Main

MUTYH encodes a DNA glycosylase which is expressed in nucleus as well as in mitochondria1 where reactive oxygen species are produced as by-products of cellular respiration. Changes in mitochondrial functions, associated with reactive oxygen species production, have recently been associated with perturbation of cellular senescence process.2 MUTYH takes part in this process since the accumulation of oxidative products in nuclear and mitochondrial DNAs results into the activation of two distinct cell-death pathways, both initiated by MUTYH-generated single-strand breaks.3

MUTYH-associated-polyposis is a recessive form of inherited polyposis associated with germline inactivating mutations of MUTYH; this gene encodes a base excision repair protein counteracting the DNA damage induced by the oxidative stress.4, 5 Although clinically variable, MUTYH-associated-polyposis phenotype resembles that of APC-linked attenuated familial polyposis with the onset in the fourth-fifth decade of life, a limited number of adenomas (generally 30–100) and an increased susceptibility to colorectal cancer. However, unlike attenuated familial adenomatous polyposis, hyperplastic and sessile serrated polyps can develop;6, 7 in addition, approximately 60% of patients with polyposis have colorectal cancer at presentation.8

Colorectal cancer cells of MUTYH-associated-polyposis patients contain an excess of c.34 G>T transversions in KRAS gene; this is due to failure to repair mismatches induced by 8-oxo-guanine (8-oxoG) variant base, a widely recognized hallmark of oxidative stress.8, 9, 10 Although mutations of BRAF gene have been recognized as early genetic events in the initiation of the ‘serrated neoplastic pathway’,11 limited BRAF analyzes on MUTYH-associated-polyposis colorectal cancers have been performed.9 On the other hand, the pathogenetic role of BRAFV600E mutation in sebaceous gland hyperplasia, a well-known extracolonic manifestation of MUTYH-associated-polyposis syndrome, has been demonstrated in affected individuals from MUTYH-associated-polyposis pedigrees.12

On the whole, MUTYH driven carcinogenesis is only partially known. Overall, MUTYH-associated-polyposis cancers appear to follow a distinct pathway compared to other colorectal cancers: some features overlap with chromosomal instability colorectal cancer phenotype, including frequent APC and KRAS mutations,9, 13 some others with microsatellite instability phenotype, including loss of HLA class I protein expression.14 Recently, Nieuwenhuis and collaborators15 have suggested the occurrence of an accelerated disease progression in MUTYH-associated-polyposis, underlying the high risk for patients to develop colorectal cancer even under surveillance.

Several studies support the involvement of respiratory chain defects in the development of tumors, including colorectal cancer, indicating that progression of these lesions can be affected by a disturbed oxidative phosphorylation.16 In addition to rRNA and tRNA genes, the phylogenetically conserved region of mitochondrial DNA includes sequences encoding for some of the polypeptides which have a role in the control of electron chain transport, such as cytochrome oxidase I and II subunits (MT-CO1/MT-CO2). Mutations affecting these genes have been identified in colorectal cancer cell lines as well as in tumor specimens from colorectal cancer patients.17, 18, 19 In particular, by sequencing the whole mitochondrial DNA in gastrointestinal precancerous lesions, Sui and collaborators20 found the highest mutation frequency in three of the conserved coding sequences (MT-CO1, MT-ND4 and MT-ND5 subunits of NADH dehydrogenase).

To date, no data concerning mitochondrial DNA alterations affecting MUTYH-associated-polyposis or classical/attenuated familial polyposis lesions have been reported. This study was therefore aimed at identifying molecular drivers of MUTYH-associated-polyposis with respect to classical/attenuated familial polyposis carcinogenesis. Due to the active role of MUTYH at both nuclear and mitochondrial level, we focused on the possible relationship between nuclear and mitochondrial alterations, searching for mutations in KRAS and BRAF genes as well as in phylogenetically conserved coding regions of mitochondrial DNA; in particular, we analyzed MT-CO1/MT-CO2 genes, involved in oxidative phosphorylation control, and the contiguously located tRNA aspartic acid (MT-TD) gene.

