of significant uncertainty and worry for patients and their health care providers.⁴⁸ For example, an individual with a pathogenic germline CDKN2A variant and a family history of pancreatic cancer would be eligible for clinical surveillance programs based on current American Society of Clinical Oncology (ASCO) guidelines.^15,49 In contrast, a similar individual with a CDKN2A VUS would not meet the threshold required for screening, assuming no other criteria were met. The identification of CDKN2A VUS, however, is not an uncommon occurrence in patients with PDAC undergoing germline genetic testing. Several research studies, utilizing Clinical Laboratories Improvement Amendments (CLIA) certified and research laboratories, have reported CDKN2A VUS in 1.9–4.3% of patients with FPC and 0.1–2.9% of patients with PDAC without a family history of the disease or unselected for family history (Table 1 and Table 2).^{10,12,23,41,46} In these studies, CDKN2A VUS are predominantly missense variants (40 of 50 reported VUS; 80%), with noncoding single base substitutions (8 of 50 reported VUS; 16%) and a frameshift deletion in p16^INK4A (1 of 50 reported VUS; 2%).^{10,12,23,41,46}

Germline CDKN2A Variants in Diverse Populations

The prevalence of pathogenic germline variants in pancreatic cancer susceptibility can be notably different in different populations, as is the case for the Ashkenazim where the prevalence of pathogenic germline variants in BRCA1 and BRCA2 is higher than in the general population, and indeed, malignancies including pancreatic cancer, is higher.⁵⁰ However, the majority of studies investigating inherited risk in patients with PDAC have had a study population that is predominantly non-Hispanic White.⁴⁰ Few studies have presented data on the prevalence pathogenic germline variants in other populations (Table 3).^12,46,51,52 This bias leads to an incomplete understanding of inherited pancreatic cancer risk.^12,46 McWilliams and colleagues reported the prevalence of pathogenic germline CDKN2A variants in 220 African Americans, 119 Hispanic, and 11 Native American patients with PDAC.⁴⁶ In this study, pathogenic germline CDKN2A variants were found in three (1.4%) African American, four (3.4%) Hispanic, and one Native American (9.1%) patient with PDAC. Others have reported germline CDKN2A VUS in 1.9–4.5% of Africa-American patients with PDAC.^12,46 Takai and colleagues, however, found no pathogenic germline CDKN2A variants in 135 Asian patients with FPC.^51,52 While these data suggest that the prevalence of pathogenic germline CDKN2A variants may be higher or lower in different populations, the small size of these studies and ability to classify variants in different populations mean that firm conclusions cannot be made. Regardless, it is clear that further study of diverse populations of patients with PDAC is necessary to a complete understanding of inherited pancreatic cancer and ensure that all patients benefit from recent advances in patient care.

TABLE 3.

Prevalence of Germline CDKN2A Variants in Patients With PDAC by Ancestry

Patients	No. Germline CDKN2A Variants, n (%)					Study, Year
	Non-Hispanic White	African American	Asian	Native American	Hispanic
FPC	4 (3.3)	–	–	–	–	McWilliams et al, 201140
	13 (3.2)	0 (0)	0 (0)	0 (0)	–	Zhen et al, 201510
	4 (2.2)	0 (0)	–	–	–	Chaffee et al, 201841
	–	–	0 (0)	–	–	Takai et al, 201651
	–	–	0 (0)	–	–	Takai et al, 202052
Nonfamilial PDAC	5 (0.4)	–	–	–	–	McWilliams et al, 201140
	1 (0.1)*	0 (0)†	–	–	–	Shindo et al, 201712
Unselected PDAC	9 (0.3)	0 (0)	0 (0)	–	0 (0)	Hu et al, 20188
	–	3 (1.4)‡	–	1 (9.0)§	4 (3.4)‖	McWilliams et al, 201846

^*0 (0) CDKN2A VUS reported.

^†1 (1.9) CDKN2A VUS reported.

^‡10 (4.5) CDKN2A VUS reported.

^§0 (0) CDKN2A VUS reported.

^‖0 (0) CDKN2A VUS reported.

