Profiles of Prostaglandin Biosynthesis in Normal Lung and Tumor Tissue from

(CANCER RESEARCH 48. 3140-3147. June I. 1988]
Profiles of Prostaglandin
Lung Cancer Patients
Biosynthesis in Normal Lung and Tumor Tissue from
Theodore L. McLemore,1 Walter C. Hubbard, Charles L. Litterst, Mark C. Liu, Stephan Miller,
Noreen A. McMahon, Joseph C. Eggleston, and Michael R. Boyd
Program Research and Development Group, Office of the Associate Division Director for Developmental Therapeutics, Division of Cancer Treatment, National Cancer
Institute—Frederick Cancer Research Facility, Frederick, Maryland 21701 ¡T.L. M., W. C. H., C. L. L., M. K. B.J, and Departments of Medicine (Pulmonary Division)
and Surgical Pathology, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 [M. C. L., S. M.,N. A. M.,J. C. E.J
ABSTRACT
Prostaglandin (PC) biosynthetic profiles from endogenous arachidonic
acid were determined by capillary gas chromatography-mass spectrometry in matched fresh normal lung (NL) and lung cancer (LC) tissue
fragments obtained from 42 individual LC patients at the time of diag
nostic thoracotomy. The histológica!diagnoses represented were squamous cell carcinoma (/V = 20), adenocarcinoma (/V = 7), small cell
carcinoma ( V = 4), mixed cell carcinoma (.V = 2), bronchioloalveolar
cell carcinoma (.V = 2), large cell undifferentiated carcinoma (/V = 3),
bronchial carcinoid ( V = 1), and metastatic tumors (A/ = 3). When PG
biosynthesis was determined in NL tissue separately, low mean levels of
PGEj and PGFi, (<2 pmol/mg protein/15 min), intermediate levels of
PGD2 and 6-keto-PGF,,, (6KPGF,.) (2-7 pmol/mg protein/15 min), and
high levels of thromboxane B2(TXB2) (>7 pmol/mg protein/15 min) were
observed. There was no particular correlation with cigarette smoking
history and PG biosynthesis in NL. When PG production in LC tissue
was evaluated separately, high levels of !'(.!'. I'M-.., and 6K1'(,I ,., as
well as I \H.. and low levels of PGD2 were noted. In addition, LC tissue
from cigarette smokers demonstrated elevated levels of PGE2, 6K I'M-',.,,
and IAlt., when compared to current nonsmokers with LC (/' < 0.05 in
all instances). Simultaneous comparison of PG production in matched
LC and NL tissue from individual patients indicated increased biosyn
thesis of 1'Mv. and I'M-'..., and low levels of PGD2 in LC compared to
NL tissue (/' < 0.05 in all instances; paired, two-tailed. Student's t test).
Individual comparison of PG biosynthesis according to LC histological
cell type revealed that I'M , and PGF&, were consistently elevated in all
four common primary LC histological cell types, the only exception being
large cell undifferentiated carcinoma. Interestingly, this latter LC histo
logical cell type presented a unique profile with lower levels of I'M- . and
I'M)., in LC than in NL tissue (/' < 0.05 in both instances). In addition,
the biosynthesis of all 5 PGs studied was consistently higher in primary
than metastatic adenocarcinomas of the lung (/' < 0.05 in all instances).
No differences were observed in NL and LC tissue for the major LC
histological cell types when I'M »...
IAB , or hkl'M ,. biosyntheses were
compared. These findings indicate that the profiles of PG biosynthesis
in LC and NL tissue from individual patients may differ substantially.
These differences may reflect, in part, contributions to the PG biosyn
thetic profile unique to malignant cells.
INTRODUCTION
Pulmonary carcinomas are a major cause of adult morbidity
and mortality in the United States population (1-5). A more
detailed understanding of the biochemistry of human LC2 could
contribute to the further development and refinement of better
approaches to earlier, more specific diagnoses as well as poten
tially lead to improvements in the therapeutic intervention and
treatment of this disease. Studies characterizing human lung
carcinomas according to selected neuroendocrine and cellular
properties have provided valuable initial information regarding
certain biochemical features of pulmonary tumors, particularly
small cell carcinomas (see Refs. 6-9 for review). The PGs and
related eicosanoids are important mediators of NL function
(10-12) and might also be important in the pathophysiology of
certain human LC (13-17). Since the lung is a tissue rich in
enzymes that biosynthesize PGs and TX as well as inactivate
these compounds (18-23) the biosynthetic capability of lung
tumor tissue to synthesize this family of compounds might be
useful in the classification of LC and provide a more complete
understanding of the role of PG and TX in the pathophysiology
of this disease (13-17, 24-49).
There is little information currently available regarding the
profiles of 20-carbon fatty acid cyclooxygenase products in
either NL or LC tissue. In the present studies, the profiles of
five such products synthesized from endogenous arachidonic
acid in matched NL and LC tissue from individual patients
representing several histological classes of human LC were
simultaneously compared.
MATERIALS
AND METHODS
Study Subjects. LC and NL tissue were obtained at the time of
diagnostic thoracotomy at the Johns Hopkins University School of
Medicine from 42 LC patients (23 males and 19 females). A summary
of the pertinent clinical information for these patients is provided in
Table I. Ninety-three % (39 of 42) of the subjects had presented initially
with primary lung tumors. The histological classification of both pri
mary and metastatic LC is provided in Table 2.
Collection, Preparation, and Incubation of Tissue Fragments. Proce
dures for tissue collection, preparation for experimentation, and incu
bation for PG from endogenous precursor in vitro are identical to those
described previously (50). Triplicate specimens of biopsy fragments of
NL and LC tissues were used for determinations of intersample varia
tions of PG biosynthesis in the tissues studied.
PG Analysis. The profiles of PGF2„,
PGD2, PGE2, TXB2, 6KPGF,,,,
and the 15-keto-13,14-dihydrometabolites of PGE2, PGF2„,
6KPGF,,,
synthesized from endogenous arachidonic acid were performed via
capillary gas chromatography-mass spectrometry as described previ
ously (50). This method of analysis offers a very high degree of selec
tivity, sensitivity (detection limits -0.1 pg of individual analyte per
injection), and reproducibility.
