Serum sphingosine-1-phosphate levels are associated with brain metastasis in EGFR-mutant lung adenocarcinoma

ABSTRACT

Aim

To investigate the association between serum sphingosine-1-phosphate (S1P) levels and brain metastasis in EGFR-mutant lung adenocarcinoma (LUAD).

Methods

Serum S1P levels were analyzed in 103 LUAD patients. The baseline characteristics of the 103 patients in this study included the following: the overall cohort consisted of 50.49% males and 49.51% females. The average age of the cohort was 69.50 years (SD = 59.12). Regarding EGFR mutations, 52 patients (50.49%) had wild-type EGFR, 19 patients (18.45%) had EGFR Ex19Del, and 32 patients (31.07%) had the L858R mutation. Logistic regression models and competing risk Cox analyses were used to evaluate the association between S1P levels and brain metastasis. Kaplan-Meier curves assessed cumulative brain metastasis incidence over time.

Results

Serum sphingosine-1-phosphate (SIP) levels were measured with the following results (mean ± SD): wild-type EGFR, 970.44 ± 344.37 nmol/L; Ex19Del, 1,246.41 ± 306.93 nmol/L; and L858R, 1,333.21 ± 385.08 nmol/L (p < 0.001). In EGFR-mutant patients, S1P levels were independently associated with increased risk of brain metastasis (OR = 8.2, p = 0.003; HR = 105, p < 0.001), whereas no significant association was observed in EGFR wild-type patients. Kaplan-Meier analysis revealed that high S1P levels were linked to earlier brain metastasis in EGFR-mutant patients (p = 0.0034). The relationship between S1P levels and brain metastasis was not significantly influenced by the presence of bone metastasis (p > 0.1).

Conclusion

Elevated serum S1P levels are significantly associated with brain metastasis in EGFR-mutant LUAD patients. S1P may serve as a biomarker for brain metastasis risk and a potential therapeutic target.

Plain Language Summary

This study highlights the potential of serum S1P, a lipid signaling molecule, as a biomarker for brain metastasis in EGFR-mutant lung adenocarcinoma. Patients with higher S1P levels had an increased risk of developing brain metastasis, suggesting that monitoring S1P levels could improve early detection and management. Moreover, targeting S1P signaling pathways may offer novel therapeutic options to reduce brain metastasis risk.

1. Introduction

Patients with metastatic non-small cell lung cancer (mNSCLC) harboring epidermal growth factor receptor mutations (EGFR Mut) exhibit longer progression-free survival (PFS) and overall survival (OS) compared to those with EGFR wildtype (WT) when treated with EGFR tyrosine kinase inhibitors (TKIs) [1,2]. However, EGFR-TKIs have limited ability to penetrate the blood-brain barrier, leading to a higher incidence of brain metastases (BM). BM are a common complication in patients with lung cancer and are associated with poor prognosis. Additionally, when BM are detected early, while still small and few in number, stereotactic radiosurgery can be employed to target individual metastases without the need for whole-brain radiotherapy (WBRT) [3,4]. Timely detection of BM in NSCLC can facilitate management by enabling the initiation of curative or palliative treatment. Currently, no biomarker exists to predict the occurrence of BM in patients with advanced EGFR-mutant lung adenocarcinoma undergoing oral TKI treatment.

A recent study published in Nature Cancer highlights the potential reliance of metastatic breast cancer cells in the brain on de novo lipid synthesis and demonstrates the therapeutic benefit of inhibiting fatty acid synthase (FASN) in a mouse model [5]. In this study, Ferraro et al. aimed to determine whether such metabolic alterations could serve as therapeutic targets for BM. Additionally, another recent study by Jin et al., published in Nature, employed a barcoding strategy to evaluate the organ-specific metastatic potential of 500 cancer cell lines [6], by integrating their findings with genomic, transcriptomic, and metabolomic data, as well as results from a genome-wide CRISPR-Cas9 viability screen, the authors highlighted the critical role of fatty acid metabolism in breast cancer cells within the brain. Additionally, patients with EGFR-mutant metastatic NSCLC have a higher incidence of BM compared to those with EGFR wildtype [7,8]. Abnormal expression of fatty acid synthase (FASN) has been shown to drive brain-specific metastasis of primary HER2+ breast tumors [5,6]. Therefore, our group hypothesized that BM from EGFR-mutant lung adenocarcinoma is associated with lipid metabolism.

