Introduction

Gene therapy was first carried out in patient by Anderson in 1990 to treat a girl with SCID (Severe Combined Immunodeficiency Disease) that results from the deficiency of the ADA gene. To date more than 1000 gene therapy protocols have been used in clinic, among which 63-67% were for cancer treatment. Nevertheless, significant breakthrough has been lacking in the field. Viral therapy was used about a hundred years ago when some cancer patients got remission after getting virus infection. In 1970, there were 38 different viruses being used in patients. Now many viruses—especially the modified viruses—are formally used in clinic, such as G207 derived from HSV-1, CV706 from Ad5, CV787 also from Ad5, HSV-1716, NV1020 from HSV-1/HSV-2, PV701 from NDV and vaccinia GM-CSF. The term virotherapy was suggested in 2001 by Kirn 1. For instance, when used in combination with 5FU and cisplatin, the ONYX-015 led to a 63% anti-tumor effect in a phase II trial 2, but the anti-tumor efficacy was only 15-20% when ONYX-015 was used alone 2. ONYX-015 had passed phase II and entered phase III clinical trial, but the trial was eventually stopped. Generally, there has also been no big breakthrough in virotherapy.

The ability to specifically target cancer cells (Targeting) would be crucial for a successful cancer therapy. To this end, we have initiated a strategy, designated as Targeting Gene-Virotherapy of Cancer, by taking the advantage of both gene therapy and virotherapy 3. Our first targeting virus constructed was ZD55, which was derived by deleting the E1B 55kDa gene from adenovirus 4. ZD55 is similar to ONYX-015 in its ability to selectively target and infect p53 negative (or p53 signaling pathway defective) cancers, but ZD55 was significantly different from ONYX-015 in many other features. In addition to differences in the construction method, our ZD55 was derived from Ad5 while ONYX-015 was from the fusion virus Ad2/Ad5; furthermore, ZD55 contained a cloning site designed to insert foreign gene(s), whereas no cloning site was available in ONYX-015. Our strategy of Targeting Gene-Virotherapy can be divided to three stages.

Using one therapeutic gene

The first stage of our strategy involves the use of one therapeutic gene (Targeting Gene-Virotheapy of Cancer, patent No. 02157662.9). We cloned many genes individually into ZD55 to form ZD55-gene(s) and examined the anti-cancer efficacy of the resulting ZD55-gene(s). A number of papers have been published from our group in recent years on different gene(s), such as CD 4, sFlt11, 2, 3, (an inhibitor of VEGF consisting of the first three extra cellular domain of soluble Flt-1, the hVEGF receptor-1) 5, TRAIL (tumor-necrosis factor-related apoptosis-inducing ligand) 6, K5 (kringle 5 of plasminogen) 7, Smac (Second mitochondria-derived activator of caspases) 8, XAF-1 (XIAP associated factor-1) 9, MnSOD 10, IL-24 11, and hSSTr2 (somatostatin receptor gene subtype 2) 12, etc. We have recently expanded this strategy to include a gene coding for an siRNA against X-IAP (X-linked inhibitor of apoptosis) (submitted). The anti-tumor effects of all these ZD55-gene constructs are much better than the respective gene therapy or virotherapy alone as shown in Figure 1 for ZD55-TRAIL 7. Another special case is ZD55-IL-24 (patent No. 2005 10026151.5). Its anti-tumor effect is about 100-fold more potent than ZD55 or ONYX-015, or the gene therapy product Ad-IL-24 11 which has entered phase II clinical trial in USA 13. We think that the following reasons may account for the success of our strategy. ZD55 is a replicative and targeting vector, which can amplify several hundred to several thousand times in the targeted tumor cells while its carried gene(s) can also simultaneously be amplified several hundred to several thousand times. These characteristics of ZD55-gene overcome the disadvantage of traditional gene therapy involving a replication-deficient adenovirus that results in lower expression of the therapeutic gene with no targeting effect, and also improve the killing effect of traditional virotherapy (such as ZD55 or ONYX-015). It is our prediction that a future trend of gene therapy or virotherapy for cancer will be the strategy of Targeting Gene-Virotherapy. We first proposed the term “Targeting Gene-Virotherapy”, and our group in recent years made major efforts to extensively test this strategy; although there were a few previous reports on delivering therapeutic genes using oncolytic viruses, such work on ONYX-015 was restricted with the tk or cd gene only 14, 15.

