(e) / 983 by Grune & Stratton, Inc. Analysis of Human Myelogenous Leukemia Cells in the Fluorescence-Activated Cell Sorter Using a

Transcription

1 Analysis of Human Myelogenous Leukemia Cells in the Fluorescence-Activated Cell Sorter Using a Tumor-Specific Antiserum By Andrew J. Malcolm, Patricia M. Logan, Robert C. Shipman, Reinhard Kurth, and Julia G. Levy The properties of a rabbit antiserum (anti-ami) raised to a purified protein from membranes of human acute myelogenous leukemia () cells is described. and peripheral blood leukocytes () from either normal mdividuals or patients with either myeloproliferative or other disorders were analyzed in a fluorescence-activated cell sorter (FACS IV) after labeling with anti-. normal rabbit serum (NRS). or antiserum raised to normal human membrane antigens. Of 40 cell samples from patients with acute myelogenous leukemia, 39 reacted strongly with the anti- antiserum. Similarly. all of 1 9 specimens from patients with chronic granulocytic leukemia reacted with the anti-ami. When 42 bone marrow or samples from A CONSIDERABLE amount of work has been done on the characterization of cell surface markers that may classify leukemic cells. The majority of these described constitutive normal surface markers that may be expressed with greater frequency or density on the surface of leukemic cell populations. In studies on acute nonlymphocytic leukemias (myelogenous), there has been some evidence that antigens unique for the leukemia cells may also be present. Baker and coworkers2 have demonstrated a leukemiaassociated antigen (LAA) on myelogenous leukemia cells using a specific antiserum raised in mice. They were able to predict imminent relapse of remission patients by demonstrating an increased number of seropositive cells in bone marrow aspirates detected by immunofluorescence. Similarly, simian heteroantisera have been used to identify what appear to be LAA on myelogenous leukemia cells.3 In this laboratory, we reported the properties of a rabbit anti- that exhibited apparently complete specificity in the enzyme-linked immunosorbent assay (ELISA) for membrane extracts of cells from patients with either acute or chronic myelogenous leukemia and had no reactivity with equivalent preparations from From the Department of Microbiology. University of British Columbia, Vancouver, British Columbia, Canada, and the Paul- Ehrlich-Institut, Frankfurt, Germany. Supported in part by Grant 65-#{243}O48from the N.C.I. of Canada and Grant from the British Columbia Health Science Research Foundation. Submitted May ; accepted November 9, Address reprint requests to Dr. Andrew J. Malcolm, Department of Microbiology, University of British Columbia, Vancouver, British Columbia, Canada V6T I W5. (e) / 983 by Grune & Stratton, Inc l/83/ $01.00/0 patients with a variety of lymphoproliferative disorders were examined, 2 specimens reacted with the antiserum. both from individuals with diagnoses of acute Iymphocytic leukemia (ALL). None of the 14 normal bone marrow or donor specimens tested reacted with the antiserum. It was also found that essentially all samples from patients in clinical remission from had high numbers of cells reactive with the anti-. When cells from such individuals were labeled and sorted on the FACS IV. it was found that cells fluorescing strongly with the anti- contained cells of both myeloid and lymphoid origin. The implications of these results are discussed. cells of either normal individuals or patients with lymphoproliferative disorders (ALL, CLL, lymphoma). The antiserum was reactive in the ELISA with preparations made from either bone marrow or from patients.4 More recently, using preparative polyacrylamide gels, we have isolated an antigen with which the antiserum reacted and prepared a high titer antiserum to it in rabbits. This antiserum subsequently showed apparent specificity for cell extracts in the ELISA. Preliminary studies using a fluorescenceactivated cell sorter (FACS IV) showed that this antiserum, when developed with fluorescein-labeled goat anti-rabbit IgG, reacted strongly with cells from bone marrow aspirates of 9/9 patients and 6/6 chronic granulocytic leukemia (CGL) patients. When bone marrow aspirates from patients with a variety of lymphoproliferative disorders (ALL, CLL, lymphomas, myeloma) were tested, 0/ I 2 showed any reactivity with the antiserum. Similarly, 0/4 normal bone marrow aspirates reacted.5 The present report is an extension of the previous publication in which both bone marrow cells and have been examined from patients presenting with in both the active disease state and in clinical remission. Analogous studies have been done on cells from normal individuals, patients with CGL, and patients with various lymphoproliferative disorders. The data reported show that this antigen may constitute a malignancy marker for a malignant clone in patients with myelogenous leukemias, since it was found to be present not only on blast cells but also on other cell populations both in patients with active disease and in remission. Subjects MATERIALS AND METHODS We studied 40 patients with acute myelogenous leukemia () and 19 patients with chronic granulocytic leukemia (CGL). Of the 858 Blood, Vol. 61, No. 5 (May), 1983: pp

