Abstract
The gut mycobiota has never been studied either during pregnancy or in patients with gestational diabetes (GDM). This study aimed to analyze the fecal mycobiota of GDM patients during the second (T2) and third (T3) trimester of pregnancy and to compare it with the mycobiota of pregnant normoglycemic women (controls). Forty-one GDM patients and 121 normoglycemic women were studied. GDM mycobiota was composed almost exclusively by the Ascomycota phylum; Basidiomicota accounted for 43% of the relative frequency of the controls. Kluyveromyces (p < 0.001), Metschnikowia (p < 0.001), and Pichia (p < 0.001) showed a significantly higher frequency in GDM patients, while Saccharomyces (p = 0.019), were more prevalent in controls. From T2 to T3, a reduction in fungal alpha diversity was found in GDM patients, with an increase of the relative frequency of Candida, and the reduction of some pro-inflammatory taxa. Many associations between fungi and foods and nutrients were detected. Finally, several fungi and bacteria showed competition or co-occurrence. Patients with GDM showed a predominance of fungal taxa with potential inflammatory effects when compared to normoglycemic pregnant women, with a marked shift in their mycobiota during pregnancy, and complex bacteria-fungi interactions.
Similar content being viewed by others
Introduction
The human mycobiota is a neglected component of the microbiome, including all the different fungal species living in the human host1,2,3. More than 300 different fungal species are living in the human digestive tract with predominance of the Ascomycota, Basidiomycota and Zygomycota phyla4 and, among the genera, Candida, Paecilomyces, Penicillum, Aspergillus, Trichosporon, Rhodotorula, Cladosporium, Aureobasidium, Saccharomycetales, Fusarium and Cryptococcus1,5,6. Overall, there is a low fecal fungal diversity with a high degree of within- and between-subjects variability7. Most species derive from the oral cavity or from diet, being natural food contaminants, and dietary habits, age, gender, and drugs influence the mycobiota composition6,8. A low fungal abundance has been suggested to characterize the gut of healthy individuals9, but, on the other hand, higher gut fungal abundance has been reported to favorably influence childhood growth10, and the gut mycobiota may play beneficial effects for the human host, such as nutrient extraction, vitamin production, modulation of metabolism, and immunoregulation and defense against pathogens3,11,12. Conversely, an unbalance in the gut fungal composition (“fungal dysbiosis”) and colonization by fungal species has been reported in specific diseases, above all in immunocompromised patients1,5. Recently, a growing number of papers is reporting associations among specific gut fungal genera and chronic diseases, such as Alzheimer’s disease13, irritable bowel syndrome (IBS)14, inflammatory bowel diseases (IBD)2,4,15,16,17, alcoholic liver diseases18, colorectal adenoma and cancer19,20, the progression of pancreatic cancer 21, mold-induced asthma 22, bronchiectasis 23, carotid atherosclerosis 24, diabetes mellitus 25,26,27, and even weight loss 28 as well as many other post-surgical, nosocomial conditions or immunocompromise-induced diseases 3,4.
It is well known that the human microbiome changes in physiological conditions, such as pregnancy, and might play a role in gestational dysmetabolic conditions as well, for example, gestational diabetes mellitus (GDM)29. Indeed, the potential role of the gut mycobiota in GDM remains largely unexplored. Similarly, very few and contrasting data are available about the associations between dysmetabolic conditions and the mycobiota. Relationships have been reported between Candida and diabetes mellitus25,26, but not between Candida and obesity28.
The aims of the present observational study were, therefore, (i) analyzing the fecal mycobiota of patients with GDM during the second and the third trimester of pregnancy and (ii) comparing it with the mycobiota of pregnant normoglycemic women.
Results
Out of 50 GDM patients, 9 women were lost at follow-up or did not return the stool sample. Out of 150 normoglycemic women, 29 women did not return the stool sample or collected and/or stored the stool sample incorrectly. The baseline clinical characteristics of the participants did not differ from those of the women who were lost (data not shown). Therefore, 41 GDM patients and 121 normoglycemic women were analyzed. Their clinical characteristics are reported in Supplementary Table S1. Most GDM patients were overweight women, with excessive fat intake (> 40% total daily energy instead of the recommended < 35%) and low fiber consumption (15 g/day, i.e., ~ 9 g/1000 kcal instead of the recommended 12.6–16.7 g/1000 kcal). From enrolment (T2) to the study end (T3), weight and Body Mass Index (BMI) increased, and metabolic and inflammatory patterns of participants worsened, as usually occurs during the third trimester of pregnancy (CRP from 4.1 to 4.5 mg/L, glycated hemoglobin from 4.6 to 4.9%, total cholesterol from 234 to 257 mg/dL, triglycerides from 173 to 259 mg/dL, Supplementary Table S1). Normoglycemic women were significantly younger and showed, as expected, lower weights, BMI, and fasting glucose values.
Mycobiota composition in GDM patients and in normoglycemic women during the second trimester of pregnancy
A total of 14.007.462 paired reads were obtained by sequencing. After quality filtering, a total of 13.356.188 reads were used, with an average value of 61.266 reads/sample, and a mean sequence length of 465 bp. Alpha diversity index showed a satisfactory coverage for all samples with an average value of 99%. Alpha rarefaction curves reached a plateau, indicating that most biodiversity was captured by the applied analysis (Supplementary Fig. 1).
Beta and alpha diversity analyses in GDM and normoglycemic women are shown in Fig. 1. Bray–Curtis distance matrix showed a significant separation between GDM and controls (Fig. 1A, pairwise PERMANOVA p = 0.001). Alpha-diversity by the Shannon index (Fig. 1B) was significantly higher in GDM patients than controls (p < 0.001).
