| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
into Established Brain Tumors Represses Growth by Antiangiogenesis
Departments of
*
Neurology and
Anesthesia and Pain Management, and
Center for Immunology, University of Texas Southwestern Medical Center, Dallas, TX 75235
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
into a mouse model of an established metastatic brain tumor.
Temperature-sensitive replication-defective adenovirus was generated
for gene transfer of IFN- (AdIFN) and ß-galactosidase (AdBGAL)
cDNAs in vivo. In this model, treatment with AdIFN elicits prolonged
survival times and brain tumor rejection. Evidence against an
immune-mediated response accounting for this result include: 1) absence
of a memory immune response upon challenge, 2) lack of antitumor
effects at sites distal to inoculation of AdIFN, and 3) preservation of
the therapeutic effects of AdIFN in scid and beige mice
and in inducible NO synthase (iNOS) knockouts. High concentrations of
IFN- do not inhibit tumor growth in vitro making it unlikely that
the antitumor effect of this treatment acts directly on the growth of
the tumor cells. However, gene transfer of IFN- inhibits
neovascularization of the tumor in a 3LL-Matrigel assay in vivo, and
AdIFN induces apoptosis of endothelial cells in vivo, supporting the
idea that AdIFN represses tumor growth by inhibiting angiogenesis. The
substantial non-immune-mediated therapeutic benefits of AdIFN in
animals paves the way for devising novel strategies for treating human
brain tumors. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Tumor growth in a live organism is dependent not only on its ability to evade immune surveillance but also on its capacity to maintain a continuous and stable supply of nutrients and oxygen for proliferation. Tumor angiogenesis, or the capacity to form new capillaries from preexisting microcapillaries, is a common characteristic and a limiting step in the growth of most solid tumors. The ability to induce neovascularization either 1) is produced by molecules secreted by the tumor itself or from a host cell recruited by the tumor, 2) may arise from molecules derived from the extracellular matrix, or 3) may be caused by loss of physiological inhibition of endothelial cell proliferation (3, 4, 5).
Data from this laboratory showed that priming intracerebrally with a
rat malignant glioma cell line (RT2), genetically modified to secrete
"high amounts" of IFN-, elicits prolongation of survival times
and tumor rejection (6). The animals acquired a memory
response characterized by systemic and CNS antitumor immunity,
suggesting that the antitumor effects in this rat model are at least in
part immune mediated. Unlike RT2, most human tumors are not immunogenic
(7). The following preclinical gene therapy experiments
were designed to examine the potential benefits of adenoviral-mediated
gene transfer of IFN- into a model of established metastatic brain
tumors in mice. A mouse Lewis lung carcinoma (3LL) was selected because
it is malignant and poorly immunogenic. Unexpectedly, we were led to
the findings that gene transfer of IFN- in vivo represses growth of
established 3LL brain tumors not because of an immune response, but
rather by inhibiting angiogenesis.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
C57BL/6 wild-type (wt),3 scid, and beige mice were purchased from The Jackson Laboratory (Bar Harbor, ME). 3LL (Lewis lung carcinoma, C57BL/6) was a generous gift from Dr. I. Fidler (M. D. Anderson Cancer Center, Houston, TX).
Generation of shuttle plasmids
Subcloning was conducted by standard recombinant DNA techniques.
The expression cassettes that drive the expression of IFN- and
lacZ were digested out of the plasmids pGIFN and pGBGAL
(6), and cloned into pXCJL.2 to generate pXIFN and pXBGAL,
respectively (Fig. 1
).
|
Shuttle plasmids and pBHGts were cotransfected via liposomes into 293 cells grown at 32°C. Successful recombination produces infectious virions that induce total lysis of 293 cells within 5–8 wk. Viral DNA isolated from 100–200 µl of supernatant from the wells with complete lysis is examined by PCR to determine whether virions contain the appropriate cDNA and to detect wt virus contamination using E1 region primers. The recombinant virus is plaque purified and grown on 293 cells at 32°C. Tissue culture supernatants containing adenovirus are concentrated by centrifugation over cesium chloride using the method of Graham and Prevec. Viral titer is determined by counting plaques using an agarose overlay method as described by Graham and Prevec (8). The titers of AdIFN and AdBGAL of viral particles/PFU ratios are 43 and 81, respectively.
