`
`Brain (2004), 127, 1370–1378
`
`Multiple sclerosis: glatiramer acetate inhibits
`monocyte reactivity in vitro and in vivo
`
`Martin S. Weber,1 Michaela Starck,2 Stefan Wagenpfeil,3 Edgar Meinl,1,4 Reinhard Hohlfeld1,4 and
`Cinthia Farina4
`
`Correspondence to: Dr R. Hohlfeld, Institute for Clinical
`Neuroimmunology, Marchioninistrasse 15, D-81377
`Munich, Germany
`E-mail: hohlfeld@neuro.mpg.de
`
`Downloaded from
`
`by guest on October 30, 2014
`
`(detected by cytofluorometry), and by production of
`(TNF)-a
`monocyte-derived tumour necrosis
`factor
`(detected by enzyme-linked immunospot assay). GA had
`a broad inhibitory effect on all measures of monocyte
`reactivity, regardless of which stimulator was used. It is
`unlikely that this reflects a simple toxic effect, because
`monocyte viability and CD14 expression were unaf-
`fected. In a second series of experiments, we investi-
`gated the properties of monocytes cultured ex vivo from
`eight GA-treated multiple
`sclerosis patients,
`eight
`untreated multiple sclerosis patients and eight healthy
`subjects. We found that LPS-induced SLAM expression
`and TNF-a production were significantly reduced in
`monocytes from GA-treated patients compared with
`controls. These results demonstrate for the first time
`that GA inhibits monocyte reactivity in vitro and in vivo,
`significantly extending the current concept of the mech-
`anism of action of GA.
`
`1Institute for Clinical Neuroimmunology, Klinikum
`Großhadern, Ludwig Maximilians University, Munich,
`2Marianne-Strauss-Klinik, Berg, 3Institute of Medical
`Statistics and Epidemiology, Technical University, Munich
`and 4Department of Neuroimmunology, Max Planck
`Institute of Neurobiology, Martinsried, Germany
`
`Summary
`It is widely assumed that glatiramer acetate (GA), an
`approved agent for the immunomodulatory treatment
`of multiple sclerosis, acts primarily as an antigen for T
`lymphocytes. Recent studies, however,
`indicated that
`in vitro, GA directly inhibits dendritic cells, a rare but
`potent
`type of professional antigen-presenting cell
`(APC). To investigate whether these in vitro observa-
`tions are relevant to the actions of GA in vivo, we
`studied the effects of GA on monocytes, the major type
`of circulating APC. In a first series of experiments, we
`investigated the effects of GA on monocyte reactivity
`in vitro. Monocytes were stimulated with ligands for
`Toll-like receptor (TLR)-2 (peptidoglycan and lipo-
`teichoic acid), TLR-4 [lipopolysaccharide (LPS)] and
`TLR-5 (flagellin), as well as two proinflammatory cyto-
`kines (interferon-g and granulocyte–monocyte colony-
`stimulating factor). Monocyte activation was measured
`by induction of the surface markers signalling lympho-
`cytic activation molecule (SLAM), CD25 and CD69
`
`Keywords: multiple sclerosis; glatiramer acetate; immunotherapy; monocyte; ex vivo assay
`
`Abbreviations: APC = antigen-presenting cell; DC = dendritic cell; EDSS = Expanded Disability Status Scale;
`Elispot = enzyme-linked immunospot; FACS = fluorescence activated cell sorting; GM-CSF = granulocyte–monocyte
`colony-stimulating factor; GA = glatiramer acetate; IFN = interferon; IL = interleukin; LPS = lipopolysaccharide;
`LTA = lipoteichoic acid; PBMC = peripheral blood mononuclear cell; PGN = peptidoglycan; SLAM = signalling
`lymphocytic activation molecule; TLR = Toll-like receptor; TNF = tumour necrosis factor
`
`Received December 3, 2003. Revised February 4, 2004. Accepted February 5, 2004. Advance Access publication April 16, 2004
`
`Introduction
`sclerosis lesions (Johnson et al., 1995, 2000; Mancardi et al.,
`Glatiramer acetate (GA, Copaxone), a random copolymer
`1998; Ge et al., 2000; Comi et al., 2001; Filippi et al., 2001;
`composed of alanine, glutamic acid, lysine and tyrosine, is an
`Wolinsky et al., 2001; Ziemssen et al., 2001). Over recent
`approved agent for the immunomodulatory treatment of
`years, a number of studies have addressed the immunological
`multiple sclerosis. Clinical and magnetic imaging studies
`basis of the clinical effects of GA in animal models and
`indicated that GA treatment reduces the activity of multiple
`Brain Vol. 127 No. 6 ª Guarantors of Brain 2004; all rights reserved
`
`Page 1 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Downloaded from
`
`by guest on October 30, 2014
`
`human multiple sclerosis (reviewed in Aharoni et al., 2000;
`Neuhaus et al., 2001; Yong, 2002). These studies consistently
`showed that GA induces a ‘TH1 to TH2 cytokine shift’ in
`GA-reactive CD4+ T cells. In the immune system of untreated
`multiple sclerosis patients and healthy subjects, the majority
`of GA-reactive CD4+ T cells belong to the TH1 subset. These
`cells characteristically produce pro-inflammatory cytokines
`such as interleukin (IL)-2 and interferon-g (IFN-g) (Murphy
`and Reiner, 2003). During treatment with GA, the cytokine
`profile of the GA-reactive T cells shifts towards the TH2 type,
`which is characterized by production of anti-inflammatory
`‘TH2’ cytokines, including IL-4 (Neuhaus et al., 2000; Farina
`et al., 2001; Dhib-Jalbut et al., 2002), IL-5 (Duda et al., 2000)
`and IL-13 (Wiesemann et al., 2003).
