Modulation of butyrate anticancer activity by solid lipid nanoparticle delivery: an in vitro investigation on human breast cancer and leukemia cell lines.

PURPOSE
Histone modification has emerged as a promising approach to cancer therapy. The short-chain fatty acid, butyric acid, a histone deacetylase (HD) inhibitor, has shown anticancer activity. Butyrate transcriptional activation is indeed able to withdraw cancer cells from the cell cycle, leading to programmed cell death. Since butyrate's clinical use is hampered by unfavorable pharmacokinetic and pharmacodynamic properties, delivery systems, such as solid lipid nanoparticles (SLN), have been developed to overcome these constraints.


METHODS
In order to outline the influence of butyrate delivery on its anticancer activity, the effects of butyrate as a free (sodium butyrate, NB) or nanoparticle (cholesteryl butyrate solid lipid nanoparticles, CBSLN) formulation on the growth of different human cancer cell lines, such as the promyelocytic leukemia, HL-60, and the breast cancer, MCF-7 was investigated. A detailed investigation into the mechanism of the induced cytotoxicity was also carried out, with a special focus on the modulation of HD and cyclin-dependent kinase (CDK) mRNA gene expression by real time PCR analysis.


RESULTS
In HL-60 cells, CBSLN induced a higher and prolonged expression level of the butyrate target genes at lower concentrations than NB. This led to a significant decrease in cell proliferation, along with considerable apoptosis, cell cycle block in the G0/G1 phase, significant inhibition of total HD activity and overexpression of the p21 protein. Conversely, in MCF-7 cells, CBSLN did not enhance the level of expression of the butyrate target genes, leading to the same anticancer activity as that of NB.


CONCLUSIONS
Solid lipid nanoparticles were able to improve butyrate anticancer activity in HL-60, but not in MCF-7 cells. This is consistent with difference in properties of the cells under study, such as expression of the TP53 tumor suppressor, or the transporter for short-chain fatty acids, SLC5A8.


INTRODUCTION
Butyric acid is the main short-chain fatty acid (SCFA) produced by bacterial fermentation of dietary fiber in the colon. It is well known that butyrate plays an important role in the homeostasis of colonic mucosa by inducing pathways for cell maturation, as well as cell cycle arrest, differentiation and apoptosis (1). Butyrate-mediated regulation of apoptotic pathways also occurs in cancer cells, including myeloid leukemia and breast cancer cells, and is not limited to the gastrointestinal tract (2)(3)(4). Transcription is the primary target of butyrate, since it causes histone hyperacetylation through non-competitive and reversible inhibition of histone deacetylase (HD).
Butyrate enables DNA binding of several transcription factors, leading to higher genomic activity (5,6). Apoptosis induced by histone deacetylase inhibitors (HDI) is generally associated with various disarrays of cell signaling pathways (7). Indeed, various studies have demonstrated the importance of epigenetic alteration, leading to gene silencing or abnormal expression as a cancer hallmark, and have proposed 'epigenetic' therapies capable of controlling transcription, as a promising approach in cancer therapeutics (8)(9)(10). The efficacy of HDI in cancer therapy may well stem from their restoration of silenced gene expression. _________________________________________ One such example is the fact that the powerful tumor suppressor p53 is the most commonly inactivated protein in human tumors (11). Its inactivation may be the result of TP53 gene alterations and approximately 50% of human cancers carry mutations in this gene (12).
It has now been well established that HDI treatment induces the expression of the cyclindependent kinase (CDK) inhibitor, p21. This plays a critical role in the regulation of cell cycle arrest and apoptosis, causing G1 arrest (13)(14)(15)(16). HDIs have also been shown to induce p16 (CDKN2 or INK4) and p27, but they attenuate levels of cyclin A and cyclin D, leading to decreased CDK4 and CDK2 activity (17). HDI treatment triggers both the intrinsic pathway and sensitizes tumor cells to the extrinsic death ligand-induced pathway of apoptosis (17,18). HDIs also have non-histone substrates and can modulate transcription by directly acetylating/deacetylating transcription factors, as well as associated cofactors such as Sp1 transcription factors or proteins, including p53 (19). Butyric acid seems to be particularly selective in regulating gene expression, leading to transcriptional activation of certain genes, such as CDKN1A coding p21 in a p53independent manner (20,21). However, although the precise molecular mechanism involved in the cancer therapeutics of butyric acid has not yet been fully clarified, it does seem to be highly dependent on the cellular context.
The clinical use of butyrate as an anticancer agent is hampered by unfavorable pharmacokinetic and pharmacodynamic properties. These include its short half-life, which requires administration of continuous parenteral infusions to maintain therapeutic concentrations (22,23). By extension, factors affecting intracellular accumulation of butyrate could potentially influence its availability to modulate gene expression and, hence, processes such as proliferation, differentiation and apoptosis (24). Although free diffusion of fatty acids may take place, it has very little physiological relevance in the uptake of butyrate. There is now a considerable body of evidence suggesting that SCFAs are predominantly taken up via a facilitated process involving a number of transport proteins.
The relatively recent characterization of a number of trans-membrane proteins has led to two well-defined routes of uptake, both monocarboxylate transport proteins: MCT1 also known as SLC16A1, a hydrogen coupled transporter, and SMCT1 (SLC5A8), a sodium-coupled transporter (25). Thus, varying levels of transporter expression may result in a varied cellular uptake of butyrate. In particular, SLC5A8 is the first plasma membrane transporter postulated to function as a tumor suppressor (26). Silencing of its expression by epigenetic mechanisms is an early event in the progression of colorectal cancer and re-expression of the gene in colon tumor cell lines induces apoptosis and prevents colony formation (27,28).
The newly developing field of nanomedicine offers very promising applications for nanotechnology as a drug delivery platform. The aim of nanomedicine can be broadly defined as the comprehensive monitoring, control, construction, repair, defense and improvement of all human biological systems, working from the molecular level using engineered devices and nanostructures, ultimately to achieve medical benefits. Thanks to their small size, nanoparticulate drug delivery systems are encouraging goals in cancer therapy approaches. These include selective or targeted drug delivery, enhanced drug transport across biological barriers (leading to an increased bioavailability of the entrapped drug) or intracellular drug delivery, which minimizes systemic side effects and improves the solubility of poorly water-soluble drugs (29). Thus, cholesteryl butyrate solid lipid nanoparticles (CBSLN) have been put forward as a special delivery system to improve butyrate anticancer activity. This is because they are also a pro-drug, with the lipid matrix of solid lipid nanoparticles made up of cholesteryl butyrate i.e. the ester of cholesterol and butyric acid (30,32). Previous studies have shown that CBSLN are able to prevent premature degradation of the incorporated butyrate and enhance cell membrane crossing (including the blood-brain barrier) (33), oral bioavailability and anticancer efficacy (31,(34)(35)(36). Moreover, CBSLN have shown very low systemic toxicity (37,38) as their matrix contains physiological components and/or excipients of accepted status (FDA-approved constituents) (39,40). Thus, CBSLN as a butyrate drug delivery system might be a valuable anticancer tool, since it improves several aspects of butyrate pharmacology. However, to date, there have been no reports on the effects of CBSLN on butyrate anticancer activity at molecular level.
Our aim was to investigate the anticancer activity of butyric acid as a lipid nanoparticle formulation, with special attention to mRNA gene expression modulation in different human cancer cell lines such as the promyelocytic leukemia, HL-60, and the breast cancer, MCF-7.

