An 8-year-old female slipped on the stairs and had a bruise on her right knee. However, since the pain persisted, the patient visited a nearby orthopedic clinic. Magnetic resonance imaging (MRI) revealed the possibility of a bone malignancy (Figure 1). The patient was then admitted to the hospital for further examination. Biochemical testing revealed significantly elevated levels of both serum AFP (AFP: 79016 ng/ml, L3: 3555 ng/ml (4.5%), PIVKA-II: 128 IU/l), which suggested possible malignancy of the liver. A serum AFP level as high as 79,000 ng/ml was also indicative of a liver tumor. Antibodies against hepatitis virus B and C were not detected. The development of hepatoblastoma was atypical in terms of the age of onset, given that hepatoblastomas typically arise in patients under 3 years of age. Pathological diagnosis was transitional liver cell tumor (TLCT). A subset of liver cell tumors in older children and adolescents may develop as an intermediate between blastomatous tumors and adult-type tumors. This type of tumor, referred to as TLCT, is highly dependent on Wnt signaling than other categories of liver tumors 13. MRI imaging showed the simultaneous development of osteosarcoma and liver cancer (Figure 1), both of which were confirmed by biopsy. After two courses of standard neoadjuvant chemotherapy for hepatoblastoma (a combination of cisplatin (CDDP) and tetrahydropyranyl-adriamycin (THP-ADR)), the residual liver tumor was removed completely by surgical extended lobectomy. After treatment with the standard neoadjuvant chemotherapy protocol for osteosarcoma (a combination of CDDP, ifosfamide (IFO), and methotrexate (MTX)), the diminished osteosarcoma was totally resected and an artificial joint was placed. The levels of transitional serum biomarkers for the liver tumor and osteosarcoma are shown (Figure 2), in which the blue arrows indicate repeated neoadjuvant chemotherapy. To date, there has been no sign of relapse or latent metastasis for about two years after last surgery. Combined therapy with surgical resection after chemotherapy was, therefore, successful in this patient. Given that we found the simultaneous development of a liver tumor and osteosarcoma, we performed genetic testing to evaluate the patient’s susceptibility to malignancy. Despite the absence of a familial history of hereditary cancer, a germline mutation in p53 (c.722 C>T, p.Ser241Phe) was identified. Notably, the frequency of de novo mutations in LFS is between 7% and 20% 14; it is, therefore, conceivable that the p53 mutation occurred de novo during embryonic development.

Immunohistochemical (IHC) analysis of the patient’s tumors prior to and following chemotherapy was performed using an antibody against the human CD44 variant isoform. CD44v8-10 was not expressed in both the liver tumor and the osteosarcoma before chemotherapy, but was ectopically expressed after chemotherapy (Figure 3).

CD44 exists in as many as 16 different isoforms, which are generated through alternative mRNA splicing. CD44v8-10 is generated by epithelial splicing regulatory protein 1 (ESRP1), an RNA-binding protein 15. Whereas the standard CD44 isoform (CD44s), which contains exons 1–5 and 16–20, is expressed predominantly in hematopoietic cells and normal epithelial cell subsets, CD44 variant isoforms with insertions in the membrane-proximal extracellular region are abundant in epithelial-type cancers, including liver tumors. Recently, CD44v8-10 was reported to have a novel function; it inhibits oxidative stress in tumor cells by promoting the GSH synthesis 12. However, exactly transcriptional factors or epigenetic mechanisms controlling the induction of ESRP1, the master regulator of CD44v8-10, remain unclear 9. We have already performed IHC using the specific antibody against human ESRP1 and ESRP2. Unfortunately, there was no antibody which was suitable for IHC due to reactions to the non-specific antigens, so that we could not obtain data that shows ESRP1 or ESRP2 exclusively in the nucleus (data not shown).

Osteosarcoma originates from mesenchymal tissue, not epithelial tissue, meaning that the CD44 variant isoforms expressed in osteosarcoma may have a completely different function from those expressed in liver cancer. Whereas the function of CD44v8-10 in osteosarcoma is unknown, the only v6 expression is negatively correlated with 5-year metastasis-free survival. Therefore, overexpression of the variant isoform, CD44v6, is considered a poor prognostic marker in patients with osteosarcoma 15. However, no specific data has been reported regarding the biological significance of CD44v8-10 in osteosarcoma. The cytotoxic drugs CDDP and THP-ADR increase the ROS level in cancer cells, thereby inducing DNA damage and apoptosis. It is possible that CD44v8-10 is induced following chemotherapy in osteosarcoma to counteract this oxidative stress.

Considering the lack of familial history of malignancy in this case, it is supposed that the de novo p53 germline mutation occurred during early development. The p53 protein suppresses expression of CD44 by binding to the promoter region of CD44 10, and CD44v8-10 was not expressed in both tumors before chemotherapy. The absence of p53 likely enabled the upregulation of CD44v8-10 in response to chemotherapy. However, this finding does not rule out the possibility that other CD44 isoforms were fully expressed prior to chemotherapy. Although cancer cells may acquire resistance to oxidative damage by upregulating CD44v8-10, there has been no sign of relapse or metastasis so far in this patient. The fact that CD44v8-10 expression was not correlated with increased malignant potential in this case seems to be contradictory to our previous research 12; however, the effect of CD44v8-10 expression may depend on the origin of the tumor. It is possible that CD44v8-10 may function differently in hepatic tumors and osteosarcomas with p53 mutations than in gastric and colorectal cancers.

Potential mechanisms for the ectopic induction of CD44v8-10 expression in osteosarcoma are shown (Figure 4):

a) Cancer cells expressing CD44v8-10 may have existed in the original tumor, but were too rare to be detected by IHC. These cells survived even under conditions of oxidative stress caused by cytotoxic chemotherapy.

b) Cancer cells expressing CD44v8-10 did not exist originally, but the oxidative stress caused by chemotherapy induced the expression of CD44v8-10 in the tumor.

Differentiating between these possibilities will require further investigation into how the alternative splicing machinery for CD44 is affected by chemotherapy. There has been no sign of relapse or metastasis so far in this patient, but it will be crucial to gather additional data regarding CSC molecular dynamics during the course of chemotherapy. Therefore, a larger number of cancer cases need to be analyzed, and longer-term follow-up studies must be conducted, to confirm the results of this study.

Written informed consent was obtained from the patient for publication of this case report and the accompanying images. A copy of the written consent is available for review upon request.