POINT OF VIEW
`
`How to Successfully Apply Animal Studies
`in Experimental Allergic Encephalomyelitis
`to Research on Multiple Sclerosis
`
`Lawrence Steinman, MD,1 and Scott S. Zamvil, MD, PhD2
`
`In their Point of View entitled “Experimental Allergic Encephalomyelitis: A Misleading Model of Multiple Sclerosis,” Sriram and
`Steiner1 wrote, “The most disappointing aspect of EAE [experimental allergic encephalomyelitis] as a potential model for MS is
`its almost total inability to point toward a meaningful therapy or therapeutic approach for MS.” Actually, EAE has led directly
`to the development of three therapies approved for use in multiple sclerosis (MS): glatiramer acetate, mitoxantrone, and natali-
`zumab. Several new approaches to MS are in clinical trials based on positive indications in preclinical work relying on EAE. New
`clues to the pathogenesis of MS and new potential surrogate markers for MS are shown from research involving EAE when it
`is critically coupled with actual findings in MS. There are pitfalls in overreliance on the EAE model, or on any animal model
`for any human disease. Nevertheless, over the past 73 years, the EAE model has proved itself remarkably useful for aiding
`research on MS.
`
`Any discussion of the pros and cons of the animal
`models of multiple sclerosis (MS), collectively known
`as experimental allergic encephalomyelitis (EAE), must
`address our present state of knowledge about MS. MS
`is a complicated disease, the cause and pathogenesis of
`which are incompletely understood. Though we have
`made progress in therapy of MS, treatment is imper-
`fect. Current therapies reduce the frequency of relapse,
`somewhere between 33 and 66%, and delay disease
`progression to a modest extent in relapsing-remitting
`and secondary progressive MS.2 There is no single test
`we can run to determine whether someone has “MS,”
`and there is no surrogate marker for us to measure to
`assess whether MS is worsening. Whether MS is actu-
`ally a single disease or whether it is primarily or ini-
`tially an “immune disease,” “an infectious disease,” “an
`inflammatory disease,” or a “degenerative disease,” or a
`combination of all of these types are all questions with
`answers that are currently unknown. A few genetic fac-
`tors have been associated with MS, most prominently
`genes
`of
`the major histocompatibility
`complex
`(MHC).3 A genetic basis for MS is clearly only part of
`the story because concordance in identical twins is less
`than even 50%. Many environmental factors have been
`associated with MS, although none can be considered
`definitively linked. Therefore, set in this context in
`
`From the 1Department of Neurology and Neurological Sciences,
`Stanford University, Stanford, CA; and 2Department of Neurology,
`University of California San Francisco, San Francisco, CA.
`Received May 2, 2006. Accepted for publication May 9, 2006.
`Published
`online
`June
`26,
`2006
`in Wiley
`InterScience
`(www.interscience.wiley.com). DOI: 10.1002/ana.20913
`
`Ann Neurol 2006;60:12–21
`
`which nearly all of the major questions about human
`MS remain unanswered, this critique addresses how
`our understanding of MS has been aided by studies on
`a collection of animal models known as EAE, first de-
`scribed almost three quarters of a century ago. Given
`all these uncertainties about MS, it is remarkable that
`studies on EAE have culminated thus far in three MS
`therapies and have led to a better understanding of the
`biology of MS. Clever applications of the EAE model
`will be a valuable tool for understanding the pathology
`of MS, for making better biomarkers for its diagnosis
`and prognosis, and for creating ever improved and safe
`therapies for this disease. To study a disease such as
`MS, without support from available animal models, is
`to unnecessarily create obstacles in a task that is com-
`plicated enough.
`
`A Brief History of Experimental Autoimmune
`Encephalomyelitis
`In the 1930s, workers at Rockefeller University discov-
`ered an animal model, now known as EAE.4 The first
`experiments were aimed at understanding episodes of
`paralysis that sometimes accompany vaccination. Three
`years ago, on the 70th anniversary of the first publica-
`tion on EAE, we wrote in the Journal of Experimental
`Medicine of Rockefeller University:
`
`Address correspondence to Dr Steinman, Beckman Center for Mo-
`lecular Medicine, Stanford University, Stanford, CA 94305.
