Inhibition of Interleukin 8 Attenuates Angiogenesis
in Bronchogenic Carcinoma
By Daniel R. Smith,* Peter J. Polverini,~ Steven L. Kunkel,$
Mark B. Orringer, II Richard I. Whyte, II Marie D. Burdick,*
Carol A. Wilke,* and Robert M. Strieter*
From the *Department of Internal Medicine, Division of Pulmonary and Critical Care
Medicine, University of Michigan Medical School, Ann Arbor, Michigan 48109-0360; the
*Section of Oral Pathology, University of Michigan Dental School, Ann Arbor, Michigan
48109-I078; the SDepartment of Pathology, University of Michigan; and the UDepartment of
Surgery, Section of Thoracic Surgery, University of Michigan, Ann Arbor, Michigan 48109
We investigated the role of interleukin 8 (IL-8) in mediating angiogenesis in human bronchogenic
carcinoma. Increased quantities of IL-8 were detected in tumor tissue as compared with normal
lung tissue. Immunohistochemical staining of tumors revealed primary localization of IL-8 to
individual tumor cells and demonstrated the capacity of tumor to elaborate IL-8. Functional
studies that used tissue homogenates of tumors demonstrated the induction of both in vitro
endothelial cell chemotaxis and in vivo corneal neovascularization. It is important to note that
the addition of neutralizing antisera to IL-8 to these assays resulted in the marked and specific
attenuation of these responses. Our observations definitively establish IL-8 as a primary mediator
of angiogenesis in bronchogenic carcinoma and offer a potential target for immunotherapies against
incidence of many other malignancies has declined or remained
stable, the occurrence of bronchogenic carcinoma has esca-
lated to near-epidemic proportions. Over 150,000 new cases
are diagnosed and an equal number of deaths are attributable
to bronchogenic carcinoma in the U.S. annually (1). Despite
attempts to advance early diagnosis and employ combination
therapies, the clinical response of this tumor yields an overall
5-yr survival rate for lung cancer patients of <15%. A new
approach to therapy is dearly needed.
Tumor vascularization is crucial for the pathogenesis of
all solid malignancies. In the absence of local capillary
proliferation, tumors may not grow beyond 2-3 mm in
diameter (2). Folkman first proposed in 1972 that tumors
are angiogenesis-dependent, with tumor growth requiring
concomitant increases in vascular supply. Previous studies have
attempted to investigate the complexities of tumor neovas-
cularization (3). Certain tumors have been found to produce
factors that are directly angiogenic, whereas others may de-
pend upon vascularization induced by products of responding
inflammatory cells (4-6). IL-8 is a cytokine of the C-X-C
supergene family that has been classified primarily by pro-
inflammatory and leukocyte chemotactic properties (7-9). A
variety of cell populations, including neutrophils (10),
mononuclear (11, 12), stromal (13), endothelial (14), and ep-
he leading cause of malignancy-related mortality in the
United States is bronchogenic carcinoma (1). While the
ithelial cells (15), and hepatocytes (16) may serve as sources
for IL-8. The recent characterization of potent angiogenic
properties has renewed interest in the role of IL-8 in a variety
of settings (17, 18). In the present study we demonstrate IL-8
is a primary mediator of angiogenesis in human broncho-
Materials and Methods
Antigenic Determination of lL-8 Content of Normal Lung and Bron-
chogenic Tumor Tissue. Tissue specimens were obtained from con-
senting individuals undergoing thoractomy for suspected primary
bronchogenic carcinoma in accordance with the University of
Michigan I.R.B. approval. Samples of tumor, and normal lung distal
to tumor, were homogenized in PBS upon recovery from the oper-
ating room. Specimens were then filtered through 0.45-micron
Sterile Acrodiscs (Gelman Sciences Inc., Ann Arbor, MI) and frozen
at -70~ until thawed and then assayed. A specific ELISA for
Ib8 was employed to determine IL-8 content of tissue samples as
previously described (19). Samples were run in parallel for total pro-
tein content (Pierce Chemical Co., Rockford, IL). Determinations
were expressed as nanograms of I1.,8 per milligrams total protein.
Results were tabulated based on final histological diagnosis as de-
termined by University Hospital pathologists.
