| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Division of Cellular Immunology, German Cancer Research Center, Heidelberg, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To identify genes of potential importance for the described tumor resistance, the two strains, DBA/2 (D) and B10.D2 (D2), were crossed, and new recombinant inbred (RI)4 mouse strains were established over many generations of inbreeding and ESb tumor resistance selection (9). In previous studies we tested the capacity of immune cells from seven resistant lines, selected over >23 breeding generations, to transfer graft-vs-leukemia (GVL) and graft-vs-host (GVH) reactivity to ESb tumor-bearing DBA/2 mice. The results demonstrated that the various sublines differed in their capacity to transfer GVL without GVH (9, 10, 11).
We demonstrate in this study that several D2xD RI lines, when typed for mouse mammary tumor virus (MMTV) pro-viruses in their genome, showed a selective loss of one long terminal repeat (LTR) hybridization band that corresponds to Mtv-7. An open reading frame (ORF) in the LTR of Mtv-7 codes for the viral superantigen (SAG) 7 (Mls-1a), an autoantigen of DBA/2 that causes deletion of SAG-reactive T lymphocytes such as Vß6-positive T cells during thymus maturation (12, 13, 14, 15, 16, 17, 18, 19, 20, 21). Suggestive evidence for a link between tumor resistance and absence of endogenous viral SAG is given by a backcross segregation analysis as well as by reintroduction of the SAG into the resistant line and the reappearance of tumor susceptibility.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The tumor line ESb is a highly metastatic spontaneous variant of the methylcholantrene-induced DBA/2 T lymphoma L5178Y/E (22). The cells, derived from the standard batch ESb 289 or from the more aggressive liver metastasis derived subline ESb-L, were grown in suspension culture as previously described (23).
Mice and tumor challenge
DBA/2J mice were obtained from Bomholtgard (Ry, Denmark), and
B10.D2/nSn, CBA/J, and BALB/c mice were obtained from Harlan Olac
(Bicester, U.K.). For challenge, mice were injected s.c. with
105 of three times washed tumor cells in 100 µl of
PBS. While syngeneic DBA/2 mice died from development of visceral
metastases (liver, spleen) within 2 wk (22), in tumor-resistant lines a
local tumor started to grow within the first 10 days and then
regressed. Mtv-7 (Mls-1a) congenic BALB/c
(BALB/D2.Mls-1a) mice, originally bred in H. Festenstein’s
laboratory (The London Hospital Medical College, London, U.K.), were
obtained from the Ludwig Institute for Cancer Research (Lausanne,
Switzerland). The experiment shown in Figure 4
was performed on
contract at BRL (Fullingsdorf, Switzerland). The BxD RI lines were
obtained from The Jackson Laboratory (Bar Harbor, ME).
|
The breeding protocol that led to the establishment of tumor-resistant RI lines has been described (9). Breeding was performed in the specific pathogen-free animal facilities at the German Cancer Research Center.
DNA preparation and Southern blot analysis
Cell pellets were homogenized and DNA was prepared as previously described (24). Genomic DNA (10–20 µg) was digested to completion overnight with 10 U of EcoRI or PvuII restriction enzymes (Life Technologies, Grand Island, NY)/µg of DNA. Samples were precipitated, lyophilized, and run on a 0.8% Tris-acetate-EDTA agarose gel. For Southern blot analysis, the gel was washed for 15 min with 0.27 M HCl and three times for 20 min each time with 0.5 M NaOH in 1.5 M NaCl and then was blotted onto a GeneScreen nylon membrane (NEF-983, DuPont, Bad Homburg, Germany) overnight. After washing the nylon membranes for 20 min with 20x SSC/0.5 M Tris-HCl, pH 7.0, the DNA was UV cross-linked to the nylon. Membranes were prehybridized for 2 h at 42°C in a solution consisting of 50% formamide, 0.01% RNA from yeast (Boehringer Mannheim, Mannheim, Germany), 5x SSC, 0.1% Denhardt’s solution (polyvinylpyrrolidone (Serva, Heidelberg, Germany), Ficoll 400 (Sigma, Munich, Germany), BSA (Sigma)), 50 mM NaPP (Na2HPO4 and NaH2PO4), and 1% SDS and then hybridized with the same solution with the addition of 50 ng of random primed labeled MMTV LTR (19, 25) and MMTV env probes (26) for 16 to 18 h at 42°C. Filters were washed three times with 2x SSC/0.1% SDS for 30 min at 68°C and exposed for 1 to 3 days at -80°C on x-ray film Kodak X-OMAT AR (Sigma).
