Novel Biomarker NEDD1 and Its Analysis in Hepatocellular Carcinoma

Xiaopeng Ding Xijing Hospital, Fourth Military Medical University Jiahao Yu Xijing Hospital, Fourth Military Medical University Xin Shi Xijing Hospital, Fourth Military Medical University Kangwei Li Xijing Hospital, Fourth Military Medical University Shuoyi Ma Xijing Hospital, Fourth Military Medical University Pengwei Ren Xijing Hospital, Fourth Military Medical University Guoyun Xuan Xijing Hospital, Fourth Military Medical University Ying Han Xijing Hospital, Fourth Military Medical University Xinmin Zhou (  zhouxmm@fmmu.edu.cn ) Xijing Hospital, Fourth Military Medical University


Introduction
HCC constitutes approximately 90% of primary liver cancer worldwide 1 , which ranks fourth leading cause of cancer-related mortality and sixth among cancer diagnoses 2 . Despite the development of clinical managements of HCC, the prognosis of patients with HCC is dismal. The overall survival (OS) of HCC patients varies around the world, but it is still rather low with the 5-year survival rate 14.1% in China at present 3 . With the exploration of the treatment of HCC, the research of multikinase inhibitors combined with immune checkpoint inhibitors for patients at advanced stages has gradually become a hot spot.
Additionally, atezolizumab plus bevacizumab was admitted for the rst-line treatment in unresectable HCC on June 1, 2020 by Food and Drug Administration (FDA) because the combination therapy can extraordinarily prolong progression free survival (PFS) than sorafenib. However, the PFS is 6.8 months (5.7-8. 3) for atezolizumab plus bevacizumab in unresectable HCC, which is still far from satisfaction.
Unfortunately, there are few molecular classi cations predicting progression or recurrence of HCC 4 .
Consequently, it is urgently demanded to investigate robust prognostic indicator and promising targets for therapy of HCC patients. NEDD1, also known as GCP-WD, a crucial component in controlling gamma-tubulin levels at the mammalian centrosome, promotes initiation of mitosis in human beings 5 . Previous studies indicated that NEDD1 probably be directly associated with cell cycle regulation 6,7 . Cell cycle has been reported to play a crucial impact in the development of HCC 8, 9 , NEDD1 could be a potential novel target to inhibit cell proliferation in HCC. Previous report has also demonstrated that small interfering RNA (siRNA) targeting NEDD1 is able to reduce cell numbers, including cervix carcinoma cells HeLa, prostate carcinoma cells DU145, colon carcinoma cells DLD-1, ovarian adenocarcinoma cells SKOV-3, breast carcinoma cells MDAMB-231, pancreas adenocarcinoma cells BxPc-3 and lung carcinoma cells A549 in vitro, which reveals that NEDD1 is a potential therapeutic target for lots of cancer types 7 . Besides, in the scirrhous gastric cancer model mice, intraperitoneal delivery of a siRNA against NEDD1 can extend its survival 10 .
However, there has not been a report analyzing the clinical signi cance of NEDD1 in HCC so far. Our ndings indicates that NEED1 play an essential impact on the development of HCC.

Materials & Methods
Data source and processing The dataset of LIHC patients were downloaded from the TCGA database, at the level 3 HTSeq-FPKM format. 373 HCC cases and 50 patients with paired normal tissues specimens were enrolled for analysis.
Since complete clinical information on the patient is required to build the nomogram, we followed up by screening the patients, excluding cases with missing or de cient data on age, gender and OS. Meanwhile, data was transformed to TPM (transcripts per million reads), and the amount of normal liver tissue in TCGA was not enough, so we used the UCSC XENA database (https://xenabrowser.net/datapages/) for differential expression analysis of NEDD1 in addition to TCGA data for pairwise difference analysis 11 .

Assessment of immune cell in ltration
To analyze the association of NEDD1 with the in ltrating abundances of immune cells in HCC, the associated data about the immune in ltration levels were downloaded from tumour immune assessment resource (TIMER) (https://cistrome.shinyapps.io/timer/) 12 . Likewise, a single-sample gene set enrichment analysis (ssGSEA) was conducted to quantify the abundances of 24 categories of immune in ltration cells in tumor specimens 13 . We calculated normalized enrichment score (NES) by ssGSEA, which was employed in the "GSVA" and "GSEABase" R package 14 (version3.14.3, https://www.bioconductor.org/packages/release/bioc/html/GSVA.html). Spearman rank correlation method was conducted to investigate the relevance of NEDD1 expression with the in ltration levels. We also employed the GEPIA website (Gene Expression Pro ling Interactive Analysis) to assess the correlation of NEED1 with immune checkpoint genes 15 .

