The Therapeutic Potential and Molecular Mechanism of Isoflavone Extract against Psoriasis

Psoriasis is a common inflammatory disease. It affects 1–3% of the population worldwide and is associated with increasing medical costs every year. Typical psoriatic skin lesions are reddish, thick, scaly plaques that can occur on multiple skin sites all over the body. Topical application of imiquimod (IMQ), a toll-like receptor (TLR)7 agonist and potent immune system activator, can induce and exacerbate psoriasis. Previous studies have demonstrated that isoflavone extract has an antioxidant effect which may help decrease inflammation and inflammatory pain. Through in vivo studies in mice, we found that the topical application to the shaved back and right ear of mice of isoflavone extract prior to IMQ treatment significantly decreased trans-epidermal water loss (TEWL), erythema, blood flow speed, and ear thickness, while it increased surface skin hydration, and attenuated epidermal hyperplasia and inflammatory cell infiltration. Through in vitro experiments, we found that isoflavone extract can reduce IL-22, IL-17A, and TNF-α-induced MAPK, NF-κB, and JAK-STAT activation in normal human epidermal keratinocytes. At the mRNA level, we determined that isoflavone extract attenuated the increased response of the TNF-α-, IL-17A-, and IL-22- related pathways. These results indicate that isoflavone extract has great potential as an anti-psoriatic agent and in the treatment of other inflammatory skin diseases.


Results
Isoflavone extract attenuates skin inflammation and hyperplasia in a murine IMQ-induced model. We studied the potential beneficial effect of isoflavone extract in a murine IMQ-induced psoriasis-like skin inflammation model in which IMQ, a TLR7/8 agonist, was applied to the whole back of the mice daily over 6 days. This repetitive application of IMQ onto mouse skin led to psoriasis-like inflammation with significant thickening, redness, and scaling caused by keratinocyte hyperproliferation and leukocyte infiltration into the skin ( Fig. 2A,B).
Histological analysis of lesion skin biopsies using H&E staining showed a decrease of inflammatory cell infiltrate in isoflavone extract-pretreated mice compared with IMQ-treated mice. In addition, isoflavone extract treatment prevented the characteristic epidermal hyperplasia induced by IMQ (Fig. 2C).
Topical pre-treatment with isoflavone extract prior to IMQ application ameliorated lesion formation in a dose-dependent manner and significantly reduced the development of erythema at the site of application (Fig. 3B). There were no significant changes after treatment with isoflavone extract alone. Moreover, our data showed that the treatment with isoflavone extract inhibited the TEWL, erythema, blood flow, ear thickness, and increased a corneometer in a dose-dependent manner. There were no significant differences in melanin or pH value between the IMQ treatment alone and mice treated with the isoflavone extract before IMQ treatment ( Fig. 3A-E).
Isoflavone extract is not cytotoxic to NHEKs. The cytotoxic effect of isoflavone extract on NHEKs was determined via MTT, trypan blue, and crystal violet assays. Treatment with isoflavone extract in a concentration range of 1-10 μg·mL −1 for 24 hours did not reduce cell viability (Fig. 4). However, the trial of 30 μg·mL −1 isoflavone extract showed slight cytotoxicity. We, therefore, used isoflavone extract at concentrations of 1-10 μg·mL −1 in our experiments.
Isoflavone extract inhibits TNF-α-, IL-17A-, and IL-22-induced MAPK pathway phosphorylation in NHEKs. MAP kinase signalling is critically involved in TNF-α-activated pathways and regulates various downstream effects in skin cells. This study, therefore, investigated possible signalling pathways modulated by isoflavone extract. A western blot analysis of NHEKs performed after they were pre-incubated with isoflavone extract for 24 h and stimulated with TNF-α (50 ng·ml −1 ) showed that the changes in phosphorylated ERK, p38, and c-Jun N-terminal kinase (JNK) expression were induced by TNF-α alone. Isoflavone extract-pretreatment suppressed the production of TNF-α-induced ERK, p38, and JNK phosphorylation in a dose-dependent manner (Fig. 5A).
To assess the activation of MAPK following exposure to IL-17A (50·ng mL −1 ), we examined phosphorylation levels of p38, ERK, and JNK as the relevant downstream molecules via western blot. Our results showed a significant increase in the phosphorylation of p38, ERK, and JNK after treatment with IL-17A alone. These increases could be significantly lowered with isoflavone extract in a dose-dependent manner (Fig. 5B). Our results demonstrate that isoflavone extract can effectively reduce the activation of MAPK.