GVL test
Recipient DBA/2 mice were sublethally irradiated on day -1 with 5 Gy and inoculated on day 0 with 104 ESb cells i.v. Immune spleen cells (ISPL; 2 x 107) from donor mice were transfered on day 1 to test for their GVL activity. Donor mice were immunized by i.v. inoculation of 105 ESb tumor cells 7 days before removal of the spleens.
Biostatistics
The log-rank test was applied to evaluate whether the survival curves of experimental groups differed significantly. Exact tests were performed by means of the software package Stat Xact (Cytel, San Diego, CA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
New RI lines that are resistant to the DBA/2 mouse lymphoma ESb were derived from a cross between the ESb tumor-resistant strain B10.D2 (D2) and DBA/2 (D). While susceptibility was dominant in the F1 generation, segregation of susceptibility and resistance occurred in F2. Ten RI lines were eventually established after repeated colony breeding, tumor challenge, and selection. From the best breeding line (D2xD-1) we established 10 sublines by strict brother-sister mating. Sublines derived from D2xD-1 in the F17 generation differed in coat color; while D2xD-1/2 and D2xD-1/8 were dilute black, D2xD-1/9 and the other lines were dilute brown, like the parental line DBA/2. No linkage of ESb tumor resistance was seen to coat color genes (2). In the present study 3 of the 10 sublines were selected for detailed analysis and immortalized by embryo freezing.
Mtv provirus typing reveals selective loss of Mtv-7 in D2xD RI lines
Mtv provirus typing of the new tumor-resistant lines was performed
from mouse tail DNA after restriction enzyme digestion (Figs. 1 and 2).
Southern blot hybridization was performed with two different Mtv
probes, one containing the LTR region and the other containing the
envelope gene (env) of MMTV. Figure 1 shows
the Southern blots obtained when using the restriction enzyme
PvuII for DNA digestion. Different band patterns were
obtained with DNA from the two parental lines (Fig. 1, lanes
1 and 5). From their m.w. and published data
(27, 28), correlations could be made with known Mtv provirus types.
Thus, in DBA/2 mice, bands corresponding to Mtv-1, -6, -7, -8, -11,
-13, -14, and -17 could be identified from PvuII-digested
DNA. Bands corresponding to Mtv-1, -6, -7, and -13 were lacking in DNA
from the ESb tumor-resistant strain B10.D2 and from the RI strain
BxD-6, derived from a cross between C57 BL/6 and DBA/2 (BxD; Fig. 1, lane 6). When rehybridizing the washed filters with
the env probe, only five bands (instead of 10 with the LTR
probe) were seen. Interestingly, B10.D2 and BxD-6 mice, which are both
ESb tumor resistant, showed differences in LTR and env
patterns (Fig. 2). DNA from the three D2xD RI lines, 2
(lane 2), 8 (lane 3),
and 9 (lane 4), showed hybridization patterns
with Mtv LTR and env probes very similar to those of the
susceptible parental strain DBA/2. A single band, however, was lacking
when typing was performed with the LTR probe. This band corresponds to
Mtv-7, the provirus coding for the strong DBA/2 SAG Mls-1a
(12, 13, 14, 15, 16, 17, 18, 19). These results suggest that resistance to the ESb tumor in the
RI strains correlates with the absence of Mtv-7.
