Extended kalman filter for estimation of parameters in nonlinear state-space models of biochemical networks.

It is system dynamics that determines the function of cells, tissues and organisms. To develop mathematical models and estimate their parameters are an essential issue for studying dynamic behaviors of biological systems which include metabolic networks, genetic regulatory networks and signal transduction pathways, under perturbation of external stimuli. In general, biological dynamic systems are partially observed. Therefore, a natural way to model dynamic biological systems is to employ nonlinear state-space equations. Although statistical methods for parameter estimation of linear models in biological dynamic systems have been developed intensively in the recent years, the estimation of both states and parameters of nonlinear dynamic systems remains a challenging task. In this report, we apply extended Kalman Filter (EKF) to the estimation of both states and parameters of nonlinear state-space models. To evaluate the performance of the EKF for parameter estimation, we apply the EKF to a simulation dataset and two real datasets: JAK-STAT signal transduction pathway and Ras/Raf/MEK/ERK signaling transduction pathways datasets. The preliminary results show that EKF can accurately estimate the parameters and predict states in nonlinear state-space equations for modeling dynamic biochemical networks.

Role and Regulation of EphA2 in Pancreatic Cancer

Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cancer cause of death in the US. Gemcitabine is the first-line therapy for this disease, but unfortunately it shows only very modest benefit. The focus of the current study was to investigate the role and regulation of EphA2, a receptor tyrosine kinase expressed in PDAC, to further understand this disease and identify new therapeutic targets. The role of EphA2 was determined in PDAC by siRNA mediated silencing. In combination with gemcitabine, silencing of EphA2 caused a dramatic increase in apoptosis even in highly resistant cells in vitro. Furthermore, EphA2 silencing was found to be useful in 2 orthotopic models in vivo: 1) shRNA-pretreated Miapaca-2 cells, and 2) in vivo delivery of siRNA to established MPanc96 tumors. Silencing of EphA2 alone reduced tumor growth in Miapaca-2 cells. In MPanc96, only the combination treatment of gemcitabine plus siEphA2 significantly reduced tumor growth, as well as the number of lung and liver metastases. Taken together, these observations support EphA2 as a target for combination therapies for PDAC. The regulation of EphA2 was further explored with a focus on the role of Ras. K-Ras activating mutations are the most important initiating event in PDAC. We demonstrated that Ras regulates EphA2 expression through activation of MEK2 and phosphorylation of ERK. Downstream of ERK, silencing of the transcription factor AP-1 subunit c-Jun or inhibition of the ERK effector RSK caused a decrease in EphA2 expression, supporting their roles in this process. Further examination of Ras/MEK/ERK pathway modulators revealed that PEA-15, a protein that sequesters ERK to the cytoplasm, inhibited expression of EphA2. A significant inverse correlation between EphA2 and PEA-15 levels was observed in mouse models of PDAC. In cells where an EGFR inhibitor reduced phospho-Erk, expression of EphA2 was also reduced, indicating that changes in EphA2 levels may allow monitoring the effectiveness of anti-Ras/MEK/ERK therapies. In conclusion, EphA2 levels may be a good prognostic factor for anti-EGFR/anti-Ras therapies, and EphA2 itself is a relevant target for the development of new therapies.

Antibodies targeting hepatoma-derived growth factor as a novel strategy in treating lung cancer.

Hepatoma-derived growth factor (HDGF) is overexpressed in lung cancer and the overexpression correlates with aggressive biological behaviors and poor clinical outcomes. We developed anti-HDGF monoclonal antibodies and tested their antitumor activity in lung cancer xenograft models. We also determined biological effects in tumors treated with the antibody alone or in combination with bevacizumab/avastin (an anti-vascular endothelial growth factor antibody) and/or gemcitabine (a chemotherapeutic agent). We found the anti-HDGF was effective to inhibit tumor growth in non-small cell lung cancer xenograft models. In the A549 model, compared with control IgG, tumor growth was substantially inhibited in animals treated with anti-HDGF antibodies, particularly HDGF-C1 (P = 0.002) and HDGF-H3 (P = 0.005). When HDGF-H3 was combined with either bevacizumab or gemcitabine, we observed enhanced tumor growth inhibition, particularly when the three agents were used together. HDGF-H3-treated tumors exhibited significant reduction of microvessel density with a pattern distinctive from the microvessel reduction pattern observed in bevacizumab-treated tumors. HDGF-H3-treated but not bevacizumab-treated tumors also showed a significant increase of apoptosis. Interestingly, many of the apoptotic cells in HDGF-H3-treated tumors are stroma cells, suggesting that the mechanism of the antitumor activity is, at least in part, through disrupting formation of tumor-stroma structures. Our results show that HDGF is a novel therapeutic target for lung cancer and can be effectively targeted by an antibody-based approach.

