CLINICAL REVIEW 131: Gene Therapy for Thyroid Cancer: Where Do We Stand?

Although most thyroid cancers are cured by surgery and 131I therapy, 10–20% of patients die from advanced differentiated and anaplastic tumors. A higher proportion of medullary thyroid cancers (MTC) has an adverse prognosis. Chemotherapy has been relatively ineffective to date, although there is always hope for the future. Thus, alternative approaches, such as gene therapy, are of interest, and thyroid tumors are an especially good target, for two reasons. First, certain gene promoters expressed in the tumors[ thyroglobulin (TG), TSH receptor (TSH-R), and calcitonin (CT)] have no or very limited expression elsewhere, and second, if the therapy leads to destruction of all normal thyroid tissue, this would be inconsequential in contrast, for instance, to destruction of the liver or lung. So the crucial questions are whether gene therapy can be effective, and whether it is safe. We believe that the answer to both questions is yes. Many methods are available, including 1) direct injection of an expressible DNA into tumor, 2) blocking expression of an oncogene by expression of antisense messenger ribonucleic acid, 3) transfection of biopsied host tumor cells in vitro with DNA and reinjection, 4) transduction of cells in vitro with a retrovirus expressing the desired gene and reinjection to the host, and 5) infection in vitro or in vivo of tumor cells using a nonreplicative (or replicative) adenovirus to reexpress a gene lost in dedifferentiation such as p-53, p-10, or sodium/iodide symporter (1). A currently common strategy is to express herpes simplex virus (HSV) thymidine kinase (tk) in the tumor, and then administer ganciclovir (GCV), which is converted by tk into a cell-lethal metabolite. A second common strategy is to use a virus to transduce expression in the tumor of a cytokine that augments antitumor immunity. Interferon-γ, interleukin-2 (IL-2), IL-12, granulocyte-macrophage colony-stimulating factor (GM-CSF), and lymphotactin are among the genes employed. Other strategies target induction of apoptosis or antiangiogenesis factors. Before reviewing some of the current preclinical and clinical studies, it may be appropriate to consider the problems associated with gene therapy. In general, adenovirus must be directly injected into tumor if a cytotoxic gene is employed, because when given iv, most of the virus goes to the liver and can in high dosage be fatal (in animals). When injected into a tumor mass, only a small proportion of cells is infected. Thus, it is unlikely that a single injection of a suicide gene can kill all of a tumor. However, expression of a gene that augments immunity to the tumor might lead to rejection of the entire tumor and distant tumor as well. Another benefit of the use of genes that augment immunity is that their effect may persist for months or longer. Virally mediated expression of genes in nontarget organs after iv administration or leakage from a tumor can be a serious problem. This can be overcome by using cell-specific promoters such as the CT or TG promoters, which have very highly regulated expression but are often weaker than nonspecific promoters, such as the commonly employed cytomegalovirus (CMV) promoter. This is the method we have employed with the TG and CT promoters for treatment of thyroid and medullary thyroid tumors. Similarly, a prostate-specific antigen (PSA)-specific promoter has been used to drive gene expression in PSA-positive prostate cancer cells (2). Targeting virus specifically to tumor would be very desirable and has been partially accomplished in some systems. Targeting can be achieved to some extent by employing an agent bound to the virus to direct it to a specific receptor on the tumor cell surface. For example, an antibody can be bound to the viral rod protein, and this antibody can be coupled with a tumor-targeting molecule, such as a growth factor or second antibody (3). In another strategy, a phage library displaying linear dodecapeptides was used for biopanning differentiated ciliated airway epithelial cells. After repeated rounds, a peptide with effective binding was recovered, and this peptide was coupled to the surface of an adenovirus using polyethylene glycol. The chemically modified vector could affect gene transfer to human airway epithelial cells (4). Many approaches to this are possible, and in the thyroid field one obvious target is the TSH receptor. Several laboratories are currently investigating targeting vectors to thyroid tumor cells bearing the TSH receptor. Another problem involves immunity to the vector, such as adenovirus. Most adults are immune to adenovirus, the commonly used vector. However, it appears that this may prevent iv distribution of the virus (actually a benefit), but only somewhat reduces, rather than obliterating, expression in the tumor after direct injection of the vector. Bramson et al. (5) studied the effect of injecting IL-12-expressing adenovirus in animals with preexisting immunity. In naive animals, virus was found disseminating into liver 72 h after injection. Previously, immunized animals had much less expression of the virus in liver because of neutralization by serum antibodies. Expression of the transgene after direct injection of tumor was approximately 50% reduced in previously immunized animals compared with that in nonimmune mice. The data suggest that immunity could actually be advantageous in that it is not a complete block to gene transfer, but