Abstract
Germline mutations in p27kip1 are associated with increased susceptibility to multiple endocrine neoplasias (MEN) both in rats and humans; however, the potential role of common polymorphisms of this gene in endocrine tumor susceptibility and tumorigenesis remains mostly unrecognized. To assess the risk associated with polymorphism rs2066827 (p27-V109G), we genotyped a large cohort of Brazilian patients with sporadic endocrine tumors (pituitary adenomas, n=252; pheochromocytomas, n=125; medullary thyroid carcinoma, n=51; and parathyroid adenomas, n=19) and 885 population-matched healthy controls and determined the odds ratios and 95% CIs. Significant associations were found for the group of patients with pituitary adenomas (P=0.01), particularly for those with ACTH-secreting pituitary adenomas (P=0.005). In contrast, no association was found with GH-secreting pituitary tumors alone or with the sporadic counterpart of MEN2-component neoplasias. Our in vitro analyses revealed increased colony formation and cell growth rate for an AtT20 corticotropin mouse cell line overexpressing the p27-V109G variant compared with cells transfected with the WT p27. However, the genotypic effects in genetic and in vitro approaches were divergent. In accordance with our genetic data showing specificity for ACTH-secreting pituitary tissues, the overexpression of p27-V109G in a GH3 somatotropin rat cell line resulted in no difference compared with the WT. Pituitary tumors are one of the major clinical components of syndromes associated with the p27 pathogenic mutations MENX and MEN4. Our genetic and in vitro data indicate that the common polymorphism rs2066827 may play a role in corticotropinoma susceptibility and tumorigenesis through a molecular mechanism not fully understood thus far.
Introduction
In the last two decades, genetic studies of endocrine tumor syndromes have identified a number of susceptibility genes including several tumor suppressors (multiple endocrine neoplasia type 1 (MEN1), VHL, NF1, SDHA, SDHB, SDHC, SDHD, SDHAF2, PRKAR1A, AIP, TMEM127, MAX, p15INK4B (CDKN2A), p18INK4C (CDKN2C), p21CIP1 (CDKN1A), and p27Kip1 (CDKN1B)) and two protooncogenes (RET and HIF2A (EPAS1)) (Pellegata et al. 2006, Agarwal et al. 2009, Zhuang et al. 2012, Toledo et al. 2013 and Supplementary refs 1–13 see section on supplementary data given at the end of this article). The genes predisposing to endocrine tumors encode transcription factors or molecules that are involved in a variety of processes such as hypoxia, cell-cycle inhibition, the citric acid cycle and the electron transfer chain, cAMP and Ras pathways, and in both mTOR and myc and glial cell line-derived neurotrophic tyrosine kinase signaling. Among the recently described tumor susceptibility genes for endocrine tumors is the CDKN1B gene, which codes for the cyclin kinase inhibitor p27Kip1 (p27). p27 inhibits cyclin and cyclin-dependent kinase (Cdk) complexes which are essential for cell-cycle progression from G1 to S phases. p27 is mutated in a MEN syndrome both in rats and human patients (Pellegata et al. 2006, Marinoni & Pellegata 2010). The germline mutations in CDKN1B so far identified in patients (12 in total) affect the stability of p27 protein, its localization, its ability to interact with partner proteins, or the efficiency of transcription of the mutated CDKN1B allele (Pellegata 2012). Usually, p27 is not somatically mutated, but its expression is reduced or lost in more than 50% of all human tumors, including endocrine tumors (Chu et al. 2008). In addition to the pathogenic mutations, CDKN1B harbors single-nucleotide polymorphisms (SNPs) within or in the vicinity of the coding sequence, several with a low allelic frequency (<5%) and three of which are potentially functional: −838C>A (rs36228499), −79C>T (rs34330), and 326T>G (V109G, rs2066827). Indeed, patients with the variant AA genotype at position −823 of CDKN1B present low risk of restenosis following coronary stenting, and the variant −838A allele is associated with augmented gene transcription (González et al. 2004), whereas the −79C>T allelic variant is related to reduced transcription in in vitro studies (Landa et al. 2010). Furthermore, the coding 326T>G (V109G) allelic variant has been associated with cancer risk and progression of different tumors including prostate cancer (Kibel et al. 2003), oral squamous cell carcinoma (Li et al. 2004), invasive epithelial ovarian cancer (Gayther et al. 2007), high-grade breast tumors (Tigli et al. 2005), and pancreatic carcinoma (Chen et al. 2010). A recent study has shown that the variant G allele at position 326 protects against thyroid cancer development (Pasquali et al. 2011), so that the role of the V109G polymorphism in increasing/decreasing tumor risk remains to be determined.
In the present study, we investigated the potential influence of the p27 rs2066827 (V109G) polymorphism on the risk of developing endocrine tumors in a large cohort of 447 Brazilian patients presenting the following sporadic endocrine tumors: pituitary adenomas (n=252), pheochromocytomas (n=125), medullary thyroid carcinoma (MTC; n=51), and parathyroid adenomas/hyperparathyroidism (HPT; n=19). The results were compared with those for 885 healthy patients, representative of Brazilian controls. In vitro analyses were performed to further assess the role of p27-V109G in pituitary adenoma cells.
Subjects and methods
Study population
All samples were collected in research centers in the Sao Paulo region of Brazil. Patients and controls were derived from matched population ancestry, composed of majority White and White-Latins, and also Asians, African-Americans, and individuals with mixed (undetermined) ancestry.
Demographic information
The patients were treated in three different hospitals (Hospital das Clinicas, Hospital Brigadeiro, and Hospital Santa Casa) all located within Sao Paulo city. The controls were collected from two different centers: one in the same hospital where the patients were treated (Hospital das Clinicas) and the other one located in the Basic Science Building of the University of Sao Paulo (IB-USP).
