Are emerging PGD2 antagonists a promising therapy class for treating asthma?

KEYWORDS:

• Asthma
• prostaglandin D₂
• CRTH2 antagonists
• DP2 receptor antagonists

1. Introduction

Until very recently, the management of asthma has centered around a handful of bronchodilators and corticosteroids developed empirically decades ago. The lack of therapeutic innovation is all the more surprising given the pressing clinical need: over 3000 people die of asthma a year in the United States alone, and ~50% of patients report exacerbations necessitating increased treatment in the last year.

That is all set to change. Respiratory medicine is entering a new era of biological therapies—treatments that selectively target specific inflammatory mediators and cellular pathways critical in disease pathophysiology. These treatments have already revolutionized patient care in rheumatology, gastroenterology, dermatology, and oncology. Almost all current and emerging biologic treatments for asthma target ‘type 2’ inflammation and require subcutaneous or intravenous administration. However, several pharmaceutical companies have recently developed inhibitors of prostaglandin D₂ (PGD₂) signaling offering an oral alternative capable of suppressing the type 2 inflammatory cascade. This editorial focuses on the rationale and efficacy of blocking PGD₂ signaling in asthma.

2. Asthma pathophysiology

Most asthma is characterized by ‘type 2’ inflammation, so-called because it is thought to be mediated by type 2 helper (Th2) cells and type 2 innate lymphoid cells (ILC2s) [1]. Th2 cells secrete the type 2 cytokines interleukin (IL)-4, IL-5 and IL-13, which in turn bring about the archetypal features of asthma: IgE production, mucus hypersecretion, airway hyperreactivity, and eosinophilia (Figure 1). Based on this understanding of asthma immunology, monoclonal antibodies directed against these mediators have been developed as novel therapies, specifically anti-IgE (omalizumab), anti-IL-5 (mepolizumab, reslizumab), anti-IL-5Rα (benralizumab), anti-IL-13 (lebrikizumab, tralokinumab), and anti-IL-4Rα (dupilumab).

Figure 1. Asthma immunology.

Meanwhile, attention has turned to describing the initiating events in the cascade of type 2 inflammation. Several candidates have subsequently been identified with the potential to stimulate Th2 cells and ILC2s to release type 2 cytokines, at least in vitro. These include IL-25, IL-33, thymic stromal lymphopoietin and PGD₂.

3. PGD₂ biology

PGD₂ is a lipid inflammatory mediator produced by the sequential action of cyclooxygenases (particularly COX-2) and PGD₂ synthases (PGDS)—either hematopoietic PGDS in circulating hematopoietic-derived cells, or lipocalin PGDS in the brain, heart, and adipose tissue (Figure 2). Mast cells are traditionally thought to be the principal source of PGD₂ as they release vast quantities in response to IgE binding [2], but Th2 cells [3] and macrophages [4] may also produce biologically significant amounts (eosinophils and basophils can also produce PGD₂, but in concentrations roughly 1000 times more dilute than mast cells).

Figure 2. Prostaglandin D₂ biology.

PGD₂ binds the D prostanoid (DP) 1 and Chemoattractant receptor-homologous molecule expressed on Th2 cells (CRTH2) receptors. DP1 is found on a variety of cell types and has broadly anti-inflammatory effects. CRTH2, by contrast, is expressed selectively on immune cells, specifically eosinophils, basophils, Th2 cells, and ILC2s. Indeed, although it has long been known that PGD₂ has bronchoconstricting effects when inhaled, it was only after the relatively recent discovery of the CRTH2 receptor that a pro-inflammatory role for PGD₂ in allergic conditions has been described. Thus, following CRTH2 receptor binding in vitro, PGD₂ triggers chemotaxis and, in the case of ILC2s and Th2 cells, the release of the type 2 cytokines IL-4, IL-5, and IL-13 [5,6]. CRTH2 binding is therefore hypothesized to be important in diseases characterized by type 2 inflammation, such as asthma, atopic dermatitis, and allergic rhinitis.

There is good evidence that the PGD₂-CRTH2 pathway is upregulated in asthma. Both COX-2 [7] and hematopoietic PGDS [8] are more highly expressed in asthmatic lungs, suggesting a greater capacity for producing PGD₂, and there are greater numbers of CRTH2⁺ cells [9], increasing sensitivity to PGD₂. These observations are more pronounced in patients with poor asthma control [8].

