Mechanisms of Loss of Heterozygosity in Neurofibromatosis Type 1-Associated Plexiform Neurofibromas

Transcription

1 ORIGINAL ARTICLE Mechanisms of Loss of Heterozygosity in Neurofibromatosis Type 1-Associated Plexiform Neurofibromas Katharina Steinmann 1, Lan Kluwe 2, Reinhard E. Friedrich 2, Victor-Felix Mautner 2, David N. Cooper 3 and Hildegard Kehrer-Sawatzki 1 Plexiform neurofibromas constitute a serious burden for patients with neurofibromatosis type 1 (NF1), a common autosomal dominant disorder characterized by pigmentary changes and tumorous skin lesions (neurofibromas). Despite the prominence of these benign tumors in NF1 patients, the mechanisms underlying the tumor-associated loss of heterozygosity (LOH) in plexiform neurofibromas have not been extensively studied. We performed LOH analysis on 43 plexiform neurofibromas from 31 NF1 patients, the largest study of its kind to date. A total of 13 (30%) plexiform neurofibromas exhibited LOH involving 17q markers. In three tumors, LOH was found to be confined to the NF1 gene region. However, in none of the tumors was a somatic NF1 microdeletion, mediated by non-allelic homologous recombination between either NF1-REPs or SUZ12 genes, detected. Thus, NF1 microdeletions do not appear to be frequent somatic events in plexiform neurofibromas. Determination of NF1 gene copy number by multiplex ligation-dependent probe amplification indicated that although tumors with smaller regions of LOH were characterized by 17q deletions, no NF1 gene copy number changes were detected in six plexiform neurofibromas with more extensive LOH. To our knowledge, mitotic recombination has not previously been reported to be a frequent cause of LOH in plexiform neurofibromas. Journal of Investigative Dermatology (2009) 129, ; doi: /jid ; published online 18 September 2008 INTRODUCTION Neurofibromatosis type 1 (NF1; MIM#162200) is a common autosomal dominant disorder caused by mutations of the NF1 gene located at 17q11.2. Whereas the vast majority of NF1 patients have inherited subtle mutations within the NF1 gene (Messiaen et al., 2000), large deletions in 17q11.2, which result in the loss of the entire NF1 gene and its flanking regions, occur in B5% of NF1 patients (Kluwe et al., 2004). Two types of such recurring gross deletion have been noted: type-1 deletions span 1.4 Mb and have breakpoints within low-copy repeats, the NF1-REPs a and c (Dorschner et al., 2000; Jenne et al., 2001; López-Correa et al., 2001; De Raedt et al., 2006a). Type-2 deletions are somewhat smaller, encompassing only 1.2 Mb, and their breakpoints are located 1 Institute of Human Genetics, University of Ulm, Germany; 2 Department of Maxillofacial Surgery, University Hospital Eppendorf, Hamburg, Germany and 3 Institute of Medical Genetics, School of Medicine, Cardiff University, Cardiff, UK Correspondence: Dr Hildegard Kehrer-Sawatzki, Institute of Human Genetics, University of Ulm, Albert-Einstein-Allee 11, Ulm, Germany. hildegard.kehrer-sawatzki@uni-ulm.de Abbreviations: LOH, loss of heterozygosity; MLPA, multiplex ligationdependent probe amplification; NAHR, non-allelic homologous recombination; NF1, neurofibromatosis type 1 Received 16 April 2008; revised 25 June 2008; accepted 22 July 2008; published online 18 September 2008 in the SUZ12 gene and its pseudogene adjacent to the NF1- REPs (Kehrer-Sawatzki et al., 2004; Steinmann et al., 2007). Such large deletions of the NF1 gene region have been termed NF1 microdeletions because they are too small to be visible by cytogenetic analysis. The mutational mechanism underlying the majority of NF1 microdeletions is non-allelic homologous recombination (NAHR). Two preferred regions of NAHR in NF1 type-1 deletions have been noted within the NF1-REPs, viz. the paralogous recombination sites PRS1 and PRS2 spanning 1.8 and 3.4 kb (kilobase pairs), respectively (Forbes et al., 2004; De Raedt et al., 2006a). In contrast to the situation with type-1 deletions, preferred breakpoint regions that are locally restricted to a few kilobase pairs have not been observed in type-2 deletions. Instead, the breakpoint regions of the 13 type-2 deletions so far characterized are dispersed over some 32 kb of the 64 kb-spanning SUZ12 gene (Steinmann et al., 2007). Clinically, NF1 is characterized by pigmentary anomalies such as café-au-lait spots and freckling, as well as the eponymous skin neurofibromas derived from the cutaneous or subcutaneous nerve sheath (summarized in Huson and Hughes, 1994). Dermal neurofibromas are benign tumors with a bump-like appearance, growing in or under the skin. By contrast, plexiform neurofibromas grow in and along the peripheral nerves and may involve multiple branches, nerve roots and plexi. Some 30% of NF1 patients have plexiform & 2009 The Society for Investigative Dermatology 615

