Isoflavones inhibit poly(I:C)-induced serum, brain, and skin inflammatory mediators - relevance to chronic fatigue syndrome

Background Chronic Fatigue Syndrome (CFS) is a neuroimmunoendocrine disease affecting about 1% of the US population, mostly women. It is characterized by debilitating fatigue for six or more months in the absence of cancer or other systemic diseases. Many CFS patients also have fibromyalgia and skin hypersensitivity that worsen with stress. Corticotropin-releasing hormone (CRH) and neurotensin (NT), secreted under stress, activate mast cells (MC) necessary for allergic reactions to release inflammatory mediators that could contribute to CFS symptoms. Objective To investigate the effect of isoflavones on the action of polyinosinic:polycytidylic acid (poly(I:C)), with or without swim stress, on mouse locomotor activity and inflammatory mediator expression, as well as on human MC activation. Methods Female C57BL/6 mice were randomly divided into four groups: (a) control/no-swim, (b) control/swim, (c) polyinosinic:polycytidylic acid (poly(I:C))/no swim, and (d) polyinosinic:polycytidylic acid (poly(I:C))/swim. Mice were provided with chow low or high in isoflavones for 2 weeks prior to ip injection with 20 mg/kg poly(I:C) followed or not by swim stress for 15 minutes. Locomotor activity was monitored overnight and animals were sacrificed the following day. Brain and skin gene expression, as well as serum levels, of inflammatory mediators were measured. Data were analyzed using the non-parametric Mann-Whitney U-test. Results Poly(I:C)-treated mice had decreased locomotor activity over 24 hours, and increased serum levels of TNF-α, IL-6, KC (IL-8/CXCL8 murine homolog), CCL2,3,4,5, CXCL10, as well as brain and skin gene expression of TNF, IL-6, KC (Cxcl1, IL8 murine homolog), CCL2, CCL4, CCL5 and CXCL10. Histidine decarboxylase (HDC) and NT expression were also increased, but only in the skin, over the same period. High isoflavone diet reversed these effects. Conclusion Poly(I:C) treatment decreased mouse locomotor activity and increased serum levels and brain and skin gene expression of inflammatory mediators. These effects were inhibited by isoflavones that may prove useful in CFS. Electronic supplementary material The online version of this article (doi:10.1186/s12974-014-0168-5) contains supplementary material, which is available to authorized users.

patients are particularly vulnerable to stress [14]. Many CFS symptoms could derive from the possible release of inflammatory mediators that could affect brain function [19,20]. As of today, there are no FDA approved drugs for the treatment of CFS [21]; psychological, physical, and pharmacological interventions used currently are not very effective [22].
An abnormal immune component may be involved in CFS [23][24][25], but the neuroimmune and neuroendocrine interactions involved are still unknown [26]. Mast cells (MC) and their mediators have been implicated in all diseases that are comorbid with CFS [9]. There is higher number of skin MC in patients with CFS [27,28], and such patients also show increased skin hypersensitivity [29]. Furthermore, CFS also occurs more often in patients with chronic urticaria, that also involves MC [30]. In fact, there is hyperresponsiveness in the bronchi of CFS patients, implying MC activation [31].
Activated MC release a number of chemokines and cytokines that could contribute to CFS symptoms [32,33]. MC are located perivascularly in close proximity to neurons [34], especially in the diencephalon [35,36], where functional MC-neuron interactions have been documented [36,37] in response to corticotropinreleasing hormone (CRH) [38]. In vivo activation of MC by CRH is augmented by neurotensin (NT) [39]. Moreover, NT is induced in the hypothalamus in response to bacterial lipopolysaccharide (LPS) and regulates the HPA axis [40].
Unfortunately, there are neither effective CFS treatments nor human MC inhibitors clinically available that may also be used in CFS. Flavonoids are natural compounds with strong antioxidant and anti-inflammatory activity [41]. Certain flavonoids also inhibit MC [42] and have neuroprotective effects [41,43].
Here we report that treatment of mice with poly(I:C) results in reduced locomotor activity and increased serum levels, as well as brain and gene expression, of inflammatory mediators, all of which are reversed by treatment with the isoflavones daidzein and genistein.

