Is there still hope for single therapies: How do we set up experimental systems to efficiently test combination therapies?

Heterogeneity of asthma

Asthma is a chronic inflammatory disease that is characterized by recurrent episodes of wheezing, breathlessness, chest tightness and cough. The symptoms are usually associated with increased airway hyperresponsiveness (AHR) and variable airflow obstruction, which is often reversible, either spontaneously or after treatment.1 Many asthma patients—especially those with early onset—are atopic, meaning they are genetically predisposed to generate immunoglobulin E (IgE) antibodies upon exposure to environmental allergens, such as house dust mite (HDM), pollen and animal dander. Repeated inhalation of such allergens can subsequently induce an IgE-mediated hypersensitivity reaction that is characterized by increased vascular permeability, vasodilatation, smooth muscle contraction and airway inflammation. Asthmatic airways also undergo structural changes (airway remodeling), with epithelial changes, increases in smooth muscle mass, deposition of extracellular matrix proteins and goblet cell hyperplasia.2, 3

For many years, asthma has been considered as a single disease, characterized by variable airways obstruction and a T helper 2 (T_H2)-driven eosinophilic, corticosteroid-responsive inflammation. As a consequence of this, and after successful clinical trials in the 1990s and early 2000s, inhaled cortico steroids (ICSs) and bronchodilators have become the golden standard of chronic asthma treatment. It has however become evident that asthma is a heterogeneous disease and that asthma is an ‘umbrella term’ that encompasses multiple subgroups or ‘phenotypes’.4 Using microarray and polymerase chain reaction analyses on airway epithelial brushings, patients with mild to moderate asthma could be divided in at least two distinct molecular phenotypes defined by the degree of T_H2 inflammation. Non-T_H2-driven asthma was found in nearly half of the patients, and these patients responded poorly to inhaled steroids.5 Using induced sputum, Simpson and colleagues found that asthma could be categorized into four inflammatory subtypes based on sputum eosinophil and neutrophil proportions. These subtypes were neutrophilic asthma, eosinophilic asthma, mixed granulocytic asthma and paucigranulocytic asthma.6 In a longitudinal study, McGrath and colleagues found that approximately half of the patients with mild-to-moderate asthma have persistently non-eosinophilic disease, a phenotype that responds poorly to the classical anti-inflammatory therapy.7

Single versus combination therapy for asthma

ICSs form the cornerstone of the treatment of asthma. Addition of a long-acting β₂-agonist (LABA) is the first choice in the next treatment step. Both drugs have broad and ‘non-specific’ actions and work synergistically.8 In fact, they are an example of an optimized form of administration of pharmacological principles (corticosteroids and adrenergic agonists) that have been known for more than half of a century. Several pivotal clinical trials have shown the clinical effects of ICS given alone, or in combination with LABA.9, 10 Treatment with ICS at low doses reduces asthma symptoms, increases lung function, improves quality of life and reduces the risk of exacerbations and asthma-related hospitalizations and death. Adding LABA to the same dose of ICS provides additional improvements in symptoms and lung function with a reduced risk of exacerbations.1

In the past 20 years, many breakthroughs have been made in understanding the immunological and inflammatory mechanisms that initiate and mediate the development of asthma, specifically with regard to allergic airway responses.11 This has led to the development of targeted therapeutics such as the anti-leukotrienes and the cytokine antibodies anti-tumor necrosis factor (TNF)-α, anti-interleukin (IL)-4, anti-IL-5 and anti-IL-13.

The initial hope that by interfering specifically with one pathway, a major clinical impact on asthma would be possible proved to be rather unrealistic. An example is inhibition of the leukotriene pathway by 5-lipoxygenase inhibitors or by blockade of the cysLT₁ leukotriene receptor. One of the big lessons from the development of anti-leukotrienes was that, although leukotrienes were found to be potent mediators of bronchoconstriction and plasma extravasation, the clinical anti-asthma effects of anti-leukotrienes were rather modest. Leukotriene receptor antagonists have an activity that is at best comparable to a low dose of ICSs and have consequently been positioned as an alternative to steroids in mild asthma or as an add-on therapy in more symptomatic patients. To be noted is that they were found to be more active in specific clinical situations, such as exercise-induced bronchoconstriction and aspirin-induced asthma, probably because in these situations leukotriene production is stimulated.1, 12, 13