Sphingolipids are a family of membrane lipids that play structural roles in regulating the fluidity and subdomain organization of lipid bilayers. Additionally, bioactive sphingolipids, such as sphingosine-1-phosphate (S1P), function as signaling molecules involved in pathways directly relevant to carcinogenesis [9]. S1P is a key metabolite in the sphingolipid metabolism pathway and has been recognized for its ability to promote tumor progression [10]. While the biological roles of S1P and its related receptors in lung cancer have been well-characterized through in vitro and in vivo studies, particularly in tumor proliferation, migration, the potential involvement of S1P in brain metastases remains largely unexplored. We selected sphingosine-1-phosphate (S1P) for investigation based on our previous research [11], which demonstrated a significant association between S1P and brain metastasis in patients with advanced EGFR-mutant lung adenocarcinoma. Additionally, using an in vitro blood-brain barrier (BBB) model, we found that S1P can disrupt BBB integrity, suggesting a potential mechanistic role in facilitating brain metastasis. Moreover, S1P signaling is a critical determinant of blood-brain barrier (BBB) permeability and is considered a potential pathogenic factor or therapeutic target in diseases characterized by BBB dysfunction [12–14]. Studies on ischemia-reperfusion injury have shown that S1P can enhance STAT3 activity, leading to BBB dysfunction [15]. These findings suggest that inhibiting S1P could be a strategy to prevent BBB dysfunction and reduce the risk of BM.

Based on previous basic research demonstrating that S1P signaling plays a pivotal role in the proliferation and invasion of lung adenocarcinoma cells [16], we hypothesized that serum S1P levels increase with the onset of BM in patients with EGFR-mutant lung adenocarcinoma. Currently, there are no standard measures to predict the risk of BM in NSCLC patients. Therefore, detecting S1P levels might be incorporated into clinical practice as a prognostic indicator to guide individualized treatment and follow-up strategies. The aim of the present study was to investigate the relationship between serum S1P concentrations and brain metastasis in EGFR-mutant lung adenocarcinoma.

2. Patients & methods

2.1. Serum samples from LUAD patients

From September 2022 to December 2022, 130 Chinese patients with lung adenocarcinoma (LUAD) underwent tyrosine kinase inhibitor (TKI) or chemotherapy combined with immunotherapy in Harbin Medical University Cancer Hospital. During this period, serum samples were collected from 130 LUAD patients after obtaining informed consent. Inclusion criteria included confirmed diagnosis of LUAD with EGFR mutation or EGFR wild-type, IV stage, age ≥18 years, and no history of prior cancer. Exclusion criteria included patients with other malignancies or incomplete clinical data. We analyzed serum samples from 103 out of 130 patients. A total of 27 patients were excluded: 10 patients with stage III disease, 8 patients with concurrent other malignancies, and 9 patients who did not undergo genetic testing. The CONSORT flow chart is shown in Figure 1. All tumors were staged and classified according to the Lung Cancer Stage Classification, Eighth Edition [17]. Serum samples were collected from all patients at baseline (before treatment initiation) and stored at −80°C until analysis. This study was approved by the the Third Clinical Ethics Committee of Harbin Medical University.

Figure 1.

CONSORT flow diagram illustrating patient selection.

A total of 130 LUAD patients with collected serum samples were initially assessed. After exclusion of 27 patients (10 with stage III disease, 8 with concurrent other malignancies, and 9 without genetic testing), 103 patients were included for serum sample analysis.

2.2. ELISA assays

Human S1P ELISA kits were purchased from MEIMIAN Industrial Co., Ltd. (China). The assays were performed according to the manufacturer’s instructions. Written informed consent was obtained from all participants. The use of serum samples from the Third Hospital of Harbin Medical University was approved by the Third Clinical Ethics Committee of Harbin Medical University. All patients provided informed consent.

2.3. Statistical analysis

All statistical analyses in this report were performed using R version 4.4.1, with a two-sided p-value < 0.05 considered statistically significant.