`These observations laid the basis for the current concept of
`the mechanism of action of GA. It is thought that activated,
`GA-reactive TH2 T cells migrate across the blood barrier. In
`the CNS,
`they are re-activated locally by cross-reacting
`myelin auto-antigens (Neuhaus et al., 2001). After local re-
`stimulation, the GA-reactive TH2 cells release their anti-
`inflammatory cytokines, and also certain neurotrophic factors
`(Ziemssen et al., 2002). In this way, the recruited TH2 T cells
`can suppress neighbouring autoaggressive TH1 cells. This
`process is called ‘bystander suppression’ (Aharoni et al.,
`1998). In addition, the GA-reactive T cells can deliver brain-
`derived neurotrophic factor
`(BDNF)
`to neurons, which
`upregulate the corresponding full-length signalling tyrosine
`kinase receptor gp145 trkB in multiple sclerosis lesions
`(Stadelmann et al., 2002; Ziemssen et al., 2002).
`The mechanism(s) of the therapeutically induced TH1 to
`TH2 cytokine shift of GA-reactive T cells is unknown.
`Theoretically,
`there are two (not necessarily exclusive)
`possibilities. First, GA might have a primary effect on T
`cells, for example by virtue of its properties as an ‘altered
`peptide ligand’, or by the special conditions of GA presen-
`tation in the skin. Secondly, GA might exert a primary effect
`on antigen-presenting cells (APCs), e.g. by altering their
`properties in such a way that they preferentially induce TH2
`cells. Indeed, there is recent evidence that GA might affect
`the properties of APCs in vitro (Hussien et al., 2001; Vieira
`et al., 2003). In this regard, the findings reported by Vieira
`et al. (2003) are especially intriguing. Using in vitro cultures,
`these authors found that GA affects the T-cell-stimulating
`properties of dendritic cells (DCs). After in vitro treatment
`with GA, DCs have an impaired capacity to secrete TH1-
`polarizing factors, and therefore preferentially induce TH2
`cells (Vieira et al., 2003). Whether these observations extend
`to other types of APC, and whether they are relevant in vivo
`presently is unknown.
`The main aim of our present study was to search for
`evidence that GA affects the function of APCs in vivo. We
`focused our analysis on monocytes, because DCs constitute a
`very rare subset of APCs which can only be studied after
`prolonged expansion in culture, and which are therefore
`inaccessible for direct ex vivo analysis. In contrast, mono-
`cytes represent a major subset of professional APCs which
`
`Glatiramer acetate therapy affects monocytes
`
`1371
`
`can be easily obtained from peripheral blood and are readily
`accessible for ex vivo investigations.
`In a first step of analysis, we performed a detailed series of
`experiments to identify the possible effects of GA on different
`patterns of monocyte activation in vitro. For monocyte
`activation, we used (i) four ligands for three distinct Toll-like
`receptors (TLRs); and (ii) the proinflammatory cytokines
`IFN-g and granulocyte–monocyte colony-stimulating factor
`(GM-CSF). As markers of monocyte activation, we looked at
`(i) the expression of activation-related surface molecules
`including
`signalling
`lymphocytic
`activation molecule
`(SLAM) (CD150), CD25 and CD69; and (ii) production of
`tumour necrosis factor (TNF)-a. These experiments revealed
`that GA broadly affects monocyte activation by different
`ligands and pathways.