RESULTS
Reproducibility of PG Measurements
Fragments
in Human Lung Biopsy
Capillary gas chromatography-mass spectrometry measure
ments of PG synthesized in biopsy fragments of NL tissue and
lung carcinoma tissue have been demonstrated previously to be
Received 11/18/87; revised 2/26/88; accepted 3/4/88.
highly reproducible with coefficients of intrasample variations
The costs of publication of this article were defrayed in part by the payment
ranging
from 2.2 to 18.5% (50). The cellular heterogeneity of
of page charges. This article must therefore be hereby marked advertisement in
NL tissue and LC tissue may contribute significantly to varia
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1To whom requests for reprints should be addressed, at NCI-FCRF, Bldg.
tions in the quantitative and qualitative profiles of PG synthe
428, Rm. 63, Frederick, MD 21701.
sized in these tissues. Examples of the variability of PG pro
:The abbreviations used are: LC, lung cancer; NL. normal lung; PC, prostaglandin; TX, thromboxane; 6KPGF,„,6-keto-PGF,„.
duction in triplicate samples of biopsy fragments of tissues (NL
3140
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
and LC) from three patients are summarized in Table 3. The
mean coefficients of variation of PC measurements in biopsy
fragments of NL and LC tissues obtained in triplicate from
individual patients ranged from 7 to 32% for the 5 PG evalu
ated. The range of the intersample coefficients of variation in
the measurements are approximately twice those reported ear
lier for intrasample replicate analysis (50). These findings in
dicate variations in cellular heterogeneity of NL as well as LC
biopsy fragments used in these studies are not extreme and are
within the expected range for human subjects.
PG Biosynthesis in NL Tissue
Appreciable mean levels of all 5 PG metabolites were present
in NL tissue from both smokers and current nonsmokers. Low
mean levels (<2 pmol/mg protein/15 min) were noted for PGE2
and PGF2„production in NL; intermediate levels (2-7 pmol/
mg protein/15 min) were noted for PGD2 and 6KPGF,„bio
synthesis; and highest levels (>7 pmol/mg protein/15 min)
were observed for TXB2. No significant differences were noted
for levels of production of the PGs in NL tissue obtained from
smokers and current nonsmokers or when PG biosynthesis was
compared in NL tissue obtained from male and female patients
(P > 0.30 in all instances; nonpaired, two-tailed Student's t
test).
The profiles of production of 5 different eicosanoids from
endogenous arachidonic acid in NL tissue fragments from 27
smokers and 15 current nonsmokers are shown in Fig. \A.
Table 1 Lung cancer patient population
No.
42
23(60± 11)
19(59± 10)
27 (58 ±30)
15 (45 ±33)
Total
Male (mean age ±SD)
Female (mean age ±SD)
Smokers (mean pack/years ±SD)
Current nonsmokers (mean former
pack/years ±SD)°
" Current nonsmokers are defined as those patients who have not actively
smoked tobacco products for >6 months or who have never smoked. Only 1
patient with primary lung cancer and 1 patient with a metastatic lung tumor had
no previous active smoking history. Mean time since cessation of smoking for
former smokers was 4.5 years (range, 0.5-20 years).
Table 2 Histological classification of human lung tumors
type"Primary*Squamous
Cell
PG Biosynthesis in Human LC Tissue
The profiles of endogenous PG biosynthesis in unstimulated
LC tissue from 27 smokers and 15 current nonsmokers are
shown in Fig. IB. High PG biosynthesis (>7 pmol/mg protein/
15 min) was demonstrated for PGE2, PGF2„,and 6KPGF,,, as
well as TXB2 in LC tissue. Conversely low mean levels of PGD2
(<2 pmol/mg protein/15 min) were observed consistently in
lung tumor specimens. In addition, higher PGE2, 6KPGF,„,
and TXB2 mean levels were apparent in LC tissue obtained
from cigarette smokers compared with those from current
nonsmokers (P < 0.05 in all instances) and similar PGD2 and
PGF2„
levels were observed for LC tissue regardless of smoking
status (P > 0.10 in all instances). Additionally, no significant
differences in PG biosynthesis were observed in LC tissue
obtained from male and female patients (P > 0.30).
Comparison of PG Biosynthesis in NL and LC Tissue
(47)'7(14)4(12)3(7)2(5)2(5)1(3)39
The
cellAdenocarcinomaSmall
profiles of PG biosynthesis from endogenous arachidonic
acid in matched NL and LC tissue from individual patients are
shown in Fig. 1. Elevated levels of PGE2 and PGF2„
production
were demonstrated in LC tissue compared to NL specimens
irrespective of the patient's smoking status (P < 0.05 in all
instances; paired, two-tailed Student's t test). Higher levels of
cellLarge
cellMixed
tumorsBronchioloalveolar
cellBronchial
carcinoidSubtotalMetastatic''Adenocarcinoma
(93)2(5)1(3)3(7)42
(prostate)Transitional
bladder)SubtotalTotalNo.20
carcinoma (urinary
(100)
" All diagnoses were histologically confirmed.
* Primary pulmonary tumors.
' Numbers in parentheses, percentage.
d Tumors metastatic to the lung from a nonpulmonary source.
6KPGFi„were synthesized in LC tissues from individual smok
ers compared to those in NL tissue (P < 0.05). TXB2 levels
were similar in lung tumor and normal pulmonary tissue from
smokers (P > 0.05 in all instances), but lower TXB2 production
was observed in LC than in NL tissue from current nonsmokers
(P < 0.05). In contrast, levels of PGD2 production were higher
in NL tissue fragments than in LC tissue regardless of the
cigarette smoking status (P < 0.05 in all instances).
Table 3 Variability in prostanoid biosynthesis in different biopsy fragments of normal lung and lung carcinoma tissue from three patients
fragments*PGE2ND*001.90.764016.64.93023PGDjND000.590.12205.51.62816PGFfc,ND001.60.3193.80.651711TXB22.9'0.85290.40.17438.71.416296KPG
tumor
fragments*PCD;4.71.5314.50.8191.40.21522PGFj,0.960.27281.90.75391.0
lung
Patient"ABCStatistical
analysisMean'SDcv7MeanSDCVMeanSDCVMean
D003.41.13215
CVLung
" Patients diagnosed with adenocarcinoma (A) and squamous cell carcinomas (B and C).
