Costimulatory pathway targets for autoimmune and inflammatory conditions: clinical successes, failures, and hope for the future

KEYWORDS:

• Autoimmune diseases
• CD28
• CD40
• costimulatory signals
• ICOS
• OX40

1. Introduction

There are many products in early-phase clinical development for autoimmune and inflammatory (A/I) conditions that target various molecules within the costimulatory pathway. This pathway is involved in the activation of naïve T cells, which requires two separate signals [1]. The first is the binding of T-cell receptor and antigen peptide: major histocompatibility complex on antigen-presenting cells. The second signal, the so-called costimulatory signal, involves interactions between specific receptors and ligands on T cells and antigen-presenting cells. These costimulatory signals modulate pathways to help balance the immune system. Therefore, it is not surprising that this area has been a focus of drug development across a wide range of A/I-related diseases.

Several costimulatory pathways have been described, including the well-characterized B7 and tumor necrosis factor (TNF)-related families. Figure 1 illustrates important drug targets within these families that are being developed for A/I diseases. The B7 family includes activating signals, such as CD28 and induced costimulatory molecule (ICOS), as well as the inhibitory signal, cytotoxic T-lymphocyte-associated antigen 4 (CTLA4) [2].

Evaluating the history of costimulatory drug development can provide insight into the therapeutic potential for this area. Using the Informa Pharma Intelligence Trialtrove® (Informa Business Intelligence, Inc., New York, USA) and Pharmaprojects® (Informa Business Intelligence, Inc., New York, USA) databases to identify candidates and review their clinical trial development, this analysis highlights the successes and failures of pipeline drugs targeting costimulatory molecules within the A/I therapeutic area.

2. B7 family drug targets

2.1. CD28

The 2005 approval of the first-in-class costimulatory blocker abatacept, a cytotoxic T-lymphocyte-associated antigen 4 (CTLA4)-immunoglobulin G1 fusion protein, validated the strategy of targeting CD80 and CD86 to compete with CD28, thus inhibiting T-cell activation. Because CTLA4 is involved in regulatory T-cell function, therapeutic strategies directly targeting CD28 were developed to mitigate adverse effects [3]. Seemingly paradoxically, preclinical models also show that, not just inhibition, but agonism of CD28 results in attenuation of the immune response, possibly through expansion of regulatory T cells [1].

Theralizumab is in a unique class of CD28-targeting products, considered to have ‘super agonistic’ activity by inducing activation of T cells without the need for the first (T cell receptor: antigen) signal [4]. Trialtrove® data shows that the first TheraMAB-sponsored rheumatoid arthritis study for this drug initiated in 2013 and evaluated three doses (up to 10 µg per kilogram), with infusions that gradually decreased to a minimum of 1 h [5]. The low-dose and slow infusion time was selected as a result of an unfortunate ‘trial and error’ Phase I study, sponsored by TeGenero, in which theralizumab (then known as TGN1412) ended with catastrophic multi-organ failure in six healthy volunteers [6]. Despite the dire safety concerns, this mechanistic approach and the drug were not abandoned. TheraMAB gained rights to the drug and very carefully re-evaluated it, adjusting the dosing in preclinical and early-phase trials [7]. TheraMAB has since completed two small studies in patients, one each in rheumatoid arthritis and psoriasis, and has terminated a trial in patients with systemic lupus erythematosus for administrative reasons (NCT01990157, NCT02796053, NCT02711813). This leaves only one ongoing trial designed to determine dosing of theralizumab for the treatment of advanced solid malignancies (NCT03006029). Although CD28 can produce pro-inflammatory or anti-inflammatory signals, theralizumab’s current development profile targeting exclusively cancer now aligns with the other CD28 agonists.