Germline CDKN2A Variants and Risk of Pancreatic Cancer

While age-specific pancreatic cancer risk estimates are not available for carriers of a pathogenic germline CDKN2A variant, estimates of risk are available from studies of melanoma-prone families and cohorts of patients with PDAC (Table 4).^{25,26,53–56} Relative risk of pancreatic cancer in relatives of melanoma patients with CDKN2A mutation was 7.4–14.8 when compared relatives without a mutation and 39–52 when compared to the general population.^{25,26,53–56} Similarly, cohort studies of patients with pancreatic cancer have reported that the hazard ratio was 4.9 in carriers compare to non-carriers. Furthermore, risk was higher in carriers who had ever smoked (hazard ratio, 25.8).^8,40 Risk of PDAC in melanoma-prone families may be dependent on the specific pathogenic germline CDKN2A variant. Goldstein and colleagues analyzed 446 melanoma-prone families that were part of the Melanoma Genetics Consortium (GenoMEL) and found that of 178 families with a pathogenic germline CDKN2A variant, 66 (15%) had a history of PDAC. The prevalence of pathogenic germline CDKN2A variants was significantly lower in families without PDAC. There were also significant differences in the prevalence of PDAC in melanoma-prone families based on the specific pathogenic germline CDKN2A variant identified (Table 5).⁵⁷ Specifically, more than 60% of families with the pathogenic variants p.A76fs and p.R112L113insR had a history of PDAC. Whereas, less than 11% of families with the pathogenic variants p.M53I and c.IVS2–10A>G had a history of PDAC.⁵⁷ Importantly, PDAC can occur in individuals with a pathogenic germline CDKN2A variant without a personal or family history of melanoma. For example, McWilliams and colleagues reported that seven (78%) out of nine PDAC patients with pathogenic germline CDKN2A variants had no personal and family history of melanoma.⁴⁰ Similarly, Harinck and colleagues reported that three (50%) out of six patients with FPC and a pathogenic germline CDKN2A variant had no family history of melanoma.⁵⁸

TABLE 4.

The Risk of CDKN2A Variant Carrier for Cancer in Melanoma-prone Family

Carriers*	Types of Cancer	Cancer Cases*	Relative Risk (95% CI)	Variant*†	Study, Year
221 (22)	Pancreatic	9	47.9 (21.7–90.0)‡	p.A76fs	Snoo et al, 200856
	Skin (non-melanoma)	10	37 (17.8–68.1)‡
	Lip, mouth, pharynx	3	15.5 (3.3–46.1)‡
	Respiratory	6	4.9 (1.8–10.6)‡
	Digestive	6	3.7 (1.4–8.1)‡
120 (28)	Melanoma	35	64.8 (36.9–117.9)‡	p.R112_L113insR	Helgadottir et al, 201425
	Pancreatic	7	43.8 (13.8–139.0)‡
	Upper digestive	6	17.1 (6.3–46.5)‡
	Respiratory	5	15.6 (5.4–46.0)‡
	All non-melanoma	39	5.0 (3.7–7.3)‡
	Skin (non-melanoma)	3	3.3 (1.0–10.7)‡
	Breast	3	3.0 (0.9–9.9)‡
80 (9)	Pancreatic	6 (4)	39 (14–84)‡	p.R112_L113insR	Borg et al, 200054
	Breast	8 (6)	3.8 (1.6–7.5)‡
252 (14)	Pancreatic	4 (4)	14.8 (4.0–37.9)§	p.G101W	Ghiorzo et al, 199953
	Breast	7 (3)	1.9 (0.4–5.6)§
117 (15)	Pancreatic	4 (3)	52 (13–132)‡	p.R87P (1), p.V126D (2)	Goldstein et al, 200455
	Gastrointestinal	6 (4)	9.1 (3.3–19.8)‡	p.R87P (1), p.G101W (1), p.V126D (2)
	All non-melanoma	12 (7)	2.3 (1.1–4.4)‡	p.M1_S8dup (1), p.N71S (1), p.R87P (1), p.G101W (1), p.V126D (3)
429 (65)	Pancreatic	5	7.4 (2.3–18.7)§	NR	Mukherjee et al, 201226
	Gastrointestinal	20	2.4 (1.4–3.7)§
	Colorectal cancer	10	1.9 (0.9–3.4)§
	All non-melanoma	56	1.5 (1.2–2.0)§

^*Number of families given in parentheses.

