A hazardous and not fully understood aeroelastic phenomenon is the transonic dip,the decrease in flutter dynamic pressure that occurs for most aircraft configurationsin transonic flows. The difficulty of predicting this phenomenon forces aircraft manufacturersto run long and costly flight test campaigns to demonstrate flutter-free behaviourof their aircraft at transonic Mach numbers.In this project, subsonic and transonic flutter calculations for the KTH-NASA genericfighter research model have been performed and compared to existing experimentalflutter data from wind tunnel tests performed at NASA Langley in 2016. For the fluttercalculations, industry-standard linear panel methods have been used together with afinite element model from NASTRAN.Further, an alternative approach for more accurate transonic flutter predictions usingthe full-potential solver Phi has been investigated. To predict flutter using this newmethodology a simplified structural model has been used together with aerodynamicmeshes of the main wing. The purpose of the approach was to see if it was possibleto find a method that was more accurate than panel methods in the transonic regimewhilst still being suitable for use during iterative design processes.The results of this project demonstrated that industry-standard linear panel methodssignificantly over-predict the flutter boundary in the transonic regime. It was alsoseen that the flutter predictions using Phi showed potential, being close to the linearresults for the same configuration as tested in Phi. For improved transonic accuracy inPhi, an improved transonic flow finite element formulation could possibly help .Another challenge with Phi is the requirement of an explicit wake from all liftingsurfaces in the aerodynamic mesh. Therefore, a method for meshing external storeswith blunt trailing edges needs to be developed. One concept suggested in this projectis to model external stores in "2.5D", representing external stores using airfoils withsharp trailing edges. / Ett farligt och inte helt utrett aeroelastiskt fenomen är den transoniska dippen, minskningeni dynamiska trycket vid fladder som inträffar för de flesta flygplan i transoniskaflöden. Svårigheten i att prediktera detta fenomen tvingar flygplanstillverkare attbedriva tidskrävande och kostsam flygprovsverksamhet för att demonstrera att derasflygplan ej uppvisar fladderbeteende i transonik inom det tilltänkta användningsområdet.I detta projekt har fladderberäkningar genomförts i både underljud och transonikför en generisk stridsflygplansmodell i skala 1:4 ämnad för forskning, byggd som ettsamarbete mellan KTH och NASA. Beräkningarna har också jämförts med fladderresultatfrån vindtunnelprov genomförda vid NASA Langley under sommaren 2016. Förfladderberäkningarna har industri-standarden linjära panelmetoder används tillsammansmed en befintlig finit element modell för användning i NASTRAN.Vidare har ett alternativt tillvägagångssätt för att förbättra precisionen i transoniskafladderresultat genom att använda potentiallösaren Phi undersökts. En förenkladstrukturmodell har använts tillsammans med aerodynamiska nät av huvudvingen föratt prediktera fladder. Syftet med denna metodik var att undersöka om det var möjligtatt hitta en metod som i transoniska flöden var mer exakt än panelmetoder men somfortfarande kunde användas i iterativa design processer.Resultaten från detta projekt visade att linjära panelmetoder, som de som används iindustrin, är signifikant icke-konservativa gällande fladdergränsen i transonik. Resultatenfrån Phi visade potential genom att vara nära de linjära resultaten som räknadesfram med hjälp av panelmetoder för samma konfiguration som i Phi. För ökad transonisknoggrannhet i Phi kan möjligen en förbättrad transonisk element-formuleringhjälpa.En annan utmaning med Phi är kravet på en explicit vak från alla bärande ytor idet aerodynamiska nätet. Därför behöver det utvecklas en metodik för nätgenereringav yttre laster med trubbiga bakkanter. Ett koncept som föreslås i denna rapport är attmodellera yttre laster i "2.5D", där alla yttre laster beskrivs genom att använda vingprofilermed skarpa bakkanter.

This study analyses the ways in which gender relations are expressed and articulated through the use of folk proverbs amongst Pashto-speaking people of Pakistan. Previous work on Pashto proverbs have romanticised proverbs as a cultural asset and a source of Pashtun pride and ethnic identity, and most studies have aimed to promote or preserve folk proverbs. However, there is little recognition in previous literature of the sexist and gendered role of proverbs in Pashtun society. This study argues that Pashto proverbs encode and promote a patriarchal view and sexist ideology, demonstrating this with the help of proverbs as text as well as proverbs performance in context by Pashto speakers. The analysis is based on more than 500 proverbs relating to gender, collected from both published sources and through ethnographic fieldwork in the Khyber Pakhtunkhwa province of Pakistan. Qualitative data was collected through 40 interviews conducted with Pashto-speaking men and women of various ages and class/educational backgrounds, along with informal discussions with local people and the personal observations of the researcher. The study is informed by a combination of theoretical approaches including folkloristics, feminist sociology and sociolinguistics. While establishing that patriarchal structures and values are transmitted through proverbs, the study also reveals that proverbs’ meanings and messages are context-bound and women may, therefore, use proverbs in order to discuss, contest and (sometimes) undermine gender ideologies. More specifically, it is argued that: (1) Proverbs as ‘wisdom texts’ represent the viewpoint of those having the authority to define proper and improper behaviour, and as such, rather than objective reality represent a partial and partisan reality which, in the context of the present research, is sexist and misogynist. (2) While proverbs as ‘texts’ seem to present a more fixed view of reality, proverbs as ‘performance in context’ suggest that different speakers may use proverbs for different strategic purposes, such as to establish and negotiate ethnic and gendered identities and power which varies on the basis of gender, age, ethnicity, and class of the interlocutors. The thesis concludes that, rather than considering folk proverbs as ‘factual’ and ‘valuable’ sources of cultural expression, scholars should pay more attention to their ‘performatory’, ‘derogatory’ and ‘declaratory’ aspects as these often relegate women (and ‘other’, weaker groups) to a lesser position in society.