2 FACS ANALYSIS OF MYELOGENOUS LEUKEMIA CELLS 859 patients, 19 were tested when newly diagnosed, 16 patients were studied when in clinical remission, and 7 as relapsed patients. Two other patients were analyzed after bone marrow transplantation. When attainable, samples from the same patient as newly diagnosed patient, as a remission patient, and/or as an in relapse were tested. Forty-two patients with leukemias of nonmyelogenous origin or with other malignant and nonmalignant conditions were used as controls. Fourteen laboratory personnel and bone marrow transplant donors were also used as normal Antiserum controls. The preparation of blast cell or normal peripheral blood leukocyte () membrane extracts and the isolation of specific bands on 7.5% polyacrylamide gel electrophoresis gels was described previously in detail.45 The resulting materials shown to have antigenic activity in the ELISA were used to immunize young adult female albino rabbits to produce an anti- and an antinormal-human-serum. These antisera were assayed by the ELISA, and the specificity of both were determined in an earlier reported study using fluorescently labeled bone marrow cells in the FACS IV.5 More recently, we have established that the antigen used to raise the anti- antiserum appears to be a homogeneous protein that produces a single spot on two-dimensional gel electrophoresis, has an isoelectric point (p1) of 7.1, and a molecular weight of 68,000 daltons (Shipman, Malcolm and Levy, submitted to Br J Cancer). Thus, the antiserum used here was prepared against a homogeneous pure protein, according to the standard biochemical criteria. Antisera to this antigen (-l ) have now been raised in six rabbits. All antisera produced show analogous specificity for the antigen and may be used interchangeably. The ability of one of these antisera to precipitate -l antigen was tested by immunoprecipitation. lodinated -l was precipitated with rabbit anti--l and protein A and run on a polyacrylamide gel as previously described. The results (Fig. 1) show that the antiserum precipitates predominantly -l material and is active up to a dilution of 1:8000. The normal rabbit serum (NRS) used in these studies was taken from unimmunized rabbits that were used subsequently for immunization with the antigen. Consequently, they served as an appropriate negative control in these experiments. Fluorescent Antibodies Fluorescein-labeled DEAE-purifIed goat anti-rabbit lgg was prepared according to a standard procedure.6 Briefly, the goat antibody at 10 mg/mi was dialyzed exhaustively against 0.15 M NaCI, after which it was dialyzed for 5 hr against 0.5 M bicarbonate-buffered saline, ph 8.5, and finally against 0.05 M bicarbonatebuffered saline, ph 9.2, for 3 hr. all at 4#{176}C. The IgG was then dialyzed for 24 hr against 100.tg/ml of fluorescein isothiocyanate (FITC, BBL) in a 0.05 M bicarbonate-buffered saline at ph 9.2. The reaction was stopped by dialysis against 0.02 M PBS, ph 7.0, at 4#{176}C for 4 hr. Unbound FITC was removed from FITC-lg conjugates by passage of the material over Sephadex G-25. A ratio of fluorochrome to protein (F:P) of 4 was determined by spectrophotometric analysis.7 Cells and Cell Labeling aspirate and peripheral blood samples were obtained from the Division of Hematology, Vancouver General Hospital, or the Cancer Control Agency of British Columbia (Terry Fox Laboratory). Samples obtained were from patients whose diagnoses (, CGL, or otherwise as referred to in Tables 1-10) were established either prior to or following our testing. The clinical diagnoses were made by hematologists at, or affiliated with, the S 10 Fig. i. Immunoprecipitation of 21-labeled -i with rabbit antiserum. (Lane i ) Normal rabbit serum cpm of antigen; (2) 1 :2 dilution of anti--i cpm of antigen; (3) 1 :8 dilution of anti--i cpm of antigen; (4) i :32 dilution of anti- i cpm of antigen; (5) i :1 28 dilution of anti- i cpm of antigen; (6) 1:512 dilution of anti ,000 cpm of antigen; (7) 1 :2048 dilution of anti cpm of antigen; (8) 1 :81 92 dilution of anti- I cpm of antigen; (9) -i alone, cpm; (10) -i alone cpm K

