PLGA-Alendronate Conjugate as a New Biomaterial to Produce Osteotropic Drug Nanocarriers

Targeting bone tissue to treat specific diseases, such as primary tumour or bone metastases from peripheral malignancies is a current and pressing research field for modern pharmaceutical technology, and nanomedicine in particular. Bone is the third most common site of metastasis and the incidence of bone metastases in patients died of cancer is reported to be around 70% [1,2]. Primary carcinomas of the lung, breast, prostate, kidney, and thyroid may develop skeletal metastases. In children affected by disseminated neuroblastoma bone metastases are also highly frequent. Tumour osteolysis is responsible for pathologic fractures, intractable pain, nerve compression syndrome, and severe hypercalcemia. In addition, patients with bone metastases may have neurologic impairment from spinal lesions, anaemia, and complications exacerbated by immobilization. Bone matrix resorption is minimally due to direct cancer cell activity. In fact, bone-colonizing tumour cells also stimulate osteoclastmediated bone resorption via the secretion of potent osteolytic agents [3]. The increased bone resorption that follows releases bone-derived growth factors into the extracellular milieu and systemic circulation, thereby further enhancing bone resorption, promoting tumour growth, and altering the tumour microenvironment. Vice versa, metastatic cancer cells are largely influenced by messages embedded within the bone matrix [4]. Under the influence of bone microenvironment, tumour cells proliferate and interact with osteoblasts and osteoclasts leading to lytic or sclerotic lesions. Clinical management of metastatic bone disease is really hard. The probability of long-term survival decreases dramatically in patients with skeletal metastases. Current treatments are based both on systemic therapy (chemotherapy, immunotherapy, hormone therapy) [5,6] and local therapy (surgery and radiotherapy) [7,8]. The restoration of functionality and mobility, along with pain relief can improve patients’ quality of life, but is unable to affect the negative progression of the disease. Because of the unique characteristics of tumours growth within bone tissues, conventional therapeutic strategies often lack efficiency in cure bone metastases. Moreover, most of the drugs used for the adjuvant therapy of osteolytic metastases interfere with signalling that induce osteoclast differentiation and osteoclast-mediated bone resorption, but they are not effective on tumour growth in the bone and their use is obviously associated with heavy side effects.


Introduction
Targeting bone tissue to treat specific diseases, such as primary tumour or bone metastases from peripheral malignancies is a current and pressing research field for modern pharmaceutical technology, and nanomedicine in particular.Bone is the third most common site of metastasis and the incidence of bone metastases in patients died of cancer is reported to be around 70% [1,2].Primary carcinomas of the lung, breast, prostate, kidney, and thyroid may develop skeletal metastases.In children affected by disseminated neuroblastoma bone metastases are also highly frequent.Tumour osteolysis is responsible for pathologic fractures, intractable pain, nerve compression syndrome, and severe hypercalcemia.In addition, patients with bone metastases may have neurologic impairment from spinal lesions, anaemia, and complications exacerbated by immobilization.Bone matrix resorption is minimally due to direct cancer cell activity.In fact, bone-colonizing tumour cells also stimulate osteoclastmediated bone resorption via the secretion of potent osteolytic agents [3].The increased bone resorption that follows releases bone-derived growth factors into the extracellular milieu and systemic circulation, thereby further enhancing bone resorption, promoting tumour growth, and altering the tumour microenvironment.Vice versa, metastatic cancer cells are largely influenced by messages embedded within the bone matrix [4].Under the influence of bone microenvironment, tumour cells proliferate and interact with osteoblasts and osteoclasts leading to lytic or sclerotic lesions.Clinical management of metastatic bone disease is really hard.The probability of long-term survival decreases dramatically in patients with skeletal metastases.Current treatments are based both on systemic therapy (chemotherapy, immunotherapy, hormone therapy) [5,6] and local therapy (surgery and radiotherapy) [7,8].The restoration of functionality and mobility, along with pain relief can improve patients' quality of life, but is unable to affect the negative progression of the disease.Because of the unique characteristics of tumours growth within bone tissues, conventional therapeutic strategies often lack efficiency in cure bone metastases.Moreover, most of the drugs used for the adjuvant therapy of osteolytic metastases interfere with signalling that induce osteoclast differentiation and osteoclast-mediated bone resorption, but they are not effective on tumour growth in the bone and their use is obviously associated with heavy side effects.167 [11,12,18,19].Bisphosphonates (BP) have been for instance conjugated to small drugs [20,21], and proteins [22] with the aim at optimizing the treatment of osteoporosis, osteoarthritis, and bone cancer, and to radiopharmaceuticals to obtain novel agents for bone scintigraphy [23].The conjugation of bisphosphonates to polymer backbones has been studied as a mean for bone targeting [24,25].Recently, cholesteryl-trisoxyethylenebisphosphonic acid (CHOL-TOE-BP), a new tailor-made BP derivative, has been used as a bone targeting moiety for liposomes [26].In other works, the amino-bisphosphonate drug alendronate (ALE; Fosamax ® ) was co-conjugated to HPMA polymer backbones together with an anticancer agent.Thereby, passive targeting was achieved by extravasation of the nanoconjugates from the tumour vessels via the EPR effect, while active targeting to the calcified tissues was achieved by ALE affinity to HA [27].
