Genetic and epigenetic aberrations are responsible for the development of colorectal cancer (CRC) that are inherited or gained during life, or both [1]. A number of risk factors for CRC have been determined in epidemiologic studies which include age, obesity, race, diet, cigarette smoking, inflammatory bowel disease, family history of colon cancer, intake of alcohol and physical inactivity [2]. In colorectal epithelial cells, these risk factors cause epigenetic and genetic changes, and together with the inherited genetic makeup, may result in the development of a tumour [1,3,4]. In colorectal cancer there are three pathways of genomic instability which include:(i) DNA microsatellite instability (MSI) phenotype due to mutations in DNA mismatch repair genes, (ii) chromosomal instability (CIN) phenotype due to mutations in APC and other genes that activate Wnt pathway, and (iii) CpG island methylator phenotype (CIMP) due to global genome hypermethylation, resulting in switch off of tumour suppressor genes [5].

In Uganda microsatellite instability is responsible for 41% of all CRC cases in Uganda, and in the West is responsible for 15-20% of all CRC cases [6,7]. Mismatch repair genes, play a major role in repairing DNA replication errors and hence promulgating genetic stability, inhibiting recombination between non-identical DNA sequences and interfering in responses to DNA damage. Hence, inactivation of mismatch repair genes results in MSI tumorigenesis. Promotor hypermethylation of the MLH1 mismatch repair gene MLH1 results in inactivation of the gene resulting in sporadic MSI tumours [8]. Hereditary nonpolyposis CRC (HNPCC or Lynch syndrome) is the familial form of MSI CRC, which accounts for 3-5% of all CRC cases in the West and is caused by germline mutations in mismatch repair genes MLH1, MSH6, PMS2 and MSH2 [9,10].

In contrast to colorectal tumours in Uganda, in the West, MSI tumousr are more often located in the proximal colon, poorly differentiated and of mucinous or signet ring histopathological subtype (Wismayer R, 2024). MSI CRC have an abundance of tumour-infiltrating T-cells and are associated with a better patient prognosis compared with MSS tumours [11,12].

An important mechanism in human carcinogenesis is the transcriptional inactivation of cystosine methylation at promotor CpG islands of tumour suppressor genes [13]. In colorectal cancer, a number of tumour suppressor genes, which include MGMT, CDKN2A, (p16 gene), MLH1, among others, are silenced by promotor methylation [14-16]. A characteristic feature for the serrated pathway of colorectal tumorignesis is the CpG island methylator phenotype (CIMP) [17]. Promotor methylation in multiple genes, referred to as CIMP is exhibited by a subset of colorectal cancers [18-21,29].

In the West, CIMP colorectal tumours have been found to have a distinct profile which includes lower TP53 mutation rates, higher BRAF and a proximal tumour location [20,21,22-26]. Some studies have observed higher KRAS mutation rates in CIMP tumours [18,22], while other studies showed lower KRAS mutation frequencies in CIMP tumours [20,23,24]. A lack of uniform criteria for CIMP, methylation detection methods and the panel of markers used for CIMP, emphasizes the importance of carefully evaluated criteria for CIMP and methylation detection methods.

Controversies exist as to whether CIMP is a distinct biological subtype of CRC, in contrast to the high degree of microsatellite instability (MSI-H) which is a distinct subtype of CRC [27]. Clinicopathological features have been found to be associated with CIMP in colon cancer [18,20-22,25,26]. However, previous studies have not been able to demonstrate a bimodal distribution of the number of methylated loci. In some studies, the case numbers were small and most studies were not population based. There are a few large prospective cohort studies which have been performed to characterise CIMP in CRC.

Due to the qualitative nature of the assay, methylation specific polymerase chain reaction (MSP) cannot really distinguish low levels of methylation from high levels of methylation. It has previously been shown that most colorectal cancers with low levels of methylation in CDKN2A (p16), MLH1 or MGMT do not silence corresponding gene expression by quantitative DNA methylation analysis [28]. With low levels of methylation in CRC, methylation specific polymerase chain reaction (MSP) may give positive results with little significance biologically. The frequency of CIMP in colorectal cancer, may be overestimated based on results by MSP and therefore any clinicopathological associations with CIMP may be obscured by misclassified tumours. Therefore, in cancer research, quantitative measurements of DNA methylation are important [29]. Real time polymerase chain reaction (PCR) and a variety of quantitative assays to measure DNA methylation in tumour tissue have been developed [30-34]. The aim of this study was to determine the genetic profile of CIMP colorectal cancer in Ugandan patients.

This was a cross-sectional study design. Consecutive participants were recruited prospectively between the 16^thSeptember 2019 to the 16^thSeptember 2021 from the Surgical Departments of Masaka Regional Referral Hospital, Mulago National Referral Hospital, Uganda Martyrs’ Hospital Lubaga, and Mengo Hospital. Retrospective CRC FFPE tissue blocks were obtained from participants diagnosed with CRC from the 1^stJanuary 2008 to the 15^thSeptember 2019, from the archives of the Department of Pathology, School of Biomedical Sciences, College of Health Sciences, Makerere University. The histopathologic diagnosis was confirmed as colorectal adenocarcinoma by a consultant pathologist at the Department of Pathology, School of Biomedical Sciences, College of Health Sciences, Makerere University and by another consultant pathologist at the Institute of Genetics and Cancer at the University of Edinburgh.

The inclusion criteria included cases with a histological diagnosis of colorectal adenocarcinoma. Cases with recurrent colorectal cancer and poor-quality CRC FFPE tissue block samples were excluded. Poor quality CRC FFPE tissue samples included those having a low concentration of extracted DNA, poor quality DNA, or insufficient tissue for the extraction of DNA. All cases meeting the selection criteria were included, during the study period.

The radiology reports from the participants’ medical case files were retrieved to obtain the stage of CRC. The stage of CRC was defined according to the AJCC 8^thedition (AJCC 8th edition). The medical case files were used to abstract data for stage, topography of tumour, and demographics of the participants. The histopathological subtype, grade and lymphovascular status (LVI) of CRC were obtained on haemaotoxylin and eosin staining of CRC tissue. The grade of CRC was obtained using the three-tier grading system as follows: G1: well-differentiated CRC with >95% glandular formation; G2: moderately differentiated CRC with 50-95% glandular formation; and G3: poorly-differentiated CRC with <50% glandular formation [35].