Sequential updating via dynamic hierarchical models: A Bayesian approach to longitudinal data analysis

Most statistical analysis, theory and practice, is concerned with static models; models with a proposed set of parameters whose values are fixed across observational units. Static models implicitly assume that the quantified relationships remain the same across the design space of the data. While this is reasonable under many circumstances this can be a dangerous assumption when dealing with sequentially ordered data. The mere passage of time always brings fresh considerations and the interrelationships among parameters, or subsets of parameters, may need to be continually revised. ^ When data are gathered sequentially dynamic interim monitoring may be useful as new subject-specific parameters are introduced with each new observational unit. Sequential imputation via dynamic hierarchical models is an efficient strategy for handling missing data and analyzing longitudinal studies. Dynamic conditional independence models offers a flexible framework that exploits the Bayesian updating scheme for capturing the evolution of both the population and individual effects over time. While static models often describe aggregate information well they often do not reflect conflicts in the information at the individual level. Dynamic models prove advantageous over static models in capturing both individual and aggregate trends. Computations for such models can be carried out via the Gibbs sampler. An application using a small sample repeated measures normally distributed growth curve data is presented. ^

Culturally sensitive cancer education: Effect on knowledge, cancer fatalism, and participation in mobile mammography screening among African-American women

Purpose. To evaluate the effectiveness of a culturally sensitive educational intervention that used an African American lay survivor of breast cancer to increase knowledge of breast cancer, decrease cancer fatalism, and increase participation in mobile mammography screening among African American women. ^ Design. Experimental pretest/posttest design. ^ Setting. Two predominantly African American churches in a large southwestern metropolitan city. ^ Sample. Participants included 93 African American women, 40 years of age and older. Participants were randomly assigned to an intervention group (n = 48) or a control group (n = 45). ^ Methods. Pretest and post-test measures included the Breast Cancer Knowledge Test and the Powe Fatalism Inventory. In addition, demographic and breast screening practices were collected by questionnaire. The intervention group received a breast cancer educational testimonial from an African American lay survivor of breast cancer, who answered questions and addressed concerns, while stressing the importance of taking responsibility for one's own health and spreading disease prevention messages throughout the African American community. The control group viewed the American Cancer Society “Keep In Touch” video prepared specifically for African American women. Participants in both groups were given culturally sensitive educational materials designed to increase knowledge about breast cancer, and were instructed on breast self-examination by an African American registered nurse, using ethnically appropriate breast models. In addition, after the post-test, all eligible participants were given an opportunity to have a free mammogram via a mobile mammography unit parked at the church. ^ Findings. Participants in the intervention group had a significant increase (p = .03) in knowledge of breast cancer and a significant decrease (p = .000) in fatalism scores compared to those individuals in the control group. The intervention group had a 61% participation rate in screening, while the control group had a 39% participation rate in screening. However, the difference was not statistically significant at the .05 level (p = .07). ^ Conclusions. Results demonstrate that culturally sensitive breast cancer education is successful in increasing knowledge and decreasing cancer fatalism. While there was a trend toward behavior change in the intervention group, more research needs to be done in this area. ^

