Lipid signature changes of women with gestational diabetes mellitus in response to puerperal exclusive breastfeeding

Abstract

ABSTRACT

Objective

We here investigated whether lactation during puerperium could help to reverse the diabetogenic effect of gestation and further explored the lipid profiling changes upon breastfeeding.

Methods

Thirty‐five women diagnosed with GDM were recruited, and fasting plasma samples were collected at ~6 weeks postpartum. Maternal metabolic parameters were determined, and an untargeted lipidomic analysis was performed. The relationship between underlying lipidomic responses and lactation was explored.

Results

Improved glucose homeostasis and insulin sensitivity were observed in GDM women who adopted breastfeeding during the puerperium. Further lipidomics analysis revealed prominent correlations between lipid constitution changes and breastfeeding in women with GDM. A total of 766 lipid species were identified, 33 of which were found to be significantly altered in response to lactation. Significant associations between dysregulated lipids and maternal metabolic parameters were also shown. Subsequently, we identified a panel of three lipids that were strongly associated with breastfeeding, from which we constructed a predictive model with higher discriminating power.

Conclusions

We generally revealed that lactation during puerperium appears to have favorable effects on diabetogenic risk factors for GDM women. We also discovered that lipidomic changes related to lactation could elucidate the mother's recovery from GDM pregnancy.

Abbreviations

• AUC
• area under the curve
• BMI
• body mass index
• Cer
• ceramide
• FFQ
• food frequency questionnaire
• FPG
• fasting plasma glucose
• GDM
• gestational diabetes mellitus
• HDL
• high-density lipoprotein
• HOMA-IR
• homeostasis model assessment of insulin resistance
• IADPSG
• international association of diabetes pregnancy study groups
• IDF
• international diabetes federation
• LPC
• lysophosphatidylcholine
• MUFA
• monounsaturated fatty acid
• OGTT
• oral glucose tolerance test
• OPLS-DA
• orthogonal projections to latent structures-discriminate analysis
• PC
• phosphatidylcholine
• PE
• phosphatidylethanolamine
• PUFA
• polyunsaturated fatty acid
• ROC
• receiver operating characteristic
• SHexCer
• sulfur hexosylceramide hydroxyfatty acid
• SFA
• saturated fatty acid
• SM
• sphingomyelin
• T2D
• type 2 diabetes
• TAGs
• triacylglycerols
• VIP
• variable importance in the projection

INTRODUCTION

Gestational diabetes mellitus (GDM) is the most common metabolic disorder occurring during pregnancy, affecting ~16.7% of pregnant women worldwide, as estimated by the International Diabetes Federation (IDF) in 2021¹. Although most women with GDM commonly return to standard glucose tolerance after pregnancy, they still have a 7-fold higher risk of developing type 2 diabetes (T2D) later in life than normoglycemic pregnancies². About 5% of women with a history of GDM will progress to T2D within the first 6 months after pregnancy, and this burden expands to 10% within the subsequent 1–2 years postpartum³. Moreover, these women are also prone to developing non-alcoholic fatty liver and cardiovascular diseases that may lead to early mortality⁴. Therefore, it is critical to understand the physiological and metabolic changes during the transition from GDM to T2D.

Breastfeeding has many health benefits and is widely recommended by the World Health Organization and public health initiatives^5–7. The relationship between breastfeeding and lower rates of T2D in women with a history of GDM has been well demonstrated. Specifically, the duration of exclusive lactation has an inverse association with the low incidence of T2D and improved maternal glucose homeostasis among women with GDM⁸. Meta-analyses of the lactation protective effects have yielded summary estimates of 9%–11% for each additional year of lactation⁹. In prospective studies of lactation duration and incidence of T2D in women with GDM, 6 or more months of lactation contributed to a graded 25%–47% risk reduction in the progression to T2D^10,11. Another study proposed that 3 months or longer confers a 45% lower 15-year incidence of diabetes¹². However, these studies mainly focus on the relationship between long-term breastfeeding duration (>3 months) and the incidence of T2D in women with GDM. Very few studies have examined the favorable effects of a shorter breastfeeding time on improving maternal metabolism after delivery.