Enrichment Analysis.
Through the LinkedOmics database, we investigate the differentially expressed genes associated with NEDD1 16 . Pearson correlation coe cient was employed to test for the differences. Gene set enrichment analysis (GSEA) was conducted to assess KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway enrichment 17 and GO (Gene Ontology) analysis, which is composed of cellular component (CC), biological process (BP) and molecular function (MF) 18 . GSEA was employed to explore whether a priori de ned group of genes have statistically signi cant and concordant differences between two biological status 19 (20). In this research, GSEA produced a list of all genes, which was generated based on their association with NEDD1 expression. The R package "clusterPro ler" (version 4.1.0) was employed to conduct GSEA in high-and low-NEDD1 expression groups 20 . For each analysis, gene set permutation was performed 1000 times. Meanwhile, Gen sets with normalized enrichment score (NES) 1, P 0.05, and false discovery rate (FDR) q value 0.25 were was judged as enriched signi cantly.

Construction and validation of prognostic nomogram in HCC patients
We constructed the nomogram, a quanti able analysis tool which integrate the value of NEDD1 expression with other variables, to infer the mortality rate of individual HCC patients. Harrell's concordance index (C-index) was measured to assess the differentiation performance of the constructed nomogram. Meanwhile, we plotted the calibration curves to assess the accuracy of nomograms by comparing the predicted values of the nomogram with the observed actual survival rates. R package "survival" and "rms" were conducted to plot the nomogram and calibration curves.

Statistical Analysis
Statistical analyses were performed by R-4.1.0. Wilcoxon rank sum tests was employed to assess NEDD1 expression in tumor and normal groups. Meanwhile, Wilcoxon signed-rank tests was conducted to distinguish NEDD1 expression in paired specimens. We employed the Kaplan-Meier method to investigate the relevance of NEDD1 expression and OS in Pan-cancer analysis. Cox regression analysis was conducted to explore the role of NEDD1 and other associated varies on survival. We chose a univariate analysis p-value of less than 0.01 as the threshold for inclusion in the multivariate analysis. The analysis was done through R package "survival" and visualization was done through R package "Forestplot". P 0.05 indicated statistically signi cant, if not speci ed. Time-dependent ROC curves was plotted by R package "survivalROC" 21 . R package "ggplot2" was employed to visualize the results.

Results
The differential expression of NEDD1 in HCC We assessed the differential expression analysis of NEDD1 in HCC patients through the Wilcoxon ranksum. The results indicated that NEDD1 expression was obviously higher in LIHC, LUSC, et al. than in normal tissue by comparing NEDD1 expression in normal samples of Genotype-Tissue Expression (GTEx) and TCGA, and tumor samples of TCGA (Figs. 1A,1C). We further demonstrated our ndings in paired samples and found that NEDD1 expression also higher in CHOL, KICH, LIHC et al. (Fig. 1B). The results showed that NEDD1 was closely associated with cancer.

The relationship between NEDD1 expression and HCC prognosis
Kaplan-Meier survival curve was conducted to explore the relevance of NEDD1 expression and OS of patients in pan cancer, which indicated that NEDD1 was relevant to poor OS in HCC(P=0.002), Kidney renal clear cell carcinoma(P=0.046) and lung adenocarcinoma (P=0.020) (Fig. 2). Subsequently, forest plot regarding OS of the univariate and multivariate Cox regression analysis in HCC demonstrated that high NEDD1 expression was an independent risk indicator for HCC patients (HR 1.643, 95%CI 1.125-2.398; P = 0.01) (Fig. 3). The nding indicated that NEDD1 was signi cantly related to HCC prognosis and probably be a novel prognostic marker for HCC.