The early catastrophic failure has left industry sponsors with products targeting CD28, even pathway antagonists, concentrating extra efforts on establishing preclinical and clinical safety measures. Bristol-Myers Squibb (BMS) and Atox Bio have drugs directly targeting CD28 that have undergone Phase II clinical trials. BMS was evaluating lulizumab pegol in patients with primary Sjögren’s syndrome and systemic lupus erythematosus. The 349-patient lupus trial did not demonstrate a significant difference between lulizumab and placebo at 24 weeks [8]. A few months before the completion of the lupus trial, BMS terminated the Sjögren’s syndrome trial after only three subjects had completed the treatment period, citing an inability of the trial to meet protocol objectives (NCT02265744, NCT02843659). Another pegylated Fab’ antibody antagonist of CD28 has completed a Phase I study in healthy volunteers [9]. OSE Immunotherapeutics’ partner, Janssen Biotech, is preparing for a Phase II trial in rheumatoid arthritis patients with FR104, a pegylated monovalent humanized Fab’ antibody fragment [10].

As opposed to targeting A/I conditions, Table 1 shows that Atox Bio is instead looking at using their CD28 antagonist peptide, reltecimod, to attenuate the dysregulated cytokine production following severe infections. The company reported positive results from a Phase II study of reltecimod [11] and, in 2015, a Phase III study was initiated that aims to demonstrate improved organ function as well as improved local signs of infection (NCT02469857). In the first half of 2018, a Phase II study evaluating the drug in sepsis-associated acute kidney injury commenced (NCT03403751).

Table 1. B7 Family: CD28 antagonists and ICOS ligand inhibitor in phase II+ clinical development.

Mechanism of action	Drug name	Originator, licensee	Diseases currently in phase II development
CD28 antagonist	Reltecimod	Atox Bio	Skin and skin structure infection^a Sepsis-associated acute kidney injury
ICOS ligand inhibitor	Prezalumab	Amgen, AstraZeneca	Sjögren’s syndrome

^ain Phase III development.

Source: Pharmaprojects® (September 2018) [12].

2.2. ICOS

Like CD28, ICOS provides a positive T-cell activation signal and is specifically involved in the activation of T-cell to B-cell functions, including immunoglobulin production [1]. This might make ICOS and its ligand an attractive target for diseases like Sjogren’s syndrome, in which T cells have been thought to initiate the autoimmune response and where B-cell hyperactivity is a hallmark of the disease [13]. Although structurally similar, CD28 induces the production of IL-2 while ICOS is more potent in IL-10 production [3].

Amgen’s fully human monoclonal anti-ICOS ligand antibody, prezalumab, recently completed a Phase IIa study in primary Sjögren’s syndrome, with results yet to be released (NCT02334306). A small Phase I exploratory lupus arthritis study reported potential clinical benefit in this underserved indication with a dose that had previously resulted in respiratory failure in a subject in the multiple ascending dose study, an event considered to be related to prezalumab [14]. On the contrary, the lupus arthritis study showed that treatment-emergent adverse events were similar between the drug and placebo [15]. Phase I studies in psoriasis (NCT01493518) and subacute cutaneous lupus erythematosus (NCT01389895) were terminated due to the sponsor’s decision and poor enrollment, respectively, about a year after a partnership with AstraZeneca was announced [16]. Amgen is also evaluating AMG 570, a bispecific antibody–peptide conjugate targeting ICOS ligand and B-cell activating factor [17], in a Phase I multiple ascending dose study enrolling rheumatoid arthritis patients (NCT03156023).

3. TNF-related family drug targets

Among the current drugs focused on the TNF super-family class, there are multiple monoclonal antibodies aiming to treat A/I diseases, as shown in Table 2.

Table 2. TNF-related Family: CD40/CD40 ligand antagonists and OX40 antagonists in Phase II+ clinical development.

