Retroviral transfer of a chimeric multidrug resistance-adenosine deaminase gene

Retroviral transfer of a chimeric multidrug resistance-adenosine deaminase gene URSULA

A. GERMANN,

Laboratory

of Molecular

Maryland

20892,

KHEW-VOON

Biology,

National

CHIN,

Cancer Institute,

fusion

between

(MDR1)

a

cDNA

cDNA

an

concomitantly

ADA

activity

Harvey

Infection

the

MDR-ADA

resistance

therefore

Coichicine-resistant

amounts

of

fusion

protein

tional

activities.

gene

on in the

cells

a membrane-associated that

preserved

both

in vivo,

Kirsten

MDR

of collarge

MDR-ADA

and

ADA

of the

virus-transformed

of

drug-sensitive presence

synthesized

expression

chi-

variety conferred

210-kDa

To monitor

a

this

a great

retrovirus

selection

chicine.

transduce

into

phenotype

allowing

and

produced

replication-de-

to

efficiently

with

multidrug

cells,

virus-derived,

gene

(ADA)

resistance

We have

retrovirus

MDR-ADA

cells.

cells.

sarcoma

resistance

deaminase

multidrug

transfected

recombinant

meric

multidrug

adenosine

confers

on

murine

fective,

the

selectable

and

NIH

func-

chimeric cells

were

infected with the MDR-ADA retrovirus, and after drugselection, injected into athymic nude mice. Tumors developed ADA

that

fusion

contained

protein.

the bifunctionally

When

placed in tissue culture did

not

protein.

lose

the

The

these

without

bifunctionally

mouse

active tumor

the selecting active

replication-defective,

MDR-ADA recombinant

MDR-

cells

were

drug, they fusion MDR-

retrovirus should be useful to stably introduce the chimeric MDR-ADA gene into a variety of cell types for biological experiments in vitro and in vivo.GERMANN, U. A.; CHIN, K. -V.; PASTAN, I.; GOTTESMAN, M. M. Retroviral transfer of a chimeric multidrug resistance-adenosine deaminase gene. FASEB j 4:

ADA

1501-1507; Key

Words:

deaminase

1990. multidrug fusion

gene

resistance .

P-glycoprotein

recombinant

adenosine

retroviru.s

TO MULTIPLE

0892-6638/90/0004-1501/$01

Institutes

AND

MICHAEL

of Health,

M.

GOTTESMAN’

Bethesda,

MDR1 gene, as well as the corresponding mouse or hamster mdr genes, encodes a 170-kDa protein, known as P-glycoprotein or multidrug transporter, which contains 12 transmembrane domains and functions as an ATP-driven effiux pump for many hydrophobic drug substrates (3-5). Both DNA-mediated and retroviral transfer of a full-length mdr cDNA confer the complete multidrug resistance phenotype to a great variety of drug-sensitive rodent and human cell lines (6-10). Transfected mdr and cotransfected DNA sequences are efficiently expressed and amplified during stepwise selection for drug resistance with drugs such as colchicine (11). We are investigating the usefulness of the human MDR1 gene as a dominant selectable, potentially amplifiable marker in a new approach for gene transfer experiments involving chimeric genes. Our approach is to construct gene fusions involving a selectable marker and an unselected gene, which under selective pressure should ensure concomitant and efficient expression of both encoded protein moieties. If the functional activity of the two protein constituents is not impaired by the fusion, this should allow a simple and easy way to enrich for properly expressing cells within a transfected cell population. We have recently designed a fusion between a human MDR1 cDNA and a human adenosine deaminase (ADA)2 cDNA (12). In the encoded chimeric MDR-ADA gene product, the initiator-Met of the ADA moiety is connected to the carboxyl terminus of the multidrug transporter by the tripeptide Gly-Arg-Pro. We used a recombinant retroviral expression vector to stably introduce the MDR-ADA fusion gene into drug-sensitive human KB3-1 cells by calcium phosphate-mediated coprecipitation, followed by selection in the presence of colchicine. The stably transfected cells efficiently expressed the chimeric gene and synthesized a membrane-associated MDR-ADA fusion protein with the expected apparent mol wt of

ITO

drugs has prevented the successful treatment of human tumors with chemotherapeutic agents (reviewed in refs 1, 2). By growing cells in the presence of selecting drugs, a similar phenotype can be induced in vitro, which is associated with enhanced expression and amplification of members of the multidrug resistance (mdr) gene family (1, 2). The human RESISTANCE

PASTAN,

National

USA

ABSTRACT A

IRA

CYTOTOXIC

.50. © FASEB

whom

correspondence

should

be addressed,

at Laboratory

of

Molecular Biology, National Cancer Institute, NIH, Bldg. 37, 2E18, 9000 Rockville Pike, Bethesda, MD 20892, USA. 2Abbreviations: ADA, adenosine deaminase; DMEM, Dulbecco’s modified Eagle’s medium; PBS, phosphate-buffered saline; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; KNIH, Kirsten virus-transformed NIH cells; MDR, multidrug resistance; PNP, purine nucleoside phosphorylase; TS, tumor sample, TCC, tumor cell culture.

