Glycogen Synthase Kinase-3β Inhibition with 9-ING-41 Attenuates the Progression of Pulmonary Fibrosis

Idiopathic pulmonary fibrosis (IPF) is a progressive interstitial lung disease with a median survival of 3 years after diagnosis. Although the etiology of IPF is unknown, it is characterized by extensive alveolar epithelial cell apoptosis and proliferation of myofibroblasts in the lungs. While the origins of these myofibroblast appear to be diverse, fibroblast differentiation contributes to expansion of myofibroblasts and to disease progression. We found that agents that contribute to neomatrix formation and remodeling in pulmonary fibrosis (PF); TGF-β, Factor Xa, thrombin, plasmin and uPA all induced fibroblast/myofibroblast differentiation. These same mediators enhanced GSK-3β activation via phosphorylation of tyrosine-216 (p-Y216). Inhibition of GSK-3β signaling with the novel inhibitor 9-ING-41 blocked the induction of myofibroblast markers; α-SMA and Col-1 and reduced morphological changes of myofibroblast differentiation. In in vivo studies, the progression of TGF-β and bleomycin mediated PF was significantly attenuated by 9-ING-41 administered at 7 and 14 days respectively after the establishment of injury. Specifically, 9-ING-41 treatment significantly improved lung function (compliance and lung volumes; p < 0.05) of TGF-β adenovirus treated mice compared to controls. Similar results were found in mice with bleomycin-induced PF. These studies clearly show that activation of the GSK-3β signaling pathway is critical for the induction of myofibroblast differentiation in lung fibroblasts ex vivo and pulmonary fibrosis in vivo. The results offer a strong premise supporting the continued investigation of the GSK-3β signaling pathway in the control of fibroblast-myofibroblast differentiation and fibrosing lung injury. These data provide a strong rationale for extension of clinical trials of 9-ING-41 to patients with IPF.

While some studies have reported that inhibition of GSK-3β promotes myofibroblast differentiation and increased expression of mesenchymal markers, others show that GSK-3β inhibition blocks fibroblast activation and reduces indices of lung injury 5,13,16,[23][24][25][26][27][28] . Further, our laboratory has shown that 9-ING-41 reduced myofibroblast differentiation of pleural mesothelial cells in our empyema model and improved lung function and pleural injury outcomes 23 . Here we provide evidence that GSK-3β regulates differentiation of lung myofibroblasts, that inhibition of the process can be achieved by 9-ING-41 and that administration of this inhibitor is well-tolerated and effectively blocks PF in two murine preclinical PF models.

Models of PF. All experiments involving animals were approved by the Institutional Animal Care and Use
Committee at the University of Texas Health Science Center at Tyler. All experiments relating to animals were performed in accordance with relevant guidelines and regulations.
Bleomycin-induced PF in mice. PF was induced by intratracheal administration of bleomycin sulfate (Teva), as previously reported 29 with some modifications. Briefly, C57BL/6 mice (10-12 weeks of age, ≈20 g, Jackson Laboratory, Bar Harbor ME) were first lightly anesthetized with ketamine/xylazine. Anesthetized mice were next intubated using a 20 G canula. Bleomycin (0.8U/kg in 40 µl 0.9% saline) was then administered dropwise into the canula until the entire volume had been administered. Mice were then maintained for up to 28d. For GSK-3β inhibition studies, 9-ING-41 (30 mg/kg) or vehicle control (DMSO) was administered daily by IP injection. Treatment began 14d after bleomycin administration and continued for the next 14d until the completion of the time course 28d after initiation of bleomycin-induced PF. At the conclusion of the time course, mice were assessed with pulmonary function testing and CT imaging analyses as previously reported 12,23,30,31 .
