Introduction
Acute myeloid leukemia (AML) is a heterogeneous and highly aggressive hematologic malignancy [
1]. It is the most prevalent form of acute leukemia in adults, with an annual incidence of 4.3/100,000 and an increasing incidence risk with older age [
2]. Existing standard treatments include intensive chemotherapy and allogeneic hematopoietic stem cell transplantation (HSCT). Since 2017, the U.S. Food and Drug Administration (FDA) has approved eight new drugs for the treatment of AML, including the FLT3 inhibitors midostaurin and gicitinib, the IDH inhibitors efusitinib and enalcitinib, the anti-CD33 monoclonal antibody gemtuzumab ozogamicin, CPX351, the hedgehog pathway inhibitor glasdegib, and BCL-2 inhibitor venetoclax. Targeted therapies are more effective and less toxic than the conventional chemotherapy [
3]. Despite significant advances in these new therapies, primary and secondary resistance remains a major obstacle in treating AML. Therefore, pre-evaluation of the likelihood of resistance is vital for treatment options [
4,
5]. Allogeneic HSCT has become an effective tool for AML cure by preventing AML relapse through high-dose chemotherapy and graft-versus-leukemia effects. However, how to screen the population for transplantation is still a challenge in AML treatment. According to the current cytogenetic risk stratification, patients are classified into three risk groups (favorable group, intermediate group, and adverse group). However, patients classified as favorable or intermediate group still have high relapse and drug resistance rates, suggesting that the current risk stratification system is still inadequate. Therefore, the creation of novel and trustworthy prognostic biomarkers for AML is urgently required.
Cysteine-rich protein 1 (
CSRP1) is a member of the CSRP family. This gene family contains a group of LIM domain proteins, which are proposed to be involved in regulatory processes essential for development and cellular differentiation.
CSRP1 is located on human chromosome 1q32.1 [
6]. The LIM protein, CRP1, is a general marker for smooth muscle lineages [
7]. CRP1 localizes to the nucleus and the cytoplasm with different functions depending on its location. When CRP1 is in the nucleus, it regulates interactions between transcription factors and promotes the upregulation of smooth muscle-specific genes [
8]. When CRP1 is in the cytoplasm, it localizes to the adhesion patch and the actin cytoskeleton to regulate actin filament bundles [
9].
Only a few reports have been made about
CSRP1's connection to cancer thus far.
CSRP1 is associated with poor clinicopathological features in adrenocortical carcinoma (ACC) [
10]. Hepatocellular carcinoma (HCC) causes abnormal methylation to inactivate
CSRP1, which may be a key biomarker for cancer [
11].
CSRP1 was used to forecast when benign prostatic hyperplasia will turn into prostate cancer and may have an effect on disease-free survival [
12]. In addition,
CSRP1 is also associated with colorectal cancer [
13] and breast cancer [
14].
There has been no research to investigate the expression profile and function of CSRP1 in AML. In this study, we examined the database and our cohort to investigate the expression of CSRP1 in AML and its prognostic significance. We also studied the differentially expressed genes related to CSRP1 expression and explored their potential roles in AML through GO and KEGG enrichment, immune infiltration, protein interaction analysis, and drug sensitivity analysis.
Discussion
In this study, we explored the relationship between CSRP1, the clinicopathological features and prognosis of AML. We found that CSRP1 was highly expressed in adult AML, which was associated with a higher proportion of bone marrow blasts, a higher frequency of DNMT3A mutation and a poor prognosis in AML patients. In addition, we constructed a nomogram to predict OS for AML based on age, cytogenetic risk stratification, and CSRP1 expression levels and explored the possible mechanisms of CSRP1 function.
Through pan-cancer analysis, we found that, unlike other tumor markers that are always highly or lowly expressed in different tumors,
CSRP1 shows different expression patterns between different tumors, with
CSRP1 highly expressed in 7 cancers and lowly expressed in 18 tumors. It suggests that
CSRP1 has a complex mechanism of regulation and can function as either an oncogene or oncogene suppressor in different cancer species or under different circumstances. In addition, the expression of
CSRP gene family in AML is inconsistent, with the
CSRP1 highly expressed and
CSRP3 lowly expressed in AML. We have systematically investigated the role of the
CSRP2 gene in AML and found that its low expression is associated with poor prognosis in AML [
15]. At the same time, the knockdown of
CSRP2 promotes proliferation and cycle progression in AML cell lines [
15]. In contrast, no studies on
CSRP1 in AML have been reported. We confirmed the high expression of
CSRP1 in AML by comparing 224 adult AML patients with 23 healthy controls, which is consistent with the database results. All of these suggest that
CSRP1 may play an essential role in AML.
Next, we investigated the relationship between
CSRP1 gene expression, the clinicopathological features, and gene mutations using data from 224 adult AML cases.
CSRP1 expression did not correlate with gender, age, risk stratification, or WBC at diagnosis. The analysis of baseline data showed a higher proportion of bone marrow blasts and a higher frequency of
DNMT3A mutations in the high
CSRP1 group.
