American Journal of Hematology 60:36–47 (1999)
`
`Evidence for Involvement of Tumor Necrosis Factor-a in
`Apoptotic Death of Bone Marrow Cells in
`Myelodysplastic Syndromes
`
`Suneel D. Mundle,* Ambereen Ali, Jonathan D. Cartlidge, Samina Reza, Sairah Alvi,
`Margaret M. Showel, B. Yifwayimare Mativi, Vilasini T. Shetty, Parameswaran Venugopal,
`Stephanie A. Gregory, and Azra Raza
`Rush Cancer Institute, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, Illinois
`
`We previously reported excessive apoptosis and high levels of tumor necrosis factor-
`alpha (TNF-a) in the bone marrows of patients with myelodysplastic syndromes (MDS),
`using histochemical techniques. The present studies provide further circumstantial evi-
`dence for the involvement of TNF-a in apoptotic death of the marrow cells in MDS. Using
`our newly developed in situ double-labeling technique that sequentially employs DNA
`polymerase (DNA Pol) followed by terminal deoxynucleotidyl transferase (TdT) to label
`cells undergoing apoptosis, we have characterized DNA fragmentation patterns during
`spontaneous apoptosis in MDS bone marrow and in HL60 cells treated with TNF-a or
`etoposide (VP16). Clear DNA laddering detected by gel electrophoresis in MDS samples
`confirmed the unique length of apoptotic DNA fragments (180–200 bp). Surprisingly,
`however, phenotypically heterogeneous population of MDS cells as well as the homog-
`enous population of HL60 cells showed three distinct labeling patterns after double
`labeling—only DNA-Pol reaction, only TdT reaction, and a combined DNA Pol + TdT
`reaction, albeit in different cohorts of cells. Each labeling pattern was found at all mor-
`phological stages of apoptosis. MDS mononuclear cells, during spontaneous apoptosis
`in 4 hr cultures, showed highest increase in double-labeled cells (DNA Pol + TdT reac-
`tion). Interestingly, this was paralleled by TNF-a–induced apoptosis in HL60 cells. In
`contrast, VP16 treatment of HL60 cells led to increased apoptosis in cells showing only
`TdT reaction. The double-labeling technique was applied to normal bone marrow and
`peripheral blood mononuclear cells after treatment with known endonucleases that spe-
`cifically cause 3* recessed (BamHI), 5* recessed (PstI), or blunt ended (DraI) double-
`stranded DNA breaks. It was found that the DNA-Pol reaction in MDS and HL60 cells
`corresponds to 3* recessed DNA fragments, the TdT reaction to 5* recessed and/or blunt
`ended fragments, and a combined ‘‘DNA Pol + TdT reaction’’ corresponds to a copres-
`ence of 3* recessed with 5* recessed and/or blunt ended fragments. Clearly, therefore,
`apoptotic DNA fragments, in spite of a unique length, may have differently staggered
`ends that could be cell (or tissue) specific and be selectively triggered by different in-
`ducers of apoptosis. The presence of TNF-a–inducible apoptotic DNA fragmentation
`pattern in MDS supports its involvement in these disorders and suggests that anti–TNF-a
`(or anticytokine) therapy may be of special benefit to MDS patients, where no definitive
`treatment is yet available. Am. J. Hematol. 60:36–47, 1999.
`© 1999 Wiley-Liss, Inc.
`
`Key words: human disorders; myelodysplastic syndromes (MDS); HL60cells; apoptosis;
`DNA fragmentation; in situ labeling; tumor necrosis factor-a (TNF-a); etoposide (VP16);
`hematopoietic disorders
`
`A. Ali is a Research Fellow of the National Cancer Institute of Canada
`supported with funds provided by the Terry Fox Run.
`
`Luke’s Medical Center, 2242 West Harrison Street, Suite 108, Chi-
`cago, IL 60612. E-mail: smundle@rush.edu
`
`*Correspondence to: Suneel D. Mundle, Ph.D., Assistant Professor,
`Rush Medical College, Rush Cancer Institute, Rush-Presbyterian-St.
`
`Received for publication 8 December 1997; Accepted 2 September
`1998
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 1
`
`