MATERIALS AND METHODS
Cholesteryl butyrate solid lipid nanoparticles CBSLN were prepared using the microemulsion method reported elsewhere (30,31). Briefly, CBSLN were prepared from cholesteryl butyrate (0.54 mmol) (Asia Talent Chemical, Shenzen, China) and Epikuron 200 (0.66 mmol) (Cargill, Milan, Italy). Epikuron 200 and cholesteryl butyrate were melted at 85 °C and a warm solution of sodium glycocholate (0.39 mmol) (PCA, Basaluzzo, Italy) was added. 2phenylethanol (Sigma Aldrich, Milan, Italy) was used as a preservative (0.25% w/v final dispersion). The microemulsion was then immediately dispersed in cold water (dispersion ratio 1:5 v/v) and washed by tangential flow filtration using a Vivaflow50 membrane (Sartorius Stedim Biotech GmbH, RC, Cut-off 100,000 Da) for purification of hydrophilic molecules (four washing steps were repeated by addition and removal of the same volume of water). The aqueous dispersion of CBSLN was then sterilized by 0.2 µm filtration before use. Dispersion of CBSLN was characterized according to average diameter (Zave) and polydispersity index (PI) (Photon Correlation Spectroscopy (PCS), Malvern Zetasizer -Nano ZS -176 Deg) and by determination of chemical composition (HPLC-UV analysis, Agilent 1260 and 1200). The zeta potential of the chol-but CBSLN dispersion after final sterilizing filtration was also characterized by laser Doppler velocimetry (LDV) (Malvern Zetasizer -Nano ZS).
Since the whole lipid matrix of CBSLN themselves acts as a prodrug of butyrate, solutions of sodium butyrate were freshly prepared in sterile water before each experiment and used as a drug reference.
The effect of butyrate, as NB or CBSLN, on HL-60 and MCF-7 cell growth was evaluated by WST-1 cell proliferation assay (Roche Applied Science, Penzberg, Germany). 1 × 10 3 HL-60 and 2.5 × 10 3 MCF-7 cells were seeded in 100 μL of growth medium in replicates (n = 8) in a 96-well culture plate. After 72 h of cell growth, the medium was removed and the cells were incubated with the experimental medium containing different butyrate concentrations (0.01, 0.10, 0.50 and 1.00 mM). At 24 h and 48 h, WST-1 reagent (10 μL) was added and the plates were incubated at 37 °C in 5% CO 2 for 2 h. Well absorbance was measured at 450 nm and 620 nm (reference wavelength) in the Asys UV 340 microplate reader (Biochrom, Cambridge, UK). Cell proliferation data were expressed as a percentage of untreated cells.

