Synthesis and characterization of polymer nanoparticles based on poly-(D, L-lactic-co-glycolic acid)
Poly-(D, L-lactic-co-glycolic acid) nanoparticles (PLGA) possessing both fluorescent and cytotoxic properties were synthesized by the double water-oil-water emulsion method, as shown in Figure 1.
Nile Blue ([9-(diethylamino)benzo[a]phenoxazin-5-ylidene] azanium sulfate, also known as Nile Blue A) was incorporated in the nanoparticles as a fluorescent dye that allows to track the particles inside cells and use them for diagnostic purposes. This dye is successfully used in a wide range of biological applications, such as gel electrophoresis, staining of histological sections, labeling of neutral lipids and fatty acids, and visualization of cancer cells [42]. Nile Blue is a biocompatible dye with absorption and emission maxima in the near-infrared optical window (635 and 674 nm in water solution, respectively), which makes it optimal for labeling target cells in vitro and in vivo.
Doxorubicin was incorporated into the nanoparticles that allow using them for therapeutic applications. Doxorubicin is an anthracycline antibiotic that causes cell death by the interaction with DNA and inhibition of topoisomerase II, which leads to suppression of nucleic acids synthesis, and by the formation of free radicals, that destroy cellular membrane and biomolecules [43, 44].
Nanoparticles were synthesized by the double emulsion “water-oil-water” method with subsequent evaporation of the solvent as shown in Figure 1.
The first emulsion was obtained by the addition of water doxorubicin solution to the solution of PLGA and Nile Blue in chloroform, followed by a short sonication. The second emulsion was obtained by the addition of the first emulsion to the PVA solution containing 1 g/L of chitosan oligosaccharide lactate, followed by the second short sonication. After chloroform evaporation by the slow mixing, nanoparticles were centrifugated and resuspended in PBS.
The morphology of as-synthesized nanoparticles loaded with Nile Blue and doxorubicin was studied by scanning electron microscopy (MAIA3 microscope, Tescan) at an accelerating voltage of 15 kV using an in-beam secondary electron detector. The received images (Fig. 2a) illustrate that synthesized PLGA nanoparticles are spherical monodisperse structures. Image processing with ImageJ software shows that the average size and standard deviation of nanoparticles are 218 ± 59 nm (Fig. 2b). The hydrodynamic size of nanoparticles, measured by the dynamic light scattering method, was determined as 201 ± 38 nm (Fig. 2c) thus completely corresponding to the value of the physical size of nanoparticles. These data indicate that nanoparticles preserved colloidal stability and did not form aggregates in saline solution. Moreover, visual observation showed that nanoparticles were stable for at least 6 months; further observations were not carried out. The ζ-potential of nanoparticles, measured by the electrophoretic light scattering method, was –1.64 mV (Fig. 2d) thus slightly deviating from zero. Such surface charge at pH 7.4 was due to the presence of both negatively charged carboxyl groups –COOH (within the composition of PLGA) and positively charged amino groups –NH2 (within the composition of chitosan) on the surface of the nanoparticles.
The effective incorporation of the fluorescent dye Nile Blue was investigated by fluorescence spectroscopy by measuring the excitation and fluorescence emission spectra of nanoparticles. The excitation spectra were measured in the range from 350 to 675 nm (with emission at 700 nm). Four PLGA nanoparticles with different Nile Blue concentrations used in the synthesis were investigated. The excitation spectra (Fig. 2e) and emission spectra (Fig. 2f) demonstrate that the most effective Nile Blue concentration during the synthesis is 1.7 g/L, the further scaling up of Nile Blue concentration leads to the decrease in fluorescence intensity. It is most probably caused by non-fluorescent H-aggregates formation with an absorption shifted to the blue region of the spectrum. The nanoparticle tracking analysis showed the size of nanoparticles to be equal to 230.7 nm, 227.0 nm, 188.5 nm, 180.5 nm for particles with 5 g/L, 1.7 g/L, 0.5 g/L, 0.18 g/L of Nile Blue used in the synthesis, respectively (Fig. S1). These data show that the Nile Blue concentration significantly affects the nanoparticle size and must be taken into account when developing scalable methods for the synthesis of nanoparticles for in vivo injection.
