Bromodomain Inhibition Blocks Inflammation-Induced Cardiac Dysfunction and SARS-CoV2 Infection in Pre-Clinical Models

Richard J Mills, Sean J Humphrey, Patrick RJ Fortuna, Gregory A Quaife-Ryan, Mary Lor, Rajeev Ruraraju, Daniel J Rawle, Thuy Le, Wei Zhao, Leo Lee, Charley Mackenzie-Kludas, Neda R Mehdiabadi, Lynn Devilée, Holly K Voges, Liam T Reynolds, Sophie Krumeich, Ellen Mathieson, Dad Abu-Bonsrah, Kathy Karavendzas, Brendan Griffen, Drew Titmarsh, David A Elliott, James McMahon, Andreas Suhrbier, Kanta Subbarao, Enzo R Porrello, Mark J Smyth, Christian R Engwerda, Kelli PA MacDonald, Tobias Bald, David E James, James E Hudson

Posted on: 27 November 2020

Preprint posted on 16 October 2020

‘In the eye of the (Cytokine) Storm’: COVID-19 inflammation and heart function in 3D-engineered cardiac tissues

Selected by Alexander Ward, Osvaldo Contreras

Categories: bioengineering, bioinformatics, clinical trials, genomics, molecular biology, pharmacology and toxicology, physiology, systems biology

Background

The ongoing novel coronavirus (SARS-CoV-2) pandemic has infected over 60 million people worldwide, with more than 1.4 million fatalities¹. Aside from the occurrence of well-characterised respiratory symptoms of coronavirus disease (COVID-19), 20-50% affected patients are presenting with cardiac complications, including myocarditis (inflammation of the heart), cardiac cell death, heart failure (acute and chronic types) and arrhythmias².

The reasons for the increased cardiac complications in COVID-19, as compared with other coronavirus diseases, are as yet unknown. It has been speculated that both direct and indirect consequences of infection contribute to diverse cardiac pathologies. With respect to the direct consequences, SARS-CoV-2 shows a significantly greater affinity for its primary target, the angiotensin-converting enzyme 2 (ACE2) receptor, than SARS-CoV-1, potently and directly targeting cardiac tissues high in ACE2². Exactly which cells of the heart SARS-CoV-2 affects remains unclear; however, direct infection of cardiomyocytes, the beating muscular, conducting cells, has also been observed³. In addition to direct effects on cardiomyocytes, the secondary inflammatory response and increased clotting risk associated with COVID-19 are thought to damage cardiac tissue indirectly². In severe cases, acute systemic inflammation induces a ‘Cytokine Storm’, the effects and composition of which are poorly understood.

Despite the recent news of efficacy of a number of novel COVID-19 vaccines^4,5, treatments which can alleviate the secondary symptoms remain behind vaccination efforts. Anti-inflammatory agents, inhibiting diverse inflammatory pathways, have made it to clinical trials; the corticosteroid, Dexamethasone, has been shown to reduce mortality in ventilated patients⁶.

Given the difficulties of studying the effects of SARS-CoV-2 on the human heart and the sheer speed of dissemination of COVID-19 related preprints, there is a necessity to quickly understand the molecular mechanisms of the disease firstly in an in vitro setting. The Hudson laboratory are world leaders in the generation of high-throughput bioengineered heart tissues (human cardiac organoids: hCO), using induced pluripotent stem cell (iPSC)-derived cardiac cells and their patented Heart-Dyno system⁷. Their system provides them with a unique opportunity to screen vast numbers of exogenous molecules and therapeutics, in mature cardiac-like tissues⁸, and, due to spontaneous beating, measure functional contractile parameters comparable to those measured in the heart. In the current preprint, Mills et al., use their tissue engineered hCO model to address the effects of the cytokine storm on cardiac function and cardiac cell composition following stimulation.

