Chronic and delayed neurological manifestations of persistent infections

Purpose of the review

Persistent infections capable of causing central nervous system (CNS) complications months or years after the initial infection represent a major public health concern. This concern is particularly relevant considering the ongoing coronavirus disease 2019 pandemic, where the long-term neurological effects are still being recognized.

Recent findings

Viral infections are a risk factor for the development of neurodegenerative diseases. In this paper, we provide an in-depth exploration of the prevalent known and suspected persistent pathogens and their epidemiological and mechanistic links to later development of CNS disease. We examine the pathogenic mechanisms involved, including direct viral damage and indirect immune dysregulation, while also addressing the challenges associated with detecting persistent pathogens.

Summary

Viral encephalitis has been closely associated with the later development of neurodegenerative diseases and persistent viral infections of the CNS can result in severe and debilitating symptoms. Further, persistent infections may result in the development of autoreactive lymphocytes and autoimmune mediated tissue damage. Diagnosis of persistent viral infections of the CNS remains challenging and treatment options are limited. The development of additional testing modalities as well as novel antiviral agents and vaccines against these persistent infections remains a crucial research goal.

INTRODUCTION

The central nervous system (CNS) is comprised of long-lived cells and is capable of modulating inflammatory processes [1] providing an ideal location in which pathogens can persist [2]. However, not all pathogens require presence in the CNS to cause neurologic disease. Emerging evidence from the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic (COVID19) has reinforced the concept that infections, irrespective of their anatomical location, can contribute to neurological disease in both acute and late phases [3].

Delayed onset of neurological issues due to infections can occur months or years after the initial infectious insult as highlighted by the recent striking finding that viral encephalitis is associated with significant risk of later development of dementia [4^▪▪] and that vaccination against some viruses can decrease the risk of developing dementia [5]. However, several barriers exist in distinguishing if the neurological symptoms are due to persistent infectious agents or postinfectious inflammatory sequelae. First, persistent infections, especially those in the CNS, are often underdiagnosed as they are not readily detected by conventional tests. Clinically, these may account for a significant proportion of the approximately 40–50% of cases of encephalitis that have no identifiable cause [6–9]. Second, during persistent infections there is ongoing inflammation which can mimic autoimmune or autoinflammatory processes leading to misdiagnosis of autoimmune encephalitis [10^▪]. However, both persistent infections and dysbiosis can also influence the development of autoimmune diseases, including CNS autoimmune diseases, through several different mechanisms including molecular mimicry and bystander activation [11–15]. It is important to understand the molecular mechanisms driving associations between persistent infections and neuro-immunologic diseases. Defining their overlap will allow for the development of diagnostic tools to help distinguish these processes. 

Box 1.

no caption available

INFECTIONS ASSOCIATED WITH LATE CENTRAL NERVOUS SYSTEM DISEASE

Multiple pathogens can cause pathology years after acute infection via direct damage through viral-induced cell death and dysfunction and indirectly by triggering immune-mediated damage (recently reviewed in [16,17]). The clinical phenotypes resulting from these processes are often distinct from the pathologies observed during acute infection. In this section we briefly review some of the known and suspected persistent infections associated with later development of CNS disease.

Human immunodeficiency virus

Human immunodeficiency virus (HIV) penetrates the CNS early during acute infection and infects microglia, macrophages, and to a lesser extent astrocytes, which can result in a macrophage-mediated encephalitis [18,19]. Clinically this is associated with a rapidly progressing subacute subcortical dementia [20] which is treatable with antiretroviral therapies (ART). However, even with adequate ART and suppression of virus both in the periphery and the CNS, HIV can cause CNS disease years after initial infection, collectively known as HIV-associated neurocognitive disorders (HAND). One risk factor for HAND is severe immune depletion during acute infection. This may be due to the relationship between lower CD4+ T-cell nadirs and a larger viral reservoir and more residual viral activity [21]. This larger and more active viral presence can lead to CSF viral escape and T-cell mediated encephalitis, which presents clinically with headache, tremors, cognitive impairment, confusion, focal neurologic deficits, and seizures [20]. However, even in the absence of viral replication, expression of viral proteins can lead to CNS disease. Viral proteins, such as Tat, gp120, and Nef, are produced despite ART, even from replication defective virus [22–24]. These proteins are toxic and alter neuronal and glial cell functioning (reviewed in [25–27]). Patients present with varying degrees of cognitive and psychomotor impairments that slowly progress over time, a spectrum disorder collectively known as HIV-protein associated encephalopathy [20]. As ART does not impact the production of viral proteins from integrated virus, the development of new therapies to inhibit viral transcription and translation are needed.

