Activating mutations in FGFR3 and HRAS reveal a shared genetic origin for congenital disorders and testicular tumors

Abstract

Genes mutated in congenital malformation syndromes are frequently implicated in oncogenesis, but the causative germline and somatic mutations occur in separate cells at different times of an organism's life. Here we unify these processes to a single cellular event for mutations arising in male germ cells that show a paternal age effect. Screening of 30 spermatocytic seminomas for oncogenic mutations in 17 genes identified 2 mutations in FGFR3 (both 1948A>G, encoding K650E, which causes thanatophoric dysplasia in the germline) and 5 mutations in HRAS. Massively parallel sequencing of sperm DNA showed that levels of the FGFR3 mutation increase with paternal age and that the mutation spectrum at the Lys650 codon is similar to that observed in bladder cancer. Most spermatocytic seminomas show increased immunoreactivity for FGFR3 and/or HRAS. We propose that paternal age-effect mutations activate a common 'selfish' pathway supporting proliferation in the testis, leading to diverse phenotypes in the next generation including fetal lethality, congenital syndromes and cancer predisposition.

Figure 1

Strategy used to quantify mutation levels at the FGFR3 K650 codon. (a) DNA sequence around the K650 codon, the relative positions of XbaI and BpiI restriction enzyme sites used for selection, the PCR primers used for amplification (black arrows, 1^st set of PCR primers; orange arrows, nested PCR primers), the 9 potential single nucleotide substitutions of the K650 codon (located at cDNA position 1948-1950) and the associated germline defects (HCH, hypochondroplasia; TDII, thanatophoric dysplasia type II; Fam AN, familial acanthosis nigricans; SADDAN, severe achondroplasia with developmental delay and acanthosis nigricans; -, not reported as germline mutation; * indicates that the change has also been reported as a somatic mutation in cancer (Supplementary Table 3). Note that all substitutions at the K650 codon abolish the BpiI recognition site (pink box) and hence are selected by digestion with this enzyme. (b) Flow diagram summarizing the strategy adopted for quantification of mutation levels using tagged-oligonucleotide pooled PCR and massively parallel sequencing (see also Supplementary Fig. 2).

Figure 2

Mutation levels at the FGFR3 K650 codon in sperm and blood quantified by massively parallel sequencing. (a) Mutation levels estimated in reconstruction dilution series of 3 heterozygous control samples (TD3 and TD6, 1948A>G; SAD4, 1949A>T) mixed with normal blood DNA. (b) Estimated levels of the 1948A>G (K650E) mutation in blood (red circles, n = 8) and sperm (blue diamonds, n = 78) in relation to the age of the sample donor. r_s refers to the Spearman’s rank correlation coefficient based on the sperm samples. Vertical bars indicate the 95% equal tail probability interval. (c) Cumulative levels of each of 9 different nucleotide substitutions at the K650 codon in sperm (left) and blood (right) samples. (d) Correlation of the relative mutation levels in sperm with other pathological features of the FGFR3 K650 codon. From left to right are shown the overall proportions of the 9 different nucleotide substitutions at the K650 codon in sperm (this work); the relative proportions of mutations in germline disorders, bladder tumors and seborrheic keratoses, each compiled from the literature (Supplementary Table 3); and the relative level of TK activation by each substitution as measured by in vitro kinase assay. The Pearson coefficient of correlation r between data for sperm mutations and each of the other four datasets is indicated above the corresponding panel. Color coding of mutations is identical to key in (c).

Figure 3

Age distribution in spermatocytic seminomas and immunohistochemical staining of FGFR3 and HRAS. (a) Age distribution of subjects at the time of removal of their spermatocytic seminomas. The mutation positive samples are filled in blue (FGFR3 K650E) or green (HRAS Q61R/K), while the yellow bars indicate the samples without identified mutation. Age data were unavailable for 2 tumors. (b) Representative immunohistochemical staining for FGFR3 (left) and HRAS (right) antibody in samples from spermatocytic seminomas mutant for FGFR3 K650E (2 top rows) or HRAS Q61R/K mutations (remaining rows). The mutation status of each tumor is indicated. Insets show staining of adjacent normal testicular parenchyma demonstrating ongoing spermatogenesis: note that FGFR3 is normally present in the cell membrane and cytoplasm of spermatogonia at the base of the seminiferous tubule, whereas HRAS is found in the nucleus of some basal spermatogonia and in the cytoplasm of primary spermatocytes and round spermatids at the tubular lumen (scale bars, 50 μm).

Figure 4

Pathways and phenotypic consequences of selfish mutations in the testis. (a) Simplified depiction of components of the growth factor receptor-RAS/MAPK signaling pathway (ovals), and associated congenital disorders (boxes). The disorders and proteins for which a paternal age effect has been documented in the causative gene are labeled in blue; red highlights those cases with additional experimental evidence for increased levels of causative mutations in testes. (CFC, cardio-facio-cutaneous syndrome; MEN, multiple endocrine neoplasia; RTK, receptor tyrosine kinase). (b) Proposed consequences of mutations depending on the level of mutant protein activation. The original mutations are rare events that occur randomly in spermatogonial stem cells or progenitors. Strongly activating mutations (left column), such as FGFR3 K650E, lead over time to the formation of large clones within the testis (red ovals), potentially (with additional mutational events) causing spermatocytic seminomas. The high relative enrichment increases the risk of fertilization by mutant sperm, causing lethal germline phenotypes such as thanatophoric dysplasia. For moderately activating mutations (central column), absence of mutations in spermatocytic seminoma suggests that clonal expansions (orange ovals) occur to a lesser degree and are more likely to be self-limiting. These mutations lead to classical congenital disorders. For weakly activating mutations (right, yellow ovals), the limited clonal expansions would be undetectable using current experimental methods, but could contribute to an increased burden of many new, potentially oncogenic, mutations in the next generation.