In this study we perform multiphoton laser scanning microscopy and measure α-synuclein levels in the living brain using a previously described enhanced Green Fluorescent Protein (GFP)-tagged human α-synuclein (Syn-GFP) transgenic mouse line called PDNG78 
. This initial characterization demonstrated that fusion of GFP to human α-synuclein's C-terminus and expression under the human Platelet Derived Growth Factor promoter lead to robust expression in a subset of cortical neurons. In addition, this previous work demonstrated increased transgene expression at the mRNA level at ~3-fold higher levels compared to that found in human brain 
Before starting in vivo
imaging we used high resolution confocal imaging in fixed cortical tissue to determine the cellular localization of Syn-GFP. This analysis demonstrates that Syn-GFP is detectable in the cell bodies of a sparse subset (~1–3%) of layer 2/3 cortical neurons and in multiple neuropil puncta (). Visual inspection shows that the vast majority (>99%) of these puncta are contained within axon-like structures and when paired with previous work showing colocalization of Syn-GFP puncta with the presynaptic terminal marker synaptophysin and electron microscopic localization of Syn-GFP at presynaptic terminals 
strongly suggests that most neuropil puncta represent presynaptic accumulations of Syn-GFP. Confocal analysis of fixed Syn-GFP tissue also reveals that immunohistochemical staining for human α-synuclein and GFP colocalize as would be expected for this fusion protein ().
Syn-GFP is present in neuronal cell bodies and presynaptic terminals.
Multiphoton imaging demonstrates that a similar pattern of Syn-GFP localization in a sparse subset of neuronal cell bodies and in neuropil puncta can be detected in vivo () as in fixed tissue. In the in vivo case, however, we are unable to visualize individual axons given the lower signal-to-noise ratio of these structures. Another difference between the two techniques is the greater number of neuropil puncta visualized within the plane of focus in vivo compared to fixed tissue within a unit area (). This is likely because of the decreased z-axis resolution of in vivo multiphoton imaging compared to confocal microscopy.
In order to determine if the density of Syn-GFP positive neurons or presynaptic terminals changes as animals age we measured their respective densities in layer 2/3 of cortex over time. A loss of Syn-GFP positive cells or terminals over time could be a manifestation of α-synuclein-mediated neurodegeneration. In general genetic models of PD have not shown large amounts of frank cell loss 
but previous studies have not used similar in vivo
techniques to follow cell or synapse number, which may be more sensitive. In our first cross-sectional study we found that over a period of more than 1 year the density of Syn-GFP expressing neurons (age 2 months: 9830
6 animals, age 3 months: 10090
5 animals, age 14–17 months: 10200
5 animals; ) and presynaptic terminals (age 4 months 4.56
5 animals, age 9–12 months 4.75
4 animals; ) did not change significantly. In addition, to further characterize the Syn-GFP neuropil puncta we plotted the distribution of mean fluorescence intensity for each punctum within a field of view in an animal 4 months old (, top panel) and this distribution is positively skewed with a tail towards higher mean puncta intensities. This indicates that Syn-GFP levels can vary by several-fold in different presynaptic terminals. Even with the wide range in Syn-GFP terminal levels detected the general shape of this distribution also did not vary with age (). Next we performed chronic imaging of the same region of cortex over time. This second longitudinal analysis shows that the pattern of expression of Syn-GFP in particular neuronal cell bodies is stable over a period of weeks (). A total of 42 individual Syn-GFP positive cell bodies (n
2 animals) were followed within a total volume of cortex of 4×106
over 49 days and all 42 cells present on day 0 were also present on day 49. In addition, no additional new Syn-GFP positive neurons were detected within this volume. These results suggest that there is no large scale neurodegeneration or synapse loss in Syn-GFP expressing neurons at the ages tested.
Syn-GFP positive cell body and presynaptic terminal density does not vary with age.
Syn-GFP cell body and high intensity presynaptic terminal expression is relatively stable over time.
The analysis of individual Syn-GFP positive presynaptic terminals over days to months is more complicated than that for Syn-GFP positive cell bodies (described above) since the higher density of terminals makes it difficult to follow the same terminals over the course of days. This is in part because of the small differences in imaging conditions that are inherently present from day to day and the small amount of movement of individual terminals relative to each other that likely occurs. However, we have found that if only the high intensity terminals (those expressing the most Syn-GFP) are selected their density is low enough to follow individual high intensity terminals over time (). This analysis of high intensity terminals showed that individual terminals could be followed over a period of months and that some terminals were stably present while others were lost over time ().
