- Email: [email protected]
Intraarticular application of superparamagnetic nanoparticles and their uptake by synovial membrane—an experimental study in sheep
ARTICLE IN PRESS
Journal of Magnetism and Magnetic Materials 293 (2005) 419–432 www.elsevier.com/locate/jmmm
Intraarticular application of superparamagnetic nanoparticles and their uptake by synovial membrane—an experimental study in sheep Katja Schulzea, Annette Kochb, Bernhard Scho¨pfa, Alke Petric, Benedikt Steitzc, Mathieu Chastellainc, Margarethe Hofmannd, Heinrich Hofmannc, Brigitte von Rechenberga, a
Musculoskeletal Research Unit, Equine Hospital, Vetsuisse Faculty, University of Zurich, Winterthurerstrasse 260, 8057 Zurich, Switzerland b Department of Chemistry and Applied BioSciences, Swiss Federal Institute of Technology Zurich (ETH Zurich), Winterthurerstrasse 190, 8057 Zurich, Switzerland c Laboratory of Powder Technology, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland d MatSearch, Chemin Jean Pavillard 14, 1009 Pully, Switzerland Available online 28 March 2005
Abstract A superparamagnetic iron oxide nanoparticle, coated with polyvinyl alcohol, (PVA-SPION) and its fluorescently functionalized analogue (amino-PVA-Cy3.5-SPION) were compared in vivo as proof of principle for future use in magnetic drug targeting in inflammatory joint diseases. They were injected either intraarticularly or periarticularly and their uptake by cells of the synovial membrane was evaluated. Uptake was completed in 48 h and was enforced by an extracorporally applied magnet. r 2005 Elsevier B.V. All rights reserved. Keywords: Magnetic drug targeting; Sheep; Knee joint; Inflammation; SPION; Intraarticular medication; Osteoarthritis; Fluorescent nanoparticles; Cy3.5; Biocompatibility; Synovial cells; Phagocytosis; PVA
1. Introduction Chronic aseptic inflammatory diseases of the joint such as rheumatoid arthritis or osteoarthritis Corresponding author. Tel.: +41 1 6358 410; fax: +41 1 6358 950. E-mail address: [email protected] (B. von Rechenberg).
require long-time therapy of patients with analgesic, anti-inflammatory, immune-modulating or chondroprotective drugs to release pain and modulate the degree of disease symptomatically [1]. Prolonged medication periods are commonly associated with negative side effects of the drugs, such as nephrotoxicity, hepatitis or gastrointestinal ulceration leading to forced discontinuation of
0304-8853/$ - see front matter r 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2005.02.075
ARTICLE IN PRESS 420
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
the medication [2]. In the last years several attempts were made to find new successful therapy strategies reducing the unwanted side effects while at the same time leading to optimal well being of the patients [3–5]. If possible, replacement of systemic drug administration through intraarticular injections is the most effective route to treat joint diseases [6]. However, the efficiency of the injected drug may be limited through a relatively short drug persistence or failure to maintain adequate drug concentrations in the joint cavity [7–9]. Different approaches are made to optimize drug persistence and efficiency in the joint either with biodegradable drug carriers, conjugated drug formulations or targeted drug systems [10–13]. Of those, superparamagnetic iron oxide nanoparticles (SPIONs) seem highly interesting for clinical applications because of their superparamagnetic properties and their possible use in magnetic drug targeting including gene therapy [14,15]. Up to this time of writing, the use of SPIONs for joint therapy has not been reported. The goal of this study was to determine whether intraarticularly (i.art.) or periarticularly (p.art.) injected SPIONs with PVA coatings are biocompatible and successfully taken up by synovial cells in vivo. In addition, the effect of an extracorporally applied magnet on nanoparticle persistence within the synovial membrane was tested. Last but not least, a fluorescent molecule, Cy3.5, was covalently attached to the surface of the nanoparticles for (i) tracking the nanoparticles within the tissues and at the same time (ii) acting as a model cargo to evaluate nanoparticle stability, performance and biocompatibility with respect to drug targeting applications after i.art. injection using sheep as an experimental model.
2. Materials and methods 2.1. Preparation of nanoparticles Polyvinyl alcohol (PVA; (Mowiols 3-83, average molecular weight Mw: 14 000 g/mol, hydrolysis degree: 83% supplied by courtesy of CLARIANT)) stabilized iron oxide nanoparticles were produced as described elsewhere [16]. For this
work, 10 wt% of the PVA was exchanged by vinyl alcohol/vinyl amine copolymer (M12, average molecular weight of 80 000–140 000 g/mol, supplied by courtesy of Erkol, (E)). To limit the number of amine group per nanoparticle, a vinyl alcohol/vinyl alcohol copolymer mass ratio of 45 was chosen. The overall ratio by weight of iron to polymer content was for all samples fixed at 13. The concentration of the amino-PVA functionalized nanoparticles (amino-PVA-SPIONs) in aqua’s solution was 4.6 mg Fe/ml. The results of a more detailed characterization of the nanoparticles are given elsewhere [17]. Amino-PVA-SPIONs were further derivatized by covalent coupling of a fluorescent dye to the nanoparticles via the polymer. For this work Cy 3.5: derivatized CyDyeTM (NHS ester; Eurogentec) was chosen that carries one reactive group on each dye molecule for accurate labelling of amine groups. The absorption maximum of this dye is at 581 nm, the maximal emission at 596 nm.The final dispersion contains 0.3 mg iron/ml dispersion and approx. 3 dye molecules/nanoparticle. 2.2. Experimental animals and surgical technique A total of 14 Swiss Alpine sheep between 2 and 4 years of age was used for application of PVASPIONs and amino-PVA-Cy3.5-SPIONs into the stifle and carpometaphalangeal joints. Eight additional sheep without SPIONs as well as without magnets and slaughtered for other reasons than infectious or systemic disease served as controls. Two main experiments were conducted with the nanoparticles, such that in experiment 1, PVASPIONs were injected either intraarticularly (i.art.; into the joint cavity) or periarticularly (p.art.;into joint capsular tissue) and in experiment 2, aminoPVA-Cy3.5-SPIONs were injected only i.art. The sheep were checked for their overall health, routinely wormed as well as vaccinated. After an adaptation period of 14 days prior to surgery, they were kept in groups in a stall, routinely fed with hay and with free access to water. All experiments were conducted under general anesthesia. The animals were fasted 24 h before sedation with medetomidine (5 mg/kg, i.m., Domitors, OrionFarmos, Turku, Finland) and induction with
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
ketamine (2 ml/kg Narketans, Chassot AG, Bern, Switzerland) in combination with valium i.v. (Diazepam, 0.01 mg/kg, Roche, Basel, Switzerland). Anesthesia was maintained with isoflurane (Forenes, Abbott AG, Baar, Switzerland) in 100% oxygen. Analgesia was achieved using buprenorphine (Temgesics 0.03 ml/kg i.v., ESSEX Chemie AG, Luzern, Switzerland) peri- and postoperatively every 4 h as needed. The skin of both carpal and stifle joints of the sheep were aseptically prepared in a surgical routine manner. Intraarticular injections: At the carpometaphalangeal joint 1.0 ml of the nanoparticle suspension was injected into the synovial cavity using a surgical sterile technique, while in the stifle joint 2.0 ml of the nanoparticle suspension were administered into the lateral joint cavity. Periarticular injections (only experiment 1): The nanoparticle suspension was injected into the capsular and subdermal connective tissue at the lateral aspect of the joints. There, 1.0 ml was placed at the carpal joint region and 1.5 ml at the stifle. 2.3. Extracorporal magnets NdFeB-permanent magnets possessing a remanence field of 1.3 T (Neomag S, Magnequench, Lupfig, Switzerland) were placed in pockets of specially designed bandages that were sutured to the skin (2–0 Supramid, B.Braun, Tuttlingen, Germany) at the craniolateral aspect of the joints immediately after injection of the nanoparticle suspension (Fig. 1). The magnets were removed 12 h after injection. 2.4. Group distribution (see Table 1) Experiment 1: For the application of PVASPIONs eight sheep were randomly split into two groups with four animals each. The first group was used for intraarticular and the second for periarticular injections. Of each group there was one sheep sacrificed at 3, 24, 72 and 120 h after surgery. Experiment 2: For the amino-PVA-Cy3.5SPIONs, six sheep were used and only i.art. injections were performed. The sheep were split into two groups with three sheep each. The first
421
Fig. 1. Sheep with extracorporal magnet sutured to the skin of the carpal and stifle joint after surgery.
group was treated with an extracorporal magnet at the craniolateral aspect of both carpal and stifle joints using the same protocol as before. The second group was left without a magnet. The sheep were sacrificed after 24, 72 and 120 h. Both experiments were conducted according to the Swiss regulations of animal welfare and protection and authorized by the local authorities and Ethical Committee (application No. 59/02).
2.5. Macroscopic evaluation of synovial tissue After sacrifice of the experimental and control animals the front and hind limbs of the animals were collected immediately. The regions of the carpometaphalangeal or stifle joint were exposed through removing the overlying skin. The subcutaneous and periarticular region was judged for yellow–brownish discolouration indicating presence of iron containing nanoparticles and signs of inflammation (reddish, edematous tissue). Thereafter, the joint capsules were opened to expose the synovial membranes, synovial fluid and the cartilage surfaces that were also checked for signs of degradation as well as discolouration due to nanoparticle injection. Cartilage degradation is recognized if there is a cobble-stone appearance of the cartilage surface present as well as fibrillation, cleft formation and/or erosion
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
422
Table 1 Group distribution of animals in experiment 1 and 2 Sheep
Joints
Sacrifice
Injection
Time in h
of SPION
left carpus
left stifle
right carpus
right stifle
Experiment 1
1 2 3 4 5 6 7 8
3 3 24 24 72 72 120 120
i.art. p.art. i.art. p.art. i.art. p.art. i.art. p.art.