1501

210,000. The fusion protein retains the functional activity of both constituents: it confers the complete multidrugresistance phenotype to the cells and exhibits ADA activity. We therefore consider the bifunctional MDR-ADA chimera an attractive candidate for gene therapy of ADA deficiency, which represents an autosomal recessive genetic disorder and is associated with severe combined immunodeficiency (13-15). To study the biological properties of the MDR-ADA fusion protein in vitro and in vivo, we wanted to be able to transfer the MDR-ADA fusion gene easily and at high frequency into a wide variety of target cells. Genetic engineering has allowed the rapid development of retrovirus technology, which offers an efficient means of gene transfer (reviewed in ref 16). Highly transmissible recombinant retroviruses have been shown to infect a great variety of cells in culture (e.g., refs 17, 18). Occasionally, depending on the chromosomal site of retroviral integration, the presence of certain viral transcriptional control signals, or the differentiation process of infected host cells, instability of the transferred genes and/or transcription products, as well as low levels of expressed proteins, may cause a problem, particularly in long-term in vitro and in vivo studies (19-22). In gene transfer experiments involving the human MDR-ADA fusion gene, however, these problems should be overcome, since the system offers the possibility to select directly for its expression. Here we describe the generation of a Harvey murine sarcoma virus-based recombinant retrovirus carrying the MDR-ADA fusion gene as the only functional eukaryotic gene. We demonstrate efficient retroviral transfer and stable expression of the bifunctionally active MDR-ADA fusion gene product in vitro in NIH 3T3 and Kirsten virus-transformed NIH (KNIH) mouse cells, as well as in vivo, in mouse tumors induced by subcutaneous injection of MDR-ADA-expressing KNIH cells. MATERIALS

AND

tor pGEMSBX contains only three unique cloning sites in the order SacII-BamHI-X/zol. A 1.09-kb BanzHI-DdeI ADA cDNA fragment (encoding amino acids 20-363 and containing 54 noncoding bases at the 3’ end, but no poly(A) site) was inserted into BamHI-X/zol doubledigested pGEMSBX, together with a DdeI-XhoI adapter containing universal translation terminator sequences (i.e., TAATI’AATTAA). Then the fusion gene was reconstructed by inserting a 4.1-kb SacII-BamHI fragment obtained from pMDRADAXS (12) (encoding full-length P-glycoprotein, the linker tripeptide Gly-Arg-Pro, and amino acids 1-19 of ADA). Finally, the fusion gene, with the poly(A) site deleted, was excised as a 5.2-kb SacIIXhoI fragment and inserted into the single Sac! I and XhoI sites between the long-terminal repeats of the Harvey murine sarcoma virus-derived expression vector pCO1 (10). The final plasmid was designated pHaMDRADA/A (Fig. 1) and used to transfect the amphotropic packaging mouse cell line PA-12 (23) and the ecotropic packaging mouse cell line Psi-2 (24) by the calcium phosphate coprecipitation method, as described in ref 12. Transfected cells were selected in the presence of 60 ng/ml colchicine, and single colonies were isolated. Fresh growth medium was added when the cells reached near confluency. The medium was harvested 20-22 h later and titered for the presence of recombinant retrovirus by infecting NIH 3T3 indicator cells as described previously for a recombinant retrovirus carrying an MDR1 cDNA (10). The cell line PA-12-MDR-ADA/A-4 (viral titer of 6 x iO cfu/ml) was used to infect all cell lines described here. Host NIH 3T3 and KNIH mouse cells were infected at a multiplicity of infection (MOl) of < 0.1 in the presence of 2 tg/ml Polybrene, using a diluted 20-h viral harvest. After 48 h the growth medium was replaced with selective medium containing 80 ng/ml colchicine to enrich for drug-resistant cells. pBR322

METHODS

Cells PA-12 (23) and Psi-2 (24) mouse packaging cells were provided by Dr. D. Lowy (Nd, NIH) and maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 2 mM L-glutamine, 50 U/ml penicillin, 50 tg/ml streptomycin, and 10% fetal bovine serum. NIH 3T3 and KNIH mouse cells were obtained from Dr. C. Scher (University of Pennsylvania School of Medicine) and grown in DMEM containing 2 mM L-glutamine, 50 U/ml penicillin, 50 tg/ml streptomycin, and 10% calf serum.

Recombinant retrovirus production, infection of cultured cells

titration,

and

The poly(A) addition signal in the previously described (12) retroviral expression vector pHaMDRADA was deleted as follows. By using a synthetic oligonucleotide adapter, a new cloning vector (pGEMSBX) was derived from the commerically available pGEM2 vector. The vec1502

Vol. 4

March

1990

MP1 cDNA

KpnI

Figure 1. Structure of retroviral vector carrying human MDR-ADA fusion gene. Ha-MuSV LTR marks the long-terminal repeats of Harvey murine sarcoma virus; the direction of transcription is indicated by the arrow.

The FASEB Journal

GERMANN

ET AL.

