To the Editor: We read, with great interest, the report by Delnatte et al.1 published in the June 2006 issue of The American Journal of Human Genetics. The authors describe four patients who presented with a severe form of juvenile polyposis of infancy (JPI), with interstitial deletions of chromosome 10q23 encompassing the BMPR1A and PTEN genes. The authors hypothesize that this genetic defect is specific to JPI and that such a severe phenotype reflects cooperation between these two tumor-suppressor genes. We report a patient who had a similar interstitial deletion of chromosome 10 that was, however, associated with a significantly milder phenotype. The patient, an 8-year-old girl, was born in the 36th gestational week to nonconsanguineous parents with a family history negative for JPI. Birth weight was 2,850 g (75th percentile), length was 50 cm (75th–90th percentile), and head circumference was 34 cm (50th–75th percentile). Perinatality and early development were normal. Mild developmental delay was noticed during the 2nd year of life. She was first seen at our clinic at age 3 years and 7 mo. Head circumference was 49 cm (−0.5 SD), weight was 14.5 kg (10th–25th percentile), and height was 97.5 cm (50th percentile). She had mildly dysmorphic features: sparse hair, epicanthic folds, hypoplastic nasal bone, protruding lower lip, small ears, and high-arched palate, with normal tooth development. A single 1.5×1.5–cm cafe-au-lait spot on the left thigh was present; no other cutaneous lesions were noticed. Heart sonogram revealed a small atrial septal defect. Results from electroencephalogram, brain magnetic resonance imaging, audiometric tests, and funduscopic examination were normal. Our patient had never complained of gastrointestinal symptoms, and she did not have her first episode of mild rectal bleeding until age 5 years. Colonoscopy performed at age 6 years detected multiple (>15) polyps throughout the entire length of the colon. Histology of resected lesions was compatible with juvenile polyps, and the largest polyp also had areas of low-grade focal dysplasia in the head of the lesion (fig. 1). Upper-gastrointestinal-tract endoscopy and videoendoscopic examination of the small intestine were normal. She never experienced major bleeding, iron-deficient anemia, or diarrhea. Figure 1. Histology of resected polyps. Hematoxylin and eosin stain is shown at 50× (A and C) and 200× (B and D) magnification. A, Sessile-peduncolated lesion with features of a typical juvenile polyp, characterized by large cysts containing mucus. ... A high-resolution karyotype detected an interstitial deletion of chromosome 10q23. To define the deletion at the molecular level, we proceeded with FISH experiments, using BAC clones spanning the region from 10q22 to 10q23 (table 1 and fig. 2). The final interpretation of the rearrangement was a 12-Mb deletion from 10q22.3 (BAC clone RP11-36D19, at 82.1 Mb) to 10q23.3 (BAC clone RP11-469M1, at 94.4 Mb). We genotyped eight STS loci in the 10q region of the patient and of her parents. Results are summarized in table 2. The deletion was de novo and involved the paternal chromosome. Analysis of the deleted region revealed that it contains 87 genes, many still uncharacterized. BMPR1A, PTEN, and FAS are the most notable genes involved. Figure 2. FISH analysis. Probes RP11-36D19 (A), RP11-74B2 (B), and RP11-469M1 (C) show weaker signals at the breakpoints (A and C) and absence of signal on del(10) (B). Short arrow = normal chromosome 10; long arrow = del(10). Table 1. BAC Clones Spanning the Region from 10q22 to 10q23 Used in FISH Experiments[Note] Table 2. Results of Genotyping of STS Loci in the 10q Region[Note] Juvenile polyposis syndrome (MIM 174900) is composed of three main subtypes: JPI, juvenile polyposis coli, and generalized juvenile polyposis.2 It is still debated whether these three subtypes represent different clinical entities or are variants of the same clinical form.3 Delnatte et al.1 suggest that JPI is a contiguous-gene syndrome and that the deletion of both PTEN and BMPR1A is the specific genetic defect of this form. They propose that these two genes have a synergistic effect on the development of this severe phenotype. The clinical picture of our patient, who harbors a similar deletion, is remarkably different from those of the four patients described by Delnatte et al.1 and from that of the patient described by Tsuchiya et al.,4 all of whom have a proven deletion of both genes. Our patient did not have any of the features of JPI (i.e., onset before age 2 years, severe bleeding, diarrhea, protein-losing