Preparation of carrier-free prodrug nanoparticles for preclinical development
The carrier-free doxorubicin (DOX) prodrug nanoparticles were designed as an alternative formulation to overcome several problems of conventional nano-sized drug delivery system in the terms of technical- and industrial-aspects. First, the cancer-specific prodrug was simply prepared by conjugating cathepsin B-specific cleavable tetrapeptide (Phe-Arg-Arg-Gly; FRRG) to DOX via one-step reaction (Additional file 1: Fig. S1). This absolutely simplified one-step synthesis protocol allowed 100 g batch of preparation as described in “Methods” section. FRRG peptide have high-specificity towards the target bioenzyme of cathepsin B to trigger drug release from prodrug in the targeted tumor cells and maintain non-toxic inactive state in normal cells with innately low cathepsin B expression, leading to enhanced antitumor therapeutic potential with less toxicity [22, 25, 34, 35]. In addition, their precise and concise structure allow easy quality control (QC) after synthesis; thus, we could verify the successful preparation of FRRG-DOX by confirming chemical structure, exact mass and purity via 1H NMR, MALDI-TOF (calculated mass: 1102.17 Da, measured mass: 1102.595 m/z) and HPLC (99%), respectively (Additional file 1: Fig. S2). Importantly, FRRG-DOX molecules self-assembled into prodrug nanoparticles by its intermolecular π-π stacking hydrophobic interactions without any additional carrier materials, resulting in high drug loading (> 50%) [23, 24]. To enhance the in vivo stability, FRRG-DOX nanoparticles were further stabilized with clinically validated pharmaceutical excipient, Pluronic F68 (30% w/w) through simple drop casting method, resulting in F68-FDOX (Fig. 1a). The conventional nano-formulations are normally prepared by resolving raw material medicine and excipient in the solvent with good solubility, followed by lyophilization and re-dispersed in the aqueous condition [36]. However, F68-FDOX was prepared in the aqueous condition that FRRG-DOX molecules exist as a prodrug nanoparticle, for stabilization by surface coating with Pluronic F68. The resulting F68-FDOX was prepared by adding FRRG-DOX solution to the Pluronic F68 solution under the distilled water condition; this simple procedure allowed us to accomplish large scale batch up to 200 g in 2 L volume (Fig. 1b). The F68-FDOX in aqueous condition showed spherical structure with average size of 91.5 ± 17.61 nm, which became smaller after surface coating of FRRG-DOX nanoparticles (321.29 ± 30.36 nm) with Pluronic F68 (Fig. 1c). This is attributable to the formulation with nonionic emulsifier Pluronic F68 that provides an additional steric stabilization effect to prevent aggregation of fine particles, resulting in narrow size distribution and smaller particle size [37]. In addition, the zeta potential of F68-FDOX was also significantly increased than FRRG-DOX owing to the presence of positively charged Pluronic F68 layer on the particle surface (Fig. 1d). As a result, FRRG-DOX nanoparticles were dissociated in mouse serum within 3 days of incubation, while F68-FDOX showed high stability without significant changes of the size and polydispersity index for 6 days (Fig. 1e and Additional file 1: Fig. S3). Importantly, the particle structure of F68-FDOX in the mouse serum was also highly stable in comparison to the FRRG-DOX stabilized with hyaluronic acid or glycine (30% w/w), indicating great suitability of Pluronic F68 as a pharmaceutical excipient to improve the stability of FRRG-DOX. These stable characteristics of F68-FDOX in the physiological condition is suitable to accumulate within tumor tissues via EPR effect in vivo [14]. Next, cathepsin B-specific cleavage of F68-FDOX was confirmed in various conditions. When the F68-FDOX was incubated with MES buffer (pH 5.5) including cathepsin B at 37 °C, 99.53% of F68-FDOX was cleaved to glycine-conjugated DOX (G-DOX) within 9 h post-incubation (Fig. 1f and Additional file 1: Fig. S4); the enzymatic cleavage after incubation of F68-FDOX with cathepsin B was slightly delayed compared to FRRG-DOX owing to reduced enzyme accessibility by surface coating with Pluronic F68 (Additional file 1: Fig. S5). This was clearly supported by MALDI-TOF measurement, wherein the molecular weights of G-DOX (calculated mass: 600.58 Da, measured mass: 656.4 m/z [M + Li] and 657.4 m/z [M + Li + H]) were confirmed at the newly appeared peak (13 min) in the HPLC spectrum after incubation of F68-FDOX with cathepsin B (Additional file 1: Fig. S6). It was already reported that one or two glycine (G) or leucine (L) peptide sequences are cleaved by intracellular lysosomal proteases when they are chemically conjugated to the DOX molecules [38]. Furthermore, previous studies showed that G-DOX cleaved from FRRG-DOX efficiently metabolized into free DOX in cultured cancer cells [21, 32]. In contrast, F68-FDOX was not cleaved when incubated with cathepsin E, D, L or caspase-3 for 24 h (Fig. 1g).
