Lymphoid hyperplasia in transgenic mice over-expressing a secreted form of the human interleukin-1β gene product

Abstract

To evaluate the biological effects of over-expression of interleukin-1β (IL-1β) on the immune system we have generated transgenic mice, expressing the IL-1β gene fused to a heterologous signal sequence under the control of the mouse immunoglobulin enhancer (Eμ). A prominent hyperplasia and a disturbed microarchitecture of lymphoid tissues were observed in the transgenic mice. The CD4⁺ T cells in the hyperplastic lymphoid organs seemed to invade the majority of the lymphoid organs including B-cell restricted areas. Analysis of lymph node cells revealed an increased frequency of CD4⁺ CD44^high CD62L⁻ T cells and local secretion of IL-2 and IL-4, compatible with an elevated number of activated T cells. Furthermore, significant levels of human IL-1β in sera and high concentrations of serum immunoglobulin G (IgG) were observed in the transgenic mice. The data suggest a role for IL-1β in controlling lymphoid microarchitecture and, when over-expressed, breaking the threshold in T-helper–B-cell interaction.

INTRODUCTION

Interleukin (IL)-1β is one of the most potent and pleiotropic cytokines. It is mainly produced by activated antigen-presenting cells (APC) and plays an important role both in normal immunoregulatory processes and in pathophysiological inflammatory responses.¹ The expression of IL-1β is rigorously regulated at different levels, probably as a result of its toxicity and inflammatory properties when excessively released. The control mechanisms include transcriptional and translational regulation as well as release of neutralizing IL-1 type II decoy receptors and IL-1 receptor antagonists (ra).^1–3 IL-1β is synthesized as a 31–34 000 MW inactive proform and is subsequently processed by interleukin 1-β converting enzyme (ICE) into a 17 000 MW bioactive form.¹ In contrast to most other cytokines IL-1β lacks a typical signal sequence. The mechanism mediating the release of the bioactive protein is still elusive although it has been suggested that IL-1β release is correlated with apoptosis.^4,5 The finding that ICE is homologous to the Caenorhabditis elegans cell-death gene ced-3,^6,7 provides further support for the assumed relation between IL-1β release and apoptosis of the producer cell.

The biological effects of IL-1β and tumour necrosis factor (TNF) have been extensively characterized. These two cytokines are able to induce several similar functions, such as induction of mitogen-activated protein (MAP) kinases,^8,9 upregulation of adhesion molecules (intracellular adhesion molecule-1; ICAM-1, vascular cell adhesion molecule; VCAM-1)¹⁰ and induction of inflammatory responses. Extensive studies have been performed to unravel the physiological role of IL-1 in vivo, but still many issues remain elusive. TNF and lymphotoxin (LT)-α, in addition to their role in inflammation and cytotoxicity, have recently been shown to have a pivotal role in lymphoid development and tissue organization.^11–14 This suggests that inflammatory cytokines may in a physiological setting, regulate tissue architecture and during pathological responses influence tissue remodeling.

We have in an earlier study developed a biological model in which IL-1β can be directed extracellularly without affecting the viability of the producer cell. In order to obtain a released form of IL-1β, the mature human IL-1β cDNA was fused to a signal sequence, derived from the structurally related IL-1ra. Transfection of different cell lines with the hybrid signal sequence-IL-1β construct, ssIL-1β, resulted in the release of large amounts of IL-1β, whereas the mature IL-1β without the signal sequence accumulated intracellularly.¹⁵ Moreover, we have shown that subcutaneous tumour growth of ssIL-1β transduced B16 F10 mouse melanoma cells was significantly reduced compared with mock- and IL-1β transduced controls and immunohistological analysis of tumour biopsies revealed strong infiltration of macrophages and moderate infiltration of CD4⁺ T cells. This indicates that IL-1β either directly or indirectly influences the migration of CD4⁺ T cells and monocytes in vivo.¹⁶

