Colonic epithelial response to injury requires Myd88 signaling in myeloid cells

Abstract

Proper colonic injury response requires myeloid-derived cells and Toll-like receptor/Myd88 signaling. However, the precise role of Myd88 signaling specifically in myeloid-derived cells that occurs during tissue damage is unclear. Therefore, we created a mouse line with Myd88 expression restricted to myeloid lineages (Myd88(-/-); LysM(Cre/+); ROSA26(Myd88/+); herein Mlcr). In these mice, Myd88 was appropriately expressed and mediated responses to bacterial ligand exposure in targeted cells. Importantly, the severe colonic epithelial phenotype observed in dextran sodium sulfate-injured Myd88(-/-) mice was rescued by the genetic modification of Mlcr mice. During injury, myeloid cell activation and enrichment of Ptsg2-expressing stromal cells occurred within the mesenchyme that surrounded the crypt bases of Mlcr and Myd88(+/-) mice but not Myd88(-/-) mice. Interestingly, these cellular changes to the crypt base mesenchyme also occurred, but to a lesser extent in uninjured Mlcr mice. These results show that Myd88 expression in myeloid cells was sufficient to rescue intestinal injury responses, and surprisingly, these cells appear to require an additional Myd88-dependent signal from a non-myeloid cell type during homeostasis.

Figure 1. Generation of Mlcr mice

(a) Targeting of Myd88 cDNA into the ROSA26 locus. The targeting construct included two homology arms indicated as SA (the ‘short arm’ at the 5′ position of the targeting construct) and LA (the ‘long arm’ at the 3′ position), the Myd88 cDNA, as well as two flox sites (denoted by filled triangles) that flank the neomycin resistance gene plus its three tandem polyadenlyation sites (3pA). The map indicates proper targeting of the construct into the ROSA26 locus by homologous recombination as well as the predicted effects of Cre mediated excision on a properly targeted locus. EcoRV digestion of genomic DNA was used to map clones and anticipated sizes of DNA fragments were indicated. (b, c) Southern blots showed clones that are either not targeted (WT) or properly targeted (KI) using (b) 5′ and (c) 3′ probes. The targeted ROSA26 locus contained a 4.0 kb EcoRV DNA fragment recognized by Probe1 and a 9.2 kb EcoRV DNA fragment recognized by Probe2.

Figure 2. Mlcr mice expressed Myd88 mRNA in colonic isolated myeloid cells

Graphs of the percent fold-difference of Myd88 mRNA expression of isolated colonic cell populations. (a) Isolated colonic myeloid cells. (b) Isolated colonic fibroblasts. For each cell type, Y-axis values were calculated as percentage of a fold difference of Myd88 mRNA expression by qRT-PCR analysis. The fold difference was calculated as the difference between the average ΔC_T of the experimental cell type and average ΔC_T of the control cells (untreated Myd88^+/−) divided by the control ΔC_T. Mean values ± SEM were plotted for each group (N=4-6 samples/group performed in 2 separate experiments). Triple asterisks indicate a statistically significant difference between two groups (P<0.001) as determined by a one way ANOVA with a Tukey’s post-test.

Figure 3. Bone marrow derived macrophages and dendritic cells from Mlcr mice showed elevated Ptgs2 and TNFα production in response to LPS treatment

(a) Graphs of the percent fold-difference of Myd88 mRNA expression in bone marrow (BM) derived macrophages and dendritic cells. For each cell type, Y-axis values were calculated as percentage of a fold difference of Myd88 mRNA expression by qRT-PCR analysis. The fold difference was calculated as the difference between the average ΔC_T of the experimental cell type and average ΔC_T of the control cells (untreated Myd88^+/−) divided by the control ΔC_T. (b-e) Graphs of the relative mRNA expression levels of (b, d) TNF-α and (c, e) Ptsg2 in bone marrow derived macrophages and dendritic cells. Cells were either (b, c) untreated or (d, e) treated with 10 ng of LPS for 1 hr. For each cell type, Y-axis values were a fold difference between the average ΔC_T of the experimental cell type and average ΔC_T of the control cells (untreated Myd88^+/−) divided by the control ΔC_T. Mean values ± SEM were plotted for each group (N=9-12 samples/group performed in 2-3 separate experiments). Asterisks indicate a statistically significant difference between two indicated groups (*P<0.05, **P<0.01, ***P<0.001) as determined by a one-way ANOVA and Tukey’s post-test.

Figure 4. Mlcr mice maintained crypt morphology during DSS treatment

H&E stained sections of descending colons from (a, d) Myd88 ^+/−, (b, e) Myd88 ^−/− , and (c, f) Mlcr mice. (a-c) Water treated controls and (d-f) mice treated with 2.5% DSS in drinking water for 7 days. Crypts from DSS-treated Myd88^−/− mice were atrophic (arrows). Descending colonic crypts from DSS-treated Myd88^+/− and Mlcr mice were similar in morphology to corresponding untreated mice. Bar= 100μm.