Characterization of the control group
The samples were obtained from two DNA banks of healthy individuals at the Medical School (FM-USP, Department of Oncology) and at the Biosciences Institute (IB-USP, Human Genome Research Center) at the University of Sao Paulo.
The control group comprised 885 tumor-free adult/elderly subjects distributed as 54% females, 46% males, with a mean age of 65.2 years (677 of them were 30 years old or older). In order to prevent the analysis of ‘controls’ that had no tumors at a young age but who could eventually develop them later, we collected samples of people still healthy at the average age of 65 years. Thus, the cohort of tumor-free controls is 25 years older than the group of patients.
To exclude from our analysis a possible effect of sex hormones, the sex distribution of the patients (59% females and 41% males) and controls (54% females and 46% of nine males) was similar.
We further characterized our study population with regard to the ethnicity and genotype/allele frequencies comparing with public databases and previous genetic studies of rs2066827. The frequencies of White/White Latinos (79%) and African/Mulatos (10%) in our control group were very similar to those assessed by The Brazilian Institute of Geography and Statistics (IBGE, Instituto Brasileiro de Geografia e Estatística), which reported frequencies of 78 and 13% of White/White Latinos and African/Mulatos, respectively, in the Sao Paulo area in the 2000 Census (www.ibge.gov.br). Allele frequencies found in our population study were compared with those reported in eight papers, which genotyped rs2066827 from controls of Caucasian and Asian ethnicity and from the USA, Canada, Malaysia, and China, as well as from four populations analyzed by the 1000 Genomes Project (CAUC1, AFR1, HISP1, and PAC1). As expected, the frequencies of the Brazilian population (G, 36% and T, 64%) were more similar to those for Caucasians (G, 24% and T, 76%) and Hispanics (G, 30% and T, 70%), than to those for Asians (G, 12% and T, 88%) and Africans (G, 36% and T, 34%).
Patients
The study analyzed 447 patients from matched population ethnicity that were clinically diagnosed with several types of sporadic endocrine tumors, including two different types of pituitary adenomas (n=252): GH-secreting/acromegaly (n=161) and adrenocorticotropic hormone (ACTH)-secreting/Cushing's disease (n=91); parathyroid adenomas/HPT (n=19); MTC (n=51); and adrenomedullary tumors/pheochromocytomas (n=125). Gender distribution was of 59% females and 41% males and the mean age at diagnosis was 40.7 years. They were diagnosed and treated at the Hospital das Clinicas of University of Sao Paulo Medical School (Endocrine Genetics Unit, Neuroendocrinology Unit, and Adrenal Unit), as well as at the Endocrine Divisions of the Hospital Brigadeiro (Sao Paulo) and Hospital Santa Casa (Sao Paulo).
The diagnosis of endocrine tumors was performed by standardized clinical, biochemical, and imaging examinations (Toledo et al. 2009, 2010a, Lourenco et al. 2010). Other possible endocrine tumors were ruled out in each patient. After surgery, tumors were confirmed through pathological criteria and immunostaining for endocrine-specific antibodies. We should mention that all cases with MTC included in the study were examined for the presence of mutations in the hot-spot exons of the RET proto-oncogene (exons 10, 11, and 13–16), and none were found (Santos et al. 2006, Toledo et al. 2010b). TMEM127 and RET mutations were ruled out in the 51 cases with pheochromocytomas as previously described (Yao et al. 2010). The previously described cases from our MEN1 cohort (Toledo et al. 2010a) were not included in this study.
Written informed consents were obtained from all subjects in accordance with Institutional Review Board-approved protocols from each center. The protocols from the University of Sao Paulo are numbered as 0425/08, 0549/09, 1231/09, 0050/10, and 0031/10.
SNP genotyping
The genomic DNA of each subject was isolated from peripheral blood using standard salting-out protocol or commercially available DNA extraction kits (DNeasy Blood & Tissue Kit, Qiagen). For the initial 140 DNA samples, the SNP 326T>G (V109G) was genotyped by direct sequencing. PCRs were performed using primers previously described (Pellegata et al. 2006) and both DNA strands were sequenced from purified PCR products using Big Dye Terminator v3.1 (Life Technologies) and an automated sequencer (ABI Prism 3130xl DNA Analyzer, Life Technologies).
The presence of the polymorphic G allele creates a restriction site for the enzyme BglI (5′…GCCNNNN^NGGC…3′). Thus, we genotyped the rest of the DNA samples using a PCR–restriction fragment length polymorphism approach as previously described (Kibel et al. 2003) after having verified that this method had 100% accuracy compared with sequencing (Supplementary Fig. 1, see section on supplementary data given at the end of this article). Amplicons of 762 bp containing codon 109 were generated by using a PTC-200 DNA thermocycler (Peltier thermal cycler; MJ Research, Inc., Waltham, MA, USA) with 10 ng of gDNA as template and the following primers: p27-V109G_F: GTCGGGGTCTGTGTCTTTTG and p27-V109G_R: GCCAGGTAGCACTGAACACC. The digestion step was carried out at 37 °C for 3 h with the enzyme BglI (New England BioLabs, Beverly, MA, USA). After digestion, the enzyme was inactivated at 85 °C for 20 min and the genotypes were identified according to the bands visualized on 3% agarose gels (Supplementary Fig. 1).
Statistical analyses
Hardy–Weinberg equilibrium of the genetic variants at codon 109 in CDKN1B (p.V109G) was assessed using χ2 statistics, and the best fitting model was determined according to the P values using parsimony. Assessment of tumor risk was performed through comparison of genotype frequencies between cases and controls using χ2 statistics and odds ratios (ORs) with 95% CIs using logistic regression models. A modified Bonferroni's-corrected nominal threshold of P=0.05/n* was used to correct for multiple hypothesis testing, of which n* is the number of independent comparisons (ACTHoma, GHoma, pheochromocytoma, and MTC) considering G and T as the risk allele. The variants, ‘all sporadic endocrine tumors’ and ‘sporadic counterpart of MEN1/2-component neoplasias’; were not considered independent and were not included in the threshold correction. Age and sex adjustments were performed in the statistical analyses.
