In vivo activities of acidic fibroblast growth factor-Pseudomonas

Jun 18, 1993 - Susan L. Gawlak, Dana F. Chace, June R. Merwin,+ and Ira Pastan*. Bristol-Myers Squibb, Pharmaceutical Research Institute, Molecular ...
0 downloads 0 Views 1MB Size
Bioconjugate Chem. 1994, 5, 77-83

77

In Vivo Activities of Acidic Fibroblast Growth Factor-Pseudomonas Exotoxin Fusion Proteins Clay B. Siegall,' Susan L. Gawlak, Dana F. Chace, June R. Merwin,+ and Ira Pastant Bristol-Myers Squibb, Pharmaceutical Research Institute, Molecular Immunology Department, 3005 First Avenue, Seattle, Washington 98121, and Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892. Received June 18, 1993"

Fibroblast growth factor receptors are highly expressed in a variety of cancer cells and activated vasculature. Using chimeric toxins targeted to cell-surface aFGF receptors, we have demonstrated specific cytotoxic activity to these cell types. These molecules, aFGF-PE40 and aFGF-PE4E KDEL, are fusion proteins containing acidic FGF and either a 40- or a 66-kDa binding defective form of Pseudomonas exotoxin, respectively. Both aFGF-toxin fusion proteins were able to inhibit protein synthesis in vitro in a variety of carcinoma cell lines. The half-life of aFGF-PE40 in serum was found to be 41 min when coadministered with heparin. Administration of aFGF-PE40 or aFGF-PE4E KDEL with heparin inhibits the growth of established KB and preestablished A431 epidermoid carcinoma xenografts in athymic mice. The antitumor activities of the two aFGF-toxin fusion proteins were equivalent against the KB tumor xenografts. While we were able to slow the growth of the KB tumor xenografts, we were unable to cause tumor regressions. Histochemical analysis of treated versus untreated tumor tissue revealed a difference in tumor size but not of vascularity. We conclude that aFGF-PE40 and aFGF-PE4E KDEL have in vivo antitumor activity that targets the tumor cell mass rather than vascular structures in mice xenografted with human epidermoid carcinoma.

INTRODUCTION

Pseudomonas exotoxin (PE) is a single polypeptide made up of three structurally and functionally distinct domains ( 1 , 2 ) . Domain I contains the cell-binding activity, domain I1 the translocation activity, and domain I11 the ADP-ribosylation activity. P E kills cells by ADPribosylating elongation factor 2, resulting in inhibition of cellular protein synthesis. Alteration in the binding domain of P E results in a loss of cytotoxic activity against cells displaying P E receptors (3-5). Replacing the lost binding activity with a growth factor through fusion with the mutant P E molecule restores its cytotoxic activity but redirects the toxin to a specific growth factor receptor (6-8). We have recently reported on the cytotoxic activity of fusion proteins composed of acidic fibroblast growth factor (aFGF) and two mutant forms of Pseudomonas exotoxin A (PE) against a variety of cancer cell lines (9). FGFtoxin fusion proteins were produced to target FGF receptors that have been found in high numbers on cancer cells (10, 11). The specific aFGF-PE chimeric molecules were termed aFGF-PE40 and aFGF-PE4E KDEL. PE40 is a truncated form of P E that is missing the binding domain (domain I) but retains domains I1 and I11 (3,121. PE4E is a mutated form of P E in which four glutamates in domain I have been substituted in place of four positively charged residues. This form of PE is the functional equivalent of PE40 (13). The KDEL sequence, which encodes the endoplasmic reticulum retention sequence, is substituted a t the carboxyl end of aFGF-PE4E KDEL in place of the native REDLK residues to increase the cytotoxicity of the molecule (14,15). Chimeric toxins of

* To whom correspondence should be addressed.

+ Present address: TargeTech Inc., 290 Pratt St., Meriden,

CT 06450. * National Cancer Institute. Abstract published in Aduance ACS Abstracts, December 1, 1993. @

