High Frequency of Mosaicism among Patients with Neurofibromatosis Type 1 (NF1) with Microdeletions Caused by Somatic Recombination of the JJAZ1 Gene

Abstract

Detailed analyses of 20 patients with sporadic neurofibromatosis type 1 (NF1) microdeletions revealed an unexpected high frequency of somatic mosaicism (8/20 [40%]). This proportion of mosaic deletions is much higher than previously anticipated. Of these deletions, 16 were identified by a screen of unselected patients with NF1. None of the eight patients with mosaic deletions exhibited the mental retardation and facial dysmorphism usually associated with NF1 microdeletions. Our study demonstrates the importance of a general screening for NF1 deletions, regardless of a special phenotype, because of a high estimated number of otherwise undetected mosaic NF1 microdeletions. In patients with mosaicism, the proportion of cells with the deletion was 91%–100% in peripheral leukocytes but was much lower (51%–80%) in buccal smears or peripheral skin fibroblasts. Therefore, the analysis of other tissues than blood is recommended, to exclude mosaicism with normal cells in patients with NF1 microdeletions. Furthermore, our study reveals breakpoint heterogeneity. The classic 1.4-Mb deletion was found in 13 patients. These type I deletions encompass 14 genes and have breakpoints in the NF1 low-copy repeats. However, we identified a second major type of NF1 microdeletion, which spans 1.2 Mb and affects 13 genes. This type II deletion was found in 8 (38%) of 21 patients and is mediated by recombination between the JJAZ1 gene and its pseudogene. The JJAZ1 gene, which is completely deleted in patients with type I NF1 microdeletions and is disrupted in deletions of type II, is highly expressed in brain structures associated with learning and memory. Thus, its haploinsufficiency might contribute to mental impairment in patients with constitutional NF1 microdeletions. Conspicuously, seven of the eight mosaic deletions are of type II, whereas only one was a classic type I deletion. Therefore, the JJAZ1 gene is a preferred target of strand exchange during mitotic nonallelic homologous recombination. Although type I NF1 microdeletions occur by interchromosomal recombination during meiosis, our findings imply that type II deletions are mediated by intrachromosomal recombination during mitosis. Thus, NF1 microdeletions acquired during mitotic cell divisions differ from those occurring in meiosis and are caused by different mechanisms.

Introduction

Neurofibromatosis type 1 (NF1) is one of the most common autosomal dominant disorders and is caused by defects of the tumor suppressor gene, NF1 (MIM 162200), in 17q11.2. NF1 is characterized by café-au-lait spots, neurofibromas, axillary/inguinal freckling, and Lisch nodules (Huson 1989; Marchuk and Collins 1994). Optic gliomas, bone malformations, and malignant peripheral nerve sheath tumors are also within the clinical spectrum of NF1. Furthermore, patients with NF1 often have problems in visual-spatial tasks, memory, sustained attention, coordination, language, and behavior (Gutmann et al. 1997).

Large deletions of the NF1 gene region have been observed with a frequency of 4.4% and 4.8%, respectively, in two studies screening 500 and 84 unselected patients with NF1 (Cnossen et al. 1997; Kluwe et al. 2004). The majority of NF1 microdeletions span ∼1.4 Mb, and, in ∼50% of these cases, the breakpoints were mapped to low-copy repeats termed “NF1 LCRs” (López Correa et al. 1999, 2001; Dorschner et al. 2000; Jenne et al. 2000, 2001). Thus, these NF1 microdeletions belong to the group of genomic disorders that are caused by a nonallelic homologous recombination (NAHR) between region-specific LCRs. Well-known genomic disorders caused by NAHR are Smith-Magenis syndrome (Chen et al. 1997), DiGeorge syndrome (Edelmann et al. 1999; Shaikh et al. 2000), Williams-Beuren syndrome (Peoples et al. 2000; Valero et al. 2000), Charcot-Marie-Tooth type 1A (CMT1A), and hereditary neuropathy with liability to pressure palsies (Pentao et al. 1992; Chance et al. 1994; Reiter et al. 1996). Predisposition to the deletions or duplications underlying these disorders is caused by LCRs and thus the genomic architecture of the corresponding regions (reviewed by Inoue and Lupski [2002] and Shaw and Lupski [2004]).

The NF1 LCRs, which predispose the NF1 gene region to deletions, have a rather complicated structure. They are composed of multiple pseudogene exons of the WI-12393 (KIAA0563-related) gene and segments with homology to chromosome 19. The proximal NF1 LCR encompasses 110 kb, whereas the distal NF1 LCR spans 74 kb. These LCRs are separated by ∼1.4 Mb (Dorschner et al. 2000; Jenne et al. 2000, 2001, 2003; López Correa et al. 2001). The NF1 LCRs probably originated from several duplication and transposition events during primate evolution, since they are absent in the NF1 gene region of the mouse (Jenne et al. 2003).

In addition to the NF1 LCRs, other duplicated sequences, namely the JJAZ1 gene and its pseudogene, are located in the NF1 gene region. The functional JJAZ1 gene spans 16 exons and maps 20 kb centromeric to the distal NF1 LCR (fig. 1). The JJAZ1 pseudogene, however, encompasses only exons 1–9 and is found 41 kb distal to the proximal NF1 LCR (Jenne et al. 2003). Within 46 kb of homology between the JJAZ1 and its pseudogene, the overall identity is 97%. The JJAZ1 gene and its pseudogene are located in direct orientation to one another. Recently, we identified an NF1 microdeletion with breakpoints in the JJAZ1 gene and its pseudogene (Petek et al. 2003). However, this has been the only deletion with breakpoints in the JJAZ1 gene reported so far. The frequency of JJAZ1-mediated microdeletions, however, has remained unclear.

Figure 1.

Map of the NF1 gene region indicating the duplicated sequences, the genes included in type I and type II deletions, and the probes used for FISH analyses. The black horizontal arrows represent the 14 genes between and within the proximal and distal NF1 LCRs (highlighted by gray rectangles). The functional genes are underlined; pseudogenes are indicated with “P.” The green rectangles represent the regions with homology to segments on chromosome 19. The positions of FISH probes DJ1576/1578, DJHK10/11, and DJ1686/1863 are indicated by black vertical lines. BACs and PACs that build up the contig spanning the NF1 gene region are shown by horizontal bars. Those BACs/PACs marked by asterisks were used as FISH probes. The corresponding accession numbers are RP11-271K11 (GenBank accession number AC005562); CTD-2349P21 (GenBank accession number AC127024); RP11-785C15 (GenBank accession number AC109516); RP11-142O6 (GenBank accession number AC079915); RP11-805L22 (GenBank accession number AC007923); HCIT-307A16 (GenBank accession number AC003041); and RP11-55J8 (GenBank accession number AC015941).

