Observations Regarding the Effect of Targeted Gene Deletions (Knockouts) on Graft Rejection

The field of immunology has become large and complex so that no one person can hope to master every aspect fully, and it is often necessary to focus on one small area in extreme detail. But the risk of such narrow specialization is that wider observations may be missed, so it may be fruitful occasionally to “step back” and view the overall picture. This may be relevant to the recent use of gene knockout mice to investigate the immune response. This technological advance has brought much new information, and, without a doubt, has greatly enhanced our understanding of the immune system. Yet some aspects of those knockouts, which have targeted crucial molecules within the immune system, seem particularly perplexing. The most remarkable example is the investigation of the role of cytokines and costimulatory molecules in graft rejection. In vitro studies have established the importance of these molecules in the T-cell-mediated immune response, and they form the mechanism for our understanding of the currently favored “Danger” paradigm, as proposed by Polly Matzinger. However, the pace and severity of the graft rejection response in vivo is little changed by most knockouts of cytokine or costimulatory molecules in the recipient, which is in strong contrast to the effect of monoclonal antibody therapy targeted on the same molecule (which hereafter will be termed the “specific blocking approach”) (Table 1) {tabft}Abbreviations: “−”, minor or nil effect; “+”, weak effect; “++”, moderate effect; “+++”, strong effect. 1). This striking finding has not gone unnoticed and the concept of “redundancy” within the immune system has been put forward to explain the results (1).Table 1: Comparison of the effect on rejection of costimulatory molecule gene knockout in the recipient versus antibody blockadeIt is certainly true that multiple costimulatory molecules and cytokines with similar actions have been described, so the argument is feasible, but how does it actually work? The obvious answer would be that there are parallel pathways and molecules that can take over the function of a pathway that is blocked. But if there is such easy redundancy, why is it inoperative when monoclonal antibodies are used to specifically inactivate just one cytokine or accessory molecule? An obvious explanation would be that monoclonal antibodies are more efficient than knockout techniques at actually reducing the effective concentration of the target molecule. There certainly are some knockout models that are “leaky,” for example β2-microglobulin knockout for MHC Class 1. However, the knockouts of costimulatory molecules and cytokines generally have not been “leaky,” according to published analysis of specific expressed RNA or protein. A second possibility is that monoclonal antibodies somehow have an extra effect on graft rejection, mediated by an unknown pathway, perhaps related to the fact that they are immune molecules themselves. This is hard to disprove, but against this is the fact that, where cytokine production can be blocked using drugs (for example calcineurin blockers, such as cyclosporin and tacrolimus blocking IL-2 production), this is also as effective as monoclonal therapy (Table 1), although, admittedly, cytokines other than IL-2 are also affected by these drugs. Furthermore, although some knockouts such as CD3 and CD4 do have profound effects on immune function, knockouts of a whole variety of other immune-related molecules show a similar finding of little or no effect on graft rejection, although there are fewer direct comparisons with antibody blockade (Table 2). It is interesting that, although direct comparisons are relatively few, knockout of a molecule in the donor is often as effective, or even has more effect, than knockout of the same molecule in the recipient (Table 3).Table 2: Comparison of effect on rejection of various gene knockouts in recipients versus antibody blockadeTable 3: Comparison of effect on rejection of gene knockout in donor versus recipientThe surprising outcome of the knockout studies may be revealing a fundamental point about the development of the normal immune system. The one major difference between the knockout and the specific blocking approaches is the presence of the defect during the period when the immune system is developing. An explanation for these findings would be the proposal that, during normal development, the immune system is being “set” to a “normal immune response,” which is effectively the set of signals that the immune cell expects to receive when its receptor makes contact with activating cognate ligand. For this to occur, the immune system makes use of all the immune related cells, molecules, and signaling pathways available, perhaps by running a form of “test run” immune response. Where exactly would this process be likely to occur? Because the thymus is the origin of the T-cells, which are the major players in cellular immunity, it would be a good candidate site. How would this explain the difference between knockout and specific blocking approaches? Because the immune defect is present during the development of the immune system, the setting of the baseline “normal immune response” effectively cancels out the absence of that molecule. A more specific example is that, in the presence of a normal compliment of costimulatory and cytokine molecules, these give strong signals to T-cells recognizing cognate ligand through their T-cell receptors and will be used as a major parameter to set the level of normal immune response signal. However, in the absence of one or more costimulatory or cytokine molecules, the immune system still goes ahead and sets the normal level of the immune response, this time using whatever other immune molecules are available. Subsequent major immune responses, therefore, will proceed more or less normally. However, it should be noted that this does not mean that absolutely no effect of the knockout will be seen, but that its effect will tend to be less than might be expected from the effect of monoclonal antibodies or immunosuppressive agents used to block the same molecules. These latter agents contrast to knockouts because the immune system develops normally, and when it is subsequently targeted by specific blocking agents, the deviation from the normal level of signaling is major and, consequently, the effect seen is large. The above explanation fits the available data reasonably well, but how to obtain proof and determine the exact mechanism is not clear. The development of “conditional knockouts,” such as those based on the Cre-Lox switch, has been used successfully in several nonimmunological molecular systems, and some groups are now developing this biotechnology to re-examine the role of cytokines and other molecules in immune responses (78). If much greater effects are seen by operating these conditional knockouts in the adult animal than are seen with the same deficiency present throughout development, this provides strong support for the “normal immune response setting” being a real phenomenon. However, if the conditional knockouts prove to be no more effective than permanent knockouts, the above explanation becomes unlikely, and alternative explanation(s) would need to be sought, the nature of which are not, at present, clear to this author. Derek W. R. Gray 1