Investigational therapies targeting the granulocyte macrophage colony-stimulating factor receptor-α in rheumatoid arthritis: focus on mavrilimumab

Abstract

Mavrilimumab (formerly CAM-3001) is a high-affinity, immunoglobulin G4 monoclonal antibody (mAb) against the granulocyte macrophage colony-stimulating factor (GM-CSF) receptor-α chain. Phase I and II trials in patients with rheumatoid arthritis (RA) treated with mavrilimumab have shown encouraging results with respect to both safety and efficacy. No significant adverse events have so far been noted. The trials have demonstrated significant clinical benefit, meeting primary endpoints. Furthermore, for RA patients treated with mavrilimumab, who were tumour necrosis factor (TNF) inhibitor-inadequate responders, there are encouraging preliminary data indicating benefit and identifying potential biomarkers predictive of patients likely to find benefit. Here, we review the clinical trial data for mavrilimumab and discuss its potential as a treatment for RA in light of the competitive landscape in which it resides.

Introduction

Rheumatoid arthritis (RA) is a chronic, systemic, autoimmune disease of unknown aetiology, which primarily affects joints. With the advent of targeted biological treatments, such as those directed against tumour necrosis factor (TNF), the management of RA has been revolutionized.¹ However, there are still patients who are refractory to such treatments, show inadequate responses or experience a loss of efficacy after a primary response; therefore, there is still the need to develop new biologics for additional targets. One such target is granulocyte macrophage colony-stimulating factor (GM-CSF; also known as CSF2) or its receptor (GM-CSFR).

GM-CSF was originally defined by its ability to generate colonies of granulocytes and macrophages in vitro from immature bone marrow cells.² Subsequently, it was shown that it could also affect more mature populations in these lineages, promoting their survival/activation/differentiation.^3,4 GM-CSF can be produced by a number of haemopoietic and nonhaemopoietic cell types upon stimulation, and it can activate/‘prime’ myeloid populations to produce inflammatory mediators, such as TNF and interleukin 1β (IL1β). GM-CSF can therefore be considered a pro-inflammatory cytokine that acts at the interface between innate and adaptive immunity.^5–7 As a result, GM-CSF ‘networks’ have been proposed to attempt to explain its role, including in chronicity.^8–14

GM-CSF binds and signals through its specific GM-CSFR which comprises two subunits, a unique ligand binding α chain that contains three extracellular domains and a signalling β chain (Figure 1). The β-common (β_c) chain is also a component of the IL-3 and IL-5 receptors. The GM-CSFR is expressed mainly on myeloid populations, including monocytes, macrophages, eosinophils and neutrophils. The crystal structure of human GM-CSFR coupled to GM-CSF shows a higher-order assembly comprising both an α hexamer, consisting of two molecules each of ligand, the receptor α chain and the receptor β chain, as well as an unexpected dodecamer in which two hexameric complexes associate.^15,16 The binding of GM-CSF to its receptor activates the JAK2/STAT5 pathway.^15,17 Other pathways, including the Ras–Raf-mitogen-activated protein kinase (MAPK), nuclear factor (NF)-κB and phosphoinositide 3-kinase (PI3K)-Akt pathways, have been reported to be activated as a result of GM-CSFR engagement.¹⁷

Figure 1.

Mode of action of mavrilimumab. (A) Highly schematic representation of GM-CSF binding to the GM-CSF receptor. The GM-CSF receptor subunit (depicted) consists of a unique ligand binding α chain that contains three extracellular domains and a β-common (β_c) chain through which signalling occurs. Binding of GM-CSF to its receptor shows a higher-order assembly comprising both an α hexamer, consisting of two molecules each of ligand, the receptor α chain and the receptor β chain, as well as a dodecamer in which two hexameric complexes associate.^15,16 (B) Mavrilimumab competes with GM-CSF for binding to the GM-CSF receptor α chain and prevents subsequent intracellular signalling via the β_c chain.

