Gestational diabetes mellitus (GDM) is a pregnancy-specific disorder of glucose metabolism, distinct from pre-existing diabetes, and is increasingly prevalent in China. It poses significant risks to maternal health, including preeclampsia, higher rates of cesarean delivery, and a markedly increased likelihood of developing type 2 diabetes mellitus (T2DM) later in life. Offspring of affected mothers are also at risk of adverse outcomes such as macrosomia, neonatal hypoglycemia, and long-term metabolic complications (1–3). Established risk factors for GDM include advanced maternal age, obesity, family history of diabetes, and genetic susceptibility (4).

Although the pathophysiology of GDM is not fully understood, it shares essential features with T2DM, including insulin resistance, impaired glucose tolerance, and β-cell dysfunction (2, 3). Increasing evidence points to the contribution of genetic variation, particularly single-nucleotide polymorphisms (SNPs), in modulating GDM risk. Among candidate loci, the macrophage migration inhibitory factor (MIF) gene and the ARAP1 locus have been implicated in diabetes-related traits. MIF rs1007888 has been associated with insulin resistance and β-cell dysfunction (5, 6), while ARAP1 rs1552224 has been linked to reduced insulin secretion (7). However, their roles in GDM remain insufficiently investigated.

MIF is a pleiotropic cytokine expressed abundantly by placental trophoblasts and upregulated in GDM placental tissue, correlating with fasting glucose levels and insulin resistance (8, 9). Functionally, MIF influences insulin secretion and glucose metabolism and exerts pro-inflammatory effects by regulating cytokines such as TNF-α, IFN-γ, and IL-10, thereby contributing to hyperglycemia (9). Elevated systemic MIF levels are a feature of T2DM (10), but its mechanistic role in pregnancy-related glucose dysregulation remains poorly defined (8).

Meanwhile, genome-wide association studies (GWAS) consistently implicate ARAP1 variants in T2DM susceptibility (11, 12). In particular, rs1552224 has been associated with fasting glucose levels and impaired glucose-stimulated insulin secretion, with evidence suggesting the A allele contributes to β-cell dysfunction (11). Interestingly, preliminary data indicate that this allele may be protective against GDM (13), highlighting potential differences between pregnancy-related and non-pregnancy-related diabetes risk mechanisms.

Despite these insights, data on the association between MIF rs1007888 and ARAP1 rs1552224 with GDM remain sparse, especially in Asian populations, and no systematic synthesis of existing findings has been undertaken. Given the high burden of GDM and the distinct genetic architecture of the Chinese population, further investigation is warranted.

Therefore, in this study, we examined the associations of MIF rs1007888 and ARAP1 rs1552224 with GDM in a Han Chinese cohort and performed a meta-analysis to contextualize our findings with the broader literature. We aimed to clarify the potential contribution of these variants to GDM risk and provide new insights into their role in pregnancy-specific glucose metabolism.

Baseline characteristics from our cohort showed that women who developed gestational diabetes mellitus (GDM) had significantly higher systolic and diastolic blood pressure, fasting plasma glucose, 1-hour and 2-hour postprandial glucose levels, pre-pregnancy BMI, and mean age compared with controls (all p < 0.05, several p < 0.001). These observations are consistent with previous reports and point to several mechanistic underpinnings (18, 19).

Although none of the GDM participants met the diagnostic threshold for hypertension, their higher blood pressure values suggest early manifestations of insulin resistance–related endothelial dysfunction and subclinical inflammation (20). This pathophysiological background helps to explain the increased risk of preeclampsia and gestational hypertension in GDM pregnancies (4, 21).

The most pronounced group differences were seen in postprandial glucose excursions particularly at 1 and 2 hours supporting the concept that postprandial hyperglycaemia represents the core metabolic abnormality of GDM (4). This likely reflects inadequate first-phase insulin release and reduced peripheral glucose uptake.

Pre-pregnancy BMI emerged as another critical determinant. Even within the “normal” range (18.5–24 kg/m²), women who later developed GDM had higher pre-BMI values, suggesting that mild elevations in adiposity may confer risk. Adipokines such as leptin and TNF-α, secreted from adipose tissue, may further aggravate insulin resistance and impair β-cell function (22, 23).

