Aqueous Extract of Paeonia lactiflora and Paeoniflorin as Aggregation Reducers Targeting Chaperones in Cell Models of Spinocerebellar Ataxia 3

Spinocerebellar ataxia (SCA) types 1, 2, 3, 6, 7, and 17 as well as Huntington's disease are a group of neurodegenerative disorders caused by expanded CAG repeats encoding a long polyglutamine (polyQ) tract in the respective proteins. Evidence has shown that the accumulation of intranuclear and cytoplasmic misfolded polyQ proteins leads to apoptosis and cell death. Thus suppression of aggregate formation is expected to inhibit a wide range of downstream pathogenic events in polyQ diseases. In this study, we established a high-throughput aggregation screening system using 293 ATXN3/Q75-GFP cells and applied this system to test the aqueous extract of Paeonia lactiflora (P. lactiflora) and its constituents. We found that the aggregation can be significantly prohibited by P. lactiflora and its active compound paeoniflorin. Meanwhile, P. lactiflora and paeoniflorin upregulated HSF1 and HSP70 chaperones in the same cell models. Both of them further reduced the aggregation in neuronal differentiated SH-SY5Y ATXN3/Q75-GFP cells. Our results demonstrate how P. lactiflora and paeoniflorin are likely to work on polyQ-aggregation reduction and provide insight into the possible working mechanism of P. lactiflora in SCA3. We anticipate our paper to be a starting point for screening more potential herbs for the treatment of SCA3 and other polyQ diseases.


Introduction
Spinocerebellar ataxias (SCAs) are a large, complex group of heterogeneous autosomal dominant neurodegenerative disorders characterized by cerebellar dysfunction alone or in combination with other neurological abnormalities [1]. Among them, the expansions of CAG trinucleotide repeats encoding a polyglutamine (polyQ) stretch have been shown to cause dominantly inherited SCA1, SCA2, SCA3, SCA6, SCA7, SCA17, and dentatorubropallidoluysianatrophy (DRPLA) [2][3][4][5][6][7][8]. These polyQ-mediated genetic disorders in SCAs have shown selective progressive degeneration of the cerebellum, brainstem, and spinal tract, with prominent pathological hallmark of intranuclear and cytoplasmic accumulation of aggregated polyQ proteins inside degenerated neurons [9]. Different polyQ tractcontaining proteins ultimately lead to the dysfunction and degeneration of specific neuronal subpopulations [10]. The aggregated polyQ proteins may cause dysfunction of mitochondria, chaperone, and ubiquitin proteasome system, leading to apoptosis and cell death [11][12][13]. As misfolding of the polyQ protein is likely the initial event in the pathogenic cascade, suppression of protein misfolding is expected to inhibit a wide range of downstream detrimental events, and to rescue neuronal dysfunction.
Increasing evidence suggests that some herbs may potentially attenuate the deterioration of neurodegenerative diseases. Paeonia lactiflora (P. lactiflora), belonging to the 2 Evidence-Based Complementary and Alternative Medicine Paeoniaceae family, is a perennial herb frequently used as an important ingredient in many traditional prescriptions. It has been commonly used for nourishing blood, alleviating pain, reducing irritability, as well as treating liver disease and cancer [14]. Paeoniflorin, one of the main compounds extracted from P. lactiflora, has been reported to ameliorate neurodegenerative process in Parkinson's disease (PD) and Alzheimer's disease (AD) models [15][16][17]. However, the effect of P. lactiflora herb extract and paeoniflorin in treating SCA remains unraveled.
In the present study, we firstly built up an aggregation screening cell model by overexpressing CAG-expanded ATXN3, the causative mutation in SCA3 [2], in 293 cells, and then examined the anti-aggregation effect of P. lactiflora aqueous extract and paeoniflorin. We further demonstrated that the anti-aggregation activity of P. lactiflora extract and paeoniflorin was contributed by the enhancement of heat shock transcription factor 1 (HSF1)-heat shock protein (HSP) 70 chaperone system. These findings provide evidence that P. lactiflora and paeoniflorin may be a novel alternative therapeutic agent for the treatment of SCAs.