2.3.1. Data source

This study analyzed data from 103 LUAD patients.

2.3.2. Methods

1. For normally distributed data, the mean and standard deviation were used to describe continuous variables, while counts and percentages were used for categorical variables.

2. Analysis of variance (ANOVA) was applied to compare normally distributed continuous variables among three groups. For two-group comparisons, independent sample t-tests were used.

3. For categorical variables, the chi-squared test or Fisher’s exact test was employed.

2.3.3. Logistic regression analysis

Baseline diagnostic data were analyzed to explore the relationship between S1P levels and brain metastasis under different EGFR mutation statuses. The analysis was conducted using three models:

• Model 1: Unadjusted model.

• Model 2: Adjusted for age, sex, BMI, and smoking status based on Model 1.

• Model 3: Further adjusted for bone metastasis and tumor location based on Model 2.

2.3.4. Competing risk Cox proportional hazards model

For patients without brain metastasis at baseline diagnosis, a competing risk Cox proportional hazards model was used to evaluate the relationship between S1P levels and brain metastasis under different EGFR mutation statuses and bone metastasis conditions, with death considered a competing event. The analysis included the following three models:

• Model 1: Unadjusted model.

• Model 2: Adjusted for age, sex, BMI, and smoking status based on Model 1.

• Model 3: Further adjusted for bone metastasis and tumor location based on Model 2.

2.3.5. Kaplan-Meier analysis

S1P levels were divided into two groups based on the mean value (1134.1 nmol/L). The cumulative incidence of brain metastasis was estimated using the Kaplan-Meier method, and differences between groups were compared using the log-rank test.

3. Results

3.1. Patient characteristics

A total of 103 patients were included in the analysis. Table 1 summarizes patient characteristics stratified by EGFR mutation type. No significant differences were observed in gender, age, smoking status, or brain metastasis incidence at diagnosis. However, S1P levels were significantly higher in the EGFR L858R mutation group (p < 0.001). The mean S1P concentrations in the EGFR wild-type, EGFR 19del, and EGFR L858R groups were 970.44, 1,246.41, and 1,333.21, respectively. There was no significant difference in the incidence of brain metastasis among different mutation types (p = 0.56).

Table 1.

Patient characteristics across the mutation type of EGFR.

Characteristic	Overall	Wild type	EGFR Ex19Del and Other	EGFR L858R	p-valuea
	N = 103	N = 52	N = 19	N = 32
Gender, n (%)					0.334
Male	52.00 (50.49%)	30.00 (57.69%)	8.00 (42.11%)	14.00 (43.75%)
Female	51.00 (49.51%)	22.00 (42.31%)	11.00 (57.89%)	18.00 (56.25%)
Age, Mean±SD	69.50±59.12	75.31±82.72	65.05±9.29	62.72±8.66	0.742
Smoking status, n (%)					0.109
Yes	45.00 (43.69%)	28.00 (53.85%)	6.00 (31.58%)	11.00 (34.38%)
No	58.00 (56.31%)	24.00 (46.15%)	13.00 (68.42%)	21.00 (65.63%)
BMI, Mean (SD	22.76 (3.04	21.61 (2.98	24.42 (2.75	23.66 (2.59	<0.001
Other gene mutation, n (%)					0.026
No	96.00 (93.20%)	45.00 (86.54%)	19.00 (100.00%)	32.00 (100.00%)
Yes	7.00 (6.80%)	7.00 (13.46%)	0.00 (0.00%)	0.00 (0.00%)
Sphingosine-1-Phosphate Levels(nmol/L), Mean±SD	1,134.05 ±386.56	970.44 ±344.37	1,246.41 ±306.93	1,333.21 ±385.08	<0.001
Brain metastases at diagnosis, n (%)					0.447
No	72.00 (69.90%)	37.00 (71.15%)	15.00 (78.95%)	20.00 (62.50%)
Yes	31.00 (30.10%)	15.00 (28.85%)	4.00 (21.05%)	12.00 (37.50%)
Brain metastases incidence, n (%) (during follow up)					0.56
No	86.00 (83.50%)	44.00 (84.62%)	17.00 (89.47%)	25.00 (78.13%)
Yes	17.00 (16.50%)	8.00 (15.38%)	2.00 (10.53%)	7.00 (21.88%)
Bone metastasis, n (%)					0.154
No	62.00 (60.19%)	36.00 (69.23%)	9.00 (47.37%)	17.00 (53.13%)
Yes	41.00 (39.81%)	16.00 (30.77%)	10.00 (52.63%)	15.00 (46.88%)
Treatment, n (%)					<0.001
TKI	42.00 (40.78%)	5.00 (9.62%)	17.00 (89.47%)	20.00 (62.50%)
Chemotherapy&immunotherapy	29.00 (28.16%)	29.00 (55.77%)	0.00 (0.00%)	0.00 (0.00%)
Other	32.00 (31.07%)	18.00 (34.62%)	2.00 (10.53%)	12.00 (37.50%)
Location of tumor, n (%)					0.549
Right Lung	57.00 (55.34%)	28.00 (53.85%)	9.00 (47.37%)	20.00 (62.50%)
Left lung	46.00 (44.66%)	24.00 (46.15%)	10.00 (52.63%)	12.00 (37.50%)

