In the present study we found significant differences in the fecal microbial composition between healthy children and children with type 1 diabetes. We are unaware of any other similar studies in children with type 1 diabetes using simultaneously DGGE molecular profiling, unweighted pair group method with arithmetic mean algorithm dendrogram construction, sequencing and real-time qPCR analysis. To determine the characteristics of the gut microbiota based on the condition of just type 1 diabetes, we excluded the influence of physiological factors such as age, gender, dietary habits and race. In addition, we also controlled for the mode of delivery at birth and the duration of breastfeeding. This was because the first year of life has a crucial impact on gut microbiota composition and epidemiological studies in humans at genetic risk for type 1 diabetes have suggested that a short duration of breastfeeding and early feeding in infancy with complex dietary proteins such as cow's milk proteins can modulate the development of beta cell autoimmunity, clinical type 1 diabetes, or both [40–42]. No significant differences were found between the two groups of children (type 1 diabetes and controls).
The DGGE analysis of the fecal microbiota revealed a significantly lower intra-group similarity index in children with diabetes than in healthy children. In other words, the DGGE profiles in healthy children were more similar to each other, whereas in children with diabetes they were less similar. A similar result was found by Giongo et al. . These data suggest that diabetic status may influence specific bacterial groups of the gut microbiota community.
Sequence analysis of the DGGE bands cloned enable the association of specific bacterial genotypes with health or diabetes situations. Consistent with previous human and animal studies [11, 21, 39, 37, 43], the gut microbiota of healthy children and children with diabetes was predominately composed of Firmicutes and Bacteroidetes and the main difference lies in the proportion of genus-division bacteria within this two phyla and the Actinobacteria phylum between both the group with diabetes and the healthy group. These results suggest that the dominant microbiota genera are different in children with type 1 diabetes compared with healthy children. Recently, three robust clusters, referred to as "enterotypes", which are not nation or continent specific have been identified. Assignment of an individual microbiome into a given enterotype is based upon the relative enrichment of that microbiome in one of three genera: Bacteroides (enterotype 1), Prevotella (enterotype 2) or Ruminococcus (enterotype 3) . In this study, within Bacteroidetes, the Bacteroides genus was prevalent in the diabetic group, whereas the Prevotella genus was associated with the healthy group. Thus, the type 1 diabetic gut microbiomes could be classified into enterotype 1 and the healthy microbiomes could be classified into enterotype 2.
As DGGE is considered a semiquantitative tool for monitoring the dynamics of the predominant bacterial species of fecal microbiota, an additional analysis with real-time qPCR was performed to obtain a quantitative estimation of the changes found in the gut microbiota between children with diabetes and healthy children. We noted significant quantitative differences between the major microbial phyla present in the feces of healthy children and those with diabetes. In contrast to the situation in healthy children, we found a significant increase in the quantity of Bacteroidetes and a significant decrease in the number of Firmicutes and Actinobacteria in children with type 1 diabetes. Our data showed a significantly lower Firmicutes to Bacteroidetes ratio in children with type 1 diabetes compared with healthy children. Moreover, we saw a negative correlation between this ratio and both the glucose and the HbA1C levels in children with diabetes, which could help to explain the significantly higher glycemic level in this group. In agreement with this, Giongo et al. observed that the Firmicutes to Bacteroidetes ratio in study participants with type 1 diabetes was changing during the first 6 months after birth before the development of the autoimmune disease. These authors showed a successive decline in Firmicutes and an increase in Bacteroidetes number in the gut microbiome over time until the children became diabetic . Moreover, this imbalance observed at the phylum level between Bacteroidetes and Firmicutes has been previously described in several human disorders. A decline in the Firmicutes to Bacteroidetes ratio compared with controls has been described in human type 2 diabetes , whereas in Crohn´s disease, both Bacteroidetes and Firmicutes seem to decline . The opposite happens in obesity, where the imbalance is due to the increase in the Firmicutes to Bacteroidetes ratio , indicating that obesity and diabetes are associated with different groups of intestinal microbiota.
