RADIATION RESEARCH
182, 310–315 (2014)
0033-7587/14 $15.00 Ó2014 by Radiation Research Society. All rights of reproduction in any form reserved. DOI: 10.1667/RR13738.1
Leukemogenesis in Heterozygous PU.1 Knockout Mice Paula C. Genik,a,1 Irina Vyazunova,b,1 Leta S. Steffen,b Jeffery W. Bacher,b,c Helle Bielefeldt-Ohmann,d Scott McKercher,e Robert L. Ullrich,f Christina M. Fallgren,a Michael M. Weila,2 and F. Andrew Raya a Department of Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, Colorado; b Promega Corporation, Madison, Wisconsin; c University of Wisconsin, Madison, Wisconsin; d School of Veterinary Science, University of Queensland, St. Lucia, Qld 4072, Australia; e Del E. Webb Center for Neuroscience, Aging, and Stem Cell Research, Sanford-Burnham Medical Research Institute, La Jolla, California; and f University of Texas Medical Branch Cancer Center, Galveston, Texas
strains such as CBA, the incidence of rAML increases to 15–25% with increasing radiation dose up to about 3 Gy and then decreases (1–3). Sfpi1, the mouse homologue of human PU.1, is a lineage-specific transcription factor involved in myeloid differentiation that behaves as a tumor suppressor gene in murine rAML. (Following common practice in the literature, we will use the human nomenclature for the mouse gene in this report.) Both copies of murine PU.1 are usually lost in radiation leukemogenesis, one through a large-scale chromosome 2 deletion and the other by point mutation, usually at a CpG dinucleotide in codon 235 (4, 5). The result is a block in myeloid differentiation. Chromosome 2 deletions encompassing PU.1 are evident in about 5% of bone marrow cells 24 h after 3 Gy gamma irradiation and are almost certainly the consequence of radiation exposure (6). The point mutation in PU.1 codon 235 is generally a C to T transition at a CpG dinucleotide, a mutation associated with spontaneous deamination of methylcytosine rather than with radiation exposure (7). rAML data from radiation leukemogenesis experiments fit well with a two-mutation carcinogenesis model, which is based on the assumption that there are two rate-limiting events in rAML. The sequential radiation-induced deletion of one PU.1 allele in an appropriate target cell followed by a spontaneous point mutation in the remaining allele are candidates for those events (4, 5, 8). Germline inactivation of both copies of the PU.1 gene in the mouse is embryonic or neonatal lethal (9, 10). However, conditional knockout of both alleles in adult mice leads to AML at high incidence (95%) with a short latency of 13–34 weeks (11), as does biallelic promoter knockdown that decreases expression levels to 20% (12). This suggests that once both copies of PU.1 are lost, additional events required for leukemogenesis occur frequently and spontaneously. It is curious then that spontaneous AML in PU.1 heterozygous knockout mice (PU.1þ/–) have not been reported in the literature, since one allele of PU.1 is already constitutionally inactive in all target cells in these animals as compared to 5% of cells carrying PU.1 deletions in 3 Gy irradiated mice. There are several possible explanations for this observation.
Genik, P. C., Vyazunova, I., Steffen, L. S., Bacher, J. W., Bielefeldt-Ohmann, H., McKirchner, S., Ullrich, R. L., Fallgren, C. M., Weil, M. M. and Ray, F. A. Leukemogenesis in Heterozygous PU.1 Knockout Mice. Radiat. Res. 182, 310– 315 (2014).
Most murine radiation-induced acute myeloid leukemias involve biallelic inactivation of the PU.1 gene, with one allele being lost through a radiation-induced chromosomal deletion and the other allele affected by a recurrent point mutation in codon 235 that is likely to be spontaneous. The short latencies of acute myeloid leukemias occurring in nonirradiated mice engineered with PU.1 conditional knockout or knockdown alleles suggest that once both copies of PU.1 have been lost any other steps involved in leukemogenesis occur rapidly. Yet, spontaneous acute myeloid leukemias have not been reported in mice heterozygous for a PU.1 knockout allele, an observation that conflicts with the understanding that the PU.1 codon 235 mutation is spontaneous. Here we describe experiments that show that the lack of spontaneous leukemia in PU.1 heterozygous knockout mice is not due to insufficient monitoring times or mouse numbers or the genetic background of the knockout mice. The results reveal that spontaneous leukemias that develop in mice of the mixed 129S2/SvPas and C57BL/6 background of knockout mice arise by a pathway that does not involve biallelic PU.1 mutation. In addition, the latency of radiation-induced leukemia in PU.1 heterozygous mice on a genetic background susceptible to radiation-induced leukemia indicates that the codon 235 mutation is not a rate-limiting step in radiation leukemogenesis driven by PU.1 loss. Ó 2014 by Radiation Research Society
INTRODUCTION
Radiation-induced acute myeloid leukemia (rAML) has been extensively studied in the mouse. In susceptible inbred 1
These two authors contributed equally. Address for correspondence: Department of Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, CO 80525; e-mail: [email protected]. 2
310
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
The latency for radiation-induced AML is long, requiring about 200 days for the first cases to appear and extending past two years (1), so one possibility is that the PU.1þ/– mice may not have been monitored for a sufficient length of time or insufficient numbers to detect AML. Another possibility is that the strain backgrounds on which the PU.1 knockout alleles were made are not rAML sensitive. While it is possible that the codon 235 mutation is, in fact, directly or indirectly radiation induced and its spontaneous mutation rate is too low for leukemogenesis to be observed in PU.1þ/– mice (13). To test these various possibilities, we irradiated or shamirradiated PU.1þ/– mice on their original B6;129S genetic background and on an rAML-susceptible (CBA3B6;129S)F1 background and monitored them to 800 days of age for AML development. The AMLs that arose were then characterized for chromosome 2 deletions and PU.1 codon 235 mutations, as well as for Flt3-ITD mutations and microsatellite instability, which have been reported in some rAMLs (14– 17). The results provide evidence for two distinct, strain dependent leukemogenic pathways, one spontaneous and the other radiation induced and, unexpectedly, indicate that the PU.1 codon 235 point mutation is not a major rate-limiting step in rAML. METHODS Mice and Irradiation Mice heterozygous for the PU.1 knockout allele, Sfpi1tm1Ram (9), were maintained on a mixed 129S2/SvPas and C57BL/6 (B6;129S) background. These mice were mated to CBA/CaJ mice obtained from the Jackson Laboratory to generate progeny that we refer to as (CBA3B6;129S)F1 in this report, regardless of the maternal strain. Unanesthetized male mice were exposed to 3 Gy whole-body 137Cs gamma radiation at 8 weeks of age. The mice were monitored twice daily and beginning at about 8 months postirradiation were palpated weekly for enlarged spleens. Mice that became moribund, symptomatic for AML or reached 800 days of age were euthanized and their spleens collected for histopathology. AML is rapidly lethal in mice and despite daily monitoring for morbidity some mice died between observations. Tissues from these mice were recovered promptly enough for histopathologic determination of AML and they are included in the incidence determinations, but they were not included in the DNA based assays. The spleens from mice with AML were comprised almost entirely of leukemic cells. All animal work was approved by the Institutional Animal Care and Use Committee at Colorado State University under protocols 12-3477A. The animal care facilities at Colorado State University are AAALAC accredited. PU.1 Genotyping PU.1 genotype was determined by PCR using the following primers: (A) 5 0 -GCCCCGGATGTGCTTCCCTTATCAAACC-3 0 and (B) 5 0 TGCCT CGGCCCTGGGAATGTC-3 0 and ( C) 5 0 -CGCACGGGTGTTGGGTCGTTTGTTGG-3 0 (10). Tail DNA was amplified using the primer mix of A, B and C at 1 lM each, 13 GoTaqt (Promega, Madison, WI), and 0.2 units of GoTaq Hot Start Polymerase (Promega) in a Veritit 96-well fast thermal cycler (Applied Biosystemst). The PCR cycling protocol was: 1 cycle of 958C for 2 min, 35 cycles of 928C for 10 s, 608C for 30 s and 728C for 90 s; final
311
