Protein interacting with C-kinase 1 is correlated to prognosis and immune infiltrates of gastric cancer

Abstract

Background

Protein interacting with C-kinase 1 (PICK1) has been proved to be involved in many malignant tumors, such as neurological tumors, digestive system tumors and breast cancer. However, its biological role in tumor immune microenvironment of gastric cancer (GC) is still unclear.

Methods

Public datasets from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases were acquired for the purpose of examining the relationship between the expression of PICK1 mRNA and clinical characteristics, as well as the survival of patients with GC. The utilization of CIBERSORT and TIMER web servers allowed for the exploration of the association between the expression level of PICK1 mRNA and the level of immune infiltrates in GC tissues. Additionally, the implementation of Gene Set Enrichment Analysis (GSEA) provided evidences for this connection. Finally, correlation analysis was conducted to assess the relationship between the expression of PICK1 mRNA and specific classical immune cell markers.

Results

We discovered that decreased PICK1 mRNA expression in GC tissues indicated a poor TNM stage and shorter overall survival duration. In addition, expression level of PICK1 mRNA was demonstrated to be relevant to the relative levels of immune infiltrates in GC tissues, especially the macrophages. Furthermore, our findings demonstrated that PICK1 mRNA was adversely linked with M2 macrophage markers.

Conclusion

The decreased PICK1 expression was believed to benefit the infiltration of macrophages and the polarization of M2 macrophages, and finally lead to an unfavorable prognosis for GC patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-025-03783-7.

Introduction

Cancer is a life-threatening disease that poses a serious risk to human health and survival. Tumor genesis involves remarkably complex mechanisms, with studies demonstrating a close link to oncogene activation and tumor suppressor gene inactivation [1]. Traditional cancer treatments include surgery, radiation therapy, and chemotherapy. With deeper investigation into cancer biology, numerous emerging therapies have gradually been developed, such as targeted therapy, immunotherapy, and gene therapy [1].

In decades, gastric cancer (GC) has been recognized as one of the most commonly diagnosed malignant diseases. The latest cancer statistics of the United States shows that the estimated incidence and mortality rates of GC have declined compared with the previous rankings [2]. However, according to the latest cancer burden statistics of China, GC remains the third leading cause of cancer mortality both for males and females [3]. Despite significant advancements in surgical resection and the use of recently developed immunosuppressive medication in patients with gastric cancer, the 5-year overall survival rate remains suboptimal [4]. It’s not sufficient to only concentrate on improving therapeutic approaches, and identifying significant early diagnostic or prognostic biomarkers is in urgent need. Uncovering novel underlying molecular functions is of great interest as the carcinogenesis and progression of gastric cancer are usually accompanied by the dysregulated expressions of genes [5].

The protein interacting with C-kinase 1 (PICK1) is well-known since it is currently the only known protein that has both the PSD-95/DlgA/ZO-1 (PDZ) domain and the Bin-Amphiphysin-Rvs (BAR) domain [6]. These two domains allow PICK1 to bind to numerous membrane proteins and lipid molecules, therefore mediating their trafficking and location [7]. According to current studies, PICK1 is largely involved in neurological disorders [8–10] and male infertility [11–13] as well as other diseases. Recently, the PICK1 expression is found to be dysregulated in neoplastic diseases and benefit the invasion and metastasis of tumor cells. However, the expression pattern and function of PICK1 are different or even reversed in different tumor types. It is found that expression of PICK1is significantly reduced in astrocytic tumor cell lines as well as clinical astrocytic tumor samples, and is negatively correlative to the invasive ability of tumor cells [14]. On the contrary, PICK1 expression shows an upward trend in breast cancer cells and higher level of PICK1 expression indicates an unfavorable prognosis in breast cancer patients [15]. Our previous study showed that expression of PICK1 mRNA and protein were downregulated in GC, and PICK1 has the potential to be a molecular biomarker for prognosis of GC [16]. However, the role of PICK1 in the tumor immune microenvironment of GC still remains largely unclear.

Currently, the relationship between expression of PICK1 and clinicopathological parameters, as well as overall survival in GC were explored by analyzing the sequencing data from TCGA and transcriptome dataset from the GEO database. We also investigated the correlation between PICK1 expression and levels of tumor-infiltrating immune cells. This study suggested that the downregulation of PICK1 expression could relate to poor outcome for GC patients and may facilitate the polarization of M2 macrophage, which has been indicated to promote tumor metastasis and thus induce progression of GC [17].

