Why clinicians should be interested in interleukin-3

Interleukin-3 (IL-3), a product of activated immune cells has recently been cloned and introduced in preclinical and clinical trials. The biological target-cell spectrum of IL-3 is broad and includes progenitor cells of various hematopoietic lineages as well as multiple stages of stem cell differentiation. IL-3 also induces growth of most primitive hemopoietic progenitors (CFU-blast). Synergistic effects on growth of myeloid cells (i.e. macrophages, eosinophils and blood basophils) are obtained by sequential use of IL-3 and later-acting myelopoietic cytokines. In addition, IL-3 supports terminal maturation, prolongs survival and enhances the functional properties of myeloid cells through high-affinity binding sites. In vivo administration of IL-3 is followed by an increase in peripheral white blood cell counts as well as by an increase in the number of circulating progenitor cells giving rise to mature hemopoietic cells in response to more lineage-restricted growth factors. IL-3 also regulates growth of leukemic cells and primes them to become more sensitive to cell cycle specific cytotoxic drugs. IL-3 apparently represents a novel and unique hemopoietic growth factor. Its clinical use should offer new strategies in the treatment of cytopenia, leukemic disease and in stem cell transplantation.     

Why should we be concerned about loss of biodiversity

Over the past century, documented extinctions in well-studied taxonomic groups have been at rates one hundred-fold to one thousand-fold above the average extinction rates seen over the half billion year sweep of the fossil record. But for most groups, particularly invertebrates, we are very uncertain how many species there are on Earth today, much less rates of extinction.     

Why dispersal should be maximized at intermediate scales of heterogeneity

Dispersal is a fundamental biological process that results in the redistribution of organisms due to the interplay between the mode of dispersal, the range of scales over which movement occurs, and the scale of spatial heterogeneity, in which patchiness may occur across a broad range of scales. Despite the diversity of dispersal mechanisms and dispersal length scales in nature, we posit that a fundamental scaling relationship should exist between dispersal and spatial heterogeneity. We present both a conceptual model and mathematical formalization of this expected relationship between the scale of dispersal and the scale of patchiness, which predicts that the magnitude of dispersal (number of individuals) among patches should be maximized when the scale of spatial heterogeneity (defined in terms of patch size and isolation) is neither too fine nor too coarse relative to the gap-crossing abilities of a species. We call this the “dispersal scaling hypothesis” (DSH). We demonstrate congruence in the functional form of this relationship under fundamentally different dispersal assumptions, using well-documented isotropic dispersal kernels and empirically derived dispersal parameters from diverse species, in order to explore the generality of this finding. The DSH generates testable hypotheses as to when and under what landscape scenarios dispersal is most likely to be successful. This provides insights into what management scenarios might be necessary to either restore landscape connectivity, as in certain conservation applications, or disrupt connectivity, as when attempting to manage landscapes to impede the spread of an invasive species, pest, or pathogen.     

Why cytoplasmic signalling proteins should be recruited to cell membranes

It has been suggested that localization of signal-transduction proteins close to the cell membrane causes an increase in their rate of encounter after activation. We maintain that such an increase in the first-encounter rate is too small to be responsible for truly enhanced signal transduction. Instead, the function of membrane localization is to increase the number (or average lifetime) of complexes between cognate signal transduction proteins and hence increase the extent of activation of downstream processes. This is achieved by concentrating the proteins in the small volume of the area just below the plasma membrane. The signal-transduction chain is viewed simply as operating at low default intensity because one of its components is present at a low concentration. The steady signalling level of the chain is enhanced 1000-fold by increasing the concentration of that component. This occurs upon 'piggyback' binding to a membrane protein, such as the activated receptor, initiating the signal-transduction chain. For the effect to occur, the protein translocated to the membrane cannot be free but has to remain organized by being piggyback bound to a receptor, membrane lipid(s) or scaffold. We discuss an important structural constraint imposed by this mechanism on signal transduction proteins that might also account for the presence of adaptor proteins.     

Why sources of heterogeneity in meta-analysis should be investigated.

Although meta-analysis is now well established as a method of reviewing evidence, an uncritical use of the technique can be very misleading. One common problem is the failure to investigate appropriately the sources of heterogeneity, in particular the clinical differences between the studies included. This paper distinguishes between the concepts of clinical and statistical heterogeneity and exemplifies the importance of investigating heterogeneity by using published meta-analyses of epidemiological studies of serum cholesterol concentration and clinical trials of its reduction. Although not without some dangers of speculative conclusions, prompted by overzealous inspection of the data to hand, a sensible investigation of sources of heterogeneity should increase both the scientific and the clinical relevance of the results of meta-analyses.     

Why blind diabetics with renal failure should be offered treatment.

During May 1978 to April 1983 this renal dialysis unit treated 65 patients by continuous ambulatory peritoneal dialysis. Of these, 24 had type I (insulin dependent) diabetes, of whom 20 were blind; eight had type II (non-insulin dependent) diabetes, of whom three were blind; and 33 did not have diabetes and were not blind. The cumulative actuarial survival rates of these patients at five years were 60% for blind diabetics, 40% for sighted diabetics, and 46% for the non-diabetics. Of the 23 blind patients, 22 successfully achieved self care, including the self administration of insulin into the peritoneal dialysis solution. Blind patients had the least peritonitis and fewest complications of continuous ambulatory peritoneal dialysis, and none objected to the treatment or requested to be taken off it or be allowed to die. It was concluded that blind diabetic patients with renal failure showed both the will and the ability to stay alive and that their treatment was worth while.     

Why neural networks should not be used for HIV-1 protease cleavage site prediction

SUMMARY: Several papers have been published where nonlinear machine learning algorithms, e.g. artificial neural networks, support vector machines and decision trees, have been used to model the specificity of the HIV-1 protease and extract specificity rules. We show that the dataset used in these studies is linearly separable and that it is a misuse of nonlinear classifiers to apply them to this problem. The best solution on this dataset is achieved using a linear classifier like the simple perceptron or the linear support vector machine, and it is straightforward to extract rules from these linear models. We identify key residues in peptides that are efficiently cleaved by the HIV-1 protease and list the most prominent rules, relating them to experimental results for the HIV-1 protease. MOTIVATION: Understanding HIV-1 protease specificity is important when designing HIV inhibitors and several different machine learning algorithms have been applied to the problem. However, little progress has been made in understanding the specificity because nonlinear and overly complex models have been used. RESULTS: We show that the problem is much easier than what has previously been reported and that linear classifiers like the simple perceptron or linear support vector machines are at least as good predictors as nonlinear algorithms. We also show how sets of specificity rules can be generated from the resulting linear classifiers. AVAILABILITY: The datasets used are available at http://www.hh.se/staff/bioinf/