+ + + + + + + +
+ + + + + + + +
Experiment 2
1 2 3 4 5 6
24 24 72 72 120 120
i.art. i.art. i.art. i.art. i.art. i.art.
+ + +
+ + +
+ + +
+ + +
Controls
8 animals
none
with magnet: +, without magnet:
down to the subchondral bone. A digital camera was used to document the results. 2.6. Histological evaluation Synovial membranes including capsular tissue as well as cartilage samples were harvested. In experiment 2 samples from the lateral and the medial aspects of the joints were taken. Tissue probes were embedded and processed for standard paraffin histology. Slides of the paraffin blocks were stained with Haematoxylin–Eosin (H.E.) and Pearl’s Prussian blue (Fe3+) using histological standard protocols of our laboratory. Evaluation of biocompatibility and presence of nanoparticles was performed using light microscopy (DMR, Leica, Glattbrugg, Switzerland) and digital imaging (DC 200, Leica). Sections without stains were used for fluorescence microscopy. To perform confocal microscopy nuclei were counter-stained with Hoechst 33342 (3 mg/ml PBS, Molecular Probes, Bedford, MA, USA). There, the slides were incubated with Hoechst 33342 for 30 min at 37 1C, thereafter washed with ddH2O and covered using Dakos Fluorescent Mounting Medium (DAKO Corporation, Carpinteria, CA, USA).
Preliminary evaluation was performed with an inverse microscope (Axiovert 35, Zeiss, Oberkochen, Germany) equipped for fluorescence microscopy. CLSM (confocal laser scanning microscopy, Zeiss 410 inverted microscope, Zurich, Switzerland) was used for final evaluation and confocal microscopy. 2-D multi-channel image processing was performed using the IMARIS software (Bitplane AG, Zurich, Switzerland). Background fluorescence of cells was determined by analysing untreated cells. 2.7. Semiquantitative evaluation The intima (layer of synovial cells facing the synovial cavity), fibrous and adipose connective tissues (both underlying the intima as part of the joint capsule) of the synovial membrane was evaluated separately. Evaluation of nanoparticle localization and concentration was done using slides stained with Pearl’s Prussian blue, whereas biocompatibility was examined using those stained with H.E. A scoring system was developed to assess nanoparticle localization, nanoparticle concentration and inflammatory response (Table 2). Two different types of scores with a range of 0–3
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
were used. Type a score represented nanoparticle concentration as well as nanoparticle localization within the tissue. The higher the scores, the more nanoparticles were visible and evenly distributed, the lower the score, the less nanoparticles were present within the tissue and distributed more heterogenously. As for the cellular reaction, a type b score was developed with high scores corresponding to absence and low scores to presence of an inflammatory response. The inflammatory response was judged by the infiltration of inflammatory cells
423
within the tissue, such as macrophages, neutrophils, eosinophils, basophils, lymphocytes and plasma cells. The total score of the inflammatory response that could be reached was 15 points if no signs of inflammation could be detected. Analysis of the results was performed semiquantitatively by comparison and interpretation of total score points. The tendency of development and behavior over time was examined by comparison of carpal and stifle joint left/right, i.art./p.art. injection, with
Table 2 Score settings for histological evaluation of the synovial membrane Characteristics
Score type
Score grades 0
1
2
3
extracellular local 1st layer mild small spots lining membrane only spots/no agglomeration
450% intracellular local and periphery 1st layer moderate thick spots lining membrane
450% intracellular evenly distributed 41st layer large and many spots
spots in cytoplasm/ extracellular agglomeration
intra-/extracellular agglomeration
Particle localization Intra-/extracellular Distribution Blue uptake intima Superficial stain
a a a a
none none none none
Singular/agglomerate
a
none
Particle concentration Stain detection Blue uptake cytoplasm
a a
none none
20/40 objective small spots
10 objective spots all over cytoplasm
Extracellular aggregation Cellular debris Activation
a
none
local
diffuse
1.25 objective thick spots, small nucleus spots widely distributed
b b
severe 45 cell layers
moderate and periphery 4–5 cell layers
mild and local 3–4 cell layers
none 1–3 cell layers
Inflammatory response Perivascular infiltration
b
4–10 cells
1–3 cells
none
b b b
cells surrounding vessels severe severe all over
moderate moderate local and periphery
mild mild local
none none none
a a a a
none none none none
local mild mild o30% cell associated
local and periphery moderate moderate 30–50% cell associated
all over severe severe 450% cell associated
Cellular infiltration Pathology Inflammatory distribution Fluorescence Distribution Intensity Autofluorescence Cell association
(a) nanoparticle concentration as well as localization. High scores represent large numbers of particles with even distribution, low scores low numbers with heterogenous distribution. (b) cellular reaction: high scores represent absence and low scores presence of inflammation.
ARTICLE IN PRESS 424
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
magnet/without magnet, lateral/medial aspect of the joint, and intima/fibrous/adipose connective tissue. Finally the two nanoparticles applied i.art. were compared at 24, 72 and 120 h time of application after surgery.
3. Results 3.1. Experimental animals i.art. and p.art. injection, as well as recovery from anesthesia was uneventful for all animals. They were weight bearing immediately after surgery. Some degree of pain was noticed which disappeared after magnet removal at 12 h. Sheep injected with PVA-SPIONs showed no clinical signs of joint inflammation or lameness after injection and magnet removal. Only sheep injected with the amino-PVA-Cy3.5-SPIONs revealed mild clinical signs of inflammation. At 72 h heat at the joint area could be detected and all sheep examined were mildly painful on palpation. Slight joint effusion could be seen in only one sheep at 72 h after injection. Joint effusion, heat and pain were considerably less at 120 h. Control
animals were normal without any lameness and inflammation. 3.2. Macroscopic examination Similar findings were reported after opening of the joint. Sheep treated with the PVA-SPIONs showed no obvious, clinical signs of inflammation with both applications (i.art., p.art.) (Fig. 2a). The periarticular connective tissues as well as the synovial membranes were whitish, clear and moist. Topical brown discolouration could be detected locally at the p.art. sites of injection and in the synovial membranes after i.art. injection (Fig. 2b). The synovial fluid was coloured brown, but of normal consistency. In contrast, sheep injected with amino-PVACy3.5-SPIONs developed signs of inflammation at the synovial membrane at 72 h after i.art. injection (Fig. 2c). The synovial membranes appeared yellow–brown, moist, thickened and the tips of the synovial tufts showed localized red colouration between the femoral condyles as well as in the proximal aspects of the stifle joint. The carpal joints rarely showed red discolouration, but thick amber-coloured flocks were visible within the joint. Discolouration was more obvious in joints
Fig. 2. Macroscopic view: joint after application of PVA-SPIONs i.art. (a), p.art. (b) and of the amino-PVA-Cy3.5-SPIONs (c) at 72 h after injection at the stifle joint. In (a) no inflammatory signs of the synovial membrane are seen, whereas in (c) signs of inflammation are visible, indicated through redness of the membrane. Thick flocks of fibrin within the synovial fluid confirm inflammation and fibrin exsudation (c). Brown discolouration of the synovial fluid indicates presence of iron particles. SM ¼ synovial membrane; FC ¼ femoral condyle; arrows indicating particle discolouration.
