SARSCoV2 (COVID-19) inflammatory pathway and potential for repurposed therapeutics
John P. Hussman, Ph.D., April 3, 2020
The following is drawn (with periodic updates to reflect current research) from correspondence sent to research colleagues, medical institutions, and contacts at NIH/NIAID.
Jump to section:
Notes on CoV molecular pathway and acute respiratory distress (Hussman)
Interim guidance on repurposed therapeutics in the inflammatory pathway
April 3, 2020
Good morning all,
A month ago, I wrote “As of March 1, 2020, the growth rate of COVID-19 cases in countries other than China is running at 25% daily. The expansion in the number of locations is also problematic, because the first person in each new location is local ‘patient zero.’ Unless serious containment efforts reduce this growth rate substantially, a continuation would produce a 1000-fold increase in cases over the coming month. Even that might not seem so bad, until the end of April, when the same rate of growth would push the number of COVID-19 cases beyond the 45 million U.S. flu cases observed in 2017-2018. That is exactly how this little coronavirus with twice the R0 of the seasonal flu, and a mortality rate that is evidently an order of magnitude higher, will produce utter chaos if containment efforts are not taken seriously.”
I do believe that recent containment efforts have brought prospective end-of-April U.S. reported cases into the range of 2-10 million, based on current growth rates, with the potential for further reductions if these efforts persist. Still, a best-case estimate for U.S. fatalities appears to be near 60,000 (based on my own adaptive model and others). In the event these containment efforts are relaxed, the prospect of 100,000 U.S. fatalities is not one that we would face “in the coming months” but potentially in the coming weeks.
Even if daily case growth peaks in mid-April, the majority of U.S. fatalities appear likely to be ahead rather than behind us. My hope is that an urgent request might help to blunt the otherwise hard impact that appears likely in the coming days.
We arrived in this crisis with a rich literature detailing aspects of the molecular pathway affected by prior CoV serotypes (SARS, MERS) as well as ARDS (acute respiratory distress syndrome). Much of the emerging literature on SARSCoV2 is strongly consistent with these findings. In the absence of clear evidence from randomized, controlled clinical trials, the patients currently under ICU care need therapeutic options that are informed by the biological pathway that can already be inferred from the literature, as well as interim odds-ratios from new research and ongoing clinical trials.
My view on noise-reduction has always been that common signals drawn from multiple, weakly correlated sensors can be still very informative, even when stronger signals (in this case RCTs) remain unavailable. This noise-reduction perspective may be useful when therapeutics reported in smaller clinical studies have a mode of action that overlaps plausible biological pathways.
To the extent that all of us are trying to use our “particular set of skills” in this crisis, mine are in statistical genetics and molecular pathway analysis. I have attached my updated pathway notes, intended to connect the dots in the existing research literature, in hope that they might be useful in provoking thought among front-line health professionals.
Notes on CoV molecular pathway and acute respiratory distress (Hussman)
Studies may not be specific to SARS-CoV-2. Reported interventions remain suggestive until demonstrated in RCTs. Numbers PubMed IDs. Those with names pre-publication (medrxiv.org)
- T-cell dysregulation: Novel CoVs cause fatal acute lung injury via T-lymphocyte dysregulation and cytokine driven inflammation (32161940). As T-cell disruption can be associated with thrombocytopenia, this mechanism seems consistent with the five-fold increase in mortality where low platelet count is observed (32178975). Suggests potential to repurpose therapeutics that act on proinflammatory response. Favalli (32205186), though at first glance appearing to address rheumatoid arthritis, aligns nicely with this pathway and also discusses RA therapeutics in addressing hyperinflammation in SARS-CoV-2.
- ACE2: The RBD of CoV uses membrane-bound ACE2 to gain access to respiratory cells. ACE2 is highly expressed on alveolar type-II pneumocytes (32310915), which serve as immune cells of the alveolar epithelium and respond to infection by producing cytokines including TNF-α, IL-6 and IL1-β (23383221). SARSCoV2 preferentially infects type-II cells, resulting in a marked increase in cytokine expression, including a 10-fold increase in mRNA encoding IL-6 (23418343). ACE converts angiotensin I (Ang I) to Ang II, ACE2 degrades Ang II to Ang(1-7). While early evidence did not suggest an effect of ACE inhibitors (i.e. blocking Ang I =>Ang II), or Ang II receptor blockers (ARBs) in CoV19 (32120458), more recent findings (e.g. Yingxia Liu, G. Yang) report protective effect of ARBs (OR 0.343) among elderly. Use of exogenous Ang(1-7) reduces inflammation and improves lung function in ARDS models (22009550). Recombinant ACE2 also suggested (29237475, 17558469).
