Transient adenovirus-mediated gene transfer of IL-1β (AdIL-1β), a proinflammatory cytokine, induces marked inflammation and severe and progressive fibrosis in rat lungs. This is associated with an increase in TGF-β1 concentration in bronchoalveolar lavage (BAL) fluid. TGF-β1 is a key cytokine in the process of fibrogenesis, using intracellular signaling pathways involving Smad2 and Smad3. In this study we investigate whether inflammation induced by IL-1β is able to independently induce lung fibrosis in mice deficient in the Smad3 gene. Seven days after AdIL-1β administration, similar levels of IL-1β transgene are seen in BAL in both wild-type (WT) and knockout (KO) mice, and BAL cell profiles demonstrated a similar marked neutrophilic inflammation. Phospho-Smad2 staining was positive in areas of inflammation in both WT and KO mice at day 7. By day 35 after transient IL-1β expression, WT mice showed marked fibrosis in peribronchial areas, quantified by picrosirius red staining and morphometry. However, there was no evidence of fibrosis or collagen accumulation in IL-1β-treated KO mice, and peribronchial areas were not different from KO mice treated with the control adenovector. TGF-β1 and phospho-Smad2 were strongly positive at day 35 in fibrotic areas observed in WT mice, but no such staining was detectable in KO mice. The IL-1β-induced chronic fibrotic response in mouse lungs is dependent on Smad3. KO and WT animals demonstrated a similar inflammatory response to overexpression of IL-1β indicating that inflammation must link to the Smad3 pathway, likely through TGF-β, to induce progressive fibrosis.
The role of inflammation in the pathogenesis of progressive pulmonary fibrosis (PF)3 remains a controversial question (1, 2). The prominent hypothesis is that PF is caused by chronic inflammation in response to an unknown etiologic agent, leading to tissue destruction, ongoing wound healing responses, and fibrosis. It has recently been proposed that inflammation is necessary to trigger the initiation of the fibrotic process, but subsequently plays a minor role in the progression of the disease, particularly in idiopathic PF (IPF). According to this hypothesis, IPF results from epithelial injury and abnormal wound repair and may progress in the absence of significant inflammation (3). The poor response of IPF to anti-inflammatory therapies supports the concept that inflammation is not the key process in the pathogenesis of the progressive disease. Research to further examine the relative roles of inflammation or fibrogenesis in IPF have been hampered by a lack of good animal models. Most models, such as the use of bleomycin, involve tissue damage and inflammation, at least for initiation of the process. Furthermore, examination of human tissue obtained from patients with lung fibrosis is merely a “snapshot,” and may not reflect the pathogenesis very well.
We have developed animal models of PF by using transient overexpression of cytokine genes in the lung of adult rodents (7–10 days) through adenoviral-mediated gene transfer. With this approach, we have shown that overexpression of TGF-β1 resulted in progressive fibrosis without a major inflammatory component, either in the initiation or progression of disease (4). Overexpression of TNF-α resulted in the opposite effect: substantial acute inflammation, with predominantly mononuclear cells and less frequent neutrophils, which resolved completely within 2 wk, and was not followed by significant progressive fibrosis (5). In contrast, overexpression of IL-1β caused severe acute inflammation, with more neutrophils than in the TNF-α model, which was associated with tissue destruction and subsequent progressive lung fibrosis (6). Interestingly, the first evidence of fibrosis was not observed until more than 1 wk after the exogenous proinflammatory cytokine had disappeared and the initial inflammatory response had subsided. The fibrotic remodeling was associated with a persistent up-regulation of endogenous TGF-β in this model, suggesting that progressive fibrosis is more related to an impairment of the tissue repair process and not to chronic inflammation.
The profibrotic effects of TGF-β are numerous, including induction of myofibroblasts, increase of matrix synthesis, and inhibition of collagen breakdown. Most of these effects are thought to be mediated through the Smad signaling pathway, because Smad3 null mice are protected from progressive fibrosis mediated by overexpression of TGF-β1 (7), do not develop lung fibrosis induced by bleomycin (8), and they are also protected against radiation-induced fibrosis of the skin (9). Smads are a family of cytoplasmic signal transducer proteins, and Smad2 and Smad3 predominantly mediate signals from activated TGF-βR (10) and interact with numerous TGF-β-responsive promoters (11).
