Lysophosphatidic acid receptor 1 inhibition: a potential treatment target for pulmonary fibrosis

ATX

ATX is an enzyme with lysophospholipase D activity and is the predominant enzyme that converts lysophosphatidylcholine (LPC) to LPA in the plasma [26, 27]. LPA is a bioactive phospholipid that mediates a host of physiologic processes (e.g. cell migration, proliferation and differentiation, cytoskeleton regulation, macrophage survival, inflammation, and alteration of cell–cell adhesions) and signals through a six-member family of G-coupled protein receptors, LPAR1–6 [21, 28]. Because ATX largely contributes to LPA availability, ATX activity is associated with outcomes of LPA signalling independent and dependent of LPAR, including LPAR-mediated fibrosis and inflammation [9, 11, 21, 28–30]. In ATX-dependent LPA synthesis, phospholipase A-1 or A-2 removes a fatty acid from the sn-1 or sn-2 position, respectively, from membrane phospholipids resulting in LPC, which is subsequentially hydrolysed by the lysophospholipase D activity of ATX to produce LPA [27, 28]. However, ATX is not the sole LPA producer, as LPA is also produced via ATX-independent mechanisms, albeit to a lesser extent. At least four intracellular ATX-independent pathways are capable of LPA production [28].

A recent study encompassing biochemical, cellular, and murine assays suggests that ATX also acts as an LPA chaperone to facilitate delivery of LPA to LPAR, independent of its catalytic activity [31]. This study revealed that ATX bound to LPA favours activation of LPAR6 more than LPAR1. Although in vitro and in vivo studies support the notion that LPA species and ATX-mediated chaperoning may influence LPAR isoform binding preference, questions remain regarding how ATX/LPA/LPAR signalling specificity affects biologic functions.

LPAR1 pathway

LPAR1 belongs to a six-member family of LPARs, each with numerous and sometimes redundant physiologic effects, including cell migration, proliferation and differentiation, cytoskeleton regulation, inflammation, and alteration of cell–cell adhesions, all of which are processes that contribute to fibrosis [8, 28]. LPA, the primary ligand of LPAR1, contains a glycerol backbone with a fatty acid chain at the sn-1 or sn-2 and a phosphate group at the sn-3 position [28]. Species of LPA are differentiated by the degree of saturation and length of the fatty acid chain and have varying affinities for LPA receptor isoforms [28]. LPAR1 has broad LPA species selectivity and binds both saturated and unsaturated LPA [32].

LPAR1, expressed by macrophages and fibroblasts amongst others cell types, promotes inflammation and fibrosis in various tissues (e.g. kidney, skin and lung) [4, 11, 12, 33–35]. After tissue injury (e.g., bleomycin exposure) or during blood clotting, LPA levels increase locally to activate LPAR1 and promote a profibrotic state [4, 36]. The LPAR1-associated G-protein (Gα_i, Gα_q or Gα_12/13) is activated upon binding of LPA to LPAR1 and this activation is associated with the aforementioned cellular events that contribute to fibrosis [3, 12, 37–42]. Preclinical evidence also suggests that LPAR1 indirectly increases ECM deposition through activation of fibroblast migration [4]. Additionally, LPAR1 increases activity of key inflammatory players such as NF-κB and c-jun N-terminal kinase; it also increases expression of chemoattractants and proinflammatory cytokines (e.g. interleukin-6 (IL-6), IL-8, chemokine C-X-C motif ligand 1 (CXCL1), CXCL8, C-C motif chemokine ligand 2 (CCL2) and CCL3) [12, 20, 33, 34, 43, 44].

