Bone Marrow Transplantation, (1998) 22, 367–374  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt Differential growth patterns in SCID mice of patient-derived chronic myelogenous leukemias

J McGuirk1,2,5, Y Yan1,4, B Childs1,2, J Fernandez1, L Barnett1, C Jagiello1, N Collins1 and RJ O’Reilly1,2,3,4

1Bone Marrow Transplantation Service, 2Department of Medicine, 3Department of Pediatrics, and 4Research Animal Laboratory, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

Summary: The product of this rearrangement is a bcr-abl hybrid gene, which encodes a p210 protein product with tyrosine kinase The development of an in vivo model for the study of activity and is thought to have leukemogenic properties.4 CML would be of significant importance in studying its CML typically begins as an initial chronic phase (CP) biological behavior and developing novel therapeutic which lasts approximately 4–6 years with eventual pro- strategies. We examined the ability of human leukemic gression to blastic transformation commonly heralded by cells derived from patients in either chronic (CP), accel- the acquisition of additional chromosomal abnormalities by erated (AP) or blast phase (BP) CML to grow and dis- Ph+ stem cells including duplication of the Ph chromosome, seminate in CB17-SCID mice by subcutaneous (s.c.) isochromy 17q and trisomy of chromosomes 8, 19 or 21.5,6 inoculation without conditioning treatment or adminis- The relationship of these non-random chromosomal tration of cytokines. Additionally, samples derived from changes to the progression of this disease is not well under- patients with CP-CML were injected s.c. into CB17- stood. The majority of patient-derived CML cells have lim- SCID mice treated with anti-Asialo GM1 (an anti-NK ited proliferative capacity in vivo and available in vitro cell antibody) and NOD-SCID mice (absent NK cell assays are only capable of detecting progenitors with activity) to study the potential role of NK cell-mediated limited proliferative and replating potential.7,8 anti-leukemic activity in preventing the propagation of An in vivo model of CML would potentially allow for CP-CML cells. We observed a significant differential the development of novel therapeutic strategies in the treat- growth pattern of CML cells in the mice such that BP- ment of this disease as well as possibly lead to the eluci- CML grew rapidly as s.c. tumors and disseminated, dation of the molecular events involved in the evolution of while AP-CML or CP-CML cells grew temporarily as the CP to blastic phase (BP). small nodules that spontaneously regressed and did not Several recent studies have shown the growth of some disseminate. This differential growth pattern suggests human leukemia cells in the severe combined immunodefi- possible important biological differences. Furthermore, ciency (SCID) mouse model in a manner analogous to their no significant difference in s.c. growth or dissemination clinical characteristics in patients.9–12 This has allowed the of CP-CML samples derived from newly diagnosed propagation of human leukemic cells that in prior studies patients in untreated CB17-SCID mice and CB-17 SCID failed to grow in vitro or in nude mice. Although several mice treated with Anti-Asialo GM1 and NOD-SCID reports demonstrate the feasibility of consistently growing mice occurred, suggesting that factors other than NK human acute myeloid and lymphoid leukemias in SCID cell anti-leukemic activity may be important. mice the growth of CML cells in this model have mostly Keywords: microenvironment; blast phase; accelerated been limited to cell lines or samples derived from patients phase; chronic phase; subcutaneous in the BP of the disease.13–16 Only one group has been able to demonstrate even limited growth of CP-CML cells in SCID mice; this growth was limited to the bone marrow and Ph+ cells were identified in only a small minority of Chronic myelogenous leukemia (CML) is an acquired the animals studied.17 malignant disorder of the hematopoietic stem cell and is Recently, we developed a SCID mouse model that per- characterized by the proliferation of clonal myeloid cells mits the s.c. growth of primary human acute leukemia 1 and their progenitors. The hallmark of this disorder is the cells.9 Furthermore, the s.c. growth in this model is associa- presence of the Philadelphia (Ph) chromosome which ted with dissemination of the leukemia reflecting a pattern results from a specific reciprocal translocation, of growth similar to the usual clinical presentation. To t(9;22)(q34;q11), of the abl oncogene on chromosome 9 to determine whether and to what degree different phases of 2,3 the break point cluster region (bcr) on chromosome 22. CML would grow or disseminate within the SCID mouse model, we subcutaneously inoculated SCID mice with samples of cells derived from patients with CP, AP or BP- 5Correspondence at his present address: Dr JP McGuirk, Yale School of Medicine, Bone Marrow Transplantation, 333 Cedar Street Room 211 CML. Additionally, several samples derived from patients WWW, PO Box 208032, New Haven, Connecticut 06520–8032, USA with CP-CML were injected subcutaneously into SCID Received 5 February 1998; accepted 20 April 1998 mice treated with anti-Asialo GM1 (an anti-NK cell CML growth in SCID mice J McGuirk et al 368 antibody) and NOD-SCID mice (absent NK cell activity) to growth was assessed by weekly measurements of the study the potential role of NK cell-mediated anti-leukemic dimensions of subcutaneous nodules. Tumor size was cal- activity in preventing the propagation of CP-CML in culated by taking the square of the mean radii and multiply- SCID mice. ing by ␲. Animals inoculated with BC-CML leukemic cells In this report we demonstrate a differential growth pat- were sacrificed by cervical dislocation when the dimensions tern of patient-derived CP and AP vs BP CML in a SCID of subcutaneous nodules were at least 1.5 cm2 or more than mouse by s.c. inoculation and present the characteristic and 3 months after inoculation. Animals inoculated with AP- biologic behavior of the human chronic myelogenous CML and CP-CML leukemic samples were sacrificed in leukemias in this model. the same manner when tumor nodule size reached maximal size or more than 3 months after inoculation. The gross anatomy was evaluated and samples from peripheral blood, Materials and methods sternum, femur, spleen, liver, lung–heart complex, kidney, brain and tumor nodules were subsequently removed for Leukemic samples analysis by flow cytometry (FACS), fluorescent in situ hybridization (FISH), polymerase chain reaction (PCR) for Samples were obtained, with informed consent, during rou- detection of bcr-abl transcripts, and histologic analysis. tine diagnostic blood studies or bone marrow (BM) aspir- ates from patients with newly diagnosed or previously diag- nosed and treated chronic myeloid leukemias at Memorial Analysis of tissue from xenografted animals Sloan-Kettering Cancer Center (MSKCC). Seventeen CP- Histopathology: Tissue sections from killed SCID mice CMLs, two AP-CMLs and eight BP-CMLs were studied. were fixed in 10% neutral-buffered formalin, dehydrated, Blast-enriched mononuclear cells were isolated by Ficoll– embedded in paraffin, sectioned and stained according to Hypaque density gradient separation and washed in RPMI standard histologic techniques. 