Modulation of bacterial ghosts – induced nitric oxide production in macrophages by bacterial ghost-delivered resveratrol

Authors


Correspondence

P. Kudela, BIRD-C GmbH&CoKG, Hauptstrasse 88, A-3420 Kritzendorf, Austria

Fax: +43 2243 28514

Tel: +43 1 890 4350

E-mail: pavol.kudela@bird-c.at

Abstract

The present study aimed to investigate the capacity of resveratrol (RV) delivered into macrophages by bacterial ghosts (BGs), representing intact empty nonliving envelopes of Gram-negative bacteria, to modulate nitric oxide (NO) production related to the presence of the pathogen-associated molecular patterns on the surface of BGs. Incubation of the murine macrophage cell line RAW 264.7 with BGs leads to a dose-dependent activation of inducible NO synthase. To modify BG-induced NO formation in RAW 264.7 cells by RV; BGs were loaded with RV (RV-BGs) and incubated with murine macrophages in a dose-dependent manner. RV-BGs delivering RV to the target macrophages significantly reduced BG-induced NO production with concentration of RV more than one order of magnitude lower than the amount of RV capable of reducing NO formation when applied directly. Moreover, no cytotoxic impact of BGs on the viability of RAW 264.7 cells added to macrophages alone or loaded with RV was detected after a mutual 24 h incubation, whereas cell viability slightly decreased (~ 10%) when RV concentrations of 30 μm alone were applied. The results obtained in the present study clearly indicate that the intracellular delivery of RV by BGs significantly enhances the total RV effect.

Abbreviations
Ag

antigen

APC

antigen-presenting cell

BG

bacterial ghosts

DC

dendritic cell

FITC

fluorescein isothiocyanate

IMC

immature myeloid cell

iNOS

inducible nitric oxide synthase

LPS

lipopolysaccharide

MDSC

myeloid-derived suppressor cell

NO

nitric oxide

NOS

nitric oxide synthase

RV

resveratrol

Introduction

The cells and soluble factors of the innate immune system are able to rapidly react on infectious agents using killing mechanisms in a nonspecific manner or by their immunological inactivation. Phagocytosis of bacteria, viruses or protozoa by macrophages, monocytes, neutrophils and dendritic cells (DCs) results in the release of antimicrobial peptides, hydrolytic enzymes and/or reactive oxygen species or reactive nitrogen species, including the production of nitric oxide (NO) through the inducible nitric oxide synthase (iNOS), as well as stimulation of antigen (Ag)-specific immune responses. However, generation of iNOS-dependent NO is a feature not only of phagocytes, but also of epithelial cells or keratinocytes [1-4].

Nitric oxide is a versatile molecule exhibiting an ambiguous endogenous role with a rapid half-life and acts as a crucial mediator molecule for distinct cellular functions. This highly reactive molecule is implicated in the pathophysiology of many diseases, including cancer, and is capable of causing cytotoxic and mutagenic effects when produced in excessive amounts (e.g. under oxidative burst conditions) [5, 6]. NO also plays an important role in the immune system as a result of its various effector and immunoregulatory functions, including antimicrobial, antitumourigenic and apoptotic activity, or its capacity to modulate the function of cytokines, suppressor cells and T cell differentiation [7, 8]. iNOS (or NOS2), which accounts for NO production in macrophages, myeloid-derived suppressor cells (MDSCs) and keratinocytes, oxidizes one molecule of l-arginine at a guanidine nitrogen to an intermediate that is oxidized to yield one molecule of NO and l-citrulline [2, 9]. Bacterial lipopolysaccharide (LPS) represents one of the most important stimuli for iNOS induction after binding to Toll-like receptor 4 [10].

Bacterial ghosts (BGs) are empty cell envelopes of Gram-negative bacteria devoid of cytoplasmic content and free of nucleic acids. They are produced by the controlled expression of plasmid-encoded lysis gene E. Protein E leads to the fusion of inner and outer membranes and the formation of a tunnel structure through which the cytoplasmic content is expelled as a result of the osmotic pressure difference between the cytoplasm and the exterior of the bacteria [11, 12]. Scanning and transmission electron micrographs taken from many different BGs show that the rod-shaped morphology is not altered and that the E-specific lysis hole is a distinct opening with a diameter varying in the range 40–200 nm [12]. Although initiation of the formation of an E-specific transmembrane tunnel structure can take place at several locations within potential division sites of growing bacteria, the first complete tunnel wins the race and the cytoplasmic content is expelled within < 1 s per single bacterium [11, 13]. In rare events, two E-specific tunnels can be determined per bacterium where the opening of the holes has occurred simultaneously. Because the rigid peptidoglycan layer is not affected by the E-lysis procedure, BGs stay as a hollow particle, similar to an open empty bottle, during and after the freeze-drying process that is normally applied at the end of production cycle [14]. These nonliving, empty bacterial envelopes maintain the full cellular morphology of the native bacteria with all the cell surface structures intact, including the outer membrane proteins, adhesins, LPS and peptidoglycans, making BGs attractive natural adjuvants for vaccination [15]. Because BGs share the antigenic determinants from their living counterpart and additional foreign proteins can be expressed on or within the cell envelope, and because DNA or drugs also can be loaded inside the cytoplasmic lumen of the ghosts, this emphasizes their role as a novel vaccination system for animals and humans. A more detailed discussion of BG platform technology and the properties of BGs is provided in recent reviews [16-18]. Incubation of mouse and human professional antigen-presenting cells (APCs) (e.g. macrophages and DCs) with BGs efficiently stimulates their maturation, and enhances the expression of Ag-presentation and co-stimulatory molecules [19-22]. Moreover, Ag delivery by BGs into APCs leads to efficient Ag presentation and stimulation of the Ag-specific immune response [23].

