REVIEW

Use of suicide
Pierre
France

genes
and Immuno-Motecutar

in gene
Therapeutics

therapy
Laboratory, Centre R#{233}gionate de Transfusion Sanguine, Besan#{231}on,

Tiberghien

Histocompatibitity

Abstract:

Gene

therapy

encompasses

a broad

range

of

treatment modalities that may eventually be applied to a variety of genetic as well as acquired diseases. In addition to compensating for a defective gene - and enhancing a cellular function - gene transfer can allow for a conditional negative selection of target cells. Indeed, the transfer of a gene encoding a susceptibility factor can make a cell specifically sensitive to a drug. After genomic integration, such a potentially destructive gene is endogenously expressed and has therefore been coined a suicide gene. This review describes current experimental approaches and prospects for using suicide genes for the treatment of human diseases. J. Leukoc. Biol. 56: 203-209; 1994. Key
. bone

Conferring drug resistance to a target of another form of gene therapy/transfer. efforts in this field are focused predominantly troduce the multi-drug-resistance-i hematopoietic progenitor cells in cancer
promote various in vivo selection of hematopoietic

is the objective Present clinical on trying to ingene into early patients in order to
cells resistant to

cell

chemotherapeutic agents [12]. The last form of gene transfer/therapy is the most distant from the initial goal of replacing a defective gene and involves, in contrast, the transfer ofa gene encoding a susceptifactor, therefore making a cell specifically sensitive to an exogenous pharmacological agent. After genomic integration, such a potentially destructive gene can be endogenously expressed, resulting in host cell death, and has therefore been called a suicide gene. After describing the mechanisms of action of various suicide genes presently used, this report reviews the current experimental approaches as well as future prospects involving the use of suicide genes for the treatment of human diseases.
bility

Words: marrow

gancictovir transplantation

. thymidine

kinase

. cancer

AIDS

INTRODUCTION Much therapy quently, disease realm. progress has been made in gene transfer and gene in various experimental conditions [1-3]. Consegene therapy as a treatment modality for human has now moved from the theoretical to the practical Over 35 protocols have been approved by the Recom-

Mechanisms

of action

binant DNA Advisory Committee of the National Institutes of Health and more than 80 patients have received gene transfer/therapy with no adverse reactions [4]. Gene therapy encompasses a broad range of therapies that may eventually be applied to a variety of genetic as well as acquired diseases. Historically, gene therapy was mainly, if not exclusively, perceived as a replacement therapy in which a defective gene would be replaced (or compensated for) by a functional gene. Indeed, the first federally approved gene therapy protocol (September 1990) involved the transfer of the adenosine deaminase (ADA) gene for the correction of a genetic disease: ADA deficiency [5]. However, the first clinical trial involving the transfer of a gene (May 1989) was not a gene therapy trial at all, but involved gene after infusion marking to track tumor-infiltrating into patients with melanoma [6]. lymphocytes Several such

Most firmly established among the suicide genes is the enzyme thymidine kinase, from the herpes simplex virus (HStk) [13, 14]. This enzyme alone is not harmful to cells, and in tk-negative cells it can allow cell survival 115]. However, in contrast to mammalian tk, HS-tk is capable of phosphorylating specific nucleoside analogues, such as acyclovir (ACV) or ganciclovir (GCV), to nucleoside monophosphate (MP) [16-18] (Table 1). The nucleoside MP is then phosphorylated by cellular kinases to nucleoside triphosphate (TP) and incorporated into DNA, leading to inhibition of DNA synthesis and cell death 119]. The high selectivity of the viral thymidine kinase for these nucleoside analogues and the very low affinity ofmammalian thymidine kinases for the same nucleosides make it possible to selectively kill HStk-expressing cells among a population of dividing mammalian cells [20-22]. Enzymatic studies have demonstrated that GCV exhibits higher affinity for HS-tk than does ACV [13, 23, 24]. At present, most of experimental approaches and all the clinical trials tk gene (Table 2). involving suicide genes use the HS-

gene marking studies are now in progress in the setting of bone marrow transplantation as well [7]. Another form of gene transfer/therapy experiments presently under way involves the transfer of genes to stimulate immunity against, or otherwise cause the destruction of, tumor cells. The first approach of this type involved the transfer of a tumor necrosis factor (TNF)-a gene to tumorinfiltrating lymphocytes, in the hope that localized secretion of TNF-a would enhance immune destruction of tumor cells [8]. Several additional teams have initiated clinical studies involving the transfer ofother cytokine or histocompatibility genes in tumor cells with the goal of stimulating a tumorspecific immune response that would result in systemic tumor destruction [9-11].

Abbreviations: isiethoxypu cvtosine

ganciclovir:

AC\, one arabinucleoside; 5-FC,

acvclovir; BMI,

ADA, bone
(lisease: virus;

adenosi tisarrow 5-FU.

Cvi.,

ne

deaminase;

AraM,
CD, CCV, leukemia; from herpes

transplantation; 5-fluorouracil;
graft versus

(leaminase;
Cv human

5-fluorocytosine;

H I),
I .3R factor

graft-versus-host

Hl\.
siiiiplex

immunodeficiencv

HS-ik.
repeat ;

thymidine

kinase
varicella-zosier

vi rus;
OCCOSiS

long a

terminal 6TX,

M P.

nl000ph(aphate;

IN F-ct,
virus;

tumor X(; PRL Reprint

6-thioxanthine;

VZV, Histocompatibility

xanthine-guanine-phosphoribosyltransfcrase. requests: Pierre Tiberghien,

and

Immuno-

Molecular
guine, Received

1herapeutics
Besancon, August 25000, 8,

Laboratory,
France. 199:3; accepted

Centre
April

Regional
4, 1994.

de Transfusion

San-

Journal

of Leukocyte

Biology

Volume

56,

August

1994

203

TABLE
Suicide
gene

Mechanisms

of

Action

of

Suicide

Genes

acid 6-TX
effects

and
[37].

xanthine

and

negatively

selected

by

exposure

to

Toxic Substrate metabolite Cellular

Cancer
HS-tk VZV-tk CD GCV-GCV-MP AraM-AraM-MP 5-FCyt--5-FUra GCV-TP AraATP 5-FUra 6TX-nMP
Inhibition of DNA synthesis