This schematic shows the process of generation of human cardiac organoids (hCO) using human induced pluripotent stem cells (iPSC) and the bioengineered Heart Dyno plate setup. Once the hCOs form, they begin to spontaneously beat, which permits the measurement of contractile force and other parameters associated with the heartbeat. Credit: (7) Mills et al., PNAS, 2017; 114 (40) E8372-E8381 (approved by authors)

Key findings

Much like the adult human heart, the dynamic, beating hCO tissues include diverse cardiac cell-types, including cardiomyocytes, endothelial cells, fibroblasts and pericytes. Using single nuclei RNA-seq (snRNA-seq), Mills et al. showed that the expression levels of several proinflammatory genes linked to the cytokine storm were very similar to those seen in the human heart. Using their high-throughput system, they next screened several pro-inflammatory factors, including cytokines, TNFa, IL-1b, IL-6, IL-8, IL-8, IL-17 and IFN-g, and pathogen associated molecule polyIC and LPS. A number of the components of the cytokine storm resulted in profound alterations to cardiac function in stimulated hCOs. TNFa induced a broad reduction in force (equating to systolic dysfunction), as has been suggested previously⁹. Whilst IL-1b, IFN-g and polyIC caused an increase in the relaxation time of tissues in between beats, which is clinically representative of diastolic dysfunction in heart failure, where the left ventricle fails to relax normally. This prolonged relaxation time of their cardiac tissues also resulted in an arrhythmic beating, a manifestation which is commonly seen in COVID-19 patients.

Video 1. Spontaneously contracting hCO, beating against the poles of the Heart Dyno. Credit to Richard J Mills.

Video Player

Media error: Format(s) not supported or source(s) not found

Download File: https://prelights.biologists.com/wp-content/uploads/2020/11/Beating-cardiac-organoid.mp4?_=1

00:00

Use Up/Down Arrow keys to increase or decrease volume.

The profound effects of the cytokine storm on cardiac dysfunction led the authors to investigate the potential downstream effectors of this inflammatory response in heart tissue. Given the critical role of phosphoproteins in transducing inflammatory signals, they interrogated the phosphoproteome using the high-sensitivity EasyPhos technique. The authors identified many differentially phosphorylated effectors, which included transcription factors and chromatin-binding proteins, as well as a handful of other transcription-related machinery, including STAT1 and BRD4. Many of these effectors appeared to be involved in cell proliferation. Upon cytokine storm stimulation, there was a shift in cell populations; fibroblasts became more activated, and cardiomyocyte populations were drastically altered. This activated, proliferative signature hint that the cytokine storm may be inducing cell-fate changes in cardiac tissue, possibly predisposing to heart failure-like cardiac dysfunction.

Next, Mills and colleagues performed a drug screening of potential small molecules that could prevent cardiac dysfunction after TNFa and cytokine storm-driven diastolic dysfunction. The authors provide evidence that suggests a protective role for the CDK8 inhibitors SEL120-34A and BI-1347 in cardiac dysfunction. Having observed induced phosphorylation of the Bromo- and Extra-Terminal domain (BET) family member BRD4 following cytokine storm exposure, they evaluated the potential protective role of BET inhibitors against cytokine storm-derived cardiac dysfunction. Among them, INCB054329 was the most effective in preventing cytokine storm-induced diastolic dysfunction, and remarkably, it also prevented diastolic dysfunction caused by COVID-19 patient serum with high CTNI and BPN, but not with low levels of these cardiac-derived proteins.

In order to recapitulate the cytokine storm observed in humans following SARS-CoV-2 infection and translate these previous findings into a clinically relevant animal model, the authors utilized a murine model of lipopolysaccharide (LPS)-induced inflammation. The treatment with INCB054329 BRD inhibitor blocked the LPS-mediated inflammation and also restricted the LPS-induced fibrotic response of fibroblasts. Remarkably, the BRD inhibitor also extended the life of LPS-challenged mice. Overall, these data suggest that INCB054329 represents a potent, robust and promising small molecule BRD inhibitor for improving cardiac function and cytokine storm-mediated mortality.

Using 2D cultured human-induced pluripotent stem cell derived cardiomyocytes, the authors also indicated that INCB054329 protects cardiac cells against SARS-CoV-2 infection and sarcomere degradation. Owing to the fact that the BRD inhibitor reduced the expression of ACE2, the authors speculated that reduced viral infiltration could be the cause for diminished viral load and infection in response to INCB054329 treatment. These favourable findings suggest that pre-treatment with BRD inhibitors could protect patients’ hearts against SARS-CoV-2 infection.

