Azadiradione ameliorates polyglutamine expansion disease in Drosophila by potentiating DNA binding activity of heat shock factor 1

Abstract

Aggregation of proteins with the expansion of polyglutamine tracts in the brain underlies progressive genetic neurodegenerative diseases (NDs) like Huntington's disease and spinocerebellar ataxias (SCA). An insensitive cellular proteotoxic stress response to non-native protein oligomers is common in such conditions. Indeed, upregulation of heat shock factor 1 (HSF1) function and its target protein chaperone expression has shown promising results in animal models of NDs. Using an HSF1 sensitive cell based reporter screening, we have isolated azadiradione (AZD) from the methanolic extract of seeds of Azadirachta indica, a plant known for its multifarious medicinal properties. We show that AZD ameliorates toxicity due to protein aggregation in cell and fly models of polyglutamine expansion diseases to a great extent. All these effects are correlated with activation of HSF1 function and expression of its target protein chaperone genes. Notably, HSF1 activation by AZD is independent of cellular HSP90 or proteasome function. Furthermore, we show that AZD directly interacts with purified human HSF1 with high specificity, and facilitates binding of HSF1 to its recognition sequence with higher affinity. These unique findings qualify AZD as an ideal lead molecule for consideration for drug development against NDs that affect millions worldwide.

Keywords: Gerotarget; HSF1; azadiradione; heat shock factor 1; neurodegenerative diseases; small molecule.

Figure 1. Azadiradione (AZD) activates HSF1 and its target heat shock chaperone genes

A. Flowchart of stepwise purification of AZD accompanied by an increase in specific activity measured by GFP expression (next to phase contrast images of cells) and luciferase assays as indicated. H, n-hexane; values are mean±SEM, n ≥ 3. B. Chemical structure of AZD. C. AZD induces DNA-binding ability of HSF1 in cells determined by EMSA using γ-³²pATP-labeled HSE (~1 ng/reaction) as described in the materials and methods. Whole cell extract (WCE) (50 μg) prepared from HEK293 cells expressing FLAG-HSF1 pre-treated with AZD or DMSO or HS (at 42°C/1 h) were used as indicated. The reactions were resolved on a 4% acrylamide and bis-acrylamide (29:1) gel and autoradiographed. D. AZD treatment induces HSF1 hyperphosphorylation like HS in WCE used in (C) which was erased upon phosphatase treatment determined by immunoblot with α-FLAG antibody. E. AZD treatment induces HSP70 and HSP27 transcription by HS in HEK293 cells that is sensitive to shRNA-mediated HSF1 downregulation (shHSF1) as determined by RT-PCR assay. Representative agarose gels of PCR products obtained using equivalent amounts of cDNAs as templates prepared from transcripts isolated from cells treated as indicated. F. Densitometric quantitation of HSP70 and HSP27 bands in agarose gels as shown in panel E from three independent experiments. **p < 0.01;***p < 0.001. G. AZD activates HSP70 only in MEF cells carrying wild type HSF1 MEF^WT gene but not in MEF cells that have been deleted both copies of HSF1 gene (MEF^HSF1−/−). Expression of HSP70 mRNA was measured by semiquantitative RT-PCR assay. The PCR products were resolved in agarose gels visualized by ethidium bromide staining.

Figure 2. AZD reduces protein aggregation and associated toxicity in the cell

A. Effect on ataxin130Q-GFP aggregation by AZD, or celastrol, (Cel) or DMSO as indicated in Neuro-2a cells. Scale bar 10 μm. B. The aggregates were counted and plotted after 48 h treatment, 450 cells were counted for each treatment, ***p < 0.001. C. Ethidium bromide stained agarose gel showing relative levels of HSP70 (mHSP70) transcripts isolated from cells subjected to identical treatments in duplicate as represented in panel (A) estimated by semiquantitative RT-PCR assays. D. Comparison of sensitivity of Neuro-2a cells to AZD or celastrol after 24 h treatment estimated by MTT assay. Data are mean±SEM, n ≥ 3

