715
REVIEW
ARTICLE
Gene therapy for cancer
The molecular basis of
cancer
is
now
understood
to involve activation of dominant oncogenes and inactivation of tumour suppressor genes, and these genetic events may represent novel targets for cancer
therapy. This review focuses on the potential use and ethical implications of gene transfer to alter the behaviour of somatic cells in cancer patients. Antisense nucleic acids and ribozymes represent informational drugs that may be used to modulate the expression of selected genes and suppress malignant behaviour in cancer cells. Genetic immunomodulation by introducing genes for cytokines into cancer cells or lymphocytes can stimulate a cytotoxic immune response against the tumour. Gene transfer techniques can be applied to target prodrug activation specifically to tumour cells and also to protect normal tissues against toxic chemotherapy. Gene replacement therapy could be used to restore the function of defective tumour suppressor genes. Lancet 1992: 339: 715-21. even
Introduction is a disorder of the genetic resulting in a clone with an abnormal pattern of growth control. Conventional cancer therapy aims to destroy the tumour, while leaving as much as possible of the normal host tissue intact. The trouble with both surgery and radiotherapy is tumour invasion and spread outside areas directly accessible to these treatments; and the trouble with chemotherapy is its low therapeutic ratio for many tumours and the fact that drug resistance may rapidly develop or even be present from the start. Despite successes in a variety of rare tumour types and tremendous efforts over the past twenty years, chemotherapy has had disappointingly little impact on survival in the vast majority We
now
know that
make-up of somatic
cancer
cells
mechanism.
Despite clear-cut evidence for the immunogenicity of tumours in rats and mice, evidence for an effective immune mechanism against most human tumours has proved elusive. The past decade has seen the use of molecular biology to isolate genes and then to produce, in pharmaceutical quantities, cytokines, many of which are involved in the control of the immune system. Some of these, such as the interferons and interleukins, have been demonstrated to have limited efficacy against specific human tumours,’ although the mechanism of action is unclear.2 Against a background of relative stagnation in terms of successful cancer therapy, our understanding of the molecular basis of cell growth control has improved remarkably. Oncogenes and tumour suppressor genes are part of the normal genome; they are vital in the transduction (see Glossary) of physiological signals from outside the cell to its nucleus.3 Subversion of this apparatus can cause cancer. The picture is confusing: for no one type of cancer is molecular pathogenesis precisely understood. Almost certainly a series of cooperating events is necessary for a tumour to emerge. During the next decade it is likely that much of the human genome will have been mapped out and sequenced, permitting comparisons between the genetic constitution of cells exhibiting abnormal growth patterns and their normal counterparts. Furthermore the mechanisms of transcriptional control will be elucidated in detail to explain how different blocks of genes are expressed in differentiation and proliferation. We are already beginning to see prognostic predictions by the use of genetic analysis in breast and ovarian cancer (amplification of erb-B2),4 neuroblastoma (N-myc amplification)5 and adenocarcinoma of the lung (ras mutations) 6 Further analysis of oncogenes and tumour suppressor genes will almost certainly reveal targets for forms of therapy at the molecular level. Gene manipulation is becoming easier. Homologous recombination, retroviral shuttle vectors, and antisense technology,
of solid cancers.
Biological approaches to the treatment of cancer began with the realisation that under certain circumstances there may be not only recognition of tumour cells by the host immune response but also an effective destructive
ADDRESS: ICRF
Oncology Group, Department of Clinical Oncology, Royal Postgraduate Medical School, Hammersmith Hospital, London W12 ONN, UK (A. A. Gutierrez, MD, N R. Lemoine, MB, PhD, Prof K. Sikora, FRCP). Correspondence to Prof Karol Sikora.
716
TABLE I-APPROVALS GRANTED BY NIH RECOMBINANT DNA ADVISORY COMMITTEE FOR GENE THERAPY PROTOCOLS FOR CANCER PATIENTS
Ti L tumour
filtrating
!ymphocytes; Ca= carcinoma.
with the power of the
analytical tools such as the together chain will revolutionise our ability to reaction, polymerase and in monitor cells culture and perhaps whole manipulate Several clinical organisms. protocols for human gene have been therapy given approval by the Recombinant DNA Advisory Committee of the US National Institutes of Health (table l)—and others are now under consideration. The goal of these protocols is to enhance the patient’s anti-tumour immune response, but the potential of gene transfer in cancer goes further than this and could result in novel approaches to halt the malignant behaviour of tumour cells directly (fig 1).
Principles of gene therapy This review focuses on somatic therapy-the insertion of exogenous normal gene into somatic cells to correct an abnormality or deficiency of a specific protein. This can be carried out by transfecting the new gene in the presence of the abnormal gene (addition therapy) or by attempting to replace a defective gene by inserting a new one at the same site, using homologous recombination. The genetic information will not be passed on to future generations. Formal selection criteria for gene therapy have been established and these include: (1) a life-threatening disease; (2) the gene responsible has been cloned; (3) its precise an
Intracellular action
Effect
on
Clinical benefit
regulation is not required; and (4) a suitable delivery system is available. Considerable effort is going into somatic gene therapy for single gene disorders such as adenosine deaminase immunodeficiency (fig 2) and thalassaemia.7 These diseases might seem to be more amenable to this approach than a complex polygenic disorder such as cancer. However, there are several systems that have been explored in malignancy to enhance the selective destruction of normal cells. The other type of gene therapy, germ cell transfection, poses considerable ethical dilemmas. It would be specifically excluded by the proposals of the UK Committee on the Ethics of Gene Therapy, which reported in January of this year. Germ-cell therapy probably has a lesser part to play in cancer therapy. At some point in the future it might be used to prevent cancers in individuals carrying defective tumour suppressor genes (eg, the RB gene in retinoblastoma and p53 in Li-Fraumeni syndrome). tumour versus
Methods The object is to get a functional new gene into a cell. There are main methods, physical and virus mediated (table n). Physical methods include transfection by calcium phosphate precipitation, two
electroporation, microinjection, protoplast fusion, liposornal transfer or receptor-mediated delivery.8 More recently interesting studies using the direct injection of vectors containing dystrophin into murine muscle have shown at least short-term expression of this protein in cells surrounding the injection site.9 Viral transfer with ecotropic or amphotropic retroviruses have been the most widely used due to their efficiency of infection and the fact that the dosage of gene transfer can be controlled more effectively than with the physical methods. Modified adenovirus and herpes simplex virus have also been investigated because they infect specific tissues such as epithelial cells. 10 A cell line can be transfected quite readily in vitro but to do this for a solid tumour in vivo is less easy. Gene transfer in vivo has been achieved, with variable success, by various routes, including
intra-arterial, intravenous, intraperitoneal, intrahepatic, intramuscular, and intratracheal injection, by both physical and viral methods.l Much effort has gone into improving the expression of genes once they are inserted into their target cells. These have included the use of selective promoters to drive transcription of the transfected genes efficiently. A reporter gene system is normally used to trace the efficiency of the insertion,12 one example being a resistance gene for neomycin. This may allow for a selective process, so enhancing the success of transfer. There are two major limitations to mammalian cell transfection. The first is a much lower efficiency of gene expression in comparison with prokaryotic systems, with considerable differences between eukaryotic cell lines. Unlike rodent cells, most primate and human cells can integrate only a small amount of foreign DNA (about 6 kilobases); as a result only 10-30% of clones selected for the expression of one transcription unit will also contain a second unit in intact form. 13 The second problem is the short-lived response after successful transfection (a few months at most) regardless of the method used. 14 We know very little about the processing steps within the cell. Clearly, there are problems of degradation by
717
peripheral T lymphocytes from child with ADA deficiency (severe combined immunodeficiency, SCID)
TABLE II-DELIVERY SYSTEMS FOR GENE TRANSFER
Harvest
Physical Calcium-phosphate precipitation Electroporation Microinjection Protoplast fusion Liposomal transfer Receptor-mediated delivery Tissue injection
I
Virus mediated Retroviruses Adenovirus
Herpes simplex virus reach tumour cell targets in vivo. Major problems here include cell permeability, chemical stability, targeting, scale-up, and toxicity.15 Antisense oligodeoxynucleotides are small synthetic nucleotide sequences formulated to be complementary to specific DNA or RNA sequences. By the binding of these nucleotides to their targets the transcription or translation of a single gene can be selectively inhibited. If that gene is responsible for a disease process then its down-regulation could result in a reversal of the clinical abnormalities. The cytoplasmic location of mRNA provides an easier target for oligodeoxynucleotides than DNA. One example of an antisense oligomer with antiproliferative activity is c-myc in lymphoma cell lines;ib others are N-mye in neurectodermal cell lines;17 c-myb in colon adenocarcinoma,18 type 1 regulatory subunit of the cAMP receptor protein kinase in K-ras transformed NIH3T3 cells19 and in neuroblastoma, leukaemia, breast, colon, and gastric carcinoma cells, bcr-abl in chronic myeloid leukaemia blast cells,zo and c-raf 1 in ras and raf-transformed NIH3T3 ce11s.21In McManaway and colleagues’ in-vitro study, the antisense oligonucleotide was only 21 bases long.16 Growth of two Burkitt’s lymphoma cell lines known to produce an abnormal mRNA was strikingly inhibited, and this finding suggested an "Achilles heel" in the growth of this neoplasm. Antisense oligonucleotides also decrease the tumorigenicity or metastatic potential of cell lines. A K-ras proto-oncogene in antisense orientation transduced into a small cell lung cancer line inhibited of K-ras expression and also suppressed tumour growth in nude mice. There was no alteration in growth kinetics of the cell line in vitro. Other experiments have shown that the antisense sequence to pre-prourokinase significantly reduced the ability of murine melanoma cells to colonise lung. By contrast, if the antisense oligomer is designed to inhibit a putative metastasis suppressor gene, such as E-cadherin, the resulting downregulation renders the cells invasive.22 These results demonstrate the power of antisense technology and the potential application it has for manipulation of abnormal growth and behaviour in tumour cells.
