(A) Phenotype of Sca-1 KO mice. FACS analysis indicates the absence of Sca-1. (B) Ductulogenesis is diminished in Sca-1 KO mice. Whole mounts of the mammary gland of seven-week old WT (a) and KO (b) indicates a lesser degree of duct elongation in KO mice (0.6 cm) vs. WT mice (1.0 cm). Treatment of six-week old WT (c, e) and KO (d, f) mice with 15 mg medroxyprogesterone (MP) resulted in less proliferation in KO mice after 7 days. The line indicates the extent of ductal invasion. Inset indicates 5× magnification. The difference in ductal invasion between WT and KO mice differed significantly (P<0.01, N=5) by the two-sided Student’s t test.
(C) The Sca-1 KO mammary gland exhibits a lower proliferative potential. KO mice exhibit increased expression of the antiproliferative marker, p21^Cip1, and less expression of the proliferative marker, pRB, than WT mice. Magnification 400×. (D) The Sca-1 KO mammary gland exhibits an impaired proliferative response to MP. Animals were treated as in (B) and proliferation was monitored after 7 days by Ki-67 labeling. Sca-1 KO mice showed less Ki-67 labeling after MP treatment. Magnification 400×. Bar graph indicates the number of Ki-67-positive cells per duct based on a total of 40–50 ducts. Although, there was no signficiant difference in Ki-67 expression between control WT and KO mice (P>0.05), their response to MP differed significantly (P<0.001, N=16) by the two-sided Student’s t test.

Sca-1 deficiency enhances the chemopreventive effect of a PPARγ agonist, and attenuates the tumor promoting action of a PPARδ agonist. (A) Survival curve following treatment with MP/DMBA and maintenance on a diet supplemented with 0.005% GW7845. Only GW7845-treated KO mice exhibited a significant reduction in tumorigenesis (P=0.001 by the log rank test), which developed more slowly in KO mice (P=0.033). The number (N) of mice per group was: WT (N=10), WT+GW7845 (N=8), KO (N=19), KO+GW7845 (N=9). (B) Cumulative tumor incidence in GW7845-treated mice. Animals were treated as in (A). KO mice treated with GW7845 exhibited a significant reduction in tumorigenesis vs. control KO mice (P=0.001 by the log rank test). (C) Survival curve following treatment with MP/DMBA and maintenance on a diet supplemented with GW501516. Carcinogenesis was induced as in A, except that animals were administered a diet supplemented with 0.005% GW501516. Only WT mice treated with GW501516 exhibited a significant increase in tumorigenesis (P=0.03 by the log rank test), which developed more slowly in KO mice vs. WT mice (P=0.009). The number (N) of mice per group was: WT (N=10), WT+GW501516 (N=5), KO (N=10), KO+GW501516 (N=6). (D) Cumulative tumor incidence in GW501516-treated mice. Animals were treated as in (C). Only KO mice treated with GW501516 exhibited a significant increase in tumorigenesis (P=0.03) by the log rank test).

Reduction of Sca-1 increases PPARγ expression. (A) 34T mammary tumor cells were transduced with either a GFP shRNA (34T) or a Sca-1 shRNA (D8) and probed by western blotting for expression of Sca-1, pSer84PPARγ, PPARγ, pERK1/2, ERK1/2, PTEN and β-actin. Reduction of Sca-1 increased PPARγ and reduced the percentage of pSer84PPARγ and pERK. (B) Bar graphs of the western blots shown in (A). (C) GW7845 increases PPARγ expression. 34T and D8 cells in (A) were treated overnight with 0.1 µM GW7845. Bar graph indicates that GW7845 enhanced PPARγ levels to a greater extent in D8 cells than in 34T cells. (D) Effect of proteasome inhibition on pSer84PPARγ and PPARγ expression. 34T and D8 cells were treated overnight with 1 µM lactacystin. Bar graph indicates that PPARγ expression is increased to a greater extent in 34T cells than in D8 cells. (E) Reduction of Sca-1 increases PTEN and FABP3 mRNA expression. 34T and D8 cells were analyzed for PTEN and FABP3 mRNA levels by qRT-PCR and all values were normalized to actin. There was a significant difference between 34T and D8 cells in PTEN mRNA levels (P<0.01, N=3) and FABP3 mRNA levels (P<0.001, N=3) by the two-sided Student’s t test. (F) Schematic of the regulation of PPARγ by Sca-1. Sca-1 is envisioned to suppress PTEN in part through ERK-mediated inhibition of PPARγ. Phosphatidylinositol 3-kinase (PI3K)-dependent Ras/ERK signaling results in increased inhibition of PPARγ by ERK-mediated phosphorylation on Ser84, which in turn reduces PTEN transcription. The opposite effect occurs during Sca-1 deficiency, which results in reduced progenitor cell expansion, delayed tumorigenesis and increased sensitivity to PPARγ agonists.

Sca-1 deficiency increases PPARγ, and reduces pSer84PPARγ, PPARδ and pERK. (A) Mammary gland lysates from two wild-type (WT) and two KO (KO) mice were probed by western blotting for pSer84PPARγ, PPARγ, pERK1/2, ERK1/2, PPARδ, PTEN and β-actin. (B) Bar graph of PPARγ (PPARg) levels shown in (A). (C) Bar graph of the ratio of pSer84PPARγ/PPARγ (pPPARg/PPARg) shown in (A). (D) Bar graph of PPARδ (PPARd) levels shown in (A). (E) Bar graph of the ratio of pERK1/2/ERK1/2 shown in (A). All values were normalized to actin.

Sca-1 deficient adenocarcinomas express increased PPARγ, and reduced pSer84PPARγ, PPARδ and pERK. (A) Lysates from adenocarcinomas induced in two wild-type (WT) and two KO (KO) mice by MP/DMBA were probed by western blotting for pSer84PPARγ, PPARγ, pERK1/2, ERK1/2, PPARδ, PTEN and β-actin. (B) Bar graph of PPARγ levels shown in (A). (C) Bar graph of the ratio of pSer84PPARγ/PPARγ shown in (A). (D) Bar graph of PPARδ levels shown in (A). (E) Bar graph of the ratio of pERK1/2/ERK1/2 tumors shown in (A). All values were normalized to actin.