Leukemia (1999) 13, 44–53
`(cid:211) 1999 Stockton Press All rights reserved 0887-6924/99 $12.00
`http://www.stockton-press.co.uk/leu
`
`Fas ligand expression in the bone marrow in myelodysplastic syndromes correlates
`with FAB subtype and anemia, and predicts survival
`P Gupta1, GA Niehans2, SC LeRoy1,3, K Gupta1, VA Morrison1, C Schultz1, DJ Knapp2 and RA Kratzke1
`
`Departments of 1Medicine and 2Pathology, and 3School of Public Health, Veterans Administration Medical Center and University of
`Minnesota Medical School, Minneapolis, Minnesota, USA
`
`Increased apoptosis in the bone marrow (BM) may contribute
`to the cytopenias that occur in myelodysplastic syndromes
`(MDS). The Fas receptor, Fas ligand (FasL) pathway is a major
`mechanism of apoptosis. Since hematopoietic progenitors can
`express the Fas receptor, they may be susceptible to apoptosis
`induced by FasL-expressing cells. We examined FasL
`expression in the BM of patients with MDS (n= 50), de novo
`acute myeloid leukemia (AML; n= 10), AML following prior MDS
`(n= 6), and normal controls (n= 6). Compared to controls, FasL
`expression was increased in MDS, and was highest in AML. In
`MDS, FasL expression was seen in myeloid blasts, erythro-
`blasts, maturing myeloid cells, megakaryocytes and dysplastic
`cells, whereas in AML, intense expression was seen in the
`blasts. FasL expression correlated with the FAB subtype
`groups of MDS, and also correlated directly with the percentage
`of abnormal metaphases on cytogenetic analysis. The FasL
`expressed in MDS BM inhibited the growth of clonogenic hema-
`topoietic progenitors. This inhibition could be blocked by a sol-
`uble recombinant FasFc protein. In MDS, FasL expression in
`the initial diagnostic BM was higher in patients who were more
`anemic, correlated directly with red cell transfusion require-
`ments over the subsequent course of the disease, and was pre-
`dictive of survival. These studies indicate that FasL expression
`in MDS is of prognostic significance, and suggest that pharma-
`cological blockade of
`the Fas-FasL pathway may be of
`clinical benefit.
`Keywords: myelodysplastic syndromes; acute myeloid leukemia;
`apoptosis; prognostic factors; Fas ligand
`
`Introduction
`
`The myelodysplastic syndromes (MDS) are a group of dis-
`orders characterized by cytopenias, often in the presence of
`adequate or increased numbers of cells in the bone marrow,
`and an increased propensity to develop acute leukemia.1
`Death occurs more frequently from the consequences of
`chronic cytopenias than from transformation to acute myeloid
`leukemia (AML).2–4
`Several recent reports indicate that a high proportion of
`bone marrow (BM) cells are undergoing apoptosis in patients
`with MDS,5–15 which may contribute to the cytopenias that
`occur in these disorders. A correlation between apoptosis and
`cytopenias in MDS has been observed in some studies but not
`in others.6–14 There are several possible reasons for variability
`in correlation between the extent of apoptosis detected in the
`marrow and the clinical features of MDS. It is possible that in
`situ end labeling may not detect a proportion of apoptotic
`cells that can be identified by other methods.16 Further, since
`apoptosis is a dynamic process, assays for apoptosis at one
`time point in complex tissues, such as the marrow, may not
`accurately reflect the total number of cells having undergone
`
`Correspondence: P Gupta, Hematology/Oncology Section 111E, VA
`Medical Center, One Veteran’s Drive, Minneapolis, MN 55417, USA;
`Fax: +1 612 725 2149
`Received 18 March 1998; accepted 8 September 1998
`
`apoptosis over time. Certain assays may fail to detect cells at
`early stages of apoptosis, whereas cells at
`later stages of
`apoptosis may already have been rapidly removed by reticulo-
`endothelial cells in the marrow. Moreover,
`the extent of
`apoptosis seen may depend on the proportions of residual nor-
`mal hematopoietic cells and of blasts in the sample, since the
`former but not the latter may be apoptotic. Thus, paradoxi-
`cally, a higher degree of apoptosis may be seen in marrows