IPF

IPF is characterised by progressive loss of lung function, dyspnoea and poor prognosis [1]. Approximately 70% of patients with IPF are males, with disease presentation usually occurring at an age of >60 years [16]. Risk factors for IPF include cigarette smoking, air pollution, other environmental exposures, genetics and older age [16]. An analysis of IPF survival in the literature estimates a median IPF survival of 3.2 years, with the majority of studies included in the analysis reporting median survival rates of 2–5 years [45]. Real-world data, although inconsistent, have demonstrated that approved IPF therapies (i.e. pirfenidone and nintedanib) are associated with improvements in survival [46–49]. For example, one study of 457 patients reported transplant-free survival rates of 3.4 versus 2.2 years with approved treatment versus without, respectively (p=0.005) [46]. However, although current pharmacologic therapies slow the decline of lung function in IPF, they do not halt disease progression and may have poor tolerability (table 1) [50, 51]. Further, the failure and termination of recent phase 3 clinical trials (Zephyrus-1 and ISABELA-1/2) that evaluated promising antifibrotic molecules (pamrevlumab (anti–connective tissue growth factor) and ziritaxestat (ATX inhibitor)) highlights challenges in drug development for IPF [24, 52, 53]. Thus, there remains an urgent need for better treatments for IPF with several phase 2 and 3 trials currently ongoing [54–59].

Preclinical and clinical studies have implicated ATX and LPA/LPAR1 signalling in IPF pathogenesis (figure 1). In one study, increased protein levels of ATX were observed in lung tissue and alveolar macrophages of the fibrotic interstitium from patients with IPF and in bronchoalveolar lavage fluid (BALF) from mice with bleomycin-induced lung fibrosis. Moreover, conditional ATX knockout in bronchiolar epithelial cells and macrophages protected mice from bleomycin-induced lung fibrosis [60]. Another study in a mouse model of bleomycin-induced lung fibrosis found that ATX protein levels and activity increased in BALF and lung homogenates (but not plasma) after bleomycin challenge and that the increase was dependent on vascular leak rather than increased mRNA expression of ATX [29]. Therefore, lung LPA production via ATX may be primarily dependent on ATX localisation via vascular leak and not local production of ATX. Collectively, these data suggest that ATX may contribute to pulmonary fibrosis and inflammation.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Model of autotaxin (ATX) and lysophosphatidic acid receptor (LPAR) 1 in systemic sclerosis and pulmonary fibrosis. Model is derived from empirical data collected from preclinical dermal and pulmonary fibrotic models. After injury, local levels of ATX rise. In the extracellular space, ATX catalytically converts lysophospholipids such as lysophosphatidylcholine (LPC) to lysophosphatidic acid (LPA), thus elevating local levels of LPA. LPA activates the LPAR family of G protein-coupled receptors. Additionally, ATX can chaperone LPA to LPARs, with heightened specificity for LPAR6 compared with LPAR1. Activation of LPAR1/2 and their G protein-coupled receptors leads to downstream signalling that results in cellular processes involved in fibrosis. Bleomycin-induced dermal fibrotic mouse models support that LPAR1, but not LPAR2, contributes to cellular events that culminate in dermal fibrosis. Bleomycin-induced pulmonary fibrotic mouse models support that LPAR1 and LPAR2 contribute to cellular events that culminate in pulmonary fibrosis. The effect of LPAR3–6 is not as well characterised in the development of dermal and pulmonary fibrosis.

Regarding LPAR1 specifically, in a study of mice with bleomycin-induced pulmonary fibrosis, LPAR1 deficiency reduced fibroblast chemotaxis, vascular leakage and mortality [4]. Although LPAR1 deficiency did not affect fibroblast to myofibroblast differentiation, LPAR1-deficient mice exhibited reduced collagen accumulation following bleomycin injury, suggesting that LPAR1 indirectly drives collagen accumulation through fibroblast chemotaxis [4]. In another study of bleomycin-induced pulmonary fibrosis, LPAR1 knockout mice exhibited less bronchial epithelial cell apoptosis than LPAR1 wild-type mice. Additionally, serum-deprived primary mouse lung fibroblasts treated with LPA were more resistant to apoptosis than were untreated cells and inhibitors targeting LPAR1 attenuated this apoptotic resistance [3]. Collectively, these in vivo and in vitro observations suggest that LPAR1 contributes to vascular leakage, epithelial cell apoptosis and fibroblast chemotaxis and survival, all of which are contributory mechanisms of pulmonary fibrosis [3, 4] Further, LPAR1 mRNA expression is the highest among the LPAR family in pulmonary fibroblasts, supporting the importance of this isoform in LPA signalling outcomes in pulmonary fibrosis [4]. Clinically, elevated levels of LPA have been observed in BALF and plasma from patients with IPF [4, 12, 61]. Further, higher levels of plasma LPA associated with greater fibrosis in the lower lung and decline in lung function as measured by diffusing capacity of carbon monoxide (D_LCO) [61]. These observations suggest that LPA/LPAR signalling is also relevant for human pathology.

