| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Neurology, University of California, San Francisco, CA 94143;
Roche Molecular Systems, Alameda, CA 94501; and
Rocky Mountain Multiple Sclerosis Center, Englewood, CO 80110-2790
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
transcripts were dramatically increased in MS as compared
with controls. This reveals a robust MHC class II up-regulation and
suggests that Ag is being presented locally to activated T cells.
Although analysis of cytokine and cytokine receptor genes
expression showed predominantly increased levels of several Th1
molecules (TGF-
, RANTES, and macrophage-inflammatory protein
(MIP)-1) in MS samples, some Th2 genes (IL-3, IL-5, and IL-6/IL-6R)
were found to be up-regulated as well. Similarly, both proinflammatory
type (CCR1, CCR5) and immunomodulatory type (CCR4, CCR8) chemokine
receptors were differentially expressed in the MS brain. Overall, our
data suggest a complex regulation of the inflammatory response in human
autoimmune demyelination. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The pathological hallmark of MS is the plaque, a well-demarcated white matter lesion characterized histologically by inflammation, particularly macrophages and T cells, demyelination, and gliosis, and different degrees of axonal loss (3, 4). Immunohistochemical and molecular (i.e., RT-PCR) analyses of CNS samples, particularly dissected MS plaques from autopsy tissue and cerebrospinal fluid cells, provide support for a model of lesion development driven by a Th1 type inflammatory response (5, 6, 7, 8, 9, 10, 11). However, although patterns of local proinflammatory cytokine production correlate fairly well with disease in models of MS, particularly in rodents, the dogmatic application of the Th1/Th2 paradigm to human demyelination is considered somehow simplistic (12). Indeed, most studies of gene expression in MS were limited to the analysis of single or a few molecular targets, preventing the formulation of a unifying hypothesis of MS pathogenesis. In contrast, the recent application of comprehensive methods such as single-pass sequencing of cDNA libraries or high capacity microarrays was limited to very few specimens (13, 14, 15).
A PCR-based method for sensitive nucleic acid quantification has been described by Higuchi and colleagues (16). Reverse transcriptase-initiated real-time PCR (kinetic RT-PCR, kRT-PCR) provides a way to monitor product formation as the reaction proceeds, allowing for precise quantification of the initial amount of mRNA present in the sample tube (17, 18). Furthermore, kRT-PCR allows for automation and the simultaneous analysis of multiple samples with a moderately large number of transcripts. Compared with microarrays, the greater sensitivity of PCR assures that most of the transcripts assayed will be detected. Here we use kRT-PCR to analyze the expression profile of 56 genes in brain samples from eight MS patients with active demyelinating lesions and eight controls. Selected targets include several cytokines, chemokines, their receptors, and other immunologically relevant genes as well as genes encoding for myelin components. To our knowledge, this study constitutes to date the most comprehensive gene expression analysis using kRT-PCR. Results show complex, but consistent patterns of predominantly inflammatory immune pathways.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Postmortem MS and non-MS control brain specimens were obtained
from The Rocky Mountain Tissue Bank (Englewood, CO). Frozen sections
from eight individuals with clinical history of MS were divided in two
for both molecular and immunohistochemical analyses. Each analyzed MS
brain sample was stained with luxol fast blue (LFB) and counterstained
with Harris hematoxylin for myelin integrity, trichrome (Tri) for
astrocytes, oil red O (ORO) counterstained with hematoxylin for neutral
lipids, and the anti-macrophage Ab EBM 11 for macrophages. These
techniques allow for standard grading of MS plaques (19).
With the exception of samples MS-7 (type III) and MS-8 (type IV), all
of the MS specimens chosen for this study represent the most active
stages of plaque activity (Table I
).
