Fast-gelling injectable blend of hyaluronan and methylcellulose for intrathecal, localized delivery to the injured spinal cord

Abstract

Strategies for spinal cord injury repair are limited, in part, by poor drug delivery techniques. A novel drug delivery system (DDS) is being developed in our laboratory that can provide localized release of growth factors from an injectable gel. The gel must be fast-gelling, non-cell adhesive, degradable, and biocompatible as an injectable intrathecal DDS. A gel that meets these design criteria is a blend of hyaluronan and methylcellulose (HAMC). Unlike other injectable gels, HAMC is already at the gelation point prior to injection. It is injectable due to its shear-thinning property, and its gel strength increases with temperature. In vivo rat studies show that HAMC is biocompatible within the intrathecal space for 1 month, and may provide therapeutic benefit, in terms of behavior, as measured by the Basso, Beattie and Bresnahan (BBB) locomotor scale, and inflammation. These data suggest that HAMC is a promising gel for localized delivery of therapeutic agents to the injured spinal cord.

Introduction

Of the several therapeutic strategies investigated for spinal cord injury repair [1], [2], [3], only systemic delivery of methylprednisolone is used clinically; however, results from its clinical trial have been openly criticized [4]. Other therapeutic strategies that have been investigated hold great promise [5], yet prolonged systemic delivery is notorious for side effects, and the more promising therapeutic proteins degrade when delivered systemically. Moreover, many therapeutic molecules are unable to cross the blood–spinal cord barrier. These difficulties suggest that local delivery strategies are required.

Two intrathecal techniques have been used to test localized delivery: (1) bolus injection, the effects of which are short-lived because the therapeutic is washed away by the cerebrospinal fluid (CSF) flow [6], [7] and (2) minipump delivery, which is invasive and can lead to complications of the catheter being blocked and/or infection [8]. A third technique has the therapeutic molecule dispersed in an injectable gel that localizes release to the site of injection. We have previously demonstrated that intrathecal injection of a collagen is safe [9] and can provide localized release of growth factors into the injured spinal cord [10]. However, the collagen gel previously used was not ideal because it was not sufficiently fast-gelling, and caused cellular build-up in the intrathecal space when growth factors were incorporated within [11]. To overcome these limitations, a new injectable gel was required. The design criteria included: (1) fast gelling to ensure delivery would be localized to the site of injection, and that the polymer would not spread with the CSF flow; (2) injectable through a 30G needle to allow for a minimally invasive surgery; (3) non-cell adhesive to minimize the possibility of scar formation in the intrathecal space; (4) degradable to obviate the need for removal after release; and (5) biocompatible to minimize foreign body reaction.

There are several injectable gels that can be described as either physical or chemical gels [12], [13]. While both are effective, chemical gels require in situ crosslinking which can involve cytotoxic crosslinkers, free radicals, and/or immobilization of the therapeutic during this reaction. Physical gels, while often not as stable or robust as chemical gels, were investigated for intrathecal injection because “weak” gels were thought to be suitable for this application and the gelation system was simpler, with the use of potentially cytotoxic crosslinking agents obviated. Two physical gels that are known to be generally non-cell adhesive are methylcellulose (MC) and hyaluronan (HA), due, in part, to their hydrophilicity [14].

MC has inverse thermal gelling properties. As the temperature increases, hydrogen bonds between the polymer and surrounding solvent break, and hydrophobic junctions form to produce a gel [15]. MC forms weak gels at 37 °C when in water. The gelation temperature decreases as the salt concentration increases [16] because water molecules surround the salts, thus reducing the solubility of MC in water [17]. Although regenerated cellulose is known to activate the complement system [18], MC has previously shown good biocompatibility when used as scaffolds in traumatic brain injury and peripheral nerve regeneration [19], [20]. Inverse thermal gelling polymers such as MC are not sufficiently fast gelling for the injectable drug delivery system (DDS).

HA has found widespread use because it is non-immunogenic and biocompatible [21]. HA is known to promote wound healing by reducing inflammation and minimizing tissue adhesion and scar formation [22]. HA has unique rheological properties [23] because its long polymeric chains form random coils and gel due to molecular entanglements [22]. Under shear force, the molecules align with the direction of stress and flow [24]. However, due to the high water solubility of HA, it quickly disperses when injected into fluid-filled cavities. Few HA derivatives maintain the injectable nature of HA unless mildly crosslinked prior to injection [25].

A fast-gelling, injectable material was created, for the first time, by blending HA and MC (HAMC), and specifically 2% HA with 7% MC. The objectives of this study were to test HAMC against the design criteria, and to better understand its potential for intrathecal delivery. The gelation mechanism, degradation profile, and cell adhesion of HAMC were studied in vitro and the injectability, biocompatibility and therapeutic efficacy were studied in vivo. To better understand the mechanisms of gelation and degradation, the 2% HA/7% MC blend (HAMC) was compared to: 7% MC, 9% MC, and a blend of acetic hydrazide-modified HAMC (acet-HAMC). The acetic hydrazide modification sheds light on the importance of free carboxylic acid groups on HA for HAMC gelation. The biocompatibility of HAMC within the intrathecal space was further examined in vivo in both uninjured and spinal cord injured rat animal models relative to controls of artificial CSF (aCSF) injections.