In this study we used Smad3 null mice to address the question whether it is inflammation or the presence of an intact TGF-β/Smad3 signaling pathway that is crucial in mediating progressive lung fibrosis. We demonstrate that transient overexpression of IL-1β caused similar marked inflammation in Smad3-deficient and wild-type (WT) control mice. In contrast to the comparable acute events between the mice, progressive fibrosis developed only in WT but not in Smad3 null mice, providing direct evidence to our earlier contention that an intact TGF-β/Smad3 pathway is the key element in determining the progressive nature of PF in this model.
Materials and Methods
The construction of adenoviral vectors is described in detail elsewhere (12). For construction of AdhIL-1, human IL-1β (hIL-1β) cDNA (a gift from British Biotech Pharmaceuticals, Oxford, United Kingdom) was cloned into a p73 shuttle vector with a human CMV promoter and cotransfected with a virus-rescuing vector. The resulting replication-deficient virus (AdhIL-1) was amplified and purified by cesium chloride gradient centrifugation and PD-10 Sephadex chromatography, and plaque titerd on 293 cells (6). The control vectors, AdDL, with no insert in the deleted E1 region were produced in the same way (4).
Exon 8 of the Smad3 gene was disrupted in mice of background 129SV/EV × C57B/6, and Smad3 heterozygous mice (courtesy Dr. A. Roberts, National Institutes of Health, Bethesda, MD) were bred under special pathogen-free conditions. The genotypes of both WT and Smad3 knockout (KO) mice were determined by PCR analysis on tail DNA obtained from 3-wk-old animals. Rodent laboratory food and water were provided ad libitum. The animals were treated in accordance with the guidelines of the Canadian Council of Animal Care. All animal procedures were performed under inhalation anesthesia with isofluorane (MTC Pharmaceuticals).
Eight-week-old animals received AdhIL-1β or AdDL (1 × 109 PFU) in a volume of 20 μl of PBS by intranasal administration. Mice were sacrificed at day 7, 14, and 35 after adenoviral administration by abdominal aorta bleeding. Bronchoalveolar lavage (BAL) was then performed as described previously. For histological examination, lungs were inflated and fixed by intratracheal instillation of 10% neutral-buffered formalin at a constant pressure of 20 cm of water for 5 min. For RNA after washing with PBS, the lung was removed and frozen immediately in liquid nitrogen. Frozen tissue samples were ground and stored at −70°C until further processing.
Determination of cytokine levels in BAL fluid
A human IL-1β ELISA (R&D Systems) was used to determine levels of active IL-1β, according to the manufacturer’s recommendations.
Histology and histomorphometry
After fixation in 10% buffered formalin for 24 h, longitudinal sections of the lung were paraffin embedded, sectioned, and stained with either H&E or picrosirius red. Immunohistochemistry was conducted using phospho-Smad2 and TGF-β1 Abs as described previously (13).
For quantitative lung collagen histomorphology, 25–30 random fields (magnification ×25, picrosirius red) were image digitized in polarized light with Leica DMR microscope and analyzed using Leica Qwin Image Processing Software (Leica Imaging Systems) as described previously (14).
Frozen lungs were homogenized in 5 ml of TRIzol (Invitrogen Life Technologies), and RNA was extracted using standard methods. RNA integrity and concentration were determined with a microgel bioanalyzer (Agilent 2100; Agilent). RNA (1 μg) was DNase treated reverse transcribed using a standard protocol (Invitrogen Life Technologies). Quantitative real-time PCR was conducted using an ABI Prism 7700 Sequence Detector (Applied Biosystems). Negative control samples (no template or no reverse transcriptase) were run concurrently. Results were normalized to 18S. Primers (Mobix) and probes (Applied Biosystems) are shown in Table I.
|.||Forward Primer .||Reverse Primer .||Probe .|
|.||Forward Primer .||Reverse Primer .||Probe .|
Data are shown as mean ± SEM. For evaluation of group differences, we used the Student’s t test. Values of p < 0.05 were considered significant.
Transient hIL-1β overexpression induces similar inflammatory responses in Smad3 WT or KO lungs
An ELISA specific for human IL-1β was used to detect and quantify the overexpressed transgenic human IL-1β. By day 7 there were equally high hIL-1β levels in Smad3 WT and KO BAL fluid (Fig. 1). The level of hIL-1β in BAL fluid from control vector-treated animals was below the limit of detection. By day 14, hIL-1β levels were undetectable (<10 pg/ml) in AdhIL-1β-treated animals in both WT and KO animals.