The mycobiota of patients with GDM was composed almost exclusively by the Ascomycota phylum (Fig. 2A); Basidiomicota accounted for 43% of the relative frequency of the mycobiota of normoglycemic women. Saccharomycodaceae, Pichiaceae, Debaryomycetaceae, Aspergillaceae, Dipodascaceae, Cladosporiaceae and Metschnikowiaceae were the most prevalent families (Fig. 2B). The relative frequency of Pichiaceae and Metschnikowiaceae was significantly higher, while that of Saccharomycetaceae and several minor ASVs were lower in GDM patients than in normoglycemic women; the between-group significant differences at family level are described in Fig. 2C. The predominant genera were: Saccharomyces (54% and 68% of the relative frequency in GDM and normoglycemic women, respectively), Pichia (16% and 2%), Debaryomyces (4.2% and 3.8%), Metschnikowia (3.6% and 0%), Penicillium (2.3% and 3.0%), Cladosporium (2.3% and 3.2%), Geotrichum (0.1% and 0.9%), Candida (2.5% and 1.96%), Kluyveromyces (3.1% and 0.4%), Torulaspora (2.2% and 0.3%), and Aspergillus (0.9% and 1.7%) (Fig. 3A, Supplementary Figure S2 and Supplementary Table 2). At genus level, the Principal Component Analysis (PCA) showed a clear separation between GDM and normoglycemic women (Fig. 3B, p = 0.005). Aspergillus, Hanseniaspora, Kluyveromyces, Metschnikowia, and Pichia showed a significantly higher frequency in GDM patients, while Saccharomyces, Clavispora, Cystobasidium, Debaryomyces, Fusarium and other minor ASVs were more prevalent in normoglycemic women (Fig. 3C). In a multiple regression model, adjusted for age and BMI, Kluyveromyces (ß = 3.18; ES = 0.44, p < 0.001), Metschnikowia (ß = 4.17; ES = 0.56, p < 0.001), and Pichia (ß = 16.8; ES = 1.37, p < 0.001) remained directly associated with GDM and the relative frequency of Saccharomyces (ß = − 12.9; ES = 5.42, p = 0.019), Clavispora (ß = − 0.62; ES = 0.26, p = 0.018), and Cystobasidium (ß = − 2.23; ES = 0.66, p < 0.001) inversely associated.
Machine Learning Procedure for fungal discrimination between normoglycemic women and GDM patients during the second trimester of pregnancy
The random forest model was used to assess the predictive ability of the ASVs at genus level in discriminating between GDM patients and normoglycemic women. The ROC analysis, performed in order to verify the accuracy of the model, showed that gut mycobiota differentiated GDM from normoglycemic women, with an AUC equal to 0.91. The relative frequency and the importance of the first 17 ASVs in discriminating between the two groups are shown in Fig. 4. The importance of the ASVs was not related with their relative frequency. Several ASVs with low frequency showed a high discriminative potential (such as Metschnikowia, Candida, Pichia, Kluyveromyces, Debaryomyces and Torulaspora for GDM patients, and Cladosporium, Saccharomyces, Penicillium, Aspergillus and Galactomyces for normoglycemic women).
Changes in the mycobiota composition in GDM patients from the second to the third trimester of pregnancy
Beta diversity differed between the second and the third trimester of pregnancy (Fig. 5A, PERMANOVA p = 0.001), and a reduction in alpha diversity measure by Shannon index was found during pregnancy (Fig. 5B, p = 0.003). The relative frequency of genera in the second and third trimesters was respectively Saccharomyces (53% and 57%), Pichia (18% and 4%), Metschnikowia (4% and 0.01%), Kluyveromyces (3% and 0.01%), Debaryomyces (2% and 3%), Cladosporium (2% and 6%), Penicillium (2% and 3%), Torulaspora (2% and 0.01%), Candida (1% and 4%), Geotrichum (0.01% and 3%), and Galactomyces (0.01% and 4%) (Fig. 5C and Supplementary Table S2). During the third trimester of pregnancy, a significant increase in the relative frequency of Candida, Cladosporium, Starmerella, Wickerhamiella and Wickerhamomyces was detected (Fig. 5D). On the other hand, the relative frequency of Cystobasidium, Filobasidium, Kluyveromyces, Metschnikowia, Pichia and Torulaspora was reduced with respect to the second trimester of pregnancy.
Associations between mycobiota composition and dietary and clinical variables in GDM patients in the second trimester of pregnancy
In GDM patients, Starmerella was associated with intakes of total carbohydrates (Rho = 0.37; p = 0.01) and sugars (Rho = 0.53; p < 0.001) (Fig. 6A). Penicillium showed an inverse relationship with dietary fiber (Rho = − 0.45; p = 0.002), monounsaturated fats (Rho = − 0.43; p = 0.004) and total energy intake (Rho = − 0.40; p = 0.009). The weekly frequency of meat consumption was directly correlated with Kazachstania (Rho = 0.38; p = 0.01) and Yarrowia (Rho = 0.40; p = 0.008), while Hanseniaspora was inversely associated with the daily consumption of bread (Rho = − 0.53; p < 0.0001), and vegetables (Rho = − 0.46; p = 0.002) and directly with daily consumption of pasta (Rho = 0.30; p = 0.049). Metschnikowia, Torulaspora and Aspergillus were related to the daily consumption of pasta and cereals (Rho = 0.31; p = 0.046; Rho = 0.32; p = 0.038; Rho = 0.35; p = 0.02, respectively). Furthermore, Geotrichum was directly associated with weight (Rho = 0.34; p = 0.028), Yarrowia and Geotrichum with diastolic blood pressure (Rho = 0.31; p = 0.045, and Rho = 0.41; p = 0.007; respectively), Yarrowia and Metschnikowia (Rho = 0.33; p = 0.031, and Rho = 0.41; p = 0.007, respectively) with fasting blood glucose values, Aspergillus with serum cholesterol (Rho = 0.38: p = 0.01) and Penicillium was inversely associated with BMI (Rho = − 0.36; p = 0.02). In a multiple regression model, many of these associations remained statistically significant (Table 1).
