Analytical and physicochemical investigations of anthracycline antibiotics

MISSING IMAGE

Material Information

Title:
Analytical and physicochemical investigations of anthracycline antibiotics
Physical Description:
Book
Creator:
Runyan, Anita Karen, 1955-
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 17864958
System ID:
AA00020011:00001

Table of Contents
    Title Page
        Page i
    Acknowledgement
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
    List of Tables
        Page vi
        Page vii
    List of Figures
        Page viii
        Page ix
        Page x
    Abstract
        Page xi
        Page xii
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
    Chapter 2. Research objectives
        Page 24
        Page 25
        Page 26
        Page 27
    Chapter 3. Electrochemistry
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
    Chapter 4. Analytical method development
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
    Chapter 5. Applications of analytical methods
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
    Chapter 6. Conclusions and suggestions for future work
        Page 130
        Page 131
        Page 132
        Page 133
    References
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
    Biographical sketch
        Page 141
        Page 142
        Page 143
Full Text













ANALYTICAL AND PHYSICOCHEMICAL INVESTIGATIONS
OF ANTHRACYCLINE ANTIBIOTICS






BY






ANITA KAREN RUNYAN


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1986













ACKNOWLEDGMENTS


I would like to express my appreciation to my mother for her

enduring faith and support which have always given me a firm

foundation from which to grow.

I want to thank Dr. Chris Riley, for taking me in as his first

graduate student. I appreciate his interest and help in initiating my

research project and his continued support, encouragement and advice

which allowed me to pursue my academic goals.

Individual thanks are given to each committee member. I want to

thank Dr. Perrin for his concern and job-finding services for which I

am most grateful. I appreciate the advice and assistance with

experimental methods I received from Dr. Derendorf. I would like to

thank Dr. Schulman for his expert advice in spectroscopy. Special

thanks are extended to Dr. Dorsey for the use of his laboratory and

equipment and for his excellent advice.

I would also like to thank Dr. Anna Brajter-Toth for her time and

explanations of electrochemistry. I am grateful to Dr. C.W. Young for

his interest, financial support and valuable discussions of my

research.

To all my friends at school, I would like to say that they have

made this time a most memorable and pleasant experience. I wish them

all the best. Finally, I extend my sincere gratitude to all my








acquaintances at the University of Florida who have contributed to my

education.













TABLE OF CONTENTS


Page

ACKNOWLEDGMENTS.................................................... ii

LIST OF TABLES..................................................... vi

LIST OF FIGURES.................................................. viii

ABSTRACT...................................................................... xi

CHAPTERS

1 INTRODUCTION...................... ..... .................... 1

1.1 Clinical Uses of Anthracycline Antibiotics............
1.2 Physicochemical Properties........................... 1
1.3 Biological Activity, Toxicity and Pharmacokinetics....5
1.4 Interactions with Blood Components and Tissues.......11
1.5 Review of Analytical Methods........................ 13
1.5.1 Chromatography............................... 13
1.5.2 Sample Preparation........................... 21
1.5.3 Electrochemistry............................. 22

2 RESEARCH OBJECTIVES...................................... 24

2.1 Analytical Method Development....................... 24
2.2 Static and Flow Electrochemistry.................... 26
2.3 Applications of Analytical Methods.................. 26

3 ELECTROCHEMISTRY......................................... 28

3.1 Materials and Methods............................... 28
3.1.1 Chemicals and Reagents....................... 28
3.1.2 Cyclic Voltammetry........................... 29
3.1.3 HPLC-EC...................................... 29
3.2 Results and Discussion..................... .......30
3.2.1 Cyclic Voltammetry........ .............. ..30
3.2.2 HPLC-EC...................................... 36

4 ANALYTICAL METHOD DEVELOPMENT............................ 41

4.1 Materials and Methods............................... 41
4.2 Results and Discussion.............................. 45









4.2.1 HPLC .......................................... 45
4.2.2 Sample Preparation............................60
4.2.3 Assay Validation ............................. .. 67

5 APPLICATIONS OF ANALYTICAL METHODS....................... 76

5.1 Materials and Methods............................... 76
5.1.1 Stability Studies............................ 76
5.1.2 Protein Binding.............................. 77
5.1.3 Erythrocyte Binding.......................... 79
5.1.4 Clinical Pharmacokinetics of Doxorubicin
and Deoxydoxorubicin....................... 81
5.2 Results and Discussion.............................. 82
5.2.1 Stability Studies............................ 82
5.2.2 Distribution of Anthracyclines Between the
Components of Blood........................... 85
5.2.3 Protein Binding............................... .. 88
5.2.4 Erythrocyte Binding........................... 95
5.2.5 Clinical Pharmacokinetics of Doxorubicin
and Deoxydoxorubicin........................ 118

6 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK..............130

REFERENCES......................................................... 134

BIOGRAPHICAL SKETCH............................................... 141














LIST OF TABLES


Table Page

1-1 Normal Phase HPLC Systems Used for the Analysis of
Anthracycl ines ..................... ........................ .... 15

1-2 Reversed Phase HPLC Systems for the Analysis of
Anthracyclines in Biological Fluids......................... 16

4-1 Summary of Chromatographic Conditions for the
Analysis of Anthracycline Antibiotics by HPLC-EC............44

4-2 Chromatographic Parameters of CN on Zorbax ODS (a)
and ODS Hypersil (b) Columns............................... 46

4-3 Chromatographic Parameters of CN on SAS Hypersil Column.....47

4-4 Effect of pH on the Chromatographic Parameters of CN on
a SAS Hypersil Column..................................... 49

4-5 Recovery of Six Anthracyclines from Borate Buffer
(1 ml)(pH 8.1-9.0) Extracted with 10 ml of Chloroform:
Isopropanol (1:1)..............................................63

4-6 Recovery of DEOXY from 1 ml of Borate Buffer (pH 8.6)
with Different Organic Solvent Mixtures.................... 63

4-7 Chromatographic Conditions Used for Buffer and Plasma
Extraction with Chloroform:Isopropanol (1:1)............... 64

4-8 Recoveries of Six Anthracyclines (500 ng/ml) from Buffer
and Plasma (1 ml) Extracted with 10 ml of Chloroform:
Isopropanol (1:1) ........................................... 66

4-9 Limits of Detection (SNR=2) of the Anthracycline
An ti bi o tics ................................................. 68

4-10 Statistical Parameters of Sample Calibration Curves
from Extracted Plasma Samples............................... 70

4-11 Absolute Recoveries of the Anthracycline Antibiotics
from Plasma................................................. 71









4-12 Accuracy and Precision Data for the Analysis
of Anthracycline Antibiotics in Plasma by HPLC-EC...........72

4-13 Statistical Parameters of Sample Calibration Curves
for Extracted Urine Samples of DOX......................... 74

4-14 Absolute Recoveries of Doxorubicin from Urine...............75

4-15 Accuracy and Precision Data for the Analysis of
Doxorubicin in Urine........................................ 75

5-1 Stability of Anthracycline Antibiotics in pH 7.4
Physiological Buffer at 25C............................... 83

5-2 Plasma Protein Binding of the Anthracyclines
Determined by Ultrafiltration at 250C.......................90

5-3 Plasma Protein Binding Determined by Red Blood Cell
Partitioning at 25C........................................ .. .. ... ..96

5-4 Partition Coefficients of Anthracyclines Between Red
Blood Cells and Plasma Water at 25C ....................... 99

5-5 Partition Coefficients of the Anthracyclines Between
Red Blood Cells and pH 7.4 Buffer at 25C..................100

5-6 First Order Influx (kinf) and Efflux (keff) Rate
Constants of the Anthracyclines at 25C ....................111

5-7 Effect of Loading Concentration on keff and kinf
of DEOXY................................................... 112

5-8 Pharmacokinetic Parameters of DEOXY After IV Injection
of 52 mg Determined by Electrochemical (HPLC-EC) and
Fluorescence (HPLC-FL) Detection.......................... .....120

5-9 Pharmacokinetic Parameters of DOX after IV Administration
to 2 Patients, RB and HH, at 3 Week Intervals (1 & 2,
First and Second Dose) ..................... .............. 126













LIST OF FIGURES


Figure Page

1-1 Structures and abbreviations for the anthracyclines
investigated in this study................................... 2

3-1 Cyclic voltammograms of EPI (a) and DOX (b) from 0.0 to
+1.0 V..................................................... 32

3-2 Cyclic voltammograms of five anthracyclines from 0.0
to +1.0 V at pH 4.......................................... 33

3-3 Cyclic voltammograms of CN, DEOXY and DMDR from 0.0
to -1.0 V ................................................... 35

3-4 Hydrovoltammograms of five anthracyclines, showing
peak height as a function of applied potential..............37

3-5 Comparison of hydrovoltammograms obtained by coulometric
(a) and amperometric (b) detection of DOX and DEOXY.........38

4-1 Effect of adding sodium dodecyl sulfate (SOS) to the
mobile phase on the retention of CN (k') ................... 52

4-2 Effect of adding sodium dodecyl sulfate (SDS) on the
chromatography of CN ........................................ 53

4-3 Effect of temperature on the micellar chromatography
of CN ....................................................... 54

4-4 Chromatograms of six anthracyclines using the same
chromatographic conditions................................. 57

4-5 Optimized separation of seven anthracyclines from
plasma..................................................... 58

4-6 Diagram of chromatographic system.......................... 61

5-1 Stability of the two isomers of CN......................... 84

5-2 Relationship between the ratio of the protein bound
and free fractions of the anthracyclines (f /fp u)
and their chromatographic capacity ratios (k '..............91


viii









5-3 Relationship between the ratio of the protein bound
and free fractions of DEOXY (f b/fn ) and the fraction
of plasma in pseudoplasma (m), quidligrated with red
blood cells................................................. 94

5-4 Relationship between the apparent partition coefficient
(D) of DOX between plasma water (o---o ) and physiological
buffer (o-- ) and time.................................. ..... 98

5-5 Histogram showing the effect of concentration on the
apparent partition coefficient (D) of the anthracyclines
between plasma water and red blood cells (13 ) and
physiological buffer and red blood cells (D )............. 101

5-6 Disappearance of CN1 (A) and CN2 (A) from a red blood
cell suspension in physiological buffer....................103

5-7 Relationship between the apparent red blood cell partition
coefficients (D) of CN1 (A) and CN2 (A) and time.........104

5-8 Relationship between the red blood cell partition
coefficients (D) of the anthracyclines and their
chromatographic capacity ratios........................... 107

5-9 Relationship between the fractional cellular content
of the anthracyclines in red blood cells and time......... 110

5-10 Relationship between the logarithms of the first order
efflux (kff) and influx (kinf) rate constants for the
anthracyclines and logarithms of their chromatographic
capacity ratios (k ) ...................................... 114

5-11 Simulated free plasma concentrations (C ,) for the
anthracyclines............................................116

5-12 Simulated free plasma concentrations (Cp ) for ?he
anthracyclines after dose normalization /OQ mg/m )
of the curves shown in Figure 5-11.........................117

5-13 Plasma levels of DEOXY following IV administration of
52 mg of the drug, determined by HPLC-FL (*-* ) and
HPLC-EC (o----o).......................................... 119

5-14 Chromatograms of plasma (a) and urine (b) taken from
a patient (RB) who had received 60 mg of DOX by IV
injection.................................................. 122

5-15 Plasma levels of DOX following two IV doses of 60 mg
of drug given to patient RB, three weeks apart.............123










5-16 Plasma levels of DOX following two IV doses of 63 mg
of drug given to patient HH, three weeks apart.............124

5-17 Cumulative amounts (Ut) of DOX excreted in the urine by
RB and HH expressed as a percentage of the given dose......129













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ANALYTICAL AND PHYSICOCHEMICAL INVESTIGATIONS
OF ANTHRACYCLINE ANTIBIOTICS

BY

ANITA KAREN RUNYAN

May, 1986

Chairman: Dr. Christopher M. Riley
Major Department: Pharmacy

A method of analysis for six anthracyclines has been developed

which uses high performance liquid chromatography with electrochemical

detection (HPLC-EC). The analytical methodologies were used in

pharmacokinetic studies and to investigate the interactions of

anthracyclines with blood components such as plasma proteins and

erythrocytes.

All the compounds of interest may be analyzed using a reversed

phase column (ODS Hypersil or Zorbax ODS) and a mobile phase of

acetonitrile-isopropanol-O.1 M phosphate buffer (pH 4.5)(25:3:72).

The use of short columns (7.5 cm) allows short analysis times and high

sensitivities so that detection limits in plasma of between 1 and 2

ng/ml can be achieved. Sample preparation involved an alkaline

extraction into chloroform, followed by a back extraction into acid.

Final clean-up of the samples was achieved by shaking the final acidic

extract with an organic solvent. The precision of the assay for the








compounds studied ranged between 1.22% and 6.46% and the accuracy

ranged between 94.5% and 106.0%.

The analytical methodology developed was applied to the clinical

pharmacokinetics of 4'-deoxydoxorubicin. The correlation between

plasma levels found by HPLC-EC and those found by HPLC with

fluorescence was excellent (r2=0.990). Pharmacokinetic studies of

doxorubicin were initiated in two patients undergoing long term

treatment for osteosarcoma which includes concurrent administration of

doxorubicin and cisplatin.

Several methods which produced comparable results were used to

determine the extent of plasma protein binding of the

anthracyclines. All of the anthracyclines show a high degree of

binding to plasma proteins (92.4% to 98.8%). Based on quantitative

structure activity relationships (QSARs) relating the extent of

binding to hydrophobicity, anthracyclines with different chromophore

substituents exhibit altered protein binding characteristics from

analogues with different aminosugar substituents.

The extent of anthracycline partitioning into red blood cells was

found to increase with increasing hydrophobicity. The kinetics of

anthracycline interactions with red blood cells were investigated by

determining the rates of efflux of the compounds out of red blood

cells. The rates of efflux and rates of influx, determined by

calculations based on the efflux rates, of the anthracyclines were

also shown to be related to their hydrophobicities. The rates of

influx were significantly faster than the rates of efflux for all the

compounds. Results from these studies may help to explain the

differences in the biological activity of the anthracyclines.













CHAPTER 1
INTRODUCTION


1.1 Clinical Uses of Anthracycline Antibiotics

The anthracycline antibiotics are a class of chemotherapeutic

agents originally isolated from the fungus Streptomyces peucetius

(1). Daunorubicin and doxorubicin (Adriamycin) are the most widely

used of the anthracycline antibiotics (2) and play a major role in the

effective treatment of a wide range of cancers including sarcomas,

breast cancer, Hodgkin's disease, non-Hodgkin's lymphomas, and acute

leukemia (2). Due to their significant acute and chronic toxicities,

one of the worst being cardiac toxicity (3), there is a continuous

search for active anthracyclines with less severe side effects and/or

increased potency.



