PPARα, PPARδ, or both—that is the question!

Cholestatic pruritus may affect 2 out of 3 individuals with, for example, primary biliary cholangitis (PBC) and primary sclerosing cholangitis (PSC), the most common chronic cholestatic liver diseases.1 Cholestatic pruritus shows a diurnal variation with peak intensity early at night and may be localized particularly at the limbs but can also be generalized. It can be mild, but also most severe, dramatically impairing quality of life.1 Potential pruritogens in cholestasis are thought to activate receptors on unmyelinated C fibers (itch neurons) in the skin, thereby triggering a complex neural network leading to the sensation of itch with the desire to scratch. Proposed candidate pruritogens in cholestatic pruritus include certain lysophospholipids,2 sulfated progesterone metabolites, and endogenous opioids, among others.1 A number of observations raised doubts in the role of bile acids as potential pruritogens in cholestasis. For example, effective reduction of serum bile acids by treatment with the potent bile acid sequestrant colesevelam did not reduce itch perception in subjects with PBC and PSC with moderate to severe itch when compared with placebo in a randomized placebo-controlled trial3 and individuals with a rare genetic disorder, Na+-taurocholate cotransporting polypeptide deficiency, who live with serum bile acids in the millimolar range (more than 100-fold higher than the upper limit of normal) were reported not to suffer from pruritus at all.1 Thus, total serum bile acids appear unlikely to be dominant pruritogens in cholestatic disorders. Medical therapy of cholestatic pruritus has clearly advanced in recent years with the introduction of the peroxisome proliferator-activated receptor (PPAR) agonist bezafibrate as an antipruritic strategy in Europe and Asia. In addition to case series, 2 academic randomized, placebo-controlled trials showed beneficial effects of bezafibrate versus placebo on cholestatic pruritus: in the BEZURSO trial,4 individuals with PBC (n = 100) were treated with bezafibrate+ursodeoxycholic acid (UDCA) or placebo+UDCA for 2 years. During this period, mostly mild pruritus as a secondary study end point improved together with markers of cholestasis when compared to placebo.4 In the FITCH trial,5 the so far largest itch trial in individuals with PBC and PSC (n = 74) suffering from most severe to moderate pruritus (5–10/10, visual analog scale), treatment for 3 weeks with bezafibrate added to UDCA (in most cases) markedly improved pruritus in the majority of included individuals in comparison to placebo (ambitious primary end point visual analog scale improvement ≥ 50% in 45% vs. 11%, p = 0.003). Secondary end points, including morning and evening itch intensity, the validated 5D-Itch questionnaire, and the pruritus domain of the Liver Disease Symptom Index 2.0 clearly confirmed results for the primary end point. The pan-PPAR agonist bezafibrate has meanwhile been recommended as first-line treatment of cholestatic itch in the most recent EASL clinical practice guidelines for sclerosing cholangitis. An equivalent update of PBC guidelines is awaited. Bezafibrate stimulates PPARα, PPARδ(=β), and PPARγ. It has been questioned whether selective PPARα or PPARδ agonists could be equally effective in the treatment of cholestatic pruritus when compared to pan-PPAR agonists. While selective PPARα agonists like fenofibrate (or ciprofibrate) show antipruritic activity in PBC, this effect may be weaker than that of the pan-PPAR agonist bezafibrate.6 Recently, the selective PPARδ agonist seladelpar (formerly MBX-8025; 10 mg daily) was reported to be more effective than placebo with regard to markers of cholestasis, but also pruritus after 3 months of treatment in the industry-driven ENHANCE trial, a phase 3 randomized, placebo-controlled trial including 265 patients with PBC under treatment with UDCA when appropriate.7 Although the mean difference of 10 mg Seladelpar daily (but not 5 mg daily) versus placebo in alleviating pruritus after 3 months (as determined by number rating scale (NRS)) was moderate at best (mean Δ −1.59, p = 0.02) and only restricted to a subgroup of 56 (21%, those with NRS ≥ 4 at baseline) of the 265 included patients (restriction to subgroup missed to mention in the abstract of the Hepatology publication7), the data are of clinical interest. Still, the 5D-itch scale and PBC-40 itch domain tended to be, but were not documented to be significantly improved in this subgroup analysis of patients with NRS ≥ 4,7 making the clinical relevance of these findings less convincing under the conditions tested when compared to those achieved previously under pan-PPAR agonists in Japan and Europe.4–6 We read with great interest the results of post hoc analyses of the above-introduced ENHANCE trial published in the present issue of Hepatology.8 This paper, authored by Kremer et al further elaborated on the potential moderate antipruritic effects of seladelpar at the only effective but well-tolerated dose of 10 mg daily. Screening of serum samples for (i) potentially itch-inducing cytokines such as IL-4, IL-17, IL-31, or IL-33 by ultrasensitive immunoassays and single molecule array technology, (ii) serum autotaxin levels by ELISA, and (iii) 15 serum bile acid species by liquid chromatography-tandem mass spectrometry was performed. The results of the determinations were correlated with each other and with itch intensity, whether causal or not. IL-31 turned out to be the only elevated serum cytokine of those screened, which reacted to 3 months of seladelpar treatment at 10 mg daily (Supplemental Figure 1). Therefore, the focus of the presented work was directed to this potentially itch-inducing cytokine. IL-31 leads to itch perception through activation of the