Resveratrol (RV) is a polyphenolic stilbene compound naturally occurring either in trans- or cis-isomeric forms in various plant species, and is especially present in the skins of grapes, peanuts and berries. RV shows a broad spectrum of immunomodulating activities, possessing anticancer, antioxidant and cardioprotective properties [24-27]. Furthermore, RV has great therapeutic potential, especially in the treatment of a variety of human and animal infectious diseases [28-30]. Moreover, RV possesses the capacity to suppress iNOS expression and subsequently NO formation by inhibition of nuclear factor-kappa B activation stimulated by LPS [31, 32].

Previous studies showed that BGs promote innate and adaptive immunity through the secretion of pro-inflammatory cytokines and the expression of antimicrobial peptides by effector cells, as well as by the stimulation of effective humoral and cellular immune responses [17]. Taking into consideration that iNOS-dependent production of NO (i.e. a key-player molecule in innate immunity and within the tumour microenvironment) is stimulated by LPS, the present study aimed to examine the extent of BG-stimulated NO release in the model murine macrophage cell line RAW 264.7, as well as investigate the impact of RV loaded within the BGs on NO production induced by BGs after its delivery into the cytosol of target myeloid-derived Ag-presenting cells.

Results

BGs can be successfully loaded with RV

The loading of lyophilized Escherichia coli NM522 BGs with RV was performed by simple resuspension of BGs within RV solutions of different concentrations (1–35 mg·mL−1 RV). To determine the stability and the amount of RV associated with the BGs, four independent ethanol extractions were performed and analyzed separately via HPLC. No RV-related metabolic processes occurred within the BG envelope after loading BG envelopes with RV solution followed by extensive washing steps. HPLC analysis confirmed only the presence of the pure compound and no other RV metabolites. These results clearly indicate the stability of RV entrapped within the BG lumen and also that the substance is efficiently protected from UV damage during handling and storage because no cis-RV was detected (Fig. 1A).

Figure 1.

Quantification of RV loaded within BGs. Correlation of RV loading solution and RV recovered from loaded BGs and correlation of the applied amounts of BGs and BGs associated (attached or internalized) with murine macrophages. HPLC chromatogram of the RV loaded inside BG envelopes confirmed the presence of pure compound, indicating that RV entrapped within the BGs remains stable with no RV-related metabolic processes detected. One representative experiment out of four is depicted (A). The correlation between the RV concentration (mg·mL−1) used for the loading of BGs and the amount of RV (μg/1 × 1010 BGs) recovered from loaded E. coli NM522 BGs after ethanol extraction and HPLC analysis is indicated by linear regression and its associated r2 of 0.9933 (B). The correlation between the numbers of BGs applied to the murine macrophages for a short period of mutual incubation (20 min) and the number of BGs associated with target cells (attached or internalized) is indicated by linear regression and its associated r2 of 0.9980 (C). Each point represents the mean of four independent experiments performed in quadruplicate.

A clear correlation between the loading concentration of RV and recovered RV was observed (correlation coefficient r2 = 0.9933) (Fig. 1B). The highest loading efficiency was obtained after resuspension of lyophilized BGs in solution with the highest RV concentration (35 mg·mL−1). The results acquired during HPLC analysis showed that the amount of RV loaded within BG envelopes after incubation of BGs in solution with the highest RV concentration corresponds to 46 μg RV/1 × 1010 BGs (4.6 fg RV/BG).