`In the second step, we addressed the possible effects of GA
`on monocytes in vivo. To this end, we compared the
`properties of ex vivo monocytes from untreated and GA-
`treated subjects. ‘Ex vivo’ means that the monocytes were
`exposed to GA only in vivo, but not in vitro. Based on the
`experiments from the first step, we chose lipopolysaccharide
`(LPS)-induced SLAM expression and TNF-a secretion as
`‘read-out’ to test the monocyte stimulation thresholds in vitro.
`Our main observation is that monocytes cultured ex vivo from
`GA-treated patients were indeed significantly less susceptible
`to activation than monocytes from untreated patients and
`normal controls. These results demonstrate, we believe for
`the first time, that GA treatment in vivo leads to a systemic
`alteration of the properties of circulating monocytes, raising
`important new questions regarding the mechanism of action
`of GA.
`
`Materials and methods
`Subjects and cell samples
`Blood was drawn from healthy individuals, GA-treated multiple
`sclerosis patients and untreated multiple sclerosis patients after their
`informed consent. This study has been approved by the local ethics
`commitee of the Ludwig Maximilians University of Munich. All
`patients had definite multiple sclerosis (McDonald et al., 2001). All
`GA-treated patients (n = 8) had a relapsing–remitting disease course.
`At
`the time of sampling,
`they had injected 20 mg of GA
`subcutaneously (s.c.) daily for at least 1 year, with a mean treatment
`duration of 34.4 6 16.9 months [mean Expanded Disability Status
`Scale (EDSS) at time of sampling 1.75 6 1.2; mean age 31.8 6 8.4
`years] (Table 1). The untreated group included four patients with
`relapsing–remitting multiple sclerosis, two with primary progressive
`multiple sclerosis and two with secondary progressive multiple
`sclerosis (five women and three men; mean EDSS 3.5 6 1.9; mean
`age 47.1 6 10.5 years). None of these patients was treated with
`immunosuppressive or immunomodulatory therapy during at least 3
`months preceding the study. The group of healthy donors included
`four men and four women with a mean age of 35.1 6 11.9 years.
`Peripheral blood mononuclear cells (PBMCs) were isolated on a
`discontinuous density gradient
`(Lymphoprep, Nycomed, Oslo,
`Norway). Viable cells were counted by trypan blue (Sigma-
`Aldrich) exclusion and resuspended in culture medium [RPMI
`
`Page 2 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Downloaded from
`
`by guest on October 30, 2014
`
`1372
`
`M. S. Weber et al.
`
`Table 1 Characteristics of GA-treated MS patients and untreated MS patients.
`
`Subjects
`
`Gender
`
`Age
`(years)
`
`Year of
`diagnosis
`
`Disease
`course
`
`EDSS*
`
`GA treatment
`(months)*
`
`GA-treated multiple sclerosis patients
`MS-GA-1
`MS-GA-2
`MS-GA-3
`MS-GA-4
`MS-GA-5
`MS-GA-6
`MS-GA-7
`MS-GA-8
`Untreated multiple sclerosis patients
`MS-NT-1
`MS-NT-2
`MS-NT-3
`MS-NT-4
`MS-NT-5
`MS-NT-6
`MS-NT-7
`MS-NT-8
`
`F
`F
`F
`M
`F
`F
`F
`F
`
`F
`F
`F
`M
`F
`M
`M
`F
`
`34
`21
`13
`62
`28
`32
`28
`57
`
`43
`30
`33
`25
`21
`33
`45
`25
`
`31
`36
`47
`39
`54
`56
`57
`57
`
`1981
`1991
`1995
`1988
`1999
`1999
`1993
`1993
`
`1996
`2000
`1997
`1982
`1975
`1976
`1996
`2000
`
`RR
`RR
`RR
`RR
`RR
`RR
`RR
`RR
`
`RR
`RR
`RR
`RR
`SP
`SP
`PP
`PP
`
`3.5
`1.5
`1.0
`1.5
`3.5
`1.0
`2
`0
`
`2.0
`2.5
`2.0
`1.0
`4.5
`6.5
`5.5
`4.0
`
`*At the time of sampling; MS-GA = GA-treated multiple sclerosis patient; MS-NT = untreated multiple sclerosis patient; RR = relapsing–
`remitting multiple sclerosis; SP = secondary progressive multiple sclerosis; PP = primary progressive multiple sclerosis.
