by Ayisha Sharma in London and Manasi Vaidya in New York.
Anti-granulocyte-macrophage colony-stimulating factor (GM-CSF) antibody treatments in development for Covid-19, such as by Roivant Sciences, Humanigen) and I-Mab Biopharma, are expected to have better therapeutic effects than cytokine-specific treatments such as interleukin (IL)-6 and IL-1 inhibitors.
As these antibodies target an earlier point in the inflammation pathway than inhibitors that target cytokines like IL-6 and IL-1, the former carries the potential to have a broader and better therapeutic response. It is also unclear which cytokines are at play at a certain point in the Covid-19 pathogenesis, so an upstream treatment of this kind that is not limited to targeting just one cytokine could have higher chances of success.
As for the ideal design of an anti-GM-CSF antibody Covid-19 trial, crucial primary endpoints that are also reflective of IL-6 inhibitor studies should include the incidence of patients moving onto mechanical ventilation, the incidence of hospitalized patients going to the ICU and mortality numbers. Some of the ongoing anti-GM-CSF antibody studies have used such measures. The number of ventilator-free days at 28 days is also a strong measure as it captures the data from various other important endpoints. The use of biomarkers like ferritin, cytokine levels and C-reactive protein can determine the potential responders.
Even though past data show GM-CSF-targeted treatments in acute respiratory distress syndrome (ARDS) have been unsuccessful, it is unclear on the impact of these results on Covid-19 studies. Covid-19-associated pneumonia and ARDS may be unique compared to ARDS caused by sepsis or trauma.
In terms of anti-GM-CSF investigations, Kinevant Sciences, a subsidiary of Basel, Switzerland-headquartered Roivant Sciences, has a Phase II GM-CSF BREATHE trial with gimsilumab (NCT04351243). Other anti-GM-CSF antibody studies include Humanigen’s lenzilumab Phase III (NCT04351152) trial, and I-Mab’s Phase I/II TJ003234 trial. Oxford, UK-based Izana Biosciences has an expanded use program for its antibody namilumab in two centres in Italy, and is about to enter into a major UK trial as part of the national phase strategy and is in discussions with international partners on a large pivotal study, said a company spokesman. Kiniksa Pharmaceuticals is also evaluating a Phase II/III program for mavrilumab, as per a 31 March press release.
Kiniksa, Roivant and I-Mab did not respond to requests for comment before press time.
Upstream mechanism of anti-GM-CSF antibodies to outperform IL-6 inhibitors
GM-CSF is an inflammatory mediator and a stimulator of other cytokines like IL-1, IL-6 and TNF, said Dr Gailen Marshall, director, Division of Clinical Immunology and Allergy, University of Mississippi Medical Center, Jackson, who is an investigator on the 144-patient Phase I/II (NCT04341116) TJ003234 study. The anti-GM-CSF mechanism targets an earlier point in the activation pathway of the cell compared with mechanisms that target specific cytokines such as IL-6 inhibitors, said Dr Carlo Perricone, rheumatologist, Department of Medicine, University of Perugia, Italy, adding the earlier action should prevent IL-6 production. This is why anti-GM-CSF treatments can be used at an earlier stage in hospitalised Covid-19 patients, while IL-6 inhibitors are mainly used as rescue therapy for Covid-19 patients on invasive mechanical ventilation, he explained.
It is difficult to differentiate Covid-19 patients based on which cytokine is more important at the time of treatment—IL-1, IL-6 or IL-33—so an upstream treatment is likely to target more patients, said gimsilumab investigator Dr Gerard Criner, chair and professor, Thoracic Medicine and Surgery, Lewis Katz School of Medicine, Temple University, Philadelphia, Pennsylvania. As the target of anti-GM-CSF antibodies is broader, the clinical responses are thought to be better, Perricone noted. Previously published data supports that an anti-GM-CSF approach should have a better effect than IL-6 inhibitors, as animal models have shown a GM-CSF blockade decreases downstream IL-6 and many other pro-inflammatory cytokines, a Humanigen spokesperson said.
GM-CSF is necessary to trigger granulocytes and macrophages stimulated as part of an immune response, which then releases inflammatory cytokines leading to cytokine storm syndrome (CSS), said Marshall. Macrophages and neutrophils are major components of the Covid-19 disease pathogenesis, and it is unlikely that only one cytokine is at play, Perricone said. The rationale for targeting GM-CSF is to delay or prevent rapid deterioration in severe Covid-19 patients by inhibiting the cytokine storm, said the Izana spokesperson.
Producing cytokines is a normal response to infection which may not always be negative, said Dr E Scott Halstead, associate professor, Division of Critical Care Medicine, Pennsylvania State University, Hershey. Inflammation in Covid-19 patients appears controlled and patients are not dying simply due to it, but also due to endothelial activation, ARDS and vascular complications such as strokes and pulmonary embolism, he explained.
While the well-established efficacy and safety profile of IL-6 inhibitors adds confidence to their mechanism, they inhibitors could nonetheless produce unexpected side effects in Covid-19, Perricone said. It is best to test anti-GM-CSF drugs in moderate-to-severe Covid-19 patients, as these treatments are known to produce side effects such as infection which could endanger patients with milder symptoms, Perricone said.
