DNA AND CELL BIOLOGY Volume 11, Number 8, 1992 Mary Ann Lieber), Inc., Publishers
Pp. 587-592
Long-Term Consequences of Interleukin-6 Overexpression in Transgenic Mice CATHERINE WOODROOFE, WERNER
MÜLLER,* and ULRICH RÜTHER
ABSTRACT With the aim of using interleukin-6 (IL-ó)-inducible promoters to express transgenes, we investigated the long-term consequences of high levels of IL-6 in mice. As a first step, we generated transgenic mice constitutively expressing the murine IL-6 at a level sufficient to induce IL-6-responsive genes. These mice were analyzed with respect to the indirect and direct consequences of elevated IL-6 expression over a time period of about 2 years. Although biologically active IL-6 was expressed from the transgene and different alterations could be documented (less immature B cells in bone marrow, expression of IL-6-inducible liver genes), the mice appeared healthy and could easily be used for breeding. Only in mice older than 18 months did we find a high incidence of lymphomas associated with different tissues. These results indicate that the side effects of long-term treatment with IL-6 are relatively moderate, and that IL-6 might be used to mediate the expression of heterologous genes in the context of functional studies.
INTRODUCTION useful system study Today, regulation, development, and tumorigenesis. sometimes certain be to
transgenic animals are a
gene
However,
aspects
can
only
analyzed
when the gene of interest is expressed from a tight but inducible promoter, e.g., in functional studies or when the overexpression of a product is growth inhibitory. Therefore, different inducible promoters, such as the mouse mammary tumor virus long terminal repeat and the mouse and the human metallothionein promoters (for review, see Palmiter and Brinster, 1986), have been tested over the past 5 years. Although these promoters are inducible up to 200-fold, they exhibit a basal level of expression prior to induction. In particular experiments, a constitutive expression of certain gene products is not acceptable, e.g., when the analysis of the specificity of the transgene activity is the aim of the experiment. Therefore, in addition, to the above-mentioned promoters, the interferon-(IFN)-inducible Mx promoter has been characterized (Arnheiter et al., 1990). The advantage of the Mx promoter is its very low constitutive expression and its very broad and high expression in almost every tissue upon induction (Arnheiter et ai, 1990; Bachiller and Rüther,
1990).
Despite the good transient inducibility of these promoters, the real problem becomes obvious in the context of long-term induction and by the induction through endogenous products. The metallothionein promoters, for example, can only be induced efficiently by the application of heavy metals, which do induce a series of other genes including stress genes such as heat shock proteins. On the other hand, there is a strong induction of the metallothionein genes by endogenous factors in mouse development that could make it impossible to use them. Problems might also be associated with the IFN-inducible Mx promoter. It has been shown that elevated expression of IFN causes male sterility (Iwakura et ai, 1988). On the basis of all these observations, we were interested to find an inducible promoter system in which the genes are tightly regulated and the effects of long-term induction are reduced to a minimum. Recently, we have found that the human C-reactive protein gene (CRP) does fulfill the first criteria in transgenic mice (Ciliberto et ai, 1987).
Consequently, we are investigating the long-term problems of induction. Since interleukin-6 (IL-6) is the direct mediator of CRP induction (Ganter et ai, 1989), we have generated transgenic mice that overproduce IL-6. These mice were used as a means of evaluating the consequences of constitutive IL-6 expression.
European Molecular Biology Laboratory, 6900 Heidelberg, Germany. •Institut für Genetik, Universität zu Köln, 5000 Köln 41, Germany. 587
588
WOODROOFE ET AL.
MATERIALS AND METHODS IL-6 construct and
microinjection
A 695-bp cDNA fragment of the murine IL-6 gene was isolated from pEP.B-IL-6 (Blankenstein et ai, 1991) and was then inserted into the Eco Rl-Nhe I sites of plasmid pi63/7 (see Fig. 1). This construct should enhance the expression of the cDNA due to the presence of intron 2 of ßglobin. For microinjection, the Hind lll-Sst I fragment was gel-purified. DNA was microinjected into fertilized eggs of a cross between F, (C57BL/6 x SJL).
Expression and flow cytometry analysis RNA isolation and Northern analysis was performed as described (Rüther et ai, 1987). As probes, we have used the IL-6 cDNA (Van Snick et ai, 1988), the alpha-1-acid
cDNA (Ricca et ai, 1981) and the mouse SAA2 cDNA (kindly provided by Hugh Rienhoff). Human recombinant IL-6 was a gift of Gennaro Ciliberto and mouse recombinant IL-6 was purchased from UBI. For the quantification of serum IL-6, the ENDOGEN murine IL-6 ELISA was used. Staining and analysis of thymocytes, splenocytes, and bone marrow cells was performed as described in Müller et ai (1989).
glycoprotein
lized eggs isolated from a cross of Ft (C57BL/6 x SJL) mice. A total of 72 eggs survived the injection and were implanted into two pseudopregnant foster animals. Thirteen pups were born, and 5 of these were identified as positive for the injected transgene. They were used for breeding and all transmitted the transgene to offspring. Expression analyses were performed with all five H2IL-6 lines by isolation of RNA from 11 different organs. As documented in Table 1, all lines but one (822-5) expressed the IL-6 gene in several organs. The pattern of expression resembled the endogenous H2 expression pattern: high in spleen and thymus, intermediate in several other organs, and absent in brain and pancreas. In all mice with high IL-6 expression, RNA isolated from spleen was degraded (see Fig. 2B). Thus, these new transgenic lines provide a useful system to investigate the consequences of constitutive IL-6 expression.