Inhibition Inhibition
Inhibition

of DNA of DNA
of DNA

synthesis synthesis
synthesis

In June

1992,

the

NIH

Recombinant

DNA

Advisory

Comstrategy otherwise with to veccells

XGPRT

6TX-6TX-nMP

mittee approved for selectively

a protocol killing tumor brain murine

involving an ingenious cells in patients with tumors. retroviruses Gene transfer is restricted

incurable malignant tors derived from In vitro studies have also been performed with the varicella-zoster virus (VZV) tk gene. 6-Methoxypurine arabinucleoside (AraM) is selectively toxic to VZV-infected cells [25]. This selectivity results from the fact that AraM is a good substrate for VZV-tk but a poor substrate for mammalian nucleotide kinase. Once monophosphorylated in a VZV-infected cell, AraM-MP can be further metabolized to produce the cytotoxic anabolite AraATP. This approach was developed in a hepatocarcinoma model in which the expression of the VZV-tk gene was associated with selective sensitivity to AraM [26]. Another gene capable of conferring a phenotype of sensitivity to a drug is the gene coding for the enzyme cytosine deaminase (CD) that catalyzes the deamination of cytosine to uracil [27]. It derives its therapeutic specificity from the presence of this enzyme in some fungi and bacteria but not in mammalian cells. Microorganisms that express the CD gene convert 5-fluorocytosine (5-FC) to 5-fluoruracil (5-FU), a highly toxic metabolite that is lethal to the cell [28, 29]. Since mammalian cells do not express significant amounts of CD, they do not deaminate 5-FC [30]. 5-FC is an antifungal agent used in vivo and is nontoxic to cells at concentrations that result in strong antimicrobiol activity [31]. However, 5-FU has potent cytotoxic effects on mammalian cells and is a potent cancer chemotherapeutic agent. It is metabolized to 5-fluorouridine 5-TP and 5-fluoro-2-deoxyuridine 5-MP, resulting in RNA and DNA synthesis inhibition and cell death [32]. After engineering the bacterial CD gene into a eukaryocytic expression vector, investigators have demonstrated that the transfer and expression of the bacterial CD gene to mammalian cells does indeed confer lethal sensitivity to 5-FC [33]. A possible advantage of using the CD versus the HS-tk gene is that 5-FCyt is not often used clinically and that many alternative antifungal agents are available for treatment offungal infections [34]. In contrast, GCV is cornmonly used for systemic cytomegalovirus infection and alternative treatment modalities are scarce [35]. Lastly, xanthine-guanine phosphoribosyltransferase (XGPRT), an enzyme found in Escherichia coli but not in mammalian cells, could be used as a suicide gene. XGPRT catalyzes the conversion of xanthine to its ribonucleotide form (xantine - xanthine ribonucleotide monophosphate guanosine monophosphate), which can contribute to purine synthesis via the salvage pathway. An analogue, 6-thioxanthine (6-TX), can also XGPRT, resulting in the formation serve as a of6-thioxanthine substrate nucleofor
I)escrmption

that are proliferating and synthesizing DNA [38]. This suggested to Culver et al. [39] that retrovirus-mediated suicide gene transfer might permit targeting of gene integration into malignant cells in organs composed mainly ofquiescent nonproliferating cells, such as the brain. To enhance transduction efficiency, they developed on approach in which retroviral particle-producing mouse fibroblasts were directly injected in situ and served as virus factories by continuously producing retroviral particles into the immediate local environment. Thus, when an individual tumor cell enters DNA synthesis, a vector particle will be available to transduce it [39]. In addition, the investigators reasoned that because the brain is a partially immunologically privileged site, longer could quency Rats survival of the xenogeneic be expected, resulting of the growing with a cerebral tumor. gliorna in murine greater were given cells in the transduction an intratumoral brain fre-

stereotaxic injection of murine fibroblasts producing a retroviral vector in which the HS-tk gene had been inserted [39]. After 5 days, during which the murine fibroblastproduced HS-tk retroviral vector transduced the neighboring proliferating glioma cells, the rats were treated with GCV. Control rats bearing tumors mixed with control fibroblasts or fibroblasts producing only a NeoR vector exhibited no tumor response. Tumors in mice injected with mixtures of tumor and nonproducing HS-tk fibroblasts showed a modest slowing of growth, suggestive of a mild antitumor effect. However, gliomas in the animals having received both the fibroblast-producing HS-tk vectors and GCV regressed completely, both macroscopically and microscopically. This impressive result was obtained despite the fact that the efficiency of in vivo gene transfer to tumor was less than 100% [39, 40]. These results suggested that additional mechanisms, other than GCV inhibition of HS-tk-transduced cells were responsible for the antitumor effect. To explore this observation further, Culver et al. [39] injected normal mice subcutaneously with wild-type tumor cells mixed at various ratios with tumor cells expressing the HS-tk gene. Treatment with GCV produced complete tumor regression in nearly all animals containing 50:50 tumor mixtures and in some animals bearing tumor mixtures consisting of as few

1ABLE

2.

Human

Gene

Therapy

Protocols

Involving

Suici

de

Genes

Principal

investigators

Disease

side monophosphate (6-TXnMP) and subsequent inhibition of nucleid acid synthesis. 6-TX is a metabolite of the cornmonly used antineoplastic agent 6-thioguanine. Investigators have established that cells from a mouse teratocarcinoma transfected ceptible with to 6-TX XGPRT toxicity DNA [36]. can indeed Of interest be rendered susis the observation

In situ injection of tumor HS-tk retrovirus-producing

fibroblasts
+

Oldfiel

et al.

[421

Brain

GCV

treatment
Freeman [52] Ovarian

Intraperitoneal infusion of HS-tktransduced ovarian cancer cells + GCV treatment

Intravenous transduced cells infusion HIV-specilIc of HS-tkcytotoxic T

that XGPRT can provide a dual tion. Indeed, investigators have expressing tumor cells could be the basis of resistance to a regimen

sensitivity-resistance funcestablished that XGPRTboth positively selected on containing mycophenolic

Riddell

et

al.

