Garlic (Allium sativum) Stimulates Lipopolysaccharide‐induced Tumor Necrosis Factor‐alpha Production from J774A.1 Murine Macrophages

Garlic (Allium sativum) is known to have many beneficial attributes such as antimicrobial, antiatherosclerotic, antitumorigenetic, and immunomodulatory properties. In the present study, we investigated the effects of an aqueous garlic extract on macrophage cytokine production by challenging the macrophage J774A.1 cell line with the garlic extract in the absence or presence of lipopolysaccharide (LPS) under different conditions. The effect of allicin, the major component of crushed garlic, was also investigated. Using enzyme‐linked immunosorbent assay and reverse transcriptase‐quantitative polymerase chain reaction, it was found that garlic and synthetic allicin greatly stimulated tumor necrosis factor‐alpha (TNF‐α) production in macrophages treated with LPS. The TNF‐α secretion levels peaked earlier and were sustained for a longer time in cells treated with garlic and LPS compared with cells treated with LPS alone. Garlic acted in a time‐dependent manner. We suggest that garlic, at least partially via its allicin component, acts downstream from LPS to stimulate macrophage TNF‐α secretion. © 2014 The Authors. Phytotherapy Research published by John Wiley & Sons, Ltd.


INTRODUCTION
Garlic has been used as a home remedy for a myriad of diseases for thousands of years. To date, garlic and its constituents are known to have many beneficial attributes such as antimicrobial (Casella et al., 2013;Goncagul and Ayaz, 2010), antiatherosclerotic (Asdaq et al., 2009;Chan et al., 2013;Gonen et al., 2005;Malekpour-Dehkordi et al., 2013), antitumorigenetic (Iciek et al., 2009;Kaschula et al., 2011), and immunomodulatory (Iciek et al., 2009) properties. Of particular interest for our studies is the modulation of cytokine production by garlic as cytokines are important immune system mediators. Different types of garlic extracts have been shown to directly modulate cytokine production in lymphocytes (Dong et al., 2011;Liu et al., 2009;Zamani et al., 2009) and macrophages (Chang et al., 2005;Dong et al., 2011;Hodge et al., 2002;Keiss et al., 2003;Romano et al., 1997). Garlic is largely recognized as having antiinflammatory properties; however, not all reports provide evidence indicating that garlic increases antiinflammatory [i.e. interleukin-10 (IL-10)] cytokines while decreasing pro-inflammatory (i.e. tumor necrosis-alpha (TNF-α), IL-1β, and IL-6) cytokines. Thus, we set out to investigate the effect of a water garlic extract and allicin, the major component of crushed garlic (Rybak et al., 2004), on murine macrophage cytokine production induced by lipopolysaccharide (LPS). The LPS is found on the cell wall of Gram-negative bacteria (Rietschel and Brade, 1992) and has been widely used to cause an immune response in vivo and in vitro from various immune cells including the activation of macrophages (Fiorentino et al., 1991;Parameswaran and Patial, 2010;Ulmer et al., 2000). Of the cytokines investigated (IL-1β, IL-6, IL-10 and IL-12, and TNF-α), we found that a water garlic extract consistently stimulated LPS-induced TNF-α secretion from macrophages. Transcription of TNF-α was also significantly enhanced by garlic, and allicin, in LPS-treated cells. In addition, macrophages treated with the aqueous garlic extract alone secreted small, albeit significant amounts of TNF-α.

MATERIALS AND METHODS
Preparation of water garlic extract. Garlic was purchased at the local grocery store. The cloves of garlic from two garlic bulbs were peeled, and any discolored parts or blemishes were cut off. Garlic was weighed (56.77 g) and added to a waring blender containing 113.54 mL of pyrogen-free water (HOSPIRA, Inc., Lake Forest, IL). The garlic to water ratio was modeled after the aqueous garlic extraction method reported by Gamboa-León et al. (2007). The garlic was then blended for 1 min, and the homogenate was passed through grade 50 cheesecloth (Lymex, Chicopee, MA). In a tissue culture hood, the flow-through was then filter-sterilized using a 150-mL Nalgene 0.2-μm pore-sized filter unit (Nalge Nunc International, Rochester, NY) connected to a vacuum pump. The flow-through from the Nalgene filter was aliquoted into microcentrifuge tubes and stored at À80°C until used. To determine endotoxin contamination, the garlic extract was sent for endotoxin testing using the kinetic chromogenic limulus amebocyte lysate assay (Charles River Laboratories, Charleston, SC).