Transgenic IL-6 is biologically active Several acute-phase genes in the liver can be induced by either by IL-1 and/or IL-6 (Fey and Fuller, 1987). To test which of the acute-phase genes could be used as a marker for biologically active IL-6, we have injected several nontransgenic mice either with human or murine recombinant
Histopathology and staining of plasma cells
RESULTS Generation
of IL-6 transgenic mice
To facilitate high expression of the murine IL-6 gene, the cDNA was inserted into the vector pl63/7 (Fig. 1). The MHC class I promoter H2 is known for its broad expression (Rüther et ai, 1988) and the intron 5' of the IL-6 cDNA should enhance the expression (Palmiter et ai, 1991). The gel-purified fragment was injected into ferti-
y ////;///;////
Line
H +++ +++
in
p 163/7
Different Organs
Lg
K
M
B
Sg
+ +++
nd
++ ++
P.L
/'
FIG. 1. The vector (pl63/7) into which the IL-6 cDNA was inserted as an Eco Rl-Nhe I fragment is illustrated. The broad striped area represents the MHC class I H2 promoter and the fine dotted drawing defines the rabbit ßglobin sequences. This globin fragment contains the last 20 bp of exon 2, the entire intron 2, and the first 50 bp of exon 3 is followed by a polylinker (PL). The rest of exon 3 of |8-globin (including the poly(A) addition signal) is located 3' of the polylinker. In addition the SV40 early gene poly(A) addition signal (fine stripped) is present in this construct. Abbreviations: H, Hind III; Sa, Sal I; B, Bam HI; E, Eco RI; N, Nhe I; K, Kpn I; S, Sst I; P.L., polylinker.
Table 1. H2-IL-6 Expression
822-5 822-8 822-9 822-10 822-12
Sa B
H
Fixation, embedding, and staining of tissue samples were performed as described in Rüther et al. (1987). For the staining of cells in lymphomas, a cytospin was made from a single-cell suspension and than fixed with methanol. Subsequently, cells were double-stained with goat anti-mouse kappa/goat anti-mouse IgG or goat anti-mouse kappa/goat anti-mouse IgM (XRTC/FITC) (SBA).
+++
++
+ + +
The relative abundances of exogenous IL-6 expression are shown. S, Spleen; P, pancreas; L, liver; T, thymus; H, heart; Lg, lungs; G, gonads; K, kidneys; M, skeletal muscle; B, brain; Sg, salivary glands.
589
IL-6 IN TRANSGENIC MICE
S
P
L
T
H
Lg
K
M
B
Sg Co
SPLTHLgGKMBSgCo
B
IL6
SAA
SAA
FIG. 2. Expression analysis of IL-6 and SAA2 mRNA in H2-IL-6 mice. RNA from various tissues was analyzed by Northern blotting using the mouse IL-6 probe or by reprobing the same filters with the mouse SAA2 cDNA. RNA was isolated from a female mouse of line 822-12 (A) or from a male mouse of line 822-9 (B). Abbreviations: S, spleen; P, pancreas; L, liver; T, thymus; H, heart; Lg, lungs; G, gonads; K, kidneys; M, skeletal muscle; B, brain; Sg, salivary glands; Co, control J558 cells transfected with pEB.B-IL-6. (Blankenstein et ai, 1991). IL-6. RNA of livers of these mice were analyzed for the expression of alpha-1-acid glycoprotein and serum amyloid protein SAA2. Strongest induction was found for the SAA gene (data not shown). Therefore, we have reprobed all Northern blots of H2-IL-6 transgenic mice with the mouse SAA2 cDNA. As shown in Fig. 2, we have found no SAA expression in mice with no or low IL-6 expression (Fig. 2A). However, in mice with high IL-6 expression, the SAA2 gene was found to be strongly expressed (Fig. 2B). In general, we have always found this correlation between the level of IL-6 expression and the induction of SAA2.
2
3
10
The same correlation was also found for the induction of the alpha-1-acid glycoprotein (data not shown). To correlate the RNA expression data with the concentration of IL-6, we measured in several mice the IL-6 levels at the time point of killing for the isolation of RNA. Those mice with a reasonable RNA expression of IL-6 (as in Fig. 2B) had serum IL-6 levels between 200 and 700 pg/ml
(data
not
shown).
The result shown in Fig. 2 suggested that not all transgenic mice synthesize enough active IL-6 to be suitable for further analysis. Therefore, we have grouped the transgenic mice based on the expression of SAA into SAA* and SAA" mice. One example is shown in Fig. 3. These first analyses already indicated that the levels of transgenic IL-6 are sufficient for the induction of certain liver genes.
Consequences of IL-6 overexpression hematopoietic system
on
the
The effect of IL-6 on different cell types of the hematopoietic system is well characterized (Kishimoto and Hirano, 1988). In a recent study, transgenic mice were generated in which human IL-6 was strongly overexpressed in B cells. As a consequence, these mice developed massive
FIG. 3. Expression analysis of RNA isolated from liver of six transgenic mice of line 822-9 (lanes 1-6) and three nontransgenic mice (lanes 7-9) using as a probe the SAA2 cDNA. As a control, RNA from the liver of the animal shown in Fig. 2B was chosen. To document the presence of RNA in the SAA" lanes, the filter was overexposed (lower part of figure) to show the background expression of the SAA gene.
590
WOODROOFE ET AL.
plasmacytosis but no plasmacytoma (Suematsu et ai, 1989). Therefore, we have measured the concentration of
different antibodies in the sera of the H2-IL-6 mice. Howserum levels of IgM, IgGt, and IgG3 were similar to litter mate controls (data not shown). To analyze the influence of IL-6 in more detail, cells from thymus, spleen, and bone marrow were isolated, stained for different surface markers, and quantified by flow cytometry. A representative result from such an experiment is shown in Table 2. This analysis was performed with the animals analyzed for SAA expression (Fig. 3). The thymocyte population did not display an obvious alteration. However, the relative cell number in the B-cell lineage was slightly altered in spleen (mature B cells, B220*) and strongly reduced in bone marrow (pre-B cells, B220+, IgM"). The very same picture emerged when mice of two other H2-IL-6 transgenic mouse lines (822-10, 82212) were analyzed. Also, in SAA* mice of these lines, we found this change in the proportion of cells of the B lineever,
age.
