Melorheostosis: Exome sequencing of an associated dermatosis implicates postzygotic mosaicism of mutated KRAS

Abstract

Melorheostosis (MEL) is the rare sporadic dysostosis characterized by monostotic or polyostotic osteosclerosis and hyperostosis often distributed in a sclerotomal pattern. The prevailing hypothesis for MEL invokes postzygotic mosaicism. Sometimes scleroderma-like skin changes, considered a representation of the pathogenetic process of MEL, overlie the bony changes, and sometimes MEL becomes malignant. Osteopoikilosis (OPK) is the autosomal dominant skeletal dysplasia that features symmetrically distributed punctate osteosclerosis due to heterozygous loss-of-function mutation within LEMD3. Rarely, radiographic findings of MEL occur in OPK. However, germline mutation of LEMD3 does not explain sporadic MEL. To explore if mosaicism underlies MEL, we studied a boy with polyostotic MEL and characteristic overlying scleroderma-like skin, a few bony lesions consistent with OPK, and a large epidermal nevus known to usually harbor a HRAS, FGFR3, or PIK3CA gene mutation. Exome sequencing was performed to ~100× average read depth for his two dermatoses, two areas of normal skin, and peripheral blood leukocytes. As expected for non-malignant tissues, the patient's mutation burden in his normal skin and leukocytes was low. He, his mother, and his maternal grandfather carried a heterozygous, germline, in-frame, 24-base-pair deletion in LEMD3. Radiographs of the patient and his mother revealed bony foci consistent with OPK, but she showed no MEL. For the patient, somatic variant analysis, using four algorithms to compare all 20 possible pairwise combinations of his five DNA samples, identified only one high-confidence mutation, heterozygous KRAS Q61H (NM_033360.3:c.183A>C, NP_203524.1:p.Gln61His), in both his dermatoses but absent in his normal skin and blood. Thus, sparing our patient biopsy of his MEL bone, we identified a heterozygous somatic KRAS mutation in his scleroderma-like dermatosis considered a surrogate for MEL. This implicates postzygotic mosaicism of mutated KRAS, perhaps facilitated by germline LEMD3 haploinsufficiency, causing his MEL.

Keywords: Dysostosis; Hyperostosis; LEMD3; Linear epidermal nevus; Malignancy; Nevus sebaceous; Osteopoikilosis; Osteosclerosis; Scleroderma; TGFβ.

Figure 1. Patient

(A,B) At 14 years-of-age our patient has slight knock-knee deformity from an enlarged and elongated right lower limb. (C) The right foot is disproportionately big in keeping with melorheostosis.

Figure 2. Epidermal Nevus

A) A classic linear epidermal nevus involves his right neck and back. B) This epidermal nevus extends to within his scalp. ​Scleroderma-like Dermatosis:​ C) This extensive dermatosis of his right thigh lies over the melorheostosis present in his right femur. It has clinical and histopathological changes (Figure 6) consistent with the scleroderma-like lesion associated with MEL, and is not an elastoma or collagenoma as found in Buschke-Ollendorff syndrome.

Figure 3

Tissue designations, locations, and characteristics selected for DNA extraction and exome sequencing.

Figure 4. Exome Sequencing and Somatic Variant Detection

A) Sequencing depth and total bases sequenced in the target region by exome sequencing. B) To demonstrate that all samples are derived from the same patient, the variant allele frequencies (VAFs) were plotted for 24 bi-allelic single nucleotide polymorphisms (SNPs) previously identified by Pengelly et al⁽⁴⁴⁾ to discriminate exome samples from multiple individuals. The highly correlated VAFs indicate all samples were derived from the same patient. C) All genes with candidate somatic variants passing manual review (see Methods), along with mutation type and translational effect, are shown for each sample. D) The scatter plots compare VAFs of somatic variants identified in each of the four tissue samples (y-axis) to their respective VAFs in the peripheral blood sample (x-axis). Variants with ≥ 8% VAF are labeled with their respective gene name. KRAS mutations were only identified in the EN and SL samples, as labeled.

Figure 5. Patient’s Skeletal Radiographs

A) Anteroposterior right chest: right ribs 4–9 (arrows) are sclerotic and slightly expanded. The fifth and eight ribs are less involved. B) Postero-anterior left hand: an osteosclerotic lesion in the capitate (arrow) is in keeping with osteopoikilosis (OPK). C) Anteroposterior pelvis: sclerosis affects the entire right iliac bone except near its crest (arrows) and has remarkable, almost “fluffy”, large, expanded, sclerotic areas. There is expansion of the body with an irregular lateral margin. D) Anteroposterior femurs: linear sclerosis in the neck (C) with subtrochanteric cortical thickening, expansion, and medullary sclerosis (arrows) (B and C). E) Anteroposterior right knee: a cluster of osteosclerotic lesions consistent with OPK is in the epiphysis of the distal femur (black ellipse). F) Anteroposterior (left) and lateral (right) left tibia: the mid tibia has sharply demarcated oblique cortical thickening that extends into the medullary cavity (arrow, left panel), and is slightly expanded with anterior bowing (arrows, right panel). Cortical thickening extends both out and into the medullary cavity, which is obliterated in the distal fibula.

Figure 6. Scleroderma-like Dermatosis

A, B) The patient’s scleroderma-like lesion shows a proliferation of normal-appearing collagen bundles within the reticular dermis extending into the subcutis that is C, D) associated with slightly decreased elastic fibers. A) Hematoxylin and eosin, 40×. B) Hematoxylin and eosin, 100×. C) Elastin, 40×. D) Elastin, 100×.

Figure 7. Detailed View Of Relevant Somatic Variants

A) Integrative Genomics Viewer (IGV) screenshot of LEMD3, exon 1, 24 bp heterozygous deletion. B) IGV screenshot of KRAS Q61H heterozygous mutation in EN and SL samples. C) Schematic representations of the KRAS protein structure comparing the KRAS mutation observed in EN and SL samples to common KRAS mutations documented in cancer (COSMIC database). Each blue circle represents one of the top-ranked amino acid changes reported in KRAS.