Determination of allicin content in garlic. Allicin in the garlic extracts was quantified using a reversed-phase HPLC method that employed the Agilent-1100 HPLC system (Agilent, CA) with a Zorbax Eclipse XDB-C18 column. The garlic extracts were eluted through the column by the isocratic solvent of water/methanol (50/50) at a flow rate of 0.75 mL/min, and the absorbance of the eluate was monitored at 254 nm.
Allicin synthesis. Allicin was synthetized by the peroxyacid oxidation of diallyl disulphide (Alfa Aesar, Ward Hill, MA) according to the procedure of Small et al. (1947). Pre-purification of diallyl disulphide by fractional vacuum distillation to remove diallyl sulphide and diallyl trisulphide was critical for obtaining pure allicin after oxidation. Because of the thermal instability of allicin, the crude product was purified by silica gel chromatography rather than distillation.
Before treatment, the conditioned media was removed from the cells 24 h after plating them. Fresh media was then added to the cells, and they were treated with the garlic or synthetic allicin. The garlic extract was added to the cells at varying dilutions in pyrogen-free water (1:40, 1:100, 1:200, 1:500, 1:1000, and 1:2000). Alternatively, allicin (1-10 μg/mL) also diluted in pyrogen-free water was added to the cells. Garlic or allicin was added in the presence or absence of 0.1 μg/mL LPS (LPS from Escherichia coli 055:B5, Sigma, St. Louis, MO). The cells were then incubated at 37°C and 5% CO 2 for 24 h. After this incubation period, cell supernatants were collected and stored at À80°C until assaying for cytokine secretion. Analysis of cytokine levels revealed an effective dilution of garlic (1:500) that modulated cytokine production.
To determine the minimum LPS concentration that would cause the maximum TNF-α stimulation in the presence of garlic, J774A.1 macrophages were plated in 96 well plates as described earlier and treated with the garlic extract (G1:500) and different concentrations of LPS (0.001 to 10 μg/mL). To further characterize the interaction between LPS and the effective garlic extract dilution (G1:500), J774A.1 macrophages were plated in 96 well plates as described earlier. The cells were then treated with G1:500 for different periods before, after, or at the same time as LPS (0.1 μg/mL). The cells were incubated at 37°C for 24 h (or for the time specified in the figure legend). At the end of this incubation period, the supernatants were collected from the wells and stored at À80°C until assaying for cytokine secretion. In some experiments, G1:500 was removed prior to the addition of the LPS. In these experiments, the conditioned media in the wells was replaced with fresh media prior to adding the LPS. The cells were incubated at 37°C for 24 h after the addition of LPS. At the end of the 24 h, the supernatants were collected from the wells and stored at À80°C until assaying for cytokine secretion. Alternatively, in other experiments, LPS was added to the cells for different periods and removed prior to the addition of G1:500. In these experiments, the conditioned media in the wells was replaced with fresh media prior to adding G1:500. These cells were then incubated at 37°C for 24 h after the addition of G1:500. At the end of the 24 h, the supernatants were collected from the wells and stored at À80°C until assaying for cytokine secretion.
Cell proliferation XTT {2,3-bis (2-methoxy-4-nitro-5sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide} assay. A cell proliferation assay was carried out to determine the effect of garlic or allicin treatment on macrophage cell viability. Thus, macrophages were plated (2.5 × 10 5 cells/mL, 100 μL/well) in a 96 well plate and incubated at 37°C. After incubation for 24 h, the complete media was removed, and new complete media was added (100 μL/well). The cells were then treated with different dilutions of the garlic extract or synthetic allicin in the absence or presence of LPS (0.1 μg/mL). The effect of the garlic extract on cell viability was assayed using the Cell Proliferation Kit II (XTT) from Roche (Indianapolis, IN) as described by the manufacturer.
mRNA quantification. To determine the effect of garlic and allicin on TNF-α and toll-like receptor 4 (TLR4) mRNA levels, J774A.1 macrophages were plated at a concentration of 1.25 × 10 5 cells/mL, 5 mL/well in complete media in six well tissue culture plates. The conditioned media was removed from the cells 24 h after plating them, and fresh media was added. The cells were then treated for 2 h with the garlic or synthetic allicin in the absence or presence or LPS (0.1 μg/mL). Preparation of RNA, cDNA, and qPCR was carried out with reagents from Bio-Rad (Hercules, CA). Total RNA extraction and cDNA synthesis were carried out as indicated by the manufacturer using the Aurum Total RNA Mini kit and the iScript reverse transcriptase, respectively. Real-time PCR was performed in the CFX96 Touch Real-Time PCR Instrument using Sso Advanced Universal SYBR Green and primers (PrimePCR SYBR Assay) for TNF-α, Actin-β, or TLR4.

