- Email: [email protected]
Irinotecan drug eluting beads for use in chemoembolization: In vitro and in vivo evaluation of drug release properties
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
Irinotecan drug eluting beads for use in chemoembolization: In vitro and in vivo evaluation of drug release properties Rachel R. Taylor, Yiqing Tang, M. Victoria Gonzalez, Peter W. Stratford, Andrew L. Lewis ∗ Biocompatibles UK Ltd., Farnham Business Park, Weydon Lane, Farnham, Surrey GU9 8QL, UK
a r t i c l e
i n f o
a b s t r a c t
Article history:
Drug eluting beads that release irinotecan in a controlled manner may be useful for appli-
Received 8 May 2006
cation in the chemoembolization of colorectal cancer metastases to the liver. In this study,
Accepted 5 September 2006
irinotecan drug eluting beads were prepared with loadings up to 50 mg drug/mL hydrated
Published on line 15 September 2006
beads. Drug loading was via an ion-exchange mechanism with sulfonate binding sites in the bead. Release in vitro was shown to be sustained and dependent upon the presence of
Keywords:
ions in the elution medium, drug loading and bead size. Drug elution in PBS was controlled
Irinotecan drug eluting beads
by solute diffusion within the beads and gave rise to values for the diffusion coefficient,
Chemoembolization
D, of between 2.4 × 10−9 and 1.4 × 10−7 cm2 s−1 . The beads were shown to decrease in size
Drug release
(by a maximum 25–30%), and concomitantly their modulus of compression increased (from
In vitro:in vivo correlation
∼27 kPa to a maximum of about 49 kPa), with increasing drug loading. This did not however, influence their ability to be suspended homogeneously in contrast agent or delivered through a microcatheter. Following porcine hepatic artery embolization, maximum plasma levels were 70–75% lower for both irinotecan and SN-38 compared to intraarterial bolus administration, with peak levels observed at 2 and 5 min after completion of the embolization procedure. The in vivo data were shown to correlate well with the in vitro release measured using a T-apparatus model of embolization. © 2006 Elsevier B.V. All rights reserved.
1.
Introduction
Despite the existence of excellent screening and preventive strategies, colorectal carcinoma remains a major public health problem in western countries. The American Cancer Society estimates there will be 145,290 new cases diagnosed in 2005, and 56,290 people will die of the disease. In the UK, about 36,000 new cases were diagnosed in 2002, and 17,190 people died of colorectal cancer (Ferlay et al., 2004). Colorectal carcinoma is the third leading cause of death from cancer in both males and females in the USA. It is also the third most common malignancy in both men (after prostate and lung cancers) and women (after breast and lung cancers).
∗
Corresponding author. Tel.: +44 1252 732819. E-mail address: [email protected] (A.L. Lewis). 0928-0987/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2006.09.002
Five-year survival rates have improved in recent years for patients with stages I–III colorectal carcinomas; for patients with early, localized disease the survival rate is 90%. Survival has improved through the use of adjuvant chemotherapy for colon cancer and adjuvant chemoradiation for rectal cancer. By the time they are diagnosed, some 25% of colon cancers will have extended through the bowel wall, whereas cancers of the rectum will have spread through the bowel wall in 50–70% of patients and metastasized to lymph nodes in 50–60%. The most common site of extralymphatic involvement is the liver, with the lungs the most frequently affected extra-abdominal organ. Patients with metastatic colorectal tumours frequently die of hepatic failure due to liver metastases.
8
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Treatment options for patients with metastatic colorectal cancer are limited and clinical outcome is generally poor. Systemic chemotherapy can palliate symptoms and improve survival. 5-Fluorouracil (5-FU) based chemotherapy has been the cornerstone of treatment of CRC for more than 40 years, and new drugs with a definite activity have recently broadened the options for treatment. Irinotecan, a topoisomerase inhibitor, was the first drug other than 5-FU to demonstrate significant activity. This product is approved for use in combination with 5-FU/folinic acid in patients without prior chemotherapy and for the second-line treatment of this disease as a single agent in patients who have failed an established 5-FU containing treatment regimen. There is renewed interest in locoregional treatment of the liver in an attempt to more effectively target liver metastases, an example of which is transarterial chemoembolization (TACE). The technique consists of delivering high concentrations of chemotherapeutic agents emulsified in an oil-based medium directly to the tumour bed followed by some form of embolization. The purpose of the embolization is two-fold: nutrient and oxygen starvation of the tumour, and minimization of chemotherapy wash-out with prolonged contact with tumor tissue. These factors combine to significantly reduce systemic toxicity of the chemotherapy. Such an approach has also been investigated in a number of metastatic settings with unresectable liver tumours, including colorectal cancer (Tellez et al., 1998; Sanz-Altamira et al., 1997; Wallace et al., 1990; Stuart, 2003; Giroux et al., 2004; Lorenz et al., 2005; O’Toole and Ruszniewski, 2005). Irinotecan drug eluting beads (DEB) combine the drug with the embolization device and can be administered intraarterially in the same manner as TACE. This drug-device combination may offer the possibility of precisely controlling the release and dose of the drug into the tumor bed. These beads obviate the need for an oil/chemotherapy suspension in addition to the embolic device and facilitate handling and delivery, achieving TACE in a simpler one-step procedure (Lewis et al., 2006a). This study presents both in vitro characterization of irinotecan DEB, with respect to how the drug influences the physical properties and handling of the device, and how the device matrix is able to modulate release of the drug over a therapeutically meaningful timeframe. Moreover, these data are correlated to the plasma pharmacokinetics obtained from a porcine model of hepatic arterial embolization.
The compressibility of 700–900 m and 900–1200 m beads with and without drug loading was assessed using an Instron 4411 with Series IX software (Instron, Canton, MA, USA). The general procedure involved distributing the beads onto a specially fabricated sample holder on the Instron to form a monolayer of about 7–10 mm in width. The residual water around the beads was carefully removed by wicking. The diameter of the probe used was 5.0 mm with a crosshead speed of 5.0 mm/min. The load cell used was 50 N and the specimen gauge length was set as 1.0 mm. All the experiments were carried out at room temperature.
2.
Materials and methods
2.5. Suspension and catheter deliverability of irinotecan DEB
2.1.
Materials
Irinotecan HCl solution (Campto® , Pfizer) of 100 mg/5 mL containing 45 mg/mL sorbitol and 0.9 mg/mL lactic acid at pH 3.5 adjusted by adding sodium hydroxide or hydrochloric acid, was used directly as received. DC BeadTM (hydrogel beads from Biocompatibles UK Ltd., Farnham, UK) were used in the size ranges 100–300 m, 300–500 m, 500–700 m, and 700–900 m (and 900–1200 m made available specifically for compressibility studies). Other solutions used included phosphate buffer
saline solution (PBS) (Inverclyde, Bellshill, UK), and Omnipaque 350 (Amersham, Oslo, Norway).
2.2. Preparation of irinotecan DEB and determination of drug loading level In a typical loading procedure, the packing solution was removed from the beads. Subsequently, the required volume of Campto® solution was added to the bead slurry, and the mixture was roller-mixed. The drug-loaded beads were stored in the dark at 4–8 ◦ C. The dose of drug used in this study ranged from 10 to 90 mg/mL. The loading of irinotecan was calculated by measuring the remaining drug concentration in the solution over time. The measurements were carried out with a Perkin-Elmer Lambda 25 UV spectrophotometer at the wavelength of 369 nm and the irinotecan concentration was determined by comparing with a standard curve. The loading efficiency was calculated as below: loading efficiency (%) =
2.3.
initial drug in solution − residual drug in solution × 100 initial drug in solution
Bead sizing and appearance
The size and appearance of beads during irinotecan loading and elution were examined by use of an Olympus BX50F4 microscope with a Colorview III camcorder (Olympus, Japan), under an aqueous solution. Sizing data were analysed by AnalySIS (Soft Imaging System GmbH) and converted to a frequency histogram with 20 m intervals across at least a 200 m range.
2.4.
Compressibility of irinotecan beads
Bead suspension and catheter deliverability were studied by mixing 1 mL of beads, containing various doses of irinotecan, with 6 mL of contrast medium (Omnipaque 350)/saline (50:50 ratio) in a 10 mL syringe. After mixing through a three-way connector, the homogeneity of this suspension was assessed visually. A mixture was considered homogeneous when the beads were suspended for more than 1 min. The deliverability of DEB was assessed by rating the ease of their delivery down a microcatheter (Progreat® , Terumo Corp., Tokyo, Japan). Microcatheters of 2.4–3 Fr were employed to
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
deliver the different sized beads, with the acceptance criteria of pass/failure based on whether the lumen was occluded during delivery.
