Loading studies of the anticancer drug camptothecin into dual stimuli-sensitive nanoparticles. Stability scrutiny

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in International Journal of Pharmaceutics on 25 October 2018, available at: https://doi.org/10.1016/j.ijpharm.2018.08.026.” 1 Loading studies of the anticancer drug Camptothecin into dual stimuli-sensitive nanoparticles. Stability scrutiny Nieves Iglesiasa, Elsa Galbisa, M. Jesús Díaz-Blancob, M.-Violante de-Paza,*, Juan A. Galbisa [a]. Dpto. Química Orgánica y Farmacéutica, Facultad de Farmacia, Universidad de Sevilla, 41012-Seville, Spain. [b]. PRO2TECS. Departamento de Ingeniería Química, Facultad de Ciencias Experimentales, Campus El Carmen – 21071 - Huelva, Spain. * E-mail: [email protected]; Telephone number: +34954556740; Fax number: +34954556737 Abstract In recent years, the preparation of valuable drug delivery systems (DDS) from selfassembled amphiphilic copolymers has attracted much attention since these nanomaterials provide new opportunities to solve problems such as the lack of solubility in water of lipophilic drugs, improve their bioavailability, prolong their circulation time and decrease the side effects associated with their administration. In the current study two types of biocompatible pH-responsive nanoparticles derived from poly(2-hydroxyethyl methacrylate) (pHEMA) have been used as drug nano-carriers, being one of them core cross-linked to circumvent their instability upon dilution in human fluids. The present paper deals with the optimization of the loading process of 2 the labile, hydrophobic and highly active anticancer drug, Camptothecin (CPT) into the nanoparticles with regard to four independent variables: CPT/polymer ratio, sonication, temperature and loading time. Forty experiments were carried out and a Box–Behnken experimental design was used to evaluate the significance of the independent variables related to encapsulation efficiency and drug retention capacity. The enhanced drug loading and encapsulation efficiency values (58% and >92%, respectively) of CPT were achieved by the core cross-linked NPs in 2 hours at 32 ºC at CPT/polymer ratio 1.5:1 w/w and 14 min of sonication. The optimized CPT-loaded NPs were studied by dynamic light scattering and scanning electron microscopy, and an increase in size of the loaded-NP compared to the unloaded counterparts was found. Other twenty experiments were conducted to study the enability to retain CPT into the conjugates at different ionic strength values and times. The stability studies demonstrated that the core cross-linked nanocarriers displayed an excellent drug retention capacity (> 90%) at 25ºC for 15 days in every ionic-strength environments whereas the non-cross-linked ones were more stable at physiological ionic strength. The optimized systems proved to be a major step forward to encapsulate and retain CPT in the NP nuclei, what makes them ideal devices to control the delivery of CPT upon the triggered acidic conditions of solid tumors. Keywords: Kinetic drug loading, polyHEMA, pH-responsive nanoparticles, drug delivery systems, anticancer therapy, core cross-linked nanoparticles, experimental design. 3 1. Introduction Cancer is a major public health problem worldwide and is the second leading cause of death in the developed countries, which accounts for more than 8.8 million deaths per year. The lifetime probability of developing an invasive cancer is currently above 40% for men and close to it for women and in 2018, 1,735,350 new cancer cases and 609,640 cancer deaths are projected to occur in the United States (Siegel et al., 2018). Treatment is dictated by the cancer type, stage at diagnosis, and the patient's tolerance to the prescribed therapy (Wolinsky et al., 2012). Thus, for solid tumors, surgery is the local treatment of choice as the damage is confined in a limited area of the body. However, most patients require the use of two or more treatments due to the potential spread of the disease as well as to effectively prevent the evolution of the disease from early to advanced stages. Chemotherapy may be combined with surgery or radiotherapy to increase the effectiveness of these treatment modalities (Gavhane et al., 2011). On the other hand, conventional chemical treatments with proven effectiveness in early stages may lack of success in advanced stages. Related to drug administration, the methods used has been traditionally limited to making the drug accessible to the blood stream, relying on the irrigation and the drug affinity for the tissues for the access to the target. In fact, bioavailability is still measured from drug levels in the bloodstream, not in the target surroundings. As a consequence, cancer treatments generally involve the administration of relatively high doses of the drug in the hope that a portion, although minor, will go to damaged tissues (Alvarez-Lorenzo and Concheiro, 2014). Therefore, there is a need to increase the drug concentration in the cancerous tissue while reducing the side effects associated with chemotherapeutic molecules. This requirement is even more compelling in the case of highly toxic anticancer drugs, which may also present too-deficient physico-chemical and stability features such as the labile camptothecin. 