Anti-TNF therapy has been very successful in inflammatory diseases such as rheumatoid arthritis and inflammatory bowel disease. In asthma (and COPD) however, initial enthusiasm on the possibilities of anti-TNF therapy in severe asthma has rapidly waned. TNF-α is an important cytokine in the innate immune response and plays a key role in immediate host defence against invading microorganisms before activation of the adaptive immune system. It is principally produced by macrophages in response to membrane-bound pattern recognition molecules such as Toll-like receptors, which detect common bacterial cell surface products, such as lipopolysaccharide. Moreover, it is produced by monocytes, dendritic cells, B cells, CD4+ cells, neutrophils, mast cells and eosinophils, and structural cells (fibroblasts, epithelial cells, smooth muscle cells). In preclinical models, its involvement in asthma was suggested and it was developed for treatment in severe, refractory asthma.14 Small studies of an antibody to TNF-α (infliximab) or the soluble TNF-α receptor (etanercept)15 reported mixed results in patients with a range of asthma severity. The initial enthusiasm was however tempered by the results of a large study on the safety and efficacy of golimumab, a fully human monoclonal antibody to TNF-α similar to infliximab. The study involved 309 patients with uncontrolled, severe persistent asthma. An unfavourable risk–benefit profile observed in patients who received golimumab led to early discontinuation of study after the week 24 database lock. Moreover, treatment with golimumab failed to achieve significant treatment effect on either of the two coprimary endpoints (forced expiratory volume in 1 s (FEV₁) or severe asthma exacerbation).16

Nonanaphylactic, humanized anti-IgE antibodies (e.g. omalizumab) have become a new therapeutic option for patients with severe allergic asthma. The therapeutic rationale beyond the idea derives from the ability of the anti-IgE antibodies to bind to the same domains on the IgE molecule that interact with the high affinity IgE receptor, thereby interfering with the binding of IgE to this receptor without cross-linking the IgE on the receptor (nonanaphylactic anti-IgE antibodies). Treatment with anti-IgE antibodies leads primarily to a decrease in serum IgE levels. As a consequence thereof, the number of high-affinity IgE receptors on mast cells and basophils decreases, leading to a lower excitability of the effector cells reducing the release of inflammatory mediator such as histamine, prostaglandins and leukotrienes. Experimental studies in mice indicate that the injection of some monoclonal anti-IgE antibodies also inhibited IgE production in vivo.17 Anti-IgE, omalizumab, inhibits the early and late allergen response in patients with asthma, and this is paralleled by a reduction in eosinophils and a decline in IgE-bearing cells post-allergen without changing the nonspecific AHR (PC20 methacholine).18 One of the pivotal studies was the ‘innovate trial’ performed in patients with severe allergic asthma. The patients included were treated with high-dose ICSs and LABA, and had a reduced lung function and a recent history of clinically significant exacerbations. In those patients with inadequately controlled severe persistent asthma, addition of omalizumab significantly reduced the rate of clinically significant asthma exacerbations, severe exacerbations and emergency visits.19

IL-5 is a key cytokine involved in eosinophil differentiation, recruitment, activation and survival. In recent years, persistent eosinophilic inflammation has been linked to the occurrence of exacerbations in a group of patients with refractory asthma.20 Humanized monoclonal antibodies against IL-5 are now in full clinical development, after years of hesitation as to whether their development would be clinically meaningful. Indeed, in a small dose finding study, a single dose of the anti-IL-5 monoclonal antibody mepolizumab decreased blood eosinophils for up to 16 weeks and sputum eosinophils at 4 weeks. However, in the allergen challenge model, no significant effect of anti-IL-5 on the late asthmatic response or on AHR to histamine was observed.21 Further on a study on a large group of asthmatic patients experiencing persistent symptoms despite ICS therapy (400–1000 mg of beclomethasone or equivalent) was negative on clinical outcomes such as lung function, symptom scores and exacerbation rates.22 However, in a small proof-of-concept trial, a more favourable outcome was found. In this study, asthmatic subjects with refractory eosinophilic asthma and a history of recurrent exacerbations were included. Treatment with mepolizumab significantly reduced the number of asthma exacerbations that resulted in the prescription of corticosteroid therapy and increased asthma-related quality of life.23 The effect of mepolizumab was confirmed in a large multicentric trial (Dose Ranging Efficacy And safety with Mepolizumab in severe asthma).24 It is important to note that in these two recent trials, there was no significant improvement in symptoms or in FEV₁, measures that are commonly used for quantifying asthma control, illustrating the importance of choosing the right outcomes when designing clinical studies.

