Colorectal and anal neoplasms following liver transplantation

OBJECTIVE

Liver transplantation (LT) is the treatment of choice for end-stage liver disease. The required immunosuppression increases the risk for developing malignancies. Some viruses play a crucial role. Data on neoplasms of the colon, rectum and anus in LT are limited.

METHOD

A retrospective evaluation of the incidence and clinical course of colorectal and anal malignancies and colonic polyps in a series of 467 consecutive LTs in 402 individuals between 1998 and 2001 was performed. Standard immunosuppression included Tacrolimus, Mycophenolic acid and steroids.

RESULTS

During a median follow up of 5.2 years, three colon adenocarcinomas, one EBV associated cecal posttransplant lymphoproliferative tumour and two HPV associated anal tumours were identified. Pre-LT colonoscopy was performed in 161 patients (40%), and of 153 evaluable individuals, 53 (34.9%) had polyps. Colonoscopy was performed in 186 patients (46.3%) median 14.8 (range 0.2-77.8) months post-LT and 55 (29.3%) had polyps. Post-LT adenomatous polyps were detected in 47.3% of patients with pre-LT polyps vs 6.7% of patients without pre-LT polyps (P < 0.001). Patients with alcoholic liver disease had a significantly higher rate of adenoma formation (50.0% vs 11.1%, P < 0.001). No patient died from colorectal/anal malignancy.

CONCLUSION

The incidence of metachronous and new polyp formation in our study is similar to people who are not immunocompromised, but subgroups are at increased risk. Viral-associated malignancies, including post-transplant lymphoproliferative disorders and anal cancer, are important entities in the LT population suggesting that complete screening of the colon, rectum and anus including pre-LT and post-LT colonoscopy should be utilized.

Management of Wilms tumors in Drash and Frasier syndromes

BACKGROUND

Children with WT1 gene-related disorders such as Denys-Drash syndrome (DDS) and Frasier syndrome (FS) are at increased risk of Wilms tumor and end-stage renal disease. We investigated whether Wilms tumors in these patients displayed a specific phenotype or behavior and whether nephron-sparing surgery was beneficial.

PROCEDURE

We retrospectively studied all patients with DDS, FS, or other WT1 mutations treated at our institutions between 1980 and 2007.

RESULTS

We identified 20 patients, of whom 18 had benign or malignant tumors. Wilms tumors occurred in 15 patients, being unilateral in 10 and bilateral in 5 (20 tumors). Median age at Wilms tumor diagnosis was 9 months. No patients had metastases. According to the International Society of Pediatric Oncology Working Classification, there were 19 intermediate-risk tumors and one high-risk tumor; no tumor was anaplastic. In patients with nephropathy who underwent unilateral nephrectomy for Wilms tumor or nephron-sparing surgery for bilateral Wilms tumor, mean time to dialysis was 11 or 9 months, respectively. Other tumors included three gonadoblastomas (in two patients), one retroperitoneal soft-tissue tumor, and one transitional cell papilloma of the bladder. Two patients, both with stage I Wilms tumor, died from end-stage renal disease-related complications. The median follow-up time for the 18 survivors was 136 months (range, 17-224 months).

CONCLUSION

Most Wilms tumors in children with WT1-related disorders were early-stage and intermediate-risk tumors, with a young age at diagnosis. In patients without end-stage renal disease, nephron-sparing surgery should be considered for delaying the onset of renal failure.

Surgical management and genotype/phenotype correlations in WT1 gene-related diseases (Drash, Frasier syndromes)

BACKGROUND/PURPOSE

The WT1 gene plays a role in urogenital and gonadal development. Germline mutations of this gene have been observed in patients with Drash or Frasier syndrome (Sd). The purpose of this report is to compare phenotype and genotype of these patients.

METHODS

Retrospective study of 12 patients treated since 1980 for WT1 gene-related disorders was conducted.

