TGFα-PE40 inhibits non-small cell lung cancer growth

Life Sciences, Vol. 54, No. 7, pp. 445~53, 1994 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0024-3205/94 $6.00 + .00

Pergamon

TGFa-PE40 INHIBITS NON-SMALL CELL LUNG CANCER GROWTH Muriel Draoui 1, Clay B. Siegall 2'3, David FitzGerald 2, Ira Pastan 2 and Terry W. Moody 1'4. tDepartment of Biochemistry and Molecular Biology, The George Washington University School of Medicine and Health Sciences, Washington, D.C. 20037 and 2Laboratory of Molecular Biology, National Cancer Institute, NIH, Bethesda, MD 20892 (Received in final form November 24, 1993) Summary The ability of a chimeric toxin containing transforming growth factor c¢ (TGFa) and truncated Pseudomonas exotoxin A to inhibit NSCLC growth was investigated. TGFa-PE40 inhibited binding of I~I-EGF to NSCLC cell lines with an ICso value of 0.5-3 lag/ml. Similarly, other forms of the fusion protein, TGFa-PE38 and TGFa-PE40 ~p553, which have active T G F a binding domains, inhibited specific 125I-EGF binding to NSCLC cells with ICso values of 0.1-2 and 0.05-0.5 ~g/ml respectively. TGFa-PE40 inhibited 35S-methionine uptake by NSCLC cells with an EDs0 value of 1-30 ng/ml. TGFa-PE38, which has one of the two disulfide pairs of PE40, inhibited amino acid uptake with EDs0 values of 3-50 ng/ml whereas TGFaPE40 A~p55a,which lacks ADP ribosylation activity, had an EDso > 100 ng/ml. TGFaPE40 inhibited colony formation of NSCLC cells with an LDso value of 0.008-0.1 ng/ml. Similarly, TGFc~-PE38 inhibited NSCLC colony formation with LDs0 values of 0.002-0.1 ng/ml whereas TGFa-PE40 ~p~3 had an LD~0 > 10 ng/ml. Also, TGFa-PE40 and TGFa-PE38 inhibited NSCLC xenograft formation in nude mice whereas TGFaPE40 ~ p ~ was inactive. These data suggest that TGFa-PE40 and TGFa-PE38 may be useful agents to inactivate NSCLC cells. Transforming growth factor ¢z (TGFc0, a 50 amino acid protein that is derived from a 170 amino acid precursor protein, is a non-small cell lung cancer (NSCLC) growth factor (1,2). TGFa, which has 3 essential disulfide bonds, bears structural homology to epidermal growth factor (EGF), which has 53 amino acids. The growth of NSCLC cells in vitro is slowed by a neutralizing antibody against T G F a (3). T G F a binds with high affinity to the EGF receptor, which contains 1186 amino acids and has structural similarities to the v-erbB oncogene (4,5). High affinity EGF binding sites (100,000/cell) are present on almost all NSCLC cell lines examined (6). When EGF or T G F a binds to the EGF receptor tyrosine kinase activity is stimulated resulting in autophosphorylation of the 170 kDalton EGF receptor. Also, which contains a 90 and 110 kDalton subunit, serves as a tyrosine kinase substrate. EGF causes internalization of the EGF receptor and stimulates the clonal growth of numerous NSCLC cell lines. Recently monoclonal antibody (mAb) 108 was identified which binds with high affinity to the EGF receptor but does not intemalize. The clonal growth of NSCLC cells and xenograft formation in nude mice was inhibited by mAb 108 (7).While mAb 108 slowed NSCLC growth, it unfortunately did not completely regress NSCLC cells. ~Present address: Bristol-Myers Squibb, Mol. Immunol. Dept., Seattle, WA 98121 4Present address:BPRB, DCPC, NCI, Rockville, MD 20850

