Highlights
-
•
Recurrence/metastasis-associated proteins modulate the effects of DOX/DDP in OS.
-
•
DIA proteomics revealed signaling pathways contributing to DOX/DDP efficacy.
-
•
High CTSG expression suggests poor outcomes after DOX/DDP treatment in OS patients.
-
•
NET formation plays a pivotal role in the low efficacy of DOX/DDP treatment in OS.
Keywords: Osteosarcoma, Doxorubicin/cisplatin chemotherapy, CTSG, Neutrophil extracellular traps (NETs), 4D-DIA proteomics, Parallel reaction monitoring
Abstract
Background
Doxorubicin and cisplatin are both first-line chemotherapeutics for osteosarcoma (OS) treatment. However, the efficacy of doxorubicin/cisplatin chemotherapy varies considerably. Thus, identifying an efficient diagnostic biomarker to distinguish patients with good and poor responses to doxorubicin/cisplatin chemotherapy is of paramount importance.
Methods
To predict the efficacy of doxorubicin/cisplatin chemotherapy, we analyzed the differentially expressed proteins in 37 resected OS samples, which were categorized into the primary group (PG), the recurrent group (RG) and the metastatic group (MG). The characteristics of the enriched differentially expressed proteins were assessed via GO and KEGG analyses. Protein‒protein interactions were identified to determine the relationships among the differentially expressed proteins. Receiver operating characteristic (ROC) curve analyses were performed to explore the clinical significance of the differentially expressed proteins. Parallel reaction monitoring (PRM) was used to validate the candidate proteins. Immunohistochemical (IHC) staining was performed to confirm the expression of cathepsin (CTSG) in patients with good and poor response to doxorubicin/cisplatin.
Results
A total of 9458 proteins were identified and quantified, among which 143 and 208 exhibited significant changes (|log2FC|>1, p < 0.05) in the RG and MG compared with the PG, respectively. GO and KEGG enrichment led to the identification of neutrophil extracellular traps (NETs). ROC curve analyses revealed 74 and 86 proteins with areas under the curve greater than 0.7 in the RG and MG, respectively. PRM validation revealed the statistical significance of CTSG, which is involved in NET formation, at the protein level in both the RG and MG. IHC staining of another cohort revealed that CTSG was prominently upregulated in the poor response group after treatment with doxorubicin/cisplatin.
Conclusion
CTSG and its associated NETs are potential biomarkers with which the efficacy of doxorubicin/cisplatin chemotherapy could be predicted in OS patients.
Background
Osteosarcoma (OS) accounts for approximately 44.6 % of malignant bone tumors and is the most common primary malignant bone tumor in the clinic. OS is commonly diagnosed in children and adolescents, progresses rapidly, and seriously endangers the health of adolescents[1,2]. OS is relatively insensitive to radiation therapy and immunotherapy, and surgical treatment alone has a low survival rate[3]. Chemotherapy, among which doxorubicin/cisplatin is the most commonly used regimen, is commonly used to improve the long-term survival rate of OS patients. However, the efficacy of this treatment varies considerably among patients, and some patients do not benefit from doxorubicin/cisplatin treatment. Chemotherapy resistance is a problem in the clinical treatment of OS and biomarkers for predicting the efficacy of doxorubicin/cisplatin chemotherapy have yet to be identified.
Previous investigations to identify potential biomarkers to facilitate the prediction of the efficacy of chemotherapy for OS have focused mostly on one specific chemotherapeutic, such as doxorubicin, cisplatin, methotrexate (Methotrexate) or ifosfamide, in cell lines, animal models or xenografts[[4], [5], [6], [7], [8], [9], [10]]. Microarray gene expression profiles from different human osteosarcoma xenografts that exhibited different sensitivities to doxorubicin, cisplatin, and ifosfamide were obtained, suggesting that some tumor markers can be used to identify patients who do or do not respond to the above chemotherapeutics but have not been validated in clinical OS specimens[8]. In addition, several potential noncoding RNAs have been found to contribute to the resistance of OS cells to doxorubicin and cisplatin [6,7,9,10]. In contrast, some studies have used open biopsy samples with oligonucleotide microarrays to identify biomarkers of chemotherapy response in OS patients, which has been shown to have greater value [11]. A study in Japan used a proteomics technique to identify potentially valuable proteins in OS patients. However, two-dimensional difference gel electrophoresis (2D-DIGE) was used, and the patient sample size was small[12].
With the rapid development of quantitative proteomic analysis techniques, the discovery and validation of biomarkers have accelerated. Recently, emergent 4-D data-independent acquisition (DIA) mass spectrometry has become an attractive quantitative proteomics method, which can achieve higher throughput, identification depth, qualitative accuracy, and stability[[13], [14], [15]]. Parallel reaction monitoring (PRM) represents a targeted peptide/protein quantification method with high mass accuracy and high resolution[[16], [17], [18]]. The combination of untargeted 4D-DIA and targeted PRM has demonstrated great applicability in the identification and validation of predictive molecular biomarkers for human diseases. In this study, we aimed to identify potential biomarkers that could be used to predict the efficacy of doxorubicin/cisplatin treatment at the protein level using primary, recurrent, and metastatic OS samples.
Methods
Clinical specimens and grouping
Paraffin-embedded OS tissues were collected from 37 OS patients who underwent doxorubicin/cisplatin chemotherapy and tumor resection at Tangdu Hospital, Air Force Medical University, between January 2020 and December 2022. Patients were divided into 3 groups. Twenty patients were categorized into the good response group, which is defined as the primary group (PG) without recurrence or metastasis; fourteen patients were categorized into the recurrent group (RG); and three patients in which their OS metastasized to the lung were categorized into the metastatic group (MG). All patients or their relatives provided signed informed consent, and the authors had no access to the patients’ identities. Studies concerning clinical specimens were performed in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Air Force Medical University.