Figure 1
figure 1

Enhanced anti-tumor effect of targeting gene-virotherapy (ZD55-hTRAIL) in vivo. Tumors were established in nude mice by implantation of SW620 cells. When tumor size reached 60-80 mm3, animals were treated by intratumor injection of PBS or Ad-hTRAIL, ZD55-GFP, ONYX-015, or ZD55-hTRAIL at 2×108 pfu/animal daily for 5 consecutive days. Tumor size was measured every 5 days and is presented as mean ± SD (n=6). The inhibitory effect of ZD55-hTRAIL at the end of 50 days on tumor growth was better than that of virotherapy alone (P<0.05, versus ONYX-015 or ZD55-GFP) and control animals receiving PBS (P<0.001). Adapted with permission from Mol Ther 2005; 11:531-541 7.

Full size image

Using two therapeutic gene

The second stage of our strategy involves the use of two anti-tumor genes (Targeting Dual Gene-Virotherapy of Cancer, patent No. 0210569.9). Although a lot of work has been done in our lab for ZD55-genes and in some cases the tumor mass can be completely eliminated in some mice, the use of any single ZD55-gene could never be able to eliminate all the tumors in all mice. Therefore we developed the Dual Gene strategy. We reason that the use two therapeutic genes together may have additive, complementary or synergetic effect, leading to a much stronger anti-tumor effect than using one gene alone. For example, K5 is a strong anti-angiogenesis factor and TRAIL is a very strong apoptotic factor. When ZD55-TRAIL and Ad-K5 and were used together, their complementary effect led to the complete elimination of the SW620 colorectal tumor xenograft in all examined nude mice as shown in Figure 2 7.

Figure 2
figure 2

The superiority of targeting dual gene-virotherapy strategy. Anti-tumor effect of targeting dual gene-virotherapy by combination of ZD55-hTRAIL and Ad-k5. When SW620 tumor size reached 70–100 mm3, animals were treated by intratumor injection of PBS or Ad-k5 (2×108 pfu daily) or ZD55-hTRAIL (2×108 pfu daily) alone or a combination of ZD55-hTRAIL (1.8×108 pfu daily) and Ad-k5 (0. 2×108 pfu daily) for 5 consecutive days. Tumor size was measured every 5 days and is presented as mean ± SD (n=6). Adapted with permission from Mol Ther 2005; 11:531-541 7.

Full size image

We have performed a number of studies using this strategy of Targeting Dual Gene-Virotherapy of Cancer. Our main results can be summarized as follows. The combination of ZD55-TRAIL and Ad-K5, or ZD55-TRAIL and ZD55-MnSOD, or ZD55-TRAIL and ZD-IL-24, could completely eradicate all SW620 colorectal cancer xenograft in all nude mice 7, 10, 16. ZD55-TRAIL + ZD55-Smac could completely eliminate all BEL7404 hepatoma xenograft in all nude mice 8. ZD55-TRAIL + ZD55-hSSTr2 could completely eliminate all BxPC-3 pancreatic adenocarcinoma xenograft in nude mice 12. AAV-hTERT-TRAIL + AAV-hTERT-IFN-β could completely eliminate all A549 lung cancer xenograft in all nude mice (submitted).

