Proc. Nati. Acad. Sci. USAVol. 89, pp. 1785-1789, March 1992Medical Sciences

Selective tumor uptake of a boronated porphyrin in an animalmodel of cerebral glioma

(brain tumor/boron/henatoporphyrin/neutron capture/photodynamic therapy)

JOHN S. HILL*t, STEPHEN B. KAHL*, ANDREW H. KAYE*, STANLEY S. STYLLI*, MYOUNG-SEO Koot,MICHAEL F. GONZALES§, NICHOLAS J. VARDAXISI, AND CHRISTOPHER I. JOHNSON"I*Higginbotham Neuroscience Research Institute, Department of Surgery, and §Department of Anatomical Pathology, Royal Melbourne Hospital, University ofMelbourne, Victoria, Australia, 3050; tDepartment of Pharmaceutical Chemistry, School of Pharmacy, University of California, San Francisco,CA 94143-0446; IDepartment of Applied Sciences, Philip Institute of Technology, Bundoora, Victoria, Australia 3083; and IlBio-Rad Pty. Ltd.,Hawthorn, Victoria, Australia 3122

Communicated by M. Frederick Hawthorne, November 12, 1991

ABSTRACT The prognosis for patients with high-gradecerebral glioma is poor. Most treatment failures are due to localrecurrence of tumor, indicating that a more aggressive localtherapy could be beneficial. Adjuvant treatments such asporphyrin-sensitized photodynamic therapy (PDT) or boronneutron capture therapy (BNCT) have the potential to controllocal recurrence. The selective tumor uptake of a boronatedporphyrin was studied in CBA mice bearing an implantedintracerebral glioma. Biopsy samples of tumor, normal brain,and blood were analyzed by a fluorometric assay followingintraperitoneal and intravenous administration of boronatedprotoporphyrin (BOPP). This compound was selectively local-ized to tumor at ratios as high as 400:1 relative to normal brain.Confocal laser scanning microscopy of glioma cells in vitro andin vivo showed that BOPP was localized within mitochondriaand excluded from the nucleus of these cells. This discretesubcellular localization was confirmed by density gradientultracentrifugation after homogenization of mouse tumor bi-opsies. The selective discrete localization of these compoundswithin the tumor suggests that this compound may be used asa dual PDT/BNCT sensitizer.

Primary cerebral tumors are responsible for -2% of allcancer deaths, with =10,000 persons dying per annum in theUnited States (1). The majority of these deaths are due to thehigh-grade gliomas-anaplastic astrocytoma and glioblas-toma multiforme. At present there is no satisfactory treat-ment for these tumors. Surgery provides a definitive histo-logical diagnosis and relief of symptoms of raised intracranialpressure. Radiotherapy and adjuvant chemotherapy are oflimited value and most studies utilizing these treatmentsreport median survival times of <1 year (1, 2). Most treat-ment failures are due to local recurrence of the tumor,suggesting that more aggressive local therapy could be ben-eficial. Two adjuvant therapies with the potential to controllocal recurrence are photodynamic therapy (PDT) and boronneutron capture therapy (BNCT).PDT relies on the selective uptake or retention of a

photosensitizing chemical in the tumor relative to surround-ing normal tissue, followed by treatment with light of theappropriate wavelength to activate the photosensitizer (3).The photoactivation of this sensitizer results in generation ofa cytotoxic chemical species, probably singlet excited stateoxygen, which leads to selective tumor necrosis (3). Photo-sensitizers that have been used in most clinical and experi-mental studies to date are hematoporphyrin derivative (HpD)and its enriched commercial preparation Photofrin II (3), bothof which have been shown to selectively localize in glioblas-

toma multiforme (4, 5). Reports of PDT in the treatment ofanimal (5-7) and human (5, 7-9) gliomas have been encour-aging, although the use of a more tumor-selective photosen-sitizer than HpD or Photofrin II would be desirable.Like PDT, BNCT is based on selective tumor localization