Lipid metabolism is significantly changed during pregnancy and contributes to the pathogenesis of GDM. The dysregulated lipid metabolism could be maintained after pregnancy even when glycemia returned to normal, and these alterations increase the risk of later T2D and cardiovascular disease¹³. A recent well-characterized prospective cohort study followed women with GDM for up to 8 years and found that lipidomic dysregulation during the postpartum in GDM women may be the underlying cause of the transition from no diabetes to incident T2D¹⁴. Breastfeeding behavior in the early postpartum period is critical in reversing the lipid profile in women with GDM. In the past decade, several pieces of research using longitudinal cohorts or animal models of GDM reveal that 3–6 months of lactation attributes the lower plasma triacylglycerols (TAGs) and higher high-density lipoprotein (HDL) in lactating women^15–17. Recently, Zhang et al. reported that intensive lactation could significantly alter the circulating lipid profile at early postpartum¹⁸. Women with GDM who metabolically fail to respond to lactation are more prone to develop T2D¹⁸. However, whether a shorter lactation duration in the primal postpartum period still has an early favorable recovery of the postpartum metabolomics profile was unclear.

To address this gap, we determined the association between breastfeeding and metabolic profiles in women with GDM pregnancy during the puerperium. The puerperium generally lasts 6 weeks and is the period of adjustment after delivery when the anatomic and physiologic changes of pregnancy return to the non-pregnant state. Thus, our objective was to investigate whether a 6-week lactation could help to reverse some of the diabetogenic effects of gestation. We also intended to construct a model of lipid markers that may facilitate the early prediction of the metabolic recovery of women with GDM after delivery.

MATERIALS AND METHODS

Study design

Women who delivered at the Women's Hospital of Nanjing Medical University (Nanjing, China) were enrolled in this study. All the participants received 75 g OGTT at 24–28 weeks of gestation and were diagnosed with GDM based on the IADPSG criteria: fasting blood glucose levels ≥5.1 mmol/L, 1-h glucose levels ≥10.0 mmol/L, or 2 h glucose levels ≥8.5 mmol/L. Pregnant women diagnosed with GDM would receive diet and exercise interventions to treat GDM and prevent further complications. Those who taking medications that affect glucose tolerance and lipid regulation were excluded. All participants had a routine postnatal health check-up after ~6 weeks postpartum to estimate the genital and bodily recovery after delivery. Plasma samples (fasting) were collected at this time. Finally, 20 women with GDM had exclusively breastfed, and 15 formula-feeding controls were screened out. All the subjects were matched by age, gestational weeks, and pregnancy BMI, and met the minimum sample size in each group calculated using PASS. The frequency and volume of exclusive human milk feeding and formula feeding for each woman were assessed by trained research staff via telephone calls, mailed surveys, and questionnaires during in-person visits. The baseline characteristics such as maternal age, gestational age, delivery mode, body mass index (BMI), and neonatal birth weight were obtained from electronic medical records with permission from participants. BMI was calculated as weight in kilograms divided by the square of body height in meters.

Dietary assessment

Dietary intake was assessed using a semi-quantitative food frequency questionnaire (FFQ) tool based on dietary guidelines¹⁹. Well-trained interviewers administered the FFQ according to the average habitual dietary intake. A total of 25 food items in the FFQ were categorized based on nutrient composition as follows: (1) rice, (2) porridge, (3) flour food, (4) desert, (5) fried food, (6) stuffing foods, (7) coarse grains, (8) tuber crop, (9) dairy and its products, (10) eggs, (11) red meat, (12) poultry, (13) processed meat products, (14) freshwater fishes, (15) seafood, (16) bean products, (17) nut fruits, (18) dark vegetables, (19) light vegetables, (20) mycorrhizae, (21) fruits, (22) beverages, (23) beer, (24) yellow rice wine, and (25) white wine. The frequency of consumption reported for each food was converted to daily portions and multiplied by the food's energy content.

Blood biochemical parameters

Blood samples obtained at 6 weeks postpartum were centrifuged at 3000 rpm for 10 min at 4°C. The supernatants were divided into 200 μL of each tube and stored in aliquots at −80°C for subsequent testing. Blood biochemical parameters, including fasting plasma glucose (FPG), insulin, TAG, and C-peptide, were determined using commercial ELISA kits according to the manufacturer's instructions. Specifically, the Glucose GLU Content Determination ELISA kit (Catalog# E1010-200) and TAG enzyme assay ELISA kit (Catalog# E1003-125) were purchased from Applygen Technologies (Beijing, China), while Insulin ELISA kit (Catalog# RAB0327, RRID: AB_2942043) and C-peptide ELISA kit (Catalog# CSB-E05066h, RRID: AB_2942044) were purchased from Sigma-Aldrich (St. Louis, MO, USA) and Cusabio (Wuhan, China), respectively. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as [fasting serum insulin (mIU/L) × FPG (mmol/L)/22.5] and was used to estimate insulin sensitivity.

