A new in silico model has been developed that can predict which COVID-19 patients need which treatments when, depending on their symptoms and the course of their disease. The model, developed by an international team, can predict how different patients are likely to respond to specific interventions based on their co-morbidities, age, and innate and adaptive immune responses.
One key prediction the model makes is that results of any treatment depend on “A sustained activation of CD8+ T cells along with the control of the populations of neutrophils and macrophages,” the authors write. Another observation is that the effects of antiviral and anti-inflammatory may depend on the stage of the disease. Other drugs that may improve outcomes are heparin and interferon, with the later likely to have a better effect when initiated earlier.
The senior author of the paper on this work is Rakesh K. Jain, PhD, director of the Edwin L. Steele Laboratories at Massachusetts General Hospital and Harvard Medical School. The paper was published yesterday online in PNAS.
One of the most challenging aspects of COVID-19 is the wide range of symptoms patients present with, ranging from almost none to rapid decline with multiorgan failure then death. In between, some suffer lung complications, systemic inflammation, and disseminated microthrombosis, which can cause stroke, myocardial infarction, or pulmonary emboli. The reasons for this variation, the researchers write, “results from a poorly understood combination of patient factors, viral dynamics, antiviral and immune modulating therapies, and dynamics of the innate and adaptive immune responses.”
Until now this has left clinicians sometimes guessing how to manage individual patients, especially now that there are finally a range of treatments available. This range in symptoms has also made clinical trials more challenging and may be the reason that trials of the antiviral Remdesivir (a nucleoside analogue prodrug), for example, failed in China, succeeded in the US and other countries, and then failed again in the World Health Organization Solidarity trial. Other therapies for COVID-19 have similarly shown conflicting results. “Given the large range of patient comorbidities, disease severities, and variety of complications such as thrombosis, it is likely that patients will have heterogeneous responses to any given therapy,” the model’s developers write.
Their model was built on a “comprehensive mathematical framework” based on what is known about how severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection proceeds. It incorporates elements of the immune response, the renin-angiotensin system and ACE2 (through which the virus enters cells), rates of viral replication, inflammatory cytokines, and the coagulation cascade.
The model predicts the progress of viral load, immune cells, cytokines, thrombosis, and oxygen saturation based on the patient’s baseline condition and comorbidities, including obesity, diabetes, and hypertension. The researchers validated its predictions with clinical data from healthy people and COVID-19 patients. Those results were used to determine how factors such as older age, co-morbidities, and a dysregulated immune response may impact disease progression.
They also simulated treatment with various drug classes to identify optimal therapeutic protocols for patients with specific symptoms. The important role of T cell activation, is interesting for a couple of reasons they point out. For one thing, other mathematical models for diseases such as hepatitis C, HIV-1, and simian HIV have shown that interactions between CD8 T cells and antigens are linked to progression of multiple viral infections. T cell activation also plays a critical role in the progression of other coronavirus infections, including SARS-CoV and Middle East respiratory syndrome CoV.
The authors suggest their “systems biology-based mathematical model could be used as the basis for “personalized, optimal management of COVID-19.”
