Anatomical correlates of recovery in single pellet reaching in spinal cord injured rats

Highlights

• Reaching recovery is not correlated with lesion size or extent of CST or RST injury.

• Dorsolateral lesions provide the greatest potential to observe training effects.

• Compensation maintains higher reaching success in rats with larger lesions.

• Reaching successes of zero percent contribute to high variability in recovery.

• High variability means small opportunity to observe treatment effects.

Abstract

Modeling spinal cord injury (SCI) in animals is challenging because an appropriate combination of lesion location, lesion severity and behavioral testing is essential to analyze recovery of motor function. For particular tests such as single pellet reaching, the contribution of individual descending tracts to recovery has been investigated using specific tract ablation or graded lesions. However, it has not been established whether single pellet reaching is sufficiently sensitive for assessing the efficacy of treatments for cervical SCI (e.g., one of the currently most successful treatment approaches: rehabilitative training). To address this issue, we trained adult rats in single pellet reaching before and after a cervical (C4) spinal lesion. Animals with lesions of increasing severity were grouped into categories based on damage to anatomical structures such as the corticospinal tract (CST) and the rubrospinal tract (RST), two descending motor tracts that have been implicated in fine motor control of the forelimb. We related lesion extent to spontaneous recovery and plasticity-promoting post injury training and found that reaching performance was not correlated with lesion size or the extent of CST or RST injury. Interestingly, the dorsolateral quadrant (DLQ) lesion category, in which the unilateral dorsal CST and most of the unilateral RST are lesioned, was the only category that showed a clear effect of plasticity-promoting treatment (i.e., training), indicating its usefulness as a lesion model for this testing paradigm. The DLQ lesion likely strikes a balance between tissue sparing and functional impairment and is, therefore, best suited to maximize the potential to observe treatment effects of plasticity-promoting treatments using single pellet reaching. Because of the specific lesion size that is necessary to observe treatment effects, the single pellet skilled reaching task can be considered a stringent behavioral test and therefore may be useful for predicting translational success of potential treatments. However, due to the variability in the success rate, the labor-intensive nature, and the limited usefulness to test functional outcome of a wide range of lesion severities, we are hesitant to continue to use single pellet reaching to assess the effectiveness of currently available treatments for cervical SCI.

Introduction

Modeling spinal cord injury (SCI) in animals to test potential treatments is challenging, partly due to the variables involved, including different lesion levels, severities and causes of injury (e.g., contusion, compression, section). A number of combinations of these variables have been applied, driven by different treatment questions ranging from neuroprotection to axonal regeneration. Much of the early research in the field of SCI focused on lesions at the thoracic level using locomotor recovery as functional readout (Barbeau and Rossignol, 1987, Behrmann et al., 1992, Little et al., 1988). Because of its clinical relevance and convenient neuroanatomical access (relatively easy to trace and examine because it descends in the dorsal funiculus), the corticospinal tract (CST) was the preferred descending system involved in many of these studies (Eidelberg and Yu, 1981, Liebscher et al., 2005, Metz et al., 1998, Muir and Whishaw, 1999). However, it became clear that (at least in rodents) it is not lesions to the CST, but to the reticulospinal tract, projecting in the ventrolateral funiculus, that determine the degree of locomotor deficits (Loy et al., 2002, Schucht et al., 2002, Steeves and Jordan, 1980). Another challenge in the interpretation of locomotor recovery is that locomotion is orchestrated by so called spinal central pattern generating networks. Thus, locomotor patterns can be initiated without direct descending input (Barbeau and Rossignol, 1987, Lovely et al., 1986), making the interpretation of treatment effects even more difficult.

A shift in the field of SCI research towards cervical lesions occurred as a result of a number of factors. These factors included a strong base of knowledge that hand/paw function involves CST function (Lawrence and Kuypers, 1968, Schwartzman, 1978), an increase in the frequency of incomplete, cervical SCIs in humans (National Spinal Cord Injury Statistical Centre, 2012), and the assumption that reaching movements are volitional, and directly orchestrated by the brain and not only by spinal networks. The majority of these studies involved rats, which can be trained to grasp for food pellets in tests such as a single pellet skilled reaching task (Whishaw and Pellis, 1990, Whishaw et al., 1993), tray reaching (Whishaw et al., 1986) or the Montoya staircase test (Montoya et al., 1991). Studies using these reaching tests showed that damage to both the lateral and dorsal funiculus, housing the rubrospinal tract (RST) and CST respectively, results in permanent functional deficits in reaching recovery (Anderson et al., 2007, Morris et al., 2011, Whishaw et al., 1998). Although spinal tracts involved in the control of reaching have been identified, an ideal combination of efficient functional readout and lesion type is still lacking. The current study attempts to address this issue by evaluating the reaching recovery of rats with various lesion severities to highlight the ideal lesion to be used to test plasticity-promoting treatments with single pellet skilled reaching.

To be useful for demonstrating the benefits of regeneration and/or plasticity-promoting treatments, an animal model of SCI must fulfill various requirements. First, considering the currently limited treatment options for SCI and the fact that effective treatments will likely consist of combinatory approaches, moderate treatment effects should be detectable. By assessing the effects of rehabilitative training we utilize one of the currently most effective treatments for promoting recovery following SCI, thereby providing a good estimation of how sensitive a current testing approach needs to be. Second, the injury has to induce significant functional deficits, but should not be so severe that a floor effect is observed. A floor effect describes the effect when data cannot be recorded lower than a designated value (e.g., success of 0% in the single pellet test). This can result in an accumulation of animals with success rates of 0 although they still might vary in their reaching abilities. It also has to be kept in mind that plasticity-inducing treatments require a certain amount of tissue sparing. Lastly, the variability in spontaneous functional recovery should be small because the margin between spontaneous recovery, treatment effects and pre-lesion motor function are often very small.

To define an optimal cervical lesion model for testing effects of regeneration and/or plasticity-promoting treatments (including training), we investigated functional deficits caused by unilateral lesions, ranging from a minimal dorsolateral lesion to a complete lateral hemisection. To also investigate the effect of lesion location (with respect to projection of descending tracts) we distributed the lesions into categories and studied spontaneous recovery as well as the effect of reaching training (as a potent plasticity-promoting treatment) to identify the most appropriate lesion type to use when testing treatments following cervical SCI.