3 860 MALCOLM ET AL. Vancouver General Hospital. The tests used included: morphological examination of peripheral blood and bone marrow smears, biochemical tests on bone marrow (Sudan black, PAS, combined esterases, and acid phosphatase tests where necessary), colony growth studies, and chromosomal analyses (in cases of CGL). Lymphocyte surface Ig and rosette tests were performed on samples from patients with lymphoid leukemias. Buffy coats were collected from bone marrow samples by sedimentation at I g, and from the heparinized peripheral blood samples using Plasmagel (Laboratoire Roger Bellon, Neuilly, France) or by the Ficoll-Hypaque technique.89 In the case of peripheral blood samples, we obtained similar results to those of Boyum* and English and Anderson9 regarding the relative purities of mononuclear-cell-enriched (96%) and granulocyte-enriched (98%) cell suspensions recovered, respectively. The cells were washed in PBS containing 5% FCS, and the cell pellet was suspended in M ammonium chloride and M Tris buffer at ph 7.2 to lyse contaminating erythrocytes, followed by a subsequent wash. Then, 10 cells were incubated for 1.5 hr on ice in 0.2 ml of antiserum (anti-normal human, anti-, or normal rabbit serum diluted to 1/10 in PBS). All antisera used were centrifuged briefly before use at 20,000 g to eliminate aggregates. Cells were washed 3 times in PBS, and 0.2 ml of fluoresceinated goat anti-rabbit IgG at I /20 was added to each cell pellet. Cells were incubated another 1.5 hr at 0#{176}C, washed once in PBS, and then centrifuged through 100% FCS. They were finally suspended in 1.0 ml PBS and 5% FCS for FACS IV analysis. FACS IV Analysis Twenty-five thousand cells from each sample were analyzed on a Becton-Dickinson FACS IV using the 488-nm wavelength of the Spectra Physics Model Aragon laser at a power setting of 400 mw. The standard filter for FITC analysis was used (520-long pass filter). The FACS IV was standardized by using glutaraldehyde-fixed chicken red blood cells 0 and fluorescent monodispersed carboxymethylated microspheres (d - I.75 m ± 0.02 SD; cat. no. 9847, Polysciences Inc., Warrington, PA). The results reported here include all those samples that were analyzed and gave appropriate results in the FACS IV with the positive (anti-normal human) and negative (NRS) controls. Occasionally, we were unable to obtain samples or book FACS time until the specimens were 48 hr old. Sometimes these samples yielded either extremely high backgrounds with NRS or did not react strongly with the positive control antiserum. Thus, it was impossible in these circumstances to evaluate the reactivity ofcells with the anti-, and these results were discounted. We observed a certain degree of fluctuation in negative test samples in which numbers of cells fluorescing were compared to the NRS control. The anti- frequently reacted at a lower level with such cells than did the NRS. It was ascertained that the variation in the range of ±5% in comparison to NRS results was within the margins of error for these studies. FACS IV Cell Sorting Technique An APML remission sample that had been analyzed as significantly positive with anti- was sorted using the FACS IV. Two sort windows were determined, one containing 25% of the sample showing no or minimal relative fluorescence and one containing 25% of the sample showing the highest relative fluorescence. The two cell populations were collected in separate tubes containing I ml PBS/ FCS. The cells between these two windows, representing the remaining 50% of the sample cell population, were discarded into the reservoir flask. The head drive frequency was set at 36 KHz, and 2000 V were applied across the electrostatic deflection plates. The Eput counters recorded the number ofcells collected in the right and left deflection tubes, and the FACS was run at 5 droplets per deflection pulse with the abort on; the droplet delay was set at 14 drops. Cooling water (2#{176}C)was circulated around the collection and sample tubes. The cells were then washed once in PBS to remove protein the FCS and the sample loaded on a cytospin (John s Scientific) at 800 rpm for 8 mm to obtain slide preparations. The slides were stained routinely with Wright s stain and examined by light microscopy. RESULTS As established in earlier work we showed the specificity of the anti- on a series of bone marrow samples from patients with and from patients with a variety of other disorders. In the experiments reported here, we further show this specificity on additional types of bone marrow samples and also on peripheral blood samples. Table I shows the results of analyses of a bone marrow aspirate and peripheral blood leukocytes from an APML (acute promyelocytic APML Table 1. FACS Analyses of Cells From a Patient With APML and Her Identical Twin Number of Cells Fluorescing With Various Treatments Percent Anti-Normal Percent Diagnosis Blasts Human NRS Anti- Positive CelIst (1) ,610 3,136 22, (2) Peripheralblood ,796w 63.0 (3) Peripheral blood (mononuclear-enriched- ND. 23, % pure) (4) Peripheral blood (granulocyte-enriched- ND. 23, , % pure) Normal (identical twin) (5) Peripheral blood 0 20, <0 (6) Peripheral blood (mononucleer-enriched % pure) (7) Peripheral blood (granulocyte-enriched , % pure) A total of 25,000 cells were analyzed in each case. tpercent cells showing positive fluorescence over background NRS controls with anti--serum. la, lb. 2a, 2b: See fluorescence intensity profiles in Fig. 2.