`
`1640 supplemented with 5% fetal calf serum (FCS), 1% glutamine
`and 1% penicillin/streptomycin; Gibco]. One batch of FCS was used
`throughout the study.
`
`Reagents and antibodies
`The following reagents were used: human IFN-g (Roche, Mannheim,
`Germany), GM-CSF (R&D, Wiesbaden-Nordenstadt, Germany),
`flagellin from Helicobacter pylori (IBT, Reutlingen, Germany),
`peptidoglycan (PGN) from Staphylococcus aureus (Fluka, Sigma-
`Aldrich, Schnelldorf, Germany),
`lipoteichoic acid (LTA) from
`Staphylococcus aureus, and LPS from Escherichia coli 0111:B4
`(Sigma-Aldrich). LPS-free reagents, water (BioWhittaker, Verviers,
`Belgium), phosphate-buffered saline (PBS; Gibco, Karlsruhe,
`Germany) and bovine serum albumin (BSA; Sigma) were used to
`prepare the aliquots. GA (Batch-No. 242992899) was from Teva
`Pharmaceutical Industries Ltd, Petah Tiqva, Israel.
`The following antibodies were used for fluorescence activated
`cell sorting (FACS) analyses: phycoerythrin (PE)-labelled anti-
`SLAM, peridinin chlorophyll protein (PerCP)-labelled anti-human
`CD14, and fluorescein isothiocyanate (FITC)-labelled anti-human
`CD25 and CD69 (all
`from Becton Dickinson, Heidelberg,
`Germany). All
`the corresponding isotype controls were from
`Becton Dickinson.
`
`Enzyme-linked immunospot (Elispot) assay
`96-well polyvinylidene difluoride plates (Millipore, Eschborn,
`Germany) were coated at 4(cid:176)C overnight with the capture antibody
`(anti-TNF-a antibody; Mabtech, Nacka, Sweden). After the wells
`were washed and blocked with culture medium for 1 h at 37(cid:176)C, the
`cells (1 3 103/well for the in vitro and 3 3 103/well for the ex vivo
`TNF-a assay) were seeded and stimulated for 18 h at 37(cid:176)C and 5%
`CO2. The experiments were performed in triplicate. After culture,
`the plates were washed and incubated first with the biotinylated
`detector antibody (Mabtech),
`then with streptavidin–alkaline
`phosphatase (Mabtech), and finally with BCIP/NBT (Sigma-
`Aldrich). The Elispot plates were analysed with an automated
`imaging system and appropriate computer software (KS ELISPOT
`automated image analysis system, Zeiss, Jena, Germany).
`
`Statistical analysis
`The GA dose dependence of the percentage of SLAM-positive
`monocytes and the frequency of TNF-a-producing cells were
`analysed with linear regression. The t test for independent samples
`was used to compare GA-treated multiple sclerosis patients,
`untreated multiple sclerosis patients and healthy controls. All P
`values given are two-sided and subject to a significance level of 5%.
`
`FACS staining and analysis
`The cells were labelled with the predetermined appropriate antibody
`dilution or with the corresponding isotype controls. FACS stainings
`were analysed on a FACScan using Cell-Quest software (Becton
`Dickinson). Monocytes were gated in forward/side scatter. The
`quadrants were set on the relative isotype controls. The in vitro
`experiments were repeated at least three times.
`
`Results
`In vitro culture with GA inhibits monocyte
`activation via different TLR ligands and
`inflammatory cytokines
`In previous experiments, we have characterized monocyte
`responses to (i) different bacterial TLR ligands (TLR-2
`ligands PGN and LTA; TLR-4 ligand LPS; and TLR-5 ligand
`
`Page 3 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Glatiramer acetate therapy affects monocytes
`
`1373
`
`Fig. 1 GA-mediated inhibition of LPS-stimulated monocyte responses in vitro (dose–response analysis). In (A), monocyte reactivity was
`measured as LPS-induced TNF-a production by Elispot assay. In (B), monocyte reactivity was measured as LPS-induced SLAM
`expression by FACS.
`
`Downloaded from
`
`by guest on October 30, 2014
`
`Fig. 2 GA-mediated inhibition of monocyte responses stimulated
`with different TLR ligands (LTA, PGN and flagellin) or
`inflammatory cytokines (GM-CSF and IFN-g). Monocyte reactivity
`was measured as TNF-a production by Elispot assay.
