| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Histology, Microbiology, and Medical Biotechnologies (L.B., G.P.), University of Padova, I-35121 Padova, Italy; and Department of Internal Medicine, Division of Endocrinology (M.B.), University of Ancona, 60100 Ancona, Italy
Correspondence: Address all correspondence and requests for reprints to: Giorgio Palù, M.D., Department of Histology, Microbiology and Medical Biotechnologies, University of Padova, via Gabelli 63, I-35121 Padova, Italy. E-mail: giorgio.palu{at}unipd.it
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. How to Exploit...III. Endocrine Cell-Specific...IV. Endocrine Side Effects...V. Oncolytic Virus Infection...VI. Endocrine Response to...VII. Risk of Germ...VIII. SummaryReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. How to Exploit...III. Endocrine Cell-Specific...IV. Endocrine Side Effects...V. Oncolytic Virus Infection...VI. Endocrine Response to...VII. Risk of Germ...VIII. SummaryReferences |
|---|
After a brief introduction on efficacy and toxicity of gene therapy in clinical trials for the treatment of cancer and the strategies developed so far to target endocrine and endocrine-related tumors, this review will focus on the endocrine aspects of cancer gene therapy. These include the possibility to exploit molecular mechanisms of regulation of hormone activity to control therapeutic gene expression, the use of endocrine cell-specific genes as therapeutic tools, the potential side effects of cancer gene therapy on the endocrine system, the neuroendocrine response to vector delivery, the effects of ectopic expression of cytokines as therapeutic genes, and side effects related to hormone or growth factor inhibition. The risk of germ line cell transduction and present ethical concerns will be also addressed.
A. Efficacy and toxicity of gene therapy
1. Overview.
The concept of cancer gene therapy derives from new understandings of the molecular biology of cancer and the complex interactions between tumor cells and the immune system. This knowledge has been exploited to develop strategies to selectively target tumor cells or to stimulate the immune response against tumor antigens. Current therapeutic approaches and results from clinical trials of cancer gene therapy, which have been reviewed recently (1, 2), are summarized in Fig. 1
and Table 1. It is apparent that, since the development of the first cancer gene therapy clinical trials in the early 1990s, clinical results are still largely unsatisfactory, notwithstanding efforts to improve gene transfer tools and therapeutic genes.
|
|
The first event was the tragic death, in September 1999, of an 18-yr-old patient with ornithine transcarbamylase (OTC) deficiency enrolled in a clinical trial at the University of Pennsylvania (4). This event focused the attention of the scientific world on the potential risks associated with gene transfer and the need to accurately record and report serious adverse events in gene therapy trials. In the Penn study, patients with partial OTC deficiency were treated via the hepatic artery with escalating doses of a nonreplicating adenovirus carrying the gene encoding for OTC (5). Intravascular administration of the vector resulted, in some patients, in some of the clinical symptoms also reported in other studies using adenoviral vectors, i.e., transient fever, myalgias, elevation in liver enzyme and cytokine levels, reversible hypophosphatemia, thrombocytopenia, and anemia (6). However, in a male patient enrolled in the highest dose cohort who received approximately 4 x 1013 viral particles, the initial mild symptoms progressed to acute respiratory distress syndrome and subsequent multiorgan failure, resulting in the patient’s death 4 d after treatment. A report from the National Institutes of Health Recombinant Advisory Committee (7) attributed the death to a systemic, adenovirus vector-induced shock syndrome, due to a cytokine cascade that led to disseminated intravascular coagulation, acute respiratory distress, and multiorgan failure. Moreover, postmortem examination revealed bone marrow red cell aplasia. It was suggested that the high dose of adenoviral vector quickly saturated available receptors for the vector and then spilled to other organ systems, including the bone marrow, thus inducing a systemic immune response. Adenoviral vector capsid proteins likely contributed to the patient’s immune response (8, 9).
The positive events were the first conclusive evidence that gene therapy can be successful in humans. These regarded a clinical protocol in patients with hemophilia based on im injection of recombinant adeno-associated viral (AAV) vectors (10) and a gene therapy clinical trial in children with X-linked severe combined immunodeficiency (X-SCID) via retrovirus transduction of hematopoietic stem cells (11, 12).
In the first study, after preclinical evidence of correction of the hemophilic phenotype in animal models (13, 14), efficient transduction and expression of factor IX was achieved in three hemophilia B patients receiving im injection of an AAV vector encoding blood coagulation factor IX. The procedure was safe, and a mild improvement of clinical condition was achieved (10). In the second study, four of five children with X-SCID due to a deficiency of the 
c chain, who were treated with autologous bone marrow stem cells and transduced ex vivo with the c gene, showed evidence of long-term correction of the immune deficiency (11, 12). However, the great optimism for the potential clinical benefits of gene transfer generated by these results was frozen by the announcement of serious adverse events observed in the clinical trial of gene therapy for X-SCID, which led the U.S. Food and Drug Administration and the American Society of Gene Therapy to the decision to put a clinical hold on similar trials in the same disease in the United States (15). Two of 11 boys treated so far in this clinical trial developed T cell leukemia-like illness about 3 yr after the gene therapy procedure, one in September 2002 and the other in December 2002. The first case involved the replication of a single clone of 
T cells, and the second involved an excess of three clones of 
ß T cells. In both cases, leukemia was probably a consequence of insertional oncogenesis, because the vector inserted itself within or near the same gene, LMO-2, which has been linked to leukemia T cells (16, 17, 18, 19, 20). However, because the complication of leukemia has not occurred in any other clinical trial (20), multiple factors could have contributed to the development of leukemia in the patients involved in this trial. These include the high level of engraftment and expansion of the genetically modified cells, unique properties of the hematopoietic stem and progenitor cells in bone marrow of X-SCID patients, the immune deficiency of the X-SCID patients, and/or the transferred gene itself (19).
2. Safety in cancer gene therapy clinical trials.
Although significant therapeutic benefit has not yet been demonstrated from cancer gene therapy clinical trials, a positive remark for gene therapy derives from the substantial safety of this approach in cancer patients treated so far (Table 2).
|
Different side effects are observed by using different types of vector, and toxicity is related to vector dose and site of administration. Fever and chills are common side effects observed after intratumoral administration of high doses of adenoviral vectors (31, 32, 33, 34, 35); subarachnoidal hemorrhage and aseptic vasculitis may occur after intracerebral inoculation of retroviral vector-producing packaging cells for recurrent glioblastoma multiforme (36); confusion, hyponatremia, seizure, and signs of central nervous system toxicity occurred in patients with end-stage malignant brain tumor injected with adenoviral vectors (37, 38); and mild transient flu-like symptoms and local inflammation with pustule formation occurred after intratumoral injection of high doses of recombinant vaccinia virus vectors (39, 40). Administration of large doses of plasmid DNA appeared to be well tolerated without evidence of the development of anti-DNA antibodies (41). Mild to moderate toxicity was reported after direct intralesional delivery of Allovectin-7, an HLA-B7/ß2-microblobulin DNA-liposome complex, in patients with metastatic melanoma (23). Symptoms included pruritus and erythema at the injection site and general aches and pains (23).