^†Variant position given in reference to NP_000068.1.

^‡Relative risks were calculated among the relatives of melanoma patients with CDKN2A variants and the general population.

^§Relative risks were calculated among the relatives of melanoma patients with CDKN2A variants and that of melanoma patients without CDKN2A variants.

CI indicates confidence interval; NR, not reported.

TABLE 5.

Prevalence of Pancreatic Cancer in Melanoma-prone Families for 10 Pathogenic CDKN2A Variants

No. Families	Variant*	No. Families With a Member Affected With PDAC
21	p.A76fs	>60%
19	p.M53I	<11%
16	p.G101W	35% to 50%
11	p.R112_L113insR	>60%
11	c.−34G>T	15% to 25%
11	c.IVIS2–105A>G	<11%
9	p.R24P	15% to 25%
7	p.V126D	35% to 50%
6	p.L32P	35% to 50%
6	c.32_33ins9–32	<11%

^*Variant position given in reference to NP_000068.1 for protein change (p.) and NM_000077.4 for transcript change (c.). PDAC – pancreatic adenocarcinoma.

Germline CDKN2A Variants and Risk of Other Cancers

Other than PDAC, pathogenic germline CDKN2A variants are associated with an increased risk of melanoma, gastrointestinal cancers (oral cavity, tongue, pharynx, esophagus, stomach, liver, gall bladder, and colorectal), breast cancer, respiratory cancer, and non-melanoma skin cancer.^{25,26,53–56} The relative risk of developing one of these cancers for an individual in a melanoma-prone family with a pathogenic CDKN2A variant is 64.8 for melanoma, 3.7–17.1 for gastrointestinal cancers, 3.0 – 3.8 for breast cancers, 4.9–15.6 for respiratory cancers, and 3.3–37 for non-melanoma skin cancer when compared to the general population (Table 4).^{25,26,53–56,59} These risks highlight opportunities for early detection of cancers and precancers at these other sites.

Diagnostic Implications

Detection of early, curable, pancreatic cancers and precursor lesions has the potential to reduce mortality.^60,61 Clinical surveillance for pancreatic cancer in high-risk individuals (HRIs) is critical to early detection efforts.^15,61,62 As screening is an intensive and often invasive medical procedure, identifying HRIs who may benefit from intervention is important. The International Cancer of the Pancreas Screening (CAPS) Consortium recommend pancreatic screening for HRIs with an estimated lifetime risk of developing pancreatic cancer of 5% or higher.⁶³ Similarly, the National Comprehensive Cancer Network (NCCN) and ASCO guidelines recommend that first-degree relatives of patients with FPC and/or carriers of pathogenic germline variants in pancreatic cancer susceptibility genes with any family history of pancreatic cancer are considered for clinical surveillance.⁴⁹

Several studies have demonstrated the benefit of clinical surveillance in patients with FPC and pathogenic germline variants in pancreatic cancer susceptibility genes, including CDKN2A. Vasen and colleagues reported a multicenter study of long-term outcome in 411 HRIs including 177 with the p16-Leiden mutation, a 19-base pair deletion of exon 2 of the CDKN2A gene that is a Dutch founder mutation and one carrier of a CDKN2A p.G23R mutation, undergoing screening with magnetic resonance imaging (MRI), magnetic resonance cholangiopancreatography (MRCP), and endoscopic ultrasonography (EUS).^60,64 PDAC was detected in 13 (7.3%) of the individuals with a pathogenic germline CDKN2A variant and, importantly, 9 out of the 13 patients were detected early and underwent surgery. Both the overall resection rate and the 5-year survival rate, at 75% and 24%, respectively, were higher than those previously reported for patients with PDAC.² Others have shown similar detection rates of PDAC or PDAC or a high-grade precursor neoplasm in HRIs.⁶⁵ However, in these studies the number of patients with pathogenic germline CDKN2A variants was small and none developed PDAC or a high-grade precursor lesion. Collectively, these studies highlight the importance of the germline genetic testing and the potential of clinical surveillance to detect pancreatic cancer in carriers of a pathogenic variant. However, further studies are needed to define the clinical utility of screening to carriers of a pathogenic germline CDKN2A variant.

