Ovarian cancer (OC), a highly lethal malignancy, significantly impacts the female reproductive system and is associated with high mortality rates globally. It is frequently diagnosed in advanced stages due to the general absence of early symptoms and the lack of effective screening strategies [3, 19]. Liver enzymes are routine biomarkers in physical examinations and are associated with a variety of diseases. They serve as potential biomarkers for the early diagnosis of OC. Consequently, this study employed Mendelian randomization analysis, which is a statistical method that uses genetic variants as instrumental variables (commonly SNPs) to infer causal relationships between an exposure and an outcome, to identify enzymes with a significant causal relationship to OC. SNPs can influence gene expression through multiple mechanisms such as altering protein sequence/function via coding region SNPs, modulating transcriptional activity by regulatory region SNPs, and affecting RNA splicing patterns from splice site SNPs, and so on [20, 21].

Further analysis was conducted at the single-cell level to examine the expression of genes encoding exposure factors within cells, providing a preliminary investigation into the regulatory mechanisms affecting OC. Additionally, retrospective clinical data were utilized for validation purposes.

The MR analysis results from this study revealed a causal relationship between ALP and AST with OC, and both were identified as protective factors. As a univariate factor, AST was significant, but in the multivariate analysis, its effect was corrected by ALP, transitioning from a significant risk factor to a non-significant factor. In other words, this suggests that the impact of ALP as a risk factor for OC is more direct compared to the influence of AST. While some degree of heterogeneity (p-value < 0.05) existed in the MR analysis, the IVW method used in this study requires the SNPs to fully meet the three principles of MR research in order to obtain a correct causal estimate. Therefore, the use of this method has not had a significant impact on the MR results. In the future, we will also conduct more analyses from various perspectives, such as subgroup analysis, to identify potential confounding factors, and adopt appropriate corrective measures to ensure the validity and accuracy of the final experimental conclusions.

ALP is a zinc-dependent dimeric metalloenzyme. Research into the enzymatic characteristics of ALP has elucidated that the active site encompasses a serine residue, with a proximate amino acid sequence reminiscent of serine environments [22]. In healthy adults, serum ALP levels are conventionally found to span from 40 to 150 U/L; deviations from this range are frequently indicative of pathological conditions [23]. Elevated serum ALP levels, for instance, have been correlated with Multiple Myeloma and osteoblastic bone tumors [24, 25]. The human body synthesizes a variety of amino transferases, with AST being predominantly localized in the mitochondrial cytoplasm of hepatocytes. Clinically, elevated serum AST levels can signify several conditions, encompassing viral hepatitis, alcoholic liver disease, cirrhosis, cholestatic syndrome, acute myocardial infarction, or skeletal muscle damage [26]. Recent investigations have uncovered a correlation between ALT/AST ratios and insulin resistance, particularly in early to middle-aged women [27]. Nevertheless, the potential causal link between the two liver enzymes’ elevation and OC remains to be definitively elucidated.

In our study, females with elevated levels of ALP had a relatively lower risk of developing OC. Conversely, elevated levels of ALT and GGT were associated with a higher risk of OC. And the clinical data have also substantiated discernible differences in ALP levels in patients up to two years prior to an OC diagnosis. The ALP levels measured two years before the diagnosis in OC patients were found to be lower than those in the healthy population during a comparable timeframe. Interestingly, ALT and GGT showed higher levels in pre-OC patients than those in the healthy population. Moreover, an inverse correlation was observed between the levels of ALP and the ovarian cancer biomarker CA125. In several cancer types, including colorectal cancer (CRC), breast cancer, and non-small cell lung cancer (NSCLC), low levels of alkaline phosphatase (ALP) have been associated with a worse prognosis for patients [28]. Some studies have observed a relationship between low ALPL expression and chemo-resistance of high-grade serous ovarian cancer (HGSOC) cells to paclitaxel. Low expression of ALPL was found to be inversely related to the FIGO stages and histological grades in a cohort of 90 patients with serous ovarian cancer (SOC). Moreover, they demonstrated that ALPL overexpression might decrease migration and invasion of HGSOC cells by inhibiting the WNT5A-FZD2-STAT3 signaling axis [29]. AST is encoded by GOT2. GOT2 can assume a tumor-suppressive role in certain oncogenic contexts, particularly in hepatocellular carcinoma. However, the precise mechanisms by which GOT2 exerts its effects in OC remain unclear at this time.