With regard to adenoviral vectors, a revision of safety parameters and long-term follow-up in 102 subjects receiving local administration (i.e., to the nasal and bronchial epithelium, metastatic tumors, skin, myocardium, and skeletal muscle) of low (<109 particle units) and intermediate (109 to 1011 particle units) doses of replication-deficient adenoviral vectors demonstrated an incidence of 0.7% major adverse events, but no deaths related to an adenoviral vector (35). The incidence of malignancy was within that expected for the population. Most adverse events associated with administration of the vector to the respiratory epithelium were fever and/or leukocytosis associated with the bronchoscopy procedure, and, occasionally, increase in fibrinogen and liver function tests. Vector administration to colon cancer metastatic to the liver was sometimes accompanied by mild transient increase of transaminases, mild fever, leukopenia, and thrombocytopenia (35, 42). Mild transient increase in transaminases and transient hypotension were reported also after intratumor adenoviral vector administration in patients with mesothelioma (43, 44) or head and neck cancer (32). Analysis of risk factors for adverse events in patients treated with local delivery of adenoviral vectors demonstrated that vector-related parameters, including dose, route, transgene, or number of vector administrations, did not predict the occurrence of major adverse events (34).
Replication-competent viral vectors may show side effects related to local viral replication. Most clinical experience with replicating viruses has been achieved with the E1B-deleted adenovirus dl1520. Lacking E1B activity, this vector should selectively replicate, and thereby kill, in cells with unpaired p53 function (45), although replication in cells with wild-type p53 has been demonstrated (46, 47, 48). Several clinical protocols with this agent have been completed so far, including trials in patients with head and neck cancer (49), pancreatic adenocarcinoma (50), hepatocellular carcinoma (51, 52), ovarian cancer (30), and gastrointestinal malignancies metastatic to the liver (51, 53, 54, 55). Phase I studies were based on intratumoral injection of a single dose of dl1520 ranging from 107 to 1011 plaque-forming units (pfu). All of these showed that escalation of virus dose was possible without dose-limiting toxicity. Most patients reported flu-like symptoms after dl1520 administration. Symptoms generally started after the first virus dose and consisted of generalized malaise, headaches, nausea, myalgias, pyrexias, and rhinorrhea. Pain at injection site was reported in patients with head and neck cancer receiving intratumoral application of up to 1011 viral particles (49). Doses of dl1520 ranging from 2 x 108 to 2 x 1012 pfu administered through the hepatic artery were well tolerated, with only transient fever and elevation of liver enzymes (51, 55). Dose-escalation was possible, and the maximum dose, which was based on manufacturing capabilities, was shown to be well tolerated in treated patients (55). Intravenous administration of up to 2 x 1012 dl1520 viral particles was also well tolerated, except for mild to moderate constitutional symptoms and transient dose-dependent increase of serum aminotransferase (51, 54). Intraperitoneal injection of 109 to 1011 pfu for ovarian cancer was associated with abdominal pain, consistent with peritonism, diarrhea, heartburn, and vomiting (30). Interestingly, no cytopathic effects suggesting viral replication have been observed in normal tissues surrounding injected tumor tissues (55).
Other replicating viruses recently introduced to the clinic include the conditionally replicating adenovirus CV706, in which the adenoviral E1A gene is driven by prostate-specific antigen (PSA) promoter/enhancer elements and, therefore, it can selectively replicate in prostate tissue (56). A recently completed phase I/II trial in patients with locally recurrent prostate cancer showed that doses of up to 1013 particles of the CV706 vector, administered using brachytherapy techniques, appear to be safe, although biochemical (PSA) responses were observed in a minority of patients (57). A more potent oncolytic adenovirus, CV787, which contains the prostate-specific rat probasin promoter driving E1A expression and the human prostate-specific enhancer/promoter driving the E1B gene (58), is currently being studied in phase I and II clinical trials. This virus, unlike CV706, maintains a wild-type E3 region, which encodes proteins that play a role in cell lysis and evasion of host immune response (59). Other oncolytic agents currently under evaluation in phase I and II clinical trials include replication competent G207, HSV1716, and NV1020 herpes simplex viruses (HSV). The double mutant G207 HSV harbors deletions of both copies of the 34.5 gene and contains an insertional inactivation of the ICP6 gene, which encodes a subunit of viral ribonucleotide reductase. The virus has been administered stereotactically in patients with recurrent gliomas at doses up to 3 x 109 pfu without any significant toxicity being encountered (60). In particular, no patient has developed HSV encephalitis (60). The HSV1716 virus, which lacks both copies of 34.5 has been safely administered intratumorally up to doses of 105 pfu in patients with recurrent high-grade glioma (61, 62) or metastatic melanoma (63). NV1020 is currently being evaluated in a phase I trial as a vaccine in the treatment of patients with colorectal carcinoma liver metastases. This virus has only one copy of 34.5 deleted and maintains sensitivity to acyclovir and ganciclovir (GCV). In addition, it contains an exogenous copy of the thymidine kinase gene under control of the powerful HSV-1 4 promoter. Genetic stability and safety have been demonstrated in extensive rodent and primate studies as well as in limited human vaccine trials (64, 65). Other replicating viruses currently under evaluation in phase I/II clinical trials include reovirus, a virus that replicates in malignant cells with activation of the Ras signaling pathway (66), the animal pathogen Newcastle disease virus (63, 67), vaccinia virus (39, 69), and autonomous parvoviruses (70, 71). Preliminary results from clinical trials are encouraging, and no serious adverse events have been demonstrated so far. The most common adverse events were flu-like symptoms occurring principally after administration of the first dose of the Newcastle disease virus PV701 (68) or after intratumoral injection of vaccinia/granulocyte macrophage-colony stimulating factor (GM-CSF) recombinant virus (39).
Side effects directly related to therapeutic transgenes are less frequent, such as in the case of mild flu-like symptoms observed after administration of IL-2 cDNA in liposome complex (21), or in the case of fever, fatigue, or change in mental status in patients receiving intratumor injection of the E1A adenovirus gene as a lipid complex (24). Deep vein thrombosis was reported in patients with malignant astrocytoma, who underwent im implantation of autologous glioma cells, treated ex vivo with an antisense oligodeoxynucleotide directed against the IGF type I receptor (IGF-IR) (72). Side effects due to the prodrug used in suicide gene therapy also have been reported, i.e., rise of liver enzymes after GCV administration (43).
