Anosmia has emerged as the paradigmatic symptom of COVID-19. Data are still rolling in, but it looks like most (and as we get more objective metrics, perhaps even all) patients report some degree of smell loss, and conversely loss of smell is the most specific predictor of having COVID – more predictive than fever, shortness of breath, or a cough by as much as 10-fold. But how does the virus attack olfaction? And why do most (but not all) patients regain their sense of smell quickly?
David Brann, Tatsuya Tsukahara and Caleb Weinreb took advantage of pre-COVID datasets they had developed in the lab to ask which cells in the olfactory epithelium express ACE2 and other key cell entry genes for SARS-CoV2, the causal agent in COVID-19. Together with our amazing collaborators from across the globe (John Ngai, Darren Logan, Brad Goldstein, Hiro Matsunami, Matthew Grubb and their colleagues in the lab), they came to the conclusion (see paper here) that support cells in the olfactory epithelium – but not primary olfactory sensory neurons – seem to express ACE2 and are therefore infectable by the virus. A similar analysis in the olfactory bulb revealed that, again, neurons don’t seem to express ACE2 but vascular cells – in particular pericytes – express high levels of ACE2 and of the proteases required for SARS-CoV2 spike protein cleavage and cell entry.
These findings lead to a working model for how the virus hits olfaction. Given its unusual timecourse and clinical manifestations (patients lose their sense of smell rapidly, have total anosmia, don’t have a runny nose, and most recover quickly), one possible explanation is that support cell dysfunction causes temporary changes in mucus composition, local energetics, ion gradients, or local cytokine concentrations that end up indirectly causing reversible sensory neuron dysfunction. However, there is a subset of patients that seem have prolonged anosmia (with accompanying parosmias, suggesting sensory neuron cell death and renewal); in these patients, particularly severe infection or support cell dysfunction might ultimately lead to sensory neuron death, which requires the slow process of neural regeneration to reverse. It is also possible that viral infection of the vasculature of the olfactory bulb causes central effects, either through changing perfusion or through local inflammation.
This paper got a crazy amount of press coverage (ranging from USA Today to the Times of India), so it is doubly important to state: 1. These are just models. Although emerging animal and human data are consistent with these models (for now), we need way more work to tease apart what is actually happening with COVID and smell; and 2. Just because we think (based upon current data) that support cells are the primary targets for the virus, it doesn’t mean that the sensory neurons aren’t dramatically affected or in some cases dying. Long term olfactory loss is devastating for patients who have it, and although our model suggests a possible explanation for the clinical observation that many patients (60 percent or more) seem to have rapidly reversible anosmia, it is critical to recognize that there are probably multiple mechanisms at play that act in parallel to cause a spectrum of symptoms from mild to severe, from transient to prolonged. All of which is to say: this is the very first piece in a complex puzzle, and much more data (from better animal models, from human patients and from autopsies) is required. Kudos to David, Tatsuya and Caleb for taking this first important step – extremely proud of the rigor of the paper, the circumspection of their conclusions, and the directions they point to that, we hope, will help unravel this important medical mystery.