Fig. 1.The strategies described in the literature to achieve targeting to bone tumours: 1) an anticancer drug (D) is encapsulated/loaded into a nanocarrier (e.g., polymeric or lipid nanoparticles, liposomes, micelles, etc.) and passive targeting is expected (EPR effect); 2) a biologically active molecule is covalently linked to an osteotropic targetor moiety (e.g., a bisphosphonate, BP); 3) nanoparticles are made by an osteotropic polymer (e.g., a polymer-BP conjugate) and then loaded with a bioactive agent: in this way, both passive and active targeting possibilities can be achieved.

The scientific and technical rationale for a new targeting strategy
Targeted DDS should be preferable over drug-BP conjugates due to different factors, such as drug protection from biodegradation in the bloodstream, transport duration, and drugpayload.Thereby, in a recent and partially still on-going research, the two strategies depicted in Figure 1 have been merged, leading to an innovative solution for active targeting of drugs to the bone.The working hypothesis was to realize a biocompatible nanocarrier showing high affinity to bone (i.e., an osteotropic nanocarrier), which can be loaded with different classes of drugs active against bone diseases, such as anticancer, antiangiogenic, antibiotics, or anti-osteolytic agents.To this aim, in a first step a new polymeric biomaterial showing osteotropic properties was produced through the conjugation of a poly(lactide-co-glycolide) (PLGA) to the BP agent alendronate (ALE).PLGA copolymers are diffusely used biocompatible and biodegradable materials for controlled drug release systems [28], including anticancer agents, and they have also been recognized as GRAS by US FDA.In this study, a copolymer made of 50:50 polylactic-co-glycolic acid was used (Resomer ® RG 502 H; Figure 2) because of the presence of free carboxyl end groups susceptible to chemical derivatisation.The typical weight composition of this copolymer allows it to remain in the human body enough to let the bound BP moiety to recognize the bone proteins.ALE (4-amino-1-hydroxybutyldiene-1,1-phosphoric acid, Figure 2) is an amino-BP, approved for the treatment and prevention of osteoporosis, treatment of glucocorticoidinduced osteoporosis in men and women, and therapy of Paget's disease of bone [29].ALE was selected among BP agents because of some peculiar properties: a) the presence of a free amine group, able to create a stable covalent link with the carboxyl group present in the used PLGA; b) such amine group is not essential for the interaction of ALE with HA, and thus for its pharmacological effects, but exerts only a supportive role [30]; c) the possibility of easily converting the commercial sodium salt of ALE into its free acid form.The choice of an amide bond between the targeting BP and the polymer was also driven by the known relatively high resistance to enzymatic hydrolysis in plasma of this linkage, that should ensure the intact conjugate to reach the target (bone) tissues.The haemo-and cytocompatibility of the PLGA-ALE conjugate was confirmed in vitro.Therefore, a nanoparticle system (NP) was produced starting from this new biomaterial.The NP were studied for their technological properties, as well as for their biocompatibility.Finally, PLGA-ALE NP were then loaded with a model cytotoxic drug, doxorubicin (DOX) and tested in vitro and in vivo.