`

`INTRODUCTION
`
`In the past we showed that hematopoietic as well as
`stromal cells in the bone marrows (BMs) of patients with
`myelodysplastic syndromes (MDS) demonstrate exces-
`sive apoptotic death, thus forming the basis for ineffec-
`tive hematopoiesis observed in these disorders [1–3].
`Others have confirmed our observation of high intramed-
`ullary apoptosis in MDS [4,5]. Such a widespread inci-
`dence of apoptosis affecting all types of bone marrow
`cells was indicative of a possible involvement of cyto-
`kine(s) with a broad range of target cells. Using immu-
`nohistochemistry, we indeed detected high levels of tu-
`mor necrosis factor-alpha (TNF-a) correlating signifi-
`cantly with the levels of apoptosis detected in situ in the
`same BM biopsies of MDS patients [3]. In our attempts
`to further understand the association between TNF-a and
`apoptosis in MDS, in the present studies we examined
`the patterns of DNA fragmentation during spontaneous
`apoptosis in MDS and in TNF-a–induced apoptosis in
`HL60 cells.
`Internucleosomal DNA fragmentation constitutes one
`of the most salient features of apoptotic cell death. The
`four endonucleases most prominently implicated in
`apoptotic DNA fragmentation are Ca++ Mg++ dependent,
`Mg++ dependent, Mn++ dependent, and acidic endonucle-
`ase. These endonucleases have been shown to maintain
`the property of internucleosomal cleavage [6–8]. No spe-
`cific differences with respect to substrate requirement or
`products formed have been noted among different apop-
`totic endonucleases. The DNA fragments formed by such
`endonucleolytic activity that typically appear to have
`58-P and 38-OH terminals [9–11], could be detected using
`two specific enzymatic reactions that label the ends of
`DNA fragments. One of these techniques uses DNA
`Polymerase (DNA Pol) or Klenow fragment of DNA Pol
`[10] whereas the other uses terminal deoxynucleotidyl
`transferase (TdT) for this purpose [9]. We have exten-
`sively used these enzymatic reactions to detect apoptosis
`in situ in a variety of clinical samples [1–3,12,13]. In one
`of these studies, we compared the labeling of apoptosis
`by the two enzymatic reactions performed separately in
`serial sections of different types of solid tumors. It was
`found that some tissues like breast tumors or primary
`brain tumors did not show labeling with DNA Pol, but
`showed positivity with TdT labeling [12,13]. On the
`other hand, tissues like nonHodgkin’s lymphoma or head
`and neck squamous cell carcinoma showed comparable
`labeling by the two enzymes in serial biopsy sections
`[12]. Gold et al. [14] in 1994 reported that whereas DNA
`Pol preferentially labels necrosis, TdT was more specific
`for apoptosis. Contrastingly, we found that both methods
`label apoptosis as well as necrosis except that necrosis is
`labeled with an extremely low intensity by both methods
`[12,15].
`
`Involvement of TNF-a in Myelodysplasia
`
`37
`
`To explain the tissue-specific differences in labeling
`with DNA Pol and TdT noted in our studies, we hypoth-
`esized that the apoptotic DNA fragments may have a
`unique length of integral multiples of 180–200 bp, but
`have differently staggered ends—38 recessed, 58 re-
`cessed, or blunt—perhaps reflecting the specificity of the
`endogenous endonuclease(s) involved. To test this pos-
`tulate we developed an in situ double-labeling technique
`in which the DNA-Pol reaction is performed first, fol-
`lowed by the TdT reaction. Using this technique, as de-
`scribed in the present paper, we have now been able to
`confirm our hypothesis and characterize and compare the
`ends of apoptotic DNA fragments in spontaneous apop-
`tosis in MDS with that induced by TNF-a in HL60 cells.
`The findings of these studies reported in the present pa-
`per highlighting their therapeutic implications suggest a
`novel approach to the treatment of MDS.
`
`MATERIALS AND METHODS
`
`Patients
`
`Sixteen bone marrow aspirate specimens from 15 pa-
`tients with a confirmed diagnosis of MDS [diagnosed
`according to the French-American-British classification
`as refractory anemia (RA)-7, RA with ringed sideroblasts
`(RARS)-3, RA with excess of blasts (RAEB)-3, RA with
`excess of blasts in transformation (RAEBt)-1, and
`Chronic myelomonocytic leukemia (CMMoL)-1] were
`studied for the incidence of spontaneous apoptosis. One
`MDS patient (RARS) was studied on two occasions. Six
`normal bone marrows from healthy donors were studied
`for comparison. The protocols, MDS 90 02 and MDS 95
`01, under which clinical specimens were obtained, were
`approved by the local Institutional Review Board (IRB)
`and informed consent was obtained from the donors.
`Bone marrow aspirates were subjected to Ficoll-Hypaque
`density gradient centrifugation to separate mononuclear
`cells. After confirming the viability by trypan blue dye
`exclusion test, cells were suspended (1 × 106 cells/ml) in
`RPMI-1640 medium (GIBCO-BRL Life Technologies
`Inc., Grand Island, NY) supplemented with 10% fetal
`bovine serum (FBS) (GIBCO-BRL), 100 U/ml penicillin,
`100 mg/ml streptomycin, and 2 mM/l glutamine (com-
`plete medium). One aliquot was used for 0 hr tests and
`the other was incubated for 4 hr at 37°C in the presence
`of 5% CO2. At both time points, cells were washed and
`divided into two aliquots. One was fixed on alcian blue
`coated coverslips using 4% buffered paraformaldehyde
`(pH 7.1) overnight at 4°C and stored in 70% ethanol until
`in situ tests for detection of apoptosis were performed.
`The other aliquot was used to extract DNA following cell
`lysis in guanidine isothiocyanate (GITC) as described
`previously [12].
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 2
`
`

