Aneuploidy in neuroblastoma tumors is not associated with inactivating point mutations in the STAG2 gene

Background Chromosomal instability is a hallmark of human cancer caused by errors in mitotic control and chromosome segregation. STAG2 encodes a subunit of the cohesion complex that participates in mitotic chromatid separation and was recently found to show low expression and inactivating mutations in Ewing’s sarcoma, melanoma and glioblastoma. In the childhood tumor neuroblastoma (NB) segmental chromosomal alterations are associated with poor prognosis whereas tumors displaying whole chromosome gains and losses have a much better prognosis. Method As the genetic contribution to aneuploidy is unknown in NB, we investigated the presence of STAG2 mutations through sequence analysis of all 33 coding exons in 37 primary NB tumors. Results and conclusion As no STAG2 mutation was detected in this study, we conclude that inactivating mutation of STAG2 is not likely causative to neuroblastoma aneuploidy.


Background
Chromosome instability leading to numerical and segmental alteration is a hallmark of human cancer genomes, contributing to malignant transformation. Whereas segmental rearrangements directly can promote cancer through activating oncogenes and inactivating tumors suppressor genes, aneuploidy could provide a tumor driving function through dosage-dependent mechanisms [1]. Aneuploidy is caused by inaccurate chromosome segregation, possibly due to genetic alteration of various participants of the segregation or mitotic checkpoint machinery. Recently, Solomon et al., showed the presence of inactivating mutations in the STAG2 gene located at Xq25 in primary tumors and cancer cell lines originating from glioblastoma, melanoma and Ewing's sarcoma. The STAG2 mutations reported by Solomon et al., represented an array of different alterations (missense-, nonsense-, splice site mutations as well as intragenic deletions), mainly concentrated to hotspot regions corresponding to exon 9, 11, 12 and 20 [2]. STAG2 encodes a subunit of the multimeric cohesion complex required in regulating sister chromatid separation during mitosis [3][4][5]. Inactivation of STAG2 in cell lines with stable karyotype resulted in increased number of abnormal mitotic events (e.g. lagging chromosomes and anaphase bridges) ultimately leading to increased aneuploidy rate while mutational correction of endogenous STAG2 mutations in aneuploid cell lines resulted in enhanced chromosomal stability. Further, loss of STAG2 protein was observed in several human cancer cell lines, primary tumors and xenografts of different origin [2]. Neuroblastoma (NB), a childhood cancer of the postganglionic nervous system show broad biological and clinical heterogeneity, ranging from highly aggressive tumors with fatal outcome to tumors with spontaneous regression. Advanced disease and poor prognosis in NB patients is commonly associated with either MYCN amplification or 11q-deletion, representing two separate NB subgroups with different genomic profiles. Whereas MYCN amplified tumors commonly have few segmental or numerical aberrations besides 17q-gain and 1p-deletions, 11q-deleted NB generally contains many more segmental alterations. In contrast, NB tumors with numerical aberrations (i.e. only whole chromosome gains and/or losses) without other segmental alterations instead have excellent outcome [6] and the underlying mechanisms causing aneuploidy in NB tumors are yet unknown.
As the findings of Solomon et al [2] indicate that loss of functional STAG2 protein could contribute to increased chromosomal instability we wanted to investigate the presence of inactivating mutations in STAG2 in prognostically favorable NB with common aneuploidy pattern and prognostically unfavorable NB with high degree of segmental alteration.

Cell material
Genomic DNA of 37 primary NB tumors and cell lines was extracted with DNeasy blood and tissue kit (Qiagen, Hilden Germany) according to the manufacturer's instructions. Mutation analysis through Sanger sequencing of the 33 coding exons (3-35) of STAG2 was performed on 18 favorable NB tumor samples (stage 1, 2 and 4S), whereas additional 19 tumors of different risk groups were subjects for whole exome sequencing whereof one was also evaluated through Sanger sequencing (Table 1).
In addition, mutation analysis with Sanger sequencing was performed on eleven NB cell lines ( Table 2)

High resolution SNP array analysis
Genomic profiling of the 37 primary NB tumors (Table 1, Figure 1) were performed with Genechip Human Mapping 250K NspI (Affymetrix Inc., Santa Clara, CA) according to the experimental procedure previously described [9].

Sanger sequencing
Sanger sequencing was performed as earlier described [10] and primers were designed manually or with the use of ExonPrimer (http://ihg.gsf.de/ihg/ExonPrimer.html), primer sequences are available upon request. The sequence data were analyzed using Seqscape v2.5 (Applied Biosystems).