Patients and methods

Patients and Specimens

The study cohort consisted of 12 subjects carrying different biallelic MUTYH germline mutations (MUTYH-associated-polyposis patients) and 13 subjects with different APC germline alterations (11 classical-familial adenomatous polyposis and 2 attenuated-familial adenomatous polyposis). Among patients there were 2 pairs of MUTYH-associated-polyposis affected sisters (M1 and M2; M3 and M4, Table 1) and one pair of classical-familial adenomatous polyposis affected brothers (F1 and F2, Table 1). All patients were of Italian origin and had been previously undergone to MUTYH/APC genetic testing. Overall, we analyzed 55 colorectal adenomas and 5 colorectal carcinomas from MUTYH-associated-polyposis patients and 41 colorectal adenomas and 2 colorectal carcinomas from classical/attenuated-familial polyposis subjects. Adenomas were cross-matched for histotype, size (range: 2–10 mm), and grade of dysplasia. Grading and staging of colorectal adenocarcinomas were in accordance with the TNM system. In most cases, multiple adenomas (2–13) from the same patient were examined. For one MUTYH-associated-polyposis patient (M6) only adenocarcinoma was available (Table 1). All germline mutations as well as clinico-pathological data are reported in Table 1. This study was conducted according to the research code of our institutional medical ethical committee on human experimentation and appropriate informed consent was obtained from all individuals included in this study.

Table 1 Clinicopathologic features and MUTYH/APC germline mutations of MUTYH-associated-polyposis and classical/attenuated adenomatous polyposis patients
Full size table

The molecular assessment was performed using 5–10 μm thick, microdissected histological sections. After deparaffinization, DNA was extracted from formalin–fixed, paraffin–embedded tissue sections according to manufacturer’s instructions (QIAamp DNA FFPE Tissue Kit, Qiagen, Hilden, Germany). We used DNA from each patient’s lymphocytes as germline control.

Mutational Analysis and Quantification of KRAS and BRAF Genes

For KRAS gene, exon 2 was amplified and sequenced by using internal primers (F1 5′-aaaaggtactggtggagtatttga-3′, R1 5′-ttgaaacccaaggtacatttca-3′, F2 5′-ttaaccttatgtgtgacatgttctaa-3′, R2 5′-tcatgaaaatggtcagagaaacc-3′). The same nested PCR approach was carried out for the amplification and sequencing of BRAF exon 15 (F1 5′-ttgactctaagaggaaagatgaagt-3′, R1 5′-agaacactgatttttgtgaatactg-3′, F2 5′-tcataatgcttgctctgatagga-3′, R2 5′-ggccaaaaatttaatcagtgga-3′). For both exons, thermal cycling was carried out in a final volume of 50 μl containing 1X Buffer, 1.5 mM MgCl2, 100 mM each dNTP, 10 pmole of primers, 150 ng of DNA extract and 5U of Taq polymerase (AccuPrime Taq DNA Polymerase, Invitrogen, Carlsbad, CA). Briefly, samples were denaturated at 94 °C for 5 min and amplified for 40 cycles as follows: 94 °C for 30 s, 56 °C for 30 s (KRAS) or 52 °C for 30 s (BRAF) and 72 °C for 30 s; final elongation at 72 °C for 7 min. Sequencing reactions were carried out according to manufacturer’s instructions for the BigDye Terminator v3.1 (Applied Biosystems, Foster City, CA) with minor modifications. All samples were directly sequenced forward and reverse on a 3730 ABI DNA Analyzer (Applied Biosystems). The reference sequences for KRAS (NM_004985.3) and BRAF (NM_004333.4) were obtained from www.ncbi.nlm.nih.gov/gene.