[54]

AIDS

GCV

treatment

204

Journal

of Leukocyte

Biology

Volume

56,

August

1994

as 10% HS-tk-expressing cells [39]. This phenomenon was reproduced in vitro with a variety oftumor cell lines [40, 41]. Thymidine incorporation assays also revealed that the supernatants collected after GCV exposure could significantly inhibit thymidine incorporation in all the tumor types tested [42]. Of interest, such a bystander effect was not found with HS-tk-tranduced primary human T lymphocytes or with an HS-tk-transduced CD4 lymphoma cell line (HUT-78) [43], suggesting that this bystander effect is dependent on the cell type and might need specific intercellular channels such as gap junctions [44]. A recent report has suggested that the bystander effect could be related to a process of apoptotic cell death when HS-tk-expressing transfer of toxic compounds through apoptotic vesicles in vivo and in vitro bystander cells are exposed to GCV to HS-TK-nonexpressing Other authors confirmed effect on glioma cells but with cells the were

higher than the relevant human GCV dose. Based on these various findings, a clinical protocol involving direct in situ injection in otherwise incurable brain tumors of HS-tk vector producing murine fibroblasts followed by systemic GCV treatment [42] was initiated in December 1992. As ofDecember 1993, 10 patients had received this novel therapy dure well [51]. Another clinical GCV/HS-tk-mediated and involves cancer line i.p. followed and had apparently at effect tolerated the proceon a in 1992

trial relying, bystander

least partly, was approved

infusion of an HS-tk-expressing ovarian by GCV administration in patients with

unable to demonstrate a toxic effect mediated by substances present in the culture supernatant of GCV-treated HStk-transduced tumor cells [46]. In this last case, antitumor efficacy was enhanced if the tumor cells were coinfected with wild-type virus [46, 1#{149}he T mechanism of this bystander effect is poorly understood at present and could involve the release of enzymes or other substances such as GCV-triphosphate. Although vivo studies showed no evidence of transduction ofthe normal neural tissue surrounding the tumor, there was evidence for occasional transduced blood vessel endothelial cells within or adjacent to the tumor [40], and GCV-induced destruction of the tumor vasculature could also contribute to the in vivo (but not the in vitro) antitumor efficacy. However, this hypothesis is not entirely compatible with the finding that, in vivo, HS-tk-expressing arterial walls can be insensitive to GCV [48] and that overall, nondividing HS-tk-expressing cells will be resistant to GCV treatment. Interestingly, in the brain tumor study, the occasional transduced endothelial to the tumor, cells were suggesting be found that within or immediately angiogenesis factors adjacent released by

residual ovarian cancer cells after surgery or chemotherapy [50, 52]. This last approach is supported by experimental data suggesting that i.p. administration of HS-tk tumor cells followed by GCV treatment prolonged the survival of i.p. tumor-bearing mice [50].. As of December 1993, one patient has been entered on this protocol with no side effects. The experimental approach pioneered by Culver et al., if successful in treating brain tumors in humans, will certainly be developed and evaluated in other tumors associated with a poor prognosis and a low metastatic potential such as hepatocarcinomas, esophageal cancer, and various other cancers of the gastrointestinal tract. As mentioned previously, in vivo experiments have indeed established that macroscopic liver metastases could significantly regress after in situ HS-tk transduction and GCV treatment [49]. Although liver cells are quiescent in a normal adult liver, this might not always be the case in a pathological liver, especially after liver surgery. One can therefore not exclude that normal liver cells could be transduced with a tk gene and be sensitive to GCV treatment. One way to address this issue could be to exploit the transcriptional differences between normal and neoplastic cells to achieve selective killing of neoplastic cells as reported by Huber et al. [26]. In the present case, the transduction of a VZV-tk gene transcriptionally regulated afetoprotein, resulting in selective in vitro AraM-induced killing of hepatoma cells. More recent data have also demonstrated that HS-tk expression could be restricted to melanoma cells and melanocytes by using a HS-tk gene under the transcriptional control of the 5-flanking sequence of the tyrosinase gene [53]. Additional in vitro and in vivo studies are needed to evaluate further the potential of such approaches.

the tumor could HS-tk transduction

indirectly responsible for both a selective and subsequent GCV sensitivity of the

tumor-associated dividing endothelial cells. Another factor that may contribute to the observed in vivo antitumor effect is the generation of a host-derived antitumor immune response. Indeed, in an experimental approach similar to the brain tumor model and involving systemic GCV treatment after in situ HS-tk transduction of established macroscopic liver metastases, Caruso et al. [49] observed an active local immune response evidenced by massive infiltration of the tumors by macrophages and both CD4
hypothesis

Acquired
To reconstitute (HIV)-specific undergoing Greenberg in vivo

immunodeficiency
or augment immunity allogeneic bone of

disease
human immunodeficiency for HIV-seropositive marrow transplantation [54] initiated host-derived a protocol CD8 virus individuals (BMT), involving gag-specific

and

antitumor

CD8 is the effect

lymphocytes. Also compatible with this finding, in an ovarian tumor model, that following intraperitoneal (i.p.) administraovarian was reduced in these tumor cell in nude studies

last the

tion of an HS-tk-expressing sequent GCV treatment mice [50]. The GCV dose used

line and subor irradiated was high (300

and co-workers administration

in vivo

cytotoxic T cell clones Because of the nature HIV, the investigators

after allogeneic BMT for lymphoma. of the host cells potentially infected by were worried that the recognition and

mg/kg/d) and is highly toxic in humans. In order to address this potential problem, studies with a relevant human dose of GCV (10 mg/kg/d) as well as two higher doses, 20 and 30 mg/kg, were evaluated in the brain tumor model [40]. In rats inoculated with HS-tk-transduced 9L tumor cells, all three doses induced similar regressions of tumors compared with the control group. Persistent viable tumor cells were, however, observed after 7 days of GCV treatment in each group. Unfortunately, a direct comparison of the antitumor efficacies of the previously used GCV dose (300 mg/kg) and these report, lower doses was not further evaluations were performed reported. Furthermore, of GCV/HS-tk-induced with 30 mg/kg GCV, in the same tumor

regressions

threefold

elimination of infected cells by the transferred CD8 cytotoxic cells could result in unique toxicity in these patients. To address this concern, the T cell clones are modified by retrovirus-mediated gene transfer to contain an HS-tk gene. The retrovirus used in this study encodes a gene that is a fusion product of the hygromycin phosphotransferase and HS-tk genes [55]. This hybrid selectable marker gene (HyTK gene) encodes a bifunctional protein that confers to transduced cells both resistance to hygromycin B and sensitivity to GCV. Thus, introduction of the HyTK gene in the CD8 clones could provide a mean of eliminating the clones in vivo if toxicity occurs. The clinical trial is under way and data on tolerance and efficacy are forthcoming.