Induction of tumor

of tumors on nude mice and establishment cell cultures

Each of the three KNIH cell lines (parental KNIH, retrovirus-infected KNIH-MDR, or KNIH-MDR-ADA cell line) was injected at a concentration of 2 x 106 cells in 200 il phosphate-buffered saline (PBS) into a single site in the flank of 10 athymic nude mice (NIH Swiss nu/nu). Tumors developed rapidly at the site of injection and reached an average diameter of 13-16 mm within 14 days. Two weeks after injection the mice were killed, and the tumors were removed and separated from the skin. Freshly collected tumors were either immediately frozen on dry ice and stored at -70#{176}C,before extraction and preparation of crude membranes, or were kept on ice and tumor cell cultures were generated as follows. The tumors were minced with sterile scissors and forceps into small pieces (average diameter 1-2 mm). Ten pieces per sample were transferred to a 60-mm tissue culture dish, covered with one drop of growth medium (DMEM, 2 mM L-glutamine, 50 U/mi penicillin, 50 tg/ml streptomycin, 10% calf serum), and incubated at 37#{176}C, 5% CO2. After 3 and 6 h another drop of medium was carefully added, and the plates were incubated overnight. The next day growth medium was added to a final volume of 5 ml. As soon as colony formation was visible to the eye (within 2-3 days), the tumor pieces were carefully removed. When the cultures reached near confluency, they were trypsinized and maintained by passaging twice a week.

DNA analysis DNA was prepared from confluent 75-cm2 flasks of cells, digested to completion with EcoRI, sizefractionated on a 0.7% agarose gel by electrophoresis, transferred to nitrocellulose membranes, and hybridized to the 32P-labeled MDR1-specific DNA probe HDR5A (25) as described (12). Genomic

Protein

analysis

Viable cells were photoaffinity-labeled with [3H]a.zidopine, extracted, and immunoprecipitated as previously described (12). Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed according to the method of Laemmli (26), and fluorography was done according to Bonner and Laskey (27). Radiometric ADA assays were carried out as described (28), using crude membrane preparations (12). Frozen tumor samples were pulverized on a metal surface on a bed of dry ice and extracted with 3 volumes (1 ml volume corresponding to 1 g wet weight) of extraction buffer (10 mM Tris-HC1, pH 7.5, 10 mM NaC1, 1 mM MgCl2, 1 mM phenylmethanesulfonylfluoride), as described previously (12). RESULTS

Generation

of transmissible

MDR-ADA

retrovirus

We have described the construction of a Harvey murine sarcoma virus-derived retroviral expression vector, called EXPRESSION OF AN MDR-ADA

FUSION

GENE

pHaMDRADA (12), which carries an MDR-ADA fusion gene as the only functional eukaryotic gene between the two long-terminal repeats. We stably introduced this plasmid into PA-12 packaging cells (23) by transfection, followed by seiection at a concentration of 60 ng/ml colchicine. The pHaMDRADA-transfected PA-12 cells, however, did not produce any detectable levels of recombinant retrovirus, presumably because the presence of the ADA poly(A) signal within the insert drastically reduced levels of full-length viral RNA (29). Therefore, we prepared a new retroviral expression vector as outlined in Materials and Methods by deleting part of the 3’ nontranslated region of the ADA cDNA (including the poly(A) addition signal) and replacing it with universal translation terminator sequences. The final piasmid, designated pHaMDRADA/A (Fig. 1), was used to transfect both the amphotropic packaging cell line PA-12 (23) and the ecotropic packaging cell line Psi-2 (24). Negative control cells received no DNA, whereas positive control cells were transfected with pHaMDR1/A, a plasmid that has recently been used to produce MDR1 retrovirus (10). As observed earlier (12), the transfection efficiencies for the construct containing the MDR-ADA fusion gene (2 x 10 for PA-12 cells, and 1.3 x iO for Psi-2 cells) were lower than for the one carrying the MDR1 gene alone (2.6 x 102 for PA-12 cells, and 4.7 x i0 for Psi-2 cells, respectively). Three isolated pHaMDRADA/A-transfected PA-12 clones and two Psi-2 clones, which grew at a concentration of 60 ng/ml colchicine, were analyzed by Southern hybridization (30) of genomic DNA with an MDR1specific probe, to confirm the intactness of integrated provirus (Fig. 2A). Furthermore, the relative amounts of expressed MDR-ADA fusion protein were compared by [3H]azidopine photoaffinity labeling of viable cells (31), and by analysis of total extracts by SDS-PAGE (26) and subsequent fluorography (27) (Fig. 2B). The viral titers were determined by testing 24-h viral harvests for their ability to confer colchicine resistance to NIH 3T3 indicator cells (Table 1). The analysis revealed higher viral titers for amphotropic producers. Furthermore, the viral titers and relative levels of MDR-ADA fusion protein varied considerably in the investigated producer clones, whereas the copy number of integrated provirus remained more or less constant.