We hypothesized that transgenic over expression of IL-1β in a lymphoid environment might offer opportunities to elucidate the biological effects of IL-1 in cell-mediated immune responses, and to serve as a useful tool when studying the influence of IL-1β on lymphoid tissue. We now demonstrate that transgenic expression of human IL-1β fused to a heterologous signal sequence (ssIL-1β) under the control of an immunoglobulin heavy chain enhancer results in lymphoid hyperplasia, hypergammaglobulinaemia and altered architecture in lymphoid tissues. Our results are in accordance with earlier findings which have shown that injection of recombinant IL-1β induced lymphoid hyperplasia and expression of IL-2 and IL-4·¹⁷ This indicates a role for IL-1β in triggering changes of the microenvironment in lymphoid organs, and/or modifying the balance of T helper cell–B-cell interactions.

MATERIALS AND METHODS

Generation of ssIL-1β transgenic mice

A 170-bp Hin fI/DdeI fragment from the immunoglobulin μ enhancer, which includes μE1 to μB, was duplicated eight times and cloned into a blunted Hin dIII site of pGEM-3Z vector with a Not I linker in Eco RI (Promega, Madison, WI). A duplicated SP1 binding site and a TATA-box sequence were cloned in blunted Sph I and Pst I sites, respectively. A 2000-bp Bam H1/blunted Xba I rabbit β-globin poly A-fragment, including the end of exon 2, intron 2, exon 3 and the polyadenylation signal was further cloned into Bam HI/blunted Sac I site. Finally the 550 bp construct consisting of the mature human IL-1β cDNA, fused with the signal sequence derived from the structurally related IL-1ra,¹⁵ was exised from the retroviral vector pLXSN¹⁵ with Bam HI and was recloned into the BamHI site of pGEM-3Z vector (Fig. 1a). The resulting construct was used for pronuclear injection of fertilized mouse eggs of (C57BL/6×CBA) F1 mice. Resulting litters were screened for incorporation of the transgene by polymerase chain reaction (PCR) analysis of tail DNA. DNA preparations were performed according to manufacturer's instructions (QIAamp Tissue Kit, QIAGEN, GmbH, Hilden, Germany). One female founder was found. The founder was bred with a C57BL/6 male. The founder and the F₂-offspring carrying the transgene were bred and their offspring developed the phenotypic characteristics described in this paper.

Figure 1.

The construction of the human IL-1β expression vectors used for (a) pronuclear injections (pGEM-3Z-8E ssIL-1) and (b) transfection of embryonic stem cells (pGEM-3Z-E ssIL-1) (The vectors are not drawn in scale). (c) Serum levels of human IL-1β in the transgenic mice at the age of 33 (mouse 99), 22 (mouse 102) and 8 (mouse 1) weeks, respectively. Serum was prepared and analysed for IL-1β by ELISA technique.

A second ssIL-1β vector was constructed by exchanging the eight times duplicated 170 bp enhancer, see above, with a single 1-kb Xba Eμ enhancer (Fig. 1b). This vector was used for transfection of I-129 originated embryonic stem cells (E14) which were further injected into blastocysts. Host blastocysts were obtained from natural mating (C57BL/6) and transplanted into pseudopregnant females (C57BL/6/CBA F1). Live-born offspring were scored shortly after birth for chimeras on the basis of coat colour. The animals were maintained under conventional conditions at Transgenic Facility, University of Lund.

IL-1β enzyme-linked immunosorbent assay (ELISA)

Circulating levels of huIL-1β in serum of the transgenic mice were determined using the ELISA technique as described earlier.^15,18 Absorbance was measured at 450 nm by a Multiscan MC reader (Labsystem, Helsinki, Finland), and the samples were analysed by DELTA SOFT II software (BioMetallics, Inc., Princeton, NJ). Measurement of IL-1β content was performed in the linear part of the standard curve. This ELISA records huIL-1β with a detection limit down to 15 pg/ml but does not react with huIL-1α or muIL-1.