`with low blast counts and a higher content of normal hemato-
`poietic cells.8
`One major mechanism for apoptosis in both hematopoietic
`and non-hematopoietic cells is the Fas pathway. Fas is a tumor
`necrosis factor-receptor (TNF-R) family cell surface receptor
`that induces apoptosis when ligated by Fas ligand (FasL), a
`cell surface TNF family molecule.17 Since normal hematopo-
`ietic cells express Fas receptor,18–24 they may be susceptible
`to apoptosis triggered by FasL-expressing cells. Mature periph-
`eral blood cells including neutrophils express high levels of
`Fas and are susceptible to Fas-mediated apoptosis,23 whereas
`normal progenitors express low levels of Fas and are not sus-
`ceptible to Fas-mediated apoptosis.25 However, on normal
`CD34+ hematopoietic cells, Fas expression is synergistically
`induced by TNF-a and interferon-g (IFN-g),
`resulting in
`increased susceptibility of clonogenic, as well as primitive
`progenitors to Fas-mediated apoptosis.18,19,26 In aplastic ane-
`mia, the increased expression of TNF-band IFN-gin the BM is
`accompanied by increased Fas on CD34+ cells and increased
`susceptibility of clonogenic progenitors to Fas-mediated kill-
`ing.27 In addition, TNF-a expression is increased in the BM of
`patients with MDS, and correlates with the degree of
`apoptosis.28
`A proportion of BM and peripheral blood cells are derived
`from normal progenitors in MDS,29,30 indicating that residual
`normal hematopoiesis contributes to the maintenance of per-
`ipheral blood cells in patients with MDS. The finding that
`CD34+ cells, differentiating hematopoietic cells and stromal
`cells undergo increased apoptosis6,9,10 suggests that cells
`derived from both the abnormal clone and those derived from
`normal progenitors may be apoptotic in MDS. Recent studies
`indicate that in MDS, the Fas pathway is functional and cap-
`able of inducing apoptosis in clonogenic progenitors when
`expressed on CD34+, CD33+
`activated.
`Fas
`is
`and
`glycophorin+ cells in MDS, and both erythroid and granulo-
`cyte–macrophage colony-forming cells are susceptible to
`apoptosis induced by an agonist anti-Fas antibody.7,8
`A variety of tumor cell types express FasL, the ligand for the
`Fas receptor.31–35 Malignant cells appear to possess mech-
`anisms to protect themselves against FasL-mediated apoptosis,
`while maintaining the ability to induce apoptosis in normal
`cells.6,35,36 Recent reports indicate that FasL expression is
`increased in MDS.15,37 We hypothesized that the expression
`of FasL (1) is increased in the bone marrow of patients with
`MDS; (2) contributes to the cytopenias observed in these dis-
`orders by its effects on hematopoietic progenitors; and (3) cor-
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 1
`
`

`

`relates with cytopenia and prognosis. We demonstrate that
`FasL expression is significantly increased in MDS and AML,
`that FasL expressed by MDS marrow cells directly inhibits the
`growth of hematopoietic progenitors in vitro, and that FasL
`expression correlates with anemia and predicts transfusion
`requirements and survival in patients with MDS.
`
`Methods
`
`Patients
`
`A cohort of 50 consecutive patients diagnosed with MDS and
`10 with AML at the Minneapolis VA Medical Center from
`1992–1997 were identified by retrospective review of BM
`pathology records. Clinical information, including features at
`presentation, clinical course and outcome, was obtained by
`chart review. The demographic characteristics of the patients
`and clinical features at presentation are shown in Table 1. All
`the patients were males, reflecting the patient population at
`the Minneapolis VA Medical Center. The French–American–
`British (FAB) classification of the 50 patients with MDS was:
`RA (n = 4), RA with ringed sideroblasts (RARS; n = 6), RA with
`excess blasts (RAEB; n = 10), RAEB in transformation (RAEBT;
`n = 7), chronic myelomonocytic leukemia (CMML; n = 13),
`and unclassified MDS (n = 10). Forty patients required blood
`product transfusions: RBC (n = 19), platelets (n = 2), or both