4 Twenty-(S)-camptothecin is a strong cytotoxic molecule with excellent antitumor activity in a broad spectrum of human cancers. However, its clinical use is currently limited by its poor solubility in water, low plasma stability due to the cleavage of its lactone ring at physiological pH and severe toxicity. Nowadays, one of the most effective research strategies to achieve a safe and efficient release of camptothecin to target cells is the use of nano-vehicles and a recently published review includes the most innovative approaches (Botella and Rivero-Buceta, 2017). Nanoparticulate drug delivery systems (DDS) have the potential to improve current disease therapies because of their ability to overcome multiple biological barriers and releasing a therapeutic molecule within the optimal dosage range. Angiogenesis during tumor growth results in a defective hypervascularization and a deficient lymphatic drainage system, which has given rise to the concept of passive targeting of nanoparticles (NPs) to tumors through the “enhanced permeability and retention” (EPR) effect (Vivek et al., 2013; Wolinsky et al., 2012). The EPR is a unique feature which allows drug delivery nanocarriers (cutoff size of >400 nm) to accumulate and diffuse preferentially in tumor tissues. Some factors such as composition, size, charge and targeting ligand functionalization, can substantial and positively affect the biodistribution and blood circulation half-life of circulating NP by reducing the level of non-specific uptake, thus delaying opsonization, and increasing the extent of tissue specific accumulation (Alexis et al., 2008). Once the NPs are concentrated in the damage tissues, the drug can be release by the trigger of a specific stimulus and, therefore, boosting the drug concentration where it is required. One of these stimuli is a decrease in pH, feature commonly encountered in the cancerous tissues of solid tumors with regards to the pH of healthy tissues. Regarding the composition of the polymeric NPs, the presence of a highly hydrophilic corona layer makes the micelles very stable in aqueous media (Li et al., 5 2006; Zhao and Liu, 2015). Among the most common hydrophilic blocks, poly(ethylene glycol) (PEG) (Salmaso et al., 2007; Zhao and Liu, 2015), chitosan (Chang and Xiao, 2010), and some polymers obtained from methacrylate esters such as 2-hydroxyethyl methacrylate [HEMA, (Cheng et al., 2012)] and N,N-dimethylaminoethyl methacrylate [DMAEMA, (Dinu et al., 2016)] can be found. In the hydrophobic blocks, the presence of the biocompatible polycaprolactone [(PCL, (Zhao and Liu, 2015)], poly(lactic acid) [PLA, (Karimi et al., 2016)], poly(lactic glycolic acid) [PLGA, (Doppalapudi et al., 2014)], poly(propylene oxide) [PPO, (Biswas et al., 2016)], polydimethylsiloxane [PDMS, (Dinu et al., 2016)], ciclodextrins (Alvarez-Lorenzo and Concheiro, 2014; Salmaso et al., 2007), and other polymers based on methacrylate derivatives such as N,N-diethylaminoethyl methacrylate [DEAEMA, (Liu et al., 2002)] are the most common options. Our research group have designed biocompatible pH-sensitive NPs capable of loading lipophilic molecules and releasing them upon acidic environments (Galbis et al., 2018, 2017). As far as we are aware, there is no precedent in the literature of a finely-tuned design to find the optimal loading conditions of lipophilic molecules into NPs. In the present paper, we aim to investigate the loading of an anticancer drug, CPT, in freshly prepared non- and cross-linked pHEMA-based NPs under different experimental conditions in order to find the optimal values for the four variables studied. For the optimization of the CPT-loading conditions in the pH-responsive systems, we study the effect of CPT/polymer concentration, sonication time, temperature and loading time. Furthermore, stability experiments were carried out to stablish the influence of the ionic strength of the medium in the drug retention capacity of the uploaded conjugates. 6 2. Experimental Section 2.1. Materials Camptothecin was purchased from TCI Europe (Tokyo Chemicals Industry). All other chemicals used were purchased from Sigma-Aldrich and used as received. The 1kDa cut-off mini-dialysis tubes used in the present work were purchased from GE Healthcare. 2.2. General Methods Measurements of UV and visible light absorbance were carried out at an Agilent 8453 UV-visible spectrophotometer (Palo Alto, USA), which presents diode array detection (DAD). The data were the result of, at least, three measurements. A 0.4 kW ultrasonic processor (UP400S, Hielscher Ultrasonics GmbH, Teltow, Germany) with 3 mm of diameter probe was used for sonication. Samples were processed at a constant frequency of 24 kHz. The energy input was controlled by setting the amplitude of the sonicator probe. Extrinsic parameters of amplitude (50%) and time (0, 7 and 14 min), were varied with total duration per pulse cycle On/Off 1 second. Corresponding ultrasonic intensity level was 230 W/cm2. Micelle samples of 12.5 mL were placed in a 100 mL jacketed vessel through which water at 25 ± 1.0 ºC with a flow rate of 0.5 L/min was circulated. The ultrasound probe was submerged to a depth of 12 mm in the sample. The morphology and distribution of the NPs were characterized by scanning electron microscopy (SEM). Before SEM observations, the dispersions were deposited and allowed to dry on a carbon coated grid usually used for transmission electron microscopy. Then, the images were performed by SEM using a field emission HITACHI S5200 microscope operating at 5 kV at the CITIUS Service (University of Seville). 