FFPE Tissue Block Sampling

Out of the 404 FFPE colorectal adenocarcinoma tissue blocks used for histopathological subtype analysis, 277 FFPE tissue blocks were excluded due to poor quality DNA extracted from these FFPE tissue blocks (Figure 1). Poor quality DNA samples included those with a Phred score of less than 20. The Phred score was used to indicate the measure of base quality in DNA sequencing. This resulted in 127 FFPE tissue blocks which had adequate quality and quantity of DNA extracted for MSI and MSS molecular subtype analysis (Figure 1). Due to inadequate quantity of DNA left for CIMP analysis a further 35 DNA samples were excluded resulting in 92 DNA samples analysed for CIMP analysis (Figure 1).

Genomic DNA Extraction

The protocol used for DNA extraction followed the Roche FFPE extraction kit protocol. During incubation with proteinase K and tissue lysis buffer, the FFPE tissue samples were homogenized and disrupted. In the presence of a chaotropic salt, the nucleic acids bind to the surface of glass fibres in the high pure purification filter tube. A series of rapid “wash-and-spin” steps to remove the cellular components, proteins and salts was carried out to purify the DNA. The DNA was then released from the glass fibre using a low-salt elution. The DNA samples passed quality control with a phred score of 20. The DNA samples were extracted at the Institute of Genetics and Cancer at the University of Edinburgh.

Figure 1: Flow diagram of FFP tissue blocks of participants selected for MSI and CIMP analysis.

Next Generation Sequencing For MLH-1, MSH-2, MSH-6, BRAF and KRAS

DNA NGS (next generation sequencing) was used to interrogate genomic variants which were present in the MMR genes (MLH-1, MSH-2 and MSH-6), BRAF V600E and KRAS [36,37]. The next generation sequencing was done at the Institute of Cancer and Genomic Sciences at the University of Birmingham. All exons, flanking and intronic sequences, structured rearrangements (duplications, delections, insertions and translocations), single nucleotide variants (SNVs) and copy number variants (CNVs) were included in MMR sequencing, BRAF and KRAS sequencing.

The MSI-NGS method was used to determine mutations in MLH-1, MSH-2 and MSH-6 and immunohistochemistry was used to determine expression of MMR genes, MLH-1 and PMS2. The colorectal cancer cases used in the study were unlinked and de-identified to patient information.

Genetic sequencing: UW-GSTL personnel performed clinically validated ColoSeq^TM tumour assay [38]. DNA was extracted from ten unstained (10µm thickness) FFPE slides from 127 samples, to perform targeted NGS. A board-certified anatomic and molecular pathologist assessed a matching H&E slide for adequate CRC tumour content. As previously described DNA was extracted and the DNA quality was assessed [39]. The sample DNA quality control had been checked using the Qubit picogreen DNA assay.

ColoSeq^TM tumour testing after DNA extraction and NGS library preparation was performed as had been previously described by Haraldsdottir S et al [38,40]. Briefly, using Hyperprep (Kapa Biosystems, Wilmington, MA, USA), the DNA was sonicated to obtain fragments of appropriate size for library preparation. Using Illumina chemistries which utilize MiSeqv2 500 (Illumina, San Diego, CA, USA), sequence adaptors and barcodes were ligated to captured DNA fragments to perform sequencing-by-synthesis. The panel used a single primer extension and the co-ordinates associated were respective to the 3’ or 5’ location. Using standard bioinformatics tools, data was multiplexed post-sequencing. Using a clinically validated custom bioinformatics pipeline, structural variant determination, alignment, variant calling, copy-number analysis and MSI were carried out [36,37,41]. Using CONTRA and PICARD, average sequencing coverage was estimated with coverage of less than 100x considered suboptimal. Data was considered to have failed assessment based on overall sequencing quality, loci-specific sequence depth and a manual review of the selected sequences.

The variant allele fraction (VAF) of neoplasm-associated variants and loss of heterozygozity assessment by haplotype analysis of mutations in the cancer was used to estimate the neoplastic content. Each colorectal cancer sample was assessed for the presence of mutations in BRAF, KRAS, MLH-1, MSH-2 and MSH-6. The greater impact on protein expression was found in variants with multiple MMR gene alterations.

Variants that were likely to affect protein expression were defined as pathogenic mutations (eg: translocation mutations, frameshift or stop gain). Colorectal cases without identified MSH-2 and MSH-6 MMR mutations were denoted as “No variant pathologically identified” and hence were likely to have MMR protein expression.

For all the colorectal cancer samples the MSI status was denoted as either MSI-positive or MSI- negative (MSI-stable (MSS)) using Next Generation Sequencing method for MLH-1, MSH-2 and MSH-6 [36]. The sequencing of genes was carried out at the Institute of Cancer and Genomic Science at the University of Birmingham.

Immunohistochemistry for MLH-1 and PMS2Expression

MMR immunohistochemistry was performed on all the 127colorectal cancer samples, utilizing the Novolink Max polymerdetection system Kit and Tris buffered saline (Thermo Scientific)according to the instructions of the manufacturer. The primarymonoclonal antibodies used were against MLH-1 (clone ES05;Ref: IR079) in a dilution of 1:100 and PMS-2 (clone EP51; Ref:IR087) in a dilution of 1:100 (DAKO, Agilent, USA) according tothe manufacturer kit instructions.These monoclonal antibodies were applied to 3-μm deparaffinizedformalin-fixed paraffin embedded (FFPE) tissue slide sections.Blockage of the endogenous peroxidase activity was carried out with the Novolink peroxidase-blocking reagent (Novolinkmaxpolymer) and antigen retrieval performed at 125 degreescentigrade for 36 seconds at a pH9. The antigen-antibody reactionwas visualized with Novolink DAB (diaminobenzidine) solution.Haematoxylin was subsequently used as a counterstain.Each tissue slide sample was screened by two consultantpathologists for the presence or absence of expression of each ofthe two MMR proteins. The two deficient MMR proteins assessedwere MLH-1 and PMS-2 and complete absence of nuclear stainingof the tumour cells was defined as negative protein expression.Deficient MMR (dMMR) was defined if the MMR protein wasnegatively expressed whilst proficient MMR (pMMR) if the MMRprotein was positively expressed in the colorectal cancer tissueslides. Deficient MMR MLH-1 and PMS-2 were subjectively givena staining intensity of 0(-). Proficient MMR MLH-1 and PMS-2were subjectively given a staining intensityof1(+).

