| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Intramural Research Support Program, SAIC Frederick and Laboratory of Molecular Immunoregulation, Division of Basic Sciences, NCI-FCRDC, Frederick, Maryland;
Institute of Toxicology, ETH, Schwerzenbach, Switzerland;
Laboratory of Molecular Immunology, Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, and Belozersky Institute of Physico-Chemical Biology, Moscow State University, Russia; and
Ludwig Institute for Cancer Research, New York Branch at Sloan-Kettering Memorial Cancer Center
Correspondence: Alexander N. Shakhov, IRSP, SAIC Frederick, Bldg. 560, Rm. 31-33, NCI-FCRDC, Frederick, MD 21702. E-mail: shakhova{at}mail.ncifcrf.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(LT) and tumor necrosis
factor (TNF) deficiencies show defects in the structure of peripheral
lymphoid organs such as spleen, lymph nodes, and gut-associated
lymphoid tissues. To identify genes associated with this defective
phenotype in spleen, we applied a gene profiling approach, including
subtractive cloning and gene array hybridizations, to mice with
combined TNF/LT deficiency. The differentially expressed genes
identified by these techniques was then evaluated by Northern blot
analysis for splenic expression in knockout mice with single LT or
single TNF deficiency. Most of the genes detected in this analysis are
directly or indirectly associated with disrupted LT and not TNF
signaling.
Key Words: gene arrays • subtractive cloning • knockout mice • spleen
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(LT) and combined tumor necrosis
factor/lymphotoxin (TNF/LT) deficiencies result in a complex phenotype
with several tissues and histological compartments affected
[1
2
3
4
5
].
The main features of the splenic deficiencies (disrupted
microarchitecture, lack of FDC networks, and defective immune response)
are plastic: i.e., they can be induced in wild-type mice within days
and hours after injection of LTßR antagonist (LTßR-Ig [see ref. 6
]) or by reciprocal bone marrow transfers. It is conceivable that the
genes associated with phenotypic defects are actively expressed in
spleens of adult wild-type mice (as opposed to temporary expression in
embryogenesis) and are quickly shut off during treatment with LTßR
antagonists. Such genes may be under direct control of LT/LTß
signaling through LTßR or may reflect altered migration or
mobilization of certain cell types into splenic compartments.
Presumably, alterations in the expression of such genes between
wild-type and mutant mice is at least partly occurring at the level of
transcription. Before starting this analysis for spleens of LT- or
TNF-deficient mice we expected to detect differences in the expression
of cytokines/chemokines or their receptors, cell adhesion molecules,
and intracellular signaling molecules.
Gene profiling is an approach to uncover differences between two RNA subsets. In this study we used a combination of several techniques, including hybridizations to gene arrays, subtractive cloning, and Northern blot analysis to identify genes whose expression is substantially different in spleens of TNF/LT-deficient mice. Although the analysis may still be incomplete, we report here on the first 30 known genes that show 2- to 10-fold difference at mRNA level and that have been verified on single TNF- and LT-deficient mice, and could be therefore tentatively referred either to TNFR or LTßR signaling or both.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
knockout (K/O) mice, TNF K/O mice, and mice with combined
TNF/LT deficiency were described previously [1
,
3
, 7
, 8
]. TNF and TNF/LT
double K/O mice were on mixed C57BL/6-sv129 background. LT K/O mice
[1
] were purchased from Jackson Laboratory (Bar Harbor,
ME) and back-crossed six times to C57BL/6 background. All mouse strains
were maintained under specific pathogen-free conditions.
RNA preparation
Spleens were frozen in liquid nitrogen and then ground with a
pestle in a pre-cooled mortar to obtain powdered tissue homogenate.
Total cellular RNA was extracted with TrizolTM reagent (GIBCO-BRL,
Gaithersburg, MD) and used for Northern analysis. Poly A+
RNA for cDNA synthesis was prepared by using mRNA Separator from
CLONTECH Laboratories, Inc. (Palo Alto, CA), following the
manufacturer’s protocol.
