A Drosophila model of cigarette smoke induced COPD identifies Nrf2 signaling as an expedient target for intervention

Introduction

Chronic obstructive pulmonary disease (COPD) is the third leading cause of death by now. It is characterized by high morbidities and mortalities, both in high-income and low-income countries, thus leading to substantial socioeconomic burdens. As smoking and aging are the two most important risk factors for various chronic lung diseases including COPD, it is anticipated that the prevalence of COPD will rise substantially with the smoking exposure and increase in life expectancies [1,2]. Clinically, COPD is characterized by progressive airflow limitations and the development of emphysema, which describes the loss of functional parenchymal lung tissue [3]. This loss of functional alveolar structures that is one major hallmark of COPD [4], seems to be caused by inappropriate repair mechanisms operative in the diseased lung [5] and the inability to cope with increased levels of reactive oxygen species (ROS) in the organ. Premature functional aging of the lung seems to be a major aspect associated with or underlying COPD development [5,6]. Long-term cigarette smoke (CS) exposure is by far the most important risk factor, but the confrontation to other air pollutants, also contributes to disease development. Despite the great body of work invested in recent years, the understanding of the molecular mechanisms underlying COPD development is unsatisfactory. This lack of understanding comes along with a lack of causative therapeutic strategies, which sets an urgent need for the development of novel, alternative therapeutic options [3,7].

Beside the study of patient material, tailored animal models are usually the most informative tools to enhance our understanding of the molecular frameworks underlying the development of human diseases in general. For COPD related research, a series of complex rodent models based on chronic long-term exposure to cigarette smoke served for this purpose [8–10]. Using these toolboxes, pathological alterations causally related to disease development have been identified. The apparent inability to control tissue regeneration in the lung and to cope with increased levels of ROS have been identified as major disease promoting alterations [11]. In this context, findings about modifications in signaling systems associated with tissue repair and ROS protection have been most important. Wnt signaling appears to take a central position in this context, as it is directly associated with tissue repair processes. Using tailored animal models, it has been demonstrated that COPD development comes along or depends on the switch from canonical towards non-canonical Wnt signaling, thus reducing the repair capacity of the lung [12]. Moreover, other signaling systems such as the JAK/STAT pathway that are also relevant for repair mechanisms, have been identified as potential targets for anti-COPD therapies [13]. IL5 and IL6, which are cytokines leading to activation of STAT transcription factors, have been identified as valuable targets for novel intervention strategies, because they have been associated with COPD development [14–16]. In addition to systems that are directly involved in tissue repair mechanisms, those that react towards high ROS and/or xenobiotic levels are also deeply involved in COPD pathogenesis. Among these signaling systems, the Nrf2 pathway takes a central position as it is strongly activated by CS exposure and its impairment has been unequivocally linked to COPD development in patients [17].

Despite the substantial efforts that have been made to develop suited animal models and to understand the molecular framework underlying COPD development, an urgent need for novel tailored models and, based on them, novel therapeutic strategies derived therefrom is obvious [11,18]. Very simple animal models may represent promising new candidates that complete the currently available toolbox for COPD research. Most important among these simple models for human diseases are tailored, transgenic Drosophila systems that have been highly informative for enhancing our understanding of numerous human diseases [19,20]. The usage of Drosophila to understand the molecular framework underlying human diseases has recently been expanded to chronic lung disease including asthma and lung cancer [21–24]. Here, we present a simple Drosophila model of COPD that depends on chronic low-level CS exposure. It recapitulates the most important pathological hallmarks of the disease, such as premature death, reduced physical activity, enhanced metabolic rates and a significant reduction of respiratory surfaces. Moreover, we could show that CS exposure regulates especially TGF-β, Nrf2 and JAK/STAT signaling in the airway epithelium and based on this, we demonstrated that pharmacological activation of Nrf2 signaling increases survival significantly in those animals chronically treated with CS.

Results

To evaluate if the fruit fly Drosophila melanogaster can serve as a useful model for the analysis of CS induced morbidities, we confronted animals of different genotypes and developmental stages with cigarette smoke using different exposure regimens. Chronic short period exposure (app. 30 min per day) had a substantial effect on lifespan in wild type flies (female Drosophila melanogaster, w¹¹¹⁸). This treatment reduced median lifespan statistically significantly by about 70% if compared with matching controls (Fig. 1A; p<0.001). In addition, we quantified the outcomes of CS exposure on different physiological and metabolic parameters. One major metabolic parameter is the body fat content. After 7 days of CS exposure, animals (females) had strongly reduced body fat levels if compared with untreated controls. This reduction was more than 50% of the control levels and it was statistically highly significant (Fig. 1B, p<0.001). We observed that the resting metabolic rates [25,26] of CS treated flies were almost twice as high as in non-treated matching controls (Fig. 1C, p<0.001). Moreover, animals confronted with daily doses of cigarette smoke showed a substantially reduced overall physical activity. Movement activities of CS treated and untreated animals were analyzed using the Drosophila activity monitoring system [27]. We observed sex-specific differences in the reaction towards CS exposure. Exposure to CS of only 1 week reduced the daily physical activity in males significantly (Fig. 2A, B) by about 60-70%. This reduction in movement activities was even more pronounced after 14d of CS exposure (Fig. 2C, D). In females, 7d exposure to CS led to a marginal decline in activity, which was not statistically significant (Fig. 2E, F). In contrast, 14d exposure induced a massive decline in physical activities also in females (Fig. 2G, H). A closer evaluation revealed that reduced activities are mainly observed in the major activity periods of flies, the so-called morning and evening peaks (Fig. 2A, C, E, G).