2.6.
In vitro release of irinotecan into various media
The elution profiles of irinotecan from DEB were evaluated in water, PBS, and contrast media, respectively. To measure drug release, irinotecan DEB were first suspended in 200 mL of elution medium and drug release over time was determined from sample aliquots by UV spectroscopy at 369 nm. In studies involving contrast medium, 1 mL of irinotecan DEB were rollermixed with 1 mL saline and 2 mL contrast medium (Omnipaque 350, Amersham). The aliquot of solution was removed and diluted with PBS for UV measurement.
2.7. Release of irinotecan using an in vitro embolization model A T-apparatus was designed to study the in vitro release of drug from DEB, as previously described (Amyot et al., 2002). One milliliter of irinotecan DEB was placed carefully in the bottom of the well in the T-apparatus together with 300 mL PBS and maintained at 37 ◦ C. The solution was circulated using a Watson-Marlow 505 U peristaltic pump (Watson-Marlow, Falmouth, UK) at 25 rpm (∼50 mL/min), and passed through a Perkin-Elmer Lambda 25 UV spectrophotometer (Pekin-Elmer, Shelton, CT, USA). The change of absorbance with time was recorded at 369 nm.
2.8.
Preclinical model of hepatic arterial embolization
Hepatic arterial embolization was performed in four groups of crossbred pigs of mixed sex (n = 5 per group): 100–300 m control beads with no drug, 100–300 m irinotecan DEB, 700–900 m irinotecan DEB and intra-arterial bolus injection of drug alone. Each animal underwent a single interventional procedure on day 0, during which a catheter was advanced into the left hepatic lobe or segment, and up to 4 mL of either loaded or unloaded beads in a total volume of 24 mL (including saline and contrast media) were slowly administered until embolization was achieved. Three animals per group were euthanized on day 32 (±2 days) and two animals per group were euthanized on day 90 (±2 days). All animals underwent a comprehensive necropsy, and selected organs and tissues were collected for histopathological evaluation. Plasma samples for the measurement of irinotecan and SN-38 plasma levels were taken over 90 days and analysed according to a reported literature method (Escoriaza et al., 2000).
3.
Results and discussion
3.1.
Maximum capacity of irinotecan loading
Irinotecan drug uptake into the beads increased linearly up to ∼50 mg of drug, with loading efficiencies >99% when loaded for 24 h. Above 60 mg, the loading efficiency was <90% for both the 100–300 m and 500–700 m bead sizes used for this study. The loss of loading efficiency around 50 mg for both sizes indicates that the sulfonate sites in the beads were satu-
9
rated with irinotecan through charge–charge interaction, and further addition of Campto® could not increase the loading significantly. The mechanism of bead uptake of irinotecan is attributed to the interaction between the positively charged irinotecan salt which bears a protonated amine group, and negatively charged sulfonic acid groups (from the 2-acrylamido2-methylpropane sulfonate sodium salt, AMPS) on the beads. A similar interaction has been observed in the system containing doxorubicin and sulfonated dextran microspheres (Liu et al., 2001). According to the quantity of AMPS added during bead manufacture and assuming that there is no material loss during synthesis, a preliminary estimation indicates that there are about 7.7 × 10−5 mol of AMPS sites in 1 mL of beads with the size range of 500–700 m. The value is equivalent to 52 mg irinotecan bound, assuming 100% occupation of sulfonate sites and no interaction between bound irinotecan. This theoretical value corresponds well with the observed limit of linearity of ∼50 mg.
3.2.
Irinotecan loading profiles
For clinical application it was desirable to determine the loading characteristics over a 2 h period. The loading of low doses (10 mg/mL) of irinotecan was rapid with >98% of the drug loaded within 10 min. For drug loading levels between 10 and 40 mg/mL, loading efficiencies >99.4% were obtained for all bead sizes within 2 h. For the optimal limit of 50 mg/mL as determined in Section 3.1, the loading percentage after 2 h was 98.1 ± 1.4% for 100–300 m, 98.0 ± 1.3% for 300–500 m, 95.0 ± 0.2% for 500–700 m, and 93.5 ± 0.7% for 700–900 m, respectively. It was observed that the rate of drug uptake for larger bead sizes was slower than for smaller ones. This effect has been observed previously with doxorubicin, and is attributed to decrease in surface area with increase in bead diameter (Lewis et al., 2005).
3.3. Appearance and size change post irinotecan loading Fig. 1 shows the change in appearance of 300–500 m beads followed by light microscopy during irinotecan loading (target loading 50 mg per mL beads). These images demonstrate that the loading is a rapid procedure, and the blue transparent beads turned a green-turquoise non-transparent hue after about 10 min, which is consistent with the data in Section 3.1. During drug loading, the beads retained their spherical shape, and no aggregation of the beads was observed. Fig. 2 shows the mean diameter change of beads with increasing dose of irinotecan. All sizes show a tendency to decrease in size with increasing drug loading. Although the mean diameter reduced by a maximum of 25–30% in all cases, for the most relevant clinical size (100–300 m) the diameter changes was insignificant. The size (diameter) change associated with drug loading is related to the displacement of water from the beads by the more hydrophobic drug. The relatively tight standard deviation of each mean size measurement demonstrates that the size distribution of beads before and after irinotecan loading does not change significantly. The spherical shape of the beads is retained upon drug loading.
10
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Fig. 1 – Micrographs of morphology of DEB (300–500 m) during irinotecan loading (50 mg/mL). (A) Without drug loading, (B) after 7 min loading, and (C) after 20 min. The scale bar shown is 200 m.
3.4.
Compressibility of irinotecan DEB
Fig. 3 shows the force required to compress beads (modulus) of 700–900 m and 900–1200 m size ranges, with increasing drug loading. The DEB shows a slight trend for increase in modulus with the drug content. The trend is consistent with the decreasing water content and bead size as described above. The higher drug loading results in a denser structure, which exhibits a compressive response more akin to elastic materials. It is noted however, that even the highest dose of irinotecan did not increase the modulus significantly, which is still within the same order of magnitude as unloaded beads, unlike the behaviour observed with these beads when loaded with doxorubicin (Lewis et al., 2006a). The beads loaded with drug maintained their compressible hydrogel structure, which should facilitate the delivery through microcatheters and hence not compromise their use in embolization.
Fig. 3 – Compression modulus of 700–900 m ( ) and 900–1200 m () beads with increasing irinotecan loading.
3.5.
Suspension and catheter deliverability studies
A homogeneous suspension was achieved for all bead sizes and doses in a time ranging from 1 to 13 min. The time to suspension increased with the bead size and decreased as the drug loading increased. For the optimal loading of 50 mg/mL, a homogeneous suspension was achieved in <1 min for the smallest beads and approximately 5 min for the largest bead size range. Suspensions of 100–300 m and 300–500 m beads in saline-contrast medium were easily delivered through a 2.4 Fr catheter, whereas 500–700 m and 700–900 m beads required a 2.7 Fr and 3 Fr catheter, respectively for successful delivery. No bead aggregation or related catheter blockage were observed.
3.6. Fig. 2 – Mean diameter change of DEB with and without irinotecan loading: 100–300 m (); 300–500 m (); 500–700 m (䊉); 700–900 m ().
