LAMP3 is upregulated in a variety of human cancers
LAMP3 is upregulated in a variety of human cancers, so we conducted a study to explore its effects and mechanisms of action in pan-cancer. First, to investigate the expression of LAMP3 in human cancers, we utilized the SangerBox3.0 online database to analyze the expression of LAMP3 in the TCGA and GTEx datasets. It has been observed that LAMP3 is up-regulated in various human cancers, including glioblastoma multiforme (GBM), lower grade glioma (LGG), uterine corpus endometrial carcinoma (UCEC), breast invasive cancer (BRCA), cervical and endocervical cancer (CESC), esophageal carcinoma (ESCA), stomach and esophageal carcinoma (STES), pan-kidney cohort (KIPAN), colon adenocarcinoma (COAD), stomach adenocarcinoma (STAD), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), liver hepatocellular carcinoma (LIHC), bladder urothelial carcinoma (BLCA), thyroid carcinoma (THCA), ovarian serous cystadenocarcinoma (OV), pancreatic adenocarcinoma (PAAD), uterine carcinosarcoma (UCS), acute lymphoblastic leukemia (ALL), acute myeloid leukemia (LAML), adrenocortical carcinoma (ACC), kidney chromophobe (KICH), and cholangiocarcinoma (CHOL). In contrast, LAMP3 was down-regulated in lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), and kidney renal papillary cell carcinoma (KIRP) (Fig. 1A). Then, the TIMER2.0 online database was used to analyze the expression of LAMP3. The results showed that LAMP3 was significantly increased in tumor tissue compared to normal tissue in various human cancers, including BLCA, BRCA, CESC, CHOL, COAD, esophageal carcinoma (ESCA), GBM, HNSC, KICH, KIRC, LIHC, STAD, THCA, and UCEC. However, LAMP3 was downregulated in LUAD (lung adenocarcinoma), LUSC (lung squamous cell carcinoma), and PRAD (prostate adenocarcinoma) (Fig. 1B). We then analyzed the relationship between LAMP3 and the stage of pan-cancers using GEPIA2. LAMP3 was associated with the stage of THCA, skin cutaneous melanoma (SKCM), PAAD, HNSC, OV, KIRP, KIRC, and BLCA, and the expression of LAMP3 was significantly positively correlated with the stages of PAAD, KIRP, and KIRC (Fig. 1C).
The Diagnostic and Prognostic Value of LAMP3 In Pan-Cancer
We then used the ROC curve to analyze the diagnostic value of LAMP3 in different types of cancer. As shown in Figures S1A–L, LAMP3 may act as a reliable diagnostic marker in various types of cancer, including CESC (AUC = 0.988), LUSC (AUC = 0.986), LUAD (AUC = 0.983), ESCA (AUC = 0.967), SKCM (AUC = 0.889), SARC (AUC = 0.768), STAD (AUC = 0.828), HNSC (AUC = 0.822), PCPG (AUC = 0.805), UCEC (AUC = 0.806), KIRC (AUC = 0.769), and PRAD (AUC = 0.706) (Fig.S1A-L) (Table. S1).
Furthermore, an increase in LAMP3 level was associated with poor overall survival (OS) in KIRC (p = 0.0065), KIRP (p = 8.4e-07), LUSC (p = 0.042), PAAD (p = 0.011), THYM (p = 0.039), UCEC (p = 0.0022), and TGCT (p = 0.041). Increased expression of CENPL was associated with improved overall survival (OS) in several cancer types, including BRCA (p = 0.002), CESC (p = 0.044), LUAD (p = 0.0068), OV (p = 0.0014), and READ (p=0.042) (Fig.2A-L).
Characteristics of LAMP3 mutations in pan-cancer
Genetic mutation of oncogenes or tumor suppressor genes is associated with growth and progression of several tumors [25], so we analyzed the mutational status of LAMP3 in TCGA cohorts. As observed in Figure 3A, there were 32 types of cancers with LAMP3 mutations, and lung squamous cell carcinoma had the highest frequency at 41.07%. The sites, types, and numbers of mutations in LAMP3 are shown in Fig. 3B. Next, we explored the correlation between LAMP3 mutations and clinical prognosis. Fig 3C-F shows that pancreatic cancer with LAMP3 mutations tends to have worse overall survival (p = 5.418e-6), disease-specific survival (p = 1.075e-3), progress-free survival (p = 0.0268), but except disease-free survival (p = 0.0988).
