Identification of common genetic features and pathways involved in pulmonary lymphangioleiomyomatosis and ER-positive breast cancer

Abstract

Accumulating evidence suggests that patients with pulmonary lymphangioleiomyomatosis (PLAM) have a markedly higher prevalence of breast cancer (BC) than the general population. However, the underlying pathophysiological mechanisms remain unclear. Therefore, in this study, we employed a bioinformatics approach to understand the association between PLAM and estrogen receptor (ER)-positive BC. The PLAM (GSE12027) and ER-positive BC (GSE42568, GSE29044, and GSE29431) datasets were obtained from the Gene Expression Omnibus database, and GEO2R was used to identify common differentially expressed genes (DEGs) between them. Functional annotation was performed, and a protein–protein interaction (PPI) network was constructed. Hub genes were identified and verified using western blotting and immunohistochemistry. We conducted an immune infiltration analysis; based on the results, selected 102 common DEGs for follow-up analysis. Functional analyses revealed that the DEGs were mostly enriched in cell proliferation, gene expression regulation, and tumor-related pathways. Four hub genes—ESR1, IL6, PLA2G4A, and CAV1—were further analyzed, and CAV1 was revealed to be associated with clinical outcomes and immune infiltration in ER-positive BC. This study proposes a common, possible pathogenesis of PLAM and ER-positive BC. These common pathways and pivotal genes may provide new directions for further mechanistic studies.

1. Introduction

Pulmonary lymphangioleiomyomatosis (PLAM) is a low-grade malignant metastatic neoplastic lung disease.^[1] This disease is characterized by diffuse cystic lung destruction, resulting in symptoms, such as dyspnea, wheezing, coughing, and recurrent pneumothorax, and in severe cases, respiratory failure can lead to death.^[2] PLAM is classified into 2 types: tuberous sclerosis-associated PLAM (TSC-LAM), caused by germline mutation in TSC2/TSC1 genes, and sporadic PLAM (S-LAM), triggered by somatic mutations in TSC2 gene.^[1,2] The prevalence of S-LAM is approximately 1/0.1 to 0.4 million, and that of TSC-LAM is 1/0.025 million, of which 30 to 80% have combined PLAM.^[3,4] Currently, no cure exists for PLAM; rapamycin is the only drug that can stabilize the disease; however, the condition continues to progress after cessation of the drug and requires lung transplantation in the final stages.^[2] Hence, exploring the pathogenic features and mechanisms of PLAM to identify other therapeutic targets is necessary.

The onset of PLAM has a strong sex bias and occurs almost exclusively in women. Notably, PLAM cells were positive for estrogen and progesterone receptor (ER and PR) expression^.[1] Research has shown that estrogen promotes the growth and metastasis of TSC2-deficient cells to the lungs.^[5] Moreover, pregnancy and exogenous estrogen have been confirmed to similarly promote disease progression in patients with PLAM.^[2] Although the exact mechanism by which PLAM predominates in women is not fully understood, estrogen evidently plays a vital role in the progression of the disease.

Breast cancer (BC) accounts for approximately 30% of all cancers in women, and approximately 73% of patients with BC are ER-positive.^[6,7] Previous studies have found that the standardized incidence ratio of invasive BC in patients with PLAM is higher than that in women without PLAM (4.88 > 2.10); immunohistochemistry showed that all patients were ER-positive.^[8] In addition, LAM lesions express molecules that promote the metastasis of BC cells to the lungs.^[9] BC patients with decreased expression of hamartin and tuberin proteins (encoded by TSC1 and TSC2, respectively) had a worse prognosis than control^.[10] These findings indicate that PLAM and ER-positive BC are similar at the clinical, pathological, and molecular levels. Furthermore, cases of PLAM combined with ER-positive BC have been reported.^[11] Although an intricate relationship between PLAM and ER-positive BC has been recognized, the specific molecular crosstalk remains unclear. Therefore, further investigation is necessary into the possible pathogenesis of PLAM combined with ER-positive BC and the potential treatment options.