Immunohistochemical Localization of Antigenic IL-8. Fresh tissue
specimens obtained at time of thoractomy were fixed in 4% para-
formaldehyde for 24 h before transferring to 70% ethanol. Paraffin-
embedded tissue sections were dewaxed with xylene and rehydrated
j. Exp. Med. y The Rockefeller University Press i 0022-1007/94/05/1409/07 $2.00
Volume 179 May 1994 1409-1415
through graded concentrations of ethanol. Samples were then stained
for IL-8 using a modification of our previously described technique
(20). Briefly, nonspecific binding sites were blocked with normal
goat serum (BioGenex Laboratories, San Ramon, CA) before
washing and applying a 1:1,000 concentration of either control
(rabbit) or rabbit anti-human IL-8 serum. Slides were then rinsed
and overlaid with secondary biotinylated goat anti-rabbit IgG (1:35)
and then incubated. After washing twice with Tris-buffered saline,
slides were overlaid with a 1:35 dilution of alkaline phosphatase
conjugated to streptavidin (BioGenex Laboratories), and incubated.
Fast Red (BioGenex Laboratories) reagent was used for chromo-
genic localization of Ib8 antigen. After optimal color development,
sections were immersed in sterile water, counterstained with Mayer's
hematoxylin, and coverslipped using an aqueous mounting solution.
Endothelial Cell Chemotaxis Responses to IL.8 Standard and Tissue
Specimens. Endothelial cell chemotaxis was performed in 48-well,
blind well chemotaxis chambers (Nucleopore Corp., Pleasanton,
CA) as previously described (21). Nucleopore Chemotaxis mem-
branes (5 micron pore size; Nucleopore Corp.) were prepared by
soaking them sequentially in 3% acetic acid overnight and for 2 h
in 0.1 mg/ml gelatin. Membranes were rinsed in sterile water, dried
under sterile air, and stored at room temperature for up to 1 too.
Bovine adrenal gland capillary endothelial cells (BCE), 1 maintained
in gelatin-coated flasks in DME with 10% FCS were used as the
target cells. 24 h before use BCE were starved in DME with 0.1%
BSA. 25 #1 of cells, suspended at a concentration of 106 per ml
in DME with 0.1% BSA were dispensed into each of the bottom
wells. A chemotaxis membrane was positioned atop the bottom
wells, chambers were sealed, inverted, and incubated for 2 h to allow
cells to adhere to the membrane. Chambers were then reinverted,
50/~1 test media was dispensed into the top wells and reincubated
for an additional 2 h. Membranes were then fixed and stained with
a Dill-Quick staining kit (American Scientific Products, McGraw
Park, IL) to enumerate membrane-bound cells, and cells that had
migrated through the membrane to the opposite surface were
counted. Four replicates, 10 fields/replicate, were tested for each
sample, and experiments were repeated at least three times. Results
were expressed as the total number of endothelial cells that migrated
across the filter in 10 high power (400 x) fields. IL-8 neutralization
studies used our previously well-characterized polyclonal rabbit
anti-human IL-8 antibody (13, 18, 20). Additional neutralization
studies employed the commercially available neutralizing goat
anti-human TGF-c~ antibody and rabbit anti-basic fibroblast growth
factor (bFGF) (R&D Systems, Minneapolis, MN).
Corneal Micropocket Model of Angiogenesis. In vitro angiogenic
activity was assayed in the avascular cornea of F344 female rats eyes,
as previously described (22, 23). Briefly, 5 mg total protein of each
specimen was combined with a equal volume of sterile Hydron
casting solution, and 5-ml aliquots were pipetted onto the surface
of l-ram Teflon rods glued to the surface of a glass petri dish. Pellets
were air-dried in a laminar flow hood (1 h) and refrigerated over-
night. Before implantation pellets were rehydrated with a drop of
lactated Ringer's solution. Animals were anesthetized with meto-
lane and injected with sodium pentobarbital intraperitoneally. A
retrobulbar injection of 0.1 ml of 2% lidocaine was made before
intracorneal implantation of the Hydron pellet into a surgically
created intracomeal pocket "~1.5 mm from the limbus. The animals
were examined daily with a stereomicroscope. 7 d after implanta-
tion, animals were reanesthetized and perfused sequentially with
1 Abbreviations used in this paper: BCE, bovine adrenal gland capillary
endothelial cells; bFGF, basic fibroblast growth factor; TP, total protein.
lactated Ringer's solution followed by colloidal carbon. Corneas
were harvested, flattened, and photographed (33 x). Positive neo-
vascularization responses were recorded only if sustained directional
ingrowth of capillary sprouts and hairpin loops towards the im-
plant were observed. Negative responses were recorded when ei-
ther no growth was observed or when only an occasional sprout
or hairpin loop displaying no evidence of sustained growth was
detected. Animals were handled in accordance with the University
of Michigan U.L.A.M. protocols.