extension of 728C for 7 min. The PCR products were resolved by 1.0% agarose/ethidium bromide gel electrophoresis and photographed under UV illumination. The 1,170 bp PCR amplification product from primers A and B was used to identify mice with the wild-type PU.1 gene, and the 970 bp product of primers A and C was used to identify mice with the PU.1 deletion. Loss of Heterozygosity (LOH) Analysis for Chromosome 2 Deletion Microsatellite markers located on chromosome 2 that are polymorphic between CBA/CaJ, 129S2/SvPas and/or C57BL/6J mouse strains, were used to determine genomic sequence loss in AML samples by LOH analysis. The CA-22 locus (Chr2: 92,491,636 to 92,491,772), containing a (CA)22 dinucleotide repeat located within a PU.1 intron, was amplified using primers: 5 0 -GCTGGGTCCTGTGGTACTTCTTTCT-3 0 and 5 0 -CACCCACATTAGGCAGCTCATAACTTC-3 0 . The mBAT-44 locus (Chr2: 92,718,229 to 92,718,567) containing an (A)44 mononucleotide repeat was amplified using primers: 5 0 GCGCTAAGTGGGCCTCTAAGAACA-3 0 and 5 0 -AGTCGTGGAAAGGTGGTAGTGGTTG-3 0 . The D2Mit44 locus (Chr2: 105,313,078 to 105,313,203) containing a (CA)23 dinucleotide repeat was amplified using primers: 5 0 -AAAGCCCGTGTGTCAAAAAC-3 0 and 5 0 -AAATTACCTCCAGTAGATTGGCA-3 0 [Mus musculus for all assemblies (GCF_000001635.21 GCF_000002165.2)]. For detection of LOH in the AML samples, 1 ng of splenic DNA was amplified with 13 primer pair mix (1 lM), 13 Gold ST*R Buffer (Promega), and 0.2 units GoTaq Hot Start Polymerase (Promega) in Veriti 96-well fast thermal cycler. The PCR cycling protocol was: 1 cycle of 958C for 2 min, 35 cycles of 958C for 10 s, 588C for 15 s and 728C for 60 s; final extension of 728C for 7 min; hold at 48C. Products were denatured in deionized formamide with Internal Lane Standard 600 (Promega) for allele sizing and analyzed on a 3130xl Genetic Analyzer using GeneMappert 4.0 software (Applied Biosystems). High-Resolution Melt (HRM) Analysis for the Codon 235 Mutation Genomic DNA was extracted from spleens of AML affected mice using the Wizardt Genomic DNA Purification Kit (Promega) as per manufacturer’s instructions. For HRM analysis, a 73 bp fragment including PU.1 codon 235 was amplified from 2.5 ng DNA with GoTaq qPCR MasterMix (Promega) and 250 nM primers (5 0 GCAACCGCAAGAAGATGACCT-3 0 , 5 0 -TTCACCTCGCCTGTCTTGCC-3 0 ) (4). Amplification and melt were performed in duplicate on a CFX96e Real-Time Detection System (Bio-Rad, Hercules, CA) with the following conditions: 1 cycle of 958C for 2 min; 40 cycles of 958C for 10 s, 608C for 1 min; melt at 0.18C intervals. Precision Melt Analysise Software (Bio-Rad) was used to cluster melt profiles with default settings. Mouse genomic DNA (Promega) and rAML cell line 8016 carrying a PU.1 R235C mutation were used for wild-type and mutant controls, respectively (18). Sequencing PU.1 Gene Exon 5 Genomic DNA was amplified from 4 ng DNA with GoTaq Hot Start Colorless Master Mix (Promega) and 250 nM PCR primers (Exon 5a: 5 0 -CTACCATGGCTGGGAGGAG-3 0 and 5 0 -GACTCTAGCCCGCTCCAA-3 0 ; Exon 5b: 5 0 -GACAGGCGAGGTGAAGAAAG-3 and 5 0 -CCAGCTTTGGAAACATGCTC-3 0 ) on a Veriti 96-well fast thermal cycler (Applied Biosystems) with the following conditions: 1 cycle of 958C for 2 min; 40 cycles of 958C for 10 s, 608C for 15 s; 728C for 75 s and 728C for 7 min. Products were isolated using the Promega Wizardt SV Gel and PCR Clean Up System and sequenced by the University of Wisconsin Biotechnology Center (Madison, WI) Sequencing Facility using nested sequencing primers (Exon 5a: 5 0 -GGGTCCCCATAAAATCCTTG-3 0 and 5 0 -GGACACAGGGAGGACATTCC-3 0 ; Exon 5b: 5 0 -CTCACCTACCAGTTCAGCGG-3 0 and 5 0 -TAAGAGATCGTGCACACGGG-3 0 ).
312
GENIK ET AL.
Microsatellite Instability and Flt3-ITD Analysis For detection of microsatellite instability (MSI), a multiplexed panel of markers designed specifically for CBA/Ca and C57BL/6 mouse strains was used (17). The CBA/Ca multiplex contains mononucleotide repeat loci mBAT-56, mBAT-57, mBAT-58, mBAT-59, mBAT64 and mBAT-66; and the C57BL/6 multiplex contains mBAT-56, mBAT-56h, mBAT-57, mBAT-58, mBAT-59 and mBAT-66. For MSI analysis 1 ng of spleen DNA was amplified with 13 CBA/Ca multiplex or 13 C57BL/6 multiplex, 13 Gold ST*R Buffer (Promega), and 0.2 units GoTaq Hot Start Polymerase (Promega) in a GeneAmp PCR System 9700 Thermocycler (Applied Biosystems) with the following amplification conditions: 958C for 2 min; 10 cycles of 948C for 30 s, 588C for 30 s (29% ramp), 658C for 1 min (28% ramp); 20 cycles of 908C for 30 s, 588C for 30 s (29% ramp), 658C for 1 min (28% ramp), 608C for 30 min. Products were denatured in deionized formamide with Internal Lane Standard 600 (Promega) for allele sizing and analyzed on a 3130xl Genetic Analyzer using GeneMapper 4.0 Software (Applied Biosystems). Size shifts in microsatellite marker of 3 bp from the wild-type allele was considered MSI positive. The presence or absence of internal tandem duplications in the Flt3 receptor tyrosine kinase gene was assayed by PCR (12). Flt3-ITD mutations were detected using primers amplifying the region covering exons 14 and 15 (5 0 -Fluorescein-GCAATTTAGGTACGAGAGTCAGC-3 0 and 5 0 -CTTTTAGCATCTTCACCGCCACC-3 0 ) using the same PCR and analysis conditions described for MSI analysis. The presence of Flt3-ITD is detectible by an increase in amplicon size of 3–400 bp (16). Statistical Analyses Differences in AML incidences between groups were evaluated using Fisher’s exact test (two-sided) and differences in latency were examined using unpaired t tests. Analyses were run using GraphPad Prism software (GraphPad Software Inc., La Jolla, CA).
RESULTS
AML Incidence
Most engineered laboratory mouse strains are maintained in small numbers, and only for their useful breeding life or for use in short-term experiments, so the lack of reports of spontaneous AMLs in PU.1þ/– mice might simply be the consequence of too few mice monitored for too short a time. Therefore, we monitored PU.1þ/þ and PU.1þ/– mice on the original B6;129 mixed background described by McKercher et al. (9) to 800 days of age (Table 1). Unirradiated PU.1þ/þ mice had a 6% incidence of AML that was not significantly increased by irradiation (P ¼ 0.3199). PU.1þ/– mice had a higher incidence of spontaneous AML than PU.1þ/þ mice (15% and 6%, respectively, P ¼ 0.0446). However, we characterized 16 of the 17 leukemias that arose in the unirradiated PU.1 þ/– mice by LOH assays and HRM analysis (Table 1) and found that these leukemias did not contain PU.1 codon 235 mutations. No evidence was found for mutation of the wild-type PU.1 allele in 15 of the 16 AMLs analyzed. The remaining AML lost the wild-type allele (reported as KO/Del2 in Table 1, though the mechanism of loss has not been determined). Thus, spontaneous leukemias in the B6;129S PU.1þ/– mice, with
one exception, did not arise through the classical PU.1 biallelic inactivation pathway. While irradiation did not increase the AML incidence in PU.1þ/þ mice, it more than doubled the incidence in PU.1þ/– mice (P ¼ 0.0002). To determine if radiation was responsible for the PU.1 codon 235 mutations typically observed in rAML, we assayed 36 leukemias from irradiated PU.1þ/– mice and found that 30 of them could be explained by biallelic PU.1 loss: 29 by LOH for the wildtype allele and one by deletion of the knockout allele and a codon 235 point mutation in the wild-type allele. These results indicate that radiation does not cause the PU.1 codon 235 mutation in rAML or, alternatively, that the codon 235 mutation cannot cooperate with the knockout allele to cause AML. Since only one of the 52 characterized AMLs arising in the B6;129S PU.1þ/– mice had a PU.1 codon 235 mutation, we next considered if AMLs with that mutation develop only in strains that are susceptible to rAML, such as CBA. The Sfpi1tm1Ram knockout allele used in this study is maintained on a mixed 129S2/SvPas and C57BL/6 background. C57BL/6 mice are rAML resistant and the 129S2/SvPas substrain has not been phenotyped for rAML susceptibility. Since F1 hybrid mice of a rAML-resistant and sensitive strain (i.e., CBA and C57BL/6) have intermediate rAML susceptibility (4.5–9%) (19, 20), we determined the incidence of AML and the frequency of AMLs bearing PU.1 codon 235 mutations in (CBA3B6;129S)F1 PU.1þ/þ and PU.1þ/– mice on the assumption that this background would be moderately rAML sensitive. None of the unirradiated (CBA3B6;129S)F1 PU.1þ/þ mice developed AML, an expected finding in light of the very low incidence of spontaneous AML in the CBA strain (,1 in 1,000) (20). Only 2 of 100 unirradiated (CBA3B6;129S)F1 PU.1þ/– mice developed AML which was not significantly different than the incidence in (CBA3B6;129S)F1 PU1þ/þ mice (P ¼ 0.2093) and demonstrates that strain background cannot fully account for the lack of reports of AMLs in PU.1 þ/– mice, since (CBA3B6;129S)F1 are sensitive to radiation-induced AML. After irradiation, 8% of the (CBA3B6;129S)F1 PU.1þ/þ mice developed rAML, indicating that this genetic background has intermediate susceptibility to radiation-induced AML, as expected. LOH and HRM analyses of eight of these leukemias revealed that they arose by chromosome 2 deletions and that six of them also had codon 235 mutations in the remaining PU.1 allele, consistent with previous reports of biallelic PU.1 inactivation by deletion and point mutation in rAML. In agreement with a previous report by Clark et al. (14), no bias was observed in loss of CBA/CaJ or C57BL/6J alleles in these leukemias, suggesting that strain-specific chromosome 2 fragile sites or cis-acting polymorphisms are not involved in strain differences in rAML susceptibility. The codon 235 point mutations detected by HRM analysis were verified by sequencing. Irradiated (CBA3B6;129S)F1 PU.1þ/– mice had a 28% incidence of rAML, significantly higher than that of
313
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
TABLE 1
Sample
PU.1
Dose (Gy)
Mice
B6;129S B6;129S B6;129S B6;129S (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1
þ/þ þ/– þ/þ þ/– þ/þ þ/– þ/þ þ/–
0 0 3 3 0 0 3 3
100 110 109 129 118 100 106 98
AML (%) 6 17 11 48 0 2 9 27
(6) (15) (10) (37) (0) (2) (8) (28)
Latency in days (range) 581 579 570 452
(478–707) (373–813) (324–693) (160–741) nd 706 (700–712) 548 (450–620) 516 (224–729)
PU.1 genotype/number of leukemias tested KO þ Del2
KO þ wt
Del2 þ R235
Del2 þ wt
na 1/16 na 29/36 nd 0/1 0/8 20/26
na 15/16 na 6/36 nd 0/1 0/8 2/26
0/6 0/16 0/9 1/36 nd 0/1 6/8 3/26
na 0/16 na 0/36 nd 1/1 2/8 1/26
Notes. Chromosome 2 deletions (Del2) were determined by LOH analysis, presence of the PU.1 knockout (KO) allele by PCR genotyping and presence of the codon 235 mutation by HRM analysis and sequencing. Del2 status could not be determined for B6;129 þ/ þ AML samples because these tumors were homozygous for the B6 allele of the LOH markers. na ¼ not applicable; nd ¼ no data.