In a study using the bleomycin-induced pulmonary fibrosis mouse model, LPAR1 deficiency did not attenuate the inflammatory leukocyte response or leukocyte activation in BALF [4]. However, in humans with IPF, serum LPA levels positively correlate with inflammation-related biomarkers (i.e. CCL17 and CCL18) that also have profibrotic properties [61]. Thus, LPA may contribute to pathogenic fibrosis and inflammation within IPF.

SSc and SSc-associated ILD

SSc is a chronic connective tissue disease characterised by autoimmunity, vasculopathy and fibrosis of the skin and internal organs [5]. Categorised into two clinical subsets on the basis of cutaneous involvement (limited cutaneous SSc (lcSSc) and diffuse cutaneous SSc (dcSSc)), SSc is a heterogeneous disease and no therapeutic regimen has been found to be effective across various SSc phenotypes [5]. Patients with dcSSc experience profound skin fibrosis that is proximal as well as distal to the elbows/knees and generally progresses at a more rapid rate than those with lcSSc. However, both cutaneous subtypes are associated with serious internal organ involvement [62, 63]. Further, patients with dcSSc have a poor prognosis and low survival rate, estimated at 88 and 74% at 5 and 10 years, respectively [64].

Positive clinical trials of therapies for SSc and SSc-ILD are reviewed in table 1. Although these therapies may slow disease progression and improve symptoms and certain organ manifestations, treatments are not generally curative and do not arrest fibrosis [5, 17, 65]. Further, current treatments are ineffective in comprehensively treating the broad spectrum of SSc clinical manifestations. For example, the IL-6 receptor antagonist tocilizumab and the antifibrotic, nintedanib, are both approved for slowing the rate of pulmonary function decline in adults with SSc-ILD; however, neither was associated with improvement in skin fibrosis compared with placebo [66–69]. To fulfil the unmet need of the current treatment landscape, novel agents that target underlying mechanisms of both inflammation and fibrosis are being developed as potential therapies for SSc (e.g. ATX and LPAR1 inhibitors).

Pulmonary involvement is a potentially severe SSc manifestation that may manifest as ILD and is associated with high rates of morbidity and mortality [1, 70, 71]. An estimated prevalence of ILD in patients with SSc (n=1168) from the Canadian Scleroderma Research Group registry was 52% (95% CI 46–59%) [71]. Although ILD is common in patients with dcSSc, ILD occurs in patients with lcSSc, as well (dcSSc: 70%; lcSSc: 39% (Registry of the Spanish Network for SSc)) [72]. These estimates are possibly underestimated due to lack of uniform screening for early SSc for ILD [73]. The prevalence of progressive pulmonary fibrosis (PPF) among patients with SSc-ILD varies by patient population, but studies report that approximately 27–39% of SSc-ILD cases can be classified as PPF [74, 75].

As previously described, two studies of bleomycin-challenged pulmonary fibrotic mice have demonstrated an association between ATX and pulmonary fibrosis [29, 60]. Because bleomycin mouse models are used for both SSc-ILD and IPF, these studies suggest a link between ATX and SSc-ILD, while another study demonstrated that ATX is additionally involved in dermal fibrosis. In a bleomycin-induced dermal fibrosis mouse model, dermal ATX mRNA and protein levels were enhanced compared with saline-injected mice [76]. Inhibition of ATX reduced collagen and myofibroblast accumulation in the dermis as well as dermal thickening after bleomycin challenge [76]. This involvement of ATX in dermal fibrosis is also observed in humans. Messenger RNA ATX levels are elevated in skin from patients with dcSSc compared with healthy skin [76]. Further, ATX, LPA and IL-6 participate in an amplification loop in dermal fibroblasts, wherein LPA produced by ATX increases IL-6 expression and IL-6, in turn, enhances ATX expression [76]. This feedback mechanism is more pronounced in SSc versus healthy fibroblasts in vitro and implicates ATX in fibrotic-related inflammation [76]. Therefore, ATX may contribute to both dermal and pulmonary fibrosis and may thus be a mediator in SSc and SSc-ILD pathogenesis (figure 1).