does diminish viral dissemination. A last problem for thyroid researchers has been the lack of good animal tumor models for differentiated thyroid cancer. Fortunately, the opposite situation exists for studies on MTC, as cell lines producing tumors in both rats and mice are available. Many other new vectors are under development, including vectors based on adeno-associated virus and canarypox (6). Amalfitano et al. (7) constructed a new adenoviral vector with deletions of early 1a region (E1a), E1b, and E3 genes. They note that use of high titers of E1-deleted vectors can result in the production of replication-competent adenovirus derived by recombination with the E1 sequences present in 293 cells used for virus production. The more extensively deleted vectors replicate well and are thought to have much less potential for generation of recombinant-competent virus. As noted, studies of differentiated thyroid cancer treatment have been hampered by the lack of good animal models. Nevertheless, work goes on at a rapid rate in this field. With differentiated thyroid cancer cells it might be possible to direct treatment that would be entirely specific for the tumor cells or residual normal cells using a promoter expressed only in thyroid tissue. Following this approach, Braiden et al. (8) developed a retrovirus expressing HSV-tk under control of the TG promoter. In rat FRTL-5 cells, transduction by the virus plus GCV treatment caused cell death. However, in FRO cells, a human anaplastic thyroid cancer cell line that does not express TG, the viral/GCV treatment had (appropriately) no effect. When an in vitro transduced FRTC cell line (a malignant cell line developed from FRTL-5 cells), was used to produce tumors in severe combined immune deficiency mice, the tumors were rejected during GCV treatment. Nagayama et al. (9) have enhanced the function of the tk/GCV system by combining it with the Cre-LoxP system to enhance the expression of tk. They used the TG promoter to target expression. Zhang et al. developed an adenoviral vector transducing tissue-specific expression of genes under control of the rat TG promoter (10). Using this promoter to express luciferase activity, infected FRTL-5 (thyroid) cells produced luciferase activity, whereas HEPG2, COS-1, rat and human MTC, HeLa, GH3, P98G, and CA77 cells did not, demonstrating a high degree of specificity of TG promoter function. When the adenovirus with TG promoter driving expression of tk (Ad-TG-tk) was used to transduce FRTL-5 cells, and GCV was added, more than 90% of the cells were killed, and a minimal effect was observed in the other cell lines. When a virus using the nonspecific CMV promoter to express tk (Ad-CMV-tk) was given iv to rats, and animals were treated with GCV, severe liver damage was induced using 1 × 109 to 1 × 1010 plaque-forming units (pfu) of the virus. In striking contrast, when the same amount of adenovirus expressing tk under control of the TG promoter was given iv, no adverse effect on the animals was detected, SGOT levels were equal to those in normal controls, and the liver, lungs, and kidneys showed no histological abnormality. Thus, use of the TG promoter to express a suicide gene (or a cytokine) appears to have a very beneficial effect to lower toxicity. Clearly this vector has to be given by direct injection into the tumor, but a virus expressing a cytokine can be expected to have an effect on tumors at a distance as well as the injected site. Another strategy, employed by Moretti et al. and Nagayama et al. (11, 12) in a preclinical study, is already being tested in many clinics in phase I or II studies in patients with metastatic tumors. This treatment takes advantage of the fact that most poorly differentiated tumors have lost expression of normal p53 tumor suppressor gene as part of the process of becoming unregulated. A replicating adenovirus (or a nonreplicative form) is used to infect the tumor cells, and expression of the p53 protein leads to apoptosis of tumor cells. Apparently, expression of the gene in normal cells, which already have an intact and expressed p53, causes less, or in some cases no, harm. Nagayama et al. used a replication-defective adenovirus to reexpress p53 in four human anaplastic thyroid cell lines that do not normally express the gene. The tumor cells were destroyed in vitro and partially in vivo in nude mice, whereas cultured normal thyroid cells were more resistant. They also found that the treatment induced apoptosis of the tumor cells and sensitized two cell lines to the chemotherapeutic effect of doxorubicin both in vitro and in vivo. This approach clearly offers potential for the therapy of anaplastic cancers, with the caveat that it may be difficult to transduce all cells by direct tumor injection. Nagayama et al. elaborated on this system during a recent presentation at the 12th International Thyroid Conference (13). To develop an adenovirus system that can replicate exclusively in wt-p53-deficient cells, they have infected the cells with two recombinant viruses. One has an expression unit of the synthetic p53-responsive promoter and the cAMP response element (Cre) recombinase gene. The other has a constitutive CA promoter and E1A gene flanked by a pair of loxP elements, followed by E1B19K gene under control of the CMV promoter. When the viruses are coinfected into cells with wild-type p53 expression, Cre would be activated, excise out the E1A gene, and switch off E1A expression, so that the second virus would not