Retinoid signalling is increased in Cyp26b1^−/− limbs and to a lesser extent in Prrx1Cre⁺/Cyp26b1^fl/fl limbs. (A) X-gal staining of E12.5 and E14.5 limbs from wild-type, Prrx1Cre⁺/Cyp26b1^fl/fl and Cyp26b1^−/− mice heterozygous for the RARE-lacZ transgene. A dorsal view of the left limb is shown. (B) Analysis of Cyp26b1 and Rarb expression in the proximal and distal regions of the fore and hind limbs by qPCR. The regions collected for analysis are shown in a schematic representation (inset) of an E12.5 limb bud showing the cuts made to generate forelimb proximal (FP) and distal (FD) and hindlimb proximal (HP) and distal (HD) regions. Analysis was performed on single embryos and repeated with similar results. (C) qPCR analysis of Cyp26a1, Aldh1a2 and Rarb expression in PLM cultures after 1 or 3 days of culture. (D) RARE-LUC reporter activity in PLM cultures after 1, 3 or 8 days of culture, in the presence or absence of DEAB (10 μM). Cells were transfected with the RARE-LUC reporter at the time of seeding (considered day 0) and DEAB was added ~1 hour later. Control was set as 100% for Cyp26b1⁺ on day 1. Scale bar: 2 mm. Error bars represent 1 s.d. Significance was evaluated relative to wild-type untreated controls on the same day and is represented as follows: *P<0.05; **P<0.01; ^#P<0.001. FL, forelimb; HL, hindlimb; P, proximal; D, distal. Nd, 40 cycles of qPCR and transcript not detected; Rel. Express., relative expression; RLU, relative light units.

Cyp26b1-null mice exhibit significantly reduced expression of the chondroblastic marker Acan, but limited change in Sox9 expression. qPCR analysis of Sox9 and Acan expression in forelimb proximal (FP), forelimb distal (FD), hindlimb proximal (HP) and hindlimb distal (HD) regions of E11.5 and E12.5 limb buds. The regions are the same as those denoted in Fig. 2B. Bars indicate mean values. Significance was evaluated relative to Cyp26b1⁺ limbs; ^#P<0.001.

Chondrogenesis is negatively impacted in Cyp26b1-null mesenchyme and this defect is rescued effectively by an RAR antagonist but neither by BMP4 addition nor Cyp26a1 overexpression. (A) A Col2a1-derived reporter gene (Col2-LUC) was used to follow SOX5, SOX6 and SOX9 activity in PLM cultures after 1, 3 or 8 days of culture with or without DEAB treatment (10 μM). Cells were transfected with the Col2-LUC reporter at the time of seeding (considered day 0) and DEAB was added 1–2 hours later. Control was set as 100% for Cyp26b1⁺ on day 1. (B) Col2-LUC and RARE-LUC were co-transfected with an expression plasmid for Cyp26a1, and luciferease activity was measured 48 hours after transfection. For comparison, atRA (100 nM) was added to the cultures 24 hours after transfection. (C) Cyp26b1⁺ and null cultures were established, and treated with an RAR antagonist 4310 (100 nM) 24 hours later. Cultures were stained with Alcian Blue after 4 days of culture. (D) PLM cells from E11.5 Cyp26b1⁺ and Cyp26b1^−/− embryos were transfected with either Col2-LUC or RARE-LUC and treated within 1–2 hours of seeding with BMP4 (20 ng/ml). Luciferase activity was measured 72 hours after transfection. (E) Similar experimental conditions were used as described in C, except that BMP4 (B4; 20 ng/ml) was added shortly after culture initiation. (F) Cultures were prepared as described in D, and qPCR was used to quantify the expression of Sox9, Col2a1 and Acan following 4 days of culture. Error bars represent 1 s.d. Significance was evaluated relative to wild-type untreated controls on the same day and is represented as follows: *P<0.05; **P<0.01; ^#P<0.001. Scale bars: 1 mm (C); 1.5 mm (E).

Cells from Cyp26b1^−/− limb mesenchyme express/retain markers indicative of pre-cartilaginous condensations. (A) Peanut agglutinin (PNA) staining of PLM cultures treated with or without DEAB at culture initiation (10 μM) for 1 or 5 days. Day 1 cultures were photographed with a higher exposure than day 5 cultures. (B) qPCR analysis of condensation markers Sox9, Vcan, and Tnc in PLM cultures after 1 or 3 days of culture. Scale bar: 200 μm. Error bars represent 1 s.d. Significance was evaluated relative to wild-type controls on the same day and is represented as follows: *P<0.05; **P<0.01; ^#P<0.001.

Prechondrogenic cells from Cyp26b1^−/− animals exhibit reduced differentiation. (A) Alcian Blue staining of 4- and 8-day-old PLM cultures with or without DEAB treatment (10 μM). DEAB was added ~1 hour following culture initiation. (B) qPCR analysis of differentiation markers Sox5, Sox6, Col2a1, Acan, Hapln1, Comp and Matn1 in PLM cultures after 1 or 3 days of culture. Scale bar: 1 mm. Error bars represent 1 s.d. Significance was evaluated relative to wild-type controls on the same day: *P<0.05; **P<0.01; ^#P<0.001. nd, 40 cycles of qPCR and transcript not detected.

Cyp26b1^−/− cells show little change in hypertrophy. (A) qPCR analysis of hypertrophic and osteoblast markers Pthrp, Ihh, Ptch1, Col10a1, Runx2, Alp1, Spp1 and Mmp13 in PLM cultures after 1 or 3 days of culture. (B) The distribution of MMP13 was evaluated using immunofluorescence in 8-day-old cultures. A control with no primary antibody shows that staining is specific for MMP13 expression. For both cultures, the image was taken from the centre of the culture where cartilage nodules are present. (C) Cultures were stained for alkaline phosphatase following 8 days of culture. Scale bars: 200 μm (B); 1 mm (C). Error bars represent 1 s.d. Significance was evaluated relative to wild-type controls on the same day: *P<0.05; **P<0.01; ^#P<0.001.

Manipulation of endogenous retinoid signalling with ketoconazole inhibits chondroblast differentiation and accelerates chondrocyte hypertrophy. PLM cultures established from the whole limb buds of wild-type mice and treated with ketoconazole (keto, 1 μM) on day 0 exhibit similar changes in gene expression, evaluated by qPCR, as cultures established from Cyp26b1^−/− mice. Error bars represent 1 s.d. Significance was evaluated relative to wild-type controls on the same day: *P<0.05; **P<0.01; ^#P<0.001.