Credibility, Replicability, and Reproducibility in Simulation for Biomedicine and Clinical Applications in Neuroscience.

Modeling and simulation in computational neuroscience is currently a research enterprise to better understand neural systems. It is not yet directly applicable to the problems of patients with brain disease. To be used for clinical applications, there must not only be considerable progress in the field but also a concerted effort to use best practices in order to demonstrate model credibility to regulatory bodies, to clinics and hospitals, to doctors, and to patients. In doing this for neuroscience, we can learn lessons from long-standing practices in other areas of simulation (aircraft, computer chips), from software engineering, and from other biomedical disciplines. In this manuscript, we introduce some basic concepts that will be important in the development of credible clinical neuroscience models: reproducibility and replicability; verification and validation; model configuration; and procedures and processes for credible mechanistic multiscale modeling. We also discuss how garnering strong community involvement can promote model credibility. Finally, in addition to direct usage with patients, we note the potential for simulation usage in the area of Simulation-Based Medical Education, an area which to date has been primarily reliant on physical models (mannequins) and scenario-based simulations rather than on numerical simulations.

Simulating in vitro epithelial morphogenesis in multiple environments.

In vitro studies of epithelial cell morphogenesis have demonstrated the influence of environment composition and orientation in the development of multicellular epithelial structures such as tubules and cysts. We have constructed a low resolution, discrete event simulation model and report on its use to explore how experimentally observed morphogenetic phenomena under four growth conditions might be generated and controlled. We identified simulation attributes that may have in vitro counterparts. We studied how changes in the logic governing simulated epithelial cell behavior might cause abnormal growth. Simulation results support the importance of a polarized response to the environment to the generation of a normal epithelial phenotype and show how disruptions of tight mechanistic control lead to aberrant growth characteristics.

Simulating leukocyte-venule interactions--a novel agent-oriented approach.

Leukocyte recruitment into sites of inflammation involves a complex cascade of molecular interactions between the leukocyte and the endothelial cells of the inflamed venule. This report proposes a novel agent-oriented approach for simulating leukocyte-venule interactions during inflammation. We focus on modeling and simulating the initial steps of rolling, activation, and firm adhesion of neutrophils on TNF-alpha-treated mouse cremaster muscle venules.

Modeling transport kinetics with StarLogo.

StarLogo, an agent-based modeling and simulation platform, was used to simulate adsorption-mediated transcytosis of a molecule from the lumen side of a cell membrane to the abluminal extra-cellular fluid (ECF). The model contains small nondiffusible substrate molecules, transporters, and substrate-transporter agents. The "reaction" is a transporter combining with the substrate which then crosses the cell cytoplasm. The substrate that is deposited on the ECF side becomes the "product". Results showed characteristics consistent with Michaelis-Menten enzyme kinetics. The model can serve as an example of agent-based modeling and simulation.

Vascular injuries of the upper extremity.

Vascular injuries of the upper extremity represent approximately 30% to 50% of all peripheral vascular injuries. The majority of injuries are to the brachial artery, and 90% of injuries are due to penetrating trauma. Return of function is often related to concomitant injury to peripheral nerves. However, timely restoration of blood flow is essential to optimize outcome. The diagnosis is made by physical examination and limited Doppler ultrasonography. Arteriography may be helpful if there are multiple sites of injury. Anticoagulation with heparin should be given if not otherwise contraindicated. Revascularization should be completed within the critical ischemic time: 4 hours for proximate injuries and 12 hours for distal injuries. Revascularization methods include resection and primary repair or resection with an interposition graft. The sequence of repair of multiple injuries to the extremity begins with arterial revascularization followed by skeletal stabilization and nerve and tendon repair.

A forecasting approach to accelerate drug development.

