IGF-I: from diagnostic to triple-helix gene therapy of solid tumors

of the promotor of IGF-I gene (RNA-IGF-I DNA triple helix). Human tumor cells transfected in vitro become down-regulated in the production of IGF-I and present immunogenic (MHC-I and B7 expression) and apoptotic characteristics. Similar results were obtained when IGF-I antisense strategy was applied. In both strategies the transfected cells reimplanted in vivo lose tumorigenicity and elicit tumor specific immunity which leads to elimination of established tumors.


IGF-I AND DIAGNOSTIC
There is a convergence between ontogenesis and cancerogenesis and the same specific antigens (oncoproteins) like a-fetoprotein (AFP), growth hormone (GH), and growth factors, such as IGF, FGF or EGF are present in embryo / fetal tissues and in neoplastic developing tissues. It was demonstrated that AFP, an oncoprotein present in different cancer tissues, especially in liver cancer (Abelev, 1971) is also present in normal developing tissues, and particularly in the central nervous system (CNS) (Trojan & Uriel, 1979). Similarly it was demonstrated that IGF-I, insulin-like growth factor I, is present both in normal developing CNS and in neoplastic glial cells (Kiess et al., 1989;Ayer-le-Lievre et al., 1991;Trojan et al., 1992;Sandberg et al., 1998). The presence of IGF-I was confirmed in different neoplastic derivatives, including hepatic tissues, and also using the model of murine teratocarcinoma (Trojan et al., 1994).
IGF-I is a 70-amino-acid polypeptide involved in cell and tissue differentiation (Daughaday et al., 1972;Froesch et al., 1985;Han et al., 1987;Trojan et al., 1994). IGF-I plays an important role in growth as a mediator of growth hormone (Froesch et al., 1985;Johnson et al., 1991;Le Roith et al., 2001). The action of IGF-I on cellular metabolism depends on binding proteins, IGFBP, which prolong the half life of this factor and modify its interaction with a receptor (Jones & Clemmons, 1995;Collet & Candy, 1998;Rosen, 1999;Hva et al., 1999). Binding to a specific IGF-I receptor and subsequent activation of a protein tyrosine kinase signal transduction cascade, is similar to that of insulin (Werner & Le Roith, 2000;Adams et al., 2000). IGF-I, mediated by IGF-I receptor, has been reported to block the apoptosis pathway in a variety of cell lines (Rodriguez-Tarduchy et al.,1992;D'Mello et al., 1993;Muta & Krantz, 1993;. Conversely, blocking IGF-I synthesis induces apoptotic and immunogenic phenomena (Upegui- Gonzalez et al., 1998). However in general, the molecular mechanism of control of IGF-I expression is poorly defined. Although the IGF-I gene consists of six exons (Daughaday & Rotwein, 1989;Sussenbach et al., 1992), the mature human IGF-I peptide is encoded only by exons 3 and 4, and a similar situation was found also in the widely studied rat model (Adamo et al., 1994). Deregulated expression of growth factors and/or their receptors, and especially of IGF-I, is associated with growth as well as with different diseases, including tumors (Heldin & Westermark, 1989;Antoniades et al., 1992;Trojan et al., 1993;Rubin & Baserga, 1995).
In the past few years, both laboratory investigations and population studies have provided strong circumstantial evidence that IGF physiology influences cancer risk (Yu & Robin, 2000;Pollak, 2000). Evidence further suggests that certain lifestyles, such as one involving a high-energy diet, may increase IGF-I levels, a finding that is supported by animal experiments indicating that IGFs may abolish the inhibitory effect of energy restriction on cancer growth (Yu & Robin, 2000).
According to Baserga (1995), IGF-I is one of the most important growth factors related to normal and neoplastic differentiation. IGF-I is expressed in 17 different tumors (for references see Trojan et al., 1993). Since the last symposium "IGFs and Cancer", held in Halle in Germany (15-17.09.2000), IGF-I is considered as a diagnostic marker and a biological modulator in different types of tumors, especially in brain tumors (Zumkeller & Westphal, 2001;Zumkeller, 2002).