Finally, we developed F68-FDOX as a lyophilized powder form and evaluated the long-term storage stability of lyophilized F68-FDOX powder stored for 3, 6, 12 months in the low (− 4 °C), room (37 °C) or accelerated (60 °C) condition; for these studies, size distribution, chemical structure and purity were analyzed after reconstitution of lyophilized powder stored at each condition (Additional file 1: Fig. S7–S9). The results showed homogeneous size distribution without chemical degradation and impurity formation similar to those of freshly prepared F68-FDOX, in all different conditions, indicating excellent storage stability of lyophilized power form. We also performed same experiment after 24 h of reconstitution using lyophilized F68-FDOX power stored at 12 months in the low temperature, which are considered as a similar condition with clinical use of DOXIL®; no significant changes were observed in size distribution, chemical structure and purity (Fig. 1h). Taken together, the manufacturing operation for mass production of F68-FDOX was optimized for preclinical development, and their physicochemical characterization, such as size distribution, particle stability, target enzyme-specificity, and even the long-term storage stability was successfully evaluated in vitro.
Cellular uptake and cancer cell-specific cytotoxicity of F68-FDOX
The cellular uptake of F68-FDOX was assessed in three types of cancer cells (HT29, human colon adenocarcinoma; MDA-MB231, human breast adenocarcinoma; KPC960, human pancreatic ductal adenocarcinoma) and normal cell (H9C2, rat cardiomyocytes). As expected, three types of cancer cells expressed a 4.78–8.04-fold higher amount of cathepsin B than H9C2 cells (Fig. 2a) [39]. The F68-FDOX and FRRG-DOX showed robust cellular uptake in a time-dependent manner in all types of cells (Fig. 2b and Additional file 1: Fig. S10). Importantly, a strong DOX fluorescence signals (red color) were observed limited to the nuclei of three types of cancer cells owing to internalization of DOX molecules into the nuclei after rapid cleavage by cathepsin B (Additional file 1: Fig. S11). In addition, molecular weight of free DOX in all cancer cells treated with F68-FDOX for 48 h was clearly detected by MALDI-TOF (calculated mass: 543.53 Da, measured mass: 568.2 m/z [M + Na + H]), indicating successful metabolism of G-DOX cleaved from F68-FDOX into free DOX (Additional file 1: Fig. S12). In contrast, F68-FDOX was mainly observed in the perinuclear compartment and cytosol of the cathepsin B-deficient H9C2 cells. Quantitatively, the DOX fluorescence signals in nuclei of F68-FDOX-treated cancer cells (HT29, MDA-MB231 and KPC960) were 7.7–8.0-fold stronger than H9C2 normal cells treated with F68-FDOX after 48 h of incubation (Fig. 2c). Since DOX induces a potent cytotoxicity by DNA intercalation in the nucleus, these intracellular behaviors of F68-FDOX can lead to the cancer-cell specific cytotoxicity, which minimize side effects toward off-target tissues by cathepsin B-specific cleavage mechanism. Next, cellular uptake mechanism of F68-FDOX was assessed in HT29 cells, which express Rab5a–RFP (a marker for early endosomes) or Lamp1–RFP (a marker for lysosomes), respectively. When the HT29 cells were incubated with F68-FDOX (1 μM) for 6 h at 37 °C, approximately 40% of F68-FDOX was observed in the endosomes, and that of 60% localized in the lysosomes (F68-FDOX observed in the endosomes or lysosomes were marked with white arrows, Fig. 2d). These results indicate that F68-FRRG-DOX internalize into the cells through endosomal/lysosomal pathway. Since a lysosomal protease, cathepsin B exhibits the highest enzymatic activity in acidic environment (pH 4–5) of lysosomes, this endocytosis route of F68-FDOX is suitable to enhance cathepsin B-specific drug release [40]. In agreement with the above in vitro results, the IC50 values of F68-FDOX were measured to be 10.62, 8.23 and 10.86 μM in HT29, MDA-MB231 and KPC960 after 48 of incubation, respectively (Fig. 2e); as a control, the cytotoxicity of FRRG-DOX in each cell was similar with F68-FDOX (Additional file 1: Fig. S13). In contrast, F68-FDOX exhibited > 200 μM of IC50 value in H9C2 cells, showing about a 20-fold difference between cancer and normal cells. As a control, DOX induced indiscriminate cytotoxicity with similar IC50 values in all cancer and normal cells (Fig. 2f and g). These results clearly demonstrate that F68-FDOX induce cytotoxicity preferentially in the cancer cells by cathepsin B-specific cleavage after endosomal/lysosomal uptake, while maintain inactive state in cathepsin B-deficient normal cells.
