Goldilocks and the Three Trials: Clinical Trials Targeting the α_vβ₆ Integrin in Idiopathic Pulmonary Fibrosis

In this issue of the Journal, we report the eagerly awaited phase 2a and 2b trials conducted by Raghu and colleagues (pp. 1128–1139 and 1166–1168) using a humanized monoclonal antibody targeting the α_vβ₆ integrin (1, 2). The α_vβ₆ integrin is a cell surface protein found specifically on epithelial cells and is upregulated after injury with a primary function to activate TGF-β (transforming growth factor β), a key profibrogenic cytokine in the lung and other organs (3). Preclinical studies demonstrated that gene deletion (4), antibody inhibition (5), and small molecular inhibition of the α_vβ₆ integrin (6) protected against the development of pulmonary fibrosis, and data from humans found that high concentrations of the α_vβ₆ integrin were present in the lungs of patients with idiopathic pulmonary fibrosis and were associated with worse outcomes (7). Together these studies suggest that uncontrolled α_vβ₆ integrin-mediated TGF-β activation promotes lung fibrosis and may be “too hot” for normal repair!

TGF-β is a potent profibrotic molecule, but several strategies to globally inhibit it have been limited by toxicity, which may reflect the role of TGF-β in homeostatic and physiological functions (8). The current studies were on the basis of the hope that epithelial cell-specific inhibition of TGF-β activation would limit local and systemic toxicity. Unfortunately, the results of these phase 2 studies suggest systemic administration of an anti-α_vβ₆ monoclonal antibody (BG0001) is associated with a lack of clinical benefit and increased adverse events, including acute exacerbations with fatalities leading the Data Safety Monitoring Board to advise premature termination of the 2b trial. There were no antidrug antibodies detected in any patients; it is therefore unlikely that the adverse effects were off-target effects related to antibody administration and probably reflect an on-target effect of inhibition of the α_vβ₆ integrin.

Given what is known about the α_vβ₆ integrin, can we offer a potential explanation? Several lines of investigation have indicated that α_vβ₆ integrin-mediated TGF-β activation plays an important homeostatic function. Genetic deletion studies demonstrated that completely switching off this pathway can lead to activation of alveolar macrophages (9), increased alveolar inflammation, increased MMP12 generation, and emphysema (10). Therefore, insufficient α_vβ₆ integrin-mediated TGF-β activation leaves the lungs “too cold” and unable to regulate macrophage behavior, which may promote inflammatory responses in the event of a lung injury or infection.

Thus, the precise regulation of α_vβ₆-mediated TGF-β activation is likely critical for its biological outcome. This raises the question of the ideal approach to targeting the integrin. Data from the phase 2a study showed a dose response with an increased risk of acute exacerbations at the highest concentration; the 3.0 mg/kg arm was terminated early in the phase 2a study because of excess acute exacerbations. High concentrations of TGF-β inhibition were observed in BAL macrophages at BG0001 doses over 1.0 mg/kg, and the half-life was 6 days. The phase 2b study was conducted with a dose of 56 mg (approx. 0.7 mg/kg) administered every 10 days on the basis of the data from the phase 2a study. However, it is clear from the exposure–response analysis that there is a narrow therapeutic window for BG0001, and with a long half-life, even a small increase in steady-state could lead to high concentrations of prolonged TGF-β suppression, and with it the risk of acute exacerbations.

Would small molecular inhibitors be more successful and provide inhibition that was “just right”? There is theoretical evidence to support this theory (11), as well as recent clinical data via a press release from Pliant Therapeutics. This release showed data from their phase 2 study in which patients were treated with an increasing concentration of PLN-74809, an oral small molecule, dual inhibitor of α_vβ₆ and alphavbeta1 (a total of 67 patients were treated with 40, 80, or 160 mg once daily for 12 wk). They reported no treatment-related serious adverse drug events and no treatment withdrawals even at the highest concentrations. This is in contrast with BG0001, in which four of six patients withdrew from the highest dose arm, which was ultimately terminated.

Why the difference between these studies? The two obvious differences reflect the chemistry of the drugs and the spectrum of inhibition. PLN-74809 is a small molecule, and BG0001 is a monoclonal antibody. In the absence of human receptor binding studies, it remains a concern that monoclonal antibodies may struggle to penetrate the dense fibrotic matrix associated with pulmonary fibrosis; thus, BG0001 may preferentially inhibit integrins in nonfibrotic regions of the lung and disproportionately affect homeostatic rather than pathological α_vβ₆ integrin function.

PLN-74809 is a dual alphavbeta1 and α_vβ₆ inhibitor and is likely to have a broader range of TGF-β biologic inhibition being able to inhibit both epithelial and fibroblast TGF-β activation (12). However, whereas this might be expected to enhance antifibrotic efficacy, it is, if anything, likely to increase on-target adverse effects, assuming a similar efficacy for inhibiting α_vβ₆-mediated TGF-β compared with BG0001. It is unfortunate that no data are presented relating to TGF-β inhibition observed with PLN-74809. However, on the basis of a published abstract, 160 mg PLN-74809 is likely to have submaximal α_vβ₆ receptor occupancy on the basis of computed tomography–positron emission tomography data (13). It is, therefore, theoretically possible that lower concentrations of epithelial TGF-β inhibition are required for an antifibrotic effect if fibroblast TGF-β inhibition is also inhibited. However, there are no data to confirm or refute this hypothesis.

What is the necessary concentration of epithelial TGF-β inhibition needed to prevent fibrosis, and how much will lead to adverse effects? Although TGF-β inhibition data are shown in the phase 2a study using BG0001, there were insufficient data from the 2b studies to determine if the adverse effects were related to TGF-β inhibition, although it would seem likely given that BG0001 leads to substantial and prolonged inhibition of TGF-β at concentrations of BG0001 greater than 1 mg/kg. Clearly, longer-term studies providing markers of TGF activity such as Smad2 phosphorylation or other biomarkers are needed before reaching conclusions about the effect of integrin inhibition by small molecule inhibitors, but longer phase 2b studies using all the tools available to determine pharmacokinetics and receptor occupancy will be important in determining the merits of therapeutically targeting the α_vβ₆ integrin.

These three studies highlight important considerations in targeting novel molecules in chronic progressive diseases such as idiopathic pulmonary fibrosis. First, it is important to recognize that every mammalian molecule is likely to have a physiological function and that complete suppression of that function will have an adverse consequence, so a detailed understanding of the pharmacokinetics of the drug is crucial before developing the definitive trial. Similarly, having a reliable pharmacodynamic biomarker to determine target engagement and the extent of drug inhibition will be vital. Finally, different methods of targeting the same molecule may have different therapeutic and adverse effects. One unsuccessful approach should not necessarily rule out different strategies to try and satisfy the challenges of biology and the delicate balance Goldilocks dictates!