446

TGFct-PEO40 Inhibits NSCLC Growth

Vol. 54, No. 7, 1994

Pseudomonas exotoxin (PE) is composed of three structural domains encoding binding, translocation and ADP-ribosylation activities (8). The removal of the binding domain results in a molecule PE40 that does not bind to the PE receptor but retains its other activities. Addition of a cell binding moiety to PE40 results in a cell specific cytotexic reagent (9). One such molecule TGFa-PE40 is cytotoxic towards A431 and KB cells which have 3 x 106 and 2 x 105 EGF receptors/cell respectively (10). TGFa-PE40 prolongs the survival of nude mice with colon cancer xenografts (11) and delays the growth of epidermoid and prostate cancer (12,13). In this report the effects of TGFa-PE40 and TGFa-PE38, a variant with one of the two disulfide pairs deleted from PE40 (14) were investigated on NSCLC cells. Methods TGFa-PE40 and TGFa-PE38 were induced in E. coli BL21 (kDE3 cells) transformed with pVC387 or pCS6) using 1 mM isopropylthiogalactoside. Cultures were harvested after 90 min and centrifuged at 3000 rpm for 15 min. The cell pellet was fractionated and inclusion bodies isolated from the spheroplast fraction as described previously (15). The fusion proteins were denatured with guanidine-HC1 and refolded in PBS. Refolded TGFatoxin was purified first by anion exchange column, the partially purified material pooled and purified to homogeneity using a TSK-250 gel filtration column. Fractions containing ADPribosylation and cell killing activity were used for biological experiments. H u m a n lung cancer cell lines were cultured in serum supplemented medium (RPMI1640 (Fisher Scientific) containing 10% heat inactivated fetal bovine serum (Gibco)) at 37°C (16). The NSCLC cell lines were adherent and were split weekly 1:20 using trypsin/EDTA (Gibco). Routinely, the cells had greater than 90% viability and were mycoplasma free. The cells were used when they were in exponential growth phase. The ability of TGFa-PE40 to inhibit EGF receptor binding was investigated. The NSCLC cells were harvested aider treatment with trypsin]EDTA and 5 x 104 cells were added to 24 well plates that were pretreated with 10 ~g of human fibronectin (Sigma Chemical Co.). After a monolayer of cells had formed (3 days) receptor binding studies were conducted. The cells were washed 4x with SIT media (RMPI-1640 containing 3 x 10~ M Na2SeOa, 5 pg/ml insulin, 10 pg/ml transferrin; Sigma Chemical Co.) and incubated with 12~IEGF (105 cpm) in 200 pl of receptor binding buffer (SIT containing 0.1% bovine serum albumin; Calbiochem). EGF was iodinated using the chloramine T procedure and I~I-EGF (20 laCi/pg) purified using gel filtration techniques. After 30 rain at 37°C, free radiolabeled protein was removed and the cells that contained bound growth factor dissolved in 0.2 N NaOH and counted in a g a m m a counter. EGF and T G F a were purchased from Collaborative Research Inc. The ability of TGFa-PE40 to inhibit protein synthesis in NSCLC cells was investigated. Cells were incubated with TGFcc-PE40 for 16 hrs in methionine free RPMI1640 containing 10% FBS. Then a~S-methionlne was added for 1 hr. The plates were then washed 3x with cold PBS. The cells were then treated with 5% cold trichloroacetic acid followed by ethanol/ether (2/1). The cells were then solubilized in 0.2 N NaOH followed by 0.2 N HC1. The cells were then solubilized with 0.2 N NaOH for I hr at room temperature and counted in a B-counter. NSCLC cell lines were harvested and tested in the agarose cloning system described previously (17). The base layer consisted of 3 ml of 0.5% agarose in SIT medium containing 5% fetal bovine serum in 6 well plates (Falcon, Oxnard, CA). The top layer consisted of 3 ml of SIT medium in 0.3% agarose, TGFa-PE40 and 1 x 104 single viable cells. For each cell