As we have observed previously (6), IL-1β overexpression resulted in severe acute inflammation. By day 7, the total cell count from BAL fluid was increased (Fig. 2,A) with a predominance of neutrophil granulocytes (Fig. 2 B) and no significant difference between WT and KO mice (neutrophil content 65 and 51% of BAL cells, respectively). After control vector administration, the total cell number was significantly lower compared with AdhIL-1β-treated animals with <5% neutrophils in both WT and KO mice. We have shown in a previous study (7) that, by day 14 after control vector AdDL, neutrophil percentage in WT and KO is similar to that observed in untreated mice. The amount of neutrophils 14 days after AdhIL-1β was mildly increased (12 and 15%, respectively) with no significant difference between WT and KO mice. By day 21, the total cell counts returned back to untreated baseline values in all of the animal groups (data not shown).
A mild inflammatory response 7 days after control vector injection was present in both WT or KO mice as assessed by histology, similar to the data seen in our earlier study, reflecting an antiviral immune response maximum at this time (7). In contrast, AdhIL-1β caused severe inflammation in the airways and parenchyma of both WT and KO lungs with no difference between the two groups as shown in Fig. 2 C.
Transient IL-1β overexpression induces similar early activation of the TGF-βR I/Smad pathway in Smad3 WT and KO
There was a high content of phosho-Smad2 in both Smad3 WT and KO mice 7 days after AdhIL-1β administration as seen by immunohistochemistry. The positive staining was predominantly present in areas with inflammatory infiltrates. No major Smad2 phosphorylation was detected in WT or KO animals after control vector administration (Fig. 3).
Transient IL-1β overexpression induces a fibrotic tissue response in Smad3 WT but not KO mouse lungs
Smad3 WT and KO animals treated with control adenovectors were not distinguishable from age-matched naive animals 35 days following vector administration (Fig. 4,A). Airspace enlargement was found in naive Smad3 KO lungs as described previously (7). In contrast, transient overexpression of IL-1β resulted in a fibrotic pattern seen in lungs of WT mice after 35 days, characterized by mesenchymal cell and extracellular matrix accumulation predominantly in peribronchial and perivascular areas of the tissue (Fig. 4,A). Positive staining with picrosirius red, which is a collagen-specific staining method, demonstrated that this matrix accumulation was predominantly composed of collagen (Fig. 4,A). In contrast to WT mice, lungs from AdIL-1β-treated Smad3 KO mice did not show any increase in extracellular matrix deposition or collagen accumulation (Fig. 4 A). They were not different from Smad3 KO lungs treated with control virus.
The morphological impression of increased collagen accumulation was quantified by morphometric analysis using image processing software (Fig. 4 B). Transient IL-1β overexpression induced a significant collagen accumulation by day 35 in WT (82% increase compared with control vector-treated WT; p < 0.05), whereas no increase was present in KO mouse lungs. There was no difference in collagen deposition between Smad3 WT and KO lungs following control vector administration.
Moreover, procollagen 3A1 gene expression was examined by quantitative RT-PCR in Smad3 WT and KO mouse lungs at 7 and 14 days after AdhIL-1β exposure and compared with naive mice. The level of procollagen mRNA was similar in nontreated Smad3 WT and KO mice (Fig. 4,C). At day 7 and 14 after AdhIL-1β administration, procollagen 3A1 gene expression was increased in WT (1.6- and 2.24-fold, respectively) compared with untreated control animals, whereas there was no increased expression in KO animals (Fig. 4 C).
Transient IL-1β overexpression is associated with prolonged TGF-β1 up-regulation in Smad3 WT, but not KO lungs
TGF-β1 and Smad2 phosphorylation was assessed by immunohistochemistry at day 35 postadenovector to examine for TGF-β protein expression and indirectly for biological activity at this late fibrotic stage. Both TGF-β1 and phospho-Smad2 were strongly positive in fibrotic areas in WT mice after transient IL-1β overexpression (Fig. 5). In contrast, no positive staining for TGF-β1 or phospho-Smad2 was seen in Smad3 KO mice 35 days after AdIL-1β as well as in Smad3 WT and KO lungs after control vector administration.