1640 medium. After separation, most of the leukemic cells were freshly inoculated into SCID mice. In some patients, Immunophenotyping: Cell suspensions derived from leukemic cells were cryopreserved in liquid nitrogen before tissues of killed animals and primary CML cells were their injection into the animal. washed in phosphate-buffered saline (PBS) and stained with directly labeled fluorescein isothiocyanate (FITC)- SCID mice conjugated and phycoerythrin (PE)-conjugated monoclonal antibodies for 30 min at 4°C. The cells were then washed All SCID (CB17-SCID/SCID and NOD-SCID) mice were twice and analyzed by FACS (Becton Dickinson, Mountain purchased from Taconic Farms (Germantown, NY, USA) View, CA, USA). The following antibodies were used for and maintained in micro-isolator cages under sterile con- characterization of lineage surface markers: CD13, CD14, ditions with a specific pathogen-free environment without CD15, CD33, CD34, CD7, DR (Becton Dickinson). the use of any antibiotics. Female SCID mice between 6 Additionally, FACS analyses of samples of blood, BM, and 8 weeks of age were used. This work was approved spleen and liver from animals were performed by using by the MSKCC Animal Review Board and all animals were two-color immunofluorescence staining and pairwise com- sacrificed in accordance with MSKCC Animal Research binations of a selected panel of antibodies (CD3-PE, CD19- regulations. PE, CD33-PE and CD45-FITC) against human antigens and an antibody mCD45-FITC (Boehringer Mannheim, Indian- Inoculation of human leukemic cells into SCID mice apolis, IN, USA) against a murine antigen to discern the origins of the cells. A median of 2.0 × 107 (range 0.2–10 × 107) viable leukemic cells were resuspended in 200 ␮l of ice-cold Matrigel FISH: FISH was performed as previously described.9 In matrix (Becton Dickinson Labware, Bedford, MA, USA) brief, single-cell suspensions from tumor nodules or organs liquid and injected subcutaneously with a tuberculin were washed and fixed in a mixture of methanol and acetic syringe into the right flank of the SCID mice. acid (3:1 v/v). The cells were put on to slides and air-dried. The number of SCID mice inoculated and the cell dose Subsequently, the slides were washed at room temperature per mouse were dependent on the number of leukemic cells (RT) in successive jars containing 0.1 n HCl solution con- available from patient samples. The animals did not receive taining 0.5% Triton, PBS, 1% formaldehyde, PBS × 2, any conditioning treatment before inoculation. Thirty 2 × SSC and denatured at 72°C in 70% form- microliters of anti-Asialo GM1 (Wako Pure Chemical amide/2 × SSC. Thereafter, the slides were dehydrated by Industries, Osaka, Japan) were administered to CB-17 2 min of exposure to graded ethanol solutions at 4°C. SCID mice intraperitoneally on days −1 (prior to the s.c. Denatured probes (10 mg/ml biotin-labeled; Oncor, Gai- inoculation), 4 and 8, and every 5–7 days thereafter for the thersburg, MD, USA) in a hybridization solution (Hybrisol duration of the study.18 Treatment of Mouse NK cells have VI; Oncor) were deposited on slides and covered with a the immunophenotype asialo GM1+, NK-2.1+, Thy-1.2−, coverslip. The slides were sealed with rubber cement, and L3T4−, and Lyt-2−, and are thought to play an important then the hybridization took place in a moistened chamber role in endogenous in vivo resistance against transplanted at 37°C for 12–18 h. After hybridization, the slides were cells and this resistance has been overcome in SCID mice washed in 65% formamide/2 × SSC at 43°C and then in by the administration of anti-Asialo GM1.19 Leukemic cell 2 × SSC at 37°C and rinsed in 1 × SSC RT. Biotinylated CML growth in SCID mice J McGuirk et al 369 probes for human X chromosome were detected by apply- Table 1 Characteristics of patients with CML and their leukemic cells ing 20 ml detection reagent containing avidin/FITC engraftment and growth pattern in SCID mice (Oncor), the plastic coverslip was added, and the slide was incubated in a humidified chamber at 37°C for 30 min. The No. Age Disease Blasts (%) No. of cells Engraftment × (year)/Sex status in sample inoculated in SCID slides were then washed in 1 SSC at RT three times and ×107 propidium iodide (2 mg/ml; Sigma, St Louis, MO, USA) was applied for counterstain. A Zeiss fluorescence micro- CML(BC) scope (Axioskop; Carl Zeiss, Pelham, NY, USA) was used BC1 35/M BC (rel) PB (59) 2.0 3/3 for visualization and at least 200 cells were scored. BC2 18/M BC (rel) PB (67) 2.0 2/2 BC3 15/M BC (rel) PB (90) 2.0 4/4 BC4 30/M BC (rel) BM (53) 1.0 6/6 PCR: To detect bcr-abl transcripts, RNA was prepared BC5 53/M BC (rel) PB (65) 0.2 1/1 from blood, BM, liver, spleen, and tumor cells. To prepare BC6 45/M BC (new) PB (77) 2.0 1/1 the cDNA, between 2 and 5 ␮g of total RNA was added BC7 40/M BC (new) PB (33) 2.0 2/3a to a reverse transcriptase reaction. BC8 32/M BC (new) PB (83) 2.0 0/4 A nested PCR technique was used, as previously CML(AP) described. PCR product was run on an ethidium bromide- AC1 38/F AC (new) PB (17) 2.0 2/2a stained 2% agarose gel and visualized under UV light. AC2 16/F AC (rel) BM (13) 1.0 3/3a Using this approach, we were able to detect a single con- AC2n 1.0 2/2a trol-positive (K562) cell in 106 normal cells. In order to CML(CP) minimize contamination, the recommendations of Kwok CP1 54/F CP BM (5) 1.0 1/1a 19 and Higushi were adopted. CP2 59/M CP BM (9) 1.0 2/3a CP3 48/M CP PB (14) 2.0 3/4a Cytogenetic analysis: Samples of cells recovered from leu- CP4 39/M CP PB (5) 6.0 1/1a CP5 49/F CP (new) BM (13) 10.0 4/4a kemic nodules and from patient-derived tissues were exam- CP5a 10.0 4/4a ined for their karyotype by standard techniques. A mini- CP5n 5.0 3/3a mum of 20 metaphases was examined for each sample. CP6 26/M CP (new) BM(12) 10.0 4/4a CD7 34/M CP (new) BM(8) 5.0 2/2a CP7n 5.0 2/2a CP8 44/M CP PB (4) 1.0 0/2 Results CP9 37/M CP PB (10) 1.0 0/4 CP10 63/M CP PB (5) 1.0 0/1 Characteristics of patients CP11 11/M CP PB (4) 2.0 0/1 CP12 42/F CP PB (10) 1.0 0/1 Table 1 summarizes the clinical characteristics of the 27 CP13 34/M CP BM (4) 1.0 0/2 patients with CML evaluated in this study. The ages of CP14 48/M CP BM (10) 1.0 0/1 CP15 35/M CP BM (8) 2.0 0/2 patients ranged from 11 to 63 years. Patients were classified CP16 45/M CP BM (6) 2.0 0/2 according to CML phase: CP-CML (17 cases), AP-CML CP17 59/M CP PB (9) 2.0 0/2 (two cases), BP-CML (eight cases). One and three of the patients with AP-CML and CP-CML, respectively, were aTemporary engraftment. newly diagnosed and untreated. new = newly diagnosed; rel = relapse; n = nod mouse; a = Anti-Asialo GM1.