PK/PD and tumor targeting of F68-FDOX
To evaluate enhanced biodistribution and tumor targeting of F68-FDOX, their pharmacokinetics (PK) profile was compared to DOX and FRRG-DOX in BALB/c nu/nu mice. For this analysis, equivalent 4 mg/kg dose based on DOX contents of free DOX, FRRG-DOX or F68-FDOX were intravenously injected into the mice, and blood samples were collected at pre-determined times. Interestingly, DOX showed fast in vivo clearance with a short half-life (t1/2) of 1.33 ± 0.23 h, whereas FRRG-DOX exhibited a significantly extended t1/2 of 7.96 ± 4.59 h (Fig. 3a). Notably, F68-FDOX showed greatly prolonged t1/2 of 25.83 ± 0.8 h, which is attributable to the steric stabilization effect by stabilization with Pluronic F68. In addition, a detectable amount of the F68-FDOX remained for 96 h in the body, showing the dramatically extended residence time in vivo. The various PK parameters, such as area under the curves (AUC), clearance (CL) and volume of distribution (Vd) of F68-FDOX were also greatly improved compared to those of DOX and FRRG-DOX, thereby further confirming longer blood plasma half-life (Fig. 3b). Motivated by the greatly improved PK profiles of F68-FDOX, we assessed tumor targeting in the HT29 tumor-bearing mice, which were prepared by subcutaneous inoculation of 1 × 107 of HT29 cells. When the tumor volumes were approximately 200 mm3, free DOX (4 mg/kg), FRRG-DOX (4 mg/kg based on DOX contents) or F68-FDOX (4 mg/kg based on DOX contents) were intravenously injected into the mice, followed by noninvasive near-infrared fluorescence imaging (NIRF). The NIRF images showed the significantly high tumor accumulation of F68-FDOX after 9 h of injection, wherein the fluorescence intensity of F68-FDOX in the tumor tissues was 6.33–6.82-fold and 2.42–2.71-fold stronger than DOX and FRRG-DOX, respectively (Fig. 3c). In addition, the ex vivo fluorescence imaging of major organs and tumor tissues after 9 h of injection further confirmed the enhanced tumor targeting of F68-FDOX (Fig. 3d and Additional file 1: Fig. S14). The histological analysis of major organs and tumor tissues was further performed after 9 h of injection for confirming more reliable pharmacodynamics (PD) of F68-FDOX; this is because the NIRF intensity of DOX is not large enough in vivo to precisely assess the biodistribution. The results exhibited that DOX was non-specifically distributed in all the major organs and low tumor accumulation, whereas FRRG-DOX highly accumulated in the tumor tissues with less distribution in the off-target tissues (Fig. 3e). Most importantly, F68-FDOX showed most high tumor accumulation owing to the favorable PK with prolonged in vivo residence time for EPR effect, wherein the 14.12–15.01-fold and 1.5–1.580-fold higher DOX fluorescence was observed in the tumor tissues of mice treated with F68-FDOX compared to that of DOX and FRRG-DOX, respectively. Taken together, F68-FDOX efficiently improve the PK/PD profiles of DOX, which significantly enhance the tumor accumulation and mitigate the distribution in the off-target tissues.
In vivo antitumor activity of F68-FRRG-DOX
The antitumor activity of F68-FDOX was assessed in the mice models bearing three types of refractory tumors because one of the key challenges commonly encountered in drug discovery is that antitumor therapeutic potential evaluated with one tumor models do not necessarily translate across different tumor models [41]. The colon, breast and pancreatic tumor models were prepared by subcutaneous inoculation of 1 × 107 of HT29, MDA-MB231 or KPC960, respectively; then, DOX (4 mg/kg), FRRG-DOX (4 mg/kg based on DOX) or F68-FDOX (4 mg/kg based on DOX) were intravenously injected once every three days when the tumor volumes were approximately 80 mm3. As expected, F68-FDOX (137.67 ± 21.61 mm3) significantly delayed the colon tumor growth compared to saline (608.65 ± 210.67 mm3, P < 0.001), DOX (478.75 ± 49.87 mm3, P < 0.01) and FRRG-DOX (347.29 ± 107.48 mm3, P < 0.01) on day 9 after treatment (Fig. 4a). In case of DOX-treated group, all the mice were dead within 9 days owing to the severe systemic toxicity. In addition, the potential antitumor activity of F68-FDOX was also observed in the breast and pancreatic tumor models, showing significantly inhibited tumor progression compared to saline (breast tumor, P < 0.01; pancreatic tumor, P < 0.001), DOX (breast tumor, P < 0.01; pancreatic tumor, P < 0.001) and FRRG-DOX (breast tumor, P < 0.01; pancreatic tumor, P < 0.001; Fig. 4b and c). These results demonstrate the broad therapeutic spectrum of F68-FDOX for the refractory tumors in clinic. The Annexin V staining of single tumor cells from colon tumor tissues further confirmed enhanced antitumor activity of F68-FDOX on day 9 