Just because you plant some seeds doesn’t mean they’ll grow. In the case of the α-synuclein protein thought to seed toxic aggregates in Parkinson’s disease, the brains of non-transgenic mice failed to provide fertile soil, according to recent findings. Researchers led by Todd Golde and Benoit Giasson at the University of Florida in Gainesville came to this conclusion by injecting mice with α-synuclein proteins and monitoring the spread of toxic aggregates throughout the brain. Synuclein aggregates did spread in mice that already harbored an α-synuclein mutation and hence were prone to this pathology, but not in non-transgenic mice, the researchers reported March 23 in Acta Neuropathologica. The latter finding contradicts a previous study that described a prion-like spread of injected synuclein in normal mice. Giasson and colleagues also reported that the antibody used to monitor such synuclein spreading cross-reacts with another protein in the brain, a finding that questions the validity of previous studies that relied heavily upon the antibody, Golde said. 

The idea of spread and templated seeding of misfolded proteins has become a powerful trend line in neurodegeneration research of late, but Golde urges caution. “A simplistic view of synuclein spreading makes for a nice headline, but it’s probably far from the truth,” he told Alzforum. “Our study raises more questions than it answers, but it’s a sobering assessment of the challenges we need to think about.”

Some researchers approached the new data with caution, pointing to differences in methodology as the reason for the contradictory results.

Alpha synuclein is the primary component of Lewy body inclusions found in the brains of PD patients. The aggregates are suspected to trigger the loss of dopaminergic neurons in the substantia nigra, which leads to PD symptoms. However, evidence suggests that misfolded, pathogenic α-synuclein travels through other regions of the brain before reaching the substantia nigra. This spreading phenomenon has been observed for protein pathologies in other neurodegenerative diseases, as well (see Jucker et al., 2013). 

How α-synuclein inclusions spread from one brain region to another is the subject of intense research (see George et al., 2013). Previous studies revealing inclusions in fetal-cell grafts placed in the brains of PD patients hinted that pathogenic α-synuclein from the host could corrupt healthy cells (see Apr 2008 news story). This could have been due to the toxic microenvironment of the PD brain or to a prion-like mechanism. In 2012, two papers from Virginia Lee’s lab at the University of Pennsylvania in Philadelphia strengthened the prion proposition. First author Kelvin Luk and colleagues reported that injected α-synuclein triggered a spread of aggregate pathology throughout the brains of mice overexpressing the mutant A53T synuclein protein (see Luk et al., 2012). The researchers subsequently reported that the seeds could corrupt endogenous α-synuclein in wild-type mice, too, a finding that suggested the bad seeds could corrupt native α-synuclein  (see Nov 2012 news story). Luk and colleagues observed the aggregates spreading to new regions of the brain between one and six months after injection.

To replicate this in their own mice, first author Amanda Sacino and colleagues first went for the low-hanging fruit, that is, A53T mice. They injected α-synuclein fibrils into the hippocampus or cortex and stained sections of those brain areas one, two, and four months after injecting the fibrils. The investigators used an antibody (pSyn) specific for phospho-Ser129 α-synuclein, which has been widely used to measure toxic aggregates in previous studies, including Luk’s. The researchers observed that widespread dissemination of α-synuclein aggregates within the hippocampus as well as other regions, including the brainstem, spinal cord, thalamus, striatum, and cortex. Other antibodies against α-synuclein also revealed the spread.

The researchers next injected non-transgenic mice with the α-synuclein fibrils. They saw no spreading beyond the area of injection using the pSyn antibody. This held true when they looked for aggregates with other antibodies in their arsenal, as well. Contrary to the expanding α-synuclein pathology described in Lee’s study, Sacino only saw inclusions near the injection site, and those waned over time. The researchers saw no spreading in mice overexpressing a different α-synuclein mutant, E46K. This strain expresses α-synuclein at the same level as the A53T mice, indicating that something inherent to the A53T mutation, rather than expression level, caused the differences in spreading.

Several researchers attributed the conflicting results in wild-type mice to technical differences in the preparation of the α-synuclein fibrils. Sacino formed the fibrils by incubating α-synuclein protein for two days at 37oC with continuous shaking, whereas Luk incubated for five days. Both groups used sonication to break up the fibrils into fragments thought to be more readily taken up by neurons, but Sacino disrupted the fibrils in a waterbath sonicator, while Luk inserted a probe directly into the mixture to break up the fragments. Both groups used electron microscopy (EM) to visualize the shape of the fibrils before injection. Several researchers commented that those pictures looked dissimilar between the two studies. “If you look at the EM of the protofibrils they’re putting in, they look quite different,” Jeffrey Kordower of Rush University in Chicago, who was not involved in the work, told Alzforum. “They may be less potent, and if they’re less potent, maybe they spread less.”