TGF-β-induced PF in mice. PF was initiated by intratracheal instillation of constitutively active TGF-β adenoviral vectors (Ad-TGF-β) bearing C223S/C225S mutations, as previously reported 32 with minor modifications. Briefly, 3 × 10 8 pfu of Ad-TGF-β or eGFP adenovirus control vector (Ad-eGFP) were administered intratracheally in a volume of 40 µl. Mice were then monitored daily until the completion of the 14d time-course. For GSK-3β inhibition studies, mice received daily 9-ING-41 or vehicle treatment as described in the previous section. 9-ING-41 was given IP 7d after administration of the adenoviral vectors. At the conclusion of the 14d time-course, pulmonary function testing and CT imaging studies were performed as previously described 12,23,30,31 . Lung histology, immunostaining, confocal, bright field microscopy and morphometry. Lung tissue sections from normal and IPF patients harvested at the time of transplantation were provided through a Material Transfer Agreement with the University of Michigan (Ann Arbor, Michigan). Five normal and IPF patient sections were analyzed. All tissue sections were first deparaffinized and subjected to antigen retrieval using a citrate buffer at 95 °C for 20 minutes. Tissue histology, collagen deposition and localization were initially assessed by Trichrome staining as previously described 33,34 . Immunofluorescence was used to visualize GSK-3β (Cell Signaling) in the lung tissues of bleomycin treated mice and lung fibroblasts cultured on glass coverslips 23 . Differential interference contrast (DIC) and fluorescence images were obtained using a Leica TSC SP8 confocal laser scanning microscopy system (Leica Microsystems, Inc., Heidelberg, Germany). A series of optical sections were collected at 1 μm intervals in Z-axis (17 μm). These multiple Z-series sections were then projected onto one plane at 25X optical zoom as previously described 30,34,35 . Nuclear GSK-3β expression was quantified using ImageJ as previously reported 29 . Fibroblast culture conditions and treatment. Normal primary human lung fibroblasts (NF) and primary human lung fibroblasts cultured from lungs harvested at the time of transplantation from patients with IPF were provided through a Material Transfer Agreement with the University of Michigan (Ann Arbor, Michigan). These cells were maintained in DMEM (Corning, Carlsbad CA) containing 10% fetal bovine serum (Life Technologies), 2% antibiotic-antimycotic (Life Technologies) and GlutaMAX (Life Technologies) as previously reported 23,29,30,[34][35][36][37] . All cells were used with passage number of less than 10. Four normal and four IPF lines were used to assess responses to selected stimuli and for group comparisons. (10 -0.5 µm) for 18 hours in serum-free conditions. The cells were then treated with PBS vehicle or TGF-β for 24 (RNA) or 48 h (protein). For reversal studies, serum-starved cells were first treated with TGF-β for 24 h. 9-ING-41 was then added to the cells in the presence of TGF-β. RNA was isolated after 24 h incubation, while proteins were isolated after 48 h. To confirm the role of GSK-3β in fibroblast to myofibroblast transition, the tool compound GSK-3 inhibitor TDZD-8 (40-5 µM) was also used in blockade analyses.
Apoptosis staining. Tissue sections from the bleomycin and TGF-β models of pulmonary fibrosis were stained for apoptosis using the ApopTag ® Fluorescein In Situ Apoptosis Detection Kit according the manufacturer's directions. This kit recognizes and labels nicks in the DNA due to apoptosis.
Statistics. All statistics were performed using the Mann Whitney U test or Student t-test using GraphPad Prism 8. A p-value of less than 0.05 was considered significant. Mediators implicated in pulmonary organization induce myofibroblast differentiation of normal and IPF fibroblasts. Serum starved human fibroblasts were treated with various mediators to induce myofibroblast differentiation (TGF-β, FXa, thrombin (THB), plasmin (PLN) and uPA; see Materials and Methods). Cell lysates and conditioned medias, collected after 48 h, were then resolved by SDS-PAGE and western blotted for α-SMA, total GSK-3β, tyrosine 216 phosphorylated GSK-3β (pTyr-GSK-3β) and collagen 1 (Col-1), in NF (A) and IPF cells (C). β-actin was the loading control. α-SMA and collagen 1 expression were then quantified by densitometric analyses. Plotted data are the mean ± SEM of n = 3 independent experiments. Collagen was most prominently induced by TGF-β and FXa. Images are representative of three independent experiments. NF (B) and IPF (D) cells were treated PBS, TGF-β, Xa, thrombin, plasmin and uPA for 24 h incubation. RNA was then collected, and qPCR analyses were then performed for α-SMA and collagen 1 expression. GAPDH was the loading control. Plotted data are the mean ± SEM of n = 3-4 independent experiments.