The TJ003234 trial is in subjects 60 years of age or over with severe coronavirus disease such that at rest, finger blood oxygen saturation ≤ 93% or PaO2/FiO2 ≤ 300 mmHg or they require non-invasive or invasive mechanical ventilation, according to ClinicalTrials.gov. The gimsilumab trial is in subjects with lung injury or ARDS secondary to Covid-19 and the lenzilumab trial is in hospitalised patients with Covid-19 pneumonia.
Mortality and ventilator-free days are relevant endpoints
While the different endpoints in the studies could make a cross-trial comparison of these antibodies difficult when the primary and secondary endpoints are seen in combination, there is an overlap, said Marshall.
Patient deterioration in anti-GM-CSF trials for Covid-19 can be measured through endpoints such as the incidence of patients moving onto mechanical ventilation, the incidence of hospitalised patients going to the ICU and mortality numbers, Perricone said. The gimsilumab study has a primary endpoint of mortality at day 43, while the lenzilumab trial’s endpoint is the incidence of invasive mechanical ventilation and/or mortality over a 28-day time frame that excludes patients requiring invasive ventilation on hospitalisation.
While the TJ003234 trial has a co-primary endpoint of treatment-emergent adverse events and the proportion of subjects experiencing deterioration in clinical status according to the 8-category ordinal scale, the latter’s timeframe of just 14 days is too short, Perricone explained, adding 30 days would have been sufficient.
Mortality is a key endpoint, Halstead agreed. The number of ventilator-free days at 28 days is also a strong measure because it acts as an umbrella for several other endpoints, such as whether a patient progresses to mechanical ventilation, how long they stay there and whether they die on it, he explained.
Clinically significant results would see at least 50% drops in all three of his preferred endpoints in the anti-GM-CSF arm of a controlled trial, Perricone said. But 50% is quite a high benchmark given the heterogeneity of the baseline characteristics of Covid-19 patients, including age and comorbidities, Halstead noted. The aforementioned trials are recruiting male and female patients of at least 18 years, with the gimsilumab trial capping age at 79 years, according to ClinicalTrials.gov.
However, experts said the trials characterised disease severity in different ways with no one standardised measure. Severity can be defined based on patients’ respiratory states, as measured by Sp02, pO2 and CO2 levels or requirements for high oxygen flow, continuous positive airway pressure (CPAP) ventilation or other noninvasive forms of ventilation, said Perricone. Namilumab will be ideally evaluated in hospitalised patients to delay or prevent the rapid deterioration and requirement for critical care and invasive ventilation, said the Izana spokesman. Any investigational trials focusing on intubated patients with an endpoint of ventilator-free days may tell whether that stage is too late for an intervention, said Marshall.
Identifying ideal patients for treatment with an anti-GM-CSF antibody is a challenge, however, since not everyone who gets the infection gets full-blown CSS, even if they develop pneumonia and have trouble with oxygen, said Marshall.
The BREATHE study is enrolling patients with elevated serum C-reactive protein (CRP) or ferritin, while the TJ003234 trial is looking at the impact on cytokine levels as per its trial description. Using biomarkers like CRP and ferritin to assess the therapy efficacy, and ideally individual cytokines or chemokines to identify responders, would be ideal, but those test panels don’t have a quick turnaround, said Criner. The anti-GM-CSF approach may not work in all Covid-19 patients, but recognising those that it may work in could be done by reviewing systemic inflammatory markers and chemokines, said Marshall. Patients at Criner’s hospital get high-resolution CT scans once admitted, which helps them make an educated guess on how important any anti-inflammatory like an anti-GM-CSF therapy could be in each patient, he said.
Debate on past anti-GM-CSF ARDS data for Covid-19
While GM-CSF has been evaluated as a target in past ARDS studies, the results have not been positive, experts said. In one example, a randomised, placebo-controlled Phase III trial (NCT00201409), 130 ARDS patients were given an intravenous infusion of anti-GM-CSF but it resulted in no difference in ventilator-free days between the GM-CSF group and the placebo group (Paine, R. et al. Crit Care Med. 2012; 40(1):90-7). Preclinical mouse studies have indicated the lung-protective properties of GM-CSF in influenza models, which makes the anti-GM-CSF approach unclear, Halstead said. Previous ARDS trial data could well indicate that treating Covid-19 with anti-GM-CSF drugs may not be the best idea, he added.
However, while in the past GM-CSF was trialled for its lung protective properties, blocking it to address the CSS leading to ARDS is rational, said a researcher involved with the Phase III study. Covid-19-associated acute lung injury may be different than other types of injury and despite no difference being observed in that 130-patient trial, it is possible that blocking GM-CSF may have some therapeutic value based on its role in the inflammatory pathways, said the researcher. With Izana’s namilumab, the aim is to bring abhorrently high GM-CSF signalling down to baseline, leaving basal levels of activity intact so it would not inhibit the normal functions of GM-CSF, said the company spokesman.
When compared to other kinds of pneumonia, Covid-19 presents a much more hyper-inflammatory and macrophage-deficient state, Perricone explained. Covid-associated pneumonia is likely a severe inflammatory state compared to other pneumonia, which makes the investigation of these anti-GM-CSF antibodies different from past approaches, said Criner.
Manasi Vaidya and Ayisha Sharma are Reporters for Clinical Trials Arena parent company GlobalData’s investigative journalism team. A version of this article originally appeared on the Insights module of GlobalData’s Pharmaceutical Intelligence Center. To access more articles like this, visit GlobalData.