The IGF system consists of IGF-I and IGF-II, the type I and type II IGF receptors, and specific IGF binding proteins (IGFBP-1 to -6). These factors regulate both normal and malignant brain growth. Enhanced expression of IGF-I and IGF-II mRNA transcripts has been demonstrated in gliomas, meningiomas, and other tumors. Abnormal imprinting of IGF-II occurs in gliomas, medulloblastomas, and meningiomas (Zumkeller & Westphal, 2001). Both types of IGF receptor are expressed in gliomas and, in particular, the type I IGF receptor appears to be upregulated in malignant brain tissue. Meningiomas and, to a lesser degree, malignant gliomas were found to synthesise IGFBP-1, supporting the notion that IGFBPs contribute towards the growth of CNS tumors in humans; glioblastoma cell lines were found to express mRNA for IGFBP-1 (42% of cell lines), IGFBP-2 (65%), IGFBP-3 (97%), IGFBP-4 (3%), IGFBP-5 (74%) and IGFBP-6 (94%) as determined by polymerase chain reaction (Zumkeller, 2002). The relationship between IGF-I and IGFBP is starting to be introduced in clinical diagnostics as one of the indications of precancerous development. Among the numerous examples of cancer growth and metastasis monitored by the IGF-I marker (serum level) one should mention colorectal cancers (Mishra et al., 1998;Manousos et al., 1999;Giovanucci et al, 2000;Wu et al., 2002;Bustin et al., 2002), of breast (Vadgama et al., 1999;Campbell et al., 2001;Helle et al., 2001;Eppler et al., 2002;Bajetta et al., 2002), prostate cancer ( Wolk et al., 2000;Mita et al., 2000) and of lung (Lee et al., 1999;Yu et al., 1999;Olchovsky et al., 2002). The role of IGFs in cancer is supported by epidemiologic studies which have found that high levels of circulating IGF-I and low levels of IGFBP-3 are associated with increased risk of several common cancers, including those of the prostate, breast, colorectum, liver and lung (Giovanucci, 1999;Yu & Robin, 2000).
IGF-I is an important mitogen required by some cell types to progress from the G1 to the S phase of the cell cycle. IGFBPs can have opposing actions, in part by binding IGF-I, but also by direct inhibitory effects on target cells. Because the tissue determinants of IGF bioactivity appear to be regulated in parallel with circulating IGF-I level, it is reasonable to hypothesize that the substantial intraindividual variability in circulating levels of IGF-I and IGFBP-3 may be important in determining risk of some cancers. In general, twoto four-fold elevated risk has been observed for prostate cancer in men in the top quartile of IGF-I relative to those in the bottom quartile, and low levels of IGFBP-3 were associated with an approximate doubling of risk. For colorectal neoplasia, four-fold elevated risk was observed in men and women with low Vol. 49 IGF-I, diagnostic and gene therapy of glioblastoma IGFBP-3, whereas high IGF-I was associated with a doubling of risk (Giovanucci, 1999).

IGF-I AND ANTI-GENE THERAPIES
As far as the role of oncoproteins in tumorigenesis is concerned, different approaches of anti-tumor treatment have been considered. The most classical was treatment using antibodies, i.e. treatment of liver cancer with injection of antibodies directed toward oncoproteins like AFP. Unfortunately, this type of technique was not specific enough for the treated tissues (AFP like other oncoproteins is present in different types of differentiating cells). Following the hypothesis that neoplastic differentiation is related to the presence of oncoproteins, the arrest of oncoprotein synthesis at the gene level seemed to be the best way to stop tumor development. "Anti-gene" strategies offer new possibilities for cancer therapy: antisense technique ( arresting protein synthesis at the transcription level) (Rubinstein et al., 1984;Weintraub et al., 1985;Green et al., 1986) or triple helix technique ( stopping the synthesis at translation level) (Dervan, 1992;Hélène, 1994).
In the past twenty years, it has been shown that natural antisense RNA which is transcribed from one strand of DNA could hybridize to the sense RNA. This natural physiological regulation represents the basics of the antisense approach to inhibit artificially the expression of particular genes involved in human diseases. Using this antisense strategy, the translation of messenger RNA (sense RNA) can be blocked by binding a complementary strand to this mRNA. The achievement of this artificial regulation could be done using either short oligonucleotides delivered to cells via appropriate carriers or by using plasmid constructs transcribing intracellular antisense RNA (Trojan et al., 1992).
IGF-I antisense gene therapy (Trojan et al., 1993) was introduced in clinical trials to treat hepatoma (Shanghaï, China) and glioblastoma (Cleveland, U.S.A., and Bydgoszcz, Poland) (Anthony et al., 1998). Glioblastoma is the most frequent and usually fatal tumor. Recently some interesting results concerning glioblastoma treatement were published by American-Asian cooperation (Wongkajornslip et al., 2001). A subject inflicted with glioblastoma who underwent partial tumor resection and radiotherapy, after subcutaneous injections of IGF-I antisense transfected glioma cells for 8 weeks, developed peritumor necrosis. The latter lesion was infiltrated by lymphocytes containing both CD8 and CD4 cells. The functional activity of these lymphocytes was demonstrated by the active production of interferon gamma and tumor necrosis factor alpha.