Epithelial factors are crucial in the initiation of an allergic response. Thymic stromal lymphopoietin (TSLP) is an epithelial-derived cytokine. TSLP overexpression in murine bronchial epithelial cells boosts T_H2 immunity in the lungs. However, in mouse models of asthma, driven by natural allergens (such as HDM), the neutralization of TSLP does not necessarily lead to reduced disease (reviewed in Lambrecht and Hammad11). Keeping this in mind it is of interest to note that in the human allergen-induced bronchoconstriction model, a humanized monoclonal antibody against TSLP (AMG 157) did reduce the allergen-induced bronchoconstriction, as well markers of inflammation (i.e. levels of blood and sputum eosinophils).25 As the authors rightly state, further investigation of the potential clinical benefit is needed.

Another pathway of development concerns the effects of macrolides in asthma. Macrolides such as erythromycin, clarithromycin and azithromycin are antibiotics that have anti-inflammatory and immunomodulatory effects in addition to their antibacterial effects. Maintenance treatment with low-dose macrolides has been proven to be effective, well tolerated and safe in several chronic neutrophilic airway diseases, including cystic fibrosis (CF), diffuse panbronchiolitis, exacerbation-prone non-CF bronchiectasis and COPD. Both antimicrobial and anti-inflammatory effects are supposed to mediate the efficacy of macrolides in neutrophilic chronic airway diseases, including the neutrophilic phenotype of severe asthma. The first studies of macrolides have been performed in small numbers of patients with mild-to-moderate asthma, had short study durations (less than 12 weeks) and focused on lung function as primary outcomes. Taking these limitations into account, it is not surprising that most studies were negative.26 Indeed, the role of neutrophilic airway inflammation becomes more prominent in severe asthma, and the optimal primary endpoint in longer term studies of severe asthma should be the effect on asthma exacerbations (expressed as the rate of exacerbations, the total number of exacerbations or the time to the first exacerbation).26 In the ‘azithromycin for prevention of exacerbations in severe asthma (AZISAST)’ study, we performed a randomized double-blind, placebo-controlled trial in adult patients with exacerbation-prone severe asthma. Although maintenance treatment with low-dose azithromycin (250 mg three times a week) for 6 months did not affect the rate of primary endpoints compared with placebo, azithromycin treatment was associated with a significantly lower rate of exacerbations in the predefined subgroup of patients with noneosinophilic severe asthma.27

The story of macrolide treatment in asthma is reminiscent of the development of anti-IL-5 monoclonal antibodies as add-on treatment in asthma.26 The first randomized clinical trials with anti-IL-5 studied patients with mild-to-moderate asthma, the enrolled patients were not phenotyped, and lung function was the primary outcome. Thanks to targeting anti-IL-5 treatment to patients with severe asthma and refractory airway eosinophilia, and thanks to choosing the correct primary outcome (i.e. asthma exacerbations), several investigators have now shown that anti-IL-5 significantly reduces exacerbations. The clinical development pathway of anti-IL-5 and macrolides, thus, underlines the importance of subphenotyping patients with severe asthma and of targeting specific add-on therapies to the correct severe asthma phenotype.26

Potential pitfalls of experimental models

Mouse models have traditionally focused on allergic asthma, using most frequently as sensitizer, the antigen ovalbumin and, more recently, HDM. The three classical outcome parameters that have been used are sensitization (total IgE and antigen-specific IgE), airway inflammation (bronchoalveolar lavage fluid and airway histology) and bronchial hyperresponsiveness. A great advantage of mouse models is their well-characterized immunology and the availability of powerful experimental tools such as the genetically modified mice (transgenic, knockouts, conditionally knockouts). It has however proved very difficult to develop real chronic asthma models mimicking the human disease. One of the reasons is the development of tolerance to the antigen(s). Moreover, the mouse airway anatomy is very different from the human airways. In contrast to mice, the human airways have extensive airway branching, do have submucosal glands (not present in mice) and the smooth muscle cells are organized in spiral bundles. Translation of data found in mouse models has proved to be more difficult than expected, for instance in the development of anti-IL-42 and anti-IL-5.28 This is in part because asthma is a heterogeneous disease that is frequently non-allergic, and in part because wrong outcomes were chosen for clinical trials (e.g. use of FEV₁ instead of a biomarker such as sputum or blood eosinophils for anti-IL-5). Moreover, in real life, interactions between allergen and infections or between allergen and air pollution (cigarette smoke, diesel exhaust particles) take place. These combined exposures are now also used in the mouse model.29

