The Wnt receptor Ryk reduces neuronal and cell survival capacity by repressing FOXO activity during the early phases of mutant huntingtin pathogenicity

Abstract

The Wnt receptor Ryk is an evolutionary-conserved protein important during neuronal differentiation through several mechanisms, including γ-secretase cleavage and nuclear translocation of its intracellular domain (Ryk-ICD). Although the Wnt pathway may be neuroprotective, the role of Ryk in neurodegenerative disease remains unknown. We found that Ryk is up-regulated in neurons expressing mutant huntingtin (HTT) in several models of Huntington's disease (HD). Further investigation in Caenorhabditis elegans and mouse striatal cell models of HD provided a model in which the early-stage increase of Ryk promotes neuronal dysfunction by repressing the neuroprotective activity of the longevity-promoting factor FOXO through a noncanonical mechanism that implicates the Ryk-ICD fragment and its binding to the FOXO co-factor β-catenin. The Ryk-ICD fragment suppressed neuroprotection by lin-18/Ryk loss-of-function in expanded-polyQ nematodes, repressed FOXO transcriptional activity, and abolished β-catenin protection of mutant htt striatal cells against cell death vulnerability. Additionally, Ryk-ICD was increased in the nucleus of mutant htt cells, and reducing γ-secretase PS1 levels compensated for the cytotoxicity of full-length Ryk in these cells. These findings reveal that the Ryk-ICD pathway may impair FOXO protective activity in mutant polyglutamine neurons, suggesting that neurons are unable to efficiently maintain function and resist disease from the earliest phases of the pathogenic process in HD.

Conflict of interest statement

The authors have declared that no competing interests exist. Dr. Robert Ferrante is listed as an author of our paper, but at the time of acceptance was not reachable or able to confirm details of his author contributions to the manuscript. The corresponding author, Christian Neri, has therefore supplied the information regarding his contribution to the manuscript and his competing interests and it is correct to the best of Christian Neri's knowledge.

Figures

Figure 1. Analysis of Ryk in polyQ nematodes and striatal cells derived from HdhQ111 mice.

(A) Modulation of touch response of polyQ nematodes by LOF of lin-18/RYK. Shown are the results in C. elegans transgenics expressing expanded (128Q) or normal (19Q) exon-1–like htt transgenes in touch receptor neurons. The 128Q-mediated loss of touch response is ameliorated by LOF of lin-18/RYK. No effects are detected in 19Q nematodes. Data are mean ± SEM with more than 200 animals tested per genotype. *p<0.001 versus 128Q controls. LOF of lin-18/RYK does not modify 128Q transgene expression levels as tested by qRT-PCR (right panel, data are mean ± SEM with n = 5). Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. (B) Mutant htt (109Q/109Q) striatal cells have increased Ryk levels. Data for qRT-PCR are mean ± SEM (n = 7). **p<0.01 versus normal htt (7Q/7Q) cells. Data for Western blotting are mean ± SD (n = 3). **p<0.01 versus normal htt (7Q/7Q) cells. Significance was tested using paired t tests. (C) Reducing Ryk levels decreased the mortality of 109Q/109Q striatal cells induced by serum deprivation compared to scramble situation, with no effect detected in 7Q/7Q cells. Western blotting was used to test expression levels as 109Q/109Q cells do not display HTT aggregation. Mutant htt levels were unchanged by knockdown of Ryk. Data are mean ± SD (n = 4). *p<0.01 versus scramble. The RNA sequences shown are indicated in the Materials and Methods section. Significance was tested using paired t tests.

Figure 2. Neuroprotection by lin-18/Ryk LOF is a cell-autonomous process that requires daf-16/FoxO activity.

(A) Expression of lin-18 cDNA in touch receptor neurons using the mec-3 promoter abolishes the neuroprotective activity of lin-18 LOF in 128Q nematodes with no effect detected in 19Q nematodes as tested in two independent extrachromosomal arrays (A1, A2) per polyQ genotype. The expression of lin-18 cDNA was confirmed by RT-PCR followed with enzymatic restriction for all of the arrays generated (unpublished data). Data are mean ± SEM (more than 200 animals tested). *p<0.001 compared to 128Q transgenics; **p<0.001 compared to 128Q;lin-18 nematodes. (B) Neuron dysfunction is aggravated by bar-1/β-catenin or daf-16/FoxO LOF in 128Q nematodes, and protection from 128Q toxicity by lin-18 LOF is reduced by LOF of bar-1 and suppressed by LOF of daf-16. Data are mean ± SEM (more than 200 animals tested). **p<0.001 compared to 128Q transgenics; ***p<0.001 compared to 128Q;lin-18 nematodes. (C) lin-18, bar-1 and daf-16 LOF alone or in combination do not change transgenic protein expression levels in 128Q nematodes. Data are mean ± SD (n = 3). Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test.

