Introduction

Multiple myeloma (MM), the second most common hematological malignant tumor after non-Hodgkin’s lymphoma1, is a malignancy characterized by the abnormal proliferation of plasma cells and has an overall survival (OS) ranging from a few months to over a decade2,3. The pathogenesis of MM is closely associated with oncogene activation, metabolic abnormalities and immune system dysfunction4,5,6. In addition, the abnormal bone marrow microenvironment under these conditions promotes the growth of MM cells and contributes to their resistance to conventional therapies. Therefore, exploring effective adjuvant treatments, combination therapies and novel biomarkers is essential for predicting disease progression and improving outcomes in newly diagnosed and relapsed/refractory MM patients.

A strong association has been reported between oxidative stress and the onset and progression of MM7,8. Research has suggested that certain antioxidants help prevent tumors by inhibiting DNA damage caused by excessive reactive oxygen species (ROS)9,10,11. Conversely, other studies have argued that antioxidants may also support cancer cell growth and proliferation by shielding them from ROS-induced death12,13,14. Thus, the role of antioxidants in cancer remains controversial and warrants further investigation. Peroxiredoxins (PRDXs) are a superfamily of antioxidative proteins that protect against oxidative stress and maintain cellular homeostasis. They are also dysregulated in various cell types, where they regulate critical biological processes, including cell proliferation, apoptosis, lipid metabolism and the immune response15,16,17. PRDX1 is detected in oral squamous cell carcinoma, and cervical, ovarian and colon cancer, which is closely related to proliferation, invasion, and metastasis. PRDX1 deficiency has a negative effect on the viability of philadelphia-positive (Ph) B-cell acute lymphoblastic leukemia (B-ALL) cells and makes them sensitive to tyrosine kinase inhibitors (TKIs) and nonhomologous end-joining (NHEJ) inhibitors18. The increased expression of PRDX2 in mantle cell lymphoma (MCL) is significantly correlated with the overexpression of SOX11, increased ROS, and poorer OS19. In addition, the peroxidase activity of PRDX2 is activated by CDK2, which inhibits the differentiation of acute myeloid leukemia (AML) cells20. PRDX3 is the target gene of c-myc, miR-383, and miR-23b, and is one of the essential molecules for mitochondrial homeostasis and tumor transformation. PRDX4 plays a tumor-promoting role in several kinds of cancers, such as renal papillary cell carcinoma, glioblastoma, lung adenocarcinoma, and hepatocellular carcinoma. While it has been studied that PRDX4 is strongly decreased in acute promyelocytic leukemia (APL)21. PRDX5 is significantly increased in primary macrophages exposed to LPS and IFN-γ22. PRDX6 plays a key role in mitochondrial dysfunction and ferroptosis, and promotes tumor invasion and metastasis23,24.

However, to date, a comprehensive analysis of PRDX expression as well as its clinical significance and function in MM is lacking. In this study, various online analysis tools and a series of experiments were used to investigate the role of PRDXs in MM. In addition, the coexpression network of the PRDXs was analysed before potential biological functions and signaling pathways associated with these genes were identified. Finally, their correlation with tumor microenvironment (TME) components was explored. The flowchart of this study illustrated in Fig. 1. Overall, these findings revealed that PRDXs, which link ferroptosis and the immune microenvironment, could serve as novel biomarkers for MM and provide a basis for developing new therapeutic strategies.

Fig. 1
figure 1

The schematic illustration of PRDXs in MM.

Full size image

Materials and methods

Bioinformatics analysis of gene expression and clinical data

The Cancer Cell Line Encyclopedia (CCLE) (https://portals.broadinstitute.org/ccle) provides public access to genomic data, analysis and visualization for 1457 cell lines, with multiple data identifiers available for each gene. This study used mRNA expression (RNAseq) data for the PRDX family across various cancer cell lines. Transcriptomic and clinical data were publicly retrieved from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). The PRDX gene expression levels were then measured as log2 values before the relative expression levels were presented as the mean ± standard deviations (SDs). The analyses were eventually conducted via R (version 4.4.2) and SPSS (version 26). We applied Benjamini–Hochberg false discovery rate (FDR) correction via the adjust = ‘fdr’ parameter in limma’s ‘topTable()’ function. The differentially expressed genes were filtered based on FDR-adjusted P-value < 0.05 combined with log₂ fold-change > 1.

Receiver operating characteristic (ROC) curves were generated for the RNA sequencing data from the GSE2658 and GSE5900 datasets via the ‘pROC’ package in R to estimate the diagnostic value of the PRDXs in MM. The Kaplan–Meier plotter (http://kmplot.com/analysis/) was able to assess the survival impact of 54,000 genes across 21 types of cancer. In this case, survival curves were plotted, and the risk ratios with 95% confidence intervals (CIs) were calculated along with log-rank P values. The effects of the PRDX family on the OS of 1416 MM patients, the event-free survival (EFS) of 801 MM patients and the postprogression survival (PPS) of 659 MM patients were examined. OS is defined as the time from initial diagnosis (or randomization) to death from any cause. EFS refers as the time from diagnosis (or randomization) to the first event (disease progression, relapse, new therapy initiation or death). PPS is defined as the time from the first documented disease progression to death from any cause. The characteristics of all bioinformatics datasets analyzed in this study are comprehensively detailed in Supplementary Table 1.

Enrichment analysis

The PRDX network and its associated genes were then constructed using STRING (https://cn.string-db.org/), and the hub genes were identified with the CytoHubba plugin in Cytoscape 3.6.1 (https://cytoscape.org/). These hub genes were subsequently submitted to DAVID25 (https://cytoscape.org/) for Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis26,27. The results were eventually visualized via lollipop plots in GraphPad Prism 9. To explore the potential functions of PRDXs in MM, samples from the GSE24080 and The Cancer Genome Atlas (TCGA)-Multiple Myeloma Research Foundation (MMRF) cohorts were divided into high- and low-expression groups on the basis of mean PRDX expression levels. Gene Set Enrichment Analysis (GSEA)28 (www.gsea-msigdb.org/gsea/index.jsp) was then performed to identify the biological functions enriched in these genes, with the analysis conducted via the ‘clusterProfiler’ and ‘ggplot2’ packages in R. In this case, an FDR (q value) of < 0.25 and a NOM (P value) of < 0.05 were selected as thresholds for statistical significance.