In vitro release of irinotecan DEB
Fig. 4 shows a release profile of irinotecan DEB that were first placed in deionised water and subsequently in PBS buffer after
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Fig. 4 – Comparison of irinotecan elution in water and PBS at 25 ◦ C (300–500 m DEB, loading 20 mg/mL).
removal from the water. There was no drug released into the deionised water, but release was immediate and rapid when the beads were exposed to PBS. This demonstrates the need for ions in the elution medium in order to displace the protonated irinotecan from the sulfonate binding sites, further supporting the mechanism of ionic interaction between drug and bead. Note that the size of the DEB in pure water increased significantly due to a water content increase under the osmotic pressure in the beads; this could facilitate the more rapid elution of drug from the DEB when subsequently placed in PBS compared to DEB placed in PBS alone. Using a free-flowing in vitro elution method, it was demonstrated that all of the loaded drug could be eluted from the DEB in all cases. The rate of irinotecan elution from the beads into
11
PBS buffer is shown to be dependent upon bead size (Fig. 5). Smaller beads tend to have much faster release rate compared to larger beads with equivalent loading (Fig. 5A versus D). This can be explained by the effect of the larger surface area of smaller beads for the same volume, enhancing the ionexchange rate. Moreover, for the larger bead sizes the distance for drug diffusion within the microspheres is greater (Jantzen and Robinson, 1996). Beads with higher irinotecan loading become more hydrophobic compared to those with low loading, which in turn reduces the ion-exchange and diffusion rates of both salt and drug ions. The influence of the change in hydrophobicity of loaded beads can be evaluated from the volume change of microspheres with and without irinotecan loading. In the previous measurements of size change, for example, the mean size of unloaded 700–900 m beads changed from 830 to 806 m with 10 mg/mL loading, and to 584 m with 50 mg/mL loading. Therefore, the coincident volume decrease is 8.4%, and 65.2%, respectively, which means that at least 2/3 of the original water volume was expelled from the microspheres with 50 mg/mL loading. Despite the fact that there is a decrease in the diffusional distance related to the reduced bead size, the loss of water results in a denser structure after drug loading which significantly retards the drug elution rate. Diffusion in ion-exchange resins has been studied in detail by Boyd and Reichenberg (Boyd et al., 1947; Reichenberg, 1953). The rate of diffusion in ion-exchange resins can be controlled by chemical exchange or solute diffusion, which happen within resins, or film diffusion in which the in-going ion is a macrocomponent of the system. It has been demonstrated that indomethacin release from ion-exchange microspheres was controlled by solute diffusion (particle diffusion) (Chretien et al., 2004). When solute diffusion was the rate-controlling
Fig. 5 – Release profiles into PBS buffer of irinotecan DEB with different loadings: 10 (), 20 (), 40 (䊉), and 50 mg/mL (). (A) 100–300 m, (B) 300–500 m, (C) 500–700 m, and (D) 700–900 m. Temperature at 25 ◦ C.
12
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
effect of irinotecan loading level on drug elution. The calculated effective diffusion coefficients are listed in Table 1. For the loading of 10 mg/mL, the value of F > 0.85 and Eq. (2) was used; for the high loading over 20 mg/mL, the value of F < 0.85 and Eq. (3) was used. The resulting Di values range from 2.4 × 10−9 to 1.4 × 10−7 cm2 s−1 , depending on drug loading and bead size. It is observed that in each size range of beads, the diffusion coefficients decrease with increasing drug loading, which is attributed to the denser structure of the DEB with more hydrophobic drug. The effect of size on Di however, is not as obvious due to the fluctuation of error in the values.
3.7.
Fig. 6 – Bt–t plots of drug loading effect on irinotecan release from 700 to 900 m DEB. (䊉) 20 mg/mL; () 40 mg/mL; () 50 mg/mL.
Fig. 7A shows the irinotecan release from beads (100–300 m at 34.1 mg/mL and 700–900 m at 48.3 mg/mL) into PBS at 37 ◦ C measured by T-apparatus. The elution profile of 100–300 m beads shows a significant burst of about 25% elution in the first 10 min, compared with the 700–900 m beads that only show a burst of less than 5% elution within 6 min. After the initial burst, the trend of the drug elution rate for the two size beads was similar. After 18 h only ∼50% and ∼65% of drug was released from the large and small beads, respectively. This is thought to be more akin to the conditions in vivo where the beads are restrained within an artery.
process, the fraction of solute released, F, could be expressed as below: Mt 6 e−n Bt =1− 2 M∞ n2 ∞
F=
2
(1)
n=1
where Mt is the amount of solute released after time t; M∞ the amount of solute released after infinite time; B = 2 Di /r2 , and Di is the effective diffusion coefficient of solute inside the resin particle. At high values of F (>0.85), Reichenberg defined: Bt = − ln
2 (1 − F) 6
3.8. Animal test data and correlation with data from T-apparatus
(2)
Irinotecan doses ranged from 92 to 200 mg for Group 2 and from 104 to 200 mg for Group 3. The variation in drug dose administered was due to variability in the volume of beads that could be delivered before embolization was judged to be complete. Group 4 animals received 164 mg of irinotecan suspended in saline/contrast media at a dose equivalent to the mean dose received by Group 3 animals. For the plasma level analysis, all data were normalized to the maximum 200 mg dose for the purpose of comparison. Fig. 7B shows the irinotecan plasma levels of the different groups over time. In Group 2 small beads (100–300 m) loaded with irinotecan were used, and the Tmax was achieved at 2 min with a Cmax of 1791 ± 520 ng/mL (mean ± S.D.). For large beads with irinotecan DEB (700–900 m, Group 3) the Tmax was achieved at 5 min with a Cmax of 1479 ± 703 ng/mL. Group 4 had an intra-arterial bolus of irinotecan without beads, for which the Tmax was at 0 min with Cmax of 4916 ± 516 ng/mL, much higher than DEB groups. A similar pattern was observed over time for SN-38 (the main active metabolite of irinotecan),
As F < 0.85, Eq. (3) was used: Bt = 2 −
2 F F − 2 1 − 6 3
In vitro irinotecan release assessed in T-apparatus
1/2 (3)
A linear Bt versus t indicates that the solute diffusion within the resin particle, “particle diffusion”, is the rate-controlling step. From the slope of Bt–t, the Di can be determined by using the mean size of particle (or beads). Fig. 6 shows the Bt–t plot derived from the irinotecan elution data of 700–900 m beads with doses of 20, 40, and 50 mg/mL (Fig. 5D), in which the elution values less than 0.85 were used. The linear Bt–t plot suggests that irinotecan elution from DEB into PBS is controlled by the drug diffusion within the beads, which is consistent with the observation of drug release from other ion-exchange resins (Chretien et al., 2004). The slope change with dose indicates a significant
Table 1 – Fitted effective diffusion coefficientsa of irinotecan release from DEB with different size range and loading Campto® loading (mg/mL) 10 20 40 50 a b
100–300 m (×10−8 ) 1.32 0.97 0.53 0.24
In cm2 s−1 . Elution values larger than 0.85 were used.
± ± ± ±
4.21b 1.16 0.12 0.13
300–500 m (×10−8 ) 8.10 3.07 0.95 0.57
± ± ± ±
4.28b 1.23 0.39 0.55
500–700 m (×10−8 ) 14.46 4.40 1.35 1.10
± ± ± ±
6.67b 1.95 0.19 0.31
700–900 m (×10−8 ) 13.16 5.91 0.94 0.86
± ± ± ±
5.85b 1.35 0.15 0.25
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
13
Fig. 8 – Multiple irinotecan DEB (100–300 m) showing uniform basophilia and artefactual folding in the portal zone. Abundant periportal fibrosis with granuloma formation and lymphocytic and eosinophilic infiltration (scale bar = 200 m).
results confirm that the T-apparatus test is a useful model to predict the in vivo release of drug from irinotecan DEB. Upon histopathological examination of the liver, the injection of drug alone produced minimal vascular changes and no significant parenchymal injury. Necrotising vasculopathy in the arteries/arterioles containing beads was observed and subsequently a granulomatous inflammatory reaction around the beads, as typical of embolization (Fig. 8). This did not result in any significant liver damage as indicated by the normal levels of liver enzymes, unlike the pan necrosis and elevated aspartate transferase (AST) observed with doxorubicin DEB (Lewis et al., 2006b). The hepatic pathology with injected beads only (small) was of a similar nature and severity to that found with beads and drug (small or large).
4.