Gene arrays
Two types of cDNA arrays were used in the study. CLONTECH
ATLASTM containing 588 murine genes was used for hybridization with
radioactively labeled first-strand cDNA prepared using total poly
A+ mRNA isolated from five to six spleens of naive
wild-type and mutant mice according to the manufacturer’s protocol.
InCyte (Genome System) gene microchips (Mouse GEM Microarray) containing 10,000 murine genes and ESTs were used as a custom service. Splenic polyA+ mRNA (0.6 µg) from both wild-type and K/O mice were supplied for custom analysis. After hybridization, the data were posted at the company web site and the results were viewed using the software provided.
Subtractive cloning
A Clontech PCR-SelectTM cDNA Subtraction kit (Palo Alto, CA) was
applied to splenic mRNA from naive wild-type (WT) and TNF/LT
double-deficient mice [3
]. Hybridization was performed
in three separate tubes for 18, 32, and 72 h, and then these three
time points were combined. After cloning into the BlueScript vector
(Xma III-Eag I), clones were pre-screened by colony hybridization with
gel-purified products of the second polymerase chain reaction (PCR),
and only clones that appeared differentially positive in this test were
further studied.
Northern blots
Ten micrograms of total or one microgram of poly A+
mRNA was separated on 1.5% denaturing agarose gel and transferred to
Supported Nitrocellulose-1 (GIBCO-BRL) membrane. Hybridization with
32P-labeled probes was performed in ExpessHybTM solution
(Clontech) and washed following the recommended protocol. Radioactivity
was quantified using Molecular Dynamics screens and ImageQuant
software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The main advantage of subtractive cloning is that this technique is unbiased and does not depend on a pre-selected set of genes. On the negative side, this technique is time-consuming and even with several modifications still produces many false-positive clones that can only be verified by quantitative Northern analysis. Another disadvantage is the variable frequency of different genes in the library that reflects mRNA abundance, in spite of the fact that the PCR-based subtraction was expected to produce a normalized library.
None of the techniques integrates all advantages and is free of limitations. For these reasons, appreciation of a combination of the different methods to the same system shows greater promise in identification of physiologically relevant differences in gene expression.
Genes identified by subtractive cloning
Initially, subtractive cDNA cloning was applied to poly
A+ mRNA preparations from spleens of naive mice with
combined TNF/LT deficiency [3
] and wild-type
controls. Several known differentially expressed genes are listed in
Table 1
, and the corresponding Northern blot analysis is shown in Figure 1
. Genes with substantially lower expression in mutant mice included
previously reported lymphoid-specific group IIG phospholipase
A2, called SPLASH [9
, 10
,
11
], a lymphoid tissue chemokine SLC, lactotransferrin,
myeloperoxidase, IgA heavy chain, receptor expressed on macrophages of
marginal zone (MARCO) [12
], endothelial cell-specific
molecules such as vascular endothelial zinc finger 1 (Vezf1), and milk
fat globule membrane protein E8 (MFG-E8), a set of pancreas-specific
genes, and several others (see Table 1
). As a control, reverse
subtraction identified the neo gene used in the targeting
construct and that therefore should have appeared as K/O-specific. The
modifications that we have introduced (see Materials and Methods) seems
to have increased the number of truly differential clones (unpublished
observation). However, the frequency of false-positive clones still
remained at about 50%, underscoring the necessity of verification of
each and every clone by Northern analysis, and making it difficult to
draw any conclusion about genes expressed at the level beyond the
sensitivity of Northern analysis.
|
|
-deficient mice compared with wild-type controls).
|
Genes identified by hybridizations to cDNA gene microchip from
InCyte (Mouse GEM Microarray)
This type of array has an important advantage because it
contains a much larger number of immobilized probes, including many
ESTs corresponding to as yet unknown genes. This technique when applied
to spleens of wild-type mice and mice with combined TNF/LT deficiencies
allowed us to identify additional differentially expressed genes. Seven
clones out of the first 13 with the biggest differences, ranging from
10- to 2-fold, showed true differences after Northern blot analysis.