irradiated (CBA3B6;129S)F1 PU.1þ/þ (P ¼ 0.0004) mice. Twenty-three of the 26 leukemias analyzed arose through biallelic inactivation of PU.1, most commonly through a chromosome 2 deletion that resulted in loss of the wild-type PU.1 allele, but in three cases through deletion of the knockout allele and codon 235 mutation on the wild-type allele. The remaining three leukemias had inactivation of one PU.1 allele either through deletion or retention of the knockout allele, but no codon 235 or other PU.1 mutations were detected on the wild-type alleles. However, there is of course the possibility that the wild-type alleles were inactivated by a mutation in a regulatory region or epigenetic gene silencing. AML Latency
The latencies for leukemias resulting from chromosome 2 deletions and PU.1 codon 235 mutations in irradiated (CBA3B6;129S)F1 PU.1þ/þ and PU.1þ/– mice (N ¼ 9, mean 6 SEM, 584.9 6 25.28) were not significantly different (P ¼ 0.0646) than those resulting from a chromosome 2 deletion and a PU.1 knockout allele (N ¼ 20, mean 6 SEM, 484.2 6 32.93). Microsatellite Instability
Microsatellite instability has been implicated in human leukemia and lymphoma (21, 22), as well as in rAML in CBA mice (14, 15). Therefore, the MSI status of AML tumors was investigated as a possible alternate mutator mechanism to help explain leukemogenesis in B6;129S mice that did not exhibit the expected biallelic PU.1 mutations. MSI in one or more of the microsatellite markers was found in 11 of 35 (31%) of AMLs from (CBA3B6;129S)F1 mice (Table 2), in close agreement with the 33% reported in AMLs from CBA/Ca mice (17). In contrast, MSI in spontaneous and radiation-induced AML from B6;129S mice occurred in only 2 of 66 cases (3%) (P ¼ 0.0001). Lymphomas from the same cohort of B6;129S mice did not exhibit any evidence of MSI, indicating that MSI was not
simply a result of clonal expansion of radiation-induced mutations in the microsatellite loci tested. The low level of MSI observed in tumors from B6;129S mice suggest that MSI is not a major contributor to leukemogenesis in this genetic background. Flt3 Mutations
PU.1 expression is decreased in about 20–25% of human AML cases due to internal tandem duplications in the FLT3 gene (23) and decreased expression has also been found in some murine rAML lacking PU.1 mutations (16). Therefore, AML samples from both the (CBA3B6;129S)F1 and B6;129S mice were screened for presence of Flt3-ITD by PCR. No evidence of Flt3 mutations was observed in any of the tumors. DISCUSSION
The lack of reports of PU.1þ/– mice developing AML challenges a model of radiation leukemogenesis in which biallelic inactivation of PU.1 results from radiation-induced deletion of one allele and spontaneous mutation at codon 235 on the other. We tested whether insufficient monitoring times or a rAML-resistant strain background could account for the lack of reports, or if radiation exposure was required for the codon 235 mutation. PU.1þ/– mice on their original B6;129S background developed AML if they were monitored long enough, however, codon 235 mutations were not involved. Irradiating B6;129S PU.1þ/– mice increased AML incidence by 2.5-fold, but rather than radiation exposure causing leukemias with codon 235 mutations, most of the resulting AMLs lost the wild-type allele of PU.1 by deletion. Only a single codon 235 mutation was detected in 36 leukemias analyzed. This indicates that radiation does not cause the codon 235 mutation, at least not on this strain background. Alternatively, the codon 235 mutation may not be able to complement the PU.1 knockout allele for the induction of leukemia.
314
GENIK ET AL.
TABLE 2 Sample
PU.1
Dose (Gy)
AML tested
B6;129S B6;129S B6;129S B6;129S (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1
þ/þ þ/– þ/þ þ/– þ/þ þ/– þ/þ þ/–
0 0 3 3 0 0 3 3
6 16 10 34 0 1 8 26
To examine the effect of strain background, we generated hybrid mice that were susceptible to rAML and carried one copy of the PU.1 knockout allele. Of the 100 unirradiated mice of this genotype that were monitored, only two developed AML. One AML was recovered and assayed and was found to lack the codon 235 mutation. Once again, the conclusion would be that the codon 235 mutation does not arise spontaneously or that it does not complement the PU.1 knockout allele for the induction of leukemia. Irradiating mice of this genotype increased the AML incidence by 14fold. Most of the leukemias harbored the knockout allele and a chromosome 2 deletion, although three of the 26 leukemias assayed resulted from deletion of one copy of PU.1 coupled with codon 235 mutation on the remaining allele. Taken together these results indicate that leukemogenesis involving codon 235 is strain background-dependent (1 of 67 AMLs in B6;129S mice compared to 9 of 35 AMLs in the F1 hybrids, P ¼ 0.0002). Because no examples of leukemia resulting from the R235 mutation in conjunction with the knockout allele were observed, the results do not resolve whether the codon 235 mutation is spontaneous or radiation induced. It is possible that radiation may have caused codon 235 mutations, but they were not detected because they could not complement the knockout allele in leukemogenesis. Two additional findings emerged from the results of these experiments. First, AML arises by different pathways in (CBA3B6;129S)F1 and B6;129S mice and second, the PU.1 codon 235 mutation commonly found in rAML does not appear to alter latency and therefore is not a major ratelimiting step in radiation leukemogenesis. While we showed that B6;129S PU.1þ/– mice have a higher spontaneous incidence of AML than PU.1þ/þ mice on the same background, however, with only one exception, we found no evidence for biallelic PU.1 loss. AMLs in B6;129S mice are characterized by microsatellite stability, an absence of Flt3 internal tandem duplications and long latencies (160–813 days, mean of 504 days). The failure to identify PU.1 mutations does not rule out the possibility that other mechanisms decrease or alter the function of PU.1 encoded by the wild-type PU.1 allele in these tumors or that haploinsufficiency for PU.1 enables an alternate leukemogenesis pathway. For example, Walter et al. (24) crossed
MSI (%) 0 1 0 1
(0) (6) (0) (3) – 1 (100) 3 (38) 7 (27)
Lymphomas tested 10 9 9 6 1 7 8 14
MSI (%) 0 0 0 0 0 1 2 2
(0) (0) (0) (0) (0) (14) (25) (14)