4D-DIA proteomics analysis
Sample preparation
Paraffin-embedded OS tissues were treated with heptane, vortexed for 10 s, and incubated for 1 h at room temperature. Methanol was then added to the tissues, which were then centrifuged at 9000 × g for 2 min. The supernatant was discarded. Then, 100 μL of SDT buffer was added, and the lysate was incubated on ice for 5 min, boiled for 20 min, and sonicated. The supernatant was centrifuged at 12,000 × g for 15 min and filtered through 0.22 µm spin filters. The filtrate was quantified using the BCA Protein Assay Kit (Pierce, USA). The samples were stored at −80 °C until use. A total of 20 µg of protein from each sample was mixed with 6 × loading buffer and boiled for 5 min. The proteins were separated on a 12 % SDS‒PAGE gel. Protein bands were visualized with Coomassie brilliant blue R-250 staining.
For filter-aided sample preparation (FASP), 200 µg of protein from each sample was mixed with 100 mM DTT for 5 min at 100 °C. DTT and other low-molecular-weight components were subsequently removed. Iodoacetamide was added, and the samples were incubated for 30 min in darkness. The filters were washed with UA buffer three times and with NH4HCO3 buffer twice. The protein suspensions were digested with trypsin (Promega, USA) in 40 μL of 50 mM NH4HCO3 buffer overnight at 37 °C, after which the peptides were collected. The peptide content was estimated by the UV spectroscopy at 280 nm.
Peptides were fractionated by reverse-phase chromatography using a 1260 infinity II HPLC (Agilent, USA). The peptide mixture was diluted with buffer A (10 mM HCOONH4, 5 % ACN, pH 10.0) and loaded onto an XBridge Peptide BEH C18 column. The peptides were eluted at a flow rate of 1 mL/min with a gradient of 0 %–7 % buffer B (10 mM HCOONH4, 85 % ACN, pH 10.0) for 30 min, 7–40 % buffer B for 30–58 min, 40 %–100 % buffer B for 58–70 min, and 100 % buffer B for 70–85 min. The elution was monitored at 214 nm, and the fractions were collected every 1 min for 5–50 min. The collected fractions were dried via vacuum centrifugation at 45 °C.
4D-DIA MS analysis
Peptide samples (1000 ng) mixed with iRT were analyzed on an Evosep One system (Evosep, Denmark) coupled to an Orbitrap Exploris 480 (Thermo Fisher Scientific, USA) instrument equipped with a CaptiveSpray source. Peptides were separated on an analytical column (Evosep, Denmark) maintained at 50 °C using an integrated column oven (Bruker, Germany). The LC separation method was provided by Evosep One.
The mass spectrometry parameters were as follows. MS: scan range (m/z) =350–1500, resolution=60,000, automatic gain control (AGC) target=300 %, maximum injection time=20 ms, charge state=2–7, and exclusion duration of filter dynamic exclusion=45 s; data-dependent MS2: isolation window=1.6 m/z, resolution=15,000, AGC target=75 %, maximum injection time=22 ms, and normalized collision energy (NCE) =30 %.
The raw DIA data were processed and analyzed by Spectronaut (Biognosys AG, Switzerland). The retention time prediction type was set to dynamic iRT. Spectronauts can dynamically determine the ideal extraction window depending on the iRT calibration and gradient stability. The cutoff for the false discovery rate of the peptides was set to 1 %, and the differentially expressed proteins were filtered according to the following criteria: p value<0.05 and |log2FC|>1.
GO, KEGG and GSEA analyses
All protein sequences were aligned to the human database sequences using NCBI BLAST+ (ncbi-blast-2.3.0+), and sequences with an E value≤1e-3 were retained. Gene Ontology (GO) terms associated with the aligned sequence with the highest bit score were selected by the Blast2GO Command Line. After annotation, InterProScan (interproscan-5.30–69.0) was used to search the EBI database for conserved motifs that matched the target protein, and motif-related information was annotated to the target proteins. To improve the accuracy, we supplemented the annotated information and made connections between different GO terms via the ANNEX module.
Orthologous genes with similar functions in the same pathway were grouped together and assigned the same Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthology (KO) label. The KEGG Orthology And Links Annotation (KOALA) software V3.0 [19] was used to align the KEGG GENES database (version: KO_INFO_END.txt (2023.10.17)) with the target protein sequences. These protein sequences were then classified by KO labels, and pathway information was obtained.
Gene set enrichment analysis (GSEA) was performed for all identified genes GSEA software. Human datasets were obtained from the Msigdb database.
PPI network analysis
The STRING database (https://www.string-db.org/) was utilized to investigate the interactions among the differentially expressed proteins. Degrees of connectivity for each protein were calculated to assess the importance of different proteins in the network analysis, which could help to identify probable target proteins involved in different efficacies of chemotherapy treatment for OS.
Receiver operating characteristic (ROC) curve analysis
For the expression matrices of all proteins, empty values were filled using the knn method (Python 3.10) based on the scikit-learn machine learning framework to obtain the complete protein abundance matrix. The abundance values were obtained for proteins based on differential expression analysis, and the area under the curve (AUC) values of each protein were calculated based on the R package pROC (R4.3.1). For proteins with AUC values ≥0.7, the logistic equation was calculated with the R package mlr3verse (R4.3.1) machine learning framework, and AUC values were calculated using the predicted results of the equation. The ROC curves were then plotted and statistically analyzed.