Ferroptosis-related genes and CIBERSORT analyses

The relationships between PRDX expression and ferroptosis-related genes were analysed in the GSE24080 and TCGA-MMRF datasets, and correlation heatmaps were subsequently generated using GraphPad Prism (version 9.0). CIBERSORT is a computational tool for characterizing the cell composition of complex tissues based on gene expression profiles29. CIBERSORT algorithm with the Leukocyte signature matrix (LM22) gene signature matrix can accurately estimate the immune landscape and therefore, was employed to investigate the correlation between immune cell populations and key genes. To deconvolve the cellular composition of the bone marrow, the CIBERSORT algorithm with the LM22 gene signature matrix was applied to the mRNA expression data from GSE276561 dataset to determine the expression scores of 22 immune cell types. CIBERSORT analysis outputs the relative proportions of 22 immune cell types for each sample, with all fractions collectively summing to 1.0. A total of 704 MM samples were covered. In addition, we used Pearson correlation analysis to assess the correlation between PRDX expression and the proportion of immune cells in MM samples via R (P < 0.05). Expression data were also examined in relation to chemokines, cytokines and various immunosuppressive molecules, with the results eventually visualized via the ‘ggboxplot’ package in R.

Collection of clinical specimens

Bone marrow aspirates were collected from 28 MM patients who were treated at the Department of Hematology at the Second Hospital of Lanzhou University, while 18 aspirates from patients without hematologic malignancies served as controls. Furthermore, clinical data from MM patients were also collected. All experimental procedures were performed in accordance with the Declaration of Helsinki, and ethical approval for the bone marrow collection was obtained from the Ethics Committee of the Second Hospital of Lanzhou University, with informed consent also provided by all participants prior to the study.

Cell culture and establishment of cell lines

Human MM cell lines (RPMI-8226, MM1.S, AMO1, and U-266), purchased from Procell Life Science and Technology (Wuhan, China), were cultured in RPMI-1640 medium (Haoyang Huake, HY1640, China) supplemented with 10% fetal bovine serum (ABW, AB-FBS-0050S, Germany). PRDX2-knockdown lentiviral vectors were generated by GeneChem Gene (Shanghai, China), while a lentiviral vector without shRNA fragments (or containing unrelated RNA fragments) served as the control (shNC). ShPRDX2-1: 5'-CCTCTTTATCATCGATGGCAA-3', shPRDX2-2: 5'-AGATGAGGGCATTGCCTAC-3'. Cell transfection was performed when the cell culture reached approximately 70% confluence. The experimental procedure involved seeding 5 × 105 cells into 6-well plates containing shPRDX2 lentiviruses (MOI = 30), with the RPMI-1640 medium replaced with fresh medium after 24 h. Following a 72-h incubation, puromycin was added at a final concentration of 2.5 µg/mL, and stable cell lines were eventually established after two weeks of screening.

Total RNA isolation and quantitative real time-PCR (qRT-PCR)

We used the Steadypure Quick RNA Extraction Kit (AG, AG21022, China) for RNA isolation, followed by cDNA synthesis using 5× SweScript ALL-in-one SuperMix for qPCR (Servicebio, G3337-1, China) and gDNA Remover (Servicebio, G3337-2, China). Subsequently, qRT-PCR analysis was performed using 2× SYBR Green qPCR Master Mix (Servicebio, G3321, China). GAPDH was used as an internal reference. All samples were run in triplicate, and the mean values were used for quantitative analyses. The primer sequences are listed in Supplementary Table 2.

Western blotting (WB)

RIPA buffer (Solarbio, R0010, China) and PMSF 100 mM (Solarbio, P0100-01, China) were combined with protease inhibitors to extract proteins. The protein lysates were separated on 12% SDS-PAGE gels and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, IPVH00010, Italy). After the membranes were blocked with 5% skim milk for 2 h, they were incubated overnight at 4 °C with the following specific primary antibodies. Finally, the membranes were exposed after treatment with secondary antibodies using the Enhanced Chemiluminescence (ECL) Kit (Biosharp, BL520B, China). Images were captured with a Tanon 5200 Multi image processing system and quantified with Image Processing and Analysis in Jave (ImageJ) software version 1.54 program (https://imagej.net/ij/download.html). β-Tubulin or β-actin was used as an internal reference. Each sample was performed in triplicate to ensure accuracy and reproducibility. Antibodies used in this study. Rabbit anti-PRDX2 (1:1000, Proteintech, 10,545–2-AP), rabbit anti-GPX4 (1:10,000, HuaAn, ET1706–45), rabbit anti-SLC7A11 (1:1000, Proteintech, 26864-1-AP), mouse anti-β-actin (1:20,000, Proteintech, 66009-1-Ig), mouse anti-β-Tubulin (1:1000, Servicebio, GB15140), goat anti-rabbit (1:5000, Proteintech, SA00001-2), and goat anti-mouse (1:5000, Proteintech, SA00001-2) antibodies were used.

Detection of cell proliferation by the Cell Counting Kit-8 (CCK-8) assay

A total of 5000 cells were seeded into each well of 96-well plates, with 100 μL of medium per well. Four experimental groups (RPMI-8226: shPRDX2, shNC; AMO1: shPRDX2, shNC) were then established, each with five replicate wells. After 0, 24, 48 and 72 h, 10 μL of CCK-8 reagent was added, followed by a 2.5-h incubation before the OD values were recorded at 450 nm via a microplate reader. For each group, the highest and lowest OD values were excluded, and the average OD value of the remaining three wells was calculated.