`products (n = 19). To normalize RBC transfusion requirements
`over time, the total number of units of RBC transfused between
`the diagnosis of MDS and either death or transformation to
`AML were converted to the average number of units trans-
`fused per 6 months, as follows: Number of units transfused
`between diagnosis of MDS and death or transformation to
`AML/Time in months from diagnosis of MDS to death or trans-
`formation to AML · 6. Patients included in this study were
`treated with pyridoxine (n = 7), androgens (n = 1), prednisone
`(n = 1), hydroxyurea (n = 2), 6-thioguanine + cytosine arabino-
`side (n = 1), 5-azacytidine (n = 2), erythropoietin (EPO; n = 2),
`and folic acid (n = 4). The remaining patients received only
`
`Table 1
`
`Clinical features of 50 patients with MDS
`
`Age (years), median (range)
`Hemoglobin (g/dl), median
`(range)
`WBC (per mm3), median
`(range)
`Platelet count (per mm3), median (range)
`Absolute neutrophil count (per mm3), median
`(range)
`Absolute monocyte count (per mm3), median
`(range)
`Blasts in peripheral blood (%), median (range)
`Reticulocytes (%), median (range)
`LDH (number of patients)
`Elevated
`Normal
`BM cellularity (number of patients):
`hypercellular; normocellular; hypocellular
`BM blasts (%), median (range)
`Bournemouth prognostic score (number of
`patients)
`0; 1; 2; 3; 4
`Cytogenetic analysis
`Abnormal
`Normal
`
`73 (26–87)
`10.2
`(7.5–13.8)
`4.3
`(1.1–112.6)
`120 (6–693)
`2.5
`(0.3–44.3)
`0.4
`(0.0–22.6)
`0.0 (0.0–9.0)
`2.0 (0.1–8.8)
`
`13
`19
`
`36; 7; 7
`2.6 (0.2–24)
`11; 11; 11;
`11; 6
`
`15
`23
`
`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`supportive treatment. One patient underwent allogeneic bone
`marrow transplantation. Five of 50 patients (10%) have trans-
`formed to AML to date. Twenty-nine of 50 patients (58%) have
`died,
`largely of complications
`related directly to MDS
`(infection in 12, hemorrhage in 7, AML in 4).
`
`45
`
`Immunohistochemical staining for FasL
`
`FasL expression was determined on BM samples that had been
`obtained at diagnosis of MDS, prior to treatment, by modifi-
`cation of a previous method.31 Sections (5 mm thick) were cut
`from bone marrow aspirate clots embedded in paraffin in B-
`5 fixative and mounted on poly-l-lysine-coated slides. The
`sections were deparaffinized in xylene, rehydrated through
`graded alcohols and demercurized in alcoholic iodine.
`Endogenous peroxidase activity was quenched by immersing
`the slides in 1.5% hydrogen peroxide for 10 min at room tem-
`perature, followed by rehydration through three changes of
`distilled water. To induce heat-induced epitope retrieval,
`slides were immersed in 10 mm citrate buffer (pH 6.0) in a
`pressure cooker which was heated to boiling at 12 psi for
`10 min and then cooled to room temperature over 20 min.
`Nonspecific binding was blocked by incubating sections with
`15 mg/ml horse serum for 15 min at room temperature. The
`sections were then incubated for 1 h with 0.6 mg/ml mouse
`MAb against residues 116–277 of human FasL (MAb clone
`33; Transduction Laboratories, Lexington, KY, USA). Negative
`control sections were incubated with nonimmune mouse
`serum or ascites fluid diluted 1:400 in PBS. Following incu-
`bation for 30 min with horse anti-mouse secondary antiserum,
`sections were immunolabeled using an avidin-biotin-immuno-
`peroxidase method (Vector Laboratories, Burlingame, CA,
`USA). The chromogen was a solution of 0.5 mg/ml 3,39-diami-
`nobenzidine tetrahydrochloride (DAB) and 0.009% hydrogen
`peroxide. Color development was terminated after 5 min by
`immersion in water. To differentiate iron deposits in hemato-
`poietic cells, sections were immersed in equal parts of 5%
`potassium ferrocyanide and 5% HCl for 5 min. Sections were
`finally counterstained with hematoxylin, dehydrated and
`coverslipped with permount. FasL expressing cells were ident-
`ified by granular intracytoplasmic staining. In each section,
`1000 cells were scored for FasL expression by a pathologist
`(GAN) who was blinded to the FAB subtype and clinical infor-
`mation regarding the patients.