GARLIC STIMULATES LPS-INDUCED TNF-α FROM MACROPHAGES
The reaction was run at 95°C for 2 min for 1 cycle, 95°C for 5 s for 40 cycles, 60°C for 30 s for 40 cycles, and 65-95°C (0.5°C increments) for 5 s/step for 1 cycle. The data, expressed as 2 ÀΔΔCq , were obtained by normalizing against the actin gene and then comparing the treated cells against the untreated cells.
Statistical analysis. Statistical analysis was performed using GraphPad Prism Version 6.0. An unpaired oneway analysis of variance (ANOVA) was performed with a p < 0.05 considered significant. Post hoc testing was performed on data considered statistically significantly different in the ANOVA, including a Tukey's multiple comparison test to compare all groups against each other. p < 0.05 was considered statistically significant.

RESULTS
The aqueous garlic extract (G) was tested at different dilutions to investigate whether it had any effects on cytokine levels. Of the cytokines tested (TNF-α, IL-1β, IL-6, IL-10, and IL-12), G at dilutions of 1:200 to 1:1000 consistently increased TNF-α secretion from the macrophages (Table 1) compared with pyrogen-free water (PFW). Furthermore, G was very effective at stimulating LPS (0.1 μg/mL)-induced TNF-α (Table 1) from macrophages. Thus, we focused on the production of TNF-α. To rule out that the stimulation of the garlic extract on TNF-α secretion was due to endotoxin contamination, the garlic extract was sent out for testing. The levels were found to be 0.0025 endotoxin units/mg in 5 mg/mL of garlic extract. These endotoxin levels were negligible, and thus, we concluded that component(s) within the garlic extract was responsible for stimulating TNF-α production from the macrophages. We next looked at cell proliferation using the XTT assay to determine whether garlic had a mitogenic effect. We found that garlic did not stimulate cell proliferation, indicating that the increase in TNF-α was not because of an increase in cell number. However, garlic at a dilution of 1:40, in the absence of LPS, decreased metabolic activity by 10% as compared with untreated cells (data not shown). This suggests that at this dilution, garlic had a slightly toxic effect on the cells. This finding is similar to that by Shin et al. (2013) who also reported that at high raw garlic extract concentrations, RAW cell viability was reduced (Shin et al., 2013).
For subsequent studies, we used the highest dilution of garlic that caused the maximum stimulation of LPSinduced TNF-α secretion, which was seen using the 1:500 dilution (G1:500). We next varied the concentration of LPS to determine the LPS concentration to use in subsequent experiments. As seen in Fig. 1A, garlic stimulated LPS-induced TNF-α macrophage secretion at all LPS concentrations studied. Furthermore, although TNF-α secretion increased in a concentration-dependent manner   when the cells were treated with only LPS, in the presence of garlic (G1:500) and LPS, TNF-α secretion reached a plateau at 0.1 μg/mL LPS. Analysis of the TNF-α transcription levels after a 2-h G1:500 and 0.1 μg/mL LPS cell treatment also revealed a significant increase in TNF-α production (Table 2). Thus, we decided to continue to use 0.1 μg/mL LPS in our subsequent studies.
To further characterize the interaction between the garlic extract and LPS, the cells were treated with G1:500 for different periods before, after, or at the same time as LPS (0.1 μg/mL). G1:500 stimulated LPSinduced TNF-α at all time points studied, although the maximum stimulation by G1:500 occurred when the extract was added 15 min before to 1 h after LPS cell treatment (Fig. 1B). These results suggest the presence of a garlic component that has a maximum stimulatory effect when it is added around the same time as LPS. We next investigated whether preincubating the cells with garlic had an effect on LPS-induced TNF-α secretion. Thus, the cells were pretreated with G1:500 for different times; however, this time, garlic was removed prior to LPS addition. We found that only cells treated with G1:500 in conjunction with LPS showed TNF-α secretion ( Fig. 2A), suggesting that either a labile garlic compound was responsible for the stimulatory effect on LPS-induced TNF-α secretion or that garlic acted downstream from LPS. Thus, we investigated whether preincubating the cells with LPS had an effect on G1:500 stimulation of TNF-α. Hence, the cells were pretreated with LPS for different times, and then, the LPS was removed. The cells then received either the PFW control or G1:500. Cells receiving PFW after LPS removal showed that only cells pretreated with LPS for 0.5 to 1 h induced TNF-α secretion (Fig. 2B, set of white bars). However, these TNF-α levels were smaller than the levels observed when LPS remained with the cells for 24 h (Fig. 2B, last white bar). Interestingly, we found that addition of G1:500 after LPS removal further enhanced cellular TNF-α production when compared with the cells pretreated with LPS followed by PFW (Fig. 2B, set of black bars compared with white bars). This effect was seen at all LPS pretreatment times studied but was greatest after cells were preincubated with LPS for 0.5 to 1 h (Fig. 2B, set of black bars). These results suggest that garlic acts downstream from LPS to enhance LPSinduced TNF-α secretion or that a labile garlic component is responsible for garlic's effect. To confirm that the garlic extract acts downstream of the LPS-signaling cascade, we investigated the effect of garlic on TLR4 mRNA expression. The TLR4 is known to be the receptor for LPS (Parameswaran and Patial, 2010). As is shown in Table 2, garlic did not significantly alter TLR4 expression in the absence or presence of LPS.