Immune in ltration analysis of NEDD1 in HCC patients
Using TIMER, we assessed the association of NEDD1 expression with the abundance of immune in ltration in HCC. The results reveled that tumor purity is not directly associated with the NEDD1 expression, but a positive relevance existed between NEDD1 expression and the abundance of in ltrating B cells(r = 0.36, P = 6.08e-12), CD8+T cells(r = 0.334, P = 2.41e-10), CD4+ T cells (r =0.471, P = 2.27e-20), Macrophages (r = 0.548, P = 3.90e-28), Neutrophils (r = 0.579, P = 2.62e-32), and DCs (r = 0.492, P = 3.93e-22) in HCC (Fig. 4A). Meanwhile, the results reveled that the immune in ltrating levels were not directly relevant to OS of HCC patients (Fig. 4B). Subsequently, Spearman correlation was performed to explore the relevance of NEDD1 with the immune cells' in ltration levels quanti ed by ssGSEA method.
Our results showed that there were positive association of NEDD1 expression with Th2 cells and Th cells, and the negative correlation of NEDD1 with cytotoxic cells, pDC and DC was found in Fig. 4C. As show in Previous studies suggested that immune checkpoint inhibitors are a promising treatment modality for the effective treatment of cancer (23). We used GEPIA to explore the relevance of NEED1 with the immune checkpoint genes, which indicated that NEDD1 was closely associated with immune checkpoint genes in pan cancer (Fig. 5A). Meanwhile, our ndings reveled that NEDD1 expression was related to the expression of CTLA4 (r=0.260, P 0.001), CD274(PD-L1) (r=0.400, P 0.001) and PTCD1(PD-1) (r=0.270, P 0.001). (Fig. 5B,5C)

GO and KEGG analyses of NEDD1
To analyze the biological function of NEDD1, we employed the LinkedOmics database to explore the coexpressed genes of NEDD1 in LIHC from TCGA, rstly (Fig. 6A). It indicated that the expression of PPP1R12A, GTE3, et al were positively associated with NEDD1 in Fig. 6B. On the contrary, ZNHIT1, DCI, et al were negatively associated with NEDD1 in Fig. 6C. Subsequently, based on GSEA in LinkedOmics, we conducted GO analysis. The results demonstrated that NEDD1 was signi cantly associated with the Cell cycle, Homologous recombination and MicroRNAs in cancer signaling pathways (Fig. 6D). Meanwhile, NEDD1 plays an important part in the process of histone binding, helicase activity and is located in chromosomal region and microtubule organizing center part (Figs. 6E,6F).
In addition, on the basis of GSEA, we performed KEGG analysis demonstrating that NEDD1 expression is positively associated with the following signaling pathways, including Cell cycle, MicroRNAs in cancer and Ribosome. It also showed that NEDD1 was negatively correlated with Ribosome pathway (Figs. 7A-7D). These results demonstrated that NEDD1 played a signi cant impact in HCC development.  (Fig. 8A). There are 373 original data, 208 cases with missing variable information, and the nal number of enrolled samples is 165. The consistency test (Concordance, C-index) of the nomogram was 0.712 (0.663-0.761). As well, the calibration curve illustrated that there was excellent agreement between the predicted and actual survival in the prognostic nomogram, suggesting that it had good predictive value. (Fig. 8B).

Construction and validation of prognostic nomogram in HCC patients
Subsequently, primary liver cancer data from ICGC was employed as an external validation of the nomogram. As shown Figs.9A-9C, it demonstrated that in the ICGC database, this modal has good predictive power and can clearly distinguish between low, medium and high risks HCC patients.

