The S1PL^C317A knock-in (KI) mice used in this study carry a point mutation in exon 11 of the sphingosine-1-phosphate lyase 1 (Sgpl1) gene, substituting cysteine (codon TGT) at amino acid position 317 with alanine (codon GCT). This site-directed mutation disrupts S1PL catalytic activity and was introduced using the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/Cas9 genome-editing system in Professor Shau-Ku Huang’s laboratory at the National Institutes of Health, Taiwan (Supplemental Fig. S1A). Oxidation normally inhibits S1PL activity; however, the C317A mutation renders S1PL resistant to oxidation. Consequently, the C317A mutant exhibits increased enzymatic activity relative to wild-type (WT) S1PL, resulting in reduced S1P levels.

Single-guide RNAs (sgRNAs) targeting exon 11 of Sgpl1 were designed to generate the KI model. The sgRNAs and Cas9 mRNA were synthesized in vitro and co-injected into fertilized C57BL/6 mouse zygotes. The resulting embryos were transferred to pseudo-pregnant female mice to generate founder animals.

For genotyping, genomic DNA was extracted from neonatal mouse toe biopsies. A 579 bp DNA fragment containing the mutation site was amplified by polymerase chain reaction (PCR) using these primers: Forward: 5ʹ-GTTTCTCCTGTGGTGGTTGG-3ʹ and reverse: 5ʹ-AATGCTGGTCACACCTTTCA-3ʹ. PCR products underwent restriction fragment length polymorphism (RFLP) analysis using the Sal I enzyme (New England Biolabs, Ipswich, MA, USA; #R0138). A silent mutation (GTC→GCT), introduced alongside the C317A substitution, created a Sal I recognition site within the KI allele. Digestion of PCR products yielded two fragments (480 and 99 bp) in KI alleles, whereas WT alleles remained intact at 579 bp (Supplemental Fig. S1B).

To further validate the mutation, direct Sanger sequencing was performed. The chromatograms confirmed the intended nucleotide substitution (TGT→GCT) at codon 317, clearly distinguishing WT from KI genotypes (Supplemental Fig. S1C).

Together, these two complementary methods—PCR-RFLP and Sanger sequencing—provided robust molecular evidence confirming the successful generation of the Sgpl1^C317A KI mouse line. Plasma S1P concentrations from WT and SGPL KI mice were measured using an enzyme-linked immunosorbent assay (ELISA) kit (MyBioSource, San Diego, CA, USA; #MBS1602779). Additionally, 6-week-old male C57BL/6JNarl mice were purchased from Taiwan’s National Laboratory Animal Center. All animal protocols were approved by the Institutional Animal Care and Use Committee of Taipei Medical University Wan Fang Hospital, Taiwan (protocol #: wan-111-024), complying with animal management regulations. The mice were housed under controlled conditions (22°C, 50% relative humidity) with a 12-hour light/dark cycle and free access to food and water.

The mice were acclimated to the environment for 1 week prior to the experiment. The sample size was determined by an a priori power analysis using G*Power software version 3.1.9.2 (Franz, Universität Kiel, Kiel, Germany), specifying an effect size of 2, alpha level of 0.05, and a power of 0.90, requiring seven mice per group (14 mice total). Mice were then allocated into four groups (n>7 per group) by an investigator blinded to experimental protocols. The control group received a normal diet (ND; standard chow diet), and the diabetes model group received a high-fat diet (HFD; Research Diets, New Brunswick, NJ, USA; D12331, 58% fat energy/sucrose). T2DM was induced by HFD feeding for 3 months [25]. Mice with high body weight and fasting blood glucose levels ≥11.1 mmol/L were classified as diabetic. Groups included WT mice (ND, n=7; HFD, n=10; total, 17 mice) and SGPL1 KI mice (ND, n=7; HFD, n=10; total, 17 mice). Mouse body weights were recorded weekly throughout the experiment. At the study’s conclusion, the mice were euthanized.

After a 13-week diet period, oral glucose tolerance tests (OGTTs) and insulin measurements were conducted following an 8-hour fast. Mice received an oral glucose load of 1 g/kg, and blood glucose was measured via tail bleeding at 0, 15, 30, 60, 90, and 120 minutes using a Contour Plus blood glucose meter (Ascensia Diabetes Care, Basel, Switzerland). The area under the curve (AUC) was calculated from these measurements. Upon sacrifice, plasma samples, liver, and kidney tissues were collected. Plasma insulin concentrations were measured with an ELISA kit (Mercodia Mouse Insulin ELISA, Uppsala, Sweden), and glycated hemoglobin (HbA1c) levels were determined using the Apex Bio Eclipse Hemoglobin A1c reagent kit (Houston, TX, USA). Serum and hepatic triglyceride (TG) concentrations were quantified using the Triglyceride Colorimetric Assay Kit (#10010303, Cayman Chemical, Ann Arbor, MI, USA), according to the manufacturer’s instructions. Serum samples were diluted 1:2 before analysis. For hepatic TG quantification, approximately 160–200 mg of liver tissue was homogenized in 1× NP-40 substitute buffer and centrifuged at 10,000 ×g for 10 minutes at 4°C. The resulting supernatant was analyzed spectrophotometrically at 540 nm. TG concentrations were calculated using a standard curve and expressed as milligrams per gram of liver tissue (mg/g tissue).