Synthesis and characterization of the PLGA-ALE conjugate
Polyl(D,L-lactic-co-glycolic acid) (50:50) containing a free carboxylic acid end group [Resomer ® RG 502 H; inherent viscosity: 0.16-0.24dl/g (0.1% in chloroform, 25°C)] was used (Boehringer Ingelheim, Ingelheim am Rhein, Germany).Sodium alendronate was converted into the acid form by treatment with aqueous acetic acid and lyophilisation.The conjugate was synthesized by two alternative methods, i.e., carbodiimide-assisted direct conjugation or preparation of an activated intermediate through N-hydroxysuccinimide [31].In the first case, a solution of Resomer ® RG 502 H in DMSO and dichloromethane (DCM) (1:1) was activated at 0°C for 2 h by N'-(3-dimethylaminopropyl)-N-ethyl carbodiimide hydrochloride (EDAC), in the presence of 1-hydroxy-benzotriazole (HOBt) and triethylamine.Alendronic acid was dissolved in DMSO and added to the reaction mixture, which was stirred for 2 h at 2°C and then at r.t. for 8 h.The solvent was partially removed under vacuum and the remaining solution was purified by dialysis water (CelluSep H1 MWCO 2000; M-Medical s.r.l., Cornaredo, Italy).In the alternative procedure, the PLGA was previously activated with N-hydroxysuccinimide (NHS) and dicyclohexylcarbodiimide in anhydrous dioxane at 15°C under stirring for 3 h [32].The formed dicyclohexylurea was filtered off and the solution was poured in anhydrous diethyl ether.The solvent was decanted and the oily residue was purified by dissolution in anhydrous dioxane and precipitation with anhydrous diethyl ether (3 times), and finally dried in vacuo.A solution of NHS-PLGA in anhydrous DMSO was treated with triethylamine and sodium alendronate under stirring at r.t. for 12 h.At the end of the reaction the solvent was partially removed under vacuum and the remaining solution was purified by dialysis.The dialysed samples were frozen into using liquid nitrogen and lyophilised.Both methods resulted in similar production yields (70-75%) and purity of the final conjugate, therefore the first method could be better proposed as more direct and simple.The chemical structure of the PLGA-ALE conjugate was confirmed by MALDI-TOF MS and 1 H-NMR; analytical details are available [31].In the view of using the PLGA-ALE copolymer to prepare bone-targeted NPs, the conjugate was evaluated for blood and cyto-compatibility, to individuate any negative effect which might have precluded any further biological investigation.Haemolysis was evaluated because erythrocytes are among the first cell lines that come into contact with injected materials.Experimental results did not show haemolytic effects of the conjugate; the plasmatic phase of coagulation was measured by the activated partial thromboplastin time (APTT) and prothrombin activity; they respectively evaluate the intrinsic and extrinsic phases of coagulation.In basal conditions, prothrombin activity was 140.2 ± 2.5 and APTT was 35.6 ± 0.3.In plasma incubated with PLGA-ALE at different dilutions, prothrombin activity and APTT were not significantly different from the plasma incubated with PBS (Figure 3A and 3B).DMSO at the same dilutions did not affected either prothrombin activity or APTT.In the bloodstream, nanoparticles come in contact with endothelium before passing through the vessel wall and reaching tissues.Therefore, the effect of PLGA-ALE on endothelial cells was tested to verify the lack of cytotoxicity.As expected, PLGA-ALE was not cytotoxic for human umbilical vein endothelial cells (HUVEC), as proven by the neutral red test (Figure 3C).Absence of cytotoxicity was also shown in cultures of human primary trabecular osteoblasts (BMSC) (Figure 3D).
In conclusion, PLGA-ALE conjugate did not cause either haemolysis on human erythrocytes, or alterations of the plasmatic phase of coagulation or cytotoxic effects on endothelial cells and trabecular osteoblasts [31].

