delGA (rs67491583) Variant and Colorectal Cancer Risk in an Indigenous African Population

Abstract

Objective

A recent study showed a higher frequency of GA deletion at rs67491583 in African American colorectal cancer (CRC) patients compared to controls, suggesting a likely contribution of this allele to racial disparity in CRC risk predisposition. We conducted a pilot study in an indigenous African population to evaluate this potential CRC risk variant.

Methods

We collected epidemiological data and biological specimen from consenting consecutive CRC cases and controls presenting at the Oncology Clinic of University College Hospital, Ibadan from 2001 to 2007. We examined germline DNA for delGA by PCR-amplification of two overlapping fragments using standard primers. The products were directly sequenced using Applied Biosystems BigDye v3.1 sequencing chemistry and AB13730 automatic DNA sequencer.

Results

There were 45 cases and 45 controls of which genotyping was successful in 39 cases and 38 controls. There were 5 heterozygous and 2 homozygous GA deletions with frequency of 11.54% (9/78) among cases whereas there were 8 heterozygous and 1homozygous GA deletions among controls with frequency of 13.15% (10/76). (p= 0.79, OR 0.88, 95% CI 0.34-2.28)

Conclusion

This study suggests that there is no association between the delGA (rs67491583) variant and CRC risk in this indigenous African population. However our sample size was small and the participants were not ethnically homogenous. Further studies are required to evaluate this marker in African CRC.

Introduction

About 35% of colorectal cancers (CRC) have been attributed to inherited predisposition. Of these, familial adenomatous polyposis (FAP) and hereditary non-polyposis colorectal cancer (HNPCC) syndromes constitute less than 10%. These are due to high-penetrance mutated genes such as APC, MLH1, MSH2, MSH6, PMS2 and MYH. APC, a tumour suppressor gene located on the long arm of chromosome 5 (5q21), is responsible for degradation of intracellular β-catenin. β-catenin is a positive regulator of the Wnt pathway which contributes to the survival of cancer stem cells. APC mutated colorectal cancer develops from chromosomal instability which eventually ends in the adenoma-carcinoma sequence [1, 2]. The DNA mismatch repair genes ( MLH1, MSH2, MSH6, etc.) proofread and repair defects that are overlooked or missed by DNA polymerase during cell replication. Mutations in these genes leave the genome with microsatellites that are shorter or longer than the parent cells. This phenomenon, called microsatellite instability is the hallmark defect in HNPCC, also called Lynch syndrome [1, 3].

A large proportion of the familial cancers are probably due to gene mutations and polymorphisms of low penetrance which are yet to be fully elucidated [3–6].The past two decades have witnessed an explosion in knowledge of molecular biology and discovery of novel genomics technologies. Scanning the whole human genome has now been made possible following the complete sequencing of the human genome and additional information has been provided by the International HapMap projects [7, 8, 9]. These have led to the development of relevant high-throughput genomic technology, which has further enhanced the identification of associations that could explain the remaining low-penetrance risk factors for CRC [10]. Genome wide association studies (GWAS) in Caucasians have identified many common genetic variants that increase susceptibility to CRC risk. The first published loci included, among others, rs6893267 at 8q24, rs4779584 at 15q13, rs4939827 at 18q21 and rs16892766 at 8q23.3 [11, 12]. Of these, the single nucleotide polymorphism (SNP) rs6983267, located at region 8q24.21, has been studied most extensively. The 8q24 region has also been associated with other cancers including cancers of the breast and prostate [13, 14]. The SNP rs6983267 has been shown to function by changing the activity of a regulatory element upstream of the MYC oncogene (MYC-335) [15, 16].

In the past, colorectal cancer was considered a very rare disease in Africans compared to Caucasians. However, recent reports and reviews from across Nigeria have documented an increasing number of patients with CRC presenting for diagnosis and treatment [17–21].A pilot report on genetic risk for colon cancer in Nigeria over a decade ago documented the presence of microsatellite instability phenotype and germ line mutations in MSH2 in the tumours of two patients among five unselected patients with CRC [22]. That report, the first of its type from Nigeria, opened up the search for genetic factors underpinning the pathogenesis of CRC in Nigeria.

