Studies were undertaken to test the hypothesis that the JNK pathway is an important regulator of hyperoxia-induced pulmonary responses. These studies demonstrate that hyperoxia inhibits cell proliferation, stimulates cell death, and alters myofibroblast transdifferentiation in a dose-dependent manner. These effects are significantly impacted by inhibiting the JNK pathway. Furthermore, there is increased expression of TGF-β1 and CTGF, on exposure to hyperoxia, both of which are known to signal via the JNK pathway. Our studies also demonstrate that hyperoxia-induced mortality and alveolarization are improved when the JNK pathway is inhibited, in the presence of excess TGF-β1 and CTGF, in the developing lung. When viewed in combination, these studies demonstrate that hyperoxia-induced cell death and TGF-β1-mediated pulmonary responses are mediated via signaling, at least in part, through the JNK pathway.
In our evaluation of varying concentrations of hyperoxia on cell death, all 3 concentrations of hyperoxia significantly reduced cell survival, compared to RA. We, however, could not discern any significant differences between the 3 concentrations of O2 on cell death. Use of JNKi was able to restore cell survival to RA values, in all 3 O2 concentrations. MLE-12 cells exposed to 95% O2 had increased cell survival on inhibition of the JNK pathway [7, 23]. Using siRNA against JNK1 in A549 cells exposed to 95% O2 decreased interleukin-8 expression, a pro-inflammatory cytokine . Data on cell viability were not reported in that study . Our data supports the contention that JNK pathway inhibition has a significant protective response in lung epithelial cells exposed to 95% O2. In addition, we report the novel finding that such an effect is also noted at lower concentrations (40% and 60%) of O2 in our in vitro model.
While hyperoxia-exposed decrease in cell proliferation in AIF was not impacted upon by JNKi, in line with the findings of Hashimoto et al in human lung fibroblasts, we found that AIF to MYF transdifferentiation was blocked by inhibiting JNK activation .
Our studies show that the death-receptor pathway and the executioner caspase 3 are involved in the process of hyperoxia-induced epithelial cell death. These data are supported by an earlier report using MLE-12 cells exposed to 95% O2 . In addition, we also report the novel observation that JNKi also impacts FAS-L and caspase-3 protein in A549 lung epithelial cells exposed to 95% O2.
To begin to understand the in vivo relevance of our findings, we selected TGF-β1 as our cytokine of interest for a variety of reasons. Firstly, the JNK pathway has been implicated in TGF-β1 signaling in lung cells [19, 24, 26–29], specifically CTGF [19, 26, 28] and cell death [19, 27] as well as myofibroblast transformation . Secondly, hyperoxia has been shown to upregulate TGF-β1 in premature rat lungs . Hence, we first confirmed increased expression of TGF-β1 and CTGF in our in vitro model, before proceeding to test our hypothesis in the NB TGF-β1 TG mice.
In the hyperoxia-induced acute lung injury model, JNKi administration was significantly protective in terms of survival in NB TGF-β1 TG mice. Importantly, the survival of the NB WT mice treated with JNKi was 100% after 7 days of 100% O2 exposure. This suggests that non-TGF-β1-dependent, but hyperoxia-induced molecular mediators signaling via the JNK pathway, are also involved.
Since TGF-β1 has been implicated in BPD, we used the lung-specific overexpression model to evaluate the impact of JNKi on alveolarization. We selected PN7 as the starting point as the mouse lung is in the alveolar phase at this time, and hyperoxia-induction of TGF-β1 was noted then . We used a 3-day treatment duration, as it takes about 48 hours for the TGF-β1 induction by dox to be sustained in our TG model, as previously described . Expectedly, as previously described , activation of TGF-β1 resulted in impaired alveolarization. Inhibition of the JNK pathway was able to improve this to a significant extent, compared to appropriate controls (NB WT mice on regular or dox water, and TGF-β1 on regular water). The improvement in alveolar architecture, however, was only partially corrected to appropriate control levels (NB WT mice on regular or dox water, and TGF-β1 on regular water, all treated with JNKi), as noted in Figure 5. This could be reflective of the short duration of treatment with JNKi that was employed.
It is important to point out that JNKi did not significantly alter the TGF-β1 levels in the BAL fluid of the TG mice on dox water (Figure 7). Hence, the effects noted above with JNKi were due to effects downstream of TGF-β1 pathway activation.
To further assess the potential for clinical translation, we used the NB WT murine BPD model and found significant, but partial, improvement in lung morphometry with JNK pathway inhibition.
Interestingly, a recent publication has reported lung-targeted conditional overexpression of CTGF to have a phenotype of BPD . This data supports our findings of increased CTGF on TGF-β1 activation in the NB lung, which also has a phenotype of BPD . Importantly, we noted decreased expression of CTGF and cell death pathway regulators with JNK pathway inhibition association with improvement of the BPD phenotype in the NB TGF-β1 TG mice lungs.
In addition to inhibiting JNK, SP600125 also inhibits ERK 1/2. The evaluation of the role of ERK 1/2 in our modeling systems was beyond the scope of the present manuscript. There are few limitations of our study: First, while A549 cells mimic lung epithelial cells, it is a transformed cell line, and hence, may not mimic the effects of freshly-isolated lung epithelial cells or in vivo. Second, the use of different strains of mice, at variable gestational ages, and different doses of JNK inhibitors can lead to significantly different results. To illustrate, in contrast to our findings, the JNK inhibitor SP600125 at a dose of 10 μM, used in lung explants obtained from CD1 mice at embryonic day 12.5, induced endogenous CTGF expression, TGFβ1-induced CTGF expression, increased DNA fragmentation and cleaved caspase 3 . Adult JNK1 null mutant mice have been reported to have increased susceptibility to hyperoxia . Interestingly, in contrast, the same group of investigators reported that adult JNK1 null mutant mice were resistant to ventilation-induced lung injury . In addition, adult rats treated with a JNKi were protected from LPS-induced lung injury . Hence, it is obvious that depending on the specific experimental conditions, the response to JNK inhibition can be quite variable.
In addition, another important factor to consider is the significant developmental regulation in the response of the developing lung to hyperoxia, versus the adult lung, as shown by us [34, 35] and other investigators . This can be the potential explanation of the seemingly conflicting results of our studies versus those in JNK1 and JNK2 null mutant adult mice having increased lethality on exposure to hyperoxia . We did, however use developmentally-appropriate NB mice for our in vivo work, in an attempt to mimic human BPD. This brings into focus the fact that independent confirmation of findings under appropriate and clinically relevant conditions must be undertaken, instead of extrapolating from in vitro or adult lung experimental results [36, 37].