PLGA-ALE nanoparticles
NP were produced by the PLGA-ALE conjugate by a nanoprecipitation method, using an opportunely adapted emulsion/solvent evaporation technique [33].The conjugate was dissolved in acetone, DMSO or their 1:1 mixture (v:v).The organic solution was added drop wise into phosphate buffered saline (PBS), pH=7.4,containing Pluronic F68.After stirring at r.t. for 10 min, the solvent was partially removed at 30 °C under reduced pressure and the concentrated suspension was purified by dialysis against water.In Figure 4 the size analysis of the formed systems, obtained by photon correlation spectroscopy is illustrated.DMSO was used in the NP production due to the low solubility of the conjugate in acetone (from which low production yields were achieved), but it did not appear to be an ideal solvent; in fact, NP obtained from pure DMSO solution showed a higher mean size with respect to those obtained using an acetone/DMSO mixture.These NP showed a homogeneous distribution, with an average size of 198.7 nm and a polidispersity index of about 0.3.However, in all cases mean sizes between 200-300 nm were obtained, an interesting feature for a further development of these systems as injectable drug carriers.Also a dialysis method using a DMSO solution of the conjugate was attempted to prepare the NP [34], but it resulted in much larger particles (around 400 nm) (not shown).Fig. 4. Size distribution patterns of PLGA-ALE NP prepared using different organic phases (adapted from [31]. The PLGA-ALE NPs showed a net negative surface charge (-potential of -38 mV), close to the value given by the NPs obtained from pure PLGA (-41.8 mV).SEM analysis revealed spherical particles with a smooth surface (Figure 5).The system was also shown to be sterilisable by gamma irradiation at 10 kGy, showing only minimal particle size changes [33].This finding prospects the possibility of using sterilized NPs for further in vivo studies of the system.To assess the affinity for HA, PLGA-ALE and PLGA NPs were loaded with a lipophilic probe (Red Oil O) and incubated with two different HA concentrations (1 and 5 mg/ml) for 15 or 30 min.PLGA-ALE NPs showed a relative increase of affinity towards the phosphate salt compared to pure PLGA NPs (Figure 6).The affinity increased with the incubation time and was proportional to the concentration of HA in the suspension.This latter observation would suggest that some form of chemical interaction occurred between PLGA-ALE and HA, reinforcing the mere physical absorption of the phosphate on NP surface, which justifies the affinity already measured for pure PLGA NPs.The biocompatibility profile of PLGA-ALE NPs was assessed by means of several in vitro assays, able to demonstrate their effects on biological systems which use to come in contact with a material injected into the body [33,35,36].The use of NPs for drug delivery necessitates an accurate assessment of their biocompatibility [37,38].For their nanoscale size, NPs may have a reduced blood compatibility in comparison with the starting material: even if the biocompatibility of a macromolecule is well-established, the enormous increase of its surface when in the form of NPs may bring on negative effects that are not given by the bulk material.

