2.1 Data collection

This study involved the analysis of multiple datasets retrieved from various sources. We utilized the GSE158494 dataset, containing microarray data derived from cabazitaxel-sensitive and -resistant DU145 and PC-3 CRPC cells. Genes showing differential expression between cabazitaxel-sensitive and -resistant PC-3 and DU145 cells were considered hub genes crucial for the occurrence of cabazitaxel resistance. Additionally, two more datasets, GSE33455 and GSE36135, were acquired, both providing microarray data of docetaxel-resistant CRPC cells. These datasets were sourced from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). Furthermore, we analyzed the GSE35988 dataset, encompassing information on patients diagnosed with PCa and CRPC. In addition to these datasets, patient data from two distinct databases were examined: The Cancer Genome Atlas (TCGA; http://cancergenome.nih.gov/) and Chinese Prostate Cancer Genome and Epigenome Atlas (CPGEA; http://www.cpgea.com).

2.2 Data handling

Raw microarray data retrieved from various databases underwent normalization using the “limma” package in the R software. During this process, the probe ID was replaced with the corresponding gene ID. Genes meeting the criteria of |log2FC > 1| and P < 0.05 were classified as cabazitaxel resistance-related genes. Additionally, clinical data were obtained from diverse databases.

2.3 Pathway analysis

Pathway analysis was conducted through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The online tool Metascape, accessible at http://metascape.org/, was used for this analysis, and a bubble map was generated using the “ggplot2” R package.

2.4 Survival analysis

The online web tool “GEPIA” facilitated the exploration of the correlation between cabazitaxel resistance-related genes and the disease-free survival (DFS) status in PCa patients. Univariate and multivariate Cox regression analyses were performed to determine the variables necessary to develop the nomogram. According to the median gene expression of each gene in PCa patients from TCGA database, patients were divided into high and low gene expression groups. Additionally, the “forest plot” package in the R software was employed to display P-values, hazard ratios (HRs), and 95% confidence intervals (CIs) for each variable. Subsequently, a nomogram was constructed based on the outcomes of the multivariate Cox proportional hazards analysis, serving as a predictive tool for overall recurrence.

2.5 Immune infiltration analysis

Gene Set Cancer Analysis (GSCA; http://bioinfo.life.hust.edu.cn/GSCA/#/) is a public online web tool used to download the immune cell infiltration information of PCa patients from TCGA database. We first downloaded this immune infiltration information from GSCA. Using data from TCGA database, Spearman correlation analysis was conducted to evaluate the relationship between the expression of cabazitaxel resistance-related genes and immune cell infiltration (Additional file 1).

2.6 Clinical specimen collection

Specimens were obtained from individuals diagnosed with PCa and CRPC at Tongji Hospital, School of Medicine, Tongji University. The collection procedure was approved by the Ethics Committee of Tongji Hospital, School of Medicine, Tongji University (SBKT-2021-220). Comprehensive information about the study was provided to all participating patients, and they provided informed consent for the utilization of their samples in the research.

2.7 Cell culture and drug treatment

PCa cell lines were obtained from the Chinese Academy of Science Cell Bank (Shanghai, China). The human CRPC cell lines PC-3 and DU145 were cultured in Roswell Park Memorial Institute (RPMI) 1640 media (Sigma, Darmstadt, Germany; Catalog No. R8758) containing 10% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA, Catalog No. 10091). Cell cultures were maintained at 37 ℃ in a humidified atmosphere with 5% CO2 and 95% air. Docetaxel and cabazitaxel were acquired for the study (SelleckChem, Houston, TX, USA, Catalog No. S1148 and S3022, respectively). The CRPC cells were exposed to docetaxel or cabazitaxel at a dose of 2 nmol/L for 24 h. Furthermore, both PC-3 and DU145 cells were treated with docetaxel (2 nmol/L) for 2 weeks to induce resistance. Upon observing that docetaxel did not affect cell growth, we concluded that these cells were resistant to docetaxel (Additional file 2).

2.8 Cell transfection and lentivirus production

Cell transfection experiments were conducted using Lipofectamine 2000 (Thermo Fisher Scientific, Catalog No. 11668019). Specifically, shGOLGA8B lentivirus was designed to facilitate specific gene knockdown. Gene-specific shRNAs and a control lentivirus (shControl) were procured from Youze Biotechnology Company (Hunan, China).

2.9 RNA extraction and qRT-PCR

Total RNA was isolated from clinical samples and CRPC cell lines using the TRIzol reagent (Sigma-Aldrich, St. Louis, MO, USA, Catalog No. T9424). Subsequently, cDNA synthesis was performed using a reverse transcription kit (Advantage® RT-for-PCR Kit, Takara Bio Inc., Kusatsu, Japan, Catalog No. 639505). Gene expression analysis was conducted using a real-time PCR kit (TB Green® Premix Ex Taq™ II, Takara Bio Inc., Catalog No. RR420A) following the manufacturer’s instructions. The PCR primer sequences for GOLGA8B and GAPDH (reference gene) are listed in Table 1. mRNA expression was quantified using the 2−ΔΔCt technique.

Table 1 The primers used in the qRT-PCR2.10 Antibodies

Rabbit polyclonal anti-GOLGA8B antibody (Catalog No. ab155806) and rabbit monoclonal anti-GAPDH antibody (Catalog No. ab9485) were purchased from Abcam UK (Cambridge, UK).

2.11 Western blotting

Proteins were extracted from tissues and cell lines using RIPA lysis buffer. Following extraction, protein samples were processed with treated with Dual Color Protein Loading Buffer (Thermo Fisher Scientific). Protein separation was achieved through SDS-PAGE using a 10% gel, followed by protein transfer onto nitrocellulose membranes sourced from Merck KGaA (Darmstadt, Germany). The Protein-Free Rapid Blocking Buffer (Thermo Fisher Scientific) was used to block the membranes. The membranes were exposed to primary antibodies against GOLGA8B (dilution 1:1000) and GAPDH (dilution 1:1000; Abcam UK) overnight at 4 ℃. The following day, the membranes were washed thrice using 1 × TBST (10 min/cycle) and then incubated at room temperature for 1.5 h with the corresponding secondary antibody [Catalog No. A0208, HRP-labeled Goat Anti-Rabbit IgG (H + L), obtained from Beyotime Biotechnology, Shanghai, China]. Finally, after exposing the membranes to X-ray film, protein identification was conducted.

2.12 Cell proliferation assay

The Cell Counting Kit 8 (CCK 8; Dojindo, Japan) was used to assess the cell proliferation capacity. Briefly, cells were placed in 96-well plates at a density of 3000 cells/well. They were then cultivated for specific lengths of time (0, 24, 48, or 72 h) in 200 µL of RPMI 1640 supplemented with 10% FBS. Subsequently, cell viability was determined using the CCK8 assay following the provided protocol. A spectrophotometer (LD942, Beijing, China) was used to measure the absorbance at 450 nm.

2.13 Statistical analysis

R v4.0.3 (Institute for Statistics and Mathematics, Vienna, Austria; https://www.r-project.org) was used to analyze the matrix data. The Wilcoxon test was applied for comparisons between two groups, and the Kruskal–Wallis test was utilized for comparisons involving more than two groups. Statistical metrics included P values, 95% CIs, and HRs. Statistical significance was set at P < 0.05 (two-tailed).