LAMP3 is associated with immune infiltration in pan-cancer.
The ESTIMATE score is calculated by adding the immune and stromal scores. The immune score indicates the percentage of immune cells present in the tumor tissue, while the stromal score represents the percentage of stromal tissue within the tumor tissue. A high ESTIMATE score indicates a high tumor purity. To understand the cellular composition of the tumor microenvironment, we analyze the relationship between the ESTIMATE scores and LAMP3 expression in pan-cancers. LAMP3 expression was significantly positively correlated with the ESTIMATE score of 36 human cancers, including CESC (p = 3.4e-6), BRCA (p = 5.5e-78), LUAD (p = 5.9e-7), ESCA (p = 6.1e-3), KIRP (p = 4.1e-14), etc. (Fig. 4A). However, it can be observed from our results that there is a negative correlation between LAMP3 expression and ESTIMATE, immune, and stromal scores in Thymoma (THYM). This suggests that LAMP3 expression is associated with a decrease in immune cells and stromal cells (Table.S2). Antitumor immunity is a strong predictor of the efficacy of tumor immunotherapy and correlates with tumor mutation burden (TMB), and mutant-allele tumor heterogeneity (MATH) [26]. Then, we analyzed the relationship between LAMP3 expression and TMB or MATH to explore whether LAMP3 can predict immunotherapy responses. The obtained results show that LAMP3 has a positive correlation with TMB in COAD (p = 4.54e-4), COADREAD (p = 1.98e-3), and BRCA (p = 0.0235). However, LAMP3 has a negative correlation with TMB in CESC (p = 0.0107), LUAD (p = 8.58e-8), STES (p = 0.0225), KIRP (p = 2.91e-4), HNSC (p = 0.0437), and LUSC (p = 1.16e-3) (Fig. 4B). The expression of LAMP3 also has a positive correlation with MATH in LAML (P = 3.39e-4), KIRP (p = 7.78e-4), KIPAN (p = 2.58e-6), and UCEC (p = 9.30e-3). However, LAMP3 has a negative correlation with MATH in BMLGG (p = 2.47e-8), LGG (p = 0.0437), LUAD (p = 2.49e-3), STAD (p = 1.83e-3), PAAD (p = 4.00e-3), and TGCT (p = 2.22e-4) (Fig. 4C) (Table S3). The most potential targets for cancer immunotherapeutic therapies are the immunological checkpoint (ICP) blocking proteins because they control the infiltration of immune cells into the tumor microenvironment. We analyzed the relationship between LAMP3 expression and immune checkpoint (ICP). Currently, immune checkpoint blockade therapy has shown promising results in the treatment of tumors in clinical practice, making it one of the most prominent tumor immunotherapies today. LAMP3 was found to have a positive correlation with ICP genes in all human cancers studied (Fig. 4D). Furthermore, the Sangerbox database was utilized to examine the correlation between LAMP3 expression and tumor-infiltrating immune cells (TIICs). These TIICs encompassed memory B cells, plasma cells, CD8+ T cells, memory CD4+ and CD8+ T cells, NK cells, regulatory T cells, monocytes, macrophages, dendritic cells, mast cells, eosinophils, and neutrophils. It could be observed that the association between LAMP3 expression and immune cells in each type of tumor changes significantly. It is obvious that the expression of LAMP3 showed a strong positive relationship with dendritic cells activated(Fig. 4E) (Table.S4).
Next, we utilized TIMER 2.0 to investigate the potential involvement of LAMP3 in immune cell infiltration. It can be observed from the results that there is a significant positive correlation between DC infiltration level and LAMP3 expression in most human cancers (Fig. 5A). The top six tumors showing this correlation were TGCT (rho = 00.638, p = 3.69e-18), PAAD (rho =00.601, p = 33.78e-18), COAD (rho =0.589, p = 4.47e-27), STAD (rho =0.577, p = 4.73e-35e-05), PRAD(rho = 0.535, p =3.19e-32), and HNSC(rho = 0.507, p = 4.06e-07) (Fig. 5B). Data also demonstrated a positive correlation between the infiltration level of Treg cells and the expression of LAMP3 in most human cancers (Fig. 5C). The top 6 tumors with the highest correlation are CHOL (Rho = 0.617, p = 7.94e-05), PRAD (Rho = 0.58, p = 8.24e-39), TGCT (Rho = 0.528, p = 6.46e-12), COAD (Rho = 0.524, p = 8.03e-21), BRCA-LumA (Rho = 0.521, p = 2.53e-37), and STAD (Rho = 0.517, p = 3.07e-27) (Fig. 5D).