this type include TGFa-PE40, IGF TYPE I-PE40, and aFGF-PE40 (9,16-181, all employing Pseudomonas toxin, and EGF-DAB, utilizing a mutated form of Diphtheria toxin (19). Growth factor-toxins have been shown to be effective antitumor reagents against human tumor xenografts implanted in mice (20-22). In this report, we have explored the effects of the two aFGF-PE forms in the treatment of established and preestablished tumor xenografts in athymic mice. Our findings indicate that it is possible to target tumor cells for elimination using toxins directed at the aFGF receptor. EXPERIMENTAL PROCEDURES Animals and Cell Lines. For all experiments performed in animals, 6-week-old athymic mice weighing 1620 g were used (strain BALB/c, Harlan Sprague-Dawley, Indianapolis, IN). MCF-7 (breast carcinoma cells) and A431 and KB (epidermoid carcinoma cells) were purchased from American Type Culture Collection (Rockville, MD). L2987 (lung carcinoma cells), A2780 (ovarian carcinoma cells), and RCA (colon carcinoma cells) were from I. Hellstrom (Bristol-Myers Squibb, Pharmaceutical Research Institute, Seattle, WA). Plasmids and Constructions. The plasmids pCS F40 F(+)T and pCS F4K F(+)T were previously described (9). The plasmid pCS F4KD553encodes a mutant form of aFGF-PE4E KDEL that has no ADP-ribosylation activity due to its single amino acid deletion a t residue 553 (glutamate) in the P E sequence (23, 24). Plasmid pCS F4KD553was prepared by digesting pCSF4K a t its BamHI and EcoRI sites, removal of the resulting 0.45-kb fragment, and replacement with a similar 0.45-kb fragment from P C S F ~ O D(9). ~~~ Expression and Purification of aFGF-Toxins. Plasmids were transformed with bacterial strain BL21 (XDE3) and cultured as previously described (9). The fusion toxins were induced a t OD650 = 2.0 with 1 mM isopropyl 1-thio-P-D-galactopyranoside (IPTG) and har-

1043-1802/94/2905-0077$04.5~l0 0 1994 American Chemical Society

78

Bloconjugate Chem., Vol. 5, No. 1, 1994

vested 90 min later. The protein was purified from inclusion bodies by denaturation in guanidine-HC1 and renaturation by rapid dissolution in phosphate-buffered saline supplemented with 0.4 M L-arginine. The renatured proteins were purified by a two-step chromatography procedure involving anion-exchange followed by heparinaffinity, as previously described (9). Protein Synthesis Inhibition Assay. The tumor cells were cultured in RPMI 1640 supplemented with 105% fetal bovine serum, 2 mM L-glutamate, and 50 units/mL of penicillin/streptomycin. The cell monolayers were plated into 96-welltissue culture plates at 1X 104/welland allowed to adhere and grow for 16 h a t 37 "C. Chimeric aFGFtoxins were diluted in growth media and 0.1 mL added to each well for 20 h a t 37 "C. The cells (assay done in triplicate or quadruplicate) were pulsed with PHIleucine (1&i/well) for an additional 4 h a t 37 "C. The cells were harvested using a 96-well automated system (Tomtec, Orange, CT) after freeze-thawing of the plate. Protein synthesis was determined by counting the incorporation of [3H]leucine into cellular protein using an LKB-BetaPlate liquid scintillation counter. Serum Level of aFGF-PE40 (+ Heparin). Five pairs of athymic mice were injected intravenously with a single dose of aFGF-PE40 (0.5mg/kg, 100 units/mL of heparin). At indicated time points, each pair of injected animals was bled (200 pL each bleeding). The blood was kept on ice for 15 min and centrifuged at 2000 rpm in a 4 "C microfuge. The serum was retained and tested in an inhibition of protein synthesis assay with MCF-7 breast carcinoma cells to determine the amount of aFGF-PE40 protein present in each sample. The amount of protein synthesis inhibition found in the serum samples a t various time points was compared to a standard curve of aFGFPE40 (+ heparin) against the same MCF-7 cells. The protein synthesis assay was performed essentially the same as above. T 1 p ~was calculated using the MKModel Version 4 software package (Biosoft, Milltown, N. J.). Lethality of aFGF-Toxins i n Athymic Mice. Groups of five mice were injected intravenously with either aFGFPE40 or aFGF-PE4E KDEL. The aFGF-toxin protein was injected in PBS supplemented with 100 units/mL of heparin. The animals were observed for 21 days. Necropsy Analysis. Tissues from major organs, including lungs, kidney, liver, stomach, duodenum, ileum, cecum, and spleen were removed from sacrificed animals 48 h after injection of 1.5 mg/kg of aFGF-PE40 in PBS supplemented with 100 units/mL of heparin. The tissues were fixed in 10% formalin, blocked in paraffin, sectioned at 5 pm, and stained with hematoxylin and eosin. Antitumor Activity of aFGF-PE40 a n d aFGF-PE4E KDEL. KB and L2987 tumor fragments were implanted into female athymic mice (nu/nu) a t 4-6 weeks of age with tumor fragments maintained from serial in vivo passage of established xenografts. Approximately 2 weeks following subcutaneous implantation of the tumor fragments onto the back of the mice, the mice were randomized for those with tumor sizes ranging from 75 to 100 mm3. The tumor-bearing animals were intravenously injected with the chimeric aFGF-toxins (with 100 units/mL of heparin) as per specific administration schedule via the tail vein. There were five animals in each treatment group, and tumors were measured on a 3-4-day regular schedule with calipers. A431 cells (1X 109 weresubcutaneously injected into groups of athymic mice. For treatment a t day 1post implant, animals could not be randomized due to tumor size. A t day 5 post implant, animals were randomized

Slegall et al.