* Endogenous prostanoid biosynthesis was performed on 0.5-1.0-mm3 normal lung and lung tumor tissue fragments as described in "Materials and Methods.'
' n = 3.
d ND, not detectable; for purposes of statistical analyses, ND values were assumed to be 0.
' pmol/mg protein/15 min.
'CV, coefficient of variation
x 100.
mean
3141
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID
20
A.
Normal
BIOSYNTHESIS
IN HUMAN LUNG CANCERS
mous cell carcinomas or adenocarcinomas of the lung. Primary
lung carcinomas demonstrated significantly higher PGE2 pro
duction than did metastatic tumors (P < 0.05 in all instances;
nonpaired, two-tailed Student's t test). PGE2 production was
Lung
I Smoker
I Nonsmoker
consistently higher in tumor tissue for all histological cell types
of primary lung carcinomas studied (P < 0.05 in all instances;
paired, two-tailed Student's t test) with the exception of large
15
10
I
B.
Lung Tumor
15
10
ÃŽ T
Ili
POE,
PGD,
PCF,
6KPGF,
L
TxB,
Fig. I. Profiles of endogenous PC biosynthesis in NL and LC tissue from
current nonsmokers and cigarette smokers with LC. A, comparison of PGE2,
PGD2, PGF^, 6KPGF,., and TXB2 endogenous biosynthesis in NL tissue from
smokers or current nonsmokers. No differences were noted among levels of the
S different PCs in NL tissue from either group (nonpaired, two-tailed Student's
t test; P> 0.30 in all instances). B, comparison of endogenous biosynthesis of the
S different PCs in LC tissue from smokers and current nonsmokers. Higher levels
of products of PGE2,6KPGFi„ and TXB2 were noted in LC tissue obtained from
smokers than nonsmokers. Comparison of A and B revealed increased production
of PGE2, PGFj., and 6KPGF,. in LC tissue from smokers than in NL tissue (P
< 0.05 in all instances; paired, two-tailed Student's / test). PGE2 and PGF^ were
increased in tumor tissue from both smokers and current nonsmoker tissues;
decreased levels of PGD2 were noted in LC than in NL tissue for individual
patients regardless of smoking status; and decreased TXB2 levels were observed
in LC compared with NL from current nonsmokers (P < 0.05 in all instances).
Columns, mean; bars, SE. All data were derived from 15-min incubations. Current
nonsmokers are defined as those patients who have not smoked tobacco products
for »6months or who have never smoked.
Relationship between LC Histological Cell Type and PG Biosyn
thesis
In order to further assess eicosanoid production from endog
enous arachidonic acid precursor in unstimulated NL and LC
tissue, patients were categorized according to tumor cell histo
lógica!cell type and were then separately evaluated for produc
tion of each of the 5 PCs (Figs. 2-6).
PGE2. Profiles of endogenous PGE2 production in unstimu
lated NL tissue and LC tissue from individual LC patients
according to histologie-aïcell type are shown in Fig. 2. The
highest PGE2 production was observed in patients with squa-
cell undifferentiated carcinomas, where higher levels of PGE2
production were observed in NL tissue (P < 0.05) (Fig. 2).
PGF^. Profiles of endogenous PGF2„production in unstim
ulated NL tissue and LC tissue from individual LC patients
according to histological cell type are shown in Fig. 3. Small
cell carcinomas of the lung, large cell undifferentiated carcino
mas of the lung, and metastatic tumors demonstrated similar
levels of PGF2„production in NL and LC tissues (P > 0.05 in
all instances). PGF2n production was higher in all other LCs
studied than in the corresponding NL tissue from the individual
patients [P > 0.05 in all instances (Fig. 3)]. In addition, the
highest PGF2o production occurred in lung adenocarcinoma
tissue.
PGD2. Profiles of endogenous PGD2 biosynthesis in unstim
ulated NL and LC tissue from individual patients according to
histological cell type are demonstrated in Fig. 4. Similar levels
of PGD2 production occurred in NL and LC tissue (P > 0.10
in all instances) with the exception of bronchioloalveolar cell
carcinomas and large cell undifferentiated carcinomas where
greater PGD2 production occurred in NL than in LC tissue (P
< 0.05). In addition, elevated PGD2 production was evident in
one bronchial carcinoid tumor in comparison with NL tissue
(P< 0.05) (Fig. 4).
TXB2 and 6KPGF... Higher levels of TXB2 and lower levels
of 6KPGF),, biosynthesis were observed for a bronchial carci
noid tumor removed from a patient compared with levels in
NL tissue (Figs. 5 and 6; P < 0.05 in both instances). No other
identifiable differences were noted between NL and LC tissue
for TXB2 and 6KPGFlo production when these were evaluated
in individual patients according to LC histological cell type (P
> 0.10 in all cases).
When primary and metastatic lung tumors were compared,
higher levels of all 5 PGs were consistently observed in primary
compared to metastatic adenocarcinomas of the lung (Figs. 26; P < 0.05 in all instances).
DISCUSSION
Many biochemical pathways may be altered in human malig
nant disease (for review, see Refs. 51 and 52). Previous studies
have suggested that there may be characteristic perturbations
in PG biosynthesis in certain human lung carcinomas in vivo
(13) and in lung carcinoma tissues in vitro (14). More recently,
it has been demonstrated that characteristic profiles of PG
biosynthesis occur in certain established human lung carcinoma
cell lines (15-17). Since PGs may be important as mediators of
normal pulmonary function as well as in the modulation of
certain components of the immune response (38-43), tumor
promotion (27-37), cellular proliferation (29-37), and tumor
cell metastasis (24-26,49), comparisons of the PG biosynthetic
profiles in NL and LC tissues may provide additional insight
into the role of this family of compounds in the pathophysiology
of human LC. A major limitation in the assay methods for
assessment of PG biosynthesis in earlier studies of LC patients
(13) and in LC tissue in vitro (14) has been the requirement for
large quantities of biological material. The recent development
of highly sensitive mass spectrometric assays (50) permits rapid
3142
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
Squamous
Cell
N = 20
Adenocarcinoma
N= 7
'.