Flow cytometry
Cell death was evaluated using the Dead Cell Apoptosis Kit with (allophycocyanin) APC-Annexin V and Sytox ® Green (Life Technologies, Milan, Italy). Briefly, 5 x 10 5 HL-60 and MCF-7 cells were treated with butyrate (NB or CBSLN) 0.50 mM and 1.00 mM, respectively. The cells were then washed twice with 1x annexin-binding buffer at 375 rcf for 5 min and were stained with APC-Annexin V and Sytox ® Green at 12, 24 and 48 h. Samples were run on the flow cytometer at a 640 nm excitation to measure APC-Annexin V (FL4) and at 488 nm to measure Sytox ® Green (FL1), respectively. Any cell debris with low FSC and SSC was excluded from the analyses. A total of 10,000 events were analyzed for APC-Annexin V (FL4) and Sytox ® Green (FL1) staining to discriminate apoptotic (Annexin V positive) or late apoptotic/necrotic cells (Annexin V and Sytox ® Green positive) from viable cells (Annexin V and Sytox ® Green negative).
Cell cycle distribution was monitored by adding 2 µM of Vybrant Dye Cycle Green (Invitrogen) live cell staining to 10 6 cells for 30 min at 37 °C. Briefly, both the HL-60 and the MCF-7 cell cycle distributions were evaluated after the treatment with butyrate (NB or CBSLN) 0.50 mM at 12 h and 1.00 mM at 48 h, respectively. The samples were run on the flow cytometer with a 488 nm excitation to measure the Vybrant Dye Cycle Green staining (FL2). Flow cytometry was carried out by a C6 flow cytometer (Accurri Cytometers, Milan, Italy) and the analysis was performed by a FCS Express 4 (BD Bioscience, Milan, Italy).
Cellular uptake studies of CBSLN were carried out by flow cytometry with 3,3'dioctadecyloxacarbocyanine perchlorate (DiO) used as a fluorescence probe. The experiment was conducted on 1 × 10 5 cells in a six-well culture plate. Cells were incubated at 37 °C, 5% CO 2 in supplemented RPMI-1640 medium. Then, HL-60 and MCF-7 cells were treated with 1 mL of 0.50 mM of DiO-tagged CBSLN in RPMI-1640 medium for 1, 5, 15, 30 and 60 min. After each incubation period, the cells were washed three times with PBS, and resuspended in 300 µl PBS and run on the flow cytometer with a 488 nm excitation to measure the DiO staining (FL1). Intracellular fluorescence was expressed as the integrated median fluorescence intensity (iMFI), which was the product of the frequency of DiO-tagged CBSLN positive cells and the median fluorescence intensity of the cells. Results were expressed as the iMFI ratio, i.e. the difference between the iMFI of treated and untreated cells and the iMFI of untreated cells (41).

Fluorescence microscopy
HL-60 and MCF-7 cellular uptake of 3,3'dioctadecyloxacarbocyanine perchlorate (DiO)tagged CBSLN was investigated by fluorescence microscopy. Briefly, the MCF-7 cells were left to attach for 24 h on glass coverslips in 24-well plate, 2 x 10 5 HL-60 and MCF-7 cells were then incubated for 30 min with DiO-tagged CBSLN 0.50 mM. Five minutes before the programmed stop time, the cells were incubated with 1 µg/mL of 4',6-diamidino-2phenylindole (DAPI) for nuclear staining. Once the cells had been washed with PBS, the HL-60 cell suspensions and the MCF-7 on the coverslip were mounted onto a glass slide, observed under a DMI4000B fluorescence microscope (Leica, Wetzlar, Germany) and photographed. A total of 50 cells per slide were analyzed at 40x magnification in three separate areas, three times in three individual experiments.