2 neurofibromas (Huson et al., 1988, reviewed by Ferner and Gutmann, 2002), whereas multiple plexiform neurofibromas are observed in 9 21% of NF1 patients (Huson et al., 1988; Friedman and Birch, 1997; Waggoner et al., 2000). In contrast to dermal neurofibromas, plexiform neurofibromas have the potential to undergo transformation into malignant peripheral nerve sheath tumors (MPNSTs) and it has been estimated that B4 5% actually do so (Ducatman et al., 1986; Huson and Hughes, 1994). Indeed, most NF1-associated malignant peripheral nerve sheath tumors arise in preexisting plexiform neurofibromas. Although the molecular changes accompanying the malignant transformation of plexiform neurofibromas are still largely unknown, the hemizygous inactivation of the TP53 gene, involving either intragenic mutation or loss of heterozygosity (LOH), has been noted in MPNSTs. It has been suggested that the hemizygous inactivation of TP53 may be sufficient to promote malignant progression since the biallelic inactivation of TP53 in MPNSTs appears to be rare (Lothe et al., 2001; Upadhyaya et al., 2008). Plexiform neurofibromas have not however been studied comprehensively with respect to LOH of the TP53 gene. Indeed, only one study has so far addressed this question, and LOH in 17p was not observed in the 10 plexiform neurofibromas investigated (Rasmussen et al., 2000). Many plexiform neurofibromas are likely to be congenital lesions since, in a clinic-based study, 44% of plexiform neurofibromas were found to have been diagnosed before the age of 5 years (Waggoner et al., 2000). Plexiform and dermal neurofibromas contain the same cell types, including Schwann cells, fibroblasts, perineurial cells, axons and inflammatory mast cells, which are embedded in an excessive extracellular matrix. An estimated 60 80% of cells within neurofibromas constitute Schwann cells, the tumor progenitor cells (Kluwe et al., 1999; Rutkowski et al., 2000; Serra et al., 2000; Maertens et al., 2006). The NF1 gene product, neurofibromin, acts as a tumor suppressor, serving as a negative regulator of ras proteins by stimulating their intrinsic GTPase activity. Cells lacking neurofibromin have higher levels of activated ras and exhibit increased proliferation (Basu et al., 1992; DeClue et al., 1992). According to the two-hit model of tumorigenesis, both copies of the NF1 gene are inactivated in benign and malignant tumors from NF1 patients (Sawada et al., 1996 and references in Table 1; Legius et al., 1993). The somatic inactivation of the NF1 gene associated with tumorigenesis can be caused either by mechanisms leading to LOH or by intragenic mutation. Although LOH has been studied extensively in dermal neurofibromas, only some 39 plexiform neurofibromas have been previously investigated in relation to LOH at the NF1 gene locus (as summarized in Table 1). Further, the paucity of polymorphic markers employed in these latter studies precluded a more precise estimation of the extent of LOH. Here, we have investigated the extent of LOH in 43 plexiform neurofibromas. The primary aims were to identify the mechanisms leading to LOH in these tumors and to determine whether or not NAHR between duplicated sequences in the NF1 gene region (either between the NF1- REPs or between the SUZ12 gene and its pseudogene) is a Table 1. Summary of previous studies of LOH in dermal and plexiform neurofibromas together with the results obtained in the present analysis Number of Dermal neurofibromas investigated (with LOH) Plexiform neurofibromas investigated (with LOH) Reference 43 (13) (30%) This study 23 (5) (22%) Spurlock et al. (2007) 192 (60) (31%) Serra et al. (2007) 38 (3) (8%) Maertens et al. (2006) 28 (6) (21%) 1 De Raedt et al. (2006b) 3 (1) Frahm et al. (2004) 80 (14) (29%) 3 (0) Upadhyaya et al. (2004) 33 (3) (9%) Wiest et al. (2003) 126 (32) (25%) Serra et al. (2001a, b) 6 (1) (17%) 1 (1) Eisenbarth et al. (2000) 74 (10) (14%) 3 (0) John et al. (2000) 15 (2) (13%) 10 (4) (40%) Rasmussen et al. (2000) 14 (8) (57%) Kluwe et al. (1999) 60 (15) (25%) Serra et al. (1997) 38 (0) 5 (1) (20%) Däschner et al. (1997) 22 (8) (36%) Colman et al. (1995) 15 (0) Lothe et al. (1995) LOH, loss of heterozygosity. 