Expert Opinion on Pharmacotherapy
`
`ISSN: 1465-6566 (Print) 1744-7666 (Online) Journal homepage: http://www.tandfonline.com/loi/ieop20
`
`Glatiramer acetate for the treatment of multiple
`sclerosis
`
`Jerry S Wolinsky
`
`To cite this article: Jerry S Wolinsky (2004) Glatiramer acetate for the treatment of multiple
`sclerosis, Expert Opinion on Pharmacotherapy, 5:4, 875-891, DOI: 10.1517/14656566.5.4.875
`
`To link to this article: http://dx.doi.org/10.1517/14656566.5.4.875
`
`Published online: 02 Mar 2005.
`
`Submit your article to this journal
`
`Article views: 99
`
`View related articles
`
`Citing articles: 1 View citing articles
`
`Full Terms & Conditions of access and use can be found at
`http://www.tandfonline.com/action/journalInformation?journalCode=ieop20
`
`Download by: [UCSF Library]
`
`Date: 15 September 2016, At: 08:33
`
`Page 1 of 18
`
`Coalition Exhibit 1060
`Coalition v. Biogen
`IPR2015-01993
`
`

`
`Drug Evaluation
`
`Glatiramer acetate for the
`treatment of multiple sclerosis
`
`Jerry S Wolinsky
`University of Texas Health Science Center at Houston, 6431 Fannin Street, Houston, TX 77030, USA
`
`Glatiramer acetate (Copaxone®, Teva Pharmaceuticals Ltd) is a collection of
`immunomodulatory, synthetic polypeptides indicated for the treatment of
`relapsing-remitting multiple sclerosis (RR MS). Preclinical and clinical studies
`provide an evolving understanding of the mechanisms by which glatiramer
`acetate exerts both immunological and potential neuroprotective effects
`that account for its clinical efficacy. The results of pivotal controlled clinical
`trials and long-term data, derived from organised extension studies, are eval-
`uated in detail and supportive data from open-label comparison, combina-
`tion treatment and therapeutic switch studies are considered in order to
`determine the place of glatiramer acetate among other approved therapies
`for RR MS. The efficacy of glatiramer acetate is stable or may increase over
`time and the drug has a favourable side effect profile. Glatiramer acetate is
`an appropriate first-line immunomodulatory therapy for RR MS.
`
`Keywords: Copaxone, glatiramer acetate, immunomodulatory therapy, magnetic resonance
`imaging, multiple sclerosis
`
`Expert Opin. Pharmacother. (2004) 5(4):875-891
`
`1. Introduction
`
`Multiple sclerosis (MS) is an inflammatory disease of the CNS that leads to myelin
`destruction and axonal loss. It is the most common, non-traumatic, disabling neuro-
`logical disorder in young adults [1]. Although the aetiology of MS remains unknown,
`its pathogenesis is believed to include autoimmune reactivity to myelin components,
`placing it among the organ-specific autoimmune diseases. In relapsing-remitting
`(RR) MS, the clinical course is punctuated by exacerbations or episodic neurological
`worsening. These attacks or relapses, are followed within weeks to a few months by
`remissions, often with full recovery from clinical symptoms. However, recovery from
`up to 40% of all relapses is incomplete, leaving measurable neurological deficits or
`disability [2]. Progressive forms of MS are characterised by a gradual downhill course
`over many months to years without remissions, so that the patient acquires increas-
`ing clinical deficits, either beginning at presentation (primary-progressive [PP] MS)
`or after a period of RR disease (secondary-progressive [SP] MS).
`Currently approved immunomodulator therapies for RR MS include glatiramer
`
`
`acetate (GA) and the recombinant IFNs, (IFN-β1a, Avonex®, Biogen, Inc.; IFN-β1a,
`Rebif®, Serono, Inc.; IFN-β
`1b, Betaseron®, Berlex Laboratories), all of which modify
`the course of this progressively disabling neurological disease. Immunomodulatory
`treatments reduce disease activity and the accumulation of disability in RR MS. The
`National Multiple Sclerosis Society recommends initiation of therapy with an immu-
`nomodulator as soon as possible following diagnosis of RR MS [201]. Results of mag-
`netic resonance spectroscopy (MRS) and pathology studies show that inflammatory
`activity can cause irreversible axonal damage in the early phases of the disease, rein-
`forcing the need for early and aggressive treatment [3,4]. Mitoxantrone (Novantrone™),
`an antineoplastic agent, is also approved for the treatment of relapsing MS, but is gen-
`erally reserved for secondary progressive and severe RR forms of the disease [5].
