Ribozymes as Therapeutic Agents: Are We Getting Closer?

The biology of ribozymes has come far since the initial seminal observations of self-cleaving molecules in Tetrahymena (1) and ribonuclease P (2). With the discovery of the hammerhead ribozyme (3) came the realization that it could be utilized to potentially disrupt the expression of any target gene in trans (i.e., acting from outside the target gene sequence) (4). As different types of catalytic RNAs have been discovered [reviewed in (5)], the biologic utility of ribozyme-mediated genetic manipulation has been explored. Ribozymes are being increasingly utilized to suppress target-specific gene expression. The major attraction of ribozymes is their ability to cleave their target messenger RNA (mRNA) in vitro, their specificity (ribozymes have been shown to discriminate among targets that differ by a single nucleic acid substitution), and, as their name implies, their potential to act as an RNA enzyme, thereby cleaving several target RNAs (6). The burden is to demonstrate in every case and with every new target that ribozymes can live up to this potential. The utility of ribozymes in the realm of target validation has been well described. Hammerhead and hairpin ribozymes have been used to inhibit genes involved in the phenotypic expression of human immunodeficiency virus (7,8), drug resistance to cancer chemotherapy agents (9,10), activated oncogene-mediated malignant transformation (11,12, among others), and the angiogenic stimulus (13). In this issue of the Journal, Yamazaki et al. (14) extend the utility of ribozymes for target validation in human cancer. They demonstrate that a hammerhead ribozyme targeting an aberrantly spliced epidermal growth factor receptor (abEGFR) transcript (found in a subset of gliomas) is capable of cleaving its target in vitro. Stable transfectants of an abEGFR-transformed 3T3 cell line (termed ERM5-1) expressing the ribozyme under the control of the b-actin promoter had an altered morphology but did not exhibit reduced cell growth in vitro. In nude mice, the tumorigenicity of ERM5-1 cells was apparently unaffected by ribozyme expression. However, the growth of ribozymeexpressing cells was reduced compared with control cells as evidenced by tumor size at day 19. The only control utilized was a plasmid containing a disabled ribozyme with a mutation in the catalytic subunit, shown to be devoid of catalytic activity using in vitro cleavage assays. Ribozyme expression was shown to decrease abEGFR mRNA levels in vitro. Finally, the authors posited a potential mechanism of action for the anti-abEGFR ribozyme, namely, diminished mitotic activity, by demonstrating reduced bromodeoxyuridine incorporation in resected ribozymeexpressing ERM5-1 tumors. This article is intriguing because the investigators examine a new class of potentially tumor-selective targets, namely aberrantly spliced transcripts, for ribozyme targeting. In this regard, these studies are similar to those performed using ribozymes to inhibit the chimeric bcr-abl transcript (12) created by the translocation between chromosomes 9 and 22 to yield the Philadelphia chromosome in chronic myelogenous leukemia. One major concern that should be addressed in future studies is the potential activity of the ribozyme against the normal EGFR transcript or in parental 3T3 cells. In the case of the bcr-abl transcript, the ribozymes against the breakpoint also suppressed the normal bcr message (15), raising concerns about potential toxicity if expressed in normal cells. Another important concern about the efficacy of ribozymes is the relationship between level of expression and change in the biologic phenotype. Given that several clones were analyzed following transfection and selection, it would have been interesting to attempt to correlate level of ribozyme expression with tumor size. While quantitating ribozyme expression by reverse transcription-polymerase chain reaction (RT-PCR) can at times represent a daunting task, ribozyme expression can also be analyzed by northern analysis since in plasmid-based systems, ribozymes are transcribed as part of a larger RNA molecule. As a corollary to the aforementioned concern, there is no determination of ribozyme cleavage. This can be difficult to achieve for technical reasons, even though cleavage products have been demonstrated by both RT-PCR and northern analysis [reviewed in (16,17)]. In some (although not all) cases, cleavage has been shown both in vitro and in cellular extracts (16,17). There is presumptive evidence of cleavage in vivo, since the disabled ribozyme was inactive, suggesting that the phenotypic reversion was accomplished by the catalytic potential of the ribozyme. This point is of special concern to antisense researchers, who frequently question the added benefit of ribozymes compared with antisense. In several systems where these two modalities have been compared directly, as one would expect, there is a range of findings from no added benefit to spectacular differences in the biologic phenotype [reviewed in (16)]. In the model systems where ribozymes have been shown to have an added benefit, an antisense activity is frequently present but to a lesser extent. Depending on the target gene's specifications (half-life of mRNA and protein) and the biologic system being investigated, this added benefit may translate into a partial or a complete reversal of the phenotype in question when compared with the antisense controls. Beyond target validation, ribozymes are being increasingly investigated for their therapeutic potential. Indeed, retroviral delivery of a hairpin ribozyme is being explored in a phase I trial in patients with human immunodeficiency virus. As potential therapeutics, several concerns remain, whose discussion is beyond the scope of this editorial. Briefly, these pertain to problems of delivery, specificity, and expression. If ribozymes are delivered as oligonucleotides, then problems encountered by antisense molecules may be of import, such as nuclease sensitivity, inaccessibility to target genes due to cellular compartment localization and sequestration, and nonspecific effects. Nonetheless, a pharmacokinetic examination of exogenously delivered ribozymes in vivo has been recently performed (18). Finally, with respect to optimizing ribozyme catalysis, the use of combinatorial chemistry and in vitro selection techniques (19,20) may lead to the development of ribozymes with enhanced catalytic potential. This may also assist with flanking sequence optimization, another area in which divergent results have been reported in the literature. If ribozymes are delivered as gene therapeutics, then the problems plaguing current gene transfer techniques come to the fore. The concerns of efficient gene transfer and high-level expression are especially relevant to ribozyme/antisense strategies because an excess of antigene over gene is required to observe efficacy. Also, as a cancer therapeutic, sustained inhibition of target gene expression will likely be desirable, requiring repeated systemic drug administration. While some preliminary results have been demonstrated using adenovirus-mediated delivery of ribozymes to tumor cells (21,22), further advances in the gene transfer field will likely be required to advance ribozymes as a therapeutic agent. Finally, what makes studies such as the study by Yamazaki et al. (14) elegant scientifically (i.e., the tumor-selective nature of the target) reduces their potential as a broad-based therapeutic agent, since such mutational events occur in a subset of patients with a particular cancer. It is hoped that future research will be able to solve the problems of delivery, expression, and specificity with broad applicability to advance ribozyme therapeutics into the clinical arena.