Luk and Laura Volpicelli-Daley, who also worked in Lee’s lab, now each run their own labs at the University of Pennsylvania, Philadelphia, and the University of Alabama, Birmingham, respectively. They expressed similar concerns in a joint comment to Alzforum (see below). 

Giasson countered that his group selected their α-synuclein preparation method only after observing that it produced the most potent fibrils in cell culture systems, including primary neurons.  “In all the other systems, our preparation is more potent at inducing pathology,” he said. However, the researchers did not perform a comparison of fibril potency in mice, Giasson said.

While it’s clear that certain synuclein preps spread better than others, the reasons why are still unknown, said Patrik Brundin of the Van Andel Institute in Grand Rapids, Michigan. “The precise nature of these preparations is not fully understood. This is still an empirical science,” Brundin said. “Lee’s lab did it in a certain way that works.” Brundin added that his group replicated α-synuclein spreading in non-transgenic mice (albeit after injection into the olfactory bulb), but only when meticulously following the same preparation procedures used in Lee’s lab.

“If those little technical details matter so much, then we should perhaps consider this spreading phenomenon isn’t so robust in non-primed mice,” said Golde.

Not all researchers were quick to criticize Giasson’s methods. “I think this is a very carefully done study, and the authors are to be congratulated,” Marc Diamond of Washington University in St. Louis wrote to Alzforum. “They have appropriately raised caution regarding the interpretation of some of the data regarding ‘prion-like’ spread of protein seeds among neurons in neuronal circuits.” (See his full comment below.)

Sacino and colleagues discovered something else in their new study: The pSyn antibody stained white matter tracts from α-synuclein knockout mice just as readily as it stained those from A53T mice, suggesting that the antibody was binding something other than α-synuclein. The researchers performed biochemical analyses, including immunoprecipitation experiments from brain extracts, and determined that pSyn bound to phosphorylated neurofilament subunit L (NFL). The pSyn antibody poorly stained white matter tracts in NFL-/- mice, suggesting that NFL, rather than α-synuclein, was the primary target of the antibody.

Other researchers considered the NFL staining background noise. This is common with antibodies that recognize phosphate groups, Brundin said. Nolwen Rey, a postdoc in Brundin’s lab who measured spreading in non-transgenic mice with the pSyn antibody, claimed that optimizing the staining, including the use of very low antibody concentrations, was crucial.  “Optimizing the protocol was a very long process,” Rey said. She noted that Luk and colleagues did not show such NFL-like staining in α-synuclein knockout mice in their 2012 paper, suggesting that the labs used different protocols.

Giasson insisted that his group did perform the staining at multiple concentrations and the NFL staining endured. In Giasson’s injection experiments, the NFL staining pattern correlated with α-synuclein pathology in the A53T mice, suggesting that if the antibody is in fact binding to NFL, this protein may play a role in pathology. “For us this was a lucky finding,” Giasson said. “It hints that what’s happening there is more than just α-synuclein toxicity.” Giasson’s lab is investigating whether NFL or other proteins play a role in α-synuclein spread.

That α-synuclein pathology does not always readily spread in non-transgenic mice is a good thing, Golde said, “If we induce pathology and the brain clears it, perhaps we can harness that mechanism to clear pathology in humans.”—Jessica Shugart

Comments

Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. I think this is a very carefully done study, and the authors are to be congratulated. The authors confirm some essential observations that have been made by others regarding the injection of α-synuclein seeds into mouse brain.

    However, they have appropriately raised caution regarding the interpretation of some of the data regarding “prion-like” spread of protein seeds among neurons in neuronal circuits. Specifically, they find that a phospho-antibody used to detect synuclein accumulation can cross-react with neurofilament proteins enriched in white matter, and this may indicate that apparent spread of synuclein aggregates through white matter tracts could be due to an artifact. Further, they observe that reactivity to this antibody can be induced by non-amyloidogenic synuclein injection, which is not consistent with induction of endogenous synuclein fibrils.

    Finally, they raise the important caveat to all existing work with synuclein injection experiments that it is critical to rule out spread of injected seeds through circuits. This could occur by seed uptake from neuronal projections directly into projections to the injection site (e.g., axons) with retrograde transport to distant cell bodies, or it could occur by actual trafficking of injected aggregates into one cell, down axons, and across to connected cells, as happens with agents such as wheat germ agglutinin or viruses. The point is that in the future, in order to fully confirm “prion-like” spread, endogenous synuclein must be shown to move from cell to cell, spreading pathology, and the possibility of injected material doing this must be ruled out.