Gene expression profiling and bioinformatics analysis based on microarray data are being increasingly used to screen genetic variations at the genome level.^[12] This study aimed to analyze the common differentially expressed genes (DEGs) between PLAM and ER-positive BC and their associated biological mechanisms. Our results showed 102 differential genes in the PLAM and ER-positive BC dataset, which are mainly related to biological functions, such as cell proliferation, differentiation, migration, hormone response, and cancer-related pathways, which may be a common mechanism between PLAM and ER-positive BC. In addition, 4 of these 102 differential genes, ESR1, IL6, PLA2G4A, and CAV1, were selected as pivotal genes, among which the Caveolin 1 (CAV1) gene is particularly noteworthy as it may play an important role in the development of PLAM combined with ER-positive BC. The research workflow is shown in Figure 1.

Figure 1.

Methodological workflow of the present study.

2. Material and methods

2.1. Collection of datasets

The GSE12027, GSE42568, GSE29044, and GSE29431 datasets were downloaded from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/).^[13] The GSE12027 dataset contains 14 human PLAM samples and 7 controls. Three ER-positive BC datasets were collected: the GSE42568 dataset containing 67 human ER-positive BC samples and 17 controls, GSE29044 dataset containing 47 human ER-positive BC samples and 36 controls and GSE29431 dataset containing 33 human ER-positive BC samples and 12 controls.

2.2. Identification of DEGs

Gene expression differences between PLAM and ER-positive BC samples were compared and analyzed using GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/). The DEGs with |log2FoldChange (log2FC)| >1.0 and adjusted P values < .05 were selected from the upregulated or downregulated expression of genes. The Venn tool was used to identify overlapping DEGs.

2.3. DEGs enrichment analysis

Kyoto encyclopedia of genes and genomes (KEGG) and Gene Ontology (GO) enrichment analysis of DEGs was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/).^[14] KEGG is widely used to analyze information on gene related functions, biological pathways, diseases, etc. GO^[15] allows the annotation and resolution of gene functions, including molecular function, biological processes, and cellular components. GO terms and KEGG pathways with P values < .05 were selected. Visualization was performed using an online website (http://www.bioinformatics.com.cn/) and Microsoft Excel.

2.4. Construction of protein–protein interaction network

A protein–protein interaction (PPI) network was constructed using the Search Tool for Retrieval of Interacting Genes/Proteins (STRING) (https://string-db.org/) to predict the functional association between the DEGs.^[16] Cytoscape 3.9.1 was used to present the results. Hub genes were identified using CytoNCA. The GeneMANIA algorithm (https://genemania.org/) was used to construct a hub gene co-expression network.^[17]

2.5. Prediction analysis

Receiver operating characteristic curves for the 3 ER-positive BC datasets were constructed, and the area under the curve (AUC) for the corresponding hub genes in the datasets was calculated to evaluate the diagnostic performance of the hub genes. The closer the AUC value of the hub gene was to 1, the better it was for disease prognosis. The Kaplan–Meier Plotter (https://kmplot.com/analysis/) was used for survival analysis.

2.6. Immune infiltration analysis

Differential alterations in 22 immune cell subtypes of ER-positive BC and PLAM samples were assessed using the CIBERSORT package.^[18] The relationship between hub genes and immune cells in ER-positive BC was analyzed using TIMER2 (TIMER2.0 (cistrome.org)).

2.7. Cell culture

MCF-10A and MCF-7 human BC cells were provided by Cheng Yan (The Second Xiangya Hospital of Central South University). TSC2-/- and TSC2+/+ mouse embryonic fibroblasts (MEF) cells were provided by Zhang Hongbing (Peking Union Medical College). The MCF-7 and MEF cells were maintained in Dulbecco Modified Eagle Medium (Gibco, America) supplemented with 10% fetal bovine serum (VivaCell, China, C04001-050). MCF-10A cells were cultured in Mammary Epithelial Cell Growth Medium (Lonza, Switzerland, CC4136). All cells were cultured at 37°C with 5% CO₂ and subcultured at 70% to 80% growth density.