Immunoglobulin isotype-specific ELISA

ELISA was performed utilizing rat antimouse immunoglobulin antibodies specific for each isotype (ISO-2 kit, Sigma, St Louis, MO) and horseradish peroxidase (HRP)-labelled rabbit antimouse immunoglobulin (Dakopatts AB, Älvsjö, Sweden). In brief, 96-well assay plates were coated overnight at 4° with each isotype specific antibody. Duplicates of sera were serially diluted in phosphate-buffered saline (PBS) and were incubated for 1 hr at room temperature. After washing, the peroxidase-labelled antibody was added to each well and allowed to incubate for 30 min at room temperature. Finally, after washing, the freshly prepared substrate (20 mm tetrametylbenzidine in acetone and 0·1 m potassium citrate, 0·1 m citric acid in ddH₂O, pH 4·25 in a 1:20 ratio and finally 2 mm H₂O₂) was added and the reaction was further stopped by 1 m H₂SO₄. The absorbance was measured at 450 nm by a Multiscan MC reader (Labsystem, Helsinki, Finland). The samples were analysed by the software DELTA SOFT II.

Immunohistology

Biopsies of spleen and lymph node were fixed in 10% formalin before embedded in paraffin. Sections (5 μm) of these organs were stained with eosin/haematoxylin. Spleens and lymph nodes from the animals were removed after death and snap-frozen in isopentan (precooled to −55°). Cryostat sections (5 μm) were fixed in ice-cooled acetone and air-dried. Sections were blocked for 15 min in avidin, washed in PBS and then blocked for 15 min in biotin (Avidin/Biotin blocking kit; Vector Labs, Burlingame, CA, 4 droplets/ml diluted in PBS) before staining. The following primary antibodies were used: rat immunoglobulin G2a (IgG2a) antimouse CD4 (clone RM4–5), rat IgG2a antimouse CD8a (clone 53–6.7), rat IgG2a antimouse CD45R/B220 (clone RA3–6B2) and isotype control rat IgG2a (all from PharMingen, San Diego, CA). Staining with primary antibodies was followed by further incubation with biotinylated goat antirat IgG (Jackson Immunoresearch Laboratories Inc., West Grove, PA) at a 1:400 dilution, for 30 min. The sections were then incubated for 30 min with ABC Elite horseradish peroxidase (Vector Labs). The antigen–antibody complexes were made visible using diaminobenzidine (DAB) (Vector Labs) for 5 min. Finally, the slides were counterstained in methyl green and mounted in DPX medium (Kebo Lab AB, Stockholm, Sweden).

For detection of cytokines the above procedure was slightly altered. The fixative used was 2% formaldehyde in PBS for 20 min at 20°. Slides were washed in PBS for 5 min. Endogenous peroxidase was blocked with 1% H₂O₂/0·3 m NaN₃ in Eagle's balanced salt solution (EBSS) buffer (GibcoBRL, Gaithersburg, MD), supplemented with 0·1% saponin, which reversibly permeabilizes cell membranes. Saponin was present through out the experiment. The slides were washed three times in EBSS-buffer supplemented with 1% HEPES buffer. The sections were then blocked for 1 hr in avidin, washed and then blocked for 20 min in biotin (avidin and biotin were diluted in EBSS-buffer supplemented with 1% HEPES) before incubated with anticytokine monoclonal antibodies (mAbs; 1–5 μg/ml) at room temperature, over night. The following primary antibodies were used: rat IgG2a antimouse IL-2 (clone S4B6), rat IgGb antimouse IL-4 (clone BVD4-1D11), isotype control rat IgG2a and rat IgG2b (all from PharMingen). Staining with primary antibodies was followed by a 1-hr incubation with biotinylated goat antirat IgG (Jackson Labs) at a 1:400 dilution. The sections were then incubated for 1 hr with extrAvidin horseradish peroxidase (1:3000 dilution) (Sigma). The antigen–antibody complexes were made visible using diaminobenzidine (DAB, Vector Labs) for 5 min. Finally, the slides were counterstained in methyl green and were mounted in DPX medium.