Associations between mycobiota composition and dietary and clinical variables in GDM patients in the third trimester of pregnancy
The relative frequency of Kluyveromyces was directly associated with dietary fats (Rho = 0.42; p = 0.005), monounsaturated (Rho = 0.41; p = 0.007) and saturated fats (Rho = 0.33; p = 0.032), total energy intake (Rho = 0.38; p = 0.01), and starch (Rho = 0.32; p = 0.040) (Fig. 6B); Metschnikowia was associated with fiber (Rho = 0.35; p = 0.02), Aureobasidium with the daily consumption of vegetables (Rho = 0.44; p = 0.003), while Aspergillus with the weekly consumption of legumes (Rho = 0.31; p = 0.045) (Fig. 6B). Fasting insulin levels were associated with the relative frequency of Hanseniaspora (Rho = 0.31; p = 0.049), and fasting glucose with Saturnispora (Rho = 0.35; p = 0.023). In a multiple regression model, the association between Kluyveromyces and monounsaturated fats and the association between Hanseniaspora and fasting insulin remained statistically significant (Table 1).
Bacteria–fungi associations in GDM patients
We analyzed the gut bacteria-fungi associations; the gut microbiota was previously characterized30. A different pattern of bacteria-fungi associations was observed in GDM patients at T2 (Fig. 6A) and T3 (Fig. 6B). A complex relationship between bacteria and fungi in the correlation network was more evident at T2 than at T3 (Fig. 6A, FDR < 0.02). At T2, Aspergillus displayed co-exclusion pattern with Lactobacillus while co-occurred with Bacteroides. Debaryomyces co-excluded Bacteroides, while Galactomyces co-occurred with Blautia; Dorea and Ruminococcus. Kluyveromyces and Metschnikowia co-excluded Lachnospira. A significant inverse association between Hanseniaspora and Blautia and between Lachnospiraceae with Hanseniaspora, Kluyveromyces and Pichia was detected while Finegoldia co-occurred with Galactomyces (Fig. 6A, FDR < 0.02). At T3, the direct associations between Kluyveromyces and Eubacterium, Fusarium and Collinsella, Penicillium and Roseburia, and the inverse relationship between Penicillium and Butyricimonas; between Bacteroides and Vishniacozyma; between Christensenellaceae and Kurtzmaniella were identified (Fig. 6B, FDR < 0.02).
Discussion
GDM patients showed a different mycobiota composition than normoglycemic women, with increased alpha diversity and predominance of Ascomycota, in particular, Metschnikowia, Pichia and Kluyveromyces, and reduced relative frequency of Saccharomyces. From the second to the third trimester of pregnancy, the mycobiota composition and alpha diversity changed in GDM patients. Many associations between fungi and dietary and clinical variables were detected. Finally, several fungi and bacteria showed inverse or direct relationships, suggesting respectively, competition or co-occurrence.
Gut mycobiota signature between GDM patients and normoglycemic women
An increase in alpha diversity was observed in GDM patients when compared to normoglycemic women. No data on fungal alpha diversity are available, at present, in pregnant women, and very few studies have assessed this topic. Individuals with overweight or obesity showed an increased fungal count, but a slightly decreased diversity6,28, as well as patients with IBD31, IBS14, Clostridium difficile infection32, and psychiatric diseases33,34. Indeed, these data are contrasting, since the same authors did not outline differences in alpha diversity in samples of patients with Crohn disease31 and of IBS-like mice14. Furthermore, other authors failed to find differences between patients with ulcerative colitis35 or neurodevelopmental conditions36 and healthy subjects. A possible explanation for the increased alpha diversity might be the increased relative frequency of the taxa potentially proinflammatory in our GDM patients when compared to normoglycemic controls (see below). Otherwise, the dietary habits of GDM women, which have been reported to differ from those of normoglycemic pregnant women29,37, might be implicated. Unfortunately, lack of reliable nutritional data in our normoglycemic women has not allowed us to make such a comparison.
GDM patients were characterized by the predominance of Ascomycota, differently from the normoglycemic women, who displayed ~ 50% relative frequency of Basidiomycota. Accordingly, Ascomycota phylum has been shown both to predominate in individuals with obesity and to be negatively correlated with the Basidiomycota phylum28. Several ASVs showed a high discriminative potential between GDM and normoglycemic women, in particular, Metschnikowia, Pichia and Kluyveromyces in GDM patients. Metschnikowia is an environmental, plant-related yeasts previously isolated from human samples (saliva, blood, vagina, and rectum) and related to several skin pathologies38. Strains belonging to Metschnikowia are able to exert an antagonist effect against microbes and fungi due to the production of pulcherrimin, a pigment closely linked to its antagonistic capacity, the production of which is promoted by the presence of simply sugars, such as glucose, galactose and disaccharides39. Therefore, it could be hypothesized that in the presence of increased concentrations of glucose, Metschnikowia might exert antagonist effects against other taxa, thus justifying its predominance in GDM towards normoglycemic women. Pichia, a member of the oral and gut microbiome40, was shown to be increased in Clostridium difficile infection when compared to healthy individuals41 and has implicated in infant gut mycobial dysbiosis associated with atopic disease42. P. kudriavzevii belonging to Pichia genus was associated with an increased inflammatory response and a reduced abundance of short chain fatty acids (SCFAs) in mice43. It was shown that the gut dysbiosis with depletion in SCFA-producing bacteria, which characterizes GDM patients44, can increase the predominance of inflammatory fungal taxa, such as Pichia, that release an inhibitory molecule with antagonistic effects against the other fungi40, thus explaining its increased relative frequency in the GDM patients. Kluyveromyces (a common dairy yeast) was a discriminant taxon in our GDM patients. It was reported to be also discriminant in Crohn’s disease35. Members from this genus are able to produce volatile sulfur compounds via methionine aminotransferase45. Higher levels of sulfuric compounds are considered to play a role in chronic inflammatory diseases, such as IBDs and cancers46.
Aspergillus, Hanseniaspora, Candida, and Torulaspora showed an increased, though not significant, higher frequency in GDM patients than in normoglycemic women. In particular, no significant difference was found in the relative frequency of Candida spp., which have been suggested to increase in diabetes mellitus and inflammatory gut diseases, and to be linked to decrease SCFA levels in the gut6,25,26,28. Indeed, other studies, similarly to our findings, did not point out differences in this taxon in the presence of conditions associated with insulin resistance, such as obesity6,28.