7 The average diameter (Dh) and size distribution (polydispersity index, PdI) of the samples were determined with a Malvern Zetasizer Nano ZS (Malvern Instruments, Malvern, UK) at 25 °C, with a particle size analysis range of 0.6 nm to 6 µm. The intensity of the scattered light (expressed in kilo counts per second) was measured by dynamic light scattering (DLS). The instrument was provided with 4 mW He–Ne laser (λ = 633 nm), digital correlator ZEN3600, and non-invasive backscatter (NIBS®) technology. Measurements were performed with a scattering angle of 173º to the incident beam, and data analyzed using CONTIN algorithms (Malvern Instruments). Data for each dispersion were collected from at least three runs. 2.3. Synthetic procedures and micelle formation The nanoparticles were synthesized following a modified procedure recently described by us (Galbis et al., 2018) and can be summarized as follows: For the preparation of the non cross-linked micellar dispersions, the freshly prepared amphiphilic block-copolymer poly[(DMAEMA31%-HEMA19%)-block-(DEAEMA45%- FMA5%)] (DMAEMA: N,N-dimethylaminoethyl methacrylate; HEMA: 2-hydroxyethyl methacrylate; DEAEMA: N,N-diethylaminoethyl methacrylate; FMA: furfuryl methacrylate. Mn = 34,700; Mw = 45,100; Mw/Mn = 1.3, Figure 1) was dissolved in tetrahydrofuran (THF, 10 mL, polymer concentration = 20 mg/mL), and then the solution was added dropwise into double-distilled water (790 mL, final polymer concentration 0.25 mg/mL) at 30 ºC and stirred for 12 hours. The dispersion adopted the slightly bluish appearance typical of nanosized particle suspensions. The DLS studies revealed that quasi-monodisperse systems were achieved at a polymer concentration of 0.25 mg/mL (PdI = 0.14; Dh = 210 nm). 8 For the synthesis of the core cross-linked micellar dispersions, the cross-linker used was 1,8-dimaleimide-3,6-dioxaoctane (DMDOO, Figure 1) and was freshly prepared following the procedure described by us elsewhere (Galbis et al., 2014). Two separated solutions of the polymer and the crosslinker in THF were prepared. The solution (A) was obtained by dissolving the amphiphilic block-copolymer poly[(DMAEMA31%- HEMA19%)-block-(DEAEMA45%-FMA5%)] in THF (final polymer concentration: 2.04 %w/v); the solution (B) was prepared so that the concentration of the crosslinker in THF was 0.1 %w/v. The solution (A) (0.062 mequiv of FMA units) and the solution (B) (0.0064 mequiv of maleimide units) were mixed and gently stirred at 25 ºC for 5 min (mol ratio FMA/maleimide = 5:1; degree of cross-linked aimed, based on FMA residues: 20%), and then added dropwise into double-distilled water (790 mL, final polymer concentration 0.25 mg/mL) at 30 ºC and stirred for 24 hours. As previously, the dispersion adopted the typical slightly-bluish appearance of NP suspensions. DLS studies revealed that nano-sized systems were achieved (PdI = 0.33; Dh = 130 nm). 2.4. Preparation of camptothecin-loaded NPs The encapsulation studies of camptothecin (CPT, Figure 2) by the freshly prepared non- and core cross-linked NPs were conducted by means of UV spectroscopy varying several experimental parameters such as CPT/polymer ratio, sonication time, temperature and loading time. The optimized CPT-loaded NPs were also studied by dynamic light scattering (DLS) and scanning electron microscopy (SEM). 15 PDMAEMA is another biocompatible polymer used in the co-delivery of paclitaxel and DNA (Guo et al., 2010) with pH responsive behavior (Dinu et al., 2016). This polymer forms part of graft co-polymers or block-copolymers with other biocompatible blocks such as polycaprolactone (PCL) or polyethylene glycol (PEG). An major challenge in the design of smart NPs is to avoid the instability associated to the micelle-unimers equilibrium due to dilution phenomena in the human fluids and the serious side effects concomitant with the premature release of the drug in normal tissues (Zhao and Liu, 2015). That is the reason why one of the nano-sized samples studied was core cross-linked to circumvent not only its potential disintegration but also some inter-micellar cross-linking (Galbis et al., 2017). The presence of furfuryl methacrylate (FMA) moieties in the polymer structure would enable the stabilization of the NPs formed by cross-linking reaction when required. The chemical stabilization of the core cross-linked micelles was carried out by a Diels-Alder coupling reaction between the furan rings in the core with the maleimide rings of the crosslinker. The amount of DMDOO added was set so that 20% of furan rings in the NP core reacted in pairs with maleimide rings of DMDOO, leading to a degree of cross-linking of 20%. This degree of cross-linking was chosen based on our previous findings related to the stability and size of the 20% reticulated core NPs (Galbis et al., 2017), which displayed enhanced features (Figure 4). To optimize the loading of CPT into the freshly prepared NPs, the variables investigated were those that most frequently display a marked effect on drug loading in the literature, i.e., drug/polymer ratio, sonication time, temperature and loading time. The experimental part was carried out by choosing two out of the four variables 16 mentioned above and their influence on the encapsulation efficiency of CPT was studied. 