Experimental Autoimmune Encephalomyelitis
Experimental autoimmune encephalomyelitis (EAE) is the oldest and most frequently used model system for studying MS in laboratory animals. Rather than a single model, EAE is a family of models in which central nervous system inflammation occurs after immunization against CNS-specific antigen. In its classic form, EAE is a CD4+ T cell–mediated autoimmune disease in which immunization with myelin proteins or peptides induces the migration of activated autoreactive T cells across the blood-brain barrier and into the CNS; alternatively, transfer of autoreactive T cells activated by such antigens can achieve the same result (Miller and Karpus, 2007). In the past 20 years, researchers have used transgenic techniques to develop new versions of the model and to expand its utility. For example, T-cell receptor (TCR) transgenic mice carry T cells engineered to respond to specific brain antigens. Similarly, genetic engineering also allows researchers to induce EAE in animals that lack a particular cytokine or immune molecule to test whether it is involved in the disease.
Initially, rats, guinea pigs, and nonhuman primates comprised the largest contingent of EAE animals. Today, however, most researchers use mice—and two strains in particular power the bulk of EAE studies. C57BL/6 mice (sometimes called B6), the most commonly used strain in the lab, are usually used to created transgenic animals, so if your plans include genetic engineering, stick with them. These mice, when immunized with peptides of myelin oligodendrocyte glycoprotein (MOG) along with an adjuvant (usually Complete Freund’s Adjuvant, pertussis toxin, or both), generally develop a chronic form of EAE, with inflammation, demyelination, and oligodendrocyte and neuronal death occurring about 2 weeks after disease induction.
A different mouse, the SJL/J strain, has played a crucial role in research because it develops a relapsing-remitting form of the disease—the most commonly seen version of MS. In these animals, immunization with myelin basic protein (MBP) or with proteolipid protein (or peptides of it)—also with an adjuvant—spurs the first signs of disease within 7 to 14 days. After the initial inflammatory attack subsides, the animals go into remission and later experience episodes that involve inflammation, demyelination, and axonal loss (Merrill, 2008). Other animals can experience different clinical and pathological manifestations of the disease; in Lewis rats, a commonly used lab strain, for example, EAE provokes no demyelination, inflammation localizes primarily to the spinal cord, and the animals experience acute paralytic disease and then recover completely (Baxter, 2007; Croxford et al., 2011).
Active induction—direct immunization with myelin protein or peptides thereof—is the easiest and fastest way to induce EAE. In the alternative approach, called adoptive transfer, animals are immunized against a particular antigen and then sacrificed. Their T cells are then harvested, reactivated with the immunizing antigen in culture, and injected into recipient animals. Which strategy you use depends on the question you’re investigating, says immunologist Burkhard Becher of the University of Zurich in Switzerland. If you’re studying the initial phase of the immune response—how T cells are activated—you’ll probably want to employ an active induction disease model, in which the injected antigen kicks off this event.
If you’re primarily interested in the effector phase—that is, how the activated T cells encounter and attack cells they deem pathogenic—an adoptive transfer model is appropriate. Because an experimenter can choose the type of introduced T cell, the latter approach offers the opportunity to tease apart the roles of different T cell lineages in the immune response, for example. A particular mouse strain might resist EAE induced by one approach but succumb to the other—and figuring out why can reveal important molecular mechanisms, Becher adds. “But one way or the other, if you wanted to look at therapeutic efficacy, then you’d have to look at both for sure,” he says.
Mice engineered to carry TCRs specific for certain brain proteins often develop spontaneous EAE, with varying frequency, depending on the particular TCR and the animal’s genetic background (Krishnamoorthy et al., 2007). Removing the need to inject a bolus of antigen or T cells into the animals, as in active induction and passive transfer, makes the process of disease onset much less artificial, says Vijay Kuchroo, an immunologist at Harvard Medical School in Boston. However, cautions Stefanie Kuerten, an anatomist at the University Hospitals of Cologne in Germany, a spontaneous model doesn’t allow researchers to predict when—or if—mice will develop disease, so if you want to study the proliferation, turnover, pathogenicity, and other characteristics of different T cell clones, then mice that spontaneously develop EAE might not be ideal subjects because you won’t know when to look. One model—a cross between two different transgenic C57B/6 lines, one engineered to carry MOG-specific T cells and the other to carry MOG-specific B cells—is becoming the spontaneous EAE model of choice (Krishnamoorthy et al., 2006; Bettelli et al., 2006).. Additionally, unlike most EAE models, in which T cell responses drive pathology, B cell activation in these animals contributes significantly to disease development.