to
.. Select for antibiotic resistance encoded by neo gene and for ADA gene function in vitro
Expand positive
clones and infuse back into patient
Fig 2-Application of gene transfer to
single gene disorder. Included hereto illustrate an application of gene therapy that isfurther advanced clinically than the cancer protocols reviewed here. a
extracellular nucleases, absorption onto and uptake into cells, transport from cytoplasm to nucleus, integration into host
chromosomes, mutation, the expression of non-integrated DNA, and transcription control of the transgene. Also, there is a small theoretical hazard of toxicity from the transfected DNA. It is calculated that the cumulative probability of having a harmful effect is less than 1 in 1016 per DNA molecule from a cell without activated proto-oncogenes or active viral oncogenes. Such a risk is likely to be acceptable with terminal cancer, but the evolution of safe, controllable gene delivery and expression systems is vital for successful therapy.
Biological approaches to cancer Recombinant drugs
Although drugs produced by recombinant DNA methods are not directly a form of gene therapy, they rely on the same technology. These agents include cytokines, growth factors and growth factor antagonists, humanised monoclonal antibodies, toxin-antibody immunoconjugates, toxin-ligand conjugates, and single-chain antigen-binding proteins. They have limited utility in cancer because the development of selectively toxic molecules has proved difficult. The selectivity of even the most specific of monoclonal antibodies with high tumour affinity is limited, and this almost certainly explains the poor clinical results so far. Improvements in molecular design and the identification of novel specific molecular targets may enhance their role.
Informational drugs Informational drugs are synthesised molecules that carry biological information which allows them to act in a specific manner. Oligonucleotides, ribozymes, and specific proteases act at different but specific levels of the
transcription system. These compounds can be of low molecular weight and therefore have considerable potential
Genetic immunomodulation
Cytokines can have a significant effect on tumour growth clinically but have systemic side-effects and very short half-lives. Many of the current prospects for gene therapy involve the vectoring of genes encoding for cytokines, with cells that home specifically on tumours. The anti-tumour immune response to the host can be improved by modifications of the immune system. Such modifications include the transfection of pleiotropic cytokine genes into the immune cells of the host or into tumour cells (eg, interleukin 2 or 4, interferon, tumour necrosis factor). The expression of cytokines in transfected cells can reduce their tumorigenicity and/or metastatic
718
TABLE III-SELECTIVE GENE EXPRESSION FOR POTENTIAL USE IN VIRALLY DIRECTED ENZYME PRODUCING THERAPY
GI=gastromtestinal,
SCLC= small cell
lung
cancer.
potential. The anti-tumour response can be enhanced by cytotoxic lymphocytes, macrophages, or antibodies which can in turn be induced by the expression of appropriate cytokine by tumour cells. In this way it may be possible to convert a weakly immunogenic tumour that elicits only a minor response to a strongly immunogenic one that could provoke a powerful destructive reactivity. Enhanced expression of major histocompatibility (MHC) molecules of class I and II can be achieved by transfected cytokine genes or by direct MHC gene transfection.23 The best characterised model so far has been the transfection of the tumour necrosis factor (TNF) gene into tumour-infiltrating lymphocytes. TNF has produced very encouraging antitumour responses in mice yet clinically it has proved disappointing, perhaps because toxicity limits the doses that can be used in man. Toxicity is common at doses above 8 Ilg/kg in man whereas in the mouse 400 ug/kg can be achieved. If a higher local concentration could be reached clinically, the response rate might increase. In a pilot study in patients with melanoma Rosenberg and colleagues24 demonstrated that tumour-infiltrating lymphocytes (TIL) transfected with a reporter gene (neomycin phosphotransferase) had localised in the tumour and continued to express foreign genes for up to ten months. A clinical trial with TNF-transfected cells is under way: TNF-transfected TIL cells will be infused into fifty patients with melanoma. It has previously been demonstrated that TIL cells produce responses in up to 30% of melanoma patients.zs The results from this clinical trial should be available some time this year. Another approach to genetic immunomodulation is vaccine development. Human papillomavirus (HPV), especially subtype 16, has been implicated in the aetiology of cervical cancer. Recombinant vaccines expressing HPV nuclear proteins E6 or E7 have already been demonstrated to reduce the development of HPV-related tumour in animals. Fibroblast-like cells transfected with HPV-16 E7 genes have been used to immunise syngeneic mice,26 and these cells confer protection against tumorigenicity of HPV-16 E7 positive tumour cells in these animals. Normal tissue protection The production of cytotoxic-drug-resistant stem cells may be a mechanism by which normal cells could be protected from treatment-related toxicity, allowing much larger doses to be used. The proposed transgenes can render stem cells resistant to some specific drugs—eg, methotrexate by use of dihydrofolate reductase (DHFR)27—or confer on them more widespread multidrug resistance (MDR) by the expression of the MDR1 gene.2$ An exciting clinical area is the infusion of haemopoietic
Fig 3-Virally directed
enzyme
prodrug therapy.
(a) Transcription factors in host cell results in production of viral enzyme from mtegrated vector, activating prodrug (b) Best example of selective toxicity is discrimination of normal liver and hepatoma cells by retroviral vectors containing AFP and albumin gene promoters. Normal liver cells are only destroyed when transfected with construct containing albumin promoter and hepatoma cells only by construct containing AFP promoter.
colony-stimulating factors (CSF) to support bone marrow function in cancer patients undergoing chemotherapy. These molecules include CSF for granulocytes, and macrophages, granulocyte-macrophage, and a relatively unchartered interleukins, erythropoietin, stem-cell factor. An alternative to infusion would be the genetic manipulation of normal bone marrow cells before chemotherapy. This might achieve a more continuous effect than the infusion of a drug with a short half-life. One study has demonstrated that the stable transfection of human G-CSF into fibroblasts in vitro produced neutrophilia on implantation into nude mice. The number ofhaemopoietic progenitor cells was increased in the spleen as well as in the bone marrow. The GM-CSF gene has also been transfected and expressed, showing increased tolerance to cytotoxic drugs.29 The efficiency of gene insertion into stem cells is low, however-10-20% in murine stem cells and only 1-5% in primates. Drug targeting The most promising selective genetic mechanism for drug targeting is virally directed enzyme prodrug therapy (VDEPT). This is based on a vector being expressed specifically in cells of a particular tissue or specifically in tumour
cells but
not
in normal cells. There
are
several
examples where tumours show tumour-specific transcription of certain host genes that are not essential for survival (table in). As our understanding of the mechanism of gene transcription in eukaryotes increases, it is likely that
719
TABLE IV-EXAMPLES OF SUPPRESSION OF MALIGNANT PHENOTYPE BY GENE TRANSFER IN VITRO
Fig 4-Strategy for VDEPT system tumour
to
produce 5-FU within
cells.
we will be able to
develop more and better VDEPT systems. The best current example of VDEPT working in vitro comes from the discrimination between normal liver and hepatoma cells by the differential expression of vectors containing promoters for albumin and ot-fetoprotein.1 In normal liver a herpes simplex thymidine kinase gene is expressed when coupled to the albumin promoter but not when coupled to the a-fetoprotein promoter (fig 3). The converse applies in hepatoma cells. Thymidine kinase can convert the prodrug 6-methoxypurine arabinonucleoside (Ara-M) which has minimum effects in normal cells, to the phosphates Ara AMP, ADP, and ATP, which are potent cytotoxic agents. In vivo studies of murine systems are being tested with this system. Another promising viral prodrug system is the conversion of the anti-fungal drug 5-fluorocytosine to the cytotoxic 5-fluorouracil (5FU) by cytosine deaminase driven by a selective promoter. One attraction of this system is that 5-fluorocytosine and 5-FU are in routine clinical use, and much pharmacological information is available for both. A strategy for a 5-FU VDEPT system is outlined in fig 4. There are several steps for potential selectivity. The first is based on selective infection by the virus. This may be achieved by targeting with immunoliposomes or a virus with predetermined tissue specificity. The second, and perhaps most important, selection is the expression of the vector only in the target cells that possess the appropriate transcription machinery. Although 5-FU may not be the most effective cytotoxic drug it is an excellent starting-point since so much is known of its pharmacokinetics. Gene replacement therapy Some human tumours overexpress dominant oncogenes fail to express normal tumour suppressor genes. Point mutations may also activate the transforming function of both oncogenes and suppressor genes.