`Additional sections from some of the same BM samples
`were immunostained using another anti-FasL primary anti-
`body (Q20, a rabbit polyclonal directed against N-terminal
`amino acid residues 2–21; Santa Cruz Biotechnology, Santa
`Cruz, CA, USA). Heat-induced epitope retrieval was induced
`the sections in 10 mm citrate buffer for
`by incubation of
`20 min at 100(cid:176) C, at atmospheric pressure. Sections were incu-
`bated with 1 mg/ml of the primary antibody for 45 min, and
`developed with a multi-link secondary antibody using the
`Vector Elite kit (Vector). To further confirm the specificity of
`the staining for FasL, blocking experiments were performed
`by pre-incubation of the sections with a four-fold excess of a
`blocking peptide (amino acid residues 2–21 of FasL).
`The immunohistochemistry technique described can be
`used on BM clot specimens which have been fixed in either
`B5 or formalin, but not in Zenker’s or Bouin’s fixatives. Also,
`the process used for decalcification of BM trephine specimens
`renders FasL undetectable by this method.
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 2
`
`

`

`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`46
`
`Immunoblotting for FasL
`
`BM was aspirated in preservative-free heparin from the pos-
`terior iliac crest of patients with MDS or healthy volunteers
`after obtaining written informed consent. The study was
`approved by the VA Medical Center Human Subjects Sub-
`committee. The BM aspirates were immediately subjected to
`Ficoll–Hypaque (sg 1.077) centrifugation. BM mononuclear
`cells (BMMNC) at the interface were aspirated, washed three
`times, counted and resuspended in Iscove’s modified Dulbec-
`co’s medium (IMDM) + 20% fetal calf serum + 20% dimethyl-
`sulfoxide, and stored in liquid nitrogen following controlled
`rate
`freezing, until used in co-culture with normal
`CD34+/HLA-DR+ cells or for immunoblotting. For immuno-
`blotting, the frozen cells were rapidly warmed to 37(cid:176) C, diluted
`in warm IMDM,
`and immunoblotting performed as
`described.31 Briefly, BMMNC were pelleted and lysed in 1 ml
`of 50 mm Tris-HCl (pH 7.5), 250 mm NaCl, 5 mm EDTA, 0.1%
`NP40, 50 mm NaF and 1 mm phenylmethylsulfonylfluoride.
`Approximately 30 mg of total protein from each sample was
`electrophoresed on 12.5% SDS-PAGE gel, and electroblotted
`on to a nitrocellulose membrane. The membranes were
`blocked using 5% dry milk + 1% BSA in PBS, incubated with
`1:1000
`dilution
`of murine
`anti-human
`FasL MAb
`(Transduction Laboratories) at 4(cid:176) C overnight and then with a
`rabbit anti-murine IgG antibody (PharMingen, San Diego, CA,
`USA). Protein bands were visualized by enhanced chemilumi-
`nescence as recommended by the manufacturer (Amersham,
`Arlington Heights, IL, USA).
`In parallel, membranes were incubated with 1.5 mg/ml of a
`different primary antibody (anti-FasL IgG MAb clone G-247-
`4; PharMingen) for 1 h at room temperature. Following incu-
`bation with a secondary antibody, FasL was detected using an
`enhanced chemiluminescence kit (Amersham). A pGEX2TK-
`FasL fusion protein (size approximately 60 kDa), possessing
`the full-length recombinant human FasL protein with a GST
`leader was run as a positive control, to confirm the capability
`of the antibody to recognize human FasL.
`
`Effect of FasL expressed by MDS BM cells on growth
`of clonogenic progenitors
`
`CD34+/HLA-DR+ (DR+) cells were purified from bone marrow
`aspirates taken from normal adult donors, by sequential Fic-
`oll–Hypaque (sg 1.077) centrifugation, Ceprate-LC (CellPro,
`Bothell, WA, USA) CD34-biotin immunoadsorption column
`selection and flow cytometry, as previously described.38 To
`simulate the effect of increased levels of TNFa found in the
`marrow in MDS,28 the DR+ cells were suspended at a concen-
`tration of 100 000/ml
`in long-term bone marrow culture
`(LTBMC) medium (12.5% FCS, 12.5% horse serum, 2 mm l-
`glutamine, 1000 U/ml penicillin, 100 U/ml streptomycin and
`10- 6 m hydrocortisone) supplemented with 100 ng/ml TNFa
`(Sigma) and 10 ng/ml interleukin (IL)-3 (R&D Systems, Minne-
`apolis, MN, USA) and incubated at 37(cid:176) C in 5% CO2 atmos-
`phere for 48 h.
