Objective: To utilize preclinical and phase I PK/PD data from a new quinolone (Q) and relevantpublic domain data to develop an exposure-response model for serum creatinine level increase byQ to support dose selection for subsequent clinical studies. Background: Reversible serumcreatinine elevations were observed during development of a novel Q that may confound clinicalsafety monitoring. Glomerular filtration rate (GFR) remained constant while creatinine urinaryclearance decreased suggesting the Q selectively inhibits creatinine renal tubular secretion.Methods: A population PK model was linked to a PD model of creatinine dynamics assumingcompetitive inhibition, consistent with preclinical data suggesting competitive inhibition ofcreatinine transport by Q. The PD model consisted of the following equation: d[Crn]/dt = ([Crn].GFR + RateCrnIn - RateCrnSec*[Crn])/VolCrn; where [Crn], GFR, RateCrnIn,RateCrnSec and VolCrn denote serum creatinine concentration (mg/dL), glomerular filtration rate(dL/Hour), zero order creatinine production rate (mg/Hour), creatinine tubular secretion rate(dL/Hour) and creatinine volume of distribution (dL). RateCrnSec was described as RateCrnSec =Vmax*[Crn]/(Km*(1 + [Q]/Ki) + [Crn])) where [Q] denotes the Q serum concentration. Theresulting model was used to simulate Q dose dependent increase in serum creatinine. Creatininedynamics parameters were derived from the literature. Results: Model supported competitiveinhibition of serum creatinine secretion (Ki 156 ng/mL, ED50, 40 mg) by Q. Simulations showedthat maximum serum creatinine increase occurred at Q doses of 200mg IV QD. Conclusion: Q maycompetitively inhibit serum creatinine secretion with near maximum increase at 200mg IV QD. Hence IV Q doses above 200mg will not produce major additional increases in serum creatininelevel.

ABSTRACT

SUMMARY OF MAJOR FINDINGS

SCHEMATIC OF GENERAL APPROACH

METHODS

Model Based Dose Selection of a Quinolone to Minimize Drug Induced Serum Creatinine ElevationS. Song1, S. Rohatagi1, P.K. Wickremasingha1, T. Khariton2, S. Kshirsagar2, T.J. Carrothers2, J. Kuwabara-Wagg2 1 – Daiichi Sankyo Pharma Development, Edison, NJ; 2 – Pharsight Corporation, Mountain View, CA

BACKGROUND

Serum creatinine elevations occur in association with Q exposure

• Mean maximal elevation above baseline was approximately 35%• Majority of elevation occurs within the first 24 hours of exposure• In most cases, creatinine levels remain within the normal range

This effect is not expected to be dose-limiting since

●Elevation appears to saturate at doses greater than 200 mg IV QD●Elevations are completely reversible upon cessation of Q dosing and return to

normal within 6 days of Q last dose

Preclinical data suggest that serum creatinine elevation results from Q mediated, competitive inhibition of creatinine renal tubular secretion

Simulations based on a physiological based model of human creatinine dynamics demonstrate that the proposed mechanism of elevation is consistent with the general time course of clinically observed creatinine elevations

Q is a novel quinolone antibiotic in phase I development

Serum creatinine elevations in response to Q administration were observed preclinically in rats, rabbits and monkeys

Preclinical in-vitro studies demonstrate that:

• 14C-creatinine is transported predominantly via hOCT2 (a human organic cation transporter), moderately, by hOCT2-A, and to a negligible extent by other organic cation transporters (hOCT1 & hOCT3) and organic anion transporters (hOAT1, hOAT2, hOAT3 & hOAT4)

• Q inhibits creatinine transport by hOCT2 and hOCT2-A transporters• The kinetics of this inhibition are competitive• Ki values for Q mediated inhibition of creatinine transport are estimated to be 1.8 and

0.7 μM for hOCT2 and hOCT2-A mediated transport, respectively• Genetic variants of the hOCT2 and hOCT2-A genes associated with decreased

transporter substrate affinity and/or increased turnover rates have been identified

An early multiple dose phase I investigation of Q was stopped due to observed moderate elevations of serum creatinine (400 and 800 mg dose groups)