Parallel reaction monitoring (PRM) validation
PRM validation was performed to verify the results of the 4D-DIA MS analysis. Each eluent was injected for nanoLC‒MS/MS analysis twice. The peptide mixture was loaded onto a C18 reversed-phase analytical column (Thermo Fisher Scientific, USA) in buffer A (0.1 % formic acid) and separated with a linear gradient of buffer B (80 % acetonitrile and 0.1 % formic acid) at a flow rate of 300 nL/min. The 60 min liquid gradient was as follows: 0–5 min, 1∼3 % buffer B; 6–45 min, buffer B from 3 % to 28 %; 46–50 min, buffer B from 28 % to 38 %; buffer B from 38 % to 100 % at 51–55 min; and 100 % at 56–60 min.
The targeted PRM LC‒MS/MS analysis was performed on a Q Exactive HF‒X mass spectrometer (Thermo Fisher Scientific, USA) coupled to an Easy nLC (Thermo Fisher Scientific, USA) for 60 min. The mass spectrometer was operated in positive ion mode. MS data were acquired using a data-dependent top10 method to dynamically choose the most abundant precursor ions from the survey scan (350–1800 m/z) for higher energy collisional dissociation (HCD) fragmentation. Survey scans were acquired at a resolution of 60,000 at m/z 200 with an automatic gain control (AGC) target of 3e6 and a maximum inject time (IT) of 50 ms. MS2 scans were acquired at a resolution of 3000 for HCD spectra at m/z 200 with an AGC target of 2e5 and a maximum IT of 50 ms, and the isolation width was 1.6 m/z. Only ions with a charge state between 2 and 6 and a minimum intensity of 8e3 were selected for fragmentation. The dynamic exclusion for selected ions was 30 s. The NCE was 27 eV. SpectroDive software and the UniProt database (UniProt_homo_20,230,312_20,423_ 9606_ swiss_prot) were used to analyze the data off-line.
Western blotting
Western blotting was performed via standard procedures. Total protein extracts (20 μg) were separated by 10 % SDS‒PAGE and transferred to a PVDF membrane. To block nonspecific binding, the membranes were incubated with 5 % nonfat milk powder at room temperature for 2 h. The membranes were then incubated with a primary antibody against CTSG overnight (1:1000, Abcam) and subsequently incubated with a horseradish peroxidase-labeled secondary antibody (Proteintech) for 1 h. The signal intensity of each band was detected by chemiluminescence and quantified with Image-Pro Plus. An anti-β-tubulin antibody (1:20,000, Abcam) was used as a control.
Hematoxylin-eosin and immunohistochemical staining
A total of 40 additional OS specimens that underwent doxorubicin/cisplatin chemotherapy were collected at Tangdu Hospital, Air Force Medical University, between Jan 2010 and Dec 2019. The specimens were divided into two groups (sensitive and resistant) based on the clinicopathological data. In total, 20 specimens were categorized into the sensitive group, and the other 20 were categorized into the resistant group. Formalin-fixed paraffin-embedded tissue specimens were cut into 3 μm slices (Leica, Germany) and heated at 60 °C for 2 h. The following procedures were conducted according to the manufacturer's instructions. An H&E staining kit (Abcam, Cat# ab245880) was used to stain the tumor cells. The anti- cathepsin G (CTSG) (Abcam, Cat# ab282105) antibody was used at a dilution of 1:1000. DAB staining (GK6007, Gene Tech) was performed to visualize the proteins. Two different pathologists independently scored the expression of proteins based on both the number of positive cells and the intensity of staining.
Statistical analysis
Python (version 3.9.5) and R software (version 3.6.0) were used for bioinformatics analyses and visualization. The experiments were independently repeated at least three times. 4D-DIA MS data were normalized using MaxLFQ [20]. Statistical analyses were performed with SPASS 26.0 and GraphPad Prism 9 software. The data are presented as the mean ± standard deviation. Student's t-test was used to compare the data between two groups, and a p value < 0.05 was considered to indicate statistical significance.
Results
Quantitative proteomics analysis of OS specimens with different responses to doxorubicin/cisplatin chemotherapy
We collected the data of 37 OS patients who received doxorubicin/cisplatin chemotherapy (see Additional file 1: Table S1). The patients were divided into 3 groups according to chemotherapy efficacy. The primary tumors without recurrence or metastases group (PG) included 20 patients, the recurrence in situ group (RG) included 14 patients, and the distant metastases group (MG) included 3 patients. 4-D DIA-based quantitative MS was performed to investigate differences in protein expression and abundance.
A total of 97,953 peptides and 9458 proteins (with at least one specific peptide segment) were identified and quantified and the false discovery rate was lower than 1 %. The molecular weights of the proteins ranged from 1.39 to 3816.03 kDa. A total of 143 proteins (112 upregulated; 31 downregulated) were differentially expressed between the RG and PG (Fig. 1a, c, see Additional file 2a: Table S2a), and 208 proteins (140 upregulated; 68 downregulated) were differentially expressed between the MG and PG (Fig. 1b, d, see Additional file 2b: Table S2b). The above proteins were significantly up- or downregulated by >2-fold between the different groups (p < 0.05). According to the Venn analysis, 28 proteins were common among the proteins differentially expressed between the PG group and the other two groups (Fig. 1e).
Fig. 1.