3.2. Modelling and influence of Sonication time and CPT/polymer ratio on the encapsulation efficiency of the NPs Sonication is a technique widely used in the preparation of NP suspensions to obtain samples with consistent properties. Thus, this method has been used for the preparation of graphene oxide NPs as anticancer drug carrier (sonication time: 3 h) (Hashemi et al., 2017) and some magneto-fluorescent nanoparticles based on carboxymethyl chitosan have been sonicated during the preparation procedure (sonication time: 1 h) (Bhattacharya et al., 2011). The sonication procedure can be part of the one-pot micelle-formation / drug-loading as for example in the preparation of chitosan nanoparticles loaded with oxaliplatin, an anticancer agent (sonication time: 30 minutes) (Vivek et al., 2014). However, in general terms, the effect of sonication time on drug loading has been scarcely investigated. One exception is the simultaneous preparation of poly lactic-co-glycolic-acid (PLGA) NPs and load of rifampicin by an optimized method in which sonication time was one of the variables involved (from 4 min to 24 min). The authors found that the best conditions for obtaining stable NPs with maximum percentages of drug loading was achieved after sonicating the formulation for 20 minutes (Tripathi et al., 2010). When methacrylate derivatives are involved, as is the case of the present work, necessary sonication times to guarantee the formation of stable NPs suspensions are, in general terms, quite short. For example, during the preparation of some methacrylate-based cross-linked NP by miniemulsion polymerization, sonication times of 10 minutes were adequate (Griset et al., 2009). When the micelle formation of pHEMA-based co-polymers was attempted by the solvent exchange method, 17 sonication times of 30 minutes were used (Cheng et al., 2012). In contrast, in the case of pDMAEMA-based copolymers, stable NP suspensions were obtained in aqueous media with sonication times as short as just 1 min (Dinu et al., 2016). In the present work, we decided to check the effect that the sonication of the prepared methacrylate NPs could exert onto their CPT load and encapsulation efficiency. The time range chosen in this experimental design for the independent variable sonication time (0, 7 or 14 minutes, depending on the trial) was selected taking into account that our systems are DMAEMA- and HEMA-based NPs. On the other hand, one of the main parameters with a weight influence on drug loading and encapsulation efficiency of hydrophobic molecules into polymer nanoparticles is the drug/polymer ratio in the feed. For example, tamoxifen-loaded chitosan NPs were prepared at various feeding ratios of drug to polymer (Vivek et al., 2013). The load of the anticancer drug honokiol into poly(ethylene glycol)- poly(ε-caprolactone) (MPEG-PCL) micelles increased in line with the honokiol/micelle mass ratio (Gou et al., 2009). Moreover, the loading of rifampicin in poly lactic-co- glycolic-acid (PLGA) nanoparticles was studied at several drug/polymer ratios (Tripathi et al., 2010) as well as was investigated the encapsulation of levofloxacin in poly(lacticco-glycolic acid) (PLGA) NPs, displaying maximum drug loading (86%) at drug/polymer ratio1:10 (Gupta et al., 2011). In general terms, it was observed that the bigger the drug concentration in the feed, the higher was its EE (Chouhan and Bajpai, 2009). This is especially accurate as long as the drug was soluble in the media. For example, Friedrich et al. (Friedrich et al., 2008) observed this trend in the encapsulation of dexamethasone up to a certain drug concentration, in which a drop in EE was observed due to the insolubility of the therapeutic molecule. This was attributed to the overload of the drug (lack of solubility) followed by its crystallization in the external aqueous phase. To confirm this hypothesis, the authors observed the presence of 18 dexamethasone crystals in the suspension by optical microscopy. Hence, the initial drug contents should be in its solubility range to prevent drug precipitation/crystallization. Regarding the choice of CPT concentration range, we took into account the concentrations in which a lineal relationship between drug concentration and absorbance at 369 nm were found (Figure S1, Supplementary Information document). In that way, Table 1 shows the values of independent variables and the experimental values of DL and EE obtained for both, non and cross-linked NPs. 19 Table 1. Experimental drug loading and encapsulation efficiency values of CPT-loaded NPs prepared at 25 C at different CPT/polymer ratios and sonication times following the experimental design1. Non cross-linked loaded NPs Cross-linked loaded NPs Sample Formulation code DL (%) EE (%) Sample Formulation code DL (%) EE (%) 1 Non-CPT/Pol0.5-S0 14.05 32.70 11 Xr-CPT/Pol0.5-S0 24.87 66.21 2 Non-CPT/Pol0.5-S7 9.66 21.39 12 Xr-CPT/Pol0.5-S7 22.95 59.56 3 Non-CPT/Pol0.5-S14 8.45 18.46 13 Xr-CPT/Pol0.5-S14 20.01 50.03 4 Non-CPT/Pol1-S0 32.77 48.75 14 Xr-CPT/Pol1-S0 40.33 67.58 5 Non-CPT/Pol1-S7 31.50 45.98 15 Xr-CPT/Pol1-S7 41.85 71.97 6 Non-CPT/Pol1-S7 33.64 50.69 16 Xr-CPT/Pol1-S7 42.20 73.00 7 Non-CPT/Pol1-S14 32.73 48.66 17 Xr-CPT/Pol1-S14 38.26 61.97 8 Non-CPT/Pol1.5-S0 47.82 61.10 18 Xr-CPT/Pol1.5-S0 52.83 74.66 9 Non-CPT/Pol1.5-S7 46.06 56.93 19 Xr-CPT/Pol1.5-S7 53.23 75.89 10 Non-CPT/Pol1.5-S14 49.02 64.11 20 Xr-CPT/Pol1.5-S14 54.20 78.90 1Each value is the average of three samples (p<0.05). DL = drug loading; EE = encapsulation efficiency; CPT/Pol = CPT/polymer ratio (0.5, 1.0 or 1.5); S = sonication time (0, 7 or 14 minutes). 