Fluorescence-activated cell sorting (FACS) analysis of spleens and lymph nodes

Biopsies of spleens and cervical lymph nodes were rubbed through a 70-μm nylon cell strainer. The cells were washed in PBS/0·5% bovine serum albumin (BSA) before stained with specific antibodies. Flow cytometric analysis was performed using standard settings on a FACScan flow cytometer (Becton Dickinson, CA). The following fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-conjugated monoclonal antibodies were used: rat antimouse CD4 PE (clone RM4-5), rat antimouse CD44 FITC (clone IM7) and rat antimouse CD62L FITC (clone MEL-14), all purchased from PharMingen.

RESULTS

Production of transgenic mice secreting IL-1β

We have earlier shown that the addition of a signal sequence, from the structurally related IL-1ra, to IL-1β cDNA results in a hybrid gene directing high levels of bioactive IL-1β extracellularly in vitro.¹⁵ Based on this finding we wanted to investigate the effect of aberrant expression of this cytokine in vivo by construction of a transgenic mouse strain. In order to target the expression of IL-1β to the lymphoid system we inserted the cDNA into an expression vector containing a SP1 binding site and a TATA box as promoter and eight repetitive 170 bp fragments of the immunoglobulin heavy chain enhancer 5′ region as control elements (Fig. 1a).¹⁹ This construct was microinjected into the male pronucleus of fertilized eggs from CBA×C57BL/6 F1 mice and one founder carrying the transgene was identified. Upon breeding the transgenic mice had normal littersizes and were apparently healthy at birth. However, 10–15% of the transgenic progeny developed the phenotypic and pathological abnormalities described in this paper at the age of 2–5 months. Three of these have here been studied in greater detail, designated mouse 99, mouse 102 and mouse 129. Since the phenotypic penetrance was low and the phenotype was derived from one single founder mouse we constructed an additional strain of huIL-1β transgenic mice. For this purpose a modified vector containing a single complete immunoglobulin heavy chain intron enhancer was used (Fig. 1b). This would exclude that the phenotype was related to the integration site. Chimeric mice were generated by blastocyst injection of transfected embryonic stem (ES)-cells instead of pronuclear microinjection. Because the ES-cells are derived from I129 mice, this experiment would also clarify whether the phenotype was independent of the genetic background. Fifteen chimeric mice were obtained after injection of two separate ES-cell clones carrying the transgenic construct at independent integration sites. Two of the chimeras developed similar phenotypic and pathological abnormalities as mouse 99, 102 and 129. One of the two chimeras, designated mouse 1, was studied in detail. Furthermore, upon breeding, the transgenic progeny with the complete immunoglobulin heavy chain intron enhancer construct developed phenotypic changes at an early age (1–3 months) and with a much higher frequency.

In order to examine whether huIL-1β (huIL-1β) was expressed in the transgenic mice, sera were analysed using an ELISA assay that specifically reacts with huIL-1β. Substantial huIL-1β levels were recorded in the transgenic animals as well as in chimeras that developed a phenotype (Fig. 1c), while the huIL-1β serum levels in healthy transgenic littermate controls were below the level of detection. The phenotype seems to be independent of transgene integration site and genetic background and is also established in a chimeric environment. Below, the detailed analysis of the phenotype of the affected mice will be presented.

Pathological findings in tissues of huIL-1β transgenic mice

The most obvious phenotype of huIL-1β mice was splenomegaly and hyperplasia of the lymph nodes. The cervical, inguinal and mesenteric lymph nodes were enlarged and a diameter of 7–10 mm was recorded for cervical and inguinal lymph nodes at the age of 1–3 months compared to less than 2 mm in normal age-matched controls (Fig. 2). The size of the lymph nodes increased progressively during the life span of the mice and were macroscopically visible at 1–3 month of age. Twenty transgenic mice showed similar lymph node enlargement identified by macroscopical observation and lymph nodes were allways three- to fivefold enlarged compared to control animals at 3 months of age. To study the pathological changes in the transgenic mice, animals were killed between 2 to 7 months of age. Eosin/haematoxylin staining of spleen and lymph node of the transgenic mice revealed an aberrant microarchitecture. The white pulps of the spleen were no longer clearly separated from the red pulp, as seen in the control littermates (Fig. 3a,B). Similarly, the lymphoid follicles of the lymph nodes were partly absent in the transgenic mice (Fig. 3d) while the follicles were easily distinguishable in the control animals (Fig. 3c).