The assessment of the percentage of KRAS and BRAF mutant alleles was performed by pyrosequencing analysis as we previously reported.21 PCR reactions were carried out as before but using the following 5′-biotynylated primers: 5′-ggcctgctgaaaatgactg (forward) and 5′-biotin-gctgtatcgtcaaggcactct (reverse) for KRAS (codon 12 and 13); 5′-biotin-cttcataatgcttgctctgatagg-3′ (forward) and 5′-gcatctcagggccaaaaat-3′ (reverse) for BRAF (codon 600). Samples were denaturated at 94 °C for 5 min and amplified for 35 cycles consisting of 94 °C for 30 s, 57 °C for 30 s (KRAS) or 55 °C for 30 s (BRAF), 72 °C for 30 s and a final elongation at 72 °C for 7 min. PCR products were then analyzed according to manufacturer’s instructions using PyroMark Vacuum Prep Workstation (Qiagen). The primed single-stranded DNA templates were subjected to real-time sequencing by using the following primer: KRAS 5′-cttgtggtagttggagct-3′, BRAF 5′-ccactccatcgagatt-3′. Pyrosequencing analysis was carried out using PyroMark Q24 Instrument (Qiagen). Quantification of mutant vs wild-type alleles was calculated by PyroMarkQ24 software (Qiagen).

Mutational Analysis of Mitochondrial DNA

This analysis was carried out by comparing mitochondrial DNA from lymphocytes and adenomas/carcinomas for selected coding sequences of the following mitochondrial genes: MT-CO1, MT-CO2 and MT-TD. All these regions were contiguous and comprised in a 590 bp fragment (nt. 7333–7860), allowing to amplify them by a single pair of primers: 5′-cttcgaagcgaaaagtcctaata-3′ (forward) and 5′-tcgttgacctcgtctgttatgt-3′ (reverse); mitochondrial DNA was amplified with Accuprime Taq DNA Polymerase System (Invitrogen) in a final volume of 50 μl containing 1X Buffer, 1.5 mM MgCl2, 100 mM each dNTP, 10 pmole of primers, 150 ng of DNA extract and 5U of Taq polymerase. Samples were denaturated at 94 °C for 4 min and amplified for 35 cycles as follows: 94 °C for 45 s, 58 °C for 45 s, and 72 °C for 1 min 30 s. Sequencing was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit on a 3730 ABI DNA Analyzer (Applied Biosystems).

Sequences were compared against human mitochondrial DNA sequence, NCBI accession # NC_012920.1, (www.ncbi.nlm.nih.gov) as well as a comprehensive mitochondrial databank, MITOMAP (www.mitomap.org). All mitochondrial DNA-positive samples were resequenced to confirm the results.

Statistical Analysis

The chi-square and Fisher‘s exact tests (when appropriate) were used to examine the significance of the association between KRAS, mitochondrial DNA mutations and the polyposis subgroups. Statistical significance was assumed with P<0.05. All tests were carried out with SPSS software, Statistics v. 17.0 (SPSS Inc., Chicago, IL).

Results

Frequency and Allelic Quantification of KRAS and BRAF Mutations in Adenomas from MUTYH-Associated-Polyposis and Adenomatous Familial Polyposis Patients

We analyzed 96 histologically comparable adenomas, 55 from 11 MUTYH-associated-polyposis and 41 from 13 classical/attenuated familial patients (2 up to 13 adenomas for each patient). Exon 2 activating mutations of KRAS gene were identified in 13/55 (24%) MUTYH-associated-polyposis and in 6/41 (15%) classical/attenuated familial polyposis adenomas, respectively (Table 2). Eight of the 13 mutated MUTYH-associated-polyposis adenomas (62%) exhibited codon 12 c.34G>T transversions, typically associated with oxidative DNA damage; the remaining 5 mutated adenomas showed mutations at codon 13, four transitions (3 c.38G>A and 1 c. 37G>A) and one transversion c.37G>C (Table 2). None of these mutations was found in classical/attenuated familial polyposis adenomas where we detected 6 substitutions at codon 12, namely 3 c.35G>A transitions and 3 c.35G>T transversions (Table 2) (P<0.001). Overall, taking into account only adenomas, KRAS activating mutations were found to affect predominantly MUTYH-associated polyposis patients (64 vs 31% classical/attenuated familial polyposis subjects; P=0.115), clustering in some individuals of both categories (patient M1, M3 and F1 of Table 2).