2.8. Western blot analysis

A freshly configured cell lysate containing a protease inhibitor cocktail (Beyotime, Changsha, China) and radioimmunoprecipitation assay (RIPA) buffer (1:100) (Beyotime, Changsha, China) was used to lyse the cells. The extracted proteins were denatured in a water bath at 100°C for 5 minutes. The proteins were analyzed using sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and blocked with 5% skimmed milk for 1 hour, then incubated overnight with primary antibody raised against CAV1 (Abcam, ab32577, 1:3000) and GAPDH (Proteintech, 10494-1-AP, 1:10,000) at 4°C. The proteins were then incubated with HRP-conjugated antibody for 1 hour at 25°C. The results were analyzed using the Image Lab software.

2.9. Immunohistochemistry

This study was approved by the Ethics Committee of The Second Xiangya Hospital of Central South University. Paraffin-embedded BC tissues and paired paracancerous tissue samples were collected from patients with PLAM having ER-positive BC at the Pathology Department of The Second Xiangya Hospital. An immunohistochemistry assay was performed as described in a previous study.^[19] Formalin-fixed, paraffin-embedded tissue sections (5-μm-thick) were prepared. Antigens were extracted from these sections by boiling in sodium citrate buffer, and endogenous HRP-activity was blocked with 3% H₂O₂. nonspecific staining of the sections was blocked, followed by overnight incubation with anti-CAV1 antibody at 4°C. Finally, the samples were incubated with the secondary antibody (anti-GAPDH) at room temperature for 30 minutes, stained with the 3,3-diaminobenzidine (DAB) substrate, and counterstained with hematoxylin.

3. Results

3.1. Common DEGs in PLAM and ER-positive BC

In total, 4124 DEGs were obtained from the GSE12027 dataset, including 2835 upregulated and 1289 downregulated genes (Fig. 2A). In addition, 784 DEGs were identified from the 3 ER-positive BC datasets, including 288 co-upregulated and 496 co-downregulated genes (Fig. 2B and 2E). DEGs that overlapped in the PLAM- and ER-positive BC datasets were recorded; 55 upregulated and 47 downregulated genes were identified (Fig 2F, Supplemental Digital Content Table S1, http://links.lww.com/MD/J649).

Figure 2.

Volcano plot and Venn diagram of DEGs obtained from patients with PLAM and ER-positive BC. (A) DEGs among PLAM versus control. (B–D) DEGs among ER-positive BC versus control. (E) Upregulated and downregulated genes among ER-positive BC versus control. (F) Upregulated and downregulated genes among the intersection of PLAM and ER-positive BC. BC = breast cancer, DEGs = differentially expressed genes, PLAM = pulmonary lymphangioleiomyomatosis.

3.2. Enrichment analysis of DEGs

For GO and KEGG enrichment analysis, the top 5 significantly enriched terms were selected using a threshold of P < .05. GO functional analysis showed that the dominant enrichment of DEGs in biological processes mainly consists of regulating gene expression, cell proliferation, wound healing, cell differentiation, and response to estrogen (Fig 3A, Table 1). Cellular components were mainly associated with the plasma membrane, cell surface, axons, dendrites, and extracellular exosomes (Fig 3B, Table 1). Regarding molecular function, DEGs were mainly enriched for heparin binding, identical protein binding, calcium-ion binding and bridging, and protein binding (Fig 3C, Table 1). The KEGG analysis results revealed that the top 5 important pathways of DEGs were enriched in cancer, mitogen-activated protein kinase (MAPK) and gonadotropin-releasing hormone (GnRH) signaling pathways, proteoglycans in cancer, and cytokine–cytokine receptor interactions (Fig 3D, Table 2).

Figure 3.

