Predictive models of hepatotoxicity using gene expression data from primary rat hepatocytes.

With the aim of evaluating the usefulness of an in vitro system for assessing the potential hepatotoxicity of compounds, the paper describes several methods of obtaining mathematical models for the prediction of compound-induced toxicity in vivo. These models are based on data derived from treating rat primary hepatocytes with various compounds, and thereafter using microarrays to obtain gene expression 'profiles' for each compound. Predictive models were constructed so as to reduce the number of 'probesets' (genes) required, and subjected to rigorous cross-validation. Since there are a number of possible approaches to derive predictive models, several distinct modelling strategies were applied to the same data set, and the outcomes were compared and contrasted. While all the strategies tested showed significant predictive capability, it was interesting to note that the different approaches generated models based on widely disparate probesets. This implies that while these models may be useful in ascribing relative potential toxicity to compounds, they are unlikely to provide significant information on underlying toxicity mechanisms. Improved predictivity will be obtained through the generation of more comprehensive gene expression databases, covering more 'toxicity space', and by the development of models that maximize the observation, and combination, of individual differences between compounds.

B-lineage regulated polyadenylation occurs on weak poly(A) sites regardless of sequence composition at the cleavage and downstream regions.

Early/memory and plasma B-cell lines and fibroblasts were analyzed for their ability to use a 5' proximal (variant) versus a 3' distal (constant) poly(A) site, in the absence of a competing splice, from a set of related constructs. The proximal:distal poly(A) site use (P:D ratio) of the resulting cytoplasmic poly(A)+ mRNA is a measure of poly(A) site strength. In this context the immunoglobulin gamma2b secretory-specific poly(A) site showed a P:D ratio of 1:1 in plasma cells, 0.43:1 in early/memory B-cells and an intermediate value in fibroblasts. Meanwhile, a construct with a proximal SV40 early-like poly(A) site produced mRNA with a P:D ratio of >>50:1 in all cell types. Alterations in the region downstream of the proximal poly(A) addition site and at the site itself resulted in changes in the P:D ratio. However, these poly(A) sites, all with a P:D ratio of < or = 5:1, were used most efficiently in plasma cells. Constructs totally devoid of immunoglobulin sequences, but containing heterologous poly(A) sites producing mRNA with P:D ratios of < or = 5:1, were also used more efficiently in plasma cells. We therefore conclude that weak poly(A) sites, regardless of sequence composition, are used more efficiently in plasma cells than in the other cell types.

Natural killer cells suppress human erythroid stem cell proliferation in vitro.

To determine the role of natural killer (NK) cells in the regulation of human erythropoiesis, we studied the effects of NK-enriched cell populations on the in vitro proliferation of erythroid stem cells at three different levels of maturation (day 14 blood BFU-E, day 5-6 marrow CFU-E, and day 10-12 marrow BFU-E). NK cells were enriched from blood by Percoll density gradient centrifugation and by fluorescence-activated cell sorting (FACS), using the human natural killer cell monoclonal antibody, HNK-1. The isolated enriched fractions were cocultured with autologous nonadherent marrow cells or blood null cells and erythropoietin in a methylcellulose erythroid culture system. Cells from low-density Percoll fractions (NK-enriched cells) were predominantly large granular lymphocytes with cytotoxic activity against K562 targets 6-10-fold greater than cells obtained from high-density Percoll fractions (NK-depleted cells). In coculture with marrow nonadherent cells (NA) at NK:NA ratios of 2:1, NK-enriched cells suppressed day 5-6 CFU-E to 62% (p less than 0.025) of controls, whereas NK-depleted cells slightly augmented CFU-E to 130% of controls (p greater than 0.05). In contrast, no suppression of day 10-12 marrow BFU-E was observed employing NK-enriched cells. The NK CFU-E suppressor effects were abolished by complement-mediated lysis of NK-enriched cells with the natural killer cell antibody, HNK-1. Highly purified HNK-1+ cells separated by FACS suppressed marrow CFU-E to 34% (p less than 0.025) and marrow BFU-E to 41% (p less than 0.025) of controls. HNK- cells had no significant effect on either BFU-E or CFU-E growth. NK-enriched cells were poor stimulators of day 14 blood BFU-E in comparison to equal numbers of NK-depleted cells or T cells isolated by E-rosetting (p less than 0.01). Interferon boosting of NK-enriched cells abolished their suboptimal burst-promoting effects and augmented their CFU-E suppressor effects. These studies provide evidence for a potential regulatory role of NK cells in erythropoiesis. The NK suppressor effect is maximal at the level of the mature erythroid stem cell CFU-E. These findings may explain some hypoproliferative anemias that develop in certain NK cell-activated states.