Design considerations for two-stage enrichment clinical trials

When there is a predictive biomarker, enrichment can focus the clinical trial on a benefiting subpopulation. We describe a two-stage enrichment design, in which the first stage is designed to efficiently estimate a threshold and the second stage is a “phase III-like” trial on the enriched population. The goal of this paper is to explore design issues: sample size in Stages 1 and 2, and re-estimation of the Stage 2 sample size following Stage 1. By treating these as separate trials, we can gain insight into how the predictive nature of the biomarker specifically impacts the sample size. We also show that failure to adequately estimate the threshold can have disastrous consequences in the second stage. While any bivariate model could be used, we assume a continuous outcome and continuous biomarker, described by a bivariate normal model. The correlation coefficient between the outcome and biomarker is the key to understanding the behavior of the design, both for predictive and prognostic biomarkers. Through a series of simulations we illustrate the impact of model misspecification, consequences of poor threshold estimation, and requisite sample sizes that depend on the predictive nature of the biomarker. Such insight should be helpful in understanding and designing enrichment trials.     

Evidence for Nerve Growth Factor-Potentiating Activities of the Nonpeptidic Compound SR 57746A in PC12 Cells

Abstract: SR 57746A {1-[2-(naphth-2-yl)ethyl]-4-(3-trifluoromethylphenyl)-1,2,5,6-tetrahydropyridine hydrochloride} exhibits neurotrophic activities in vivo and in vitro. We used the rat pheochromocytoma PC12 cell line to investigate in vitro cellular changes induced by SR 57746A. A significant increase in the percentage of cells bearing neurite-like processes was obtained in cells treated by SR 57746A and nerve growth factor (NGF) compared with NGF treatment alone. SR 57746A added alone, however, had no effect on morphogenesis or on survival of cells in serum-free medium. In contrast, SR 57746A induced a "priming" effect on PC12 cells for neurite outgrowth within 6 h of addition of the protein tyrosine kinase inhibitor genistein. An increase in α-actinin content resulted from treatment with SR 57746A. Expression of NGF-mediated acetylcholinesterase and choline acetyltransferase was enhanced within 5 days by SR 57746A. The molecule also induced rapid F-actin redistribution. Within 2 min of incubation, outgrowth of F-actin-containing filopodia was clearly visible at the cell periphery, as previously shown with NGF. It is interesting that this effect of SR 57746A could be mimicked by protein tyrosine kinase inhibitors and abolished by preincubation with sodium orthovanadate, a protein tyrosine phosphatase inhibitor.     

The structure and function of p55PIK reveal a new regulatory subunit for phosphatidylinositol 3-kinase.

Phosphatidylinositol 3-kinase (PI-3 kinase) is implicated in the regulation of diverse cellular processes, including insulin-stimulated glucose transport. PI-3 kinase is composed of a 110-kDa catalytic subunit and an 85-kDa regulatory subunit. Here, we describe p55PIK, a new regulatory subunit that was isolated by screening expression libraries with tyrosine-phosphorylated insulin receptor substrate 1 (IRS-1). p55PIK is composed of a unique 30-residue NH2 terminus followed by a proline-rich motif and two Src homology 2 (SH2) domains with significant sequence identify to those in p85. p55PIK mRNA is expressed early during development, remains abundant in adult mouse brain and testis tissue, and is detectable in adult adipocytes and heart and kidney tissues. p55PIK forms a stable complex with p110, and it associates with IRS-1 during insulin stimulation. Moreover, the activated insulin receptor phosphorylates p55PIK in Sf9 cells, and insulin stimulates p55PIK phosphorylation in CHOIR/p55PIK cells. The unique features of p55PIK suggest that it is important in receptor signaling.     

The Bacillus subtilis ureABC operon.

The Bacillus subtilis ureABC operon encodes homologs of the three subunits of urease enzymes of the family Enterobacteriaceae. Disruption of ureC prevented utilization of urea as a nitrogen source and resulted in a partial growth defect in minimal medium containing limiting amounts of arginine or allantoin as the sole nitrogen source.     

Rapid Plasma Membrane Anchoring of Newly Synthesized p59 fyn: Selective Requirement for NH2-Terminal Myristoylation and Palmitoylation at Cysteine-3

The trafficking of Src family proteins after biosynthesis is poorly defined. Here we studied the role of dual fatty acylation with myristate and palmitate in biosynthetic transport of p59^fyn. Metabolic labeling of transfected COS or NIH 3T3 cells with [³⁵S]methionine followed by analysis of cytosolic and total membrane fractions showed that Fyn became membrane bound within 5 min after biosynthesis. Newly synthesized Src, however, accumulated in the membranes between 20– 60 min. Northern blotting detected Fyn mRNA specifically in soluble polyribosomes and soluble Fyn protein was only detected shortly (1–2 min) after radiolabeling. Use of chimeric Fyn and Src constructs showed that rapid membrane targeting was mediated by the myristoylated NH₂-terminal sequence of Fyn and that a cysteine at position 3, but not 6, was essential. Examination of Gα_o-, Gα_s-, or GAP43-Fyn fusion constructs indicated that rapid membrane anchoring is exclusively conferred by the combination of N-myristoylation plus palmitoylation of cysteine-3. Density gradient analysis colocalized newly synthesized Fyn with plasma membranes. Interestingly, a 10–20-min lag phase was observed between plasma membrane binding and the acquisition of non-ionic detergent insolubility. We propose a model in which synthesis and myristoylation of Fyn occurs on soluble ribosomes, followed by rapid palmitoylation and plasma membrane anchoring, and a slower partitioning into detergent-insoluble membrane subdomains. These results serve to define a novel trafficking pathway for Src family proteins that are regulated by dual fatty acylation.     