GO and KEGG enrichment analysis of the overlapping DEGs. (A–C) GO analysis of down- and upregulated DEGs. (D) KEGG pathway enrichment analysis of downregulated and upregulated DEGs. DEGs: differentially expressed genes. GO = Gene Ontology, KEGG = kyoto encyclopedia of genes and genomes.

Table 1.

GO enrichment analysis of DEGs between PLAM and ER-positive BC.

Category	Term	Description	Genes	P value
	GO:0010628	Positive regulation of gene expression	RET, IL33, PID1, ANGPT1, CAV1, LEF1, FGF2, BMP2, IL6, ERBB3, ID1, AGR2, VIM, EZR	1.63E-06
	GO:0042493	Response to drug	VAV3, RET, SFRP1, ANXA1, STAT1, MAP1B, MYO6, ABAT, GATA3, TGFBR2, S100A10	1.98E-06
	GO:0042060	Wound healing	IL33, ERBB3, SDC1, CELSR1, FGF2, TGFBR2, DCBLD2	1.25E-05
GO biological process	GO:0008284	Positive regulation of cell proliferation	ARNT2, LEF1, FGF2, PRLR, TGFBR2, CXCL10, BMP2, IL6, SFRP1, ERBB3, CLDN7, CD248, SOX4	1.79E-05
	GO:0043627	Response to estrogen	MME, CAV1, GATA3, ESR1, TGFBR2	4.32E-04
	GO:0042802	Identical protein binding	CFH, ABAT, GATA3, FGF2, RASGRP1, ERBB3, MYO6, SLAMF8, AGR2, AOX1, IGFBP6, WLS, ANXA1, ANGPT1, STAT1, CAV1, ESR1, SFRP1, ID1, CLDN7, TFF3, SDC1, VIM, EZR, KCNK1	3.54E-06
	GO:0005509	Calcium ion binding	RET, TPD52, ANXA1, ANXA3, PLA2G4A, CELSR1, RASGRP1, FSTL1, SGCE, EFEMP1, CD248, S100A14, ADGRL4, S100A10	1.09E-04
	GO:0008201	Heparin binding	CXCL10, SFRP1, PCOLCE2, CFH, COL11A1, FGF2, FSTL1	2.80E-04
GO molecular function	GO:0005515	Protein binding	RET, CNTNAP2, SRPX, PID1, MT1X, FGF2, LAPTM4B, EFEMP1, SOX4, WLS, VAV3, TPD52, ANXA1, CERS6, PTGIS, NUP210, MME, BGN, RRAS2, EMP1, KNL1, PRLR, TGFBR2, SFRP1, MAP1B, FEZ1, CLDN7, TFF3, EZR, TNFRSF21, LLGL2, CFH, LEF1, CXCR4, CACNA1D, GATA3, NOD2, PLD1, FSTL1, FAM171A1, MUC1, ERBB3, MYO6, SLAMF8, AGR2, IGFBP6, S100A14, MEST, AP1M2, S100A10, SPDEF, FARP1, ARNT2, IL33, ANGPT1, CAV2, STAT1, CAV1, AKR1C1, ESR1, LARP6, MAPK13, DCBLD2, ARHGAP32, CXCL10, BMP2, IL6, MLPH, ID1, BHLHE40, GNAS, NMT2, SDC1, CD248, VIM, KCNK1	0.008875
	GO:0060090	Binding, bridging	CAV2, CAV1, CXCR4, VIM, TGFBR2	0.009638
	GO:0009986	Cell surface	SRPX, CNTNAP2, ANXA1, MME, BGN, CXCR4, SLC1A4, NOD2, TFPI, PRLR, DCBLD2, BMP2, SFRP1, SLAMF8, SDC1	1.76E-06
GO cellular component	GO:0005886	Plasma membrane	RET, CXCR4, CACNA1D, SLC1A4, NOD2, CELSR1, PLD1, ADD3, TFPI, RASGRP1, SGCE, FAM171A1, LAPTM4B, MUC1, ERBB3, MYO6, S100A10, WLS, VAV3, ANXA1, MME, ANGPT1, CAV2, ANXA3, CAV1, RRAS2, EMP1, PRLR, ESR1, TGFBR2, BMP2, SFRP1, MAP1B, FEZ1, CLDN7, GNAS, SDC1, NMT2, VIM, EZR, ADGRL4, KCNK1, ADGRL2, TNFRSF21, LLGL2	6.94E-06
	GO:0070062	Extracellular exosome	CFH, CXCR4, SLC1A4, FSTL1, MUC1, EFEMP1, MYO6, AOX1, S100A14, MEST, S100A10, WLS, ANXA1, MME, ANGPT1, ANXA3, AKR1C1, BGN, RRAS2, SFRP1, MLPH, MGAT4A, GNAS, SDC1, CD248, VIM, EZR	8.31E-06
	GO:0030425	Dendrite	RET, SGCE, FARP1, CNTNAP2, MLPH, MME, ANXA3, STAT1, MAP1B, FEZ1, SLC1A4, KCNK1	1.10E-05
	GO:0030424	Axon	RET, CNTNAP2, MME, ANXA3, STAT1, MAP1B, FEZ1, VIM, ADGRL2, TNFRSF21	5.27E-05