Tiberghien

Use

of suicide

genes

in gene

therapy

205

Continuous and stable in vivo expression of the HS-tk gene is essential in this last approach as well as in several other approaches described later in this review. Loss of HStk expression has been observed in vitro in HS-tk-transfected cell lines. DNA mutation, deletion, or methylation can be responsible studies involving tk-positive tumor trast to sarcomas for decreased TK gene expression [56]. In in vivo GCV treatment of various Hscells, Moolten et al. [57] found that in con(K3T3), lymphomas (Lyl8) were quite

of a HIV-infected patient, the authors suggested that the presence of TK-harboring cells could reduce viral spread with the result that HIV infection might even abort. However, for such an approach to be successful, in vivo expression of the transgene in the target cells would have to be very close to 100%. Furthermore, the importance ofthe viral spread relative to the multiple HIV-associated immune dysregulations in determining the severity of AIDS is uncertain [62]. Although elegant, this approach obviously needs further in vitro and in vivo evaluation.

resistant to GCV treatment, with rapid regrowth after initial partial remission. Analysis of the tumors that regrew revealed loss of HSV-tk activity, suggesting to the authors that decreased gene stability might be responsible for these findings. Alternatively, this finding could result from a GCV/HS-tk-induced bystander effect present with the sarcoma cells but absent with the lymphoma cells [43]. In the last case, the absence ofa bystander effect could have allowed a post-GCV outgrowth of an initially small percentage of HS-tk-nonexpressing lymphoma cells. In vitro, HS-tk producer cells [40] and HS-tk-transduced HUT-78 (T cell lymphoma) cells [43] have remained sensitive to GCV for several months in culture. However, these in vitro findings may clearly change after in vivo administration and only careful analysis of the first patients receiving HS-tk-transduced lymphocytes followed by GCV treatment will reveal whether gene instability is indeed a critical and limiting issue. The contribution of suicide genes to eliminating the initial pool of HIV-infected cells in seropositive patients with the hope of preventing disease progression is also being evaluated by several investigators. Venkatesh et al. [58] imagined an attractive strategy involving the engineering and use of a replication-defective adenovirus vector (Ad-tk) whose action is dependent on the targeted expression of the HS-tk gene, cloned downstream of the HIV-1 long terminal repeat (LTR), in cells expressing the HIV-l transcriptional activator Tat. In the initial phase of viral replication, mRNAs encoding HIV regulatory proteins such as Tat are produced, whereas the mRNAs encoding HIV structural proteins are not significantly expressed until 24 h after viral infection Therefore, a toxic gene placed under the control of the HIV LTR would be significantly expressed only in HIVinfected cells. Because the gene would be switched on in the initial phase of HIV replication, before the initiation of HIV structural protein synthesis, ACV or GCV treatment could result in cell death before the release of newly synthesized viral particles and thus prevent virus spread [60]. Infection of Tat-expressing human HeLa or Jurkat cells with Ad-tk resulted in high-level HS-tk expression, which was not deleterious to the viability of these cells. However, in the presence of GCV, Ad-tk infection resulted in a massive and specific reduction in the viability of these Tat-expressing cell lines. By expressing a similar LTR-driven HS-tk gene in CD4 cells (HUT-78 cell line), Caruso et al. [60, 61] were able to examine the effect of ACV treatment on the spread of an HIV-infected cell population. ACV treatment successfully eliminated the HIV-infected cells and, most important, resulted in the arrest of HIV spreading in the culture. Complete protection of HS-tk-expressing HUT-78 (tk-HUT) cells from HIV infection was obtained using ACV concentrations achieved in the plasma of patients treated for HSV-1 infection. When HIV-infected tk-HUT cells, first grown with ACV, were cocultured with parental HUT-78 cells, no infectious HIV particles could be rescued. Furthermore, HUT-tk cells initially infected with HIV and grown in ACV could be cultured indefinitely without any toxic effects. In the context

Allogeneic

bone marrow

transplantation

Allogeneic BMT is associated with a severe complication, graft-versus-host disease (GvHD) [63]. While effectively preventing GvHD, ex vivo T lymphocyte marrow depletion unfortunately increases graft rejection and reduces the graftversus-leukemia (GvL) effect [64]. Furthermore, a GVL effect persists even in the absence of GVHD in patients receiving a non-T cell depleted marrow graft [65]. Unfortunately, GvHD occurrence probability and severity are difficult to assess prior to BMT. The ex vivo transfer of the HS-tk gene into T cells before their infusion with hematopoietic stem cells could allow selective in vivo depletion of these T cells with GCV if subsequent GvHD was to occur. Thus, one could preserve the beneficial effects of the T cells on engraftment and tumor control in patients not experiencing severe GvHD. In addition, the GCV-induced selective immunosuppression restricted to the donor mature T cells infused with hematopoietic stem cells could result in less toxicity than the broad immunosuppressive agents presently used for GvHD treatment. Studies in our laboratory have investigated the feasibility of such an approach [43]. Phytohemagglutininor alloantigen-stimulated purified primary T cells were successfully transduced with a retroviral vector containing the HS-tk and neomycine resistance (NeoR) genes. Subsequent culture in G418 for 1 week allowed the selection of transduced cells. Importantly, GCV treatment of IL-2-responding transduced and selected cells results in over 80% growth inhibition, whereas GCV treatment ofcontrol cells had no effect [43]. Similarly, the allogeneic reactivity of HS-tk-transduced cells was specifically inhibited by GCV. The inhibition of HS-tk-transduced cells was obtained with GCV concentrations (0.1 to 3 jg/ml) achievable in vivo during GCV treatment for cytomegalovirus infection [66]. Combining transduced and nontransduced T cells did not reveal a bystander effect, implying that all of Lastly, the cells studies inhibited involving by GCV transduction were indeed of HUT-78 transduced. cells sug-

I 59].

gested that stable expression of HS-tk can be maintained over 3 months in vitro in the absence of G418. These studies established the feasibility of generating HS-tk-transduced T cells for subsequent transfer with hematopoietic stem cells and, if necessary, specific in vivo GCV-induced T cell depletion in allogeneic BMT recipients. A phase I trial in allogeneic BMT will soon be initiated.