1 In addition to these 28 neurofibromas from patients with intragenic mutations, De Raedt et al. (2006b) investigated 40 neurofibromas from patients with constitutional gross deletions in the NF1 gene region. However, none of the 40 tumors exhibited LOH. frequent event leading to LOH in the plexiform neurofibromas of NF1 patients. RESULTS A total of 43 plexiform neurofibromas from 31 patients were investigated for LOH on chromosome 17. Of the 43, 13 tumors exhibited LOH on 17q but not on 17p (30%; Figure 1). Five tumors exhibited LOH of at least 52 Mb including almost the entire length of chromosome 17q, whereas eight tumors exhibited more restricted LOH. Three tumors displayed LOH that was confined to markers located within the NF1 microdeletion region (317-1, 390 and 337-5) (Figure 1). All 43 plexiform neurofibromas were investigated with respect to whether they were positive for PCR fragments spanning the breakpoints of NF1 microdeletions comparable to those previously observed in patients with constitutional type-1 or mosaic type-2 deletions (De Raedt et al., 2006a; Steinmann et al., 2007). Although we detected LOH that was confined to the NF1 microdeletion interval in only three tumors, all 43 neurofibromas were tested in an attempt to estimate the frequency of such deletions. Investigating the presence of somatic NF1 microdeletions 616 Journal of Investigative Dermatology (2009), Volume 129

3 Position (Mb) Marker 6.19 D17S p53 17p cen p D17S1353 D17S720 D17S804 D17S520 D17S783 D17S975 D17S1307 D17S2237 IVS27TG24.8 IVS27TG28.4 D17S1166 D17S1800 D17S1666 D17S1880 D17S907 D17S1788 D17S1861 D17S1809 D17S668 D17S827E D17S928 D17S632 D17S1402E LOH of almost the entire 17q Figure 1. LOH analysis in 13 plexiform neurofibromas. Tumor acronyms are indicated above. In those seven tumors highlighted with gray rectangles, LOH was caused by a deletion according to MLPA. The markers located in the NF1 microdeletion interval are underlined. Black circles: LOH, white circle: no LOH, striped circles: marker not informative, and gray filled circle: not determined. by breakpoint-junction PCR offers the advantage of being able to detect even quite small numbers of cells carrying these deletions, if present, despite a possible high background of cells lacking the deletion. We tested all 43 neurofibromas in order to include those tumors where LOH could have been obscured by the presence of cells lacking LOH. The primers used for these analyses were designed so as to span those regions of the SUZ12 gene or the NF1-REPs already known to harbor breakpoints of constitutional or somatic deletions in patients with NF1 (Table S3). These analyses included breakpoint-spanning PCRs used to amplify across the deletion hotspots noted in type-1 deletions (PRS1 and PRS2) and a total of seven PCRs designed to detect type-2 deletions as observed in patients with somatic mosaicism for type-2 NF1 microdeletions. However, in none of the tumor-derived DNA samples were deletion-junction PCR products amplified. The presence of such fragments would have been indicative of somatic NF1 microdeletions with breakpoints in the respective regions. In order to determine whether the LOH observed in the 13 neurofibromas was caused by deletions, multiplex ligationdependent probe amplification (MLPA) analysis was performed using probes located within the NF1 gene region. Seven of the thirteen tumors were found to exhibit deletions (Table S4), mostly those tumors that were characterized by the more restricted LOH 317-1, 335-3, 337-5, 338-2, 390, 913-5, and (Figure 1). No NF1 gene copy number changes were detected in the other six tumors with LOH. DISCUSSION Several studies have previously identified an LOH, within the region of 17q harboring the NF1 gene, in plexiform and dermal neurofibromas. The percentage of dermal neurofibromas exhibiting LOH ranges from 8 to 36%, whereas in plexiform neurofibromas, LOH has been detected in 20 57% of tumors (Table 1). In this study, 13 of 43 (30%) plexiform neurofibromas exhibited LOH. The rationale of the LOH studies of plexiform neurofibromas performed here was to characterize the extent of LOH on chromosome 17, to determine the mechanism underlying LOH and to measure the frequency of somatic duplicon-mediated NF1 microdeletions in these tumors. We have further investigated the frequency of loss of the TP53 gene (located on 17p13.1) in order to explore whether it is possible at an early stage to identify those plexiform neurofibromas that have the potential to progress towards malignancy. The proportion of plexiform neurofibromas exhibiting LOH in our study (30%) may represent a conservative estimate of the actual prevalence of LOH in NF1-associated plexiform neurofibromas because the presence of LOH could have been concealed in some neurofibromas by nontumor cells that lack LOH. Both plexiform and dermal neurofibromas are heterogeneous with respect to their cellular components but only the Schwann cells experience the somatic inactivation of the wild-type NF1 allele and are consequently considered to be the tumor progenitor cells (Kluwe et al., 1999; Rutkowski et al., 2000; Maertens et al., 2006; Serra et al., 2000, 2001a, b). The heterogeneous cellular composition of neurofibromas, and in particular the nontumor cell background (of NF1 ( þ / ) genotype), hampers the analysis of somatic NF1 gene inactivation events. De Raedt et al. (2006b) observed that the percentage of cells exhibiting LOH in dermal neurofibromas varied between 30 and 75%. Consequently, LOH 617

4 may be difficult to detect if the number of NF1 ( þ / ) cells is high. Our inability to detect LOH in some neurofibromas may however also have been a consequence of the inactivation of the wild-type NF1 allele by intragenic mutation rather than by LOH. Indeed, Maertens et al. (2006) identified subtle intragenic point mutations in 26 of 38 neurofibroma-derived Schwann cell cultures and only three of these exhibited LOH. Similarly, subtle somatic NF1 gene mutations were noted in 13 of 33 dermal neurofibromas from two NF1 patients, whereas LOH was detected in only three dermal neurofibromas (Wiest et al., 2003). In our study, seven patients were included from whom two or more plexiform neurofibromas were analyzed (Table S1). In four of these patients (nos. 1221, 290, 390, and 784), LOH was found in one but not the other plexiform neurofibroma investigated. LOH was not, however, observed in any of the plexiform neurofibromas from the remaining three patients from whom we analyzed multiple tumors. At present, the number of patients from whom we could obtain different plexiform neurofibromas is too small for any firm conclusions to be drawn with respect to a possible individual predisposition to acquire preferentially specific type(s) of NF1 gene mutations (whether subtle gene mutations or LOH involving 17q) as a second hit. LOH of multiple markers distributed along 17q was observed in 5 of the 13 LOH-positive plexiform neurofibromas identified here, and was indicative either of the presence of large deletions or mitotic recombination involving extended regions of 17q. However, in none of the plexiform neurofibromas investigated here did the LOH include 17p. Thus, in none of the 13 LOH-positive tumors was the LOH due to the loss of the entire chromosome 17. In three plexiform neurofibromas, LOH appeared to be confined to the NF1 microdeletion interval (Figure 1, tumors 337-5, 390, and 317-1). In a previously studied series of 126 dermal neurofibromas, which included sufficient markers from within the NF1 gene region to permit investigation with high definition, LOH of the region bounded by the NF1-REPs a and c was observed in three dermal neurofibromas (Serra et al., 2001b). It is possible that in the tumors with LOH confined to the NF1 microdeletion interval, the mechanism resulting in LOH could have been NAHR, either between the NF1-REPs or the SUZ12 sequences, as already described in constitutional and mosaic NF1 microdeletions (De Raedt et al., 2006a; Steinmann et al., 2007). To investigate the frequency of LOH in neurofibromas caused by recurring NF1 microdeletions, deletion breakpoint-spanning PCRs were performed. In these experiments, we included not only the tumors in which LOH was confined to the NF1 microdeletion region, but also all 43 available tumors in order to determine the overall frequency of this type of somatic deletion in plexiform neurofibromas. The deletion breakpoint-spanning PCRs were designed so as to amplify across the deletion breakpoints previously identified in NF1 microdeletions. The advantage of using this PCR-based approach to identify somatic NF1 microdeletions lies in its ability, in principle, to detect deletion events in small numbers of cells. However, in none of the 43 tumor samples examined, even those in which LOH was confined to the NF1 microdeletion interval, were positive deletion breakpoint-spanning PCR products detected. This implies that somatic NF1 microdeletions, mediated by NAHR in regions previously identified as harboring deletion breakpoints in patients with constitutional and mosaic NF1 microdeletions, are not common events