Fig. 7 – (A) T-apparatus study of irinotecan release from DEB 100 to 300 m (), and 700–900 m () into PBS buffer at 37 ◦ C. (B) In vivo irinotecan plasma levels (n = 5 per group) from group of intra-arterial bolus of irinotecan (), 100–300 m group (), and 700–900 m group (). (C) In vitro–in vivo correlation data of irinotecan release from DEB with the size range of 100–300 m () and 700–900 m ().
except as anticipated, the levels detected were some 300–400fold lower (data not shown). Fig. 7C shows the in vitro–in vivo (AUC) correlation data of irinotecan DEB with the size ranges 100–300 m and 700–900 m. The drug elution data from T-apparatus correlate well with the in vivo data resulting in a good linear IVIVC (in vitro–in vivo correlation) with correlation coefficients of 0.98 and 0.99 for small and large bead sizes, respectively. These
Conclusions
It has been demonstrated in this study that DEB of sizes ranging from 100 to 900 m can load irinotecan up to a maximum bound capacity of 50 mg/mL of beads, by direct immersion of the device in irinotecan solution such as the commercially available Campto formulation. Drug loading and subsequent release is controlled by an ion-exchange mechanism between the positively charged irinotecan hydrochloride salt and the sulfonate-containing hydrogel beads. Absorption of drug is accompanied by a decrease in volume and increase in the force required to compress the beads, although neither significantly impact upon the handling or delivery through microcatheters, crucial properties for devices used in transarterial chemoembolization. Elution of the drug was measured by two different in vitro methods. Using an infinite sink approach, drug release was shown to be related to bead size, the smaller beads eluting the drug faster primarily due to the increased surface area effect. Release was also modelled using a T-apparatus which has been developed to emulate the diffusion and convection effects experienced by such drug-eluting devices when placed
14
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
within the vessel at the site of embolization. This in vitro release was shown to correlate very well with the levels of drug measured in plasma in vivo post embolization in a porcine hepatic arterial embolization model with the irinotecan DEB. The Cmax and AUC were smaller and the Tmax later for irinotecan released from the DEB when compared to an intraarterial bolus of the drug. We have also reported a linear IVIVC for doxorubicin DEB comparing T-apparatus elution with in vivo plasma levels from a similar porcine hepatic arterial embolization model (Lewis et al., 2006b), which demonstrates the usefulness and utility of the T-apparatus as a tool for predicting initial plasma levels that may be expected post embolization with DEB. This is of importance when considering systemic exposure to very toxic agents such as doxorubicin where car¨ diac toxicity is lifetime dose-limiting (Muller et al., 2003). Histopathology revealed that embolization with irinotecan DEB results in no unexpected effects, with minimal damage to the associated liver tissue. This is in contrast to the effects observed with small doxorubicin DEB which can cause extensive pan necrosis (Lewis et al., 2006b). A recent study of the chemoembolization of rat colorectal liver metastases demonstrated that both irinotecan and doxorubicin DEB showed significant anti-tumoral activity, although irinotecan was preferred because of its wider therapeutic range (Eyol et al., 2006). There is sufficient evidence therefore, to support the clinical evaluation of irinotecan DEB in the treatment of colorectal metastases to the liver, and to this end this drug-device combination is currently being evaluated in a pilot clinical study (Aliberti et al., 2006).
Acknowledgements We would like to thank the team at the Institute of Medical and Veterinary Services in Adelaide, Australia, for their assistance in performing the in vivo studies.
references
Aliberti, C., Tilli, M., Benea, G., Fiorentini, G., 2006. Trans-arterial chemoembolization (TACE) of liver metastases from colorectal cancer using irinotecan-eluting beads: preliminary results. Anticancer Research 26, 3779–3782. Amyot, F., Boudy, V., Jurski, K., Counord, J.-L., Guiffant, G., Dufaux, J., Chaumeil, J.-C., 2002. A new experimental method for the evaluation of the release profiles of drug-loaded microbeads designed for embolisation. ITBM-RBM 23, 285–289. Boyd, G.E., Adamson, A.W., Myers, L.S., 1947. The exchange adsorption of ions from aqueous solutions by organic zeolites. II. Kinetics. JACS 69, 2836–2848. Chretien, C., Boudy, V., Allain, P., Chaumeil, J.C., 2004. Indomethacin release from ion-exchange microspheres: impregnation with alginate reduces release rate. J. Control. Rel. 96, 369–378. ´ Escoriaza, J., Aldaz, A., Castellanos, C., Calvo, E., Giradles, J., 2000. Simple and rapid determination of irinotecan and its metabolite SN-38 in plasma by high performance liquid-chromatography: application to clinical
pharmacokinetic studies. J. Chromatogr. B: Biomed. Sci. Appl. 740, 159–168. Eyol, E., Lewis, A.L., Taylor, R.R., Berger, M., 2006. Chemoembolization of rat colorectal liver metastases with drug eluting beads loaded with irinotecan or doxorubicin. In: Proceedings of the AACR 97th Annual Meeting, Washington. Ferlay, J., Bray, F., Pisani, P., Parkin, D.M., 2004. GLOBOCAN 2002: cancer incidence, mortality and prevalence worldwide. In: IARC CancerBase No. 5. Version 2.0. IARC Press, Lyon. Giroux, M.-F., Baum, R.A., Soulen, M.C., 2004. Chemoembolization of liver metastases from breast carcinoma. J. Vasc. Interv. Radiol. 15, 289–291. Jantzen, G.M., Robinson, J.R., 1996. Sustained and controlled-release drug delivery systems. In: Banker, G.S., Rhodes, C.T. (Eds.), Modern Pharmaceutics, 3rd ed. Marcel Dekker, Inc., New York. Lewis, A.L., Gonzalez, M.V., Hall, B., Willis, S.L., Leppard, S.W., Tang, Y., Palmer, R.R., Wolfenden, L.C., Stratford, P.W., Phillips, G.J., Lloyd, A.W., 2005. Doxorubicin-eluting beads for chemoembolization, Transactions of the Controlled Release Society 32nd Annual Meeting & Expositions, #021. Lewis, A.L., Gonzalez, M.V., Lloyd, A.W., Hall, B., Tang, Y., Willis, S.L., Leppard, S.W., Wolfenden, L.C., Palmer, R.R., Stratford, P.W., 2006a. DC Bead: in vitro characterization of a drug-delivery device for transarterial chemoembolization. J. Vasc. Interv. Radiol. 17, 335–342. Lewis, A.L., Taylor, R.R., Hall, B., Willis, S.L., Stratford, P.W., 2006b. Pharmacokinetic and safety study of doxorubicin eluting beads in a porcine model of hepatic arterial embolisation. J. Vasc. Interven. Radiol. 17, 1335–1343. Liu, Z., Cheung, R., Wu, X.Y., Ballinger, J.R., Bendayan, R., Rauth, A.M., 2001. A study of doxorubicin loading onto and release from sulfopropyl dextran ion-exchange microspheres. J. Control. Rel. 77, 213–224. Lorenz, K., Brauckhoff, M., Behrmann, C., Sekulla, C., Ukkat, J., Brauckhoff, K., Gimm, O., Dralle, H., 2005. Selective arterial chemoembolization for hepatic metastases from medullary thyroid carcinoma. Surgery 138, 986–993. ¨ Muller, H.J., Port, R.E., Grubert, M., Hilger, R.A., Scheulen, M., Mross, K., 2003. The influence of liver metastases on the pharmacokinetics of doxorubicin- a population-based pharmacokinetic project of the CESAR-APOH. Int. J. Clin. Pharmacol Therapeut. 41, 598–599. O’Toole, D., Ruszniewski, P., 2005. Chemoembolization and other ablative therapies for liver metastases of gastrointestinal endocrine tumours. Best Pract. Res. Clin. Gastroenterol. 19, 585–594. Reichenberg, D., 1953. Properties of ion-exchange resins in relation to their structure. III. Kinetics of exchange. JACS 75, 589–597. Sanz-Altamira, P.M., Spence, L.D., Huberman, M.S., Posner, M.R., Steele Jr., G., Perry, L.J., Stuart, K.E., 1997. Selective chemoembolization in the management of hepatic metastases in refractory colorectal carcinoma: a phase II trial. Dis. Colon. Rectum. 40, 770–775. Stuart, K., 2003. Chemoembolization in the management of liver tumors. Oncologist 8, 425–437. Tellez, C., Benson, A.B., Lyster, M.T., Talamonti, M., Shaw, J., Braun, M.A., Nemcek Jr., A.A., Vogelzang, R.L., 1998. Phase II trial of chemoembolization for the treatment of metastatic colorectal carcinoma to the liver and review of the literature. Cancer 82, 1250–1259. Wallace, S., Carrasco, C.H., Charnsangavej, C., Richli, W.R., Wright, K., Gianturco, C., 1990. Hepatic artery infusion and chemoembolization the management of liver metastases. Cardiovasc. Intervent. Radiol. 13, 153–160.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
Irinotecan drug eluting beads for use in chemoembolization: In vitro and in vivo evaluation of drug release properties Rachel R. Taylor, Yiqing Tang, M. Victoria Gonzalez, Peter W. Stratford, Andrew L. Lewis ∗ Biocompatibles UK Ltd., Farnham Business Park, Weydon Lane, Farnham, Surrey GU9 8QL, UK
a r t i c l e
i n f o
a b s t r a c t
Article history:
Drug eluting beads that release irinotecan in a controlled manner may be useful for appli-
Received 8 May 2006
cation in the chemoembolization of colorectal cancer metastases to the liver. In this study,
Accepted 5 September 2006
irinotecan drug eluting beads were prepared with loadings up to 50 mg drug/mL hydrated
Published on line 15 September 2006
beads. Drug loading was via an ion-exchange mechanism with sulfonate binding sites in the bead. Release in vitro was shown to be sustained and dependent upon the presence of
Keywords:
ions in the elution medium, drug loading and bead size. Drug elution in PBS was controlled
Irinotecan drug eluting beads
by solute diffusion within the beads and gave rise to values for the diffusion coefficient,
Chemoembolization
D, of between 2.4 × 10−9 and 1.4 × 10−7 cm2 s−1 . The beads were shown to decrease in size
Drug release
(by a maximum 25–30%), and concomitantly their modulus of compression increased (from
In vitro:in vivo correlation
∼27 kPa to a maximum of about 49 kPa), with increasing drug loading. This did not however, influence their ability to be suspended homogeneously in contrast agent or delivered through a microcatheter. Following porcine hepatic artery embolization, maximum plasma levels were 70–75% lower for both irinotecan and SN-38 compared to intraarterial bolus administration, with peak levels observed at 2 and 5 min after completion of the embolization procedure. The in vivo data were shown to correlate well with the in vitro release measured using a T-apparatus model of embolization. © 2006 Elsevier B.V. All rights reserved.