RESULTS

End-stage renal disease (ESRD) occurred in 9 patients, mostly because of diffuse mesangial sclerosis (DMS) or focal and segmental glomerular sclerosis (FSGS). Seven patients underwent kidney transplantation, and 2 died. Eleven tumors occurred: 8 Wilms' tumors, one soft tissue tumor, one bladder papilloma, and one gonadoblastoma. Wilms' tumors occurred at a younger age than expected. Eight patients had a 46,XY karyotype. One of these XY patients had female phenotype (Frasier syndrome); she was raised as a girl with bilateral gonadectomy. Seven XY patients had ambiguous phenotype; 4 have been raised as boys and 3 as girls. Four patients had a 46,XX karyotype; they had female genitalia and were raised as girls. WT1 gene analysis was performed in 10 patients and showed heterozygous germline mutations in exon 9 (n = 6), intron 9 (n = 1), exon 3 (n = 1), exon 4 (n = 1), or exon 7 (n = 1).

CONCLUSIONS

ESRD was secondary to DMS when exon 9 was mutated, and secondary to FSGS when intron 9 was mutated. When exon 3, 4, and 7 were mutated, no nephropathy has been observed. Wilms' tumors occurred with any kind of WT1 mutation except intron 9. Abnormal sexual differentiation has been observed in all XY patients with WT1 mutation, and the most profound inversion of phenotype was observed with mutation in intron 9. Correlation between phenotype and genotype provides better understanding of the role of WT1, and can help the surgeon in the management of these patients.

Profiling of differential gene expression in Wilms tumor by cDNA expression array

In order to identify genes or pathways involved in Wilms tumor etiology, we used the Atlas Cancer cDNA expression array to compare the gene expression profiles of five tumors, one Wilms tumor cell line (SK-NEP1), and normal mature and fetal kidneys. Of 588 genes tested, 153 had a different expression pattern in tumors compared with mature kidney. Ninety-six genes were differentially expressed in tumors compared with both normal mature and fetal kidney, and 57 genes had expression profiles similar to that of fetal kidney, which may reflect the developmental stage of the tumor cells. Comparison of the expression patterns of tumors shows that only 13% of the differentially expressed genes are constantly up- or downregulated in the five tumors tested, and this provides molecular evidence of tumor heterogeneity. We then confirmed the differential expression by an independent method, using quantitative reverse transcriptase polymerase chain reaction for two of the differentially expressed genes, MMP-14 and cyclin D2. Analysis of expression levels in a panel of 40 tumors showed that 30% overexpressed MMP-14 and 80% overexpressed cyclin D2. Profiling of gene expression using cDNA arrays in a large tumor panel will ultimately lead to the molecular classification of tumors, the identification of prognosis markers, and the design of targeted therapy.

WT1 splicing alterations in Wilms' tumors

Hereditary and sporadic forms of tumors are generally related to germ-line and somatic mutations of the same tumor suppressor gene. Unexpectedly, in Wilms' tumor, somatic mutations of the WT1 gene were found only occasionally in sporadic cases, although constitutional mutations of this gene are clearly associated with predisposition. It has been suggested that abnormal splicing may be another mode of somatic WT1 alteration. However, this idea was based on the analysis of a small series of tumors, precluding accurate evaluation of the frequency of such changes. To investigate WT1 changes at the somatic level in more detail, we analyzed the levels of the four isoform transcripts produced by alternative splicing events in a large series of 50 tumors, normal mature kidneys, and fetal kidneys. We characterized splicing alterations in 63% of sporadic Wilms' tumors. Moreover, taking into account the decreased and increased overall levels of WT1 mRNA, the percentage of sporadic tumors with changes in WT1 expression reached 90%. Whether and how these alterations of expression play a role in the tumorigenic process remain to be evaluated.

Spectrum of hSNF5/INI1 somatic mutations in human cancer and genotype-phenotype correlations

