UNIVERSITY OF HAWAI'I LIBRARY

CANCER CHEMOPREVENTION BY WATER SOLUBLE ASTAXANTHIN

DERIVATIVES

A THESIS SUBMITTED TO THE GRADUATE DIVISION OF THEUNIVERSITY OF HAWAtI IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

IN

BIOMEDICAL SCIENCES(CELL AND MOLECULAR BIOLOGY)

MAY 2004

ByLauraM. Hix

Thesis Committee:

Tom Humphreys, Co-ChairpersonJohn Bertram, Co-Chairperson

Pratibha Nerurkar

TABLE OF CONTENTS

List of Figures , .IV

Chapter 1: Introduction 1

Chapter 2: Materials and Methods 7

2.1. Materials 72.1.1. Compounds 72.1.2. Cel/lines and culture conditions 9

2.2. Methods 92.2.1. Analysis ofCx43 expression 92.2.2. Immunofluorescence 102.2.3. Gapjunctional communication assay 102.2.4. MCA-induced neoplastic transformation assay 112.2.5. Statistical analysis 12

Chapter 3: Results 13

3.1. Formulation and Solubility Studies 133.2. Molecular Analysis ofCx43 , 153.3. Cellular Analysis ofCx43 203.4. Induction of Gap Junctional Communication 223.5. Inhibition of MCA-induced Neoplastic Transformation 23

Chapter 4: Discussion 26

Appendix 31

References 32

III

LIST OF FIGURES

Figure Page

I. Structures of the disodium salt astaxanthin derivatives 8

2. UV/vis absorption spectra ofracemic dAST and AST 14

3. Western blot demonstrating Cx43 protein expression in lOTI/2 cells treated with

two different formulations of racemic dAST 15

4. Western blot demonstrating induction ofCx43 protein expression in IOT1I2 cells

treated with dAST derivatives 17

5. Western blot demonstrating induction ofCx43 protein expression in IOTII2 cells

treated with pAST, AST and CTX 19

6. Immunofluorescence demonstrating dAST increases assembly of Cx43

immunoreactive junctional plaques .21

7. Micrographs demonstrating dAST enhances functional GJIC in lOTI/2 cells .23

8. Dose-dependent effects ofpAST and AST on MCA-induced neoplastic

transformation 25

IV

CHAPTER 1. INTRODUCTION

It is now well established that the major contributing factor to cancer risk in

Western societies is lifestyle rather than genetics. Current epidemiological data suggests

that approximately 70% of the estimated 1.37 million cancer cases in 2004 will be linked

to preventable lifestyle factors such as tobacco use, alcohol, diet, infections (including

sexually transmitted diseases), occupational exposures, pollution, sunlight exposure (UV

radiation), and stress (Cancer prevention and early detection, American Cancer Society,

2004); [1]. Despite considerable improvements in the detection and treatment of cancer,

it is estimated that 563,700 Americans will die ofcancer in the year 2004 (Cancer facts

and figures, American Cancer Society, 2004). Cancer chemoprevention, defined as the

reduction in cancer incidence through the use of pharmaceuticals, vitamins, minerals and

other chemicals, has emerged as a powerful strategy in the fight against cancer. Potential

targets for chemoprevention include the general population, subgroups at risk due to

lifestyle or other environmental factors, individuals with precancerous lesions, and cancer

survivors at risk for secondary tumors.

A growing body of epidemiological evidence has linked increased intake and/or

blood levels of several dietary carotenoids, plant pigments found in green, yellow and

orange fruits and vegetables, to decreased risk of cancer at many anatomic sites. This

activity appears to be independent of the pro-vitamin A activity of these carotenoids [2].

Studies in experimental animal and cell culture models ofcarcinogenesis have confirmed

the cancer chemopreventative activity of many dietary carotenoids. In these systems,

activity does not correlate with the pro-vitamin A properties of these compounds, nor

does it correlate with their ability to act as lipid-phase antioxidants [3].

In the C3H/lOTl/2 mouse embryo fibroblast (lOTl/2) cell system developed in

the lab of the late Charles Heidelberger, cells undergo neoplastic transformation in

response to many chemical and physical carcinogens in a dose-dependent manner [4].

This cell system has been shown to effectively mimic the initiation and transformation

events of tumor formation in whole animals, and represents one of the most widely

characterized in vitro models for carcinogenesis [5]. Studies conducted in the 10Tl/2 cell

system revealed that carotenoids active in inhibiting neoplastic transformation

upregulated gap junctional intercellular communication (GJlC) in direct relationship to

their activity as chemopreventative agents [3].

GJIC permits the transfer of ions and small hydrophilic molecules <1 kiloDalton

(kDa) in size by passive diffusion through aqueous channels (connexons) that span the

plasma membranes of adjoining cells. Individual connexons are comprised of proteins

known as connexins. Communicating cells typically form several thousand connexons

that assemble into plaques. A family of approximately 20 connexin members is

expressed in mammals. These connexins exhibit developmental- and differentiation­

specific expression, allowing the formation of communicating compartments between

compatible family members. There is growing evidence that these compartments are

vital to normal growth and development.

Interest in gap junctions and carcinogenesis stems from several independent lines

of evidence. Following the discovery ofjunctional communication in cultured normal

cells, it was soon discovered that neoplastically-transformed cells were deficient in GJlC.

Moreover, inhibition ofGJIC was a very early event after activation ofthe oncogene src

[6]. Human and animal tumors of many pathologic types were found to be deficient in

2

GJIC, either as a consequence of decreased connexin expression or faulty assembly into

the plasma membrane. Restoration ofjunctional communication between transformed

cells and growth-inhibited normal cells resulted in growth arrest of the transformed cells

in direct proportion to the extent of heterologous cell coupling [7]. Tumor promoters­

agents that accelerate the process of carcinogenesis-were found to inhibit GJIC, while

cancer preventive agents such as retinoids and carotenoids-agents that delay

carcinogenesis-had the opposite effect. .Finally, transfection of molecular expression

constructs into human and animal neoplastic cells, utilizing either constitutive or

inducible promoters, demonstrated that connexin re-expression in these cells decreased

their neoplastic potential in in vitro and in vivo assays. Collectively, these lines of

evidence support the conclusion that connexins can function as tumor-suppressor genes

[8].

Studies of human and animal cells have demonstrated that connexin 43 (Cx43),

the most widely expressed connexin in human and animal tissues, is upregulated at the

message and protein level by chemopreventive retinoids and carotenoids. The ability of

carotenoids to regulate Cx43 expression appears to be independent of their conversion to

retinoids. This is exemplified by compounds such as lycopene, a straight-chain

hydrocarbon that is not cleaved to vitamin A in mammals, yet possesses this regulatory

ability. It is unclear how much the parent compound contributes to the observed increase

in Cx43 expression; for example two oxidation products of lycopene are active in this

respect [9;10]. Moreover, the potential conversion of carotenoids to

retinoids--chemopreventive agents which are highly potent inducers ofCx43 and GJIC

[11;12]-eannot be ignored. Indeed, there is evidence that conversion of canthaxanthin

3

to 4-oxo-retinoic acid, an active retinoid, may be in part responsible for increased Cx43

expression in lOTl/2 cells [13], although we found no evidence of this conversion in the

same system [14].

In the 10Tl/2 system, upregulated GJIC as a consequence of increased Cx43

expression is highly correlated with the ability of carotenoids to prevent neoplastic

transformation of these cells [15;16]. Due to these multiple lines ofevidence linking

GJIC with inhibition of carcinogenesis and decreased neoplastic potential of established

transformed cells, Bertram and others have proposed that GJIC allows for the

transmission of growth controlling signals between normal and transformed cells [17]. In

this model, the ability of carotenoids and retinoids to inhibit tumor progression is a result

of enhanced GJIC between carcinogen-initiated cells and surrounding growth-controlled

normal cells. By extension, the ability to upregulate GJIC may be an important indicator

of chemopreventative potential.

