New Research: H5N1 influenza virus induces PD-relevant neurodegeneration

Dr. Richard Smeyne spoke with Dr. Kirsten Carlson of MJFF about his new publication in PNAS: "Highly pathogenic H5N1 influenza virus can enter the central nervous system and induce neuroinflammation and neurodegeneration".

KC: What are the most important takeaways from your results?

RS: The most important finding of this work is that we show that certain viruses can enter the CNS and directly induce pathological changes that are reminiscent of those seen in Parkinson’s disease.  This includes loss of dopaminergic neurons in the SNpc, long-lasting activation of microglia in areas infected by virus, as well as upregulation of alpha-synuclein (with aggregation) in these same regions.

KC: What do your data indicate from an epidemiological standpoint? Specifically, how might these results relate to the Spanish flu outbreak of 1918?

RS: Following the 1918 Spanish flu, which resulted from an H1N1 virus, there were clear cases of encephalitis.  Although controversial, there is a clear epidemiological tie to development of von Economo’s encephalopathy and later development of post-encephalic parkinsonism.   Although work on the reconstituted H1N1 Spanish flu virus did not show direct neurotropism, infection with this agent did induce a cytokine storm in the CNS with activation of microglia.  On a broader level, it has long been hypothesized that viruses may be an etiological factor in Parkinson’s disease.  This hypothesis is offered due to the presence of parkinsonian clusters, where small groups of people in close contact develop Parkinson’s disease in significantly higher incidences that would be predicted from the general population.  Our findings provide direct proof that this particular influenza virus is the etiological cause of the parkinsonian pathologies.

In this study we infected mice with the A/Vietnam/1203,1204 strain of the H5N1 virus.  Thus, while there can be no direct tie since a different virus was used there are some general principles that can be taken from our work that ties to the 1918 H1N1 influenza pandemic.  First, naturally occurring viruses can enter and affect the CNS.  At least with the virus we studied, the pattern of infection mimics the hierarchical progression of the disease as described by Braak, with infection starting in the enteric nervous system, traveling into the CNS via the vagus nerve where it first infects that solitary and dorsal motor nucleus of X, and then spreading from the brain stem into higher regions of the CNS.  Secondary to the entrance of the virus from the enteric system, we also have some evidence that the virus can enter the CNS through the olfactory nerve, which makes sense since mice are inoculated with influenza via the nose.

KC: Given the recent attention to various influenza strains and their potential for harm, what, in your opinion, are the broad implications for risk of neuropathogenesis as a result of airborne threats?

RS: I don’t want to alarm the public by overstating the risk of developing Parkinson’s disease from all strains of influenza virus, as each must be examined as their own entity.  That being said, it is clear that the vast majority of Parkinson’s disease occurs due to an unknown etiology (non-familial PD).  Epidemiology has shown that exposure to organics such as herbicides, pesticides and pescicides can increase the risk of developing PD and these studies are supported by experimental modeling of the disease in species ranging from mice and non-human primates.  I think our work has shown that viruses can be an etiological risk factor for PD.   However, each virus strain will need to be examined to determine its neuropathological consequences.

KC: What are the clinical implications of your findings?

RS: This is a tough question to answer.  On a broad level, as stated before, I think our research has shown that viruses can be an initiating factor in PD.  What we don’t want to do is generalize our findings to every strain of influenza.  This being said, perhaps it would be worthwhile to examine neurotropism (and its consequences) in each major emerging influenza strain to identify those that could have repercussions beyond the respiratory symptoms common to each of them.  If we were able to identify the strains that were neurotropic perhaps we could better identify genetic fingerprints that would allow us to predict what strains of viruses could potentially impact the CNS.

KC: What are your priorities for next steps in this research?

RS: We have three major aims that we are exploring.  First, in research supported by the Michael J. Fox Foundation for Parkinson’s Research we are examining, in detail, the pattern, timing and transport of H5N1 virus from the enteric route as well as the olfactory route.  Second, we are examining the profile of cytokine expression in the brain at different times following infection with H5N1, ranging from short term infection (days post infection 1-21) as well as longer term studies (says post infection, 30, 60, 90 and 120).  Third, we are exploring the mechanism by which infection with H5N1 virus kills cells.  We have shown in our recent paper that regions infected with H5N1 express activated caspase-3.  What we don’t know at this time is if the cells that are dying are those infected by H5N1 or are bystanders. 

