Characteristics of the study population
We randomly enrolled 30 healthy mother-infant pairs in this study by following up to two years old with the normal range of the growth and development indexes (BMI, LAZ, WAZ and BMI Z: -1 to 1) at birth, 6 months, one year, one and a half year, and two years respectively at the First Hospital of Tsinghua University (Fig. 1). The characteristics of the study population was shown in Table 1, the infants were born at the gestational age (38.91 ± 1.35 week), in which 18 (60.00%) subjects were produced by the cesarean and 17 (56.67%) were boy. Meanwhile, the mother age at the time of specimen collection was (31.09 ± 3.44) year with the (12.98 ± 3.07) kg weight and (4.91 ± 1.15) kg/m2 BMI gain during the whole pregnancy. Moreover, Table 1 demonstrated there were no significant differences of these characteristics at baseline including the maternal age, height, pre-pregnancy weight, pre-pregnancy BMI, prenatal weight, prenatal BMI, weight gain, BMI gain, gestational age, sex of the infants and mode of delivery between all subjects (n = 254) and the enrolling 30 healthy mother-infant pairs (n = 30) in this research (P > 0.05).
Table 1
Descriptive data of the basic information in this study (n = 30)
Index
|
All subjects (n = 254)
|
Enrolling subjects (n = 30)
|
t/x2
|
P
|
Pregnant woman
Age (years)
|
31.88 + 3.97
|
31.09 + 3.44
|
0.363
|
0.729
|
Height (cm)
|
162.32 + 7.86
|
162.58 + 14.52
|
0.027
|
0.980
|
Pre-pregnancy weight (kg)
|
57.91 ± 11.11
|
55.54 + 6.07
|
0.460
|
0.662
|
Pre-pregnancy BMI (kg/m2)
|
21.68 ± 3.76
|
21.01 + 2.77
|
0.351
|
0.738
|
Prenatal weight (kg)
|
70.81 ± 9.06
|
68.52 + 6.54
|
0.484
|
0.645
|
Prenatal BMI (kg/m2)
|
26.55 ± 2.77
|
25.92 + 3.10
|
0.344
|
0.742
|
Weight gain (kg)
|
12.91 ± 3.36
|
12.98 + 3.07
|
0.035
|
0.973
|
BMI gain (kg/m2)
|
4.87 ± 1.37
|
4.91 + 1.15
|
0.127
|
0.903
|
Infant
Gestational age (week)
|
39.08 + 1.40
|
38.91 + 1.35
|
0.251
|
0.810
|
Sex-boy (%)
|
132 (51.97%)
|
17 (56.67%)
|
0.237
|
0.626
|
Mode of production-Natural (%)
|
161 (63.39%)
|
18 (60.00%)
|
0.132
|
0.716
|
Diversity of the gut microbiota between the different age subgroups
To characterize the dynamic colonization of intestinal flora at different ages, we shotgun the sequenced stool samples at 6 months, one-year and two years by 16SrRNA sequencing in the V3-V4 region. As shown in the results of Fig. 2A and 2B, the diversity of the gut microbiota was significantly increased with the age (Two-year group vs One-year group, Two-year group vs Six-month group, One-year group vs Six-month group), which was expressed by the Shannon and Simpson index (P < 0.05). However, no significant difference in the richness of the gut microbiota (Chao and Ace index) was found among the different age subgroups of 6 months, one year and two-years old ( (P > 0.05, Fig. 2C and 2D). To assess the overall structure of gut microbiota, the score plots of non-metric multidimensional scaling (NMDS) based on unweighted UniFrac distances were constructed. The results in Fig. 2E, 2F and 2G showed that the structures and compositions of the gut microbiota successfully partitioned into three distinct sections from the phylum and genus levels at 6 month, one-year and two-year subgroups. It was also demonstrated by the network ternary-plot in Fig. 2H, 2I and 2J at 6 month, one-year and two-year subgroups from the phylum to genus levels (P < 0.05).