Animals
C57BL/6 female mice, nine to twelve weeks old, (Jackson Laboratories, Bar Harbor, ME, USA) were kept in virus-free sections of a modern animal facility and were allowed ad libitum access to food and water. They were maintained on a 14:10 hour light-dark cycle (the standard light-dark cycle used by the Department of Animal Health). Female mice were chosen because published reports indicate a female to male ratio of 4:1 [5], while the US Centers for Disease Control and Prevention (CDC) specify a female to male ratio of 4:1 [5]. Mice were kept in cages of five mice/cage until the day of the experiments. Both low and high isoflavone diets (2018X and 2020X) were sterile with similar ingredients other than isoflavone content. We monitored weight changes for 21 days prior to the beginning of the experiments. Only poly(I:C)/no swim-treated mice showed a slight decrease in weight change. Poly(I:C)/swim-treated mice, as well as their corresponding control mice, slightly increased their weight over the three-week observation period. However, by the end of this period, there was no statistical difference in weight change. The protocol was approved by Tufts Medical Center IACUC under number B 2011-88.

Treatment conditions
Mice were provided with chow containing either nondetectable to low (ND-20 mg/kg, Teklad 2920X) or high (150 to 250 mg/kg, Teklad 2918X) isoflavone (daidzein plus genistein) levels for two weeks. Conditions included four groups: (a) control (normal saline intraperitoneal (ip) injection)/no swim, (b) control/swim, (c) poly(I:C)/ no swim, and (d) poly(I:C)/swim (n = 5 to 7/group). Mice were injected ip with 20 mg/kg of poly(I:C) or normal saline the first day. Subsequently, they were subjected to swim for 15 minutes, individually in a transparent plastic cylindrical jar (17 cm × 25 cm) containing 15 cm-deep water at room temperature (23 ± 1°C). This approach reflects both exercise and the stress of water immersion. Mice were then placed individually into specific cages and locomotor activity was monitored overnight.

Assessment of behavioral parameter-locomotor activity
After the experimental procedures, animals were placed individually into standard plastic housing cages with food and water available ad libitum and overnight locomotor activity (for a total of 16 hours) was monitored with the Neuroscience Behavior Core's mouse SmartFrame® Cage Rack System (Kinder Scientific, Poway, CA, USA). This system consists of 20 PC-interfaced horizontal photobeam frames. The frame (containing 12 photocells; arranged on a 8 L × 4 W grid) surrounds one home cage environment and continuously tracks the animal's movement. This fully automated system allows the user to quantify horizontal ambulation by counting breaks in infrared photocell beams using MotorMonitor® software (Hamilton-Kinder Scientific, Poway, CA, USA). Data were collected and subsequently analyzed in time bins (every hour) or as a total over the course of collection to the 'Total Distance Travelled' (in cm) parameter for each zone.

Sample collection
Mice were euthanized 24 hours post poly(I:C) ip injection using isoflurane overdose and thoracotomy. Blood was collected by cardiac puncture and was used to determine inflammatory mediator levels in the serum. Brain (diencephalon) and skin (back shaved with a electric shaver the day before) samples were collected and immersed into RNAlater (catalog# AM7021) purchased from Invitrogen (Grand Island, NY, USA). Samples were stored at -80°C.

Quantitative PCR
Total RNA from mouse tissues was extracted using RNeasy Plus Mini kit (catalog# 74134) and RNeasy Fibrous Tissue Mini Kit (catalog# 74704), purchased from QIAGEN (Valencia, CA, USA). Reverse transcription was performed with 300 ng of total RNA using the iScript cDNA synthesis kit (catalog# 170-8891) purchased from Bio-Rad (Hercules, CA, USA). Real-time quantitative polymerase chain reaction (qPCR) was carried out in a 7300 Sequence Detector, according to TaqMan Gene Expression Assay instructions from Life Technologies, Applied Biosystems (Grand Island, NY, USA) using Taqman primer/probe sets (Additional file 1: Table S1). Samples were analyzed for Tnf (TNF-α), Il4, Il6 (IL-6), KC (Cxcl1, IL-8 homolog in the mouse), Ccl2 (CCL2), Ccl4 (CCL4), Ccl5 (CCL5), Cxcl10 (CXCL10), histidine decarboxylase-Hdc (HDC), and Nts (NT) gene expression using Gapdh as internal control.Thermal cycling proceeded at 50°C for 2 minutes, 95°C for 10 minutes, 95°C for 15 seconds, for 40 cycles, and 60°C for 1 minute. Negative controls included samples with water instead of template. Assays were performed in triplicate for each data point. Results were normalized against the endogenous gene GAPDH and were expressed relative to the mean of the control for each gene (relative fold change).