Astaxanthin, found predominantly as a dietary source in shrimp, lobster and

salmon, has been associated with reduced risk of diseases such as age-related macular

degeneration and ischemic diseases, effects attributed to its potent antioxidant activity

[18]. In addition, the antioxidant activity of astaxanthin has been reported to be 10 times

stronger than that of other carotenoids, namely, zeaxanthin, lutein, canthaxanthin, and ~­

carotene [19;20]. In experimental animal studies astaxanthin has been shown to be

capable of inhibiting chemically induced oral and bladder carcinogenesis [21 ;22]; its

usefulness as an immunomodulating agent in experimental cancer studies is also well

documented (reviewed [23;24]). Based on these findings, astaxanthin has significant

cancer chemopreventative potential.

4

One problem confronting researchers investigating the effects of carotenoids such

as astaxanthin in cell culture and/or whole animals has been the delivery of these highly

lipophilic molecules to target cells. For cell culture studies, the Bertram lab developed

the use of tetrahydrofuran (THF) as a delivery solvent, the use of which creates a highly

bioavailable and non-toxic pseudo-solution of carotenoids in cell culture medium [25].

However, caution must be taken to protect THF from oxidation to toxic species, and this

method is unsuitable for use in whole animal studies. Delivery of carotenoids to

experimental animals or humans is usually achieved by delivery of carotenoids as a

suspension in oil, frequently as a micro-dispersed emulsion. Unfortunately,

bioavailability is usually low in animals (e.g. rodents) and can be variable in humans.

To circumvent these problems of drug delivery Hawaii Biotech, Inc. has

synthesized a set of novel carotenoid derivatives, disodium salt disuccinate (dAST) and

disodium salt diphosphate (PAST) derivatives of synthetic astaxanthin (3,3'-dihydroxy-p,

p-carotene-4,4'-dione), in all-trans (all-E) form. Synthetic astaxanthin can be

commercially obtained most economically as the statistical mixture of stereoisomers

[optically active (38,3'8) and (3R,3'R) and optically inactive (3R,3'S; meso) in a 1:1:2

ratio]; the statistical mixture of stereoisomers is known as (3RS,3 ,RS)-astaxanthin or

"racemic"astaxanthin (Buckton Scott, India). The individual stereoisomers are also

available commercially from Hoffman-LaRoche (Switzerland) and BASF (Germany).

The racemic mixture of the novel derivative, as well as purified stereoisomeric forms of

the derivative, were successfully synthesized and tested in the current study. The

derivatives exhibit several unique characteristics that increase their utility for use in the

cancer chemoprevention setting. They are water soluble (critical micelle concentration =

5

0.3 mg/mL) and water dispersible, with the maximum aqueous dispersibility greater than

8 mg/mL (approximately 10 mM), allowing them to be introduced into cell culture

without a co-solvent and delivered parenterally to experimental animals [27;28];

(Lockwood, unpublished results). Additionally, the intact synthetic derivatives retain

antioxidant activity prior to enzymatic cleavage to mono-succinates and non-esterified,

free astaxanthin [29]. Derivatives that enhance astaxanthin's water solubility and

bioavailability should prove invaluable in assessing the carotenoids' abilities in vitro and

in vivo.

This thesis evaluates the ability of the novel derivatives delivered in several

aqueous formulations to upregulate Cx43 protein expression, induce functional GllC, and

inhibit carcinogen-induced neoplastic transformation in the lOTI/2 cell culture system

described above. The compounds were found to induce expression of functional Cx43

protein, significantly increase GllC and significantly inhibit MeA-induced neoplastic

transformation with enhanced ability over the parent carotenoid astaxanthin. These

results indicate that the major hurdle ofdelivering hydrophobic carotenoids to biological

tissues in model systems has been overcome, and that evaluation of these highly

bioavailable astaxanthin derivatives in in vivo models of cancer chemoprevention should

be pursued.

6

CHAPTER 2. MATERIALS AND METHODS

2.1 Materials

2.1.1. Compounds

The racemic disodium salt diphosphate derivative ofastaxanthin (racemic pAST; >

90% purity by HPLC) was synthesized from commercial astaxanthin and its structure

verified. It was prepared in a fonnulation of20% EtOH and sterile water (0.2% final

EtOH cone.) to minimize supramolecular aggregation. The racemic disodium salt

disuccinate derivative ofastaxanthin (racemic dAST; > 90% purity by HPLC) was

synthesized from commercial astaxanthin and its structure verified. It was prepared in a

fonnulation of33% EtOH and sterile water (0.33% final EtOH cone.). Non-esterified,

all-E synthetic astaxanthin (AST) (> 96% purity by HPLC; Sigma, St. Louis, MO) was

added to tetrahydrofuran (THF). Synthetic canthaxanthin (CTX) (CaroteNatur,

Switzerland) was added to THF. The synthetic retinoid TTNPB [p-«E)-2-(5,6,7,8­

Tetrahydro-5,5,8,8-tetramethyl-2-napthyl) propenyl>benzoic acid] (Hoffman-LaRoche,

Nutley, NJ) and retinol acetate (Sigma, St. Louis, MO) were prepared in acetone (Sigma,

St. Louis, MO). TTNPB, canthaxanthin and astaxanthin concentrations were confinned

by comparing their UV absorption and their published extinction coefficients. The

chemical structures of the three stereoisomers are shown in Fig. IA. These are:

(3R,3'R)-, (3S,3'S)-, and (3R,3'S; meso)-astaxanthin disuccinate, disodium salt. The

structure of the all-E-astaxanthin diphosphate, disodium salt is shown in Figure. IB. Due

to the sensitivity ofcarotenoids to light, heat and oxygen, special precautions were taken

7

throughout the study. All compounds were stored under nitrogen at -70° C and care was

taken to ensure minimal exposure to direct sunlight or UV light.

A

3R,3'R disodium salt disuccinateastaxanthin derivative

o

3S.3'S disodium salt disuccinateastaxanthin derivative

o3R,3'S disodium salt disuceinateastaxanlltin derivative

BRacemic disodium diphosphateastaxanthin derivative

oNaO-~...,.,

NaO' U

o

oONa

O~P/-ONaIIo

Fig. 1. (A) Structures of the three stereoisomers of the disodium salt disuccinate astaxanthin derivatives

(clAST) evaluated in the current study (shown as the all-E geometric isomers). The racemic mixture of

stereoisomers, referred to as ''racemic clAST" in the text, contains (38,3'S)-astaxanthin disuccinate,

disodium salt; (3R,3'R)-astaxanthin disuccinate, disodium salt; and (3R,3'S; meso)-astaxanthin disuccinate,

disodium salt in a 1: 1:2 ratio. (B) Structure of the racemic disodium salt diphosphate astaxanthin derivative

8

(pAST), containing (38,3'S)-astaxanthin diphosphate, disodium salt; (3R,3'R)-astaxanthin diphosphate,

disodium salt; and (3R,3'S; meso)-astaxanthin diphosphate, disodium salt in a 1:1:2 ratio. Racemic pAST,

racemic dAST and all individual dAST stereoisomers (38,3'8 dAST, meso dAST, and 3R,3'R dAST) were

separated to > 90% purity by HPLC.

2.1.2. Cell lines and culture conditions

Mouse embryonic fibroblast lOTl/2 cells were cultured in Eagle's basal medium

with Earle's salts supplemented with 4% fetal calf serum (Atlanta Biologicals, Norcross,

GA), 25 J.lg/mL gentamicin sulfate (Sigma, St. Louis, MO), and passaged by

trypsin/EDTA. Cells were incubated at 37° C in 5% C02 as described previously [1;2],

and confluent cells were treated on the 7th day after seeding, unless otherwise indicated in

methods.