Barnase*barstar protein interface for the targeted two-step delivery of PLGA particles to HER2-overexpressing cancer cells in vitro
One of the central problems of modern chemotherapy is its relative non-specificity. The surface of nanoparticles is modified with targeting molecules to deliver them to specific cells and tissues. This can help to reduce non-specific toxicity of drugs to normal non-transformed cells. To make this kind of modification universal for any target on the cell surface and include the possibility to "cancel the action on demand”, we propose to mediate the interaction between toxic nanoparticles and molecules recognizing cancer cells using protein adaptors, the barnase*barstar protein pair. Barstar (10 kDa) is a natural inhibitor of bacterial ribonuclease barnase (12 kDa) [38]. The N- and C-terms of both proteins are available for chemical conjugation and genetic engineering and are not located in the active site of both enzymes [38, 40].
We used scaffold protein DARPin9_29 that recognizes the receptor HER2 on the surface of cancer cells with high affinity (KD = 3.8 nM) for the targeted delivery of synthesized polymer PLGA nanoparticles to cancer cells. This modular DDS based on PLGA nanoparticles, protein adaptors barnase*barstar, and scaffold proteins is schematically illustrated in Figure 3.
The surface of the nanoparticles was modified by one of the components of the pair – Figure 3 shows PLGA nanoparticles covalently modified with barnase. During the pre-targeting process, Bs-DARPin9_29 bifunctional protein was added to the cells with HER2 overexpression leading to the selective binding of the anti-HER2 molecule to the cancer cell surface. Next, self-assembly with the second component of the pair was carried out, namely, with PLGA nanoparticles conjugated with barnase. The resulting supramolecular structure selectively interacted with the cells with overexpression of receptor HER2: DARPin9_29 mediated the internalization of PLGA nanoparticles in the cells, while chemotherapy drug induced cell death.
Chemical modification of PLGA nanoparticles was carried out by using the sodium salt of 1-ethyl-3- (3-dimethyl aminopropyl) carbodiimide, EDC, and the sodium salt of N-hydroxysulfosuccinimide, sulfo-NHS, as cross-linking agents through the formation of amide bonds between the carboxyl groups of proteins and amino groups on the surface of nanoparticles. The amino groups are presented on the surface of nanoparticles because of chitosan oligosaccharide lactate surface stabilization. In the first stage of the reaction, proteins were activated with EDC/sulfo-NHS mixture in an acidic buffer with pH 5.0, then nanoparticles were added to the buffer with pH 6.0 or pH 8.0.
The efficacy of conjugation of PLGA nanoparticles to barnase was measured by the enzymatic ability of conjugated nanoparticles, namely, their ability to hydrolyze RNA due to the presence of functionally active barnase on the nanoparticle surface. The measurement was performed by the commonly used method of the acid-insoluble precipitate [39]. First, the solution of conjugated PLGA nanoparticles was mixed with yeast RNA and incubated at 37°C to digest RNA. Then, the reaction was stopped by the addition of sulfuric acid, and the supernatant containing uncleaved RNA was separated by centrifugation. The optical density of the solution corresponding to the concentration of free mononucleotides and proportional to the activity of the enzyme was measured by the microplate reader in 96-well UV-Vis transparent plates. The value of sample absorbance at 260 nm is proportional to the concentration of free mononucleotides in the solution, thus reflecting the RNAse activity of the tested sample, either ribonuclease activity of the tested proteins or nanoparticles conjugated to proteins.
The efficiency of conjugation of PLGA nanoparticles to barstar was measured in a similar way by testing the ability of barstar to inhibit the RNAse activity of free barnase added to the sample of nanoparticles. Nanoparticles conjugated with barstar were pre-incubated with barnase, and the enzymatic activity of the mixture was measured as described above.
