<%BANNER%>

Semi-Physiological Population Pk/Pd Model of Adc Neutropenia

Permanent Link: http://ufdc.ufl.edu/UFE0045141/00001

Material Information

Title: Semi-Physiological Population Pk/Pd Model of Adc Neutropenia
Physical Description: 1 online resource (60 p.)
Language: english
Creator: Tatipalli, Manasa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: friberg -- neutropenia -- pkpd
Pharmaceutics -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, M.S.P.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Cancer is the second leading cause of mortalityin the United States, resulting in >500,000 American deaths annually. AntibodyDrug Conjugates (ADCs) are a class of therapeutics where the monoclonalantibodies bearing cytotoxic drugs covalently bound via a chemical linker.Neutropenia is one of the most common dose limiting hematologic toxicityassociated with ADCs often resulting inthe dose reductions and delays. TK data from 10 different ADC molecules incynomolgus monkeys is used to model the neutropenia. The 10 different ADCs usedin the analysis had similar structure and chemical characteristics (similardrug i.e. MMAE, linker i.e. vc and similar antibody IgG )  but differ only in terms of the target. Consideringthe similar structure, chemical characteristics and similar DARs, it is assumedthat all the ADCs have similar potency. But it was noticed that the PK as wellas the neutropenia was different across different ADCs. Hence the objectiveswere 1) to determine if the differences in the total antibody PK can explainthe neutropenia across different ADCs and 2) Use model to predict incidence ofneutropenia in future studies which can aid in clinical development foroptimizing the doses. Friberg’s neutropenia model was used for this analysis
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Manasa Tatipalli.
Thesis: Thesis (M.S.P.)--University of Florida, 2012.
Local: Adviser: Derendorf, Hartmut C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045141:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045141/00001

Material Information

Title: Semi-Physiological Population Pk/Pd Model of Adc Neutropenia
Physical Description: 1 online resource (60 p.)
Language: english
Creator: Tatipalli, Manasa
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: friberg -- neutropenia -- pkpd
Pharmaceutics -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, M.S.P.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Cancer is the second leading cause of mortalityin the United States, resulting in >500,000 American deaths annually. AntibodyDrug Conjugates (ADCs) are a class of therapeutics where the monoclonalantibodies bearing cytotoxic drugs covalently bound via a chemical linker.Neutropenia is one of the most common dose limiting hematologic toxicityassociated with ADCs often resulting inthe dose reductions and delays. TK data from 10 different ADC molecules incynomolgus monkeys is used to model the neutropenia. The 10 different ADCs usedin the analysis had similar structure and chemical characteristics (similardrug i.e. MMAE, linker i.e. vc and similar antibody IgG )  but differ only in terms of the target. Consideringthe similar structure, chemical characteristics and similar DARs, it is assumedthat all the ADCs have similar potency. But it was noticed that the PK as wellas the neutropenia was different across different ADCs. Hence the objectiveswere 1) to determine if the differences in the total antibody PK can explainthe neutropenia across different ADCs and 2) Use model to predict incidence ofneutropenia in future studies which can aid in clinical development foroptimizing the doses. Friberg’s neutropenia model was used for this analysis
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Manasa Tatipalli.
Thesis: Thesis (M.S.P.)--University of Florida, 2012.
Local: Adviser: Derendorf, Hartmut C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0045141:00001


This item has the following downloads:


Full Text

PAGE 1

1 SEMI PHYSIOLOGICAL POPULATION PK/PD MODEL OF ADC NEUTROPENIA By MANASA TATIPALLI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIR EMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN PHARMACY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Manasa Tatipalli

PAGE 3

3 To God, my parents and my s ister

PAGE 4

4 ACKNOWLEDGMENTS My journey through the graduate school has been a fabulous one. Through this experience I got to meet various people who encouraged me to complete my tasks by believing in my efforts. During a point of time I believed that I would not acquire my dream, but the continuous support and guidance of my teachers and peers lead me to achieve my success. First I would like to acknowledge my a dvisor and professor Dr. Hartmu t Derendorf. Without his advice and support this work could not have been accomplished. He ha s supported me in each and every step of my work by encouraging me to fulfill my task with zeal and enthusiasm. I am grateful to him for all his support. I would like to express my gratitude to the member of my committee, Dr. Anthony Palmieri, for his endl ess patience and time. I am grateful to him for all his support Pharmaceutics department office staff. I would also like to thank my manager at Genentech, Jay Tibbitts for his e ndless patience, support and opportunity to work on the project. I would like to thank all my colleagues for their endless support and assistance in the journey of my research experience. I would also like to thank the Department of Pharmaceutics for their financial support without which my journey would not have been accomplished. Last but not the least I would like to thank my family and friends for their endless support and encouragement through all my academics.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURE S ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Antibody Drug Conjugate ................................ ................................ ........................ 12 Drug to Antibody Ratio ................................ ................................ ............................ 13 Monoclonal Antibodies ................................ ................................ ............................ 13 Linkers ................................ ................................ ................................ .................... 14 Nonclevable L inkers: ................................ ................................ ........................ 15 Cytotoxic compounds used in ADC ................................ ................................ ........ 16 Mechanism of Action ................................ ................................ ............................... 16 Assays ................................ ................................ ................................ .................... 18 Pharmacokinetics of ADC ................................ ................................ ....................... 21 Absorption ................................ ................................ ................................ ........ 22 Distribution ................................ ................................ ................................ ....... 22 Metabolism/Catabolism and Elimination ................................ ........................... 23 Why Cynos ................................ ................................ ................................ ....... 23 ADC and T oxicities ................................ ................................ ........................... 24 Neutropenia: Dose Limiting Toxicity ................................ ................................ 25 POPPK Model: Explains the Problem ................................ ................................ ..... 26 Friberg: Established Model for Neutropenia ................................ ............................ 26 2 HYPOTHESIS ................................ ................................ ................................ ......... 35 Introduction to the P roblem ................................ ................................ ..................... 35 Hypothesis ................................ ................................ ................................ .............. 35 Objectives ................................ ................................ ................................ ............... 35 Methods ................................ ................................ ................................ .................. 36 Patients (cynos) and Measurements: ................................ ............................... 36 Treatments ................................ ................................ ................................ ....... 36 Blood counts ................................ ................................ ................................ ..... 36 Pharmacokinetic Pharmacodynamic A nalyses ................................ ....................... 37 Pharmacokinetic A nalysis ................................ ................................ ....................... 37 odel: ................................ ................................ ................................ 38 Data A naly sis ................................ ................................ ................................ .......... 39

PAGE 6

6 3 RESULTS ................................ ................................ ................................ ............... 44 4 DISCUSSIONS ................................ ................................ ................................ ....... 49 Procedural ................................ ................................ ................................ ........ 51 Biology P erspective ................................ ................................ .......................... 52 REFERENCES ................................ ................................ ................................ .............. 58 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 60

PAGE 7

7 LIST OF TABLES Table P age 3 1 Parameters of pharmacokinetic model describing intravenous data Mean values and corresponding i nterindividual variability (CV%). ............................... 47 3 2 Typical population parameter estimates (relative SE %) for neutrophils with a linear Concentration Effect Model. ................................ ................................ ..... 48 4 1 Clearance of Anti ETBR in cynomolgus monkey. ................................ ............... 54 4 2 Various PD models explored. ................................ ................................ ............. 55