`GA (formerly known as copolymer 1 or Cop 1) is indicated for the reduction of
`the frequency of relapses in RR MS. The drug is approved in 42 countries
`
`2004 © Ashley Publications Ltd ISSN 1465-6566
`
`875
`
`1. Introduction
`
`2. Immunopharmacology
`
`3. Clinical efficacy
`
`4. Tolerability
`
`5. Conclusion
`
`6. Expert opinion
`
`For reprint orders, please
`contact:
`reprints@ashley-pub.com
`
`Ashley Publications
`www.ashley-pub.com
`
`Page 2 of 18
`
`

`
`Glatiramer acetate
`
`Systemic
`
`Blood–brain
`barrier
`
`CNS
`
`Pre-Rx
`
`On-Rx
`
`Macrophage
`
`MHC
`
`'Myelin' Ag
`
`TCR
`
`GA
`
`'Myelin'
`Ag
`
`Microglia
`
`Neuron
`
`CD4+ Th1
`IL-2
`IFN-γ
`TNF-α
`LT
`
`CD4+ Th2
`IL-4
`IL-5
`IL-13
`IL-10
`BDGF
`
`BDNF
`
`Myelin internode
`
`Figure 1. A simplified diagrammatic representation of the immunopharmacology of GA in MS therapeutics. The Pre-Rx portion
`of the panel emphasises the baseline state in MS with CD4+ Th1 myelin antigen-reactive cells being activated by systemic antigen
`processing cells, including macrophages, that present foreign antigens that are myelin-like (‘myelin’ Ag) in the context of surface MHC to
`the TCR; invoking the concept of molecular mimicry. Stimulated CD4+ Th1 ‘myelin’ Ag-reactive cells secrete a number of pro-inflammatory
`cytokines (IL-2, IFN-γ, TNF-α and LT). WIth GA therapy, GA may displace some ‘myelin’ Ag. More importantly, on presentation and
`stimulation of GA and ‘myelin’ Ag-reactive CD4+ Th1 cells, GA silences crossreacting CD4+ Th1 ‘myelin’ Ag-reactive cells through anergy,
`apoptosis or antigen-specific mechanisms. Concomitantly, GA stimulates and expands a population of GA-reactive CD4+ Th2 cells that
`secrete anti-inflammatory cytokines (IL-4, -5, -13 and -10) to systemically inhibit ‘myelin’ Ag-reactive CD4+ Th1 cells (red arrow). With
`continued therapy the net result is a reduced proportion of CD4+ Th1 and an increased proportion of GA and ‘myelin’ crossreactive CD4+
`Th2 cells. When these GA and ‘myelin’ crossreactive CD4+ Th2 cells gain access to the CNS by trafficking across the blood–brain barrier,
`they are restimulated by true myelin Ags processed and presented by microglia, a brain-resident macrophage. On restimulation, the GA-
`reactive CD4+ Th2 cells secrete anti-inflammatory cytokines to inhibit ‘myelin’ Ag reactive CD4+ Th1 cells within the CNS and also secrete
`tropic factors, such as BDNF that may facilitate neuronal survival (green arrow). Modified from Neuhaus [14] and other sources.
`Ag: Antibody; BDNF: Brain-derived neurotrophic factor; BDGF: Brain-derived growth factor; GA: Glatiramer acetate; LT: Leukotriene; MHC: Major histocompatibility
`complex; MS: Multiple sclerosis; Rx: Treatment; TCR: T cell receptor.
`
`worldwide, including the US, Canada, Australia, Europe and
`Israel. A comprehensive review of GA was published in this
`journal in 2001 [6]. New data on the long-term clinical experi-
`ence with GA, comparative studies, therapy switching studies
`and magnetic resonance imaging (MRI) findings have since
`become available.