GM-CSF, granulocyte macrophage colony-stimulating factor.

Preclinical development of anti-granulocyte macrophage colony-stimulating factor receptor α monoclonal antibody for the treatment of rheumatoid arthritis

Evidence for a role for GM-CSF in RA comes from many studies (reviewed in Hamilton,^4,7 Hamilton et al.,⁵ and Wicks and Roberts¹⁸). Raised GM-CSF levels in RA synovial fluid and plasma, and overexpression of the GM-CSFR within cells of RA synovial tissue have been reported.^19–21 Depletion of GM-CSF has been shown to suppress arthritis in a number of mouse models^7,18,22–27 and, as a result, a number of monoclonal antibodies (mAbs) targeting GM-CSF are being developed for the treatment of RA, which have recently been reviewed elsewhere⁵ and are summarized below.

As discussed above, the GM-CSFRα chain is specific for the GM-CSFR and responsible for binding GM-CSF with high specificity and low affinity.^15,16 Given the lack of a GM-CSFRα chain specific gene-deficient mouse and, until recently, a specific anti-mouse GM-CSFRα mAb, data specifically neutralizing GM-CSF receptor signalling has been lacking. However, the cloning of the gene encoding the human GM-CSFα chain, CSF2A, led to the development of mavrilimumab (formerly CAM-3001). Mavrilimumab is a high-affinity, immunoglobulin (Ig)G4 mAb which has been developed by MedImmune against GM-CSFRα. It was isolated by phage display, is a poor complement activator due to its being an IgG4 isotype and binds the GM-CSFRα chain with high affinity. Thus mavrilimumab competes with GM-CSF for binding to the GM-CSFRα chain and prevents subsequent intracellular signalling via the βc chain (Figure 1).Using mavrilimumab, it was found that there was increased expression of GM-CSFRα by cluster of differentiation (CD)68⁺ and CD163⁺ synovial macrophages from RA and psoriatic arthritis patients compared with osteoarthritis patients and healthy controls.²⁸ An analogous mAb, CAM-3003, against murine GM-CSFRα, was effective at ameliorating murine collagen-induced arthritis (CIA)²⁸ and reducing the numbers of F4/80⁺ macrophages in the inflamed synovium of both CIA and antigen-induced arthritis (AIA).^28,29 GM-CSFRα blockade was shown to modulate inflammatory responses differently to TNF or IL-6 blockade with more of an effect on depleting inflammatory macrophage/monocyte-derived dendritic cell numbers.²⁹

Comprehensive preclinical toxicology studies evaluating the safety of mavrilimumab have been carried out in nonhuman primates.³⁰ In particular, the development of lung toxicity and pulmonary alveloar proteinosis (PAP) were of potential concern with regards to long-term blockade of GM-CSFR signalling. PAP is a heterogenous lung condition characterized by the accumulation of surfactant lipids and protein in pulmonary alveolar macrophages resulting in respiratory insufficiency.³¹ Some forms of hereditary PAP are associated with mutations in the genes encoding the GM-CSFR, while other forms are characterized by high levels of anti-GM-CSF autoantibodies. Weekly intravenous or subcutaneous dosing studies, ranging from 4 weeks to 26 weeks’ duration, in nonhuman primates showed mavrilimumab to have an acceptable safety profile with no changes in any parameters apart from microscopic findings in the lung, where slight accumulation of foamy macrophages was observed in the animals dosed for at least 11 weeks.³⁰ However, this was reversible upon cessation of treatment and considered to be nonadverse. In a 26-week repeat intravenous dose study, the presence of lung foreign material, cholesterol clefts and granulomatous inflammation were also present in a few animals dosed with high levels of mavrilimumab (⩾30 mg/kg/week).³⁰ At lower doses (3 and 10 mg/kg/week for 26 weeks) no mavrilimumab-related effects were seen in the lung, nor for any other parameter assessed. The dose- and time-related changes noted in the lung macrophages following mavrilimumab treatment were expected, based upon the known role of GM-CSFRα signalling in the development and function of alveolar macrophages. Despite this, there were high clinical exposure safety margins compared with the doses that caused these adverse changes in the lung.³⁰ A recent study modelled systemic versus pulmonary pharmacodynamics of CAM-3003 (anti-mouse GM-CSFRα mAb) and determined that dosing with CAM-3003 using regimes that saturate circulating cells, and shown to be efficacious in inflammatory arthritis models²⁸, did not lead to complete blockade of the alveolar macrophage response to GM-CSF,³² suggesting that a significant therapeutic window is possible with GM-CSFR inhibition.