GSK-3β is activated in fibroblast derived myofibroblasts.
Because of the enhanced expression of GSK-3β in the lungs of mice with induced PF, we next determined the activity of GSK-3β in fibroblast-myofibroblast differentiation. Normal and IPF fibroblasts were treated with TGF-β, Factor Xa, thrombin, plasmin and uPA, mediators previously shown to induce myofibroblast transition in other cell types 34 . As anticipated, TGF-β robustly induced α-SMA expression in both normal ( Fig. 2A) and IPF cells (Fig. 2C). Xa and thrombin, likewise, significantly induced α-SMA expression in both cell types. Conversely, only TGF-β and FXa significantly increased collagen 1 expression. GSK-3β expression was also enhanced in TGF-β, Xa, thrombin and plasmin treated cells. Phosphorylation of the GSK-3β activating tyrosine 216 motif was comparably enhanced by TGF-β in both NF and IPF cells. While uPA induced collagen expression in normal and IPF fibroblasts, induction of α-SMA was minimal. qPCR analyses showed significant increases in α-SMA by treatment with TGF-β, Xa and thrombin (Fig. 3A,D). TGF-β alone significantly increased Col-1 mRNA (p < 0.05).
In our recent report, we found that activated GSK-3β localizes to the nucleus in myofibroblasts 23 . To determine if GSK-3β is similarly activated in NF and IPF cells during fibroblast to myofibroblast transition, we visualized GSK-3β in normal and IPF fibroblasts in the presence and absence of TGF-β. Similar to our previous work, TGF-β treatment significantly increased GSK-3β localization to the nucleus in normal fibroblasts (Fig. 3, p < 0.01). IPF fibroblasts demonstrated significantly increased nuclear GSK-3β localization under control conditions, when compared to similarly treated NF cells. TGF-β treatment did not potentiate this affect in IPF cells. These findings suggest that GSK-3β activity is increased in IPF fibroblasts.
We next induced PF using the intratracheal bleomycin model. In these experiments, mice were injured 14 days prior to initiating treatment with DMSO vehicle or GSK-3β inhibitor; 9-ING-41. Bleomycin injured mice were then treated with 9-ING-41 or DMSO vehicle control for 14 days (Fig. 7). At 28 days post injury, both untreated and vehicle-treated mice demonstrated significant reductions in lung compliance compared to uninjured mice (Fig. 7A). However, 9-ING-41 treated mice showed significantly improved lung compliance compared to saline and vehicle-treated controls (p < 0.05).
Lung tissue sections from DMSO and 9-ING-41 treated mice were next trichrome-stained to show changes in lung architecture and collagen/neomatrix deposition. While DMSO treated mice showed pronounced injury and areas of intense collagen deposition, 9-ING-41 treated mice had significantly fewer areas of injury and demonstrated less collagen deposition (Fig. 7B). These findings were confirmed by immunological staining for collagen 1 and imaging by confocal microscopy that showed marked collagen deposition in DMSO treated mice (Fig. 7C). Conversely, 9-ING-41 treatment significantly reduced areas of intense collagen deposition (p = 0.04). Because apoptosis of epithelial cells is strongly implicated as a driver of PF in the bleomycin model and other forms of PF 29 , we next analyzed apoptosis in bleomycin mediated PF in the presence and absence of 9-ING-41 (Fig. 7D, p < 0.05). Bleomycin treatment significantly increased the number of TUNEL positive cells lining and within the alveolar walls. Conversely, 9-ING-41 treatment significantly reduced the number of apoptotic cells 40,41 in bleomycin-treated mice compared to vehicle treated controls (p < 0.05). These findings strongly support our hypothesis that the therapeutic targeting of GSK-3β with the novel inhibitor, 9-ING-41, reduces myofibroblast differentiation, collagen deposition, and subsequent PF in vivo.