Small
Cell
i Bronchioloalveolar Cell
N= 2
N= 4
Metastatic
Tumors
N = 3
Fig. 2. Profiles of endogenous ['(.I',, production in matched unstimulated NI, and carcinoma (CA) tissue according to tumor histological cell type. }'(,!.. levels
were higher in all primary lung tumor histological cell types (/' < 0.05 in all instances; paired, two-tailed Student's t test) with the exception of large cell undifferentiated
carcinomas where PGEi levels were higher in NL than in LC tissue (P < 0.05). No differences were noted in PGE2 production between metastatic tumors and NL
tissue from individual patients (P > 0.03). Columns, mean; Aurv.SE. All data were derived from 1S-min incubations.
Squamous
Cell
N = 20-
Adenocarcinoma
N=7
Small
Cell
N= 4
Mixed
Cell
N=2
! Bronchiolo| alveolar Cell
;
N= 2
Bronchial
Carcinoid
N= 1
Large
Cell
N= 3
Metastatic
Tumors
N= 3
Fig. 3. Profiles of endogenous l'<il ....production in matched unstimulated NL and carcinoma (CA) t ssue according to histological cell type. Higher I'( .1 ... levels
were noted in LC than NL tissue for squamous cell carcinomas, adenocarcinomas, mixed cell tumors, and a bronchial carcinoid (P < 0.05 in all instances). Columns,
mean; bars, SE. All data were derived from 15-min incubations.
determinations of the biosynthetic profiles of PG in small (0.5synthesis in NL and LC tissue fragments. Moderate variations
1.0 mm3) biopsy fragments of NL and LC tissues. The repro[7-32% (Table 3)] in PG production were observed in separate
ducibility of the assays (coefficients of intrasample variation ^
NL and LC tissue fragments from individual patients or ap15%) has also facilitated the current investigations of PG bio- proximately twice the range in intrasample variation observed
3143
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
Squamous
Cell
N = 20
Adeno-
Small
Cell
N= 4
carcinoma
N= 7
Bronchiolealveolar Cell
N= 2
Mixed
Cell
N= 2
Bronchial
Carcinoid
N= 1
Large
Cell
N= 3
i Metastatic
'
Tumors
N= 3
Fig. 4. Profiles of endogenous PGD¡ production in matched unstimulated NL and carcinoma (CA) tissue according to histological cell type. Higher PGD;
production was noted in NL than in LC tissue Tor bronchioloalveolar cell carcinomas and large cell undifTeremiated carcinomas (/' < 0.05 in both instances) and in
LC than NL tissue for one bronchial carcinoid (P < 0.05). All other comparisons were not significantly
from 15-min incubations.
different. Columns, mean: bars, SE. All data were derived
777A CA
17.5
15.0
Fig. 5. Profiles of endogenous 6KPGF,.,
production in matched unstimulated NL and
LC tissues according to histological cell type.
6KPGFi„ levels were higher in NL from a
patient than in a bronchial carcinoid tumor i '
< 0.05). Otherwise, no differences were noted
between carcinoma (CA) and NL tissue for the
various histological cell types (/' > 0.05 in all
instances). Columns, mean; bars, SE. All data
were derived from 15-min incubations.
I
12.5
£
•
~;
10.0
uT
C3
Q.
O
a.
2.5
Metastatic
Tumors
N= 3
for the current assay methods (50).
The relative contribution of tumor cells to the biosynthetic
profiles of PCs in LC tissue is difficult to assess. An experi
mental design with tissue fragments cannot definitely distin
guish the relative contribution of tumor cells to the total PG
profiles from that of surrounding normal immune, stromal,
hematopoietic, and/or parenchyma! cells that may be present
within tumor tissue fragments. The observed differences in the
biosynthetic profiles of PG in matched NL and LC tissue from
individual LC patients do, however, strongly suggest that these
discrepancies may be related in part to selective PG biosynthesis
in lung tumor cells. Recent investigations of PG biosynthesis
in established cell lines derived from human lung carcinomas
(15-17) which retain in vivo tumorigenicity in immunodeficicnt
animals (53) have clearly demonstrated enhanced PGE2, PGF2„,
and/or I \H biosynthetic capabilities which further support
the postulate that human tumor cells may contribute signifi
cantly to the profiles of PG biosynthesis in LC tissue.
In addition to the contribution of different cell populations
to the PG biosynthetic profiles, the heterogeneity of the LC
cells and the relative contribution of PG production in various
LC subtypes must also be considered. Tumor cell heterogeneity
3144
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
17.5r
15.0
12.5
Fig. 6. Profiles of endogenous I \H.. bio
synthesis in matched NL and LC tissue ac
cording to histológica!cell type. T\B„.levels
were higher in a bronchial carcinoid tumor
than in NL tissue from the same patient (P <
0.05). No other differences were noted between
NL and carcinoma (CA) tissue from individual
patients (P > 0.10 in all instances). Columns,
mean: bars, SE. All data were derived from 15min incubations.