Engraftment and growth pattern of BP-CML in SCID mice and liver by FACS analysis. Leukemic cells from patient Leukemic cells from seven of eight patients with BP-CML BC8 failed to either engraft as an s.c. nodule or disseminate (87.5%) were able to grow as s.c. nodules in SCID mice to distant organs in SCID mice recipients. (Table 1). The mean time for leukemia nodule growth to reach 0.8 cm2 (diameter = 1.0 cm) and 2.0 cm2 = Engraftment and growth pattern of BC-CML in (diameter 1.6 cm) of surface area in mice bearing these secondary SCID mice passages seven BP-CML was 9.4 (Ϯ 4.9) and 15.6 (Ϯ 5.8) weeks after inoculation, respectively (Figure 1a). Leukemic cells After adoptive growth of leukemic samples derived from from five of seven (BC6 has not yet been sacrificed) patients BC1, BC2, BC3 and BC4 in SCID mice, specimens patient-derived samples that grew as subcutaneous nodules were obtained from leukemic s.c. tumors and passaged in disseminated to and infiltrated extensively the BM, periph- secondary SCID mice by s.c. inoculation using the same eral blood (PB), spleens and livers of the animals as amount of cells and same methodology as utilized during assessed by FISH, FACS and PCR analysis (Table 2). Leu- the first passages. All four specimens generated s.c. tumor kemia dissemination was also found in lung, kidney, brain growth in the secondary SCID mice. The blast cells derived and spleen of the animals on pathologic examination. One from two patients (BC1 and BC3) displayed a similar patient-derived specimen (BC7) grew as a small s.c. tumor growth and dissemination pattern when compared with and demonstrated minimal systemic dissemination, their leukemia growth in the first SCID passages (Figure 2 although there was gross liver and splenic enlargement, and Table 2). However, leukemic cells derived from only rare leukemic cell detection was demonstrated in BM patients BC2 and BC4 displayed more rapid leukemic sub- CML growth in SCID mice J McGuirk et al 370 a 10 aneous nodules without subsequent dissemination to distant BC1 9 organs (Table 1). Leukemic cells derived from patient AC1 BC2 ) 8 gradually generated an s.c. tumor with a maximal size of 2 BC3 2 7 0.5 cm approximately 5–6 weeks after inoculation, there- BC4 6 after, the tumor size slowly regressed (Figure 1b). BC5 Additionally, samples derived from patient AC2 grew as 5 BC6 s.c. tumors in both CB17 and NOD-SCID mice. The time 4 BC7 for the leukemia nodule to reach 0.8 cm2 in CB17 and NOD BC8 3 mice was 14.5 weeks and 7 weeks, respectively, however, 2 the tumor sizes in these animals did not exceed 1 cm2. Sev- 1 enteen and 12 weeks after inoculation, the tumors began to 0 spontaneously regress in the CB17 and NOD-SCID mice, 0 5 10 15 20 25 30 35 40 b 2 respectively (Figure 3). In none of these animals were spleen or liver enlargement or other clinical signs that indi- AC1 cated leukemia development evident. Furthermore, there AC2 ) Surface area (cm