Lipid extraction

Plasma samples were sent to Shanghai BioTree Biomedical Technology Co., Ltd. (Shanghai, China) for lipidomics analysis. To extract lipids from plasma, 100 μL of the sample was added to an EP tube, mixed with 480 μL of extract solution (methyl tertiary butyl ether: methanol = 5:1), and vortexed for 30 s. The mixture was then sonicated for 10 min in an ice-water bath, incubated at −40°C for 1 h, and centrifuged at 3000 rpm for 15 min at 4°C. Three hundred and fifty microliters of supernatant was transferred to a clean tube and evaporated to dryness in a vacuum concentrator at 37°C. Then, the dried sample was reconstituted in 100 μL of 50% methanol in dichloromethane by sonication for 10 min in an ice-water bath. The constitution was then centrifuged at 13,000 rpm for 15 min at 4°C, and 75 μL of supernatant was transferred to a fresh glass vial for further analysis. The quality control (QC) sample was prepared by mixing an equal aliquot of 20 μL of the supernatants from all the samples. Ultra-high-performance Liquid Chromatography coupled to Q Exactive Orbitrap Mass spectrometry (UHPLC-QE-MS) was performed to provide qualitative and quantitative information on the lipid components^20,21.

UHPLC-QE-MS analyses

The UHPLC-QE-MS analyses were performed using a UHPLC system (1,290, Agilent Technologies in conjunction with a Kinetex C18 column (2.1 × 100 mm, 1.7 μm, Phenomenex)). The mobile phase consisted of component A, which consisted of 40% water, 60% acetonitrile, and 10 mmol/L ammonium formate, and component B, which consisted of 10% acetonitrile and 90% isopropanol. Additionally, 50 mL of 10 mmol/L ammonium formate was added to every 1,000 mL of mixed solvent. The gradient elution method was applied as follows: 0 –1.0 min, 40% B; 1.0–12.0 min, 40%–100% B; 12.0–13.5 min, 100% B; 13.5–13.7 min, 100%–40% B; 13.7–18.0 min, 40% B. Two-microliter sample was injected for both positive and negative models via auto-sampler at 4°C. MS/MS spectra were obtained from the Q Exactive Orbitrap Mass Spectrometer on data-dependent acquisition mode following the previous strategy^20,21. Briefly, the electrospray ionization source was operated at +5,000 V in the positive ion model and −4,500 V in the negative ion model. The capillary temperatures were 320 and 300 °C, and the sheath gas and Aux gas flow rates were 30 and 10 Arb, in positive and negative mode, respectively. The full MS resolution was 70,000, the MS/MS resolution was 17,500, and the collision energy was 15/30/45 in NCE mode.

Data preprocessing and annotation

The raw data were converted to files in mzXML format using the “msconvert” program from ProteoWizard, and the CentWave algorithm in XCMS was then used for peak detection, extraction, alignment, and integration, the minfrac for annotation was set at 0.5, the cutoff for annotation was set at 0.3. When the X peaks were detected, and X metabolites were left after relative standard deviation de-noising, then, the missing values were filled up by half of the minimum value. Lipid identification was achieved through a spectral match using LipidBlast library, which was developed using R and based on XCMS. Results from raw LC–MS files were separately processed to extract profiles of positive and negative ionization modes. In this study, 569 lipid species in positive ion mode and 278 lipid species in negative ion mode were picked out, respectively. The details for lipid metabolite identification including retention time (RT), mass-to-charge ratio (m/z), molecular formula, identification score, and mass error for the identified metabolites were also provided (Data S1). A combination of both positive and negative data was performed to conduct statistical analyses. Data were first filtered by the relative standard deviation (RSD) of QC samples less than 20%, and duplicated lipids with low MS2 scores and peak area would be removed. After this filtering step, 847 lipid species were reduced to 766 for further analysis.

Bioinformatics analysis

The dataset containing the information on peak number, sample name, and normalized peak area was imported to the SIMCA16.0.2 software package (Sartorius Stedim Data Analytics AB, Umea, Sweden) for multivariate analysis. Data were scaled and logarithmically transformed to minimize the impact of both noise and high variance of the variables. After these transformations, supervised orthogonal projections to latent structures-discriminate analysis (OPLS-DA) were applied to visualize group separation and find significantly changed metabolites. Furthermore, the value of variable importance in the projection (VIP) of the first principal component in OPLS-DA analysis was obtained. It summarizes the contribution of each variable to the model. The metabolites with VIP > 1 and P < 0.05 (Student's t-test) were considered as significantly changed lipids.

