|Home | About | Journals | Submit | Contact Us | Français|
The objective of this study was to formulate and evaluate freeze-dried black raspberry (FBR) ethanol extract (RE) loaded poly(DL-lactic-co-glycolic acid) (PLGA) and poly(DL-lactic acid) (PLA) injectable millicylindrical implants for sustained delivery of chemopreventive FBR anthocyanins (cyanidin-3-sambubioside (CS), cyanidin-3-glucoside (CG) and cyanidin-3-rutinoside (CR)).
Identification and quantitation of CS, CG, and CR in RE was performed by mass spectroscopy and HPLC. RE:triacetyl-β-cyclodextrin (TA-β-CD) inclusion complex (IC) was prepared by a kneading method and characterized by X-ray diffraction (XRD), nuclear magnetic resonance spectroscopy (NMR) and UV-visible spectroscopy. RE or RE:TA-β-CD IC-loaded PLGA or PLA implants were prepared by a solvent extrusion method. In vitro and in vivo controlled release studies were conducted in phosphate-buffered saline Tween-80 (pH 7.4, 37°C) and after subcutaneous administration in male Sprague-Dawley rats, respectively. Anthocyanins were quantified by HPLC at 520 nm.
The content of CS, CG, and CR in RE was 0.2, 1.5, and 3.5 wt%, respectively. The chemical stability of anthocyanins in solution was determined to be pH-dependent, and their degradation rate increased with an increase in pH from 2.4 to 7.4. PLGA/PLA millicylindrical implants loaded with 5 or 10 wt% RE exhibited a high initial burst and short release duration of anthocyanins (35–52 and 80–100% CG + CR release after 1 and 14 days, respectively). The cause for rapid anthocyanins release was linked to higher polymer water uptake and porosity associated with the high osmolytic components of large non-anthocyanin fraction of RE. XRD, 1H NMR and UV-visible spectroscopy indicated that the non-anthocyanin fraction molecules of RE formed an IC with TA-β-CD, decreasing the hydrophilicity of RE. Formation of an IC with hydrophobic carrier, TA-β-CD, provided better in vitro/in vivo sustained release of FBR anthocyanins (16–24 and 97–99% CG + CR release, respectively, after 1 and 28 days from 20 wt% RE:TA-β-CD IC/PLA implants) over 1 month, owing to reduced polymer water uptake and porosity.
PLA injectable millicylindrical implants loaded with RE:TA-β-CD IC are optimal dosage forms for 1-month slow and continuous delivery of chemopreventive FBR anthocyanins.
The American Cancer Society estimates that over 35,000 Americans will be diagnosed with oropharyngeal cancer in 2009 (1). Due to the aggressive nature of the disease, the potential for local tumor recurrence and the possibility for development of a second primary cancer, 5-year survival rates for patients with oral squamous cell carcinoma (oral SCC) remain among the lowest of solid tumors (2–4). Cancer chemoprevention, which entails the use of naturally or synthetically derived substances to prevent, inhibit or reverse malignant transformation, could positively impact oral SCC outcomes in two respects. First, chemoprevention could prevent progression of the recognized premalignant lesions of oral epithelial dysplasia to overt oral SCC. Secondly, for persons with previously treated oral SCC, chemoprevention could inhibit local tumor recurrence or development of a new, second primary tumor (5). As numerous complex molecular and biochemical perturbations accompany oral cancer development, current strategies for oral cancer chemoprevention propose use of cocktails comprised of agents with complementary mechanisms of action instead of reliance on a single agent.
Previously, we demonstrated that glutathione and N-acetylcysteine inhibit activation and function of an enzyme that enables progression of oral epithelial dysplasia to invasive oral SCC, i.e. matrix metalloproteinase-9 (6). In light of the potential for oral cancer chemopreventive applications, we subsequently formulated and characterized PLGA implants capable of sustained release of N-acetylcysteine (5). Additional, natural-product-focused studies from our labs showed that freeze-dried black raspberries possess chemopreventive properties that are potentially additive or complementary to those provided by NAC (6-10).