In this study we clearly demonstrate that the progression of IL-1β-induced inflammation toward fibrosis is Smad3 dependent, because it developed only in WT control mice and not in Smad3 null mice. This observation implies that there is a common pathway to tissue fibrosis, and other, non-Smad signaling pathways are not sufficient for the induction of fibrosis. This finding also supports our previous hypothesis that it is not so much the inflammation but more the presence and stimulation of an intact TGF-β/Smad3 pathway that is the key element in determining the progressive nature of PF.
Whether inflammation is necessary for the progression of fibrosis in PF is a controversial issue. There is evidence that both supports and refutes the inflammatory hypothesis of PF, resulting in part from differences in the animal models that are used for investigating the pathogenesis of PF.
The best studied fibrosis model is the administration of bleomycin, which causes a severe acute inflammatory response, followed by chronic inflammation and fibrosis. It has been repeatedly shown that the degree of inflammation in bleomycin injury models is associated with the intensity of development of fibrosis. One very recent publication demonstrated that the cysteinyl leukotriene 1 receptor controls the severity of chronic pulmonary inflammation and fibrosis, and it was suggested that interference with the axis of cysteinyl leokotriene synthesis may be beneficial in IPF (15). Similarly, mice lacking superoxide dismutase develop more inflammation and more fibrosis in response to bleomycin (16). However, it is possible to reduce bleomycin-induced fibrosis without affecting the inflammatory component. This can be achieved by interfering with active TGF-β either by administration of TGF-β Abs, or overexpression of the TGF-β binding glycoprotein decorin (17). More support for the key role of TGF-β, but not inflammation is derived from studies in αvβ6 integrin-deficient mice. These mice are unable to activate TGF-β from the latent form and generate marked inflammation on exposure to bleomycin but fail to develop PF (18). Likewise, if the platelet-derived growth factor receptor tyrosine kinase is blocked by imatinib, there is no alteration in the inflammation, but a marked reduction in bleomycin fibrosis (19, 20), suggesting the Abl pathway from the TGF-βR may also be involved in fibrogenesis. Despite recognition of all the knowledge that was acquired using this animal model, it is widely accepted that it in some respects is a limited model of PF because of the intense inflammatory component not usually seen in human disease, and the fibrosis is partly resolving after prolonged observation (21).
We have established models in which we transiently overexpress individual cytokines in the lung by adenoviral-mediated gene transfer. With this approach, we have shown that overexpression of TGF-β1 resulted in progressive fibrosis without a major inflammatory component, either in the initiation or progression of disease (4). Overexpression of TNF-α resulted in the opposite effect: substantial acute inflammation, which resolved completely within 2 wk, and was not followed by significant fibrosis (5). Novel models with inducible transgenic animals have provided further support that the uncontrolled production of growth factors may play a more prominent role in determining fibrosis than the presence of inflammation. One example is the inducible IL-13 transgenic mouse, in which the development of airway and lung fibrosis is related to persistent up-regulation and activation of TGF-β, but not inflammation (22). Taken together, these data indicate that inflammation and fibrosis can be, but do not have to be, interrelated.
To further delineate the relation between inflammation and fibrosis, we have investigated the pulmonary effects of the highly proinflammatory cytokine IL-1β. In contrast to TNF-α, overexpression of IL-1β caused severe acute inflammation, which was associated with tissue destruction and subsequent progressive lung fibrosis (6). Interestingly, fibrosis developed 1 wk after the disappearance of the initial inflammatory response. The fibrotic remodeling was associated with a persistent up-regulation of endogenous TGF-β in this model, strongly suggesting that progressive fibrosis is more related to an impairment of the repair process and less to chronic inflammation.
The work presented in this study provides further evidence that inflammation acts through a common pathway in the induction of fibrosis. Inflammation alone is unable to produce significant fibrosis in the lung in the absence of a functional fibrogenic signaling pathway, such as mediated through Smad3. The Smad pathway is believed to be the major signaling mechanism through which active TGF-β stimulates the induction of profibrotic genes. Smad3 KO mice are resistant to development of lung fibrosis but not inflammation caused by bleomycin, and are also resistant to the profibrotic AdTGFβ. In this study, we have exposed Smad3 KO mice to AdIL-1β and examined the expression of TGF-β and related genes. We show that both Smad3 WT and KO mice have similar inflammation at an early time when the IL-1β transgene is highly expressed (day 7). Both also demonstrate similar degree in early Smad2 phosphorylation. Although IL-1β does not involve the Smad pathway directly (23) but has been shown to induce TGF-β, there are likely comparable levels of bioactive TGF-β generated following IL-1β overexpression in this model. Despite this, Smad3 KO mice did not develop progressive tissue fibrosis, whereas WT mice did.
