Synonymous mutation in TP53 results in a cryptic splice site affecting its DNA binding site in an adolescent with two primary sarcomas

Abstract

Pathologic variants in TP53 are known risk factors for the development of cancer. We report a 17-year-old male who presented with two primary sarcomas. Germline sequencing revealed a novel TP53 c.672 G>A mutation. Sequencing revealed wild-type TP53 in the parents, and there was no history of cancer in first-degree relatives. This de novo synonymous germline mutation results in a 5′ cryptic splice site that is bound by U1, resulting in a shift of the splice site by 5 base pairs (bp). The frame shift results in a truncated protein at residue 246, which disrupts the DNA binding domain of p53.

Introduction

TP53 is the most frequently mutated gene in human cancer [1], Encoded by 13 exons, this gene has many genetic polymorphisms leading to more than 100 haplotypes and through alternative splicing 12 non-pathogenic protein isoforms [2] [3]. To date, none of the non-pathogenic isoforms have splicing alterations in the exons 6 and 7. Most TP53 mutations associated with cancer result from a missense substitution (73%) leading to an amino acid substitution in the DNA binding domain. Other changes include frameshift insertions and deletions (9%), nonsense mutations (8%), silent mutations (4%), and splice site mutations (2%) (IARC TP53 database) [4], These mutations typically result in a full-length protein with either a gain-of-function or loss of wild-type tumor suppressive function. Exons 4 through 8 of TP53 comprise the DNA binding domain with exons 5 through 7 constitutes a hotspot for cancer-associated mutations.

The classic definition of Li-Fraumeni syndrome includes one proband with sarcoma under 45 year of age, a first-degree relative with cancer prior to the age of 45 years, and a first- or second-degree relative with any cancer before 45 years of age or a sarcoma at any age [5]. Birch and Eeles modified the criteria, leading to the nomenclature of “Li-Fraumeni-like syndrome”, which still requires two relatives with cancer. 7–24% of TP53 mutations are de novo mutations[6][7]. Carriers with TP53 pathogenic variants have a 40% probability of cancer by 20y and greater than 90% by 70 years. These mutations produce an 83-fold risk of developing multiple tumors [8].

We describe an adolescent who presented with two different sarcomas. Sequencing revealed a germline synonymous mutation in the DNA binding domain of TP53, which produces a 5′ cryptic splice site that results in a frame shift in the amino acids translated resulting in a premature termination at amino acid 246, disrupting the DNA binding domain of p53.

Case Report

A 17 year-old fraternal twin male, the product of in vitro fertilization (IVF), presented with hematuria. He was initially diagnosed with pleomorphic sarcoma, most likely a leiomyosarcoma, of the pelvis (Figure 1A+B). During the workup for metastasis, he was found to have a left humeral mass, which was biopsied and determined to be a telangiectatic osteosarcoma (Figure 1C+D). His bone marrow at diagnosis had no evidence of disease. He completed five cycles of ifosfamide/doxorubicin and local radiation, followed by surgical resection, of the pelvic tumor. His disease progressed post-operatively, and he died 13 months from his initial diagnosis. There was no family history of cancer, other than a maternal uncle diagnosed with a glioblastoma multiforme in his seventh decade.

Figure 1.

Histopathology of the two primary tumors (A-D). Hematoxylin and eosin stain. (A, B) The abdominal tumor showed both round cell (A) and pleomorphic spindle cell (B) morphology. No osteoid was identified. Smooth muscle actin immunohistochemical stain was strongly positive, favoring leiomyosarcoma. (C,D) The bone tumor consisted of spindle cells with blood-filled cystic spaces arrow (C) and osteoid production arrow (D), consistent with an aggressive telangiectatic osteosarcoma.

Methods and Results

Following informed consent, the patient and his parents provided peripheral blood lymphocytes for TP53 DNA sequencing, performed by GeneDx. Sequencing of DNA from peripheral blood lymphocytes from the patient, but not his parents, revealed c.672 G>A germline variant preserving the glutamate at p. 224. However, this mutation is located at the last base pair of exon 6. An in silico algorithm (http://in-silico.net/tools/biology/sequence_conversion) predicts that this novel mutation produces a 5′ cryptic splice site at an alternative location.

