|Home | About | Journals | Submit | Contact Us | Français|
LKB1/STK11 germline inactivations are identified in the majority (66–94%) of Peutz–Jeghers syndrome (PJS) patients. Therefore, defects inother genes or so far unidentified ways of LKB1 inactivation may cause PJS. The genes encoding the MARK proteins, homologues of the Par1 polarity protein that associates with Par4/Lkb1, were analyzed in this study because of their link to LKB1 and cell polarity. The genetic defect underlying PJS was determined through analysis of both LKB1 and all four MARK genes. LKB1 point mutations and small deletions were identified in 18 of 23 PJS families using direct sequencing and multiplex ligation-dependent probe amplification analysis identified exon deletions in 3 of 23 families. In total, 91% of the studied families showed LKB1 inactivation. Furthermore, a MARK1, MARK2, MARK3 and MARK4 mutation analysis and an MARK4 quantitative multiplex polymerase chain reaction analysis to identify exon deletions on another eight PJS families without identified LKB1 germline mutation did not identify mutations in the MARK genes. LKB1 defects are the major cause of PJS and genes of the MARK family do not represent alternative PJS genes. Other mechanisms of inactivation of LKB1 may cause PJS in the remaining families.
Peutz–Jeghers syndrome (PJS) is an autosomal dominant disorder characterized by mucocutaneous melanin pigmentation, hamartomatous polyps and an increased cancer risk (1, 2). The discovery of underlying mutations in the tumor suppressor gene LKB1/STK11, has provided further insight into this disorder. However, the precise function of LKB1 remains elusive as is the exact molecular mechanism responsible for the phenotypic characteristics of PJS. We recently hypothesized that loss of the polarity function of LKB1 results in mucosal prolapse, ultimately leading to PJS polyp formation, and tumor growth (3).
Although LKB1 was identified as the PJS gene, germline mutations were found in only 30–70% of patients using conventional mutation analyses (4). LKB1 might, however, be alternatively inactivated and recently exonic deletions have been described, resulting in 66–94% of PJS patients with LKB1 inactivation (5–7). A subset of PJS patients remains with seemingly no LKB1 inactivating mutation and consequently a second PJS gene may exist. Several possible candidates have been studied, including genes encoding LKB1 interacting proteins LIP1, BRG1, STRAD and its co-activator MO25, but to date no second PJS gene has been identified (8–11). In search of a second PJS locus, linkage to chromosome 19q13.4 was found in one Indian PJS family (12) and a 6-day-old patient presenting with a hamartoma with the histology of a PJS polyp had a translocation in the same region (11). Several genes within 0.5 Mb of the breakpoint were sequenced (including BRSK1/KIAA1811), but none was mutated in PJS patients without LKB1 mutation. As the region on chromosome 19q13.4 was implicated in two different PJS families, it may harbor a second PJS gene.
One of the genes located in the 19q13 region is the MARK4 gene. The MARK proteins are part of the family of AMPK-related kinases of which LKB1 is an upstream activator (13, 14). These four microtubule affinity-regulating kinases play a role in microtubule dynamics during polarization of cells (15). MARK2 knockout and heterozygous mice were also described to present with a phenotype of colorectal prolapse (16). Interestingly, PJS polyps histologically resemble mucosal prolapse (3). However, the MARK2 knockout mice also develop characteristics not linked to PJS-like immune system dysfunction, overall proportionate dwarfism and a peculiar hypofertility (16, 17). The MARK genes have also been implicated in tumorigenesis since in two colorectal tumors a mutation and 1 bp insertion were reported in MARK3 (18).
The MARK proteins are the human homologues of Par1, which is, like LKB1/Par4, a member of the par family of polarity proteins. This family is conserved during evolution and the six members of the family regulate epithelial polarity in Drosophila melanogaster, Caenorhabditis elegans and vertebrates, by involvement in cell migration and the establishment of the anterior–posterior axis. Due to their relation to LKB1 and their role in polarity, the MARK genes make interesting PJS candidates.
To further investigate germline alterations related to PJS, we screened for LKB1 defects and for mutations in the MARK1, MARK2, MARK3 and MARK4 genes. LKB1 mutation analysis was performed on 23 PJS patients from different families using sequence analysis and multiplex ligation-dependent probe amplification (MLPA) to detect point mutations and exon deletions. Furthermore, a mutation analysis of the four MARK genes and an exon deletion screen of MARK4 on eight PJS patients from families without LKB1 germline mutations was performed including the Indian PJS family where linkage was found at chromosomal region 19q13.4 (19).