Among the above reported cases, ZD55-TRAIL proves to be an important drug which can be combined with many other ZD55-genes to completely eradicate all the xenograft cancers. TRAIL is a good anti-tumor factor, as it shows no toxicity to liver when it is expressed in a targeting vector under the hTERT promoter 17 and it possesses the bystander effect 18 which is very important for the killing all tumor mass. Thus, it might be desirable to further test ZD55-TRAIL on a larger scale. We plan to conduct phase I/II clinical trials for ZD55-IL-24 in 2007 and for ZD55-TRAIL in 2008. It is possible that the combination of ZD55-TRAIL and ZD55-IL-24 will lead to a very good protocol for the treatment of cancer patients, since together they could completely eliminate all the colorectal cancer xenograft as shown in Figure 3 16. The mechanism of their action was shown in Figure 4 16. It is our hope that the Targeting Dual Gene-Virotherapy strategy will produce excellent clinical therapeutical effect for patients in the future. However, it should be noted that delivering two genes in the replication-deficient vector under conventional gene therapy was rarely found to completely eliminate all the xenograft tumor mass in all nude mice.

Figure 3
figure 3

Complete eradication of human SW620 xenograft tumor in nude mice by the co-administration of ZD55-IL-24 and ZD55-TRAIL. When tumor size reached 100-150 mm3, subcutaneous tumor-bearing mice were divided into four groups and treated with four consecutive daily intratumoral injections of PBS or with ZD55-IL-24, ZD55-TRAIL and the combination at 5×108 pfu/dose per day (treatment indicated by arrow). (A) The tumor size was measured using calipers and tumor volume was calculated. Data are presented as means of tumor volume ± SD. (n=8). (B) The death of animals was monitored. Long-term survival of animals was observed after treatment with oncolytic adenoviruses compared with control animals receiving saline. Adapted with permission from Cancer Gene Ther 2006; 13:1011-1022 16.

Full size image
Figure 4
figure 4

Immunohistochemical staining and TUNEL assay. (A) SW620 tumors subcutaneously receiving various treatments were harvested 3 days after infection and tumor sections were analyzed for the IL-24 (a–c) and TRAIL (d–f) expression by immunohistochemistry (400 original magnification). The stained signals are indicated with a brownish black arrow, and no expression of IL-24 or TRAIL was seen in mock infection (a, d). The IL-24 expression in the tumor sections was observed in the case of the tumors injected with ZD55-IL-24 (b) or with ZD55-IL-24 and ZD55-TRAIL (c). Similar results were also obtained for the detection of TRAIL expression (e), ZD55-TRAIL infection; (f), ZD55-TRAIL and ZD55-IL-24 infection. (B) TUNEL assay was also performed to detect apoptosis in the sections of tumors that received different treatments. (a) tumors treated with PBS as control; (b) tumors treated with ONYX-015; (c) Tumors treated with ZD55-TRAIL; (d) tumors treated with ZD55-IL-24; (e) tumors treated with ZD55-IL-24 and ZD55-TRAIL. Arrows denote cells undergoing apoptosis. Percentage of TUNEL-positive cells to all cells is shown for each of the indicated treatments. Data are presented as mean ± SD. (**P<0.05; **P<0.01). Adapted with permission from Cancer Gene Ther 2006; 13:1011-1022 16.

Full size image

Two therapeutic genes can also be linked together and be cloned into the same viral vector, but there are several potential issues with this approach. First, it is not yet clear whether our SFDA would approve this kind of product with two linked genes for clinical use. The second issue is how to select the appropriate linker to join the two genes into the same vector. IRES and (glyglyglyglyser)3 are considered out of date for use as the linker. For example, the expression level of gene B in the construct of A-IRES-B is usually 5-10 times lower than that of gene A. We are currently exploring the possibility of harnessing the self-cleavage property of the 2A peptide from FMDV or using the four residue IETD motif as the TRAIL linker. For example, a TRAIL-IETD-Smac fusion will be specifically cleaved at D↓ of the IEPD motif by caspase 8 whose activity can be induced by TRAIL itself. Our animal studies have already shown that two anti-tumor genes delivered in the targeting vector (such as ZD55-gene 1 and ZD55-gene 2) are usually sufficient to kill all the tumor mass if the appropriate genes with complementary or synergetic effect were chosen. Given this, it seems not necessary to pursue the combinatorial use of three genes as it would be practically much more difficult and/or costly to achieve.