Sample MS-5 is mostly grade II but it has a marginal region displaying
a beginning of a grade III plaque. Control samples consisted of eight
non-MS, noninflammatory white matter brain specimens. Total RNA was
isolated from 100–200 mg of frozen brain tissue homogenized with
Trizol reagent (Life Technologies, Gaithersburg, MD) according to the
manufacturer’s recommendations.
|
One-step RT-PCRs (20) were performed in low-transmissiveness 96-well reaction plates (MicroAmp Optical; PE Applied Biosystems, Foster City, CA) containing 50 ng total RNA and the following: 50 mM Bicine, pH 8.3; 125 mM KOAc, pH 7.5; 2.5% glycerol (including contribution from the enzyme); 200 µM each of dATP, dCTP, dGTP, and 400 µM dUTP; 0.2 µM each oligonucleotide primer, 0.2x SYBR (Molecular Probes, Eugene, OR) in DMSO (1% final concentration); 1 U uracyl N-glycosilase (AmpErase; PE Applied Biosystems); 4 mM Mn(OAc)2; and 5 U rTth polymerase (GeneAmp; PE Applied Biosystems). Total reaction volume was 50 µl. An initial incubation of 10 min at 45°C was performed to activate uracyl N-glycosilase followed by a reverse transcription step of 30 min at 60°C. All reactions were conducted for 50 cycles of 95°C for 15 s and 55°C for 30 s on an ABI-5700 Sequence Detector (Applied Biosystems). To cover the entire battery of primers, each RNA sample required nine 96-well plates. Cellular GAPDH was amplified from all samples on each plate as a housekeeping gene to normalize expression between different samples and to monitor assay reproducibility. A control without added template was included for each target analyzed. To calibrate quantitation among different runs, a 10-fold dilution series of a GAPDH run-off transcript (106 to 102 initial GAPDH mRNA copies) was included in each reaction plate.
Quantitative analysis
Output data was generated by the instrument on-board software Geneamp 5700 SDS (PE Applied Biosystems) and subsequently transferred to a custom designed MS Excel 2000 spreadsheet for analysis. A log-linear calibration graph was generated by plotting GAPDH copy number (from run-off transcript) for each of the six 10-fold dilutions of GAPDH mRNA against the number of cycles it took for each reaction product to exceed a preset fluorescence threshold (Ct). Ct values for each sample were then compared with those obtained in the standard curve fitted to the points of the calibration graph to obtain quantitative measurements. Finally, all readings were standardized to the amplification values obtained for the housekeeping gene so that the copy number of each transcript is expressed as relative to GAPDH. A cut-off value of Ct = 30 was set as this is the number of cycles it usually took for the fewest GAPDH molecules (100) to reach the threshold in the calibration curve. After normalization for cellular GAPDH content, this Ct corresponds approximately to a value of 10 GAPDH units. Therefore, expression values of <10 were considered background. Although potentially entailing some loss of information, this procedure eliminates variability arising from the assay itself or from tissue handling, and allows the comparison of transcription patterns among different samples. Replicate reactions were conducted to confirm statistically significant differences in gene expression.
Statistical analysis
Parametric (Student’s t) and nonparametric (Mann-Whitney rank-sum) tests were conducted to challenge the hypothesis that there were no differences in the expression levels of the analyzed genes between samples and controls. There was a close agreement between both tests. No adjustment for multiple comparisons was performed in our data to avoid overlooking potentially true differences in gene expression. This type of correction was considered too conservative an approach because it usually mechanizes the interpretive problem negating the value of much of the information in large bodies of data (21). Statistical analysis was performed in consultation with the University of California at San Francisco statistical unit.
Clustering
Expression data was analyzed using Gene Cluster and visualized with Tree View software tools (22) (http://rana.stanford.edu/software). A different weight was assigned to targets according to their level of significance for the Rank-Sum Test. A weight of 0.99 was given to genes showing p < 0.01, 0.95 to those with p < 0.05, and 0.5 to those that did not exhibit statistically significant expression changes. A weight of 0.8 was assigned to genes that reached significance only on the Student t test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Aberrant
MHC expression and well-defined areas of demyelination were also
identified.