Statistical analysis
Data were analyzed using the non-parametric Mann-Whitney U-test. Results are presented as mean ± SE. Statistical significance is defined as P < 0.05. We analyzed separately the different treatment subgroups within each dietary group, using Mann-Whitney U-test to compare poly(I:C)/no swim and poly(I:C)/swimtreated mice with the control/no swim and control/ swim-treated groups, accordingly. We then analyzed the same treatment subgroups between the two dietary groups, using Mann-Whitney U-test to compare, for example, poly(I:C)/swim-treated mice with low isoflavone diet to poly(I:C)/swim-treated mice that were provided with high isoflavone diet. Due to space limitations, we submitted figures composed of all the experimental conditions from the two different diets and all the statistical significant results were marked in these figures.

Effect of poly(I:C) and isoflavones on locomotor activity
Poly(I:C)/swim and poly(I:C)/no swim-treated mice on low isoflavones had reduced maximum (max) locomotor activity (over the 10-hour night period), denoted by the dark bar (P = 0.008 and P = 0.036) compared to the control/swim and control/no swim-treated mice, respectively ( Figure 1A). High isoflavone chow reversed this decrease ( Figure 1B), and the difference between the two diets was statistically significant (P = 0.032).

Effect of poly(I:C) and isoflavones on serum inflammatory mediators
Poly(I:C)-treated mice on low isoflavones had increased serum levels of TNF-α (Figure 2A Table S2). High isoflavones reduced the poly(I:C)-increased serum levels of all the inflammatory markers. All IFNγ, IL-1β, IL-4, IL-9, IL-10, IL-12p70, IL-17 and VEGFα serum levels were below the detection limit, while IL-1α serum levels were similar between the different treatment groups. Swim stress augmented the poly(I:C) effect on TNF-α, CCL4, and CCL5 serum levels.
Poly(I:C) treatment also increased CCL2, CCL3, CCL5 and CXCL10 serum levels and high isoflavones reduced those increases (Additional file 3).
Effect of poly(I:C) and isoflavones on brain gene expression of inflammatory mediators TNF-α ( Figure 3A), IL-6 ( Figure 3B), KC ( Figure 3C), CCL4 ( Figure 3D), as well as CCL2, CCL5 and CXCL10 brain gene expression were increased in the poly(I:C)treated mice. High isoflavones reduced the increased TNF, IL6, KC, CCL4 and CCL2 brain gene expression. There was no difference in HDC and NT gene expression between the different treatment groups (Additional file 4: Table S3).