`
`50 mg/ml GA, SLAM induction was inhibited by 66.8 6 37.4
`(PGN stimulation), 80.4 6 30.3 (flagellin) and 87 6 18.3%
`(LTA) (average values 6 SD; data not shown). Similarly, GA
`inhibited TNF-a production by monocytes stimulated with
`half-maximal doses of the different TLR ligands (Fig. 2;
`P = 0.002 for PGN, P < 0.0001 for LTA and flagellin).
`GA inhibited not only TLR-stimulated, but also IFN-g- and
`GM-CSF-stimulated TNF-a release (Fig. 2, P = 0.001 for
`IFN-g and P < 0.0001 for GM-CSF), although in this case
`high concentrations of inflammatory cytokines were neces-
`sary for monocyte stimulation. In three different experiments,
`the inhibitory effect of GA on TNF-a production was
`consistently greater when TLR ligands were used as
`activators (87–97% inhibition), compared with stimulation
`with IFN-g or GM-CSF (47–58% inhibition) (Fig. 2).
`We also investigated the effects of GA on two additional
`markers of monocyte activation, CD25 and CD69. These
`molecules are newly induced on monocytes by stimulation
`with TLR ligands or inflammatory cytokines. Table 2 shows
`the results of one (out of
`three)
`representative FACS
`experiment. Pre-incubation with 50 mg/ml GA reduced the
`
`flagellin), and (ii) inflammatory cytokines (IFN-g and GM-
`CSF). Monocyte activation was measured by (i) induction of
`surface activation markers (CD25, CD69 and SLAM); and
`(ii) cytokine [TNF-a production (Farina et al., 2004)]. We
`found that SLAM was induced preferentially by stimulation
`with TLR ligands but not inflammatory cytokines, whereas
`CD25 and CD69 were induced by both TLR ligands and
`inflammatory cytokines. Further, TNF-a production was
`induced more strongly by TLR ligands than by IFN-g and
`GM-CSF (Farina et al., 2004). These results laid the basis for
`our present study of
`the effects of GA on monocyte
`activation.
`In a typical dose–response curve, a plateau of activation
`was reached with LPS concentrations >1000 pg/ml. At
`plateau, GA had no detectable inhibitory effect on monocyte
`reactivity. For optimal detection of the inhibitory effect of
`GA, we found that low to intermediate concentrations (150–
`1000 pg/ml) of LPS had to be used. PBMCs from healthy
`donors were pre-incubated for 1.5 h with four different
`concentrations of GA, and stimulated overnight with different
`concentrations of LPS. Figure 1 shows one of
`three
`representative experiments for TNF-a production (Fig. 1A)
`and SLAM induction (Fig. 1B). At each LPS concentration,
`the percentage of SLAM-positive monocytes and the fre-
`quency of TNF-a-producing cells was strongly reduced
`(P < 0.05 for TNF-a production by 300, 600 and 1000 pg/ml
`LPS; P < 0.05 for SLAM induction by 600 and 1000 pg/ml
`LPS). As shown in Fig. 1, the inhibitory effect of GA was
`dose dependent. We conclude that GA inhibits monocyte
`responses induced with the TLR-4 ligand LPS.
`We also investigated the effects of GA on monocyte
`activation stimulated with the TLR-2 ligands PGN and LTA,
`and the TLR-5 ligand flagellin. As with LPS, GA had no
`detectable inhibitory effect at very high concentrations of
`these ligands. We measured dose–response curves for each
`ligand and determined the half-maximal stimulating concen-
`trations. Again, the monocyte response was measured in
`terms of SLAM induction and TNF-a production by FACS
`and Elispot assay, respectively. After pre-incubation with
`
`Page 4 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Downloaded from
`
`by guest on October 30, 2014
`
`we assessed LPS-stimulated SLAM induction and TNF-a
`production by monocytes from GA-treated multiple sclerosis
`patients and controls. We compared PBMCs from eight GA-
`treated multiple sclerosis patients, eight untreated multiple
`
`1374
`
`M. S. Weber et al.
`
`fraction of activated monocytes expressing CD25 or CD69,
`regardless of which activator was used for stimulation
`(Table 2). Although the magnitudes of monocyte activation
`and of the inhibitory effect of GA varied between the
`different stimuli, GA consistently reduced the induction of
`CD25 and CD69 by all the different stimuli (CD69: 70%
`average inhibition for all activators; CD25: 70% average
`inhibition for LPS, LTA and PGN, 50% average inhibition for
`flagellin and IFN-g, 30% inhibition for GM-CSF). In contrast,
`CD14, which is constitutively expressed on monocytes, was
`unaffected by GA (Fig. 3).