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
treated with an extracorporal magnet. The amount of synovial fluid was increased in all joints. The articular cartilage matrices were whitish–grey with a consistency softer than normal. However, in sheep sacrificed 120 h after surgery the clinical signs of inflammation had almost subsided. Most importantly, the cartilage matrices were almost normal, at least its consistency if cut. The joint tissues revealed no macroscopic signs of clinical inflammation in the control animals. 3.3. Semiquantitative histological evaluation of PVA-SPIONs Localization after intraarticular application: Evaluation of the score of nanoparticle localization within the intima revealed that, in combination with a magnet, higher scores could be reached than without a magnet (Table 3). Scores were generally increasing over time. A peak could be detected at 120 h in most cases. In the carpal joint total scores were higher than in the stifle joints. In the fibrous connective tissue underlying the intima, the scores of nanoparticle localization were lower with a magnet than without a magnet (Table 4). There, nanoparticle distribution had a tendency to be focused more locally with a magnet in place. In the adipose connective tissue results were comparable to those of the fibrous tissue, whereas there was a tendency that total scores where higher with a magnet than without a magnet. Localization after periarticular application: Since p.art. injections did not involve the synovial membranes, no nanoparticles were detected within the intima. Peaks of nanoparticle localization in the fibrous tissue were variable with a trend to be highest at 72 h. Scores for nanoparticle localization were generally higher compared to i.art. injections with a magnet, whereas scores were lower compared to results without magnets. The tendency of development over time was variable but increased towards the peak that occurred mostly at 72 h and decreased thereafter. In the adipose tissue, the scores for nanoparticle localization were very low. Concentration after intraarticular application: Nanoparticle concentrations in the intima
425
Table 3 Histological evaluation of the intima: mean values and standard deviation (SD) over all times and joints in summary Characteristic
Intima Cy with +
Intima-Cy without +
Intracellular: extracellular Particle distribution Blue stain uptake of intima Superficial stain Singular: agglomerated Stain detection Blue uptake of cytoplasm Cellular debris Activation Cellular infiltration Pathology Inflammatory distribution Superficial fluorescence Distribution fluorescence Intensity fluorescence Autofluorescence tissue Cell associated fluorescence
1.6571.53 0.9571.0 0.5570.76 0.0570.22 0.7070.80 0.8570.93 0.7070.73 2.8970.46 2.9570.23 2.0070.88 2.7970.54 1.5371.17 0.070.0 0.7070.92 1.071.17 0.2070.52 0.6570.88
1.5771.40 1.2471.14 1.2471.26 0.070.0 1.1071.09 0.9070.89 0.9571.02 2.9070.30 2.8670.36 1.6770.80 2.5770.81 0.7671.09 0.070.0 1.071.26 0.9071.18 0.5270.93 0.9071.18
increased when a magnet was applied, whereas in the subintimal fibrous tissue the application of a magnet lead to a decreasing scores for nanoparticle concentrations (Fig. 3a). The development over time was highly variable. Concentration after periarticular application: Scores in the fibrous tissue were higher without than with the magnets, but generally scores were found higher compared to i.art. injections. Concentrations over time were variable, but a peak was noticed at 72 h. Negligible amounts of nanoparticle concentrations were found in the adipose tissue. Inflammatory response after intraarticular application: Generally signs of inflammation were mild and not present regularly (Fig. 3b). No difference was found in the intima and its inflammatory responses between the application of an external magnet or not. Inflammatory response after periarticular application: Similar to i.art. injections, signs of inflammation were mild. Scores for inflammation were lower (low scores ¼ increase of inflammation) if no magnet was applied, whereas with a magnet, the signs of inflammation had a tendency to have
ARTICLE IN PRESS 426
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Table 4 Histological evaluation of the fibrous connective tissue (FCT): mean values and standard deviation (SD) over all times and joints in summary Particle in the fibrous connective tissue
amino-PVA-Cy3.5-SPION
PVA-SPION
Route of injection Number of slides evaluated Application of a +
i.art. 24 with +
i.art. 24 without +
i.art. 8 with +
i.art. 8 without +
p.art. 8 with +
p.art. 8 without +
Parameter Particle localization Intracellular:extracellular Particle distribution
mean7SD
mean7SD
mean7SD
mean7SD
mean7SD
mean7SD
0.5071.14 0.2570.61
1.0471.30 0.7070.82
1.1371.55 0.6370.92
1.8871.36 1.7571.16
1.7571.49 1.1370.99
1.8871.25 1.1370.83
Particle concentration Singular:agglomerated Stain detection Blue uptake of the cytoplasm Extracellular aggregations
0.2570.68 0.2970.75 0.3370.82 0.0070.00
1.3571.47 1.1371.36 0.8371.30 0.9171.24
0.6371.06 0.6371.06 0.7571.16 0.3870.74
1.5071.20 1.7571.39 1.5071.20 1.0071.20
1.7571.49 1.7571.49 1.7571.49 1.2571.16
2.0071.41 1.8871.36 1.7571.39 1.1371.36
Inflammatory response Perivascular infiltration Cellular infiltration Pathology Inflammatory distribution
2.2970.81 1.9270.97 2.5070.66 1.8371.13
2.2270.74 1.7071.15 2.2271.09 1.6171.20
2.7170.76 2.6371.06 3.0070.00 2.6371.06
2.8670.38 2.7570.46 3.0070.00 2.7570.46
2.8670.38 2.6370.52 2.8870.35 2.6370.52
2.8670.38 2.8870.35 3.0070.00 2.6371.06
Fluorescence Distribution of fluorescence Intensity of fluorescence Autofluorescence of tissue Cell association of fluorescence
0.2170.59 0.2570.74 0.2570.68 0.0070.00
0.4870.79 0.5770.99 0.6170.89 0.3570.78 15.12
20.74
22.13
21.14
Total for mean values (without SD 10.16 and fluorescence)
13.71
more local character. Perivascular infiltration could be detected only occasionally. 3.4. Semiquantitative histological evaluation of amino-PVA-Cy3.5-SPIONs (i.art.) Localization: Higher scores were reached in the intima without than with extracorporal magnets. All parameters increased over time and showed a peak at 120 h (Table 5). In the fibrous connective tissues, nanoparticle localization reached higher scores without a magnet as well. An increasing tendency for particle localization or concentration towards the lateral (where the magnet was located) or the medial aspect of the joint was not detected; neither in the intima nor in the fibrous tissues. The development
over time was variable and no common peak was located. In the adipose tissue, almost no nanoparticles were detected. Concentration: Generally, scores were higher in the intima than in the fibrous tissue after i.art. injection of nanoparticles. Without a magnet, mean values of all parameters examined were slightly higher than with a magnet in both types of tissue (Fig. 4a). No tendency for increased concentrations over time or development of peaks was recorded; neither in the intima nor in the fibrous tissue. Concentration of nanoparticles in the adipose tissue was considerably less. Inflammatory response: Inflammation of the intima seemed to be milder after injection of the amino-PVA-Cy3.5-SPION, when a magnet was applied as compared to experiments where no
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
427
Fig. 3. Histology samples after injection of PVA-SPIONs in Pearls’s Prussian blue reaction (stains Fe3+) (a) and haematoxylin-eosin stain H.E. (b); in (a) iron is blue and nanoparticles show cellular association (arrows); in (b) iron is brown, no inflammatory cells are visible. Table 5 Development of SPIONs uptake as well as biocompatibility by the time; scores summarized for all joints and aspects of the joint; only i.art. injections were compared. At 72 h: significant differences between the application of a magnet or not Characteristic
ACT-Cy with +
ACT-Cy without +
Intracellular: extracellular Particle distribution Singular: agglomerated Stain detection Blue uptake of cytoplasm Extracellular aggregation Perivascular infiltration Cellular infiltration Pathology Inflammatory distribution Distribution fluorescence Intensity fluorescence Autofluorescence tissue Cell associated fluorescence
0.2570.85 0.0870.28 0.2570.85 0.1770.56 0.2570.85 0.070.0 2.9670.2 2.9270.28 3.070.0 2.8870.45 0.070.0 0.070.0 0.070.0 0. 070.0
0.3070.76 0.2270.52 0.4871.08 0.2670.69 0.1770.65 0.2670.69 2.7070.56 2.3070.93 2.6570.78 2.1771.07 0.070.0 0.070.0 0.0470.21 0.070.0
magnet was present (Fig. 4b). Using a magnet, clearly more signs of inflammation were found on the lateral aspect of the joint (where the magnet was located), whereas without a magnet no difference between the lateral and medial aspects were found. The peak of the inflammatory response detected histologically appeared to be at 24 h in most cases. At this time, cellular infiltration scored from mild to severe, however, was highly variable within the same histology slide. Inflammatory responses in the fibrous connective tissue showed no difference between application of a magnet or not. The same was true for the lateral or the medial aspects of the joints. In most cases, a peak was noticed at 24 h after injection. However, development over time was variable, such that in four cases inflammatory parameters were improving, in three cases they varied, and in one case they became worse. Cellular infiltration
ARTICLE IN PRESS 428
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Fig. 4. Histology samples after injection of amino-PVA-Cy3.5-SPIONs in Pearls’s Prussian blue reaction (Fe3+) (a) and H.E. (b); cellular infiltration in the fibrous connective tissue in (c); arrows indicate inflammatory cells.
varied from mild to severe without any preferences for the joint, body site or magnet application (Fig. 4c). Evaluation of fluorescence: Stronger fluorescence signals were recorded in all tissues examined without a magnet than with a magnet; although with a magnet, the fluorescence was more intense
at the lateral aspect of the joint than at the medial aspect of the joints (Fig. 5). There was an overall tendency for a peak to be reached between 72 and 120 h. Cell association of fluorescence was significantly higher in the intima than in the fibrous connective tissue. Autofluorescence of the tissue showed only mild differences between both types
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
429
Fig. 5. Confocal microscopy of the synovial membrane of amino-PVA-Cy3.5-SPIONs injected joints (a); blue visulaizes the nucleus stained with Hoechst 33342, red indicates fluorescence. Fluorescence is detectable intracellularly; less cell association of the fluorescence in the connective tissue (b).
of tissue, but was mainly detected in the connective tissue. Local distribution was not very prominent and intensity was mild. Control animals: The Pearl’s Prussian blue reaction was completely negative in all synovial tissues examined. The same was true for fluorescence. Nevertheless, local inflammatory reactions were found in four control sheep, especially in the fibrous connective tissue and less prominent in the intima of the stifle joint (Fig. 4b). 3.5. Comparison of both nanoparticle-types (only i.art. injection) In all types of tissue examined the PVA-SPIONs showed better results than the amino-PVA-Cy3.5-
SPIONs. Generally, more PVA-SPIONs accumulated in the tissues and were found inside the cells as compared to amino-PVA-Cy3.5-SPIONs (Table 5). At the same time, inflammation was milder with the non-functionalized PVA-SPIONs. Amino-PVA-Cy3.5-SPIONs reached higher scores without a magnet than with a magnet concerning concentration and localization within the intima, as well as in the fibrous connective tissue. By contrast, PVA-SPIONs uptake in the intima as well as in the adipose connective tissue was increased with the application of an extracorporal magnet. In the fibrous connective tissue, similarly as with amino-PVA-Cy3.5-SPIONs, the mean values for concentration as well as for localization were less with a magnet than without a magnet.
ARTICLE IN PRESS 430
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
In case of PVA-SPIONs, the inflammatory response was milder without application of a magnet as compared to animals where a magnet was applied. In contrast, inflammation appeared to be more severe after application of the amino-PVACy3.5-SPIONs without the application of a magnet (Tables 3 and 4).