or
There is now good evidence that in vitro reversion of malignant morphology can be achieved by a variety of different sequences (eg, normal H-ras, raplA [K-n], and c-jun).Other sequences can increase the immunogenicity of tumour cells making them more susceptible to the cellular antitumoural response of the host. Further genes can selectively inhibit the metastatic potential of cells, examples being nm23, &bgr;-actin, fibronectin receptor, connexin, and E-cadherin. The nm23 gene encodes a product that is probably involved in the control of ras-related proteins, and decreased
I
I
’
I
expression allelic deletion of this gene is associated with high metastatic potential in breast and colorectal cancers. Transfection of the nm23 gene with highly metastatic mouse melanoma cells reduced their metastatic potential and abrogated their responsiveness to the cytokine TGF-&bgr;.31 Finally, there are genes with recognised tumour suppressor properties such as RB and wild type p53 (table IV). The wild type p53 gene (WTp53) encodes a short half-life nuclear phosphoprotein with many properties consistent with tumour suppressor activity. WTp53 can promote cell differentiation and suppress the proliferation rate of transfected tumour cells. Expression of this gene can abolish tumorigenicity in vivo when such cells are placed in mude mice. The expression of WTp53 (but not mutant p53) inhibits proliferation of different human cell lines lacking p53 expression or expressing a mutated p53 allele. These include colorectal, breast, osteosarcoma, glioblastoma, and ovarian carcinoma cells.32 p53 mutations have now been implicated as a late event in tumour progression in many human cancers, and in the germline of at least some families with the Li-Fraumeni syndrome it may be possible to develop homologous recombination systems to insert functional p53 in place of the mutated gene. The loss or inactivation of both RB alleles has been observed in different types of tumour, apart from retinoblastoma itself--eg, breast cancer, small cell lung cancer, prostate, bladder, osteosarcoma, and other soft tissue tumours. RB gene transfection into RB negative or RB mutated cells from retinoblastoma, osteosarcoma, and prostate carcinoma lines can inhibit the malignant phenotype both in vitro and in vivo. The mechanisms by which this gene exerts its action are complex. There is evidence that pl05 Rb acts as a transcription factor binding to DNA. Mutation, deletion, or sequestration by adenovirus Ela, simian virus 40 T, and HpVH 16 E7 protein may result in a uncontrolled growth. The demonstration that the insertion of this gene in appropriate model systems can restore normal growth control provides an encouraging avenue for further development.33 Another intriguing molecule is GC factor (GCF). This is the product of a gene that encodes a transcription factor that downregulates the expression of certain genes associated with human cancer. In vitro assays have shown that this factor inhibits the transcription of epithelial growth factor receptor, p-actin, and calcium-dependent protease by interacting with their promoters. This is due to its affinity for specific GC-rich sequences found in these and other regulatory sites in DNA.34 The significance in or
720
cancer
suppression in vivo remains to be elucidated. It may well act system for studying downregulation of growth by similar
REFERENCES
as a model
mechanisms.
With gene replacement therapy, however, we lack suitable control technology for expression vectors, and it is unlikely that this strategy will have direct consequences in clinical care for some time.
Conclusion There have been remarkable developments in our understanding of the molecular basis of human cancer. This is likely to continue as we learn more about the human genome and about transcriptional control. The recognition of a system of positively acting growth signals and negatively acting growth inhibitory signals working in parallel will lead
rationally designed pharmacological agents active against cancer. Many hurdles, ethical and technical, remain, but selection criteria for gene therapy are being established’ and the ethical issues have been extensively reviewed.47-49 The development of secondary tumours and/or autoimmune disorders after gene therapy is a theoretical hazard, but the risk may be comparable to that which we now recognise for chemotherapy regimens. Gene therapy for cancer will require the cooperation of physicians, molecular biologists, and virologists to develop safe, robust systems for clinical use. The ethical issues will be resolved by rigorous review of protocols (in the US by the Recombinant DNA Advisory Committee in the UK by the Committee on the Ethics of Gene Therapy, until a more formal structure is devised) and by public education. The transition from theoretical possibility to practical reality may be only beginning but we could be at the dawn of a new age to new
of cancer treatment. We thank Dr Helen Hurst for helpful suggestions. A. A. G. has been partly supported by a grant from the Instituto Mexicano del Seguro Social, IMSS.
1. Mastrangelo MJ, Schultz S, Kane M, et al. Newer immunologic approaches to the treatment of patient with cancer. Semin Oncol 1988; 15: 589-94. 2. Smith KA. Interleukin 2: inception, impact and implications. Science 1988; 240: 1169-70. 3. Bishop JM. Molecular themes in oncogenesis. Cell 1991; 64: 235-48. 4. Borg A, Baldetrop B, Ferno M, et al. Erb-B2 amplification in breast cancer with a high rate of proliferation. Oncogene 1991; 6: 137-43. 5. Brodeur G, Seeger RC, Schwab M, et al. Amplification of N-myc in untreated neuroblastoma correlates with advanced disease stage. Science 1984; 24: 1121-24. 6. Slebos RT, Kiberlaan RE, Paledio O, et al. K-ras oncogene activation as a prognostic marker in adenocarcinoma of the lung. N Engl J Med 1990; 323: 561-65. 7. Weatherall DJ. Gene therapy in perspective. Nature 1991; 349: 255-76. 8. Benvenisty N, Reshef L. Director introduction of genes into rats and expression of the genes. Proc Natl Acad Sci USA 1986; 83: 275-76. 9. Ascadi G, Dicson G, Love DR, et al. Human dystrophin expression in mdx mice after intramuscular injection of DNA constructs. Nature 1991; 352: 815-18. 10. Geller AI, Keyomarsi K, Bryan J, Pardee AB. An efficient deletion mutant packaging system for defective herpes simplex virus vectors: potential applications to human gene therapy and neuronal physiology. Proc Natl Acad Sci USA 1990; 87: 8950-54. 11. Wu Ch, Wilson JM, Wu Gy. Targeting genes: delivery and persistent expression of a foreign gene driven by mammalian regulatory elements in vivo. J Biol Chem 1989; 264: 16985-87. 12. Kohn DB, Kantoff PW, Rglitis MA, McLachlin JR, Moen RC, et al. Retroviral mediated gene transfer into mammalian cells. Blood 1987; 13: 285-98. 13. Temin HM. Overview of biological effects of addition of DNA molecules to cells. J Med Virol 1990; 31: 13-17. 14. Johnson P, Gray D, Mowat M, Benchimol S. Expression of wild-type p53 is not compatible with continued growth of p53-negative tumor cells. Mol Cell Biol 1991; 11: 1-11. 15. Miller AD. Retrovirus packaging cells. Human-Gene-Ther 1990; 1: 5-14. 16. McManaway ME, Neckers LM, Loke SL, et al. Tumour specific inhibition of lymphoma growth by an antisense oligodeoxynucleotide. Lancet 1990; 335: 808-11. 17. Whitesell L, Rosolen A, Neckers LM. Episome-generated N-myc antisense RNA restricts the differentiation potential of primitive neuroectodermal cell lines. Mol Cell Biol 1991; 11: 1360-70. 18. Melani C, Rivoltini L, Parmiani G, Calabretta B, Colombo MP. Inhibition of proliferation by c-myb antisense oligodeoxynucleotides in
Glossary and abbreviations
’
Small Antisense oligodeoxynucleotides. to be formulated nucleotide sequences synthetic complementary to specific DNA or RNA sequences. They are targeted to inhibit selectively the transcription or translation of a gene. Cloning vector. A plasmid or bacteriophage used to "transport" inserted DNA for the purposes of producing more material or a protein product. Homologous recombination. Recombination between homologous sequences in a chromosome and allows direct replacement of the original DNA sequence with an exogenous segment. Long-terminal repeat. A promoter sequence directly repeated at both ends of retroviral DNA. It regulates the transcription of a gene. Oncogene. Gene where product has ability to transform eukaryotic cells to neoplastic phenotype. Promoter. A region of DNA involved in binding of RNA polymerase to initiate transcription of a gene. Reporter gene. A coding unit whose product is easily assayed (eg, neomycin resistance) in transfected cells, used to evaluate transfection efficiency. Retrovirus vectors. Viruses equipped to incorporate themselves into host DNA. These are "ecotropic" if the host range is limited to rodents or "amphotropic" are if the host range is wide, including man.