PAGE 8

8 LIST OF FIGURES Figure P age 1 1 Selected immunoconjugates in clinical development for cancer therapy. ........... 27 1 2 Schematic of ADC structure. ADCs are monoclonal antibodies bearing cytotoxic drugs covalently bound via a chemical linker. ................................ ...... 27 1 3 Schematic of ADC heterogeneity. ADCs are heterogenous mixture of different DAR species, with individual molecules exhibiting a range of DARs. ... 28 1 4 Therapeutic antibodies approved by the FDA for cancer treatment. ................... 28 1 5 MAb doxorubicin hydrazone linker. ................................ ................................ .... 29 1 6 MAb DM1 disulfide linker. ................................ ................................ ................... 29 1 7 MAb MMAE peptide linker. ................................ ................................ ................. 29 1 8 MAb MMAF thioether linker. ................................ ................................ ............... 29 1 9 Cy totoxic drugs used in Antibody D rug conjugates. ................................ ........... 30 1 10 Compared stability of Enzyme Labile and chemically labile linkers in ADCs w ith MMAE conjugated to cBR96mAb ................................ ................................ 30 1 11 Internalisat ion of antibody drug conjugates ................................ ........................ 31 1 12 Ant ibody drug conjugate processing ................................ ................................ ... 31 1 13 Typic al ADC formats for ADC analytes ................................ ............................... 32 1 14 Conjugated antibody assay ................................ ................................ ................ 32 1 15 Diagram of theoretical ADC catabolism ................................ .............................. 33 1 16 Grades of neutropenia. ................................ ................................ ....................... 33 1 17 Treatment related deaths, by chemotherapy cycle, in patients with aggressive no n hodgkins leukemia who were treated with cyclophosphamide, doxorubicin vincristin and prednisone ............................... 34 2 1 Antibodies armed with Auristatins. ................................ ................................ ...... 41 2 2 Neutrophil observations by dose and by molecule across different ADCs. ......... 41 2 3 Median baseline of the neutrophils acr oss different ADCs with distribution of bas eline of each cynomolgus monkey ................................ ................................ 42

PAGE 9

9 2 4 Two compartment pharmacokinetic model. ................................ ........................ 42 2 5 ................................ ................................ ............... 43 3 1 PK diagnostic plots ................................ ................................ ............................. 45 3 2 PK visual predictive checks ................................ ................................ ................ 46 3 3 PD diagnostic plots by molecule ................................ ................................ ......... 47 4 1 Clearance distribution across Molecules. ................................ ........................... 54 4 2 Clearance values of a ntibodies in cynomolgus monkeys ................................ ... 54 4 3 Slope estimates of 10 ADC molecules. ................................ .............................. 56 4 4 Correlation between CL of ADC Concentration vs Slope estimated. In conclusion there is a 10 fold difference in the slopes across the 10 molecules. ................................ ................................ ................................ .......... 56 4 5 Correlation between cmax of ADC Concentration vs Slope estimated. .............. 57 4 6 Correlation between AUCSS of ADC Concentration vs Slope estimated. .......... 57

PAGE 10

10 LIST OF ABBREVIATIONS ADCs Antibody drug conjugates CYNOS Cynomolgus monkeys DAR Drug to Antibody Ratio IIV InterIndividual Variability MMAE Monomethyl Auristatin E MMAF Monomethyl Auristatin F PD Pharmacodynamics PK Pharmacokinetics PKPD Pharmacokinetic phar madynamic s PO PPK Population pharmacokinetics TK Toxicokinetic Vc Valine citrulline

PAGE 11

11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Require ments for the Degree of Master of Science in Pharmacy SEMI PHYSIOLOGICAL POPULATION PK/PD MODEL OF ADC NEUTROPENIA By Manasa Tatipalli December 2 012 Chair: Hartmut Derendorf Major: Pharmaceutical Sciences Cancer is the second leading cause of mortality in the United States, resulting in >500,000 American deaths annually. Antibody Drug Conjugates (ADCs) are a class of therapeutics where the monoclonal antibodies bearing cytotoxic drugs covalently bound via a chemical linker. Neutropenia is one of the most common dose limiting hematologic toxicity associated with ADCs often resulting in the dose reductions and delays. TK data from 10 different ADC molecules in cynomolgus monkeys is used to model the neutropeni a. The 10 different ADCs used in the analysis had similar structure and chemical characteristics (similar drug i.e. MMAE, linker i.e. vc and similar antibody IgG) but differ only in terms of the target. Considering the similar structure, chemical character istics and similar DARs, it is assumed that all the ADCs have similar potency. But it was noticed that the PK as well as the neutropenia was different across different ADCs. Hence the objectives were 1) to determine if the differences in the total antibody PK can explain the neutropenia across different ADCs and 2) Use a model to predict the incidence of neutropenia in future studies which can aid in clinical development for

PAGE 12

12 CHA PTER 1 INTRODUCTION Cancer is the second leading cause of mortality in the united states, resulting in >500,000 American deaths annually 1 .Treatment of cancer is a double edged sword: it should be as aggressive as possible to completely destroy the tumor but it is precisely this aggressiveness which often causes severe side effects this is one reason why some promising therapeutics cannot be applied systematically. An elegant way to accumulate therapeutic agents at the tumor site is their conjugation /fu sion to tumor specific antibodies. The selective targeting of the cytotoxic drugs to tumors by conjugating the drug to tumor specific antibodies might, provide a solution. Antibody based therapeutics are of growing significance for cancer therapy as eviden ced by 12 such drugs approved for oncologic indications since 1995, including 9 in the United States 2 Antibody Drug Conjugate Antibody Drug Conjugates (ADCs) are a class of therapeutics where the monoclonal antibodies bearing cytotoxic drugs covalently b ind via a chemical linker. ADCs combine the target specificity of the antibody with the potency of a chemotherapeutic 4 agent to increase the therapeutic index of the anticancer drugs and also reducing the systemic exposure and related toxicity of the chemo therapeutic drug while maximizing the delivery of the drug to the target. When compared to the conventional anticancer drugs the toxins used in the ADC can be 100 to 1000 times more cytotoxic. The increase in the number of antibody targets, along with the advanceme nt in the antibody engineering and the chemistry of conjugation led to the further interest in the development of ADCs.

PAGE 13

13 Drug to Antibody R atio ADCs consist of a mixture of different species (antibody with different DARs). The heterogeneity is due to many reasons primarily the manufacturing of ADCs and their deconjugation. The drug is conjugated to the antibody to form an ADC which is a controlled chemical reaction involving specific amino acids on the antibody. Most often used sites on the antibod y are the lysine side chain residues or through cysteine sulfhydryl groups that are activated by reduction of interchain disulfide bonds. Depending upon the type of linker used and the site of conjugation the deconjugation occurs which may be an enzymatic process or a chemical process which results in the loss of the drug. Drug conjugation in the presence of four sets of interchain disulfide bonds gives rise to a heterogeneous ADC mixture that is described in terms of a drug to antibody ratio (DAR) distribu tion and an average DAR. For example let us consider an ADC with an average DAR of 4 ( illustrated in Figure 1 2) is just one possible molecular species of a mixture that may be composed of 0 to 8 drugs per antibody covalently attached via the vc linker. This heterogeneity causes different PK, efficacy, and toxicity properties of each fraction, in some instances it has been reported that fractions with higer DAR are cleared more rapidly and contributed to more severe toxicity while other reports have demon strated, similar efficacy, tolerability and PK between preparations having heterogeneous (0 8) and homogeneous ( 4) DAR. 5 The heterogeneity also causes difficulties in ADC quantitation and optimization. Monoclonal Antibodies The therapeutic antibodies approved by FDA for cancer treatment are illustrated in figure 1 1

PAGE 14

14 Linkers Linkers and linker technology are very important as they have an impact on the safety, potency and specificity of ADC. Linkers are the small and central part of the ADC that conne cts the cytotoxic drug to the mAb. The main function of the linkers is maintaining the stability of ADC in the circulation by keeping it intact, to provide favorable pharmacokinetics in circulation and for effectively releasing the drug at the target .The cytotoxic drug is generally attached to the Fc portion or the constant region of the mAb to prevent any interference in the detection and binding to the antigen. 6 Linkers can be classified into many types. One of the Classificati ons is : 1. Cle a vable linkers A ) Chemically labile linkers B ) Enzyme labile linkers 2. Nonclevable linkers Cle a vable linkers : A. Chemically labile linkers: The cleavage of the linker selectively and release of the drug at the target site depends on the differences in the properties betwee n the plasma and the cytoplasmic compartments such as differences in pH between plasma and cytoplasmic compartments. a. Acid clevable Hydrazone linkers : Linkers were chosen 7.5) but cleaves and releases the drug by the hydrolysis of the hydrazone once they enter the slightly acidic endosomes (pH 5.0 6.5) and lysosomes (pH 4.5 5.0) Eg Doxorubicin was attached to the cysteine residues of the Lewis Y mAb through a hydrazone linker (6 maleimidocaproyl) However the pH based linkers are not stable as there is a lot of drug loss in the bloodstream as reported from the pharmacokinetic data and lower mean half life of few of the ADC.