3. Efficacy in cancer gene therapy clinical trials.
Despite anecdotal reports of therapeutic responses in several patients, unequivocal proof of the clinical efficacy of cancer gene therapy is still lacking (Tables 1
and 2).
Of the different approaches to cancer gene therapy, including immunotherapy, tumor suppressor gene replacement, and suicide gene/prodrug activation therapy, immunotherapy showed better clinical results, being less affected by the limitations related to vector titer and transduction efficiency. Partial responses were observed after IL-2 (21, 26), HLA-B7 (23, 73), IL-7 (74), or GM-CSF (39) gene transfer in patients with advanced solid tumors, including renal cell carcinoma, melanoma, and soft-tissue sarcomas. At variance, no evidence of tumor response was seen at sites distal from the injected tumor in a phase I trial of interferon- (IFN-) retroviral vector administered intratumorally to patients with metastatic melanoma (75).
Clinical and radiological improvements were observed in patients with malignant astrocytoma, after ex vivo treatment of autologous tumor cells with an antisense oligodeoxynucleotide directed against the IGF-IR (72). Minor tumor responses also were demonstrated in two of 16 evaluable patients with recurrent breast and head and neck cancer, receiving intratumoral liposome E1A gene therapy (24). In this strategy, the E1A adenovirus gene functions as a tumor suppressor gene by inhibiting expression of HER-2/neu and other oncogenes, inducing apoptosis in cancer cells and sensitizing cancer cells to chemotherapeutic drugs (76, 77).
Using tumor suppressor gene-replacement approaches, transient local disease control and partial tumor responses were observed after viral vector-mediated delivery of wild-type TP53 in phase I and pilot studies in patients with lung cancer (27, 78, 79), head and neck cancer (32), bladder cancer (25), or metastatic malignant liver tumor (51). However, a controlled phase II study in patients with newly diagnosed advanced non-small-cell lung cancer failed to demonstrate a significant clinical benefit from local TP53 gene transfer by intratumoral vector injection in combination with effective first-line chemotherapy (31). A phase I clinical trial in end-stage ovarian cancer patients treated with ip administration of retroviral vectors expressing the BRCA1 tumor suppressor gene reported tumor reduction in three of 12 treated patients, vector stability, and minimal antibody response (29). At variance, a subsequent phase II protocol on six patients with less extensive disease showed no response, no disease stabilization, and rapid clearance of the vector due to antibody development (80). Conceivably, patients’ immune system status played a major role in conditioning gene therapy effectiveness.
Regarding prodrug activation therapy, phase I/II studies in patients with recurrent brain tumors receiving intratumor stereotactic administration of packaging cells producing a retroviral vector encoding for the thymidine kinase gene of HSV type 1 (HSV-TK), followed by treatment with GCV reported up to 30% objective responses (36, 38, 81, 82, 83, 84, 85). However, no significant therapeutic benefit over radiotherapy was obtained in a phase III study in newly diagnosed patients with glioblastoma multiforme (86). A combined approach, based on stereotactic intratumor injection of packaging cells producing a retroviral vector carrying the human IL-2 and the HSV-TK genes (87), followed by GCV administration, led to tumor regression in four treated patients, with partial response in one case (88).
An improvement of gene therapy efficacy has been observed in association with conventional radiotherapy and chemotherapy. Chemosensitization of a variety of cancers after wild-type TP53 delivery has been demonstrated in in vitro and in vivo preclinical studies (89) and confirmed in a clinical trial in patients with non-small-cell lung cancer treated with a nonreplicating p53 adenoviral vector (90). Radiosensitization has also been demonstrated after TP53 or HSV-TK gene transfer (91, 92, 93, 94). A phase I/II trial of radiation therapy in combination with three biweekly intratumor injections of p53 adenoviral vector in patients with advanced non-small-cell lung cancer documented a 1-yr progression-free survival of 45.5%, superior to historic controls (27). These results have been recently confirmed in a phase II protocol, reporting a response rate of 12 of 19 treated patients (95).
First results from clinical trials with conditionally replicating oncolytic viruses are available. In phase I/II dose-escalation protocols of intratumoral injection of dl1520, objective tumor responses, generally minor, were reported in 10–25% of cases, even at high virus doses (96). In particular, tumor necrosis at the site of single dl1520 injection was demonstrated in three of 22 patients with head and neck cancer enrolled in a phase I trial (97), whereas in a subsequent phase II study of repeated virus injection, three complete responses and two partial responses were observed out of 40 treated patients (79). Mild tumor responses also were reported in six of 23 patients with pancreatic adenocarcinoma enrolled in a phase I study of intratumor administration of dl1520 (98).
Patients who received the highest viral doses (1012 pfu) experienced better survival than patients treated with the lower doses in a phase I study of intraarterial dl1520 administration in patients with colorectal carcinoma liver metastases (99). On the other hand, no significant tumor response was achieved with dl1520 as a single therapeutic agent in patients with hepatocellular carcinoma (52), recurrent ovarian cancer (30), or advanced solid cancers metastatic to the lung (54).
Treatment with the conditionally replicating HSV mutant G207 at doses of 106 to 3 x 109 pfu led to a decrease in tumor volume in eight of 20 patients with recurrent malignant gliomas enrolled in a phase I trial, including two long-term survivors (60, 100). In a phase I study of replication competent HSV1716 at doses of 103 to 105 pfu in nine patients with recurrent malignant gliomas, four cases of long-term survival were documented, although no tumor responses were detected (61). Two complete tumor responses and three partial responses were observed, with evidence of viral replication and immune infiltration in injected lesions in a phase I clinical protocol of intralesional administration of replication competent vaccinia virus carrying the GM-CSF gene in patients with refractory, recurrent melanoma (39).
Conventional chemotherapy and radiotherapy also have been associated with delivery of replication-competent viruses, resulting in improved clinical response. In a clinical trial in head and neck cancer patients treated with dl1520 as a single therapeutic agent, objective responses were observed in 15% of cases compared with 60% of patients when the treatment was combined with 5-fluorouracil/cisplatin chemotherapy (49). Stabilization of disease and two cases of objective response were achieved in patients with multiple colorectal liver metastases undergoing intraarterial dl1520 and 5-fluorouracil infusion (51) or intrahepatic artery dl1520 infusion plus iv 5-fluorouracil and leucovorin (55).
4. Comment.
A decade of clinical trials for cancer has demonstrated disappointing results, with minimal antitumor efficacy of currently available gene therapy tools. On the other hand, treatment modalities have been demonstrated to be safe, with only minor gene therapy-related toxicities. On a cost-benefit analysis, even anecdotal reports of cases of response to gene therapy in patients with tumors refractory to conventional treatment still favor gene therapy intervention, considering the substantial safety of the procedure.
