A Long and Winding Road: Drug Development in Idiopathic Pulmonary Fibrosis

Idiopathic pulmonary fibrosis (IPF) is a relentlessly progressive fibrotic lung disease of unknown origin associated with high morbidity, poor survival, and substantial healthcare use (1). Two drugs, nintedanib and pirfenidone, have been shown to slow the rate of functional decline and disease progression and are approved worldwide. However, the condition of most patients continues to decline despite treatment, and both drugs are associated with substantial tolerability issues. The IPF community is therefore in desperate need of more efficacious and better tolerated therapies.

In recent years, several high-quality clinical trials have been conducted in IPF. These studies have provided important insights into trial design and choice of endpoints. Despite high hopes of success, three promising potential IPF therapies, ziritaxestat, pentraxin-2, and pamrevlumab, have all recently failed. This has led to considerable disappointment and frustration.

All three drugs had a substantial preclinical evidence base supporting their potential as antifibrotic therapies. Ziritaxestat is a small-molecule, selective inhibitor of autotaxin, an enzyme responsible for the production of lysophosphatidic acid, a mediator, which drives fibroblast recruitment and vascular leak (2). Pentraxin-2 is a plasma protein that regulates responses to tissue damage and fibrosis (3). Pamrevlumab is a fully human monoclonal antibody against connective tissue growth factor, a secreted glycoprotein produced by fibroblasts, myofibroblasts, and endothelial cells, which modulates several biological activities associated with tissue fibrosis (4).

These failures highlight a number of issues, including our limited understanding of the systems biology of pulmonary fibrosis, the inefficiency of current IPF trial design and endpoints, and the heterogeneity of IPF trial populations (a problem that has increased further with the inclusion of patients in trials who are taking pirfenidone or nintedanib). With this in mind, how can the performance of clinical trials in IPF be improved?

IPF results from an aberrant wound-healing response after recurrent alveolar epithelial microinjury in genetically susceptible individuals. An intricate network of profibrotic pathways, with feedforward activation, is believed to be involved in disease pathogenesis. At the same time, pathways that play a role in resolution of fibrosis are downregulated (5). The process of fibrosis is biologically complex and involves multiple simultaneously activated pathways with considerable inbuilt redundancy. As such, a disease such as IPF is unlikely to be effectively targeted by any single compound alone unless that drug is highly pleiotropic in its mechanism of action. Such pleiotropism likely accounts for the (limited yet clinically relevant) efficacy of pirfenidone and nintedanib. Alternative approaches to identifying pleiotropic molecules are to combine therapies, which target specific complementary pathways, or to better identify individuals in whom the specific pathway is active in driving disease.

Development of More Reliable Disease Models

The murine model of bleomycin-induced fibrosis is the most widely used preclinical model for testing potential IPF therapies (6). However, this model fails to recapitulate a number of important aspects of IPF biology. The bleomycin model is triggered by an acute lung injury, the fibrosis is nonprogressive and partially resolves over time, and histologically the fibrotic lesion is temporally homogeneous and does not reproduce the salient features of usual interstitial pneumonia. Accordingly, its utility in predicting the clinical efficacy of candidate therapeutics in humans is limited. Alternative animal models exist, but these exhibit similar deficiencies and are probably no better in discriminating the potential efficacy of pharmacological agents (6). Induced pluripotent stem cells are able to reproduce the molecular characteristics of a number of diseases and are increasingly used for predicting drug safety and efficacy. Recently, Suezawa and colleagues developed a model of pulmonary fibrosis in vitro using bleomycin-exposed surfactant protein C–positive cell-derived fibroblast-dependent alveolar organoids (7). The fibroblast-dependent alveolar organoids recapitulate in vitro several important mechanisms that are observed in human IPF and demonstrate potential for screening novel therapeutic agents for IPF.

Pathway-driven Therapy

Historically, drugs for IPF have been developed on the basis of the “one disease, one drug” dogma with little consideration given to the mechanisms underlying disease pathophysiology. Although this approach has often proved successful in monogenic/monotarget diseases, it falls short in complex/multifactorial diseases such as IPF. One strategy for improving the performance of clinical trials in IPF is to move away from the “one disease, one drug” approach and instead embrace the principle of causative molecular mechanisms, generally referred to as “endotypes.” Identification of endotypes has the potential to explain disease heterogeneity, thus enabling treatment tailored to the individual patient. Recently, Kraven and colleagues, by applying machine learning to multiple gene expression datasets, identified three clusters of patients with IPF with distinct clinical features and survival (8). Patients in different clusters demonstrated activation of different underlying biological pathways, including immune system response and metabolic changes that could potentially be modulated therapeutically (8). Precision medicine trials will be necessary to ensure that potential beneficial effects of drugs in specific endotypes are not overlooked because of heterogeneous effects in an unselected population.

Patient Stratification by Genetic Background

Genetic polymorphisms play an important role in the development of pulmonary fibrosis. It is increasingly clear that specific genetic defects, such as those associated with maintenance of telomeres, are associated with distinct clinical characteristics (9). The use of genomic signatures as an inclusion criterion may improve trial performance through enrichment for individuals at highest risk of progression. Alternatively, genetic polymorphisms may influence response to specific therapies or could form the target for precision therapy. No completed trials to date have stratified patients on the basis of genetic profile. However, post hoc analysis of the PANTHER trial (Prednisone, Azathioprine, and N-Acetylcysteine: A Study that Evaluates Response in Idiopathic Pulmonary Fibrosis) suggested an interaction between TOLLIP rs3750920 and response to N-acetylcysteine (10). This observation has led to the first randomized, placebo-controlled precision medicine study to be performed in individuals with IPF (NCT04300920) (11). Additional critical aspects to consider include novel approaches to trial design (i.e., adaptive trials, external control arms, and Bayesian analyses [12]) and selection of endpoints that explore the effect of novel therapies on a patient’s symptoms and quality of life (13).

The road to drug development in IPF has proved long and winding. Better understanding of disease pathobiology, strategies to secure the enrollment of homogeneous populations of patients in clinical trials, and innovative approaches to trial design are instrumental to developing truly efficacious and better tolerated drugs for this devastating disease.