Transwell assay

To evaluate the migration and invasion capabilities of MM cells, Transwell chambers (8 μm pore size; Corning, 3397, USA) were pre-coated without or with Matrigel (Corning, 356,234, USA). Transfected MM cells were collected and resuspended in serum-free RPMI-1640 medium. A total of 1 × 105 cells were seeded into the upper chamber, while RPMI-1640 medium containing 10% FBS was added to the lower chamber as a chemoattractant. After incubation at 37 °C for 48 h, cells that had migrated or invaded to the lower chamber were stained with 1% crystal violet. The number of cells was counted using an inverted microscope (Olympus, Tokyo, Japan), and images were captured for cell quantification.

Flow cytometry

Whole blood staining was performed within 24 h of blood collection, using one million cells per tube. The following fluorescently labelled monoclonal antibodies (mAbs) were added to the tubes and used at the concentrations given in the instructions: anti-CD3-FITC, anti-CD45-PerCP, anti-CD8-PE and anti-CD4-APC, anti-CD19-APC, anti-CD16 + CD56-PE, anti-CD4-FITC, anti-CD25-APC, anti-CD127-PE (all from BD Pharmingen, USA). The percentages of immune cells (Tregs, total T cells, CD4 + T cells, CD8 + T cells, NK cells, and total B cells) were evaluated by flow cytometry. The serum levels of IL-1, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, TNF-α, IFN-γ, IFN-α, and IL-17A were assayed via the Cytokine Detection Kit (JIANGXI CELLGENE BIOTECH, China). The experiments were performed on a BD FACSCanto™ II (Becton Dickinson) flow cytometer.

ROS, glutathione (GSH), malondialdehyde (MDA), and ferrous ion detection

The ROS Assay Kit (S0033S) and MDA Assay Kit (S0131S) were purchased from Beyotime Biotechnology, while the GSH assay Kit (A006-2–1) was obtained from Nanjing JianCheng Biotechnology Research Institute. Moreover, the Ferrous Ion Content Assay Kit (BC5415) was supplied by Beijing Solarbio Science and Technology. DCFH-DA was loaded into the cells, which were subsequently incubated at 37 °C for 20 min. The residual reagent was then washed off with PBS, after which the cells were observed under a fluorescence microscope (Olympus, IX53, Japan). Furthermore, the cells were lysed via ultrasonication, and the resulting supernatant was collected. The reagents were then sequentially added to 96-well plates according to the kit instructions before the absorbance values were recorded at 405 nm, 530 nm and 593 nm using a microplate reader, with three replicates per group.

Statistical analysis

Two-group comparisons were subsequently performed via the two-tailed unpaired T-test or the Wilcoxon test. In addition, categorical variables were analysed via the Chi-square test. Correlations between variables were then evaluated via Pearson correlation analysis, with all data analyses performed with GraphPad Prism (version 9.0), R (version 4.4.2), and SPSS (version 26). Statistical significance was set as P < 0.05.

Results

Expression levels, diagnostic and prognostic value of PRDXs in MM as determined from the analysis of different databases

PRDX expression in MM cell lines and patients at different stages of disease progression was analysed. For this purpose, the mRNA levels of PRDXs in various solid and hematological tumors were retrieved from the CCLE database. Overall, it was found that PRDX2 and PRDX5 were relatively highly expressed in MM, suggesting their potential roles in MM pathogenesis (Fig. 2A, B and Supplementary Fig. 1). The transcript levels of PRDXs in normal and MM samples from GEO datasets were analysed. Initially, two GEO datasets (GSE19784 and GSE5900), processed using the same microarray platform (GLP570), were merged into a single MM database. The mRNA expression levels of the PRDXs were then calculated for 22 healthy controls (HDs) and 328 newly diagnosed multiple myeloma (NDMM) patients. The results revealed that the expression of PRDX2/3/4/5/6 was significantly different between multiple myeloma cells (MMCs) and normal plasma cells (NPCs) (P < 0.05), while no significant differences in PRDX1 expression were observed (Fig. 2C). Publicly available data (GSE68891) were further analysed to compare the PRDX expression levels between mature CD138 + CD19- plasma from 126 myeloma patients and CD138-CD19 + MM precursor cells from 17 myeloma patients. The findings revealed greater expression of PRDX1/2/3/4/5 in CD138 + MM cells than in MM precursor cells (P < 0.05, Fig. 2D). In contrast, PRDX6 expression did not differ significantly between the two groups. As shown in Fig. 2E, analysis of the GSE6477 dataset revealed that the levels of PRDX2 and PRDX6 were significantly increased in smoldering multiple myeloma (SMM, n = 24), and MM (n = 73) compared with HD (n = 15) (P < 0.05). Among the HD, monoclonal gammopathy of undetermined significance (MGUS, n = 22), SMM, and MM groups, PRDX6 expression exhibited a progressive increase, while PRDX5 was undetected. Furthermore, analysis of the GSE31161 dataset revealed significantly increased levels of PRDX1/2/3/6 in bone marrow clonal plasma cells (BMPCs) from relapsed/refractory multiple myeloma (RRMM) patients (n = 258) compared with those in baseline MMCs (n = 780) (Fig. 2F). These comprehensive analyses of multiple publicly available datasets indicated that PRDX2/4/5/6 were significantly upregulated in MM patients, with relatively high levels of PRDX1/2/3/6 in RRMM patients.

Fig. 2
figure 2

Expression, prognostic and diagnostic value of PRDXs in multiple myeloma. Transcriptional levels of (A) PRDX1 and (B) PRDX2 in various tumor cell lines as obtained from the CCLE database, with the solid line in the box representing the average levels. (C) PRDXs in HD and NDMM. (D) PRDXs in CD138 + MM cells (n = 126) and CD138-CD19 + B cells (n = 17) from myeloma patients. (E) PRDXs in HD (n = 15), MGUS (n = 22), SMM (n = 24), and MM (n = 73) patients. (F) PRDXs in baseline MM cells (n = 780) and relapsed MM cells (n = 258). (G) ROC curve of the diagnostic value of the PRDX family in MM, in which the x-axis is 1-specificity and the y-axis is sensitivity. The AUCs were recorded in the right corner. (H) Kaplan–Meier plotter analysis of the relationship between PRDX1 and OS, EFS, and PPS. (I) Kaplan–Meier analysis of the impact of PRDX2 expression on EFS. *P < 0.05, **P < 0.01, ***P < 0.001.