4 FACS ANALYSIS OF MYELOGENOUS LEUKEMIA CELLS 861 lb 2b Fluorescence intensity profiles of peripheral blood leukocytes from a patient with APML (1 ) and her identical twin (2). Vertical number of cells in log scale. Horizontal axis. relative fluorescence intensity. Nonspecific fluorescence is demonstrated by the with NRS. and these profiles are shown in (a) and (b) in all cases. Superimposed on the NRS controls are the fluorescence profiles of cells treated with anti-normal human serum (positive control) shown in (a) and the fluorescence intensity profiles of anti- serum (b). patient and the peripheral blood leukocytes her normal identical twin. The APML bone aspirate, peripheral blood leukocyte sample, Ficoll-Hypaque-enriched populations of either or granulocytic cells (samples 1, 2, 3, all showed a high percentage of cells reacting with the anti-. Hence, both bone marand the peripheral blood cells from this APML the malignancy marker. It is clear,, that the anti- did not bind to the blood leukocytes, nor to the enriched monoor granulocyte populations (samples 5, 6, and of the normal identical twin. Represenfluorescence profiles of peripheral blood leukothe APML patient (sample 2) and from the identical twin (sample 5) are shown in Fig. 2. test demonstrates that the normal identical twin not have the antigen and hence would have been an ideal bone marrow donor for the APML sibling (58 yr. female), age being the limiting factor. Subsequently, either and/or bone marrow cells from a number of patients presenting with, APML, or AMML (acute myelomonocytic leukemia) were tested. The results of representative individual tests as well as the averaged results of the total number of tests run are shown in Table 2. It can be seen that in all cases, a high percentage of the cells analyzed fluoresced positively with the anti-. The results obtained with bone marrow aspirates and were essentially the same. It is of interest that the number of cells fluorescing with anti- shows no correlation with the number of blast cells found in either bone marrow or, but is consistently higher. A similar series of analyses was carried out on patients who were in clinical remission from acute myelogenous leukemia after chemotherapy. The

5 862 MALCOLM ET AL. Table 2. FACS Analyses of Cells From Untreated Patients With Acute Myelogenous Leukemia. A Total of 25,000 Cells Were Diagnosis Cell Population APML Analyzed With Each Sample Percent Blasts > With Anti ± 23.7 (average)t (13) AMML AMML ± 26.2 (average) (1 1) Percents are based on the following equation: [(cells fluorescing with anti- - cells fluorescing with NRS)/(cells fluorescing with antinormal - cells fluorescing with NRS)] x 100. taveraced results ± standard deviation of 1 3 bone marrow samples Averaged results ± standard deviation of 1 1 samples. results are shown in Table 3. It can be seen that in both bone marrow and, there is still a significant number of cells showing positive fluorescence with the anti-, even though the clinical state in each case is clearly one of remission. The variation between cases is somewhat greater than that seen in clinical disease Table 3. FACS Analyses of Cells From Patients in Clinical Remission From Acute Myelogenous Leukemia After Chemotherapy. A Total of Cells Were Analyzed Diagnosis Cell Population Bone marrow Bone Bone Bone Bone marrow marrow marrow marrow for Each Sample Tested Percent Blasts Occasional >100 With Anti ± 28.4 (average)t None Occasional None None None Occasional >100 > ± 29.7 (average) Percent positive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with antinormal - number fluorescing with NRS)] x 100. taveraed results ± standard deviation of 1 3 bone marrow samples from remission patients. Averaged results ± standard deviation of 6 samples from remission patients (i.e., in some patients the percent positive cells is lower than 25%; this is essentially never seen in active leukemia). Whether or not this is related to prognosis is uncertain at this time. These data indicate that, in remission patients, normally differentiating cells are expressing this cell surface antigen (malignancy marker). The observations that some samples that are enriched for mononuclear cells also showed a high number of fluorescing cells suggested the possibility that lymphoid as well as myeloid cells might be expressing this antigen. Since the involvement of lymphoid cells has never been implicated in human acute myelogenous leukemia, we wished to establish the gross morphology of those cells showing the highest relative positive fluorescence with the anti- and to determine whether this population did, in fact, include lymphocytes. To