Vol. 54, No. 7, 1994

TGFa-PEO40 Inhibits NSCLC Growth

447

line and growth factor concentration, triplicate wells were plated. After 2 weeks 1 ml of 0.1% p-iodonitrotetrazolium violet was added and after 16 hours at 37°C the plates were screened for colony formation; the number of colonies larger than 120 pin in diameter were counted. The ability of TGFa-PE40 to inhibit xenograft formation in nude mice was investigated. Female athymic Balb/c nude mice, 4-5 weeks old, were housed in a pathogenfree temperature controlled isolation room and the diet consisted of autoclaved rodent chow and autoclaved water given ad libit~m. NCI-H157 or H838 cells (1 x 107) were injected into the right flank of each mouse by subcutaneous injection. Palpable tumors were observed in approximately 90% of the mice after 2 weeks at which time PBS (100 pl), or TGFa-PE40 (7 pg/100 lxl) were placed in Alzet micropumps and the pumps implanted i.p. The tumor volume (height x width x depth) was determined weekly by calipers and recorded. When the tumor became necrotic, the growth studies were terminated. Results Cell specific bindinu of TGF-t0xins. The specificity of binding was investigated. Figure 1 shows that little specific I~IEGF binding was inhibited by 1 ng/ml EGF whereas almost all specific I~I-EGF binding was inhibited by 100 ng/ml EGF. The IC~0 for EGF was 10 ng/ml for cell lines NCI-H1264 (Fig. 1, left) and 12 ng/ml for cell line NCI-H838 (Fig. 1, right). T G F a was slightly less potent with ICs0 values of 22 and 48 ng/ml respectively for cell lines NCI-H1264 and NCI-H838. A m u t a n t form of TGFa-PE40 that does not inhibit protein synthesis because of a defective ADP-ribosylation domain, TGFa-PE40 ~p~a had an ICso value of 90 ng/ml for both cell lines whereas TGFa-PE40 had an ICs0 value of 1100 ng/ml for both NCI-H1264 and H838 cell lines. These data suggest that the T G F a binding domain of the chimeric toxins binds with high affinity to the EGF receptor. NCI-H1264

NCI-H838

10C

100

.o I,u .L

½ 50

50

8o. ~o

ffl

,

½

I

-8

-7

-F

(Factor), Log g/ml

I -9

I -8

~ -7

I -6

(factor), Log g/ml

Fig. i . Specificity of binding. The ability of EGF (o), T G F a (o),TGFa-PE40 ~ (A) and TGFa-PE40 (A) to inhibit specific I~I-EGF binding to cell line NCI-H1264 (left) and NCI-H838 (right) is indicated. The mean value + S.E. of 3 determinations each performed in quadruplicate is indicated.

448

TGFa-PEO40 Inhibits NSCLC Growth

Vol. 54, No. 7, 1994

Table I shows that TGFa-PE40 also inhibited EGF binding to cell lines NCI-H23, H1299, H157, H226, H460 and H727 with high affmity (IC~o = 500-3000 ng/ml). Similar binding results were obtained for TGF~-PE38 and TGFa-PE40 ~ fusion proteins. These data indicate that TGFa-PE40 analogues bind to EGF receptors present on adenocarcinoma, large cell carcinoma, squamous cell carcinoma, NSCLC-neuroendocrine and lung carcinoid cell lines. Table I. Effect of TGFa-PE analogues on 125I-EGF binding to NSCLC cell lines Cell line (TGFa-PE40) Adenocarcinoma NCI-H23 NCI-H838 NCI-H1264 Large Cell carcinoma NCI-H1299 Squamous cell carcinoma NCI-H157 NCI-H226 NSCLC-Neuroendocrine NCI-H322 NCI-H460 Lung carcinoid NCI-H727

ICs0,nM (TGFa-PE38)

(TGFa-PE40 ~p55a)

500 1000 1000

250 n.d. n.d.

300 n.d. n.d.

600

420

150

800 3000

400 100

110 50

n.d. 500

1000 250

400 n.d.