In the triple helix (TH) technology the oligonucleotides that block gene expression are triple-helix forming oligonucleotides (TFOs). They block RNA polymerase transit by forming a triple-helical structure on DNA (Dervan, 1992;Hélène, 1994).
The TH strategy was applied to the ras oncogenes which are the most frequently activated oncogenes in human cancer. In vitro transcription of human Ha-ras was inhibited by TFOs targeted to sequences recognized by the Sp1 transcription factor (Mayfield et al., 1994). Using transient transfection assays, it was demonstrated that a purine-rich TFO could also inhibit the transcription of murine c-Ki-ras gene in NIH 3T3 cells (Alluni-Fabroni et al, 1996).
Growth factors are known to play a role in tumorigenesis and thereby represent convenient targets for anti-gene therapies. The synthesis of human tumor necrosis factor (TNF), which acts as an autocrine growth factor in various tumor cell lines including neuroblastoma and glioblastoma, could be blocked by TFO treatment (Aggarwal et al., 1996). TFOs were also shown to bind in vitro to human EGF receptor promoter (Durland et al., 1991), and to inhibit in vitro transcription of the HER2/neu gene (Ebbinghaus et al., 1993). The transcription of endogenous human HER2/neu oncogene, which is overexpressed in breast cancer and other human malignancies, was inhibited by TFO treatment of a breast carcinoma MCF-7 cell line (Porumb et al., 1996).
More examples of the inhibitory activity of TFO on target genes involved in tumorigenesis are now available (Maher, 1996;Chan & Glazer, 1997;Giovannangeli & Hélène, 1997;Vasquez & Wilson, 1998). Most of the TFOs are targeted to polypurine and/or polypyrimidine sequences located in control regions of the gene of interest and are cell delivered via transfection with various chemical carriers. An alternative way to introduce the TFOs in the cells is to use a plasmid vector that can drive the synthesis of the TFO RNA inside the cells. This TFO generated in situ is therefore protected from degradation by nucleases and could reach its DNA target with-out being trapped in lysosomal vesicles. Obviously, it could be transfected into cells via either standard cell transfection procedures or via ways used in virus-based gene therapy. An application of this triplex-based approach was used for the inhibition of the IGF-I protein in tumorigenesis of glioblastoma and hepatocarcinoma (Shevelev et al., 1997;Upegui-Gonzalez et al., 2001;Ly et al., 2001). The IGF-I triple helix (IGF-I TH) strategy shows that an RNA strand containing a 23-nucleotide (nt) oligopurine sequence may be capable of forming a triple helix in cultured human primary glioma or rat C6 cells with an oligopurine and/or oligopyrimidine sequence of the IGF-I gene. Although we can not exclude other mechanisms, triple helix formation remains the most plausible explanation for the inhibition in expression of the IGF-I gene (Shevelev et al., 1997). The 23-nucleotide target regions are also present in other growth factors, such as IGF-II or FGF. These target regions are composed only of A and G bases, but their sequence is different in different growth factors. For example, the IGF-I gene expressed in glioblastoma has a 23-nucleotide target region different from the IGF-II gene in neuroblastoma (Trojan et al., non published data).
Another way of using TFO delivered inside the cells via transcription of a plasmid vector was recently described with the IGF-I receptor gene in glioblastoma cells .

IGF-I TRIPLE-HELIX THERAPY: APOPTOTIC AND IMMUNE MECHANISMS
To demonstrate that IGF-I, and not another factor, plays a really important role in neoplastic diseases, gene therapy based on IGF-I TH approach was applied for experimental gliomas (Shevelev et al., 1997;Ly et al., 2001). This method gave as good results as the IGF-I antisense strategy applied previously for ex- Vol. 49 IGF-I, diagnostic and gene therapy of glioblastoma perimental glioma, teratocarcinoma and hepatoma treatement (Trojan et al., 1993;1994;Lafarge-Frayssinet et al., 1997;Upegui-Gonzalez et al., 1998;Ellouk-Achard et al., 1998).
In the IGF-I antisense strategy, the transfectants lost tumorigenicity and induced a T-cell mediated immune reaction both against themselves and against their non transfected tumorigenic progenitor cells in syngeneic animals. Consequently, these cells were shown to elicit a curative anti-tumor immune response with tumor regression at distal sites.