4. Discussion In this study, two types of superparamagnetic nanoparticles, PVA-SPIONs and its fluorescently functionalized analogue, amino-PVA-Cy3.5SPIONs were assessed concerning biocompatibility and applicability for magnetic drug targeting. Therefore, nanoparticles were applied either intra(both types of nanoparticles) or periarticularly (only PVA-SPION) in the stifle and carpometaphalangeal joints of sheep. To increase local concentration of the superparamagnetic nanoparticles, an extracorporal magnet was applied. Tracking of nanoparticles in the synovial membrane was performed using Pearl’s Prussian blue reaction for iron ions (both types of nanoparticles), as well as fluorescence microscopy (aminoPVA-Cy3.5-SPIONs). It could be demonstrated that PVA-SPIONs showed better biocompatibility and increased nanoparticle accumulation than the amino-PVA-Cy3.5-SPIONs, when speed of uptake, accumulation of nanoparticles and inflammatory reaction of the tissues were compared. Uptake of the amino-PVA-Cy3.5-SPIONs was initially slow and increased substantially between 24 and 120 h. Both types of nanoparticles could be detected successfully at any time between 3 and 120 h with the methods established in vitro and applied in this study. With amino-PVA-Cy3.5-SPIONs, fluorescence localization, distribution and intracellular appearance was generally similar as the blue staining for iron ions resulting from the Pearl0 s Prussian blue reaction with the iron oxide core of the nanoparticles. This indicates that the linkage between the nanoparticle and the fluorescent dye was stable over the whole experimental period and therefore, the superparamagnetic nanoparticles, together with the fluorescent dye, were simulta-
neously internalized into the synovial membrane tissue. In view of future clinical employment of magnetic targeted drugs in joint diseases, an experimental animal model with sheep was chosen for proof of principle with plain and functionalized PVA coated SPIONs in joint structures [18]. In sheep, bone and joint structures are very similar to humans and results can be directly transferred for human applications [19]. For costs and ethical reasons, the number of animals was limited and thus, statistical analysis of the results could not be performed and had to be restricted to merely qualitative and semi-quantitative descriptions. However, the changes of cell morphology and nanoparticle uptake over time could be followed in vivo in short time intervals and different anatomical locations that outweighed by far the disadvantages of the small animal groups. In addition, the 3 h group of animals was not repeated with the amino-PVA-Cy3.5-SPIONs. Results from analysis of the plain PVA-SPION showed that there was only minimal nanoparticle uptake at 3 h after injection. Furthermore, due to the small amount of synovial membrane that could be harvested in the carpophalangeal joints and the varying anatomical structure within the joints, statistical objective analysis was not possible and appeared not to be adequate. Our purpose was to evaluate each slide qualitatively and apply a semiquantitative scoring system to the various structures and local appearances of the tissue. However, this system was not used to judge if the characteristics evaluated were ‘‘positive’’ or ‘‘negative’’. Technical errors in surgery were minimal. p.art. injections were successful in all cases as was easily documented during the macroscopic evaluation after sacrifice of animals. In case of the i.art. injections, it was not always possible to document the initial location of the needle during injection. While it was always possible to withdraw synovial fluid after puncturing the joint cavity with a needle in the stifle, this was more difficult in the carpometaphalangeal joints. There, the joint cavity is much smaller and the amount of fluid less and thus, it was not possible to withdraw synovial fluid in all cases. In the few cases, where microscopic
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
nanoparticle detection was negative, these results may have been related to failure of the injection technique. Detection of nanoparticles was not possible in every slide evaluated. Especially in the stifle joints the nanoparticles were not regularly found if the lateral and the medial aspects were compared. The synovial membrane is the inner layer of the joint capsule and it forms many pouches in different sizes where the nanoparticles may have been caught after injection. This may have been the reason for the heterogenous distribution of the nanoparticles within the synovial membrane in conjunction with the different morphological joint structures and extracorporal magnets rather than different uptake of nanoparticles through the synoviocytes. Within the same areas of the pouches and periarticular joint structures uptake was very homogenous. The differences between the presence and absence of an extracorporal magnet were expected to be more obvious, especially after injection of the amino-PVA-Cy3.5-SPIONs. Partly, this may have been due to the lower concentrations of injected nanoparticle suspension. The limited suspension stability of the amino-PVA-Cy3.5-SPIONs did not allow the application of higher concentrations. Therefore, results and comparison to the plain PVA-SPIONs have to be interpreted with some caution. Apart from the concentration, the inflammatory reaction may have also contributed to lower nanoparticle detection with the amino-PVACy3.5-SPIONs. Despite lower concentrations, the inflammatory response of the tissue was more pronounced and thus, faster elimination of the nanoparticles could be expected. Using higher strength magnets (41.3 T) or an optimized magnetic field for the specific type of joint tissue, the SPIONs may have been retained longer within the synovial membrane for both types of nanoparticles also in face of an increased inflammatory response. As synovial cells are said to have macrophagic properties, phagocytosis appears to be a reasonable mechanism of intracellular uptake, as assumed for polymerized albumin nanospheres after i.art. injection [12]. Mild signs of inflammation were seen in conjunction with nanoparticle uptake and localization in both types of nanoparticles.
431
Similar localized inflammations of synovial membranes were also detected in control animals, although they never received either nanoparticles, sham injections or magnets. Nevertheless, the inflammatory response of the tissue was more consistent and pronounced after injection of amino-PVA-Cy3.5-SPIONs compared to controls and injection of PVA-SPIONs. The fluorescent dye, Cy3.5, has been applied before in many in vitro and in vivo studies with a variety of cells and animal species and, therefore, was considered to be safe to use for intraarticular applications. The application of this dye in joints was not found in the literature and thus, cannot be compared to the results of this current study. In a recent study, biodegradable nanoparticles that were applied to synovial membranes were not analysed concerning their biocompatibility, as injection of particles into joints where inflammation is already present cannot show any reactions caused by the particles [12]. Similar research work with viral vectors in the area of gene therapy for arthritic diseases were shown to induce specific immune responses resulting in early reduction of gene expression [20]. There, the inflammatory response was attributed to the viral vectors. The inflammatory response of synovial membranes seen in vivo in the current study could also be related to the amino-groups used for functionalization of the PVA coated nanoparticles. Preliminary work in the laboratory has shown that amino functionalized, PVA coated nanoparticles in high concentrations were detrimental for cell viability of the musculoskeletal system, although it was dependent on concentrations and availability of free amino-groups (unpublished data). We, therefore, concluded that the reaction could be caused by a combination of both, the fluorescent dye and the remaining free amino-groups. However, this question was not further pursued, since the fluorescent dye served only as proof of principle for magnetic drug delivery and for easier detection of the nanoparticles in combination with a functionalized molecule in the synovial tissue. Future functionalization of PVA coated SPION with other molecules for i.art. applications, such as antiinflammatory drugs, plasmids or proteins, will have to be investigated on an individual basis.
ARTICLE IN PRESS 432
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Bindings to the nanoparticles may vary with each individual molecule and thus, basic biocompatibility tests will have to be conducted for each type of medical application, at least in the musculoskeletal system. Nevertheless, as the binding between the functionalized nanoparticle and the fluorescent dye appeared stable when crossing the cell membrane, it is save to assume that this may also be the case if a drug is attached to the nanoparticles based on the same binding properties.
5. Conclusions PVA coated SPIONs were shown to be a promising delivery system for magnetic drug targeting in synovial membrane tissue as they were taken up intracellularly in vitro and in vivo. Furthermore, the nanoparticles and the fluorescent dye remained within the synovial membrane for at least five days indicating that they could prolong the action of intraarticularly applied medication for treating acute or chronic joint diseases. Current work of our groups is focused on functionalization of PVA-coated SPIONs with specific drugs, proteins and plasmids directed at reducing acute and chronic joint inflammation as well as inhibiting cartilage matrix degradation.
Acknowledgements The authors thank the Department of Prof. Dr.med.vet. Andreas Pospischil for support in preparing the histology slides, Kati Zlinszky and Sabina Wunderlin for their help in establishing the special staining reactions. This work was
supported by the European Community, 5th Framework Program, ‘‘Magnanomed’’ G5RD2000-00375.
References [1] R. Jain, P.E. Lipsky, Med. Clin. North. Am. 81 (1997) 57. [2] A.A. Kalla, A.F. Tooke, E. Bhettay, et al., Drug Saf. 11 (1994) 21. [3] T. Tomita, H. Hashimoto, H. Yoshikawa, Curr. Drug Targets 4 (2003) 609. [4] D.A. Putnam, Am. J. Health Syst. Pharm. 53 (1996) 151. [5] S.S. Davis, I.M. Hunneyball, L. Illum, et al., Drugs Exp. Clin. Res. 11 (1985) 633. [6] J.A. Hunter, T.H. Blyth, Drug Saf. 21 (1999) 353. [7] R.E.L. Manin H.J., in: D.J. Mc Carty, (Ed.) Structure and function of joints, Lea Fiebiger, Philadelphia, 1993. [8] H. Derendorf, H. Mollmann, A. Gruner, et al., Clin. Pharmacol. Ther. 39 (1986) 313. [9] M.H. Bonanomi, M. Velvart, M. Stimpel, et al., Rheumatol. Int. 7 (1987) 203. [10] S. Bozdag, S. Calis, H.S. Kas, et al., J. Microencapsul. 18 (2001) 443. [11] C.J.P. Williams A.S., Goodfellow R.M.,Williams B.D., Brit. J. Rheumat. (1996) 719. [12] E. Horisawa, T. Hirota, S. Kawazoe, et al., Pharm. Res. 19 (2002) 403. [13] J. Highton, D. Guevremont, J. Thomson, et al., Clin. Exp. Rheumatol. 17 (1999) 43. [14] C.C. Berry, A.S.G. Curtis, J. Physics. D: Appl. Phys. (2003) R198. [15] F. Scherer, M. Anton, U. Schillinger, et al., Gene Ther. 9 (2002) 102. [16] M. Chastellain, A. Petri, H. Hofmann, J. Coll. Interface Sci. 2004, submitted for publication. [17] A. Fink-Petri, M. Chastellain, L. Juillerat-Jeanneret, et al., J. Biomater. 2004, submitted for publication. [18] L.F.v.d. Kraan PM van der, Berg WB van den, Biomaterials (2004) 1497. [19] D. Apelt, F. Theiss, A.O. El-Warrak, et al., Biomaterials 25 (2004) 1439. [20] E.C. Robbins PD, Chernajovsky Y, Gene Therapy (2003) 902.