1000

2000

500

The mean value of 3 determinations to inhibit specific I~I-EGF binding (ICs0, ng/ml) is indicated; not determined, n.d. Protein svnthesis inhibition activitv of TGFa-toxin fusion uroteins. The TGFa-PE40 complex is internalized into endocytic vesicles and acidification of the vesicles enables the toxin to be processed and enter the cytosol, blocking protein synthesis (10). Here 0.1 ng/ml TGFa-PE40 had little effect on a~S-methionine incorporation into newly synthesized proteins (Fig. 2) whereas almost all aSS-methionine incorporation was inhibited by 100 ng/ml TGFa-PE40. The EDs0 of TGFa-PE40 was 11 ng/ml and 20 ng/ml for cell lines NCI-H1264 and NCI-H727 respectively. TGFa-PE38 was slightly more potent with ED~0 values of 5 and 6 ng/ml for cell lines NCI-H1264 and H727 respectively. TGFaPE40 ~ p ~ had no effect on ~S-methionine uptake at 10 ng/ml and slightly inhibited amino acid uptake at 100 ng/ml. Similar results were obtained for cell lines NCI-H23, H838, H1299, H157 and H322. TGFa-PE40 inhibition of colony formation, The effect of TGFa-PE40 on NSCLC growth was investigated. Figure 3 (left) shows t h a t numerous colonies stained with 0.1% p-iodonitrotetrazolium violet. One ng/ml TGFaPE40 abolished colony formation of NCI-H157 (Fig. 3, middle). The effect of TGFa-PE40 on colony formation was dose dependent. One pg/ml TGFa-PE40 had little effect on NCI-H727 colonies, whereas few colonies were present at 1 ng/ml. The LD~owas 0.008 ng/ml for TGFccPE40 using NCI-H727 whereas for NCI-H157 and NCI-H838 the LDso was 0.03 and 0.002 ng/ml respectively. Similarly, the LD60for TGFa-PE38 ranged from 0.002-0.1 ng/ml whereas TGFa-PE40 ~p6~ was inactive (LDso > 10 ng/ml). Similar data were obtained for NCI-H460, H1264 and H1299.

Vol. 54, No. 7, 1994

TGFa-PEO40 Inhibits NSCLC Growth

A

449

B 100

100 q)

G)

t-

t-

"E

.2 t-

0 J~

50

50

I t~ t.O tO

I O) tO ¢0

-10

-9

(factor),

(Factor), Log g/ml

-8

-7

Log g/ml

Fig. 2. Inhibition of protein synthesis. The ability of increasing doses of TGFa-PE40 (o), TGFa-PE38 (o) and TGFa-PE40 ~ a (A) to inhibit ~S-methionine incorporation into newly synthesized proteins was determined using cell lines NCI-H1264 (A) and NCIH727 (B). The mean value + S.E. of 4 determinations is indicated.

7~4

-::

' CONTROL .

Oe I

* Q"

•

•

+1 ng/ml TGFa-PE4o

.

, ~•. m O

×

4 .0*Oe

e-

5.(

._=

• . •

"" I, • .*we

i °

2.1

Q-

el.

o

Q:

-12

-11

-10

"-9

(TGF alpha-PE 40), Log g / m l

Fig. 3. Inhibition of colony formation. Large numbers of NCI-H157 colonies formed in the absence (left) but not presence (middle) of i ng/ml TGFa-PE40. (Right) The number of NCI-H838 (e), H157 (o) and H727 (A) colonies was determined as a hmction of TGFcc-PE40 concentration. The mean value + S.E. of 3 determinations is indicated.

450

TGFa-PEO40 Inhibits NSCLC Growth

Vol. 54, No. 7, 1994

Anti-tumor activitv of TGFa-toxin fusion uroteins. The effect of TGFcc-PE40 was investigated in vivo. Xenografts formed 2 weeks after injection of NCI-H838 cells. The tumors grew exponentially and wereapproximately 1700 m m a after 7 weeks (Fig 4, left). TGFa-PE40 (1 ~g/day) was infused using an Alzet micropump during week 3 and t u m o r growth was significantly inhibited at weeks 6 and 7 relative to control ,nimals. Thus TGFa-PE40 slowly inhibited t u m o r growth so t h a t xenograft formation was significantly decreased several weeks after injection. Similar results were obtained using TGF(~-PE38 (Fig. 4, right). Administration of TGFa-PE40 ~ did not result in any anti-tumor activity. As a control, ~nlmals were sacrificed at the end of the experiment and weighed. Table II shows t h a t TGFa-PE38 but not TGFa-PE40 ~pS~ significantly inhibited t u m o r weight but had no effect on body weight.