`E-mail: steinman@stanford.edu
`
`12 © 2006 American Neurological Association
`Published by Wiley-Liss, Inc., through Wiley Subscription Services
`
`Apotex v. Novartis
`IPR2017-00854
`NOVARTIS 2051
`
`

`

`One of the most enduring models of human disease
`now celebrates the seventieth anniversary of its publi-
`cation in The
`Journal
`of Experimental Medicine.
`Thomas Rivers, working at the Hospital of the Rock-
`efeller Institute for Medical Research, along with his
`colleagues D.H. Sprunt and G.P. Berry, submitted the
`article entitled, “Observations on Attempts to Produce
`Disseminated Encephalomyelitis in Monkeys,” on Feb.
`21, 1933 (4). Rivers established this model to try to
`understand what caused neurological reactions to cer-
`tain viral infections like smallpox and in some circum-
`stances to vaccinations like rabies: the very first sen-
`tence
`of
`this
`landmark
`paper
`reads,
`“During
`convalescence from certain diseases notably smallpox,
`vaccinia and measles, and during or following vaccina-
`tion against
`rabies, an occasional patient develops
`symptoms and signs referable to the central nervous
`system.”5
`Thus, the EAE model was initially constructed to un-
`derstand acute disseminated encephalomyelitis, not
`MS. Acute paralysis was observed in the first reported
`models with inflammatory changes in the central ner-
`vous system. Later versions of more chronic EAE have
`been developed with pathology including demyelina-
`tion and axonal damage and clinical events such as re-
`lapsing and remitting episodes of paralysis,6 all of
`which are features common to MS. We must remem-
`ber then that EAE is a collection of various models
`reflecting features of acute disseminated encephalomy-
`elitis, as well as MS. The relation between acute dis-
`seminated encephalomyelitis
`and MS remains
`an
`enigma itself.
`Many refinements and variations have been devel-
`oped in the past 75 years. Even the name EAE has
`evolved from experimental allergic encephalomyelitis to
`experimental autoimmune encephalomyelitis. Research-
`ers have developed numerous variations of EAE includ-
`for optic neuritis,7 relapsing-remitting
`ing models
`MS,8 –10 and progressive MS.9 Some of these models
`reflect certain aspects of the pathology seen in MS in-
`cluding axonal degeneration together with demyelina-
`tion.10 Researchers have constructed EAE models with
`essential genes,
`such as human leukocyte antigen
`(HLA) DR2 associated with susceptibility to MS, in-
`stalled into mice as transgenes.11,12 Others have de-
`vised forms of EAE in nonhuman primates such as the
`marmoset that reflect essential aspects of the pathology
`of MS with high fidelity.13 Numerous mouse versions
`of EAE exist where important components of the im-
`mune system have been “knocked out” by homologous
`recombination.14 Thus, there is no single model of
`EAE that we refer to in this critique, rather, it is the
`ensemble of EAE models, which have been reported in
`more than 5,000 publications since 1933.
`
`Experimental Autoimmune Encephalomyelitis for the
`Development of Approved Therapies of Multiple
`Sclerosis: Three Case Studies
`Sriram and Steiner1 wrote, “The most disappointing
`aspect of EAE as a potential model for MS is its almost
`total inability to point toward a meaningful therapy or
`therapeutic approach for MS.” We take a position
`nearly diametrically opposite to that perspective: In-
`deed, six medications have received approval from the
`US Food and Drug Administration for treatment of
`MS, and three of them, glatiramer acetate, mitox-
`antrone and natalizumab, were developed after showing
`promise in EAE. Moreover, glatiramer acetate and na-
`talizumab were invented after a set of
`logical and
`forward-looking experiments in the EAE model, which
`elucidated key targets in the pathogenesis of MS. Here,
`we review how experiments in EAE led to the devel-
`opment of three approved drugs for MS and how the
`model has been useful in helping us to understand the
`disease. Approved therapies that have been developed
`with the EAE model and new targets of interest devel-
`oped using the EAE model are shown in the Figure.