Journal of Magnetism and Magnetic Materials 293 (2005) 419–432 www.elsevier.com/locate/jmmm
Intraarticular application of superparamagnetic nanoparticles and their uptake by synovial membrane—an experimental study in sheep Katja Schulzea, Annette Kochb, Bernhard Scho¨pfa, Alke Petric, Benedikt Steitzc, Mathieu Chastellainc, Margarethe Hofmannd, Heinrich Hofmannc, Brigitte von Rechenberga, a
Musculoskeletal Research Unit, Equine Hospital, Vetsuisse Faculty, University of Zurich, Winterthurerstrasse 260, 8057 Zurich, Switzerland b Department of Chemistry and Applied BioSciences, Swiss Federal Institute of Technology Zurich (ETH Zurich), Winterthurerstrasse 190, 8057 Zurich, Switzerland c Laboratory of Powder Technology, Swiss Federal Institute of Technology (EPFL), Lausanne, Switzerland d MatSearch, Chemin Jean Pavillard 14, 1009 Pully, Switzerland Available online 28 March 2005
Abstract A superparamagnetic iron oxide nanoparticle, coated with polyvinyl alcohol, (PVA-SPION) and its fluorescently functionalized analogue (amino-PVA-Cy3.5-SPION) were compared in vivo as proof of principle for future use in magnetic drug targeting in inflammatory joint diseases. They were injected either intraarticularly or periarticularly and their uptake by cells of the synovial membrane was evaluated. Uptake was completed in 48 h and was enforced by an extracorporally applied magnet. r 2005 Elsevier B.V. All rights reserved. Keywords: Magnetic drug targeting; Sheep; Knee joint; Inflammation; SPION; Intraarticular medication; Osteoarthritis; Fluorescent nanoparticles; Cy3.5; Biocompatibility; Synovial cells; Phagocytosis; PVA
1. Introduction Chronic aseptic inflammatory diseases of the joint such as rheumatoid arthritis or osteoarthritis Corresponding author. Tel.: +41 1 6358 410; fax: +41 1 6358 950. E-mail address: [email protected] (B. von Rechenberg).
require long-time therapy of patients with analgesic, anti-inflammatory, immune-modulating or chondroprotective drugs to release pain and modulate the degree of disease symptomatically [1]. Prolonged medication periods are commonly associated with negative side effects of the drugs, such as nephrotoxicity, hepatitis or gastrointestinal ulceration leading to forced discontinuation of
0304-8853/$ - see front matter r 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jmmm.2005.02.075
ARTICLE IN PRESS 420
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
the medication [2]. In the last years several attempts were made to find new successful therapy strategies reducing the unwanted side effects while at the same time leading to optimal well being of the patients [3–5]. If possible, replacement of systemic drug administration through intraarticular injections is the most effective route to treat joint diseases [6]. However, the efficiency of the injected drug may be limited through a relatively short drug persistence or failure to maintain adequate drug concentrations in the joint cavity [7–9]. Different approaches are made to optimize drug persistence and efficiency in the joint either with biodegradable drug carriers, conjugated drug formulations or targeted drug systems [10–13]. Of those, superparamagnetic iron oxide nanoparticles (SPIONs) seem highly interesting for clinical applications because of their superparamagnetic properties and their possible use in magnetic drug targeting including gene therapy [14,15]. Up to this time of writing, the use of SPIONs for joint therapy has not been reported. The goal of this study was to determine whether intraarticularly (i.art.) or periarticularly (p.art.) injected SPIONs with PVA coatings are biocompatible and successfully taken up by synovial cells in vivo. In addition, the effect of an extracorporally applied magnet on nanoparticle persistence within the synovial membrane was tested. Last but not least, a fluorescent molecule, Cy3.5, was covalently attached to the surface of the nanoparticles for (i) tracking the nanoparticles within the tissues and at the same time (ii) acting as a model cargo to evaluate nanoparticle stability, performance and biocompatibility with respect to drug targeting applications after i.art. injection using sheep as an experimental model.
2. Materials and methods 2.1. Preparation of nanoparticles Polyvinyl alcohol (PVA; (Mowiols 3-83, average molecular weight Mw: 14 000 g/mol, hydrolysis degree: 83% supplied by courtesy of CLARIANT)) stabilized iron oxide nanoparticles were produced as described elsewhere [16]. For this
work, 10 wt% of the PVA was exchanged by vinyl alcohol/vinyl amine copolymer (M12, average molecular weight of 80 000–140 000 g/mol, supplied by courtesy of Erkol, (E)). To limit the number of amine group per nanoparticle, a vinyl alcohol/vinyl alcohol copolymer mass ratio of 45 was chosen. The overall ratio by weight of iron to polymer content was for all samples fixed at 13. The concentration of the amino-PVA functionalized nanoparticles (amino-PVA-SPIONs) in aqua’s solution was 4.6 mg Fe/ml. The results of a more detailed characterization of the nanoparticles are given elsewhere [17]. Amino-PVA-SPIONs were further derivatized by covalent coupling of a fluorescent dye to the nanoparticles via the polymer. For this work Cy 3.5: derivatized CyDyeTM (NHS ester; Eurogentec) was chosen that carries one reactive group on each dye molecule for accurate labelling of amine groups. The absorption maximum of this dye is at 581 nm, the maximal emission at 596 nm.The final dispersion contains 0.3 mg iron/ml dispersion and approx. 3 dye molecules/nanoparticle. 2.2. Experimental animals and surgical technique A total of 14 Swiss Alpine sheep between 2 and 4 years of age was used for application of PVASPIONs and amino-PVA-Cy3.5-SPIONs into the stifle and carpometaphalangeal joints. Eight additional sheep without SPIONs as well as without magnets and slaughtered for other reasons than infectious or systemic disease served as controls. Two main experiments were conducted with the nanoparticles, such that in experiment 1, PVASPIONs were injected either intraarticularly (i.art.; into the joint cavity) or periarticularly (p.art.;into joint capsular tissue) and in experiment 2, aminoPVA-Cy3.5-SPIONs were injected only i.art. The sheep were checked for their overall health, routinely wormed as well as vaccinated. After an adaptation period of 14 days prior to surgery, they were kept in groups in a stall, routinely fed with hay and with free access to water. All experiments were conducted under general anesthesia. The animals were fasted 24 h before sedation with medetomidine (5 mg/kg, i.m., Domitors, OrionFarmos, Turku, Finland) and induction with
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
ketamine (2 ml/kg Narketans, Chassot AG, Bern, Switzerland) in combination with valium i.v. (Diazepam, 0.01 mg/kg, Roche, Basel, Switzerland). Anesthesia was maintained with isoflurane (Forenes, Abbott AG, Baar, Switzerland) in 100% oxygen. Analgesia was achieved using buprenorphine (Temgesics 0.03 ml/kg i.v., ESSEX Chemie AG, Luzern, Switzerland) peri- and postoperatively every 4 h as needed. The skin of both carpal and stifle joints of the sheep were aseptically prepared in a surgical routine manner. Intraarticular injections: At the carpometaphalangeal joint 1.0 ml of the nanoparticle suspension was injected into the synovial cavity using a surgical sterile technique, while in the stifle joint 2.0 ml of the nanoparticle suspension were administered into the lateral joint cavity. Periarticular injections (only experiment 1): The nanoparticle suspension was injected into the capsular and subdermal connective tissue at the lateral aspect of the joints. There, 1.0 ml was placed at the carpal joint region and 1.5 ml at the stifle. 2.3. Extracorporal magnets NdFeB-permanent magnets possessing a remanence field of 1.3 T (Neomag S, Magnequench, Lupfig, Switzerland) were placed in pockets of specially designed bandages that were sutured to the skin (2–0 Supramid, B.Braun, Tuttlingen, Germany) at the craniolateral aspect of the joints immediately after injection of the nanoparticle suspension (Fig. 1). The magnets were removed 12 h after injection. 2.4. Group distribution (see Table 1) Experiment 1: For the application of PVASPIONs eight sheep were randomly split into two groups with four animals each. The first group was used for intraarticular and the second for periarticular injections. Of each group there was one sheep sacrificed at 3, 24, 72 and 120 h after surgery. Experiment 2: For the amino-PVA-Cy3.5SPIONs, six sheep were used and only i.art. injections were performed. The sheep were split into two groups with three sheep each. The first
421
Fig. 1. Sheep with extracorporal magnet sutured to the skin of the carpal and stifle joint after surgery.
group was treated with an extracorporal magnet at the craniolateral aspect of both carpal and stifle joints using the same protocol as before. The second group was left without a magnet. The sheep were sacrificed after 24, 72 and 120 h. Both experiments were conducted according to the Swiss regulations of animal welfare and protection and authorized by the local authorities and Ethical Committee (application No. 59/02).
2.5. Macroscopic evaluation of synovial tissue After sacrifice of the experimental and control animals the front and hind limbs of the animals were collected immediately. The regions of the carpometaphalangeal or stifle joint were exposed through removing the overlying skin. The subcutaneous and periarticular region was judged for yellow–brownish discolouration indicating presence of iron containing nanoparticles and signs of inflammation (reddish, edematous tissue). Thereafter, the joint capsules were opened to expose the synovial membranes, synovial fluid and the cartilage surfaces that were also checked for signs of degradation as well as discolouration due to nanoparticle injection. Cartilage degradation is recognized if there is a cobble-stone appearance of the cartilage surface present as well as fibrillation, cleft formation and/or erosion
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
422
Table 1 Group distribution of animals in experiment 1 and 2 Sheep
Joints
Sacrifice
Injection
Time in h
of SPION
left carpus
left stifle
right carpus
right stifle
Experiment 1
1 2 3 4 5 6 7 8
3 3 24 24 72 72 120 120
i.art. p.art. i.art. p.art. i.art. p.art. i.art. p.art.