`

`38
`
`Mundle et al.
`
`HL60 Cell Cultures
`HL60 cells were maintained in RPMI-1640 complete
`medium supplemented with 20% FBS. Freshly harvested
`cells, after confirming viability by trypan blue dye ex-
`clusion test, were resuspended (1 × 106 cells/ml) in com-
`plete medium and treated with human recombinant
`TNF-a (0.01 ng/ml; Promega Inc., Madison, WI) for 8 hr
`or with VP16 (etoposide—35 mM/l; Bristol Laboratories,
`Princeton, NJ) for 4 hr. As both these agents were dis-
`solved in RPMI 1640, cells suspended in plain complete
`medium and incubated for a maximum interval, i.e., 8 hr
`served as controls. The dose of TNF-a was chosen based
`on the ED50 dose (0.016 ng/ml) suggested by the manu-
`facturer and according to our initial dose response ex-
`periments, which showed a linear apoptotic response at
`0.01 ng/ml until 8 hr (data not shown). The concentration
`and incubation time of VP16 was based on our previous
`studies [12]. Cells were fixed on alcian blue coated cov-
`erslips using 4% paraformaldehyde before and after the
`designated time of incubation, and stored in 70% ethanol
`for in situ detection of apoptosis. Experiments were re-
`peated four times and the results represent the average of
`four experiments.
`
`Single Labeling of Apoptosis With DNA Pol I
`or TdT
`These techniques to detect apoptotic cell death in situ
`have been described in detail in previous reports by us
`[1–3,12] and others [9,10]. Briefly, cells were rehy-
`drated, postfixed in 0.23% periodic acid, and pretreated
`with SSC solution (30 mM/l sodium citrate + 0.3 M/l
`sodium chloride, pH 7.0) at 80°C for 20 min. The cells
`were then treated with either a mixture of deoxyribo-
`nucleotides, one of which was biotinylated (11 bio-
`dUTP; Sigma, St. Louis, MO) and E. coli DNA Pol I
`(Promega, Madison, WI) or with 11 bio-dUTP and TdT
`(Promega). Incorporation of labeled nucleotide was vi-
`sualized using avidin-biotin-horseradish peroxidase
`(Vectastain Elite ABC kit, Vector, Burlingame, CA) and
`diaminobenzidine tetrahydrochloride (DAB). Dark
`brown nuclear staining indicated cells that were under-
`going apoptotic death.
`
`Double Labeling With DNA Pol I and TdT
`
`Cells were rehydrated, postfixed with 0.23% periodic
`acid, and pretreated with SSC solution. Subsequently,
`cells were washed in buffer A, pH 7.5 (50 mM/l Tris
`hydrochloride, 5 mM/l magnesium chloride, 10 mM/l b
`mercaptoethanol, and 0.005% bovine serum albumin,
`fraction V) (Sigma) and treated with a cocktail of four
`deoxynucleotides, one of which was biotinylated (bio
`dUTP) and DNA Pol I prepared in buffer A [0.01 mM/l
`of dATP, dCTP, and dGTP (Promega) + 0.001 mM/l bio
`dUTP (Sigma) and 20 U/ml Escherichia coli DNA Pol I
`
`(Promega)] at 18–19°C for 2 hr. The DNA-Pol reaction
`was visualized with ABC and DAB giving brown nuclear
`staining. Subsequently, cells were washed with buffer B,
`pH 6.8 (100 mM/l sodium cacodylate, 0.1 mM/l dithio-
`threitol, 5 mM/l cobalt chloride) (Sigma) and treated with
`a mixture of TdT (10 U/ml) and 0.001 mM/l digoxigenin-
`dUTP (Boehringer Mannheim, Germany) prepared in
`TdT buffer. The reaction was performed at 37°C for 1 hr
`and was terminated by immersing the coverslips in SSC
`solution. Incorporation of digoxigenin-dUTP was de-
`tected using antidigoxigenin Fab fragment conjugated
`with alkaline phosphatase (1:100 diluted; Boehringer
`Mannheim, Germany). The cells were then immersed in
`a solution freshly prepared as follows: Naphthol As-Mx
`Phosphate (20 mg; Sigma) was dissolved in 2 ml of
`N-N8-dimethylformamide (Sigma) and this was added to
`100 ml of 0.1 M/l Tris buffer, pH 8.2, at 20°C. Next, 0.1
`ml of 1 M/l levamisole was added to the solution to
`inhibit endogenous alkaline phosphatase, followed by
`100 mg of Fast blue BB salt (Sigma). This mixture was
`stirred for 2 min and then filtered. The coverslips were
`then immersed in this solution for 10–12 min. The cov-
`erslips were finally washed in distilled water and
`mounted in fluoromount. Positive TdT reaction thus
`stained nuclei blue.
`
`Controls for In Situ Labeling Procedures
`
`Specimens treated with the in situ end labeling reac-
`tion mixtures prepared in respective buffers with labeled
`nucleotide (and nucleotides for DNA Pol), but devoid of
`the enzyme (DNA Pol or TdT), served as negative assay
`controls.
`Treatment of normal human bone marrow and pe-
`ripheral blood mononuclear cells with known endo-
`nucleases. Normal bone marrow was collected from a
`resected rib procured during a thoracic surgical proce-
`dure and peripheral blood was obtained from a healthy
`donor under IRB approved protocols and with donors’
`written consent. Mononuclear cells were separated by
`density centrifugation and fixed on alcian blue coated
`coverslips with 4% paraformaldehyde and stored in 70%
`ethanol. The cells were rehydrated, postfixed with 0.23%
`periodic acid and pretreated with SSC solution as de-
`scribed above. At this point cells were subjected to dif-
`ferential treatments with known sequence specific endo-
`nucleases (Boehringer Mannheim) in respective reaction
`buffers provided along with the enzymes and at the over-
`digestion concentrations recommended by the manufac-
`turer as follows:
`
`1. BamHI (causes 38 recessed breaks with 58 over-
`hangs—58-Gfl GATCC-38) at 40 U/100 ml for 16 hr at
`37°C.
`2. PstI (causes 58 recessed breaks with 38 overhangs—
`58-CTGCAfl G-38) at 120 U/100 ml for 16 hr at 37°C.
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 3
`
`