`For co-cultures of DR+ cells with cell
`lysates of MDS
`BMMNC, the frozen cells were rapidly warmed to 37(cid:176) C and
`diluted in warm IMDM. The cells were pelleted gently, resus-
`pended in IMDM at a concentration of 25 · 106/ml and lysed
`by snap-freezing and thawing three times. Lysates from
`2 · 105 or 2 · 106 MDS BMMNC were mixed with 2000 DR+
`cells in IMDM + 10% FCS + 10 ng/ml TNFa + 10 ng/ml
`interleukin-3 (IL-3) and incubated at 37(cid:176) C in 5% CO2 for 18 h.
`
`In some cultures, the MDS cell lysates were incubated with
`either 10 mg/ml recombinant FasFc protein (R&D Systems),
`50 mg/ml neutralizing anti-TNFa antibody (R&D Systems), or
`both, for 1 h at room temperature, to block the FasL39 and
`TNFapresent, prior to addition of the DR+ cells to the lysates.
`For control cultures, DR+ cells were incubated with the cyto-
`kines in the absence of cell lysates. Following the incubation,
`cells were washed once with IMDM + 20% FCS to remove
`cytokines, cell debris and FasFc, and plated in semisolid
`methylcellulose cultures supplemented with 3 U/ml EPO and
`5 ng/ml IL-3. In cultures examining the effect of FasFc protein
`and anti-TNFa antibody, additional FasFc (0.5 mg/ml) and
`anti-TNFa antibody (2.5 mg/ml) were added directly to the
`methylcellulose medium. Colonies were scored after 14
`days.40
`
`Statistics
`
`Data were entered on and analyzed using an SPSS software
`package. Mean differences between groups were compared
`using either the t-test or ANOVA. Correlations of interval scale
`variables were analyzed using linear regression. Overall sur-
`vival by FasL expression (FasL <12% and FasL .12%) was
`analyzed by the Kaplan–Meier method41 and compared using
`the log-rank statistic. In addition, the correlation between %
`FasL expression and overall survival was analyzed using Cox
`regression. Results were expressed as the mean –
`s.e.m.
`
`Results
`
`FasL expression is increased in MDS and AML
`
`We examined by immunohistochemistry (Figures 1 and 2) the
`expression of FasL in the BM of patients with MDS, AML
`developing in patients with MDS (secondary AML), de novo
`AML and age- and sex-matched normal controls. Immunohis-
`tochemistry was performed using two separate anti-FasL anti-
`bodies, directed against a C-terminal epitope (MAb clone 33;
`Figure 1) or an N-terminal epitope of FasL (polyclonal anti-
`body Q20; Figure 2). There was a high degree of correlation
`(r = 0.97) between the percentage of FasL positive cells
`(normal controls, MDS and AML) detected by the two anti-
`bodies, MAb clone 33 and polyclonal Q20 (data not shown).
`In the BM of normal controls (n = 6), a uniformly low level of
`FasL expression was detected (6 – 1% positive cells; range 3–
`9%) and was mostly confined to lymphocytes (Figures 1a and
`3a). In contrast, a wide range of FasL expression was seen in
`the BM of patients with MDS (n = 50; 17 – 2% positive cells;
`range 2–51%; Figures 1b, 1c, 2a, 2c and 3a), which was
`higher than the controls (p = 0.02). FasL expression in MDS
`was seen in erythroblasts, myeloblasts, megakaryocytes, mat-
`uring myeloid precursors and dysplastic cells, and was not
`confined to lymphocytes (Figures 1 and 2). Mature neutrophils
`and erythroid normoblasts did not express detectable FasL.
`The staining in both MDS and AML BM cells by the Q20 anti-
`body was almost completely blocked by pre-incubation of the
`sections with a peptide (amino acids 2–12 of FasL) comprising
`the epitope for the Q20 antibody (Figures 2b, d and f).
