Energy balance and the AKT pathway are important in
colorectal cancer development and regulate p27 (
cyclin-dependent kinase inhibitor-1B/CDKN1B/KIP1), which plays a role in preventing cell cycle progression. However, little is known on the clinical outcome or prognostic significance of p27 alterations in relation to patient body mass index (BMI). Among 630
colon cancers (stage I-IV) in two prospective cohort studies, we detected p27 alterations (cytoplasmic p27 localization or p27 loss) in 500
tumors (79%) by immunohistochemistry. The remaining 130 (21%)
tumors were "p27-nuclear+." Cox proportional hazard models computed hazard ratios (HR) of deaths, adjusted for patient and tumoral characteristics, including p53, p21,
cyclin D1, KRAS, BRAF, PIK3CA,
cyclooxygenase-2,
fatty acid synthase (FASN),
beta-catenin,
microsatellite instability (MSI), CpG island methylator phenotype (CIMP), and long interspersed
nucleotide element-1 (LINE-1) hypomethylation. Compared with p27-nuclear+ patients, p27-altered patients experienced low
colon cancer-specific [adjusted HR, 0.63; 95% confidence interval (95% CI), 0.42-0.94] and overall mortality (adjusted HR, 0.70; 95% CI, 0.51-0.95), independent of FASN, MSI, CIMP, LINE-1 methylation, and other potential confounders. The effect of p27 alteration on overall mortality significantly differed by BMI (P(interaction) = 0.013); adjusted HR (p27-altered versus p27-nuclear+
tumors) was 0.28 (95% CI, 0.13-0.59) for BMI >or=30 kg/m(2), 0.67 (95% CI, 0.40-1.14) for BMI 25 to 29 kg/m(2), and 0.91 (95% CI, 0.57-1.46) for BMI <25 kg/m(2).
Obesity was associated with inferior overall survival among p27-nuclear+ cases (adjusted HR, 3.07; 95% CI, 1.49-6.32; versus nonobese cases), but not among p27-altered cases (adjusted HR, 1.08). In conclusion, p27 alterations in
colon cancer are associated with superior prognosis. Adverse prognostic effect of
obesity seems limited to patients with nuclear p27 expression, suggesting a host-
tumor interaction.