|
|
shows the corresponding Mtv patterns obtained with
EcoRI-digested DNA. Again, different patterns were seen with
DNA from either DBA/2 or B10.D2. There were few LTR bands in common,
while many additional bands were seen in DBA/2 that were not seen in
B10.D2. Interestingly, the LTR pattern of BxD-6 mice, which after
PvuII digestion looked very similar to that of B10.D2, were
seen to share some bands (no. 2, 8, and 10) with DBA/2 that were not
seen in B10.D2. EcoRI typing of the DNA from the new
resistant lines clearly revealed a pattern very close to that of the
susceptible parental strain DBA/2. However, all resistant lines
(lanes 2–4) were lacking one particular band
(no. 3) in the range of 14.8 kb containing Mtv-7 that was detected with
both Mtv LTR and env probes in EcoRI-digested
DBA/2 DNA. In addition, the new resistant lines had several DNA
fragments (bands 2, 9, and 10) in common with DBA/2 that are lacking in
B10.D2. Taken together, our results of Mtv typing of three independent
lines using two different probes and two different restriction enzymes
revealed a very close similarity of the new resistant lines with DBA/2
and a selective loss of Mtv-7. Backcross analysis of tumor resistance and Mtv type
To further substantiate a possible linkage between tumor
resistance and the presence or the absence of Mtv-7, we performed a
backcross analysis (Fig. 3). D2xD-1/9
mice were crossed with the parental line DBA/2, and offspring from the
F2 generation were typed by Southern blot analysis of
tail DNA for the presence or the absence of Mtv-7. They were then
challenged with ESb tumor cells to test for tumor resistance. The
survival curves of tumor-injected mice of the two parental strains
DBA/2 and D2xD-1/9 as well as those of the F1 and
F2 generation are shown in Figure 3A. While
DBA/2 and F1 mice were highly susceptible, D2xD-1/9 mice
and F2 generation mice were mostly resistant upon challenge
with ESb 289, the standard batch of ESb tumor cells. Figure 3B shows the results of an analysis in which the
F2 generation mice were grouped according to their Mtv-7
type. The Mtv-7 negative mice were significantly
(p = 0.039) more resistant than the
Mtv-7-positive mice. The reason for the reduced resistance of the
Mtv-7-negative mice compared with that of the F2 generation
mice in Figure 3A is most likely that they were challenged
with a more aggressive ESb variant (ESb-L instead of ESb 289). The
backcross and Mtv-7 segregation analysis corroborates the original
finding of a link between tumor resistance and absence of Mtv-7, but
also clearly indicates the contributions of other genes to the overall
resistance of B10.D2 mice.
|
If the genetic basis for tumor resistance in B10.D2, BxD-6, and
the new D2xD lines is different, a cross between these strains might
lead to genetic complementation. Different crosses between the
resistant lines were therefore generated and tested for ESb tumor
resistance. No complementation and regeneration of tumor susceptibility
was seen in any of the crosses between these mouse strains (data not
shown). Additional crosses were performed between the resistant line
BxD-6 and either CBA/J mice, which express Mtv-7 or BALB/c mice which
do not express Mtv-7. The cross with CBA/J, but not that with BALB/c,
led to genetic complementation so that tumor resistance was reduced
from 100 to 10% (Table I).
|
).
This line is an Mls-1 congenic mouse line to BALB/c
(Mls-1b). The Mls-1a (Mtv-7) gene was derived
from DBA/2. When the tumor-resistant line D2xD-1/9 was crossed with
this Mtv-7-positive line, a partial complementation effect was
observed. Upon challenge with ESb 289, 11 of 20 (60%) of the
F1 mice developed tumor compared with 1 of 6 (17%) of the
D2xD-1/9 mice and all six (100%) of the DBA/2 mice. Three of the 11
tumor-bearing mice were eventually able to reject the tumor, showing
that complementation by Mtv-7 was only partial. This is also obvious
from the fact that 9 of the 20 challenged mice did not develop tumors
(Fig. 4A).
Further analysis of the F1 mice was performed to test for a
possible basis for the heterogeneity seen in tumor susceptibility.
Response heterogeneity was not due to sex influence. All mice were
Mtv-7 typed and found to be positive as expected (Fig. 4B,
group I).
In Table I we have summarized the main results about Mtv segregation in
the newly established RI lines and from the backcross and genetic
complementation analysis and compared them to the segregation in
already established BxD RI lines. It can be seen that in the already
established BxD RI lines from The Jackson Laboratory the six different
Mtv types of DBA/2 shown are distributed differently. None of the 10
BxD RI lines tested (only four shown) was like the newly selected
D2xD lines, which differ from DBA/2 only by a selective loss of
Mtv-7.