Kinetic RT-PCR provides an efficient experimental approach for the
comparison of expression levels of a relatively large number of
transcripts from multiple samples. The primary data output in kRT-PCR
is the Ct, defined as the number of PCR cycles needed to exceed a
predefined fluorescence threshold for each sample. When this threshold
is set within the exponential phase of the reaction, the Ct is
inversely proportional to the log of the starting number of copies of
the target RNA. A 10-fold dilution series of a GAPDH run-off transcript
of known copy number is included in each experiment to build a standard
curve for quantification (Fig. 1). Such a
quantification would allow the accurate assessment of the actual
starting copy number of a given target RNA assuming that RT and PCR
efficiency are the same than those of GAPDH transcripts. Because the
fact that these efficiencies are the same cannot be proven for all
transcripts, expression values are provided as GAPDH equivalents rather
than true copy number (17).
|
).
Thirteen targets showed no difference in expression between MS and
non-MS specimens. Twelve and 17 transcripts in MS and control samples,
respectively, showed a value <10 GAPDH units, which was the cut-off
value for detection,. Average differences in expression are displayed
in Fig. 2. The most dramatic difference
in expression was detected in CD4 transcripts, which showed an average
fold increase (AFI) of 8.51 for MS samples (maximum fold increase,
70.26; Table II). This finding is consistent with the presence of
activated T helper cells in the brain parenchyma of MS patients and
correlates well with our histological data (Table I). Conversely, CD8
transcripts were not detected in the MS or in the control brain. Other
markedly up-regulated genes (AFI >4) included the IL-2 receptor
and 
genes (IL-2R and IL-2R), the C-C chemokine RANTES, IL-6
(B cell stimulation factor 2), and the complement component C1r. Most
of the remaining transcripts considered significant had an AFI between
2 and 4. Targets in this group included the Th1 markers TNF-
receptor 1 (TNF--R1), IL-6R, CCR1, CCR5, IL-12R1, caspase-1,
IL-1, prolactin, and IL-18, all with p values of 0.01
(Table II). Also in this set of targets, TGF1, TGF3, and CCR4
were marginally significant (p = 0.05).
|
|
(AFI =
3.18, p < 0.01) as well as CD4 suggest that the local
microenvironment is enriched with MHC-activating factors, and that Ag
is possibly being presented to T cells. Unexpectedly, IFN-, a known
up-regulator of MHC molecules, was not detected consistently in all
samples in repeated amplification attempts.
Transcripts that fell below the established cut-off included the CD8
receptor and the Th2-related transcripts IL-4, IL-7, IL-10, IL-11,
IL-13, and CCR8 (Table II). Surprisingly, transcripts for IL-12 (p35
and p40 subunits) failed to be detected even though differential
expression was previously reported in PBMC and cerebrospinal fluid from
MS patients (23). Because IL-12-R2 seems to play a
pivotal role in the generation of pathogenic autoreactive Th1 cells,
this transcript was expected to be up-regulated in MS samples. However,
IL-12-R1 but not IL-12-R2 transcripts were found increased in the
MS sample panel. IFN-, HLA-class I genes C and E, IL-1R, and IL-8
were detected at similar levels in MS and controls.
Although no significant differences between MS samples and controls
were observed, myelin basic protein (MBP) transcripts displayed
the highest level of expression relative to GAPDH, being almost 10
times more abundant than the next most abundant target (Table II).
Of interest, high transcriptional levels of the principal myelin
component, MBP, and glial acidic fibrillary protein (GFAP, a marker for
astrocytes) have also been found in the analysis of single-pass
sequenced cDNA libraries from MS patients (S.E.B. and J.R.O.,
unpublished observation). Myelin oligodendrocyte glycoprotein (MOG),
MOG, and myelin-associated glycoprotein showed, on average,
decreased levels of expression in MS samples (Fig. 2, Table II),
although these differences were statistically not significant. Another
transcript found to be less expressed in MS samples was Calbindin
(27K), a vitamin D-dependent calcium binding protein (-4.3 on average,
not statistically significant, Table II). However, this target was
detected at levels very close to the lower limit of the dynamic range
of the assay (10 GAPDH units).
