(NON SOLL’S

ELSEVIER

01010

0010101010010

COMPUTATIONAL

001-01010010201010101011

101001

0101016

ANDSTRUCTURAL

0101001

1010101

0101001

1010101 10

10 0101001

0101010

BIOTECHNOLOGY

00 10100

6101010

010101010011010107

110101010010010000010010

JOURNAL

journal homepage: www.elsevier.com/locate/csbj

COMPUTATIONAL

ANDSTRUCTURAL

BIOTECHNOLOGY JOURNAL

01010100100101 0010101
1010101011001_0101001.201010
101010101110	0101001	01010
010101001001	0101001		10101
010101001011	0101001		10101
)101010101110	0101001		01010
)101010101100	0101001	J1010
[010101001001010101001		10101

Identifies microtubule-binding protein CSPP1 as a novel cancer biomarker associated with ferroptosis and tumor microenvironment

Check for updates

Wenwen Wang, Jingjing Zhang, Yuqing Wang, Yasi Xu, Shirong Zhang *

Translational Medicine Research Center, Key Laboratory of Clinical Cancer Pharmacology and Toxicology Research of Zhejiang Province, Affiliated Hangzhou First People’s Hospital, Zhejiang University School of Medicine, Cancer Center, Zhejiang University, Hangzhou, China

ARTICLE INFO

Article history: Received 24 April 2022 Received in revised form 19 June 2022 Accepted 21 June 2022 Available online 24 June 2022

Keywords:

CSPP1 Biomarker Ferroptosis

Tumor microenvironment Pan-cancer

ABSTRACT

Centrosome and spindle pole-associated protein (CSPP1) is a centrosome and microtubule-binding pro- tein that plays a role in cell cycle-dependent cytoskeleton organization and cilia formation. Previous studies have suggested that CSPP1 plays a role in tumorigenesis; however, no pan-cancer analysis has been performed. This study systematically investigates the expression of CSPP1 and its potential clinical outcomes associated with diagnosis, prognosis, and therapy. CSPP1 is widely present in tissues and cells and its aberrant expression serves as a diagnostic biomarker for cancer. CSPP1 dysregulation is driven by multi-dimensional mechanisms involving genetic alterations, DNA methylation, and miRNAs. Phosphorylation of CSPP1 at specific sites may play a role in tumorigenesis. In addition, CSPP1 correlates with clinical features and outcomes in multiple cancers. Take brain low-grade gliomas (LGG) with a poor prognosis as an example, functional enrichment analysis implies that CSPP1 may play a role in ferroptosis and tumor microenvironment (TME), including regulating epithelial-mesenchymal transition, stromal response, and immune response. Further analysis confirms that CSPP1 dysregulates ferroptosis in LGG and other cancers, making it possible for ferroptosis-based drugs to be used in the treatment of these can- cers. Importantly, CSPP1-associated tumors are infiltrated in different TMEs, rendering immune check- point blockade therapy beneficial for these cancer patients. Our study is the first to demonstrate that CSPP1 is a potential diagnostic and prognostic biomarker associated with ferroptosis and TME, providing a new target for drug therapy and immunotherapy in specific cancers.

@ 2022 Published by Elsevier B.V. on behalf of Research Network of Computational and Structural Bio- technology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

1. Introduction

Cancer is a leading cause of global health problems [1]. Although cancer treatment methods have recently improved,

clinical outcomes remain unsatisfactory due to side effects and drug resistance issues. Therefore, it is urgent to identify new sensi- tive biomarkers for the diagnosis and treatment of these cancer patients.

Abbreviations: TME, tumor microenvironment; CAF, cancer-associated fibroblasts; EMT, epithelial-mesenchymal transition; CSPP1, centrosome and spindle pole- associated protein; DLBC, diffuse large B-cell lymphoma; TCGA, The Cancer Genome Atlas; GTEx, Genotype-Tissue Expression; CPTAC, Clinical Proteomic Tumor Analysis Consortium; ROC, receiver operating characteristics; CNA, copy number alteration; KM, Kaplan-Meier; OS, overall survival; DSS, disease-specific survival; PFS, progression- free survival; ENCORI, Encyclopedia of RNA Interactomes; TISIDB, Tumor-Immune System Interactions DataBase; LGG, low-grade gliomas; PFI, progression-free interval; C- index, concordance index; DEGs, differentially expressed genes; GO, Gene Ontology; CC, cellular component; MF, molecular functions; BP, biological pathways; GSEA, Gene Set Enrichment Analysis; KEGG, Kyoto Encyclopedia of Genes and Genomes; GSVA, gene set variation analysis; TIMER, Tumor Immune Estimation Resource; MHC, major histocompatibility complex; TMB, tumor mutation burden; MSI, microsatellite instability; TIDE, Tumor Immune Dysfunction and Exclusion; CTL, cytotoxic T lymphocyte; ICB, immune checkpoint blockade; TGCT, testicular germ cell tumors, STAD, stomach adenocarcinoma; BRCA, breast invasive carcinoma; CHOL, cholangiocarcinoma; COAD, colon adenocarcinoma; ESCA, esophageal carcinoma; GBM, glioblastoma multiforme; HNSC, head and neck squamous cell carcinoma; LAML, acute myeloid leukemia; LIHC, liver hepatocellular carcinoma; PAAD, pancreatic adenocarcinoma; READ, rectum adenocarcinoma; THYM, thymoma; ACC, adrenocortical carcinoma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; KICH, kidney chromophobe; KIRC, renal clear cell carcinoma; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; OV, ovarian serous cystadenocarcinoma; PRAD, prostate cancer; SKCM, skin cutaneous melanoma; THCA, thyroid cancer; UCEC, endometrial cancer uterine corpus endometrial carcinoma; UCS, uterine carcinosarcoma; FAG, ferroptosis-associated gene; FDG, ferroptosis-driver gene; FSG, ferroptosis-suppressor gene.

* Corresponding author. E-mail address: shirleyz4444@zju.edu.cn (S. Zhang).

https://doi.org/10.1016/j.csbj.2022.06.046

Ferroptosis is a novel iron-dependent programmed cell death, that differs from typical cell death processes, mediated by lethal accumulation of lipid peroxides [2,3]. It involves a series of meta- bolic pathways and lipid peroxidation signaling pathways and is characterized by increased lipid peroxidation and reactive oxygen species, smaller mitochondria, and higher mitochondrial mem- brane density, but the change in nuclear morphology is not obvious [4]. Ferroptosis is essentially an antitumor mechanism that sup- presses tumor growth and kills these cells. Cancer cell ferroptosis not only promotes the antitumor response of immune cells but also affects the ferroptosis of immune cells themselves. Therefore, ferroptosis plays an important role in tumor occurrence, progres- sion, and prognosis [5,6].

Tumor microenvironment (TME) is the surrounding microenvi- ronment for tumor cells, mainly including peripheral blood vessels, stromal cells (cancer-associated fibroblasts (CAFs), endothelial cells, etc.), immune cells, and non-cellular components (cytokines, growth factors, hormones, and the extracellular matrix) [7-10]. Stromal components typically form a microenvironment conducive to tumor cell growth, including influencing metabolic pathways, inhibiting ferroptosis, inducing epithelial-mesenchymal transition (EMT), and regulating immune cell infiltration. Meanwhile, in the early stage, immune cells are recruited and activated by tumor cells to form an antitumor immune microenvironment and delay tumor development. With the continuous activation by tumor antigens, the relevant effector cells enter the depletion or remodel- ing stage, resulting in an immunosuppressive microenvironment. Different microenvironmental components interact and regulate each other, and are closely related to tumor progression and prog- nosis. Therefore, novel targets and biomarkers can be identified by identifying genes that influence ferroptosis and TME, leading to the selection of effective drugs and immunotherapy strategies to improve the prognosis of cancer patients.

Centrosome and spindle pole-associated protein (CSPP1), encoded by chromosome 8q13.2, is initially identified as a highly expressed proto-oncogene in diffuse large B-cell lymphoma (DLBC) [11]. It localizes to the interphase centrosome and mitotic spindle, migrates to the central spindle at the end of mitosis, and concen- trates at the midbody during telophase and cytokinesis, thus func- tioning throughout cell cycle progression. Overexpression or suppression of CSPP1 causes cell-cycle defects [11-14]. Interest- ingly, CSPP1 is not only localized to the centrosome and spindle in cycling cells but also interacts with Nephrocystin 8 to extend to the cilia axoneme in postmitotic resting cells, thus playing an important role in ciliogenesis. E3 ubiquitin-protein ligase UBR5- mediated ubiquitylation of CSPP1 is an underlying requirement for cilia localization. Meanwhile, interacting with the centrosomal protein of 104 kDa (CEP104), CSPP1 regulates axoneme length and cilia formation in the Hedgehog signaling pathway [15-17]. Muta- tion or loss of function in CSPP1 causes primary cilia abnormalities and ciliopathy, including Joubert syndrome and Meckel-Gruber syndrome [18-22]. Beyond cell cycle control and ciliogenesis, CSPP1 displays microtubule-independent but desmoplakin- dependent desmosome localization in apical-basal polarized epithelial cells and it is necessary for normal spheroid formation [23]. More importantly, CSPP1 has also been identified as a candi- date oncogene in luminal breast cancer; meanwhile, nuclear CSPP1 expression can define subtypes and clinical subgroups of basal-like breast carcinoma [24]. In addition to DLBC, CSPP1 is also identified as a putative hallmark gene associated with the malignancy of oral squamous cell carcinoma [25]. However, no comprehensive analy- ses of the expression, function, and clinical significance of CSPP1 as well as its correlation with ferroptosis and TME components have been performed.

In this study, we systematically analyzed CSPP1 expression and found that its aberrant expression is driven by genetic alterations,

DNA methylation, and miRNAs. Phosphorylation of CSPP1 protein may regulate its activity, especially at Ser424. In addition, CSPP1 strongly correlated with ferroptosis and TME components, poten- tially serving as a diagnostic and prognostic biomarker. Patients with CSPP1-associated tumors may benefit from ferroptosis- based drug therapy and immunotherapy by modulating ferroptosis and TME in several types of cancer.

2. Material and methods

2.1. Data collection

Transcriptomic data containing mRNA, miRNA, and associated clinical information were retrieved from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/repository) and UCSC XENA (https://xenabrowser.net) databases [26]. mRNA sequencing data in level 3 HTSeq-FPKM format was converted into TPM format, and miRNA sequencing data in level 3 BCGSC format was converted into RPM format; all data were downloaded from TCGA. UCSC XENA database contained both TCGA and Genotype-Tissue Expres- sion (GTEx) data, which were processed by the Toil process into TPM format [27]. The Oncomine (https://www.oncomine.org/re- source/login.html) image for CSPP1 expression was downloaded (threshold: p < 0.0001; fold change >2; gene rank: top 10%). R soft- ware (Version 3.6.3, https://cran.r-project.org/bin/windows/base/ old/3.6.3/) and ggplot2 R package (Version 3.3.3, https://cran.r- project.org/web/packages/ggplot2/index.html) were used to statis- tical analyses and visualization, respectively.

2.2. Differential expression analysis

RNA sequencing data in TPM or FPKM format for CSPP1 expres- sion in tissues and single cells were recorded from the Human Pro- tein Atlas portal (HPA, https://www.proteinatlas.org/) and visualized by radar plots. To compare CSPP1 and miRNA expression in normal and tumor tissues, an unpaired Wilcoxon rank-sum test was performed using TCGA and GTEx datasets. Histograms of CSPP1 protein expression and site phosphorylation levels in nor- mal and primary tumor tissues were downloaded from the UAL- CAN portal (https://ualcan.path.uab.edu/analysis-prot.html) using Clinical Proteomic Tumor Analysis Consortium (CPTAC) data [28].

2.3. Diagnostic analysis

The diagnostic value of CSPP1 was estimated using RNA- sequencing data from TCGA and visualized by a receiver operating characteristics (ROC) curve using the pROC R package (Version 1.17.0.1, https://cran.r-project.org/web/packages/pROC/index. html).

2.4. Genetic alterations and DNA methylation analysis

Histograms of CSPP1 mutation and copy number alteration (CNA) frequency in pan-cancer (TCGA, PanCancer Atlas) were downloaded from the cBioPortal (https://www.cbioportal.org/). Histograms of CSPP1 promoter methylation in normal and primary tumor tissues were downloaded from the UALCAN portal (https:// ualcan.path.uab.edu/index.html) [29]. Correlations between CSPP1, CNAs, and DNA methylation (TCGA, Firehose) were recorded from cBioPortal and visualized by heatmaps. Dot size together with transition color represented the degree of correlation. The larger the dot, the stronger the correlation. Red and blue dots represented positive and negative correlations, respectively. Kaplan-Meier (KM) plots of these alterations on survival probability, including

that of overall survival (OS), disease-specific survival (DSS), and progression-free survival (PFS), were downloaded from cBioPortal.

2.5. Correlations between CSPP1 and associated miRNAs

Spearman correlations between CSPP1 and associated miRNAs were recorded from the Encyclopedia of RNA Interactomes portal (ENCORI, https://rna.sysu.edu.cn/encori/index.php) (parameter setting: assembly, hg38; miRNA: all; CLIP-Data ≥ 3; pan- Cancer ≥ 1; programNum ≥ 2; target, CSPP1) and visualized by a heatmap [30].

