|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
Prepublished online as a Blood First Edition Paper on December 19, 2002; DOI 10.1182/blood-2002-06-1800.
NEOPLASIA
From the Comprehensive Cancer Center and the
Department of Laboratory Medicine, University of California, San
Francisco; Section of Hematology/Oncology, University of Chicago, IL;
Division of Hematology/Oncology, Brigham and Women's Hospital and
Howard Hughes Medical Institute, Harvard Medical School, Boston, MA;
and Sugen Inc, South San Francisco, CA.
The PML-RAR Acute promyelocytic leukemia (APL) is defined by
the presence of a PML-RARA fusion gene or, rarely, by
alternative RARA fusions.1 PML-RARA
fusion is created as a result of a chromosomal translocation, t(15;17)(q22;q12).2 RARA encodes retinoic acid
receptor The FMS-like tyrosine kinase 3 (FLT3 [FLK-2, STK-1]) receptor
enhances the proliferation and survival of hematopoietic progenitors in
response to FLT3 ligand.11,12 Activating mutations of FLT3 have been identified in 25% to 30% of human patients with AML and are
common in APL (18%-37% of APL patients).13-17 Many
activating mutations of FLT3 are the result of an internal tandem
duplication (ITD) of the intracellular juxtamembrane region of this
molecule.18 These ITD mutations result in constitutive
tyrosine kinase activity.19,20 It is of interest that
though activated tyrosine kinases are often associated with human
chronic myeloproliferative diseases (MPDs), commonly
BCR-ABL,1 and infrequently such fusions as
TEL-PDGFBR,21 HIP1-PDGFBR,22
D10S170-PDGFBR,23,24 and
RAB5EP-PDGFBR,25 activating mutations
of FLT3 have not yet been similarly identified in human MPD.
Genetic manipulation of mice is a powerful tool for understanding the
contributions that particular genetic changes make to malignant
transformation. Studies of PML-RARA transgenic mice revealed
that PML-RAR Given the clinical efficacy of the selective tyrosine kinase inhibitor,
imatinib mesylate, to treat BCR-ABL- and
TEL-PDGFBR-positive chronic myeloproliferative
diseases,32,33 we investigated whether the inhibition of
activated FLT3 would affect the growth of acute leukemias that
expressed PML-RAR Mice
Retroviral transductions
Peripheral blood counts and bone marrow differential counts Blood was obtained from the retro-orbital sinuses of anesthetized mice. White blood cell (WBC) count, hemoglobin level, and platelet count were measured with the Hemavet 850 FS cell counter (CDC Technologies, Oxford, CT). Bone marrow was flushed from the tibias and femurs of mice with buffered saline supplemented with 2% heat-inactivated fetal bovine serum (FBS) and 2.5% cell dissociation buffer (Invitrogen, Carlsbad, CA). Blood smears, marrow smears, and marrow and spleen cytospins were prepared according to standard hematologic techniques and were stained with Wright-Giemsa stain. Peripheral blood differential WBC counts (200 cells each) and bone marrow differential counts (400 cells each) were in accord with published guidelines.40Histopathology Sternum, liver, spleen, kidney, lungs, lymph nodes, and intestine were initially fixed in either Bouin fixative or a buffered formalin solution before embedding in paraffin. Sternums fixed in formalin were decalcified for 3 hours before embedding (11% formic acid, 8% formaldehyde). Sections were prepared according to standard protocols and were stained with hematoxylin and eosin.Cytochemical stains Suspensions of viable bone marrow and spleen cells were depleted of red blood cells with the use of Histopaque 1119 (Sigma, St Louis, MO) at 4°C according to the manufacturer's instructions. Cytospins of bone marrow and spleen were prepared by standard technique. Cytospins were stained with chloroacetate esterase or alpha-naphthyl acetate esterase according to the manufacturer's instructions (Sigma).Immunophenotyping For flow cytometric immunophenotyping, 300 000 