Recently, Tuupanen et al performed a follow-up study for the 8q24 CRC enhancer work and characterized in detail the genetic variation withinMYC-335 using DNA samples of people from Europe, Africa and Asia [23]. They identified a 2-base pair (bp) deletion, rs67491583, located 927 bp downstream of rs6983267 that affected a Growth Factor Independent (GFI) binding site and which was present only in people of African descent. To elucidate a likely contribution of this allele to racial disparity in CRC risk predisposition, they further genotyped 1027 African American CRC cases and 1773 healthy controls for this deletion. Their results showed that the frequency of GA deletion was higher in cases (9.9%) than in controls (8.6%). Though their finding did not reach statistical significance, it demonstrated that there was a trend of association of this variant with CRC predisposition. As a follow up to this novel discovery, blood samples from CRC cases and controls from Ibadan were analysed for this potential colon cancer risk variant and the results are hereby presented.

Materials and Methods

Study participants

This was a pilot study of the genetic epidemiology of colorectal cancer in the Division of Surgical Oncology of the University College Hospital (UCH) Ibadan, Nigeria. Patients with histologically diagnosed CRC who provided consent to participate were recruited to the study at presentation at the surgical outpatients department of the hospital. Questionnaires were used to elicit epidemiological information from the patients. Blood was then drawn from the patients for DNA extraction at the surgical research laboratory. For this analysis, germ line DNA samples from 45 cases and 45 controls were available for genotyping. Controls were recruited from the general population and were individuals who were free of any known cancer.

Genotyping

Genomic sequencing and genotyping of cases and controls were done in the Department of Medical Genetics, University of Helsinki, Finland by Tuupanen et al and the full procedure has been described in a previous publication [17]. The samples were genotyped for delGA with direct sequencing. The samples were PCR-amplified using the following primers: F: 5’-ATTCCTGACCTACCCCCAAA-3’ , R: 5’-GTTTTCAGGTGCGTGTGTGT-3’ and the products were sequenced directly using Applied Biosystems BigDye v3.1 sequencing chemistry and AB13730 automatic DNA sequencer. Sequence reads were analysed using Mutation Surveyor v3.24 software (Softgenetics).

Association analysis

Association analysis was performed using SISA (Simple Interactive Statistical Analysis; http://www.quantitativeskills.com/sisa/). Pearson’s chi square test was used to calculate P-value, odds ratio (OR) and the associated confidence interval (CI).

Results

Table 1 contains selected characteristics of the cases and controls. The male to female ratio was 3:2 in cases and 1:4 in controls. The mean age was 46.5 years and 48.9 years in cases and controls respectively. Most, 71.1% of the cases and 88.9% of the controls were from the Yoruba ethnic group in southwest Nigeria. Moreover, most cases (40.0%) and controls (35.6%) completed tertiary school education. The mean weight and BMI were 51.9kg and 17.2kg/m² in cases and 62.9kg and 24.7kg/m² in controls, respectively.

Table 1. Selected Characteristics of Colorectal Cancer Cases and Controls in Ibadan, Nigeria.

Characteristics	Cases (n=45)	Controls (n=45)
Age in years
Minimum	22	20
Maximum	74	79
Mean (SD)	46.5 (14.6)	48.9 (15.5)
Sex, n (%)
Male	28 (62.2)	9 (20.0)
Female	17 (37.8)	36 (80.0)
Ethnicity, n (%)
Yoruba	32 (71.1)	40 (88.9)
Ibo	5 (11.1)	4 (8.9)
Hausa	2 (4.5)	0 (0)
Others	6 (13.3)	1 (2.2)
Education
No formal	6 (13.3)	6 (13.3)
Elementary	5 (11.1)	9 (20.0)
Secondary	12 (26.7)	12 (26.7)
Tertiary	18 (40.0)	16 (35.6)
Others	4 (8.9)	2 (4.4)
Height (m), mean (SD)	1.4 (0.6)	1.6 (0.7)
Weight (kg), mean (SD)	51.9 (20.9)	62.9 (13.7)
BMI (kg/m2), mean (SD)	17.2 (9.4)	24.7 (5.9)