Data was further analyzed by a hierarchical clustering algorithm
that distributed genes and samples along a two-dimensional colored
chart according to their transcriptional profile (Fig. 3). Most of the differentially expressed
targets were linked in the computer-generated dendrogram by shorter
branches, suggesting closer distances and therefore a similar
expression profile. Genes showing no differential expression were
automatically placed at the top of the picture and connected to each
other through longer branches in the dendrogram. In this group, two
families of transcripts ( and ) from the alternative spliced
myelin gene MOG, were placed next to each other linked with very short
branches, reflecting their coordinated expression. Other myelin genes,
such as myelin-associated glycoprotein and MBP, were also located
nearby due to a similar expression profile.
|
) and were joined
by shorter branches than those linking the controls, reflecting a
higher and more homogeneous expression profile in the MS group. The
relative wide range of maximum fold increase (Table II) reflects a
certain degree of variability among the MS samples due to, most likely,
pathogenic heterogeneity. Interestingly, the less active plaques DA
(type III) and RR (type IV) were linked together in the same cluster;
within this cluster was also included sample ME, classified as type II
but with an emerging type III histology at the edge of the plaque. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
From all targets analyzed, CD4 was the most clearly overexpressed in MS
samples (AFI = 8.51). This observation provides evidence of helper
T lymphocytes having homed to the brain of MS patients. Equally
relevant is the elevated expression of the DR transcripts.
Up-regulation of MHC class II genes has been proposed as a marker of
plaque activity (19). Because the level of cell surface
expression of MHC class II molecules directly affects the nature and
magnitude of the immune response, the study of the mechanisms involved
in the regulation of class II expression in the MS plaque is essential
for understanding the inflammatory response in the affected brain. It
is also possible that enhanced or deregulated MHC class II expression
by APCs within the CNS mediates the autoimmune process, as is the case
in experimental allergic encephalomyelitis (24, 25).
However, it is important to note that in many silent plaques devoid of
T cell infiltrates, class II MHC may be expressed at high levels on
reactive microglia. In addition, up-regulation of MHC class II Ags is
not unique to MS tissue, as it has also been detected in other
neurodegenerative diseases or following trauma (2).
Although the costimulatory molecules ICAM-1 and B7 have been identified
on microglia from humans and mice (26), the case for an
effective Ag presentation in the brain requires further
experimentation.
The very high level of expression detected for MBP in both MS and
control samples is of special interest. This transcript showed a
concentration high above the rest of the targets analyzed (
10-fold
higher than the next most abundant target). Although
transcriptional activity does not necessarily correlate with protein
abundance, this could be indicative of either a high mRNA stability or
a continuous synthesis that keeps an internal pool to readily suffice
the need of such proteins and regenerate the myelin sheath during
normal turnover or under a demyelination event (27, 28).
Most of the knowledge on the role of cytokines in autoimmune
demyelination has been obtained through studies in experimental models
or isolated lymphocytes activated in vitro with myelin Ags. Although
the majority of the studies suggest a bias toward a Th1 type of
response, others insinuate a less than strict polarization of the
cytokine profile (29). Our analysis shows a predominant
expression pattern of Th1 cytokines mainly represented by the
MIP-1/RANTES/CCR5 and Caspase-1/IL-1/IL-18 axes. Surprisingly,
key inflammatory-type molecules such IL-2, IFN-, and TNF- did not
display consistent and reproducible expression patterns. In contrast,
concurrently with elevated expression of IL-5 and IL-6/IL-6R,
prototypic Th2-type molecules such as IL-4, IL-10, IL-13, and CCR8 were
undetected. Altogether, this particular pattern of cytokine expression
suggests a complex, not fully polarized regulation of the local immune
response in human autoimmune demyelination.