Effect of poly(I:C) and isoflavones on skin gene expression of inflammatory mediators
TNF-α ( Figure 4A), IL-6 ( Figure 4B), KC ( Figure 4C), CCL4 ( Figure 4D), as well as CCL2, CCL5 and CXCL10 (Additional file 5: Table S4) skin gene expression were increased in the poly(I:C)-treated mice. Moreover, HDC Figure 2 Polyinosinic:polycytidylic acid (poly(I:C)) and isoflavone effect on serum levels of inflammatory mediators. Mice were provided with either low or high isoflavone diet ad libitum. Mice were injected intraperitoneally (ip) with 20 mg/kg poly(I:C) and were forced to swim for 15 minutes. Following this, they were individually placed into specific cages and locomotor activity was monitored overnight. The next day, serum samples were collected and analyzed for TNF-α (Figure 2A and NT gene expression were also increased in the skin ( Figure 5A and B). High isoflavone diet reduced the increased TNF-α, IL-6 and KC skin gene expression.
CCL4 gene expression was increased in the skin of poly (I:C)-treated mice on low isoflavones (24 ± 15 and 21 ± 19) compared to their controls (0.4 ± 0.3, P = 0.0021 and 1 ± 1.7, P = 0.0023, respectively) ( Figure 4D). CCL4 gene expression was also increased in the skin of poly(I:C)-treated mice on high isoflavones (20 ± 24 and 27 ± 20) compared to their controls (0.8 ± 0.7, P = 0.0079 and 1 ± 1.3, P = 0.0079, Figure 3 Polyinosinic:polycytidylic acid (poly(I:C)) and isoflavone effect on brain gene expression of inflammatory mediators. Mice were provided with either low or high isoflavone diet ad libitum. Mice were injected intraperitoneally (ip) with 20 mg/kg poly(I:C) and were forced to swim for 15 minutes. Following this, they were individually placed into specific cages and locomotor activity was monitored overnight. The next day, brain samples were collected and analyzed for TNF-α ( Figure 3A respectively) and high isoflavones did have an effect on reducing this increase.
Finally, poly(I:C) treatment also increased CCL2, CCL5 and CXCL10 gene expression in the skin of the mice. High isoflavones did not have any statistical effect on CCL4, CCL2, CCL5, CXCL10, HDC and NT skin gene expression (Additional file 5: Table S4).