Normoglycemic women were characterized by an increased relative frequency of Saccharomyces, and other minor taxa, such as Clavispora and Cystobasidium. Saccharomyces is a stable yeast in the gut of healthy individuals, even after consumption of Saccharomyces free-diet9,47,48. Strains belonging to this genus have been used to treat gastrointestinal disorders such as IBD and IBS49,50. Patients with IBD and obesity showed highest levels of anti S. cerevisiae antibodies and lower abundance of Saccharomyces in the gut31. Furthermore, Saccharomyces coexists with beneficial SCFA-producing bacteria, such as Faecalibacterium, Lachnospiraceae and Ruminococcaceae51, thus suggesting a favorable anti-inflammatory role. Finally, physiologically relevant concentrations of SCFAs seem to be fungistatic52.
Gut mycobiota changes during GDM
During the course of pregnancy, we have observed both a reduction in the alpha diversity value and changes in the mycobiota composition. The reduction of alpha diversity within the patients with GDM is in line with the remarkable hormonal, immunological, and metabolic changes taking place during the third trimester to promote maternal weight gain, increasing circulating pro-inflammatory cytokines, and insulin resistance, which are accentuated in the presence of gestational hyperglycemia29. Among the changes in the mycobiota composition occurring between the second and the third trimester of pregnancy, the reduction in the relative frequency of Pichia, Metschnikowia, Kluyveromyces, and the increase in Candida, Cladosporium and Starmerella are the most relevant variations we observed. An inverse relationship between Candida and Pichia has already been reported53, with a reduction of the latter coinciding with an increase in Candida colonization, and vice versa. Recently, Candida spp have been demonstrated to play a key role in induction host-protective antifungal IgG antibodies that are protective against systemic fungal infection12. Thus, competitive effects among fungal taxa may be hypothesized.
Associations between mycobiota, nutrient intakes, and metabolic variables
Our GDM women ate an excessive amount of fat and less than the recommended dietary fiber intake. Most gut fungi derive from foods4, in particular, Penicillium, and Aspergillus were associated with a plant-based diet54. On the other hand, an enrichment of Penicillium was observed in humans consuming an animal-based diet55 and in mice fed with a high-fat diet56. Moreover, Aspergillus has been linked to adiposity and related metabolic disorders including insulin resistance, arterial blood pressure, and inflammation28. Accordingly, in our GDM patients, the relative frequency of Aspergillus was associated with the consumption of pasta and cereals, while Penicillium was inversely associated with fiber and monounsaturated fat intake and with serum cholesterol levels.
Furthermore, we found a direct relationship between Yarrowia and meat consumption; strains belonging to Yarrowia (Y. lipolytica) are commonly found in poultry products, ham, beef, and sausages57 as well as in dairy foods. Kluyveromyces is a dairy potential probiotic58 whose relative frequency was found to be associated with the intake of monounsaturated fats in the third trimester. Hanseniaspora resulted inversely correlated with plant-derived foods and, in the third trimester of pregnancy, with fasting insulin. In line, six-weeks of a high-fat diet induced an increase in this taxon in mice56. Finally, Metschnikowia (a plant related yeasts) was associated with fasting glucose, since increased concentrations of glucose might favor its growth and the exertion of antagonist effects against other taxa39.
Bacteria–fungi interactions in GDM patients
The chronic inflammatory status associated with gestational hyperglycemia59 impacts on the microbiota composition60, potentially modifying the interaction between bacteria and fungi, by causing changes in their relative abundances. We have previously found several associations between pro-inflammatory bacteria (Bacteroides, Collinsella, and Sutterella) and metabolic/inflammatory variables across pregnancy in our GDM patients30.
Complex relationships between bacteria and fungi have been reported in humans, whose biological significance is still uncertain, but might be an opportunity to modulate the gut taxa7,31,48. Accordingly, we found that fungi likely associated with a pro-inflammatory state, such as Metschnikowia, excluded beneficial bacteria such as Lachnospira and Lactobacillus. Aspergillus co-occurred with Bacteroides, probably because members of the genus Aspergillus can produce extracellular polysaccharides and, consequently, favor the presence of Bacteroides, which are polysaccharide degraders61. Furthermore, Hanseniaspora, which is associated with the consumption of fruits and sugary foods, was inversely associated with Lachnospiraceae family and in particular with Blautia, whose relative abundance was related to a higher insulin sensitivity in our GDM patients30, and Debaryomyces, microorganisms with potential health benefit deriving from dairy and fermented foods62 co-exclude Bacteroides. Lachnospiraceae family co-excludes several yeasts, and an abnormal increase in the former was reported to lead to excessive energy uptake from indigestible polysaccharides63.
However, the bacteria-fungi relationships were quite complex, and our GDM patients showed the concomitance of pathogenic fungi and pro-inflammatory bacteria, but also the co-occurrence of beneficial fungi and bacteria. Thus, the opportunistic pathogen Fusarium (a plant pathogen from environment40,64,65) co-occurred with the pro-inflammatory Collinsella66 which displays the ability to detoxify trichothecene mycotoxins from Fusarium67. Similarly, the beneficial Galactomyces, deriving from dairy fermented foods64, co-occurred with beneficial bacteria, such as Blautia. The health-associated taxa, Christensenellaceae co-exclude the presence of Kurtzmaniella which was reported to be associated with high-fat diet68,69.
This area of research is at the beginning, but it is quite intriguing owing its potential implications on health. In particular, the study of known microbiome interactions with human health, comprising pregnancy and preterm birth, inflammatory bowel diseases, and stressors that affect individuals with prediabetes, have been considered to serve as models of ‘typical’ microbiome-associated conditions of broad interest to the research community70.
Clinical implications
The transfer of fungal phenotypes from the mother to the offspring has been reported47 and early fungal colonization was hypothesized to impact on the development of diseases later in life3. Increased fungal abundance in children was associated with increased height velocity and reduced BMI10, and the mycobiome characteristics have been implicated in obesity and many inflammatory and allergic diseases during childhood71,72,73,74.
The development of strategies that promote a beneficial mycobiota in pregnancy might be an opportunity to modulate the inflammatory status and insulin resistance characterizing gestational hyperglycemia, and the future health of the offspring11.
Limitations and strengths
The limitations of the present study should be acknowledged. Finding one method able of sufficiently extracting DNA from all fungal types is challenging and a great variation between the methods’ performance was already highlighted.