This research was carried out in accordance with the ethical guidelines of the research review committee of our institution. All PJS patients fulfilled the clinical criteria for PJS as described by Tomlinson et al. (20).
PJS patient material was obtained from pathology archives of several Dutch hospitals including the Academic Medical Center (Amsterdam, The Netherlands), Erasmus University (Rotterdam, The Netherlands) and the Free University (Amsterdam, The Netherlands), and the Johns Hopkins Hospital (Baltimore, MD, USA). Use of anonymous or coded leftover material for scientific purposes is part of the standard treatment contract with patients in our hospital (21).
Formalin-fixed paraffin-embedded tissue was available from 23 PJS patients from different families. Also, genomic DNA isolated from blood and cell lines was available from eight other PJS families without LKB1 germline mutation (22). One of these families was the Indian PJS family PJS07 as described by Mehenni et al. (19) kindly provided by Dr S. Antonarakis.
Genomic DNA was isolated from paraffin-embedded hamartomatous tissue. Tissue was deparaffinized and DNA was isolated using the Puregene DNA Isolation kit (Gentra Systems, Minneapolis, MN, USA). DNA concentrations were measured using the PicoGreen Double Stranded DNA Quantitation kit (Molecular Probes, Leiden, The Netherlands).
LKB1 coding exons and exon/intron boundaries were amplified by polymerase chain reaction (PCR) as previously described (23) (GenBank accession numbers: exon 1, AF032984; exons 2–8, AF032985; exon 9, AF032986). All coding exons and exon/intron boundaries of MARK1, MARK2, MARK3 and MARK4 were also amplified by PCR (GenBank accession numbers: NM_018650, NM_017490, NM_002376, NM_031417 respectively). Primer sequences and PCR conditions are available upon request.
PCR products were purified using the Qiagen PCR purification kit (QIAGEN Benelux B.V., Venlo, The Netherlands), and the sequencing reaction was performed using the ABI Big Dye Terminator Cycle Sequencing Kit (Applied Bio-systems, Foster City, CA, USA). Samples were run on an ABI 3100 genetic analyzer and analyzed using SEQUENCE NAVIGATOR and CODONCODE ALIGNER.
Deletion of LKB1 exons was studied using the MLPA kit P101 (MRC-Holland, Amsterdam, The Netherlands). Results were analyzed using the MRC COFFALYSER software (www.MLPA.com). For controls, genomic DNA samples from six normal tissues were used. Results were normalized on all control probes present in the kit and on all six normal tissues. Deletions and duplications were defined as ratios of <.55 and <1.45, respectively and were repeated at least twice.
Deletions of MARK4 were determined by quantitative multiplex PCR of short fluorescent fragments (QMPSF) by a pairwise combination of MARK4 exons and as an internal reference exon 13 of the household gene HMBS. Primer sequences and PCR conditions are available upon request.
Analysis was carried out using an automated ABI 3100 sequencer (Applied Biosystems) with a GENESCAN™ 500 ROX size standard (Applied Biosystems) and the manufacturer’s GENESCAN® 5.1 software. The intensity of the genescan peak for a specific exon for a patient sample was normalized for HMBS in the same reaction. The same was calculated for the normal control. The normalized value of the patient sample was divided by the normalized control sample. Loss of an exon was assumed if the ratio between these two values was less than 0.6.
All coding exons and adjacent exon–intron boundaries of LKB1 were amplified from genomic DNA of all 23 PJS patients in our study group and used for direct sequencing. LKB1 germline mutations were detected in 18 of the 23 (78%) patients; 12-point mutations (of which 2 in intron–exon transition), 2 deletions, 4 insertions of a few basepairs and one silent mutation (Table 1). Of these, 13 mutations have been described previously (19, 22, 24–28). Five novel mutations are reported here: c.551T>C, c.712A>T, c.762delC, c.829–830insGGGCG and c.547C>T leading to two missense (p.Leu184Pro, p.Ile238Phe) mutations, two frameshift mutations (p.Pro245Pro fsX33, p.Asp277Gly fs X12) and one silent mutation (p.L183L) respectively. For some patients, non-affected family members were used as controls. Furthermore, the described mutations were not identified in a control group of over 250 tumor samples (18, 29, 30).