Murine macrophages efficiently recognize and internalize fluorescently-labelled BGs

The capacity of RAW 264.7 cells to recognize and internalize BGs was determined using a quantitative fluorometric assay. The amount of BGs attached and phagocytosed by murine macrophages was measured immediately after short period (20 min) of mutual incubation (Fig. 1C). The results obtained showed that the number of BGs associated with target cells (attached or internalized) clearly depends on the ratio of BGs per cell (correlation coefficient r2 = 0.9980). Furthermore, no differences were detected when the experiments were performed either in 96- or 24-well plates (data not shown). The results obtained showed that a mean of 25% of the applied BGs was internalized or attached to the cell surface of macrophages already after the short period of incubation. In para-llel, confocal laser scanning microscopy studies were conducted to confirm the phagocytosis of BGs by murine macrophages within this short incubation period and to discriminate between attached and internalized BGs. Accordingly, z-stacks were performed after incubation of fluorescein isothiocyanate (FITC)-labelled E. coli NM522 BGs with target cells after 20 and 40 min. Macrophages were stained with Texas-Red Phalloidin to visualize F-actin cytoskeleton. FITC-labelled BGs that were efficiently phagocytosed by Texas-Red Phalloidin stained RAW 264.7 cells (red) appear yellow (~ 65%), whereas non-internalized but attached BGs appear green (less than 35%) (Fig. 2A). Treatment of RAW 264.7 cells pre-incubated with FITC-BGs for both 20 and 40 min with trypsin led to the successful removal of BGs attached to the cell surface (Fig. 2B).

Figure 2.

Efficient internalization of fluorescently-labelled BGs by murine RAW 264.7 macrophages. The cells were incubated for 20 or 40 min with FITC-labelled E. coli NM522 BGs (103 per cell), extensively washed, fixed and stained with Texas Red-X Phalloidin, followed by examination using confocal laser scanning microscopy. Fluorescently-labelled BGs located within the murine macrophages are depicted in yellow, which represents direct overlays of green and red fluorescent structures; non-internalized BGs appear green. Images display a representative single z-stack of various optical sections. Images were taken with a ×20 objective after 20 min (A) or with a ×63 oil objective after 40 min of incubation (B). One representative experiment for each setting is depicted.

BGs stimulate the production of NO by murine macrophages in a dose-dependent manner

Treatment of RAW 264.7 cells with LPS for 24 h led to NO release, with significant NO production already starting at a concentration of 1 ng·mL−1 (< 0.0001) (Fig. 3A). No significant difference in NO production by macrophages was detected in response to 1, 5, 10 or 100 ng·mL−1 of purified LPS, whereas 1000 ng·mL−1 LPS led to a more than three-fold increase of accumulated nitrite. Intact LPS molecules present on the surface of BGs are unlike the free LPS used for stimulation of NO production tightly integrated into the outer membrane. Single nonliving BG comprises the same amount of LPS as its single living counterpart [33]. Therefore, we tested the capacity and the efficacy of BG-bound LPS to stimulate NO production after incubation of BGs with murine macrophages. The results obtained showed that BGs surface-bound LPS induced a signal in macrophages at a ratio of BG : macrophage of 10 : 1, which led to a significant stimulation of NO formation (= 0.0017) comparable to the level detected after incubation of target cells with 1 ng·mL−1 of free LPS. Based on the results from previous studies, the amount of cell-bound LPS delivered by BGs to macrophages at the ratio of 10 BGs per cell corresponds to an LPS concentration of ~ 615 ng·mL−1 [33-35]. The data confirm a more than one order of magnitude higher activity of free LPS than cell-bound LPS, as reported previously [33, 36]. Moreover, a clear BG particle-dependent induction of NO formation was detected up to a multiplicity of infection of 500. Additional increase of BG particles per single cell (up to 1 × 104) had no significant impact on the synthesis and production of NO by RAW 264.7 cells (Fig. 3B).

Figure 3.

BGs induce NO formation in RAW 264.7 macrophages in a dose-dependent manner. Incubation of RAW 264.7 cells with various concentrations of LPS for 24 h led to a dose-dependent formation of NO, with significant production starting already at 1 ng·mL−1 LPS (P < 0.0001) (A). Intact LPS present on the surface of BGs significantly stimulates NO formation at the ratio of BG : cell of 10 : 1 (= 0.0017) to a similar extent as that seen for macrophages incubated with 1 ng·mL−1 LPS. Particle (dose)-dependent production of NO by macrophages after co-culture with BGs was detected up to a multiplicity of infection rate of 500. A further increase in the amount of BG added to the cell culture system had no significant impact on the synthesis and formation of NO (B). NO production was determined by measurement of the nitrite concentration using the Griess assay, as described in the 'Materials and methods'. Each bar represents the mean ± SD of four independent experiments performed in triplicate. P < 0.05 was considered statistically significant (*< 0.05; **< 0.01; ***< 0.001). ns, not significant.

