Transcriptome analysis identifies pathways associated with enhanced maternal performance in QSi5 mice

Abstract

Background: Highly fecund mouse strains provide an ideal model to understand the factors affecting maternal performance. The QSi5 inbred strain of mice was selected for high fecundity and low inter-litter interval, and is very successful at weaning large numbers of offspring when compared to other inbred strains.

Results: Post-natal pup weight gain was used to estimate mammary gland output and to compare the performance of QSi5 mice to CBA mice. Cumulative litter weights and individual pup weight gain was significantly higher throughout the first eight days of lactation in QSi5 mice compared to CBA mice. Morphometric analysis of mammary glands during pregnancy in QSi5 mice revealed a 150 percent greater ductal side branching compared to CBA mice (P < 0.001). Ontology and pathway classification of transcript profiles from the two strains identified an enrichment of genes involved in a number of pathways, including the MAPK, tight junction, insulin signalling and Wnt signalling. Eleven of these genes, including six genes from the MAPK signalling pathway, were identified as associated with postnatal growth. Further, positive mediators of Wnt signalling, including Wnt4, Csnk2a1 and Smad4, were over-represented in the QSi5 strain profile, while negative regulators, including Dkkl1, Ppp2r1a and Nlk, were under-represented. These findings are consistent with the role of Wnt and MAPK signalling pathway in ductal morphogenesis and lobuloalveolar development suggesting enhanced activity in QSi5 mice. A similar pattern of phenotype concordance was seen amongst 12 genes from the tight junction pathway, but a pattern did not emerge from the insulin signalling genes. Amongst a group of differentially expressed imprinted genes, two maternal imprinted genes that suppress growth induced via the IGF signalling pathway, Grb10 and Igf2r, were under-represented in QSi5 mice. Whereas Peg3 and Plagl1, both paternally imprinted genes that enhance neonatal growth, were over-represented in QSi5 mice.

Conclusion: We propose that the combined action of at least three major signalling pathways involved in mammary gland development and milk secretion, namely Wnt, MAPK and tight junction pathways, contribute to the superior maternal performance phenotype in QSi5 mice. Additionally, favourable expression patterns of the imprinted genes Peg3, Plagl1, Grb10 and Igf2r may also contribute.

Figure 1

Comparison of litter size and pup weight at birth in the two inbred strains. oth number of pups born alive and mean pup weight at birth at the first two parities in CBA (n = 20) and QSi5 (n = 20) were significantly higher in the QSi5 strain of mice (p < 0.0001). Note: all error bars are + SEM.

Figure 2

Assessment of maternal performance in two inbred mice strains. (a) Cumulative litter weights in two strains – Cumulative litter weights measured using 6 pups/dam during the first 8 days of lactation in CBA (n = 16) and QSi5 (n = 20) strains; Weights are reported as Mean ± SEM. Lactation performance was significantly greater in QSi5 dams compared to CBA. (b) Individual pup weight gain at 8 days – Mean pup weight gain at the first 2 parities in CBA (n = 16) and QSi5 (n = 20) strains of mice. Pup weight gain was significantly higher in QSi5 strain relative to CBA strain of mice in both parities.

Figure 3

Histological analysis of mammary glands from pregnant mice. (a) Whole mounts of the inguinal mammary gland from day 12 pregnant mice A. CBA strain and B. QSi5 strain. (b) Comparison of relative mammary gland weights of the two different strains (n = 5 each). (c) Quantitation of ductal branching Both relative mammary gland weights (p < 0.001) and density of ductal branching (p < 0.001) were significantly higher in the QSi5 strain compared to CBA strain of mice.

Figure 4

Significant GO categories classified by DAVID. Pie chart showing the significantly enriched molecular function categories (Level 5) based on gene ontological classification of the differentially expressed genes (p < 0.01) between the two strains of mice using DAVID.

Figure 5

Hierarchical clustering and heat map of differentially expressed Wnt signalling genes. Clustering was performed on all ten transcript profiled samples (QSi5 – 1^st 5 vertical columns and CBA – last 5 vertical columns) for ten genes up-regulated in QSi5 (bottom ten horizontal rows) and 15 down-regulated genes (top 15 horizontal rows) relative to the CBA strain of mice. Expression levels are colour coded with shades of red, green and black corresponding to an increase, decrease and no change in gene expression, respectively.

Figure 6

Comparison of fold change differences in microarrays and qRT-PCR for selected genes. Quantitative RT-PCR validation of the microarray results showing the observed fold change differences in both methods for three up regulated genes in QSi5 strain (Kif5b, Ptger2, Wnt4) and three down regulated genes (Dkkl1, Grb10, Tnc).