Comparative analysis of central nervous system persistence of morbilliviruses across host species reveals shared features with subacute sclerosing panencephalitis

Abstract

Background

Subacute sclerosing panencephalitis (SSPE) is a rare but always fatal neurological complication of measles, associated with persistent measles virus (MeV) infection of the central nervous system (CNS). MeV is detected in the brains of deceased SSPE patients as hypermutated viral genomes, which often contain premature stop codons in the matrix (M) gene and encode a hyperfusogenic fusion (F) protein. As a result, infected cells in the CNS cannot produce virions but rather transmit the viral ribonucleoprotein from cell-to-cell in a slow but progressive manner, which ultimately results in widespread MeV infection of the brain. However, our understanding of the pathogenesis of SSPE is constrained by the lack of animal models combined with a clinically silent incubation period of several years.

Main text

In a recent publication in PNAS, Daniel Roji et al. analysed animal morbillivirus sequences detected in the brains of persistently infected hosts, including a lynx (canine distemper virus, CDV), two seals (phocine distemper virus, PDV) and a dolphin (cetacean morbillivirus, CeMV). The authors describe that the morbillivirus sequences detected in these animals contained unique mutational signatures. The CDV, CeMV and one of the PDV sequences encoded nonfunctional M proteins and hyperfusogenic F proteins, reminiscent of the mutational profile of MeV detected in SSPE patients. Therefore, the authors conclude that human and animal morbilliviruses associated with chronic neurological infections share key genotypic and phenotypic traits, and that thorough analysis of naturally occurring animal morbillivirus infections could improve our understanding of SSPE pathogenesis.

Conclusion

The comparison of neurological complications associated with human and animal morbillivirus infections as described by Daniel Roji et al. provides a unique One Health approach to understanding the drivers of SSPE. These findings may support the development of early diagnostics or an animal model for SSPE, providing an opportunity to test novel intervention strategies.

Background

Measles virus (MeV), an enveloped virus with a negative strand RNA genome, is one of the most infectious viruses of humans. MeV is transmitted via the respiratory route, and uses at least two cellular receptors to infect myeloid, lymphoid and epithelial cells [1]. After an incubation time of approximately two weeks, patients develop cough, high fever and a typical skin rash. Opportunistic infections are common as a result of measles-associated immune suppression, leading to complications such as pneumonia, diarrhoea and otitis media [2]. Recovery from measles usually leads to lifelong immunity, but in rare cases MeV can persist in the central nervous system (CNS) and resurface years later as the cause of the fatal neurological complication subacute sclerosing panencephalitis (SSPE) [3]. MeV belongs to the genus Morbillivirus of the family Paramyxoviridae. Several closely related animal morbilliviruses are known to cause acute disease in their respective host species, including canine distemper virus (CDV) in carnivores, phocine distemper virus (PDV) in seals, cetacean morbillivirus (CeMV) in whales and dolphins, and different morbilliviruses of bats [4, 5].

Morbilliviruses cause acute infections in immunologically naive hosts. During the early subclinical incubation stage, morbilliviruses infect myeloid and lymphoid cells by targeting the cellular receptor CD150 (SLAMF1). A cell-associated viremia rapidly leads to widespread infection of the lymphoid organs, followed by infiltration of infected immune cells into peripheral tissues. Infection of epithelial cells and keratinocytes detected late in the pathogenesis is mediated by interaction with the second cellular receptor nectin-4, a protein expressed at the adherens junction. Infection and depletion of memory immune cells leads to lymphopenia and immune amnesia, but if the adaptive system is not overwhelmed a morbillivirus-specific immune response will develop and clear the infection. However, in some cases morbilliviruses can infect and persist in the CNS. The prevalence of neurological complications differs per morbillivirus: in some host species (e.g. humans and ruminants) these are rare, whereas in others (e.g. carnivores and cetaceans) these are relatively common. The route of morbillivirus entry into the CNS remains unclear, also because neurons are receptor-negative. Nevertheless, the CNS infection persists despite high MeV-specific antibody levels in serum and cerebrospinal fluid. Currently available therapeutic options can lead to transient clinical improvement, but the outcome of the disease is always fatal.

In a recent publication, Daniel Roji et al. analysed genotypic and phenotypic properties of morbilliviruses identified in the brain of different animal hosts and compared these with known properties of MeV associated with SSPE [6]. They concluded that similar evolutionary pathways across different animal species drive morbilliviruses to evolve analogous mechanisms favouring persistence in the CNS, followed by the development of chronic neurological disease [6]. This study illustrates the importance of a One Health approach for understanding the complex pathogenesis of neurological complications of measles.

Properties of MeV strains detected in SSPE patients

Although the route of MeV entry into the CNS remains unclear, genotypic and phenotypic analysis of MeV sequences detected in the brains of deceased SSPE patients have provided several important clues. Whereas MeV genomes detected during acute MeV infection are remarkable stable, MeV genomes detected in SSPE patients usually show biased hypermutation, often associated with ADAR1-profiles. Although the mutational profiles between MeV sequences detected in different patients are often divergent, they functionally converge in two major patterns: SSPE-associated MeV genomes encode non-functional matrix (M) proteins and hyperfusogenic fusion (F) proteins.