`
`Treatment with GA in vivo leads to a reduction
`of monocyte reactivity ex vivo
`The experiments described in the previous section indicated
`that GA inhibits monocyte activation by TLR ligands and
`inflammatory cytokines in vitro. To investigate whether
`in vivo treatment with GA also affects monocyte properties,
`
`Table 2 GA blocks induction of activation markers on
`monocytes
`
`CD25*
`
`CD69*
`
`–GA
`
`+GA**
`
`–GA
`
`+GA**
`
`No stimulus
`LPS 0.15 ng/ml
`LTA 0.25 mg/ml
`PGN 0.25 mg/ml
`Flagellin 0.6 mg/ml
`IFN-g 1000 U/ml
`GM-CSF 100 ng/ml
`
`0.4%
`24.1%
`17.4%
`24.6%
`5.3%
`5.9 %
`19.8 %
`
`1.2%
`0.8%
`7.8%
`6.7%
`3.3%
`3.5 %
`15.3 %
`
`4.1%
`16.2%
`13.3%
`15.4%
`11.1%
`13.6 %
`8.5 %
`
`2.4%
`0%
`3.1%
`2.4%
`1.4%
`5.9 %
`0.5 %
`
`*The expression of the indicated activation markers on monocytes
`was determined by FACS and expressed as percentage of positive
`cells; **pre-incubation with 50 mg/ml GA.
`
`Fig. 3 GA-mediated effects on CD25 (A) and CD14 (B)
`expression by LPS-stimulated monocytes. LPS (0.15 ng/ml) and
`GA (50 mg/ml) were used. Upper panels (A) represent dot-blots of
`CD25 expression (ordinate) versus forward scatter (abscissa). GA
`reduced the proportion of CD25+ monocytes from 34.7% (without
`GA, upper left) to 3.1% (with GA, upper right). Lower panels (B)
`represent dot-blots of CD14 expression (ordinate) versus forward
`scatter (abscissa). There was no detectable change of CD14
`expression in the presence of GA (lower right). The two markers
`were analysed in parallel in one experiment using PBMCs from
`one healthy donor.
`
`Fig. 4 LPS-stimulated SLAM induction on monocytes cultured ex vivo from eight healthy donors, eight untreated multiple sclerosis
`patients and eight GA-treated multiple sclerosis patients. SLAM induction was measured by FACS. The monocytes had been exposed to
`GA only during in vivo treatment, not in vitro. Monocytes from GA-treated patients showed significantly reduced responses. *P = 0.002
`healthy controls versus GA-treated patients; P = 0.017 untreated patients versus GA-treated patients; P = 0.185 healthy controls versus
`untreated patients. The right panel shows means from each of the left panels. The P values at lower LPS concentrations were not
`significant.
`
`Page 5 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Glatiramer acetate therapy affects monocytes
`
`1375
`
`Fig. 5 LPS-stimulated TNF-a production by monocytes cultured ex vivo from eight healthy donors, eight untreated multiple sclerosis
`patients and eight GA-treated multiple sclerosis patients. TNF-a production was measured by Elispot. The monocytes had been exposed to
`GA only during in vivo treatment, not in vitro. Monocytes from GA-treated patients showed significantly reduced responses. *P = 0.015
`healthy controls versus GA-treated patients; P = 0.013 untreated patients versus GA-treated patients; P = 0.69 healthy controls versus
`untreated patients; **P = 0.002 healthy controls versus GA-treated patients; P = 0.033 untreated patients versus GA-treated patients;
`P = 0.41 healthy controls versus untreated patients; ***P < 0.0001 healthy controls versus GA-treated patients; P = 0.001 untreated
`patients versus GA-treated patients; P = 0.29 healthy controls versus untreated patients. The right panel shows means from each of the left
`panels.
`
`Downloaded from
`
`by guest on October 30, 2014
`
`sclerosis patients and eight healthy donors. Note that in these
`experiments, the monocytes were not exposed to GA in vitro.
`Figure 4 compares LPS-induced SLAM expression on
`monocytes from each group of patients and control subjects.
`Despite some inter-individual variability, cells from GA-
`treated patients were clearly less susceptible to SLAM
`induction than cells from the control groups. For example,
`at 1000 pg/ml LPS, the mean proportion of SLAM-expressing
`monocytes was 22.1 6 4.4% (mean 6 SD) for GA-treated
`patients compared with 37.2 6 9.5% for healthy subjects
`(P = 0.002), and with 31.1 6 8.0% for the untreated multiple
`sclerosis group (P = 0.017). The two control groups were not
`significantly different (Fig. 4).