Testing for BRAF and KRAS

Tumour DNA were tested for the p.V600E BRAF mutation using a fluorescent allele-specific PCR assay. This mutation accounts for approximately 90% of BRAF mutations in colorectal carcinoma in Western literature. Mutations in Kras codons 12 and 13 were identified through forward and reverse sequencing of amplified tumour DNA. Mutations in the hotspot region account for approximately 80% of Kras mutations in colorectal carcinoma.

Analysis of Sequencing Results for MLH1, MSH2, MSH6, BRAF and KRAS

The quality of the raw fastq files was first assessed using FastQC and MultiQC software (Ewels, 2016) which generated html quality reports. Any bases below quality of 25, as well as adapter sequences were trimmed of using Trimgalore tool [41]. The processed reads were aligned to the human genome reference version38 (hg38) using BWA-MEM (Burrows Wheeler Aligner) [42,43] generating the alignment files. The GATK4 (Genome Analysis Tool Kit version 4) pipeline using the best practices guideline [44] was used to variant discovery using the HaplotypeCaller option. Only variants having an overall read depth greater than 20X and having a variant allele depth of at least 10X were filtered for downstream analysis. The resultant variants were annotated using ANNOVAR [45]. Only pathogenic variants located in genes MLH1, MSH2, MSH6, BRAF and KRAS were filtered out for consideration in this study.

Methylation Array Testing

Targeted Nextgen Bisulphite Sequencing (Tngbs)

Assay Design: Each regulatory element of a requested gene was carefully evaluated before beginning the process of assay design. Gene sequences containing the target of interest were acquired from the Ensembl genome browser and annotated. The target sequences were re-evaluated against the UCSC genome browser for repeat sequences including LINE, SINE, and LTR elements. Sequences containing repetitive elements, low sequence complexity, high thymidine content, and high CpG density were excluded from the in-silico design process. 2. Methods a. Sample Digestion Dissolution of paraffin was accomplished by the addition of 1mL of HistoChoice® Clearing Agent (Sigma-Aldrich; St. Louis, MO; cat# H2779) and incubation at 65°C for 30 min. Samples were digested by the addition of 200µL of digestion buffer consisting of 20µL 10X Target Retrieval Solution high pH (Agilent; Santa Clara, CA; cat# S2375), 160µL of Buffer ATL (Qiagen; Hilden, Germany; cat# 939011), and 20µL of protease K, followed by incubation overnight at 65°C. Samples were vortexed and checked for complete digestion. b. Bisulfite Modification 20µL of the supernatant from the sample extracts were bisulfite modified using the EZ-96 DNA Methylation-Direct Kit™ (ZymoResearch; Irvine, CA; cat# D5023) as per the manufacturer’s protocol with minor modification. The bisulfite modified DNA samples were eluted using Melution buffer in 46µL. c. Multiplex PCR i. All bisulfite modified DNA samples were amplified using separate multiplex or simplex PCRs. PCRs included 0.5 units of HotStarTaq (Qiagen; Hilden, Germany; cat# 203205), 0.2µM primers, and 3µL of bisulfite-treated DNA in a 20µL reaction. All PCR products were verified using the Qiagen QIAxcel Advanced System (v1.0.6). Prior to library preparation, PCR products from the same sample were pooled and then purified using the QIAquick PCR Purification Kit columns or plates (cat# 28106 or 28183). ii. Recommended PCR cycling conditions: 95°C 15 min; 45 x (95°C 30s; Ta°C 30 s; 68°C 30 s); 68°C 5 min; 4°C ∞ iii. Customer samples were run alongside established reference DNA samples with a range of methylation. They were created by mixing high- and low-methylated DNA to obtain samples with 0, 50, and 100% methylation. The high-methylated DNA was in vitro enzymatically methylated genomic DNA with >85% methylation. The low-methylated DNA was chemically and enzymatically treated with <5% methylation. They were first tested on numerous gene-specific and global methylation assays using pyrosequencing.

Library Preparation and Sequencing Libraries were prepared using a custom Library Preparation method created by EpigenDx. Next, library molecules were purified using Agencourt AMPure XP beads (Beckman Coulter; Brea, CA; cat# A63882). Barcoded samples were then pooled in an equimolar fashion before template preparation and enrichment were performed on the Ion Chef™ system using Ion 520™ & Ion 530™ ExT Chef reagents (Thermo Fisher; Waltham, MA; cat# A30670). Following this, enriched, template-positive library molecules were sequenced on the Ion S5™ sequencer using an Ion 530™ sequencing chip (cat# A27764). The Targeted NextGen Bisulphite sequencing (tNGBS) was carried out at the EpigenDx Laboratory, Hopkinton, Massachusetts, USA.

Data Analysis FASTQ files from the Ion Torrent S5 server were aligned to a local reference database using the open-source Bismark Bisulfite Read Mapper program (v0.12.2) with the Bowtie2 alignment algorithm (v2.2.3). Methylation levels were calculated in Bismark by dividing the number of methylated reads by the total number of reads. An R-squared value (RSQ) was previously calculated from controls set at known methylation levels to test for PCR bias.

CIMP testing was completed based on a validated quantitative DNA methylation assay using a fourteen (14) gene panel which included APC, CACNA1G (MINT31), CDKN2A, CRABP1, IGF2, IGFBP3, MGMT, RASSF1, SEPTIN9, SFRP2, SOCS1, SV2C(MINT1), TMEFF1 (HPP1), and WIFI. One of the genes, IGFBP3 failed all assays and therefore was excluded, resulting in 13 gene markers being used for analysis.