Statistical analysis

Data were described by means and standard deviations (mean ± SD). Student's t-test was used to compare the differences in each parameter, and P-values <0.05 were considered statistically significant. Spearman correlation analysis was carried out to evaluate the correlation between significantly dysregulated lipids and clinical parameters. To identify independent predictors, plasma lipids showed significant differences between the breastfeeding and formula-feeding groups and were further adjusted in conditional logistic regression models with a stepwise method (including forward and backward). Prediction performance was presented as receiver operating characteristic (ROC) curves. Model evaluation in the hold-out testing set was presented as the area under the curve (AUC), sensitivity, and specificity. All the analyses above were performed in open-source, statistical software SPSS (Version 26.0), and MedCalc (Version 19.0). The power analysis was addressed by the software PASS (Version 11.0) to calculate the power of a statistical test with the given means, standard deviations, and sample size. Given the current sample size with 20 GDM women in the breastfeeding group and 15 in the formula-feeding group, we already have >80% power to detect a difference in FPG (0.86), insulin (0.91), HOMA-IR (0.84), and C-peptide (0.98) with a Type I error level of 0.05.

RESULTS

Puerperal exclusive breastfeeding improves maternal metabolism in women with GDM

The basal characteristics of 35 GDM participants are summarized in Table 1. In the prenatal period, there was no significant difference in age, gestational age, pre-pregnancy BMI, and BMI collected at enrollment between lactating and non-lactating women. Diabetic indicators that included 2-h 75 g OGTT glucose levels and glycosylated hemoglobin (HbA1c) also showed comparable performance among GDM women who breastfed or not. Additionally, we observed no difference in gestational weight gain and neonatal birth weight between the two groups. During the puerperium (~6 weeks postpartum), compared with controls, lactating women displayed a lower mean FPG, insulin, HOMA-IR, and C-peptide than non-lactating women after controlling for postpartum BMI and weight decrease (Figure 1). Further, no significant difference in dietary intake was found between these two groups during pregnancy and the postpartum period (Table 1). Taken together, these results support that short-term lactation during the puerperium still favors maternal metabolism in GDM women.

Table 1 Clinical characteristics of participants

	Breastfeeding (n = 20)	Formula feeding (n = 15)	P-value
Prenatal characteristics
Age (years)	29.90 ± 2.64	30.73 ± 2.68	0.119
Gestational age (weeks)	39+3 ± 1.73	39+1 ± 0.96	0.809
Pre-pregnancy BMI (kg/m2)	21.66 ± 2.31	23.45 ± 3.67	0.085
BMI collected at enrollment (kg/m2)	24.14 ± 2.63	26.08 ± 3.56	0.073
FPG (mmol/L)	4.97 ± 0.53	5.02 ± 0.65	0.769
1-h 75 g OGTT (mmol/L)	9.41 ± 1.72	10.16 ± 0.49	0.279
2-h 75 g OGTT (mmol/L)	9.00 ± 1.74	9.18 ± 1.26	0.809
HbA1c (%)	5.32 ± 0.48	5.18 ± 0.29	0.340
Gestational weight gain (kg)	5.71 ± 3.44	4.87 ± 3.31	0.470
Neonatal birth weight (g)	3,309.5 ± 453.60	3,512.00 ± 506.60	0.220
Energy intake (kcal/day)	1,614.16 ± 843.86	1,799.89 ± 792.71	0.513
Characteristics at ~6 weeks postpartum
BMI postpartum (kg/m2)	22.35 ± 2.56	24.29 ± 3.50	0.070
Weight decrease (kg)	7.97 ± 3.15	8.11 ± 3.11	0.899
Energy intake (kcal/day)	1,937.04 ± 750.49	2,084.84 ± 757.02	0.570

[IMAGE OMITTED. SEE PDF]