Expression vectors, cell lines, and transfections
The p27V109G mutations were introduced by site-directed mutagenesis (Quikchange II Site-Directed Mutagenesis Kit; Stratagene, Waldbronn, Germany) in the WT human CDKN1B cDNA cloned in a pYFP backbone as described previously (Pellegata et al. 2006). AtT-20/D-16v (kindly provided by Dr M C Zatelli) and HeLa cells (LGC Standards, Wesel, Germany) were maintained in DMEM, respectively, supplemented with 10% FCS, 20 mM l-glutamine, 100 units/ml penicillin G sodium, and 100 μg/ml streptomycin sulfate. GH3 cells (ATCC, Manassas, VA, USA) were grown in F12 medium supplemented with 15% horse serum, 2.5% FCS, 20 mM l-glutamine, 100 units/ml penicillin G sodium, and 100 μg/ml streptomycin. Transient transfection was carried out as described previously (Pellegata et al. 2006).
Protein extraction and western blotting
To extract total proteins, the cells were collected, washed twice in PBS, and lysed in RIPA buffer essentially as previously reported (Pellegata et al. 2006). Protein concentration was assessed by BCA assay (Pierce Chemical Co., Rockford, IL, USA). Total extracts were subjected to PAGE using Bis–Tris 4–12% NuPAGE gels, blotted, and probed with the following antibodies: monoclonal, against p27 (BD Biosciences, Franklin Lakes, NJ, USA); and monoclonal anti-α-tubulin (Sigma). Immunoreactive proteins were visualized using West Pico chemiluminescent substrates (Pierce Chemical Co.). The bands that we obtained by western blotting were quantified using the Molecular Imager ChemiDoct XRS.
Clonogenic assay and growth curve
To examine their clonogenic activity, AtT-20/D-16v were plated in six-well plates (500 000/well) and transfected with p27-wt, p27-V109G, or GFP empty vector using Fugene-6 (Roche). Subsequently, the cells were trypsinized, diluted 1:6, and replated in 10 cm plates. Twenty-four hours later, the cells were selected for 6 weeks by adding 300 μg/ml to the culture medium and stained with 0.3% crystal violet in 30% ethanol. Colonies containing a minimum of 50 cells were scored.
Clones derived from AtT-20/D-16v cells stably expressing exogenous p27-wt, p27-V109G, or GFP were plated in duplicate (100 000/well) in 12-well plates in full medium plus 300 μg/ml G418. The number of cells was counted every 2 days over a 7-day period using a cell counter.
Immunofluorescence
Immunofluorescence was carried out on AtT20 cells grown on coverslips and transfected with p27-wt, p27-V109G, or GFP vector. Twenty-four hours later, the transfected cells were fixed in 2% paraformaldehyde in PBS for 30 min at room temperature. The cell nuclei were stained with 1 μg/ml Hoechst for 5 min at room temperature and mounted on glass slides. Images were generated using a Zeiss Axiovert 200 epifluorescence microscope including an Apotome unit (Zeiss, Jena, Germany) using the YFP and the DAPI channel and processing was carried out using Zeiss Computer Software (AIM 3.2).
Results
p27-V109G in controls
Genotype frequencies of the SNP rs2066827 were assessed in the control group of 885 representative healthy Brazilian individuals and were distributed as follows: GG=116, 13.1%; GT=406, 45.9%; and TT=363, 41.0%. The allele frequencies (G=0.36 and T=0.64) of our cohort of controls were more similar to those reported in the SNP databases and in previous genetic studies for Hispanic (G=0.30 and T=0.70) and Caucasian (G=0.24 and T=0.76) populations than those for Asians (G=0.12 and T=0.88) or Africans (G=0.72 and T=0.28) (Table 1). No significant deviation from the Hardy–Weinberg equilibrium was observed for the two alleles.
Allele and genotype distribution of the rs2066827 V109G variant of p27 in controls and patients diagnosed with sporadic endocrine tumors investigated
| Total | G | T | GG | GT | TT | |
|---|---|---|---|---|---|---|
| Sporadic counterpart of MEN1-component neoplasias | 271 | 168 | 349 | 30 | 108 | 133 |
| sporadic counterpart of MEN2-component neoplasias | 195 | 143 | 275 | 21 | 101 | 73 |
| Sporadic pituitary adenomas | 252 | 155 | 323 | 28 | 99 | 125 |
| Sporadic ACTHomas | 91 | 50 | 111 | 10 | 30 | 51 |
| Sporadic GHomas | 161 | 105 | 212 | 18 | 69 | 74 |
| Sporadic HPT | 19 | 13 | 26 | 2 | 9 | 8 |
| Sporadic MTCs | 51 | 42 | 69 | 8 | 26 | 17 |
| Sporadic pheochromocytomas | 125 | 88 | 180 | 11 | 66 | 48 |
| All tumors | 447 | 298 | 598 | 49 | 200 | 198 |
| Healthy controls | 885 | 638 | 1132 | 116 | 406 | 363 |
Sporadic counterpart of multiple endocrine neoplasia type 1 (MEN1)-component neoplasias – types of sporadic endocrine tumors that are included in the clinical phenotype of the MEN1 syndrome as hyperparathyroidism and pituitary adenomas; sporadic counterpart of MEN2-component neoplasias – types of sporadic endocrine tumors that are included in the clinical phenotype of the MEN2 syndrome medullary thyroid carcinoma (MTC) and pheochromocytomas; ACTHomas, ACTH-secreting pituitary adenomas; GHomas, GH-secreting pituitary adenomas, also know as somatotropinomas; HPT, parathyroid adenomas/hyperparathyroidism; G and T represent the alleles that observed in the V109G p27Kip1 SNP; and GG, GT, and TT are the observed genotypes.
