River Inc., St. Constant, Quebec, Canada, respectively. Newborn mice (less than ... The cells were cultured in DME medium (Dulbecco modified Eagle me- dium ...
JOURNAL OF VIROLOGY, Aug. 1993, p. 4722-4731
Vol. 67, No. 8
0022-538X/93/084722-10$02.00/0 Copyright X 1993, American Society for Microbiology
Graffi Murine Leukemia Virus: Molecular Cloning and Characterization of the Myeloid Leukemia-Inducing Agent MING RU,' CHAIM SHUSTIK,2 AND ERIC RASSARTl* Departement des Sciences Biologiques, Universite du Quebec a Montreal, 1 Montreal, Quebec H3C 3P8,1 and Division of Hematology, Royal Victoria Hospital, and McGill Cancer Center, McGill University, Montreal, Quebec H3A lA1, 2 Canada Received 19 January 1993/Accepted 29 April 1993
The Graffi murine leukemia virus (MuLV) is a retroviral mixture that induces predominantly myeloid leukemia in several inbred strains of mice. To analyze the viral component responsible for the myeloid leukemogenesis, we cloned several proviruses from a Graffi MuLV-infected cell line. Several infectious molecular clones were obtained that could be classified into two distinct groups of infectious MuLV. Both types of MuLV were nondefective, ecotropic, and NB tropic and induced granulocytic leukemia in BALB/c and NFS mice. Restriction enzyme analysis and molecular hybridization with several MuLV probes on one molecular clone from each group revealed that both groups are closely related to each other but are clearly distinct from all known retroviruses. One component of MuLV, however, induced leukemia with a shorter latency period and harbored a lengthier long terminal repeat. The long terminal repeat of the more leukemogenic component of MuLV had acquired a 60-bp perfect duplication in the U3 region. Analysis of the tumor DNAs with probes for the mouse T-cell receptor and immunoglobulin heavy chain genes revealed frequent rearrangements with one or both probes. This concomitant expression by leukemic cells of markers of different lineages, observed in human leukemias, has been termed "lineage infidelity" and confirms that the latter rearrangements are not restricted to hematopoietic precursors committed to lymphoid differentiation. AKR mice induced predominantly myeloid leukemias (21). Also, by biological cloning, Fischinger et al. (22) identified a highly leukemogenic ecotropic component which induced thymic lymphoma but not myeloid leukemia in C57BL mice. These studies, which have been limited to morphological and histochemical analysis of the leukemic cell phenotype, suggested that the Graffi virus could be a mixture of viral components. In order to identify and analyze the virus responsible for the myeloid leukemia in the preparation, we cloned several infectious genomes derived from the Graffi MuLV stock, characterized them biologically, and sequenced the long terminal repeats (LTR) of two distinct infectious clones.
Murine leukemia retroviruses (MuLV) that induce lymphocytic leukemia or lymphoma (1, 38, 44, 45) and erythroid leukemia (23, 47) have been extensively investigated (for a review, see reference 56). By contrast, studies of myeloid leukemia-inducing retroviruses have been limited. Leukemias of myeloid origin have been induced by irradiation or by injection of cell-free tumor filtrates (55, 57). The Friend MuLV and the Cas-Br-E MuLV occasionally induce myeloid leukemia (13, 14, 33). In the BXH-2 recombinant inbred mouse strain, a B-tropic recombinant MuLV is associated with a high incidence of granulocytic leukemia (4). In pristane-pretreated adult mice, promonocytic leukemia developed following Moloney MuLV (Mo-MuLV) and amphotrophic 4070A MuLV injection (39, 59). However, none of these studies characterized a naturally occurring retrovirus that induces exclusively myeloid leukemia in mice. Graffi initially reported in 1957 that cell-free supernatants of Ehrlich sarcoma and derivative tumors induced chloroleukemias with a high frequency in various inbred and outbred mouse strains (24). In subsequent studies, Graffi et al. observed a spectrum of leukemic cell types after repeated passage of the putative viral agent; this phenomenon was described as hematological diversification (25). The leukemic cell type observed in these experiments was influenced by the corresponding lineage of the tumors used for the viral extract preparation. The influence of host factors on leukemic lineage was seen in experiments by Fiore-Donati et al. (20, 21). Inoculation of AKR mice with Graffi virus caused lymphocytic leukemia with a short latent period, but cellfree extracts of these leukemias could induce a variety of leukemic cell types, including myeloid and undifferentiated leukemias, in C57BL, DBA/2, and C3Hf/Gs recipients. Interestingly, inoculation of Graffi virus into thymectomized