`Michael Sela and his colleagues Ruth Arnon and
`Dvora Teitelbaum15–17 first conceived glatiramer ace-
`tate in the early 1970s. They made a series of random
`copolymers based on the molar ratios of four amino
`acids, glutamate, alanine, tyrosine, and lysine, that are
`present in myelin basic protein. Sela and McDevitt18
`had shown 5 years earlier that the antibody response to
`ordered copolymers of tyrosine and glutamate on a
`backbone of alanine and lysine was under strict genetic
`control linked to the MHC. McDevitt and Sela’s work
`opened the field of the genetic control of the immune
`response. Their discovery that such control was linked
`to the MHC had widespread implications for immu-
`nology. More than just a coincidence, genes within the
`MHC are the most critical for imparting genetic sus-
`ceptibility to MS. Moreover, the MHC HLA class I
`and class II gene products, HLA-A, -B, -DR, and -DQ,
`are the likely targets for glatiramer. Interactions of
`glatiramer with the MHC turned out to be critical in
`understanding its mechanism of action (see Fig).
`In 1971, Sela and colleagues15 showed that the ran-
`dom copolymer composed of glutamate, tyrosine, ala-
`nine, and lysine, termed Copolymer 1, was able to sup-
`press the induction of acute EAE. They then showed
`that Copolymer 1 blocked relapsing EAE in the guinea
`pig and EAE in the nonhuman primate.19 –21 Initial
`clinical testing of glatiramer was undertaken in Jerusa-
`lem under the direction of Abramsky22 in patients with
`MS and acute disseminated encephalomyelitis. Clinical
`testing of glatiramer by Bornstein and colleagues23
`showed that glatiramer was effective in reducing re-
`lapses in relapsing-remitting MS. A pivotal trial leading
`to FDA approval, under the leadership of Johnson,24
`showed that relapses were reduced by 29% in patients
`
`Steinman and Zamvil: Applying EAE Research to MS
`
`13
`
`

`

`Fig. General scheme for pathogenesis of MS. T and B cell homing to the central nervous system is followed by inflammation medi-
`ated by antibodies, complement and the toxic effect of cytokines. Medications approved for multiple sclerosis (MS) that arose from
`studies on experimental allergic encephalomyelitis (EAE) are shown in black. Promising therapies for MS elucidated from a creative
`interplay of work in MS and in EAE are shown in red. ECM ⫽ extracellular matrix; IFN ⫽ interferon; IL ⫽ interleukin;
`MHC ⫽ major histocompatibility complex; TNF ⫽ tumor necrosis factor; VCAM ⫽ vascular cell adhesion molecule.
`
`trade-
`receiving daily injections. Glatiramer acetate,
`marked Copaxone from the original name Copolymer
`1, was approved in 1996 for treatment of relapsing-
`remitting MS. Glatiramer acetate was thus first derived
`from a preclinical conception and invention in the
`acute EAE model, and was then taken through proof
`of concept in various acute and relapsing models of
`EAE. Success in the EAE model was followed by dem-
`onstration of clinical efficacy in relapsing-remitting
`MS. Glatiramer acetate currently is one of the most
`popular medications
`for
`treatment of
`relapsing-
`remitting MS, and more than 100,000 individuals with
`MS worldwide have received glatiramer acetate treat-
`ment.25 It took a quarter of a century for the develop-
`ment of glatiramer from the publications of the first
`results in EAE to its approval for relapsing-remitting
`MS!