To date, no pulmonary changes have been observed following treatment with efficacious doses of mavrilimumab in clinical trials (see below). This finding is consistent with the recent proposal³³ that PAP is unlikely to be induced by mAbs to GM-CSF but only by polyclonal Abs as found in autoimmune PAP.

Phase I clinical trials with mavrilimumab

The first in-human phase I trial was a randomized, double-blind, placebo-controlled, dose-escalating study involving 32 patients with mild-to-moderate RA. Patients received a single intravenous dose of 0.01 (n = 1), 0.03 (n = 1), 0.1 (n = 5), 0.3 (n = 5), 1.0 (n = 5), 3.0 (n = 5) or 10.0 (n = 5) mg/kg mavrilimumab or placebo (n = 5) with stable methotrexate treatment [ClinicalTrials.gov identifier: NCT00771420]³⁴ (Table 1). All doses were well tolerated over the observation period of 24 weeks, with no significant changes in peripheral blood cell counts or other laboratory parameters. Lung function tests at 12 weeks showed no significant changes and no viral infectons were reported. In addition, no patients developed Abs to mavrilimumab. The half-life of mavrilimumab for doses > 1.0 mg/kg, at which the target is saturated, was calculated to be 5–15 days. While this study was not designed to assess efficacy, it was observed that 33% of patients (7/21) with a disease activity score 28-joint assessment (DAS28) of >2.6 at baseline achieved remission (DAS28 < 2.6) at week 4.³⁵ These results supported further clinical studies in RA.

Table 1.

Clinical trials involving mavrilimumab in rheumatoid arthritis.

