Exp Neurobiol 2018; 27(3): 217-225
Published online June 30, 2018
© The Korean Society for Brain and Neural Sciences
1Division of Neurobiology, Department of Molecular and Cell Biology, Helen Wills Neuroscience Institute, University of California, Berkeley, CA 94720, USA, 2Molecular Neurobiology Lab, Department of Structure and Function of Neural Network, Korea Brain Research Institute, Daegu 41062, Korea
Correspondence to: *To whom correspondence should be addressed.
TEL: 82-53-980-8450, FAX: 82-53-980-8339
Deficient BDNF signaling is known to be involved in neurodegenerative diseases such as Huntington's disease (HD). Mutant huntingtin (mhtt)-mediated disruption of either BDNF transcription or transport is thought to be a factor contributing to striatal atrophy in the HD brain. Whether and how activity-dependent BDNF secretion is affected by the mhtt remains unclear. In the present study, I provide evidence for differential effects of the mhtt on cortical BDNF secretion in the striatum during HD progression. By two-photon imaging of fluorescent BDNF sensor (BDNF-pHluorin and -EGFP) in acute striatal slices of HD knock-in model mice, I found deficient cortical BDNF secretion regardless of the HD onset, but antisense oligonucleotide (ASO)-mediated reduction of htts only rescues BDNF secretion in the early HD brain before the disease onset. Although secretion modes of individual BDNF-containing vesicle were not altered in the pre-symptomatic brain, the full-fusion and partial-fusion modes of BDNF-containing vesicles were significantly altered after the onset of HD symptoms. Thus, besides abnormal BDNF transcription and transport, our results suggest that mhtt-mediated alteration in activity-dependent BDNF secretion at corticostriatal synapses also contributes to the development of HD.
Huntington's Disease (HD) is a neurodegenerative disorder caused by the mutation of the huntingtin gene (
Neurotrophin signaling is known to have a protective function in HD by promoting striatal neuronal survival [5,6]. Deficiency in brain derived-neurotrophic factor (BDNF) is linked to diverse brain dysfunctions, and the close relationship between HD pathology and BDNF loss has been extensively demonstrated [5,6,7,8]. In order to reverse the lowered striatal BDNF level and striatal atrophy in the HD brain, various manipulations for increasing BDNF signaling have been found to enhance cell survival and alleviate HD symptoms in HD mouse models. One of approaches to recover the striatal BDNF level is overexpressing
In this study, I have utilized a fluorescent BDNF sensor, that is expressed by a viral vector encoding
Animal protocols were approved by the Animal Care and Use Committee of University of California, Berkeley. All mice were purchased from Jackson Laboratory except the Q140 and Q175 mice, both of which were provided by the CHDI Foundation and shipped via Jackson Laboratory.
AAV-DIO-BDNF-pHluorin was used as described previously . AAV-DIO-BDNF-EGFP was constructed by replacing the pHluorin fragment into the PCR-amplified EGFP fragment. Sequencing and restriction enzyme reactions were performed to verify the plasmid. Custom packaging and purification of both BDNF sensors were performed through UNC Vector Core. 500 nl (per each hemisphere) of AAV-packaged BDNF sensors were bilaterally injected into the M1 of
Injection of 700 µg of ASO (in 200~500 nl of PBS) targeting htt [21,22], which was provided by ISIS (now IONIS) pharmaceutical via the exclusive material transfer agreement (MTA) among the UC Berkeley, CHDI, and ISIS pharmaceutical, was performed in the right lateral ventricle of Q175 mice. As a control, the same volume of PBS was injected. One week later a single bolus ASO or PBS injection, the same mice were again injected with AAV-hSyn-BDNF-EGFP into the motor cortex (M1). After additional three more weeks to achieve full BDNF-EGFP expression in the M1 cortex, striatal slices were then prepared to examine activity-induced changes in BDNF-EGFP fluorescence using a two-photon microscope.