The JAK-STAT pathway is a classical signal transduction pathway for numerous cytokines and growth factors 24 . IL-22 activates the JAK/STAT, ERK, JNK, and p38 MAPK pathways 25 . Therefore, we assessed how IL-22-induced activation is affected by the presence of isoflavone extract in NHEKs using western blot analysis. The analysis of IL-22-induced molecular cascades demonstrated that JAK/STAT and MAPK pathways were rapidly induced in NHEKs and that the phosphorylations of STAT3, JAK2, ERK, JNK, and p38 were inhibited by treatment with isoflavone (Fig. 5C).
Isoflavone extract inhibits TNF-α-and IL-17A-induced NF-κB pathway activation in NHEK. Since phosphorylation and degradation of the IκB protein, especially the α isoform (IκBα) are critical steps in the activation of the NF-κB signalling pathway 26 , we measured the effect of isoflavone extract on the TNFα-and IL-17A-induced phosphorylation and degradation of IκBα. Once NHEKs are stimulated by TNF-α or IL-17A, IκBαs are phosphorylated and degraded rapidly. As shown in Fig. 6, treatment of NHEKs with TNF-α or IL-17A increased the phosphorylation and degradation of IκBα, and this stimulation was prevented significantly by pre-treatment with isoflavone extract (Fig. 6). These results indicate that isoflavone extract influences NF-κB activation by regulating the phosphorylation and degradation of IκBα.
Isoflavone extract inhibits TNF-α-, IL-17A-, and IL-22-induced mRNA expression in NHEKs. Antimicrobial peptides and proteins (AMPs) are a diverse group of small molecules (12-50 amino acids), which are ubiquitously distributed in mammals, insects, frogs, and which can even be found in plants and invertebrates 27 . The main function of AMPs is to resist and kill pathogenic microorganisms such as bacteria, fungi, and viruses. In general, AMPs are concentrated in animal body surfaces, such as the skin and gastrointestinal tract 28,29 . In patients with psoriatic lesions, there are three subclasses of AMPs, including cathelicidin, S100 proteins, and defensins, which are highly expressed, and which are considered to play an important role in the pathogenesis of psoriasis 30,31 . RT-qPCR analysis of NHEKs revealed that during the development of psoriatic inflammation induced by TNF-α, IL-17A, or IL-22, increased levels of the AMPs CCL20, S100A7, S100A8, S100A9, hBD2, and LL-37 modulated and triggered host immune response. Moreover, we found that expression of CCL20, S100A7, S100A8, S100A9, hBD2, and LL-37 significant decreased after pretreatment with isoflavone extract, compared with the group treated with TNF-α, IL-17A, or IL-22 alone (Fig. 7). These data indicate that isoflavone extract attenuated the TNF-α-, IL-17A-, or IL-22-induced pathway at the mRNA level.

Discussion
The topical application of isoflavone extract prior to IMQ treatment significantly decreased the levels of TEWL and erythema, the speed of blood flow, and the thickness of the ear. The in vitro study demonstrated that, in human keratinocytes, cytokines such as TNF-α and IL-17A trigger the phosphorylation of the MAPK and NF-κB pathways and that IL-22 induces JAK/STAT and MAPK pathways, all of which induces inflammation. In the MAPK pathway, we found that the phosphorylation of signal transduction molecules such as ERK, p38 kinase, and JNK is upregulated by TNF-α, IL-17A, and IL-22. IκBα expression in the NF-κB pathway was also significantly upregulated after stimulation with TNF-α and IL-17A, and STAT3 and JAK2 expression in the JAK/STAT pathway was rapidly induced by IL-22. In contrast, pre-treatment with isoflavone extract downregulated the phosphorylation levels of p38 kinase, ERK, JNK, IκBα, STAT3, and JAK2.  Furthermore, the levels of CCL20 and the AMPs, includind S100A7, S100A8, S100A9, hBD2, and LL-37 alternate mRNA were analysed via qRT-PCR after treatment with the recombination proteins TNF-α, IL-17A, and IL-22 with or without pre-treatment with isoflavone extract. The results show that pre-treatment with isoflavone extract downregulated the levels of the AMPs.