Intracerebral tumor/viral implantation
C57BL/6 mice were anesthetized by i.p. xylazine (10 mg/kg) and ketamine (90 mg/kg) and then placed in a stereotactic frame. The bregma was identified, a burr hole was drilled and the dura was exposed. Using a Hamilton syringe, the dura was punctured, and a needle was inserted 3.5 mm (1 mm anterior and 2 mm to the right of the bregma) into the brain. After 3 min, 3 µl containing 1500 tumor cells mixed 50:50 with 30% methylcellulose, or 10 µl of viral particles suspended in 10 µl of PBS, were injected over 3 min. After another 3 min, the needle was withdrawn slowly. Where indicated, 2 days later, the mice were injected at the same coordinate with viral particles (AdIFN or AdBGAL) in a total volume of 10 µl.
ELISA
Supernatants of transduced cells or sera of animals treated by
AdIFN were assayed for murine IFN- (R&D, Minneapolis, MN) by ELISA
following the manufacturer’s specifications.
Tissue staining
Abs used are against mouse IFN- (American Type Culture
Collection, Manassas, VA), mouse CD31 (anti-platelet endothelial
cell adhesion molecule (PECAM)-1, PharMingen, San Diego, CA), human
factor VIII (Dako Corporation, Carpinteria, CA). For
immunohistochemistry, anti-IFN- and anti-CD31 were reacted
with biotin following the manufacturer’s specifications (Vector,
Burlingame, CA). Fresh frozen sections of 6–10 µm thickness were
reacted with biotinylated Abs, developed, and then counterstained in
1% methyl green as previously described (6). Reactions
with anti-IFN- were performed in the presence of saponin. For
immunofluorescence, tissue sections were sequentially reacted with
rabbit anti-factor VIII then Cyan3-coupled anti-rabbit IgG
(Jackson ImmunoResearch Laboratories, West Grove, PA). For the TUNEL
reaction, sections were processed using the Fluorescein-FragEL DNA
fragmentation detection kit following the manufacturer’s
specifications (Oncogene Research Product, Cambridge, MA). Processed
sections were viewed by epifluorescence microscopy (Axioplan 2; Carl
Zeiss, Thornwood, NY).
Matrigel assay for angiogenesis
Angiogenesis assays were performed as described elsewhere (9); briefly C57BL/6 mice were injected with 0.5 ml of Matrigel (Collaborative Biomedical Products, Bedford, MA) mixed with 1 x 106 3LL cells and AdIFN or AdBGAL (22 x 109 viral particles) in the abdominal midline (day 0 in all experiments). To quantify angiogenesis, Matrigel pellets were harvested on day 6, all surrounding tissue was dissected away, 0.3 ml Matrisperse was added (Collaborative Biomedical Products), and the gel was liquefied by incubation at 4°C overnight. Hemoglobin content of the liquefied pellets was determined by the Drabkin method (Sigma Diagnostic, St. Louis, MO) following the manufacturer’s specifications.
Statistical analysis
Statistical calculations were performed by the JMP software (SAS Institute, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene
into normal brain and tumor cells
Shuttle plasmids pXIFN and pXCLON were constructed to contain the
mouse IFN- and LacZ cDNAs cloned 3' to CMV promoters and
SV40 splicer donor sites and 5' to the adenovirus recognition sequence,
respectively (Fig. 1
A). The recombinant
temperature-sensitive replication-defective adenoviral vectors AdIFN
and AdBGAL were generated by cotransfecting pXIFN or pXCLON with pBHGts
into 293 cells respectively. Cultured 3LL cells transduced by AdBGAL or
AdIFN express LacZ or secrete IFN- in vitro, respectively
(Fig. 1B). AdIFN and AdBGAL transduce normal brain to
express IFN- and ß-galactosidase in vivo, respectively (Fig. 2). Furthermore, IFN- secreted in the
brain is detected in serum; 2 days after receiving 11.5 x
109 viral particles of AdIFN intracerebrally,
mice sera (n = 2) contain 1.44 ng/ml of IFN- as
measured by ELISA. LacZ expression in the brain is
persistent for longer than 6 wk. AdIFN-treated mice show no overt signs
of neurological dysfunction.