accomplish this we used the FACS IV to collect and sort from a remission patient who had been analyzed as significantly positive with the anti-. Cytology was performed on that 25% of the cell sample showing the lowest positive relative fluorescence and on that 25% of the same sample showing the highest positive relative fluorescence. Similar sorts were carned out with the positive control antiserum (antinormal-human). The results shown in Table 4 were obtained using from a patient with APML in remission, which had been subjected to a one-step Ficoll- H ypaque separation (mononuclear enriched). As can be seen, a majority of the cells showing the highest level of fluorescence were lymphoid. These results indicate that in remission patients, cells of both lymphoid and myeloid lineage are expressing this antigen. When bone marrow aspirates or of patients in relapse with acute myelogenous leukemia were exammed with the anti-, it was again seen that a majority of the cells showed positive fluorescence, and again, the percent positive cells did not correlate with the number of blasts present. The results are shown in Table 5. In an earlier study5 we showed that this antigenic marker was also present on bone marrow aspirate cells of patients with CGL. Comparative studies were done between bone marrow and of a number of these patients. The results are shown in Table 6, in which it can be seen that both bone marrow and cells showed positive fluorescence in all cases examined. We had also shown previously that the anti- did not react with bone marrow aspirates taken from normal individuals. That study was extended and includes a number of from individual donors. The results are shown in Table 7, in which it can be seen that no significant levels of fluorescence could be

6 FACS ANALYSIS OF MYELOGENOUS LEUKEMIA CELLS 863 Table 4. Microscopic Analysis of Cells From a Remission Patient With APML After Labeling With Either Anti-Normal Human or Anti- and After Analysis and Sorting on the FACS IV Test Anti serum Used Anti- Normal Human Anti- 20, (>100%) Percent Cells in Percent Cells in Percent Cells in Percent Cells in Test Performed: High Relative Low Relativet High Relative Low Relativet FACS IV Analyses Fluorescence Fluorescence Fluorescence Fluorescence (Number of Cells Fluorescing): Population Population Population Population Neutrophils 2 Occasional 3 - Lymphocytes Monocytes Red blood cells - Occasional - - Eosinophils - - Occasional - Those cells sorted as the 25% of the population having the highest relative fluorescence. tmose cells sorted as the 25% of the population having the lowest relative fluorescence. detected with the anti- in any of the samples tested. Our previous data had shown that bone marrow cells from patients with nonmyelogenous malignancies (ALL, CLL, lymphomas) did not react with the anti- in FACS IV analyses. Further studies shown here (Table 8) substantiate our previous findings in that both bone marrow cells and of these patients do not react significantly with the antiserum. We have had the opportunity to examine cells from two patients who received allogeneic bone marrow transplants from tissue-matched sibling donors as a treatment for (Table 9). Donor cells had not been treated in any way to reduce the mature T-cell population. Cells from patient 1 were taken 20 mo after transplantation and the patient was doing well with no disease recurrence. The cells from this patient did not show significant fluorescence with the test anti-. The cells from patient 2 were followed from shortly after transplantation. At no time (either 2 wk after transplantation, when the patient was in clinical remission, or 4 mo later, when the patient was in relapse) did cells show negative reactivity with the Table 5. FACS Analyses of Cells From Relapsed Patients With Acute Myelogenous Leukemia Following Chemotherapy. A Total of Cells Analyzed for Each Sample Tested Diagnosis Cell Population Percent Blasts With Anti AMML 60 >100 AMML AMML 9 >100 Percent positive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with antinormal human - number fluorescing with NRS)J x 100. antiserum. The donor, on the other hand, had no reactive cells in his bone marrow. By and large, the anti- under test here has shown absolute specificity with regard to reactivity with cells of patients with myeloproliferative disorders. In all specimens tested, only 3 cases have not shown the expected reactions. These are shown in Table 10. Patient 1 was diagnosed as having ALL. At the time the bone marrow aspirate was tested, the patient was in clinical remission. However, a high percentage of the patients cells fluoresced