Assessment of transgene expression in target cells has demonstrated poor transduction efficiency of gene transfer vectors, which conceivably accounts for most therapeutic failures. Thus, key issues to be considered are the improvement of vectors to achieve high levels of therapeutic gene expression and transduction of a sufficient number of target cells to result in clinical benefits. To overcome the need to infect all tumor cells to achieve complete response, suicide or cytokine genes should be inserted into oncolytic vectors to increase tumor cell killing and antitumor immunity.
B. Gene therapy strategies for targeting endocrine and endocrine-related tumors
1. Overview.
An important issue in the development of gene therapy protocols is the need to target therapeutic gene delivery. Indeed, safety is a primary concern of gene therapy, and targeted vectors are required both to minimize the risk of germ line cell transduction and to prevent side effects to the surrounding healthy tissues. Moreover, targeting can reduce vector wastage and the amount of vector stocks that need to be produced and administered in vivo to achieve therapeutic levels of transduction.
Targeting of vectors can be obtained in many ways (Table 3). The easiest one is to administer the vector directly at the target site. For systemic administration, molecular engineering is required to target either gene expression (transcriptional targeting) or gene delivery (transductional targeting). Transcriptional targeting can be attempted by the introduction of tissue-specific or tumor-specific enhancers/promoters that control the expression of therapeutic genes (101, 102). Transductional targeting is based on enhanced interaction between the vector and target cell surface. It may exploit the natural tropism shown by some viruses for specific tissues or be achieved by modification of viral envelope protein sequences, by insertion of ligand molecules, by viral envelope pseudotyping, or by expression of antibodies on the viral particle surface to confer new binding specificity toward target cell receptors (103).
|
2. Transductional targeting approaches.
Attempts to improve efficacy and selectivity of vector targeting to endocrine and endocrine-related tumor cells have been reported only for breast, ovary, and prostate cancer. An example of vector retargeting was the engineering of retroviral vector envelope glycoproteins were to display single chain antibodies recognizing Her2/neu, which is a member of the epidermal growth factor (EGF) receptor (EGFR) family of receptors, overexpressed in 20–30% of breast and ovarian cancers (105). Targeting of the EGFR in tumor cells was also performed with adenoviral vectors, by using a bifunctional crosslinker, i.e., the Fab fragment of an antiknob monoclonal antibody conjugated with an anti-EGFR monoclonal antibody (106), or by using a single chain Fv antibody fragment specific to the fiber, linked to EGF as a fusion protein (107). Other efforts of adenoviral retargeting by using bifunctional crosslinkers include conjugates between antiviral knob monoclonal Fab fragment and fibroblast growth factor (108) or folate (109) to target fibroblast growth factor receptor and folate receptor, respectively, both overexpressed on the surface of a variety of tumor cells, including ovarian and breast carcinoma cells.
An alternative transductional targeting approach is based on engineering of surface viral molecules to modify viral tropism. An example of this approach is the generation of a modified AAV vector displaying a 15-amino acid peptide, which binds to the human LH receptor (LH-R), to selectively transduce LH-R-bearing cells, such as ovarian cancer cells (110). Transduction was shown to be LH-R-mediated and to be increased by progesterone treatment, via induction of LH-R expression (110). Similarly, the fiber of an oncolytic adenovirus was modified by incorporating an integrin binding motif to increase transduction of ovarian cancer cells, which generally do not express the coxsackie adenovirus receptor (111, 112). At variance with adenoviral vectors, engineering of envelope glycoproteins of ecotropic and amphotropic retroviruses to redirect virus tropism may markedly impair transduction efficiency, such as is the case of modified retroviruses targeting EGFR, IGF receptor, and folate receptor (113, 114, 115). It is conceivable that retargeted retroviral particles bind to the target cells in an envelope-independent manner and that the modification of cellular factors incorporated into the lipid envelope plays a dominant role in promoting initial adsorption of viral particles to cells. The receptor binding domain of the envelope glycoprotein would then function in a secondary recognition step essential for intracellular translocation of the virus particle.
Hypothetically, transductional targeting strategies could be feasible in a variety of endocrine tumors, exploiting targeting moieties such as antibodies and hormones directed to receptor molecules selectively expressed on the surface of the target endocrine tumor cell. Moreover, innovative methodologies, involving screening of phage display libraries, are available to find specific ligands with a high degree of specificity for the target cancer cell, without requiring that the molecules against which they are targeted be identified (116, 117).
3. Transcriptional targeting approaches.
Transcriptional targeting strategies have been used largely to selectively express cytokine or suicide genes in endocrine and endocrine-related tumor cells (Table 3). Targeting has been attempted by using enhancer/promoter sequences of genes that are selectively expressed in endocrine and endocrine-related tissues or tumors. Endocrine tumors favor this targeting approach because, typically, they express a variety of specific genes.
a. Thyroid cancer.
Transcriptional targeting of differentiated thyroid carcinomas was achieved by using thyroglobulin (TG) promoter to control therapeutic gene expression in either retroviral or adenoviral vectors (118, 119, 120). Strategies to enhance promoter activity included the use of a synthetic TG enhancer/promoter sequence (121) or a tandemly repeated TG promoter in an adenoviral vector (122), the replacement of viral enhancer with TG enhancer in a retroviral vector (123), or the use of histone deacetylase inhibitors (121, 124). Another strategy to improve TG promoter activity was the use of a Cre-loxP system, in which the Cre recombinase was controlled by TG promoter to switch on transgene expression in target thyroid carcinoma cells (125). With regard to undifferentiated and anaplastic thyroid carcinomas, which do not produce TG, targeting of transgene expression was obtained by coexpressing thyroid transcription factor-1, which activates the TG promoter, together with TG promoter-driven therapeutic genes (126).
A tissue-specific approach was attempted also for medullary thyroid carcinomas, exploiting the promoter sequence of the calcitonin gene/calcitonin gene-related peptide gene to drive selective expression of therapeutic genes in target tumor cells (127). As for the TG promoter, also in the case of transcriptional control elements of the calcitonin gene/calcitonin gene-related peptide gene, increased efficacy and specificity were achieved by engineering a chimeric sequence containing a tandemly repeated enhancer sequence and a minimal promoter (128, 129).
b. Pituitary adenomas.