Real Time RT-PCR
Depending on the effects obtained on cell growth, total RNA was isolated from the HL-60 cells at 6 and 12 h after NB or CBSLN 0.50 mM treatment and from the MCF-7 cells at 24 and 48 h after NB or CBSLN 1.00 mM treatment. Briefly, HL-60 and MCF-7 cells were collected in an RNA Cell Protection Reagent (Qiagen, Milan, Italy) and stored at -80 °C. Total RNA was then obtained using the RNeasy Plus Mini Kit (Qiagen). The total RNA concentration (µg/mL) was determined using the Qubit fluorometer (Invitrogen, Milan, Italy) and the Quant-IT RNA Assay Kit (Invitrogen) was used. Calibration was performed applying a two-point standard curve, according to the manufacturer's instructions. The integrity of the RNA samples was determined using the total RNA 6000 Nano Kit (Agilent Technologies, Milan, Italy) and the Agilent 2100 Bioanalyzer (Agilent Technologies).
Real-time RT-PCR analysis was carried out using 500 ng of total RNA, which was reverse transcribed in a 20 µL cDNA reaction volume using the QuantiTect Reverse Transcription Kit (Qiagen). In accordance with the manufacturer's instructions, 12.5 ng of cDNA was used for each 10 µL real-time RT-PCR reaction. Quantitative RT-PCR was performed using SsoFast EvaGreen (Bio-Rad, Milan, Italy) and the QuantiTect Primer Assay (Qiagen) was used as the gene-specific primer pair for the panel of genes studied ( Table 1). The transcript of the reference glyceraldehyde-3phosphate dehydrogenase gene (GAPDH) was used to normalize mRNA data and real-time PCR was performed using a MiniOpticon Real Time PCR system (Bio-Rad). The PCR protocol conditions were as follows: HotStarTaq DNA polymerase activation step at 95 °C for 30 s, followed by 40 cycles at 95 °C for 5 s and 55 °C for 10 s. All runs were performed with at least three independent cDNA preparations per sample and all samples were run in duplicate. At least two non-template controls were included in all PCR runs. The quantification data analyses were performed using the Bio-Rad CFX Manager Software version 1.6 (Bio-Rad) according to the manufacturer's instructions. These analyses were performed in compliance with MIQE guidelines (Minimum Information for Publication of Quantitative Real-time PCR Experiments) (42).

HD activity quantification
Depending on the effects on cell growth, the HL-60 cells were collected at 12 h after NB or CBSLN 0.50 mM treatment and the MCF-7 cells at 48 h after NB or CBSLN 1.00 mM treatment. In order to evaluate deacetylation by HD enzymes, nuclear extracts were prepared using the Nuclear Extract Kit (Active Motif, Rixensart, Belgium) to obtain the nuclear proteins, starting from 3 x 10 6 cells.
The protein concentration (µg/mL) was quantified using the Qubit fluorometer and the Quant-IT Protein Assay Kit. Calibration was performed applying a two-point standard curve, according to the manufacturer's instructions. We then used the colorimetric HD Assay Kit (Active Motif) according to the manufacturer's instructions. This colorimetric HD assay uses a short peptide substrate containing an acetylated lysine residue that can be deacetylated by class I, II and IV HD enzymes. Well absorbance was measured at 405 nm in the microplate reader Asys UV 340. p21 protein quantification HL-60 cells were collected at 12 h after NB or CBSLN 0.50 mM treatment and the MCF-7 cells at 48 h after NB or CBSLN 1.00 mM treatment. 1 x 10 6 cells were washed twice with PBS, collected in PBS and stored at -80 °C. The p21 in the cell lysates was analyzed using the p21 EIA kit (Enzo Life Sciences, Milan, Italy) according to the manufacturer's instructions. The protein concentration (µg/mL) was quantified using the Qubit fluorometer and the Quant-IT Protein Assay Kit. Calibration was performed applying a two-point standard curve, according to the manufacturer's instructions.

STATISTICAL ANALYSIS
Data are shown as the mean values ± SD of three independent experiments. Statistical analyses were performed with GraphPad Prism 6.0 software (La Jolla, CA, USA) using the analysis of variance (twoway ANOVA). Bonferroni's test was used to calculate the threshold of significance. Statistical significance was set at P < 0.05.

Characterization of cholesteryl butyrate solid lipid nanoparticles
The average diameter (Zave) and polydispersity index (PI) of CBSLN after 0.2 µm filtration were 78.13 nm and 0.240, respectively. The concentration of cholesteryl butyrate and consequently of butyrate, determined by reversed-phase HPLC after washing four times, were 38.6 mM and 38.6 mM, respectively and, after 0.2 µm filtration, 36 mM and 36 mM, respectively.

Effects of SLN butyrate delivery on cell proliferation
CBSLN were observed to enhance butyrate antiproliferative activity in HL-60 cells but not in MCF-7 cells, as shown in Figure 1. In HL-60 cells, CBSLN determined a significant dose-dependent decrease in cell proliferation starting from 24 h of exposure, whereas NB induced only a slight reduction in cell proliferation at 48 h of exposure to the highest concentration tested (1.00 mM). In MCF-7 cells, CBSLN did not enhance the cytotoxic effects of butyrate, since it determined the same NB activity with a significant decrease in cell proliferation starting from 24 h of exposure to the highest concentration tested (1.00 mM). CBSLN were able to increase butyrate activity in leukemic cells but not in breast cancer cells, as compared to free butyrate.