2 was no evidence of leukemic cell infiltration of distant AC2n organs detectable by FISH, FACS, histopathological or RT- PCR analysis of the tissues performed at 5–10 weeks 1 post inoculation.

Engraftment and growth pattern of CP-CML in SCID mice 0 Samples derived from 17 CP-CML patients were included 0 5 10 15 20 25 30 35 40 c in this study. Leukemic cells from seven of 17 patients 2 (CP1–7) grew transiently as subcutaneous nodules without CP1 resultant dissemination to distant organs as determined by

)CP2 Surface area (cm FISH, FACS and histopathological analysis (Tables 1 and 2 2 CP3 2). No leukemic nodule reached a size of 0.8 cm in the CP4 mice bearing these seven CP-CML samples (Figure 1c), CP6 1 with the exception of leukemic cells derived from patient CP5 and inoculated into NOD-SCID mice (CP5n) which generated a maximal tumor size of 0.8 cm2 5 weeks after inoculation (Figure 3). Three of these seven patient-derived Surface area (cm samples were obtained from newly diagnosed patients (CP5, CP6, CP7) who had not yet received any treatment. 0 The s.c. nodules resultant from the seven inoculated speci- 0 5 10 15 20 25 30 35 40 mens which generated tumors all spontaneously regressed Weeks post-inoculation starting at approximately 5 weeks. Subcutaneous tissue Figure 1 The curves of subcutaneous growth of different phases of analysis by FACS, FISH and histopathology at the site of patient-derived CML cells in SCID mice. (a) Leukemic blasts from BP- tumor cell injection revealed minimal or no residual tumor CML patients BC1, BC2, BC3, BC4 displayed an aggressive growth pat- cells present at the time of animal sacrifice. RT-PCR analy- tern, whereas cells from patient BC5 grew as a small tumor with slow sis for the bcr-abl fusion gene transcription product in dif- regression and cells from patient BC6 did not induce tumor growth. (b) Leukemic blasts from AP-CML patients AC1 and AC2 initiated prompt ferent organs in the mice inoculated with higher numbers of tumor growth with subsequent spontaneous regression. Furthermore, injec- leukemic cells derived from newly diagnosed and untreated tion of cells from patient AC2 into NOD-SCID mice (AC2n) did not dem- patients (CP5, CP6, CP7) indicated that minimal dissemi- onstrate a growth pattern different from that observed in CB17 SCID mice nation to distant organs of leukemic cells occurred in these (AC1,AC2). (c) Samples derived from patients CP1, CP2, CP3, CP4, CP6 also demonstrated temporary engraftment as subcutaneous tumors with a animals (Table 2). rapid subsequent spontaneous regression. Note: y-axis for a differs from Furthermore, analysis for the bcr-abl transcript by RT- 2 b and c. Tumor size is the mean surface area (cm ). PCR analysis with primer B2A2 revealed positive findings in the s.c. tumor and liver in CB17 SCID mice (CP6a) sac- rificed at 5 weeks, whereas animals sacrificed at 7 weeks cutaneous tumor growth in the secondary SCID mice in (CP6b) displayed positive findings in bone marrow and contrast to the first passage (Figure 2). liver with no positive findings in the tumor. None of the tissues from these animals revealed positive RT-PCR find- Engraftment and growth pattern of AP-CML in SCID ings when the control primer B3A2 was utilized for leukemic cell detection. mice With the exception of the PCR results, there was no sig- Leukemic cells derived from two AP-CML patients were nificant difference in s.c. tumor growth characteristics inoculated into SCID mice. The leukemic cells from two between samples obtained from either newly diagnosed or of two patients with AP-CML grew transiently as subcut- previously treated patients with CP-CML. Ten CP-CML CML growth in SCID mice J McGuirk et al 371 Table 2 Engraftment and dissemination of human CML leukemic cells in SCID mice