Dynamic colonization of the changes in the gut microbiota between the different age subgroups
We further performed the dynamic colonization of the changes in the gut microbiota between the different age subgroups at the phylum (Fig. 3A) and genus levels (Fig. 4A and 4B). From the phylum level in Fig. 3B, Actinobacteria constituted the dominant phylum in the infant stool at the Six-month group, while Firmicutes was the dominant phylum in the infant faeces at one-year and two-year groups (P < 0.05). However, there was no significant difference at the percents of Bacteroietes among six-month (2.65 ± 1.72)%, one-year (1.54 ± 1.67)% and two-year groups (3.21 ± 4.05)% (P > 0.05). Exactly, the Firmicutes (21.77% vs 41.67% vs 65.27%) (Fig. 3C),Bacteroidetes (2.65% vs 1.54% vs 3.21%) (Fig. 3F) and Verrucomicrobia (0.019% vs 0.18% vs 0.45%) (Fig. 3G) were increased, while the compositions of Actinobacteria (38.44% vs 30.26% vs 23.49%) (Fig. 3D) and Proteobacteria (36.92% vs 26.25% vs7.45%) (Fig. 3E) were decreased from Six-month to Two-year subgroups (Six-month group vs One-year group vs Two-year group) in Fig. 3H (P < 0.05).
To structurally organize and taxonomically annotate the gut microbiota from the genus levels at the six-month, one-year and two-year subgroups (Fig. 4C and 4D), the gut microbiota of Bifidobacterium, Enterobacteriaceae, Klebsiella and Veillonella were much lower, and higher compositions of unidentified_Lachnospiraceae, Faecalibacterium, Blautia, Enterococcus, Subdoligranulum, Agathobacter, Erysipelotrichaceae, Staphylococcus, Acinetobacillus, Ruminococcaceae and Fusicatenibacter in the Two-year group than those in the Six-month and/or One-year subgroups (P < 0.05). While lower levels of Bifidobacterium, Enterobacteriaceae and Klebsiella, with higher levels of unidentified_Lachnospiraceae, Faecalibacterium, Veillonella, Blautia, Subdoligranulum, Agathobacter, Unidentified_ Erysipelotrichaceae, Unidentified_Ruminococcaceae and Fusicatenibacter in the One-year group than those in the six-month group (P < 0.05). However, there was no significant difference at the percent of Lactobacillus among six-month (1.34 ± 0.40)%, one-year (1.21 ± 0.15)% and two-year subgroups (0.90 ± 0.18)% (P > 0.05). Moreover, the significant trends of gut microbiota at the genus level were that the percent of Bifidobacterium(36.93% vs 28.94% vs 21.92%), Enterobacteriaceae (26.03% vs 21.06% vs 4.94%) and Klebsiella (6.98% vs 1.55% vs 0.26%) were significantly decreased, while Faecalibacterium (0.40% vs 3.38% vs 6.47%), Blautia (0.65% vs 2.79% vs 8.29%), Enterococcus (0.48% vs 0.40% vs 2.94%), Subdoligranulum (0.12% vs 1.33% vs 2.61%), Agathobacter (0.15% vs 0.72% vs 4.93%), Unidentified_Erysipelotrichaceae (0.09%vs 0.48% vs 2.85%), Staphylococcus (0.015%, 0.023%, 0.53%), Acinetobacillus (0.012%, 0.019%, 0.51%), Unidentified_ Ruminococcaceae (0.21% vs 0.82% vs2.52%) and Fusicatenibacter (0.12% vs 0.79% vs 3.32%) was increased from six months to two years (Six-month vs One-year vs Two-year group) (Fig. 4E) (P < 0.05).
Functional maturation of the gut microbiota
To determine the functional capacity and trends of the infants gut microbiota developed during the first two year of life, we analyzed the functional maturation of gut microbiota using the KEGG pathways. As shown in Fig. 5B, the gut microbiome evolved into the more complex and adult like configuration in the two-year group, which was significantly reduced inter individual differences than those in the six-month and one-year groups (P < 0.05). As the result of the functional maturing infant gut microbiota in Fig. 5A and 5C, we observed that the functional maturation of gut microbiomes was enriched in Chemoheterotrophy, Fermentation, Animal_parasites_or_symbionts, Human_gut, Mammal_gut, Nitrate_ reduction, Aerobic_chemoheterotrophy, Human_pathogens_all and Human_pathogens _diarrhea at the six-month, one-year and two-year subgroups, which was consisted with more than 75% gut microbiomes in the faeces (P < 0.05). Comparing with the succession of bacterial metabolic function in the Two-year group, Chemoheterotrophy, Fermentation and Animal_parasites_or_symbionts were all higher in the Six-month and One-year groups, which were more than 50% microbiomes in the faeces (P < 0.05). Meanwhile, all intestinal flora related functions in the One-year group was in agreement with those in the Six-month group (P > 0.05).