Trial, ClinicalTrials.gov identifier	Drug regimen	Endpoint	Outcome	Ref
Phase I, NCT00771420	Escalating single intravenous doses of mavrilimumab (0.01, 0.03, 0.1, 0.3, 1.0, 3.0, 10.0 mg/kg) versus placebo, with stable methotrexate	Incidence and severity of adverse event (24 weeks of follow up)a	Minor adverse events profile No serious adverse events	34,35
Phase IIa, NCT01050998 Eastern European cohort	10, 30, 50, 100 mg subcutaneously doses of mavrilimumab given every second week versus placebo, with stable methotrexate for 12 weeks	>1.2 point reduction in the DAS28-CRPa EULAR moderate or good response ACR20/50/70	55.7% mavrilimumab versus 34.7% placebo, p = 0.003 [10 mg, 41.0% (p = 0.543); 30 mg, 61.0% (p = 0.011); 50 mg, 53.8% (p = 0.071); 100 mg, 66.7% (p = 0.001)] 67.7% mavrilimumab versus 50.7% placebo, p = 0.025 ACR20, 69.2% versus 40.0%, p = 0.005; ACR50, 30.8% versus 12.0%, p = 0.021; ACR70, 17.9% versus 4.0%, p = 0.030, mavrilimumab versus placebo	36
Phase IIa, NCT01050998 Japanese cohort	10, 30, 50, 100 mg subcutaneously doses of mavrilimumab given every second week versus placebo, with stable methotrexate for 12 weeks	>1.2 point reduction in the DAS28-CRPa ACR20	50.0% mavrilimumab versus 23.5% placebo, p = 0.081 (30 and 100 mg doses, 75%, p = 0.028) 50.0% mavrilimumab versus 23.5% placebo, p = 0.081 (100 mg dose, 75%, p = 0.028)	37
Phase IIb, NCT01706926 EARTH EXPLORER 1	30, 100, 150 mg subcutaneously doses of mavrilimumab given every other week versus placebo with stable methotrexate	DAS28-CRP change from baseline at 12 weeksa ACR20 response (24 weeks)a	Mavrilimumab 30 mg, −1.37 (0.14); 100 mg, −1.64 (0.13); 150 mg, −1.90 (0.14) versus placebo −0.68 (0.14), p < 0.001 [change from baseline (SE)] Mavrilimumab 30 mg, 51%; 100 mg, 61%; 150 mg, 73% versus placebo 25%, p < 0.001	38
Phase IIb, NCT01715896 EARTH EXPLORER 2	100 mg mavrilimumab subcutaneously given every other week or 50 mg golimumab subcutaneously every 4 weeks, with stable methotrexate	ACR20/50/70 responses at 24 weeksa DAS28-CRP < 2.6 at 24 weeksa HAQ-DI improvement > 0.22 at 24 weeksa	Mavrilimumab 62.0, 34.8, 16.1% (ACR20,50,70); golimumab 65.6, 43.4, 25.9% (ACR20,50,70) Mavrilimumab 17.4%; golimumab 29.0% (DAS28-CRP < 2.6) Mavrilimumab 58.7%; golimumab 69.0% (HAQ-DI improvement > 0.22)	39
Phase II, NCT01712399 Open-label extensionb	100 mg mavrilimumab sub-cutaneously every other week for up to 158 weeks (median 122 weeks), with stable methotrexate	Incidence and severity of adverse eventsa ACR20/50/70 responses at week 122 DAS28-CRP < 3.2/<2.6 at week 122	83 patients reported >1 pulmonary adverse event; bronchitis most frequent (34 patients) One serious adverse event (acute bronchitis) 80, 50, 30% (ACR20,50,70) DAS28-CRP < 3.2, 65%; <2.6, 41%	40,41

^aPrimary endpoints.

^bOpen-label extension study: rheumatoid arthritis patients were enrolled after completing either the EARTH EXPLOER 1 or 2 trials.

Abbreviations: ACR, American College of Rheumatology; CRP, C-reactive protein; DAS28, disease activity score 28; HAQ-DI, Health Assessment Questionnaire Disability Index; SE, standard error.

Phase II clinical trials with mavrilimumab

Next, mavrilimumab was extensively evaluated in the EARTH development program. The first trial was a phase IIa multi-centre, randomized, double-blind, placebo-controlled study in RA patients from Eastern Europe (n = 233) with moderate-to-severe disease activity and who were on stable methotrexate therapy [ClinicalTrials.gov identifier: NCT01050998] (Table 1). Sequentially increasing doses of 10 (n = 39), 30 (n = 41), 50 (n = 39) or 100 (n = 39) mg mavrilimumab or placebo (n = 75) were given subcutaneously every second week, in conjunction with methotrexate for 12 weeks. Patients were observed for an additional 12-week period following the end of treatment. At week 12, 55.7% of mavrilimumab-treated patients met the primary endpoint of a ⩾1.2 point reduction in the DAS28-CRP score compared with baseline, as did 34.7% of placebo-treated subjects (p = 0.003).³⁶ A European League Against Rheumatism (EULAR) moderate or good response was seen in 67.7% of mavrilimumab-treated patients compared with 50.7% of placebo-treated patients (p = 0.025). Patients receiving the 100 mg dose of mavrilimumab had the largest effect compared with placebo, with significantly more patients achieving American College of Rheumatology (ACR)20, ACR50 and ACR70 responses compared with placebo (Table 1). Response rate differences from placebo were observed at week 2 and increased throughout the treatment period and adverse reactions were mild to moderate. No significant hypersensitivity reactions, serious infections or changes in pulmonary parameters were observed.