Table 1. Cytotoxic Activities of aFGF-PE40 and aFGF-PE4E KDEL Chimeric Proteins on Various Cell Lines

ECW.' ndmL aFGFcell line (type) PE4EKb PE4EKbDm PE A2780 (ovarian Ca) 60 40 >500 90 200 150 L2987 (lungCa) >500 20 60 50 >500 35 KB (epidermoid Ca) A431 (epidermoid Ca) 8 9 6 >500 15 RCA (colon Ca) 10 >500 32 MCF-7 (breast Ca) 40 10 >500 25 ECm is the amount of aFGF-toxin needed to inhibit 50% of protein synthesis. K = KDEL. aFGFPE40

aFGF-

into groups of five based on the appearance of small A431 tumor nodules. Histologic Examination of Treated a n d Untreated Tumor Tissue. Athymic mice were implanted with KB tumor sections as described above. Mice (A treatment with aFGF-PE40 (+heparin), 0.3785 mg/kg, 1d X 3) were sacrificed 24 days post-tumor implant, and the tumors were removed. The tumor tissue was rinsed with PBS and cut into three longitudinal sections which were then fixed in 10% formalin. The samples were blocked in paraffin, sectioned at 8 pm, and stained with hemotoxylin and eosin prior to being mounted onto slides and inspected with light microscopy a t (13X and 1OOX). RESULTS In VitroProtein Synthesis Inhibition Activity. The sensitivity of a variety of cancer cells to aFGF-PE40 and aFGF-PE4E KDEL was assessed. The two forms of aFGF-PE were active in inhibiting protein synthesis in A431, L2987, RCA, A2780, MCF-7, and KB carcinoma cell lines (Table 1). A431 cells were the most sensitive of the cell lines to both aFGF-toxin forms. A mutant form of aFGF-PE4E KDEL, termed aFGF-PE4E KDEL D663, in which there is no ADP-ribosylation activity, did not inhibit protein synthesis (EC50 > 500 ng/mL) in any of the cell lines tested (Table 1). Serum Level Analysis of aFGF-PE40. Since heparin has been shown to potentiate the biological effects of aFGF (25-28) and potentiate the cytotoxic activity of aFGFtoxins (29), heparin was coinjected with aFGF-PE40 in a serum level assay. Acidic FGF-PE40, with 100 units/mL of heparin, was injected via the tail vein into athymic mice. Blood was obtained at various times following injection, and the serum was isolated as described in the Experimental Procedures. The serum was diluted in growth media and compared to known amounts of aFGF-PE40 (and heparin) in a cytotoxicity assay against MCF-7 breast carcinoma cells. By comparing the protein synthesis inhibition activity of the serum samples with known amounts of aFGF-PE40, the amount of functional aFGFPE40 present in the serum a t different time points was determined. The data show that aFGF-PE40 can be recovered from blood when coadministered with heparin (Figure 1). The T1pa of aFGF-PE40, with heparin coadministration, is 41 min. At 6 h post administration, there was no aFGF-PE40 detectable in the serum. Lethality i n Athymic Mice. Before assessing the ability of aFGF-PE to inhibit tumor growth in uiuo, the maximum tolerated dose of the chimeric toxin was estimated by injecting the two different molecules intravenously in groups of five athymic mice. Different administration schedules and dose regimens were tested, and the results are listed in Table 2. Acidic FGF-PE4E KDEL was approximately 2-3-fold more lethal to mice

Bloconjugate Chem., Vol. 5, No. 1, 1994 79

aFGF-PE Fusion Proteins

5

15

30

60

120

240

360

TIME, min

Figure 1. Serum levels and T 1 p of aFGF-PE40inmice. Acidic

FGF-PE40 levels were determined by measuring inhibition of protein synthesis with serum samplestaken at various time points following injection of 0.5 mg/kg of aFGF-PE40 against MCF-7 breast carcinoma cells. These data were compared to an inhibition of protein synthesis assay using known amounts of aFGF-PE40 against the same cell line. Heparin was added to the injected aFGF-PE40 fusion protein at 100 units/mL. The Tlpa for aFGF-PE40 was 41 min. Table 2. Maximum Tolerated Doses of aFGF-Toxin Fusion Proteins (Intravenous Injections) in Athymic Mice total dose lethality (%) dose administrn (mg/kg) (dead/total) (mg/kg) schedule aFGF-PE4E KDEL single dose 0.25 0 0.25 0.375 single dose 0.375 0 0.5 single dose 0.5 40 single dose 0.625 100 0.625 qld x 3 0.75 0 0.25 0.375 qld X 3 1.125 40 q4d X 3 1.125 0 0.375 q4d X 3 1.5 80 0.5 0.5 1.0 1.5 2.0 0.5 1.0