Retroviral fusion

gene

transfer and expression of the MDR-ADA in NIH 3T3 and KNIH mouse cells

To evaluate retrovirus transmission and to determine whether the MDR-ADA cDNA sequences retained their biological activity, drug-sensitive NIH 3T3 fibroblasts and cells were infected with the MDR-ADA retrovirus obtained from the producer cell line PA-12-MDRADA/A-4 (6 x iO cfu/ml). Control experiments were performed using an MDR1 retrovirus produced by the previously isolated cell line PA-12-MDR1/A-1 (8 x 10 cfu/ml) (10). Two days after infection, the mouse cells were selected for the multidrug resistance phenotype by placing them in growth media containing 80 ng/ml colchicine. Drug-resistant cell populations were expanded 1503

12345678 23.1

-

9.46.64.4-

2.3

MDR-ADA MDR

-

12345678 200-

MDR-ADA

-

MDR

97-

68-

infected with the MDR-ADA retrovirus synthesized a larger protein with an apparent mol wt of 210,000, which was labeled by [3H]azidopine and reacted specifically both with multidrug transporter-specific and ADA-specific antisera, but not with preimmune serum. Similar results were obtained for the KNIH cell populations (data not shown). To confirm that the chimeric protein is membraneassociated and to test whether the ADA portion shows enzyme activity, cytosolic and crude membrane fractions were prepared by differential centrifugation of cell extracts. The parental cell lines as well as transduced cell populations were analyzed. As summarized in Table 2, radiometric ADA assays (28) of cytosolic fractions revealed similar levels of ADA activity in all NIH 3T3 derivatives, and somewhat higher levels in KNIH derivatives, due to the presence of endogenous ADA. Substantially increased levels of ADA activity, however, were detected in crude membrane fractions isolated from cells that had been infected with MDR-ADA retrovirus, when compared with crude membrane preparations from noninfected or MDR1 retrovirus-infected cells.

-

In vivo expression Figure 2. Characterization of MDR-ADA retrovirus producer cells. Parental (#1: PA-l2; #6: Psi-2) and various pHaMDRADA/Atransfected PA-l2 (#2- #4) or Psi-2 (#7, #8) clones, as well as the producer cell line PA-12-MDR1/A-l (# 5), were analyzed for the presence of provirus (A) and the levels of MDR-ADA fusion protein (B). Lanes refer to numbers of cell lines in Table 1. A) Southern blot analysis of EcoRI-digested genomic DNA with the MDRI-specific hybridization probe 5A (25). Genomic DNA (5 tg) was loaded in each lane, except for lane 5, which received only 1.5 gig. Markers at the left indicate sizes (kb) of HindIll fragments of phage X DNA. B) [3HlAzidopine photoafilnity labeling of expressed MDR-ADA fusion protein and multidrug transporter. Extracts from 1 x 106 cells were applied per lane. Only the high molecular weight range of the fluorogram (> 60 kDa) is shown, because photoaffinity labeling of smaller proteins was similar in all lanes. In the left- and rightmost lanes a molecular-size marker is shown, and the sizes of standard proteins are given in kDa.

and investigated for the presence of active multidrug transporter and MDR-ADA fusion protein, respectively. Cells were photoaffinity labeled with [3H]azidopine (31). Total cell extracts were immunoprecipitated in three different ways: by using 1) polyclonal rabbit antiserum 4007 directed against the human multidrug transporter (32), 2) polyclonal rabbit antiserum against bovine ADA (which cross-reacts with human ADA), or 3) preimmune serum. The immunoprecipitates were analyzed by SDS-PAGE (26) and subsequent autofluorography (27). Figure 3 shows the analysis of the NIH 3T3 cell populations. No [3H]azidopine-labeled protein was immunoprecipitated in the negative control (noninfected parental NIH 3T3 cells) with any of the antisera tested. In the MDR1 retrovirus-infected cells, a 170-kDa [3H]azidopinelabeled protein, corresponding to the size of the MDR1 gene product, was immunoprecipitated with multidrug transporter-specific antiserum, but did not react with ADA-specific antiserum or preimmune serum. Cells

1504

Vol. 4

March 1990

of the MDR-ADA

fusion

gene

To assess the expression and stability of the chimeric multidrug transporter-ADA protein in vivo, 2 x 106 colchicine-selected KNIH-MDR-ADA cells were injected subcutaneously into each of three athymic nude mice. KNIHMDR and parental KNIH cells served as controls. All mice developed tumors at the site of injection within a few days. Two weeks after injection of the cells, the mouse tumors (which had a diameter between 10 and 16 mm) were

analyzed

as follows.