Herpes viruses

Multiple human herpes viruses can infect the CNS and acutely cause encephalitis. These viruses also establish persistent infections from which they can reemerge and cause viral shedding with or without recrudescence of disease [28–31]. Repeat reemergence of virus is associated with CNS pathology and the development of cognitive impairment in animal models [32]. However, even during latency herpesviruses express viral genes which can contribute to immune activation [31,33]. These viral-driven processes may contribute to the development of CNS diseases not associated with acute infection including multiple sclerosis [34^▪] and dementia [32,35]. Recent work suggests that these viruses also impact the development of CNS complications associated with other infections. For example, reactivation of Epstein-Barr virus (EBV) is associated with the development of post-acute sequelae of COVID-19 (PASC) [36]. Mechanistically herpesviruses are thought to initiate the formation of Aβ plaques [37–39] through stimulation of the innate immune response as well as drive the expansion of autoreactive T and B cells [40,41^▪▪]. However, data from epidemiological, pathology, and animal and in vitro models both support and refute the roles of herpesviruses in late-stage CNS disease [38,42–45].

Flaviviruses

Flaviviruses are typically vector-borne although sexual, transplacental, and transfusion transmission have all been documented [46,47]. With a world-wide distribution, a high rate of infection, and the ability to cause significant morbidity and mortality, this group of viruses are of great medical importance [48]. Flaviviruses such as Zika virus [49,50], Dengue virus [51–53], West Nile virus [54], and Chikungunya virus [55] have all been associated with CNS disease years after initial infection. In animal models, T cells post-viral clearance have been demonstrated to contribute to synaptic loss, neuronal apoptosis, and cognitive impairments [56]. These brain resident T cells, mostly CD8+ cells, produce interferon-γ, which triggers microglial cell activation. This, in turn, activates astrocytes, leading to glial cell dysfunction and specific patterns of neuronal and synaptic loss. The T cell is the main driver in this process, as cognitive impairments were prevented when CD8+ T-cells or microglial interferon-γ receptor levels were reduced after viral infection [56]. This ongoing T-cell induced damage to neurons following viral infection might explain the correlation between viral infections and the later onset of neurodegenerative processes [57].

Additionally, several flaviviruses are suggested to be present months or years after acute infection, although data demonstrating viral persistence from humans are limited. Preliminary evidence from animal models and case reports suggests that flaviviruses could reside in the CNS for extended times and are associated with progressive cognitive impairments [54,58–62]. A significant proportion, up to 80%, of flavivirus infections are asymptomatic [48]. This poses a challenge as the subsequent development of cognitive impairment may not be linked to a prior infection. To fully comprehend the long-term neurological effects of these viruses, it is imperative that future research investigates the frequency and persistence of flavivirus CNS infections.

Ebola

The recent outbreaks of Ebola have revealed that the majority of survivors have significant sequelae [63–65], including neurological impairments such as vocal outbursts, cognitive deficits, hallucinations, numbness, and fatigue [64,66,67]. In some patients this may be due to persistent virus [68]. Ebola can infect the CNS and can reemerge months to years later to cause significant pathology [69,70,71^▪]. However, how often this occurs or for how long the virus endures in the CNS are unknown.

Measles

Measles virus infection can lead to subacute sclerosing panencephalitis (SSPE), a fatal CNS complication which typically occurs 6 to 10 years after infection [72]. A mutated fusion protein allows for cell-to-cell spread within the CNS with significant viral damage resulting in cognitive decline, behavioral changes, motor dysfunction, and seizures [73]. Patients progress quickly to coma and death. As vaccinations have declined, a rise in SSPE has been observed [74^▪,75] and is predicted to increase over the coming years. SSPE occurs at an increased rate, with a shorter incubation period and a more fulminant form in children who acquire measles prior to two years of age [75–78].