The clinical phase of drug development should be concluded sooner and at a lower cost if primarily only the pivotal and supportive studies were to be conducted. Such improved efficiency requires development of a decision support system that delivers five new capabilities: (i) it enables one to predict a result of a clinical study and to identify those studies that are expected to have an acceptable probability of success; (ii) it will allow one to optimally utilize available pharmacokinetic and pharmacodynamic (PK/PD) data and improve its predictive capability as more data become available; (iii) it will enable one to project useful population results, not just mean results; (iv) predictions will be accompanied by a measure of reliability; and (v) expected initial clinical results will be predictable from animal and related drug class data. With such a tool population targets could be specified very early in the drug development programme, challenged, and then rationally revised at each step during the development process. This report describes progress in developing and testing a clinical trials Forecaster, a prototype for such a system. The Forecaster generates estimates of the joint density for a population of combined PK/PD parameters. That population then serves as a surrogate for the population of individuals. When the resulting joint density is sampled, the obtained sets of parameters may be used to generate data that is statistically indistinguishable from the original experimental data. Such simulated data can be used to validate assumptions, and make inferences on specified population targets that are accompanied by a measure of prediction reliability. We demonstrate use of the forecaster by employing N = 22 PK/PD parameter sets for an orally administered analgesic.

Bootstrapping for pharmacokinetic models: visualization of predictive and parameter uncertainty.

PURPOSE: We explore use of "bootstrapping" methods to obtain a measure of reliability of predictions made in part from fits of individual drug level data with a pharmacokinetic (PK) model, and to help clarify parameter identifiability for such models. METHODS: Simulation studies use four sets (A-D) of drug concentration data obtained following a single oral dose. Each set is fit with a two compartment PK model, and the "bootstrap" is employed to examine the potential predictive variation in estimates of parameter sets. This yields an empirical distribution of plausible steady state (SS) drug concentration predictions that can be used to form a confidence interval for a prediction. RESULTS: A distinct, narrow confidence region in parameter space is identified for subjects A and B. The bootstrapped sets have a relatively large coefficient of variation (CV) (35-90% for A), yet the corresponding SS drug levels are tightly clustered (CVs only 2-9%). The results for C and D are dramatically different. The CVs for both the parameters and predicted drug levels are larger by a factor of 5 and more. The results reveal that the original data for C and D, but not A and B, can be represented by at least two different PK model manifestations, yet only one provides reliable predictions. CONCLUSIONS: The insights gained can facilitate making decisions about parameter identifiability. In particular, the results for C and D have important implications for the degree of implicit overparameterization that may exist in the PK model. In cases where the data support only a single model manifestation, the "bootstrap" method provides information needed to form a confidence interval for a prediction.

Validation of a decision support system for use in drug development: pharmacokinetic data.

PURPOSE: Single dose pharmacokinetic data from several individuals can be used to predict the fraction of the population that is expected to be within a therapeutic range. Without having some measure of its reliability, however, that prediction is only likely to marginally influence critical drug development decision making. The system (Forecaster) described generates an approximate prediction interval that contains the original prediction and where, for example, the probability is approximately 85% that a similar prediction from a new set of data will also be within the range. The goal is to validate that the system functions as designed. METHODS: The strategy requires having a Surrogate Population (SP), which is a large number (> or = 1500) of hypothetical individuals each represented by set of model parameter values having unique attributes. The SP is generated so that a sample taken from it will give data that is statistically indistinguishable from the available experimental data. The automated method for building the SP is described. RESULTS: Validation studies using 300 independent samples document that for this example the SP can be used to make useful predictions, and that the approximate prediction interval functions as designed. CONCLUSIONS: For the boundary conditions and assumptions specified, the Forecaster can make valid predictions of pharmacokinetic-based population targets that without a SP would not be possible. Finally, the approximate prediction interval does provide a useful measure of prediction reliability.

Non-peptide glycoprotein IIb/IIIa inhibitors. 17. Design and synthesis of orally active, long-acting non-peptide fibrinogen receptor antagonists.

The synthesis and pharmacological evaluation of 5 (L-738, 167), a potent, selective non-peptide fibrinogen receptor antagonist is reported. Compound 5 inhibited the aggregation of human gel-filtered platelets with an IC50 value of 8 nM and was found to be > 33000-fold less effective at inhibiting the attachment of human endothelial cells to fibrinogen, fibronectin, and vitronectin than it was at inhibiting platelet aggregation. Ex vivo platelet aggregation was inhibited by > 85% 24 h after the oral administration of 5 to dogs at 100 micrograms/kg. The extended pharmacodynamic profile exhibited by 5 appears to be a consequence of its high-affinity binding to GPIIb/IIIa on circulating platelets and suggests that 5 is suitable for once-a-day dosing.