Either

crude

membrane

fractions

were isolated by differential centrifugation of cell extracts and ADA activity was determined (28) (as summarized in Table 3), or tumor cell cultures were established, expanded, and grown in nonselective medium. After propagation for 2 wk, ADA activity was measured in cytosolic and crude membrane fractions, as summarized in Table 3. These data show that membrane-associated ADA activity levels both in KNIH-MDR-ADA-derived tumors or cul-

TABLE Lane

1. Retrovirus no.’

titrations

on NIH

Cell line

3T3

cells” Viral

titer,

(parental)

cfu/ml’

1

PA-l2

2

PA-12-MDR-ADA/A-4

6 x

3

PA-12-MDR-ADA/A-9

5 x l0

4

PA-12-MDR-ADA/A-l2

2 x i0

5

PA-12-MDR1/A-1

8 x

6

Psi-2

7

Psi-2-MDR-ADA/A-l4

0

8

Psi-2-MDR-ADAIA-16

1 x 102

0

(parental)

l0 0

“Viral 20-h harvests were serially diluted with growth taining 2 ig/ml Polybrene and titrated on NIH 3T3 cells described (10). 1’Lane numbers refer to Fig. I. ‘The were derived from the number of colchicine-resistant colonies 8 days after infection.

The FASEB Journal

i0

medium conas previously cfu/ml values formed within

GERMANN

ET AL.

A

B

123

123456

789 -200

200-

-4

-97

9768-

-68

-

43-.

‘-43

29-.

-29

Figure 3. Expression of human MDR-ADA fusion protein after retroviral infection of NIH 3T3 cells. The chimeric MDR-ADA protein was detected by 13H]azidopine photoaffinity labeling (A) and identified by immunoprecipitations (B). Whole cells were labeled with (3H]azidopine, and extracts from 1 x 106 cells were analyzed by SDS-7% polyacrylamide gel electrophoresis and fluorography. Lane 1) MDRI retrovirusinfected, colchicine-resistant (80 ng/ml) NIH 3T3 cells; lane 2) parental NIH 3T3 cells; lane 3) MDR-ADA retrovirus-infected, colchicineresistant (80 ng/ml) NIH 3T3 cells. B) Extracts from 1 x 106 [3H]azidopine-labeled NIH 3T3-MDR cells (lanes 1-3), parental NIH 3T3 cells (lanes 4-6), or NIH 3T3-MDR-ADA cells (7-9) were immunoprecipitated with either ADA-specific polyclonal antiserum 3652 (lanes 3, 6, 9), multidrug transporter-specific polyclonal antiserum 4007 (lanes 2, 5, 8), or preimmune serum (lanes 1, 4, 7). Immunoprecipitates were analyzed on a SDS-7% polyacrylamide gel. Arrows indicate the MDR-ADA fusion protein, arrowheads indicate the MDR1 gene product

P-glycoprotein.

tures

of these

tumor

cells

were

significantly

enhanced

DISCUSSION

when compared with the two controls. The presence of the MDR-ADA fusion protein was confirmed by immunoprecipitations of [3H]azidopine-labeled crude membrane fractions with both antiserum raised against the MDR1 or ADA gene product (data not shown). Furthermore, as presented in Fig. 4, similar amounts of MDRADA fusion protein were detected by [3H]azidopine labeling of KNIH-MDR-ADA cells and cell lines derived from KNIH-MDR-ADA-induced mouse tumors, indicating that neither in vivo passage nor the absence of a selecting agent for 4 wk caused any decrease of MDRADA fusion gene expression.

TABLE

2. Analysis

of ADA

activity

in retrovirus-infected

NIH

3T3

and

Retroviruses provide an excellent vehicle for gene transfer because of their ability to carry foreign genetic information into host cells easily and at high frequency. Expression of the delivered gene, however, appears to be variable, and is sometimes very low. To enforce concomitant expression of both a dominant selectable marker, the human MDR1 gene, and a nonselective gene, the human adenosine deaminase gene, we had previously constructed a fusion gene that encodes a bifunctionally active chimeric protein (12). To overcome the inefficiency of calcium phosphate-mediated DNA transfections, we now describe a TABLE 3. Analysis of ADA activity in KNIH-MDR-ADA-induced and reestablished in vitro cultures”

KNIH cells”

Specific Specific

Cell line

ADA

NIH

activity

Cytosol nmol inosine

3T3

(parental)

3.8

Tumor sample (TS) or tumor cell culture (TCC)

Membrane min’

nd.

16.5

nd.

61.0

TCC-KNIH

87.6

9.8

TCC-KNIH-MDR

86.3

7.7

TCC-KNIH-MDR-ADA

75.1

157.0

4.1

2.8 58.0

72.4

12.5

57.7

14.7 195.0

“NIH 3T3 and KNIH cells were infected with PA-12-MDR-ADA/ A-4 retrovirus or control PA-12-MDRI/A-1 retrovirus and expanded in the presence of 80 ng/ml of coichicine. Crude membrane and cytosolic fractions were prepared and assayed for ADA activity. The values are means of two determinations.

EXPRESSIONOF AN MDR-ADA FUSION GENE

mg’

TS-KNIH-MDR-ADA

9.4

64.4

min’

TS-KNIH-MDR

3T3-MDR-ADA

KNIH-MDR-ADA

nmol inoszne

Membrane

3.0

3T3-MDR

KNIH-MDR

Cytosol

TS-KNIH

mg’

NIH

(parental)

ADA activity

nd.6

NIH

KNIH

tumors

19.0

“Parental and MDR1 or MDR-ADA retrovirus-infected, colchicineselected KNIH cells were injected into nude mice. Tumors developed within 2 wk were used for preparation of crude membranes and ADA assays. Alternatively, tumor cells were grown in vitro in the absence of colchicine for 2 wk before measurements of ADA activity in cytosolic or crude membrane fractions. The values are means for three independent tumors. not determined.