To evaluate if cigarette smoke exposure induces structural changes in the airway system, we used Drosophila larvae as their airway system is highly amenable to a detailed structural analysis and modified the protocol for cigarette smoke exposure accordingly. Prolonged confrontation with cigarette smoke (2 or 3 days with 3 periods of 30 min daily) had a significant impact on the terminal structures of the airway system, which are those regions of the airway system devoted to gas exchange. These terminal cells have long gas filled protrusions to deliver oxygen to the different organs of the fly (Fig. 3A). In response to two different regimens of cigarette smoke exposure, the corresponding animals showed a reduced respiratory surface. To demonstrate this effect, we analyzed the overall numbers as well as the total lengths of type II and type III branches using an ImageJ plugin (Fig. 3B, C). For all experiments we observed similar trends, likewise whether we analyzed the parameters after 2 or 3 days, with more pronounced reductions after 3 days. The number of type II branches was not changed by this intervention (Fig. 3D), whereas the number of type III branches was significantly reduced after both, 2 and 3 days of CS exposure (Fig. 3E). After 3 days of CS exposure the reduction was almost 20%. Quantification of the total lengths of type II branches revealed a decrease from controls to animals treated for 2 days t. Moreover, a further decrease was observed if the animals treated for 3 days were included (Fig. 3F). Type III branches show an even greater reduction in total lengths after 2 and 3 days of treatment, which was up to 20%. All values were significantly different (Fig. 3G).