0.5

1.0

aFGF-PE40 0.5

single dose single dose single dose single dose qld X 3 qld X 3 q4d X 3 q4d X 3

1.0 1.5

2.0 1.5 3.0

1.5

3.0

0 0 20 80 20 80 0 60

than was aFGF-PE40. No lethality was observed a t singledose amounts of 0.375 mg/kg for aFGF-PE4E KDEL and 1.0 mg/kg for aFGF-PE40. Administration of larger total doses without lethality was possible using multiple injections of the aFGF-toxin fusion proteins (Table 2). Higher total doses of the fusion proteins could be given if an administration schedule of q4d X 3 was utilized rather than a schedule of q l d X 3. Toxicity of aFGF-PE40 toward Mouse Tissue. The effects of aFGF-PE40 on normal cells was determined in mice since FGF receptors are widespread (30-32). Mice were injected with a single dose of a sublethal amount of aFGF-PE40 (1.5mg/kg) and the mice sacrificed 48 h later so that necropsies could be performed. Histological analyses indicated significant organ damage in both the liver and spleen (data not shown). The liver cells showed marked polyploidy characterized by a wide variation in nuclear size and chromatin content. There were numerous mitotic figures, and many hepatocytes were binucleate. Additionally, there was a substantial number of nuclear and cytoplasmic remnants in a generalized distribution

within the liver sections analyzed. The spleen was enlarged, and there was excessive hematopoiesis. The red pulp was mildly congested, and the lymphoid follicleswere enlarged as compared to nontreated mice. No toxicity was found in any of the other tissues analyzed. Antitumor Activity of aFGF-PE40. To assess the in vivo antitumor activity of aFGF-toxins, KB cells were xenografted into athymic mice. While KB cells were not the most sensitive cell line tested in this study, aFGFtoxin fusion proteins were still quite cytotoxic to the cells, with EC50 values of 50-60 ng/mL (Table 1). Additionally, when passaged into animals using trocar-implanted tumor tissue, KB tumor xenografts grew into highly vascularized tumors in a very tight size range making for extremely reproducible experiments. KB tumor cells were implanted into nude mice and allowed to grow until the tumor size was between 200 and 400 mm3. The tumor xenografts were serially sectioned and passaged into additional mice using a trocar. When the tumor xenografts were established and vascularized (approximately 75-100 mm3, 15-17 days following implantation), the mice were randomized and set into groups of five per cage. Mice were treated with intravenous injections of chimeric toxin supplemented with 100 units/ mL of heparin, using a variety of administration schedules. Acidic FGF-PE4E KDEL or aFGF-PE40 was administered using a treatment schedule of q l d X 3 (Figure 2A and B), respectively. Higher doses of aFGF-PE40 were used, as it was less toxic to the mice (Table 2). In Figure 2A, the experimental control was treatment with aFGFPE4E KDEL D553, a noncytotoxic mutant form of aFGFPE4E KDEL. In Figure 2B, untreated animals were the control. Tumor xenografts from these two control groups grew at essentially the same rate, indicating that the noncytotoxic aFGF-toxin form did not inhibit tumor growth. Administration of aFGF-PE4E KDEL a t 0.25 mg/kg was able to inhibit tumor growth while a dose of 0.125 mg/kg had very little antitumor effect (Figure 2A). Surprisingly, a t the lower dose of aFGF-PE4E KDEL, there was a slight inhibition of tumor growth, matching that of the higher dose from days 33-40. Administration of aFGF-PE40 was able to generate a stronger antitumor response as compared to aFGF-PE4E KDEL (Figure 2B). Mice with the KB xenografts were injected with both 0.375 and 0.25 mg/kg of aFGF-PE40, using the same q l d X 3 schedule. The antitumor effect was approximately the same for both doses. Higher doses of chimeric toxin could be administered to mice without toxicity by prolonging the time between doses. Figure 3 shows the effects of the fusion proteins using a q4d X 3 schedule. The antitumor activity of both aFGF-PE4E KDEL and aFGF-PE40 against the KB tumor xenografts was similar. With aFGF-PE4E KDEL a t doses of 0.375 and 0.25 mg/kg, the antitumor activity was nearly identical (Figure 3A). The same was found for aFGF-PE40 (0.5 and 0.375 mg/kg) (Figure 3B). As shown in both Figures 2 and 3, the KB tumor xenograft growth was significantly retarded by the administration of aFGF-PE40 or aFGF-PE4E KDEL. The effect on the tumor was most profound during the course of administration and for a short period following the treatment. This usually was found between days 15 and 30 post tumor implant. Following this period, the tumors grew back and had the appearance of a normal growing tumor. The most significant antitumor response came from the q4d x 3 administration of both aFGF-toxin forms,

00

Siegall et al.