Full size image

ROC curve analysis was performed using the PRDX family data from GSE2658 and GSE5900. ROC curve analysis showed area under the curves (AUCs) > 0.6 for all 6 genes, with PRDX2 showing high diagnostic potential with an AUC of 0.881 (Fig. 2G). These results highlight the potential of PRDXs as biomarkers, with PRDX2 emerging as a particularly promising diagnostic biomarker for MM patients due to its highest AUC value. Kaplan–Meier plotter was used to evaluate the impact of the PRDX family on survival (Fig. 2H, I and Supplementary Fig. 2). Overall, high expression of PRDX1 was associated with shorter OS and PPS, whereas those exhibiting higher PRDX1 or PRDX2 expression experienced poorer EFS.

Relationships between PRDXs and the clinical characteristics of MM patients

The relationships between PRDX expression and the clinical features of NDMM patients were also analysed using the GSE24080 dataset. GSE24080 includes pretreatment bone marrow plasma cells from 559 NDMM patients enrolled in the TT2 and TT3 clinical trials. According to the average PRDX cut-off, we divided the 559 MM patients into high- and low-expression groups and then uncovered that PRDX1 and PRDX2 were correlated with the M protein type (P = 0.012, P < 0.001), the number of focal lesions defined by magnetic resonance imaging (MRI) (P < 0.001, P = 0.006) and the lactate dehydrogenase (LDH) level (P < 0.001, P = 0.007), as shown in Table 1. Additionally, PRDX2 expression was associated with serum β2-microglobulin (β2-MG) (P = 0.003) and creatinine (P = 0.032) levels (Table 1). Statistical analysis revealed no significant differences in sex, age or cytogenetic abnormalities between the groups. Furthermore, PRDX5 increased with C-reactive protein (CRP) and LDH levels, while MM patients with low PRDX6 expression exhibited lower creatinine and LDH levels (P < 0.05).

Table 1 Correlations between PRDXs and the clinical characteristics of MM patients.
Full size table

Enrichment analyses of coexpressed genes in MM

To further elucidate the functions of the PRDXs, a protein–protein interaction network was constructed using STRING before identifying the hub genes with the CytoHubba plugin in Cytoscape (Supplementary Fig. 3A). Subsequent functional and pathway enrichment analyses via DAVID revealed significant enrichment of selected genes related to antioxidant activity, protection against oxidative stress, glutathione peroxidase activity and cancer formation, with these results presented in lollipop charts (Fig. 3A–C and Supplementary Fig. 3B–D). Additionally, PRDX1, PRDX2 and PRDX6 were found to be involved in ferroptosis.

Fig. 3
figure 3

Function of PRDXs in multiple myeloma. Lollipop charts showing the functional and pathways enrichment analyses of (A) PRDX1, (B) PRDX2 and (C) PRDX6. GSEAs identifying the signaling pathways associated with (D) PRDX1 and (E) PRDX2. The relationship of PRDX1 to ferroptosis (F) inducers and (G) inhibitors was analysed separately in the GSE24080. TCGA-MMRF was used to analyse the correlation between PRDX2 and ferroptosis (H) inducers and (I) inhibitors. (J) The lollipop plots show the relationship between PRDXs mRNA levels and immune cell infiltration in the GSE276561. (K) Differential expression of chemokines, cytokines, and immunosuppressive factors between high- and low-expression groups. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Full size image

GSEAs using the GSE24080 and TCGA-MMRF datasets linked high PRDX1 expression with the "B cell receptor signalling pathway", "ErbB signalling pathway", "mTOR signalling pathway", "RIG-I-LIKE receptor signalling pathway", and "FCγR-mediated phagocytosis" (Fig. 3D). The genes positively associated with PRDX2 were enriched in “metabolism”, “oxidative phosphorylation”, “cell cycle”, “adhesion”, “proteasome”, "calcium signalling pathway", "chemokine signalling pathway", "JAK/STAT signalling pathway", and "FCγR-mediated phagocytosis", as shown in Fig. 3E. The GSEAs of the other members are presented in Supplementary Fig. 3E–H.

PRDXs related to ferroptosis and immune infiltration in MM

To investigate the relationship between PRDXs and ferroptosis, Pearson correlation analyses were performed using the GSE24080 and TCGA-MMRF datasets. The analyses included ferroptosis inducers (ACSL1, ACSL4, ACSL6, ALOX15, ATG5, ATG7, FTL, LPCAT3, MAP1LC3A, MAP1LC3B, NCOA4, SAT1, SLC39A14, TF, TFRC, TP53, VDAC2, and VDAC3) and inhibitors (ACSL3, FTH1, GCLC, GCLM, GPX4, GSS, HMOX1, PCBP1, SLC3A2, SLC40A1, and SLC7A11), with the results revealing a strong association between PRDXs and ferroptosis regulators in MM (Fig. 3F–I).