C6 glioma cells transfected with an IGF-I TH vector displayed morphological changes, upregulation of MHC class I antigens and B7 antigen, followed by apoptosis similarly to the IGF-I antisense transfectants (Trojan et al., 1996;Ly et al., 2000). Moreover, they increased expression of the protease nexin I (Shevelev et al., 1997). Dramatic inhibition of tumor growth occurred in nude mice following injection of the transfected C6 cells (Shevelev et al., 1997). Similar results of the IGF-I TH strategy were obtained using a syngeneic model of PCC3 derived mouse teratocarcinoma (Ly et al., 2000). We have concluded that the IGF-I TH strategy can parallel the antisense approach and should be very useful in anti-tumor gene therapy.
The role of both B-7 and MHC-I antigens in the induction of T cell immunity against tumors has been extensively investigated (Linsley et al., 1990;Freeman et al., 1991;Chen et al., 1992;Harding et al., 1992;Guo et al., 1994). The explanation of B-7 appearance in the IGF-I TH transfected cells would be as follows: in our work, the transfected cells were growing in a culture medium containing a high concentration of a fetal calf serum (15-20%), while non transfected cells were maintained in a low concentration (5-8%). This could lead to a higher activation of IGF-I receptor (a tyrosine kinase); IGF-I and -II present in a fetal calf serum, as well as intracellular IGF-II act via the type-1 IGF-I re-ceptor (Baserga, 1995;Lafarge-Frayssinet et al., 1997). There is a relation between the signal transduction pathway of a tyrosine kinase and the induction of B7 molecules (Schwartz, 1992;Satoh et al., 1995;Angelisova et al., 1996); the enhancement in B7 co-stimulation through a cAMP mechanism linked to the tyrosine kinase activity of the CD28 receptor has been demonstrated (Schwartz, 1992).
As to the MHC-I expression, down-regulation of MHC-I due to the action of IGF-I has been reported for experiments with rat thyroid cells (Saji et al., 1992). This would be in agreement with the results reported here concerning the inverse correlation between IGF-I and MHC-I protein expression in glioma cells. Moreover, using tumor cells transfected with a IGF-I TH vector, we found increased level of TAP-1 and -2 in these cells, explaining also the presence of MHC-I .
The mechanism of apoptosis is related to the receptor of IGF-I (a tyrosine kinase), itself related to phosphorylation of IRS-1 (insulin receptor substrate) (D'Ambrosio et al., 1996). For this reason different researchers have tried to stop the apoptotic effect using the antisense approach to IGF-I receptor (Sell et al.,1993;Baserga et al., 1994;Resnicoff et al., 1994;Valentinis et al., 1994). In glioma cells, the absence of IGF-I, caused by IGF-I antisense or triple-helix technologies, is associated with massive apoptosis (Ly et al., 2000;. There is a relationship between the immune process, related to the MHC-I or the HLA system (Blanchet et al., 1992), and the apoptotic process -both phenomena simultaneously increasing or decreasing in IGF-I TH transfectants . Recently it was demonstrated that dendritic cells which are involved in tumor-immunogenicity mechanisms by activation of lymphocytes CD8 in the context of MHC-I, recognize apoptotic cells (Matthew et al., 1998).
The IGF-I TH technology was also investigated using the model of mouse hepatoma (Upegui-Gonzalez et al., 2001). The IGF-I TH transfected hepatoma cells also stopped pro-ducing IGF-I, and recovered MHC-I expression accompanied by apoptosis but were down-regulated in the production of IL-10 and TNF-a. Injection of the transfected cells into mice bearing hepatoma at terminal-phase significantly prolonged their survival. The results suggest that injection of hepatoma cells transfected using the TH approach could constitute a vaccine against hepatoma. To our knowledge, we are the first to obtain in vivo results with RNA-DNA triple helix produced after plasmid vector transfection in vitro. The TH strategy for clinical gene therapy of tumors is currently being introduced in University Hospitals of Krakow (digestive tube cancer and liver cancer) and of Bydgoszcz (glioblastoma), Poland. Our IGF-I TH treatment is registered by the international Wiley Gene Therapy Databases -No. 635 and 636 (Gene Therapy Clinical Trials, updated September, 2001 by the Journal of Gene Medicine). The first clinical results obtained with glioblastoma and colon cancer are very promising. The results related to IGF-I TH treatment of colon cancer (five cases to date) are undergoing evaluation and will be published next year. For comparative purposes we have also explored the possibility of using our TH strategy for the therapy of other tumors expressing IGF-I such as breast, prostate and ovary cancers. The latter cancers will be treated at the University Hospital of Bydgoszcz in 2003. Vol. 49 IGF-I, diagnostic and gene therapy of glioblastoma