Retroviral shuttle vector. A construct with two hosts so that it can be
origins for replication for
used to carry a foreign sequence in prokaryotes or eukaryotes. Transcription. Synthesis of RNA from a DNA template. Transduction (signal). Transfer and processing of growth factor or other signals via cytoplasmic messengers to influence nuclear events. Transduction (gene). Acquisition and transfer of eukaryotic genetic material be retroviruses. Transfection. In eukaryotic cells this is the acquisition of new genetic material by incorporation of added DNA through physical or virus-dependent methods. AFP
ot-fetoprotein. CSF=co)ony-stimu!ating =
factors.
DHFR=dihydrofolate
LTR
=
long terminal
repeat.
MDR=multidrug resistance.
MHC= major histocompatibility. EGFR=epithelium derived growth factor receptor. RB retinoblastoma. 5- FU = 5-fluorouracil. TIL=tumour-infiltrating HPV= human lymphocytes. TNF=tumour necrosis papillomavirus. reductase.
=
IL=interleukin. LASN vector50 = vector carrying a human adenosine deaminase gene
factor.
VDEPT=virally directed enzyme prodrug therapy.
721
colon adenocarcinoma cell lines that express c-myb. Cancer Res 1991; 51: 2897-901.
33.
YS, Clair T, Tortora G, Yokozaki H, Pepi S. Suppression of malignancy targeting the intracellular signal transducing proteins of cAMP: the use of site selective cAMP analogs antisense strategy and gene transfer. Life Sci 1991; 48: 1123-32. 20. Szczylik C, Skorski T, Nicolaides NC, et al. Selective inhibition of BCR-ABL antisense cell leukemia proliferation by oligodeoxynucleotides. Science 1991; 253: 562-68. 21. Kolch W, Heidecker G, Lloyd P, Rapp UR. Raf-1 protein kinase is required for growth of induced NIH/313 cells. Nature 1991; 349: 19. Cho-Chung
426-28.
22. Vleminck K, Vakaet L, Mareel M, Fiers W, Van Roy F. Genetic of E-Cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 1991; 66: 107-19. 23. James RF, Edwards S, Hui KM, et al. The effect of Class II gene transfection on the tumorigenicity of the H-2K negative mouse leukemia cell line. Immunology 1991; 72: 213-18. 24.Rosenberg SA, Albersold P, Cornetta K, et al. Gene transfer into humans: immunotherapy of patients with advanced melanoma, using tumor-infiltrating lymphocytes modified by retroviral gene transduction. N Engl J Med 1990; 323: 570-78. 25. Aebersold P, Hyatt C, Johnson S, et al. Lysis of autologous melanoma cells by tumor-infiltrating lymphocytes: association with cLnical response. J Natl Cancer Inst 1991; 83: 932-37. 26. Chen L, Thomas EK, Hu SL, Hellstrom I, Hellstrom KE. Human papilloma virus type 16 nucleoprotein E7 is a tumour rejection antigen. Proc Natl Acad Sci USA 1991; 88: 110-14. 27. Corey CA, Desilva AD, Holland CA, Williams DA. Serial transplantation of methotrexate-resistance bone marrow: protection of murine recipients from drug toxicity by progeny of transduced stem cells. Blood 1990; 75: 337-43. 28. Galski H, Sullivan M, Willingham M, et al. Expression of a human multidrug resistance CDNA(MDRI) in the bone marrow of transgenic mice. Mol Cell Biol 1989; 9: 4357-63. 29. Keith WN, Brown R, Pragnell IB. Retrovirus mediated transfer and expression of GM-CSF in haematopoietic cells. Br J Cancer 1990; 62: 388-94. 30. Huber BE, Richards CA, Krenitsky TA. Retroviral-mediated gene therapy for the treatment of hepatocellular carcinoma: an innovative approach for cancer therapy. Proc Natl Acad Sci 1991; 88: 8039-43. 31. Leone A, Flatow U, King CR, et al. Reduced tumor incidence, metastatic potential and cytokine responsiveness of nm23-transfected melanoma cells. Cell 1991; 65: 25-35. 32. Mercer WE, Shields MT, Lin D, Appella E, Ulrich SJ. Growth suppression induced by wild-type p53 protein is accompanied by selective down-regulation of proliferating-cell nuclear antigen
manipulation
expression. Proc Natl Acad Sci USA 1991; 88: 1958-62. Huang HJS, Yee JK, Shew JY, et al. Suppression of the neoplastic phenotype by replacement of the RB gene in human cancer cells. Science 1988; 242: 1563-66.
Johnson A, Kageyama R, Popescu NC, Pastan I. Expression and chromosomal localisation of the human transcriptional repressor GCF. J Biol Chem (in press). 35. Weissman BE, Saxon PJ, Pasquale SR, et al. Introduction of a normal
34.
human chromosome 11 into
a
Wilms
tumour
cell line controls its
tumorigenic expression. Science 1987; 175: 181. 36. Declue JE, Zhang KE, Redford P, et al. Suppression of transformation by over-expression of full length GTPA activating protein (GAP) or of the GAP C terminus. Mol Cell Biol 1991; 11: 2819-25. 37. Wallich R, Bulbuc N, Hammerling J, et al. Abregation of metastatic properties of tumour cells by de novo expression of H2K antigens following H2 gene transfection. Nature 1985; 315: 301-05. 38. Sadano H, Taniguchi S, Baba T. Newly identified types of &bgr;actin reduces invasiveness of mouse p16 melanoma. FEBS Lett 1990; 271: 23-27. 39. Vleminck K, Vakaet L, Mareel M, et al. Genetic manipulation of E cadherin expression by epithelial tumour cells reveals an invasion suppressor role. Cell 1991; 66: 107-19. 40. Giancotti FG, Ruoslahti E. Elevated levels of the alpha5 beta 1 fibrinectin receptor suppress the transformed phenotype of Chinese Hamster ovary cells. Cell 1990; 60: 849-59. 41. Leone A, Flatow U, King CR, et al. Reduced tumour incidence metastatic potential and cyclokine responsiveness of nn23 transfected melanoma cells. Cell 1991; 65: 25-35. 42. Rollins B, Sunday M. Suppression of tumour formation in vivo by expression of the JE gene in malignant cells. Mol Cell Biol 1991; 11: 3125-31. 43. Kitayama H, Sugimoto Y, Matsuzaki T, Ikawa Y, Noda M. A ras-related gene with transformation suppressor activity. Cell 1989; 63: 77-84. 44. Friedman T, Roblin R. Gene therapy for human genetic disease? Science 1972; 175: 949-55. 45. Friedman T. Progress towards gene therapy. Science 1989; 244: 1275-80. 46. Anderson WF. Prospects for human gene therapy. Science 1984; 226: 401-09. 47. Motulsky AG. Impact of genetic manipulation on society and medicine. Science 1983; 219: 135-40. 48. Recommendations of European Medical Research Councils. Gene therapy in man. Lancet 1988; 1: 1271-72. 49. Murray TH. Ethical issues in human genome research. FASEB J 1991; 5: 55-60. 50. Culver KW, Osborne WRA, Miller AD, et al. Correction of ADA deficiency in human T lymphocytes using retroviral-mediated gene transfer. Transplantation Proc 1991; 23: 170-71.
STROKE OCTET Complications of acute stroke
Acute stroke is remarkable not for the damage it inflicts but for the recovery it allows. Nevertheless, both recovery and survival can be jeopardised by cerebral, systemic, and cardiac complications. The causes vary with time after the stroke (table I). Rigorous attention to general medical factors can attenuate some of these effects (table n).
Cerebral
complications
Overall stroke-related case-fatality is about 20%; the percentage ranges from 15% for supratentorial and brainstem infarcts to 58% for supratentorial haemorrhage.1 Case-fatality from infratentorial haemorrhage is 31 %.1
Transtentorial herniation and cerebral oedema
During the first week,
transtentorial herniation is the of death (table i); the incidence of this peaks within 24 hours for cerebral and at 4-5 days for cerebral infarction.1
commonest cause
complication haemorrhage
Brainstem compression with subsequent haemorrhage and infarction accounts for the serious morbidity and mortality associated with herniation. Herniation is the result of raised intracranial pressure caused by cerebral oedema, and is seen with large strokes.2 Postmortem studies show microscopic cerebral oedema in 93% of stroke patients;2 this high figure may reflect selection bias in necropsy-based investigations. In an antemortem study, 41 % of 22 patients showed computed tomographic (CT) evidence of a mass effect associated with oedema.2There was a good correlation between infarct size, mass effect, midline shift, neurological ADDRESSES: Cerebrovascular Program, Johns Hopkins Hospital and Johns Hopkins University School of Medicine, Baltimore, Maryland, USA (S Oppenheimer, FRCPC), and Department of Clinical Neurological Sciences, University of Western Ontario, London, Ontario Canada (Prof V. Hachinski, FRCPC). Correspondence to Dr Stephen Oppenheimer, Johns Hopkins Hospital, Meyer 5-181, 600 N Wolfe Street, Baltimore MD 21205, USA.