Plasma non-esterified fatty acid (NEFA) concentrations were measured using the Randox NEFA assay kit (Randox Laboratories, Antrim, UK; #FA 115) following the manufacturer’s instructions. Serum samples were incubated with assay reagents at 37°C for 10 minutes, and absorbance was recorded at 550 nm. NEFA levels were determined using a standard curve and expressed as millimoles per liter (mmol/L).

Total RNA was extracted from mouse liver tissues using TRIzol^TM Reagent (#15596026; Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. Subsequently, 1 μg of total RNA was reverse-transcribed into cDNA using the ProtoScript^® First Strand cDNA Synthesis Kit (#E6560L; New England Biolabs). Quantitative real-time PCR (qRT-PCR) was performed using Fast SYBR^TM Green Master Mix (#4385612; Applied Biosystems, Waltham, MA, USA) following the manufacturer’s protocol. The expression levels of target genes were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primer sequences are listed in Supplemental Table S1. Relative gene expression levels were calculated using the 2−^ΔΔCt method.

Liver tissues were harvested, fixed in 10% formalin, and embedded in paraffin. Embedded tissues were dehydrated through graded alcohol solutions. Sections (5 μm thickness) were prepared and stained with hematoxylin and eosin (H&E), then examined using light microscopy. The histological examination of liver tissues included assessment of the MASLD activity score (NAS), ranging from 0 to 8. This score integrates three critical histological parameters: lobular inflammation (score 0–3), steatosis (score 0–3), and ballooning degeneration (score 0–2) [26].

Mouse kidney specimens were fixed in paraformaldehyde and embedded in paraffin. Coronal sections of kidney tissues were prepared and stained using periodic acid–Schiff staining (Sigma-Aldrich, St. Louis, MO, USA). Histological images were scanned and analyzed using the Motic Digital Slide Assistant (Motic China Group Co., Ltd., Xiamen, China; 2017) to quantify the areas of glomeruli, mesangium, and renal tubular lumens. For each mouse, five glomeruli were randomly selected to measure average glomerular and mesangial areas, and five high-power fields were randomly chosen to calculate the mean tubular lumen fraction.

Total RNA was extracted from lung tissue samples of ND WT mice and SGPL1 KI mice using the RNeasy Mini Kit (Qiagen, Germantown, MD, USA), following the manufacturer’s protocol. RNA concentration and purity were assessed using a SimpliNano^TM Biochrom Spectrophotometer (Biochrom, Holliston, MA, USA). RNA integrity and degradation were evaluated using the Qsep 100 DNA/RNA Analyzer (BiOptic Inc., New Taipei City, Taiwan).

High-quality RNA samples were sent to BGI Americas (San Jose, CA, USA) for RNA-Seq transcriptome sequencing. Libraries were prepared using the KAPA mRNA HyperPrep Kit (KAPA Biosystems, Roche, Basel, Switzerland). Briefly, mRNA was isolated from total RNA using oligo-dT magnetic beads and fragmented using KAPA Fragment, Prime, and Elute Buffer. First-strand cDNA synthesis was performed using random hexamer primers, followed by second-strand synthesis incorporating deoxyuridine triphosphate to ensure strand specificity during sequencing. The resulting double-stranded cDNA was end-repaired, A-tailed, and ligated to adapters. cDNA fragments (300 to 400 bp) were purified using KAPA Pure Beads. Libraries were amplified using KAPA HiFi HotStart ReadyMix, and quality and size distribution were validated on the Qsep 100 DNA/RNA Analyzer.

Prepared libraries were sequenced on the Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, USA), generating 150 bp paired-end reads. Raw sequencing data were processed and analyzed by BGI Americas (Cambridge, MA, USA) using their bioinformatics pipeline. Differentially expressed genes (DEGs) were identified using the DESeq2 algorithm, with significance determined by a false discovery rate-adjusted P<0.05 and log₂-fold change (log₂FC) ≥1 or ≤−1. Functional enrichment analysis of DEGs was performed using the Gene Ontology (GO) and Disease Ontology (DO) databases to elucidate biological processes, cellular components, molecular functions involved in gene expression changes, and associations with diseases.