1.2 Physicochemical Properties

The basic structure of the anthracycline antibiotics (Figure 1-1)

is a glycoside composed of a tetracycline quinoid aglycone linked to

an aminosugar. Studies on the relationship between the molecular

structure and biological activity of these drugs give guidelines for

the design of improved drugs. The structures of six anthracyclines

which have been chosen for study are given in Figure 1-1. The only

structural difference between doxorubicin (DOX) and daunorubicin (DNR)

is the substitution of a hydroxyl group at position 14. Epimerization






















R1 R2 R3


CH30

CH30


H NH2

OH NH2


H H NH2

CH30 H N H2
J CN
CH30 H

CH3O H NH2


OH DOX


H OH EPI


H DMDR


H OH DEOXY


OH

OH


OH CN

H DNR


Figure 1-1.


Structures and abbreviations for the anthracyclines
investigated in this study. DOX = doxorubicin. EPI = 4'-
epidoxorubicin, DMDR = 4-demethoxydaunorubicin, DEOXY =
4'-deoxydoxorubicin, CN = 3'-deamino-3'-(3-cyano-4-
morpholinyl)doxorubicin, DNR = daunorubicin.


R4 Rs









of the hydroxyl group in the 4' position gives 4'-epidoxorubicin (EPI)

and removal of the 4'-hydroxy gives 4'-deoxydoxorubicin (DEOXY).

Replacement of the 3'-amino group with a 3-cyano-4-morpholino group

gives 3'-deamino-3'-(3-cyano-4-morpholinyl )doxorubicin (CN) which

exists as a pair of diastereomers due to the lack of chiral control in

the attachment of the cyano group to the morpholino ring during

synthesis (4). Removal of the 4-methoxy group from the chromophore of

daunorubicin gives 4-demethoxydaunorubicin (DMDR).

X-ray crystallographic studies of N-Br-acetyl-daunorubicin show

that the cyclohexane ring is in the half-chair conformation and that

the sugar moiety is almost perpendicular to the plane of the

chromophore (5). A difference in the electronic charge within the

quinonic chromophore is indicated by a slightly smaller distance of
0
2.45 A between the 0-5 and 0-6 than the distance between the 0-11 and

0-12 of 2.67 A. The molecular conformation of the molecule is also

affected by intramolecular hydrogen bonding (6). The stability of the

half-chair conformation of the cyclohexane ring may be increased by

hydrogen bonding between the 0-9 and 0-7. The 0-9 may also bond with

the oxygen in the sugar ring, which helps to stabilize the orientation

of the sugar to the aglycone.

The aminosugar is an important structural requirement for the

biological activity of these compounds (7). The aglycone is inactive

and is highly insoluble in water. The amino group on the carbohydrate

allows the glycoside to dissolve in aqueous solutions mostly in the

cationic form at neutral pH. Hydrogen bonding by the protonated amino

group and the 0-4 hydroxyl contributes to the stabilization of









intermolecular systems. The carbohydrate residue contains four of the

six chiral centers in the molecules (Figure 1-1). Many analogues of

doxorubicin have alterations in the aminosugar stereochemistry and

substituents which modify the efficacy and toxicity of the compounds

(6). Active transport and enzyme reactions are also known to depend

on the carbohydrate structure (8).

Doxorubicin shows characteristic ultraviolet, visible and

fluorescent spectra due to its chromophore. The ultraviolet and

visible maxima are 233 nm, 253 nm, 290 nm, 477 nm, 495 nm, and

530 nm. In alkaline media the maxima are shifted to longer

wavelengths. Around pH 9 an orange-red aqueous solution will turn to

a blue-violet color (9). The chromophore is also responsible for the

electrochemical behavior of the anthracycline antibiotics as it

contains two oxidizable hydroquinone centers and two reducible quinone

groups.

Due to the planar aromatic ring system strong interactions

between doxorubicin molecules result in the formation of dimers and

higher order association species which influence the solution behavior

of doxorubicin at concentrations greater than 10-6 M (10). Log

formation constants for dimers and tetramers of doxorubicin in aqueous

solution are about 4.5 and 12, respectively, over the pH range of 4 to

6 (10). Menozzi et al. (11) have also studied the self-association of

doxorubicin, daunorubicin, and several of their analogues in aqueous

solution. Substitution on the chromophore, buffer composition and

ionic strength are known to influence self-association of

anthracycline antibiotics (11).








Doxorubicin contains several ionizable functional groups (Figure

1-1). The degree of protonation of the phenolic groups and the amino-

sugar group changes significantly over the pH range 8 to 13 (12). The

microscopic dissociation constants have been reported by Sturgeon and

Schulman (12). The protonated aminosugar group is slightly more

acidic than the phenolic group. The pKa of the amino group has been

reported to be 8.22 by Sturgeon and Schulman (12) and 8.34 by DiMarco

et al. (13). These values are greater than the value of 8.08 found

for 4'-epidoxorubicin but less than 8.69 found for 4'-deoxydoxorubicin

(13). A pKa of 8.46 has been reported for the amino group of

daunorubicin (13).

It is believed that the cationic form of the anthracyclines binds

to DNA and a linear relationship has been shown between the apparent

DNA binding constants and pKas for a set of daunorubicin analogues in

which no steric effects are involved (13). In addition, the observed

differences in in vitro inhibitory activity of the analogues have been

attributed to their differences in pKa (13).



1.3 Biological Activity, Toxicity and Pharmacokinetics

Precise relationships between physicochemical properties and

antitumor activity have been very difficult to establish. Activity

appears to be related to DNA binding although the cyano analogue shows

weaker DNA binding but much greater potency (4). Also, comparison of

the DNA binding of daunorubicin and some of its analogues showed no

correlation between the drugs' DNA affinity and their antitumor

activity (14). Many other factors affect activity including protein








binding, cellular uptake and metabolism. It has been observed that

cellular uptake may be inversely related to drug polarity (15). The

cellular uptake of doxorubicin and some analogues show significant

differences (13).

A wide range of biochemical effects which are potentially toxic

to both host and tumor cells are produced by the anthracyclines.

Three known mechanisms of action are DNA intercalation, lipid

peroxidation and membrane binding. The best documented action is the

interaction of these agents with DNA (16). The aglycone portion of

the molecule inserts into the DNA double helix between the adjacent

base pairs and parallel to them (16). The aminosugar binds

electrostatically to the sugar phosphate backbone of DNA (2). It is

well known that positively charged intercalating molecules interact

strongly with DNA in the fully cationic form (17). Therefore it is

reasonable to believe that anthracyclines must be in the cationic form

for effective DNA binding. The binding may cause blockage of

synthesis of DNA, RNA and protein, inhibition of DNA repair and

fragmentation of the DNA (14). Anthracyclines have been shown to

cause free radical formation and Handa and Sato (18) have demonstrated

that microsomal P reductase catalyzed the reduction of daunorubicin or

doxorubicin to a semiquinone free radical which rapidly reduced

molecular oxygen in the superoxide ion. It has been suggested that

this oxygen radical could cause the membrane damage seen in cardiac

tissue. This effect may contribute to the mutagenicity and cardiac

toxicity of the anthracyclines but appears to have no antitumor

influence (2).








Cell membrane binding of doxorubicin alters membrane function and

may increase sodium permeability and alter calcium handling (19).

Binding to spectrin and cardiolipin have been demonstrated (20). It

appears that malignant cells and cardiac mitochondria have an

increased cardiolipin content which might explain why the

anthracyclines damage both tumor cells and cardiac tissue (2).

After intravenous (IV) administration, doxorubicin shows a

triphasic disappearance curve with the initial phase representing

tissue uptake, the second phase metabolism and the final phase the

release of the drug from various binding sites. The half lives of

each phase are approximately 10 to 30 minutes, 10 hours, and 24 to 48

hours (21). The terminal elimination rate constant may be difficult

to determine if data are collected for only the first 24 to 36 hours

after administration. The predominant metabolite of doxorubicin is

doxorubicinol, 13-dihydrodoxorubicin, produced by reduction of the 13-

keto group (22). Similarly daunorubicin is metabolized to

daunorubicinol. Both are reduced by an aldo-keto reductase which is

present in most human tissues (23). Both metabolites have some

antitumor activity (24). Other doxorubicin metabolites which have

been identified include the doxorubicin aglycone, doxorubicinol

aglycone, 7-deoxy-doxorubicinol, and the 4-0-glucuronide and 4-0-

sulfate conjugates (24).

The major route of drug elimination is biliary excretion which

accounts for elimination of approximately 40% of the dose (25). Renal

excretion is usually considered a minor route of total drug

elimination (26), although a wide range of values for the percentage








of doxorubicin excreted in the urine have been reported (27, 28).

Benjamin et al. (28) found only 3.4% of a single dose of doxorubicin

was excreted in 5 days after IV administration to patients with solid

tumors or leukemia having normal renal function. In a study by Rosso

et al. (27) doxorubicin was administered to patients having normal

renal function with solid tumors for three consecutive days. After up

to 14 days, the percent of dose in the urine ranged from 28 to 51.1%,

while only 9.6% was recovered from one patient with impaired renal

function (27). The analogues have been shown to undergo similar

metabolic and elimination processes (29, 30).

Comparison of the activity of the analogues have shown that

epidoxorubicin has a higher efficacy than doxorubicin in vivo against

Lewis lung tumor, MS-2 sarcoma, and murine sarcoma virus-induced

sarcoma (31). Epidoxorubicin and doxorubicin have comparable activity

against L1210 and P388 leukemias and ascitic and solic sarcoma 180

(32). Epidoxorubicin has also shown activity against renal cell

carcinoma, malignant melanoma and colon carcinoma (33).

Studies relating tissue distribution and toxicity in mice showed

that administration of equal doses of doxorubicin and epidoxorubicin

resulted in lower peak levels and area under the curve (AUC) up to 48

hours in both the heart and spleen for epidoxorubicin compared to

doxorubicin. Activity and toxicity have been correlated with AUC and

maximum plasma concentration. Epidoxorubicin had less cardiotoxicity

and less general toxicity than doxorubicin (29). In rats 40 to 45% of

both drugs undergo biliary excretion either as the unchanged compound

or as the reduced metabolite. Epidoxorubicin was found to be more








extensively metabolized over the time period studied indicating a

faster metabolic rate than doxorubicin (30).

In humans, comparison of plasma levels of patients given the same

dose of doxorubicin and epidoxorubicin at a three-week interval showed

that plasma levels of epidoxorubicin were significantly lower (20 to

40%) than those of doxorubicin. Both followed a three-compartment

open body model with epidoxorubicin having a shorter terminal half-

life (30 h) than doxorubicin (43 h). Epidoxorubicin showed a higher

plasma clearance than doxorubicin, which could be due to higher renal

clearance of epidoxorubicin and/or extrahepatic metabolism (31).

Deoxydoxorubicin has a spectrum of antitumor activity different

from that of doxorubicin (35). Results from five different types of

solid tumors studied in mice showed that at the optimal dose, defined

as the dose that produces the maximal antitumor activity without

serious toxic effects, deoxydoxorubicin was more active than

doxorubicin on colon 38 and 26 carcinoma, about as active on mammary

carcinoma and on MS-2 sarcoma, and less active than doxorubicin on B16

melanoma (35). Similar results from studies on human tumors

heterotransplanted into nude mice found deoxydoxorubicin more active

than doxorubicin on colon carcinoma (36) and slightly less active on

melanoma (37).

In a study on tissue distribution and toxicity in mice by

Formelli and Casazza (29) deoxydoxorubicin had higher spleen peak

levels and AUC at 48 hours than doxorubicin which correlated with its

higher general toxicity than doxorubicin. In the heart the AUC up to

48 hours was lower for deoxydoxorubicin than doxorubicin which was








attributed to its faster elimination from that tissue. At the

tolerated doses in mice, the minimal cumulative cardiac dose (MCCD) of

deoxydoxorubicin could not be established and minimal cardiac lesions

were observed indicating almost no cardiotoxicity of deoxydoxorubicin

(29).

In humans deoxydoxorubicin appears to have some differences in

pharmacokinetics compared with doxorubicin. Both follow a three-

compartment model with deoxydoxorubicin having a very rapid initial

phase with a half-life of 1 to 2 minutes compared to approximately 20

minutes for doxorubicin. The terminal half-life of deoxydoxorubicin

was estimated to be 90 hours which is much longer than that of

doxorubicin (24 to 48 hours) (38).

Of the analogues chosen for this study, the most potent

synthesized so far is 3'-(deamino)-3'-(3-cyano-4-morpholinyl)doxo-

rubicin (CN) (Figure 1-1). From early studies of largely separated

diastereomers, it appears that the chirality of the cyano group does

not affect the activity. This drug is 600 times more potent than

doxorubicin against P388 leukemia in mice (39). It is up to 1000

times more potent against tumors in cell culture and in mice and

appears not to produce any chronic myocardial lesions (4). This

analogue shows somewhat weaker binding to DNA than doxorubicin but

does inhibit nucleic acid synthesis (4). The cyano analogue may have

an altered base-pair specificity or an additional non-intercalative

mode of binding (39). The morpholino group of the cyano analogue is

very weakly basic and the compound cannot be extracted into aqueous

acid or converted to the hydrochloric salt (4). This is due to the








influence of the alpha-CN on the tertiary amine which decreases the

pKa of the nitrogen by about six units (40). Other non-basic

analogues where the nitrogen was acylated or removed showed some

activity, but the cyano-substituted analogue shows a level of potency

attained by few other antitumor agents.

The demethoxy analogue of daunorubicin has about eight times the

antineoplastic activity of its parent compound; however, its toxicity

is also increased (6). In the study by Formelli and Casazza (29)

demethoxydaunorubicin had higher spleen peak levels and AUC up to 48

hours than daunorubicin, which correlates with its higher general

toxicity. Demethoxydaunorubicin had a lower MCCD than daunorubicin

and has been shown to be retained in the heart for longer periods of

time.



1.4 Interactions with Blood Components and Tissues

Doxorubicin has been shown to be substantially bound to plasma

proteins and tissues in animals. Using equilibrium dialysis,

doxorubicin was found to be 78% bound to rat plasma proteins at a

concentration of 500 ng/ml at 37C (41). Other investigators (42)

report doxorubicin 65.6% plasma protein bound in rats, 58.5% bound in

rabbits, and 47.1% bound in guinea pigs at concentrations up to 2

ug/ml at 37C using equilibrium dialysis. Apparently few studies have

been done on human plasma. A value of 50% protein binding for human

plasma determined by ultracentrifugation and equilibrium dialysis has

been reported (43). However, re-evaluation of the equilibrium








dialysis results by Chan et al. (21) indicated that doxorubicin is 90%

bound in the therapeutic concentration range in humans.

The interaction of doxorubicin with human erythrocytes has been

studied to determine the type of transport process involved (44,

45). By measuring the net efflux of doxorubicin from loaded human red

blood cells, the transport of doxorubicin showed apparent saturation

kinetics and a dependence on temperature and concentration (44).

