In recent years, there has been increasing interest in the discovery of correlations between bacterial communities in the human gut and disease [2
]. Several studies have shown a correlation between gut bacteria communities and the development of autoimmune diseases. In Crohn’s disease, a T-RFLP fingerprinting method was used to shown that bacterial diversity is higher in stool from healthy subjects compared to their twin diseased subjects. Two rodent models for type 1 diabetes have shown that feeding antibiotic or probiotics can delay the onset of the disease [1
]. In addition, a germ-free environment increases the incidence of diabetes in NOD mice [28
These recent results encourage the use of human stool samples to search for bacteria that are correlated, negatively or positively, with the onset of type 1 diabetes in children. Two studies have collected such samples that could be used for this purpose. The Diabetes Prediction and Prevention study (DIPP) in Finland has been collecting stool and blood samples quarterly (beginning soon after birth) from children at high risk for the disease since 1994. A larger prospective study in the United States, Germany, Sweden, and Finland called The Environmental Determinants of Diabetes in the Young (TEDDY) (seeking to determine possible environmental triggers of autoimminty) has been collecting samples since 2004 from genetically high risk children on a monthly basis [30
]. In both studies, some have seroconverted and developed type 1 diabetes.
The stool samples from the DIPP and TEDDY studies could be very useful in determining whether any bacteria are correlated with both the triggering of the autoimmune process and/or the onset of type 1 diabetes. However, to be useful, the conditions under which they were collected need to be examined to be certain that collection protocols to not impact the bacterial contents of these samples.
Four samples were taken from healthy children enrolled in the DIPP study. Each sample was divided into five aliquots. One was immediately frozen at -80°C while the others were frozen after 12, 24, 48, and 72h. The 72h period is the maximum time required to collect a sample from the infant or child until it is frozen at a DIPP or TEDDY repository. During that period of transit, the sample sits at or near room temperature. In order for DIPP or TEDDY samples to be useful, it is important to know whether the bacterial community composition changes over the 72-hour period.
Over a 72-hours, we showed that bacterial communities only change by ~10% with most change beginning between 12 and 24 hours. As expected, there was enormous variability from person to person in both bacterial community composition and in the rate of change in composition over time. This suggests that nutrient availability in stool is very different from sample to sample, thereby affecting growth rates at room temperature over time.
In addition, changes are observed only in the most abundant taxa. The samples are so diverse that no statistical changes are observed over 72hr if the whole community is considered. However, careful examination of the dominant taxa shows that changes are occurring.
These data suggest that new experiments may be needed to examine these issues further. For example, what is the effect of storage temperature over time? That is, could a delivery time of 2-3 days from the time of collection to the time of permanent storage be tolerated if the samples were immediately placed on ice packs? Is storage at 4°C or -20°C during transit cold enough to minimize changes in community composition?
Finally and most importantly, are the changes observed here over time, large enough to mask changes that may occur over time in one person during the several month period of seroconversion? TEDDY metadata shows that 23.9% of all samples collected in the Georgia, Florida, Colorado, and Washington State clinics are frozen within 24h of collection. Thus, if AMOVA shows that the changes that occur in bacterial community composition during seroconversion are much greater than 10%, the current DIPP and TEDDY samples may be very useful, particularly the large number of samples frozen within 24h.