Dr. Richard Smeyne's new publication in PNAS that influenza infection results in neuronal degenration similar to PD is very interesting.  They demonstrate that neuroinflammation in the absence of viral infection was sufficient to cause neuronal degenration.  Dr. Isacson's response (Responses: 1 Influenza virus triggers Parkinson's like conditions and potential pathology By: Ole Isacson, Harvard)  briefly summarised how inflammatory responses enhance neuronal death.

These observations as well as recent studies by Reynolds et al and Brochard et al who clearly established CD4 T cell contribution (in addition to others) to the progression of PD bring out new ways of looking at PD.  Is there a possibility of viral-induced molecular mimicry in PD?  If so, is there an autoimmune component contributing to neuronal death in PD?  Are the autoantigens derived from altered forms of alpha-synuclien? Many interesting questions with a few answers..

Author recognized relevant financial relationship in this content.

           The interview with Richard Smeyne and the associated response by Ole Isacson point to a greater role for peripheral factors in Parkinson’s disease.  Indeed, the long-standing implied relationship with the 1918 panencephalitis and the recent study by Smeyne and colleagues suggesting that neurotropic viruses lead to neuroinflammation and dopamine cell loss suggest an environmental cause.  These studies are also consistent with those by John Hong’s group demonstrating that in males, a single high dose endotoxin (lipopolysaccharide; LPS) treatment or multiple high dose exposures in females in early life produce delayed losses of dopamine neurons accompanied by traditional neuroinflammation. Since both the viral exposure and the endotoxin exposures produced similar DA neuron loss, perhaps the issue is more related to chronic neuroinflammation as implied by Isacson.  We also showed that rats with long-standing neuroinflammation due to prenatal LPS were more sensitive to rotenone induced cell loss in later life.  All these studies are consistent with the multiple hit hypothesis that Parkinson’s disease is a consequence of environmental events that set up a neurodegenerative process in the CNS and that high dose or multiple low dose exposures may be needed to initiate the cascade.  These studies are not inconsistent with the notion that genetic risk factors may interact with environmental factors as well.       

            Both Smeyne’s and Hong’s studies point to a greater role of the periphery in Parkinson’ disease, whether it be through vagal nerve or nasal entry or passage across the blood brain barrier (BBB).  One of the major problems with this notion, however, becomes “why is it specific to the motor pathways?”  Although the long-standing argument has been the higher numbers of microglia in the nigra, it remains difficult to fathom such a highly directed response unless we assume that the virus and endotoxin are motor pathway specific, which is clearly not the case with LPS.  An alternative is that the neuroinflammation leads to the formation of neoantigens resulting from free radical attack on normal DA neuron constituents that then leak into the periphery to produce a directed adaptive immune response. Indeed recent studies suggest such a role for adaptive immunity in such a cascade.  Thus, studies by Brochard et al and Benner and colleagues point to a role for peripheral T-cells that would be specific for a targeted neuronal population.  Again this would be consistent with the multiple hit hypothesis where an initial inflammatory response leads to the formation of neoantigens and T-cell response that then enter the brain as a second hit.  This notion is also consistent with recent studies suggesting involvement of a compromised or dysfunctional BBB that several authors have now shown to occur in animal models of Parkinson’s disease.  Although Isacson’s group has recently shown that the BBB is intact in MPTP monkeys using imaging, this does not negate the possibility of small areas of punctate leakage which characterizes the animal models or alterations in function of the barrier shown thus far.  In regard to the BBB issue, since all neurodegenerative diseases thus far studied exhibit BBB dysfunction, why would Parkinson’s disease be different?  Regardless, taken together, the studies alluded to above, point to a greater role for the periphery in Parkinson’s disease and whether it is viral, bacterial, toxin induced, or an interaction with a genetic risk factor, the common denominator remains neuroinflammation.  What then causes the specific pattern of degeneration seen in Parkinson’s disease remains the central issue.  The study by Smeyne suggests the involvement of elements in the periphery and suggest that those of us who study Parkinson’s disease should become less neurocentric.  Perhaps peripheral factors are responsible for the targeted central response.