These included transcripts from IgA heavy chain gene (10-fold
difference, see segments G7 on plate 021S4908 and C5 on 021B3432 in
Fig. 3A
), two other lymphoid tissue chemokines, BLC and ELC (Fig. 3A , segments D4
on plate 021S4908 and F4 on plate 021O3443; 6.4- and 3.5-fold,
respectively) complement factor B precursor (C3/C5 convertase; see
segment F7 on plate 02193446, Fig. 3A
), phosphodiesterase I/nucleotide
pyrophosphatase (Npps2; segment G2 on plate 021B3432, Fig. 3A
), and an
unknown gene operationally called clone no. 3, corresponded to a novel
gene sharing homology with several cell adhesion molecules, in
particular with MadCAM-1, and possessed characteristic mucin and Ig
domains (segment B10 on plate 021L4907 in Fig. 3A
). Differences in
splenic expression for all these genes between wild-type and mutant
mice were confirmed by Northern blot analysis (Fig. 3B)
. It was
surprising to note that neither BST-1 nor clusterin were identified as
differentially expressed genes by this type of array, even though the
Northern blot analysis clearly confirmed approximately twofold
difference in each case. On the other hand, some of the
apparently differentially expressed clones identified by GEM microarray
were not confirmed by Northern blot analysis (the second 10 clones with
differences between 2.1- and 1.8-fold, data not shown). These findings
further support the conclusion that only a combination of techniques
may provide full and reliable gene profiling.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
deficiencies. This result is consistent with
the recent findings by Ngo et al. [15
]. Splenic
expression of BLC seems to be reduced in both TNF- and LT-deficient
mice, indicating that normal expression is dependent on both TNF and
LT signaling. Additional analysis has shown that down-regulation of
chemokine transcription is much more pronounced in spleens of naive and
immunized LT-/- mice compared with TNF or
TNFRp55-deficient mice, even though the splenic microarchitecture in
the two latter cases is also abnormal [Shakhov et al., unpublished
results]. In spleen, B and T cells predominantly migrate in response
to BLC and SLC, respectively. Therefore, deficient levels and
dislocated positioning of these chemotactic signals may explain
abnormal localization of lymphocytes in spleens of LT-deficient mice.
Because ELC is produced by dendritic cells, its lower splenic
expression may be related and may be secondary to the reported
deficiency in dendritic cells in LT-deficient mice
[16
]. DC migrate in response to ELC and SLC that are
both produced by stromal or stroma-associated cells in spleen. Why stromal cells do not produce these chemokines is not clear, but recent studies suggest that they may need signals coming from B cells, and mediated by LTßR and TNFRp55 [see ref. 17 ]. It would be tempting to speculate that in such stromal cells chemokine genes are under direct transcriptional control by the signals coming from LTßR (and to lesser extent from TNFRp55), however, this has to be experimentally tested in the future.
Probes for many other chemokine and chemokine receptor genes were present on the GEM microarray, but failed to provide differential signal when wild-type and mutant mice were compared, suggesting that only these lymphoid tissue chemokines are affected. However, this result should be interpreted with caution, because we came across genes whose expression is differential by Northern blot analysis, but this difference could not be confirmed by GEM microarrays.
SPLASH, a new lymphoid-specific sPLA2 homolog
This gene product was initially identified in 1997 by subtractive
cloning using mice with combined TNF/LT deficiency, and was later
shown to be expressed by splenic stromal or stroma-associated cells and
not by lymphocytes [11
]. Although PLA2s have
been originally implicated in TNF signaling, the expression of this
particular sPLA2 is linked to LT signaling because both TNF
and TNFRp55-deficient mice show normal levels of SPLASH in spleen
[11
]. Although we cannot rule out direct effects of
LT/LTßR signaling on SPLASH expression, our preferred hypothesis is
that cells expressing SPLASH do not efficiently migrate to spleen (like
dendritic cells [16
]). Therefore, the deficiency in
SPLASH expression may be secondary to the disrupted or dislocated
expression of other gene products, such as chemokines. Alternatively,
in mutant mice cells may not mature into SPLASH-producing cells from
their precursors.