Enrichment analysis further suggested that PRDXs may influence immune-related pathways. Hence, to explore the link between PRDX expression and the TME, CIBERSORT, a whole bone marrow gene expression-based computational technique, was used to determine the proportion of tumor-infiltrating immune cells (TICs) in 704 NDMM samples from GSE276561. A comprehensive analysis was then cconducted using Pearson correlation to examine the relationship between PRDX gene expression and different cell types. We performed immune infiltration analysis on 704 NDMM samples from GSE276561. CIBERSORT analysis of the bone marrow biopsy GEP data revealed PRDX1 expression was positively correlated with that of M1 macrophages (R = 0.25, P < 0.001), M2 macrophages (R = 0.17, P < 0.001), dendritic cells (R = 0.12, P = 0.002), and T cells regulatory (R = 0.09, P = 0.023) but negatively correlated with neutrophils (R = − 0.27, P < 0.001), naive CD4 + T cells (R = − 0.25, P < 0.001) and monocytes (R = − 0.23, P < 0.001). The other PRDX family members also exhibited the following correlations: PRDX2/3/4/5/6 with T cells, mast cells, NK cells and monocytes, PRDX2/3/4/6 with macrophages, neutrophils, and PRDX2/4 with dendritic cells (P < 0.05, Fig. 3J). The relationships between PRDXs and chemokines, cytokines and immunosuppressive factors, including CCL2, CCL3, CXCL12, CXCL13, CCL19, CXCL10, IL-4, IL-6, IL-10, TNF-α, PD-1, and PD-L1, were also examined. PRDX1 was correlated with CCL2, CCL3, CXCL13, CCL19, CXCL10, IL-4, TNF-α, PD-1, and PD-L1; and PRDX2 with CCL2, CCL3, CXCL12, CXCL13, CXCL10, TNF-α, PD-1, and PD-L1; PRDX3 with increased CXCL10, IL-10, TNF-α and decreased PD-L1; PRDX4 with increased CCL2, CCL3 and IL-6; PRDX5 with increased PD-1 but decreased CCL3, CXCL12, CCL19 and PD-L1; and PRDX6 with CCL2, TNF-α PD-1, and PD-L1. These results are presented in Fig. 3K.

Validation of PRDX2 in MM

Previous studies have investigated the roles of PRDX1/4/5/6 in MM30,31,32, and it was found that PRDX2 expression could increase progressively during MM progression, hence indicating its potential as a biomarker. Subsequent functional enrichment analysis further linked PRDX2 to redox homeostasis and ferroptosis, with both pathways being critical for MM survival and resistance. Therefore, PRDX2 was selected for further investigation. PRDX2 expression in bone marrow samples from 28 MM patients and 18 HD was examined. qRT-PCR analysis subsequently revealed significantly higher PRDX2 mRNA levels in MM patients compared with HD, with relative expression evaluated via the 2 − ΔΔCt method (P < 0.0001, Fig. 4A). WB analysis of PRDX2 expression in five cases of NDMM and two cases of HD further confirmed this result (Fig. 4B). Moreover, PRDX2 expression in MM cell lines (RPMI-8226, AMO1, MM1.S and U-266) was assessed, with the findings showing significantly higher PRDX2 expression in RPMI-8226 and AMO-1 cells than in MM1.S and U-266 cells (Fig. 4C, D).

Fig. 4
figure 4

The expression and function of PRDX2 in MM. (A) Relative mRNA expression of PRDX2 in MM and HD. (B) PRDX2 in MM (n = 5) and HD (n = 2) were detected. T: MM, N: HD. (C, D) The expression and quantitative analysis of PRDX2 in myeloma cell lines (RPMI-8226, AMO1, MM1.S, and U-266). Relative quantification was normalized to U-266 as the reference. Data shown represent mean ± SD (N = 3). (E, F) Transfection effects of two different sequences were detected by qRT-PCR and WB in RPMI-8226 and AMO1 cells. Data shown represent mean ± SD (N = 3). (G, H) Results of CCK8 after targeting PRDX2 in RPMI-8226 and AMO1 cells (N = 3). (I, J) Invasion and migration of MM cells measured by Transwell assay. Scale bar: 100 μm. Data shown represent mean ± SD (N = 3). Full-length blots are presented in Supplementary Fig. 4.

Full size image

In vitro experiments involving lentiviral transfection of shPRDX2 and shNC were also conducted to establish stable MM cell lines (RPMI-8226 and AMO1). Transfection effects of two different sequences were detected by qRT-PCR and WB (Fig. 4E, F). WB analysis of RPMI-8226 cells subsequently showed that shPRDX2-2 had a superior interference effect compared with shPRDX2-1 (shPRDX2-1, P = 0.005; shPRDX2-2, P = 0.005). This transfection efficiency was then validated in the AMO-1 cell line, and sequence 2 was selected for further experiments (P = 0.015). CCK-8 assay performed on RPMI-8226 and AMO-1 cells revealed lower cell proliferation rates in the shPRDX2 group than in the shNC group (Fig. 4G, H). Finally, the number of invading and migrating MM cells was decreased in the shPRDX2 group (Fig. 4I, J).

Validation of the correlation between PRDX2 expression and clinical characteristics and immune-related indices

Clinical information was collected from 28 patients with MM who were treated in the Department of Hematology at the Second Hospital of Lanzhou University. The 28 MM patients included 19 males and 9 females, with an average age of 61 ± 8 years. Six patients had light chain MM, 12 had IgA MM, 8 had IgG MM, 1 had IgD MM, and 1 had non-secreting MM. Based on the median relative expression of PRDX2, MM patients were divided into low- and high-expression groups, with the clinical characteristics and immune-related indicators of these groups presented in Table 2. Notably, patients with high PRDX2 expression were more likely to exhibit elevated β2-MG levels (P < 0.001). There were no differences in other clinical features, probably because of the small sample size.