2.2. Methods

2.2.1. Analysis ofCx43 expression

Expression of Cx43 protein in murine fibroblasts was assessed by immuno­

(Western) blotting essentially as described [1;2]. Briefly, lOTl/2 cells were treated with

the novel astaxanthin derivatives and/or retinoids at confluence on the 7th day after

seeding in 100 mm dishes (Fisher Scientific, Pittsburgh, PA). Cells were harvested after

4 days and total protein content was measured using the Protein Assay Reagent kit

(Pierce Chemical Co., Rockford, IL) according to the manufacturer's instructions. Cell

lysates containing 50 J.lg of protein were analyzed by Western blotting using the NuPage

Western blotting kit (Invitrogen, Carlsbad, CA). Cx43 was detected using a rabbit

polyclonal antibody (Zymed, San Francisco, CA) raised against a synthetic polypeptide

9

corresponding to the C-terminal domain common to mouse, human and rat Cx43. Cx43

immunoreactive bands were visualized by chemiluminescence using an anti-rabbit HRP­

conjugated secondary antibody (Pierce Chemical Co., Rockford, IL). Images were

obtained by exposure to X-ray film as previously described [1 ;2] or obtained with a

cooled CCD camera, and quantitative densitometry was performed (Bio-Rad, Richmond,

CA). Equal protein loading of the lanes was confirmed by staining with Coomassie blue

protein (Sigma, S1. Louis, MO) and digital image analysis.

2.2.2. Immunofluorescence

Expression and assembly ofCx43 into plaques was assessed by

immunofluorescence staining essentially as described [3]. Briefly, confluent cultures of

10Tl/2 cells were grown on Permanox plastic 4-chamber slides (Nalge Nunc

International, Naperville, IL) and treated for 4 days with (1) racemic dAST dissolved in

33% EtOH; (2) TTNPB at 1 x 10-8 M in acetone (0.1 % final acetone concentration in

assay) as a positive control; or (3) 33% EtOH (0.33% EtOH final concentration) as a

solvent-only control. Cells were fixed with - 20°C methanol overnight, washed in buffer,

blocked in 1% bovine serum albumin (Sigma, S1. Louis, MO) in PBS, incubated with the

rabbit anti-Cx43 antibody, and finally detected with Alexa568 conjugated anti-rabbit

secondary antibody (Molecular Probes, Eugene, OR). Slides were illuminated with 568

nm light and images were acquired at a wavelength of 600 nm using a Zeiss Axioplan

microscope and a Roper Scientific cooled CCD camera.

10

2.2.3. Gap junctional communication assay

Junctional permeability was assayed by microinjection of the fluorescent dye Lucifer

Yellow CH (10% in 0.1 M LiCI) (Sigma, St. Louis, MO) into confluent cells as described

previously [4]. Confluent cultures of lOTl/2 cells were treated for 4 days with: (1)

racemic dAST at lO-5 M in 33% EtOH; (2) TTNPB in acetone at lO-8 M as positive

control; or (3) 33% EtOH as solvent-only control as described above. Single cells were

injected with the fluorescent dye and digital images were taken under UV illumination

after 2 minutes. The number of fluorescent cells adjacent to the injected cell was later

determined by digital image analysis, and this number was used as an index ofjunctional

communication, as described previously [4].

2.2.4. MeA-induced neoplastic transformation assay

In this experiment, the novel derivatives were assessed for their ability to prevent

carcinogenic neoplastic transformation in the C3H/1 OT1I2 cell assay developed in the

Bertram lab [4;5]. Using the protocols previously described, cells were initiated with the

potent carcinogen methylcholanthrene (MCA, Sigma, St. Louis, MO) in acetone at a final

concentration of 5 f,lg/mL, and media was replaced 24 hours later. Cells were treated

with either the novel carotenoid derivative (pAST) in 20% EtOH (0.2% final EtOH

concentration) or the parent carotenoid astaxanthin (AST) in THF (0.1 % final THF

concentration) beginning 7 days after removal of carcinogen and continuing to the end of

the four-week culture period [6]. The ability to inhibit neoplastic transformation was

assessed at concentrations of lO-5, lO-6, lO-7 and 10-8M. Negative controls include MCA­

treated cells followed by media only treatments, Acetone-treated cells followed by media

11

only treatments, and MCA-treated cells followed by treatment with THF as a solvent

control. A total of 24 dishes/treatment has previously been shown to yield an expected 50

transformed foci in MCA-only controls and 10 foci in canthaxanthin 10-5 M treated cells,

and allows statistical significance to be met over most of the dose range [7].

2.2.5 Statistical Analysis

Statistical analyses were conducted with the NCSS 2001 and PASS 2002

statistical software (J. Hintze, Number Croneher Statistical Systems, Kaysville, UT). All

tests were conducted at the a = 0.05 level.

12

CHAPTER 3. RESULTS

3.1. Formulation and Solubility Studies

In pure aqueous formulation, racemic dAST and the purified stereoisomeric dAST

derivatives demonstrate supramolecular assembly, a form of 3-dimensional aggregation

that can limit solubility and bioavailability [27;31]. The aggregation is of the so called

H- or "card-pack" type, and may be disrupted into a completely monomeric solution with

the addition of a less polar solvent (such as EtOH) [32]. We observed precipitation of red

solid onto the cellular monolayer several days after introduction of dAST from purely

aqueous stock solutions, without effect on the cellular viability (data not shown). This

phenomenon has also been observed after introduction of non-esterified, free astaxanthin

in EtOH to cell culture systems [Lockwood, unpublished results]. To enhance both

solubility and bioavailability, several EtOH/water formulations were tested for

aggregation with UVIvis spectroscopy. The "solubility" of the derivatives was

significantly enhanced by the use of 1:1 (50% EtOH) and 1:2 (33% EtOH) EtOH/water

formulations. These formulations have been previously demonstrated to maintain the

carotenoid derivatives in monomeric form [26]. The addition of EtOH at 50% final

concentration caused the expected red-shifted, hyperchromic changes in the UVIvis

absorption spectrum of the racemic dAST derivative, from 423 nm (card-pack aggregate)

to 484 nm (Amax of the non-esterified, free astaxanthin chromophore measured in

acetone), due to the disaggregation of the compound to molecular monomers (Figs. 2A

and 2B) [33]. Therefore, ethanolic formulations were utilized in subsequent analyses in

the current study. Non-esterified, synthetic all-E astaxanthin exhibited a lack of

13

solubility in both the water and EtOH/water fonnulations as assessed

spectrophotometrically (Figs. 2C and 2D), demonstrating the significant enhancement of

water solubility achieved by the derivatives over the parent carotenoid.

A 1.0 C 1.0

c c0 .~';:. 0.5 Q, D.S'" ..i <:l..

.Q .c-< '(

0.0 I65fj 200 6SO

WIl'v"len:tb (11m) WavtleDtth (om)

1.0B D 1.0

-I c~

',ca. 0.5 e- t.S..j 0

ill•< ~

0.1) «..0200 65ft 200 6S0

Waveleagth (nm) 'W'.vtl~llth(nrn)

Fig, 2, UVivis absorption spectra ofracemic dAST dissolved in water, 2A, compared to a solution in 50%

EtOH, 2B (all solutions at 10-5 M racemic dAST). The Amax in 50% EtOH (484 nm) is red-shifted and

hyperchromic relative to the aggregated state in water (423 nm), expected spectral changes for carotenoids

which demonstrate supramolecular assembly in aqueous solution, Panel2C: lack ofUV/vis absorption of

non-esterified, synthetic all-E astaxanthin in water, or as shown in 2D, in 50% EtOH, illustrating the dramatic

increase in aqueous solubility achieved with the disodium disuccinate derivatization ofnon-esterified, free

astaxanthin.