mice transgenic for human RARa-PML that occasionally develop acute promyelocytic leukemia (APL) with PU.1þ/– mice and found that progeny with the knockout allele had a greatly increased APL incidence. Additional experiments provided evidence that RARa-PML also reduced PU.1 mRNA levels from the wild-type allele. B6;129S mice can be manipulated to develop AML through the biallelic PU.1 inactivation pathway. We found that irradiated B6;129S PU.1þ/– mice had a high incidence of AML resulting from biallelic PU.1 inactivation which reinforces previous findings that leukemogenesis can be forced down this pathway even in a strain in which it does not normally occur (11, 12). Conversely, (CBA3B6;129S)F1 mice do not develop spontaneous AML, even if one allele of PU.1 is inactivated, the AML incidence is very low (2 of 100 PU.1þ/– mice followed on this background). However, AMLs can be induced in these mice by radiation exposure and the majority of these tumors have biallelic PU.1 inactivation. Interestingly, these leukemias also lack Flt3-ITD mutations, have long latencies, and they are frequently MSI positive. Interestingly, we found no leukemias that had biallelic PU.1 codon 235 mutations, perhaps because the encoded protein retains partial function (5). If so, levels of the partially functional protein could be high enough to prevent leukemogenesis if two copies of the mutant gene are present, but not if only a single copy is present, similar to class 2 low-penetrance RB1 mutations found in some retinoblastoma families (25). Also interesting is that none of the leukemias in PU.1þ/– mice that retained the knockout allele lost the wild-type PU.1 allele through codon 235 mutation. This suggests that hemizygosity for another gene (or other genes) in the minimally deleted region of chromosome 2 is required for rAML development involving the codon 235 mutation. Irradiated (CBA3B6;129S)F1 PU.1þ/– mice have a higher AML incidence than irradiated (CBA3B6;129S)F1 PU.1þ/þ mice. This is not surprising since most of these AMLs arise by biallelic PU.1 inactivation, with either a chromosome 2 deletion or a codon 235 mutation complementing the knockout allele. More cells are at risk for leukemic transformation from a radiation-induced chromosome 2 deletion in PU.1þ/– mice in which all of the target cells have
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
an inactivated PU.1 allele than in PU.1þ/þ mice in which only those cells that have a codon 235 mutation concurrently with the deletion are at risk. Finally, the knockout allele increased rAML incidence it does not significantly decrease its latency. Since germline inactivation of one copy of a rate limiting tumor suppressor gene (in this case the knockout allele) should decrease tumor latency (26), one conclusion is that the PU.1 codon 235 mutation is not a major rate limiting step in rAML. This observation raises the possibility that codon 235 has a high mutation rate and the mutation is present around the time of irradiation. ACKNOWLEDGMENTS This work was supported by grant NNX09AM08G from the National Aeronautics and Space Administration entitled, ‘‘NASA Specialized Center of Research on Radiation Carcinogenesis.’’ We wish to thank the staff at Laboratory Animal Resources (Colorado State University) for excellent animal care. Received: March 17, 2014; accepted: May 30, 2014; published online: July 30, 2014
REFERENCES 1. Major IR, Mole RH. Myeloid leukaemia in x-ray irradiated CBA mice. Nature 1978; 272:455–6. 2. Mole RH, Papworth DG, Corp MJ. The dose-response for x-ray induction of myeloid-leukemia in male CBA/H mice. Br J Cancer 1983; 47:285–91. 3. Weil MM, Bedford JS, Bielefeldt-Ohmann H, Ray FA, Genik PC, Ehrhart EJ, et al. Incidence of acute myeloid leukemia and hepatocellular carcinoma in mice irradiated with 1 GeV/nucleon (56)Fe ions. Radiat Res 2009; 172:213–9. 4. Cook WD, McCaw BJ, Herring C, John DL, Foote SJ, Nutt SL, et al. PU.1 is a suppressor of myeloid leukemia, inactivated in mice by gene deletion and mutation of its DNA binding domain. Blood 2004; 104:3437–44. 5. Suraweera N, Meijne E, Moody J, Carvajal-Carmona LG, Yoshida K, Pollard P, et al. Mutations of the PU.1 Ets domain are specifically associated with murine radiation-induced, but not human therapy-related, acute myeloid leukaemia. Oncogene 2005; 24:3678–83. 6. Peng Y, Brown N, Finnon R, Warner CL, Liu X, Genik PC, et al. Radiation leukemogenesis in mice: loss of PU.1 on chromosome 2 in CBA and C57BL/6 mice after irradiation with 1 GeV/nucleon 56Fe ions, X rays or gamma rays. Part I. Experimental observations. Radiat Res 2009; 171:474–83. 7. Duncan BK, Miller JH. Mutagenic deamination of cytosine residues in DNA. Nature 1980; 287:560–1. 8. Dekkers F, Bijwaard H, Bouffler S, Ellender M, Huiskamp R, Kowalczuk C, et al. A two-mutation model of radiation-induced acute myeloid leukemia using historical mouse data. Radiat Environ Biophys 2011; 50:37–45. 9. McKercher SR, Torbett BE, Anderson KL, Henkel GW, Vestal DJ, Baribault H, et al. Targeted disruption of the PU.1 gene results in multiple hematopoietic abnormalities. Embo Journal 1996; 15:5647–58.
315
10. Scott EW, Simon MC, Anastasi J, Singh H. Requirement of transcription factor Pu.1 in the development of multiple hematopoietic lineages. Science 1994; 265:1573–7. 11. Metcalf D, Dakic A, Mifsud S, Di Rago L, Wu L, Nutt S. Inactivation of PU.1 in adult mice leads to the development of myeloid leukemia. Proc Natl Acad Sci U S A 2006; 103:1486–91. 12. Rosenbauer F, Wagner K, Kutok JL, Iwasaki H, Le Beau MM, Okuno Y, et al. Acute myeloid leukemia induced by graded reduction of a lineage-specific transcription factor, PU.1. Nature Genet 2004; 36:624–30. 13. Ban N, Kai M. Implication of replicative stress-related stem cell ageing in radiation-induced murine leukaemia. Br J Cancer 2009; 101:363–71. 14. Clark DJ, Meijne EI, Bouffler SD, Huiskamp R, Skidmore CJ, Cox R, et al. Microsatellite analysis of recurrent chromosome 2 deletions in acute myeloid leukaemia induced by radiation in F1 hybrid mice. Genes Chromosomes Cancer 1996; 16:238–46. 15. Curik N, Burda P, Vargova K, Pospisil V, Belickova M, Vlckova P, et al. 5-azacitidine in aggressive myelodysplastic syndromes regulates chromatin structure at PU.1 gene and cell differentiation capacity. Leukemia 2012; 26:1804–11. 16. Finnon R, Brown N, Moody J, Badie C, Olme CH, Huiskamp R, et al. Flt3-ITD mutations in a mouse model of radiation-induced acute myeloid leukaemia. Leukemia 2012; 26:1445–6. 17. Steffen LS, Bacher JW, Peng Y, Le PN, Ding LH, Genik PC, et al. Molecular characterisation of murine acute myeloid leukaemia induced by 56Fe ion and 137Cs gamma ray irradiation. Mutagenesis 2013; 28:71–9. 18. Hirouchi T, Takabatake T, Yoshida K, Nitta Y, Nakamura M, Tanaka S, et al. Upregulation of c-myc gene accompanied by PU.1 deficiency in radiation-induced acute myeloid leukemia in mice. Exp Hematol 2008; 36:871–85. 19. Boulton E, Cleary H, Papworth D, Plumb M. Susceptibility to radiation-induced leukaemia/lymphoma is genetically separable from sensitivity to radiation-induced genomic instability. Int J Radiat Biol 2001; 77:21–9. 20. Meijne E, Huiskamp R, Haines J, Moody J, Finnon R, Wilding J, et al. Analysis of loss of heterozygosity in lymphoma and leukaemia arising in F1 hybrid mice locates a common region of chromosome 4 loss. Genes Chromosomes Cancer 2001; 31:373– 81. 21. Wimmer K, Etzler J. Constitutional mismatch repair-deficiency syndrome: have we so far seen only the tip of an iceberg? Hum Genet 2008; 124:105–22. 22. Wimmer K, Kratz CP. Constitutional mismatch repair-deficiency syndrome. Haematologica 2010; 95:699–701. 23. Schnittger S, Schoch C, Dugas M, Kern W, Staib P, Wuchter C, et al. Analysis of FLT3 length mutations in 1003 patients with acute myeloid leukemia: correlation to cytogenetics, FAB subtype, and prognosis in the AMLCG study and usefulness as a marker for the detection of minimal residual disease. Blood 2002; 100:59–66. 24. Walter MJ, Park JS, Ries RE, Lau SK, McLellan M, Jaeger S, et al. Reduced PU.1 expression causes myeloid progenitor expansion and increased leukemia penetrance in mice expressing PMLRARalpha. Proc Natl Acad Sci U S A 2005; 102:12513–8. 25. Harbour JW. Molecular basis of low-penetrance retinoblastoma. Arch Ophthalmol 2001; 119:1699–704. 26. Knudson AG, Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A 1971; 68:820–3.
