Autoimmune triggers in COVID-19 as a key to managing symptom diversity

Autoimmune triggers in COVID-19 pathology as a key to deciphering symptom diversity and enabling tailored therapy with repurposed drugs. 

Marinos Kallikourdis

COVID-19 and, more importantly, post-COVID-19, display a wide array of diverse symptoms, rendering their clinical management difficult. It is this difficulty, and the burden it imposes on heath systems, that acts as a driver for the need for lockdowns, with their consequent dire socioeconomic effects.

The symptoms differ from patient to patient (even in twins); they worsen with age and pre-existing comorbidities; they rarely occur in children; and they are often cardiovascular, as we and others have shown (Stefanini, Heart, 2020) but also neurological or dermatological.

Are there drugs that are in use today that could have a good chance to be effective therapeutics against the diverse COVID-19 and post-COVID-19 symptoms? How can we identify them? How can we tailor their choice to the different symptoms?

Is there a way, beyond trial-and-error, to answer these questions combining both hard evidence-based hypotheses and data mining approaches?

Could this be extended to possible future viral pandemics?

Our proposal stems from an evidence-based hypothesis for the mechanism of generation of pathology that underlies the multiple symptoms of COVID-19 and post-COVID-19. This is important as it potentially may affect any of the 133M humans (as of 4/2021) that have been through COVID-19, who may face these clinical issues on the years to come. It also affects health systems world-wide, who will have to deal with the clinical burden of COVID-19-triggered long-lasting symptoms, that do not disappear (or indeed only become clinically detectable) after the end of infection.  

Our approach may reveal evidence for the mechanism driving the multiple facets of COVID-19. It may help identify profiles of the different facets, in a way that opens the door for the targeted testing of existing autoimmunity drugs that can be re-purposed to treat the different COVID-19 facets.  

We work in a research center embedded in one of the central hospitals in Lombardy, that has treated several thousand COVID-19 patients. We are actively working with post-COVID patients, whose progress we are monitoring clinically and analyzing in great (molecular) depth. This brings us very close to the target audience of this proposal.

Our primary expected outcome is the identification of possible candidate autoimmunity drugs that may be used in repurposing trials for treating COVID-19/post-COVID-19. This mirrors our own successful repurposing of an autoimmunity drug in an experimental heart disease context, and the clinical repurposing of Dexamethasone for COVID-19 (RECOVERY trial, N Engl J Med 2020). The fact that some attempts for autoimmunity drug repurposing for COVID-19 did not work (e.g. tocilizumab, British Med J 2021) highlights the utility of this data-driven approach, which can leverage existing clinical data and help predict the candidate drugs most likely to work. As this also applies to post-COVID-19, the (hopeful) termination of the pandemic phase is not likely to eliminate the need for these solutions.

Our findings on candidate drugs will be written up for publication in open-access journals, in as short a time-frame as possible, as is standard within academia. This will enable anyone within the entire scientific community to act upon them, whilst we will, in parallel, pursue their clinical testing.

The same dissemination strategy will be used for any findings leading to the suggestion of new diagnostic biomarkers for monitoring the progression of disease.

The identification of drugs suitable for repurposing for the treatment of COVID-19 (and post-COVID-19) syndrome is of immediate relevance for both the acute syndrome patients, as well as the >130M convalescent patients. Emerging data suggest that the latter may face a series of secondary pathologies (post-COVID-19), triggered by the infection but leading to symptoms long after the elimination of the virus.

The availability of cheap, easy-to-produce drugs is a crucial issue, as many developing countries lack the resources to afford many of the innovative reagents developed to address the COVID-19 healthcare emergency.

The best example of this has been the finding (by trial-and-error) that a very old, very cheap, and very widely available immunosuppressant drug, Dexamethasone, used extensively in autoimmunity, can be used in severe COVID-19 patients with demonstrable clinical benefit (RECOVERY trial, New England Journal of Medicine https://doi-org.ezproxyberklee.flo.org/10.1056/NEJMoa2021436 )

This drug offers a financially viable option, applicable to health systems regardless of lack of resources. There are likely to be other drugs that can fit this profile. It will take an approach that combines the power of Big Data analytics with the subtlety of deciphering complex immune-mediated pathophysiology to help uncover them, for common benefit.

The outcome of our approach (identification of novel therapy solutions), communicated via open-access peer-reviewed publications/pre-prints, can reach millions of patients. This process is usually slow, taking a decade from concept to cure. COVID-19 accelerated all timelines involved, demonstrating that concept to clinic can be obtained in months. Thus our default impact may be wide and fast.

This leaves space for scaling up the approach itself. Currently our team (2 academics) curate iteratively the data searches of in-house computational scientists/data scientists managing the data set of the Trinity Challenge Partner who has shown interest in the approach. This can be scaled up by:  

-Increasing the number of partners owning Big Data, willing to apply the approach. This requires exposure, and would be the main benefit of becoming one of the Trinity Challenge winners. This could be fully running within a year.  

-Add more members to the team of 2 curators. This requires specialist knowledge of the complex dynamics of immune-mediated pathophysiology, but a suitable recruitment can be made if funds are available.  

-Train the collaborating institutions to perform the approach independently. This would take more time (3 years), due to the complexity of the curation process, but would exponentially increase impact.  

The initial metric for success is the number of entities whose clinical data are being analyzed via our approach (currently 2, our in-house hospital and one TC Partner). This number will grow, if the project is successful.  

A follow-up metric is the distillation of actionable conclusions from each data set analyses. A consequent, intermediate, and more elaborate indicator is the generation of pre-prints and, subsequently, open-access peer-reviewed publications derived the most promising of conclusions (such as the previously mentioned Kallikourdis M et al (2017) T cell costimulation blockade blunts pressure overload-induced heart failure. Nat. Commun. 8, 14680 https://doi-org.ezproxyberklee.flo.org/10.1038/ncomms14680 ). A third milestone in this conceptual path is the initiation of clinical testing designed to validate the conclusion.  

A third category of indicators, the most long-term, is the number of institutions able to perform the approach independently, without the active involvement of the team, so as to guarantee the self-sustained growth of the solution, until it becomes common practice, for the benefit of public health.

Our solution can be applied to any entity or institution that holds Big Data of clinical nature. This mainly includes public health systems and private health sector entities in countries with existing digitalization of health records.

The benefits for each institution are easily apparent – but decision makers within each institution must know about the option to apply our solution. Thus this is a barrier of lack of exposure. It is our hope that being selected as one of the winning teams of the Trinity Challenge may help surpass this barrier.

The funding requirements and technical obstacles beyond this are minimal, and the solution, if fully applied, can be self-propagating.

Humanitas University, Milan, Italy

Humanitas Research Hospital, Milan, Italy

COVID-19 and post-COVID-19 are in dire need of clinical solutions.

Many institutions and public and private entities posesss unrivalled datasets and analytics to answer complex queries; but they may not be posing the correct questions.

We have pioneered a novel way of examining physiopathological data, that may apply to the COVID-19 and post-COVID-19 symptoms. Yet access to the datasets above, and interest from the entities to explore this solution can only come through initiatives such as the Trinity Challenge. This is why we are applying.

We have established a connection and are attempting to apply our solution with TC Partner Optum. Other Partners that hold clinical data (or proxies of clininal data, e.g. hospital payment data) would be suitable and could benefit from the application of our solution.