^aPearson’s Chi-squared test; anova test; Fisher’s exact test.

3.2. The distribution of S1P levels across groups with different clinical and pathological features

S1P levels varied significantly by EGFR mutation status, with the EGFR 21 mutant group exhibiting significantly higher S1P levels compared to the other groups (wild-type and EGFR 19 mutant, p < 0.01, Figure 2(c)). While S1P levels were slightly higher in patients with bone metastases compared to those without, the difference did not reach statistical significance (p = 0.066, Figure 2(d)). Similarly, there was no significant difference in S1P levels between patients with and without brain metastases, with a high degree of overlap in the distributions (p = 0.14, Figure 2(f)). Each subplot in the figure illustrates the point distribution of S1P levels and their statistical significance across groups, with annotated p-values indicating that most groups did not differ significantly, except for EGFR mutation subtypes, which showed significant differences. These findings suggest that S1P levels may be associated with specific EGFR mutation subtypes but show less variation across other clinical and pathological features.

Figure 2.

Comparative analysis of S1P levels across different clinical and pathological groups: (a) age, (b) sex, (c) EGFR subtype, (d) bone metastasis, (e) tumor location, and (f) brain metastasis. The y-axis represents serum S1P levels.

3.3. Association between S1P and BM risk in EGFR-mutant patients

Logistic regression analysis (Table 2) revealed that elevated S1P levels significantly increased the risk of brain metastasis in EGFR-mutant patients (OR = 8.2, 95% CI: 2.59–44.13, p = 0.003), independent of age, sex, BMI, bone metastasis, and tumor location. In contrast, no significant association was observed in wild-type patients. Among EGFR-mutant patients, S1P levels were strongly associated with an increased risk of brain metastasis, and this association remained significant after multivariable adjustment. Conversely, in EGFR wild-type patients, S1P levels were not significantly associated with the risk of brain metastasis. These findings suggest that EGFR mutations may modulate the relationship between S1P levels and brain metastasis, highlighting the need for further investigation.

Table 2.

Logistic regression analysis of S1P levels (per-1-sd increase) and brain metastasis in LUAD stratified by EGFR mutation status.

Characteristic	N	Event N	Model 1			Model 2			Model 3			P for interaction
			OR	95% CI1	p-value	OR	95% CI	p-value	OR	95% CI1	p-value
EGFR wild type	52	15	1.29	0.65, 2.70	0.473	1.1	0.51, 2.47	0.804	0.83	0.18, 3.23	0.780	0.082
EGFR mutation type(19del/L858R)	51	16	5.03	2.15, 14.90	<0.001	6.02	2.36, 20.73	<0.001	8.2	2.59, 44.13	0.003

Model 1: Crude model.

Model 2 adjusts for age, sex, smoking status, and BMI.

Model 3 adjusts for age, sex, smoking status, BMI, bone metastasis and tumor location.