Patients with NF1 and large deletions often suffer from a more severe and atypical manifestation of disease than do the general group of patients who have NF1 with intragenic mutations. The severe and atypical phenotype is characterized by lowered mean IQ or mental retardation, facial dysmorphism, and increased risk for malignant peripheral nerve sheath tumors (Kayes et al. 1994; Wu et al. 1995; Leppig et al. 1997; Tonsgard et al. 1997; Wu et al. 1997; De Raedt et al. 2003). However, it is difficult to distinguish between patients with deletions and those without deletions solely on the basis of clinical observations. Furthermore, the majority of previous deletion studies included only preselected patients with facial anomalies and mental retardation, whereas patients without such clinical features have usually been excluded from the deletion screening. Thus, solid evidence has been missing for the widely accepted association of NF1 microdeletion with facial anomalies and mental retardation. Additionally, it is not known whether deletion breakpoint heterogeneity influences the manifestation of symptoms in the patients.

Recently, we identified 24 NF1 gene deletions by screening unselected patients with NF1 for polymorphic markers and by using intragenic cosmid probes in FISH (Kluwe et al. 2004; L. Kluwe, R. Siebert, S. Gesk, R. Freidrich, S. Tinschert, H. Kehrer-Sawatzki, V.-F. Mautner, unpublished data). Samples of 16 of these 24 patients were available for further analysis presented here. Additionally, we included five further deletions in our investigations. The extent of these 21 deletions was determined by use of completely sequenced and unambiguously mapped BACs, PACs, and cloned PCR products as FISH probes. Furthermore, PCR and sequence analysis was performed to identify and characterize the regions of strand exchange during the recombination underlying the deletions.

Material and Methods

Patients

The main phenotypic features of 22 patients with NF1 investigated in this study are summarized in table 1. All patients fulfill the diagnostic criteria for NF1 (Gutmann et al. 1997). Twelve patients examined here were among the 22 identified by an unselected screening of patients with NF1 performed by use of combined microsatellite marker typing and subsequent FISH (Kluwe et al. 2004). The deletions in patients 1180, 1277, 270, and 284 were identified in an updated unselected screen. Thus, in total, 16 deletions were identified on the basis of unselected screening. We investigated the deletions of six additional patients who were identified by a selective screening because of a severe phenotype (patients Z41, SR, HC, and SB) or during loss-of-heterogeneity (LOH) analysis of neurofibromas from different parts of the body (patient KCD). Patient WB is the mother of patient SB and is mildly affected, whereas the daughter has a severe phenotype. Among the 22 patients indicated in table 1, 20 patients were sporadic cases, and patients 1277 and SB were familial cases. The study protocol was approved by the institutional review board, and all participants provided informed consent.

Table 1.

Physical Features of Patients with NF1 Microdeletions^[Note]

Patient	Age(Years)	Sex	No. ofNeurofibromas	DysmorphicFacialFeaturesa	MentalRetardation	Other Features	Type ofDeletion	Mosaic
442b	25	M	>1,000	Yes	Yesc	High burden of subcutaneous tumors, spinal tumors	I	No
619b	15	F	10	Yes	Yesd	Astrocytoma	I	No
270	21	M	>600	Yes	Yesc	Plexiform neurofibroma	I	No
801b	38	F	>1,000	Yes	Yes	Plexiform neurofibroma	I	No
Z41	9	F	…	Yese	Yesc	Delayed motoric development, ataxia, hamartomas of thalamus	I	No
SR	4	M	1	Yes	Yesc	Large ears, short stature, hyperactivity, hypogonadism, brachydactyly, camptodactyly, kyphoscoliosis	I	No
1180	15	F	…	…	Yes	MPNST	I	No
752b	22	F	…	?	Yes	…	I	No
450b	9	F	…	No	Yesc	…	I	No
284	40	F	…	…	Yes	…	I	No
1277	3	M	…	…	?	Family history	I	No
HC	9	M	>30	…	Yesc	Delayed motoric development, hamartomas of thalamus and globus pallidus	II	No
SB	35	F	>500	Yes	Yes	Kyphoscoliosis, large feet	II	No
WBf	65	F	>20	No	No	…	II	Yes
800b	49	F	>100	No	No	…	I	No
659b	47	F	>1,000	…	No	…	I	Yes
928b	35	F	12	No	No	…	II	Yes
697b	11	F	0	No	No	Slightly handicapped in reading and writing	II	Yes
488b	33	F	220	No	No	…	II	Yes
938b	31	F	3	No	No	MPNST	II	Yes
KCD	34	F	>100	No	No	…	II	Yes
736b	68	F	…	No	No	…	II	Yes

Note.— MPNST = malignant peripheral nerve sheath tumors.

^aDysmorphic features are characterized by hypertelorism, downslanting palpebral fissures, broad nose, and coarse facial appearance.

^bThese 12 patients were identified in a previous study of 500 unselected patients (Kluwe et al. 2004). The deletions in patients 1180, 1277, 270, and 284 were identified in an updated unselected screen.

^cMild mental retardation.

^dSevere mental retardation.

^eDysmorphism in epicanthus.

^fMother of patient SB.

Cell Lines

Somatic cell hybrids were generated by polyethyleneglycol-mediated fusion of thymidine-kinase–deficient mouse cell line B82 and either Epstein-Barr virus (EBV)–transformed lymphoblastoid cells of patient SB or skin fibroblasts of patient KCD. To confirm that the hybrids retained only one chromosome 17, FISH was performed with BAC 142O6 covering the 5′ end of the NF1 gene and BAC 55A13 on 17q24; in addition, analyses were performed of marker D17S849, located in the deleted interval, and markers D17S1880 and D17S1841, located outside the deleted interval.