Standard artificial cerebral spinal fluid (ACSF) consisted of (in mM) 130 NaCl, 3.5 KCl, 1.25 NaH2PO4, 24 NaHCO3, 2 CaCl2, 2 MgCl2, and 10 glucose (pH 7.3). Mice were deeply anesthetized with isoflurane, and then transcardially perfused with ~20 ml of slicing ACSF (ACSF containing 10 mM Mg2+ and 0.5 mM Ca2+) before the brain was dissected. Parasagittal striatal slices (400 µm thick) were prepared using a vibratome (Leica) using ice-cold slicing ACSF (below 4℃) and maintained at 30~32℃ in normal ACSF for 1 hour before electrophysiological recording or two-photon imaging.
Two-photon laser-scanning microscopy was performed using an LSM 510 META/NLO Axioimager system (Zeiss; Molecular Imaging Center at UC Berkeley) equipped with a Spectra-Physics MaiTai HP DeepSee laser (700 to 1,020 nm) and 403 water-immersion infrared objective (NA 0.8). BDNF-pH or BDNF-EGFP was excited by the 880 nm laser. The emission signals of BDNF-pH or BDNF-EGFP were acquired by using 500~550 nm band-pass filter. The field of view (512×512 pixels, 0.21 mm/pixel, 0.8 ms pixel time) was chosen in the striatal slice where cortical projections remained intact and BDNF-pH or BDNF-EGFP was significantly expressed at synaptic bouton-like structures (1~2 µm). Slices were placed in a recording chamber, submerged, and continuously perfused (2~3 ml/min) with oxygenated ACSF (containing 100 µM picrotoxin to isolate the glutamatergic synaptic transmission) at room temperature (20~25℃).
To record changes in BDNF-pH or BDNF-EGFP intensity in response to electrical stimulation, I acquired at least 100 consecutive images (at 1 Hz) as a baseline, then applied electrical stimulation (HFS: 4 trains of stimuli spaced at 10 s intervals, with each train containing bursts of 100 spikes at 100 Hz) using a tungsten bipolar electrode (WPI) placed on the cortical layer 6 close to the white matter, and then at least additional 200 images at 1 Hz after stimulation were taken. In some experiments, iso-osmotic ACSF containing 50 mM NH4Cl (pH 7.4) was applied to identify axonal BDNF-pH or BDNF-EGFP at the end of each experiment.
Images were processed and analyzed with ImageJ software (NIH). I presented data as normalized fluorescence changes (ΔFt/F0), in which fluorescence changes (Δ
Statistical analyses were performed by using Prism 6.0 software (GraphPad). Unpaired Student's t-test and one-way ANOVA with post-test were used for testing significance between two groups and among three or more groups, respectively.
To test whether activity-dependent secretion of BDNF is altered in the brain of HD model mice, BDNF secretion was monitored with acute striatal slices containing corticostriatal projections as previously described , using a knock-in type of HD model mice (
After applying electrical stimulation to cortical axons with the high-frequency stimulation (HFS) protocol (Fig. 1C,
Next, to explore the mechanism underlying differential effects of mhtts on BDNF secretion from cortical axons vs. striatal MSN dendrites, I analyzed behaviors of individual BDNF-containing vesicles by performing a population analysis of BDNF-pH puncta, which are indicative of BDNF-pH containing secretory vesicles [17,24]. I was able to categorize fluorescence changes in BDNF-pH puncta into three differential secretion modes: partial-, full-, and no-fusion mode (Fig. 1F, above; see Materials and Methods for more details) .