Effects of SLN butyrate delivery on cell death
On the basis of cell proliferation data (Figure 1), HL-60 cell death was evaluated at 12, 24 and 48 h of exposure to 0.50 mM NB or CBSLN. At 0.50 mM, CBSLN induced a significant increase in the percentage of apoptotic and necrotic cells, which reached 50% and 30%, respectively of the whole cell population at 12 h. The increase in apoptotic and necrotic cell percentage was maintained for up to 48 h after treatment ( Figure 2). Conversely, 0.50 mM NB induced an increase in the apoptotic cell percentage which reached about 30% of the whole cell population only at 48 h (Figure 2). On the basis of cell proliferation data, MCF-7 cell death was evaluated at 24 and 48 h after 1.00 mM NB or CBSLN treatment. A slight increase in the apoptotic cell percentage was detectable only at 48 h with NB treatment, while the same formulation did not induce a significant increase in the percentage of necrotic cells (Figure 2). These data demonstrate that CBSLN were able to induce a significant increase in apoptotic and necrotic cell percentage, compared to free butyrate, at 12 h, while CBSLN were not able to improve necrotic or apoptotic cell death in MCF-7 cells.

Effects of SLN butyrate delivery on cell cycle distribution
The lowest butyrate concentrations able to influence cell proliferation, i.e. 0.50 mM for HL-60 cells and 1.00 mM for MCF-7 cells (Figure 1), were evaluated for their effects on cell cycle distribution. Owing to the severe decrease in HL-60 cell viability after 0.50 mM CBSLN treatment, cell cycle distribution was only evaluated at 12 h of incubation. A statistically significant difference in the percentage of HL-60 cells in each cell cycle phase was observed after exposure to 0.50 mM CBSLN compared to both untreated and NB treated cells (Table 2). At 0.50 mM, CBSLN showed a statistically significant increase in HL-60 cells in the G0/G1 phase and a reduction in the S and G2/M phases, which was compatible with an arrest in the G0/G1 phase (Table  2). However, little difference was observed, in both 1.00 mM NB and CBSLN treated cells, in the proportion of MCF-7 cells in any cell cycle phase (Table 2). Thus, CBSLN were able to increase butyrate activity on the leukemic cell cycle, while the two butyrate formulations had similar effects on the breast cancer cell cycle.

CBSLN cellular uptake
Since differences in growth, death and cell cycle distribution were observed between HL-60 and MCF-7 cells exposed to the same CBSLN concentrations, their cellular uptake was investigated first by means of fluorescence microscopy. Both the HL-60 and MCF-7 cells seemed to have internalized the DiO-tagged CBSLN 0.50 mM after 30 min of incubation in more than 80% of the whole cell population (Figure 3a and b). In order to explore the DiO-tagged CBSLN internalization, a flow cytometric analysis was performed. Specifically, leukemia cells were able to internalize DiO-tagged CBSLN 0.50 mM with a high iMFI ratio 1 min after incubation, reaching maximum internalization at 30 min (P < 0.001). The MCF-7 DiO-tagged CBSLN uptake was slower and almost constant during the time of incubation ( Figure  3c). These data showed that the CBSLN internalization rate was higher and the kinetic uptake was faster in leukemia cells than in breast cancer cells.