2.6. Correlations between CSPP1 and clinical features

Correlations between CSPP1, pathologic stage, and histologic grade from TCGA database were analyzed using the Kruskal- Wallis test and visualized by violin plots. Correlations between CSPP1 and molecular subtype were also analyzed with the Kruskal-Wallis test and the violin plots were downloaded from Tumor-Immune System Interactions DataBase (TISIDB, https://cis. hku.hk/TISIDB/index.php) [31].

Correlations between CSPP1 and clinical features in brain low- grade gliomas (LGG) were analyzed using the Chi-squared test or Fisher’s exact test and visualized by a baseline datasheet.

2.7. Prognostic analysis of CSPP1 and associated miRNAs, model construction, and evaluation

Survival differences analyses of CSPP1 and associated miRNAs, including OS, DSS, and progression-free interval (PFI), were visual- ized by forest plots based on KM analyses. The Survivin R package (Version 3.2-10, https://cran.r-project.org/web/packages/sur- vivalAnalysis/index.html) was used for statistical analysis, and the survminer R package (Version 0.4.9, https://cran.r-project. org/web/packages/survminer/index.html) was used for visualization.

Univariate and multivariate Cox regression analyses were visu- alized by forest plots. Based on multivariate Cox regression, risk score plots were constructed using the ggrisk R package (Version 1.3, https://cran.r-project.org/web/packages/ggrisk/index.html). Nomograms were also designed using the rms R package (Version 6.2-0, https://cran.r-project.org/web/packages/rms/index.html) and survival R package [26]. Calibration curves and the concor- dance index (C-index) were evaluated by comparing predicted probabilities with observed events.

2.8. Functional enrichment analysis in LGG

CSPP1-associated differentially expressed genes (DEGs) in LGG were identified using the limma R package (Version 3.40.2, https://bioconductor.org/packages/release/bioc/html/limma.html) and visualized by a volcano plot [32]. Spearman correlations between CSPP1 and the top 20 DEGs were assessed and visualized by a heatmap.

DEGs were used for Gene Ontology (GO) enrichment analyses, including cellular components (CCs), molecular functions (MFs), and biological pathways (BPs). Gene Set Enrichment Analysis (GSEA) was also conducted to detect phenotypes and signaling pathways. Hallmark v7.2, GO c5 v7.2 (BPs, CCs, MFs), and Kyoto Encyclopedia of Genes and Genomes (KEGG) c2 v7.2 gene sets were used. Statistical analysis and graphical charting were performed using the clusterProfiler R package (Version 3.14.3, https://biocon- ductor.org/packages/release/bioc/html/clusterProfiler.html) [32,33].

2.9. Gene mutation and ferroptosis correlation analysis

Somatic mutations in LGG from TCGA database were analyzed using the maftools R package (Version 3.14, https://bioconductor. org/packages/release/bioc/html/maftools.html) and visualized by an oncoplot [34,35]. Ferroptosis-associated score was calculated with the gene set extracted from KEGG with the ssGSEA algorithm in the gene set variation analysis (GSVA) package (Version 1.34.0, https://bioconductor.riken.jp/packages/3.0/bioc/html/GSVA.html), and the difference between the driver score minus suppressor score was defined as the ferroptosis score to represent the ferrop- tosis status of samples [36]. Spearman correlations between CSPP1 and ferroptosis-related genes and ferroptosis scores were analyzed and visualized by heatmaps.

2.10. TME analysis and immune checkpoint blockade therapy prediction

TCGA datasets were used to analyze the Spearman correlations between CSPP1 and immune cells using the ssGSEA algorithm in the GSVA package. They were also used to estimate the stromal score, immune score, and ESTIMATE score using the ESTIMATE package (Version 1.0.13, https://bioinformatics.mdanderson.org/ estimate/index.html) [37,38]. Purity-adjusted Spearman correla- tions between CSPP1 and CAFs, endothelial cells, and immune checkpoints were recorded from the Tumor Immune Estimation Resource 2 portal (TIMER2, https://timer.cistrome.org) with XCELL or TIMER algorithm [39,40]. The Spearman correlations between CSPP1 and major histocompatibility complex (MHC) molecules, immune stimulator genes, immune inhibitor genes, tumor muta- tion burden (TMB) score, and microsatellite instability (MSI) score from the TCGA database were analyzed [41,42]. All corresponding correlations were visualized by heatmaps.

Tumor Immune Dysfunction and Exclusion (TIDE, https://tide. dfci.harvard.edu/) is a comprehensive score for tumor immune dysfunction and immune escape, including tumor-infiltrating cyto- toxic T lymphocyte (CTL) dysfunction and rejection by immune checkpoints. RNA-sequencing raw count data and corresponding clinical information from TCGA database were estimated using the TIDE algorithm to predict the potential immune checkpoint blockade (ICB) response. A low score indicated good efficacy [43,44].

3. Results

3.1. Aberrant expression of CSPP1 serves as a diagnostic biomarker among cancers

CSPP1 was widely present in all the tested tissues and cells. It was highly expressed in skeletal muscle, testis, and fallopian tube, as well as in respiratory ciliated cells, endometrial ciliated cells, and early spermatids; meanwhile, high expression was observed in testicular germ cell tumors (TGCT), stomach adenocarcinoma (STAD), and breast invasive carcinoma (BRCA) (Fig. 1A-C).

To compare CSPP1 expression in human adjacent normal versus 33 types of tumor tissues, TCGA datasets were used. CSPP1 was sig- nificantly upregulated in ten cancer types and downregulated in five from TCGA (Fig. 1D). In order to expand the sample size, we also introduced normal samples from the GTEx database. CSPP1 expression was increased in BRCA, cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), DLBC, esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), acute myeloid leukemia (LAML), LGG, liver hepatocellular carcinoma (LIHC), pancreatic adenocarcinoma (PAAD), rectum adenocarcinoma (READ), STAD, and thymoma

Fig. 1. Aberrant expression of CSPP1 serves as a diagnostic biomarker among cancers. (A) Radar Plot of CSPP1 expression in normal tissues basedon GTEx datasets from HPA portal. (B) Radar Plot of CSPP1 expression in single cells based on single-cell types dataset from HPA portal. (C) Radar Plot of CSPP1 expression in tumor tissues based on TCGA dataset from HPA portal. (D, E) Histogram of CSPP1 expression in 33 types of unpaired normal and tumor tissues from TCGA and TCGA plus GTEx database using Wilcoxon rank-sum test. ns: p > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001. (F) Heatmap of CSPP1 expression from Oncomine portal. (G, H) ROC analyses of differential CSPP1 expression in 27 types of upregulated (G) and downregulated (H) cancer from TCGA and GTEx databases. AUC > 0.9 was considered a high diagnostic value, 0.9 > AUC > 0.7 was median, and 0.7 > AUC > 0.5 was low.

Low tissue specificity: GTEx datasets

skeletal muscle Retinaele usTestis

RNA single cell type specificity: Cell type enhanced

Adrenal gland

Fallopian tube Pancreas

Distal tubular cells Paneth cells Urothelial cells Kupffer cells

Respiratory ciliated cells

Endometrial ciliated cells Early spermatids Rod photoreceptor cells Spermatocytes

Pituitary gland

Spleen

Lung

Melanocytes

Collecting duct cells

250

Excitatory neuronsInhibitory neurons Oligodendrocytes

Skin

Hofbauer cells

Ovary

Basal respiratory cells lonocyles

Late spermatids

Oligodendrocyte precursor cells Cone photoreceptor cells Astrocytes

Esophagus

200

Liver

Hepatocytes

Salivary gland

nTPM

Kidney

Langerhans cells

Microglial cells

Intestinal goblet cells Distal enterocytes

150

Spermatogonia

Prostate

Cervix

Cardiomyocytes

Thyroid gland

Exocrine glandular cells

100

Heart muscle

Granulosa cells

Adipose tissue

Breast

Pancreatic endocrine cells

Club cells

Vagina

Proximal enterocytos

Skeletal myocytos

Urinary.Blandfastine

Stomach Endometrium

Colon

Prostatic glandular cells

Cholangiocytes

Basal prostatic cells

Endometrial stromal cells

Low cancer specificity: TCGA datasets

Gastric mucus-secreting cells

nTPM

Glandular and luminal cells

Enteroendocrine cells

Sertoli cells

SKCM

TGCT

STAD

Alveolar cells type 2

Granulocytes

LIHC

3.5

BRCA

Hepatic stellate cells

Alveolar cells type 1

THCA

2.5

COADREAD

Macrophages

Dendritic cells

Monocytes

Proximal tubular cells

1.5

Undifferentiated cells

Horizontal cells

PRAD

Muller glia cells

Cytotrophoblasts

FPKM

Extravillous trophoblasts

Leydig cells Basal keratinocytes

HNSC

GBM

Theca cells Ductal cells

B-cells

Adipocytes Bipolar cells Smooth muscle cells

PAAD

UCEC

Squamous epithelial cells

Suprabasal keratinocytes Basal squamous epithelial cells

T-cells

Peritubler

KICHKIRCKIRP

LUADLUSC

Erythroid cells Endothelial cells

plasmasependular cells

Plasma cells

NK-cells Breast myoepithelial cells Syncytiotrophoblasts

CESC

BLCA

The expression of CSPP1 Log2 (TPM+1)

TCGA-ALL

Oncomine

Cancer VS. Normal

Normal

Analysis Type by Cancer

Tumor

Bladder Cancer

Brain and CNS Cancer

Breast Cancer

Cervical Cancer

ACC (79)

BLCA (19)

BLCA 408)

BRCA (113

BRCA (1090)

CESC (3)

CESC (304

CHOL (9

CHOL (36)

COAD (41)

COAD (454)

DLBC (48)

ESCA (11)

ESCA (161)

GBM (160)

HNSC (44)

HNSC (500)

KICH (24)

KICH (65)

KIRC (72)

KIRC (530)

KIRP (32)

KIRP (288)

LAML (151)

LGG (510)

LIHC (50)

LIHC (371)

LUAD (59)

LUSC (49)

LUAD (512)

LUSC (501)

MESO (86)

OV (376)

PAAD (4)

PAAD (177)

PCPG (3)

PCPG (179)

PRAD (52)

PRAD (492)

READ (10)

READ (165)

SARC (2)

SARC (259

SKCM (1

SKCM (468)

STAD (32

STAD (375)

TGCT (134)

THCA (58)

THCA (502)

THYM (2)

THYM (119)

UCEC (23)

UCEC (543)

UCS (56)

UVM (80)

Colorectal Cancer

Esophageal Cancer

Gastric Cancer

Head and Neck Cancer

XENA-TCGA_GTEx

Kidney Cancer

The expression of CSPP1 Log2 (TPM+1)

Leukemia

…

Liver Cancer

Lung Cancer

Lymphoma

Normal

Melanoma

Tumor

Myeloma

Other Cancer

Ovarian Cancer

Pancreatic Cancer

Prostate Cancer

Sarcoma

ACC (128)

ACC (77

BLCA (28)

BLCA (407

BRCA (292

BRCA (1099

CESC (13

CESC (306

CHOL (9

CHOL (36

COAD (349

COAD (290

DLBC (444

DLBC (47

ESCA (666

ESCA (182

GEM (1157

GBM (166

HNSC (44

HNSC (520

KICH (53

KICH (66

KIRC (100

KIRC (531

KIRP (60

KIRP (289 LAML (70

LAML (173

LGG (1152

LGG (523

LIHC (160

LIHC (371

LUAD (347)

LUSC (338

LUAD (515

LUSC (498

MESO (87)

OV (88)

OV (427

PAAD (171

PAAD (179

PCPG (3

PCPG (182

PRAD (152

PRAD (486

READ (318

READ (93

SARC (2

SARC (262

SKCM (813

SKCM (469

STAD (210

STAD (414 TGCT (165)

TGCT (154)

THCA (338)

THCA (512

HIM (446)

THYM (446

THYM (119

UCEC (101 UCEC (181)

UCS (78)

UCS (57)

UVM (79)

Significant Unique Analyses Total Unique Analyses

17 4

392

1 5 10

10 5

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

Sensitivity (TPR)

1.0

Sensitivity ( IPR)

1.0

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

0.8

0.6

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

Upregulation

0.2

AUC: 0.552

0.2

AUC: 0.997

0.2

AUC: 0.671

0.2

AUC: 0.702

0.2

AUC: 0.853

0.2

AUC: 0.621

0.2

AUC: 0.749

0.0

CI: 0.519-0.505

0.0

CI: 0.908-1.000

0.0

CI: 0.620-0.715

0.0

CI: 0.646-0.757

0.0

CI: 0.816-0.091

0.0

CI: 0.583-0.660

0.0

CI: 0.676-0.023

0.0 0.2 0.4 0.6 0.8 1.0

1-Specificity (FPR)

1-Specificity (FPR) READ

0.0 0.2 0.4 0.6 0.8 1.0 1-Specificity (FPR) STAD

0.0 0.2 0.4 0.6 0.8 1.0 1-Specificity (FPR) THYM

1.0

LAML

Sensitivity (TPR)