No. Time of (%) Human cells FISH PCR sacrifice (week) Histopathology Flow cytometry TU PB BM SP LI TU PB BM SP LI

BM SP LI KI BR TU PB BM SP LI

CML blast crisis BC1 31(S) +++++ 90 50 21 15 63 78 48 8 10 82 +++++ BC1b 30(D) ND ND ND ND ND 95 42 35 26 57 ND ND ND ND ND +++++ BC2 21(S) −++−− 96 29 2 38 55 98 29 2 0 16 + ND +++ BC2b 12(S) −++−− 92 52 15 6 42 ND ND ND ND ND +++++ BC3 22(S) +++++ 98 27 89 12 76 96 7 29 5 32 + ND +++ BC3b 24(D) +++++ 99 42 75 17 82 ND ND ND ND ND ND ND ND ND ND BC4 36(D) +++−− 96 18 4 12 11 96 36 4 10 2 +++++ BC5 23(S) +++−− 98 21 11 9 35 ND ND ND ND ND NE ND ND ND ND BC7 43(S) −−−−− 7 Ͻ12Ͻ12 141400 +−+ND ND BC8 40(S) −−−−−NG Ͻ1 Ͻ1 Ͻ1 Ͻ1NG0010 NG −−−−

CML accelerate phase AC1 38(S) −−−−− 7 Ͻ10 0Ͻ1 50000 +−+++ AC2 30(S) −−−−− 12 Ͻ1 Ͻ1 Ͻ1 Ͻ1ND0000 ND −−−− AC2n 40(S) −−−−−NDNDNDNDND00000 −−−−−

CML chronic phase CP1 12(S) −−−−− 40Ͻ1 Ͻ10 80000 ND −−−− CP2 21(S) −−−−− 0 Ͻ1 Ͻ100 00000 −−−−− CP3 18(S) −−−−− 000Ͻ10 00000 NDNDNDNDND CP4 40(S) −−−−− 0 Ͻ1000 00000 −−−−− CP5 21(S) ND ND ND ND ND ND ND ND ND ND 00000 −−−−− CP5a 24(S) ND ND ND ND ND ND ND ND ND ND 00000 −−−−− CP5n 11(S) −−−−− 250000 120000 +−+−− CP6 5(S) −−−−−ND Ͻ1 Ͻ1 Ͻ1 Ͼ1830000 +−−+− CP7 4(S) −−−−−ND0000 370000 +−−+− CP7n 4(S) −−−−−ND0000 790000 +−−−− CP8 20(S) −−−−−NG Ͻ1 Ͻ10Ͻ1NG0000 NG −−−− CP9 11(S) −−−−−NG 0 0 0 0 NG ND ND ND ND NG −−−− CP10 13(S) −−−−−NG ND ND ND ND NG 0 0 0 0 NG −−−− CP11 7(D) −−−−−NG 0 0 Ͻ10 NG0 0 0 0 NGNDNDNDND CP12 12(S) −−−−−NG 0 Ͻ100NG0000 NG −−−− CP13 22(S) −−−−−NG Ͻ10 0Ͻ1NG0000 NG −−−− CP14 13(S) −−−−−NG ND ND ND ND NG ND ND ND ND NG ND ND ND ND CP15 19(S) −−−−−NG Ͻ10Ͻ1 Ͻ1NGNDNDNDNDNGNDNDNDND CP16 10(S) −−−−−NG0000NG0000 NG −−−− CP17 12(S) −−−−−NG0000NG0000 NG −−−−

ND = not done; NG = no growth; S = sacrifice; D = died; TU = tumor; BM = bone marrow; PB = peripheral blood; SP = spleen; LI = liver; KI = kidney; BR = brain; n = nod scid mouse.