Table 2 Clinical characteristics and immune-related indicators in MM patients.
Full size table

The differences in immune cells and cytokines between the two groups were also compared. Significant differences were also observed between the two groups in terms of the percentages of CD4 + T cells, CD8 + T cells and CD4 + /CD8 + T cells as well as in the levels of IL-6, IL-8 and IL-12p70 (P < 0.05, Table 2). Flow cytometry showed that CD8 + T cells accounted for 52.62% of the total lymphocytes with high expression of PRDX2, which was significantly higher than 35.27% of those with low expression of PRDX2 (Fig. 5A, B). However, the percentages of CD4 + T and CD4 + /CD8 + T cells were lower than those in patients with low expression of PRDX2 (32.69%, 1.27 VS 44.82%, 0.62). Therefore, the relationships between PRDX2 expression and immune cells as well as cytokine levels were further explored. Correlation analysis revealed that PRDX2 expression was negatively correlated with the proportion of CD4 + T cells (R = − 0.448, P = 0.017), CD4 + /CD8 + T cells (R = − 0.455, P = 0.015) and the level of IFN-γ (R = − 0.404, P = 0.033), and was positively correlated with the expression of IL-6 (R = 0.484, P = 0.009) and IL-8 (R = 0.727, P < 0.001) (Fig. 5C).

Fig. 5
figure 5

Role of PRDX2 in ferroptosis and immune infiltration. Percentage of CD45 + , CD3 + , and CD4 + /CD8 + T cells in high (A) and low (B) PRDX2 expression groups as detected by flow cytometry. (C) The expression of PRDX2 was negatively correlated with the proportion of CD4 + T cells, CD4 + T/CD8 + T cells, and the expression of IFN-γ, and positively correlated with the IL-6 and IL-8. (D–F) PRDX2 regulates ROS levels in RPMI-8226 and AMO1 cells (20x). (G–I) GSH, MDA, and ferrous ion were measured in RPMI-8226 and AMO1 cells. (J, K) Expression of GPX4 and SLC7A11 in MM patients and normal samples. (L) Expression levels and quantitative analyses of GPX4 and SLC7A11 in RPMI-8226 and AMO1 cells of the shNC and shPRDX2 groups. Data shown represent mean ± SD (N = 3). Full-length blots are presented in Supplementary Fig. 5.

Full size image

The role of PRDX2 in ferroptosis in MM

Ferroptosis is a distinct type of programmed cell death characterized by excessive accumulation of intracellular iron as well as an increase in ROS levels. It is regulated by oxidative stress, iron overload and lipid metabolism, with MDA being a byproduct of lipid peroxidation. We investigated key indicators of ferroptosis by interfering with PRDX2 expression to determine the role of PRDX2 in ferroptosis. The ROS, GSH, MDA, and Ferrous Ion Content Assay Kits were used to investigate the role of PRDX2 in ferroptosis. DCFH-DA staining first demonstrated that PRDX2 could regulate ROS levels in RPMI-8226 and AMO1 cells (P = 0.006, P = 0.012) (Fig. 5D-F). GSH is an antioxidant that maintains redox balance and protects cells against oxidative stress. After the knockdown of PRDX2, the GSH level was significantly reduced (P = 0.028, P = 0.043), the levels of intracellular MDA and ferrous ion were significantly increased by approximately twofold (Fig. 5G–I).

More importantly, the mRNA levels of glutathione peroxidase 4 (GPX4) and solute carrier family 7 member 11 (SLC7A11), which are ferroptosis-related genes, were significantly elevated in patients (n = 28) compared with normal individuals (n = 18), suggesting an inhibitory effect on ferroptosis in MM (P < 0.001, P = 0.006) (Fig. 5J, K). Since GPX4 and SLC7A11 are key regulators of ferroptosis, their relationships with PRDX2 were further studied. Overall, GPX4 and SLC7A11 expression decreased after the inhibition of PRDX2 (P < 0.05, Fig. 5L). Collectively, these findings indicated that PRDX2 deficiency can induce ferroptosis in MM cells by inhibiting the GPX4-SLC7A11 signaling pathway, thereby effectively suppressing tumor growth.

Ferrostatin-1 negated the protective effect of PRDX2 knockdown on RSL3-induced ferroptosis

Cell proliferation was assessed using the CCK-8 assay in shNC or shPRDX2 groups after inducing ferroptosis with RSL3, a GPX4 inhibitor and a ferroptosis inducer. The shPRDX2 group exhibited a lower half-maximal inhibitory concentration (IC50) compared to the shNC group, suggesting that PRDX2 downregulation enhanced the cytotoxic effects of RSL3 on MM cells (Fig. 6A). Cell proliferation assays were also performed for both groups under RSL3 treatment with or without ferrostatin-1 (Fer-1, a ferroptosis inhibitor) pretreatment. In this case, Fer-1 significantly decreased RSL3-induced growth inhibition in MM cells (Fig. 6B). In addition, the impaired invasion and migration abilities observed after PRDX2 knockdown were restored by Fer-1 treatment (Fig. 6C, D). Phenotypic analysis of ferroptosis under the above experimental conditions further revealed that the combined Fer-1 and RSL3 treatment resulted in elevated GSH levels as well as reduced ROS, MDA and Fe2⁺ levels compared with RSL3 treatment alone (Fig. 6E–H). Hence, Fer-1 could effectively reverse the protective effect of PRDX2 knockdown against RSL3-induced ferroptosis. Finally, western blot analysis of key proteins in the ferroptosis signaling pathway, performed on shNC and shPRDX2 MM cells treated with or without Fer-1, showed that Fer-1 could rescue ferroptosis that was induced by PRDX2 knockdown (Fig. 6I).

Fig. 6
figure 6

PRDX2 enhanced the resistance to ferroptosis in MM cells. (A) Dose–response curves of the RSL3 in RPMI-8226 and AMO1 cells transfected with either shNC or shPRDX2. MM cells in the two transfection groups were further divided into four treatment groups: vehicle control, Fer-1 treatment alone, RSL3 treatment alone and combination treatment with both RSL3 and Fer-1. Cell proliferation (B), invasion (C), migration (D), and ferroptosis phenotypes (E–H) were assessed across these groups. (I) Key ferroptosis-related signaling proteins were also assessed for the above groups. Data shown represent mean ± SD (N = 3). Full-length blots are presented in Supplementary Fig. 6.