Bioconjugate Chem., Vol. 5, No. 1, 1994

1000 7

A.

8.

h

E E

750-

3

V

Y

ia f+

500-

250-

.

U '

.

.

.

,

,

,

.

,

I

I

10 15 20 25 30 35 40 45 50

1250 -

E E a

I

-

I

.

I

.

I

.

I

.

I

.

I

.

i

50

-

1500 h

.

10 15 20 25 30 35 40 45

1000-

V

v

W

750-

E

500-

; 0 5c

250: 0

,

I

,

I

,

,

1

,

1

,

1

,

1

,

,

1

.

1

'

I

.

1

'

l

.

l

'

l

.

i

aFGF-PE Fusion Proteins

Bloconjugate Chem., Vol. 5, No. 1, 1994 81

A.

B.

'"1

'"3

A431

f

lo00

i

l5OI 500

250

-

00

10

20

30

40

0

10

20

30

40

DAYS P O S T TUMOR I M P L A N T

Figure 4. Antitumor activity of aFGF-toxin fusion proteins in mice xenografted with (A) A431 epidermoid carcinoma cells and (B) L2987 lung carcinoma cells. In both models, aFGF-toxin fusion protein was administered intravenously with 100 units/mL of heparin: (A) aFGF-PE40 treatment (0.5 mg/kg, q2d X 3) was initiated day 5 post implant (A),day 1 post implant (A),nontreated controls ( 0 ) ;(B) aFGF-PE4E KDEL treatment was initiated day 14 post tumor implant using a schedule of q4d X 3,0.375 mg/kg ( O ) , 0.25 mg/kg (w), nontreated controls ( 0 ) .

underneath the highly vascularized edge. The central sections of the treated and untreated tumor mass had areas of necrosis (N), fibrosis (F), and a large infiltrate of polymorphonuclear cells (Figure 5, panels C and E). There appeared to be less microvasculature in the treated tumor samples; however, this was not quantitated. The untreated tumor contained more hemorrhagic areas. DISCUSSION We have prepared many growth factor-toxins using a variety of forms of Pseudomonas exotoxin. In this report, we compare two forms, aFGF-PE40 and aFGF-PE4E KDEL, and demonstrate that they both have antitumor activity against established human KB tumor xenografts and preestablished A431 tumor cells in athymic mice. The antitumor activity resulted in an inhibition of continued KB tumor growth during and for a short time following administration of the chimeric toxin. The treatment did not result in regression of the tumor xenografts (Figures 2 and 3). While a more dramatic antitumor effect was found using A431 tumor cells in uiuo, these experiments used tumors that were not fully established (Figure 4A). Against established L2987 tumor xenografts that were weakly sensitive to the protein synthesis inhibitory effects of aFGF-toxin fusion protein in uitro there was no or minimal antitumor activity (Figure 4B). Heparin was coadministered (100 units/mL) with aFGFtoxins since it potentiates the effects of FGF on the growth of cells by prolonging its biological lifetime (25-28) as well as potentiating the cytotoxic effects of aFGF-toxin fusion protein (29). The serum half-life of aFGF-toxin fusion protein coinjected with heparin was 41 min (Figure 1). Acidic FGF-PE4E KDEL was devised to increase the cell-specific cytotoxic activity (14, 15), and as shown in Table 1, it is more cytotoxic on certain cell lines than aFGF-PE40 (Table 1). However, it also is more toxic to athymic mice (Table 2). In fact, it would appear that the therapeutic window with aFGF-PE40 is larger than the window with aFGF-PE4E KDEL against most carcinoma cell lines due to the increased toxicity of aFGF-PE4E KDEL. This type of difference was also found for another chimeric toxin which was composed of interleukin 6 and various forms of P E ( 3 3 ) . In the IL6 study, IL6-PE4E