Lipid profiling changes in GDM mothers in response to breastfeeding

Lactation requires the mobilization of lipids for milk synthesis and may benefit the postpartum reversal of abnormal lipid status in women with GDM. To explore lipid profiling changes in GDM mothers in response to breastfeeding during the puerperium, a broad-spectrum lipidomics analysis was performed with plasma samples. A total of 766 lipids were detected in both ionization modes among the collected samples (Table S1)²². Of these lipids, 34.46% were from TAG, 25.72% from phosphatidylcholine (PC), 7.57% from sphingomyelin (SM), 6.27% from phosphatidylethanolamine (PE), 4.7% from Ceramide (Cer), and so on (Figure 2a). OPLS-DA was used to visualize the separability between breastfeeding and formula-feeding groups based on the lipid pattern. The score scatter plot of the OPLS-DA model indicated a distinct difference between samples from breastfeeding and formula-feeding GDM women (Figure 2b). Further, a volcano map was then applied to quantify the contribution of each lipid and variables that satisfied VIP values than 1 and P values < 0.05 were identified as statistical significance (Figure 2c). In particular, 33 lipid species (VIP > 1 and P < 0.05) were significantly altered between breastfeeding and formula-feeding groups, including 18 neutral lipids, 14 phospholipids, and 3 sphingolipids (Figure 2d and Table S2)²².

[IMAGE OMITTED. SEE PDF]

Overview of lactating-associated lipids during puerperium

We then constructed a bubble chart to illustrate the changes in the lipids within each subclass. It is not difficult to notice that TAG, PC, SM, LPC (lysophosphatidylcholine), and SHexCer (sulfur hexosylceramide hydroxyfatty acid) had undergone significant changes during puerperium following exclusive breastfeeding (Figure 3a). The distinct lipid species with a VIP >1.0 and P < 0.05 are summarized in Figure 3b,c. Of note, lipids clustered in the TAG category were downregulated in the breastfeeding group relative to the formula-feeding group, while lipids clustered in PCs were upregulated (Figure 3b,c). This is consistent with previous clinical research that blood TAG concentrations decline more rapidly within 3–6 months after delivery and these levels stabilize thereafter⁶.

[IMAGE OMITTED. SEE PDF]

Lipids in relation to improved maternal metabolic parameters

Lactation has lasting health benefits on biochemical risk factors (e.g., glucose, insulin, HbA1c) that determine subsequent development of T2D in GDM women. We, thus, performed a comprehensive correlation analysis and linked the 33 differently expressed lipids to the common metabolic parameters (FPG, insulin, HOMA-IR, and C-peptide) of GDM. Among these differential lipids, almost all the TAGs displayed a negative association with FPG (r = [−0.56 to −0.25]), while half of the dysregulated PCs showed a positive relation with maternal FPG (r = [0.05 to 0.53]; Figure 4). Besides, SHEexCer (d37:3) also exhibited a positive correlation with FPG. Moreover, as shown in Figure 4, we found that some TAG and PC family members showed moderate correlations with C-peptide. Specifically, TAG (16:0/18:0/18:1), TAG (18:0/18:0/18:1), and TAG (18:0/18:0/18:0) were negatively correlated with C-peptide, while PC (16:0/20:4), PC (18:0/20:4) and PC (16:0/22:5) were positively correlated with expression of C-peptide. Additionally, TAG (16:0/18:0/18:1) and TAG (18:0/18:0/18:1) were shown to have positive correlations with HOMA-IR; while PC (16:0/22:5) showed a negative correlation.

[IMAGE OMITTED. SEE PDF]

Association between breastfeeding and lipid biochemical configuration

Lipidomics profiling also provides comprehensive coverage of plasma lipids for us to help better understand lipid species' biochemical structure (i.e., chain length, numbers of carbon atoms, double bonds) with breastfeeding¹⁴. TAGs significantly altered by breastfeeding were clustered in the range of carbon atoms 48–60 and double bond 0–5, specifically with even carton atoms 48, 54, and 56 (Figure 5a). Coincidentally, PCs have different chain lengths, typically between 36 and 40 carbons, and most contained an even number of carbon atoms 36 and 38 (Figure 5b). We also revealed the relationship between lipid biochemical configuration and puerperal breastfeeding from the perspective of specific fatty acid chains (Figure 5c). For total fatty acids, four SFAs (saturated fatty acid; FA12:0, FA 17:0, FA 21:0 and FA 22:0), two MUFAs (monounsaturated fatty acid; FA 17:1 and FA 19:1), as well as two PUFAs (polyunsaturated fatty acid; FA 19:3 and FA 22:6), were negatively associated with puerperal breastfeeding. In comparison, three long-chain SFAs (FA 14:0, FA 24:0, and FA 26:0) and seven PUFAs (FA 18:2, FA 18:3, FA 20:2, FA 20:3, FA 24:2, FA 22:5, and FA 37:3) were positively associated with puerperal breastfeeding. When referring to classes, positively associated fatty acids were mainly from SMs and PCs including long-chain SFAs (C14-C16, C18, C24, and C26), MUFAs (C16, C18), and PUFAs (C18, C20, C22, and C24). In contrast, the negatively associated fatty acids mainly belonged to TAGs and two still belonged to PCs such as SFAs (C12, C15-18, and C21-22), MUFAs (C16-C19), and PUFAs (C19-20 and C22). Moreover, we found that lauric acid (FA 12:0), palmitic acid (FA 16:0), and stearic acid (FA 18:0) which were associated with pancreatic β cytotoxicity were significantly reduced in response to exclusive breastfeeding. However, we failed to identify a clear pattern associated with structures in the other dysregulated lipids.