bone marrow or spleen cells were suspended in 100 to 200 µL buffered saline (with 2% heat-inactivated FBS and 2.5% cell dissociation buffer). Cells were incubated on ice with unlabeled anti-CD16/CD32 antibodies (Fc block) for 15 minutes before the addition of antibodies (obtained from Cal-Tag Laboratories, Burlingame, CA or BD PharMingen, San Diego, CA). Phycoerythrin, Tri-Color, or biotin-conjugated antibodies to mouse CD3, CD11b (Mac-1), CD19, CD31 (PECAM), CD34, CD61, CD86, CD117 (c-kit), Ly-6G (Gr-1), Ly-71 (F4/80), or Ly-76 (Ter119) were added and incubated with the cells for 20 minutes in the dark on ice. Biotin antibodies were subsequently washed and incubated with streptavidin-allophycocyanin. Cells were washed with buffered saline and resuspended in a final volume of 200 µL. Stained cells were analyzed on a FACScan or FACScalibur (Becton Dickinson), and at least 10 000 events were collected for each sample. FACS data were analyzed with CellQuest (Becton Dickinson). We used forward-scatter (FSC), side-scatter (SSC), GFP, and CD45 (when applicable) to select gated cells for analysis, as described in the legends to Figures 3 and 8.Western blotting Whole cell lysates were prepared by lysing 1 × 107 cells in 500 µL 2 × sample buffer, heating at 90°C to 95°C for 5 minutes, and shearing through a 20-gauge needle. Western blot analysis was performed as previously described8,41 using a rabbit polyclonal antiserum raised against a glutathione-S-transferase (GST)-fusion protein that encompassed amino acids 420 to 462 of the human RAR protein. FLT3
expression was detected using anti-FLT3 SC479 (Santa Cruz
Biotechnology, Santa Cruz, CA). Bands were visualized using the
enhanced chemiluminescence (ECL) reagent (Amersham Biosciences, Piscataway, NJ).
Karyotyping Cytogenetic and spectral karyotyping analysis of spleen cells was performed as described.30Southern blotting Southern blots of DNA derived from leukemic spleens was performed as described with a probe for retroviral DNA.31In vitro analysis of FLT3 phosphorylation The MV 4;11 human leukemia cell line, which expresses FLT3-ITD, was maintained as previously described.42 To assess FLT3 phosphorylation, cells were treated with SU11657 for 2 hours in medium containing 0.1% FBS and lysed as described.42 Equivalent amounts of protein from each sample were immunoprecipitated overnight at 4°C with an agarose-conjugated anti-FLT3 antibody (Santa Cruz Biotechnology). Immune complexes were washed (150 mM NaCl, 1.5 mM MgCl2, 50 mM HEPES [N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid], pH 7.5, 10% glycerol, 0.1% Triton X-100, and 1 mM EGTA [ethyleneglycotetraacetic acid]), separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose membranes. Membranes were probed with an antiphosphotyrosine antibody 4G10 (Upstate Biotechnology, Lake Placid, NY or Transduction Laboratories, Lexington, KY) and were stripped with Restore Western Blot Stripping Buffer (Pierce, Rockford, IL). Membranes were reprobed with an anti-FLT3 antibody.Transplantation of leukemic cells Leukemic cells isolated from the bone marrow and spleen of experimental animals were resuspended in buffered saline. After a sublethal (4.5 Gy) dose of irradiation, 1 × 106 cells in 100 µL buffered saline were injected into the tail veins of each FVB/N recipient mouse. Recipients of serially transplanted leukemia 1127 were treated with ATRA, SU11657, or both.Treatment of mice with ATRA and SU11657 Retinoic acid (10 mg, 21-day release) or placebo pellets (Innovative Research of America, Sarasota, FL) were implanted subcutaneously using a 12-gauge trochar. SU11657 was administered daily at 20 mg/kg/d in a carboxymethylcellulose (CMC) suspension by oral gavage. Mice were observed twice daily for the development of disease.Statistical analyses The Student t test (2-sided, unequal variance) was used except as otherwise noted.