Figure 5. Mlcr mice maintained epithelial proliferation during DSS treatment

Images of descending colonic crypts from (a, d) Myd88^+/−, (b, e) Myd88^−/−, and (c, f) Mlcr mice. Images were from mice that were either (a-c) water treated or (d-f) DSS-treated. Each figure includes sections that were stained with either (left panel) H&E or (right panel) goat-anti-BrdU antisera, Alex-Fluor 594-labeled donkey anti-goat IgG (red) and bis-benzamide (blue). The crypt epithelial-mesenchyme junctions (white dotted lines) and the epithelial crypt-surface junctions (white dashed lines) are indicated. Bar=15 μm. Quantification of epithelial proliferation as determined by (g) S-phase (BrdU incorporation) and (h) M-phase analysis. Quantification of (i) crypt cell census and (j) epithelial apoptosis (apoptotic body counts on H+E). Mean values ± SEM were plotted for each group (N=8 mice/group analyzed in a total of three separate experiments; 100 well oriented crypt units analyzed/mouse). Asterisks indicate a statistically significant difference between DSS-treated and corresponding untreated controls (*P<0.05, **P<0.01, ***P<0.001) by one way ANOVA and Tukey’s post-test.

Figure 6. Mlcr macrophages express elevated CD86 during DSS treatment

(a, b) Image of the basal half of a descending colonic crypt and associated mesenchyme from a Myd88^+/− mouse. Sections were stained with rat-anti-mouse CD86 antisera directly conjugated PerCP (red), rat anti-mouse F4/80 antisera directly conjugated to Alex-Fluor 488 (green) and bis-benzamide (blue). The epithelial-mesenchyme junction is delineated by a white dotted line. Inset in (b) shows CD86 positive aggregates that either co-localized with F4/80 (yellow, arrows) or did not co-localize with F4/80 (red, double head arrow). Bar=15 μm. (c, d) Graph of (c) the number of CD86 aggregates that co-localize with F4/80 per crypt associated mesenchymal area and (d) the percentage of CD86 aggregates that co-localized with F4/80. Mean values ± SEM were plotted for each group (N=4 mice/group analyzed from 2 separate experiments; N=100 crypt associated mesenchyme units were evaluated per mouse). Asterisks indicate statistically significant differences between indicated groups (*P<0.05, **P<0.01, ***P<0.001) as determined by ANOVA and Tukey’s post-test.

Figure 7. Colonic fibroblasts express Ptgs2 in Mlcr mice

(a-f) Images of descending colonic crypts and surrounding mesenchyme from (a, d) Myd88^+/−, (b, e) Myd88^−/− and (c, f) Mlcr mice. Images from mice that were either (a-c) water treated or (d-f) DSS-treated mice. Sections were stained with mouse-anti-Ptgs2 antisera labeled with Alex-Fluor 594 (red), rat anti-mouse F4/80 antisera directly conjugated to Alexa-Fluor 488 (green), rat anti-mouse CD11c antisera directly conjugated to Alexa-Fluor 647 (pink) and bis-benzamide (blue). The crypt epithelialmesenchyme junctions (white dotted lines) are indicated. The arrows indicate mesenchymal cells that were Ptgs2 positive. Bars=15 μm. (g) Graph of the number of Ptgs2-positive cells (fibroblasts) per 100 crypt-mesenchyme units for the indicated genotypes and treatments (N=6 mice/group analyzed from two separate experiments; N=100 crypt-mesenchyme units evaluated/mouse). (h) Graph of the relative mRNA expression levels of Ptsg2 in isolated colonic fibroblasts and myeloid cells. Y-axis values were a fold-difference between the average ΔC_T of the experimental cell type and average ΔC_T of the control cells (Myd88^+/− myeloid cells) divided by the control ΔC_T. Mean values ± SEM were plotted for each group (N=4 mice/group analyzed from 2 separate experiments). Asterisks indicate a statistically significant difference between indicated groups (**P<0.01, ***P<0.001) as determined by ANOVA and Tukey’s post-test.

Figure 8. Ptgs2-expressing stromal cells are preferentially distributed in the crypt base associated mesenchyme of Mlcr mice

(a) Map of three mesenchymal zones associated with distinct populations of epithelial cells (Panel A; Upper=red, surface, middle=blue, post-mitotic crypt, lower=green, proliferative crypt base). Bar=5 μm. (b) Graph of the fractional representation of Ptgs2-espressing stromal cells located in each mesenchymal zone for each genotype calculated with and without DSS treatment (N=6 mice/group analyzed from two separate experiments; N=100 crypt-mesenchyme units analyzed/mouse). Asterisks indicate a statistically significant difference in fractional representation in the lower mesenchyme when comparing treated to untreated for a given genotype (*P<0.05, Student’s t test).

Figure 9. Untreated Mlcr mice contain mildly activated myeloid cells

Graph of the relative mRNA expression levels of multiple markers of myeloid cell activation. For each gene, Y-axis values were a fold-difference between the average ΔC_T of the experimental cell type and average ΔC_T of the control cells (Myd88^+/− myeloid cells) divided by the control ΔC_T. Mean values ± SEM were plotted for each group (N=4 mice/group analyzed from 2 separate experiments). Asterisks indicate a statistically significant difference between indicated groups (**P<0.01, ***P<0.001) as determined by ANOVA and Tukey’s post-test.