Association of p27-V109G and sporadic counterpart of MEN1-component neoplasia
Genotyping analysis of the overall group including 447 patients with endocrine tumors provided the following genotypic frequencies: GG=49, 11%; GT=200, 44.7%; and TT=198, 44.3%. We observed a significant overrepresentation of the T allele (valine) in the patients with sporadic counterpart of MEN1-component neoplasias (n=271, parathyroid neoplasias and pituitary adenomas), compared with the 885 controls (OR=1.25; 95% CI=1.02–1.54; P=0.03). Accordingly, the presence of the polymorphic allele G (GT or GG genotypes) was significantly associated with decreased susceptibility to sporadic counterpart of MEN1-component neoplasias (protection), fitting both recessive (OR=0.72; 95% CI=0.55–0.94; P=0.019) as well as dominant (OR=0.72; 95% CI=0.54–0.97; P=0.03) models (Table 2).
Endocrine tumor susceptibility and rs2066827 V109G variant of p27
| Group | n | Allele frequency (G×T) | Heterozygous (GG×GT) | Homozygous (GG×TT) | Allele positivity ((GG+GT)×TT) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | OR | 95% CI | P | ||
| Controls | 885 | 1.000 | 1.000 | 1.000 | 1.000 | ||||||||
| All sporadic endocrine tumorsa | 455 | 1.127 | 0.951–1.335 | 0.16622 | 1.166 | 0.802–1.696 | 0.42069 | 1.291 | 0.886–1.881 | 0.18222 | 1.225 | 0.859–1.748 | 0.26174 |
| 0.887 | 0.749–1.051 | 0.16622 | 0.903 | 0.709–1.151 | 0.40950 | 0.774 | 0.532–1.128 | 0.18222 | 0.875 | 0.695–1.100 | 0.25252 | ||
| Sporadic counterpart of MEN1-component neoplasiasb | 195 | 1.255 | 1.021–1.542 | 0.03091 | 1.029 | 0.653–1.620 | 0.90322 | 1.417 | 0.905–2.217 | 0.12622 | 1.212 | 0.791–1.857 | 0.37704 |
| 0.797 | 0.648–0.980 | 0.03091 | 0.726 | 0.543–0.971 | 0.03054 | 0.706 | 0.451–1.105 | 0.12622 | 0.722 | 0.549–0.948 | 0.01899 | ||
| Sporadic counterpart of MEN2-component neoplasiasc | 269 | 0.973 | 0.775–1.222 | 0.81713 | 1.374 | 0.822–2.296 | 0.22354 | 1.111 | 0.655–1.884 | 0.69651 | 1.250 | 0.763–2.046 | 0.37450 |
| 1.027 | 0.818–1.290 | 0.81713 | 1.237 | 0.887–1.725 | 0.20972 | 0.900 | 0.531–1.527 | 0.69651 | 1.162 | 0.844–1.600 | 0.35620 | ||
| Sporadic pituitary adenomasd | 250 | 1.269 | 1.026–1.570 | 0.02787 | 1.010 | 0.633–1.612 | 0.96605 | 1.427 | 0.901–2.260 | 0.12878 | 1.207 | 0.778–1.871 | 0.40055 |
| 0.788 | 0.637–0.975 | 0.02787 | 0.708 | 0.525–0.955 | 0.02348 | 0.701 | 0.442–1.110 | 0.12878 | 0.707 | 0.533–0.936 | 0.01512 | ||
| Sporadic ACTHomas | 91 | 1.488 | 1.059–2.090 | 0.02115 | 0.857 | 0.407–1.805 | 0.68476 | 1.630 | 0.802–3.313 | 0.17362 | 1.222 | 0.616–2.425 | 0.56605 |
| 0.672 | 0.479–0.944 | 0.02115 | 0.526 | 0.338–0.844 | 0.00696 | 0.614 | 0.302–1.247 | 0.17362 | 0.545 | 0.353–0.843 | 0.00575* | ||
| Sporadic GHomas | 159 | 1.165 | 0.905–1.499 | 0.23592 | 1.095 | 0.627–1.914 | 0.74937 | 1.314 | 0.754–2.291 | 0.33489 | 1.198 | 0.707–2.031 | 0.50093 |
| 0.859 | 0.667–1.105 | 0.23592 | 0.834 | 0.583–1.192 | 0.31788 | 0.761 | 0.437–1.327 | 0.33489 | 0.818 | 0.583–1.146 | 0.24184 | ||
| Sporadic MTCs | 51 | 0.805 | 0.536–1.208 | 0.29474 | 0.929 | 0.409–2.106 | 0.85918 | 0.679 | 0.286–1.614 | 0.37841 | 0.811 | 0.372–1.768 | 0.59732 |
| 1.242 | 0.827–1.100 | 0.29474 | 1.367 | 0.730–2.561 | 0.32667 | 1.473 | 0.619–3.501 | 0.37841 | 1.391 | 0.765–2.528 | 0.27726 | ||
| Sporadic pheochromocytomas | 125 | 1.038 | 0.786–1.369 | 0.79432 | 1.714 | 0.876–3.353 | 0.11171 | 1.394 | 0.701–2.774 | 0.34152 | 1.563 | 0.817–2.991 | 0.17395 |
| 0.964 | 0.731–1.272 | 0.79432 | 1.229 | 0.826–1.830 | 0.30842 | 0.717 | 0.361–1.426 | 0.34152 | 1.116 | 0.759–1.639 | 0.57718 | ||
Significant associations were initially observed in the sporadic MEN1-related, sporadic pituitary adenoma and sporadic ACTH-secreting pituitary adenomas (corticotropinomas). After Bonferroni's correction of multiple testing, association with corticotropinomas remained significant, indicating a robust finding. For each tumor type, the upper row represents the results for the T allele and the lower row the results for the G allele. P values <0.05 are shown in bold. *Values that remained significant after use of a modified Bonferroni's-corrected nominal threshold of P=0.05/n* to correct for multiple hypothesis testing where n* is the number of the independent comparisons (ACTHoma, GHoma, pheochromocytoma, and MTC), and taking account G and T as the risk allele (corrected threshold of P=0.05/8=0.00625). Age and sex adjustments were performed in the statistical analysis. OR, odds ratio; MTC, medullary thyroid carcinoma.