The hSNF5/INI1 gene which encodes a member of the SWI/SNF chromatin ATP-dependent remodeling complex, is a new tumor suppressor gene localized on chromosome 22q11.2 and recently shown to be mutated in malignant rhabdoid tumors. We have searched for hSNF5/INI1 mutations in 229 tumors of various origins using a screening method based on denaturing high-performance liquid chromatography. A total of 31 homozygous deletions and 36 point alterations were identified. Point mutations were scattered along the coding sequence and included 15 nonsense, 15 frameshift, three splice site, two missense and one editing mutations. Mutations were retrieved in most rhabdoid tumors, whatever their sites of occurrence, indicating the common pathogenetic origin of these tumors. Recurrent hSNF5/INI1 alterations were also observed in choroid plexus carcinomas and in a subset of central primitive neuroectodermal tumors (cPNETs) and medulloblastomas. In contrast, hSNF5/INI1 point mutations were not detected in breast cancers, Wilms' tumors, gliomas, ependymomas, sarcomas and other tumor types, even though most analyzed cases harbored loss of heterozygosity at 22q11.2 loci. These results suggest that rhabdoid tumors, choroid plexus carcinomas and a subset of medulloblastomas and cPNETs share common pathways of oncogenesis related to hSNF5/INI1 alteration and that hSNF5/INI1 mutations define a genetically homogeneous family of highly aggressive cancers mainly occurring in young children and frequently, but not always, exhibiting a rhabdoid phenotype.

WT1 and PAX-2 podocyte expression in Denys-Drash syndrome and isolated diffuse mesangial sclerosis

Denys-Drash syndrome is a rare disorder of urogenital development characterized by the association of early onset glomerulopathy caused by diffuse mesangial sclerosis, gonadal dysgenesis leading to pseudohermaphroditism in males, and a high risk of developing Wilms' tumor. The syndrome is caused by dominant negative point mutations in the WT1 gene that encodes a tumor suppressor transcription factor normally expressed in podocytes. Mutations usually affect the zinc fingers of the WT1 protein. The basic defect is unknown in most cases of isolated diffuse mesangial sclerosis, a disease characterized by the same glomerular changes as in Denys-Drash syndrome but possibly transmitted as an autosomal recessive trait. Here we show that the distribution of WT1 is abnormal in most patients with Denys-Drash syndrome : WT1 nuclear staining of podocytes is decreased or absent. This finding is consistent with the decreased DNA binding capacity of the mutated protein. One target gene of WT1 is PAX2, the expression of which is down-regulated in podocytes during early stages of nephrogenesis. We demonstrate that WT1 mislocalization is associated with abnormal podocyte expression of PAX2 protein and RNA. We suggest that persistent expression of PAX2 is likely to result from the loss of WT1 dependent transcriptional repression and may participate in the pathological mechanisms leading to glomerular dysfunction. Abnormal distribution of WT1 and PAX2 was also observed in isolated diffuse mesangial sclerosis suggesting that a defect in WT1 could also be operative in isolated diffuse mesangial sclerosis. Primary involvement of PAX2 is an alternative hypothesis because persistent expression of PAX2 in transgenic mice is associated with the occurrence of early and severe glomerulopathy.

Identification of constitutional WT1 mutations, in patients with isolated diffuse mesangial sclerosis, and analysis of genotype/phenotype correlations by use of a computerized mutation database

Constitutional mutations of the WT1 gene, encoding a zinc-finger transcription factor involved in renal and gonadal development, are found in most patients with Denys-Drash syndrome (DDS), or diffuse mesangial sclerosis (DMS) associated with pseudohermaphroditism and/or Wilms tumor (WT). Most mutations in DDS patients lie in exon 8 or exon 9, encoding zinc finger 2 or zinc finger 3, respectively, with a hot spot (R394W) in exon 9. We analyzed a series of 24 patients, 10 with isolated DMS (IDMS), 10 with DDS, and 4 with urogenital abnormalities and/or WT. We report WT1 heterozygous mutations in 16 patients, 4 of whom presented with IDMS. One male and two female IDMS patients with WT1 mutations underwent normal puberty. Two mutations associated with IDMS are different from those described in DDS patients. No WT1 mutations were detected in the six other IDMS patients, suggesting genetic heterogeneity of this disease. We analyzed genotype/phenotype correlations, on the basis of the constitution of a WT1 mutation database of 84 germ-line mutations, to compare the distribution and type of mutations, according to the different symptoms. This demonstrated (1) the association between mutations in exons 8 and 9 and DMS; (2) among patients with DMS, a higher frequency of exon 8 mutations among 46, XY patients with female phenotype than among 46,XY patients with sexual ambiguity or male phenotype; and (3) statistically significant evidence that mutations in exons 8 and 9 preferentially affect amino acids with different functions.