Vaccinia virus (VACV) is the most intensively studied orthopoxvirus and acts as an excellent model to investigate host-pathogen interactions. VACV encodes about 200 proteins, many of which modulate the immune response. This study focusses on two of these: C16 and C4, that share 43.7 % amino acid identity. Given the sequence similarity, we explored whether C16 and C4 have any shared functions, whilst also searching for novel functions. To gain mechanistic insight, we sought to identify binding partners and determine the residues responsible. C16 has two reported functions. Firstly, it inhibits DNA-PK-mediated DNA sensing, and this study found that C4 can perform this function as well. Like C16, C4 associates with the Ku heterodimer to block its binding to DNA leading to reduced production of cytokines and chemokines. For both proteins, the function localised to the C termini and was abrogated by mutating three residues. Secondly, C16 induces a hypoxic response by binding to PHD2. This function was mapped to the N-terminal 156 residues and a full length C16 mutant (D70K,D82K) lost the ability to induce a hypoxic response. In contrast, C4 did not bind PHD2. C4 inhibits NF-κB signalling by an unknown mechanism. Reporter gene assays showed that C16 also suppresses NF-κB activity and, intriguingly, this was carried out by both the N and C termini. C16 acts at or downstream of p65 and the N terminus of C16 associated with p65 independently of PHD2-binding. Conversely, C4 acted upstream of p65, did not display an interaction with p65, and the function was restricted to its C-terminal region. Novel binding partners were identified by a screen utilising tandem mass tagging and mass spectrometry, and selected hits were validated. The C terminus of C16 associated with VACV protein K1, a known NF-κB inhibitor. Additionally, C16 bound to the transcriptional regulator ARID4B. C4 did not interact with these proteins, but the N-terminal region of C4 associated with filamins A and B. The functional consequences of these interactions remain to be determined. In vivo, C4 and C16 share some redundancy in that a double deletion virus exhibits an attenuated virulence phenotype that is not observed by single deletion viruses in the intradermal model of infection. However, non-redundant functions also contribute to virulence in that both single deletion viruses display attenuated virulence compared to a wild-type Western Reserve virus in the intranasal model of infection. Data presented also reveal that C4 inhibits the recruitment of immune cells to the site of infection, as was previously described for C16. Overall, this investigation highlights the complexity of host-pathogen interactions showing that VACV encodes two multifunctional proteins with both shared and unique functions. Moreover, their inhibition of DNA-PK emphasises the importance of this PRR as a DNA sensor in vivo.

Biofilmes são comunidades bacterianas complexas encapsuladas em matrizes poliméricas autoproduzidas e podem se desenvolver em superfícies inertes ou tecidos vivos. A formação do biofilme é um importante fator de virulência, pois permite à bactéria resistir às respostas do hospedeiro e à terapia antimicrobiana. Devido a essa elevada resistência aos antimicrobianos, é difícil estabelecer uma estratégia eficaz para o tratamento de infecções com formação de biofilmes, levando a falhas na erradicação das mesmas. Nesse contexto, o objetivo do presente estudo é desenvolver um modelo farmacocinético/farmacodinâmico (PK/PD) para descrever o efeito do ciprofloxacino (CIP) na presença de biofilmes de Pseudomonas aeruginosa (ATCC 27853), visto que a modelagem PK/PD de antimicrobianos é uma ferramenta útil na escolha de regimes posológicos que atinjam o efeito bactericida máximo, minimizando o desenvolvimento de resistência. Para atingir esse objetivo, inicialmente um método analítico por CLAE/fluorescência foi desenvolvido para quantificar o CIP em amostras de plasma e microdialisado. O método desenvolvido foi simples, rápido e com sensibilidade adequada para corretamente caracterizar a farmacocinética plasmática e pulmonar do CIP. Posteriormente, um modelo animal de infecção pulmonar crônica foi adaptado da literatura e padronizado, permitindo a investigação da distribuição pulmonar do CIP em ratos Wistar sadios e infectados. Para tal, bactérias foram imobilizadas em beads de alginato a fim de manter a infecção por até 14 dias com cargas bacterianas superiores à 108 UFC/pulmão. Estudo de microdiálise foi então conduzido para avaliar as concentrações livres de CIP após administração intravenosa de 20 mg/kg. A análise não-compartimental (NCA) e a modelagem farmacocinética populacional (PopPK) dos dados foram realizadas nos softwares Phoenix® e NONMEM®, respectivamente. Diferenças significativas foram observadas no clearance plasmático (1,59 ± 0,41 L/h/kg e 0,89 ± 0,44 L/h/kg) e na constante de eliminação (0,23 ± 0,04 h-1 e 0,14 ± 0,08 h-1) para ratos sadios e infectados, resultando em uma exposição plasmática maior nos animais infectados (ASC0-∞ = 27,3 ± 12,1 μg·h/mL) quando comparados com os animais sadios (ASC0-∞ = 13,3 ± 3,5 μg·h/mL) ( = 0,05). Apesar da maior exposição plasmática, quando comparados com os animais saudáveis (fT = 1,69), animais infectados apresentaram uma penetração pulmonar quatro vezes menor (fT = 0,44). Diferenças na constante de eliminação pulmonar não foram observadas. Dados plasmáticos e pulmonares foram simultaneamente descritos por modelo PopPK constituído de compartimentos venoso e arterial, dois compartimentos representativos de duas regiões pulmonares distintas e dois compartimentos periféricos, representando outros tecidos que não os pulmões. Um clearance pulmonar foi adicionado ao modelo apenas para os dados de microdiálise dos animais infectados (CLlung = 0,643 L/h/kg) afim de explicar a exposição tecidual diminuída. O modelo desenvolvido descreveu, com sucesso, os dados plasmáticos e teciduais de animais sadios e infectados, permitindo a correta caracterização das alterações observadas na disposição plasmática e pulmonar do CIP decorrentes da infecção com biofilme. Para os estudos de farmacodinâmica, o efeito bactericida do CIP frente a biofilmes e células planctônicas de P. aeruginosa foi simultaneamente avaliado através do uso de curvas de morte bacteriana. Para a construção destas curvas, biofilmes de P. aeruginosa foram formados na superfície de blocos de acrílico e sua formação foi confirmada pelo ensaio cristal violeta e por microscopia eletrônica de varredura. Os blocos foram expostos a concentrações constantes de CIP (de 0,0625 a 10 μg/mL) e, em tempos pré-determinados, células planctônicas e de biofilmes eram amostradas para quantificação. Um modelo semi-mecanístico que incorpora um modelo Emax sigmoidal foi utilizado para descrever o efeito do CIP frente a ambos estilos de vida bacteriano. Uma subpopulação pré-existente com menor suscetibilidade ao CIP foi incluída no modelo e o efeito do CIP nesta subpopulação também foi descrito pelo modelo Emax sigmoidal. A comparação dos parâmetros estimados pelo modelo demonstrou que o efeito in vitro do CIP é maior para as células planctônicas (EC50 = 0,259 mg/L e 0,123 mg/L e Emax = 2,25 h-1 e 5,59 h-1 para biofilmes e planctônicas, respectivamente). A potência estimada do CIP para a subpopulação resistente foi muito menor para ambos estilos de vida bacteriano (EC50 = 2,71 mg/L e 1,15 mg/L para biofilmes e planctônicas, respectivamente). Os modelos desenvolvidos podem ser utilizados para a simulação de cenários não testados e servir como