Tumors were classified as CIMP-positive if the percentage of methylated reference (PMR) ratio was >10 for at least six of the 13 markers and as non-CIMP if the PCR ratio was >10 for fewer than six markers. PMR was calculated as the amount of methylated tumor DNA at a specific locus divided by the ALU-normalised amount in a methylated reference sample, multiplied by 100. If a specific site on a gene was more than 50% methylated, it was considered a methylated position. If a sample failed an assay, it was considered normal or not methylated. If 60% of the sites of a gene were methylated, the gene was considered highly methylated as well. If 6 or more (≥6) genes out of 13 genes were methylated then it was considered CIMP positive. If less than 6 genes out of 13 genes were methylated then it was considered CIMP negative.

A total of 127 CRC patients diagnosed between 1^st January 2008 to 15^th September 2019 were analysed for molecular subtypes of CRC. Of these cases, the mean(SD) age of the MSI cases was 55.6(16.9) years and of the MSS cases was 55.4(15.5) years. The median(IQR) age for MSI cases was 53(43-68.5) years whilst for MSS was 54(46-67) years. Left-sided colon tumours were more common at 84(66.1%) whilst right-sided colon tumours constituted 43(33.9%) of all cases. AJCC stage (III+IV) was found in the majority, 109(85.8%) of cases whilst early-stage (I+II) CRC constituted 18(14.2%) cases. AC was the most common histopathological subtype at 114(89.8%) followed by 11(8.66%) invasive mucinous adenocarcinoma and 2(1.57%) SRCC.

MSI Colorectal Adenocarcinoma Molecular Characterization

The majority were MSS, 75(59.06%) followed by MSI, 52(40.9%). MSI gene mutation analysis for MLH-1, MSH2 and MSH-6 was performed on all the 127 cases. The analysis demonstrated 14(11.02%) MLH-1 mutations; 30(23.62%) MSH2 mutations and 26(20.47%) MSH6 mutations. There were 4(7.7%) MLH1 only mutations, 14(26.9%) MSH6 only mutations and 18(34.6%) MSH2 only mutations. Of the MSI positive cases, MSH2+MSH6 mutations made up 6(11.5%); MLH1+MSH2 mutations constituted 4(7.7%) and MLH1+MSH6 mutations constituted 4(7.7%). Out of the MSI positive cases 2(3.8%) cases had mutations in MLH1+MSH2+MSH6.

MLH-1 and PMS2 Expression Defects on Immunohistochemistry

Of the 127 participants with CRC, that had MLH-1immunohistochemistry, 85(66.9%) tumours were MMR protein proficient in MLH1. There were 42(33.1%) tumours that had an MMR protein defect in MLH-1. Among the 42(33.1%) tumours that were MMR MLH-1 deficient, 14(11%) had a mutation inMLH-1 on next generation sequencing, whilst 28(22%) tumours had no MLH-1 mutation on next generation sequencing. Out of the 85(66.9%) tumours that were MLH-1 protein proficient,84(66.1%) had no MLH-1 mutation on next generation sequencing, whilst 2(1.6%) tumours had a mutation in MLH-1 on next generation sequencing. Of the 127 participants with CRC that had MLH-1 and PMS2immunohistochemistry, 37(29.1%) had at least one MMR proteindefect in either MLH1 or PMS2. Among them, 21(16.5%) participants were identified with isolated loss of PMS2 expression; 26(20.5%) had a defect in the expression of both MLH-1 andPMS2. The 21(16.5%) participants with isolated PMS2 MMRdefect expression had no pathologic mutations in MLH-1, MSH-2 and MSH-6.

MSS and MSI Subtypes

MSI positive tumours were found in 52(40.9%) of all CRCs in Ugandans. Compared to 46.7% MSS cases which were <54 years, there were 26(50%) MSI cases which were <54 years. Although there were more MSI cases <54 years this finding did not reach statistical significance (p=0.712). AC was more common in MSS tumours compared to MSI tumours (56.1% versus 43.9%) and this reached statistical significance (p=0.039). The presence of LVI was more common compared to the absence of LVI in MSI positive tumours (42.1% versus 30.8%) however this did not reach statistical significance (p=0.431). Other clinicopathological features in particular topography, stage and grade did not show any significant differences between MSI positive tumours and MSS tumours (Table 1).

Table 1: Clinicopathological characteristics of MSI and MSS CRC subtypes.

Table 17 illustrates a log binomial regression analysis for risk factors for MSI CRC. Patients >54 years were 8% less likely to have MSI however this did not reach statistical significance (p=0.712). The risk of LVI was 1.4x higher in MSI positive cases than MSS cases, however, this did not reach statistical significance (p=0.466). There was also a higher risk of SRCC (1.14x), stage III (2.33x) and stage IV(1.57x) having MSI compared to MSS, however, none of these findings were statistically significant (Table 17). The other variables particularly, sex, grade and location were not significantly associated with MSI status (Table 2).

Table 2:  Log Binomial Regression analysis for risk factors for MSI CRC.

BRAF and KRAS Mutation Analysis

BRAF mutational analysis was performed on all the 127 cases with only 5(3.9%) having pathologic missense BRAF V600 mutations and 1(0.8%) was a benign synonymous variant. Only 1(0.9%) case had a BRAF mutation in all MSI positive cases. The other 3(2.4%) BRAF mutations were in MSS(MSI negative) patients. The two pathologic variants included Glu269Gly and Gly506Glu. The commonest pathologic BRAF mutation variant was Glu269Gly.

KRAS mutational analysis was performed on all 127 cases with only 8(6.3%) having pathologic missense KRAS mutations and 1(0.8%) case was a benign synonymous variant. The pathologic variants included Gly13Asp, Gly12Asp, Gly12Ser, Gln61His and Asn116Ser. These included two Gly12Asp pathogenic variants, 2 Gly12Ser pathogenic variants and 2 Asn116Ser pathogenic variants. There was one Gln43Gln benign KRAS synonymous variant.