4. The case for CRTH2 antagonists

Given the pathogenic roles of IL-4, IL-5, and IL-13 in asthma, a broader treatment that suppresses all three mediators is likely to be more effective than monoclonal antibodies targeting each individually. Of the various candidates that might trigger type 2 inflammation, PGD₂-CRTH2 activation is particularly attractive as there is evidence to support it being a dominant pathway: whilst IL-25 and IL-33 can also induce type 2 cytokine release by ILC2s in vitro, their stimulatory effect is inhibited by a selective CRTH2 antagonist [6].

On a practical level, selective CRTH2 antagonists are small molecules (rather than monoclonal antibodies) and can therefore be produced relatively cheaply, stored without refrigeration, and administered orally. As with all chronic disorders, non-adherence is a major problem in asthma, and an oral alternative could improve adherence and hence efficacy. There is therefore a compelling case for clinical development of selective CRTH2 antagonists.

5. Evidence to date

There have been several trials of CRTH2 antagonists in stable asthma and in the asthmatic response to an inhaled allergen challenge, a model with good positive and excellent negative predictive value for drug efficacy (Table 1). These have universally found CRTH2 antagonists to be safe and well-tolerated, even at the highest doses. The overall results in terms of efficacy, however, have been underwhelming, with inconsistent reports of statistically significant but small (and potentially not clinically significant) improvements in lung function and quality of life measures (it is worth noting that the bronchoconstriction effects of PGD₂ are via the activity of its metabolite 11β-PGD₂α on the thromboxane receptor, so CRTH2 blockade would not be expected to affect lung function). How do we explain the lack of clinical efficacy? Much of it may come down to two aspects of clinical trial design, selection of the appropriate population and outcome measures.

Table 1. Trials of CRTH2 antagonists in asthma.

Study	Year	Population	Intervention	Main findings
OC459
Barnes et al.	2012	ICS-free, allergic asthma n = 132	OC459 vs. placebo	Small but significant improvement in quality of life (AQLQ) and nighttime symptom scores Small improvement in FEV1, significant in per protocol analysis only (+9.2% vs. +1.8% placebo; p = 0.037)
Singh et al.	2013	ICS-naive allergic asthma FEV1 > 65% n = 16	OC459 vs. placebo	Effect on response to bronchial allergen challenge Reduced late but not early asthmatic response, reduced sputum eosinophils No effect on FEV1, FeNO
Pettipher et al.	2014	ICS-free, FEV1 60–85% n = 476	OC459 vs. placebo	Small but significant improvement in FEV1, ACQ, and AQLQ Nonsignificant trend toward reduced exacerbations
BI671800
Miller et al.	2012	ICS-treated, FEV1 60–85%, ACQ ≥1.5 n = 108	BI671800 + ICS vs. placebo + ICS	No significant difference in trough FEV1 or ACQ
Sutherland et al.	2012	ICS-naive, FEV1 60–85%, ACQ ≥1.5 n = 388	BI671800 vs. ICS vs. placebo	Small but significant improvement in FEV1 and ACQ
Hall et al.	2015	FEV1 60–85%, ACQ ≥1.5 (a) ICS-free (n = 388) (b) ICS-treated (n = 243)	(a) BI671800 vs. ICS (b) BI671800 + ICS vs. montelukast + ICS vs. placebo + ICS	For (a) small significant improvement in trough FEV1, no effect on ACQ For (b) small significant improvement in trough FEV1 and ACQ vs. placebo + ICS but not vs. montelukast + ICS
QAW039
Gonem et al.	2016	ICS-treated, ACQ ≥ 1.5, sputum eosinophil  ≥ 2% n = 61	QAW039 vs. placebo	Significant reduction in sputum eosinophils, and small significant improvements in FEV1 (+0.06 L vs. −0.10L placebo; p = 0.021) and AQLQ (+0.27 vs. −0.33 placebo; p = 0.008)—but not ACQ In subgroup with uncontrolled asthma at baseline, ACQ clinically and statistically significantly lower
Erpenbeck et al.	2016	Atopic, FEV1 60–85%, ACQ ≥ 1.5 n = 170	QAW039 vs. placebo	No significant differences in FEV1 or ACQ overall (both improved if baseline FEV1 < 70%)
Bateman et al.	2016	ICS-treated (low dose), FEV1 40–80%, ACQ ≥ 1.5 n = 1058	QAW039 + ICS vs. montelukast + ICS vs. placebo + ICS	Significant reduction in trough FEV1 vs. placebo, no effect on ACQ or AQLQ
ARRY-502
Wenzel et al.	2014	ICS-free, FEV1 60–85%, ACQ ≥ 1.5 n = 184	ARRY-502 vs. placebo	Patients with elevated Th2 associated biomarkers (e.g. FeNO) had improved spirometry, measures of asthma control and quality of life (unspecified)
Discontinued (AMG853, ACT-129968, AZD1981)
Busse et al.	2013	ICS-treated (200–1000μg/d fluticasone), FEV1 50–85%, ACQ ≥ 1.5 n = 396	AMG853 + ICS vs. placebo + ICS	No significant difference in ACQ, FEV1, symptoms, exacerbations, AQLQ, serum IgE, FeNO; discontinued
Diamant et al.	2014	ICS-free, house dust mite allergy, FEV1 > 70% n = 18	Setipiprant (ACT-129968) vs. placebo	Reduced late but not early asthmatic response No effect on serum eosinophils, IgE, and FeNO
NCT01225315 (unpublished)	2012	ISC-free, FEV1 ≤ 85%, ACQ ≥ 1.5 n = 438	Setipiprant (ACT-129968) vs. placebo	Did not replicate efficacy of allergen challenge model (no details available); discontinued
NCT00758589 (unpublished)	2013	ICS-treated, FEV1 40–85% n = 368	AZD1981 + ICS vs. Placebo + ICS	Significant improvement in ACQ, FEV1 only improved for middle of three doses; post-hoc analysis showed responders were atopic
NCT01197794 (unpublished)	2013	Atopic, ICS-LABA-treated, FEV1 40–85% n = 1144	AZD1981 + ICS-LABA vs. placebo + ICS-LABA	Only patients on second lowest dose demonstrated statistically significant improvement in FEV1; discontinued