Figure 2.

The IL-1β transgenic mice developed hyperplasia in lymphoid organs. The lymph nodes grew progressively, the photograph shows 22-week-old mice. Note the massive enlargement of the cervical lymph nodes of the transgenic mice (left) compared to control littermates (right). The diameter of the cervical lymph nodes were 7–10 mm when the mice were killed. Twenty transgenic mice showed similar lymph node enlargement identified by macroscopical observation.

Figure 3.

Altered intra-architecture of spleen and lymph node of the transgenic mice. Sections of paraffin-embedded biopsies were stained with eosin and hematoxylin. The white pulpes of the spleen is clearly visible in the control (indicated by arrows) (a) in contrast to the transgenic mice in which these partly are deformed or absent (b). Similarly, the B-cell follicles of the transgenic lymph nodes (d) are destructed, compared to the lymph-node follicles of the control littermates, indicated by arrows (c). The lymph nodes are here represented by the inguinal lymph nodes. The same findings were seen in the cervical lymph nodes (data not shown). (a,b, bar=500 μm; c,d, bar=300 μm). The results presented are representative of several sections from two different animals.

We subsequently expanded our analysis by immunohistology (Fig. 4). The CD4⁺ T cells in the spleen and in the lymph nodes of the transgenic mice seemed to have a changed pattern of localization. Serial sections of lymph node from the transgenic mice showed a more homogeneous distribution of CD4⁺ T cells through out the lymph node including normally B-cell restricted areas (Fig. 4d), which was not seen in the control lymph nodes (Fig. 4c). The B-cell follicles were less well defined in the huIL-1β transgenic animals (Fig. 4b) as compared to controls (Fig. 4a). However, the B cells still seemed to form identifiable follicular structures and the less distinct pattern could be a secondary effect because of the infiltration of CD4⁺ T cells. Staining for CD8⁺ T cells showed a similar picture in the huIL-1β transgenic mice and controls. Only a minor infiltration of CD8⁺ T cells in B-cell areas was noted in the huIL-1β transgenic animals (data not shown).

Figure 4.

Altered localization of CD4⁺ T cells in lymph node of the transgenic mice. Parallel sections of lymph node from control (a, c) and transgenic (b, d) mice were stained with B220 mAb (a and b) and CD4 mAb (c and d), respectively. (e) Staining of transgenic lymph node with an isotype-matched (IgG2a) control antibody. Staining of control lymph node with isotype-matched control antibody were negative as well (data not shown). The bound antibodies were made visible by use of diaminobenzidine (DAB). Sections were counterstained with methyl green. The positively stained cells are brown. Note that the CD4⁺ T cells in the transgenic organs are intervening the majority of the lymph node including normally B-cell restricted areas which is not seen in the control (a,b, bar = 50 μm; c–e, bar = 100 μm). The results are representative of several sections from three different animals.

Immunohistological staining of the colon of transgenic mice showed infiltration of CD4⁺ T cells (Fig. 5e) and to a lesser extent of CD8⁺ T cells (Fig. 5f). The CD4⁺ T cells were predominantly located in the lymphatic nodules but also between the crypts (Fig. 5e). Few CD8⁺ T cells and only moderate numbers of CD4⁺ T cells were detected in the lymphatic nodules in control mice (Fig. 5b,c). In similarity to the lymph nodes, the colonic B-cell areas were less distinct in huIL-1β transgenic animals (Fig. 5d) than in control animals (Fig. 5a). Thus, lymphoid tissue from huIL-1β transgenic animals displays overt pathological changes in lymphoid tissue, including prominent hyperplasia and CD4⁺ T-cell infiltration.

Figure 5.