Total RNA was then extracted from formaldehyde-fixed paraffin-embedded bone marrow and leiomyosarcoma, [9]. RNA was then transcribed into cDNA using Superscript VELO master mix containing random primers (Invitrogen). The RNA yield was low 0.02 μg/pl and 0.01 μg/pl as well as the quality. Short fragments of <150 bp were amplified and the best sets amplified 131 bp using nested PCR. Primers were designed to amplify the region spanning exons 6 and 7 junction (Figure 2A). Amplified fragments were resolved by MetaPhor agarose gel, which demonstrated doublets (Figure 2B). The higher molecular weight band seen in the samples is consistent with intronic retention (arrow in Figure 2B). We believe the mutation is germline due to its presence in the bone marrow where he had no evidence of disease. However, the expression of the mutant mRNA seems to more stable in the bone marrow whereas the doublet in the tumor has less definition. The two amplified fragments were isolated from the gel and sequenced, using Sanger sequencing (GENEWIZ, South Plainfield, NJ). Sequencing of the 131 bp fragment proved challenging. To better delineate the mutation in such a small fragments, the fragments were inserted into the pcDNA3 plasmid (Invitrogen, Carlsbad, CA) resulting in sequencing of the entire fragment. Figure 2C shows the insertion of 5 bp from the intron in the mutated strand. Supplemental figure shows sequencing from the controls samples.

Figure 2.

Genomic organization of TP53 and strategy of primer design. (A) Diagram to illustrate exons 2–11 of TP53 and where the primers are in relation to the mutation. (B) Nested PCR product lane 1 the DNA ladder, lanes 2–5 bone marrow (unaffected tissue) PCR done in quadruplicate, lanes 6–9 Tumor-leimyosarcoma in quadruplicate, lane 10 human cDNA and lane 11 human genomic DNA from an unaffected individual. A doublet is present in the bone marrow and tumor but the tumor sample has shifted up, indicating a larger fragment. (C) Sanger sequence of the tumor depicting cryptic splice site after 5 bp intronic insertion, green is the 5′ of exon 6, red is the mutation, blue are the ends of the intron, purple is the 3′ of exon 7, and all three primers.

Discussion

We performed a search and surveyed the online TP53 data in IARC TP53, NCBI dbSNP, and ExAc for mutations and their effects on the structure of p53 and its DNA binding site. Other patients harbored synonymous mutations G>T and G>C at this site. His novel germline mutations (G>A), we predicted, would result in a splice variant based on the conserved sequences for splicing at the end of the exon, but has not been previously described in the literature. The U1 splicing complex binds to the 5′ intron-exon border based on conserved sequences. A mutation in these conserved sequences can result in a shift of the spliceosome binding site and insertion of intronic sequence into the mRNA transcript. We have found no reports of our patient’s mutation. Review of the literature found a reported family with a G>A at +1 of the intron 6 resulting in an 18 bp insertion[10]. Given the conserved nature of this region and that it is a hotspot for oncogenic mutations, we believe our patient’s pathogenic variant would have truncated the p53 protein in the DNA binding site resulting in a change in the binding affinity to the DNA. Truncation mutations 5′ and 3′ of this site have been reported in adrenocortical carcinoma, sarcomas, and breast cancer [11][12] (http://p53.iarc.fr). Most TP53 mutations result in an amino acid substitution in the DNA binding domain. The mutations result in a full length protein produced with a gain of function. Most splice variants lead to a loss of function due to exon skipping or early terminations.

Interestingly, this de novo pathologic variant occurred in an adolescent, who was a product of IVF, but most evidence points to no increased risk of cancer in children conceived by IVF [13].

Limited availability of tissue prevents us from further investigation whether there are more splice variants in the tumors. The cDNA tumor bands in the agarose gel where not as defined as they were in the unaffected bone marrow. With the limited amount and quality of the tissue we have not been able to answer this question. However based on the bioinformatic analysis, we believe that the de novo synonymous germline mutation in TP53 resulted in a shift of the exon 6 splice site by 5 bp, producing a ffameshift and premature stop codon at residue 246 in exon 7. This pathogenic variant likely served as a driver for the two primary tumors in this patient. A recent report suggests that ~9% of pediatric cancers have a predisposition due to pathogenic variants. Of the genes affected, TP53 is the most common [14]. Thus, identification of TP53 germline mutations are important in pediatric cancers and cancer surveillance. Any child or adolescent with two primary cancers should undergo genetic testing and subsequent genetic counseling.

Supplementary Material

Supplemental Figure S1. depicting the Sanger sequencing results of TP53 from the control cDNA and Genomic DNA.

Abbreviations

bp

base pair

bm

bone marrow