Full size image

Discussion

MM is a genetically and clinically heterogeneous malignancy. While some patients achieve effective remission during the early stages of the disease33,34,35, drug resistance and relapse/refractory conditions remain the most significant challenges in treatment36,37. Advances in genomics and proteomics have therefore enhanced the current understanding of MM pathogenesis. In particular, the PRDX family is found to be involved in the regulation of cell proliferation, apoptosis, lipid metabolism, and the immune response. In terms of the number of active-site cysteine (Cys) residues, PRDXs are classified into three types (typical 2-Cys, 2-Cys, and 1-Cys)38,39, all of which are 22–27 kDa proteins. PRDX2 is a typical 2-Cys antioxidant that protects cells from oxidative stress by scavenging H₂O₂ and modulating ROS levels40. There are few detailed studies on PRDX2’s role, expression-regulation relationships and the immune microenvironment in MM. We characterized PRDX2 expression profiles in MM cohorts and evaluated its survival impact. Ferroptosis markers were quantified using standardized assays to determine correlations with PRDX2. Furthermore, PRDX2 overexpression drives immune evasion through T cell exhaustion and M2 macrophage recruitment.

The results, derived from a relatively large cohort of participants, revealed increased PRDX2/4/5/6 in NDMM patients as well as relatively higher PRDX1/2/3/6 in RRMM patients, thus suggesting their potential involvement in MM development. Leveraging on a substantial dataset of MM samples with detailed clinical information, this study revealed that certain members of the PRDX family may contribute to relapse and adverse clinical outcomes. Demasi et al.41 demonstrated that PRDX1/4 are highly expressed in MM and in lymphomas characterized by high levels of immunoglobulin synthesis, such as Burkitt lymphoma and diffuse large B cell lymphoma (DLBCL). Ren et al.31 reported that the overexpression of PRDX5 protects IM9 myeloma cells from apoptosis induced by (−)-epigallocatechin 3-gallate (EGCG). Gao et al.32 reported that PRDX6 is elevated in MM and is negatively correlated with prognosis. PRDX6 protects MM cells from oxidative stress and maintains mitochondrial homeostasis, which is similar to our findings. However, the potential function of PRDX2 in MM has not been demonstrated. This study further demonstrated that knockdown of PRDX2 may inhibit MM cell proliferation, invasion and migration. Moreover, patients with high PRDX2 expression exhibited greater tumor burdens and their bone marrow plasma cells presented more malignant characteristics.

Ferroptosis is blocked by metal ion chelation, inhibition of lipid peroxidation, and scavenging of free radicals but not by apoptosis inhibitors or necroptosis inhibitors42. Its primary cytoprotective axis is the SLC7A11-GSH-GPX4 pathway43. SLC7A11 mediates cystine uptake, facilitating the production of Cys, which is essential for GSH biosynthesis. GSH acts as an essential cofactor for GPX4, enabling the efficient detoxification of lipid peroxides and suppression of oxidative stress damage. The overexpression of PRDX2 in adipose-derived mesenchymal stem cells (ADSCs) inhibited ferroptosis by regulating the GPX4/ACSL4 axis44. Recent studies45 indicate that hyperoxidized PRDX3 can translocate to the plasma membrane and inhibit the function of the cystine transporter SLC7A11, thereby compromising GSH synthesis and promoting ferroptosis. Under erastin treatment, PRDX1 depletion significantly elevates iron content and MDA levels, increases COX2 protein expression, and concurrently reduces GSH levels along with GPX4 and SLC7A11 protein expression in DLBCL46. Although PRDX2 exhibits distinct subcellular localization compared to PRDX1/PRDX3, all belong to the 2-Cys PRDX subfamily, sharing conserved structural features and susceptibility to hyperoxidation. PRDX2 may modulate cystine uptake via direct protein interactions or oxidative signaling cascades, thereby maintaining the expression of downstream antioxidant components GPX4 and SLC7A11 and preventing the accumulation of lipid peroxides. This hypothesis positions PRDX2 as an upstream regulatory node within the SLC7A11-GSH-GPX4 axis, providing a novel conceptual framework for targeting the PRDX2-ferroptosis pathway. We first found that the downregulation of PRDX2 in MM cells reduced GPX4, SLC7A11 and GSH levels while increasing the levels of ROS, MDA and ferrous ion to induce ferroptosis. Ferrostatin-1 abrogated the protective effect of PRDX2 knockdown against RSL3-induced ferroptosis.

In fact, various studies have suggested that ferroptosis could indeed play a critical role in MM. Since MM cells are particularly sensitive to iron overload, promoting the efficacy of bortezomib and carfilzomib through MM cell death via ferroptosis has been shown47,48. Furthermore, Bordini et al.49 also pointed out that iron overload could trigger MM cell death through lipid peroxidation. In addition, GPX4 and SLC7A11 are highly expressed in MM cells. FTY720, an immunosuppressant, can promote ferroptosis and autophagy by reducing the levels of GPX4 and SLC7A11 in U266 cells50. These findings align with previous observations in which the mRNA levels of GPX4 and SLC7A11 were upregulated in MM patients. Taken together, these results underscore the potential of ferroptosis-targeted therapies as novel treatment strategies for MM.

