Whole exome sequencing in relapsed or refractory childhood cancer: case series

Four patients diagnosed with relapsed or refractory childhood cancer and with available tissue samples for DNA extraction were included and underwent WES. One was affected with hematologic malignancy (relapsed acute lymphoblastic leukemia [ALL]), and 3 had solid tumors (relapsed osteosarcoma, recurrent Wilms’ tumor, recurrent colorectal cancer [CRC]). The clinical manifestations and tumor pathology are summarized in Table 1. WES was performed in both tumor and normal samples from all 4 patients. Six variants were identified from the tumor samples obtained from all patients (Table 2). WES of the normal samples revealed no significant cancer-related variants in all cases according to the American College of Medical Genetics and Genomics/Association for Molecular Pathology (ACMG/AMP) guidelines for variant classification and the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists Recommendations [5, 6].

Patient 1 was a 12-year-old girl with relapsed ALL. She presented with fever and anemic symptoms, hepatosplenomegaly, and multiple lymphadenopathies at the age of 3 years. A complete blood count (CBC) showed anemia with thrombocytopenia and hyperleukocytosis. Bone marrow examination was performed and abnormal lymphoblasts were seen, with the immunophenotype as a Pre-B cell lineage shown by flow cytometry. A chromosome study of the bone marrow sample revealed 46, XX. The diagnosis of Pre-B cell ALL was made, and chemotherapy was started. After the complete course of chemotherapy, a bone marrow examination was performed and it revealed no abnormal cells. After 56 months, the patient developed fever, anemia, and pancytopenia. Bone marrow examination showed lymphoblasts with almost 100% of cellularity. Chemotherapy for relapse was given. The patient however developed multiple episodes of central nervous system (CNS) relapse. WES performed in the bone marrow samples identified the variants, c.5890_5906del [p.(Ser2492Leu)] in the NCOR2 gene.

Patient 2 was a 15-year-old girl with recurrent osteosarcoma. She developed a palpable mass on the left side of the ninth rib at the age of 10 years. The pathology of the excised tissues revealed osteogenic osteoma. Imaging studies showed no evidence of metastasis. The patient received adjuvant chemotherapy. About 45 months later, the patient developed pulmonary metastases in the right upper lung (RUL) by computerized tomography (CT) scan. RUL lobectomy was performed and the tissue pathology revealed osteosarcoma. WES identified the variant, c.7475C>T [p.(Ser2492Leu)] in COL6A3.

Patient 3 was a 14-year-old boy with recurrent colon cancer. He presented with acute right lower abdominal pain with fever and vomiting without a history of abnormal defecation at the age of 12 years. The patient underwent an exploratory laparotomy with a right hemicolectomy. A mass in the mid-transverse colon with the size of 3 cm × 3 cm was detected. The tissue pathology revealed a poorly differentiated adenocarcinoma, a type of signet ring cells with foci of neuroendocrine differentiation. The patient subsequently received chemotherapy. During the treatment, an abdominal CT scan showed multiple peritoneal nodules. The biopsy was performed, and tissue pathology showed poorly differentiated adenocarcinoma, with signet ring cells. The diagnosis of recurrent colon cancer was therefore made after 19 months after the first diagnosis. WES identified 3 candidate variants from the tissue samples, c.757dupA [p.(Thr253Asnfs*11)] (Tier III) in TP53, c.1082G>A [p.(Thr253Asnfs*11)] (Tier III) in SMAD4, and c.1082G>A [p.(Thr253Asnfs*11)] (Tier III) in GNAS.

Patient 4 was a 6-year-old boy with a recurrent Wilms’ tumor. A palpable abdominal mass was noted at the age of 5 years. A whole abdominal CT scan showed a right renal mass with internal calcification. The patient underwent exploratory laparotomy with tumor removal. Wilms’ tumor stage III was diagnosed, and chemotherapy was given. The tissue pathology revealed a nephroblastoma with favorable histology. After 6 months of follow-up, the patient developed severe abdominal pain, nausea, and vomiting. An abdominal CT scan showed an increase in tumor size with intestinal obstruction. WES of the tissue samples revealed a heterozygous variant, c.96+1G>C in TP53 (Tier III).