First, we tested the functional activity of free barnase and barstar before conjugation to nanoparticles. The enzymatic activity of free barnase and barstar proteins is shown in Figure 4a. The purple curve corresponds to the activity of the free barnase and has a concentration-dependent manner achieving saturation. As a positive control in the experiment directed towards the investigation of the enzymatic activity of PLGA nanoparticles conjugated with barnase, a sample of free barnase at a concentration of 2.5 nM was used. This point corresponds to the middle of the linear range of the barnase enzymatic activity curve. The enzymatic activity of barstar, namely the ability to inhibit barnase, was investigated similarly by its pre-incubation with 2.5 nM of barnase (green curve in Figure 4a) and measuring the enzymatic activity of the sample. As a positive control in the investigation of the enzymatic activity of PLGA nanoparticles conjugated with barstar, a sample with barstar at a concentration of 15 nM (+ barnase 2.5 nM) was used.
We obtained three types of PLGA nanoparticles modified with barnase by three different methods: i) carbodiimide conjugation at pH 8.0, ii) carbodiimide conjugation at pH 6.0, iii) non-covalent protein adsorption on the particle surface. For the modification via protein adsorption, PLGA nanoparticles were incubated with barnase or barstar in PBS at the same concentrations used for covalent coupling for 5 h at room temperature.
Data presented in Figure 4b indicate that the highest efficiency of modification of PLGA nanoparticles with barnase is achieved during chemical conjugation at pH 6.0. Similar data were obtained for PLGA nanoparticles conjugated with barstar: the highest inhibition of the barnase activity is achieved for conjugates obtained at pH 6.0. Therefore, the possibility of obtaining functionally effective PLGA nanoparticles in terms of enzymatic activity with both barnase and barstar has been demonstrated.
2.3. Targeted delivery of polymer PLGA nanoparticles to the HER-overexpressing cells in vitro
We demonstrated the efficiency of two-step DDS vs. one-step DDS for the creation of targeted nanostructures for delivery to the cancer cells for their selective destruction. Namely, conjugates of PLGA nanoparticles with barnase, PLGA-Bn, were obtained and self-assembled on the cancer cell surface using barstar fused with DARPin9_29. Thus, supramolecular structures PLGA-Bn*Bs-DARPin9_29 were assembled on the cell surface using pre-targeting concept via Bs-DARPin9_29 protein and subsequent binding with PLGA-Bn. These structures were delivered to the cells overexpressing receptor HER2.
For the in vitro experiments, we selected two cell lines with various levels of HER2 expression, namely SK-BR-3 and CHO cells. SK-BR-3 is a mammary adenocarcinoma cell line with overexpression of HER2 (about 106 receptors per cell), while CHO, Chinese hamster ovary cells, do not express any receptor of the EGFR family. Expression of HER2 receptor on these cells was confirmed by confocal microscopy (Fig. 5a) and by flow cytometry (Fig. 5b) by imaging cells with fluorescence-labeled full-length antibody against HER2 – Trastuzumab-FITC. Also, the binding of DARPin9_29 was confirmed by cell labeling with DARPin9_29-FITC (Fig. 5b). Data from confocal microscopy and cytometry assays presented in Figure 5a,b indicate that SK-BR-3 cells do express HER2 and are effectively labeled with full-length anti-HER2 antibody Trastuzumab and anti-HER2 scaffold protein DARPin9_29.
The functional activity of DARPin within the composition of fusion proteins with barnase and barstar – Bn-DARPin9_29 and Bs-DARPin9_29 was confirmed by flow cytometry (Fig. 5b) by labeling cells with HER2 receptor overexpression. It was demonstrated that the presence of barnase and barstar in fusion protein does not affect the interaction of DARPin9_29 with HER2-positive cells. Thus, both components of the barnase*barstar protein pair did not influence the functional activity of recognizing scaffold DARPin9_29 and can be used as adaptor proteins mediating two-step targeted drug delivery.
The functional polymer PLGA nanostructures were used for selective targeting of cells with HER2 overexpression. Labeling was carried out by the two-stage targeted delivery method. At the first stage, cells were incubated in suspension with Bs-DARPin9_29 in two different concentrations, followed by washing from unbound protein. Next, the cells were labeled by PLGA-Bn, followed by washing from unbound particles. The binding between nanostructures and cells was estimated by flow cytometry with excitation with a 640 nm laser in the fluorescence channel corresponding to the Nile Blue fluorescence.