I found that the major secretion mode of pools of BDNF-containing vesicles (BDNF-vesicles) in cortical axons and MSN neurites of WT mice was full-fusion (Fig. 1F; the fraction of the full-fusion mode, cortical, 0.60±0.09; striatal, 0.49±0.06). However, in Q140 mice the major secretion mode of cortical BDNF-vesicles was the partial fusion, and I found that an increase in the partial fusion mode of BDNF-vesicles in Q140 mice was statistically significant (Fig. 1F; the fraction of the partial-fusion mode, WT, 0.20±0.07; Q140, 0.52±0.09; p<0.05, unpaired t-test). On the other hand, in Q140 mice striatal BDNF-vesicles rather showed the moderate increase in the fraction of either the full-fusion or partial-fusion mode compared to WT (the fraction of the partial-fusion mode: WT BDNF-vesicles=0.23±0.07 vs. Q140 BDNF-vesicles=0.36±0.02; the fraction of the full-fusion mode: WT BDNF-vesicles=0.49±0.06 vs. Q140 BDNF-vesicles=0.59±0.02), and this was likely due to the significant decrease in the no-fusion mode compared to WT ones (Fig. 1F; the fraction of the no-fusion mode, WT, 0.29±0.07; Q140, 0.05±0.01; p<0.05, unpaired t-test). These results indicate a differential alteration of cortical and striatal mechanisms for activity-dependent BDNF in the HD brain and support the idea that cortical neurons express molecular machinery that is either essential for BDNF secretion or sensitive to the mhtt.
There is evidence for early changes in the striatum in HD model mice and human patients before the onset of motor symptoms [19,25,26,27]. It is thus possible that cellular components required for activity-dependent BDNF secretion are already altered before striatal degeneration is observed. To test this idea, I next compared BDNF secretion from cortical axons of 10 week old (presymptomatic age) with that of 1 year old (symptomatic age)
Consistent with the finding of impaired BDNF-pH secretion from cortical axons in Q140 (Fig. 1), a replicated BDNF secretion assay still showed a significant reduction of HFS-induced BDNF secretion from cortical axons of symptomatic Q140 mice (Fig. 2A and B; averaged ΔFt/F0 over 100~200 sec duration after stimulation, symptomatic age: WT, −0.10±0.04; Q140, −0.00±0.02; p<0.05, unpaired t-test). I also found that Q140 mice at the presymptomatic age showed a decrease in cortical BDNF secretion in response to HFS when compared to wild-type mice (Fig. 2A and B; averaged ΔFt/F0 over 100~200 sec duration after stimulation, presymptomatic age: WT, −0.43±0.05; Q140, −0.23±0.04; p<0.001, unpaired t-test).
However, a population analysis of BDNF-vesicles in each condition revealed substantial changes in BDNF-vesicle fusion modes with the progress of the disease. At the presymptomatic age, reduced activity-dependent BDNF secretion was not correlated with any change in fusion modes of BDNF-vesicles (Fig. 2C), suggesting that molecular mechanisms involved in activity-dependent cortical BDNF secretion are relatively intact before the HD onset. By contrast, impaired BDNF secretion at the symptomatic age was accompanied by both a significant increase in the partial-fusion mode (fraction: WT, 0.27±0.06; Q140, 0.49±0.04; p<0.01, unpaired t-test) and decrease in the full-fusion mode (fraction: WT=0.65±0.06 vs. Q140=0.48±0.04; p<0.05, unpaired t-test), indicating reduced activity-dependent full-fusion of BDNF-containing vesicles. These results suggest that mhtts not only affect BDNF transcription and transport but also inhibit BDNF secretion by disrupting mechanisms for activity-dependent exocytosis of BDNF-containing secretory granules.
My results demonstrated an association of the mhtt with disrupted activity-dependent BDNF secretion from cortical axons. I next examine whether the mhtt is directly responsible for impaired cortical BDNF secretion in the HD brain, by reducing the htt level with a htt-targeted antisense nucleotide (ASO), which was shown previously to down-regulate mhtt expression [21,22]. Another knock-in HD line with ~180 CAG repeats in the htt gene (Q175 mice;
I found that changes in fluorescence intensity of BDNF-EGFP in response to electrical stimulation in all five ASO-injected symptomatic Q175 mice were similar with those observed from PBS-injected Q175 mice (Fig. 3), indicating that ASO-mediated htt knock-down unable to prevent the impairment of BDNF secretion. However, ASO injection to the presymptomatic Q175 (8-week old) mice was effective in preventing defective activity-dependent BDNF secretion. I found that overall cortical BDNF secretion was significantly recovered in three of five presymptomatic Q175 mice, which were injected with the same amount of htt ASO with that injected in the symptomatic Q175 mice (Fig. 3; ***p<0.001). These results indicate that a mhtt reduction during the early stage of HD could be beneficial for restoring BDNF level in the striatum through recovering activity-dependent BDNF secretion, although same application was not effective for reversing impaired BDNF secretion in the symptomatic HD brain.