Discussion
The aberrant accumulation of matrix-producing myofibroblasts in fibrotic foci and areas of remodeling characterizes interstitial lung diseases, including IPF. Fibroblast to myofibroblast differentiation is believed to significantly contribute to this process. Although GSK-3β had been shown to be important in the treatment of some cancers and the pathogenesis of fibrosis in other organs [13][14][15][16]23,24,42,43 , its role in myofibroblast differentiation and the progression of PF has remained unclear until the present time. In this study, we show that GSK-3β activation is critical for fibroblast to myofibroblast differentiation. Further, a novel clinically relevant GSK-3β inhibitor; 9-ING-41, now demonstrating promise in human neoplasia 18-20 effectively reversed lung injury in two distinct models of PF.
The role of GSK-3β in fibroblast to myofibroblast transition has been controversial. Studies by Baarsma and colleagues show that GSK-3 inhibition attenuates induction of α-SMA and fibronectin expression. Further, they show that this mechanism is dependent on GSK-3-mediated activation of CREB signaling 13 . These studies however were limited to analyses in a fibroblast cell line. Similar studies by Liu et al. showed that GSK-3 inhibition protected against PF by increasing cellular autophagy 26 . Conversely, studies by Xia et al. found that GSK-3β is inactivated in IPF-derived fibroblasts and that subsequent aberrant activation of β-catenin may contribute to cellular differentiation. Similar results were found in peritoneal mesothelial cells, in which activation of GSK-3β protected against MesoMT 42-44 while promoting cell death 45 . These disparate findings in the aggregate support our hypothesis that GSK-3β promotes PF and offer a strong rationale to investigate its role in this condition, as we did in this study.
Our data here clearly show that inhibition of GSK-3β improves physiologic outcomes in two models of PF and that the effects are associated with inhibition of fibroblast to myofibroblast transition and subsequent matrix deposition. To our knowledge, this represents the first evidence along these lines and suggests that this approach merits further consideration for further translational investigation. We first showed that GSK-3β expression was enhanced in the lungs of mice with bleomycin induced PF. Similar changes were not observed in tissue sections of IPF patients or in the TGF-β PF model. We posit the chronicity of the disease in IPF patients may have muted the changes in GSK-3β expression. The variable GSK-3β expression we observed in our subacute models likely reflects the different agents we used to induce PF, possible differences in the durability or intensity of the response, or differences in the contributions of levels of GSK-3β to the different forms of PF. We next showed that diverse mediators could induce myofibroblast differentiation of NF and IPF cells. TGF-β, however, was the most potent inducer of α-SMA and collagen in these cells.
Although GSK-3β is a constitutively active kinase, its activity can be increased by phosphorylation of Tyr-216 and localization to the nucleus and 46,47 . In our previous work, we found that a diverse group of mediators (TGF-β, FXa, thrombin, plasmin and uPA) potently induced tyr-216 GSK-3β phosphorylation in HPMCs. In both NF and IPF cells, we found that TGF-β enhanced GSK-3β Tyr-216 phosphorylation to a much greater extent than any of the other tested mediators. We also found that GSK-3β localized primarily to the nucleus of NF and IPF. Further, TGF-β enhancement of GSK-3β localization to the nucleus was limited to the IPF cells. Because we previously Lung tissue sections from GFP and TGF-β adenovirus infected mice were stained for apoptosis by fluorescent TUNEL stain (green) and nuclei (blue). Images were taken at 20X optical zoom. Images are representative of 15 fields/slide/condition and n = 5-6 samples/condition. Data are expressed as mean ± SEM. p denotes a p < 0.05. Bar indicates 50 µm. www.nature.com/scientificreports www.nature.com/scientificreports/ showed that GSK-3β was localized primarily to the cytoplasm of unstimulated PMCs, we now hypothesize that the role of GSK-3β in myofibroblast differentiation may vary by cell type. These findings will be studied in greater detail in subsequent studies. They further support our hypothesis that GSK-3β phosphorylation at Tyr-216 and localization to the nucleus are drivers of myofibroblast differentiation.