We next investigated the length of time that garlic could stimulate LPS-induced TNF-α secretion. To this end, cells were treated with G1:500 + LPS or LPS alone, and cell supernatants were collected at different time points after LPS addition. We found, as before, that in the presence of garlic, LPS-induced TNF-α secretion was significantly greater in cells treated with LPS alone (Fig. 3). Furthermore, in cells treated with Cells were plated at a concentration of 1.25 × 10 5 cells/mL, 5 mL/well in complete media in six well tissue culture plates. The conditioned media was removed from the cells 24 h after plating them, and fresh media was added. The cells were then treated for 2 h with garlic (G) or synthetic allicin (A) at the indicated concentrations in the absence or presence or LPS (0.1 μg/mL). Total RNA was extracted, reverse transcribed, and the cDNA amplified to determine TNF-α and TLR4 mRNA levels. The data were normalized against the actin gene and are expressed as fold expression compared with untreated cells. *p < 0.02 compared with LPS 0.1 μg/mL. :500 + LPS, TNF-α levels began and peaked earlier than in cells treated with LPS alone. Thirdly, TNF-α secretion remained elevated longer in cells treated with G1:500 + LPS as compared with cells treated with LPS alone (Fig. 3). Thus, garlic enhances and prolongs the cellular response to LPS in terms of TNF-α production.
To determine if the stimulating component in the garlic was allicin, we quantified the allicin content in our garlic extract using HPLC. More than five peaks, because of different components in the garlic extract, were observed on the chromatogram. We identified the peak for allicin by its migration time (1.286 min) and by running it alongside the synthetic allicin. Upon integrating all peaks on the chromatogram, the peak area of allicin accounts for~24% of overall peak areas, indicating that allicin is one of the major components of our aqueous garlic extract. We calculated the allicin content in garlic to be 60 ± 14 μM. Using synthetic allicin, we found that allicin [2.5 (15 μM), 5 (31 μM), and 10 μg/mL (62 μM)] enhanced LPS-induced TNF-α secretion (Fig. 4). Furthermore, allicin (5 μg/mL) increased LPS-induced TNF-α mRNA levels as well (Table 2). These findings suggest that allicin, at least in part, is responsible for garlic's stimulation of TNF-α production from the macrophages.

DISCUSSION
Tumor necrosis factor-α is a pleiotropic cytokine that, depending on the cellular context, can regulate a number of cellular functions, including inflammation, proliferation, differentiation, and cell death or survival (Parameswaran and Patial, 2010;Varfolomeev and Ashkenazi, 2004). The major producers of TNF-α are macrophages, important players in the innate immune response (Parameswaran and Patial, 2010). In the present study, we demonstrate that an aqueous garlic extract and synthetic allicin, in the absence or presence of LPS, stimulate TNF-α secretion from the murine macrophage cell line J774A.1.
In our experiments, we have been unable to observe an effect of garlic on the secretion of other cytokines whether pro-inflammatory (i.e.  or antiinflammatory (i.e. IL-10). Others have conflicting reports on the effect of garlic or its derivatives on these cytokines. It was found that DADS stimulated the proinflammatory cytokines IL-1β and IL-6, and that AMS stimulated the antiinflammatory cytokine IL-10 (Chang et al., 2005;Hodge et al., 2002;Liu et al., 2009). It was also reported that DADS inhibited IL-10 (Chang et al., 2005). An aqueous garlic extract was also shown to inhibit IL-12 from monocytes (Hodge et al., 2002), whereas another group reported garlic lectins as stimulating IL-12 from monocytes (Dong et al., 2011). Others have reported garlic or its derivatives as inhibiting the pro-inflammatory cytokines IL-1β (Keiss et al., 2003) and IL-6 (Hodge et al., 2002).