strongly with the anti-. We have no explanation for this observation. Patient 2 was diagnosed as ALL. However, the clinical reports show that while bone marrow colony growth suggested ALL, peripheral blood growth better fitted some type of granulocytic disease. It would appear that this patient s diagnosis may be somewhat questionable. Patient 3 has been diagnosed as having. This is a juvenile (10-yr-old), and we have no rationale for this anomalous result. Table 6. FACS Analyses of Cells From Patients in the Chronic Phase of Chronic Granulocytic Leukemia. A Total of Cells Were Analyzed for Each Sample Tested Diagnosis Cell Population With Anti- CGL CGL CGL > CGL CGL ± 26.6t CGL 52.2 ± 23.3 Percent positive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with antinormal - number fluorescing with NRS)] x 100. taveracied results ± standard deviation of 8 bone marrow samples. Averaged results ± standard deviation of 1 5 samples.

7 864 MALCOLM ET AL. Table 7. FACS Analyses of Cells From Normal Individuals. A Total of Cells Were Analyzed for Each Sample Tested Cell Population With Anti <0 0.8 <0 <0 <0 <0 <0 <0 #{149}Percentpositive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with antinormal - number fluorescing with NRS)] x 100. DISCUSSION A number of questions are raised by the results reported here. We have shown that an antiserum raised in rabbits to an antigen purified on preparative polyacrylamide gels from membrane extracts from cells of patients reacts strongly in FACS IV analysis with a cell surface antigen found on 39 of 40 samples taken from patients diagnosed as having acute myelogenous leukemias. Cells taken from individuals with no known disorders did not react with this antiserum. It would thus appear that the isolated antigen represents what could be termed a tumor-associated antigen (TAA) or malignancy marker, in that it does not appear to be present, at detectable levels, on normal cells. This is in agreement with previous findings in this laboratory that the antigen could not be detected in extracts from normal cell membranes when tested with the antiserum in the ELISA.4 Whether or not the antigen is present on a small population of normal bone marrow cells is the subject of ongoing studies in this laboratory. We have also reported here that this associated antigen is also detectable on the cells of patients with CGL. Of 19 samples taken from patients with CGL, all samples showed significant numbers of Table 8. FACS Analyses of Cells From Individuals With Leukemias of Nonmyelogenous Origin. Lymphomas. or Other Hematologic Disorders. A Total of Cells Were Analyzed for Each Sample Tested Diagnosis Cell Population ALL-92% blasts <0 With Anti- ALL-congenital 8% blasts 6.2 CLL 0.32 Non-Hodgkin s lymphoma- 6% blasts <0 Idiopathic thrombocytopenic purpura 2.6 ALL-remission <0 ALL-remission 0.5 ALL-remission 1.3 ALL-remission 0.1 Idiopathic thrombocytopenic purpura <0 ALL-remission 27% blasts 7.4 ALL-remission (same patient when in remission) <0 Percent positive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with antinormal - number fluorescing with NRS)] x 100. positively fluorescing cells (at least 30% positive) in both bone marrow and materials. These results indicate that a common TAA is present on the cells of patients diagnosed as having either acute myelogenous leukemia or CGL. These results may imply a common event in the onset of both of these conditions. This has not been suggested in previous studies, even though both conditions involve proliferative disorders of myeloid cells. It has been a consistent observation in these studies that the numbers of cells reacting with the anti- antiserum in either bone marrow aspirates or in patients do not correlate with the number of blast cells present in the sample. This would imply that the antigen is present on cells other than blasts. This is explicitly demonstrated by the observation that patients in clinical remission from (Table 3) showed a high percentage of positively fluorescing cells. In patients with active disease, these findings indicate that this marker (antigen) may be present on a Table 9. FACS Analyses of Either Bone Marrow Cells or From Patients With or AMML Following Bone Marrow Transplantations. A Total of 25,000 Cells Were Analyzed for Each Sample Tested Patient No. Diagnosis Time After Transplant Condition With Anti- 1 bone marrow 20 mo Remission wk Remission (no blast cells) mo Relapse 23% blasts mo Relapse 23% blasts Normal donor sibling of patient 2 (bone marrow) - Normal <0 Percent positive [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with anti-normal - number fluorescing with NRS)] x 100.