Inactivation of the M gene can be mediated by introduction of premature stop codons, corruption of the start codon, or biased hypermutation that results in generation of a biologically inactive protein. In addition, the cytoplasmic tails of the F protein often contain mutations. Combined, this leads to a phenotype in which the generation of virus particles is impossible, as the M protein interacts with the cytoplasmic tail of F when virions bud from the cell membrane of infected cells.

A second common property of SSPE-associated MeV sequences is that they encode destabilized F trimers, resulting in a hyperfusogenic phenotype. Interestingly, this often allows cell-to-cell fusion with membranes of cells that do not encode one of the high affinity cellular receptors (CD150 or nectin-4). In neurons the cell adhesion molecules CADM1 and CADM2, located at the synaptic cleft of neurons, can trigger the MeV fusion process in cis [7]. This may explain the slow but progressive spread of the MeV ribonucleoprotein (RNP) between neurons across the CNS.

Animal parallels of SSPE

The pathogenesis of animal morbilliviruses is similar to that of measles. As described earlier, infections of carnivores with CDV are frequently associated with acute neurological complications, which are more reminiscent of measles inclusion body encephalitis (MIBE) than SSPE. Whereas in vitro infection studies with MeV and CDV in organotypic brain slice cultures did not reveal an intrinsically higher neurotropic phenotype of CDV [8], CDV often overwhelms the adaptive immune system of the host [9, 10]. As a result, the host cannot clear the infection or prevent virus dissemination to peripheral tissues, leading to prolonged opportunities for the virus to enter the CNS.

Interestingly, late neurological complications reminiscent of SSPE have been described in morbillivirus-infected dogs, dolphins and seals [11–13]. Old dog encephalitis is a chronic CDV encephalitis associated with severe demyelination several years after the initial infection. Like in SSPE, the causative virus cannot be isolated from brain tissue [11], despite abundant presence of viral antigen [14]. Recently, similar persistent morbillivirus infections were described in harbour seals [13] and dolphins [15]. As described by Daniel Roji et al., these veterinary cases may act as natural experiments for understanding human neuropathogenesis in the situation where animal models that reflect SSPE pathogenesis are lacking.

Nonfunctional M proteins and hyperfusogenic F proteins in animal morbilliviruses

Daniel Roji et al. have characterized animal morbilliviruses that persisted in the CNS of their respective host species for several years after the initial infection: CDV from a lynx, CeMV from a striped dolphin and PDV from two harbour seals stranded in 2001 or 2014 (13 and 12 years after the PDV epizootics of 1988 and 2001, respectively; strikingly similar to the interval between acute measles and the development of SSPE). The viral genomes contained complex mutational profiles with several parallels to SSPE-associated MeV genomes: biased hypermutation or premature stop codons leading to inactivation of the M protein and mutations causing F protein thermal instability and hyperfusogenicity. These findings suggest a conserved pan-morbillivirus intra-host evolutionary mechanism leading to long-term persistence in the CNS.

In their study, the authors used several approaches to link the observed mutations to functional changes. They started by performing fusion assays in Vero cells in the presence or absence of expression of the relevant CD150 of seal, dolphin, dog or human origin, and demonstrated that in three out of four viruses fusion in receptor-negative cells was enhanced. Subsequently, the authors assessed the impact of hyperfusogenic F proteins on virus entry using vesicular stomatitis virus (VSV) particles pseudotyped with animal morbillivirus (CDV, PDV, and CeMV) or MeV glycoproteins with truncated cytoplasmic tails, demonstrating decreased entry in CD150-expressing cells.

Infection studies in CNS culture models

In follow-up experiments, the authors focused on the CDV detected in a lynx and assessed the role of three mutations in the F gene on the hyperfusogenic phenotype: the authors identified a single amino acid change (P479A) that fully explained the hyperfusogenicity. Virus isolation attempts from the brain tissue of the lynx were unsuccessful. The authors therefore generated recombinant CDV strains, demonstrating that the CDV strain containing P479A was indeed attenuated in Vero-CD150 cells and could only be passaged as infected canine B-cells. Despite reduced dissemination in these lymphoid cells, the virus showed enhanced infectivity in organotypic brain slice cultures from ferrets. These observations link directly to studies previously performed with SSPE-associated MeV strains: Schmitz et al. demonstrated that a chimeric wildtype MeV with the F gene from a SSPE case was no longer lymphotropic, but instead efficiently disseminated in human neural network cultures [16].

Conclusions

SSPE is a late neurological complication of measles, characterized by persistent MeV infection of the CNS. Although we have a good understanding of the pathogenesis of acute measles, and have several highly detailed case reports of the end stage of SSPE, information about progression during the clinically silent incubation time of several years is absent. Due to the lack of adequate animal models these events remain hard to study, hampering the efforts to design early diagnostic tools or identify biomarkers. The emergence of novel organoid-based brain culture models has increased options to analyze the phenotype of specific mutations, but does not provide information on whether these mutations are required for morbillivirus entry into the CNS (seeding mutations) or if these emerge during the long-term persistent CNS infection. A One Health approach to comparative pathogenesis of morbilliviruses in different host species could lead to novel options to study morbillivirus evolution and CNS persistence in natural infections, by identification of molecular signatures with predictive value. The ultimate goal of such studies would be to enable early diagnosis of persistent MeV infection of the CNS, allowing new intervention strategies using antivirals.

Author contributions

Both authors contributed equally to writing this comment.

Funding

Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethical approval

Not applicable.

Competing interests

The authors declare no competing interests.