? o

~

Pump

E E

Implinta t i ° n

PBS

/

/:

~~~ E •xE

E * ~

p-.

2

3

4 5 Time, weeks

6

TGFa-PE40

I

PBS ~TGFaPE40Asp553

Pump

E L~.--qr _ A 2 3 4 Time,weeks

7

TGFa-PE38

5

Fig. 4. Inhibition of NSCLC growth in vivo. (Left) NCI-H838 xenograft volume was determined weekly in nude mice treated with PBS (o) and TGFa-PE40 (e). (Right) NCI-H157 xenograft volume was determined weekly in nude mice treated with PBS (o), TGFa-PE38 (m) and TGFa-PE40 m ~ (A). The m e a n value of 3 determinations is indicated; p < 0.05, * Table II. Effect of TGFa-PE40 analogues on nude mice.

Agent injected

t u m o r weight, g

body weight, g

PBS TGFc~-PE40A~p55s TGFa-PE38

4.3 + 0.7 4.8 + 0.5 2.1 + 0.9*

25.0 + 1.4 24.0 + 1.4 23.7 + 0.6

The m e a n weight _+ S.E. of 3 determinations is indicated; p < 0.05,*.

Vol. 54, No. 7, 1994

TGFa-PEO40 Inhibits NSCLC Growth

451

Discussion The present study indicates that TGFa-PE40 inhibits the growth of NSCLC cells in vitro and in vivo. The effects of TGFa-PE40 are receptor mediated. TGFa-PE40, which is comprised of a 360 amino acid toxin and a 50 amino acid TGFa, bound with high affinity to NSCLC cells (ICs0 = 500-1000 ng/ml). In contrast the ICso for TGFa, which is comprised of 50 amino acids, is 20-50 ng/ml. These data suggest that on a molar basis T G F a and TGFa-PE40 bind with similar affinities to EGF receptors (ICs0 = 7 nM). Similar results were obtained for TGFa-PE38, whereas TGFa-PE40 A ~ bound with slightly higher affinity (ICso = 7 and 2 nM respectively). Previously it was found that TGFa-PE40 but not PE40T G F a bound with high affinity to A431 and KB cells (10). Thus, it is best to delete domain Ia of Pseudomonas toxin, and couple it to the C-terminal of T G F a to obtain a chimeric protein which binds with high affinity to EGF receptors (18). Other hybrid toxins have been generated for interleukin-2 and acidic fibroblast growth factor (19,20). TGFcc-PE40, similar to EGF, is internalized into NSCLC cells. Previously, it was demonstrated that domain II of Pseudomonas exotoxin is important in translocation (13). This is followed by proteolytic cleavage of PE near arginine 279 after internalization into acidic endosomes (21). The 37 kDalton fragment that results from proteolytic cleavage is reduced and transferred by endocytotic vesicles to the trans Golgi. The C-terminal of domain III of PE directs the toxin into the endoplasmic reticulum where the 37 kDalton fragment utilizes protein transport to gain access to the cytosol. Domain III can then ADPribosylate elongation factor 2 (EF2) inactivating protein synthesis. Here 35S-methionine uptake into new proteins was blocked by 100 ng/ml TGFa-PE40 and half maximally reduced by 10 ng/ml TGFa-PE40. In vitro. TGFa-PE40 was cytotoxic for NSCLC cells. Because 0.05 ng/ml TGFa-PE40 half maximally inhibited amino acid uptake in A431 ceils which have 3 x 106 EGF receptors/cell, TGFa-PE40 potency is a hmction of EGF