1505

1

2

3

4

5

6

7

8

IU’

MDR-ADA:

9

10

11 12

200

-68

Figure 4. Expression

of MDR-ADA fusion protein in cultured mouse tumor cells after in vivo passage. Retrovirus-infected, colchicineselected KNIH cells were injected into athymic nude mice to produce tumors. These were dissected 2 wk later and used to reestablish cell cultures. Tumor cells grown in vitro for 2 wk in the absence of colchicine were labeled with [3H]azidopine. Cell extracts from 1 x cells were analyzed by SDS-7% PAGE and autofluorography. Lane 2) Parental KNIH cells; lane 2) MDR1 retrovirus-infected; and lane 3) MDR-ADA retrovirus-infected, colchicine-selected KNIH cells. Lanes 4-12 represent tumor cell lines. Tumors for lanes 4-6 were derived from parental KNIH cells, tumors for lanes 7-9 from KNIHMDR cells, and tumors for lanes 10-12 from KNIH-MDR-ADA cells. MDR1 and MDR-ADA gene products are indicated by arrows. On the right, a molecular-size standard is applied; sizes of proteins are given in kilodaltons.

replication-defective retrovirus carrying the MDR-ADA fusion gene. This recombinant Harvey murine sarcoma virus derivative transduces various types of cells at high efficiency and confers both the multidrug resistance phenotype and plasma membrane-associated ADA activity. The constructed retroviral vector was based on one we had previously used to produce recombinant retrovirus to efficiently transfer and express the human MDR1 gene (10). The vector contains pBR322 sequences necessary for propagation in Escherichia coli and several kilobases of Harvey murine sarcoma virus-specific DNA sequences, including the long-terminal repeats necessary for integration, as well as for initiation and polyadenylation of viral transcripts. In agreement with data reported by others (e.g., ref 29), we found that the presence of an additional polyadenylation site within the eukaryotic insert between the long-terminal repeats leads to prematurely polyadenylated transcripts and drastically affects retroviral titers. Replication-defective retrovirus was produced by using the packaging cell lines Psi-2 (23) and PA-12 (24) that release virions with an ecotropic or amphotropic host range, respectively. In different isolates of either transfected packaging cell line containing a similar copy number of intact integrated provirus, both the viral titers and levels of expressed MDR-ADA fusion protein were highly variable. Our data suggest no clear correlation between the amount of MDR-ADA retrovirus and fusion protein production. It is believed that retroviral DNA sequences integrate randomly into the host cell genome. The capacity of a packaging cell to produce translatable mRNA to yield recombinant MDR-ADA fusion protein and/or to express viral genomic RNA for encapsidation into a virion might therefore depend on the insertion site of the introduced retroviral sequences. Similarly, depending on the chromosome site of retrovirus integration, the problem of inadequate levels of expression and instability of transferred genes after many rounds of cell division may arise. In our system, however, once the MDR-ADA

1506

Vol. 4

March

1990

fusion gene is delivered to a recipient cell, sure can be applied to ensure proper and sion for a long period of time. The use of centrations of substrates for the multidrug (e.g., colchicine, vinblastine, Adriamycin)

selective presstable exprescytotoxic contransporter allows enrich-

ment of a cell population for cells infected with the MDRADA retrovirus at a multiplicity of infection of < 0.1. Simultaneously, the selection guarantees physiological levels of ADA activity from the bifunctionally active MDR-ADA fusion protein. The ability to select may therefore compensate for the relatively modest viral titers obtained from the various producer isolates. Efforts to increase the viral titers by a series of cocultivation experiments involving mixtures of Psi-2 cells and MDRADA virus-producing PA-12 cells (33) have so far failed. Both our in vitro and in vivo studies indicate that after retroviral transfer and colchicine selection, the MDRADA fusion protein is stably expressed in large amounts and exhibits bifunctional biological activity. Even in vivo passage and growth under nonselective conditions for several weeks do not seem to affect chimeric MDR-ADA protein levels. Therefore we consider the amphotropic MDR-ADA retrovirus a useful candidate to transfer the MDR-ADA fusion gene into a great variety of cultured cells for biological experiments in vitro and in vivo. We have recently generated transgenic mice in which expression of the human MDRI cDNA in bone marrow confers resistance in vivo to leukopenia induced by daunomycm (34). This indicates that expression of the MDR1 cDNA (or MDR-ADA fusion gene) in bone marrow should give a selective advantage to such marrow in animals challenged with chemotherapeutic agents. Preliminary data indicate that functional expression of the MDR1 cDNA occurs in mouse bone marrow with high efficiency after infection with an MDR1 retroviral vector (J. McLachlin, W. F. Anderson, I. Pastan, and M. M. Gottesman, unpublished results). It is also conceivable that in retroviral vectors the human MDR1 gene may serve as a selectable marker for the expression of many other novel fusion genes. The human MDRI gene product is a plasma membrane protein that contains 12 predicted transmembrane domains (3). Since both of its termini are located in the cytosol, any cytosolic protein or plasma membrane-associated polypeptide with free, cytosol-oriented ends should represent a potential fusion partner. So far, we have investigated only the MDR-ADA fusion system, in which the amino terminus of the unselected constituent (ADA) is connected with the carboxyl terminus of the selectable human multidrug transporter. It remains to be elucidated whether constructs in which the carboxyl terminus of an unselected constituent is attached at the amino terminus of the MDR1 gene product will also preserve bifunctional activity. We thank Lovelace graphic Swiss