The amplicon sequencing technique may lead to possible biases deriving from DNA extraction, PCR amplification, as well as the failure in discriminating between live or death cell. Furthermore, fungi produce different types of cells, such as simple cells, hyphae or spores, which may lead to biases during DNA extraction and either over- or underestimation. In addition, fungal sexual or asexual forms can be classified as different taxa with possible discrimination. The study of gut fungi is still a new science and remains a main challenge. Indeed, the metataxonomic approach is at presence a largely used method for characterizing the gut mycobiota. However, given the nature of PCR, bias can be introduced based on the target region chosen for the amplifications. Due to the observational design of this study, the presence of unmeasured confounding factors cannot be excluded. Reliable data on dietary habits of normoglycemic women were not available. The fecal samples were used as proxies for the microbiome of the entire gastrointestinal tract. The limitations of the food questionnaires must be recognized, even if the dietary intakes of our patients resembled those previously reported in pregnancy75. Finally, to the best of our knowledge, no previous studies have investigated the composition of the mycobiota neither in pregnancy nor in patients with GDM.
Conclusions
Patients with GDM showed a predominance of fungal taxa with potential inflammatory effects, with a marked shift in their mycobiota during pregnancy, and complex bacteria-fungi interactions. If these data will be confirmed by further studies, the possibility to modulate the gut microbiome during pregnancy should be tested as an intriguing challenge to impact on the health of the mother and the offspring.
Methods
Participant enrolment
The first 50 patients consecutively diagnosed with GDM at the “Città della Salute e della Scienza” Hospital of Turin from April 2016, were enrolled. The characteristics of this cohort were already reported30,76. Briefly, inclusion criteria were: gestational age between 24 and 28 weeks, Caucasian race, GDM diagnosed by a 75 g oral glucose tolerance test (OGTT), performed in the morning, after at least 8 h-overnight fast, and interpreted according to international criteria30. The exclusion criteria were: twin pregnancy, use of prebiotics/probiotics, antibiotics or any drug during pregnancy, any pathological conditions before or during pregnancy (known diabetes mellitus, hypertension, cardiovascular, pulmonary, autoimmune, joint, liver or kidney diseases, thyroid dysfunction, cancer, any other disease/condition), no compliance to the study protocol.
A group of 150 normoglycemic women were then consecutively enrolled starting from October 2020, at the same Hospital. Inclusion criteria were: gestational age between 24 and 28 weeks, Caucasian race, normoglycemia diagnosed by a 75 g OGTT, according to the same criteria. Exclusion criteria were the same as for GDM patients.
Ethical aspects
Each participant gave her written informed consent to participate in the study. The study protocol was approved by the Ethics Committee of the “Città della Salute e della Scienza” Hospital of Torino. All research was performed in accordance with the Helsinki Declaration principles.
GDM patients
All GDM patients routinely received dietary counselling and nutritional and exercise recommendations in line with guidelines77. Questionnaires, anthropometric values, fasting blood samples and stool samples were collected for all participants both at 24–28 weeks of gestational age at the time of GDM diagnosis (second trimester, T2), and at 38 weeks, or before delivery, in the case of preterm delivery (third trimester, T3). Participants completed a 3-day food record (2 weekdays and 1 weekend day) at T2 and T3. Detailed information on how to record food and drink consumed by using common household measures was provided to all participants. Two dieticians checked all questionnaires for completeness, internal coherence, and plausibility.
Data relative to pre-pregnancy weight was self-reported; weight, height, and arterial blood pressure (BP) were measured at T2 and weight and BP at T3. Body weight was measured to the nearest 0.1 kg, and height was measured to the nearest 0.1 cm with a stadiometer (SECA model 711, Hamburg, Germany), with the participants wearing light clothes and no shoes. Arterial BP was measured from the left arm, in a sitting position, after at least 10 min of rest, with a mercury sphygmomanometer with appropriate cuff sizes (ERKA Perfect-Aneroid, Germany). Two measurements were taken by trained personnel with arm supported at heart level and the values reported were the means of the two.
A fasting blood sample was collected both at T2 and T3 from all the participants for the detection of serum glucose, glycated hemoglobin, insulin, total and HDL cholesterol, triglycerides, and C-reactive protein. Laboratory methods were previously reported30. All laboratory measurements were centralized. The HOMA-IR was calculated according to the published algorithm78.
Normoglycemic women
Normoglycemic women were evaluated during the OGTT only (T2). Their weight, height, and arterial BP were measured as in the GDM patients.
Stool samples
Stool samples were self-collected by the patients as previously described30. Briefly, the subjects were instructed to self-collect the samples, and all materials were provided in a convenient, refrigerated, specimen collection kit. Patients were provided with sterile containers to collect the feces (VWR, Milan, Italy). Fecal samples were collected at home and transferred into the sterile sampling containers using a polypropylene spoon (~ 10 g) and immediately stored at 4 °C. The specimens were transported to the laboratory within 8 h of collection under refrigerated temperature. Containers were immediately stored at − 80 °C for DNA extraction. No storage medium was used.
DNA extraction and meta-taxonomic amplicon sequencing
Total DNA was extracted by using the RNeasy Power Microbiome KIT (Qiagen, Milan, Italy) following the manufacturer’s instructions. In order to digest the RNA, 5uL of RNAse-A was added to the mixture and then incubated for 1 h at 37 °C. DNA was quantified by using the QUBIT ds Kit and normalized at 5 ng/uL. 2.5 uL of the DNA was amplified by using the primers NL4R (5′-GGTCCGTGTTTCAAGACGG-3′) and LS2-MF (5′-GAGTCGAGTTGTTTGGGAAT-3′) as previously described79. Standard Illumina overhang adapter sequences were added to locus‐specific sequences. PCR consisted of 25 cycles (95 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s) plus one additional cycle at 72 °C for 10 min as a final chain elongation. PCR products were then purified by using the Agencourt AMPure XP beads (Beckman Coulter Genomics) and tagged according to the Illumina Sequencing Library Preparation. Amplicons were quantified using Qubit dsDNA assay kit diluted to 4 nM, denatured with 0.2 N NaOH and spiked with 20% (v/v) of PhiX. The combination of pool library and PhiX were diluted to 8 pM, and paired end sequencing was performed on the MiSeq platform using the MiSeq Reagent Kit V2 (2 × 250 bp) (Illumina, San Diego, USA), following the standard Illumina sequencing protocol.
