31 Januari 2022 Episode 233 : Blocking GM-CSF as a therapeutic strategy to mitigate the inflammatory COVD phase

Dear colleagues,

For this episode, I will focus on the potential benefit of blocking the activity of an “upstream” pro-inflammatory cytokine, called granulocyte-monocyte colony-stimulating factor (GM-CSF). 

As the name suggests, it was originally discovered as a factor to stimulate early myeloid development, but it acts also on more mature neutrophils, monocytes, macrophages, enabling their survival, proliferation and differentiations.

As such, GM-CSF is used to grow monocyte-derived dendritic cells in immune therapeutic trials and  it is used as an “adjuvant” in experimental vaccines.

However, GM-CSF apparently also has a role in pathological inflammation processes and therefore GM-CSF antagonists (i.e. monoclonal antibodies neutralizing GM-CSF itself or blocking the cellular receptor) are being used in auto-immune-inflammatory diseases, such as rheumatoid arthritis.

COVID pneumonia is characterized by an infiltration of neutrophils and pro-inflammatory monocytes and macrophages, which creates an alveolitis that strongly impairs respiration.  Therefore, it was a logical step to investigate whether GM-CSF antagonists could be beneficial to counter this process, reduce the inflammatory markers, the need for mechanical ventilation and improve survival.  

Par 1 Principles

Ep 233-1: Plaçais in Muc. Imm. Oct 2021 Immune intervention in COVID-19 is a matter of timing

This review leaves anti-virals out of the picture. It first proposes a schematic time course of COVID

And a scheme of immune responses

Followed by immune interventions

As can be seen, type 1 interferon and anti-Spike mAb should be applied very early to block the virus. Anti-cytokine therapies are proposed later, at the time of symptoms onset and Dexamethasone only at the time of clinical deterioration, when the hyper-inflammation is already in progress.

Clearly, there is a time-balance to strike:

• Generalized immunosuppressive treatment (such as corticosteroids) should not be started too soon, because they might suppress the initiation of adaptive immunity, mainly T cells that clear the viral infection; while applying it too late might no more be able to avoid immune exhaustion and tissue damage. Therefore dexamethasone should be given when hyper-inflammation becomes clinically (and biochemically evident).   Dexa was not favorable in patients with dyspnea, but without oxygen need, it was harmful if  C-reactive protein levels < 100 mg/L, but beneficial if CRP was > 200mg/L.
• Anti-Cytokine treatment could be started earlier to block the onset of the hyper-inflammation.  These include monoclonal antibodies against the inflammatory cytokines interleukin-1 and -6 and against GM-CSF.
• In this context, GM-CSF (granulocyte-monocyte-colony-stimulating factor) is considered as an “upstream” factor, because it promotes the proliferation and activation of the inflammatory neutrophils and monocytes/ macrophages (including alveolar macrophages).  Hence blocking of GM-CSF should occur rather early.     

Ep 233-3: Bonaventura Front Imunol 2020 explains the (patho-)physiology in depth. 

• In fact, these authors first raise the possibility of using GM-CSF itself (Sagramostim) to stimulate the function of alveolar macrophages (panel B), but that idea has been abandoned:  at the symptoms onset, these cells are already over-activated and there is influx of neutrophils and monocytes into the lungs. 
• This leads to a “positive feed-forward loop”: GM-CSF leading to more inflammation (panel A).
• Therefore, the more rational approach is to block GM-CSF by monoclonals, either by direct “neutralization” of this cytokine (Gimsilumab, Lenzilumab, TJ003234,Namilumab or Otilimab) or indirectly by blocking the receptor by Mavrilimab.  Several of these monoclonals have been used previously with some success in the treatment of severe rheumatoid arthritis.

Par 2 Three early open label trails

Ep 233-3: De Luca Lancet Rheumatology 2020: Mavrilimumab in 13 treated (mean age 57 years) versus 26 matched controls (mean 60 years) all non-mechanically ventilated patients, apparently mostly in WHO class (5 and 6).

As can be seen, the Mavrilimumab-treated subjects, clinically recovered better and showed also more pronounced reduction of C-reactive protein inflammation marker.

