Studies of Neural Injury and Disease
A wide variety of animal disease models are used in neuroscience research to study the causes and treatments of neurologic and psychiatric diseases. They include models of degenerative diseases (such as Alzheimer’s disease, Parkinson’s disease, and frontotemporal dementia); of traumatic injury of the head, spinal cord, or peripheral nervous system; of infectious diseases (such as immunodeficiency viruses, prion diseases, and viral, bacterial, or parasitic meningitis or encephalitis); of neuroimmunologic disorders (such as multiple sclerosis, myasthenia gravis, and polymyositis); of neurodevelopmental disorders (such as autism, Asperger’s syndrome, and Williams’ syndrome); of pain (from tissue injury and nerve injury); of neurologic problems that are secondary to primary medical conditions (such as diabetic neuropathy, nutritional disorders, and hepatic and renal encephalopathy); and of psychiatric disorders (such as schizophrenia and affective disorders).
Major considerations in the evaluation of research protocols and the management of animals experiencing those conditions are assessment of animal well-being, provision of appropriate nursing care and pain management, limitation of the duration and severity of the condition to be consistent with the experimental goal, and, in some cases, assessment and minimization of potential human health risks.
Assessment of animal well-being is discussed in detail in Chapter 2, and the same principles apply here. Personnel who are knowledgeable about the species-typical behavior of the animals under study and the clinical symptoms of the
disease should evaluate animal well-being at appropriate intervals. Clinical symptoms could involve such diverse markers as musculoskeletal abnormalities (tremor, reduced ambulation, and paralysis), decreased appetite (anorexia or aphagia), adipsia, signs of pain, fever, seizures, disorientation, and/or self-mutilation.
Before initiation of the research project, the research team, in consultation with the veterinary staff, should determine a course of clinical intervention or management based on the observed or expected clinical signs (see Table 6-1). The clinical plan should prevent the development of unintended pain and distress; but in instances where pain and/or distress is an intended outcome (as in pain research), the adverse consequences to the animal should be minimized as much as possible without jeopardizing the research goals. Potential interventions include the provision of easy access to water and perhaps to highly palatable food, promotion of urination and defecation, avoidance of decubital ulceration, maintenance of fluid balance, administration of appropriate analgesic or tranquilizing drugs, and, for some species, human contact to soothe and comfort the animal. Close clinical observation may also be necessary during periods of disease exacerbation. For example, in studies of seizure induction or treatment, continuous animal observation during the seizure is essential to prevent injury to the animal, although the seizure itself may not be painful.
The duration and severity of the condition should be managed within predetermined limits (humane endpoints) that reflect experimental goals. For example, studies of the mechanisms of disease development might require a shorter post-procedural duration than would studies of treatment interventions. Endpoints should be defined in terms of both the experimental goals (such as development of the syndrome or recovery of function according to some objective or subjective standard) and animal well-being (such as clinical deterioration that indicates that euthanasia is warranted). Endpoints applied to a specific model can be modified as more is learned about the model. For example, although death was used as an endpoint in some early studies of tumor metastasis in mice, later studies used hind limb paralysis to indicate that euthanasia should be performed, because paralysis was shown to be a valid indicator of death (Huang et al., 1993, 1995). Seizure studies should be designed to minimize the number, duration, and severity of seizures, without jeopardizing the scientific goals of the study.
Occupational Health and Safety
The production of some animal disease models requires the use of substances that pose health risks to humans. They include neurotoxins (such as MPTP), infectious agents (such as prions, bacterial, and viral pathogens), human cell lines, and noise exposure (free field). Appropriate risk assessment and the development of safe standard operating procedures is essential for the review and use of these models.
TABLE 6-1 Animal Welfare Considerations Associated with Disease Modelsa
Neuroscientists often induce lesions to learn about normal brain function (NIH, 1991). A classic example involves the study of the hippocampus and recognition memory, in which rats and monkeys with lesions limited to the hippocampus are impaired in tests of recognition memory (Zola and Squire, 2001).
