Can immunological imprinting drive neurological dysfunction in long COVID?

Abstract

This scientific commentary refers to ‘Neurologic sequelae of COVID-19 are determined by immunologic imprinting from previous coronaviruses’ by Spatola et al. (https://doi.org/10.1093/brain/awad155).

This scientific commentary refers to ‘Neurologic sequelae of COVID-19 are determined by immunologic imprinting from previous coronaviruses’ by Spatola et al. (https://doi.org/10.1093/brain/awad155).

The long-term consequences of SARS-CoV-2 infection in some individuals, known as ‘long COVID’, are driven by host responses that are only now being identified. Current estimates for the prevalence of neurological symptoms associated with long COVID—including what is commonly referred to as ‘brain fog’, which can persist with varying severity for years—range from ∼3% to ∼30%.^1,2 These and other prevalence estimates are based upon meta-analyses of studies that incorporate varying diagnostic criteria and assessment approaches. Patients referred for neurological evaluation often describe symptoms such as difficulty with concentration and word-finding, short-term memory impairment, inattentiveness and impaired reading comprehension. However, the lack of universally applied criteria for defining long COVID CNS symptoms means that our ability to diagnose and treat them is evolving only slowly.

The most informative evaluations of long COVID patients with self-reported neurological symptoms, also referred to as neuroPASC (post-acute sequelae of COVID-19), include a thorough patient interview, physical examination and neurocognitive screening. Associative or predictive biomarkers that could enhance diagnostic accuracy are needed but are not yet available. Our inability to identify SARS-CoV-2-driven neuropathogenic pathways linked to specific symptoms of neurological dysfunction contributes to the difficulty in developing a working definition and accurate diagnostic criteria for long COVID. To this end, both validated neurological criteria and, if possible, complementary laboratory markers are needed to move towards more effective diagnosis and management of long COVID complications.

In this issue of Brain, Spatola and co-workers³ provide tantalizing insights into a potential immunological biomarker and associated pathological mechanisms for the development of long COVID-associated CNS dysfunction. The study does not focus on a single symptom (such as brain fog), but instead includes a broad spectrum of neurological symptoms. The potential biomarker/pathogenesis link is an immunological imprinting phenomenon, previously known as an ‘original antigenic sin’, in which an individual’s immune responses to previous infections—in this case by coronaviruses—shape their immune response to an antigenically related virus, here SARS-CoV-2.⁴ Immunological imprinting associated with early-life infections by influenza A subtypes has been shown to determine susceptibility to severe influenza infection complications later in life.⁵ Spatola and co-workers³ now propose that the same principal may also apply to long COVID CNS dysfunction (Fig. 1).

Figure 1.

Proposed contribution of coronavirus immunologic imprinting to the development of long COVID neurological dysfunction. Previous (ancestral) infection with common coronaviruses induces B cell immune responses, clonal expansion and production of antibodies against these viruses. Subsequent (years) infection with SARS-CoV-2 triggers enhanced production of these ancestral antibodies, and relatively limited production of anti-SARS-CoV-2 antibodies, in patients who go on to develop long COVID neurological dysfunction. Selective transfer of anti-SARS-CoV-2 IgG1 (and not other immune globulin classes) across the blood–brain barrier by the binding and transport functions of neonatal Fc receptors (FcRn) contributes to a limited immune response to SAR-CoV-2 in the CNS. It has also been proposed that the robust expression of ancestral coronavirus antibodies limits the ability of naïve B cells to generate antibody responses to SARS-CoV-2 (a phenomenon consistent with immunologic imprinting). This may result in decreased SARS-CoV-2 clearance and increased inflammation within the CNS compartment, leading to neurological dysfunction. Created with BioRender.com.

Spatola and co-workers³ addressed the potential association between CSF and plasma immunoglobulin responses to SARS-CoV-2 infection, prior infection with common coronaviruses, and the presence of persistent neurological symptoms post-COVID in 112 patients (18 with long COVID/PASC neurological symptoms; 94 without). The investigators profiled functional immune responses in the peripheral blood (serum) compartment of both groups, as well as in the CSF of patients with neurological symptoms.

Unexpectedly, while all SARS-CoV-2 immunoglobulin isotypes (IgG, IgM, IgA) and subclasses (IgA1–2; IgG1–4) were found in serum, only a limited and focused IgG1 SARS-CoV-2 response was found in the CSF of neurological patients. Furthermore, some serum SARS-CoV-2 antibody functional responses [e.g. antibody-dependent complement deposition (ADCD), antibody-dependent natural killer cell activation (ADNKA), and ability to bind to Fcγ receptors] were lower in patients with neurological symptoms than in those without. The investigators convincingly hypothesize selective transfer of specific antibodies (IgG1) across the blood–brain barrier (in an FcγR-dependent manner associated with enhanced neonatal FcRn binding), without robust intrathecal synthesis of SARS-CoV-2 antibodies. This, along with the lower serum functional antibody responses, may result in increased CNS vulnerability to SARS-CoV-2-associated injury and/or dysfunction. The selectively transferred IgG1 would be expected to provide more limited CNS compartment responses against SARS-CoV-2 than those provided by a more diverse CNS SARS-CoV-2 antibody transfer or by robust and diverse intrathecal antibody production. Notably, individuals with higher serum SARS-CoV-2 specific IgG1 responses had better neurological symptom status. Reduced serum antibody responses to SARS-CoV-2 and selective IgG1 transfer to the CNS may therefore be directly linked to the genesis of long COVID neurological symptoms.