[IMAGE OMITTED. SEE PDF]

Lipidomics signature to predict metabolic recovery of lactating GDM women

We then tested whether the 33 lipids that were significantly associated with puerperal exclusive breastfeeding have predictive properties to indicate maternal metabolism reversal. By using a binary logistic regression model, we found a combination of three lipids (one PC and two TAGs) showed an excellent ability to predict metabolic recovery of lactating GDM women (Figure 6a). When using these lipids alone, we achieved a median AUC (area under the curve) value of 0.783 (Figure 6b), which was slightly superior to the predication ability of common metabolic parameters FPG (AUC, 0.758) and HOMA-IR (AUC, 0.747). When we combined lipids with FPG and HOMA-IR to establish new models, the median AUC values reached 0.850 and 0.870, respectively. Startlingly, the classical clinic predictive parameter C-peptide showed a prediction power of AUC 0.888 which was improved to AUC 0.933 by adding lipids (Figure 6b,c). The predictive performance was strongly improved after combining any two clinical variables with the three-analyte signature. Notably, combining the three-lipid panel outcomes with FPG, HOMA-IR, and C-peptides, the discriminative power was significantly improved to AUC 0.933 (Figure 6b,c). These data confirmed that specific circulating lipids in women with GDM pregnancy could be applied to assist with routine clinical indicators to estimate favorable recovery of metabolism among women who breastfeed in the early postpartum period.

[IMAGE OMITTED. SEE PDF]

DISCUSSION

It is well recognized that a longer duration of lactation has favorable roles among women with a history of GDM and is associated with a lower risk of type 2 diabetes. However, data are lacking on the association between short-term lactation and benefits for mothers' health. Here, we demonstrated that short-term exclusive breastfeeding during puerperium could improve glucose homeostasis and insulin sensitivity in women with recent GDM. We also revealed that breastfed women showed clear differences in lipidome compared with non-lactating mothers and constructed a three-lipid model to predict postpartum recovery from GDM pregnancy.

Breastfeeding is recommended and encouraged for GDM mothers. Previous studies have reported favorable recovery of postpartum glucose homeostasis among women who adopt breastfeeding behavior in the early postpartum period²³. Moreover, a clinical study of six women with recent GDM showed that 3 months of breastfeeding in mothers with previous GDM is associated with improved beta-cell function without affecting adiposity²⁴. We performed biochemical analyses of metabolic parameters in GDM women to explore the health-promoting effects of breastfeeding during puerperium. Except for conventional glucose metabolism parameters, we specifically found that plasma C-peptide levels were remarkably reduced in the exclusively breastfeeding group. Insulin and C-peptide are secreted by β cells into the circulation by exocytosis in an equimolar ratio, but C-peptide is less impacted by liver metabolism and peripheral blood than insulin, making it a more accurate indicator of the islet β-cell activity²⁵. The current findings suggest that the protective effects of lactation occur immediately at the very early postpartum after lactation.

Maternal metabolic adaptations to lactogenesis are obvious to meet the large demand for lipids required for milk production. The diversion of metabolic fuels results in lower lipid concentrations despite the higher rate of lipolysis observed in lactating women. We noted significant differences in the type and content of lipids in GDM women induced by lactating during the early postpartum period. These results support our hypothesis that lactating has a favorable effect on maternal blood lipids that may reflect a less diabetic risk profile. The most striking findings were the downregulation of TAG metabolism and the upregulation of PC metabolism. TAG is the most abundant lipid in the human species and acts as the primary energy source. Increased TAGs have been independent risk factors associated with various diseases²⁶. This is consistent with previous clinical research that blood TAG concentrations decline more rapidly within 3–6 months after delivery and these levels stabilize thereafter⁶. Thus, the dynamic balance of TAGs in plasma is important for maintaining health.