Figure 3. Ryk ICD binds to β-catenin.

(A) β-catenin binds to Ryk. Constructs of Myc-tagged Ryk or uncleavable Ryk (Ryk:EGFR Rc) were transfected into 293T cells. Ryk proteins were immunoprecipitated with anti-Myc antibody, and beta-catenin associated with Ryk was determined by immunoblotting. Ryk ICD can be detected in the cells expressing wild-type Ryk. (B) The ICD of Ryk binds to β-catenin. Cells expressing Flag-tagged Ryk-ICD were used for anti-Flag immunoprecipitation. The ICD and associated β-catenin were determined by Western blot. NE, nuclear extract; FUIGW-FLAG, vector FUIGW plus a Flag sequence.

Figure 4. The Ryk ICD represses the transcriptional activity of FOXO3a, a protein that protects from mutant HTT.

(A) Foxo3a siRNA treatment enhances the mortality of mutant htt striatal cells subjected to serum deprivation, whereas FOXO3a overexpression (O/E) has the opposite effect. Data are mean ± SD (n = 4). *p<0.001 compared to scramble; **p<0.001 compared to empty vector control. (B) Representative Western blots showing decreased (si-Foxo3a) or increased (FOXO3a O/E) FOXO3a levels and no change in HTT protein levels. (C) FOXO transcriptional activity was measured in normal htt mouse striatal cells. Cells were cultured in normal conditions and co-transfected with a construct encoding FOXO3a together with the reporter FHRE-luciferase, which contains three canonical FOXO binding sites, and an internal Renilla luciferase reporter construct. Luciferase and Renilla luciferase activities were measured and the ratio (luciferase/Renilla luciferase)10,000 calculated. Data are mean ± SD of four independent experiments performed in triplicate. Treatment with β-catenin siRNA, full-length Ryk cDNA, and Ryk-ICD cDNA reduces luciferase activity to similar levels, whereas treatment with uncleavable Ryk showed no effect. *p<0.001 compared to FHRE-luc, **p<0.001 compared to scramble RNA and ***p<0.001 compared to FOXO3a O/E. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. (D) Representative Western blots showing increased levels of FOXO3a and decreased levels of β-catenin, and expression of Myc-tagged Ryk, Myc-tagged Ryk-ICD, and Myc-tagged γ-secretase–uncleavable Ryk (all proteins with a Myc tag at the C-terminal end). The Myc-tagged Ryk and γ-secretase–uncleavable Ryk proteins were detected as two fragments, one corresponding to the full-length Ryk precursor (Ryk) and one corresponding to a Ryk CTF (Ryk CTF) resulting from proteolytic cleavage in the extracellular domain near the transmembrane domain. The full-length Ryk precursor is less abundant for wild-type Ryk expression compared to mutant Ryk expression (see Results for the discussion of Ryk expression profiles).

Figure 5. The Ryk ICD is cytotoxic in C. elegans neurons and mouse striatal cells expressing expanded polyQs.