Associations between the α-diversity and compositions of the gut microbiota over the first two-year of the life
In the regression models (shown in Table 2), the gut microbiota in the stool samples was negatively associated with the significant differences in shannon and simpson at 6 month (r=-0.495, P = 0.0120; r=-0.639, P < 0.001), one-year (r=-0.774, P < 0.001; r=-0.746, P < 0.001) and two-year (r=-0.740, P = 0.00105; r=-0.792, P < 0.001), while the positive correlations were shown between the shannon index at 6 month (r = 0.455, P = 0.0222) and one-year (r = 0.445, P = 0.432) with the maternal BMI increment during the whole pregnancy.
Table 2
Rank-based regression models between the alpha diversity and compositions of the gut microbiota at different age groups (r, P)
Index (n = 30)
|
Shannon
|
Simpson
|
Chao
|
ACE
|
Six-month
|
|
|
|
|
Gut microbiota in the milk
|
-0.495, 0.012
|
-0.639, < 0.001
|
0.181, 0.387
|
0.124, 0.555
|
BMI at six-month
|
0.078, 0.712
|
-0.220, 0.291
|
0.288, 0.162
|
0.279, 0.177
|
BMI at birth
|
-0.160, 0.445
|
-0.237, 0.254
|
0.175, 0.404
|
0.103, 0.624
|
BMI increment of the infant
|
0.302, 0.143
|
0.172, 0.410
|
0.089, 0.671
|
0.142, 0.500
|
Gestational weeks
|
-0.055, 0.795
|
0.071, 0.735
|
0.031, 0.886
|
0.020, 0.924
|
Maternal age at pregnancy
|
0.112, 0.596
|
0.325, 0.113
|
-0.259, 0.212
|
-0.196, 0.348
|
BMI at the pre-pregnancy
|
0.004, 0.984
|
0.159, 0.448
|
0.037, 0.859
|
-0.005, 0.981
|
BMI increment during the
whole pregnancy
|
0.455, 0.022
|
0.206, 0.323
|
0.394, 0.051
|
0.406, 0.044
|
One-year
|
|
|
|
|
Gut microbiota in the milk
|
-0.774, < 0.001
|
-0.746, < 0.001
|
0.014, 0.951
|
0.076, 0.743
|
BMI at one-year
|
0.073, 0.754
|
0.177, 0.444
|
0.167, 0.468
|
0.186, 0.420
|
BMI at birth
|
-0.096, 0.679
|
-0.113, 0.626
|
0.226, 0.325
|
0.261, 0.253
|
BMI increment of the infant
|
-0.021, 0.929
|
0.103, 0.658
|
-0.183, 0.427
|
-0.203, 0.378
|
Gestational weeks
|
0.174, 0.452
|
-0.102, 0.659
|
0.366, 0.103
|
0.423, 0.0559
|
Maternal age at pregnancy
|
0.151, 0.515
|
0.185, 0.421
|
-0.066, 0.775
|
-0.031, 0.894
|
BMI at the pre-pregnancy
|
0.222, 0.335
|
-0.094, 0.687
|
0.213, 0.354
|
0.238, 0.299
|
BMI increment during the
whole pregnancy
|
0.445, 0.043
|
0.335, 0.137
|
0.417, 0.060
|
0.424, 0.056
|
Two-year
|
|
|
|
|
Gut microbiota in the milk
|
-0.740, 0.001
|
-0.792, < 0.001
|
0.033, 0.905
|
-0.128, 0.638
|
BMI at two-year
|
-0.044, 0.871
|
-0.103, 0.704
|
-0.050, 0.854
|
0.027, 0.922
|
BMI at birth
|
0.159, 0.557
|
0.179, 0.506
|
-0.171, 0.528
|
-0.138, 0.610
|
BMI increment of the infant
|
-0.246, 0.359
|
-0.306, 0.249
|
0.081, 0.766
|
0.146, 0.591
|
Gestational weeks
|
-0.061, 0.824
|
0.012, 0.964
|
0.051, 0.851
|
-0.068, 0.801
|
Maternal age at pregnancy
|
-0.047, 0.862
|
-0.025, 0.926
|
-0.208, 0.439
|
-0.120, 0.659
|
BMI at the pre-pregnancy
|
-0.474, 0.064
|
-0.400, 0.135
|
-0.459, 0.074
|
-0.547, 0.028
|
BMI increment during the
whole pregnancy
|
-0.362, 0.169
|
-0.418, 0.107
|
0.224, 0.405
|
0.153, 0.572
|
Associations between quantitative characteristics and the compositions of the gut microbiota in the infant , faeces over the first two-year of the life