Lesions of the nervous system can be produced by various means. Surgical or vacuum ablations, stereotaxic administration of neurotoxins, electric lesioning, and vascular occlusions require opening the cranial cavity and are considered major survival surgery that requires aseptic technique and, depending on the circumstances, the use of dedicated facilities (NRC, 1996). Noninvasive techniques can include radiation, blunt trauma, and intravenous administration of neurotoxins, although some of these methods may also be applied directly to the brain after surgical opening of the skull.
The use of lesions can establish an essential role of a structure, but because the processes of learning and memory have many steps, permanent lesions cannot be used to determine which step of learning and memory depends on the structure under study. To address the latter question, reversible lesions are used: the target structure can be temporarily deactivated during different stages of the assessment of learning and memory (for example, at the time of initial learning, during the delay interval, or at the time of retrieval).
Reversible lesions now allow assessment of cognitive function during different phases of learning and memory. Two types of reversible human amnesias have been studied in animal models using reversible lesions: transient global amnesia (Kritchevsky and Squire, 1989) and the amnesia associated with electroconvulsive therapy (Squire, 1986). Three approaches for making reversible lesions are cooling, chemical treatments, and transcranial magnetic stimulation (Lomber, 1999). The first two involve placing implants in the brain (cooling probes or cannulae). The use of implants to produce reversible lesions should be consistent with the guidelines of asepsis and sterility previously discussed (“Animal Care and Use Concerns Associated with Introduction of Probes into Neural Tissue” in Chapter 4).
In some studies, lesions are induced to produce animal models of naturally occurring diseases. For example, a lesion of the nigrostriatal pathway leads to motor deficits associated with Parkinson’s disease and was helpful in development of new dopaminergic agents for treatment and in demonstrating the effectiveness of neural transplantation (Tolwani et al., 1999). Such a lesion is usually produced by stereotaxic injection of 6-hydroxydopamine or intravenous administration of MPTP. Parkinsonian lesions have been induced in various animals, including mice, rats, cats, dogs, sheep, and nonhuman primates (Zigmond and Stricker, 1989). The clinical symptoms of the lesions depend on the species used but can include hypokinesis, circling behavior, aphagia, adipsia, bradykinesia,
rigidity, balance impairment, and resting tremor (Schneider and Kovelowski, 1990; Stern and Langston, 1985; Taylor et al., 1997).
A common strategy in neuroscience research is to induce lesions that produce specific structural or functional deficits. The deficits then can be studied to develop treatments that lead to recovery of function, as in spinal cord or peripheral nerve injury research. Common, clinically relevant lesion models include spinal cord contusion (Allen, 1911; Dohrmann et al., 1978; Wrathall et al., 1985), spinal cord transection (Khan et al., 1999), and neurectomy (Bouyer et al., 2001). Neurectomy and deafferentation surgery can result in autotomy (self-mutilation of the denervated limb) (Blumenkopf and Lipman, 1991); however, the use of local anesthetic at the time of nerve section will reduce autotomy (Magnusson and Vaccarino, 1996), and may be appropriate if the study of autotomy or dysesthesia is not the goal of the experiment.
Monitoring and Care Plan
Each nervous system lesion model has the potential for unique animal care issues that need to be fully investigated so that a monitoring and care plan can be drawn up before experimentation.
Preoperative health assessment and postoperative care are particularly important in lesion studies. In many cases, it is also useful to document the performance of an animal in a behavioral task to create a baseline assessment. For example, changes in exploratory behavior can be used as measures of chronic pain after spinal cord contusion in rats (Mills et al., 2001), although such measures alone are not sufficient to define an experimental manipulation as nocifensive.