`GA is the acetate salt of a synthetic mixture of polypeptides
`that consists of random sequences of four naturally-occurring
`amino acids: L-glutamic acid, L-lysine, L-alanine and L-tyrosine
`in racemeric form at a defined molar ratio of 1.4:3.4:4.2:1.0,
`respectively. The copolymer was first synthesised in 1967 by
`Arnon et al. at the Weizmann Institute of Science [7] in an
`attempt to simulate some of the then known physicochemical
`
`properties of myelin basic protein (MBP) to induce and then
`dissect experimental allergic encephalomyelitis (EAE). EAE is a
`laboratory animal model of organ-specific autoimmune CNS
`inflammatory disease with some similarities to MS. Even
`though GA proved incapable of inducing EAE, it did demon-
`strate a marked effect in suppressing EAE when animals were
`subsequently challenged with MBP [8,9].
`
`2. Immunopharmacology
`
`The mechanisms of action of GA in humans remain uncertain
`but substantial preclinical data support both immunomodula-
`tory and neuroprotective effects of the drug. At least five
`
`876
`
`Expert Opin. Pharmacother. (2004) 5(4)
`
`Page 3 of 18
`
`

`
`interdependent processes are thought to contribute to the
`effects of GA (Figure 1):
`
`(cid:127) High affinity binding to the major histocompatibility
`complex (MHC) within the antigen binding pocket.
`(cid:127) Competition with MBP at the antigen-presenting cell
`(APC) level for binding to MHC and subsequent inhibi-
`tion of MBP-specific T cell activation through competition
`with MBP–MHC complexes for the T cell receptor.
`(cid:127) Induction of a shift in GA-reactive T cells from a T helper
`Type 1 (Th1) to a T helper Type 2 (Th2) phenotype.
`(cid:127) Migration of GA-specific T cells into the CNS.
`(cid:127) Neuroprotection induced via promotion of neurotrophic
`factors.
`
`As would be expected of repeated injection of any foreign pro-
`tein, immunisation with GA consistently induces GA anti-
`bodies; whether these may contribute or detract from the
`clinical effects of the drug are also considered in this section.
`
`2.1 High-affinity binding to major histocompatibility
`complex
`Fundamental to the development of an antigen-specific,
`T-cell-dependent immune response is the processing and
`presentation of a fragment of the antigen by an APC to a
`T cell precursor. This occurs when an appropriately processed
`antigen is bound by physicochemical interactions within the
`antigen-binding cleft of the MHC of an APC. The resulting
`unique structure is presented on the cell surface of the APC
`where it can interact with the complementary hypervariable
`portions of the T-cell receptors of appropriate T cells. Forma-
`tion of this trimolecular complex is a critical, although not
`necessarily sufficient, prerequisite to stimulating signals that
`activate and condition the behaviour of the T cell.
`Intact GA binds directly to MHC displayed on fixed APCs
`[10]. This binding can be blocked by anti-DR, but not anti-
`DQ, or anti-Class I antibodies, and GA binding to Class II
`must occur at, or very near to, the peptide-binding cleft. Iso-
`lated DR molecules exposed to GA form covalently linked
`complexes [10] and this interaction is not easily blocked by the
`staphylococcal B antigen. Staphylococcal B antigen has a
`known binding site to the Class II antigen that resides out-
`side of the antigen-binding cleft. The binding of GA to a
`Class II antigen is of high avidity and has been demonstrated
`for all common MS-associated DR haplotypes. Based on the
`crystallographic structure of the immunodominant peptide
`of MBP and DR2 [11], it appears that the repeated alanines
`and tyrosines in GA may facilitate the anchoring of GA
`within binding pockets of the Class II binding cleft. Further,
`variation in amino acid sequence inherent to GA could
`account for its ability to efficiently bind to a wide array of
`different Class II haplotypes. However, the high-affinity
`interaction between GA and Class II antigen alone is not suf-
`ficient to explain the mechanism of action of the drug, as the
`immunobiologically inert dextrorotatory form of GA binds
`with similar avidity to DR.
`
`Wolinsky
`
`2.2 Competition with myelin basic protein
`In vitro studies have shown that GA competes with MBP at
`the level of APC for binding to the MHC [12]. GA appears to
`have greater MHC-binding affinity than MBP and other
`myelin-associated proteins (e.g., proteolipid protein [PLP]
`and myelin oligodendrocyte glycoprotein [MOG]). Thus, GA
`efficiently displaces MBP-, PLP- and MOG-derived peptides
`from the MHC binding site, but is not displaced by these
`antigens once it is bound to the MHC [6]. GA isomers have
`the same effect on MHC but do not suppress EAE [13]. After
`binding to MHC, the GA–MHC complex competes with
`available MBP/MHC molecules for binding to T-cell recep-
`tors. As a consequence, some of the myelin-specific, patho-
`genic T cells may become anergic or otherwise altered [14].