Infiltration of T lymphocytes in colon of the transgenic mice. Parallel sections of colon from control (a–c) and transgenic (d–f) mice were stained with B220 mAb (a and d), CD4 mAb (b and e) and CD8 mAb (c and f), respectively. (g) Staining of transgenic colon with an isotype-matched (IgG2a) control antibody. Staining of control lymph node with isotype-matched control antibody were completely negative as well (data not shown). The bound antibodies were made visible by use of diaminobenzidine (DAB). Sections were counterstained with methyl green. The positively stained cells are brown. ((a–g) Scale bar 100 μm.) The results are representative of several sections from two different animals.

Phenotypic characterization of spleen and lymph node cells

To further analyse the cell composition of the hyperplastic lymphoid organs, we prepared single-cell suspensions from spleen and lymph nodes and performed flow cytometric analyses. Dual staining of spleen and lymph node cells with CD4 mAb and CD62L (MEL-14) or CD44 mAb showed that the large majority of the CD4⁺ T cells in huIL-1β transgenic mice had downregulated CD62L and expressed increased amounts of CD44 (Fig. 6e–h, respectively). This indicates that the CD4⁺ T cells in the transgenic mice were activated. In contrast, a low frequency of CD62L negative cells and low levels of CD44^high cells were detected in normal lymph nodes, and moderate numbers were detected in the spleen of normal mice (Fig. 6a–d, respectively). Moreover, the total number of B and T cells in the lymph nodes of huIL-1β transgenic mice were greatly increased, although the frequency of T cells were decreased (Fig. 6i). Hence, both by surface marker expression and cell numbers our findings were compatible with a significant hyperplasia in lymph nodes from huIL-1β transgenic mice.

Figure 6.

The majority of the CD4⁺ T cells present in spleen and lymph node of the transgenic mice were activated. Single-cell suspensions of spleens and lymph nodes were prepared and were analysed by FACS. The cells were double-stained with CD4 mAb and either CD62L or CD44 mAb. Note the downregulation of CD62L (e and g) and the upregulation of CD44 (f and h) on CD4 ⁺ T cells in both spleen and lymph node of the transgenic mice, compared to control (a and c) and (b and d), respectively. One representative experiment of two performed. (i) Total number of CD4⁺, CD8⁺ and B220⁺ cells in transgenic cervical lymph nodes were highly increased compared to wild type (wt), compatible with lymphoid hyperplasia.

Induction of T-cell derived growth factors in lymphoid organs

IL-1 is known to induce several T-cell related growth factors, including IL-2, IL-4 and IL-6,²⁰ and because the CD4⁺ T cells seemed to have a changed pattern of migration and displayed an activated phenotype, frozen sections of the spleens and lymph nodes were stained for expression of a number of cytokines. Lymph nodes of the IL-1β transgenic mice contained significant levels of IL-4 and high levels of IL-2-positive cells (Fig. 7c,d, respectively). In contrast, control lymph nodes lacked or contained very low numbers of IL-4 and IL-2 producing cells (Fig. 7a,b, respectively). Similarly, an increased frequency of IL-4- and IL-2-producing cells were seen in transgenic spleens compared to control animals (data not shown). In contrast, no significant differences in expression of other cytokines (including interferon-γ (IFN-γ), TNF-α and IL-6) in lymph nodes and spleens were observed between control and transgenic mice (data not shown).

Figure 7.

Cytokine expression in lymph nodes of transgenic mice. Sections of lymph node from control (a,b) and transgenic (c,d) mice were stained with IL-4 mAb (4 μg/ml) (a and c) and IL-2 mAb (1 μg/ml) (b and d), respectively. The bound antibodies were made visible by use of diaminobenzidine (DAB). Stainings with isotype-matched control antibodies to anti-IL-4 and anti-IL-2 were negative (data not shown). Sections were counterstained with methyl green. The positively stained cells are brown. (Scale bar 100 μm.) The results are representative of several sections from three different animals.