*
MATERIALS AND METHODS Mice. NFS and BALB/c mice were used in leukemia development studies and were purchased from the Frederick Cancer Research Facility, Frederick, Md., and from Charles River Inc., St. Constant, Quebec, Canada, respectively. Newborn mice (less than 48 h old) were inoculated intraperitoneally with approximately 5 x 104 PFU in 0.1 to 0.2 ml of filtered viral supernatant. The inoculated mice were observed daily and sacrificed when they showed signs of advanced disease. Cell lines and viruses. Graffi virus-maintaining cell line (Graffi cell line) was kindly given by Natalie Teich (Imperial Cancer Research Fund, London, England). The mouse fibroblast cells NIH 3T3 and BALB/c, the rat XC cells and Rat-1 fibroblasts (American Type Culture Collection), and mink lung cell lines (American Type Culture Collection) were used for viral host tropism analysis. The cells were cultured in DME medium (Dulbecco modified Eagle medium; GIBCO) supplemented with 10% calf serum, penicillin G (100 U/ml), and streptomycin (100 ,ug/ml). They were grown at 37°C in a humidified atmosphere with 5% CO2.
Corresponding author. 4722
VOL. 67, 1993
Virus infection was done by using fibroblasts that were pretreated with 20,ug of Polybrene per ml for 1 h before virus infection. Molecular cloning of integrated viral DNA genomes. The DNA of the Graffi cell line was digested with EcoRI. DNA fragments were sedimented in a sucrose density gradient (5 to 20%), and fractions containing the 9- to 20-kbp fragments were pooled. DNA (0.5,ug) was then ligated to the EcoRIcleaved lambda-DASH vector (Stratagene) and packaged. The MuLV-containing clones were identified by using a 32P-labeled U3 probe derived from Mo-MuLV and purified. Transfection procedure. Molecularly cloned viral DNAs from the Graffi cell line were transfected onto NIH 3T3 fibroblasts by the calcium phosphate procedure (26). Emerging viruses were detected by the XC assay (48) or the reverse transcriptase assay and recovered from the supernatant by rapid harvest, and the titers were determined by the XC assay. Probes. Four probes were used from Mo-MuLV: an 8.2kbp full-length (representative) probe comprising a permuted DNA molecule cloned in the HindIII site of pBR322 (obtained from D. Baltimore), a 380-bp U3-specific fragment corresponding to the U3 region of the Mo-MuLV LTR (45), a 850-bp gag probe derived from PstI to XhoI (positions 739 to 1560), and a 1.05-kbp env probe derived from BamHI-toClaI sites (positions 6537 to 7674). The E/U probe describes a 1.2-kbp YI4vnI fragment derived from Graffi molecular clone GV-1.2 and included the 3' part of gp70, the complete plSE, and the U3 LTR region. The murine T-cell receptor n-chain (TCR-13) gene was analyzed with a 730-bp PstI DNA fragment (RBL-5-17) containing most of the murine C 1-coding sequence (10). The immunoglobulin heavy chain (IgH) region was analyzed with a 0.6-kbp EcoRI-XhoI JH probe (3) and the immunoglobulin kappa light chain region with a 2.7-kbp HindIII JK probe (37). Labeling of purified fragments was done by the random primer extension method with oligohexamers (18) (Pharmacia P-L Biochemicals, Montreal, Canada). DNA analysis and hybridization. DNAs from molecular clones or extracted from tumors were digested with restriction endonucleases and analyzed by Southern transfer (52). Specific DNA fragments were detected by hybridization with appropriate 32P-labeled DNA probes. Briefly, nylon membranes (Hybond-N; Amersham) were hybridized in 50% formamide-10% dextran sulfate-1% sodium dodecyl sulfate (SDS)-1 M sodium chloride at 42°C. After hybridization, the filters were washed sequentially in 2x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at 20°C for 5 min and 0.2x SSC-0.1% SDS at 65°C for 1 h. Membranes were exposed at -80°C to Kodak XAR-5 film by using a Cronex lightning-plus intensifying screen (DuPont Co., Wilmington, Del.). DNA sequence