GO = Gene Ontology.

Table 2.

Significantly enriched KEGG pathways of DEGs.

Category	Term	Description	Genes	P value
	hsa05205	Proteoglycans in cancer	VAV3, ERBB3, CAV2, CAV1, RRAS2, SDC1, EZR, ESR1, FGF2, MAPK13	7.78E-06
KEGG pathway	hsa05200	Pathways in cancer	RET, ARNT2, STAT1, LEF1, CXCR4, PLD1, FGF2, RASGRP1, ESR1, TGFBR2, BMP2, IL6, GNAS	1.68E-04
	hsa04010	MAPK signaling pathway	ERBB3, ANGPT1, RRAS2, PLA2G4A, CACNA1D, FGF2, RASGRP1, MAPK13, TGFBR2	7.17E-04
	hsa04060	Cytokine-cytokine receptor interaction	IL33, CXCL10, IL6, BMP2, CXCR4, PRLR, TNFRSF21, TGFBR2	.003432
	hsa04912	GnRH signaling pathway	GNAS, PLA2G4A, CACNA1D, PLD1, MAPK13	.003454

KEGG = kyoto encyclopedia of genes and genomes.

3.3. PPI network and hub genes identification

To construct a PPI network, we submitted 102 common DEGs to STRING 11.5 tool. The results from the web tool were downloaded and imported into Cytoscape software to visualize and identify the hub genes using the CytoNCA plug-in (Fig. 4). Based on betweenness centrality, the top 5% of genes were selected as potential hub genes, and the functions and interactions of these genes were analyzed using GeneMANIA (Fig. 4). The results showed that the genes were in a PPI network with physical interactions of 77.64%, co-expression of 8.01%, predicted of 5.37%, colocalization of 3.63%, shared protein domains of 0.60%, and genetic interactions of 2.87% and were associated with steroid hormone response, hemostasis, coagulation, and regulation of T-cell activation.

Figure 4.

Protein–protein interaction (PPI) network construction and hub gene identification.

3.4. Prognostic value and survival prediction

We assessed the diagnostic efficacy of 4 hub genes in ER-positive BC by plotting receiver operating characteristic curves. (Fig. 5). The AUC for IL6, ESR1, PLA2G4A, and CAV1 in ER-positive BC and controls were 0.866, 0.839, 0.941, 0.909 (in GSE42568); 0.768, 0.811, 0.907, 0.946 (in GSE29044); and 0.803, 0.927, 0.917, 0.980 (in GSE29431), respectively. Next, we estimated the effects of the hub genes on the survival and prognosis of patients with ER-positive BC. Kaplan–Meier analysis revealed that low expression of CAV1 and PLA2G4A was correlated with shorter recurrence-free survival (RFS). However, only CAV1 was associated with the overall survival (OS) of patients with ER-positive BC (Fig. 6).