8 FACS ANALYSIS OF MYELOGENOUS LEUKEMIA CELLS 865 Table 1 0. FACS Analyses on Cells From Three Patients Whose Cells Yielded Anomalous Results With Regard to Their Diagnosis. A Total of 25,000 Cells Were Analyzed for Each Sample Tested Patient No. Diagnosis Percent Blasts Cell Population With Anti ALLt ALL None-Remission <0 Percent positive = [(number of cells fluorescing with anti- - number fluorescing with NRS)/(number of cells fluorescing with anti-normal - number fluorescing with NRS)] x 100. tthis patient has been diagnosed as a typical ALL in remission. tthis patient shows that although bone marrow growth suggests ALL, peripheral blood cell colony growth better fits some type of granulocytic disease. This patient is a 1 0-yr-old juvenile case. pluripotent stem cell. The observation, in remission patients, that lymphoid as well as myeloid cells may be expressing this antigen suggest that: a pluripotent stem cell is expressing the antigen, and, cells expressing this antigen can differentiate, albeit transiently, in normal fashion. This finding also supports the possibility that there may be clonal dominance of a potentially malignant stem cell, probably pluripotential, in the bone marrow and peripheral blood of The question could be raised as to whether the antigen, detected on lymphocytes of remission patients, is actually synthesized by these cells or whether it is adsorbed from other sources. At this time, we do not have definitive proof as to the origin of the antigen on lymphocytes. Because the patients are in clinical remission, it is unlikely that the amounts of antigen observed in these cells are derived from a few occult malignant cells. However, the possibility that the antigen is derived from a normally differentiating myelocytic population cannot be ruled out; ongoing research is being carried out to address these possibilities. The results of extensive FACS IV sorting studies on peripheral blood and bone marrow cells from patients with myeloproliferative disorders are the subject of a manuscript in preparation. Further discussion of the possible significance of these results will be presented at that time. These findings are in agreement with those of Metzgar and Mohanakumar,3 who found that simian anti- antisera reacted positively with 40% of cells from a patient in remission from AMML. In 39 of 40 diagnosed cases of, the anti- used here detected antigen on the surface of a significant number of either bone marrow cells or of these patients. The one case of diagnosed whose cells did not react with the antiserum was a 10-yr-old juvenile case. We have no explanation for this. It may be that a small number of patients do not express this antigen on their bone marrow cells, or it may be that in children involves a series of cellular changes distinct from the adult disease. It is hoped that subsequent studies will clarify these possibilities. It was also found that cells from two patients diagnosed as ALL had cells (either bone marrow or ) that reacted strongly with the antiserum. One patient, who had a significant number of blast cells in both bone marrow and, had a disease picture that was somewhat ambiguous in that bone marrow cultures indicated ALL whereas cultures indicated a myeloproliferative disorder. The second patient was diagnosed as ALL but was in clinical remission at the time that cells were tested. Further testing of this patient s cells may clarify this case. At this time, it is only possible to state that the antigen is present on cells from a high percentage of patients diagnosed as (or other myeloproliferative leukemias) and may be expressed on cells of a small number of patients with ALL. It has not been observed on cells of patients with other lymphoid disorders. To date, we have only been able to examine cells from two patients who have received bone marrow transplants as treatment for. The patient who was doing well and was still in clinical remission after 20 mo had essentially no cells in a bone marrow aspirate that reacted with the anti-. The second patient, who was followed from 2 wk after transplantation, at all times showed a significant number of cells reactive with the antiserum. This patient relapsed at 4 mo