`

`3. DraI (causes blunt ended breaks—58-TTTfl AAA-38)
`at 120 U/100 ml for 16 hr at 37°C.
`4. PstI (120 U/100 ml) for 16 hr at 37°C followed by
`BamHI (20 U/100 ml) for 2 hr at 37°C.
`5. Mixture of PstI +DraI (120 U/100 ml each) for 16 hr
`at 37°C.
`
`At the end of treatments, cells were washed thoroughly
`with phosphate buffered saline and continued with
`double labeling as described above. Experiments were
`repeated to confirm the results.
`Determination of labeling index. Every specimen
`was carefully observed under light microscopy and 1–2
`thousand cells from several randomly selected 100× ob-
`jective fields were counted to determine the percent of
`labeled cells in each case. The significance of differences
`in the mean percent labeling indices in various test
`groups was determined by the Student’s t test. Also, the
`paired t test was used to determine the significance of
`differences in relative percent increase in labeling index
`of individual labeling patterns within each study group.
`
`RESULTS
`Spontaneous Apoptosis in MDS
`Our earlier studies showed an excessive incidence of
`spontaneous apoptosis in the bone marrows of MDS pa-
`tients in which a significant number of mononuclear cells
`from the bone marrow aspirates of these patients under-
`went apoptosis in 4-hr cultures in complete medium con-
`taining 10% serum [1,2]. Therefore, in the present study
`comparative labeling of Ficoll-separated mononuclear
`cells from 16 MDS aspirates and six normal aspirates,
`with DNA Pol and TdT was examined at 0 and 4 hr
`following incubation in vitro in a complete medium.
`Staining was performed after paraformaldehyde fixation
`at each time point. As shown in Figure 1, the mean
`labeling index of MDS cells at 4 hr by either enzymes
`was significantly higher than that at 0 hr. Furthermore,
`the labeling by the two enzymes at each time point was
`comparable (0.8% ± 0.4% at 0 hr vs. 3.9% ± 0.9% at 4
`hr, n 4 16, P 4 0.004, by DNA Pol, and 1.3% ± 0.4%
`at 0 hr vs. 3.8% ± 1.0% at 4 hr, n 4 10, P 4 0.046, by
`TdT). Interestingly, normal cells also showed a marginal
`but significant increase in labeling index in 4 hr (0.1% ±
`0.07% at 0 hr vs. 0.8% ± 0.1% at 4 hr, n 4 6, P 4 0.001,
`by DNA Pol, and 0.2% ± 0.08% at 0 hr vs. 1.5% ± 0.2%
`at 4 hr, n 4 5, P 4 0.005, by TdT). It is evident that at
`4 hr, the labeling indices of MDS cells were 2–3 times
`higher than those of normal cells (P 4 0.003 for DNA
`Pol and P 4 0.05 for TdT), hence confirming the in-
`creased propensity of MDS bone marrow cells to un-
`dergo spontaneous apoptosis in vitro.
`We subsequently performed our newly developed en-
`zymatic double labeling with DNA Pol/diaminoben-
`
`Involvement of TNF-a in Myelodysplasia
`
`39
`
`Fig. 1. Comparative detection of apoptosis in MDS and
`normal cells using DNA Pol or TdT single labeling: Bone
`marrow aspirate mononuclear cells from MDS and normal
`subjects were incubated in RPMI 1640 containing 10% FBS
`for 4 hr. Spontaneous apoptotic death was detected in these
`cells following fixation in 4% buffered paraformaldehyde,
`using single labeling with DNA Pol or TdT and bio-dUTP.
`Percent positively labeled cells (dark brown staining in the
`nuclei) were determined in each case. By either technique,
`the labeling index was significantly higher at 4 hr in MDS as
`well as in normal, but those in the former were 2–3 times
`higher than the latter, indicating higher propensity of MDS
`cells to undergo spontaneous apoptosis. No significant dif-
`ference was noted in the labeling efficiencies of the two
`enzymes at either time points within each group.
`
`zidine (brown staining) and TdT/fast blue (blue staining)
`systems, at the two time points on nine MDS and five
`normal specimens. Three distinct labeling patterns were
`observed both in MDS and normal specimens—cells
`with nuclei stained only brown (only DNA-Pol reaction),
`only blue (only TdT reaction), and double labeled (DNA
`Pol + TdT reaction). Interestingly, as depicted in Figure
`2, each labeling pattern was found at all phases of apop-
`tosis ranging from the early-stage nuclear margination to
`intermediate-stage chromatin condensation and clump-
`ing, to the end-stage karyorrhexis, indicating the main-
`tenance of labeling pattern throughout the process of ap-
`optosis at a single cell level. The increase in mean label-
`ing index in 4 hr, for each individual pattern was
`significant in MDS (Fig. 3a). In contrast, normal cells did
`not show a significant increase in individual patterns, but
`the increase in total index reached statistical significance
`(Fig. 3b).
`Parallel to these experiments, DNA fragmentation was
`also studied by agarose gel electrophoresis (MDS, n 4 8
`and normal, n 4 2). As illustrated in Figure 4, at 0 hr
`neither the MDS specimen (lane 2) nor the normal cells
`(lane 5) showed low molecular weight DNA fragments.
`However, after 4 hr MDS cells showed an intensely
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 4
`
`

