Idiopathic Pulmonary Fibrosis: Time to Get Personal?

Despite significant progress in understanding the mechanisms underlying pulmonary fibrosis in general, our understanding of how these mechanisms contribute to individual and patient-specific manifestations of idiopathic pulmonary fibrosis (IPF) remains poor. Within interstitial lung diseases, IPF remains the deadliest of the pneumonitides (1), with a course that is highly variable (2, 3). The iterative nature of diagnosis (4) and the unpredictable course of the disease have made treatment planning extremely difficult. This has complicated the design of drug studies and raised a noticeable debate regarding patient-relevant outcomes.

Until recently, other than lung physiology in such scoring systems as the GAP (gender, age, and physiology) score (5), we did not have replicated molecular and genetic markers that would allow better prediction of outcome and classification of disease. However in the last 2 years, several studies that incorporated derivation and replication cohorts in their design suggested that peripheral blood proteins (6), changes in gene expression (7), or gene variants (8, 9) may be used to better classify patients with IPF as belonging to groups with high risk for early mortality compared with those with significantly more stable disease. Although these studies strongly suggest the notion that molecular and genetic information could be used to personalize management in IPF, they still lack a direct link to such a therapeutic option.

In this issue of the Journal, O’Dwyer and colleagues (pp. 1442–1450) go a step further toward implementing personalized medicine approaches in IPF by identifying a gene variant that is indicative of distinct outcomes in patients with IPF, is mechanistically relevant, and at the same time, may have some very practical therapeutic implications (10). The authors chose to explore a candidate polymorphism in the Toll-like receptor 3 (TLR3) gene, which is known to be important in pathogen recognition and activation of the innate immune response to viruses through recognition of molecular patterns such as the double-stranded RNA. The L412F polymorphism of TLR3 has been previously noted to be functionally important in epithelial cell lines (11). The authors started their investigation by looking at the function of TLR3 in primary human lung fibroblasts from the lungs of patients with IPF with and without the L412F polymorphism. They divided these samples on the basis of presence or absence of the polymorphism and demonstrated that fibroblasts homozygous for the polymorphic allele produce less IL-8 as a readout of nuclear factor-κB activity, and RANTES (regulated upon activation, normal T-cell expressed and secreted) as a readout of interferon regulatory factor 3. IPF fibroblasts that were homozygous or heterozygous for the polymorphic allele also exhibited an impaired IFN-β response and as a consequence exhibited enhanced fibroproliferation after TLR3 stimulation. Impressively, this effect was ameliorated in the presence of recombinant IFN-β.

They also provided a circumstantial evidence that TLR3-mediated signaling was potentially important in fibrosis by comparing the response to bleomycin between TLR3 knockout mice and wild-type control mice. Compared with wild-type mice, TLR3 knockouts exhibited significantly increased mortality, with increased levels of transforming growth factor–β1, IL-13, IL-4, and hydroxyproline.

Finally they proceeded to determine the clinical implications of the polymorphism in two independent cohorts of patients with IPF. The first consisted of a United Kingdom population of 170 subjects, and the second consisted of 138 subjects from the placebo arm of the INSPIRE clinical trial of IFN-γ, where patients were followed for 24 months. They assessed mortality within 12 months in these two groups. The analysis demonstrated a hazard ratio of 4.98 (95% confidence interval, 1.25–20.00; P = 0.023 corrected for age, FVC, and diffusing capacity of carbon monoxide) for patients that carried the minor allele. There seemed to be an additive effect for each allele in that the homozygous minor (TT for CC) carried a hazard ratio of 9.71 (95% confidence interval, 2.70–34.93; P = 0.001) in the United Kingdom cohort. Furthermore, they also demonstrated an association with decline in FVC from baseline over time in the INSPIRE cohort that again seemed to have a dose–response relationship to the polymorphism.

One of the strongest attributes of this manuscript beyond the use of two cohorts is the significant effort to provide a strong mechanistic link between a hypothesis-driven candidate polymorphism and altered fibrosis, the immune response, and survival in patients. It must be mentioned here that the bleomycin results in this case cannot be really considered more than circumstantial evidence for a role of TLR3 in fibrosis, because the mechanisms and downstream effects of bleomycin in TLR3-deficient mice are very, very different then those active in patient carrying a variant in the sequence of TLR3 and exposed to mild environmental injury.

Another strong attribute of this study is that it may and should revive an interest in a therapy previously tried unsuccessfully in IPF. At the International Conference of the American Thoracic Society in 2000, Raghu and colleagues (12) presented the results of a multicenter, randomized double-blind placebo-controlled study and reported no therapeutic benefit, though the study itself was never fully published. The results of the current study, which demonstrate an impaired IFN-β response in cells from patients carrying the rare allele and that IFN-β supplementation blunted aberrant fibroblast activation seen in these patients, suggest potentially revisiting Raghu and colleagues’ results and maybe even retrying IFN-β in patients that carry the polymorphic allele, especially because they are such a high-risk group.

Given the significance, reproducibility, and ease of measurement of the proteins, genes, or genetic variants recently demonstrated to be predictive of outcome in IPF, one must wonder why there is not stronger adoption of the use of molecular and genetic markers in IPF, as it seems only logical to follow the evidence and implement personalized medicine approaches in our referral for lung transplant or decision about drug studies based on molecular and genetic markers. Not only does the article by O’Dwyer and colleagues move us a significant step closer to this goal, it also provides a clue about a potential approach to personalizing therapy in IPF, in which a drug (IFN-β) will only be tried in those patients in whom a positive disease-modifying effect on their fibroblasts exists. Adoption of such approaches has the potential to shorten the duration of drug studies, make them cheaper to execute and most importantly more likely to benefit our patients, a benefit they at last deserve.