Data analysis was conducted using GraphPad Prism software version 9.0.2 (GraphPad, La Jolla, CA, USA). Results are expressed as mean±standard error of the mean or median with interquartile range (IQR). The Mann-Whitney U test was used for continuous variables not following a parametric distribution. Statistical significance was defined as a P value less than 0.05. One-way analysis of variance was conducted for comparisons involving three or more groups, with the Bonferroni correction applied for multiple comparisons.

To evaluate the impact of reduced S1P on T2DM, we conducted an experiment using WT and S1PL^C317A mutant (KI) mice. Plasma S1P concentrations were measured initially, and as expected, KI mice showed significantly lower S1P levels than WT mice (Fig. 1A). To induce diabetes mellitus, mice were fed an HFD, a widely used model for studying impaired glucose tolerance and T2DM. Four groups were established: WT mice on an ND, WT mice on an HFD, KI mice on an ND, and KI mice on an HFD. After 8 weeks, diabetes mellitus was successfully induced in the HFD-fed groups. Mice were sacrificed after 13 weeks of feeding. Throughout the experiment, mouse body weights were closely monitored. After introducing the HFD, KI mice had significantly higher body weights than WT mice as early as the first week. Over time, KI mouse body weights progressively increased with continued HFD feeding. Both groups demonstrated similar growth patterns, continuously gaining weight for the initial 12 weeks and plateauing thereafter (Fig. 1B). In contrast, the ND groups showed comparable weights throughout (Fig. 1B). Visually, HFD-fed S1PL^C317A KI mice appeared physically larger than HFD-fed WT mice (Fig. 1C). Final body weights for the HFD-fed KI mice (median 43.56 g [IQR, 42.41 to 48.05]) were significantly greater (P<0.05) compared to the HFD-fed WT mice (median 40.50 g [IQR, 37.6 to 45.82]).

An OGTT was conducted to compare the metabolic phenotypes among the four mouse groups. Blood glucose (Fig. 2A) and insulin (Fig. 2B) levels over time during the OGTT are shown in Fig. 2. Typically, glucose-tolerant mice exhibit a sharp increase in glucose levels, peaking at 15 to 30 minutes post-glucose challenge, followed by rapid clearance. WT mice fed an ND served as the glucose-tolerant control and displayed expected glucose kinetics, peaking around 371 mg/dL at 30 minutes and rapidly decreasing by 90 minutes, indicating efficient glucose clearance. Conversely, WT mice fed an HFD peaked at approximately 404 mg/dL with delayed glucose clearance, indicative of glucose intolerance. KI mice on an HFD reached a higher peak glucose level (approximately 454 mg/dL) and demonstrated a clearance pattern similar to HFD-fed WT mice, indicating exacerbated glucose intolerance. The AUC of the time-blood glucose curve of the OGTT of the HFD-fed S1P lyase knock-in (S1PL^C317A KI) mice was notably higher than that of the respective control group (Fig. 2C). Although glucose responses at 15 and 30 minutes post-glucose administration appeared similar between ND and HFD-fed WT groups, the calculated glucose AUC was higher in HFD-fed WT mice, though not statistically significant. However, KI mice showed a significantly greater effect of the HFD on glucose tolerance, suggesting increased susceptibility to glucose intolerance and warranting further investigation.

Insulin levels were also measured to explore underlying metabolic disruptions across groups. Mice on ND showed unchanged insulin responses, whereas HFD-fed mice exhibited significantly elevated insulin responses (Fig. 2B). The AUC of the time blood insulin curve of the OGTT was significantly greater in HFD-fed KI and WT mice compared to their ND-fed counterparts. Furthermore, the AUC of the time-blood Insulin curve of OGTT in HFD-fed KI mice significantly exceeded that of HFD-fed WT mice (Fig. 2D). Notably, ND-fed mice across both genotypes displayed comparable glucose and insulin levels, confirming appropriate baseline controls (Fig. 2A, B). HbA1c levels measured in all groups showed no significant differences (Fig. 2E). Plasma TG and NEFA were significantly higher in mice fed an HFD compared to ND-fed mice (TG: WT P<0.05, KI P<0.005; NEFA: WT P<0.05, KI P<0.005) (Fig. 2F, G). Although HFD-fed KI mice had higher plasma TG and NEFA levels than HFD-fed WT mice, these differences were not statistically significant (P=0.4814) (Fig. 2F, G).