CIMP Analysis

The DNA samples were transferred from the Institute of Genetic and Cancer at the University of Edinburgh, UK to EpigenDx Laboratory, Hopkinton, Massachusetts, USA for targeted NextGen Bisulphite sequencing (tNGBS).CIMP was defined when more than ≥6 genes out of the 13 gene panel were methylated. One of the genes, IGFBP3 failed an assay in all the samples and therefore thirteen (13) genes in the panel were used.  Out of 92 cases which had an adequate quantity of DNA to carry out CIMP analysis, 11(11.9%) were CIMP positive and 81(88.0%) were CIMP negative. CIMP positive tumours represented 3(5.8%) of MSI positive tumours compared with 8(10.7%) of MSS tumours. There were 7(11.1%) CIMP positive tumours in the left colon and 4(13.8%) CIMP positive tumours in the right colon and this did not reach statistical significance (p=0.713). Table 3 demonstrates the CIMP status of all the CRC cases analysed.

Table 3: CIMP status of each CRC tissue sample.

CIMP Status summary

CIMP- 81

CIMP+ 11

In the present study the prevalence of CIMP positive tumours in Uganda has been found to be low compared to Wester populations [46-48]. The prevalence of CIMP positive CRC has been found to be 9.8% in Uganda which is comparable to the prevalence of 6.4% and 11.6% in two different studies from South Korea [49,50].

Studies have shown that colorectal cancer with CpG island methylator phenotype (CIMP) has a distinct genetic profile (Ogino S et al., 2006). The CIMP tumours have significantly higher BRAF mutation frequencies than non-CIMP counterparts. Independent of MSI status, CIMP is a distinct biological subtype of colorectal cancer [28]. In MSI-H tumours., using quantitative DNA methylation analysis, a clear bimodal distribution of MSH-H tumours has been found according to the number of methylated promoters, and CIMP has been found to be associated with high BRAF and low KRAS mutation rates [28].

The subset of colorectal cancers with promotor methylation in multiple genes are CIMP colorectal cancers [18, 20-22, 25, 26]. Unimodal distributions of the numbers of methylated loci were shown in most previous studies [20, 25, 51, 52], however, to be classified as CIMP, it has been unclear how many loci should be methylated for a given tumour. Some studies have suggested that a spectrum of predicted distributions of stochastic methylation events, constitutes CIMP [51,52].

MSP was used in previous studies to determine promotor methylation in multiple genes [18,20-22,25,26,51,52]. A major limitation of MSP based methylation is that it cannot reliably distinguish low levels from high levels of methylation with no biological significance. Most colorectal cancers with low levels of promotor methylation in MGMT, CDKN2A (p16) or MLH1 (PMR<4) have been shown to have an intact protein expression, resulting in little biological significance of low levels of DNA methylation in these loci [28].

The choice of gene promoters to determine the CIMP status is also important. The pattern of methylation is tumour specific and aberrant CpG island methylation occurs in a non-random fashion [53]. Therefore, to determine the CIMP status, a carefully selected and validated panel of methylation markers should be used. The five gene promotors, CACNA1G, CRABP1, CDKN2A, NEUROG1 and MLH1 as well as MGMT have been evaluated in previous studies for suitability in the CIMP panel. Data from other studies has shown that MGMT has a low specificity (66%) and low sensitivity (62%) for the determination of CIMP and may therefore be excluded from a CIMP panel. The other five markers correlate well with the overall CIMP status and have a very high specificity and sensitivity (>90%). MGMT methylation has been found to be associated with G>A mutations in TP53 and KRAS [54], resulting in field effects which lead to the development of CRC [55,56].

Previous studies were not population based and the number of participants was limited. One large retrospective population-based study showed unique associations of CIMP with various clinicopathological features [22]. Using MSP based assays, previous studies have shown that 30-35% of colorectal cancers are CIMP positive [18,20,21,26]. The reported frequencies of CIMP from these studies may be overestimated as MSP may detect low levels of methylation. A study by Ogino S et al., found that CIMP is less frequent at 17% [28]. However, when in the same study. CIMP was defined as methylation in ≥3/6 loci, the CIMP frequency increased to 32%. Threse results are consistent with previous reports [20,25,51,52] which shows that 10% or more tumours might have been misclassified as CIMP. The associations of CIMP with various clinicopathological features may have been obscured due to misdiagnosis of CIMP in previous studies. Therefore, for research in cancer epigenetics, quantitative DNA methylation analysis should be validated.

CIMP has been found to be more commonly associated with sporadic MSI-H tumours than in MSI-H tumours in the setting of hereditary non-polyposis colorectal cancer [18,20,22,25,52,57]. Our study did not find this association as we had a few cases of sporadic MSI positive tumours in our Ugandan population. The majority of MSI-positive tumours in the Ugandan setting are in the context of hereditary non-polyposis colorectal cancer (7). Our study showed that all the BRAF mutated tumours were CIMP positive. These findings are consistent with another study from the USA which showed that CIMP positive tumours were associated with 63% BRAF mutation frequencies, in contrast to CIMP negative tumours which were not associated with BRAF mutations (0%) [28]. As our study showed a low BRAF mutation frequency in Uganda, we were unable to determine whether a difference in BRAF mutation frequency exists for MSS CIMP positive tumours and MSS CIMP negative tumours. However, a study from the USA showed that BRAF mutations were present in only 6.6% of non-CIMP MSS tumours, in contrast to 54% of CIMP positive MSS tumours. These findings support an association between CIMP and BRAF mutations in the West similar to previous studies [22,23,24].

An association has been found in previous studies between CIMP and KRAS mutation [18,22]. However, our study as in other previous studies has shown that CIMP positive tumours have a lower KRAS mutation frequency than CIMP negative tumours [20,23,24,25,58]. Therefore, our data supports an inverse association of CIMP with KRAS mutations.

CIMP positive tumours have been associated with a poorer clinical outcome in many studies [28,56]. However, studies have also shown controversial results on the independent role of CIMP on clinical outcome [50,62]. Brae JM, found that CIMP is an independent prognostic factor for rectal tumours but not for colon tumours on survival analysis [49]. A poorer clinical outcome was found for CIMP positive rectal tumours and this poor survival outcome decreases from the rectum to the proximal colon [49]. In the present study, only a few patients were CIMP positive, with only two CRC tumours being CIMP positive in the right colon, and the majority of CRCs in our patient population were in the rectum and sigmoid colon. Hence our study had a small sample size to determine whether a poorer outcome was associated for rectal tumours compared to right-sided colon tumours. Therefore, the survival outcome of CIMP positive CRC tumours according to bowel subsite could not be determined.