Mechanism of action	Drug name	Originator, licensee	Diseases currently in Phase II development
CD40 ligand antagonist	Dapirolizumab pegol	UCB, Biogen, Nektar Therapeutics	Systemic lupus erythematosus
CD40 ligand antagonist	Letolizumab	Bristol-Myers Squibb	Immune thrombocytopenic purpura
CD40 antagonist	FFP104	PanGenetics, FF Pharma	Primary biliary cirrhosis Crohn’s disease
CD40 antagonist	Bleselumab	Kyowa Hakko Kirin, Astellas Pharma	Transplant rejection Glomerulosclerosis
CD40 antagonist	CFZ-533	Novartis, Xoma	Sjögren’s syndrome Transplant rejection Myasthenia gravis Rheumatoid arthritis Graves’ disease Lupus Nephritis
CD40 antagonist	BI 655064	Boehringer Ingelheim, AbbVie	Lupus nephritis
OX40 antagonist	GBR-830	Glenmark	Atopic dermatitis
OX40 antagonist	KHK-4083	Kyowa Hakko Kirin	Ulcerative colitis

Source: Pharmaprojects® (September 2018) [12].

3.1. CD40

Among the best-studied members of the costimulatory TNF-related family are CD40 and CD40 ligand [1]. CD40 was identified as a B-cell receptor and is important in the antibody-mediated immune response [18]. Since CD40 signaling effects are pleiotropic, affecting a variety of cells and biological processes, drug developers must be vigilant to minimize unwanted side effects [19].

UCB and partner Biogen’s Phase II systemic lupus erythematosus study with dapirolizumab pegol, a CD40 ligand antagonist [20], completed in early 2018 (NCT02804763). BMS’ letolizumab, another anti-CD40 ligand candidate [21], has completed a Phase I/II trial in immune thrombocytopenic purpura patients (NCT02273960). Both of these antibodies have been engineered specifically to prevent platelet activation, which may have caused adverse effects observed in previous anti-CD40 ligand drug trials. For example, ruplizumab (developed by Biogen) and toralizumab (developed by IDEC, which is now Biogen) were both discontinued in clinical trials due to thromboembolic events [22,23]. Viela Bio, a Medimmune spinoff, is now responsible for an anti-CD40 ligand-Tn3 fusion protein, engineered to be Fc deficient [24], which recently completed a Phase I study that included adults with rheumatoid arthritis (NCT02780388).

An alternative to targeting the CD40 ligand is direct inhibition of the CD40 receptor. This approach may also avoid the thromboembolism observed during treatment with CD40 ligand antibodies. A Phase I trial evaluating FF Pharma’s anti-CD40 antibody, FFP104, in patients with active psoriatic arthritis was previously terminated due to poor enrollment (NCT00787137). Trials for primary biliary cirrhosis (NCT02193360) and Crohn’s disease (NCT02465944) initiated in 2015 and 2016, respectively, which use the highest dose from the psoriatic arthritis trial as the lowest dose arms in these studies [25].

Although thromboembolic events have not been observed with Astellas and Kyowa Hakko Kirin’s fully human anti‐CD40 monoclonal recombinant immunoglobulin G4, results from a Phase IIa psoriasis study revealed that bleselumab did not improve Psoriasis Area and Severity Index scores, static Physician’s Global Assessment scores, or the percentage body surface area compared to placebo [26]. Results from a Phase IIa study in de novo kidney transplant recipients demonstrated that bleselumab with a calcineurin inhibitor minimization regimen had similar efficacy to standard of care in the prevention of biopsy-proven acute rejection [27]. A Phase IIa study in de novo kidney transplant recipients is ongoing to evaluate bleselumab for the prevention of recurrent focal segmental glomerulosclerosis (NCT02921789).

Novartis’ strategy is to run pilot studies with their fully human anti-CD40 monoclonal antibody, CFZ-533, in primary Sjögren’s syndrome (NCT02291029), de novo renal transplant recipients (NCT02217410), myasthenia gravis (NCT02565576), Graves’ disease (NCT02713256), and a first-in-human study that included a small number of rheumatoid arthritis patients (NCT02089087). More recent studies include patients with lupus (NCT03610516, NCT03656562). Results from the primary Sjögren’s syndrome study suggest that, in the higher intravenously dosed cohort, CFZ-533 strongly favored improvement in the EULAR Sjögren’s Syndrome Disease Activity Index over placebo [28]. Interim results from another study showed that CFZ-533 may be an effective calcineurin-free alternative treatment for kidney transplant recipients by preventing graft rejection without nephrotoxic and other calcineurin-related adverse effects [29]. In the Graves’ disease trial, none of the patients achieved normalization of thyroid stimulating hormone by week 12, however, CFZ-533 led to significant reduction in thyroid hormones during this time frame [30].