REVIEW
ARTICLE
Gene therapy for cancer
The molecular basis of
cancer
is
now
understood
to involve activation of dominant oncogenes and inactivation of tumour suppressor genes, and these genetic events may represent novel targets for cancer
therapy. This review focuses on the potential use and ethical implications of gene transfer to alter the behaviour of somatic cells in cancer patients. Antisense nucleic acids and ribozymes represent informational drugs that may be used to modulate the expression of selected genes and suppress malignant behaviour in cancer cells. Genetic immunomodulation by introducing genes for cytokines into cancer cells or lymphocytes can stimulate a cytotoxic immune response against the tumour. Gene transfer techniques can be applied to target prodrug activation specifically to tumour cells and also to protect normal tissues against toxic chemotherapy. Gene replacement therapy could be used to restore the function of defective tumour suppressor genes. Lancet 1992: 339: 715-21. even
Introduction is a disorder of the genetic resulting in a clone with an abnormal pattern of growth control. Conventional cancer therapy aims to destroy the tumour, while leaving as much as possible of the normal host tissue intact. The trouble with both surgery and radiotherapy is tumour invasion and spread outside areas directly accessible to these treatments; and the trouble with chemotherapy is its low therapeutic ratio for many tumours and the fact that drug resistance may rapidly develop or even be present from the start. Despite successes in a variety of rare tumour types and tremendous efforts over the past twenty years, chemotherapy has had disappointingly little impact on survival in the vast majority We
now
know that
make-up of somatic
cancer
cells
mechanism.
Despite clear-cut evidence for the immunogenicity of tumours in rats and mice, evidence for an effective immune mechanism against most human tumours has proved elusive. The past decade has seen the use of molecular biology to isolate genes and then to produce, in pharmaceutical quantities, cytokines, many of which are involved in the control of the immune system. Some of these, such as the interferons and interleukins, have been demonstrated to have limited efficacy against specific human tumours,’ although the mechanism of action is unclear.2 Against a background of relative stagnation in terms of successful cancer therapy, our understanding of the molecular basis of cell growth control has improved remarkably. Oncogenes and tumour suppressor genes are part of the normal genome; they are vital in the transduction (see Glossary) of physiological signals from outside the cell to its nucleus.3 Subversion of this apparatus can cause cancer. The picture is confusing: for no one type of cancer is molecular pathogenesis precisely understood. Almost certainly a series of cooperating events is necessary for a tumour to emerge. During the next decade it is likely that much of the human genome will have been mapped out and sequenced, permitting comparisons between the genetic constitution of cells exhibiting abnormal growth patterns and their normal counterparts. Furthermore the mechanisms of transcriptional control will be elucidated in detail to explain how different blocks of genes are expressed in differentiation and proliferation. We are already beginning to see prognostic predictions by the use of genetic analysis in breast and ovarian cancer (amplification of erb-B2),4 neuroblastoma (N-myc amplification)5 and adenocarcinoma of the lung (ras mutations) 6 Further analysis of oncogenes and tumour suppressor genes will almost certainly reveal targets for forms of therapy at the molecular level. Gene manipulation is becoming easier. Homologous recombination, retroviral shuttle vectors, and antisense technology,
of solid cancers.
Biological approaches to the treatment of cancer began with the realisation that under certain circumstances there may be not only recognition of tumour cells by the host immune response but also an effective destructive
ADDRESS: ICRF
Oncology Group, Department of Clinical Oncology, Royal Postgraduate Medical School, Hammersmith Hospital, London W12 ONN, UK (A. A. Gutierrez, MD, N R. Lemoine, MB, PhD, Prof K. Sikora, FRCP). Correspondence to Prof Karol Sikora.
716
TABLE I-APPROVALS GRANTED BY NIH RECOMBINANT DNA ADVISORY COMMITTEE FOR GENE THERAPY PROTOCOLS FOR CANCER PATIENTS
Ti L tumour
filtrating
!ymphocytes; Ca= carcinoma.
with the power of the
analytical tools such as the together chain will revolutionise our ability to reaction, polymerase and in monitor cells culture and perhaps whole manipulate Several clinical organisms. protocols for human gene have been therapy given approval by the Recombinant DNA Advisory Committee of the US National Institutes of Health (table l)—and others are now under consideration. The goal of these protocols is to enhance the patient’s anti-tumour immune response, but the potential of gene transfer in cancer goes further than this and could result in novel approaches to halt the malignant behaviour of tumour cells directly (fig 1).
Principles of gene therapy This review focuses on somatic therapy-the insertion of exogenous normal gene into somatic cells to correct an abnormality or deficiency of a specific protein. This can be carried out by transfecting the new gene in the presence of the abnormal gene (addition therapy) or by attempting to replace a defective gene by inserting a new one at the same site, using homologous recombination. The genetic information will not be passed on to future generations. Formal selection criteria for gene therapy have been established and these include: (1) a life-threatening disease; (2) the gene responsible has been cloned; (3) its precise an
Intracellular action
Effect
on
Clinical benefit
regulation is not required; and (4) a suitable delivery system is available. Considerable effort is going into somatic gene therapy for single gene disorders such as adenosine deaminase immunodeficiency (fig 2) and thalassaemia.7 These diseases might seem to be more amenable to this approach than a complex polygenic disorder such as cancer. However, there are several systems that have been explored in malignancy to enhance the selective destruction of normal cells. The other type of gene therapy, germ cell transfection, poses considerable ethical dilemmas. It would be specifically excluded by the proposals of the UK Committee on the Ethics of Gene Therapy, which reported in January of this year. Germ-cell therapy probably has a lesser part to play in cancer therapy. At some point in the future it might be used to prevent cancers in individuals carrying defective tumour suppressor genes (eg, the RB gene in retinoblastoma and p53 in Li-Fraumeni syndrome). tumour versus
Methods The object is to get a functional new gene into a cell. There are main methods, physical and virus mediated (table n). Physical methods include transfection by calcium phosphate precipitation, two
electroporation, microinjection, protoplast fusion, liposornal transfer or receptor-mediated delivery.8 More recently interesting studies using the direct injection of vectors containing dystrophin into murine muscle have shown at least short-term expression of this protein in cells surrounding the injection site.9 Viral transfer with ecotropic or amphotropic retroviruses have been the most widely used due to their efficiency of infection and the fact that the dosage of gene transfer can be controlled more effectively than with the physical methods. Modified adenovirus and herpes simplex virus have also been investigated because they infect specific tissues such as epithelial cells. 10 A cell line can be transfected quite readily in vitro but to do this for a solid tumour in vivo is less easy. Gene transfer in vivo has been achieved, with variable success, by various routes, including
intra-arterial, intravenous, intraperitoneal, intrahepatic, intramuscular, and intratracheal injection, by both physical and viral methods.l Much effort has gone into improving the expression of genes once they are inserted into their target cells. These have included the use of selective promoters to drive transcription of the transfected genes efficiently. A reporter gene system is normally used to trace the efficiency of the insertion,12 one example being a resistance gene for neomycin. This may allow for a selective process, so enhancing the success of transfer. There are two major limitations to mammalian cell transfection. The first is a much lower efficiency of gene expression in comparison with prokaryotic systems, with considerable differences between eukaryotic cell lines. Unlike rodent cells, most primate and human cells can integrate only a small amount of foreign DNA (about 6 kilobases); as a result only 10-30% of clones selected for the expression of one transcription unit will also contain a second unit in intact form. 13 The second problem is the short-lived response after successful transfection (a few months at most) regardless of the method used. 14 We know very little about the processing steps within the cell. Clearly, there are problems of degradation by
717
peripheral T lymphocytes from child with ADA deficiency (severe combined immunodeficiency, SCID)
TABLE II-DELIVERY SYSTEMS FOR GENE TRANSFER
Harvest
Physical Calcium-phosphate precipitation Electroporation Microinjection Protoplast fusion Liposomal transfer Receptor-mediated delivery Tissue injection
I
Virus mediated Retroviruses Adenovirus
Herpes simplex virus reach tumour cell targets in vivo. Major problems here include cell permeability, chemical stability, targeting, scale-up, and toxicity.15 Antisense oligodeoxynucleotides are small synthetic nucleotide sequences formulated to be complementary to specific DNA or RNA sequences. By the binding of these nucleotides to their targets the transcription or translation of a single gene can be selectively inhibited. If that gene is responsible for a disease process then its down-regulation could result in a reversal of the clinical abnormalities. The cytoplasmic location of mRNA provides an easier target for oligodeoxynucleotides than DNA. One example of an antisense oligomer with antiproliferative activity is c-myc in lymphoma cell lines;ib others are N-mye in neurectodermal cell lines;17 c-myb in colon adenocarcinoma,18 type 1 regulatory subunit of the cAMP receptor protein kinase in K-ras transformed NIH3T3 cells19 and in neuroblastoma, leukaemia, breast, colon, and gastric carcinoma cells, bcr-abl in chronic myeloid leukaemia blast cells,zo and c-raf 1 in ras and raf-transformed NIH3T3 ce11s.21In McManaway and colleagues’ in-vitro study, the antisense oligonucleotide was only 21 bases long.16 Growth of two Burkitt’s lymphoma cell lines known to produce an abnormal mRNA was strikingly inhibited, and this finding suggested an "Achilles heel" in the growth of this neoplasm. Antisense oligonucleotides also decrease the tumorigenicity or metastatic potential of cell lines. A K-ras proto-oncogene in antisense orientation transduced into a small cell lung cancer line inhibited of K-ras expression and also suppressed tumour growth in nude mice. There was no alteration in growth kinetics of the cell line in vitro. Other experiments have shown that the antisense sequence to pre-prourokinase significantly reduced the ability of murine melanoma cells to colonise lung. By contrast, if the antisense oligomer is designed to inhibit a putative metastasis suppressor gene, such as E-cadherin, the resulting downregulation renders the cells invasive.22 These results demonstrate the power of antisense technology and the potential application it has for manipulation of abnormal growth and behaviour in tumour cells.