The compositions of the infant gut microbiota in the faeces over the first two-year of the life at the phylum and genus levels (Fig. 6) showed some associations with the maternal and infant characteristics. As shown in Fig. 6A and Fig. 6B when the infants were 6 months old, the compositions of gut microbiomes (Actinobacteria and Cyanobacteria-positively, Firmicutes and undentlfied_Bacteria -negatively) were significantly associated with that in the breast milk (GM) (P < 0.05), while the BMI at birth (B.BMI) could affect the percents of Firmicutes (negatively), Actinobacteria (positively) and Verrucomicrobia (positively) (P < 0.05), and the percent of Tenericutes was negatively correlated with the BMI at the 6 months old (S.BMI) and positively correlated with the maternal pre-pregnancy BMI (M.BMI) in the faeces at the phylum level (Fig. 6A) (P < 0.05). What is more, positively correlation between the composition of Chloroflexi and gestational week (G), negatively correlation between the composition of Undentlfied_Bacteria and maternal age (MA) (P < 0.05), and positively correlations between the composition of Bacteroidetes and BMI increment during the pregnancy (I.BMI) were also demonstrated in Fig. 6A (P < 0.05). While at the genus level, the compositions of gut microbiomes (Bifidobacterium and Rothia-positively, undentlfied_ Lachnospiraceae, Blautia, Delftia and undentlfied_Erysipelotrichaceae- negatively) were significantly associated with the GM (P < 0.05), while the compositions of Bacteroidetes, Collinsella and Flavonifractor were positively correlated with that in I.BMI (P < 0.05). And the percent of gut microbiota were negatively associated with the MA (Veillonella, Granulicatella and Subdoligranulum) and M.BMI (Blautia, Faecalibacterium and Agathobacter) (P < 0.05). What is more, B.BMI could affect the percents of Bifidobacterium (positively), Rothia (positively) and undentlfied_ Lachnospiraceae (negatively) in the Fig. 6B (P < 0.05).
When the subjects were one year old, the compositions of gut microbiomes at the phylum level (Firmicutes-negatively and Actinobacteria-positively) were significantly associated with GM (P < 0.05), while there were positive correlations between maternal and infant characteristics and the compositions of the gut microbiota in the infant, faeces (Fusobacteria and B.BMI, Verrucomicrobia and G)(P < 0.05), which were also shown of the negative correlations (Fusobacteria and I.BMI, Chloroflexi and G. BMI) in the faeces (Fig. 6C) (P < 0.05), while at the genus level, the compositions of gut microbiomes (Bifidobacterium, Klebsiella, Streptococcus, Enterococcus and Rothia-positively, Blautia, Hungatella, Undentlfied_ Lachnospiraceae, Undentlfied_Erysipelotrichaceae and Anaerostipes-negatively) were significantly associated with that in GM (P < 0.05), while the compositions of Klebsiella and Fusobacterium were negatively correlated with the I.BMI, Agathobacter and Butyricicoccus were positively associated with M.BMI (P < 0.05). What is more, B.BMI could affect the percents of Fusobacterium (positively), Subdoligranulum (positively) and Undentlfied_Clostridiales (negatively)(P < 0.05). Furthermore, the compositions of gut microbiomes (Agathobacter, Akkermansia, Collinsella, Undentlfied_Ruminococcaceae, Fusicatenibacter and Butyricicoccus- positively, Megasphaera and Lachnoclostridium-negatively) were significantly associated with that in the G in the Fig. 6D (P < 0.05).