Preclinical and clinical evidence also support LPAR1 as a contributor to SSc. Two studies involving bleomycin-induced pulmonary fibrotic mice with LPAR1 deficiency, one of which also involved primary mouse lung fibroblasts treated with LPAR1 inhibitors, concluded that LPAR1 contributes to physiologic mechanisms associated with pulmonary fibrosis and thus SSc-ILD (reviewed in the IPF section) [3, 4]. Mice with bleomycin-induced dermal fibrosis and LPAR1 knockout have reduced dermal thickening as well as collagen and myofibroblast accumulation compared with wild-type bleomycin-induced dermal fibrotic mice. Skin fibroblasts from patients with SSc had elevated LPA-activated CI⁻ current activity, a phenomenon involved in lung fibroblast differentiation [77]. Dermal fibroblasts and skin biopsies from patients with SSc contain more LPAR1 mRNA than mRNA encoding other LPAR isoforms [12]. Inhibition of LPAR1 in SSc dermal fibroblasts led to decreased CCL2, CXCL1 and IL-6 protein secretion in response to LPA, supporting a proinflammatory role of LPAR1 [12]. Moreover, patients with SSc have elevated total serum LPA:lypophosphatidylcholine ratios compared with healthy controls [78]. Together, these observations associate LPAR1 activity with SSc.

In addition to its fibrotic and immunomodulatory effects, LPAR1 activation exerts vascular effects such as intimal hyperplasia in response to vascular injury due to interactions with smooth muscle and vasoconstriction [79, 80]. Therefore, LPAR1 activity may also contribute to SSc-related vasculopathy, which may be improved with LPAR1 inhibition. In summary, evidence of ATX and LPAR1 involvement in fibrotic disease and potential involvement in vasculopathy suggests that modulation of ATX or LPAR1 may attenuate fibrosis and that this antagonism may be an effective strategy in the development of therapies for IPF and SSc.

ATX inhibition

The contribution of ATX in pathological fibrosis has motivated the design of multiple ATX inhibitors that have been reviewed extensively elsewhere [81, 82]. Although ATX inhibition may decrease activation of LPAR1 (via LPA depletion), because LPA engages LPARs 1–6, ATX inhibition is not specific to LPAR1 and may indirectly inhibit LPARs 2–6, as well (figure 2) [28]. Several ATX inhibitors are under preclinical development and three ATX inhibitors have entered clinical trials for IPF, namely BLD-0409 (cudetaxestat) [55], GLPG1690 (ziritaxestat) [83] and BBT-877 [84, 85]. In mice with bleomycin-induced lung fibrosis, all three inhibitors reduced the fibrosis versus vehicle [85–87].

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Theoretical mechanism of action of a) autotaxin (ATX) and b) selective lysophosphatidic acid receptor (LPAR) 1 inhibitors. In panel a), inhibition of ATX leads to decreased LPA and thus decreased activation of all LPARs. Decreased LPAR1 signalling leads to decreased dermal and lung fibrosis, while decreased LPAR2 signalling leads to decreased pulmonary fibrosis. Inhibition of LPAR3–6 is not as well characterised because data for LPAR3–6 signalling in dermal and pulmonary fibrotic models are lacking. Inhibition of ATX chaperone function may lead to a more pronounced inactivation of LPAR6 and thus decreased pulmonary fibrosis. In panel b), during LPAR1 selective inhibition, upstream signalling and activation of LPAR2–6 is maintained. Inhibition of LPAR1 leads to decreased cellular events that improve dermal and pulmonary fibrosis. Because LPAR1 is expressed at higher levels compared with other LPARs in pulmonary fibroblasts and systemic sclerosis dermal fibroblasts, this selective inhibition targets cell types involved in dermal and pulmonary fibrosis. LPA: lysophosphatidic acid; LPC: lysophosphatidylcholine.