Sequence-independent inhibition of RNA transcription by DNA dumbbells and other decoys.

DNA dumbbells are stable, short segments of double-stranded DNA with closed nucleotide loops on each end, conferring resistance to exonucleases. Dumbbells may be designed to interact with transcription factors in a sequence-specific manner. The internal based paired sequence of DNA dumbbells in this study contains the X-box, a positive regulatory motif found in all MHC class II DRA promoters. In electrophoretic mobility shift assays (EMSAs), dumbbells and other oligonucleotides ('decoys') with the core X-box sequence were found to compete with the native strand for binding to X-box binding proteins (including RFX1). However, only the X-box dumbbell was capable of forming detectable complexes with such proteins using EMSA. In a model cell system, dumbbells were tested for their ability to block RFX1VP16 activation of a plasmid containing multiple repeats of the X-box linked to the CAT gene. While it appeared that dumbbells could block this activation, the effect was non-specific. This and further evidence suggests an inhibition of transcription, most likely via an interaction with the general transcriptional machinery.

Measures of uncertainty of pharmacokinetic and pharmacodynamic parameter estimates: a new computerized algorithm.

Numerous algorithms exist to fit data to nonlinear models of the type used in chemistry, pharmacology, physiology, etc. Most include modules that provide some measure of the reliability of the estimated model parameters. The variance-covariance matrix (VCM) is the common tabulation of information that is used to quantify the parameter uncertainty as well as correlations between parameters. The VCM has its mathematical foundation in the linear regression world, where the dependent variable is a linear function of the parameters. However, when the model is not linear in its parameters, then the VCM is no longer an absolute quantitative measure of reliability of the parameter estimates and should be interpreted with caution. If the goal is to obtain a realistic and quantitative rather than a qualitative measurement of the parameter reliability, then it is necessary to have an alternative approach to describe the parameter likelihood region. We present a computerized algorithm that fills that need, and we compare its performance with the traditional VCM approach for different data sets. We also discuss criteria that may be used to determine when the VCM approach should and should not be used.

An oligonucleotide blocks interferon-gamma signal transduction.

Interferon (IFN)-gamma is an important mediator of transplant graft rejection. It induces endothelial cell expression of HLA-DR and intercellular adhesion molecule-1, which render transplant grafts more susceptible to rejection by the host. Oligonucleotide 5'-GGG GTT GGT TGT GTT GGG TGT TGT GT-RNH2 (oligo I) blocks multiple IFN-gamma effects in human K562 cell cultures. A systematic approach revealed that oligo I has a novel, and potentially important, mode of action--it blocks the binding of IFN-gamma to its receptor, thus preventing activation of the IFN-gamma signal transduction pathway. The results are consistent with an aptamer mechanism of action, because oligo I exerts its inhibitory effects by interacting with protein, not intracellular nucleic acid targets, such as mRNA or genomic DNA.

PSD-95 is associated with the postsynaptic density and not with the presynaptic membrane at forebrain synapses.

PSD-95, a prominent protein component of the postsynaptic density (PSD) fraction from rat forebrain, has been localized by light microscopy to dendrites of hippocampal neurons (Cho et al., 1992) and to the presynaptic plexus of cerebellar basket cells (Kistner et al., 1993). Here we extend these studies to show that an affinity-purified antibody to PSD-95 labels the dendrites of most neurons in the forebrain and of a subset of neurons in the cerebellum. To confirm that PSD-95 is associated with the PSD at forebrain synapses and to clarify whether it is also associated with the presynaptic membrane, we employed immunogold electron microscopy of forebrain synaptosomes. Gold-labeled antibodies to PSD-95 labeled postsynaptic densities in both intact and lysed forebrain synaptosomes but did not label presynaptic terminals or the presynaptic membrane. The asymmetric distribution of PSD-95 at synapses contrasts with that of its homologs, disks-large and ZO-1, which are arranged symmetrically at septate and tight junctions, respectively.

Screening and characterization of estrogenic activity from a hydroxystilbene library.