20 In Table 2, the equations obtained using polynomial regression and statistical parameters (R2, df and F) are shown. Table 2. Equations yielded for the dependent variable (EE) as a function of the independent variables (CPT/polymer ratio and sonication time, normalized values) for the first experimental design. Equation R2 df F 𝑋𝑟𝐸𝐸 = 67.99 + 8.94 𝐶 − 2.93 𝑆 + 5.10 𝐶 𝑆 0.91 3.6 29.11 𝑁𝑜𝑛𝐸𝐸 = 48.52 +18.26 𝐶 − 6.07 𝐶2+ 4.31 𝐶 𝑆 0.97 3.6 57.04 C = CTP/polymer ratio normalized value; S = sonication normalized value; Xr-EE = encapsulation efficiency in percentage for core cross-linked NPs; Non-EE = encapsulation efficiency in percentage for non cross-linked NPs Concerning the response equation, in most cases an acceptable R2 (>0.90) and F (>29) values have been found. Both equations contain complex terms that involve interactions between the independent variables. Identifying the influence of the relative independent statistical variables on the dependent variable in the displayed equations is not straightforward, nor are the calculations to obtain the two values of independent variables at which the maximum percentage of drug absorption could be achieved. To overcome this drawback, the response surface for the dependent variable in each system studied is shown in Figure 5. As can be seen in Figure 5, the relative influence of the cross-linked treatment is readily visible. The EE percentages obtained for cross-linked NPs are, in the studied ranges, higher than those achieved by the non cross-linked ones. This is in line with our findings related to the therapeutic molecule pilocarpine which is used clinically as co-drug in glaucoma and xerostomia and presents low water solubility (Galbis et al., 2018). 21 Among the independent variables studied, the CPT/polymer ratio is the most influential and positive variable on EE percentage for the two systems investigated, being more relevant in the non reticulated NPs. In addition, the lower the CPT/polymer ratio tested, the more marked were the differences in EE encountered between the two types of NPs. The use of higher drug/polymer ratios resulted in an increase in load values under every experimental conditions, similar to the observations found by other research teams in other drug-polymer systems (Biswas et al., 2016). Conversely, sonication time not only exerted a slighter influence on CPT loading than that exerted by the CPT/polymer ratio, but also displayed a dual effect on CPT load depending on the drug/polymer ratio. Thus, for example, sonication showed a negative influence on the loading of the drug in non- and cross-linked NPs at low and medium CPT/polymer ratios, whereas this trend was reversed when the CPT/polymer ratio used was the highest one (1.5:1 w/w); to the latter concentration, the longer the sonication time the higher the EE of CPT. This is probably due to concomitant enhancement of the total micellar surface with the reduction of aggregates in the NP suspensions so that the encapsulation of the drug at a high drug content was facilitated. In summary, the best DL and EE values of CPT were achieved by core cross-linked NPs at the highest CPT/polymer ratio (1.5:1 w/w) and sonication time (14 min). 22 3.3. Modelling and influence of Temperature and Time on NP Uptake This second experimental design was tailored to investigate the influence of temperature and CPT loading time in DL and EE under the optimized conditions disclosed in the first batch of trials, i.e., CPT/polymer ratio =1.5:1 w/w and sonication time = 14 min. For this experimental design, Table 3 shows the values of independent variables —temperature and loading time— and DL and EE of CPT for the two nanocarriers studied. 23 Table 3. Experimental drug loading and encapsulation efficiency values of CPT-loaded NPs at different temperatures and loading times following the experimental design1. Non cross-linked loaded NPs Cross-linked loaded NPs Sample Formulation code DL (%) EE (%) Sample Formulation code DL (%) EE (%) 21 Non-T25-t1 50.05 66.8 31 Xr-T25-t1 54.48 79.78 22 Non-T25-t2 50.06 66.82 32 Xr-T25-t2 54.56 80.05 23 Non-T25-t3 49.90 66.39 33 Xr-T25-t3 54.46 79.72 24 Non-T32-t1 55.50 83.13 34 Xr-T32-t1 57.76 91.18 25 Non-T32-t2 55.50 83.15 35 Xr-T32-t2 57.77 91.21 26 Non-T32-t2 55.81 84.19 36 Xr-T32-t2 57.90 91.67 27 Non-T32-t3 56.58 86.87 37 Xr-T32-t3 57.85 91.48 28 Non-T39-t1 41.19 46.69 38 Xr-T39-t1 44.50 53.45 29 Non-T39-t2 41.43 47.16 39 Xr-T39-t2 43.97 52.32 30 Non-T39-t3 42.95 50.2 40 Xr-T39-t3 44.25 52.92 1Each value is the average of three samples (p<0.05). DL = drug loading; EE = encapsulation efficiency; T = temperature (25, 32 or 39 °C); t = time (1, 2 or 3 hours). 