RV delivered within BGs into the cytoplasm of murine macrophages modulates their capacity to produce NO after stimulation with intact surface structures of BGs

Resveratrol externally added to the culture system containing murine macrophages and BGs had no significant impact on NO formation when applied at low concentrations (0.3 and 3 μm). However, 30 μm externally added RV was able to significantly decrease the production of NO (= 0078) by target macrophages stimulated with BGs (Fig. 4A). The data showed that BGs are able to stimulate NO production in a dose-dependent manner when applied to RAW 264.7 cells (Fig. 3B). Furthermore, it was shown previously that RV is capable of reducing LPS-induced NO production [32, 37, 38]. Therefore, the capacity of RV loaded within BGs and delivered directly into the cytoplasm of the target cells to modulate BG-induced NO formation was examined. The results demonstrate that murine macrophages reach the plateau of NO production after incubation with BGs at a ratio of BG : macrophage of 500 : 1 (Fig. 3B). Based on these observations, the RAW 264.7 cells were stimulated with this boundary amount of BGs (500 per cell) or with a reduced concentration of BGs (100 per cell) that is still capable of inducing significant NO formation and the impact of RV loaded inside these BGs corresponding to 15.115 μm and 3.023 μm RV, respectively, on NO production was detected. NO formation by murine macrophages after incubation with RV-loaded BGs was significantly reduced after incubation with a lower particle number of RV-BGs (100 BGs per cell; 3.023 μm RV within BGs added to the cells; = 0.0006) compared to supplementing the macrophage cell culture with empty BGs (100 per cell) and extra-added pure RV (3 μm) (Fig. 4B). Similar results were detected when a boundary amount of RV loaded BGs was used (500 BGs per cell; 15.115 μm RV within BGs added to the cells; = 0.0095) (Fig. 4C).

Figure 4.

Internalization of RV-loaded BGs by RAW 264.7 macrophages allows intracellularly delivered RV to modulate NO production induced by intact surface structures of BGs. Low doses of pure RV added to murine macrophages pre-incubated with BGs (500 per cell) had no impact on NO formation by stimulated cells. The only significant decrease of BG-induced NO production by RAW 264.7 cells (= 0.0078) was detected after using a high dose of RV (30 μm) (A). RV delivered within BGs into the cytoplasm of murine macrophages efficiently reduces their capacity to produce NO in response to intact surface structures of BGs after incubation with various doses of BGs (B, C). NO production was determined by measurement of the nitrite concentration using the Griess assay, as described in the Materials and methods. Each bar represents the mean ± SD of four independent experiments performed in triplicate. P < 0.05 was considered statistically significant. ns, not significant.

BGs have no cytotoxic effect on murine macrophage cell viability

The influence of empty BGs, RV loaded BGs and pure RV alone on the viability of murine macrophages was investigated in a dose-dependent manner after 24 h of mutual incubation. The data showed no toxic impact of empty BGs in all concentrations analyzed from ten BGs per cell to 1000 BGs per cell and no significant difference in the viability of cells incubated with BGs compared to cells incubated without BGs. A significantly decreased viability of RAW 264.7 cells was detected only in the presence of the cytotoxic agent Triton X-100, which was used as positive control (Fig. 5A). Similar results were detected after a mutual short period (20 min) of incubation of cells with BGs followed by an additional 20 h of incubation in the absence of tested stimuli (data not shown). Moreover, the data obtained showed no cytotoxic impact of different amounts of RV-loaded BGs and pure RV up to a concentration of 15 μm on the viability of murine macrophages added directly to the culture system followed by an extended 24 h of incubation (Fig. 5B). However, a high dose of pure RV (30 μm) caused a significant decrease of RAW 264.7 cell viability (< 0.0001) (Fig. 5B).

Figure 5.

Neither empty BGs, nor RV-loaded BGs have a negative impact on the viability of RAW 264.7 cells. The viability of murine macrophages was determined after 24 h of co-culture with tested stimuli using the colorimetric Neutral Red uptake assay. No cytotoxic impact of empty BGs incubated with RAW 264.7 cells was detected in all analyzed concentrations from 10 BGs per cell to 1000 BGs per cell (A). Similarly, loading of RV into the BGs and their follow-up incubation with murine macrophages using various BGs concentrations did not change the viability of cells (B). Incubation of RAW 264.7 cells with 15 μm pure RV did not affect the viability of murine macrophages, whereas 30 μm RV significantly reduced cell viability (< 0.0001) (B). A substantial decrease of RAW 264.7 cell viability was also detected after incubation of cells in the presence of the cytotoxic agent Triton X-100 serving as a positive control (A). Each bar represents the mean ± SD of four independent experiments performed in triplicate. P < 0.05 was considered statistically significant. ns, not significant.