Using two targeting promoters and two therapeutic genes

The third stage of our strategy involves the use of two targeting promoters and two anti-tumor genes, and we designate it as Double Controlled Targeting Virus-Dual Gene Therapy of Cancer. In this case, the eventual tumor-targeting oncolytic viruses were constructed by the use of two different tumor-specific promoters to control the essential viral genes, thus resulting in greater specificity against the targeted tumors.

For example, the use of hTERT (human telomerase reverse transcriptase, which is up-regulated in about 90% of tumors) promoter to control adenovirus E1A results in Ad-hTERT, which could target a broad range of cancers such as hepatoma or breast cancers (Figure 5) as shown in our previous paper 19; this is in contrast to the PSA promoter which can only target prostate cancer. The Ad-hTERT is nevertheless still a targeting virotherapy reagent. After insertion of an anti-tumor gene as TRAIL to form Ad-hTERT-TRAIL, it became Targeting Gene-Virotherapy with a more potent anti-tumor effect 20, 21, 22, 23, 24, 25. To take advantage of features of both the Ad-hTERT vector and the ZD55-vector, a hybrid vector consisting of Ad-hTERT-E1A-ZD55 (E1B) was constructed and was named as TD55. Cloning of the TRAIL gene into TD55 generated TD55-TRAIL. Interestingly, both TD55 and TD55-TRAIL demonstrated better safety for normal cells such as MRC5 and WI38 (Figure 6); and TD55-TRAIL showed good anti-tumor effect against SW620 or A549 tumor cells (Figure 6) 26. TD55-TRAIL can be considered as a vector for Double Controlled Targeting Virus-Gene Therapy, as it delivers a therapeutic gene (TRAIL) under a double-controlled virus (E1A is controlled by the hTERT promoter, and the deletion of E1B offers another control).

Figure 5
figure 5

(A) Construction of Ad-hTERT-Gene for anti-tumor therapy. (a) In wild type Adenovirus, E1A is controlled by its own promoter. (b) hTERT promoter was used to replace the E1A promoter to form Ad-hTERT. (c) An anti-tumor gene (such as Trail) expression cassette was inserted into Ad-hTERT to form Ad-hTERT-gene. (B) Anti-tumor effect of Ad-hTERT for Hepatoma BEL 7404. (C) Anti-tumor effect of Ad-hTERT for Breast cancer Bcap 37. Adapted with permission of Ai Zheng 2006; 23:385-392 26.

Full size image
Figure 6
figure 6

(A) Replication of TD55 and TD55-TRAIL in MRC5 and WI38 cells. (B) Selective effects of TD55 and TD55-TRAIL on SW620 and A549 cells, Adapted with permission of Ai Zheng 2006; 23:385-392 26.