Dr. M. Blaese

for providing

ADA

antiserum

3652,

for help with the tissue culture, and Steven Neal assistance. U.A.G. is the recipient of a fellowship National

Science

Betty

for photofrom the

Foundation.

REFERENCES 1. Gottesman, M. M., and Pastan, porter, a double-edged sword.

The FASEB Journal

I. (1988)

The

multidrug

trans-

J. Biol. Chein. 263, 12163-12166

GERMANN

ET AL.

2. Bradley, G., Juranka, P. F., and Ling, V. (1988) Mechanism of multidrug resistance. Biochim. Biophys. Acta 948, 87-128 3. Chen, C. -j., Chin, J. E., Ueda, K., Clark, D. P., Pastan, I., Gottesman, M. M., and Roninson, I. B. (1986) Internal duplication and homology with bacterial transport proteins in the mdrl (P-glycoprotein) gene from multidrug-resistant human cells. Cell 47, 381-389 4. Gros, P., Croop, J., and Housman, D. (1986) Mammalian multidrug-resistance gene: complete eDNA sequence indicates strong homology to bacterial transport proteins. Cell 47, 371-380 5. Gerlach, J. H., Endicott, J. A., Juranka, P. F., Henderson, G., Sarangi, F., Deuchars, K. L., and Ling, V. (1986) Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance. Nature (London)

324, 485-489 6. Gros, P., Ben Neriah, Y., Croop, J. M., and Housman, D. E. (1986) Isolation and expression of a complimentary DNA that confers multidrug resistance. Nature (London) 323, 728-731 7. Deuchars, K. L., Du, R. -P., Naik, M., Evernden-Porelle, D., Kartner, N., Van der Bliek, A. M., and Ling, V. (1987) Expression of hamster P-glycoprotein and multidrug-resistance in DNAmediated transformants of mouse LTA cells. Mo!. Cell. Biol. 7, 718-724 8. Ueda, K., Cardarelli, C., Gottesman, M. M., and Pastan, I. (1987) Expression of a full-length eDNA for the human “MDR1” gene confers resistance to colchicine, doxorubicin, and vinblastine. Proc. Nat!. Acad. Sci. USA 84, 3007-3008 9. Guild, B. C., Mulligan, R. C., Gros, P., and Housman, D. E. (1988) Retroviral transfer of a murine eDNA for multidrug resistance confers pleiotropic drug resistance to cells without prior drug selection. Proc. Nat!. Acad. Sci. USA 85, 1595-1599 10. Pastan, I., Gottesman, M. M., Ueda, K., Lovelace, E., Rutherford, A. V., and Willingham, M. C. (1988) A retrovirus carrying an MDR1 eDNA confers multidrug resistance and polarized expressionof P-glycoprotein in MDCK cells. Proc. Nat!. Acad. Sci. USA 85, 4486-4490 11. Kane, S. E., Troen, B. R., Gal, S., Ueda, K., Pastan, I., and Gottesman, M. M. (1988) Use of a cloned multidrug resistance gene for coamplification and overexpression of major excreted protein, a transformation-regulated secreted acid protease. Mo!. Cell. Biol. 8, 3316-3321 12. Germann, U. A., Gottesman, M. M., and Pastan, I. (1989) Expression of a multidrug resistance-adenosine deaminase fusion gene. j Biol. C/tern. 264, 7418-7424 13. Giblett, E. R., Anderson, J. E., Cohen, F, Pollara, B., and Meuwissen, H. J. (1972) Adenosine-deaminase deficiency in two patients with severely impaired cellular immunity. Lancet ii, 1067-1069 14. Kredich, N. M., and Hershfield, M. S. (1982) Immunodeficiency diseases caused by adenosine deaminase deficiency and purine nucleoside phosphorylase deficiency. In The Metabolic Basis of Inherited Disease (Stanbury, J. B., and Wyngaarden, J. B., Fredrickson, D. S., Goldstein,J. L., and Brown, M. S., eds) 5th ed., pp. 1157-1183, McGraw-Hill, New York 15. Hirschhorn, R. (1986) Inherited enzyme deficiencies and immunodeficiency: adenosine deaminase (ADA) and purine nucleoside phosphorylase (PNP) deficiencies. C!in. Immunopathol. 40, 157-165 16. Gilboa, E., Eglitis, M. A., Kantoff, P. W., and Anderson, W. F. (1986) Transfer and expression of cloned genes using retroviral vectors. BioTechniques 4, 504-512 17. Cepko, C., Roberts, B. E., and Mulligan, R. C. (1984) ConstructiOn and applications of a highly transmissible murine retrovirus shuttle vector. Cell 37, 1053-1062 18. Guild, B. C., Finer, M. H., Housman, D. E., and Mulligan, R. C. (1988) Development of retrovil-us vectors useful for expressing genes in cultured murine embryonal cells and hematopoietic cells in vivo. j Virol. 62, 3795-3801