Table 3 presents the results of a competing-risk Cox regression analysis assessing the association between S1P levels and brain metastases in LUAD, with death treated as a competing risk. Among EGFR-mutant patients, elevated S1P levels significantly increased the hazard of brain metastasis after adjusting for confounders (HR = 105, 95% CI: 7.77–1,414, p < 0.001). This strong association remained significant after multivariable adjustment, with a relatively high hazard ratio, indicating that S1P levels substantially increase the competitive risk of brain metastasis in EGFR-mutant patients. In contrast, no statistically significant relationship was observed between S1P levels and brain metastasis in EGFR wild-type patients. These findings suggest that EGFR mutation status may be a critical modulatory factor in the relationship between S1P levels and the risk of brain metastasis

Table 3.

Relationship between S1P levels and brain metastasis in LUAD under different EGFR mutation status.

Characteristic	Model 1			Model 2			Model 3
	HR	95% CI	p-value	HR	95% CI	p-value	HR	95% CI	p-value
EGFR wild type	1.06	0.41, 2.70	0.910	1.21	0.44, 3.29	0.710	1.42	0.48, 4.19	0.520
EGFR mutation type(19del/L858R)	10.2	3.59, 28.7	<0.001	32.5	10.6, 100	<0.001	105	7.77, 1,414	<0.001

Model 1: Crude model.

Model 2 adjusts for age, sex, smoking status, and BMI.

Model 3 adjusts for age, sex, smoking status, BMI, bone metastasis, and tumor location.

3.4. Association between S1P and bone metastasis risk in EGFR-mutant patients

Table 4 examines the relationship between S1P levels and brain metastasis in patients with and without bone metastasis. The results show that the association between S1P levels and brain metastasis was not statistically significant in either subgroup (p > 0.1). This indicates that bone metastasis does not modulate or influence the relationship between S1P levels and the risk of brain metastasis.

Table 4.

Relationship between S1P levels and brain metastasis in LUAD with or without bone metastasis.

Characteristic	Model 1			Model 2			Model 3
	HR	95% CI	p-value	HR	95% CI	p-value	HR	95% CI	p-value
Bone metastasis	2.98	0.82, 10.9	0.099	4.37	0.63, 30.4	0.14	1.42	0.48, 4.19	0.52
Have no bone metastasis	2.36	0.86, 6.50	0.096	2.73	0.64, 11.7	0.18	2.82	0.61, 13.2	0.19

Model 1: Crude model.

Model 2 adjusts for age, sex, smoking status, and BMI.

Model 3 adjusts for age, sex, smoking status, BMI, and tumor location.

3.5. Effect of sphingosine-1-phosphate (S1P) level on the timing of brain metastasis

Kaplan-Meier curves revealed that no significant difference was observed in the EGFR wild-type group (p = 0.87, Figure 3(a)). However, high S1P levels were associated with earlier brain metastasis in EGFR-mutant patients (p = 0.0034, Figure 3(b)). This figure reveals the different effects of S1P levels on the timing of brain metastasis in LUAD patients with EGFR mutation and EGFR wild type. In the EGFR mutant group, high S1P levels were significantly associated with earlier brain metastases, whereas in the EGFR wild-type group, S1P levels were not significantly associated with time to onset of brain metastases.

Figure 3.

Kaplan–Meier curves for time to brain metastasis in LUAD patients stratified by S1P levels and EGFR mutation status. (a) EGFR wild-type Group: No significant difference was observed in the time to brain metastasis between patients with high and low S1P levels (p = 0.87). (b) EGFR mutation Group: patients with high S1P levels had a significantly shorter time to brain metastasis compared to those with low S1P levels (p = 0.0034).

4. Discussion

The prevalence of BM in NSCLC is reported to be increasing, possibly due to advancements in brain imaging diagnostic tools and prolonged survival associated with new systemic therapies [18]. Some reports suggest that the incidence of BM may be higher in EGFR-mutant patients and that EGFR mutation might be an independent prognostic factor for survival compared to EGFR wild-type patients [19,20]. This study demonstrates that elevated serum S1P levels are strongly associated with brain metastasis in EGFR-mutant LUAD patients. These findings provide new insights into the role of S1P in the metastatic cascade and highlight its potential as a biomarker for identifying high-risk patients. S1P has been implicated in cancer metastasis through several mechanisms [13,21–25]. Moreover, S1P can activate STAT3 to cause BBB dysfunction [15]. S1P promotes vascular permeability, enabling tumor cells to cross the BBB [12]. Current evidence highlights the significant role of S1P signaling as a key determinant of BBB permeability, suggesting its potential as both a pathogenic factor and a therapeutic target in diseases associated with BBB dysfunction [12–14].