`We also compared LPS-induced TNF-a production by
`monocytes from the different groups. Figure 5 shows the
`dose–response curves for each subject as detected by Elispot
`assay. Spontaneous release was similar in the three groups
`(P = 0.41 for healthy controls versus GA-treated patients;
`P = 0.81 for healthy controls versus untreated patients;
`P = 0.23 for untreated patients versus GA-treated patients).
`However, PBMCs from GA-treated individuals were less
`susceptible to in vitro activation even at a low concentration
`of LPS (P = 0.015 and P = 0.013 at 150 pg/ml LPS when
`compared with healthy subjects and untreated multiple
`sclerosis patients, respectively). At increasing LPS concen-
`trations, the differences became more evident (P = 0.002 and
`P = 0.0001 when compared with healthy donors, P = 0.033
`and P = 0.001 when compared with untreated multiple
`sclerosis patients at 600 and 1000 pg/ml LPS, respectively).
`There were no significant differences between the two control
`groups at any of the different LPS concentrations (P values
`ranging from 0.29 to 0.69).
`the curves
`As shown in Figs 4 and 5 (right panels),
`representing the means for GA-treated patients were different
`from the control curves.
`
`Discussion
`The major new observation reported herein is that in treated
`multiple sclerosis patients, GA induces systemic functional
`changes of circulating monocytes. Monocytes from GA-
`treated multiple sclerosis patients are significantly less
`susceptible to LPS stimulation than monocytes
`from
`untreated multiple sclerosis patients and normal control
`subjects. This reduced monocyte reactivity was observed in
`terms of both LPS-induced expression of
`the surface
`activation molecule SLAM, and also LPS-induced secretion
`of the cytokine TNF-a.
`These changes were observed with ‘ex vivo’ monocytes,
`which had been exposed to GA only in vivo, but not in vitro.
`The results are fully consistent with our parallel observations
`on monocytes cultured in the presence of GA in vitro. After
`pre-incubation with GA in vitro, monocytes were less
`susceptible to a broad range of diverse stimuli, including
`ligands of different TLRs, and also different stimulatory
`cytokines (IFN-g and GM-CSF).
`structural
`recognition of
`TLRs are responsible for
`components conserved among classes of microorganisms
`(for a review see Underhill, 2003). Ligand binding induces
`the expression of various host defence molecules, including
`inflammatory cytokines, chemokines, co-stimulatory mol-
`ecules and major histocompatibility complex (MHC) mol-
`ecules. We recently have characterized the complex spectrum
`of human monocyte responses triggered by stimulation with
`TLRs, and compared it with the classical activation induced
`by inflammatory cytokines (Farina et al., 2004). We found
`that TLR ligands are stronger inducers of TNF-a production
`than are the cytokines IFN-g and GM-CSF, and that TLR
`ligands selectively trigger SLAM expression on the monocyte
`surface (Farina et al., 2004). Other activation markers such as
`CD25 and CD69 are induced both by TLR ligands and by
`inflammatory cytokines. In the present study, pre-incubation
`
`Page 6 of 9
`
`YEDA EXHIBIT NO. 2076
`MYLAN PHARM. v YEDA
`IPR2015-00643
`
`
`
`Downloaded from
`
`by guest on October 30, 2014
`
`1376
`
`M. S. Weber et al.
`
`with GA blocked the induction of all investigated surface
`markers (SLAM, CD25 and CD69), and also of TNF-a
`release. Several experiments with enriched monocyte popu-
`lations showed that the inhibitory in vitro effect of GA on
`SLAM induction occurs in the absence of T cells.
`Taken together, these results suggest that GA broadly
`affects monocyte reactivity. The exact mechanism(s) of the
`inhibitory effect is presently unknown, and deserves further
`study. For several reasons, it appears unlikely that GA exerts
`a (trivial) toxic effect on monocytes. First, monocyte viability
`was not altered at the GA concentrations used in our in vitro
`experiments, as seen by propidium iodide uptake. Secondly,
`although GA reduced the induction of the activation-related
`surface molecules SLAM, CD25 and CD69, another surface
`marker of monocytes, CD14 [which is part of the LPS
`receptor complex (Underhill, 2003)] was completely un-
`affected by GA treatment in vitro. Furthermore, T cells from
`GA-treated patients showed the typical, previously described
`Elispot cytokine responses when stimulated with a wide range
`of GA concentrations in vitro (Farina et al., 2001, 2002).