Discussion

The present study demonstrates that NO formation induced after the internalization of BGs by macrophages can be effectively modulated by RV loaded inside BGs and delivered directly to the cytosol of target cells. Low doses of RV presented within the BG envelope significantly decreased the generation of NO without having a cytotoxic impact on the viability of phagocytic cells. Moreover, intracellular delivery of RV by BGs significantly enhanced the total RV effect, which might imply the existence of an unknown intracellular and/or nuclear RV receptor capable of amplifying the biological signal provided by RV (Fig.  6).

Figure 6.

Model modulation of NO production in macrophages by resveratrol delivered by BGs directly into the cytosol. RV-BGs (A) or empty BGs (B) are recognized by pattern recognition receptors on the surface of macrophages (MΦ), which leads to their activation and the internalization of RV-BGs or BGs into early endosomes (E). Endosomes are fusing with lysosome (L) forming a hybrid organelle called an endosome–lysosome fusion (ELF). The presence of intact PAMPs on the surface of BGs leads to activation of iNOS and results in the production and release of NO. RV has the capacity to modulate NO production, although it has to be transferred into the cytosol of cells to be able to execute its regulatory activities. RV is rapidly metabolized into several distinct sulfate or glucuronide metabolites (mRV), although it is still unclear whether RV accomplishes its biological functions directly as a native non-modified substance or through its metabolites. Intracellular delivery of RV by BGs and its release from ELF as an intact molecule and/or mRV significantly enhances the total RV effect, which might indicate the existence of an unknown intracellular and/or nuclear endogenous RV receptor that is capable of amplifying the biological signal provided by RV.

Intact surface morphological, structural and antigenic components of BGs actively participate in the efficient recognition and internalization of BGs by professional APCs, including DCs and macrophages, as well as various types of tumour cells [16, 17]. Toll-like receptor signalling in macrophages is required for anti-pathogen responses, including the biosynthesis of NO radicals, which is extremely beneficial for curing pathogen-induced diseases [5, 39-41]. However, increased levels of iNOS expressed by tumour-associated macrophages and MDSCs within the tumour microenvironment promote immune suppressive activities [42]. MDSCs represent a heterogeneous cell population containing immature myeloid cells (IMCs) and precursors of monocytes, macrophages and DCs at early differentiation stages, and are considered as a critical cell population responsible for tumour evasion from immunosurveilance [43]. The presence of suppressor cells within the tumour microenvironment is associated with a bad prognosis and increased mortality [44]. Down-regulation of ARG1 and iNOS significantly reduces the suppressive potential of MDSCs, and enhances the effectiveness of the anti-tumour immune response [45, 46]. Modulation of the suppressive behaviour of MDSCs via inhibition of NO production by specific drug delivery targeted directly to the tumour microenvironment using BGs might lead to increased tumour infiltration by effector Ag-specific T cells. Moreover, natural adjuvant properties of BGs could positively affect the maturation of IMCs within the tumour microenvironment and hence improve the anti-tumour immune response. The data obtained in the present study indicate that RV delivered in the cytosol of macrophages is able to significantly decrease the levels of NO produced and therefore possibly affect NO production by other types of myeloid-derived cells present within the tumour microenvironment. However, additional detailed studies need to be performed aiming to determine whether RV delivered by BGs into the MDSCs (in addition to modulation of NO formation after activation of IMCs by intact BG surface structures) could also inhibit ARG1 and thus have a positive impact on the maturation of IMCs.

Resveratrol belongs to the group of antioxidants capable of killing rapidly dividing cells without resulting in side-effects such as chromosomal alterations and mutagenesis. Moreover, RV possesses antioxidative and chemoprotective properties against inflammatory processes related to cancer and vascular diseases, as well as viral and bacterial infections [24-27, 47]. In addition to its antitumour activity, RV has the capacity to modulate multiple cell-signalling pathways, and hence effectively sensitize tumour cells to the chemotherapeutics and increase the impact of a chemotherapeutic drug, even for chemoresistant tumour cells [25, 26, 48-51]. Novel delivery systems for targeted drug delivery should not only be safe, noncytotoxic and nongenotoxic, but also capable of highly efficient drug and/or gene medicine delivery to target cell populations. Furthermore, regarding tumour therapy, these systems should be able to modulate the function of suppressor cells within the tumour microenvironment and enhance the vulnerability of tumour cells to both chemotherapeutics and effector immune cells of the immune system.