Studies conducted using the disodium diphosphate derivative (pAST) yielded

similar disaggregation effects with the use of 20% EtOH. This allowed the compound to

be directly dissolved in media without precipitation for up to 24 hours, upon which

14

phosphatases present in the media and serwn presumably resulted in cleavage of the

diphosphates to the parent moiety and subsequent precipitation in solution.

3.2. Molecular Analysis of Cx43

Initial treatments of the lOT1/2 cells were performed with the goal of comparing the

relative biological efficacy of the racemic dAST in both water and in 50% EtOH to

enhance connexin 43 expression. Cells were treated with racemic dAST in water at 10-5,

10-6, and 10-7 M, as well as in a 50% EtOH formulation at 10-5 M (0.5% final EtOH

concentration in assay). 50% EtOH was used as a solvent-only control. Racemic dAST

caused induction of Cx43 in comparison with solvent-only treated controls in a dose-

dependent manner when dissolved in sterile water (Fig. 3). Additionally, induction levels

were higher with the EtOH formulation at 10-5 M than for the formulations in sterile

water alone, demonstrating enhanced biological efficacy using EtOH as a co-solvent, as

suggested by the physico-chemical studies described in Section 3.1 above. Racemic

dAST in pure aqueous formulation was able to upregulate Cx43 expression with

equivalent or greater potency than that previously observed for other carotenoids in

organic vehicle [15;16]. Importantly, this activity was achieved via direct delivery in

sterile water alone, eliminating the absolute need for introduction of a co-solvent into cell

culture.

A

2 4 5

15

B

Fold Induction

4.00 ...---------------------------.......

3.S03.00 ­

2.S0

2.00

1.S0

1.00O.SO0.00

dAST 10-SH20

dAST 10-6H20

dAST 10-7H20

dAST 10-SEtOH

EtOH

Fig. 3. Western blot demonstrating Cx43 protein expression in IOTl/2 cells treated for four days with two

different formulations ofracemic dAST. (A) Lanes 1-3: racemic dAST in water at IO-s, 10-6, and 10-7 M,

respectively; Lane 4: racemic dAST at 10-s M in 50% EtOH; Lane 5: 50% EtOH solvent-only control. (B)

Relative Cx43 induction levels by racemic dAST obtained by digital analysis ofdata shown in Panel A;

Cx43 protein levels in 50% EtOH solvent control set to 1.0. The upper immunoreactive bands in A are

believed to represent phosphorylated forms ofthe protein assembled into gap junctions; lower bands

unphosphorylated proteins [34]. Figure is representative of three separate replicates. Immunoblot stained

with Coomassie blue and confirmed by densitometry for equal protein loading and transfer.

Racemic dAST and the purified stereoisomers 38,3'8 dAST, 3R,3'S (meso) dAST,

and 3R,3'R dAST were also added to cell cultures in a formulation of33% EtOH at I x

10-5 M. The potent chemopreventive retinoids TTNPB at 10-8 M in acetone (0.1 % fmal

acetone concentration in assay) and retinol acetate ("RetAC") at 10-5 M in acetone (0.1 %

final acetone concentration in assay) were included as positive controls. All compounds

induced increased expression ofCx43 in comparison to cell cultures treated with 33%

16

EtOH as a solvent-only control (Fia. 4). TreatIneaM with racemic clAST induced the

highest level ofCx43 ofall novel derivatives tested (::::3-fold). As previously observed

with unmodified carotenoids, connexin 43 induction by the astaxanthin derivatives (::::3-

fold or less) was several-fold lower than induction levels observed with the retinoids (up

to 9-fold induction). Interestingly, in cells treated with the novel astaxanthin derivatives,

the majority of the Cx43 protein was located in the higher molecular weight regions of the

gel, whereas the majority ofconnexin 43 iDduced by the retinoids remained in the lower

regions (Fig. 4A). These higher molecular weight isoforms are believed to represent

phosphorylated forms of the protein assembled into gap junctions, whereas the lower

blinds are comprised ofnon-phosphorylated, unassembled proteins [39].

A1 2 7

-mlUI . \", GIlD,

[.'4~ '. ", . ,,~:t.

B

fold Inductieft

10.009.008.007.008.005.004.003.002.001.000.00

Media RetAC TTNPB Rac R,R S,S Meso

17

Fig.4. Western blot demonstrating induction ofCx43 protein expression in 10Tl/2 cells treated for four

days by novel astaxanthin derivatives delivered in 33% EtOH, versus positive and negative controls. (A)

Lane 1: 33% EtOH (final EtOH concentration in assay 0.3%) as solvent-only control. Lane 2: TTNPB at

10-8 M in acetone (final acetone concentration in assay 0.1%) as positive control. Lane 3: Retinol acetate

("RetAC") at 10-5 M in acetone (final acetone concentration in assay 0.1 %) as positive control. Lane 4:

racemic clAST at 10-5 M; Lane 5: 3R,3'R clAST at 10-5 M; Lane 6: 38,3'8 clAST at 10-5 M; Lane 7: meso

clAST at 10-5 M. (B) Relative induction levels ofCx43 expression versus solvent-only control, as in Fig.

3B. Figure is representative ofthree separate replicates. Racemic clAST at 10-5 M in 33% EtOH

demonstrates greatest induction ofCx43 protein levels (> 3-fold), similar to that obtained with racemic

dAST at 10-5 M in 50% EtOH vehicle (Fig. 3B). Figure is representative of three separate replicates.

Immunoblot stained with Coomassie blue and confirmed by densitometry for equal protein loading and

transfer.

Separate treatments of the 10TI12 cells were also performed with the goal of

comparing the relative biological efficacy of the racemic pAST in a 20% EtOH

formulation to enhance connexin 43 expression as compared to the parent carotenoid

astaxanthin, the related xanthophyll canthaxanthin and the retinoid control. Cells were

treated with racemic pAST in 20% EtOH (0.2% final EtOH cone.), AST in THF (0.1 %

final THF cone.), and CTX in THF (0.1 % final THF cone.) at 10-5, 10-6

, and 10-7 M.

TTNPB in acetone at 10.8 M was used as a positive control and media without compound

was used as a negative control. Racemic pAST caused induction of Cx43 in comparison

with non-treated controls in a dose-dependent manner at concentrations of 10-6 and 10-7

M, although Cx43 induction was not observed at 10-5 M (Fig. 5). Cx43 induction by the

astaxanthin derivatives (::::3.5-fold or less) was several-fold lower than induction levels

observed with the retinoid (up to 12-fold induction). CTX exhibited potent induction of

Cx43 in comparison with non-treated controls in a dose-dependent manner, with strong

18

induction observod at 10-5 M (::::7-fold). NeitberpAST nor AST at 10-5 M demonstrated

sipificant Cx43 induction. This finding was corroborated by Lucifer dye transfer

studies, where treatment ofcells with pAST and AST at 10-5 M demonstrated a lack of

induced gap junctional communication (data not shown). CTX induced greater

expression 0 Cx43 at 10-6 M (:::4.5-fold as compared to pAS at 10-6 M (::::3.5-fold)

and comparable induction at 10.7 M (::::2-fold). AST exhibited very weak induction at 10­

5 and 10-6 M (::::2-fold or less), and no induction at 10-7 M.