Effects of SLN butyrate delivery on gene expression
Based on the cell proliferation data (Figure 1), HL-60 RNA was extracted at 6 and 12 h after 0.50 mM NB or CBSLN treatment and MCF-7 RNA was extracted at 24 and 48 h after 1.00 mM NB or CBSLN treatment. HL-60 samples (n = 18) had an RNA integrity number (RIN) of 10.00 ± 0.00 and MCF-7 samples (n = 18) had an RIN of 9.95 ± 0.05. The fold change in mRNA expression levels of genes involved in the butyrate mechanism of action, as compared to the control cells, i.e. untreated cells (expression level = 1), is reported for HL-60 cells in Figure 4 and for MCF-7 cells in Figure 5.
We analyzed the gene expression profile of HL-60 at 6 and 12 h after 0.50 mM NB (Figure 4a) or CBSLN (Figure 4b) treatment. Of note is that the TP53 gene was not expressed in the HL-60 cell line (Figure 4a and b). No significant modifications was observed in the mRNA expression of the HD genes of class I, i.e. HDAC1 and 2 (Figure 4a and b). Conversely, the HD gene of class II, HDAC4, was significantly over-expressed at 12 h after CBSLN treatment ( Figure 4b).
As it is well known that butyrate is able to induce cell cycle arrest and apoptosis, we analyzed the expression levels of some genes involved in these pathways.
CDK4, which regulates the cell cycle during G1/S transition, was slightly over-expressed at 6 h after NB (Figure 4a) and 12 h after CBSLN treatment (Figure 4b). Furthermore, the mRNA expression level of CDKN2A, which codes for the CDK4 inhibitor p16, was increased at 6 h after NB treatment (4.8 ± 0.1, Figure 4a) and, to a greater extent, at 12 h after CBSLN treatment (9.8 ± 0.2, Figure 4b). CDKN1A, which codes for the CDK inhibitor p21characterized by a specific sequence on its promoter region recognized by butyrate -was significantly over-expressed both at 6 and 12 h after NB ( Figure  4a) and CBSLN treatment (Figure 4b). The highest mRNA expression level of CDKN1A was observed at 12 h after CBSLN treatment (64.00 ± 5.00, Figure   4b). This expression profile was in line with the considerable G0/G1 block and apoptosis observed in HL-60 cells after 0.50 mM CBSLN treatment (Table 2 and Figure 2). As treatment with HDI triggers cancer cells in both the extrinsic and intrinsic death ligand-induced pathways for apoptosis, we also analyzed genes such as AKT1 and APAF1. Specifically, it was observed that, AKT1, which is involved in the transduction pathway of the tyrosinekinase receptor, was significantly over-expressed at 6 h after NB treatment ( Figure 4a) and at 12 h after CBSLN treatment (Figure 4b). APAF1, which mediates the cytochrome c-dependent autocatalytic activation of pro-caspase-9, was significantly overexpressed at 6 h after NB treatment (Figure 4a) and at 12 h after CBSLN treatment (Figure 4b). When the pro-apoptotic BAX gene and the anti-apoptotic BCL-2 gene were taken into consideration, no statistically significant differences were observed in the mRNA expression levels, either after NB (Figure 4a) or after CBSLN (Figure 4b) treatment. Furthermore, the expression of MAP3K15 and NFKB1 gene encoding transcriptional factors, which play an essential role in apoptotic cell death triggered by cellular stresses, was affected by neither butyrate, NB (Figure 4a) nor CBSLN (Figure 4b) treatment. The expression levels of two genes, SLC5A8 and SLC16A1 encoding butyrate transporters were also analyzed. There was no SLC5A8 expression in HL-60 cells (Figure 6a and b), whilst the mRNA expression level of SLC16A1 remained unaffected by butyrate both as NB ( Figure  6a) and as CBSLN (Figure 6b).
We analyzed the MCF-7 gene expression profile at 24 and 48 h after 1.00 mM NB (Figure 5a) or CBSLN (Figure 5b) treatment. Noteworthy is the fact that the TP53 gene in the MCF-7 cell line was expressed and differentially modulated by NB and CBSLN at 24 h, reaching control level at 48 h (Figure 5a and b). HDAC1, 2 and 4 were downexpressed at 24 h and over-expressed at 48 h by NB (Figure 5a). Conversely, these genes were overexpressed at 24 h and down-expressed at 48 h after CBSLN treatment (Figure 5b).  Figure b: Representative fluorescence image of MCF-7 cells exposed to 0.50 mM DiO-tagged CBSLN (in green) for 30 min; DAPI (in blue) was used as nuclear counterstain, at 63x magnification. Figure c: Cellular uptake by flow cytometric analysis, cells was exposed to DiO-tagged CBSLN for different incubation time (1, 5, 15, 30, 60 min). Fluorescent signal was detected by a flow cytometer with a 488 nm excitation to measure the DiO staining (FL1) and expressed as integrated median fluorescence intensity (iMFI) ratio calculated as reported in Materials and Methods. *** P < 0.001, ****P < 0.0001 statistical significance versus untreated cells.
mRNA expression levels of CDK2 and CDK4 had significantly decreased at 24 h after NB reaching control level at 48 h (Figure 5a). CBSLN modulated CDK2 expression only at 24 h (Figure 5b). Although CDKN2A was down-regulated by both NB and CBSLN at 48 h, CBSLN treatment led to an upregulation at 24 h (Figure 5a and b). CDKN1A was similarly modulated by NB and CBSLN reaching the same extent of over-expression at 48 h (Figure 5a and b). Moreover, AKT1 was down-regulated at 48 h by NB and CBSLN treatment, as was NFKB1, whilst BAX was slightly over-expressed (Figure 5a and b). Regarding the expression of the butyrate transporter genes, at 48 h we observed overexpression of SLC5A8 after NB treatment and overexpression of SLC16A1 after CBSLN treatment (Figure 6c and d).

Effects of SLN butyrate delivery on HD activity
Data are expressed as the HD activity fold change compared to control cells, i.e., untreated cells. HD activity was 235.27 ± 36.71 nmol/min/mg in untreated HL-60 cells and 192.97 ± 25.42 nmol/min/mg in untreated MCF-7 cells.  CBSLN (b). Values are relative to untreated cells represented as the dash line, i.e. a value of 1; nd: not detectable; * P < 0.05, ** P < 0.01, *** P < 0.001 statistical significance versus untreated cells.
There was a significant reduction in total HL-60 HD activity after 12 h of exposure to 0.50 mM CBSLN, while NB had no effect (Figure 7a). Conversely, there was a slight decrease in the HD activity of MCF-7 nuclear extract samples after 48 h of exposure to 1.00 mM NB, while CBSLN had no effect (Figure 7b).
These alterations in total HD activity are consistent with the HDAC4 mRNA over-expression observed in HL-60 cells at 12 h of exposure to 0.50 mM CBSLN (Figure 4b) and in MCF-7 cells at 48 h of exposure to 1.00 mM NB (Figure 5a), suggesting some kind of compensatory feedback (40).