The liver is an insulin-responsive organ crucial for glucose homeostasis, lipid metabolism, and the pathogenesis of T2DM. Hepatic steatosis, linked to T2DM and obesity, exacerbates insulin resistance in the liver [27,28]. Thus, we investigated S1P’s impact on hepatic steatosis. Liver sections from HFD-fed S1PL^C317A KI mice stained with H&E revealed greater lipid deposition than in HFD-fed WT mice (Fig. 3A). KI mice fed an ND exhibited baseline lipid accumulation comparable to WT mice. Further histological analysis employed the NAS, revealing significantly higher NAS scores in HFD-fed KI mice compared to HFD-fed WT mice (median 7.5 vs. 5.0, P<0.0001) (Fig. 3B). Hepatic TG concentrations were significantly increased in HFD-fed KI mice compared to HFD-fed WT mice (P<0.05) (Fig. 3C). Liver mRNA expression of lipogenesis-related genes (peroxisome proliferator-activated receptor γ [PPARG], apolipoprotein A2 [APOA2], apolipoprotein C4 [APOC4]) and a lipolysis-related gene (patatin-like phospholipase domain containing 2 [PNPLA2]) were analyzed. KI mice on HFD demonstrated upregulated lipogenesis genes and downregulated lipolysis genes, though only APOC4 reached statistical significance (P<0.0005) (Fig. 3D-G). These findings indicate that reduced S1P levels significantly enhance hepatic steatosis.

Histological examination showed that the two groups fed an ND exhibited normal glomerular morphology, characterized by intact capillary lumina and minimal mesangial expansion. Interestingly, KI mice on the ND showed significantly increased glomerular cross-sectional areas compared to WT mice on the same diet (Fig. 4A). WT mice fed an HFD readily developed diabetic nephropathy (DN), demonstrated by marked mesangial expansion within the glomerulus (Fig. 4A). KI mice fed an HFD exhibited significantly greater glomerular area and mesangial fraction compared to WT mice on the HFD (Fig. 4A). These findings suggest that reduced circulating S1P levels exacerbate mesangial hypertrophy in DN. Additionally, decreased S1P levels alone induce glomerular hypertrophy even in the absence of diabetes mellitus (Fig. 4B, C). Furthermore, mice fed an HFD displayed significant renal tubular dilation (Fig. 4D). KI mice on an ND did not exhibit increased renal tubular dilation compared to WT mice on the ND diet (Fig. 4D). However, KI mice on an HFD demonstrated significantly greater renal tubular dilation compared to HFD-fed WT mice (Fig. 4D, E). The differences observed between WT and KI mice on the same diet can be attributed to alterations in S1P levels, whereas the differences within the same mouse strain fed ND versus HFD reflect diabetes-induced effects. Collectively, these results suggest that decreased circulating S1P significantly exacerbates DN severity. Nonetheless, lower S1P alone appears sufficient to induce glomerular hypertrophy even without diabetes mellitus.

To compare gene expression between S1PL^C317A KI and WT mice, next-generation sequencing (NGS) was employed. A total of 4656 DEGs between KI and WT samples were identified, using a statistical significance threshold of P<0.05 and a log₂FC >1 for upregulated genes or log₂FC <−1 for downregulated genes. Specifically, 2,985 genes were upregulated, and 1,671 genes were downregulated. GO and DO enrichment analyses were conducted to characterize the functions of DEGs and identify associations with diseases. The enrichment of DEGs across various GO categories was analyzed, and the top 30 GO categories with the highest overall enrichment, determined by adjusted P values, were identified. The numbers of positively and negatively regulated genes within these categories were then reported.

Fig. 5 illustrates the top 30 GO terms, comprising 21 biological processes and nine cellular components. Among biological processes, the five most significantly enriched GO terms were: ‘ATP metabolic process,’ ‘purine ribonucleoside monophosphate metabolic process,’ ‘purine ribonucleoside triphosphate metabolic process,’ ‘ribonucleoside monophosphate metabolic process,’ and ‘ribonucleoside triphosphate metabolic process.’ For cellular components, the five most enriched GO terms included: ‘inner mitochondrial membrane protein complex,’ ‘respiratory chain complex,’ ‘respiratory chain,’ ‘mitochondrial respiratory chain,’ and ‘mitochondrial inner membrane.’ Overall, the GO enrichment analysis demonstrated that the top enriched categories primarily related to mitochondrial function and bioenergetics, with most of these genes downregulated in the S1PL^C317A KI group (Fig. 5). Notably, genes involved in cell division and DNA repair were generally upregulated in KI mice. DO, a standardized human disease ontology, was used to interpret interactions between DEGs and diseases. Based on the DO database, the 10 most enriched terms related to human diseases were identified, with ‘diabetes mellitus’ ranking highest. Within this category, 131 diabetes mellitus-related genes were found to be differentially expressed (Fig. 6, Supplemental Table S2).