Interleukin-22 (IL-22) is a member of the IL-10 family, which consists of 179 residues and has 25% identity with human IL-10 36 . Human IL-22 is characterized by six alpha-helices that help maintain its secondary structure 37 . IL-22 was originally found to be stimulated by IL-19 in mice with T lymphocytes using the cDNA subtraction method, and is commonly considered an immunosuppressive cytokine, even though it can induce an immune response, as well. IL-22 has been shown to induce rheumatoid arthritis, along with an increase in IL-17 in patients' plasma and synovial fluid, resulting in serious joint damage 38 . IL-22 also prevents inflammatory bowel  (IBD) by producing a pancreatitis-associated protein in pancreatic acinar cells 39,40 , and can prevent autoimmune liver diseases, such as primary biliary cirrhosis (PBC). IL-22 also shows protective effects that ameliorate liver inflammation by decreasing T cell and B cell infiltration in PBC mice 41 . After IL-22 binds to the receptor complex on the cell membrane, which consists of two receptor chains, the IL-22R1 chain and the IL-10R2 chain 42 , it activates the JAK-STAT pathway 43,44 . Further, IL-22 regulates the expression of genes responsible for antimicrobial defence, cellular differentiation, mobility in keratinocytes 45 , and inhibition of epidermal differentiation 46 . These bioactivities suggest that IL-22 plays an important role in inflammatory skin processes and wound healing, but may be harmful to patients with psoriasis. The level of IL-22 was positively correlated with the production of IL-22 in plasma and the severity of psoriasis, indicating that IL-22 has a significant effect on psoriasis 47 . Besides that, high IL-22 levels in psoriatic skin were associated with strongly upregulated cutaneous S100A7, S100A8, S100A9, and MMP1 expression 45 .
TNF-α is the primary mediator of inflammatory skin disease pathology. It is a multifunctional cytokine that plays an important role in inflammation, immune response, and apoptosis 48 . Patients with psoriasis have higher levels of TNF-α and translation factors such as NF-κB in their skin lesions 49,50 . This is due to TNF-α's ability to trigger NF-κB activation, which leads to the activation of other cytokines and upregulation of TNF-α itself, in a positive feedback loop 49,51 . In addition, TNF-α causes mTOR activation and ROS generation, which leads to IκB degradation, NF-κB translocation, and, finally, the production and secretion of inflammatory cytokines 52 . Moreover, TNF-α was shown to activate the Ras/Raf/MEK/ERK and protein kinase B/Akt pathways. In a human epidermal cell line, phosphatidylinositol (PI)3-kinase/Akt transduction was activated by TNF-α, which subsequently triggered the activation of NF-κB, which, in turn, regulated gene expression, and induced innate and adaptive immune responses and inflammation 53 . TNF-α also induces phosphorylation of JNK and reduces the expression of filaggrin and loricrin-two major proteins expressed by terminally differentiated epidermal keratinocytes 54 . Therefore, the use of a TNF-α antagonist may increase the expression of the skin barrier and ameliorate the symptoms of psoriasis by improving the barrier protein expression 55 . Antioxidants, also, have been shown to reduce the TNF-α-induced production of cytokines 52 , and may, therefore, play an important role in inflammatory skin diseases, especially psoriasis 52 . It is reasonable to assume that antioxidants have the potential to be developed into a new drug which would have both anti-inflammatory and barrier repair properties. Skin is the body's largest organ 56 and serves as the main barrier to the body's internal and external environment in order to maintain body homeostasis 57 , and to protect the human body from external damage. In addition, the skin can recognize, discriminate, and integrate various signals from the outside environment 58 , choose appropriate responses, and trigger a variety of physiological reactions including skin immune, pigmentary, epidermal, adnexal systems, systemic immune, neural, and endocrine systems [59][60][61][62] . In previous studies, it was reported that many skin-related responses are mediated via the cutaneous neuroendocrine system 63 . For example, in response to environmental changes, the skin can generate signals to produce neural, humoral, or immune responses at the local and systemic levels. In this study, we explored the effect of inflammatory-related molecules on the mechanisms of psoriasis. In the future, we hope to study the cutaneous neuroendocrine system in more detail.