Neutrophil products
At least three neutrophil products have been identified as
expressed at decreased levels in spleens from mutant mice:
myeloperoxidase, lactotransferrin, and chemotactic protein CP-10
[18
] (Table 1) . Our findings suggest that recruitment of
neutrophils into lymphoid organs may be deficient in
LT-/- mice, consistent with recent data obtained on mice
with combined TNF/LT deficiency [19
]. Deficiency in
myeloperoxidase in both LT and TNF single-deficient mice was also
confirmed by measurements of enzymatic activity [A. Shakhov and M.
Marino, unpublished observation]. Myeloperoxidase expression was more
strongly reduced in spleens from mice with combined TNF/LT deficiency
compared with single LT deficiency (data not shown), suggesting that
for normal neutrophil production (or recruitment) both TNF and LT
signaling pathways are required. Based on this hint from gene profiling
we are currently analyzing the status of neutrophil-mediated host
defense functions of the LT-deficient mice.
Cell adhesion molecules
Two gene products identified by subtractive cloning and gene
arrays fall into a broadly defined family of cell adhesion molecules.
The first is milk fat globule membrane protein E8 (MFG-E8), also known
as lactadherin, reported to be expressed by dendritic cell-derived
exosomes [20
] and on high endothelial venules of
lymphatic tissues. The second (clone no. 3, Table 1
) is a novel gene
with 48% identity (68% homology) to rat kidney injury molecule-1
(KIM-1) [21
] whose function is unknown. We are currently
raising antibodies against the latter molecule in order to determine
tissue localization of the protein-producing cells by
immunohistochemical analysis.
Other membrane-associated molecules
Comparative Northern blot analysis with splenic RNA from mice with
combined and single TNF and LT deficiencies indicated that scavenger
receptor MARCO and CRP-ductin, another protein with scavenger motif,
showed decreased splenic expression in combined and single LT
deficiency, but not in TNF-/- mice [Shakhov et al.,
unpublished results]. Normal MARCO expression in TNF-deficient mice is
in good correlation with the fact that marginal zone macrophages,
producers of MARCO [12
], are normal in these mice
compared with LT-/- mice, where these cells are absent
(data not shown). Another membrane-associated molecule, BST-1, is a
close homolog of CD38 and is produced by bone marrow stromal cells to
facilitate pre-B cell growth [13
]. Our data suggest that
the reduced BST-1 expression in LT-/- (but not in
TNF-/-) mice may be related to reduced production of
certain Ig isotypes by B cells [22
]. Indirectly these
findings are in agreement with published observations that oral
immunization of BST-1-/- mice with thymus-dependent
antigens resulted in poor production of Ag-specific antibodies by the
intestinal mucosa accompanied by the reduced number of Ag-specific
IgA-producing cells in the lamina propria [23
]. These
results suggest that BST-1 has a role in B cell development and Ab
production in vivo.
Pancreas-specific genes
Subtractive cloning unexpectedly revealed several gene products
that are associated with the function of the pancreas. These included
-amylase, elastase, preprotrypsin, precarboxypeptidase, triglyceride
lipase, and carboxyl ester lipase (Table 1)
. The mRNA for these genes
can be detected on Northern blots with pooled RNA prepared from spleens
of normal and (at lower levels) mutant mice. Initially we thought that
our spleen preparations might be contaminated with pancreatic material
because the splenic artery is known to run through the pancreas in
mice. However, we did not observe any difference in the level of
expression of the above-mentioned genes when pancreas from wild-type
and mutant mice were compared, arguing against such contamination.
Our data may be interpreted as suggestive of active tissue remodeling processes occurring in spleens of normal mice that require the activity of remodeling enzymes. For unknown reasons, in mutant mice these processes are less active. Additional immunohistochemical studies on splenic sections will be performed to evaluate this possibility. Expression of trypsin in spleens of normal mice has been previously reported [24 ].