Key pathological features
Most variants of the EAE model are thought to reflect a CD4+-mediated immune response. Pathological elements manifest in about 7 to 14 days when EAE is actively induced but appear more quickly upon adoptive transfer of activated myelin-specific T cells. EAE generally targets the spinal cord and sometimes the cerebellum, causing inflammation followed by demyelination and axonal damage. In monophasic or relapsing-remitting EAE, varying degrees of remyelination occur. The specific pathological features vary dramatically depending on the animal, genetic strain, induction method, and autoantigen used (Miller and Karpus, 2007; Pachner 2011; Baxter 2007).
Key clinical features
Classic EAE in mice causes so-called flaccid paralysis characterized by decreased muscle tone that progresses from the tail upward along the body. A six-point scale (0-5) reflects the severity of symptoms, with a score of 1 being tail paralysis, a score of 4 indicating quadriplegia, and a score of 5 defined as death. In some models, strains, and species—particularly those in which disease pathology reaches the cerebellum—animals may instead or additionally experience lack of coordination, or ataxia. Symptoms can be chronic, monophasic, or relapsing-remitting. This disease-progression pattern and other clinical features, such as the type of paralysis that occurs, depend on the animal, genetic strain, induction method, and autoantigen used (Merrill, 2008; Miller and Karpus, 2007; Baxter, 2007).
EAE provides a means for investigating mechanisms of autoimmune-related CNS damage and demyelination. It has broad similarities to MS—for example, the characteristics of demyelination and partial remyelination, the distribution of lesions around blood vessels, and the presence of immunoglobulin G in the cerebrospinal fluid (Baxter, 2007). As the most widely used model system for studying MS, with many variations (see above for types of variation), it offers an array of possibilities for matching an experimental question with a physiological profile.
The biology of all EAE models diverges significantly from that of MS. For example, clear autoimmunity underlies EAE, but MS lacks some features of classic autoimmunity, such as a known autoantigen that sparks the disease (Denic et al., 2011). And whereas CD4+ T cells play the primary role in prompting pathology in most versions of EAE, a wider spectrum of T cells as well as other immune cells such as B cells contribute significantly to MS (see "B Cells Step Into the Limelight").
Researchers therefore caution against drawing direct parallels between MS and EAE. This caveat applies especially in drug testing, they say. Work in EAE has led to the approval of four drug therapies for MS: natalizumab, mitoxantrone, glatiramer acetate, and fingolimod. However, many prospective treatments that relieve symptoms in EAE mice have not worked in human trials, and some, such as tumor necrosis factor-α, have exacerbated MS (Baxter, 2007). “ ’Oh, the mice do better, so this could be a therapy’—that is something I don’t even look at anymore,” Becher says.
Part of the mismatch, he says, stems from the nature of the experimental manipulations: They create a situation that is so extreme on the cellular level that the condition they trigger differs in crucial ways from MS. The spontaneous B- and T-cell model described above, for example, “has been discussed as the supermodel for spontaneous-onset EAE, but that’s a bit unfair because if all of your B cells and all of your T cells recognize brain antigen, then it isn’t surprising that these mutants develop EAE,” Becher says. “This is a very valuable addition to the toolbox available to date, but discoveries based on one single EAE model don’t necessarily translate to the human disease.”
Regardless of which EAE model you use, it’s not just the strain, the antigen, and the induction method that determines animals’ disease course but also the vivarium in which they are raised. It’s very tough to induce EAE in animals raised in a “dirty” environment; pathogen-free status is best. “But even if you have such a facility, there are differences depending on animal houses”—for example, in the time to disease onset, Kuerten says. Becher agrees that such environmental variations are a major issue but notes that if EAE experiments more often included the necessary controls, such as tests in both actively induced and adoptive transfer models, the noise generated by animal husbandry variations would matter less (see also Tips section below).
Disease processes that can be studied
EAE provides a powerful framework for investigating the inflammatory elements of MS. Researchers have used the model to study processes such as tolerance, immune surveillance, molecular mimicry, epitope spreading, environmental triggers and genetics of autoimmune disease, inflammation, lymphocyte entry into the CNS through the blood-brain barrier, functional differences between T-cell clones; relapse mechanisms, and immune-mediated demyelination and tissue injury (Pachner, 2011; Krishnamoorthy and Wekerle, 2009). A handful of EAE models, including the TCR transgenic relapse-remitting mouse model described above, also allow researchers to explore the role of B cells in EAE and, by extension, in MS.