`FasL expression in MDS was further confirmed by immuno-
`blot analysis of BMMNC. A protein of the expected size for
`FasL (37 kDa) was recognized by a murine anti-human FasL
`antibody (MAb clone 33)
`in lysates from MDS BMMNC
`(Figure 1f). The level of FasL protein expression was signifi-
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 3
`
`

`

`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`47
`
`(a–e) Immunohistochemical detection of FasL expression in BM clot sections. (a) Normal control with occasional FasL+ lymphocytes
`Figure 1
`(arrow); (b) MDS, unclassified, with 31% FasL+ cells. FasL in blast cells (arrow) and in a megakaryocyte (arrowhead); (c) MDS, RAEB, with 51%
`FasL+ cells; (d) AML M5 subtype, with intensely FasL+ blasts and negative erythroblasts (arrow); (e) AML M5, negative control using preimmune
`serum in place of the anti-human FasL MAb (Transduction Laboratories MAb clone 33). Original magnification · 800. (f) Immunoblotting analysis
`of FasL expression in MDS. FasL expression was examined in cell lysates (30 mg total protein per lane) of BMMNC from MDS patients and from
`normal controls. The lung cancer cell line H2009 which expresses FasL31 was used as a positive control. A 37 kDa protein was detected by
`an anti-human FasL MAb (Transduction Laboratories MAb clone 33).
`
`cantly higher in MDS than in normal control BMMNC, using
`equivalent quantities of total protein from cell lysates. FasL
`(37 kDa) was also detected in MDS BMMNC by immunoblot-
`ting with another primary antibody (anti-FasL IgG MAb clone
`G-247-4, PharMingen), which also recognized the 60 kDa
`recombinant FasL fusion protein (Figure 2g). Taken together,
`these results indicate that FasL is expressed in BM cells in both
`MDS and AML.
`In the BM of patients with de novo AML (n = 10), an
`extremely high proportion of cells showed FasL expression
`(56 – 9%; range 10–97%; Figures 1d, 2e and 3a) which was
`higher than that found in controls (P = 0.001) and patients
`with MDS (P , 0.001). FasL expression in AML was largely
`seen in the blast cells, which showed intense positivity for the
`protein (Figures 1d and 2e). In all three patients studied, FasL
`
`expression increased at least three-fold upon transformation
`of MDS to secondary AML (Figure 3b). In three additional
`patients with secondary AML, FasL expression ranged from 56
`to 86% (initial diagnostic MDS marrows were not available
`for FasL staining in these patients). As in de novo AML,
`expression of FasL was seen largely in the blasts in secondary
`AML. The overall FasL expression in secondary AML (n = 6;
`48 – 13%; range, 9–86%) was comparable to de novo AML.
`
`FasL expression correlates with FAB subtypes of MDS
`
`A significant correlation was observed between FasL
`expression and FAB subtype groups of MDS:RA/RARS (9 – 3%
`FasL positive cells), CMML (14 – 3%), RAEB/RAEBT (20 – 3%)
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 4
`
`

`

`48
`
`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`Figure 2
`(a–f) Immunohistochemical detection of FasL expression in BM clot sections. Sections were incubated with Q20, a rabbit polyclonal
`directed against N-terminal amino acid residues 2–21 (Santa Cruz Biotechnology). (a) MDS, unclassified (c) MDS, RAEB-T (e) AML M5 subtype.
`Photomicrographs b, d and f are sections from the same marrows, which were pre-incubated with a blocking peptide (amino acid residues 2–
`21 of FasL) before staining with the antibody, demonstrating that the staining is blocked by the FasL peptide containing the epitope for the
`antibody. Original magnification · 320 (a,b,e,f) and · 800 (c,d). (g) Immunoblotting of FasL expression in MDS. FasL expression was examined
`in cell lysates (30 mg total protein per lane) of BMMNC from three MDS patients, using a third primary antibody (anti-FasL IgM MAb clone G-
`247-4, PharMingen). A pGEX2T/K-FasL fusion protein (60 kDa), with the full-length recombinant human FasL protein and a GST leader was
`used as a positive control. A 37 kDa protein corresponding to FasL was detected in all three MDS samples.
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 5
`
`

`

`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`49
`
`Figure 3
`(a) FasL expression in MDS and AML. FasL expression
`was determined as described in Methods in the BM of patients with
`morphologically normal marrows (normal controls) and in the initial
`diagnostic BM of patients with MDS, de novo AML and secondary
`AML (AML developing in patients with known prior MDS). Compari-
`son between normal controls and other categories: *P = 0.02, §P =
`0.009, ¶P = 0.001; comparison between MDS and other categories:
`**P , 0.001. (b) FasL expression following transformation of MDS to
`secondary AML. FasL expression in three patients at diagnosis of MDS
`(RA = 1, RARS = 1, RAEBT = 1) and at transformation to secondary
`AML.