PAGE 15

15 b. Disulfide linkers: Other class of chemically labile linkers is the disulfide linkers .The release of the drug depends on the concentrations of glutathione. Generally the concentrations of glutathione in the cytoplasm are in mill i molar that is approximately 1000 fold higher when compar ed with the blood stream which is in micro molar. This ratio is even higher in the tumor cell since the tumor has higher concentrations of the reductive enzymes when compared with the normal cell due to decreased supply of oxygen to the tumor cell. In the absence of these reductive enzymes the disulfide bond is stable. Eg.Taxoid ADC payloads are attached to the mAb via a disulfide bearing 4 mercaptopentanoate linker against epidermal growth factor receptor B. Enzyme labile linkers : Chemical labile linkers are less stable which allows alternate approaches like connecting the drug to the mAb via a peptide linkage. The drug may be released from the intact ADC by the action of proteases in the lysosomes such as cathepsin or plasmin which are present at elevated le vels in the tumor tissues. Apart from that the peptide linkage has good stability in the plasma as the proteases are not present outside the cell due to the pH conditions. some of the enzyme labile linkers are a ) Valine citrulline b ) Phenylalanine lysine Comparing the stability of the enzyme labile linkers such as Valine citrulline and Phenylalanine lysine linkers with the chemically labile linker hydrazone linker the Valine citrulline linker was 100 times more stable than the hydrazone linker which resu lts in increased stability and specificity. Nonclevable linkers: The Noncle a vable linkers are the most stable linkers that are being used in ADC. Thioether linker is one example of nonclevable linkers. The mechanism involves internalization of the ADC fol lowed by the degradation of the mAb component in the lysosome, resulting in the release of the drug along with the linker. This altered drug f orm is charged as a result will not

PAGE 16

16 noncle a vable linkers are w ell tolerated resulting in improved therapeutic index. The invivo stability of the ADC with noncle a vable linkers is greater in terms of the half life when compared to the other linkers. The released drug is attached to an amino acid i.e either lysine for maytansinoid conjugates or cysteine for auristatin conjugates. This modified form of the drug is equally potent when compared with the drug itself. 7 Eg.Transtuzumab DM1 (TDM1) for HER2 positive metastatic breast cancer. Cytotoxic compounds used in ADC Different cytotoxic drugs conjugated to the antibody in ADC have very diverse intracellular targets allowing killing the cancer cells specifically. Inspite of the diversity of their targets the mechanism of killing of the cells is arresting the cell cycle in either the S or G2/M phase depending on the type of the drug used. The cells in the G0 phase or the nondividing phase are less affected when compared to the cells in the dividing phase such as cancer cells. Most of the anticancer drugs target the tubuli n, microtubules, or DNA and the cells in the nondividing phase are resistant to these kinds the normal cells. Some of the cytotoxic compounds that have been used in antibody d rug conjugates are listed in figure Mechanism of Action The primary mechanism of action for the cytotoxicity of ADC is internalization of the ADC. It can be summarized as follows. 1) Internaliz ation of ADC 2) Intracellular trafficking of the internal ized conjugates 3) Activation of ADCs into cytotoxic compounds 4) Bystander effect of ADC

PAGE 17

17 Internaliz ation of ADC: The antibody component of the ADC is a very important co mponent that contributes to the selective cytotoxicity .The ADC binds to the target antig en present on the cell surface which results in the internalization of the ADC antigen complex. When comparing the internalization rates of the ADC antigen complex with the antibody alone it was found that the ADC antigen complex is equally internalized w hen compared to the antibody itself or som etimes even more efficiently the anti CD20 antibody rituximab is an example of these. As the antibodies will not penetrate the cellular membranes they get into the mammalian cells through 3 different internalizat ion routes as follows: a) Clathrin mediated endocytosis b) Caveolae mediated endocytosis c) Pinocytosis The first two mechanisms are antigen mediated while the pinocytosis is antigen independent mechanism. a) Clathrin mediated endocytosis : o ne of the major routes of i ntracellular uptake side of the plasma membrane called the clathrin coated pits. These pits lead to the formation of clathrin coated vesicles which follow the endosomal lyso somal pathway. The antigens may be concentrated on the clathrin coated pits or eventually migrate towards the pits after binding with the ADC.

PAGE 18

18 b) Caveolae mediated endocytosis result in the activation of the ADC as it accumulates in the endoplasmic reticulum or Golgi complex which lack the proteolytic enzymes that can release the drug c) Pinocytosis: It is a non specific and antigen in dependent method of uptake of ADC. It occurs via the uptake of the dissolved conjugate in the surrounding medium by pinocytosis. The pinocytosis of the ADC causes the necrotic lysis of the cell. Intracellular trafficking of the internalized conjugates: T he internalized complex later enters the endosomal lysosomal pathway .The Clathrin coated complex are first uncoated of the Clathrin which later fuse with the early endosomes which later enter the late e ndosomes and then the lysosomes Activation of ADCs i nto cytotoxic compounds: The acidic environment and the endosomes and lysosomes and proteolytic enzymes causes the proteolytic degradation of the ADC resulting in the release of the cytotoxic drug. The released cytototoxic drug is later effluxed into the c ytoplasm which binds to its molecular target resulting in the arrest of the cell cycle followed by apoptosis. Bystander effect of ADCs: The drug in the cytoplasm effluxes from the cell either by passive diffusion, leakage from the dying cells or active transport. If the released drug is permeable it enters into the neighboring cell which is called the bystander cell killing. 8 Assays Different bioanalytical strategies are developed for measurement of analytes quantitatively aiding in understanding the beh avior of the drugs. Since the ADC consists of both the antibody component and the drug component i.e. both the large and the

PAGE 19

19 small molecule component and both the components of the ADC are significant for its activity the analytical strategies are to be de veloped so that they can measure both the components. The different analytes are measured using various assays examples of these assays are listed below. 1) Total Antibody (Conjugated and unconjugated antibody) 2) Conjugated Antibody 3) Conjugated Drug 4) Unconjugate d antibody 5) Unconjugated Drug (free drug) Each analyte provides unique information about the ADC behavior. Total antibody concentrations are determined by using enzyme linked immunosorbent assay (ELISA) .In the total antibody assay the ADC antibody was ca ptured with the aid of an antigen or Extracellular Domain which was later detected using a labeled antibody as shown in the figure 1 13 .It includes measurement of both the conjugated and unconjugated antibody. The Total Antibody PK helps in understanding the effect of conjugation on the antibody which helps in optimizing the ADC payloads. The figure 1 14 shows the plasma concentration time profile for both the unconjugated antibody and t otal antibody from ADC. It was observed that for most of the ADC the clearance of the antibody was increased due to the conjugation. The difference in the curves illustrates the impact of conjugation on the clearance. Higher DAR species tend to have higher clearance when co mpared to the lower DAR species.