Biocompatibility studies
Blood-biomaterial interactions are complex events that involve erythrocyte and leukocyte damage, and activation of platelet, coagulation, and complement.Damage of red blood cells, as well as complement activation may favour haemolysis and other forms of general toxicity.Analogously, NPs should not activate platelets and the plasmatic phase of coagulation, nor damage endothelial cells to avoid thrombogenesis [39,40].In the meantime, they should not reduce the levels of the plasmatic factors of coagulation, to prevent haemorrhagic accidents.Finally, with the aim of bone targeting, when NP will 0 20 As a further proof of their compatibility with blood components, at all the tested concentrations PLGA-ALE NPs did not induce any significant effect either on the total leukocyte number or on their subpopulations percentage.Analogously, the NPs did not cause platelet adhesion or activation (release reaction), as assessed by the platelet factor 4 measurement (Figure 7), all process that may induce thrombotic phenomena after NP injection [33].To assess the lack of alterations of blood proteins, we evaluated the effects of PLGA-ALE NPs on the plasmatic phase of coagulation and on complement.A decrease of the coagulation factors levels favours haemorrhage, while an increase of their activity may induce thrombotic phenomena.In our experiments, only the highest tested concentration of NPs caused a decrease of the prothrombin activity, while all the lower concentrations increased it, with changes however always ranging within normal physiological values (Figure 8).Since APTT (i.e., the intrinsic and the common phase of coagulation) was also not significantly affected by the NPs (not shown), the observed changes could be due to an alteration of Factor VII of the extrinsic pathway of coagulation.Probably, Factor VII was adsorbed on the NPs surface at high concentration and was less available for coagulation; conversely, lower concentrations of PLGA-ALE NPs could activate factor VII, in a way similar to the effect of tissue factor.Another hypothesis could involve the activation of Factors XII and XI by the NPs, similarly to silica gel or glass (cf.[33] for more details).Activation of complement induces the production and release of small biologically active peptides.C3a and C5a are chemo-attractive for leukocytes and favour their aggregation.Furthermore, C5a induces the adhesion of granulocytes and monocytes to the endothelium, their migration into the external tissues, enzyme release and production of pro-coagulant or platelet aggregating compounds.C3a and the C5b-9 complex can directly activate platelets.Therefore, complement activation should not be considered as a strictly local phenomenon, but results also in systemic effects.Complement activation may occur by the classical and the alternative pathways.In the classical pathway, the protein C1q recognises activators (usually immune complexes) and binds to them.The activation via the alternative pathway starts by the binding of C3b to the activator surface (such as microbial polysaccharides or lipids or surface antigens present on some viruses, parasites and cancer cells) and then follow the same events of the classical pathway.NPs predominantly activate the complement through the alternative pathway.In our experiments PLGA-ALE NPs did not activate complement by none of the two pathways, as shown by a non-significant different complement consumption from PBS (Figures 9A and 9B).Conversely, the positive control zymosan induced a 70% complement consumption, twice the amount consumed by PBS.Also the Bb fragment, which is produced during the complement activation by the alternative pathway, was not significantly affected by the PLGA-ALE NPs (not shown).In the bloodstream, NP come rapidly in contact with vessel endothelium before passing to the outer tissues; therefore, absence of damage to endothelial cells must be ensured.Moreover, bone oriented NPs should not affect the vitality and function of normal osteoblasts.The cytotoxicity of the prepared PLGA-ALE NPs was excluded on both endothelial (HUVEC) cells and osteoblasts derived by bone marrow stromal cells (BMSC).Cell viability was always higher than 80% upon 24 h-exposure to the various concentrations of NPs or to PBS.Phenol, used as a positive control, reduced the cell viability to 19.0% (HUVEC) and to 27.5% (BMSC) (Figure 10).In conclusion, PLGA-ALE NPs did not affect platelets, leukocytes and complement, did not induce haemolysis and did not exert cytotoxic effects on endothelial cells and osteoblasts [33].To assess if the PLGA-ALE NPs are able to retain the antiosteoclastic properties of the bisphosphonate, osteoclast cultures obtained from human peripheral blood mononuclear cells (PBMC) were incubated with either PLGA-ALE or pure PLGA NPs, at an equivalent concentration of 0.64 M or 6.4 M of ALE; free ALE was tested as a positive control.Experiments showed that ALE retained the ability of inhibiting the osteoclast-mediated degradation of type I human bone collagen, determined a dose-dependent reduction of osteoclast number, and induced apoptosis in osteoclast cultures, also when conjugated with the copolymer PLGA and in the form of NPs [41].Interestingly, pure PLGA NPs also showed similar effects, and this can be considered an useful phenomenon in the view of the overall aim of this study: since the conjugation of ALE to PLGA, and the resulting NP formation was aimed at targeting antitumor drugs to osteolytic bone metastases, the additional antiosteoclastic effect observed for polymer-bound ALE and also of PLGA could even contribute to the inhibition of the associated osteolysis.

