(701h) Cyclodextrin-Crosslinked Injectable Hydrogels for Prolonged Hydrophobic Drug Delivery

Introduction

Hydrogels have many favorable properties in the context of drug
delivery such as their chemical, physical, and mechanical similarities to
native extracellular matrix as well as their low interfacial energy, leading to
generally good biocompatibility responses. 
However, in the context of delivering hydrophobic drugs, hydrogels
present significant challenges in terms of both the drug loading and drug release.
If drug loading is attempted following gel formation, the amount of drug
incorporated is minimal and rarely reaches therapeutic levels, given the lack
of affinity between the highly hydrated hydrogel and the hydrophobic
therapeutic.  If a sufficient degree of
loading is somehow achieved, the absence of interactions between the
hydrophobic drug and the hydrophilic matrix typically results in a rapid release
profile in vivo via partitioning into
lipid bilayers.  Nanocomposite hydrogels in which hydrophobic
drugs are incorporated inside liposomes, lipid
droplets, or polymeric micro or nanoparticles have been reported to address
this challenge.  However, such materials are
often opaque owing to the scattering of the embedded phases (problematic in,
for example, ocular drug delivery); in addition, the presence of an additional
phase can significantly alter the swelling properties and/or the mechanical
properties of the hydrogel, potentially problematic in some biological
environments (e.g. at joints).  As an
alternative, hydrogels in which cyclodextrin (CD) is incorporated may
facilitate hydrophobic drug delivery while avoiding these issues. The
hydrophobic cavity of CDs can facilitate hydrophobic drug binding, while the
hydrophilic exterior and the small dimensions (2-3nm) of cyclodextrins maintain
a highly swollen and transparent hydrogel structure. 

In this work, a dexamethasone-loaded dextran/β-cyclodextrin (βCD)-based hydrogel was designed (Dex-βCD).
Gelation of these materials occurs in situ upon injection through the
reaction of an aldehyde-functionalized dextran with a hydrazide-functionalized βCD molecule, creating a (slowly) hydrolytically
degradable, covalently-crosslinked hydrogel network
formed from liquid-like precursors.  The
dual role of βCD as a crosslinker and host for
the hydrophobic drug allows for the adjustment of hydrogel properties and
release behavior based on the number of crosslinks between hydrazide-βCD and the hydrogel. 
In this way, drug release kinetics and drug loading can easily be
manipulated by changing the concentration of βCD
incorporated into the hydrogel and/or the ratio of entrapped versus crosslinked
βCD molecules. 
We show that drug release can be facilitated through both hydrophobic
partitioning (with bound βCD) and simple
diffusion (of the free βCD-drug complex through
the bulk hydrogel).

Materials and Methods

βCD was functionalized with multiple hydrazide groups by reacting 5 g
of carboxymethylated βCD
(3.81 COOH/ βCD) with 12.6 g of ADH in 120 mL of
water. The pH of the solution was adjusted to 4.75 and the reaction was started
with the addition of 13.85 g of EDC.  The
pH was maintained at 4.75 until it stabilized, which took approximately 4
hours.  The solution was neutralized to a
pH of 7.0 and water was removed from the product under vacuum in a rotary
evaporator at 60°C. The product was precipitated with a large excess of acetone
and collected through vacuum filtration. 
The degree of substitution of the hydrazide-functionalized product was
determined through potentiometric titration (~3.1 hydrazide groups/βCD).  A
hydrazide-functionalized dextran polymer (Dex-Hzd, ~600
hydrazide groups per polymer chain) was synthesized in a similar fashion,
except the modified polymer was purified through dialysis against water and
isolated via lyophilization. Aldehyde-functionalized
dextran (Dex-Ald) was synthesized by adding 10 mL of
an 80 mg/mL aqueous solution of sodium periodate to 1.5 g of dextran [Mr- 500,000] in 150mL of water.  The reaction was allowed to proceed for two
hours and was stopped with the addition of 0.4 mL of ethylene glycol.   The modified polymer was dialyzed and
isolated through lyophilization. Drug loading was
achieved by adding a large excess of dexamethasone to 2 mL of a βCD-Hzd solution and shaking for 3 days at ambient
temperature.  Drug uptake by βCD-Hzd was quantified using gravimetric methods.  Dexamethasone loading in the control dextran
hydrogel was achieved by suspending the hydrazide-functionalized dextran in a
90 µg/mL solution of dexamethasone in water. The hydrogel was formed using a
double-barrel syringe in which each barrel contained the contents of the gel
precursor solutions, composed of either a hydrazide phase or an aldehyde
phase.  Equal amounts of both phases were
extruded through the mixing device and injected into silicone rubber moulds.
Six hydrogels were used for each release study and were placed in cell culture
inserts in a 12-well culture plate. The release study was carried out at 37°C over
the course of 21 days.  Quantitative
determinations of dexamethasone in PBS release solutions were performed using
high-performance liquid chromatographic (HPLC). 