PAGE 20

20 Conjugated antib ody (antibody conjugated to at least to one drug) concentrations are measured using an ELISA technique. In conjugated antibody assay the conjugated antibody was captured with the aid of anti cytotoxic drug antibody which was later detected using a labeled antigen for ECD as shown in the figure 1 14 Since this assay measures the antibody along with the drug, it can be used for the measurement of the active ADC concentration. Total antibody has faster clearance when compared to the conjugated antibody as sh own in the figure as the change in the concentrations of the conjugated antibody is driven by two processes i) Elimination of the intact AD C ii) Deconjugation clearance (d econjugation to DAR0).The extent of divergence of the two curves indicates the rate o f drug loss from the ADC. The effect of the conjugation site and linker stability can also be studied. This assay canno t differentiate the different DAR species. Irrespective of the similar PK properties of the different ADC species they might differ in th eir potencies which make it difficult to relate the concentrations and their associated pharmacological activity. The conjugated d rug assay measures the amount of cytotoxic drug that is covalently bound to the antibody. The drug that is conjugated to the antibody is cleaved which is late r quantified. The concentration of the conjugated drug decline rapidly when compared to the Total antibody since to processes drive the decline of the drug concentrations. The initial difference in t he molar concentrations of the conjugated drug and t otal antibody at dosing gives the starting average DAR and at a particular time after dosing both the concentrations i ntersect indicating the DAR is same.

PAGE 21

21 But the assay does no t provide much information concerning the concentration of the antibody with bound drug It canno t differentiate between low concentrations of high DAR and high concentrations of low DAR. The c ytotoxic drug assay measures the drug released from the ADC by either by chemical or enzymatic process. The measurement of free drug helps in understanding both the safety and efficacy issues. The free drug is measured by using LC MS methods or the ELISA methods.LC MS technique is highly specific when compared to the ELISA technique as they measure all the a nalytes with similar structure. The cytotoxic drug concentration assists in understanding the systemic exposure of the drug along with the drug drug interactions however since the measured concentrations are very low the interpretation is difficult. Among all the assays the affinity capture LC MS is the most powerful assay .In this assay the intact ADC is extracted from the plasma and later subjected to the LC MS/MS. This assay directly measures the average concentrations of drug associated with the antib ody. This assay was utilized to understand the effect of conjugation site on the stability of the linker. It was also used to determine the DAR distribution for TDM1. 9 Pharmacokinetics of ADC The antibody is the backbone of the ADC and contributes approxi mately 98% of the total ADC Molecular weight. The ADC possess many properties similar to that of unconjugated antibodies such as slow clearance, long half life ,low volume of distribution and proteolysis mediated catabolism. They also have undesirable prop erties possessed by unconjugated antibodies such as poor oral bioavailability, incomplete absorption, immunogenicity and nonlinear distribution and elimination.

PAGE 22

22 The ADME i.e. the absorption, distribution, metabolism and elimination properties of ADCs are important for interpreting the ADC PK and PK/PD relationship. Absorption Most of the ADCs are administered through an intravenous or an intraperitoneal route (mostly in preclinical). The absorption properties are similar to that of unconjugated antibodie s. Distribution The antibody forms th e backbone of the ADC structure t he distribution of ADC is similar to that of unconjugated antibody. The distribution of ADC is important in understand ing the pharmacological properties and toxic effects. Initially it is distributed in the vascular space having a volume of distribution of approximately 50mL/kg. Further it distributes into the interstitial space with a steady state volume of distribution of 150 200ml/kg approximately. The ADC is transported from the pla sma to the interstitial fluid through convection similar to that of unconjugated antibody. The distribution of ADC depends on the target expression as well as internalization. In a few targets the antigen is shed and released into the systemic circulation The ADC binds to the shed ded antigen in circulation changing the distribution of ADC and the elimination. The complex is then cleared by the live r thereby increasing the chance of liver toxicity. The distribution of antibodies with the MMAE was increase d into the liver as compared to the unconjugated antibodies in rodents. The distribution of the cytotoxic drug (MMAE) in all tissue was similar to that of the unconjugated antibody except the liver which has increased concentration of the drug when compare d to that of antibody due to the clearance property of the liver. The distribution of ADC and the cytotoxic drug

PAGE 23

23 may be different to normal as well as the tumor tissue which later affects the pharmacological activity or toxicity. Metabolism/Catabolism and Elimination The predominant method of elimination for the antibodies is target mediated elimination .It is also eliminated through non specific uptake into the cells which is followed by proteolytic degradation. The catabolism of the ADC by the lysosomal degradation or linker cle a vage results in the products which are highly cytotoxic which includes the free drug or drug containing products unlike the unconjugated antibodies wherein the catabolic properties are amino acids and small peptides. The ADC can be cleaved by deconjugation or catabolism. This deconjugation takes place by enzymatic process or chemical process with the intact antibody. The catabolism process involves the proteolytic degradation resulting in the cytotoxic drug containing products. T he catabolites formed are further metabolized by CYPs with elimination through drug transporters similar to that of small molecule drugs. Why Cynos Predicting the human PK from the preclinical studies is one of the important features of the translational PK studies. Prediction of human PK for ADCs is not much reported indicating narrow clinical experience about new drug linker combinations. At this point prediction of human PK of ADCs, catabolic products with the cytotoxic drug, along with the deconjugati on products from the translational studies can be useful in predicting the dose, dosage regimen clinically and also helps in assessing the safety issues associated with use of ADC clinically. Non human primates, cynomolgus monkeys were used in the preclini cal studies in the development of ADCs. Cynomolgus monkeys were used for various reasons as follows I ) The target antigen have similar

PAGE 24

24 homology between cynos and humans mo stly II ) The cross reactivity to the antibodies is similar across both the species. Both the target and non target mediated antibody clearance can be compared. The above aspects make the cynomolgus monkeys suitable for the human PK prediction i.e. the human c learance of ADCs can be predicted from the cynomolgus monkey PK data and an all ometric scaling factor of 0.85. However if any structural changes takes place in the mAb due to the conjugation of the drug the above approach may not be applicable due to changes in t he ADC clearance, distribution as well as susceptibility to the nonspecific proteolytic degradation. The total antibody and the unconjugated antibody are guided by the mAb component of the ADC than the small molecule. The prediction of the human PK from the translational studies worked well for the antibody based PK parameters but it was very challenging in terms of predicting the drug related products (cytotoxic drug containing products and free drug itself). There were species differences in the formation and disposition of these small molecule drug containing products causing problems in the prediction of PK of these small molecules analytes. E.g. ADCETRIS (brentuximab vedotin) showed increased free MMAE levels in the plasma in patients when compared to the concentrations predicted fr om the cynomolgus monkeys. Further examination in the mechanistic differences across species in terms of ADC catabolism, disposition and PK of the products formed can help in the development of the current ADCs and design the new ADCs. ADC and toxicities T he most common adverse effects associated with ADCs are peripheral sensory neuropathy (47%), fatigue (46%), nausea (42%), upper respiratory tract infection (37%),

PAGE 25

25 and diarrhea (36%).The most common grade 3 or 4 adverse events were neutropenia (20%), periph eral sensory neuropathy (8%), thrombocytopenia (8%), and anemia (6%). The most common drug associated adverse events i.e MMAE related adverse effects were peripheral neuropathy, nausea, fatigue, diarrhea, dizziness, and neutropenia; most were grade 1 or 2 in severity. Neutropenia: Dose Limiting Toxicity Neutropenia is one of the most common adverse effects associated with cytotoxic anticancer drugs and the most serious hematologic toxicity 10 It is a dose limiting toxicity often resulting in dose reductions and delays which may compromise the treatment outcomes. The degree and duration of the neutropenia determines the risk of infection. According to the National cancer institute the scale u sed for grading the neutropenia consists of 4 grades as shown in figure 1 15 11 The absolute neutrophilic count at the nadir is used as a surrogate marker for the selection of a dose in order to maximize the efficacy. It was also shown that during the canc er chemotherapy lack of hematological toxicity was associated with lower anticancer efficacy. 12 Currently there are no reliable in vivo preclinical models to describe the chemotherapy induced myelosuppression. 13 The risk of infection is more for patients w ith prolonged neutropenia than the patients with the same ANC nadir who recover back immediately. The chemotherapy regimen is one of the important determinants of the risk of neutropenia. When considering the effect of treatment cycle on the number of deat hs it was reported that 67% of the deaths occurred due to neutropenia during the first treatment cycle while 37% were reported during cycles >=2. 14 .The CSFs (Colony stimulating