To evaluate the underlying molecular mechanisms, we performed a RNAseg analysis of untreated as well as CS exposed airways from 3^rd instar larvae of the same genotype. In total, 143 genes were statistically significantly upregulated and 36 genes were statistically significantly downregulated (Supplementry Table 1). Displayed is a heatmap of the top 50 differentially transcribed genes with regard to their statistical support (Fig. 4A, p<0.05). To further evaluate the validity of these RNAseq results, we selected a small group of genes showing increased transcript levels. The transcript level quantification was performed in independent experiments by qRT-PCR experiments. Both GST genes (GstD2, GstD5) that were analyzed showed several hundredfold transcript levels in response to CS exposure. A similar increase was observed for Prx2540 and for the ligands of the JAK/STAT signaling pathway upd2 and upd3 (Fig. 4B). This observed increased expression was also seen in experiments using a promoter reporter line, where the presumptive Prx2540 promoter drives expression of Gal4. Based on the Gal4/UAS system [28], the promoter activity could be monitored by utilizing GFP as the reporter. As expected, CS exposure induced expression of GFP that is mainly confined to the tracheal system (Fig. 4C, D). Given the complexity of the cohorts of significantly regulated genes, we identified common signatures using various approaches. We compared the cohorts of regulated genes identified in the current study with those sets of genes that were regulated in response to infection [29] or in response to ectopic activation of the IMD pathway [30] in the airways of larval Drosophila (Fig. 4E). Moreover, we searched for transcription factor binding sites in the presumptive promoter regions of these genes using Pscan [31] and performed a detailed gene ontology analysis using the DAVID program package [32,33]. Using Pscan, only 4 different transcription factor binding sites were significantly enriched in the cohorts of regulated genes. These are Trl (p=6X10^-7), Mad (p=0.0052), opa (p=0.014) and CTCF (p=0.044). Moreover, we performed a GO analysis and retrieved significantly enriched KEGG pathways and GO terms. These are listed in SupplementaryTable 1 and comprise the enriched KEGG pathways: glutathione metabolism (p=1.1X10^-15), metabolism of xenobiotics by p450 (p=7.9X10^-16), and drug metabolism p450 (p=2.2X10^-14). The GO term analyses (Supplementary Table S2) revealed some suspicious categories including glutathione metabolic processes (p=2.6X10^-18), response to oxidative stress (p=0.0044), and regulation of JAK-STAT cascade (p=0.015) among others (Supplementary Table S1). Regulation of xenobiotic and glutathione metabolism is usually under the control of Nrf2 signaling. Thus, we focused on the two most suspicious signaling pathways, the Nrf2 and the JAK/STAT signaling systems for in-depth analyses. Thus, we employed reporter fly lines that specifically reflect activation of either pathway within the fly. We analyzed reporter lines that monitor activation of the JAK/STAT signaling pathway by subsequent production of GFP (1oXSTAT-GFP) and observed specific expression mostly confined to the trachea in CS exposed animals and here to those regions that are close to the spiracular openings (Fig. 5A, B). Moreover, we analyzed if the expression of the natural cytokines that activate the JAK/STAT signaling pathway are regulated. Here, no change could be observed for unpaired (upd, Fig. 5C, D), but strong activation in those parts of the tracheal system that are close to the spiracular openings could be observed for upd2 (Fig. 5E, F) and upd3 (Fig. 5G, H). We observed CS induced upd3 expression also in distal parts of the tracheal system (Fig. 5E, F). In addition to the JAK/STAT system, we focused on the Nrf2 signaling system and used reporter lines that demonstrate a specific activation of this pathway (Fig. 6). In the so-called ARE-GFP line, CS exposure resulted in significant induction of GFP expression especially in the dorsal trunks (Fig. 6A, B), a very similar pattern was observed for the GstD2-Gfp line, which serves as a reliable reporter of Nrf2 activation (Fig. 6C, D). To manipulate this signaling system experimentally, we employed oltipraz, a specific Nrf2 activator, and treated larval and adult Drosophila. Oral application of this compound induced activation of the ARE-GFP reporter both in larvae (Fig. 6E, F) and adults (Fig. 6G, H). Quantitative evaluation of the transcript levels of two Nrf2 responsive genes, namely of GstD2 and Prx2540 revealed a specific increase in the tracheal tissue, implying for a specific activation of the corresponding signaling system (Fig. 6J). Taking this information into account, we evaluated if manipulation of this signaling pathway interferes with the organism’s sensitivity to CS exposure. To evaluate the effects of chronic oltipraz application independent of CS exposure, we treated control animals with this compound for their entire adult life and compared the effects with those animals that were only treated with the solvent. In this case, the application of oltipraz led to a significantly reduced life span (median survival reduced from 38d for untreated controls to 30d for oltipraz treated ones), implying that this oltipraz concentration had a slight but significant (p<0.0001) negative effect on life span in general (Fig. 6K). Moreover, we measured the life spans of cohorts of flies fed with oltipraz as well as of matching controls that were chronically exposed to CS. We find that administration of oltipraz prolonged median life spans in these CS treated animals from 14d to 16d, which is an increase of approximately 15% (Fig. 6K, p<0.0001).

Discussion

Chronic cigarette smoke exposure is the major cause of COPD. Although substantial steps forward have been made in the last years, our understanding of the molecular and cellular mechanisms that underlie disease progression is still not satisfactory [18]. This lack of understanding is mirrored by the lack of causal and successful therapeutic interventions [7]. To improve this unedifying situation, novel research strategies are mandatory. Thus, informative and druggable animal models are highly appreciated. The Drosophila model conceived here, shows a complex set of pathological phenotypes in response to chronic, low level CS exposure that resemble those seen in human COPD patients. Flies chronically exposed to CS have a strongly reduced life span, they show reduced physical activities, increased metabolic rates and a decreased body fat content. The occurrence of this set of phenotypes implies that CS exposure induces very similar alterations in flies and in human patients suffering from the disease [34]. Surprisingly, even the respiratory surfaces of CS exposed flies were significantly reduced, which is one major hallmark of COPD [4,35]. In patients, the gradual decline of respiratory net surface in the alveolar space finally leading to the development of emphysema, is the major pathological development in the disease [3,18,36].

Based on this surprising number of phenotypical similarities that encompass all major hallmarks of the disease, it is reasonable to utilize this seemingly unsuited model for further studies. Based on the transcriptomic studies performed with airway epithelia of CS exposed and non-treated flies, we observed very robust regulation of a restricted cohort of genes that comprise various genes that are deeply involved in responses towards xenobiotics and ROS. The abundant presence of both, ROS and xenobiotics, in the lungs of smokers are thought to be intimately connected to disease development [36,37]. Very impressive was the regulation of sets of genes in the airway system of Drosophila that are directly associated with the glutathione biology. The regulation and performance of the complex glutathione metabolism has been shown to be highly relevant in smoke-induced lung diseases. Glutathione and especially its biologically active form, reduced glutathione, is depleted upon CS exposure [38,39]. Presumably to counteract this depletion of biologically active glutathione metabolites, the expression of numerous genes coding for enzymes involved in glutathione metabolism is upregulated substantially in response to CS exposure. Moreover, failures to produce sufficient amounts of glutathione cause worsening of the inflammatory state of the lung, and numerous associations of genes of the glutathione metabolism have been linked with CS induced lung diseases or cancer [40,41].