Preemptive
Inadequate

gene insertion
selectivity (i.e., nonspecific toxicity) severely

limits the antitumor efficacy of most cancer chemotherapeutic agents. In an effort to devise therapeutic strategies more tumor-specific than proliferation-specific, Moolten [44] suggested that the prophylactic insertion, in a mosaic fashion, of one or several genes altering cell sensitivity to chemotherapeutic could agents improve into hosts at high chemotherapeutic risk ofsubsequent selectivity by neoplasm enhancing

206

Journal

of Leukocyte

Biology

Volume

56,

August

1994

the prospect gene. GCV

that a subsequent tumor will express the suicide can eradicate transplanted murine tumors expressing the HS-tk gene [22, 57]. Because autochthonous tumors are obviously an experimental model more relevant to human applications, Moolten and co-workers chose to test their
tk

provements sion modulation, although the potential

tional by

in areas initial such

such and

as gene

delivery,

in situ

or expres-

stability clinical as helper

the induction

are needed. experience is virus genes

of

In addition, encouraging, inserdisease in mind

hazards
mutagenesis,

or

approach by using transgenic mice harboring the HSgene under the transcriptional control of an immunoglobulin heavy-chain enhancer and kappa light-chain promoter, a combination that restricts gene expression to mainly lymphoid cells [22]. They hypothesized that if lymphomas were generated in these mice, they might be susceptible to GCV treatment. The host, in contrast, might tolerate the
tk

the

introduction

of foreign

contamination, autoimmune need be kept

and addressed in future studies. Nevertheless, in a short period of time, the use of suicide genes as a form of therapy for human disease has jumped from theoretical speculations to practical reality.
Note added in proof

treatment and the

because normal
by HS-tk

nonlymphoid GCV-sensitive
negative

cells do lymphoid
precursors.

not express HScells could be

treatment of

of Epstein-Barr ing the HS-tk lymphoproliferative

Since submission virus (EBV)-specific gene in a patient with disease has been

of this review, the use T-lymphocytes expresspost-BMT EBV-related reported [69].

replenished

GCV

the lymphoma mice resulted

(induced in complete

by Abelson leukemia tumor regression

in

virus)-bearing 11 of 12 mice an obserthe lymmutant they had model

REFERENCES
1.

[ 67]. Five of these mice remained tumor-free over vation period that exceeded 100 days. Among phomas that recurred, most appeared to represent subpopulations that were GCV insensitive because lost HS-tk. Despite the limitations ofthe experimental ( tissue-specific expression, germline

Miller,

D.A.
455-460. W.F.

(1992)
(1992)

Human
Human

gene gene basic

therapy therapy. science

comes Science of gene

of

age.

Nature

357,
2. Anderson,

256,
therapy.

3. Mulligan,

R.C.

(1993)

The

808-813. Science Report to

260,
4.

926-932.
W.F., McGarrity, Recombinant DNA retrovirus

gene insertion), the authors felt that these results lent plausibility to the hypothesis that preemptive somatic insertion of suicide genes in a mosaic fashion might be of interest for patients with diseases associated with a very high risk of fatal neoplasms, such as myelodysplastic syndromes.

Anderson, the NIH

G.J.,

Moen,
(RCR)

R.C.
Committee assays.

(1993)
on

Advisory

murine

replication-competent

Hum.

Gene Ther. primary

A., Mor-

4, 311-321. 5. Blaese, R.M.,

immunodeficiency 6. Rosenberg, gan, R.A., Topalian, S.A.,

Culver,
disease.

K.W.

(1992) Gene therapy for Immunodefic. Rev. 3, 329-349.

P., Cornetta, EM., Culver, Lotze, K., K., Kasid, M.T, Miller,

Aebersold, Karson, M.J.,

Fail-safe

device
the in vivo use of ex vivo so after in vivo gene transfer cells either by activation of inappropriate insertion of enofantioncogenes [68]. Fur-

Moen, R., S.L., Merino,

Yang, J.C., D., Blaese,

M.,

Anderson,

W.F.
of N

(1990)

Gene
with

transfer
advanced by

into
melanoma, retroviral

humansusing gene

One concern when considering transduced cells or even more is the potential to transform proto-oncogenes through the hancer elements or by disruption

immunotherapy tumor-infiltrating transduction.

7. Brenner, M.K.,

patients

lymphocytes Engt. j Med.

Rill,

modified 323, 570-578.

DR.,
Ihle,J.N. autologous

Moen,

R.C.,

Krance,

R.A.,

Mirro,
on-

J.,

Anderson, gin of relapse

WE., after

(1993) Gene-marking bone-marrow Anderson, R., Delgado,

to trace transplantation. WE, C., Perez, Mule,

with of cancers

thermore, in addition to the neoplastic potential, transduced cells could potentially mediate a variety of other toxicities in relation to inappropriate expression or unexpected toxicities in relation to the expression of the transgene. Inclusion of a suicide gene as a component of the gene therapy vectors could create the potential for eradicating any transduced cells that subsequently exhibit neoplastic or other unwanted behavior. Therefore, the use of a suicide gene could not only contribute to reducing the side effects associated with the therapeutic use of cells such as allogeneic T lymphocytes in the setting of BMT but also provide a failsafe this option respect, for innovative gene the use of suicide interesting and therapy genes strategies [37, such as XGPRT both the 48]. In could

Lancet 8. Hwu, Chiang,

tion necrosis of

341, 85-86. P., Yannelli,

Y., Schwarz,

J.,
(1993)

Kniegler, M., S., Cowherd,

C.,

J.,

Rosenberg,

S.A.
tumor-infiltrating factor-alpha

Functional
lymphocytes eDNA for the

and
gene

molecular
transduced therapy

chanacterizatumor in

humans. 9. Rosenberg,

J.

Immunot. 150, S.A., Anderson, Hwu,

4104-4115. WE, P., D.J.,

Asher,

AL., Mule, S.L.,

Blaese,

MR.,

Ettinghausen, SE., DR., Schwartzentruber, nelli, cancer tion

57-73.