These phenomena were found to be caused by changes in the

physicochemical properties of doxorubicin with concentration and

temperature, mainly self-association, and not changes in doxorubicin

interactions with cell membrane components (e.g., saturation) (44).

Doxorubicin transport also changes with pH (45). Doxorubicin appears

to cross cell membranes by simple Fickian diffusion through the lipid

portion of the membrane (45). It should be noted that many of these

binding studies were done at concentrations well in excess of those

found in blood following conventional dosing (45). No data exist in

the literature for the binding of the newer analogues to the

components of human blood and if one accepts that only the free

fraction in blood is biologically active, then future studies of this

nature are clearly indicated.

Both doxorubicin and daunorubicin have been shown to rapidly

distribute into tissues and to different extents. There are also

significant differences in the tissue to plasma partition coefficients

among different tissues for each drug (46). Doxorubicin has a high

affinity for kidney and spleen but low affinity for adipose and muscle

tissue in rabbits and rats. Interspecies variations in tissue









distribution also exist (47). Determining factors in tissue

distribution could be membrane permeability, DNA affinity, and DNA

concentration. A fairly good correlation between tissue DNA

concentration and the tissue partition coefficient exists (47).

Further work has shown there is no significant difference in the

binding characteristics of doxorubicin to DNA from different tissues

(48). Therefore, it appears a predominant factor in tissue

partitioning is the DNA concentration in tissue which does vary among

tissues and species (48).



1.5 Review of Analytical Methods

1.5.1 Chromatography

More effective chemotherapy can be achieved with a greater

understanding of the pharmacology, mechanisms of action, metabolism,

and pharmacokinetics of drugs. Highly sensitive and selective methods

of analysis are required to study plasma, tissue and cellular levels

of the drugs and their metabolites.

The earliest analytical methods for the determination of

doxorubicin and daunorubicin used total extractable fluorescence from

plasma (49), urine (49), and tissue (50). Since the fluorescence of

the anthracyclines is due to the aglycone chromophore any metabolite

retaining this structure may have similar fluorescence and contribute

to the overall signal. Separation of metabolites by TLC followed by

fluorescent scanning of the TLC plate allowed quantitation with

detection limits as low as 2 ng/ml in plasma (51). Because of the

acidic nature of silica gel and the acid content in some solvent








mixtures, formation of aglycones has led to erroneous estimations by

TLC (52).

Radioimmunoassay has limited application because of its

variability and cross reactivity with metabolites (53). Using HPLC to

separate doxorubicin and its metabolite, Langone et al. (1) developed

a sensitive HPLC-RIA procedure for the determination of doxorubicin

and its metabolites in urine.

With the advent of HPLC, many analytical methods were developed

employing normal (Table 1-1) and reversed phase (Table 1-2)

chromatography. The methods of detection include fluorimetry (52, 54,

55-61), which is the most commonly used, ultraviolet (62, 63-67) and

electrochemistry (68, 69).

A normal phase system was developed by Hulhoven and Desager (62)

for the determination of daunorubicin and daunorubicinol in plasma

using a UV detector at 490 nm. The column packing was 5 an silica

(Zorbax Sil), and the mobile phase was methylene chloride-methanol-25%

ammonia water (90:9:0.1:0.8). Concentrations as low as 10 ng/ml could

be detected by this method (62).

Israel et al. (52) developed a method to determine doxorubicin in

plasma, urine, and bile with both normal and reversed phase systems

used in parallel to characterize the parent compound and its

metabolites. A fluorescence detector was used in both systems with

the excitation wavelength at 484 nm and emission at 550 nm. The

normal phase column was a Whatman Partisil-lO PAC, a cyanoaminoalkyl

derivative of silica gel. A two-component gradient mixture of

chloroform, methanol, acetic acid and water (850:150:50:15) and













Table 1-1.


Normal Phase HPLC Systems Used for the Analysis of
An thracycl ines.


Column Mobile Phase Method of Ref.
Detection


Zorbax Sil Methylene chloride-methanol Uv 62
5 um silica 25% ammonia water (90:9:0.1:0.8)

Whatman Gradient of chloroform-methanol Fluorescence 52
Partisil-lO-PAC Acetic acid-water (850:150:50:15)
cyanoaminoalkyl and chloroform

7 pm silica gel Chloroform-methanol-acetic acid Fluorescence 54
0.3 mM magnesium chloride water
solution (720:210:40:30)

Zorbax Sil Chloroform-isopropanol-acetic Fluorescence 71
5 pm silica acid-water-sodium acetate
buffer (pH 4.5) (100:100:14:14:1)

10 I1m Chloroform-methanol-acetic acid Fluorescence 72
aminocyano- water (80:20:2:3)
silica











Table 1-2.


Reversed Phase HPLC Systems for the Analysis of
Anthracyclines in Biological Fluids.


Column Mobile Phase Method of Ref.
Detection


C-18 ODS "Hi-Eff"
5 jim


LiChrosorb RP-8
5 Vm


U-Bondapak C-18



Partisil ODS
10 Pm

p-Bondapak phenyl




u-Bondapak phenyl




u-Bondapak phenyl


p-Bondapak C-18


p-Bondapak phenyl


Acetonitrile-0.01 M
phosphoric acid (pH 2.3)
(50:50)

Acetonitrile-water-0.1 M
phosphoric acid (pH 2.3)
(31:61:8)

Acetonitrile-0.1% ammonium
format buffer (pH 4.0)
(30:70)

Acetonitrile-0.03 M
phosphoric acid (40:60)

Gradient from 100% A (0.05 M
KH POh, pH 3) to 60% B
(65:35 acetonitrile, 0.05 M
(KH2PO4, pH 3)

Gradient from 70% A (ammonium
format buffer) and 30% B
(ammonium format buffer/THF
(1:1) to 0% A, 100% B

Acetonitrile-0.05 M KH2PO4
(35:65)

Acetonitrile-water-acetic
acid (28:71:1) adjusted to
pH 4 with 20% sodium acetate

Acetoni tri le-water-acetic
acid (27:72:1) adjusted to
pH 4.3 with 20% sodium acetate


Fluorescence



Fluorescence



Fluorescence



Fluorescence


Fl uorescence




Fluorescence




Fluorescence


Electro-
chemical


Electro-
chemical








chloroform was used. A gradient was also used with a reversed phase

P-Bondapak phenyl column starting with 30% acetonitrile and 70% pH 4

ammonium format buffer and ending with 35% acetonitrile and 65%

buffer after 5 minutes. The reversed phase system gave more rapid and

cleaner separations than the normal phase method. Typical problems of

normal phase chromatography were encountered: retention times varied

from day to day and re-equilibration was lengthy (52).

A similar normal phase system was developed by Baurain (54) using

a 7 um silica gel column and a mobile phase of chloroform, methanol,

acetic acid, and 0.3 mM magnesium chloride water solution

(720:210:40:30). Magnesium chloride is known to form complexes with

daunorubicin (70) and caused the retention time of doxorubicin to

decrease by a factor of three. Fluorescence detection was used with

excitation and emission wavelengths of 480 nm and 560 nm,

respectively. The lowest amount detectable was reported as 1.5 ng/ml

in biological fluids.

Shinozawa and Oda (71) used a normal phase system to determine

doxorubicin in lymph and gall. A Zorbax Sil 5 im stationary phase

with a mobile phase of chloroform-isopropanol-acetic acid-water-sodium

acetate buffer (pH 4.5) (100:100:14:14:1) was used. For fluorescence

detection, excitation and emission wavelengths were 470 nm and

585 nm. The limit of detection was 1 ng/ml.

Anthracycline concentrations in plasma and in vitro enzyme

studies were determined by Averbuch et al. (72) using a 10 pm amino-

cyanosilica column and a mobile phase of chloroform-methanol-acetic








acid-water (80:20:2:3). Fluorescence excitation and emission

wavelengths were 480 nm and 560 nm.

Many reversed phase systems have been developed for the analysis

of doxorubicin and daunorubicin in pharmaceutical preparations, body

fluids, and metabolite studies (Table 1-2). Ultraviolet and visible

light detection have been used to some extent but fluorescence is more

commonly used because of its greater sensitivity and selectivity.

Eksborg (63) used UV/VIS absorption to study the separation of

doxorubicin, daunorubicin, and their hydroxy metabolites on different

aliphatic side chain stationary phases with different organic

modifiers. LiChrosorb RP-2, RP-8 and RP-18 which are silica gels

reacted with dimethyldichlorosilane (RP-2), octylchlorosilane (RP-8)

and octadecyldichlorosilane (RP-18) were tested. Acetone,

acetonitrile, methanol, ethanol and isopropanol were used as organic

modifiers in dilute phosphoric acid. The retention of the solutes

increased with increasing length of the alkyl chains of the stationary

phase but the selectivity was not influenced to any great extent. Low

concentrations (20%) of acetonitrile gave the highest separation and

selectivity, but the shorter analysis times were obtained with 40 to

60% organic solvent in the mobile phase.

Eksborg and Ehrsson (64) used a RP-2 column with a mobile phase

of acetonitrile-water-0.1 M phosphoric water (25:65:10) to measure

plasma levels of daunorubicin and daunorubicinol using UV detection at

254 nm.

Other reversed phase systems which use UV detection are for the

determination of doxorubicin and daunorubicin in bulk pharmaceutical








preparations (65-67). The FDA accepted method for doxorubicin assay

as described in the Code of Federal Regulations uses a UV detector at

254 nm, a P-Bondapak C18 column, and acetonitrile-water (31:69) mobile

phase adjusted to pH 2.0 with phosphoric acid. The sample is

dissolved in an internal standard solution of 2 ng/ml 2-

naphthalenesulfonic acid in mobile phase. The accepted assay for

daunorubicin uses the same conditions with a mobile phase of

acetonitrile-water (38:62) adjusted to pH 2.2 with phosphoric acid

(65). Neither method allows simultaneous determination of doxorubicin

and daunorubicin and possible impurities such as the aglycones. It is

important to determine the potency of these drugs for accurate

dosage. Since the aglycones are mutagenic and have no antitumor

activity their content is also important. For these reasons Haneke et

al. (66) designed a system in which doxorubicin, daunorubicin, and

their aglycones could be quantified simultaneously. This system uses

254 nm UV detection, a P-Bondapak C18 column, and a mobile phase of

methanol-ammonium phosphate pH 4.0 (65:35) or another useful mobile

phase of methanol-0.005 M 1-heptanesulfonic acid pH 3.5 (62.5-37.5).

Relatively low levels of the aglycones (0.5-3.0% of the parent

compound) can be determined.

To determine maximum production yield, doxorubicin, daunorubicin,

and other anthracyclines present have been determined in fermentation

broth (37). The broth is extracted at pH 1.5 and the extract analyzed

on a u-Bondapak C18 column using acetonitrile-potassium monobasic

phosphate (7:18) adjusted to pH 3 with citric acid with UV detection








at 254 nm (67). Another mobile phase which has been used is methanol-

water (65:35 or 60:40) adjusted to pH 2.0 with phosphoric acid (73).

The most common application of HPLC analysis of the

anthracyclines is measurement of the drug and its metabolites in

biological fluids, usually plasma. Most reversed phase assays use

some type of C18 column and fluorescence detection which gives a limit

of detection ranging from 1 to 10 ng/ml. Mobile phases are composed

of 20 to 50% organic modifiers, usually acetonitrile or methanol and a

buffered aqueous solution at pH 2 to 4 (Table 1-2).

Doxorubicin and its metabolites in human plasma have been

measured with a C18 OOS "Hi-eff" 5 Um column and a mobile phase of

acetonitrile-0.01 M phosphoric acid pH 2.3 (50:50) (55). The same

compounds have been measured in rat plasma using a LiChrosorb RP-8 5

pm column and a mobile phase of acetonitrile-water-0.1 M phosphoric

acid pH 2.3 (31:61:8) (56).

Determination of 4'-deoxydoxorubicin and its metabolites has been

performed with a u-Bondapak C18 column and an acetonitrile-0.1%

ammonium format buffer pH 4.0 (30:70) mobile phase (57).

Epidoxorubicin and its metabolites have been measured using a Partisil

ODS 10 pm column with acetonitrile-0.03 M phosphoric acid (40:60)

(58). A gradient system has also been used with a U-Bondapak phenyl

column. The 25-minute linear gradient ran from 100% A (0.05 M

potassium monobasic phosphate, pH 3.0) to 60% B (65:35 acetonitrile,

0.05 M KH2PO4, pH 3.0). Identification of 4'-epidoxorubicin-a-

glucuronide conjugates was also possible with this system (59).








The u-Bondapak phenyl column with a gradient system has also been

used to determine doxorubicin, daunorubicin, and their metabolites.

The linear gradient ran from 70% A (ammonium format buffer) and 30% B

(ammonium format buffer/THF, 1:1) to 0% A, 100% in B in 10 minutes.

Eight metabolites of doxorubicin were separated and characterized

(60). Another method uses the u-Bondapak phenyl column to determine

plasma levels of 4-demethoxydaunorubicin and its 13-dihydro metabolite

with a mobile phase of acetonitrile-0.05 M KH2PO4 (35:65) (61).

All of the above mentioned reversed phase systems use

fluorescence detection. A method of analysis for daunorubicin and its

metabolites using electrochemical detection has been developed by

Akpofure et al. (68). Using an applied oxidative potential of +0.65

volts the limit of detection was 8 ng/ml in plasma. With the

fluorescence and electrochemical detectors in series, the amperometric

sensitivity was greater than that of fluorescence. A u-Bondapak C18

column was used with a mobile phase of acetonitrile-water-acetic acid

(28:71:1) adjusted to pH 4 with 20% sodium acetate. Kotake et al.

(69) developed a similar HPLC-EC method for doxorubicin and its

metabolite doxorubicinol with detection limits in plasma 2 ng/ml for

doxorubicin. The column was a u-Bondapak phenyl column, the mobile

phase was acetonitrile-water-acetic acid (27:72:1) adjusted to pH 4.3

with 20% sodium acetate, and the applied potential was +0.7 volts.

1.5.2 Sample Preparation

The extraction of doxorubicin, daunorubicin, and their hydroxyl

metabolites using an organic phase of chloroform and 1-pentanol (9:1)

has been studied by Eksborg (10). The optimum extraction pH ranged








from 8.0 to 8.5 depending on the dissociation constants of the

compound. The distribution ratio was affected by the degree of self-

association in the aqueous phase. Most extraction procedures adjust

the pH of the sample within this range and use a 5- to 10-fold excess

of organic solvent. Some of the common organic solvents used include

chloroform:heptanol (9:1), chloroform:isopropanol in ratios of 4:1,

2:1, or 1:1, methylene chloride:isopropanol (1:1), and

chloroform:methanol (9:1). Some procedures re-extract into 0.1 M

phosphoric acid to separate the aglycones but usually the organic

phase is separated, evaporated and reconstituted with mobile phase.

Other sample cleanup procedures involve the use of C18 cartridges

such as Bond-Elut (Analytichem International) or Sep-Pak (Waters).