Drug-loaded nanoparticles
Doxorubicin (DOX) is an anticancer agent of wide clinical use, from leukaemia and Hodgkin's lymphoma to symptomatic metastatic breast cancer, from neuroblastoma to many other cancers (prostate, thyroid, bladder, stomach, lung, ovary).Its therapeutic applications are nevertheless limited by the strong cardiac and bone marrow toxicity.DOX effectiveness has been greatly improved when specific targeting at the tumour sites has been achieved, for instance by loading the drug into liposomes (e.g., Caelyx ® /Doxil ® ) [42,43].DOX was loaded in the previously described PLGA-ALE NPs and the anti-tumour effect of the carrier was assessed in vitro and in vivo [44].In a first instance, the intracellular accumulation and distribution of DOX-loaded NPs was assessed by fluorescence and confocal microscopy, in comparison with free DOX.A panel of potential target cells was used in these experiments: MDA-MB-231 and MCF7 breast adenocarcinoma cell lines, Saos-2 and U-2 OS osteosarcoma cell lines, SH-SY5Y neuroblastoma cell line, and ACHN renal adenocarcinoma cell line.The above tumour histotypes were chosen because all originate from or can metastasize to the bone.The final localization of DOX in cell nucleus is important because of it mechanism of action [45].We observed that the incubation of free DOX with cells resulted in its accumulation in the nuclei, while it was absent in the cytoplasm (Figure 11).Conversely, cells treated with DOX-loaded NPs showed also fluorescent spots localized in the cytoplasmic vacuoles.Polymeric NPs typically accumulate in lysosomal vesicles, from which the drug is released into the cytoplasm [46,47].Therefore, it is likely that the observed cytoplasmic fluorescence was due to the same phenomenon, and that the DOX-loaded NPs were trafficked through the endo-lysosomal compartment.The incubation of the above cell lines with DOX-loaded PLGA-ALE NPs (48 and 72 h) gave a cell growth inhibition profile very similar to that one of the free drug [44].To evaluate the in vivo activity, DOX-loaded NPs were injected into a mouse model of breast cancer bone metastases: osteolytic lesions were induced by intratibial inoculation of the human breast carcinoma cells MDA-MB-231 that can induce prominent bone metastases.Histological analysis confirmed that the NP did not cause any organ abnormality, hence they should not exert any systemic cytotoxic effect.Control mice treated with PBS developed pronounced  Mice were weekly treated for six weeks with PBS, free DOX, unloaded NPs, or drug-loaded NPs (DOX-NP) at the dose of 0.2 or 1 mg/kg.A, X-ray of hind limbs at day 42 of MDA-MB-231-injected mice and treated with PBS, free DOX, unloaded NPs or DOX-NP at an equivalent drug dose of 0.2 mg/kg (arrows: osteolytic areas).B, incidence of osteolytic bone metastases after the same treatments at the lower (left panel) and higher dose (right panel).Means ± SE, n = 8 (modified from ref. [44]).osteolytic lesions detectable by X-ray analysis starting from the 28th day after the tumour cell inoculation (Figure 12A).The free drug was effective in retarding the onset of metastases and reducing their incidence compared to the animals treated with PBS or blank NPs (Figure 12B).However, a significant effect was observed only in the DOX-loaded NPtreated group at the dose of 0.2 mg/kg (Figure 12B, left panel; P= 0.028 vs. vehicle).The extension of tumour size was significantly smaller in animal groups which received both the free drug and DOX-loaded NPs at a drug concentration of 0.2 mg/kg or 1 mg/kd (Figure 13A), whereas unloaded NPs were ineffective.A trend of reduction of the osteolytic areas was measured in mice treated with either free DOX, DOX-loaded NPs or unloaded NPs.However, DOX-loaded NPs induced a higher inhibition of osteolysis than unloaded Fig. 13.Effects of DOX-loaded NPs on tumour area and osteolysis.Mice were treated with either PBS, free DOX, unloaded NPs, or DOX-loaded NPs (DOX-NP) at 0.2 mg/kg (dose 1) or 1 mg/kg (dose 2).At the end point, the extension of the osteolytic area was determined by X-ray analysis; sections of the tibiae were histologically examined to evaluate tumour area and the