P. lactiflora Extract Preparation and HPLC Analysis.
Aqueous extract from P. lactiflora was provided by Sun-Ten Pharmaceutical Company (Taipei, Taiwan). Briefly, 100 g of dried P. lactiflora was boiled with 1500 mL of water at 100 ∘ C for 30 min and was sieved using a 100-mesh sieve. The extract was concentrated to 100 mL and filtered through a 200-mesh sieve. The extract was then dried by speed vacuum concentration and then stored at −20 ∘ C until used.

Isogenic 293 and SH-SY5Y Cell Lines.
Human 293derived Flp-In-293 cells (Invitrogen) were cultivated in DMEM containing 10% FBS as described. The cloned pcDNA5/FRT/TO-ATXN3/Q 14 and Q 75 plasmids were used to generate the isogenic ATXN3/Q 14∼75 cell lines by targeting insertion into Flp-In-293 cells, according to the supplier's instructions. The repeats in these ATXN3 cell lines were examined by PCR and sequencing. These cell lines were grown in medium containing 5 g/mL blasticidin and 100 g/mL hygromycin (InvivoGen). Human SH-SY5Yderived Flp-In host cell line was constructed as described [23]. The SH-SY5Y host cells were used to generate isogenic ATXN3/Q 14∼75 lines and maintained as described above.
2.6. ATXN3/Q 75 Aggregation Assay. 293ATXN3/Q 75 -GFP cells were plated into 96-well (2 × 10 4 /well) dishes, grown for 24 hr and treated with different concentrations of the P. lactiflora extract (2∼200 g/mL) or suberoylanilide hydroxamic acid (SAHA, Cayman Chemical), paeoniflorin (Sigma), gallic acid, and albiflorin (Chromadex) (100 nM∼5 M) for 8 hr. Then doxycycline (10 g/mL, BD) was added to the medium to induce ATXN3/Q 75 -GFP expression for 6 days. Oxaliplatin (5 M, Sigma) was also added for aggregate accumulation through inhibition of cell division [24]. Then cells were stained with Hochest 33342 (0.1 g/mL, Sigma) and aggregation percentage was assessed by HCA system, with excitation/emission wavelengths at 482/536 (GFP). SH-SY5Y ATXN3/Q 75 -GFP cells were seeded in 6-well (2×10 5 /well) plate, with all trans retinoic acid (10 M, Sigma) added at seeding time. At day 2, cells were treated with paeoniflorin (100 nM) or the P. lactiflora extract (10 g/mL) for 8 hr, and then doxycycline (5 g/mL) was added to induce ATXN3/Q 75 -GFP expression. The cells were kept in the medium containing 10 M trans retinoic acid, doxycycline and paeoniflorin/P. lactiflora extract for 7 days. After that,  Cytotoxicity of the aqueous extract of P. lactiflora, paeoniflorin, garlic acid, albiflorin, and SAHA against HEK-293 and SH-SY5Y cells using MTT viability assay. The IC 50 of each herb/compound was shown under the columns. To normalize, the relative viability in untreated cells is set as 100%. The red line represents 50% viability. cells were stained with Hochest 33342 (0.1 g/mL) and aggregation percentage was assessed as described.

Real-Time PCR.
Total RNA from 293 ATXN3 lines was extracted using Trizol reagent (Invitrogen). The RNA was DNase (Stratagene) treated, quantified, and reversetranscribed to cDNA as described. Real-time quantitative PCR experiments were performed in the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). Amplification was performed on 100 ng cDNA with genespecific TaqMan fluorogenic probes Hs0024525 ml for ATXN3 and 4326321E for HPRT1 (endogenous control) (Applied Biosystems). Fold change was calculated using the formula 2 ΔC , ΔC = C (control) − C (target), in which C indicates cycle threshold.

Statistical Analysis.
For each set of values, data were expressed as the means ± standard deviation (SD). Three independent experiments were performed and non-categorical variables were compared using the Student's -test. All values were two-tailed, with values of < 0.05 considered significant.