The perioperative care and monitoring of an animal with a lesion are similar to standard surgical care and monitoring. As suggested by the Guide, surgical monitoring may include monitoring core temperature, cardiovascular and respiratory function, and postoperative pain or discomfort (p. 63). A heightened level of monitoring is beneficial in these models to determine whether the lesion caused unusual or unexpected pain and/or distress postoperatively. The difficulty of predicting whether a lesion will compromise an animal’s health or well-being (NIH, 1991) reinforces the need for frequent and comprehensive monitoring.
Impaired Physiologic Functioning
A lesion may cause changes in physiologic function. For example, animals with thoracic spinal cord transection often need to have their bladders manually expressed, and can have inhibited gastrointestinal motility. Close monitoring of bowel and urinary bladder status of these animals is required; additional care, such as the administration of laxatives (Khan et al., 1999), may be necessary in some cases. Information about the physiologic difficulties associated with well-characterized models is readily available in the literature and should be incorpo-
rated into the long-term care plan described in the animal-use protocol. In new models that are being developed, a specific monitoring plan should be developed to assess possible changes in physiologic status.
Lesions that impair animals’ mobility or alter their motivation can reduce their ability to care for themselves. They may be anorexic or adipsic and may not groom adequately. Adequate monitoring (scheduled weighing and assessment of appearance) and detailed record keeping will alert researchers and animal care staff to administer extra care (NIH, 1991), possibly even euthanasia in accordance with predetermined endpoint criteria. To ensure the consistency of monitoring between observers and experiments, it may be beneficial to use a quantitative scoring system for monitoring appearance and grooming (Ullman-Cullere and Foltz, 1999).
The additional animal care provided may include administration of nutritional support (Ungerstedt, 1968), fluid replacement, or provision of soft, rather than hard, food (NIH, 1991). In some instances, such as with stereotaxic injections of 6-hydroxydopamine, the lesion may be administered unilaterally rather than bilaterally. That approach is sometimes adequate for research purposes and reduces the impairment (Tolwani et al., 1999).
ANIMAL MODELS INVOLVING PAIN
The approaches used to recognize and treat unintended pain originating in neuroscience studies and in studies of pain are basically the same and are discussed elsewhere (“Pain and Distress” in Chapter 2). This section focuses on models of inflammation and nerve injury that produce pain so that its underlying mechanisms can be studied. Experiments with animals have mostly used stimuli that produce acute pain of short duration and moderate intensity; these models have become standards in the screening of putative analgesics. More recently, investigators have begun to develop nonhuman animal models that mimic persistent pain conditions seen in humans. Tissue injury and inflammation are commonly associated with clinical conditions that lead to persistent pain. Accordingly, new animal models to study these conditions differ in important ways from earlier, acute pain models.
Animal models of pain and hyperalgesia (excessive sensitivity to pain) have been developed to study the functional changes produced by the injection of inflammatory agents into the rat or mouse hindpaw (for review see Ren and Dubner, 1993; Dubner and Ren, 1999). The animals withdraw their limbs reflexively but also exhibit more complex organized behaviors, such as paw-licking and guarding (Hargreaves et al., 1988). A paw-withdrawal latency measure and withdrawal duration can be used to infer pain and hyperalgesia in response to
thermal or mechanical stimuli (Ren and Dubner, 1993). Methods of measuring nocifensive behavior have also been applied to the orofacial region (Imamura et al., 1997; Ren and Dubner, 1993). In the above studies, most of the nocifensive behaviors provide an animal with control of the intensity or duration of the stimulus in that the behaviors result in removal of the aversive stimulus.