Targeting therapeutic genes to specific cell types is particularly relevant for gene therapy of pituitary adenomas to spare normal pituitary cells and neighboring tissues. Cell type-specific expression of the therapeutic gene was achieved using the promoters of the GH, glycoprotein hormone -subunit, prolactin (PRL), and POMC genes (130, 131, 132, 133, 134). As for thyroid tumors, also in the case of pituitary adenomas, Cre-mediated activation of loxP-repressed transcriptionally targeted therapeutic genes was demonstrated to be an efficacious strategy for targeted suicide gene therapy. In an in vitro and in vivo model of GH-secreting adenomas, coinfection with adenoviruses carrying either Lox-P-repressed diphtheria toxin gene under GH promoter regulation or the Cre recombinase gene under the control of GH promoter caused a marked tumor regression (135). In addition to cell type-specific expression of the therapeutic gene, regulated transgene expression was pursued by using a pharmacologically regulated gene expression system, such as the tetracycline-inducible system. By driving the expression of the tetracycline transactivator through the PRL-specific promoter, expression of the inducible transgene was restricted to both lactotrophic tumor cell lines and PRL-positive cells in primary anterior pituitary cultures and within the pituitary gland in vivo (136).
c. Adrenocortical carcinoma.
Transcriptionally targeted gene therapy for adrenocortical carcinoma was attempted by using a chimeric enhancer/promoter element, containing both the CYP11B1 promoter and the P450SCC enhancer to drive transgene expression (137). In this tumor model, expression of HSV-TK in stably transfected cells was enhanced by treatment with factors acting through the cAMP pathway, such as ACTH (137).
d. Prostate carcinoma.
Due to the presence of well-characterized prostate-specific markers, such as PSA and a variety of prostate-unique genes, prostate carcinoma has represented an ideal model for targeted gene therapy treatment (138). Several studies used PSA regulatory regions to drive expression of therapeutic genes (139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149). Tandem duplication of the PSA enhancer increases expression approximately 50-fold while retaining tissue-specific control (150). A higher efficiency was achieved by coupling the PSA promoter to a yeast promoter (142, 151) or by complex engineering of the enhancer sequence (152). A minimal composite PSA promoter/enhancer element was used to drive expression of adenoviral E1A in the attenuated replication-competent vector CN706 (56, 153). The PSA enhancer region used in this vector contained a functional androgen response element capable of up to 100-fold induction of transgene expression in PSA-expressing cells in the presence of testosterone or the steroid analog R1881 (154). At variance, a long PSA promoter allowed efficient transgene expression both in the presence and absence of androgens (143). Coexpression of a partial androgen receptor gene and PSA-driven therapeutic gene allowed activation of the PSA enhancer/promoter even in the absence of androgens (155). Other tissue-specific promoters used to target prostate cancer include the human kallikrein gene promoter, which is expressed predominantly in the prostate and transcriptionally up-regulated by androgens (156, 157); prostate-specific membrane antigen, which is highly expressed in metastatic or poorly differentiated prostate cancer and up-regulated by androgen deprivation (158, 159, 160, 161); the probasin promoter, selectively expressed in prostate cells (162, 163, 164, 165); or the osteocalcin promoter to target metastatic lesions to the bone (166, 167). Tissue-specific, inducible systems were developed by using a prostate-specific chimeric promoter, based on the probasin gene promoter and two copies of the androgen response region, which was induced by activation of caspase after administration of a chemical inducer of dimerization (168). The same chimeric promoter sequence was used in a tetracycline-regulated expression system (169).
e. Breast carcinoma.
Transcriptional targeting of the mammary tissue has been pursued by using either the human -lactalbumin or ovine ß-lactalbumin promoter to drive therapeutic gene expression (170). Tumor targeting was attempted by using promoters of tumor-specific genes, such as the DF3/MUC-1 gene, which encodes a high molecular weight mucin-like glycoprotein overexpressed in the majority of breast cancers (171, 172, 173, 174), and the HER-2/neu oncogene (also named c-erbB-2), which is overexpressed in a variety of human cancers, including breast and ovarian carcinomas (173, 175, 176, 177). A clinical trial was conducted for patients with recurrent breast carcinoma expressing the HER-2/neu gene (178). These patients were treated by intratumor injection of a plasmid containing the cytosine deaminase gene driven by the tumor-specific erbB-2 promoter. Efficiency of cancer cell killing was proportional to cellular HER-2/neu expression.
Estrogen-responsive elements, which allow modulation of transgene expression by estrogens and tamoxifen, have been used to develop conditionally replicating adenoviral vectors to target estrogen receptor (ER)-positive breast cancer (179), or, in combination with hypoxia-responsive elements, to develop a targeted and regulated adenoviral vector (180). Gene therapy for breast carcinoma may also be approached by tailoring a virus with affinity to this tissue, such as the mouse mammary tumor virus. The glucocorticoid-responsive long terminal repeats (LTRs) of this retrovirus have been used as promoters for dexamethasone-inducible oncolytic cytokine expression (181).
f. Ovarian carcinoma.
Several tumor-specific or tissue-specific promoters have been investigated to target gene delivery to ovarian cancer. These include the promoter/enhancer sequences of the secretory leukoproteinase inhibitor gene (182, 183), the human epithelium-specific ets transcription factor gene hESE1 (184), the L-plastin gene (185), and the MUC1/DF3 gene (186), which are highly expressed in several epithelial tumors, including ovarian cancers.
g. Neuroendocrine tumors.
After positive results from clinical studies of immunoscintigraphy and radioimmunotherapy with targeted monoclonal antibodies, several tumor/tissue-specific transcriptional regulatory sequences or antigens are being exploited to target transgene expression or vector delivery to neuroendocrine tumors. Tissue/tumor-specific genes under investigation include chromogranin A and B (187, 188, 189, 190, 191), calcitonin (192), neuron-specific enolase (193), arginine vasopressin (194), somatostatin receptor (195), and the RET protooncogene (196).
4. Comment.
A number of possibilities to target endocrine and endocrine-related tumors by means of either transductional or transcriptional targeting are available and make these tumors ideal candidates for gene therapy. However, because malignant transformation is generally accompanied by cell dedifferentiation and loss of tissue-specific features, targeting is often not feasible. It is thus conceivable that effective endocrine tumor targeting could be achieved only by using oncolytic viruses characterized by selective tropism for endocrine glands. Indeed, as discussed in the following sections, several viruses have been shown to preferentially infect and replicate in endocrine cells, such as reovirus in thyroid and pancreas, adenovirus in adrenal cortex, and HSV in adrenal cortex and adrenal medulla. These viruses could be engineered as oncolytic agents for endocrine tumors or, more generally, for gene therapy applications in the endocrine system.
|
II. How to Exploit the Endocrine System for Regulating Therapeutic Gene Expression |
|---|
|
TopAbstractI. IntroductionII. How to Exploit...III. Endocrine Cell-Specific...IV. Endocrine Side Effects...V. Oncolytic Virus Infection...VI. Endocrine Response to...VII. Risk of Germ...VIII. SummaryReferences |
|---|
Tight regulation of gene expression is a typical feature of the endocrine system. The increasing understanding of tumor biology and the molecular mechanisms involved in gene expression control (i.e., activators, repressors, coregulators) (201) has led to the development of new molecular switches that could be exploited for gene therapy applications and functional genome research.