Prime and boost vaccination against HER2/neu in breast cancer

Breast cancer is the most common cancer among women with approximately 180,000 new cases being diagnosed yearly in the United States (1). HER2/neu gene amplification and subsequent protein overexpression is found in 20–30% of breast cancer patients and can lead to the promotion of various metastasis-related properties (2–4) and/or resistance to cancer therapies such as chemotherapy and radiation (5). ^ The protein product of the HER2/neu gene, p185, is a proven target for immunological therapy. Recently, passive immunotherapy with the monoclonal antibody Trastuzumab® has validated an immunological approach to HER2/neu+ breast cancer. Immunity to HER2/ neu, when found in breast cancer patients, is of low magnitude. Vaccination-induced HER2/neu-specific antibodies and HER2/neu-specific cytotoxic T cells could result in long-lived immunity with therapeutic benefit. Many features of DNA vaccines and attenuated viral vectors may contribute to the efficacy of prime-boost vaccination. In particular, vaccines capable of eliciting strong cell-mediated immunity are thought to hold the greatest promise for control of cancer (6–9). ^ To optimize cellular immunization to HER2/neu in my study, the HER2/neu gene was presented to the immune system using a priming vector followed by a second vector used as the boost. In both animals and humans, priming with DNA and boosting with a poxviruses, vaccinia or canarypox appears to be particularly promising for induction of a broad immune responses (10). ^ I tested three gene vaccines encoding the HER2/neu gene: (1) a plasmid, SINCP, that contains part of the genome of Sindbis virus; (2) Viral Replicon Particles (VRP) of Venezuela Equine Encephalitis virus (VEE) and (3) E1/E2a-deleted human Type 5 Adenovirus. In SINCP and the VRP, the caspid and envelope genes of the virus were deleted and replaced with the gene for HER2/neu. SINCP-neu, VRP- neu and Adeno-neu when used alone were effective vaccines protecting healthy mice from challenge with a breast cancer cell line injected in the mammary fat pad or injected i.v. to induce experimental lung metastasis. However, SINCP-neu, VRP-neu or Adeno-neu when used alone were not able to prolong survival of mice in therapeutic models in which vaccination occurred after injection of a breast cancer cell line. ^ When the vaccines were combined in a mixed regimen of a SINCP- neu prime VRP-neu or Adeno-neu boost, there was a significant difference in tumor growth and survival in the therapeutic vaccine models. In vitro assays demonstrated that vaccination with each of the three vaccines induced IgG specific for p185, the gene product of HER2/neu, induced p185-specific T lymphocytes, as measured by tetramer analysis. Vaccination also induced intracellular INF-γ and a positive ELISPOT assay. These findings indicate that SINCP-neu, VRP-neu and Adeno-neu, used alone or in combination, may have clinical potential as adjuvant immunotherapy for the treatment of HER2/neu-expressing tumors. ^

Mechanism of IFN-induced apoptosis and sensitization by the proteasome inhibitor bortezomib in bladder cancer cells

Bladder cancer is the fifth most common cancer with more than 50,000 cases diagnosed each year. Interferon-α (IFNα) is mostly used in combination with BCG for the treatment of transitional cell carcinoma (TCC). To examine the effects of IFNα on bladder cancer cells, I analyzed a panel of 20 bladder cancer cell lines in terms of their sensitivity to IFNα-induced apoptosis and the underlying mechanisms. I identified three categories: cells that die after 48hr, after 72h, and cells resistant even after 72hr of IFNα treatment. Examination of the IFN-signal transduction pathway revealed that the defect was not due to abrogation of IFN signaling. Further analysis demonstrated dependency of IFN-induced apoptosis on caspase-8, implicating the role of death receptors in IFN-induced cell death. Of the six most-IFN-sensitive cell lines, the majority upregulated Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) at the mRNA and protein level and IFN-induced cell death was mediated through TRAIL, while a minority of the most IFN-sensitive cells undergo apoptosis through a TNFα-dependent mechanism. IFNα resistance was due to either absence of TRAIL upregulation at the mRNA or protein level, resistance to exogenous rhTRAIL itself or lack of sensitization to IFN-induced cell death. Downregulation of XIAP, or XIAP inactivation through its regulator NFκB has been reported to sensitize tumor cells to death receptor-induced cell death. Baseline and IFN-inducible XIAP levels were examined in the most and least IFN-sensitive cells, knocking down XIAP and the p65 subunit of NFκB enhanced IFN-induced cell death, implicating XIAP downregulation as a mechanism through which bladder cancer cells are sensitized to IFN-induced apoptosis. To determine whether or not the proteasome inhibitor Bortezomib (BZ) sensitizes bladder cancer cells to IFN-induced cell death, the combined effects of IFN+BZ and the underlying molecular mechanisms were examined both in vitro and in vivo using two bladder xenograft models. In both models, tumor growth inhibition was the result of either increased cell death of tumor cells exerted by the two agents and/or inhibition of angiogenesis. In vitro, MAP downregulation in response to the combined treatment of IFN+BZ accounts for one of the mechanisms mediating IFN+BZ cell death in bladder cancer cells. ^

