CDKN2A: Two Tumor Suppressors from One Locus

Alternative Reading Frames: One Gene, Two Proteins

The CDKN2A locus encodes p16INK4a and p14ARF through a remarkable genetic arrangement: both proteins share exons 2 and 3 but use different first exons (1α for p16INK4a, 1β for p14ARF) and are translated in different reading frames. The two proteins are therefore completely unrelated in amino acid sequence — p16INK4a is a CDK inhibitor with four ankyrin repeats, while p14ARF lacks enzymatic activity but has potent protein-binding properties. This overlapping reading frame configuration is not unique to CDKN2A but is rare in the mammalian genome, suggesting strong evolutionary pressure to co-localise these two tumour suppressor functions.

Despite sharing no protein sequence, p16INK4a and p14ARF converge on a single cellular outcome through different mechanisms: G1 cell cycle arrest. p16INK4a achieves this by inhibiting CDK4/CDK6, preventing pRb phosphorylation and E2F release. p14ARF achieves this by stabilising p53 through MDM2 sequestration, driving p53-mediated p21 transcription and CDK2 inhibition. A deletion removing both reading frames simultaneously eliminates both G1 arrest mechanisms and both tumour suppressor pathways from a single genomic event.

Epigenetic Silencing by Promoter Methylation

CDKN2A deletion is not the only mechanism of inactivation. The p16INK4a promoter contains a dense CpG island that is one of the most commonly hypermethylated regions in human cancer, with methylation-mediated silencing detectable in ~30% of sporadic cancers across many tissue types. p16INK4a promoter methylation is an early event in carcinogenesis — detectable in premalignant lesions including oral leukoplakia, Barrett's oesophagus, and colonic adenomas — making it a potential early cancer detection biomarker in liquid biopsies.

Importantly, epigenetic silencing is reversible in principle: DNMT inhibitors (azacitidine, decitabine) can restore p16INK4a expression in methylation-silenced cancer cells in vitro, though clinical efficacy specifically through p16 restoration is difficult to isolate from global demethylation effects. The p14ARF promoter is less commonly methylated than p16INK4a in sporadic cancers, meaning isolated p16INK4a silencing can occur without simultaneous ARF inactivation.

Cancer-Type Distribution and Deletion Mechanisms

CDKN2A is recurrently altered across many cancers, including melanoma, pancreatic ductal adenocarcinoma, glioma and squamous carcinomas. The prevalence and alteration class differ substantially by tumour type and cohort. In pancreatic cancer, CDKN2A disruption is one of several recurrent driver events alongside KRAS, TP53 and SMAD4 alterations; these events are common but are not individually required in every tumour.

Deletion, promoter methylation and sequence variants answer different questions. A homozygous 9p21 deletion can remove both CDKN2A alleles and may extend into neighbouring genes such as MTAP, whereas smaller or transcript-specific events may not affect the same products. MTAP loss has motivated PRMT5- and MAT2A-directed research, but the dependency and clinical evidence are agent- and setting-specific. Copy number, methylation and sequence assays should therefore be interpreted separately.

Key Takeaways

• ·CDKN2A encodes two completely unrelated tumour suppressor proteins from overlapping reading frames: p16INK4a (four-ankyrin-repeat CDK4/6 inhibitor) and p14ARF (MDM2-sequestering nucleolar protein), sharing exons 2 and 3 but different first exons and reading frames.
• ·A broad CDKN2A deletion can affect both p16INK4a–RB and p14ARF–MDM2–p53 control, but the deletion boundaries and transcript consequences must be established.
• ·p16INK4a promoter hypermethylation is among the earliest detectable epigenetic changes in carcinogenesis, serving as a potential liquid biopsy biomarker for cancer early detection in at-risk populations.
• ·CDKN2A deletion frequently co-deletes MTAP at the 9p21.3 locus, creating a PRMT5 inhibitor synthetic lethality being exploited in clinical trials for MTAP-null cancers.
• ·CDKN2A, KRAS, TP53 and SMAD4 are recurrently altered in pancreatic ductal adenocarcinoma, but frequency varies and no fixed four-event combination is required in every tumour.