Oncogenic FGFR1 mutation and amplification in common cellular origin in a composite tumor with neuroblastoma and pheochromocytoma

Abstract Neuroblastoma (NB) and pheochromocytoma (PCC) are derived from neural crest cells (NCCs); however, composite tumors with NB and PCC are rare, and their underlying molecular mechanisms remain unknown. To address this issue, we performed exome and transcriptome sequencing with formalin‐fixed paraffin‐embedded (FFPE) samples from the NB, PCC, and mixed lesions in a patient with a composite tumor. Whole‐exome sequencing revealed that most mutations (80%) were shared by all samples, indicating that NB and PCC evolved from the same clone. Notably, all samples harbored both mutation and focal amplification in the FGFR1 oncogene, resulting in an extraordinarily high expression, likely to be the main driver of this tumor. Transcriptome sequencing revealed undifferentiated expression profiles for the NB lesions. Considering that a metastatic lesion was also composite, most likely, the primitive founding lesions should differentiate into both NB and PCC. This is the first reported case with composite‐NB and PCC genetically proven to harbor an oncogenic FGFR1 alteration of a common cellular origin.


| INTRODUC TI ON
PCC and NB are the most common NCC-derived tumors in adults and children, respectively. [1][2][3] Composite pheochromocytoma refers to tumors with morphologic features of PCC and NCC-derived tumors, such as malignant peripheral nerve sheath tumor and neuroendocrine carcinomas, within the same tumor. 4,5 Composite tumors are rare and most often combined with ganglioneuroma in composite PCC; therefore, composite tumors comprising PCC and NB are even rarer. [5][6][7][8][9][10] The genetic mechanism of composite tumors with PCC and NB remains unclear; a single nucleotide polymorphism array analyzed only 1 case. 5 Here, we applied exome and transcriptome analyses to a patient case with a composite tumor with NB and PCC to investigate whether the NB and PCC lesions arose from a common cellular origin and how this tumor developed.

| MATERIAL S AND ME THODS
Detailed methods are provided in the Supporting information section of this paper and include the following:  however, the treatment was quickly terminated due to a rash. He then received MIBG radiotherapy and pazopanib; however, he died from disease progression ( Figure 1A).

| Composite-NB and PCC have the same cellular origin
Using FFPE samples at the fourth surgery, we extracted DNA and RNA from a total of 6 lesions as follows: 1 from each of the NB, PCC, and mixed lesions in the primary adrenal gland and of the metastatic intra-abdominal lymph node. We identified somatic mutations and copy number alterations (CNA) in all DNA samples using WES.
Consequently, 108 somatic mutations were detected and validated by amplicon deep sequencing (median: 18 per sample; range: [17][18][19] ( Table S1). Fifteen mutations were found in all 6 samples, indicating NB, PCC, and mixed lesions evolved from the same neoplastic clone ( Figure 2A,B). Additionally, there were no NB or PCC-specific mutations, suggesting the composite characteristic was not driven by distinct gene mutations (Figure 2A,B).

| Composite-NBs transcriptionally contain larger fractions of early normal fetal adrenal neuroblasts
We further illustrated the molecular basis of this composite tumor by performing whole-transcriptome sequencing of the 6 samples.  (Table   S2). Most of the NBs, including composite-NBs, were confirmed to transcriptionally match the normal neuroblasts ( Figure 2E).
Unexpectedly, in composite-NBs, only a few late neuroblasts (differentiated neuroblasts) were detected; however, the abundance of neuroblasts (early neuroblasts) was higher than that in TARGET NB cohort ( Figure  Therefore, tumors are believed to consist of a heterogeneous mixture of functionally distinct cancer cells and that their subpopulations vary widely in their responses to therapeutic agents. 23 In the present case, it seems possible that the clone with del (7) was present as a minor clone in other lesions, albeit this clone subsequently became dominant in this mixed lesion. Nevertheless, whether the significance of this del (7) is related to drug resistance or malignancy remains unclear.
Third, the expression profile of composite-NBs was mainly similar to that of early normal neuroblasts, which suggests that the FGFR1 alteration may be acquired in less differentiated progenitor cells. This speculation is supported by the fact that, in a mixed phenotype acute leukemia, which is also a composite of lymphoid and myeloid hematopoietic lineages, mutations are acquired in early hematopoietic progenitor cells, which then drives the bi-phenotypic nature. 24 To the best of our knowledge, this is the first case with composite-NB and PCC, genetically proven to harbor a common cellular origin. Furthermore, our results suggest a possible mechanism for the formation of this composite characteristic in line with a previous report 14 (Figure 3). However, as there were no viable cells left, in vivo validation using xenograft models or cell lines derived from this composite tumor was impossible; therefore, further studies, including those on other composite tumors, are needed to investigate whether the stemness is related to the formation of composite tumors.