24 Similar to those obtained in the previous section, the estimates of the model coefficients and statistical parameters (R2, df and F) were calculated by means of a polynomial regression between the analytical response and the independent variables (Table 4). Table 4. Equations yielded for the dependent variable (EE) as a function of the independent variables (temperature and loading time, normalized values) for the second experimental design. Equation R2 df F 𝑋𝑟𝐸𝐸 = 91.02 −13.4 𝑇 − 24.90 𝑇2+ 0.37 𝑡2 0.99 3.6 747.2 𝑁𝑜𝑛𝐸𝐸 = 84.51 − 9.33 𝑇 − 27.17 𝑇2− 0.93 𝑡 0.99 3.6 443.4 T = temperature normalized value; ...... t = time normalized value; Xr-EE = encapsulation efficiency in percentage for core cross-linked NPs; Non-EE = encapsulation efficiency in percentage for non cross-linked NPs. Regarding the response equations, suitable R2 (0.99) and F (>400) values have been found. As can be seen in the equations recorded in Table 4, temperature turned to be the most influential variable on EE for both systems (non- and cross-linked NPs) and a low statistical effect of time was found. To obtain the optimal values of EE for the independent variables in the systems studied, the response surfaces for this dependent variable were plotted against temperature and time (Figure 6). Similar to the observations found in the first experimental design, the influence of the cross-linked treatment is fairly visible and the EE obtained for cross-linked NPs are, in the ranges studied, higher than those achieved by the non cross-linked ones. The interval of temperature chosen at the present experimental design is established based on the most common laboratory working temperature (25 ºC) and to avoid reaching the lower critical solution temperature (LCST) of polyDMAEMA (42 ºC (Dong 31 Similar to those calculated in the previous sections, the estimates of the model coefficients and statistical parameters (R2, df and F) in the obtained equations are displayed in Table 7. Table 7. Equations yielded for each dependent variable as a function of the independent variables (normalized values) for the third experimental design. Equation R2 Df F 𝑋𝑟𝐷𝑅𝐶 = 94.50 − 3.19 𝑡 − 0.99 𝑁𝐴 − 2.58 𝑡2− 2.93 𝑁𝐴2 0.99 4.5 67.9 𝑁𝑜𝑛𝐷𝑅𝐶 = 86.82 − 8.77 𝑡 + 3.58 𝑁𝐴 − 1.92 𝑡2− 3.04 𝑁𝐴2+ 1.49 𝑡 𝑁𝐴 0.99 5.4 116.3 NA = NaCl concentration normalized value; ...... t = time normalized value; Xr-DRC = drug retention capacity in percentage for core cross-linked NPs; Non-DRC = drug retention capacity in percentage for non cross-linked NPs. In this case, good R2 (0.99) and F (> 67) values have been found. As can be seen in equations, time is the most influential variable on DRC percentage for both non- and cross-linked NPs. Good drug retention capacities, under both cross-linked treatment, could be observed when using the response surfaces (Figure 8). The most relevant feature of Figure 8 is the excellent DRC (> 90%) displayed by the core cross-linked nanocarriers during 15 days at different ionic strength and under bench conditions. Conversely, a clear influence of time on DRC was observed for the non-cross-linked NPs, being this effect dimished at high salt concentrations. The equilibrium that may have been stablished between the micelles and the unimers in the aqueous media led to the unwanted partial release of the drug to the media (< 15% in 8 days). Consequently, for good DRC of the non cross-linked systems, the CPT-loading nano-vehicles could be freshly prepared and then suspended and stored under solutions with physiological ionic strength for times no longer than 8 days. 32 4. Conclusions The present work have stablished the optimal conditions to upload the anticancer drug camptothecin into NPs, cross-linked or not, with hydrophilic, biocompatible and pH-responsive pHEMA-based shells. The variables investigated were drug/polymer ratio, sonication time, temperature and loading time, being temperature and CPT/polymer ratio the most influencial variables. In the global ranges studied of the four independent variables, the encapsulation efficiency percentages obtained for cross-linked NPs were higher than those achieved by the non cross-linked ones in every trial. Between sonication time and CPT/polymer ratio, the latter was the most influential and positive variable on encapsulation efficiency of CPT for the two systems investigated with an optimal value of 1.5:1 w/w. Sonication time, although not as relevant, exerted a positive effect on drug load at the highest CPT/polymer ratio, with optimized times of 14 min. When temperature and loading time were investigated, the most critical experimental parameter turned to be temperature, with 32 C being the optimum one. Regarding time, CPT was loaded efficiently in the first 2 hours. The optimized CPT-loaded NPs were studied by dynamic light scattering and scanning electron microscopy, and an increase in size of the loaded-NP compared to the unloaded counterparts was found. The stability studies demonstrated that the core cross-linked nanocarriers displayed excellent drug retention capacities (> 90%) at 25ºC for 15 days and in different ionicstrength environments. In addition, the non cross-linked conjugates could behave as good CPT-loaded nano-vehicles when suspended into aqueous media at physiological ionic strength and for periods of time of up to 8 days. The optimized systems proved to be a major step forward to encapsulate and retain CPT in the NP nuclei, what makes 33 them ideal devices to control the delivery of drugs under acidic triggering conditions such as those of solid tumors. 