IPF is a chronic, progressive parenchymal lung disease for which no effective therapy has yet been developed. A better understanding of the molecular mechanisms underlying the pathogenesis and progression of the disease is required for the development of novel therapeutic regimens for IPF. Recent studies suggested a significant contribution of SPARC to the pathogenesis of pulmonary fibrosis. However, the roles of SPARC have not been fully elucidated. In the present study, we demonstrated that SPARC enhances H2O2 production in fibroblasts treated with TGF-β.
Consistent with our observations, deletion of the SPARC gene significantly reduces the levels of urinary and renal reactive oxygen species, inflammation, and tubulointerstitial fibrosis in angiotensin II-infused mice . It is well known that increased ROS levels can cause epithelial cell apoptosis in culture [25, 26]. Moreover, activated myofibroblasts, which produce significant amounts of extracellular ROS, are sufficient to induce apoptosis of adjacent epithelial cells . Alveolar epithelial injury is considered to be one of the main characteristics of the lung in IPF, and recurrent epithelial damage is thought to cause fibrotic changes, and eventually result in fatal respiratory dysfunction [2, 4]. Inhibition of ROS production by NOX4 gene deletion [27, 28] and administration of the radical scavenger NAC  were shown to have protective effects against alveolar epithelial injury in the bleomycin-induced lung fibrosis model. A recent clinical trial indicated that NAC monotherapy may have some beneficial effects in the early stages of IPF although it failed to significantly change forced vital capacity . These reports indicated that elevated ROS production is one of the causative factors of recurrent epithelial damage in fibrotic lungs. Therefore, SPARC may be involved in epithelial cell injury through enhanced H2O2 production from activated fibroblasts. This hypothesis is supported by our results indicating that knockdown of SPARC expression level by siRNA mitigated the decrease in viability of A549 epithelial cells in coculture with TGF-β-stimulated fibroblasts. This reduction in A549 cell viability was alleviated in the presence of NAC. In addition, interference with SPARC expression by siRNA reduced H2O2 release from fibroblasts treated with TGF-β. SPARC has been shown to play an important role in ECM accumulation [15, 31]. In addition to this role of SPARC in the pathogenesis of fibrosis, our findings indicated a possible contribution of SPARC to epithelial cell damage through regulation of ROS production.
We demonstrated the involvement of ILK in the mechanism underlying enhanced ROS production by SPARC, which was supported by a number of observations. First, knockdown of SPARC with siRNA diminished ILK activation in TGF-β-stimulated fibroblasts. Second, siRNA against ILK significantly reduced extracellular H2O2 generation in TGF-β-stimulated fibroblasts. Our findings were consistent with those of previous studies indicating that SPARC activates ILK in fibroblasts  and that activation of ILK by high pressure leads to ROS production in vessels through Rac-1-mediated NAD(P)H oxidase activation . In isolated cardiomyocytes, ILK is activated by stromal cell-derived factor 1 (SDF-1) and is necessary for SDF-1-triggered activation of Rac-1, NAD(P)H oxidase, and release of ROS . ILK interacts with the cytoplasmic domain of the integrin β1/β3 subunits, which is important for cell adhesion, differentiation, and survival . Blocking of SPARC-integrin β1 interaction by function-blocking anti-integrin β1 antibody impairs ILK activation , suggesting that SPARC-ILK signaling is mediated at least in part by integrin β1.
