Overlapping SETBP1 gain-of-function mutations in Schinzel-Giedion syndrome and hematologic malignancies

Abstract

Schinzel-Giedion syndrome (SGS) is a rare developmental disorder characterized by multiple malformations, severe neurological alterations and increased risk of malignancy. SGS is caused by de novo germline mutations clustering to a 12bp hotspot in exon 4 of SETBP1. Mutations in this hotspot disrupt a degron, a signal for the regulation of protein degradation, and lead to the accumulation of SETBP1 protein. Overlapping SETBP1 hotspot mutations have been observed recurrently as somatic events in leukemia. We collected clinical information of 47 SGS patients (including 26 novel cases) with germline SETBP1 mutations and of four individuals with a milder phenotype caused by de novo germline mutations adjacent to the SETBP1 hotspot. Different mutations within and around the SETBP1 hotspot have varying effects on SETBP1 stability and protein levels in vitro and in in silico modeling. Substitutions in SETBP1 residue I871 result in a weak increase in protein levels and mutations affecting this residue are significantly more frequent in SGS than in leukemia. On the other hand, substitutions in residue D868 lead to the largest increase in protein levels. Individuals with germline mutations affecting D868 have enhanced cell proliferation in vitro and higher incidence of cancer compared to patients with other germline SETBP1 mutations. Our findings substantiate that, despite their overlap, somatic SETBP1 mutations driving malignancy are more disruptive to the degron than germline SETBP1 mutations causing SGS. Additionally, this suggests that the functional threshold for the development of cancer driven by the disruption of the SETBP1 degron is higher than for the alteration in prenatal development in SGS. Drawing on previous studies of somatic SETBP1 mutations in leukemia, our results reveal a genotype-phenotype correlation in germline SETBP1 mutations spanning a molecular, cellular and clinical phenotype.

Fig 1. Genetic and clinical characteristics of individuals with germline SETBP1 mutations and Schinzel-Giedion syndrome.

A. Schematic representation of the SETBP1 protein, indicating changes found in SGS and in hematologic malignancies. The residues of the canonical degron are highlighted with arrows. Protein domains of SETBP1 are shown in different colors with green corresponding to three AT hooks, purple to the SKI homologous region, blue to the SET binding domain and orange to a repeat domain (modified from Piazza et al.). B. Sequence alignment of the region containing the degron of SETBP1 (in bold) in human (Uniprot accession number Q9Y6X0), chimpanzee (H2QEG8), mouse (Q9Z180), chicken (A0A1D5PT15), african clawed frog (F6TBV9) and zebrafish (B0R147). The consensus motif for βTrCP1 substrates is shown on top, with φ representing a hydrophobic residue and X any amino acid. Residues in which pathogenic germline mutations have been identified in classic SGS are highlighted in blue, while residues in which novel mutations leading to an atypical form of SGS are shown in green. C. Distinctive facial features encountered in classic SGS (current case 9 at 1,5 years of age). D. Typical question mark-shaped ear observed in current case 18. E. Characteristic hand posture with clenched fingers from current case 16. F. Facial features of current case 27 with a mutation in SETBP1 residue S867 at 4 years of age. Note the clenched fingers. G. Facial features of current case 28 with a mutation in SETBP1 residue E862 at 5 years of age. H. Facial features of current case 29 with a mutation in SETBP1 residue T873 at the age of 23 months.

Fig 2. Functional analysis of SETBP1 mutations identified in SGS.

A. Fluorescence measurements in live HEK293 cells expressing YFP-tagged SETBP1 variants. (*** p<0.001 versus wild-type and all mutants, ANOVA). All SETBP1 mutations studied displayed a statistically significant difference compared to wild-type and to all other mutations. This graph is representative of 3 independent experiments performed, with 6 technical replicates per experiment. Bars represent the standard error. B. Relative expression of SETBP1 protein variants in live HEK293 cells treated with MG132 proteasome inhibitor or vehicle only. Bars represent the standard error. (*** p<0.001, * p<0.05, NS: not significant, Student’s T test and Mann-Whitney U test). C. ΔΔG values for degron-βTrCP1 interaction for all germline mutations reported in SETBP1 per residue (** p<0.01 D868 versus other residues; ANOVA). D. Immunoblot of whole cell lysates of HEK293 cells expressing FLAG-tagged SETBP1 variants probed with anti-FLAG antibody. E. Immunoblot of whole-cell lysates of fibroblasts probed with anti-SETBP1 antibody. Fibroblasts were derived from two cases of SGS, one carrying the I871T variant and the other carrying the D868N variant, as well as from two unrelated controls. In D and E, blots were stripped and re-probed with anti-β-actin antibody.

Fig 3. On average, SETBP1 mutations seen in cancer are more severe than those observed in SGS.

A. Distribution of mutations within the SETBP1 degron in SGS and in hematological malignancies. (** p<0.01, Fisher’s test and Bonferroni correction for multiple testing). B. ΔΔG values for protein stability (x-axis) and degron-βTrCP1 interaction (y-axis) for all mutations reported in SETBP1. The size of each circle is proportional to the frequency of the mutation in each condition. C. Difference in free energy of binding in the interaction between βTrCP1 and the degron of variants arising from germline or somatic SETBP1 mutations compared to that of the interaction between βTrCP1 and the wild-type degron (* p <0.05, Mann-Whitney’s U test). The median is highlighted by an arrow head.

Fig 4. Increased SET protein and proliferation in cell lines of patients with SGS.

A. Immunoblot of whole cell lysates of lymphoblastoid cell lines (LCLs) derived from three SGS patients and two age-matched controls. The blot was probed with anti-SET, before stripping and re-probing with β-actin to confirm equal loading. * denotes full-length SET protein (37kDa), arrowhead denotes cleaved SET protein (22kDa). B. SET protein levels as determined by densitometric analysis of Western blot data and normalized to actin (n = 2 independent experiments). Bars represent the standard error. C. Proliferation assay of LCLs derived from SGS individuals and age-matched controls over 3 days (*** p< 0.001 versus controls, ANOVA). Bars represent the standard error. D. Mean doubling time for LCLs derived from individuals with germline SETBP1 mutations compared to controls over 3 days (* p<0.05 versus controls, ANOVA). The mean of two experiments is shown, bars represent the standard error.

Fig 5. Functional effects of germline SETBP1 mutations and risk of malignancy.

Degron-βTrCP1 interaction ΔΔG for SETBP1 mutations in individuals with SGS who did not develop a malignancy versus those who did (*p<0.05, Mann-Whitney’s U test). The median for each group is marked by an arrowhead. The criteria to be considered negative for the development of a malignancy was either reaching the age of 60 months or dying without developing a malignant tumor or leukemia.