All
hematopoietic organs as well as liver, kidney, and among others, were examined histopathologically, but no alterations were detectable. Thus, the overexpreslungs,
0
wA
sion of IL-6, which is sufficient to induce liver gene expression, does not cause a plasmacytosis but has an influence on the maturation of B cells.
H2-IL-6 mice
can
develop lymphomas
i-
H2-IL-6 mice from all five lines were bred for several generations and inspected for changes on a weekly basis. Most of the mice were healthy up to an age of 2 years; however, some of them, in particular of lines 822-8, 9, and 12, but also to a certain extent of line 822-10, developed lymphomas between 18 and 24 months at different sites in the body. Most frequently, these tumors were associated
Thymus H2-IL-6 SAA +a H2-IL-6 SAA -a coa
Spleen H2-IL-6 SAA H2-IL-6 SAA
+a -a
coa
Bone
Histology of lymphoma from mouse #822-12-34. magnification of intestine-associated lymphoma. magnification of lymphoma showing an islet of regularly shaped lymphocytes surrounded by massive plasmacytosis. of the
and
Cellular Composition
of
Bone Marrow
CD4VCD8+
CD4-/CD8-
CD8VCD4-
CD4/CD8-
75
75 75
8.2 10.1 10.1
7.6 7.6 7.6
9.0 6.4 6.5
CD4*/CD8-
CD8*/CD4-
CD4-/CD8-
B220*
23.2 25.7 25.7
7.3 9.5 9.5
68.2 63.0 63.0
35.1 46.7 46.7
marrow
H2-IL-6 SAA H2-IL-6 SAA
«r
FIG. 4. A. Low B. High
Table 2. Flow Cytometric Analysis
Thymus, Spleen,
I
B220VIgM*
B220*/IgM-
13.5 14.1 14.1
14.3 34.6 34.6
+a -a
coa
aThe animals are subgrouped by SAA expression positive (+, 2 animals) or negative (-, 4 animals) as documented in Fig. 3. Three animals have been used as controls (co). Lymphocytes were gated by forward and 90° scatter. The percentage of cells positive for the various
cell-surface markers in this gate
are
shown.
IL-6 IN TRANSGENIC MICE
with the lymph tissue surrounding the intestine (Fig. 4A), but were also apparent in lymph nodes and in kidney. These lymphomas were composed of big plasma cells characterized by eccentrically located nucleus and by regularly shaped lymphocytes (Fig. 4B). In some mice, the lymphomas were accompanied by amyloid deposits in the
spleen.
To estimate which cell type is primarily affected, the lymphoma cells were stained with antibodies directed against light and heavy chains of immunoglobulins. Most of the cells (about 90%) could be stained on the cell surface by anti-kappa and/or anti-IgG (Fig. 5). In addition, about 10% of the cells could be stained for cytoplasmic immunoglobulins, indicating that these lymphomas indeed included plasma cells. Most of the plasma cells could be identified as double-positive for kappa and IgG (Fig. 5A, B) and only some were double-positive for kappa and IgM (Fig. 5C, D). Thus, the lymphoma is of B-cell origin and is composed of different B-cell types, including IgM- and IgG-producing plasma cells. Growth of these lymphomas after transplantation into syngenic or nude mice has not been analyzed in this study. The overall frequency of lymphoma development was 22.4% (13 out of 58 transgenic mice). Only one out of more then 100 control littermates developed in a comparable age a lymphoma of similar phenotype. Thus, there is good evidence that the lymphoma development in the transgenic mice is a consequence of IL-6 overexpression.
591
DISCUSSION We generated transgenic mice expressing the murine IL-6 gene. This was possible by making use of an artificial splice cassette in front of the cDNA encoding IL-6. Although we found a variation of the expression levels of IL-6 in individuals and in lines, expression of IL-6 was high enough to cause an induction of IL-6-inducible liver genes. We chose the expression of the inducible liver gene SAA as a physiological marker for our further analyses. However, even in SAA* mice we could not detect an increased expression of IgG and have not found a plasmacytosis as has been reported (Suematsu et ai, 1989, 1992). This can be explained by differences in the constructs resulting in different expression levels of IL-6 protein (about 10-fold higher in Suematsu et ai, 1989, 1992) or by the different mouse strains (C57BL/6 versus Ft C57BL/6 x SJL) used in the experiments. In contrast, the H2-IL-6 mice had a normal life span (about 2 years) and therefore represented a source for studying long-term effects of IL-6. In this study, we could show that B-cell maturation was shifted from immature to a more mature state as documented by the distribution of B-cell markers in bone marrow and spleen. These changes, however, can obviously be tolerated by the mouse. We have never found any indication of immunological problems before 18 months of age. Thereafter, some mice developed lymphomas of B-cell origin.
FIG. 5. Cytoplasmic immunoglobulin staining of cells from mouse #822-9-2H-13. Cells were isolated from lymphoma and suspended to single cells. After cytospin and fixation, cells were double-stained with different antibodies and analyzed by fluorescence microscopy. Cells were stained with an anti-kappa (A, C) and either with an anti-IgG- (B) or with
an
anti-Ig-M (D) specific antibody.
592
WOODROOFE ET AL.