Finally, it has been shown that B cell activation and Ab responses play an important role in the development of demyelination both in human and experimental disease (30, 31, 32). Abs may participate in myelin destruction through different mechanisms such as myelin opsonization facilitating phagocytosis by macrophages, and/or complement fixation (33). Also, CNS Igs may induce myelinolysis via activation of a Ca2+-dependent myelin-associated protease acting on MBP (34). The up-regulation in MS plaques for transcripts encoding IL-6 (AFI = 4.13) and C1r (AFI = 2.22) is consistent with the potential pathogenic role of Ab-producing plasmocytes in the injured MS tissue. Interestingly, in the Theiler’s virus experimental model of autoimmune demyelination, mAbs directed against the oligodendrocyte surface promote remyelination (35, 36). Hence, a pathogenic as well as reparative role for the humoral immune response could be postulated.
In conclusion, our data indicate a multifaceted cellular and humoral immune response underlying autoimmunity in MS. The comprehensive analysis of tissue-specific transcriptional programs in MS, together with the development of advanced computational algorithms to integrate the descriptive data into metabolic and regulatory circuits, will reveal the molecular fingerprint of the demyelinating process and help identify the complete array of MS disease genes.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Jorge R. Oksenberg. Department of Neurology, University of California, San Francisco. 513 Parnassus Avenue, Medical Science Building, Room S-256, San Francisco. CA 94143-0435.
3 Abbreviations used in this paper: MS, multiple sclerosis; MBP, myelin basic protein; kRT-PCR, kinetic RT-PCR; Ct, fluorescence threshold; AFI, average fold increase; MOG, anti-myelin/oligodendrocyte glycoprotein; ORO, oil red O; LFB, luxol fast blue.
Received for publication July 7, 2000. Accepted for publication August 29, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and IP-10 are expressed in demyelinating brain lesions. Proc. Natl. Acad. Sci. USA 96:6873.1, -2, and -3 isoforms and TGF- type I and type II receptors in multiple sclerosis lesions and human adult astrocyte cultures. J. Neuropathol. Exp. Neurol. 58:174.[Medline]
and interleukin-1 receptor antagonist genes with disease severity in MS. Neurology 52:595. and lymphotoxin in demyelinating disease. Ann. Rheum. Dis. 58:(Suppl. 1):I121.
This article has been cited by other articles:
|
|
 |
|
  H. P. Carroll, V. Paunovic, and M. Gadina Signalling, inflammation and arthritis: Crossed signals: the role of interleukin-15 and -18 in autoimmunity Rheumatology, September 1, 2008; 47(9): 1269 - 1277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Goertsches, S. E Baranzini, C. Morcillo, C. Nos, M. Camina, J. R Oksenberg, X. Montalban, and M. Comabella Evidence for association of chromosome 10 open reading frame (C10orf27) gene polymorphisms and multiple sclerosis Multiple Sclerosis, April 1, 2008; 14(3): 412 - 414. [Abstract] [PDF]
|
|
|
|
 |
|
 A. C. Anderson, D. E. Anderson, L. Bregoli, W. D. Hastings, N. Kassam, C. Lei, R. Chandwaskar, J. Karman, E. W. Su, M. Hirashima, et al. Promotion of Tissue Inflammation by the Immune Receptor Tim-3 Expressed on Innate Immune Cells Science, November 16, 2007; 318(5853): 1141 - 1143. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhang, C. Taveggia, C. Melendez-Vasquez, S. Einheber, C. S. Raine, J. L. Salzer, C. F. Brosnan, and G. R. John Interleukin-11 Potentiates Oligodendrocyte Survival and Maturation, and Myelin Formation. J. Neurosci., November 22, 2006; 26(47): 12174 - 12185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L Lopatinskaya, J Zwemmer, B Uitdehaag, K Lucas, C Polman, and L Nagelkerken Mediators of apoptosis Fas and FasL predict disability progression in multiple sclerosis over a period of 10 years Multiple Sclerosis, November 1, 2006; 12(6): 704 - 709. [Abstract] [PDF]