Increase in serum immunoglobulin of huIL-1β transgenic mice

To investigate whether the hyperactivity of the CD4⁺ T cells in huIL-1β transgenic mice also had a functional impact on B cells we measured serum immunoglobulin levels in huIL-1β transgenic mice by isotype-specific ELISA. A profound increase in both IgM and IgG serum levels was recorded in the transgenic mice compared to control animals (Table 1). The most significant increase was seen in the IgG1 isoforms, with 30-fold elevated levels compared to controls. In contrast, the levels of IgG2b and IgA were only slightly increased. The data suggest that the CD4⁺ T cells that are expanded in huIL-1β transgenic mice functionally contribute to activate B-cell immunoglobulin secretion.

Table 1.

Relative levels of isotype-specific immunoglobulins in sera of IL-1β transgenic mice

DISCUSSION

In the present study we have studied the biological effects of extracellular release of IL-1β in a lymphoid microenvironment. The huIL-1β transgenic mice developed a number of abnormal features such as hyperplasia, aberrant architecture of lymphoid organs, hypergammaglobulinaemia, appearance of activated CD4⁺ T cells and spontaneous secretion of IL-2 and IL-4 in lymphoid organs. It is previously known that IL-1 has the ability to enhance T-cell dependent antibody production in part by increasing T helper cell activity,²¹ including enhanced IL-2 production²² and IL-4 production.²³

Only a low frequency of transgene-positive offspring from the first founder mouse developed the characteristic phenotype described herein. This was most likely a result of the design of the transgenic construct as well as the integration site, because our second transgenic construct that contained the intact immunoglobulin heavy chain intron enhancer gave a higher penetrance of the phenotype in two separate founders. We are currently investigating various means to provoke the onset of phenotype in the low penetrance transgenic mouse by immune or pharmacological assault which could generate a useful model for the study of the establishment of lymphoid hyperplasia.

Immunohistological analyses of spleens and lymph nodes from the transgenic mice showed an altered architecture with partly absent white pulps in the spleen and B-cell follicles in the lymph node, respectively. A changed distribution of CD4⁺ T cells in huIL-1β transgenics compared to the control animals was evident. The transgenic CD4⁺ T cells seemed to invade most of the lymphoid tissue and to be present also in normally B-cell restricted areas. Earlier findings have suggested that IL-1β is a potent chemotactic factor that is relatively specific for helper T cells.²⁴ Whether IL-1β acts directly or indirectly as a chemotactic factor, by induction of certain T helper specific chemotactic proteins, remains to be determined. Analysis of spleen and lymph node cells from the transgenic mice revealed that the majority of the CD4⁺ T cells had down-regulated CD62L (l-selectin) and upregulated CD44. Because downregulation of CD62L and upregulation of CD44 is a hallmark of activated T cells, it is conceivably that the transgenic CD4⁺ T cells are in an activated state. In contrast, the transgenic CD8⁺ T cells did not show an significant altered distribution compared to control animals (data not shown). This is in agreement with earlier studies showing that IL-1R expression and IL-1 responsiveness is seen primarily in CD4⁺ and not in CD8⁺ T cells.²⁵ We have in an earlier study demonstrated that release of the genetically modified huIL-1β locally in a tumour area, has the potential of converting an aggressive, poorly immunogenic tumour into a less tumorigenic form. Part of this antitumour effect seemed to correlate with an infiltration of CD4⁺ T but not CD8⁺ T cells into the tumour area,¹⁶ indicating an important role for IL-1β in the migration of CD4⁺ T cells. Thus, taken collectively, IL-1β seems to function as an activation signal and possibly also as a chemotactic factor for CD4⁺ T cells. Studies in TNF and LT deficient mice have shown a disturbed B-cell migration from outer periarteriolar lymphoid sheath (PALS) towards follicles in spleen and disordered T/B-cell organization. Furthermore, these cytokines are indispensable for the formation of germinal centers.^26–28 Although the molecular mechanisms responsible for the development of normal microarchitecture in lymphoid tissues, including development of primary and secondary (germinal centres) lymphoid follicles as well as distinct T- and B-cell areas, are still elusive, it is likely that they involve regulated expression of various cytokines. It is tempting to speculate that a fine-tuned homeostasis between different inflammatory cytokines, including IL-1β, TNF-α and LT-α, as well as adhesion molecules is crucial in order to achieve a proper migration of lymphoid cells. Thus, inflammatory cytokines may control homeostasis of the microenvironment in a physiological immune responses and influence tissue and cell remodelling during chronic pathological inflammatory responses.