replicate. However, in cells with a mutant p53, the p53 response element would not be activated, E1A would be expressed and lead to virus replication and cell lysis. They use this system to show that the double infection caused viral replication and a cytopathic effect only in cells not expressing wild-type p53. Scala et al. developed an adenovirus carrying the HMGI(Y) gene in an antisense orientation. This virus induced cell death in two human thyroid anaplastic carcinoma cell lines (ARO and FB-1), but not in normal thyroid cells, and also killed a variety of other tumor cell lines (14). Although it is logical to use a thyroid-specific promoter such as TG in directing gene therapy to thyroid cancer cells, clearly some thyroid tumors lose expression of TG as they progressively undifferentiate. Chun et al. and Zeiger transfected non-TG-expressing ARO and WRO thyroid cell lines, which had low intrinsic TG, thyroid transcription factor-1, and PAX-8 gene expression. Transfection of these cells with the two genes known to be involved in thyroid cell differentiation, thyroid transcription factor-1 and PAX-8, resulted in increased function of the TG promoter. This is a model that might be important in gene therapy, because it suggests a way to reintroduce expression of a specific gene in thyroid cells (15). Progress with animal models of MTC has been more rapid. Soler et al. studied the effect of the tk/GCV system and IL-2 secretion in the Wag/Rij rat MTC model (16). These researchers used a retrovirus to transduce expression of the desired genes in vitro, and then injected the cells into the host. Presumably, such cells are nearly totally transduced by the virus during in vitro infection, in contrast to the less effective transduction when the virus is injected into the tumor in vivo. When the animals were treated with GCV, tumors were eradicated in 60% of animals. In subsequent studies, cells that had been transduced with the tk-producing retrovirus, and cells that had been transduced with an IL-2-producing virus were mixed and inoculated. The combination of the suicide gene and IL-2 gene produced superior tumor inhibition. This system is obviously complicated, as it requires in vitro transduction of the tumor cells by the retrovirus and subsequent reinjection. Lausson et al. (17) also reported effective control of this model tumor using a vector producing IL-2. Zhang and co-workers studied adenovirally mediated treatment of a rat MTC model tumor in Wag/Rij rats. An adenovirus expressing the viral tk gene under control of the CMV promoter was used for treatment. Adenovirus (2 × 109 pfu) was injected into MTC tumors of less than 100 mm3 volume, and animals were treated with GCV ip 24 h later at a dose of 100 mg/kg/day for 7 consecutive days. In these small tumors, the viral/GCV treatment caused destruction or stabilization of the tumor. However, in larger tumors, averaging 400 mm3, the treatment was not effective (18). These tumors allow little bystander killing of neighboring cells (18). To get around this problem, the researchers developed a replication-defective adenovirus that produced the mouse IL-2 (mIL-2) gene under control of the CMV promoter. The studies were carried out in a mMTC tumor system in syngeneic BALB/c mice. Tumors less than 30 mm3 were destroyed after injection of 1 × 109 pfu of virus, and larger tumors were stabilized (19). Mice that were cured of their tumor by treatment with the Ad-CMV-mIL2 virus did not develop tumors after reinjection of wild-type MTC cells, indicating that long-term immunity developed. Histological and immunological studies revealed massive CD4+ and CD8+ T cell infiltration in the tumors. Thus, IL-2 was shown to induce effective antitumor immunity in this model system. Similar responses were found using a rat medullary tumor line (20). To study dissemination of the virus and toxicity, adenovirus expressing lac-z under control of the CMV promoter was injected into tumors at a dose of 1 × 109 pfu. Functional virus was detected by β-galactosidase activity in the tumor during the subsequent 5 days and in the liver on the fifth and seventh days after but not in or is known to be the at which viral in the are However, after iv administration of of × 109 to 1 × 1010 pfu of the vector, there was no of SGOT or Thus, administration of appears to be the minimal of liver toxicity, after administration of 1 × 1010 pfu directly iv, which is to lead to of more than of the virus in the liver (20). As both the tk/GCV system and the virus have antitumor it is logical to the two The tk/GCV suicide gene treatment is to by killing transfected cells, killing neighboring cells the bystander and to lead to the of antitumor immunity of tumor The cytokine treatment expression of to targeting of cytotoxic T cells and cells to the of the tumor and to of these cells the tumor. To express both genes in one vector, both Soler et al. and Zhang and developed an adenoviral vector tk and human IL-2 under control of CMV promoter in two expression the same virus. The genes were into the E1 region of adenovirus MTC cells infected with the virus showed high in vitro to ganciclovir and produced of of treated tumors were destroyed when this vector was injected into tumors. This an the function of the or the virus used Many have been tested for antitumor activity to These tumor IL-12, and is a cytokine that causes of cells and CD8+ T cells. also and of naive CD4+ T cells into cells. administration of recombinant has been shown to tumor growth