Results and Discussion
IL-8 Is Increased in Tumor Tissue.
tion using an ELISA to detect IL-8 in human tissue homo-
genates of both normal lung and nonsmall cell bronchogenic
carcinomas. As shown in Fig. 1, antigenic IL-8 was observed
in a fourfold excess in tumor tissue as compared with normal
lung tissue (expressed as nanograms per milligrams of total
protein [TP]). Normal lung tissue contained 2.604 _+ 0.729
ng/mg TP oflL-8, as compared with 8.499 _+ 2.327 ng/mg
TP for tumor samples (n = 52). There were similar eleva-
tions of antigenic IL-8 for tumors of adenocarcinoma histology
as compared with those of squamous cell phenotype, 6.986 +
3.314 ng/mg TP and 9.752 _+ 2.956 ng/mg TP, respectively.
Immunohistochemistry Localizes IL-8 to Individual Tumor Cells.
We next used immunohistological localization to demonstrate
the cellular origins of antigenic IL-8 in tumor tissue. Previous
studies have convincingly demonstrated that human bron-
chogenic tumor cell lines may directly elaborate IL-8 (20,
24). We exploited immunohistochemical localization to iden-
tify cellular sources of the increased IL-8 in tumor tissue.
Stains of tumor sections typically demonstrate heterogeneous
tumor cell production of IL-8 in both adenocarcinomas and
squamous cell carcinomas of the lung (Fig. 2). In addition,
our results reveal that stromal cells within the local des-
moplastic response may also serve as cellular sources for IL-8,
especially in squamous cell cardnomas. It is important to note
We began our evalua-
Normal Lung Tumor Tissue AdenoCA
Figure 1. IL-8 content of tissue samples of normal lung and broncho-
genic lung tumors. Subgroup analysis by tumor histological diagnosis.
1410 Inhibition of II.-8 Attenuates Angiogenesis in Bronchogenic Carcinoma
Figure 2. Immunohistochemical
staining of 1I.-8 in tumor tissue. (A)
Normal lung with preimmune serum
(200 x). (B) Normal lung with anti-IL-8
(200 x). (C) Squamous cell carcinoma
with preimmune serum (200 x). (D, E,
and F) Squamous cell carcinoma with
anti-lL-8 antibody (100, 200, and 400x,
respectively). (G) Adenocarcinoma with
preimmune serum (200 x). (/4, I,J) Ad-
enocarcinoma with anti-IL-8 antibody
(100, 200, and 400• respectively).
1411 Smith et al.
that these specific findings may be reflected in the different
clinical behaviors of squamous cell and adenocarcinomas. The
more aggressive course of adenocarcinomas could result from
its capacity to generate suf~cient angiogenic signal by tumor
IL-8 production, independent from contributions from the
surrounding tissue response. The observation of a minimal
inflammatory cell infiltrate in the tumor samples, despite the
presence of significant quantities of IL-8, is unexpected and
of particular interest.
Endothelial Cell Chemotaxis Activity to Tumor Homogenates
Is Blocked by Anti-IL-8.