Differentially expressed proteins in the RG, MG, and PG of OS patients. a Volcano plot of differentially expressed proteins between the RG and PG. b Volcano plot of differentially expressed proteins between the MG and PG. c Heatmap of differentially expressed proteins between the RG and PG (|log2FC|>1, p < 0.05). d Heatmap of differentially expressed proteins between the MG and PG (|log2FC|>1, p < 0.05). e Venn diagram showing the proteins common and specific to the groups.
Functional classification and signaling pathway analysis of differentially expressed proteins
GO and KEGG analyses were subsequently performed to determine differences in tumor response to doxorubicin/cisplatin. The significantly enriched GO terms were mostly associated with the extracellular matrix (GO: 0030198, GO: 0005201, GO: 0030023, GO: 0062023, GO: 0030021, GO: 0050840) and cytoskeleton (GO: 0007016, GO: 0005200, GO: 0014731, GO: 0005884, GO: 0030863) in the RG and MG compared with the PG (Fig. 2a, b, see Additional file 3: Table S3). KEGG analyses revealed that extracellular matrix (ECM)-receptor interaction, cytoskeleton regulation, inflammatory reactions, and neutrophil extracellular trap (NET) formation were more common in the RG and MG than in the PG (Fig. 2c, d, see Additional file 3:Table S3).
Fig. 2.
Analyses of GO, KEGG, and GSEA enrichment between different groups of OS. a, b Significantly enriched GO terms for the biological process, molecular function, and cellular component categories between the recurrent group (RG) and primary group (PG)(a) and between the metastatic group (MG) and PG (b). c, d, Enriched KEGG pathways between the RG and PG (c) and between the MG and PG (d). The horizontal axis demonstrates the significance of pathways in the form of –log10 (p value). The numbers beside each bar represent the enrichment factors of the pathways. e, f GSEA between the RG and PG (e) and between the MG and PG (f).
The GSEA results showed that the ECM receptor interaction, complement and coagulation cascades, regulation of the actin cytoskeleton, arachidonic acid metabolism, drug metabolism cytochrome P450, the MTOR signaling pathway, and ABC transporters were enriched in the KEGG terms of the RG compared with those of the PG (Fig. 2e, see Additional file 4: Fig. S1a). Compared with those in the PG, cell adhesion molecules, cytokine‒cytokine receptor interactions, complement and coagulation cascades, ABC transporters, ECM receptor interactions and arachidonic acid metabolism were enriched in the MG (Fig. 2f, see Additional file 4: Fig. S1b). In addition, the differentially abundant proteins identified in both RG vs. PG and MG vs. PG comparisons were mostly involved in ECM receptor interactions, complement and coagulation cascades, and drug metabolism. These results suggested that drug metabolism and inflammatory responses may play a role in the poor response of OS to doxorubicin/cisplatin chemotherapy.
Protein‒protein interaction (PPI) network analysis
PPI network analysis was performed to identify the relationships among the differentially expressed proteins between the RG and PG and between the MG and PG using the online database STRING (https://string-db.org/). The highly enriched connected proteins between the RG and PG included PIP5K1C (UniProtKB-O60331), PLCH1 (UniProtKB-Q4KWH8), PIP4K2C (UniProtKB-Q8TBX8), and CPA3 (UniProtKB-P15088), which interact with two significantly differentially expressed proteins (Fig. 3a, see Additional file 5: Table S4). The highly connected proteins between the MG and the PG included EPB42 (UniProtKB-P16452), SLC4A1 (UniProtKB-P02730), SPTA1 (UniProtKB-P02549), ITGA2B (UniProtKB-P08514), EPB41 (UniProtKB-P11171), and ANK1 (UniProtKB-P16157), which interact with ≥5 differentially expressed proteins (Fig. 3b, see Additional file 5: Table S4). Interestingly, the above highly connected proteins were enriched in actin cytoskeleton dynamics, cytoskeletal network, extracellular space, and neutrophil chemotaxis, which was consistent with the GO and KEGG enrichment results.
Fig. 3.
Relationships of differentially expressed proteins between the RG/MG and PG in OS. a between the RG and PG; b between the MG and PG. Protein–protein interaction (PPI) network analysis was performed using the STRING database. The purple line represents interactions that were experimentally confirmed; the blue line represents interactions from the curated database; the light green line represents text mining interactions; the dark green line illustrates the gene neighborhood; the red line represents gene fusions; and the light violet line represents protein homology.
Analysis of the discriminative ability of candidate proteins by AUC
To further explore the clinical significance of the differentially expressed proteins, ROC curve analyses were performed, and the predictive accuracy of the 100 proteins with the highest AUC-values that were also enriched in GO and KEGG terms and identified in the highly connected nodes in the PPI analysis were assessed in the RG (Fig. 4a) and MG (Fig. 4b) compared with the PG. The AUC values of the assessed proteins ranged from 0.557 to 0.9, with 11 proteins having an AUC value above 0.7. The protein with the highest AUC value was PI51C (UniProtKB-O60331, 0.838, encoded by the PIP5K1C gene), followed by PLCH1 (UniProtKB-Q4KWH8, 0.786, encoded by the PLCH1 gene), CATG (UniProtKB-P08311, 0.779, encoded by the CTSG gene), and MASP1 (UniProtKB-P48740, 0.764, encoded by the MASP1 gene) in the comparison between RG and PG. In the comparison between MG and PG, the four proteins with high AUC values were PLCH1 (0.9), MASP1 (0.883), CATG (0.867), and SPTA1 (UniProtKB-P02549, 0.867). Compared with the results observed in the, PLCH1 was downregulated, while the other proteins were upregulated in the RG and MG comparison. Interestingly, PLCH1, CTSG, and MASP1 had high AUC values in both groups.