number of osteoclasts on the bone surface.A, tumour area (mean ± SD); B, osteolytic areas quantified on X-ray images of mice treated with dose 1 (mean± SD); C, number of osteoclasts found on the bone surface (OcN/BS) in TRAP stained sections of the tibiae (mean ± SD). * P < 0.05, ** P < 0.005 (modified from ref. [44]).NPs (0.090±0.009 vs. 0.310±0.040mm 2 , P = 0.033); both free and NP-loaded DOX reached a similar effect at the end of the experiment (day 42) (Figure 13B).Similarly, the histomorphometric analysis of sections stained for TRAP activity showed a trend of reduction for the number of osteoclasts found at the bone surface by either free DOX, unloaded NPs, and DOX-loaded NPs (Figure 13C).However there was a significant inhibition only for unloaded NPs and DOX-loaded NPs at the highest dose (P= 0.04 and P = 0.014 vs. vehicle, respectively).In conclusion, loading of DOX in the osteotropic PLGA-ALE did not affect the drug efficacy on bone metastases formation; the reduction of the incidence of metastases induced by DOX-loaded NPs was significantly higher than that allowed by the free drug.The enhanced efficacy of the drug when loaded in these NPs can be used as an indirect demonstration of the successful delivery of DOX to the bones.The modest activity registered with unloaded PLGA-ALE NPs could be related to the inhibitory effect of ALE conjugated to PLGA on osteoclast activity, that indirectly reduced tumour expansion, and of a direct effect exerted by ALE on tumour cells [48].Treatment of m i c e w i t h b o t h f r e e o r N P -l o a d e d D O X significantly reduced also the tumour area.All the above experimental findings provide in vitro in vivo evidences of the effectiveness of a new osteotropic delivery system for DOX, and possibly other antitumour agents, in which a synergism between the antineoplastic activity of the drugs and the antiosteolytic activity of ALE can afford a better inhibition of tumour development and progression.In the meantime, loading of DOX in biodegradable and biocompatible NPs, made from a conjugate between PLGA and ALE, can allow a site-specific delivery of the drug to osteolytic areas, possibly reducing its systemic side effects.A potential tool for the development of innovative regimens for metastatic bone diseases can thus be exploited by this novel nanocarrier.

Fig. 3 .
Fig. 3. Biocompatibility results of the PLGA-ALE conjugate: (A) mean prothrombin activity and (B) APTT of human plasma incubated with different concentrations of the conjugate; (C) viability of human umbilical vein endothelial cells (HUVEC) and (D) human primary osteoblasts from trabecular bone, respectively, after incubation with the conjugate.No haemolytic activity was given by the tested compound (not shown) (adapted from ref. [31]).

Fig. 6 .
Fig. 6.Affinity of PLGA-ALE and PLGA NPs for hydroxyapatite (HA).Oil Red O-loaded NP suspensions were incubated at r.t. for either 15 or 30 min with an aqueous suspension containing 5 mg/ml of the phosphate salt.Results are expressed as the percentage decrease of light absorbance at 523 nm with respect to the corresponding NPs incubated without HA [cf.ref. 33].

Fig. 9 .
Fig. 9. Percentage consumption of human serum complement activity via the classical pathway (A) or the alternative pathway (B), after incubation with the PLGA-ALE NPs (adapted from ref. [33]).

Fig. 12 .
Fig. 12.Effect of DOX-loaded NPs on the incidence of osteolytic bone metastases in vivo.BALB/c-nu/nu mice were injected intratibially with a suspension of MDA-MB-231 cells.Mice were weekly treated for six weeks with PBS, free DOX, unloaded NPs, or drug-loaded NPs (DOX-NP) at the dose of 0.2 or 1 mg/kg.A, X-ray of hind limbs at day 42 of MDA-MB-231-injected mice and treated with PBS, free DOX, unloaded NPs or DOX-NP at an equivalent drug dose of 0.2 mg/kg (arrows: osteolytic areas).B, incidence of osteolytic bone metastases after the same treatments at the lower (left panel) and higher dose (right panel).Means ± SE, n = 8 (modified from ref.[44]).

Table 1 .
Hemolytic activity (arithmetic mean ± S.E. of 6 experiments) of PLGA-ALE NPs.Their concentration was expressed as the amount of ALE in each sample.