In-vivo diffusion tensor imaging of rat spinal cord with a phased array coil at 7T

Magnetic resonance imaging (MRI) is a non-invasive technique that offers excellent soft tissue contrast for characterizing soft tissue pathologies. Diffusion tensor imaging (DTI) is an MRI technique that has shown to have the sensitivity to detect subtle pathology that is not evident on conventional MRI. ^ Rats are commonly used as animal models in characterizing the spinal cord pathologies including spinal cord injury (SCI), cancer, multiple sclerosis, etc. These pathologies could affect both thoracic and cervical regions and complete characterization of these pathologies using MRI requires DTI characterization in both the thoracic and cervical regions. Prior to the application of DTI for investigating the pathologic changes in the spinal cord, it is essential to establish DTI metrics in normal animals. ^ To date, in-vivo DTI studies of rat spinal cord have used implantable coils for high signal-to-noise ratio (SNR) and spin-echo pulse sequences for reduced geometric distortions. Implantable coils have several disadvantages including: (1) the invasive nature of implantation, (2) loss of SNR due to frequency shift with time in the longitudinal studies, and (3) difficulty in imaging the cervical region. While echo planar imaging (EPI) offers much shorter acquisition times compared to spin-echo imaging, EPI is very sensitive to static magnetic field inhomogeneities and the existing shimming techniques implemented on the MRI scanner do not perform well on spinal cord because of its geometry. ^ In this work, an integrated approach has been implemented for in-vivo DTI characterization of rat spinal cord in the thoracic and cervical regions. A three element phased array coil was developed for improved SNR and extended spatial coverage. A field-map shimming technique was developed for minimizing the geometric distortions in EPI images. Using these techniques, EPI based DWI images were acquired with optimized diffusion encoding scheme from 6 normal rats and the DTI-derived metrics were quantified. ^ The phantom studies indicated higher SNR and smaller bias in the estimated DTI metrics than the previous studies in the cervical region. In-vivo results indicated no statistical difference in the DTI characteristics of either gray matter or white matter between the thoracic and cervical regions. ^

Characterization of cytochrome P450 4F subfamily: Response in traumatic brain injury and gene regulation

CYP4F enzymes metabolize endogenous molecules including arachidonic acid, leukotrienes and prostaglandins. The involvement of these eisosanoids in inflammation has led to the hypothesis that CYP4Fs may modulate inflammatory conditions after traumatic brain injury (TBI). In rat, TBI elicited changes in mRNA expression of CYP4Fs as a function of time in the cerebrum region. These changes in CYP4F mRNA levels inversely correlated with the cerebral leukotriene B4 (LTB4) level following injury at the same time points. TBI also resulted in changes in CYP4F protein expression and localization around the injury site, where CYP4F1 and CYP4F6 immunoreactivity increased in surrounding astrocytes and CYP4F4 immunoreactivity shifted from endothelia of cerebral vessels to astrocytes. The study with rat primary astrocytes indicated that pro-inflammatory cytokines TNFα and IL-1β could affect the transcription of CYP4Fs to a certain degree, whereas the changing pattern in the primary astrocytes appeared to be different from that in the in vivo TBI model.^ In addition, the regulation of CYP4F genes has been an unsolved issue although factors including cytokines and fatty acids appear to affect CYP4Fs expression in multiple models. In this project, HaCaT cells were used as an in vitro cellular model to define signaling pathways involved in the regulation of human CYP4F genes. Retinoic acids inhibited CYP4F11 expression, whereas cytokines TNFα and IL-1β induced transcription of CYP4F11 in HaCaT cells. The induction of CYP4F11 by both cytokines could be blocked by a JNK specific inhibitor, indicating the involvement of the JNK pathway in the up-regulation of CYP4F11. Retinoic acids are known to function in gene regulation through nuclear receptors RARs and RXRs. The RXR agonist LG268 greatly induced transcription of CYP4F11, whereas RAR agonist TTNPB obviously inhibited CYP4F11 transcription, indicating that the down-regulation of CYP4F11 by retinoic acid was mediated by RARs, and that inhibition of CYP4F11 by retinoic acid may also be related to the competition for RXR receptors. Thus, the CYP4F11 gene is regulated by signaling pathways including the RXR pathway and the JNK pathway. In contrast, the regulation mechanism of other CYP4Fs by retinoic acids appears to be different from that of CYP4F11.^