after treatment, wherein the percentage of apoptotic cells was significantly higher in the F68-FDOX group (55.93 ± 4.46%) than in saline (0.4 ± 0.02%), DOX (17.7 ± 1.51%) and FRRG-DOX (35.87 ± 1.87%) groups (Fig. 4d). Tumor tissues stained with TUNEL also showed greatly elevated apoptosis region in tumor tissues of mice treated with F68-FDOX compared to saline (P < 0.001), DOX (P < 0.01) and FRRG-DOX (P < 0.05; Fig. 4e and Additional file 1: Fig. S15). Finally, we examined the in vivo cathepsin B-specificity of F68-FDOX with two groups of colon tumor models: (i) F68-FDOX treatment once every three days along with the local injection with the cathepsin B-inhibitory siRNA 7 times with 2 days-intervals; and (ii) F68-FDOX treatment under the same protocol. Interestingly, co-treatment with cathepsin B-inhibitory siRNA significantly inhibited the antitumor activity of F68-FDOX; as a result, the volumes of tumors (2123.87 ± 171.56 mm3) rapidly increased compared to those of mice treated with F68-FDOX only (438.26 ± 22.55 mm3), on day 15 after treatment (Fig. 4f). These results clearly indicate that F68-FDOX have a broad spectrum of antitumor activity against refractory tumor types and their high in vivo cathepsin B-specificity can be expected to mitigate the DOX-related side effects by maintaining inactive state in cathepsin B-deficient normal tissues.
Safety of F68-FDOX treatment
The safety of F68-FDOX treatment was evaluated in the BALB/c mice after single-/multi-dosage. The DOX (10 mg/kg), FRRG-DOX (10 mg/kg based on DOX) or F68-FDOX (10 mg/kg based on DOX) were intravenously injected into the mice. First, body weight of the mice treated with DOX gradually reduced after treatment due to their severe systemic toxicity (Fig. 5a). In contrast, F68-FDOX- and FRRG-DOX-treated mice showed no significant body weight loss compared to saline-treated group. Consequently, mice in the DOX group were all dead within 9 days of treatment, whereas F68-FDOX-treated mice were survived for up to 30 days (Additional file 1: Fig. S16). Thus, we performed the hematological and histological analyses to compare the toxicity of the treatment on day 9. The serological examination showed severe cardiac, renal and hepatic toxicity in the DOX group, as confirmed by significant change in the hematological parameters, such as blood urea nitrogen (BUN), alanine transaminase (ALT) and troponin-I (Fig. 5b and Additional file 1: Fig. S17). In addition, mice treated with DOX also exhibited severe leukopenia, oligocythemia and thrombocytopenia in the complete blood count (CBC) analyses (Fig. 5c and Additional file 1: Fig. S18). In contrast, all the hematological parameters of F68-FDOX-treated mice were in normal range, which was similar with saline group, indicating greatly minimized DOX-related side effects. Finally, major organ tissues stained with H&E or TUNEL showed elevated structural abnormalities with apoptosis in DOX group, whereas F68-FDOX treatment did not induce noticeable tissue damages (Fig. 5d and Additional file 1: Fig. S19).
Next, we also assessed in vivo toxicity after each drug treatment five times with 3 days-intervals. As expected, systemic toxicity of DOX was more worsen owing to repetitive dose than in the single-dosage, showing severe body weight loss of the mice; accordingly, mice were all dead within 7 days of treatment (Fig. 6a and Additional file 1: Fig. S20). Even though same dose of equal drugs was administered into the mice, the outcomes can be varied according to the not only the number and intervals of the administration, but also different reactivity by species, feeding environment and age of the mice. Similar results by which the mice were all dead within two weeks after 5 mg/kg treatment of free DOX two times (Total 10 mg/kg) were reported, thus these observations are typical results when compared to other studies [42]. In contrast, F68-FDOX treatment showed high safety without significant body weight changes even with high doses of repeated injection. Hematological parameters that are confirmed on day 7 after DOX treatments remarkably got out from the normal range by severe organ dysfunction, while those of mice treated with F68-DOX were similar with saline group (Fig. 6b and Additional file 1: Fig. S21). Finally, histology of liver, spleen and heart tissues on day 7 showed severe tissue damages by DOX treatment, but F68-FDOX efficiently minimized the DOX-related systemic toxicity without damage to the normal organs (Fig. 6c). These results clearly demonstrate that F68-FDOX greatly minimize the DOX-related systemic toxicity accompanying severe cardiotoxicity by maintaining inactive state in normal tissues with innately low cathepsin B expression, improving safety of DOX-based chemotherapy.