Supplemental Figure 1. Histomorphometry in osteoblast-specific Notch gain of function transgenic mice. a, Transgenic construct expressing Notch1 ICD. b, Expression of the transgene by Q-RT-PCR, and western blot of calvarial extracts showing increased Notch1 ICD RNA and protein in transgenic mouse line Tg1. c, Skeletal preparations from 3 different transgenic lines (Tg1, Tg2, Tg3) expression Notch1 ICD. Arrows indicate the thickening of the ribs, clavicles and humerus. Scale bar, 500 mm. d, Calcein labeling of 4 week-old spinal trabeculae of WT and Tg mice showing dramatically increased calcein labeling due to increased but disorganized woven bone formation. Scale bar, 100 μm. e, TRAP stained hind limbs of 4 week-old transgenic and wild type mice show absolute increase in relative staining consistent with absolute increase in number of osteoclasts. Scale bar, 500 μm. However, histomorphometry of 4 week-old transgenic hind limbs (n=3) stained with TRAP shows decreased osteoclast number (Oc.N) and surface (Oc.S) per bone surface (BS) in secondary spongiosa of tibia consistent with relative decrease in osteoclast per area of bone in transgenic vs. wild type mice. * p<0.05 between WT and Tg. f, Notch1 ICD co-immunoprecipitates with Runx2. HeLa cells were transfected with plasmids expressing either Myc-His epitope tagged Notch1 or Flag-tagged Runx2. Upper panel: Immunoprecipitation (IP) performed using anti-Myc antibody followed by Western blot (WB) with anti-Flag antibody. Lower panel: IP with anti-Flag antibody followed by WB with anti-Myc antibody. g, Notch1 ICD represses DNA binding of Runx2 in EMSA. COL10A1 promoter element was bound to: : Lane1, no protein; lane 2, RUNX2 alone; lanes 3–5, RUNX2 with increasing amounts of in vitro transcribed-translated Notch1 ICD (lane 3: 1X, lane 4: 2X; lane 5: 3X); lane 6, in vitro transcribed-translated luciferase; lane 7, mutated probe which is unable to bind Runx2. Runx2-DNA protein complex is shown by arrow. Free probe is below. Increasing concentration of Notch1 ICD decreases formation of Runx2-DNA complex. Negative control luciferase protein had no effect on this complex.
Supplemental Figure 2. Normal osteoblast formation in Ps1/Ps2 deleted mice. a, Genomic PCR showing Ps1, Ps2, and Cre alleles from tail DNA of various genotypes. b, Semi- quantitative RT-PCR for Ps1 expression from Ps1f/f deletion using RNA from cultured P7 osteoblastic cells. cDNA product from floxed (undeleted) Ps1 allele is noted by the band (Ps1) at increasing PCR cycles. cDNA content was normalized with β-actin. C; Control, D; DKO. c, Western blot shows the expression of Notch1 ICD in Ps1/Ps2 DKO or single knockout Ps1f/f; Ps2−/− (SKO) mice.
Supplemental Figure 3. a, 6 month-old (n=7) lumbar vertebral sections stained with toluidine blue and doubly-labeled with calcein showed no significant changes in osteoblast surface (Ob.S), mineral apposition rate (MAR), and bone formation rate (BFR), respectively. b, 3 month-old (n=6) lumbar vertebral sections stained with toluidine blue and doubly-labeled with calcein showed no significant changes in mineral apposition rate (MAR), bone formation rate (BFR) and mineralized surface per bone surface (MS/BS), respectively. c, Histomorphometry of 6 month-old (n=6) mice lumbar vertebrae revealed decreased bone volume (BV/TV), osteoclast number per bone surface (N.Oc/BS), and osteoclast surface per bone surface (Oc.S/BS) in DKO mice. * p<0.05 between WT and DKO.
Supplemental Figure 4. a, Western analysis using anti-FLAG antibody on lysates of 293 cells infected with lentivirus expressing NICD (right lane) tagged with FLAG or with control GFP expressing virus (left lane). b, ELISA measurement of OPG levels in pooled medium of Ps1/Ps2 null (DKO) vs. wild type (WT) calvarial osteoblasts from co-culture experiments with wild type splenocytes after 2 and 4 days of culture.