To determine the relative contribu-
tion of IL-8 to the total angiogenic signal from broncho-
genic carcinomas we employed a strategy using both in vitro
and in vivo models of angiogenesis. First, concentrated tissue
homogenates from normal lung, adenocarcinoma, and squa-
mous cell carcinomas, as well as tissue extracts from suspen-
sions of a bronchoalveolar carcinoma cell line (ATCC-A549;
American Type Culture Collection, R.ockville, MD), were
evaluated for chemotactic activity toward BCE. Results for
tissue samples were expressed as a percentage of chemotaxis
activity induced by a standard of 50 ng/ml recombinant human
IL-8. As illustrated in Fig. 3 A, samples from adenocarci-
noma, squamous cell carcinoma, and A549 cells demonstrated
62, 84, and 86% respectively, of the maximal BCE chemotaxis
induced by the IL-8 standard. BCE chemotaxis in response
to media alone was 16% of the IL-8 standard. The addition
of neutralizing antibodies to IL-8 reduce the endothelial cell
chemotactic response to each of the test samples by '~66%
(to a background level of 33% of the standard activity) (13,
20). Specifically, BCE chemotaxis in response to tissue samples
plus anti-IL-8 was attenuated to similar background levels
of 37, 39, and 32% of the standard, respectively, for adeno-
carcinoma, squamous cell carcinoma, and A549 tissue and
cell homogenates. This data suggested that a significant por-
tion of tumor-generated endothelial cell chemotaxis is medi-
ated directly by IL-8. To further examine the angiogenic signal
from tumor specimens in an attempt to characterize the rela-
tive contributions of other known angiogenic factors such
as bFGF and TGF-cr a second series of experiments were
performed. Endothelial cell chemotactic studies were repeated
as before, with addition of neutralizing antibodies to either
IL-8, bFGF, or TGF-o~. Results are demonstrated in Fig. 3
B, standardized to the response generated from a standard
of 25 ng/ml bFGE Normal lung tissue generated only 40%
of standard chemotactic activity, and was not significantly
affected by the addition of any of the neutralizing antisera.
Again, tumor samples alone generated brisk chemotactic re-
sponses, with 198, 130, and 104% standard activity, respec-
tively, for samples of adenocarcinoma, squamous cell carci-
noma, and A549 cell line tissue. The addition of neutralizing
antibodies to IL-8 again resulted in significant reductions of
chemotactic responses to tumor tissue, with decreases to 75,
39, and 61% standard activity, respectively, for adenocarci-
noma, squamous cell carcinoma, and A549 samples. Anti-
bFGF antibodies had no significant effect on responses to
samples of A549 or squamous cell carcinoma tissue, but did
yield a reduction in chemotactic activity of adenocarcinoma
A549 + IL-8 Ab
SCC A + IL-8 Ab
Adenoca. + IL-8 Ab
Normal Lung + IL-8 Ab
IL-8 + Ab
scc A ~/#y
Ad ..... ~~.~ :
Normal Lung ~ :
IL-S (50ng~ml) ~////~/~~
OME + 0,1% BSA
i I , , i I , , ,
, , , I , , , i
Endothelial cell migration/10HPF
40 60 80 100
, , , !
0 ............... ,
'J O. *.s e, es
= += *- .-o .-
Figure 3. Endothelial cell chemotaxis. (A) Demonstrates responses stan-
dardized to 50 ng/ml IL8, of specimens of normal lung, adenocarcinomas,
squamous cell carcinomas, and A549 tissue in the presence or absence of
specific neutralizing antisera to Ib8. Also demonstrates response to pro-
rein control. (B) Endothelial cell chemotactic responses to normal and tumor
tissue, standardized to 25 ng/ml bFGF. Effect of addition of specific neu-
tralizing antisera to IL-8, bFGF, and TGF-~ is also demonstrated.
tissue from 198 to 137% standard activity. It is interesting
to note that the neutralization of TGF-o~ effected no change
in responses to adenocarcinoma or A549 tissue, but demon-
strated a significant reduction in the chemotactic response
1412 Inhibition of IL-8 Attenuates Angiogenesis in Bronchogenic Carcinoma
to squamous cell carcinoma tissue, from 130 to 71% stan-
Neovascularization Induced by Tumor Homogenates Is Attenu-
ated by Anti-IL-8.
We next evaluated the angiogenic signal
mediated by tumor-associated IL-8 in an in vivo model. The
previously well-characterized corneal micropocket modal in
the rat was employed (22, 23). Tissue samples and controls
were incorporated into Hydron (Interferon Sciences Inc., New
Brunswick, NJ) pellets and then embedded into the normally
avascular rat cornea (Fig. 4 and Table 1). Control samples
of either 50 ng recombinant IL-8 or 50 ng recombinant bFGF
induced positive corneal angiogenic responses of 80 and 100%,
respectively. Samples from normal lung, adenocarcinoma, squa-
mous cell carcinoma samples, and A549 cells produced posi-
tive angiogenic responses in 13 (1/8), 86 (6/7), 83 (5/6), and
100% (7/7), respectively. Anti-IL-8 antibodies completely
abrogated the angiogenic response to IL-8 controls (zero of
four corneas positive), but had no effect upon angiogenesis
induced by bFGF (three of three corneas positive). The addi-
tion of antibodies to IL-8 resulted in significant reduction
of angiogenic responses to tumor tissue samples. A549 cell
homogenates, which had yielded a 100% neovascularization
rate, demonstrated no angiogenic activity in the presence of
neutralizing antibodies to IL-8 (zero of five samples positive).