Construction of 293 Cells
Expressing ATXN3/Q 75 Aggregates. For therapies toward the polyQ diseases, we aimed to screen herbs/compounds potentially inhibiting polyQ aggregation. As removal of the N-terminus of polyQ-expanded ATXN3 is required for aggregation in vitro and in vivo [25], we cloned GFP-tagged ATXN3 C-terminal Q 14∼75 -containing fragment to establish Flp-In 293 cells with ATXN3/Q 14∼75 -GFP expression in an inducible fashion. As shown in Figure 1(a), the GFP antibody detected 40 kDa ATXN3/Q 14 -GFP and 57 kDa ATXN3/Q 75 -GFP proteins in doxycycline (Dox) induced ATXN3 cells. ATXN3-RNA levels were then examined by real-time PCR using ATXN3-specific probe and primers. As shown in Figure 1(b), in the presence of Dox, the two ATXN3 lines expressed ∼20 times more ATXN3 RNA than in the absence of Dox. While the expressed ATXN3/Q 14 was mainly diffused, the expressed ATXN3/Q 75 -GFP formed aggregates (Figure 1(c)).

Aqueous Extract of P. lactiflori and Constituents.
To examine the potential active compounds in P. lactiflori, the chemical profile of aqueous extract was analyzed and quantified by full-spectrum analytic HPLC. Chromatographic patterns showed peaks in 230 nm corresponding to the retention time compatible with paeoniflorin, garlic acid and albiflorin (Figure 2(a)). The amounts of paeoniflorin, garlic acid, and albiflorin in aqueous extract of P. lactiflori were 2.27%, 0.30%, and 0.73%, respectively, corresponding to 47.33 mM, 18.06 mM, and 15.16 mM, respectively, in 1 g/mL aqueous extract (Figure 2(b)).
MTT assays were performed with human embryonic kidney 293 and human neuroblastoma SH-SY5Y cells after treatment with extract of P. lactiflora and the its three constituents, respectively, for 24 hr. The histone deacetylase inhibitor suberoylanilide suberoylanilide hydroxamic acid (SAHA) known to reduce SDS-insoluble polyQ aggregate [26] was included for comparison. The IC 50 of the herb and compounds were calculated using the interpolation method. Both P. lactiflora extract and its constituents paeoniflorin and albiflorin had an IC 50 higher than the highest concentration tested ( >30 mg/mL for P. lactiflora and >1 mM for paeoniflorin and albiflorin), suggesting their very low cytotoxicity (Figure 2(c)).

P. lactiflori Extract and Paeoniflorin Reduce ATXN3/Q 75
Aggregation on 293 Cell Model. To screen if herb/compounds potentially inhibit aggregation, we used ATXN3/Q 75 -GFP cells to examine aqueous extract of P. lactiflora and its constituents for their potentials to reduce the ATXN3/Q 75 aggregation. The experiment flow chart is shown in Figure 3(a) and representative fluorescence microscopy images of aggregation after treatment with paeoniflorin and the aqueous extract of P. lactiflora are shown in Figure 3(b). As a positive control, HDAC inhibitor SAHA reduced the ATXN3/Q 75 aggregation to 85% (at 100 nM) as compared to untreated cells (Figure 3(c)). While garlic acid did not display good aggregation-inhibitory potential (90∼95% at 100 nM∼1 M), P. lactiflora (81∼82% at 2∼50 g/mL), paeoniflorin (73% at 100 nM) and albiflorin (78% at 5 M) had greater aggregation reduction potential than SAHA (Figure 3(c)). The IC 50 cytotoxicity/effective (reduced the ATXN3/Q 75 aggregation to 85% or lower) dose ratio of SAHA, paeoniflorin, albiflorin, and extract of P. lactiflora were 3800, >10000, >200, and >15000, respectively. Considering 2 g/ml of P. lactiflora extract contained 95 nM paeoniflorin and 30 nM albiflorin and tested greatest aggregation reduction potential of 100 nM for paeoniflorin and 5 M for albiflorin, paeoniflorin was regarded as a major active component for the aggregation inhibition in P. lactiflora.

