Introduction
Multiple myeloma (MM) is the second most common hematological malignancy characterized by malignant terminally differentiated plasma cells [
1]. During the last decades, advances in anti-myeloma therapeutics, including proteasome inhibitors, immunomodulatory drugs, and anti-CD38 monoclonal antibodies have considerably improved the treatment outcome in myeloma [
2]. Despite these improvements, most patients still face relapse and/or drug resistance. Thus, exploring a new mechanism of MM and searching for novel targets are urgently needed.
Inflammation predisposes to the development of cancer and promotes all stages of tumorigenesis [
3]. According to early research reported by Lindqvist, a personal history of all infections combined was associated with a significantly increased risk of MM [
4]. At present, a growing number of evidence indicates that inflammation plays a crucial role in the recurrence and drug resistance of MM. Because inflammatory immune cells not just participated in the formation of the tumor microenvironment (TME), but involved in tumor survival and immune modulation [
5,
6]. Recently, single-cell transcriptomic datasets of MM tracking stromal inflammation in individuals over time revealed that successful antitumor induction therapy is unable to revert bone marrow inflammation [
6]. Relapsed/refractory multiple myeloma (RRMM) cells can also shape an immune suppressive bone marrow microenvironment by up-regulation of inflammatory cytokines [
7]. Furthermore, clinical and laboratory research suggested that use of anti-inflammatory agents is a promising approach for cancer prevention and treatment [
8]. Currently, several studies have demonstrated that inflammatory response-related genes (IRRGs) could be used to predict the prognosis of various cancers, such as pancreatic ductal adenocarcinoma, hepatocellular carcinoma and transitional bladder cancer [
9‐
11]. However, the relationship between IRRGs and the prognosis of MM patients remains unclear.
In this study, we collected the transcriptional expression and matched clinical data of patients with MM from Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases. Subsequently, we constructed and validated a risk model with differentially expressed genes (DEGs) related to inflammatory response. Then, the potential mechanism was explored in high-risk group by gene set enrichment analysis (GSEA) and the correlation analysis between IRRGs and infiltration of immune cells was carried out. Finally, we identified CD81 as one of the IRRGs has impacted the immune status extensively and displayed a bad prognosis as well.
Discussion
MM is an incurable and highly heterogeneous malignant tumor, despite recent advances in therapy, approximately 20% of people newly diagnosed have substantially worse outcomes [
12]. An increasing number of improved biomarkers are used for determining the overall prognosis of MM patients, but patients are still managed in a similar manner regardless of individual risk factors or disease characteristics. It means currently there is no sufficient information to routinely utilize predictive biomarkers to select initial treatment for MM or intensify treatment for high-risk MM. With the exception of tumor treatment, people pay close attention to TME. Inflammatory immune cells are an important component. Thus, we hope to explore the prognosis of MM from the perspective of inflammation and provide several evidence for the adjustment of MM therapeutic strategy.
Inflammatory response-related serum biomarkers such as interleukin-1(IL-1), sIL-2R, IL-6, IL-10 and IL-17 have a good performance in predicting prognosis of MM [
13]. In addition, a bidirectional mendelian randomization study indicated that higher genetically determined monocyte-specific chemokine 3, vascular endothelial growth factor, IL-10, and IL-7 were associated with increased risk of MM [
8]. However, the inflammatory response-related gene signature as prognostic marker for MM has not been reported. In our study, three MM datasets from GEO and TCGA were used to identify 24 DE-IRRGs, six of them screened out by LASSO and Cox regression for constructing a predictive model.
Our risk model consisted of VCAM1, RGS1, KIT, CD81, BLNK and BIRC3. VCAM1 is a member of the immunoglobulin super family and is produced by the expression of vascular endothelial cells, which can not only mediate inflammatory effects, but also promote angiogenesis of tumors, and lead to their growth and metastasis [
14,
15]. It is also used by cancer cells to escape immune detection as its expression is up-regulated in multiple cancers, including MM, acute myeloid leukemia, where high expression associates with poor prognosis [
16,
17]. RGS1, as a type of GTPase-activating protein, regulates T cell trafficking to tumors by attenuating chemokine-mediated signals and increased expression of RGS1 was associated with poorer survival of patients with breast cancer [
18]. Meanwhile, reports from Korea and Egypt revealed RGS1 overexpression were associated with lower response rate and inferior OS in MM patients [
19,
20]. KIT is one of the type III receptor tyrosine kinases and is involved in several signaling pathways, including the PI3K pathway, MAPK pathway, etc., responsible for cellular growth and proliferation. In medicinal chemistry and drug discovery, KIT(c-Kit) is considered one of the key targets for the management of various types of cancer, including melanoma, gastrointestinal stromal tumors, small cell lung carcinomas and acute myeloid leukemia [
21]. However, relevant studies on KIT in MM are scarce. CD81 belongs to the tetraspanin family of proteins [
22]. The expression of CD81 in melanoma in humans was shown to promote tumor growth and metastasis [
23], and knockdown of human CD81 in osteosarcoma and breast cancer cells reduced tumor progression and dissemination [
24,
25]. A previous study revealed that CD81 positive expression could potentially contribute to stratify minimal residual disease-positive MM patients after treatment and predicted inferior outcomes [
26]. B cell linker protein (BLNK) is a major downstream effector of the B cell receptor signaling. On phosphorylation, it recruits multiple effectors modulating several signaling pathways, including MAPKs, ERK1/2, JNK, p38 and PLCg2 signaling, contributing to B cell development, maturation, and differentiation [
27]. An experimental study has suggested that BLNK were up-regulated in Waldenström’s macroglobulinemia but not in MM, and it contributed to the regulation of Met receptor signaling in non-small cell lung cancer [
28]. BIRC3 is one of the eight members of the human inhibitors of apoptosis proteins family. Many evidences point to the pro-survival and antiapoptotic role of BIRC3 in cancer cells [
29]. A previous study has indicated BIRC3 inactivation is consistently associated to shorter PFS and poor OS in chronic lymphocytic leukemia patients [
30]. The latest research shows the low expression of BIRC3 is correlated with poor prognosis in MM [
31].