uma ferramenta para guiar a escolha dos regimes posológicos adequados, contribuindo para o sucesso terapêutico no tratamento de infecções associadas à biofilmes. / Biofilms are complex bacterial communities enclosed in self-produced polymeric matrices that can develop in inert surfaces or living tissues. Biofilm formation is an important virulence factor that allows bacteria to resist host responses and antibacterial agents. Due to this high resistance to antibiotics, it is difficult to establish an efficacious strategy for treatment of infections with biofilm formation leading to failure in infection eradication. In this context, the goal of this study was to develop a pharmacokinetic/pharmacodynamic (PK/PD) model to describe the antimicrobial effect of ciprofloxacin (CIP) in the presence of biofilms of Pseudomonas aeruginosa (ATCC 27853), since PK/PD modeling for antibacterial agents can be a useful tool to choose dosing regimens and to achieve the maximum bactericidal effect, minimizing the development of resistance. To reach this goal, firstly an analytical method based on HPLC/fluorescence was developed in order to quantify CIP in plasma and lung microdialysate. The developed method was simple, fast and with enough sensibility to proper characterize CIP plasma and lung pharmacokinetics. Secondly, an animal model of chronic lung infection was adapted from literature and standardized, allowing the analysis of CIP lung distribution in infected and healthy Wistar rats. Bacteria were immobilized in alginate beads prior to inoculation to Wistar rats in order to sustain the pneumonia for 14 days, maintaining a bacterial load superior to 108 CFU/lung. A microdialysis study was then conducted to evaluate free CIP concentrations after an intravenous administration of 20 mg/kg. Non-compartimental analysis (NCA) and populational PK modeling (PopPK) of the data were performed in Phoenix® and NONMEM®, respectively. Statistical differences were observed in the plasma clearance (1.59 ± 0.41 L/h/kg and 0.89 ± 0.44 L/h/kg) and elimination rate constant (0.23 ± 0.04 h-1and 0.14 ± 0.08 h-1) for healthy and infected rats, respectively, resulting in a significantly higher CIP plasma exposure in infected rats (AUC0-∞ = 27.3 ± 12.1 μg·h/mL) compare to healthy animals (AUC0-∞ = 13.3 ± 3.5 μg·h/mL) ( = 0.05). Besides the plasma exposure, a four times lower pulmonary penetration was observed in infected rat’s lungs (fT = 0.44) in comparison to healthy animals (fT = 1.69), with no significant differences in the lung elimination rate constant. Plasma and lung data were simultaneously fitted using a PopPK model consisting of an arterial and a venous compartment, two compartments representing different regions of the lungs and two peripheral distribution compartments, representing tissues other than lungs. A lung clearance was added to the model for infected animals (CLlung = 0.643 L/h/kg) to explain the lower tissue exposure. The model successfully described the plasma and microdialysis data from both, healthy and infected rats and allowed to correctly describe the changes in CIP plasma and lung disposition in biofilm infections. For the pharmacodynamic studies, CIP bactericidal effect against Pseudomonas aeruginosa biofilms and planktonic shedding cells were simultaneously evaluated using the time-kill curves approach. For the time-kill curves construction, P. aeruginosa biofilms were formed in acrylic blocks, which was confirmed by the crystal violet assay and scanning electron microscopy. The blocks were placed in flasks containing Mueller-Hinton growth medium and exposed to constant CIP concentrations (ranging from 0.0625 to 10 μg/mL). At pre-determined time points, biofilm and planktonic cells were sampled for bacterial counting. A mechanism-based model which incorporates a sigmoidal Emax model was used to describe the CIP effect against P.aeruginosa in both llifestyles, biofilm and planktonic. The presence of a pre-existing resistant subpopulation was included in the model and also modeled with a sigmoidal Emax model to describe CIP effect in this subpopulation. Comparison of the parameter estimates showed that the in vitro effect of CIP is higher for planktonic cells (EC50 = 0.259 mg/L and 0.123 mg/L and Emax = 2.25 h-1 and 5.59 h-1 for biofilm and planktonic cells, respectively). CIP potency was much lower for the resistant subpopulation, for both bacteria lifestyles (EC50 = 2.71 mg/L and 1.15 mg/L for biofilm and planktonic, respectively). The developed models can be used to simulate untested scenarios and serve as a tool to guide dosing regimen selection, contributing for the therapeutic success of treatments of biofilm-associated infections.

本研究主要在於設計一種新式生物反應器，並應用於生產豬瘟病毒疫苗。首先根據所培養細胞的生長特性與原有生物反應器之缺點，改良成新式的生物反應器，並評估此新式生物反應器適用性、效能，以及所培養豬腎細胞之生長代謝情形與豬瘟病毒力價。整個實驗過程大致分為兩個部分，第一個部分探討細胞固定化培養之最適化培養條件與生長代謝情形，第二個部分探討豬瘟病毒培養之最適化培養條件與生長代謝情形。實驗結果發現豬腎細胞(PK-15)以批次方式培養於新式生物反應器，搭配著FIBRA-CEL®載體，成功的進行擴大培養，豬腎細胞最高的生長量達到2.29×109cells/300mL的細胞量。因此，改良之新式生物反應器可提供細胞優越的生長環境，具有擴大規模培養之潛力，可藉由此簡單設備、操作容易、成本低且低能源消耗之新式生物反應器達成細胞製品之生產基座。 / In this study, the production of PK-15 cell using immobilized animal cell culture in a novel bioreactor was investigated.We evaluated the serviceability and efficiency of a design-improved novel bioreactor for the growth and metabolic states of cultured PK-15 cells and the production of HC virus. The entire experiment includes two major stages: (1) investigation of the optimal conditions and metabolic states for the growth of immobilized cells, (2) investigation of the optimal conditions for the production of HC virus. Our results showed that immobilized PK-15 cells on the fibra-cell carries in the design-improved novel bioreactor exhibited their best growth of 2.29×109 cells/300mL.The immobilized conditions for cell culture, can provide a shearing stress of growth state, easy separation of cells from the culture mediu, and a operation of continuously feeding medium, leading to possibility growth of the high density cell and a long period of production;as a result, the efficiency of producing process is promoted. Here,our design-improved novel bioreactor is expected to provide an optimal growth environment of both the cells and viruses for the production of high-yielded, stable, and consistent cellular biological preparations. Furthermore, it will also provide the basis for the production of cell products with advantages of simple-equipped, easy-to-operate, low cost, and low energy consumption. / 致謝 i 中文摘要 ------------------------------------------------------------------------ ii 英文摘要 ------------------------------------------------------------------------ iii 目錄 ------------------------------------------------------------------------ iv 表目錄 ------------------------------------------------------------------------ vi 圖目錄 ------------------------------------------------------------------------ vii 第一章 緒論------------------------------------------------------------------ 1 第二章 文獻回顧------------------------------------------------------------ 3 2-1 豬腎傳代細胞(PK-15 cell) -------------------------------------- 3 2-2 豬瘟病毒------------------------------------------------------------ 4 2-2.1 豬瘟之歷史背景--------------------------------------------------- 4 2-2.2 豬瘟病毒之特性--------------------------------------------------- 7 2-2.3 豬瘟發生原因之探討--------------------------------------------- 11 2-3 生物反應器--------------------------------------------------------- 12 第三章 實驗材料與方法--------------------------------------------------- 21 3-1 細胞------------------------------------------------------------------ 21 3-2 