ACQ: Asthma Control Questionnaire score; FEV; forced expiratory volume; ICS: Inhaled Corticosteroids.

Asthma is increasingly recognized as a heterogeneous disease with various clinical phenotypes and molecular endotypes. Given this, it is unrealistic to expect a single drug to be a panacea for all asthmatics. The type 2 inflammatory endotype, which theoretically should respond best to CRTH2 antagonism, is only one of these endotypes, albeit the most common representing around half of all asthma. However, most clinical trials of CRTH2 antagonists have selected patients on the basis of severity (e.g. use of inhaled corticosteroids or forced expiratory volume in 1 s, FEV₁) rather than phenotype or endotype (see Table 1). It makes more sense to select participants on the basis of biomarkers of type 2 inflammation, such as blood eosinophilia or exhaled nitric oxide (FeNO) levels. Indeed, when these study participants are analyzed separately, the results are more impressive. For example, there were significantly greater increases in FEV₁ in subgroups with: elevated serum eosinophils [10], positive skin prick tests [11], and raised FeNO [12].The choice of end points assessed may also be painting an unduly negative picture of CRTH2 antagonist efficacy in clinical trials of asthma. In particular, no trial to date has been powered to detect an effect on asthma exacerbations, an outcome responsible for the majority of morbidity and mortality in asthma and around half the health-care costs. Moreover, type 2 inflammation is particularly prominent during exacerbations, as demonstrated both by the marked reduction in exacerbation frequency using anti-IL-5 therapies [13,14] and experimentally using rhinovirus challenge models in asthma [15]. Our group has recently shown that PGD₂ also rises during asthma exacerbations, with levels correlating with increases in IL-5, IL-13, and measures of exacerbation severity [16]. It is therefore biologically plausible that CRTH2 antagonists could be particularly effective in preventing or attenuating exacerbations, whilst simultaneously having a limited effect on stable disease. Interestingly, this appears to be the case for the emerging monoclonal antibody treatments, which produce only modest improvements in lung function and symptom scores in stable asthma, but crucially are effective in preventing ~40–50% of exacerbations (see Table 2).

Table 2. Trials of biologicals in asthma.