Methods

Public data collection

The sequencing data as well as the clinical data of GC samples in TCGA database were obtained from the “DATASETS” module on UCSC Xena browser (https://xenabrowser.net/) [18]. We also selected one chip of transcriptome gene expression dataset (GSE15459) containing 200 GC tissues from GEO database (https://www.ncbi.nlm.nih.gov/geo/). Gene expression data and clinical information were both downloaded. Eight samples in GSE15459 were excluded as they failed quality control or were proven not to be GC. “Affy” R package (version: 1.66.0) was used to process the raw data files of GSE15459 and then we obtained the normalized gene expression matrix file.

Patients were divided into high and low expression groups by the median value of PICK1 mRNA. Then, the relationship between PICK1 mRNA expression and clinicalpathological parameters (age, sex, tumor location, G/T/N/M stage, pathological TNM stage, recurrence) was analyzed.

Prognosis and immune infiltration analysis

After excluding those patients lacking survival data, all of the remaining GC patients from TCGA and GSE15459 were separated into two expression groups using the method described above respectively, and then we investigated the prognostic significance of PICK1 expression in GC.

Cell Identification by Estimating Relative Subsets of RNA Transcripts (CIBERSORT) employs linear support vector regression to compare RNA sequencing (or microarray) data from mixed tissue samples against a known immune cell gene expression signature matrix, thereby estimating the relative proportions of 22 immune cell subtypes within the sample [19]. The CIBERSORT web server (https://cibersort.stanford.edu/) was employed to quantify the relative infiltration levels of 22 different types of immune cells using gene expression data of GC from TCGA and GSE15459. Then the correlation was identified between PICK1 expression and the relative levels of those 22 immune cell types.

TIMER applies a deconvolution method to analyze RNA-seq data of cancer samples in the TCGA database to estimate the infiltration levels of six immune cell types (B cell, CD4⁺ T cell, CD8⁺ T cell, neutrophil, macrophage, and dendritic cell) within the tumor microenvironment [20]. Using the “Gene” module of TIMER web server (https://cistrome.shinyapps.io/timer/), we investigated the association between PICK1 mRNA expression and infiltration proportions of these six types of immune cells in GC. We also explored the prognostic significance of the immune-infiltration levels through the “Survival” module of TIMER. Moreover, the “Gene” module of TIMER2.0 vision [21] was employed to examine the relationship between PICK1 expression and infiltration of subtypes of the immune cells.

Infiltration and polarization of macrophages

The Molecular Signatures Database (MSigDB) of the GSEA web server (https://www.gsea-msigdb.org/gsea/index.jsp) was searched for the gene sets relevant to macrophages. The reliability of GSEA results is ensured through three key steps: calculation of the enrichment score (ES), estimation of the statistical significance of the ES, and adjustment for multiple hypothesis testing [22]. Using the approach mentioned above, two expression groups were created for the GC samples in the TCGA and GSE15459. Then we performed GSEA on the two groups according to the default settings using the macrophage-related gene sets [22, 23]. It was deemed statistically significant at |NES|≥1.0, NOM P-val < 0.05, and FDR q-val < 0.25. Additionally, the correlations between expression of PICK1 mRNA and classical biomarkers of M1 macrophage, M2 macrophage, and tumor-associated macrophage (TAM) were evaluated with the gene expression profile of GC tissues from TCGA and GEO.

Statistical methods

Statistical analysis was conducted on R software (version 4.3.0). The Chi-square (χ²) statistic was conducted to explore the relationship of PICK1 and clinical features of GC patients. Pearson correlation tests (r) were conducted to explore the correlations of gene expression. Survival analysis was evaluated using the Kaplan-Meier method and statistical difference was examined by the Log-rank test. P < 0.05 was considered as statistically significant.

Results

PICK1 expression is correlated with clinicopathological parameters

The results showed that low PICK1 mRNA expression was associated with poor differentiation degree (P = 0.005) and pathological TNM stage (P = 0.004) for GC patients, but not with sex, age, tumor location, T stage, N stage, M stage, or recurrence state in the TCGA database (Table 1). In accordance, a similar result was obtained by the same analysis of a GC dataset from the GEO database. It showed that lower expression level of PICK1 mRNA only indicated a poorer TNM stage, while had no connection with sex, age, differentiation degree, T stage, and N stage in GC patients (Table S1). Taken together, results from both TCGA and GEO databases confirmed that down-expression of PICK1mRNA was correlated with poor TNM stage in GC.

Table 1.