SD, Standard Deviation

Samples from all the cases and controls were analysed for a potential colon cancer risk variant rs67491583.Altogether, 39 cases and 38 controls were successfully genotyped. PCR amplification was not achieved in 6 cases and 7 controls. Among the 39 CRC cases, 7 patients had a GA deletion (5heterozygotes and 2 homozygotes), producing an allele frequency of 11.54% in the cases. In the control group, there were 8 heterozygote and 1homozygote deletions and the allele frequency of the deletion was 13.15% (Table 2). These numbers indicate that the GA deletion is not associated with increased risk of CRC in Nigerian samples p=0.79 (OR= 0.88, 95% CI: 0.34-2.28)

Table 2. Frequency of rs67491583 in Ibadan CRC cases and controls.

	Total	Genotype						OR (95% CI)
		Wt	delGA/-	-/-	AF	%	P
CRC	39	32	5	2	9/78	11.54	0.79	0.88 (0.34-2.28)
Male	25	20	3	2
Female	14	12	2	0
Controls	38	29	8	1	10/76	13.15
Male	6	4	2	0
Female	32	25	6	1

CRC, Colorectal cancer; AF, Allele Frequency; P, p value; OR, Odds ratio; CI, Confidence Interval

Discussion

In this pilot study, germ line DNA samples from colorectal patients and controls from Ibadan, Nigeria were analysed for rs6749153 as a potential colon cancer risk variant in an indigenous African population. Significant evidence has accumulated in support of the contribution of rs6983267 to CRC risk mainly in peoples of European ancestry. Such evidence of an association is lacking for CRC predisposition in populations of African descent. In fact, reports have suggested that other variants at 8q24, and not rs6983267, might contribute to CRC in African Americans [24, 25]. The variant, rs67491583, had been suspected to play a role in CRC predisposition in African Americans. It has also been shown to be another functional variant, in addition to rs6983267, in the CRC associated enhancer element in 8q24. However, compelling evidence for an association has not been obtained, in part due to the rarity of the variant, as well as the genetic heterogeneity in the African American population.

Although most of the samples in this study were successfully genotyped, the result did not support our hypothesis, as the variant was more common in controls than in cases. The CRC risk allele G at rs6983267 and the GA deletion at rs67491583 enhance the activity of the MYC-335 enhancer, leading to increased expression of MYC. Interestingly, the HapMap data show a higher frequency of G allele of rs6983267 compared to other ethnic groups that were involved in the HapMap project. Similarly, the GA deletion allele of rs67491583, which is found in African Americans, is rare in Caucasian populations.

This phenomenon fuels the speculation that these two functional alleles might be controlled by environmental and dietary factors, as their frequency reduces when moving away from Africa. They might be important in tolerance to gastrointestinal infections in Africa, whereas in countries with western lifestyle they contribute to the predisposition to colorectal neoplasia. However, larger studies must be conducted to study this interesting hypothesis.

This small study has not demonstrated an association between GA deletion of rs67491583 allele and CRC risk in Africans. The study was beset by some limitations. The unexpected result might just be due to the small sample size, thus reliable conclusions to the research question cannot be drawn. Moreover, the previous studies that generated the hypothesis for this work were in African Americans. The different origin might also play a role, as we know that heterogeneity exists among African Americans. Nevertheless, both the previous works and this present effort in peoples of African descent bring up an intriguing postulate about racial disparity in CRC predisposition that requires further elucidation in larger studies.