The data presented in Fig. 5c indicates highly effective labeling of HER2-overexpressing cells by polymer PLGA nanostructures, assembled on the cells’ surface, PLGA-Bn*Bs-DARPin9_29. Non-specific labeling of cells by PLGA-Bn conjugates is not observed, and binding of PLGA-Bn*Bs-DARPin9_29 with cells has a concentration-dependent manner. With an increase in the concentration of Bs-DARPin9_29 by 5 times from 0.5 µg/mL to 2.5 µg/mL, the median fluorescence intensity of the cell population increases by 17551/2904 = 6 times.
Cytotoxicity of targeted supramolecular structures PLGA Bn*Bs DARPin9_29
The synthesized PLGA polymer nanoparticles contain a fluorescent dye, Nile Blue, and a chemotherapeutic drug, doxorubicin, which induces cell death via apoptosis.
The efficient incorporation of doxorubicin was investigated by fluorescent spectroscopy on nanoparticles that do not contain Nile Blue. Nanoparticles were solved in DMSO and then fluorescence was measured using a fluorescence calibration curve for doxorubicin samples in the same solutions (Fig. 5d). The measurement of the fluorescence of the samples showed that doxorubicin incorporation was 0.9 nmol doxorubicin per 1 mg of nanoparticles. The stability of fluorescence of the synthesized nanoparticles loaded with doxorubicin and Nile Blue was investigated for 1 week both in water and DMSO, the data presented in Fig. 5e confirm that the particles do not bleach during storage at room temperature in a plastic tube without light protection at least for one week.
The cytotoxicity of two-step DDS PLGA-Bn*Bs-DARPin9_29 was investigated by standard MTT-test three days after adding the nanostructures in different concentrations to the HER2-overexpressing cells. The therapeutic efficacy of two-step DDS was compared with i) one-step DDS, namely, PLGA directly conjugated to DARPin9_29, ii) free doxorubicin, which was added to the cells under similar conditions, iii) non-targeted PLGA nanoparticles loaded with doxorubicin. The results of the cytotoxicity study are presented in Fig. 5f, which shows the molar concentration of free doxorubicin and molar concentration of doxorubicin incorporated inside PLGA particles in different formulations.
Half-maximal inhibitory concentration (IC50) calculated for different formulations was found to be:
-
IC50 = 43 ± 3 nM for doxorubicin delivered via two-step DDS;
-
IC50 = 4972 ± 1965 nM for doxorubicin delivered via one-step DDS;
-
IC50 = 441 ± 61 nM for free doxorubicin;
-
IC50 = 134 ± 51.2 nM for doxorubicin delivered via non-targeted PLGA.
Consequently, the incorporation of doxorubicin in the composition of targeted two-step DDS nanoparticles decreases its IC50 by 10.3 times vs. free doxorubicin and more than 100 times vs. one-step DDS. At the same time, cells exposed to non-targeted PLGA nanoparticles or one-step DDS were not affected by the cytotoxic properties of doxorubicin loaded inside nanoparticles and survived by more than 82% even at the highest concentrations of PLGA, namely 1 g/L (Fig. 5f). Hence, including a chemotherapeutic drug in the composition of polymer PLGA nanoparticles assembled on the surface of the cancer cells via barnase*barstar interface significantly decreases the concentration of chemotherapeutic drug doxorubicin, needed to receive the same cytotoxic effect in comparison with either doxorubicin in molecular form or one-step DDS.