It has been speculated that certain autoimmune diseases such as rheumatoid arthritis are associated with IL-6 overexpression (Hirano et ai, 1988). All H2-IL-6 mice, when examined for pathology, were inspected for signs indicative of arthritis but without any success. Therefore, we would like to argue that either the overexpression of IL-6 alone is insufficient or the expression in the H2-IL-6 mice was not strong enough to cause this disease. In conclusion, the findings of this study provide evidence that IL-6 (in concentrations up to 500 pg/ml) can be considered as a potential inducer of IL-6-inducible promoters to drive heterologous genes without strongly affecting the physiological state of the mouse. Whether this finding can be generalized to other animal species remains open. At the present, we are testing the IL-6-inducible CRP promoter to express different genes for a defined length of time by serial injection of recombinant IL-6. The outcome of this study will show the usefulness of the IL-6inducible promoters.
ACKNOWLEDGMENTS We thank Prof. D. Komitowski for the pathology of lymphomas, Dr. C. Murphy for a critical reading of the manuscript, Thomas Pohl for construction of the vector for IL-6 expression, Gennaro Ciliberto for the human recombinant IL-6, and Barbara Stace for taking care of the mice.
REFERENCES ARNHEITER, H., SKUNTZ, S., NOTEBORN, M., CHANG, S., and MEIER, E. (1990). Transgenic mice with intracellulaT immunity to influenza virus. Cell 62, 51-61. BACHILLER, D„ and RÜTHER, U. (1990). Inducible expression of the proto-oncogene c-fos in transgenic mice. Arch. Geschwulstforsch. 60, 357-360. BLANKENSTEIN, T., QIN, Z., ÜBERLA, K., MÜLLER, W., ROSEN, H., VOLK, H.D., and DIAMANTSTEIN, T. (1991). Tumor suppression after tumor cell-targeted tumor necrosis factor a gene transfer. J. Exp. Med. 173, 1047-1052. CILIBERTO, G., ARCONE, R., WAGNER, E.F., and RÜTHER, U. (1987). Inducible and tissue-specific expression of human C-reactive protein in transgenic mice. EMBO J. 6, 4017-4022.
FEY, G.H., and FULLER, G.M. (1987). Regulation of acute phase gene expression by inflammatory mediators. Mol. Biol. Med. 4, 323-338. GANTER, U., ARCONE, R., TONIATTI, C, MORRONE, G., and CILIBERTO, G. (1989). Dual control of C-reactive protein gene expression by interleukin-1 and interleukin-6. EMBO J. 8, 3773-3779.
HIRANO, T., MATSUDA, T., TURNER, M., MIYASAKA, N.,
BUCHAN, G., TANG, B., SATO, K., SHIMIZU, M., MAINI, R., FELDMANN, M., and KISHIMOTO, T. (1988). Excessive production of interleukin 6/B cell stimulatory factor-2 in rheumatoid arthritis. Eur. J. Immunol. 18, 1797-1801. IWAKURA, Y., ASANO, M., NISHIMUNE, Y., and KAWADE, Y. (1988). Male sterility of transgenic mice carrying exogenous mouse interferon-/3 gene under the control of the metallothionein enhancer-promoter. EMBO J. 7, 3757-3762. KISHIMOTO, T., and HIRANO, T. (1988). Molecular regulation of B lymphocyte response. Annu. Rev. Immunol. 6, 485512.
MÜLLER, W., RÜTHER, U., VIEIRA, P., HOMBACH, J., RETH, M., and RAJEWSKY, K. (1989). Membrane-bound IgM obstructs B cell development in transgenic mice. Eur. J. Immunol. 19, 923-928. PALMITER, R.D., and BRINSTER, R.L. (1986). Germ-line transformation of mice. Annu. Rev. Genet.
20,
465-499.
PALMITER, R.D., SANDGREN, E.P., AVARBOCK, M.R., ALLEN, D.D., and BRINSTER, R.L. (1991). Heterologous introns can enhance expression of transgenes in mice. Proc. Nati. Acad. Sei. USA 88, 478-482. RICCA, G.A., HAMILTON, R.W., McLEAN, J.W., CONN, A., KALINYAK, J.E., and TAYLOR, J.M. (1981). Rat a-l-acid glycoprotein 10362-10368.
mRNA.
Biol.
J.
Chem.
256,
RÜTHER, U., GARBER, C, KOMITOWSKI, D., MÜLLER,
R., and WAGNER, E.F. (1987). Deregulated c-fos expression interferes with normal bone development in transgenic mice. Nature 325, 412-416.
RÜTHER, U., MÜLLER, W., SUMIDA, T., TOKUHISA, T., RAJEWSKY, K., and WAGNER, E.F. (1988). c-fos expression interferes with thymus development in transgenic mice. Cell 53, 847-856.
SUEMATSU, S., MATSUDA, T., AOZASA, K., AKIRA, S., NAKANO, N., OHNO, S., MIYAZAKI, J., YAMAMURA, K., HIRANO, T., and KISHIMOTO, T. (1989). IgGl plasmacytosis in interleukin 6 transgenic mice. Proc. Nati. Acad. Sei. USA 86, 7547-7551. SUEMATSU, S., MATSUSAKA, T., MATSUDA, T., OHNO, S., MIYAZAKI, J., YAMAMURA, K., HIRANO, T., and KISHIMOTO, T. (1992). Generation of plasmacytomas with the chromosomal translocation t(12;15) in interleukin 6 transgenic mice. Proc. Nati. Acad. Sei. USA 89, 232-235. VAN SNICK, J., CAYPHAS, S., SZIKORA, J.P., RENAULD, J.C., VAN ROOST, E., BOON, T., and SIMPSON, R.J. (1988). cDNA cloning of murine interleukin-HPl: Homology with human interleukin 6. Eur. J. Immunol. 18, 193-197. Address
reprint requests
to:
Dr. Ulrich Rüther
MHH P.O. Box 610180 3000 Hanover 61, Germany Received for publication March 10, 1992; 1992.