Scientific RepoRtS
To determine the role of GSK-3β in fibroblast to myofibroblast transition, we initially attempted to directly reduce its expression with siRNA-and GSK-3β shRNA-encoding adenoviral vectors. However, we were unable to sufficiently reduce GSK-3β expression for subsequent analyses. As such, our studies focused on the use of GSK-3β-targeting pharmacologic inhibitors to assess the contribution of GSK-3β in fibroblast to myofibroblast transition. For these studies, we used the highly efficacious GSK-3β inhibitor; 9-ING-41. We previously showed that 9-ING-41 strongly blocked and reversed MesoMT and reduced Tyr-216 phosphorylation 23 . Further, 9-ING-41 effectively blocked MesoMT at lower concentrations than TDZD-8. In the studies presented here, we found similar results, as 9-ING-41 treatment blocked TGF-β-mediated myofibroblast differentiation at doses 8 times less than TDZD-8. While the same concentration of 9-ING-41 (10 µM), was effective for pleural mesothelial cells, higher concentrations of TDZD-8 were required to effectively block lung fibroblast to myofibroblast differentiation (40 µM versus 20 µm). This difference is most likely a consequence of the increased efficacy and half-life of 9-ING-41 compared to other GSK-3β inhibitors 48 .
We next confirmed that 9-ING-41 could reverse established myofibroblast differentiation. TGF-β can induce detectable markers (α-SMA and collagen) of fibroblast to myofibroblast transition in as little as 24 h 29,34,39 . We found that induction of these biomarkers was reversed by treatment with 9-ING-41. Western blotting and qPCR analyses showed that the highest two concentrations of 9-ING-41 effectively reversed α-SMA and collagen 1 induction. Further, GSK-3β phosphorylation was also markedly reduced. These findings strongly support our hypothesis that GSK-3β inhibition by reducing Tyr-216 phosphorylation represents a therapeutic target for the treatment of PF, providing the rationale for in vivo analyses.
Because of our prior experience with 9-ING-41 in mice 23 , we chose to use a daily 30 mg/kg dose of 9-ING-41. In this study, 9-ING-41-treated mice tolerated this dose and showed no overt signs of toxicity. 9-ING-41 significantly improved lung function in both models of PF. In the bleomycin study, 9-ING-41 treatment significantly improved lung function and lung volumes compared to vehicle treated mice. 9-ING-41 treatment did not, however, restore lung function to control levels. While collagen deposition was reduced by 9-ING-41 treatment, residual areas of lung injury and hyperplasia were observed. These findings suggest that 9-ING-41 treatment prevented scarring resulting from fibroblast to myofibroblast transition but was not fully protective at the doses we used. Further, the fact that 9-ING-41 reduced overall lung and, presumably, epithelial apoptosis in both the TGF-β and bleomycin models strongly suggests that preservation of epithelial cell viability contributes to protection of the lung architecture. Additional studies with longer treatment courses may show continued improvement compared to vehicle control with values approaching normal levels.
In summary, this report provides evidence that GSK-3β plays a critical role in myofibroblast differentiation and subsequent PF. While numerous factors may contribute to disease progression, TGF-β likely plays a central role in this process. As TGF-β is locally expressed and implicated in the pathogenesis of PF 49,50 , we now demonstrate that it too can activate GSK-3β during induction of fibroblast to myofibroblast transition. We also found that other mediators of neomatrix remodeling can drive myofibroblast differentiation. As in other cell types 25,31,34,39,51 , these factors may likewise activate GSK-3β in vivo, thereby promoting PF. While inhibition of GSK-3β with 9-ING-41 can attenuate myofibroblast differentiation in vitro, our findings show that pharmacologic targeting of GSK-3β in vivo is effective and generally well tolerated. Our findings also provide a strong rationale for advancement of 9-ING-41 as a candidate new therapeutic for the treatment of patients with PF.

Data availability
No datasets were generated or analyzed during the current study.