To determine if LKB1 was inactivated in the remaining five PJS patients, an MLPA analysis was performed to identify whole exon deletions. For one patient, MLPA analysis was not possible due to poor quality of the DNA. In total, three patients showed heterozygous exon losses: one patient showed loss of the entire gene, one loss of exon 2 and the last loss of exons 2–7. In total, LKB1 inactivation is observed in 21 of the 23 (91%) patients analyzed.
Similarly, all exons and intron–exon boundaries from the MARK1, MARK2, MARK3 and MARK4 genes were amplified and used for direct sequencing. DNA from eight other PJS patients without LKB1 mutation (22) including the PJS07 family with linkage to 19q14.3 (19) was analyzed. No mutations were found in any of these genes. Several known polymorphisms were found (MARK1: rs3737296, rs3737297; MARK2: rs224174; MARK3: rs2273702, rs2273700, rs2273699, rs4281653, rs1951393, rs10137161, rs1058546; MARK4: rs2240672, rs173179) as well as an alteration in exon 14 of MARK4 in one patient (c.1553C>T, p.518Pro>Leu). This alteration was also detected in one of 50 healthy controls, and, therefore, was considered a (novel) polymorphism.
Since in one specific family the chromosomal region 19q13 might be associated with PJS, we reasoned that if MARK4 located at 19q13 was involved, it could also be inactivated by loss of complete exons as was the case for LKB1 in the three PJS families described above. We, therefore, performed an MARK4 exon deletion analysis based on QMPSF, in which several exons, including an internal control, were co-amplified. However, no MARK4 exon deletions were identified using this method in any of the 8 PJS families without LKB1 germline mutation.
In the present study, LKB1 was inactivated in 21 of the 23 (91%) PJS patients in our study group. This included 78% point mutations and 13% exon deletions or even whole gene deletions. Five novel mutations were identified, two frameshift mutations, two missense and one silent mutation. The silent mutation has been included since it has been shown that silent mutations can be pathogenic (31) and, furthermore, this mutation was not identified in the control group. These results are consistent with previous reports where the total percentage of LKB1 inactivation ranged from 66% to 94% (5–7). Therefore, the question remains whether a second PJS gene exists to explain the percentage of PJS patients without an LKB1 germline mutation. Although affected in a minority of cases, inactivation of such a gene could result in the same phenotype if it affected the same pathways as LKB1. In sporadic colorectal cancer, a majority of the patients have mutations in the primary CRC gene APC. However, in about 10% of cases, APC is unaffected, but its target β-catenin is mutated in the domain that binds to APC (32). This may also be the case for PJS. Presently, however, no mutations in genes other than LKB1 have been discovered.
Due to the close association with LKB1, the MARK proteins may be involved in LKB1 signaling and therefore, be of importance in the etiology of PJS. In this study, no mutations or deletions were identified in the MARK genes in PJS patients without LKB1 germline mutation. This indicates that the MARK genes are unlikely to be second PJS genes. Also, protein expression of the MARK proteins was investigated by performing immunohistochemistry on paraffin material of PJS patients, but, unfortunately, the staining was not specific and provided no further information.
The MARK proteins are members of the AMPK family of kinases, all involved in energy metabolism. Since LKB1 functions upstream of these kinases, a role for the AMPK kinases has been suggested in PJS. AMPK itself is a multi-subunit protein; knockouts of both the α2 and the γ2 subunits have been published but neither have a phenotype comparable with PJS. The phenotype of the α2 knockout is glucose intolerance and that of the γ2 knockout electrocardiographic failure. In humans, a germline AMPK subunit γ2 mutation results in the Wolff–Parkinson–White syndrome (33). Although inactivation of AMPK has significant effects on energy metabolism, these are not similar to PJS symptoms, suggesting that the effect on energy metabolism is not the main cause of PJS symptoms.
If no candidate genes that are associated or related to LKB1 can be found with germline defects in patients without LKB1 mutations or deletions, LKB1 may be the only gene affected in PJS. Presently, germline mutations and exonic losses have been described for 66–94% of PJS patients. Here, we report 9% of the patients without detectable alterations in LKB1. The question remains whether the syndrome in the remaining patients can be genetically explained by the existence of a second PJS gene or that LKB1 might also be inactivated via intronic mutations or deletions that have not been studied thus far.
The authors thank Stylianos E Antonarakis (Division of Medical Genetics, Centre Médical Universitaire, Geneva, Switzerland) for providing us with DNA from PJS family PJS07. The study was supported by the Netherlands Digestive Disease Foundation (WS01-03), the John G Rangos, Sr. Charitable Fund, the Clayton Fund and NIH grant P50 CA 62924-10.