+ + + + + + + +
+ + + + + + + +
Experiment 2
1 2 3 4 5 6
24 24 72 72 120 120
i.art. i.art. i.art. i.art. i.art. i.art.
+ + +
+ + +
+ + +
+ + +
Controls
8 animals
none
with magnet: +, without magnet:
down to the subchondral bone. A digital camera was used to document the results. 2.6. Histological evaluation Synovial membranes including capsular tissue as well as cartilage samples were harvested. In experiment 2 samples from the lateral and the medial aspects of the joints were taken. Tissue probes were embedded and processed for standard paraffin histology. Slides of the paraffin blocks were stained with Haematoxylin–Eosin (H.E.) and Pearl’s Prussian blue (Fe3+) using histological standard protocols of our laboratory. Evaluation of biocompatibility and presence of nanoparticles was performed using light microscopy (DMR, Leica, Glattbrugg, Switzerland) and digital imaging (DC 200, Leica). Sections without stains were used for fluorescence microscopy. To perform confocal microscopy nuclei were counter-stained with Hoechst 33342 (3 mg/ml PBS, Molecular Probes, Bedford, MA, USA). There, the slides were incubated with Hoechst 33342 for 30 min at 37 1C, thereafter washed with ddH2O and covered using Dakos Fluorescent Mounting Medium (DAKO Corporation, Carpinteria, CA, USA).
Preliminary evaluation was performed with an inverse microscope (Axiovert 35, Zeiss, Oberkochen, Germany) equipped for fluorescence microscopy. CLSM (confocal laser scanning microscopy, Zeiss 410 inverted microscope, Zurich, Switzerland) was used for final evaluation and confocal microscopy. 2-D multi-channel image processing was performed using the IMARIS software (Bitplane AG, Zurich, Switzerland). Background fluorescence of cells was determined by analysing untreated cells. 2.7. Semiquantitative evaluation The intima (layer of synovial cells facing the synovial cavity), fibrous and adipose connective tissues (both underlying the intima as part of the joint capsule) of the synovial membrane was evaluated separately. Evaluation of nanoparticle localization and concentration was done using slides stained with Pearl’s Prussian blue, whereas biocompatibility was examined using those stained with H.E. A scoring system was developed to assess nanoparticle localization, nanoparticle concentration and inflammatory response (Table 2). Two different types of scores with a range of 0–3
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
were used. Type a score represented nanoparticle concentration as well as nanoparticle localization within the tissue. The higher the scores, the more nanoparticles were visible and evenly distributed, the lower the score, the less nanoparticles were present within the tissue and distributed more heterogenously. As for the cellular reaction, a type b score was developed with high scores corresponding to absence and low scores to presence of an inflammatory response. The inflammatory response was judged by the infiltration of inflammatory cells
423
within the tissue, such as macrophages, neutrophils, eosinophils, basophils, lymphocytes and plasma cells. The total score of the inflammatory response that could be reached was 15 points if no signs of inflammation could be detected. Analysis of the results was performed semiquantitatively by comparison and interpretation of total score points. The tendency of development and behavior over time was examined by comparison of carpal and stifle joint left/right, i.art./p.art. injection, with
Table 2 Score settings for histological evaluation of the synovial membrane Characteristics
Score type
Score grades 0
1
2
3
extracellular local 1st layer mild small spots lining membrane only spots/no agglomeration
450% intracellular local and periphery 1st layer moderate thick spots lining membrane
450% intracellular evenly distributed 41st layer large and many spots
spots in cytoplasm/ extracellular agglomeration
intra-/extracellular agglomeration
Particle localization Intra-/extracellular Distribution Blue uptake intima Superficial stain
a a a a
none none none none
Singular/agglomerate
a
none
Particle concentration Stain detection Blue uptake cytoplasm
a a
none none
20/40 objective small spots
10 objective spots all over cytoplasm
Extracellular aggregation Cellular debris Activation
a
none
local
diffuse
1.25 objective thick spots, small nucleus spots widely distributed
b b
severe 45 cell layers
moderate and periphery 4–5 cell layers
mild and local 3–4 cell layers
none 1–3 cell layers
Inflammatory response Perivascular infiltration
b
4–10 cells
1–3 cells
none
b b b
cells surrounding vessels severe severe all over
moderate moderate local and periphery
mild mild local
none none none
a a a a
none none none none
local mild mild o30% cell associated
local and periphery moderate moderate 30–50% cell associated
all over severe severe 450% cell associated
Cellular infiltration Pathology Inflammatory distribution Fluorescence Distribution Intensity Autofluorescence Cell association
(a) nanoparticle concentration as well as localization. High scores represent large numbers of particles with even distribution, low scores low numbers with heterogenous distribution. (b) cellular reaction: high scores represent absence and low scores presence of inflammation.
ARTICLE IN PRESS 424
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
magnet/without magnet, lateral/medial aspect of the joint, and intima/fibrous/adipose connective tissue. Finally the two nanoparticles applied i.art. were compared at 24, 72 and 120 h time of application after surgery.
3. Results 3.1. Experimental animals i.art. and p.art. injection, as well as recovery from anesthesia was uneventful for all animals. They were weight bearing immediately after surgery. Some degree of pain was noticed which disappeared after magnet removal at 12 h. Sheep injected with PVA-SPIONs showed no clinical signs of joint inflammation or lameness after injection and magnet removal. Only sheep injected with the amino-PVA-Cy3.5-SPIONs revealed mild clinical signs of inflammation. At 72 h heat at the joint area could be detected and all sheep examined were mildly painful on palpation. Slight joint effusion could be seen in only one sheep at 72 h after injection. Joint effusion, heat and pain were considerably less at 120 h. Control
animals were normal without any lameness and inflammation. 3.2. Macroscopic examination Similar findings were reported after opening of the joint. Sheep treated with the PVA-SPIONs showed no obvious, clinical signs of inflammation with both applications (i.art., p.art.) (Fig. 2a). The periarticular connective tissues as well as the synovial membranes were whitish, clear and moist. Topical brown discolouration could be detected locally at the p.art. sites of injection and in the synovial membranes after i.art. injection (Fig. 2b). The synovial fluid was coloured brown, but of normal consistency. In contrast, sheep injected with amino-PVACy3.5-SPIONs developed signs of inflammation at the synovial membrane at 72 h after i.art. injection (Fig. 2c). The synovial membranes appeared yellow–brown, moist, thickened and the tips of the synovial tufts showed localized red colouration between the femoral condyles as well as in the proximal aspects of the stifle joint. The carpal joints rarely showed red discolouration, but thick amber-coloured flocks were visible within the joint. Discolouration was more obvious in joints
Fig. 2. Macroscopic view: joint after application of PVA-SPIONs i.art. (a), p.art. (b) and of the amino-PVA-Cy3.5-SPIONs (c) at 72 h after injection at the stifle joint. In (a) no inflammatory signs of the synovial membrane are seen, whereas in (c) signs of inflammation are visible, indicated through redness of the membrane. Thick flocks of fibrin within the synovial fluid confirm inflammation and fibrin exsudation (c). Brown discolouration of the synovial fluid indicates presence of iron particles. SM ¼ synovial membrane; FC ¼ femoral condyle; arrows indicating particle discolouration.