(A) Expression of LIN-18 ICD cDNA (4 ng/µl) in touch receptor neurons using the mec-3 promoter is sufficient to abolish the neuroprotective activity of lin-18 LOF in 128Q nematodes with no effect detected in 19Q nematodes as tested in two independent extrachromosomal arrays (128Q: A1 is ID1333, A2 is ID1334; 19Q: A1 is ID1331, A2 is ID1332; see also Table S8) per polyQ genotype. The expression of LIN-18 ICD cDNA was confirmed by RT-PCR for all of the arrays generated. Data are mean ± SEM (more than 200 animals tested). **p<0.001 compared to 128Q animals, **p<0.001 compared to 128Q;lin-18 animals. ns, not significant. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. (B) Expression of LIN-18 ICD cDNA (4 ng/µl) in touch receptor neurons using the mec-3 promoter is also sufficient to abolish the protective activity of lin-18 LOF on axonal swelling in the PLM neurons of 128Q nematodes as tested in two independent extrachromosomal arrays (Lin-18 ICD: A1 is ID1333, A2 is ID1334; Lin-18: A1 is ID1325, A2 is ID1326). The expression of LIN-18 ICD cDNA was confirmed by RT-PCR for all of the arrays generated. Expression of empty vector (4 ng/µl) showed no effect as tested in two independent extrachromosomal arrays (A1 is ID1339, A2 is ID1340). Data are mean ± SEM (more than 200 animals tested). *p<0.001 compared to 128Q animals, **p<0.001 compared to 128Q;lin-18 animals. ns, not significant. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. The lower panel shows a representative image of axonal swelling in the anterior process of posterior touch receptor neurons of 128Q nematodes co-expressing HTT1-57::CFP and YFP . Swelling (white arrows, YFP signals are pseudocolored in green) and HTT::CFP aggregation (yellow arrows, CFP signals are pseudocolored in red) are shown. Magnification is 100× and scale bar is 5 µM. (C) Overexpressing either V5-tagged β-catenin or Myc-tagged Ryk-ICD or both has no effect on the mortality induced by serum deprivation in normal htt striatal cells. Overexpressing β-catenin reduces the mortality induced by serum deprivation in mutant htt striatal cells, whereas overexpressing the Ryk-ICD aggravates cell mortality. Co-expressing Ryk-ICD and β-catenin resulted in cell mortality levels that are similar to those induced by empty vector overexpression. Data are mean ± SD (n = 4). *p<0.01 and **p<0.05 compared to empty vector. ns, not significant. Significance was tested using paired t tests. (D) Representative Western blot showing increased V5-tagged β-catenin and Myc-tagged Ryk-ICD levels after transfection of 7Q/7Q and 109Q/109Q striatal cells.

Figure 6. Reducing presenilin 1 expression counteracts the cytotoxicity of full-length Ryk overexpression in mutant htt striatal cells.

Assays were performed using caspase 3/7 activity in cells subjected to serum deprivation. (A) PS1 and PS2 siRNA treatment enhances the viability of mutant htt striatal cells. Data are mean ± SD (n = 3), *p<0.01 compared to scramble RNA. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. (B) Representative Western blots showing decreased levels of PS1/PS2. (C) Knockdown of PS1 reduces the cytotoxic effects of overexpressing full-length Ryk in mutant htt striatal cells, with no effect detected on the cytotoxic effects of overexpressing Ryk-ICD. Data are mean ± SD (n = 4), *p<0.05 and **p<0.01 compared to scramble RNA. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. EV, empty vector; ns, not significant. (D) Representative Western blots showing decreased levels of PS1 and expression of Myc-tagged Ryk species and Myc-tagged Ryk-ICD. *Nonspecific signal.

Figure 7. The ICD of Ryk is increased in the nucleus of mutant htt striatal cells.

(A) Representative confocal microscopy images showing the pattern of Ryk-ICD immunoreactivity in normal htt (7Q/7Q) and mutant htt (109Q/109Q) striatal cells under normal culture conditions (no serum starvation) as detected using the rabbit polyclonal antibody anti-RykICD. (B) Quantification of Ryk-ICD immunoreactivity in mouse striatal cells. 7Q/7Q and 109Q/109Q cells were grown on the same slides. Comparisons were performed for cells with a nucleus size in the range of 150–250 pixels. Ryk-ICD immunoreactivity was increased in the nucleus of 109Q/109Q cells compared to 7Q/7Q cells. Data are mean ± SEM for the ratio Intensity/Area as detected in either nucleus or cytoplasm (n = 3 for a total of at least 100 cells analyzed), ***p<0.0001 compared to normal htt cells; n.s., not significant. Significance was tested using one-way ANOVA, with correction for multiple testing by Tukey's Multiple Comparison Test. Ryk siRNA treatments were observed to reduce nuclear Ryk-ICD immunoreactivity in 109Q/109Q cells (see Figure S6). (C) Representative confocal microscopy images showing the pattern of Ryk-ICD immunoreactivity in normal htt (7Q/7Q) and mutant htt (109Q/109Q) striatal cells under normal culture conditions (no serum starvation) as detected using the rabbit polyclonal antibody anti-RykICD and mouse Pol2 antibody 7C2. (D) Quantification of Ryk-ICD and Pol2 immunoreactivity in mouse striatal cells. 7Q/7Q and 109Q/109Q cells were grown on the same slides. Comparisons were performed for cells with a nucleus size in the range of 150–250 pixels. The ratio for Ryk-ICD/Pol2 immunoreactivity was increased in the nucleus of 109Q/109Q cells compared to 7Q/7Q cells. Data are mean ± SEM for the ratio Intensity Ryk-ICD/Intensity Pol2 as detected in the nucleus (n = 4 for a total of at least 100 cells analyzed), **p<0.002 compared to normal htt cells. Significance was tested using Welch's t test.