in plexiform neurofibromas. Interestingly, somatic type-2 NF1 microdeletions were recently identified in leukemia. Screening of 103 patients without NF1 but with pediatric T-cell acute lymphoblastic leukemia and 71 patients with acute myeloid leukemia indicated type-2 NF1 deletions in three patients with T-cell acute lymphoblastic leukemia and two patients with acute myeloid leukemia (Balgobind et al., 2008). Thus, although type-2 NF1 microdeletions occur as somatic mutations in hematopoietic cells, it would appear as if they are infrequent events in plexiform neurofibromas. Differently sized regions of LOH on 17q have been observed in dermal neurofibromas, but the majority of these studies did not include enough markers to allow assessment of whether LOH extended over the entire 17q. In the most comprehensive study of its kind, 14 microsatellite markers were analyzed including those located in the most proximal and distal portions of 17q (Serra et al., 2001b). A total of 27 dermal neurofibromas with LOH were identified, with 19 of them exhibiting LOH over almost the entire length of 17q. In another study that employed multiple well-mapped microsatellite markers, LOH over the entire 17q was observed in one of five dermal neurofibromas (Spurlock et al., 2007). If the results of both studies are taken together, LOH encompassing almost the entire 17q has been observed in a total of 20 of 32 dermal neurofibromas (63%) (Serra et al., 2001b; Spurlock et al., 2007). By comparison, plexiform neurofibromas have not been investigated extensively with regard to the extent of LOH. In one of four plexiform neurofibromas, LOH was observed to extend over virtually the entire length of 17q (Rasmussen et al., 2000). The LOH analysis in plexiform neurofibromas performed here has the highest resolution of all studies so far performed; LOH spanning almost the entire length of 17q was noted in 5 of 13 tumors with LOH (Figure 1). Thus, LOH encompassing 17q would appear to occur at a similar frequency in plexiform neurofibromas as in dermal neurofibromas. Loss or mutation of the TP53 gene at 17p13.1 has been frequently observed in NF1-associated MPNSTs and is hence likely to contribute to the progression towards malignancy (Upadhyaya et al., 2004, 2008 and references therein). By contrast, LOH of TP53 has never been found in dermal neurofibromas. All plexiform neurofibromas investigated by us retained heterozygosity for markers in 17p, indicating that LOH did not encompass the TP53 gene. This is in line with the observations of Rasmussen et al. (2000) who did not observe LOH at 17p in a total of 10 plexiform neurofibromas. Taken together, no striking difference exists, with respect to the extent of LOH on chromosome 17, between dermal and plexiform neurofibromas. Our study therefore suggests that the similarity between both subtypes of neurofibroma extends 618 Journal of Investigative Dermatology (2009), Volume 129

5 to the underlying mechanism of LOH. In addition to deletions, LOH in dermal neurofibromas has been shown to be frequently caused by mitotic recombination resulting in homozygosity of the NF1 germline mutation (Serra et al., 2001b; De Raedt et al., 2006b). In our study, 7 of 13 plexiform neurofibromas exhibited LOH due to deletions (Figure 1). In the other six plexiform neurofibromas with LOH (605-1, 385, 952-8, 389-2, 374-4, and 604-4), no reduction of NF1 copy number was observed, indicating that the LOH in these tumors was due to mitotic recombination. Mitotic recombination has also been reported in NF1-associated myeloid malignancies and gastrointestinal stromal tumors resulting in reduction to homozygosity of the NF1 germline mutation (Stephens et al., 2006; Stewart et al., 2007). Our study clearly indicates that mitotic recombination in 17q is a frequent cause of LOH in plexiform neurofibromas. We conclude that there is no significant difference between dermal and plexiform neurofibromas with respect to either the extent of LOH on 17q or the underlying mutational mechanisms. Since mitotic recombination was observed in 46% of plexiform neurofibromas with LOH, it may be regarded as a frequent mutational mechanism in these benign tumors. The progression of a fraction of plexiform neurofibromas towards malignancy and their transformation to MPNSTs may thus occur in only a small subpopulation of tumor cells but could be initiated by genetic changes that are consequent to LOH on 17q. MATERIALS AND METHODS Tumor samples A total of 43 plexiform neurofibromas (listed in Table S1) derived