1.0

LGG

1.0

LIHC

1.0

PAAD

0.8

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

Sensitivity (TPR)

1.0

0.8

0.6

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.2

AUC: 0.705

0.2

AUC: 0.644

0.2

AUC: 0.579

0.2

CI: 0.641-0.769

CI: 0.617-0.670

AUC: 0.741

CI: 0.688-0.794

AUC: 0.697

0.2

0.0

CI: 0.531-0.627

0.0

CI: 0.625-0.769

AUC: 0.853

AUC: 0.762

0.0

CI: 0.824-0.883

0.0

CI: 0.723-0.801

0.0 0.2 0.4 0.6 0.8 1.0

1-Specificity (FPR)

0.0 0.2 0.4 0.6 0.8 1.0 1-Specificity (FPR)

1.0

ACC

1.0

CESC

1.0

KICH

1.0

KIRC

1.0

LUAD

1.0

LUSC

1.0

0.8

0.6

0.4

Downregulation

Sensitivity (TPR)

CSPP1

0.2

AUC: 0.739

0.2

AUC: 0.720

0.2

AUC: 0.803

0.2

AUC: 0.667

0.2

AUC: 0.791

0.2

AUC: 0.813

0.2

AUC: 0.828

0.0

CI: 0.657-0.820

0.0

CI: 0.563-0.878

0.0

CI: 0.718-0.888

0.0 0.2 0.4 0.6 0.8 1.0

0.0

CI: 0.600-0.725

0.0

CI: 0.760-0.822

CI: 0.783-0.842

CI: 0.791-0.864

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.0 0.2 0.4 0.6 0.8 1.0

0.0

0.0 0.2 0.4 0.6 0.8 1.0

1-Specificity (FPR)

1-Specificity (FPR) UCEC

1-Specificity (FPR)

1.0

PRAD

Sensitivity (TPR)

1.0

SKCM

Sensitivity (TPR)

1.0

TGCT

1.0

THCA

1.0

UCS

Sensitivity (TPR)

0.8

Sensitivity (TPR)

1.0

0.8

0.6

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.4

CSPP1

0.2

AUC: 0.652

0.2

CI: 0.604-0.700

AUC: 0.627

CI: 0.592-0.661

0.2

AUC: 0.979

CI: 0.966-0.993

0.2

AUC: 0.958

0.2

CI: 0.942-0.973

AUC: 0.700

0.2

AUC: 0.800

0.0

CI: 0.639-0.761

0.0

CI: 0.713-0.866

0.0 0.2 0.4 0.6 0.8 1.0

1-Specificity (FPR)

(THYM). In contrast, its expression was decreased in adrenocortical carcinoma (ACC), cervical squamous cell carcinoma and endocervi- cal adenocarcinoma (CESC), kidney chromophobe (KICH), renal clear cell carcinoma (KIRC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), ovarian serous cystadenocarci- noma (OV), prostate cancer (PRAD), skin cutaneous melanoma (SKCM), TGCT, thyroid cancer (THCA), endometrial cancer uterine corpus endometrial carcinoma (UCEC), and uterine carcinosarcoma (UCS) (Fig. 1E). Thereafter, we used the Oncomine datasets to val- idate CSPP1 expression patterns. Significantly elevated CSPP1 expression was observed in most cancer types; however, its expression was reduced in bladder cancer and kidney cancer (Fig. 1F). Combined with these databases, 27 cancers with differen- tial CSPP1 expression from GTEx plus TCGA database were used in subsequent analyses.

ROC analyses with TCGA or TCGA plus GTEx datasets indicated that the diagnostic values of CSPP1 were median or high in CHOL, DLBC, ESCA, HNSC, LAML, PAAD, STAD, THYM, ACC, CESC, KICH, LUAD, LUSC, OV, TGCT, THCA, and UCS (Fig. 1G, H). Our results indi- cated that aberrant CSPP1 expression serves as a diagnostic bio- marker among cancers.

3.2. Multi-dimensional mechanisms involving genetic alterations, DNA methylation, and miRNAs underly CSPP1 dysregulation

To determine the cause underlying CSPP1 dysregulation, we comprehensively analyzed the factors related to the expression of CSPP1, including genetic variation, DNA methylation, and associ- ated miRNAs. We first used cBioPortal to study the genetic varia- tion and found that among these 26 cancer types (COAD and READ were combined into COADREAD in the Portal), 21 contained mutations and 21 had CNAs (Fig. 2A). Except for UCEC (>5%), CSPP1 mutation frequencies were relatively low in most cancers (<5%). These mutations caused poor PFS in SKCM and good PFS in UCEC (Fig. S1A). CNA is the genetic variation most closely associated with CSPP1 expression [45]. It occurred more frequently in UCS, PRAD, LIHC, BRCA, and OV (>5%). CSPP1 positively correlated with CNAs (Fig. 2B; Table S1) and had worse DSS and PFS in COADREAD, as well as worse OS, DSS, and PFS in PAAD, PRAD, STAD, and UCEC (Fig. S1B).

Besides CNAs, DNA methylation also affects gene expression [46]. Promoter methylation was reduced with upregulated CSPP1 expression in BRCA, HNSC, and READ, while it was increased with downregulated CSPP1 expression in KIRC and LUSC from UALCAN portal (Fig. 2C; Fig. S2). CSPP1 expression negatively correlated with DNA methylation from cBioPortal (Fig. 2D and Table S2).

In addition to CNAs and DNA methylation, miRNAs also play important roles in regulating mRNA expression [47]. The ENCORI portal was used to search for miRNAs negatively associated with CSPP1. These miRNAs were found in most cancers except for ACC (Fig. 2E; Table S3). We further conducted differential expression analyses of these miRNAs across 14 cancers with data using TCGA datasets. CSPP1 upregulation was associated with the downregula- tion of miR-221-3p and miR-377-3p in BRCA; miR-145-5p and miR-125b-5p in STAD (Fig. 2F). Meanwhile, CSPP1 downregulation may be affected by upregulation of miR-222-3p in KICH; miR-425- 5p, miR-221-3p, miR-222-3p, miR-340-5p, miR-150-5p, miR-708- 5p in KIRC; miR-135b-5p, miR-222-3p, miR-27a-3p, miR-708-5p in LUAD; miR-135b-5p in LUSC; miR-135a-5p in PRAD; miR-105- 5p, miR-221-3p and miR-222-3p in THCA; miR-135b-5p and miR-27a-3p in UCEC (Fig. 2G). A prognosis-related forest plot indi- cated that in CSPP1-overexpressed cancers, downregulated miR- 221-3p in BRCA caused favorable DSS and PFI; meanwhile, down- regulated miR-145-5p and miR-125-5p in STAD had favorable OS, DSS, and PFI. Among cancers with low CSPP1 expression, upregu- lated miR-425-5p, miR-221-3p, miR-222-3p, and miR-708-5p in

KIRC caused poor OS, DSS, and PFI; whereas, upregulated miR- 27a-3p in LUAD had poor OS and DSS (Fig. 2H; Fig. S3). Together, CSPP1 dysregulation involves multi-dimensional mechanisms, including genetic alterations, DNA methylation, and miRNAs.

3.3. Phosphorylation of CSPP1 at specific sites may play a role in tumorigenesis, especially at Ser424

With datasets available in BRCA, GBM, HNSC, KIRC, LIHC, LUAD, OV, PAAD, and UCEC, CSPP1 protein expression was significantly increased in LIHC, PAAD, and UCEC compared with normal tissues; meanwhile, it was decreased in BRCA and HNSC (Fig. 3A; Fig. S4A).

Post-translational modification (PTM) is a key molecular mech- anism associated with the activity of the protein [48]. A higher S31 phosphorylation level was observed in BRCA, GBM, LIHC, and LUAD; meanwhile, a lower level was observed in KIRC and PAAD. S424 phosphorylation was increased in BRCA, HNSC, KIRC, and LIHC. S847 phosphorylation was increased in HNSC but decreased in PAAD. Increased S866 phosphorylation was observed in HNSC and LIHC but decreased in KIRC, LUAD, and PAAD. S885 phosphory- lation was increased in LIHC and decreased in KIRC. S931 phospho- rylation was increased in LIHC and decreased in KIRC and LUAD (Fig. 3B, C; Fig. S4B). No threonine or tyrosine phosphorylation was identified in these cancer types with UALCAN database. Together, these findings suggested that phosphorylation of CSPP1 at specific sites may play a role in tumorigenesis, especially at Ser424.

3.4. CSPP1 correlates with clinical features and outcomes in multiple cancers

Thereafter, we investigated CSPP1 expression at different patho- logic stages, histologic grades, and molecular subtypes. CSPP1 over- expression significantly correlated with advanced pathologic stage in READ, ACC, and KICH, and advanced histologic grade in HNSC and LIHC; however, high CSPP1 expression correlated with low his- tologic grade in KIRC (Fig. 4A-D). In addition, CSPP1 expression sig- nificantly differed with respect to molecular subtypes in BRCA, COAD, ESCA, LGG, LIHC, READ, LUSC, and UCEC (Fig. 4E, F). How- ever, no association was observed in other cancers (Fig. S5).

To monitor the clinical outcomes of CSPP1 differential expres- sion, a Cox regression analysis was performed with respect to patients’ prognoses. Results indicated that in CSPP1-upregulated cancers, CSPP1 overexpression was associated with poor OS, DSS, and PFI in LGG and LIHC. In CSPP1-downregulated cancers, decreased CSPP1 expression was associated with favorable OS, DSS, and PFI in ACC (Fig. 4G; Fig. S6). Overall, excessive CSPP1 expression is unfavorable in several cancer types, especially LGG, LIHC, and ACC.

To further evaluate whether CSPP1 was an independent risk fac- tor for prognosis, we used LGG as an example (p < 0.001 for OS, DSS, and PFI). The baseline datasheet showed that CSPP1 was sig- nificantly correlated with the WHO grade, IDH status, 1p/19q co- deletion, and histological type (Table S4). Univariate Cox regres- sion analyses further indicated that CSPP1 correlated with poor prognosis. Furthermore, multivariate Cox regression analyses con- firmed that CSPP1 overexpressed as an independent factor associ- ated with OS, DSS, and PFI (Fig. 4H, I; Fig. S7A, B, E, F). Based on the multivariate Cox regression analyses, nomogram prediction models were established (Fig. 4J; Fig. S7C, G). We performed cali- bration analysis on the nomograms to verify the validity of the pre- dictive models. The C-indexes of OS, DSS, and PFI indicate median accuracy (Fig. 4K; Fig. S7D, H). These results confirmed CSPP1 as an independent risk factor for LGG survival.

** * * ** * ៛ *
:* ** * * ** * ** * ** * ** * ** * * ** * * **	* p < 0.05 ** p < 0.01 Correlation 1.0 0.5 0.0 -0.5 -1.0
A * K * + * * ** * * . ** * * * *	* p < 0.05 ** p < 0.01 Correlation 1.0 0.5 0.0 -0.5 -1.0

Alteration Frequency

10%

Mutation

Amplification

Deep Deletion

Multiple Alterations

miR-145-5p

miR-425-5p

miR-135a-5p

miR-135b-5p

miR-381-3p

miR-194-5p

miR-19b-3p

2%-

miR-205-5p

miR-495-3p

221

31 +

miR-641

Mutation +

miR-802

CNA +

miR-105-5p

UCS (57)

UCEC (529) +

PRAD (494) +

BRCA (1084) +

LIHC (372) +

STAD (440) +

SKCM (444) +

LUAD (566) +

OV (584) +

PAAD (184) +

COAD (411) +

HNSC (523) +

DLBC (48) +

LUSC (487) +

ESCA (182) +

CESC (297) +

KIRC (511) +

ACC (91) +

GBM (592) +

LGG (514) +

TGCT (149) +

THCA (500) +

LAML (200) +

CHOL +

KICH +

THYM +

miR-532-5p

miR-221-3P

miR-222-3p

miR-377-3p

miR-27a-3p

miR-27b-3p

miR-410-3p

miR-340-50

miR-1197

Correlation with CNA

“p<0.05

miR-129-5p

PSDOT

miR-125b-5p

miR-150-50

UCS

PRAD

LIHC

BRCA

PAAD

LUAD

ESCA

HNSC

STAD

UCEC

COAD

READ

DLBC

SKCM

LUSC

CESC

KIRC

TGCT

LAML

GBM

LGG

miR-361-5p

miR-324-5p

miR-493-50

miR-28-50

Comelapior

“ps 0.00 ** p<8.45

miR-708-5p

miR-125a-5p

with mothylation

ACC BRCA

CESC

CHOL

READ

BOAD

ESCA

GBM

HNSC

KICH

KIRC

LAML

LGG

LIHC

LUAD

LUSC

PAAD

PRAD

SKCM

STAD

TOCT

THCA

THYM

UCEC

UCS

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

Promoter methylation

0.04

BRCA

6 0%-03

LUSC

0.045-

HNSC

0.045-

KIRC

0.00

LUAD

0.04

0.045

PAAD

30-03

0.04

READ

THICA

UCEC

6.452-04

1.8Je-05

0.035

2.260-11

0.04

20-04

0.045-

2.42e-04

0.035

0.04

0.07

0.035

0.04

0.035

1.480-02

0.035

0.04

Beta yalue

0.03

$0.035

Beta value

80.035

Beta value

Beta val

0.06

0.00

$0.035

Beta value

value

0.03

Beta value

80.035

Bota valu

025

0.03

0.05

0.03

0.02

0.025

0.04

10.025

0.025

30 025

0.025-

0.015

0.02

0.03

0.02

0.025

0.02-

0.01

Normal (n = 97)