When the subjects were two years old, the compositions of gut microbiomes at the phylum level (Firmicutes and Bacteroidetes- negatively, Actinobacteria and Proteobacteria-positively) were significantly associated with the GM, while the G and M.BMI could respectively affect the percents of Tenericutes and Cyanobacteria in the faeces (Fig. 6E) (P < 0.05). From the genus level, there were positively correlations between the maternal and infant characteristics and the compositions of the gut microbiota (GM and Bifidobacterium/Undentlfied_ Enterobacteriaceae, B.BMI and Faecalibacterium/Roseburia, G.BMI and Undentlfied _Rhizobiaceae, M.BMI and Streptococcus) (P < 0.05), which were also proved the negative associations (GM and Bacteroidetes/Subdoligranulum/Fusicatenibacter/Undentlfied_Erysipelotrichaceae/ Undentlfied_Ruminococcaceae/Parabacteroides/Lachnospira/Sarcina/Dorea/Butyricicoccus, T.BMI and Romboutsia/Undentlfied_Prevotellaceae, G.BMI and Blautia/ Roseburia, M.BMI and Undentlfied_Lachnospiraceae/Romboutsia/Butyricicoccus, I.BMI and Anaerostipes) in the faeces (Fig. 6F) (P < 0.05).
Effects of qualitative characteristics on the profiles of faecal gut microbiota in the infants
Significant shifts in the composition of profiles of faecal gut microbiota in the infants were observed through stratified intervention including the sex (Fig. 7) and mode of production (Fig. 8) using a permutational ANOVA of unweighted and weighted Unifrac distance in this birth cohort. Concretely, the α-diversity index (Shannon, Simpson, Chao and ACE, Fig. 7A), β diversity index (PCA analysis) and the compositions of gut microbiota from the phylum and genus levels levels were not significant differences between the Boy and Girls groups when they were six months (Fig. 7B, 7E, 7H and 7k), one year (Fig. 7C, 7F, 7I and 7L) and two years old (Fig. 7D, 7G, 7J and 7M) (P > 0.05), in which all above indicators were not significantly different between the caesarean (CB) and natural birth groups (NB) when they were six months (Fig. 8) (P > 0.05).
Compositions of the gut microbiota in the breast milk and their correlations with the maternal characteristics
As shown in Fig. 9, the percents of Proteobacteria (60.94 ± 5.11)%, Firmicutes (23.81 ± 4.30)%, Bacteroidetes (9.56 ± 2.98)%, Cyanobacteria (1.36 ± 0.61)% and Actinobacteria (0.76 ± 0.12)% constituted the dominant gut microbiota in the breast milk at the phylum level, which was accounted for more than 90% of the community (Fig. 9A and 9B). Meanwhile, Bacillus (3.59 ± 2.99)%, Sphingomonas (12.61 ± 4.64)%, Staphylococcus(4.52 ± 2.21)%, unidentified_Alphaproteobacteria (4.69 ± 2.74)%, Brevundimonas (9.52 ± 2.40)% and Streptococcus (8.45 ± 2.13)% were the main dominant percents of the gut microbiota at the genus level in the breast milk (Fig. 9C, 9D and 9E). Then the associations between maternal characteristics (G, MA, M.BMI and I.BMI) and the gut microbiomes in the milk were discussed in Fig. 9H, 9I and Table 3, which were proved that the percents of Proteobacteria were negatively associated with the M.BMI and I.BMI. Meanwhile, there were significantly positive correlations between the I.BMI and the compositions of Firmicutes and Actinobacteria at the phylum level by the regression models (Fig. 9F) (P < 0.05). From the genus level (Fig. 9G), the M.BMI was positively correlated with the percents of Sphingomonas and Bacillus, and negatively correlated with Limnobacter and Brevundimonas (P < 0.05). Meanwhile there was still positive significance between the composition of unidentified_Bacteroidales and I.BMI .