I
I
t 10.0
7.5
5.0
2.5
Squamous
Call
N = 20
is specifically addressed when the PG profiles are evaluated
files in NL and LC tissue from individual LC patients. These
according to LC histológica! classification. Enhanced I'GF
results corroborate and extend previous investigations which
and PGF2.. production was a characteristic feature of the ma
have reported elevated PGE2 production in vivo in certain
jority of LC histological cell types in the present study with the human lung tumors, predominantly in primary squamous cell
highest levels of these two PCs being isolated from adenocarcarcinomas of the lung (13-15). The present studies confirm
cinomas and squamous cell carcinomas of the lung. It is well that squamous cell carcinomas may have enhanced potential
established that considerable variation in histological type is for the production of PGE2 but also demonstrate that all human
often observed in individual lung tumor specimens (6, 7, 54) LC histological cell types (with the exception of large cell
and selected studies have demonstrated that all 4 major LC cell undifferentiated carcinoma of the lung) have enhanced PGE2
types are present within individual lung tumors (6, 7). Varia
production when compared with NL tissue from individual LC
tions in PGEi and PGF2n production among different LC patients. Also of interest is the observation that tumors mehistological classifications might, therefore, be the result of the tastatic to the lung demonstrate levels of PGE2 production
relative distribution of specific tumor cell subtypes (i.e., ade- similar to that of NL tissue, suggesting an organ specific
nocarcinoma and squamous cell carcinoma) within a given relationship between PGE2 production and tumor origin. Be
tumor. In contrast, the relative contribution of the large cell cause PGE2 has been postulated to be involved in certain aspects
of immunoregulation (38-43), tumor promotion (27-37), pul
undifferentiated tumor cell population could have the opposite
effect on the levels of PGE2 and PGF2„biosynthesis since the monary blood flow (10-12), and determination of tumor me
tastatic potential (24-26, 49), elevated PGE2 levels might pro
formation of these compounds in this tumor type is either
vide a selective advantage for tumor cell survival in the microensimilar to that observed in NL tissue or decreased. The potential
importance of LC histological cell types to the relative contri
vironment of the lung. Further in vivo studies, with the use of
bution of the biosynthesis of different PG metabolites appears
the recently developed orthotopic animal model for the propa
even more relevant in view of recently published studies in gation of human lung tumors (53), may provide a suitable
established human lung tumor cell lines. These studies have animal model for further investigations of the role of PGE2 in
demonstrated considerable variation in PG biosynthesis in dif
LC pathobiology.
While PGE2 production has been investigated previously in
ferent cell lines representing different lung tumor histological
cell types (15-17). Thus, there is a substantial independent
human LC, the biosynthesis of other cyclooxygenase products
has not, heretofore, been studied in this disease. The present
body of evidence suggesting that certain tumor cell types present
in human lung tumors may contribute directly to the currently
investigations offer the first documentation of PGF2„,PGD2,
6KPGFi„,and TXB2 production in individual human LC tissue
observed profiles of PG biosynthesis in fresh LC tissue.
The current data provide documentation of the profiles of 5 fragments. The present report also provides the first direct
comparison of production of these 5 PGs in NL and LC tissue
PGs synthesized endogenously in individual NL tissue frag
ments via the fatty acid cyclooxygenase pathways. High levels from individual LC patients. PGF2„production was greatly
of PGD2, 6KPGF,.,, and TXB2 and low levels of PGE2 and enhanced in pulmonary carcinoma tissue, but the mechanisms
PGF2„
production were shown in the normal pulmonary tissue
for enhanced PGF2„biosynthesis in tumor tissue are not cur
fragments studied. These data confirm and extend previous
rently known. When PGF2a production was analyzed according
to histological cell type, squamous cell carcinomas, adenocarreports which have measured the biosynthesis of individual PG
in a given tissue specimen and have demonstrated low but cinomas, bronchioloalveolar carcinomas, mixed cell carcino
detectable PGE2 levels (10-12) as well as production of PGD2 mas, and a bronchial carcinoid demonstrated enhanced PGF2„
production. Increased PGF2„
biosynthesis may result from sterand TXB2 (10-12, 50) in NL tissue.
eospecific reduction of PGE2 or could reflect other alterations
The present studies also clearly demonstrate that the profiles
of PG biosynthesis are altered in many human lung tumors and in PG biosynthetic pathways in lung tumor tissue. The capabil
provide the first matched comparison of PG biosynthetic pro
ity of LC to synthesize PGF2o may be an important aspect of
3145
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
tumor invasiveness and metastatic potential. I'GF >„
has recently
been shown to enhance type IV collagenase production in tumor
cells in vitro, which, in turn, increases their ability to invade
through basement membrane matrices (55).
The present studies also demonstrate that profiles of PGD2,
6KPGF,„,and TXB2 biosynthesis are altered in certain LC
tissues from selected patients. The biochemical mechanisms for
alteration of these specific PGs in selected human lung tumors
are not presently known, but in the case of 6KPGFi„,or TXB2
production, these could be related to the effects of cigarette
smoking on tumor cell metabolic pathways. The elevation in
TXB, biosynthesis could also result directly from enhanced
TXB: production in tumor cells. Alternatively, enhanced plate
let PG biosynthesis and/or increased platelet concentrations in
LC tissues from smokers could partially account for the ob
served increase in TXB2 levels. Recent reports by Hubbard et
al. (16, 17) have demonstrated TXB2 production in vitro in
certain cell lines derived from human lung carcinomas support
ing the concept that TX biosynthesis may occur in selected
human lung tumor cells and that TXB2 biosynthesis in LC cells
could be important in the metastatic process.
The effect of active cigarette smoking on biosynthesis of PG
is not currently understood. It is interesting that detectable
differences in PG biosynthesis in LC tissue were observed
between active smokers and patients who were not smoking at
the time of study, even though the former pack/year history of
smoking for the current nonsmokers was similar to that of the
actively smoking patient group (Table 1). Certain metabolic
pathways are enhanced by cigarette smoking (see Refs. 56 and
57 for review) and PG biosynthesis might also be enhanced as
reflected by increased production of PGE2,6KPGFi„,and TXB2
in LC tissue from smokers. It is well known that cigarette
smoking induces specific cytochrome P-450 enzymes in human
lung tissues (see Ref. 56 for a review) and previous studies have
demonstrated alterations in certain cytochrome P-450 associ
ated enzyme activities in patients with LC (58-62). Enhanced
PG metabolism (via induction of cytochrome P-450 systems or
via induction or repression of other enzymes involved in bio
synthesis of these compounds) might, therefore, explain the
elevated levels of selected PGs observed in LC tissue from
smokers. Additional investigations, using relatively homoge
neous human lung tumor cell lines, will, however, be required
to delineate the specific enzyme(s) involved in this enhanced
PG biosynthesis in human LC tissues. The effects of cigarette
smoking on PG biosynthesis in LC might also have a significant
secondary impact on tumor growth and metastatic potential,
especially when PGE2 and TXB.. levels are affected (see previous
discussion).