In this study, no significant difference was observed in the incidence of brain metastases between EGFR wild-type and EGFR-mutant patients. The incidence of brain metastases was 28.85% (15/52) in EGFR wild-type patients and 31.37% (16/51) in EGFR-mutant patients. These findings are inconsistent with the results of previous studies [19,20]. Several factors may account for this finding. Among patients with EGFR wild-type LUAD who presented with brain metastases at initial diagnosis, five cases were found to have concurrent genetic alterations, including ALK fusion (two cases), KRAS G12c mutations (two cases), and HER2 exon 20 insertions. These co-mutations may have influenced the metastatic behavior and contributed to the lack of a statistically significant difference in brain metastasis rates between the EGFR wild-type and mutant groups. However, in the subgroup analyses, we found that patients with EGFR 19del mutations exhibited higher SIP levels compared to those with EGFR wild-type, yet had a lower incidence of brain metastases. The relatively small number of patients with EGFR 19del mutations in our study (n = 19) may have limited our ability to detect brain metastases within this subgroup. Moreover, the incidence of brain metastases was significantly higher with EGFR L858R mutations than with EGFR 19del mutations. This result is consistent with that reported in the article [26,27]. The S1P concentration in the EGFR L858R mutant group was the highest, at 1,333.21 nmol/L. Our findings are consistent with previous studies showing elevated S1P levels in metastatic cancers. For instance, high S1P concentrations have been linked to peritoneal metastasis in lung cancer [16]. Moreover, S1P treatment significantly enhances EGF-stimulated colony formation, proliferation, and invasion of lung adenocarcinoma cells. Sphingosine is converted into S1P through the catalytic activity of sphingosine kinase 1 (SK1) [28]. It has been shown that increased SK1 expression and S1P formation, transactivate EGFR signaling pathway [29,30]. In EGFR-overexpressing glioma cells, extracellular S1P produced by SK1 contributes to increased invasiveness through the activation of the EGFRvIII-ERK-SK1-S1P pathway via the S1P1 receptor [31]. The strong association between S1P levels and brain metastasis in EGFR-mutant patients suggests that EGFR signaling may interact with S1P pathways. EGFR activation is known to upregulate SK1, the enzyme responsible for S1P synthesis, thereby increasing S1P production and promoting metastasis. This interaction could account for the elevated risk of brain metastasis observed in patients with concurrent EGFR mutations and high S1P levels.

To further explore the relationship between S1P levels and the occurrence of brain metastases, logistic regression analysis was performed, revealing that S1P levels were significantly associated with an increased risk of brain metastases in patients with EGFR-mutant lung adenocarcinoma. Death was treated as a competing risk for brain metastases in the context of EGFR mutation status, and the relationship between S1P levels and brain metastases in LUAD was analyzed using competing risk Cox regression. The analysis showed that elevated S1P levels significantly increased the competing risk of brain metastases in EGFR-mutant patients. Notably, S1P is already used as a prognostic marker in various solid tumors [32–34]. Brain metastases are the most common cause of oral TKI treatment failure in EGFR-mutated lung adenocarcinoma. Currently, there are no reliable biomarkers to predict brain metastases during oral TKI therapy. In this study, high S1P levels were found to be associated with early brain metastases in EGFR-mutant patients, highlighting the distinct impact of S1P levels on the timing of brain metastases between EGFR-mutant and EGFR wild-type LUAD patients. Based on these findings, S1P may serve as a biomarker to identify EGFR-mutant LUAD patients at high risk for brain metastases.

From the study: Beyond the brain, bone is a common site of metastasis in NSCLC [35]. Does bone metastasis influence S1P production? In other words, can S1P serve as a predictor for the occurrence of bone metastases? Existing studies have shown that the incidence of brain and bone metastases does not differ significantly between EGFR-mutant and EGFR-wildtype patients [36]. However, as shown in Table 1, the incidence of bone metastases is higher in EGFR-mutant patients compared to EGFR-wildtype patients, at 49.02% (25/51) and 30.77% (16/52), respectively. Nonetheless, bone metastases do not alter or influence the observed relationship between S1P levels and the risk of brain metastases (Table 4). This suggests that bone metastases do not modulate or influence the relationship between S1P levels and the risk of brain metastases.