Activated FLT3 accelerates the appearance of leukemia in MRP8 PML-RARA transgenic mice To test the ability of activated FLT3 to cooperate with PML-RAR
in leukemogenesis, we used retroviral transduction of bone marrow from
MRP8 PML-RARA transgenic mice. Bone marrow cells of control
and PML-RARA transgenic mice were transduced with control or
activated FLT3 retroviruses. These cells were subsequently introduced
into lethally irradiated recipient mice that were monitored for the
development of disease.
Recipients of PML-RARA transgenic marrow transduced with an
activated FLT3 retrovirus (FLT3W51, an ITD mutation from a
human patient) developed leukemia beginning at 62 days after
transplantation (median latency, 105 days; range, 62-299 days; Figure
1). Latency until disease was shorter
than that observed in recipients of PML-RARA marrow
transduced with control retroviruses (Figure 1) or in recipients of
PML-RARA marrow that had not undergone retroviral
transduction.28 Time to illness of activated
FLT3/PML-RAR
Although our transductions of bone marrows with the MIG retrovirus typically is associated with transduction efficiencies of approximately 30%, transduction efficiencies with the wild-type and FLT3W51 retroviruses were lower (2%-17%), as was observed in previous studies (Kelly et al31,43 and data not shown). These low levels of transduction were nevertheless associated with high penetrance of leukemia in recipients of activated FLT3-transduced PML-RARA bone marrow. Of note, in the moderate number of mice we studied, there was no correlation between transduction efficiency and latency of disease (r2 = 0.4; 12 leukemias from 6 independent transductions). Activated FLT3/PML-RAR recipients was an aggressive leukemia. When compared with healthy blood, the peripheral blood contained numerous immature forms/blasts, including occasional cells with deep
nuclear indentations similar to the bilobed forms that typify the
microgranular variant of human APL1 (Figure
2A-B). Similarly, in contrast to healthy
mice, bone marrows contained numerous immature forms/blasts (Table 2;
Figure 2C-D). The cytoplasm of these
cells contained azurophilic granules, but Auer rods were not seen. Many of the immature leukemic cells stained weakly to strongly with the
neutrophil marker chloroacetate esterase. A variable number of marrow
myeloid cells stained with the monocyte marker -naphthyl acetate
esterase, but the number of such cells did not exceed 20% of marrow
nonerythroid cells (data not shown). Flow cytometric immunophenotyping
of 7 leukemias demonstrated that leukemic cells expressed GFP, and it
corroborated morphology and cytochemistry in demonstrating the myeloid
character of the leukemias (Figure 3).
The leukemias expressed the combination of Ly-6G (Gr-1) and CD117
(c-kit), indicative of immature myeloid cells. Consistent with this was
the expression of CD61 (which in mice is not restricted to
megakaryocytic cells but is also expressed by granulocytes) and CD31
(PECAM). Cells lacked the lymphocyte markers CD3 and CD19 and the
erythroid marker Ly-76 (Ter119). There was some heterogeneity among the
leukemias: CD11b (Mac-1) varied from weak to moderate; Ly-6G (Gr-1)
expression varied from moderate to strong; and 2 leukemias expressed
low levels of Ly-71 (F4/80) and CD86. In addition to expressing GFP, a
surrogate marker for FLT3 expression, the leukemias expressed
PML-RAR and FLT3, as assessed by Western blot analysis (data not
shown).