Primary humor (n = 793)

0.015

Nonmal (n = 50)

Primary tumor (n =‘528)

0.015

Normal (n = 160)

Primary tumor (n = 324)

0.02

Normal (n = 32)

Primary tumor (-473)

0.015

Normal (n = 47)

Primary tumor (n = 370)

0.015

Normal (n = 10)

0 015

Normal (n=7)

Primary tumor (n = 164)

Primary tumor (n = 98)

0.02

Normal (n =50)

Primary tumor (n = 507)

0.015

Normal (n = 46)

Primary tumor (n = 430)

BRCA

The expression levels Log (RPM+1)

The expression levels

CHOL

The expression levels

COAD

ESCA

HNSC

LIHC

The expression levels Log2 (RPM+1)

The expression levels

Log (RPM+1)

Log2 (RPM+1)

The expression levels

Log2 (RPM+1)

E Normal

Normal

E Norma

Normal

5 Normal

E Tumor

Tumor

5 Tumor

E Tumor

Tumor

E Tumor

miR-135b-5p

miR-105-5p

miR-221-3p

miR-377-3p

miR-425-5p

miR-135a-5p

miR-205-5p

miR-27a-3p

miR-27b-3p

miR-221-3p

miR-361-5p

miR-194-5p

miR-105-5p

The expression levels Log2 (RPM+1)

STAD

KICH

KIRC

The expression levels

Log2 (RPM+1)

The expression levels Log2 (RPM+1)

UCEC

E Normal

Normal

₿ Normal

È Normal

Tumor

Ę Tumor

miR-145-5p

miR-125b-5p

miR-222-3p

miR-708-5p

miR-145-5p

miR-425-5p

miR-381-3p

miR-194-5p

miR-495-3p

miR-802

miR-221-3p

miR-222-3p

miR-377-3p

miR-410-3p

miR-340-5p

miR-150-5p

miR-361-5p

miR-493-5p

miR-708-5p

miR-1355-5p

miR-27a-3p

The expression levels Log2 (RPM+1)

LUAD

The expression levels Log2 (RPM+1)

LUSC

The expression levels Log2 (RPM+1)

PRAD

The expression levels Log2 (RPM+1)

THCA

Normal

E Normal

Normal

# Normal

Tumor

E Tumor

5 Tumor

Tumor

miR-1350-5p

miR-221-3p

miR-222-3p

miR-27a-3p

miR-708-5p

miR-145-5p

miR-135b-5p

miR-27a-3p

miR-145-5p

miR-1358-5p

miR-205-5p

miR-641

miR-221-3p

miR-222-3p

miR-27b-3p

miR-105-5p

miR-221-3p

miR-222-3p

miR-129-5p

H		OS		DSS		PFI
Cancers	miRNA	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
BRCA	hsa-miR-221-3p	1.30 (0.93-1.80)	0.124	1.88 (1.20-2.93)	0.005 **	1.44 (1.03-2.01)	0.032*
	hsa-miR-377-3p	1.20 (0.85-1.69)	0.299	1.54 (0.98-2.41)	0.061	1.37 (0.97-1.93)	0.074
STAD	hsa-miR-145-5p	1.83 (1.27-2.64)	0.001*	2.01 (1.24-3.26)	0.005 **	1.93 (1.28-2.90)	0.002 **
	hsa-miR-125b-5p	1.70 (1.22-2.35)	0.002*	1.88 (1.21-2.92)	0.005 **	1.64 (1.15-2.34)	0.006 **
KICH	hsa-miR-222-3p	0.61 (0.13-2.93)	0.535	0.35 (0.04-2.88)	0.326	0.46 (0.10-2.13)	0.319
KIRC	hsa-miR-425-5p	1.91 (1.40-2.60)	< 0.001 ***	2.64 (1.55-4.49)	< 0.001 ***	1.90 (1.26-2.87)	0.002 **
	hsa-miR-221-3p	2.36 (1.74-3.19)	< 0.001 ***	2.74 (1.86-4.02)	< 0.001 ***	1.88 (1.37-2.57)	< 0.001 ***
	hsa-miR-222-3p	2.24 (1.63-3.08)	< 0.001 ***	2.68 (1.76-4.07)	< 0.001 ***	1.82 (1.32-2.51)	< 0.001 ***
	hsa-miR-340-5p	0.89 (0.66-1.21)	0.466	0.80 (0.54-1.17)	0.253	0.78 (0.57-1.07)	0.122
	hsa-miR-150-5p	0.84 (0.62-1.15)	0.284	1.24 (0.85-1.81)	0.265	1.20 (0.85-1.69)	0.305
	hsa-miR-708-5p	1.50 (1.09-2.05)	0.012*	1.92 (1.31-2.82)	0.001 **	1.64 (1.18-2.27)	0.003 **
LUAD	hsa-miR-135b-5p	0.76 (0.56-1.03)	0.073	0.73 (0.50-1.07)	0.111	0.81 (0.62-1.07)	0.135
	hsa-miR-222-3p	1.17 (0.88-1.57)	0.28	1.35 (0.91-2.01)	0.132	1.38 (1.05-1.82)	0.021*
	hsa-miR-27a-3p	1.51 (1.08-2.12)	0.016*	1.59 (1.08-2.33)	0.019*	1.31 (0.99-1.73)	0.059
	hsa-miR-708-5p	0.88 (0.66-1.18)	0.387	0.72 (0.49-1.05)	0.089	1.16 (0.85-1.58)	0.349
LUSC	hsa-miR-135b-5p	1.26 (0.94-1.68)	0.119	1.49 (0.96-2.32)	0.074	1.25 (0.85-1.83)	0.252
PRAD	hsa-miR-135a-5p	0.39 (0.09-1.65)	0.202	0.00 (0.00-Inf)	0.998	0.66 (0.43-1.01)	0.055
THCA	hsa-miR-221-3p	0.38 (0.14-1.01)	0.053			1.74 (1.00-3.02)	0.051
	hsa-miR-222-3p	0.48 (0.18-1.29)	0.145			1.85 (1.03-3.33)	0.041*
UCEC	hsa-miR-135b-5p	1.45 (0.92-2.29)	0.106	1.44 (0.83-2.50)	0.2	1.23 (0.86-1.76)	0.246
	hsa-miR-27a-3p	1.25 (0.81-1.94)	0.317	1.27 (0.74-2.17)	0.38	1.21 (0.79-1.84)	0.379

Fig. 2. Multi-dimensional mechanisms involving genetic alterations, DNA methylation, and miRNAs underly CSPP1 dysregulation. (A) Histogram of genetic alteration frequency of CSPP1 from cBioPortal portal. (B) A heatmap of correlations between CSPP1 and CNAs. Dot size together with transition color represented the degree of correlation. The larger the dot, the stronger the correlation. Red and blue represented positive and negative correlations, respectively. * p < 0.05, ** p < 0.01. (C) Histograms of CSPP1 promoter methylation in normal and primary tumors with significant differences from UALCAN portal. 0.7 > Beta value > 0.5 was considered hyper-methylation, 0.3 > Beta value > 0.25 was hypo-methylation. (D) A heatmap of correlations between CSPP1 and DNA methylation from cBioPortal portal. (E) A heatmap of correlation between CSPP1 and predicted miRNAs from ENCORI portal. Red and blue words indicated upregulated and downregulated cancers, respectively. (F, G) Differential expression of negatively associated miRNAs from TCGA database. Red stars represent negatively correlated miRNAs of CSPP1. (H) A forest plot of the correlations between CSPP1- negatively associated miRNAs expression and survival probability, including OS, DSS, and PFI. * p < 0.05, ** p < 0.01, *** p < 0.001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Phosphorylation of CSPP1 at specific sites may play a role in tumorigenesis, especially at Ser424. (A) Histograms of CSPP1 expression in nine types of the normal and primary tumors with significant differences using CPTAC samples from UALCAN portal. (B) The schematic diagram of CSPP1 phosphorylation sites. Red and blue words indicated high and low protein expression, respectively. (C) Histograms of the phosphorylation site of CSPP1 in normal and primary tumors with significant differences. p < 0.05 was considered statistically significant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Protein expression in BRCA

2 -

1E-12

Protein expression in HNSC

2.424E-11

Protein expression in PAAD

3.791E-03

Protein expression in LIHC

1.252E-02

Protein expression in UCEC

1.399E-03

1 -

Z-value

0 -

-1-

-1

-1-

-2

-2-

-3

-2

Normal (n = 18)

Primary tumor (n = 125)

-3

Normal (n = 71)

Primary tumor (n = 108)

-4

Normal Primary tumor (n = 165) (n = 165)

-4

Normal Primary tumor (n = 74) (n = 137)

-4

Normal Primary tumor (n = 31) (n = 100)

CPTAC samples

BRCA/GBM/ HNSC/KIRC /LIHC/LUAD/PAAD S31

BRCA/HNSC /KIRC/LIHC/LUAD S424

GBM/HNSC/

KIRC/LIHC/ LUAD/PAAD

HNSC/KIRC/

BRCA/GBM/ HNSC/KIRC/

HNSC/PAAD

S847

S866

LIHC/LUAD LIHC/LUAD/PAAD

S885

S931

CSPP1

200

400

600

800

1000

1221

Phosphorylation of S31 in BRCA Phosphorylation of S424 in BRCA

Phosphorylation of S31 in GBM

Phosphorylation of $424 in HNSC Phosphorylation of S847 in HNSC Phosphorylation of S866 in HNSC

6.059E-07

2.599E-06

3.802E-02

2.915E-02

9.364E-10

3.771E-02

Z-value

-1

-2.

-1

-2

-1

-3

Normal (n = 18)

Primary tumor (n = 125)

-3

Normal (n = 18)

Primary tumor (n = 125)

-3

Normal (n = 10)

Primary tumor (n = 99)

-3

Normal (n = 70)

Primary tumor (n = 108)

-2

Normal (n = 70)

Primary tumor (n = 108)

-2

Normal (n = 70)

Primary tumor (n = 108)

CPTAC samples

Phosphorylation of S31 in KIRC

Phosphorylation of S424 in KIRC

Phosphorylation of S866 in KIRC

Phosphorylation of S885 in KIRC

Phosphorylation of S931 in KIRC

9.936E-35

5.122E-04

8.804E-22

3.335E-10

2.392E-07

Z-value

-2

-1

-2

-3

-2

-1-

-3

Normal (n = 83)

Primary tumor (n = 110)

-4

Normal (n =83)

Primary tumor (n = 110)

-3

Normal (n = 83)

Primary tumor (n = 110)

-2

Normal (n =83)

Primary tumor (n =‘110)

-2

Normal (n = 83)

Primary tumor (n = 110)

CPTAC samples

Phosphorylation of S31 in LIHC

Phosphorylation of S424 in LIHC

3.452E-13

Phosphorylation of S866 in LIHC

Phosphorylation of S885 in LIHC

Phosphorylation of S931 in LIHC

6.527E-24

1.827E-04

1.761E-30

1.358E-06

Z-value

-1

-2

-1.

-1

-2

-1

-3

-2

-3

-2

-4

Normal (n = 165)

Primary tumor (n = 165)

-3

Normal (n = 165)

Primary tumor (n = 165)

-3

Normal (n = 165)

Primary tumor (n = 165)

-4

Normal (n = 165)

Primary tumor (n = 165)

-3

Normal (n = 165)

Primary tumor (n = 165)

CPTAC samples

Phosphorylation of S31 in LUAD Phosphorylation of S866 in LUAD Phosphorylation of S931 in LUAD Phosphorylation of S31 in PAAD Phosphorylation of S847 in PAAD Phosphorylation of S866 in PAAD

1.6E-02

5.941E-04

3.5E-02

6.807E-03

2.535E-15

3.210E-05

Z-value

-1.

-1

-2

-3

Primary tumor (n =111)

Normal (n = 102) CPTAC samples

-3

Primary tumor (n =111)

Normal (n = 102) CPTAC samples

-3

Normal (n = 102)

Primary tumor (n =111)

-4

Normal (n = 74)

Primary tumor (n =‘137)

-3

Normal (n = 74)

Primary tumor (n = 137)

-3

Normal (n = 74)

Primary tumor (n = 137)

CPTAC samples

3.5. Functional enrichment indicates CSPP1 is potentially associated with ferroptosis and TME in LGG

Based on its unfavorable prognosis, LGG patients were divided into high and low CSPP1 expression groups, and mRNA expression patterns were compared. A total of 14 upregulated and 67 down- regulated genes were identified (Fig. 5A). Correlations between CSPP1 and the top 20 DEGs were visualized by a heatmap (Fig. 5B; Table S5).