Animals in persistent-pain models do not have control of stimulus intensity or duration. For example, the writhing response is produced in rodents by injecting pain-producing chemical substances intraperitoneally. The acute peritonitis resulting from the injection produces a response characterized by internal rotation of one foot, arching of the back, rolling on one side, and accompanying abdominal contractions. The writhing response is considered a model of visceral pain (Vyklicky, 1979). Not only does the animal lack stimulus control with this method, but the experimenter cannot control the duration of the stimulus. In another test, formalin is injected beneath the footpad of a rat or cat (Abbott et al., 1995; Dubuisson and Dennis, 1977). The chemical produces complex response patterns that last for about an hour. Many response measures are used for assessing pain after formalin injection. They include single measures such as flinching, shaking, and jerking—or complex scores that are derived from several nocifensive behaviors, such as licking or guarding (Clavelou et al., 1995). However, the animals do not have complete control over the aversiveness of the persistent stimulus. Vocalization is another common, unlearned reaction to painful stimuli (Kayser and Guilbaud, 1987), and the stimulus intensity necessary to elicit a vocal response from the animal can be determined. The stimulus can be applied to any part of the body; again, the animals cannot control the intensity or duration of the stimulus.
Nerve-injury models that mimic neuropathic pain in humans have been developed recently (Dubner and Ren, 1999). Partial nerve injury in the rat results in signs of hyperalgesia and spontaneous pain. In one model, loose ligatures are placed around the sciatic nerve; demyelination of the large fibers and destruction of some unmyelinated axons result (Bennett and Xie, 1988). In another model, ligation and severing of the dorsal one-third to one-half of the sciatic nerve produce similar behavioral changes (Seltzer et al., 1990). Kim and Chung (1992) have developed a third model, in which the L5 and L6 spinal nerves are tightly ligated on one of the rat’s sides. All three models mimic clinical conditions of painful neuropathy and yield evidence of persistent spontaneous pain, allodynia (pain resulting from a nonnoxious stimuli), and hyperalgesia. These nerve-injury models of neuropathic pain have been adapted for use in mice (Malmberg et al., 1997; Ramer et al., 1997), in which they can be used to study pain mechanisms in transgenic models.
Ethical Considerations Associated with Pain Research
Anesthetic and pain-relieving methods and drugs generally act on the system under study—the nervous system—and neuroscientists and IACUCs must make
difficult choices in selecting the means by which pain and distress are controlled and how much pain and/or distress is acceptable.
Several ethical issues have been proposed for IACUC consideration when reviewing protocols involving pain and/or distress in animals (Tannenbaum, 1999):
Of the animal-use protocols reviewed by the IACUC, those which include pain and/or distress should be subject to a full committee review rather than review by a designated member or with an expedited review process. If necessary, the committee should involve an outside consultant to understand better the ramifications of the study.
The protocol should provide a compelling justification for the work, a description of the qualifications of the personnel who will perform the work and provide care for the animals, and a rationale for withholding analgesics or other pain-relieving or distress-relieving methods.
The protocol should contain a complete and accurate description of the severity of pain and/or distress that will potentially be experienced by the animals.
When it is not in conflict with the scientific goals of a well-designed study, pain relief should be provided by anesthetizing the animals; giving them analgesics; allowing them to escape or avoid the pain; or control the experimental trials.
A humane endpoint for the use of an animal should be determined as an element of the protocol, before work begins.
Ethically, models of persistent pain present a particular challenge because they produce pain that most guidelines for the use of animals in research state should be avoided. Scientists should demonstrate a continuing responsibility for the proper treatment of the animals involved in these experiments. Because some models produce persistent pain that the animals cannot control, it is important that investigators assess the level of pain in these animals and provide analgesic agents when they do not interfere with the purpose of the experiment. A reduction in body weight or a significant deviation from normal behavior—such as a change in normal activity patterns, social adjustment, feeding behavior, and sleep-wake patterns—suggests that an animal is in severe and possibly intolerable pain.
In animal models of inflammation and nerve injury, the IACUC should ensure that steps are taken to safeguard animal welfare. The steps may include use of fail-safe devices to avoid excessive exposure to painful stimuli (for example, monitoring stimulus intensity and duration), having in place well-established humane endpoints to deal appropriately with intractable conditions (such as self-mutilation), and postprocedure monitoring of animal well-being.