The leukemias were aggressive, disseminated diseases effacing the
spleens and infiltrating nonhematopoietic tissues including liver,
kidneys, and lung (Figure 2E-J and data not shown). Spleens and livers
of leukemic animals were enlarged (mean weight of activated FLT3/PML-RAR Overall, the predominant disease that arose in activated
FLT3/PML-RAR The survival curve for activated FLT3/PML-RAR Activated FLT3/PML-RAR ,
either as an MRP8 transgene or by retroviral transduction of
PML-RARA transgenic bone marrow, accelerated leukemogenesis
and was associated with complex karyotypic abnormalities. In contrast,
though all the leukemias that developed in activated FLT3/PML-RAR
mice exhibited clonal karyotypic abnormalities (Table
3), the cytogenetically abnormal clones
were simpler than what we had seen in PML-RAR and BCL-2/PML-RAR
mice.30 The abnormalities observed were gain of chromosome
15 in 3 mice, gain of chromosome 10 in 3 mice, and gain of chromosome 7 in 2 mice. Six mice showed loss of an X chromosome. All 7 mice had a
single abnormal clone. Southern blotting of leukemic samples was
consistent with the cytogenetic results in demonstrating 1 or 2 predominant retroviral integration sites in the samples (data not
shown).
Activated FLT3 does not readily initiate myeloproliferative disease in FVB/N mice Previously, the FLT3W51 retrovirus we used in these studies was shown to induce myeloproliferative disease when BALB/c bone marrow was transduced using a protocol similar to that used in the current study.31 This MPD was characterized by increased numbers of myeloid cells with retained maturation, and it contrasted with the AMLs we observed in FLT3W51/PML-RARA mice in the FVB/N strain background. In contrast to what was observed in BALB/c mice, none of the recipients of control FVB/N bone marrow transduced with the FLT3W51 construct developed myeloproliferative disease.Five of 12 recipients of control bone marrow transduced with activated FLT3 did become ill. One such mouse became ill at 75 days because of intestinal distention, and other recipients appeared ill at 117, 168, and 308 days, at which times pathology examination did not reveal any specific abnormalities. Of particular interest was one animal that became ill 175 days after transplantation because of a thoracic T-cell malignancy (presumptive diagnosis, T-cell acute lymphoblastic leukemia/lymphoma) that showed strong expression of GFP. We also observed T-cell acute lymphoblastic leukemia lymphomas when activated FLT3 was introduced into a mixed B6 × C3H strain of mice.43 Remaining activated FLT3/control recipients do not show evidence of MPD in their peripheral blood 250 to 419 days after transplantation. Activated FLT3/PML-RAR and activated FLT3, we treated sublethally irradiated mice that had been injected with activated FLT3/PML-RAR leukemic cells. Mice received either placebo or ATRA pellets (10 mg,
21-day release). In one experiment, ATRA pellets were implanted 12 days
after the injection of leukemic cells. ATRA prolonged survival (median
prolongation of survival, 29 days; Figure
5A). In an ensuing experiment, treatment
was begun earlier (6 days after injection of leukemic cells).
Not surprisingly, the increase in survival with ATRA was longer when
therapy was begun on day 6 (median prolongation in survival, 52 days;
Figure 5B). These results were not markedly different from those of a
previous study on PML-RAR leukemia without the FLT3 mutation, for
which we observed that implantation of 10-mg ATRA pellets on day 12 after injection prolonged survival a median of 41 days.47
SU11657, a tyrosine kinase inhibitor, blocks autophosphorylation of FLT3 with ITD mutation MV 4;11 cells express FLT3 with an ITD mutation and were treated with SU11657, a novel receptor tyrosine kinase inhibitor. Immunoprecipitation with an FLT3-specific antibody, followed by Western blot analysis with an antiphosphotyrosine antibody, demonstrated dose-dependent inhibition of autophosphorylation of FLT3 with an IC50 of approximately 50 nM (Figure 4).
SU11657 alters the response of leukemias to ATRA and rapidly restores normal hematopoiesis To move toward the development of a therapeutic regimen in which combination therapy with ATRA and a tyrosine kinase inhibitor, SU11657, would be more efficacious than with ATRA alone, we investigated the effects of these agents in vivo after a short course of therapy. Matched recipients of FLT3/PML-RAR leukemic cells were
treated with ATRA, SU11657, or both beginning on day 12 after
injection. After 4 days of therapy, animals were killed and analyzed.