This study provides strong evidence that SARS-CoV-2-associated neuropathogenic responses occur on a background of enhanced immune responses to other coronavirus infections. Neurologically symptomatic patients had stronger antibody responses to other coronaviruses (variants 229E, HKU1, NL63, OC43) in both serum and CSF. However, serum antibody responses to other commonly encountered non-coronaviruses (EBV, influenza, HSV1) were similar between patients with and without neurological symptoms. Thus, neurological symptoms were observed in the presence of relatively weak serum antibody functional responses to SARS-CoV-2 and relatively strong responses to other coronaviruses, without evidence for deficient antibody responses to other viruses. This suggests that relative levels of serum antibodies to common coronaviruses and to SARS-CoV-2 could serve as predictive biomarkers for neurological outcomes. This will be important to confirm.

The investigators conclude that neurologically symptomatic long COVID patients demonstrate immunologic imprinting, whereby prior coronavirus infections impact antibody responses to subsequent SARS-CoV-2 infection; this is a reasonable conclusion. Strong antibody responses to prior coronavirus infections (ancestral antigens) induced by SARS-CoV-2 infection in these patients is reminiscent of immunologic imprinting in influenza A infections. Robust immunity to the original influenza strain exists along with relatively weaker immunity to subsequently acquired new influenza strains.⁶ In the case of COVID-19, the heightened serum response to other coronaviruses could inhibit the expression of new B lymphocyte responses to SARS-CoV-2, resulting in less CNS virus clearance.⁷ Spatola and co-workers³ extend their observations and conclusions to suggest that the SARS-CoV-2 immune response in the CNS compartment is incomplete (IgG1 selection and response intensity, less diversity) and less focused on SARS-CoV-2 than on other coronaviruses. This may result in reduced SARS-CoV-2 clearance and persistent inflammation within the CNS compartment, which could indeed provide a pathogenic drive for the development of neurological symptoms. They convincingly discuss potential roles for the heightened responses to other coronaviruses and for weaker CSF SARS-CoV-2 antibody responses. This combination of compartmentalized antibody responses is also consistent with immunologic imprinting and modulation of antibody responses as a potential factor in neuropathogenesis.

These investigators might be on to something: reduced SARS-CoV-2 specific antibody responses may drive the emergence of long COVID CNS symptoms. However, a conundrum in implicating reduced brain viral clearance as a driver of long COVD is the consistent lack of detection of replication-competent SARS-CoV-2 within the brains of decedents who are beyond the 1–2 month post-acute infection window, despite occasional detection of subgenomic virus sequences.⁸ One might speculate that persistence of uncleared SARS-CoV-2 RNA fragments within the brain serves as a substrate for production of some viral proteins, which could trigger inflammatory and/or immunological responses. Some, but not all, studies have reported detection of nucleocapsid antigen and/or antibody in the CSF of some long COVID patients.⁹ It has also been suggested that persistence of heightened CNS coronavirus antibody levels combined with reduced levels of SARS-CoV-2 antibodies might directly induce neuroinflammation, which is an interesting possibility. Large, prospective studies of rigorously diagnosed long COVID patients that include CSF analyses and tracking of outcomes are clearly needed.

The hypothesis that immunologic imprinting may ultimately contribute to neurological dysfunction in SARS-CoV-2 infection is important not only for developing therapeutic targeting strategies but also for avoiding risk in patients receiving B lymphocyte-targeting therapies. These questions are certainly worthy of investigation. In this era of highly successful immune-modulating, B lymphocyte-targeting treatment of neuroinflammatory disorders (multiple sclerosis, neuromyelitis optica spectrum disorders), one can speculate about the effects of such treatment on the natural history of long COVID neurological disorders. Would such treatments increase or decrease long COVID risk? Would reducing immune responses to common coronaviruses reduce long COVID risk in new SARS-CoV-2 infection? The use of monoclonal antibodies that deplete B lymphocyte precursors in patients with multiple sclerosis may increase acute COVID severity¹⁰ and prospective studies of long COVID risk in these patients are necessary.

Much remains to be done, and there is much to build upon from this provocative study. Although the criteria for the diagnosis of ‘neuroPASC’ were somewhat broad and not the same as those in some other long COVID neurological studies, the observations are enlightening and likely relevant to the pathogenesis of long COVID neurological dysfunction. Establishing consistent and widely accepted diagnostic criteria for long COVID would help all future investigations. Long-term prospective studies of immunologic imprinting in long COVID patients with neurological manifestations will hopefully follow this important study.

Funding

D.L.K has received funding from the National Institutes of Health (1R01NS122570-02).

Competing interests

The author reports no competing interests.