We studied 4 pediatric patients with relapsed or refractory cancer. With WES in tumor–normal pairs of patients’ tissues, we successfully identified significant somatic variants in all cases. Genetic alterations have been demonstrated to be associated with disease progression and prognosis, which could explain the clinical courses of our patients. The findings could be useful for designing experiments to develop novel therapeutic targets.

Patient 1 had relapsed ALL. WES performed in the bone marrow samples identified the variants, c.5890_5906del [p.(Ser1964Leufs*57)] in the NCOR2 gene. The NM_006312.6 (NCOR2): c.5890_5906del [p.(Ser1964Leufs*57)] variant resulted in a frameshift, predicted to cause loss of function (LoF) through protein truncation. The p.S1964Lfs*57 variant is located 1,474 base pairs upstream from the penultimate exon junction of NCOR2. The nuclear receptor corepressor 2 (NCOR2) encodes proteins that function in B cell development pathways, which include STAT5, FOFOX1, and p53. Lee et al. [7] have reported on the role of NCOR in genomic integrity and transformation. NCOR2 defects increase the expression of FOFOX1 and RAG, which leads to genomic instability in B-cell ALL. This finding could explain patient phenotypes. Further experiments are needed to verify the effect of the NCOR2 variant on tumor progression.

Patient 2 was affected with recurrent osteosarcoma. WES of the tissue samples identified the variant, c.7475C>T [p.(Ser2492Leu)] in COL6A3. This variant is a missense located in exon 36 of COL6A3. There is a large physicochemical difference between serine and leucine, which is likely to impact the secondary protein structure as these residues differ in polarity, charge, and size. The p.Ser2492Leu variant occurs in 4 samples in COSMIC. This missense variant is predicted to be damaging by both SIFT and PolyPhen2. The nucleotide c.7475 in COL6A3 is predicted to be conserved by GERP++ and PhyloP across 100 vertebrates. The p.Ser2492Leu variant also occurred in an active binding site. This variant is therefore classified as likely oncogenic.

Patient 3 was diagnosed with recurrent colon cancer. WES performed in the tissue samples identified 3 variants, c.757dupA [p.(Thr253Asnfs*11)] (Tier III) in TP53, c.1082G>A [p.(Arg361His)] (Tier III) in SMAD4, and c.1082G>A [p.(Arg361His)] (Tier III) in GNAS. The NM_000546.6 (TP53):c.757dupA [p.(Thr253Asnfs*11)] variant is a frameshift, which is predicted to cause loss of protein function through protein truncation. The p.(Thr253Asnfs*11) variant occurred in 4 samples in COSMIC. The p.(Thr253Asnfs*11) variant has been previously classified as pathogenic in ClinVar. The p.(Thr253Asnfs*11) variant is 343 base pairs upstream from the penultimate exon junction of TP53. There are 113 downstream pathogenic LoF variants, with the furthest variant being 142 residues downstream of the variant, p.(Thr253Asnfs*11). The p.(Thr253Asnfs*11) variant results in a LoF of the TP53 gene, as indicated by the following supporting evidence: the presence of 9 overlapping LoF regions in CiVIC, the presence of existing pathogenic LoF variant NM_000546.6:c.−29+1G>C and 415 others, and 18.76% of variants in TP53 are LoF in COSMIC. It is therefore classified as oncogenic. TP53 mutations occur in approximately 40%–50% of sporadic CRC [8]. The development of CRC involves the activation of oncogenes and the inactivation of tumor suppressor genes. TP53 is a key tumor suppressor gene [9]. TP53 mutations in CRC occur in 34% of proximal colon tumors and 45% of distal colorectal tumors [10, 11] and are closely related to the progression and outcome of sporadic CRC. Common mutations occur in the DNA binding domain at the hotspot codons, which include 175, 245, 248, 273, and 282 causing the disruption of specific DNA binding and sequential transactivation [12, 13]. The most common TP53 mutations are at codon p.Arg248 (8%) and at Arg273 (6%) [14]. The NM_005359.6 (SMAD4):c.1082G>A [p.(Arg361His)] variant is a missense in exon 9 of SMAD4. There is a small physicochemical difference between arginine and histidine, which is not likely to impact the secondary protein structure as these residues share similar properties. The p.(Arg361His) variant occurs in 183 samples in COSMIC. The p.(Arg361His) variant has been previously classified as pathogenic in ClinVar.