In bleomycin-challenged mice, cudetaxestat reduced expression of ATX/LPA target genes in RNAseq experiments, plasma LPA levels, mRNA expression of key drivers of lung fibrosis and eventually lung fibrosis in a dose-dependent manner [86]. The safety of cudetaxestat was evaluated in four phase 1 clinical trials of healthy volunteers (n>200) [88]. A single/multiple ascending dose study of oral cudetaxestat solution reported only transient, mild gastrointestinal treatment-emergent adverse events (TEAEs), while a bioavailability study of oral cudetaxestat tablets reported no gastrointestinal TEAEs. A drug–drug interaction study reported that cudetaxestat is safe to use in combination with nintedanib or pirfenidone with no significant changes in drug exposure during concomitant use. No serious adverse events (AEs) were reported in the phase I studies [88].

Ziritaxestat inhibits ATX catalytic and chaperone function and has higher selectivity for LPAR6 than for LPAR1 [31]. Ziritaxestat was shown to be more efficient at protecting mice from radiation-induced pulmonary fibrosis than was an ATX catalytic inhibitor that did not inhibit chaperone activity, supporting the premise that ATX chaperone inhibition may contribute to pulmonary fibrosis protection [31]. Ziritaxestat also reduced LPA 18:2 levels in BALF of bleomycin-challenged mice [85]. In the first in-human study of ziritaxestat in healthy men, headache was the only reported TEAE and no serious AEs were reported [89].

BBT-877 reduced LPA 18:2 and 20:4 with higher potency than GLP1690 in ex vivo enzymatic assays that used human plasma. BBT-877 also reduced body weight loss, lung weight, Ashcroft score and collagen content in bleomycin-challenged mice compared with the vehicle. A phase I trial of BBT-877 in 80 healthy volunteers resulted in only mild AEs and no serious AEs [87].

LPAR1 inhibition

Although ATX and LPAR1 inhibitors both reduce fibrosis via inhibition of LPA signalling, their MOA vary considerably. Unlike ATX inhibitors, which deplete LPA production, LPAR1 inhibitors bind LPAR1 and specifically inhibit its activation by LPAs [12, 90]. Hence, LPAR1 inhibitors may have no or a limited effect on LPARs 2–6, depending on binding specificity for LPAR1 within the LPAR family (figure 2) [12]. Examples of LPAR1 inhibitors are fipaxalparant (HZN-825), BMS-986020 and BMS-986278 [23, 54, 91]. Although these inhibitors target the same receptor, in vitro and in vivo evidence suggests that their binding mechanisms differ, which may lead to differences in efficacy and safety profiles in the clinic [90, 92].

Fipaxalparant behaves as a selective allosteric LPAR1 inhibitor [12, 90]. In vitro, fipaxalparant inhibited LPAR1-mediated activation by five different LPA species, including 20:4 LPA, which is increased in serum of patients with SSc [12, 78]. In cellular assays, fipaxalparant blocked calcium response in Chinese hamster ovary cells treated with serum from patients with SSc, prevented LPA-mediated differentiation of IPF fibroblasts, decreased Wnt signalling and decreased secretion of inflammatory markers (IL-6, CCL2 and CXCL1) in SSc dermal fibroblasts [12]. Additionally, in Tsk1 mice, fipaxalparant reduced hypodermal thickening, myofibroblast accumulation, hydroxyproline content and secretion of CCL2 and CXCL1 [12]. Notably, fipaxalparant was inactive on LPAR2, 3 and 5 and had some activity with LPAR6 [12]. Thus, fipaxalparant exerts its effect primarily on LPAR1, but it has limited effect on other LPARs, whose contributions to dermal and pulmonary fibrosis are not as well established [3, 4, 12, 19]. For example, evidence from mice subcutaneously injected with bleomycin indicates that LPAR2 does not contribute to dermal fibrosis, while evidence from a bleomycin-induced fibrotic mouse model supports that LPAR2 does contribute to pulmonary fibrosis [19, 93]. Together, these data suggest that fipaxalparant selectively inhibits LPAR1 and attenuates fibrotic processes in preclinical models. Five phase 1 safety trials of fipaxalparant in healthy volunteers (n=94) reported no severe AEs. The most frequently reported TEAEs were headache, orthostatic hypotension, postural dizziness, flatulence and abdominal pain (unpublished data, Horizon Theraputics).