As an add-on to the phase IIa trial described above [ClinicalTrials.gov identifier: NCT01050998], mavrilimumab was evaluated in a cohort of Japanese patients with RA (n = 51)³⁷ (Table 1). Sequentially increasing doses of 10 (n = 9), 30 (n = 8), 50 (n = 9) or 100 (n = 8) mg mavrilimumab or placebo (n = 17) were again given subcutaneously every second week in conjunction with methotrexate for 12 weeks. For all mavrilimumab-treated patients, at week 12, a DAS28-CRP response (>1.2 point reduction) (50.0% versus 23.5%, p = 0.081) and an ACR20 response rate (50.0% versus 23.5%, p = 0.081) were observed more frequently than in placebo-treated patients; while these differences did not reach significance, the positive trends are encouraging, noting that the Japanese arm of the study was not powered to formally evaluate efficacy. At week 12, there were, however, significantly more patients who achieved an ACR20 response in the 100 mg mavrilimumab-treated group compared with the placebo-treated group (Table 1).³⁷ The safety profile of mavrilimumab in the Japanese cohort was in line with the findings from the Eastern European population. Adverse events were mild to moderate; one patient developed pneumonia which was considered to be possibly treatment related.³⁷

A subsequent 24 week, phase IIb, randomized, double-blind, parallel-group, placebo-controlled study (EARTH EXPLORER 1) [ClinicalTrials.gov identifier: NCT01706926] was performed in RA patients with moderate-to-severe disease (n = 326) and on background methotrexate³⁸ (Table 1). Patients received 30 mg (n = 81), 100 mg (n = 85) or 150 mg (n = 79) mavrilimumab or placebo (n = 81) given subcutaneously every other week, in combination with stable doses of methotrexate (7.5–25.0 mg/week) for 24 weeks. Mavrilimumab treatment significantly reduced the DAS28-CRP score from baseline compared with placebo at week 12 (Table 1).³⁸ Differences from placebo were detected at week 1, with benefit increasing until week 12. At week 24, significant ACR20 responses were achieved at all doses of mavrilimumab compared with placebo (Table 1). Dose-dependent, rapid (week 1) and sustained (week 24) reductions were seen in both CRP and ESR levels. One patient developed pneumonia and another developed angioedema, both of which were considered to be treatment related. Rates of pulmonary adverse events for mavrilimumab were similar to placebo. The phase IIb study met its coprimary outcomes. Mavrilimumab treatment significantly resulted in greater reductions from baseline in the DAS28-CRP score at week 12 and a greater percentage of ACR20 responders at week 24 compared with placebo. The highest dose (150 mg) of mavrilimumab was most effective.³⁸