Maternal age was also independently associated with GDM: the mean age of affected women was 31 years compared with 29 years in controls (p < 0.001), and stratified analysis showed that those aged ≥30 years had more pronounced metabolic impairments. This is consistent with evidence that both insulin sensitivity and β-cell function decline progressively with advancing age (4, 24).

Collectively, these findings underscore that even modest elevations in blood pressure, adiposity, and maternal age contribute to GDM risk. They highlight the need for pre-pregnancy counselling and weight optimisation in women of reproductive age as a preventive strategy to reduce the burden of GDM and its complications.

The human macrophage migration inhibitory factor (MIF) gene, located on chromosome 22q11.2, is constitutively and abundantly expressed in immune cells, particularly T lymphocytes and macrophages (25). MIF exerts diverse biological functions, and altered expression has been reported in several inflammatory and metabolic disorders, including rheumatoid arthritis, atherosclerosis, and diabetes (26). In the placental decidua, macrophages and dendritic cells regulate T-cell activation, initiating immune responses. Subsequent differentiation of T cells into Th1 or Th2 subtypes—driven by both intrinsic and extrinsic factors alters cytokine synthesis and distribution. These shifts directly influence MIF regulation and are essential for immunological adaptation across pregnancy stages (15).

Beyond its role in immunity, MIF functions as a pleiotropic cytokine linking inflammation to glucose metabolism. At the molecular level, it regulates glucose transporter 4 and fructose-bisphosphatase 2 expression (8). Moreover, MIF promotes TNF-α production (27, 28), thereby influencing glycolysis and inflammatory signalling. In adipocytes, MIF-driven TNF-α secretion contributes to insulin resistance (8), while in endothelial cells, MIF disrupts insulin-mediated nitric oxide release, fostering endothelial insulin resistance (8).

These mechanisms align with the lipid overload hypothesis, which posits that ectopic lipid deposition in insulin-sensitive tissues such as skeletal muscle, pancreas, and liver underpins insulin resistance (29–31). Supporting this, both adipose tissue MIF mRNA and circulating MIF concentrations are elevated in obese women with type 2 diabetes (32). Elevated adipose-derived MIF in obesity and diabetes further impairs insulin signalling and amplifies pro-inflammatory cascades, thereby reinforcing insulin resistance (33).

Despite previous findings linking MIF variants to T2D, our study found no significant association between the MIF rs1007888 variant and GDM. This divergence suggests that pregnancy-specific genetic or physiological factors might influence the role of this genetic locus in GDM development (6, 32, 33). Moreover, despite the known involvement of MIF in inflammatory and metabolic disorders (34), its genetic influence via rs1007888 on GDM risk within the context of higher pre-BMI appears to be negligible or absent in our specific cohort. This suggests that the genetic impact of rs1007888 on GDM risk may be either minor or does not significantly interact with the metabolic environment of pre-existing obesity, at least in the population we investigated. It is plausible that other genetic or environmental factors exert a more predominant influence on GDM development within this particular BMI subgroup.

ARAP1, which encodes the protein Arap1 (ARF-GAP, Rho-GAP Anchoring Protein Repeat Sequence, and Pleckstrin Homology Domain Protein 1, also known as centaurin delta 2), functions as a phosphatidylinositol 1,4,5-trisphosphate-regulated Arf GTPase-activating protein. This protein primarily targets the small GTP-binding protein Arf6, which is known to play a critical role in the modulation of insulin exocytosis (6). The rs1552224 SNP is situated within an intronic region of the ARAP1 gene.

Research indicates that ARAP1 is significantly associated with elevated blood sugar levels and reduced insulin secretion in response to glucose stimulation, suggesting that impaired pancreatic β-cell function may be a mechanism through which this genetic locus contributes to the development of diabetes (35). Recent findings propose that the expression of the ARAP1 locus is predominantly influenced by markedly increased levels of STARD10 expression in the pancreas (36).