FISH Analysis

Chromosome spreads were prepared according to standard methods from whole blood samples, after culturing for 72 h in the presence of phytohemagglutinin (PHA) (Invitrogen), which mainly stimulates T lymphocytes. Therefore, we can not rule out potential selection for specific T-lymphocyte populations. After 72 h, colcemid was added and the cells were incubated for 15 min in hypotonic solution (0.075 M KCl). After centrifugation, lymphocytes were fixed with methanol and glacial acetic acid at a ratio of 3:1. The slides prepared from these suspensions were used for FISH analyses of metaphase chromosomes and interphase nuclei. Clones used as FISH probes were purchased from the BACPAC Resource Center and have been characterized elsewhere (Jenne et al. 2003). FISH was also performed with probe DJ1576/1578, which covers the region from positions 67241 to 79694 of PAC 2349P21 (GenBank accession number AC127024); probe DJHK10/11, which spans positions 4242–8373 of BAC 307A16 (GenBank accession number AC003041); and probe DJ1686/1863, which covers positions 98021–104095 of AC090616. These FISH probes were generated by PCR by use of the Expand Long Template PCR System (Roche Molecular Biochemicals) and were cloned with the TOPO TA Cloning Kit (Invitrogen). Two-color FISH was performed by use of probes labeled with either biotin-16-dUTP or digoxygenin-11-dUTP (Roche Diagnostics). Hybridization signals were visualized with fluorescein isothiocyanate–avidin and biotinylated anti-avidin (Vector) or with anti-digoxygenin antibodies coupled to Texas Red (Dianova). Slides were counterstained with diamidinophenylindole (DAPI) and mounted with Vectashield antifade solution (Vector). Probes from the NF1 gene region were labeled with biotin and cohybridized with the digoxygenin-labeled BAC 55A13 (GenBank accession number AC015651) from 17q24. To investigate mosaicism with normal cells, slides were hybridized with biotin-labeled BAC RP11-142O6 and digoxygenin-labeled BAC RP11-55A13. In blood lymphocytes, at least 200 interphase nuclei were evaluated, and, in buccal smears, at least 100 interphase cells were evaluated.

PCR with Somatic Cell Hybrids to Narrow the Deletion Boundaries

PCR was performed with primers P1–P4 (listed in supplementary table A1 [online-only]) to narrow the deletion boundaries by use of somatic cell hybrids that contain only the chromosome 17 with the microdeletion but not the normal chromosome 17 of the respective patient.

Oligonucleotides summarized in supplementary table A2 (online-only) were used for PCR experiments to evaluate paralogous sequence variants (PSVs) (also termed “cis-morphisms”), which distinguish the functional JJAZ1 gene from its pseudogene. PCR products were sequenced by use of an ABI Prism 3100 Genetic Analyzer (Applied Biosystems).

Deletion-Junction PCR to Detect Breakpoints Located in the 2-kb Hotspot Region

To determine the frequency of type I deletions with breakpoints in the 2-kb region of preferred strand exchange described elsewhere (Jenne et al. 2001; López Correa et al. 2001), PCR was performed with primers DJ2290 (5′-TCAACCTCCCAGGCTCCCGAA-3′) and DJ2314 (5′-TTTGCACGTGTGACCTTCCACAG-3′), which detect PSVs between the proximal and the distal LCRs at their very 3′ end. Furthermore, these primers are phosphorothioated at the 3′ residue to avoid correction of the mismatches by the polymerase. As this PCR is not very robust and can lead to false-positive results, PCR products were sequenced to confirm the authenticity of a true deletion-junction fragment.

Deletion-Junction PCR to Identify Breakpoints in the JJAZ1 Sequences

To evaluate whether the breakpoints of other type II deletions map to one of the regions of strand exchange identified, PCR was performed with genomic DNA isolated from peripheral blood samples of patients with type II deletions and with the primers listed in supplementary table A3 (online-only). The primers detect PSVs at the 3′ end and are phosphorothioated at the 3′ residue. Thus, the primers discriminate between the JJAZ1 gene and its pseudogene. PCR was performed by use of the Expand Long Template PCR System (Roche Molecular Biochemicals).

Genotyping to Determine the Mechanism Underlying the Deletions in Patients KCD and WB

Markers proximal to the NF1 gene, D17S1873, D17S1841, D17S975, and D17S1863, as well as markers distal to the NF1 gene, D17S1880, D17S907, D17S1833, and D17S1788, were investigated by 6FAM-labeled primers and capillary electrophoresis by use of an ABI Prism 3100 (Applied Biosystems). Markers RsaI and SNP2 detect SNPs located in exon 5 of the NF1 gene and the RAB11-FIB4 gene, respectively. These latter markers were amplified with primers described by Eisenbarth et al. (2000).

Marker Analysis to Evaluate Mosaicism in Patient 928

Markers D17S2237 and NF1PCR were amplified by PCR with primer pairs 5′-CAAGAAAAGCTAATATCGGC-3′ and 5′-GGAACCTTAAGTTCACTTAG-3′ (D17S2237), and 5′-CCTCCCAAAATGCTGGGATTACAG-3′ and 5′-TGAGAGGCCAAGGTGAGAGAATTGCTGGA-3′ (NF1PCR), from neurofibroma tissue sections of patient 928. Heterozygosity of these markers was detected by electrophoresis by use of an ABI Prism 3100 (Applied Biosystems).

RNA In Situ Hybridization

Pregnant C57Bl/6 mice from overnight matings were killed by cervical dislocation. The embryos and the adult mouse brain were fixed with 4% paraformaldehyde at 4°C. Serial cryosections of 10 μm were mounted on slides. RNA in situ hybridization was performed with antisense and sense riboprobes generated by in vitro transcription with T7 or Sp6 polymerase in the presence of α³⁵S-UTP (Amersham Biosciences) from linearized DNA of IMAGE cDNA clone 6513835 (GenBank accession number BQ964374), which covers exons 8–16 of the mouse Jjaz1 mRNA. Hybridizations were performed overnight as described elsewhere (Schmid et al. 1989).

Sequence Comparisons

Sequences were aligned by use of the BLAST search facilities of the National Center for Biotechnology Information (NCBI) server and the Genetics Computer Group (GCG) software package (Wisconsin package, version 10.2). Inverted repeats were identified by use of the program Dotter (see the Web site of the Karolinska Institutet Center for Genomics Research).

Expression Analyses of the Chimeric JJAZ1 Sequence

RNA was isolated from the hybrid cell lines SB-B9, KCD-3, and IL-39 by use of the RNeasy isolation kit (Qiagen). Reverse transcription was performed with the SuperScript transcriptase (Invitrogen). Primers located in exon 10 (5′-TACGGCTCCTATTGCCAAAC-3′) and exon 14 (5′-GCCATTCAGGATCCTTTTCA-3′) of the JJAZ1 gene were used for RT-PCR. Products were then sequenced to confirm their authenticity.

Results

Characterization of Deletions by FISH and Deletion-Junction PCR

The extent of 21 NF1 gene deletions was determined by FISH on metaphase chromosomes by use of seven BACs and PACs and three PCR products as probes. The position of these FISH probes in the completely sequenced contig of the NF1 gene region is indicated in figure 1. Of the 21 deletions, 13 include the region flanked by BAC 307A16 and PAC 2349P21 (summarized in table 2). In these 13 deletion cases, BAC 271K11 shows a reduced hybridization signal. This BAC spans the proximal NF1 LCR and cohybridizes to the distal NF1 LCR. Thus, the breakpoints of these 13 deletions are located in the NF1 LCRs.