post-transplant. It is possible that such monitoring may be of prognostic value in this form of treatment. We have shown that the antiserum described here is detecting an antigen that is present on the surface of a high percentage of both bone marrow cells and of most patients with or CGL, whether they are presenting with the disease, in remission, or in relapse. It is not present, at detectable levels, on normal bone marrow cells or normal, nor is it present, at detectable levels, on the cells of most patients with lymphoproliferative disorders. It is possible that this antigen may be expressed on a small population of normal cells of bone marrow origin, and this is the

9 866 MALCOLM ET AL. subject of ongoing studies. At present, work is underway using monoclonal antibodies raised to this antigen to clarify this question, as well as others regarding the nature and origin of this antigen. At this point, however, we conclude that this marker (antigen) may be of clinical significance in both prognosis and diagnosis of myeloproliferative diseases. ACKNOWLEDGMENT The authors are grateful to Dr. J. W. Thomas and Dr. J. Denegri, Vancouver General Hospital, and Drs. C. and A. Eaves, Cancer Control Agency of British Columbia, for their help in supplying patient material for our analyses. We are also indebted to Dan Zecchini for his expertise and his time spent in operating the FACS IV. REFERENCES I. Mahen M, Baker MA, Falk JA, Taub RN: Immunological diagnosis and monitoring of human acute luekemias. Am J Pathol 103: , Baker MA, Falk JA, Carter WH, Taub RN: Early diagnosis of relapse in acute myeloblastic leukemia: Serological detection of leukemia-associated antigens in human marrow. N Engl J Med 301 : , Metzgar RS, Mohanakumar T: Tumor-associated antigens of human leukemia cells. Semin Hematol 15: , Al-Rammahy A Kh, Shipman R, Jackson A, Malcolm A, Levy JG: Evidence for a common leukemia-associated antigen in acute myelogenous leukemia. Cancer Immunol Immunother 9: , Malcolm AJ, Shipman RC, Levy JG: Detection of a tumorassociated antigen on the surface of human myelogenous leukemia cells. J Immunol 128: , Wofsy L, Henry C, Kimura J, North J: Modification and use ofantibodies to label cell surface antigens, in Mishell BB, Shiigi SM (eds): Selected Methods in Cellular Immunology. San Francisco, W.H. Freeman, 1980, pp Wells AF, Miller CE, Nadel MK: Rapid fluorescein and protein assay method for fluorescent-antibody conjugates. Appl Microbiol 14: , Boyum A: Isolation of lymphocytes, granulocytes and macrophages. Scand J Immunol S (Suppl 5):9-15, English D, Anderson BR: Single step separation of red blood cells, granulocytes and monoclear leukocytes and discontinuous density gradients of Ficoll-Hypaque. J Immunol Meth 5: , Herzenberg LA, Herzenberg LA: Analysis and separation using the fluorescence activated cell sorter (FACS), in Weir DV (ed): Handbook of Experimental Immunology, vol 2. Oxford, Blackwell Scientific, 1978, pp Wiggans RG, Jacobson Ri, Fialkow PJ, Woolley PV, Macdonald is, Schein PS: Probable clonal origin of acute myeloblastic leukemia following radiation and chemotherapy of colon cancer. Blood 52: , Fialkow PJ, Singer JW, Adamson JW, Berkow RL, Friedman JM, Jacobson Ri, Moohr JW: Acute non-lymphocytic leukemia. Expression in cells restricted to granulocytic and monocytic differentiation. N Engl J Med 301:1-5, 1979 I 3. Kurth R, Huesgen A, Katz F, Lower J: Comparison of radioimmunoprecipitation assays for the detection of human antitumor virus antibodies. J Immunol Meth 30: , 1979

10 : AJ Malcolm, PM Logan, RC Shipman, R Kurth and JG Levy Analysis of human myelogenous leukemia cells in the fluorescence- activated cell sorter using a tumor-specific antiserum Updated information and services can be found at: Articles on similar topics can be found in the following Blood collections Information about reproducing this article in parts or in its entirety may be found online at: Information about ordering reprints may be found online at: Information about subscriptions and ASH membership may be found online at: Blood (print ISSN , online ISSN ), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC Copyright 2011 by The American Society of Hematology; all rights reserved.