細胞繼代培養------------------------------------------------------ 21 3-3 細胞冷凍保存------------------------------------------------------ 22 3-4 解凍細胞培養------------------------------------------------------ 22 3-5 病毒感染------------------------------------------------------------ 23 3-6 收集病毒------------------------------------------------------------ 24 3-7 豬瘟病毒力價測試------------------------------------------------ 24 3-8 細胞滾瓶培養------------------------------------------------------ 26 3-9 生物反應器操作--------------------------------------------------- 27 3-10 載體上細胞數的測定--------------------------------------------- 33 3-11 葡萄糖的測定------------------------------------------------------ 33 3-12 培養過程中pH值測定------------------------------------------- 34 第四章 結果與討論--------------------------------------------------------- 35 4-1 測試細胞貼附的材料--------------------------------------------- 35 4-2 細胞固定時間的比較--------------------------------------------- 36 4-3 測試不同比例的載體量培養豬腎細胞------------------------ 37 4-4 測試不同接種量--------------------------------------------------- 40 4-5 測試培養基流速對豬腎細胞生長的影響--------------------- 44 4-6 測試培養基停留於培養槽時間對豬腎細胞生長的影響--- 45 4-7 測試豬腎細胞暴露空氣時間對於生長的影響--------------- 47 4-8 測試Bellocell培養豬腎細胞(PK-15)可行性----------------- 49 4-9 測試利用新式生物反應器培樣豬瘟病毒--------------------- 50 第五章 結論與建議--------------------------------------------------------- 53 參考文獻 ------------------------------------------------------------------------ 55 表目錄 表1. 兔化豬瘟疫苗與組織培養豬瘟疫苗的比較------------------ 6 表2. 急性、慢性與遲發型豬瘟比較---------------------------------- 10 表3. .Growth of Various cell Lines in bellocell-500----------------- 18 表4. Comparison of SF-9 cell Growth and BEV production in Various Laboratory bioreators------------------------------------ 19 表5. Comparison of HEK293 Cell growth and Receptor X production in Cell Factory®/20 roller bottles and BelloCell-500Bioreactor------------------------------------------ 20 表6. Reed-Muench Methods法計算方法----------------------------- 26 表7. 比較不同材料培養PK-15 cell所用的載體量---------------- 50 表8. 細胞固定時間的比較所接細胞量與載體量------------------ 51 圖目錄 圖1. Liau提出以潮汐生物反應器圖--------------------------------- 17 圖2. Operation principle of Bellocell system------------------------- 18 圖3. 新式生物反應器(novel reactor)-潮汐式生物反應器(tidal typereactor)之運作流程圖--------------------------------------- 30 圖4. 比較不同材料培養PK-15 cell ---------------------------------- 63 圖5. 比較不同時間細胞的貼附量------------------------------------ 64 圖6. 測試的不同比例載體量培養豬腎細胞生長曲線------------ 65 圖7. 測試的不同比例載體量培養豬腎細胞培養過程glucose消耗趨勢------------------------------------------------------------ 66 圖8. 測試的不同比例載體量培養豬腎細胞培養過程pH變化------------------------------------------------------------------------ 67 圖9. 測試不同接細胞量培養在10g carrier生長曲線------------- 68 圖10. 測試不同接細胞量培養在10g 載體glucose消耗趨勢----- 69 圖11. 測試不同接細胞量培養在10g carrier pH趨勢--------------- 70 圖12. 測試流速對豬腎細胞生長的影響------------------------------ 71 圖13. 測試流速對細胞影響的葡萄糖消耗--------------------------- 72 圖14. 測試流速對豬腎細胞生長影響pH值-------------------------- 73 圖15. 測試培養基停留時間對豬腎細胞生長影響------------------ 74 圖16. 測試培養基holding時間對豬腎細胞生長的影響之葡萄糖趨勢--------------------------------------------------------------- 75 圖17. 測試培養基holding時間對細胞的影響之pH值趨勢------- 76 圖18. 測試豬腎細胞暴露空氣時間對生長的影響------------------ 77 圖19. 測試豬腎細胞暴露空氣對生長的影響葡萄糖消耗趨勢--- 78 圖20. 測試豬腎細胞暴露空氣對細胞生長的影養pH值趨勢---- 79 圖21. Bellocell反應器培養豬腎細胞---------------------------------- 80 圖22. Bellocell培養豬腎細胞葡萄糖消耗趨勢---------------------- 81 圖23. Bellocell培養豬腎細胞pH值趨勢----------------------------- 82 圖24. 測試利用新式生物反應器培養豬瘟病------------------------ 83

Antibiotics are among the most commonly prescribed drugs. Although the majority of these drugs were developed several decades ago, optimal dosage (dose, dosing interval and treatment duration) have still not been well defined. This thesis focuses on the development and evaluation of pharmacometric models that can be used as tools in the establishment of improved dosing strategies for novel and already clinically available antibacterial drugs. Infectious diseases are common causes of death in preterm and term newborn infants. A population pharmacokinetic (PK) model for gentamicin was developed based on data from a prospective study. Body-weight and age (gestational and post-natal age) were found to be major factors contributing to variability in gentamicin clearance and therefore important patient characteristics to consider for improved dosing regimens. A semi-mechanistic pharmacokinetic-pharmacodynamic (PKPD) model was also developed, to characterize in vitro bacterial growth and killing kinetics following exposure to six antibacterial drugs, representing a broad selection of mechanisms of action and PK as well as PD characteristics. The model performed well in describing a wide range of static and dynamic drug exposures and was easily applied to other bacterial strains and antibiotics. It is, therefore, likely to find application in early drug development programs. Dosing of antibiotics is usually based on summary endpoints such as the PK/PD indices. Predictions based on the PKPD model showed that the commonly used PK/PD indices were well identified for all investigated drugs, supporting that models based on in vitro data can be predictive of antibacterial effects observed in vivo. However, the PK/PD indices were sensitive to the study conditions and were not always consistent between patient populations. The PK/PD indices may therefore extrapolate poorly across sub-populations. A semi-mechanistic modeling approach, utilizing the type of models described here, may thus have higher predictive value in a dose optimization tailored to specific patient populations.

Adenoviruses have been extensively studied in the field of gene regulation, since their genes are subjected to a tightly controlled temporal expression during the virus lifetime. The early-to-late shift in adenoviral gene expression distinguishes two completely different programs in gene expression. The adenoviral L4-33K protein, which is the subject of this thesis, was previously implicated to be a key player in the transition from the early to the late phase of infection. Here we show that L4-33K activates late gene expression by functioning as a virus-encoded alternative RNA splicing factor activating splicing of transcripts containing weak 3’ splice sites; a feature common to the viral genes expressed at late times of infection. The splicing enhancer activity of L4-33K was mapped to a tiny arginine/serine (RS) repeat in the carboxyl-terminal domain of the protein. Also, the subcellular distribution to the nucleus with enrichment in the nuclear membrane and subnuclear redistribution to viral replication centers during a lytic infection was observed to depend on this motif. RS repeats are common features for the cellular splicing factors serine/arginine-rich (SR) proteins, which in turn are regulated by reversible phosphorylation. We further show that L4-33K is phosphorylated by two cellular protein kinases, the double-stranded DNA-dependent protein kinase (DNA-PK) and protein kinase A (PKA) in vitro. Interestingly, DNA-PK and PKA have opposite effects on the control of the temporally regulated L1 alternative RNA splicing. DNA-PK functions as an inhibitor of the late specific L1-IIIa pre-mRNA splicing whereas PKA functions as an activator of L1-IIIa pre-mRNA splicing. In summary, this thesis describes L4-33K as an SR protein related viral alternative splicing factor. A tiny RS repeat conveys splicing enhancer activity as well as redistribution of L4-33K to replication centers. Finally, DNA-PK and PKA that phosphorylates L4-33K are suggested to be novel regulatory factors controlling adenovirus alternative splicing.