Table 2.

Characterization of the Deletion Breakpoint Boundaries by FISH and Breakpoint-Spanning PCR Analysis^[Note]

	Results of FISH Analysis with BACs/PACs								% of Cells with Deletion in
Patient	271K11	2349P21	785C15	142O6	805L22	307A16	55J8	Deletion-Junction PCR	PeripheralBlood	OtherTissue
442	P	D	D	D	D	D	N	−	100	ND
619	P	D	D	D	D	D	N	−	100	ND
270	P	D	D	D	D	D	N	+	100	ND
801	P	D	D	D	D	D	N	−	100	ND
Z41	P	D	D	D	D	D	N	ND	100	ND
SR	P	D	D	D	D	D	N	+	100	ND
1180	P	D	D	D	D	D	N	−	100	ND
752	P	D	D	D	D	D	N	+	100	ND
450	P	D	D	D	D	D	N	+	100	ND
284	P	D	D	D	D	D	N	+	100	100a
1277	P	D	D	D	D	D	N	+	100	100a
800	P	D	D	D	D	D	N	+	100	ND
659	P	D	D	D	D	D	N	−	96	52a
SB	N	P	D	D	D	P	N	+	100	ND
WBb	N	P	D	D	D	P	N	+	94	ND
HC	N	P	D	D	D	P	N	−	100	ND
928	N	P	D	D	D	P	N	−	100	55a; 80c
697	N	P	D	D	D	P	N	−	97	59a
488	N	P	D	D	D	P	N	−	98	56a
938	N	P	D	D	D	P	N	−	91	80a
KCD	N	P	D	D	D	P	N	+	92	51d
736	N	P	D	D	D	P	N	−	94	59a

Note.— The plus sign (+) indicates positive deletion-junction PCR, as determined by sequence analysis; the negative sign (−) indicates negative deletion-junction PCR, as determined by sequence analysis. D = deleted; N = not deleted; P = partially deleted; ND = not determined.

^aBuccal smears.

^bMother of patient SB.

^cNeurofibroma.

^dSkin fibroblasts cultured from a skin biopsy.

To determine whether the WI-12393 gene is affected by the deletion, FISH was performed with probe DJ1686/1863, covering exon 1 and the promoter region of the WI-12393 gene. No signal was obtained by this probe in all 13 cases, indicating that the corresponding sequence is deleted. Therefore, we concluded that the 13 type I NF1 microdeletions span ∼1.4 Mb and lead to hemizygosity of 14 genes, including the WI-12393 gene.

Breakpoint-spanning PCRs were performed to evaluate how many of these type I deletions have breakpoints in the 2-kb region of preferred strand exchange described elsewhere (Jenne et al. 2001; López Correa et al. 2001). A true deletion-junction fragment was identified in 7 of 12 type I deletions for which DNA was available.

The breakpoints of the remaining eight deletions are not located in the NF1 LCRs, since BAC 307A16 and PAC 2349P21 are only partially deleted and BAC 271K11 does not decrease in signal intensity (table 2). To further narrow the breakpoints of these eight type II deletions, FISH was performed with cloned PCR products DJ1576/1578 and DJHK10/11. In all eight type II deletions, the corresponding sequences were deleted. These results imply that the breakpoints map to the JJAZ1 gene and its pseudogene.

Identification of the Deletion Boundaries in the JJAZ1 Sequences by PCR

In two type II deletions, the breakpoints were identified by use of DNA of somatic cell hybrids that retain only the microdeleted chromosome 17 (cell lines SB-B9 and KCD-3). We also investigated hybrid cell line IL-39, which contains the chromosome 17 with the type II deletion described elsewhere (Petek et al. 2003).

The deletion boundaries were narrowed by PCR amplification of products P1–P4 from the DNA of the hybrid cell lines. The position of the primers with respect to the JJAZ1 gene and its pseudogene is indicated in figure 2. Products P2, P3, and P4 were specifically amplified from all three hybrid cell lines (IL-39, SB-B9, and KCD-3). This indicates that the corresponding regions are not deleted. The region covered by the product P1, however, turned out to be deleted in all three hybrids.

Figure 2.

Position of the deletion breakpoints in patients SB and KCD (investigated in this study) and in patient IL (described elsewhere) (Petek et al. 2003). The precise breakpoints in the JJAZ1 gene and its pseudogene (JJAZ-P) were identified by use of somatic cell hybrids that contain only the deleted chromosome 17 (cell lines KCD-3, IL-39, and SB-B9). Black bars indicate the nondeleted regions; white bars represent the deletion intervals. Exon numbers are given above, and the regions covered by PCR products P1–P4 are marked by gray rectangles.

PSVs (paralogous sequence variants), which distinguish the JJAZ1 gene from the pseudogene, were evaluated to identify the regions of strand exchange. These PSVs were discriminated from polymorphisms by the comparison to published BAC sequences, by sequence analysis of PCR products amplified from three somatic cell hybrids that contain a nondeleted chromosome 17 and by sequence analysis of PCR products amplified from those three hybrids with the recombinant JJAZ1 gene.

In cell line KCD-3, the transition of proximal to distal PSVs, which indicates the region of strand exchange, was identified in intron 4 between nucleotides 29666 and 29776 (BAC 640N20 [GenBank accession number AC090616]) (fig. 3). The deletion breakpoint in hybrid IL-39 was mapped in intron 5 between nucleotides 51928 and 51976. The crossover region of the deletion in patient SB was identified in intron 8 between nucleotides 61762 and 62015. This latter region includes a partial L1 and an Alu element. In contrast, the regions of strand exchange in introns 4 and 5 occurred in single-copy segments.

Figure 3.

Areas of strand exchange in the recombinant JJAZ1 sequences of patients KCD (A), SB (B), and IL (C), as determined by use of somatic cell hybrids that contain only the chromosome 17 with the deletion found in the respective patient. The regions of strand exchange, marked by black rectangles, were identified by the analysis of nucleotide differences between the JJAZ1 gene located distal and its pseudogene located proximal to the NF1 gene. The nucleotides typical for the JJAZ1 gene are given above the respective lines, and those characteristic for the pseudogene are given below. White ovals indicate that the sequence amplified from the hybrid cell line corresponds to the JJAZ1 pseudogene sequence; black ovals denote matches to the functional JJAZ1 gene. Noninformative polymorphisms (sequence differences among individuals) are represented as gray ovals. The numbers indicate the position of the respective PSV according to AC090616 (BAC 640N20).