Neutralizing antibodies to IL-8 caused a reduction in tumor
sample angiogenesis of both adenocarcinoma and squamous
cell carcinoma with 80% (four of five) and 80% (four of five)
inhibition of angiogenesis, respectively. The addition of anti-
IL-8 antibodies to normal lung samples did not induce a posi-
tive angiogenic response (zero of four positive), whereas
normal lung samples with control antibodies had a 17% posi-
tive angiogenic response (one of six positive). It is important
to note that there was no infiltration of the corneal tissue
by inflammatory cells in any of the test samples or controls,
suggesting that the angiogenic responses were mediated en-
tirely by factors present in tumor tissue, rather than by any
additional contributions from infiltrating inflammatory cell
Figure 4. In vivo angiogenesis demonstrating inhibition of lung tumor extract-induced corneal neovascularization by IL-8 neutralizing antibody.
Representative photomicrographs of corneal neovascular responses. (A) Positive angiogenic response induced by A549 cellular homogenate. (B) Positive
angiogenic response induced by tissue homogenate from human bronchogenic squamous cell carcinoma. (C) Negative angiogenic response from normal
lung homogenate. (D) Markedly suppressed angiogenic response induced by squamous cell carcinoma tissue homogenate preincubated with neutralizing
1413 Smith et al.
Table 1. In Vivo Angiogenesis
Proportion of positive responses
Test sample (-) anti-IL-8 Ab (+) anti-IL-8 Ab
IL-8 (50 ng)
bFGF (25 ng)
A549 cell line
Composite of data demonstrating inhibition of lung tumor extract-induced
corneal neovascularization by IL-8 neutralizing antibody.
These studies demonstrate that a primary angiogenic signal
for bronchogenic carcinoma neovascularization is directly
mediated by tumor-associated IL-8. Other previously described
factors appear to also be involved to a lesser extent in tumor
angiogenesis, with perhaps some variation by tumor histology.
Although previous studies have demonstrated that various
tumors and tumor cell lines elaborate IL-8, the significance
of tumor production of a neutrophil chemotaxin remained
enigmatic (16, 20, 24-27). Our data, coupled with the re-
cent characterizations of IL-8 bioactivity, suggest that tumor
production of this potent angiogenic factor may be crucial
for the neovascularization necessary for the initiation and main-
tenance of tumor growth. Interestingly, other investigations
have noted increased angiogenic activity in hyperplastic le-
sions which precedes subsequent malignant transformation,
suggesting that increased IL-8 production may play a func-
tional role in, or possibly serve as a marker for, cytogenic
dysregulation (28, 29). The observation of minimal inflam-
matory cell infiltrate, despite an appropriate chemotactic signal,
remains perplexing and may reflect the attenuation of inflam-
matory cell chemotactic signal by other tumor-derived factors.
Our recent description of the production of IL-1 receptor
antagonist protein by human bronchogenic carcinoma lends
support to this concept (30). Finally, as the demand for fur-
ther vascular supply increases with continued tumor growth,
successful maintenance of vascularization may require exploi-
tation of normal inflammatory responses in order to provide
adequate angiogenic signal. Accordingly, tumors may also
elaborate additional factors to stimulate or enhance IL-8
production by other cells in a paracrine fashion, which may
be reflected in the immunohistochemical localization of IL-8
in stromal cells within the tumor matrix. Future studies will
explore the consequences of attenuation of tumor IL-8 produc-
tion as a possible therapeutic intervention for bronchogenic
This work was supported in part by National Institutes of Health grants HL-39926 (to P. J. Polverini),
HL-50057, HL-02401, and 1P50 HL-46487 (to R. M. Strieter), HL-31693 and HL-35276 (to S. L. Kunkel).