182, 310–315 (2014)
0033-7587/14 $15.00 Ó2014 by Radiation Research Society. All rights of reproduction in any form reserved. DOI: 10.1667/RR13738.1
Leukemogenesis in Heterozygous PU.1 Knockout Mice Paula C. Genik,a,1 Irina Vyazunova,b,1 Leta S. Steffen,b Jeffery W. Bacher,b,c Helle Bielefeldt-Ohmann,d Scott McKercher,e Robert L. Ullrich,f Christina M. Fallgren,a Michael M. Weila,2 and F. Andrew Raya a Department of Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, Colorado; b Promega Corporation, Madison, Wisconsin; c University of Wisconsin, Madison, Wisconsin; d School of Veterinary Science, University of Queensland, St. Lucia, Qld 4072, Australia; e Del E. Webb Center for Neuroscience, Aging, and Stem Cell Research, Sanford-Burnham Medical Research Institute, La Jolla, California; and f University of Texas Medical Branch Cancer Center, Galveston, Texas
strains such as CBA, the incidence of rAML increases to 15–25% with increasing radiation dose up to about 3 Gy and then decreases (1–3). Sfpi1, the mouse homologue of human PU.1, is a lineage-specific transcription factor involved in myeloid differentiation that behaves as a tumor suppressor gene in murine rAML. (Following common practice in the literature, we will use the human nomenclature for the mouse gene in this report.) Both copies of murine PU.1 are usually lost in radiation leukemogenesis, one through a large-scale chromosome 2 deletion and the other by point mutation, usually at a CpG dinucleotide in codon 235 (4, 5). The result is a block in myeloid differentiation. Chromosome 2 deletions encompassing PU.1 are evident in about 5% of bone marrow cells 24 h after 3 Gy gamma irradiation and are almost certainly the consequence of radiation exposure (6). The point mutation in PU.1 codon 235 is generally a C to T transition at a CpG dinucleotide, a mutation associated with spontaneous deamination of methylcytosine rather than with radiation exposure (7). rAML data from radiation leukemogenesis experiments fit well with a two-mutation carcinogenesis model, which is based on the assumption that there are two rate-limiting events in rAML. The sequential radiation-induced deletion of one PU.1 allele in an appropriate target cell followed by a spontaneous point mutation in the remaining allele are candidates for those events (4, 5, 8). Germline inactivation of both copies of the PU.1 gene in the mouse is embryonic or neonatal lethal (9, 10). However, conditional knockout of both alleles in adult mice leads to AML at high incidence (95%) with a short latency of 13–34 weeks (11), as does biallelic promoter knockdown that decreases expression levels to 20% (12). This suggests that once both copies of PU.1 are lost, additional events required for leukemogenesis occur frequently and spontaneously. It is curious then that spontaneous AML in PU.1 heterozygous knockout mice (PU.1þ/–) have not been reported in the literature, since one allele of PU.1 is already constitutionally inactive in all target cells in these animals as compared to 5% of cells carrying PU.1 deletions in 3 Gy irradiated mice. There are several possible explanations for this observation.
Genik, P. C., Vyazunova, I., Steffen, L. S., Bacher, J. W., Bielefeldt-Ohmann, H., McKirchner, S., Ullrich, R. L., Fallgren, C. M., Weil, M. M. and Ray, F. A. Leukemogenesis in Heterozygous PU.1 Knockout Mice. Radiat. Res. 182, 310– 315 (2014).
Most murine radiation-induced acute myeloid leukemias involve biallelic inactivation of the PU.1 gene, with one allele being lost through a radiation-induced chromosomal deletion and the other allele affected by a recurrent point mutation in codon 235 that is likely to be spontaneous. The short latencies of acute myeloid leukemias occurring in nonirradiated mice engineered with PU.1 conditional knockout or knockdown alleles suggest that once both copies of PU.1 have been lost any other steps involved in leukemogenesis occur rapidly. Yet, spontaneous acute myeloid leukemias have not been reported in mice heterozygous for a PU.1 knockout allele, an observation that conflicts with the understanding that the PU.1 codon 235 mutation is spontaneous. Here we describe experiments that show that the lack of spontaneous leukemia in PU.1 heterozygous knockout mice is not due to insufficient monitoring times or mouse numbers or the genetic background of the knockout mice. The results reveal that spontaneous leukemias that develop in mice of the mixed 129S2/SvPas and C57BL/6 background of knockout mice arise by a pathway that does not involve biallelic PU.1 mutation. In addition, the latency of radiation-induced leukemia in PU.1 heterozygous mice on a genetic background susceptible to radiation-induced leukemia indicates that the codon 235 mutation is not a rate-limiting step in radiation leukemogenesis driven by PU.1 loss. Ó 2014 by Radiation Research Society
INTRODUCTION
Radiation-induced acute myeloid leukemia (rAML) has been extensively studied in the mouse. In susceptible inbred 1
These two authors contributed equally. Address for correspondence: Department of Environmental and Radiological Health Sciences, Colorado State University, Fort Collins, CO 80525; e-mail: [email protected]. 2
310
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
The latency for radiation-induced AML is long, requiring about 200 days for the first cases to appear and extending past two years (1), so one possibility is that the PU.1þ/– mice may not have been monitored for a sufficient length of time or insufficient numbers to detect AML. Another possibility is that the strain backgrounds on which the PU.1 knockout alleles were made are not rAML sensitive. While it is possible that the codon 235 mutation is, in fact, directly or indirectly radiation induced and its spontaneous mutation rate is too low for leukemogenesis to be observed in PU.1þ/– mice (13). To test these various possibilities, we irradiated or shamirradiated PU.1þ/– mice on their original B6;129S genetic background and on an rAML-susceptible (CBA3B6;129S)F1 background and monitored them to 800 days of age for AML development. The AMLs that arose were then characterized for chromosome 2 deletions and PU.1 codon 235 mutations, as well as for Flt3-ITD mutations and microsatellite instability, which have been reported in some rAMLs (14– 17). The results provide evidence for two distinct, strain dependent leukemogenic pathways, one spontaneous and the other radiation induced and, unexpectedly, indicate that the PU.1 codon 235 point mutation is not a major rate-limiting step in rAML. METHODS Mice and Irradiation Mice heterozygous for the PU.1 knockout allele, Sfpi1tm1Ram (9), were maintained on a mixed 129S2/SvPas and C57BL/6 (B6;129S) background. These mice were mated to CBA/CaJ mice obtained from the Jackson Laboratory to generate progeny that we refer to as (CBA3B6;129S)F1 in this report, regardless of the maternal strain. Unanesthetized male mice were exposed to 3 Gy whole-body 137Cs gamma radiation at 8 weeks of age. The mice were monitored twice daily and beginning at about 8 months postirradiation were palpated weekly for enlarged spleens. Mice that became moribund, symptomatic for AML or reached 800 days of age were euthanized and their spleens collected for histopathology. AML is rapidly lethal in mice and despite daily monitoring for morbidity some mice died between observations. Tissues from these mice were recovered promptly enough for histopathologic determination of AML and they are included in the incidence determinations, but they were not included in the DNA based assays. The spleens from mice with AML were comprised almost entirely of leukemic cells. All animal work was approved by the Institutional Animal Care and Use Committee at Colorado State University under protocols 12-3477A. The animal care facilities at Colorado State University are AAALAC accredited. PU.1 Genotyping PU.1 genotype was determined by PCR using the following primers: (A) 5 0 -GCCCCGGATGTGCTTCCCTTATCAAACC-3 0 and (B) 5 0 TGCCT CGGCCCTGGGAATGTC-3 0 and ( C) 5 0 -CGCACGGGTGTTGGGTCGTTTGTTGG-3 0 (10). Tail DNA was amplified using the primer mix of A, B and C at 1 lM each, 13 GoTaqt (Promega, Madison, WI), and 0.2 units of GoTaq Hot Start Polymerase (Promega) in a Veritit 96-well fast thermal cycler (Applied Biosystemst). The PCR cycling protocol was: 1 cycle of 958C for 2 min, 35 cycles of 928C for 10 s, 608C for 30 s and 728C for 90 s; final
311