• Most subjects experienced such elevation with a mean elevation from baseline of 35%

Comparable elevations were observed in an earlier single dose study (100 and 200 mg dose groups)

• Smaller elevations were also observed in lower dose groups (single dose study), e.g., 25 and 50 mg

Serum creatinine elevations were also noted in at least two other phase I studies

EXPLORATORY ANALYSIS

To utilize preclinical and phase I PK/PD data for a new quinolone, Q, and relevant public domain data to develop a exposure-response model for serum creatinine elevation

Exploit the model to establish whether the clinically observed time course and extent of serum creatinine elevation in response to Q administration is consistent with Q mediated competitive inhibition of creatinine renal tubular secretion

The general approach consisted of the following steps:

• Exploratory data analysis of phase I data• Utilize phase I PK data to develop a PK model for Q• Implement an established, public-domain, creatinine turnover model• Based on preclinical in-vitro and phase I data link the PK model to the creatinine turnover model• Simulate the relationship between Q dosing regimen and serum creatinine elevation• Compare the observed to the predicted serum creatinine elevations

Left Panel: Median % change of serum creatinine from baseline as a function of day for Q treatment (upper curve) and placebo (lower curve) arms of a phase I renal function study. Note that 800 mg Q was administered via a 1-hourinfusion on days 1, 2, 3 and 4. These days are denoted by the space between the solid vertical lines. Right Panel: Corresponding median % change of serum creatinine from placebo as a function of day for the phase I renal function study.

Q Dosing Regimen

Phase I PK DataPhase I PK Data

PK ModelPK Model[Q] Plasma (time) Physiological

CreatinineTurnover Model

[Creatinine]Serum(time)

Preclinical In-Vitro DataPreclinical In-Vitro Data

Phase I Renal Function StudyPhase I Renal Function Study

Public Domain Creatinine

Dynamics Model

Phase I Studies:

Serum Creatinine

Elevations

Phase I Studies:

Serum Creatinine

Elevations

Are predicted &

observed

creatinine profiles

consistent?

OBJECTIVES

Exploratory analysis of phase I renal function study data completed

●Mean and median serum creatinine percentage changes from baseline plotted as a function of time for both placebo and Q treatment arms●Difference between the median % change in serum creatinine from baseline for the Q treatment group and the comparable change from baseline for the placebo group calculated and plotted as a function of time●Concurrent changes in estimated Glomerular Filtration Rate (GFR) and Creatinine Urinary Clearance before, during and after Q dosing, plotted as a function of time.

Pooled phase I study data then used to calculate and plot the median % change from baseline of serum creatinine as a function of dose and an Emax function of dose fitted to the data

An established model of creatinine dynamics implemented (PD model)

●Parameter values obtained from the literature

Phase I PK data used to developed a PK Model for Q

The PK and PD models were linked under the assumption of Q mediated competitive inhibition of creatinine renal tubular secretion

Linked PK/PD model then used to simulate the time course of serum creatinine elevation as a function of Q dosing regimen

Simulated and observed serum creatinine profiles qualitatively compared

Serum Creatinine Elevations at 800 mg QD average about 30% above baseline

CREATININE SIMULATION RESULTS

Q effects on renal creatinine clearance were characterized in a phase I renal function study

The objective was to assess whether Q-mediated serum creatinine elevation reflect Q-mediated decreases of glomerular filtration rate (GFR) and/or renal tubular creatinine secretion

800 mg Q was administered daily for 4-days as an I.V. infusion

Creatinine clearance was calculated using a 24-hour urine specimen and estimated two days prior to dosing, on the last day of dosing, and, 1-week after the final dose

Glomerular Filtration Rate (GFR) was measured using the cold iohexol “gold standard” two days prior to dosing, on the last day of dosing, and, 1-week after the final dose

Physiologically, urinary creatinine clearance consists of two components, one reflecting glomerular filtration (GFR) and another reflecting renal tubular secretion. This is because renal tubular creatinine reabsorption is negligible.