1.
Introduction
Despite the existence of excellent screening and preventive strategies, colorectal carcinoma remains a major public health problem in western countries. The American Cancer Society estimates there will be 145,290 new cases diagnosed in 2005, and 56,290 people will die of the disease. In the UK, about 36,000 new cases were diagnosed in 2002, and 17,190 people died of colorectal cancer (Ferlay et al., 2004). Colorectal carcinoma is the third leading cause of death from cancer in both males and females in the USA. It is also the third most common malignancy in both men (after prostate and lung cancers) and women (after breast and lung cancers).
∗
Corresponding author. Tel.: +44 1252 732819. E-mail address: [email protected] (A.L. Lewis). 0928-0987/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ejps.2006.09.002
Five-year survival rates have improved in recent years for patients with stages I–III colorectal carcinomas; for patients with early, localized disease the survival rate is 90%. Survival has improved through the use of adjuvant chemotherapy for colon cancer and adjuvant chemoradiation for rectal cancer. By the time they are diagnosed, some 25% of colon cancers will have extended through the bowel wall, whereas cancers of the rectum will have spread through the bowel wall in 50–70% of patients and metastasized to lymph nodes in 50–60%. The most common site of extralymphatic involvement is the liver, with the lungs the most frequently affected extra-abdominal organ. Patients with metastatic colorectal tumours frequently die of hepatic failure due to liver metastases.
8
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Treatment options for patients with metastatic colorectal cancer are limited and clinical outcome is generally poor. Systemic chemotherapy can palliate symptoms and improve survival. 5-Fluorouracil (5-FU) based chemotherapy has been the cornerstone of treatment of CRC for more than 40 years, and new drugs with a definite activity have recently broadened the options for treatment. Irinotecan, a topoisomerase inhibitor, was the first drug other than 5-FU to demonstrate significant activity. This product is approved for use in combination with 5-FU/folinic acid in patients without prior chemotherapy and for the second-line treatment of this disease as a single agent in patients who have failed an established 5-FU containing treatment regimen. There is renewed interest in locoregional treatment of the liver in an attempt to more effectively target liver metastases, an example of which is transarterial chemoembolization (TACE). The technique consists of delivering high concentrations of chemotherapeutic agents emulsified in an oil-based medium directly to the tumour bed followed by some form of embolization. The purpose of the embolization is two-fold: nutrient and oxygen starvation of the tumour, and minimization of chemotherapy wash-out with prolonged contact with tumor tissue. These factors combine to significantly reduce systemic toxicity of the chemotherapy. Such an approach has also been investigated in a number of metastatic settings with unresectable liver tumours, including colorectal cancer (Tellez et al., 1998; Sanz-Altamira et al., 1997; Wallace et al., 1990; Stuart, 2003; Giroux et al., 2004; Lorenz et al., 2005; O’Toole and Ruszniewski, 2005). Irinotecan drug eluting beads (DEB) combine the drug with the embolization device and can be administered intraarterially in the same manner as TACE. This drug-device combination may offer the possibility of precisely controlling the release and dose of the drug into the tumor bed. These beads obviate the need for an oil/chemotherapy suspension in addition to the embolic device and facilitate handling and delivery, achieving TACE in a simpler one-step procedure (Lewis et al., 2006a). This study presents both in vitro characterization of irinotecan DEB, with respect to how the drug influences the physical properties and handling of the device, and how the device matrix is able to modulate release of the drug over a therapeutically meaningful timeframe. Moreover, these data are correlated to the plasma pharmacokinetics obtained from a porcine model of hepatic arterial embolization.
The compressibility of 700–900 m and 900–1200 m beads with and without drug loading was assessed using an Instron 4411 with Series IX software (Instron, Canton, MA, USA). The general procedure involved distributing the beads onto a specially fabricated sample holder on the Instron to form a monolayer of about 7–10 mm in width. The residual water around the beads was carefully removed by wicking. The diameter of the probe used was 5.0 mm with a crosshead speed of 5.0 mm/min. The load cell used was 50 N and the specimen gauge length was set as 1.0 mm. All the experiments were carried out at room temperature.
2.
Materials and methods
2.5. Suspension and catheter deliverability of irinotecan DEB
2.1.
Materials
Irinotecan HCl solution (Campto® , Pfizer) of 100 mg/5 mL containing 45 mg/mL sorbitol and 0.9 mg/mL lactic acid at pH 3.5 adjusted by adding sodium hydroxide or hydrochloric acid, was used directly as received. DC BeadTM (hydrogel beads from Biocompatibles UK Ltd., Farnham, UK) were used in the size ranges 100–300 m, 300–500 m, 500–700 m, and 700–900 m (and 900–1200 m made available specifically for compressibility studies). Other solutions used included phosphate buffer
saline solution (PBS) (Inverclyde, Bellshill, UK), and Omnipaque 350 (Amersham, Oslo, Norway).
2.2. Preparation of irinotecan DEB and determination of drug loading level In a typical loading procedure, the packing solution was removed from the beads. Subsequently, the required volume of Campto® solution was added to the bead slurry, and the mixture was roller-mixed. The drug-loaded beads were stored in the dark at 4–8 ◦ C. The dose of drug used in this study ranged from 10 to 90 mg/mL. The loading of irinotecan was calculated by measuring the remaining drug concentration in the solution over time. The measurements were carried out with a Perkin-Elmer Lambda 25 UV spectrophotometer at the wavelength of 369 nm and the irinotecan concentration was determined by comparing with a standard curve. The loading efficiency was calculated as below: loading efficiency (%) =
2.3.
initial drug in solution − residual drug in solution × 100 initial drug in solution
Bead sizing and appearance
The size and appearance of beads during irinotecan loading and elution were examined by use of an Olympus BX50F4 microscope with a Colorview III camcorder (Olympus, Japan), under an aqueous solution. Sizing data were analysed by AnalySIS (Soft Imaging System GmbH) and converted to a frequency histogram with 20 m intervals across at least a 200 m range.
2.4.
Compressibility of irinotecan beads
Bead suspension and catheter deliverability were studied by mixing 1 mL of beads, containing various doses of irinotecan, with 6 mL of contrast medium (Omnipaque 350)/saline (50:50 ratio) in a 10 mL syringe. After mixing through a three-way connector, the homogeneity of this suspension was assessed visually. A mixture was considered homogeneous when the beads were suspended for more than 1 min. The deliverability of DEB was assessed by rating the ease of their delivery down a microcatheter (Progreat® , Terumo Corp., Tokyo, Japan). Microcatheters of 2.4–3 Fr were employed to
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
deliver the different sized beads, with the acceptance criteria of pass/failure based on whether the lumen was occluded during delivery.