from 31 NF1 patients were investigated by PCR analysis to search for LOH. The patients fulfilled the diagnostic criteria for NF1 (Stumpf et al., 1988) and all gave their informed consent. Clinical data from the patients as well as the germline NF1 gene mutation, if known, are given in Table S1. Ethical approval for this study was obtained by the local Ethics Committee. Patients included in this study gave written informed consent. The study was conducted according to the Declaration of Helsinki Principles. The plexiform neurofibromas were ascertained by radiological and clinical evaluation. Tumor tissue was frozen in liquid nitrogen immediately after excision and stored at 70 1C. DNA was extracted from frozen tissue and patient blood samples using the Blood & Cell Culture DNA Kit (Qiagen, Hilden, Germany), following the sample preparation and lysis protocol for tissue samples. LOH analysis LOH was investigated using 25 microsatellite markers from chromosome 17. The sequences of the primers used for this analysis, as well as their precise genomic positions on chromosome 17, are summarized in Table S2. Seven markers are located within the NF1 microdeletion region between the NF1-REPs a and c (Figure 1). PCR products were analyzed on an ABI 3100 sequencer using a standard fragment analysis protocol. All markers were analyzed at least twice by independent PCR experiments. LOH in a tumor sample was scored when the height ratio of the two marker alleles was at least 20% different from the allele ratio observed in patient peripheral blood DNA and in blood samples from unaffected probands with the same allele sizes as those detected in the tumor DNAs. Breakpoint-junction PCRs A total of nine primer pairs (listed in Table S3) were used to amplify across the deletion breakpoints identified in constitutional type-1 NF1 microdeletions and mosaic type-2 microdeletions as described in previous studies (Jenne et al., 2001; Forbes et al., 2004; De Raedt et al., 2006a; Steinmann et al., 2007). The positions of the respective breakpoint regions are shown in Figure S1. To ensure good quality of the tumor-derived DNA samples, a control PCR was performed with each tumor DNA (expected size of the PCR product: 3 kb) using primer BO11f (5 0 -atctttccaccacatggtccc-3 0, chromosomal position on 7q34: 141,580, ,580,615; hg18) and primer BO10r (5 0 - ccctcaactggattgaaatcacc-3 0, chromosomal position: 141,583, ,583,452). As positive controls for the deletion junction PCRs, DNAs were used from patients with known type-1 (with breakpoints located within PRS1 or PRS2) and type-2 NF1 microdeletions as indicated in Table S3. Putative breakpoint-junction PCR products amplified from tumor DNA samples were cloned using the TOPO TA cloning system (Invitrogen, Karlsruhe, Germany) and at least 20 clones were sequenced in order to avoid false positive PCR products derived from the proximal or distal duplications (NF1-REPs or SUZ12 sequences). In one case of deletion breakpoint-junction PCR (primer combination hks4033 and jtü8r), a restriction enzyme assay using Tsp45I was performed to determine whether the PCR fragments amplified from the tumor material were derived from an authentic breakpoint-spanning fusion between the SUZ12 pseudogene and the SUZ12 gene or whether the PCR products were instead false positives having been amplified from either the SUZ12 gene or the SUZ12 pseudogene. Multiplex ligation-dependent probe amplification To determine the NF1 gene copy number in the tumors, the MLPA assay SALSA P122 NF1 area (version 01, MRC Holland, Amsterdam, The Netherlands) was used. Tumor DNA samples were analyzed according to the manufacturer s instructions using 200 ng genomic DNA. After hybridization, ligation, and amplification, the PCR products were separated on an ABI Prism 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) by capillary electrophoresis. The relative probe signal was then determined by means of a normalization procedure as described by Wimmer et al. (2006). In those cases where two copies of a sequence are present in a given sample, this process is expected to yield a value of 1.0. A decrease in the peak area values to o0.91 for probes in the NF1 gene region was considered to be indicative of a deletion. CONFLICT OF INTEREST The authors state no conflict of interest. ACKNOWLEDGMENTS We thank Helene Spöri for technical assistance. The work was funded by the Deutsche Krebshilfe grant no SUPPLEMENTARY MATERIAL Table S1. Phenotype information on NF1 patients with plexiform neurofibromas