The fluorescent reporter substance calcein was used to fill freeze-dryed BGs and inside-out membrane vesicles and seal the hole. The present study showed that the E-lysis tunnel stays open and that the carrier BGs could be filled with calcein. Calcein was released within eukaryotic cells after uptake of the loaded BGs and their degradation in the endo-lysosomal compartments, as detected by fluorescence of the cells by dilution of the self-quenched calcein within the cytoplasm [52]. Scanned electron micrographs showed the rod-shaped BG containers and their morphological identity with the living bacteria that they were derived from [52]. Because RV is largely hydrophobic by nature, it can be assumed that the aromatic rings of the RV molecules and the saturated carbon bridge in between them interact with the lipid environment of the inner and outer membranes of BGs. If the substance is smaller than 500 Da, it should be assumed that the drug can also diffuse through the pores of the outer membrane. The major entrance for the drug, however, is the opening of the E-specific tunnel within the envelope complex and, as the solution is filling the inner lumen of BGs, it is absorbed by the phospholipids of the inner membrane. Extensive washing steps of BGs to remove unabsorbed RV were carried out until free RV was no longer detected in the supernatant. The free space inside BGs is ~ 0.3–1.7 × 10−9 mL, and so the effect of residual drug during BG uptake time (< 10 min) through the E-lysis tunnel can be neglected because the diffusion rate through E-tunnel is smaller than of the free substance through dialysis tubing. Only 12% of BG-associated doxorubicin was released within 10 h compared to 100% of free doxorubicin released from dialysis tubing [53]. Moreover, because RV is more hydrophobic than doxorubicin, any consideration of the important effect of free RV released from BGs can be neglected.

Resveratrol has to be transferred into the cytosol of cells to be able to execute its regulatory activities. To the best of our knowledge, the exact mechanism of RV cellular uptake is still unclear. To date, no specific receptor has been found and defined for RV, except integrin ανβ3, which partially contributes to RV uptake and the activation of the apoptosis signalling pathway [54]. RV is rapidly metabolized into several distinct sulfate or glucuronide metabolites [55]. It is still unclear whether RV accomplishes biological functions directly as a native non-modified substance or through its metabolites. Passive transport of RV to the cytosol through phospholipid bilayers or carrier-mediated transfer using a human hepatic model can be limited by RV interaction and coupling with serum proteins that significantly decrease its transport [30]. Moreover, passive transport of RV might enhance its metabolism and decrease the total effect of native RV on specific signalling pathways.

By contrast, direct transport and release of RV within the cytosol could effectively potentiate the biological effects of RV. Because the rates of BGs internalization showed that a mean of 25% of the BGs applied were successfully endocytosed or attached to the cell surface of macrophages, the results of the present study confirm the exceptional properties of BGs with respect to their use as a targeted delivery vehicle. We demonstrated that even a lower amount of RV loaded within BGs was capable of inducing the same effect as that observed after incubation of target cells with pure RV. The RV concentration of 30 μm can be reduced by 40- and eight-fold when RV is associated and loaded inside BGs and used at a ratio of 100 and 500 RV-BGs per cell, respectively. Questions still remain about the proper dosing of RV for clinical application because RV has the capacity to protect cells as well as inhibit proliferation and induce apoptosis [30, 55, 56]. The data obtained in the present study indicate that the direct transfer of RV using suitable carriers can significantly decrease the dose of RV required to reach the desired effects. However, it is very important to take the application route into consideration, as well as the properties of the carrier used for RV delivery to target cells. Extensive studies need to be performed to determine the effective dose of RV delivered using specific carriers to mediate cell protection and induce growth inhibition or elicit apoptosis, as well as for disease prevention in accordance with previously reported studies before translation to the clinic [31, 48, 50, 51, 56-59].

The uptake of RV-loaded BGs can lead to an increased RV concentration inside the cell and hence to improved detection of the RV-mediated effect. However, the release of RV delivered by BGs within the cytosol from endosome–lysosome fusion after carrier (BGs) and cargo (RV) degradation might also imply the existence of a novel intracellular RV-specific receptor responsible for the detected RV-mediated effect. The present study provides no evidence to support or refute either the presumed RV-specific intracellular receptor theory, nor simply the delivery of a higher RV concentration by BGs into the cytosol. However, after taking into consideration the complex intracellular degradation process occurring after internalization of RV-loaded BGs compared to transport of RV to the cytosol by passive diffusion via lipid rafts or by binding to surface receptors (e.g. integrin αvβ3) [60], we hypothesize the existence and presence of a novel internal RV-specific receptor within the cytoplasm and/or nuclear area of macrophages. Based on our results, we propose the term ‘endogenous resveratrol receptor’ for such a hypothetical receptor (Fig.  6). However, this concept has to be investigated using both in vitro and in vivo models.

The results reported in the present study show the potential of BGs to be used as a carrier for the delivery of natural antioxidants to myeloid-derived cells present within the tumour microenvironment, as well as the capacity to modulate their suppressive functions. Moreover, the intact surface structures of BGs and their natural adjuvant properties, together with targeted antioxidant delivery, might stimulate the maturation of IMCs, increase the chemosensitivity of tumours and help induce the immune response against tumour cells.