2.6.
In vitro release of irinotecan into various media
The elution profiles of irinotecan from DEB were evaluated in water, PBS, and contrast media, respectively. To measure drug release, irinotecan DEB were first suspended in 200 mL of elution medium and drug release over time was determined from sample aliquots by UV spectroscopy at 369 nm. In studies involving contrast medium, 1 mL of irinotecan DEB were rollermixed with 1 mL saline and 2 mL contrast medium (Omnipaque 350, Amersham). The aliquot of solution was removed and diluted with PBS for UV measurement.
2.7. Release of irinotecan using an in vitro embolization model A T-apparatus was designed to study the in vitro release of drug from DEB, as previously described (Amyot et al., 2002). One milliliter of irinotecan DEB was placed carefully in the bottom of the well in the T-apparatus together with 300 mL PBS and maintained at 37 ◦ C. The solution was circulated using a Watson-Marlow 505 U peristaltic pump (Watson-Marlow, Falmouth, UK) at 25 rpm (∼50 mL/min), and passed through a Perkin-Elmer Lambda 25 UV spectrophotometer (Pekin-Elmer, Shelton, CT, USA). The change of absorbance with time was recorded at 369 nm.
2.8.
Preclinical model of hepatic arterial embolization
Hepatic arterial embolization was performed in four groups of crossbred pigs of mixed sex (n = 5 per group): 100–300 m control beads with no drug, 100–300 m irinotecan DEB, 700–900 m irinotecan DEB and intra-arterial bolus injection of drug alone. Each animal underwent a single interventional procedure on day 0, during which a catheter was advanced into the left hepatic lobe or segment, and up to 4 mL of either loaded or unloaded beads in a total volume of 24 mL (including saline and contrast media) were slowly administered until embolization was achieved. Three animals per group were euthanized on day 32 (±2 days) and two animals per group were euthanized on day 90 (±2 days). All animals underwent a comprehensive necropsy, and selected organs and tissues were collected for histopathological evaluation. Plasma samples for the measurement of irinotecan and SN-38 plasma levels were taken over 90 days and analysed according to a reported literature method (Escoriaza et al., 2000).
3.
Results and discussion
3.1.
Maximum capacity of irinotecan loading
Irinotecan drug uptake into the beads increased linearly up to ∼50 mg of drug, with loading efficiencies >99% when loaded for 24 h. Above 60 mg, the loading efficiency was <90% for both the 100–300 m and 500–700 m bead sizes used for this study. The loss of loading efficiency around 50 mg for both sizes indicates that the sulfonate sites in the beads were satu-
9
rated with irinotecan through charge–charge interaction, and further addition of Campto® could not increase the loading significantly. The mechanism of bead uptake of irinotecan is attributed to the interaction between the positively charged irinotecan salt which bears a protonated amine group, and negatively charged sulfonic acid groups (from the 2-acrylamido2-methylpropane sulfonate sodium salt, AMPS) on the beads. A similar interaction has been observed in the system containing doxorubicin and sulfonated dextran microspheres (Liu et al., 2001). According to the quantity of AMPS added during bead manufacture and assuming that there is no material loss during synthesis, a preliminary estimation indicates that there are about 7.7 × 10−5 mol of AMPS sites in 1 mL of beads with the size range of 500–700 m. The value is equivalent to 52 mg irinotecan bound, assuming 100% occupation of sulfonate sites and no interaction between bound irinotecan. This theoretical value corresponds well with the observed limit of linearity of ∼50 mg.
3.2.
Irinotecan loading profiles
For clinical application it was desirable to determine the loading characteristics over a 2 h period. The loading of low doses (10 mg/mL) of irinotecan was rapid with >98% of the drug loaded within 10 min. For drug loading levels between 10 and 40 mg/mL, loading efficiencies >99.4% were obtained for all bead sizes within 2 h. For the optimal limit of 50 mg/mL as determined in Section 3.1, the loading percentage after 2 h was 98.1 ± 1.4% for 100–300 m, 98.0 ± 1.3% for 300–500 m, 95.0 ± 0.2% for 500–700 m, and 93.5 ± 0.7% for 700–900 m, respectively. It was observed that the rate of drug uptake for larger bead sizes was slower than for smaller ones. This effect has been observed previously with doxorubicin, and is attributed to decrease in surface area with increase in bead diameter (Lewis et al., 2005).
3.3. Appearance and size change post irinotecan loading Fig. 1 shows the change in appearance of 300–500 m beads followed by light microscopy during irinotecan loading (target loading 50 mg per mL beads). These images demonstrate that the loading is a rapid procedure, and the blue transparent beads turned a green-turquoise non-transparent hue after about 10 min, which is consistent with the data in Section 3.1. During drug loading, the beads retained their spherical shape, and no aggregation of the beads was observed. Fig. 2 shows the mean diameter change of beads with increasing dose of irinotecan. All sizes show a tendency to decrease in size with increasing drug loading. Although the mean diameter reduced by a maximum of 25–30% in all cases, for the most relevant clinical size (100–300 m) the diameter changes was insignificant. The size (diameter) change associated with drug loading is related to the displacement of water from the beads by the more hydrophobic drug. The relatively tight standard deviation of each mean size measurement demonstrates that the size distribution of beads before and after irinotecan loading does not change significantly. The spherical shape of the beads is retained upon drug loading.
10
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Fig. 1 – Micrographs of morphology of DEB (300–500 m) during irinotecan loading (50 mg/mL). (A) Without drug loading, (B) after 7 min loading, and (C) after 20 min. The scale bar shown is 200 m.
3.4.
Compressibility of irinotecan DEB
Fig. 3 shows the force required to compress beads (modulus) of 700–900 m and 900–1200 m size ranges, with increasing drug loading. The DEB shows a slight trend for increase in modulus with the drug content. The trend is consistent with the decreasing water content and bead size as described above. The higher drug loading results in a denser structure, which exhibits a compressive response more akin to elastic materials. It is noted however, that even the highest dose of irinotecan did not increase the modulus significantly, which is still within the same order of magnitude as unloaded beads, unlike the behaviour observed with these beads when loaded with doxorubicin (Lewis et al., 2006a). The beads loaded with drug maintained their compressible hydrogel structure, which should facilitate the delivery through microcatheters and hence not compromise their use in embolization.
Fig. 3 – Compression modulus of 700–900 m ( ) and 900–1200 m () beads with increasing irinotecan loading.
3.5.
Suspension and catheter deliverability studies
A homogeneous suspension was achieved for all bead sizes and doses in a time ranging from 1 to 13 min. The time to suspension increased with the bead size and decreased as the drug loading increased. For the optimal loading of 50 mg/mL, a homogeneous suspension was achieved in <1 min for the smallest beads and approximately 5 min for the largest bead size range. Suspensions of 100–300 m and 300–500 m beads in saline-contrast medium were easily delivered through a 2.4 Fr catheter, whereas 500–700 m and 700–900 m beads required a 2.7 Fr and 3 Fr catheter, respectively for successful delivery. No bead aggregation or related catheter blockage were observed.
3.6. Fig. 2 – Mean diameter change of DEB with and without irinotecan loading: 100–300 m (); 300–500 m (); 500–700 m (䊉); 700–900 m ().