Most identified genes are specifically associated with LT, and not
with TNF deficiency
For many of the genes that have been identified by one of the
techniques the comparison of the expression levels in spleens from
naive wild-type mice or mice with single LT, single TNF, or combined
TNF/LT deficiencies was performed. The majority of the differentially
expressed genes described above turned out to be specifically
associated with LT deficiency, i.e., the decreased expression levels
compared with controls were detected in spleens of mice with LT or
combined TNF/LT deficiency, and not with single TNF deficiency. This
observation is in agreement with the fact that alterations in spleens
of TNF-deficient mice, reflected in the pattern of constitutively
expressed genes, are far less prominent compared with LT-deficient
mice. Indeed, even though both LT- and TNF-deficient mice are devoid
of FDC clusters and do not develop germinal centers, TNF-deficient mice
retain segregated T and B cell zones. The splenic marginal zone is
present in TNF-deficient but not in LT-deficient mice.
In conclusion, gene profiling appears to be a promising approach for characterization of alterations occurring in distinct histological compartments of knockout mice at the molecular level. Unbiased approaches, such as subtractive cloning, appear to have an advantage as a gene discovery tool. However, in the near future this advantage may be matched with introduction of the next-generation gene arrays that will contain dozens or even hundreds of thousands of murine genes, including orphan ESTs.
|
ACKNOWLEDGEMENTS |
|---|
The contents of this publication do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The publisher or recipient acknowledges right of the U.S. Government to retain a nonexclusive, royalty-free license in and to any copyright covering the article. Animal care was provided in accordance with the procedures outlined in the Guide for the Care and Use of Laboratory Animals [NIH Publication No. 86-23, 1985].
This article reports data presented at the 8th Conference of the International Cytokine Society, Hilton Head, SC, December 5–10, 1999.
Received February 7, 2000; revised February 24, 2000; accepted February 25, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Heikenwalder, M. Prinz, N. Zeller, K. S. Lang, T. Junt, S. Rossi, A. Tumanov, H. Schmidt, J. Priller, L. Flatz, et al. Overexpression of Lymphotoxin in T Cells Induces Fulminant Thymic Involution Am. J. Pathol., June 1, 2008; 172(6): 1555 - 1570. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Huber, C. Thielen, H. Seeger, P. Schwarz, F. Montrasio, M. R. Wilson, E. Heinen, Y.-X. Fu, G. Miele, and A. Aguzzi Lymphotoxin-{beta} Receptor-Dependent Genes in Lymph Node and Follicular Dendritic Cell Transcriptomes J. Immunol., May 1, 2005; 174(9): 5526 - 5536. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Banno, A. Gazel, and M. Blumenberg Effects of Tumor Necrosis Factor-{alpha} (TNF{alpha}) in Epidermal Keratinocytes Revealed Using Global Transcriptional Profiling J. Biol. Chem., July 30, 2004; 279(31): 32633 - 32642. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Abe, F. O. Yarovinsky, T. Murakami, A. N. Shakhov, A. V. Tumanov, D. Ito, L. N. Drutskaya, K. Pfeffer, D. V. Kuprash, K. L. Komschlies, et al. Distinct contributions of TNF and LT cytokines to the development of dendritic cells in vitro and their recruitment in vivo Blood, February 15, 2003; 101(4): 1477 - 1483. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V. Kuprash, M. B. Alimzhanov, A. V. Tumanov, S. I. Grivennikov, A. N. Shakhov, L. N. Drutskaya, M. W. Marino, R. L. Turetskaya, A. O. Anderson, K. Rajewsky, et al. Redundancy in Tumor Necrosis Factor (TNF) and Lymphotoxin (LT) Signaling In Vivo: Mice with Inactivation of the Entire TNF/LT Locus versus Single-Knockout Mice Mol. Cell. Biol., December 15, 2002; 22(24): 8626 - 8634. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Hjelmström Lymphoid neogenesis: de novo formation of lymphoid tissue in chronic inflammation through expression of homing chemokines J. Leukoc. Biol., March 1, 2001; 69(3): 331 - 339. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