analysis. Sequencing of the LTR of the Graffi MuLVs was performed by subcloning appropriate restriction fragments from infectious molecular clones (GV1.2 and GV-1.4) in Bluescript vector. The sequence was determined according to the method of Sanger et al. (50) by using universal and reverse primers (Pharmacia, Montreal, Canada). Characterization of leukemia. The diagnosis of leukemia was based on the morphology and cytochemical stains (5, 54, 60) of peripheral blood smears and leukemic tissue imprints. Smears of peripheral blood and imprints of fresh tissue were stained with Wright-Giemsa for morphological examination. Myeloperoxidase (60) and dual-esterase with a-naphthylbutyrate (Sigma) and naphthol AS-D chloroacetate (Sigma)
GRAFFI MuLV AND MYELOID LEUKEMIA
4723
staining was done according to the method described by Swirsky (54). Cell surface antigen markers were analyzed by cell flow cytometry. Single cell suspensions prepared from spleen and/or lymph nodes were obtained by grinding the organs through a fine wire mesh. One million cells were washed three times with phosphate-buffered saline (PBS) and then incubated with fluorescein isothiocyanate-conjugated monoclonal antibodies in an ice bath for 30 min. Cells were then washed with 500 ,ul of ice cold PBS three times. The cells (104) were analyzed by flow cytometry (fluorescence-activated cell sorter; Becton Dickinson). The monoclonal antibodies used in this study were Thy-1.2 (rat anti-mouse; Becton Dickinson) as a pan-T-cell marker and goat anti-mouse surface immunoglobulin and Ly-5 (B220) as B and pre-B cell markers, respectively (for a review, see reference 32). RESULTS Molecular cloning and restriction endonuclease analysis of Graffi MuLV genomes. Since an ecotropic virus is thought to be the leukemogenic agent in the Graffi MuLV preparation, we cloned the ecotropic proviruses of the Graffi cell line. High-molecular-weight DNA was extracted from the Graffi cell line and digested with EcoRI, and fragments ranging in length between 9 and 20 kbp were pooled and ligated to the EcoRI-digested lambda-DASH vector. Because of the nonecotropic endogenous MuLV sequences present in the mouse cell line, we had to choose a specific MuLV probe. For this purpose, we analyzed the unintegrated viral genomes purified from the Hirt supernatant of newly infected NIH 3T3 cells. The unintegrated DNA was analyzed with a series of probes that have been shown to be specific to ecotropic MuLV genomes. The probes used were a gp7O ecotropic MuLV-specific fragment (12), a TaqI-BamHI fragment specific to the Cas-Br-E envelope (44), and two LTRderived probes: one from VL3 radiation leukemia virus (46) and one from Mo-MuLV (46). Only the latter probe (U3Moloney) hybridized well to the unintegrated viral genome extracted from the Graffi cell line (results not shown) and was used to screen the lambda-DASH library. Thirty-nine lambda clones were obtained from this library, and physical mapping of the retroviral inserts was performed to determine the differences among the viral genomes. These could be classified into four groups of retroviral genomes. By restriction enzyme analysis and Southern hybridization, we found three groups of 12, 11, and 2 distinct clones that all contained two LTRs. One representative clone was chosen from each group for further study: GV-1.2, GV-1.4, and GV-2.5. The fourth group was composed of 14 clones that contained only one LTR and therefore represented incomplete viral genomes. These are most probably mink cell focus-forming virus-like or endogenous retroviral genomes since the molecular clones contained either the 5' end (eight clones) or the 3' end (six clones) of a typical retroviral structure carrying an EcoRI site at position 6.9. This last group was not further studied. Clones from the three groups of ecotropic MuLV were digested with several restriction enzymes, assayed on agarose gel, transferred to nylon membranes, and hybridized with either full-length MoMuLV (representative) or U3 LTR-derived probes (Fig. 1). When necessary, the membranes were also hybridized with gag and env probes derived from Mo-MuLV, and a E/U probe derived from Graffi clone GV-1.2 env and U3 region. A summary of the restriction maps is shown in Fig. 2. Clone GV-2.5 was found to be very similar to clone
J. VIROL.
RU ET AL.
4724 A a
tI
eL
;
;.