After multiple analysis and verification of multiple datasets, our model has proven patients with high-risk scores not only exhibited significantly advanced and high tumor burden, but also decreased survival rates. Therefore, this risk model could be considered as a tool to stratify patients in risk categories and it is an independent prognostic factor. Moreover, we conducted GSEA analysis based on this risk signature, and noted that the pathway in autophagy was more suppressed in up-regulated genes subgroup of the high-risk group. Autophagy mediates the delivery of various cellular cargoes to lysosomes for degradation and recycling, so that it is a quality-control, metabolic, and innate immunity process [
32]. When autophagy is perturbed, this has repercussions on diseases with inflammatory components, including infections, autoimmunity and cancer. Many studies demonstrate important protective roles for autophagy against disease. However, in cancer, it seems that opposing roles of autophagy are observed in the prevention of early tumor development versus the maintenance and metabolic adaptation of established and metastasizing tumors [
33]. It similarly has a dual role in the autophagy of myeloma cells. Once tumors are formed, tumor cells use autophagy to ensure their survival under nutrient-deficient and hypoxic conditions [
34]. So far autophagy has been also demonstrated to be involved in the induction of MM-drug resistance [
35]. Beside, we also found PI3K-Akt signaling pathway was activated and enriched in up-regulated genes subgroup, which is a conventional inflammatory signaling pathways involved in cancer development [
36]. Intriguingly, enrichment scores related to GO function of cell migration and cell motility, were also elevated in up-regulated genes subgroup, suggesting that higher tumor cell activity and more potential to tumor metastasis were observed in MM patients with high-risk group. However, enrichment scores related to GO function of regulation of apoptotic process and regulation of programmed cell death were decreased. This indicated that the elimination of tumor cells might be inhibited.
A comparison of immune cells types between high-and low-risk MM groups was performed here. The results showed high-risk participants had higher proportions of B cells naïve, T cells follicular helper (Tfh), NK cells resting, monocytes and Eosinophils. These results implied that these immune cells were associated with poor prognosis. Among them, previous study revealed that the Tfh17/Tfh ratio was significantly elevated in newly diagnosed patients and even higher in relapsed patients with MM. In addition, the Tfh17/Tfh ratio was reduced in post-ASCT patients compared to that in non-ASCT patients [
37]. Bone marrow monocytes are primarily committed to osteoclast formation and in the MM condition exhibit dysfunction status [
38,
39]. As to Eosinophils, a study indicated that microbiota-driven interleukin-17- producing cells and eosinophils synergized to accelerate multiple myeloma progression [
40].
In the study, correlation analysis between each signature and 22 immune cells showed that CD81 had a wide influence on the infiltration of various immune cells, especially on the infiltration of monocytes and macrophages M2. Our study showed that initial MM patients with CD81 positive expression by flow cytometry had poorer OS (68.06 months
vs. 100.61 months). However, the role of CD81 in MM has not been elucidated. Furthermore, macrophages M2 can be polarized from ordinary macrophages and it can support cancer progression by several mechanisms including immune suppression, growth factor production, promotion of angiogenesis and tissue remodeling [
41]. Several clinical studies confirmed macrophages M2 were associated with increased microvessel density, chemoresistance and reduced survival, independently of the MM stage [
42]. A recent study pointed out that JNK signaling pathway may be involved in the growth suppression mediated by CD81 overexpression in hepatocellular carcinoma can cell [
43]. Meanwhile, preclinical experiments have indicated upregulating expression of JNK and its downstream transcription c-Myc may induce macrophages toward an M2 phenotype [
44]. But currently, it is unclear whether CD81 affects the prognosis of MM by promoting M2 macrophage polarization and infiltration. Thus, further study is worth to conduct in the future and CD81 might be regarded as a potential target for anti-MM therapy.
For our study, some limitations must be acknowledged. First, since MM is highly heterogeneous, some important clinical variables were not available from the public datasets. For instance, the revised-ISS stage or Mayo Stratification for Myeloma and Risk-Adapted Therapy (mSMART) risk stratification in MM patients were not assessed in our model. Therefore, greater number of clinical variables should therefore be included in future studies. Second, the relevant information of patients involved in this study was not analyzed in vitro samples. Therefore, more experiments in cell and animal model will be performed to elucidate how the gene signatures modulate the outcome of MM. Third, more independent MM cohorts should be used to validate the identified prognostic IRRGs.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.