extension of 728C for 7 min. The PCR products were resolved by 1.0% agarose/ethidium bromide gel electrophoresis and photographed under UV illumination. The 1,170 bp PCR amplification product from primers A and B was used to identify mice with the wild-type PU.1 gene, and the 970 bp product of primers A and C was used to identify mice with the PU.1 deletion. Loss of Heterozygosity (LOH) Analysis for Chromosome 2 Deletion Microsatellite markers located on chromosome 2 that are polymorphic between CBA/CaJ, 129S2/SvPas and/or C57BL/6J mouse strains, were used to determine genomic sequence loss in AML samples by LOH analysis. The CA-22 locus (Chr2: 92,491,636 to 92,491,772), containing a (CA)22 dinucleotide repeat located within a PU.1 intron, was amplified using primers: 5 0 -GCTGGGTCCTGTGGTACTTCTTTCT-3 0 and 5 0 -CACCCACATTAGGCAGCTCATAACTTC-3 0 . The mBAT-44 locus (Chr2: 92,718,229 to 92,718,567) containing an (A)44 mononucleotide repeat was amplified using primers: 5 0 GCGCTAAGTGGGCCTCTAAGAACA-3 0 and 5 0 -AGTCGTGGAAAGGTGGTAGTGGTTG-3 0 . The D2Mit44 locus (Chr2: 105,313,078 to 105,313,203) containing a (CA)23 dinucleotide repeat was amplified using primers: 5 0 -AAAGCCCGTGTGTCAAAAAC-3 0 and 5 0 -AAATTACCTCCAGTAGATTGGCA-3 0 [Mus musculus for all assemblies (GCF_000001635.21 GCF_000002165.2)]. For detection of LOH in the AML samples, 1 ng of splenic DNA was amplified with 13 primer pair mix (1 lM), 13 Gold ST*R Buffer (Promega), and 0.2 units GoTaq Hot Start Polymerase (Promega) in Veriti 96-well fast thermal cycler. The PCR cycling protocol was: 1 cycle of 958C for 2 min, 35 cycles of 958C for 10 s, 588C for 15 s and 728C for 60 s; final extension of 728C for 7 min; hold at 48C. Products were denatured in deionized formamide with Internal Lane Standard 600 (Promega) for allele sizing and analyzed on a 3130xl Genetic Analyzer using GeneMappert 4.0 software (Applied Biosystems). High-Resolution Melt (HRM) Analysis for the Codon 235 Mutation Genomic DNA was extracted from spleens of AML affected mice using the Wizardt Genomic DNA Purification Kit (Promega) as per manufacturer’s instructions. For HRM analysis, a 73 bp fragment including PU.1 codon 235 was amplified from 2.5 ng DNA with GoTaq qPCR MasterMix (Promega) and 250 nM primers (5 0 GCAACCGCAAGAAGATGACCT-3 0 , 5 0 -TTCACCTCGCCTGTCTTGCC-3 0 ) (4). Amplification and melt were performed in duplicate on a CFX96e Real-Time Detection System (Bio-Rad, Hercules, CA) with the following conditions: 1 cycle of 958C for 2 min; 40 cycles of 958C for 10 s, 608C for 1 min; melt at 0.18C intervals. Precision Melt Analysise Software (Bio-Rad) was used to cluster melt profiles with default settings. Mouse genomic DNA (Promega) and rAML cell line 8016 carrying a PU.1 R235C mutation were used for wild-type and mutant controls, respectively (18). Sequencing PU.1 Gene Exon 5 Genomic DNA was amplified from 4 ng DNA with GoTaq Hot Start Colorless Master Mix (Promega) and 250 nM PCR primers (Exon 5a: 5 0 -CTACCATGGCTGGGAGGAG-3 0 and 5 0 -GACTCTAGCCCGCTCCAA-3 0 ; Exon 5b: 5 0 -GACAGGCGAGGTGAAGAAAG-3 and 5 0 -CCAGCTTTGGAAACATGCTC-3 0 ) on a Veriti 96-well fast thermal cycler (Applied Biosystems) with the following conditions: 1 cycle of 958C for 2 min; 40 cycles of 958C for 10 s, 608C for 15 s; 728C for 75 s and 728C for 7 min. Products were isolated using the Promega Wizardt SV Gel and PCR Clean Up System and sequenced by the University of Wisconsin Biotechnology Center (Madison, WI) Sequencing Facility using nested sequencing primers (Exon 5a: 5 0 -GGGTCCCCATAAAATCCTTG-3 0 and 5 0 -GGACACAGGGAGGACATTCC-3 0 ; Exon 5b: 5 0 -CTCACCTACCAGTTCAGCGG-3 0 and 5 0 -TAAGAGATCGTGCACACGGG-3 0 ).
312
GENIK ET AL.
Microsatellite Instability and Flt3-ITD Analysis For detection of microsatellite instability (MSI), a multiplexed panel of markers designed specifically for CBA/Ca and C57BL/6 mouse strains was used (17). The CBA/Ca multiplex contains mononucleotide repeat loci mBAT-56, mBAT-57, mBAT-58, mBAT-59, mBAT64 and mBAT-66; and the C57BL/6 multiplex contains mBAT-56, mBAT-56h, mBAT-57, mBAT-58, mBAT-59 and mBAT-66. For MSI analysis 1 ng of spleen DNA was amplified with 13 CBA/Ca multiplex or 13 C57BL/6 multiplex, 13 Gold ST*R Buffer (Promega), and 0.2 units GoTaq Hot Start Polymerase (Promega) in a GeneAmp PCR System 9700 Thermocycler (Applied Biosystems) with the following amplification conditions: 958C for 2 min; 10 cycles of 948C for 30 s, 588C for 30 s (29% ramp), 658C for 1 min (28% ramp); 20 cycles of 908C for 30 s, 588C for 30 s (29% ramp), 658C for 1 min (28% ramp), 608C for 30 min. Products were denatured in deionized formamide with Internal Lane Standard 600 (Promega) for allele sizing and analyzed on a 3130xl Genetic Analyzer using GeneMapper 4.0 Software (Applied Biosystems). Size shifts in microsatellite marker of 3 bp from the wild-type allele was considered MSI positive. The presence or absence of internal tandem duplications in the Flt3 receptor tyrosine kinase gene was assayed by PCR (12). Flt3-ITD mutations were detected using primers amplifying the region covering exons 14 and 15 (5 0 -Fluorescein-GCAATTTAGGTACGAGAGTCAGC-3 0 and 5 0 -CTTTTAGCATCTTCACCGCCACC-3 0 ) using the same PCR and analysis conditions described for MSI analysis. The presence of Flt3-ITD is detectible by an increase in amplicon size of 3–400 bp (16). Statistical Analyses Differences in AML incidences between groups were evaluated using Fisher’s exact test (two-sided) and differences in latency were examined using unpaired t tests. Analyses were run using GraphPad Prism software (GraphPad Software Inc., La Jolla, CA).
RESULTS
AML Incidence
Most engineered laboratory mouse strains are maintained in small numbers, and only for their useful breeding life or for use in short-term experiments, so the lack of reports of spontaneous AMLs in PU.1þ/– mice might simply be the consequence of too few mice monitored for too short a time. Therefore, we monitored PU.1þ/þ and PU.1þ/– mice on the original B6;129 mixed background described by McKercher et al. (9) to 800 days of age (Table 1). Unirradiated PU.1þ/þ mice had a 6% incidence of AML that was not significantly increased by irradiation (P ¼ 0.3199). PU.1þ/– mice had a higher incidence of spontaneous AML than PU.1þ/þ mice (15% and 6%, respectively, P ¼ 0.0446). However, we characterized 16 of the 17 leukemias that arose in the unirradiated PU.1 þ/– mice by LOH assays and HRM analysis (Table 1) and found that these leukemias did not contain PU.1 codon 235 mutations. No evidence was found for mutation of the wild-type PU.1 allele in 15 of the 16 AMLs analyzed. The remaining AML lost the wild-type allele (reported as KO/Del2 in Table 1, though the mechanism of loss has not been determined). Thus, spontaneous leukemias in the B6;129S PU.1þ/– mice, with
one exception, did not arise through the classical PU.1 biallelic inactivation pathway. While irradiation did not increase the AML incidence in PU.1þ/þ mice, it more than doubled the incidence in PU.1þ/– mice (P ¼ 0.0002). To determine if radiation was responsible for the PU.1 codon 235 mutations typically observed in rAML, we assayed 36 leukemias from irradiated PU.1þ/– mice and found that 30 of them could be explained by biallelic PU.1 loss: 29 by LOH for the wildtype allele and one by deletion of the knockout allele and a codon 235 point mutation in the wild-type allele. These results indicate that radiation does not cause the PU.1 codon 235 mutation in rAML or, alternatively, that the codon 235 mutation cannot cooperate with the knockout allele to cause AML. Since only one of the 52 characterized AMLs arising in the B6;129S PU.1þ/– mice had a PU.1 codon 235 mutation, we next considered if AMLs with that mutation develop only in strains that are susceptible to rAML, such as CBA. The Sfpi1tm1Ram knockout allele used in this study is maintained on a mixed 129S2/SvPas and C57BL/6 background. C57BL/6 mice are rAML resistant and the 129S2/SvPas substrain has not been phenotyped for rAML susceptibility. Since F1 hybrid mice of a rAML-resistant and sensitive strain (i.e., CBA and C57BL/6) have intermediate rAML susceptibility (4.5–9%) (19, 20), we determined the incidence of AML and the frequency of AMLs bearing PU.1 codon 235 mutations in (CBA3B6;129S)F1 PU.1þ/þ and PU.1þ/– mice on the assumption that this background would be moderately rAML sensitive. None of the unirradiated (CBA3B6;129S)F1 PU.1þ/þ mice developed AML, an expected finding in light of the very low incidence of spontaneous AML in the CBA strain (,1 in 1,000) (20). Only 2 of 100 unirradiated (CBA3B6;129S)F1 PU.1þ/– mice developed AML which was not significantly different than the incidence in (CBA3B6;129S)F1 PU1þ/þ mice (P ¼ 0.2093) and demonstrates that strain background cannot fully account for the lack of reports of AMLs in PU.1 þ/– mice, since (CBA3B6;129S)F1 are sensitive to radiation-induced AML. After irradiation, 8% of the (CBA3B6;129S)F1 PU.1þ/þ mice developed rAML, indicating that this genetic background has intermediate susceptibility to radiation-induced AML, as expected. LOH and HRM analyses of eight of these leukemias revealed that they arose by chromosome 2 deletions and that six of them also had codon 235 mutations in the remaining PU.1 allele, consistent with previous reports of biallelic PU.1 inactivation by deletion and point mutation in rAML. In agreement with a previous report by Clark et al. (14), no bias was observed in loss of CBA/CaJ or C57BL/6J alleles in these leukemias, suggesting that strain-specific chromosome 2 fragile sites or cis-acting polymorphisms are not involved in strain differences in rAML susceptibility. The codon 235 point mutations detected by HRM analysis were verified by sequencing. Irradiated (CBA3B6;129S)F1 PU.1þ/– mice had a 28% incidence of rAML, significantly higher than that of
313
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
TABLE 1
Sample
PU.1
Dose (Gy)
Mice
B6;129S B6;129S B6;129S B6;129S (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1
þ/þ þ/– þ/þ þ/– þ/þ þ/– þ/þ þ/–
0 0 3 3 0 0 3 3
100 110 109 129 118 100 106 98
AML (%) 6 17 11 48 0 2 9 27
(6) (15) (10) (37) (0) (2) (8) (28)
Latency in days (range) 581 579 570 452
(478–707) (373–813) (324–693) (160–741) nd 706 (700–712) 548 (450–620) 516 (224–729)
PU.1 genotype/number of leukemias tested KO þ Del2
KO þ wt
Del2 þ R235
Del2 þ wt
na 1/16 na 29/36 nd 0/1 0/8 20/26
na 15/16 na 6/36 nd 0/1 0/8 2/26
0/6 0/16 0/9 1/36 nd 0/1 6/8 3/26
na 0/16 na 0/36 nd 1/1 2/8 1/26
Notes. Chromosome 2 deletions (Del2) were determined by LOH analysis, presence of the PU.1 knockout (KO) allele by PCR genotyping and presence of the codon 235 mutation by HRM analysis and sequencing. Del2 status could not be determined for B6;129 þ/ þ AML samples because these tumors were homozygous for the B6 allele of the LOH markers. na ¼ not applicable; nd ¼ no data.