Figure 5.

ROC curves for hub genes in 3 ER-positive BC datasets. BC = breast cancer, ROC = receiver operating characteristic.

Figure 6.

Prognostic value analysis of hub genes for ER-positive BC patients. BC = breast cancer.

3.5. Analysis of immune infiltration

The proportions of the 22 immune cells in the PLAM and BC groups are shown in (Fig. 7A and 7B). In PLAM, the percentages of mast cells, macrophages, dendritic cells, B cells, and T cells changed considerably, whereas in ER-positive BC, only the percentages of macrophages and T cells increased. These results suggest that the patient immune system was activated. Interestingly, analysis of hub genes using GeneMANIA revealed an association of tumor immune cell infiltration with T-cell activation (Fig. 4). Further analysis revealed that immune infiltration by neutrophils, macrophages, mast cells, CD8 + T cells, and other immune cells was correlated with the expression of CAV1 in BC (Fig. 7C). This finding indicates that CAV1 may be related to the immune status of patients with BC; however, further studies are required to reveal the exact underlying mechanism.

Figure 7.

Analysis of immune infiltration of 22 immune cells in GSE12027 (A) and GSE42568 (B). (C) The correlation between the expression level of CAV1 and the infiltration level of immune cells in BC. BC = breast cancer.

3.6. Validation of CAV1 gene in PLAM and ER-positive BC

The western blot results revealed that CAV1 protein expression was lower in the PLAM cell model (TSC2-/-) and ER-positive BC cells (MCF-7) than in TSC2+/+ cells and human breast epithelial cells MCF-10A (Fig. 8A and 8B), consistent with the results of the data analysis in this study. The immunohistochemistry analysis revealed that the expression of CAV1 was lower in ER-positive BC tissues than in normal tissues (Fig. 8C).

Figure 8.

Experimental validation of CAV1 protein. Western blotting analysis of CAV1 expression in PLAM cell model (A) and ER-positive BC cell line (B). Immunohistochemical staining of CAV1 in human ER-positive BC tissue (C) (×100). BC = breast cancer, PLAM = lymphangioleiomyomatosis.

4. Discussion

Pulmonary lymphangioleiomyomatosis (PLAM) is a rare, hormone-sensitive, low-grade metastatic lung malignancy.^[1] Unlike most tumors, PLAM is a monogenic disease triggered by mutations in the TSC1/TSC2 gene, resulting in excessive activation of the mammalian target of rapamycin complex 1 (mTORC1), causing abnormalities in cell growth, metabolism, and senescence. It occurs almost exclusively in women of childbearing age and is considered an estrogen-sensitive disorder.^[2] Interestingly, the prevalence of invasive BC is higher in patients with PLAM than in the general female population, particularly in pre-menopausal women.^[8] Incidentally, a causative mutation in the TSC2 germline was found during a study of 1000 BC patients.^[20] These results suggest that the 2 diseases—PLAM and ER-positive BC—may have a common cellular origin and/or genetic risk factors. Therefore, exploring the molecular signaling network between PLAM and BC might help understand the complex link between its female-specific lung pathology, estrogen dependence, and TSC gene mutations—all of which have important clinical implications.

This study used bioinformatic analysis to identify common DEGs in PLAM and ER-positive BC. The results showed that these genes’ primary cell functions are wound healing regulation, gene expression, and cell proliferation. KEGG analysis showed that tumor-associated and cell proliferation signaling pathways are the common, important signaling pathways between PLAM and ER-positive BC, further indicating that PLAM has tumor-associated properties.^[2] In addition, the results of PPI network analysis suggest that CAV1, PLA2G4A, IL6, and ESR1 are hub genes for PLAM and ER-positive BC, all responding to steroid hormones, which substantiate the role of hormones in PLAM and ER-positive BC.^[21,22] However, only CAV1 expression was associated with both OS and RFS in patients with ER-positive BC.

