SMPD1-mediated sphingolipid metabolism dysregulation in predicting and preventing in-stent restenosis of coronary heart disease patients

Graphical abstract

Highlights

• •Network-based metabolomic biomarkers accurately predicts ISR in CTO patients.
• •Sphingolipid metabolism dysregulation is a key feature of ISR pathogenesis.
• •SMPD1 is identified as central regulator in ISR development.
• •Quinapril hydrochloride inhibits SMPD1 and reduces neointimal formation.
• •Combined biomarker-based prediction and SMPD1 targeting offers new ISR preventation management.

Coronary heart disease (CHD) represents a leading cause of mortality worldwide, with a rising incidence. Chronic total occlusion (CTO) is an increasingly recognized and remarkable subgroup of CHD [1]. In-stent restenosis (ISR) remains a significant complication following percutaneous coronary intervention (PCI) in CHD patients, particularly in CTO patients [2,3]. In recent years, although drug-eluting stents (DES) have decreased overall ISR rates and bioresorbable scaffolds have shown distinct patterns of vascular healing and inflammatory response, the heightened incidence of ISR in CTO lesions underscores the imperative for both enhanced early predictive methodologies and novel prevention interventions [4,5].

To address this unmet clinical need, four parts were designed and conducted (Fig. S1A). First, plasma samples of CTO patients were collected 48 h post-procedure underwent metabolomic profiling, with 9 months followed-up. Second, metabolomic network analysis and machine learning were employed to construct biomarkers for ISR prediction. Third, network topology and pathway analyses were conducted to identify key metabolic alterations. Fourth, the prevention potential of identified target was evaluated through virtual screening, in vitro, and in vivo validation. The detailed methods are shown in the Supplementary data. Our clinical study protocol was approved by the Ethics Committee of Beijing Anzhen Hospital, China (IRB#: AZHEC2016-0516, AZHEC2017-0813, and AZHEC2022-0724). Animal care and experimental protocols were approved by the Animal Ethics Committee of Beijing Anzhen Hospital, China (Approval No.: 0124-1-1-ZX(Y)−4).

In the first part, we conducted a two-phase prospective study (Supplementary data). In the discovery cohort, we enrolled a total of 161 CTO patients who underwent successful PCI, among whom 68 developed ISR during follow-up. Of the 162 CTO patients assessed in the additional validation cohort, 69 cases developed ISR following PCI (Tables S1 and S2).

In the second part, we constructed correlation-based networks separately for ISR and non-ISR cohorts (Figs. S1B and C). To ensure robust comparison of network complexity while controlling for statistical power variations due to unequal sample sizes, we implemented an iterative random resampling approach (Fig. S2A). In the discovery cohort, ISR networks demonstrated significantly enhanced complexity compared to non-ISR networks across all network metrics (Figs. 1A, S2B, and S2C). This finding was independently validated in the second cohort (Figs. S2D−F), which exhibited concordant patterns of increased metabolic network complexity in ISR patients. Cross-validation analysis identified a core network of 175 metabolites connected by 247 edges that were consistently present across all five bootstrap sizes (Fig. 1B). Following the identification of dysregulated metabolite pairs in ISR patients, a core interaction network as a refined biomarker panel was constructed for early ISR prediction, designated as NetMet-ISR (Fig. 1C). Subsequently, to evaluate the predictive performance, we conducted a comparative analysis between NetMet-ISR and conventional single-metabolite biomarkers (Supplementary data). In the discovery cohort, NetMet-ISR showed superior predictive performance with area under the curve (AUC) values of 0.935 (Figs. 1D and E). These findings were subsequently validated in an independent cohort (Fig. S3).

Fig. 1.

Sphingomyelin (SM) phosphodiesterase 1 (SMPD1)-mediated sphingolipid metabolism dysregulation in predicting and preventing in-stent restenosis (ISR) of coronary heart disease (CHD) patients. (A) Edges of significant pairs in the metabolic network of ISR (n = 68) and non-ISR (n = 93) in discovery cohort. (B) Cross-validation analysis across all five bootstrap sizes. (C) A core interaction network, designated as NetMet-ISR (node: one metabolite, which was colored and shaped by classification of metabolic pathways; gray edge: positive correlation; blue edge: negative correlation; and red edge: correlation reversed). (D) Receiver operating characteristic (ROC) curve of NetMet-ISR, butyrylcarnitine, indoxyl sulfate, and dihydroceramide (DHC) (18:1/24:0) in discovery cohort. (E) Comparative assessment of NetMet-ISR predictive performance in discovery cohort, using area under the curve (AUC), recall, precision, F1 score, sensitivity, specificity, or accuracy as a metric to quantify performance. (F) Pathway enrichment analysis of the dysregulated metabolites pairs correlated with ISR occurrence. Point size indicated the count of metabolites in pathway. (G) Box plots showing differential levels of ceramide (d18:1/12:0) and SM (d18:1/22:2) between ISR and non-ISR patients. (H) Schematic illustration of sphingolipid metabolism pathway in lysosomes. SMPD1 catalyzes the hydrolysis of SM to ceramide, which can be further metabolized to sphingosine and sphingosine-1-phosphate through the action of ceramidase (CDase) and sphingosine kinase (SphK), respectively. (I) Dose-response curve demonstrating SMPD1 inhibition by quinapril. The half-maximal inhibitory concentration (IC₅₀) value was determined to be 9.1 μM. Three independent experiments were carried out, each including three technical replicates. (J) Representative hematoxylin and eosin staining on the cross-sections of stented coronary arteries from porcine models with or without quinapril hydrochloride (QHCL) treatment (n = 5 for each group). (K, L) Quantification of the neointimal area (K) and percentage area stenosis (L) in hematoxylin and eosin–stained cross-sections of stented coronary arteries. ^∗P < 0.05 and ^∗∗∗P < 0.001.