Deletion-Junction PCR to Amplify across the Regions of Strand Exchange in the JJAZ1 Sequences

To evaluate whether the breakpoints of the other six type II deletions map to one of the identified regions of strand exchange, deletion-junction PCRs were performed. These PCRs amplify the breakpoint regions identified in introns 4, 5, or 8, respectively. Since the breakpoint regions in introns 5 and 8 are separated by only 10 kb, we also investigated the segment between both regions by deletion-junction PCR. None of the remaining six type II deletions was positive for one of these PCRs.

Sequence Analysis of the Breakpoint Regions

The breakpoint regions in introns 4, 5, and 8 of the JJAZ1 gene were screened for the presence of recombinogenic motifs (summarized by Abeysinghe et al. [2003]). Only in the 253-bp segment of strand exchange in intron 8 was such a motif found—namely, a translin binding site.

To identify structural features that might have mediated double-strand breaks during recombination, the JJAZ1 sequences were scanned for inverted repeats. These analyses indicated that the region of strand exchange in intron 5 is flanked by an inverted repeat of 127 bp, with 88% identity. Also, the region of strand exchange in intron 8 is flanked by a small inverted repeat of 75 bp, with 74% identity. However, inverted repeats are not present in the crossover region of intron 4.

The average GC content of the JJAZ1 gene and its pseudogene is 41%. The GC content of the regions of strand exchange in introns 4, 5, or 8 is not above this average.

Expression Analysis of the Recombinant JJAZ1 Fusion Gene

The recombination between the JJAZ1 pseudogene and the functional gene leads to a chimeric gene with the pseudogene at the 5′ end and part of the functional gene at the 3′ end. RT-PCR and sequencing of the respective products indicated that the recombinant JJAZ1 gene is expressed in hybrid cell lines SB-B9, KCD-3, and IL-39. Correspondingly, the fusion transcript of the JJAZ1 pseudogene and the functional JJAZ1 gene is not prone to nonsense-mediated mRNA decay. It is not very likely that the chimeric mRNA is efficiently translated, as stop codons are found in exon 2 of the pseudogene, truncating the ORF prematurely. Additionally, internal methionine codons located in exons 3 and 4 are not embedded in a KOZAK sequence.

Mosaicism in Patients with NF1 Microdeletions

Analysis of peripheral blood, buccal mucosa, skin fibroblasts, or neurofibroma tissue revealed mosaicism for microdeletions in eight cases (table 2 and supplementary table A4 [online-only]). Six of the mosaic microdeletions were found among the 24 patients with deletions who were identified by an unselected screen (Kluwe et al. 2004; L. Kluwe, R. Siebert, S. Gesk, R. Freidrich, S. Tinschert, H. Kehrer-Sawatzki, V.-F. Mautner, unpublished data). This gives a frequency of mosaicism of at least 26% among the 23 patients with no family history for NF1 (6/23). In seven of the eight mosaic deletions, the breakpoints were assigned to the JJAZ1 gene (table 2). A very peculiar form of somatic mosaicism was observed in patient 928. In peripheral blood lymphocytes, the deletion was detected by FISH in 100% of the cells. However, in sections of a dermal neurofibroma, 20% of the nuclei, likely nontumor cells, had two FISH signals of BAC 142O6. Mosaicism was further confirmed in this patient by heterozygosity of two markers located in the NF1 gene in DNA extracted from this tumor. Furthermore, normal cells were detected by FISH in 45% of buccal cells.

These findings prompted us to search for mosaicism in families with a microdeletion in the offspring. Patient WB is mildly affected, whereas her daughter, SB, has a severe manifestation and an NF1 microdeletion. Attentive analysis of blood lymphocytes of patient WB revealed mosaicism for a type II deletion she passed on to her daughter, SB (table 2).

Mechanism Underlying the Deletion in Patients KCD and WB

The recombination mechanism underlying the mosaic type II deletions identified in patients KCD and WB was evaluated by marker analysis (fig. 4). The haplotypes of patient KCD were identified by use of somatic cell hybrids that contain only one chromosome 17. Hybrid-line KCD-101 retains the normal chromosome 17, whereas hybrid KCD-25 contains the normal chromosomes 17 prior to the deletion. Hybrid KCD-3, however, contains the chromosome 17 with the microdeletion. In patient SB, the phase of markers was determined with hybrids that contained either the normal chromosome 17 (SB-B10) or the microdeleted chromosome 17 (SB-B9). Marker analysis in the mother, patient WB, was performed with DNA of an EBV-transformed lymphoblastoid cell line, with 30% normal cells and 70% microdeleted cells. According to parsimony considerations, the deletion in patient KCD, as well as that in patient WB, occurred by intrachromosomal recombination during mitosis.

Figure 4.

Genotype analysis to determine that the type II deletion in patients KCD (A) and WB (B) occurred by intrachromosomal recombination. Eight polymorphic markers that flank the NF1 gene region and are located outside of the deletion interval were investigated. In addition, two markers within the deletion interval (RsaI and SNP2) were analyzed. A, The haplotypes were determined by the analysis of hybrid cell line KCD-3 with the microdeleted chromosome 17, hybrid cell line KCD-25 with chromosome 17 prior to the deletion, and hybrid cell line KCD-101, which retains the unaffected chromosome 17. B, Patient WB shows mosaicism for the type II deletion. The phase of markers was determined in her daughter, SB, by use of hybrid cell line SB-10 with the normal chromosome 17 and cell line SB-B9 with the microdeleted chromosome 17. The gray rectangles highlight the haplotype of the chromosome with the microdeletion.

Expression Analysis of the JJAZ1 Gene by RNA In Situ Hybridization

In view of the paucity of knowledge about the function of the JJAZ1 gene, we performed RNA in situ hybridization by use of mouse tissue sections to identify the sites of highest Jjaz1 expression (fig. 5A–5H). At an early developmental stage (embryonal day [ED] 9,5), Jjaz1 is ubiquitously expressed in the embryo proper and in the labyrinth of the placenta (fig. 5F). During later development, at ED14,5, prominent Jjaz1 expression is noticed in the liver, the lung, and the kidneys, as well as in the developing neocortex in subventricular zones (fig. 5H). In the adult mouse brain, Jjaz1 is highly expressed in the hippocampus, the pyriform cortex, the habenula, the granular cell layer of the cerebellum, and the surrounding Purkinje cells (fig. 5B and 5D).

Figure 5.