`

`40
`
`Mundle et al.
`
`Fig. 2. Double labeling of bone marrow
`aspirate mononuclear cells with DNA Pol/
`biotin-dUTP/diaminobenzidine (brown
`staining) and TdT/digoxigenin-dUTP/Fast
`blue (blue staining): As depicted here,
`both in MDS and normal specimens, three
`distinct labeling patterns were recog-
`nized: only-brown, only-blue, and dou-
`ble labeled. Furthermore, interestingly
`enough, each pattern was seen at early as
`well as late stages of apoptosis ranging
`from nuclear margination at early stage to
`larger clumping and fragmentation at late
`stages, e.g., double-labeled cells at early
`(B), intermediate (C), and late (D) stages
`or brown labeling at early (C) and late (A)
`stages seen in these micrographs. Origi-
`nal magnification, ×1000.
`
`stained, characteristic ladder of low molecular weight
`DNA fragments (lane 3), whereas the two normal speci-
`mens studied showed only a faint ladder (lane 6). One of
`the MDS specimens studied showed laddering at 0 hr
`also. The presence of clear laddering and lack of smear-
`ing in each case shows the absence of necrosis and con-
`firms apoptosis.
`
`TNF-a Vs. VP16-Induced Apoptosis in
`HL60 cells
`
`(Fig. 5b). In contrast, VP16 treatment showed the highest
`labeling indices with a remarkable increase in cells with
`only-TdT reaction and only a marginal increase in the
`other two patterns (Fig. 5c), with a significant increase in
`the total labeling index. Thus, at the end of the designated
`period of incubation, among individual labeling patterns,
`untreated cells and VP16-treated cells showed the high-
`est labeling indices in cells with only-TdT reaction,
`whereas TNF-a treated cells showed the highest indices
`in double-labeled cells (DNA Pol + TdT reaction).
`
`DNA fragmentation patterns were also studied in
`HL60 cells induced to undergo apoptosis by treatment
`with TNF-a (0.01 ng/ml for 8 hr) or VP16 (0.35 mM/l for
`4 hr). Cells incubated with vehicle only (culture medium)
`for the highest incubation period of 8 hr, served as con-
`trols. The experiments were repeated four times. Con-
`trols however were available from three experiments.
`DNA Pol/TdT double labeling was performed after para-
`formaldehyde fixation of cells before and after treatment
`in each experiment. The results described here represent
`the average of four experiments. Surprisingly, like MDS
`or normal cells, even the HL60 promyelocytic cells
`showed three distinct labeling patterns in each group. In
`the untreated group, during the 8-hr incubation, there was
`a slight increase in cells with only-DNA Pol reaction, a
`notable increase in cells with only-TdT reaction, and vir-
`tually no increase in double-labeled cells (Fig. 5a). On
`the other hand, TNF-a treatment induced an appreciable
`increase in the number of cells with only-DNA Pol re-
`action, virtually no increase in cells with only-TdT reac-
`tion, and a significant increase in double-labeled cells
`(DNA Pol + TdT reaction) and in total labeling index
`
`Comparison of Relative Percent Increase in
`Different Labeling Patterns in MDS, Normal, and
`HL60 Cells
`
`Considering the total net increase in percent positively
`labeled cells as 100%, relative percent increase in indi-
`vidual pattern was calculated in each case. Figure 6 com-
`pares the relative percent increase in individual labeling
`patterns during spontaneous apoptosis of MDS and nor-
`mal cells, and in HL60 cells with or without treatment.
`Normal cells showed a comparable increase in the three
`labeling patterns (only-DNA Pol reaction—30.6% ±
`19.5%; only-TdT reaction—39.8% ± 18.8%, and double
`labeled (DNA Pol + TdT reaction)—29.6% ± 11.2%, n
`4 5). On the other hand, MDS cells showed twice as
`much increase in double-labeled cells and in cells with
`only-DNA Pol reaction as compared with the cells with
`only-TdT reaction (39.2% ± 6.8%, P 4 0.08 and 38.6%
`± 5.3%, P 4 0.06 vs. 22.3% ± 6 %respectively, n 4 9).
`Surprisingly, TNF-a–treated HL60 cells exhibited a
`similar pattern to spontaneous apoptosis in MDS, show-
`ing the highest increase in double-labeled cells and the
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 5
`
`