In contrast, PCs appeared to be upregulated in GDM women with exclusive breastfeeding, which may directly affect the postpartum protective effect of GDM women. As cell membrane integrity, PC is linked to cell signal transduction, and lack of PCs may lead to impaired insulin receptor signal transduction and insulin resistance, leading to the occurrence of diseases such as diabetes²⁷. Specifically, we discovered that PC (38:5), an effective predictor of GDM, was also significantly decreased in the exclusive breastfeeding group, and showed great correlation with lower levels of FPG and HOMA-IR. Additionally, as a typical cis fatty acid of MUFA, oleic acid (18:1) is increased upon exclusive breastfeeding. Higher amounts of MUFA are hypothesized to improve blood sugar management by decreasing the risk of cholesterol oxidation, lowering the concentration of low-density lipoprotein, and reducing blood viscosity in the vessel wall of coagulation²⁸. In addition, the enhancement of linoleic acid (FA 18:2) and docosapentaenoic acid (FA 22:5) was especially noticeable among PUFA in the exclusive breastfeeding group, which belong to the ω-3 and ω-6 unsaturated fatty acid series and could help to reduce inflammatory cytokine secretion, reverse glucose intolerance, and regulate lipid metabolism²⁹. Thus, these findings indicated that PCs are essential to transitioning from pregnancy to lactation.

CONCLUSION

In general, our study provides a unique perspective on short-term breastfeeding and maternal metabolism recovery in women with recent GDM. We further linked these positive impacts to maternal lipid profiling changes upon short-term breastfeeding. However, some limitations in our study should be raised. First, while we distinguished lipid signature using untargeted lipidomics, these detected metabolites have not been externally validated, for example, using purified compounds. Second, our study was not designed to demonstrate the superiority of a lipid biomarker compared to glucose-factor-based prediction alone, thus we lacked data on quantifying and validating the identified metabolites in a larger cohort population. Last, although the energy intake was equal, we still could not rule out the effects of diet and exercise management during pregnancy and lactation. Nevertheless, our study provides direct evidence demonstrating that a short-term duration of breastfeeding still has favorable effects on the metabolic parameters of GDM mothers.

ACKNOWLEDGMENTS

This research was supported by funding from the National Natural Science Foundation of China (Grant Number: 81770866, 81900783, 82070879, 82170880, and 82470879), the Nanjing Medical Science and Technique Development Foundation (ZKX23041), and the Natural Science Foundation of Jiangsu Province (BK20241734).

DISCLOSURE

The authors declare no conflict of interest.

Approval of the research protocol: The study protocol was recorded in with identifier NCT05629403 and has been approved by the Ethics Committee of Women's Hospital of Nanjing Medical University (Ethics batch Number: 2020KY075).

Informed consent: All participants signed the ethical informed consent.

Approval date of registry and the registration no. of the study/trial: The approval date of Registry is 27 November 2022; the registration number is NCT05629403.

Animal studies: N/A.

PRIOR PRESENTATION

A part of this study was presented at the 20th World Congress of Gynecological Endocrinology Annual Meeting 2022.

DATA AVAILABILITY STATEMENT

The raw data that support the findings of this study are available from the corresponding author upon request.

References

Wang H, Li N, Chivese T, et al. IDF diabetes atlas: Estimation of global and regional gestational diabetes mellitus prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group's Criteria. Diabetes Res Clin Pract 2022; 183: [eLocator: 109050].

Johns EC, Denison FC, Norman JE, et al. Gestational diabetes mellitus: Mechanisms, treatment, and complications. Trends Endocrinol Metab 2018; 29: 743–754.

Vounzoulaki E, Khunti K, Abner SC, et al. Progression to type 2 diabetes in women with a known history of gestational diabetes: Systematic review and meta‐analysis. BMJ 2020; 369: [eLocator: m1361].

Ajmera VH, Gunderson EP, VanWagner LB, et al. Gestational diabetes mellitus is strongly associated with non‐alcoholic fatty liver disease. Am J Gastroenterol 2016; 111: 658–664.

Much D, Beyerlein A, Roßbauer M, et al. Beneficial effects of breastfeeding in women with gestational diabetes mellitus. Mol Metab 2014; 3: 284–292.

Gunderson EP. Impact of breastfeeding on maternal metabolism: Implications for women with gestational diabetes. Curr Diab Rep 2014; 14: 460.

Rameez RM, Sadana D, Kaur S, et al. Association of Maternal Lactation with Diabetes and Hypertension: A systematic review and meta‐analysis. JAMA Netw Open 2019; 2: [eLocator: e1913401].

Ley SH, Chavarro JE, Li M, et al. Lactation duration and long‐term risk for incident type 2 diabetes in women with a history of gestational diabetes mellitus. Diabetes Care 2020; 43: 793–798.