Bioinformatics Analysis
After sequencing, fastq files were imported in QIIME version 2. Sequence adapters and primers were trimmed by using cutadapter, while DADA2 algorithm was used to trim low quality reads, to remove chimeric sequences, and joined sequences shorter than 300 bp by using the DADA2 denoise paired plug-in of QIIME2. Amplicon sequence variants (ASVs) obtained by DADA2 were used for taxonomic assignment using the QIIME feature-classifier plugin against the manually build database for the mycobiota79. Briefly the database was constructed using the large subunit rRNA gene sequences, 23.381 sequences were downloaded from Silva database and from NCBI. Taxonomy assignment for 26S was double checked on BLAST suite tools to confirm the taxonomic assignment.
Statistical analyses
QIIME2 diversity script was used to perform alpha and beta diversity analysis. Bray–Curtis distance matrix was generated by QIIME2 and was used both to build the principal coordinate analysis (PCoA) and to perform pairwise PERMANOVA by the “vegan” package in R environment. Principal Component Analysis (PCA) based on ASVs at genus level on GDM dataset was performed and plotted through the function dudi.pca in R environment. Not-normally distributed variables were presented as median (range interquartile). Within-differences in GDM patients (T2 vs T3) were evaluated by paired-sample t-test, or Wilcoxon matched pairs test, as appropriate. Differences between categorical variables were computed by the chi-square test. Between-differences between GDM patients and normoglycemic women were assessed by t-Student’s test or Mann–Whitney test.
q2-sample-classifier plugin of QIIME2 was used to predict sample type (GDM/normoglycemic) as a function of the amplicon sequence variants detected in the samples. Briefly, the input samples were randomly split into a training set and a test set. The test set was held out until the end of the pipeline, allowing to assess accuracy. The model was trained to predict a specific target variable (GDM vs normoglycemic women) based on the Random Forest approach. The accuracy of the results was calculated from the qiime2 software using the Receiver Operating Characteristic (ROC) curves as well as by measure the area under the curve (AUC) in order to verify the classification accuracy of the model.
Pairwise Spearman’s non-parametric correlations were used to study the relationships between the relative frequency of fungal taxa, bacteria, and dietary and clinical variables. Bacteria composition of GDM cohort previously published30 was used to performed the correlation analysis. Correlation network was visualized by using Cytoscape v. 3.8.2. P-values were adjusted for multiple testing using the Benjamini–Hochberg procedure, which assesses the false discovery rate (FDR). Multiple regression analyses were performed to evaluate the associations between fungal relative frequency and nutrient and clinical variables in GDM patients, after adjusting for age and BMI. The same model was used to evaluate the association of specific taxa with GDM in the whole sample of women (Statistica, ver. 7.0; StatSoft Inc., Tulsa, OK, USA).
Data availability
Data generated by sequencing were deposited in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) and are available under the BioProject Accession Number PRJNA762265 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA762265).
References
Gouba, N. & Drancourt, M. Digestive tract mycobiota: A source of infection. Méd. Mal. Infect. 45, 9–16 (2015).
Hager, C. L. & Ghannoum, M. A. The mycobiome: Role in health and disease, and as a potential probiotic target in gastrointestinal disease. Dig. Liver Dis. 49, 1171–1176 (2017).
Wu, X., **a, Y., He, F., Zhu, C. & Ren, W. Intestinal mycobiota in health and diseases: From a disrupted equilibrium to clinical opportunities. Microbiome 9, 1–18 (2021).
Huseyin, C. E., O’Toole, P. W., Cotter, P. D. & Scanlan, P. D. Forgotten fungi—the gut mycobiome in human health and disease. FEMS Microbiol. Rev. 41, 479–511 (2017).
Dworecka-Kaszak, B., Dabrowska, I. & Kaszak, I. The mycobiome – a friendly cross-talk between fungal colonizers and their host. Ann. Parasitol. 62, 175–184 (2016).
Borges, F. M. et al. Fungal diversity of human gut microbiota among eutrophic, overweight, and obese individuals based on aerobic culture-dependent approach. Curr. Microbiol. 75, 726–735 (2018).
Nash, A. K. et al. The gut mycobiome of the Human Microbiome Project healthy cohort. Microbiome 5, 153 (2017).
Strati, F. et al. Age and gender affect the composition of fungal population of the human gastrointestinal tract. Front. Microbiol. 7, 1–16 (2016).
Auchtung, T. A. et al. Investigating colonization of the healthy adult gastrointestinal tract by fungi. mSphere 3, 1–16 (2018).
Schei, K. et al. Early gut fungal and bacterial microbiota and childhood growth. Front. Pediatr. 8, 572538 (2020).
Li, X. V., Leonardi, I. & Iliev, I. D. Gut mycobiota in immunity and inflammatory disease. Immunity 50, 1365–1379 (2019).
Doron, I. et al. Human gut mycobiota tune immunity via CARD9-dependent induction of anti-fungal IgG antibodies. Cell 184, 1017-1031.e14 (2021).
Alonso, R., Pisa, D., Fernández-fernández, A. M. & Carrasco, L. Infection of fungi and bacteria in brain tissue from elderly persons and patients with alzheimer ’ s disease. Front. Aging Neurosci. 10, 1–20 (2018).
Botschuijver, S. et al. Intestinal fungal dysbiosis is associated with visceral hypersensitivity in patients with irritable bowel syndrome and rats. Gastroenterology 153, 1026–1039 (2017).
Iliev, I. D. et al. Interactions between commensal fungi and the C-type lectin receptor dectin-1 influence colitis. Science 336, 1314–1317 (2012).
Seed, P. C. The human mycobiome. Cold Spring Harb. Perspect. Med. 5, 1–10 (2015).
El Mouzan, M. et al. Fungal microbiota profile in newly diagnosed treatment-naïve children with crohn’s disease. J. Crohns. Colitis 11, 586–592 (2017).