Expression pattern of the mouse Jjaz1 mRNA in coronal sections of the adult mouse brain (A–D), horizontal sections of ED9,5 (E and F), and sagittal sections of ED14,5 embryos (G and H). Bright-field images are stained with Giemsa (A, C, E, and G). The corresponding dark-field images (B, D, F, and H) indicate as white areas the regions of highest Jjaz1 expression. Amy = amygdala; CA1–3 = regions of the hippocampus; db = decidua basalis; DG = dentate gyrus; e = embryo; fv = primordium of follicle of vibrissae; ge = ganglionic eminence; gtr = trophoblast giant cell; GrCb = granular cell layer of cerebellum; k = kidney; l = labyrinth; li = liver; lu = lung; MHb = medial habenular nucleus; nc = neocortex; Hy = hypothalamus; sb = submandibular gland; to = upper and lower molar tooth; Pur = Purkinje cells; Pir = piriform cortex; ys = yolk sac; and Zi = zona incerta.

Discussion

Two Types of NF1 Microdeletions

Previous studies showed that, in ∼50% of NF1 microdeletions, the breakpoints are located in the NF1 LCRs (López Correa et al. 1999, 2001; Dorschner et al. 2000; Jenne et al. 2000, 2001). In our sample, 13 (62%) of 21 deletions have breakpoints in the NF1 LCRs (table 2). FISH analyses indicated that the WI-12393 (KIAA0563-related) gene is included in the deleted region in the 13 type I deletions. Thus, all of the 13 type I deletions span 1.4 Mb and encompass 14 functional genes.

Our study further revealed a second major type of NF1 microdeletion. In 8 (38%) of the 21 deletions, the breakpoints occurred in the JJAZ1 gene and its pseudogene (table 2). These type II deletions span 1.2 Mb and encompass only 13 genes, since the functional WI-12393 gene is retained (fig. 1). It is interesting that seven (88%) of eight type II deletions are mosaic deletions and are found in female patients.

Somatic Mosaicism in NF1 Microdeletions

Among the 24 deletions identified in a previous (Kluwe et al. 2004) and updated screening of unselected patients with NF1 (L. Kluwe et al., unpublished data), 6 were found to be mosaic deletions in this study, giving a minimum frequency of mosaicism of 26% (6/23 cases without family history). Until now, mosaicism of NF1 microdeletions has been observed in only a few cases (summarized in Petek et al. 2003). Rasmussen et al. (1998) detected two mosaic NF1 microdeletions by LOH analysis and suggested that mosaicism might be more frequent than was previously assumed. The most obvious reason for the underestimation of mosaicism is that deletion screening is mostly restricted to patients with mental retardation and facial dysmorphism. However, none of the patients with mosaic deletions in our study had these features. These deletions would not have been identified by restricted screens. Our recent screening included all patients, even those with no clinical indication of an NF1 microdeletion, which enabled the identification of patients with mosaicism.

Another reason for the underestimation of mosaicism in NF1 microdeletions might be that the number of cells with the deletion is extremely high in peripheral blood, the tissue most commonly used for deletion analysis. Our results demonstrate that, in comparison to peripheral blood, normal cells are far more frequent in other peripheral tissues, like buccal smears or skin fibroblasts. An extreme case of mosaicism is patient 928. The deletion was detected by FISH in 100% of the lymphocytes. However, normal cells were identified in 20% of neurofibroma cells and in 45% of buccal cells. These findings suggest that hematopoietic stem cells with an NF1 microdeletion have a significant growth advantage, compared with normal cells.

Positional Preference of NAHR in NF1 Microdeletions

Specific sequences or genomic architectural features predispose to NAHR between duplicated sequences (for review, see Inoue and Lupski 2002; Stankiewicz and Lupski 2002; Lupski 2003; Shaw and Lupski 2004). In some of the LCR-mediated genomic disorders, hotspots of positional preference for strand exchange during recombination were identified. For instance, a 557-bp region in the 24-kb CMT1A-REPs in 17p11.2 is involved in ∼78% of crossovers in CMT1A (Reiter et al. 1998). In Williams-Beuren syndrome, >27% of deletion breakpoints occurred in a 12-kb region of the much-larger LCRs in 7q11.23 (Bayes et al. 2003). Recently, Bi et al. (2003) observed that, despite ∼170 kb of high homology, 50% of the recombinant junctions of the 4-Mb deletions in Smith-Magenis syndrome map to a 12-kb region. As predicted by reciprocal recombination, junction fragments in patients with dup(17)(p11,2p11.2) were also mapped to this hotspot region.

Positional preference for recombination was also found in NF1 LCR–mediated microdeletions in a segment of 2 kb (Jenne et al. 2001; López Correa et al. 2001). This hotspot is confirmed in our study, since 58% of the type I deletions had breakpoints in this interval. Our results further indicate that, in addition to the NF1 LCRs, the JJAZ1 gene is also a preferred region of NAHR leading to NF1 microdeletions. In three JJAZ1-mediated type II deletions, the regions of strand exchange were identified in introns 4, 5, and 8 of the JJAZ1 gene, respectively (figs. 2 and 3). The breakpoint regions in introns 5 and 8 are separated by only 10 kb.

Structural Motifs or Sequence Elements That Stimulate Strand Exchange between the JJAZ1 Sequences

DNA motifs—like χ elements, minisatellite core sequences, and several other motifs summarized by Abeysinghe et al. (2003)—are well known to promote recombination. We screened the breakpoint regions in JJAZ1 introns 4, 5, and 8 for such motifs. Only in the region of strand exchange in intron 8 was a recombinogenic motif identified—namely, a translin binding site. Translin is known to bind to single-stranded ends of staggered breaks during DNA repair (Sengupta and Rao 2002). Translin binding sites were found to be significantly overrepresented near both translocation and deletion breakpoints (Abeysinghe et al. 2003). It remains to be experimentally proven whether strand exchange is indeed promoted by this mechanism.

Among the structural genomic features that are able to stimulate recombination are inverted repeats. They cause double-strand breaks by forming hairpin-like structures (Gordenin and Resnick 1998), thus initiating nonhomologous recombination. The breakpoint region in intron 5 is flanked by an inverted repeat of 127 bp. Also, the region of strand exchange in intron 8 is flanked by an inverted repeat of 75 bp. Hypothetically, these inverted repeats form hairpin loops that cause double-strand breaks during the recombination process. It has been suggested that a certain length of perfect identity of 300–500 bp is required for effective NAHR (Reiter et al. 1998; López Correa et al. 2001; Bi et al. 2003). To investigate a possible correlation of strand exchange and regions of perfect identity, we compared the proximal and distal JJAZ1 sequences by pairwise alignment. In the regions of strand exchange in introns 4 and 5, no segments of perfect identity larger than 100 bp were found. In the region of crossover in intron 8, however, a segment of 252 bp, with 100% identity, was detected.