Table 2 KRAS and BRAF gene mutations in MUTYH-associated-polyposis and classical/attenuated adenomatous polyposis adenomas and colorectal cancers (CRCs)
Full size table

MUTYH-associated-polyposis and classical/attenuated familial polyposis adenomas were also evaluated for the percentage of KRAS mutated allele by pyrosequencing technology: MUTYH-associated-polyposis polyps exhibited a higher mean value compared to classical/attenuated familial polyposis lesions (23 vs 13%; P=0.329) (Table 3a).

Table 3 Percentage of KRAS and BRAF mutated alleles in MUTYH-associated-polyposis and classical/attenuated adenomatous polyposis adenomas (a) and colorectal adenocarcinomas (b)
Full size table

We further tested the adenomas for the presence of BRAF exon 15 mutations. V600E mutation (c.1799A>T transversion) was found in 1/55 MUTYH-associated-polyposis and 1/41 classical/attenuated familial polyposis adenomas (Table 2). The two BRAF mutated adenomas were wild type for KRAS gene and showed comparable level of mutant allele (9 and 10%) (Table 3a). KRAS and BRAF mutations never coexisted in the same lesion, even if the two oncogenes were mutated in different polyps belonging to the same patient (see case M12 in Table 2).

Frequency and Allelic Quantification of KRAS and BRAF Mutations in Carcinomas from MUTYH-Associated-Polyposis and Familial Adenomatous Polyposis

We searched for KRAS exon 2 and BRAF exon 15 mutations in 7 adenocarcinomas, 5 derived from MUTYH-associated-polyposis and 2 from classical/attenuated familial polyposis patients. KRAS exon 2 activating mutations were present in 4/5 MUTYH-associated-polyposis colorectal cancers: all base substitutions were c.34G>T transversions. Three KRAS mutated colorectal cancers also contained BRAFV600E (Table 2). The following quantification of KRAS c.34G>T and BRAF c.1799A>T alleles showed that KRAS mutations were quantitatively more represented compared to BRAF in all three lesions (mean value: 38 vs 14%; P=0.141) (Table 3b and Figure 1a). On the other hand, colorectal cancers from the two classical/attenuated familial polyposis patients showed KRAS c.35G>A mutation but no BRAF genetic alterations (Tables 2 and 3b)

Figure 1
figure 1

Mutational analysis of KRAS, BRAF and MT-CO2 in MUTYH-associated-polyposis colorectal cancers and adenomas (a) sequencing and pyrosequencing evaluation of KRAS c.34G>T transversion and BRAFV600E mutation in colorectal cancer of M1 patient; pyrosequencing detection for KRAS c.34G>T and BRAFV600E mutated alleles is 45 and 16%, respectively. (b) comparison of MT-CO2 sequencing in lymphocyte and adenoma mitochondrial DNAs derived from M8 patient; mitochondrial DNA from adenoma shows non-synonymous coding mutation m. 7742A>C.