of a sensitizing agent, a compound containing 10B atoms,followed by activation ofthe sensitizer. However, in contrastto PDT, the activating beam of thermal neutrons can reachdeep-seated tumor sites. The high linear energy transferparticles produced in the reaction, 4He and 7Li, have meanfree paths of9 and 5 ,um, respectively, and deposit substantialenergies within approximately one cell diameter. The shortrange of these products results in severe dependency of thecells on boron microlocalization. For a typical cell, it iscalculated that 90% less boron is needed if it is locatedintracellularly as opposed to extracellularly (10). Thus, bothPDT and BNCT are binary treatments, the individual com-ponents of which are nontumoricidal but when combinedhave tumoricidal potential. While previous studies havesuggested that BNCT might be a useful treatment for cerebraltumors (11), data obtained from patients and animals injectedwith a variety of boron-containing compounds indicate thatthese chemicals do not exhibit high tumor selectivity relativeto surrounding normal brain (12, 13). Porphyrins present apotential solution to the problems of sensitizer uptake en-countered in the application of BNCT for cerebral gliomasince they have a much higher tumor selectivity and retention(4) than Na2B12H11SH, the agent currently used, and have ahigh carrying capacity for boron (14, 15). To that end, severalboronated porphyrins have been synthesized and we reporthere on the pharmacokinetics of one of these compounds,known as BOPP, in mice bearing the C6 intracerebral gliomaxenograft (16). BOPP is a tetrakiscarborane carboxylate esterof 2,4-(a,4-dihydroxyethyl)deuteroporphyrin IX (14). Thefour closo-carborane cages, bearing 10 boron atoms each,provide BOPP with a boron weight percentage of nearly 30o.The compound is highly water soluble and is stable tophysiologic conditions (17). Unlike HpD, BOPP is a singularchemical entity and does not aggregate in aqueous solution.

MATERIALS AND METHODSMaterials. All reagents used were of analytical grade or

better. Hepes was obtained from Sigma, and cetyltrimethyl-ammonium bromide (CTAB), sucrose, and K+EDTA werefrom BDH. Chloroform and methanol were purchased fromMillipore and Dulbecco's phosphate-buffered saline (PBS)

Abbreviations: HpD, hematoporphyrin derivative; PDT, photody-namic therapy; BNCT, boron neutron capture therapy; BOPP,2,4-(a,,-dihydroxyethyl)deuteroporphyrin IX tetrakiscarborane car-boxylate ester; CTAB, cetyltrimethylammonium bromide; CLSM,confocal laser scanning microscopy.tTo whom reprint requests should be addressed.

1785

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Apr

il 19

, 202

1

Proc. Natl. Acad. Sci. USA 89 (1992)

(pH 7.3) was from Commonwealth Serum Laboratories(Parkville, Australia). Percoll was purchased from Pharma-cia-LKB and stored at 40C. BOPP was synthesized as de-scribed (14) and dissolved in 0.9o NaCl at a concentration of10 mg/ml. This stock solution was stored at 40C in the darkand was routinely used within 24 hr.

Cells. The C6 rat glioma cell line was obtained from theAmerican Type Culture Collection and the cells were grownat 370C in monolayer culture in RPMI 1640 medium supple-mented with 10% fetal calf serum (FCS), both obtained fromCommonwealth Serum Laboratories. Cells were harvestedduring the logarithmic phase of growth.

Intracranial Implantation of C6 Glioma. The C6 gliomamodel was grown as a xenograft in adult CBA mice asdescribed (16).

Administration of BOPP. The boronated porphyrin BOPPwas dissolved in 0.9o NaCl before either intraperitoneal(i.p.) or intravenous (i.v.) tail vein administration into CBAmice bearing the intracranial C6 glioma. To study the kineticsof uptake, mice were injected with BOPP solution at theappropriate time points such that all were sacrificed exactly14 days postimplantation of the tumor cells. To study thedose dependence of uptake, mice were injected with theappropriate dose of BOPP 13 days postimplantation of thetumor cells and were sacrificed 24 hr later.Measurement of BOPP Uptake. On sacrifice, triplicate