Immunohistological staining of the colon revealed a prominent infiltration of CD4⁺ T cells and to a lesser extent of CD8⁺ T cells compared to controls. The infiltrating cells were predominantly situated in the lymphatic nodules but also in between the crypts. This is in agreement with earlier findings showing that overproduction of IL-1 in the colonic mucosa results in infiltration of predominantly CD4⁺ T cells and influence the early stages of development of chronic colitis.²⁹

Analysis of the immunoglobulin titre in sera of the transgenic mice revealed an enhanced production of several immunoglobulin classes, but most pronounced the IgG1 isotype. IL-1β is known to induce gene expression of a number of cytokines such as IL-2, IL-4, IL-6 and TNF.¹ These cytokines all possess the capability of mediating immunoglobulin class-switch from IgM to IgG. Immunohistochemical analysis of the lymph nodes showed induction of both IL-2 and IL-4, and because there is no evidence for IL-1 to directly induce class-switch to IgG1, it is a reasonable assumption that the IgG1 levels were a secondary effect of elevated IL-4 production. This is in agreement with earlier reports showing that IL-1 increases IgG production, in part, by stimulating T-cell helper activity.^21,30 In contrast, we were not able to detect any elevated levels of IL-6 either in sera, spleens or lymph nodes. This was unexpected because IL-1 is a known IL-6-inducer and since several of the biological effects in our transgenic mice, such as splenomegaly, lymph node enlargement and hypergammaglobulinaemia are similar to those seen in IL-6 transgenic mice.³¹ Thus, IL-1 and IL-6 may both have the ability to directly or indirectly regulate lymphocyte expansion and homeostasis.

Is the lymphoid hyperplasia as well as hyperimmunoglobulinemia seen in the transgenic mice induced by the activated T cells? Several factors suggest that this may be the fact. The transgenic mice expressed significant levels of both IL-4 and IL-2 in the lymph nodes, which are known growth factors for B cells. The immunoglobulins in sera were predominantly of IgG isotypes, implicating class switch, which is a T-cell dependent event. Hence, it makes it tempting to speculate that the activated CD4⁺ T cells are the key mediator of several pathological features seen in the transgenic mice, including altered lymphoid architecture, lymph node hyperplasia and hyper gammaglobulinemia. Cross-breeding between IL-1β transgenic mice and T-cell deficient strains, such as CD4 knock-out mice, is presently performed to address this issue.

In conclusion, IL-1β seems to play an important role in regulating lymphoid microarchitecture and migration of CD4⁺ T cells in lymphoid tissues. In spite of the existence of several control mechanisms for IL-1β expression, including transcriptional and translational regulation, neutralizing IL-1 type II decoy receptors and IL-1 receptor antagonists,^1–3 this study shows that the protective mechanisms can be bypassed by overexpression of a secreted IL-1β protein. Various diseases such as systemic lupus erythematosus (SLE),³² rheumatoid arthritis (RA),^33,34 and chronic colitis²⁹ are associated with upregulation of IL-1β gene expression which has been suggested to contribute to the chronic inflammatory response. In chronic conditions the elevated level of IL-1β expression seems to be capable of breaking the threshold in T helper cell–B-cell interactions leading to pathological conditions such as unregulated proliferation and activation of lymphocytes. This makes IL-1 an obvious target for pharmacological intervention in inflammatory disorders. Our huIL-1β transgenic mice may be an useful tool when studying the role of IL-1β in pathological conditions such as chronic inflammatory and immunoproliferative disorders.

Abbreviations

ra

receptor antagonist

ICE

interleukin-1β converting enzyme

ced-3

C. elegans cell death gene-3

ssIL-1β

signal sequence IL-1β

DAB

diaminobenzidine

SLE

systemic lupus erythrematosus

RA

rheumatoid arthritis