Influenza

Influenza viruses can cause encephalitis and encephalopathy and are associated with significant levels of mortality and enduring neurological deficits [79–82]. Recent work demonstrated a striking association between influenza infections and later development of Alzheimer's disease, amyotrophic lateral sclerosis, generalized dementia, vascular dementia, and Parkinson's disease [4^▪▪]. As influenza is not thought to persist in immunocompetent hosts, this association may be due to damage sustained during acute infection. For example, influenza replication in neurons can lead to α-synuclein aggregation [83], which may set the stage for later Parkinson's disease development. However, an alternative hypothesis is that influenza, much like SARS-CoV-2, could remain persistent in a small fraction of cases. Although there are conflicting reports in the literature [84], influenza A was detected in macrophages in the CNS of patients with Parkinson's disease postmortem [85], although if this was a recently acquired or chronic infection remains unknown.

Severe acute respiratory syndrome coronavirus 2

Ongoing investigations have revealed that a subset of patients may have persistent SARS-CoV-2 infections. Biopsy and postmortem studies have demonstrated viral presence months after initial infection in the small intestine, heart, lung, and brain [86^▪▪,87–90] and continuous affinity maturation of immune responses, which require viral proteins, have been detected 6 to 14 months after infection [88,91,92].

This viral persistence is suggested to be one of the causes of PASC [93]. PASC is part of a spectrum of disorders known as post-acute infection syndromes [94] and shares some clinical features with myalgic encephalomyelitis (ME)/chronic fatigue syndrome [94]. One of the chief complaints among people with PASC is fatigue, a complex and poorly understood symptom associated with activation of the immune response which can occur in both autoimmune and infectious processes [94]. Patients with PASC also develop several other neurological complications including cognitive impairments, anosmia, myalgia, and autonomic dysfunction [95,96].

In patients with PASC, there are increased levels of both viral RNA and the spike protein as compared to patients with past COVID infection without PASC [97^▪▪,98^▪]. Viral RNA and spike protein levels increased from acute infection levels in patients who developed PASC. This is in notable contrast to patients who fully recover from COVID in which both viral protein and RNA significantly decreased as compared to acute infection [97^▪▪]. Collectively these data, along with case reports [99–101], suggest that in some patients with PASC persistent viral infection is occurring, and that they may benefit from combination antiviral therapies. Although the exact causes of PASC are likely complex and multifactorial [93,95], a deeper exploration of the subgroup of PASC patients with indications of persistent infection could offer crucial insights not only for those afflicted by PASC, but also for individuals experiencing other variants of post-acute infection syndromes.

PERSISTENT INFECTIONS AND AUTOIMMUNE PROCESSES

Infections are recognized as a susceptibility factor for autoimmune diseases, including autoimmune diseases of the CNS. For example, herpes simplex virus (HSV) infections are associated with the development of N-methyl-D-aspartate receptor-encephalitis [14,15]. Initially, it was believed that CNS-specific self-reactive T-cells remained ignorant, as antigens from the CNS were thought to be sequestered. However, this notion has been challenged following the rediscovery of the lymphatic system in the CNS [102] and the identification of deep cervical lymph nodes as the site for CNS antigen presentation [103]. Indeed, T-cells activated at the cervical lymph nodes can develop tropism for the CNS. These brain resident cells express less perforin and granzyme B and higher levels of inhibitory molecules programmed cell death protein 1 and cytotoxic T-lymphocyte associated protein 4 as compared to cells in the periphery [104]. Although the barrier for overcoming inhibition is quite high, upon antigen stimulation and recognition, these cells do produce effector molecules efficiently. Furthermore, brain resident T-cells can reactivate via cytokine signaling without stimulation at the lymph node [105,106^▪▪] and can be both protective and detrimental in the CNS (reviewed in [107]).

Persistent infections can improve the immune response to the pathogen by driving continued affinity maturation and refinement of antibody responses. However, this process can also result in autoimmunity. A recent example of this was described in the context of MS where continued affinity maturation of antibodies against the EBV protein EBV nuclear antigen 1 (EBNA-1) generated antibodies that cross-reacted with a host protein GlialCAM, a glial cell adhesion molecule [106^▪▪]. Given this finding, in combination with another recent large study which strongly supports the involvement of EBV in the pathophysiology of MS [34^▪], prevention of MS could potentially be achieve through vaccination. EBV vaccine trials are currently underway.