biopsy samples of tumor, normal brain, and venous bloodwere taken from all mice. Typically, 40-80 mg of tissue and100 ul of blood were sampled. The level of BOPP wasdetermined by an adaptation of a published method (4).Briefly, samples of known wet weight (or 100-pA volume inthe case of blood) were homogenized in 6 ml of a solutioncontaining 50 mM Hepes (pH 7.4), 10 mM CTAB using anYstral type X1020 homogenizer (H.D. Scientific, Bayswater,Australia). Triplicate 2-ml aliquots of the homogenate wereremoved, and each was mixed with 5 ml of a chloroform/methanol mixture (1:1; vol/vol), thoroughly Vortex mixed,and centrifuged at room temperature for 5 min at 2000 x g ina bench centrifuge (Clements GS200; Selby-Anax, Notting-hill, Australia). The upper phase and a layer of cell debris atthe interface between upper and lower phases were dis-carded, and the lower phase was retained. All extractedBOPP was present in the lower phase, with no detectableporphyrin in the upper phase. Similarly, no BOPP wasdetectable in the debris layer after reextraction. The chloro-form-rich lower phase was then evaporated to dryness undera stream of N2 gas, and the resulting residue was suspendedin 1 ml of the 50 mM Hepes, pH 7.4/10 mM CTAB solution.The absorbance of the solution at 400 nm was then deter-mined with a Beckman DU65 spectrophotometer (Beckman)relative to a control blank. Those samples with higher ab-sorbance values were diluted with 50 mM Hepes, pH 7.4/10mM CTAB such that their final absorbance was equal to 0.15absorbance unit in a 10-mm path length cell. This dilution stepovercame the problem of concentration-dependent quench-ing of the BOPP fluorescence emission by either the ex-tracted porphyrin or hemoglobin that was coextracted withBOPP from the tissue samples. A standard curve of BOPPcoextracted in the presence ofadded hemolyzate showed thathemoglobin neither enhanced nor quenched the fluores-cence, provided the absorbance measured at 400 nm in a10-mm light path was below this limiting value of 0.15absorbance unit.

Fluorescence of the samples was determined with a PerkinElmer LS-30 spectrofluorimeter with the emission and exci-tation wavelengths set at 400 and 625 nm, respectively. Thelevel of BOPP in each sample was determined by comparingthe emission at 625 nm to that of BOPP standards preparedby this procedure. A standard curve of BOPP fluorescenceanalyzed in this manner was linear over the. concentration

range 0-0.3 jig of BOPP per ml, with quenching of thefluorescence at concentrations >0.3 ug/ml.

Confocal Laser Scannng Microscopy. For analysis of BOPPuptake in vitro, a single cell suspension of C6 cells harvested inthe logarithmic phase of growth was seeded into 9-cm2 slideflasks with a detachable microscope slide base (A/S Nunc,Ramstrup, Denmark) and grown in RPMI 1640 medium sup-plemented with 10% FCS. The cells were grown to 40%oconfluence at 37CC, and then the flask was wrapped in aluminumfoil to shield the cells from light; an aliquot ofBOPP solution (2mg/ml in 0.9%o NaCl) was added such that the final concentra-tion in the flask was 20 Ag/ml. The flask was then reincubatedat 370C for a further 18 hr, then the medium was removed, andthe adherent cells were washed twice with RPMI 1640 mediumcontaining 10%o FCS to remove exogenous BOPP. The slidebase ofthe flask was then removed and a coverslip (18 x 50mm)was placed on the culture area to make a wet mount. Theintracellular BOPPwas then detected by using eitheraMRC 500orMRC 600 confocal laser scanning microscope (Bio-Rad) withexcitation from an argon ion laser at 488 and 514 nm and theemission monitored above 600 nm. For analysis of the in vivouptake, CBA mice bearing the C6 glioma xenograft wereadministered BOPP at a dose of50mg per kg ofbody weight bythe i.v. route 24 hrbefore sacrifice. After sacrifice, the brain wasremoved and placed in OCT Tissue-Tek embedding compound(Miles) and then frozen in isopentane cooled with liquid N2.Unfixed and unstained 10-,m-thick cryosections of tumor andsurrounding normal brain were then taken and placed on glassslides and sealed with a glass coverslip. The slides were thenwrapped in foil to minimize photobleaching ofBOPP until theywere analyzed on the confocal microscope.