receptor concentration (13). Similar results were obtained for TGFa-PE38 whereas TGFa-PE40 ~pS~ had no effect. These data indicate that even though TGFcc-PE40~p~Sa binds with high affmity to the EGF receptor it does not have toxic activity. Because cytesolic PE can be utilized to inactivate EF2, protein synthesis is rapidly blocked resulting in death of rapidly growing cells. Here NSCLC colonies formed in the presence of i pg/ml TGFa-PE40 but not 1 ng/ml TGFa-PE40. Similar data were obtained using TGFa-PE38 whereas TGFa-PE40 A~53 was inactive at 10 ng/ml. Surprisingly little TGFa-PE40 was required to half maximally inhibit colony formation (LDso = 0.01 ng/ml) whereas more TGFa-PE40 was required to inhibit amino acid uptake (ED~0 = 10 ng/ml) and 125I-EGF binding (ICso = 500 ng/ml). This may result because the clonogenic assay took 2 weeks to perform whereas the amino acid uptake and receptor assay took 16 and 0.5 hr to perform respectively. As a result, in the clonogenic assay the TGFcc-PE40 had ample time to undergo internalization, enzymatic catalysis and ADP ribosylation of EF2. TGFa-PE40, but not TGFa-PE40 ~ slowed NSCLC xenograi~ formation. TGFaPE40 was released continuously by Alzet micropumps (1 lag/day) for 7 days. Previously, it was found that this dose of TGFa-PE40 inhibited A431 tumors in nude mice whereas 4.5 lag/day was lethal in most ~nimals (13). TGFa-PE40 was stable in the micropump and a steady state concentration of 8 ng/ml of TGFa-PE40 was obtained in the serum. TGFa-PE40 did not reduce NSCLC xenografts but slowed their growth. In contrast, TGFa-PE40 reduced A431 xenografts possibly because A431 cells have 3 x 106 EGF receptors/cell resulting in more TGFcc-PE40 internalization (13). Nonetheless TGFa-PE40 and TGFcc-PE38 significantly inhibited the growth of NSCLC cells which have 105 EGF receptors/cell. The effects of TGFa-PE38 were specific in that tumor weight but not total body weight was significantly reduced in nude mice.

452

TGFa-PEO40 Inhibits NSCLC Growth

Vol. 54, No. 7, 1994

Previously we found that mAb 108 slows NSCLC growth in vitro and in vivo (7). This antibody, which is not readily internalized, functions as an EGF receptor antagonist. Upon removal of the antibody, xenograft growth in nude mice readily increases. In contrast, TGFa-PE40 is designed to be internalized and kill NSCLC cells. In this regard, TGFa-PE40 may have killed many NSCLC cells during the week of its release from Alzet micropumps resulting in reduced xenograft growth during the succeeding weeks. These studies suggest that EGF receptors may be utilized to develop novel therapies for NSCLC. Additional growth factors for NSCLC include insulin-like growth factor I (IGF-I) and vasoactive intestinal peptide (VIP) (22). The growth of NSCLC cells is inhibited by aiR-3, a monoclonal antibody against the IGF-I receptor, and VIPhybrid, a synthetic peptide VIP receptor antagonist. I~I-IGF-I and ~I-VIP bind to approximately 10,000 and 100,000 sites/cell (23). Because IGF-I and VIP binding sites are readily internalized at 37°C, it may be possible to kill NSCLC cells using IGF-I or VIP toxin conjugates. AcknowledL,ments This research is supported in part by NCI grants CA-48071 and CA-53477. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