The correlation between PICK1 mRNA expression and clinicopathological parameters of GC patients in TCGA

Parameter	Total	PICK1 expression		P-value
		Low (N, %)	High (N, %)
Sex
Man	241	118 (49.0)	123 (51.0)	0.455
Woman	134	71 (53.0)	63 (47.0)
Agea
≤ 67	187	102 (54.5)	85 (45.5)	0.108
> 67	184	85 (46.2)	99 (53.8)
Tumor locationa
Cardia/gastroesophageal Junction	88	38 (43.2)	50 (56.8)	0.322
Fundus/body	131	69 (52.7)	62 (47.3)
Antrum	138	72 (52.2)	66 (47.8)
G stagea
Well/medium	147	60 (40.8)	87 (59.2)	0.005
Poor	219	122 (55.7)	97 (44.3)
T stagea
T1 + T2	99	4 5(45.5)	54 (54.5)	0.335
T3 + T4	268	137 (51.1)	131 (48.9)
N stagea
N0	111	50 (45.0)	61 (55.0)	0.236
N1 + 2 + 3	247	128 (51.8)	119 (48.2)
M stagea
M0	330	162 (49.1)	168 (50.9)	0.293
M1	25	15 (60.0)	10 (40.0)
pTNMa
I + II	171	72 (42.1)	99 (57.9)	0.004
III + IV	190	109 (57.4)	81 (42.6)
Recurrencea
Yes	244	123 (50.4)	121 (49.6)	0.948
No	86	43 (50.0)	43 (50.0)

^aThe missing values were removed from the analysis

Downregulated PICK1 expression predicts poorer prognosis in GC patients

The remaining 368 samples were divided into low (N = 185) and high (N = 183) PICK1 mRNA groups, and survival analysis was carried out after eliminating those GC samples without survival data in the TCGA database. The result showed that individuals with GC had shorter overall survival time when PICK1 mRNA expression was decreased (Fig. S1). A similar result was seen in the survival analysis of a GC dataset from GEO (Fig. S1). Overall, there was an unfavorable correlation between decreased PICK1 expression and the clinical outcome of GC patients.

PICK1 expression is relevant to immune infiltration level

Through The CIBERSORT online server, we were able to determine the relative proportions of immune infiltration in GC tissues. In a result, PICK1 expression was found to be linked with infiltration levels of M0 macrophages (P = 0.011), M2 macrophages (P = 0.035), resting dendritic cells (P = 0.032), resting mast cells (P = 0.002), and eosinophils (P = 0.003) (Fig. 1) in GC samples from TCGA. Similarly, by the same analysis using the trancriptomes data of GC from GEO, PICK1 expression was also shown to be correlated with infiltration levels of immune cells, including memory B cells (P = 0.012), plasma cells (P = 0.025), gamma delta T cells (P = 0.04), resting NK cells (P < 0.001), M2 macrophages (P = 0.001), and resting mast cells (P = 0.001) (Fig. S2). These findings revealed a relationship between PICK1 expression and the immune cell infiltration in GC.

Fig. 1.

PICK1 mRNA expression is correlated with the infiltration levels of immune cells in GC revealed by TCGA. * P < 0.05, ** P < 0.01

PICK1 acts through the activation of macrophages in GC

Through the CIBERSORT online server, T cells and macrophages were discovered to be the two most prevalent types of tumor-infiltrating immune cells in GC tissues (Fig. 2A). At the same time, we looked at the correlation between PICK1 expression and infiltration levels of immune cells on the TIMER web server. Our findings demonstrated that PICK1 expression was negatively correlated with the infiltration levels of CD8⁺ T cells (r = − 0.13, P = 0.012), neutrophils (r = − 0.127, P = 0.0147) and dendritic cells (r = − 0.186, P = 3.20 × 10^− 4), and was especially significantly negatively correlated with macrophages (r = − 0.3, P = 3.78 × 10^− 9) (Fig. 2B), while had no correlation with B cells and CD4⁺ T cells in GC tissues. What’s more, survival analysis indicated that only the macrophage infiltration level was significantly related to the overall survival for patients with GC (P = 0.004), whereas the infiltration proportions of the other five types of immune cells did not have any prognosis significance (Fig. 2C). The aforementioned findings led us to believe that downregulation of PICK1 mRNA may trigger macrophage infiltration and finally negatively impact the clinical outcome of GC patients.

Fig. 2.