Subcellular Fractionation. Tumor and normal brain wereremoved 24 hr after i.v. administration of BOPP (50 mg perkg of body weight). The tissues were homogenized at 4°C ina buffer containing 0.32 M sucrose, 1 mM EDTA, 10 mMHepes (pH 7.4) and the various subcellular fractions wereisolated by ultracentrifugation in a TL-100 ultracentrifugeusing a TLA-100.3 rotor (Beckman) on Percoll (Pharmacia-LKB) gradients as described (18). The level of BOPP in thesubcellular fractions from tumor and normal brain was thendetermined by extraction as described above, and the totalprotein content in each fraction was determined by themethod of Lowry et al. (19). At all times during isolation ofthe subcellular fractions, care was taken to minimize lightexposure, which could cause photobleaching of the BOPP.

RESULTSThe selective uptake of BOPP into the intracerebral C6xenograft is shown in Fig. 1 with the fluorescence emissionfrom BOPP confined to the histologically proven tumor andno detectable fluorescence in normal brain. To quantitate theBOPP levels in tumor, the kinetics and dose dependency ofuptake were determined after injection into mice bearing theC6 glioma. The methods of administration of the sensitizerswere by either i.p. or i.v. routes with porphyrin levels intumor, normal brain, and blood determined by fluorescenceassay after extraction from the tissue. This assay detects theporphyrin component of the boron porphyrin conjugate.Previous studies have shown that the chemical bonds be-tween the carborane cages and the porphyrin ring are stablein vivo (17). Thus, since these molecules are =30%o boron byweight, an estimate of the total boron present in the tissue canbe made by measurement of total porphyrin content. It isapparent that BOPP is a highly effective tumor-localizingagent when administered by either i.p. or i.v. routes. Max-imal tumor concentrations were observed after 24 hr at eitherlow (10 mg per kg of body weight; Fig. 2A) or high (100 mgper kg of body weight; Fig. 2B) doses. However, the uptakekinetics into blood were dependent on the route of adminis-tration. Maximum levels were detected 2 hr after i.v. injec-

1786 Medical Sciences: Hill et al.

Dow

nloa

ded

by g

uest

on

Apr

il 19

, 202

1

Proc. Natl. Acad. Sci. USA 89 (1992) 1787

A

20

t

10

5

l.vi.p.

..

.1. .

1O-If ...... * J0 20 40 60 80 0 20 40 60 8(

B

20-;

0 AA0 20 40 60 80 0 20 40 60 80

Time after injection, hr

40

30

20

10

0

360

300

240

180120

600

FIG. 1. BOPP uptake into the C6 glioma xenograft. (Upper)Coronal section of mouse brain through the tumor implanted in theleft cerebral hemisphere (16). The section was taken 24 hr afteradministration of BOPP (40 mg per kg of body weight) via the i.v.route and was photographed under UV light as described (20).(Lower) A section adjacent to that shown in Upper stained withhematoxylin/eosin as described (16) to show the discrete intracere-bral tumor. The operative procedure of tumor implantation andsacrifice of the animals was as described (16, 20) and was inaccordance with approved guidelines of the Royal Melbourne Hos-pital Animal Ethics Committee.

tion for both low and high doses, whereas for i.p. injectionmaximum levels were noted 24 and 18 hr after administrationfor the low and high doses, respectively. The rates ofclearance from blood were also dependent on the route ofadministration, with much more rapid clearance after i.v.injection. Tumor levels remained relatively high throughoutthe experimental time course following high dose injection byeither route, indicating the retention of BOPP in the tumor,even though blood levels fell substantially. The tumor/normal brain and tumor/blood ratios for the data in Figs. 2and 3 are shown in Table 1. It is clear that BOPP is a highlyselective tumor sensitizer. Ratios as high as 400:1 in tumorcompared to normal brain were obtained 48 hr after i.v.injection of 100 mg of BOPP per kg of body weight, while atthe same time point the tumor/blood ratio is 11:1.The dose dependence of BOPP uptake 24 hr after adminis-

tration is shown in Fig. 3 and the ratios derived from these dataare shown in Table 1. At doses >100 mg/kg, the tumor/brainand tumor/blood ratios decrease markedly, suggesting that thecirculating blood plasma levels of BOPP are so high that itsapparent selectivity is decreased. However, examination ofthese ratios at longer time points shows subsequent increasesreflecting clearance ofBOPP from plasma. Doses as high as 200mg/kg were readily tolerated with no apparent signs of mor-bidity and no mortality of animals.