To detect early stage pancreatic cancers, it is important that screening identifies pancreatic abnormalities, such as pancreatic cystic lesions, mild and focal branch pancreatic duct irregularities, sacculation, and main pancreatic duct strictures, which are either associated with pancreatic cancer or with precursor lesions with a high risk of progression to invasive cancer.^66,67 Poley and colleagues reported that screening detected PDAC in 6.8% (3/44) and branch-type intraductal papillary mucinous neoplasia (IPMN) in 15.9% (7/44) of asymptomatic HRIs, including 13 CDKN2A mutation carriers, with one IPMN detected in a carrier of a pathogenic CDKN2A variant.⁶⁸ Similarly, DaVee and colleagues reported that screening detected a pancreatic abnormality in 26.7% (23/86) of HRIs with a pathogenic germline variant, including 37 patients with a family history of PDAC.⁶⁹ In another multicenter prospective cohort study, pancreatic abnormalities were detected in 92 of 216 (42.6%) of HRIs, including 195 relatives of patients with FPC, 19 patients with familial breast-ovarian cancer, and two patients with Peutz-Jeghers syndrome with a family history of PDAC.⁷⁰ However, these studies did not include carriers of a pathogenic CDKN2A variant. The reported prevalence of pancreatic abnormalities in HRIs in these studies is higher than the prevalence of pancreatic cystic lesions is in asymptomatic individuals (8%)⁷¹ and clearly highlight the benefit of clinical surveillance to detect high-risk precancerous lesions and reduce mortality in HRIs, including those with a pathogenic germline variant in a pancreatic cancer susceptibility gene.

In 2019 updated NCCN guidelines changed to strongly recommend that germline testing should be considered for any patient with PDAC using a comprehensive gene panel for hereditary cancer syndromes.⁷² Furthermore, the ASCO provisional clinical opinion recommends that all patients diagnosed with PDAC should undergo an assessment of risk for hereditary syndromes known to be associated with an increased risk for PDAC and that germline genetic testing for cancer susceptibility may be discussed with individuals diagnosed with PDAC, even if family history is unremarkable.⁴⁹ It is expected that based on these guidelines, surveillance of PDAC and its precursors in asymptomatic HIRs will be increasingly performed worldwide and potentially lead to improved prognosis.

Therapeutic Implications

The CDKN2A gene product p16^INK4a inhibits CDK4/6 and regulates the cell cycle.¹⁷ Loss of p16^INK4a function, through pathogenic germline variants combined with a second somatic hit to the wild-type allele, or more commonly, by two somatic alterations of CDKN2A is a common event in pancreatic tumorigenesis and results in increased CDK4/6 activity and cell proliferation. Therefore, inhibition of CDK4/6 is a potential target for anticancer therapy in patients with PDAC. CDK4/6 inhibitors have shown efficacy in patients with hormone receptor (HR)-positive breast cancer, and three inhibitors, palbociclib (PD0332991), ribociclib (LEE011), and abemaciclib (LY2835219), have been approved by the U.S. Food and Drug Administration (FDA) for the treatment of patients with HR-positive, epidermal growth factor receptor 2 (HER2)-negative advanced or metastatic breast cancer in 2015, 2017, and 2018, respectively.^73–75 In 2019, Giuliano and colleagues reported that patients with HR-positive, HER2-negative metastatic breast cancer treated with CDK4/6 inhibitors and hormone therapies such as anastrozole and letrozole have longer progression-free survival compared to patients treated with hormone therapies alone.⁷⁶ Monotherapy with CDK4/6 inhibitor has also shown promising activity in patients with liposarcoma⁷⁷ and CDKN2A-mutated non-small cell lung cancer.⁷⁸ However, no effect was observed in patients with squamous cell lung cancer with cell cycle gene abnormalities⁷⁹ nor was any seen in patients with urothelial cancer.⁸⁰ Therefore, for some types of cancer, combination therapies may be more required for antitumor efficacy.⁸¹