Fig. 4.
Receiver operating characteristic (ROC) curves for the selected differentially expressed proteins. a between the RG and PG; b between the MG and PG.
PRM and western blotting validation of 4D-DIA quantitative mass spectrometry
A total of 24 differentially expressed proteins that were enriched in the common KEGG terms in the RG and MG in comparison to the PG or showed high AUC values both in the RG and MG were selected for further verification using PRM. The proteins included ITGA2B, CTSG, ITGB3, COL6A3, C3, SPTA1, ANK1, FGA, FGB, SLC4A1, LAMC1, FGG, EPB41L2, LAMA4, SERPINC1, EPB42, C4BPA, CPA3, ITGA5, ITGA11, MGST3, GP1BA, PLCH1, and ITGA4. The results of quantitative validation of these proteins in different groups correlated well with 4D-DIA results. Eighteen proteins exhibited trends similar to those observed in the 4D-DIA results. Interestingly, according to the DIA and PRM analyses, CTSG was significantly upregulated in both the RG and MG (Fig. 5c). However, according to the 4D-DIA and PRM analyses, the levels of other proteins, such as ITGA2B, FGA, and SLC4A1, did not significantly differ between the RG and MG (Fig. 5a-b, d-f, see Additional file 6: Table S5). Western blotting was subsequently performed, and higher expression of CTSG in the RG and MG than in the PG was detected (see Additional file 7: Fig.S2).
Fig. 5.
Expression of six candidate proteins by 4D-DIA MS and PRM validation. a ITGA2B, b SLC4A1, c CTSG, d FGA, e FGB, f FGG. Each group contained at least three samples. Student's t-test. *p < 0.05; **p < 0.01; ***p < 0.001. ns: not significant.
Immunohistochemical staining of clinical specimens
A total of forty OS specimens, including 20 doxorubicin/cisplatin chemotherapy-sensitive and 20 doxorubicin/cisplatin chemotherapy-resistant patient specimens, were further examined to confirm the 4D-DIA and PRM results. The clinical data of the patients are recorded in see Additional file 8 (Table S6). H&E staining of each specimen revealed tumor cells in the OS tissues (Fig. 6a, b). IHC staining demonstrated that CTSG expression was significantly greater in the chemotherapy-resistant specimens than in the chemotherapy-sensitive specimens (Fig. 6c-e).
Fig. 6.
Expression of CTSG in doxorubicin/cisplatin-sensitive and doxorubicin/cisplatin-resistant OS specimens. a,b Hematoxylin-eosin (H&E) staining of doxorubicin/cisplatin-sensitive (a) and doxorubicin/cisplatin-resistant (b) OS specimens; scale bar=200 μm. c, d Representative immunohistochemical image of OS specimens; scale bar=100 μm (upper) and 50 μm (lower). e Quantification of CTSG expression was achieved via immunohistochemical staining. Student's t-test.
Discussion
DIA mass spectrometry technology was named the most noteworthy technology by Nature Methods in 2015; this technique divides the entire full scan range of mass spectrometry into several windows and selects, fragments, and cyclically detects all ions in each window at high speed to avoid omissions and obtain all fragment information [21]. Based on these characteristics, 4D-DIA has several advantages, such as high data repeatability and traceability, high sensitivity, and increased detection depth, making the prediction of biomarkers more efficient [22]. In our study, 143 proteins were differentially expressed between the RG and the PG, and 208 proteins were differentially expressed between the MG and the PG (|lg2FC|≥1, p < 0.05); 74 and 86 of these proteins had AUC values greater than 0.7 in the RG and MG, respectively (p < 0.05). PRM analysis was performed to validate the identified candidate biomarkers, and 19 proteins showed similar increased or decreased abundance trends, as demonstrated by the 4D-DIA analysis.
Proteins are direct reflections of various life activities and are the ultimate executors of physiological functions within cells. The different expression patterns of proteins in tumor cells and the surrounding microenvironment might be critically important for the efficacy of chemotherapeutic treatments. However, no protein biomarker has been successfully used to predict the efficacy of different chemotherapeutic agents for OS in clinical practice. The emergence of 4D-DIA represents a technological advance in protein identification and quantification due to the advantages mentioned above. The combination of untargeted 4D-DIA and targeted PRM demonstrated great potential for investigating candidate biomarkers for predicting response in cancer patients.
The expression of the candidate protein CTSG was 4.54- and 3.20-fold greater in the RG and MG, respectively, than in the PG (p < 0.05). CTSG is involved in NET formation, which was significantly enriched in the KEGG analyses in both the RG and MG when compared with the PG. The AUC values of CTSG were 0.779 and 0.867 in the RG and MG, respectively, which shows the ability to differentiate the RG and MG from the PG. PRM validation further demonstrated significantly increased levels of CTSG expression in the RG and MG.
CTSG is an immune-related protein that has been reported to be correlated with poor overall survival in patients with gastric cancer and lung cancer [23,24]. In addition, CTSG was identified as an independent biomarker in oral squamous cell carcinoma (OSCC) and head and neck squamous cell carcinoma (HNSCC) [25,26], and CTSG expression in the PANC-1 pancreatic adenocarcinoma cell line promoted tumor metastasis in mice [27]. On the other hand, CTSG suppresses cancer progression in colorectal cancer [28]. In breast cancer, CTSG can induce cell migration and aggregation through insulin-like growth factor-1 signaling [29]. Nevertheless, the role of CTSG in OS, especially in the process of doxorubicin/cisplatin chemotherapy, is unclear. Our proteomics study demonstrated that CTSG is a prominently increased protein in the RG and MG compared with that in the PG, indicating that this protein may predict poor prognosis after doxorubicin/cisplatin treatment in OS patients.