`
`2.3 Induction of shift from T helper Type 1 to T helper
`Type 2 lymphocytes
`Data from animal models and ex vivo studies of human lym-
`phocytes show that exposure to GA induces a relative anti-
`inflammatory state by causing a shift in the GA-reactive lym-
`phocyte population from a dominant Th1 state to a Th2 dom-
`inant state [14,15]. Th1 cells produce IL-2, IL-12, IFN-γ and
`TNF-α, which generally behave as pro-inflammatory
`cytokines, whereas Th2 cells produce IL-4, -5, -6, -10 and -13,
`which generally exert anti-inflammatory effects. GA-reactive
`peripheral blood lymphocytes from untreated MS patients
`mostly express TNF-α mRNA, whereas those harvested from
`GA-treated patients mainly express IL-10, transforming
`growth factor (TGF)-β and IL-4 mRNA [16]. The shift towards
`Th2 bias is also demonstrated by the diminished ratio of
`IFN-γ/IL-5 secretion of GA-reactive T cell lines isolated from
`MS patients both before and during GA therapy [15,17]. The
`GA-reactive Th2 cells are believed to act as regulatory cells to
`modulate the pathogenic immune reaction. Many T cell lines
`reactive to a number of potentially encephalitogenic myelin
`proteins, when stimulated in vitro with GA, do not proliferate
`but do secrete cytokines with a predominant IL-5 pattern [18].
`It is important to realise that two phenomena are occurring
`in concert as patients begin therapy with GA. Both naive and
`memory GA-reactive CD4+ T cells are part of the resident
`T cell repertory of mice and men [19]. With initiation of ther-
`apy, the numbers of GA-reactive T cells that can be found using
`proliferation assays progressively falls after a transient increase
`within the first month of therapy. The reduced response is evi-
`dent within 3 – 6 months, substantial at 12 months and is
`decreased by 75% from baseline at 24 months after initiating
`treatment [20]. The proportion of patients with a negative
`in vitro proliferative response to GA increased from 5% at base-
`line to 40% at 2 years. Moreover, treatment with GA results in
`increased apoptosis of a substantial percentage of activated
`(CD69+) CD4+ T cells [21]. Thus, as the shift from a Th1 to a
`Th2 state is being established, the numbers of Th1 GA-reactive
`cells is also falling. Many of the above effects are likely to occur
`systemically, resulting in reduced availability of autoaggressive
`cells for entry into the CNS over time.
`
`Expert Opin. Pharmacother. (2004) 5(4)
`
`877
`
`Page 4 of 18
`
`

`
`Glatiramer acetate
`
`The Th2-biased immunological response seen with GA is
`sustained over long-term treatment. Chen et al. [22] isolated
`48 GA-reactive T cell lines from 10 RR MS patients who had
`taken GA for 6 – 9 years. Proliferative responses, cytokine pro-
`duction and crossreactivity with whole MBP and the MBP
`immunodominant peptide 83-99 were compared with
`responses obtained from 10 MS patients tested before GA
`treatment and after shorter treatment periods (1 – 10 months).
`Long-term treatment with GA resulted in a 2.9-fold decrease
`in the estimated precursor frequency of GA-reactive T cells.
`Nevertheless, the sustained response to GA remained Th2-
`biased and, in part, crossreactive with MBP and MBP (83-99),
`as measured by proliferation and cytokine-release assays [22].
`
`2.4 Migration of activated T cells into the central
`nervous system
`Areas of active CNS demyelination and axonal loss in MS
`white matter lesions contain substantial numbers of inflamma-
`tory cells. GA does not appear to directly inhibit the transmi-
`gration of these inflammatory cells into the CNS. Rather,
`repeated systemic activation by daily GA treatment stimulates
`GA-reactive T cells, which increasingly become Th2-like [23].