All sporadic counterpart of MEN1- and MEN2-component neoplasias.
Sporadic counterpart of MEN1-component neoplasias, including hyperparathyroidism and pituitary adenomas.
Sporadic counterpart of MEN2-component neoplasias, including hyperparathyroidism, MTC, and pheochromocytomas.
ACTH- and GH-secreting pituitary adenomas.
Absence of association of p27-V109G and sporadic counterpart of MEN2-component neoplasia
No significant association was observed between the p27-V109G polymorphism and the group of sporadic tumors belonging to the MEN2 spectrum (MTC and pheochromocytoma; P=0.35), as well as the groups containing exclusively MTC (P=0.27) or pheochromocytoma (P=0.57). Also, no significant differences were found between the genotypic and allelic frequencies of the controls and the group of patients with sporadic counterpart of MEN1- and MEN2-component neoplasias (P=0.25; Table 2). Primary HPT was initially included in both MEN1 and MEN2 groups, but its exclusion did not change the results.
Absence of association of p27 rs36228499 and sporadic tumors
In addition to SNP V109G, we analyzed in our controls and patient cohorts the allelic frequencies of another SNP located in the noncoding region of p27 (rs36228499), which has been studied previously and shown not to be associated with endocrine/thyroid tumors (Landa et al. 2010), and, according to the HapMap, is not in linkage disequilibrium with the SNP V109G (rs2066827). The genotyping was carried out by PCR and Sanger sequencing (primers −838 F: TGGCCTCGGAGAAATTAAAA and −838 R: TTAAGGCTGAGCGAACCATT) and the frequencies observed in our controls (G=0.51 and T=0.49) and patients (G=0.54 and T=0.46) were similar between each other, and both were more similar to the frequencies reported for the European/HapMap-CEU population (G=0.56 and T=0.44) than to those found in African/HapMap-YRI populations (G=0.75 and T=0.25).
The role of p27-V109G specifically in ACTH–pituitary tumors
Protection by the G allele (GT and GG genotypes) was seen overall in the sporadic pituitary adenoma group (OR=0.70; 95% CI=0.53–0.93; P=0.01) and more especially in the ACTH-secreting pituitary adenoma group (OR=0.54; 95% CI=0.35–0.84; P=0.005). The observed association between rs2066827 and corticotropinomas was still significant after correction for multiple testing using the conservative Bonferroni's method, indicating that this finding is robust. In contrast, the presence of the G allele was not associated with protection from acromegaly (P=0.24).
To determine whether the T or G allelic variants at codon 109 of p27 may be associated with specific functions in ACTH-secreting cells, we performed in vitro analyses comparing the effects of p27-wt (TT) or p27-V109G (GG) overexpression in a mouse corticotropinoma cell line (AtT20). Following transfection of p27-wt and p27-V109G in these cells, we observed that the p27-V109G variant protein localizes to the nucleus, as the WT protein does (Supplementary Fig. 1). We then assessed the effects of both variants on cell growth by performing colony formation assays. The results were compared with the rat somatotroph adenoma-derived GH3 cell line. We found that AtT20 cells transfected with p27-V109G show a 50% increase in the number of colonies when compared with cells either transfected with p27-wt or with empty vector (Fig. 1). In contrast, both variant proteins show similar growth suppression of GH3 cells (Fig. 1 and Supplementary Figs 2 and 3, see section on supplementary data given at the end of this article). This phenotype is in agreement with the increase in cell proliferation observed in AtT20-derived clones stably expressing exogenous p27-V109G when compared with clones expressing p27-wt or transfected with empty vector (data not shown). Thus, in vitro, the two p27 variant proteins show a different behavior in AtT20 corticotrope adenoma cells when compared with GH3 somatotrope adenoma cells.
The effect of overexpression of both variants, p27-wt (TT) and p27-V109G (GG), on cell growth in a mouse corticotropinoma cell line (AtT20) and the rat somatotroph adenoma-derived GH3 cell line (A). AtT20 cells transfected with p27-V109G show a 50% increase in the number of colonies when compared with either cells transfected with p27-wt or with empty vector, while both variant proteins show similar growth suppression of GH3 cells (B).
Citation: Endocrine-Related Cancer 21, 3; 10.1530/ERC-13-0486
Discussion
In the last few years, mutations in the CDKN1B p27 gene have been identified as the cause of a novel syndrome (named MEN4) that is characterized by multiple endocrine tumors with a spectrum similar to the MEN1 syndrome (Pellegata et al. 2006, Georgitsi et al. 2007, Agarwal et al. 2009, Marinoni & Pellegata 2010, Pellegata 2012). While this article was under revision, somatic mutations and deletions of CDKN1B were reported in a subset of neuroendocrine tumors of the small intestine (Francis et al. 2013). In addition to these rare pathogenic mutations, CDKN1B contains numerous SNPs, including a frequent polymorphism (rs2066827, V109G), which has been found to be associated with the risk of or protection from cancer, depending on the tumor type. Specifically, the V109G polymorphism has been shown to have a protective effect on overall survival of patients with sporadic pancreatic carcinoma (Chen et al. 2009), while it is associated with increased risk of progression and poor prognosis for prostate and breast cancer patients (Kibel et al. 2003, Tigli et al. 2005).