`

`Involvement of TNF-a in Myelodysplasia
`
`41
`
`Increase in individual label-
`Fig. 3.
`ing patterns in MDS and normal
`cells: Percentage of cells labeled
`positively for each individual stain-
`ing pattern were determined in MDS
`(a) and normal specimens (b). Note
`the significant increase observed in
`4 hr in each individual labeling pat-
`tern in MDS cells as compared with
`the normal cells, which showed
`marginal increase only in the total
`labeling index.
`
`least increase in cells with only-TdT reaction (double
`labeled—47.3% ± 4.4%, P 4 0.015 and only-DNA Pol
`reaction—32.2% ± 8.2%, P 4 0.3 vs. only-TdT reac-
`tion—20.5% ± 6.6%, n 4 4). The untreated and VP16-
`treated HL60 cells qualitatively showed a comparable
`pattern with relatively highest increase in cells with only-
`TdT reaction (69.9% ± 8.9%, P ł 0.01, n 4 3 and
`62.6% ± 15.5%, P ł 0.07, n 4 4 respectively), albeit
`quantitatively, the labeling indices after VP16 treatment
`were 10 times higher than those of untreated cells.
`
`Determination of Specific DNA Breaks Labeled by
`Individual Staining Patterns
`
`To determine if the individual staining patterns ob-
`served in our double-labeling experiments were related
`to specific types of DNA breaks, we treated normal hu-
`man bone marrow—and peripheral blood—mononuclear
`cells with different known sequence-specific endonucle-
`ases (also see Methods section and the legend to Figure
`7 for experimental details). Figure 7 shows staining in
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 6
`
`

`

`42
`
`Mundle et al.
`
`Fig. 4. DNA ladder in MDS and normal cells: Agarose gel
`electrophoresis of DNA extracted from MDS and normal
`mononuclear cells, incubated for 4 hr in RPMI-1640 medium
`containing 10% FBS. (Ethidium bromide staining photo-
`graphed in ultraviolet light). At 0 hr neither MDS (lane 2) nor
`normal cells (lane 5) showed the presence of low molecular
`weight DNA bands. At 4 hr, whereas MDS cells showed in-
`tensely stained low molecular weight bands (lane 3), the
`normal cells showed only a faint ladder (lane 6), suggesting
`the internucleosomal DNA cleavage causing fragments of
`unique length.
`
`normal bone marrow mononuclear cells under different
`conditions, which was paralleled by that in peripheral
`blood cells (data not shown). These bone marrow mono-
`nuclear cells virtually showed no labeling in the absence
`of nucleolytic treatment (Fig. 7). Upon treatment with the
`three endonucleases, however, these cells showed a spe-
`cific staining pattern in each case, following in situ
`double labeling with DNA Pol and TdT. As illustrated in
`Figure 7, treatment with BamHI, which gives rise to 38
`recessed double-stranded DNA fragments with 58 over-
`hangs, showed only-DNA Pol specific brown staining of
`the nuclei. On the other hand, treatment with PstI that
`gives rise to double-stranded 58 recessed DNA fragments
`with 38 overhangs, or with DraI which generates blunt
`ended double-stranded DNA fragments, exclusively
`showed only-TdT specific blue staining. Interestingly,
`when cells were treated with both PstI and BamHI, as
`shown in Figure 7, the majority of positive cells were