Flavonoids have been widely investigated in the past. One flavonoid, epigallocatechin gallate (EGCG), has been reported to possess anti-inflammatory effects in the skin. EGCG has been found effective against UVB-induced prostaglandin metabolism, and also to reduce UVB-induced erythema, myeloperoxidase activity, hydrogen peroxide generation, and leukocyte infiltration 64,65 . In addition, EGCG inhibits UVB-induced AP-1 upregulation and the expression of apoptosis-regulatory genes (p53-p21) 66 , and possesses a protective effect on UVA-induced damage in human keratinocytes 67 . Similar to EGCG, (−)-epicatechin-3-gallate (ECG) can protect against UVA-induced cell death of human keratinocytes by reducing the generation of H 2 O 2 and hypoxanthine-xanthine oxidase 68,69 . Our previous research also found that (+)-catechin provides an anti-oxidation effect after UVB exposure, but the mechanism of (+)-catechin differs from that of EGCG, since (+)-catechin protects against UV and hydrogen peroxide damage via inhibition of JNK phosphorylation 70 . Genistein, found in soybeans, can prevent UVB-induced H 2 O 2 generation, lipid peroxidation, and oxidative DNA damage, and inhibits the progression and promotion of cancer via the inhibition of tyrosine protein kinase 71 . Further studies have shown that genistein can also suppress the expression of UVB-induced c-fos and c-jun proto-oncogenes, inhibiting the transformation of a normal cell into a tumour cell 72 .
Previous studies in our laboratory have shown that isoflavone extract works better against photo-aging (UVB) than genistein and is almost non-toxic. These two properties give isoflavones excellent potential for therapeutic applications and development in the future 73,74 . Theaflavins, the main component of black tea, are also reported to possess excellent antioxidant activity along with the ability to downregulate UV-induced lipid oxidation to reduce skin damage further 75 . Theaflavins also suppress the activation of transcription factor AP-1 76 , UVB-induced phosphorylation of STAT1 at Ser727 77 , and arsenite-induced apoptosis 78 . In another study, 12-o-tetradecanoylphorbol-13-acetate (TPA)-induced inflammation in a rodent model showed that theaflavins reduced ornithine decarboxylase (ODC) levels, which led to an anti-inflammatory effect 79 , and decreased NF-κB phosphorylation, thereby inhibiting tumour production 76 .
Another plant, indigo naturalis has been used as a medicine for the treatment of skin diseases in the past. In recent years, more and more reports have shown that indigo naturalis and indirubin can be effectively and safely used for the treatment of psoriasis 80 , especially nail psoriasis and plaque-type psoriasis [81][82][83] . Indigo naturalis upregulates claudin-1 expression and tight-junctions in human keratinocytes and also has a strong inhibitory effect on the inflammatory response of human neutrophils 84,85 . Therefore, an exploration of the relationship between natural products and activity will be of great value to the development of anti-inflammatory and barrier repair skin treatments.
The use of steroids can suppress both the immune system and inflammation, but it does not cure the source of the inflammation. When the use of steroids is discontinued, the inflammatory response will flare up again, unless the allergens causing the skin inflammation have been neutralized. Moreover, long-term and significant steroid use inevitably leads to systemic side effects, since steroids absorbed by the body inhibit the normal secretion of epinephrine. The long-term result is a reduction in pituitary and adrenal gland secretions, which causes dermis layer thinning that does not return to normal even after cessation of steroid use, microvascular proliferation, loss of melanin, and skin that begins to fade 86 . Thus, the use of steroids involves many limitations, while the use of natural products and Chinese herbal medicines is possible at a low cost, without significant restrictions, or side effects, which is a significant advantage in the treatment of inflammatory skin diseases such as psoriasis.

Materials and Methods
Ethics Statement. The foreskins used in the experiments was provided by the Mackay Memorial Hospital to the Fu Jen Catholic University with a letter providing an exemption from the institutional review board (IRB) (#13MMHIS022), as no interaction occurred with the foreskin donors themselves and no identifiable information was made available to the researchers. All animal tests were approved by the Fu Jen Catholic University's Institutional Animal Care and Use Committee (IACUC) policy (approval #A10367).
