NEDD9 overexpression: Prognostic and guidance value in acute myeloid leukaemia

Abstract It has been demonstrated that neural precursor cell expressed developmentally downregulated protein (NEDD) plays crucial roles in tumorigenesis and may serve as potential biomarkers in cancer diagnosis and prognosis. However, few studies systematically investigated the expression of NEDD family members in acute myeloid leukaemia (AML). We systemically determined the expression of NEDD family members in AML and determined their clinical significance. We identified that NEDD9 expression was the only member among NEDD family which was significantly increased in AML. NEDD9 overexpression was more frequently classified as FAB‐M4/M5 (p = 0.008 and 0.013, respectively), hardly as FAB‐M2/M3. Moreover, NEDD9 overexpression was significantly associated with complex karyotype and TP53 mutation. The significant association between NEDD9 overexpression and survival was also observed in whole‐cohort AML and non‐M3 AML patients. Notably, AML patients with NEDD9 overexpression may benefit from hematopoietic stem cell transplantation (HSCT), whereas those cases without NEDD9 overexpression did not. Finally, a total of 822 mRNAs and 31 microRNAs were found to be differentially expressed between two groups. Among the microRNAs, miR‐381 was also identified as a microRNA that could direct target NEDD9. Taken together, our findings demonstrated that NEDD9 overexpression is associated with genetic abnormalities as well as prognosis and might act as a potential biomarker guiding the choice between HSCT and chemotherapy in patients with AML after achieving complete remission.

For instance, Fujita et al revealed that NEDD1 expression silencing by siRNA might provide a new opportunity in the treatment of the peritoneal metastasis of scirrhous gastric cancer. 9 NEDD4 is widely studied and mostly functions as an oncogene in human cancers, such as gastric cancer, colorectal cancer, lung adenocarcinoma, non-small-cell lung carcinoma, hepatocellular carcinoma, breast cancer, and endometrial cancer. 5,10,11 Oncogenic role of NEDD8 has been demonstrated in diverse human cancers. 6,[12][13][14] Notably, NEDD8-activating enzyme inhibitor MLN4924 (pevonedistat) has been used in clinical treatment of AML. 6,[15][16][17] NEDD9, as a member of the CAS family of adhesion docking proteins, plays a key role in regulating several signalling cascades related to multiple activities, including migration, adhesion, cell death or proliferation. 18,19 Overexpression of NEDD9 has now been strongly linked to poor prognosis in various types of cancers, as well as resistance to first-line chemotherapeutics. 20 However, NEDD9 exhibits opposite effects regarding migratory capacity on myeloid cells as compared to epithelial or lymphoid cells, which block migration and dissemination of neoplastic cells of the myeloid lineage, while stimulating them in solid tumours or lymphoid neoplasias. NEDD9 −/− mice show an increased number of macrophages, and a simultaneous reduction of B lymphocytes in peripheral blood 21 and secondary lymphoid organs, yielding an almost complete loss of marginal zone B cells in the spleen. 22 NEDD9-deficient p210-BCR/ABL transgenic mice show an increased number of granulocytes in peripheral blood, a hyperplasia of myeloid and megakaryocytic cells in the bone marrow and a diffuse myeloid infiltration in the spleen, lung and liver, leading to earlier progression and shorter mouse survival, which support NEDD9 capacity to block chronic myeloid leukaemia (CML) progression. 23 There are few studies on the association of NEDD9 and AML. 24 Herein, as far as known, we for the first time identified and verified that NEDD9 expression, among NEDD family members, was significantly increased in AML. NEDD9 overexpression was correlated with specific cytogenetic and genetic abnormalities of AML.
Moreover, NEDD9 overexpression predicts poor clinical outcome in AML and might act as a potential biomarker guiding treatment selection between chemotherapy and hematopoietic stem cell transplantation (HSCT) as consolidation therapy.

| GEPIA analysis
The Gene Expression Profiling Interactive Analysis (GEPIA) database (http://gepia.cance r-pku.cn/) provides RNA-sequencing expression data of 9,736 tumours and 8587 normal samples from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) projects, using a standard processing pipeline. 25 The expression of NEDD family members between AML and control was identified by GEPIA.

| TCGA data
A total of 173 AML patients with RNA-sequencing data from the databases of TCGA were included in this study. 26 NEDD family member expression data of these patients were obtained by mRNA sequencing. Mutation data of these patients were also obtained by DNA sequencing. Clinical characteristics and treatment regimens of these patients were also obtained.

| Bioinformatic analysis
To obtain the differential expressed genes (DEGs), analysis of RNAsequencing (mRNA and microRNA) data was calculated using the raw read counts with the R/Bioconductor package 'edgeR'. All analyses were controlled for the false discovery rate (FDR) by the Benjamini-Hochberg procedure. Functional and signalling pathway enrichments were analysed through the STRING (http://strin g-db.

| Statistical analysis
Student t/Mann-Whitney U/Kruskal-Wallis H test and Pearson's χ 2 /Fisher's exact test were applied for the comparison of continuous and categorical variables, respectively. The effect of NEDD family member expression on overall survival (OS) and leukaemia-free survival (LFS) was analysed by the Kaplan-Meier method (log-rank test). Two-sided p-values less than 0.05 in all statistical analyses were considered as statistically significant differences.

| Expression of NEDD family in AML
In order to investigate the NEDD family (NEDD1/4/8/9) expression patterns in AML, we first used the AML cohort from public databases by GEPIA online website. The AML patients were from the TCGA data sets, whereas normal controls were from the GTEx projects. The expression of NEDD1/4/8 showed no significant differences between AML and controls ( Figure 1A-C). However, NEDD9 expression was markedly increased in AML ( Figure 1D).