accepted
June
24,
Pp. 587-592
Long-Term Consequences of Interleukin-6 Overexpression in Transgenic Mice CATHERINE WOODROOFE, WERNER
MÜLLER,* and ULRICH RÜTHER
ABSTRACT With the aim of using interleukin-6 (IL-ó)-inducible promoters to express transgenes, we investigated the long-term consequences of high levels of IL-6 in mice. As a first step, we generated transgenic mice constitutively expressing the murine IL-6 at a level sufficient to induce IL-6-responsive genes. These mice were analyzed with respect to the indirect and direct consequences of elevated IL-6 expression over a time period of about 2 years. Although biologically active IL-6 was expressed from the transgene and different alterations could be documented (less immature B cells in bone marrow, expression of IL-6-inducible liver genes), the mice appeared healthy and could easily be used for breeding. Only in mice older than 18 months did we find a high incidence of lymphomas associated with different tissues. These results indicate that the side effects of long-term treatment with IL-6 are relatively moderate, and that IL-6 might be used to mediate the expression of heterologous genes in the context of functional studies.
INTRODUCTION useful system study Today, regulation, development, and tumorigenesis. sometimes certain be to
transgenic animals are a
gene
However,
aspects
can
only
analyzed
when the gene of interest is expressed from a tight but inducible promoter, e.g., in functional studies or when the overexpression of a product is growth inhibitory. Therefore, different inducible promoters, such as the mouse mammary tumor virus long terminal repeat and the mouse and the human metallothionein promoters (for review, see Palmiter and Brinster, 1986), have been tested over the past 5 years. Although these promoters are inducible up to 200-fold, they exhibit a basal level of expression prior to induction. In particular experiments, a constitutive expression of certain gene products is not acceptable, e.g., when the analysis of the specificity of the transgene activity is the aim of the experiment. Therefore, in addition, to the above-mentioned promoters, the interferon-(IFN)-inducible Mx promoter has been characterized (Arnheiter et al., 1990). The advantage of the Mx promoter is its very low constitutive expression and its very broad and high expression in almost every tissue upon induction (Arnheiter et ai, 1990; Bachiller and Rüther,
1990).
Despite the good transient inducibility of these promoters, the real problem becomes obvious in the context of long-term induction and by the induction through endogenous products. The metallothionein promoters, for example, can only be induced efficiently by the application of heavy metals, which do induce a series of other genes including stress genes such as heat shock proteins. On the other hand, there is a strong induction of the metallothionein genes by endogenous factors in mouse development that could make it impossible to use them. Problems might also be associated with the IFN-inducible Mx promoter. It has been shown that elevated expression of IFN causes male sterility (Iwakura et ai, 1988). On the basis of all these observations, we were interested to find an inducible promoter system in which the genes are tightly regulated and the effects of long-term induction are reduced to a minimum. Recently, we have found that the human C-reactive protein gene (CRP) does fulfill the first criteria in transgenic mice (Ciliberto et ai, 1987).
Consequently, we are investigating the long-term problems of induction. Since interleukin-6 (IL-6) is the direct mediator of CRP induction (Ganter et ai, 1989), we have generated transgenic mice that overproduce IL-6. These mice were used as a means of evaluating the consequences of constitutive IL-6 expression.
European Molecular Biology Laboratory, 6900 Heidelberg, Germany. •Institut für Genetik, Universität zu Köln, 5000 Köln 41, Germany. 587
588
WOODROOFE ET AL.
MATERIALS AND METHODS IL-6 construct and
microinjection
A 695-bp cDNA fragment of the murine IL-6 gene was isolated from pEP.B-IL-6 (Blankenstein et ai, 1991) and was then inserted into the Eco Rl-Nhe I sites of plasmid pi63/7 (see Fig. 1). This construct should enhance the expression of the cDNA due to the presence of intron 2 of ßglobin. For microinjection, the Hind lll-Sst I fragment was gel-purified. DNA was microinjected into fertilized eggs of a cross between F, (C57BL/6 x SJL).
Expression and flow cytometry analysis RNA isolation and Northern analysis was performed as described (Rüther et ai, 1987). As probes, we have used the IL-6 cDNA (Van Snick et ai, 1988), the alpha-1-acid
cDNA (Ricca et ai, 1981) and the mouse SAA2 cDNA (kindly provided by Hugh Rienhoff). Human recombinant IL-6 was a gift of Gennaro Ciliberto and mouse recombinant IL-6 was purchased from UBI. For the quantification of serum IL-6, the ENDOGEN murine IL-6 ELISA was used. Staining and analysis of thymocytes, splenocytes, and bone marrow cells was performed as described in Müller et ai (1989).
glycoprotein
lized eggs isolated from a cross of Ft (C57BL/6 x SJL) mice. A total of 72 eggs survived the injection and were implanted into two pseudopregnant foster animals. Thirteen pups were born, and 5 of these were identified as positive for the injected transgene. They were used for breeding and all transmitted the transgene to offspring. Expression analyses were performed with all five H2IL-6 lines by isolation of RNA from 11 different organs. As documented in Table 1, all lines but one (822-5) expressed the IL-6 gene in several organs. The pattern of expression resembled the endogenous H2 expression pattern: high in spleen and thymus, intermediate in several other organs, and absent in brain and pancreas. In all mice with high IL-6 expression, RNA isolated from spleen was degraded (see Fig. 2B). Thus, these new transgenic lines provide a useful system to investigate the consequences of constitutive IL-6 expression.