While autoreactive GlialCAM antibodies may contribute to pathology, other autoreactive antibodies may not. During CNS infections lymphocytes entering the CNS may undergo clonal expansion, driven in part due to ample secondary signaling from the recognition of pathogens in the CNS, a process known as bystander activation [108]. This may also give rise to autoreactive lymphocytes in the CNS and the detection of autoantibodies in the CSF that do not cross-react with the pathogen, despite concurrent presence of a persistent infection. For example, in a patient with chronic dengue virus encephalitis [61] multiple lines of evidence indicated the patient had high levels of CSF autoantibodies to microtubule-associated proteins that were confirmed biochemically (Fig. 1). As no pathogen could be detected in this patient initially, a diagnosis of autoimmune encephalitis was considered. Indeed, the misdiagnosis of autoimmune encephalitis is increasing [10^▪] some which may be due to immune abnormalities driven by persistent infections. It will be important to elucidate the factors that drive the development of autoantibodies and to determine the specific consequences of autoreactivity in each case.

FIGURE 1.

Autoantibodies are detectable from the sera and CSF of a patient with chronic dengue encephalitis. (A) Immunofluorescence images from human neuronal cultures were fixed with paraformaldehyde, permeabilized with acetone and then stained with mouse MAP6 antibody (red) and human CSF (1 : 25) from a patient with chronic dengue encephalitis (green). Dapi was used to visualize the nuclei (blue). Colocalization of mouse MAP6 antibodies with antibodies in the CSF was observed (bottom right, merged image). (B) Immunoblot images from reciprocal immunoprecipitations (IP). After approval by the Institutional Review Board at the NIH, adult human brain homogenate extracts (temporal cortex, 20 μg) from patients undergoing surgery for epilepsy, were incubated with three microliters of patient sera or mouse MAP6 antibodies for 1 h at 4°C. Antibodies were then captured with protein G beads and eluted from the beads by heating in LDS buffer. Proteins were resolved by gel electrophoresis and probed with either patient sera (1 : 1,000) or mouse anti-MAP6 (1 : 1,000) overnight at 4°C. Horseradish peroxidase (HRP) secondary antibodies (antihuman HRP or antimouse HRP; 1 : 5000) were incubated at room temperature for 1 h. Membranes were developed with enhanced chemiluminescence and visualized on a ProteinSimple FluorChem.

DIAGNOSTICS AND TESTING

Diagnosing CNS infections remains a challenge. Common modes of testing currently in use for the detection of CNS infections primarily utilize the CSF (Table 1). While biopsied tissue samples are often considered optimal for diagnosis, they are invasive and oftentimes risky for the patient. Serum testing has limited value as it does not typically yield CNS-specific information.

Table 1.

Sensitivity and specificity of detecting representative pathogens from the CSF.

Test	Fluid		Sensitivity (%)	Specificity (%)	Other results/comments	References
BioFire ME panel (Multiplex PCR)	CSF	S. pneumoniae	87.5	98.5		[109▪]
		H. influenzae	64.9	99.4
		S. agalactiae	71.5	99.5
		E. coli	70.9	99.6
		N. meningitidis	74.5	99.1
		L. monocytogenes	70.4	98.9
		Enterovirus	93.8	99.3
		HSV-1	75.5	99.9
		HSV-2	94.4	99.9
		VZV	91.4	99.8
mNGS (UCSF)	CSF		N/A	N/A	13/58 (22%) CNS infections identified by mNGS that were not identified by clinical testing; 19/58 concurrently diagnosed by mNGS; 11/58 diagnosed by serologic testing only; 7/58 diagnosed by tissue samples; 8/58 negative on mNGS due to low CSF pathogen titers	[119]
VirCapSeq	CSF		N/A	N/A	VirCapSeq-VERT detected virus in the CSF of 16/38 (42%) of patients with no pathogen identified previously; In at least 50% of these cases the virus identified was possibly causative.	[123]
JC virus PCR	CSF		74–92	92–100		[124,125]
JC virus PCR (ultrasensitive)	CSF		>95	N/A		[126]
HSV PCR	CSF		98–100	94–100		[110,127]
VZV IgG	CSF		93	N/A		[128]
VZV PCR	CSF		30	N/A
SARS-CoV-2 PCR	CSF		N/A	N/A	Positive in 0/76 patients	[129]
					Positive in 0/21 patients	[130]
Ebola PCR	CSF		N/A	N/A	PCR was negative in 0/7 Ebola survivors in 2017 publication	[131]
VDRL	CSF		67–72	80–98		[132]