Immune cells from tumor-resistant D2xD mice transfer GVL reactivity to tumor-susceptible D mice
Mtv typing revealed that the new immunoresistant D2xD RI strains
were more similar to the parental strain D than to D2. Therefore, we
were interested to test whether antitumor immune cells from D2xD RI
strains could still transfer, like D2, GVL reactivity into the
tumor-bearing parental strain D. We chose a test system in which DBA/2
recipients were first sublethally irradiated with 5 Gy and then
injected with 104 ESb tumor cells i.v. One day later,
different groups of such mice (10 animals each) were treated i.v. by
ISPL from either B10.D2 or D2xD-1/9 mice that had been preimmunized by
i.v. inoculation of 105 ESb tumor cells 1 wk previously. As
shown by the survival curves in Figure 5,
nontreated control mice died within 10 days, while immune cells from
D2xD mice were capable of transferring GVL reactivity, leading to long
term (>100 days) survival. Adoptive immunotherapy with ISPL from
B10.D2 mice also led to effective GVL reactivity, but some mice were
affected later by chronic and fatal GVH disease (9, 10, 11). Adoptive
immunotherapy with syngeneic anti-ESb immune peritoneal effector
cells (Fig. 5, DBA/2 PEC) was less effective, with 50% of the mice
dead after 30 days.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MMTVs exist either as exogenous infectious virus or as endogenous
provirus (Mtv). They are lymphotropic and depend on SAG-mediated T cell
activation in their live cycle (12, 21). Mtv SAGs are type 2
transmembrane glycoproteins encoded by an ORF of the 3' LTR of the
provirus and associate with class II MHC molecules. B cells transfected
with a cDNA construct containing the LTR ORF from Mtv-7 stimulate the
proliferation of Mls-1a-responsive T cells expressing TCR Vß6 (B.
Huber, personal communication). A polymorphic region in the carboxyl
terminus of Mtv SAGs determines its Vß domain specificity (13, 14).
About 28 endogenous Mtv loci are found in all the common mouse strains
(28). Since most mouse strains contain about 20 different Vß TCR
chains, each Mtv-SAG shows reactivity with 3 to 30% of all 
,ß CD4
T cells, which often belong to one of four different Vß
subpopulations (29). This extraordinarily high frequency of
SAG-responsive T cells in nonexpressing mice is 10- to 50-fold higher
than the response to allogeneic MHC class II molecules and 500- to
2000-fold higher than the response to conventional protein Ags. All of
the 19 or more tested laboratory mouse strains express one to eight
endogenous Mtvs (28, 29, 30, 31), all of which can affect the TCR repertoire.
As self proteins, Mtv SAGs alter the peripheral T cell repertoire by
intrathymic deletion of Vß-reactive T cells. Since without such
reactive T cells exogenous homologous MMTV cannot propagate in these
mice, Mtv SAGs can be considered as determinants of resistance to
homologous MMTV. While the presence of germline proviruses may reflect
an evolutionary mechanism of protection against horizontal transmission
of MMTV, it also may carry a potential risk of increasing
susceptibility to unrelated infectious agents and diseases.
It has been shown that polyoma tumor susceptibility in C3H mice is due to the presence of Mtv-7 SAG. This abrogates Vß6 T lymphocytes, which play a role in polyoma tumor immunosurveillance (32). In another system, a novel Mtv SAG contributes to development of spontaneous B cell lymphomas in SJL mice. In this case, the Mtv SAG stimulates CD4 Vß-16 T cells to release cytokines required for tumor growth (33). Thus, in the two tumor systems of C3H and SJL mice, endogenous Mtv SAGs are required for tumor development, but the strategies are different, involving stimulation of CD4 T cells in one case and deletion of CD8 T cells in the other. MMTV can influence tumors in many ways, either directly, as in mammary carcinogenesis, or indirectly, as in many other tumor systems. Mtv SAGs can influence the TCR repertoire in vivo by deletion or stimulate peripheral T cells expressing certain vß chains. The genetic event (rearrangement of Mtv provirus genes, possibly through recombination and chromosome exchange) that led to the selective loss of Mtv-7 in this study is presently unknown. It is also not known when this event happened and whether it happened once or several times during the selection of the tumor-resistant lines. It is likely that the repeated ESb tumor challenge and resistance selection that preceded this analysis played a decisive role in this result. ESb cells have recently been shown to express Mtv-7, and Mtv-7 SAG-reactive CTLs could also be recently identified (our unpublished observations).