In this study, we analyzed the role of the V109G genetic variant in CDKN1B in a case–control study, including 447 Brazilian patients with different types of sporadic endocrine tumors and 885 adult healthy representative Brazilian controls. We found an association between rs2066827 and susceptibility to tumors belonging to the MEN1 tumor spectrum (i.e. pituitary adenomas), while no role in the risk of developing tumors belonging to the MEN2 tumor spectrum (i.e. MTC and pheochromocytoma) could be demonstrated (Table 2). We are aware that the ethnic intermix of the Brazilian population can be an important issue for case–control studies like ours, so we addressed this point by further characterizing our cohorts. According to the Brazilian government database (IBGE) for the Sao Paulo area, where all the samples were collected and the study took place, our cohort is composed for the most part of whites and presented no sampling bias. In accordance, the allele frequencies we found for this genetic variant and for a second SNP in the noncoding sequence of CDKN1B (rs36228499) were similar to those reported in public databases and in previous genetic studies for Caucasians and Hispanics.
Positive association with rs2066827 was found in the pituitary adenoma group, including both somatotropinomas and corticotropinomas (P=0.01); however, more specific association was observed with the corticotropinomas (P=0.005) (Table 2). The association between rs2066827 and corticotropinomas remained significant after the calculation of the adjusted P value cut-off using the conservative Bonferroni's method as a correction for multiple testing (P=0.00625). In contrast, rs2066827 was not associated either with sporadic counterpart of MEN2-component neoplasias or with MTC or pheochromocytoma groups (P>0.05 in all models tested). To our knowledge, this is the first study that investigated p27 polymorphisms in pheochromocytoma, while two independent reports that were recently published analyzed the role of p27 variants in European MTC cohorts (Landa et al. 2010, Pasquali et al. 2011). In accordance to our results for Brazilian MTC patients, no association of p27-V109G with increased risk of sporadic MTC was observed in a large cohort of Spanish MTC cases, while a borderline significance was observed in the analysis of tumor risk of Italian MTC patients (Landa et al. 2010, Pasquali et al. 2011).
We could demonstrate that rs2066827 associates with a specific molecular phenotype in corticotrope cells but not in somatotrope cells. Indeed, the ectopic overexpression of p27-V109G in AtT20 cells, as opposed to p27-wt, does not inhibit cell growth and actually seems to promote cell proliferation. In contrast, no differences in the cell growth characteristics of p27-wt and p27-V109G were observed in somatotroph cells. As the expression of p27-V109G is associated with increased cell growth and proliferation in AtT20 cells, it is not easy to reconcile these findings with the results of the association studies. The fact that AtT20 cells express high endogenous levels of p27 and still show sustained proliferation may indicate that this experimental model, the only one available for corticotrope adenomas, might not be ideal for reproducing the situation of the expression of p27 variants in vivo in individuals.
Patients with germline pathogenic mutations in p27 (MEN4 syndrome) may develop ACTH-secreting adenoma, although rarely (Lee & Pellegata 2013). Previous studies have shown that p27 is consistently and frequently downregulated in corticotropinomas, but not in other types of pituitary adenomas (Dahia et al. 1998). Moreover, a crucial role for p27 in the regulation of the proliferation of pituitary proopiomelanocortin (POMC) cells producing ACTH has also been demonstrated in animal studies, and p27-deficient mice develop POMC tumors at high frequencies (Fero et al. 1996).
The exact molecular mechanism involved in p27-V109G change is still not clear. While the structure of the N- and C-terminal domains of p27 has been resolved and the interactions with key partners such as cyclins and CDKs have been mapped, the structure of the central domain of the protein (amino acids 97–151) is still unknown. The amino acid valine 109 is located in the central portion of the protein, within a domain that mediates the binding of p27–p38Jab1 (Tomoda et al. 1999). The interaction of the two proteins mediates the nuclear export of p27 and its subsequent degradation. Interestingly, the substitution of the aspartic acid at position 108 by glycine impairs the formation of this complex (Tomoda et al. 2002) and it has been speculated that this variant might interfere in the interaction with p38Jab1, and thereby increase the stability of p27 in the nucleus (Schöndorf et al. 2004). Experimental evidence is nevertheless required to confirm this hypothesis.
In addition to our finding of a potential role of rs2066827 in corticotropinomas, another p27 polymorphic variant, −79C>T (rs34330), was specifically correlated with a high risk of developing the follicular variant of papillary thyroid carcinoma (Landa et al. 2010), indicating that reduced CDKN1B gene transcription could be involved in the molecular mechanism that mediates the pathogenic effects of this variant.
There are conflicting findings in the literature regarding the risk/protection associated with rs2066827 T/G alleles. A recent meta-analysis has evaluated the association data of eight studies encompassing 3799 controls and 3591 patients with non-endocrine tumors (oral squamous cell, prostate, breast cancer, and pancreatic cancer) and found no correlation between the rs2066827 variant and the overall cancer risk in the general population (Wei et al. 2012). Since p27 is a tumor susceptibility gene for multiple endocrine tumors in both humans and rats, we decided to investigate specifically endocrine tumors. Based on our results, it seems worthwhile to assess whether the V109G genetic variant may modify individual susceptibility to additional endocrine tumors, i.e. entero-pancreatic tumors, prolactinomas, nonfunctioning pituitary adenomas, parathyroid carcinomas, or adrenocortical carcinomas, and to investigate cohorts of patients from different populations.
Supplementary data
This is linked to the online version of the paper at http://dx.doi.org/10.1530/ERC-13-0486.
Declaration of interest
M D Bronstein declares an association with the following companies: Ipsen, Novartis, and Pfizer (consultant, speaker, and grant/research support). The remaining authors have nothing to disclose.