(CP8–17) samples, all derived from patients previously while CP7n mice demonstrated positive findings in the treated, failed to grow as subcutaneous nodules or to tumor only; none of the animals demonstrated positive disseminate to distant organs. findings when the control primer B2A2 was utilized for To address the possibility of leukemia rejection by leukemic cell detection. mouse NK cells, we compared the leukemic cell engraftment and growth in CB17 SCID mice treated with Cytogenetic and phenotypic analysis of recovered Anti-Asialo GM1 or NOD-SCID mice with leukemic cell leukemic cells from mouse tissues growth in CB17 SCID mice without any treatment. There were no discernible differences in subcutaneous tumor Cytogenetic analysis of leukemic cells recovered from leu- growth in CB17 SCID mice treated with Anti-Asialo GM1 kemic nodules growing in SCID mice inoculated with BP- (CP5a), NOD-SCID mice (CP5n and CP7n) and untreated CML demonstrated significant evolutionary changes CB17 SCID mice (CP5 and CP7) (Figure 3). including chromosomal duplications and additional com- As mentioned, sample CP7 induced subcutaneous tumor plex translocations (Table 3). Phenotypically, the leukemic growth in both CB17 SCID (CP7) mice and NOD-SCID cells recovered from leukemic nodules induced by speci- (CP7n) mice and there was no demonstrable difference in mens BC1 and BC2 demonstrated an evolution to a less nodule growth characteristics or dissemination patterns. mature phenotype than was predominant in the original RT-PCR analysis for bcr-abl breakpoint B3A2 revealed samples, however, leukemic cells recovered from leukemic positive findings in the spleen and tumor of CP7 mice, nodules induced by specimen BC3 demonstrated a more CML growth in SCID mice J McGuirk et al 372 a b 8 8

7 7

6

) 6 ) 2 2 BC2(P2) BC2(P1) 5 BC1(P1) BC1(P2) 5

4 4

3 3 Surface area (cm Surface area (cm

2 2

1 1

0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35

8 8

7 7

6 BC3(P1) ) 6 ) 2 BC3(P2) 2 5 5

BC4(P2) 4 4

3 3 Surface area (cm Surface area (cm BC4(P1) 2 2

1 1

0 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 Weeks post-inoculation Weeks post-inoculation

Figure 2 Differential growth of BP-CML-induced s.c. tumors in SCID mice comparing initial and secondary passage of leukemic cells derived from patients or subcutaneous tumors, respectively. (a) Primary (P1) and secondary (P2) passage of samples BC1 and BC3 displayed similar growth patterns. (b) In contrast, secondary passage of tumor cells derived from patients BC2 and BC4 resulted in more rapid tumor growth and greater maximal tumor size (mean surface area (cm2)) than occurred with initial passage of cells from the patients.

mature phenotype than was predominant in the original tumor harvest. Additionally, second passage of BP-CML sample (Table 4). The leukemic cells recovered from cells (BC2, BC4) derived from the subcutaneous tumors of leukemic nodules in animals inoculated with sample BC4 first passage SCID mice demonstrated a more rapid growth displayed the same phenotype as the original sample. pattern and reached a greater maximal tumor size than was found in the first passage mice. This may have represented the selective growth of a leukemic population of cells not Discussion detected in the initial cytogenetic analysis, or the original leukemic cells may have undergone an evolution to more In this study we have demonstrated for the first time that complex chromosomal abnormalities expressed as the dem- BP-CML cells, like primary acute leukemias,9 can engraft onstrated cytogenetic and phenotypic abnormalities. The after s.c. inoculation in SCID mice without conditioning latter explanation would be consistent with previously treatment or the administration of growth-promoting cyto- described evidence for a multi-step pathogenesis in the kines. Moreover, the xenografted tumor cells subsequently progression of CP-CML to BP-CML.5,20 disseminate to distant organs in a pattern similar to that We observed distinct differences in the growth character- seen in the patients from whom they were derived. istics of leukemic cells derived from patients in BP-CML BP-CML cells demonstrated remarkable cytogenetic and as compared to CP-CML or AP-CML. Seven of eight BP- phenotypic changes from preimplantation to the time of CML samples grew in a manner analogous to human AML CML growth in SCID mice J McGuirk et al 373 2 and ALL in the SCID mice.9 However, CP-CML and AP- CP5 CML only grew either transiently as s.c. tumors which CP5a spontaneously regressed, or not at all. We did not observe )

2 CP5n a difference between the ability of samples derived from CP7 either newly diagnosed and untreated patients vs patients CP7n already receiving treatment. Although subcutaneous tumor 1 cells generated from CP-CML samples initially demon- strated bcr-abl transcripts by RT-PCR analysis, animals sacrificed at a later date demonstrated disappearance of the