Comparison of Bayesian and frequentist approaches in clinical outcomes between hospitals

Many public health agencies and researchers are interested in comparing hospital outcomes, for example, morbidity, mortality, and hospitalization across areas and hospitals. However, since there is variation of rates in clinical trials among hospitals because of several biases, we are interested in controlling for the bias and assessing real differences in clinical practices. In this study, we compared the variations between hospitals in rates of severe Intraventricular Haemorrhage (IVH) infant using Frequentist statistical approach vs. Bayesian hierarchical model through simulation study. The template data set for simulation study was included the number of severe IVH infants of 24 intensive care units in Australian and New Zealand Neonatal Network from 1995 to 1997 in severe IVH rate in preterm babies. We evaluated the rates of severe IVH for 24 hospitals with two hierarchical models in Bayesian approach comparing their performances with the shrunken rates in Frequentist method. Gamma-Poisson (BGP) and Beta-Binomial (BBB) were introduced into Bayesian model and the shrunken estimator of Gamma-Poisson (FGP) hierarchical model using maximum likelihood method were calculated as Frequentist approach. To simulate data, the total number of infants in each hospital was kept and we analyzed the simulated data for both Bayesian and Frequentist models with two true parameters for severe IVH rate. One was the observed rate and the other was the expected severe IVH rate by adjusting for five predictors variables for the template data. The bias in the rate of severe IVH infant estimated by both models showed that Bayesian models gave less variable estimates than Frequentist model. We also discussed and compared the results from three models to examine the variation in rate of severe IVH by 20th centile rates and avoidable number of severe IVH cases. ^

Advanced Protein Modeling Method: Benchmarking and Applications in Computer-Aided Drug Discovery

Development of homology modeling methods will remain an area of active research. These methods aim to develop and model increasingly accurate three-dimensional structures of yet uncrystallized therapeutically relevant proteins e.g. Class A G-Protein Coupled Receptors. Incorporating protein flexibility is one way to achieve this goal. Here, I will discuss the enhancement and validation of the ligand-steered modeling, originally developed by Dr. Claudio Cavasotto, via cross modeling of the newly crystallized GPCR structures. This method uses known ligands and known experimental information to optimize relevant protein binding sites by incorporating protein flexibility. The ligand-steered models were able to model, reasonably reproduce binding sites and the co-crystallized native ligand poses of the β2 adrenergic and Adenosine 2A receptors using a single template structure. They also performed better than the choice of template, and crude models in a small scale high-throughput docking experiments and compound selectivity studies. Next, the application of this method to develop high-quality homology models of Cannabinoid Receptor 2, an emerging non-psychotic pain management target, is discussed. These models were validated by their ability to rationalize structure activity relationship data of two, inverse agonist and agonist, series of compounds. The method was also applied to improve the virtual screening performance of the β2 adrenergic crystal structure by optimizing the binding site using β2 specific compounds. These results show the feasibility of optimizing only the pharmacologically relevant protein binding sites and applicability to structure-based drug design projects.