6 Table S2. Primers and genomic positions of the polymorphic markers on chromosome 17 used for LOH analysis of the plexiform neurofibromas. Table S3. Sequence and genomic locations of primers used for breakpointspanning PCRs with DNA of 43 plexiform neurofibromas. Table S4. Results of the MLPA analysis of 13 neurofibromas exhibiting LOH. Figure S1. Scheme of the NF1 gene region. REFERENCES Balgobind BV, Van Vlierberghe P, van den Ouweland AM, Beverloo HB, Terlouw-Kromosoeto JN, van Wering ER et al. (2008) Leukemiaassociated NF1 inactivation in patients with pediatric T-ALL and AML lacking evidence for neurofibromatosis. Blood 111: Basu TN, Gutmann DH, Fletcher JA, Glover TW, Collins FS, Downward J (1992) Aberrant regulation of ras proteins in malignant tumour cells from type 1 neurofibromatosis patients. Nature 356:713 5 Colman SD, Williams CA, Wallace MR (1995) Benign neurofibromas in type 1 neurofibromatosis (NF1) show somatic deletions of the NF1 gene. Nat Genet 11:90 2 Däschner K, Assum G, Eisenbarth I, Krone W, Hoffmeyer S, Wortmann S et al. (1997) Clonal origin of tumor cells in a plexiform neurofibroma with LOH in NF1 intron 38 and in dermal neurofibromas without LOH of the NF1 gene. Biochem Biophys Res Commun 234: De Raedt T, Maertens O, Chmara M, Brems H, Heyns I, Sciot R et al. (2006b) Somatic loss of wild-type NF1 allele in neurofibromas: comparison of NF1 microdeletion and non-microdeletion patients. Genes Chrom Cancer 45: De Raedt TD, Stephens M, Heyns I, Brems H, Thijs D, Messiaen L et al. (2006a) Conservation of hotspots for recombination in low-copy repeats associated with the NF1 microdeletion. Nat Genet 38: DeClue JE, Papageorge AG, Fletcher JA, Diehl SR, Ratner N, Vass WC et al. (1992) Abnormal regulation of mammalian p21ras contributes to malignant tumor growth in von Recklinghausen (type 1) neurofibromatosis. Cell 69: Dorschner MO, Sybert VP, Weaver M, Pletcher BA, Stephens K (2000) NF1 microdeletion breakpoints are clustered at flanking repetitive sequences. Hum Mol Genet 9:35 46 Ducatman BS, Scheithauer BW, Piepgras DG, Reiman HM, Ilstrup DM (1986) Malignant peripheral nerve sheath tumors. Cancer 57: Eisenbarth I, Beyer K, Krone W, Assum G (2000) Toward a survey of somatic mutation of the NF1 gene in benign neurofibromas of patients with neurofibromatosis type 1. Am J Hum Genet 66: Ferner RE, Gutmann DH (2002) International consensus statement on malignant peripheral nerve sheath tumors in neurofibromatosis. Cancer Res 62: Forbes SH, Dorschner MO, Le R, Stephens K (2004) Genomic context of paralogous recombination hotspots mediating recurrent NF1 region microdeletion. Genes Chrom Cancer 41:12 25 Frahm S, Mautner VF, Brems H, Legius E, Debiec-Rychter M, Friedrich RE et al. (2004) Genetic and phenotypic characterization of tumor cells derived from malignant peripheral nerve sheath tumors of neurofibromatosis type 1 patients. Neurobiol Dis 16:85 91 Friedman JM, Birch PH (1997) Type 1 neurofibromatosis. A descriptive analysis of the disorder in 1728 patients. Am J Med Genet 70: Huson SM, Harper PS, Compston DAS (1988) Von Recklinghausen neurofibromatosis: a clinical and population study in South East Wales. Brain 111: Huson SM, Hughes RAC (1994) The Neurofibromatoses. London: Chapman and Hall, Jenne DE, Tinschert S, Reimann H, Lasinger W, Thiel G, Hameister H et al. (2001) Molecular characterization and gene content of breakpoint boundaries in patients with neurofibromatosis type 1 with 17q11.2 microdeletions. Am J Hum Genet 69: John AM, Ruggieri M, Ferner R, Upadhyaya M (2000) A search for evidence of somatic mutations in the NF1 gene. J Med Genet 37:44 9 Kehrer-Sawatzki H, Kluwe L, Sandig C, Kohn M, Wimmer K, Krammer U et al. (2004) High frequency of mosaicism among patients with neurofibromatosis type 1 (NF1) with microdeletions caused by somatic recombination of the JJAZ1 gene. Am J Hum Genet 75: Kluwe L, Friedrich RE, Mautner VF (1999) Allelic loss of the NF1 gene in NF1- associated plexiform neurofibromas. Cancer Genet Cytogenet 113:65 9 Kluwe L, Siebert R, Gesk S, Friedrich RE, Tinschert S, Kehrer-Sawatzki H et al. (2004) Screening 500 unselected neurofibromatosis 1 patients for deletions of the NF1 gene. Hum Mutat 23:111 6 Legius E, Marchuk DA, Collins FS, Glover TW (1993) Somatic deletion of the neurofibromatosis type 1 gene in a neurofibrosarcoma supports a tumour suppressor gene hypothesis. Nat Genet 3:122 6 López-Correa C, Dorschner M, Brems H, Lazaro C, Clementi M, Upadhyaya M et al. (2001) Recombination hotspot in NF1 microdeletion patients. Hum Mol Genet 