Immunogenicity study of barnase and barstar
To study the immunogenicity of barnase and barstar BALB/c mice (18-22 g) were injected intraperitoneally with 10 µg of proteins in 100 µL of sterile pyrogen-free PBS on study days 1, 3, 5, 7, 9, 11, 13. Proteins were pre-purified from lipopolysaccharides using Pierce High Capacity Endotoxin Removal Spin Columns, 0.5 mL. A group of animals that were injected with PBS without proteins was used as a “negative” control. As a “positive” control, was used a group of animals that were injected with proteins according to the same scheme, but at the same time, on day 1, a protein was injected in a mixture with 50 µL of complete Freund's adjuvant, on day 13 – mixed with 50 µL of incomplete Freund's adjuvant. Before and 21 days after the first injection, the blood samples were taken, serum was isolated, and the amount of protein-specific antibodies in the serum was analyzed in the enzyme-linked immunosorbent assay format as described in the Methods section.
The results of the enzyme-linked immunosorbent assay, namely the amount of antibodies depending on the serum dilution for mice injected with the proteins barnase or barstar for the main and control groups are shown in Fig. 6c. The data presented in Fig. 6c indicate the absence of a specific immune response in mice for the tested proteins on day 21 after the first injection even with the "boosting" the immune response by injecting proteins in a mixture with complete and then incomplete Freund's adjuvant. At the same time, the weight of the animals (Fig. 6b) did not change significantly throughout the entire experiment for all experimental groups. The obtained data indicate the possibility of multiple administration of the studied proteins without serious risks associated with the specific B-cell immune response of the organism.
In vivo bioimaging
To compare the effectiveness of targeted two-step and one-step therapy mediated by scaffold polypeptides DARPins, the following in vivo tests were performed. For the one-step targeted delivery, the injection of the conjugate of PLGA nanoparticles with DARPin9_29 was used. For the two-step targeted delivery in vivo, the injection of DARPin9_29-Bs followed by the injection of the conjugate of PLGA-Bn with barnase was used.
To assess the effectiveness of these two approaches for tumor treatment, we developed a mouse tumor model with HER2 overexpression. For this, mouse mammary cancer cells EMT6/P were transduced with transmembrane receptor HER2 via lentiviral transfection, and the single clone of the resultant cells was selected and grown. The HER2 expression on as-obtained cells, EMT-HER2, was confirmed with flow cytometry (Fig. 6d) and confocal laser scanning microscopy (Fig. 6e). Next, these cells were s.c. injected into the right flank of BALB/c mice in full culture medium and after tumor size reaches ~200 mm3 the cryosections of the tumor were performed and stained with anti-HER2 IgG and Hoeschst33342 (Fig. 6f). Thus we show that these cells stably express receptor HER2 both in vitro and in vivo in immunocompetent mice (BALB/c).
As-obtained EMT-HER2 tumors were used in the bioimaging and tumor growth inhibition studies. Mice were divided into three groups that received the following injections: 1) i.v. injection of 500 µg of the conjugate of PLGA nanoparticles with DARPin9_29 on days 8 and 10 of the treatment, 2) i.v. injection of 150 µg of DARPin9_29-Bs and 2h later i.v. injection of 500 µg of the conjugate of PLGA nanoparticles with Bn on days 8 and 10 of the treatment, 3) the control group received no treatment.
The accumulation of PLGA nanoparticles in the tumor area was monitored 4 h after the nanoparticles injection with the Lumotrace FLUO bioimaging system (Fig. 6h). We showed that the two-step targeted drug delivery is much more effective in terms of nanoparticle accumulation in tumors for bioimaging purposes.
In vivo therapy
To evaluate the efficacy of two-step targeted delivery in comparison with one-step delivery for the tumor treatment, we measured the tumor size in all experimental groups. The tumor growth dynamics are presented in Fig. 6g.
The tumor growth inhibition indexes at 21 day calculated as %TGI = (1 − {Tt/T0}/{Ct/C0})/(1 − {C0/Ct}) × 100 where Tt = median tumor volume of treated at time t, T0 = median tumor volume of treated at time 0, Ct = median tumor volume of control at time t and C0 = median tumor volume of control at time 0 were found to be equal to TGI1 = 68.4% for one-step DDS and TGI2 = 94.9% for two-step DDS. The obtained data show that the two-step delivery of the same dose of doxorubicin-loaded nanoparticles is much more effective than one-step both in terms of diagnostics and therapy.