The single-cell data analysis reveals the cellular composition of ovarian tissue, including macrophages, ovarian cells, endothelial cells, epithelial cells, tumor fibroblasts, T cells, and others. AST is encoded by genes such as GOT1, GOT2, while ALP is encoded by ALPL, ALPP, ALPI, ALPG, with these genes primarily expressed in epithelial cells. Furthermore, genes associated with TWIK-related alkaline pH activate, Hydroxycarboxylic acid-binding receptor, Intracellular oxygen transport, Regulation of thyroid hormone activity, Sterols are 12-hydroxylated by CYP8B1 are enriched in epithelial cells.

The cell communication results indicate that endothelial cells have the most interactions with epithelial cells. Additionally, receptors and ligands associated with epithelial cells and having a p < 0.05 include TIMP1_FGFR2 between CAF (Cancer-Associated Fibroblast) and Epithelial cells, FGFR2_CD83 between Epithelial and Macrophage cells, C5AR1_RPS19, CD74_APP, CD74_COPA, and CD74_MIF between Macrophage and Epithelial cells, and CD74_MIF between T cells and Epithelial cells. Furthermore, clustering analysis of epithelial cells (divided into two clusters) and pseudotime analysis reveal three distinct differentiation states in epithelial cells, with Cluster 2 (C2) showing a higher degree of differentiation compared to Cluster 1 (C1). Subsequently, transcription factors encoding the exposure factors' genes were predicted using online databases. Finally, our pathological examinations of tissue samples revealed that the expression of ALPL and GOT2 in cancerous tissues was lower than that in adjacent non-tumorous tissues.

Recent literature extends the significance of ALP to oncological contexts. Rao et al. demonstrated that an ALPL knockdown could attenuate migration in prostate cancer cell lines [30]. Concurrently, another investigation posited that ALPL specifically mitigates lung adenocarcinoma (LUAD) cell metastasis by interacting with the p-ERK/c-Myc/RhoA signaling axis [31]. Moreover, overexpression of ALPL was observed to curtail migration and invasion in high-grade serous ovarian cancer (HGSOC) cell models [29].

GOT2 has been implicated in the augmentation of tumorigenicity. A recent study suggests that cutaneous melanoma (CM) patients with heightened GOT2 expression exhibit diminished survival rates and reduced immune cell infiltration [32]. Concurrently, elevated GOT2 levels have been linked to the accelerated proliferation of breast cancer cells [33]. Paradoxically, GOT2 can also assume a tumor-suppressive role in certain oncogenic contexts. For instance, GOT2 expression is inversely regulated in hepatocellular carcinoma (HCC) tissues, where its downregulation is associated with adverse prognostic outcomes in HCC patients. Functionally, GOT2 silencing has been shown to enhance proliferation, migration, and invasion of HCC cell lines [34]. While the underlying mechanisms of GOT2's dualistic role remain under active investigation, current evidence points to its involvement in the reprogramming of glutamine metabolism, a process that ostensibly supports cancer progression.

Given these findings, the emerging evidence concerning the role of ALPL and GOT2 in oncological pathophysiology, particularly with respect to their influence on OC, necessitates additional research. This is imperative to reconcile current discrepancies in the literature and to provide a deeper understanding of the molecular mechanisms at play.

However, it is noteworthy that a single biomarker often cannot fully account for a complex health condition, and typically, a comprehensive assessment requires the integration of multiple indicators and clinical information. This study was limited by a relatively small sample size and an observational period of only two years, which may introduce a certain degree of bias. In addition, the patients selected for clinical trials are of a wide age, and heterogeneity in this segment of the population may affect the study results. Further prospective studies should entail multi-center, long-term studies to expand the sample size and extend the duration of observation. Moreover, the underlying mechanisms by which ALP and GOT2 are involved in ovarian cancer (OC) pathogenesis warrant further investigation through in-depth cellular and animal studies.

In conclusion, our study leveraged MR analysis, single-cell data, and clinical data analysis to provide a theoretical basis for the mechanisms associated with ALP and AST in ovarian cancer.