to
.. Select for antibiotic resistance encoded by neo gene and for ADA gene function in vitro
Expand positive
clones and infuse back into patient
Fig 2-Application of gene transfer to
single gene disorder. Included hereto illustrate an application of gene therapy that isfurther advanced clinically than the cancer protocols reviewed here. a
extracellular nucleases, absorption onto and uptake into cells, transport from cytoplasm to nucleus, integration into host
chromosomes, mutation, the expression of non-integrated DNA, and transcription control of the transgene. Also, there is a small theoretical hazard of toxicity from the transfected DNA. It is calculated that the cumulative probability of having a harmful effect is less than 1 in 1016 per DNA molecule from a cell without activated proto-oncogenes or active viral oncogenes. Such a risk is likely to be acceptable with terminal cancer, but the evolution of safe, controllable gene delivery and expression systems is vital for successful therapy.
Biological approaches to cancer Recombinant drugs
Although drugs produced by recombinant DNA methods are not directly a form of gene therapy, they rely on the same technology. These agents include cytokines, growth factors and growth factor antagonists, humanised monoclonal antibodies, toxin-antibody immunoconjugates, toxin-ligand conjugates, and single-chain antigen-binding proteins. They have limited utility in cancer because the development of selectively toxic molecules has proved difficult. The selectivity of even the most specific of monoclonal antibodies with high tumour affinity is limited, and this almost certainly explains the poor clinical results so far. Improvements in molecular design and the identification of novel specific molecular targets may enhance their role.
Informational drugs Informational drugs are synthesised molecules that carry biological information which allows them to act in a specific manner. Oligonucleotides, ribozymes, and specific proteases act at different but specific levels of the
transcription system. These compounds can be of low molecular weight and therefore have considerable potential
Genetic immunomodulation
Cytokines can have a significant effect on tumour growth clinically but have systemic side-effects and very short half-lives. Many of the current prospects for gene therapy involve the vectoring of genes encoding for cytokines, with cells that home specifically on tumours. The anti-tumour immune response to the host can be improved by modifications of the immune system. Such modifications include the transfection of pleiotropic cytokine genes into the immune cells of the host or into tumour cells (eg, interleukin 2 or 4, interferon, tumour necrosis factor). The expression of cytokines in transfected cells can reduce their tumorigenicity and/or metastatic
718
TABLE III-SELECTIVE GENE EXPRESSION FOR POTENTIAL USE IN VIRALLY DIRECTED ENZYME PRODUCING THERAPY
GI=gastromtestinal,
SCLC= small cell
lung
cancer.
potential. The anti-tumour response can be enhanced by cytotoxic lymphocytes, macrophages, or antibodies which can in turn be induced by the expression of appropriate cytokine by tumour cells. In this way it may be possible to convert a weakly immunogenic tumour that elicits only a minor response to a strongly immunogenic one that could provoke a powerful destructive reactivity. Enhanced expression of major histocompatibility (MHC) molecules of class I and II can be achieved by transfected cytokine genes or by direct MHC gene transfection.23 The best characterised model so far has been the transfection of the tumour necrosis factor (TNF) gene into tumour-infiltrating lymphocytes. TNF has produced very encouraging antitumour responses in mice yet clinically it has proved disappointing, perhaps because toxicity limits the doses that can be used in man. Toxicity is common at doses above 8 Ilg/kg in man whereas in the mouse 400 ug/kg can be achieved. If a higher local concentration could be reached clinically, the response rate might increase. In a pilot study in patients with melanoma Rosenberg and colleagues24 demonstrated that tumour-infiltrating lymphocytes (TIL) transfected with a reporter gene (neomycin phosphotransferase) had localised in the tumour and continued to express foreign genes for up to ten months. A clinical trial with TNF-transfected cells is under way: TNF-transfected TIL cells will be infused into fifty patients with melanoma. It has previously been demonstrated that TIL cells produce responses in up to 30% of melanoma patients.zs The results from this clinical trial should be available some time this year. Another approach to genetic immunomodulation is vaccine development. Human papillomavirus (HPV), especially subtype 16, has been implicated in the aetiology of cervical cancer. Recombinant vaccines expressing HPV nuclear proteins E6 or E7 have already been demonstrated to reduce the development of HPV-related tumour in animals. Fibroblast-like cells transfected with HPV-16 E7 genes have been used to immunise syngeneic mice,26 and these cells confer protection against tumorigenicity of HPV-16 E7 positive tumour cells in these animals. Normal tissue protection The production of cytotoxic-drug-resistant stem cells may be a mechanism by which normal cells could be protected from treatment-related toxicity, allowing much larger doses to be used. The proposed transgenes can render stem cells resistant to some specific drugs—eg, methotrexate by use of dihydrofolate reductase (DHFR)27—or confer on them more widespread multidrug resistance (MDR) by the expression of the MDR1 gene.2$ An exciting clinical area is the infusion of haemopoietic
Fig 3-Virally directed
enzyme
prodrug therapy.
(a) Transcription factors in host cell results in production of viral enzyme from mtegrated vector, activating prodrug (b) Best example of selective toxicity is discrimination of normal liver and hepatoma cells by retroviral vectors containing AFP and albumin gene promoters. Normal liver cells are only destroyed when transfected with construct containing albumin promoter and hepatoma cells only by construct containing AFP promoter.
colony-stimulating factors (CSF) to support bone marrow function in cancer patients undergoing chemotherapy. These molecules include CSF for granulocytes, and macrophages, granulocyte-macrophage, and a relatively unchartered interleukins, erythropoietin, stem-cell factor. An alternative to infusion would be the genetic manipulation of normal bone marrow cells before chemotherapy. This might achieve a more continuous effect than the infusion of a drug with a short half-life. One study has demonstrated that the stable transfection of human G-CSF into fibroblasts in vitro produced neutrophilia on implantation into nude mice. The number ofhaemopoietic progenitor cells was increased in the spleen as well as in the bone marrow. The GM-CSF gene has also been transfected and expressed, showing increased tolerance to cytotoxic drugs.29 The efficiency of gene insertion into stem cells is low, however-10-20% in murine stem cells and only 1-5% in primates. Drug targeting The most promising selective genetic mechanism for drug targeting is virally directed enzyme prodrug therapy (VDEPT). This is based on a vector being expressed specifically in cells of a particular tissue or specifically in tumour
cells but
not
in normal cells. There
are
several
examples where tumours show tumour-specific transcription of certain host genes that are not essential for survival (table in). As our understanding of the mechanism of gene transcription in eukaryotes increases, it is likely that
719
TABLE IV-EXAMPLES OF SUPPRESSION OF MALIGNANT PHENOTYPE BY GENE TRANSFER IN VITRO
Fig 4-Strategy for VDEPT system tumour
to
produce 5-FU within
cells.
we will be able to
develop more and better VDEPT systems. The best current example of VDEPT working in vitro comes from the discrimination between normal liver and hepatoma cells by the differential expression of vectors containing promoters for albumin and ot-fetoprotein.1 In normal liver a herpes simplex thymidine kinase gene is expressed when coupled to the albumin promoter but not when coupled to the a-fetoprotein promoter (fig 3). The converse applies in hepatoma cells. Thymidine kinase can convert the prodrug 6-methoxypurine arabinonucleoside (Ara-M) which has minimum effects in normal cells, to the phosphates Ara AMP, ADP, and ATP, which are potent cytotoxic agents. In vivo studies of murine systems are being tested with this system. Another promising viral prodrug system is the conversion of the anti-fungal drug 5-fluorocytosine to the cytotoxic 5-fluorouracil (5FU) by cytosine deaminase driven by a selective promoter. One attraction of this system is that 5-fluorocytosine and 5-FU are in routine clinical use, and much pharmacological information is available for both. A strategy for a 5-FU VDEPT system is outlined in fig 4. There are several steps for potential selectivity. The first is based on selective infection by the virus. This may be achieved by targeting with immunoliposomes or a virus with predetermined tissue specificity. The second, and perhaps most important, selection is the expression of the vector only in the target cells that possess the appropriate transcription machinery. Although 5-FU may not be the most effective cytotoxic drug it is an excellent starting-point since so much is known of its pharmacokinetics. Gene replacement therapy Some human tumours overexpress dominant oncogenes fail to express normal tumour suppressor genes. Point mutations may also activate the transforming function of both oncogenes and suppressor genes.