`There are multiple mechanisms of action associated
`with glatiramer, and many of these mechanisms were
`first unveiled in the EAE model. Antigen-specific mod-
`ulation of the immune response to myelin basic pro-
`tein has been described.21,26 Modulation of the im-
`mune response with glatiramer leads to deviation of
`cytokine production in response to myelin basic pro-
`tein from so-called Th1 cytokines such as ␥ interferon
`
`14 Annals of Neurology Vol 60 No 1
`
`July 2006
`
`to Th2 cytokine production.26 Another mechanism of
`action centers on the random chemical structures in-
`herent in this random copolymer that allow it to bind
`to molecular targets with a wide combinatorial array of
`peptides based on four amino acids. Glatiramer binds
`to MHC molecules derived from most genetic back-
`grounds.27–29 With its capacity to bind to a broad ar-
`ray of MHC molecules, glatiramer could compete with
`many proteins for these critical molecules responsible
`for presentation of antigen to T lymphocytes.
`Given its widespread binding to diverse HLA mole-
`cules, it is not surprising that glatiramer may have non-
`specific effects on the immune system. Glatiramer has
`been shown not only to block EAE, but is active in
`preventing models of inflammatory bowel disease and
`even amyotrophic lateral sclerosis.30 –32 This implies
`that the effect on MS may not even be specific for this
`disease, but that it may have a more general effect on
`immune modulation. Even though there are multiple
`potential mechanisms of action for glatiramer, and
`even though this drug may have potential uses in other
`diseases, the undisputed history of the development of
`glatiramer shows that it emanated from experiments on
`a treatment for EAE. One might reasonably conclude
`
`

`

`that without EAE, there would not have been a glati-
`ramer for treatment of MS.
`Mitoxantrone was developed for treatment of MS,
`after promising results published in the mid-1980s in
`reversing paralysis in the EAE model. The first article
`on mitoxantrone in EAE described mitoxantrone as a
`novel “anthracenedione that has shown antineoplastic
`activity against a variety of experimental tumors.”33 At
`the time, there was great interest in this class of drugs,
`because numerous cytotoxic agents, including azathio-
`prine, had shown promise in MS.34 Thus, mitox-
`antrone was tried in the EAE model in the rat. Ongo-
`ing paralysis was reversed, and the number and extent
`of perivascular lesions in brain was reduced after treat-
`ment with mitoxantrone. Given its promise in the EAE
`model, clinical development of mitoxantrone was taken
`forward in the clinic, leading to its approval for use in
`MS. FDA approval in 2000 for its use in secondary
`progressive MS and progressive or worsening relapsing-
`remitting MS was granted for reducing frequency of
`relapses and slowing clinical progression of disease.
`Success in the EAE model clearly spurred translation of
`this approach to the clinic in MS.35,36
`Natalizumab is yet another example of a drug that
`was developed directly from work in the EAE model.
`In the early 1990s, immunologists had developed the
`working hypothesis that there were specific “molecular
`addresses” for lymphocyte homing to various organs.
`Some referred to this as the “Zip Code Hypothesis” for
`lymphocyte homing. In collaboration with Yednock
`and colleagues37 at a small biotechnology company,
`Athena Neurosciences,
`the Steinman laboratory at
`Stanford determined the precise molecule involved in
`lymphocyte adhesion to inflamed brains taken from
`rats with EAE.38 Using an assay where frozen sections
`were cut on brains with EAE, the researchers bound
`human and rodent lymphocytes to the inflamed EAE
`sections. Monoclonal antibodies were then applied to
`these sections to see whether this lymphocyte binding
`could be blocked, and by inference, what molecules
`were involved in the association of lymphocytes to in-
`flamed brain tissue. More than 20 monoclonal anti-
`bodies to most of the adhesion molecules known at
`that time were tried on these sections of EAE brain,
`and only monoclonal antibodies binding ␣4 or ␤1 in-
`tegrin molecules inhibited adhesion of lymphocytes to