- IL6 axis inhibition: IL-6 levels predictive of respiratory failure in CoV19 (Herold, Zhang). SARS associated coronavirus produces a high level of proinflammatory cytokines (21964025), which is dependent on the presence of RELA (p65) and IkB kinases, and is dependent on p65 phosphorylation via IkB-α degradation (17532082). The viral nucleocapsid protein of SARS-CoV also promotes IL6 via NFkB activation (17490702). Inhibition of NFkB activation may also increase survival (24198408). Inhibition of JAK signaling may block p65 phosphorylation and attenuate proinflammatory cascade (28642467). IL-6 concentrations are increased 2.9-fold in patients with complicated CoV19 vs uncomplicated (Coomes). Increased IL-6 is an early indicator of cytokine release syndrome in CoV19 patients (Wang). Use of low molecular weight heparin reported to be associated with improvement in aberrant coagulation and reduced IL-6 levels (Shi) and may increase survival (Negri, 32220112). Tocilizumab, a well tolerated blocker of the IL-6 receptor, may have potential to dampen cyotokine release syndrome in CoV19 (Zhang). Because catecholamines augment the production of IL-6 and other inflammatory cytokines, α1 adrenergic receptor antagonists (e.g. prazosin) may provide prophylactic benefit against cytokine storm (Konig).
- TNF-α inhibition: The SARS spike protein induces (TNF-α converting enzyme) TACE-dependent shedding of the extracellular ACE2 domain. NL63-S, a CoV common cold serotype, does not induce similar ACE2 shedding or TNF-α (18490652). TACE antagonists may block this pathway in SARS-CoV and attenuate disease severity (14741070). T-cell profiling of CoV19 patients shows an inflammatory signature with contributing cell-cell interactions mediated largely by TNF/LTB, IFNγ and IL-17 related genes (Wen). TNF-α downregulates FCγIIB expression. In presence of IL4 & IL13, TNF-α skews FcR to inhibitory phenotype and production of alternatively activated macrophages (w/regulatory and wound-healing functions). TNF-α + IL10 synergistically upregulate FCγIIA (15703199). Interestingly, FCγIIB is selectively upregulated in dendritic cells from RA patients with quiescent disease (19734236). TNF-α inhibition has been suggested as a potential therapeutic in SARS-CoV (14741070).
- FcR modulation: Antibody-dependent enhancement (ADE) may be triggered by antibodies against the CoV spike protein (anti-spike S-IgG) (25073113). ADE is dependent on activation of FCγ receptor II. Among FcR subtypes, FcγR IIA (CD32A) appears to mediate infectivity most efficiently (21775467). Neutralizing antibodies can bind to the spike protein and enable alternative entry into FcγR IIA expressing cells. Spike-based subunit vaccines lacking RBD appear desirable (31826992). SARS macaque models produce skewed inflammatory cytokine production (including chemoattractants IL8 and MCP1) and absence of wound-healing response similar to that observed in fatal human cases. Blockade of activating subtypes of FCγR (e.g. ITAM-containing FcγR I and/or FCγR IIA) reduced these effects (Liu). Blockade of FcR activation via IVIG has been suggested for severe pulmonary inflammation and lung injury in SARS-CoV-2 (32125642). The anti-inflammatory effect is associated with its ability to recruit surface expression of the inhibitory Fc receptor FcγR IIB (11161202). Among potentially repurposed therapeutics, IVIG is not without dangers (renal, thrombosis), and effectiveness is not established in MERS (28864360). Alternatively, human polyclonal immunoglobulin G from bovines has been reported inhibit MERS-CoV in vivo (26888429). FcγRIIA-R/R131 (rs1801274) & CD14-159CC (rs2569190) are risk-genotypes both for severe SARS (17913858, 16185324) and related hyperinflammatory conditions (18180796, 14597109, 21193288, 16387800, 30691461), suggesting investigation as potential factors in CoV19. 23andMe genotypes rs1801274 directly. CD14-159CC (rs2569190) is in a strong LD block (we checked) extending ~10kb that includes rs201979630, rs201545309, rs143761293 in Europeans.