PICK1 mRNA expression is associated with the infiltration of macrophages which is correlated with the prognosis of GC patients. A T cells and macrophages were the two mostly infiltrated immune cell types in GC tissues. B PICK1 mRNA level was significantly negatively correlated with the infiltration number of macrophages. C Macrophage was the only immune cell type which was connected with the prognosis of GC patients

We performed GSEA between the two PICK1 expression groups to get more proof of a possible association between PICK1 and macrophages. Firstly, a total of 19 gene sets related to macrophages were downloaded from the MSigDB database. Next, GSEA was carried out on GC datasets obtained from TCGA and GEO databases, respectively. As a consequence, several macrophage-related gene sets which were important for the activation, differentiation, or cytokine production of macrophages were highly enriched in the low PICK1 expression group, whereas none were enriched in the high PICK1 mRNA group (Fig. 3A, B). Therefore, we supposed that there could be an association between PICK1 expression and macrophage activation in GC tissues.

Fig. 3.

PICK1 may function through macrophages in GC. Macrophage activation or differentiation-related gene sets were enriched in the lower PICK1 mRNA group revealed by TCGA (A) and GSE15459 (B)

Correlation analysis between PICK1 expression and markers of macrophages

Through the TIMER2.0 vision web server, we found that PICK1 mRNA was negatively connected with infiltration level of M2 macrophages, while positively correlated with M0 macrophages and M1 macrophages (Fig. 4A). Later, we explored the relationship between PICK1 expression and biomarkers of macrophages, and discovered that PICK1 mRNA was moderately negatively correlated with three classical markers of M2 macrophages (MS4A4A, CD163, and MRC1), and weakly negatively correlated with two of three TAM markers (IL10, CCL2, and CD68), while weakly negative connected with one of three markers of M1 macrophages (ARG2, PTGS2, and NOS2) in TCGA (Fig. 4B). Furthermore, similar results were got from the examination of the GC dataset from GEO database (Fig. S3). These findings suggested that downregulation of PICK1 expression in GC tissues may benefit the polarization of M2 macrophages and play a certain role in a biological process, during which macrophages transform into tumor-associated macrophages (TAMs).

Fig. 4.

The correlation between PICK1 mRNA and markers of subtypes of macrophages in TCGA. A PICK1 mRNA expression was negatively correlated with the infiltration level of M2 macrophages. B The correlation between PICK1 mRNA expression and markers of M1 macrophage, M2 macrophage, and tumor-associated macrophage. STAD: stomach adenocarcinoma

Discussion

As a unique protein composed of both PDZ and BAR domains, PICK1 is demonstrated to role as an adaptor which binds to a wide range of proteins, and is found to be involved in many diseases including cancers [6]. In our previous study, PICK1 expression showed a downregulated trend in GC tissues and was adverse to the prognosis of patients with GC [16]. Correspondingly, we demonstrated in this study that decreased PICK1 mRNA expression was correlated with a worse pathological TNM stage and a worse survival for GC patients. It demonstrated that the downregulated expression of PICK1 could facilitate the tumor progression and indicate an unfavorable prognosis in GC. Similarly, a prior study had proved that PICK1 expression was downregulated in astrocytic tumor and was correlated with an enhanced invasive ability [14]. And another research showed that downregulation of PICK1 expression may facilitate the bone metastasis in prostate cancer [24]. Taken together, PICK1 may serve as a tumor suppressor in certain tumor types including GC.

Tumor-infiltrating immune cells (TIICs) are considered as a key component of the tumor microenvironment, and have been proved to be closely linked with the tumorigenesis, invasion, and outcome of cancers [25]. Accumulating investigations have shown that the immune cell is a double-edged sword which could exert anti-tumor immunity or promote cancer cell survival [26]. Aberrant infiltration levels of immune cells may facilitate tumor progression by triggering tumor immune escape [27, 28]. It showed that there were differences in the relative infiltration degrees of immune cells in GC tissues as compared to control samples and among different tumor stages [26]. Furthermore, it also demonstrated that enhanced infiltration of M2 macrophages, resting dendritic cells and monocytes were correlated with shorter overall survival time, while increased infiltration degree of CD8⁺ T cells indicated better overall survival [26]. Our results revealed that among all of the infiltrated immune cells in GC tissues, T cells ranked first, followed by macrophages. Through TIMER web server, we also found that, out of the six main TIICs types, only the infiltration level of macrophages had a significant impact on the prognosis of GC patients. The higher the infiltration level of macrophages, the more negative the outcome for GC patients. Similarly, a prior study had indicated that increased macrophages demonstrated poor prognosis for GC patients [29]. It was proved that the higher degree of macrophage infiltration in GC tissues may result in an elevated expression of CD44 by inhibiting miR-328, which ultimately results in the advancement of tumors [30].