`the inflamed blood vessels in the brain.37
`We then proceeded to test whether a monoclonal
`antibody to ␣4␤1 integrin could block paralysis in-
`duced by T-cell clones that recognized myelin basic
`protein. These clones caused clinical paralysis and brain
`inflammation in a classic acute EAE model. The anti-
`body to ␣4␤1 integrin inhibited the development of
`paralysis when given at a dose of 4 to 6.4mg/kg in the
`Lewis rat. An article published in 1992 in Nature
`stated:
`
`Previous work on alpha-4 beta-1-dependent cell adhe-
`sion has mainly involved studies with endothelium that
`has been grown and stimulated in culture. The in vitro
`section assay described here extends those observations
`by showing that alpha-4 beta-1 integrin is crucial for
`the adhesion of leukocytes to vessels that have been
`activated in vivo. Furthermore, in vivo administration
`of anti-alpha 4 integrin prevented paralysis associated
`with the pathogenic inflammation of EAE. Therapy
`based on inhibiting alpha-4 beta-1 integrin, or the li-
`gand for this receptor on brain endothelium may prove
`effective in treating inflammatory disease in CNS.37
`From these experiments in EAE, it was recognized that
`blockade of ␣4␤1 integrin might be useful for MS. In
`1995, pathologists demonstrated the vascular cell adhe-
`sion molecule-1, the binding partner for ␣4␤1 inte-
`grin, was expressed in MS lesions. Over the next 10
`years, clinical development of a humanized monoclonal
`antibody to ␣4␤1 integrin showed that it had remark-
`able efficacy in blocking relapses of MS and even de-
`laying disease progression.39,40 Phase 3 studies showed
`that over a 2-year period, injection of 300mg of mono-
`clonal ␣4␤1 integrin reduced the relapse rate by two
`thirds. The dose was directly in the range shown to be
`effective in the experiments
`in EAE reported in
`1992.37 Indeed, the development of natalizumab was a
`tangible result of research in the EAE model.38
`Natalizumab has had a bipolar existence: It was ap-
`proved in November 2004 for use after 1 year of data
`were available in the 2-year phase 3 clinical
`trial.
`Within 3 months, three cases of Progressive Multifocal
`Leukoencephalopathy (PML) were observed, with 2
`deaths. The drug was voluntarily withdrawn. Unfortu-
`nately, PML does not occur in the animal species used
`in the EAE model, and this usually fatal complication
`was neither observed nor could it have been even tested
`in any EAE model or any available animal model for
`that matter.
`Investigators
`searched to see whether
`blockade of ␣4␤1 was associated with increased risk for
`infection to microbes such as cytomegalovirus and
`Borna virus. No increased risk for opportunistic infec-
`tions with either of these viruses was observed. The
`EAE model has limitations, and it is not particularly
`useful for examining the issue of opportunistic infec-
`tions, especially when the microbe in question has a
`species barrier.
`It should also be mentioned that ␤ interferons, the
`other major category of drugs approved for treatment
`of MS, have shown success when tested in EAE mod-
`els.41 However, the ␤ interferons were not developed
`initially because they showed promise in EAE, but
`rather because of the interest in development of anti-
`viral therapies for MS. Therefore, we do not count the
`development of ␤ interferons for therapy of MS as a
`triumph for applications of research on EAE. We do,
`however, consider the development of glatiramer, mi-
`
`Steinman and Zamvil: Applying EAE Research to MS
`
`15
`
`

`

`toxantrone, and natalizumab a direct consequence of
`research in the EAE model. So far, research on EAE
`has given three gifts of new therapies for treatment of
`MS.
`
`Problems and Promise of Using Experimental
`Autoimmune Encephalomyelitis for
`Development of Therapies for
`Multiple Sclerosis
`There is a long list of drugs that have shown promise
`in EAE models that are now being taken forward into
`the clinic. Other approaches including an orally avail-
`able sphingosine inhibitor,42 statins,43– 46 an orally
`available carboxamide,47,48 and a monoclonal antibody
`to IL-2 receptor49 –51 have shown great promise in
`phase 2 trials based first on success in the EAE model
`(see Fig).