10: Lothe RA, Slettan A, Saeter G, Brøgger A, Børresen AL, Nesland JM (1995) Alterations at chromosome 17 loci in peripheral nerve sheath tumors. J Neuropathol Exp Neurol 54:65 73 Lothe RA, Smith-Sørensen B, Hektoen M, Stenwig AE, Mandahl N, Saeter G et al. (2001) Biallelic inactivation of TP53 rarely contributes to the development of malignant peripheral nerve sheath tumors. Genes Chrom Cancer 30:202 6 Maertens O, Brems H, Vandesompele J, De Raedt T, Heyns I, Rosenbaum T et al. (2006) Comprehensive NF1 screening on cultured Schwann cells from neurofibromas. Hum Mutat 27: Messiaen LM, Callens T, Mortier G, Beysen D, Vandenbroucke I, Van Roy N et al. (2000) Exhaustive mutation analysis of the NF1 gene allows identification of 95% of mutations and reveals a high frequency of unusual splicing defects. Hum Mutat 15: Rasmussen SA, Overman J, Thomson SA, Colman SD, Abernathy CR, Trimpert RE et al. (2000) Chromosome 17 loss-of-heterozygosity studies in benign and malignant tumors in neurofibromatosis type 1. Genes Chrom Cancer 28: Rutkowski JL, Wu K, Gutmann DH, Boyer PJ, Legius E (2000) Genetic and cellular defects contributing to benign tumor formation in neurofibromatosis type 1. Hum Mol Genet 9: Sawada S, Florell S, Purandare SM, Ota M, Stephens K, Viskochil D (1996) Identification of NF1 mutations in both alleles of a dermal neurofibroma. Nat Genet 14:110 2 Serra E, Ars E, Ravella A, Sánchez A, Puig S, Rosenbaum T et al. (2001a) Somatic NF1 mutational spectrum in benign neurofibromas: mrna splice defects are common among point mutations. Hum Genet 108: Serra E, Pros E, García C, López E, Gili ML, Gómez C et al. (2007) Tumor LOH analysis provides reliable linkage information for prenatal genetic testing of sporadic NF1 patients. Genes Chrom Cancer 46:820 7 Serra E, Puig S, Otero D, Gaona A, Kruyer H, Ars E et al. (1997) Confirmation of a double-hit model for the NF1 gene in benign neurofibromas. Am J Hum Genet 61:512 9 Serra E, Rosenbaum T, Nadal M, Winner U, Ars E, Estivill X et al. (2001b) Mitotic recombination effects homozygosity for NF1 germline mutations in neurofibromas. Nat Genet 28:294 6 Serra E, Rosenbaum T, Winner U, Aledo R, Ars E, Estivill X et al. (2000) Schwann cells harbor the somatic NF1 mutation in neurofibromas: evidence of two different Schwann cell subpopulations. Hum Mol Genet 9: Spurlock G, Griffiths S, Uff J, Upadhyaya M (2007) Somatic alterations of the NF1 gene in an NF1 individual with multiple benign tumours (internal and external) and malignant tumour types. Fam Cancer 6: Steinmann K, Cooper DN, Kluwe L, Chuzhanova NA, Senger C, Serra E et al. (2007) Type 2 NF1 deletions are highly unusual by virtue of the absence of nonallelic homologous recombination hotspots and an apparent preference for female mitotic recombination. Am J Hum Genet 81: Stephens K, Weaver M, Leppig KA, Maruyama K, Emanuel PD, Le Beau MM et al. (2006) Interstitial uniparental isodisomy at clustered breakpoint 620 Journal of Investigative Dermatology (2009), Volume 129

7 intervals is a frequent mechanism of NF1 inactivation in myeloid malignancies. Blood 108: Stewart DR, Corless CL, Rubin BP, Heinrich MC, Messiaen LM, Kessler LJ et al. (2007) Mitotic recombination as evidence of alternative pathogenesis of gastrointestinal stromal tumours in neurofibromatosis type 1. J Med Genet 44:e61 Stumpf DAJ, Annegers J, Brown S, Conneally P, Housman D, Leppert M et al. (1988) Neurofibromatosis. Conference Statement. National Institutes of Health Consensus Development Conference. Arch Neurol 45: Upadhyaya M, Han S, Consoli C, Majounie E, Horan M, Thomas NS et al. (2004) Characterization of the somatic mutational spectrum of the neurofibromatosis type 1 (NF1) gene in neurofibromatosis patients with benign and malignant tumors. Hum Mutat 23: Upadhyaya M, Kluwe L, Spurlock G, Monem B, Majounie E, Mantripragada K et al. (2008) Germline and somatic NF1 gene mutation spectrum in NF1- associated malignant peripheral nerve sheath tumors (MPNSTs). Hum Mutat 29:74 82 Waggoner DJ, Towbin J, Gottesman G, Gutmann DH (2000) A clinic-based study of plexiform neurofibromas in neurofibromatosis 1. Am J Med Genet 92:132 5 Wiest V, Eisenbarth I, Schmegner C, Krone W, Assum G (2003) Somatic NF1 mutation spectra in a family with neurofibromatosis type 1: toward a theory of genetic modifiers. Hum Mutat 22:423 7 Wimmer K, Yao S, Claes K, Kehrer-Sawatzki H, Tinschert S, De Raedt T et al. (2006) Spectrum of single- and multiexon NF1 copy number changes in a cohort of 1,100 unselected NF1 patients. Genes Chrom Cancer 45: