COVID-19 and silent hypoxemia in a minimal closed-loop model of the respiratory rhythm generator

Biological Cybernetics https://doi.org/10.1007/s00422-024-00989-w ORIGINAL ARTICLE COVID-19 and silent hypoxemia in a minimal closed-loop model of the respiratory rhythm generator Casey O. Diekman1 · Peter J. Thomas2 · Christopher G. Wilson3 Received: 12 April 2023 / Accepted: 28 March 2024 © The Author(s) 2024 Abstract Silent hypoxemia, or “happy hypoxia,” is a puzzling phenomenon in which patients who have contracted COVID-19 exhibit very low oxygen saturation (SaO2 < 80%) but do not experience discomfort in breathing. The mechanism by which this blunted response to hypoxia occurs is unknown. We have previously shown that a computational model of the respiratory neural network (Diekman et al. in J Neurophysiol 118(4):2194–2215, 2017) can be used to test hypotheses focused on changes in chemosensory inputs to the central pattern generator (CPG). We hypothesize that altered chemosensory function at the level of the carotid bodies and/or the nucleus tractus solitarii are responsible for the blunted response to hypoxia. Here, we use our model to explore this hypothesis by altering the properties of the gain function representing oxygen sensing inputs to the CPG. We then vary other parameters in the model and show that oxygen carrying capacity is the most salient factor for producing silent hypoxemia. We call for clinicians to measure hematocrit as a clinical index of altered physiology in response to COVID-19 infection. Keywords Silent hypoxemia · Breathing control · Central pattern generator · Computational modeling · COVID-19 · Polycythemia · Sensory feedback 1 Introduction 1.1 Background The global COVID-19 pandemic led to over 1,003,000 deaths in the USA, and over 6,881,000 worldwide, from its onset in late 2019 through March, 2023 (Johns Hopkins University Communicated by Benjamin Lindner. B Casey O. Diekman Peter J. Thomas Christopher G. Wilson 1 Department of Mathematical Sciences, New Jersey Institute of Technology, University Heights, Newark, NJ 07102, USA 2 Department of Mathematics, Applied Mathematics and Statistics, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106, USA 3 Department of Pediatrics and Basic Sciences, Lawrence D. Longo, MD Center for Perinatal Biology, Loma Linda University, 11223 Campus St, Loma Linda, CA 92350, USA Coronavirus Research Center 2023). COVID-19 can cause profoundly low levels of oxygen in the blood (hypoxemia) with near normal arterial carbon dioxide (Pa CO2 ) levels. Although some individuals with COVID-19-induced hypoxemia experience dyspnea (breathing discomfort), many do not (Dhont et al. 2020). During surges of the pandemic, patients arriving at already overcrowded emergency rooms (ERs) who were not in obvious respiratory distress were often triaged (Dhont et al. 2020). However, some of these patients may have had reduced oxygen saturation despite their lack of dyspnea (Simonson et al. 2021; Berezin et al. 2021; Chandra et al. 2020). This subpopulation of COVID-19 patients present with a novel condition known as silent hypoxemia or “happy hypoxia” (Simonson et al. 2021). Silent hypoxemia can result in tachypnea (rapid, shallow breathing), and with severe hypoxemia, changes in ventilation can occur (Easton et al. 1986; Easton and Anthonisen 1988), but in general there is an absence of increased alveolar ventilation (Dhont et al. 2020). The mechanism underlying this condition is poorly understood but has been hypothesized to depend upon high expression levels of angiotensin converting enzyme 2 (ACE2) in the lungs, carotid body, and, perhaps, in the central breathing control circuitry within 123 Biological Cybernetics the medulla oblongata (Simonson et al. 2021). ACE2 is the cellular entry point for SARS-CoV-2 (Yuki et al. 2020). Additionally, recent work has shown that there is a shift in the oxyhemoglobin dissociation curve1 in COVID-19 patients (Vogel et al. 2020; Ceruti et al. 2022). Since carotid body chemoreceptors respond to both low O2 and high CO2 , a primary problem in these patients may be dysregulation of these sensors and chemosensory reflexes in general. COVID19 infection has been shown to increase ACE2 expression, leading to changes in sensitivity to both CO2 and O2 ; changes in blood gases lead to a concomitant change in activity within the nucleus tractus solitarii (NTS). Recent work has shown that ACE2 is present within the carotid bodies of humans (Porzionato et al. 2021; Villadiego et al. 2021) and there is evidence of altered chemosensation across multiple systems with SARS-CoV-2 infection (Caretta and Mucignat-Caretta 2022). The absence of dyspnea—even though patients exhibit low oxygen saturation—suggests that changes in carotid body inputs to the NTS are a key feature of SARS-CoV-2 infection. Additionally, there may be changes in NTS activity that contribute to the blunted ventilatory response but this has not yet been reported. 1.2 Altered chemosensory function and silent hypoxemia After four years of the COVID-19 pandemic and ongoing endemic infection, a few key pathophysiologies have become apparent. First, ACE2 expression is correlated with the location and severity of infection (Zou et al. 2020). Because ACE2 is, based on current knowledge, the main vector by which SARS-CoV-2 enters the body’s cells, changes in ACE2 expression should have an impact on the severity and time course of COVID-19 symptoms. Second, changes in NTS signaling may play a key role in altering the normal, physiological response to changes in oxygenation during COVID-19, and that information may be carried by the glossopharyngeal nerve (innervating the carotid body) or lung afferents via the vagus nerve. Information sensed at the carotid bodies (and lung interoceptors) ultimately reaches the NTS via the vagus and glossopharyngeal nerves. From the NTS, these signals are distributed to local visceral integration circuits within the medulla, including the cardivascular control regions (rostral and caudal in the ventral medulla) and the preBötzinger complex and associated regions of respiratory control within the brainstem. Based on the clinical observations reported so far, it appears that there is a change in gain in the pathway from carotid body, to NTS, to the breathing rhythm generator and 1 The oxyhemoglobin dissociation curve gives the steady-state fraction of hemoglobin capacity occupied by oxygen, as a function of oxygen tension in the blood. 123 pattern formation network. These observations in patients have provided the motivation for us to focus on assessing the effect of changes in sensitivity/gain in this signaling pathway. This change in gain may be more prevalent in any one of these circuit elements and further work needs to be done to determine the exact mechanism by which sensitivity of the control circuit is impacted. Given the low partial pressure of oxygen in arterial blood (Pa O2 ) of patients infected with SARS-CoV-2 virus (Sartini et al. 2020; Chen et al. 2020) (...truncated)