Objetivo: O objetivo geral deste trabalho foi estabelecer modelo farmacocinéticofarmacodinâmico (modelo PK/PD) para descrever o perfil temporal do efeito bactericida do levofloxacino e do gatifloxacino contra Streptococcus pneumoniae. Método: Para alcançar este objetivo as seguintes etapas foram realizadas: i) foram validadas metodologias analíticas de SPE-HPLC para o gatifloxacino e HPLC para o levofloxacino e o gatifloxacino para quantificação destes em amostras de plasma, microdialisado tecidual e caldo de cultura; ii) foi avaliada a farmacocinética do gatifloxacino em roedores nas doses de 6 e 12 mg/kg via oral e 6 mg/kg via intravenosa (i.v.) e a biodisponibilidade oral foi determinada; iii) foram estabelecidas as condições ideais para microdiálise do gatifloxacino e as taxas de recuperação in vitro, por diálise (EE), retrodiálise (RD) e fluxo líquido zero (NNF) e in vivo, em tecido pulmonar e muscular, por retrodiálise e fluxo líquido zero. Essas recuperações foram utilizadas para determinar a penetração pulmonar do gatifloxacino após a administração i.v. bolus de 6 mg/kg a ratos Wistar sadios; iv) foram simuladas as concentrações livres pulmonares esperadas para humanos após tratamento com diferentes regimes de dosagem para o levofloxacino e o gatifloxacino em modelo de infecção in vitro frente a Streptococcus pneumoniae ATCC® 49619. Simulações de concentrações constantes múltiplas do MIC de cada fármaco também foram realizadas. As curvas de morte bacteriana por tempo obtidas foram modeladas com modelo PK/PD de Emax modificado, com auxílio do programa Scientist® v 2.01. Resultados e Conclusões: i) Os métodos analíticos por SPE-HPLC e HPLC para quantificação do gatifloxacino e do levofloxacino foram validados. As curvas foram lineares na faixa de 20 a 600 ng/mL para plasma e microdialisado tecidual de gatifloxacino e na faixa de 250 a 6000 ng/mL para caldo de cultura para ambos os fármacos, com r > 0,99, independente do método desenvolvido. Em plasma e microdialisado, a exatidão foi ≥ 94,3 %. A recuperação do gatifloxacino dos cartuchos de extração em fase sólida variou entre 95,6 e 99,7 %. A precisão não excedeu 5,8 % do CV. Em caldo de cultura, a exatidão foi ≥ 92,0 % e 93,4 % para o gatifloxacino e o levofloxacino, respectivamente. A precisão não excedeu 3,2 % e 4,2 % do CV para o levofloxacino e o gatifloxacino, respectivamente; ii) A avaliação farmacocinética demonstrou que os modelos abertos de dois compartimentos e de um compartimento com absorção de primeira ordem descreveram adequadamente os perfis plasmáticos após administração do gatifloxacino pelas vias i.v. e oral nas doses de 6 e 12 mg/kg, com CL de 0,9 ± 0,2 e 1,0 ± 0,3 L/h/kg, t½ de 3,3 ± 0,8 e 3,7 ± 0,3 h e Vd de 2,8 ± 0,4 e 3,1 ± 1,0 L/kg, respectivamente. Os parâmetros determinados por abordagem compartimental e não compartimental não diferiram significativamente para as duas vias investigadas (α = 0,05). A ASC0-∞ foi de 4,1 ± 1,6 e 6,6 ± 1,3 μg.h/mL após administração oral e i.v. das doses de 12 e 6 mg/kg, respectivamente, levando a uma biodisponibilidade de 31%. A constante de velocidade de absorção foi alta (5,0 ± 1,8 h-1) e a farmacocinética mostrou-se linear na faixa de doses investigada; iii) A recuperação das sondas de microdiálise in vitro por EE e RD para 80, 160 e 400 ng/mL de gatifloxacino foi de 33,5 ± 1,3%, 33,1 ± 1,2%, 31,8 ± 2,7% e 31,4 ± 2,6%, 33,1 ± 2,2%, 30,6 ± 3,3%, respectivamente. In vivo a recuperação por RD no músculo esquelético e pulmão de ratos Wistar foi de 29,1 ± 1,0% e 30,7 ± 1,4%, respectivamente. A recuperação por NNF in vitro e in vivo foi de 30,9 ± 2,9% e 29,0 ± 0,8%, respectivamente. Desse modo, concluiu-se que a recuperação foi constante e independente do método ou meio utilizado. Os perfis de concentração livre no músculo, pulmão e plasma de ratos Wistar foram virtualmente superpostos após dose de 6 mg/kg i.v., resultando em ASC similares de 3888 ± 734 ng.h/mL, 4138 ± 1071 ng.h/mL e 3805 ± 577 ng.h/mL, respectivamente (α = 0,05). O fator de distribuição tecidual foi de 1,02 e 1,08 para músculo e pulmão, respectivamente; iv) O modelo PK/PD empregado foi capaz de descrever o efeito do levofloxacino e do gatifloxacino contra o Streptococcus pneumoniae in vitro para todas as simulações investigadas. O EC50 médio para o levofloxacino (3,57 ± 2,16 mg/L) foi significativamente maior que o do gatifloxacino (0,95 ± 0,56 mg/L) quando regimes de doses múltiplas foram simulados. O mesmo foi observado para concentrações constantes, sendo o EC50,levofloxacino = 2,75 ± 0,45 mg/L e EC50,gatifloxacino = 1,03 ± 0,52 mg/L. O kmax foi estatisticamente semelhante para ambos os fármacos independente se foram simuladas concentrações flutuantes (kmax,levofloxacino = 0,40 ± 0,19 h-1; kmax,gatifloxacino = 0,48 ± 0,15 h-1) ou concentrações constantes (kmax,levofloxacino = 0,34 ± 0,06 h-1; kmax,gatifloxacino = 0,39 ± 0,23 h-1). Nenhum dos índices PK/PD foi capaz de prever o desfecho da infecção para todas as situações investigadas. O modelo PK/PD desenvolvido permitiu a comparação entre as duas fluorquinolonas e de diferentes posologias para cada fármaco, podendo ser utilizado para simular o efeito temporal de regimes de dosagem alternativos