Software and database for the analysis of mutations in the human WT1 gene

The WT1 gene, located at 11p13, encodes a zinc finger transcription factor involved in renal and gonadal development and in Wilms' tumor. Constitutional mutations of this gene have been described in most patients with Denys Drash syndrome (mesangial sclerosis associated with male pseudohermaphrodism and/or Wilms' tumor), but also in patients with genitourinary abnormalities and Wilms' tumor (WT) or presenting with only unilateral or bilateral WT. Moreover, approximately 10% of Wilms' tumors carry WT1 mutations at the somatic level. To facilitate the genotype-phenotype correlation analyses, we have created a software package along with a computerized database of germline (70 entries) and somatic (28 entries) mutations reported in the literature.

Correlations of allelic imbalance of chromosome 14 with adverse prognostic parameters in 148 renal cell carcinomas

To investigate cumulative genetic alterations during development and progression of renal cell carcinoma (RCC), we examined DNAs that were isolated from 148 RCCs for allelic imbalance (AI) at four loci on chromosome arm 3p and at 26 loci on chromosome arm 14q by using polymorphic microsatellite markers and densitometric scanning. Because the analysis of solid tumor unbalanced rearrangements remains difficult due to the large proportion of cells that infiltrate from the stroma, we developed a method for the detection and quantification of AI between control and tumor samples by using polymerase chain reaction (PCR) amplification of microsatellite markers. This technique allows detection down to 20% of contaminating cells with good accuracy. We detected AI on 3p and 14q in 57 and 28% of RCC, respectively. A comparison of genetic changes with clinicopathological data showed that, in marked contrast to AI on 3p, AI on 14q was correlated significantly with the stage and grade of the tumors, with 56 and 58% of RCC in Stage IV and Grade 4, respectively, showing AI. Our results suggest that tumor suppressor genes on 3p, including the von Hippel-Lindau gene, may be involved in early steps of carcinogenesis in clear cell carcinoma and that AI on 14q may play an important role in the progression of clear cell and papillary chromophilic cell carcinomas. Loss of heterozygosity (LOH) on 14q may be a new prognostic factor in RCC. Despite the size of the series of tumors and the number of markers used, only rearrangements that involved the whole length of the long arm of chromosome 14 were observed in the present study. The localization of the putative tumor suppressor gene on 14q will require further investigation of RCC with structural rearrangements of 14q.

Characterization of regions of chromosomes 12 and 16 involved in nephroblastoma tumorigenesis

There are at least three loci involved in Wilms' tumor (WT) tumorigenesis: WT1 in 11p13, WT2 in 11p15.5, and WT3, as yet unmapped. A compilation of cytogenetic data published for 107 WT revealed that deletion of chromosome 16 and duplication of chromosome 12 occur as frequently as the well-documented 11p deletions. Allelic imbalance for chromosomes 16 and 12 was investigated in a series of 28 WT. By use of a large panel of restriction fragment length polymorphisms and (CA)n probes, we demonstrated loss of heterozygosity (LOH) for 16q in seven (25%) of the tumors. The whole length of 16q was involved in six of the tumors. Moreover, consistent with a previous report of 16q13 LOH in a sporadic WT and a constitutional breakpoint with a Beckwith-Wiedemann patient, we map a region of particular interest to between D16S308 and D16S320. The assumption that 16q LOH may be an early event was based on: 1) the detection of 16q LOH in one case of nephroblastomatosis; 2) the presence of a complete (clonal) 16q LOH in a tumor with partial (mosaic) 11p LOH; and 3) 16q LOH as the sole abnormality in one WT. By quantification of chromosome 12 allelic imbalance, we detected duplication in 18% of the total series and in 25% of the sporadic unilateral cases. The common region extended from the centromere to D12S7 in 12q21.1-q23. We also suggest that the various pathogenetically important loci are not equally involved in the different forms of WT and that their sequential involvement may differ.

