Characterization of gene expression profiles for different types of mast cells pooled from mouse stomach subregions by an RNA amplification method

Abstract

Background: Mast cells (MCs) play pivotal roles in allergy and innate immunity and consist of heterogenous subclasses. However, the molecular basis determining the different characteristics of these multiple MC subclasses remains unclear.

Results: To approach this, we developed a method of RNA extraction/amplification for intact in vivo MCs pooled from frozen tissue sections, which enabled us to obtain the global gene expression pattern of pooled MCs belonging to the same subclass. MCs were isolated from the submucosa (sMCs) and mucosa (mMCs) of mouse stomach sections, respectively, 15 cells were pooled, and their RNA was extracted, amplified and subjected to microarray analysis. Known marker genes specific for mMCs and sMCs showed expected expression trends, indicating accuracy of the analysis. We identified 1,272 genes showing significantly different expression levels between sMCs and mMCs, and classified them into clusters on the basis of similarity of their expression profiles compared with bone marrow-derived MCs, which are the cultured MCs with so-called 'immature' properties. Among them, we found that several key genes such as Notch4 had sMC-biased expression and Ptgr1 had mMC-biased expression. Furthermore, there is a difference in the expression of several genes including extracellular matrix protein components, adhesion molecules, and cytoskeletal proteins between the two MC subclasses, which may reflect functional adaptation of each MC to the mucosal or submucosal environment in the stomach.

Conclusion: By using the method of RNA amplification from pooled intact MCs, we characterized the distinct gene expression profiles of sMCs and mMCs in the mouse stomach. Our findings offer insight into possible unidentified properties specific for each MC subclass.

Figure 1

Comparisons of three round-amplified products starting with very small quantities of RNA. (a-c) Amplification biases in the products starting from a small quantity of RNA. Scatter plots of signal intensity obtained from 5 μg of BMMC RNA prepared by the standard protocol and from 30 pg (a), 10 pg (b) and 2 pg (c) of BMMC RNA prepared by three rounds of amplification are shown. (d, e) Reproducibility of the three-round amplification of a small quantity of RNA. Scatter plots of signal intensity between two independent products from 30 pg of BMMC RNA (BMMC 30 pg-1 and BMMC 30 pg-2) (d) or from 2 pg of BMMC RNA (BMMC 2 pg-1 and BMMC 2 pg-2) (e), are shown. Red dots show probe sets judged as "Presence", and yellow dots represent probe sets judged as "Absence" in both arrays. Blue dots show probe sets judged as "Presence" only in either array. The correlation coefficients (r) are presented. The same, four-fold induction and suppression thresholds are indicated as diagonal lines. Genes judged as "Presence" are placed in groups corresponding to pairwise overlaps shown in the accompanying Venn diagrams.

Figure 2

Gene expression profiles of sMCs and mMCs from stomach tissue. (a) Isolation of toluidine blue-stained MCs in the submucosa (sMC; upper panels) and the mucosa (mMC; lower panels) of stomach sections. A sMC (arrow) and mMC (arrowhead) that was metachromatically stained with toluidine blue before microdissection (left panels) disappeared after microdissection with a patch pipette (right panels). Bars, 10 μm. (b) Outline of the experimental strategy. (c) Labeled and fragmented antisense RNAs of three individual sMC samples, three individual mMC samples and the 'no cell' samples were hybridized to a Murine Array. Scatter plots for 'no cell' (x axis) and sMC₁ (y axis) (upper left), 'no cell' (x axis) and mMC₁ (y axis) (lower left), sMC₁ (x axis) and sMC₂ (y axis) (upper center), mMC₁ (x axis) and mMC₂ (y axis) (lower center), sMC₁ (x axis) and mMC₁ (y axis) (upper right) are shown. The correlation coefficients (r) for comparison within sMC_1–3, within mMC_1–3 and between sMCs and mMCs are presented as means ± S.D. Red dots show probe sets judged as "Presence", and yellow dots represent probe sets judged as "Absence" in both arrays. Blue dots show probe sets judged as "Presence" only in either array. The same, two-fold induction and suppression thresholds are indicated as diagonal lines.

Figure 3

Global gene expression analysis of sMC_1–3 and mMC_1–3. (a) Hierarchical clustering of global gene expression of various preparations of MCs and non-MCs. Three-round amplified products of sMC_1–3, mMC_1–3, skin MCs and BMMCs, and the standard products of BMMCs, peritoneal MCs, macrophages and fibroblasts were analyzed. (b) The principal component analysis (PCA) reveals different gene expression profiles of sMC_1–3, mMC_1–3, and two preparations of BMMCs. The blue dotted square indicates mMCs, the red dotted square indicates sMCs, and the black dotted square indicates BMMCs.

Figure 4

Validation of the differentially expressed genes between sMCs and mMCs. (a) sMC-specific (Cma2, Mcpt4), mMC-specific (Mcpt1, Mcpt2) and MC-common markers (Fcer1a and Kit) (left panel) and six randomly selected genes (Ces3, Cnn1, Cpe, Anxa10, Ctse and Fos) (right panel) are indicated in the representative scatter correlation graphs between sMC₁ and mMC₁. The same, two-fold induction and suppression thresholds are indicated as a yellow, blue and red line, respectively. (b) The expression levels of the genes in (a) were verified by real-time RT-PCR. The values represent the ratio of relative expression levels of mMCs to sMCs, and are shown as mean ± S.D. (n = 3). The specificity of the PCR product was confirmed by gel electrophoresis and analysis of the melting temperature. The expression level of each gene was normalized to 28S ribosomal RNA.

Figure 5

Clustering analysis of the gene expression profiles between sMCs and mMCs. (a) Representation of mRNA expression levels of sMC_1–3 and mMC_1–3 compared with BMMCs. The color of the bars represents the ratio of signal intensity between independent samples and BMMCs, according to the scale shown on the top right. Genes with significantly different expression between sMCs and mMCs (p < 0.05, Limma's t test) were selected (1,272 genes) and classified into 7 clusters using the k-means algorithm (CL1-7). (b) Functional categorization of representative genes from (a).

Figure 6

Immunohistochemical analysis of Notch4 and Ptgr1 in sMCs and mMCs in stomach tissue. (a) Stomach submucosa (sMCs; left panels), stomach mucosa (mMCs; middle panels) and skin (skin MCs; right panels) sections were stained with an anti-Notch4 antibody (lower panels) and with toluidine blue (upper panels). sMCs stained with the anti-Notch4 antibody in the gastric submucosa and skin dermis are indicated by arrows. No staining was observed in mMCs (arrowheads) localized in the gastric mucosa. sMCs and mMCs were metachromatically stained with toluidine blue. (b) Stomach submucosa (sMCs; left panels) and stomach mucosa (mMCs; right panels) sections were stained with an anti-Ptgr1 antibody (lower panels) and with toluidine blue (upper panels). No staining with the anti-Ptgr1 antibody was found in the sMCs (arrow). Small signals were observed in the mMCs (arrowheads). sMCs and mMCs were metachromatically stained with toluidine blue. Bars, 25 μm (a, b).