`and unclassified MDS (22 – 4%); P = 0.028 (Figure 4a). The
`high expression of FasL in unclassified MDS was somewhat
`surprising. However, patients included in the unclassified cat-
`egory had significant dysplasia in all three lineages, even
`though the proportion of blasts in the BM was ,5%. The pres-
`ence of dysmegakaryopoiesis or dysgranulopoiesis has pre-
`viously been shown to adversely impact survival in MDS.42
`The majority of patients with unclassified MDS experienced a
`short survival in our study. BM cytogenetic analysis was avail-
`able for 38 patients. In the patients who had detectable abnor-
`malities, FasL expression correlated directly with the percent-
`age of abnormal metaphases on cytogenetic analysis
`(P = 0.022; Figure 4b). This suggests that FasL expression may
`correlate directly with the proportion of the marrow replaced
`by the progeny of the abnormal clone in MDS.
`
`The FasL expressed by MDS cells inhibits the growth
`of clonogenic progenitors
`
`report,15 we observed FasL
`In agreement with a recent
`expression in myeloid, erythroid and mononuclear cells in
`
`Figure 4
`(a) Correlation of FasL expression with subtype groups of
`MDS. Patients were combined into four groups on the basis of FAB
`categories and prognosis: RA + RARS, CMML, RAEB + RAEBT and
`unclassified (patients who could not be assigned to a specific FAB
`category). Comparison between RAEB/RAEBT or unclassified and
`RA/RARS: *P = 0.028. (b) Correlation of FasL expression with the pro-
`portion of abnormal metaphases. FasL expression in the BM was cor-
`related with the percentage of abnormal metaphases on cytogenetic
`analysis performed on the same marrow aspirate obtained at diagnosis
`of MDS. n = 13, r = 0.62, P = 0.022.
`
`MDS. We therefore examined if the FasL expressed by cells
`in MDS BM is functionally active in inhibiting the growth of
`hematopoietic progenitors. A significant and dose-dependent
`inhibition of the growth of clonogenic progenitors was seen
`in the presence of MDS BMMNC lysates (Figure 5a). The
`growth of erythroid BFU-E as well as myeloid CFU-GM pro-
`genitors was inhibited, as reflected in inhibition of the total
`numbers of colonies formed. To examine if this inhibition was
`mediated by TNFa, FasL or both, we evaluated the effect of
`addition of neutralizing anti-TNFa antibody, soluble FasFc
`protein (which specifically binds and blocks FasL)31,39 or both,
`in presence or absence of MDS cell lysates (Figure 5b). In the
`absence of MDS cell lysates, neither the anti-TNFa antibody
`nor FasFc altered colony formation significantly. However, the
`inhibition of progenitor growth by MDS cell lysates was larg-
`ely prevented by pre-incubation of the MDS cell lysates with
`FasFc, but not anti-TNFa. The simultaneous addition of both
`FasFc and anti-TNFa did not
`further augment progenitor
`growth over that seen with FasFc alone. Using DR+ cells from
`the same donors and cell lysates from the same MDS patients,
`we observed that the frequency of apoptotic cells in the DR+
`population (detected by the TUNEL assay or by annexin V
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 6
`
`

`

`50
`
`Fas ligand expression in myelodysplastic syndromes
`P Gupta etal
`
`Figure 5
`Effect of FasL expressed by MDS BM cells on the growth
`of normal clonogenic progenitors. (a) Dose-dependent inhibition of
`normal progenitor growth in presence of MDS cell lysates. Lysates of
`BM mononuclear cells from four separate MDS patients were each
`cultured with DR+ cells from three separate BM donors, as described
`in Methods. Two thousand DR+ cells were cultured with lysates from
`2 · 105 MDS cells (effector:target (E:T) ratio, 100:1) or 2 · 106 cells
`(E:T ratio, 1000:1). As controls, DR+ cells were cultured in the absence
`of MDS cell lysates (no MDS cells). Comparison between control cul-
`tures and other conditions: *P , 0.001, §P = 0.02. (b) Effect of anti-
`TNFa antibody and FasFc protein on MDS cell
`lysate-mediated
`growth inhibition. DR+ cells (2000 cells each) were cultured in pres-
`ence or absence of MDS cell lysates (2 · 105 cells), with or without
`a neutralizing anti-TNFa antibody or soluble FasFc protein. As con-
`trols, DR+ cells were cultured in the absence of MDS cell lysates, anti-
`TNFa or FasFc. n = 3. Comparison between controls and DR+ with
`MDS lysates: *P = 0.027, §P = 0.006. Comparison between DR+ with
`MDS lysates and other conditions: ¶P = 0.040, **P , 0.02, §§P = 0.02,
`§§P , 0.05.