CFU, colony forming units; CSF, cerebrospinal fluid; HSV, Herpes simplex virus; mNGS, Metagenomic next-generation sequencing; PCR, polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; VDRL, venereal disease research laboratory test; VirCap Seq, virome capture sequencing; VZV, varicella-zoster virus.

Polymerase chain reaction (PCR) testing for microbial genetic material in the CSF is often considered the gold standard for diagnosis of neurological infections. However, while this is true for certain organisms, caution must be taken as this assay has suboptimal sensitivity for a number of pathogens. The BioFire FilmArray Meningitis/Encephalitis (ME) Panel is widely used in cases where infectious meningitis and/or encephalitis are suspected. It employs multiplex PCR of CSF samples to detect multiple bacterial and viral organisms. However, a recent meta-analysis showed suboptimal detection for several organisms in the panel, including <75% for five bacterial organisms and 75.5% for HSV-1 [109^▪], which has a sensitivity of 98% on conventional PCR testing [110,111]. Another example that illustrates the shortcomings of PCR is varicella-zoster virus (VZV) vasculopathy, in which IgG detection in the CSF demonstrated a three-fold sensitivity over PCR [112]. A more nuanced case of PCR limitations is the diagnosis of PML. Conventional PCR methods are often used for confirmation in suspected PML cases and have a sensitivity of 74–92% [113,114]. However, ultrasensitive PCR methods (sensitivity >95%) [115] have recently been developed, enabling earlier therapeutic intervention and avoiding the need for invasive brain biopsies.

Antibody detection and quantification represent another testing modality. However, the assays used often lack the accuracy desired for clinical diagnosis. For example, the current Lyme antibody testing paradigm is labor intensive and yields both low sensitivity and specificity, and West Nile Virus immunoassays demonstrate poor specificity due to cross-reactivity with other flaviviruses [116]. However, antibody testing may be invoked more frequently for SSPE in coming years, given the expected rise in incidence of this pathology. Established prior data dictates that SSPE may be diagnosed with specific antibody titers of 1:256 or greater in serum, and 1:4 or greater in CSF [117]. Further, antibody testing by VirScan, a phage-display screening tool, has been useful in identifying immune responses to viruses in the CSF which can implicate etiologic agents in cases which have no identified cause [61,118].

Metagenomic next generation sequencing (mNGS) represents a new frontier in detection of CNS microbial infections. The clear benefit of this analysis is seen in cases where uncommon or novel pathogens are implicated in perplexing cases. However, like the BioFire ME panel, sensitivities for certain organisms via mNGS can be lower compared to standard tests for specific organisms, especially when the pathogen load is low [119,120]. Improvements in sensitivity have been demonstrated by using cell-free nucleic acids from CSF as compared to cell-associated DNA [121,122]. Making metagenomics and Virscan diagnostics more accessible could have a significant impact on the diagnosis and management of infectious diseases, especially in resource-poor settings, potentially improving health outcomes and saving lives.

CONCLUSION

In conclusion, the long-term neurological sequelae of persistent infections represent a significant public health concern, particularly in the context of emerging infections such as COVID-19. Despite advancements in our knowledge of the immune response and pathogenesis of these infections, much remains unknown about how persistent infections result in later CNS disease. However, recent findings suggests that there is a window in which intervening by targeting the persistent infections could prevent later development of autoimmunity or neurodegeneration. This is a complex and challenging area of study that requires further research in order to fully understand the mechanisms and develop effective treatments and prevention strategies to mitigate long-term neurological damage. In addition, there is a need for improved diagnostic tools and biomarkers to aid in the identification and monitoring of these infections.

Acknowledgements

This work was supported by the Intramural Research Programs of the NIH National Institute of Neurological Disorders and Stroke.

Financial support and sponsorship

None.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest

• ▪▪ of outstanding interest