PAGE 26

26 factors) can be used for the treatment of neutropenia but the costs associated are high. 15 POPPK Model: Explai ns the P roblem The ANC at the nadir and the survival fraction of the neutrophils do no t alone explain the time course of neutrophils as a result the PK/PD Models were developed in order to explain and quantify the hematologi cal toxicity i.e. neutropenia after the administration of anticancer drugs 16 The pharmacodynamic models that use the percentage decrease in the ANC or the immediate models canno t be used to describe the time course of neutropenia. The models developed mus t include the production and the destruction of the target cells 17 Hence the models were developed that accounted for the underlying processes responsible for the fate of circulating neutrophils such as proliferation of progenitor cells, maturation, degra dation and a feedback mechanism. The concentrations of drug in the plasma were used for modeling the drug effects. 18 Friberg: Established Model for Neutropenia mechanistic PK/PD Model of neutropenia is considered as the gold standard that describes the time course of Absolute neutrophil counts. It consists of drug specific parameters and system related parameters. 19 Across various anticancer drugs and studies the model showed consistency in terms of system related parameters. It was also us ed to make human predictions from animals, to identify the covariate effects and to develop a tool for neutrophil guided dose adaptation in chemotherapy. 20

PAGE 27

27 Figure 1 1. Selected immunoconjugates in clinical development for cancer therapy. Figure 1 2. Schematic of ADC structure. ADCs are monoclonal antibodies bearing cytotoxic drugs covalently bound via a chemical linker.

PAGE 28

28 Figure 1 3. Schematic of ADC heterogeneity. ADCs are heterogenous mixture of different DAR species, with individual molecules exhibiting a range of DARs. Figure 1 4. Therapeutic antibodies approved by the FDA for cancer treatment.

PAGE 29

29 Figure 1 5. MAb doxorubicin hydrazone linker. Figure 1 6. MAb DM1 disulfide linker. Figure 1 7. MAb MMAE peptide linker. Figure 1 8. MAb MMAF thioether linker.

PAGE 30

30 Figure 1 9. Cytotoxic drugs used in Antibody drug conjugates. Figure 1 10. Compared stability of Enzyme Labile and chemically labile linkers in ADCs with MMAE conjugated to cBR96mAb.

PAGE 31

31 Figure 1 11. Internalisation of antibody drug conjugates. Figure 1 12. Antibody drug conjugate processing.

PAGE 32

32 A B Figure 1 13 Typic al ADC formats for ADC analytes. A) Total antibody assay: capture of ADC antibody using antigen or target extracellular domain (ECD), with detection using labeled antibody to ADC antibody. B) Typical Concentration time profiles of unconjugated antibody and ADC analytes following an intravenous bolus dose. Comparison of plasma concentrat ion profile of Total antibody ( following unconjugated antibody administration) with Total antibody (following ADC administration). Faster decrease in Tab concentrations suggests that pharmacokinetics of the antibody are affected by conjugation A B Figure 1 14 Conjugated antibody assay: A ) capture of ADC using anti cytotoxic drug antibody, with detection using labeled antigen or extr acellular domain. B ) comparison of plasma concentration profiles following ADC administration Tab (blue) has multi exponential profile typical of antibody. Conjugated antibody (grey) shows more rapid decrease in concentrations as a result of antibody e limination and cytotoxic drug deconjugation. C onjugated drug (orange) starts at higher concentrations than Tab reflecting its DAR then decreases more rapidly than Tab due antibody elimination and cytotoxic drug conjugation. Arrows indicate effect of deconj ugation on clearance. Free drug (green) concentrations are much lower, increase with time to reflect delay in deconjugation from ADC,and decline over time

PAGE 33

33 Figure 1 15. Diagram of theoretical ADC catabolism: The formation of cytotoxic drug containing products from ADCs may occur by two concurrent processes deconjugation and catabolism. The deconjugation process results in the release of cytotoxic drug containing products from the ADC, via enzymatic or chemical processes, and unconjugated antibody.The catabolism processes includes proteolytic catabolism of the antibody and formation of cytotoxic drug containing catabolites Figure 1 16 Grades of neutropenia.

PAGE 34

34 Figure 1 17 Treatment related deaths by chemotherapy cycle in patients with aggressive non hodgkins leukemia who were treated with cyclophosphamide, doxorubicin vincristin and prednisone There were 35 deaths in 265 consecutive patients (13%) includ ing 22 deaths (63%) that occurred in the first cycle, and 29 of the 35 deaths (83%) were infection related.

PAGE 35

35 CHAPTER 2 HYPOTHESIS Introduction to the problem The basic structure of the ADC is as follows : The drug used across all the ADCs is the MMAE (Mono Methyl Auristatin E) and the linker used is the vc is valine citrulline. analysis had similar structure and chemical charact eristics .Consi dering the similar structure chemical characteristics and similar Drug to antibody ratios across different ADCs it is assumed that all the ADCs have similar potency. However, it was noted that the PK values as well as t he neutropenia were different across different ADCs. Hypothesis Considering similar chemical structure, potency and target independent effect of the ADCs on neutropenia it is assumed that the neutropenia across different ADCs is across the different ADCs can be Objectives Determine if the differences in the total antibody PK can explain the neutropenia across different ADCs. Use model to describe concentration respons e of neutropenia to: Inform design of future studies of vc MMAE ADCs. Inform investigations into mechanisms of toxicity. Develop model for interspecies translation and clinical development.

PAGE 36

36 Methods Patients (cynos) and Measurements: The Total antibody plasma concentrations of the 10 ADC molecules (including the clinical and surr ogate) and absolute neutrophil count for the cynomolgus monkeys were obtained from the 10 different studies. Total of 222 cynomolgus monkeys were used for the TK studies with 7 P K samples for the first dose and 2 6 samples for the rest of the doses. A total of 3934 PK samples were used for PK analysis. Treatments The different treatments ranged from 1 through 6 mg/kg as 1, 3, 5 and 6 mg/kg. All the treatments were given every th ree weeks .The number of treatment cycles ranged from 1 to 6. Blood counts Neutrophils(2380 observations from 172 cynomolgus monkeys) with different cancer forms who received the different ADCs (i.e. molecules 1 ,2,3,4,5,6,7,81 and 82) from 1 to 6 cycle s(varying between one and 6 cycles per cynomolgus monkey ) were analysed. Absolute neutrophil counts were measured at least twice per week for the first cycle and every other week after first cycle .Some of the molecules had the measurements at baseline, 3 7 days for the first cycle and every other week after that .Some of them had only one measurement per cycle Median baseline values were 1.2 to 18 (.1000/uL) for molecules 1 2, 3,4,5,6,7,81 and 82 respectively. All the ADCs were administered intravenous ly every 3 rd week. Individual concentration time profiles were obtained using empirical bayes estimates from a population pharmacokinetic model.

