Discussion of Hartman et al. (2009)

This web page was produced as an assignment for an undergraduate course at Davidson College.



Human intestines house a diverse population of microbes necessary for food digestion.  The dynamics of microbial ecology are especially critical in patients in need of small bowel transplants (SBTs), a procedure to lengthen the small bowel in order to facilitate sufficient nutrient absorption. Despite immunosuppression therapies, SBT patients are at high risk for graft rejection and bacterial sepsis, but the role of microbiota in these patients is still unknown. Researchers do know that mutations in the NOD2 gene increase the likelihood of graft rejection and sepsis in SBT patients. NOD2 encodes a protein that senses microbial products and causes production of antimicrobial peptides. The purpose of this study was to characterize the microbial population that re-colonizes the small bowel after separation from the donor and transfer to a new environment, and to find correlations between a population shift and clinical observations like the NOD2 genotype.


Illustration of an ileostomy with an ileostomy bag. A loop of the ileum is pulled out through the abdominal wall. An ileostomy bag is attached to collect waste. Photo courtesy of CancerHelp UK.

In order to investigate this, Hartman and colleagues designed a novel method of data collection. After a SBT, doctors gain access to the transplant site for observation by creating an ileostomy, a surgical opening exposing a loop of the ileum for endoscopy, biopsies and waste removal. Hartman and colleagues collected ileal effluent and biopsied tissue samples from SBT patients over at least 10 post-transplant time points. From these samples, they were able to study microbial diversity by quantitative PCR (qPCR) of 16S rRNA. This brief webpage summarizes the figures presented in this paper, and presents my conclusions about the article.

You can access the full text article for Hartman et al. (2009).


Figure Summaries

Figure 1. Bacterial populations in 17 patients were measured by qPCR at the time points (in days) given on the x axes. In all patients the bacterial population is composed of 4 different bacterial orders – Lactobacilli, Enterobacteria, Clostridia, and Bacteroides. The relative proportions of the four orders and the total number of bacteria vary over time and between patients.  Even so, in patients with open ileostomies, the fractional majority of bacteria is generally composed of Lactobacilli and Enterobacteria, which Hartman explains are facultative anaerobes. Clostridia and Bacteroides, strict anaerobes, usually constitute the majority of the ileal bacterial population, indicating that there is an inversion of population dynamics in these patients. In patients that had their ileostomy closed, Clostridia and Bacteroides reclaim the majority of the population, suggesting ileostomies cause a dynamic shift in the microbial ecology in the small bowel.


Figure 2. Bacterial population fractions were compared in clinical conditions involving an ileostomy and not involving an ileostomy to visualize any correlation. Comparison of transplant patients to non-transplant patients all having ileostomies shows that the majority of bacteria are Lactobacilli and Enterobacteria regardless of transplant procedure. Moreover, in patients that underwent an ileum biopsy but lacked an ileostomy, Clostridia and Bacteroides constitute the majority; this dynamic is inverted in a biopsy involving a postsurgical ileostomy. Finally, after ileostomy closure, the majority returns to normal – Clostridia and Bacteroides being the majority. This figure indicates that the population shift seen in these SBT patients is due to the presence of the ileostomy, not the transplant procedure or biopsy procedure.

Hartman et al. Figure 2.


Figure 3. For this figure, the information given in the caption seems incongruent with information in the figure and the text. The purpose of this figure is to illustrate whether the bacterial populations sampled from ileal effluent is representative of the population found on host tissue. The ratio of each bacterial order found in biopsy sample compared to effluent samples was reportedly calculated by log10(biopsy percentage / effluent percentage), so a negative value should indicate a higher population was found in the effluent, and a positive value should indicate a higher population in the biopsy. Thus, the figure would indicate that higher populations of Enterobacteria and Clostridia were found in biopsy samples compared to effluent samples, and higher populations of Bacteroides were found in effluent samples compared to biopsy samples. Similar populations of Lactobacilli were obtained from both collection methods. Graft rejection does not seem to correspond with bacterial population in either the effluent or biopsied tissue. This reading of the figure would suggest that Bacteroides occupies a more planktonic niche in the intestine. However, the figure labels the negative side of the graph as “biopsy” and the positive side as “effluent”; in correspondence with this, the results section states that Bacteroides show higher levels in biopsy specimens.