BMS-986020 and BMS-986278 are structurally related yet distinct inhibitors of LPAR1 [92, 94]. An in vitro study suggested that BMS-986020 binds LPAR1 differently than fipaxalparant [90]. BMS-986020 also inhibits hepatic resident transporters, particularly bile salt export pump (BSEP) and mitochondrial function in hepatic cells and, therefore, has off-target effects that likely caused the hepatobiliary toxicity observed in a phase 2 clinical trial [23, 92]. Conversely, BMS-986278 inhibition of BSEP was weak in in vitro assays, suggesting that BMS-986278 may have a better safety profile than BMS-986020 [92]. BMS-986278 did not cause hepatobiliary toxicity in rat or monkey models [95]. In addition, a phase 1 trial of BMS-986278 in healthy individuals did not report hepatic toxicity. Most AEs were mild and the most frequent one was headache. Decreased blood pressure was also reported (ClinicalTrials.gov identifier: NCT03429933) [96, 97].

ATX inhibitors

Three ATX inhibitors have entered phase 2 of clinical development: BBT-877, cudetaxestat and ziritaxestat (table 2) [55, 83, 84]. ISABELA 1 and 2 were phase 3, double-blind, placebo-controlled, global, randomised clinical trials that evaluated the safety and efficacy of ziritaxestat in participants with IPF (n=525; n=781) [24]. Participants were randomised 1:1:1 to receive ziritaxestat 600 mg, ziritaxestat 200 mg or placebo for at least 52 weeks. Participants were permitted, but not required, to receive standard-of-care treatment (nintedanib, pirfenidone) during the study. Both trials were terminated early after a review by an independent data safety monitoring committee that concluded there was increased mortality in the 600 mg ziritaxestat group and lack of efficacy in all treatment groups [24]. Despite early termination, most patients were included in the analysis (1300/1306) and mean (sd) total treatment duration for ISABELA-1 and ISABELA-2 were 345 (188) and 339 (161) days, respectively. Ziritaxestat treatment did not result in benefits in the primary end-point (annual rate of forced vital capacity (FVC) decline versus placebo) or in key secondary end-points versus the placebo group. Outcomes for time to first respiratory-related hospitalisation were worse in the ziritaxestat groups versus placebo. Pooled all-cause mortality rates for ziritaxestat 600 mg, ziritaxestat 200 mg and placebo were 8.9, 7.0 and 5.5%, respectively, and respiratory-specific mortality rates were 3.6, 3.5 and 1.8%, respectively. The most common TEAEs were gastrointestinal disorders. Serious AEs were reported similarly across treatment groups (ISABELA 1: 22, 22 and 21%; ISABELA 2: 25, 24 and 16%; for 600 mg, 200 mg and placebo, respectively). Of participants treated with ziritaxestat, 74% had reduced LPA C18:2 plasma levels and 26% had increased LPA C18:2 levels, indicative of target engagement by ziritaxestat in a portion of participants [24]. Clinical development of ziritaxestat was stopped after termination of the ISABELA trials [83].

NOVESA was a randomised, double-blind, placebo-controlled phase 2A clinical trial for ziritaxestat in participants with dcSSc (n=33). This trial included patients with early dcSSc (first manifestation of SSc other than Raynaud's phenomenon within the last 5 years) and excluded patients with FVC ≤45%. Most participants (90%) were receiving background immunosuppressant standard-of-care therapy. The NOVESA trial met its primary end-point of reduced mean modified Rodnan skin thickness score (mRSS) at week 24 (p=0.04; ziritaxestat versus placebo groups), with no significant differences in the change from baseline in FVC between groups. Diarrhoea and headaches were the most frequently reported TEAEs in the ziritaxestat group and 9.5 and 8.3% of patients experienced serious AEs in the ziritaxestat and placebo groups, respectively. In ziritaxestat-treated participants, circulating LPA C18:2 was significantly reduced (p<0.0001 versus placebo group), demonstrating ATX inhibition. Multiple fibrosis biomarkers in the blood were stabilised in the ziritaxestat group and increased in the placebo group (between-group difference (p<0.04)), while no differential gene expression was observed in the skin of ziritaxestat versus placebo groups. Further, membrane-spanning 4-domains A4A RNA, a biomarker of M2 macrophages, was significantly reduced 1.9-fold in blood samples from patients treated with ziritaxestat compared with placebo (p=0.03) [83].