Transgenic IL-6 is biologically active Several acute-phase genes in the liver can be induced by either by IL-1 and/or IL-6 (Fey and Fuller, 1987). To test which of the acute-phase genes could be used as a marker for biologically active IL-6, we have injected several nontransgenic mice either with human or murine recombinant
Histopathology and staining of plasma cells
RESULTS Generation
of IL-6 transgenic mice
To facilitate high expression of the murine IL-6 gene, the cDNA was inserted into the vector pl63/7 (Fig. 1). The MHC class I promoter H2 is known for its broad expression (Rüther et ai, 1988) and the intron 5' of the IL-6 cDNA should enhance the expression (Palmiter et ai, 1991). The gel-purified fragment was injected into ferti-
y ////;///;////
Line
H +++ +++
in
p 163/7
Different Organs
Lg
K
M
B
Sg
+ +++
nd
++ ++
P.L
/'
FIG. 1. The vector (pl63/7) into which the IL-6 cDNA was inserted as an Eco Rl-Nhe I fragment is illustrated. The broad striped area represents the MHC class I H2 promoter and the fine dotted drawing defines the rabbit ßglobin sequences. This globin fragment contains the last 20 bp of exon 2, the entire intron 2, and the first 50 bp of exon 3 is followed by a polylinker (PL). The rest of exon 3 of |8-globin (including the poly(A) addition signal) is located 3' of the polylinker. In addition the SV40 early gene poly(A) addition signal (fine stripped) is present in this construct. Abbreviations: H, Hind III; Sa, Sal I; B, Bam HI; E, Eco RI; N, Nhe I; K, Kpn I; S, Sst I; P.L., polylinker.
Table 1. H2-IL-6 Expression
822-5 822-8 822-9 822-10 822-12
Sa B
H
Fixation, embedding, and staining of tissue samples were performed as described in Rüther et al. (1987). For the staining of cells in lymphomas, a cytospin was made from a single-cell suspension and than fixed with methanol. Subsequently, cells were double-stained with goat anti-mouse kappa/goat anti-mouse IgG or goat anti-mouse kappa/goat anti-mouse IgM (XRTC/FITC) (SBA).
+++
++
+ + +
The relative abundances of exogenous IL-6 expression are shown. S, Spleen; P, pancreas; L, liver; T, thymus; H, heart; Lg, lungs; G, gonads; K, kidneys; M, skeletal muscle; B, brain; Sg, salivary glands.
589
IL-6 IN TRANSGENIC MICE
S
P
L
T
H
Lg
K
M
B
Sg Co
SPLTHLgGKMBSgCo
B
IL6
SAA
SAA
FIG. 2. Expression analysis of IL-6 and SAA2 mRNA in H2-IL-6 mice. RNA from various tissues was analyzed by Northern blotting using the mouse IL-6 probe or by reprobing the same filters with the mouse SAA2 cDNA. RNA was isolated from a female mouse of line 822-12 (A) or from a male mouse of line 822-9 (B). Abbreviations: S, spleen; P, pancreas; L, liver; T, thymus; H, heart; Lg, lungs; G, gonads; K, kidneys; M, skeletal muscle; B, brain; Sg, salivary glands; Co, control J558 cells transfected with pEB.B-IL-6. (Blankenstein et ai, 1991). IL-6. RNA of livers of these mice were analyzed for the expression of alpha-1-acid glycoprotein and serum amyloid protein SAA2. Strongest induction was found for the SAA gene (data not shown). Therefore, we have reprobed all Northern blots of H2-IL-6 transgenic mice with the mouse SAA2 cDNA. As shown in Fig. 2, we have found no SAA expression in mice with no or low IL-6 expression (Fig. 2A). However, in mice with high IL-6 expression, the SAA2 gene was found to be strongly expressed (Fig. 2B). In general, we have always found this correlation between the level of IL-6 expression and the induction of SAA2.
2
3
10
The same correlation was also found for the induction of the alpha-1-acid glycoprotein (data not shown). To correlate the RNA expression data with the concentration of IL-6, we measured in several mice the IL-6 levels at the time point of killing for the isolation of RNA. Those mice with a reasonable RNA expression of IL-6 (as in Fig. 2B) had serum IL-6 levels between 200 and 700 pg/ml
(data
not
shown).
The result shown in Fig. 2 suggested that not all transgenic mice synthesize enough active IL-6 to be suitable for further analysis. Therefore, we have grouped the transgenic mice based on the expression of SAA into SAA* and SAA" mice. One example is shown in Fig. 3. These first analyses already indicated that the levels of transgenic IL-6 are sufficient for the induction of certain liver genes.
Consequences of IL-6 overexpression hematopoietic system
on
the
The effect of IL-6 on different cell types of the hematopoietic system is well characterized (Kishimoto and Hirano, 1988). In a recent study, transgenic mice were generated in which human IL-6 was strongly overexpressed in B cells. As a consequence, these mice developed massive
FIG. 3. Expression analysis of RNA isolated from liver of six transgenic mice of line 822-9 (lanes 1-6) and three nontransgenic mice (lanes 7-9) using as a probe the SAA2 cDNA. As a control, RNA from the liver of the animal shown in Fig. 2B was chosen. To document the presence of RNA in the SAA" lanes, the filter was overexposed (lower part of figure) to show the background expression of the SAA gene.
590
WOODROOFE ET AL.
plasmacytosis but no plasmacytoma (Suematsu et ai, 1989). Therefore, we have measured the concentration of
different antibodies in the sera of the H2-IL-6 mice. Howserum levels of IgM, IgGt, and IgG3 were similar to litter mate controls (data not shown). To analyze the influence of IL-6 in more detail, cells from thymus, spleen, and bone marrow were isolated, stained for different surface markers, and quantified by flow cytometry. A representative result from such an experiment is shown in Table 2. This analysis was performed with the animals analyzed for SAA expression (Fig. 3). The thymocyte population did not display an obvious alteration. However, the relative cell number in the B-cell lineage was slightly altered in spleen (mature B cells, B220*) and strongly reduced in bone marrow (pre-B cells, B220+, IgM"). The very same picture emerged when mice of two other H2-IL-6 transgenic mouse lines (822-10, 82212) were analyzed. Also, in SAA* mice of these lines, we found this change in the proportion of cells of the B lineever,
age.