Eksborg et al. (74) mixed the plasma sample with diatomeceous earth

which was placed in an empty column and eluted with chloroform:

heptanol (8:2). This was then extracted with phosphate buffer pH 2.2

which was injected into the analytical column.

The technique of column switching has been used in place of

extraction (75). The sample is injected onto a precolumn, which is

connected to the analytical column, where it is washed with water then

eluted onto the analytical column with mobile phase similar to those

described previously (68, 75).

1.5.3 Electrochemistry

The anthracyclines would appear good candidates for

electrochemical detection due to the presence of the two different

electroactive sites on the chromophore. Rao et al. (76) have studied

the cathodic electrochemistry of doxorubicin and several other









anthracyclines using DC polarography, cyclic voltammetry, differential

pulse polarography and chronopotentiometry in an attempt to relate

their electrochemistry to their antitumor activities. Two metabolic

pathways which involve reduction are the reduction at the 13-carbonyl

and reductive glycosidic cleavage. Reductive activity is also known

to cause free radical formation (18).

Two sets of reduction waves were reported, the first occurring at

approximately -0.6 V vs. SCE (saturated calomel electrode) and the

other around -1.2 V. The first wave involved the 2 electron, 2 proton

reduction of the quinone centers and could be separated into multiple

waves depending on pH and potential scan rate. The second set was

attributed to the irreversible reduction of the side chain carbonyl.

From studies of the daunorubicin aglycone, the quinone reduction

appeared not to be influenced by the presence of the sugar moiety or a

hydroxyl group at C-7 (77).

The total redox behavior of doxorubicin at a carbon paste

electrode was examined by Baldwin et al. (78). Cathodic behavior

similar to Rao's observations (76) was found as well as an apparently

reversible anodic and cathodic pair at approximately +0.5 V vs. SCE at

pH 4.5. A 1.0 x 10-8 M doxorubicin sample could be easily detected

using differential pulse polarography (DPP) at +0.5 V. Another system

using DPP was developed by Sternson and Thomas (79) with a dropping

mercury electrode. Using a reduction potential of -0.6 V, 400 ng/ml

of total doxorubicin species could be determined in untreated plasma

samples.













CHAPTER 2
RESEARCH OBJECTIVES


2.1 Analytical Method Development

Reversed phase high performance liquid chromatography (HPLC) has

been used extensively (55-61, 68, 69) for the analysis of

anthracycline antibiotics in plasma. Almost exclusively, HPLC methods

for the analysis of this group of drugs employ fluorescence detectors

which give practical limits of detection around 5 ng/ml in plasma.

Quantitation by fluorescence may be affected by light exposure and

possible alterations of the chromophore in biological systems.

Possible nonfluorescent metabolites are undetectable. In contrast to

fluorescence detectors, the use of electrochemical detectors in

combination with HPLC for the analysis of anthracycline antibiotics

has received much less attention (68, 69). This is despite the fact

that all the compounds in this class contain both oxidizable

(phenolic) and reducible (quinone) functional groups and are ideally

suited to electrochemical detection (68, 69, 76-79). The high

performance liquid chromatography with electrochemical detection

(HPLC-EC) for daunorubicin and doxorubicin have been described by

Akpofure et al. (68) and Kotake et al. (69), respectively. The limits

of detection were 8 ng/ml for daunorubicin (68) and 2 ng/ml for

doxorubicin (69). Some advantages of electrochemical detection which








have been cited are selectively based on electrode potential, low cost

and simplicity of operation (78).

Present studies are concerned with investigating the potential of

HPLC-EC for the analysis of anthracyclines as a class of compounds

using doxorubicin, daunorubicin and four of their analogues as

representative examples (Figure 1-1). The cyano analogue is of

special interest because it is the newest and least well defined,

analytically and pharmacologically. Preliminary investigations

concentrated on the analysis of the cyano isomers. Subsequently, the

studies were expanded to include analogues of doxorubicin and

daunorubicin in an attempt to develop more generally applicable

methodologies and to characterize their physicochemical and

biopharmaceutical properties. High sensitivity is especially

important for the cyano compound because of its increased potency

which reduces its projected dose and therapeutic plasma concentrations

compared with doxorubicin.

To optimize the HPLC-EC system, the chromatography of the

compounds was studied. Chromatographic conditions were examined to

determine the system that offers a rapid analysis time with a high

sensitivity and selectively separates each compound from other

interfering peaks arising from the biological matrix. Investigations

included the effect of mobile phase composition, pH, and temperature,

column length and column packing on the chromatographi-c behavior of

the compounds. Methods of sample preparation were examined for plasma

and urine.








2.2 Static and Flowing Electrochemistry

Amperometric and coulometric electrochemical detectors may be

used in a flowing system (e.g., HPLC). The electrochemical response,

detected as the signal and background current, is dependent on

temperature, flow rate, and mobile phase composition, pH, and ionic

strength (80). Comparison of an amperometric and coulometric detector

was performed under various chromatographic and detector conditions.

The optimum applied oxidative potential for both detectors was

determined from hydrovoltammograms. At the optimum applied potential,

chromatographic behavior of the compounds was observed at different

chromatographic conditions to optimize the HPLC-EC system.

Cyclic voltammetry is a static system where the electrochemical

response of a sample to a changing applied potential can be

observed. Both the oxidative and reductive behavior of the compounds

were observed over the pH range of 2 to 8. The results of these

studies were used to support the choice of an optimum HPLC-EC system.



2.3 Applications of Analytical Methods

The developed analytical methods were applied to investigations

of the compounds' behavior in in vivo and in vitro systems. The

extents of human plasma protein and erythrocyte binding were

investigated. In addition to its biological significance, erythrocyte

binding has been used as a measure of partition coefficients and

membrane permeability of drugs (18). Studies on the effect on

pharmacokinetic parameters of concurrent long-term doxorubicin and

cisplatin administration to pediatric patients with osteogenic sarcoma









were initiated. Previous pharmacokinetic studies have been restricted

to single dose investigations with one notable exception (27).

Doxorubicin was administered to patients with solid tumors for three

consecutive days and the treatment was repeated three times at 10-day

intervals (27). Each patient showed different patterns in the plasma

disappearance curves. Over the first three-day course of therapy the

general trend was for plasma levels to become progressively higher

each day and last longer while plasma clearance became slower. The

authors (27) attributed the higher plasma levels to the accumulation

of doxorubicin in tissues. Other factors such as changes in protein

binding, liver function, and metabolism could have played a role in

the observed behavior (34).

Results from the biological studies will give insight into the

relative biological distribution of the compounds. Although the

analogues are similar in many respects, differences found in their

physicochemical and biopharmaceutical properties will help to explain

their different biological activities. Correlations of pKas and

partition coefficients with biological activities have already been

shown (13). However, these parameters have not been related to the

binding of these drugs to blood components. Analysis of these latter

studies should give a greater insight into the pharmacological

activities of these compounds and aid in the optimization of dosage

regimens.













CHAPTER 3
ELECTROCHEMISTRY


3.1 Materials and Methods

3.1.1 Chemicals and Reagents

The doxorubicin was a gift from Dr. M.J. Williamson of Adria

Laboratories, Columbus, Ohio. The analogues of DOX and DNR (EPI,

DEOXY, DMDR and CN) were kindly provided by Dr. C.W. Young (Sloan-

Kettering Memorial Cancer Center, New York, NY) who obtained them from

Pharmatalia. The daunorubicin was purchased from Shands Hospital

Pharmacy, Gainesville, FL, in the form of 20 mg vials of Cerubidine

(Ives, New York, NY) which were reconstituted with 4 ml of deionized

water to give a working solution of 5 mg/ml. Stock solutions of the

drugs were prepared in water with the exception of CN which was

dissolved in methanol:water (50:50). HPLC grade organic solvents and

ACS grade KH2PO4 were obtained from Fisher Scientific, Fair Lawn,

NJ. All other solvents and chemicals were of reagent grade and

obtained from Fisher Scientific. Distilled deionized water was used

throughout. Silanized glassware was used for all solutions containing

the anthracycline antibiotics since our preliminary investigations

indicated extensive binding of the drugs to glass, especially at low

concentrations.








3.1.2 Cyclic Voltammetry

Cyclic voltammetry (CV) was performed on the compounds of

interest using a BioAnalytical Systems (BAS, West Lafayette, IN) CV-IA

cyclic voltammeter and a microcell (50 pl) assembly (BAS). The solute

concentrations were approximately 10-4 M in KH2PO4/H3PO4 or

KH2P04/NAOH (0.1 M, pH 2 to 8). Since response varied with supporting

electrolyte composition, the same buffer used in the chromatographic

system was employed in the cyclic voltammninetry experiments. The

working electrode was carbon paste, the counter electrode was platinum

wire and the reference electrode was Ag/AgCl. The recording system

was a digital voltammeter and a X-Y recorder, model RE0074 from

Princeton Applied Research, Princeton, NJ. Nitrogen purging was used

to degas the solutions prior to scanning in the reduction mode. Scan

rates of 25, 50, 100, 200, 500 and 1000 mV/s were used.

3.1.3 HPLC-EC

The experiments were performed using a liquid chromatograph

consisting of a Constametric IIG pump (LDC, Riviera Beach, FL), a

manual injection valve (Negretti and Zamba Aviation, Southampton, UK)

with a 20 ul loop, a SAS Hypersil (5 pm, 15 cm x 4.6 mm, Shandon

Southern, Sewickley, PA) analytical column, and a Fisher Recordall

5000 series strip chart recorder. The BAS electrochemical detector

system was a LC-4A amperometric controller which was connected to a 3-

electrode TL-5A thin layer flow cell. The working electrode was

glassy carbon and the reference electrode was Ag/AgCl. The ESA

Coulachem (Environmental Sciences Association, Inc., Bedford, MA)

system had dual coulometric detectors with a precolumn guard cell.








The aqueous portion of the mobile phase was filtered under vacuum

through a 0.45 Um Millipore filter and degassed by sonication prior to

use.



3.2 Results and Discussion

3.2.1 Cyclic Voltammetry

The microcell used for the CV studies was designed to allow

measurements with 50 ul of solution. The design of the microcell

differed from the standard cell type in that the sample solution is in

an electrode jacket with a Vicor frit at the end, similar to a

reference electrode, which separates the sample from the supporting

electrolyte solution. The working electrode is placed in the sample

solution. All of the electrodes are in close proximity to each

other. One possible difference in results from the microcell compared

to a standard type of cell is an increase in IR drop, which is a

factor dependent on the current and the resistance in the cell (81).

An increase in IR drop is evidenced by shifting of the anodic peaks

anodically and cathodic shifting of the cathodic peaks (81). This can

be minimized by increasing the ionic strength of the supporting

electrolyte outside the sample cell (82).

Although of small volume, the system may be considered to be

diffusion controlled with possibly less turbulence and a thicker

diffusion layer than the standard electrode type (82).- The anodic

peak in a positive potential scan (or the cathodic peak in a negative

scan) is the result of two competing factors. One factor is the

increase in rate of oxidation as the potential becomes more positive








and the other is the formation of a thickening depletion layer across

which reactant must diffuse. When the reactant concentration at the

electrode is small compared with the concentration farther away from

the electrode, the current is controlled by the rate of diffusion of

reactant through the depletion layer (83). A supporting electrolyte

is used to prevent migration currents and ensure conductivity (81).

To study the oxidative behavior of the compounds, a forward scan

from an initial potential of 0.0 V was performed to a switching

potential of +1.2 V and back to 0.0 V (Figures 3-1 and 3-2). At high

scan rates (greater than 200 mV/s) the IR drop shifted the anodic

peaks outside the scan range. At low pH (2 to 4) there was a

reversible anodic and cathodic pair (Figure 3-1). In the pH 6 to 7

range the cathodic peak began to decrease in height at slow scan rates

for DOX. At pH 8 the cathodic peaks were smaller for all compounds

(Figure 3-1). For example, at a scan rate of 25 mV/s the cathodic

peak for DOX was approximately one-tenth that of the anodic peak

(Figure 3-1). This would indicate that the reversibility of the

oxidative process is affected at higher pHs. The response of all the

compounds, determined by the peak heights, was greater at lower pH.

The electrochemical selectivity of each compound was

approximately the same over the pH range studied. For example, cyclic

voltammograms (Figure 3-2) of a positive potential scan were very

similar for each of the compounds at pH 4. At pH 2 the oxidative

half-wave potential was approximately +0.7 V for all compounds. The

half-wave potential was shifted cathodically with increasing pH. The








(a) (b)
0V 0.0
I.o I v IU I.O 0.0 v


/







Figure 3-1. Cyclic voltammograms of EPI (a) and DOX (b) from 0.0 to
+1.0 V. a) EPI at pH 2, b) DOX at pH 8, c) scan rate
25 mV/s. d) scale 0.79 pA/cm.




















































Figure 3-2.


Cyclic voltammograms of five anthracyclines from 0.0 to
+1.0 V at pH 4.








average half-wave potential was +0.6 V and +0.5 V around pH 4 and 6,

respectively.

The reductive behavior of the compounds was studied from an

initial potential of 0.0 V to -1.2 V and back. At approximately pH 2

an apparently reversible anodic and cathodic pair was observed which

was similar to that seen in the oxidative mode for all compounds

(Figure 3-3). At low pH the peak heights were greatest, almost twice

those at pH 8. The peak height of the CN compound did not increase

with scan rate to the extent of the other compounds.

Beginning around pH 4, at scan rates below 50 mV/s the cathodic

peak was a large broad peak but the anodic peak was very small (Figure

3-3). This indicates a quasi-reversible or irreversible reductive

reaction with possibly two indistinguishable cathodic processes

occurring. The same occurrence was noted in the pH 6 to 7 region. No

evidence of an anodic peak for CN and DMDR was observed below scan

rates of 100 mV/s. The cathodic peak of DMDR was broad as was the

anodic peak. This probably indicates that more than one

electrochemical reaction was occurring.

At pH 8 all the compounds showed a small cathodic shoulder at

approximately -0.5 V (Figure 3-3). The anodic peak appeared at faster

scan rates but was about half the peak height of the cathodic peak.

This again indicates some type of irreversible process.

The reductive half-wave potential was also approximately the same

for all compounds at a constant pH. The half-wave potential was

shifted cathodically with increasing pH from -0.35 V at low pH to

-0.6 V at pH 5 to 7, to -0.7 V to -0.75 V at pH 8.

















(a)


0 -lV


Figure 3-3.


Cyclic voltammograms of CN, DEOXY and DMDR- from 0.0 to
-1.0 V. a) CN, pH 2, scan rate 100 mV/s, scale
3.94 PA/cm; b) DEOXY, pH 4, scan rate 25 mV/s, scale
1.97 uA/cm; c) DMDR, pH 8, scan rate 100-mV/s, scale
3.94 PA/cm.