R. DERYNCK, A.B. ROBERTS, M.E. WINKLER, E.Y. CHEN and D.V. GOEDDEL, Cell 38 287-297 (1984). J.M. SIEGFRIED and S.M. OWENS, Cancer Res. 48 4976-4981 (1988). K. IMANISHI, K. YAMAGUCHI, M. KURANAMI, E. KYO, T. HOZUMI and K. ABE, J. Natl. Cancer Inst. 81 220-223 (1989). T.A. LIBERMANN, H.R. NUSBAUM, N. RAZON, R. KRIS, I. LAX, H. SOREQ, N. WHITTLE, M.D. WATERFIELD, A. ULLRICH and J. SCHLESSINGER, Nature ~13 144-147 (1985). T HUNTER and J.A. COOPER, Ann. Rev. Biochem. 54 897-930 (1985). T.W. MOODY, M. LEE, R.M. KRIS, F. BELLOT, G. BEPLER, H. OIE and A. GAZDAR, Journal of Cellular Biochemistry 48 139-147 (1990). M. LEE, M. DRAOUI, F. ZIA, A.F. GAZDAR, H. OIE, C. TARR, F. BELLOT, R. KRIS and T.W. MOODY, J. Natl. Cancer Inst. Monographs 13:117-123 (1992). J. HWANG, D.J. FITZGERALD, S. ADHYA and I. PASTAN, Cell 48 129-136 (1987). T. KONDO, D. FITZGERALD, V.K. CHAUDHARY, S. ADHYA and I. PASTAN, J. Biol. Chem. 26~ 9470-9475 (1988). I. PASTAN, V. CHAUDHARY and D.J. FITZGERALD, Ann. Rev. Biochem. 61 331-354 (1992). C.B. SIEGALL, Y.H. XU, V.K. CHAUDHARY, S. ADHYA, D. FITZGERALD and I. PASTAN, FASEB J. ~ 2647-2652 (1989). D.C. HEIMBROOK, S.M. STIRDIVANT, J.D. AHERN, N.L. BALISHIN, D.R. PATRICK, G.M. EDWARDS, D. DEFEO-JONES, D.J. FITZGERALD, I. PASTAN and A. OLIFF, Proc. Natl. Acad. Sci. USA 87 4697-4701 (1990). L.H. PAI, M.G. GALLO, D.J. FITZGERALD and I. PASTAN, Cancer Res. 51 2808-2812 (1991). C.B. SIEGALL, V.K. CHAUDHARY, D.J. FITZGERALD and I. PASTAN, J. Biol. Chem. 264 14256-14261 (1989). C.B. SIEGALL, V.K. CHAUDHARY, D.J. FITZGERALD and I. PASTAN, Proc. Natl. Acad. Sci. USA 85 9738-9742 (1988). D.N. CARNEY, A.F. GAZDAR, G. BEPLER, J.G. GUCCION, P.J. MARANGOS, T.W. MOODY, M.H. ZWEIG and J.D. MINNA, Cancer Res. 45 2913-2923 (1985).

Vol. 54, No. 7, 1994

17. 18. 19. 20. 21. 22. 23.

TGFa-PEO40 Inhibits NSCLC Growth

453

S. MAHMOUD, J. STAT,EY, J. TAYLOR, A. BOGDEN, J.P. MOREAU, D. COY, I. AVIS, F. CUTTITTA, J.L. MULSHINE and T.W. MOODY, Cancer Res. 51 1298-1802 (1991). C.B. SIEGALL, D.J. FITZGERALD AND I. PASTAN, Sem. Cancer Biol. 1 345-350 (1990). H. LORBERBOUM-GALSKI, D. FITZGERALD, V. CHAUDHARY, S. ADHYA and I. PASTAN, Proc. Natl. Acad. Sci. USA 85 1922-1926 (1988). C.B. SIEGALL, S. EPSTEIN, E. SPEIR, T. HLA, R. FOROUGH, T. MACIAG, D.J. FITZGERAT,D and I. PASTAN, FASEB J. 5 2843-2849 (1991). M. OGATA, V.K. CHAUDHARY, I. PASTAN and D.J. FITZGERALD, J. Biol. Chem. 265 20678-20685 (1990). T.W. MOODY, F. ZIA, D. BRENNEMAN, M. FRIDKIN, A. DAVIDSON and I. GOZES, Proc. Natl. Acad. Sci. USA 90:4345-4349 (1993). F. ZIA, F. CUTTITTA, J. MULSHINE and T. MOODY, Cancer Res. Proc. 33, 170 (1992).