Study	Year	Drug	n		Selected outcomes (vs placebo)
				Exacerbations	Lung function / FEV1	Quality of life
		Anti-IgE
Cochrane meta-analysis	2014	Omalizumab	25 trials	38% reduction (OR 0.55)	Not significant	AQLQ +0.31
		Anti-IL-5/-5R
Haldar et al	2009	Mepolizumab	61	43% reduction (p=0.02)	Not significant	AQLQ +0.35 (p=0.02)
Nair et al	2009	Mepolizumab	20	Reduced (p=0.002) (12 exacerbations in 11 patients on placebo vs 1 in 9)	+300mL 4 weeks after last dose Not sustained 8 weeks after last dose	Not significant
Castro et al	2011	Reslizumab	106	Not significant	+240mL (p=0.002)	Not significant
Pavord et al	2012	Mepolizumab	621	39-52% reduction (p<0.001)  (range for different doses)	Not significant	Not significant
Ortega et al	2014	Mepolizumab	576	47-53% reduction (p<0.001)  (range for iv vs sc administration)	+98-100mL (p<0.05)	SGRQ +6.4-7.0 (p<0.001)
						ACQ -0.42-0.44 (p<0.001)
Bel et al	2014	Mepolizumab	135	32% reduction (p=0.04)	Not significant	ACQ +0.52 (p=0.004)
Castro et al	2014	Benralizumab	609	Not significant (Significant for subset with serum eosinophils >0.3x10^9/L)	Not significant (Significant for subset with serum eosinophils >0.3x10^9/L)	Not significant
Castro et al	2015	Reslizumab	953	54% reduction (p<0.0001)	+110mL (p<0.0001)	AQLQ +0.23 (p<0.0001) ACQ -0.25 (p<0.0001)
		Anti-IL-13
Corren et al	2011	Lebrikizumab	219	Not significant	+110mL (p<0.02)	Not significant
Piper et al	2013	Tralokinumab	194	Not significant (relatively few events)	+200mL (p<0.041) in highest dose subset only	Not significant
Noonan et al	2013	Lebrikizumab	212	79% reduction (p<0.001)	Not significant	Not significant
De Boever et al	2014	GSK679586	198	Not significant	Not significant	Not significant
Hanania et al	2015	Lebrikizumab	463	60% reduction in periostin high group (24% reduction whole cohort) (p values not given, assume <0.05)	+9.1% (periostin-high) +2.6% (periostin-low)	Not significant
		Anti-IL-4Rα
Wenzel et al	2013	Dupilumab	104	87% reduction (p<0.001)	+250mL (p=0.0002)	ACQ -0.37 (p=0.256)
Wenzel et al	2016	Dupilumab	769	33-71% reduction (ns to p=0.0001)	+100-200mL (p=0.03 to <0.0001)If EOs≥0.3/mL, +170-260mL (p=0.0002)	AQLQ + 0.23 to +0.36 (not significant for 200mg 4wkly; otherwise p<0.02) AQQ if 2wkly dosing -0.31 to -0.35 (p<0.005); not significant if 4 wkly

Notes: FEV1 = forced expiratory volume in 1 second, minimum clinically important difference (MCID) +230mL (+10%); AQLQ = Asthma Quality Life Questionnaire, MCID ≥+0.5; ACQ = Asthma Control Questionnaire, MCID ≤-0.5; SGRQ = St George’s Respiratory Questionnaire, MCID >+4.0

Legend: green = clinically and statistically significant (p<0.05); red = not statistically significant; none = statistically but not clinically significant or equivocal for some other reason (given)

6. Future directions

There are a number of clinical trials ongoing of CRTH2 antagonists in asthma that will address the shortcomings outlined earlier. These include studies that restrict inclusion to asthmatics with evidence of type 2 inflammation (NCT02560610, NCT02660489, and NCT01836471) and those powered to assess an effect on exacerbations. The latter studies include those of sufficient length and size to study naturally occurring exacerbations (NCT02563067 and NCT02555683), or precipitating exacerbations following withdrawal of oral corticosteroid maintenance therapy (NCT02560610) or experimentally following rhinovirus challenge (NCT02660489).

Should the results be positive, studies comparing CRTH2 antagonists to existing treatments and other novel monoclonal antibodies targeting type 2 pathways will be required. Their cost to health-care systems will also likely determine where they fit into existing management pathways. In addition, studies in pediatric asthma and potentially in those with other clinical phenotypes, such as aspirin-exacerbated respiratory disease, will be needed to establish benefit across the asthmatic spectrum. As is the case for other treatments targeting type 2 inflammation, the discovery of a biomarker that identifies patients most likely to benefit from CRTH2 blockade would be invaluable, particularly in those whose FeNO and serum eosinophils are suppressed by inhaled corticosteroid treatment. Nonetheless given the positive findings with CRTH2 antagonists in the subset of asthmatics with evidence of type 2 inflammation, as well as the benefit of other type 2-targeted therapies in reducing exacerbations, we believe the outstanding studies are warranted and hopeful that they yield positive results.

Declaration of interest

S Johnston is a National Institute of Health Research Senior Investigator. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Additional information

Funding

H Farne was supported by a Medical Research Council Programme Grant (MR/M025330). S Johnston is supported by an Asthma UK Chair (CH11SJ) and a Medical Research Council Centre Grant (G1000758).