Heterogeneity of COPD

According to the latest Global Initiative for Chronic Obstructive Lung Disease guidelines, COPD is defined as ‘a common preventable and treatable disease, characterized by persistent airflow limitation that is usually progressive and associated with an enhanced chronic inflammatory response in the airways and the lung to noxious particles or gases. Exacerbations and co morbidities contribute to the overall severity in individual patients’.30 COPD is characterized physiologically by a poorly reversible airflow limitation and pathologically by bronchiolitis and emphysema.31 In western countries, COPD is mainly caused by cigarette smoking, is incurable and is progressive, even after smoking cessation.32 In addition to well-characterized innate immune responses, recent findings on auto-antibodies (against lung matrix and epithelium) in patients with COPD suggest a contribution of an ongoing adaptive immune response to the persistence of chronic airway inflammation.33 In recent years, data coming out of large cohort studies, such as the ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints) study, have pointed to the heterogeneity of COPD. Poor correlations between FEV₁, symptoms, quality of life, functional outcomes and biomarkers was found. Systemic inflammation was present in only a limited proportion of patients, and did not relate to baseline characteristics or disease progression.34 Some patients with COPD exacerbate frequently, while others do not. In the ECLIPSE study, exacerbations were tracked over time: a frequent exacerbator phenotype (i.e. two or more exacerbations per year) appeared to be relatively stable over a period of 3 years.35 The disease course was very variable, the rate of change in FEV₁ is highly variable, and close to a third of patients was not progressing at all.36

Single versus combination therapy for COPD

The cornerstone of the present pharmacological treatment for COPD consists of long-acting bronchodilators (β₂-agonists, anticholinergics) to which ICSs are added in more severely obstructed patients with an exacerbation history.37, 38 The bronchodilator treatment is intended to be symptomatic, as, per definition, COPD is characterized by a largely irreversible airways obstruction. Both inhaled long-acting bronchodilators and ICSs have effects on patient-reported outcomes and on exacerbations, and, as for asthma, an additive effect of both drugs has been documented.30

The current bronchodilator therapy in COPD is the result of an improvement of existing β₂-agonists and anticholinergic agents, and has resulted in development of long-acting bronchodilators. The LABA and the long-acting muscarinic antagonists (LAMA) allow a twice or once-daily dosage.38 The additive bronchodilator effect of two short-acting bronchodilators (i.e. salbutamol/ipratropium bromide) is well known, and this proved, not unexpectedly, to be also the case for long-acting bronchodilators, that is, the combination of formoterol and tiotropium bromide.39 So, while the combination of LABA/ICS is already well established, the introduction of several combinations of long-acting bronchodilator treatment (LABA/LAMA, i.e. indacaterol/glycopyrronium bromide; vilanterol/umeclidinium) is expected for 2014/2015. The dual bronchodilator QVA149 has a better effect on lung function than the isolated components indacaterol and glycopyrronium bromide. Moreover, in two 6 months studies therapy, dual combinations of bronchodilators (indacaterol/glycopyrronium bromide; vilanterol/umeclidinium) were shown to be superior to the single bronchodilators, with regard to clinically meaningful outcomes such as dyspnoea and quality of life scores.40, 41 Whether there will be an additive effect of these long-acting bronchodilators on the occurrence of exacerbations is not yet clear.42

COPD is associated with an inflammatory response in the lungs, and suppression of the inflammatory response is a logical approach to the treatment of COPD. Two approaches are being explored: either drugs that directly influence the cellular components of inflammation or drugs that inhibit inflammatory mediators.43 An example is phosphodiesterase-4 (PDE4) inhibition. Drugs that inhibit PDE4 have a wide range of anti-inflammatory actions in vitro and in vivo, and provide a novel approach to the treatment of COPD. Roflumilast, a new PDE4 inhibitor, reduces airway inflammation in COPD, as assessed with sputum neutrophil and eosinophil counts. In two placebo-controlled, double-blind, multicentre trials with identical design, patients with COPD older than 40 years, with severe airflow limitation, bronchitic symptoms and a history of exacerbations were randomly assigned to oral roflumilast or placebo for 52 weeks. Roflumilast reduced exacerbation frequency and induced small but significant improvements in FEV₁.44-46 Patients with COPD are less responsive to ICSs than patients with asthma, and this has been explained by the presence of neutrophilic inflammation. As in severe neutrophilic asthma, macrolides have the potential to reduce the frequency of COPD exacerbations.47-49