Full size image

KRAS Mutations in Adenomas and Colorectal Cancers from MUTYH-Associated-Polyposis Affected Sisters

Our set included adenomas and carcinomas from two pairs of sisters with MUTYH-associated-polyposis: two sisters were homozygous carriers of p.Y179C MUTYH germline mutation (patients M3 and M4 of Table 2) while the two others were p.G396D/c.229insGG MUTYH compound heterozygotes (patients M1 and M2 of Table 2). In both cases, in addition to adenomas, one of the two sisters had also developed a colorectal cancer (patients M1 and M3 of Table 2). As far as p.Y179C homozygous sisters is concerned, we analyzed the colorectal cancer and 6 adenomas from patient M3, and 9 adenomas from patient M4; as far as p.G396D/c.1229insGG mutation carriers, we investigated the colorectal cancer and 11 adenomas from patient M1, and 6 adenomas from patient M2. KRAS c.34G>T transversions were found in 5/6 (83%) adenomas from patient M3 and in 1/11 (9%) adenomas from patient M1 (Table 2). This same transversion was also present in combination with BRAF V600E mutation in colorectal cancer cells of both M1 and M3 subjects. Interestingly, no c.34G>T KRAS mutation was present in the adenomas of M2 and M4 subjects who had not developed carcinomas (Table 2).

Mutational Analysis of Mitochondrial DNA Sequences in MUTYH-Associated-Polyposis and Familial Adenomatous Polyposis Lesions

By sequencing DNA extracted from lymphocytes and colorectal lesions of the same patients, we compared phylogenetically conserved MT-CO1, MT-CO2 and MT-TD. By only taking into account sequence changes absent in mitochondrial DNA from lymphocytes and present in mitochondrial DNA from the adenoma cells of the same patient, we found DNA variants in 18/47 (38%) MUTYH-associated-polyposis adenomas and in 6/29 (21%) classical/attenuated familial polyposis adenomas (Table 4 and Figure 1b); overall, variants were found in 82% (9/11) MUTYH-associated-polyposis vs 38% (5/13) classical/attenuated familial polyposis patients (P=0.040).

Table 4 MT-CO1/MT-CO2 and MT-TD gene mutations in MUTYH-associated-polyposis and classical/attenuated adenomatous polyposis adenomas and colorectal cancers (CRCs)
Full size table

Taking into account both adenomas and colorectal cancers, 6/12 (50%) MUTYH-associated-polyposis subjects were carriers of both KRAS and mitochondrial DNA variants vs 3/13 (23%) classical/attenuated familial polyposis individuals (P=0.163) (Tables 2 and 4).

The majority of the identified mitochondrial DNA variants were G>A transitions, although other rare substitutions were also detected (transition C>T; transversions A>C and C>G), prevalently in MUTYH-associated-polyposis lesions (Table 4). Interestingly, variants affecting MT-CO1/MT-CO2 subunits were only found in MUTYH-associated-polyposis adenomas; with the exception of nt. 7768A>G recurrent base change, most variants caused aminoacid substitutions (Table 4). Variants causing an aminoacid change were also found in 2/5 of MUTYH-associated-polyposis carcinomas (nt. 7763G>A); these variants were detected in both adenomas and carcinomas of the same patients (M1 and M3) (Table 4). Similarly, the variants involving MT-TD were more frequent in MUTYH-associated-polyposis than in familial polyposis adenomas (10 vs 6). Most of the identified changes (15/16) caused G>A transition of nt.7521 (Table 4). MUTYH-associated-polyposis patients with MT-TD variants did not show any MT-CO1/MT-CO2 mutations. Moreover, in MUTYH-associated-polyposis lesions, only MT-CO1/MT-CO2 coding mutations were significantly associated with KRAS mutations (P=0.0085; Table 5).

Table 5 Association between KRAS and mitochondrial DNA mutations in MUTYH-associated-polyposis adenomas and colorectal adenocarcinomas
Full size table

Discussion

The preponderance of c.34G>T transversion in KRAS gene is a well-assessed molecular feature of MUTYH associated carcinogenesis. However, a detailed analysis of the role of this oncogenic mutation throughout the MUTYH-associated-polyposis progression has never been investigated.