In vitro release of irinotecan DEB
Fig. 4 shows a release profile of irinotecan DEB that were first placed in deionised water and subsequently in PBS buffer after
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
Fig. 4 – Comparison of irinotecan elution in water and PBS at 25 ◦ C (300–500 m DEB, loading 20 mg/mL).
removal from the water. There was no drug released into the deionised water, but release was immediate and rapid when the beads were exposed to PBS. This demonstrates the need for ions in the elution medium in order to displace the protonated irinotecan from the sulfonate binding sites, further supporting the mechanism of ionic interaction between drug and bead. Note that the size of the DEB in pure water increased significantly due to a water content increase under the osmotic pressure in the beads; this could facilitate the more rapid elution of drug from the DEB when subsequently placed in PBS compared to DEB placed in PBS alone. Using a free-flowing in vitro elution method, it was demonstrated that all of the loaded drug could be eluted from the DEB in all cases. The rate of irinotecan elution from the beads into
11
PBS buffer is shown to be dependent upon bead size (Fig. 5). Smaller beads tend to have much faster release rate compared to larger beads with equivalent loading (Fig. 5A versus D). This can be explained by the effect of the larger surface area of smaller beads for the same volume, enhancing the ionexchange rate. Moreover, for the larger bead sizes the distance for drug diffusion within the microspheres is greater (Jantzen and Robinson, 1996). Beads with higher irinotecan loading become more hydrophobic compared to those with low loading, which in turn reduces the ion-exchange and diffusion rates of both salt and drug ions. The influence of the change in hydrophobicity of loaded beads can be evaluated from the volume change of microspheres with and without irinotecan loading. In the previous measurements of size change, for example, the mean size of unloaded 700–900 m beads changed from 830 to 806 m with 10 mg/mL loading, and to 584 m with 50 mg/mL loading. Therefore, the coincident volume decrease is 8.4%, and 65.2%, respectively, which means that at least 2/3 of the original water volume was expelled from the microspheres with 50 mg/mL loading. Despite the fact that there is a decrease in the diffusional distance related to the reduced bead size, the loss of water results in a denser structure after drug loading which significantly retards the drug elution rate. Diffusion in ion-exchange resins has been studied in detail by Boyd and Reichenberg (Boyd et al., 1947; Reichenberg, 1953). The rate of diffusion in ion-exchange resins can be controlled by chemical exchange or solute diffusion, which happen within resins, or film diffusion in which the in-going ion is a macrocomponent of the system. It has been demonstrated that indomethacin release from ion-exchange microspheres was controlled by solute diffusion (particle diffusion) (Chretien et al., 2004). When solute diffusion was the rate-controlling
Fig. 5 – Release profiles into PBS buffer of irinotecan DEB with different loadings: 10 (), 20 (), 40 (䊉), and 50 mg/mL (). (A) 100–300 m, (B) 300–500 m, (C) 500–700 m, and (D) 700–900 m. Temperature at 25 ◦ C.
12
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
effect of irinotecan loading level on drug elution. The calculated effective diffusion coefficients are listed in Table 1. For the loading of 10 mg/mL, the value of F > 0.85 and Eq. (2) was used; for the high loading over 20 mg/mL, the value of F < 0.85 and Eq. (3) was used. The resulting Di values range from 2.4 × 10−9 to 1.4 × 10−7 cm2 s−1 , depending on drug loading and bead size. It is observed that in each size range of beads, the diffusion coefficients decrease with increasing drug loading, which is attributed to the denser structure of the DEB with more hydrophobic drug. The effect of size on Di however, is not as obvious due to the fluctuation of error in the values.
3.7.
Fig. 6 – Bt–t plots of drug loading effect on irinotecan release from 700 to 900 m DEB. (䊉) 20 mg/mL; () 40 mg/mL; () 50 mg/mL.
Fig. 7A shows the irinotecan release from beads (100–300 m at 34.1 mg/mL and 700–900 m at 48.3 mg/mL) into PBS at 37 ◦ C measured by T-apparatus. The elution profile of 100–300 m beads shows a significant burst of about 25% elution in the first 10 min, compared with the 700–900 m beads that only show a burst of less than 5% elution within 6 min. After the initial burst, the trend of the drug elution rate for the two size beads was similar. After 18 h only ∼50% and ∼65% of drug was released from the large and small beads, respectively. This is thought to be more akin to the conditions in vivo where the beads are restrained within an artery.
process, the fraction of solute released, F, could be expressed as below: Mt 6 e−n Bt =1− 2 M∞ n2 ∞
F=
2
(1)
n=1
where Mt is the amount of solute released after time t; M∞ the amount of solute released after infinite time; B = 2 Di /r2 , and Di is the effective diffusion coefficient of solute inside the resin particle. At high values of F (>0.85), Reichenberg defined: Bt = − ln
2 (1 − F) 6
3.8. Animal test data and correlation with data from T-apparatus
(2)
Irinotecan doses ranged from 92 to 200 mg for Group 2 and from 104 to 200 mg for Group 3. The variation in drug dose administered was due to variability in the volume of beads that could be delivered before embolization was judged to be complete. Group 4 animals received 164 mg of irinotecan suspended in saline/contrast media at a dose equivalent to the mean dose received by Group 3 animals. For the plasma level analysis, all data were normalized to the maximum 200 mg dose for the purpose of comparison. Fig. 7B shows the irinotecan plasma levels of the different groups over time. In Group 2 small beads (100–300 m) loaded with irinotecan were used, and the Tmax was achieved at 2 min with a Cmax of 1791 ± 520 ng/mL (mean ± S.D.). For large beads with irinotecan DEB (700–900 m, Group 3) the Tmax was achieved at 5 min with a Cmax of 1479 ± 703 ng/mL. Group 4 had an intra-arterial bolus of irinotecan without beads, for which the Tmax was at 0 min with Cmax of 4916 ± 516 ng/mL, much higher than DEB groups. A similar pattern was observed over time for SN-38 (the main active metabolite of irinotecan),
As F < 0.85, Eq. (3) was used: Bt = 2 −
2 F F − 2 1 − 6 3
In vitro irinotecan release assessed in T-apparatus
1/2 (3)
A linear Bt versus t indicates that the solute diffusion within the resin particle, “particle diffusion”, is the rate-controlling step. From the slope of Bt–t, the Di can be determined by using the mean size of particle (or beads). Fig. 6 shows the Bt–t plot derived from the irinotecan elution data of 700–900 m beads with doses of 20, 40, and 50 mg/mL (Fig. 5D), in which the elution values less than 0.85 were used. The linear Bt–t plot suggests that irinotecan elution from DEB into PBS is controlled by the drug diffusion within the beads, which is consistent with the observation of drug release from other ion-exchange resins (Chretien et al., 2004). The slope change with dose indicates a significant
Table 1 – Fitted effective diffusion coefficientsa of irinotecan release from DEB with different size range and loading Campto® loading (mg/mL) 10 20 40 50 a b
100–300 m (×10−8 ) 1.32 0.97 0.53 0.24
In cm2 s−1 . Elution values larger than 0.85 were used.
± ± ± ±
4.21b 1.16 0.12 0.13
300–500 m (×10−8 ) 8.10 3.07 0.95 0.57
± ± ± ±
4.28b 1.23 0.39 0.55
500–700 m (×10−8 ) 14.46 4.40 1.35 1.10
± ± ± ±
6.67b 1.95 0.19 0.31
700–900 m (×10−8 ) 13.16 5.91 0.94 0.86
± ± ± ±
5.85b 1.35 0.15 0.25
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
13
Fig. 8 – Multiple irinotecan DEB (100–300 m) showing uniform basophilia and artefactual folding in the portal zone. Abundant periportal fibrosis with granuloma formation and lymphocytic and eosinophilic infiltration (scale bar = 200 m).
results confirm that the T-apparatus test is a useful model to predict the in vivo release of drug from irinotecan DEB. Upon histopathological examination of the liver, the injection of drug alone produced minimal vascular changes and no significant parenchymal injury. Necrotising vasculopathy in the arteries/arterioles containing beads was observed and subsequently a granulomatous inflammatory reaction around the beads, as typical of embolization (Fig. 8). This did not result in any significant liver damage as indicated by the normal levels of liver enzymes, unlike the pan necrosis and elevated aspartate transferase (AST) observed with doxorubicin DEB (Lewis et al., 2006b). The hepatic pathology with injected beads only (small) was of a similar nature and severity to that found with beads and drug (small or large).
4.