The genetic variant STARD10 rs11603334 is functionally linked to the expression of ARAP1 and is in complete linkage disequilibrium with rs1552224. The risk allele (C) of STARD10 rs11603334 for type 2 diabetes is positioned near the ARAP1 promoter. This specific placement interferes with the collaborative complex formation between the transcriptional regulators PAX6 and PAX4. This interference leads to an increased amplification of ARAP1 P1 promoter transcription, resulting in higher ARAP1 transcription levels within the Islets of Langerhans (37).

Conversely, research by Carrat et al (36) suggests that compromised insulin secretion from pancreatic and β-cells is associated with decreased STARD10 levels, rather than elevated ARAP1. Furthermore, the rs1552224C allele might offer a protective effect against type 2 diabetes. This protective role could be mediated by the overexpression of STARD10, which in turn enhances Ca²⁺ dynamics and glucose-triggered insulin release from pancreatic and β-cells (34). Despite these findings, the precise operational mechanism of this process in GDM requires further extensive investigation.

The protective influence of the C allele at rs1552224 warrants particular attention. ARAP1 (ArfGAP with RhoGAP domain, ankyrin repeat, and PH domain 1) plays a critical role in insulin secretion and glucose homeostasis by regulating membrane trafficking and cytoskeletal dynamics within pancreatic β cells (38). While previous research has linked ARAP1 polymorphisms to an elevated risk of type 2 diabetes (39, 40), its specific contribution to GDM, especially in younger populations, remains underexplored. Our findings suggest that the C allele may modify ARAP1’s function, potentially enhancing β-cell compensation or insulin sensitivity to meet the increased metabolic demands of pregnancy, thereby lowering GDM risk. This aligns with existing evidence that genetic variations affecting β-cell function are significant predictors of GDM susceptibility (41).

In the context of the allelic model for the ARAP1 gene, the presence of each C allele is associated with a 38% reduction in the risk of developing GDM. This observation is consistent with the established role of ARAP1 in insulin granule trafficking and the functionality of βcells (37). Notably, the absence of CC homozygotes among GDM cases (0% compared to 0.8% in controls) indicates a potential dosage-dependent protective effect. However, the low minor allele frequency (MAF = 5–8%) (42) may limit the statistical power of analyses based on recessive models.

Furthermore, our results corroborate previous functional studies that demonstrate that ARAP1 knockdown adversely affects glucose-stimulated insulin secretion (GSIS) in β cells Benjamin (31). The protective C allele may enhance the expression of ARAP1, thereby improving the first-phase insulin response, which is crucial for regulating postprandial glucose levels. This is consistent with the elevated 1-hour and 2-hour post-glucose challenge levels presented in Table 1 .

The observation that the rs1552224 C allele exhibits a protective association in individuals with elevated pre-BMI is a compelling finding. Given that obesity is a well-established driver of insulin resistance and a significant risk factor for GDM (43), our data imply that the C allele may offer a unique advantage by ameliorating the adverse metabolic consequences of obesity during gestation (38).

The SNP rs1552224 has the potential to enhance the prediction of GDM risk in women with high BMI and nulliparous status. However, its modest effect size, with an odds ratio of approximately 0.6, limits its utility as a standalone predictive marker. We hypothesize that this allele could modulate ARAP1 function in a manner that either bolsters β-cell compensatory mechanisms or enhances insulin signaling pathways. This modulation would be particularly beneficial in counteracting the heightened insulin resistance characteristic of a higher pre-BMI, especially within a population already predisposed to GDM due to their initial weight. This interpretation aligns with the broader concept that genetic variants can influence an individual’s susceptibility to obesity-induced metabolic dysfunction (44).

A comparison of findings in Table 3 with the data in Table 5 , which utilized age as a stratification factor, indicates that the protective effect of the rs1552224 C allele is consistently observed across various risk stratifications, including age and BMI. This consistency enhances the potential clinical significance of this genetic variant as a biomarker for assessing the risk of GDM. Our research highlights the critical interaction between genetic factors and environmental influences, such as pre-BMI, in shaping susceptibility to GDM. By identifying genetic variants that may either increase or decrease the risk associated with obesity, we can develop more tailored strategies for the prevention and management of GDM. Future research should investigate the functional implications of rs1552224 and its potential role in targeted interventions for pregnant women with elevated BMI.