`double labeled (DNA Pol + TdT reaction), whereas treat-
`ment with PstI and DraI demonstrated only-blue labeling
`(only-TdT reaction). Surprisingly, in the former case,
`double labeling was seen only when cells were treated
`sequentially with PstI first followed by BamHI, while a
`concomitant treatment with the two enzymes showed
`dominance of BamHI giving only-brown labeling (only-
`DNA Pol reaction; data not shown). On the other hand,
`sequential as well as concomitant treatment with PstI and
`
`DraI showed blue labeling only (only-TdT reaction).
`These experiments thus suggest that the brown labeling
`in MDS, normal, or HL60 cells could be due to the
`presence of 38 recessed DNA fragments, whereas the
`blue labeling may be due to the presence of 58 recessed
`and/or blunt-ended DNA fragments. Furthermore, the
`double labeling could be due to the copresence of 38
`recessed DNA fragments with 58 recessed or blunt-ended
`DNA fragments.
`
`DISCUSSION
`The mainstay in the therapy of MDS continues to be
`supportive care. Only through an in-depth understanding
`of the pathobiology of this intriguing disorder that we
`can hope to develop novel approaches in its treatment.
`Our present report not only provides a number of inter-
`esting insights into the biology of MDS and also into the
`process of apoptotic DNA fragmentation in general, but
`also shows a new ray of hope in the development of
`treatment for MDS. The most salient findings of these
`studies are summarized below and are subsequently dis-
`cussed in detail:
`1. The prominent DNA fragmentation pattern observed
`in spontaneous apoptosis of bone marrow cells in
`MDS is similar to that brought about by TNF-a treat-
`ment of HL60 cells.
`2. Apoptotic DNA fragments may have a unique length
`but have differently staggered ends.
`3. The two commonly used enzymes, DNA Pol and TdT,
`in a sequential double-labeling technique are now
`shown to differentiate the end patterns of apoptotic
`DNA fragments. DNA Pol can detect only 38 recessed
`DNA fragments, whereas TdT applied subsequently,
`detects 58 recessed and blunt-ended fragments. This
`observation defining the detection specificities of the
`two enzymes may warrant a caution in the current
`indiscriminate use of either enzyme singularly for de-
`tection of apoptosis.
`4. The characteristic DNA fragmentation pattern at a
`single cell level may be conserved throughout the pro-
`cess of apoptotic DNA disintegration.
`5. Different apoptotic stimuli could cause distinct DNA
`fragmentation patterns.
`Apototic death in the BMs of MDS patients dem-
`onstrates TNF-a–inducible DNA fragmentation pat-
`tern. In the present study apoptosis was studied in den-
`sity separated mononuclear cells from bone marrow as-
`pirates of MDS patients. The extent of apoptosis
`determined in bone marrow aspirate cells during very
`short-term (only 4 hr) in vitro incubation conditions ap-
`pears to be lower than the estimates reported previously
`in bone marrow biopsies by in situ studies by us and
`others [1,3,5]. However, our present estimates are in con-
`cordance with the estimates reported previously by us
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 7
`
`