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
treated with an extracorporal magnet. The amount of synovial fluid was increased in all joints. The articular cartilage matrices were whitish–grey with a consistency softer than normal. However, in sheep sacrificed 120 h after surgery the clinical signs of inflammation had almost subsided. Most importantly, the cartilage matrices were almost normal, at least its consistency if cut. The joint tissues revealed no macroscopic signs of clinical inflammation in the control animals. 3.3. Semiquantitative histological evaluation of PVA-SPIONs Localization after intraarticular application: Evaluation of the score of nanoparticle localization within the intima revealed that, in combination with a magnet, higher scores could be reached than without a magnet (Table 3). Scores were generally increasing over time. A peak could be detected at 120 h in most cases. In the carpal joint total scores were higher than in the stifle joints. In the fibrous connective tissue underlying the intima, the scores of nanoparticle localization were lower with a magnet than without a magnet (Table 4). There, nanoparticle distribution had a tendency to be focused more locally with a magnet in place. In the adipose connective tissue results were comparable to those of the fibrous tissue, whereas there was a tendency that total scores where higher with a magnet than without a magnet. Localization after periarticular application: Since p.art. injections did not involve the synovial membranes, no nanoparticles were detected within the intima. Peaks of nanoparticle localization in the fibrous tissue were variable with a trend to be highest at 72 h. Scores for nanoparticle localization were generally higher compared to i.art. injections with a magnet, whereas scores were lower compared to results without magnets. The tendency of development over time was variable but increased towards the peak that occurred mostly at 72 h and decreased thereafter. In the adipose tissue, the scores for nanoparticle localization were very low. Concentration after intraarticular application: Nanoparticle concentrations in the intima
425
Table 3 Histological evaluation of the intima: mean values and standard deviation (SD) over all times and joints in summary Characteristic
Intima Cy with +
Intima-Cy without +
Intracellular: extracellular Particle distribution Blue stain uptake of intima Superficial stain Singular: agglomerated Stain detection Blue uptake of cytoplasm Cellular debris Activation Cellular infiltration Pathology Inflammatory distribution Superficial fluorescence Distribution fluorescence Intensity fluorescence Autofluorescence tissue Cell associated fluorescence
1.6571.53 0.9571.0 0.5570.76 0.0570.22 0.7070.80 0.8570.93 0.7070.73 2.8970.46 2.9570.23 2.0070.88 2.7970.54 1.5371.17 0.070.0 0.7070.92 1.071.17 0.2070.52 0.6570.88
1.5771.40 1.2471.14 1.2471.26 0.070.0 1.1071.09 0.9070.89 0.9571.02 2.9070.30 2.8670.36 1.6770.80 2.5770.81 0.7671.09 0.070.0 1.071.26 0.9071.18 0.5270.93 0.9071.18
increased when a magnet was applied, whereas in the subintimal fibrous tissue the application of a magnet lead to a decreasing scores for nanoparticle concentrations (Fig. 3a). The development over time was highly variable. Concentration after periarticular application: Scores in the fibrous tissue were higher without than with the magnets, but generally scores were found higher compared to i.art. injections. Concentrations over time were variable, but a peak was noticed at 72 h. Negligible amounts of nanoparticle concentrations were found in the adipose tissue. Inflammatory response after intraarticular application: Generally signs of inflammation were mild and not present regularly (Fig. 3b). No difference was found in the intima and its inflammatory responses between the application of an external magnet or not. Inflammatory response after periarticular application: Similar to i.art. injections, signs of inflammation were mild. Scores for inflammation were lower (low scores ¼ increase of inflammation) if no magnet was applied, whereas with a magnet, the signs of inflammation had a tendency to have
ARTICLE IN PRESS 426
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Table 4 Histological evaluation of the fibrous connective tissue (FCT): mean values and standard deviation (SD) over all times and joints in summary Particle in the fibrous connective tissue
amino-PVA-Cy3.5-SPION
PVA-SPION
Route of injection Number of slides evaluated Application of a +
i.art. 24 with +
i.art. 24 without +
i.art. 8 with +
i.art. 8 without +
p.art. 8 with +
p.art. 8 without +
Parameter Particle localization Intracellular:extracellular Particle distribution
mean7SD
mean7SD
mean7SD
mean7SD
mean7SD
mean7SD
0.5071.14 0.2570.61
1.0471.30 0.7070.82
1.1371.55 0.6370.92
1.8871.36 1.7571.16
1.7571.49 1.1370.99
1.8871.25 1.1370.83
Particle concentration Singular:agglomerated Stain detection Blue uptake of the cytoplasm Extracellular aggregations
0.2570.68 0.2970.75 0.3370.82 0.0070.00
1.3571.47 1.1371.36 0.8371.30 0.9171.24
0.6371.06 0.6371.06 0.7571.16 0.3870.74
1.5071.20 1.7571.39 1.5071.20 1.0071.20
1.7571.49 1.7571.49 1.7571.49 1.2571.16
2.0071.41 1.8871.36 1.7571.39 1.1371.36
Inflammatory response Perivascular infiltration Cellular infiltration Pathology Inflammatory distribution
2.2970.81 1.9270.97 2.5070.66 1.8371.13
2.2270.74 1.7071.15 2.2271.09 1.6171.20
2.7170.76 2.6371.06 3.0070.00 2.6371.06
2.8670.38 2.7570.46 3.0070.00 2.7570.46
2.8670.38 2.6370.52 2.8870.35 2.6370.52
2.8670.38 2.8870.35 3.0070.00 2.6371.06
Fluorescence Distribution of fluorescence Intensity of fluorescence Autofluorescence of tissue Cell association of fluorescence
0.2170.59 0.2570.74 0.2570.68 0.0070.00
0.4870.79 0.5770.99 0.6170.89 0.3570.78 15.12
20.74
22.13
21.14
Total for mean values (without SD 10.16 and fluorescence)
13.71
more local character. Perivascular infiltration could be detected only occasionally. 3.4. Semiquantitative histological evaluation of amino-PVA-Cy3.5-SPIONs (i.art.) Localization: Higher scores were reached in the intima without than with extracorporal magnets. All parameters increased over time and showed a peak at 120 h (Table 5). In the fibrous connective tissues, nanoparticle localization reached higher scores without a magnet as well. An increasing tendency for particle localization or concentration towards the lateral (where the magnet was located) or the medial aspect of the joint was not detected; neither in the intima nor in the fibrous tissues. The development
over time was variable and no common peak was located. In the adipose tissue, almost no nanoparticles were detected. Concentration: Generally, scores were higher in the intima than in the fibrous tissue after i.art. injection of nanoparticles. Without a magnet, mean values of all parameters examined were slightly higher than with a magnet in both types of tissue (Fig. 4a). No tendency for increased concentrations over time or development of peaks was recorded; neither in the intima nor in the fibrous tissue. Concentration of nanoparticles in the adipose tissue was considerably less. Inflammatory response: Inflammation of the intima seemed to be milder after injection of the amino-PVA-Cy3.5-SPION, when a magnet was applied as compared to experiments where no
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
427
Fig. 3. Histology samples after injection of PVA-SPIONs in Pearls’s Prussian blue reaction (stains Fe3+) (a) and haematoxylin-eosin stain H.E. (b); in (a) iron is blue and nanoparticles show cellular association (arrows); in (b) iron is brown, no inflammatory cells are visible. Table 5 Development of SPIONs uptake as well as biocompatibility by the time; scores summarized for all joints and aspects of the joint; only i.art. injections were compared. At 72 h: significant differences between the application of a magnet or not Characteristic
ACT-Cy with +
ACT-Cy without +
Intracellular: extracellular Particle distribution Singular: agglomerated Stain detection Blue uptake of cytoplasm Extracellular aggregation Perivascular infiltration Cellular infiltration Pathology Inflammatory distribution Distribution fluorescence Intensity fluorescence Autofluorescence tissue Cell associated fluorescence
0.2570.85 0.0870.28 0.2570.85 0.1770.56 0.2570.85 0.070.0 2.9670.2 2.9270.28 3.070.0 2.8870.45 0.070.0 0.070.0 0.070.0 0. 070.0
0.3070.76 0.2270.52 0.4871.08 0.2670.69 0.1770.65 0.2670.69 2.7070.56 2.3070.93 2.6570.78 2.1771.07 0.070.0 0.070.0 0.0470.21 0.070.0
magnet was present (Fig. 4b). Using a magnet, clearly more signs of inflammation were found on the lateral aspect of the joint (where the magnet was located), whereas without a magnet no difference between the lateral and medial aspects were found. The peak of the inflammatory response detected histologically appeared to be at 24 h in most cases. At this time, cellular infiltration scored from mild to severe, however, was highly variable within the same histology slide. Inflammatory responses in the fibrous connective tissue showed no difference between application of a magnet or not. The same was true for the lateral or the medial aspects of the joints. In most cases, a peak was noticed at 24 h after injection. However, development over time was variable, such that in four cases inflammatory parameters were improving, in three cases they varied, and in one case they became worse. Cellular infiltration
ARTICLE IN PRESS 428
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Fig. 4. Histology samples after injection of amino-PVA-Cy3.5-SPIONs in Pearls’s Prussian blue reaction (Fe3+) (a) and H.E. (b); cellular infiltration in the fibrous connective tissue in (c); arrows indicate inflammatory cells.
varied from mild to severe without any preferences for the joint, body site or magnet application (Fig. 4c). Evaluation of fluorescence: Stronger fluorescence signals were recorded in all tissues examined without a magnet than with a magnet; although with a magnet, the fluorescence was more intense
at the lateral aspect of the joint than at the medial aspect of the joints (Fig. 5). There was an overall tendency for a peak to be reached between 72 and 120 h. Cell association of fluorescence was significantly higher in the intima than in the fibrous connective tissue. Autofluorescence of the tissue showed only mild differences between both types
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
429
Fig. 5. Confocal microscopy of the synovial membrane of amino-PVA-Cy3.5-SPIONs injected joints (a); blue visulaizes the nucleus stained with Hoechst 33342, red indicates fluorescence. Fluorescence is detectable intracellularly; less cell association of the fluorescence in the connective tissue (b).
of tissue, but was mainly detected in the connective tissue. Local distribution was not very prominent and intensity was mild. Control animals: The Pearl’s Prussian blue reaction was completely negative in all synovial tissues examined. The same was true for fluorescence. Nevertheless, local inflammatory reactions were found in four control sheep, especially in the fibrous connective tissue and less prominent in the intima of the stifle joint (Fig. 4b). 3.5. Comparison of both nanoparticle-types (only i.art. injection) In all types of tissue examined the PVA-SPIONs showed better results than the amino-PVA-Cy3.5-
SPIONs. Generally, more PVA-SPIONs accumulated in the tissues and were found inside the cells as compared to amino-PVA-Cy3.5-SPIONs (Table 5). At the same time, inflammation was milder with the non-functionalized PVA-SPIONs. Amino-PVA-Cy3.5-SPIONs reached higher scores without a magnet than with a magnet concerning concentration and localization within the intima, as well as in the fibrous connective tissue. By contrast, PVA-SPIONs uptake in the intima as well as in the adipose connective tissue was increased with the application of an extracorporal magnet. In the fibrous connective tissue, similarly as with amino-PVA-Cy3.5-SPIONs, the mean values for concentration as well as for localization were less with a magnet than without a magnet.
ARTICLE IN PRESS 430
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
In case of PVA-SPIONs, the inflammatory response was milder without application of a magnet as compared to animals where a magnet was applied. In contrast, inflammation appeared to be more severe after application of the amino-PVACy3.5-SPIONs without the application of a magnet (Tables 3 and 4).