CDK4/6 inhibitors have displayed antiproliferative activity in p16^INK4a-deficient PDAC cell lines and patient-derived xenografts (PDX).^82–85 In 2019, the Targeted Agent and Profiling Utilization Registry (TAPUR) Study (ClinicalTrial.gov ID; NCT02693535), reported a prospective, phase II trial of commercially available targeted agents in patients with advanced cancers that harbor genetic alterations in established drug targets with the aim of identifying antitumor efficacy.⁸⁶ In this study, 12 patients with PDAC and CDKN2A alterations confirmed by targeted next generation sequencing were treated with palbociclib monotherapy, however, there were no objective responses or stable disease at 16 weeks, indicating that palbociclib monotherapy has limited clinical activity in patients with PDAC and loss of CDKN2A. Preclinical studies, however, have reported synergistic effects for combinations of CDK4/6 inhibitors with MEK⁸⁷ and mTOR inhibitors.^87–89 Clinical trials including CDK4/6 inhibitors in combination with other therapies are also being investigated in patients with PDAC (Table 6), including: 1) a phase I trials of palbociclib with the anti-microtubule agent nab-paclitaxel (NCT02501902), cisplatin or carboplatin (NCT02897375), the mTOR inhibitor LY3023414 (NCT03065062), or the ERK inhibitor ulixertinib (NCT03454035), 2) a phase II trial of ribociclib in combination with MEK inhibitor trametinib (NCT02703571), 3) a phase I/II trial of ribociclib in combination with mTOR inhibitor everolimus (NCT02985125), and 4) a phase II trial with abemaciclib in combination with mTOR inhibitor LY3023414 (NCT02981342) (Table 6). Since PDAC is characterized, in >80% of cases, by the inactivation of CDKN2A, these trials have been conducted in patients with advanced pancreatic cancer without considering CDKN2A status.³⁴ If clinical trials of CDK4/6 inhibitors including patients with PDAC unselected for CDKN2A status show promising activity, selection of patients with verified inactivating alterations in CDKN2A may improve efficacy.

TABLE 6.

Clinical Trials of CDK4/6 Inhibitors in Patients With PDAC

CDK4/6 Inhibitor	Combination Drug	NCT Identifier	Phase	CDKN2A Status Considered	Study Title
Palbociclib	Nab-paclitaxel	NCT02501902	I	No	Dose-Escalation Study of Palbociclib + Nab-Paclitaxel in mPDAC
Palbociclib	Cisplatin or carboplatin	NCT02897375	I	No	Palbociclib With Cisplatin or Carboplatin in Advanced Solid Tumors
Palbociclib	LY3023414 (mTOR inhibitor)	NCT03065062	I	No	Study of the CDK4/6 Inhibitor Palbociclib (PD-0332991) in Combination With the PI3K/mTOR Inhibitor Gedatolisib (PF-05212384) for Patients With Advanced Squamous Cell Lung, Pancreatic, Head & Neck and Other Solid Tumors
Palbociclib	Ulixertinib (ERK inhibitor)	NCT03454035	I	No	Ulixertinib/Palbociclib in Patients With Advanced Pancreatic and Other Solid Tumors
Ribociclib	Trametinib (MEK inhibitor)	NCT02703571	II	No	Study of Safety and Efficacy of Ribociclib and Trametinib in Patients With Metastatic or Advanced Solid Tumors
Ribociclib	Everolimus (mTOR inhibitor)	NCT02985125	I/II	No	LEE011 Plus Everolimus in Patients With Metastatic Pancreatic Adenocarcinoma Refractory to Chemotherapy
Abemaciclib	LY3023414 (mTOR inhibitor)	NCT02981342	II	No	A Study of Abemaciclib (LY2835219) Alone or in Combination With Other Agents in Participants With Previously Treated Pancreatic Ductal Adenocarcinoma

Conclusion

Pathogenic germline CDKN2A variants are one of the most frequently inherited variants associated with an increased risk of pancreatic cancer. Identification of patients with a pathogenic CDKN2A variant is important for screening of at-risk relatives for pancreatic cancer and melanoma. Furthermore, patients with an inherited pathogenic CDKN2A variant and somatic loss of CDKN2A may benefit in future from treatment with targeted therapies such as a CDK4/6 inhibitor.