`In vitro models demonstrate that Th2 cells can penetrate the
`CNS [24]. Adoptively transferred GA-reactive T cells adminis-
`tered systemically to recipient mice are detectable in the ani-
`mals’ CNS [23,25]. GA-reactive Th2 cells in the CNS are
`postulated to decrease local inflammation through ‘bystander
`suppression’. It is thought that GA-specific Th2 cells within the
`CNS are restimulated by-products of myelin turnover pre-
`sented by local APCs. The antigen presented within the CNS
`cannot be GA, as the drug is rapidly metabolised in subcutane-
`ous tissue at the administration site. Local reactivation of GA-
`specific T cells stimulates the release of anti-inflammatory
`cytokines such as IL-4, -6, -10, TGF-β and brain-derived neu-
`rotrophic factor (BDNF), but not IFN-γ [25,26]. The produc-
`tion of pro-inflammatory cytokines, including IL-2 and IFN-γ,
`is inhibited through this bystander effect. The mechanism of
`bystander suppression may make GA useful in other autoim-
`mune diseases of the CNS where Th1 cells predominate [27].
`
`2.5 Neuroprotection
`A recently identified mechanism of action of GA is related to
`a potential neuroprotective effect of some autoreactive Th1
`and Th2 cells. This possibility was first raised after unex-
`pected findings were reported in a rodent optic nerve crush
`injury model, which typically results in a predictable loss of
`retinal-ganglion neurons. In early experiments, animals
`injected with MBP-reactive T cells immediately after the
`crush injury exhibited attenuated subsequent loss of retinal
`ganglion neurons, but also suffered adoptive transfer EAE [28].
`In subsequent experiments, rats subjected to optic nerve crush
`injury and then injected with GA-specific T cells showed
`increased retinal ganglion neuron survival, compared with
`injured controls, and did not develop EAE [29]. Moreover, in a
`murine model in which intraocular injection of glutamate
`
`destroys retinal ganglion neurons, glutamate toxicity was
`reduced in mice immunised with GA, but not in those immu-
`nised with either MBP or MOG [30]. The neuroprotective
`effects of GA have since been demonstrated in several animal
`models. Compared with untreated controls, GA treatment
`reduced axonal damage in C57/bl mice with chronic EAE [31],
`and increased survival time and improved motor function in a
`murine model of amyotrophic lateral sclerosis (ALS) [32].
`A variety of mechanisms underpinning the neuroprotective
`effects of GA are currently under investigation. Kayhan et al.
`[33] demonstrated that in mice, induction of EAE leads to
`fourfold elevation in nitric oxide (NO) secretion. NO is an
`inflammatory mediator thought to affect regulation of the
`immune response, permeability of the blood–brain barrier,
`trafficking of cells to the CNS and immunosuppression. NO
`has been implicated in primary demyelination via nonspecific
`damage to the myelin sheath of axons, as well as promoting
`direct oligodendrocyte death [34]. Treatment of EAE mice
`with GA leads to a significant decrease in NO secretion by the
`splenocytes in response to the encephalitogen [33].
`Another neuroprotective mechanism may involve GA-stim-
`ulated secretion of BDNF, a neurotrophic factor that plays an
`important role in plasticity and development of the nervous
`system [27,35]. Ziemssen et al. [36] determined that GA-specific
`Th2 and Th1 peripheral blood mononuclear cells produce
`BDNF. Using three GA-specific long-term T cell lines with
`phenotypes Th1, Th1/0 and Th0 derived from a healthy sub-
`ject and one T cell line with a Th2 phenotype derived from a
`GA-treated MS patient, they demonstrated that all four T cell
`lines could be stimulated to produce BDNF [36]. Similarly,
`Chen et al. [26] studied BDNF production in 73 GA-reactive,
`13 MBP-reactive and two tetanus toxoid (TT)-reactive T cell
`lines isolated from 12 MS patients treated with GA. BDNF
`levels produced by GA-specific T cells generated during treat-
`ment were higher than those generated pretreatment. Among
`the 73 GA T cell lines generated, 14% secreted levels of
`BDNF two standard deviations above the GA T cell line
`mean. All GA-reactive T cells that secreted high levels of
`BDNF were Th2 biased. A total of 26, 14 and 13 GA-, MBP-
`and TT T cell lines, respectively, originated from the same
`4 MS patients and could be compared directly. The mean
`BDNF level for the GA-reactive T cell lines was significantly
`higher than that for the MBP- and TT-reactive T cell lines.