ESb tumors are highly aggressive lymphomas of DBA/2 mice, but they
cannot grow in the new Mtv-7-negative RI strains derived from a cross
with B10.D2. This is because the tumor cells are inducing strong host
immune responses. We demonstrate that immune spleen cells from D2xD
mice can transfer immunoresistance and GVL reactivity back to the
tumor-susceptible mice. The GVL reactivity was stronger than that of
syngeneic anti-ESb-immune peritoneal cells (Fig. 5).
When Mtv-7 was reintroduced by crossing D2xD mice with BALB/D2
(Mls-1a), the mice became more tumor susceptible (Fig. 4),
while this was not the case in a cross with BALB/c (Mls-1b;
Table I). Similarly, ESb tumor resistance of BxD-6 (Mtv7-negative) RI
mice, established at The Jackson Laboratory from a cross of C57 Bl/6
and DBA/2 mice, could be reduced from 100 to 10% by reintroduction of
Mtv-7 via a cross with CBA/J mice (Table I). In the previously
established BxD RI lines that were not selected for ESb tumor
resistance, there was an apparent random segregation of Mtv types (9)
(Table I), while in the new D2xD RI lines ESb tumor resistance
selection seems to have led to a selective absence of Mtv-7.
When the resistant mice were backcrossed to DBA/2, segregation of Mtv-7
in the F2 generation correlated with tumor susceptibility
or resistance. Tumor resistance in these Mtv-7 backcross mice was not
absolute, however, and appeared to depend on the malignancy of the
tumor line used for challenge. Protection was less efficient against
the highly malignant liver metastasis-derived subline ESb-L than to the
standard line ESb-289 (Fig. 3). Also, Mtv-7 reintroduced tumor
susceptibility only partially (Fig. 4). These results suggest that
Mtv-7 is an important gene, but possibly not the only one contributing
to the tumor resistance of B10.D2 mice against DBA/2 tumors. There may
be genes in close proximity to Mtv-7, H-1, H-4, and/or Hbb (2) that
also could contribute to B10.D2 ESb tumor resistance.
With regard to the test crosses of the resistant RI lines with BALB/c (which did not increase tumor susceptibility) and BALB/D2.Mls-1a (which increased tumor susceptibility in about 50% of the mice), the introduction of additional minor Ags and the phenomenon of hybrid resistance could have influenced the results. We previously found that ESb cells that are strongly rejected by B10.D2 are only rejected by about 80% of injected BALB/c mice. Furthermore, F1 mice from crosses of DBA/2 with BALB/c or B10.D2 mice survived longer after ESb challenge than DBA/2 mice, although they all were tumor susceptible and died eventually (34).
Susceptibility to chemically induced lymphomas of H-2d
mouse strains described here correlated with the presence of Mtv-7,
just as susceptibility to polyoma-virus induced tumors of
H2k mouse strains correlated with Mtv-7. In both systems, T
cell-mediated antitumor immunity seems to play a predominant role, but
there appear to be differences with respect to the target Ags
recognized. A strongly biased usage of Vß6 by polyoma virus-specific
CD8 CTL was considered critical for anti-polyoma tumor effector
cells in vivo. Deletion of Vß6 T cells by Mtv-7 would thereby reduce
the T cell repertoire of antitumor effector cells. In the ESb lymphoma
model analyzed here, the mechanism seems to be different. DBA/2-derived
ESb tumor-specific CTLs were previously characterized as expressing
Vß5 and Vß8.1 preferentially, but not Vß6 (23). We recently found
that ESb tumor cells do not only express individually distinct class I
MHC (Kd)-restricted CTL epitopes (23) but in addition
proviral Mtv-7 at the RNA-level (unpublished observations). Upon
transfer into B10.D2 or D2xD resistant mice that do not express Mtv-7
(Table I), Vß6 and possibly other T cells may become activated and,
upon transfer into tumor-bearing DBA/2, may infiltrate ESb liver
metastases. It seems from preliminary findings that in the ESb tumor
model the tumor cells themselves express endogenous viral SAGs that can
be recognized in tumor-resistant, but not in syngeneic, mice by Vß6 T
cells.