Address correspondence to Dr. Robert M. Strieter, Department of Internal Medicine, Division of Pulmo-
nary and Critical Care, Box 0360, University of Michigan Medical Center, 3916 Taubman Drive, Ann
Arbor, MI 46109-0360.
Received for publication 14 October 1993 and in revised form 12 January 1994.
1. Garfinkel, A. 1991. Cancer statistics and trends. In American
Cancer Society Textbook of Clinical Oncology. A.I. Holleb,
D.J. Fink, and G.P. Murphy, editors. American Cancer So-
ciety, Atlanta. 2-9.
2. Folkman, J., and R. Cotran. 1976. Relation of vascular prolifer-
ation to tumor growth. Int. Rev. Exp. Pathol. 16:207.
3. Folkman, J. 1985. Tumor angiogenesis. In Advances in Cancer
Research, Vol. 43. G. Klein and S. Weinhouse, editors. Aca-
demic Press, Inc., Orlando, FU 175-203.
4. Atherton, A. 1977. Growth stimulation of endothelial cells
by simultaneous culture with sarcoma 180 cells in diffusion
chambers. Cancer Res. 37:3619.
5. Fenselau, A., and R.J. Mello. 1976. Growth stimulation of
cultured endothelial cells by tumor cell homogenates. Cancer
6. Zetter, B.R. 1980. Migration of capillary endothelial cells is
stimulated by tumour-derived factors. Nature (Lond.). 285:41.
7. Baggiolini, M., A. Walz, and S.L. Kunkel. 1989. Neutrophil-
activating peptide-1/interleukin 8, a novel cytokine that acti-
vates neutrophits. J. Clin. Invest. 84:1045.
8. Larsen, C.G., A.O. Anderson, E. Appella, J.J. Oppenheim,
and K. Matsushima. 1989. Neutrophil activating protein
(NAP-l) is also chemotactic for T lymphocytes. Science (Wash,
9. Matsushima, K., K. Morishita, T. Yoshimura, S. Lavu, Y.
Kobayashi, W. Lew, E. Appella, H.F. Kung, E.J. Leonard, and
J.J. Oppenheim. 1988. Molecular cloning of a human mono-
cyte-derived neutrophil chemotactic factor (MDNCF) and the
induction of MDCNF mRNA by interleukin 1 and tumor
necrosis factor. J. Extx Med. 167:1883.
1414 Inhibition of Ib8 Attenuates Angiogenesis in Bronchogenic Carcinoma
10. Bazzoni, F., M.A. Cassatella, F. Rossi, M. Ceska, B. Dewald, Download full-text
and M. Baggiolini. 1991. Phagocytosing neutrophils produce
and release high amounts of the neutrophil-activating peptide
1/interleukin 8. J. Exp. Med. 173:771.
11. Gregory, H., J. Young, J.M. Schroder, U. Mrowietz, and E.
Christophers. 1988. Structure determination of a human lym-
phocyte derived neutrophil activating peptide. Biochem. Bio-
phys. Res. Comm. 151:883.
12. Strieter, R.M., S.W. Chensue, M.A. Basha, T.J. Standiford,
J.P. Lynch, M. Baggiolini, and S.L. Kunkel. 1990. Human al-
veolar macrophage gene expression of interleukin-8 by tumor
necrosis factor-or, lipopolysaccharide, and interleukin-l~. Am.
J. Respir. Cell Mol. Biol. 2:321.
13. Strieter, R.M., S.H. Phan, H.J. Showell, D.G. Remick, J.P.
Lynch III, M. Genord, C. Raiford, M. Eskandari, lL.M. Marks,
and S.L. Kunkel. 1989. Monokine-induced neutrophil chemo-
tactic factor gene expression in human fibroblasts.J. Biol. Chem.
14. Strieter, K.M., S.L. Kunkel, H.J. Showell, D.G. Remick, S.H.
Phan, P.A. Ward, and K.M. Marks. 1989. Endothelial cell gene
expression of a neutrophil chemotactic factor by TNF-alpha,
LPS, and IL-l-beta. Science (Wash. DC). 243:1467.
15. Elner, V.M., K.M. Strieter, S.G. Elner, M. Baggiolini, I. Lindley,
and S.L. Kunkel. 1990. Neutrophil chemotactic factor (I1,-8)
gene expression by cytokine-treated retina pigment epithelial
cells. Am. J. Pathol. 136:745.