- 1 V









Since the redox behavior of all the compounds is similar, it is

reasonable to assume that the redox groups on the chromophore present

in all the compounds are responsible for the observed electrochemical

behavior of the compounds. The response of the oxidation and

reduction of the compounds based on anodic peak height and cathodic

peak height respectively were within *10 nA, indicating both modes

have comparable sensitivity. The signal to noise ratio (SNR) was

greater in the oxidative system at all pHs. The increased noise in

the reductive system was probably mostly due to the presence of

reducible dissolved oxygen. The results of these studies indicated

that the oxidation mode is preferable for HPLC since the SNR was

greater and no extensive measures are required to deoxygenate the

system.

3.2.2 HPLC-EC

Electrochemical detection is based on the oxidation or reduction

of a compound. In the BAS system the compound in the mobile phase

flows over the working electrode where approximately 10% of the sample

may be oxidized or reduced. This type of detection is called

amperometric (84). An ESA detector uses coulometric detection and as

the mobile phase flows through a porous electrode almost 100% of the

sample is oxidized or reduced (84).

To determine a suitable working oxidation potential for all the

compounds, a range of potentials from +0.1 V to +1.2 V was applied

using the BAS system. The mobile phase used was acetonitrile-

methanol-0.1 M KH2PO4 (pH 4.5) (28:10:62). Hydrovoltammograms were

constructed by plotting the detector response (peak height of each










































0.4 0.8 1.2
POTENTIAL (V)


Figure 3-4.


Hydrovoltammograms of five anthracyclines, showing peak
height as a function of applied potential. 0 = DEOXY,
0 = DOX, ( = EPI, = DMDR, A = CN (unresolved
isomers).















120-






E
E


w



L

o!


0/

0.1


Figure 3-5.


I I 19 1
0.3 0.5 0.7 0.6
POTENTIAL (V)


Comparison of hydrovoltammograms obtained by coulometric
(a) and amperometric (b) detection of DOX and DEOXY. Key
as in Figure 3-4.








compound) as a function of the applied potential (Figure 3-4). The

profiles of the hydrovoltammograms for each of the compounds were very

similar (Figure 3-4), indicating that it is the same functional group

which is oxidized and gives rise to the measured diffusion current.

This conclusion is supported by the cyclic voltammetry studies (Sec.

3.2.1) in which all the compounds behaved very similarly. For further

chromatographic analysis, an applied potential of +0.8 V was used for

the detection of all the compounds studied since this gave the highest

signal-to-noise ratio (SNR). It is reasonable to assume that it is

the oxidation of one or both of the aromatic hydroxyl groups on the

anthracycline moiety which is detected.

An ESA electrochemical detector was compared with the BAS

amperometric detector for the analysis of the anthracyclines (Figure

3-5). The ESA detector has a guard cell between the pump and the

injector which could be set at an applied potential slightly higher

than the working detector to oxidize any mobile phase components and

thereby reduce the background noise. The system has two working

electrodes in which the first one may be set at a low oxidizing

potential to further reduce noise by oxidizing interfering compounds

in the injected sample. It was found that increasing the potential of

the first detector decreased the response from the second detector.

The first detector can also be used for reduction of the sample before

it reaches the second electrode where it is oxidized thereby improving

sensitivity. In the case of DOX and DEOXY a reduction potential at

the first electrode caused a decrease in the response obtained at the

second electrode operated in the oxidation mode. This finding is








consistent with the results from the cyclic voltammetry studies (Sec.

3.2.1) which showed that the reduction of these compounds is not

always completely reversible.

Using the ESA detector and the same chromatographic system as

that used with the BAS detector, hydrovoltammograms for DOX and DEOXY

were obtained and compared with those from the BAS detector (Figure

3-5). Both compounds studied showed higher sensitivity at lower

applied potentials with the ESA detector. The applied potential is

compared with a proprietary reference electrode in the ESA detector

rather than with the Ag/AgCl electrode in the BAS detector. Thus, the

observed half-wave potentials were about 0.3 V lower for the ESA

detector compared with the BAS detector (Figure 3-5).

A working potential of +0.7 V on the ESA detector was used for

further comparison studies. The apparent limit of detection (LOD)

(SNR=2) for aqueous solutions of the compounds was about five times

lower with the ESA detector compared with the BAS detector. This was

mainly due to a reduced noise level obtained by using a 10 second time

constant on the ESA detector. However, the higher sensitivity also

gave a greater response to plasma constituents and a large solvent

peak which interfered with early eluting drug peaks (e.g., DOX)

increasing the LOD considerably. After comparison of the results from

the two detectors, the amperometric (BAS) detector was chosen as the

most suitable one for determination of the anthracyclines in

biological samples.













CHAPTER 4
ANALYTICAL METHOD DEVELOPMENT


4.1 Materials and Methods

The chemicals and reagents used were described in Section 3.1.1.

Initial chromatographic studies of CN were performed with a Model

RR/035 HPLC pump (HPLC Technology, Southampton, UK), a Rheodyne

injection valve fitted with a 20 il loop, a fixed wavelength (254 nm)

UV detector (LDC, Riviera Beach, FL), and a Fisher Recordall 5000

series strip chart recorder. Subsequently, the investigations were

expanded to include all the compounds described in Figure 1-1.

Analytical columns were 15 cm x 4.6 mm packed by an upward slurry

technique (85) with 5 um Zorbax ODS (DuPont, Wilmington, DE), 5 Pm ODS

Hypersil and 5 um SAS Hypersil (Shandon Southern, Sewickley, PA). A

HPLC water jacket (Alltech, Deerfield, IL) was used in conjunction

with a circulating water bath (Precision Scientific Co., Chicago, IL)

to control the column temperature in the studies involving

investigation of temperature effects. The aqueous portion of each

mobile phase was filtered under vacuum through a 0.45 vm Millipore

filter and degassed by sonication prior to use.

Preliminary sample preparation studies used a liquid

chromatograph consisting of a Constametric IIG pump (LDC, Riviera

Beach, FL), a manual injection valve (Negretti and Zamba Aviation,

Southampton, UK) equipped with a 20 ui loop, a Zorbax ODS (5 inm,








14 cm x 4.6 mm) analytical column, a BioAnalytical Systems Model LC-4A

amperometric detector (BAS, West Lafayette, IN) and a Fisher Recordall

5000 series strip chart recorder. The electrochemical detector was

equipped with a TL-5A thin layer flow cell. The working electrode was

glassy carbon and the reference electrode was Ag/AgCl. The mobile

phase used was acetonitrile-methanol-0.1 M KH2PO4 (pH 4.5) (28:10:58)

with flow rates varying from 1.0 to 2.5 ml/min depending on the

compound studied.

For the extraction of the compounds from buffer solutions, a

stock solution of the compound was added to 1 ml of 0.1 M borate

buffer (pH 8 to 9) in a 15 ml Teflon lined screw capped test tube.

For plasma extraction, a stock solution of the compound was added to

1 ml of plasma in the same type test tube to which 0.1 ml of borate

buffer (0.5 M, pH 9.1) was added to adjust the pH. Ten milliliters of

chloroform:isopropanol (1:1) were added to each sample which was

vortexed for 30 seconds and centrifuged at 2000 rpm for 5 minutes.

After aspiration of the aqueous layer, 5 ml of the organic layer were

transferred to a second 15 ml test tube and evaporated to dryness

under nitrogen at 40C. The residue was redissolved in 0.5 ml of

mobile phase and 20 ul were injected onto the analytical column.

The optimized HPLC system was comprised of a Constametric IIG

pump (LDC, Riviera Beach, FL), a manual injection valve (Negretti and

Zamba Aviation, Southampton, UK), fitted with a 200 uil- loop, a SSI

high pressure filter (Rainin Instruments, Woburn, MA) positioned

between the injection valve and the column, a Rheodyne 3-way slider

valve positioned between the column and the detector, a BAS Model








LC-4A amperometric detector (described previously), and a Fisher

Recordall 5000 series strip chart recorder. Two analytical columns,

an ODS Hypersil (3 um, 7.5 cmx 4.6 mm) and a Zorbax ODS (5 Pm, 7.5 cm

x 4.6 mm) were packed using the upward slurry technique (85).

The sample preparation procedure ultimately developed for plasma

was as follows: an appropriate volume of internal standard solution

(Table 4-1) was added to 1 ml of plasma in a 15 ml Teflon lined screw

capped test tube. For the analysis of DOX, EPI, and DMDR the pH of

the plasma was adjusted to 8.4 by the addition of 100 ul of borate

buffer (0.1 M, pH 9.1). For the analysis of DEOXY and CN the pH was

adjusted to 8.6 by the addition of 120 ul of borate buffer. Ten

milliliters of chloroform were added to each tube which was then

gently shaken for 10 minutes. After centrifugation at 2000 rpm for 10

minutes, the aqueous layer was aspirated. The chloroform layer was

transferred to a second 15 ml test tube and the solvent evaporated to

dryness at 40C under a gentle stream of nitrogen. Residues

containing DOX, EPI, or DEOXY were redissolved in 210 ul of 0.2 M

H3PO4 (pH 2.0) and shaken for 30 seconds with 1 ml of chloroform.

Residues containing DMDR or CN were redissolved in 210 ul of 0.1 M

KH2PO4 (pH 4.5) and shaken for 30 seconds with 1 ml of hexane. After

centrifugation at 2000 rpm for 2 minutes, 200 il of the aqueous phase

were injected onto the HPLC column.

A similar procedure was used for urine samples of DOX where an

appropriate volume of borate buffer (0.2 M, pH 9.2) was added to 1 ml

of urine to adjust the pH to 8.4. After extraction, residues were












Table 4-1.



Compound


DOX


EPI


DEOXY


DMDR


CN


Summary of Chromatographic Conditions
Anthracycline Antibiotics by HPLC-EC.


Internal Stationaryb Flow Rate
Standard Phase (ml/min)


EPI Zorbax ODS 0.7
(5 Um)

DOX Zorbax ODS 0.7
(5 i1m)

DNR Zorbax ODS 1.1
(5 um)

DNR Zorbax ODS 1.6
(5 um)

DNR ODS Hypersil 0.8
(3 pm)


for the Analysis of


k'c
Analyte


4.62


6.23


7.68


15.02


14.66d
16.67d


Internal
Standard


6.23


4.62


9.28


9.28


9.85


a See Figure 1-1 for key.

b Column dimensions: 4.6 x 75 mm. Mobile phase:
acetonitrile:isopropanol:0.1 M KH2PO4 (pH 4.5) (25:3:72)
c k' = (tr-to)/to, where tr and to are the retention times of the
solute and an unretained compound, respectively.
d Separated diastereomers.









redissolved in 1 ml of 0.2 M H3P04 (pH 2.0), shaken with 1 ml of

chloroform, and centrifuged for 2 minutes. Two hundred microliters of

the aqueous phase were injected onto the HPLC column.


4.2 Results and Discussion

4.2.1 HPLC

An investigation of the chromatography of the anthracyclines

began with the cyano analogue (CN) as a model compound because it has

been the least studied to date. Since the electrochemistry of the

compounds had not been investigated at the time of these studies, UV

detection was used. The effect of organic modifier composition on the

retention of CN by a Zorbax ODS, an ODS Hypersil and a SAS Hypersil

column was studied (Tables 4-2 and 4-3). The SAS column was found to

be less hydrophobic than the ODS columns and compounds eluted more

rapidly (Table 4-3). The organic modifiers chosen were acetonitrile,

methanol, isopropanol and tetrahydrofuran (THF). Various percentages

of each were mixed with 0.05 M KH2PO4 (pH 4.5) and flow rates of 1 or

2 ml/min were used. In addition, an ODS Hypersil column was compared

with the Zorbax ODS column using a mobile phase of methanol-0.05 M

KH2PO4 (Table 4-2).

The CN isomers were separated on the ODS columns (Table 4-2) but

not on the SAS column (Table 4-3). At a constant concentration of

organic modifier, isopropanol produced the shortest retention times

followed by THF, acetonitrile, and methanol. Comparisons of

acetonitrile, isopropanol, and THF at 35% organic modifier showed that

isopropanol produced the highest selectivity and resolution values.









Table 4-2.


Chromatographic Parameters of CN on Zorbax ODS (a) and ODS
Hypersil (b) Columns.


Mobile Phase Parametersc

to tri tr2 k'1 k'2 Rs
(min) (min) (min)


(a) acetonitrile/0.05 M
KH2P04

40/60
35/65
30/70

(a) THF/0.05 M KH2PO4


35/65
25/75


(a) isopropanol/0.05 M
KH2PO4

35/65
25/75

(a) methanol/0.05 M
KH2PO4


70/30
65/35
60/40
55/45
50/50


1.08
1.08
1.08


4.33
9.45
22.24


4.72
10.04
24.02


3.0
7.75
19.59


3.37
8.30
21.24


1.12
1.07
1.08


<0.5
0.5
0.75


0.98 3.74 4.33 2.82 3.42 1.21 1.0
0.98 14.37 16.73 13.66 16.07 1.18 1.33


0.98 2.17 2.95 1.21 2.01 1.66 1.33
1.18 3.15 5.12 1.67 3.34 2.00 2.5


0.98
0.98
0.98
0.98
0.98


5.12
5.71
10.04
12.80
36.02


5.91
7.09
12.80
16.54
47.83


4.22
4.83
9.24
12.06
35.76


5.03
6.23
12.06
15.88
47.81


1.19
1.29
1.37
1.32
1.34


0.80
1.40
2.0
2.71
6.00


(b) methanol/0.05 M
KH2PO4

70/30
65/35
60/40
55/45


1.57
1.57
1.57
1.57


3.54
5.12
7.87
9.25


4.33
6.30
10.34
11.81


1.25
2.25
4.01
4.89


1.76
3.01
5.52
6.52


1.41
1.38
1.33
1.33


1.0
1.2
1.71
2.17


c to = elution time of unretained compound (water); trl and tr2 =
retention times of first and second isomers of CN, respectively; k'
= (tr to)/to; a = k'2/k'l; Rs = (tr2 trl)/w1, where w, is width
of CN1 peak at base (estimated when Rs<1.0).











Table 4-3. Chromatographic Parameters of CN on SAS Hypersil Column.


Mobile Phase Parametersa

to tr k' a
(min) (min)


Acetonitrile/0.05 M
KH2PO4

40/60
30/70
25/75

Methanol/0.05 M
KH2P04

70/30
50/50
45/55

Isopropanol/0.05 M
KH2PO4

35/65
30/70
25/75


THF/0.05 M
KH2PO4

35/65
30/70

25/75


0.59
1.48
1.18


1.18
1.18
1.18


1.18
1.18
1.08


0.59
0.89

0.79


2.26
15.94
27.36


2.95
10.24
25.00


2.17
2.85
5.12


2.76
7.58
8.27
11.22
12.40


2.83
9.8
22.19


1.5
7.68
20.19


0.83
1.42
3.74


3.68
7.52
8.29
13.20
14.70


0.00
0.00
0.00


0.00
0.00
0.00


0.00
0.00
0.00


0.00
1.10

1.11


a See Table 4-2 for explanation of parameters.









