Femke van den Berg

A recently described splice variant of CD44 expressed in metastasizing cell lines of rat tumors has been shown to confer metastatic potential to a non-metastasizing rat pancreatic carcinoma cell line and to non-metastasizing sarcoma cells. Homologues of this variant as well as several other CD44 splice variants are also expressed at the RNA level in human carcinoma cell lines from lung, breast, and colon, and in immortalized keratinocytes. Using antibodies raised against a bacterial fusion protein encoded by variant CD44 sequences, we studied the expression of variant CD44 glycoproteins in normal human tissues and in colorectal neoplasia. Expression of CD44 variant proteins in normal human tissues was readily found on several epithelial tissues including the squamous epithelia of the epidermis, tonsils, and pharynx, and the glandular epithelium of the pancreatic ducts, but was largely absent from other epithelia and from most non-epithelial cells and tissues. In human colorectal neoplasia CD44 variant proteins, including homologues of those which confer metastatic ability to rat tumors, were found on all invasive carcinomas and carcinoma metastases. Interestingly, focal expression was also observed in adenomatous polyps, expression being related to areas of dysplasia. The distribution of the CD44 variants in human tissues suggests that they play a role in a few restricted differentiation pathways and that in colorectal tumors one of these pathways has been reactivated. The finding that metastasis-related variants are already expressed at a relatively early stage in colorectal carcinogenesis and tumor progression, i.e., in adenomatous polyps, suggests the existence of a yet unknown selective advantage linked to CD44 variant expression. The continued expression in metastases would be compatible with a role in the metastatic process.

A recently described splice variant of CD44 expressed in metastasizing cell lines of rat tumors, has been shown to confer metastatic potential to nonmetastasizing rat pancreatic carcinoma and sarcoma cell lines. Using antibodies raised against a bacterial fusion protein encoded by variant CD44 sequences, we have explored the expression of variant CD44 glycoproteins on human lymphoid cells and tissues and on non-Hodgkin's lymphomas. Normal lymphohematopoietic cells express barely detectable low levels of variant CD44 glycoproteins, whereas T lymphocytes, upon activation by mitogen or antigen, transiently upregulate expression of specific CD44 variant glycoproteins. The reaction pattern of various antibodies indicates that these CD44 variants contain the domain encoded by exon v6, which is part of the variant that in the rat confers metastatic capability. It is interesting that overexpression of v6 was also found in several aggressive, but not low-grade, non-Hodgkin's lymphomas.

Fungicide-resistance management would be more effective if principles governing the selection of resistant strains could be determined and validated. Such principles could then be used to predict whether a proposed change to a fungicide application program would decrease selection for resistant strains. In this review, we assess a governing principle that appears to have good predictive power. The principle states that reducing the product of the selection coefficient (defined as the difference between the per capita rate of increase of the sensitive and resistant strains) and the exposure time of the pathogen to the fungicide reduces the selection for resistance. We show that observations as well as modeling studies agree with the predicted effect (i.e., that a specific change to a fungicide program increased or decreased selection or was broadly neutral in its effect on selection) in 84% of the cases and that only 5% of the experimental results contradict predictions. We argue that the selection coefficient and exposure time principle can guide the development of resistance management tactics.

We have reviewed the experimental and modeling evidence on the use of mixtures of fungicides of differing modes of action as a resistance management tactic. The evidence supports the following conclusions. 1. Adding a mixing partner to a fungicide that is at-risk of resistance (without lowering the dose of the at-risk fungicide) reduces the rate of selection for fungicide resistance. This holds for the use of mixing partner fungicides that have either multi-site or single-site modes of action. The resulting predicted increase in the effective life of the at-risk fungicide can be large enough to be of practical relevance. The more effective the mixing partner (due to inherent activity and/or dose), the larger the reduction in selection and the larger the increase in effective life of the at-risk fungicide. 2. Adding a mixing partner while lowering the dose of the at-risk fungicide reduces the selection for fungicide resistance, without compromising effective disease control. The very few studies existing suggest that the reduction in selection is more sensitive to lowering the dose of the at-risk fungicide than to increasing the dose of the mixing partner. 3. Although there are very few studies, the existing evidence suggests that mixing two at-risk fungicides is also a useful resistance management tactic. The aspects that have received too little attention to draw generic conclusions about the effectiveness of fungicide mixtures as resistance management strategies are as follows: (i) the relative effect of the dose of the two mixing partners on selection for fungicide resistance, (ii) the effect of mixing on the effective life of a fungicide (the time from introduction of the fungicide mode of action to the time point where the fungicide can no longer maintain effective disease control), (iii) polygenically determined resistance, (iv) mixtures of two at-risk fungicides, (v) the emergence phase of resistance evolution and the effects of mixtures during this phase, and (vi) monocyclic diseases and nonfoliar diseases. The lack of studies on these aspects of mixture use of fungicides should be a warning against overinterpreting the findings in this review.

Strategies to slow fungicide resistance evolution often advocate early "prophylactic" fungicide application and avoidance of "curative" treatments where possible. There is little evidence to support such guidance. Fungicide applications are usually timed to maximize the efficiency of disease control during the yield-forming period. This article reports mathematical modeling to explore whether earlier timings might be more beneficial for fungicide resistance management compared with the timings that are optimal for efficacy. There are two key timings for fungicide treatment of winter wheat in the United Kingdom: full emergence of leaf three (counting down the canopy) and full emergence of the flag leaf (leaf 1). These timings (referred to as T1 and T2, respectively) maximize disease control on the upper leaves of the crop canopy that are crucial to yield. A differential equation model was developed to track the dynamics of leaf emergence and senescence, epidemic growth, fungicide efficacy, and selection for a resistant strain. The model represented Zymoseptoria tritici on wheat treated twice at varying spray timings. At all fungicide doses tested, moving one or both of the two sprays earlier than the normal T1 and T2 timings reduced selection but also reduced efficacy. Despite these opposing effects, at a fungicide dose just sufficient to obtain effective control, the T1 and T2 timings optimized fungicide effective life (the number of years that effective control can be maintained). At a higher dose, earlier spray timings maximized effective life but caused some reduction in efficacy, whereas the T1 and T2 timings maximized efficacy but resulted in an effective life 1 year shorter than the maximum achievable.