Antisense oligodeoxynucleotides are being widely used to interfere with specific gene activities in cell-culture systems (1,2), and there are possible analytical advantages to partial and timed interference in the whole embryo, for certain genes, as compared with targeted null mutations ("knockouts"—see e.g., ref. 3). Despite this, and the logical possibility of controlling rigorously for the sequence specificity of effects observed, use of antisense oligos will continue to attract a healthy level of controversy, because of occasional evidence for nongenetically based sequence specificity of particular cellular effects (4), and also the possibility of inadvertently targeting other, unknown genes containing matching base sequences. These matters are further discussed below, and in Subheading 4. The chick blastoderm consists of two or at most three-cell layers at the time many genes with vital roles in development of the basic body plan are being activated, and only one layer, the epiblast, resembles a true epithelium enough to constitute a potential barrier for oligo access to the other two. In fact, there is evidence that epiblast itself is good at uptake of oligos, and it should be remembered that there is evidence for extensive bulk uptake by "cell drinking" (pinocytosis) by ectoderm and neurepithelium, the equivalent cell layer of the young mammalian embryo. This blastoderm can be kept, with or without incubation, in a protein-free nutrient medium for up to 3 h, and then cultured onward in either of two ways. It proceeds, within 15 h, from gastrula/neurula stages to those where the normality or otherwise of body plan formation can readily be assessed. Even relatively transient and incomplete degradation of gene product levels by antisense will thus greatly help in analysis of a some dynamically expressed genes' early roles, and supplement "knockout" analysis in mice. Thus, control over the timing and possibly degree of interference might enable analysis of genes where the mouse null mutant develops insufficiently for much to be said about what went wrong, or halts development at a stage precluding analysis of separate and slightly later developmental functions of the same gene. Furthermore, mammalian genes are documented for which acute antisense interference in cell culture produces a functional deficit that cannot be occurring in the corresponding cells of the null mutant whole embryo, whose phenotype is much less severe (5,6). This suggests functional compensation mechanisms during the prolonged sequence of development, whereby a gene's normal role is masked.
Antisense oligos exert their functional interference by binding to their target sequence on nascent or working messenger RNA, inhibiting its translation, causing degradation through RNase H, or having both these effects. There is consensus that the half-life of normal phosphodiester DNA oligos (P-oligos), both in culture media and within the cell, tends to be minutes long and is thus inadequate to perturb levels of even the most rapidly turned over gene products enough for developmental effects. A variety of chemical modifications to the DNA structure have therefore been developed, which cause resistance to the exonucleases—widespread in serum and within cells—that degrade short DNA stretches (preferentially from the 3'-end). Of these, the currently most widely used are phosphorothioated DNA oligos (S-oligos), in which sulfur is substituted for one of the oxygen atoms in the phosphate (sugar) backbone (7,8). It is believed that their half-life in biological situations is lengthened by more than an order of magnitude, that they enter cells relatively readily though maybe by different mechanisms from natural DNA, and that their stability is traded off against somewhat less binding stability for a given DNA-RNA hybrid sequence, and possibly diminished ability of the hybrid to act as substrate for RNase H. If the last is the case, such oligos will act only by allowing decay of protein levels through nonreplacement by translation, rather than also acting as a catalytic (recycling) degrader of message within the cell. The literature reports successful cases of sequence-specific gene interference with and without the specific RNA degradation following their use (9,10). In our laboratory, most successful work has utilized this modification, which is by far the least expensive currently on the market, in 15-mer antisense sequences. Thus, in Subheading 3., it will be understood that S-oligos are used, but our experience with alternatives that may be better in the future is briefly discussed in Subheading 4.
Our recent work has been with two closely related genes encoding zinc-finger-containing putative transcription factors, with different early developmental expression patterns and apparent functions. We have achieved the following evidence for the genetically meaningful sequence-specificity of the "phenotypes" of developmental disturbance caused by the antisense treatments (see Subheading 4. for further comments):
1. A large number (>30 in all) of other oligo sequences of comparable length and overall base composition, either of random sequence, sense to the same genes, antisense to other genes, or sequence parallel to the active ones but with 6-base internal 5'- to 3'- reversed sectors, are without effect in the procedure at up to twice the normal concentrations used.
2. The "phenotypes" are different and nonoverlapping for the two genes sequence-targeted, and are comprehensible in each case in terms of the normal sites and times of dynamic expression of that gene's mRNA. Thus in one case, a particular cell-adhesive-behavioral transition normal to each expression site is inhibited (3). In the other, somite segmentation and left-right anatomical asymmetry of the body are disturbed, the gene being dynamically expressed in nascent somites and, at a separate site, with striking left-right asymmetry (11).
3. Antisense to sequences at two separate, comparable positions within the coding sequence give the same distinctive "phenotype" in the case of each gene. At least in one of the cases, there is a strong interactive effect such that a given micromolarity of antisense oligo gives much stronger interference if composed of equal amounts of the two sequences than if given as either sequence alone. This would theoretically be expected under conditions where each oligo species was independently compromising a high proportion of all mRNA molecules at a particular time.
4. Direct evidence for gene-specific mRNA degradation in the hours following treatment has been obtained in the case of one but not both of the genes, whereas none of the oligos involved appear to lessen mRNA amounts generally in the cultured embryos. The literature suggests that specific RNA degradation in the case of antisense-active phosphorothioate oligos is not necessarily to be expected (9).
Taken together, particularly with inclusion of item 3, we regard these as rather powerful evidence for the biological meaningfulness of the effects. Demonstration of specific degradation of gene product at either RNA or protein level is always valuable if the necessary reagents are available. However, insistence on it before work is accepted would seem unnecessarily restrictive, in that only genes on which much resource-intensive work has already been done (quantitative PCR or raising of an appropriate antibody) could then be considered for the approach. In itself, the approach can be used as a relatively low-cost screen for interest, applied to any new chick gene with a dynamic "early developmental" expression pattern ($600 or £400 oligo-related costs/ sequenced gene, for 10 experiments involving 60 experimental and 60 control blastoderms, in labs equipped for chick culture). The main overall evidence for validity of antisense-specific effects is that the great majority of oligo sequences, including those that are sense to known genes, fail to affect devel opment. It is known that a statistical majority of sequences within any one gene will be ineffective targets for antisense (e.g., 9), and the question of design of effective sequences is alluded to in Subheading 4.
Subheading 3. describes:
1. The preparation of ring-culture setups for the chick blastoderm by a method modified from the original owing to New (12) and the removal of blastoderms into protein-free medium with optimal physiological treatment of the embryos.
2. The administration of phosphorothioated oligodeoxynucleotides with adjunct use of a lipofection procedure, which we believe to enhance intracellular availability (see Subheading 4. for possible alternative improvements to availability).
3. Replacement of blastoderms into ring culture for development of up to 20 h longer, leading to potentially normal development of the 10-15 somite stage.
4. Replacement of blastoderms from step 2. into an alternative form of culture (13), allowing good development over 36 h to the 30+ somite stage (early limb buds and considerable craniofacial morphogenesis), as well as the possibility of continuing presence of oligos, thus maintaining gene interference, in the medium.
Was this article helpful?