A 24-week randomized, double-blind, phase IIb exploratory study (EARTH EXPLORER 2) [ClinicalTrials.gov identifier: NCT01715896] recently compared the efficacy and safety of mavrilimumab (100 mg every other week, subcutaneously, n = 70) with golimumab, an anti-TNF mAb (50 mg every 4 weeks, subcutaneously, n = 68), on top of methotrexate (7.5–25.0 mg/week), in patients with active RA who were inadequate responders to at least one DMARD and/or one to two TNF inhibitors³⁹ (Table 1). For mavrilimumab, there was an acceptable safety profile, with the only two serious treatment-related events [pneumocystis pneumonia (n = 1) and lung disorder (n = 1)] both being in golimumab-treated patients. Once again, no significant pulmonary safety signals were identified. At 24 weeks, the ACR20/50/70 DAS28-CRP < 2.6 and Health Assessment Questionnaire Disability Index (HAQ-DI) improvement > 0.22 response rates reached in the mavrilimumab- versus golimumab-treated patients are shown in Table 1, noting that this study was not powered to demonstrate statistical significance between the two mAbs.³⁹ The authors noted that the dose of 100 mg mavrilimumab every other week used in this trial was suboptimal compared with the dose of 150 mg every other week used in the EARTH EXPLORER 1 trial (see above). Peripheral biomarkers and pathophysiological pathways modulated by mavrilimumab and golimumab in the RA patients were also assessed in this study.⁴² Serum levels of CCL17 (TARC) and CCL22 (MDC), which share a common receptor, CCR4, were suppressed by mavrilimumab but not by golimumab, while CXCL13 and ICAM1 were suppressed by golimumab but not by mavrilimumab. In the DMARD-inadequate responder group, both mAbs induced early and sustained suppression of a number of mediators (CRP, SAA, MMP1, MMP3, IL-6, VEGF, IL-2R, and CD163), while RNA sequencing of peripheral blood cells showed significant changes in 1042 and 2058 transcripts at day 169 post-treatment with mavrilimumab and golimumab, respectively. In contrast, for the anti-TNF-inadequate responder group, only the mavrilimumab-induced changes were maintained until week 24, while the golimumab-induced early changes rapidly returned towards baseline levels. Mavrilimumab treatment of the anti-TNF-inadequate responders also led to long-term suppression of extracellular matrix markers (C1M, C3M, and P4NP7S), and induced significant changes in 1547 transcripts at day 169 in peripheral blood cells, whereas golimumab only induced a transient change in the extracellular matrix markers and had no impact on gene expression in the peripheral blood cells of this patient group.⁴² These results demonstrate that mavrilimumab, but not golimumab, treatment is able to induce sustained differential suppression of peripheral disease markers in anti-TNF-inadequate responders; the authors therefore concluded that treatment with mavrilimumab, but not with golimumab, may lead to greater long-term disease control in anti-TNF-inadequate responders.

In an attempt to further identify biomarkers which are predictive of patients likely to find benefit from mavrilimumab, various candidates have been measured in the sera of patients involved in the above described phase IIa [ClinicalTrials.gov identifier: NCT01050998] and IIb [ClinicalTrials.gov identifier: NCT02806926] clinical trials. Abs to peptidyl arginine deiminase 4 (PAD4) were measured in the sera of RA patients from these trials with 35% being positive.⁴³ In the DMARD-inadequate responders from the phase IIb study, those positive for anti-PAD4 were enriched for anti-citrullinated protein Abs (ACPA), had higher baseline joint erosions and higher modified total Sharp scores than patients who tested negative; following mavrilimumab (150 mg dose) treatment, those who tested negative for anti-PAD4 had significantly greater benefit from mavrilimumab compared with placebo (37.6% difference from placebo; odds ratio for ACR50 response = 17.61) relative to patients who tested positive for anti-PAD4 (8.7% difference from placebo; odds ratio for ACR50 response = 1.46).⁴³ Similarly, anti-PAD4 negative patients showed a greater change in their DAS28-CRP response compared with anti-PAD4 positive patients (−1.44 versus −0.79, anti-PAD4 negative versus anti-PAD4 positive). Similar trends were observed in the phase IIa trial and were specific to anti-PAD4 Abs, suggesting the presence of anti-PAD4 Abs may be predictive of patients less likely to benefit from mavrilimumab treatment. Further work is needed to determine whether this is the case.