NETs have been at the forefront of current research in cancer biology [[30], [31], [32], [33], [34]]. A report in Nature has shown that NETs can act as chemotactic factors to attract tumor cells and that serum NETs may predict the occurrence of liver metastases in early-stage breast cancer patients[35]. In addition, NETs promote tumor growth, and tumors in PAD4-deficient mice grow more slowly[36]. In our study, NETs formation was significantly greater in both the RG and MG than in the PG, and CTSG is an important protein involved in this process.
IHC staining was performed to further validate the expression of CTSG in an extra forty OS specimens from patients treated with doxorubicin/cisplatin chemotherapy. CTSG staining was significantly stronger in patients who were resistant to doxorubicin/cisplatin than in those who were sensitive to doxorubicin/cisplatin, which further validated the potential function of CTSG as a biomarker of poor response in OS patients receiving doxorubicin/cisplatin treatment.
CTSG might be of pivotal importance for prediction of doxorubicin/cisplatin chemotherapy efficacy and lead to new treatment strategies for OS patients with high CTSG expression. However, there are several limitations of this study. Rigid quality control in specimen collection makes the data and the following analyses more convincing, yet additional experiments using cell lines and animals could be performed in future studies to provide more solid evidence and elucidate the molecular mechanisms and specific signal transduction pathways involved. The associations and proteins described in this manuscript were acquired from high-throughput technologies and subsequent analyses. Only the results of IHC staining support the drawing of firm biological conclusions. Nevertheless, the integration of 4D-DIA proteomics and subsequent validations using OS specimens with different clinical performances in response to chemotherapy represents a substantial advance in the discovery of biomarkers that can be used to predict the efficacy of doxorubicin/cisplatin, facilitating the design of targeted therapies specific to the different OS subtypes [37,38].
Conclusion
This investigation identified differentially expressed proteins between the RG, MG, and PG in paraffin-embedded clinical OS specimens using 4D-DIA analysis and PRM verification. Further investigations elucidated the close relationship between CTSG, NETs formation and the efficacy of doxorubicin/cisplatin chemotherapy in OS patients. Identifying biomarkers and biological processes involved in the efficacy of different chemotherapies may aid in predicting the prognosis of OS patients and may pave the way for the development of new strategies for the clinical treatment of OS patients.
CRediT authorship contribution statement
Qiong Ma: Writing – original draft, Investigation. Jin Sun: Visualization. Qiao Liu: Resources. Jin Fu: Resources. Yanhua Wen: Resources. Fuqin Zhang: Investigation. Yonghong Wu: Resources. Xiaoyu Zhang: Writing – review & editing. Li Gong: Supervision. Wei Zhang: Funding acquisition, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
Acknowledgments
We sincerely thank Dr. Gang Xu for academic and technical support.
Funding
This research was funded by the Department of Science and technology of Shaanxi Province (grant number 2022-PT-47).
Availability of data and materials
The dataset supporting the conclusions of this article is available in the ProteomeXchange Consortium repository, [PXD051097] [39,40].
Ethics approval and consent to participate
This study was conducted according to the Declaration of Helsinki, and approved by the Ethics Committee of the Air Force Medical University (TDLL-202402-03).
Consent for publication
Not applicable.
Footnotes
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.tranon.2024.102098.
Contributor Information
Li Gong, Email: glzwd16@fmmu.edu.cn.
Wei Zhang, Email: zhwlyh@fmmu.edu.cn.
Appendix. Supplementary materials
Additional file 4: Fig. S1 Results of GSEA of KEGG pathways. a Extracellular matrix (ECM) receptor interaction, complement and coagulation cascades, regulation of actin cytoskeleton, arachidonic acid metabolism, drug metabolism cytochrome P450, metabolism of xenobiotics by cytochrome P450, the MTOR signaling pathway, and ABC transporters were enriched according to GSEA between the RG and PG. b ECM receptor interaction, complement and coagulation cascades, arachidonic acid metabolism, cell adhesion molecules, cytokine-cytokine receptor interaction, and ABC transporters were enriched according to GSEA between the MG and PG.
Additional file 7: Fig. S2 Western blotting revealed greater expression of CTSG in the RG and MG than in the PG. β-tubulin was used as a loading contral.