Why we chose this preprint and what we liked about it

Following on from our first prelight (https://doi.org/10.1242/prelights.23879), which addressed the role of BRD4 inhibition on the reversion of activated cardiac fibroblasts in ischaemic and hypertrophic mouse injury models, the current preprint used a similar approach to treat the COVID-19-induced cytokine storm. Here, Mills and colleagues demonstrated that BRD4 inhibition prevented fibroblast activation and cardiac dysfunction in response to the cytokine storm, in their hCO model, arriving at similar conclusions to the study from the Haldar and Srivastava laboratories¹⁰. The comprehensive efficacy of BRD4 inhibition in treating cardiac dysfunction in these two studies makes the recent preprints highly complementary and extremely exciting for the field.

This preprint utilises state-of-the-art techniques in conjunction with the unique Heart Dyno system developed in the Hudson lab, to address the critical question of how heart function can be affected directly and indirectly by SARS-CoV-2 infection. They have been able to elegantly and quickly identify a combination of cytokines with profound effects on the function of their heart tissues, as well as establish effective potential therapeutics to ameliorate cytokine-induced cardiac dysfunction. The assay-throughput and functional similarity of the hCO system to the adult heart are huge strengths of this study. As such, this preprint represents an important step in furthering our understanding of the mechanisms of COVID-19 in the heart.

Tags: bet proteins, cardiovascular disease, covid-19, cytokine storm, fibroblasts, inflammation, microtissues, organoid model

doi: https://doi.org/10.1242/prelights.25933

Read preprint

(No Ratings Yet)

Author's response

Richard J Mills and James E Hudson shared

Given systemically, BRD4 inhibition will affect all tissues, what do you suspect will be the principal off-target effects of BRD4 inhibition? And would you speculate that the inhibitory effects may outweigh the off-target effects?

Bromodomain inhibitors generally target all the BET (Bromo- and Extra-Terminal domain) family: BRD2, BRD3, BRD4 and BRDT. The drugs generally prevent the protein-protein interactions with acetylated histones and regulate transcriptional responses by binding to the bromodomains. There are 2 bromodomains (BD1 and BD2) on the proteins and the first generation of molecules target both with side effects including shrinking of testis and thrombocytopenia in some clinical trials. Newer generation drugs targeting the BD2 domain more specifically have a reduced side effect profile and have been safe in phase III clinical trials (https://www.globenewswire.com/news-release/2020/03/30/2008314/0/en/Resverlogix-Announces-Publication-of-Key-Apabetalone-Study-BETonMACE-in-the-Journal-of-the-American-Medical-Association.html). Our ambition is to repurpose these newer drugs to the clinic, and we have new data demonstrating that all BET inhibitors that have been in clinical trials (except ABBV-744 with its AR effects) are efficacious in our cytokine storm model. As apabetalone is the most specific for BD2 and has been safe in clinical trials, this is our lead candidate.

Do BRD inhibitors have a similar protection of SARS-CoV-2 infection in other tissues? We mean, are these findings scalable to other cell types like airway or lung-resident cells?

Preliminary data suggests that apabetalone also regulates ACE2 expression in lung cells and infection studies are underway.

Several differential phosphosites were found in extracellular matrix components, what are the implications of this for COVID-19 patients? Furthermore, do you think this may contribute to a prominent fibrotic response and possibly, later stage heart failure?

I certainly think that these may be important. It would be very interesting to follow-up some of the data in our phosphoproteomics and snRNA-seq because there may also be important regulators of dysfunction embedded in that data which have not yet been studied.

Why do you think there is a shift in CM populations with CS, from predominant CM1 to roughly equivalent numbers of CM1 and CM2? Moreover, what does the CM2 population represent?

CM1 seem to have higher expression of sarcomeric proteins and metabolic genes in comparison to CM2. It would be very interesting to determine specifically what is going on in these cells, but at this stage it’s hard to pinpoint with so many transcriptional changes.

It is not clear whether LPS-induced cytokine storm mice model used phenocopies or recapitulates the plethora of different manifestations of SARS-CoV2 infection in the human body. Are you trying to find a more suitable animal model?

We have struggled to determine ‘what is a good animal model for this’. We have done some testing in K18hACE2, but unlike patients these die of brain infection due to the exogenous expression of ACE2. At QIMRB, we will soon have mACE2-hACE2 mice ready for testing.