Excess Rab4 rescues synaptic and behavioral dysfunction caused by defective HTT-Rab4 axonal transport in Huntington’s disease

Huntington’s disease (HD) is characterized by protein inclusions and loss of striatal neurons which result from expanded CAG repeats in the poly-glutamine (polyQ) region of the huntingtin (HTT) gene. Both polyQ expansion and loss of HTT have been shown to cause axonal transport defects. While studies show that HTT is important for vesicular transport within axons, the cargo that HTT transports to/from synapses remain elusive. Here, we show that HTT is present with a class of Rab4-containing vesicles within axons in vivo. Reduction of HTT perturbs the bi-directional motility of Rab4, causing axonal and synaptic accumulations. In-vivo dual-color imaging reveal that HTT and Rab4 move together on a unique putative vesicle that may also contain synaptotagmin, synaptobrevin, and Rab11. The moving HTT-Rab4 vesicle uses kinesin-1 and dynein motors for its bi-directional movement within axons, as well as the accessory protein HIP1 (HTT-interacting protein 1). Pathogenic HTT disrupts the motility of HTT-Rab4 and results in larval locomotion defects, aberrant synaptic morphology, and decreased lifespan, which are rescued by excess Rab4. Consistent with these observations, Rab4 motility is perturbed in iNeurons derived from human Huntington’s Disease (HD) patients, likely due to disrupted associations between the polyQ-HTT-Rab4 vesicle complex, accessory proteins, and molecular motors. Together, our observations suggest the existence of a putative moving HTT-Rab4 vesicle, and that the axonal motility of this vesicle is disrupted in HD causing synaptic and behavioral dysfunction. These data highlight Rab4 as a potential novel therapeutic target that could be explored for early intervention prior to neuronal loss and behavioral defects observed in HD.