This meta-analysis provides a comprehensive assessment of the association between the rs1007888 variant and gestational diabetes mellitus (GDM), synthesising evidence from three independent studies (5, 45) together with data from Zhang et al. (our current study, 2024). The pooled analyses, illustrated in the accompanying forest plots (A–F), consistently demonstrate an increased risk of GDM associated with this locus, thereby offering new insights into the genetic basis of disease susceptibility.

In particular, Figure 1 highlights that Plots A and B yielded statistically significant associations, with pooled odds ratios of 1.48 and 1.45, respectively. Importantly, both models displayed very low between-study heterogeneity (I² = 0.0% for A and 17.6% for B), underscoring the stability and reproducibility of these findings. Such homogeneity suggests that the genetic effect captured in these models—likely reflecting an additive or dominant contribution of the risk allele—represents a robust and biologically meaningful determinant of GDM.

The consistency of associations across both Chinese cohorts (5, 45) and an Iranian cohort (6) further strengthens the evidence, supporting the notion of a shared genetic predisposition that transcends ethnic boundaries. This aligns with prior meta-analyses reporting that common genetic variants contribute significantly to GDM risk across diverse populations (26, 41). The low heterogeneity observed in several key models adds to the reliability of our effect estimates and suggests that rs1007888 may represent a stable genetic marker of GDM susceptibility.

Our study has several significant limitations that should be acknowledged. First, the relatively small number of studies included in the meta-analysis restricts the strength of the conclusions. This underscores the need for larger, multi-ethnic investigations to validate our findings and to explore potential gene-environment interactions. Functional studies will also be essential to elucidate the biological mechanisms by which these variants may contribute to GDM risk. Ultimately, identifying reliable genetic markers could inform personalized risk stratification, early screening, and targeted prevention strategies for GDM.

Discrepancies were observed between the results of our case–control analysis, the meta-analysis, and previous literature. We propose that a limited sample size represents a key contributor to these inconsistencies. Power calculations using G*Power software indicated that approximately 19,788 participants would be required to assess the association with rs1007888 adequately, and 2,273 participants for rs1552224, under the following parameters: test family = Exact; statistical test = Proportions: Inequality, two independent groups (Fisher’s exact test); test direction = two-tailed; power = 0.8; significance level α = 0.05.

The challenges of sample size requirements differ between variants. For rs1552224, the low minor allele frequency (0.079) limits statistical power, necessitating large cohorts to detect associations. For rs1007888, although the reference allele is more common (p2 = 0.486), the effect size is minimal (OR = 0.923), which again demands a substantial cohort to achieve reliable power. This highlights a broader difficulty in genetic association studies of weak-effect variants, where enormous sample sizes are often required to detect modest associations.

In addition to sample size, the absence of comprehensive adjustment for potential confounders may have contributed to the discrepancies observed. Significant clinical and biochemical variables such as lipid profiles, fasting glucose, HOMA indices, C-peptide concentrations, and comorbidities were not systematically included. Moreover, the lack of multicenter data collection limits generalizability, as regional differences within China may influence allele frequencies and disease risk. Addressing these issues in future studies through larger sample sizes, standardized adjustment factors, and multicenter recruitment will be crucial to resolving these inconsistencies and advancing understanding of the genetic basis of GDM.

This study investigated the relationship between the MIF rs1007888 and ARAP1 rs1552224 genetic variants and the risk of GDM. Our findings revealed no significant association between the MIF rs1007888 variant and GDM risk. ARAP1 rs1552224 was significantly linked to reduced GDM incidence. In contrast, meta-analysis results indicated that the MIF rs1007888 mutation was associated with an increased likelihood of developing GDM. Furthermore, the ARAP1 rs1552224 mutation was linked to a heightened risk of GDM. The discrepancies between our findings and those of previous studies may be attributed to limitations in sample size and ethnic diversity. To enhance the understanding of the relationship between these genetic variants and GDM risk, future research should prioritize increasing sample sizes and incorporating a broader range of adjustment factors and diverse ethnicities.