Yamauchi et al., showed no linear relationship between MSI positive, CIMP positive and BRAF mutation frequencies and tumour location [63]. These findings are in contrast to findings by Benedix F et al., which showed a linear trend for caecal tumours in terms of age, stage, grade, histological subtype, MSI status and CIMP status [64]. The present study in Uganda showed no linear trend, in favour of the right colon, due to the small sample size, with the majority of CIMP positive tumours in the left colon and only two CIMP positive tumours in the right colon.

Compared to Western developed high-income countries, in Uganda the prevalence of CIMP positive tumours is low. Colorectal patients with the BRAF mutation were CIMP positive and CIMP tumours are a distinct epigenetic subtype of colorectal cancer in Ugandan patients.

Ethical Approval

This study was part of the PhD work, which was approved by the Doctoral Committee and Higher Degrees Research and Ethics Committee of the School of Biomedical Sciences, College of Health Sciences, Makerere University for the corresponding author(SBS-HDREC-630). Final approval of this research study was obtained from the Uganda National Council for Science and Technology (HS-2574). Written informed consent was obtained from prospective participants included in the study before completing the Data Extraction Form. All the data and specimen spertaining to the research were kept confidential. For the retrospective arm of the study, data were abstracted from the case files in the respective hospitals. Therefore, a waiver of consent was obtained from the Higher Degrees Research and Ethics Committee, School of Biomedical Sciences, College of Health Sciences,Makerere University, to access the data and perform the experiments on the tissue block samples. The patient data, whichwere accessed from the medical case files in the respective hospitals, were anonymized and maintained with confidentiality.The conduct of the study was in accordance with the Declaration of Helsinki.

Consent for Publication

Consent was obtained from all the participants enrolled in this study.

Competing Interests

The authors declare that they have no competing interests.

Funding

The authors declare that they received no specific funding for this work. However, the corresponding author and grants fromProfessor Ian Tomlinson funded this part of the corresponding author’s PhD research study. The methylation array testing was supported through the funds of the corresponding author. No payment was received by theauthors to write and publish this part of the study.

Authors’ Contributions

Richard Wismayer conceived the concept and proposal, collected the data, performed data analysis and wrote the paper. Rosie Matthew sextracted DNA from colorectal cancer tissue samples. Celina Whalley designed and performed mutation analysis and DNA sequencing. Fredrick Elishama Kakembo and Steve Thorn carriedout bioinformatics analysis of the variant data. Michael Odida andHenry Wabinga interpreted all the immunhistochemical slides.Julius Kiwanuka performed data analysis and provided statisticalsupport. Henry Wabinga, Michael Odida and Ian Tomlinsoncarried out critical reviews of the manuscript for intellectualcontent. All authors approved the final manuscript.

Acknowledgements

This study was part of a collaboration between the University of Edinburgh and Makerere University. Professor Ian Tomlinson is acknowledged for funding the laboratory experiments of this study from grants pertaining to the 1000 Genomes Project. We thank all the staff and research assistants from the Institute of Genetics and Cancer, University of Edinburgh and the Institute of Cancer and Genomic Sciences, University of Birmingham. We are also grateful to Ms Dorothy Nabbale for the immune histochemistry laboratory technical work carried out for this part of the colorectal cancer research project in the Department of Pathology, School of Biomedical Sciences, College of Health Sciences, Makerere University. The corresponding author wishes to thank the laboratory research assistants at the Epigen Dx Laboratory, Hopkinton, Massachusetts, USA for carrying out the methylation array testing experiments for this research project. We thank the Department of Pathology, Makerere University forusing the database and retrieving the tissue block samples from the archives. Finally, we would like to thank the research assistants, staff and recruited patients from the Department of Surgery of Masaka Regional Referral Hospital, Mulago National Referral Hospital, Uganda Martyrs’ Hospital Lubaga and Mengo Hospital.

Availability of Data and Materials

The data supporting the findings of this study are available from the corresponding author upon reasonablerequest.
Abbreviations