KDEL was a t least 2-fold more toxic to animals than both IL6-PE40 and IL6-PE4E (C. B. Siegall, unpublished observation). Thus, it may not be advantageous to construct chimeric toxins in which KDEL is included a t the carboxyl end since although there is a slight increase in its cytotoxic effects against some cell lines the increase in the nonspecific i n vivo toxicity may be greater. Acidic FGF-toxin molecules are not only cytotoxic to tumor cells but they also are cytotoxic to rapidly proliferating vasculature-bearing aFGF receptors (9). We have also found that aFGF-PE4E KDEL is cytotoxic to endothelial cells using an in vitro model of angiogenesis ( 3 4 ) . In addition, aFGF-PE4E KDEL was cytotoxic to cardiac smooth muscle and endothelial cells isolated as primary culture from rats (35). In the current study, analysis of treated and nontreated KB tumor xenografts with histochemical techniques revealed that the tumor cells were affected by the aFGFtoxin treatment while the vascular tissue was not (Figure 5). The vascular cells that supply the xenograft are of mouse origin, while the tumor tissue is of human origin. Acidic FGF binds to both human and mouse aFGF receptor and therefore should be able to target the actively proliferating vasculature that is supplying the tumor with blood. In contrast to antibody-toxin chimeric molecules in which the antibody binds to the human tumor xenograft and not to mouse tissue (361, growth factor-toxins such as aFGF-toxin fusion proteins are valid for use in mouse models, since aFGF-PE40 binds to both mouse and human aFGF receptor. We do not know whether i n uiuo vasculature displays enough cell surface aFGF receptors to be sensitive to aFGF-toxin. Our previous results studying proliferating endothelial cells have shown that the endothelial cells are sensitive to the protein synthesis inhibition activity of aFGF-toxin chimeric molecules (35). However, the levels that are needed to reach 50% of protein synthesis is 100 ng/mL for endothelial cells in vitro. This level is slightly higher than the level of aFGF-toxin needed to affect KB tumor cells in culture (Table 1). It is possible that at the administered dose the endothelial cells of the vascular tissue supplying the blood to the tumor xenografts were not sensitive to the aFGF-toxin. Alternatively, it is

82

Bbcon$~@e chem., Vol. 5, No. 1, 1994

Sbgall et al.

Treatment of metastatic tumors is one of the most important areas in cancer chemotherapy. There is adirect correlation between metastasis and tumor angiogenesis (40). Thus, we attempted to target both the tumor tissue and the vasculature with these aFGF-toxin molecules. Although we could not definitively show that aFGF-toxins were able to target the tumor vasculature in this model system (Figure 5), this type of investigation may lead to the discovery of new and powerful proteins and/or drugs that will be potent inhibitors of human tumor growth. ACKNOWLEDGMENT We thank Drs. P. Senter, P. Friedman, and E. Wolff for critical reading of the manuscript, and Drs. P. Trail, P. Fell, and K. E. Hellstrijm for helpful suggestions. LITERATURE CITED (1) Gray, G. L., Smith, D. H., Baldridge, J. S., Harkins, R. N., Vasil, M. L., Chen, E. Y., and Heyneker, H. L. (1984)Cloning,

nucleotide sequence,and expressionin Escherichia coli of the exotoxinA structural gene of Pseudomonas aerugirwsa. Proc. Natl. Acad. Sei. U.S.A. 81, 2645-2649. (2) Allured, V. S., Collier, R. J., Carroll, S. F.,and McKay, D. B. (1986)Structure of exotoxin A of Pseudomonas aeruginosa at 3.0-A resolution. Proc. Natl. Acad. Sei. U.S.A. 83, 1320-

I

Figure 5. Histologic examination of aFGF-PE40-treated and -untreated KB tumor tissue. Mice carryingKBtumor xenografts were either treated with aFGF-PE40 (0.375 mg/kg, Id X 3) or untreated and sacrificed24 dayspost tumor implant. The tumors were removed, fixed in 10% formalin, blocked in paraffin, and sectioned at 8pm. The tissue was stained with hematoxylin and eosin, mounted onto slides,and inspectedunder light microscopy. Nontreated tumors (A-C) measured 1.7 X 1.1 cm while treated tumors (D-F) measured 5 X 7 mm. Whole mounts (A, F), edge of tumor mass (B, D), center of mass (C, E). N = necrosis; F = fibrosis. Magnification: A, F = 13X, B-E = l00X.

possible that reduced vascular development due to aFGFtoxin treatment could have resulted in reduced tumor growth. A similar type of growth factor-toxin molecule that is active against cells with FGF receptors is bFGF-SAP, a chemical conjugate composed of basic FGF and the plant toxin, saporin (37). This molecule is valuable in culturing pancreatic islet cells since bFGF-SAP can eliminate the fibroblast cells that frequently block initiation of islet cells in culture (38). Basic FGF-SAP is specifically cytotoxic to cells displaying bFGF-receptors (39) and can inhibit the growth of human tumor xenografts in nude mice (22).