Isolation of kidney complementary DNAs down-expressed in Wilms' tumor by a subtractive hybridization approach

We applied a subtractive hybridization approach to isolate genes differentially expressed between mature kidney and Wilms' tumor. We constructed a complementary DNA library from a total mature kidney complementary DNA subtracted by an excess of mRNA from a Wilms' tumor, WAGR4, with a germline deletion of 11p13 and a somatic loss of alleles at 11p15. Six clones presenting a differential pattern of expression, positive with mRNA from the mature kidney and negative with mRNA from the Wilms' tumor WAGR4, were characterized. Among these clones were two as yet unknown expressed sequences (D11S877E and D15S109E) and four sequences from known genes: renal dipeptidase (DPEP1), alpha B-crystallin (CRYA2), uromodulin (UMOD), and plasma glutathione peroxidase (GPX2). The different patterns of expression of these genes in 11 Wilms' tumors, whether or not they are hereditary, reflect the well-documented pathogenetic heterogeneity for Wilms' tumors. We propose that these clones could be helpful for an improved histological characterization of Wilms' tumors.

Physical and genetic mapping of the dipeptidase gene DPEP1 to 16q24.3

We report the subregional physical and genetic mapping on chromosome 16q of a cDNA clone selected as a potential tumor/growth suppressor sequence. By DNA sequencing and RNA expression pattern, this clone was identified as part of the renal dipeptidase gene (DPEP1). Using somatic cell hybrids carrying either different human chromosomes or chromosome 16 segments, we confirm and refine the physical mapping of DPEP1 to the chromosome 16 subregion q24.3. Two RFLPs, a biallelic polymorphism detected by TaqI and a VNTR detected by BamHI, EcoRI, and BglII, are described. Using the VNTR polymorphism, DPEP1 was shown to be linked to D16S7 with a maximum lod score of 5.8 at a recombination fraction of 0.03.

Chromosomal assignment of the uromodulin gene (UMOD) to 16p13.11

We report the chromosomal assignment on chromosome arm 16p of a cDNA clone isolated through its expression in mature kidney and lack of expression in several Wilms tumors. DNA sequencing and analysis of the pattern of RNA expression in different tissues identified this clone as a uromodulin (Tamm-Horsfall glycoprotein, uromucoid; UMOD) sequence. By hybridizing this clone to somatic cell hybrids carrying different human chromosomes or segments of chromosome 16, the gene for UMOD was localized to 16p13.11.

Subregional physical mapping of an alpha B-crystallin sequence and of a new expressed sequence D11S877E to human 11q

We report the regional assignment on Chromosome (Chr) 11q of two cDNA clones selected as sequences expressed in mature kidney and not expressed in Wilms' tumor. Clone T70 was identified as an alpha B-crystallin sequence (CRYA2). CRYA2 has previously been mapped to 11q22.3-23.1 by in situ hybridization. Clone 6.2 represents a new gene expressed in adult and fetal kidney, pancreas, and liver. In order to map sequences corresponding to clone 6.2 and to physically define the boundaries of the localization of CRYA2, we used somatic cell hybrids carrying either different human chromosomes or Chr 11 segments and a cell line established from a patient with an interstitial deletion of region 11q14.3-q22.1. We showed that CRYA2 lies proximal to the 11q23.2 breakpoint defined by the constitutional t(11;22) and distal to the 11q22.1 breakpoint (between D11S388 and D11S35) of a constitutional interstitial deletion. This is in agreement with previous data obtained by in situ hybridization and provides proximal and distal physical benchmarks for this localization. Clone 6.2-related sequence (D11S877E) was assigned to region 11q23.2-q24.2 defined by the breakpoints of the constitutional t(11;22) and of the Ewing's sarcoma neuroepithelioma t(11;22).

