Sporadic early-onset colon cancer expresses unique molecular features

Abstract

Background

The overall incidence of colon cancer (CC) has steadily declined in the last decades but has increased in patients under age 50 y. The etiology of early-onset (EO) CC is not understood. The aim of this study was to elucidate gene expression patterns in EOCC and show its uniqueness compared to late-onset (LO) disease.

Methods

Two cohorts of patients with sporadic CC were identified. Tumors and matching noninvolved tissues from six EOCC patients (<50) and six late-onset colon cancers (LOCC) patients (>65) were obtained from pathology archives. De-paraffinized tissues were macrodissected from FFPE sections, RNA isolated, and used for expression profiling of 770 cancer-related genes representing 13 canonical pathways.

Results

Among 770 genes assayed, changes in expression levels of 93 genes were statistically significant between EOCC and matching noninvolved tissues. There were also significant differences in expression levels of 118 genes between LOCC and matching noninvolved tissues. Detailed comparative gene expression analysis between EOCC and LOCC normalized to their matching noninvolved tissues revealed that changes in expression of 88 genes were unique to EOCC using the cutoff criteria of expression levels difference >2 fold and P value <0.01. From these differentially expressed genes specific to EOCC, 28 genes were upregulated and 60 genes downregulated. At the pathway level, RAS, MAPK, WNT, and DNARepair pathways were similarly deregulated in both age groups, whereas PI3K-AKT signaling was more specific to EOCC and cell cycle pathway to LOCC.

Conclusions

These results suggest that sporadic EOCC is characterized by distinct molecular events compared to LOCC.

Introduction

Colon cancer (CC) is an important contributor to worldwide cancer mortality and morbidity as there are about 1.2 million new cases diagnosed and 0.6 million deaths from this disease reported every year around the world.¹ In the United States, 45,890 new colon cancer cases are expected to be diagnosed in men and 47,200 in women in 2015.² Although the overall incidence and mortality of colon cancer have declined over the past 30 years in both men and women,³ the number of early-onset colon cancer (EOCC), in individuals younger than 50 years, continued to rise alarmingly.⁴

Using Surveillance Epidemiology and End Results Program Cancer Registries database, we demonstrated in our previous clinical study that incidence rates of CC are continuously increasing in every age group (5-y intervals) from 20 to 49 y, with the most impressive increase seen within the age group 40-44 y. As expected, incidence rates of CC decreased in all age groups from 55 to 85 y.⁵ We also used Arizona Cancer Registry database to analyze the incidence rates of EOCRC in Arizona.⁶ This study revealed that there is a significant increase (102%) in incidence of CC seen in the age group 10-29.⁶

Overall declines in late-onset colon cancers (LOCC) in patients older than 50 years have been mostly attributed to changes in risk factors and the introduction of screening primarily in the form of colonoscopy.⁷ However, evidence accumulating steadily over the last decade now threatens this optimistic expectation.⁸ Trends toward poorer survival have been seen in EOCC patients.⁹ Most (70%-80%) of these EOCC are sporadic and not attributed to any hereditary cause.4, 10, 11, 12 They tend to have more aggressive features like mucinous, poorly differentiated histopathology, and signet ring morphology, and are often diagnosed at a more advanced stage.4, 5, 10, 13 The most common location of early-onset colon cancer is observed in sigmoid colon and recto-sigmoid colon.5, 14

Colon cancer is a highly heterogeneous disease with clinically and pathologically similar tumors differing in molecular signatures and in response to therapy and patient survival.¹⁵

Only a few pilot genetic studies focused on EOCC have been performed up to date despite the fact that it is a growing problem, still remains poorly understood, and the etiology is unclear.4, 10, 11, 12 It is believed sporadic EOCC pathogenesis involves accumulation of various genetic and/or epigenetic modifications that regulate proliferation and apoptosis in addition to telomere maintenance and genomic integrity. Cell structure, cell-cell interactions, and polarity, angiogenesis and changes in the tumor microenvironment also play an important role in EOCC pathogenesis.¹⁶

To date, only a few studies have tried to uncover critical genes and pathways important in the initiation and progression of CRC, especially looking at WNT, RAS-MAPK, PI3K, TGFβ, P53, and DNA mismatch-repair pathways.¹⁷ It is also known that most EOCRC tumors do not seem to harbor activating BRAF or KRAS mutations,¹³ whereas a genome-wide LINE-1 hypo-methylation seems to be a frequent epigenetic alteration found in EOCRC tumors.14, 18

The biggest limitation of current genomic studies focusing on EOCRC is that they all are designed to use downstream genetic applications such as PCR-based techniques and microarrays requiring high-quality RNA, thus limited to analysis of fresh frozen samples only. As there is insufficient amount of fresh frozen tissues especially for the EOCRC cases, these studies use only limited number of samples to assess gene expression profiles. There are, however, more samples of EOCRC tissues available in FFPE (formalin fixed, paraffin embedded) format. Thus, there is a need to find novel methods to analyze these samples. To overcome the limitation of low-quality RNA usually isolated from FFPE samples, methods which do not include any reverse transcription and amplification steps need to be tested. The NanoString nCounter platform is a novel method suitable for gene expression analysis of samples archived as FFPE blocks or as tissue sections preserved on slides, thus might potentially allow to analyze larger cohorts of EOCRC samples. The aim of our study was to attempt to elucidate the molecular uniqueness of sporadic EOCC using a more advanced platform that uses archived FFPE samples.