or
There is now good evidence that in vitro reversion of malignant morphology can be achieved by a variety of different sequences (eg, normal H-ras, raplA [K-n], and c-jun).Other sequences can increase the immunogenicity of tumour cells making them more susceptible to the cellular antitumoural response of the host. Further genes can selectively inhibit the metastatic potential of cells, examples being nm23, &bgr;-actin, fibronectin receptor, connexin, and E-cadherin. The nm23 gene encodes a product that is probably involved in the control of ras-related proteins, and decreased
I
I
’
I
expression allelic deletion of this gene is associated with high metastatic potential in breast and colorectal cancers. Transfection of the nm23 gene with highly metastatic mouse melanoma cells reduced their metastatic potential and abrogated their responsiveness to the cytokine TGF-&bgr;.31 Finally, there are genes with recognised tumour suppressor properties such as RB and wild type p53 (table IV). The wild type p53 gene (WTp53) encodes a short half-life nuclear phosphoprotein with many properties consistent with tumour suppressor activity. WTp53 can promote cell differentiation and suppress the proliferation rate of transfected tumour cells. Expression of this gene can abolish tumorigenicity in vivo when such cells are placed in mude mice. The expression of WTp53 (but not mutant p53) inhibits proliferation of different human cell lines lacking p53 expression or expressing a mutated p53 allele. These include colorectal, breast, osteosarcoma, glioblastoma, and ovarian carcinoma cells.32 p53 mutations have now been implicated as a late event in tumour progression in many human cancers, and in the germline of at least some families with the Li-Fraumeni syndrome it may be possible to develop homologous recombination systems to insert functional p53 in place of the mutated gene. The loss or inactivation of both RB alleles has been observed in different types of tumour, apart from retinoblastoma itself--eg, breast cancer, small cell lung cancer, prostate, bladder, osteosarcoma, and other soft tissue tumours. RB gene transfection into RB negative or RB mutated cells from retinoblastoma, osteosarcoma, and prostate carcinoma lines can inhibit the malignant phenotype both in vitro and in vivo. The mechanisms by which this gene exerts its action are complex. There is evidence that pl05 Rb acts as a transcription factor binding to DNA. Mutation, deletion, or sequestration by adenovirus Ela, simian virus 40 T, and HpVH 16 E7 protein may result in a uncontrolled growth. The demonstration that the insertion of this gene in appropriate model systems can restore normal growth control provides an encouraging avenue for further development.33 Another intriguing molecule is GC factor (GCF). This is the product of a gene that encodes a transcription factor that downregulates the expression of certain genes associated with human cancer. In vitro assays have shown that this factor inhibits the transcription of epithelial growth factor receptor, p-actin, and calcium-dependent protease by interacting with their promoters. This is due to its affinity for specific GC-rich sequences found in these and other regulatory sites in DNA.34 The significance in or
720
cancer
suppression in vivo remains to be elucidated. It may well act system for studying downregulation of growth by similar
REFERENCES
as a model
mechanisms.
With gene replacement therapy, however, we lack suitable control technology for expression vectors, and it is unlikely that this strategy will have direct consequences in clinical care for some time.
Conclusion There have been remarkable developments in our understanding of the molecular basis of human cancer. This is likely to continue as we learn more about the human genome and about transcriptional control. The recognition of a system of positively acting growth signals and negatively acting growth inhibitory signals working in parallel will lead
rationally designed pharmacological agents active against cancer. Many hurdles, ethical and technical, remain, but selection criteria for gene therapy are being established’ and the ethical issues have been extensively reviewed.47-49 The development of secondary tumours and/or autoimmune disorders after gene therapy is a theoretical hazard, but the risk may be comparable to that which we now recognise for chemotherapy regimens. Gene therapy for cancer will require the cooperation of physicians, molecular biologists, and virologists to develop safe, robust systems for clinical use. The ethical issues will be resolved by rigorous review of protocols (in the US by the Recombinant DNA Advisory Committee in the UK by the Committee on the Ethics of Gene Therapy, until a more formal structure is devised) and by public education. The transition from theoretical possibility to practical reality may be only beginning but we could be at the dawn of a new age to new
of cancer treatment. We thank Dr Helen Hurst for helpful suggestions. A. A. G. has been partly supported by a grant from the Instituto Mexicano del Seguro Social, IMSS.
1. Mastrangelo MJ, Schultz S, Kane M, et al. Newer immunologic approaches to the treatment of patient with cancer. Semin Oncol 1988; 15: 589-94. 2. Smith KA. Interleukin 2: inception, impact and implications. Science 1988; 240: 1169-70. 3. Bishop JM. Molecular themes in oncogenesis. Cell 1991; 64: 235-48. 4. Borg A, Baldetrop B, Ferno M, et al. Erb-B2 amplification in breast cancer with a high rate of proliferation. Oncogene 1991; 6: 137-43. 5. Brodeur G, Seeger RC, Schwab M, et al. Amplification of N-myc in untreated neuroblastoma correlates with advanced disease stage. Science 1984; 24: 1121-24. 6. Slebos RT, Kiberlaan RE, Paledio O, et al. K-ras oncogene activation as a prognostic marker in adenocarcinoma of the lung. N Engl J Med 1990; 323: 561-65. 7. Weatherall DJ. Gene therapy in perspective. Nature 1991; 349: 255-76. 8. Benvenisty N, Reshef L. Director introduction of genes into rats and expression of the genes. Proc Natl Acad Sci USA 1986; 83: 275-76. 9. Ascadi G, Dicson G, Love DR, et al. Human dystrophin expression in mdx mice after intramuscular injection of DNA constructs. Nature 1991; 352: 815-18. 10. Geller AI, Keyomarsi K, Bryan J, Pardee AB. An efficient deletion mutant packaging system for defective herpes simplex virus vectors: potential applications to human gene therapy and neuronal physiology. Proc Natl Acad Sci USA 1990; 87: 8950-54. 11. Wu Ch, Wilson JM, Wu Gy. Targeting genes: delivery and persistent expression of a foreign gene driven by mammalian regulatory elements in vivo. J Biol Chem 1989; 264: 16985-87. 12. Kohn DB, Kantoff PW, Rglitis MA, McLachlin JR, Moen RC, et al. Retroviral mediated gene transfer into mammalian cells. Blood 1987; 13: 285-98. 13. Temin HM. Overview of biological effects of addition of DNA molecules to cells. J Med Virol 1990; 31: 13-17. 14. Johnson P, Gray D, Mowat M, Benchimol S. Expression of wild-type p53 is not compatible with continued growth of p53-negative tumor cells. Mol Cell Biol 1991; 11: 1-11. 15. Miller AD. Retrovirus packaging cells. Human-Gene-Ther 1990; 1: 5-14. 16. McManaway ME, Neckers LM, Loke SL, et al. Tumour specific inhibition of lymphoma growth by an antisense oligodeoxynucleotide. Lancet 1990; 335: 808-11. 17. Whitesell L, Rosolen A, Neckers LM. Episome-generated N-myc antisense RNA restricts the differentiation potential of primitive neuroectodermal cell lines. Mol Cell Biol 1991; 11: 1360-70. 18. Melani C, Rivoltini L, Parmiani G, Calabretta B, Colombo MP. Inhibition of proliferation by c-myb antisense oligodeoxynucleotides in
Glossary and abbreviations
’
Small Antisense oligodeoxynucleotides. to be formulated nucleotide sequences synthetic complementary to specific DNA or RNA sequences. They are targeted to inhibit selectively the transcription or translation of a gene. Cloning vector. A plasmid or bacteriophage used to "transport" inserted DNA for the purposes of producing more material or a protein product. Homologous recombination. Recombination between homologous sequences in a chromosome and allows direct replacement of the original DNA sequence with an exogenous segment. Long-terminal repeat. A promoter sequence directly repeated at both ends of retroviral DNA. It regulates the transcription of a gene. Oncogene. Gene where product has ability to transform eukaryotic cells to neoplastic phenotype. Promoter. A region of DNA involved in binding of RNA polymerase to initiate transcription of a gene. Reporter gene. A coding unit whose product is easily assayed (eg, neomycin resistance) in transfected cells, used to evaluate transfection efficiency. Retrovirus vectors. Viruses equipped to incorporate themselves into host DNA. These are "ecotropic" if the host range is limited to rodents or "amphotropic" are if the host range is wide, including man.