A. Physiologically regulated gene expression systems
Heat, hypoxia, glucose deprivation, irradiation, and chemotherapeutic agents up-regulate various genes involved in stress responses. Promoters of these genes are attractive for cancer gene therapy because they depend to a large extent on the biology of the tumor or are already induced by various therapeutic modalities. Genes up-regulated in these conditions include multidrug resistance (MDR-1), human heat-shock protein (HSP), vascular EGF (VEGF), irradiation-inducible early growth response (Egr-1), and the tissue plasminogen activator (tpa) genes.
Hypoxia is a common feature in many solid tumors and plays a significant role in the resistance of cancer to ionizing radiation and cytotoxic chemotherapy. Cellular hypoxia induces a stress response with up-regulation of many genes involved in shifting cellular respiration toward the glycolytic pathway, increasing erythropoiesis and angiogenesis (202). The promoters of the genes that mediate this adaptive response, i.e., phosphoglycerate kinase 1, erythropoietin, and VEGF, contain cis-acting hypoxia response elements capable of binding hypoxia inducible factor 1 and related proteins (203, 204, 205). The feasibility of tumor targeting by using hypoxia response elements promoters to drive therapeutic gene expression has been demonstrated in vitro (206, 207) and in vivo (208, 209, 210).
Irradiation- and chemotherapy-responsive promoter sequences were identified for tpa and Egr-1 genes. Expression of the radiosensitizing cytokine TNF- under the control of the Egr-1 promoter followed by either radiotherapy (211, 212) or chemotherapy (213) led to synergistic antitumor effects. Engineering of the CArG consensus elements derived from Egr-1 promoter allowed optimization of enhancer sensitivity to low doses of ionizing radiation (214). Chemotherapeutic agents, such as vincristine and doxorubicin, also induce the MDR-1 gene, which encodes a membrane-effluxing glycoprotein. Effective tumor-targeting has been achieved by combining MDR-1 promoter-regulated expression of therapeutic genes and chemotherapy (215).
Heat-shock proteins are induced by a variety of stressful environmental conditions, such as heat, irradiation, hypoxia, acidosis, hypoglycemia, and osmotic changes, which are generally present in poorly vascularized tumors. Inducible HSP promoters have been used to drive the expression of a variety of therapeutic genes in experimental tumor models after hyperthermia therapy or glucose starvation conditions (216, 217, 218, 219, 220, 221), as well as to enhance the oncolytic effect of replicative viruses (222, 223).
B. Pharmacologically regulated gene expression systems
A number of drug-related gene expression systems are available whereby targeted gene transcription is controlled through the use of small-molecule inducing compounds (101), such as the antibiotics tetracycline (224), streptogramin (225), and macrolides (226); the insect steroid ecdysone or its analogs (227); the antiprogestin mifepristone (RU486) (228, 229, 230, 231); and chemical dimerizers represented by the immunosuppressant rapamycin and its analogs (232, 233).
1. Mifepristone-inducible system.
At variance with other systems using regulatory proteins of nonhuman origin, the mifepristone system is based on a mutant human progesterone receptor, thus minimizing problems of potential immunogenicity. Like other nuclear receptors of the steroid-hormone superfamily, the progesterone receptor consists of three major functional domains; i.e., the DNA binding domain, the ligand binding domain, and the transactivation domain, which can be interchanged with correspondent elements of other receptors to generate chimeric molecules. The progesterone receptor used in this system has a deletion in the carboxy terminus of the ligand binding domain so that it no longer binds to the agonist progesterone but is still capable of binding to the antagonist mifepristone (234). This mutant is fused to the DNA binding domain of the yeast transcription factor gal4 and the transactivation domain of the HSV VP16 protein, yielding the GL-VP transcription factor (235). Alternatively, the transactivation domain of the chimeric transcription factor may be represented by the activation domain of the nuclear factor-
B p65 subunit. Moreover, to further reduce the risk of inducing a host immune response, the yeast gal4 DNA binding domain could be replaced by a DNA binding domain of human origin. In the presence of mifepristone, this chimeric regulator binds to genes with upstream gal4 recognition sequences and efficiently activates transcription of the target transgene. The efficiency of the mifepristone-regulated system has been demonstrated by incorporating it in recombinant viral and nonviral vectors (236, 237).
2. Ecdysone-inducible system.
This system is based on the use of the insect molting steroid hormone ecdysone and its receptor, the ecdysone receptor (EcR), which is a member of the nuclear receptor superfamily. In the ecdysone-inducible system, a chimeric protein (VgEcR) composed of the VP16 activation domain fused to an EcR with altered DNA-binding specificity heterodimerizes with the retinoid X receptor (RXR) and binds a unique synthetic response element not recognized by natural nuclear hormone receptors. Upon exposure to ecdysone or the synthetic analog muristerone, the VgEcR/RXR complex efficiently induces transgene expression (238, 239) (Fig. 2). Advantages of this system include lower basal activity and higher inducibility compared with other regulated systems and absence of ecdysone effects on mammalian cell physiology. Moreover, ecdysteroids have a lipophilic nature favoring efficient penetration into all tissues including the brain, possess short half-lives that allow for precise and potent inductions, and exhibit favorable pharmacokinetics that prevent storage and expedite clearance (227). This system has been effectively used to generate transgenic mice (227) and inducible viral vectors (240).
|
4. Tamoxifen-inducible system.
Disadvantages of transcriptionally regulated inducible systems are represented by basal activity in the absence of induction and low inducibility. As an improved hormone-dependent strategy for regulating protein expression at a posttranslational level, fusion of the hormone-binding domain of the transcriptionally inactive mutant of the murine ER has been adopted (245, 246, 247, 248). The modified receptor is unable to bind estrogen yet retains normal affinity for the synthetic ligand 4-hydroxy-tamoxifen (4-OHT). After administration of ligand, the ER fusion proteins are rapidly activated by allowing translocation from the cytosol to the nucleus (249). The effects and pharmacology of nonsteroidal antiestrogens, such as tamoxifen and its derivatives, have been well characterized in animal and human trials, confirming their suitability for gene therapeutic approaches in humans (250). This regulated system was used for the development of an inducible adenoviral vector for cancer gene therapy. Activity of the E2F1 gene, encoding a transcription factor that triggers massive apoptosis in several human cancers, was made 4-OHT-dependent by fusion to the ligand binding domain of the ER (251). Upon 4-OHT administration, the ER-E2F1 fusion protein translocated from the cytosol to the nucleus, transactivated E2F-dependent promoters, and rapidly induced cytotoxicity both in vitro and in vivo (251).