Yang, A. M. et al. Intestinal fungi contribute to development of alcoholic liver disease. J. Clin. Invest. 127, 2829–2841 (2017).
Luan, C., Miao, H. & Zhu, B. Gut mycobiota and adenomas. Gut Microbes 6, 331–333 (2015).
Coker, O. O. et al. Enteric fungal microbiota dysbiosis and ecological alterations in colorectal cancer. Gut 68, 654–662 (2019).
Aykut, B. et al. The fungal mycobiome promotes pancreatic oncogenesis via activation of MBL. Nature 574, 264–267 (2019).
Bacher, P. et al. Human anti-fungal Th17 immunity and pathology eely on cross-reactivity against Candida albicans. Cell 176, 1340-1355.e15 (2019).
Aogáin, M. M. et al. Immunological corollary of the pulmonary mycobiome in bronchiectasis: The CAMEB study. Eur. Respir. J. 52, 1800766 (2018).
Chacón, M. R. et al. The gut mycobiome composition is linked to carotid atherosclerosis. Benef. Microbes 9, 185–198 (2018).
Gosiewski, T. et al. Quantitative evaluation of fungi of the genus Candida in the feces of adult patients with type 1 and 2 diabetes—A pilot study. Gut Pathog. 6, 1–5 (2014).
Soyucen, E. et al. Differences in the gut microbiota of healthy children and those with type 1 diabetes. Pediatr. Int. 56, 336–343 (2014).
Kowalewska, B., Zorena, K., Szmigiero-Kawko, M., Wąż, P. & Myśliwiec, M. Higher diversity in fungal species discriminates children with type 1 diabetes mellitus from healthy control. Patient Prefer. Adherence 10, 591–599 (2016).
Mar Rodríguez, M. et al. Obesity changes the human gut mycobiome. Sci. Rep. 5, 1–15 (2015).
Ponzo, V. et al. Diet-gut microbiota interactions and Gestational Diabetes Mellitus (GDM). Nutrients 11, 330 (2019).
Ferrocino, I. et al. Changes in the gut microbiota composition during pregnancy in patients with gestational diabetes mellitus (GDM). Sci. Rep. 8, 12216 (2018).
Sokol, H. et al. Fungal microbiota dysbiosis in IBD. Gut 66, 1039–1048 (2017).
Cao, Y. et al. Fecal mycobiota combined with host immune factors distinguish Clostridioides difficile infection from asymptomatic carriage. Gastroenterology 160, 2328-2339.e6 (2021).
Jiang, H. Y. et al. Altered gut bacterial–fungal interkingdom networks in patients with current depressive episode. Brain Behav. 10, 1–10 (2020).
Zhang, X. et al. Analysis of gut mycobiota in first-episode, drug-naïve Chinese patients with schizophrenia: A pilot study. Behav. Brain Res. 379, 112374 (2020).
Qiu, X. et al. Alterations in the mucosa-associated fungal microbiota in patients with ulcerative colitis. Oncotarget 8, 107577–107588 (2017).
Strati, F. et al. Altered gut microbiota in Rett syndrome. Microbiome 4, 1–15 (2016).
Bo, S. et al. Dietary fat and gestational hyperglycaemia. Diabetologia 44, 972–978 (2001).
Savini, V. et al. An atypical, pigment-producing Metschnikowia strain from a leukaemia patient. Med. Mycol. 51, 438–443 (2013).
Horváth, E. et al. The antagonistic Metschnikowia andauensis produces extracellular enzymes and pulcherrimin, whose production can be promoted by the culture factors. Sci. Rep. 11, 1–14 (2021).
Kabwe, M. H., Vikram, S., Mulaudzi, K., Jansson, J. K. & Makhalanyane, T. P. The gut microbiota of rural and urban individuals is shaped by geography and lifestyle. BMC Microbiol. 20, 1–12 (2020).
Beheshti-Maal, A. et al. Gut mycobiome: The probable determinative role of fungi in IBD patients. Mycoses 64, 468–476 (2021).
Ward, T. L., Knights, D. & Gale, C. A. Infant fungal communities: Current knowledge and research opportunities. BMC Med. 15, 1–10 (2017).
Boutin, R. C. T. et al. Bacterial–fungal interactions in the neonatal gut influence asthma outcomes later in life. Elife 10, 1–22 (2021).
Fan, Y. & Pedersen, O. Gut microbiota in human metabolic health and disease. Nat. Rev. Microbiol. 19, 55–71 (2021).
Kagkli, D. M., Bonnarme, P., Neuvéglise, C., Cogan, T. M. & Casaregola, S. L-methionine degradation pathway in Kluyveromyces lactis: Identification and functional analysis of the genes encoding L-methionine aminotransferase. Appl. Environ. Microbiol. 72, 3330–3335 (2006).
Carbonero, F., Benefiel, A. C., Alizadeh-Ghamsari, A. H. & Gaskins, H. R. Microbial pathways in colonic sulfur metabolism and links with health and disease. Front. Physiol. 3, 1–11 (2012).
Schei, K. et al. Early gut mycobiota and mother-offspring transfer. Microbiome 5, 1–12 (2017).
Hoarau, G. et al. Bacteriome and mycobiome interactions underscore microbial dysbiosis in familial Crohn’s disease. MBio 7(5), e01250-e1316 (2016).
Palma, M. L. et al. Probiotic Saccharomyces cerevisiae strains as biotherapeutic tools: is there room for improvement?. Appl Microbiol Biotechnol 99, 6563–6570 (2015).
Pineton de Chambrun, G. et al. A randomized clinical trial of Saccharomyces cerevisiae versus placebo in the irritable bowel syndrome. Dig. Liver Dis. 47, 119–124 (2015).
Suhr, M. J. & Hallen-Adams, H. E. The human gut mycobiome: Pitfalls and potentials-a mycologist’s perspective. Mycologia 107, 1057–1073 (2015).
Cottier, F., Tan, A. S. M., Xu, X., Wang, Y. & Pavelka, N. MIG1 regulates resistance of Candida albicans against the fungistatic effect of weak organic acids. Eukaryot. Cell 14, 1054–1061 (2015).