|
|
|
|
 |
|
 E. S. Smith, A. K. Li, A. M. Wang, D. H. Gelfand, and T. W. Myers Amplification of RNA: High-Temperature Reverse Transcription and DNA Amplification with a Magnesium-Activated Thermostable DNA Polymerase CSH Protocols, May 1, 2006; 2006(1): pdb.prot4115 - pdb.prot4115. [Full Text]
|
|
|
|
 |
|
 L. Klotz, M. Schmidt, T. Giese, M. Sastre, P. Knolle, T. Klockgether, and M. T. Heneka Proinflammatory Stimulation and Pioglitazone Treatment Regulate Peroxisome Proliferator-Activated Receptor {gamma} Levels in Peripheral Blood Mononuclear Cells from Healthy Controls and Multiple Sclerosis Patients J. Immunol., October 15, 2005; 175(8): 4948 - 4955. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. R. Ure, T. E. Lane, M. T. Liu, and M. Rodriguez Neutralization of chemokines RANTES and MIG increases virus antigen expression and spinal cord pathology during Theiler's virus infection Int. Immunol., May 1, 2005; 17(5): 569 - 579. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Zhao and R. D. Brinton Suppression of Proinflammatory Cytokines Interleukin-1{beta} and Tumor Necrosis Factor-{alpha} in Astrocytes by a V1 Vasopressin Receptor Agonist: A cAMP Response Element-Binding Protein-Dependent Mechanism J. Neurosci., March 3, 2004; 24(9): 2226 - 2235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Tsutsui, J. Schnermann, F. Noorbakhsh, S. Henry, V. W. Yong, B. W. Winston, K. Warren, and C. Power A1 Adenosine Receptor Upregulation and Activation Attenuates Neuroinflammation and Demyelination in a Model of Multiple Sclerosis J. Neurosci., February 11, 2004; 24(6): 1521 - 1529. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I Mendez, J Alcocer-Varela, A Parra, A Lava-Zavala, D A de la Cruz, D Alarcon-Segovia, and F Larrea Neuroendocrine dopaminergic regulation of prolactin release in systemic lupus erythematosus: a possible role of lymphocyte-derived prolactin Lupus, January 1, 2004; 13(1): 45 - 53. [Abstract] [PDF]
|
|
|
|
|
|
A. Alter, M. Duddy, S. Hebert, K. Biernacki, A. Prat, J. P. Antel, V. W. Yong, R. K. Nuttall, C. J. Pennington, D. R. Edwards, et al. Determinants of Human B Cell Migration Across Brain Endothelial Cells J. Immunol., May 1, 2003; 170(9): 4497 - 4505. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Trebst, S. M. Staugaitis, P. Kivisakk, D. Mahad, M. K. Cathcart, B. Tucky, T. Wei, M. R. S. Rani, R. Horuk, K. D. Aldape, et al. CC Chemokine Receptor 8 in the Central Nervous System Is Associated with Phagocytic Macrophages Am. J. Pathol., February 1, 2003; 162(2): 427 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Vallieres, I. L. Campbell, F. H. Gage, and P. E. Sawchenko Reduced Hippocampal Neurogenesis in Adult Transgenic Mice with Chronic Astrocytic Production of Interleukin-6 J. Neurosci., January 15, 2002; 22(2): 486 - 492. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Trebst and R. M. Ransohoff Investigating Chemokines and Chemokine Receptors in Patients With Multiple Sclerosis: Opportunities and Challenges Arch Neurol, December 1, 2001; 58(12): 1975 - 1980. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D W Montgomery Prolactin production by immune cells Lupus, October 1, 2001; 10(10): 665 - 675. [Abstract] [PDF]
|
|
|
|
 |
|
 R. L. P. Lindberg, C. J. A. De Groot, L. Montagne, P. Freitag, P. van der Valk, L. Kappos, and D. Leppert The expression profile of matrix metalloproteinases (MMPs) and their inhibitors (TIMPs) in lesions and normal appearing white matter of multiple sclerosis Brain, September 1, 2001; 124(9): 1743 - 1753. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