    On a final note, this study highlights the idea that using only one marker (in this case a phospho-antibody) to monitor protein accumulation is fraught with potential difficulties, and thus that it will be important to use multiple ways of determining induction of an amyloid state in brain proteins. This applies to studies of tau as well as synuclein.

    View all comments by Marc Diamond
  2. We were happy to see that the authors were able to recapitulate formation of α-synuclein Lewy body and Lewy Neurite-like inclusions by intracranially injecting synthetic fibrils made from recombinant protein into the hippocampus and cortex of both non-transgenic mice and mice overexpressing mutant human α-synuclein. However, these authors do not report spread of pathology as reported previously by our and other labs (Fujiwara et al., 2002; Luk et al., 2012b; Luk et al., 2012a; Rey et al., 2013), thereby claiming that the α-synuclein fibrils do not lead to a prion-like spread of pathology across brain areas. Additionally, Dr. Giasson’s manuscript claims that much of the pathology reported in previous publications largely results from cross-reactivity of the antibody mAB81A, which recognizes phosphorylated α-synuclein, a marker of inclusions. However, there are several methodological differences between Dr. Giasson’s manuscript and the previous studies (Luk et al., 2012b; Luk et al., 2012a).

    In this study, the protocol used to generate fibrils differs significantly from that used by other groups. For example, Sacino and colleagues only formed fibrils for two days. However, if using wild type α-synuclein, it takes at least five to seven days for optimal fibril formation (Wood et al., 1999). Furthermore, the fibrils used in their study were sonicated by mild-bath sonication. As we previously reported (Volpicelli-Daley et al., 2011Luk et al., 2012bLuk et al., 2012a), repeated sonication with a probe tip sonicator is crucial to initiate seeding and spreading of α-synuclein pathology. This is likely because it is necessary to generate small seeds that neurons can take up by macropinocytosis (Holmes et al., 2013). These differences in fibril preparations can easily be seen by comparing electron microscopy images from the Sacino et al. paper (Supplemental Figure 15) to those presented by Luk et al. (Figure 1C, Luk et al. 2012a).

    These methodological differences likely (as the images suggest) led to morphologically and functionally distinct species of α-synuclein being ultimately injected. Data from the labs of Virginia Lee (Guo et al., 2013) and Ronald Melki (Bousset et al., 2013) have demonstrated that vastly different conformations can be generated from identical protein preparations depending on the incubation conditions. Overall, it can be concluded that Dr. Giasson’s group did not obtain the same results because of these critical methodological differences.

    With respect to the mAB81A antibody, Dr. Giasson’s group is correct that when used at a high concentration, mAB81a and other antibodies that recognize phospho-α-synuclein have background staining both in primary neurons and in brain sections. It is indeed typical for most phospho-specific antibodies to have some background staining. However, when mAB81A is diluted to 1:10,000, there is minimal background staining in controls, and intense staining of inclusions in primary neurons treated with fibrils, and sections from fibril injected brains that is well above background. These inclusions can also be visualized with other p-α-synuclein-specific antibodies such as a rabbit polyclonal antibody developed by Virginia Lee’s lab (Volpicelli-Daley et al., 2011), an Epitomics rabbit antibody provided by the Michael J. Fox Foundation (MJFR-13), and antibodies Syn506 and Syn514 that specifically recognize pathologic inclusions (Waxman et al., 2008) and were first characterized by Dr. Giasson.

    In conclusion, we believe that severe methodological discrepancies have resulted in a failure to recapitulate the prion-like spread of α-synuclein observed by many other labs. This highlights the importance for a standardized protocol to be developed.

    View all comments by Kelvin Luk

References

News Citations

  1. Dopaminergic Transplants—Stable But Prone to Parkinson’s?
  2. Toxic Synuclein Corrupts Native in Wild-Type Mice

Paper Citations

  1. . Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature. 2013 Sep 5;501(7465):45-51. PubMed.
  2. . α-Synuclein: the long distance runner. Brain Pathol. 2013 May;23(3):350-7. PubMed.
  3. . Intracerebral inoculation of pathological α-synuclein initiates a rapidly progressive neurodegenerative α-synucleinopathy in mice. J Exp Med. 2012 May 7;209(5):975-86. PubMed.

Further Reading

Papers

  1. . Modeling Lewy pathology propagation in Parkinson's disease. Parkinsonism Relat Disord. 2014 Jan;20 Suppl 1:S85-7. PubMed.
  2. . Thinking laterally about neurodegenerative proteinopathies. J Clin Invest. 2013 May 1;123(5):1847-55. PubMed.

Primary Papers

  1. . Amyloidogenic α-synuclein seeds do not invariably induce rapid, widespread pathology in mice. Acta Neuropathol. 2014 May;127(5):645-65. PubMed.