Discussion
Here we show that poly(I:C) significantly reduced locomotor activity over the first 24 hours, in comparison to control mice. Forced swim did not have any effect on its own, but augmented the effect of poly(I:C) on increasing TNF-α, CCL3 and CCL5 serum levels. We also used BALBc mice in an effort to investigate the possibility of Figure 4 Polyinosinic:polycytidylic acid (poly(I:C)) and isoflavone effect on skin gene expression of inflammatory mediators. Mice were provided with either low or high isoflavone diet ad libitum. Mice were injected intraperitoneally (ip) with 20 mg/kg poly(I:C) and were forced to swim for 15 minutes. Following this, they were individually placed into specific cages and locomotor activity was monitored overnight. The next day, skin samples were collected and analyzed for TNF-α ( Figure 4A), IL-6 ( Figure 4B), KC ( Figure 4C) and CCL4 ( Figure 4D) skin gene expression using qPCR. strain differences, but there was no difference from our findings with C57BL/6 mice (results not shown).
Use of poly(I:C) increased serum levels of molecules that have been associated with inflammation and fatigue including TNF-α, IL-6, KC, CCL2, CCL3, CCL4, CCL5, and CXCL10. Moreover, poly(I:C) also increased brain and skin gene expression of TNF-α, IL-6, KC, CCL2, CCL4, CCL5, CXCL10, while HDC and NT gene expression were only increased in the skin, suggesting that NT and histamine may explain the skin findings in CFS patients. Forced swim stress had no additional effect to that of poly(I:C), except for augmenting TNF-α serum levels. TNF-α and IL-6 serum levels were actually increased in CFS [20,44]. Such pro-inflammatory mediators can increase blood-brain barrier (BBB) permeability [45] and permit entry of circulating leukocytes leading to brain inflammation. TNF-α was shown to be released along with histamine from rat brain MC [46], and was involved in brain inflammation [47]. MC can also interact with T cells [48][49][50] and superactivate them through TNF-α [51].
Unlike our present results, one group reported that daily forced swim stress for two to three weeks, with or without an immunological trigger administered on day one (lipopolysaccharide or Brucella abortus antigen), induced 'chronic fatigue', but only in Albino laca mice [52][53][54][55] and Wistar rats [56]. In these papers, behavioral parameters, such as immobility time, time to start grooming after swim stress, roda rod test, and elevated plus maze test were increased indicating post stress fatigue and anxiety. Also, biochemical measurements indicative of brain oxidative stress were increased. In another paper, BALBc mice were injected with Brucella abortus and developed decreased running activity that lasted one week [57]. Forced swim of Charles Foster albino rats for twenty-one days also increased immobility time, anxiety as assessed by elevated plus maze test, and brain oxidative stress [58]. These findings maybe due to the differences in the strains and triggers used.
Here we also show that poly(I:C) with or without NT or CRH did not have any effect on TNF-α, CXCL8 and VEGFα release from human cultured LAD2 MC (results not shown), but increased only TNF-α gene expression when used together with CRH, NT, or SP. Poly(I:C) alone did not have any effect on human MC, but increased TNF-α gene expression at 24 hours when used together with CRH or NT.
Human umbilical cord blood-derived MC express TLR-3, activation of which produced IFNα and IFNβ in response to double-stranded RNA [73]. A recent publication reported that MC respond to intracellular, but Figure 5 Polyinosinic:polycytidylic acid (poly(I:C)) and isoflavone effect on skin gene expression of inflammatory mediators. Mice were provided with either low or high isoflavone diet ad libitum. Mice were injected intraperitoneally (ip) with 20 mg/kg poly(I:C) and were forced to swim for 15 minutes. Following this, they were individually placed into specific cages and locomotor activity was monitored overnight. The next day, skin samples were collected and analyzed for HDC ( Figure 5A) and NT ( Figure 5B) skin gene expression using qPCR. not extracellular, poly(I:C) by inducing mainly IFNα and TNF-α; moreover, infection of MC with live Sendai virus induces an anti-viral response similar to that of intracellular poly(I:C) [74].
We also used CRHR-1 knockout (KO) mice in order to investigate the possible involvement of CRHR in any stress effect. However, forced swim stress alone did not have an effect and there was no difference using these mice (results not shown). Nevertheless, rats exposed to water immersion stress had a four-fold increase in plasma histamine levels that was absent in W/W v MCdeficient rats [75]. Acute stress also increased serum histamine and IL-6 levels, both of which were also absent in MC deficient W/W v mice [76].
Here we also show that the isoflavones genistein and daidzein reversed the effect of poly(I:C) on mice. Genistein has been reported to attenuate muscle fatigue [77], protect against endothelial barrier dysfunction [78] and suppress LPS-induced inflammatory response in macrophages [79]. Isoflavones also suppress MC expression of the high affinity IgE receptor (FcεRI) [80]. However, isoflavones have estrogenic activity and may not be desirable in certain clinical settings. The flavonoids epigallocatechin, naringin, and curcumin ameliorated 'chronic fatigue' [53][54][55][56]. Other papers reported similar effects for the Astragalus flavonoids [81] and for the olive extract [82].
Flavonoids exert potent anti-inflammatory effects via various pathways [83][84][85][86][87]. A review of human randomized controlled trial studies summarized some significant benefits to cognitive function after isoflavone supplementation [88]: improvements in executive function, working memory and processing speed [88]. Specifically, two studies reported significant effects of 60 mg/day treatment with isoflavones in processing speed and psychomotor speed [89,90].
Quercetin increases exercise tolerance in mice [91]. Oral administration of quercetin leads to accumulation in brain tissue and attenuates the increased oxidative stress in the hippocampus and striatum of rats exposed to chronic forced swimming [92,93]. Quercetin has potent anti-oxidant and anti-inflammatory activity [41,42], and inhibits MC degranulation [94,95], as well as TNFα, IL-6, and IL-8 secretion [95,96]. Moreover, it reverses acute stress-induced behavioral changes and reduces brain glutathione levels in mice [97].
Recent studies with antigen-stimulated MC show that epigallocatechin gallate (EGCG) inhibits MC degranulation, leukotriene C 4 secretion, as well as the production of TNFα, IL-6 and IL-8 [94,96]. The quercetin-related flavone luteolin inhibits MC activation [98] and MC-dependent stimulation of activated T cells [51]. Luteolin also inhibits IL-6 release from microglia cells [99] and from astrocytes [100]. Recent reviews have addressed the possible use of flavonoids for the treatment of CNS diseases [101,102].
Increasing evidence implicates CNS inflammation [103], as well as MC-microglia interactions, in neuropsychiatric diseases [104,105]. MC are important for allergic reactions, but also in immunity [106,107], and inflammatory diseases [15,108]. TLRs have also been implicated in CNS dysfunction through MC and glial activation [104].