BACKGROUND: Compounds that either inhibit or induce an estrogen response in vivo are important as potential drugs and biochemical tools. Non-steroidal stilbene analogs such as tamoxifen are known to function as both estrogen agonists and antagonists depending upon the analog structure. This family of compounds is amenable to parallel-manifold synthesis because stilbene analogs are easily synthesized using a single-step olefination reaction. RESULTS: We have prepared a small 23-component hydroxystilbene library using a solid phase synthesis approach. The library was screened for estrogenic and antiestrogenic activity using a cell-based bioassay that measures estrogen receptor-mediated transcription of a reporter gene. Three of the analogs proved to have dose-dependent estrogenic activity with EC50 values between 5 microM and 15 microM. Further characterization of the hydroxystilbene-mediated estrogenic activity suggests that the agonist activity results from direct binding to the steroid site on the estrogen receptor with IC50 values of 1-10 microM. CONCLUSIONS: The results of this study show that classic olefination chemistry can be adapted to a solid-phase format for parallel synthesis of analog libraries. Although yields varied for the individual analogs, sufficient quantity of pure material was obtained directly from the resin for structural characterization and biological evaluation. This study further validates solid-phase organic synthesis as a useful approach for rapid parallel-manifold library synthesis to augment both lead compound discovery and optimization.

Suppression of interferon-gamma induction of MHC class II and ICAM-1 by a 26-base oligonucleotide composed of deoxyguanosine and deoxythymidine.

Interferon-gamma (IFN-gamma) is an important cytokine released by T lymphocytes and natural killer cells which is able to induce expression of class II MHC and ICAM-1, crucial factors in cellular immune response. HeLa S3, HS 27, and NF-71-1 are cell lines which can be induced to express HLA-DR and HLA-DP by exposure to IFN-gamma. When T2 (5'GGGGTTGGTTGTGTTGGGTGTTGTGTRNH(2)3') oligonucleotide was added at 5-20 microM every other day, cell surface induction of HLA-DR and HLA-DP by IFN-gamma was suppressed in a dose-dependent manner in HeLa S3. T2 suppressive effect on HLA class II was also observed in four different nontransformed human cell lines, HS 27 at passage 18, NF-71-1 at passage 5, human corneal endothelial cell at passage 5, and human retinal pigmented epithelial cell at passage 3. Control oligonucleotides had no suppressive effect. Northern hybridization showed that HLA-DR A mRNA induction by IFN-gamma was blocked by T2 in HeLa S3 and fibroblast 143B. The suppressive effect of T2 was also reversible as continued culture of the treated cells without further addition of the oligonucleotide allowed full re-expression of HLA-DR. Further experiments showed that T2 oligonucleotide was also able to inhibit IFN-gamma enhancement of ICAM-1 (CD54) on human corneal endothelial cell and human retinal pigmented epithelial cell. We conclude that T2 oligonucleotide is effective at suppressing HLA-DR, HLA-DP and ICAM-1 induction by IFN-gamma in transformed and nontransformed cells in vitro.

Inhibition of transcription factor binding to the HER2 promoter by triplex-forming oligodeoxyribonucleotides.

We have identified a 28-bp homopurine/homopyrimidine sequence capable of triple helix (triplex) formation with G+T-rich oligodeoxyribonucleotides (oligos) within the critical proximal promoter of the HER2/neu/c-erbB2 (HER2) proto-oncogene. To investigate the possible therapeutic potential of triplex-forming oligos in HER2 overexpressing breast cancers, we have studied the ability of triplex formation to compete with and to inhibit the binding of a transcription factor to its consensus sequence at an adjacent site. Competition binding assays demonstrate that a triplex-forming oligo can inhibit transcription factor binding in a sequence-specific manner. Moreover, we find that the addition of both nucleotide and non-nucleotide 'tails' to triplex-forming oligos do not confer any enhancement of binding affinity, but provide additional inhibition of transcription factor binding, potentially by steric hindrance.

Paradoxical production of target protein using antisense RNA expression vectors.