Effects of SLN butyrate delivery on p21 protein expression
Data are expressed as the p21 protein expression fold change compared to control cells, i.e., untreated cells. The p21 protein was 2626.67 ± 158.41 pg/mL in untreated HL-60 cells and 4760.00 ± 310.03 pg/mL in untreated MCF-7 cells. A significant increase in p21 protein expression was observed in HL-60 cells after 12 h of exposure to 0.50 mM CBSLN (Figure 8a), in line with the marked CDKN1A mRNA over-expression observed ( Figure  4b). Conversely, there was no difference in the MCF-7 p21 protein expression as compared to control cells after either NB, or CBSLN treatment (Figure 8b). This finding is in line with the slight increase in CDKN1A mRNA expression observed in MCF-7 cells after treatment with 1.00 mM NB or after CBSLN treatment (Figure 5a and b).  nd: SLC5A8 gene expression not detectable in HL-60. * P < 0.05, ** P < 0.01, *** P < 0.001 statistical significance versus untreated cells.

DISCUSSION
Solid lipid nanoparticles are able to improve not only the pharmacokinetic, but also the pharmacodynamic properties of anti-cancer agents (39,40,(43)(44)(45). Butyrate induces apoptosis and blocks in the G0/G1 phase in most cell lines at concentrations from 2 to 5 mM (14,46,47). Therefore, we investigated how solid lipid nanoparticles may influence the butyrate anticancer activity at a molecular level. CBSLNswere able to improve in vitro butyrate anticancer activity at lower concentrations in HL-60, but not in MCF-7 cells (Figure 1-2 and Table 2). These data are confirmed by the cellular uptake of fluorescent CBSLN, which was significantly higher in HL-60 than in MCF-7.
It is reasonable to suppose that the form of endocytosis involved in nanoparticle uptake could affect both intracellular localization and trafficking of the nanoparticles. To the best of our knowledge, differences in uptake extent and mechanism by hematological and epithelial cancer cell lines have not been reported to date. This is in spite of the fact that nanoparticles have been shown to exploit more than one pathway in their attempt to gain cellular entry (45).  We hypothesized that the different anticancer efficacy of butyrate as CBSLN between HL-60 and MCF-7 cells may be due to differences in the apoptotic mechanisms involved and/or butyrate cellular transport between the two cell lines. First of all, we investigated butyrate activity as NB or CBSLN in each cell line.
In HL-60 cells, NB and CBSLN had different effects on the expression of genes involved in the butyrate mechanism. In particular, NB determined a lesser and earlier mRNA over-expression (Figure 4a) than CBSLN, which determined a marked overexpression at 12 h (Figure 4b). A possible explanation for this might well be the different cell uptake kinetics as NB or CBSLN, which, in turn, leads to a different transcriptional activity. Of note is that CBSLN induced an up-regulation of HDAC4 mRNA expression at 12 h in HL-60 cells (Figure 4b) and, at the same time, a significant decrease in total HD activity (Figure 7a).
These data are in line with the hypothesis of a HDI-triggered self-regulatory loop due to a compensatory feedback pathway of HD transcription after inhibition of its enzymatic activity (40). Moreover HL-60 cells are TP53 null cells and CBSLN induced a marked CDKN1A mRNA overexpression both at 6 and 12 h (Figure 4b) and a significant increase in p21 protein expression at 12 h (Figure 8a). In keeping with these data, various studies have reported that butyrate dramatically increases p21 expression in cancer cells, such as osteosarcoma, colon carcinoma, glioma and lung cancer (15). The binding of p21 to the cyclin/cyclindependent kinase (CDK) complex is reported to be the main mechanism of butyrate induced inhibition of proliferation, which plays an essential role in G0/G1 arrest. Effectively, it has been reported that p21 forms complexes with cyclin D and CDK 4 or 6 as well as with cyclin E and CDK 2, thus reducing CDK activities and leading to inhibition of G1-S transition (48). Moreover, CDKN1A coding for p21 has a specific sequence in its promoter region, which is recognized by the Sp1/Sp3 transcription factors. Indeed, some authors have reported that the butyrate up-regulation of p21 can be explained by selective butyrate regulation of the Sp transcriptional activators, leading to a common mechanism for cell cycle arrest and apoptosis (19,49).