Animals. Mice were obtained from the National Laboratory Animal Center, Taipei, Taiwan. Animals were housed and handled according to institutional guidelines. Briefly, mice were housed one per cage in a controlled environment for 1 week with the temperature set at 21-25 °C, humidity at 60 ± 5%, and light in a 12/12 h light/ dark cycle. Alfalfa-free food and water were given ad libitum. All of the animal-experiment protocols were reviewed by the committee and conducted after obtaining an Affidavit of Approval of Animal Use Protocol from Fu Jen Catholic University.

IMQ-induced psoriasis-like skin inflammation in mice.
Male BALB/c mice (8-11 weeks) received a daily topical dose of 62.5 mg of commercially available IMQ cream (Aldara 5%; Meda AB, Solna, Sweden) on shaved areas of the dorsal and lumbar back and on the right ear for 6 consecutive days. Similarly, a vehicle cream (Vaselina Pura, Laboratorios Rida, Valencia, Spain) was applied to the IMQ-untreated mice. The trans-epidermal water loss (TEWL), amount of erythema, amount of melanin, skin hydration, pH value, ear thickness, and blood flow were measured before treatment. The surface changes in the dorsal skin were recorded using photography. The TEWL, erythema, melanin level, skin hydration, and blood flow were regularly measured using an MPA-580 cutometer (Courage & Khazaka, Cologne, Germany) and FLO-N1 laser tissue blood flowmeter (Omegawave, Tokyo, Japan). One hour before the cream was applied, isoflavone extract (10 mg/mL) or vehicle were topically administered. At the end of the experiment, the mice were killed, and tissues were collected and stored at −80 °C for subsequent homogenization and fixture in formalin. Animals were removed from the study and euthanized with CO 2 when clear suffering negated the need to continue, in accordance with the Fu Jen Catholic University's IACUC policy.
Histopathological analysis. Mouse tissues were fixed overnight with neutral buffered 4% paraformaldehyde (PFA) at 4 °C and then directly embedded in paraffin. Blocks of skin biopsies in paraffin were prepared using routine methods and sectioned to obtain consecutive levels. Five-micrometre sections were stained either with haematoxylin and eosin (H&E) or processed further. Images from H&E staining were obtained using a ZEISS Axioskop 40 Inverted System Microscope (NY, United States) and SPOT Cam software (Sterling Heights, MI).
Primary human keratinocyte isolation from foreskin and cell culture. Normal human epidermal keratinocytes (NHEKs) were obtained from normal adult human foreskins. All protocols and procedures were approved by the local ethics committee and carried out in accordance with the Declaration of Helsinki. Written consent was obtained from each donor before the experiments were performed. In brief, the skin was divided and incubated in 0.2% protease (Sigma-Aldrich, St Louis, MO) for 2 days at 4 °C. After the epidermis separated from the dermis, the keratinocytes were disaggregated into a single-cell suspension in serum-free Dulbecco's modified Eagle's medium (DMEM) and then centrifuged at 1100 × rpm for 5 minutes. The keratinocytes were then cultured in Keratinocyte-SFM (Gibco BRL/Invitrogen, Carlsbad, CA), supplemented with recombinant epidermal growth factor (0.1-0.2 ng·mL −1 ), bovine pituitary extract (20-30 μg·mL −1 ), and 1% penicillin/streptomycin in a humidified atmosphere at 37 °C and 5% CO 2 . The second-to fourth-passage cells were used in the experiments.
Stimulation of primary human keratinocytes. NHEKs were seeded in 12-well plates at a density of 5 × 10 4 cells/well. After reaching subconfluency, the keratinocyte medium was renewed, and the cells were subjected to a 24 h-pretreatment with isoflavone extract. Control keratinocytes were placed in an equal volume of vehicle. Finally, cells were stimulated with either IL-17A (50 ng·mL −1 ), IL-22 (50 ng·mL −1 ), or TNF-α (50 ng·mL −1 ) from PeproTech (Rocky Hill, NJ).

Cell viability assays (MTT, trypan blue assay, and crystal violet assay).
Cell viability was determined via MTT, trypan blue, and crystal violet assays. The MTT assay was performed according to a previously described protocol 68 . Briefly, cells were pretreated with double-distilled (dd)H 2 O or isoflavone extract and incubated for 24 hours. After being washed with the phosphate buffer solution (PBS), MTT (0.5 mg/mL in Keratinocyte-SFM) was used for the quantification of metabolically active cells. Mitochondrial dehydrogenases metabolized MTT to a purple formazan dye, which was analysed photometrically at 550 nm. Cell viability was proportional to the absorbance.
