Table 1.
Demographics of NEC and control patients.
Table 2.
Origins of control small intestinal tissue samples.
Figure 1.
Reduced proportions of γδ IEL subsets in patients with NEC compared to non-NEC surgical controls.
(A) Example of the gating strategy used to calculate proportions of γδ IEL subsets. The control sample shown is from a 4 days old 26 weeks gestation infant with spontaneous (focal) intestinal perforation and the NEC sample is from a 15 days old 28 weeks gestation infant with surgical NEC. Gates were set on “live”, CD14−, CD19− (“Dump” negative) and CD3+ cells before applying to sub-populations. Next we identified CD3+ CD8+ T cells followed by differentiating conventional CD3+ CD103+ TCRαβ from TCRγδ IEL (γδ IEL). The patient with NEC showed significant reduction in γδ IEL with a corresponding greater proportion of αE integrin (CD103) negative, conventional T cells. Dot plot of total IEL (B) and γδ IEL (C) proportions were statistically significantly reduced in NEC tissue compared to non-NEC controls, p<0.001 and p = 0.02, respectively.
Figure 2.
Developmental regulation of γδ IEL subsets in humans.
Logarithmic transformed percentages of γδ IEL were plotted against gestational age (GA), postmenstrual age (PMA = gestational age plus chronological age) and age. Using Pearson's correlation coefficient we did not detect any association of γδ IEL proportions with GA, PMA or age in either NEC or non-NEC control patients.
Figure 3.
Window of susceptibility with low γδ IEL subsets in human neonates.
Graph depicting the relationship between percentage of γδ IEL and gestational age in non-NEC surgical control samples. The black line shows the curve estimate of the U-shaped association as modeled by nonlinear regression with 95% confidence intervals of the model curve shaded in grey. The lowest percentage of γδ IEL occurs between 27 and 32 weeks gestation.
Figure 4.
γδ IEL are the predominant IEL subtype in the immature murine small intestine.
A) Flow cytometry of distal small intestinal intraepithelial cells from 1 week old (1 wk, n = 4), 2 weeks old (2 wk, n = 3), 3 weeks old (3 wk, n = 2) and adult (n = 1) C57BL/6J mice stained for CD3, CD8α, CD103, TCRγδ and TCRβ as described. Intraepithelial cells were pregated on CD103+, CD3+ to depict IEL and then further gated on TCRγδ and TCRβ as shown. B) Percent (mean ±SE) γδ IEL (defined as percent of total IEL that were TCRγδ+, TCRβ− IEL) in the distal small intestines of 1 wk, 2 wk, 3 wk, and adult mice. Data are representative of 3 independent experiments (*p<0.05 when compared to 1 wk samples).
Figure 5.
Intestinal injury in wild-type mice is not associated with a selective reduction in γδ IEL.
A) Flow cytometry of small intestinal intraepithelial cells isolated from 2 weeks old dam fed wild-type (WT control, n = 2) or wild-type mice subjected to experimental gut injury as described (WT+PAF/LPS, n = 2). C57BL/6J mice stained for CD103, CD3, CD8α, TCRγδ and TCRβ as described above. Intraepithelial cells were pregated on CD103+, CD3+ to depict IEL and then further gated on TCRγδ and TCRβ as shown. B) Percent (mean ±SE) γδ IEL (defined as percent of total IEL that were TCRγδ+, TCRβ- IEL). Data are representative of 3 independent experiments (NS indicates no statistical difference between groups).
Figure 6.
Retinoic acid-related orphan nuclear hormone receptor C (RORC) gene expression in intestinal lymphocyte subsets.
(A) We measured gene expression levels of RORC by quantitative RT-PCR in 15 NEC tissue sections and 7 controls by quantitative RT-PCR array. RORC gene expression was significantly decreased in NEC samples versus controls (p<0.001). Relative level of mRNA expression of RORC in each sample was normalized to the expression level of reference gene GAPDH. (B) To determine whether loss of IEL contributed to reduction of RORC expression, we compared IEL with lamina propria lymphocytes (LPL) from the same tissue source using 10 non-NEC controls. RORC gene expression was significantly higher in IEL compared to LPL (p = 0.01). (C) RORC gene expression and proportions of total TCRγδ IEL correlated positively with each other (p = 0.02).
Figure 7.
γδ T cells reduce experimental gut injury.
A) Representative H&E staining of distal small intestines isolated from dam fed wild-type (1) or TCRδ−/− (3) mice with normal histologic appearance; or wild-type (2) or TCRδ−/− (4) mice subjected to experimental gut injury (PAF/LPS) as described (scale marker = 100 µm). Note shortened villi and epithelial sloughing with inflammatory infiltrate in wild-type PAF/LPS mice (2) and submucosal edema with severe villous sloughing in TCRδ−/− PAF/LPS mice (4). B) Histologic severity score (mean ±SE) of distal small intestinal sections obtained from dam fed wild-type (WT control) or TCRδ−/− (tcrδ−/− control) mice; or wild-type (WT PAF/LPS) or TCRδ−/− (tcrδ−/− PAF/LPS) mice subjected to experimental gut injury as described. Data are representative of 4 independent experiments with at least 3 mice per condition per experiment (*p<0.05).
Figure 8.
More severe intestinal injury in TCRδ−/− mice was associated with increased TNFα and decreased IL17A gene expression.
Gene expression of TNFα and IL17A as measured by quantitative RT-PCR in distal small intestinal sections obtained from dam fed wild type (WT control) or TCRδ−/− (tcrδ−/− control) mice; or wild type (WT PAF/LPS) or TCRδ−/− (tcrδ−/− PAF/LPS) mice subjected to experimental gut injury injury as described. Data are representative of 3 independent experiments with at least 3 mice per condition per experiment (*p<0.05).
Figure 9.
Occludin gene expression in human NEC versus control mucosa.
(A) We measured gene expression levels of occluding by quantitative RT-PCR in 16 NEC tissue sections and 13 controls by quantitative RT-PCR array. Occludin gene expression was significantly decreased in NEC samples versus controls (p<0.0001). Relative level of mRNA expression of occludin in each sample was normalized to the expression level of reference gene GAPDH.