GO and GSEA analyses were performed to investigate the func- tional mechanisms of CSPP1. CSPP1-related DEGs were enriched in (i) CCs: presynapse, transport vesicle, and synaptic membrane; (ii) MFs: passive transmembrane transporter activity, channel activity, and substrate-specific channel activity; and (iii) BPs: signal release, vesicle-mediated transport in synapse, and regulation of trans-

porter activity (Fig. 5C; Table S6). GSEA was also used to identify CSPP1-associated pathways. The results suggested that, in hallmark gene sets, CSPP1-related DEGs were positively related to cell cycle- related pathways (E2F targets, G2/M checkpoint, and mitotic spin- dle), cancer-related pathways (Notch signaling and TGF-ß signal- ing), EMT, and inflammatory response; meanwhile, they were negatively related to KRAS signaling DN, ferroptosis-related meta- bolic pathways (fatty acid metabolism, cholesterol homeostasis, and oxidative phosphorylation). For GO and KEGG gene sets, CSPP1-related DEGs were positively associated with cell cycle- related pathways (cell cycle checkpoint, chromosome segregation, and microtubule cytoskeleton organization involved in mitosis), cancer-related pathways (Notch signaling pathway, TGF-ß signal- ing pathway, and pathways in cancer), stromal-related pathway (ECM structural constituent, extracellular structure organization,

ECM receptor interaction, and focal adhesion), and immune- related pathways (B cell-mediated immunity, adaptive immune response, positive regulation of T cell proliferation, T cell activation involved in immune response, complement and coagulation cas- cades, intestinal immune network for IgA production, cytosolic DNA sensing pathway, and toll-like receptor signaling pathway);

meanwhile, they were negatively associated with ferroptosis- related metabolic pathways (steroid metabolic process, steroid biosynthetic process, response to metal ion, terpenoid backbone biosynthesis, and oxidative phosphorylation) (Fig. 5D; Table S7). These findings implied that CSPP1 may be involved in ferroptosis and TME.

READ

ACC

0.02

KICH

3.00-03

HNSC 1.76-04

LIHC

0.03

The expression of CSPP1 Log2 (TPM+1)

3.5

The expression of CSPP1 Log2 (TPM+1)

-110-09

The expression of CSPP1 Log2 (TPM+1)

78-K

The expression of CSPP1 Log2 (TPM+1)

5.16.45

3.0

2.5

2.0

1.5

1.0

0.5

Stage | Stage II Stage IIIStage IV Pathologic stage

Stage I Stage IIStage IIIStage IV Pathologic stage

Stage | Stage II Stage IIIStage IV Pathologic stage

G3&G4

Histologic grade

Histologic grade LGG

KIRC

The expression of CSPP1

BRCA p = 1.38e-13

The expression of CSPP1

COAD

The expression of CSPP1

ESCA

The expression of CSPP1

40-03

p = 1.23e-04

p = 1.58e-03

p = 1.29e-26

The expression of CSPP1 Log2 (TPM+1)

0.04

Log2CPM

₿

Basal

Her2

LumA

LumB

Normal

CIN

HM-SNV

HM-indel

CIN

ESCC

HM-SNV

HM-indel

Classic-like

Codel

G-CIMP-high

G-CIMP-low

Mesenchymal

-like

PA-like

Histologic grade

Molecular subtype

The expression of CSPP1

LIHC

The expression of CSPP1

READ

The expression of CSPP1

LUSC

The expression of CSPP1

Molecular subtype UCEC

p = 4.23e-04

p = 2.52e-02

p = 1.68e 03

p = 1.2e-02

Log2CPM

iCluster1

iCluster2

iCluster3

CIN

HM-SNV

HM-indel-

Basel

Classical

Primitive

Secretory

CN-High

CN-Low

MSI

POLE

0.0

2.5

5.0

7.5

10.0

Points

100

WHO grade

1.0

C-index = 0.801 (0.774-0.828)

Survival time

6000

IDH status

Observed fraction survival probability

Status

Mut

non-codel

0.8-

4000

. Alive

1p/19q codeletion

Oligodendroglioma

2000

Dead

Histological type

Astrocytoma

0.6

Laterality

Loft

Right

Midline

CSPP1

High

0.4

Risk score

Low

Risk group

Total Points

. Low

120

160

200

240

· High

Linear Predictor

0.2

1-Year

-1

-1.5

-0.5

0.5

1.5

2.5

3-Year

1-year Survival Probability

5-Year

0.95

09 0.85 0.80.750.70.65

0.0

Ideal line

3-year Survival Probability

0.8

0.6

0.4

0.2

0.0

0.2

0.4

0.6

0.8

1.0

5-year Survival Probability

0.8

0.6

0.4

0.2

Nomogram predicted survival probability

G	Molecular subtype	OS	Molecular subtype		Molecular subtype
Cancers	HR (95% CI)		P value	HR(95% CI)	P value	HR(95% CI)	P value
BRCA	1.47 (1.06-2.06)	1	0.022*	0.84 (0.54-1.3)	0.425	0.83 (0.59-1.6)	0.268
CHOL	0.59 (0.21-1.67)		0.318	0.29 (0.09-0.98)	0.046*	2.27 (0.87-5.95)	0.094
COAD DLBC	0.78 (0.53-1.15) 4.67 (0.56-38.88)		0.207 0.154	0.71 (0.43-1.15) 1.89 (0.19-18.34)	0.164 0.584	0.81 (0.56-1.16) 1.77 (0.54-5.84)	0.24 0.35
ESCA	1.52 (0.81-2.86)		0.194	1.68 (0.78-3.62)	0.186	1.43 (0.91-2.24)	0.125
GBM	0.83 (0.58-1.19)		0.303	0.78 (0.53-1.14)	0.197	0.56 (0.39-0.80)	0.002 **
HNSC	0.76 (0.58-0.99)		0.044*	0.68 (0.48-0.97)	0.035*	0.87 (0.64-1.17)	0.352
LAML	0.54 (0.34-0.86)		0.009 **
LGG	2.26 (1.56-3.27)		<0.001 ***	2.49 (1.68-3.69)	<0.001 ***	2.04 (1.52-2.74)	<0.001 ***
LIHC	1.7 (1.18-2.46)		0.005 **	1.73 (1.09-2.75)	0.02*	1.43 (1.06-1.91)	0.018*
PAAD	1.42 (0.93-2.16)		0.105	1.34 (0.83-2.15)	0.225	1.35 (0.87-2.1)	0.181
READ	0.61 (0.27-1.38)		0.236	0.39 (0.13-1.22)	0.106	1.35 (0.7-2.58)	0.367
STAD	1.23 (0.86-1.77)		0.262	1.33 (0.87-2.03)	0.181	0.83 (0.57-1.2)	0.315
THYM	0.31 (0.08-1.29)		0.107	0.12 (0.01-1.31)	0.082	0.21 (0.05-0.93)	0.04*
ACC	4.06 (1.40-11.73)		0.01*	3.83 (1.31-11.15)	0.014*	3.89 (1.38-10.92)	0.01*
CESC	1.6 (0.88-2.93)		0.126	1.77 (0.86-3.63)	0.118	2.25 (1.21-4.19)	0.011*
KICH KIRC	8.79 (1.10-70.43) 0.82 (0.6-1.13)		0.041* 0.224	>100 (0-Inf) 0.72 (0.49-1.06)	0.998 0.097	5.39 (1.16-25.08) 0.61 (0.42-0.86)	0.032* 0.006 **
LUAD	0.76 (0.57-1.03)		0.078	0.85 (0.58-1.24)	0.398	0.82 (0.59-1.13)	0.228
LUSC	0.87 (0.65-1.17)		0.367	0.75 (0.49-1.15)	0.185	0.84 (0.61-1.17)	0.309
OV	1.34 (1.01-1.78)		0.041*	1.28 (0.95-1.73)	0.108	0.89 (0.7-1.13)	0.356
PRAD SKCM	4.5 (0.95-21.38) 0.72 (0.52-1.00)		0.059 0.049*	>100 (0-Inf) 0.75 (0.54-1.05)	0.999 0.096	1.75 (1.15-2.65) 0.83 (0.65-1.05)	0.009 ** 0.116
TGCT	3.71 (0.34-40.89)		0.285	3.71 (0.34-40.89)	0.285	0.71 (0.34-1.47)	0.354
THCA	3.14 (1.14-8.65)		0.027*			0.71 (0.4-1.27)	0.254
UCEC	1.46 (0.89-2.38)		0.132	1.82 (0.96-3.42)	0.065	0.72 (0.49-1.05)	0.086
UCS	1.66 (0.81-3.37)		0.163	1.81 (0.84-3.88)	0.129	1.53 (0.77-3.03)	0.222

Characteristics	Total(N)	HR(95% CI)	P value	HR(95% CI)	P value
WHO grade (G3 vs G2)	452 (237 vs 215)	3.167 (2.071-4.842)	<0.001 ***	1.764 (1.053-2.957)	0.031*
IDH status (Mut vs WT)	506 (412 vs 94)	0.155 (0.107-0.225)	<0.001 ***	0.296 (0.171-0.510)	<0.001 ***
1p/19q codeletion (Non-codel vs codel)	509 (342 vs 167)	2.562 (1.602-4.098)	<0.001 ***	0.780 (0.370-1.645)	0.514
Primary therapy outcome	439
SD vs PD	143 vs 101	0.380 (0.248-0.583)	<0.001 ***	0.386 (0.226-0.658)	<0.001 ***
PR vs PD	62 vs 101	0.141 (0.057-0.351)	<0.001 ***	0.183 (0.064-0.522)	0.001 **
CR vs PD	133 vs 101	0.120 (0.055-0.263)	<0.001 ***	0.161 (0.071-0.364)	<0.001 ***
Histological type	509
Oligoastrocytoma vs Astrocytoma	128 vs 192	0.614 (0.384-0.982)	0.042*	1.373 (0.792-2.378)	0.259
Oligodendroglioma vs Astrocytoma	189 vs 192	0.558 (0.374-0.834)	0.004 **	0.681 (0.374-1.240)	0.209
Laterality	504
Middle vs Left	6 vs 248	1.008 (0.301-3.376)	0.99
Right vs Left	250 vs 248	0.768 (0.535-1.105)	0.155
CSPP1 (High vs Low)	509 (255 vs 254)	2.169 (1.494-3.148)	<0.001 ***	1.613 (1.000-2.601)	0.05

3.6. CSPP1 dysregulates ferroptosis in LGG and other cancer types

TP53 is the most commonly mutated gene associated with can- cers and its mutations have been reported to be closely associated with ferroptosis [49-55]. To verify the correlation between TP53 mutation and CSPP1 expression, somatic mutation analysis was performed according to CSPP1 expression in LGG. From the onco- plot, higher frequencies of TP53 and ATRX mutations and lower fre- quencies of CIC, FUBP1, NOTCH1, IDH2, and ZBTB20 mutations were observed in the high CSPP1 expression group (Fig. 6A, B). However, no association between ATRX, CIC, FUBP1, NOTCH1, IDH2, and ZBTB20 mutations and ferroptosis has been reported.

To further confirm the correlation between CSPP1 and ferropto- sis, 30 ferroptosis-associated genes (FAGs) were extracted from KEGG, including 18 ferroptosis-driver genes (FDGs) of ACSL1, ACSL4, ACSL6, ALOX15, ATG5, ATG7, FTL, LPCAT3, MAP1LC3A, MAP1LC3B, NCOA4, SAT1, SLC39A14, TF, TFRC, TP53, VDAC2, VDAC3 and 12 ferroptosis-suppressor genes (FSGs) of ACSL3, FTH1, FTMT, GCLC, GCLM, GPX4, GSS, HMOX1, PCBP1, SLC3A2, SLC7A11, SLC40A1 [56-59]. We found that CSPP1 positively correlated with most FAGs, but negatively correlated with two FDGs of FTL, MAP1LC3A, and three FSGs of FTH1, GPX4, and HMOX1 (Fig. 6C; Table S8).

Next, the overall scores of driver genes and suppressor genes were calculated by the ssGSEA algorithm, and the ferroptosis score obtained from driver score minus suppressor score was used to evaluate whether the function of CSPP1 was to activated or inhib- ited ferroptosis in cancers. The gene set and its corresponding algo- rithm have been proved to be able to predict ferroptosis status [56-59]. In CSPP1-upregulated cancers, CSPP1 positively correlated with ferroptosis score in LAML; meanwhile, negative correlations were observed in BRCA, GBM, LGG, LIHC, and THYM. In CSPP1- downregulated cancers, positive correlations were exhibited in ACC and LUAD, while negative correlations existed in KIRC, OV, and PRAD (Fig. 6D; Table S8). These tumor samples were further divided into high and low CSPP1 expression groups, and the ferrop- tosis score was further compared between the two groups. From the histogram, as an antitumor mechanism, ferroptosis was overall inhibited in pan-cancer (ferroptosis score < 1), except for LAML. Lower scores represented lower ferroptosis levels in BRCA, COAD, GBM, LGG, LIHC, THYM, OV, and PRAD and higher scores repre- sented higher ferroptosis levels in LAML and ACC were observed in the high CSPP1 expression group (Fig. 6E; Table S8). Thus, our findings revealed that CSPP1 dysregulates ferroptosis in LGG and other cancer types.