bem como para otimização da posologia desses fármacos para o tratamento da pneumonia adquirida na comunidade. / Objective: The aim of this work was to establish a pharmacokinetic-pharmacodynamic model (PK/PD model) to describe the profile of bactericidal effect over time of levofloxacin and gatifloxacin against Streptococcus pneumoniae. Method: To achieve this goal the following steps were carried out: i) an analytical method of SPE-HPLC to quantify gatifloxacin in plasma and tissue microdialysates, and an HPLC method for measuring levofloxacin and gatifloxacin in culture broth samples were developed and validated; ii) the pharmacokinetics of gatifloxacin in rodents after intravenous (6 mg/kg) and oral (6 and 12 mg/kg) administration was assessed as well as the oral bioavailability of the drug was determined; iii) microdialysis conditions for gatifloxacin were established and the recovery rates in vitro by dialysis (EE), retrodialysis (RD) and no-net-flux (NNF), and in vivo in lung and skeletal muscle tissue by RD and NNF were determined. Gatifloxacin tissue penetration in lung after intravenous administration (6 mg/kg) to healthy Wistar rats was determined; iv) levofloxacin and gatifloxacin free lung concentrations expected in humans following different dosing regimens of the drugs were simulated using Streptococcus pneumoniae ATCC® 49619 in vitro model of infection. The effect of constant concentrations multiples of MIC were also investigated. The time-kill curves obtained were modeled using an Emax modified model using Scientist® v. 2.01 software. Results and Conclusions: i) The analytical methods by SPE-HPLC and HPLC for quantifying gatifloxacin and levofloxacin were validated. Calibration curves were linear between 20-600 ng/mL for gatifloxacin in plasma and tissue microdialysate samples and between 250-6000 ng/mL for broth media for both drugs, with r > 0.99 independently of the method considered. The accuracy was ≥ 94.3 % for plasma and microdialysate. Gatifloxacin recovery from the solid phase extraction cartridges ranged from 95.6 to 99.7%. The precision did not exceed 5.8% of the CV. In broth media the accuracy was ≥ 92.0% and 94.3% for gatifloxacin and levofloxacin, respectively. The precision did not exceed 3.2% and 4.2% of the CV for levofloxacin and gatifloxacin, respectively; ii) Gatifloxacin experimental plasma profiles in rats were adequately fitted to a two-compartment model after intravenous and to a one compartment model with first order absorption after oral dosing. The total clearance (0.9 ± 0.2 and 1.0 ± 0.3 L/h/kg), the terminal half-life (3.3 ± 0.8 and 3.7 ± 0.3 h) and the apparent volume of distribution (2.8 ± 0.4 and 3.1 ± 1.0 L/kg) were statistically similar (α = 0.05) after i.v. and oral administration, by both model independent and compartmental approaches. The area under the curve was reduced after oral dosing (4.1 ± 1.6 μg.h/mL) in comparison to i.v. dosing (6.6 ± 1.3 μg.h/mL) leading to an oral bioavailability of 31%. The absorption was fast, with a constant rate of 5.0 ± 1.8 h-1. The results evidenced the linear pharmacokinetics of gatifloxacin in rodents in the dose range investigated; iii) Microdialysis recoveries determined in vitro by EE and RD at 80, 160 and 400 ng/mL resulted in 33.5 ± 1.3%, 33.1 ± 1.2%, 31.8 ± 2.7% and 31.4 ± 2.6%, 33.1 ± 2.2%, 30.6 ± 3.3%, respectively. In vivo recovery by RD in Wistar rat’s skeletal muscle and lung were 29.1 ± 1.0% and 30.7 ± 1.4%, respectively. Recoveries by no-net-flux in vitro and in vivo resulted in recoveries of 30.9 ± 2.9% and 29.0 ± 0.8%, respectively. In this way, it was shown that gatifloxacin recovery was constant and independent of the method or media used. Free skeletal muscle, lung and plasma profiles were virtually superimposed after i.v. administration of gatifloxacin 6 mg/kg dose resulting in similar area under the curve of 3888 ± 734 ng.h/mL, 4138 ± 1071 ng.h/mL and 3805 ± 577 ng.h/mL, respectively (α = 0.05). The tissue distribution factors were determined to be 1.02 and 1.08 for muscle and lung, respectively; iv) The PK/PD model used was able to describe the effect of levofloxacin and gatifloxacin against Streptococcus pneumoniae in vitro for all the regimens investigated. Levofloxacin EC50 (3.57 ± 2.16 mg/L) was higher than gatifloxacin (0.95 ± 0.56 mg/L) when multiple dosing regimens where simulated. Using constant concentrations, levofloxacin EC50 was also higher than gatifloxacin (EC50,levofloxacin = 2.75 ± 0.45 mg/L; EC50,gatifloxacin = 1.03 ± 0.52 mg/L). The kmax was statistically similar for both drugs independent of whether fluctuating (kmax,levofloxacin = 0.40 ± 0.19 h-1; kmax,gatifloxacin = 0.48 ± 0.15 h-1) or constant concentrations (kmax,levofloxacin = 0.34 ± 0.06 h-1; kmax,gatifloxacin = 0.39 ± 0.23 h-1) were simulated. None of the PK/PD indices was capable of predicting the infection outcome for all the situations investigated. The PK/PD model developed allowed not only the comparison between the fluoroquinolones effect but also the comparison of different dosing regimes for the same drug and can be used for simulating alternative regimens and optimizing therapy of these drugs to treat community-acquired pneumonia.