A recent S1P inhibitor, Fingolimod, has been approved for the treatment of multiple sclerosis. In this condition, Fingolimod functions as an S1P analog, retaining T-cells in secondary lymphoid organs by inhibiting S1PRs, including S1PR1, S1PR3, S1PR4, and S1PR5 [37]. Through this mechanism, Fingolimod reduces T-cell infiltration into the brain, thereby limiting disease progression [38]. Additionally, the anti-cancer potential of Fingolimod has been demonstrated in several preclinical models across various tumor types. Fingolimod has been shown to suppress tumor growth in vivo in lung cancer models [39], and fingolimod also represses in vivo tumor progression in other cancer [40–43]. Interestingly, Fingolimod has also been shown in vitro to enhance the anti-cancer effects of EGFR-targeting therapies, Cisplatin, or Pemetrexed treatments in NSCLC models [39,44]. Therefore, targeting the S1P signaling axis using inhibitors such as fingolimod (FTY720) could represent a novel therapeutic strategy to prevent or delay brain metastasis. This study has limitations due to the small sample size and short follow-up period. Prospective studies and large sample are needed to validate these findings and explore the mechanistic interplay between EGFR mutations and S1P signaling.

5. Conclusion

Elevated serum S1P levels are significantly associated with brain metastasis in EGFR-mutant LUAD patients. In EGFR-mutant patients, S1P levels were independently associated with increased risk of brain metastasis, moreover, high S1P levels were linked to earlier brain metastasis in EGFR-mutant patients. Therefore, S1P may serve as both a biomarker for brain metastasis risk and a potential therapeutic target, warranting further investigation in clinical settings.

Funding Statement

This work was supported by Innovation Fund of the Department of Education of Heilongjiang Province [2023-KYYWF-0228] and the National Natural Science Foundation of China [82373041].

Article highlights

• Brain metastasis (BM) is a major complication in EGFR-mutant lung adenocarcinoma (LUAD) with limited predictive biomarkers.

• Sphingosine-1-phosphate (S1P), a lipid signaling molecule, is implicated in cancer metastasis and blood-brain barrier (BBB) dysfunction, but its role in BM remains underexplored.

• This study analyzed serum S1P levels in 103 LUAD patients using ELISA assays.

• Logistic regression, competing-risk Cox models, and Kaplan-Meier analysis evaluated associations between S1P levels and BM risk.

• S1P Levels by EGFR Status: EGFR L858R mutants had the highest S1P levels, followed by Ex19Del and wild-type.

• BM Risk in EGFR-Mutants: Elevated S1P independently increased BM risk, with no significant association in wild-type patients.

• Timing of BM: High S1P correlated with earlier BM in EGFR-mutants, but not in wild-type.

• Bone Metastasis: S1P’s association with BM was unaffected by bone metastasis, suggesting specificity to BM pathogenesis.

• S1P may serve as a novel biomarker for BM risk stratification in EGFR-mutant LUAD, enabling earlier detection and intervention.

• Targeting S1P signaling (e.g., with inhibitors like fingolimod) could represent a therapeutic strategy to prevent or delay BM.

Author contribution

Yan Yu: conceptualization, methodology, supervision. Dexin Jia: writing-reviewing and editing. Bo An and Bo Pan: visual-ization, investigation. YaJie Li and LiHua Shang: validation, software, data curation. Gang Xu: writing-original draft preparation, editing, data curation. All authors read and approved the final manuscript.

Disclosure statement

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosure

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research

The authors state that they have obtained appropriate institutional review board approval (Third Clinical Ethics Committee of Harbin Medical University(KY2023–68)) and/or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.

AI-Based tools and technologies

The authors acknowledge the use of ChatGPT (OpenAI, version 4.0) for language polishing and improving the clarity of the manuscript. The tool was used solely to enhance the readability of the text without altering the scientific content.