`

`Involvement of TNF-a in Myelodysplasia
`
`43
`
`Fig. 5. Double labeling of HL60 cells: Freshly
`harvested HL60 cells were treated with 0.01 ng/
`ml TNF-a (b) or 0.35 µM/l VP16 (c) in RPMI-1640
`medium containing 20% FBS. Cells incubated in
`plain medium served as controls (a). (n = num-
`ber of experiments) Percent positively labeled
`cells were determined following DNA Pol/TdT
`double labeling. Interestingly, untreated cells
`(a) and VP16-treated (c) cells showed the high-
`est increase in cells labeled with only-TdT reac-
`tion (only-blue staining), whereas following
`TNF-a treatment (b) the highest increase was
`found in cells double-labeled with DNA Pol +
`TdT reaction. The net increase in total labeling
`index was two times higher with TNF-a and al-
`most 10 times higher with VP16 treatment as
`compared with the untreated cells.
`
`and others in MDS aspirates [2,4]. Nonetheless, all pre-
`vious reports by others and us [1–5] are in complete
`agreement with the fact that MDS show higher apoptosis
`than normal in either type of specimen. In the present
`study, the MDS bone marrow cells showed 2–3 times
`higher propensity than the normal cells to spontaneously
`undergo apoptosis despite the presence of serum in the
`medium. When relative percent increase in different la-
`beling patterns of the total net increase was compared in
`normal, MDS, and HL60 cells, as shown in Figure 6, in
`normal cells the rate of increase in each individual label-
`ing pattern was comparable. Surprisingly in MDS, the
`comparative labeling pattern exhibited a similar profile
`as that demonstrated by the TNF-a–treated HL60 cells.
`Both showed the highest increase in cells double labeled
`
`with DNA Pol and TdT, and the least in those labeled by
`only-TdT reaction. Our earlier studies have shown ex-
`cessive apoptosis and higher levels of TNF-a in MDS
`marrows as compared with normal marrows [3]. Using a
`simultaneous histochemical double labeling for TNF-a
`and apoptosis, these studies also revealed high prepon-
`derance of TNF-a around cells undergoing apoptosis.
`The present studies thus provide further circumstantial
`evidence for the possible association of TNF-a with in-
`creased incidence of apoptosis in MDS.
`Apoptotic DNA fragments may have unique length
`but differently staggered ends. As mentioned earlier,
`the majority of apoptotic endonucleases appear to be to-
`pologically restricted in their action to the internucleo-
`somal linker region of DNA [6–8], which in general ap-
`
`DR. REDDY’S LABS., INC. EX. 1027 PAGE 8
`
`

`

`44
`
`Mundle et al.
`
`Fig. 6. Comparison of relative per-
`cent increase in individual labeling
`patterns of the total net increase in
`MDS, normal, and HL60 cells. As il-
`lustrated in the figure, whereas nor-
`mal cells showed comparable in-
`crease in the three patterns, MDS
`cells showed relatively highest in-
`crease in cells double labeled with
`DNA Pol + TdT reaction and least
`in cells with only-T