It is difficult to distinguish whether the caption or the figure label is incorrect, or if I am misinterpreting the data. Even so, the figure suggests that sampling the effluent alone for three of the four bacterial orders is not similar to a biopsy sample, because most of the points do not touch the 0 reference line ( log10(1) ). A statistical test on these ratio values could show that they are not significantly dissimilar, but no statistical values are available. So I do not agree with the authors’ conclusion that the effluent and biopsy compositions are “quite similar”.

Hartman et al. Figure 3.


Figure 4. Different microbial communities pre- and post-closure of the ileostomy suggested to the authors there might be an observable difference in metabolism. Metabolomic profiling identified 139 compounds that were used to construct a metabolic network diagram. Most metabolites were enriched during pre-closure of the ileostomy. Notable enriched compounds were pyruvate, glutamic acid, and alpha ketoglutaric acid, all involved in the Krebs cycle. Lactic acid was highly enriched in preclosure, which corresponds with an abundance of Lactobacilli. In post-closure, phosphoric acid and methanolphosphate were highly enriched. Hartman says that no possible role is understood for the other highly enriched compounds, suggesting microbial function in humans is extensive, but still not fully understood.


Figure 5. This figure is a schematic diagram summarizing the data found in this study by illustrating the stable states of the small bowel microbial ecosystem. In its normal state, the small bowel is dominated by strict anaerobes due to the low oxygen concentrations. In an open ileostomy, the bowel is dominated by facultative anaerobes due to a high concentration of oxygen. The less stable transition state between these resembles a closed ileostomy, in which facultative anaerobes are still the majority of the oxygenated bowel, but their fraction decreases as oxygen concentrations decrease. Thus this figure shows how the population dynamics are inverted based on the oxygen tension in the small bowel due to the ileostomy procedure.


Hartman et al. Figure 5.


Hartman et al. found that post-surgical patients with ileostomies have a microbial population inversion from strict anaerobes, Clostridia and Bacteroides, to facultative anaerobes, Lactobacilli and Enterobacteria. This microbial composition reverts to normal after the ileostomy is closed. I thought the figures adequately illustrated this dramatic shift, though Figure 3 was difficult to follow because of the caption and axis labeling. For this figure, I also think it would be beneficial to show some sort of statistical test to designate what value range denotes “similar” and what denotes “different” in the comparison. Even so, I do not think this figure affects the validity of the authors’ conclusion; since all samples were taken from effluent, the dramatic inversion of the effluent microbial population is still significant, even if that inversion may not correspond precisely with the population of all bacteria found in the ileum. Since this figure did not seem to me to validate the main conclusions of the article, and the authors spend very little time discussing the figure except regarding Bacteroides, I think it may be unnecessary to have this figure in the main article instead of supplemental material.

Overall, I agree with the conclusions of Hartman et al., that an ileostomy dramatically alters the ileal microbial composition. I did not find it surprising that a shift in oxygen tension alters the microbial composition of anaerobes and aerobes, but it was interesting that this dramatic shift in microbial composition can cause fatal sepsis after ileostomy closure. Because of this, I think the findings of this paper are critical in re-evaluating the use of ileostomies after small bowel transplants, and generally understanding the microbial shifts that can affect humans.



Hartman AL, Lough DM, Darupa DK, Fiehn O, Fishbein T, Zasloff M, Eisen JA. Human gut microbiome adopts an alternative state following small bowel transplantation. Proc Natl Acad Sci USA 2009; 106(40): 17187-92.


Return to Olivia's Genomics Page
Genomics Page
Biology Home Page

© Copyright 2010 Department of Biology, Davidson College, Davidson, NC 28035