|
To test for therapeutic effects, animals implanted intracranially
with 3LL were injected 2 days later with AdIFN or AdBGAL (Fig. 3A). Unlike those treated by
AdBGAL, mice treated by AdIFN show statistically significant
prolongation of survival times (Log-Rank test, p =
0.0003). Furthermore, 3/8 (38%) of the mice survived longer than 85
days after injection and rejected the tumor because histological
analysis revealed cavity formation and no tumor mass. To examine
whether AdIFN generates similar systemic antitumor effects, groups of
mice were implanted s.c. with 1 x 106 3LL
cells mixed with either AdIFN or AdBGAL. Unlike with AdBGAL, 3LL cells
mixed with AdIFN fail to produce palpable tumors 2 wk after injection,
although tumor growth occurred after 3 wk (Fig. 3B).
|
To study whether AdIFN generates a memory antitumor response,
animals that had rejected intracerebrally implanted 3LL were
rechallenged by wt 3LL s.c. Unexpectedly, growth in naive mice was
similar to growth in mice that had previously rejected intracerebral
3LL cells (Fig. 4A). To
further examine for antitumor effects at a distance, mice implanted by
3LL cells in both hemispheres were treated by AdIFN or AdBGAL only on
the right side. None of these animals rejected intracerebral 3LL, and
the mean survival times associated with AdIFN and AdBGAL were 26 and 21
days, respectively (Fig. 4B). At autopsy, visual inspection
revealed that, whereas brains of all mice injected with AdBGAL carried
large bilateral tumor masses (8/8), 6/9 animals implanted with AdIFN
showed tumors only on the left side. Histological examination of the
latter showed large tumors on the left and cavity formation on the
right (Fig. 4, C and D). Thus the data support
the hypothesis that AdIFN elicits "local" antitumor effects.
|
A) and beige
mice (Log-Rank, p = 0.0125; Fig. 5B). In
addition, histological analysis shows that 4/9 (44%) scid
and 3/10 (33%) beige AdIFN-treated mice reject 3LL. IFN- may
enhance nonspecific phagocytic activity in microglia by inducing
expression of inducible NO synthase (iNOS). However, the antitumor
effects generated by AdIFN are independent of iNOS expression because
they are preserved in C57BL/6 iNOS knockouts and in wt mice treated by
NG-nitro-L-arginine
methyl ester (L-NAME) (data not shown). These
results strongly suggest that AdIFN generates its antitumor effects by
a mechanism other than an adaptive cellular immune response.
|
To examine whether IFN- suppresses tumor growth in vitro, 3LL
cells were grown in the presence of titrated concentrations of IFN-,
and cell numbers were counted at variable time intervals. The log
growth curves demonstrate that physiological (12.5 ng/ml; 1x) and high
concentrations (1.25 µg/ml; 100x) of IFN- only slightly retard
growth of 3LL in vitro and only after 7 days (Fig. 6A). Thus, the direct effects
of IFN- on tumor growth in vitro do not explain the in vivo data,
namely that AdIFN elicits rejection of 3LL brain tumors in 38% and
44% of C57BL/6 and scid mice, respectively (Figs. 3A and 5A).
|
B).
Furthermore, immunohistochemical analysis revealed that, whereas 3LL
cells mixed with AdBGAL grew as a continuous tumor penetrated by blood
vessels, those mixed with AdIFN grew in spheroid-like islands devoid of
endothelial cells (Fig. 6C). Therefore, it appears that
AdIFN inhibits tumor neovascularization in vivo.