FIG. 2. Kinetics of uptake of BOPP into CBA mice bearing theC6 glioma xenograft after 10 mg (A) or 100 mg (B) per kg of bodyweight injection via i.p. or i.v. tail vein routes. Solid BOPP powderwas dissolved in isotonic saline at a concentration of 0.5 mg/ml (forthe 10 mg/kg dose) and 5 mg/ml (for the 100 mg/kg dose) andadministered to 24 mice (3 mice at each time point). At the indicatedtimes after administration, the mice were sacrificed and samples oftumor (W), normal brain (A), and venous blood (v) were taken. Thefluorescence assay was performed in a Perkin Elmer LS-30 spec-trofluorimeter with the excitation and emission wavelengths set at400 and 625 nm, respectively. The level of BOPP in each sample wasdetermined by comparing the emission at 625 nm to that of BOPPstandards prepared by the same procedure. Ordinate shows the levelof BOPP uptake [expressed as jg per g of tissue (wet weight) fortumor and brain and as Itg per ml of whole blood]. The data shownrepresent mean ± 1 SD of BOPP levels determined in 3 mice at eachtime point.

The studies detailed above indicate the selectivity of up-take of BOPP into glioma but do not indicate the site(s) oflocalization within the gross tumor or the individual tumorcells. Confocal laser scanning microscopy (CLSM) was usedto define the subcellular sites ofBOPP localization in C6 cellsgrown in vitro and tumor in vivo. The micrographs (Fig. 4)show punctate fluorescence associated with discrete sites oflocalization in the perinuclear region and minimal uptakeeither into the nucleus or throughout the cytoplasm in vitro(Fig. 4A). A similar pattern was evident in frozen sections of

oa 1500tor 120

0)0) 90L.0-z&.. )

c 4- 60

a. 30a..0co 0

0 50 100 150 200 0 50 100 150 200BOPP dose, mg per kg of body weight

FIG. 3. Uptake of BOPP 24 hr after i.p. (A) and i.v. (B) administra-tion. The assay conditions are as described in Fig. 2, except that a stocksolution of BOPP (10 mg/ml in isotonic saline) was used. This was thendiluted in isotonic saline such that each mouse was injected with theappropriate dose in a vol of 0.4 ml and sacrificed after 24 hr; tissuesamples were then taken. Symbols and effors are as described in Fig. 2.

3:0)

.e0)UC

cm 15

U)

0.0D

Medical Sciences: Hill et al.

3

Dow

nloa

ded

by g

uest

on

Apr

il 19

, 202

1

1788 Medical Sciences: Hill et al.

Table 1. Tumor/normal brain and tumor/blood ratios 24 hr afteradministration of BOPP (10-200 mg per kg of body weight)

i.p. i.v.

Dose T/NB T/BL T/NB T/BL10 157 1 137 1620 220 0.93 137 1250 206 1.5 197 8.1100 202 2.0 238 6.5150 55 1.7 98 4.2200 32 1.5 48 3.5

T/NB, tumor/normal brain; T/BL, tumor/blood. All values werecalculated as the ratio of mean levels shown in Fig. 3.

C6 glioma biopsies (Fig. 4B), with BOPP detected intracel-lularly in the tumor bulk and in isolated nests of tumor cellsinfiltrating into the brain adjacent to tumor region. To definethese sites of localization, C6 tumor tissue and normal brainfrom BOPP-sensitized mice (50 mg per kg of body weight)were homogenized and subcellular fractions were isolated bydensity gradient ultracentrifugation (18). The results in Table2 demonstrate that these discrete sites of fluorescence arewithin the mitochondria, with low levels in the cytoplasm,lysosomes, and Golgi.