`Of course, there are numerous examples of drugs
`that are effective in EAE, only to fail when tested in
`MS. One conclusion from these negative studies is that
`EAE is a poor predictor of success in MS. One must,
`however, examine the process of drug development to
`realize that preclinical research in EAE is merely just
`“exploratory,” whereas human clinical trails in MS un-
`dergo a rigorous “developmental” process,
`involving
`three phases of testing. Thus, once success is seen in
`EAE, one has to contend with issues such as formula-
`tion, dosing, and unforeseen toxicities when a drug is
`taken forward into clinical testing on humans. Given
`the high costs of clinical development of therapies in
`MS, one must make astute choices, usually on the first
`try, when translating preclinical results in EAE to clin-
`ical trials of MS. Each attempt at refining therapy in
`human clinical trials of MS is often prohibitively ex-
`pensive, so second chances are undertaken only in rare
`circumstances. In the case of natalizumab, the dose of
`monoclonal antibody in which a successful outcome
`was achieved in EAE was precisely translated to success
`in human clinical trials. However,
`for other drugs,
`problems with selecting a correct dose and dose fre-
`quency have confounded development. The case of al-
`tered peptide ligands, known as APL, which have
`shown great promise in EAE, exemplifies this problem.
`An APL from the region of myelin basic protein
`p83-99 showed promise in reducing relapse rates and
`reversing paralysis in preclinical studies of EAE.52–55 A
`version of this APL, known as NBI 5788, was designed
`and taken into clinical
`testing in patients with
`relapsing-remitting MS. This APL had an alteration in
`the main contact residues with human T-cell receptors
`recognizing this epitope of myelin basic protein. When
`NBI 5788 was given at a dosage of 50mg/week subcu-
`taneously, it was associated with exacerbations in three
`MS patients in an open-label trial.56 However, dosages
`of 5mg/week of this drug were associated with reduc-
`tion in gadolinium-enhancing lesions on magnetic res-
`
`16 Annals of Neurology Vol 60 No 1
`
`July 2006
`
`onance imaging, and there was no evidence of disease
`exacerbations.57 However, weekly dosing, at 5, 20, and
`50mg,
`led to allergic-type hypersensitivity reactions.
`The basis for these hypersensitivity reactions, seen in
`MS patients in phase 2 clinical trials, was then inves-
`tigated in the EAE model. In this new model of EAE,
`it was discovered that self-peptides of the myelin sheath
`could trigger fatal anaphylactic reactions in mice. The
`implications of this finding (ie, that even self-peptides
`were allergic) raised new and challenging questions. “A
`new version of horror autotoxicus,”58 first described a
`century ago by Paul Ehrlich, was discovered from un-
`derstanding a problem in a clinical trial of an MS drug.
`When dosing of NBI 5788 in phase 2b trials on
`relapsing-remitting MS was reduced to 5mg once a
`month instead of once a week to attempt to mitigate
`these hypersensitivity reactions, both the desirable ac-
`tivity in reducing magnetic resonance activity and the
`undesirable hypersensitivity reactions seen with APL
`disappeared.
`Continued development of APL in humans would
`require further financial
`investment if this particular
`approach is to be pursued with different dosing sched-
`ules. Negative studies in humans, on novel drugs such
`as the APL, which so far have failed to translate into an
`approved drug for MS, may not be a fault of the EAE
`model per se, but rather a reality of the huge expenses
`required for development of new drugs. When going
`from animal to human studies, revisions in dosing and
`formulation may be required that are too expensive
`and time consuming to pursue, given competing prior-
`ities. It is worth noting that the EAE model was used
`to help understand one of the clinical complications of
`this approach, allergic-like hypersensitivity reactions to
`self-constituents,
`seen with administration of
`the
`APL.58
`Another area where the EAE model has been called
`into question has been its inefficiency in predicting
`how blockade of various cytokines would work in MS.
`In rheumatoid arthritis and Crohn’s disease, we have
`seen the major triumph in therapy with the class of
`drugs known as tumor necrosis factor (TNF) blockers.