| Clinical implications of NEDD9 expression in AML
Since NEDD9 was the only one member of NEDD family to be aberrantly expressed in AML, we further analysed the correlations of abnormal NEDD9 expression (above the median level) with the clinical/biological characteristics. As shown in Table 1, AML cases with NEDD9 overexpression had a markedly older age than those without NEDD9 overexpression (p = 0.003). Interestingly, NEDD9 overexpressed patients had significantly lower peripheral blood blasts than NEDD9 underexpressed patients (p < 0.001). Moreover, significant differences in the distribu-

| Prognostic value of NEDD9 expression in AML
To explore the prognostic value of NEDD9 expression in AML, Kaplan-Meier analysis was performed and revealed that AML patients with

| Guidance value of NEDD9 expression in AML
HSCT as a consolidation treatment regimen is of great importance in AML against disease recurrence. To investigate whether HSCT might overcome the adverse prognostic effect caused by NEDD9 overexpression in AML, we analysed the prognostic impact of HSCT in NEDD9 overexpressed and underexpressed groups, respectively.
After AML patients achieved CR, cases undergoing HSCT exhibited markedly longer OS and LFS compared with that only receiving chemotherapy in NEDD9 overexpressed group. However, in NEDD9 underexpressed group, there were no significant differences regarding OS and LFS between HSCT and chemotherapy groups (Figure 4). These results suggested that AML patients with NEDD9 overexpression may benefit from HSCT, and NEDD9 expression might act as a potential biomarker guiding treatment selection between HSCT and chemotherapy in patients with AML after achieving CR by induction therapy.

| Biological network of NEDD9 expression in AML
To get better understanding of the biological network associated with NEDD9 expression in AML, we first compared the transcrip-  to be differentially expressed between two groups ( Figure 5A,B and Table S1). The top 10 upregulated genes such as FEZ1 and PDK4 are reported with proto-leukaemia effects. 27,28 Furthermore, the Gene Ontology analysis revealed that these genes involved in biologic processes, including multicellular organismal process, cell communication and signalling ( Figure 5C).  Figure 5D and Table S1). Downregulated microRNAs such as miR-135a, miR-203, miR-497, miR-381, miR-370 and miR-758 were found to be underexpressed in AML or have anti-leukaemia effects in previous reports. [29][30][31][32][33][34][35] Of these microRNAs, miR-381 was also identified as a microRNA that could direct target NEDD9 ( Figure 5E and Table S2), which suggested NEDD9 is a direct target of miR-381.

| DISCUSS ION
It has been determined that NEDD9 plays a crucial role in regulating several signalling cascades contained in multiple activities, including cell apoptosis or proliferation, migration, invasion, metastasis and adhesion. 7 Moreover, overexpression of NEDD9 correlated with cancer cell development and drug resistance in several types of solid tumours such as lung cancer, melanoma and breast cancer. 8 It is not surprising that aberrant NEDD9 expression has been linked to the prognosis of human cancers. 8,20 In this study, we for the first time revealed that NEDD9 overexpression, identified from NEDD family, was associated with poor prognosis in AML. Notably, NEDD9 expression might act as a potential biomarker predicting prognosis and guiding treatment choice between chemotherapy and HSCT in AML. Until now, few investigations have reported the links between NEDD9 and AML. In contrary to our results, Pallarès et al demonstrated that NEDD9 was an independent good prognostic factor in intermediate-risk AML patients. 24 The possible reason was that the previous report only included AML patients less than 65 years. As it is well known, AML is an ageing disease which contains larger numbers of older patients.
Accordingly, further clinical and functional studies are needed to evaluate the clinical implication and potential role of NEDD9 in AML.
Our study also found significant associations between NEDD9 expression and FAB classifications as well as cytogenetic/genetic subtypes in AML. For FAB classifications, NEDD9 overexpression was associated with FAB-M4/M5 in accordance with previous studies, 24 and results analysed by BloodSopt (https://serve rs.binf.ku.dk/ blood spot/) show that NEDD9 expression is significantly higher in monocytic lineages, suggesting it may play a crucial role in monocytic line development ( Figure S1). For cytogenetic/genetic subtypes, NEDD9 overexpression was found to be strongly correlated with complex karyotype and TP53 mutations. Since TP53 mutation is frequently occurred in the complex karyotype, 36 it is difficult to classify which is mainly the factor associated with NEDD9 overexpression.
Interestingly, previous studies have determined the association of TP53 with NEDD9 in non-small-cell lung cancer. 37 [40][41][42][43][44] Additionally, NEDD9 expression negatively associated with miR-125a/b was shown in pancreatic cancer, lung adenocarcinoma and melanoma. [45][46][47] In pancreatic cancer and hepatocellular carcinoma, NEDD9 was reported to be regulated by miR-18a playing a key role during carcinogenesis. 48,49 In our study, we for the first time found that NEDD9 expression was negatively associated with miR-381 in AML. However, the limitation in our study was that luciferase assay was not performed to verify the direct associations between miR-381 and NEDD9. Therefore, a number of investigations are needed to confirm our results in the future.
Collectively, our findings demonstrated that NEDD9 overexpression associated with genetic abnormalities as well as prognosis might act as a potential biomarker guiding the choice between HSCT and chemotherapy in patients with AML.

ACK N OWLED G EM ENTS
None.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.