U b
sz_
h
a
b
d
C
e
r
g
K2iP
,
j
k
I
m
.-.O
-
n
-i
K_,Z7
a
p
r
q
t
u
go
_
m
0
a so'4w Ai _- A_t 'm_w _m w
s
~~~~~~~~~~~~~~~~~~~~~~~~~I'-m-
Z
Z*Z
h
m
_
_, Am Mr im
Aub.
-:4-U
a S
_ _-
.
lb
:A
.d M i
P
.4
-a3Sm,
FIG. 1. Southern blot analysis of restriction enzyme-digested Graffi MuLV molecular clones. DNA (0.5 ,ug) of clones GV-1.2 (lanes a, d, g, j, m, p, and s), GV-1.4 (lanes b, e, h, k, n, q, and t), and GV-2.5 (lanes c, f, i, 1, o, r, and u) in the lambda-DASH vector were digested with PstI (lanes a to c), PstI-KpnI (lanes d to f), KpnI (lanes g to i), KpnI-EcoRV (lanes j to 1), EcoRV (lanes m to o), EcoRV-SmaI (lanes p to r), and SmaI (lanes s to u) and hybridized with a Mo-MuLV-derived full-length (representative) probe (A) or with a U3 LTR-specific probe
(B).
compare lanes a and c). The same was observed for KpnI, with a shift from 1.5 to 1.15 kbp (Fig. 1A, compare lanes g and i). Again with SmaI, the difference in mobility was detected in the 1.95-kbp fragment of clone GV-1.2 that
GV-1.2 except for an approximately 400-bp deletion which was localized in the env region of the virus since it was detected in the PstI digestion on a 1.5-kbp fragment in clone GV-1.2 and a 1.15-kbp fragment in clone GV-2.5 (Fig. 1A,
GV-1.2
PlPISSK S .11 I Y . I
PEE Sc
Pi
GV-1.4
I
Pi
Si
.
II
S -I
I
I
I
SIl
I
S SiI Xh Ha Ya a A
Sc
I
PE
PUPUSSK
If
Si
SIl
I
I
LT R U3
--
SI
SIl I
I
.4I I
PK
P K
PI. I SSK I
S
It I
P K
SIl
H C ..I
P PISSK If I Y -C ± ± PEPE Sc
E PE
Sc
P PISSK
S I I
PK
asa w-I
I
S I I
PK
ScK E
P E
SIlI
ScK E
I I
I
C
I
Xh
I
PE PE Sc
I
I
ScK E
P E
I
SIl
H C
I
---
SSK
I; * *3 PE
GV-2.5
Xh
I
nri p4ug.
P K
E
PEE
Sc
LT R tnvw n
I
gag
env elu
0
2
3
4
5
6
7
8
kbp
FIG. 2. Restriction endonuclease map of Graffi MuLV clones GV-1.2, GV-1.4, and GV-2.5. The viral genomes of these three clones with their flanking sequences have been analyzed as described in Materials and Methods and in the legend to Fig. 1. Only the viral portion is represented here. The open boxes represent the LTR. The closed boxes represent sequences missing in one clone compared with another. The probes used are indicated as thick lines under the genetic map. C, Clal; E, EcoRV; H, HindIII; K, KpnI; P, PstI; PI, PvuI; S, SmaI; SII, SacII; SI, Sall; Sc, Sacl; X, XhoI.
GRAFFI MuLV AND MYELOID LEUKEMIA
VOL. 67, 1993
TABLE 1. Biological characteristics of molecularly cloned Graffi MuLV variants Virus
GV-1.2 GV-1.4 GV-2.5 GV-parental
Reverse transcriptase activity (103 cpm) in cell linea: Mink BALB/C Rat-1 NIH 3T3
2,200 3,400
1,990 1,180
895 1,020
5B
t-_.5
r__
F_¢
A
B am
qw, _*
"U,-w
_ 4w
"
1-fo
_~a ...
_
46*40 -.111,1111W
n -q.
ti' kk
_
IF
*
*t..
il~~~~~~~~~~~~~.....
I
..
_8~~~~~~~~~~~~~~~~~~.:.'".'.., _.;~
C D
of
W
I
w
_§#4
-6,4-
W
-
.:,
4m
'5_,-
0 4
LK
3S
L
r