3.7. CSPP1-associated tumors are infiltrated in different TMEs, improving ICB therapeutic efficacy in specific cancers

Functional enrichment analysis also implied that CSPP1 may regulate the TME by influencing the immune response and stromal response. Therefore, we performed a pan-cancer analysis of the correlation between CSPP1 and these two components. We first

Fig. 4. CSPP1 correlates with clinical features and outcomes in multiple cancers. (A- D) Violin plots of correlation between CSPP1 expression and pathologic stage (A, B) and histologic grade (C, D) from TCGA database with significant differences using Kruskal-Wallis test. (E, F) Violin plots of correlation between CSPP1 expression and molecular subtype with significant differences from TISIDB portal. p < 0.05 was considered as a statistical difference between the two groups. (G) A forest plot of the correlations between CSPP1 expression and survival probability, including OS, DSS, and PFI. (H) A forest plot of univariate and multivariate Cox regression analyses with OS in LGG from TCGA database. p represented the overall difference. * p < 0.05, ** p < 0.01, *** p < 0.001. (I) Identification of CSPP1 as an independent risk factor in LGG. The upper portion scatters plot was survival time and survival status according to CSPP1 expression, and the middle portion scatters plot was risk score. (J) Construction of a prognostic nomogram in LGG. (K) Nomogram calibration analysis with prognostic data in LGG. C-index > 0.9 indicated highly accuracy, 0.9 > C-index > 0.7 was median, and 0.7 > C-index > 0.5 was low.

focused on CSPP1 and 24 types of immune cells using the ssGSEA algorithm. The results showed that CSPP1 negatively correlated with most immune cells across cancers but positively correlated with T helper cells, central memory T (Tcm) cells, and T helper 2 (Th2) cells. Of note, Tcm cells are also immunosuppressive cells. Next, stromal cell infiltration was assessed using the XCELL algo- rithm from the TIMER2 portal, mainly including CAF cells and endothelial cells [60]. There was a positive correlation between CSPP1 and CAFs in HNSC, LGG, LIHC, THYM, KICH, SKCM, and THCA, whereas an inverse correlation was observed in STAD, KIRC, LUSC, and TGCT. Moreover, CSPP1 was negatively correlated with endothelial cells in BRCA, DLBC, LGG, LIHC, STAD, THYM, PRAD, TGCT, and UCEC, whereas a positive correlation was noted in OV. Thereafter, we comprehensively calculated the TME score using the ESTIMATE package. CSPP1 was negatively associated with the stromal score, immune score, and ESTIMATE score in most cancers, while positively associated with these scores in LGG (Fig. 7A; Table S9). To sum up, CSPP1 comprehensively regulates the TME from both immune cell infiltration and stromal cell infiltration.

To further study the regulatory mechanism of CSPP1-related tumor infiltration, correlations between CSPP1 and three types of immunomodulators were investigated with TCGA datasets [61]. CSPP1 negatively correlated with MHCs and positively with immune stimulators and immune inhibitors in most cancers (Fig. 7B; Table S9). Among immune inhibitors, CD274 (PD-L1), CTLA4, HAVCR2, LAG3, PDCD1 (PD1), PDCD1LG2 (PD-L2), TIGIT, and SIGLEC15 are known immune checkpoints responsible for tumor immune escape. Combined with these immune checkpoint results and subsequent analysis using the TIMER2 portal, it was further confirmed that CSPP1 was positively correlated with immune checkpoints in BRCA, DLBC, ESCA, HNSC, LGG, LIHC, PAAD, READ, KICH, and KIRC, and negatively correlated with them in COAD, LAML, CESC, THCA, and UCEC (Fig. 7B, C; Table S9).

TMB and MSI are two emerging biomarkers associated with immunotherapy response. Tumor cells with high TMB or MSI scores have strong antigenicity and more neoantigens, thus pro- moting immune cell infiltration. Results showed that CSPP1 posi- tively correlated with TMB in LGG, STAD, and PRAD, but inversely correlated with it in COAD, LIHC, THCA, and UCEC. The correlation between CSPP1 and MSI was then investigated. LGG, READ, STAD, LUAD, and LUSC showed positive correlations, whereas DLBC presented a negative correlation (Fig. 7D; Table S9).

Tumor immunotherapy is a treatment that controls and elimi- nates tumors by reactivating and maintaining the tumor-immune cycle and restoring the normal antitumor immune response, including ICB and cell therapy. The effectiveness of ICB therapy depends not only on immune cell infiltration but also on immune checkpoints, TMB, and MSI. The close correlations between CSPP1 and immune checkpoints, TMB, and MSI implied that these CSPP1- associated tumor patients may respond well to immunotherapy. Therefore, the TIDE algorithm was used to predict the therapeutic effect of ICB from TCGA database. Results revealed that in CSPP1- upregulated cancers, the high CSPP1 expression group exhibited a lower TIDE score, including BRCA, DLBC, LGG, and STAD; mean- while, the low CSPP1 expression group exhibited a lower TIDE score in LIHC (Fig. 7E; Fig. SSA). In CSPP1-downregulated cancers, the high CSPP1 expression group exhibited a lower TIDE score, including CESC, KIRC, LUSC, PRAD, SKCM, TGCT, THCA, and UCES (Fig. 7F; Fig. S8B). It is suggested that these patients with low TIDE scores may benefit from ICB therapy.

4. Discussion

Cancer is the leading cause of morbidity and mortality world- wide. CSPP1 is a centrosome and microtubule-binding protein that

Fig. 5. Functional enrichment indicates that CSPP1 is potentially associated with ferroptosis and TME in LGG. (A) A volcano plot of CSPP1-related DEGs in LGG. Red and blue points indicated upregulated and downregulated genes, respectively. (B) A heatmap of correlation between CSPP1 and the top 20 DEGs. *** p < 0.001. (C) Bubble plots of GO enrichment. The X-axis represents the ratio of these DEGs, and the Y-axis represents the categories of DEGs. (D) Ridge plots of GSEA enrichment. p < 0.05 was considered the meaningful pathway. Red and blue indicated immune-related pathways and ferroptosis-related metabolic pathways, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

LGG

Change

CSPP1 Log2 (TPM+1)

LGG

Up: 14

Not

Low

Down: 67

High

COL11A1

-Log 10 (P.adj)

VCAM1

RPE65

COL28A1

CHST9

MSTN

IRX1

F2RL1

FOXJ1

MARCHF4

CHGA

GABRG2

PTPRN

CPLX2

SVOP

SYT4

SNCB

L1CAM

TRIM67

-1

Log2 (Fold Change)

Z-score

-2

passive

presynapse

transmembrane

signal release

transport vesicle

transporter activity

p.adjust

channel activity

p.adjust

vesicle-mediated

transport in synapse

p.adjust

2.56-07

synaptic membrane

substrate-specific

1.0e-04

regulation of

1e-07

channel activity

7.5e-05

transporter activity

2.0e-07

axon part

ion channel activity

5.0e-05

positive regulation

1.5e-07

glutamatergic

5e-08

gated channel

of ion transport

1.0e-07

2.5e-05

neurotransmitter

synapse

5.0e-08

activity

distal axon

Counts

ion gated channel

transport

Counts

regulation of

Counts

transport vesicle

activity

voltage-gated

exocytosis

membrane

synaptic vesicle

channel activity

exocytic vesicle

voltage-gated ion

cycle

neurotransmitter

channel activity

voltage-gated cation

secretion

calcium ion

synaptic vesicle

neuron projection

channel activity

regulated exocytosis

terminus

syntaxin-1 binding

glutamate secretion

0.100.120.140.160.180.200.22

0.060.080.100.120.14

0.100.120.140.160.160.20

GeneRatio

Hallmark

KEGG

HALLMARK E2F TARGETS

GO SISTER CHROMATID

KEGG CELL CYCLE

SEGREGATION

HALLMARK G2M

GO CELL CYCLE

KEGG NOTCH SIGNALING

CHECKPOINT

PATHWAY

HALLMARK MITOTIC

GO CHROMOSOME

KEGG TGF BETA

SPINDLE

SEGREGATION

SIGNALING PATHWAY

HALLMARK NOTCH

GO MICROTUBULE

CYTOSKELETON

KEGG PATHWAYS IN

SIGNALING

ORGANIZATION

CANCER

HALLMARK TGF BETA

INVOLVED IN MITOSIS

GO ATTACHMENT OF

KEGG CALCIUM SIGNALING PATHWAY

SIGNALING

SPINDLE MICROTUBULES

HALLMARK KRAS

TO KINETOCHORE

GO EXTRACELLULAR

KEGG ECM RECEPTOR

SIGNALING DN

NES

MATRIX STRUCTURAL

CONSTITUENT

NES

INTERACTION

NES

HALLMARK EPITHELIAL

GO EXTRACELLULAR

2.5

MESENCHYMAL

STRUCTURE

KEGG FOCAL ADHESION

TRANSITION

HALLMARK FATTY ACID

ORGANIZATION

0.0

METABOLISM

GO STEROID METABOLIC

-1

KEGG STEROID

-2

PROCESS

HALLMARK CHOLESTEROL

BIOSYNTHESIS

KEGG TERPENOID

-2.5

HOMEOSTASIS

GO STEROID

BIOSYNTHETIC PROCESS

BACKBONE

BIOSYNTHESIS

HALLMARK OXIDATIVE

PHOSPHORYLATION

GO RESPONSE TO METAL

KEGG OXIDATIVE

ION

PHOSPHORYLATION

HALLMARK IL6 JAK

STAT3 SIGNALING

GO B CELL MEDIATED IMMUNITY

KEGG COMPLEMENT AND

COAGULATION CASCADES

HALLMARK INTERFERON

KEGG INTESTINAL

GAMMA RESPONSE

GO ADAPTIVE IMMUNE

RESPONSE

IMMUNE NETWORK FOR

HALLMARK

GO POSITIVE

IGA PRODUCTION

INFLAMMATORY

REGULATION OF T CELL

KEGG CYTOSOLIC DNA

RESPONSE

GOT CELLACTIVATION

SENSING PATHWAY

HALLMARK INTERFERON

KEGG TOLL LIKE

ALPHA RESPONSE

INVOLVED IN IMMUNE

RECEPTOR SIGNALING

RESPONSE

PATHWAY

HALLMARK COAGULATION

GO REGULATION OF T

CELL ACTIVATION

KEGG B CELL RECEPTOR

SIGNALING PATHWAY

-1

plays a role in cell cycle-dependent cytoskeleton organization and cilia formation. Although there is increasing evidence that CSPP1 may play a role in tumorigenesis, its specific role across different cancers remains unclear. This study systematically analyzed CSPP1 expression and demonstrated that its aberrant expression in 27 cancer types is driven by multi-dimensional mechanisms. CSPP1

correlates with clinical features and serves as a potential diagnostic and prognostic biomarker as well as the target for ferroptosis- based drug therapy and immunotherapy.

To explore how CSPP1 influences the progress and prognosis of cancer, its effects on ferroptosis and TME were studied. Function enrichment demonstrated that CSPP1 was involved in ferroptosis-

Fig. 6. CSPP1 dysregulates ferroptosis in LGG and other cancer types. (A) Oncoplot of somatic mutant landscape in high and low CSPP1 expression groups in LGG. * p < 0.05, ** p < 0.01, *** p < 0.001. (B) Histograms of gene mutants comparison in high and low CSPP1 expression groups by chisq.test with significant differences. (C, D) Heatmaps of correlation between CSPP1 and FAGs and ferroptosis-associated scores. Dot size together with transition color represented the degree of correlation. The larger the dot, the stronger the correlation. Red and blue dots represented positive and negative correlations, respectively. * p < 0.05, ** p < 0.01. (E) Histograms of ferroptosis scores between high and low CSPP1 expression groups from TCGA database. p < 0.05 was considered statistically significant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

9635

Altered in 483 (95.08% ) of 508 samples.