Somatic activating mutations in GNAS have rarely been identified in the setting of colorectal adenocarcinoma. A higher proportion of GNAS mutations have been reported in tubulovillous and villous adenomas, but no GNAS mutations have been identified in tubular adenomas. These data suggest that GNAS may function as an important driver mutation during a certain phase of colorectal carcinogenesis, but then could be lost once the biological advantage provided by the mutated gene is no longer needed [15,16,17]. GNAS mutations are frequently synchronous with activating mutations in either the KRAS or BRAF oncogenes [18]. The prognostic and therapeutic significance of GNAS mutations in the setting of colorectal adenocarcinoma is currently unknown. GNAS mutations have been reported in up to 3% of colorectal carcinomas, most exhibiting a villous morphology [15, 16].

WES of the tissue samples obtained from patient 4 revealed a heterozygous variant, c.96+1G>C in TP53 (Tier III). The NM_000546.6 (TP53):c.96+1G>C variant affects a splice–donor sequence, potentially resulting in the retention of large segments of intronic DNA by the mRNA and non-functional proteins. The c.96+1G>C variant occurs in 2 samples in COSMIC. The c.96+1G>C variant has been previously classified as pathogenic in ClinVar. The c.96+1G>C variant is predicted to disrupt the existing donor splice site at this position by 3 of 4 splice site algorithms. There are 347 downstream pathogenic LoF variants, with the furthest variant being 395 residues downstream of the variant c.96+1G>C. The c.96+1G>C variant is a LoF variant in the gene TP53, which is intolerant of LoF variants, as indicated by the following supporting evidence: presence of 9 overlapping LoF regions in CiVIC, presence of the existing pathogenic LoF variant NM_000546.6:c.−29+1G>C and 415 others, and 18.76% of variants in TP53 are LoF in COSMIC. For these reasons, this variant has been classified as oncogenic. TP53 deletions occur in approximately <1% of cancers. Inactivation of TP53 plays an important role in neoplastic transformation in solid tumors and it has also been reported in hematological malignancies in association with the progression of disease [19,20,21]. Deletion of the TP53 tumor suppressor gene occurs in approximately 50% of all human cancers [22].

WES has limitations in detecting copy number variants, large gene deletions, large structural variants, and variants located deep within introns. These genetic alterations also play an important role in cancer progression and aggressiveness. Larger studies are required to assess the molecular signature of relapsed or refractory tumors.

This is the first study of WES in tumor–normal pair samples in Thailand. There have been studies of WES in newly diagnosed tumors compared to relapsed/recurrent tumors, both sharing similar mechanisms and variants. In a recent extensive cohort, TP53 and PIK3CA were identified as the most prevalent small gene alterations in somatic tissue. Additionally, copy number variations (CNVs) and fusion genes, such as the EWSR1/FLI1 fusion gene in Ewing sarcoma, play a crucial role in tumorigenesis. Larger studies are imperative to unveil molecular signatures and tumor-specific variants across various types of tumors in the Thai pediatric population.

In conclusion, using WES in tumor–normal pairs of patients’ samples, this study successfully identified somatic variants in all Thai pediatric cases with relapsed or refractory cancers. These findings could provide further insight into the mechanism underlying hard-to-treat cancers.