CAV1 is an integral transmembrane protein involved in various cellular processes, such as membrane transport, cell cycle and proliferation, cancer cell invasion, migration, and metastasis. During the development of BC, CAV1 can act as a tumor promoter or suppressor, depending on the tumor subtype and stage.^[23] CAV1 downregulation in ER-positive MCF7 cells promoted tumor cell proliferation; CAV1 (mutant P132L) mutations occurred only in ER-positive BC cells, and its deletion resulted in ER overexpression and promoted estrogen-stimulated breast tumorigenesis.^[23,24] Downregulation of CAV1 in BC activates the MAPK pathway, and overexpression of CAV1 stimulates ER translocation to the plasma membrane and inhibits estrogen-induced activation of the extracellular signal-regulated kinase (ERK) pathway.^[24]Moreover, overexpression of CAV1 also downregulated ER and PR expression, inhibiting the ERK pathway in the small-cell lung cancer H466 cell line.^[25] Reduced CAV1 expression was found in both bleomycin-treated mouse lungs and patients with idiopathic pulmonary fibrosis, whereas overexpression of CAV1 alleviated bleomycin-induced lung injury.^[26] In this study, we found that CAV1 expression was downregulated, and ESR1 (which encodes ER) expression was upregulated, consistent with the results of previous studies. Further in vitro experiments showed that CAV1 expression was downregulated in PLAM cells, ER-positive BC cells, and tissues of patients with PLAM having ER-positive BC. Our findings indicate that CAV1 may be a key bridge in understanding the pathogenesis of PLAM and its association with ER-positive BC. However, no research has been conducted to determine the role of CAV1 in PLAM.

Immune cell invasion is highly correlated with the prognosis of tumor patients, and several drug treatments are available that target immune abnormalities in tumors and improve patient survival. The tumor microenvironment comprises tumor cells, infiltrating immune cells, and cancer-associated stromal cells.^[27,28] Abnormal changes in the immune status of PLAM, BC, and related therapeutic targets have become a research focus in recent years, especially the CD8 + T-cells, mast cells, and dendritic cells.^[29,30] In contrast, B, dendritic, and CD8 + T cells are associated with better prognoses in patients with ER-positive BC.^[31] In the present study, we found that immune infiltration of tumor-associated fibroblasts, macrophages, neutrophils, and CD8 + T-cells in BC was positively correlated with CAV1 expression. These findings will facilitate further studies on the mechanism of action of CAV1 in PLAM and BC.

5. Conclusion

Although the available evidence suggests a strong association between PLAM and BC, the specific underlying mechanisms have not yet been elucidated. To our knowledge, this is the first study to use a bioinformatics approach to probe the relationship between PLAM and ER-positive BC. Through a series of enrichment analyses, we hypothesized that the role of estrogen and tumor-related characteristics can be the common pathogeneses of PLAM- and ER-positive BC. In addition, we report that changes in the expression of some common hub genes that influence the effects of estrogen, such as CAV1, play a critical role in the disease progression. Therefore, CAV1 may serve as a common, latent therapeutic target for the treatment of PLAM and ER-positive BC.

However, our study does have some limitations. First, because PLAM is a rare disease that can only be diagnosed based on serum vascular endothelial growth factor D (VEGF-D) levels ≥ 800 pg/mL, clinical tissue specimens of PLAM were not obtained in this study. Second, the present study only verified the initial CAV1 protein expression and did not explore its related mechanisms; this aspect will be investigated in our next experimental study.

Author contributions

Conceptualization: Lulu Yang, Siying Ren.

Data curation: Lulu Yang.

Supervision: Siying Ren.

Validation: Xiao Ying.

Writing – original draft: Lulu Yang.

Writing – review & editing: Xiao Ying.