Materials and methods

Cell culture and reagents

The murine macrophage cell line RAW 264.7 obtained from the American Type Culture Collection (TIB-71; ATCC, Rockville, MD, USA) was maintained in DMEM with 4.5 g·L−1 glucose (Lonza, Verviers, Belgium) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA), 2 mm l-glutamine (Lonza), 100 U·mL−1 penicillin (Invitrogen) and 100 μg·mL−1 streptomycin (Invitrogen) in a 5% CO2 humidified incubator at 37 °C. The medium was changed every 2–3 days and the cells were subcultured when they reached the confluent state. All experiments were performed in a serum-free complete culture medium. If not otherwise stated, all chemicals were obtained from Sigma Aldrich (Sigma Chemical Co., St Louis, MO, USA).

BG production

Bacterial ghosts from E. coli NM522 were produced by the controlled expression of the phage-derived lysis protein E as described previously [12]. Inactivation of the nonlysed bacteria was performed by the addition of gentamycin (50 μg·mL−1; Invitrogen) and streptomycin (100 μg·mL−1; Invitrogen). Subsequently, BGs were washed three times with NaCl/Pi (pH 7.4), resuspended in distilled water and lyophilized under sterile conditions. Lyophilized BGs were stored at room temperature. BGs were resuspended in serum-free complete culture medium before the treatment experiments.

Loading of BGs with RV

Lyophilized BGs (12–24 mg) were resuspended in different solutions of RV (1–35 mg·mL−1) dissolved in methanol and incubated under vigorous shaking (800 r.p.m.) for 30 min at 28 °C. Subsequently, the loaded BGs were collected by centrifugation at 11 300 g for 15 min and the pellets were washed five times with distilled water. BGs used as a negative control for all of the experiments were re-suspended in methanol without RV followed by the same incubation and washing protocol as that applied for the generation of BGs loaded with RV. Determination of non-metabolized drug loaded within the BGs after the extensive washing procedure was performed using HPLC analysis as described below. The loaded BGs and the BGs used as a negative control were then aliquoted and stored at –20 °C until further use.

Quantification of RV extracted from BGs

Resveratrol-loaded BGs (1 mg) were resuspended in 500 μL of 96% ethanol (Brenntag CEE GmbH, Vienna, Austria) followed by 5 min of ultrasonification. Subsequently, the ethanolic extract was diluted equally with distilled H2O (1 : 1) and immediately centrifuged at 11 300 g for 15 min at 4 °C. HPLC analysis was performed using a PE 200 Series HPLC System (Perkin Elmer Inc., San Jose, CA, USA) equipped with a Lichrospher 100 RP-18e column (5 μm; 250 × 4 mm) (Merck, Darmstadt, Germany) for separation and turbochrom navigator software (Applera Corporation, Norwalk, CT, USA) for control of the instrument and data evaluation. The quantification was carried out using the peak area method applying RV as an external standard. A gradient method was performed using aqueous acetic acid (pH 2.8) as solvent A and methanol (gradient grade: HiPerSolv for HPLC; VWR International, Vienna, Austria) as solvent B at a flow rate of 1 mL·min−1 at 25 °C. The gradient profile started from 50% B to 60% B within 10 min and a final purge with 95% B. Ten microlitres of all samples and dilutions used were injected, and the chromatograms were monitored at a detection wavelength of 305 nm. The relevant RV peak was identified by comparing the retention time (5.3 min) and the UV of the samples with those of the external RV standard.

Application and detection of FITC-labelled BGs

The labelling of BGs with fluorescent marker and the investigation of the efficiency of endocytic activity of RAW 264.7 cells was performed as described previously with minor modifications [21, 61]. Briefly, the cells were plated in flat-bottom 96-well plates (Sarstedt, Nümbrecht, Germany) 48 h before incubation with FITC-labelled BGs. After 20 min of incubation of the macrophages (~ 3 × 105 per well) with a defined number of empty fluorescently-labelled BGs, the cells were extensively washed twice with NaCl/Pi to remove excess BGs and the washing solutions were collected in empty neighbouring wells. Increased values of fluorescence related to the internalization of FITC-BGs were recorded with a Tecan GENios Pro plate reader (Tecan Group Ltd, Männedorf, Switzerland) at excitation and emission wavelengths of 485/535 nm (gain 40). The uptake was calculated from the measured fluorescence levels per macrophage, taking the fluorescence values of the applied FITC-BGs as 100%. Additionally, confocal laser scanning microscopy was used to verify internalization of BGs by murine macrophages. The cells were seeded in eight-well chamber slides (BD Falcon; BD Biosciences, Pharmingen, San Jose, CA, USA) and allowed to attach to the plastic surface. Next, the medium was replaced by 200 μL of serum-free complete culture medium containing the FITC-labelled BGs (1 × 103 per cell) followed by incubation for either 20 or 40 min at 37°C. Thereafter, the cells were washed twice with NaCl/Pi and fixed in 1% paraformaldehyde in NaCl/Pi (20 min at room temperature). After incubation with FITC-BGs, the cells were washed three times with NaCl/Pi to remove excess BGs. Finally, the cells were trypsinized, washed twice with NaCl/Pi and fixed onto glass slides by cytospinning (800 r.p.m. for 15 min) and paraformaldehyde as described above. Next, the cells were washed again twice with NaCl/Pi and permeabilized with 0.5% Triton X-100 for 20 min, followed by two additional washing procedures. Finally, 100 μL of freshly prepared PromoFluor-590 (Texas-Red)-conjugated Phalloidin (5 μL of methanolic 100 U·mL−1 dye stock; NaCl/Pi; 1% BSA; PromoCell GmbH, Heidelberg, Germany) was added to visualize cell boundaries. After 45 min of incubation, the slides were washed again twice with NaCl/Pi and imaged by confocal laser scanning microscopy.