Combination therapy with ATRA and SU11657 had a dramatic ability to
cause regression of the leukemia and to restore normal hematopoiesis.
Spleen and liver weights, correlated with the degree of involvement by
leukemia, were reduced through treatment (Table 4). Reduction in splenic involvement by
leukemia was also apparent by decreased GFP expression (Figure
6). ATRA alone decreased splenic involvement from 71% to 30%, and combined therapy nearly eliminated leukemic cells from the spleen (only 5% of splenic WBCs expressed GFP). Considering the reduction in splenic weight and the reduction of
GFP-positive leukocytes, SU11657 + ATRA eliminated, on average, 99% of leukemic cells from the spleen. Combination therapy decreased GFP-positive cells in the blood and bone marrow as well, but, interestingly, ATRA therapy did not. In fact, ATRA therapy was associated with increasing numbers of blood cells that expressed GFP,
reflecting the differentiation of leukemic cells to mature circulating
neutrophils. Cytology and histopathology examination of bone
marrow, spleen, and liver revealed the cellular effects of the
treatments (Figure 7). SU11657 had
minimal morphologic effects (compare Figure 7 column 1 with Figure 7
column 2). Although in vitro the inhibition of FLT3 has been associated
with apoptosis, by histopathology we did not observe any substantive
increase in the number of apoptotic cells compared with the moderate
numbers seen in placebo-treated mice. We cannot, however, exclude the possibility of a small increase in apoptosis in vivo. ATRA alone caused
differentiation of the leukemic cells. The bone marrow was filled with
maturing granulocytes (Figure 7C,K). Differentiation was also observed
in the spleen, accompanied by a relative increase in erythroid and
lymphoid cells (Figure 7G,O). There was regression of disease from the
liver (Figure 7S). The addition of SU11657 to ATRA dramatically
accelerated disease regression and restoration of normal bone marrow
hematopoiesis. Marrows of mice treated with ATRA and SU11657 were
filled with a heterogeneous mixture of myeloid, megakaryocytic, and
lymphoid elements, along with prominent erythropoiesis (Figure 7D,L).
Leukemia was markedly reduced in the spleen (Figure 7H,P; H shows
normal splenic lymphocytes) and was eliminated from the liver (Figure
7T).
Flow cytometric immunophenotyping of GFP-positive cells remaining in
the spleen (Figure 8) demonstrated that
untreated cells of this leukemia expressed CD117 (c-kit), low levels of
CD11b (Mac-1), and heterogeneous Ly-6G (Gr-1). SU11657 had a modest effect on surface immunophenotype (small increase in CD11b, loss of
Ly-6G-positive leukemic population), and both ATRA and SU11657 + ATRA caused immunophenotypic differentiation (loss of CD117, increased
CD11b and Ly-6G). These results, showing that a short course of SU11657
synergized with ATRA to restore normal hematopoiesis, will serve as the
basis for additional preclinical studies in which ATRA will be combined
with multiple short courses of SU11657.
Although the creation of the PML-RARA fusion gene is
the central genetic event in APL, this fusion is insufficient for
leukemogenesis. Expression of PML-RARA in myeloid cells of
transgenic mice initially causes only a modest increase in immature
neutrophilic cells, and leukemias appear only after a long latency
(median, 8.5 months in MRP8 PML-RARA transgenic mice).