A1 2 3 .. 5 8 7 8 9 10 11

B..--- .. __ .. ~43kD

Fold Induction

14.00 -....--------------------------,

12.00

10.00

8.00

6.00

4.00

2.00

0.00

/ ....~ ....rJ> ....f:1~ ....f:1+, ....t? ....~ ~+, ....J:' ....f:1~ ~.v¢~v¢',,~-!"/q$'

19

Fig. 5. Western blot demonstrating induction ofCx43 protein expression in IOTl/2 cells treated for four

days by novel astaxanthin derivatives, versus the related carotenoids astaxanthin and canthaxanthin and

positive and negative controls. (A) Lane 1: TTNPB at 10'8 M in acetone (final acetone concentration in

assay 0.1 %) as positive control. Lane 2: CTX at 10-5 M in THF (fmal THF concentration in assay 0.1 %).

Lane 3: : CTX at 10-6 M in THF (final THF concentration in assay 0.1%). Lane 4: CTX at 10-7 M in THF

(fmal THF concentration in assay 0.1%). Lane 5: AST at 10-5 M in THF (final THF concentration in

assay 0.1%). Lane 6: AST at 10,6 M in THF (final THF concentration in assay 0.1 %). Lane 7: AST at

10-7 M in THF (final THF concentration in assay 0.1 %). Lane 8: Racemic pAST at 10-5 Min 20% EtOH

(final EtOH concentration in assay 0.2%). Lane 9: Racemic pAST at 10-6M in 20% EtOH (final EtOH

concentration in assay 0.2%). Lane 10: Racemic pAST at 1O-7 M in 20% EtOH (final EtOH concentration

in assay 0.2%). Lane 11: Non-treated sample as negative control. (8) Relative induction levels ofCx43

expression versus non-treated control. Figure is representative of two separate replicates. Racemic pAST

at 10-6 M in 20% EtOH demonstrates greater induction ofCx43 protein levels «;::::3.5-fold) over the parent

AST carotenoid at 10-6 M (;::::2-fold), although less than the related carotenoid CTX at 10-6 M (;::::4.5-fold).

Figure is representative of two separate replicates. Immunoblot stained with Coomassie blue and

confirmed by densitometry for equal protein loading and transfer.

3.3. Cellular Analysis of Cx43

Racemic dAST increased assembly of immunoreactive Cx43 into plaques in regions

of the cell membrane in direct contact with adjacent cells. Such localization is consistent

with formation of functional gap junctions. Cells treated with TTNPB at 10.8 M in

acetone (0.1% final acetone concentration in assay) also exhibited more extensive

immunoreactive plaques than untreated cells. In solvent-only treated cultures, such

immunoreactive plaques were infrequent, and were smaller than those detected in cells

treated with racemic dAST or with TTNPB as a positive control. The frequency of these

plaques and their size is consistent with the functional differences in gap junctional

20

penneability as detected by the Lucifer Yellow dye transfer experiments (described

below), and with the degree of induction of Cx43 as detected in the immunoblot

experiments described in Section 3.2 above. Representative photomicrographs are shown

in Fig. 6.

E F

Fig. 6. Racemic dAST increases assembly ofCx43 immunoreactive junctional plaques. Confluent

cultures of lOT1/2 cells were treated for 4 days as in Figs. 3-5, and immunostained with a Cx43 antibody.

Panels: (A) racemic dAST at 10.5 M in 33% EtOH; (B) 33% EtOH as solvent-only control; (C) TTNPB

at 10-8 M in acetone (final acetone concentration in assay 0.1 %). Panels D, E, F: digital analysis of

micrographs in Panels A, B, and C, respectively, demonstrating pixels above a fixed set threshold positive

for fluorescent intensity. Yellow arrows: immunoreactive junctional plaques; red arrows: position of cell

nuclei. Note the greater number and intensity ofjunctional immunoreactive plaques in the cultures treated

with racemic dAST in comparison with solvent-only treated controls. The junctional plaques shown in

21

Panels B and E represent infrequent plaques seen in controls; most cells in these cultures were negative for

Cx43 staining, in contrast to the extensive immunoreactivity in racemic dAST- and lTNPB-treated cells.

3.4. Induction of Gap Junctional Communication

To assess the functional capacity of these junctional plaques for direct intercellular

communication, Lucifer Yellow dye microinjection studies were performed to assess the

ability of racemic clAST to enhance dye transfer among IOTI/2 cells. As discussed

above, this ability has been previously highly correlated with the ability of carotenoids to

inhibit carcinogen-induced neoplastic transformation [15;16]. Racemic clAST in 33%

EtOH formulation at a concentration of I x 10-5 M effectively increased the extent of

junctional communication over that seen in solvent-only treated controls. Of 22

microinjected treated cells, IS (56%) were functionally coupled by gap junctions in

contrast to only 3 out of II (27%) solvent-only treated control cells. These differences

were statistically significant (P < 0.03; Fisher's Exact test). Representative

photomicrographs are shown in Fig. 7.

A

D

B

E

22

c

F

Fig. 7. Racemic dAST enhances functional GJIC in 1OTl/2 cells as shown by transfer of microinjection

dye. Confluent cultures were treated for 4 days as described in Fig. 6, then assayed for the ability to transfer

the fluorescent dye Lucifer Yellow. Arrows indicate the cell injected with Lucifer Yellow. Panels: (A)

racemic dAST at 10-5 M in 33% EtOH; (B) 33% EtOH solvent-only control; (C) TTNPB at 10-8 M in

acetone (final acetone concentration in assay 0.1%). Panels D, E, F: digital analysis ofmicrographs in

Panels A, B, and C, respectively, demonstrating pixels above a set threshold positive for Lucifer Yellow

fluorescence. Because cell nuclei have the most volume, they accumulate the most Lucifer Yellow and

exhibit the greatest fluorescence. Racemic dAST-treated cells demonstrate significantly increased

functional coupling (56%) over solvent-only treated controls (27%); P < 0.03, Fisher's Exact test.

3.5. Inhibition of MeA-induced Neoplastic Transformation

The Bertram lab has previously reported that carotenoids such as canthaxanthin

and ~-carotene act as potent inhibitors of transformation when added in the postinitiation

phase of carcinogenesis [36]. Here we examine the ability of the pAST derivative and its

parent carotenoid moiety astaxanthin to similarly inhibit carcinogen-induced neoplastic

transformation in a dose-dependent manner. As per Bertram lab protocols, compounds

were added in all cases 7 days after the removal of the carcinogen so as not to potentially

interfere with the production of carcinogen-initiated cells. Treatments were continued at

weekly intervals after re-feeding with fresh media 24 hours after MCA treatment.

Results are presented as the average number of foci per dish (Fig. 8). Cells treated with

pAST and AST at 10-5 M demonstrated highly significant inhibition of transformation (P

< 0.0001), however visualization of the cell monolayers for both treatments by

miscroscopy revealed regions of incomplete monolayer formation. This is most likely

due to cell toxicity at high concentrations, and such toxicity acts to inhibit transformation.

23

This may also account for their lack ofCx43 protein expression and induced GJIC at

these concentrations. This effect was not observed at the lower concentrations. At 10-6

M pAST demonstrated 100% inhibition of neoplastic transfonnation, and at 10-7 M pAST

exhibited significant inhibition of transformation (P > 0.04). At 10-8 M pAST did not

demonstrate significant inhibition. AST at concentrations of 10-6 M, 10-7 M, and 10-8 M

did not significantly demonstrate inhibition of transfonnation, which corroborates earlier

findings of a significant lack of Cx43 expression and induction of GJIC for all doses of

AST.

2.50

2.00

s::.fI) 1.50C:::CJ

1.000LL

0.50

0.00 -.---~-

Fig. 8. Dose-dependent effects of pAST and AST on MCA-induced neoplastic transfonnation. Results arepresented as average number of foci per dish. As a negative control, cells were treated with acetone andfollowed by media-alone treatments as per protocol. Positive controls were treated with MCA andfollowed by either media-alone or THF as a solvent control. Statistical significance was calculated bystudent's paired t-test.