Mukherjee, P. K. et al. Oral mycobiome analysis of HIV-infected patients: Identification of Pichia as an antagonist of opportunistic fungi. PLoS Pathog. 10(3), e1003996 (2014).
Hallen-Adams, H. E. & Suhr, M. J. Fungi in the healthy human gastrointestinal tract. Virulence 8, 352–358 (2017).
David, L. A. et al. Diet rapidly and reproducibly alters the human gut microbiome. Nature 505, 559–563 (2014).
Van Der Merwe, M. et al. Time of feeding alters obesity-associated parameters and gut bacterial communities, but not fungal populations, in C57BL/6 male mice. Curr. Dev. Nutr. 4, 1–12 (2020).
Mamaev, D. & Zvyagilskaya, R. Yarrowia lipolytica: A multitalented yeast species of ecological significance. FEMS Yeast Res. 21, 1–18 (2021).
Fadda, M. E., Mossa, V., Deplano, M., Pisano, M. B. & Cosentino, S. In vitro screening of Kluyveromyces strains isolated from Fiore Sardo cheese for potential use as probiotics. LWT - Food Sci. Technol. 75, 100–106 (2017).
Pantham, P., Aye, I. L. M. H. & Powell, T. L. Inflammation in maternal obesity and gestational diabetes mellitus. Placenta 36, 709–715 (2015).
Mokkala, K. et al. Gut microbiota aberrations precede diagnosis of gestational diabetes mellitus. Acta Diabetol. 54, 1147–1149 (2017).
Rui, Y. et al. Simulated digestion and fermentation in vitro by human gut microbiota of intra- and extra-cellular polysaccharides from Aspergillus cristatus. Lwt 116, 108508 (2019).
Agarbati, A. et al. Potential probiotic yeasts sourced from natural environmental and spontaneous processed foods. Foods 9, 1–25 (2020).
Gomez-Arango, L. F. et al. Connections between the gut microbiome and metabolic hormones in early pregnancy in overweight and obese women. Diabetes 65, 2214–2223 (2016).
Ghannoum, M. et al. Effect of mycobiome diet on gut fungal and bacterial communities of healthy adults. J. Probiotics Heal. 8(215), 1–6 (2020).
Limon, J. J., Skalski, J. H. & Underhill, D. M. Commensal fungi in health and disease. Cell Host Microbe 22, 156–165 (2017).
Gomez-Arango, L. F. et al. Low dietary fiber intake increases Collinsella abundance in the gut microbiota of overweight and obese pregnant women. Gut Microbes 9(3), 189–201 (2017).
Yu, H. et al. Isolation of deoxynivalenol-transforming bacteria from the chicken intestines using the approach of PCR-DGGE guided microbial selection. BMC Microbiol. 10(182), 1–10 (2010).
Tavella, T. et al. Elevated gut microbiome abundance of Christensenellaceae, Porphyromonadaceae and Rikenellaceae is associated with reduced visceral adipose tissue and healthier metabolic profile in Italian elderly. Gut Microbes 13, 1–19 (2021).
Van Der Merwe, M. et al. Time of feeding alters obesity-associated parameters and gut bacterial communities, but not fungal populations, in C57BL/6 male mice. Curr. Dev. Nutr. 4, 1–12 (2020).
iHMP. The integrative human microbiome project. Nature 569, 641–648 (2019)
Chehoud, C. et al. Fungal signature in the gut microbiota of pediatric patients with inflammatory bowel disease. Inflamm. Bowel Dis. 21, 1948–1956 (2015).
Osborne, M. et al. Specific fungal exposures, allergic sensitization, and rhinitis in infants. Pediatr. Allergy Immunol. 17, 450–457 (2006).
Behbod, B. et al. Asthma and allergy development: contrasting influences of yeasts and other fungal exposures. Clin. Exp. Allergy 45, 154–163 (2015).
Borgo, F. et al. Relative abundance in bacterial and fungal gut microbes in obese children: A case control study. Child. Obes. 13, 78–84 (2017).
Portune, K. J., Benítez-Páez, A., Del Pulgar, E. M. G., Cerrudo, V. & Sanz, Y. Gut microbiota, diet, and obesity-related disorders—The good, the bad, and the future challenges. Mol. Nutr. Food Res. 61, 1600252 (2017).
Ponzo, V. et al. The microbiota composition of the offspring of patients with gestational diabetes mellitus (GDM). PLoS ONE 14, 1–18 (2019).
Hod, M. et al. The International Federation of Gynecology and Obstetrics (FIGO) Initiative on gestational diabetes mellitus: A pragmatic guide for diagnosis, management, and care. Int. J. Gynecol. Obstet. 131, S173–S211 (2015).
Matthews, D. R. et al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28, 412–9 (1985).
Mota-Gutierrez, J., Ferrocino, I., Rantsiou, K. & Cocolin, L. Metataxonomic comparison between internal transcribed spacer and 26S ribosomal large subunit (LSU) rDNA gene. Int. J. Food Microbiol. 290, 132–140 (2019).
Author information
Authors and Affiliations
Contributions
I.F. and S.B. were responsible for the development of the study funding acquisition and preparation of the manuscript. V.P., I.G., MA.PE., MA.PA. and C.D. recruited subjects and collected biospecimens. I.F. and IR.FR. performed the DNA extraction and the amplicon sequencing. V.P., MA.PE. and S.B., collected and analyzed the dietary information. I.F. carried out the bioinformatics and generated the manuscript figures. I.F. and S.B. carried out the statistical analyses. E.G. and L.C. supervised the data analysis and contributed to manuscript preparation. I.F., V.P. and S.B. writing the original draft. All authors revised and approved the final version of the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Ferrocino, I., Ponzo, V., Pellegrini, M. et al. Mycobiota composition and changes across pregnancy in patients with gestational diabetes mellitus (GDM). Sci Rep 12, 9192 (2022). https://doi.org/10.1038/s41598-022-13438-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-13438-0
- Springer Nature Limited
This article is cited by
-
Maternal microbiota and gestational diabetes: impact on infant health
Journal of Translational Medicine (2023)