Figure 3. AZD feeding ameliorates 127Q (polyQ) induced eye defect and vision in the fruit fly

A. Dietary supplementation of AZD improved eye morphology and structure of ommatidial arrays (a-c' vs. b-d') damaged by polyQ expression. B. AZD supplementation significantly improved the vision defect caused by polyQ expression in the fly eyes as determined by phototaxis assay. Shown are the proportion (% on Y-axis; N = 75 flies in each case) of wild type and polyQ expressing flies (as indicated on the top) moving to the illuminated chamber on different days (X-axis; ** and *** indicate P < 0.01 and < 0.001), respectively, for comparison between DMSO and AZD treated polyQ flies. C. Dietary supplementation of AZD induces levels of HSP70 transcripts (Y-axis) in the head of the flies not expressing or expressing polyQ (X-axis). Data are mean±SEM, n ≥ 3. D. Compared to DMSO (upper row) dietary supplementation of AZD induces expression of HSP70 and reduces the level of polyQ in retinal cells as detected by immunostaining as indicated. Individual or merged staining in the confocal projection images are indicated on top of each column. Scale bar 20 micrometer.

Figure 4. AZD induces HSP70 protein expression without interfering with the functions of HSP90 or proteasome

WCE (20 μg) of HEK293 cells pretreated for 16 h with bortezomib (BTZ) or geldanamycin (Gel) or various concentrations of AZD, or DMSO were subjected to immunoblot A. using the antibodies against anti-ubiquitin (P-Ub). The levels of HSP70 protein in the same samples were also determined as indicated using anti-HSP70 antibody. B. The bands in the blot were estimated and plotted as bar graph considering ß-actin levels as an internal loading control. C. Immunoblot with antibodies against two HSP90 client proteins Akt or Raf1 and HSP70 and HSP90 itself to compare their relative expression levels in those samples. D. Band intensities in the blot were estimated by densitometric scanning and plotted to compare their expression levels. ß-actin levels were used as internal loading control.

Figure 5. AZD directly interacts with the purified HSF1 protein tested by EMSA as well as fluorometric assay

A. Autoradiogram of an EMSA showing concentration-dependent enhancement of radiolabeled HSE binding of HSF1 (0.5 μM) by AZD. The binding is competed out by excess unlabeled HSE (SpCold) but not by an unrelated double-stranded (ds) DNA oligonucleotide (NSpCold) as described in the method. B. Effect of AZD (5 μM) on the binding of increasing concentration HSE by HSF1 (2 μM) measured by fluorometric assay as indicated. C. Bar diagram representing the Kd values of HSF1 binding to HSE without or with AZD. D. Concentration dependent effect of AZD as indicated on HSF1 (2 μM) fluorescence (inset: Scatchard plot of AZD binding to HSF1) and E. on HSP90 (2 μM) fluorescence. F. Effect of geldanamycin (Gel) on the indicated concentration on HSP90 fluorescence. G. Concentration dependent effect of AZD (as indicated) or heat shock (HS) on HSF1 (2 μM) multimerization in the cell-free system determined by monitoring the kinetics of polymerization by using light scattering at 350 nm. The presence of DTT in the reaction abolished the polymerization induced by AZD or HS. H. Cartoon proposing AZD targeting the free pool of monomeric HSF1, an intermediate between repressed (i) and active HSF1 (homotrimeric state) (ii) in the cell that translocates to the nucleus to execute its transcription function (iii). HSF1 is kept in inactive monomeric state by a repression complex composed of proteins including HSP90, HSP70, p23 and immunophilin (i); AZD does not disturb (indicated by a cross symbol) HSP90 chaperone function of refolding of unfolded/misfolded client proteins (iv), or proteasome that degrades irreversibly unfolded protein (v).