Page 26 - Read Online

Page 4 of 13 Huang et al. Microbiome Res Rep 2024;3:30 https://dx.doi.org/10.20517/mrr.2024.14

Scientific) by extracting ion chromatograms using a 5 ppm mass tolerance and referencing an in-house
retention-time library of chemical standards. Peak area ratios for the analytes were calculated, in which the
raw peak area for each metabolite was divided by the raw peak area of the appropriate internal standard, i.e.,
the isotopically labeled amino acids [Supplementary Tables 1 and 2].

Metabolomic data analysis
The targeted metabolomic data were analyzed using Metaboanalyst (https://www.metaboanalyst.ca/)
following the protocol . Briefly, the data (peak to internal standard ratio, Supplementary Tables 1 and 2)
[21]
were uploaded, normalized by sum, transformed (log transformation, base 10), and auto-scaled (mean-
centered and divided by the standard deviation of each variable). After that, the normalized data were
clustered using the following parameters (distance calculation, Euclidean; clustering method, Ward; color
contrast, default) and visualized as clustered heatmap.

Transcription factor analysis
A list of transcription factors was retrieved from multiple databases [22,23] and compiled in Supplementary
Tables 3 and 4. The gene expression of these transcription factors in colonic epithelial cells was retrieved
from our previous study and compared under different conditions [Supplementary Tables 3 and 4].
[15]
Briefly, a comparison was performed between GuMI with no bacteria (GuMI-NB) and Static culture and
between GuMI with F. prausnitzii (GuMI-FP) and GuMI-NB. Fold change and adjusted p values were
calculated as previously reported and used to identify the effects of microenvironment and bacterium
[15]
F. prausnitzii. A gene is deemed significantly changed in one comparison if its adjusted P-value is less than
0.05 and the absolute value of fold change exceeds 2.

RESULTS
GuMI is versatile for coculture of bacteria and colonic epithelium derived from multiple donors
[15]
We previously demonstrated a successful coculture of F. prausnitzii with human colonic epithelial cells
and innate immune cells under physiologically relevant oxygen gradient and flow. The colonic epithelium
[16]
was derived from healthy tissue. To test whether GuMI can accommodate bacteria-host coculture from
multiple donors, we prepared organoid-derived monolayers from the uninflamed colonic region of the
[20]
patient with Crohn’s disease based on the previously reported protocol . Monolayers were cultured under
three conditions for 4 days [Figure 1A], i.e., Static culture, GuMI-NB, and GuMI-FP. Interestingly, the
TEER values of monolayers in GuMI were significantly higher than those in Static culture (P < 0.001,
Figure 1B). Additionally, TEER values in GuMI-FP were higher than those in GuMI-NB (P < 0.05,
Figure 1B). In agreement, we observed dark spots in the monolayers under Static culture, whereas there
were fewer in the monolayers in GuMI-NB and GuMI-FP [Figure 1C]. Correspondingly, F. prausnitzii
actively grew in GuMI, with live bacterial cell density increasing from 5.7 × 10 to 2.3 × 10 CFU/mL at day 2
5
7
after bacterial injection [Figure 1D]. For a donor of ulcerative colitis (HC465, Figure 1E), we compared
head-to-head with a non-UC donor HC2978 in coculturing F. prausnitzii for four days in GuMI with
dendritic cells and macrophages . TEER values are not significantly different between the two donors
[16]
[Figure 1F], and F. prausnitzii grew significantly after four days of coculture [Figure 1G]. Although these
pilot data suggest that GuMI can accommodate colonic epithelium-bacterium coculture for multiple donors
for up to 4 days, these experiments are performed with limited repeats due to COVID disruption. Further
experiments are warranted to validate the broad applicability of the GuMI model.

GuMI and F. prausnitzii alter the chemical profile
As demonstrated in our previous study, the characteristic metabolite butyrate was produced by
F. prausnitzii in the apical compartment of GUMI . As a proof of concept, we confirmed that the NF-κB
[15]
pathway is suppressed by butyrate . In addition to the NF-κB pathway, more than 4,000 genes were
[15]