PAGE 37

37 Pharmacokinetic pharmacodynamic analyses Data were analyzed according to a nonlinear mixed approach by use of NONMEM 7.2 (all bug patches implemented and updated according to the official bug list).NONMEM was running with Compaq Visual Fortran 6.6 compiler (Hewlett Packard Company) on a 2.4 GHz Pentium 4 central processing unit, under Windo ws 7 operating system. First, pharmacokinetic analysis of ADC total antibody plasma concentrations versus time was performed to obtain pharmacokinetic parameters for each patient (post hoc values). Second, the pharmacokinetic pharmacodynamic analysis of AN C (absolute neutrophil count) versus time was performed using these individual pharmacokinetic parameters. 21 Pharmacokinetic analysis Individual specific pharmacokinetic parameter s were obtained by analysis of t otal antibody plasma concentrations versus t ime by use of first order conditional estimation (FOCE) and a structural pharmacokinetic model of 2 compartment model with linear elimination from the central compartment CL, central volume (V1) ,peripheral volume (V2) and intercompartmental clearance(Q).P harmacokinetic data of 10 molecules were analyzed separately to generate specific post hoc individual pharmacokinetic parameters, as well as associated interindividual and intraindividual variabilities are summarized in the table 3 1. A proportional model w as used for the residual variability .Individual pharmacokinetic parameters (post hoc values) were then included in the PD (neutrophil) dataset for pharmacokinetic pharmacodynamic analysis to generate the complete time plasma concentration profiles. Log no rmal distribution of the PK parameters was assumed. Various other models were performed such as three compartmental model and various error

PAGE 38

38 structures but the data was well explained with the two compartment model in accordance with the typical ADC PK prof ile and proportional error model .The PK diagnostics per molecule and the visual predictive checks are presented in the figures 3 1 and 3 2 respectively. : The pharmacody namic model proposed by Friberg et al for other cytotoxic compounds was employed The model mimics the maturation chain of neutrophils and was based on 5 compartments, as follows: 1 compartment that represented progenitor cells (Prol), 3 maturation compartments (Transit), and a compartme nt of circulating observed neutrophils (Circ). The generation of new cells in the Prol was dependent on the number of cells in the compartment i.e. self renewal or mitosis, a proliferation rate constant determining the rate of cell division (K prol ) A maturation chain, with the transit compartments and rate constants Ktr were used to describe the time delay between the administration and the observed effect. A feedback mechanism was modeled as (Circ 0 /Circ) where Circ is the circulating blood cell co unt at a given time and Circ 0 is the baseline circulating cell count before administration The feedback loop describes the rebound of cells( i.e. an overshoot compared with the baseline value (Circ 0 )).As for the reference model, the constant rates were assumed to be equal : K prol K tr K circ for the rate of cell division the rate of transit between compartments, and the rate of physiologic elimination of circulating cells, respectively. At steady state: dProl/dt=0, kprol=ktr kcirc= ktr MTT= (n+1)/k, n=No. of transit compartments.

PAGE 39

39 The differential equations corresponding to the change in the proliferating cells, maturating cells and circulating cells used in the model were as follows: dProl/dt =kprol.Prol.(1 Edrug).(Circ0/Ci Ktr Prol. dTransit1/dt=Ktr.Prol Ktr.Transit1. dTransit2/dt= Ktr.Transit1 Ktr.Transit2. dTransit3/dt= Ktr.Transit2 Ktr.Transit3. dCirc/dt= Ktr.Transit3 Kcirc.Circ. The drug concentration in the central compartment ( Conc) reduces the proliferation rate or induces the cell loss by a function E drug which could be modeled as a linear function ( slopeConc) or an E max model, E max Conc/(EC 50 Conc). The cell loss in the transit compartments takes place only into the next compartment. As the proliferative cells differentiate into more mature cell types, the concentration of cells is maintained by cell division. The parameters that could be estimated Ci rc 0, MTT slope (or E max and EC 50). Data analysis The structural model was used to explain the neutropenia across different ADCs. The individual PK estimates for all the cynos were obtained from the post hoc estimates of the PK analysis. All the mo lecules were fitted together, estimating the system related parameters MTT and FP together and differentiating the different molecules in terms of drug specific parameter Slope. Both the linear model ( Slope) and E max Model were used to describe the data, b ut the slope model explained the data well. the three transit compartments used in the structural model were used. The parameters estimated in the final model were MTT, FP and slopes were estimated individually per

PAGE 40

40 molecule. Estimating the Circ 0 data from all the cynos were analyzed sim ultaneously .The population model parameters estimated were the fixed effects, related to the typical individual, and random effects, with magnitude of residual variability between individual predictions and observations. Log normal parameter distributions were used for the IIV as follows: Where TVP is the typical population value, P i is the individual parameter value, i s are symmetrically distributed zero mean random variables, with a varia nce estimated as part of the model. The residual error was modeled with an additive and a propor tional component. The analyses were perfor med using NON MEM (version 7.2) The First Order M ethod (FO) was used. The First order methods implemented in NONMEM a re based on first order Taylor series linearizations of the prediction, with respect to the dependence on parameters. FOCE i.e the First order conditional estimation method was used along with the INTERACTION since these methods are most computationally complex and computer work .S Plus (version 8.2, TIBCO corp.) or R (Version 2.14.0) were used for the graphical diagnostics. The comparison of different models, using the objective function values (OFV) in the likelihood ratio test, guided the model selection. A difference in OFV of >10.83 was used for discrimination between two models differing in one parameter. All predictions (population and individual) were based on in dividual concentration time profiles.

PAGE 41

41 Figure 2 1. Antibodies armed with A uristatins. Figure 2 2. Neutrophil observations by dose and by molecule across different ADCs.

PAGE 42

42 Figure 2 3. Median baseline of the neutrophils across different ADCs with distribution of baseline of each cynomolgus monkey. E ach point represents an individual cynomolgus monkey. Figure 2 4. Two compartment pharmacokinetic model. V2 V1 Q CL

PAGE 43

43 Figure 2 5.

PAGE 44

44 CHAPTER 3 RESULTS The structural model explained the data set well (Figure 3 1 ) .Toxicity profiles were also well characterized when several treatment cycles were mod eled continuously in time (Figure 3 2 ).The mean parameters of the PK and PD model are reported in (Table 3 1) and (Table 3 2) The best residual error model was the combination (additive plus proportional model) as an additive model deteriorated the fits (data not shown). The syst em parameters i.e the MTT and FP were evaluated by fixing and estimating but found that estimating the system related parameters gave a better fit .The MTT estimated was 3.23 days and the IIV on MTT was 75.5% and the feedback parameter estimated was 0.282 and the interindividual variability on FP was 13.7 .The slopes estimated ranged about 10 fold across different molecules with lowest value of 0.00157 for molecule 21 to 0.054 for molecule 81.

PAGE 45

45 Figure 3 1. PK diagnostic plots

PAGE 46

46 Figu re 3 1. C ontinued. Figure 3 2. PK visual predictive c hecks

PAGE 47

47 Figure 3 3. PD diagnostic plots by molecule Table 3 1 Parameters of pharmacokinetic mod el describing intravenous data Mean values and corresponding interindividual variability (CV%). Molecule CL(ml/day) CL(ml/day/kg) V1(ml/day) V2(ml/day) Q(ml/day/kg) Mol1 48.5 14 146 177 108 Mol21 16.1 6 97.9 89.9 18.9 Mol22 16.4 6 91.2 71.9 62.9 Mol3 31.9 12 103 198 51.7 Mol4 22.2 8 126 161 73.7 Mol5 44.9 15 100 143 40.3 Mol6 26.2 8 119 191 47.2 Mol7 46.6 9 291 191 22.6 Mol81 50.8 18 112 117 59.1 Mol82 33.7 11 107 183 64.8