However, the major difference between the two groups was found in the number of bacteria at genus-division level. The most remarkable result was the significant increase in the number of Clostridium, Bacteroides and Veillonella in the children with diabetes, whereas the number of Lactobacillus, Bifidobacterium, the Blautia coccoides/Eubacterium rectale group and Prevotella genus were all significantly decreased in children with diabetes. Our findings concerning the microbiota of children with diabetes are in line with observations in other animal studies. Roesch et al. found higher levels of Lactobacillus and Bifidobacterium in BioBreeding diabetes-resistant rats whereas Bacteroides and Clostridium were more abundant in BioBreeding diabetes-prone rats . In contrast with this, however, Brown et al. found that Lactobacillus and Bifidobacterium were more abundant in participants with type 1 diabetes than in healthy participants .
The significant decrease in the number of Lactobacillus and Bifidobacterium observed in children with type 1 diabetes in our study was associated with their higher levels of plasma glucose, as indicated by the negative correlation found. Also, the regression analysis showed that the decrease in the number of Lactobacillus and Bifidobacterium could be associated with the plasma glucose level in the children with diabetes. In previous studies, the levels of Bifidobacterium have also been related to improved glucose metabolism, insulin resistance and low-grade inflammation [48, 49]. Moreover, Valladares et al. determined that the administration of Lactobacillus johnsonii isolated from BioBreeding diabetes-resistant rats delays or inhibits the onset of type 1 diabetes in BioBreeding diabetes-prone rats .
Both Lactobacillus and Bifidobacterium have members with probiotic characteristics and these have been associated with positive effects for the host in the large intestine . In addition, both bacterial groups have the capacity to produce the beneficial organic acid lactate, which is converted into butyrate by butyrate-producing bacteria in the gut . Barcenilla et al.  showed that most of the butyrate-producing isolates from human fecal samples are related to the Blautia coccoides-Eubacterium rectale group. Previous studies have shown that butyrate induces mucin synthesis (a glycoprotein produced by the host that could maintain the integrity of the gut epithelium) , decreases bacterial transport across the epithelium , and improves gut integrity by increasing tight junction assembly . In addition, the genera Prevotella are responsible for the degradation of this mucin ; thus, the significant decline in the numbers of the Blautia coccoides-Eubacterium rectale group and Prevotella that we found in children with type 1 diabetes compared with healthy children could indicate a reduction in mucin synthesis by the host and a lack of this mucin on the epithelial layer of the gut, which would lead to a significant alteration in intestinal permeability. Other studies have described an association between type 1 diabetes and compromised barrier permeability in humans and both the NOD mouse and BioBreeding rat models [16–18, 20].
The significant increase in the number of Clostridium, Bacteroides and Veillonella in the children with diabetes with respect to the healthy children was accompanied by a significant positive correlation between both the plasma levels of glucose and HbA1c and the quantity of Clostridium. These bacteria are able to ferment glucose and lactate to propionate, acetate and succinate. However, these short fatty acids do not induce mucin synthesis . This situation would, though, reduce the tight junction assembly, generating an increase in the gut permeability in children with type 1 diabetes .
Finally, we propose a possible mechanism to explain the relationship we have found between the gut microbiota present in children with type 1 diabetes and the glycemic levels observed. The short-chain fatty acids (such as butyrate and propionate) formed by this gut microbiota have a role in the regulation of the levels of gut hormones such as glucose-dependent insulinotropic polypeptide, glucagon-like peptide 1 and ghrelin. These hormones have important effects on carbohydrate metabolism , thus allowing gut microbiota to affect glycemic levels. In addition, Huml et al. have previously demonstrated an altered secretion pattern of gut hormones in children with type 1 diabetes that may impact on the metabolic control of diabetes in these patients . Further studies will be necessary to demonstrate this proposed mechanism.
A limitation of the 16S rRNA gene-based method is that the function of the identified bacteria is unknown. Future studies using a microbial metagenomic sequencing analysis will be carried out to obtain information about the functional diversity of the bacterial community analyzed here.