1324. (3) Hwang, J., FitzGerald, D. J. P., Adhya, S., and Pastan, I. (1987)Functionaldomainsof Pseudomonas exotoxinidentified by deletion analysis of the gene expressed in E. coli. Cell 48, 129-136. (4) Jinno, Y., Chaudhary,V. K.,Kondo,T., Adhya,S.,FitzGerald, D. J., and Pastan, I. (1988) Mutational analysis of domain I of Pseudomonas exotoxin. J. Biol. Chem. 263,13202-13207. (5) Chaudry, G. J., Wilson, R. B., Draper, R. K., and Clowes, R. C. (1989)A dipeptide insertion in domain I of exotoxin A that impairs receptor binding. J. Biol. Chem. 264, 15151-15156. (6) Pastan, I., and FitzGerald, D. (1989) Pseudomonas exotoxin: chimeric toxins. J. Biol. Chem. 264, 15157-15160. (7) FitzGerald, D., and Pastan, I. (1989)Targeted toxin therapy for the treatment of cancer. J. Natl. Cancer Inst. 81,14551463. (8) FitzGerald, D., Chaudhary, V. K., Kreitman, R. J., Siegall, C. B., and Pastan, I. (1992) in Genetically Engineered Toxins (A. E. Frankel, Ed.) pp 447-462, New York, Marcel Dekker. (9) Siegall, C. B., Epstein, S. E., Speir, E., Hla, T., Forough, R., Maciag, T., FitzGerald, D., and Pastan, I. (1991) Cytotoxic

activity of chimeric proteins composed of acidic fibroblast growth factor and Pseudomonus exotoxin on a variety of cell types. FASEB J . 5,2843-2849. (10) Armstrong, E., Vainikka, S., Patanen, JU., Korhonen, J., and Alitalo, R. (1992). Expressionof fibroblast growth factor receptors in human leukemia cells. Cancer Res. 52, 20042007. (11) Luqmani, Y. A., Graham, M., and Coombes, R. C. (1992)

Expression of basic fibroblast growth factor, FGFRl and FGFR2 in normal and malignant human breast, and comparison with other normal tissues. Br. J. Cancer 66,273-280. (12) Kondo, T., FitzGerald, D., Chaudhary, V. K., Adhya, S., and Pastan, I. (1988) Activity of immunotoxins constructed with modified Pseudomonas exotoxin a lacking the cell recognition domain. J. Biol. Chem. 263,9470-9475. (13) Chaudhary, V. K., Jinno, Y., Gallo, M. G., FitzGerald, D., and Pastan, I. (1990) Mutagenesis of Pseudomonas exotoxin in identification of sequences responsible for the animal toxicity. J. Biol. Chem. 265, 16306-16310. (14) Chaudhary, V. K., Jinno, Y., FitzGerald, D., and Pastan, I. (1990)Pseudomonas exotoxin contains a specific sequenceat the carboxyl terminus that is required for cytotoxicity. Proc. Natl. Acad. Sci. U.S.A. 87, 308-312. (15) Seetharam,S.,Chaudhary,V. K., FitzGerald,D.,and Pastan, I. (1991)Increasedcytotoxicactivity of Pseudomonas exotoxin

aFGF-PE Fusion Proteins and two chimeric toxins ending in KDEL. J.Biol. Chem. 266, 17376-17381. (16) Strom,T. B., Anderson,P. L.,Rubin-Kelley, V. E., Williams, D. P., Kiyokawa,T., and Murphy, J. R. (1991) Immunotoxins and cytokine toxin fusion proteins. Ann. N. Y.Acad. Sci. 636, 233-250. (17) Siegall, C. B., Xu, Y. H., Chaudhary, V. K., Adhya, S., FitzGerald, D., and Pastan, I. (1989) Cytotoxic activities of a

fusion protein comprised of TGFa andPseudomonasexotoxin. FASEB J. 3, 2647-2652. (18) Prior, T. I., Helman, L. J., FitzGerald, D. J., and Pastan, I. (1991) Cytotoxic activity of a recombinant fusion protein between insulin-like growth factor I and Pseudomonas exotoxin. Cancer Res. 51, 174-180. (19) Shaw, J. P., Akiyoshi, D. E., Arrigo, D. A., Rhoad, A. E., Sullivan, B., Thomas, J., Genbauffe, F. S., Bacha, P., and Nichols, J. C. (1991) Cytotoxic properties of DAB486EGFand DAB389EGF, epidermal growth factor (EGF) receptor-targeted fusion toxins. J. Biol. Chem. 266, 21118-21124. (20) Heimbrook, D. C., Stirdivant, S. M., Ahern, J. D., Balishin, N. L., Patrick, D. R., Edwards, G. M., Defeo, J. D., FitzGerald, D. J., Pastan, I., and Oliff, A. (1990) Transforming growth factor alpha-Pseudomonas exotoxin fusion protein prolongs survival of nude mice bearing tumor xenografts. Proc. Natl. Acad. Sci. U.S.A. 87, 4697-4701. (21) Pai, L. H., Gallo, M. G., FitzGerald, D. J., and Pastan, I. (1991) Anti-tumor activity of a transforming growth factor alpha-Pseudomonas exotoxin fusion protein (TGFa-PE4O). Cancer Res. 51, 2808-2812. (22) Beitz, J. G., Davol, P., Clark, J. W., Kato, J., Medina, M., Frackelton, A. R. Jr., Lappi, D. A., Baird, A., and Calabresi, P. (1992) Antitumor activity of basic fibroblast growth factorsaporin mitotoxin in vitro and in vivo. Cancer Res. 52, 227230. (23) Douglas, C. M., and Collier, R. J. (1987) Exotoxin A of Pseudomonas aeruginosa: substitution of glutamic acid 553