References
- 1.Beird H.C., Bielack S.S., Flanagan A.M., et al. Osteosarcoma. Nat. Rev. Dis. Primers. 2022;8(1):77. doi: 10.1038/s41572-022-00409-y. [DOI] [PubMed] [Google Scholar]
- 2.Eaton B.R., Schwarz R., Vatner R., et al. Osteosarcoma. Pediatr. Blood Cancer. 2021;68(Suppl 2):e28352. doi: 10.1002/pbc.28352. [DOI] [PubMed] [Google Scholar]
- 3.Borović Šunjić S., Jaganjac M., Vlainić J., Halasz M., Žarković N. Lipid peroxidation-related redox signaling in osteosarcoma. Int. J. Mol. Sci. 2024;25(8):4559. doi: 10.3390/ijms25084559. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Xiao X., Wang W., Li Y., et al. HSP90AA1-mediated autophagy promotes drug resistance in osteosarcoma. J. Exp. Clin. Cancer Res. 2018;37(1):201. doi: 10.1186/s13046-018-0880-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Niu J., Yan T., Guo W., et al. The COPS3-FOXO3 positive feedback loop regulates autophagy to promote cisplatin resistance in osteosarcoma. Autophagy. 2023;19(6):1693–1710. doi: 10.1080/15548627.2022.2150003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sun M.X., An H.Y., Sun Y.B., Sun Y.B., Bai B. LncRNA EBLN3P attributes methotrexate resistance in osteosarcoma cells through miR-200a-3p/O-GlcNAc transferase pathway. J. Orthop. Surg. Res. 2022;17(1):557. doi: 10.1186/s13018-022-03449-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kun-Peng Z., Xiao-Long M., Chun-Lin Z. Overexpressed circPVT1, a potential new circular RNA biomarker, contributes to doxorubicin and cisplatin resistance of osteosarcoma cells by regulating ABCB1. Int. J. Biol. Sci. 2018;14(3):321–330. doi: 10.7150/ijbs.24360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bruheim S., Xi Y., Ju J., Fodstad O. Gene expression profiles classify human osteosarcoma xenografts according to sensitivity to doxorubicin, cisplatin, and ifosfamide. Clin. Cancer Res. 2009;15(23):7161–7169. doi: 10.1158/1078-0432.CCR-08-2816. [DOI] [PubMed] [Google Scholar]
- 9.Lee A.M., Ferdjallah A., Moore E., et al. Long non-coding RNA ANRIL as a potential biomarker of chemosensitivity and clinical outcomes in osteosarcoma. Int. J. Mol. Sci. 2021;22(20) doi: 10.3390/ijms222011168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Hu Y., Gu J., Shen H., et al. Circular RNA LARP4 correlates with decreased Enneking stage, better histological response, and prolonged survival profiles, and it elevates chemosensitivity to cisplatin and doxorubicin via sponging microRNA-424 in osteosarcoma. J. Clin. Lab. Anal. 2020;34(2):e23045. doi: 10.1002/jcla.23045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Mintz M.B., Sowers R., Brown K.M., et al. An expression signature classifies chemotherapy-resistant pediatric osteosarcoma. Cancer Res. 2005;65(5):1748–1754. doi: 10.1158/0008-5472.CAN-04-2463. [DOI] [PubMed] [Google Scholar]
- 12.Kubota D., Mukaihara K., Yoshida A., Tsuda H., Kawai A., Kondo T. Proteomics study of open biopsy samples identifies peroxiredoxin 2 as a predictive biomarker of response to induction chemotherapy in osteosarcoma. J. Proteomics. 2013;91:393–404. doi: 10.1016/j.jprot.2013.07.022. [DOI] [PubMed] [Google Scholar]
- 13.Skowronek P., Meier F. High-throughput mass spectrometry-based proteomics with dia-PASEF. Methods Mol. Biol. 2022;2456:15–27. doi: 10.1007/978-1-0716-2124-0_2. [DOI] [PubMed] [Google Scholar]
- 14.Meier F., Brunner A.D., Frank M., et al. diaPASEF: parallel accumulation-serial fragmentation combined with data-independent acquisition. Nat. Methods. 2020;17(12):1229–1236. doi: 10.1038/s41592-020-00998-0. [DOI] [PubMed] [Google Scholar]
- 15.Zhang F., Ge W., Ruan G., Cai X., Guo T. Data-independent acquisition mass spectrometry-based proteomics and software tools: a glimpse in 2020. Proteomics. 2020;20(17–18) doi: 10.1002/pmic.201900276. [DOI] [PubMed] [Google Scholar]
- 16.Vidova V., Spacil Z. A review on mass spectrometry-based quantitative proteomics: targeted and data independent acquisition. Anal. Chim. Acta. 2017;964:7–23. doi: 10.1016/j.aca.2017.01.059. [DOI] [PubMed] [Google Scholar]
- 17.Lesur A., Schmit P.O., Bernardin F., et al. Highly multiplexed targeted proteomics acquisition on a TIMS-QTOF. Anal. Chem. 2021;93(3):1383–1392. doi: 10.1021/acs.analchem.0c03180. [DOI] [PubMed] [Google Scholar]
- 18.Kimura Y., Shin J., Nakai Y., et al. Development of parallel reaction monitoring mass spectrometry assay for the detection of human norovirus major capsid protein. Viruses. 2022;14(7) doi: 10.3390/v14071416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kanehisa M., Sato Y., Morishima K. BlastKOALA and GhostKOALA: KEGG tools for functional characterization of genome and metagenome sequences. J. Mol. Biol. 2016;428(4):726–731. doi: 10.1016/j.jmb.2015.11.006. [DOI] [PubMed] [Google Scholar]
- 20.Cox J., Hein M.Y., Luber C.A., Paron I., Nagaraj N., Mann M. Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ. Mol. Cell Proteomics. 2014;13(9):2513–2526. doi: 10.1074/mcp.M113.031591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Tsugawa H., Cajka T., Kind T., et al. MS-DIAL: data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat. Methods. 2015;12(6):523–526. doi: 10.1038/nmeth.3393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hao Y., Chen M., Huang X., Xu H., Wu P., Chen S. 4D-diaXLMS: proteome-wide four-dimensional data-independent acquisition workflow for cross-linking mass spectrometry. Anal. Chem. 2023;95(37):14077–14085. doi: 10.1021/acs.analchem.3c02824. [DOI] [PubMed] [Google Scholar]