Results

Figure 1 shows the release of dexamethasone
from a Dex-βCD gel compared to a Dex-only gel (2wt% Dex-Hzd, 2wt% Dex-Ald in both gels).  Drug release from a dextran-only hydrogel
followed conventional hydrogel kinetics, consisting of an initial (rapid) burst
followed by first order release that was largely completed after 1 day.  In comparison, when βCD
was added to the hydrogel, the burst release was nearly eliminated and only
5.4% of the total drug was released from the Dex-βCD
gel over the 20-day sampling period.  Of particular
interest, when the drug release was plotted as a function of the sample number
as opposed to time (Figure 1b), the dextran-only hydrogel shows a typical
first-order profile (characteristic of diffusion-based release) while the βCD-containing hydrogel exhibits constant release per
sample collected.  The amount of
dexamethasone released (~1mg/mL/sample) is significantly lower than the solubility of
dexamethasone in the PBS release medium (~80mg/mL) and is constant
regardless of the time between sampling points, suggesting that the slow
release from this gel is attributable to partitioning of diffusible drug
species between the hydrogel phase and the release medium.  The immobilization of βCD
via crosslinking prevents the diffusion of (soluble) βCD-dexamethasone
complex out of the hydrogel, leading to the extremely slow release
observed.  Extrapolating these results,
for use in the eye (clearance time ~2 hours for dexamethasone from vitreous
humour), this release rate would facilitate continuous release of dexamethasone
for 3-4 months via a purely partitioning mechanism, significantly longer
release than currently reported hydrogels for this purpose.

By doubling the concentration of the Dex-Hzd polymer in the gel precursor solution to 4wt%, the
hydrogel became significantly more crosslinked overall, with a ~50% increase in
G' compared to the 2wt% Dex-Hzd hydrogel.  For diffusion-based release, this would
normally result in significant slowing of the drug release kinetics, as higher
elasticity corresponds to higher crosslink density and thus lower average pore
sizes.  However, Figure 2 indicates that
drug release in hydrogels prepared with higher fractions of Dex-Hzd
is significantly faster than that from hydrogels prepared with lower Dex-Hzd concentrations; during the 20 day release study,
64% of the total contained drug was released from the 4wt% Dex-Hzd
hydrogel, an 11-fold higher dexamethasone release relative to the gels prepared
with 2wt% Dex-Hzd.  In addition, release no longer follows a
linear profile with sample number for 4wt% Dex-Hzd
hydrogels (Figure 2b), suggesting a significant diffusive contribution to the
release kinetics.  We attribute these differences
in release kinetics to the higher mobility of the complexed
drug when βCD is physically entrapped in a
hydrogel versus chemically crosslinked. 
In this case, the increased number of available hydrazide groups in hydrogel
recipes involving 4wt% Dex-Hzd can compete more
effectively with hydrazide groups on the functionalized βCD,
making βCD is less likely to crosslink to the
gel and thus more likely to become physically entrapped in the hydrogel network.
 As a result, more diffusible drug
species are present and release can occur via both partitioning and
diffusion-based mechanisms.   This result suggests that drug release from βCD-containing hydrogels can be precisely tuned, both
in magnitude and mechanism, by the number of βCD
groups added as well as the degree to which those βCD
groups are covalently attached to the hydrogel network.  Biocompatibility results for this hydrogel
will be presented to assess the potential of this hydrogel as a clinical drug
release vehicle.

Conclusions

In hydrogel systems where βCD was primarily immobilized within the polymer
network, an extremely slow, partition-based drug release rate was observed.  In hydrogel systems where βCD
was in part physically entrapped within the matrix, drug release was
significantly faster and occurred via both diffusion and partitioning.  Thus, βCD can
be used to increase or decrease the rate of drug release from hydrogels while
in both cases significantly increasing the total capacity of the gel to deliver
a hydrophobic drug payload.  Coupled with
the injectability of the hydrogel, such materials have significant potential
for long-term (i.e. several months) local hydrophobic drug delivery.

References

1. Hoare, T.R.,
Kohane, D.S. (2008).  Polymer 49:
1993-2007.
2. Bibby, D. C., N. M. Davies, et al. (2000). International Journal of Pharmaceutics
197(1-2): 1-11.

3. Peng, K., I. Tomatsu, et al. (2010). Soft
Matter 6(1): 85-87.

Acknowledgements

Financial support
was provided by the 20/20 NSERC Ophthalmic Materials Network.