TGF-β autoinduction is thought to be an important aspect in the maintenance of the fibrotic process over time (24). We have previously shown that TGF-β autoinduction is blocked in primary lung fibroblast from Smad3 KO mice (7). In the current study we looked at the late expression of TGF-β and phospho-Smad2 at day 35, and there was a persistent presence of both in the fibrotic lungs of WT mice, but not in Smad3 KO mouse lungs, indicating endogenous autoinduction of TGF-β in the fibrotic phenotype. These findings suggest that IL-1β is able to induce and/or activate TGF-β, but the progression of fibrosis requires an intact TGF-β signaling pathway when inflammation is no longer present. However, whereas strong evidence supports a prominent role of the TGF-β signaling pathway in the development of lung and skin fibrosis, its role is more controversial in the liver. Fibrogenesis in a model of schistomosiasis, which causes a chronic Th2-dominated granulomatous liver disease, appears to be independent from the induction of TGF-β, and is possibly mediated through direct profibrotic effects of IL-13 (25).
The Smad3 pathway may have an impact on inflammation because TGF-β is a known anti-inflammatory cytokine (26), and KO experiments for the three mammalian isoforms of TGF-β have indeed demonstrated their importance in regulating inflammation (27). We have already shown that naive 8-wk-old KO mice have a significant 35% increase in total cell count in BAL fluid compared with WT littermates. However, there was no significant difference in the differential cell count, except for a slight increase in neutrophils in KO animals compared with WT (7). After IL-1β overexpression there was no difference in the inflammatory response between Smad3 KO and WT or the kinetics of inflammation as assessed by cell counts and histology in both mice. This suggests that in the presence of overexpression of IL-1β, an intact Smad3 pathway is neither crucial for the inflammatory response to occur nor for its recovery.
Our data demonstrate that IL-1β can cause lung fibrosis, but not necessarily through the inflammatory component, but through induction of TGF-β, and its consequent autoinduction. It cannot be excluded that IL-1 may be involved in triggering the initiation of PF, and a significant role is suggested by its increased presence in fibrotic areas of IPF lungs (28) and by gene polymorphism studies. However, the concept of a nonessential role for IL-1 and inflammation in IPF fits with the lack of efficacy of anti-inflammatory therapy (29). Furthermore, it cannot be neglected that acute lung injury and adult respiratory distress syndrome—conditions characterized by dramatic levels of IL-1, TNF, and acute inflammation in the lungs—often resolve without chronic organ damage. If progressive fibrosis occurs, it is frequently accompanied by high levels of fibrogenic cytokines such as TGF-β (30).
In summary, we demonstrate in this study that the late progressive fibrotic response following transient IL-1β overexpression is dependent on the presence of an intact Smad pathway, and that inflammation alone is insufficient to induce significant fibrosis by another route of signaling. We further provide evidence that the progression of fibrosis in this model is likely due to persistent autoinduction of TGF-β. This suggests that there is a common pathway to tissue fibrosis, and inflammation may not be required for the progression of PF. This may explain the poor efficacy of anti-inflammatory drugs in this disease. Furthermore, these observations suggest that future promising therapeutic interventions in PF should target this common fibrogenic pathway.
We thank Jane Ann Schroeder, Carol Lavery, Duncan Chong, Xueya Feng, and Mary Jo Smith for their invaluable technical help.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work is supported by Canadian Institutes of Health Research, St. Joseph’s Healthcare and Hamilton Health Sciences. P.B. is supported by the Bourses Lavoisier and the Ligue Bourguignonne Contre le Cancer. P.J.M. is a Canadian Institutes of Health Research Clinician Scientist. M.K. is a Parker B. Francis Fellow.
Abbreviations used in this paper: PF, pulmonary fibrosis; IPF, idiopathic PF; WT, wild type; hIL-1β, human IL-1β; KO, knockout; BAL, bronchoalveolar lavage.