We suggest that our results are different from other studies carried out in macrophages because of the nature and concentration of garlic compounds found in organic extracts compared with our aqueous garlic Figure 3. Garlic stimulates LPS-induced TNF-α secretion from J774A.1 macrophages more efficiently than LPS alone. Cells were plated as described in Fig. 1. Cells were treated with PFW or the garlic extract (G) and LPS (0.1 μg/mL). Cell supernatants were then collected at different times (from 3 to 49 h after LPS treatment) and assayed for TNF-α levels via ELISA. Statistical significance: *p < 0.01 compared with the corresponding LPS controls. This is a representative experiment of at least three independent experiments. Figure 4. Allicin stimulates LPS-induced TNF-α secretion from J774A.1 macrophages. Cells were plated as described in Fig. 1. Cells were treated with LPS (0.1 μg/mL) and the indicated concentrations of allicin. Cell supernatants were collected 24 h after LPS addition and assayed for TNF-α levels via ELISA. Statistical significance: *p < 0.02, **p < 0.006 compared with LPS-only treatment. This is a representative experiment of at least two independent experiments. extract. In addition, the effective garlic concentrations used by other investigators were higher than the stimulatory concentration that we used. This concentrationdependent effect of garlic is illustrated in a work carried out by Liu et al. (2009) who found that rats fed high concentrations (200 mg/kg) of garlic oil reduced concanavalin A-induced IFN-γ production in isolated rat lymphocytes, whereas consumption of lower concentrations (50 mg/mL) of the garlic oil enhanced the IFN-γ to IL-4 ratio (Liu et al., 2009). We found that our aqueous garlic extract altered LPS-induced TNF-α production in a concentration-dependent manner. Garlic at 1:200 and 1:500 was more potent at stimulating LPS-induced TNFα than at higher dilutions (1:1000) of the extract. When we used lower dilutions of our garlic preparation (1:40 dilution), we saw a decrease in TNF-α secretion, perhaps because of the cytotoxic effects generated by the garlic. These findings suggest that the concentration of garlic components and extraction methods utilized by different investigators play a significant role in the experimental outcome. In garlic oil extracts, the main components include the oil soluble polysulphides DADS, DAS, and DATS (Brodnitz et al., 1971). On the other hand, water-soluble garlic compounds include proteins and sulphur-containing compounds such as S-allylcysteine and S-allylmercaptocysteine (Rabinkov et al., 1995).
Ours was an aqueous garlic extract in which the content of allicin (60 ± 14 μM), the major component of crushed garlic (Rybak et al., 2004), was about 24%, similar to what others have reported (Wang et al., 2010). Our results using the synthetic allicin are in agreement to those of Kang et al. (2001) who reported that allicin stimulated TNF-α from resident murine peritoneal macrophages (Kang et al., 2001), suggesting that allicin is at least partially responsible for the effects that we observe.
It is interesting that garlic alone was able to induce TNF-α from the macrophages. TNF-α is known to be produced from these cells upon ligation of TLRs (Parameswaran and Patial, 2010). The LPS is known to act via TLR4 to induce TNF-α secretion. Activation of TLR4 via LPS triggers, amongst others, the mitogen-activated protein kinase (MAPK) pathway leading to TNF-α production (Parameswaran and Patial, 2010). It is not known whether garlic compounds are able to bind TLRs or an unknown receptor that then interacts with TLR. To investigate whether garlic (or allicin) altered TLR4 expression in the J774A.1 macrophages, we determined TLR4 mRNA levels and found that neither garlic nor allicin had an effect on TLR4 transcriptional levels. This finding suggests that neither garlic nor allicin affect TNF-α production by altering LPS function at the level of its receptor. This finding also suggests that garlic and allicin act downstream from the LPS-signaling cascade. We are currently investigating the role of MAPK in the induction of TNF-α by the aqueous garlic extract.
Garlic has been recognized to have immunostimulatory effects. Although garlic has been largely recognized as having antiinflammatory response, the induction of TNF-α by garlic does not contradict the notion that this herb helps immunity. The increase in TNF-α is necessary when an organism is injured or infected. We suggest that garlic renders macrophages more efficient at producing this cytokine in response to infections as evidenced by a more rapid increase in TNF-α and prolongation of this response. As is true to a healthy pro-inflammatory response, the pro-inflammatory cytokines cannot remain elevated or they can be fatal to the organism. Thus, we find that although garlic does increase LPS-induced TNF-α, this elevation is temporary as evidenced by the subsequent decrease in the levels of this cytokine. Therefore, we believe that our findings support the notion that garlic is beneficial to the immune system.