| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Oncology, Markers, Clinical Correlates |
Departments of Molecular and Cellular Oncology [S. E., J. E., C. M. S., E. A. G.], Pathology and Medicine (Dermatology) [V. G. P.], Biostatistics [M. M.], University of Texas, M. D. Anderson Cancer Center, Houston, Texas 77030, and Centro de Oncologia, Hospital Sirio Libanes, Sao Paulo, Brazil 01308-050 [A. C. B.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Nitric oxide (NO) is an important bioactive agent and signaling molecule that mediates a variety of actions such as vasodilatation, neurotransmission, and host defense and has been proposed to contribute to the pathogenesis of cancer (2, 3, 4, 5) . NO is synthesized from arginine by NOSs.3 The genes for three isoforms of NOS have been cloned; two of these isoforms are calcium dependent and constitutively expressed: endothelial NOS (type III) and neuronal NOS (type I). The iNOS (type II) is calcium independent and cytokine inducible. It is also now known that NOS type I and NOS type III can be induced and that low levels of NOS type II are expressed in some tissues constitutively (6, 7, 8) . Only the inducible form, iNOS, once activated can produce higher levels of NO (in the micromolar range) for a long duration (days; Refs. 9 and 10 ). Recent studies have shown that NO may be involved in inhibiting cell proliferation, differentiation (11) , and apoptosis (12) . Damaging effects of NO may be attributable to reaction with superoxide anions to yield peroxynitrite, which is a potent nitrating and nitrosylating agent. Peroxynitrite can oxidize nuclear DNA and membrane phospholipids and also nitrate either free or protein-associated tyrosines to form NTs (13) . Thus, the occurrence of NT in tissues has been measured as a marker of peroxynitrite formation (14) . Some human tumors are reported to express iNOS, and the product of its enzymatic action, NO, is proposed to affect the clinical features of these tumors (3 , 4 , 8 , 15, 16, 17, 18) . Although iNOS expression by tumors in animal models has been studied for years, human tumor studies are recent.
The treatment of advanced melanoma patients with a 5-drug regimen
consisting of IL-2, IFN-
2a, dacarbazine (DTIC), cisplatin,
and vinblastine, referred to as "biochemotherapy," has shown
encouraging results (19
, 20)
. Although the mechanism of
biochemotherapy is not well understood, some biological responses that
may be related to immune effector cell activity have been investigated
during and after therapy. In the original report of a clinical trial
using biochemotherapy in the neoadjuvant setting for stage III
patients, Anderson et al. (21)
found that serum
nitrite levels were higher in patients who showed major responses to
therapy, relative to those who did not, although the statistical
significance was marginal. Levels of serum neopterin, a product of
activated macrophages, correlated weakly with nitrite levels; however,
there was no correlation of neopterin levels with response to
treatment. It was therefore concluded that the elevated serum nitrite
in responding patients was not of macrophage origin and perhaps derived
from tumor cells or endothelial cells.
In the present study, we have attempted to further explore the source of elevated nitrite levels in patients responding to biochemotherapy, using the same patient population studied by Anderson et al. (21) . The working hypothesis was that melanoma tumor cells produce iNOS and provide the source of nitrite seen in the sera of responding patients. A corollary to this would be the expectation that patients with tumor iNOS would have a higher rate of response to therapy and potentially a longer survival. We designed our study to determine whether: (a) melanoma cells express iNOS and NT; (b) tumor expression of these molecules correlates with serum nitrite levels; and (c) tumor expression of these molecules correlates with patient outcomes. We report our findings that some melanoma tumors do express iNOS and NT, but there is no correlation between tumor expression and serum nitrite levels or response to therapy. We also report the unexpected finding that tumor expression of iNOS and NT is associated with poor survival of melanoma patients.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2a, 5
M units/m2 s.c. on days
1–5. One cycle of therapy was 28 days. All patients received a minimum
of two cycles of therapy prior to surgery. Surgery was usually
performed within 2–4 weeks after the last biochemotherapy cycle. For
the purpose of the present report, patients were categorized by the
presence or absence of a major response to neoadjuvant therapy. Major
response is defined as a 50% or greater reduction in clinical or
radiographic bidimensional tumor measurements (clinical complete or
partial response), or the pathological finding of <50% residual
viable tumor in the surgical specimen, regardless of clinical
measurements.
Specimen Collection and Storage.