AC- Classical adenocarcinoma

APC- Adenomatous Polyposis Coli

AJCC- American Joint Committee on Cancer

CRC- Colorectal cancer

CIN- Chromosomal instability

CSS- Cancer specific survival

CIMP – CpG island Methylator Phenotype

DNA- Deoxyribonucleic acid

FAP- Familial Adenomatous Polyposis

FFPE- Formalin fixed paraffin embedded

HNPCC- Hereditary Non-Polyposis Colorectal Cancer

HD-REC- Higher Degrees Research and Ethics Committee

IHC- Immunohistochemistry

LS- Lynch syndrome

LVI- Lymphovascular invasion

MAC- Mucinous adenocarcinoma

MSI- Microsatellite Instability

MMR- Mismatch repair

MSS- Microsatellite stability

PCR- Polymerase chain reaction

SRCC- Signet ring colorectal cancer

TNM- Tumor Node Metastasis

1. Pancione M, Remo A, Colantuoni V. Genetic and epigenetic events generate multiple pathways in colorectal cancer progression. Patholog Res Int. 2012; 2012: 509348.
2. Haggar FA, Boushey RP. Colorectal cancer epidemiology: incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg. 2009; 22: 191-197.
3. Chan AT, Ogino S, Fuchs CS. Aspirin use and survival after diagnosis of colorectal cancer. JAMA. 2009; 302: 649-658.
4. Slattery ML, Wolff RK, Herrick J, Caan BJ, Potter JD. Leptin and leptin receptor genotypes and colon cancer: gene-gene and gene-lifestyle interactions. Int J Cancer. 2008; 122: 1611-1617.
5. Worthley DL, Leggett BA. Colorectal cancer: molecular features and clinical opportunities. Clin Biochem Rev. 2010; 31: 31-38. 
6. Wismayer R, Kiwanuka J, Wabinga H and Odida M. CDX2 as A Prognostic Marker for Colorectal Adenocarcinoma in Ugandan Patients. Int J Surg Clin Pract. 2024; 6: 170.
7. Wismayer R, Kiwanuka J, Wabinga H, Odida M. Colorectal adenocarcinoma in Uganda: are right-sided and left-sided colon cancers two distinct disease entities? World J Surg Oncol. 2023; 21: 215.
8. Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, et al. A National Cancer Institute Workshop on Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res. 1998; 58: 5248–5257.
9. Boland CR, Goel A. Microsatellite instability in colorectal cancer. Gastroenterology. 2010; 138: 2073-2087.
10. Wu CW, Ng SSM, Dong YJ, Ng SC, Leung WW, Lee CW, Wong YN, et al. "Detection of miR-92a and miR-21 in stool samples as potential screening biomarkers for colorectal cancer and polyps." Gut. 2012; 61: 739-745.
11. Pino MS, Chung DC. Microsatellite instability in the management of colorectal cancer. Expert Rev Gastroenterol Hepatol. 2011; 5: 385-399.
12. Laird PW. Cancer epigenetics. Hum Mol Genet. 2005; 14: R65-R76.
13. Issa JP. CpG island methylator phenotype in cancer. Nat Rev Cancer. 2004; 4: 988-993.
14. Rashid A, Issa JPJ. CpG island methylation in gastroenterologic neoplasia: a maturing field. Gastroenterology. 2004; 127: 1578-1588.
15. Kondo Y, Issa JPJ. Epigenetic changes in colorectal cancer. Cancer Metastasis Rev. 2004; 23: 29-39.
16. Jass JR. Serrated adenoma of the colorectum and the DNA?methylator phenotype. Nat Clin Pract Oncol. 2005; 2: 398-405.
17. Toyota M, Ohe-Toyota M, Ahuja N, Issa JPJ. Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype. Proc Natl Acad Sci. 2000; 97: 710-715.
18. Minoru T, Suzuki H, Shinomura Y. The epigenome of colorectal cancer. Current Colorectal Cancer Reports. 2009; 5: 84-89.
19. van Rijnsoever M, Grieu F, Elsaleh H, Joseph D, Lacopetta B. Characterisation of colorectal cancers showing hypermethylation at multiple CpG islands. Gut. 2002; 51: 797-802.
20. Van Rijnsoever M, Elsaleh H, Joseph D, McCaul K, Lacopetta B. CpG island methylator phenotype is an independent predictor of survival benefit from 5?fluorouracil in stage III colorectal cancer. Clin Cancer Res. 2003; 9: 2898-2903.
21. Samowitz WS, Albertsen H, Herrick J, Levin TR, Sweeney C, Murtaugh MA, et al. Evaluation of a large, population?based sample supports a CpG island methylator phenotype in colon cancer. Gastroenterology. 2005; 129: 837-845.
22. Nagasaka T, Sasamoto H, Notohara K, Cullings HM, Takeda M, Kimura K, et al. Colorectal cancer with mutation in BRAF, KRAS, and wild?type with respect to both oncogenes showing different patterns of DNA methylation. J Clin Oncol. 2004; 22: 4584-4594.
23. Kambara T, Simms LA, Whitehall VLJ, Spring KJ, Wynter CVA, Walsh MD, et al. BRAF mutation is associated with DNA methylation in serrated polyps and cancers of the colorectum. Gut. 2004; 53: 1137-1144.
24. Hawkins N, Norrie M, Cheong K, Mokany E, Ku SL, Meagher A, et al. CpG island methylation in sporadic colorectal cancers and its relationship to microsatellite instability. Gastroenterology. 2002; 122: 1376-1387.
25. Toyota M, Ahuja N, Ohe?Toyota M, Herman JG, Baylin SB, Issa JPJ, et al. CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci. 1999; 96: 8681-8686.
26. Grady WM. Genomic instability and colon cancer. Cancer Metastasis Rev. 2004; 23: 11-27.
27. Ogino S, Kawasaki T, Brahmandam M, Cantar M, Kirkner GJ, Spigelman D, et al. Precision and performance characteristics of bisulfite conversion and real?time PCR (MethyLight) for quantitative DNA methylation analysis. J Mol Diagn. 2006; 8: 209-217.
28. Issa JPJ, Shen L, Toyota M. CIMP, at last. Gastroenterology. 2005; 129: 1121-1124.
29. Cottrell SE, Distler J, Goodman NS, Mooney SH, Kluth A, et al. A real?time PCR assay for DNA?methylation using methylation?specific blockers. Nucleic Acids Res. 2004; 32: e10.
30. Thomassin H, Kress C, Grange T. MethylQuant: a sensitive method for quantifying methylation of specific cytosines within the genome. Nucleic Acids Res. 2004; 32: e168.
31. Widschwendter M, Siegmund KD, Muller HM, Fiegi H, Marth C, et al. Association of breast cancer DNA methylation profiles with hormone receptor status and response to tamoxifen. Cancer Res. 2004; 64: 3807-3813.
32. Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, et al. MethyLight: a high?throughput assay to measure DNA methylation. Nucleic Acids Res. 2000; 8: E32-00.
33. Zeschnigk M, Bohringer S, Price EA, Onadim Z, Mabhofer L, Lohmanm DR. A novel real?time PCR assay for quantitative analysis of methylated alleles (QAMA): analysis of the retinoblastoma locus. Nucleic Acids Res. 2004; 32: e125.