- 23.Qu Z., Han Y., Zhu Q., et al. A novel neutrophil extracellular traps signature for overall survival prediction and tumor microenvironment identification in gastric cancer. J. Inflamm. Res. 2023;16:3419–3436. doi: 10.2147/JIR.S417182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Li N., Li Y., Zheng P., Zhan X. Cancer Stemness-based prognostic immune-related gene signatures in lung adenocarcinoma and lung squamous cell carcinoma. Front. Endocrinol. (Lausanne) 2021;12 doi: 10.3389/fendo.2021.755805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Huang G.Z., Wu Q.Q., Zheng Z.N., et al. Bioinformatics analyses indicate that cathepsin G (CTSG) is a potential immune-related biomarker in oral squamous cell carcinoma (OSCC) Onco Targets. Ther. 2021;14:1275–1289. doi: 10.2147/OTT.S293148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Shen Y., Chen H., Gong X., Wang Z., Chen M., Chen D. Identification of lysosome-related genes in connection with prognosis and immune cell infiltration for drug candidates in head and neck cancer. Open. Life Sci. 2023;18(1) doi: 10.1515/biol-2022-0660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yu M., Chen Q., Li Q., et al. A disintegrin and metalloprotease 10 expressions modulate potential metastatic and thrombus formation in pancreatic carcinoma. Iran. J. Public Health. 2022;51(8):1778–1789. doi: 10.18502/ijph.v51i8.10262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Chan S., Wang X., Wang Z., et al. CTSG suppresses colorectal cancer progression through negative regulation of Akt/mTOR/Bcl2 signaling pathway. Int. J. Biol. Sci. 2023;19(7):2220–2233. doi: 10.7150/ijbs.82000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Morimoto-Kamata R., Yui S. Insulin-like growth factor-1 signaling is responsible for cathepsin G-induced aggregation of breast cancer MCF-7 cells. Cancer Sci. 2017;108(8):1574–1583. doi: 10.1111/cas.13286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Hu Y., Wang H., Liu Y. NETosis: sculpting tumor metastasis and immunotherapy. Immunol. Rev. 2024;321(1):263–279. doi: 10.1111/imr.13277. [DOI] [PubMed] [Google Scholar]
- 31.Zheng C., Xu X., Wu M., et al. Neutrophils in triple-negative breast cancer: an underestimated player with increasingly recognized importance. Breast. Cancer Res. 2023;25(1):88. doi: 10.1186/s13058-023-01676-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Bonilha C.S., Veras F.P., de Queiroz Cunha F. NET-targeted therapy: effects, limitations, and potential strategies to enhance treatment efficacy. Trends. Pharmacol. Sci. 2023;44(9):622–634. doi: 10.1016/j.tips.2023.06.007. [DOI] [PubMed] [Google Scholar]
- 33.Herre M., Cedervall J., Mackman N., Olsson A.K. Neutrophil extracellular traps in the pathology of cancer and other inflammatory diseases. Physiol. Rev. 2023;103(1):277–312. doi: 10.1152/physrev.00062.2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.De Meo M.L., Spicer J.D. The role of neutrophil extracellular traps in cancer progression and metastasis. Semin. Immunol. 2021;57 doi: 10.1016/j.smim.2022.101595. [DOI] [PubMed] [Google Scholar]
- 35.Yang L., Liu Q., Zhang X., et al. DNA of neutrophil extracellular traps promotes cancer metastasis via CCDC25. Nature. 2020;583(7814):133–138. doi: 10.1038/s41586-020-2394-6. [DOI] [PubMed] [Google Scholar]
- 36.Papayannopoulos V. Neutrophil extracellular traps in immunity and disease. Nat. Rev. Immunol. 2018;18(2):134–147. doi: 10.1038/nri.2017.105. [DOI] [PubMed] [Google Scholar]
- 37.Saddozai U.A.K., Wang F., Khattak S., et al. Define the two molecular subtypes of epithelioid malignant pleural mesothelioma. Cells. 2022;11(18):2924. doi: 10.3390/cells11182924. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Saddozai U.A.K., Wang F., Akbar M.U., et al. Identification of clinical relevant molecular subtypes of pheochromocytoma. Front. Endocrinol. (Lausanne) 2021;12 doi: 10.3389/fendo.2021.605797. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Perez-Riverol Y., Bai J., Bandla C., et al. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic. Acids. Res. 2022;50(D1):D543–d552. doi: 10.1093/nar/gkab1038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Deutsch E.W., Bandeira N., Perez-Riverol Y., et al. The ProteomeXchange consortium at 10 years: 2023 update. Nucleic. Acids. Res. 2023;51(D1):D1539–d1548. doi: 10.1093/nar/gkac1040. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional file 4: Fig. S1 Results of GSEA of KEGG pathways. a Extracellular matrix (ECM) receptor interaction, complement and coagulation cascades, regulation of actin cytoskeleton, arachidonic acid metabolism, drug metabolism cytochrome P450, metabolism of xenobiotics by cytochrome P450, the MTOR signaling pathway, and ABC transporters were enriched according to GSEA between the RG and PG. b ECM receptor interaction, complement and coagulation cascades, arachidonic acid metabolism, cell adhesion molecules, cytokine-cytokine receptor interaction, and ABC transporters were enriched according to GSEA between the MG and PG.
Additional file 7: Fig. S2 Western blotting revealed greater expression of CTSG in the RG and MG than in the PG. β-tubulin was used as a loading contral.
Data Availability Statement
The dataset supporting the conclusions of this article is available in the ProteomeXchange Consortium repository, [PXD051097] [39,40].