Blood was collected at 
10 a.m. on 4 separate days: before
biochemotherapy (baseline), and on days 5, 6, and 7 of treatment. The
day 5 collection took place at the termination of the IL-2 infusion,
corresponding to 24 h after the last cisplatin dose. Serum was
separated from the blood within 2 h of collection and frozen at
-80°C for later analysis. Formalin-fixed and paraffin-embedded
tissue sections were obtained from the Melanoma and Skin Cancer Core
Laboratory of the University of Texas M. D. Anderson Cancer Center for
use in immunohistochemical staining of tumor tissues, as approved in
the research arm of the protocol.
Reagents.
Anti-iNOS mouse monoclonal antibody (Transduction Laboratories,
Lexington, KY) was used for iNOS immunohistochemistry, and anti-NT
mouse monoclonal antibody (Upstate Biotechnology, Lake Placid, NY) was
used for NT staining; both antibodies were confirmed as being
cross-reactive between species. Preimmune normal mouse IgG (Vector
Laboratories, Burlingame, CA) was used as a negative control.
Anti-vimentin antibody (BioGenex Laboratories, San Ramon, CA) was used
as a positive control for all melanoma staining. Other reagents
including sulfanilamide, N-1-naphthyl-ethylenediamine,
ammonium formate, zinc sulfate, sodium nitrate, and sodium nitrite was
purchased from Sigma Chemical Co., Inc. (St. Louis, MO).
Nitrite Determination in Serum.
NO was measured as the total of its oxidation products (nitrite and
nitrate) after enzymatic reduction by Escherichia coli
nitrate reductase prepared according to published methods
(22)
. The serum assay for NO oxidation products (nitrite
plus nitrate) was performed on freshly thawed serum samples as
described previously (21
, 22)
. Briefly, frozen serum was
thawed at room temperature, and 100 µl were aliquoted into a
microcentrifuge tube; 5 µl of 30% zinc sulfate were added with
vortexing to allow protein precipitation, the tubes were centrifuged
for 12 min at 14,000 rpm. Then 56 µl of supernatant were removed and
placed in a new microcentrifuge tube to which 62 µl of nitrate
reductase mix [1 part enzyme suspension to 10 parts 2.4
M HEPES (pH 7.2) and 10 parts 1
M ammonium formate (pH 7.2)] were added; the
tubes were vortexed and incubated for 30 min at 37°C, followed by
centrifugation for 5 min. The enzyme suspension used was E.
coli nitrate reductase, used as a crude preparation of E.
coli according to published methods (23)
. Eighty µl
of supernatant were pipetted into a 96-well plate with 80 µl of
Griess reagent (1% sulfanilamide, 0.1%
N-1-naphthyl-ethylenediamine, and 2.5% phosphoric acid in
distilled water) and incubated at room temperature for 10 min, and the
absorbance was read at 540 nm against a reference of 650 nm in a
microplate spectrophotometer. A Dynatech MR700 spectrophotometer plate
reader (Chantilly, VA) was used to read the sample plates for NO
determination. Standards of known concentrations of sodium nitrate and
sodium nitrite in serial dilutions were used as positive controls to
create a standard curve. Standards and samples were subjected to
identical treatment. The final nitrite concentration, which was the sum
of the serum nitrite plus the reduced nitrate in the Griess reaction,
was reported in µM.
Immunohistochemical Staining.
Immunohistochemical staining was performed with 10% formalin-fixed,
paraffin-embedded melanoma tissue, cut 4–6 µm thick. Sections
were placed on silanized slides (Histology Control Systems, Glen Head,
NY), deparaffinized in xylene, and rehydrated in descending grades
(from 100 to 85%) of ethanol. To enhance the immunostaining and
restore the maximal antigenicity of cytokines, sections then were
placed in antigen unmasking solution (Vector Laboratories, Burlingame,
CA) and microwaved intermittently for up to 10 min to maintain a
boiling temperature. After the slides were cooled at room temperature
for 30 min, they were washed in distilled water and PBS. After this
initial preparation, the slides were removed from PBS and covered with
3% H2O2 (Sigma Chemical
Co.) in methanol to block endogenous peroxidase activity. All
incubations were carried out at room temperature in a humidified
covered slide chamber. The slides were washed in PBS before incubation
in Tris-buffered saline (TBS) containing 0.05% Triton X-100 (Sigma
Chemical Co.) for 15 min to permeabilize the cells. An
avidin-biotin-peroxidase complex (ABC) kit (Vectastain; Vector
Laboratories) was then used to detect the primary antibody staining.