Kasid, A., Topalian,

J.J.,
Weber, modified

Parkinson, J.S., Yan-

JR., Yang, J.C., Linehan, patients using autologous of the S., gene for tumor N., (1993) cells. necrosis

W.M.
cancer

(1992)
cells factor. N.,

Immunization
by Gene inserHum. Ther. Melief,

of 3,

be particularly and chemosensitivity

in combining therefore reducing

selectability number

10.

Osanto,

Brouwenstyn,

Vaessen,

Figdor,

C.G.,

of

C.J.M.,
transfected
metastatic

genes included in the vector. Indeed, extra genes are frequently associated with substantial loss of expression efficiency. Clearly, the use of suicide genes to increase the safety of various ex vivo or in vivo gene therapy approaches is a promising and valuable approach.

Schnier, P.1. melanoma

melanoma.

Hum.
Immunotherapy

11.
12.

Nabel, transfer
Sorrentino,

G.

(1992) into
B.P.,

Immunization with interleukin-2 A phase I-Il study in patients with Gene Ther. 4, 323-330. ofmalignancy by in vivo gene

tumors. Cline, A.,

marrow

Hum.
Brandt,

Gene Ther. 3, 399-410. S.J., Bodine, D. , Gottesman,

A.W.
in vivo G.N.,

M.,
of drugof Dis-

Pastan,
resistant

I.,

Nienhuis,
cells

(1992)
after

Selection
retroviral

bone

transfer DR. (1974)

CONCLUSION
13.

human
Kit, tinctive S.,

MDRJ.
Leung,

Science
W.C.,

257,
Jorgensen,

99-103.
Dubbs,

I have tried to summarize the development and use

the various of suicide

approaches genes in gene

involving therapy.
14.

man
Cooper, in cells

properties and avian

ofthymidine herpes virus.

(1973) with Phosphorylation herpes

kinase ml. j
simplex

isozymes Cancer 14,

of virus.

induced

by

hu-

598-610. Acad. Sci.

Most certainly, many other suicide genes, transfer techniques, and target tissues are being considered. As the first patients are receiving suicide gene-expressing cells, with or without GCV, valuable in vivo clinical information will soon be available. Suicide gene therapy, not unlike other forms of gene therapy, is still at the stage of early infancy. Major im-

G.M.
infected

5-bromodeoxycytidine

Proc. Natt. A., Cheng,

thymidine

USA
15. Wigler,

70,
(1977)

3788-3792.
M., Silverstein, Transfer of

S. , Lee,
purified

L., Pellier,
herpes virus

Y., Axel,
kinase

R.
16. gene Elion,

to

cultured GB.,

mouse Furman,

cells. PA.,

Gett 11, 223-232. Fyk, J.A., Dc

Miranda,

P.,

Tzberghzen

Use

of suicide

genes

in

gene

therapy

207

Beauchamp,
antiherpetic

L., Schaeffer,
agent,

H.J.

(1977)

Selectivity

ofaction
guanine.

of an

35. 36.

9-(2-hydroxyethoxymethyl)

Proc.
GB.
37.

17.

Nail. Acad. Fyfe,J.A.,

Sci. USA 74, 5716-5720. Keller, P.M., Furman,

PA.,

Miller,

R.L.,

Elion,

(1978) Thymidine kinase from herpes simplex virus phosphorylates the new compound, 9-(2-hydroxyethoxymethyl) guanine. j RioL Chem. 253, 8721-8727. 18. Field, AK., Davies, ME., DeWitt, C., Penny, H.C., Liou, R., Germershausen, J. , Karkas, J.D. , Asthon, W.T. , Johnston, D.B.R., Tolman, R.L. (1983) 9-i[2-Hydroxy-1-(hydroxymethyl)
ethoxy]methylfi
virus replication.

Progress in the treatment of Intern. Med. 109, 769-771. Besnard, C., Monthioux, J. (1987) Selection against expression of the Escherichia coli gene gpi in hprt mouse teratocarcinoma and hybrid cells. Mot. Cell. RioL 7, 4139-4141. Mroz, P.J., Moolten, EL. (1993) Retrovirally transduced Escherichia coti gpi genes combine selectability with chemosensitivity capable of mediating tumor eradication. Hum. Gene Ther. 4,
cytomegalovirus pneumonia.

Franck,

I., Friedman,

H.M.

(1988) Ann. E., Jami,

589-595.
38.

guanine:

a selective

inhibitor

of

herpes

group 39.

Proc. Nail. Acad. Sci. USA 80, 4139-4143. 19. Furman, PA., McGujint, P.V., Keller, P.M., Fyfe, J.A., Elion, GB. (1980) Inhibition by acyclovir ofcell growth and DNA synthesis 20. of cells biochemically transformed with herpesvirus genetic information. Borelli, E., Heyman, ing of an inducible Virology 102, R., Hsi, M., toxic phenotype
7572-7576.

Miller, D.G., Adam, MA., Miller, AD. (1990) Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. MoL Cell. Riot. 10, 4238-4242.
Culver, K.W., Ram, Z., Wallbnidge, S., Ishii, H., with Oldfield, retroviral

420-430.
Evans, R.M. in animal Anderson, (1988) Targetcells. Proc. Nail. D., Richman,
40.

E.H., Blaese, R.M. (1992) In vivo gene vector-producer cells for treatment tumors. Science 256, 1550-1552.
Ram, Z., Culver, K.W., Walbnidge, S.,

transfer

of

experimental
R.M.,

brain
Oldfield,

Blaese,

Acad.
21. Heyman,

Sci. Baird,

USA
R.A.,

85,

E.H. (1993) In situ retroviral-mediated treatment of brain tumors in rats.

41. Freeman, S., Whartenby, K.A.,

gene
Cancer Koeplin, Res. D.S.,

transfer for 53, 83-88.

et al.

the

Borrelli,

E.,

Lesley,J.,

(1992)

D.D.,
kinase

SM.,

Hyman,
creation Proc. Tumor kinase strategy.

R.,
Nail.

Evans,
Acad. genes: Sci.

R.M.
mice USA

(1989)
with

Thymidine
controlled
2698-2702.