Ep 233-4: Temesgen in medRxiv 2020: First trial with Lenzilumab in 12 patients with confirmed

COVID pneumonia, but without a control group. The authors show: improved clinical outcomes, oxygen requirement, and cytokine storm, but very difficult to interpret in the absence of controls.

Ep 233-5:  Fisher Lancet Resp Med 2021 CATALYST:  Namilumab or infliximab (= anti-TNF) compared with standard of care in hospitalised patients with COVID-19 an open phase 2 proof of concept trial

Primary outcome = inflammation: Namilumab, but not Infliximab, showed proof-of-concept evidence of reduction in inflammation—as measured by CRP concentration

Secondary outcome: trends to shorter stay at ICU or in hospital and less deaths in Namilumab vs standard care, but not powered (see Tables 2 and 3 p. 9)

Note: most (90 %) patients on steroids and about half on Remdesivir.

Par 3 Randomized Controlled Phase 3 trials

Ep 233-6: Patel medRxiv April 2021 Otilimab in Severe COVID-19 Pneumonia (OSCAR)

Multicenter trial with 806 patients with severe COVID (77% WHO 5 and 22 % 6) see Table 1 p. 26.

Result:

• Overall: 71% of patients receiving otilimab were alive and free of respiratory failure at Day 28 versus 67% receiving placebo = NS
• However: patients > 70 yrs had a reduction of > 19 % alive and free of respiratory failure at day 28 and > 14 % all-cause mortality at day 60

Prospect: To be confirmed in patients aged ≥70 in Part 2 of this study

Ep 233-7: Part 2 prematurely stopped and GSK announces to stop the development of Otilimab for COVID, but continuing it for rheumatoid arthritis and Sotrovimab (broad neutralizing anti-Spike) for COVID. 

Ep 233-8: Cremer Lancet Resp Med 2021 Mavrilimumab in patients with severe COVID-19 pneumonia and systemic hyperinflammation (MASH-COVID)

Patients: hospitalized with pneumonia, hypoxemia, CRP > 50 mg/L , but no mechanical ventilation (WHO 5 and 6?)

Concomitant medications included antiviral drugs related to COVID-19, corticosteroids, convalescent plasma, other immunosuppressive agents, and antimicrobial drugs related to non-COVID-19 infections. About 70 % in each group received corticosteroids

Trend for better outcome at day 14, but NS: group too small 

Ep 233-9: Temesgen Lancet Resp Med Dec 2021:  Lenzilumab in hospitalised patients with COVID-19 pneumonia (LIVE-AIR):

Patients: mostly on corticosteroids and remdesivir

Result:  Survival without invasive mechanical ventilation to day 28 was achieved in 198 (84%;) participants in the lenzilumab group and in 190 (78%) patients in the placebo group (p 0.04).

Ep 233-10: Temgesen medRxiv Jan 2022 Further multivariate statistical analysis of LIVE-AIR study

Survival without invasive mechanical ventilation (SWOV)

• If baseline CRP<150 mg/L: 152 (90%; 85to 94) lenzilumab and 183 (79%; 72 to 84) placebo and its likelihood was greater with lenzilumab than placebo (HR: 2.54;; p=0.0009)

but not in participants with CRP≥150 mg/L at baseline.

Implication: GM-CSF best in early phase (immune intervention 2 or WHO clinical 2-4), whereas in later phases, the anti-IL-1, anti-IL-6 and JAK/STAT inhibitors are most indicated.

Ep 233-11:  Economic benefit of Lenzilumab

Ep 233-12: Guan medRxiv Jan 2022 Meta-analysis on GM-CSF antagonists in severe COVID

= summary of the studies above

1. Small, but significant reduced risk of mortality

2. Risk reduction on mechanical ventilation
3. Improvement of ventilation

4. No significant difference in secondary infections

Conclusions:

Taken together, evidence that GM-CSF antagonists may be beneficial with regard to survival and improvement of respiratory function,

• Without increasing the risk for secondary infections (as was seen in trials anti-IL6).
• Maybe subgroups with lower CRP and higher age may benefit more?