These kits are specific for the species of primary antibody used and
contain a blocking serum, a secondary biotinylated antibody, and the
ABC reagent. After the slides were incubated for 30 min with the
blocking serum, the primary antibody at various dilutions (1:100 to
1:200 for iNOS and 1:50 for NT) was added, and the slides were
incubated for 60 min at room temperature. The slides were then washed,
incubated for 30 min with secondary biotinylated antibody, washed
again, and then incubated for 30 min with the ABC reagent. After the
slides were washed in PBS, the immunostaining was developed with the
use of 3-amino-9-ethylcarbazole as a chromogen for 15 min. Slides were
counterstained with hematoxylin (Vector Laboratories) and mounted with
Aqua-Mount (Lerner Laboratories, Pittsburgh, PA). Control tissues used
were the samples of formalin-fixed and paraffin-embedded normal human
nevi obtained as incidental material from our Pathology Department. The
following scores were assigned to each specimen according to the
percentage of positively stained cells in the tumor: 4+, >75% cells
were positive; 3+, 51–75% cells positive; 2+, 26–50% cells
positive; 1+, 5–25% cells are positive; +/-, <5% cells positive;
and -, no positive staining.
Statistical Analysis.
Kaplan-Meier estimates of survival were calculated from the original
date of melanoma diagnosis and also from the first day of treatment for
stage III disease. Comparisons of survival between groups of patients
who tested positive and those who tested negative for iNOS and NT were
assessed with the log-rank test. An original date of melanoma diagnosis
could not be ascertained for two patients who were negative for iNOS
and NT and one patient who was positive for iNOS and NT. These three
patients were excluded from the analysis of survival performed from
date of diagnosis. McNemar’s test of symmetry was used to evaluate the
consistency of testing positive for iNOS and for NT (24)
.
The null hypothesis for the McNemar’s test is that there is an
agreement between two parameters. Fisher’s exact test was used to
compare the proportion of treatment responders between those patients
who tested iNOS positive and those who tested negative.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
|
. Posttreatment tumors
from 12 of the 20 subjects were positive for iNOS (Fig. 1)
, and nine of these tumors were positive for NT (Fig. 2)
. An interesting observation was related to the correlation between
iNOS and NT expression within the same tumor (P = 0.248
by McNemar’s test), suggesting that there may be agreement between the
two markers (24)
. All iNOS-negative tumors were also
negative for NT by immunohistochemistry. Only 3 of 12 iNOS-positive
tumors did not express NT.
|
|
Serum Nitrite Levels.
As a surrogate measurement of NO production, serum nitrite levels for
the 20 patients during the first biochemotherapy treatment were
measured (Fig. 3)
. For all patients, serum nitrite rose higher than baseline on days 5
and 6 and declined on day 7. The levels of nitrite in the serum did not
show a correlation with tumor expression of iNOS. Therefore, tumor iNOS
does not appear to be the source of serum nitrite in these patients.
|
To address the impact of tumor iNOS and NT on survival, information on
survival times from both the original diagnosis of melanoma and from
the initiation of therapy was gathered for all patients. Two patients
who presented with stage III disease had an unknown site of primary
tumor; for another patient, records regarding the primary tumor could
not be obtained. These three patients are excluded from analysis of
survival from the date of original diagnosis but are included in the
analysis from the initiation of therapy. In contrast to the lack of
correlation of iNOS expression with response to therapy, survival
appeared to be significantly impacted (Fig. 4)
. The median survival from diagnosis of the primary tumor (Fig. 4
A) and from initiation of therapy for stage III disease
(Fig. 4
B) was 30.3 and 22.1 months, respectively, for the
patients who had iNOS-expressing tumors; the median survival times were
not reached in either case for iNOS-negative patients
(P = 0.001 and P < 0.001,
respectively). A similar pattern was found when analyzing the
association of tumor NT expression with survival (Fig. 5)
. A significantly shorter survival was found from initiation of
treatment for patients with NT-positive tumors (median survival, 24.4
months versus not reached for patients with NT negative
tumors, P = 0.020; Fig. 5
B). There was a
trend toward inferior survival from diagnosis for patients with
NT-positive tumors, but the difference did not reach significance (Fig. 5
A).
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, or IL-6. Thus,
tumor does not appear to be the source of elevated serum nitrite in
responding patients reported by Anderson et al.