`The signal transducing receptor for BDNF, the full-length
`145 tyrosine kinase receptor (trk) B, is expressed in neurons
`and astrocytes in MS lesions [37]. Therefore, BDNF secreted
`by GA-reactive Th1 and Th2 cells in the CNS could exert
`neurotrophic effects directly in the MS target tissue. GA-
`reactive T cells also appear to have a reduced ability to trans-
`form bipolar microglia into a morphologically activated ame-
`boid form [38].
`
`2.6 Glatiramer acetate antibodies
`Immunisation of mice and other laboratory animals with GA
`results in the development of polyclonal GA antibodies.
`
`878
`
`Expert Opin. Pharmacother. (2004) 5(4)
`
`Page 5 of 18
`
`

`
`There is little evidence of crossreaction of the polyclonal GA
`antibodies with MBP or other myelin proteins. However,
`some IgM monoclonal antibodies (mAb) generated against
`either GA or murine MBP do show substantial crossreactivity
`in binding and competition assays [39]. Of interest, polyclonal
`murine GA IgG antibodies stimulate remyelination in a
`murine viral model of inflammatory demyelination [40].
`Patients receiving GA in controlled trials develop demon-
`strable levels of GA-binding antibodies within 1 month of
`starting treatment that greatly exceed background levels found
`in placebo-treated patients. These antibodies are restricted to
`the IgG class, with IgG1 levels several-fold higher than IgG2
`levels and low levels of IgG4 [41] – patterns consistent with a
`Th2-driven response. The GA antibodies peak within
`3 months of treatment initiation, reaching levels of 8- to
`20-fold higher than baseline. GA antibody titres decrease by
`month 6 of therapy, but persist at detectable levels for at least
`2 years of continued treatment [20]. In patients treated for
`2 years, those who were relapse-free at 18 and 24 months of
`therapy had statistically higher GA antibody titres than
`treated patients who had one or more on-trial relapses. No
`correlation between antibody titre and either disability as
`measured by categorical change from baseline EDSS or side
`effect profile, was observed [20].
`Extensive attempts to show that IgG class GA mAbs
`inhibit cellular responses to GA either in vitro and in vivo
`have failed to do so [42]. These have included attempts to
`block the binding of GA to isolated MHC molecules, to
`inhibit the responses of GA-specific T cell lines to proliferate
`or secrete cytokines upon stimulation by GA and to inhibit
`the effect of GA in EAE, either by premixing the mAbs with
`the GA inocula or with passive transfer of the immunised
`mice. Similar negative results were found with high-titre
`human GA-antibody sera from treated MS patients. Moreo-
`ver, 53 serum samples from 34 GA-treated patients failed to
`reduce the proliferative response of a murine GA-specific
`T cell clone to GA or reduce the competitive inhibition of
`proliferation of an MBP-specific human T cell clone by GA.
`Discrepant results were reported by Zhang et al. [43]. Using
`an enzyme-linked
`immunosorbent assay (ELISA)-based
`binding assay, GA antibodies were found in 48% of
`42 RR MS patients treated with GA for 1 – 5 years, with
`most antibody-positive patients seen after 1 year on therapy.
`Of the 14 high-titred sera that were found to inhibit the pro-
`liferation of normal donor peripheral blood cells to GA,
`6 were selected for further study. At low dilution, these 6 sera
`inhibited the proliferate responses of a panel of GA-specific
`T cell lines to GA to a varying extent.
`In summary, the predominance of available data does not
`suggest GA antibodies influence the therapeutic effect of the
`drug. However, given the considerable importance of binding
`and neutralising antibodies to IFN-β that develop during ther-
`apy on longer-term clinical and MRI outcomes [44], additional
`studies of long-term GA-treated patients are appropriate.