By comparing adenomas from MUTYH-associated-polyposis and familial adenomatous polyposis patients, carrying different types of constitutional mutations in MUTYH and APC genes, we showed that KRAS activating mutations differ in frequency and type. Not only mutations affect more frequently adenomas from MUTYH-associated-polyposis than from familial adenomatous polyposis subjects (64 vs 31%), but KRAS mutated alleles are also quantitatively more represented in MUTYH-associated-polyposis adenomas compared to classical/attenuated familial polyposis lesions (23 vs 13%). A limited influence of KRAS mutations in familial adenomatous polyposis tumorigenesis was suggested by Obrador-Hevia and coworkers22 who reported that these mutations affect only 10% of FAP adenomas without achieving the WNT pathway activation. Therefore, while KRAS exerts a marginal role in classical/attenuated familial polyposis progression, it appears to have a key role in MUTYH-driven colorectal tumorigenesis.

As far as the type of KRAS mutation is concerned, we definitely showed that c.34G>T transversions are the most common alterations in MUTYH-associated-polyposis adenomas (62% of the mutated lesions). In addition to this hallmark of DNA oxidation, we also detected KRAS activating mutations at codon 13. Both types of mutations were found in carriers of different MUTYH germline alterations, but neither was detected in classical/attenuated familial polyposis adenomas (P<0.001). Van Puijenbroek and coworkers23 pointed out that, among patients with family history of polyposis and less than 10 adenomas, KRAS c.34G>T pre-screening of formalin-fixed, paraffin-embedded carcinomas may represent a valuable tool to identify MUTYH-associated-polyposis cases. Our results indicate that this approach could be adopted at the level of pre-malignant lesions in order to specifically address germline mutation analysis towards either MUTYH or APC gene.

Presently, the relation between KRAS mutations at codon 13 and MUTYH impairment is unknown. At any rate, the mutated MUTYH-associated-polyposis adenocarcinomas of our cohort showed only c.34G>T transversions (p.G12C); moreover, in adenomas, the frequency of c.34G>T mutation was two times higher than that of mutations at codon 13. Overall, this suggests that KRAS activation through either codon 12 or codon 13 mutations might have a different impact on MUTYH-associated-polyposis carcinogenesis. A variable carcinogenic potential of KRAS mutations was reported by Al-Mulla and co-workers24 who demonstrated a preferential association of p.G12V (c.35G>T) with high-grade tumors and bad prognosis compared to p.G12D (c.35G>A) mutation. More recently, De Roock and coworkers25 showed that colorectal cancer patients with p.G13D-mutated tumors treated with cetuximab had longer overall and progression-free survival.

In our set, MUTYH-associated-polyposis carcinomas differed from familial adenomatous polyposis colorectal cancers not only for the type of KRAS mutations, but also for the concomitant presence of both KRAS and BRAF mutated alleles (in 3/5 of the examined MUTYH-associated-polyposis neoplasias). By also taking into account intratumoural heterogeneity,26 KRAS and BRAF mutations are well demonstrated mutually exclusive events in colorectal cancer.27, 28 Therefore, their coexistence, although very likely in different cellular subclones, might represent a peculiar feature of MUTYH-associated-polyposis carcinogenesis. Boparai and coworkers6 reported that about 50% of MUTYH-associated-polyposis patients, besides adenomas with APC somatic mutation, show hyperplastic polyps and sessile serrated adenomas harbouring KRAS c.34G>T mutations. On the other hand, BRAF mutations are known to identify the subset of colorectal cancers following the serrated neoplastic pathway.29 On this basis, our results support the hypothesis of a MUTYH-associated-polyposis progression with a distinct genetic pattern at the interface between the Vogelstein’s classical model and the serrated carcinogenesis.