The resolution of the two isomers of CN was dependent on the nature

and the concentration of the organic modifier (Table 4-2). Comparison

of the ODS Hypersil with the Zorbax ODS column using methanol mobile

phases showed that the Hypersil gave smaller capacity factors (k'),

slightly less resolution and little difference in selectivity.

The results from the SAS column are presented in Table 4-3. The

CN isomers were not separated on the SAS column with mobile phases

containing methanol, acetonitrile or isopropanol and eluted as a

single peak. Partial resolution was obtained with THF mobile

phases. The organic modifiers produced chromatographic properties on

the SAS Hypersil column similar to those on the Zorbax ODS column with

the shortest retention times produced by isopropanol followed by THF,

methanol, and acetonitrile. In general the capacity factors on the

SAS column were smaller than those obtained on the Zorbax ODS column

(Tables 4-2 and 4-3). The SAS column would be preferable for

determination of total CN since the sensitivity would double that

achievable if the individual isomers were being quantified (the peak

heights of the two isomers were approximately the same). The

determination of the individual isomers of CN would best be performed

on an ODS column.

The effect of mobile phase pH on the separation of the CN isomers

on the SAS Hypersil column was investigated using different buffers as

the aqueous component of the mobile phase. Sulfuric acid was used to

attain pH 2, acetate buffer for pH 3.5 and phosphate buffer for pH

7.5. The pH was varied in systems of 35% acetonitrile and 30% THF and

the results are given in Table 4-4. The general trend for both













Table 4-4.


Effect of pH on the Chromatographic Parameters of CN on a
SAS Hypersil Column.


Mobile Phase Parametersa

to tr k'
(min) (min)

35/65 acetonitrile/buffer

pH 2 1.77 7.28 3.11 0.00
pH 3.5 2.07 9.45 3.57 0.00
pH 4.5 1.08 6.30 4.83 0.00
pH 7.5 1.77 10.04 4.67 0.00

30/70 THF/buffer

pH 2 1.77 8.27 3.67 1.10
9.06 4.02 1
pH 3.5 1.77 7.87 3.45 1
8.86 4.01 1
pH 4.5 0.79 7.58 8.59 1.10
8.27 9.47

a See Table 4-2 for explanation of parameters.








organic systems was an increase in k' with increasing pH. In

acetonitrile the peak height was reduced at pH 3.5 and 7.5. Very poor

peak shapes and decreased peak heights occurred in THF at pH 2 and 3.5

and no peak was observed at pH 7.5. It was later found that CN is

unstable below pH 4 which probably contributed to the apparent

decrease in sensitivity. Therefore, subsequent studies were conducted

with a mobile phase of pH 4.5.

Further studies involved the addition of ionic surfactants to the

mobile phase to manipulate retention. In ion pair chromatography, the

concentration of surfactant is usually kept below the critical micelle

concentration (CMC) (86). Above the CMC the surfactant molecules

aggregate to form micelles which solubilize the analytes in the mobile

phase (87). Increasing the surfactant concentration above the CMC

increases the micelle concentration but the free surfactant

concentration remains almost constant and equal to the CMC. Thus,

retention and selectivity can be controlled by manipulating the

concentration of micelles (86). Whereas increasing the organic

modifier concentration decreases solute retention, increasing the

concentration of ionic surfactant may result in an increase, decrease,

or no change in retention (80). Although addition of a sufficient

concentration of organic modifier to a micellar mobile phase will

disrupt the micellar environment and influence the separation

mechanism, retention of solutes is still influenced by the type and

the concentration of the surfactant in the mobile phase (86).

Sodium dodecyl sulfate (SDS) was added in increasing
concentrations from 10-4 M to 10-1 M to water and 0.01 M KHPO4 which
concentrations from 10~ M to 10 M to water and 0.01 M KH2PO4 which








were used as the aqueous components of the mobile phase. Two organic

modifiers, 35% acetonitrile and 25% THF were studied with the SAS

column and the results are shown in Figure 4-1. Increasing the SOS

concentration above 10-2 M markedly decreased retention. The

chromatograms in Figure 4-2 show that better peak shape was obtained

at higher SDS concentrations in the acetonitrile system. In the THF

system the split peaks fused at 10-1 M SDS (Figure 4-2). Similarly,

on the Zorbax ODS column in 15:85 acetonitrile-10-1 M SDS in 0.01 M

KH2PO4 the isomers were not separated. Apparently ion pairing effects

and/or incorporation of the solutes within the micelles resulted in

the two isomers of CN eluting unseparated. Micellar chromatography

generally gave broad peaks. Consequently the effect of increasing

temperature was studied in an attempt to improve the chromatography.

Temperature is known to have an effect on peak shape and

retention which generally decreases with increasing temperature

(88). Since mass transfer of the solute between the mobile phase and

stationary phase is a kinetically controlled process, increasing the

temperature of the system also increases column efficiency (86).

The effect of temperature was studied using 10:90 acetonitrile-

10-1 M SDS in 0.01 M KH2PO4 on the SAS column. A mobile phase of

15:85 acetonitrile-lO-1 M SDS in 0.01 M KH2PO4 was used with the

Zorbax column. Some typical chromatograms are shown in Figure 4-3.

The two isomers of CN were partially resolved at temperatures greater

than 45C in the acetonitrile-containing mobile phases on the SAS

column. On the Zorbax ODS column peak shape was slightly improved

with increasing temperature. In general retention decreased with













































-4 -3 -2 -1
LOG [SDS]


Figure 4-1.


Effect of adding sodium dodecyl sulfate (SDS) to the
mobile phase on the retention of CN (k'). CN isomers
were not resolved. Column: SAS Hypersil. Mobile
phases: 0--Oacetonitrile (MeCN)/SDS in water (35/65);
*-- MeCN/SOS in 0.01 M1 KH2PO4 (35/65);[J-0 THF/SDS in
water (25/75);H-- THF/SDS in 0.01 M KH2PO4 (25/75).
















































Figure 4-2.


Effect of adding sodium dodecyl sulfate (SDS) on the
chromatography of CN. Column: SAS Hypersil. Mobile
phases: a) MeCN/water (35/65); b) MeCN/O.1 M SDS in
water (35/65); c) THF/water (25/75); d) THF/O.1 m SDS in
water (35/65).







































0 5 10
mins


b






I f\


0 5 10


Figure 4-3.


Effect of temperature on the micellar chromatography of
CN.


Column


SAS Hypersil
SAS Hypersil
Zorbax ODS
Zorbax ODS


Mobile Phase


MeCN/U.i1
MeCN/0.1
MeCN/0.1
MeCN/0.1


SUS
SDS
SOS
SDS


1(U/9U)
(10/90)
(15/85)
(15/85)


Temperature(C)


Key








increasing temperature indicating that the partitioning process

between the two phases is exothermic. Although micellar chromato-

graphy of CN produced some interesting observations, it did not appear

to offer any advantages over more conventional reversed phase

techniques for the analysis of CN and related compounds.

The chromatographic conditions studied using UV detection showed

that the retention time of CN can be decreased by increasing the

organic content of the mobile phase, adding SOS, increasing the

temperature or reducing the pH of the aqueous component in the mobile

phase. The other compounds were expected to show similar behavior.

The influence of these factors and column type on resolution, peak

shape and analysis time was used to optimize the HPLC of the

anthracycline antibiotics using EC detection.

Using the electrochemical detector, binary mobile phases of

methanol-phosphate buffer (0.05 M, pH 4.5) or acetonitrile-phosphate

buffer were found suitable for the chromatography of the compounds of

interest. However, acetonitrile was preferred to methanol since it

could be used at lower concentrations which resulted in lower

background currents and less noise from the electrochemical

detector. Addition of a small percentage of methanol to the

acetonitrile-phosphate mobile phase aided in separation of the CN

isomers peaks.

The influence of ionic strength on electrochemical response was

evident. Phosphate buffer (pH 4.5) in the mobile phase at

concentrations less than 0.1 M caused decreased peak height and

increased the width of the peaks. In addition, the nature of the








buffer added to the mobile phase had an effect on chromatographic

behavior. For example, replacement of 0.1 M KH2PO4 with a 0.1 M

acetate buffer of the same pH gave similar retention characteristics

but very broad peaks. Results from the cyclic voltanmmetry studies

(Sec. 3.2.1) showed the compounds had a higher sensitivity at low pHs

(pH 2 to 4). Since the CN isomers are unstable below pH 4 and the

chromatographic studies had shown poor chromatographic behavior at

higher pH (7.5), a pH of 4.5 was determined optimal for all the

compounds. Therefore, preliminary studies were performed using a

mobile phase of acetonitrile-methanol-O.1 M KH2PO4 (pH 4.5)

(28:10:58). Figure 4-4 shows chromatograms of the compounds using

this mobile phase. Capacity factors of the compounds ranged between 2

for DOX and 14 for CN2. Slight manipulations of flow rate and mobile

phase composition (pH, organic modifier concentration) were needed for

analysis of plasma samples of some compounds to eliminate interference

from endogenous substances and eventually it was possible to obtain a

single set of chromatographic conditions suitable for determination of

all the compounds in plasma despite their wide range in hydrophobici-

ties (Figure 4-5). The chromatography of the anthracycline

antibiotics was optimized in terms of resolution from interfering

endogenous substances, peak shape, and overall analysis time. It was

found that the inclusion of a small concentration of isopropanol (3%)

instead of methanol further reduced the noise and also produced

sharper peaks. Thus the optimum mobile phase was found to be

acetonitrile-isopropanol-0.1 M KH2PO4 (pH 4.5) in the ratios 25:3:72

(Figure 4-5).










DOX


EPI


DEOXY





CN1 CN2


DMDR


-C. I ,I I

5 10 5 10
mins


Figure 4-4.


Chromatograms of six anthracyclines using the same
chromatographic conditions. Column: Zorbax ODS. Mobile
phase: MeCN/methanol/O.1 M KH2PO4 (28:10:58). Flow
rate: 2.5 ml/mmin. Detection: Amperometric at +0.8 V.






































.LJ ins
-- 10 mins


Figure 4-5.


Optimized separation of seven anthracyclines from
plasma. See Table 4-1 for conditions. Solute
concentrations: EPI--20 ng/rnl, DOX--5 ng/ml, DNR--
20 ng/ml, DMDR--25 ng/ml, OEOXY--12 ng/ml, CN (total)--
25 ng/ml.








Two short columns (7.5 cm), ODS Hypersil and Zorbax ODS, were

investigated and were found to give virtually identical retention

characteristics. However, the 3 Pm 00S Hypersil column gave slightly

narrower and more symmetrical peaks than the 5 uim Zorbax ODS column.

The use of the short columns permitted the rapid analysis of the

compounds of interest, with the exception of CN, at relatively low

back pressures (500 to 1000 psi). Changing from the more conventional

15 cm columns to columns of 7.5 cm in length permitted the analysis

times to be halved and the sensitivity to be doubled. A summary of

chromatographic conditions used for each anthracycline under

investigation is given in Table 4-1 and chromatograms of each compound

from plasma extracts under these chromatograms conditions are shown in

Figure 4-5. In the case of CN, an analysis time of 20 minutes was

required for the separation of its two diastereomers (Figure 4-5).

The analysis time for CN could be reduced by increasing the organic

modifier content of the mobile phase or increasing the flow rate.

However, this could only be achieved at the expense of incomplete

resolution of its diastereomers.

The retention of the anthracycline antibiotics was found to be

related to solute hydrophobicity and it was possible to calculate some

of the contributions of the functional groups, T (log (k'2/k'1) (89)

to retention. The value of T for the alcoholic OH at position 14

(-0.30) is obtained from the retention data for DOX and DNR and the

contribution (T=-0.18) of the same functional group at the 4' position

on the amino sugar is obtained from the data for DOX and DEOXY (Table

4-4). Interestingly, the two epimers of doxorubicin were well








separated (T=0.13) so that one could be used as an internal standard

for the other. The contribution of the methoxy group (T=-0.21) in the

4 position is obtained from the retention data for DMDR and DNR.

Replacement of the 3'-amino group by the hydrophobic 3-cyanomorpholino

group caused a tremendous increase in retention. The two isomers of

CN were separated indicating a difference in the hydrophobic

contributions of the cyanomorpholino groups to retention (T=O.50 and

T=0.56 for the two isomers). There appeared to be a qualitative

relationship between retention and solute hydrophobicity and this was

utilized later (Sec. 5.2.2 and 5.2.3) in the description of the

structure activity relationships of the anthracyclines.

The coating of the glassy carbon electrode by irreversibly bound

plasma constituents was minimized by the use of a 3-way slider valve

positioned between the column and the flow cell (Figure 4-6). The

material eluting before the peaks of interest was vented to waste.

This significantly reduced the frequency with which the surface of the

working electrode had to be regenerated by polishing. It also

prevented excessive contamination of the mobile phase which was

generally recycled. Recycling of the mobile phase resulted in lowered

background current and detector noise.

4.2.2 Sample Preparation

To determine the best extraction procedure for plasma the

extraction of CN from buffer was initially studied. Extraction

efficiency was found to be dependent on the pH of the buffer or

plasma, the organic solvent and volume ratio of organic solvent(s).

Preliminary experiments were performed on CN which was extracted from























RESERVOIR


WASTE


3 WAY
VALVE


Figure 4-6.


Diagram of chromatographic system. Three-way slider
valve was used to divert early eluting, endogenous peaks
to waste. At other times, the mobile phase was recycled.


SAMPLE








1 ml buffer solutions ranging from pH 3 to pH 11 in 1 pH unit

increments with 10 ml of chloroform:isopropanol (1:1). The layers

were vortexed, centrifuged and the aqueous layer removed and the

organic layer evaporated with nitrogen at 40C. The final residue was

reconstituted with 1 ml of mobile phase. It was found that CN was

unstable below pH 4 resulting in low recovery below pH 3. The highest

recovery was obtained around pH 9.

Extraction of DOX from plasma has been reported to be optimum at

pH 8.6 and DMDR at pH 8.4 (63) using chloroform:isopropanol (1:1) in

an aqueous/organic phase volume ratio of 1:10. Since the

anthracyclines are amphoteric their extractability was expected to be

highly dependent on pH. Comparison of the percent recovery based on

known concentrations in mobile phase from 1 ml of buffer extracted

with 10 ml of chloroform:isopropanol (1:1) for the compounds buffered

at pH 8.1, 8.4, 8.6 and 9 with borate buffer (0.1 M) showed small

differences (Table 4-5). A pH of 8.6 was chosen as a good compromise

for most compounds and was selected for further recovery experiments.

To determine the optimum extraction conditions from pH 8.6

buffer, different volume ratios of the organic solvents were

investigated. The results are given in Table 4-6 which shows that the

highest extraction from 1 ml of plasma was obtained with 10 ml of

chloroform:isopropanol (1:1). Extraction with chloroform:butanol

resulted in good recovery (71%) but poor chromatography.