It is apparent from the present studies that the use of an
individual LC patient's tumor and NL tissue provides a novel
method for biochemically profiling PG biosynthesis in small
tissue fragments and in some cases for initial characterization
of LC biochemically. This appears to be notable for both
primary large cell undifferentiated carcinomas of the lung and
adenocarcinomas metastatic to the lung, although the number
of tumors studied in each group was small (N = 3). Large cell
undifferentiated carcinoma of the lung which, heretofore, was
thought to be an undifferentiated carcinoma devoid of any
characteristic biochemical features (6, 7) demonstrated specific
alterations in PGD2 and PGE2 biosynthesis which appeared to
be characteristic of only this LC cell type. Adenocarcinomas
metastatic to the lung also demonstrated lower levels of all 5
PGs studied compared to primary adenocarcinomas of the lung
and these specific alterations in PG metabolic profiles could
prove to be useful in distinguishing these particular lung tumors
from other LC cell types. It should be emphasized that further
studies, involving large numbers of primary large cell undiffer
entiated LC and adenocarcinomas metastatic to the lung, will
be required to confirm these preliminary observations. More
over, detailed studies of endogenous PG biosynthesis in estab
lished human LC cell lines derived from fresh tumor material
may provide better characterization of the specific metabolic
pathways involved and allow further investigation of the role
of altered PG patterns in the pathogenesis of human LC.
In summary, the present studies demonstrate that biosynthetic profiles of bisenoic PG and TXB2 differ in matched NL
and LC tissues from individual LC patients. The discrepancies
in the biosynthetic profiles of the two tissues may result, in
part, from altered PG biosynthesis in lung tumor cells. Further
investigations of PG biosynthesis in human LC are required,
however, for a more complete understanding of the role of this
family of compounds in LC as well as normal pulmonary tissue
biology.
ACKNOWLEDGMENTS
The authors express their appreciation to Robert Brennan for assist
ance with data analysis and to Kathy Gill for assistance in the organi
zation and processing of this manuscript.
REFERENCES
1. Silverberg, E. Cancer Statistics, 1985. CA Cancer J. Clin., 35: 19-35, 1985.
2. Homi, J. W., Asiré,
A. J., Young, J. L., and Pollack, E. S. SEER Program:
Cancer Incidence and Mortality in the United States 1973-1981. NIH
Publication No. 85-1837. Bethesda, MD: Department of Health and Human
Services, 1984.
3. Loeb, L. A., Ernster, V. L., Warner, K. E., Abbotts, J., and Laszlo, J. L.
Smoking and lung cancer: an overview. Cancer Res., 44: 5940-5958, 1984.
4. American Cancer Society. 1986 Facts and Figures, pp. 1-32. New York:
American Cancer Society, 1986.
5. Minna, J. D., Ihde, D. C., and Glatstein, E. J. Lung cancer: scalpels, beams,
drugs and probes. N. Engl. J. Med., 315: 1411-1413, 1986.
6. Yesner, R. Spectrum of lung cancer and ectopie hormones. Pathol. Aniiu..
»(Parti): 217-240, 1978.
7. Yesner, R., and Carter, D. Pathology of carcinoma of the lung: changing
patterns. Clin. Chest Med., 3: 257-289, 1982.
8. Mathews, M. J., and Gazdar, A. F. Pathology of small cell carcinomas of the
lung and its subtypes. A clinico-pathologic correlation. In: R. Livingston
(ed.). Lung Cancer, Vol. 1, pp. 283-307. The Hague: Martinus Nijhoff, 1981.
9. Gazdar, A. F., Carney, D. N., Nau, M. M., and Minna, J. D. Characterization
of variant subclasses of cell lines derived from small cell lung cancer having
distinctive biochemical, morphological, and growth properties. Cancer Res.,
«.•2924-2930,1985.
10. Hyman, A. L., Spannhake, E. W., and Kadowitz, P. J. State of the art:
prostaglandins and the lung. Am. Rev. Respir. Dis., //": 111-136, 1978.
11. Newman, J. H., and Sheller, J. R. Review: arachidonic acid metabolites in
the healthy and diseased lung. Am. J. Med. Sci., 288: 75-80, 1984.
12. Mathe, A. A., Hedqvist, P., Strandberg, I. F., and Leslie, C. A. Aspects of
prostaglandin function in the lung (two parts). N. Engl. J. Med., 296: 850855, 910-914, 1977.
13. Seyberth, H. W., Segre, G. V., Morgan, J. L., Sweetman, B. J., Potts, Jr. J.
T., and Oates, J. A. Prostaglandins as mediators of hypercalcemia associated
with certain types of cancer. N. Engl. J. Med., 293: 1278-1283, 1975.
14. Bennett, A., Carroll, M. A., Stamford, I. F., Whimster, W. F., and Williams,
F. Prostaglandins and human lung carcinomas. Br. J. Cancer, 46: 888-893,
1982.
15. Lau, S. S., McMahon, J. B., McMenamin, M. G., Schuller, H. M., and Boyd,
M. R. Differential metabolism of arachidonic acid in human lung carcinoma
cell lines. Cancer Res., 47: 3757-3762, 1987.
16. Hubbard, W. C., Alley, M. C., McLemore, T. L., and Boyd, M. R. Evidence
for thromboxane biosynthesis in a human lung adenocarcinoma cell line,
A549. Fed. Proc., 46:692, 1987.
17. Hubbard, W. C., Alley, M. C, McLemore, T. L., and Boyd, M. R. Evidence
for thromboxane biosynthesis in established cell lines derived from human
lung adenocarcinomas. Cancer Res., in press, 1988.
18. Bakhle, Y. S., and Vane, J. R. Pharmacokinetic function of the pulmonary
circulation. Physiol. Rev., 54:1007-1045, 1974.
19. Samuelsson, B., Granstrom, E., Green, K., Hamberg, M., and 11¡miniarsi
rom,
S. Prostaglandins. Annu. Rev. Biochem., 44:669-695, 1975.
20. Samuelsson, B., Goldyne, M., Granstrom, E., Hamberg, M., Hammarstrom,
3146
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
PROSTANOID BIOSYNTHESIS IN HUMAN LUNG CANCERS
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
S., and Malmsten, C. Prostaglandins and Ihromboxanes. Annu. Rev.