All
hematopoietic organs as well as liver, kidney, and among others, were examined histopathologically, but no alterations were detectable. Thus, the overexpreslungs,
0
wA
sion of IL-6, which is sufficient to induce liver gene expression, does not cause a plasmacytosis but has an influence on the maturation of B cells.
H2-IL-6 mice
can
develop lymphomas
i-
H2-IL-6 mice from all five lines were bred for several generations and inspected for changes on a weekly basis. Most of the mice were healthy up to an age of 2 years; however, some of them, in particular of lines 822-8, 9, and 12, but also to a certain extent of line 822-10, developed lymphomas between 18 and 24 months at different sites in the body. Most frequently, these tumors were associated
Thymus H2-IL-6 SAA +a H2-IL-6 SAA -a coa
Spleen H2-IL-6 SAA H2-IL-6 SAA
+a -a
coa
Bone
Histology of lymphoma from mouse #822-12-34. magnification of intestine-associated lymphoma. magnification of lymphoma showing an islet of regularly shaped lymphocytes surrounded by massive plasmacytosis. of the
and
Cellular Composition
of
Bone Marrow
CD4VCD8+
CD4-/CD8-
CD8VCD4-
CD4/CD8-
75
75 75
8.2 10.1 10.1
7.6 7.6 7.6
9.0 6.4 6.5
CD4*/CD8-
CD8*/CD4-
CD4-/CD8-
B220*
23.2 25.7 25.7
7.3 9.5 9.5
68.2 63.0 63.0
35.1 46.7 46.7
marrow
H2-IL-6 SAA H2-IL-6 SAA
«r
FIG. 4. A. Low B. High
Table 2. Flow Cytometric Analysis
Thymus, Spleen,
I
B220VIgM*
B220*/IgM-
13.5 14.1 14.1
14.3 34.6 34.6
+a -a
coa
aThe animals are subgrouped by SAA expression positive (+, 2 animals) or negative (-, 4 animals) as documented in Fig. 3. Three animals have been used as controls (co). Lymphocytes were gated by forward and 90° scatter. The percentage of cells positive for the various
cell-surface markers in this gate
are
shown.
IL-6 IN TRANSGENIC MICE
with the lymph tissue surrounding the intestine (Fig. 4A), but were also apparent in lymph nodes and in kidney. These lymphomas were composed of big plasma cells characterized by eccentrically located nucleus and by regularly shaped lymphocytes (Fig. 4B). In some mice, the lymphomas were accompanied by amyloid deposits in the
spleen.
To estimate which cell type is primarily affected, the lymphoma cells were stained with antibodies directed against light and heavy chains of immunoglobulins. Most of the cells (about 90%) could be stained on the cell surface by anti-kappa and/or anti-IgG (Fig. 5). In addition, about 10% of the cells could be stained for cytoplasmic immunoglobulins, indicating that these lymphomas indeed included plasma cells. Most of the plasma cells could be identified as double-positive for kappa and IgG (Fig. 5A, B) and only some were double-positive for kappa and IgM (Fig. 5C, D). Thus, the lymphoma is of B-cell origin and is composed of different B-cell types, including IgM- and IgG-producing plasma cells. Growth of these lymphomas after transplantation into syngenic or nude mice has not been analyzed in this study. The overall frequency of lymphoma development was 22.4% (13 out of 58 transgenic mice). Only one out of more then 100 control littermates developed in a comparable age a lymphoma of similar phenotype. Thus, there is good evidence that the lymphoma development in the transgenic mice is a consequence of IL-6 overexpression.
591
DISCUSSION We generated transgenic mice expressing the murine IL-6 gene. This was possible by making use of an artificial splice cassette in front of the cDNA encoding IL-6. Although we found a variation of the expression levels of IL-6 in individuals and in lines, expression of IL-6 was high enough to cause an induction of IL-6-inducible liver genes. We chose the expression of the inducible liver gene SAA as a physiological marker for our further analyses. However, even in SAA* mice we could not detect an increased expression of IgG and have not found a plasmacytosis as has been reported (Suematsu et ai, 1989, 1992). This can be explained by differences in the constructs resulting in different expression levels of IL-6 protein (about 10-fold higher in Suematsu et ai, 1989, 1992) or by the different mouse strains (C57BL/6 versus Ft C57BL/6 x SJL) used in the experiments. In contrast, the H2-IL-6 mice had a normal life span (about 2 years) and therefore represented a source for studying long-term effects of IL-6. In this study, we could show that B-cell maturation was shifted from immature to a more mature state as documented by the distribution of B-cell markers in bone marrow and spleen. These changes, however, can obviously be tolerated by the mouse. We have never found any indication of immunological problems before 18 months of age. Thereafter, some mice developed lymphomas of B-cell origin.
FIG. 5. Cytoplasmic immunoglobulin staining of cells from mouse #822-9-2H-13. Cells were isolated from lymphoma and suspended to single cells. After cytospin and fixation, cells were double-stained with different antibodies and analyzed by fluorescence microscopy. Cells were stained with an anti-kappa (A, C) and either with an anti-IgG- (B) or with
an
anti-Ig-M (D) specific antibody.
592
WOODROOFE ET AL.