Long-term safety of mavrilimumab treatment

The long-term efficacy and safety of mavrilimumab have been investigated in an open-label extension study [ClinicalTrials.gov identifier: NCT01712399] (Table 1). A total of 397 RA patients were enrolled after completing either the EARTH EXPLORER 1 or 2 trials (described above), or transferred from week 12 onwards because of an inadequate response. Patients received 100 mg mavrilimumab subcutaneously every other week for up to 158 weeks (median 122 weeks).⁴⁰ Mavrilimumab treatment was not associated with substantial negative effects on pulmonary safety. As regards clinical efficacy, at week 122, ACR20 (80%), ACR50 (50%) and ACR70 (30%) response rates were reached. Furthermore, a DAS28-CRP < 3.2 was achieved in 65% of patients and a DAS28-CRP < 2.6 in 41% of patients. Of note, after 74 weeks of treatment, 68% of patients had shown no radiographic progression of disease.⁴¹

Based upon the potential risk for lung toxicity and PAP development related to mavrilimumab treatment (see above), the pulmonary safety of mavrilimumab was specifically assessed in the phase IIb trials [ClinicalTrials.gov identifiers: NCT01706926; NCT01715896] and in the open-label extension study [ClinicalTrials.gov identifier: NCT01712399]. In total, mavrilimumab was given to 442 patients with a median (range) exposure time of 2.5 (0.1–3.3) years. Overall, 83 patients reported ⩾1 pulmonary adverse event, with bronchitis being the most frequent. The rate of adverse events remained stable over time. Only one adverse event was considered serious and treatment related (acute bronchitis). No PAP cases were found, nor pulmonary-related deaths reported. Thus, mavrilimumab demonstrated sustained efficacy and an acceptable safety profile and was not associated with a substantial decline in pulmonary function or with PAP in patients treated for up to 3.3 years.⁴⁴

As GM-CSF can regulate myelopoiesis, there is a potential risk of neutropenia associated with mavrilimumab treatment. However, no significant reduced neutrophil or monocyte counts, nor associated infections, have been reported in the phase I [ClinicalTrials.gov identifier: NCT00771420],³⁵ phase IIa [ClinicalTrials.gov identifier: NCT01050998],^36,37 or phase IIb [ClinicalTrials.gov identifier: NCT01706926]³⁸ trials, nor in the open-label extension study [ClinicalTrials.gov identifier: NCT01712399].⁴⁰

Clinical trials using anti-granulocyte macrophage colony-stimulating factor monoclonal antibodies for the treatment of rheumatoid arthritis

While mavrilimumab is the only therapeutic currently directed against the GM-CSFRα chain, and the focus of this review, there are several other mAbs in RA clinical trials that target GM-CSF itself (Table 2) and which mavrilimumab may have to compete with commerically. These have been reviewed elsewhere⁵ and include the following trials.

Table 2.

Clinical trials using anti-granulocyte macrophage colony-stimulating factor monoclonal antibodies for the treatment of rheumatoid arthritis.

Name	Molecule/target	Manufacturer	Trial, ClinicalTrials.gov identifier	Ref
GSK3196165 (previously known as MOR103)	Human mAb to GM-CSF	Developed by MorphoSys AG and in-licensed by GlaxoSmithKline	Phase Ib/IIa, NCT01023256 Phase 11a, NCT02799472 Phase IIb, NCT02504671	45
KB003	High-affinity, recombinant IgG1κ mAb against GM-CSF	Kalobios Pharmaceuticals	Phase II, NCT00995449	46
Namilumab (MT203)	Human IgG1 mAb against GM-CSF	Takeda	Phase Ib, NCT01317797 Phase II, NCT02393378 Phase II, NCT02379091	47
MORAb-022	Human IgG1 mAb against GM-CSF	Morphotek/Esai	Phase I, NCT01357759	48

GM-CSF, granulocyte macrophage colony-stimulating factor; mAb, monoclonal antibody; IgG1κ, immunoglobulin G1 kappa.