11p15.5-specific libraries for identification of potential gene sequences involved in Beckwith-Wiedemann syndrome and tumorigenesis

Constitutional and somatic chromosomal abnormalities of the chromosome 11p15 region are involved in an overgrowth malformation syndrome, the Beckwith-Wiedemann syndrome (BWS), and in several types of associated tumors. The bias in parental origin for the different etiologic forms of this syndrome and for loss of heterozygosity in the tumors suggests that a gene (or genes) mapping to this region undergoes genomic imprinting. However, the precise localization of the locus (or loci) for the BWS and associated tumors is still unknown and more markers are required. We therefore isolated 11p15 markers from two libraries: the first one obtained by microdissection of the chromosome 11p15.5 region and the second one, a phage library, constructed from a hybrid cell line containing this region as its sole human DNA. Of 19 microclones isolated from the microdissection library, 11 were evolutionarily conserved. Four phage clones were isolated; one (D11S774) detected a highly informative variable number of tandem repeats (VNTR) and another (D11S773) a biallelic polymorphism. These clones were sublocalized using a panel of somatic cell hybrids that defines eight physical intervals in 11p15.5. Twenty-one clones map to the distal interval that harbors the BWS locus.

[Antioncogenes: models for tumors in children]

The chromosomal assignment of genes responsible for malformation syndromes associated with increased susceptibility to malignancy could be determined owing to specific constitutional chromosomal abnormalities or to family studies. For certain types of tumors, somatic chromosomal rearrangements (loss of alleles) occur at the same locus indicating the presence of a recessive suppressor gene or an antioncogène. For other types of tumors chromosomal rearrangements involving regions different from the locus for predisposition suggest genetic heterogeneity and/or implication of genes for tumor progression. These genes which are also involved in development and regulation of differentiation and cell growth, may undergo a differential genomic imprinting.

Constitutional and somatic deletions of two different regions of maternal chromosome 11 in Wilms tumor

Loss of heterozygosity for 11p markers and preferential loss of maternal alleles have been described in Wilms tumor. In this report we describe the molecular characterization of the constitutional and somatic 11p rearrangements in a del(11p13) WAGR patient with Wilms tumor. Both rearrangements led to loss of maternal alleles for two different regions of 11p, namely, 11p13 and 11p14----p15. This result clearly suggests that Knudson's hypothesis of two hits at the same locus does not necessarily apply to Wilms tumor. Moreover, the loss of 11p15 maternal alleles in the tumor is not incompatible with maternal inheritance of predisposition at 11p13. The putative roles of these two loci are discussed.

Characterization of a panel of somatic cell hybrids for subregional mapping along 11p and within band 11p13. Subdivision of the WAGR complex region

The short arm of chromosome 11 carries genes involved in malformation syndromes, including the aniridia/genitourinary abnormalities/mental retardation (WAGR) syndrome and the Beckwith-Wiedemann syndrome, both of which are associated with an increased risk of childhood malignancy. Evidence comes from constitutional chromosomal aberrations and from losses of heterozygosity, limited to tumor cells, involving regions 11p13 and 11p15. In order to map the genes involved more precisely, we have fused a mouse cell line with cell lines from patients with constitutional deletions or translocations. Characterization of somatic cell hybrids with 11p-specific DNA markers has allowed us to subdivide the short arm into 11 subregions, 7 of which belong to band 11p13. We have thus defined the smallest region of overlap for the Wilms' tumor locus bracketed by the closest proximal and distal breakpoints in two of these hybrids. The region associated with the Beckwith-Wiedemann syndrome spans the region flanked by two 11p15.5 markers, HRAS1 and HBB. These hybrids also represent useful tools for mapping new markers to this region of the human genome.

Mapping of a human fibrillar collagen gene, pro alpha 1 (XI) (COL11A1), to the p21 region of chromosome 1

Type XI collagen is a minor and poorly characterized structural component of cartilage. Recently, cDNA and genomic clones coding for the pro alpha 1 chain of human Type XI collagen, formerly 1 alpha collagen, have been isolated and fully characterized. Here we have used one such probe to establish the chromosomal localization of the pro alpha 1 (XI) collagen gene (COL11A1) by hybridization to filter-bound DNA isolated from flow-sorted chromosomes and by in situ hybridization on metaphase chromosomes. This combination of approaches has enabled us to locate COL1A11 in the p21 region of chromosome 1. This represents the first mapping of a Type XI collagen gene and the first assignment of a collagen locus to chromosome 1. These studies also provide additional evidence for the nearly uniform dispersion of the human fibrillar collagen genes in the human genome.