Table 3
Correlations between the compositions of gut microbiota in the breast milk and the factors of maternal characteristics (r, P)
|
Gestational age
|
Age of pregnancy
|
Pre-pregnancy
BMI
|
BMI increment
|
Phylum
|
|
|
|
|
Proteobacteria
|
0.035, 0.868
|
0.004, 0.984
|
-0.493, 0.012
|
-0.393, 0.049
|
Firmicutes
|
0.005, 0982
|
-0.197, 0.345
|
0.237, 0.253
|
0.393, 0.049
|
Bacteroidetes
|
0.129, 0.540
|
0.088, 0.674
|
0.238, 0.252
|
0.147, 0.482
|
Cyanobacteria
|
-0.137, 0.512
|
0.302, 0.142
|
0.278, 0.179
|
-0.005, 0.981
|
Actinobacteria
|
-0.058, 0.783
|
0.207, 0.314
|
0.206, 0.323
|
0.471, 0.018
|
Tenericutes
|
-0.100, 0.636
|
0.086, 0.682
|
-0.035, 0.869
|
0.203, 0.331
|
unidentified_Bacteria
|
0.303, 0.141
|
-0.002, 0.993
|
0.365, 0.073
|
0.320, 0.118
|
Spirochaetes
|
0.329, 0.109
|
0.092, 0.661
|
0.286, 0.165
|
0.245, 0.237
|
Melainabacteria
|
0.278, 0.179
|
0.103, 0.626
|
0.178, 0.395
|
0.196, 0.349
|
Fibrobacteres
|
-0.132, 0.531
|
0.131, 0.532
|
-0.004, 0.984
|
0.213, 0.307
|
Genus
|
|
|
|
|
Bacillus
|
0.057, 0.789
|
0.018, 0.929
|
0.398, 0.047
|
-0.054, 0.796
|
Sphingomonas
|
-0.075, 0.722
|
-0.131, 0.532
|
-0.312, 0.129
|
-0.272, 0.189
|
Staphylococcus
|
0.128, 0.543
|
0.022, 0.918
|
0.309, 0.133
|
0.264, 0.203
|
unidentified_Alphaproteobacteria
|
0.022, 0.916
|
0.134, 0.523
|
-0.070, 0.740
|
0.108, 0.606
|
Brevundimonas
|
-0.027, 0.897
|
-0.150, 0.473
|
-0.480, 0.015
|
-0.112, 0.595
|
Streptococcus
|
-0.002, 0.992
|
-0.167, 0.424
|
0.085, 0.688
|
0.162, 0.439
|
Limnobacter
|
-0.004, 0.984
|
-0.153, 0.466
|
-0.465, 0.019
|
-0.131, 0.532
|
Enhydrobacter
|
-0.237, 0.255
|
-0.105, 0.617
|
0.036, 0.866
|
0.111, 0.596
|
Acinetobacter
|
0.006, 0.979
|
0.002, 0.991
|
0.036, 0.864
|
-0.198, 0.342
|
Sphingobacterium
|
0.255, 0.219
|
0.052, 0.805
|
0.400, 0.037
|
-0.058, 0.783
|
Pseudomonas
|
-0.136, 0.518
|
0.017, 0.935
|
-0.203, 0.331
|
-0.032, 0.878
|
Lactobacillus
|
-0.223, 0.284
|
-0.094, 0.655
|
-0.273, 0.187
|
0.239, 0.251
|
unidentified_Cyanobacteria
|
-0.148, 0.482
|
0.278, 0.178
|
0.285, 0.167
|
-0.033, 0.877
|
Stenotrophomonas
|
0.246, 0.235
|
-0.098, 0.641
|
-0.004, 0.984
|
-0.037, 0.861
|
unidentified_Bacteroidales
|
0.032, 0.878
|
0.016, 0.942
|
0.032, 0.880
|
0.386, 0.047
|