In summary, several sequence motifs or structural features were identified that might facilitate strand exchange in the JJAZ1 sequences. However, none of the identified crossover regions reveals conspicuous features like a high density of recombinogenic motifs or inverted repeats of 2 kb. Such features have been identified in the breakpoint regions of SMS-REP–mediated deletions and duplications (Bi et al. 2003). Deletion-junction PCRs that amplify across the breakpoints in JJAZ1 introns 4, 5, or 8 were negative in the other six patients with type II deletions. Thus, a highly restricted positional preference of NAHR in the JJAZ1 gene is not very likely. The multitude of polymorphisms, however, found in introns 4, 5, and 8 indicates that allelic nonaberrant recombination occurs quite commonly in the respective regions (fig. 3).

Positional Preference for Somatic Nonallelic Recombination in the JJAZ1 Gene

Conspicuously, the breakpoints of seven of the eight mosaic deletions are located in the JJAZ1 gene and its pseudogene. Thus, the JJAZ1 gene is a preferred target of strand exchange during somatic NAHR, leading to NF1 microdeletions. Although the NF1 LCRs are also involved in somatic recombination, as indicated by mosaicism of a type I deletion in patient 659, mitotic NAHR seems to be much more frequent in the JJAZ1 gene and its pseudogene. The reason for this preference is unknown.

The function of the JJAZ1 gene, which is completely deleted in patients with type I NF1 microdeletions and is disrupted in deletions of type II, is not well characterized. It encodes an 803-aa protein with a zinc-finger domain and has extensive homology to the Drosophila protein Suppressor of zeste (Suz12), which belongs to the Polycomb group of proteins. It is interesting that JJAZ1 is highly expressed in the cerebellum, the Purkinje cells, the pyriform cortex, and the pyramidal cells of the hippocampus in the mouse brain (fig. 5). Neurons in these structures are characterized by the highest synaptic activity in the whole brain (Malenka and Nicoll 1999; Bliss et al. 2000). The hippocampus is an essential component in cognitive function and learning. The cells of the pyriform cortex are linked to the formation of the amygdala, which has been associated with a range of cognitive functions, including emotion, learning, memory, attention, and perception (Gallagher and Chiba 1996). Therefore, the JJAZ1 gene is a candidate gene whose haploinsufficiency contributes to mental impairment in patients with constitutional NF1 microdeletions.

The JJAZ1 gene has recently been shown to be disrupted by the somatically acquired translocation t(7;17) in endometrial stromal tumors. This translocation leads to the fusion of the JJAZ1 gene in 17q11.2 with the JAZF1 gene on chromosome 7, which also encodes for a protein with a zinc-finger domain (Koontz et al. 2001). The breakpoint of t(7;17) was mapped to intron 1 of the JJAZ1 gene by Koontz et al. (2001). Although the breakpoints in introns 4, 5, and 8 identified in NF1 microdeletions in the present study are located more distally, these data underscore that the JJAZ1 gene is a preferred target for somatic aberrant recombination.

Recombination Mechanism Underlying Type I and Type II Deletions

The type II deletions identified in patients KCD and WB were mediated by intrachromosomal recombination (fig. 4). Additionally, the JJAZ1-mediated deletion in patient IL arose by intrachromosomal NAHR (Petek et al. 2003). In contrast, the majority of type I deletions are generated by interchromosomal recombination between the NF1 LCRs during germ-cell development (López Correa et al. 2000). Thus, the underlying mechanism differs between type I and type II NF1 microdeletions.

Genotype/Phenotype Correlations in Patients with NF1 Microdeletions

Our study clearly indicates breakpoint heterogeneity among patients with NF1 microdeletions. The 1.4-Mb–spanning type I deletion includes the functional WI-12393 gene, which is retained in the smaller type II deletion of 1.2 Mb. Mental retardation in patient SB, who has a type II deletion, shows that WI-12393 hemizygosity does not play a key role in this special phenotype. Nevertheless, the clinical spectrum of patients with type I and type II deletions may differ. Further investigations with larger numbers of clinically well-characterized patients with NF1 microdeletions are recommended to address this issue. Furthermore, somatic mosaicism is a hitherto-unappreciated factor that influences genotype/phenotype correlations. None of the patients with mosaicism for an NF1 microdeletion investigated in our study is mentally retarded or has facial dysmorphism. Normal facial configuration and intelligence was also found in patient 800, who had a type I deletion in 100% of the lymphocytes. We also suspect mosaicism in this exceptional patient, but somatic tissues other than blood were not available for further analyses.

In our study, a total of nine patients with deletions did not exhibit the mental retardation and facial dysmorphism that was believed to be strictly associated with NF1 microdeletions. These patients would not have been identified if the deletion screen had been restricted to patients with such phenotypes, as was the case in most previous deletion studies and is still the case in some laboratories. Our results suggest a general screening for NF1 microdeletions, regardless of the presence of special phenotypes.

Appendix A

Table A1.

Oligonucleotides Used to Amplify PCR Products P1–P4

		Forward		Reverse
Product	Position(nt)	Primer Name	Sequence (5′→3′)	Primer Name	Sequence (5′→3′)
P1	13219–13534a	HK17s5	TGGCAACTTACTTCACATTTGTTT	HK17s6	AGCATGAGAACCATGGAACA
P2	74884–75463a	HK17s1	CAAGGCAACAAACTGAAGCA	HK17s3	TAAAATAGGCCAGGCACACC
P3	80374–81158a	HK17s4	TTTTCTTGGGGAGGGAAAAT	HK17s2	TACTCAACCACAGTGCTCGG
P4	13895–14090b	HK17s7	TGGTCCAGGAATAGCAGGAG	HK17s8	CTGTCCCATCTCAGGCTGTC

^aPosition on BAC 640N20 (GenBank accession number AC090616).

^bPosition on BAC 2349P21 (GenBank accession number AC127024).

Table A2.