(21)
. Finally, contrary to our expectations, we found that
the presence of iNOS and NT in posttreatment tumors predicted a poor
survival. There are several limitations to our study that must be considered when interpreting these data. The first limitation is the fact that our analysis is based on posttreatment tumor samples. These samples, which were obtained as part of the clinical trial, provided the advantage of surgically excised material of adequate quantity and preserved histology, as opposed to the scant fine-needle aspirates used in the pretreatment, diagnostic setting. However, if we are to suggest that melanoma tumors producing iNOS display an aggressive phenotype, it would be more appropriate to make this observation on pretreatment material, as one could postulate that the cytokines included in biochemotherapy might indirectly induce tumor iNOS expression. We were able to analyze pretreatment tumors in half of the patients and found good correlation between pre- and posttreatment samples; however, future studies should focus on material from untreated patients. The second limitation also derives from our use of posttreatment samples in that it forced us to exclude from analysis four patients who achieved complete resolution of their disease, thus leaving no tumor to study. This may be less important, as we were unable to demonstrate a correlation of tumor iNOS with response to treatment. Finally, the number of patients in this study was small. Although the survival data are provoking, they must be confirmed in an appropriately designed trial with a larger number of patients before strong conclusions can be drawn.
Bearing in mind the above limitations, two observations are worthy of discussion, those being the findings that some melanoma metastases produce iNOS and NT, and that stage III patients bearing these iNOS/NT-positive melanomas have a shortened survival compared with those with iNOS/NT-negative tumors. This suggests that the presence of tumor iNOS, and thus NO, provides an advantage for enhanced tumor growth and survival. Potential mechanisms include vasodilatation with improved tumor blood flow and damaging oxidative effects directed toward lymphocytes, inducing membrane and DNA damage. In agreement with our findings are reports of increased iNOS expression by primary human tumors compared with normal tissue, including colon carcinomas (18) , colon adenomas (8) , pancreatic adenocarcinomas (15) , and gastric cancers (17) . To our knowledge, only one other study has examined expression in human melanomas (25) . Those investigators studied lymph nodes and subcutaneous metastases and found that patients with a lower percentage of iNOS-expressing tumor cells had a greater risk of developing distant metastases. Treatment details and specific survival times were not provided, and a significant impact on survival could not be demonstrated. The apparent discrepancy between these data and ours may be explained in part by different methodologies used to quantitate iNOS expression. However, the two studies cannot be directly compared with respect to influence on survival, based on the information provided.
Similar examples of conflicting reports of iNOS expression exist for
other tumor types as well (26)
. One explanation for the
inconsistent findings may be a differential effect of NO on tumor
behavior, depending on the quantity of this molecule produced. Recent
studies have shown that transfecting metastatic murine melanoma cells
with the full-length iNOS gene leads to the
production of micromolar quantities of NO and suppression of metastatic
potential (27)
. Melanoma cells similarly transfected with
a truncated iNOS gene did not produce NO and retained their
metastatic potential, suggesting that high levels of NO caused
self-destruction of these tumor cells (27
, 28)
.
Conversely, transient or low levels of NO are suggested to benefit
tumor growth by increasing blood supply via its well-known,
vascular-relaxing property (29)
. Another factor that may
contribute to the conflicting observations is the cytokine inducibility
of iNOS expression. One recent study showed that treatment of murine
melanoma cells with IFN-
plus lipopolysaccharide results in the
up-regulation of iNOS transcription in nonmetastatic cells but not in a
metastatic cell line, suggesting a variable transcriptional basis for
the differences in iNOS (30)
. Taken together, these
observations indicate that the level of iNOS expression and NO
production by the tumor may be associated with tumor viability, with
the very high NO levels leading to tumor destruction but moderate
levels promoting growth. Such alterations in the levels of iNOS and NO
could be influenced by numerous patient factors, including medical and
surgical treatment received, and other cancer-related physiological
stresses such as pain, dehydration, cachexia, and infection.
In conclusion, our data suggest that iNOS and NT expression by melanoma metastases may be predictive of a poor outcome, information that may be of clinical value when making therapeutic decisions for individual patients. However, we view our findings as preliminary. Our current efforts focus on confirming these data with a larger number of patients, studying untreated samples to eliminate the impact of exogenous cytokines, and analyzing early-stage tumors for prognostic information.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by National Cancer
Institute Grant RO1 CA64906 (to E. A. G.), Cancer Center Support
(Core) Grant P30 CA16672 to M. D. Anderson Cancer Center Melanoma
Research Program Core Laboratory, and Training Grant T32 CA72371 (to
J. E.). 
2 To whom requests for reprints should be
addressed, at Department of Molecular and Cellular Oncology, Box 79,
University of Texas M. D. Anderson Cancer Center, Houston, TX 77030.
Phone: (713) 792-7477; Fax: (713) 794-4784; E-mail: egrimm{at}mdanderson.org
3 The abbreviations used are: NOS, nitric oxide
synthase; iNOS, inducible NOS; NT, nitrotyrosine; IL, interleukin.