DISCUSSIONOur data suggest that BOPP may have considerable potentialas a dual BNCT and PDT sensitizer because of its extremelyselective tumor uptake and retention. Theoretical calcula-tions indicate that useful BNCT ofglioma may be carried out

AI

FIG. 4. CLSM of BOPP in C6 glioma cells. (A) CLSM image ofglioma cells in vitro incubated with BOPP (20 ,ug/ml) for 18 hr andwashed with medium to remove exogenous BOPP. Control images ofC6 cells not exposed to BOPP showed minimal detectable fluores-cence. (B) CLSM image of BOPP in isolated nests of tumor cellsinvading into normal brain in the brain adjacent to tumor region. Anunfixed 10-,um-thick frozen section of tumor and surrounding brainwas taken from a mouse 24 hr after i.v. BOPP administration (50 mgper kg of body weight).

Table 2. Subcellular localization of BOPP in C6 tumorSubcellular fraction Normal brain Tumor

Cytoplasm (includinglysosome and Golgi) 0.002 0.012

Nucleus 0.0005 0.0005Plasma membrane 0.001 0.004Mitochondria 0.002 0.28

Tumor and normal brain tissue were removed 24 hr after i.v.administration ofBOPP (50 mg per kg ofbody weight) and were thenhomogenized; subcellular fractions were obtained on Percoll densitygradients (18) in a Beckman TL-100 ultracentrifuge with a TLA-100.3rotor. Values are expressed as ,ug of BOPP per mg of protein. Thelevel of BOPP in each fraction was determined after extraction asdescribed in Materials and Methods and the total protein in eachsample was determined by the method of Lowry et al. (19).

with tumor concentrations of 10B >20 pug/g, provided theconcentration in the blood and parenchyma of the centralnervous system is significantly lower (13). The results pre-sented in Fig. 2 show that 24 hr after administration ofBOPP,the levels in the C6 glioma xenograft are higher than in bloodand normal brain irrespective ofwhether the i.p. or i.v. routeis used at both low or high doses. Since BOPP is 30%o boronby weight, then a BOPP concentration in tumor >65 pug/gwould result in boron levels above this threshold of 20 pug/g.The data in Fig. 2B show that 24 hr after a dose of 100 mg ofBOPP per kg of body weight, these levels are reached, withcorrespondingly low levels in blood and normal brain (Table3). However, even more importantly, the intracellular loca-tion ofBOPP described in Fig. 4 has been calculated to resultin a 10-fold reduction in the amount of boron required toachieve tumor necrosis, since there is more efficient targetingof critical subcellular structures (10). Thus, the tumor levelsreported in this study are well above this threshold even at thelow BOPP dose of 10 mg per kg of body weight.

Previous studies using slow infusion of the sulfhydrylderivative of icosahedral borane, Na2B12HjjSH, in a patientwith glioblastoma multiforme showed tumor concentrationsof2-6 ,ug ofboron per g, with parenchymal concentrations of-1 ,ug/g (13). However, Na2B12H11SH has not been shownto localize intracellularly in glioma cells, indicating that analternative sensitizer is required, since these levels are belowthe clinically useful threshold value. In contrast, tumor boronlevels of =12 ,ug/g and 63 ,ug/g using i.v. doses of 10 and 100mg ofBOPP per kg ofbody weight, respectively, are reportedin this study. In addition, the selectivity of this uptake meansthat the levels in normal tissue and blood are very low, whilethe intracellular localization suggests it will be extremelydose effective. Interestingly, the maximal tumor levels wereobserved after 24 hr at either low or high dose, irrespectiveof the route of administration. This is in contrast to thekinetics of HpD uptake in the same animal model, where

Table 3. Tumor/normal brain and tumor/blood ratios after i.p.and i.v. administration of BOPP

BOPP at 10 mg/kg BOPP at 100 mg/kg

Time,* i.p. I.V. I.p. I.V.hr T/NB T/BL T/NB T/BL T/NB T/BL T/NB T/BL2 11 0.21 138 0.12 16 0.15 43 0.056 47 0.32 64 0.25 51 1.1 31 0.1512 87 0.68 124 1.8 74 0.36 40 0.6718 116 0.85 124 3.7 158 0.55 113 1.824 157 1.1 136 15 203 2.0 238 6.548 114 1.3 203 15 253 2.6 402 1172 90 5.9 253 14 215 4.9 287 9.7

T/NB, tumor/normal brain ratio; T/BL, tumor/blood ratio. Allvalues were calculated as the ratio of mean levels shown in Fig. 2.*Time after administration of BOPP.