24

CHAPTER 4. DISCUSSION

The accumulated body of evidence demonstrates that tumor cells exhibit a lack of

communication with surrounding normal cells. This is either a consequence of lack of

connexin gene expression, or a loss of functional gap junctional communication due to

faulty trafficking or assembly into the plasma membrane [37]. Unlike other tumor

suppressor genes, inactivating mutations of connexins are apparently rare and lack of

expression in tumors appears a consequence of increased DNA methylation of the

connexin promoter [38]. While fully established neoplastic cells can be expected to be

non-responsive to carotenoid or retinoid treatments, pre-neoplastic cells may retain some

responsiveness. Indeed, in the lOTl/2 cell system, isolated carcinogen-initiated cells

were shown to be less junctionally coupled than normal cells, yet able to respond to

retinoids with increased GJIC [39]. In contrast, once transformation has occurred,

retinoids--and one would expect carotenoids as well--are not able to induce junctional

communication between these and surrounding normal cells, nor suppress their ability to

proliferate and form transformed foci [11]. A similar situation may exist clinically. The

Bertram lab has shown that in cervical dysplasia and leukoplakia ofthe oral cavity, Cx43

expression is strongly downregulated in these pre-neoplastic lesions induced by HPV

infection and tobacco exposure respectively, in contrast to its strong expression in normal

epithelium [40;41]. In the cervix, retinoic acid, and in the oral cavity retinoic acid and f3-­

carotene, have separately been shown to have chemopreventative activity. In this context

it is of interest that retinoic acid only suppresses oral tumors after a delay of six months

to one year. This suggests a lack of effect on initially latent, fully transformed cells and

25

instead an inhibition of progression of pre-neoplastic cells [42;43]. In the Bertram model

of gap junctional communication and proliferation control, the restoration of OJIC by

retinoids or carotenoids would lead to suppression of aberrant proliferation and inhibition

of the development of cells bearing multiple mutations leading to full carcinogenic

potential.

Unlikecarotenoids, the use of retinoids as chemopreventative agents is limited by

toxicity. ~-carotene, a pro-vitamin A carotenoid whose consumption in foods and

presence in serum strongly correlates in epidemiological studies with decreased lung

cancer risk in smokers, has been evaluated most thoroughly. In two interventional

studies of individuals at high risk of lung cancer as a consequence of tobacco and/or

asbestos exposure, supplemental ~-carotene at levels approximately 10-fold higher than

found in a "healthy" diet, actually increased lung cancer rates over placebo controls

[42;43]. However, in a third study oflow-risk physicians, few of whom were smokers, ~­

carotene did not influence lung cancer rates [44]. In a fourth study, also in current and

former smokers, ~-carotene tended to decrease cancers of the oral cavity but increase

those ofthe lung [45]. This suggests a direct interaction between tobacco smoke and~­

carotene. This was confirmed in animal studies as in which ~-carotene or its breakdown

products were shown to enhance smoke-induced pathological changes and to interfere

with retinoic acid signaling [32]. Such interactions appear unlikely with non-pro-vitamin

A carotenoids; a conclusion confirmed in a follow-up study. Here lycopene replaced~­

carotene which resulted in protection against smoke-induced lung pathology without

associated toxicity [46].

Carotenoids may act as effective antioxidants at low oxygen tension and low

26

physiologic concentration, in comparison to other chain-breaking antioxidants such as

vitamin E. Certain carotenoids may be pro-oxidant under high oxygen tension and high

physiologic concentration environments (e.g. p-carotene) [47;48]. Therefore,

supplementation of p-carotene at high levels could add an additional oxidative stress

insult (particularly in the lung) to these individuals. In contrast, astaxanthin has been

described as a "class three" antioxidant; that is an excellent antioxidant that effectively

quenches excited molecular states as well as ground state radicals, without evidence of

pro-oxidant activity [48]. Ifpro-vitamin A activity, retinoid toxicity, and pro-oxidant

behavior are subsequently definitively implicated in the increased lung cancer rates in

supplemented human populations, then the astaxanthin derivatives described here may be

particularly well suited for clinical cancer chemoprevention. The influence on GJIC

demonstrated in the current study suggests that novel astaxanthin derivatives with

increased bioavailability may be especially useful chemopreventative agents.

Lycopene, a non-pro-vitamin A carotenoid found in tomatoes, may also have both

therapeutic and preventive properties. Epidemiological evidence suggests that tomato

consumption is protective against prostate cancer [49]. When administered as a

supplement to men scheduled for radical prostatectomy after diagnosis of prostate cancer,

a dose of 30 mg/day for approximately three weeks apparently decreased indices of

neoplasia in excised tumors and increased expression of connexin 43 [50]. When

lycopene was administered in conjunction with a-tocopherol to patients at high-risk of

liver cancer as a consequence of infection with hepatitis C, this combination reduced the

incidence of hepatocellular carcinoma after a delay of approximately one year [51].

Carotenoids have been shown to enhance antibody production, increase T-helper

27

cell numbers, T- and B-cell proliferation and even to increase the cytotoxicity ofnatural

killer cells [52]. Experimental evidence on tumor immunity suggests that ~-carotene's

suppressive effects on tumor growth may be attributed to enhanced production of tumor

specific antigens [53]. Astaxanthin has also demonstrated the ability to inhibit tumor

formation in the livers of mice through enhanced T cell-mediated immune response

[54;55]. While enhanced immune function may playa role during tumorigenesis in the

whole animal, cell culture studies demonstrating inhibition of cancer during the

promotion phase ofcarcinogenesis suggest that it is unlikely to be the central mechanism

ofaction. Rather, the current body ofevidence suggests that induction of Cx43

expression and increase in GJIC is the major mechanism by which initiated cells prevent

transformation. In the current study, astaxanthin did not significantly induce Cx43

expression, increase GJIC, or prevent carcinogen-induced neoplastic transformation, in

contrast to the novel compounds derived from astaxanthin. This apparent discrepancy

may be attributed to differences in uptake or cleavage of the derivatives as compared to

the parent astaxanthin moiety. Further studies need to be performed to elucidate the

mechanism behind this discrepancy. It is noteworthy that the diphosphate derivative was

demonstrated to have enhanced biological efficacy over the parent carotenoid despite the

precipitation of the compound in solution 24 hours post treatment.

The availability of a water-soluble, highly bioavailable derivative of astaxanthin

should facilitate the evaluation of its cancer chemopreventative properties in

experimental animals, and if successful, will simplify its evaluation as a

chemopreventative agent in the clinic. The demonstration that novel astaxanthin

derivatives are able to upregulate the expression ofCx43, increase the size and number of

28

Cx43 immunoreactive gap junctional plaques, functionally increase GIIC and inhibit

carcinogen-induced neoplastic transformation with enhanced activity over the parent

carotenoid suggests that these derivatives will have potent chemopreventive properties in

vivo. In view of this potential, further cell culture and animal studies utilizing these

derivatives are warranted.

29

APPENDIX

Hix, L.M., Lockwood, S.P., and Bertram, J.S. (2004) Upregulation of Connexin 43Protein Expression and Increased Gap Junctional Communication by Water SolubleDisodium Disuccinate Astaxanthin Derivatives. Cancer Letters, in press.