Regional administration of "high" amounts of IFN- and TNF in
combination with melphalan, by the limb perfusion technique, results in
over an 80% complete response in patients with limb melanoma. In these
patients, "high local" concentrations of TNF and IFN- cause
selective disruption of the tumor vasculature by reducing the
activation of 
vß3
integrin, leading to apoptosis of angiogenic endothelial cells
(10). This information raised the question whether, in
this model, "high local" concentrations of IFN- (Fig. 2) induce
apoptosis of endothelial cells in vivo. Three days after treating 3LL
brain tumors with virus, unlike AdBGAL, H&E staining of AdIFN-injected
brains reveals intense perivascular infiltrates and fragmented nuclei
in tumor blood vessels (Fig. 7A). Double immunofluorescence
staining for fragmented DNA (TUNEL method) and factor VIII positive
endothelial cells reveal apoptosis in blood vessels in the tumor bed
(Fig. 7, B–D). Furthermore, 10 days after
receiving adenovirus, unlike brains injected with AdBGAL that harbor
large "healthy" well-vascularized tumors (Fig. 7, E and
G), tumors treated by AdIFN are devoid of blood vessels and
show massive apoptosis (Fig. 7, F and H).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
to produce an
immune reaction because IFN- up-regulates expression of MHC class I
molecules in 3LL (data not shown) and drives microglia, a putative APC
of the brain, into an immune-active state (11). Our
findings reveal that intracranial inoculation of AdIFN elicits
prolonged survival times and rejection of tumors established in the
brain for 2 days while s.c. inoculation directly at the tumor site
retards growth of 3LL implanted peripherally. Evidence against an
immune-mediated response accounting for this result include 1) absence
of a memory immune response upon challenge, 2) lack of antitumor
effects at sites distal to inoculation of AdIFN, and 3) preservation of
the therapeutic effects of AdIFN in scid and beige mice, and
in C57BL/6 iNOS knockouts and wt mice treated by
L-NAME. High concentrations of IFN- do not
inhibit tumor growth in vitro, making it unlikely that the antitumor
effect of this treatment acts directly on the growth of tumor cells.
However, gene transfer of IFN- inhibits neovascularization of the
tumor implanted with Matrigel in the avascular s.c. space in vivo;
furthermore, AdIFN induces apoptosis of endothelial cells in vivo,
supporting the idea that AdIFN represses tumor growth by inhibiting
angiogenesis. The effects of IFN- on angiogenesis have been
described in human disease as well as many in vivo and in vitro models.
IFN- inhibits proliferation and migration of human endothelial cells
in vitro (12, 13) and represses lymphocyte-induced tumor
angiogenesis in experimental animals (14, 15). IFN-
also elicits significant inhibition of tube formation in a
HUVEC/Matrigel in vivo model (16).
In our model, treatment with AdIFN generates rejection of 3LL brain
tumors in 38% and 44% of C57BL/6 and scid mice,
respectively (Figs. 3A and 5A). The question of
how an antiangiogenic molecule eradicates a tumor is intriguing. A
biological precedence was established in experiments where
tumor-bearing animals were treated by multiple injections of
endostatin. Whereas tumors typically regrow after a single cycle of
endostatin, they fail to regrow, showing prolonged dormancy after
repeated cycles (17). It is possible that induction of
continuous "high" local concentrations of IFN- for weeks is
equivalent to multiple cycles of antiangiogenic therapy. In this model,
the end result is tumor cell apoptosis, probably caused by ischemia
(Fig. 7H). The effects of IFN- on endothelial cells may
be mediated either by direct effects or through secondary molecules
such as IFN-inducible protein 10 (IP-10) or monokine induced by IFN-
(MIG). These chemokines, which react with the CXCR3 receptor, inhibit
endothelial cell proliferation in vitro and repress tumor
vascularization in vivo (9, 18).
The therapeutic implications are important; because this model shows
that, regardless of tumor immunogenicity, the strategy of inducing high
local concentration of IFN- is successful in generating a
substantial anti-brain tumor response mediated by antiangiogenesis.
The results also suggest that, in the rat glioma model where animals
were implanted by RT2 genetically modified in vitro to secrete IFN-,
the observed antitumor response may be caused primarily by
antiangiogenesis and that the acquired immunity reaction against wt RT2
may be secondary (6). The data constitute a proof of
principle in animals that sets the stage for devising novel strategies
that may lead to clinical trials.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Hassan Fathallah-Shaykh, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, TX 75235-9036. E-mail address:
3 Abbreviations used in this paper: wt, wild type; iNOS, inducible NO synthase; L-NAME, NG-nitro-L-arginine methyl ester; H&E, hematoxylin and eosin; o.m., original magnification; AdIFN, adenovirus IFN; AdBGAL, adenovirus ß-galactosidase..