31 of the 33 participants enrolled in the NOVESA trial entered the 104-week open-label extension study. However, premature termination of the NOVESA study following results from ISABELA trials precluded analysis of the NOVESA open-label extension beyond week 52. At week 52 of the open-label extension, the mean (standard error) change from baseline in mRSS was −11.6 (3.0) and −12.2 (1.6) units in the ziritaxestat–ziritaxestat and placebo–ziritaxestat groups, respectively. All participants experienced at least one TEAE and serious AEs were reported in 29% of all participants [83].

LPAR1 inhibitors

The clinical profile of fipaxalparant is being investigated in two ongoing phase 2 trials in individuals with IPF and dcSSc (NCT05032066 and NCT04781543) (table 2). The safety and tolerability of fipaxalparant were demonstrated in 32 participants with dcSSc in a phase 2a study wherein participants were treated with fipaxalparant or placebo for 8 weeks, followed by a 16-week extension during which all participants received treatment. Most patients (24/32) were receiving background immunosuppressive therapy. Headache was the most commonly reported TEAE in the fipaxalparant-treated group and one (6.7%) participant in the fipaxalparant-treated group experienced a serious AE [22].

At the end of the extension period, participants treated with fipaxalparant for the entire 24-week study duration experienced a clinically meaningful decrease in mean total mRSS from baseline (mean: −7.36; median: −7.5) [98]. Participants treated with placebo for 8 weeks and fipaxalparant for 16 weeks also experienced a clinically meaningful decrease in mRSS from baseline (mean: −7.31; median: −7.00) [22, 99]. Clinically meaningful improvements in Health Assessment Questionnaire Disability Index were observed in fipaxalparant-treated participants at week 8 (−0.14) and week 24 (−0.15) and in placebo/fipaxalparant–treated participants at week 24 (−0.23) [22]. The most commonly reported TEAEs were infection and headache. Two (6.7%) serious AEs were reported, one in each treatment group.

Once-daily and twice-daily BMS-986020 were evaluated in a phase 2, multicentre, three-group, randomised, double-blind, placebo-controlled clinical trial in individuals with IPF. Twice-daily BMS-986020 significantly decreased FVC decline compared with placebo at week 26 (p=0.49). BMS-986020 caused hepatobiliary toxicity in participants with IPF, inciting early termination of the study after three cases of cholecystitis occurred. Of the 143 participants, 108 completed the 26-week dosing phase and 35 participants withdrew from the study, 13 because of AEs [23].

A phase 2, randomised, double-blind, placebo-controlled clinical trial evaluated BMS-986278 in patients with IPF (n=276; ClinicalTrials.gov identifier: NCT04308681). Participants were permitted background therapy (pirfenidone or nintedanib) and/or immunosuppressive therapy if they were receiving a stable dose before screening [91]. The primary end-point was rate of change in FVC% over 26 weeks. Patients randomised to receive placebo, 30 mg or 60 mg of BMS-986278 twice daily exhibited a mean rate of change in FVC of −2.7, −2.8 and −1.2%, respectively. Thus, the mean FVC% difference of 60 mg twice daily compared with placebo was +1.4%. A subgroup analysis concluded that 60 mg of BMS-986278 had an effect regardless of background antifibrotic therapy use. Serious AEs occurred in 17, 11 and 11% of patients receiving placebo, 30 mg and 60 mg, respectively. The most frequently reported AEs were diarrhoea, cough and orthostatic hypotension [25]. A parallel cohort including 123 patients with PPF (defined as having fibrotic ILD with prior progression within 2 years) exhibited a mean rate of change in FVC of −4.3, −2.7 and −1.1%, for those treated with placebo, 30 mg and 60 mg, respectively. The mean FVC difference between the BMS-986278 and placebo groups was 1.6% for 30 mg and 3.2% for 60 mg. Treatment differences in the PPF cohort were independent of background antifibrotics and usual interstitial pneumonia pattern. AEs occurred in 78, 83 and 67% of the placebo, 30 mg and 60 mg groups, respectively [100]. Findings from both the IPF and PPF cohorts of this phase 2 trial support progression to phase 3 [25, 100].