Frame_Shift_Del

= In_Frame_Del

Groups

Missense_Mutation

= Translation_Start_Site

High exp

Frame_Shift_Ins

In_Frame_Ins

Splice_Site

· Multi_Hit

Low exp

Gene mutations

# Nonsense_Mutation

384

0.0 0.2 0.4 0.6 0.8 1.0

CSPP1

IDH1

76%

High

101

150

TP53

45%

TP53

ATRX ***

33%

Low

172

p = 0

CIC ***

20%

High

143

108

ATRX

TTN

12%

p = 0

FUBP1 ***

Low

190

NOTCH1*

High

223

CIC

PIK3CA

Low

179

p = 0

MUC16

EGFR

High

245

FUBP1

NF1

Low

221

p = 0

PIK3R1

PTEN

High

239

NOTCH1

ARID1A

Low

225

p = 2.8e-2

IDH2

FLG

High

249

IDH2

ZBTB20

Low

232

p = 4e-4

NIPBL

RYR2

High

249

ZBTB20

BCOR

Low

231

p = 2e-4

CSPP1

Groups

Mutant

ACSL1

ACSL4

** **

ACSL6

ALOX15 **

ATG5

ATG7

FTL

LPCAT3

MAP1LC3A

MAP1LC3B

NCOA4

* p < 0.05

SAT1

SLC39A14

p < 0.01

Correlation

TFRC

1.0

TP53

VDAC2 **

. *

0.5

VDAC3

ACSL3

0.0

FTH1

FTMT

-0.5

GCLC **

GCLM **

-1.0

GPX4

GSS

* *

HMOX1

PCBP1

SLC3A2

SLC7A11

SLC40A1

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

Driver score

p < 0.05

** p < 0.01

Suppressor score

Correlation

Ferroptosis score **

1.0

0.5

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

0.0

-O.S

0.8

-1.0

6.8e-03

Ferroptosis score

0.4

3.9e-02

2.2e-02 ns

4.5e-02

7.4e-12

3.6e-03

4.4e-02

4.40-02

6.0e-03

1.6e-03

0.0

Low CSPP1

₿

High CSPP1

-0.4

-0.8

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

related metabolic pathways. Mutation analyses further indicated that CSPP1 was closely associated with TP53 mutation, which has been reported to be associated with cancer and ferroptosis, thus

speculating that CSPP1 might correlate with ferroptosis. At present, studies on ferroptosis-associated gene mutations are mainly lim- ited to the oncogenes and tumor suppressor genes, including onco-

Fig. 7. CSPP1-associated tumors are infiltrated in different TMEs, improving ICB therapeutic efficacy in specific cancers. (A) A heatmap of correlations between CSPP1 and 24 types of immune cells from TCGA database, CAFs and endothelial cells from TIMER2 portal using XCELL algorithm, and TME scores from TCGA database using the ssGSEA algorithm. (B) A heatmap of correlations between CSPP1 and immunomodulators, including MHC molecules, immune stimulator genes, and immune inhibitor genes from TCGA database. (C) A heatmap of correlations between CSPP1 and immune checkpoints from the TIMER2 portal. (D) A heatmap of correlations between CSPP1 and TMB score, MSI score from TCGA database. Dot size together with transition color represented the degree of correlation. The larger the dot, the stronger the correlation. Red and blue dots represented positive and negative correlation, respectively. * p < 0.05, ** p < 0.01. (E, F) Histograms of CSPP1-associated ICB therapeutic effect between high and low CSPP1 expression groups from TCGA database by TIDE algorithm with a significant difference. A low score indicated good efficacy. p < 0.05 was considered statistically significant. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

aDC

* *

B cells

CD8 T cells

Cytotoxic cells

Eosinophils

TCGA

MHC molecules

** **

iDC

Macrophages **

Mast cells

HĽ

* p < 0.05

TCGA

Immune cells

Neutrophils

NK CD56bright **

NK CD56dim

p < 0.01

NK cells **

pDC

Correlation

T cells

1.0

T helper cells

Tcm

0.5

Tem

TFH

** *

0.0

* p < 0.05

Tgd

ICC

** p < 0.01

Th1 cells

-0.5

Immune stimulators

Th17 cells

Correlation

XCELL Stromal cells

Th2 cells

-1.0

1.0

TReg

TCGA

0.5

CAFs

0.0

Endothelial cells

-0.5

TCGA

TME

StromalScore

-1.0

ImmuneScore **

ESTIMATEScore **

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

AML LA

LIHC

PAAD

REA

STAD

THYM

ACC

CESC

KICH

IRC KI

PRAD

SKCM

TGCT

THCA

UCEC

UCS

Immune checkpoints

CD274

p <0.05

CTLA4 *

p < 0.01

TIMER2

HAVCR2

Correlation

ADO

LAG3

1.0

PDCD1

PDCD1LG2

0.5

Immune Inhibitors

TIGIT

SIGLEC15

0.0

IL1

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

TCGA

-0.5

-1.0

TCGA

TMB score MSI score

*p< 0.05

Correlation

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC CESC

KICH

KIRC

LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

SIGLEC15

BRCA

CHOL

COAD

DLBC

ESCA

GBM

HNSC

LAML

LGG

LIHC

PAAD

READ

STAD

THYM

ACC

CESC

KICH

KIRC LUAD

LUSC

PRAD

SKCM

TGCT

THCA

UCEC

UCS

-4.0

Responder

BRCA

Responder

DLBC

True

184

160

True

Responder

LGG

True

120

Responder

LIHC

STAD

True

Responder

True

False

365

388

False

135

166

False

137

122

False

111

124

2.4e-04

6.7e-03

9e-04

3.3e-02

1.1e-03

TIDE score

-1

-2

-1

-2

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

-2

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

Responder

CESC

KIRC

True

Responder

True

106

Responder

LUSC

PRAD

True

Responder

True

118

False

116

False

159

184

False

169

173

False

130

156

1.1e-02

1.1e-04

3.4e-02

3.4e-03

TIDE score

2.5

TIDE score

-1

-2

-2.5

-1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

Responder

SKCM

True

Responder

TGCT

THCA

UCEC

True

Responder

True

Responder

True

119

False

186

189

False

172

193

False

153

184

6.7e-03

3.7e-02

8.5e-06

4.1e-04

TIDE score

-2

-1

-2

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

High CSPP1

Low CSPP1

genes of PIK3CA, KRAS, NEDD4, VDAC2/3, DJ1, PDK4, and tumor sup- pressor genes of TP53, BAP1, KEAP1, ARF. Activating mutations in oncogenes and inactivating mutations in tumor suppressor genes regulate the expression of FAGs, and generally tend to inhibit fer- roptosis and promote tumor progression. However, mutations in tumor suppressors of the E-cadherin-NF2-Hippo axis, VHL, and

oncogenes of EGFR and IDH1 render cancer cells vulnerable to fer- roptosis in one or more cancer types [62,63]. Our correlation anal- yses between CSPP1 and FAGs and ferroptosis-associated scores further confirmed that CSPP1 was indeed involved in the regulation of ferroptosis in pan-cancer. Combined with the prognostic data, we believed that in CSPP1-upregulated cancers, CSPP1 overexpres-

sion inhibited ferroptosis, including BRCA, GBM, LGG, LIHC, and THYM, thus promoting tumor cell growth and leading to poor prognoses in BRCA, LGG, and LIHC. Meanwhile, its overexpression promoted ferroptosis in LAML and led to cancer with a good prog- nosis. These results were consistent with the progression and prog- nosis results of CSPP1. However, there were some exceptions. For example, patients with lower expression of CSPP1 showed lower ferroptosis level but better prognosis in ACC, and higher ferroptosis levels but still led to cancer in OV and PRAD. Thus, there should be other mechanisms influencing the progression and prognosis of these cancers. For CSPP1-associated tumors with suppressed fer- roptosis, drug-induced ferroptosis through the CSPP1-FDGs or CSPP1-FSGs axis may inhibit tumor progression and thus improve prognosis. Therefore, the possibility of drug therapy for CSPP1- associated tumor patients by regulating ferroptosis is proposed.

In addition to regulating ferroptosis, CSPP1 may have a regula- tory role in the TME by affecting EMT, stromal-related pathways, and immune-related pathways. In recent years, an increasing num- ber of studies have linked microtubule-associated genes to immune infiltration. High expression of Targeting Protein for Xenopus kinesin-like protein 2 (TPX2) and Tubulin alpha-1C chain (TUBA1C) increases immune cell infiltration in LIHC, LUAD, and LGG, respectively, and is associated with poor prognosis [64-66]. A high level of microtubule-associated protein Tau is inversely cor- related with the vascular and immune contents, delaying tumor growth in gliomas [67]. Increased expression of microtubule inter- acting and trafficking domain containing 1 (MITD1) indicates a poor prognosis with decreased NK cell infiltration in LIHC and increased CD8+ T cells infiltration in KIRC [68,69]. Spindle and kinetochore-associated protein (Ska) complex negatively and pos- itively correlate with immune cell infiltration in BRCA and LIHC, respectively [70,71]. In our study, CSPP1 comprehensively regu- lated the TME from both immune cell infiltration and stromal cell infiltration. CSPP1 was negatively correlated with immune scores, stromal scores, and TME scores for most cancers. Moreover, it was also negatively correlated with MHCs and positively associ- ated with immune stimulators and immune inhibitors, including immune checkpoints. CSPP1 expression also significantly corre- lated with TMB and MSI in specific cancers. ICB therapy prediction confirmed that these cancer patients could benefit from ICB ther- apy, thus promoting a favorable prognosis. Specifically, in CSPP1- upregulated cancers, low TME scores and high levels of immune checkpoints expression indicated immune infiltration was greatly suppressed, leading to tumor growth and poor prognosis. There- fore, ICB therapy promoting immune infiltration is effective for patients in the high CSPP1 expression group of BRCA, DLBC, LGG, and STAD, but more effective for patients in the low CSPP1 expres- sion group of LIHC. In CSPP1-downregulated cancer, low CSPP1 expression had higher levels of immune cell infiltration and lower levels of immune checkpoints expression, thus these cancers them- selves were in a favorable prognostic immune microenvironment, and ICB therapy may be more effective for the early treatment of these patients, including CESC, KIRC, LUSC, PRAD, SKCM, TGCT, THCA, and UCEC.

However, several limitations still remain. At present, our study on the regulation of CSPP1 on ferroptosis and TME, as well as the subsequent potential drug treatment and ICB therapy are limited to bioinformatics, which provides a reference to basic experiments, but basic experiments are still necessary for follow-up research. In addition, the relatively small sample size is also one of the main reasons for data deviation. For example, as described above, the regulation of CSPP1 expression involves multiple factors, which leads to the inconsistency between any individual factor and CSPP1 expression, and the small sample size increases this inconsistency. In addition, the inconsistent expression between CSPP1 mRNA and protein is the same case. Therefore, further basic experiments and

more clinical samples are required to explore the direct functional mechanism of CSPP1 affecting cancer progression and prognosis through ferroptosis and TME function.

5. Conclusion

In conclusion, our study is the first to demonstrate that CSPP1 is a potential diagnostic and prognostic biomarker associated with ferroptosis and TME, providing a new target for ferroptosis-based drug therapy and immunotherapy in specific cancer types.

Author contributions

SZ contributed to the conception of the study. WW designed the study and wrote the manuscript. WW, JZ, YW, and YX had full access to all of the data in the study and take responsibility for the integrity of the data. WW, JZ, YW, and YX performed the statis- tical analyses. SZ obtained funding. All authors contributed to the article and approved the submitted version.

Funding information

This work was supported by the National Natural Science Foun- dation of China (81773242), Zhejiang Provincial Natural Science Foundation of China (LY19H160032), and Major Project of Hang- zhou Science and Technology Bureau (20180417A01).

Declaration of Competing Interest

The authors declare that they have no known competing finan- cial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We thank the members of our research center for inspiring discussion.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.csbj.2022.06.046.

References

[1] Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov 2022;12:31-46.

[2] Dixon SJ, Stockwell BR. The hallmarks of ferroptosis. Ann Rev Cancer Biol 2019;3:35-54.

[3] Moujalled D, Strasser A, Liddell JR. Molecular mechanisms of cell death in neurological diseases. Cell Death Differ 2021;28:2029-44.

[4] Stockwell BR, Friedmann Angeli JP, Bayir H, Bush AI, Conrad M, et al. Ferroptosis: A regulated cell death nexus linking metabolism, redox biology, and disease. Cell 2017;171:273-85.

[5] Wang W, Green M, Choi JE, Gijon M, Kennedy PD, et al. CD8(+) T cells regulate tumour ferroptosis during cancer immunotherapy. Nature 2019;569:270-4.

[6] Sun LL, Linghu DL, Hung MC. Ferroptosis: a promising target for cancer immunotherapy. Am J Cancer Res 2021;11:5856-63.

[7] Turley SJ, Cremasco V, Astarita JL. Immunological hallmarks of stromal cells in the tumour microenvironment. Nat Rev Immunol 2015; 15:669-82.

[8] Bruni D, Angell HK, Galon J. The immune contexture and Immunoscore in cancer prognosis and therapeutic efficacy. Nat Rev Cancer 2020;20:662-80.

[9] Zhou J, Kryczek I, Li S, Li X, Aguilar A, et al. The ubiquitin ligase MDM2 sustains STAT5 stability to control T cell-mediated antitumor immunity. Nat Immunol 2021;22:460-70.

[10] Zou W. Immune regulation in the tumor microenvironment and its relevance in cancer therapy. Cell Mol Immunol 2022; 19:1-2.

[11] Patzke S, Hauge H, Sioud M, Finne EF, Sivertsen EA, et al. Identification of a novel centrosome/microtubule-associated coiled-coil protein involved in cell- cycle progression and spindle organization. Oncogene 2005;24:1159-73.

W. Wang, J. Zhang, Y. Wang et al.

[12] Patzke S, Stokke T, Aasheim HC. CSPP and CSPP-L associate with centrosomes and microtubules and differently affect microtubule organization. J Cell Physiol 2006;209:199-210.

[13] Asiedu M, Wu D, Matsumura F, Wei Q. Centrosome/spindle pole-associated protein regulates cytokinesis via promoting the recruitment of MyoGEF to the central spindle. Mol Biol Cell 2009;20:1428-40.

[14] Zhu L, Wang Z, Wang W, Wang C, Hua S, et al. Mitotic protein CSPP1 interacts with CENP-H protein to coordinate accurate chromosome oscillation in mitosis. J Biol Chem 2015;290:27053-66.