The black raspberry (Rubus occidentalis), a perennial shrub native to North America, contains many compounds with potential chemopreventive effects, including anthocyanins (cyanidin-3-glucoside (CG), cyanidin-3-sambubioside (CS), cyanidin-3-rutinoside (CR), cyanidin-3-xylosylrutinoside, and pelargonidin-3-rutinoside) (8,11), phenolics (ellagic, ferulic, and p-coumeric acid), vitamins (ascorbic acid, α and β-carotene and folate), and minerals (12). Studies from our labs have shown that freeze-dried black raspberry (FBR) ethanol extract (RE) possesses cancer-preventing properties at both the in vitro and in vivo levels (6–10,13). Importantly, no deleterious effects were observed in any of our Phase I human clinical trials of FBR (13–15). Additional studies have shown that the anthocyanin-enriched fraction is the FBR component responsible for inhibition of tobacco-associated carcinogen, redox-mediated activation of the pleiotropic transcription activating factors NF-κB and AP-1 (16–18). Recent in vivo studies have confirmed that anthocyanin-rich extracts are the major FBR components responsible for prevention of esophageal tumors in a rodent model (18). The non-anthocyanin FBR fraction, however, may also contribute to FBR's chemopreventive effects. Non-anthocyanin fraction molecules (phenolics, vitamins, and minerals) of FBR were found to influence the chemopreventive potential of anthocyanins (12,13), signifying the necessity of this fraction in controlled release formulations. Topical application of 10 wt% FBR gel to oral premalignant lesions resulted in positive therapeutic effects, including reduction in histologic grade, decrease in loss of heterozygosity at tumor suppressor gene associated loci, and significant reduction in lesional epithelial cyclooxygenase-2 levels (6,9,10,13). Based on these preliminary in vitro-in vivo results (6,9,10,13), we formulated and characterized RE-loaded PLGA/PLA millicylindrical implants in the current study for long-term localized delivery of FBR anthocyanins.
Implantable poly(DL-lactic-co-glycolic acid) (PLGA) and poly(DL-lactic acid) (PLA) vehicles have been well-demonstrated as potential vehicles for maintaining local tissue levels of therapeutic agents for extended time periods (19–21). In addition to their ability to provide therapeutically relevant local levels without systemic effects, PLGA implants provide another significant clinical benefit. Unlike standard local delivery formulations (gel, cream and ointment for example), which require multiple, repeated dosing throughout the day, PLGA/PLA implantable vehicles alleviate concerns with multiple dosing schedules and poor patient compliance (19–21).
In this study, PLGA and PLA injectable millicylindrical implants were selected for development of controlled release anthocyanins (CS, CG, and CR) in oral SCC. First, identification and determination of the relative amount of anthocyanins in RE were performed by mass spectroscopy and HPLC. The chemical stability of FBR anthocyanins in the medium used for the release study and the effect of pH on their chemical stability was then investigated. Finally, the ability of formulation strategies, such as RE loading, polymer hydrophobicity (lactide:glycolide ratio) and formation of hydrophobic inclusion complex (IC) of RE molecules with triacetyl-β-cyclodextrin (TA-β-CD) to provide slow and continuous in vitro/in vivo release of anthocyanins, was thoroughly investigated. RE:TA-β-CD inclusion complex (RE:TA-β-CD IC) was characterized by XRD, NMR, and UV-visible spectroscopy.
Reference CS was purchased from Polyphenols Laboratories (Sandnes, Norway). Reference CG and CR were purchased from Indofine Chemical Company, Inc. (Hills-borough, NJ, USA). PLGA 50:50 (inherent viscosity, i.v. = 0.57 dl/g and weight average molecular weight, Mw = 51 kDa), PLGA 85:15 (i.v. = 0.61 dl/g and Mw = 95 kDa) and PLA (i.v. = 0.58 dl/g and Mw = 84.5 kDa) were purchased from Medisorb, Alkermes Inc. (Cambridge, MA, USA). All polymers used were lauryl ester end-capped. NMR (5 mm) tubes, phosphoric acid-d3 (D3PO4), deuterium oxide, ethanol-d6, Tween® 80 and TA-β-CD were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA). Xylazine (AnaSed®) injection was purchased from LLoyd Laboratories (Shenandoah, Iowa, USA). Ketamine hydrochloride (Ketaset) injection was purchased from Fort Dodge Animal Health (Fort Dodge, Iowa, USA). All other chemicals, reagents and solvents (analytical or purer grade) were purchased from commercial suppliers.
Black raspberries (Rubus occidentalis) were obtained from a single farmer each year (Stokes Fruit Farm, Wilmington, OH, USA). They are of the same variety (Jewel), grown in the same part of the field, and picked at about the same degree of ripeness (when the majority of the berries in a cluster have turned black—this occurs within a period of 1–2 weeks). The berries are harvested mechanically (with a picker) in a total period of 4 h. During the picking process, the berries are washed mechanically, and each batch is frozen at −20°C within 1 h of the time of harvest. The variation in CS, CG and CR content in black raspberries picked over a period of 5 years was determined to be 10–20 wt%. For the current study, the ethanol extract of freeze-dried black raspberries (RE) was obtained as described previously (8).