Clinical trial design challenges

Background therapy has been inconsistent throughout IPF/ILD trials and may affect trial outcomes. ISABELA-1/2 were the first phase 3 trials in IPF to allow pirfenidone (n=454/1281) and nintedanib (n=444/1281) background therapy [24, 101]. A drug–drug interaction between ziritaxestat and nintedanib caused elevated plasma levels of nintedanib and may have contributed to high rates of AEs (e.g. diarrhoea and nausea overall; nondiarrhoeal TEAEs (600 mg ziritaxestat arm)) [24]. Moreover, greater rates of annual FVC decline were observed in the placebo arm for patients with background therapy versus without (least squares mean, nintedanib: −163.1 mL; pirfenidone: −189.0 mL; neither: −149.1 mL) [24]. Comparatively, nintedanib and pirfenidone pivotal trials reported a greater decline for placebo-treated patients (adjusted annual rate: −207 and −240 mL; linear rate at week 52: −280 mL, respectively) [50, 51], suggesting an enrolment bias for patients with poor response to antifibrotic therapy or with less progressive disease that has also occurred in smaller IPF trials [101, 102]. Prior to nintedanib and pirfenidone trials, low rates of FVC decline in the placebo arm have proved challenging (e.g. CAPACITY 006 failure) [103]. Future trial designs should consider novel strategies to overcome these challenges, such as leveraging precision medicine or biomarkers (e.g. phase 3 PRECISIONS trial) [104] to increase the likelihood of success.

Despite permitted background antifibrotic use, a 26-week phase 2 trial of BMS-986278 in patients with IPF (n=276) and ILD (n=125) yielded positive results [25, 100]. Comparatively, the phase 2a study of ziritaxestat in IPF was smaller (n=23) and shorter (12 weeks); rapid progression into ISABELA-1/2 had uncertain risks because drug–drug interactions may not have been fully characterised [105]. Therefore, the BMS-986278 trial provides optimism for its phase 3 trial despite persisting challenges.

MOA: ATX versus LPAR1 inhibitors

The failure of the ISABELA-1/2 phase 3 trials stand in contrast to the promising results of the BMS-986278 phase 2 trial, suggesting an association between MOA and clinical outcomes. Increased levels of plasma ATX were detected in participants in ISABELA-1/2, indicative of disease progression or a regulatory feedback loop that has previously been observed in mice treated with an ATX inhibitor [106, 107]. Raised ATX levels suggest that the broad effect of ATX inhibition (e.g. cellular responses via ATX substrate accumulation and LPAR1–6 inhibition) may cause unwanted upstream effects, whereas LPAR1 selective inhibition may confer a more regulated effect. Furthermore, LPAR1 inhibitors may initially cause increased LPA levels that could trigger a feedback mechanism, resulting in reduced levels of ATX and LPA [107]. Therefore, LPAR1 inhibitors may attenuate ATX activity more than that of ATX inhibitors (i.e. reduced expression versus pharmacological inhibition). To date, no drug–drug interactions between BMS-986278 or fipaxalparant and nintedanib or pirfenidone have been reported [108]. Although association of MOA and clinical outcome is not currently confirmed, the MOA of LPAR1 inhibitors may contribute to success in forthcoming trials of BMS-986278 and fipaxalparant. Moreover, synergy between LPAR1 inhibitors and antifibrotic therapies may occur due to distinct MOAs. Forthcoming trials will assist in testing these hypotheses.