5. Acknowledgements The authors would like to thank the Ministerio de Economía y Competitividad (Grant MAT2016-77345-C3-2-P) and the Junta de Andalucía (Grant P12-FQM-1553) of Spain for financial support. 6. References Alexis, F., Pridgen, E., Molnar, L.K., Farokhzad, O.C., 2008. Factors Affecting the Clearance and Biodistribution of Polymeric Nanoparticles. Mol. Pharm. 5, 505– 515. https://doi.org/10.1021/mp800051m Alvarez-Lorenzo, C., Concheiro, A., 2014. Smart drug delivery systems: from fundamentals to the clinic. Chem. Commun. 50, 7743–7765. https://doi.org/10.1039/c4cc01429d Bhattacharya, D., Das, M., Mishra, D., Banerjee, I., Sahu, S.K., Maiti, T.K., Pramanik, P., 2011. Folate receptor targeted, carboxymethyl chitosan functionalized iron oxide nanoparticles: a novel ultradispersed nanoconjugates for bimodal imaging. Nanoscale 3, 1653–1662. https://doi.org/10.1039/c0nr00821d Biswas, S., Kumari, P., Lakhani, P.M., Ghosh, B., 2016. Recent advances in polymeric micelles for anti-cancer drug delivery. Eur. J. Pharm. Sci. 83, 184– 202. https://doi.org/10.1016/j.ejps.2015.12.031 Botella, P., Rivero-Buceta, E., 2017. Safe approaches for camptothecin delivery: 34 Structural analogues and nanomedicines. J. Control. Release 247, 28–54. https://doi.org/10.1016/j.jconrel.2016.12.023 Chang, Y., Xiao, L., 2010. Preparation and characterization of a novel drug delivery system: Biodegradable nanoparticles in thermosensitive chitosan/gelatin blend hydrogels. J. Macromol. Sci. Part A Pure Appl. Chem. 47, 608–615. https://doi.org/10.1080/10601321003742147 Cheng, R., Wang, X., Chen, W., Meng, F., Deng, C., Liu, H., Zhong, Z., 2012. Biodegradable poly(ε-caprolactone)-g-poly(2-hydroxyethyl methacrylate) graft copolymer micelles as superior nano-carriers for “smart” doxorubicin release. J. Mater. Chem. 22, 11730–11738. https://doi.org/10.1039/c2jm30700f Chouhan, R., Bajpai, A.K., 2009. An in vitro release study of 5-fluoro-uracil (5-FU) from swellable poly-(2-hydroxyethyl methacrylate) (PHEMA) nanoparticles. J. Mater. Sci. Mater. Med. 20, 1103–1114. https://doi.org/10.1007/s10856-008- 3677-x Dinu, I.A., Duskey, J.T., Car, A., Palivan, C.G., Meier, W., 2016. Engineered nontoxic cationic nanocarriers with photo-triggered slow-release properties. Polym. Chem. 7, 3451–3464. https://doi.org/10.1039/C6PY00343E Dong, J., Wang, Y., Zhang, J., Zhan, X., Zhu, S., Yang, H., Wang, G., 2013. Multiple stimuli-responsive polymeric micelles for controlled release. Soft Matter 9, 370– 373. https://doi.org/10.1039/C2SM27116H Doppalapudi, S., Jain, A., Khan, W., Domb, A.J., 2014. Biodegradable polymers-an overview. Polym. Adv. Technol. 25, 427–435. https://doi.org/10.1002/pat.3305 Friedrich, R.B., Fontana, M.C., Beck, R.C.R., Pohlmann, A.R., Guterres, S.S., 2008. Development and physicochemical characterization of dexamethasone-loaded polymeric nanocapsule suspensions. Quim. Nova 31, 1131–1136. https://doi.org/10.1590/S0100-40422008000500038 35 Galbis, E., de-Paz, M.-V., Iglesias, N., Lacroix, B., Alcudia, A., Galbis, J.A., 2017. Core cross-linked nanoparticles from self-assembling polyfma-based micelles. Encapsulation of lipophilic molecules. Eur. Polym. J. 89, 406–418. https://doi.org/10.1016/j.eurpolymj.2017.02.032 Galbis, E., de Paz, M. V, McGuinness, K.L., Angulo, M., Valencia, C., Galbis, J.A., 2014. Tandem ATRP/Diels-Alder synthesis of polyHEMA-based hydrogels. Polym. Chem. 5, 5391–5402. https://doi.org/10.1039/C4PY00580E Galbis, E., Iglesias, N., Lucas, R., Tinajero-Díaz, E., De-Paz, M.-V., Muñoz-Guerra, S., Galbis, J.A., 2018. Validation of Smart Nanoparticles as Controlled Drug Delivery Systems: Loading and pH-Dependent Release of Pilocarpine. ACS Omega 3, 375–382. https://doi.org/10.1021/acsomega.7b01421 Gavhane, Y., Shete, A., Bhagat, A., Shinde, V., Bhong, K., Khairnar, G., Yadav, A., 2011. Solid Tumors : Facts, Challenges and Solutions. Int. J. Pharma Sci. Res. 2, 1–12. Gou, M., Zheng, X., Men, K., Zhang, J., Wang, B., Lv, L., Wang, X., Zhao, Y., Luo, F., Chen, L., Zhao, X., Wei, Y., Qian, Z., 2009. Self-assembled hydrophobic honokiol loaded MPEG-PCL diblock copolymer micelles. Pharm. Res. 26, 2164– 2173. https://doi.org/10.1007/s11095-009-9929-8 Griset, A.P., Walpole, J., Liu, R., Gaffey, A., Colson, Y.L., Grinstaff, M.W., 2009. Expansile nanoparticles: synthesis, characterization, and in vivo efficacy of an acid-responsive polymeric drug delivery system. J. Am. Chem. Soc. 131, 2469– 2471. https://doi.org/10.1021/ja807416t Guo, S., Qiao, Y., Wang, W., He, H., Deng, L., Xing, L., Xu, J., Liang, X.-J., Dong, A., 2010. Poly(ε-caprolactone)-graft-poly(2-(N, N-dimethylamino) ethyl methacrylate) nanoparticles: pH dependent thermo-sensitive multifunctional carriers for gene and drug delivery. J. Mater. Chem. 20, 6935–6941. 36 https://doi.org/10.1039/C0JM00506A Gupta, H., Aqil, M., Khar, R.K., Ali, A., Bhatnagar, A., Mittal, G., 2011. Biodegradable levofloxacin nanoparticles for sustained ocular drug delivery. J. Drug Target. 19, 409–417. https://doi.org/10.3109/1061186X.2010.504268 Hashemi, M., Yadegari, A., Yazdanpanah, G., Omidi, M., Jabbehdari, S., Haghiralsadat, F., Yazdian, F., Tayebi, L., 2017. Normalization of doxorubicin release from graphene oxide: New approach for optimization of effective parameters on drug loading. Biotechnol. Appl. Biochem. 