enteropathy, inanition, and rectal prolapse5), and polyps were found only in the colon. Three of the four patients reported by Delnatte et al.1 underwent colectomy at ages 10 mo, 17 mo, and 8 years because of the severity of symptoms, such as diarrhea and bleeding. Although our patient is still oligosymptomatic, we are now considering prophylactic colectomy, because of the foci of dysplasia found in one polyp. Furthermore, our patient did not display any of the clinical signs typical of Cowden syndrome (CS [MIM 158350]) or Bannayan-Riley-Ruvalcaba syndrome (MIM 153480)—in particular, she did not have macrocephaly, cutaneous lesions, or lipomas,6 and only mild and aspecific dysmorphic signs were present. Even though some of the features of CS, as well as extracolonic polyps, may appear at a later age, the clinical picture of our patient underscores the phenotypic variability of 10q23 deletions. There are several possible explanations for this phenomenon. It may reflect the ample phenotypic variability of PTEN mutations (also intrafamilial), which is well documented.6 Conversely, a rare haplotype at either the nondeleted PTEN or BMPR1A locus in our patient might be associated with a high level of expression of the remaining allele. Other genetic modifier loci may also play an important role in determining the phenotype. Our patient harbored a deletion that extended from 82.1 to 94.4 Mb, which was larger than those of the patients reported elsewhere (fig. 3). The two deletions fully characterized by Delnatte et al.1 extended from 88.4 to 90.4 Mb (patient 1) and from 88.5 to 90.6 Mb (patient 2), whereas the deletion in the patient reported by Tsuchiya et al.4 ranged from 87.8 to 92.5 Mb. One can speculate that deletion of one or more genes in the 82.1–87.8-Mb and 92.5–94.4-Mb regions might contribute to the milder phenotype in our patient; however, this hypothesis seems unlikely, and we could not identify any obvious gene with such a role within these regions. The size of the deletion, however, may be related to the presence of mild mental retardation. This feature was present in our patient and in the patient who harbored larger deletions, reported by Tsuchiya et al.,4 whereas patients 1 and 2 with smaller deletions, reported by Delnatte et al.,1 had normal intelligence. To clarify these issues, a larger number of patients must be characterized at the molecular level. Finally, the parental origin of the deletion does not seem to contribute to the clinical presentation, since Delnatte et al.1 report both paternal and maternal deletions in their series of patients with a severe phenotype. Figure 3. Schematic representation of the 10q22-10q23 deletion in our patient (A), in patients 1 (B) and 2 (C) reported by Delnatte et al.,1 and in the patient (D) reported by Tsuchiya et al.4 Variable expressivity is not uncommon for microdeletion syndromes. An example is DiGeorge/velocardiofacial syndrome, which is associated with deletions of the 22q11.2 region.7 Patients display a marked clinical variability, even within the same family.8 This variability is probably due, in part, to nongenetic factors, as demonstrated by studies of identical twins.9 The 10q23 microdeletions probably behave in a similar fashion. Interestingly, our patient’s deletion included FAS, a gene with a pivotal role in regulating apoptosis. Analysis of the deletion breakpoints of the patient with severe JPI reported by Tsuchiya et al.4 reveals that FAS was also deleted. Mutations in FAS have been associated with a fatal autosomal dominant lymphoproliferative disorder (autoimmune lymphoproliferative syndrome [MIM 601859]).10 Our patient did not show any of the early features of this disorder (i.e., lymphadenomegaly, splenomegaly, hemolytic anemia, and hypergammaglobulinemia), and her immune function is essentially normal. No immune abnormalities were reported by Tsuchiya et al. either.4 These data suggest that FAS haploinsufficiency is not associated with a specific phenotype and confirm that heterozygous mutations in FAS probably act through a dominant negative mechanism.10 We are currently evaluating in depth the immune function of our patient, to unveil minor abnormalities of lymphocyte function. In conclusion, we believe that the phenotypic spectrum of the deletion of both PTEN and BMPR1A is not restricted to JPI and that further clinical and experimental data are needed to confirm the postulated cooperative mechanism between these two genes. To avoid ascertainment bias, deletion screening should not be limited to patients with JPI but should be performed also in patients with less severe phenotypes.