We analyzed cytochrome c oxidase subunit I (MT-CO1) cytochrome c oxidase II (MT-CO2), and tRNA aspartic acid (MT-TD) genetic variant by comparing mitochondrial DNA sequences from lymphocytes and colorectal lesions of the same patients; taking into account only adenomas, we found mutations in 82% of MUTYH-associated-polyposis with respect to 38% of classical/attenuated familial polyposis patients (P=0.040). Mostly, non-synonymous substitutions causing aminoacid change in MT-CO1 and MT-CO2 were only found in MUTYH-associated-polyposis lesions, both adenomas and carcinomas. Mutations targeting these genes and leading to aminoacid changes have previously been found in colorectal cancer,30 while data concerning mitochondrial DNA mutations in colorectal adenomas are extremely scanty. Sui and coworkers,20 performing a whole genome analysis of mitochondrial DNA from gastrointestinal precancerous lesions, including sporadic colorectal adenomas, found MT-CO1 among the most frequently mutated sequences.

Cytochrome c oxidase genes encode the enzyme acting as the final acceptor in the electron transport chain and catalyses the reduction of oxygen to water. Besides regulation of respiratory chain, the enzyme is also known to activate a cell death pathway associated with mitochondrial dysfunction. Indeed, apoptosis can occur by changes in mitochondrial integrity initiated by effectors like Ca2+ and reactive oxygen species, leading to the release of cytochrome c. In vitro analysis by Oka and coworkers3 demonstrated that the accumulation of 8-oxoG, due to oxidative stress, causes mitochondrial dysfunction and Ca2+ release, thereby activating calpain; at this context, cell death is triggered by single-strand-breaks that accumulate in the DNA, and is suppressed by knockdown of MUTYH thus indicating that excision of adenine opposite 8-oxoG leads to the accumulation of single-strand-breaks that cause mitochondrial depletion. In this scenario, the functional impairment of MUTYH and the presence of MT-CO1 and MT-CO2 mutations are expected to deeply compromise the control of mitochondrial-mediated apoptosis.

In our cohort, coding mutations targeting MT-CO2 were also found in 2/5 MUTYH-associated-polyposis colorectal cancers, thereby suggesting a contribution of this mitochondrial gene in MUTYH-associated-polyposis progression. If so, MT-CO2 alterations might be regarded as ‘driver mutations’ in MUTYH-associated-polyposis syndrome. Interestingly, MT-CO1/MT-CO2 mutations in MUTYH-associated-polyposis were inversely correlated to the presence of MT-TD alterations, thus suggesting a possible ‘mutational balance’ of mitochondrial DNA mutations in tumor progression.

In classical/attenuated familial polyposis patients, mitochondrial DNA mutations were limited to MT-TD with almost the same mutation frequency compared to MUTYH-associated-polyposis. Mitochondrial tRNA genes have previously been found to be mutation hot spots in mitochondrial DNA and their defects are expected to affect both replication and transcription;31 an altered regulation of these processes appears to be common in both MUTYH-associated-polyposis and classical/attenuated familial polyposis lesions.

In this study MUTYH-associated-polyposis subjects mutated in both KRAS and mitochondrial DNA genes were two times more that of classical/attenuated familial polyposis individuals. Moreover, only MT-CO1/MT-CO2 coding mutations were significantly associated with the presence of KRAS mutations in MUTYH-associated-polyposis lesions (P=0.0085). Some evidences indicate a role of KRAS in activating tumor cell growth by inducing mitochondrial dysfunction and increasing the production of ROS.32, 33 In MUTYH-associated-polyposis adenomas the early occurrence and the high frequency of KRAS mutations can result into the increase of mitochondrial functional mutations; indeed our results show a relationship between KRAS activation and mitochondrial DNA mutations affecting a gene specifically involved in controlling oxidative phosphorylation.

Our results clearly demonstrate, for the first time, that MUTYH-associated-polyposis carcinogenesis can be driven by the early occurrence of specific and statistically associated mutations in both KRAS and mitochondrial genes, MT-CO1/MT-CO2. These findings cast new insights into the existence of a colorectal carcinogenesis in which changes in mitochondrial functions, which are important for the control of oxidative phosphorylation, cooperate with the process of RAS-induced malignant transformation.