16. Thornton, A.J., R.M. Strieter, I. Lindley, M. Baggiolini, and
S.L. Kunkel. 1990. Cytokine-induced gene expression of a neu-
trophil chemotactic factor/IL-8 in human hepatocytes. J. Im-
17. Hu, D.E., Y. Hori, and T.P.D. Fan. 1993. Interleukin-8 stimu-
lates angiogenesis in rats. Inflammation. 17:135.
18. Koch, A.E., P.J. Polverini, S.L. Kunkel, L.A. Harlow, L.A.
Pietra, V.M. Elner, S.G. Elner, and IL.M. Strieter. 1992.
Interleukin-8 as a macrophage-derived mediator of angiogen-
esis. Science (Wash. DC). 258:1798.
19. Bacon, K.B., J. Westwick, and lL.D.K. Camp. 1989. Potent
and specific inhibition of II,-8. IL-lalpha, and IL-lbeta-induced
in vitro human lymphocyte migration by calcium channel an-
tagonists. Biochem. Biophy. Res. Comm. 165:349.
20. Standiford, T.J., S.L. Kunkel, M.A. Basha, S.W. Chensue, J.P.
Lynch III, G.B. Toews, J. Westwick, and lL.M. Strieter. 1990.
Interleukin-8 gene expression by a pulmonary epithelial cell
line: a model for cytokine networks in the lung.J. Clin. Invest.
21. Koch, A.E., S.J. Leibovich, and P.J. Polverini. 1986. Stimula-
tion of neovascularization by human rheumatoid synovial tissue
macrophages. Arthritis Rheum. 29:471.
22. Koch, A.E., J. Burrows, M. Cho, P.J. Polverini, and S.J.
Leibovich. 1991. Thiol-containing compounds inhibit the pro-
duction of monocyte/macrophage-derived angiogenic activity.
Agents Actions. 34:350.
23. Polverini, P.J., P.S. Cotran, M.A.J. Gimbrone, and E.R. Un-
anue. 1977. Activated macrophages induce vascular prolifera-
tion. Nature (Lond.). 269:804.
24. Hotta, K., K. Hayashi, J. Ishikawa, M. Tagawa, K. Hashimoto,
S. Mizuno, and K. Suzuki. 1990. Coding region structure of
interleukin-8 gene of human lung giant cell carcinoma LU65C
cells that produce LUCT/interleukin-8: homogeneity in inter-
leukin-8 genes. Immunol. Lett. 24:165.
25. VanMeir, E., M. Ceska, F. Effenberger, A. Walz, E. Grouz-
mann, I. Desbaillets, K. Frei, A. Fontana, and N. deTribolet.
1992. Interleukin-8 is produced in neoplastic and infectious
diseases of the human central nervous system. Cancer Res. 52:
26. Kasahara, T., N. Mukaido, K. Yamashita, H. Yagisawa, T. Aka-
hoshi, and K. Matsushima. 1991.11,-1 and TNF-alpha induc-
tion of IL-8 and monocyte chemotactic and activating factor
(MCAF) mKNA expression in a human astrocytoma cell line.
27. Kaashoek, J.G., lL. Mout, J.H. Falkenburg, K. WiUemze, W.E.
Fibbe, and J.E. Landegent. 1991. Cytokine production by the
bladder carcinoma cell line 5637: rapid analysis of mKNA ex-
pression levels using a cDNA-PCR procedure. Lymphokine
Cytokine Res. 10:231.
28. Folkman, J., K. Watson, D. Ingber, and D. Hanahan. 1989.
Induction of angiogenesis during the transition from hyper-
plasia to neoplasia. Nature (Lond.). 339:58.
29. Maiorana, A., and P.M. Gullino. 1978. Acquisition of angio-
genic capacity and neoplastic transformation in the rat mam-
mary gland. Cancer Res. 38:4409.
30. Smith, D.IL., S.L. Kunkel, T.J. Standiford, S.W. Chensue, M.W.
lLolfe, M.B. Orringer, lL.I. Whyte, M.D. Burdick, J.M. Dan-
forth, A.lL. Gilbert, and lL.M. Strieter. 1993. The production
of interleukin-1 receptor antagonist protein by human bron-
chogenic carcinoma. Am. J. Pathol. 143:794.
1415 Smith et al.