Biochem., 47: 997-1029. 1978.
Watanabc, K., Iguchi, Y., Iguchi, S., Arai, Y., Hayaishi, O., and Roberts, L.
J., II. Stereospecific conversion of proslaglandin D¡ to (5Z,I3£)-(I55)9«,1I#.l5-trihydroxyprosta-5,l3-dien-l-oic
acid (9it,I Irf-prostaghmdin F2)
and of proslaglandin II (o prostaglandin F¡,,by bovine lung prostaglandin
F synlhasc. Proc. Nail. Acad. Sci. USA. 83: 1583-1587, 1986.
Powell, W. S., and Solomon, S. Formation of 20-hydroxyprostaglandins by
lungs of pregnant rabbits. J. Biol. Chem. 263: 4609-4616. 1978.
Powell. W. S. Distribution of prostaglandin uj-hydroxylases in different
tissues. Prostaglandins, 19: 701-710, 1980.
Honn, K. V., Bockman. R. S., and Marnett, L. J. Prostaglandins and cancer.
A review of tumor initiation through tumor metastalis. Prostaglandins, 21:
833-850. 1981.
Honn, K. V. In: T. J. Powles. R. S. Bockman. K. V. Honn, and P. Ramwell
(eds.), Prostaglandins and Cancer: First International Conference. Vol. 2,
pp. 733-751. New York: Alan R. Liss, 1982.
Rolland, P. II., Martin, P. M., Jacqucmier. J.. Rolland. A. M.. and Toga,
M. Proslaglandins in human breast cancer: evidence suggesting that elevated
prostaglandin production is a marker of metastatic potential for neoplastic
cells. J. Nati. Cancer Inst.. 64: 1061-1070. 1980.
Bresnick, E., Meunier, P.. and Lamden, M. Epidermal prostaglandins after
topical application of a tumor promoter. Cancer Lett., 7: 121-124, 1979.
Verma, A., Ashendel, C. L., and Boutwell. R. K. Inhibition by prostaglandin
synthesis inhibitors of the induction of epidermal ornithine decarboxylase
activity, the accumulation of prostaglandins and tumor promotion caused by
12-O-letradecanoyl-phorbol-13-acetate. Cancer Res., 40:308-312, 1980.
Levine, L. Arachidonic acid transformation and tumor production. Adv.
Cancer Res., 35: 49-52, 1981.
Fisher, S. M.. Mills. G. D.. and Slaga. T. J. Inhibition of mouse skin tumor
promotion by several inhibitors of arachidonic acid metabolism. Carcinogenesis (Lond.), 3: 1243-1250, 1982.
Levine. L. Stimulation of cellular prostaglandin production by phorbol esters
and growth factors and inhibition by cancer chemopreventive agents. in: T.
J. Powles, R. S. Bockmann, K. V. Honn, and P. W. Ramwell (eds.), Prosta
glandins and Related Lipids. Vol. 2, pp. 189-205. New York: Alan R. Liss,
1984.
Furstenberger, G., Gross, M., and Marks, F. On the role of prostaglandins
in the induction of epidermal proliferation, hyperplasia and tumor promotion
in mouse skin. In: T. J. Powles, R. S. Bockman, K. V. Honn. and P. W.
Ramwell (eds.), Prostaglandins and Related Lipids. Vol. 2, pp. 239-251.
New York: Alan R. Liss, 1984.
Fischer. S. M., and Slaga, T. J. Modulation of prostaglandin synthesis and
tumor promotion. In: T. J. Powles, R. S. Bockmann, K. V. Honn, and P. W.
Ramwell (eds.), Prostaglandins and Related Lipids, Vol. 2, pp. 255-261.
New York: Alan R. Liss. 1984.
Fischer, S. M. Arachidonatc cascade and skin tumor promotion. In: H.
Thaler-Dao, A. Crastes de Paulet, and R. Paoletli (eds.). Icosanoids and
Cancer, pp. 79-90. New York: Raven Press, 1984.
Furstenberger, G., Gross, M., and Marks, F. Involvement of prostaglandins
in the process of skin tumor promotion. In: H. Thaler-Dao, A. Crasles de
Paulet. and R. Paoletti (eds.). Icosanoids and Cancer, pp. 91-100. New York:
Raven Press, 1984.
Kato, R., Nakadate, T., and Yamamoto, S. Involvement of lipoxygenase
products of arachidonic acid in tumor promoting activity of TPS. In: H.
Thaler-Dao. A. Crasles de Paulet, and R. Paoletti (eds.), Icosanoids and
Cancer, pp. 101-114. New York: Raven Press. 1984.
Levine, L., Goldstein, S. M., Snoek, G. T., and Rigas. A. Arachidonic acid
metabolism by cells in culture: effect of tumor promoters. ln:H. Thaler-Dao.
A. Crastes de Paulet, and R. Paolelli (eds.), Icosanoids and Cancer, pp. 115121. New York: Raven Press. 1984.
Sydney. E. and Salmon, M. D. Interrelationship of endogenous macrophages,
prostaglandin synthesis and tumor cell clonogenicity in human tumor biop
sies. In: T. Powles, T. J. Bockman, K. V. Honn, and P. Ramwell (eds.),
Prostaglandins and Cancer: First International Conference, pp. 633-651.
New York: Alan R. Liss, 1984.
Meerpohl, H. G., Bauknechl. T.. Tritschler, V.. and Lang, H. The influence
of prostaglandin E¡on macrophage mediated cytotoxicity. In: W. Forster
(ed.). International Workshop: Heterogeneity of Mononuclear Phagocytes,
pp. 428-461. New York: Academic Press. 1981.
Walker. C.. Kristensen. F., Beltens. F., and DeWeck, A. L. Lymphokine
regulation of activated «;,) lymphocytes. I. Prostaglandin E2-induced inhi
bition of interleukin-2 production. J. luminimi.. 130: 1770-1778, 1983.
Young, M. R., and Henderson, S. Enhancement in immunity of tumor
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
bearing time by immunization against prostaglandin E2. luminimi. Commun.,
//: 345-352. 1982.
Milincili. M. J.. Herberman, R. B.. and Holden, H. T. Inhibition of murine
natural killer cell activity by prostaglandins. J. luminimi., 124: 2682-2688,
1980.
Taffet, S. M., and Russell. S. W. Macrophage-mediated tumor cell killing:
regulation of expression of cytolytic activity of prostaglandin E. J. Immunol.,
126: 424-432, 1981.
Bennett. A., Charlier. E. M., McDonald, A. M., Simpson, J. S., Stanford, I.
F., and Zebro. T. Prostaglandins and breast cancer. Lancet, 2: 624-626.
1977.
Bennet, A. Proslaglandins and Cancer. In: S. M. M. Karim (ed.). Practical
Applications of Prostaglandins and Their Synthesis Inhibitors, pp. 149-165.
Lancaster, United Kingdom: MTP Press, 1979.
Bennet, A. Proslaglandins and inhibitors of (heir synthesis in cancer growth
and spread. In: D. P. Rose (ed.). Endocrinology of Cancer, Vol. 3, pp. 113122. Boca Raton. FL: CRC Press, 1982.
Bauknechl, T. A., Siegel, Meerpohl, H. G., and Zahradnik, H. P. Formation
of prostaglandins by ovarian carinomas. Proslaglandins, 29: 665-671, 1985.
Robertson, R. P., Baylink, D. J., Marini, J. J., and Adkison, H. W. Elevated
prostaglandins and suppressedparathyroid hormone associated with hypercalcemia and renal cell carcinoma. J. Clin. Endocrinol. Metab., 41: 164-171,
1975.
Young, M. R.. Young. M. E., and Wepsic, H. T. Effects of prostaglandin E,producing nonmetastatic Lewis lung carcinoma cells on the migration of
proslaglandin E¡—responsive
metaslatic Lewis lung carcinoma cells. Cancer
Res., 47: 3679-3683. 1987.
Hubbard, W. C. Litterst, C. L.. Liu, M. C., Bleecker, E. R., Eggleston, J.
C., McLemore.T. L.,and Boyd, M. R. Profiling of prostaglandin biosynthesis
in biopsy fragments of human lung carcinoma and normal human lung by
capillary gas chromatography-negative ion chemical ionizalion mass spectromelry. Prostaglandins, 32: 889-906. 1986.
Weber, G. Enzymology of cancer cells. Part I. N. Engl. J. Med., 296.- 486493, 1977.
Weber, G. Enrymology of cancer cells. Part II. N. Engl. J. Med., 296: 541551, 1977.
McLemore, T. L., Liu, M. C., Blacker, P. C., Gregg, M.. Alley, M. C.,
Abbott, B. J., Shoemaker, R. H., Bohlman, M. E., Litterst, C. L., Hubbard,
W. C., Brennan. R. H., McMahon, J. B., Fine, D. L.. Eggleston, J. C., Mayo,
J. G., and Boyd. M. R. A novel inlrapulmonary model for the orthotopic
propagation of human lung cancers in athymic nude mice. Cancer Res., 47:
5132-5140. 1987.
Roggli. V. L., Vollmer, R. T., Greenberg, S. D., McGavran, M. H., Spjut,
H. J., and Yesner, R. Lung cancer heterogeneity: a blinded and randomized
study of 100 consecutive cases. Hum. Pathol., 16: 569-579, 1985.
Reich, R., and Martin, G. R. Reduction of the invasive and metastatic activity
of tumor cells by inhibitors of specific pathways of arachidonic acid nu-iah
olism. J. Cell Biol., 105: 182, 1987.
McLemore, T. L.. and Martin. R. R. Chapter 1. Pulmonary carcinogenesis:
aryl hydrocarbon hydroxylase. In. R. B. Livingston (ed.), Cancer Treatment
and Research, Vol. I, pp. 1-34. The Hague: Martinus Nijhoff, 1981.
Martin, R. R., and McLemore, T. L. The role of macrophages and other
cells in human carcinogenesis. In: H. Wallers (ed.). Handbook of Cancer
Immunology, Vol. 7, Chap. 5, pp. 125-138. New York: Garland Press, 1981.
McLemore, T. L., Martin, R. R., Busbee, D. L., Richie, R. C., Springer, R.
R., Topell, K. L., and Cantrell, E. T. Aryl hydrocarbon hydroxylase activity
in pulmonary macrophages and lymphocytes from lung cancer and noncancer
patients. Cancer Res., 37: 1175-1181, 1977.
McLemore. T. L., Martin, R. R., Busbee. D. L., Richie. R. C., Springer. R.
R.. Topell, K. L., and Cantrell, E. T. Aryl hydrocarbon hydroxylase activity
in pulmonary macrophages and lymphocytes from lung cancer and noncancer
patients. Cancer Res.. 37: 1175-1181, 1977.
McLemore. T. L., Martin. R. R., Wray, N. P., Cantrell, E. T., and Busbee,
D. L. Dissociation between aryl hydrocarbon hydroxylase activity in cultured
pulmonary macrophages and blood lymphocytes from lung cancer patients.
Cancer Res.. 38: 3805-3811, 1978.
McLemore, T. L., Martin, R. R., Pickard. L. R., Springer, R. R., Wray, N.
P.. Topell, K. L., Mattox, K. L., Guinn, G. A., Cantrell, E. T., and Busbee,
D. L. Analysis of aryl hydrocarbon hydroxylase activity in human lung tissue,
pulmonary macrophages, and blood lymphocytes. Cancer (Phila.), 41: 22922300. 1978.
McLemore. T. L.. Litterst, C. L.. Hubbard. W. C., Paszek, M. C., Liu, M.
C., Eggleston. J. C., and Boyd, M. R. 4-lpomeanol metabolism in normal
lung and lung tumor expiants from lung cancer patients. Proc. Am. Assoc.
Cancer Res., 2«:1681. 1987.
3147
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.
Profiles of Prostaglandin Biosynthesis in Normal Lung and
Tumor Tissue from Lung Cancer Patients
Theodore L. McLemore, Walter C. Hubbard, Charles L. Litterst, et al.
Cancer Res 1988;48:3140-3147.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/48/11/3140
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on July 7, 2017. © 1988 American Association for Cancer Research.