`
`Wolinsky
`
`3. Clinical efficacy
`
`3.1 Controlled clinical trials
`3.1.1 Bornstein single centre study
`The first double-blind, randomised, placebo-controlled trial
`of GA in patients with RR MS was conducted in the US in
`the 1980s [45]. A total of 50 patients, 20 – 35 years of age
`(mean: 30.5 years) with ≥ 2 relapses during the 2 years before
`study enrollment (average number: 3.9 relapses) with a
`Kurtzke Disability Status Scale (DSS) score of ≤ 6 (mean
`score: 3) were studied. Patients received GA or placebo
`20 mg/day s.c. for 2 years. The treatment groups were
`matched for gender, number of prior relapses and extent of
`entry disability (DSS 0 – 2 or 3 – 6). The primary end point
`was the proportion of relapse-free patients. Secondary end
`points included the frequency of relapses, change in DSS
`score from baseline and time to progression (defined as
`≥ 1 DSS unit increase maintained for ≥ 3 months).
`The proportion of relapse-free patients was significantly
`lower in the GA-treated cohort (14 of 25 [56%] GA versus
`6 of 23 [26%] placebo; p = 0.045). The overall 2-year relapse
`frequencies were 0.6 and 2.7 in the GA and placebo groups,
`respectively. Multiple regression analysis showed baseline DSS
`score significantly influenced relapse rate; a lower disability
`score at baseline increased the likelihood that a patient would
`remain free of exacerbations (p = 0.003). At the end of the
`2-year study, 84.6% of patients in the lower DSS stratum
`(0 – 2) taking GA were stable or improved versus 30% in the
`placebo group. Mean improvement was 0.5 DSS units with
`GA, whereas those taking placebo worsened by an average of
`1.2 DSS units (p = 0.012). In the higher DSS stratum (3 – 6),
`similar proportions of patients in each treatment group were
`stable, improved or worsened; on average these subjects wors-
`ened by 0.3 (GA) and 0.4 (placebo) DSS units.
`
`3.1.2 The US multi-centre study
`This study had three phases: a placebo-controlled, 24-month,
`double-blind treatment phase (core study); a blinded exten-
`sion phase (up to 36 months) that preserved the original treat-
`ment assignments; and an open-label extension phase in
`which patients who received placebo crossed-over to active
`treatment with GA whereas patients that received GA during
`the double-blind period continued to receive GA. The open-
`label extension phase is ongoing and now in its 12th year; data
`are available from the 6-, 8- and 10-year time points. The pri-
`mary end point (for all study phases) was the number of
`relapses. Secondary end points included the proportion of
`relapse-free patients, time to first relapse, change in Expanded
`Disability Status Scale (EDSS) score relative to baseline and
`proportion of patients with sustained disease progression
`(defined as ≥ 1 point increase in EDSS persisting for
`≥ 3 months). EDSS evaluations were made every 3 months
`during the blinded phase (up to 35 months) and every
`6 months thereafter. Patients were seen within 7 days of each
`
`Expert Opin. Pharmacother. (2004) 5(4)
`
`879
`
`Page 6 of 18
`
`

`
`Glatiramer acetate
`
`suspected relapse. When possible, the same neurologist and
`nurse coordinator completed the assessments of each patient.
`3.1.2.1 Core study phase
`The original study, conducted between 1991 and 1994
`enrolled 251 RR MS patients; 18 – 45 years of age (mean age:
`34 years) [46]. In order to be eligible, patients had to have
`≥ 2 relapses in the 2 years preceding study enrollment (mean
`2.9) and an EDSS score of ≤ 5 (mean 2.5). Patients were ran-
`domised to GA or placebo. Results at 24 months showed a
`relapse rate of 1.19 ± 0.13 for patients receiving GA and
`1.68 ± 0.13 for placebo patients, a 29% reduction in favour of
`GA (p = 0.007); annualised rates were 0.59 and 0.84, respec-
`tively. The proportions of patients who improved, were
`unchanged or worsened by ≥ 1 EDSS unit between baseline
`and the end of 2 years of treatment favoured GA (p = 0.037).
`At 24 months, 33.6 and 27.0% of GA- and placebo-treated
`patients, respectively, were relapse-free (p = 0.098). A post hoc
`analysis suggested that the therapeutic effect of GA may have
`been more pronounced in patients with lower EDSS scores
`(≤ 2 units) at baseline.