4. Discussion In this study, two types of superparamagnetic nanoparticles, PVA-SPIONs and its fluorescently functionalized analogue, amino-PVA-Cy3.5SPIONs were assessed concerning biocompatibility and applicability for magnetic drug targeting. Therefore, nanoparticles were applied either intra(both types of nanoparticles) or periarticularly (only PVA-SPION) in the stifle and carpometaphalangeal joints of sheep. To increase local concentration of the superparamagnetic nanoparticles, an extracorporal magnet was applied. Tracking of nanoparticles in the synovial membrane was performed using Pearl’s Prussian blue reaction for iron ions (both types of nanoparticles), as well as fluorescence microscopy (aminoPVA-Cy3.5-SPIONs). It could be demonstrated that PVA-SPIONs showed better biocompatibility and increased nanoparticle accumulation than the amino-PVA-Cy3.5-SPIONs, when speed of uptake, accumulation of nanoparticles and inflammatory reaction of the tissues were compared. Uptake of the amino-PVA-Cy3.5-SPIONs was initially slow and increased substantially between 24 and 120 h. Both types of nanoparticles could be detected successfully at any time between 3 and 120 h with the methods established in vitro and applied in this study. With amino-PVA-Cy3.5-SPIONs, fluorescence localization, distribution and intracellular appearance was generally similar as the blue staining for iron ions resulting from the Pearl0 s Prussian blue reaction with the iron oxide core of the nanoparticles. This indicates that the linkage between the nanoparticle and the fluorescent dye was stable over the whole experimental period and therefore, the superparamagnetic nanoparticles, together with the fluorescent dye, were simulta-
neously internalized into the synovial membrane tissue. In view of future clinical employment of magnetic targeted drugs in joint diseases, an experimental animal model with sheep was chosen for proof of principle with plain and functionalized PVA coated SPIONs in joint structures [18]. In sheep, bone and joint structures are very similar to humans and results can be directly transferred for human applications [19]. For costs and ethical reasons, the number of animals was limited and thus, statistical analysis of the results could not be performed and had to be restricted to merely qualitative and semi-quantitative descriptions. However, the changes of cell morphology and nanoparticle uptake over time could be followed in vivo in short time intervals and different anatomical locations that outweighed by far the disadvantages of the small animal groups. In addition, the 3 h group of animals was not repeated with the amino-PVA-Cy3.5-SPIONs. Results from analysis of the plain PVA-SPION showed that there was only minimal nanoparticle uptake at 3 h after injection. Furthermore, due to the small amount of synovial membrane that could be harvested in the carpophalangeal joints and the varying anatomical structure within the joints, statistical objective analysis was not possible and appeared not to be adequate. Our purpose was to evaluate each slide qualitatively and apply a semiquantitative scoring system to the various structures and local appearances of the tissue. However, this system was not used to judge if the characteristics evaluated were ‘‘positive’’ or ‘‘negative’’. Technical errors in surgery were minimal. p.art. injections were successful in all cases as was easily documented during the macroscopic evaluation after sacrifice of animals. In case of the i.art. injections, it was not always possible to document the initial location of the needle during injection. While it was always possible to withdraw synovial fluid after puncturing the joint cavity with a needle in the stifle, this was more difficult in the carpometaphalangeal joints. There, the joint cavity is much smaller and the amount of fluid less and thus, it was not possible to withdraw synovial fluid in all cases. In the few cases, where microscopic
ARTICLE IN PRESS K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
nanoparticle detection was negative, these results may have been related to failure of the injection technique. Detection of nanoparticles was not possible in every slide evaluated. Especially in the stifle joints the nanoparticles were not regularly found if the lateral and the medial aspects were compared. The synovial membrane is the inner layer of the joint capsule and it forms many pouches in different sizes where the nanoparticles may have been caught after injection. This may have been the reason for the heterogenous distribution of the nanoparticles within the synovial membrane in conjunction with the different morphological joint structures and extracorporal magnets rather than different uptake of nanoparticles through the synoviocytes. Within the same areas of the pouches and periarticular joint structures uptake was very homogenous. The differences between the presence and absence of an extracorporal magnet were expected to be more obvious, especially after injection of the amino-PVA-Cy3.5-SPIONs. Partly, this may have been due to the lower concentrations of injected nanoparticle suspension. The limited suspension stability of the amino-PVA-Cy3.5-SPIONs did not allow the application of higher concentrations. Therefore, results and comparison to the plain PVA-SPIONs have to be interpreted with some caution. Apart from the concentration, the inflammatory reaction may have also contributed to lower nanoparticle detection with the amino-PVACy3.5-SPIONs. Despite lower concentrations, the inflammatory response of the tissue was more pronounced and thus, faster elimination of the nanoparticles could be expected. Using higher strength magnets (41.3 T) or an optimized magnetic field for the specific type of joint tissue, the SPIONs may have been retained longer within the synovial membrane for both types of nanoparticles also in face of an increased inflammatory response. As synovial cells are said to have macrophagic properties, phagocytosis appears to be a reasonable mechanism of intracellular uptake, as assumed for polymerized albumin nanospheres after i.art. injection [12]. Mild signs of inflammation were seen in conjunction with nanoparticle uptake and localization in both types of nanoparticles.
431
Similar localized inflammations of synovial membranes were also detected in control animals, although they never received either nanoparticles, sham injections or magnets. Nevertheless, the inflammatory response of the tissue was more consistent and pronounced after injection of amino-PVA-Cy3.5-SPIONs compared to controls and injection of PVA-SPIONs. The fluorescent dye, Cy3.5, has been applied before in many in vitro and in vivo studies with a variety of cells and animal species and, therefore, was considered to be safe to use for intraarticular applications. The application of this dye in joints was not found in the literature and thus, cannot be compared to the results of this current study. In a recent study, biodegradable nanoparticles that were applied to synovial membranes were not analysed concerning their biocompatibility, as injection of particles into joints where inflammation is already present cannot show any reactions caused by the particles [12]. Similar research work with viral vectors in the area of gene therapy for arthritic diseases were shown to induce specific immune responses resulting in early reduction of gene expression [20]. There, the inflammatory response was attributed to the viral vectors. The inflammatory response of synovial membranes seen in vivo in the current study could also be related to the amino-groups used for functionalization of the PVA coated nanoparticles. Preliminary work in the laboratory has shown that amino functionalized, PVA coated nanoparticles in high concentrations were detrimental for cell viability of the musculoskeletal system, although it was dependent on concentrations and availability of free amino-groups (unpublished data). We, therefore, concluded that the reaction could be caused by a combination of both, the fluorescent dye and the remaining free amino-groups. However, this question was not further pursued, since the fluorescent dye served only as proof of principle for magnetic drug delivery and for easier detection of the nanoparticles in combination with a functionalized molecule in the synovial tissue. Future functionalization of PVA coated SPION with other molecules for i.art. applications, such as antiinflammatory drugs, plasmids or proteins, will have to be investigated on an individual basis.
ARTICLE IN PRESS 432
K. Schulze et al. / Journal of Magnetism and Magnetic Materials 293 (2005) 419–432
Bindings to the nanoparticles may vary with each individual molecule and thus, basic biocompatibility tests will have to be conducted for each type of medical application, at least in the musculoskeletal system. Nevertheless, as the binding between the functionalized nanoparticle and the fluorescent dye appeared stable when crossing the cell membrane, it is save to assume that this may also be the case if a drug is attached to the nanoparticles based on the same binding properties.
5. Conclusions PVA coated SPIONs were shown to be a promising delivery system for magnetic drug targeting in synovial membrane tissue as they were taken up intracellularly in vitro and in vivo. Furthermore, the nanoparticles and the fluorescent dye remained within the synovial membrane for at least five days indicating that they could prolong the action of intraarticularly applied medication for treating acute or chronic joint diseases. Current work of our groups is focused on functionalization of PVA-coated SPIONs with specific drugs, proteins and plasmids directed at reducing acute and chronic joint inflammation as well as inhibiting cartilage matrix degradation.
Acknowledgements The authors thank the Department of Prof. Dr.med.vet. Andreas Pospischil for support in preparing the histology slides, Kati Zlinszky and Sabina Wunderlin for their help in establishing the special staining reactions. This work was
supported by the European Community, 5th Framework Program, ‘‘Magnanomed’’ G5RD2000-00375.
References [1] R. Jain, P.E. Lipsky, Med. Clin. North. Am. 81 (1997) 57. [2] A.A. Kalla, A.F. Tooke, E. Bhettay, et al., Drug Saf. 11 (1994) 21. [3] T. Tomita, H. Hashimoto, H. Yoshikawa, Curr. Drug Targets 4 (2003) 609. [4] D.A. Putnam, Am. J. Health Syst. Pharm. 53 (1996) 151. [5] S.S. Davis, I.M. Hunneyball, L. Illum, et al., Drugs Exp. Clin. Res. 11 (1985) 633. [6] J.A. Hunter, T.H. Blyth, Drug Saf. 21 (1999) 353. [7] R.E.L. Manin H.J., in: D.J. Mc Carty, (Ed.) Structure and function of joints, Lea Fiebiger, Philadelphia, 1993. [8] H. Derendorf, H. Mollmann, A. Gruner, et al., Clin. Pharmacol. Ther. 39 (1986) 313. [9] M.H. Bonanomi, M. Velvart, M. Stimpel, et al., Rheumatol. Int. 7 (1987) 203. [10] S. Bozdag, S. Calis, H.S. Kas, et al., J. Microencapsul. 18 (2001) 443. [11] C.J.P. Williams A.S., Goodfellow R.M.,Williams B.D., Brit. J. Rheumat. (1996) 719. [12] E. Horisawa, T. Hirota, S. Kawazoe, et al., Pharm. Res. 19 (2002) 403. [13] J. Highton, D. Guevremont, J. Thomson, et al., Clin. Exp. Rheumatol. 17 (1999) 43. [14] C.C. Berry, A.S.G. Curtis, J. Physics. D: Appl. Phys. (2003) R198. [15] F. Scherer, M. Anton, U. Schillinger, et al., Gene Ther. 9 (2002) 102. [16] M. Chastellain, A. Petri, H. Hofmann, J. Coll. Interface Sci. 2004, submitted for publication. [17] A. Fink-Petri, M. Chastellain, L. Juillerat-Jeanneret, et al., J. Biomater. 2004, submitted for publication. [18] L.F.v.d. Kraan PM van der, Berg WB van den, Biomaterials (2004) 1497. [19] D. Apelt, F. Theiss, A.O. El-Warrak, et al., Biomaterials 25 (2004) 1439. [20] E.C. Robbins PD, Chernajovsky Y, Gene Therapy (2003) 902.