sa.66. in silico modeling of intranasal insulin b:9-23 peptide therapy in the non-obese diabetic...
TRANSCRIPT
S102 Abstracts
Mechanistic mathematical modeling can inform experimen-tal design of preclinical laboratory studies to identifyoptimal treatment approaches for reversing diabetes inNOD mice. Here we use the T1D PhysioLab® platform, amodel of T1D pathogenesis in the NOD mouse, to predictoptimal dosing regimens for oral insulin/anti-CD3 or oralinsulin/exendin-4 combination therapies and identify poten-tial mechanisms of action. Various hypotheses of diseasepathogenesis are represented in the model as differentvirtual mice, calibrated to reproduce laboratory data on theefficacy of each monotherapy. Variations in the dose andtiming of administration of the individual agents weresimulated in the virtual mice to identify treatment protocolsthat maximize the likelihood of observing diabetes reversalin the laboratory. Simulations predict that optimal efficacyof oral insulin/anti-CD3 combination therapy is achieved bydelaying initiation of anti-CD3 treatment relative to oralinsulin. The delay is predicted to support a continuousinduction of regulatory T cells (Tregs). In contrast, simula-tions predict that optimal efficacy of oral insulin/ exendin-4combination therapy is achieved when the two agents aregiven simultaneously early in diabetes progression, ensuringsufficient β-cell function during expansion of Tregs. Treat-ment protocols have been developed to test these in silicopredictions in the laboratory. If confirmed, these predictionscould potentially inform future clinical trial design in humanpatients.
doi:10.1016/j.clim.2008.03.286
Sa.66. In Silico Modeling of Intranasal Insulin B:9-23Peptide Therapy in the Non-Obese Diabetic (NOD)Mouse: Mechanisms of Action, Protocol Optimization,and Implications for Clinical Trial DesignJason Chan,1 Yanan Zheng,1 Georgia Fousteri,2 Matthias vonHerrath,2 Chan Whiting.3 1Entelos, Inc., Foster City, CA; 2LaJolla Institute for Allergy and Immunology, La Jolla, CA;3Ingenuity Systems, Inc., Redwood City, CA
Effective antigen-specific strategies to treat or preventtype 1 diabetes (T1D) in humans are still lacking, due inpart to our poor understanding of the therapeuticmechanisms of action and the influence of various factorson the efficacy of treatment regimens. Intranasal insulinB:9-23 peptide therapy (i.n. B:9-23) has been reported toinduce tolerance in prediabetic NOD mice with varyinglevels of success. We used the T1D PhysioLab® platform,a large-scale mathematical model of T1D pathogenesis inthe NOD mouse, to investigate the possible mechanismsunderlying the efficacy of i.n. B:9-23 and evaluate theimpact of dose, frequency of administration, and age ontherapeutic outcome. Various hypotheses of diseasepathogenesis are represented in the model as differentvirtual mice. The simulated response of each virtualmouse to four different i.n. B:9-23 treatment protocols aswell as other published therapies (e.g., anti-CD3, anti-B7,anti-CD8, and deletion of CD4, CD8, or B cells) repro-duced published data. Simulations in these virtual micepredict that NOD mice will be protected from diabetes athigh or low, but not intermediate doses of i.n. B:9-23.
Surprisingly, a high dosing frequency is predicted toinhibit tolerance induction through deletion of regulatoryT cells before sufficient expansion has occurred to providetherapeutic benefit. Finally, treatment is predicted to bemost effective if administered early enough to reverse orhalt β-cell loss. Confirmation of these predictions in vivowill establish key criteria to consider for clinical trialdesign and translation of this therapeutic strategy tohumans.
doi:10.1016/j.clim.2008.03.287
Sa.67. Rapid Identification of MHC Class I-restrictedAntigens Relevant to Autoimmune Diabetes UsingRetrogenic T CellsRodolfo Chaparro,1 Amanda Burton,2 David Serreze,3
Dario Vignali,2 Teresa DiLorenzo.1 1Albert Einstein Collegeof Medicine, Bronx, NY; 2St. Jude Children's ResearchHospital, Memphis, TN; 3The Jackson Laboratory, BarHarbor, ME
We have developed a novel strategy for the rapididentification of CD8+ T cell epitopes relevant to type 1diabetes in the context of the nonobese diabetic (NOD) mousemodel of disease. Obtaining the large number of antigen-sensitive monospecific T cells required for conventionalantigen discovery methods has historically been problematicdue to (1) difficulties in culturing autoreactive CD8+ T cellsfrom NOD mice and/or (2) the large time and resourceinvestments required for the generation of transgenic NODmice. We circumvented these problems by exploiting therapid generation time of retrogenic (Rg) mice, relative totransgenic mice, as a novel source of sensitive monospecificCD8+ Tcells, using the diabetogenic AI4 Tcell receptor on NOD.SCID and NOD.Rag1-/- backgrounds as a model. Rg AI4 T cellsare diabetogenic in vivo, demonstrating for the first time thatRg mice are a means for assessing the pathogenic potential ofCD8+ T cell receptor specificities. In order to obtain asufficient number of Rg CD8+ T cells for antigen screens, weoptimize a method for their in vitro expansion that resulted ina ~500 fold expansion. We demonstrate the high sensitivityand specificity of expanded Rg AI4 Tcells in the contexts of (1)specific peptide challenge, (2) islet cytotoxicity, and (3) theirability to resolve previously defined mimotope candidatesfrom a positional scanning peptide library. Our method is thefirst to combine the speed of Rg technology with an optimizedin vitro Rg T cell expansion protocol to enable the rapiddiscovery of T cell antigens.
doi:10.1016/j.clim.2008.03.288
Sa.68. Ins2 Expression from NOD Bone MarrowDerived Cells Does Not Prevent or Delay DiabetesDevelopment in NOD-Ins2 Knockout MiceAinhoa Martin-Pagola, Antonello Pileggi,Francesco Vendrame, Elsie Zahr, Ruth Molano,Camillo Ricordi, Alberto Pugliese. University of MiamiLeonard M. Miller School of Medicine, Miami, FL