Full scan mass spectroscopic data were acquired in the positive mode from m/z 100–700 by using a mass spectrometer (AutoSpec Ultima, Micromass/Waters Corporation, Milford, MA, USA). The ESI voltage, capillary temperature, sheath gas pressure, and auxiliary gas were set at 39 V, 300°C, 65 psi, and 20 psi, respectively.
All HPLC assays were performed on a Waters 2695 alliance system (Milford, MA, USA) consisting of a 2996 Photodiode array (PDA) detector and a personal computer with Empower 2 Software. An Econosphere C18 (4.6 × 250 mm) reverse phase column (Alltech, USA) was used at a flow rate of 1.0 mL/min. Separation of the CS, CG, and CR was achieved using the following mobile phases: (A) 1.5% (v/v) phosphoric acid and (B) formic acid: acetonitrile: phosphoric acid:water (20:26:1.5:52.5). The solvent gradient was as follows: initial composition 80% A and 20% B, changing to 30% A and 70% B at 25 min, then 10% A and 90% B at 30 min and held until 35 min. Finally, the composition returned to 80% A and 20% B (initial conditions) at 40 min before another sample injection at 45 min. All gradient curves were linear. The sample injection volume was 20 μL, and the detection wavelength was 520 nm. Chromatography data were collected and processed using an Empower PDA Software. Standard curves of reference compounds (CS, CG, and CR) were established in 1.5% (v/v) phosphoric acid solution and concentration of CS, CG, and CR in unknown samples was calculated from their respective standard curve.
Chemical stability of anthocyanins (CS, CG, and CR) was studied in two steps. First, chromatographic and visible spectroscopic characteristics of degradation products of anthocyanins were obtained by studying the degradation of each reference compound at a concentration = 18 μg/mL in phosphate-buffered saline Tween-80 (PBST) (pH 7.4). Then, using the 0.1 mg/mL RE, the degradation profile of CS, CG, and CR in PBST and effect of pH (2.4, 5.0 and 7.0) on the anthocyanin (CR) degradation rate was studied. All the solutions were taken into separate 1-mL amber-colored ampoules and sealed under vacuum. Ampoules were placed in an incubator maintained at 37°C and shaken at 100 RPM. At predetermined time intervals, three ampoules were taken out and fraction of remaining anthocyanins was determined by HPLC.
RE:TA-β-CD IC was prepared by a kneading method. Briefly, the required amounts of RE and TA-β-CD (1:1 mass ratio) were triturated with a small amount of 80/20 ethanol/1.5% (v/v) phosphoric acid solution and kneaded thoroughly for 1 h. The resulting slurry was dried under vacuum at room temperature for 24 h and passed through a stainless steel sieve to obtain the particle size of <90 μm.
The XRD pattern of RE, TA-β-CD and RE:TA-β-CD IC was obtained using a Scintag powder X-ray diffractometer (Scintag Inc., Cupertino, CA, USA). The X-ray source was copper Kα (40 kV, 30 mA), and the scanning speed was 3 deg/min.
About 1 mg/mL RE or TA-β-CD or RE:TA-β-CD IC samples were prepared by dissolving the required amount in 1.5% (v/v) D3PO4 for RE or 80/20 (v/v) CD3CD2OD/1.5% (v/v) D3PO4 for TA-β-CD and RE:TA-β-CD IC. About 1 mL of RE, TA-β-CD and RE:TA-β-CD IC solutions were transferred into 5 mm NMR tubes, and one-dimensional 1H NMR spectra were recorded on a Bruker Avance DRX-500 spectrometer (Bruker BioSpin Corporation., Billerica, MA, USA). Typical acquisition parameters consisted of 64 K points covering a sweep width of 10330 Hz, a pulse width (pw90) of 11 μs, and a 0.5 Hz exponential function was applied to the FID before Fourier transformation. The resonances due to residual solvents (water, phosphoric acid and ethanol), present as impurities, were used as internal reference.
Stock solutions of RE (1 mg/mL) and TA-β-CD (10 mg/mL) were prepared in 80/20 ethanol/1.5% (v/v) phosphoric acid solution. At a fixed RE concentration (0.1 mg/mL), solutions of RE/TA-β-CD with varying amount of TA-β-CD (0, 1, 2, 4, 8, and 9 mg/mL) were then prepared. All the solutions in amber-colored vials with screw-caps were incubated at room temperature for 1 h with constant agitation, and then UV-visible spectra were obtained on a UV-visible spectrophotometer (Beckman Instruments Inc., Fullerton, CA, USA).