Full size image

After consideration of all the above data, we are interested in further developing a new strategy, Double Controlled Targeting Virus-Dual Gene Therapy, for the treatment of individual cancers. Some specific protocols will be designed. There are three crucial points that need to be considered, taking the hepatoma as an example. First, there should be double cancer-specific promoters to control the viral replication. For example, the E1A of adenovirus can be controlled by the hTERT promoter and the E1B can be controlled by the AFP (α-fetoprotein) promoter, a promoter specific for liver cancer. The resulting adenovirus would be a double tumor promoter controlled vector of Ad5-hTERT-E1A-AFP-E1B which is expected to have higher cancer specificity. Second, the expression of a specific cancer suppressor gene, for example, the liver cancer suppressor gene HCCS1 (or LFIRE), can be controlled by the AFP promoter to generate a construct such as Ad-hTERT-E1A-AFP-E1B-HCCS1 (or LFIRE). In a preliminary study ZD55-HCCS1 has shown a very good anti-tumor effect (data not shown). It is a specific and individualized therapy for hepatocellur cancer (HCC) by using two cancer specific promoters, with one HCC specific promoter AFP to direct the expression of a hepatoma suppressor gene (HCCS1 or LFIRE). We speculate that constructs like Ad-hTERT-E1A-AFP-E1B-HCCS1 will also have very good anti-tumor effect. Third, if constructs such as Ad-hTERT-E1A-AFP-E1B-HCCS1 (or LFIRE) could not completely eliminate all the hepatoma xenograft, we will generate additional constructs with predicted good anti-tumor effect such as Ad-hTERT-E1A-AFP-E1B-IL-24. The combined use of Ad-hTERT-E1A-AFP-E1B-HCCS1 (or LFIRE) and Ad-hTERT-E1A-AFP-E1B-IL-24 would realize the idea of Dual Gene therapy strategy as discussed above; and it is predicted that all the BEL-7404 hepatoma xenograft would be completely eliminated by this combination approach. Similarly, the colorectal cancer could be treated by the use of colorectal cancer specific suppressor gene such as ST13 and the gastroenterological promoter CEA. All the above protocols are currently being studied. The same principle can be also applied to lung cancer, prostate cancer etc. The safety of a double controlled targeting vector has been confirmed (Figure 6) 26. In the above discussion I have outlined the general principle for conducting Double Controlled Targeting Virus-Dual Gene Therapy of Cancer. It is conceivable that changes and modifications can be made to meet the needs of specific procedures. For the example, the alternative choice for the targeting control promoter is the surviving promoter (SurP) or the hTERT-E1A (Δ24), both of which were better than hTERT 27, 28; and for the killer gene, one can use TRAIL, MnSOD or others instead of IL-24. In summary, it is expected that the strategy of Double Controlled Targeting Virus-Dual Gene Therapy of Cancer could be used to design many excellent protocols to achieve ideal therapeutic effect for many individual cancers, and we hope that its successful implementation will lead to a cure for at least some cancers in the not far future.

Finally, it is worthwhile to consider our strategy in the context of the recent progress in cancer therapy. There are three critical issues in cancer therapy. First, early detection remains a key, as reflected by Laura Spinney's recent writing in Nature 29 (August 17, 2006), “The detection of cancer at an early stage in its development can be life saving”. Second, cancer stem cell is the cause of tumor recurrence; therefore it is a root problem for cancer therapy to kill the cancer stem cell. Many cancer stem cells have been discovered including those in leukemia 30, breast cancer 31, brain tumor 32, and prostate cancer 33, 34; and these have been extensively reviewed elsewhere 35, 36, 37, 38. Third, targeting therapy of cancer is another key. To develop drugs that target and kill cancer cells without impairing the normal cells is an important problem in cancer therapy. Recently, a Xiangshan conference (similar to Gorden conference in USA) on Targeting Therapy of Cancer was held (March 22-24, 2005) in China 39. There are a number of directions and approaches to target cancer cells including, for example: 1), antibody therapy (24 antibodies have been licenced to market among which 8 are anti-tumor antibodies such as Herceptin and Avastin); 2), antibody gene therapy 40, 41 which offers an attractive alternative to the traditional antibody therapy 42; 3), targeting gene-virotherapy, which can completely eliminate xenograft tumors in nude mice by the use of two appropriate therapeutic genes 7; 4), targeting cancer stem cells; 5), targeting an intracellular protein tyrosine kinase (such as Gleesvec, which was selected as one of the 10 greatest science contributions in 2001); and 6), targeting the EGFR Kinase 43. It is of no doubt that progress in our ability to target cancer will continue to contribute to better cancer therapy.