Disease processes that cannot be studied
Because it’s impossible to predict the timing and location of lesions in EAE, the model is not ideal for studying the complete cycle of demyelination and partial remyelination, says Richard Ransohoff, an immunologist at the Cleveland Clinic in Ohio. What’s more, says Stephen Miller, an immunologist at Northwestern University Feinberg School of Medicine in Chicago, Illinois, assessing whether a compound directly guards against or even ameliorates neuronal damage in EAE or other models with an immune component is challenging. Many drugs that researchers are testing as neuroprotective agents might also exert immunosuppressive effects; that characteristic makes it difficult to determine whether a compound acts directly on neurons or simply quells the inflammation that damages them. To separate the two processes, a toxic agent such as cuprizone to elicit demyelination but not inflammation might be useful and more appropriate (see "Animal Arsenal").
The sheer number of permutations of EAE models makes their use a minefield for experimental error, and diving into the literature is often not enough to develop the needed level of expertise to use EAE most effectively, Becher says. Apprentice yourself to an experienced EAE researcher who can critique your scientific plan, Becher suggests: “When I started on EAE, I latched onto a couple of people, and it was very helpful.”
If you are using EAE to test experimental therapeutic compounds, says David Baker, an immunologist at Queen Mary, University of London, make sure to treat the animals with your compound of interest after immunizing the animals rather than before. This order more accurately reflects the experience of MS patients, who don’t receive drugs until they experience symptoms. Baker and his colleague Sandra Amor at VU University Medical Center in the Netherlands laid out a set of guidelines that proposes minimally acceptable standards for publishing studies that use EAE as a preclinical model (Baker and Amor, 2011).
Researchers also say that it’s important to confirm findings in different models—especially when testing investigational compounds. Experiments should be done in an actively induced model and a passive transfer model. As with all types of experiments, controls are crucial. If you’re testing a compound that blocks some immune factor, make sure that inducing EAE in a knockout mouse that lacks the factor produces the same result, suggests Cris Constantinescu, a neurologist at the University of Nottingham in the U.K.
Utility for probing relevant biology
Researchers widely agree that EAE has played a significant role in uncovering basic immunological features of multiple sclerosis—as well as immune response more generally. In the early days of EAE, studying disease states caused by injecting brain homogenate in different species helped pinpoint how particular tissue-specific molecules produce autoimmune reactions (Krishnamoorthy and Wekerle, 2009). A landmark 1981 study reported that injecting T cells that target MBP into rats causes EAE and thus propelled T cells into their starring role in MS research (Ben-Nun et al., 1981). Later work expanded on the inflammatory mechanism such as the role of CD4+ regulatory T cells (Olivares-Villagomez et al., 1998; McGreachy et al., 2005) and epitope spreading (McRae et al., 2005).
In 2005, researchers identified a novel class of helper T cells called Th17 cells that seem to contribute to the pathogenesis of EAE (and possibly MS), although the precise mechanism remains unclear (Korn et al., 2009; Becher and Segal, 2011). Other experiments have riffed on classic EAE models to study the involvement of specific cytokines such as GM-CSF (McQualter et al., 2001; Codarri et al., 2011) and have also panned out to probe nonclassic contributors to MS such as B cells (Pollinger et al., 2009).
EAE is also the go-to model for testing the effectiveness of potential MS therapeutics. However, its contribution to this endeavor is controversial: Most researchers support its utility to at least some extent, but a few dismiss it.
EAE was born out of one investigator’s effort more than 80 years ago to determine why individuals who received an early form of the first rabies vaccine sometimes experienced attacks of paralysis. In the course of his experiments, the researcher, Thomas Rivers, injected control animals with emulsified brain tissue; they developed brain-specific antibodies and a proportion of them became transiently paralyzed. Based on Rivers’s work, other researchers began to probe the possibility that the antibodies were creating an immune response that led to demyelination. The experimental system thus became a model of demyelinating disease and human autoimmune disease (Baxter, 2007).
Mice (C57BL/6, SJL, other)
Jackson Laboratory (http://www.jax.org/)
Charles River Laboratories (http://www.criver.com/)
Harlan Laboratories (http://www.harlan.com/)
Myelin proteins or peptides
Hooke Laboratories (http://hookelabs.com/)
Adjuvant (pertussis toxin, Complete Freund’s Adjuvant)
Difco Microbiology (http://www.vgdusa.com/DIFCO.htm)
Hooke Laboratories (http://hookelabs.com/)
Thanks to David Baker, Burkhard Becher, and Stephen Miller for reviewing this article.