RE or RE:TA-β-CD IC-loaded millicylindrical implants were prepared by the solvent extrusion method as described previously (5). Briefly, PLGA or PLA was dissolved in acetone to make a final polymer concentration of 55% (w/w) with respect to solvent weight. Sieved (<90 μm) RE (5 and 10 wt%) or RE:TA-β-CD IC (20 wt%) was added to polymer solution to form RE or RE:TA-β-CD IC/polymer mixture. The mixtures were loaded into a 3-ml syringe and slowly extruded into the silicone tubing (I.D = 0.8 mm). The solvent-extruded product was dried at room temperature for 48 h and then in a vacuum oven at 40°C for another 72 h before testing.
To determine the relative amount of CS, CG and CR in RE, ~1 mg RE (n=3) was dissolved in 1.5% (v/v) phosphoric acid solution, injected into the HPLC system and concentration was determined using the calibration curve of respective reference compound. Then, the theoretical loadings of CS, CG and CR in 5 and 10 wt% RE or 20 wt% RE: TA-β-CD IC-loaded implants were calculated. To determine the actual loadings of CS, CG and CR in RE or RE:TA-β-CD IC-loaded implants, ~6 mg implants (n=3) were placed in 5 ml glass vials. To these vials, 2 mL methylene chloride and 1 mL of 1.5% (v/v) phosphoric acid solution were added, followed by spinning for 5 min. The phosphoric acid solution layer was taken out, injected into the HPLC system, and concentration of CS, CG and CR was determined using the calibration curve of respective reference compound. CS, CG and CR loading was calculated as the percentage of amount of CS or CG or CR versus the total weight of mixture (i.e., RE or RE:TA-β-CD IC and PLGA/PLA).
In vitro release studies were conducted in PBST (pH 7.4) under perfect sink conditions. About 5 mg of RE or RE:TA-β-CD IC-containing millicylinders were weighed and placed in 1-ml amber-colored ampoules (3 ampoules for each sampling interval). Exactly 1 ml PBST was added to all ampoules and sealed under vacuum. The ampoules were placed in an incubator maintained at 37°C and shaken at 100 RPM. At predetermined time intervals (1, 3, 7, 14, 21, and 28 days), three ampoules were taken out and broken, and implants were freeze-dried immediately. The amount of anthocyanins remaining in each implant was determined similarly as described in loading determination. The cumulative amount of released anthocyanins was determined by subtracting the content remaining in the implants from the initial loading.
The treatment of experimental animals was in accordance with University of Michigan animal care guidelines, and all NIH guidelines for the care and use of laboratory animals were observed. Male Sprague-Dawley rats of 7–8 weeks old (weighing 250–300 g) were housed in cages and given free access to standard laboratory food and water. Animals were anesthetized with a 9:1 mixture of ketamine (100 mg/ml) and xylazine (100 mg/ml) administered by intraperitoneal injection of 0.5–1.0 mL/kg body weight. RE:TA-β-CD IC/PLA millicylindrical implants (length = ~1 cm and weight = ~7 mg) were subcutaneously implanted into the dorsal region of rats. At different implantation duration (1, 7, 14, 21, and 28 days), the rats were euthanized, and implants (n=5) were carefully recovered by removing the skin. The tissue adhered to the implants was carefully removed with the help of forceps, and then the implants were freeze-dried. The amount of anthocyanins remaining in each implant was determined similarly as described in loading determination. The cumulative in vivo anthocyanins release was determined by subtracting the content remaining in the implants from the initial loading.
After 1, 3, 7, 14, 21, and 28 days of incubation in PBST (pH 7.4) or subcutaneous implantation in male Sprague-Dawley rats (see in vivo anthocyanin release study section for subcutaneous implantation and recovery procedure), implants were blotted with tissue paper, weighed immediately, and then freeze-dried. Freeze-dried implants were weighed again, and water content of the implant was calculated by
where W1 and W2 are the weights of the wet and dry implants, respectively.
The inner morphology of RE and RE:TA-β-CD IC-loaded implants was examined using a Hitachi S3200N scanning electron microscope (Hitachi Ltd., Tokyo, Japan). The millicylinders were fixed previously on a brass stub using double-sided adhesive tape and then were made electrically conductive by coating, in a vacuum, with a thin layer of gold (approximately 3 to 5 nm) for 60 s at 40 W. The cross-sectional view images of millicylinders were taken at an excitation voltage of 20 kV.
Anthocyanins are a group of flavonoids responsible for the red, violet, and blue colors of most of berries and fruits. Berries have been identified as one of the richest sources of anthocyanins. Among the berries, black raspberries (Rubus occidentalis) were found to contain the highest amount of anthocyanins (11,22,23). Cyanidin is one of the six most commonly found anthocyanin aglycones (pelargonidin, cyanidin, peonidin, delphinidin, pentunidin, and malvidin). Different types and numbers of sugars (e.g., glucose, rhamnose, xylose) conjugated to the cyanidin form numerous structures of anthocyanins (e.g., CG, CS, and CR). (8,11,13,22,23).