- Anti-rheumatics (DMARDs): Use of chloroquine phosphate/HCQ in SARS-CoV-2 has been reported (32074550, 32075365, 32070753) but should be evaluated carefully, including control & ex-post case exclusion. Identified as a candidate in screening of FDA-approved compounds in MERS-CoV (24841269). No clinical benefit reported in recent studies (Shamshirian), Magagnoli). Risks: retinopathy, arrhythmia. Addition of azithromycin is reported to be associated with increased risk of heart failure and cardiovascular mortality (Lane). HCQ is in use as a DMARD in (mild) arthritis. Putative mechanisms in SARS-CoV-2 include replication (15351731), ACE2 binding interference (16115318), endosomal acidification (28596841), and reduced pro-inflammatory CD4+ Th signaling (28169350). Antiviral/anti-inflammatory activity also reported for artemisinin/artesunate (31484283, 32105722, 25973011, 30334416, 29305894, 19699744). Some artemisinin-based compounds suggested to have activity against certain enveloped single-strand RNA viruses, and binds to RELA (p65) to downregulate NFkB activation (29305894).
- IFN1: During the initial replication stage, IFN-1 may act to reduce viral load by upregulating STAT1 (a differentiator of Th1 cells). STAT1 deficiency increases mortality and skews immune response toward a pro-fibrotic Th2-biased profile (20702617). STAT1 expression is protective (20386712). However, later delivery of IFN-1 may amplify risk by elevating pro-inflammatory response (26867177, 31355779).
- Steroids: Early steroid treatment (prednisone) discouraged, may have narrow use after progression to respiratory distress. Prolonged use without antimicrobials is discouraged (17597972).
- Progression: The acute stage of respiratory failure may involve an enhanced Th2 response (15784184, 20702617), with progression to fatal disease associated with expression of Th2 cytokines (18832706). The Th2 transducer STAT6 is required for the development of alt-activated macrophages that may exacerbate severity. In mice, STAT1/STAT6 double knockouts do not develop fibrosis (23015170).
- Risk factors: Age, cardiovascular disease, diabetes, smoking, maximum body temperature at admission, chronic respiratory disease or failure, albumin, and C-reactive protein cited as risk factors for progression (32091533, 32118640).
Interim guidance on repurposed therapeutics in the inflammatory pathway
On the weight of published research regarding SARS, MERS, ARDS, and the emerging literature on SARS-CoV-2 my impression is that one of the reasons SARS-CoV-2 has such a high fatality rate is that it is being treated in the context of pulmonology when it should also be treated in the context of rheumatology.
A recent survey reported that steroid use may be the most common therapeutic approach at present, even though research on prior CoV serotypes discourages steroids except in the most acute stages of inflammation. The appropriate therapeutic modality may instead require downregulating pro-inflammatory response, and part may involve antibiotic support to exogenously manage any resulting increase in the risk of secondary bacterial infection.
The request here is straightforward. That NIH/NIAID issue interim guidance on repurposed therapeutics now, on the basis of existing research and any preliminary findings and odds ratios available from ongoing RCTs. I emphatically do not think we have the luxury of waiting. The guidelines can and should be revised and updated regularly.
At present, the existing guidance from the CDC is limited to remdesivir and hydroxychloroquine, with no stepwise protocol or clear biological rationale. Even the existing literature on prior CoV serotypes provides clarity about the likelihood of T-lymphocyte dysregulation and cytokine storm, and broader potential options to address it.
The following is a suggestion of the form of the guidance needed here, not necessarily the substance - particularly to the extent that it might be revised or augmented based on preliminary evidence from ongoing randomized controlled clinical trials. I trust that we all have a very strong preference for RCTs over empiric treatment. The problem is that we have run out of time.
Emphatically, this is not a call for the use of experimental therapeutics, but rather those that are already FDA-approved, widely prescribed (though typically in the context of rheumatic or autoimmune conditions), and generally well-tolerated.