Funding
The study was supported by the DAAD/CAPES PROBRAL program (Brazil/Germany collaboration) with research grants to Drs S P A Toledo and N S Pellegata. T Sekiya and R A Toledo received fellowships from FAPESP, and S P A Toledo received research and productivity grants from CNPq.
Author contribution statement
N S Pellegata, S P A Toledo, and R A Toledo designed the project; M D Bronstein, R S Jallad, M C Machado, T D Goncalves, L H Osaki, J Viana-Jr, C Kater, B Liberman, M C B V Fragoso, and S P A Toledo provided clinical data and DNA samples from patients; G Francisco, R Chammas, M S Naslavsky, D Schlesinger, Y A O Duarte, M L Lebrão, and M Zatz provided clinical data and DNA samples from healthy controls; T Sekiya, V C Longuini, L H Osaki, and R A Toledo performed and analyzed the genetic experiments; B Katiuscia and N S Pellegata performed and analyzed the in vitro assays; O Meirelles and R A Toledo performed the statistical analysis; T Sekiya, M D Bronstein, M C B V Fragoso, P Gama, S P A Toledo, N S Pellegata, and R A Toledo contributed to discussions; and R A Toledo wrote the paper with inputs from N S Pellegata, T Sekiya, and S P A Toledo.
References
Agarwal SK, Mateo CM & Marx SJ 2009 Rare germline mutations in cyclin-dependent kinase inhibitor genes in multiple endocrine neoplasia type 1 and related states. Journal of Clinical Endocrinology and Metabolism 94 1826–1834. (doi:10.1210/jc.2008-2083).
Chen J, Li D, Killary AM, Sen S, Amos CI, Evans DB, Abbruzzese JL & Frazier ML 2009 Polymorphisms of p16, p27, p73, and MDM2 modulate response and survival of pancreatic cancer patients treated with preoperative chemoradiation. Annals of Surgical Oncology 16 431–439. (doi:10.1245/s10434-008-0220-8).
Chen J, Amos CI, Merriman KW, Wei Q, Sen S, Killary AM & Frazier ML 2010 Genetic variants of p21 and p27 and pancreatic cancer risk in non-Hispanic Whites a case–control study. Pancreas 39 1–4. (doi:10.1097/MPA.0b013e3181bd51c8).
Chu IM, Hengst L & Slingerland JM 2008 The Cdk inhibitor p27 in human cancer prognostic potential and relevance to anticancer therapy. Nature Reviews. Cancer 8 253–267. (doi:10.1038/nrc2347).
Dahia PL, Aguiar RC, Honegger J, Fahlbush R, Jordan S, Lowe DG, Lu X, Clayton RN, Besser GM & Grossman AB 1998 Mutation and expression analysis of the p27/kip1 gene in corticotrophin-secreting tumours. Oncogene 16 69–76. (doi:10.1038/sj.onc.1201516).
Fero ML, Rivkin M, Tasch M, Porter P, Carow CE, Firpo E, Polyak K, Tsai LH, Broudy V & Perlmutter RM et al. 1996 A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27Kip1-deficient mice. Cell 85 733–744. (doi:10.1016/S0092-8674(00)81239-8).
Francis JM, Kiezun A, Ramos AH, Serra S, Pedamallu CS, Qian ZR, Banck MS, Kanwar R, Kulkarni AA & Karpathakis A et al. 2013 Somatic mutation of CDKN1B in small intestine neuroendocrine tumors. Nature Genetics 45 1483–1486. (doi:10.1038/ng.2821).
Gayther SA, Song H, Ramus SJ, Kjaer SK, Whittemore AS, Quaye L, Tyrer J, Shadforth D, Hogdall E & Hogdall C et al. 2007 Tagging single nucleotide polymorphisms in cell cycle control genes and susceptibility to invasive epithelial ovarian cancer. Cancer Research 67 3027–3035. (doi:10.1158/0008-5472.CAN-06-3261).
Georgitsi M, Raitila A, Karhu A, van der Luijt RB, Aalfs CM, Sane T, Vierimaa O, Mäkinen MJ, Tuppurainen K & Paschke R et al. 2007 Germline CDKN1B/p27Kip1 mutation in multiple endocrine neoplasia. Journal of Clinical Endocrinology and Metabolism 92 3321–3325. (doi:10.1210/jc.2006-2843).
González P, Díez-Juan A, Coto E, Alvarez V, Reguero JR, Batalla A & Andrés V 2004 A single-nucleotide polymorphism in the human p27kip1 gene (−838C>A) affects basal promoter activity and the risk of myocardial infarction. BMC Biology 2 2–5. (doi:10.1186/1741-7007-2-2).
Kibel AS, Suarez BK, Belani J, Oh J, Webster R, Brophy-Ebbers M, Guo C, Catalona WJ, Picus J & Goodfellow PJ 2003 CDKN1A and CDKN1B polymorphisms and risk of advanced prostate carcinoma. Cancer Research 63 2033–2036.
Landa I, Montero-Conde C, Malanga D, De Gisi S, Pita G, Leandro-García LJ, Inglada-Pérez L, Letón R, De Marco C & Rodríguez-Antona C et al. 2010 Allelic variant at −79 (C>T) in CDKN1B (p27Kip1) confers an increased risk of thyroid cancer and alters mRNA levels. Endocrine-Related Cancer 17 317–328. (doi:10.1677/ERC-09-0016).
Lee M & Pellegata NS 2013 Multiple endocrine neoplasia type 4. Frontiers of Hormone Research 41 63–78. (doi:10.1159/000345670).
Li G, Sturgis EM, Wang LE, Chamberlain RM, Spitz MR, El-Naggar AK, Hong WK & Wei Q 2004 Association between the V109G polymorphism of the p27 gene and the risk and progression of oral squamous cell carcinoma. Clinical Cancer Research 10 3996–4002. (doi:10.1158/1078-0432.CCR-04-0089).