`This stunning advance in therapy led to the award of a
`Lasker Prize in Clinical Medicine to Profs Marc Feld-
`mann and Taini Maini to honor their achievement for
`implementing this mode of therapy in rheumatoid ar-
`thritis and Crohn’s disease. More than one million pa-
`tients with rheumatoid arthritis and Crohn’s disease
`have benefited from this approach.59 However, TNF
`blockade has been associated with worsening of
`MS,59 – 61 and a “black box” label has been placed on
`these drugs, warning against their use in MS.59 Studies
`in EAE have been equivocal, where some published ex-
`periments have shown the virtues of TNF blockade
`with anti-TNF monoclonal antibodies and soluble
`TNF-receptor constructs62– 64; in contrast, other pub-
`
`

`

`Table. The Long and Winding Road from Proof of Concept in Experimental Autoimmune Encephalomyelitis to US Food and
`Drug Administration Approval: Ranging from a Quarter of a Century to More Than a Decade
`
`Therapeutic
`
`Glatiramer Acetate
`Mitoxantrone
`
`Natalizumab
`
`Year of Publication
`of Proof of Concept in EAE
`
`Year of FDA
`Approval
`
`1971
`1987
`
`1992
`
`1996
`2000
`
`2004
`2005 withdrawn
`2006 reinstated
`
`Approved Indication
`
`Approved for RR-MS
`Approved for secondary progressive MS,
`worsening RR-MS
`Approved for RR-MS
`
`EAE ⫽ experimental autoimmune encephalomyelitis; FDA ⫽ US Food and Drug Administration; RR-MS ⫽ relapsing-remitting multiple
`sclerosis; MS ⫽ multiple sclerosis.
`
`lications have demonstrated the virtues of TNF itself,
`thus highlighting the pitfalls of blockade.65 One reason
`for this confusion is that a molecule such as TNF may
`have “janus-like effects,” giving benefit in some aspects
`of the disease, whereas imparting risk during other as-
`pects of disease. TNF is a mediator of inflammation in
`human autoimmune diseases and in animal models,59
`as well as a mediator of repair, including oligodendro-
`cyte growth, and a mediator of oligodendroglial
`death.66 – 68 Studies on EAE, which are often short
`term, must be carefully interpreted so that we pay at-
`tention to the “janus-like” propensity of cytokines:
`TNF blockade may be beneficial in inhibiting some
`inflammatory pathways, yet deleterious in blocking the
`maintenance and repair of cells in the nervous system.
`We have seen similar problems with glucocorticoids,
`which have virtues as antiinflammatory drugs, but have
`problems in inhibiting wound healing together with
`their numerous other unfortunate side effects.
`Yes, the same molecule may have different effects in
`influencing different biochemical pathways. And yes,
`EAE has its limitations. Misinterpretation and over-
`interpretation of certain results from EAE experiments
`may lead to calamitous consequences. Translation of
`animal to human studies is filled with large uncertain-
`ties. But, this should not mean that we discard the
`EAE model because of our uncertainty and ignorance
`of biology. We are only 21st century medical scientists,
`after all. We must remain humble about our under-
`standing of disease processes. Our predecessors in pre-
`vious centuries appear to us now so quaintly ignorant.
`Even though we have sequenced the human genome,
`we need to know what the products of these genes are
`doing in health and in disease. We need ever more
`clever animal models to help us understand observa-
`tions made in direct studies on human disease.
`One of the exciting directions in the development of
`therapy for MS is consideration of various combina-
`tions of medications. The EAE model has demon-
`strated potential synergies between drugs such as statins
`and glatiramer, which combine in the EAE model to
`show efficacy when used at doses that are suboptimal
`
`for these drugs when used alone.69,70 Other combina-
`tions of therapies might be tried in the EAE model, to
`search for synergies or unexpected adverse interactions.
`Drug development is not only costly, it is remark-
`ably slow. Rip Van Winkle could have a good sleep in
`the time intervals involved in drug development. The
`three drugs glatiramer, mitoxantrone, and natalizumab
`that first showed promise in EAE each took more than
`a decade before they were actually approved to treat
`MS. Drug development is a time consuming and ex-
`pensive process (Table). The fault does not necessarily
`rest with EAE, but in the harsh reality of how difficult
`it is to develop a new drug for MS.