Subsequently, the role of LAMP3 in chemokines and chemokine receptors was investigated using the TISIDB database. The results showed a positive correlation between the expression of LAMP3 and most chemokines. Regarding KIRC, it can be observed that the levels of CCL5 (rho = 0.388, p < 2.2e-16), CCL17 (rho = 0.22, p = 2.96e-07), CCL19 (rho = 0.338, p = 1.22e-15), CCL22 (rho = 0.453, p < 2.2e-16), CCL10 (rho = 0.387, p < 2.2e-16), and CCL9 (rho = 0.484, p < 2.2e-16) are significantly positively correlated with LAMP3 expression (Fig. 6A), which can increase the chemotactic effect on Treg. Furthermore, our results also suggest that LAMP3 expression is positively associated with most of the chemokine receptors. In KIRC, it was observed that the levels of CCR4 (rho = 0.412, p < 2.2e-16), CCR7 (rho = 0.493, p < 2.2e-16), and CCR8 (rho = 0.559, p < 2.2e-16) are significantly positively correlated with LAMP3 expression (Fig.6B).
Single-cell analysis found that DC cells with high expression of LAMP3 interact with Treg
Next, we examined the causes of these changes in LAMP3 expression. High expression of LAMP3 in HCC and CRC cells was found to be concentrated in immune cells, as shown by single-cell sequencing analysis (Fig. 7A). After integrating immune cells from 13 different tumor types, we conducted additional investigations and discovered that a cluster of cDC cells had high expression of LAMP3 (Fig. 7B). By re-clustering, we were able to obtain 9 DC clusters. These clusters were then divided into the following 4 subsets based on marker gene expression (Fig. 7C): DC_CD1C, DC_CLEC9A, pDC, and DC_LAMP3. According to the classic DC markers, DC_CD1C cells, which express high levels of CD1C, FCER1A, and CLEC10A, correspond to cDC2s. These cells are responsible for antigen presentation on MHC-II molecules, which is crucial for CD4+ T-cell priming. DC_CLEC9A cells, which specifically express CLEC9A and XCR1, correspond to cDC1s. cDC1s specialize in the cross-presentation of antigens onto MHC-I molecules for CD8+ T-cell priming. DC_LAMP3 corresponds to cDC3s, a newly discovered DC cell. The cDC3s in all types of cancer demonstrated migratory ability with high expression of CCR7 and FSCN1, immunomodulation ability with high expression of CD274, FAS, activation with high expression of CD40 and CD80, maturation with LAMP3 (Fig.7D). Furthermore, to explore the relationship between cDC3s and other types of cDCs, we conducted pseudotime analysis. The results show that cDC3s were derived from both cDC1s and cDC2s in a terminal manner (Fig. 7E), indicating that cDC3s were the most differentiated and mature cells. GSVA was used to identify a distinct pattern of enriched pathways within the DC cell clusters. It could be observed that the NF-κB, apoptosis, and hypoxia pathways were upregulated in cDC3s compared to the other three clusters (Fig. 7F). To investigate the cellular communication network in pan-cancers, we conducted a further analysis of the cellular interactions between Treg cells, exhausted T cells, and cDC3s using the CellChat software. CDC3 cells were predicted to interact with Treg cells through the CCL17-CCR4 and CCL22-CCR4 pathways [31], which are known to attract Treg cells. CDC3 cells were also predicted to interact with exhausted T cells through LGALS9_HAVCR2 signaling[32], an immune-suppressive pathway (Fig. 7G) (Table.S5). The SCENIC algorithm suggested that the activation of the KLF7, ETV3, JUN, REL, RELB, IRF1, and NFKB2 motifs was correlated with upregulated immunosuppressive activity in cDC3s (Fig. 7H).