and produce an antitumor but causes severe toxicity, in animals and et al. studied the use of a vector producing and showed in a tumor model and another used gene transfection of metastatic tumors with and this resulted in of tumors Zhang and developed an that has both the and of expressed under the control of CMV promoters in in the E1 region of adenovirus injection of the vector into tumors using 1 × 109 pfu eradicated most tumors and produced long-term immunity to with naive MTC tumor cells at after the This viral system to be the most effective system to date, the and vector systems. it the for use in human studies et al. used a gene to of expressing IL-12, and They studied the of the directly injected DNA to induce the production of and to of in mice. A of the produced the most antitumor et al. directly injected an adenoviral vector both of into tumors in mice. A nonreplicative and virus with a CMV promoter driving the expression was used in this A dose of 1 × 109 pfu of the virus was of mice given the specific vector had tumor and rejected their tumors. of the mice that rejected tumor were with tumor at a later and rejected the tumors. et al. used the virus in a metastatic tumor and the virus These animals were given 5 × on days. They found a of T cells, and around the tumor. The virus was also effective in animals of CD4+ or CD8+ cells and in severe combined immune deficiency mice, that the is mediated by cells, including and and could The of virus expressing was found to enhance the antitumor activity of viruses expressing tk and in a study by et al. expression of in the tumor model and expression antitumor immunity and animal The of these studies is that vectors expressing are highly effective in antitumor immunity. The is not clearly and may in to of and immunity does not prevent the effect of direct treatment with adenovirus. In these preclinical studies in cell lines in vitro and animal models clearly the of the tk/GCV suicide gene strategy and the of gene expression to tumor destruction in The use of cell-specific promoters toxicity, although requires direct tumor injection. of virus using a cell-specific promoter into the appears to have little adverse effect. To there have been no of gene therapy for human thyroid but there are many studies other tumors. et al. a phase I of gene therapy in patients with A adenoviral vector the HSV-tk gene under control of the virus was into the of patients with malignant followed by of therapy with GCV at a dose of 5 were minimal and liver and as well as a response in those the and immune responses were This study that administration of an adenoviral vector the HSV-tk gene is well and in gene transfer when at high et al. used an system in which cells were transduced by retrovirus in vitro and injected into the after of surgery for residual tumor cells always lead to After a transduction GCV was for days. patients with were treated serious adverse of the patients than months after of patients had no One was of and after treatment. Thus, of vector cells producing tk were well and were associated with et al. a phase 1 study using an and adenovirus IL-2 directly injected into of and cancer. was observed at the of injection in 60% of but were tumor at the of injection in of but no clinical responses were CD4+ and CD8+ T cell 24 h after injection. IL-2 was by in tumor sequences were detected or after indicating the of replication-competent virus. and antibody titers were 1 after injection in all This the of use of adenoviral vectors for gene in and transgene expression in the of preexisting immunity to adenovirus. et al. used a recombinant retrovirus to transduce expression of IL-2 in of cells. These tumor cells were then used to the patients were or with approximately tumor cells. was well with at of 24 injection and of and after One developed antitumor after the and showed an increased response after the second patients had for and 1 had not after Thus, with tumor cells is and safe. and cytotoxic T can be induced in some patients with such a et al. also studied direct in gene therapy for of the prostate using a adenovirus expressing tk and administration of of of virus into the prostate under GCV was then given iv for days patients 1 × to 1 × of and were for growth of adenovirus. One at the dose developed and patients achieved an by a in serum levels by 50% or that was for to 1 are to the distribution of vector the prostate and to the of of gene therapy. In an important study, et al. the treatment of patients of gene therapy using an tk by direct injection into prostate cancer. increased from to in patients from of therapy, but all were and was for patients studied a of in in of the that direct injection of followed by iv GCV, was and effective, in A problem gene therapy is to the virus into from the direct injection into a tumor has been studied for the treatment of cell cancer by et al. In a clinical phase I patients with cell cancer and a of the tumor suppressor gene p53 were treated by gene therapy. of a vector adenovirus expressing wild-type p53 were injected under The gene therapy was in all patients no adverse of gene therapy were After four of the patients showed at the treated tumor whereas other tumor are an to for gene therapy. Clearly gene therapy can be and to human tumors by injection into or tumor or by direct tumor injection. The dosage of virus in these caused at most We can in the of some of the methods in preclinical to human thyroid tumors that are control by therapy.