`
`Experimental Autoimmune Encephalomyelitis
`Provides Insights to Understand Pathology, to
`Identify Surrogate Biomarkers and
`Therapeutic Targets
`MS researchers are now using a variety of powerful
`tools to understand the pathology of MS. These tech-
`nologies include gene, protein, and lipid microarrays,
`robotic sequencers for analysis of gene expression in
`MS tissue, and mass spectroscopy to detect minute
`amounts of proteins and lipids
`in MS brain le-
`sions.71–73 Once genes, proteins, and lipids of interest
`have been identified in MS, then clever applications of
`the EAE model allow one to explore their biological
`roles. Take the case of osteopontin, first discovered as a
`major transcript in lesions of MS by the use of a robot
`to sequence gene transcripts isolated from lesion mate-
`rial.74 No one had ever considered a role for osteopon-
`tin in MS brain, and from the name of the molecule,
`its main activity suggested that it involved bone and
`not brain. However, robotic sequencing of genes ex-
`pressed in MS lesions indicated that osteopontin tran-
`scripts were highly elevated at the site of brain inflam-
`mation.
`Immunohistochemistry showed widespread
`expression of the protein in MS lesions. A mouse strain
`existed where the gene for osteopontin was deleted. It
`was discovered that when EAE was induced in this
`strain, disease was milder, disease progression was
`blunted, mortality was decreased, and the intensity of
`
`Steinman and Zamvil: Applying EAE Research to MS
`
`17
`
`

`

`inflammatory mediators such as ␥ interferon was re-
`duced in myelin reactive T cells. Osteopontin, a
`cytokine-like molecule, was thus shown to have impor-
`tant immunomodulatory properties in EAE.74
`Researchers
`then explored whether osteopontin
`might play a role in the progression of MS. Three ar-
`ticles have now reported that osteopontin is elevated in
`blood before a relapse of MS.75–77 In MS, there is no
`surrogate marker to follow disease activity, which is
`akin to measurements of C-reactive protein in rheuma-
`toid arthritis or C-peptide in type 1 diabetes mellitus.
`Research on the EAE model in conjunction with stud-
`ies on human MS brain material may be leading to the
`discovery of the first surrogate marker in blood for MS.
`There are many other examples where a clever inter-
`play of studies on MS material combined with work on
`EAE has been illuminating. For example, B cells posi-
`tive for CD20 are found in MS lesions, and immuno-
`globulin V gene transcripts are among the most highly
`expressed genes in lesions.78,79 In the marmoset model
`of EAE, CD20 cells are common in cortical lesions,80
`whereas in the mouse model of EAE, B cells are quite
`common.81 There is great interest in several aspects of
`the role of B cells. Researchers have used studies on
`EAE combined with parallel observations in MS to de-
`termine the nature of the immune response in the cen-
`tral nervous system in MS, and then to see whether
`such antibody responses may play a pathogenic role in
`MS. Studies on the biological roles of antibodies to
`myelin proteins82– 84 and to lipids such as galactocere-
`broside and sphingomyelin73,85 have shown that not
`only are such antibodies present in the brains and spi-
`nal fluid of MS patients, but such antibodies are crit-
`ical in worsening the severity of disease in EAE models.
`Yet another example of understanding the pathogen-
`esis of a molecule with an unknown role in MS comes
`from studies on cytokines in MS brain. IL-6 and IL-17
`were first reported in MS lesions using transcriptional
`profiling with gene microarrays.80 Studies in the EAE
`model have shown that IL-6 and IL-17 are critical in
`the pathogenesis of EAE.86 IL-6 governs the produc-
`tion of a key set of regulatory T cells that modulate
`another subset of effector T cells producing IL-17.86
`Insights into MS will be understood only after dissec-
`tion of the biology of these newly discovered mole-
`cules. The use of transgenics, genetic “knock-outs and
`knock-ins,” and application of gene silencing tech-
`niques of interfering RNA, all techniques that allow us
`to analyze the biology of molecules in unprecedented
`ways, can o