Lourenço DM Jr, Coutinho FL, Toledo RA, Montenegro FL, Correia-Deur JE & Toledo SP 2010 Early-onset, progressive, frequent, extensive, and severe bone mineral and renal complications in multiple endocrine neoplasia type 1-associated primary hyperparathyroidism. Journal of Bone and Mineral Research 25 2382–2391. (doi:10.1002/jbmr.125).
Marinoni I & Pellegata NS 2010 p27kip1: a new multiple endocrine neoplasia gene? Neuroendocrinology 93 19–28. (doi:10.1159/000320366).
Pasquali D, Circelli L, Faggiano A, Pancione M, Renzullo A, Elisei R, Romei C, Accardo G, Coppola VR & De Palma M et al. 2011 CDKN1B V109G polymorphism a new prognostic factor in sporadic medullary thyroid carcinoma. European Journal Endocrinology 164 397–404. (doi:10.1530/EJE-10-0929).
Pellegata NS 2012 MENX and MEN4. Clinics 67 13–18. (doi:10.6061/clinics/2012(Sup01)04).
Pellegata NS, Quintanilla-Martinez L, Siggelkow H, Samson E, Bink K, Höfler H, Fend F, Graw J & Atkinson MJ 2006 Germ-line mutations in p27Kip1 cause a multiple endocrine neoplasia syndrome in rats and humans. PNAS 103 15558–15563. (doi:10.1073/pnas.0603877103).
Santos MA, Nunes AB, Abelin N, Ezabella MC, Toledo RA, Lourenço DM Jr, Hayashida CY, Fonseca II & Toledo SP 2006 Genetic screening of multiple endocrine neoplasia type 2 experience of the USP Endocrine Genetics Unit. Arquivos Brasileiros de Endocrinologia e Metabologia 50 7–16. (doi:10.1590/S0004-27302006000100003).
Schöndorf T, Eisele L, Göhring UJ, Valter MM, Warm M, Mallmann P, Becker M, Fechteler R, Weisshaar MP & Hoopmann M 2004 The V109G polymorphism of the p27 gene CDKN1B indicates a worse outcome in node-negative breast cancer patients. Tumour Biology 25 306–312. (doi:10.1159/000081396).
Tigli H, Buyru N & Dalay N 2005 Molecular analysis of the p27/kip1 gene in breast cancer. Molecular Diagnosis 9 17–21. (doi:10.2165/00066982-200509010-00003).
Toledo SP, Lourenço DM Jr, Santos MA, Tavares MR, Toledo RA & Correia-Deur JE 2009 Hypercalcitoninemia is not pathognomonic of medullary thyroid carcinoma. Clinics 64 699–706. (doi:10.1590/S1807-59322009000700015).
Toledo RA, Mendonca BB, Fragoso MC, Soares IC, Almeida MQ, Moraes MB, Lourenço DM Jr, Alves VA, Bronstein MD & Toledo SP 2010a Isolated familial somatotropinoma 11q13-LOH and gene/protein expression analysis suggests a possible involvement of AIP also in non-pituitary tumorigenesis. Clinics 65 407–415. (doi:10.1590/S1807-59322010000400010).
Toledo RA, Wagner SM, Coutinho FL, Lourenço DM Jr, Azevedo JA, Longuini VC, Reis MT, Siqueira SA, Lucon AM & Tavares MR et al. 2010b High penetrance of pheochromocytoma associated with the novel C634Y/Y791F double germline mutation in the RET protooncogene. Journal of Clinical Endocrinology and Metabolism 95 1318–1327. (doi:10.1210/jc.2009-1355).
Toledo RA, Qin Y, Srikantan S, Morales NP, Li Q, Deng Y, Kim SW, Pereira MA, Toledo SP & Su X et al. 2013 In vivo and in vitro oncogenic effects of HIF2A mutations in pheochromocytomas and paragangliomas. Endocrine-Related Cancer 20 349–359. (doi:10.1530/ERC-13-0101).
Tomoda K, Kubota Y & Kato J 1999 Degradation of the cyclin-dependent-kinase inhibitor p27Kip1 is instigated by Jab1. Nature 398 160–165. (doi:10.1038/18230).
Tomoda K, Kubota Y, Arata Y, Mori S, Maeda M, Tanaka T, Yoshida M, Yoneda-Kato N & Kato JY 2002 The cytoplasmic shuttling and subsequent degradation of p27Kip1 mediated by Jab1/CSN5 and the COP9 signalosome complex. Journal of Biological Chemistry 277 2302–2310. (doi:10.1074/jbc.M104431200).
Wei F, Xu J, Tang L, Shao J, Wang Y, Chen L & Guan X 2012 p27Kip1 V109G polymorphism and cancer risk a systematic review and meta-analysis. Cancer Biotherapy & Radiopharmaceuticals 27 665–671. (doi:10.1089/cbr.2012.1229).
Yao L, Schiavi F, Cascon A, Qin Y, Inglada-Pérez L, King EE, Toledo RA, Ercolino T, Rapizzi E & Ricketts CJ et al. 2010 Spectrum and prevalence of FP/TMEM127 gene mutations in pheochromocytomas and paragangliomas. Journal of the American Medical Association 304 2611–2619. (doi:10.1001/jama.2010.1830).
Zhuang Z, Yang C, Lorenzo F, Merino M, Fojo T, Kebebew E, Popovic V, Stratakis CA, Prchal JT & Pacak K 2012 Somatic HIF2A gain-of-function mutations in paraganglioma with polycythemia. New England Journal of Medicine 367 922–930. (doi:10.1056/NEJMoa1205119).
R A Toledo is now at Division of Hematology and Medical Oncology, Department of Medicine, Cancer Therapy and Research Center, University of Texas Health Science Center, 7703 Floyd Curl Drive, MC 7880, San Antonio, Texas 78229-3900, USA