PCR Products Used to Evaluate PSVs in the JJAZ1 Gene and Its Pseudogene

	Primer Sequence
PCRProduct	Forward	Reverse	Position(nt)a
int3-1	TAGGCTAACACTGCCCCAAC	CTATTGAGGAGGGCCATGAA	26593–27064
int3-2	TTTGGTGTCT GCATTCAAGC	TTTGGTGTCTGCATTCAAGC	27786–28100
ex4	AGAGAAAATGAAAGGAGAGCAA	GGAGGTATTCAAGCCACAGC	28839–29076
int3-3	CGTTTTAGATGGGGCTTGAA	TCAAGGTGAT CACTAGGAAAATGA	28598–28958
int3-5	TCAGAAACTTGAACATTTGGACA	AAAACTAGGAGGAAAAGAAAAATGC	29375–29668
int4-4n	TCACAATGCATTTTTCTTTTCC	GTGCGTGCACTGCGAATA	29661–30067
int4-3	TTCCATGTCTAGGCAATCCTG	TTTCAAACCCAATGTCACAA	29924–30371
int3-4	TCATTGTGACATTGGGTTTGA	GCCAACTCACTGGTCCTAGC	30349–30564
int4-2	ATTCATGTTCTTTTAGCACTGCC	TTCTATCAGCTCCTTCCACCA	32153–32451
int4-1	CACACACACATCATCCACCA	AACAGCCAGATCCCACAGAG	40399–40724
ex5	TGTCATATTTAACAGGGAAAGCA	CACTCAGAAGCCCCACTTTT	46977–47276
int5-1	CCTGGGTAGCATTATGATTTTTG	TGGAAACAGATATCAACAGAGTTCA	49639–49971
int5-6	ACGTTGGTCAGGTTGGTCTC	CCCACTATGTTGCCCAGACT	50346–50837
int5-3	TTGGCAATTAATACTTGAAACTGC	TTGGGGGTGTGTGTGTAAAA	51108–51291
int5-7	GTGAAGCCCCGTCTCTACTG	GCACTGGGCAACACATTTTT	51446–51840
int5-2	TTTTTGTATGTTCATGGATTCATCT	AGACGCTTTCCCCAGAATTT	51858–52146
int5-5	GAGAAATTCTGGGGAAAGCA	GCAGGAGACAACCACATTCA	52143–52417
int5-4	GAGTTTGCTGTAAGTTTGGAACTG	AACAAGTCAGGGAACTTCAAACA	53979–54058
ex6	TGTGCCACTGCACTGTACCT	TGGCAAACTTTCACAAGCAG	53846–54218
ex7	CATTTGGTGATCATGGCCTA	GCACCATGACTATCATGAAGGA	56366–56805
ex8	TGTTAAGGGACCATATATCGTAACAC	GAGCCTCGCTTGTTCATCA	57448–57640
int8-1	CAGGATGGCAAGAAACACAG	CGCAGCCTGGTACAAGAAA	58947–59100
sb1/2	ATTAGAATGAGGCGCATTGG	TCCCCATCTTGTCATTTTTGA	59614–62146
sb5/6	TGTGACATTTTTCCCCCAGT	GGAGCAGAGAACGAATGCTC	60520–60952
sb7/8	GCATTCCTTT CCCTAGCTTTG	GGAGAAGGTGATTCTGGTATATCTTA	61429–61404
sb9/19	CTCAGCTCACCGCTACCTCT	GCACTCAGCCCTCTGTGAG	61516–61875
sb10/11	TGGAAGACCAAGGTGAGAGG	TCCCCATCTTGTCATTTTTGA	61924–62186
int8-2	AACCAATATTCAAAAGCCTTATAGGA	GTGGTCTTTTAAAATTAGTATGACCTG	62068–62241
sb3/4	TGACAAGATGGGGAGAAATTA	CCATTCTTCAACCATGCTCA	62173–63012
int8-3	TGAAGAGAAGCCAGTTTCTTGAT	CACACAATATAGTTTTAAGGTGTGTG	62972–63100

^aNucleotide numbers according to BAC 640N20 (GenBank accession number AC090616).

Table A3.

Primers Used to Amplify Deletion Junctions in the JJAZ1 Introns 4, 5, or 8 and Flanking Regions

Primer	Sequence (5′→3′)	Positiona	PCR Product Size(bp)
int4 (forward)	CTGTTTCAAATGGTTAATATTCAGT	26914
int4b (reverse)	AGTTGCTATTGAACCCACCA	30486	3,572
int5b (forward)	CTGGTCTCAAGTGATCCTG	49457
int5 (reverse)	TAGAGACAGAGCTATCAGAG	52193	2,736
int5-il7 (forward)	TACTTTTGACTAGCAGTGTATTATCATAAATGTTG	52791
int6-il8 (reverse)	TCATATCATCACTTGTTGTAATGTGTATGGA	55565	2,774
int6-il13 (forward)	TTAGCTATTTGTACCCATTGACATTTTAGATAAGA	54538
int7-il14 (reverse)	AGGGATGGGCAGGGGGATAAATACA	56980	2,442
int8b (forward)	ACACGTTGGAAGAAATAATCTTTGACTCTT	57195
int8 (reverse)	AAAATTAGTATGACCTGTTTACATTCAAGAATGAC	62257	5,062

^aPosition on BAC 640N20 (GenBank accession number AC090616).

Table A4.

FISH Analyses of Interphase Nuclei with BAC 142O6 (Green Signal) and BAC 55A13 (Red Signal)

		No. of Cells with
Patient andTissue	Na	1 Green/ 2 Red	2 Green/2 Red	2 Green/4 Red	4 Green/4 Red	No Signal	Ambiguous	% of Cells with Deletion
442:
Blood	208	208	0	0	0	2	4	100
619:
Blood	200	200	0	0	0	0	3	100
270:
Blood	210	210	0	0	0	3	0	100
801:
Blood	200	200	0	0	0	1	2	100
Z41:
Blood	220	220	0	0	0	1	3	100
SR:
Blood	200	200	0	0	0	0	2	100
1180:
Blood	203	202	0	1	0	2	2	100
752:
Blood	240	240	0	0	0	5	4	100
450:
Blood	200	200	0	0	0	4	2	100
284:
Blood	250	250	0	0	0	0	0	100
Buccal smear	100	99	0	1	0	4	2	100
1277:
Blood	223	223	0	0	0	5	2	100
Buccal smear	100	100	0	0	0	3	0	100
800:
Blood	300	300	0	0	0	7	4	100
659:
Blood	202	194	8	0	0	2	1	96
Buccal smear	103	54	49	0	0	9	2	52
SB:
Blood	200	200	0	0	0	0	0	100
WB:
Blood	200	188	12	0	0	2	2	94
HC:
Blood	200	200	0	0	0	1	3	100
928:
Blood	200	200	0	0	0	0	2	100
Buccal smear	101	56	45	0	0	4	1	55
Neurofibroma	215	172	43	0	0	0	7	80
697:
Blood	209	203	6	0	0	1	4	97
Buccal smear	102	60	42	0	0	5	5	59
488:
Blood	200	195	4	1	0	4	7	98
Buccal smear	102	55	45	2	0	5	3	56
938:
Blood	220	200	20	0	0	3	4	91
Buccal smear	100	80	20	0	0	3	7	80
KCD:
Blood	200	184	16	0	0	3	2	92
Skin fibroblasts	160	72	73	10	5	1	2	51
736:
Blood	210	197	13	0	0	4	1	94
Buccal smear	109	63	45	1	0	5	3	59

^aIndicates the number of cells evaluated and does not include those nuclei without signals or with ambiguous signals.