Received for publication August 13, 1999. Accepted for publication October 14, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
affect tumorigenicity and response to IL-12 therapy and antiangiogenesis. Immunity 9:25.[Medline]
vß3 in the disruption of the tumor vasculature induced by TNF and IFN-. Nat. Med. 4:408.[Medline]
-interferon. J. Cell Biol. 104:689. followed by a synthetic lipid-A subunit analogue (GLA-60). Int. J. Cancer. 51:641.[Medline]
and on in vitro model of angiogenesis. J. Cell. Physiol. 146:164.[Medline]
This article has been cited by other articles:
|
|
 |
|
  Y. Ikeda, T. Kojima, S. Kuroda, Y. Endo, R. Sakai, M. Hioki, H. Kishimoto, F. Uno, S. Kagawa, Y. Watanabe, et al. A Novel Antiangiogenic Effect for Telomerase-Specific Virotherapy through Host Immune System J. Immunol., February 1, 2009; 182(3): 1763 - 1769. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Li, F. Pradera, T. Kammertoens, B. Li, S. Liu, and Z. Qin Cross-Talk between T Cells and Innate Immune Cells Is Crucial for IFN-{gamma}-Dependent Tumor Rejection J. Immunol., August 1, 2007; 179(3): 1568 - 1576. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Mitra-Kaushik, J. Harding, J. Hess, R. Schreiber, and L. Ratner Enhanced tumorigenesis in HTLV-1 Tax-transgenic mice deficient in interferon-gamma Blood, November 15, 2004; 104(10): 3305 - 3311. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. I. Kim, K. S. Kim, H. S. Kim, D. S. Kim, Y. Jang, K. H. Chung, and Y. S. Park Inhibitory Effect of the Salmosin Gene Transferred by Cationic Liposomes on the Progression of B16BL6 Tumors Cancer Res., October 1, 2003; 63(19): 6458 - 6462. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Chhokar and A. L. Tucker Angiogenesis: Basic Mechanisms and Clinical Applications Seminars in Cardiothoracic and Vascular Anesthesia, September 1, 2003; 7(3): 253 - 280. [Abstract] [PDF]
|
|
|
|
|
|
R. Ramesh, A. M. Mhashilkar, F. Tanaka, Y. Saito, C. D. Branch, K. Sieger, J. B. Mumm, A. L. Stewart, A. Boquio, L. Dumoutier, et al. Melanoma Differentiation-associated Gene 7/Interleukin (IL)-24 Is a Novel Ligand That Regulates Angiogenesis via the IL-22 Receptor Cancer Res., August 15, 2003; 63(16): 5105 - 5113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Ikeda, Y. Nakajima, T. Tokuhara, N. Hattori, M. Sho, H. Kanehiro, and M. Miyake Clinical Significance of Aminopeptidase N/CD13 Expression in Human Pancreatic Carcinoma Clin. Cancer Res., April 1, 2003; 9(4): 1503 - 1508. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Gattacceca, Y. Pilatte, C. Billard, I. Monnet, S. Moritz, J. Le Carrou, M. Eloit, and M.-C. Jaurand Ad-IFN{gamma} Induces Antiproliferative and Antitumoral Responses in Malignant Mesothelioma Clin. Cancer Res., October 1, 2002; 8(10): 3298 - 3304. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Hayakawa, K. Takeda, H. Yagita, M. J. Smyth, L. Van Kaer, K. Okumura, and I. Saiki IFN-gamma -mediated inhibition of tumor angiogenesis by natural killer T-cell ligand, alpha -galactosylceramide Blood, August 13, 2002; 100(5): 1728 - 1733. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Carroll, L. L. Nielsen, S. B. Pruett, and J. M. Mathis The Role of Natural Killer Cells in Adenovirus-mediated p53 Gene Therapy Mol. Cancer Ther., November 1, 2001; 1(1): 49 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Ma, H. Qin, and E. N. Benveniste Transcriptional Suppression of Matrix Metalloproteinase-9 Gene Expression by IFN-{gamma} and IFN-{beta}: Critical Role of STAT-1{alpha} J. Immunol., November 1, 2001; 167(9): 5150 - 5159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Salcedo, H. A. Young, M. L. Ponce, J. M. Ward, H. K. Kleinman, W. J. Murphy, and J. J. Oppenheim Eotaxin (CCL11) Induces In Vivo Angiogenic Responses by Human CCR3+ Endothelial Cells J. Immunol., June 15, 2001; 166(12): 7571 - 7578. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Carta, S. Pastorino, G. Melillo, M. C. Bosco, S. Massazza, and L. Varesio Engineering of Macrophages to Produce IFN-{{gamma}} in Response to Hypoxia J. Immunol., May 1, 2001; 166(9): 5374 - 5380. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