34. Mito JK, Mitra D, Barysauskas CM, Mariño-Enriquez A, Morgan EA, Fletcher CDM, et al. A Comparison of Outcomes and Prognostic Features for Radiation-Associated Angiosarcoma of the Breast and Other Radiation-Associated Sarcomas. Int J Radiat Oncol Biol Phys. 2019; 104: 425-435.
35. Salipante SJ, Scroggins SM, Hampel HL, Turner EH, Pritchard CC. Microsatellite instability detection by next generation sequencing. Clin Chem. 2014; 60: 1192-1199.
36. Pritchard CC, Grady WM. Colorectal cancer molecular biology moves into clinical practice. Gut. 2011; 60: 116-29.
37. Haraldsdottir S, Einarsdottir HM, Smaradottir A, Gunnlaugsson A, Halfdanarson TR. Krabbamein i ristli og endaþarmi. [Colorectal cancer - review]. Laeknabladid. 2014; 100: 75-82.
38. Austin JD. Understanding the Relatlonships between Survivorship Care Plans and Survivor Level Outcomes among Breast and Colorectal Cancer Survivors. UTHealth School of Public Health. 2019.
39. Pearlman R, Frankel WL, Swanson B, Zhao W, Yilmaz A, Miller K, et al. Prevalence and Spectrum of Germ line Cancer Susceptibility Gene Mutations Among Patients with Early-Onset Colorectal Cancer. JAMA Oncol. 2017; 3: 464-471.
40. Friedman DB, Adams SA, Brandt HM et al. Rise Up, Get Tested, and Live: An Arts-Based Colorectal Cancer Educational Program in a Faith-Based Setting. J Canc Educ. 2019; 34: 550-555.
41. Ewels PA, Peltzer A, Filinger S, Patel H, Alneberg J, Wilm A, et al. The nf-core framework for community-curated bioinformatics pipelines. Nature biotechnology. 2020; 38: 276-278.
42. Li FY, Lai MD. Colorectal cancer, one entity or three. J Zhejiang Univ Sci B. 2009; 10: 219-229.
43. Leen V, Kenis C, Conings G, Milisen K, Prenen H, Cutsem EV, et al. Impact of Geriatric Assessment on Treatment Decisions and Follow-Up in Older Colorectal Cancer Patients. Annals of Oncology. 2013; 24: iv96.
44. Wang D, Dubois RN. The role of COX-2 in intestinal inflammation and colorectal cancer. Oncogene. 2010; 29: 781-788.
45. George JC, You YNY, Russell CA, Tierno MB, Turner M, Bennett JP, et al. Young-Onset Colon Cancer and Recurrence Risk by Gene Expression. JNCI: Journal of the National Cancer Institute. 2020; 112: 1170-1173.
46. Chang SC, Lin JK, Yang SH, Wang HS, Li AF, Chi CW. Relationship between genetic alterations and prognosis in sporadic colorectal cancer. Int J Cancer. 2006; 118: 1721-1727.
47. Hideho T, Tsunehida O, Shigeyoshi I, Ryo Y, Masataka N, Seiki I, et al. Changing site distribution of colorectal cancer in Japan. Diseases of the colon and rectum. 2020; 45: 1249-1254.
48. Bae JM, Kim JH, Cho NY, Kim Ty, Kang GH. Prognostic implication of the CpG island methylator phenotype in colorectal cancers depends on tumour location. British J cancer. 2013; 109: 1004-1012.
49. Hong BS, Cho JH, Kim HJ, Choi EJ, Rho S, Kim J, et al. Colorectal cancer cell-derived microvesicles are enriched in cell cycle-related mRNAs that promote proliferation of endothelial cells. BMC Genomics. 2009; 10: 1-13.
50. Anacleto C, Leopoldino AM, Rossi B, Soares FA, Lopes A, Rocha JC, et al. Colorectal cancer "methylator phenotype": fact or artifact?. Neoplasia. 2005; 7: 331-335.
51. Yamashita K, Dai T, Dai Y, Yamamoto F, Perucho M. Genetics supersedes epigenetics in colon cancer phenotype. Cancer cell. 2003; 4: 121-131.
52. Costello JF, Fruhwald MC, Smiraglia DJ, Rush LJ, Robertson GP, Gao X, et al. Aberrant CpG?island methylation has non?random and tumour?type?specific patterns. Nat Genet. 2000; 24: 132-138.
53. Esteller M, Risques R A, Toyota M, et al. Promoter hypermethylation of the DNA repair gene O(6)?methylguanine?DNA methyltransferase is associated with the presence of G:C to A:T transition mutations in p53 in human colorectal tumorigenesis. Cancer Res. 2001; 61: 4689-4692.
54. Giovannucci E, Ogino S. DNA methylation, field effects, and colorectal cancer. J Natl Cancer Inst. 2005; 97: 1317-1319.
55. Shen L, Kondo Y, Rosner GL, Xiao L, Hernandez NS, Vilaythong J, et al. MGMT promoter methylation and the field defect in sporadic colorectal cancer. J Natl Cancer Inst. 2005; 97: 1330-1338.
56. Yamamoto H, Min Y, Itoh F.et al. Differential involvement of the hypermethylator phenotype in hereditary and sporadic colorectal cancers with high?frequency microsatellite instability. Genes Chromosomes Cancer. 2002; 33: 322-325.
57. Whitehall VLJ, Wynter CVA, Walsh MD, Simms LA, Purdie D, Pandeya N, et al. Morphological and molecular heterogeneity within nonmicrosatellite instability?high colorectal cancer. Cancer Res. 2002; 62: 6011-6014.
58. Ogino S, Nosho K, Irahara N, Meyerhardt JA, Baba Y, Shima K, et al. Lymphocytic reaction to colorectal cancer is associated with longer survival, independent of lymph node count, microsatellite instability, and CpG island methylator phenotype. Clin Cancer Res. 2009; 15: 6412-6420.
59. Shen L, Toyota M, Kondo Y, Lin E, Zhang L, Guo Y, et al. Integrated genetic and epigenetic analysis identifies three different subclasses of colon cancer. Proc Natl Acad Sci. 2007; 104: 18654-18659.
60. Dahlin AM, Palmqvist R, Henriksson ML, Jacobsson M, Eklöf V, Rutegård J, et al. The role of the CpG island methylator phenotype in colorectal cancer prognosis depends on microsatellite instability screening status. Clin Cancer Res. 2010; 16: 1845-1855.
61. Sanchez JA, Krumroy L, Plummer S, Aung P, Merkulova A, Skacel M, et al. Genetic and epigenetic classifications define clinical phenotypes and determine patient outcomes in colorectal cancer. Journal of British Surgery. 2009; 96: 1196-1204.
62. Yamauchi M, Lochhead P, Morikawa T, Huttenhower C, Chan AT, Giovannucci E, et al. Colorectal cancer: a tale of two sides or a continuum? Gut. 2012; 61: 794-797.
63. Benedix F, Schmidt U, Mroczkowski P, Gastinger I, Lippert H, Kube R. Colon carcinoma-classification into right and left sided cancer or according to colonic subsite?. Analysis of 29,568 patients. Eur J Surg Oncol. 2011; 37: 134-139.