Fig. 7 – (A) T-apparatus study of irinotecan release from DEB 100 to 300 m (), and 700–900 m () into PBS buffer at 37 ◦ C. (B) In vivo irinotecan plasma levels (n = 5 per group) from group of intra-arterial bolus of irinotecan (), 100–300 m group (), and 700–900 m group (). (C) In vitro–in vivo correlation data of irinotecan release from DEB with the size range of 100–300 m () and 700–900 m ().
except as anticipated, the levels detected were some 300–400fold lower (data not shown). Fig. 7C shows the in vitro–in vivo (AUC) correlation data of irinotecan DEB with the size ranges 100–300 m and 700–900 m. The drug elution data from T-apparatus correlate well with the in vivo data resulting in a good linear IVIVC (in vitro–in vivo correlation) with correlation coefficients of 0.98 and 0.99 for small and large bead sizes, respectively. These
Conclusions
It has been demonstrated in this study that DEB of sizes ranging from 100 to 900 m can load irinotecan up to a maximum bound capacity of 50 mg/mL of beads, by direct immersion of the device in irinotecan solution such as the commercially available Campto formulation. Drug loading and subsequent release is controlled by an ion-exchange mechanism between the positively charged irinotecan hydrochloride salt and the sulfonate-containing hydrogel beads. Absorption of drug is accompanied by a decrease in volume and increase in the force required to compress the beads, although neither significantly impact upon the handling or delivery through microcatheters, crucial properties for devices used in transarterial chemoembolization. Elution of the drug was measured by two different in vitro methods. Using an infinite sink approach, drug release was shown to be related to bead size, the smaller beads eluting the drug faster primarily due to the increased surface area effect. Release was also modelled using a T-apparatus which has been developed to emulate the diffusion and convection effects experienced by such drug-eluting devices when placed
14
e u r o p e a n j o u r n a l o f p h a r m a c e u t i c a l s c i e n c e s 3 0 ( 2 0 0 7 ) 7–14
within the vessel at the site of embolization. This in vitro release was shown to correlate very well with the levels of drug measured in plasma in vivo post embolization in a porcine hepatic arterial embolization model with the irinotecan DEB. The Cmax and AUC were smaller and the Tmax later for irinotecan released from the DEB when compared to an intraarterial bolus of the drug. We have also reported a linear IVIVC for doxorubicin DEB comparing T-apparatus elution with in vivo plasma levels from a similar porcine hepatic arterial embolization model (Lewis et al., 2006b), which demonstrates the usefulness and utility of the T-apparatus as a tool for predicting initial plasma levels that may be expected post embolization with DEB. This is of importance when considering systemic exposure to very toxic agents such as doxorubicin where car¨ diac toxicity is lifetime dose-limiting (Muller et al., 2003). Histopathology revealed that embolization with irinotecan DEB results in no unexpected effects, with minimal damage to the associated liver tissue. This is in contrast to the effects observed with small doxorubicin DEB which can cause extensive pan necrosis (Lewis et al., 2006b). A recent study of the chemoembolization of rat colorectal liver metastases demonstrated that both irinotecan and doxorubicin DEB showed significant anti-tumoral activity, although irinotecan was preferred because of its wider therapeutic range (Eyol et al., 2006). There is sufficient evidence therefore, to support the clinical evaluation of irinotecan DEB in the treatment of colorectal metastases to the liver, and to this end this drug-device combination is currently being evaluated in a pilot clinical study (Aliberti et al., 2006).
Acknowledgements We would like to thank the team at the Institute of Medical and Veterinary Services in Adelaide, Australia, for their assistance in performing the in vivo studies.
references
Aliberti, C., Tilli, M., Benea, G., Fiorentini, G., 2006. Trans-arterial chemoembolization (TACE) of liver metastases from colorectal cancer using irinotecan-eluting beads: preliminary results. Anticancer Research 26, 3779–3782. Amyot, F., Boudy, V., Jurski, K., Counord, J.-L., Guiffant, G., Dufaux, J., Chaumeil, J.-C., 2002. A new experimental method for the evaluation of the release profiles of drug-loaded microbeads designed for embolisation. ITBM-RBM 23, 285–289. Boyd, G.E., Adamson, A.W., Myers, L.S., 1947. The exchange adsorption of ions from aqueous solutions by organic zeolites. II. Kinetics. JACS 69, 2836–2848. Chretien, C., Boudy, V., Allain, P., Chaumeil, J.C., 2004. Indomethacin release from ion-exchange microspheres: impregnation with alginate reduces release rate. J. Control. Rel. 96, 369–378. ´ Escoriaza, J., Aldaz, A., Castellanos, C., Calvo, E., Giradles, J., 2000. Simple and rapid determination of irinotecan and its metabolite SN-38 in plasma by high performance liquid-chromatography: application to clinical
pharmacokinetic studies. J. Chromatogr. B: Biomed. Sci. Appl. 740, 159–168. Eyol, E., Lewis, A.L., Taylor, R.R., Berger, M., 2006. Chemoembolization of rat colorectal liver metastases with drug eluting beads loaded with irinotecan or doxorubicin. In: Proceedings of the AACR 97th Annual Meeting, Washington. Ferlay, J., Bray, F., Pisani, P., Parkin, D.M., 2004. GLOBOCAN 2002: cancer incidence, mortality and prevalence worldwide. In: IARC CancerBase No. 5. Version 2.0. IARC Press, Lyon. Giroux, M.-F., Baum, R.A., Soulen, M.C., 2004. Chemoembolization of liver metastases from breast carcinoma. J. Vasc. Interv. Radiol. 15, 289–291. Jantzen, G.M., Robinson, J.R., 1996. Sustained and controlled-release drug delivery systems. In: Banker, G.S., Rhodes, C.T. (Eds.), Modern Pharmaceutics, 3rd ed. Marcel Dekker, Inc., New York. Lewis, A.L., Gonzalez, M.V., Hall, B., Willis, S.L., Leppard, S.W., Tang, Y., Palmer, R.R., Wolfenden, L.C., Stratford, P.W., Phillips, G.J., Lloyd, A.W., 2005. Doxorubicin-eluting beads for chemoembolization, Transactions of the Controlled Release Society 32nd Annual Meeting & Expositions, #021. Lewis, A.L., Gonzalez, M.V., Lloyd, A.W., Hall, B., Tang, Y., Willis, S.L., Leppard, S.W., Wolfenden, L.C., Palmer, R.R., Stratford, P.W., 2006a. DC Bead: in vitro characterization of a drug-delivery device for transarterial chemoembolization. J. Vasc. Interv. Radiol. 17, 335–342. Lewis, A.L., Taylor, R.R., Hall, B., Willis, S.L., Stratford, P.W., 2006b. Pharmacokinetic and safety study of doxorubicin eluting beads in a porcine model of hepatic arterial embolisation. J. Vasc. Interven. Radiol. 17, 1335–1343. Liu, Z., Cheung, R., Wu, X.Y., Ballinger, J.R., Bendayan, R., Rauth, A.M., 2001. A study of doxorubicin loading onto and release from sulfopropyl dextran ion-exchange microspheres. J. Control. Rel. 77, 213–224. Lorenz, K., Brauckhoff, M., Behrmann, C., Sekulla, C., Ukkat, J., Brauckhoff, K., Gimm, O., Dralle, H., 2005. Selective arterial chemoembolization for hepatic metastases from medullary thyroid carcinoma. Surgery 138, 986–993. ¨ Muller, H.J., Port, R.E., Grubert, M., Hilger, R.A., Scheulen, M., Mross, K., 2003. The influence of liver metastases on the pharmacokinetics of doxorubicin- a population-based pharmacokinetic project of the CESAR-APOH. Int. J. Clin. Pharmacol Therapeut. 41, 598–599. O’Toole, D., Ruszniewski, P., 2005. Chemoembolization and other ablative therapies for liver metastases of gastrointestinal endocrine tumours. Best Pract. Res. Clin. Gastroenterol. 19, 585–594. Reichenberg, D., 1953. Properties of ion-exchange resins in relation to their structure. III. Kinetics of exchange. JACS 75, 589–597. Sanz-Altamira, P.M., Spence, L.D., Huberman, M.S., Posner, M.R., Steele Jr., G., Perry, L.J., Stuart, K.E., 1997. Selective chemoembolization in the management of hepatic metastases in refractory colorectal carcinoma: a phase II trial. Dis. Colon. Rectum. 40, 770–775. Stuart, K., 2003. Chemoembolization in the management of liver tumors. Oncologist 8, 425–437. Tellez, C., Benson, A.B., Lyster, M.T., Talamonti, M., Shaw, J., Braun, M.A., Nemcek Jr., A.A., Vogelzang, R.L., 1998. Phase II trial of chemoembolization for the treatment of metastatic colorectal carcinoma to the liver and review of the literature. Cancer 82, 1250–1259. Wallace, S., Carrasco, C.H., Charnsangavej, C., Richli, W.R., Wright, K., Gianturco, C., 1990. Hepatic artery infusion and chemoembolization the management of liver metastases. Cardiovasc. Intervent. Radiol. 13, 153–160.