DNMT3A is one of the most frequently mutated genes in AML [
31] and is an independent prognostic factor used for risk stratification [
32]
. It may suggest a higher tumor burden and a higher incidence of adverse prognostic mutations in those with high
CSRP1 expression.
We next explored the impact of CSRP1 gene expression levels on overall survival in adult AML patients. We found that high CSRP1 expression was associated with poor OS through multiple adult AML database cohort studies, including TCGA-LAML, Beat-AML, and GEO databases. Moreover, we further validated this result with the ZZU cohort. To further optimize the current stratification system for AML and facilitate clinical application, we first performed a univariate analysis. Initial screening revealed that age > 60 years old, worse cytogenetic risk stratification and high CSRP1 expression were independent poor prognostic factors. The prognostic significance of age and karyotype stratification is well established. To facilitate clinical application, we further developed a nomogram and applied a calibration plot to validate the model. This model performed well for both the TCGA-LAML dataset and the ZZU cohort. This finding is beneficial for further optimizing the stratification system, especially for some patients classified as intermediate risk according to the current stratification, who can now better evaluate transplantation or chemotherapy according to the current guideline recommendation. This stratification system can further score, stratify and guide the treatment selection.
The overexpression of
CSRP1 and its associated poor prognostic value in AML suggest that it may play a role in AML. To further explore the mechanism of action of
CSRP1, we subjected patients with high and low
CSRP1 expression to differential gene expression analysis, followed by GO and KEGG functional enrichment. GO-BP enrichment analysis revealed that
CSRP1 was closely associated with neutrophil function, which was confirmed in subsequent microenvironmental correlation analysis. Upregulating neutrophil elastase (NE) promoted the growth of leukemia cells and decreased the proportion of apoptotic cells [
33]. GO-BP enrichment analysis reveals that
CSRP1 is associated with actin binding and Rho GTPs binding. AML with F
LT3-ITD mutations is characterized by RAC1-dependent actin cytoskeleton remodeling that substantially contributes to the acquisition of resistance to midostaurin in vitro [
34]. Yang et al. [
35] analyzed the expression patterns and prognostic significance of Rho family GTPases in AML and found that
RhoBTB3 was significantly downregulated in AML bone marrow compared to healthy controls and correlated with prognosis of AML. KEGG enrichment analysis revealed that
CSRP1 was associated with cell adhesion molecules, rap1 signaling pathway, HIF-1 signaling pathway, JAK-STAT signaling pathway, and apoptosis. Cellular adhesion molecules also impact the poor prognosis of AML and may be used as targets for AML-specific therapies [
36]. Also, the Rap1 signaling pathway plays a crucial role in cancer [
37]. HIF signaling has been implicated in myeloid cell survival, and PI3K/Akt is known to induce HIF-1 transcription [
38]. The signaling pathway JAK/STAT plays a critical role in the development and progression of AML [
39].
In the PPI analysis, nine out of the 19 hub genes (
CD163,
CX3CR1,
C5AR1,
THSD7A,
ADMATS18,
IL10,
THBS1,
ADAMTS15, and
LILRB2) were correlated with OS in AML. Notably,
CD163 and
THBS1 were among the 20 genes screened for the most significant differences based on
CSRP1 expression levels.
CD163 was a specific marker for macrophages of M2 type and was identified as a potential target for the therapeutic intervention of AML [
40,
41]. The association of
CSRP1 with poor prognosis and macrophages in AML may be acting through
CD163, but this requires further experimental confirmation.
THBS1 is a novel serum prognostic factors of AML [
42].
CX3CR1 was identified as one of the distinctive features of AML cells for universal MRD monitoring [
43]. For AML patients who are ineligible for standard treatment with chemotherapy and HSCT and who also experience less severe side effects, Laura Jimbu et al. propose that manipulation of both the co-inhibitory network (with anti-PD-L1 blocking antibodies) and suppressor network (with anti-IL-10 blocking antibodies) is an appealing immunotherapeutic intervention. To increase the therapeutic effectiveness of treating AML illness, the suggested combination of these two immunotherapies provides a novel strategy that can be easily used in the clinic [
44]. Endothelial cells (ECs)-derived small extracellular vesicles contained a high level of
ANGPTL2, which accelerated leukemia progression via binding to the
LILRB2 receptor [
45]. The other genes, including
C5AR1,
THSD7A,
ADAMTS18, and
ADAMTS15, have not yet been reported in AML-related studies and also require further investigation.
Finally, we also aimed to investigate whether CSRP1 expression is capable of prognosticating the patients’ response to the commonly used chemotherapeutic agents. Our findings revealed that patients with higher CSRP1 expression were more sensitive to certain chemotherapy agents including 5-fluorouracil, gemcitabine, rapamycin, and cisplatin and more resistant to fludarabine. This will help us in the selection of drugs in case of relapse and refractory patients.
In summary, CSRP1 was highly expressed in adult AML, and such high-level expression of CSRP1 was related to a poor prognosis in adult AML. CSRP1 may serve as a potential prognostic marker and a therapeutic target for AML in the future. Further verification is expected to be carried out to reveal the clinical significance and biological impacts of CSRP1 in AML.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.