Figure S6. Rab4 localizes with polyQ and HTT accumulations in HD iNeurons differentiated from HD patient iPSCs. (a)
Representative image of an axonal blockage (arrow) containing HTT and Rab4 is seen in a neurite differentiated from HD-patient iPSC with 109Q repeats, using an antibody against HTT and Rab4. Intensity plots show overlapping peak (purple arrow). X=distance (µm), Y=intensity (AU). Bar=5µm. (b) Representative images from normal (WT) neurons derived from a normal individual (25Q) and HD iNeurons derived from a HD patient (72Q) stained with antibodies to Rab4 (green arrows), polyQ (red arrows) and Hoechst (blue). Co-localization of Rab4 and polyQ is seen in WT neurons (yellow arrows). Not all Rab4 puncta (green arrows) or polyQ puncta (red arrows) co-localize. Accumulations of Rab4 (green arrows) and polyQ (red arrows) are seen at the proximal end of the HD iNeurons. Only a few puncta show co-localization (yellow arrows). Bar=5µm. Intensity plots show overlapping peaks in WT iNeurons (purple arrows). Plots from HD iNeurons show red and green peaks with fewer overlapping peaks (purple). X=distance (µm), Y=intensity (AU). (c,d) Representative high-resolution images from WT iNeurons stained for Rab4 and DIC (c), Rab4 and HIP1 (d) show Rab4 co-localizes with DIC or HIP1 (yellow arrows). The HD iNeurons show Rab4 only (red arrows) or DIC or HIP1 only (green arrows) puncta (Bar = 5µm). Intensity plots show Rab4 and DIC or Rab4 and HIP1 overlapping puncta (purple arrows) in WT iNeurons, while in HD iNeurons only Rab4 (red) and DIC or HIP1 (green) puncta are seen. X=distance (µm), Y=intensity (AU). Figure S7. The putative HTT-Rab4 synaptic complex is disrupted in HD iNeurons differentiated from HD patient iPSCs. (a) Analysis of total proteins from WT and HD iNeuronal culture homogenates using antibodies to HTT, Rab4, HIP1, DIC, KIF5C, Rab11-FIP5 and actin (loading control). Both full length (FL-HTT) and HTT fragments are observed in WT and HD iNeuronal cultures. Molecular weight ladder is indicated in Kd. Quantifications of band intensity on western blots (AU) revealed that DIC (p<0.005) and HIP1 (p<0.005) levels are significantly decreased in HD compared to WT. The level of KIF5C is also decreased in HD compared to WT although this was not significant (ns). The level of Rab4 and HTT show an increased trend in HD compared to WT. A trend towards increased levels is also seen for Rab11-FIP5 in HD compared to WT. Y axis=normalized intensity to WT (AU). Protein levels were normalized to actin and then to WT. n=3. ns=p>0.05, *p<0.05, **p<0.005. Statistical significance was determined using the two-sample two-sided Student's t-test. Figure S8. Expression of pathogenic HTT in Drosophila larval axons cause Rab4 blockages within larval axons and elevated Rab4 levels at NMJs. (a) Larvae expressing HTT103Q-eGFP and Rab4-mRFP show axonal blockages that contain both HTT and Rab4 (arrows). Bar= 5µm. (b) Expression of Rab4-mRFP with HTT16Q or HTT128Q does not affect the average number of axonal blocks (ns). n=8. (c) Representative images of NMJs from muscle 6/7 segment A4-5 in larvae expressing Rab4-mRFP alone or co-expressing Rab4-mRPF with either HTT16Q or HTT128Q stained with HRP-FITC. Note that larvae expressing Rab4-mRFP alone show homogenous Rab4-mRFP within all boutons, which is seemingly lost with HTT128Q. Bar=5µm. Quantification analysis of HRP intensity (AU) in larval NMJs show no changes (ns) in HRP intensity across all genotypes; however, quantification of Rab4-mRFP intensity (AU) revealed that larvae co-expressing Rab4-mRFP with either HTT128Q showed significantly increased Rab4-mRFP intensity compared to larvae expressing Rab4-mRFP alone (p<0.0001) or to larvae co-expressing Rab4-mRFP with HTT16Q (p<0.001). When Rab4-mRFP intensity at NMJs was normalized to HRP intensity (AU), quantification revealed a similar significant increase in larvae co-expressing Rab4-mRFP with HTT128Q compared to larvae expressing Rab4-mRFP alone (p<0.001) or to larvae co-expressing Rab4-mRFP with HTT16Q (p<0.001). Note that larvae expressing Rab4-mRFP alone and larvae co-expressing Rab4-mRPF with HTT16Q show unchanged intensities at NMJs (ns). AU=arbitrary units. n=8. ns=p>0.01, *p<0.01, **p<0.001, ***p<0.0001. Statistical significance was determined using the two-sample two-sided Student's t-test.

Figure S9. A model of the putative moving HTT-Rab4 vesicle and how it is disrupted in HD.
(a) Our observations propose a working model where a specific HTT-Rab4 vesicle complex that contains synaptic SNARE proteins synaptotagmin and synaptobrevin, and Rab11 is attached to molecular motors kinesin and dynein via HIP1 and is transported down axons for synaptic function. (b) In HD, polyQ-HTT present with this HTT-Rab4 vesicle disrupts associations between accessory proteins by decreasing the binding affinity of HIP1. Together, these aberrant associations dislocate the link between the polyQ-HTT-Rab4 vesicle and motor proteins kinesin-1 and dynein, resulting in the perturbation of its motility within axons causing synaptic and locomotion defects.

Movie S1
A representative movie from a larva expressing HTT25Q-eGFP (green) and . Note that HTT and Rab4 co-migrate (yellow puncta) within larval axons.

Movie S11
A representative movie from a normal (WT) human iNeuron transfected with mCherry-Rab4 shows that Rab4 moves bi-directionally within the human neurite.

Movie S12
A representative movie from a HD patient iNeuron (Q72) transfected with mCherry-Rab4 shows that the motility of Rab4 is decreased.