In the present study, the identification and quantitation of chemopreventive anthocyanins in RE was performed by mass spectroscopy and HPLC. In mass spectroscopy, the mass spectra of RE and reference compounds (CS, CG, and CR) were obtained (see Fig. 1). CS, CG, and CR in RE were identified by comparing the molecular masses (M+) of reference compounds with those of RE sample (see Table I). As shown in Fig. 1, the mass spectrum of all the three reference compounds showed the precursor peak at m/z 287 corresponding to cyanidin. Ions at m/z 581.2, 449.1, and 595.2 in reference compound samples corresponded to the molecular cations (M+) of CS, CG, and CR, respectively. Mass spectrum of RE also showed the peaks at m/z 581.2 (CS), 449.1 (CG), and 595.2 (CR) along with the precursor (cyanidin) peak at m/z 287, indicating the presence of three chemopreventive anthocyanins of black raspberry (8,12,13) in the encapsulated extract. Further, the molecular weight (Mw) of anthocyanins identified in RE corresponded with their theoretical Mw (24) (see Table I). MS/MS fragmentation data obtained with RE and reference compound samples was consistent with the anthocyanins structures (8,11,13,22,23). Loss of M+–294.1, M+–162, and M+–308.1 respectively from CS, CG, and CR molecules corresponded to the cleavage of sambubiosyl, glucosyl, and rutinosyl units, respectively. The current identifications are consistent with the previously published information on black raspberry anthocyanins (8,11,13,22,23).
Further, identification and determination of elution profile and relative abundance of anthocyanins in RE were performed by the chromatographic techniques (see Fig. 2). A typical HPLC profile (see Fig. 2a) displayed 5 peaks at a detection wavelength of 520 nm, obtained from 1 mg/mL RE in 1.5% (v/v) phosphoric acid. All the peaks in chromatogram are well-separated, suggesting the effectiveness of the HPLC method. The peak corresponding to CS, CG, and CR was identified by co-elution. A known amount of reference compounds (CS, CG and CR) was separately dissolved in 1-mL RE solution (1 mg/mL in 1.5% (v/v) phosphoric acid) and injected into the HPLC system. The peak (1–5) area obtained by RE (1 mg/mL) and RE (1 mg/mL) + reference compounds solutions were compared to identify each anthocyanin. Peaks 1, 2, and 3 in Fig. 2b were respectively identified as CS, CG, and CR, as co-injection of the respective reference compound led to an increase in area (data not shown) of peaks 1, 2, and 3, respectively. CS, CG, and CR exhibited a maximum UV absorption at 520 nm and eluted at the retention time of 17.3, 17.8 and 18.8 min, respectively (see Table I). To determine the relative amount of CS, CG, and CR in RE, about 1 mg (n=3) RE was dissolved in 1.5% (v/v) phosphoric acid solution and analyzed by HPLC. The relative abundance of CS, CG, and CR in RE was 0.23, 1.53, and 3.53 wt%, respectively (see Table I).
The chromatographic and UV spectroscopic characteristics of degradation products of CS, CG, and CR were first obtained using respective reference compounds by HPLC (with PDA detector) analysis. Reference compound solutions (CS, CG, and CR) were separately prepared in PBST (pH 7.4), incubated at 37°C over a period of 34 h and analyzed. During this interval, each reference compound showed 3 new chromatographic peaks, suggestive of degradation products. The chromatographic and UV spectroscopic data of degradation products for each reference compound are presented in Table II. Degradation products were assigned product I, II, and III in the order of appearance in the HPLC chromatogram during the study period. Further, degradation products (I, II and III) of all the three reference compounds (CS, CG, and CR) exhibited a close relationship, indicating a close structural similarity. The chromatographic characteristics of degradation products of reference compounds were used to identify the degradation products of RE's anthocyanins in release study polymer samples. Since all the three reference compounds exhibited the same chromatographic pattern, the chromatograms of one reference compound (CR) solution at different time points is shown in Fig. 3.