Boehringer Ingelheim’s anti-CD40 monoclonal antibody, BI 655064, remains active in Phase II development, despite reporting only moderate efficacy in a Phase I study that included rheumatoid arthritis patients [31] and termination of development in immune thrombocytopenic purpura at Phase I (NCT02009761). BI 655064 raised no relevant safety concerns and Phase II trials in lupus nephritis patients are ongoing (NCT02770170, NCT03385564). In 2016, AbbVie gained rights to BI 655064 and may elect to advance the program after completion of certain undisclosed clinical achievements, however, AbbVie currently lists another CD40 antagonist, ABBV-323, in Phase I for Crohn’s disease [32,33].

3.2. OX40

TNF-related family drug target, OX40, is also an area of interest for early-phase drug development. Since OX40 ligand is involved in the Th2 immune response, it is no surprise that Th2-driven diseases, such as atopic dermatitis, are being targeted by antagonists of this pathway [34].

Glenmark and Kyowa Hakko Kirin are in Phase II clinical trials with anti-OX40 antibodies GBR-830 and KHK-4083, respectively. Following favorable results from an exploratory study of GBR-830 administered via intravenous infusion [35], Glenmark has initiated a Phase II trial for a subcutaneous formulation in atopic dermatitis (NCT03568162). The company is also considering GBR-830 for an exploratory lupus study [36]. Like Glenmark, Kyowa Hakko Kirin is evaluating KHK-4083 in atopic dermatitis (NCT03703102) and, although Glenmark has the first-in-class product targeting this pathway, Kyowa may have an edge as their drug is a fully human monoclonal antibody that is also in Phase II for ulcerative colitis (NCT02647866) [37].

OX40 ligand is another target option for drug developers. As an example, oxelumab was under development by Roche and Genmab for allergic asthma (NCT00983658) and allergic rhinitis (NCT01152619), but has since been dropped from the pipeline [25]. Encouragingly, however, Kymab is moving their human monoclonal antibody which targets OX40 ligand, KY1005, into a Phase II atopic dermatitis trial (NCT03754309).

4. Discussion

Products targeting costimulatory pathways continue in clinical trials across a wide range of A/I indications, despite the mixed efficacy and safety outcomes of past candidates. Although approved over a decade ago for rheumatoid arthritis and more recently for psoriatic arthritis, BMS is still actively studying abatacept in this area, including investigating predictive biomarkers to maximize the drug’s benefit in sub-sets of patients [38,39]. Additionally, successful studies with a subcutaneous formulation have allowed the company to push the juvenile idiopathic arthritis label from 6 years down to a 2-year minimum age [40]. Recent trials are evaluating Sjögren’s syndrome (NCT02915159), idiopathic inflammatory myopathy (NCT02971683), myasthenia gravis (NCT03059888), interstitial lung disease (NCT03084419, NCT03215927), and Wegener’s granulomatosis (NCT00468208). One study gauged the ability of abatacept to prolong remission induced by ustekinumab in psoriasis patients (NCT01999868).

Previous study results show that abatacept was not effective in patients with inflammatory bowel disease [41]. BMS supported an investigator-initiated study that evaluated abatacept in diffuse cutaneous systemic sclerosis and, although the trial failed to demonstrate statistical significance on the primary endpoint (modified Rodnan skin score at 1 year), results favored abatacept numerically on several disease-related endpoints, and statistically on some key secondary outcomes [42]. Likewise, a study of abatacept in patients with active lupus nephritis did not achieve the primary endpoint of complete renal response at 1 year, however, earlier treatment responses were seen in proteinuria [43].