Preliminary studies using plasma concentrations of the compounds

ranging from 0.5 to 2 ug/ml required adjustment of the mobile phase

conditions to separate the compounds from the many endogenous








Table 4-5.


Recovery of Six Anthracyclines from Borate Buffer (1
ml)(pH 8.1-9.0) Extracted with 10 ml of Chloroform:
Isopropanol (1:1).


Compound % Recovery (SD)(n=3)

pH 8.1 pH 8.4 pH 8.6 pH 9.0

DOX NDa 88 (0) 86 (8) 100 (11)

EPI 80 (1) 62 (3) 85 (7) ND

DEOXY 72 (1) 73 (4) 83 (3) 76 (2)

DMDR 82 (17) 91 (15) 90 (6) ND

CN1 ND 83 (13) 79 (0) 76 (5)

CN2 91 (13) 88 (9) 78 (3)

a ND = not determined.







Table 4-6. Recovery of DEOXY from 1 ml of Borate Buffer (pH 8.6) with
Different Organic Solvent Mixtures.


Extracting Solvent % Recovery (SD)
(n=3)
Composition Volume Extraction
Ratio Volume (ml)

Chloroform:2-propanol 1:1 5 74 (3.4)
Chloroform:2-propanol 4:1 5 59 (10.4)
Chloroform:2-propanol 4:1 10 71 (2.9)
Chloroform:2-propanol 1:1 10 83 (3.0)
Chloroform:n-butanol 9:1 10 71 (18.8)











Table 4-7. Chromatographic Conditions Used for Buffer and Plasma Extraction with
Chloroform:Isopropanol (1:1).

Compound Buffer Extraction Plasma Extraction

Mobile Phase Flow Rate k' Mobile Phase Flow Rate k'
Actn/Meth/O.1 M (ml/min) Actn/Meth/0.1 M (ml/min)
KH2P04, pH 4.5 KH2PO4

DOX 28/10/58 2.0 3.3 28/10/62 1.5 3.60
(pH 3)

EPI 28/10/58 1.5 3.60 28/10/58 1.5 3.60
(pH 4.5)

DEOXY 28/10/58 2.0 5.00 28/10/62 2.5 7.00
(pH 3)

DMDR 28/10/58 2.5 11.33 28/10/62 2.5 19.33
(pH 3)

CN' 28/10/58 2.5 13.67 28/10/58 2.5 13.67
15.33 (pH 4.5) 15.33








substances also extracted (Table 4-7). Recovery from buffer and

plasma was greater than 80% for all the compounds at 0.5 ug/ml using

chloroform:isopropanol (1:1) (Table 4-8). As the chromatographic

system was improved to allow determinations of lower concentrations,

higher sensitivity was obtained for the compounds as well as the

interfering endogenous substances. This necessitated further

modifications in sample preparation to eliminate interfering peaks.

It was found that removal of the isopropanol from the organic phase

for extraction significantly reduced the number of endogenous

components co-extracted, but also decreased the absolute recoveries of

the compounds (Sec. 4.2.3) from greater than 80% to 60 to 80%.

Another method of plasma extraction which was investigated

involved the use of solid-liquid extraction techniques such as the

Sep-Pak (Waters) C18 cartridges. The Sep-Pak method has been used

with an LC technique employing fluorescence detection (38, 59) which

may or may not detect interfering plasma constituents that are

detected by electrochemical means. Many procedures employing the

Sep-Pak were attempted but resulted in decreased recovery compared

with the liquid-liquid extraction technique and excessive interfering

peaks. Therefore, this method was not pursued further and

improvements in the liquid-liquid extraction procedure were continued.

Selective isolation of the anthracyclines from plasma was

achieved by an alkaline extraction into chloroform, followed by back

extraction into an acidic aqueous solution. The remaining

interference were removed by shaking the final aqueous aliquot with

an organic solvent (n-hexane or chloroform). The highest recovery of















Table 4-8.


Recoveries of Six Anthracyclines (500 ng/ml) from Buffer
and Plasma (1 ml) Extracted with 10 ml of Chloroform:
Isopropanol (1:1).


Compound % Recovery SD RSDa % Recovery SD RSDa
from buffer (n=5) (%) from plasma (n=5) (%)


DOX 100.8 8.0 9.0 96.8 3.0 2.1

EPI 87.9 8.9 7.4 86.8 6.8 14.8

DEOXY 85.7 9.5 11.8 81.2 3.2 4.0

DMDR 89.7 8.1 7.0 92.6 5.4 4.1

CN1 92.0 6.0 4.7 96.2 3.7 5.4

CN2 101.0 2.9 3.4 95.4 0.8 1.7

a Relative standard deviation, RSD(%) = (SD/Recovery) x 100.








DOX, EPI and DMDR into chloroform was achieved at a pH of 8.4. A pH

of 8.6 was preferred for the extraction of DEOXY and CN. Removal of

interference from acidified extracts containing DOX, EPI, DEOXY, or

DMDR could be achieved at pH 2.0 by shaking with chloroform. However,

these conditions were unsuitable for the clean-up of the more

hydrophobic CN which partitioned into the chloroform, resulting in low

recoveries. Clean-up of CN was achieved by shaking with n-hexane at

pH 4.5 which was also a suitable method for DMDR. In addition, a pH

greater than 4.0 was needed for the clean-up of CN because both

diastereomers were found to be unstable at pH 2.0.

4.2.3 Assay Validation

Using the optimized HPLC-EC system, detection limits of 200 pg

are possible (Table 4-9). This permitted the anthracyclines to be

quantified in plasma at concentrations as low as 1 to 2 ng/ml (Table

4-9) which is equal to or lower than the limits of quantification by

HPLC using fluorescence detection (57, 59, 61). Optimum detector

conditions are essential for detection of ng/ml concentrations. These

include a clean working electrode surface obtained by polishing and a

reference electrode producing a constant reference potential.

Reference electrodes were replaced approximately every two months.

For each drug studied, two calibration curves were constructed,

one spanning a high concentration range of drug (0 to 4.0 ug/ml) and

one spanning a low concentration range of drug (0 to 100 ng/ml). For

the high concentration calibration curves, duplicate samples of 1 ml

of blank plasma were spiked with 0, 0.2, 0.4, 0.8, 1.6, 2.4, 3.2 and 4

ug of the analyte plus 1 pg of internal standard (Table 4-1). For the













Table 4-9. Limits of Detection (SNR=2) of the Anthracycline
Antibiotics.


Compound Limit of Detectiona Limit of Detectionb
(pg) in Plasma (ng/ml)

DOX 400 2

EPI 400 2

DEOXY 400 2

DMDR 400 2

CN1 200 1

CN2 200 1

a Amount in 200 ul injection of an aqueous solution.

b Plasma concentration measured by injection of 200 ul of plasma
extract.









low concentration range calibration curve, duplicate samples of 1 ml

of plasma were spiked with 0, 10, 20, 40, 60, 80 and 100 ng of analyte

plus 20 ng of internal standard. Each spiked plasma sample was

extracted in the manner described for the samples. In all cases, the

relationships between peak height ratio of drug to internal standard

and the concentration of drug injected were linear (r>0.990) with near

zero intercepts (Table 4-10).

The absolute recoveries of the extraction procedures were

determined by comparing the peak heights of the drug extracted from

spiked plasma (10 ng/ml) with the peak heights of the drug in aqueous

solutions at the same concentration. The absolute recoveries ranged

between 56.8% for DMDR and 83.2% for one of the two isomers of CN

(Table 4-11).

The accuracy and precision of the procedures were determined at

two concentrations (10 ng/ml and 1000 ng/ml). Five 1 ml aliquots of

blank plasma were spiked with either 10 ng or 1 ug of each compound

plus an internal standard and analyzed using the described

procedures. The relative standard deviation (coefficient of

variation), used as a measure of precision, ranged between 1.22% for

one of the CN isomers at 10 ng/ml and 6.43% for DOX at 1000 ng/ml

(Table 4-12). The accuracy of the procedures, measured by the

percentage of the actual concentration found, ranged between 94.5 and

106% (Table 4-12).

Using the extraction procedures described, an assay for urine

samples containing DOX was also developed. HPLC conditions used were

the same as those for plasma samples (Table 4-1). Concentrations as










Table 4-10. Statistical Parameters of Sample Calibration Curves from Extracted Plasma
Samples.

Compound 0-100 ng/ml 0-4 ug/ml

Slope (SE)a Y Intercept (SE)b SEyXxC Slope (SE)a Y Intercept (SE)b SEyx

DOX 0.027 (0.001) 0.02 (0.03) 0.06 0.48 (0.01) 0.00 (0.03) 0.05

EPI 0.024 (0.002) 0.14 (0.08) 0.15 0.67 (0.01) -0.01 (0.02) 0.05

DEOXY 0.019 (0.002) 0.09 (0.08) 0.03 0.80 (0.02) 0.01 (0.004) 0.07

DNR 0.036 (0.0004) 0.01 (0.02) 0.04 0.82 (0.05) -0.02 (0.02) 0.08

DMDR 0.031 (0.001) 0.01 (0.03) 0.07 0.55 (0.02) 0.04 (0.09) 0.11

CN1 0.020 (0.001) 0.06 (0.03) 0.06 0.12 (0.005) 0.02 (0.01) 0.02

CN2 0.015 (0.001) 0.05 (0.05) 0.10 0.10 (0.009) 0.10 (0.10) 0.02


Standard error of the slope.
Standard error of the Y intercept.

Standard error of estimate of y on x.














Table 4-11. Absolute Recoveries of the Anthracycline Antibiotics from
Plasma.


Compoundab Recovery ()c RSD (%)d


DOX 62.5 5.2

EPI 59.8 4.4

DEOXY 76.3 7.4

DMDR 56.8 4.5

CN1 83.2 5.1

CN2 76.4 5.9

a See Table 4-1 for chromatographic conditions.

b See Figure 4-1 for key. Concentration 10 ng/ml for each drug.

c Peak height of the drug extracted from plasma compared with that of
the drug in aqueous solution, expressed as a percentage.
d Relative standard deviation; n=5, RSD(%) = (SD/Recovery) x 100.









Table 4-12.


Accuracy and Precision Data for the Analysis
of Anthracycline Antibiotics in Plasma by HPLC-EC.


Compoundab Concentration (ng/ml) Accuracyc Precisiond

Added Found


DOX 1000 984 98.4 6.43
10.0 10.50 105.0 1.23

EPI 1000 1044 104.4 5.23
10.0 9.57 95.7 1.46

DEOXY 1000 995 99.5 6.46
10.0 9.45 94.5 3.89

DMDR 800 795 99.4 1.65
10.0 10.60 106.0 4.52

CN1 420e 421 100.2 2.38
CN2 380e 360 94.7 5.58


CN1 5.30f 5.40 101.9 1.22
CN2 4.70f 4.92 104.7 4.07


a See Table 4-1 for chromatographic conditions.

b See Figure 1-1 for key.


c Concentration found divided by concentration added,
percentage.


expressed as a


Relative standard deviation, n=5.


e Total CN concentration was 800 ng/ml. Concentrations of the
individual isomers was determined from their peak areas using an
HP 3392A integrator.
f Total CN concentration was 10 ng/ml. Concentrations of the
individual isomers was determined from their peak areas using an
HP 3392A integrator.








low as 10 ng/ml could easily be detected (SNR=2). Three calibration

curves were constructed spanning a low concentration range (0 to 1

ug/ml), a medium concentration range (1 to 10 pg/ml) and a high

concentration range (10 to 100 pg/ml). For the low concentration

calibration curve, duplicate samples of blank urine (1 ml) were spiked

with 0, 0.1, 0.2, 0.4, 0.6, 0.8 and 1 Ug of DOX and 0.5 pg of internal

standard (EPI). Similarly, for the medium concentration range

calibration curve, samples were spiked with 1, 2, 4, 6, 8 and 10 pg of

DOX and 3 pg of EPI and, for the high concentration range calibration

curve, samples were spiked with 10, 20, 40, 60, 80 and 100 pg of DOX

and 30 ug of EPI. Over all the concentration ranges, the

relationships between peak heights and DOX concentration injected were

linear (r>0.990) with near zero intercepts (Table 4-13).

The absolute recoveries of the extraction procedures were

determined by comparing the peak heights of DOX extracted from spiked

urine with the peak heights of DOX in aqueous solutions at the same

concentrations. The absolute recoveries at 0.6 pg/ml, 6 pg/ml and 60

Ug/ml were 62.94%, 80.57% and 74.29%, respectively (Table 4-14).

The precision and accuracy of the procedures were evaluated by

analyzing five samples of blank urine (1 ml) spiked with 0.6 Ug, 6 pg

or 60 pg and the appropriate amount of EPI. The relative standard

deviations were 2.48%, 6.98% and 10.18% for the three respective

concentrations (Table 4-15). Based on comparison of the found

concentration to the known concentration, the accuracy was 92.17%,

108.90% and 106.88% at 0.6 ug/ml, 6 pg/ml and 60 ug/ml, respectively

(Table 4-15).















Table 4-13. Statistical Parameters of Sample Calibration Curves for
Extracted Urine Samples of DOX.


Concentration Slope (SE)a Y Intercept (SE)b SEyXc
Range (ng/ml)


0 1 2.19 (0.11) 0.07 (0.07) 0.09

1 10 0.32 (0.02) 0.10 (0.10) 0.13

10 100 0.026 (0.002) 0.23 (0.11) 0.15


Standard error of the slope.

Standard error of the Y intercept.

Standard error of estimate of y and x.








Table 4-14. Absolute Recoveries of Doxorubicin from Urine.


Concentration Recovery (%)a RSD (%)b
(ug/ml)


0.6 62.94 1.95

6.0 80.57 3.19

60.0 74.29 6.69

a Peak height of DOX extracted from urine compared with that of DOX
in aqueous solution, expressed as percentage.
b Relative standard deviation, n=5.











Table 4-15. Accuracy and Precision Data for the Analysis of
Doxorubicin in Urine.


Concentration (ig/ml) Accuracy (%)a Precision (%)b

Added Found


0.6 0.55 91.67 2.48

6.0 6.40 106.70 6.98

60.0 64.13 106.88 10.18

a Concentration found divided by concentration added, expressed as a
percentage.
b Relative standard deviation, n=5.













CHAPTER 5
APPLICATIONS OF ANALYTICAL METHODS


5.1 Materials and Methods

Chemicals and reagents used in the studies were described in

Section 3.1.1. The HPLC-EC chromatograph used for analysis was

described in Section 4.1.

5.1.1 Stability Studies

The stability of DOX in plasma water and physiological buffer (pH

7.4) at room temperature (25C) was determined over a 6-hour period.