EXPRESSION OF AN MDR-ADA

FUSION

GENE

19. Kantoff, P. W., Gillio, A. P., McLachlin, J. R., Bordignon, C., Eglitis, M. A., Kernan, N. A., Moen, R. C., Kohn, D. B., Yu, S. F, Karson, E., Karlsson, S., Zwiebel,J. A., Gilboa, E., Blaese, R. M., Nienhuis, A., O’Reilly, R. J., and Anderson, W. F (1987) Expression of human adenosine deaminase in nonhuman primates after retrovirus-mediated gene transfer.]. Exp. Med. 166, 219-234 20. Williams, D. A., Orkin, S. H., and Mulligan, R. C. (1986) Retrovirus-mediated transfer of human adenosine deaminase gene sequences into cells in culture and into murine hematopoietic cells in vivo. Proc. Nat!. Acad. Sci. USA 83, 2566-2570 21. Mclvor, R. S., Johnson, M. J., Miller, A. D., Pitts, S., Williams, S. R., Valerio, D., Martin, D. W., and Verma, I. M. (1987) Human purine nucleoside phosphorylase and adenosine deaminase: gene transfer into cultured cells by using recombinant amphotropic retroviruses. Mo!. CelL BioL 7, 838-846 22. Magli, M.-C., Dick,J. E., Huszar, D., Bernstein, A., and Phillips, R. A. (1987) Modulation of gene expression in multiple hematopoietic cell lineages following retroviral gene transfer. Proc.

NatI. Acad. Sci. USA 84, 789-793 23. Miller, A. D., Law, M. -F., and Verma, I. M. (1985) Generation of helper-free amphotropic retrovirus that transduce a dominantacting, methothrexate-resistant dihydrofolate reductase. MoL Cell. Biol. 5, 431-437 24. Mann, R., Mulligan, R. C., and Baltimore, D. (1985) Construction of a retrovirus packaging mutant and its use to produce helper-free defective retrovirus. Cell 33, 153-159 25. Goldstein, L. J., Galski, H., Fojo, A., Willingham, M., Lai, S.-L., Gazdar, A., Pirker, R., Green, A., Crist, W., Brodeur, G. M., Lieber, M., Cossman, J., Gottesman, M. M., and Pastan, I. (1989) Expression of a multidrug resistance gene in human cancers. j Nat!. Cancer Inst. 81, 116-124 26. Laemmli, U. K. (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature (London) 227, 680-685 27. Bonner, W. M., and Laskey, R. A. (1974) A fIlm detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. j Biochern. 46, 83-88 28. Yeung, C. -Y., Riser, M. E., Kellems, R. E., and Siciliano, M. J. (1983) Increased expression of one of two adenosine deaminase alleles in a human choriocarcinoma cell line following selection with adenine nucleotides. j Bio!. Chern. 258, 8330-8337 29. Shimotohno, K., and Temin, H. M. (1981) Formation of infectious progeny virus after insertion of herpes simplex thymidine kinase gene into DNA of avian retrovirus. Ce!! 26, 67-77 30. Southern, E. M. (1975) Detection of specihc sequences among DNA fragments separated by gel electrophoresis. J. Mo!. Biol. 98, 503-517 31. Safa, A. R., Glover, C. J., Sewell, J. L., Meyers, M. B., Biedler, J. L., and Felsted, R. L. (1987) Identification of the multidrug resistance-related membrane glycoprotein as an acceptor for calcium channel blockers.]. Bio!. C/tern. 262, 7884-7888 32. Tanaka, S., Currier, S. J., Bruggemann, E. P., Ueda, K., Germann, U. A., Pastan, I., and Gottesman, M. M. (1990) Use of recombinant P-glycoprotein fragments to produce antibodies to the multidrug transporter. Biochern. Biophys. Res. Commun. In press 33. Bestwick, R. K. Kozak, S. L., and Kabat, D. (1988) Overcoming interference to retroviral superinfection results in amplified expression and transmission of cloned genes. Proc. Nat!. Acad. Sci. USA 85, 5404-5408 34. Galski, H., Sullivan, M., Willingham, M. C., Chin, K.-V., Gottesman, M. M., Pastan, I., and Merlino, G. T. (1989) Expression of a human multidrug resistance eDNA (MDR1) in the bone marrow of transgenic mice: resistance to daunomycininduced leukopenia. Mo!. Cell. Bio!. 9, 4357-4363 Received for publication October 18, 1989, Accepted for publication December 4, 1989.

1507