A study assessing patients with atopic asthma demonstrated that abatacept had no effect on the inflammatory response in the lung, on lung function, or on asthma control, however, the sample size of just 24 patients with mild asthma may not have been large enough to detect a strong signal in the treatment effect [44]. Looking at the drug in patients with ankylosing spondylitis did not show a major response either [45]. This study also had a small sample size of only 30 patients and, furthermore, did not have the gold standard blinded, randomized, placebo-controlled design. Additionally, in a recent study evaluating patients with relapsing-remitting multiple sclerosis, abatacept did not significantly reduce the average number of new inflammatory magnetic resonance imaging lesions compared to placebo, but this study only enrolled about half its target accrual, possibly contributing to the lack of observed treatment effect [46].

Another study demonstrated that the addition of abatacept to standard therapy reduced the incidence of severe acute graft-versus-host disease after human leukocyte antigen-mismatched unrelated donor transplant [47]. Also in the area of transplantation, BMS’ belatacept, with just a two amino acid difference from abatacept, was the first costimulatory blocker approved to prevent renal transplant rejection in 2011 [48]. Unfortunately, this product did not fare as well in a liver transplant study, demonstrating increased death and graft loss in the belatacept arms [49].

Many of the products listed in Figure 2 may have been dropped in additional indications after showing a lack of efficacy in results from a separate disease. For example, lulizumab pegol completed the failed lupus study (NCT02265744) around the time the Sjogren’s syndrome study was terminated for administrative reasons (NCT02843659). Likewise, Roche discontinued evaluation of oxelumab in asthma based on Phase II study data [50] and activity subsequently subsided surrounding the product’s evaluation in allergic rhinitis (NCT01152619). Alternatively, Astellas and Boehringer’s CD40 antagonists were not terminated in glomerulosclerosis (NCT02921789) and lupus nephritis (NCT02770170) after reporting negative results in psoriasis [26] and rheumatoid arthritis [31], respectively. Other products may have been affected by corporate licensing deals and development discontinued in certain indications for strategic purposes, such as FFP104, which was licensed [51] after PanGenetics terminated the psoriatic arthritis trial (NCT00787137), and prezalumab, which discontinued psoriasis (NCT01493518) and cutaneous lupus (NCT01389895) trials a year following the AstraZeneca partnership [16]. Finally, safety concerns were the source of the discontinuation of development for toralizumab and ruplizumab across a number of different diseases [22,23].

Figure 1. Costimulatory drug targets and products in phase II or higher development for A/I diseases. Important drug targets within these families that are being developed for A/I diseases. The B7 family includes activating signals, such as CD28 and induced costimulatory molecule (ICOS), as well as the inhibitory signal, cytotoxic T-lymphocyte-associated antigen 4 (CTLA4) [2].

Abbreviations: APC: antigen presenting cell, CTLA4: cytotoxic T-lymphocyte-associated antigen 4, ICOS: induced costimulatory molecule, L: ligand, MHC: major histocompatibility complex, TNF: tumor necrosis factorNote: The solid green lines designate activation secondary signals and dashed red lines designate inhibitory secondary signals.Source: Adapted from Pharmaprojects® (September 2018) [12] and Esposito et al. [1]

Figure 2. Clinical development, past and present, of costimulatory antagonists that have reached phase II, by A/I indications tested. The active and discontinued clinical development for drug antagonists, that have reached phase II at some point, and target some of the newer pathway costimulatory molecules: CD28, ICOSL, CD40/CD40 ligand, and OX40/OX40 ligand. The indications depicted include earlier phases of development where these drugs have been tested in patients. These Pharmaprojects® data reveal that the CD40/CD40 ligand target is very popular, having previously been or currently being evaluated across several different indications. Not surprisingly, lupus, an area where a high unmet need persists, is being evaluated by drugs that span across multiple costimulatory pathway targets.