The stability of black raspberry anthocyanins is highly dependent on pH and presence of other ingredients (13,25–27). Hence, the degradation profiles of FBR anthocyanins in the medium used for release study (PBST) (pH 7.4) and effect of pH on the stability of the RE anthocyanin CR was studied and displayed in Fig. 4. As shown in Fig. 4a, CR exhibited the fastest degradation (81 and 96% degradation after 12 and 24 h of incubation, respectively) over a period of 24 h, whereas CG exhibited slowest degradation (38 and 94% degradation after 12 and 34 h of incubation, respectively), and intermediate degradation rate was observed with CS (57 and 93% degradation after 12 and 24 h of incubation, respectively). To understand the effect of acidic to neutral pH on the stability of FBR anthocyanins, we then investigated the stability of CR at three different pH values (2.4, 5.0 and 7.0). The effect of pH on the stability profile of CR at 37°C is displayed in Fig. 4b. At highly acidic pH 2.4, the CR degradation rate was very slow, with 80% CR remaining after 28 days. At neutral pH, complete degradation of CR occurred in 3 days. Intermediate degradation rate was observed at pH 5.0 (27% CR remaining after 28 days).
Because of their potential oral SCC chemopreventive properties, there is an increasing implication to develop vehicles for localized controlled delivery of FBR anthocyanins (8,12–14,28–30). Implantable polymeric vehicles, particularly injectable configuration (e.g., millicylindrical), are the potential carriers to tune the release locally, because of the ability to control the loading and migration of the implant (20,21,31). RE or RE:TA-β-CD IC-loaded millicylindrical PLGA and PLA implants were prepared by a solvent extrusion method. Three FBR anthocyanins (CS, CG, and CR) were identified in encapsulated RE. We evaluated the loading efficiency of each and total anthocyanins in the implants. As shown in Table III, in each formulation, the loading efficiency of each and total anthocyanins in PLGA and PLA implants was very high (79–95%), suggesting the effectiveness of anthocyanins loading by a solvent extrusion method.
All the anthocyanins degraded completely within 40 h of incubation in PBST (pH 7.4) at 37°C. We evaluated the in vitro anthocyanin release by determining the remaining amount in the implants. In an effort to optimize the controlled release millicylindrical implant formulation parameters, we first prepared the PLGA 50:50 implants loaded with 5 and 10 wt% RE (theoretical) and evaluated the release profiles. Among the three anthocyanins, the amount of CS remaining in the polymer rapidly fell below the detection limit, and hence it was not possible to obtain an accurate release profile of CS. The effects of RE loading on sum of anthocyanins (CG + CR) release and polymer water uptake profiles of PLGA 50:50 millicylindrical implants are shown in Fig. 5. As shown in Fig. 5a, implants with 5 and 10 wt% RE exhibited high initial burst and short release duration of anthocyanins (39–52 and 89–99% CG + CR release after 1 and 14 days, respectively). Implants with 5 and 10 wt% RE exhibited high polymer water uptake (see Fig. 5b) and highly porous polymer matrices (see Fig. 6) compared to blank implants, suggesting very rapid release of RE molecules from the PLGA 50:50 implants, leading to high initial burst and short release duration of anthocyanins. On the other hand, there was no significant difference in release between individual anthocyanins (data not shown).
In an effort to reduce the high initial burst effect and short release duration of anthocyanins, we used polymers with higher lactide content (PLGA 85:15 and PLA) to prepare 10 wt% RE-loaded implants. The effect of lactide: glycolide ratio on sum of anthocyanins (CG + CR) release and polymer water uptake profiles of millicylindrical implants are shown in Fig. 7. As the lactide content increased in the polymer, the implants exhibited slightly reduced initial burst and improved prolonged release duration of anthocyanins (see Fig. 7a). Consistent with the anthocyanin release behavior, lower polymer water uptake (see Fig. 7b) and porosity (see Fig. 8) were exhibited by the higher lactide content polymer implants. Further, we found no difference in release between individual anthocyanins from higher lactide content polymer implants (data not shown).
Examination of in vitro release and polymer water uptake (see Figs. 5 and and7)7) and inner morphology (see Figs. 6 and and8)8) of 5 or 10 wt% RE-loaded PLGA 50:50, PLGA 85:15, and PLA implants indicated a difficulty to control the release of CS, CG, and CR from the implants due to high polymer water uptake and porosity. Among the four commonly found anthocyanins in RE, we identified three anthocyanins in the encapsulated RE, and their total content (CS + CG + CR) was determined to be 5.3 wt%. The remaining very large non-anthocyanin fraction of RE is known to contain other vital components, including phenolics (ellagic, ferulic, and p-coumeric acid), vitamins (ascorbic acid, α and β-carotene and folate), and minerals (12,13,32). The results of initial formulations studied above and high water solubility of non-anthocyanin fraction components (12,13,32,33) suggested the necessity of reducing the rapid release of non-anthocyanin fraction molecules from the implants to obtain an optimum sustained release of anthocyanin fraction molecules. Nevertheless, it is also advantageous to sustain the release of non-anthocyanin fraction molecules from the implants, as these components are shown to influence the chemopreventive potential of anthocyanins in oral SSC (12,13). Thus, to reduce the rapid release of non-anthocyanin fraction molecules (and hence the polymer water uptake and porosity), we formed the hydrophobic inclusion complex of RE molecules with TA-β-CD. TA-β-CD is a hydrophobic derivative of β-cyclodextrin (practically insoluble in water) and has proven to be a potential hydrophobic inclusion complex-forming carrier to obtain sustained release of highly-water soluble molecules (34,35).