However, there are limitations:

1. Low quality evidence bc of limited numbers
2. Differences in study population between studies
3. Differences in “standard of care” e.g. nursing, ventilator availability
4. Potential interactions with corticosteroids, vaccines…

Personal reflections: How can we put these results on GM-CSF inhibitors into the perspective of evolving treatment for COVID-19?

1. Presently clinically available antivirals

1. 1. The major breakthrough is the orally available Pfizer protease inhibitor Paxlovid.   The drug cuts the risk of hospitalisation or death for high-risk patients by 88%, compared with the placebo, if given within 5 days of symptom onset. (see Ep 233-13, still not a formal publication, but based on press-release).  The disadvantage is the associated Ritonavir, which  inhibits drug-metabolizing liver enzymes and therefore can increase drug toxicity of statins, steroids, sedative hypnotics, anticoagulants, and antiarrhythmic therapies (Ep 221-9).  

1. 2. With regard to the polymerase inhibitors, Remdesivir has gone through a “long and windy road”.  According to a meta-analysis of 5 RCT with over 7000 patients, it has no significant effect on mortality in hospitalized patients (see Ep 233-14).  

A very recent Canadian study (Ep 233-15) shows again a marginal but non-significant decrease in mortality, but significantly lower need for mechanical ventilation….

On the other hand Burki (Ep 233-13) states that treatment with remdesivir within 7 days of the appearance of symptoms reduced the chances of hospitalisation or death by 87%, compared with the placebo, for non-hospitalised patients with COVID-19 who were at high risk of severe disease, but the reference is not given.  In any case, a major disadvantage for out-of-hospital treatment is the intravenous application.  

1. 3. In this regard Molnupiravir has the advantage of oral administration, but the effect is moderate: treatment with molnupiravir within 3 days of symptom onset reduced the chances of hospitalisation or death by 30%, compared with the placebo (Ep 233-13).  And there are concerns about the mechanism of action, induction of mutations, which raises the possibility of carcinogenic potential.   

1. 4. Finally, there is a series of anti-Spike monoclonal antibody (combinations) with proven effect on avoidance of hospitalization, if applied early enough.  However, unfortunately, most are no longer active against omicron.  Only Sotrovimab retains full activity. The disadvantage remains their intravenous application.

2. Imunomodulators

Many modulators have been proposed, the first and widely used class are the glucocorticosteriods (dexamethasone, methylprednisolone and hydrocortisone) and the second class are the more specific blockers of IL-1, IL-6, JAK/STAT.  The most experience is with dexamethasone and anti-IL6

In this regard, two recent reviews are of interest

Ep 233-14: Crichton et al focus on a limited number of well-designed RCT investigating the effect on mortality in hospitalized COVID patients.  This is the summary:

For corticosteroids: The pooled effect estimate showed OR 0.80 (95% CI 0.64–1.01, p=0.06), indicating a mortality benefit that narrowly failed to meet the pre-specified level of statistical significance. There was no difference between dexamethasone, methylprednisolone and hydrocortisone.

However:   No benefit in those patients not requiring oxygen (rate ratio 1.19, 95% CI 0.92–1.55)

For anti-IL6 receptor: The pooled effect estimate was OR 0.87 (95% CI 0.79–0.96, p=0.007), showing

significant reduction in mortality with anti-IL-6 treatment.

Similar effect in patients not receiving ventilation support and non-invasive ventilation support, with a suggestion of lower effect in those invasively ventilated.

No difference between Tocilizumab or Sarilumab.

Ep 233-16:  A larger meta-analysis, including also case-control studies suggest that Tocilizumab is better than corticosteroids and that there is a benefit of combining both.   

So, where is the place of anti-GM-CSF treatment?

As explained, most patients in anti-GM-CSF trials (also in the control arm) were also treated with corticosteroids, therefore it is already evident that the small, but significant effect comes on top of the steroid effect.

Theoretically, GM-CSF is driving the inflammation, upstream of mediators such as IL-1, IL-6 etc., it might either replace those or be additive or have a better effect, by blocking the production of these mediators. 

Proving this point, however, would require large scale RCT….

Best wishes,

Guido