Evidence suggests that immune cells with diverse immunosuppressive phenotypes contribute to the formation of the tumor immune microenvironment, thereby promoting immune escape and facilitating myeloma progression51. Consequently, immunotherapy has emerged as an important strategy to counteract the microenvironment and enhance anti-tumor immunity in MM52. This study demonstrated a significant correlation between PRDXs and immune cell infiltration, suggesting their crucial role in immune escape mechanisms. We focused on the immune function of PRDX2. Dong et al.53 indicated that PRDX2 affects T cells in the TME and induces immune escape in lung cancer cells through the HDAC3-mediated regulation of Galectin-9. And Zhang et al.54 similarly found a significant negative correlation between PRDX2 and CD4 + T cells in colon adenocarcinoma (COAD), which is in agreement with our view. Furthermore, within the TME, inflammation plays a pivotal role in malignant cell proliferation, metastasis and survival as well as in angiogenesis and disruption of the immune system55. These processes are mediated mainly by tumor-associated macrophages (TAMs), cytokines (TNF, IL-1 and IL-6), as well as chemokines (CCL2, CCL3, CXCL12, and CXCL13). Conversely, molecules, such as CXCL4, CCL19 and CXCL10, may contribute to recruiting immune cells and enhancing their cytotoxicity against malignant cells56. Our study revealed that PRDX2 expression correlates with the proportion of CD4 + T cells, the CD4 + /CD8 + T cell ratio and the level of IFN-γ, IL-6 and IL-8. IL-6 exhibits multiple biological activities, including the modulation of cellular activity, differentiation, and growth, while concurrently participating in inflammatory processes that influence disease progression. Pellegrino et al.57 demonstrated the chemotactic and proliferative effects of IL-8 on myeloma cell lines and plasma cells isolated from patient bone marrow. Although IFN-γ derived from NK cells may exert anti-tumor effects, it concurrently upregulates inhibitory molecules—including PD-L1—on tumor cells, thereby potentially inducing immunosuppressive effects on immune cells, including NK cells themselves58. In summary, PRDX2 emerges as a critical orchestrator of tumor immune evasion through direct modulation of T cell activity and suppression of CD4 + T cell infiltration. Concurrently, inflammation-driven remodeling of the TME—mediated by TAMs, cytokines, and chemokines—establishes a permissive niche for tumor progression.

The most common assay is the PRDX2 antibody. Furthermore, mass spectrometry (MS) represents an accurate qualitative and quantitative method for PRDX2 identification, with MALDI-TOF allowing the precise molecular weight to be determined. For instance, Manta et al.59 identified the monomer and dimer of PRDX2 using 12.5% SDS-PAGE, followed by MALDI-TOF MS after purifying PRDX2 from fresh blood samples. In recent years, high-performance liquid chromatography (HPLC) has emerged as a faster and more convenient approach for quantifying PRDX260. Combination therapies offer new prospects for clinical application, especially in helping to overcome the limitations of monotherapy and enhance anti-cancer efficacy. PRDX2 has been identified as a direct target of triptolide, which induces ROS-dependent apoptosis and autophagy in gastric cancer cells61. Conoidin A (CONA), a covalent inhibitor of PRDX2, has been shown to increase the intracellular levels of ROS in glioblastoma cells. Furthermore, both menadione and celecoxib, as ROS-inducing agents, can enhance CONA activity62. N-Carbamoyl Alanine (NCA), another PRDX2 inhibitor, is a peptide-like organic compound that undergoes degradation under the acidic gastric pH. To overcome its limitations, NCA-loaded chitosan nanoparticles (NCA-Chit NPs) have been developed and were shown to enhance cell death in colorectal cancer patients with CHK2 deficiency by targeting PRDX263. Overall, PRDX2 can be of clinical value, thus further underscoring its role in disease pathogenesis as well as in the development of more sensitive, efficient detection methods and innovative therapeutic strategies.

Drug resistance in MM may involve intrinsic or extrinsic mechanisms64. Intrinsic mechanisms include genetic and epigenetic alterations, alterations in drug targets, and dysregulation of intracellular signaling pathways mediating processes such as apoptosis65, autophagy, and ferroptosis66. Extrinsic mechanisms are primarily mediated through interactions with cell adhesion to the extracellular matrix, and the tumor immune microenvironment, which activate signal transduction pathways leading to drug resistance51. We therefore propose that PRDX2 may orchestrate a cellular defense program, intrinsically suppressing ferroptosis pathways while extrinsically establishing an immune-evasive sanctuary. Inhibition of PRDX2 is anticipated to resensitize drug-resistant MM cells to proteasome inhibitors (e.g., Bortezomib, Carfilzomib) and immunomodulatory drugs, primarily by attenuating their antioxidant defense and, in particular, potentiating the ferroptosis pathway. PRDX2 inhibition to improve the efficacy of immunotherapies (e.g., Daratumumab, bispecific antibodies) by mitigating the cytokine-driven immunosuppressive milieu (reducing IL-6/IL-8, potentially improving T-cell infiltration/function), thus fostering a more responsive immune microenvironment. We explicitly link this mechanism to clinical observation that PRDX2 overexpression is associated with more aggressive disease features and inferior survival outcomes. Combining PRDX2 inhibitors with existing therapeutics could therefore lead to enhanced efficacy.

The current findings provided valuable insights into the significance of PRDX2 as a potential biomarker in ferroptosis and the immune microenvironment regulation. However, this study has several limitations. Firstly, most analyzed data were derived from publicly available databases. Owing to the diversity of data sources, some inconsistencies were noted across datasets, and thus additional experimental validation would be needed to confirm these results. Secondly, the limited retrospective clinical sample size and variables cannot exclude the possibility of biased results, and a larger number of clinical samples would be needed to assess PRDXs expression in CD138 + MM cells for robust conclusions. Thirdly, further studies are necessary to elucidate the mechanisms linking PRDXs to ferroptosis and the immune microenvironment, which will be the focus of our future work. Finally, further experiments are warranted to evaluate the efficacy of PRDX2 inhibitors in preclinical models and to assess their synergistic anti-tumor effects and capacity to overcome drug resistance when combined with proteasome inhibitors and immunotherapy.

Conclusions

This study provides comprehensive evidence on PRDXs expression and their potential as biological and prognostic biomarkers in MM, positioning their as novel therapeutic targets for inducing non-programmed cell death and overcoming the tumor immune microenvironment challenges. The results demonstrate a strong correlation between PRDX2 and poor clinical outcomes. Notably, significant relationships among PRDX2, ferroptosis and immune cell infiltration were identified. Overall, PRDX2 exerts a pivotal dual role in maintaining redox homeostasis and reshaping the TME, thereby positioning it as a highly attractive therapeutic target to counteract resistance and enhance clinical outcomes.