Oxidant Injury in PC12 Cells—A Possible Model of Calcium "Dysregulation" in Aging: I. Selectivity of Protection Against Oxidative Stress

Abstract: Previous research has suggested that the initial effects of cellular free radical neurotoxic insult involve large increases in intracellular Ca²⁺. However, the exact role of oxidative stress on the various parameters involved in these increases has not been specified. The present experiments were performed to examine these parameters in PC12 cells exposed to 5, 25, or 300 µM H₂O₂ for 30 min in growth medium alone or containing either nifedipine (L-type Ca²⁺ antagonist), conotoxin (N-type antagonist), Trolox (vitamin E analogue), or α-phenyl-n-tert-butylnitrone (nitrone trapping agent; PBN). The concentrations of H₂O₂ were chosen by examining the degree of cell killing induced by exposure to graded concentrations of H₂O₂. The 5 and 25 µM concentrations of H₂O₂ produced no significant cell killing at either 30 min or 24 h after treatment, whereas the 300 µM concentration produced a moderate degree of cell killing that did not increase between the two times. Fluorescent imaging was used to visualize intracellular Ca²⁺ changes in fura-2-loaded cells. Baseline (pre-30 mM KCI) Ca²⁺ levels were increased significantly by H₂O₂ treatment (e.g., 300 µM, 200%), but the rise in the level of free intracellular Ca²⁺ after KCI stimulation (i.e., peak) was decreased (e.g., 300 µM, 50%) and the cell's ability to sequester or extrude the excess Ca²⁺ (i.e., Ca²⁺ recovery time) after depolarization was decreased significantly. All compounds prevented baseline Ca²⁺ increases and, with the exception of conotoxin, antagonized the peak decreases in Ca²⁺. It is interesting that after 300 µM H₂O₂ exposure, only Trolox was partially effective in preventing these deficits in recovery. Conotoxin increased the decrement recovery in the absence of H₂O₂. However, in cells exposed to 5 or 25 µM H₂O₂, conotoxin as well as the other agents were effective in preventing the deficits in recovery.     

Distribution of 3α-hydroxysteroid dehydrogenase in rat brain and molecular cloning of multiple cDNAs encoding structurally related proteins in humans

3α-Hydroxysteroid dehydrogenase in the brain is responsible for production of neuroactive tetrahydrosteroids that interact with the major inhibitory gamma-aminobutyric acid receptor complexes. Distribution of 3α-hydroxysteroid dehydrogenase in different regions of the brain in rats was evaluated by activity assay and by Western immunoblotting using a monoclonal antibody against liver 3α-hydroxysteroid dehydrogenase as the probe. The olfactory bulb was found to contain the highest level of 3α-hydroxysteroid dehydrogenase activity, while moderate levels of the enzyme activity were found in other regions such as cerebellum, cerebral cortex, hypothalamus and pituitary. Some activity was found in the rest of the brain such as amygdala, brain stem, caudate putamen, cingulate cortex, hippocampus, midbrain, and thalamus. The protein levels of 3α-hydroxysteroid dehydrogenase in different regions of the brain as detected by Western immunoblotting are comparable to those of the enzyme activity. We used the rat cDNA as the probe to screen a human liver λ gt11 cDNA library. A total of four different cDNAs were identified and sequenced. One of the cDNAs is identical to that of the human chlordecone reductase cDNA except that our clone contains a much longer 5′-coding sequence than previously reported. The other three cDNAs display high degrees of sequence homology to those of both rat 3α-hydroxysteroid dehydrogenase and human chlordecone reductase. We are currently investigating the functional relationship between the enzymes encoded by these human cDNAs and 3α-hydroxysteroid dehydrogenase.     

The muscle layer of the canine gallbladder and cystic duct

The muscle layer of the canine gallbladder wall and cystic duct was found to be a three-dimensional meshwork of smooth muscle bundles which appear loosely and irregularly arranged on the mucosal aspect and consolidate to form a homogeneous plate-like layer on the serosal aspect. The muscle bundles are tightly woven around interspersed pockets of loose connective tissue in the gallbladder wall and gradually become loosely arranged with more prominent amounts of intervening connective tissue in the cystic duct. The muscle layer is thickest in the gallbladder wall and becomes progressively thinner out into the cystic duct. No anatomic sphincter was observed. Ultrastructural organization revealed individual muscle fibers to be of irregular profile, often branching, widely spaced with intervening collagen fibers, and having few cell-to-cell contacts.