irradiated (CBA3B6;129S)F1 PU.1þ/þ (P ¼ 0.0004) mice. Twenty-three of the 26 leukemias analyzed arose through biallelic inactivation of PU.1, most commonly through a chromosome 2 deletion that resulted in loss of the wild-type PU.1 allele, but in three cases through deletion of the knockout allele and codon 235 mutation on the wild-type allele. The remaining three leukemias had inactivation of one PU.1 allele either through deletion or retention of the knockout allele, but no codon 235 or other PU.1 mutations were detected on the wild-type alleles. However, there is of course the possibility that the wild-type alleles were inactivated by a mutation in a regulatory region or epigenetic gene silencing. AML Latency
The latencies for leukemias resulting from chromosome 2 deletions and PU.1 codon 235 mutations in irradiated (CBA3B6;129S)F1 PU.1þ/þ and PU.1þ/– mice (N ¼ 9, mean 6 SEM, 584.9 6 25.28) were not significantly different (P ¼ 0.0646) than those resulting from a chromosome 2 deletion and a PU.1 knockout allele (N ¼ 20, mean 6 SEM, 484.2 6 32.93). Microsatellite Instability
Microsatellite instability has been implicated in human leukemia and lymphoma (21, 22), as well as in rAML in CBA mice (14, 15). Therefore, the MSI status of AML tumors was investigated as a possible alternate mutator mechanism to help explain leukemogenesis in B6;129S mice that did not exhibit the expected biallelic PU.1 mutations. MSI in one or more of the microsatellite markers was found in 11 of 35 (31%) of AMLs from (CBA3B6;129S)F1 mice (Table 2), in close agreement with the 33% reported in AMLs from CBA/Ca mice (17). In contrast, MSI in spontaneous and radiation-induced AML from B6;129S mice occurred in only 2 of 66 cases (3%) (P ¼ 0.0001). Lymphomas from the same cohort of B6;129S mice did not exhibit any evidence of MSI, indicating that MSI was not
simply a result of clonal expansion of radiation-induced mutations in the microsatellite loci tested. The low level of MSI observed in tumors from B6;129S mice suggest that MSI is not a major contributor to leukemogenesis in this genetic background. Flt3 Mutations
PU.1 expression is decreased in about 20–25% of human AML cases due to internal tandem duplications in the FLT3 gene (23) and decreased expression has also been found in some murine rAML lacking PU.1 mutations (16). Therefore, AML samples from both the (CBA3B6;129S)F1 and B6;129S mice were screened for presence of Flt3-ITD by PCR. No evidence of Flt3 mutations was observed in any of the tumors. DISCUSSION
The lack of reports of PU.1þ/– mice developing AML challenges a model of radiation leukemogenesis in which biallelic inactivation of PU.1 results from radiation-induced deletion of one allele and spontaneous mutation at codon 235 on the other. We tested whether insufficient monitoring times or a rAML-resistant strain background could account for the lack of reports, or if radiation exposure was required for the codon 235 mutation. PU.1þ/– mice on their original B6;129S background developed AML if they were monitored long enough, however, codon 235 mutations were not involved. Irradiating B6;129S PU.1þ/– mice increased AML incidence by 2.5-fold, but rather than radiation exposure causing leukemias with codon 235 mutations, most of the resulting AMLs lost the wild-type allele of PU.1 by deletion. Only a single codon 235 mutation was detected in 36 leukemias analyzed. This indicates that radiation does not cause the codon 235 mutation, at least not on this strain background. Alternatively, the codon 235 mutation may not be able to complement the PU.1 knockout allele for the induction of leukemia.
314
GENIK ET AL.
TABLE 2 Sample
PU.1
Dose (Gy)
AML tested
B6;129S B6;129S B6;129S B6;129S (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1 (CBA3B6;129S)F1
þ/þ þ/– þ/þ þ/– þ/þ þ/– þ/þ þ/–
0 0 3 3 0 0 3 3
6 16 10 34 0 1 8 26
To examine the effect of strain background, we generated hybrid mice that were susceptible to rAML and carried one copy of the PU.1 knockout allele. Of the 100 unirradiated mice of this genotype that were monitored, only two developed AML. One AML was recovered and assayed and was found to lack the codon 235 mutation. Once again, the conclusion would be that the codon 235 mutation does not arise spontaneously or that it does not complement the PU.1 knockout allele for the induction of leukemia. Irradiating mice of this genotype increased the AML incidence by 14fold. Most of the leukemias harbored the knockout allele and a chromosome 2 deletion, although three of the 26 leukemias assayed resulted from deletion of one copy of PU.1 coupled with codon 235 mutation on the remaining allele. Taken together these results indicate that leukemogenesis involving codon 235 is strain background-dependent (1 of 67 AMLs in B6;129S mice compared to 9 of 35 AMLs in the F1 hybrids, P ¼ 0.0002). Because no examples of leukemia resulting from the R235 mutation in conjunction with the knockout allele were observed, the results do not resolve whether the codon 235 mutation is spontaneous or radiation induced. It is possible that radiation may have caused codon 235 mutations, but they were not detected because they could not complement the knockout allele in leukemogenesis. Two additional findings emerged from the results of these experiments. First, AML arises by different pathways in (CBA3B6;129S)F1 and B6;129S mice and second, the PU.1 codon 235 mutation commonly found in rAML does not appear to alter latency and therefore is not a major ratelimiting step in radiation leukemogenesis. While we showed that B6;129S PU.1þ/– mice have a higher spontaneous incidence of AML than PU.1þ/þ mice on the same background, however, with only one exception, we found no evidence for biallelic PU.1 loss. AMLs in B6;129S mice are characterized by microsatellite stability, an absence of Flt3 internal tandem duplications and long latencies (160–813 days, mean of 504 days). The failure to identify PU.1 mutations does not rule out the possibility that other mechanisms decrease or alter the function of PU.1 encoded by the wild-type PU.1 allele in these tumors or that haploinsufficiency for PU.1 enables an alternate leukemogenesis pathway. For example, Walter et al. (24) crossed
MSI (%) 0 1 0 1
(0) (6) (0) (3) – 1 (100) 3 (38) 7 (27)
Lymphomas tested 10 9 9 6 1 7 8 14
MSI (%) 0 0 0 0 0 1 2 2
(0) (0) (0) (0) (0) (14) (25) (14)