PAGE 48

48 Table 3 2. Typical population parameter estimates (relative SE %) for neutrophils with a linear Concentration Effect Model. Parameter Description Unit Estimate IIV(%) MTT Mean Transit time Days 3.23 (2.81) 75.5 (59.3) FP Feedback Parameter 0.282 (9.75) 13.7 (88.2) Slope1 ADC Drug effect on Molecule 1 L/ug 0.00844 (47.3) 114.5 (61.6) Slope 21 ADC Drug effect on Molecule 21 L/ug 0.00157 (52.5) 114.5 (61.6) Slope 22 ADC Drug effect on Molecule 22 L/ug 0.00218 (25.7) 114.5 (61.6) Slope 3 ADC Drug effect on Molecule 3 L/ug 0.00439 (27.1) 114.5 (61.6) Slope 4 ADC Drug effect on Molecule 4 L/ug 0.00193 (45.1) 114.5 (61.6) Slope 5 ADC Drug effect on Molecule 5 L/ug 0.0114 (49.5) 114.5 (61.6) Slope 6 ADC Drug effect on Molecule 6 L/ug 0.00259 56.8 114.5 (61.6) Slope 7 ADC Drug effect on Molecule 7 L/ug 0.00187 (31.4) 114.5 (61.6) Slope 81 ADC Drug effect on Molecule 81 L/ug 0.054 (33.1) 114.5 (61.6) Slope 82 ADC Drug effect on Molecule 82 L/ug 0.00255 (54.5) 114.5 (61.6) Residual Variability EPS1 5.87 (22.5) EPS2 41.4 (35.6)

PAGE 49

49 CHAPTER 4 DISCUSSIONS The derived model adequately described the observed neutrophil response after the administration of 10 different ADCs. In agreement with the aim of mechanistically based models, many features of the present model mimic physiological theories on the structu re and regulation of the granulopoietic system. The PK estimates i.e the clearances estimated from the PK model were about a threefold change from the Molecule 21 to Molecule 81.These v alues were shown in the (Figure 4 1) The clearance values were vali dated with the PK results from the PK data of same ADCs of different studies. The three fold increase in the PK values were as expected .This could be further confirmed from the (Figure 4 2) where there is a threefold difference in the PK values of a ntibod y drug conjugates due to unexplained reasons. Estimates of the system parameters were nevertheless not completely void of drug influence. S ince all the 10 ADCs had similar chemical structure and DAR and belong to the same class they are assumed to have s imilar system related parameters across all the ADCs and so the system related parameters across the 10 estimated with certainty. 22 The MTT estimated was found to be 3.23 days. Fri berg et.al used the exponent value of 0.158 to scale the estimated MTT for white blood cells of 52.8h in a 300g to a value of 125 h in a 70 kg patient. 23 .Cell turnover rate is thought to be inversely related to body size and therefore MTT is expected to be lower in rats than in patients and so considering the median body weight of cynomolgus monkeys i.e. 3kg

PAGE 50

50 and considering the exponent value of 0.158 to scale, the MTT can be estimated betwe en 2.2days(125 h) to 5.2 days (125 h) in agreement with the estimated parameter range in patients (0.121 0.239) close to the estimated value of 0.282 in cynomolgus monkey dimensionless and may imply that the regulation of production following myelosuppression occurs with a similar mechanism in rats, cynomolgus monkeys and patients A neutrophil half l ife of 6.7 hours has been reported in cynomolgus monkeys, compared with the estimated half l ife of 13.45 hours. Most likely, a lack of information about K circ in the data, as an estimated or fixed half previous analyse s, estimated half lives of circulating neutrophils have been longer than expected. The estimated i nterindividual variability on the slopes was the highest. A large interindividual variability in bone marrow cell toxicity of cytostatic drugs has been shown in vitro, which implies real differences in sensitivity of progenitor cells. Considering the fact that all the ADCs have same drug MMAE and similar mechanism of action and considering the similarities across different ADCs in terms of structure, DAR etc i t is hypothesized that the potencies are assumed to be similar across different ADCs which needs to be evaluated from the slope values estimated from the model. It was observed that there was a 10 fold difference across the slopes of 10 different ADCs i.e the esti mated slope values ranged from 0.00157 g/mL for Molecule 21 to 0.054 g/mL for Molecule 82.Different slopes could be due to many reasons Procedural and Biological reasons.

PAGE 51

51 Procedural The neutrophil data set consisted of 10 different molecul es and different sampling times It might be possible that some of the molecules had very sparse data samples ( one point per cycle ) making it difficult to estimate the accurate slope values .some of the datasets had very high rebounds randomly at di fferent c ycles. The data could no t be ignored since they are physiologically reasonable .T hey may be seen in different scenarios such as stress inflammation etc. There was no dose response (of neutrophils) was seen across differe nt ADC molecules. There was a large variability within the molecules. The above mentioned reasons were the challenges faced in terms of the dataset in estimating the slope factor accurately. T parameters together for al l the ADCs assuming similar drug MMAE used, chemical structure and mechanism of action. The rebounds could no t be explained by the model in terms of occasiona l variability since there was no trend for the random high rebounds that was seen in few cynomolgu s monkeys It could be possible that in order to capture the rebounds the model may be missing a few nadirs thereby affecting the slope value. The high inter subject variability made the model diffi cult to converge. Method 1 did no t converge due to various above reasons. better since it was very variable i.e the baseline absolute neutrophil counts ranged from 1.2 (1000/uL) to 18(1000/uL).The various steps in the development of the final model were as follows. EC50 values estimated in the Emax models were large leading towards linear models estimating slopes. Due to data limitations the system

PAGE 52

52 individual va rebounds where all the points were categorized into two different points where the points above 1. 5 times the baseline were categorized into type 1 points else type 0 points and a little more variability was allowed on the type 1 points compared to the type was disc arded. The final model consisted of the linear model with slope and single inter individual variability estimated on all the slopes of all molecules and estimating the system related parameters were estimated together for all the molecules with an additive and proportional error model. Biology perspective C onsider the similarities across the 10 ADC molecules i.e. the similar MMAE drug across the various ADCs, similar linkers, similar ADC structures, similar Drug to antibody ratios across all the ADCs i.e. the average DAR of 3.5 across all the 10 ADC molecules and considering the fact that neutropenia caused is a target independent factor the potencies of the ten ADC molecules was assumed to be simi l ar across the different ADCs. However, the estimated slope s were different from what was hypothesized indicating that the potencies of the 10 ADC molecules may differ inherently. In conclusion there is a 10 fold difference in the slopes across the 10 different ADC molecules .The hypotheses that different ADC mo lecules should have similar potency considering similar drug, chemical structure and DAR was disproved The estimated values of slopes differ by around 30 fold considering the 10 molecules. Among the 10 ADC molecules it was observed that 6 of the molecules have similar

PAGE 53

53 slopes ranging from 1 to 1.62 fold considering them into single pool The other 4 molecules had slopes as follows 0.00439, 0.00844 0.0114 and 0.0255 about 30 fold Hence further exploration is required to explore the reasons behind these f our molecules. It could be due to data related reasons or due to inherent properties. Thus the difference in the potency across different ADCs suggests us to explore reasons other than the ADC structure for differences in toxicity. Assumptions that similar DAR, linker drug should result in similar exposure response was disproved resulting in exploring into other causes. The model can be used for interspecies translation. Considering the known values i.e. exposure response in cynos and previous data in cynos along with humans regarding other ADC molecules the exposure response in humans can be predicted.

PAGE 54

54 Table 4 1. Clearance of Anti ETBR in cynomolgus monkey. Molecule Clearance Species Anti ETBR 8.99 ml/day/kg Cynologus monkeys 0.3 and 1 mg/kg. Figure 4 1. Clearance distribution across Molecules. Figure 4 2 Clearance values of antibodies in cynomolgus monkeys. Clearance values of antibodies in cynomolgus monkeys, individual animal data and geometric mean values (red bars).