Two different “window” techniques have been developed to perform in vivo
multiphoton imaging in the cortex of rodents, each with its own advantages and disadvantages 
. We tested whether the “glass coverslip” cranial window approach used in this study might produce detectible changes in Syn-GFP over time. Different changes possibly related to the window placement process itself have been suggested in some model systems 
. In our case, however, repeated imaging of cortical Syn-GFP did not produce any noticeable changes in either the pattern of expression (), density of labeled Syn-GFP cell bodies (day 0 post-window: 10030
16 animals, 6 months post-window: 9130
3 animals; ) or presynaptic terminals (day 0 post-window: 4.73
6 animals, 6 months post-window: 4.49
3 animals; ). In addition, the distribution of mean fluorescence intensity for Syn-GFP neuropil puncta maintained the same general shape with time after window placement ().
Syn-GFP expression pattern, cell body and terminal density do not vary with time post-window placement.
Together all the data presented above suggest that this experimental approach provides a powerful new paradigm for visualizing α-synuclein in the living brain, something difficult to do with other methods.
In order to better characterize the mobility of α-synuclein within neurons in our system and compare this to what has been reported in the literature for hippocampal neurons in culture 
and C. elegans
body wall muscle 
, we used the fluorescence recovery after photobleaching (FRAP) technique in vivo
to photobleach Syn-GFP in individual presynaptic terminals and measure the recovery of this signal over time. Given the positively skewed distribution of mean terminal Syn-GFP intensities we observed () and previous work from C. elegans
suggesting two different populations of inclusions in body wall muscle we decided to selectively photobleach two different populations of presynaptic terminals, high intensity terminals (those in the top decile of the intensity distribution) and “normal” terminals (those outside the top decile). After photobleaching both kinds of terminals we measure a similar rate of recovery of Syn-GFP signal with a t1/2
~2 min () in both cases. This is slower than has been reported in dissociated hippocampal cell cultures for GFP-tagged α-synuclein at presynaptic terminals 
and in C. elegans
body wall muscle for yellow fluorescent protein (YFP)-tagged α-synuclein 
. This suggests a difference in presynaptic terminal α-synuclein mobility in cortical neurons in vivo
compared to neuronal culture or in worm body wall muscle, the possible causes of which are discussed below. In our experiments fluorescence signal did not recover fully to its pre-photobleaching baseline and the fractional level of recovery was significantly different between high intensity and normal terminals (fractional recovery at 8 min normal intensity terminals: 0.76
15 terminals; high intensity terminals: 0.20
10 terminals; t-test p<0.0001; n
5 animals). These fractional levels of recovery are similar to those reported in the two previous studies and can be related to the fraction of immobile α-synuclein species present at terminals, in our case ~25% in normal terminals and ~80% in high intensity terminals.
Syn-GFP presynaptic terminal signal recovers differently after photobleaching.
Next we tested the turnover and mobility of α-synuclein within neuronal cell bodies in our system using the FRAP technique. First we photobleached Syn-GFP throughout the entire cell body and then measured its time course of recovery. The rate of α-synuclein synthesis within cells is an issue of importance since multiple mechanisms controlling this synthesis have been postulated to play a role in PD pathogenesis 
. We measure a t1/2
of recovery ~1 hr after photobleaching Syn-GFP throughout the soma (). Given this slower time course the source of Syn-GFP contributing to this recovery is not clear since it may represent the synthesis and maturation of new Syn-GFP molecules, the movement of unbleached Syn-GFP located in other regions (e.g. presynaptic terminals) back to the cell body after the establishment of a concentration difference by somatic FRAP, or a combination of both. In order to better characterize the mobility of α-synuclein within neuronal cell bodies and determine how this movement might play a role in the rate of Syn-GFP recovery after whole-cell bleaching we photobleached Syn-GFP in a region encompassing only one half of the soma and measured its recovery. Somewhat to our surprise, imaging cell bodies after half the soma had been photobleached (bleaching pulse ~5 sec duration) demonstrated that Syn-GFP within the entire cell body had been greatly reduced to a level equivalent to that seen with whole-cell bleaching (relative signal post whole-cell bleach: 0.39
20 cells; half-cell bleach: 0.38
3 cells; n
3 animals; ). In addition, the t1/2
of recovery after half-cell bleaching was essentially identical to that seen with whole-cell bleaching (). These results strongly suggest that α-synuclein is rapidly mobile within the somatic compartment since essentially all somatic Syn-GFP molecules visit the bleached half of the cell during the 5 sec long bleaching pulse. Given this rapid mobility it is possible that redistribution of Syn-GFP from unbleached regions back to the soma plays a role in the recovery of signal after whole-cell bleaching. Determining the relative contribution of this process (vs. new synthesis and maturation of Syn-GFP molecules) will be an interesting avenue for further study.
Syn-GFP somatic signal recovers slowly after whole-cell photobleaching and is rapidly mobile during half-cell photobleaching.