Biofilmes são comunidades bacterianas complexas encapsuladas em matrizes poliméricas autoproduzidas e podem se desenvolver em superfícies inertes ou tecidos vivos. A formação do biofilme é um importante fator de virulência, pois permite à bactéria resistir às respostas do hospedeiro e à terapia antimicrobiana. Devido a essa elevada resistência aos antimicrobianos, é difícil estabelecer uma estratégia eficaz para o tratamento de infecções com formação de biofilmes, levando a falhas na erradicação das mesmas. Nesse contexto, o objetivo do presente estudo é desenvolver um modelo farmacocinético/farmacodinâmico (PK/PD) para descrever o efeito do ciprofloxacino (CIP) na presença de biofilmes de Pseudomonas aeruginosa (ATCC 27853), visto que a modelagem PK/PD de antimicrobianos é uma ferramenta útil na escolha de regimes posológicos que atinjam o efeito bactericida máximo, minimizando o desenvolvimento de resistência. Para atingir esse objetivo, inicialmente um método analítico por CLAE/fluorescência foi desenvolvido para quantificar o CIP em amostras de plasma e microdialisado. O método desenvolvido foi simples, rápido e com sensibilidade adequada para corretamente caracterizar a farmacocinética plasmática e pulmonar do CIP. Posteriormente, um modelo animal de infecção pulmonar crônica foi adaptado da literatura e padronizado, permitindo a investigação da distribuição pulmonar do CIP em ratos Wistar sadios e infectados. Para tal, bactérias foram imobilizadas em beads de alginato a fim de manter a infecção por até 14 dias com cargas bacterianas superiores à 108 UFC/pulmão. Estudo de microdiálise foi então conduzido para avaliar as concentrações livres de CIP após administração intravenosa de 20 mg/kg. A análise não-compartimental (NCA) e a modelagem farmacocinética populacional (PopPK) dos dados foram realizadas nos softwares Phoenix® e NONMEM®, respectivamente. Diferenças significativas foram observadas no clearance plasmático (1,59 ± 0,41 L/h/kg e 0,89 ± 0,44 L/h/kg) e na constante de eliminação (0,23 ± 0,04 h-1 e 0,14 ± 0,08 h-1) para ratos sadios e infectados, resultando em uma exposição plasmática maior nos animais infectados (ASC0-∞ = 27,3 ± 12,1 μg·h/mL) quando comparados com os animais sadios (ASC0-∞ = 13,3 ± 3,5 μg·h/mL) ( = 0,05). Apesar da maior exposição plasmática, quando comparados com os animais saudáveis (fT = 1,69), animais infectados apresentaram uma penetração pulmonar quatro vezes menor (fT = 0,44). Diferenças na constante de eliminação pulmonar não foram observadas. Dados plasmáticos e pulmonares foram simultaneamente descritos por modelo PopPK constituído de compartimentos venoso e arterial, dois compartimentos representativos de duas regiões pulmonares distintas e dois compartimentos periféricos, representando outros tecidos que não os pulmões. Um clearance pulmonar foi adicionado ao modelo apenas para os dados de microdiálise dos animais infectados (CLlung = 0,643 L/h/kg) afim de explicar a exposição tecidual diminuída. O modelo desenvolvido descreveu, com sucesso, os dados plasmáticos e teciduais de animais sadios e infectados, permitindo a correta caracterização das alterações observadas na disposição plasmática e pulmonar do CIP decorrentes da infecção com biofilme. Para os estudos de farmacodinâmica, o efeito bactericida do CIP frente a biofilmes e células planctônicas de P. aeruginosa foi simultaneamente avaliado através do uso de curvas de morte bacteriana. Para a construção destas curvas, biofilmes de P. aeruginosa foram formados na superfície de blocos de acrílico e sua formação foi confirmada pelo ensaio cristal violeta e por microscopia eletrônica de varredura. Os blocos foram expostos a concentrações constantes de CIP (de 0,0625 a 10 μg/mL) e, em tempos pré-determinados, células planctônicas e de biofilmes eram amostradas para quantificação. Um modelo semi-mecanístico que incorpora um modelo Emax sigmoidal foi utilizado para descrever o efeito do CIP frente a ambos estilos de vida bacteriano. Uma subpopulação pré-existente com menor suscetibilidade ao CIP foi incluída no modelo e o efeito do CIP nesta subpopulação também foi descrito pelo modelo Emax sigmoidal. A comparação dos parâmetros estimados pelo modelo demonstrou que o efeito in vitro do CIP é maior para as células planctônicas (EC50 = 0,259 mg/L e 0,123 mg/L e Emax = 2,25 h-1 e 5,59 h-1 para biofilmes e planctônicas, respectivamente). A potência estimada do CIP para a subpopulação resistente foi muito menor para ambos estilos de vida bacteriano (EC50 = 2,71 mg/L e 1,15 mg/L para biofilmes e planctônicas, respectivamente). Os modelos desenvolvidos podem ser utilizados para a simulação de cenários não testados e servir como uma ferramenta para guiar a escolha dos regimes posológicos adequados, contribuindo para o sucesso terapêutico no tratamento de infecções associadas à biofilmes. / Biofilms are complex bacterial communities enclosed in self-produced polymeric matrices that can develop in inert surfaces or living tissues. Biofilm formation is an important virulence factor that allows bacteria to resist host responses and antibacterial agents. Due to this high resistance to antibiotics, it is difficult to establish an efficacious strategy for treatment of infections with biofilm formation leading to failure in infection eradication. In this context, the goal of this study was to develop a pharmacokinetic/pharmacodynamic (PK/PD) model to describe the antimicrobial effect of ciprofloxacin (CIP) in the presence of biofilms of Pseudomonas aeruginosa (ATCC 27853), since PK/PD modeling for antibacterial agents can be a useful tool to choose dosing regimens and to achieve the maximum bactericidal effect, minimizing the development of resistance. To reach this goal, firstly an analytical method based on HPLC/fluorescence was developed in order to quantify CIP in plasma and lung microdialysate. The developed method was simple, fast and with enough sensibility to proper characterize CIP plasma and lung pharmacokinetics. Secondly, an animal model of chronic lung infection was adapted from literature and standardized, allowing the analysis of CIP lung distribution in infected and healthy Wistar rats. Bacteria were immobilized in alginate beads prior to inoculation to Wistar rats in order to sustain the pneumonia for 14 days, maintaining a bacterial load superior to 108 CFU/lung. A microdialysis study was then conducted to evaluate free CIP concentrations after an intravenous administration of 20 mg/kg. Non-compartimental analysis (NCA) and populational PK modeling (PopPK) of the data were performed in Phoenix® and NONMEM®, respectively. Statistical differences were observed in the plasma clearance (1.59 ± 0.41 L/h/kg and 0.89 ± 0.44 L/h/kg) and elimination rate constant (0.23 ± 0.04 h-1and 0.14 ± 0.08 h-1) for healthy and infected rats, respectively, resulting in a significantly higher CIP plasma exposure in infected rats (AUC0-∞ = 27.3 ± 12.1 μg·h/mL) compare to healthy animals (AUC0-∞ = 13.3 ± 3.5 μg·h/mL) ( = 0.05). Besides the plasma exposure, a four times lower pulmonary penetration was observed in infected rat’s lungs (fT = 0.44) in comparison to healthy animals (fT = 1.69), with no significant differences in the lung elimination rate constant. Plasma and lung data were simultaneously fitted using a PopPK model consisting of an arterial and a venous compartment, two compartments representing different regions of the lungs and two peripheral distribution compartments, representing tissues other than lungs. A lung clearance was added to the model for infected animals (CLlung = 0.643 L/h/kg) to explain the lower tissue exposure. The model successfully described the plasma and microdialysis data from both, healthy and infected rats and allowed to correctly describe the changes in CIP plasma and lung disposition in biofilm infections. For the pharmacodynamic studies, CIP bactericidal effect against Pseudomonas aeruginosa biofilms and planktonic shedding cells were simultaneously evaluated using the time-kill curves approach. For the time-kill curves construction, P. aeruginosa biofilms were formed in acrylic blocks, which was confirmed by the crystal violet assay and scanning electron microscopy. The blocks were placed in flasks containing Mueller-Hinton growth medium and exposed to constant CIP concentrations (ranging from 0.0625 to 10 μg/mL). At pre-determined time points, biofilm and planktonic cells were sampled for bacterial counting. A mechanism-based model which incorporates a sigmoidal Emax model was used to describe the CIP effect against P.aeruginosa in both llifestyles, biofilm and planktonic. The presence of a pre-existing resistant subpopulation was included in the model and also modeled with a sigmoidal Emax model to describe CIP effect in this subpopulation. Comparison of the parameter estimates showed that the in vitro effect of CIP is higher for planktonic cells (EC50 = 0.259 mg/L and 0.123 mg/L and Emax = 2.25 h-1 and 5.59 h-1 for biofilm and planktonic cells, respectively). CIP potency was much lower for the resistant subpopulation, for both bacteria lifestyles (EC50 = 2.71 mg/L and 1.15 mg/L for biofilm and planktonic, respectively). The developed models can be used to simulate untested scenarios and serve as a tool to guide dosing regimen selection, contributing for the therapeutic success of treatments of biofilm-associated infections.