PAGE 55

55 Table 4 2 Vari ous PD models explored. Model Method Objective function Description 101 Method 0 5754.178 Linear model with estimating slopes and IIV individually on slope and estimating MTT and FP 102 Method 0 5728.987 Linear Model with estimating slopes and single IIV on the slopes and estimating MTT and FP 103 Method 0 5494.871 Linear model with estimating slopes and baseline (single ) and single variability on slopes and estimating MTT and FP 104 Method 0 5459.028 Linear model with estimating s lopes and baseline individually and IIV individually and estimating MTT and FP 105 Method 1 5508.485 Model 101 with Method 1 106 Method 1 Inter 5474.99 Model 101 with Method 1 inter 107 Method 0 1963.771 Cycle 1 data with Model 101 108 Method 0 494.748 Log conc with method 0 109 Method 1 5689.723 Model 102 with method 1 110 Method 1 Inter 5636.599 Model 102 with Method 1 Inter 111 Method 0 5574.164 Model 101 with additive proportional model 112 Method 0 5653.632 Model 102 with additive proportional model

PAGE 56

56 Figure 4 3 Slope estimates of 10 ADC molecules. Figure 4 4 Correlation between CL of ADC Concentration vs Slope estimated. In conclusion there is a 10 fold difference in the slopes across the 10 molecules 0.00000E+00 5.00000E-04 1.00000E-03 1.50000E-03 2.00000E-03 2.50000E-03 3.00000E-03 3.50000E-03 4.00000E-03 4.50000E-03 1 2 3 4 5 6 7 8 9 10 Mol vs Slope Mol vs Slope 0.00000 0.00050 0.00100 0.00150 0.00200 0.00250 0.00300 0.00350 0.00400 0.00450 0.00500 0 2 4 6 8 10 12 14 CL vs Slope CL vs Slope Slope Molecule Clearance (ml/day/kg) Slope

PAGE 57

57 Figure 4 5. Correlation between cmax of ADC Concentration vs Slope estimated. Figure 4 6. Correlation between AUCSS of ADC Concentration vs Slope estimated. 0.00000 0.00050 0.00100 0.00150 0.00200 0.00250 0.00300 0.00350 0.00400 0.00450 0.00500 0 100 200 300 400 500 600 700 Cmax vs Slope 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005 0 200 400 600 800 1000 1200 AUCSS vs Slopes c max (ug/mL ) Slope Slope AUCSS (ug/mL.day)

PAGE 58

58 REFERENCES 1. May S. Kung Sutherland et al Lysosomal Trafficking and Cysteine Protease Metabolism Confer Target specific Cytotoxicity by Peptide linked Anti CD30 Auristatin Conjugates Jo urnal of biological chemistry, 2006 pp. 10540 10547 DOI: 10.1074/jbc.M510026200. 2. Paul J.Carter,PhD, and Peter D.Senter,PhD, Antibody Drug Conjugates for Cancer Therapy 2008 May Jun;14(3):154 69 DOI:10.1097/PPO.0b013e318172d704 3. Alejandro D Ricart et al Technology Insight: cytotoxic drug immunoconjugates for cancer therapy 2007 Apr;4(4):245 55 DOI: 10.1036/ncponc0774. 4. Pharmacokinetic considerations for Antibody Drug Conjugates arch, 2012 Sep;29(9):2354 66 DOI: 10.1007/s11095 012 0800 y. 5. C. Andrew Boswell. Impact of Drug conjugation on Pharmacokinetics and tissue distribution of anti STEAP1 Antibody Drug conjugate in Rats 2011 Oct 19;22(10):1994 2004 DOI: org/10.1021/bc2002 12a 6. David Schrama Ralph A. Reisfeld and Jrgen C. Becker Antibody targeted drugs as cancer therapeutics 2006 Feb;5(2):147 59 DOI :10.1038/nrd1957 7. Laurent Ducry and Bernhard Stump Antibody Drug Conjugates: Linking Cytotoxic Payloads to Monoclonal Antibodies 2010 Jan;21(1):5 13, DOI: 10.1021/bc9002019 8. Carter, Paul J. PhD; Senter, Peter D. PhD, Antibody Drug Conjugates for Cancer Therapy The Cancer Journal, 2008 May Jun;14(3):154 69 DOI: 10.1097/PPO.0b013e318172d704 9. Laurent Ducry and Bernhard Stump Antibody Drug Conjugates: Linking Cytotoxic Payloads to Monoclonal Antibodies 2010 Jan;21(1):5 13 DOI: 10.1021/bc9002019 10. Yelena V. Kovtun Victor S. Goldmacher Cell Killing by antibody drug conjugates Cancer Letters, 2007 Oct 8;255(2):232 40. Epub 2007 Jun 5 DOI: 10.1016/j.canlet.2007.04.010 11. Elena Soto Alexander Staab Christiane Doege Matthias Freiwald Gerd Munzert and Iaki F. Trocniz Comparison of different semi mechanistic models for chemotherapy related neutropenia: application to BI 2536 a Plk 1

PAGE 59

59 inhibitor Cancer chemotherapy and pharmacology, 2 011 Dec;68(6):1517 27. Epub 2011 Apr 24 DOI: 10.1007/s00280 011 1647 3. 12. Jeffrey Crawford, M.D.,David C. Dale, M.D.,Gary H. Lyman, M.D., M.P.H., Chemotherapy Induced Neutropenia Risks, Consequences, and New Directions for Its Management Cancer, 2004 Ja n 15;100(2):228 37 DOI : 10.1002/cncr.11882. 13. Zamboni WC D'Argenio DZ Stewart CF MacVittie T Delauter BJ Farese AM Potter DM Kubat NM Tubergen D Egorin MJ Pharmacodynamic model of topotecan induced time course of neutropenia 2001 Aug;7(8):2301 8 14. G ary h. L yman, C hristopher h. L yman, O layemi agboola, Risk Models for Predicting Chemotherapy Induc ed Neutropenia 2005 Jun Jul;10(6):427 37 15. Friberg LE Henningsson A Maas H Nguyen L Karlsson MO Model o f chemotherapy induced myelosuppression with parameter consistency across drugs 2002 Dec 15;20(24):4713 21 DOI: 10.1200/JCO.2002.02.140. 16. Paul Polakis Antibody Drug conjugates for Cancer therapy 17. Frederic Leger et.al Mechanism based models for topotecan induced neutropenia Clinical Pharmacology and Therapeutics, 2004 Dec;76(6):567 78 DOI:10.1016/j.clpt.2004.08.008 18. Scaling the time course of myelosuppression from rats to patients with a semi phy siological model Investigational New Drugs, 2010 Dec;28(6):744 53. Epub 2009 Aug 27 DOI: 10.1007/s10637 009 9308 7. 19. Pharmacokinetics and Pharmacodynamics: Poster Presentations Proffered Abstracts molecular cancer therapeutics 20. A strategy for risk mitigation of antibodies with fast clearance Landes bioscience mAbs, 2012 Volume 4, DOI: 10.4161/mabs.22189 21. Model of chemotherapy induced myelosuppression with parameter consistency across drugs Journal of Clnical Oncology, 2002 Dec 15; 20(24):4713 21 DOI: 10.1200/jco.2002.02.140.

PAGE 60

60 BIOGRAPHICAL SKETCH Manasa Tatipalli was born in An dhra Pradesh, India in 1987 and was the elder daughter of Linga Murthy Tatipalli and Sujatha Tatipalli She was awarded in Pharmacy from SR College of Pharmacy, Warangal, India in April 2009 She was enrolled in D epartment of Pharmaceutics at University of Florida, Gainesville in January 2010 and received h er Master of Science in Pharmacy in December 2012 Her research work was focused on semi mechanistic neutropenia modeling in cynomolgus monkeys which is having greater importance in pharmaceutical industry.