with aspartic acid drastically reduces toxicity and enzymatic activity. J. Bacteriol. 169, 4967-4971. (24) Carroll, S. F., and Collier, R. J. (1987) Active site of Pseudomonas aeruginosa exotoxin A glutamic acid 553 is photolabeled by NAD and shows functional homology with glutamic acid 148 of Diphtheria toxin. J. Biol. Chem. 262, 8707-8711. (25) Schreiber, A. B., Kenney, J., Kowalski, W. J., Friesel, R., Mehlman, T., and Maciag,T. (1985) Interaction of endothelial

cell growth factor with heparin: Characterization by receptor and antibody recognition. Proc. Natl. Acad. Sci. U.S.A. 82, 6138-6143. (26) Gospodarowicz,D., and Cheng, J. (1986) Heparin protects basic and acidic FGF from inactivation. J.Cell. Physiol. 128, 475-484.

Bioconjugate Chem., Voi. 5, No. 1, 1994

83

(27) Rosengart, T. K., Kupferschmid, J. P., Ferrans, V. J., Casscells, W., Maciag, T., and Clark, R. E. (1988) Heparin-

binding growth factor I (endothelial cell growth factor) binds to endothelium in vivo. J. Vasc. Surg. 7, 311-317. (28) Damon, D. H., Lobb, R. R., DAmore, P. A., Wagner, J. A. (1989) Heparin potentiates the action of acidic fibroblast growth factor by prolonging its biological half-life. J. Cell. Physiol. 138, 221-226. (29) Gawlak, S. L., Pastan, I., and Siegall, C. B. (1993) Basic FGF-Pseudomonas exotoxin chimeric proteins; Comparison with acidic FGF-Pseudomonas exotoxin. Bioconjugate Chem. (in press). (30) Burgess, W. H., and Maciag,T. (1989) The heparin-binding (fibroblast) growth factor family of proteins. Annu. Rev. Biochem. 58, 575-606. (31) Basilico, C., and Moscatelli, D. (1992) The FGF family of growth factors and oncogenes. Adv. Cancer Res. 59,115-165. (32) Gospodarowicz,D. (1989) Fibroblast growth factor. Crit. Rev. Ongogen. 1, 1-26. (33) Siegall,C. B., Kreitman, R. J.,FitzGerald, D. J., and Pastan, I. (1991) Antitumor effects of interleukin 6-Pseudomonas exotoxinchimeric moleculesagainst the human hepatocellular carcinoma, PLC/PRF/5 in mice. Cancer Res. 51,2831-2836. (34) Merwin, J. R., Lynch, M. J., Madri, J. A., Pastan, I., and Siegall, C. B. (1992) Acidic fibroblast growth factor-Pseudomonus exotoxin chimeric protein elicits antiangiogenic effects on endothelial cells. Cancer Res. 52,4995-5001. (35) Biro, S., Siegall, C. B., Fu, Y-M., Speir, E., Pastan, I., and Epstein, S. E. (1992) Zn vitro effects of a recombinant toxin targeted to the fibroblast growth factor receptor on rat vascular smooth muscleand endothelial cells. Circulation Res. 71,640645. (36) Friedman, P. N., Chace, D. F., Trail, P. A., and Siegall, C. B. (1993) Anti-tumor activity of the single-chain immunotoxin

BR96 sFv-PE40 against established breast and lung tumor xenografts. J. Zmmunol. 150, 3054-3061. (37) Lappi, D. A., Martineau, D., and Baird, A. (1989) Biological and chemicalcharacterization of basic FGF-saporin mitotoxin. Biochem. Biophys. Res. Commun. 160,917-923. (38) Beattie, G. M., Lappi, D. A., Baird, A., and Hayek, A. (1990) Selective elimination of fibroblasts from pancreatic islet monolayers by basic fibroblast growth factor-saporinmitotoxin. Diabetes. 39, 1002-1005. (39) Lappi, D. A., Maher, P. A., Martineau, D., and Baird, A. (1991) The basic fibroblast growth factor-saporin mitotoxin acts through the basic fibroblast growth factor receptor. J. Cell. Physiol. 147, 17-26. (40) Weidner, N., Semple, J. P., Welch, W. R., and Folkman, J. (1991) Tumor angiogenesis and metastasis: correlation in invasive breast carcinoma. N. Engl. J. Med. 324, 1-8.