Received 5/31/00; revised 8/17/00; accepted 9/13/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, and interleukin-2 for patients with metastatic melanoma. J. Clin. Oncol., 16: 1752-1759, 1998.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  L. L. Hardy, D. A. Wick, and J. R. Webb Conversion of Tyrosine to the Inflammation-Associated Analog 3'-Nitrotyrosine at Either TCR- or MHC-Contact Positions Can Profoundly Affect Recognition of the MHC Class I-Restricted Epitope of Lymphocytic Choriomeningitis Virus Glycoprotein 33 by CD8 T Cells J. Immunol., May 1, 2008; 180(9): 5956 - 5962. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. V. Madhunapantula, D. Desai, A. Sharma, S. J. Huh, S. Amin, and G. P. Robertson PBISe, a novel selenium-containing drug for the treatment of malignant melanoma Mol. Cancer Ther., May 1, 2008; 7(5): 1297 - 1308. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Nagaraj and D. I. Gabrilovich Tumor Escape Mechanism Governed by Myeloid-Derived Suppressor Cells Cancer Res., April 15, 2008; 68(8): 2561 - 2563. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Thomas, L. A. Ridnour, M. G. Espey, S. Donzelli, S. Ambs, S. P. Hussain, C. C. Harris, W. DeGraff, D. D. Roberts, J. B. Mitchell, et al. Superoxide Fluxes Limit Nitric Oxide-induced Signaling J. Biol. Chem., September 8, 2006; 281(36): 25984 - 25993. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Su, J. Kuball, A.-P. Barreiros, D. Gottfried, E. A. Ferreira, M. Theobald, P. R. Galle, D. Strand, and S. Strand Nitric Oxide Promotes Resistance to Tumor Suppression by CTLs J. Immunol., April 1, 2006; 176(7): 3923 - 3930. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.-H. Tang and E. A. Grimm Depletion of Endogenous Nitric Oxide Enhances Cisplatin-induced Apoptosis in a p53-dependent Manner in Melanoma Cell Lines J. Biol. Chem., January 2, 2004; 279(1): 288 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Crowell, V. E. Steele, C. C. Sigman, and J. R. Fay Is Inducible Nitric Oxide Synthase a Target for Chemoprevention? Mol. Cancer Ther., August 1, 2003; 2(8): 815 - 823. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ekmekcioglu, J. A. Ellerhorst, J. B. Mumm, M. Zheng, L. Broemeling, V. G. Prieto, A. L. Stewart, A. M. Mhashilkar, S. Chada, and E. A. Grimm Negative Association of Melanoma Differentiation-associated Gene (mda-7) and Inducible Nitric Oxide Synthase (iNOS) in Human Melanoma: MDA-7 Regulates iNOS Expression in Melanoma Cells Mol. Cancer Ther., January 1, 2003; 2(1): 9 - 17. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Thomas, M. G. Espey, M. P. Vitek, K. M. Miranda, and D. A. Wink Protein nitration is mediated by heme and free metals through Fenton-type chemistry: An alternative to the NO/O2- reaction PNAS, October 1, 2002; 99(20): 12691 - 12696. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ikeguchi, T. Ueta, Y. Yamane, Y. Hirooka, and N. Kaibara Inducible Nitric Oxide Synthase and Survivin Messenger RNA Expression in Hepatocellular Carcinoma Clin. Cancer Res., October 1, 2002; 8(10): 3131 - 3136. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.-M. Simeone, S. Ekmekcioglu, L. D. Broemeling, E. A. Grimm, and A. M. Tari A Novel Mechanism by Which N-(4-hydroxyphenyl)retinamide Inhibits Breast Cancer Cell Growth: The Production of Nitric Oxide Mol. Cancer Ther., October 1, 2002; 1(12): 1009 - 1017. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Paulukat, M. Bosmann, M. Nold, S. Garkisch, H. Kampfer, S. Frank, J. Raedle, S. Zeuzem, J. Pfeilschifter, and H. Muhl Expression and Release of IL-18 Binding Protein in Response to IFN-{gamma} J. Immunol., December 15, 2001; 167(12): 7038 - 7043. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S. Haqqani, J. F. Kelly, and H. C. Birnboim Selective Nitration of Histone Tyrosine Residues in Vivo in Mutatect Tumors J. Biol. Chem., January 25, 2002; 277(5): 3614 - 3621. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, presence of iNOS in the patient’s
tumor; 
, absence of iNOS in the tumor. Columns, means
of patients’ serum nitrite levels; bars, SD.