Jager S, Jacobs S, Kroger J, et al. Breast‐feeding and maternal risk of type 2 diabetes: A prospective study and meta‐analysis. Diabetologia 2014; 57: 1355–1365.

Gunderson EP, Hurston SR, Ning X, et al. Lactation and progression to type 2 diabetes mellitus after gestational diabetes mellitus: A prospective cohort study. Ann Intern Med 2015; 163: 889–898.

Gunderson EP, Lewis CE, Lin Y, et al. Lactation duration and progression to diabetes in women across the childbearing years: The 30‐year CARDIA study. JAMA Intern Med 2018; 178: 328–337.

Ziegler AG, Wallner M, Kaiser I, et al. Long‐term protective effect of lactation on the development of type 2 diabetes in women with recent gestational diabetes mellitus. Diabetes 2012; 61: 3167–3171.

Furse S, Fernandez‐Twinn DS, Chiarugi D, et al. Lipid metabolism is dysregulated before, during and after pregnancy in a mouse model of gestational diabetes. Int J Mol Sci 2021; 22: [eLocator: 7452].

Lai M, Al Rijjal D, Röst HL, et al. Underlying dyslipidemia postpartum in women with a recent GDM pregnancy who develop type 2 diabetes. Elife 2020; 9: [eLocator: e59153].

Knopp RHWC, Wahl PW, Bergelin R, et al. Effect of postpartum lactation on lipoprotein lipids and apoproteins. J Clin Endocrinol Metab 1985; 60: 542–547.

Knopp RHBR, Wahl PW, Walden CE. Effects of pregnancy, postpartum lactation, and oral contraceptive use on the lipoprotein cholesterol/triglyceride ratio. Metabolism 1985; 34: 893–899.

Gunderson EP, Lewis CE, Wei GS, et al. Lactation and changes in maternal metabolic risk factors. Obstet Gynecol 2007; 109: 729–738.

Zhang Z, Lai M, Piro AL, et al. Intensive lactation among women with recent gestational diabetes significantly alters the early postpartum circulating lipid profile: The SWIFT study. BMC Med 2021; 19: 241.

Lv S, Qin R, Jiang Y, et al. Association of maternal dietary patterns during gestation and offspring neurodevelopment. Nutrients 2022; 14: [eLocator: 730].

Jiang D, He J, Hua S, et al. A comparative lipidomic study of the human placenta from women with or without gestational diabetes mellitus. Mol Omics 2022; 18: 545–554.

Zhong H, Zhang J, Xia J, et al. Influence of gestational diabetes mellitus on lipid signatures in breast milk and association with fetal physical development. Front Nutr 2022; 9: [eLocator: 924301].

Lu HJL, Cui Y, Jiang C, et al. Lipid signature changes of women with gestational diabetes mellitus in response to puerperal exclusive breastfeeding. 2023.

Shub A, Miranda M, Georgiou HM, et al. The effect of breastfeeding on postpartum glucose tolerance and lipid profiles in women with gestational diabetes mellitus. Int Breastfeed J 2019; 14: 46.

McManus RM, Cunningham I, Watson A, et al. Beta‐cell function and visceral fat in lactating women with a history of gestational diabetes. Metabolism 2001; 50: 715–719.

Lappas M, Jinks D, Ugoni A, et al. Post‐partum plasma C‐peptide and ghrelin concentrations are predictive of type 2 diabetes in women with previous gestational diabetes mellitus. J Diabetes 2015; 7: 506–511.

Kotronen A, Velagapudi VR, Yetukuri L, et al. Serum saturated fatty acids containing triacylglycerols are better markers of insulin resistance than total serum triacylglycerol concentrations. Diabetologia 2009; 52: 684–690.

Fazakerley DJ, van Gerwen J, Cooke KC, et al. Phosphoproteomics reveals rewiring of the insulin signaling network and multi‐nodal defects in insulin resistance. Nat Commun 2023; 14: 923.

Imamura F, Micha R, Wu JH, et al. Effects of saturated fat, polyunsaturated fat, monounsaturated fat, and carbohydrate on glucose‐insulin homeostasis: A systematic review and meta‐analysis of randomised controlled feeding trials. PLoS Med 2016; 13: [eLocator: e1002087].

D'Angelo S, Motti ML, Meccariello R. Omega‐3 and omega‐6 polyunsaturated fatty acids, obesity and cancer. Nutrients 2020; 12: [eLocator: 2751].