Specifically, we detected that PICK1 mRNA was negatively correlated with levels of CD8⁺ T cells, neutrophils as well as dendritics, and was most significantly negatively associated with the infiltration level of macrophages. Therefore, we predicted that the downregulation of PICK1 mRNA expression may cause an increased infiltration level of macrophages and finally lead to poor clinical outcome for GC patients. For obtaining additional evidence regarding the relationship between PICK1 expression and macrophages, we performed GSEA using transcriptome data of GC from TCGA and GEO databases, respectively. And the results showed that macrophage activation or differentiation related biological processes were significantly enriched in the group with low expression level of PICK1. Taken together, we believed that PICK1 expression may have a certain regulation relationship with macrophages in GC.

As we all know, macrophages are classified into M1 and M2 phenotypes according to their biological functions [31]. M1 macrophages are usually considered to have significant effect in anti-inflammatory response and antitumor immune response, while M2 macrophages are confirmed to possess the capacity to contribute to tumor progression [32, 33]. The CIBERSORT web server makes it possible for us to further explore the relationships between PICK1 expression and 22 TIICs types in GC. The findings demonstrated that there may be certain relationship between PICK1 mRNA and levels of certain types of immune cells. Particularly, PICK1 expression was found to be negatively associated with the infiltration degree of M2 macrophages in GC both in TCGA and GEO databases. Similarly, result from the TIMER web server also evidenced the negative relationship between PICK1 mRNA and M2 macrophages. It suggested that the infiltration level of M2 macrophages was negatively connected with the prognosis of GC patients, while M1 macrophages infiltration was beneficial to a longer survival time [34]. These results suggested that the decreased expression of PICK1 mRNA may be associated with the increased infiltration level of M2 macrophages, and then promoted tumor progression.

Tumor-associated macrophages (TAMs) refer to those macrophages which appear at the tumor micro-environment, and they mostly display an M2 phenotype [35]. To study the connection between PICK1 expression and phenotypes of macrophages in more depth, we analyzed the correlation between PICK1 mRNA expression and markers of M1 macrophage, M2 macrophage and tumor-associated macrophage (TAM). We found that PICK1 expression had moderate connections with markers of M2 macrophage, while weak correlations with gene markers of M1 macrophage and TAM markers. This finding revealed that decreased expression of PICK1 may fascinate the polarization of M0 macrophages to M2 macrophages, which could be further transformed into TAMs. The infiltration number of TAMs had been proved to be a negative prognostic factor for GC patients [36, 37]. Furthermore, it also showed that TAMs infiltration was associated with the epithelial–mesenchymal transition and was an independent prognostic factor for GC patients [38]. Those results suggested that the decreased expression of PICK1 may promote macrophages differentiate into TAMs, which may benefit the progression of GC.

There is a growing literature on the prognostic value of tumor-infiltrating immune cells in cancer [39]. Consistent with this, our study revealed that PICK1 expression levels exhibit a certain correlation with tumor immune cell infiltration in GC and are associated with patient prognosis. With deeper insights into tumor biology, future cancer therapies are likely to focus more on personalized and precision strategies, particularly targeting the tumor immune microenvironment and molecular mechanisms [40]. Our findings demonstrate that PICK1 may regulate macrophage polarization in GC, suggesting its potential as a novel molecular target for GC immunotherapy.

The findings of this study are all based on analyses of the TCGA and GEO databases. Theoretically, further experimental validation, such as cell line assays, clinical tissue validation and animal studies, should be conducted to further support our findings [41].

Conclusion

Our results demonstrated that downregulated expression of PICK1 predicted a poor prognosis for GC patients. PICK1 mRNA was correlated with the infiltration degrees of certain immune cell subtypes, especially the M2 macrophages, and decreased expression of PICK1 may promote the polarization of M2 macrophages. Above all, PICK1 may serve as a potential prognostic and therapeutic target which may regulate the immune cell infiltration of the tumor microenvironment in GC. However, further clinical trials and basic experiments are needed to identify our findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Ying Zhou and Yaoqiang Du contributed to the study conception and design. Ying Zhou and Biqin Zhang analyzed the data. Feng Li and Xiaohong Li contributed to the figure drawing. The first draft of the manuscript was written by Ying Zhou and Yutao Zhang. Yaoqiang Du and Biqin Zhang reviewed and modified the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by National Natural Science Foundation of China (Grant No. 82202605), Zhejiang Provincial Medical and Health Technology Project of China (Grant No. 2024KY020) and Zhejiang Provincial Special Support Program for Cultivation of High-Level Innovative Health Talents of China (Grant No. 2023, DU YAOQIANG ).

Data availability

The datasets used during the current study can be obtained from UCSC Xena browser (https://xenabrowser.net/) and GEO (https://www.ncbi.nlm.nih.gov/geo/).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.