Much further work is required to find out how SAG-reactive T cells might contribute to augment antitumor reactivity and how this can be exploited for adoptive immunotherapy. The contribution seems to be very effective, because we demonstrated that a single transfer of immune cells from tumor resistant to tumor-susceptible mice can cause complete tumor regression even in late stage advanced disease with macroscopic metastases in multiple organs (34). A prerequisite for this effectivity was a sublethal host irradiation by 5 Gy to prevent host-vs-graft reactivity. Without this, the GVL reactivity was much less pronounced (35). The therapy effects could be evaluated in individual tumor-bearing mice both quantitatively (36) and qualitatively (37) by new noninvasive methods such as 31P nuclear magnetic resonance spectroscopy (36) and 1H nuclear magnetic resonance microimaging (37). Following immune cell transfer, donor immune T cells were found to infiltrate liver metastases (36, 38) and to form clusters with a subset of host macrophages bearing the adhesion molecule sialoadhesin (38). These host macrophages possibly function as scavengers and APCs. The elucidation of basic mechanisms of this immunoresistance and cellular immunotherapy at the molecular and cellular levels may help to design new strategies for introducing effective GVL reactivity into the clinic. It will be most interesting to determine whether humans express SAGs that are encoded by endogenous retroviral genes. The observation of significant homology between a simian herpes virus gene and the MMTV CTR ORF (39) and of an EBV-associated SAG (40) may be important in this regard.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: University Hospital, Department of Surgery, Unit for Experimental Transplantation-Immunology, Würzburg, Germany.
3 Current address: Research and Development, Cancer Immunology, SmithKline Beecham Biologicals, rue de l’Institut 89, B-1330 Rixensart, Belgium.
4 Abbreviations used in this paper: RI, recombinant inbred; GVL, graft-vs-leukemia; GVH, graft-vs-host; MMTV, mouse mammary tumor virus; LTR, long terminal repeat; ORF, open reading frame; SAG, superantigen; ISPL, immune spleen cells.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
interferon induction and cytotoxic response. Invasion Metastasis 1:175.[Medline]
-Irradiation suppresses T-cell mediated protective immunity against a metastatic tumor in the afferent phase of the immune response but enhances it in the efferent phase if given before immune cell transfer. Int. J. Oncol. 4:335.
-interferon production in primary murine MLC. J. Interferon Res. 1:339.[Medline]
but fails to trigger cytotoxicity. Eur. J. Immunol. 22:2789.[Medline]
This article has been cited by other articles:
|
|
 |
|
  A. Facchinetti, S. D. Santa, S. Mezzalira, A. Rosato, and G. Biasi A Large Number of T Lymphocytes Recognize Moloney-Murine Leukemia Virus-Induced Antigens, but a Few Mediate Long-Lasting Tumor Immunosurveillance J. Immunol., May 1, 2005; 174(9): 5398 - 5406. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Müerköster, J. D. Laman, M. Rocha, V. Umansky, and V. Schirrmacher Functional and in Situ Evidence for Nitric Oxide Production Driven by CD40-CD40L Interactions in Graft-versus-Leukemia Reactivity Clin. Cancer Res., May 1, 2000; 6(5): 1988 - 1996. [Abstract] [Full Text]
|
|
|
|
 |
|
 S. Muerkoster, M. Rocha, P. R. Crocker, V. Schirrmacher, and V. Umansky Sialoadhesin-Positive Host Macrophages Play an Essential Role in Graft-Versus-Leukemia Reactivity in Mice Blood, June 15, 1999; 93(12): 4375 - 4386. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