Another approach to designing an ER-based inducible system was the construction of chimeric regulators containing the human ER ligand binding domain and a Cys (2)-His (2)-type zinc finger DNA binding domain. Cys (2)-His (2)-type zinc finger domains are common among human DNA binding proteins and can be engineered to selectively bind different DNA sequences. These chimeric regulators demonstrated a very efficient drug-dependent transgene induction in vitro and in vivo, after adenovirus-mediated gene delivery to mice (252). Moreover, specific point mutation in the ER ligand-binding domain that ablated estrogen binding enabled selective in vivo regulation by tamoxifen (252).
5. Other steroid hormone-inducible systems.
Other inducible systems based on the use of steroid hormones include the thyroid hormone-, androgen-, and vitamin D3-regulated systems. A thyroid hormone-responsive system was developed by using three copies of palindromic thyroid hormone/retinoic acid-responsive element to drive transgene expression. Variations of thyroid hormones and all-trans-retinoic acid levels within their physiological range allowed in vivo regulation of transgene expression (253). Tissue-specific, thyroid hormone-mediated expression of toxic genes for gene therapy of gliomas was achieved by using the promoter of the myelin basic protein, which contains a thyroid hormone response element (254). Tissue-specific inducible expression of therapeutic genes was also achieved by using a human osteocalcin promoter, which is activated by vitamin D3, to drive the early adenoviral E1A and E1B genes. Not only did this promoter allow selective replication of the oncolytic adenovirus in osteocalcin-expressing cells, but it also enhanced viral replication of at least 10-fold upon vitamin D3 exposure (167).
Androgen-responsive expression of therapeutic genes has been widely used for prostate cancer gene therapy. Androgen response elements have been identified and characterized in the enhancer/promoter region of a variety of androgen-inducible genes, and engineered to attain enhanced transactivation efficiency (154, 255). New inducible systems, based on these naturally evolved switches, might be developed by manipulation of steroid hormone nuclear receptors or their response elements (256, 257).
6. Comment.
As shown for gene therapy targeting strategies, endocrinology might greatly contribute to the development of regulated transgene expression systems. Regulated endocrine axes might ideally represent a paradigm of any gene switch systems, and knowledge in the field of endocrinology should be translated in the set-up and refinement of these gene expression tools. Potential applications will range from oncology to cardiovascular diseases, hormone deficiencies, and inherited diseases.
|
III. Endocrine Cell-Specific Genes as New Therapeutic Tools |
|---|
|
TopAbstractI. IntroductionII. How to Exploit...III. Endocrine Cell-Specific...IV. Endocrine Side Effects...V. Oncolytic Virus Infection...VI. Endocrine Response to...VII. Risk of Germ...VIII. SummaryReferences |
|---|
A. Sodium/iodide symporter
The iodide transporter NIS is an intrinsic plasma membrane protein that mediates the active transport of iodide in the thyroid, lactating mammary gland, stomach, and salivary glands (258, 259). The presence of NIS in the thyroid gland is exploited in diagnostic scintigraphic imaging and radioiodide therapy of thyroid cancer. The demonstration of NIS expression in breast cancer, but not in normal breast tissue (260), suggests the potential use of radioiodine as a diagnostic and therapeutic tool also for nonthyroid cancers in which NIS is functionally active. Moreover, ectopic NIS expression in cancer cells by gene transfer may be exploited for both diagnostic and therapeutic purposes. In this regard, NIS-transduced tumor cells exhibited efficient iodide accumulation, either in culture or in xenografted tumors in nude mice, and were selectively killed by radioiodide (261, 262, 263, 264). In a mouse model of intracerebral gliomas, which had been retrovirally transduced with human NIS, tumors could be imaged by 99mTcO4 and 123I scintigraphy and underwent a significant regression after treatment with 131I (265). Moreover, radioiodide uptake and NIS expression in the thyroid gland could be reduced by feeding a T4-supplemented diet (264), thus preventing thyroid toxicity. Tissue-specific, androgen-dependent iodide uptake has been induced in prostate cancer cells in vitro by PSA promoter-directed NIS expression (266). Transfected tumors showed a significant regression after single-dose radioiodide therapy in animal models of prostate cancer (267). Adenovirus-mediated NIS gene transfer followed by radioiodide administration resulted in highly active iodide uptake and significant tumor volume reduction (268). However, less enthusiastic results have been reported in other tumor models, in which a rapid radioiodide efflux due to lack of iodide organification and intracellular retention was demonstrated (269, 270, 271, 272). An amelioration of iodide kinetic was achieved by thyroid ablation and low-iodide diet, although this regimen, in combination with radioiodide therapy, did not inhibit tumor development (273). Cotransfection of non-small-cell lung cancer with both NIS and the thyroperoxidase (TPO) gene, which catalyzes iodination of proteins and subsequent iodide retention within thyroid cells, resulted in an increase in radioiodide uptake and retention and enhanced tumor cell killing (272). However, by using an adenoviral vector to deliver NIS and TPO, the levels of iodide organification achieved were too low to significantly increase iodide retention (274).
B. Noradrenaline transporter
Metaiodobenzylguanidine conjugated to 131I-iodide is an effective agent for targeted radiotherapy of tumors of neural crest origin that express the NAT, i.e., pheochromocytoma, neuroblastoma, carcinoid and medullary thyroid carcinoma. Transfer of the NAT gene into nonneuroectodermal tumors would allow targeting of 131I-MIGB to a wide range of tumor types for which no specific and targetable characteristic currently exist. An attractive feature of tumor targeting with 131I-MIGB is that cancer cells that fail to accumulate a lethal quantity of 131I-MIGB may still absorb ß-radiation from neighboring targeted cells. The feasibility of this approach was demonstrated by transfecting cell lines with NAT (275, 276, 277, 278). In particular, human glioblastoma cell lines transfected with bovine NAT showed a 15- to 25-fold enhancement of radionuclide uptake and dose-dependent cell killing (277, 279). By using cells grown as either monolayer cultures or spheroids, 131I-MIGB was twice as toxic for cells in spheroids compared with cells in monolayers, consistent with a greater radiation cross-fire effect (radiological bystander effect) from 131I ß-radiation in the three-dimensional tumor spheroids (277, 280). Moreover, efficient and tumor-selective NAT expression was achieved by placing the transgene under the transcriptional control of the telomerase RNA promoter (281).