Objetivo: O objetivo geral deste trabalho foi desenvolver um modelo farmacocinético/farmacodinâmico (PK/PD) para descrever o efeito bactericida in vitro das fluoroquinolonas levofloxacino (LEV) e moxifloxacino (MXF)contra Escherichia coli, baseando-se em dados in vivo de concentração livre prostática. Métodos: Ratos Wistar machos foram utilizados nos experimentos in vivo para determinação da farmacocinética plasmática e prostática do LEV (7 mg/kg) e MXF (6 e 12 mg/kg) após dose i.v. bolus. As concentrações livres prostáticas foram determinadas por microdiálise. A coleta das amostras de plasma e dialisado de tecido foi realizada simultaneamente nos animais previamente anestesiados com uretano para determinação do fator de distribuição tecidual (fT). Para a quantificação do LEV e MXF nas amostras de plasma e dialisado, métodos analíticos foram validados. Análise farmacocinética não-compartimental e modelagem compartimental dos dados foram realizadas utilizando o WinNonlin® e NONMEM® v. 6, respectivamente. Os experimentos de farmacodinâmica in vitro foram executados utilizando sistema composto de caldo de cultura Mueller-Hinton no qual a bactéria teste (Escherichia coli ATCC 25922) foi exposta a concentrações constantes e flutuantes dos antimicrobianos. O número de colônias bacterianas viáveis (CFU/mL) foi determinado em função do tempo e utilizado como parâmetro farmacodinâmico para construção das curvas de morte bacteriana (time-kill curves). Nos experimentos de time-kill curves estáticos, concentrações baseadas em múltiplos da MIC na faixa de 0.008–2 mg/L foram utilizadas, enquanto que no dinâmico a meia-vida de eliminação do LEV em humanos foi simulada no sistema in vitro através de diluição constante do caldo de cultura. Resultados e Discussão: Um método analítico por HPLC-fluorescência foi desenvolvido e validado para a quantificação do MXF nas amostras biológicas. Método analítico também foi validado para quantificação do LEV nas amostras. Os perfis plasmáticos e teciduais das duas fluoroquinolonas foram modelados simultaneamente utilizando modelo de três compartimentos considerando transporte linear (difusão passiva) e saturável (cinética de Michaelis-Menten). O modelo, que foi o mais adequado para descrever os dados experimentais, sugere a presença de transportadores de efluxo na próstata. A penetração prostática média do MXF foi significativamente maior que a do LEV (fT = 1.24 vs. 0.78) e foi independente da dose. Em ratos, não foi observada diferença na meia-vida plasmática média entre LEV (5.0 h) e MXF (4.9 h), embora a meia-vida tecidual foi ligeiramente maior para o MXF (3.3 vs. 2.3 h). Usando a abordagem populacional de modelagem PK/PD, modelo de Emax sigmoidal foi utilizado para descrever o efeito das duas quinolonas frente a E. coli tanto nos experimentos de concentração estática quanto dinâmica. A comparação dos parâmetros PK/PD estimados mostrou que o MXF apresenta potência superior ao LEV contra a cepa através da comparação dos valores de EC50, embora ambos tenham apresentado eficácia comparável (Emax de 1.85 e 1.83 h-1 para MXF e LEV, respectivamente). Para o LEV, os esquemas posológicos de 500 mg q12 h e 1000 mg q24 h apresentaram maior eficácia no período de 24 h, pois promoveram a inibição completa do recrescimento bacteriano observado nos outros dois regimes de dose testados. Conclusões: A correlação dos dados de farmacocinéticain vivo com os experimentos de farmacodinâmica in vitro, seguida da construção do modelo PK/PD de efeito máximo, possibilitou explorar a relação do efeito antimicrobiano em função do tempo baseada em concentrações livres esperadas na prostatite. / Objective: The aim of this study was to develop a pharmacokinetic/pharmacodynamic (PK/PD) model to describe the in vitro bactericidal effect of the fluoroquinolones levofloxacin (LEV) and moxifloxacin (MXF) against Escherichia coli based on free concentrations in prostate tissue measured in vivo. Methods: Pharmacokinetic experiments were conducted in male Wistar rats for the determination of plasma and free prostate concentrations of LEV (7 mg/kg) and MXF (6 and 12 mg/kg) after i.v. bolus administration. Blood and tissue dialysate samples were collected simultaneously in the group of rats previously anesthetized with urethane to determine the tissue distribution factor (fT). To quantify MXF and LEV in plasma and dialysate samples obtained after administration of the quinolones, analytical methods based on HPLC-fluorescence were developed and validated accordingly. Non-compartmental analysis and compartmental PK modeling of the data was performed in WinNonlin® and NONMEM® v. 6, respectively. The in vitro pharmacodynamic experiments were executed by using a system composed of Mueller-Hinton growth medium in which the test bacterial strain (Escherichia coli ATCC 25922) was exposed to constant and fluctuating antimicrobial concentrations. The number of viable colony-forming units (CFU/mL) was determined as a function of time and used as the pharmacodynamic parameter for construction of bacterial time-kill curves. In the static time-kill curves, concentrations in the range of 0.008-2 mg/L were tested based on multiples of the MIC, whereas in the dynamic time-kill curves the half-life of LEV in humans was simulated in the in vitro system by stepwise dilution of the growth medium. Results and Discussion: An HPLC-fluorescence method was developed and fully validated to quantify MXF in biological fluids. A method was also validated to determine LEV in the samples. Plasma and prostate concentrations of both drugs were simultaneously fitted using a three-compartment model considering linear (passive diffusion) and saturable transport (Michaelis-Menten kinetics), suggesting the presence of efflux transporters in the prostate. The average tissue penetration of MXF in the prostate was significantly higher than that of LEV (fT = 1.24 vs. 0.78) and was independent of the dose. In rats, differences in average plasma half-life between plasma LEV (5.0 h) and MXF (4.9 h) were not observed, even though the tissue half-life was slightly longer for MXF (3.3 vs. 2.3 h). Using a population PK/PD modeling approach, a sigmoidal Emax model was used to describe the effect of the two quinolones against E. coli both in the static as well as in the dynamic time-kill curves. Comparison of the PK/PD parameter estimates showed that the in vitro potency of MXF is higher than LEV against the strain tested as shown by EC50 values, but both presented equivalent efficacy (Emax of 1.85 and 1.83 h-1 for MXF and LEV, respectively). For LEV, the dosing regimens of 500 mg q12 h and 1,000 mg q24 h showed overall greater efficacy over the 24 h period as they resulted in complete inhibition of bacterial regrowth observed in the other two dosing regimens tested. Conclusions: The correlation of in vivo pharmacokinetic data with in vitro pharmacodynamic experiments, followed by the development of an Emax PK/PD model, allowed determining the relationship between the bactericidal effect as a function of time based on free tissue concentrations expected in the site of infection.