RE:TA-β-CD IC was prepared by kneading method at 1:1 mass ratio. The formation of the IC was studied by XRD and NMR. The X-ray diffraction pattern of RE, TA-β-CD, and RE:TA-β-CD IC is shown in Fig. 9. RE exhibited a hollow XRD pattern, indicating the presence of amorphous compounds in the extract. By contrast, TA-β-CD exhibited numerous sharp diffraction peaks, particularly at 2θ of 7.6, 7.9, 10.1, 12.1, 20.2, and 23.1°, indicating the crystalline character of TA-β-CD. Compared to the RE and TA-β-CD XRD patterns, the XRD pattern of RE:TA-β-CD IC was reasonably different. For RE:TA-β-CD IC, the major diffraction peaks of TA-β-CD, observed at 2θ of 7.6, 10.1, and 20.2° were shifted to 2θ of 9.1, 12.3, and 23.0°, respectively. In addition, the intensity of all the major diffraction peaks of TA-β-CD was significantly reduced in the XRD pattern of RE:TA-β-CD IC. Shifts in 2θ value, reduction in intensity and disappearance of some diffraction peaks of TA-β-CD are indicative of formation of inclusion complex with RE molecules, as the change in crystallinity of cyclodextrin arises due to interaction with other molecules (36–38).
To characterize RE:TA-β-CD IC further, 1H NMR spectra of RE, TA-β-CD, and RE:TA-β-CD IC were obtained, and 1H chemical shifts corresponding to RE's molecules and TA-β-CD were calculated. The induced chemical shift (Δδ) is the difference in chemical shifts in the presence and absence of the other reactants (Δδ = δ(complex) - δ(free)). The induced shift can be either downfield (positive sign) or upfield (negative sign). A chemical shift variation of specific host or guest nucleus provides direct evidence for the formation of inclusion complexes (39–42). 1H NMR spectra of RE, TA-β-CD, and RE:TA-β-CD IC are shown in Fig. 10. The protons of RE molecules (H-1 to H-11) and TA-β-CD (H-1 to H-7) involved in chemical shift (Δδ) are numbered as 1–11 or 1–7 in Fig. 10.
The 1H-chemical shifts of the RE sample in both the free and the complexed state are reported in Table IV. As shown, the larger downfield shift (Δδ = + 0.035 to + 0.188) experienced by RE protons (H-1 to H-11) indicates strong hydrophobic interactions with TA-β-CD. Upon inclusion complex formation, the cause for significant downfield shift effect by guest protons has been attributed to a dielectric change in proton's surrounding environment imposed by inclusion inside the hydrophobic cavity (39). To study further, in the presence and absence of RE molecules, we also obtained the induced chemical shift (Δδ) for TA-β-CD protons (see Table IV). Not surprisingly, in the presence of RE molecules, TA-β-CD protons, particularly the cyclodextrin cavity (interior) protons (H-3 and H-7) (39,41– 43), exhibited an upfield shift (Δδ = −0.007 and −0.009, respectively). The upfield shift induced by cyclodextrin protons can be attributed to magnetic anisotropic effects exerted by the guest molecules (42). The downfield and upfield shifts exerted respectively by RE molecules and TA-β-CD protons indicate the formation of inclusion complex of RE molecules with TA-β-CD.
To understand specifically the interaction of identified anthocyanins (CS, CG, and CR) with TA-β-CD, we then employed UV-visible spectroscopy. Changes in the UV-visible absorption spectrum (shift in the absorption maxima or band) have been observed when a molecule forms inclusion complex with cyclodextrin derivatives (42,44). In the present study, the effect of TA-β-CD on the visible absorption band of anthocyanins was qualitatively investigated by keeping the concentration of the guest constant (RE concentration = 0.1 mg/mL) and by varying the host concentration between 1 to 9 mg/mL for TA-β-CD. The absorption band of anthocyanins at 520 nm did not change at varying concentration of TA-β-CD (see Fig. 11), suggesting the absence of interaction of anthocyanins with TA-β-CD. Based on the XRD, 1H NMR and UV-visible spectroscopy results, it can be emphasized that not the anthocyanins but other unidentified hydrophilic and osmotically active molecules (non-anthocyanin fraction molecules) of the RE are involved in the formation of IC with TA-β-CD.