mice transgenic for human RARa-PML that occasionally develop acute promyelocytic leukemia (APL) with PU.1þ/– mice and found that progeny with the knockout allele had a greatly increased APL incidence. Additional experiments provided evidence that RARa-PML also reduced PU.1 mRNA levels from the wild-type allele. B6;129S mice can be manipulated to develop AML through the biallelic PU.1 inactivation pathway. We found that irradiated B6;129S PU.1þ/– mice had a high incidence of AML resulting from biallelic PU.1 inactivation which reinforces previous findings that leukemogenesis can be forced down this pathway even in a strain in which it does not normally occur (11, 12). Conversely, (CBA3B6;129S)F1 mice do not develop spontaneous AML, even if one allele of PU.1 is inactivated, the AML incidence is very low (2 of 100 PU.1þ/– mice followed on this background). However, AMLs can be induced in these mice by radiation exposure and the majority of these tumors have biallelic PU.1 inactivation. Interestingly, these leukemias also lack Flt3-ITD mutations, have long latencies, and they are frequently MSI positive. Interestingly, we found no leukemias that had biallelic PU.1 codon 235 mutations, perhaps because the encoded protein retains partial function (5). If so, levels of the partially functional protein could be high enough to prevent leukemogenesis if two copies of the mutant gene are present, but not if only a single copy is present, similar to class 2 low-penetrance RB1 mutations found in some retinoblastoma families (25). Also interesting is that none of the leukemias in PU.1þ/– mice that retained the knockout allele lost the wild-type PU.1 allele through codon 235 mutation. This suggests that hemizygosity for another gene (or other genes) in the minimally deleted region of chromosome 2 is required for rAML development involving the codon 235 mutation. Irradiated (CBA3B6;129S)F1 PU.1þ/– mice have a higher AML incidence than irradiated (CBA3B6;129S)F1 PU.1þ/þ mice. This is not surprising since most of these AMLs arise by biallelic PU.1 inactivation, with either a chromosome 2 deletion or a codon 235 mutation complementing the knockout allele. More cells are at risk for leukemic transformation from a radiation-induced chromosome 2 deletion in PU.1þ/– mice in which all of the target cells have
LEUKEMOGENESIS IN HETEROZYGOUS PU.1 KNOCKOUT MICE
an inactivated PU.1 allele than in PU.1þ/þ mice in which only those cells that have a codon 235 mutation concurrently with the deletion are at risk. Finally, the knockout allele increased rAML incidence it does not significantly decrease its latency. Since germline inactivation of one copy of a rate limiting tumor suppressor gene (in this case the knockout allele) should decrease tumor latency (26), one conclusion is that the PU.1 codon 235 mutation is not a major rate limiting step in rAML. This observation raises the possibility that codon 235 has a high mutation rate and the mutation is present around the time of irradiation. ACKNOWLEDGMENTS This work was supported by grant NNX09AM08G from the National Aeronautics and Space Administration entitled, ‘‘NASA Specialized Center of Research on Radiation Carcinogenesis.’’ We wish to thank the staff at Laboratory Animal Resources (Colorado State University) for excellent animal care. Received: March 17, 2014; accepted: May 30, 2014; published online: July 30, 2014
REFERENCES 1. Major IR, Mole RH. Myeloid leukaemia in x-ray irradiated CBA mice. Nature 1978; 272:455–6. 2. Mole RH, Papworth DG, Corp MJ. The dose-response for x-ray induction of myeloid-leukemia in male CBA/H mice. Br J Cancer 1983; 47:285–91. 3. Weil MM, Bedford JS, Bielefeldt-Ohmann H, Ray FA, Genik PC, Ehrhart EJ, et al. Incidence of acute myeloid leukemia and hepatocellular carcinoma in mice irradiated with 1 GeV/nucleon (56)Fe ions. Radiat Res 2009; 172:213–9. 4. Cook WD, McCaw BJ, Herring C, John DL, Foote SJ, Nutt SL, et al. PU.1 is a suppressor of myeloid leukemia, inactivated in mice by gene deletion and mutation of its DNA binding domain. Blood 2004; 104:3437–44. 5. Suraweera N, Meijne E, Moody J, Carvajal-Carmona LG, Yoshida K, Pollard P, et al. Mutations of the PU.1 Ets domain are specifically associated with murine radiation-induced, but not human therapy-related, acute myeloid leukaemia. Oncogene 2005; 24:3678–83. 6. Peng Y, Brown N, Finnon R, Warner CL, Liu X, Genik PC, et al. Radiation leukemogenesis in mice: loss of PU.1 on chromosome 2 in CBA and C57BL/6 mice after irradiation with 1 GeV/nucleon 56Fe ions, X rays or gamma rays. Part I. Experimental observations. Radiat Res 2009; 171:474–83. 7. Duncan BK, Miller JH. Mutagenic deamination of cytosine residues in DNA. Nature 1980; 287:560–1. 8. Dekkers F, Bijwaard H, Bouffler S, Ellender M, Huiskamp R, Kowalczuk C, et al. A two-mutation model of radiation-induced acute myeloid leukemia using historical mouse data. Radiat Environ Biophys 2011; 50:37–45. 9. McKercher SR, Torbett BE, Anderson KL, Henkel GW, Vestal DJ, Baribault H, et al. Targeted disruption of the PU.1 gene results in multiple hematopoietic abnormalities. Embo Journal 1996; 15:5647–58.
315
10. Scott EW, Simon MC, Anastasi J, Singh H. Requirement of transcription factor Pu.1 in the development of multiple hematopoietic lineages. Science 1994; 265:1573–7. 11. Metcalf D, Dakic A, Mifsud S, Di Rago L, Wu L, Nutt S. Inactivation of PU.1 in adult mice leads to the development of myeloid leukemia. Proc Natl Acad Sci U S A 2006; 103:1486–91. 12. Rosenbauer F, Wagner K, Kutok JL, Iwasaki H, Le Beau MM, Okuno Y, et al. Acute myeloid leukemia induced by graded reduction of a lineage-specific transcription factor, PU.1. Nature Genet 2004; 36:624–30. 13. Ban N, Kai M. Implication of replicative stress-related stem cell ageing in radiation-induced murine leukaemia. Br J Cancer 2009; 101:363–71. 14. Clark DJ, Meijne EI, Bouffler SD, Huiskamp R, Skidmore CJ, Cox R, et al. Microsatellite analysis of recurrent chromosome 2 deletions in acute myeloid leukaemia induced by radiation in F1 hybrid mice. Genes Chromosomes Cancer 1996; 16:238–46. 15. Curik N, Burda P, Vargova K, Pospisil V, Belickova M, Vlckova P, et al. 5-azacitidine in aggressive myelodysplastic syndromes regulates chromatin structure at PU.1 gene and cell differentiation capacity. Leukemia 2012; 26:1804–11. 16. Finnon R, Brown N, Moody J, Badie C, Olme CH, Huiskamp R, et al. Flt3-ITD mutations in a mouse model of radiation-induced acute myeloid leukaemia. Leukemia 2012; 26:1445–6. 17. Steffen LS, Bacher JW, Peng Y, Le PN, Ding LH, Genik PC, et al. Molecular characterisation of murine acute myeloid leukaemia induced by 56Fe ion and 137Cs gamma ray irradiation. Mutagenesis 2013; 28:71–9. 18. Hirouchi T, Takabatake T, Yoshida K, Nitta Y, Nakamura M, Tanaka S, et al. Upregulation of c-myc gene accompanied by PU.1 deficiency in radiation-induced acute myeloid leukemia in mice. Exp Hematol 2008; 36:871–85. 19. Boulton E, Cleary H, Papworth D, Plumb M. Susceptibility to radiation-induced leukaemia/lymphoma is genetically separable from sensitivity to radiation-induced genomic instability. Int J Radiat Biol 2001; 77:21–9. 20. Meijne E, Huiskamp R, Haines J, Moody J, Finnon R, Wilding J, et al. Analysis of loss of heterozygosity in lymphoma and leukaemia arising in F1 hybrid mice locates a common region of chromosome 4 loss. Genes Chromosomes Cancer 2001; 31:373– 81. 21. Wimmer K, Etzler J. Constitutional mismatch repair-deficiency syndrome: have we so far seen only the tip of an iceberg? Hum Genet 2008; 124:105–22. 22. Wimmer K, Kratz CP. Constitutional mismatch repair-deficiency syndrome. Haematologica 2010; 95:699–701. 23. Schnittger S, Schoch C, Dugas M, Kern W, Staib P, Wuchter C, et al. Analysis of FLT3 length mutations in 1003 patients with acute myeloid leukemia: correlation to cytogenetics, FAB subtype, and prognosis in the AMLCG study and usefulness as a marker for the detection of minimal residual disease. Blood 2002; 100:59–66. 24. Walter MJ, Park JS, Ries RE, Lau SK, McLellan M, Jaeger S, et al. Reduced PU.1 expression causes myeloid progenitor expansion and increased leukemia penetrance in mice expressing PMLRARalpha. Proc Natl Acad Sci U S A 2005; 102:12513–8. 25. Harbour JW. Molecular basis of low-penetrance retinoblastoma. Arch Ophthalmol 2001; 119:1699–704. 26. Knudson AG, Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A 1971; 68:820–3.