Results
The coding sequence of the STAG2 gene was screened with Sanger sequencing and targeted re-sequencing in a panel of 37 NB tumors and eleven NB cell lines (Tables 1  and 2). High-quality sequence of the STAG2 gene was obtained for all 33 coding exons both from Sanger-and exome sequencing. Sequences with an average coverage of 51x (range 13-117x) were acquired from all coding exons in tumors analyzed by exome sequencing. One of 37 primary tumors (11E8) and one of eleven NB cell lines (SK-N-DZ) showed a hemizygous and heterozygous base substitution respectively corresponding to C>T at mRNA position 1298-9 in intron 10 ( Figure 2), similar to a finding earlier reported in a glioblastoma primary xenograft [2]. However, this variant corresponding to SNP rs7063522, was also present in normal DNA from patient's blood ( Figure 2). No exonic variants were detected in our set of 37 NB tumors and 11 NB cell lines, neither in the hot spot positions reported by Solomon et al., nor in any other coding region of the STAG2 gene [2].

Discussion
Aneuploidy results from missegregation of sister chromatids during mitotic cell division. In NB segmental chromosomal alterations are common in aggressive tumors whereas tumors with only whole chromosome gains and losses are prognostically more favorable [6]. However, the cause of the high aneuploidy rate in this favorable group of NB tumors is currently unknown. Aneuploidy is also frequently observed in other malignancies and recently STAG2 was recognized as a tumor suppressor gene causing malfunctioning mitotic sister chromatid separation when inactivated. Inactivating mutations of STAG2 have been found in glioblastoma, Ewing's sarcoma and melanoma, possibly contributing to chromosome instability in these tumors [2] and this led us to investigate whether they might also be responsible for the high aneuploidy rate seen in favorable NB tumors.  In the present study, we sequenced all 33 coding exons of the STAG2 gene in 11 NB cell lines and 37 primary NB tumors and no deviation from the human reference sequence could be detected. This is similar to other recent studies of the STAG2 mutational hot spot region (corresponding to exon 9, 11, 12 and 20) in 90 acute leukemia samples [15] and 225 adult carcinomas of various origin (colorectal-, gastric-, breast-, prostate carcinoma and non-small lung cell cancer) [16]. Alteration of STAG2 was not detected in either study, suggesting that  STAG2 inactivation is not due to inactivating gene mutations in these malignancies. Other gene silencing mechanisms might contribute as STAG2 protein was lost in 23-30% of studied carcinomas [16]. According to the publicly accessible R2-microarray analysis and visualization platform (http://r2.amc.nl), STAG2 expression in NB can be correlated with outcome in that lower STAG2 expression is correlated to lower overall survival and relapse free survival. Other causes than STAG2 mutations are likely to be involved in this, e.g. epigenetic events.
Furthermore, previous and subsequent analyzes of copy number variations performed by us in totally 360 primary NB tumors show that segmental deletions of the chromosomal region Xq25 harboring STAG2 are not seen ( [6] and unpublished data).
Taken together these data suggest that STAG2 mutations are not likely responsible for the high aneuploidy rate in NB tumors with numerical aberrations. However, other genes participating in the mitotic checkpoint control may be responsible for the chromosomal missegregation in NB tumors with many numerical aberrations.
Three major next generation sequencing studies have together analyzed a large serie of NB tumors [17][18][19]. No STAG2 mutations are reported and overall very few recurrent amino-acid changing mutations were found. Molenaar et al. [17], found few recurrent mutations except for the genes ALK and TIAM1 and mutations in genes regulating neuritogenes. Cheung et al. [18], reported the finding of deletions and loss-of-function mutations in the ATRX gene and the association with advanced age at NB diagnosis. Furthermore, Sausen et al. [19], found mutations and deletions in the genes ARID1A and ARID1B which were associated with treatment failure and poor prognosis. Collectively more recurrent somatic mutation have been found in high risk NB patients compared to patients with stage 1, 2 and 4S tumors and in the published NGS data we cannot identify clear candidate genes that could be responsible for the high aneuploidy rate seen in favorable NB.

Conclusion
Here we show that inactivating point mutations in the STAG2 gene are not common in NB tumors and that aneuploidy seen in NB tumors is likely due to other participants of the mitotic checkpoint and chromosome segregation machinery.