In the third part, pathway enrichment analysis revealed significant dysregulation in three major sphingolipid-related pathways: ceramide metabolism, sphingomyelin (SM) metabolism, and deoxysphingolipid metabolism (Fig. 1F). Topological analysis of the metabolic network revealed a cascade of metabolic dysregulation originating from core sphingolipid metabolism, rather than random metabolic alterations across pathways (Supplementary data). Next, metabolomic profiling revealed significant perturbations in the SM-ceramide pathway, characterized by elevated ceramide (d18:1/12:0) levels (P < 0.001) and concomitantly decreased SM (d18:1/22:2) concentrations (P = 0.026) in ISR patients compared to non-ISR controls (Fig. 1G). Analysis of the sphingolipid metabolic cascade revealed that SM phosphodiesterase 1 (SMPD1) is the key lysosomal enzyme catalyzing SM hydrolysis to ceramide (Fig. 1H). The observed metabolic profile, characterized by SM depletion and ceramide accumulation, indicated enhanced SMPD1 enzymatic activity in ISR patients. These findings established SMPD1 hyperactivity as a promising prevention target.

In the fourth part, 1729 U.S. Food and Drug Administration (FDA)-approved drugs were screened against the target protein SMPD1 by structure-based virtual screening (Supplementary data). Quinapril hydrochloride (QHCL) was identified for further studies (Fig. S4A). Based on the in vitro validation experiment, quinapril exhibited a concentration-dependent inhibitory effect on SMPD1 activity (Fig. 1I). The half-maximal inhibitory concentration (IC₅₀) was determined to be 9.1 μmol/L.

Next, the effects of QHCL on neointimal formation were evaluated in a porcine coronary stent model (n = 15). The experimental design is shown in Fig. S4B. Histomorphometric analysis of stented arterial segments revealed significant differences among groups (Fig. 1J). Representative photomicrographs of hematoxylin and eosin (H&E) demonstrated marked neointimal formation in vehicle-treated controls, while both QHCL treatment groups exhibited reduced neointimal proliferation (Supplementary data). Histopathological evaluation of neointimal area demonstrated that QHCL treatment resulted in a dose-dependent reduction compared to the vehicle-treated group (Fig. 1K and Table S3). The mean neointimal area was significantly decreased from 1.205 ± 0.043 mm² in the vehicle group to 1.054 ± 0.041 mm² in the low-dose QHCL group (0.2 mg/kg/day, P = 0.040) and 0.827 ± 0.043 mm²in the high-dose QHCL group (1 mg/kg/day, P < 0.001). Moreover, the difference between the two QHCL groups was also statistically significant (P = 0.014), indicating a concentration-dependent effect. Similarly, the percentage of stenosis was significantly attenuated by QHCL treatment (Fig. 1L and Table S3). The mean stenosis rate was reduced from 36.39% ± 0.937% in the vehicle group to 31.49% ± 1.000% in the low-dose group (P = 0.015) and 28.08% ± 0.829% in the high-dose group (P < 0.001). The difference between the two QHCL doses was also statistically significant (P = 0.033), further supporting the dose-dependent inhibitory effect on restenosis.

In conclusion, we employed a comprehensive and multi-staged approach to identify and validate novel prediction biomarkers and prevention strategies for ISR targeting the SMPD1-mediated sphingolipid dysregulation in ISR pathogenesis. This comprehensive approach not only provides insights into the metabolic basis of ISR, but also helps to identify high-risk CTO patients early and further presents a promising therapeutic strategy for its prevention. Before clinical claims are made, future studies should incorporate prospective, multi-center cohorts with varied patient characteristics, procedural techniques, and stent types to confirm the robustness and clinical utility of our findings.

CRediT authorship contribution statement

Weiyu Meng: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis, Conceptualization. Yuyang Sha: Writing – original draft, Methodology, Formal analysis. Ci-Ren Zhong-ga: Resources, Investigation, Data curation. Hongxin Pan: Software, Methodology. Xiaobing Zhai: Software, Data curation. Pu Zhen: Investigation, Data curation. Henry H.Y. Tong: Supervision, Data curation. Edmundo Patricio Lopes Lao: Supervision, Data curation. Hongjia Zhang: Supervision, Resources. Wenzhi Yang: Supervision, Resources. Song Cui: Writing – review & editing, Supervision, Project administration. Xiantao Song: Writing – review & editing, Supervision, Project administration. Kefeng Li: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: