<|endoftext|>D). FACS analyses demonstrated that GFP^+^ progeny was present in both luminal and basal compartments in the control (MMTV-Cre; Sox9^wt/wt^; CAG-eGFP) mice (Figure [6](#Fig6){ref-type="fig"}E). In contrast, the mammary glands of mice with MMTV-driven Sox9 deletion showed a near complete absence of GFP^+^ (sox9-deleted) progeny in the luminal compartment (Figure [6](#Fig6){ref-type="fig"}F), consistent with staining results (Figure [6](#Fig6){ref-type="fig"}C-D). These results suggest that myoepithelial progeny of Sox9-deleted precursors are viable while their luminal progeny is not.Figure 6**Sox9 deleted luminal cells are lost over time. (A)**. LacZ staining of whole mount mammary glands from 5 week old Sox9 cKO (R26R; MMTV-Cre; Sox9^fl/fl^) mouse. **(B)**. GFP in whole mount 8 week old Sox9 cKO (CAG-EGFP; MMTV-Cre; Sox9^fl/fl^) was visualized with fluorescence microscope. **(C)**. Section of mammary glands from 5-week old Sox9 cKO (R26R; MMTV-Cre; Sox9^fl/fl^) mouse, stained with X-gal. **(D)**. Section of mammary glands from 8 week old Sox9 cKO (CAG-EGFP; MMTV-Cre; Sox9^fl/fl^) mouse, immuno-labeling with anti-GFP (green) and anti-α-SMA (red) antibodies. **(E)**. CD24, GFP FACS analysis of cells isolated from control (CAG-EGFP;MMTV-Cre; Sox9^wt/wt^) mouse mammary glands. **(F)**. CD24, GFP Flow cytometry analysis of cells isolated from Sox9 deletion (MMTV-Cre; Sox9^fl/fl^; CAG-EGFP) mouse mammary glands.
Discussion {#Sec7}
==========
Sox9 has been established as a key regulator of tissue stem and progenitor cells during development in several organ systems \[[@CR18]-[@CR22],[@CR28]\], more recently has been linked to cancer stem cell function \[[@CR24]\]. A potential role of Sox9 in regulating mammary stem/progenitor cell functions has emerged from recent studies using ectopic expression and analyses of breast cancer cell lines. However, a physiological role of Sox9 in mammary gland development and mammary stem/progenitor cells remains unknown. Here, we use a conditional gene deletion approach to provide in situ evidence for a physiological role of Sox9 as an important regulator of mammary gland development. We also carried out short and long term lineage-tracing analyses to implicate a role for Sox9 in the maintenance of mammary stem cells and luminal progenitors.
Our conclusion that Sox9 is an essential component of the mammary gland developmental program is based on a substantial delay in post-natal mammary gland ductal elongation and branching in mice with MMTV-Cre directed Sox9 deletion (Figure [1](#Fig1){ref-type="fig"}). The defects in development are particularly severe at early time points (3 weeks postnatal) but are followed by gradual recovery, with essential no defects beyond 8 weeks of age. We hypothesized that recovery from the developmental block imposed by Sox9 deletion is overcome by the expansion of mammary epithelial cells in which Sox9 deletion did not occur, due to the known mosaic expression of MMTV-Cre that we used (Figure [2](#Fig2){ref-type="fig"}). Unfortunately, direct Sox9 staining was not possible due to lack of specific anti-Sox9 antibodies. Several antibodies that we tested showed either no or non-specific reactivity in western blotting, IHC, IF (Additional file [1](#MOESM1){ref-type="media"}: Figure S3). Consequently, we used Cre expression as a surrogate for Sox9 deletion and found that Sox9 deleted cells are lost in cKO mice by 8 weeks (Figure [2](#Fig2){ref-type="fig"}). Accordingly, no subsequent defects in mammary gland development were observed during pregnancy and lactation. These results suggest that Sox9 deletion puts the precursors of mammary epithelium at a developmental disadvantage. Our efforts at conditional Sox9 deletion in the mammary gland using a K14-Cre, which is active during embryonic mammary placode stage \[[@CR34]\], were not feasible due to perinatal lethality of the successfully targeted mice. However, the absence of mammary gland in a rare short-term survivor confirmed conclusions reached based on the use of MMTV-Cre driven Sox9 knockout mice (data not shown). Our results are reminiscent of more severe phenotypes of gene deletion using K14-Cre vs. MMTV-Cre, reported for GATA-3 deletion \[[@CR30]\].