NIH/NIAID guidance could take a form something along the following lines. Again, the following suggests form and not necessarily substance, but it is based on molecular pathways inferred from the existing research literature (see pathway notes above for study references):
1) Brief discussion of potential inflammatory mechanisms, possibly including overlap of these mechanisms with pro-inflammatory mechanisms in RA, lupus, and immune-related conditions, and the exigency of informed empiric treatments based on likely biological pathways and preliminary odds-ratios, even while RCTs are still underway
2) Guidance or contraindications on candidate therapeutics, ideally with research references or brief discussion of the rationale for each in the context of cytokine driven hyperinflammation, placing milder and well-tolerated interventions first. Anti-viral therapeutics (remdesivir, lopinavir-ritonavir) might be included in this list as preliminary results from ongoing trials become available, as would interventions such as low molecular weight heparin.
a) ACE inhibitors/ARBs. Despite initial questions (possibly reflecting confusion between ACE and ACE2, which is downstream in the RAS system), findings to-date indicate that patients using ACE/ARB medications should continue using them. Indeed, recent evidence suggests that ARBs (e.g. losartan) may be protective.
b) Hydroxychloroquine (cautions and contraindications): Given the role of HCQ as a weak DMARD (disease-modifying anti-rheumatic drug), with reported effects on replication, ACE2 binding, endosomal acidification, and CD4+ Th signaling, a potential impact in COVID-19 is not wholly implausible, however several reports (as of 4/20) show no clinical benefit. To the extent that HCQ is being frequently used, and several reports are of questionable merit (e.g. ex-post case exclusion), risks such as retinal damage and arrhythmia should be noted, as well as reported increase in risk of heart failure when used in combination with azithromycin.
c) Inhibitors of IL-6 and related proinflammatory signaling:Tocilizumab (Actemra), Siltuximab (Sylvant), Sarilumab (Kevzara)
d) Inibitors of IL-17: Sekukinumab, Brodalumab
The first responsibility is to do no harm. But the fact is that the moment a patient is placed on a ventilator, the likelihood of fatality may already exceed the hope for survival. At that point, given the existing research base implicating hyperinflammatory response and cytokine storm, focusing on this pathway seems superior to ignoring it.
e) JAK inhibitors: Tofacitinib (Xeljanz), Fedratinib (Inrebic)
f) TNF inhibitors: Etanercept (Enbrel), Infliximab (Remicade), Adalimumab (Humira)
3) Given that the acute phase of treatment likely requires less than two weeks of these therapeutics, availability might be dose-matched to that horizon. Dividing longer-acting dosages in order to reach more patients might be considered.
4) Ideally, outcome data could be aggregated. Physicians I’ve spoken with seem to be largely sharing anecdotal outcomes, even though the data needed to calculate high-power odds ratios could potentially be obtained through information sharing of basic 2x2 outcome data.
Please understand - I realize that it’s not ideal for interim guidance to be driven by pathway analysis, research literature, and small-scale trials rather than the results of robust RCTs. In less urgent conditions, I would agree. So I am aware of the apparent temerity of this proposal.
Again, the guidance will ideally involve approved, well-known, preferably well-tolerated therapeutics where associated risks are monitored. The key shift is from one focused primarily on pulmonology to a broader modality that includes biologics (particularly along the IL6/TNF/Th17 axis).
I am deeply hopeful that NIH/NIAID will make interim, regularly updated therapeutic guidance along these lines available at the earliest moment possible. Frankly, the CDC guidance is a thimble of water in a forest fire.
Finally, a dear friend sent this to me yesterday, and I think it’s particularly worth sharing here.
“Hope is a state of mind, not of the world. Either we have hope within us or we don’t; it is a dimension of the soul, and it's not essentially dependent on some particular observation of the world or estimate of the situation.
Hope is not prognostication. It is an orientation of the spirit, an orientation of the heart; it transcends the world that is immediately experienced, and is anchored somewhere beyond its horizons.
Hope, in this deep and powerful sense, is not the same as joy that things are going well, or willingness to invest in enterprises that are obviously heading for success, but rather the ability to work for something because it is good, not just because it stands a chance to succeed.
The more propitious the situation in which we demonstrate hope, the deeper the hope is.
Hope is definitely not the same thing as optimism. It is not the conviction that something will turn out well, but the certainty that something makes sense, regardless of how it turns out.”
- Vaclav Havel
Wishing you, your families, and loved ones well.
Best - John
John P. Hussman, Ph.D.
Director, Hussman Foundation