Proc. Natl. Acad. Sci. USA 89 (1992)

Dow

nloa

ded

by g

uest

on

Apr

il 19

, 202

1

Proc. Natl. Acad. Sci. USA 89 (1992) 1789

maximal tumor levels are observed 6 and 24 hr after admin-istration for i.v. and i.p. routes, respectively (20). The natureof the HpD mixture, which is composed of both hydrophobicand hydrophilic components (3), is probably the basis forthese differences in uptake kinetics. The fluorescence assayused to determine HpD uptake measures total porphyrincontent in the tumor (4) but does not utilize chromatographictechniques to identify the proportions of the various compo-nents present at each time point. We have previously estab-lished that a very pure hydrophilic porphyrin, porphyrin C(21), exhibits rapid uptake kinetics into the C6 tumor. Max-imal porphyrin C levels were observed 45 min and 6 hr afteri.v. and i.p. administration, respectively, with subsequentrapid clearance from the tumor (21). Conversely, BOPP,which is far more hydrophobic than porphyrin C, exhibitsmuch slower uptake and longer retention times in the tumorand in this respect is similar to the hydrophobic componentsof HpD. The kinetics of uptake and clearance of BOPP fromblood are also similar to those of HpD. Thus, the BOPPcompound appears to combine the best kinetic aspects of thevarious fractions ofHpD. The high tumor/normal brain ratiosand low blood levels suggest that the hydrophobic BOPPmolecule probably partitions into lipid-rich membraneswithin the tumor and is not present in more aqueous envi-ronments such as the tumor stroma or blood vessels. This isin contrast to HpD, which has been proposed to localize inthe tumor vasculature and have its major site of phototoxicaction at those sites (3). The tumor/normal brain ratios aremuch greater than those obtained with similar doses ofHpDin an identical animal model or after a dose of 5 mg per kg ofbody weight in human glioma patients, where maximal ratiosof 50:1 were noted (4). The data in Figs. 2 and 3 and Table 1also show that while maximal BOPP uptake in tumor wasobserved 24 hr after administration, the optimum tumor/normal brain ratios were observed at 48 hr. This indicates thatBOPP is not only selectively taken up by the tumor relativeto normal brain, but it is also preferentially retained in tumor.The CLSM micrographs in Fig. 4 provide unequivocal evi-dence that this sensitizer localizes in the nests of glioma cellsinvading into the edematous brain adjacent to tumor region,which are thought responsible for tumor recurrence (7). Thesubcellular localization in these tumor cells in vivo is similarto the in vitro situation. The exclusion of porphyrins from thenucleus has been reported previously (22, 23), and whilesome studies have suggested that porphyrins are taken up bylysosomes (22), the data presented here and by others (23, 24)show that hydrophobic porphyrins are specifically localizedin mitochondria.While PDT has been shown to be a promising adjuvant

therapy for cerebral glioma (5-9, 21), it is apparent that afactor limiting to its efficacy is the poor penetration of lightrequired to activate the sensitizer in these nests of tumorcells. The greater penetration of a neutron beam may enableactivation of the sensitizer, resulting in necrosis of these cellsand better local control of the tumor.While it is not clear whether BOPP has as much phototoxic

potential as HpD or whether it induces transient skin pho-tosensitization in the same manner as HpD, the degree andselectivity of uptake suggest that BOPP may be useful as adual-mode sensitizer for combination PDT and BNCT as

adjuvants to conventional surgery in the control ofintractabletumors such as glioblastoma multiforme.

We thank F. Steele for typing the manuscript and P. Smith, B.Kreunen, and M. Reddick for photography and preparation of allfigures. This work was supported by grants from the National Healthand Medical Research Council (Australia), Anti-Cancer Council ofVictoria, Stroke Research Foundation, Victor Hurley Foundation,George Hicks Foundation, Royal Australasian College of Surgeons,and the U.S. Department of Health and Human Services, NationalInstitutes of Health Grant CA 37961.