`3.1.2.2 Double-blind extension phase
`A total of 203 patients entered the blinded extension. By the
`end of this phase (∼ 35 months of treatment), there was a 32%
`reduction in mean relapse rate with GA (1.34 and 1.98 GA
`versus placebo, respectively; p = 0.002); annualised relapse
`rates were 0.58 and 0.81, respectively [47]. During the entire
`extended trial, 33.6 and 24.6% of GA- versus placebo-treated
`patients remained relapse-free, respectively (p = 0.035). No
`GA-treated patients who were relapse-free during the core
`24-month phase experienced a relapse during the extension
`period. The median time to first relapse was 287 days with GA
`and 198 days with placebo (p = 0.057). Subjects who received
`placebo were significantly more likely to experience multiple
`relapses (p = 0.008). The proportion of patients who improved
`by ≥ 1 EDSS units from entry favoured GA (27.2 and 12.0%
`GA versus placebo, respectively; p = 0.001). Worsening by
`≥ 1.5 EDSS units occurred in 21.6 and 41.6% of GA and pla-
`cebo-treated patients, respectively (p = 0.001). The mean
`EDSS score improved by -0.11 in the GA group and worsened
`by +0.34 in the placebo group (p = 0.006).
`3.1.2.3 Open-label extension phase
`Of the original 251 patients, 208 patients chose to continue
`on the open-label phase of the study. The annualised relapse
`rate of patients treated from the beginning of the study
`dropped each year. At ≥ 6 years of continuous treatment with
`GA since trial entry, 26 of 101 (25.7%) remained relapse free.
`The mean annualised relapse rate for those who received GA
`from randomisation was 0.42 (95% CI = 0.34 – 0.51), a 72%
`decrease compared with their prestudy rate, with a relapse rate
`of 0.23 during year 6 of treatment [48,49]. Of those treated with
`GA from study inception, 69.3% were either neurologically
`unchanged (within 0.5 EDSS units of baseline) or had
`improved by at least 1 unit on the EDSS [49].
`Data beyond 6 years has thus far been presented only in
`abstract form. At 8 years, 142 (56.6%) of the original patients
`
`remained in the study [50]. The annual relapse rate declined to
`0.2. The mean EDSS for the entire cohort was 3.14, which
`represents an increase of 0.5 units from randomisation. It is
`difficult to draw firm conclusions from this long-term cohort
`in the absence of a true placebo group and given the patient
`attrition that occurred over time. However, natural history
`studies of comparable patients would predict a higher level of
`neurological disability, with 50% of patients reaching an
`EDSS of 6 by year 12 following the onset of MS [51]. When
`the 72 patients always on GA were compared to those origi-
`nally randomised to placebo, there were still discernible disa-
`bility differences.
`Of the 208 patients who entered the open-label phase of
`the study, 133 began year 10 of treatment and 122 completed
`that year [52]; 64 were originally randomised to GA and 69 to
`placebo. Before randomisation, their annualised relapse rates
`were 1.52 and 1.46, respectively. By year 10, annual relapse
`rates were 0.22 and 0.23. The mean EDSS score for the group
`always on GA increased from randomisation by 0.9 EDSS
`units to 3.67. However, the majority (64.4%) remained stable
`or improved from randomisation. When patients who were
`switched from placebo to active treatment were compared to
`those continuously on GA, the proportion of patients having
`confirmed progression during the entire 10-year study period
`differed significantly (50 GA from entry versus 72 originally
`randomised to placebo; p = 0.015). These observations are
`consistent with the growing conviction that early and
`extended treatment offers the best outcomes in RR MS.
`
`3.1.3 Meta-analysis of the double-blind,
`placebo-controlled clinical trials
`A meta-analysis using pooled data from all 540 patients in
`three randomised, double-blind, placebo-controlled trials was
`recently published [53]. It was designed to investigate whether
`the treatment effect varied according to disease-related varia-
`bles at baseline. In all trials, patients had RR MS for at least
`1 year, at least 1 or 2 relapses during the 2 years prior to study
`entry and a baseline DSS or EDSS score ranging from 0 to 6.
`Inclusion criteria for one study included at least one gadolin-
`ium (Gd)-enhancing lesion on a MRI screening of the brain
`[54]. Durations of the placebo-controlled phases of the trials
`were 24 [45], 35 [46] and 9 [54] months. Regression models were
`developed to estimate the annualised relapse rate, total number
`of on-trial relapses and time to first relapse. Also explored and
`recently reported in poster form were the effect of GA on