In this study, I show that cortical activity-dependent BDNF secretion is abnormal, but no alteration of striatal BDNF secretion was observed in the brain with HD (Fig. 1E). These results suggest that cortical activity-dependent BDNF secretion is more vulnerable to the mhtt than striatal BDNF secretion from MSNs. Since striatal MSN do not express a significant level of BDNF [14,15], insensitivity of activity-dependent BDNF secretion to the mhtt might result from the lack of molecular machinery essential for BDNF secretion. Overall activity-induced cortical BDNF-pH secretion in WT striatal slices was significantly higher than striatal one by same stimulation (Fig. 1E; p<0.001, unpaired t-test), but this is probably because cortical axons directly received electrical stimulation, whereas activity in striatal neurites was evoked by electrical stimulation-induced synaptic transmission. Since presynaptic glutamate release at corticostriatal synapses was shown to be already reduced at this HD stage [26,27], the reduced no-fusion mode of striatal BDNF-vesicles in the HD brain might be reminiscent of the increased postsynaptic receptor sensitivity as a compensatory mechanism caused by decreased presynaptic inputs, and this alteration may affect activity-dependent BDNF secretion from striatal neurons of HD brains.
The exact molecular mechanisms of BDNF secretion affected by mhtts are still unclear. Several studies provide molecular evidence for mhtt-dependent alteration of vesicular exocytosis. Overexpression of mutant huntingtin in PC12 cells depleted vesicular machinery such as complexin II, resulting in impaired Ca2+-triggered vesicular exocytosis . Moreover, impaired astrocytic BDNF release in Q140 knock-in mice is reported to be caused by the abnormal function of Rab3a, a small GTPase localized on membranes of dense-core vesicles . Since an ATP release from astrocytic dense-core vesicles was also significantly reduced in Q140 mice , it is possible that mhtts disrupt general Ca2+-dependent exocytosis of dense-core vesicles. Despite normal fusion modes of BDNF-vesicles in the presymptomatic HD brain (Fig. 2C), I cannot exclude a possibility that a mhtt reduction reverses a deficient BDNF transport in the presymptomatic HD brain that was found to be present during early developmental stages [6,7], still capable of increasing the chance of BDNF secretion in the striatum.
My study did not directly show whether the ASO treatment restored impaired secretion of a pro-form (proBDNF) or mature form of BDNF (mature BDNF). A line of evidence suggests that the mature BDNF is a major form of BDNF secreted from cortical axons in the striatum. A previous study demonstrated that forebrain overexpression of BDNF in the HD model mice increases striatal TrkB signaling and cell survivals , both of which are mediated by mature BDNFs. Moreover, induction of corticostriatal long-term potentiation (LTP) was dependent on mature BDNF secretion from cortical presynaptic axons [30,31] as shown at hippocampal synapses, where LTP and long-term depression (LTD) were selectively regulated by mature and proBDNF signaling, respectively [16,30,31,32,33]. Thus, I suggest that restoration of BDNF signaling by the ASO treatment is mostly mediated by mature BDNFs.
Together, not only providing additional evidence for a mhtt-mediated disruption of BDNF secretion, the present study also points to a combination of the mhtt reduction and BDNF overexpression for prevention of HD after early detection of the disease (Fig. 4). In addition to a neuronal delivery of BDNF, I expect that glial BDNF overexpression or implantation of genetically modified cells with BDNF expression would also be useful [10,11,13] in promoting striatal BDNF level and reducing HD symptoms, with the efficacy greatly enhanced if accompanied with the mhtt reduction before HD symptoms occur.