Our data are in line with several literature reports linking the rise in p21 levels to a histone-direct effect of HDI (14,22,50). Butyrate may also modulate other CDK inhibitors, such as p16 coding by CDKN2A, whose transcription is up-regulated by CBSLN in HL-60 cells (Figure 4b). This result is in line with Siavoshian's work, where the author reported that treatment of HT-29 with butyrate strongly induces cyclin D and p21 expression, but does not influence CDK4, CDK6 and P16 expression (51). The butyrate transporter SLC5A8 was not expressed in HL-60 cells (Figure 6a), whereas the SLC16A1 transporter was expressed and modulated only by CBSLN (Figure 6b). CBSLN was then able to enhance the butyrate inhibition of HL-60 cell proliferation through a p53 independent pathway as a result of HD inhibition, resulting in marked CDKN1A mRNA over-expression and an increase in p21. Therefore, the different butyrate activity in HL-60 cells as CBSLN or NB may be related to butyrate intracellular trafficking improved by SLN, since the active butyrate transporter SLC5A8 was not expressed. Figure 5 shows the expression profile of genes involved in the butyrate mechanism of action in MCF-7 cells exposed to NB or CBSLN. Noteworthy is the fact that NB and CBSLN determined upregulation of the same genes at different time points. There was a general up-regulation trend in the same genes at 48 h for NB and at 24 h for CBSLN. Conversely, NB down-regulated TP53 transcription, whilst CBSLN up-regulated it. NB led to a significant up-regulation of the mRNA expression of HDAC4 at 48 h. These data are in line with the slight NB-induced decrease in total HD activity observed in the MCF-7 cells (Figure 7b). Some authors have reported that butyrate inhibits cell proliferation in p53 wild type cancer cells, such as MCF-7, through the p53 pathway. Indeed p53 acetylation precedes the induction of BAX, suggesting that p53 mediates NB-induced apoptosis (52)(53)(54). Noteworthy was the fact that neither as NB nor as CBSLN was butyrate able to up-regulate CDKN1A mRNA expression in MCF-7 cells to such an extent as to induce an increase in p21 protein expression (Figure 8b). Moreover, NB induced a notable up-regulation of the mRNA expression of BAX at 48 h. CDK2 and CDK4 were significantly down-regulated by NB at 24 h, reaching control level at 48 h. This result is in line with the down-regulation of CDK4 gene expression in the human non-small lung carcinoma cell line H460 treated with NB (47). Butyrate as NB or CBSLN induced a different modulation of the butyrate transporters SLC5A8 and SLC16A1 in MCF-7 cells (Figure 6c and d), i.e., NB-induced SLC5A8 over-expression at 48 h, while CBSLN induced SLC16A1 over-expression at the same time point. Hence, the same butyrate activity as CBSLN or NB in MCF-7 cells may be related to the upregulation of the butyrate active transporters genes SLC5A8 and SLC16A1.
Since the primary target of butyrate may be transcription in both cell lines, CBSLN determined an improved treatment response, as compared to NB, only in HL-60 cells due to the cellular survival/death network. Such findings were also supported by other studies, where low levels of p53 expression are linked to a resistance to HDI (55). Although butyrate's ability to stimulate apoptosis in some adenocarcinoma cells has been well documented, sensitivity to butyrate-induced apoptosis is quite variable and some authors have reported that the magnitude of p21 protein induction in colon cancer cell lines was smaller in the p53 wild-type cell lines than in the p53 mutant cell lines (20). Indeed, using a retroviral delivery system, an overexpression of p21 in mammary tumor cells led to increased apoptosis (55). Furthermore, some data suggest that histone hyperacetylation is partly responsible for the induction of p21 on colon cancer cell lines, because both butyrate and the specific inhibitor of histone deacetylase as trichostatin A resulted in a similar induction of p21 (56). Moreover, as the ability of butyrate to exert its effects depends on its intracellular concentration, solid lipid nanoparticles may improve butyrate concentration, especially in those cells where active butyrate transporters, such as SLC5A8, are not expressed (e.g., HL-60 cells).
Compared to the hydrogen-coupled transporter SLC16A1, SLC5A8 appears to have an extremely high affinity for butyrate at normal concentrations (25). Moreover, this evidence, along with the fact that SLC5A8 is coupled to a Na + concentration gradient, means there is a far greater "push" of butyrate into the cells than with SLC16A1. Indeed, as SLC16A1 relies on an H + gradient and as the magnitude of this gradient across the apical membrane is negligible, there is a smaller driving force pushing butyrate across (27). Therefore, the same effects determined by NB and CBSLN in MCF-7 cells may be due to an NB-induced overexpression of SLC5A8, which may lead to improved butyrate cellular uptake. Moreover, SLC5A8 butyrate-mediated inhibition of HD has been reported as the underlying mechanism of the tumorsuppressive function of SLC5A8 (25).

CONCLUSION
Delivery of butyrate by solid lipid nanoparticles was able to considerably improve butyrate transcriptional activation in HL-60 cells only, leading to enhanced anticancer activity. According to the mRNA gene expression profile, we hypothesize that the different modulation of gene transcription was due to cell properties such as the expression of SLC5A8 with different patterns of butyrate internalization as a consequence. Indeed, HL-60 does not express SLC5A8 (Figure 6a, b), however it was expressed in MCF-7 cells (Figure 6c, d). Therefore, on the basis of these findings, pharmacological enhancement of butyrate transport as solid lipid nanoparticles, according to the specific cancer cell properties, may well represent a useful therapeutic tool in future cancer treatment protocols.