[15] Frikstad KM, Molinari E, Thoresen M, Ramsbottom SA, Hughes F, et al. A CEP104-CSPP1 complex is required for formation of primary cilia competent in hedgehog signaling. Cell Rep 2019;28. 1907-22 e6.

[16] Patzke S, Redick S, Warsame A, Murga-Zamalloa CA, Khanna H, et al. CSPP is a ciliary protein interacting with Nephrocystin 8 and required for cilia formation. Mol Biol Cell 2010;21:2555-67.

[17] Shearer RF, Frikstad KM, McKenna J, McCloy RA, Deng N, et al. The E3 ubiquitin ligase UBR5 regulates centriolar satellite stability and primary cilia. Mol Biol Cell 2018;29:1542-54.

[18] Latour BL, Van De Weghe JC, Rusterholz TD, Letteboer SJ, Gomez A, et al. Dysfunction of the ciliary ARMC9/TOGARAM1 protein module causes Joubert syndrome. J Clin Invest 2020; 130:4423-39.

[19] Ben-Omran T, Alsulaiman R, Kamel H, Shaheen R, Alkuraya FS. Intrafamilial clinical heterogeneity of CSPP1-related ciliopathy. Am J Med Genet A 2015;167A:2478-80.

[20] Tuz K, Bachmann-Gagescu R, O’Day DR, Hua K, Isabella CR, et al. Mutations in CSPP1 cause primary cilia abnormalities and Joubert syndrome with or without Jeune asphyxiating thoracic dystrophy. Am J Hum Genet 2014;94:62-72.

[21] Shaheen R, Shamseldin HE, Loucks CM, Seidahmed MZ, Ansari S, et al. Mutations in CSPP1, encoding a core centrosomal protein, cause a range of ciliopathy phenotypes in humans. Am J Hum Genet 2014;94:73-9.

[22] Akizu N, Silhavy JL, Rosti RO, Scott E, Fenstermaker AG, et al. Mutations in CSPP1 lead to classical Joubert syndrome. Am J Hum Genet 2014;94:80-6.

[23] Sternemalm J, Geimer S, Frikstad KA, Schink KO, Stokke T, et al. CSPP-L associates with the desmosome of polarized epithelial cells and is required for normal spheroid formation. PLoS ONE 2015;10:e0134789.

[24] Sternemalm J, Russnes HG, Zhao X, Risberg B, Nord S, et al. Nuclear CSPP1 expression defined subtypes of basal-like breast cancer. Br J Cancer 2014;111:326-38.

[25] Ribeiro IP, Marques F, Barroso L, Rodrigues J, Caramelo F, et al. Genomic profile of oral squamous cell carcinomas with an adjacent leukoplakia or with an erythroleukoplakia that evolved after the treatment of primary tumor: A report of two cases. Mol Med Rep 2017;16:6780-6.

[26] Liu J, Lichtenberg T, Hoadley KA, Poisson LM, Lazar AJ, et al. An integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell 2018;173. 400-16 e11.

[27] Vivian J, Rao AA, Nothaft FA, Ketchum C, Armstrong J, et al. Toil enables reproducible, open source, big biomedical data analyses. Nat Biotechnol 2017;35:314-6.

[28] Chen F, Chandrashekar DS, Varambally S, Creighton CJ. Pan-cancer molecular subtypes revealed by mass-spectrometry-based proteomic characterization of more than 500 human cancers. Nat Commun 2019; 10:5679.

[29] Chandrashekar DS, Bashel B, Balasubramanya SAH, Creighton CJ, Ponce- Rodriguez I, et al. UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses. Neoplasia 2017;19:649-58.

[30] Li JH, Liu S, Zhou H, Qu LH, Yang JH. starBase v2.0: decoding miRNA-ceRNA, miRNA-ncRNA and protein-RNA interaction networks from large-scale CLIP- Seq data. Nucleic Acids Res 2014;42:D92-7.

[31] Ru B, Wong CN, Tong Y, Zhong JY, Zhong SSW, et al. TISIDB: an integrated repository portal for tumor-immune system interactions. Bioinformatics 2019;35:4200-2.

[32] Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550.

[33] Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012; 16:284-7.

[34] Mayakonda A, Lin DC, Assenov Y, Plass C, Koeffler HP. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res 2018;28:1747-56.

[35] Hu W, Wang G, Chen Y, Yarmus LB, Liu B, et al. Coupled immune stratification and identification of therapeutic candidates in patients with lung adenocarcinoma. Aging (Albany NY) 2020;12:16514-38.

[36] Hanzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinf 2013; 14:7.

[37] Yoshihara K, Shahmoradgoli M, Martinez E, Vegesna R, Kim H, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612.

[38] Bindea G, Mlecnik B, Tosolini M, Kirilovsky A, Waldner M, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 2013;39:782-95.

[39] Yi L, Wu G, Guo L, Zou X, Huang P. Comprehensive analysis of the PD-L1 and immune infiltrates of m(6)A RNA methylation regulators in head and neck squamous cell carcinoma. Mol Ther Nucleic Acids 2020;21:299-314.

[40] Wu M, Wang Y, Liu H, Song J, Ding J. Genomic analysis and clinical implications of immune cell infiltration in gastric cancer. Biosci Rep 2020;40.

[41] Bonneville R, Krook MA, Kautto EA, Miya J, Wing MR, et al. Landscape of microsatellite instability across 39 cancer types. JCO Precis Oncol 2017:2017.

[42] Thorsson V, Gibbs DL, Brown SD, Wolf D, Bortone DS, et al. The immune landscape of cancer. Immunity 2018;48. 812-30 e14.

[43] Jiang P, Gu S, Pan D, Fu J, Sahu A, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018;24:1550-8.

[44] Wang Q, Li M, Yang M, Yang Y, Song F, et al. Analysis of immune-related signatures of lung adenocarcinoma identified two distinct subtypes: implications for immune checkpoint blockade therapy. Aging (Albany NY) 2020;12:3312-39.

[45] Zack TI, Schumacher SE, Carter SL, Cherniack AD, Saksena G, et al. Pan-cancer patterns of somatic copy number alteration. Nat Genet 2013;45:1134-40.

[46] Nishiyama A, Nakanishi M. Navigating the DNA methylation landscape of cancer. Trends Genet 2021;37:1012-27.

[47] Dvinge H, Git A, Graf S, Salmon-Divon M, Curtis C, et al. The shaping and functional consequences of the microRNA landscape in breast cancer. Nature 2013;497:378-82.

[48] Cuijpers SAG, Vertegaal ACO. Guiding mitotic progression by crosstalk between post-translational modifications. Trends Biochem Sci 2018;43:251-68.

[49] Mendiratta G, Ke E, Aziz M, Liarakos D, Tong M, et al. Cancer gene mutation frequencies for the U.S. population. Nat Commun 2021;12:5961.

[50] Liu Y, Gu W. p53 in ferroptosis regulation: the new weapon for the old guardian. Cell Death Differ 2022.

[51] Jiang L, Kon N, Li T, Wang SJ, Su T, et al. Ferroptosis as a p53-mediated activity during tumour suppression. Nature 2015;520:57-62.

[52] Wang SJ, Li D, Ou Y, Jiang L, Chen Y, et al. Acetylation is crucial for p53- mediated ferroptosis and tumor suppression. Cell Rep 2016;17:366-73.

[53] Jennis M, Kung CP, Basu S, Budina-Kolomets A, Leu JI, et al. An African-specific polymorphism in the TP53 gene impairs p53 tumor suppressor function in a mouse model. Genes Dev 2016;30:918-30.

[54] Tarangelo A, Magtanong L, Bieging-Rolett KT, Li Y, Ye J, et al. p53 Suppresses Metabolic Stress-Induced Ferroptosis in Cancer Cells. Cell Rep 2018;22:569-75.

[55] Xie Y, Zhu S, Song X, Sun X, Fan Y, et al. The Tumor Suppressor p53 Limits Ferroptosis by Blocking DPP4 Activity. Cell Rep 2017;20:1692-704.

[56] Zhou N, Bao J. FerrDb: a manually curated resource for regulators and markers of ferroptosis and ferroptosis-disease associations. Database (Oxford) 2020;2020.

[57] Liu Z, Zhao Q, Zuo ZX, Yuan SQ, Yu K, et al. Systematic analysis of the aberrances and functional implications of ferroptosis in cancer. iScience 2020; 23:101302.

[58] Tang B, Yan R, Zhu J, Cheng S, Kong C, et al. Integrative analysis of the molecular mechanisms, immunological features and immunotherapy response of ferroptosis regulators across 33 cancer types. Int J Biol Sci 2022;18:180-98.

[59] Bai D, Feng H, Yang J, Yin A, Lin X, et al. Genomic analysis uncovers prognostic and immunogenic characteristics of ferroptosis for clear cell renal cell carcinoma. Mol Ther Nucleic Acids 2021;25:186-97.

[60] Mhaidly R, Mechta-Grigoriou F. Role of cancer-associated fibroblast subpopulations in immune infiltration, as a new means of treatment in cancer. Immunol Rev 2021;302:259-72.

[61] Charoentong P, Finotello F, Angelova M, Mayer C, Efremova M, et al. Pan- cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade. Cell Rep 2017;18:248-62.

[62] Lei G, Zhuang L, Gan B. Targeting ferroptosis as a vulnerability in cancer. Nat Rev Cancer 2022.

[63] Chen X, Kang R, Kroemer G, Tang D. Broadening horizons: the role of ferroptosis in cancer. Nat Rev Clin Oncol 2021;18:280-96.

[64] Zhu H, Liu J, Feng J, Zhang Q, Bian T, et al. Overexpression of TPX2 predicts poor clinical outcome and is associated with immune infiltration in hepatic cell cancer. Medicine (Baltimore) 2020;99:e23554.

[65] Bian T, Zheng M, Jiang D, Liu J, Sun H, et al. Prognostic biomarker TUBA1C is correlated to immune cell infiltration in the tumor microenvironment of lung adenocarcinoma. Cancer Cell Int 2021;21:144.

[66] Zhu H, Hu X, Gu L, Jian Z, Li L, et al. TUBA1C is a prognostic marker in low-grade glioma and correlates with immune cell infiltration in the tumor microenvironment. Front Genet 2021;12:759953.

[67] Cejalvo T, Gargini R, Segura-Collar B, Mata-Martinez P, Herranz B, et al. Immune profiling of gliomas reveals a connection with IDH1/2 mutations, tau function and the vascular phenotype. Cancers (Basel) 2020;12.

[68] Shen H, Wang Z, Ren S, Wang W, Duan L, et al. Prognostic biomarker MITD1 and its correlation with immune infiltrates in hepatocellular carcinoma (HCC). Int Immunopharmacol 2020;81:106222.

[69] Chen C, Sheng Y. Prognostic impact of MITD1 and associates with immune infiltration in kidney renal clear cell carcinoma. Technol Cancer Res Treat 2021; 20:15330338211036233.

[70] Ding J, He X, Wang J, Cao G, Chen S, et al. Integrative analysis of prognostic value and immune infiltration of spindle and kinetochore-associated family members in breast cancer. Bioengineered 2021;12:10905-23.

[71] Yu DC, Chen XY, Li X, Zhou HY, Yu DQ, et al. Transcript levels of spindle and kinetochore-associated complex 1/3 as prognostic biomarkers correlated with immune infiltrates in hepatocellular carcinoma. Sci Rep 2021;11:11165.

** * * ** * ៛ *
:* ** * * ** * ** * ** * ** * ** * * ** * * **	* p < 0.05 ** p < 0.01 Correlation 1.0 0.5 0.0 -0.5 -1.0
A * K * + * * ** * * . ** * * * *	* p < 0.05 ** p < 0.01 Correlation 1.0 0.5 0.0 -0.5 -1.0

Quartz 4

Explorer

35832625

Identifies microtubule-binding protein CSPP1 as a novel cancer biomarker associated with ferroptosis and tumor microenvironment

ARTICLE INFO

ABSTRACT

1. Introduction

2. Material and methods

2.1. Data collection

2.2. Differential expression analysis

2.3. Diagnostic analysis

2.4. Genetic alterations and DNA methylation analysis

2.5. Correlations between CSPP1 and associated miRNAs

2.6. Correlations between CSPP1 and clinical features

2.7. Prognostic analysis of CSPP1 and associated miRNAs, model construction, and evaluation

2.8. Functional enrichment analysis in LGG

2.9. Gene mutation and ferroptosis correlation analysis

2.10. TME analysis and immune checkpoint blockade therapy prediction

3. Results

3.1. Aberrant expression of CSPP1 serves as a diagnostic biomarker among cancers

3.2. Multi-dimensional mechanisms involving genetic alterations, DNA methylation, and miRNAs underly CSPP1 dysregulation

3.3. Phosphorylation of CSPP1 at specific sites may play a role in tumorigenesis, especially at Ser424

3.4. CSPP1 correlates with clinical features and outcomes in multiple cancers

3.6. CSPP1 dysregulates ferroptosis in LGG and other cancer types

3.7. CSPP1-associated tumors are infiltrated in different TMEs, improving ICB therapeutic efficacy in specific cancers

4. Discussion

5. Conclusion

Author contributions

Funding information

Declaration of Competing Interest

Acknowledgements

Appendix A. Supplementary data

References

Graph View

Table of Contents

Backlinks