C. Somatostatin receptor
Somatostatin and its analogs suppress the growth of tumor cells that express somatostatin receptors, such as neuroendocrine tumors (282). This antiproliferative effect is mediated by somatostatin receptor subtype 1 (sst1), sst2, and sst5 (283). Tumor progression is often associated with the loss of differentiated functions, including expression of somatostatin receptors. In this regard, restoration of somatostatin responsiveness in tumor cells by sst2 gene transfer has been demonstrated to be an effective gene therapy approach for human pancreatic adenocarcinomas (284, 285, 286), which, typically, show a specific loss of sst2 expression (287). Stable transfection of these cells with human sst2 resulted in the induction of a negative-autocrine loop with secretion of endogenous ligand that activated constitutively the recombinant sst2 receptor (284, 285, 288). sst2-Expressing cells showed significant reduction of cell growth and tumorigenicity both in vitro and in vivo (284, 285, 288), and this antitumor effect was enhanced by administration of the cytotoxic somatostatin analog AN-238 (288). Moreover, a significant bystander effect and inhibition of metastatic progression was reached when only 25% of tumor cells expressed sst2 (284, 288, 289).
Aside from antitumor activity, sst2 gene transfer has been exploited for in vivo noninvasive nuclear imaging of tumors. After transduction of tumor cells with viral vectors encoding the sst2 gene, tumor masses were visualized using radiolabeled somatostatin-avid peptides (290, 291, 292, 293). Transduction of tumors with vectors coexpressing a therapeutic gene together with sst2 allows noninvasive in vivo monitoring of the efficacy of gene therapy, as demonstrated in ovarian and lung cancer models in mice injected with bicistronic adenoviruses carrying both HSV-TK and sst-2 (292, 293).
Comment.
The idea of noninvasive in vivo monitoring of the efficiency of gene therapy by radionuclide imaging has already moved to the clinic, with a protocol of gene therapy for glioblastoma multiforme based on the intratumor delivery of HSV-TK followed by radiolabeled nucleoside analog administration (294). Preclinical results in animal models indicate NIS, NAT, and sst2 gene therapy could also be effectively applied in humans for diagnostic and therapeutic purposes. Specificity and stability of the interaction between the endocrine transporter, receptor, or enzyme and the ligand or substrate are critical for the success of these therapeutic approaches.
|
IV. Endocrine Side Effects of Gene Therapy |
|---|
|
TopAbstractI. IntroductionII. How to Exploit...III. Endocrine Cell-Specific...IV. Endocrine Side Effects...V. Oncolytic Virus Infection...VI. Endocrine Response to...VII. Risk of Germ...VIII. SummaryReferences |
|---|
).
|
Nonviral gene delivery (reviewed in Refs. 295, 296, 297) refers to the use of naked DNA (298), cationic lipids formulated into liposomes and complexed with DNA (lipoplexes) (299, 300), cationic polymers complexed with DNA (polyplexes) (301), polymeric vesicles complexed with DNA (302), or a combination of both cationic lipids and cationic polymers complexed with DNA (lipopolyplexes) (303). There have also been attempts to combine the benefits of viral and nonviral systems into one delivery vehicle (304).
Cationic liposomes and cationic polypeptides are efficient reagents for the transfer of nucleic acids to cells in vitro and in vivo. These reagents have several advantages over other methods of nucleic acid transfer; however, toxicity remains a significant problem, especially in vivo. These vehicles have been used in several studies, including phase I and II clinical trials (299). In cancer gene therapy, lipoplexes are generally delivered through iv or intraarterial administration, or direct intratumor injection, to limit toxicity to the targeted tissues (reviewed in Ref. 299). One of the drawbacks of intratumor administration is the localization of the delivered nucleic acids predominantly in the needle track. When administered systemically by iv injection, the distribution is mainly in the lung, followed by liver, spleen, and kidney, whereas the intraarterial route allows selective delivery to the target lesion (305, 306, 307, 308). In cell culture, lipoplexes cause several changes to cells, including cell shrinking, growth inhibition, and vacuolization of the cytoplasm (309). Cationic lipids may also induce hemolysis (310) or fusion between erythrocytes (311). In vivo studies demonstrated toxic effects mainly in the injection site, such as inflammation of the eyes after intraocular instillation (312), epileptic seizures, and, in severe cases, death after intracerebral injection in mice (313, 314), inflammatory response after intraarticular delivery (315), epithelial cell death after intratracheal administration (316), and complement activation via the alternative pathway (317). Cationic liposomes have also been reported to induce acute systemic inflammatory reactions (299, 318) and macrophage and neutrophil infiltration into the lungs of mice when administered intratracheally (319). Nephrotoxicity (320) and hepatoxicity (321) were reported after delivery of lipoplexes via the renal artery and the portal vein, respectively. Intravascular delivery of lipoplexes may lead to embolization and microinfarctions, as demonstrated after iv, intraarterial, or intracoronary administration (322). These complexes were also found to be highly toxic when administered orally, provoking a dramatic hypothermia resulting in death in some mice (323). Systemic administration of lipoplexes stimulates the production of proinflammatory cytokines, such as TNF-, IFN-, IL-6, and IL-12 (318, 319, 324, 325). This effect is mainly related to the plasmid DNA component of these vectors and the cytosine-phosphate-guanine motifs contained within (326). Although potentially associated with significant side effects, these immunostimulatory properties may be exploited for cancer immunotherapy. Effects on the endocrine system of cytokines, including TNF-, IFN-, IL-6, and IL-12, will be discussed in Section IV.C.
Cationic polymers, such as polylysine, histones, and dendrimers, are able to interact electrostatically with the DNA molecule and condense into compact particles (polyplexes). Condensation prevents DNA degradation by nucleases and allows internalization of particles into cells by natural processes such as endocytosis, pinocytosis, and phagocytosis. To improve transfection efficiency, cell-binding ligands have been incorporated into these transfection vehicles, resulting in receptor-mediated mechanisms for cellular uptake. Positively charged polycation/DNA complexes were found to aggregate at physiological salt concentrations, to interact with components of the coagulation and complement system, and to cause aggregation of erythrocytes that can result in microembolism (301, 303, 327, 328, 329). This process may be enhanced by the avid binding of the positively charged complexes to cell membranes (330, 331, 332). Nonspecific interactions with plasma components or erythrocytes can be prevented by shielding the surface of transfection particles with hydrophilic molecules, such as polyethylene glycol (329) or transferrin (333, 334). Systemic delivery of nonshielded complexes into tumor-bearing mice resulted in high transgene expression in the lungs and lower gene expression in other organs, such as heart and liver, but was often associated with severe toxicity (328), particularly when high molecular weight complexes were used (303). At variance, shielded complexes preferentially accumulated in tumor tissues (297, 299, 333, 334).
B. Viral vectors
Endocrine side effects in the course of cancer gene therapy could be due to untoward infection of endocrine cells by nonreplicative viral vectors, active viral replication and lysis of endocrine cells in the case of replicating viruses, and toxic effects due to stimulation of cytokine production by these vectors. Endocrine side effects caused by replication-deficient viral vectors will be discussed in this section, whereas infection of endocrine cells by wild-type and oncolytic viruses will be dissected in Section V.
1. Adenoviral vectors.
Among endocrine glands, adenoviral vectors have a natural tropism