Nitrite assay

RAW 264.7 cells were seeded in flat-bottom 96-well plates, 48 h before each experiment. Approximately 3 × 105 cells per well were then stimulated either with 200 μL of a defined amount of: (a) empty BGs; (b) RV-loaded BGs (4.6 fg RV/BG); or (c) empty BGs and defined RV concentrations mixed together just before addition to the culture system. After 20 min of incubation at 37 °C, the cells were washed twice with NaCl/Pi to remove non-attached and non-internalized BGs and RV. Subsequently, the cells were incubated for additional 20 h in the dark in a 5% CO2 humidified incubator at 37 °C. In parallel, and using the same culture conditions without BGs and RV, the stimulatory effect of various doses of pure LPS (Fluka; Sigma Chemical Co.; E. coli serotype 055:B5) was investigated after 24 h of treatment of RAW 264.7 cells. The Griess reaction was used to measure the nitrite concentration as an indicator of NO production in the supernatant of the macrophages after incubation with the tested stimuli [62]. Briefly, 100 μL of supernatant from each sample was mixed with 90 μL of 1% sulphanilamide (Fluka) in 5% H3PO4 and 90 μL of 0.1% N-(1-naphthyl)ethylene diamine dihydrochloride in H2O. A550 (reference wavelength 620 nm) was measured with an ELISA reader (Tecan Group Ltd).

Cytotoxicity assay

The Neutral Red uptake assay for the estimation of cell viability/metabolic activity was used to test the impact of empty or RV-loaded BGs, as well as pure RV, on the viability of RAW 264.7 cells, as described previously with minor modifications [63]. Briefly, before treatment, the cells (1.25 × 105 per well) were plated in flat-bottom 96-well plates (Sarstedt) and allowed to attach overnight. Next, macrophages were treated with 200 μL of serum-free complete culture medium containing different amounts of empty or RV-loaded BGs (BG : macrophage, 10 : 1, 100 : 1 and 1000 : 1) or with RV alone (15 and 30 μm) for 20 min followed by the removal of BGs using extensive washing and incubation for an additional 20 or 24 h in the presence of used BG types and RV. Cells incubated with Triton X-100 (0.005%) and cells incubated in serum-free culture medium alone were used as controls. Subsequently, cells were carefully washed twice with NaCl/Pi and incubated with 100 μL of Neutral Red (80 μg·mL−1; Carl Roth GmbH + Co. KG, Karlsruhe, Germany) for 2 h at 37 °C in a 5% CO2 humidified incubator. Next, the dye was discarded and plates were extensively washed twice with NaCl/Pi. Neutral Red incorporated in macrophages was released by the addition of 100 μL of the destaining solution (1 mL of acetic acid, 73 mL of 96% ethanol and 26 mL of deionized water). After 10 min of plate shaking, A570 (reference wavelength 690 nm) was measured using a DYNEX Opsys MR plate reader (Dynex Technologies, Chantilly, VA, USA).

Statistical analysis

The results obtained were analyzed using graphpad prism 5 (GraphPad Software, La Jolla, CA, USA). Data are expressed as the mean ± SD. Statistical significance of the difference between the two groups was evaluated using Student's t-test. < 0.05 was considered statistically significant.

Acknowledgements

This work was supported by BIRD-C GmbH&CoKG, Kritzendorf, Austria and the Austrian Science Fund (FWF, P18982-B17). We thank Beate Mayr and Marek Sramko for their helpful discussions and critical comments. V.J.K. performed her PhD work under the supervision of Professor Werner Lubitz, the CEO of the BIRD-C GmbH&CoKG, Kritzendorf, Austria, in the laboratory of BIRD-C. V.J.K. was an employee of BIRD-C GmbH&CoKG, Kritzendorf, Austria; P.K. and W.L. are employees of BIRD-C GmbH&CoKG, Kritzendorf, Austria, which has licensed the rights to the Bacterial Ghosts Technology.