`The new findings have obvious implications for our
`understanding of the mechanisms of action of GA. In several
`previous studies, part of these mechanisms was elucidated in
`some detail. It is widely assumed that the beneficial effects of
`GA are mediated mainly by GA-reactive TH2 cells. Previous
`studies showed that GA treatment induces a gradual shift of
`GA-reactive T cells from TH1 to TH2 (Miller et al., 1998;
`Neuhaus et al., 2000; Duda et al., 2000; Farina et al., 2001).
`Since the GA-reactive T cells are constantly activated by
`immunization, they can gain access to the CNS (Wekerle
`et al., 1986; Hickey, 1991). Indeed, GA-reactive T cells have
`been directly demonstrated in the CNS of adoptively
`transferred animals (Aharoni et al., 2000). It is assumed
`further that after local recognition of cross-reactive myelin
`degradation products, the GA-specific TH2 cells are stimu-
`lated to secrete TH2-like cytokines, which suppress neigh-
`bouring encephalitogenic T cells via ‘bystander suppression’
`(Aharoni et al., 1997, 1998; Neuhaus et al., 2001).
`Furthermore,
`the locally activated GA-reactive T cells
`produce the neurotrophic factor BDNF (Ziemssen et al.,
`2002). Because the BDNF receptor gp145 trk B is expressed
`in neurons in the immediate vicinity of multiple sclerosis
`plaques (Stadelmann et al., 2002), these neurons should be
`responsive to the T-cell-derived BDNF.
`The basic tenets of the above scenario, especially the GA-
`induced TH2 shift of T cells, are widely accepted. It is
`unclear, however, how the cytokine shift is brought about.
`The most widely held view is that GA acts mainly on T cells,
`and that several factors related to the special antigenic
`properties of GA contribute to the observed TH2 shift. These
`include the properties of GA as a co-polymer, its route of
`administration by frequent s.c. injection and perhaps its local
`presentation to T cells by dermal Langerhans cells. In
`contrast, our results indicate that GA directly affects the
`functional properties of ‘professional’ APCs. Consistent with
`our findings, several recent publications showed that GA had
`
`inhibitory effects on TNF-a production by a monocytic cell
`line (Li et al., 1998) and on IL-12 secretion by in vitro
`generated, monocyte-derived DCs (Hussien et al., 2001;
`Vieira et al., 2003). Furthermore, our results show, as we
`believe for the first time, that GA reduces monocyte reactivity
`in treated patients, and that this effect can be observed with
`‘ex vivo’ monocyte cultures.
`What are the implications of these observations? First, GA
`may alter the properties of APCs in such a way that they
`preferentially induce TH2 responses. This is supported by the
`study of Vieira et al. (2003) who demonstrated that GA-
`treated DCs induce IL-4-secreting TH2 cells in vitro. We
`investigated the effects of GA treatment on the TH2 shift in a
`subset of patients. Five of six patients showed a pronounced
`GA-induced IL-4 production by Elispot assay, defined
`according to Farina et al. (2001) (data not shown). These
`findings indicate that the GA-induced TH2 shift and the
`inhibition of monocyte reactivity usually occur concurrently.
`Secondly, monocyte inhibition per se could be beneficial in
`multiple sclerosis, independently of the TH2 shift of T cells.
`Monocytes and macrophages have an important role in the
`pathogenesis of multiple sclerosis (for a review see Izikson
`et al., 2002). Although the composition of the inflammatory
`infiltrate in the CNS varies depending on the type, stage and
`activity of multiple sclerosis, monocytes/macrophages are
`thought to be key effectors responsible for tissue damage.
`They predominate in active multiple sclerosis lesions, and the
`presence of myelin degradation products inside macrophages
`is one of the most reliable markers of lesional activity. Toxic
`products of activated monocytes/macrophages, such as TNF-
`a, reactive oxygen species and matrix metalloproteinases
`(Bar-Or et al., 2003), contribute to myelin injury (for a review
`see Lassmann et al., 2001). Blockade of monocytes is
`therefore an obvious therapeutic goal.
`Thirdly,
`the direct effect of GA on APCs makes it
`necessary to reconsider the concept of ‘antigen selectivity’
`of GA. Although not formally proven, it is widely believed
`that GA preferentially affects myelin basic protein (MBP)-
`specific T-cell responses, at l