Here we present our results of microarray profiling of normal human lung fibroblast following siRNA mediated knockdown of NRF2 and KEAP1. We have identified a distinct gene set of anti-correlated genes in this analysis to better define NRF2 regulated genes in a lung specific cellular context.
A comparison of the 1,045 signature sequences differentially modulated by NRF2 and KEAP1 siRNA (Figure 2) with other gene expression signatures collected in the Gene Expression Omnibus (GEO. http://www.ncbi.nlm.nih.gov/geo/) data base indicates a highly significant anti-correlation with a gene signature obtained from primary human lung fibroblast treated with dithiothreitol (DTT) for 24 hours (Overlapping P = 1.7E-99) ; and a significant correlation with a gene set from dexamethasone-treated (24 hours) human primary osteoblast-like cells (HOb)(GEO series GSE10311.Overlapping P = 1.0E-46). In addition, we found two cigarette smoke-related gene signatures which are anti-correlated to our gene signature, one from a normal human bronchial epithelial (NHBE) cells exposed to a cigarette smoke condensate for 18 hours (Overlapping P = 2.5E-33, data derived from GEO series GSE18235 , and the other from a comparison of small airway epithelial cells between smokers and non-smokers (overlapping p = 2.3E-24) . Since DTT and cigarette smoke induce ER stress and oxidative stress, respectively; it appears that NRF2 is activated in both situations to confer cellular protection.
In addition to NRF2 promoting the anti-oxidant response machinery, this pathway also has profound anti-inflammatory effects . Studies with NRF2 deficient mice demonstrate an increased inflammatory response in several inflammatory disease models [58, 66, 67]. In respiratory models, the loss of Nrf2 results in increase eosinophil recruitment in the lungs of allergen challenged animals and the increase in lung macrophages upon hyperoxic lung injury . In models of COPD, Nrf2 deficient mice have increased neutrophil and macrophage recruitment to the lung . In vitro studies have demonstrated a specific effect of the NRF2 regulating cytokine and chemokine expression in neutrophils following LPS challenge . In addition, pharmacological activation of NRF2 with the triterpenoid CDDO can inhibit LPS induced inflammation in neutrophils and PBMCs .
In this study we make the novel discovery that Eotaxin-1 is uniquely inhibited by NRF2 activation. While the direct role of NRF2 on Eotaxin-1 regulation has not be reported previously, mice deficient for Nrf2 do have increased eosinphil recruitment to the lung upon allergen challenge associated with increased level of Eotaxin-1 in the BAL fluid . In addition, it has been demonstrated that mice with a deficiency of NADPH oxidase in non-hematopoietic cells have decreased lung eosinophilia during allergen challenge implicating the ROS in the production of Eotaxin-1 in the lung . Interestingly, it has been shown that dietary flavonoids inhibit Eotaxin-1 release from fibroblasts [69, 70]. Flavonoids have various anti-inflammatory properties and are potent inhibitors of NF-κB signalling  but are also potent activators of NRF2 . This inhibition of Eotaxin-1 observed is consistent with our study where we show inhibition of Eotaxin-1 with the triterpenoid CDDO. Based on our data with KEAP1 knockdown it can be concluded that the inhibitory effect that these flavonoids have on Eotaxin-1 is likely mediated directly by their activation of NRF2 and not through other anti-inflammatory mechanisms.
As the major eosinophil chemoattractant, Eotaxin-1 plays a critical role in allergic inflammation and asthma . In the lung Eotaxin-1 promotes the influx of eosinophils where activation and release of key mediators of an inflammatory response occurs . The role of the fibroblast in mediating eosinophil recruitment has long been established ; where it has been shown that fibroblasts derived from numerous sources secrete a significant amount of Eotaxin-1 in response to several pro-inflammatory stimuli [75–79]. Consistent with this, we have demonstrated in this report that IL-1β, IL-13 and TNFα all have potent effects on Eotaxin-1 secretion in fibroblasts. These factors are key inducers of Eotaxin-1 release and eosinophil recruitment in addition to contributing to fibrotic changes seen in airway disease [75, 79]. It would be of interest to evaluate an NRF2/ Eotaxin-1 relationship in fibroblasts from asthmatics to determine if Eotaxin-1 expression would be equally regulated by NRF2 activation is a disease state.
The mechanism by which Eotaxin-1 is modulated by NRF2 is not known. A detailed promoter study failed to identify a bonafide ARE  upstream of the human Eotaxin-1 gene, suggesting that this inhibition may be an indirect consequence of NRF2 activation. One way in which NRF2 has been shown to mediate its anti-inflammatory properties is through the inhibition of NF-κB. NRF2 and NF-κB have been shown to work together to modulate inflammatory gene expression [58, 60] and it has been suggested that NRF2 activation can lead to NF-κB inhibition [58, 81–83]. In addition it has been shown that the NF-κB pathway plays a critical role in Eotaxin-1 regulation in fibroblasts [30, 59]. While it is not clear if this is the case in our study, it is unlikely since we have demonstrated using pharmacological inhibition that all of the chemokines and cytokines induced by IL-1β and TNFα are NF-κB dependent, yet only Eotaxin-1 is inhibited by NRF2 activation.
Another key transcription factor that can mediate Eotaxin-1 expression is STAT6. A STAT6 binding site is present on the Eotaxin-1 promoter along with an NF-κB binding site and it is thought that Eotaxin-1 may be regulated by the concerted activity of NF-κB and STAT6 . STAT6 is of course a key mediator of Eotaxin-1 expression induced by IL-4 , but studies in fibroblasts have shown that STAT6 also is required for TNFα induced Eotaxin-1 expression . Thus, it remains feasible that in someway, NRF2 activation inhibits STAT6 activity, thus leading to the inhibition of Eotaxin-1 expression. There is no published data directly linking NRF2 activation to STAT6 activity, however, in one study using the licorice root triterpenoid Glycyrrhizin, it has been demonstrated that inhibition of Eotaxin-1 with this compound is associated with the inhibition of STAT6 phosphorylation and nuclear translocation . This data suggests that perhaps NRF2 does indeed regulate Eotaxin-1 expression through the regulation of STAT6 activity. Another potential mechanisms by which NRF2 may modulate Eotaxin-I expression is through modulation of MAPK signaling as it has been demonstrated that MAPK signaling downstream of TGFβ can synergize with IL-13 to induce Eotaxin-1 expression by interfering with negative feedback loops in the IL-13/STAT6 pathway . Interestingly it has been demonstrated that reactive oxygen species can directly augment the activity of STAT6  raising the possibility that a decrease in reactive oxygen species as a result of NRF2 activation may inhibit STAT6 activity and inhibit Eotaxin-1 expression.