RESULTS: Q RENAL EFFECT

Since creatinine clearance is given by the sum of GFR and renal tubular secretion, and, Q has minimal effects on GFR, the effects of Q on creatinine clearance are inferred to be mediated by reduced renal tubular secretion of creatinine.

Q renal effect is predominantly at the level of creatinine tubular secretion

Median % change in creatinine clearance (normalized for body surfacearea) for placebo and active treatment arms. Active treatment reducedcreatinine clearance by about 25%. The effect was reversible upon cessation of dosing. Placebo effects were negligible.

% change in median GFR (normalized for body surface) from baselinefor placebo and treatment arms. Error bars denote standard deviations. The effects of placebo and active treatment on GFR were negligible.

The difference between estimated median creatinine clearance and GFR is the estimated change in renal tubular secretion. Note thatactive treatment reduced renal tubular secretion of creatinine butthat placebo treatment effects were negligible.

The serum creatinine elevation effect is near maximum at 200 mg QD

Median % changes of serum creatinine from placebo for the 400and 800 mg treatment arms of a phase I study as a function of time. The majority of maximal effect is achieved within the first 24-hoursof dosing. The effect is maintained over the duration of dosing.Following the final Q dose, creatinine levels return to baseline levels over a period of several days.

24-hour post-first dose median change of serum creatinine frombaseline as a function of dose for data pooled across three phaseI studies (SAD and MAD studies combined).

*During PK model development, a saturable component of Q renal clearance was identified. The D50 of this effect was 31 mgcomparable to that observed for the serum creatinine elevationeffect (39 mg).

RENAL FUNCTION STUDY

Note:

1. Minimal steady-state Q accumulation

2. Dotted horizontal line denotes Q KI value of 156 ng/mL

Simulated % Change from Baseline of Serum Creatinine vs Timefor DX619 QD IV dosing regimens of 14 days duration

0.0%

10.0%

20.0%

30.0%

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time (Days)

% C

hg

Se

rum

Cre

atin

ine

fro

m B

L

DX619 dose = 400 mg

DX619 dose = 800 mg

400 mg Phase I Data

800 mg Phase I Data

Simulated serum creatinine profiles as a function of regimen are comparable to those observed in phase I studies

CREATININE DYNAMICS MODEL

Simulated serum creatinine as a function of time for administrationof 50, 100, 400 and 800 mg QD IV for 14-days. Simulations were based on a physiological model of serum creatinine dynamics in which Q mediates acompetitive inhibition of creatinine renal tubular secretion.

Simulated percentage change from baseline of serumcreatinine as a function of time for adminstration of 400and 800 mg IV Q for 14 days. Filled triangle and squaresdenote the observed median % changes of serumcreatinine from placebo for the 400 and 800 mgtreatment arms of a phase I study.

PK MODEL SUMMARY AND PK/PD LINK

VolCrn

RateCrnIn

GlomerularFiltration

TubularSecretion

VolCrn

RateCrnIn

GlomerularFiltration

TubularSecretion

The creatinine dynamics model consists of a single compartment into whichcreatinine is released at constant rate (RateCrnIn), and, from whichcreatinine is cleared via the renal mechanisms of glomerular filtration(GFR) and tubular secretion (RateCrnSec). Renal reabsorption ofcreatinine is assumed to be negligible as is extrarenal clearance ofcreatinine. The volume of creatinine distribution (VolCrn) is assumed to beequal to the volume of total body water. Under these assumptions, the rateof change of serum creatinine concentration is given by the equation below.

Creatinine Renal

Tubular Secretion

DoseI.V. infusion

Central

V1

Central

V1

CL

K12Peripheral

V2

Peripheral

V2 K21

RateCrnSec was described as a saturating function of plasma Q levels [Q] as outlinedbelow.

Based on preclinical data, Q is assumedto act as a competitive inhibitor of renaltubular secretion such that RateCrnSec isa function of plasma Q levels, [Q].

Plasma (central) Q levels weredescribed using a 2-compartmentPK model

Linked PK/PD model equations were solved for dosing regimens of interest to simulate:

• Plasma Q levels as a function of time • Serum creatinine levels as a function of time