64, 433–442. https://doi.org/10.1002/bab.1487 Jaturanpinyo, M., Harada, A., Yuan, X., Kataoka, K., 2004. Preparation of bionanoreactor based on core-shell structured polyion complex micelles entrapping trypsin in the core cross-linked with glutaraldehyde. Bioconjug. Chem. 15, 344–348. https://doi.org/10.1021/bc034149m Karimi, M., Ghasemi, A., Sahandi Zangabad, P., Rahighi, R., Moosavi Basri, S.M., Mirshekari, H., Amiri, M., Shafaei Pishabad, Z., Aslani, A., Bozorgomid, M., Ghosh, D., Beyzavi, A., Vaseghi, A., Aref, A.R., Haghani, L., Bahrami, S., Hamblin, M.R., 2016. Smart micro/nanoparticles in stimulus-responsive drug/gene delivery systems. Chem. Soc. Rev. 45, 1457–1501. https://doi.org/10.1039/C5CS00798D Li, Y., Lokitz, B.S., Armes, S.P., McCormick, C.L., 2006. Synthesis of reversible shell cross-linked micelles for controlled release of bioactive agents. Macromolecules 39, 2726–2728. https://doi.org/10.1021/ma0604035 Liu, S., Weaver, J.V.M., Tang, Y., Billingham, N.C., Armes, S.P., Tribe, K., 2002. Synthesis of Shell Cross-Linked Micelles with pH-Responsive Cores Using ABC Triblock Copolymers. Macromolecules 35, 6121–6131. https://doi.org/10.1021/ma020447n 37 Salmaso, S., Semenzato, A., Bersani, S., Matricardi, P., Rossi, F., Caliceti, P., 2007. Cyclodextrin/PEG based hydrogels for multi-drug delivery. Int. J. Pharm. 345, 42–50. https://doi.org/10.1016/j.ijpharm.2007.05.035 Siegel, R.L., Miller, K.D., Jemal, A., 2018. Cancer statistics, 2018. CA. Cancer J. Clin. 68, 7–30. https://doi.org/10.3322/caac.21442 Terry, C.A., Fernández, M.J., Gude, L., Lorente, A., Grant, K.B., 2011. Physiologically relevant concentrations of NaCl and KCl increase DNA photocleavage by an N-substituted 9-aminomethylanthracene dye. Biochemistry 50, 10375–10389. https://doi.org/10.1021/bi200972c Thavanesan, T., Herbert, C., Plamper, F.A., 2014. Insight in the phase separation peculiarities of poly(dialkylaminoethyl methacrylate)s. Langmuir 30, 5609–5619. https://doi.org/10.1021/la5007583 Tripathi, A., Gupta, R., Saraf, S.A., 2010. PLGA nanoparticles of anti tubercular drug: Drug loading and release studies of a water in-soluble drug. Int. J. PharmTech Res. 2, 2116–2123. Vivek, R., Nipun Babu, V., Thangam, R., Subramanian, K.S., Kannan, S., 2013. PH- responsive drug delivery of chitosan nanoparticles as Tamoxifen carriers for effective anti-tumor activity in breast cancer cells. Colloids Surfaces B Biointerfaces 111, 117–123. https://doi.org/10.1016/j.colsurfb.2013.05.018 Vivek, R., Thangam, R., Nipunbabu, V., Ponraj, T., Kannan, S., 2014. Oxaliplatinchitosan nanoparticles induced intrinsic apoptotic signaling pathway: A “smart” drug delivery system to breast cancer cell therapy. Int. J. Biol. Macromol. 65, 289–297. https://doi.org/10.1016/j.ijbiomac.2014.01.054 Wolinsky, J.B., Colson, Y.L., Grinstaff, M.W., 2012. Local drug delivery strategies for cancer treatment: Gels, nanoparticles, polymeric films, rods, and wafers. J. Control. Release 159, 14–26. https://doi.org/10.1016/j.jconrel.2011.11.031 38 Wu, W., Shen, J., Banerjee, P., Zhou, S., 2010. Chitosan-based responsive hybrid nanogels for integration of optical pH-sensing , tumor cell imaging and controlled drug delivery. Biomaterials 31, 8371–8381. https://doi.org/10.1016/j.biomaterials.2010.07.061 Zhao, X., Liu, P., 2015. Reduction-responsive core-shell-corona micelles based on triblock copolymers: Novel synthetic strategy, characterization, and application as a tumor microenvironment-responsive drug delivery system. ACS Appl. Mater. Interfaces 7, 166–174. https://doi.org/10.1021/am505531e 39 List of captions Figure 1. Structure of the amphiphilic block-copolymer and the cross-linker used in the present work Figure 2. Chemical structure of camptothecin (CPT) Figure 3. General scheme of the loading process of CPT into the selected nanoparticles at 25 C. Figure 4. Schematic representation of non cross-linked NP (on the left), in which formulation the crosslinker has not been added, and core cross-linked NPs (on the right). In the latter samples, some of the furan rings have reacted with the bifunctional cross-linker DMDOO leading to core cross-linked NPs by means of Diels-Alder reaction. Figure 5. Response surface for CPT encapsulation efficiency percentage on both, non- and cross-linked NPs. Figure 6. Response surfaces for encapsulation efficiency (EE in percentage) against temperature and loading time on non- and cross-linked nanocarriers for the second experimental design. Figure 7. Selected SEM images of the CPT-loaded NP systems at pH 7.0 prepared according to the optimized conditions found in the present study. (a) Non-cross-linked NP (Non-T32-t2); (b) core cross-linked NP with 20% of degree of cross-linking (Xr-T32-t2). Figure 8. Response surface for drug retention capacity (DRC, in percentage) against salt concentration (NaCl, in % w/v) and time on non- and cross-linked nanocarriers for the third experimental design. 40 Names of chemical compounds studied in the article. CPT: camptothecin DMAEMA: N,N-dimethylaminoethyl methacrylate HEMA: 2-hydroxyethyl methacrylate DEAEMA: N,N-diethylaminoethyl methacrylate FMA: furfuryl methacrylate DMDOO: 1,8-dimaleimide-3,6-dioxaoctane