Plasma water was obtained by filtration of human plasma through

Centrific Ultrafiltration Membrane Cones (Amicon Corp., Danvers,

MA). The plasma (purchased from Civitan Regional Blood Center,

Gainesville, FL) was placed in the cones and centrifuged for 30

minutes at 2000 rpm. The filtrate (plasma water) was collected and

stored at 4C until time of use. Physiological buffer was prepared by

combining 0.05 M KH2PO4 in 0.1 M NaCl and 0.05 M NaH2PO4 in 0.1 M NaCl

to obtain a solution of pH 7.4. Solutions of DOX at concentrations of

200 and 1600 ng/ml in both plasma water and buffer were maintained at

room temperature and at certain time intervals 0.5 ml was removed.

One hundred and fifty and 1000 ng of internal standard (EPI) were

added to the 200 ng/ml and 1600 ng/ml solution of DOX, respectively.

The pH was adjusted to 8.4 with 0.05 ml of borate buffer (0.1 M, pH

9.1) and extracted with 5 ml of chloroform for 10 minutes. The








samples were centrifuged for 10 minutes at 2000 rpm, the aqueous layer

aspirated and the chloroform transferred to a clean test tube. After

evaporation of the chloroform under nitrogen at 40C, the residues

were reconstituted with 0.5 ml of 0.2 M H3PO4 (pH 2.0), shaken with 1

ml of chloroform and centrifuged for 2 minutes. Two hundred

microliters were then injected onto the HPLC column. Prior to the

analysis of solutions containing DOX in physiological buffer, 100 ng

and 1000 ng of internal standard were added to the 200 ng/ml and 1600

ng/ml buffer samples, respectively. Two hundred microliters of each

sample were then injected onto the HPLC column. The concentrations of

DOX were determined from calibration curves ranging from 0 to 200

ng/ml and 0 to 1600 ng/ml.

The stability of both isomers of CN in physiological buffer at

25C was also observed starting with a total CN concentration of 2000

ng/ml. Using the procedure described for buffer solutions of DOX,

1000 ng of internal standard (DNR) were added to each 0.5 ml sample.

The logarithms of the percentages of the initial concentrations of

each isomer remaining were plotted against time. The same procedure

was also used to determine the stability of solutions containing 1000

ng/ml of EPI, DEOXY, DNR, and DMDR to which 500 ng of internal

standard (Table 4-1) were added.

5.1.2 Protein Binding

The extent of human plasma protein binding of each of the

anthracyclines was determined by ultrafiltration. An Amicon Corp.

MicroPartition System MPS-1 (Danvers, MA) with YMT membranes was

used. The extent of binding was determined over the general








therapeutic concentration range of 100 to 3000 ng/ml. An appropriate

amount of stock solution of each compound was added to plasma to make

a final volume of 1 ml at the desired concentration. The sample was

pipetted into the filtration device and centrifuged for 15 to 20

minutes at 2000 rpm. An aliquot of 0.3 ml of the ultrafiltrate was

transferred to a test tube and 0.03 ml of borate buffer was added to

adjust the pH to 8.4. To the DOX samples 100 ng of internal standard

were added while 80 ng of internal standard were added to the other

compounds. The samples were extracted as described previously (Sec.

5.1.1). Residues of DOX, EPI, and DEOXY were reconstituted with 0.220

ml of 0.2 M H3PO4 (pH 2.0) and shaken with 1 ml of chloroform.

Residues of DMDR and CN were reconstituted with 0.220 ml of 0.1 M

KH2PO4 (pH 4.5) and shaken with 1 ml of hexane. After centrifugation

for 2 minutes, 200 pl of the aqueous layer were injected onto the HPLC

column. Calibration curves were constructed by adding appropriate

amounts of stock solution of each compound to plasma water and

extracting the solutions (0.3 ml) by the method described.

Protein binding was also determined by red blood cell

partitioning based on a method by Garrett and Hunt (90) for DOX, EPI,

DEOXY and DMDR. Packed human red blood cells (purchased from Civitan

Regional Blood Center, Gainesville, FL) were washed twice with normal

saline and once with pH 7.4 buffer with subsequent centrifugation.

One milliliter of washed red blood cells was added to 1 ml of plasma

containing the compound to make a final drug concentration of 500

ng/ml. The hematocrits of each sample were determined using a

microhematocrit (Damon/IEC Division, Needham Hts., MA). After 2 hours








at 25C, the samples were centrifuged for 10 minutes at 2000 rpm and

0.5 ml of the supernate was assayed. The procedure was repeated using

pseudoplasma which contained various fractions of true plasma mixed

with pH 7.4 buffer at plasma fractions of 0.05, 0.1, 0.2, 0.3 and 0.4.

5.1.3 Erythrocyte Binding

The distribution of DOX between plasma water and red blood cells

(RBCs), and physiological buffer (pH 7.4) and red blood cells at 25C

were studied as a function of time at two concentrations, 200 and 1600

ng/ml. Packed human red blood cells were washed twice with normal

saline and once with pH 7.4 buffer with subsequent centrifugation.

The stock solution of DOX was prepared in pH 7.4 buffer immediately

prior to use. Twenty milliliters of packed RBCs and 20 ml of buffer

containing the DOX were mixed and duplicate 2 ml samples taken at 0,

0.25, 0.5, 1, 2, 4, and 6 hours. Hematocrits of each sample were

determined. Each 2 ml sample was centrifuged for 5 minutes at 2000

rpm and 0.5 ml of the supernatant was transferred to a clean test

tube. The plasma and buffer samples were assayed as described in

Section 5.1.1.

The distribution of the 2 CN isomers was also studied as a

function of time between physiological buffer and RBCs at a total CN

concentration of 1000 ng/ml at 25C. Ten milliliters of buffer

containing CN were mixed with 10 ml of washed packed RBCs. A 1 ml

sample was taken at 0, 0.33, 0.67, 1, 1.5, 2, 2.5, 3, 4, 5, and 6

hours. The hematocrit was determined and the sample centrifuged for 3

minutes. A 250 pl aliquot of the supernatant was removed to which 25

ng of internal standard (DNR) were added and 200 Il immediately








injected onto the HPLC column. The experiment was performed twice as

were similar experiments in which 1 ml samples were taken at 0, 5, 10,

15, 20, 25, 30 and 60 minutes. Calibration curves were constructed

from buffer solutions of known CN concentrations.

The apparent distribution of DOX, EPI, DEOXY and DMDR between

plasma water and RBCs was determined at concentrations of 50, 200 and

1600 ng/ml. Stock solutions were prepared in physiological buffer.

One milliliter of plasma water containing the appropriate amount of

stock solution was mixed with 1 ml of washed packed RBCs. The

hematocrits were determined and the samples allowed to equilibrate for

2 hours. The procedures described previously in Section 5.1.1 were

used to determine the sample concentrations.

The rate of efflux of each compound from loaded RBCs was

determined. The efflux rates of DEOXY from RBCs previously

equilibrated with 500, 1000 and 2000 ng/ml of drug were measured. The

efflux rates of DOX, EPI, DEOXY, DMDR and DNR were determined at 2000

ng/ml. All experiments were performed in duplicate. To load the

cells, 1 ml of washed packed RBCs was resuspended in 1 ml of

physiological buffer containing the appropriate concentration of

drug. The cells were allowed to equilibrate at 25C for 2 hours then

centrifuged for 10 minutes. The supernatant was removed from the

cells. Approximately 0.5 ml of the loaded packed cells was added to

50 ml of buffer containing no drug at 25C stirred by a Teflon coated

magnetic. At certain time intervals 1 ml samples were removed and

immediately cooled to 0C with acetone/ice which prevented further

drug efflux. The samples were centrifuged at 0C to 5C for 2 minutes








at 5000 rpm. Then 250 ul of the supernate were removed, internal

standard was added and 200 0l were injected onto the HPLC column.

5.1.4 Clinical Pharmacokinetics of Doxorubicin and Deoxydoxorubicin

A patient suffering from neoplastic disease was administered 52

mg of DEOXY by intravenous infusion. Blood samples were taken

immediately after the end of the infusion and then after the following

periods of time: 5, 10, 15, 30 and 45 minutes and 1, 2, 4, 6, 8, 12,

24, 30 and 48 hours. The blood samples were collected in heparinized

tubes, centrifuged and the plasma frozen (-20C) prior to analysis.

These samples were provided by Dr. C.W. Young (Sloan-Kettering

Memorial Cancer Center, New York, NY) and had been assayed previously

in his laboratory.

At Shands Hospital (Gainesville, FL) patients with osteogenic

sarcoma are initiated on a one-year protocol which includes

chemotherapy, radiation and surgery. Two treatments with doxorubicin

and cisplatin are given at three-week intervals prior to surgery.

Doxorubicin is administered intravenously and followed within three

hours by an intra-arterial two-hour infusion of cisplatin (CDDP). The

pharmacokinetics of DOX in two patients pretreated in this manner was

investigated. Venous blood samples were drawn into heparinized tubes

before DOX administration and at various time intervals thereafter.

The samples were centrifuged and the separated plasma divided for

analysis of DOX in this laboratory and for the determination of CDDP

at the University of Kansas. The plasma samples were frozen (-20C)

prior to analysis. The total urinary output was determined and 5 ml

samples frozen following administration of DOX until the patient was








released from the hospital, which was usually after 24 to 36 hours.

The plasma and urine samples were analyzed as described in Section

4.1.



5.2 Results and Discussion

5.2.1 Stability Studies

In order to choose suitable methods for the determination of

protein binding and erythrocyte partitioning, it was necessary to

estimate the stability of the compounds at pH 7.4 and 25C.

Preliminary investigations had shown that the isomers of CN were

unstable in aqueous solution. The stability of DOX in plasma water

and physiological buffer at 25C was observed over a 6-hour period at

200 and 1600 ng/ml concentrations. Table 5-1 shows little difference

in degradation between plasma water and buffer at both concentrations

and about 90% of the drug remained intact after 6 hours. Similar

results (Table 5-1) were seen for DNR, DMDR, EPI and DEOXY and it was

concluded that the degradation of the drugs would not seriously

compromise the protein binding and RBC partitioning experiments.

At a total concentration of 2000 ng/ml the stability of both CN

isomers was determined also over a 6-hour period (Figure 5-1). The

isomer which elutes first (CN1) off the HPLC column degraded faster

than the second isomer (CN2). The degradation of both isomers

followed first-order kinetics (Figure 5-1) according to the equation


log [CN] = log [CN]0 k obst/2.3


(5-1)












Table 5-1. Stability of Anthracycline Antibiotics in pH 7.4
Physiological Buffer at 25C.


Compound Concentration Percent Remaining
(ng/ml) After 6 hours (SD)a


DOX 200 88.77 (1.48)

DOXb 200 91.5 (2.79)

DOX 1600 92.09 (3.60)

DOXb 1600 88.92 (5.81)

EPI 1000 87.20 (3.17)

DEOXY 1000 91.64 (4.37)

DNR 1000 92.26 (1.59)

DMDR 1000 90.22 (1.72)

CN1 2000c 12.53 (0.96)

CN2 2000 20.66 (0.99)

a n=2.

b In plasma water.

c Total CN concentration.




84







100

80-


60 A




40







S20-




\A


10




0 1.5 3 4.5 6
Hours


Figure 5-1. Stability of the two isomers of CN. CN1 (n) and CN2
(A) are the isomers which elute first and second from
the HPLC column. Temperature = 25C.









where kobs is the pseudo first-order rate constant. The values of

kobs (and half-life), determined by fitting the data to equation 5-1

using linear regression, were 0.385 h-1 (1.80 h) and 0.271 h-1

(2.55 h) for CN1 and CN2, respectively. The degradation of CN1 and

CN2 did not compromise the protein binding experiments since the

ultrafiltration was completed within 15 minutes during which the loss

of either isomer from solution would be less than 10%. In addition,

since only the free fraction is unstable, the loss of drugs would be

much less than anticipated from the stability studies performed in

buffer.

5.2.2 Distribution of Anthracyclines Between the Components
of Blood

The distribution of drugs between the components of blood may be

conveniently described by Scheme 5-1 where C refers to the

concentrations of drug in each of the compartments (RBC fluid, RBC

membrane, plasma water and plasma protein). The first subscripts r,

m, and p refer to the RBC fluid, membrane and plasma, respectively.

The second subscripts u and b refer to the unbound (or free) and bound

forms of the drugs, respectively. The distribution of drugs between

plasma water and plasma protein and between plasma water and RBCs may

be studied by physically separating the various components of blood.

Red blood cells may be separated by centrifugation and purified by

washing with physiological buffer which is used to prevent hemolysis

(Sec. 5.1.3). Plasma water is produced by ultrafiltration of plasma

isolated by centrifugation from blood.


















































Scheme 5-1. The distribution of drugs between the components of
blood.








If transport across the RBC membrane occurs by diffusion of the

unbound drugs, then the rates of influx (dCp, u/dt) and efflux

(dCr,u/dt) are both given by Fick's law (91) (equations 5-2 and 5-3,

respectively):


-dC pu/dt = (P.A e.D mp/1 m)(Cpu Cr,u ) (5-2)


-dC ru/dt = (P.A.D mr/Im )(Cru C pu) (5-3)


where P is the diffusivity, 1m is the thickness of the membrane and Ai

and Ae are internal and external surface areas of the RBC membranes,

respectively. The partition coefficients of the drug between the RBC

fluid and the membrane (Dm,r) and between the plasma water and the

membrane (D ,p) are given by equations 5-4 and 5-5, respectively.


Dm,r = Cm/Cr,u (5-4)

Dmp = C /Cu (5-5)
m,p m p,u

At equilibrium


dC pu/dt = dCru/dt = 0 (5-6)

which can only occur when the concentration gradients (Cpu Cru and

Cr,u Cp,u) are zero, in which case Cr,u = Cpu. The absolute RBC

partition coefficient, K, is given by


K = C /C pu
r,u p,u


(5-7)








which must equal unity since the concentrations of the drug on either

side of the membrane are equal. The apparent RBC partition

coefficient (D) is given by:


D = C rtotal/Cpu (5-8)


which can be greater than unity if the drug is bound to the cell

membrane and/or the intracellular components (predominantly

hemoglobin). In addition, differences in pHs of the intracellular and

extracellular fluids may affect the degree of ionization of the drugs

which also affects the apparent partition coefficient (42). Of

course, this assumes that there is no protein in the plasma water.

5.2.3 Protein Binding

The three methods commonly used to determine protein binding are

equilibrium dialysis, ultracentrifugation and ultrafiltration (21,

92). Both equilibrium dialysis and ultracentrifugation require rather

long time periods (18 to 24 h). Another potential problem is binding

of the compound to either the dialysis membrane or ultrafiltration

device (membrane or cone). Adsorption of DOX to dialysis membranes

has been observed (41, 43) and based on the stability studies, it is

also evident that dialysis and ultracentrifugation would not be

suitable for CN studies. Ultrafiltration requires less than 20

minutes and was found to be suitable for the measurement of protein

binding since there was no binding of the anthracyclines to the

filtration apparatus used in these investigations. In addition the

Amicon MicroPartition System used here allows separation of free from