30

REFERENCES

1. Doll,R. and Peto,R. (1981) The causes ofcancer. Oxford University Press.

2. Britton,G. (1995) Structure and properties of carotenoids in relation to function.FASEB J., 9, 1551-1558.

3. Bertram,J.S. (1999) Carotenoids and gene regulation. Nutrition Reviews, 57, 182­191.

4. Reznikoff,C.A., Bertram,J.S., Brankow,D.W., and Heidelberger,C. (1973)Quantitative and qualitative studies of chemical transformation of cloned C3Hmouse embryo cells sensitive to postconfluence inhibition of cell division. CancerRes., 33, 3239-3249.

5. Bertram,J.S., Hossain,M.Z., Pung,A., and Rundhaug,lE. (1989) Development of invitro systems for chemoprevention research. Prev.Med., 18, 562-575.

6. Bertram,J.S. (1999) Carotenoids and gene regulation. Nutrition Reviews, 57, 182­191.

7. Loewenstein,W.R. (1979) Junctional intercellular communication and the control ofgrowth. Biochim Biochimica et Biophysica Acta, 560, 1-65.

8. Mehta,P.P., Bertram,lS., and Loewenstein,W.R. (1986) Growth inhibition oftransformed cells correlates with their junctional communication with normal cells.Cell, 44, 187-196.

9. Neveu,M. and Bertram,J.S. (2000) Gap junctions and neoplasia. In Hertzberg,E.L.and Bittar,E.E. (eds.) Gap Junctions. JAI Press, Greenwich, CT, pp 221-62.

10. Aust,O., Ale-Agha,N., Zhang,L., Wollersen,H., Sies,H., and Stahl,W. (2003)Lycopene oxidation product enhances gap junctional communication. FoodChem.Toxicol., 41, 1399-1407.

11. Bertram,J.S., King,TJ., Fukushima,L., and Khachik,F. (2000) Enhanced activity ofan oxidation product of lycopene found in tomato products and human serumrelevent to cancer prevention. In Sen,K.S., Sies,H., and Baeuerle,P.A. (eds.)Antioxidant and redox regulation ofgenes. Academic Press, San Diego, pp 410-24.

12. Mehta,P.P., Bertram,lS., and Loewenstein,W.R. (1989) The actions ofretinoids oncellular growth correlate with their actions on gap junctional communication. J. CellBioi, 108, 1053-1065.

31

13. Hossain,M.Z., Wilkens,L.R, Mehta,P.P., Loewenstein,W.R, and Bertram,J.S.(1989) Enhancement of gap junctional communication by retinoids correlates withtheir ability to inhibit neoplastic transformation. Carcinogenesis, 10, 1743-1748.

14. Hanusch,M., Stahl,W., Schulz,W.A., and Sies,H. (1995) Induction of gap junctionalcommunication by 4-oxoretinoic acid generated from its precursor canthaxanthin.Arch.Biochem.Biophys., 317, 423-428.

15. Acevedo,P. and Bertram,J.S. (1995) Liarozole potentiates the cancerchemopreventive activity of and the up-regulation of gap junctional communicationand connexin43 expression by retinoic acid and ~-carotene in 1OTl/2 cells.Carcinogenesis, 16, 2215-2222.

16. Zhang,L.-X., Cooney,RV., and Bertram,J.S. (1991) Carotenoids enhance gapjunctional communication and inhibit lipid peroxidation in C3HIlOTl/2 cells:Relationship to their cancer chemopreventive action. Carcinogenesis, 12, 2109­2114.

17. Zhang,L.-X., Cooney,RV., and Bertram,J.S. (1992) Carotenoids up-regulateconnexin43 gene expression independent of their pro-vitamin A or antioxidantproperties. Cancer Res., 52, 5707-5712.

18. Hossain,M.Z., Zhang,L.-X., and Bertram,J.S. (1993) Retinoids and carotenoidsupregulate gap junctional communication: correlation with enhanced growth controland cancer prevention. In Hall,J.E., Zampighi,G.A., and Davies,RM. (eds.) Progessin Cell Research Vol. 3: Gap Junctions. Elsiever, Amsterdam, pp 301-9.

19. Tso, M. O. and Lam, T. T. Method of retarding and ameliorating central nervoussystem and eye damage. U.S. Patent 5,527,533. 1996.

20. Cardounel,AJ., Dumitrescu,c., Zweier,J.L., and Lockwood,S.F. (2003) Directsuperoxide anion scavenging by a disodium disuccinate astaxanthin derivative:Relative efficacy of individual stereoisomers versus the statistical mixture ofstereoisomers by electron paramagnetic resonance imaging.Biochem.Biophys.Res.Commun., 307, 704-712.

21. Miki,W. (1991) Biological Functions and Activities ofAnimal Carotenoids. Pureand Applied Chemistry, 63, 141-146.

22. Tanaka,T., Morishita,Y., Suzui,M., Kojima,T., Okumura,A., and Mori,H. (1994)Chemoprevention ofmouse urinary bladder carcinogenesis by the naturallyoccurring carotenoid astaxanthin. Carcinogenesis, 15, 15-19.

23. Tanaka,T., Makita,H., Ohnishi,M., Mori,H., Satoh,K., and Hara,A. (1995)Chemoprevention of rat oral carcinogenesis by naturally occurring xanthophylls,astaxanthin and canthaxarathin. Cancer Research, 55, 4059-4064.

32

24. Kurihara,H., Koda,H., Asami,S., Kiso,Y., and Tanaka,T. (2002) Contribution of theantioxidative property of astaxanthin to its protective effect on the promotion ofcancer metastasis in mice treated with restraint stress. Life Sci., 70, 2509-2520.

25. Jyonouchi,H., Sun,S., Iijima,K., and Gross,M.D. (2000) Antitumor activity ofastaxanthin and its mode ofaction. Nutr.Cancer, 36,59-65.

26. Cooney,R.V., Kappock,T.J., Pung,A., and Bertram,J.S. (1993) Solubilization,cellular uptake, and activity of ~-caroteneand other carotenoids as inhibitors ofneoplastic transformation in cultured cells. Methods in Enzymology, 214, 55-68.

27. Cardounel,A.J., Dumitrescu,c., Zweier,J.L., and Lockwood,S.F. (2003) Directsuperoxide anion scavenging by a disodium disuccinate astaxanthin derivative:Relative efficacy of individual stereoisomers versus the statistical mixture ofstereoisomers by electron paramagnetic resonance imaging.Biochem.Biophys.Res. Commun. , 307, 704-712.

28. Zsila,F., Simonyi,M., and Lockwood,S.F. (2003) Interaction of the disodiumdisuccinate derivative of meso-astaxanthin with human serum albumin: from chiralcomplexation to self-assembly. Bioorg.Med Chem.Lett. , 13,4093-4100.

29. Cardounel,A.J., Dumitrescu,C., Zweier,J.L., and Lockwood,S.F. (2003) Directsuperoxide anion scavenging by a disodium disuccinate astaxanthin derivative:Relative efficacy of individual stereoisomers versus the statistical mixture ofstereoisomers by electron paramagnetic resonance imaging.Biochem.Biophys.Res.Commun., 307, 704-712.

30. Rogers,M., Berestecky,J.M., Hossain,M.Z., Guo,H.M., Kadle,R, Nicholson,B.J.,and Bertram,lS. (1990) Retinoid-enhanced gap junctional communication isachieved by increased levels of connexin 43 mRNA and protein. Mol. Carcinog., 3,335-343.

31. Bertram,lS., Rundhaug,J.E., and Pung,A. (1990) Carotenoids inhibit chemically­and physically-induced neoplastic transformation during the post-initiation phase ofcarcinogenesis. In Prasad,K.N. and Meyskens,F.L. (eds.) Nutrients and cancerprevention. Humana Press, Clifton, pp 99-111.

32. Bertram,J.S., Pung,A., Churley,M., Kappock,T.J., Wilkins,L.R, and Cooney,RV.(1991) Diverse carotenoids protect against chemically induced neoplastictransformation. Carcinogenesis, 12, 671-678.

