Search for LOH has provided a strong tool for gaining insight into the process of cancer neo-genesis and the involvement of a subset of deletions and mutations in the initiation and pro-gression of tumor development. Midgut carcinoids are rare malignant tumors of the small intes-tine. The primary tumor is often inconspicious in size but nevertheless associated with generally larger mesenteric lymph node or liver metastases. The rather indolent malignant behaviour as well as their low incidence has evoked an interest in the genetic events in tumorigenesis of these neoplasms.

Only few midgut carcinoids have been investigated so far (Toliat, Berger et al. 1997; Ghimenti, Lonobile et al. 1999; Jakobovitz, Nass et al. 1996; Debelenko, Emmert-Buck et al. 1997a; Gortz, Roth et al. 1999; Zhao, de Krijger et al. 2000; DŽAdda, Pizzi et al. 2002; Kytola, Nord et al. 2002). In contrast, the molecular mechanisms involved in the tumorigenesis of more common highly malignant gastrointestinal carcinomas are better characterized including frequent mutations of the Smad4/DPC4 and DCC genes on chromosome 18q21, APC on 5q21 and p16 on 9p2 .

This genome-wide screening for LOH of eight midgut carcinoid tumors revealed multiple allelic deletions with losses found in all tumors and in most cases encompassing both chromosomal arms. The rate of fractional allellic loss (FAL, in %) varied greatly between the different lesions. Even though the tumor samples still may have contained a certain amount of contaminating fibroblasts after microdissection, the LOH results were striking, especially in tumors 5216 and 5807, presenting with most deleted chromosomes.

Most conspicious were the findings of LOH on chromosome 18 with allelic losses in seven of eight lesions (88%). Only one study, using comparative genomic hybridization, could reveal LOH on chromosome 18 so far, LOH on chromosome 18p in 7/15 and on 18q in 8/15 midgut carcinoids (Zhao, de Krijger et al. 2000). In all but one tumor (2762) the deletions were large and included all informative markers on chromosome 18. LOH on chromosome 18 is a common event in a high proportion of gastroenteropancreatic carcinomas (Fearon, Cho et al. 1990; Schutte, Hruban et al. 1996; Uchida, Nagatake et al. 1996) as well as other tumor types (Papadimitrakopoulou, Oh et al. 1998; Hessman, Lindberg et al. 1999). Moreover, colorectal cancers with LOH on chromosome 18 behave clinically more aggressive than those without LOH (Fearon, Ekstrand et al. 1994). Three candidate tumor suppressor genes have been identified in this region: The DCC (deleted in colorectal cancer), Smad4/DPC4 and Smad2/MADR2/JV18-1 genes located on 18q21 (Fearon, Ekstrand et al. 1994; Eppert, Scherer et al. 1996; Schutte, Hruban et al. 1996). The Smad4/DPC4 gene has been found homozygously mutated in both exocrine and endocrine pancreatic tumors as well as colorectal cancers (Schutte, Hruban et al. 1996; Takagi, Kohmura et al. 1996; Howe, Roth et al. 1998; Bartsch, Hahn et al. 1999; Friedl, Kruse et al. 1999). DCC has also been described to be homozygously deleted in a subset of pancreatic and other cancers (Fearon, Ekstrand et al. 1994; Hilgers, Song et al. 2000) while Smad2/MADR2/JV18-1 alterations have been detected in a limited fraction of colorectal and lung cancers (Eppert, Scherer et al. 1996; Uchida, Nagatake et al. 1996). Although the deletion pattern suggested a tumor suppressor gene telomeric of the gene cluster at 18q21, we chose to analyze the Smad4/DCP4 gene in more detail since exon-specific PCR-primers were available and the described anti-serum for immunohistochemical labeling of Smad4/DPC4 has proven specificity and sensitivity for homozygous gene inactivation of 94% and 91%, respectively (Wilentz, Su et al. 2000). For Smad2/MADR2/JV18-1 and DCC, the correlation of immunohistochemical labeling and gene inactivation is unknown.

In our series of midgut carcinoids we did not identify any mutations in exon 8-11 of the Smad4/DPC4 gene, including the intron-exon boundaries. Although homozygous deletions are difficult to exclude, all exons of all tumors were PCR-amplified with the same efficiency. Immunohistochemical staining of seven tumors, revealed normal expression of the Smad4/DPC4 protein in all investigated lesions. The sequencing results and the distinct staining of Smad4/DPC4 protein strongly suggest the idea that the Smad4/DPC4 gene is unlikely to be in-volved in the development of midgut carcinoids. Tumor 2762, however, displayed a more limited deletion telomeric to D18S858 and the Smad4/DPC4 and DCC loci on 18q, allelic loss of 18q markers D18S541 and D18S844. This same tumor had remained D18S858 on 18q as well as all 18p markers. These findings suggest that this region of LOH might harbour yet another TSG distal to Smad4/DPC4 and DCC loci on 18q unknown to date and that its inactivation is more likely to be involved in the initiation of midgut carcinoid neogenesis.

Previous studies on chromosome 11 have described one somatic missense MEN1 mutation (V531 in exon 2) in one of 16 midgut carcinoid tumors (Toliat, Berger et al. 1997; Gortz, Roth et al. 1999). Two constitutional putative missense mutations, H50R and G12S on the SDHD (TSG) (succinate-ubiquinone oxidoreductase subunit D) gene locus were found in two midgut carcinoids, both mutations were associated with LOH of the other allele (Kytola, Nord et al. 2002). Microsatellite instability was detected in one of six analyzed midgut carcinoid tumors (Ghimenti, Lonobile et al. 1999). LOH on chromosome 11 was analyzed in 16of 83 midgut carcinoids (Jakobovitz, Nass et al. 1996; Toliat, Berger et al. 1997; Debelenko, Emmert-Buck et al. 1997a; Ghimenti, Lonobile et al. 1999; Gortz, Roth et al. 1999; Zhao, de Krijger et al. 2000; DŽAdda, Pizzi et al. 2002; Kytola, Nord et al. 2002. In contrast to these findings, we could not detect LOH on 11q13 in our midgut carcinoids even though both PYGM and INT2 flanking the MEN1 gene were informative in six of eight tumors and for the other two tumors markers D11S2000 at 11q22 was informative. This discrepancy might be due to different patient series and different microsatellites used in the studies. However, the lack of LOH on chromosome 11, especially at 11q13, holds with the fact that sporadic midgut carcinoids are not associated with the MEN1 syndrome and MEN1 deletions only occur in a subset of these tumors. Clearly more tumors have to be analyzed in order to gain a more true insight into involvement of 11q13 in midgut carcinoid neogenesis.