Tumor
HSV-TK 42. Oldfield, in-

obliteration:

of transgenic chemosensitivity

regression gene. j
E.H.,

when a fraction Cell RioL i6F,

Ram,
(1993)

ofthe
47. Culver,

tumor
Abstract. K.W.,

cells contains
Blaese, R.M.,

the

immune
22. Moolten,

deficiency.
FL. (1986)

86,
for

Z.,

conferred paradigm

by a prospec-

DeVroom,
tumors
kinase gene

H.L.
using
and

Gene

therapy
transduction
ganciclovir.

for the

treatment
with the

of brain
thymidine

serted
tive 23.

herpes
cancer

thymidine
control

intra-tumoral
intravenous

Smith, Radatus, against

K.O., BK.

Cancer Res. 46, 5276-5281. Galloway, KS., Kennell, W.L., Ogilvie, K.K., (1982) A new nucleoside analog, 9i[2-hydroxy-lguanine, types I and highly active 2. Aniimicrob. in vitro

Hum.
Keller,

Gene

Then.

4,

43.

39-69. Tiberghien,

P.,

Reynolds,

C.W.,

(hydroxymethyl)ethoxymethyl1 Chemoiher.

Deschaseaux,
Lyons, R.,

M.,
Chiang,

Certoux,
Y., Longo,

J.M.,
DL.,

Contassot,
Ruscetti,

J., Spence, E., Murphy,

EW (1994)

S., W.J.,
Gan-

24.

25.

herpes simplex virus Agents 22, 55-61. Smee, D.E, Martin, J.C., Venheyden, J.P.H., Matthews, T.R. (1983) Anti herpes virus activity of the acyclic nucleoside 9-(1,3-dihydroxy-2-propoxymethyl) guanine. Aniimicrob. Agents Chemother. 23, 676-682. Averett, DR., Koszalka, G.W., Fyfe, J.A., Roberts, GB., Punfoy, D.J.M., Knenitsky, TA. (1991) 6-Methoxypunine arabinoside as a selective and potent inhibitor of vanicella-zoster virus.

ciclovir

treatment

of herpes

simplex

kinase-transduced

primary Rlood,

T-lymphocytes: depletion after

44. Moolten, means munot. Freeman,

an approach bone marrow

for specific transplantation?

induced

in vivo

donor T-cell in press.

as a

EL. (1990) of improving

Mosaicism chemotherapeutic C.N.,

by gene selectivity. K.A.,

insertion

Crit.

Rev. ImPackman,

10,

203-233.
SM., Abboud, Whartenby,

45.

C.H.,
bystander tumor
46.

Koeplin,
mass

D.S.,
effect: is genetically

Moolten,
tumor

EL.,
regression

Abraham,
when a

G.N.
fraction

(1993)
of

The
the

Aniimicrob. Agents Chemoiher. 35, 851-857. 26. Huber, BE., Richards, CA., Krenitsky, TA. (1991) Retroviralmediated gene therapy for the treatment ofhepatocellular carcinoma: an innovative approach for cancer therapy. Proc. Nail. Acad. Sci. USA 88, 8039-8043. 27. Kilstrup, M., Meng, L.M., Neuhard, J., Nigaard, P. (1989) Genetic evidence for a repressor of synthesis of cytosine deaminase and punine biosynthesis enzymes in Escherzchia coti. .1 Racteriot. 171, 2124-212 7.
28.

Takamiya, Breakefield, nant brain

virus type

Res. 53, 5274-5283. Y., Short, M.P., Ezzeddine, Z.D., Moolten, EL., X.O., Martuza, R.L. (1992) Gene therapy of maligtumors: a rat glioma line bearing the herpes simplex
modified.
i-thymidine Y., kinase gene and wild

Cancer

type

retrovirus

kills

other
47. Takamiya,

tumor

Breakefield,
model j 48. of New 49. Caruso, Plautz, Riot. of G.,

cells. J. Neurosci. Res. 33, 493-503. Short, M.P., Moolten, EL., Fleet, C., Mineta, X.O., Martuza, R.L. (1993) An experimental
gene therapy Nabel, in vivo G.J. with for malignant (1991) a suicide Selective retroviral brain tumors.

T.,

Nishiyama,

T.,

Kawamura,

Y.,

Kawamoto,

K.,

Matsumura,

H.,

retrovirus

Yamamoto, N., Ito, T., Ohyama, A., Katsuragi, T., (1985) Antineoplastic effects in rats of 5-fluorocytosine bination with cytosine deaminase capsules. Cancer
1753-1759.

Sakai, T. in comRes. 45,

Neurosurg.
recombinant

79,
Nabel,

104-110.
E.G., genes elimination vector.

3, 709-715.
M. , Panis, Y. , Gagandeep, S. , Houssin, D. , Salzmann, U. (1993) Regression of established macroafter in situ transductibn of cells for a suicide the treat-

29. 30.

31.

Polak, A., Scholer, H.J. (1975) Mode of action of 5-fluorocytosine and mechanisms of resistance. Chemotherapy 21, 113-130. Koechlin, BA., Rubio, F., Palmer, S., Gabriel, T., Duschinsky, R. (1966) The metabolism of 5-fluorocytosine-2-14-C and of cytosine-14-C in the rat and the disposition of 5-fluorocytosine2-14-C in man. Biochem. Pharmacol. 15, 435-446. Bennett,J.E. (1990) Antimicrobial agents (continued). Antifungal agents. In Goodman and Gilmans The Pharmacological Basis of Therapeutics, 8th ed. (AG. Gilman, T. Rall, AS. Nies, and P. Taylor, eds), New York: Pergamon, 1165-1181.
Calabrisi, P., Chabner, BA. (1990) Antineoplastic agents. In

J. -Li,
scopic

Klatzmann,
liver SM. metastases

gene.
50. Freeman,

Proc. Nail. of cancer.

June DNA

Acad.
(1993)

Sci.

USA

90,

7024-7028.
tumor

Gene-modified

ment
Therapy, 51.

IBCs
21-22,

Second
Washington,

Annual
D.C. Committee

Conference
(RAC) Data

on

Gene

Recombinant

Advisory

Manage-

ment
543-546. 52. Freeman,

Report
SM.,

- December
McCune, C.,

1993
Angel,

(1994)
C.,

Hum.
Abraham,

Gene.

Ther.
G.N.,

5,
Ab-

32.