It has been speculated that certain autoimmune diseases such as rheumatoid arthritis are associated with IL-6 overexpression (Hirano et ai, 1988). All H2-IL-6 mice, when examined for pathology, were inspected for signs indicative of arthritis but without any success. Therefore, we would like to argue that either the overexpression of IL-6 alone is insufficient or the expression in the H2-IL-6 mice was not strong enough to cause this disease. In conclusion, the findings of this study provide evidence that IL-6 (in concentrations up to 500 pg/ml) can be considered as a potential inducer of IL-6-inducible promoters to drive heterologous genes without strongly affecting the physiological state of the mouse. Whether this finding can be generalized to other animal species remains open. At the present, we are testing the IL-6-inducible CRP promoter to express different genes for a defined length of time by serial injection of recombinant IL-6. The outcome of this study will show the usefulness of the IL-6inducible promoters.
ACKNOWLEDGMENTS We thank Prof. D. Komitowski for the pathology of lymphomas, Dr. C. Murphy for a critical reading of the manuscript, Thomas Pohl for construction of the vector for IL-6 expression, Gennaro Ciliberto for the human recombinant IL-6, and Barbara Stace for taking care of the mice.
REFERENCES ARNHEITER, H., SKUNTZ, S., NOTEBORN, M., CHANG, S., and MEIER, E. (1990). Transgenic mice with intracellulaT immunity to influenza virus. Cell 62, 51-61. BACHILLER, D„ and RÜTHER, U. (1990). Inducible expression of the proto-oncogene c-fos in transgenic mice. Arch. Geschwulstforsch. 60, 357-360. BLANKENSTEIN, T., QIN, Z., ÜBERLA, K., MÜLLER, W., ROSEN, H., VOLK, H.D., and DIAMANTSTEIN, T. (1991). Tumor suppression after tumor cell-targeted tumor necrosis factor a gene transfer. J. Exp. Med. 173, 1047-1052. CILIBERTO, G., ARCONE, R., WAGNER, E.F., and RÜTHER, U. (1987). Inducible and tissue-specific expression of human C-reactive protein in transgenic mice. EMBO J. 6, 4017-4022.
FEY, G.H., and FULLER, G.M. (1987). Regulation of acute phase gene expression by inflammatory mediators. Mol. Biol. Med. 4, 323-338. GANTER, U., ARCONE, R., TONIATTI, C, MORRONE, G., and CILIBERTO, G. (1989). Dual control of C-reactive protein gene expression by interleukin-1 and interleukin-6. EMBO J. 8, 3773-3779.
HIRANO, T., MATSUDA, T., TURNER, M., MIYASAKA, N.,
BUCHAN, G., TANG, B., SATO, K., SHIMIZU, M., MAINI, R., FELDMANN, M., and KISHIMOTO, T. (1988). Excessive production of interleukin 6/B cell stimulatory factor-2 in rheumatoid arthritis. Eur. J. Immunol. 18, 1797-1801. IWAKURA, Y., ASANO, M., NISHIMUNE, Y., and KAWADE, Y. (1988). Male sterility of transgenic mice carrying exogenous mouse interferon-/3 gene under the control of the metallothionein enhancer-promoter. EMBO J. 7, 3757-3762. KISHIMOTO, T., and HIRANO, T. (1988). Molecular regulation of B lymphocyte response. Annu. Rev. Immunol. 6, 485512.
MÜLLER, W., RÜTHER, U., VIEIRA, P., HOMBACH, J., RETH, M., and RAJEWSKY, K. (1989). Membrane-bound IgM obstructs B cell development in transgenic mice. Eur. J. Immunol. 19, 923-928. PALMITER, R.D., and BRINSTER, R.L. (1986). Germ-line transformation of mice. Annu. Rev. Genet.
20,
465-499.
PALMITER, R.D., SANDGREN, E.P., AVARBOCK, M.R., ALLEN, D.D., and BRINSTER, R.L. (1991). Heterologous introns can enhance expression of transgenes in mice. Proc. Nati. Acad. Sei. USA 88, 478-482. RICCA, G.A., HAMILTON, R.W., McLEAN, J.W., CONN, A., KALINYAK, J.E., and TAYLOR, J.M. (1981). Rat a-l-acid glycoprotein 10362-10368.
mRNA.
Biol.
J.
Chem.
256,
RÜTHER, U., GARBER, C, KOMITOWSKI, D., MÜLLER,
R., and WAGNER, E.F. (1987). Deregulated c-fos expression interferes with normal bone development in transgenic mice. Nature 325, 412-416.
RÜTHER, U., MÜLLER, W., SUMIDA, T., TOKUHISA, T., RAJEWSKY, K., and WAGNER, E.F. (1988). c-fos expression interferes with thymus development in transgenic mice. Cell 53, 847-856.
SUEMATSU, S., MATSUDA, T., AOZASA, K., AKIRA, S., NAKANO, N., OHNO, S., MIYAZAKI, J., YAMAMURA, K., HIRANO, T., and KISHIMOTO, T. (1989). IgGl plasmacytosis in interleukin 6 transgenic mice. Proc. Nati. Acad. Sei. USA 86, 7547-7551. SUEMATSU, S., MATSUSAKA, T., MATSUDA, T., OHNO, S., MIYAZAKI, J., YAMAMURA, K., HIRANO, T., and KISHIMOTO, T. (1992). Generation of plasmacytomas with the chromosomal translocation t(12;15) in interleukin 6 transgenic mice. Proc. Nati. Acad. Sei. USA 89, 232-235. VAN SNICK, J., CAYPHAS, S., SZIKORA, J.P., RENAULD, J.C., VAN ROOST, E., BOON, T., and SIMPSON, R.J. (1988). cDNA cloning of murine interleukin-HPl: Homology with human interleukin 6. Eur. J. Immunol. 18, 193-197. Address
reprint requests
to:
Dr. Ulrich Rüther
MHH P.O. Box 610180 3000 Hanover 61, Germany Received for publication March 10, 1992; 1992.
accepted
June
24,