GSK3196165 (previously known as MOR103) is a human mAb to GM-CSF developed by MorphoSys AG and in-licensed by GlaxoSmithKline (GSK),⁴⁵ which showed good tolerability and evidence of rapid efficacy and sustained clinical response up to 10 weeks beyond the 4-week treatment period in a phase Ib/IIa dose-escalation trial in patients with moderate RA [ClinicalTrials.gov identifier: NCT01023256].⁴⁵ There are several ongoing phase IIa/IIb trials assessing the efficacy and safety of GSK3196165 in combination with methotrexate therapy in patients with active RA [ClinicalTrials.gov identifiers: NCT02799472; NCT02504671], and a phase II trial in hand osteoarthritis exploring the potential of anti-GM-CSF treatment for disease modification and analgesic activity [ClinicalTrials.gov identifier: NCT02683785]. Results from these studies are pending. A safety (phase Ib) study with this mAb has also been carried out in patients with multiple sclerosis (MS) [ClinicalTrials.gov identifier: NCT01517282];⁴⁹ the treatment was generally well tolerated in these patients with relapsing–remitting MS and secondary progressive MS.

KB003 is a high-affinity, recombinant IgG1κ mAb against GM-CSF (Kalobios Pharmaceuticals). A randomized phase II clinical trial in patients with RA was terminated due to a programme refocus [ClinicalTrials.gov identifier: NCT00995449];⁴⁶ it was generally safe and well tolerated over ~3 months of repeat dosing. KB003 has also been tested in asthma [ClinicalTrials.gov identifier: NCT01603277].⁵⁰ Of note, there were no safety signals and no evidence of PAP.

Namilumab (MT203) is a human IgG1 mAb against GM-CSF. A phase Ib study in patients with mild-to-moderate RA [ClinicalTrials.gov identifier: NCT01317797] showed good tolerability and safety along with some efficacy.⁴⁷ Additional phase II trials [ClinicalTrials.gov identifiers: NCT02393378 and NCT02379091] have been completed; however, no results have been reported, to date.

MORAb-022, a human IgG1 mAb against GM-CSF (Morphotek/Esai), has undergone a phase I trial in patients with RA [ClinicalTrials.gov identifier: NCT01357759].⁴⁸ No results have been disclosed, to date.

While trials investigating the potential of mAbs against GM-CSF are not as advanced as those using mavrilimumab, they are similarly showing promise, with no significant drug-related adverse effects, particularly with regard to lung toxicity and PAP development; results from a number of trials are pending. Interestingly, the anti-GM-CSF mAbs are also being tested in indications other than RA.⁵

Conclusion

The clinical trial data on mavrilimumab with respect to safety, efficacy and speed of response are encouraging and warrant phase III trials. In particular, lung toxicity and PAP development have not been an issue so far. Furthermore, there is some encouraging data from patients treated with mavrilimumab who were inadequate responders to either a DMARD or TNF inhibitor(s).³⁹ Although golimumab treatment showed a similar response to mavrilimumab, the latter treatment appeared to sustain the suppression of several peripheral disease markers, suggesting it may lead to greater long-term disease control. However, one should be careful not to overinterpret this data as it was not sufficiently powered to demonstrate statistical significance between the two mAbs.

Because myeloid cell populations appear to be the main targets of GM-CSF activity during inflammation, the functions of GM-CSF/GM-CSFR are likely to be more restricted than those of, for example, TNF; these functional differences could mean that GM-CSF represents a unique targeting opportunity. Furthermore, in RA, some evidence suggests that GM-CSF is expressed earlier in the course of the disease than is TNF.^51–53 Preliminary data looking at the potential of various putative biomarkers indicate that it may be possible to select patients who are likely to respond to mavrilimumab, although more studies are required to verify this. With the extremely competitive market as regards current RA treatment, this type of personalized medicine is likely the way of the future.

As regards the broader potential of GM-CSF/GM-CSFR targeting, as discussed above, there are also several mAbs directed against the GM-CSF ligand in clinical trials for RA which are showing similar results to mavrilimumab with respect to safety, efficacy and response, making it a highly competitive landscape. Clarification on whether targeting the ligand or the receptor will be better or not is still needed. Interestingly, several of these mAbs are being trialled in other indications, such as osteoarthrits, MS and psoriasis, which is something that ought also be considered for mavrilimumab.