Surface area (cm positive RT-PCR signal initially found in these tumor cells, consistent with the morphologic regression of these tumors. The differential growth pattern described here is similar to 0 that previously demonstrated in SCID mice utilizing the 0 5 10 15 20 25 30 35 40 13 Weeks post-inoculation renal capsule as the site of injection. Although it has been demonstrated in several studies that Figure 3 Growth curves of CP-CML samples derived from patients CP5 CML cell lines and patient-derived BP-CML cells, utilizing and CP7 and injected into CB17 SCID mice (CP5 (n = 4), CP7 (n = 2)), varying routes of administration, can grow and disseminate CB17 SCID mice treated with Anti-Asialo GM1 (CP5a (n = 4)), and NOD- SCID mice (CP5n (n = 3), CP7n (n = 2)). No significant difference either in SCID mice, s.c. growth of patient derived BP-CML has 13–16 in rate of tumor growth or maximal tumor size (mean surface area) reached not yet been reported. Additionally, no reliable and was detected between these groups of mice. consistent in vivo model for the growth of CP-CML has been developed. Sirard et al17 have recently reported growth of normal and leukemic cells in the bone marrow of sublethally irradiated SCID mice after intravenous Table 3 Karyotypes of leukemic cells before and after engraftment infusion of patient-derived CP-CML samples. However, in into SCID mice only a small proportion of marrows growing human hema- topoietic cells was the presence of the bcr-abl transcript Patients Primary leukemia Leukemic cells recovered from identified. Additionally, in this study the CP-CML cells did SCID mice not disseminate in a manner analogous to the pattern seen in humans. BC1 46,XY,t(9;22) (q34;q11) 46,XY,t(9,22)(q34;q11),+16 It has been suggested that there may be a relationship BC2 48,XY,+8,t(9;22) 56,XY,+Xx2,+Y,der(2)t(1;2) between the inoculated cell dose of patient-derived CML (q34;q11),+der(22)t(9;22) (q21;q37),+3,+6,+7,+8, samples and the subsequent growth of both leukemic cells (q34;q11) t(9;22) (q34;q11),+der(22) 17 + + and normal hematopoietic cells in SCID mice. In this t(9;22) (q34;q11), 10, 19 study we utilized a median cell dose of 2 × 107 and may BC3 49,XY,+8t(9;22) 48,XY,dup(1) (q23q42),+8, have achieved better engraftment with higher cell doses. (q34;q11),+10,+der(22) t(9;22) (q34;q11),+der(22) However, even when CP-CML tumors grew as subcutane- t(9;22) (q34;q11) t(9;22) (q34;q11) ous nodules there was a subsequent spontaneous and + 50,XY, 8 complete tumor regression. BC4 46,XY,t(9;22) (q34;q11) 46,XY,t(6;14) (p25;q11.1) It has been postulated that intact NK cell function may t(7;9) (p13;q34),t(11;16) contribute to the lack of growth of CP-CML cells in SCID (q25;q11.2) mice.17 We did not observe any growth difference in CB17- SCID mice treated with an anti-NK cell antibody when compared to untreated CB17 SCID mice. Additionally, we were unable to propagate CP-CML cells in NOD-SCID (NK-deficient) mice. Our results suggest that factors other Table 4 Immunophenotype of CML blast cells before and after than NK cell activity are responsible for the lack of growth engraftment into SCID mice and propagation of CP-CML in SCID mice. One possible explanation for the observed differential Patients MoAb positive staining cells (%) growth is that, although propagation of BP-CML cells occur by autocrine mechanisms, continued growth of CP- CD13 CD14 CD15 CD33 CD34 CD7 DR CML or AP-CML may require stromal components or microenvironmental factors depleted or absent in the Fic- BC1 65a 1 ND 84 9 22 48 71b 18896453550 oll–Hypaque density fraction. It has been reported that BC2 58 23 86 99 64 28 90 abnormalities in the CML marrow microenvironment 15 1 9 82 94 24 2 related to the presence of malignant stromal macrophages BC3 5 1 45 59 66 7 47 may contribute to the selective expansion of leukemic pro- 96 1 99 99 68 5 19 genitors and suppression of normal hematopoiesis in BC4 87 3 ND 80 83 87 27 21 85 Ͻ11089898238 CML. Moreover, successful engraftment of SCID mice with primitive CP-CML progenitor cells has been reported aPatient-derived sample. after the implantation of human fetal bone fragments, pre- bCell recovered from SCID mice. sumably containing all the requisite stromal components.22 CML growth in SCID mice J McGuirk et al 374 Additional manipulations of the SCID mouse model to defect at the leukemic stem cell level. Proc Natl Acad Sci more closely approximate the human CML marrow USA 1992; 89: 6192–6196. microenvironment may be required to achieve durable 9 Yan Y, Salomon O, McGuirk J et al. Growth pattern and clini- transplantion of CP-CML cells. cal correlation of subcutaneously inoculated human primary In