PMC

Periostin effects on DD and PF cell proliferation in collagen culture. Proliferation assays were performed on DD and PF cells growing in collagen-enriched culture containing recombinant periostin (2 μg/ml) or vehicle with using the WST-1 assay. As shown, recombinant periostin induced a significant increase in PF cell proliferation compared to vehicle treated cells (*, p<0.05). DD cell proliferation was unaffected by addition of recombinant periostin. Normalized absorbance units (mean±standard error) are shown for six cell lines (3 patient-matched PF and DD) assessed three times in triplicate.

Periostin effects on α smooth muscle actin induction by DD and PF cells in FPCL culture. FPCLs containing DD or PF cells and either recombinant periostin (2 μg/ml, Pn) or vehicle (Veh) were lysed 48 h after release and the lysates were subjected to western immunoblotting for α smooth muscle actin. β actin levels were also assessed to confirm that equal amounts of cell-derived protein were loaded in each well. As shown, addition of 2 μg/ml periostin to the FPCL cultures of DD cells markedly induced α smooth muscle actin accumulation. Recombinant periostin treatment of FPCL cultures containing PF cells did not consistently alter α smooth muscle actin accumulation above cultures treated with vehicle. These immunoblots are representative of three PF and three DD cell lines assessed in triplicate.

Real Time PCR analysis of POSTN mRNA levels in normal palmar fascia and DD cord tissues. POSTN mRNA levels were assessed in phenotypically normal palmar fascia and DD cord samples. Matching samples of PF and DD cord are shown for patient #4 (#4*) as well as normal palmar fascia from patient #5 compared to DD cord samples from patients (#6, #7). Triplicate analyses of POSTN mRNA levels were normalized to GAPDH mRNA expression using the comparative Ct method. Relative expression units (means and 95% confidence interval) are shown. Despite the variability inherent to cord samples of varying cellularity, the mean POSTN mRNA levels in each of the three DD cord samples were significantly increased relative to the phenotypically normal palmar fascia samples tested (p<0.05).

Periostin effects on apoptosis of DD and PF cell in collagen culture. The levels of apoptosis were assessed in cells growing on collagen substrate treated with vehicle or periostin in MCDB media using the Cell Death Detection ELISA. PF cells (while bars) and DD cells (black bars) were treated with vehicle (Phosphate Buffered Saline/0.1% BSA), 0.5 or 2.0 μg/ml recombinant periostin. PF and DD cells display a significant difference in baseline apoptosis with vehicle treatment (δ, p<0.05). A significant increase in PF cell apoptosis is evident relative to vehicle treated cells for 0.5 μg/ml recombinant periostin, (*, p<0.05) and 2.0 μg/ml recombinant periostin (*, p<0.01). DD cells did not display a significant change in apoptosis relative to vehicle treatment with 0.5 μg/ml recombinant periostin (p=0.075) or 2.0 μg/ml recombinant periostin (p=0.065). Normalized absorbance units (mean±standard error) are shown for three sets of patient-matched cell lines assessed three times in triplicate.

Western immunoblotting and immunohistochemistry of periostin in normal palmar fascia and DD cord tissues. (A) Western immunoblots (a, b and c) for periostin in total protein lysates of surgically resected palmar fascia tissue (PF) and DD (DD) samples are shown. While each sample subjected to electrophoresis was adjusted to the same total protein concentration of the same, the cell:extra-cellular matrix ratio is different between densely cellular DD cord and phenotypically normal palmar fascia tissues. For this reason, intracellular β-actin levels were assessed to normalize the relative cellular contribution of each sample. As shown, periostin is readily detectable in total protein lysates of DD cord tissue but not patient-matched palmar fascia while intracellular β-actin could be detected in both. An example of variable periostin immunoreactivity is shown (c, PF, DD nodule (DDn), DD cord (DDc)). (B) Immunohistochemistry of paraffin-embedded DD cord tissue (a and c) or adjacent, phenotypically normal palmar fascia (b) with periostin polyclonal antibody (a and b) or rabbit IgG (c). Periostin is evident as brown staining resulting from di-amino benzidine precipitation and all sections were counterstained with methyl green. As shown, abundant periostin is evident in densely cellular DD cord samples while relatively low-level immunoreactivity is evident in patient-matched palmar fascia tissue.

In situ hybridization studies of POSTN mRNA expression in normal palmar fascia and DD cord tissues. In situ hybridization studies performed using S³⁵ labeled sense and antisense transcripts derived from a 391-bp fragment from the 5′ end of a POSTN cDNA. Paraffin-embedded DD cord tissues (A, B, E and F) and palmar fascia tissue adjacent to the cord (C and D) were exposed to antisense POSTN mRNA probe (A–D) or sense POSTN mRNA probe control (E and F). Bright field images of hematoxylin and eosin stained tissues are shown in A, C and E while the same fields under dark field are shown in B, D and F. Silver grains deposited due to exposure to the S³⁵ labeled transcripts are evident as uniform white specks under dark field. As shown, densely cellular DD cord tissue (B) exhibits intense antisense POSTN mRNA probe signal relative to low-level signal in adjacent palmar fascia (D) or densely cellular DD cord tissue exposed to sense POSTN mRNA probe (F). All images are 400×.

Correlation of in vivo and in vitro periostin levels. To determine appropriate levels of exogenous periostin treatment to approximate in vivo conditions, total protein extracts (12 μg) from four DD cords (DD) were compared to 0.25, 0.5, 1.0 and 2.0 μg of recombinant periostin (R&D systems, Minneapolis, MN) by western immunoblotting with periostin antibody (A). The relative sizes of recombinant periostin and endogenous periostin in DD cord are discussed in the text. The bands evident at 85–70 kDa represent low-level cross-reactivity with the serum albumin in the DD samples and BSA in the vehicle (Phosphate Buffered Saline/0.1% BSA). Based on the intensity of the most abundant molecular weight form of periostin in the DD cord samples, we estimated that the DD cord samples exhibited a range between 0.5 μg and 2 μg of periostin, or between 0.05 μg and 0.2 μg of periostin/μg of total protein. Based on this estimate and 1×10⁵ DD or PF cells routinely yielding approximately 10 μg of total protein in our hands, this data was extrapolated to conclude that a treatment range of 0.5 to 2.0 μg/ml of recombinant periostin to 1×10⁵ cells in culture would approximate in vivo levels.

Periostin effects on DD and PF cell stressed Fibroblast Populated Lattice Culture (FPCL) contraction. Stressed FPCLs containing DD or PF cells and either recombinant periostin (2 μg/ml) or vehicle were constructed as described in the Methods. Lattices were photographed at various time intervals after release (A) and the area of each lattice was determined using the freehand tool in ImageJ software. Sequential area calculations were normalized to the area immediately after release (0 h post release) for three cell lines assessed three times in triplicate (B). The horizontal axis denotes time (hours) after release and the vertical axis denotes the percentage reduction in lattice area after release. As shown, cell-free collagen lattices (No cells) do not contract significantly after release. Vehicle treated DD cells (DD-Veh) exhibit increased contractility relative to vehicle treated PF cells (PF-Veh), consistent with our previous data []. 2 μg/ml recombinant periostin induced a significant increase in DD cell contraction of the lattice (DD-Pn) compared to vehicle treated cells (*, p<0.05) at 12, 24 and 48 h after release, indicative of increased myofibroblast contractility. No statistical difference in collagen contraction was evident between periostin treated (PF-Pn) and vehicle treated PF cells (p=0.08, 48 h after release).