Thus, this highly impressive increase in expression of numerous genes associated with the glutathione metabolism is in line with observations that have been made in patient material or in mammalian models of chronic CS exposure [42,43]. In addition, the regulation of genes, whose products are tightly associated with the detoxification of xenobiotics was very intriguing. Most important in this context were several cytochrome P450 enzymes (Cyp) that are also known to take a central position as detoxifying enzymes in the human lung and that are also upregulated in response to CS exposure in human or murine systems [44,45]. Although these Cyp genes are induced in response to xenobiotics, their activity is not per se positive, as some irritants and carcinogenic compounds need to be modified by the activity of specific Cyps to unfold their pathogenic potential [46,47].

Usually, these sets of genes are under the transcriptional control of stress responsive signaling systems such as the Nrf2 pathway. We could show that Nrf2 signaling is indeed substantially upregulated in response to CS exposure in well-defined parts of the tracheal system of the fly. Impairments in the Nrf2 mediated anti oxidative response in the lung is thought to be associated with the development of COPD [48]. Thus, it is reasonable to assume that manipulation of Nrf2 signaling in the airway epithelium is a promising intervention strategy to interfere with CS exposure induced pathologies [49–51]. We used oltipraz, a specific activator of Nrf2 signaling [52] and applied it together with the food. In flies, the small body size enables the activation of Nrf2 signaling in the airways, presumably due to diffusion processes. Moreover, this pharmacological treatment partially rescued the early lethality induced by CS exposure and enhanced median lifespan by approximately 15%. Although seemingly a minor effect, translated to the human situation, it would be equivalent to an increase in life span from 70 to 84 years. Moreover, it contrasts the overall negative effect seen by oltipraz administration on the life span in otherwise untreated animals. Thus, based on these results, it can be assumed that directed activation of Nrf2 signaling in the airways is a promising strategy to prevent or delay CS induced morbidities. Attempts to follow this approach in clinical trials failed so far [53]. Sulforaphane, which was used as a Nrf2 activator was administered orally in this clinical study. Orally applied Nrf2 activators not necessarily induce significant activation of this pathway in the lung. Our result also implied that oral administration of a general Nrf2 activation throughout the entire organism has severe negative effects. Based on our results, a revitalization of the concept to use Nrf2 activators as COPD therapeutics seems to be reasonable, but local application forms, such as inhalation therapies might be much more suited to achieve a protective effect.

Two other signaling systems were identified in this study that have unequivocally been linked to COPD development. These are the JAK/STAT- and the TGF-β-pathways. JAK/STAT signaling is centrally involved in repair processes in various tissues and thus represents a key signaling pathway that is tightly connected with the genesis of COPD [13]. Deregulated activity of this pathway or of upstream cytokine systems within the lung are highly relevant for disease development [13]. Both, cytokines such as IL5 or IL6, as well as the terminal modules of the signaling cascade, namely STAT1 and STAT3 have been associated with disease development. In the Drosophila system, we observed a very similar type of activation caused by CS exposure. Again, both, the transcription factors, but also the cytokines activating the system were upregulated in the airway epithelium. Interfering with this system at different levels has been shown to be a promising strategy for the treatment of COPD (IL5, IL6). The results of the current study imply that an autocrine positive feedback loop is operative under these experimental conditions and might thus contribute to disease development.

Among other functional categories, JAK/STAT signaling has been shown to be tightly associated with tissue regeneration processes. Tissue regeneration in general appears to be specifically impaired in the lungs of COPD patients, implying that manipulation of JAK/STAT signaling holds the potential to specifically interfere with these pathological developments. Upregulation of this highly important signaling pathway can have two different underlying mechanisms, the response to cellular damage caused by CS exposure in the treated airway epithelia or a specifically induced signaling system that is launched to prevent further cellular damage. Upregulation of this signaling system usually is a double-edged sword, it can be highly effective in safeguarding the cell’s survival even under stressful conditions, but a chronically deregulated pathway may lead to massive inflammation and finally to the development of cancer.

The Drosophila COPD model presented here is a valuable extension of the animal model toolbox that is currently in use to study this disease. Animal models of COPD are mainly based on CS exposed mice or other rodents [11,54,55]. Although their use has been highly informative, they have some drawbacks where alternative models can yield complementary information. Rodent COPD models can nicely reproduce major features of COPD, but long-term studies, especially those that comprise life span as a major read-out are almost impossible to perform. Moreover, these systems are not suited for high-throughput approaches aiming to evaluate the effectiveness of novel therapeutic interventions. Here, the fly model can step in as it allows quantifying the effects on life span in high-throughput settings, which can speed up the search for new therapeutic interventions substantially.

Materials and Methods

Supplementary Materials

Supplementary Table 1

Supplementary Table 2