The Trypan Blue exclusion method was used, according to the manufacturer's protocols, to accurately determine cell viability subsequent to isolation. Briefly, NHEKs were seeded in a 35 mm culture dish. After reaching subconfluency, the Keratinocyte-SFM was renewed, and the cells were subjected to a 24-h pretreatment with isoflavone extract. The cells were then collected into cellular suspension with TrypLE Express (Gibco BRL/ Invitrogen, Carlsbad, CA) and stained with equal volumes of 0.4% trypan blue dye for 1 min. The cells were counted using a dual-chamber haemocytometer and a light microscope.
The crystal violet assay was used to determining the viability of cultured cells 87 . Crystal violet is a triarylmethane dye (purple) that stains DNA and proteins in cells that are grown into a monolayer. The cytotoxicity of a drug is determined by measuring the absorbance of the crystal violet-stained cells via spectrophotometry. Cell treatments were performed as above described. After a 24-h incubation period, the cells were washed with PBS to remove non-adherent cells. Then, methanol (500 μL) was added to each well to fix the living cells to the bottom of the plate, and the cells were incubated for 30 min at 25 °C. The methanol was discarded and 0.1% crystal violet staining solution (300 μL) was added for 1 h. The crystal violet was discarded, and the plate was washed three times with ddH 2 O. The plates were left to dry and 33% acetic acid (500 μL) was added to dissolve the stained cells. Absorbance optical density (OD) was read using a Tecan Sunrise spectrophotometer (Tecan, Crailsheim, Germany) at a wavelength of 550 nm.
Real-time quantitative reverse transcriptase -polymerase chain reaction (RT-qPCR). After these treatments, the cells were lysed and the total RNA was extracted using a total RNA isolation kit SCIentIFIC REPoRTs | (2018) 8:6335 | DOI:10.1038/s41598-018-24726-z (GeneDireX ® , Vegas, NV). First strand cDNAs were synthesized using the SuperScript ® III First-Strand Synthesis System (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. qPCR was performed using an CFX96 ™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA) and under the following conditions: heat to 95 °C for 3 min, followed by 54 denaturation cycles at 95 °C for 3 s each, primer annealing at 60 °C for 20 s, and primer extension at 95 °C for 10 s. The melting temperature curve ranged from 60 to 95 °C, increasing in increments of 0.3 °C. cDNA was amplified using SYBR green (Kapa Biosystems, Wilmington, MA) and using the following primers: human CCL20 88 , forward TACTCCACCTCTGCGGCGAATCAGAA, a n d r e v e r s e G T G A A A C C T C C A A C C C C A G C A A G G T T ; h u m a n S 1 0 0 A 7 8 9 , f o r w a r d GCATGATCGACATGTTTCACAAATACAC, and reverse TGGTAGTCTGTGGCTATGTCTCCC; human S100A8 90 , forward TGAAGAAATTGCTAGAGAC, and reverse CTTTATCACCAGAATGAGGA; h u m a n S 1 0 0 A 9 4 6 , f o r w a r d G C T C C T C G G C T T T G A C A G A G T G C A A G , a n d r e v e r s e GCATTTGTGTCCAGGTCCTCCATGATGTGT; human BD2 91 , forward CCAGCCATCAGCCATGAGGGT, and reverse GGAGCCCTTTCTGAATCCGCA; human LL-37 92 , forward GCAGTCACCAGAGGATTGTGAC, and reverse CACCGCTTCACCAGCCC; human β-Actin 93,94 , forward CGGGGACCTGACTGACTACC, and reverse AGGAAGGCTGGAAGAGTGC. The CCL20, S100A7 to 9, hBD2, and LL-37 amplification signals were normalized relative to β-actin expression and evaluated using the equation: fold change = 2 −ΔΔCT .
Statistical analysis. Unless otherwise indicated, data are expressed as mean ± standard error of the mean (SEM) using the GraphPad Prism Program 6 software (GraphPad Software, San Diego, CA). Comparison of the mean survival rates of cells with and without isoflavone extract was performed using a one-way analysis of variance (ANOVA) followed by Dunnett's t-test for multiple comparisons. We considered P < 0.05 to be statistically significant.