We used antisense RNA in a protocol designed to reduce estrogen receptor (ER) content in human breast cancer cells and observed paradoxical increases in ER levels. ER protein activity was measured using a highly sensitive reporter gene assay that relies on the ability of functional ER to bind a consensus estrogen response element (ERE) and drive the production of chloramphenicol acetyl-transferase (CAT). Upon transient transfection of ER-positive cell lines with three different vectors containing the full-length ER cDNA cloned in an antisense orientation, we observed unexpected increases in ER-driven CAT activity. To further investigate this phenomenon, expression from the antisense ER vectors was studied in an ER-negative breast tumor cell line, MDA-MB-453. ER activity was observed in these ER-negative cells upon transient transfection with each of three antisense ER vectors, but not from control vectors. Expression of ER from antisense constructs was 30-100-times less efficient than ER expression from isogenic sense constructs. The paradoxical ER activity was consistent with expected ER behavior in that it exhibited characteristic binding to the natural ligand, 17 beta-estradiol (E2), and it was inhibited by the antiestrogens, 4-hydroxy-tamoxifen (OHT) and ICI 164384 (ICI). Control vectors containing a truncated antisense ER cDNA produced no ER activity. Although the mechanism for this ER expression has not been determined, it appears likely that it is due to transcription off the opposite strand of the antisense construct.(ABSTRACT TRUNCATED AT 250 WORDS)

Characterization of apocytochrome C binding to human erythrocytes.

The binding of 125I-labeled apocytochrome c to human erythrocytes was determined for free apocytochrome c concentrations at 10(-10)-10(-6) M. At about 2 x 10(-9) M, maximum cell association of apocytochrome c occurs at 50 mM NaCl and at 22 degrees C. Intact erythrocytes at 22 degrees C have three classes of apocytochrome c binding sites: one high-affinity noncooperative site (n1 = 728 per cell, Kd1 = 1.5 x 10(-9) M) and two positively cooperative sites (n2 = 3.7 x 10(4) per cell, Kd2 = 1.2 x 10(-7) M, alpha 2 = 2.0, and n3 = 2.5 x 10(5) per cell, Kd3 = 7.1 x 10(-7) M, alpha 3 = 12). Erythrocytes at 37 degrees C, and erythrocyte ghosts at 22 degrees C, also have three classes of apocytochrome c binding sites, and most sites are positively cooperative.

Characterization of the oligodeoxynucleotide-mediated inhibition of interferon-gamma-induced major histocompatibility complex class I and intercellular adhesion molecule-1.

The major histocompatibility complex (MHC) Class I and II genes and intercellular adhesion molecule-1 (ICAM-1) are regulated by interferon-gamma in a variety of cell types. We have previously shown that the oligodeoxynucleotide 5'-GGG GTT GGT TGT GTT GGG TGT TGT GT-RNH2 (oligo I) inhibits the interferon-gamma-mediated enhancement of MHC Class I and ICAM-1 proteins in the K562 cell line. We have now investigated the mechanism of action of oligo I and report that it acts by inhibiting the binding of interferon-gamma to cells. We also show that the dose-response curves, the selectivity profile, and the kinetics of oligo I are consistent with this novel mechanism of action. The dose-response curves for oligo I, obtained using antibodies against the MHC Class I heavy chain, beta 2-microglobulin, or ICAM-1, are almost superimposable at each observation time. MHC Class I induction by 6400 units/ml interferon-alpha or interferon-beta or ICAM-1 enhancement by 800 units/ml tumor necrosis factor-alpha is not inhibited by oligo I. However, the synergistic induction of MHC Class I by mixtures of tumor necrosis factor-alpha and interferon-gamma is inhibited. Oligo I belongs to a class of active oligodeoxynucleotides that inhibits interferon-gamma-induced MHC Class I and ICAM-1 in K562 cells. The activity and potency is sequence-dependent, but remarkably different sequences can have comparable effects. The activity of oligo I in the HeLa S3 cell line inhibits the interferon-gamma-mediated enhancement of both ICAM-1 and MHC Class II DR and the interferon-gamma-mediated reduction in transferrin receptor expression. Thus, oligo I appears to specifically inhibit interferon-gamma-induced changes in protein expression, which is consistent with oligo I acting at an early step(s) in the induction process. Taken together, our results show that oligo I exerts its effects by inhibiting the association of interferon-gamma with the cell surface, which is a novel mechanism of action for oligodeoxynucleotides.