1. Walker, M. D., Green, S. B., Byar, D. P., Alexander, E.,Batzdorf, U., Brooks, W. H., Hunt, E. E., MacCarty, C. S.,Mahaley, M. S., Mealey, J., Owens, G., Ransohoff, J., Rob-ertson, J. T., Shapiro, W. R., Smith, K. R., Wilson, C. B. &Strike, T. A. (1980) N. Engl. J. Med. 303, 1323-1329.

2. Wilson, C. B. (1979) N. Engl. J. Med. 300, 1469-1471.3. Delaney, T. F. & Glatstein, E. (1988) Compr. Ther. 14, 43-55.4. Hill, J. S., Kaye, A. H., Sawyer, W. H., Morstyn, G., Megi-

son, P. D. & Stylli, S. S. (1990) Neurosurgery 26, 248-254.5. Wharen, R. E., Jr., Anderson, R. E. & Laws, E. R., Jr. (1983)

Neurosurgery 12, 446-450.6. Kaye, A. H. & Morstyn, G. (1987) Neurosurgery 20, 408-415.7. Kaye, A. H., Morstyn, G. & Apuzzo, M. L. J. (1988) J.

Neurosurg. 69, 1-14.8. Kaye, A. H. (1989) Photosensitizing Compounds: Their Chem-

istry, Biology, and Clinical Use, Ciba Foundation Symposium146 (Wiley, New York), pp. 209-224.

9. Muller, P. J. & Wilson, B. C. (1990) Lasers Med. Sci. 5,245-252.

10. Gabel, D., Foster, S. & Fairchild, R. G. (1987) Radiat. Res.111, 14-25.

11. Sweet, W. H., Messer, J. R. & Hatanaka, H. (1986) in BoronNeutron Capture Therapy for Tumors, ed. Hatanaka, H.(Nishimura Press, Niigata, Japan), pp. 59-76.

12. Coderre, J. A., Glass, J. D., Fairchild, R. G., Micca, P. L.,Fand, I. & Joel, D. D. (1990) Cancer Res. 50, 138-141.

13. Finkel, G. C., Poletti, C. E., Fairchild, R. G., Slatkin, D. N. &Sweet, W. H. (1989) Neurosurgery 24, 6-11.

14. Kahl, S. B. & Koo, M.-S. (1990) J. Chem. Soc. Chem. Coin-mun., 1769-1771.

15. Kahl, S. B., Joel, D. P., Nawrocky, M. M., Micca, P. L.,Tran, K. P., Finkel, G. C. & Slatkin, D. N. (1990) Proc. Natl.Acad. Sci. USA 87, 7265-7269.

16. Kaye, A. H., Morstyn, G., Gardner, I. & Pyke, K. (1986)Cancer Res. 46, 1367-1373.

17. Fairchild, R. G., Kahl, S. B., Laster, B. H., Kalef-Ezra, J. &Popenoe, E. A. (1990) Cancer Res. 40, 4860-4865.

18. Rendon, A. & Masmoudi, A. (1985) J. Neurosci. Methods 14,41-51.

19. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.(1951) J. Biol. Chem. 193, 265-275.

20. Kaye, A. H., Morstyn, G. & Ashcroft, R. G. (1985) Neurosur-gery 17, 883-890.

21. Kaye, A. H. (1990) in Malignant Cerebral Glioma, ed. Apuzzo,M. L. J. (Am. Assoc. Neurosurgeons, Park Ridge, IL), pp.189-202.

22. Moan, J., Berg, K., Kvam, E., Western, A., Malik, Z., Ruck,A. & Schneckenburger, H. (1989) Photosensitizing Com-pounds: Their Chemistry, Biology, and Clinical Use, CibaFoundation Symposium 146 (Wiley, New York), pp. 95-111.

23. Woodburn, K. W., Vardaxis, N. J., Hill, J. S., Kaye, A. H. &Phillips, D. R. (1991) Photochem. Photobiol. 54, 725-732.

24. Hanzlik, C. A., Knox, R. S., Gibson, S. L. & Hilf, R. (1989)Photochem. Photobiol. 50, 45-53.

Medical Sciences: Hill et al.

Dow

nloa

ded

by g

uest

on

Apr

il 19

, 202

1