As in asthma, anti-TNF treatment has not fulfilled the initial enthusiasm and has been disregarded as treatment for COPD.50, 51 Other possible specific inhibitors are being explored, for example p38 MAPK inhibition,52 oral inhibition of neutrophil elastase53 and other. It is hoped that these efforts may lead to the discovery of effective and safe anti-inflammatory treatments that might prevent the relentless progression of the disease.43

In recent years, a lot of attention has been paid to phenotyping COPD patients, and this may also impact on management strategies and drug development. For instance during exacerbations of COPD, biomarkers can be used to identify specific clinical phenotypes (specifically those associated with bacteria, virus and sputum eosinophilia). It is of interest to note that bacterial and eosinophilic clinical exacerbation phenotypes can be identified already in the stable state.54 For further discussion on phenotypes and disease characterization in COPD, the reader is referred to a recent review article Agusti et al.55

Potential pitfalls of experimental model for COPD

To study pathogenetic mechanisms underlying the development of COPD and cigarette smoke-induced inflammatory responses in the lungs and lymph modes, murine models of COPD have been developed by several groups in the past 15 years.56, 57 As for asthma, the mouse model has proven to be excellent for the study of innate and adaptive immune responses, and predictive for immunologic and inflammatory mechanisms that are also operating in humans. Adaptive immune responses can be initiated and sustained in the regional lymph modes (secondary lymphoid tissue) or alternatively in pulmonary lymphoid follicles (tertiary lymphoid tissue) which have been demonstrated to increase in lungs of COPD patients, along with disease severity.31 On the other hand, the mouse model of COPD does not easily allow to study the classical hallmark of the disease, the (progressive) airway obstruction. Moreover, emphysema does develop, but in a rather limited amount, especially in the most frequently used mouse strain C57Bl6. When using knockout mice, one should realize that the possible therapeutic target is already removed when the disease is being induced; this is very different from the clinical situation where a treatment is started in a patient that has already developed the disease.

Clinical heterogeneity and molecular patterns in pulmonary fibrosis

IPF is a progressive and fatal fibrosing interstitial disease of the lung of unknown aetiology. The lung histological finding of IPF is usual interstitial pneumonia, which is characterized by the presence of honeycombing, fibroblastic foci and areas of fibrosis alternating with normal or less affected parenchyma.58-60 Patients have poor prognosis with a median survival of 2∼5 years.1 No curative therapy is currently available except lung transplantation.61, 62 The clinical phenotype of IPF is characterized by a high inter-individual variability in terms of disease severity at presentation, disease progression and survival time; that is, while some patients remain clinically stable for a long period of time, others have rapid stepwise progression or succumb to frequent bouts of acute exacerbations with high mortality rate.58, 63, 64 The gene expression pattern has been found to differ between patients with distinct clinical variants of the disease. Microarray analysis has shown that patients with clinically accelerated disease and poor prognosis have at the time of diagnosis significant overexpression of genes involved in morphogenesis, oxidative stress, migration and proliferation in the lungs compared with patients with relatively stable disease.65 Evaluation of IPF transcriptome by serial analysis of gene expression could also identify several molecular signatures that can significantly distinguish progressive and relatively stable disease groups.66 Also, a mucin-forming gene associated with high susceptibility of contracting IPF by a wide-genome analysis in the general population was found useful to categorize IPF patients into distinct clinical phenotype groups.67-70 Overall, these studies provide suggestive evidence that the clinical variants of IPF is caused by different patterns of gene expression in the lung parenchyma. The molecular classification of the clinical variants of IPF may lead to the identification and development of novel therapeutic targets and to personalization of the clinical and pharmacological approach to the disease.

Single versus combination therapy in IPF

Abnormalities in various biological pathways including the coagulation system, apoptosis, wound repair, oxidant mechanism, inflammatory response and the immune system have been implicated in the pathogenesis of IPF.71, 72 The current hypothesis is that the disease results from a dysregulated interaction between the lung epithelium and mesenchymal cells.73 Genes and environmental factors that predispose to this aberrant epithelial/mesenchymal interaction have been identified but the mechanism by which they trigger it remains unknown.74-76 The initial belief was that inflammation was the main player in the pathophysiology of the disease but subsequent studies pointed fibrogenesis as the main causative process.71 In this connection, the fibrotic milieu in the lung was found to be rich in molecules capable of enhancing the expression of connective tissue components and promoting the differentiation and/or proliferation of myofibroblasts.72 This new line of evidence has led to a change in therapeutic strategies for pulmonary fibrosis; while early efforts have focused on developing therapies that can control the inflammatory and/or immune response, recent therapeutic approaches focus more on targeting the process of fibrogenesis. Unfortunately, even treatment with specific inhibitors of pro-fibrotic factors provided disappointing results in recent clinical trials despite encouraging good response obtained in pre-clinical studies using animal models of pulmonary fibrosis.77