33. Simonyi,M., Bikadi,Z., Zsila,F., and Deli,J. (2003) Supramolecular exciton chiralityofcarotenoid aggregates. Chirality, 15, 680-698.

34. Zsila,F., Simonyi,M., and Lockwood,S.F. (2003) Interaction of the disodiumdisuccinate derivative of meso-astaxanthin with human serum albumin: from chiralcomplexation to self-assembly. Bioorg.MedChem.Lett., 13,4093-4100.

33

35. Bikadi,Z., Zsila,F., Deli,l, Mady,G., and Simonyi,M. (2002) The supramolecularstructure of self-assembly formed by capsanthin derivatives. Enantiomer., 7, 67-76.

36. Cardounel,Al, Dumitrescu,C., Zweier,J.L., and Lockwood,S.F. (2003) Directsuperoxide anion scavenging by a disodium disuccinate astaxanthin derivative:Relative efficacy of individual stereoisomers versus the statistical mixture ofstereoisomers by electron paramagnetic resonance imaging.Biochem.Biophys.Res.Commun., 307, 704-712.

37. Salares, V. R, Young, N. M., Carey, P. R, and Bernstein, H. J. Excited state(exciton) interactions in polyene aggregates. Journal of Raman Spectroscopy 6,282-288. 1977.

38. Saez,J.C., Martinez,AD., Branes,M.C., and Gonzalez,H.E. (1998) Regulation ofgap junctions by protein phosphorylation. Braz.J Med. Bioi. Res. , 31,593-600.

39. Bertram,J.S. (1985) Neoplastic Transformation in Cell Cultures: In Vitro/In-VivoCorrelations. fARC Sci. Pub. , 67, 77-91.

40. Bertram,J.S., Pung,A., Churley,M., Kappock,T.J., Wilkins,L.R, and Cooney,R.V.(1991) Diverse carotenoids protect against chemically induced neoplastictransformation. Carcinogenesis, 12,671-678.

41. Vine,AL. and Bertram,1.S. (2002) Cancer chemoprevention by connexins. CancerMetastasis Rev., 21, 199-216.

42. Piechocki,M.P., Burk,RD., and Ruch,RJ. (1999) Regulation ofconnexin32 andconnexin43 gene expression by DNA methylation in rat liver cells. Carcinogenesis,20,401-406.

43. Mordan,L.1. and Bertram,lS. (1983) Retinoid effects on cell-cell interactions andgrowth characteristics ofnormal and carcinogen-treated C3H/I0Tl/2 cells. CancerResearch, 43, 567-571.

44. Sakr, W., Tabaska, P., Kucuk, 0., and Bertram, 1. S. Differential expression ofconnexin 43 in normal, preneoplastic and neoplastic squamous epithelium of theupper aerodigestive tract. Proc.AACR 37,269. 1996.

45. King,T.J., Fukushima,L.H., Hieber,AD., Shimabukuro,K.A., Sakr,W.A., andBertram,J.S. (2000) Reduced levels of connexin43 in cervical dysplasia: inducibleexpression in a cervical carcinoma cell line decreases neoplastic potential withimplications for tumor progression. Carcinogenesis, 21, 1097-1109.

46. Omenn,G.S., Goodman,G.e., Thornquist,M.D., Balmes,J., Cullen,M.R, Glass,A,Keogh,lP., Meyskens,F.L., Jr., Valanis,B., Williams,J.H., Jr., Barnhart,S., andHammar,S. (1996) Effects of a combination of beta carotene and vitamin A on lungcancer and cardiovascular disease. New England Journal ofMedicine, 334, 1150­1155.

34

47. Heinonen,O.P. and Albanes,D. (1994) The effect of vitamin E and beta carotene onthe incidence of lung cancer and other cancers in male smokers. New EnglandJournal o/Medicine, 330, 1029-1035.

48. Omenn,G.S., Goodman,G.e., Thornquist,M.D., Balmes,J., Cullen,M.R., Glass,A.,Keogh,J.P., Meyskens,F.L., Jr., Valanis,B., Williams,J.H., Jr., Barnhart,S., andHammar,S. (1996) Effects ofa combination of beta carotene and vitamin A on lungcancer and cardiovascular disease. New England Journal ofMedicine, 334, 1150­1155.

49. Hennekens,C.H., Buring,J.E., Manson,J.E., Stampfer,M., Rosner,B., Cook,N.R.,Belanger,C., LaMotte,F., Gaziano,J.M., Ridker,P.M., Willett,W., and Peto,R.(1996) Lack of effect of long-term supplementation with beta carotene on theincidence of malignant neoplasms and cardiovascular disease. New EnglandJournal o/Medicine, 334, 1145-1149.

50. Mayne,S.T., Cartmel,B., Baum,M., Shor-Posner,G., Fallon,B.G., Briskin,K.,Bean,J., Zheng,T.Z., Cooper,D., Friedman,C., and Goodwin,W.J., Jr. (2001)Randomized trial of supplemental ~-carotene to prevent second head and neckcancer. Cancer Research, 61, 1457-1463.

51. Liu,C., Lian,F., Smith,D.E., Russell,R.M., and Wang,XD. (2003) Lycopenesupplementation inhibits lung squamous metaplasia and induces apoptosis via up­regulating insulin-like growth factor-binding protein 3 in cigarette smoke-exposedferrets. Cancer Res., 63, 3138-3144.

52. Burton,G.W. and Ingold,K.U. (1984) beta-Carotene: an unusual type oflipidantioxidant. Science, 224,569-573.

53. Martin, H. D., Jager, C., Ruck, C., Schmidt, M., Walsh, R., and Paust, J. Anti- andprooxidant properties of carotenoids. J.Prakt.Chem. 341, 302-308. 2004.

54. Martin, H. D., Jager, C., Ruck, C., Schmidt, M., Walsh, R., and Paust, J. Anti- andprooxidant properties ofcarotenoids. J.Prakt.Chem. 341, 302-308. 2004.

55. Giovannucci,E. (2002) A review ofepidemiologic studies oftomatoes, lycopene,and prostate cancer. Expo BioI. Med. (Maywood.), 227, 852-859.

56. Kucuk,O., Sarkar,F.H., Sakr,W., Djuric,Z., Pollak,M.N., Khachik,F., Li,Y.W.,BaneIjee,M., Grignon,D., Bertram,J.S., Crissman,J.D., Pontes,E.J., and Wood,D.P.,Jr. (2001) Phase II randomized clinical trial oflycopene supplementation beforeradical prostatectomy. Cancer EpidemioloBiomarkers Prevo, 10, 861-868.

57. Nishino, H. and Jinno, K. Prevention of hepatitis C virus-induced liver cancer bynatural carotenoids and vitamin E. Proceedings ofFrontiers in Cancer PreventionResearch, AACR Conference (Oct. 14-18), 107.2002. Boston.

35

58. Bendich,A. (1991) Beta-carotene and the immune response. Proc.Nutr.Soc., 50,263-274.

59. Tomita,Y., Himeno,K., Nomoto,K., Endo,H., and Hirohata,T. (1987) Augmentationof tumor immunity against syngeneic tumors in mice by beta- carotene.JNatl.Cancer Inst., 78, 679-681.

60. Kurihara,H., Koda,H., Asami,S., Kiso,Y., and Tanaka,T. (2002) Contribution of theantioxidative property of astaxanthin to its protective effect on the promotion ofcancer metastasis in mice treated with restraint stress. Lift Sci., 70,2509-2520.

61. Jyonouchi,H., Sun,S., Iijima,K., and Gross,M.D. (2000) Antitumor activity ofastaxanthin and its mode of action. Nutr.Cancer, 36,59-65.

36