The effect of formation of hydrophobic IC of RE molecules with TA-β-CD on in vitro anthocyanin (sum of CG and CR) release and polymer water uptake characteristics of millicylindrical PLA implants is shown in Fig. 12. Formation of hydrophobic IC not only reduced the initial burst considerably, but also provided better prolonged release of FBR anthocyanins from implants (see Fig. 12a), indicating the slow release of RE (anthocyanin and non-anthocyanin fraction) molecules from the IC-loaded implants. Consistent with these results, lower polymer water uptake (see Fig. 12b) and less polymer porosity (see Fig. 13) were exhibited by the 20 wt% IC-loaded PLA implants than by the 10 wt% RE/PLA implants, thereby providing slow and continuous release of anthocyanins from IC/PLA implants over a period of 28 days. In addition, the difference in release between individual anthocyanins (CS, CG and CR), which was not affected by RE loading and lactide content of the polymer, was also noticeably affected by the formation of hydrophobic IC (data not shown).
Among the PLGA and PLA millicylindrical implant formulations tested in vitro for achieving slow and continuous anthocyanin release, 20 wt% RE: TA-β-CD IC/PLA implants exhibited better controlled release behavior. To evaluate the efficacy of this formulation further, in vivo anthocyanin release from 20 wt% RE:TA-β-CD IC/PLA implants was conducted in male Sprague-Dawley rats (see Fig. 14). As shown in Fig. 14a, when administered in vivo, IC-loaded PLA implants also exhibited low initial burst effect (16% anthocyanin (sum of CG + CR) release after 1 day) and provided 1-month slow and continuous release of anthocyanins (CG + CR). Importantly, an in vitro-in vivo correlation (level A) of anthocyanin release from IC-loaded PLA implants was performed (in vitro vs. in vivo release plot (data not shown)), and an excellent correlation (R2 =0.996) was observed. The data at each time point was further analyzed by two-tailed unpaired Student's t-test, and there was no significant difference (P>0.05) between in vitro and in vivo anthocyanin release from IC-loaded PLA implants. Consistent with this finding, in vivo polymer water uptake (see Fig. 14b) and porosity (see Fig 13) of IC-loaded implants were also very similar with in vitro results (excellent correlation (R2=0.99) and no significant difference (P>0.05) between in vitro-in vivo polymer water uptake). These data further support the potential efficacy of IC-loaded PLA implants in providing 1-month slow and continuous delivery of chemopreventive anthocyanins.
The formation of degradation products of FBR anthocyanins in the PLGA and PLA polymers was examined by analyzing the chromatograms of polymer samples from in vitro-in vivo release study at different time points. Using the retention time of degradation products of reference compounds (see previous section) and by the co-elution technique (co-injection of completely degraded reference compounds + RE solutions), we then identified and monitored the degradation peaks of CS, CG, and CR molecules in RE or RE:TA-β-CD IC-loaded implants. The formation of degradation products of anthocyanins in the PLGA/PLA polymers was negligible, as the peak area of the degradation product was 2–5% of the stable anthocyanin peak (data not shown). Note that anthocyanin (CR) exhibited slow and lowest degradation profile (~ 80% remaining after 28 days) in a very acidic (pH = 2.4) medium (see Fig. 4b). Since the microclimate (μ) pH of these polymers was also found to be very acidic (2.8 < μpH < 5.6) (45,46), it is likely that the acidic polymer microclimate acted as a suitable environment for unreleased anthocyanins to remain stable, similar to PLGA stabilization of the lactone form of camptothecins (47,48).
The goal of this study was to characterize the FBR anthocyanins in RE and to formulate and characterize RE-encapsulated millicylindrical injectable implants for slow and continuous delivery of the anthocyanins. CS, CG, and CR were identified in encapsulated RE, and their respective content in RE was 0.23, 1.53, and 3.53 wt%. PLGA/PLA implants loaded with 5 or 10 wt% RE exhibited high initial burst and short release duration of anthocyanins. High initial burst and short release duration of anthocyanins from PLGA/PLA implants was attributed to high polymer water uptake and porosity caused by very hydrophilic components of large non-anthocyanin fraction of RE. Formation of hydrophobic IC of non-anthocyanin fraction molecules with TA-β-CD provided optimal in vitro/in vivo initial burst effect and 1-month slow and continuous release of chemopreventive anthocyanins (CS, CG and CR).
This study was supported by NIH R01 CA95901 and NIH R01 CA129609. We thank Dr. Scott Woehler, College of Pharmacy, University of Michigan, for the technical assistance with the NMR analysis.