`
`staining) increased from 3% in controls to 10% in the pres-
`ence of the MDS lysate, but was reduced to 6% by pre-
`incubation of the lysate with FasFc (data not shown).
`
`Expression of FasL correlates with anemia in MDS
`
`We next examined if FasL expression in the initial diagnostic
`BM of the 50 patients with MDS correlated with features of
`the clinical presentation and the subsequent disease course.
`FasL expression in the initial diagnostic BM was higher in
`patients with an initial hemoglobin ,10 g/dl vs hemoglobin
`>10 g/dl (20 – 3% vs 13 – 2%, respectively; P = 0.03; Figure
`6a). Also,
`there was a direct correlation between FasL
`expression in the initial diagnostic BM and the average num-
`ber of units of pRBC transfusions required per 6 months over
`the subsequent disease course (P = 0.039). Thus, patients with
`high FasL expression required more pRBC transfusions than
`patients with low FasL expression. There was a trend towards
`higher FasL expression in the BM of patients with more severe
`
`Figure 6
`(a) Correlation of FasL expression with hemoglobin. FasL
`expression at diagnosis was correlated with hemoglobin at diagnosis.
`Comparison between Hb ,10 g/dl and Hb >10 g/dl: P = 0.031.
`(b) Overall survival of patients expressing low and high levels of FasL
`at diagnosis. The overall survival stratified by FasL expression (FasL
`<12% and FasL .12%) from diagnosis of MDS (n = 50) was estimated
`by the Kaplan–Meier method41 and compared using the log-rank test.
`Comparison between FasL <12% and FasL .12%: P = 0.035.
`
`neutropenia, though this did not achieve statistical signifi-
`cance (P = 0.06). No correlation was detected between FasL
`expression and age, platelet count, need for platelet trans-
`fusions, lactate dehydrogenase level, marrow cellularity, per-
`centage of blasts, type of therapy or response to therapy.
`
`FasL expression predicts survival in MDS
`
`FasL expression was predictive of survival in the 50 MDS
`patients studied (median survival 28 months for FasL <12%
`vs 13 months for FasL .12%; P = 0.035; Figure 6b). The 2-
`year survival was 60 and 21%, respectively. FasL expression
`also correlated directly with survival of these 50 patients as a
`single group (P = 0.04 by Cox regression analysis). A number
`of studies have confirmed the prognostic significance of the
`Bournemouth scoring system, in which a score from 0–4 is
`assigned based on hemoglobin, neutrophil count, platelet
`count and blast percentage at diagnosis.3 We observed a sig-
`nificant correlation between Bournemouth score and survival
`in our patients (P = 0.019 by Cox regression analysis). The
`expression of FasL directly correlated with the Bournemouth
`prognostic score (P = 0.02). That FasL expression correlates
`
`Dr. Reddy’s Laboratories, Inc. v. Celgene Corp.
`IPR2018-01504
`Exhibit 2015, Page 7
`
`

`

`51
`
`with an independent, validated prognostic scoring system
`further suggests that it may have potential value in predicting
`the survival of patients with MDS.
`
`Discussion
`
`Recent reports demonstrate that the response of cytopenias to
`EPO and granulocyte colony-stimulating factor (G-CSF) ther-
`apy43 or to inhibition of TNF-a is associated with reduction
`in the proportion of apoptotic cells in the BM.10,22,44 These
`studies indicate that apoptosis is a clinically important
`phenomenon in MDS. Our studies demonstrate a correlation
`between the expression of FasL, the molecule which triggers
`Fas-mediated apoptosis, and the clinical features and prog-
`nosis of MDS. We further show that FasL expressed in MDS
`BM directly inhibits the growth of clonogenic hematopoietic
`progenitors. A recent report15 demonstrating that the apoptotic
`(TUNEL positive) cells in MDS BM express both Fas receptor
`and Fas ligand, further supports our results.
`We observed that apoptosis of FACS sorted DR+ cells was
`increased in