Abbreviations: ICOS: induced costimulatory molecule, ITP: immune thrombocytopenic purpura, MS: multiple sclerosis, PBC: primary biliary cirrhosis, RA: rheumatoid arthritis, SLE: systemic lupus erythematosus, UC: ulcerative colitisSource: Pharmaprojects® (September 2018) [12]

Failures in drug development influence the direction of current products within the same pathways. Notably, progress in the costimulatory arena demonstrates a remarkable resilience of the industry to continue to pursue other avenues using the same drug target in the face of many early-stage failures. Companies are presently testing these products across multiple A/I conditions in order to hone in on areas where they might see the most success.

5. Expert opinion

Looking at past successes and failures identified strategies that have been improved upon in recent drug development. The adverse events seen in trials with anti-CD40 ligands informed further refinements to the currently active programs. For instance, dapirolizumab pegol and letolizumab are modified monoclonal antibodies designed to minimize thromboembolism observed in the ruplizumab and toralizumab trials. Drug development for CD40 ligand antagonists had switched to focus on immune thrombocytopenia, a disease marked by lower platelet counts. Strategies for developing direct antagonists of the CD40 receptor have undoubtedly evolved out of earlier trial failures with the ligand antagonists. Likewise, the disastrous Phase I theralizumab trial in 2006 surely influenced the pursuit of CD28 antagonists, which lack agonistic activity. Despite scientific advancements bringing about drugs adapted to circumvent adverse effects, some of the newer and more targeted drugs are failing to demonstrate efficacy in various autoimmune patients. The termination of BMS’ lulizumab pegol trials may threaten the upcoming Phase II trial with another anti-CD28 pegylated antibody, FR104. However, OSE Immunotherapeutics’ partner Janssen Biotech is expected to evaluate their drug in rheumatoid arthritis, as opposed to Sjögren’s syndrome or lupus, where it has demonstrated positive results in preclinical models of arthritis when compared to abatacept [4]. The CD28 antagonist also demonstrated positive results in preclinical models of psoriasis, but partner Janssen Biotech, being a Johnson & Johnson company, may want to avoid competition with their recently launched guselkumab and, perhaps, fill the gap left by the discontinuation of the sirukumab program in rheumatoid arthritis.

Overall, a resurgence in targeting CD28 brought about by the recent development of selective antagonists has already been met with the termination of one of only two of these antibodies to reach clinical development for autoimmune conditions. Another B7-family target is almost equally inactive. Amgen, and partner AstraZeneca, are the only sponsors to have taken an ICOS targeting product to clinical development in patients with autoimmune diseases. Setting aside lupus and psoriasis to focus on Sjögren’s syndrome with prezalumab, these partners remain confident in this pathway and even have a bispecific antibody-peptide conjugate targeting BAFF and ICOS ligand that is to be evaluated in rheumatoid arthritis.

In contrast, the TNF-related family of costimulatory drug targets has seen many products tested in autoimmune patients. Drug developers may be wary of the dual role B7-family molecules play in the immune system, such as the activation of suppressive regulatory T cells by CD28 and ICOS [2]. Within the TNF-related family, CD40 and CD40 ligand antagonists have reported positive results from studies evaluating Sjögren’s syndrome, lupus, and transplant rejection while OX40 has demonstrated proof-of-concept as a target in atopic dermatits. Initial results show promise in the strategy of engineering anti-CD40 Fab’ antibody fragments to avoid the thromboembolic events seen in CD40 ligand inhibition using full-length immunoglobulin G1 antibodies [20]. The number of anti-CD40 antibodies, as well as the wide range of indications they target, suggest that CD40 biologics may be competing to be the next major novel class of costimulatory inhibitors to treat autoimmune disease. Evaluating safety as sponsors move through clinical trials with these drugs will be critical to gain the confidence of regulatory authorities, physicians, and patients in the long term.

Declaration of interest

P Spicer and L Runkel are full-time employees at Informa. 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. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose

Acknowledgments

The authors thank Gunilla Sturrock and Dorothy Shin for reviewing the article.

Additional information

Funding

This paper was not funded.