NADPH oxidase family of proteins is comprised of five members, including NADPH oxidase 1 to 5 . In the present study, knockdown of NOX4 using siRNA almost completely blocked TGF-β-induced H2O2 production in HFL-1 cells (see Figure S3 B in Additional file 3), suggesting NOX4 is a major NADPH oxidase involved in TGF-β-induced H2O2 production. It has been known that NOX4 is a constitutively active NADPH oxidase isoform and NOX4 activity is regulated, at least in part, at the transcriptional level . NOX4 expression is increased by TGF-β stimulation in fibroblasts [28, 37]. Consistent with these reports, our study showed that NOX4 was upregulated by TGF-β in HFL-1 cells. While knockdown of SPARC prominently blocked H2O2 production induced by TGF-β stimulation, upregulation of NOX4 expression was reduced only moderately by SPARC knockdown (see Figure S3 C in Additional file 3), implying that SPARC may promote H2O2 production through regulation of NOX4 activity rather than regulation of transcriptional level of NOX4. Although activity of NOX4 is known to be regulated at the transcriptional level, more recently several reports have shown that NOX4 activity can be regulated by the mechanisms other than transcriptional regulation. P22phox and polymerase DNA-directed delta-interacting protein 2 (poldip2) modulate NOX4 activity [24, 38]. Post-translational modifications of NOX4, such as glycosylation, sumoylation or phosphorylation, are reported to be required for NOX4 activation [24, 39, 40]. In order to understand the precise mechanisms underlying enhancement of H2O2 production by SPARC, further studies are needed.
Another important finding in the present study was that SPARC expression is upregulated by TGF-β but not other profibrotic factors, such as PDGF, CTGF, TNF-α, IL-13, PGF2α, endothelin-1, angiotensin II, and IGF, in HFL-1 cells. In the bleomycin-induced lung fibrosis model, blocking of TGF-β signaling by the ALK-5 inhibitor SB-525334 significantly decreased SPARC expression as well as the degree of fibrosis. These results suggest that SPARC may be selectively upregulated by TGF-β and promote fibrotic changes via ROS production and ECM deposition. In accordance with our results, several previous studies indicate that TGF-β increases SPARC expression at both mRNA and protein levels in gingival fibroblasts, dermal fibroblasts, and pulp cells [21, 41]. In contrast to our results, angiotensin II was reported to increase SPARC level in renal mesangial cells . Thus, SPARC expression may be regulated by different factors in a cell type-specific manner. Although previous studies demonstrated regulation of SPARC by TGF-β, the signaling pathway involved in this regulation has not been explored in detail. In the present study, we showed that p38 MAPK and PI3K signaling are important for SPARC induction by TGF-β rather than the SMAD3 pathway using pharmacological inhibitors and siRNA experiments.
TGF-β signals are transduced by transmembrane Type I and Type II serine/threonine kinase receptors, which phosphorylate transcriptional factors SMAD2 and SMAD3. TGF-β also uses non-SMAD signaling pathways, such as MEK, PI3K-AKT, p38 MAPK, and JNK . We examined whether TGF-β activates PI3K-AKT, and p38 MAPK in HFL-1 cells. We found that TGF-β treatment induced AKT phosphorylation within 20 minutes (data not shown). On the other hand, p38 MAPK was phosphorylated in the basal state. Both AKT and p38 MAPK phosphorylation were reduced in the presence of specific inhibitors of these pathways. Our observations indicated that the basal activity of p38 MAPK and TGF-β-induced PI3K-AKT activation are involved in SPARC induction. With regard to the importance of PI3K and p38 MAPK in the pathogenesis of fibrosis, it was shown that phosphorylated AKT is strongly expressed in areas of pulmonary fibrosis after intratracheal administration of bleomycin in mice, and that blockade of PI3K-AKT signaling attenuates pulmonary fibrosis induced by bleomycin treatment or TGF-β overexpression [43, 44]. It has also been reported that inhibition of p38 MAPK attenuates the progression of fibrosis in the bleomycin model . SPARC may serve as one of the downstream factors of PI3K and p38 MAPK signaling in the pathogenesis of fibrosis. Although PDGF is also known to be able to activate both PI3K and p38 MAPK signalling pathways [46, 47], SPARC upregulation was not induced by PDGF stimulation in our study. Therefore, activation of PI3K and p38 MAPK is required but is not enough for SPARC induction. Other signaling pathways could also be involved in upregulation of SPARC by TGF-β.