Discussion
Although Chinese herbs have been reported to reduce pneumonia risk in elderly patients with dementia [27] and regarded as a potential treatment of Huntington's disease (HD) [28], the attempts to apply this alternative treatment in SCA are still few. Okabe et al. (2007) reported a patient with SCA6 was treated with a mixture of 18 medical herbs (modified Zhengan Xifeng Tang) and then the patient's ataxia was remarkably reduced [29]. However, the therapeutically effective compound(s) in this remedy remains unknown. In this study we identified the aqueous extract of P. lactiflora that reduced ATXN3-aggregates mainly via its active compound paeoniflorin (Figures 2 and 3). The reporter gene assay and Western blotting further indicated the aggregation-reduction effect of paeoniflorin was modulated by the up-regulation of HSF1 and its targets, HSPA8 and HSPA1A chaperone expressions ( Figure 4).
P. lactiflora, with immunomodulatory and anti-inflammatory effects [30], has been widely used as a component of traditional Chinese prescriptions to relieve pain and to treat rheumatoid arthritis, systemic lupus erythromatosus, dysmenorrhea, hepatitis, muscle spasm, and fever with a long history. Its main bioactive component, paeoniflorin, possesses wide pharmacological effects in the nervous system. It has been reported to decrease the death of rat cortical cells while exposed to H 2 O 2 -induced oxidative stress [31]. Subcutaneous injection of paeoniflorin has shown functional protection in the 6-OHDA lesion rodent model of PD [16], and reduce the MPTP-induced toxicity by activation of the adenosine A1 receptor to inhibit neuroinflammation [15]. In A (1-42) -injected AD rat model, paeoniflorin attenuates the neurotoxicity and cognitive decline by regulating calcium homeostasis and ameliorating oxidative stress [17]. Our results demonstrate both extract of P. lactiflora and paeoniflorin up-regulating HSF1 and HSP70 expressions to inhibit aggregate formation in ATXN3/Q 75 293 cells ( Figure 5). This aggregation-inhibitory effect can also be seen in neuronal differentiated SH-SY5Y cells expressing ATXN3/Q 75 (Figure 7), providing a novel mechanism of P. lactiflora and paeoniflorin to slow down the neurodegenerative process by inducing the expression of heat shock proteins.
Peoniflorin was the first reported active component of herbal medicines to induce expression of heat shock proteins in HeLa, IMR-32, and normal rat kidney cells [32]. Upon stress, HSF1 is released from the chaperone complex, selftrimerizes, and then is transported into the nucleus as a transcription factor. It activates chaperones, which play an important role in preventing unwanted protein aggregation. Overexpression of HSF1 significantly improved the life span of R6/2 Huntington's disease mouse [33]. Up-regulation of chaperone expression by HSF1 and its activating compounds 17-(allylamino)-17-demethoxygeldanamycin (17-AAG) demonstrate a strong inhibitory effect on HD aggregate formation [34]. We also showed aggregation-inhibitory effect of HSF1 overexpression in 293 cells expressing ATXN3/Q 75 ( Figure 6). Overexpression of its downstream target gene HSPA1A suppresses polyQ-mediated neurodegeneration in Drosophila and mouse models [35,36]. Reduced expression of HSPA8, another target genes of HSF1 [37], has been shown in the 293 cells overexpressing expanded TBP-Q 61 [38] and SCA17 lymphoblastoid cells [39]. The therapeutic potential of P. lactiflora and paeoniflorin in the treatment of SCA is strongly supported by the up-regulated HSF1 and HSP70 expressions.
In conclusion, our in vitro study provides strong evidence that P. lactiflora and paeoniflorin could be novel therapeutics for SCA3 and other polyQ diseases. Future application of P. lactiflora and paeoniflorin to SCA animal models would solidify their effects on aggregation reduction and disease improvement. Since the pathogenesis of the polyQ diseases is not completely clear and effective treatment is not available, our cell models are extremely valuable for identifying potential therapeutic targets in polyQ diseases. A systemic highthroughput screening of herbal and chemical compounds using ATXN3/Q 75 293 cell model is undergoing.