IL-31-Oncostatin-M-specific receptor-subunit β heterodimer on TRPA1-positive and TRPV1-positive neurons. Serum IL-31 can be elevated in diverse inflammatory conditions, which include liver diseases, such as metabolic dysfunction–associated steatohepatitis, PBC, and PSC,9 dermatological disorders, such as atopic dermatitis, psoriasis, and chronic urticaria, all accompanied by itch, but also inflammatory diseases without accompanying itch such as allergic airway diseases or axial spondyloarthritis. For dermatological diseases accompanied by itch, the question recently arose whether IL-31 is a cause or consequence of itch: a mouse model for atopic dermatitis that displays scratching behavior showed that nail clipping reduced serum IL-31 levels, possibly due to an impaired ability to damage the skin barrier, implicating that increased IL-31 levels could be the consequence rather than the cause of an increased scratch activity.10 In the present article, the authors clearly showed > 30-fold higher serum IL-31 levels in 161 subjects with PBC who completed the 3-month period of the ENHANCE trial than in 50 age-, sex- and weight-matched healthy controls, confirming the most recent data of others.9 They also observed 3-fold higher total serum bile acid levels in the PBC cohort than in healthy volunteers (as expected), mainly in the high normal range. A correlation between serum IL-31 and total serum bile acids was shown (Figure 1). For data interpretation, it would have been desirable to also correlate IL-31 serum levels to those of plasma alkaline phosphatase activities as an established marker of cholestasis and potent prognostic marker in (noncirrhotic) PBC, the improvement of which had been associated with improved itch intensity under treatment with the pan-PPAR agonist bezafibrate in the FITCH trial.5 The authors then analyzed the effect of seladelpar (5 mg vs. 10 mg daily) or placebo treatment on serum IL-31 and clearly showed an impressive decrease of mean -52% in the 10 mg seladelpar group after 3 months, markedly different from placebo and possibly the most remarkable finding of this post hoc analysis (Figure 2). Notably, farnesoid X receptor agonists discussed as alternative second-line treatment to PPAR agonists in PBC and other cholestatic cholangiopathies—and in case of obeticholic acid in PBC approved by authorities (U.S. Food & Drug Administration, European Medicines Agency)—have recently been shown to stimulate IL-31 formation in parallel with induction of itch.9 In a third step, the authors tried to link pruritus intensity to IL-31 serum levels by correlation of IL-31 with pruritus intensity at baseline in the whole study cohort (n=161), although they otherwise focused in their pruritus-related analyses only on the PBC subgroup with moderate to severe pruritus responsive to seladelpar (NRS ≥ 4).7,8 Figure 3A shows a strong accumulation of data points in the lower left, but also considerable numbers of subjects with intense itch but very low serum IL-31 levels and scattering in the upper left area where subjects with NRS ≥ 4 are localized. It would have been interesting to see the correlation of pruritus intensity and IL-31 for this selective subgroup with moderate to severe pruritus on which the authors relied when discussing a potential antipruritic effect of seladelpar.7,8 A validation cohort (Supplemental Figure 2B) showed an even weaker correlation. The graphs raise some doubts on a link between relevant itch intensity (NRS ≥ 4) and IL-31 serum levels in this subgroup. Consequently, Figure 3C shows that a "clinically relevant" decrease of itch intensity (NRS ≥ 2) was achieved by 11 of 55 (20.0%) placebo-treated subjects with PBC, by 13 of 53 (24.5%) subjects in the 5 mg seladelpar group, and by 10 of 53 (18.9%) subjects in the 10 mg seladelpar group, which is not different. Thus, a link between the clearly documented decrease of serum IL-31 by seladelpar and an antipruritic effect by the drug in PBC appears unlikely based on the data presented. Supplemental Figure 3 of the paper shows that there is a correlation of IL-31 with the level of conjugated bile acids but not with that of unconjugated bile acids. It has been established that conjugated bile acids but not unconjugated bile acids increase during cholestasis, which could in part explain this difference. A correlation of serum bile acids with itch intensity is not reported in the present paper. Finally, this study confirms that autotaxin, which forms one candidate pruritogen for cholestatic pruritus, lysophosphatidic acid,2 appears not involved in PPAR agonist-induced amelioration of cholestatic pruritus as recently documented for the pan-PPAR agonist bezafibrate in the FITCH trial.5 In conclusion, the present post hoc analysis presents clear evidence for an impressive inhibitory effect of the PPARδ receptor agonist seladelpar on serum levels of the potentially pruritogenic cytokine IL-31 while it shows a quite moderate improvement of pruritus when compared to placebo in a minor subgroup (21%) of subjects with PBC (NRC ≥ 4) included in the ENHANCE cohort. On the basis of these results, it remains unclear to what extent the effect of seladelpar on IL-31 serum levels is pathophysiologically linked to the moderate effect of seladelpar on itch intensity in the limited subgroup that responded. From a clinical perspective, the effect of the PPARδ agonist seladelpar on cholestatic itch appears minor compared to that observed with the pan-PPAR agonist bezafibrate4–6 at least during the first weeks of treatment, but head-to-head comparisons are lacking. If the minor therapeutic effect of PPARδ agonists on pruritus is confirmed in further studies, one might conclude that combined PPARα and PPARδ agonism is more effective than PPARδ agonism only in cholestatic pruritus.