Retroviral shuttle vector. A construct with two hosts so that it can be
origins for replication for
used to carry a foreign sequence in prokaryotes or eukaryotes. Transcription. Synthesis of RNA from a DNA template. Transduction (signal). Transfer and processing of growth factor or other signals via cytoplasmic messengers to influence nuclear events. Transduction (gene). Acquisition and transfer of eukaryotic genetic material be retroviruses. Transfection. In eukaryotic cells this is the acquisition of new genetic material by incorporation of added DNA through physical or virus-dependent methods. AFP
ot-fetoprotein. CSF=co)ony-stimu!ating =
factors.
DHFR=dihydrofolate
LTR
=
long terminal
repeat.
MDR=multidrug resistance.
MHC= major histocompatibility. EGFR=epithelium derived growth factor receptor. RB retinoblastoma. 5- FU = 5-fluorouracil. TIL=tumour-infiltrating HPV= human lymphocytes. TNF=tumour necrosis papillomavirus. reductase.
=
IL=interleukin. LASN vector50 = vector carrying a human adenosine deaminase gene
factor.
VDEPT=virally directed enzyme prodrug therapy.
721
colon adenocarcinoma cell lines that express c-myb. Cancer Res 1991; 51: 2897-901.
33.
YS, Clair T, Tortora G, Yokozaki H, Pepi S. Suppression of malignancy targeting the intracellular signal transducing proteins of cAMP: the use of site selective cAMP analogs antisense strategy and gene transfer. Life Sci 1991; 48: 1123-32. 20. Szczylik C, Skorski T, Nicolaides NC, et al. Selective inhibition of BCR-ABL antisense cell leukemia proliferation by oligodeoxynucleotides. Science 1991; 253: 562-68. 21. Kolch W, Heidecker G, Lloyd P, Rapp UR. Raf-1 protein kinase is required for growth of induced NIH/313 cells. Nature 1991; 349: 19. Cho-Chung
426-28.
22. Vleminck K, Vakaet L, Mareel M, Fiers W, Van Roy F. Genetic of E-Cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 1991; 66: 107-19. 23. James RF, Edwards S, Hui KM, et al. The effect of Class II gene transfection on the tumorigenicity of the H-2K negative mouse leukemia cell line. Immunology 1991; 72: 213-18. 24.Rosenberg SA, Albersold P, Cornetta K, et al. Gene transfer into humans: immunotherapy of patients with advanced melanoma, using tumor-infiltrating lymphocytes modified by retroviral gene transduction. N Engl J Med 1990; 323: 570-78. 25. Aebersold P, Hyatt C, Johnson S, et al. Lysis of autologous melanoma cells by tumor-infiltrating lymphocytes: association with cLnical response. J Natl Cancer Inst 1991; 83: 932-37. 26. Chen L, Thomas EK, Hu SL, Hellstrom I, Hellstrom KE. Human papilloma virus type 16 nucleoprotein E7 is a tumour rejection antigen. Proc Natl Acad Sci USA 1991; 88: 110-14. 27. Corey CA, Desilva AD, Holland CA, Williams DA. Serial transplantation of methotrexate-resistance bone marrow: protection of murine recipients from drug toxicity by progeny of transduced stem cells. Blood 1990; 75: 337-43. 28. Galski H, Sullivan M, Willingham M, et al. Expression of a human multidrug resistance CDNA(MDRI) in the bone marrow of transgenic mice. Mol Cell Biol 1989; 9: 4357-63. 29. Keith WN, Brown R, Pragnell IB. Retrovirus mediated transfer and expression of GM-CSF in haematopoietic cells. Br J Cancer 1990; 62: 388-94. 30. Huber BE, Richards CA, Krenitsky TA. Retroviral-mediated gene therapy for the treatment of hepatocellular carcinoma: an innovative approach for cancer therapy. Proc Natl Acad Sci 1991; 88: 8039-43. 31. Leone A, Flatow U, King CR, et al. Reduced tumor incidence, metastatic potential and cytokine responsiveness of nm23-transfected melanoma cells. Cell 1991; 65: 25-35. 32. Mercer WE, Shields MT, Lin D, Appella E, Ulrich SJ. Growth suppression induced by wild-type p53 protein is accompanied by selective down-regulation of proliferating-cell nuclear antigen
manipulation
expression. Proc Natl Acad Sci USA 1991; 88: 1958-62. Huang HJS, Yee JK, Shew JY, et al. Suppression of the neoplastic phenotype by replacement of the RB gene in human cancer cells. Science 1988; 242: 1563-66.
Johnson A, Kageyama R, Popescu NC, Pastan I. Expression and chromosomal localisation of the human transcriptional repressor GCF. J Biol Chem (in press). 35. Weissman BE, Saxon PJ, Pasquale SR, et al. Introduction of a normal
34.
human chromosome 11 into
a
Wilms
tumour
cell line controls its
tumorigenic expression. Science 1987; 175: 181. 36. Declue JE, Zhang KE, Redford P, et al. Suppression of transformation by over-expression of full length GTPA activating protein (GAP) or of the GAP C terminus. Mol Cell Biol 1991; 11: 2819-25. 37. Wallich R, Bulbuc N, Hammerling J, et al. Abregation of metastatic properties of tumour cells by de novo expression of H2K antigens following H2 gene transfection. Nature 1985; 315: 301-05. 38. Sadano H, Taniguchi S, Baba T. Newly identified types of &bgr;actin reduces invasiveness of mouse p16 melanoma. FEBS Lett 1990; 271: 23-27. 39. Vleminck K, Vakaet L, Mareel M, et al. Genetic manipulation of E cadherin expression by epithelial tumour cells reveals an invasion suppressor role. Cell 1991; 66: 107-19. 40. Giancotti FG, Ruoslahti E. Elevated levels of the alpha5 beta 1 fibrinectin receptor suppress the transformed phenotype of Chinese Hamster ovary cells. Cell 1990; 60: 849-59. 41. Leone A, Flatow U, King CR, et al. Reduced tumour incidence metastatic potential and cyclokine responsiveness of nn23 transfected melanoma cells. Cell 1991; 65: 25-35. 42. Rollins B, Sunday M. Suppression of tumour formation in vivo by expression of the JE gene in malignant cells. Mol Cell Biol 1991; 11: 3125-31. 43. Kitayama H, Sugimoto Y, Matsuzaki T, Ikawa Y, Noda M. A ras-related gene with transformation suppressor activity. Cell 1989; 63: 77-84. 44. Friedman T, Roblin R. Gene therapy for human genetic disease? Science 1972; 175: 949-55. 45. Friedman T. Progress towards gene therapy. Science 1989; 244: 1275-80. 46. Anderson WF. Prospects for human gene therapy. Science 1984; 226: 401-09. 47. Motulsky AG. Impact of genetic manipulation on society and medicine. Science 1983; 219: 135-40. 48. Recommendations of European Medical Research Councils. Gene therapy in man. Lancet 1988; 1: 1271-72. 49. Murray TH. Ethical issues in human genome research. FASEB J 1991; 5: 55-60. 50. Culver KW, Osborne WRA, Miller AD, et al. Correction of ADA deficiency in human T lymphocytes using retroviral-mediated gene transfer. Transplantation Proc 1991; 23: 170-71.
STROKE OCTET Complications of acute stroke
Acute stroke is remarkable not for the damage it inflicts but for the recovery it allows. Nevertheless, both recovery and survival can be jeopardised by cerebral, systemic, and cardiac complications. The causes vary with time after the stroke (table I). Rigorous attention to general medical factors can attenuate some of these effects (table n).
Cerebral
complications
Overall stroke-related case-fatality is about 20%; the percentage ranges from 15% for supratentorial and brainstem infarcts to 58% for supratentorial haemorrhage.1 Case-fatality from infratentorial haemorrhage is 31 %.1
Transtentorial herniation and cerebral oedema
During the first week,
transtentorial herniation is the of death (table i); the incidence of this peaks within 24 hours for cerebral and at 4-5 days for cerebral infarction.1
commonest cause
complication haemorrhage
Brainstem compression with subsequent haemorrhage and infarction accounts for the serious morbidity and mortality associated with herniation. Herniation is the result of raised intracranial pressure caused by cerebral oedema, and is seen with large strokes.2 Postmortem studies show microscopic cerebral oedema in 93% of stroke patients;2 this high figure may reflect selection bias in necropsy-based investigations. In an antemortem study, 41 % of 22 patients showed computed tomographic (CT) evidence of a mass effect associated with oedema.2There was a good correlation between infarct size, mass effect, midline shift, neurological ADDRESSES: Cerebrovascular Program, Johns Hopkins Hospital and Johns Hopkins University School of Medicine, Baltimore, Maryland, USA (S Oppenheimer, FRCPC), and Department of Clinical Neurological Sciences, University of Western Ontario, London, Ontario Canada (Prof V. Hachinski, FRCPC). Correspondence to Dr Stephen Oppenheimer, Johns Hopkins Hospital, Meyer 5-181, 600 N Wolfe Street, Baltimore MD 21205, USA.