We learned two important lessons from previous failed clinical trials in IPF: (i) that blockage of a single molecule is not sufficient to stop the complex cascade of pro-fibrotic events, and (ii) that in vivo models currently used for pre-clinical studies are not appropriate. The first assumption is supported by the fact that significant beneficial effects on some clinical parameters have been achieved only in IPF patients treated with compounds known to control multiple pathways such as pirfenidone and nintedanib.78, 79 Pirfenidone is a pyridine analogue small molecule that has shown to have significant effect on the decline of forced vital capacity and progression-free survival in a multicenter, double-blind, placebo-controlled, randomized phase III clinical trial.80 In vitro and in vivo studies have shown that pirfenidone exerts anti-inflammatory activity by inhibiting TNF-α, IL-6, IL-12 and IL-8, anti-fibrotic activity by reducing the expression of collagen and transforming growth factor (TGF)-β1, and anti-oxidant activity.80 A phase II clinical trial has shown that treatment with nintedanib, an intracellular tyrosine kinase inhibitor, was associated with significantly less decline in lung functions and significant reduction in acute exacerbations.79 Nintedanib is a multifunctional inhibitor; it can block signalling mediated by fibroblast growth factor receptor-1, -2 and -3, platelet-derived growth factor receptor-α and β, vascular endothelial growth factor-1, -2 and -3, and Lck, Lyn and Flt-3 tyrosine kinases.81 These observations illustrate the involvement of multiple and interconnected mechanisms in the fibrotic network and suggests the need of combination therapy targeting several pro-fibrotic pathways to achieve significant clinical outcome in IPF.

Experimental protocol to test combination therapy for IPF

Mouse has been the most commonly used laboratory animal for developing experimental models of human disease including pulmonary fibrosis because the mouse genome highly overlaps with that of humans and because it can be easily manipulated.82 Various approaches have been employed for the experimental induction of pulmonary fibrosis in mice including the use of environmental factors (bleomycin, radiation) and gene manipulation.83 Evaluation of the therapeutic effect of potential drugs for IPF has been mostly performed using the bleomycin-induced pulmonary fibrosis model. Unfortunately, translation of drug efficacy established in mouse models induced by bleomycin to clinical application has been so far inconsistent. More than two hundreds compounds were found to be effective against bleomycin-induced pulmonary fibrosis but almost none of them were judged effective in IPF patients.84 There is no explanation for this loss of translation to clinical outcome but the inability of currently used experimental models to recapitulate the biology of IPF appears to be the main reason.

Novel mouse models of pulmonary fibrosis have been recently developed by lung overexpression of cytokines including TGF-β1, TGF-α, TNF-α, IL-1β and IL-13.85-87 Just a few studies have applied these transgenic mice to assess efficacy of potential therapeutic drugs for IPF.85, 88 The use of the newly developed transgenic models may improve the predictive value of preclinical trials because they may provide (i) important hints to identify and tract the signalling pathways and their interaction during the process of fibrogenesis, and (ii) key tools for the experimental design of molecular targeted combination therapy.

Pro-fibrotic cytokines are known to stimulate the synthesis and secretion of matrix proteins (e.g. collagen) directly and independently by binding to specific receptors on the cell membrane, or indirectly by promoting the secretion of fibrogenic cytokines and/or growth factors by autocrine or paracrine mechanisms.89 Mice genetically engineered to overexpress a pro-fibrotic factor may be useful to explore the therapeutic impact of combining a drug that inhibits the direct mechanism with another that blocks the indirect pathway. For example, the currently available mice overexpressing human TGF-β1 can be used for this purpose. TGF-β1 strongly promotes collagen production by activating the TGF-β-receptor-smad2/3 pathway and by stimulating the secretion of other pro-fibrotic factors including platelet-derived growth factor, CCL2 and insulin growth factor-1.72 TGF-β1 transgenic mice can be used to evaluate whether an inhibitor of the TGF-β1-receptor-smad2/3 axis in combination with another inhibitor of cytokine-mediated pathways can have additive effect on outcomes in pulmonary fibrosis. Similar experimental protocols can be designed using other available pro-fibrotic cytokine transgenic mice. The results of these experiments may provide relevant information on the potential clinical benefits of molecular targeted combination therapy and shed some light of hope to improve prognosis in IPF.