Goodman apeutics,
br, 33. eds),

and Gilmans The 8th ed. (AG. Gilman,

New York: Pergamon,

Pharmacological T. Rall, AS.

1209-1263.

Basis of ThenNies, and P. Tay53.

boud, gene
342-349.

C.N. (1992) Treatment modified vaccine-regulatory

R.G., Hart, I.R. (1993)

of ovarian cancer using HSV-TK issues. Hum. Gene Ther. 3,

Use oftissue-specific expression of

Mullen, CA., Kilstrup, M., Blaese, R.M. (1992) Transfer of the bacterial gene for cytosine deaminase to mammalian cells confers lethal sensitivity to 5-fluorocytosine: a negative selection system. Pmc. Nail. Acad. Sci. USA 89, 33-37. 34. Walsh, T.J., Pizzo, A. (1988) Treatment ofsystemic fungal infections: recent progress and current problems. Eur. J. Clin. MicrobioL Infect. Dis. 7, 460-475.

Vile,

the herpes simplex growth of established

tratumoral 54. Riddell,

Gilbert,

R.W.,

virus thymidine kinase gene to inhibit munine melanomas following direct ininjection of DNA. Cancer Res. 53, 3860-3864. SR., Greenberg, PD., Overell, R.W., Loughran, T.P., M.J., Lupton, S.D., Agosti, J., Scheeler, S., Coombs, Corey, L. (1992) Phase I study of cellular adoptive im-

208

Journal

of Leukocyte

Biology

Volume

56,

August

1994

munotherapy using genetically modified CD8 HIV-specific T cells for HIV seropositive patients undergoing allogeneic bone marrow transplant. Hum. Gene Ther. 3, 319-338. 55. Lupton, S.D., Brunton, L.L., Kalberg, V.A., Overell, R.W. (1991) Dominant positive and negative selection using a hygromycin phosphotransferase-thymidine kinase fusion gene. MoL CelL RioL 11, 3374-3378. 56. Christy, B., Scangos, G. (1982) Expression of transferred thymidine kinase genes is controlled by methylation. Proc. NaiL Acad. Sci. USA 79, 6299-6303.
57. j 58. Moolten, EL., Wells, Insi. J.M. (1990) Curability of tumors bearing

63. Burakoff, S.J., Deeg, H.J., Ferrara, J., Atkinson, K. (eds) (1990) Graft-vs-Host Disease: Immunology, Paihophysiotogy, and Treatment. Marcel Dekker, New York. Hematology, Vol. 12. 64. Marmont, AM., Horowitz, MM., Gale, R.P., Sobocinski, K., Ash, R.C., Bekkum, D.W.V., Champlin, RE., Dicke, K.A., Goldman, J.M., Good, R.A., Herzig, RH., Hong, R., Masaoka, T., Rimm, A.A., Ringden, 0., Speck, B., Weiner, R.S., Bortin, MM. (1991) T-cell depletion of HLA-identical transplants in leukemia. Rlood 78, 2120-2130. Horrowjtz, MM., Gale, R.P., Sondel, P.M., Goldman, J.M., Kersey, J., Kolb, H.J., Rimm, A.A., Ringden, 0., Rozman, C., Speck, B., Truitt, R.L., Zwaan, EE., Bortin, MM. (1990) Graft-versus-leukemia reactions after bone marrow transplantation. Rlood 75, 555-562. 66. Fletcher, C., Sawchuk, R., Chinnock, B., Miranda, P., Balfour, H.H. (1986) Human pharmacokinetics of the antiviral drug DHPG. Clin. Pharmacot. Ther. 40, 281-285.
67. Moolten, FL., Wells, J.M., Heyman, R.A., Evans, R.M. (1990)

65.

herpes
Nail.

thymidine
Cancer

Venkatesh, G. (1990)
ing human

L.K., Selective

kinase genes transferred by retroviral vectors. 82, 297-300. Arens, M.Q., Subramanian, T., Chinnadurai, induction of toxicity to human cells expressvirus type 1 Tat by a condition-

immunodeficiency

ally
59.

cytotoxic
Byrn,

adenovirus
R.,

vector.

Proc.

Nail.

Acad.

Sci.

USA

87, Temimgene

8746-8750. Kim, S.,

Groopman,

60.

poral aspects of DNA and munodeficiency virus infection: expression. J. ViroL 63, 3708-3713. Caruso, M., Klatzmann, D. (1992)
cells harboring a human

D. (i989) RNA replication during evidence for differential Selective killing

virus-inducible

J., Baltimore,

Lymphoma a herpes 125-134. 68.

regression thymidine

induced kinase

by ganciclovir in mice transgene. Hum. Gene

bearing
Ther.

1,

of CD4
sui-

immunodeficiency

viral spread in an infected cell population. Proc. NaiL Acad. Sci. USA 89, i82-i86. 61. Caruso, M., Tzikas, G., Roussel, M., Alizon, M., Klatzmann, M. (1992) HIV-triggered killing of booby trapped cells prevents viral spread in an HIV-infected cell population. Rone Marrow Transplant. 9(suppl 1), 187-188. 62. Pantaleo, G., Graziosi, C., Fauci, AS. (1993) The immunopathogenesis of human immunodeficiency virus infection. N EngL j Med. 328, 327-335.

cide

gene

prevents

Donahue, RE., Kessler, SW., Bodine, D., Mc Donagh, K., Dinbar, C., Goodman, S., Agricola, B., Byrne, E., Raffeld, M., Moen, R., Bacher, J., Zsebo, KM., Nienmhuis, A.W. (1993) Helper virus induced T-cells in nonhuman primates after retroviral mediated gene transfer.]. Exp. Med. 176, 1125-1135.
Servida,
Nobili,

69.

P.,
N.,

Rossini,
Vago, L.,

R.,

Traversari,
Faravelli, A.,

C.,

Ferrari,

G.,
A.,

Bonini,
Mavilio,

C.,
E,

Vanzulli,

Bordignon, phocytes
immunity

C. (1993) for in vivo

in a patient

Gene transfer immunomodulation

affected by

into

peripheral of donor

blood lymanti-tumor
lymphoma.

EBV-induced

Rtood

82 (suppl

1), 214a

(abst).

Tiberghien

Use

of suicide

genes

in gene

therapy

209