Discussion
Because of the absence of valid approaches to early diagnosis for HCC patients, almost 70% HCC patients are diagnosed at an unresectable state 1,22,23 , leading to limited treatment options and a very poor prognosis. It is of tremendous signi cance to make an early diagnosis and effectively predict the prognosis for improving the overall survival in HCC. Consequently, it is urgently needed to investigate potential biomarkers for diagnosis and prognosis, and a novel therapeutic target of HCC patients. NEDD1, the gamma-tubulin ring complex targeting factor, plays an essential impact in regulating cell cycle. Moreover, it has been proved that appropriate doses of siRNA against NEDD1 can affect numerous types of tumor cells growth in vitro and signi cantly prolong the overall survival of scirrhous gastric cancer model mice in in vivo 7,10 . Thus, NEDD1 probably be a promising target for HCC therapy. However, according to our knowledge, the potential impact of NEDD1 in HCC is still insu cient.
In this research, NEDD1 expression pro les and survival data were downloaded from TCGA to analyze the impact of NEDD1 in HCC. The results indicated that NEDD1 expression was higher in human HCC tissues than that in adjacent normal tissues. Meanwhile, overexpression of NEDD1 was considered to be an independent prognostic factor for poor OS by Kaplan-Meier survival curve and multivariable analysis in HCC patients, which indicated that NEDD1 could be used to predict the prognostic of HCC and an alternative therapy target for HCC.
Additionally, we found positive correlations between NEDD1 expression and the abundance of in ltrating B cells (r = 0.36, P = 6.08e-12), CD8+T cells (r = 0.334, P = 2.41e-10), CD4+T cells (r =0.471, P = 2.27e-20), Macrophages (r = 0.548, P = 3.90e-28), Neutrophils (r = 0.579, P = 2.62e-32), and DCs (r = 0.492, P = 3.93e-22) in HCC. However, our study demonstrated that the relevance between immune cell in ltration and OS in HCC patients was not statistically signi cant, and that NEDD1 was negatively associated with OS in patients with HCC. This suggested that NEDD1 was not responsible for the poor prognosis of patients with HCC by affecting the abundances of in ltrating immune cells in the tumor. Consistent with us, it has been reported that HCC immune microenvironment has different impacts on the prognostic of HCC because of its heterogeneity 24 . Meanwhile, Spearman correlation analysis and ssGSEA were conducted to detect the connection of NEDD1 expression with immune cells' in ltrating levels in HCC patients. As shown in Fig. 4D and 4E, Th2 cells and Th cells were positively related with NEDD1 expression. However, how NEDD1 regulates the abundance of immune cells in HCC is still unknown, which requires a further exploration. Our research also demonstrated that NEDD1 expression was signi cantly relevant to the expression of several immune-related genes, including CTLA-4, PD-L1 and PD-1. As immune checkpoint inhibitors are increasingly used in the management of unresectable HCC, there is an urgent need for molecules that predict their therapeutic e cacy 25 . Our ndings indicated that NEDD1 was most likely a predictor of the e cacy of immune checkpoint inhibitor therapy.
Consistent with previous studies 5,7 , GO and KEGG analyses indicated that NEDD1 was involved in cell cycle regulation and microtubule protein composition. Previous reports have suggested that the cell cycle plays a crucial impact in the development of HCC and we are able to ameliorate the prognostic of HCC patients by regulating hepatoma cell proliferation negatively via affecting the key biomarkers in the cell cycle regulation, including PCK1, MCM6, PKM2 and others [26][27][28][29] . Through the Linkedomics database we analyzed related genes co-expressed with NEDD1. The results reveled that NEED1 was signi cantly relevant to PPP1R12A, GTE3, ZNHIT1, DCI, et al. PPP1R12A, also known as MYPT1, which is a member of the myosin family phosphatase-targeting protein (MYPT) family and regulates smooth muscle contraction 30,31 . It has been reported that PPP1R12A is closely associated with processes, such as cell cycle 32 , development 33 , migration and cell adhesion 34 . A study by Xiao Zheng et al. also indicated that circPPP1R12A-encoded proteins can promote colon cancer development and metastasis 35 . All of this indicated that PPP1R12A might play a crucial impacted in the development of HCC. In turn, it has been demonstrated that NEDD1 played an important part in HCC development, but higher quality evidence was needed for validation.
Finally, we established a nomogram, for there wasn't a nomogram for HCC by integrating NEDD1 expression value with clinical characteristics for now. On the one hand, it helps to more precisely estimate the survival risk of individual patients; on the other hand, it helps clinicians guide patients' assessment and therapeutic decision-making 36 . This predictive model is desired for providing help to promote the individualized therapy of HCC patients.

Conclusions
In conclusion, this study is the rst to report the role of NEDD1 in HCC, suggesting that NEDD1 can be used to predict HCC prognostic and be a potential target therapeutic molecule. Besides, NEED1 probably be able to predict the therapeutic e cacy of immune checkpoint inhibitors for hepatocellular carcinoma. Availability of data and materials: All data included in this study are available including The Cancer Genome Atlas (TCGA, https://cancergenome.nih.gov/), International Cancer Genome Consortium (ICGC, https://dcc.icgc.org/) and The Genotype-Tissue Expression (GTEx https://gtexportal.org/home/) portal.

Competing interests:
The authors declare that they have no competing interests. Authors' Contribution: DXP, YJH, SX and LKW collected the information and drafted the initial manuscript. RPW, XGY and MSY participated in the discussion of the manu-script. DXP, YJH, MSY, HY and ZXM conceived and designed the study, performed data analysis, and wrote the manuscript. All authors approved the nal manuscript as submitted and agree to be accountable for all aspects of the work.

Figure 1
Analysis of differential expression of NEDD1 in pan-cancer A. Differential expression of NEDD1 in pancancer by using the data from TCGA and GTEx databases B. Differential expression of NEDD1 in pancancer by using the matched samples data from TCGA database C. Differential expression of NEDD1 in pan-cancer by using the unpaired samples data from TCGA database. Kaplan-Meier Plot analysis of NEDD1 in cancers with differential expression