conclusion, the subcutaneously inoculated SCID acute leukemias in severe combined immunodeficiency mice. Blood 1996; 88: 3137–3146. mouse leukemia model appears to accurately reflect the 10 Uckun FM, Sather H, Reaman G et al. Leukemic cell growth intrinsic biologic behavior of BP-CML. The subcutaneous in SCID mice as a predictor of relapse in high-risk B-lineage engraftment and resultant dissemination of BP-CML in this acute lymphoblastic leukemia. Blood 1995; 85: 873–878. model may provide an important tool for the study of CML 11 Uckun FM, Downing JR, Chelstrom LM et al. Human biology and the development of novel therapeutic strategies t(4:11)(q21:q23) acute lymphoblastic leukemia in mice with such as immunotherapeutic manipulations or pharmocol- severe combined immunodeficiency. Blood 1994; 84: 859– ogic agents such as angiogenesis inhibitors without necessi- 865. tating the sacrifice of the animals to observe an effect. 12 Uckun FM, Downing JR, Gunther R et al. Human Further refinements of this model will be necessary for the t(1:19)(q23:3) pre-B acute lymphoblastic leukemia in mice establishment of a reliable in vivo model for the study of with severe combined immunodeficiency. Blood 1993; 81: 3052–3062. CP-CML. 13 Sawyers CL, Gishizky ML, Quan S et al. Propagation of human blastic myeloid leukemias in the SCID mouse. Blood 1992; 79: 2089–2098. Acknowledgements 14 Beran M, Pisa P, O’Brien S et al. Biological properties and growth in SCID mice of a new myelogenous leukemia cell line This work was supported by grants from the National Institutes (KBM-5) derived from chronic myelogenous leukemia cells in of Health (No.CA23766), the Lisa Biloti Foundation, the Andrew the blastic phase. Cancer Res 1993; 53: 3603–3610. Gaffiney Foundation, the Zelda Radow Weintraub Cancer Foun- 15 Cesano A, Hoxie JA, Lange B et al. The severe combined dation, and the Vincent Astor Chair in Clinical Research Fund. immunodeficient (SCID) mouse as a model for myeloid leuke- mias. Oncogene 1992; 7: 827–836. 16 Skorski T, Nieborowska-Skorska M, Wlodarski P et al. Anti- sense oligodeoxynucleotide combination therapy of primary References chronic myelogenous leukemia blast crisis in SCID mice. Blood 1996; 88: 1005–1012. 1 Fialkow PJ, Jacobsenn RJ, Papayannopolou T. Chronic 17 Sirard C, Lapidot T, Vormoor J et al. Normal and leukemic myelogenous leukemia: Clonal origin in a stem cell common SCID-repopulating cells (SRC) coexist in the bone marrow to the granulocyte, erythrocyte, platelet and mono- and peripheral blood from CML patients in chronic phase, cyte/macrophage. Am J Med 1977; 63: 125–130. whereas leukemic SRC are detected in blast crisis. Blood 2 de Klein A, van Kessel AG, Grosfeld G et al. A cellular onco- 1996; 87: 1539–1548. gene (c-abl) is translocated to the Philadelphia chromosome 18 Lacerda J, Ladanyi M, Jagiello C et al. Administration of rab- in chronic myelocytic leukemia. Nature 1982; 294: 765–767. bit Anti-Asialo GM1 antiserum facilitates the development of 3 Heisterkamp N, Stephenson JR, Groffen J et al. Localization human Epstein–Barr virus-induced lymphoproliferations in of the c-abl oncogene adjacent to a translocation break point xenografted C.B.-17 SCID/SCID mice. Transplantation 1996; in chronic myelocytic leukemia. Nature 1983; 306: 239–242. 61: 492–497. 4 Daley GQ, Van Etten RA, Baltimore D. Induction of chronic 19 Kwok S, Higushi R. Avoiding false positives with PCR. Nat- myelogenous leukemia in mice by the p210 bcr/abl gene of ure 1989; 339: 237–238. the Philadelphia chromosome. Science 1990; 247: 824–830. 20 Fialkow PH, Martin PJ, Najfeld V et al. Evidence for a multi- 5 Kantarjian HM, Deisseroth A, Kurzrock R et al. Chronic mye- step pathogenesis of chronic myelogenous leukemia. Blood logenous leukemia: a concise update (review). Blood 1993; 1981; 58: 158–163. 82: 691–703. 21 Bhatia R, McGlave PB, Dewald GW. et al. Abnormal function 6 Wetzler M, Talpaz M, Estrov Z et al. CML: mechanisms of of the bone marrow microenvironment in chronic myelogen- disease initiation and progression (review). Leuk Lymphoma ous leukemia: role of malignant stromal macrophages. Blood 1993; 1: 47–50. 1995; 85: 3636–3645. 7 Eaves C, Cashman J, Zoumo N et al. Biological strategies for 22 Van den Berg D, Wessman M, Murray L et al. Leukemic the selective manipulation of normal and leukemic stem cells. burden in subpopulations of CD34+ cells isolated from the Stem Cells 1993; 11: 109–121. mobilized peripheral blood of ␣-interferon-resistant or -intol- 8 Udomsakdi C, Eaves CJ, Swolin B et al. Rapid decline of erant patients with chronic myeloid leukemia. Blood 1996; 87: chronic myeloid leukemic cells in long-term culture due to a 4348–4357.