Activating mutations of FLT3 are a common additional genetic
change in APL. Our results demonstrate that the coexpression of an
activated FLT3 allele accelerates the appearance of leukemia
in mice that express a PML-RARA transgene and that
inhibiting FLT3 in combination with using ATRA to neutralize the effect
of PML-RAR PML-RAR It is notable that the leukemias that arose in activated
FLT3/PML-RAR Although we have previously identified several recurring cytogenetic abnormalities in the leukemias of MRP8 PML-RARA mice, the particular genetic changes responsible for completing transformation are unknown. That these leukemias do not usually exhibit leukocytosis suggests that activation of FLT3 or other cytokine receptors is not a common, spontaneously occurring secondary abnormality in these mice. Further investigations may reveal whether the cooperating spontaneous genetic changes in this mouse model of APL, and in human APL without FLT3 mutations, have molecular effects in common with FLT3 activation or, alternatively, reflect a distinct route to leukemic transformation. We had observed that the combination of BCL-2 and PML-RAR Although activated FLT3 can induce myeloproliferative disease in BALB/c mice, we did not observe this illness in recipients of control FVB/N bone marrow transduced with FLT3W51 retrovirus. Although this could reflect slight technical differences, rates of bone marrow transduction in the current study were comparable to those in the previous study in BALB/c (data not shown). We suspect that the difference in the 2 strains reflects a relative sensitivity in the BALB/c background to MPD. Consistent with this idea, the expression of FLT3W51 in a mixed B6 × C3H strain background also did not result in MPD.43 Elucidating the strain-specific differences that underlie susceptibility to activated FLT3-induced MPD is an area of active investigation (L.M.K., D.G.G., unpublished observations, 2002). Studies of bone marrow transduction with BCR-ABL-expressing retroviruses had previously demonstrated significant strain effects.50 Hence, it is possible that genetic variation in the human population may likewise influence susceptibility to myeloid neoplasms when potentially pathogenic mutations occur. The present study complements a study using cathepsin G
PML-RARA transgenic mice43 to show that the
cooperative effect of activated FLT3 on PML-RAR These 2 mouse models are proving useful for the development of therapeutic strategies that incorporate FLT3 inhibitors. The ability to assess the efficacy of various therapeutic approaches on independent leukemias in different strain backgrounds should enhance the likelihood that such strategies will be successful when applied to the heterogeneous human population. Activating mutations of FLT3 are common in AML, and previous work has shown that FLT3 inhibition represents a promising avenue for novel therapy.42,54-57 We have investigated the possible application of the inhibition of FLT3 to treatment for APL. We found that SU11657, a tyrosine kinase inhibitor able to block FLT3 activity, had only a modest effect when used as a single agent but that it synergized with ATRA to rapidly cause disease regression. SU11657 altered the effect of ATRA. ATRA alone caused the differentiation of leukemic blasts into numerous mature neutrophils. In the presence of SU11657, ATRA caused some differentiation, but its major effect was to rapidly reduce leukemic cell mass. Our observations suggest that combination therapy could potentially reduce the incidence of side effects of ATRA therapy, including retinoic acid syndrome, and might decrease early fatalities by hastening remission. There could also be a positive impact on long-term survival. Additional studies of FLT3 inhibition in our mouse models of APL should lead to treatment strategies that enhance cure. Use of these models will allow the development of novel treatment regimens that will facilitate subsequent human clinical trials.
We thank J. Michael Bishop, Daphne A. Haas-Kogan, H. Jeffrey Lawrence, Frank McCormick, and Kevin M. Shannon for their continuing support.
Submitted July 18, 2002; accepted December 3, 2002.
Prepublished online as Blood First Edition Paper, December 19, 2002; DOI 10.1182/blood-2002-06-1800.
Supported by National Institutes of Health grants K08-CA75986 (S.C.K), U01-CA84221 (S.C.K., M.M.L.), CA66996, (D.G.G.), and DK50654 (D.G.G.) and by a Leukemia and Lymphoma Society SCOR grant (D.G.G.). S.C.K. is the 32nd Edward Mallinckrodt Junior Scholar and is a recipient of a Burroughs Wellcome Fund Career Award. L.M.K. is a Fellow of the Leukemia and Lymphoma Society. D.G.G. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
R and TEL-JAK2.
indicates placebo; 
, SU11657; 
, ATRA; and 
,
ATRA + SU11657. Error bars represent standard deviation.
