Overexpression of Frat1 in transgenic mice leads to glomerulosclerosis and nephrotic syndrome, and provides direct evidence for the involvement of Frat1 in lymphoma progression

Article metrics


The proto-oncogene Frat1 was originally identified as a common site of proviral insertion in transplanted tumors of Moloney murine leukemia virus (M-MuLV)-infected Eμ-Pim1 transgenic mice. Contrary to most common insertion sites implicated in mouse T cell lymphomagenesis, retroviral insertional mutagenesis of Frat1 constitutes a relatively late event in M-MuLV-induced tumor development, suggesting that proviral activation of Frat1 contributes to progression of T cell lymphomas rather than their genesis. To substantiate this notion we have generated transgenic mice that overexpress Frat1 in various organs, including lymphoid tissues. Frat1 transgenic mice develop focal glomerulosclerosis and a nephrotic syndrome, but they do not exhibit an increased incidence of spontaneous lymphomas. Conversely, these mice are highly susceptible to M-MuLV-induced lymphomagenesis, and Frat1/Pim1 bitransgenic animals develop lymphomas with increased frequency compared to Pim1 transgenic littermates. These data support a role for Frat1 in tumor progression.


Cancer is a complex process in which a normal cell has to accumulate multiple genetic and epigenetic alterations in order to become a fully transformed metastasizing tumor cell (Bishop, 1991; Hunter, 1991). Compared to other tissues, lymphoid cells with intrinsic migratory capacity may require fewer steps to become highly malignant tumor cells. Still, such steps are likely to include enhanced proliferative capacity, augmented growth factor independence, and progression towards escape from terminal differentiation. In mouse T cell lymphomas, several proto-oncogenes that are implicated in early stages of transformation have been identified, most notably the Gfi1 gene and members of the Myc and Pim gene families (Berns et al., 1994). In contrast, only a limited number of genes have been isolated that may be important for lymphoma progression in vivo (Jonkers and Berns, 1996). Recently, we have described the identification of a novel gene, Frat1, that is primarily involved in progression of mouse T cell lymphomas (Jonkers et al., 1997). Activation of Frat1 via retroviral insertions is preferentially detected in transplanted tumors of M-MuLV-infected Eμ-Pim1 or H2-K-Myc transgenic mice. The initial selection for activation of Frat1 occurs in primary tumor cells that have already acquired proviral insertions near other proto-oncogenes, resulting in primary lymphomas that contain only a minor fraction of tumor cells with integrations near Frat1. Transplantation of these lymphomas allows for a further expansion of tumor cell clones carrying a mutated Frat1 allele, resulting in detectable Frat1 rearrangements in 17% of the transplanted Eμ-Pim1 tumors and 30% of the transplanted H2-K-Myc tumors, respectively. Integrations in Frat1 cluster in a small genomic region of approximately 500 bp, resulting in the enhanced expression of truncated Frat1 transcripts which lack most of the 3′ untranslated region contained in normal Frat1 mRNA. The nucleotide sequence of the Frat1 cDNA encodes a polypeptide of 274 amino acid residues. Recently, a Xenopus GSK-3 binding protein, named GBP, was found to be homologous to Frat1 in three well-conserved regions (Yost et al., 1998). Both GBP and human FRAT2 were shown to be potent inhibitors of GSK3 kinase activity, leading to accumulation of β-catenin and activation of TCF/β-catenin mediated signal transduction in Xenopus. Importantly, deletion of GBP from the maternal mRNA pool showed that GBP is required for establishment of the dorsal-ventral axis in Xenopus embryos (Yost et al., 1998). In lymphomas with a provirally activated Frat1 allele, the protein-encoding domain of Frat1 is never disrupted by the proviral insertions, suggesting that overexpression of normal Frat1 contributes to transformation. Further support for a role of Frat1 in tumor progression came from transplantation experiments with cell lines derived from spontaneous lymphomas in Eμ-Pim1 transgenic mice. These cell lines, which already overexpress Myc and Pim1, gained an additional selective advantage in vivo upon transduction with a Frat1-IRES-lacZ retrovirus (Jonkers et al., 1997).

The oncogenicity of genes implicated in early stages of tumorigenesis can be tested effectively by monitoring the frequency of spontaneous tumors in transgenic mice that overexpress the putative oncogene in the relevant tissues. Moreover, transgenic mouse models have become an invaluable tool to study oncogene cooperation, as tumorigenesis in these animals may be accelerated by introducing another trans-oncogene or by using a retrovirus as an insertional mutagen. Illustrative examples are Eμ-Myc transgenic mice, which are predisposed to pre-B cell lymphomas (Adams et al., 1985), and Eμ-Pim1 transgenic mice, which develop spontaneous T cell lymphomas with low frequency (van Lohuizen et al., 1989). Proviral tagging experiments in these animals (van Lohuizen et al., 1989, 1991; Haupt et al., 1991), or breeding of Eμ-Myc/Eμ-Pim1 doubly transgenic mice (Verbeek et al., 1991) revealed the capacity of both genes to synergize in the induction of lymphomagenesis.

Retroviral activation of the Frat1 tumor progression gene was found to be invariably preceded by the activation of at least two other loci, suggesting that this gene can only confer an additional selective advantage to cells that are already (partially) transformed (Jonkers et al., 1997). Therefore, overexpression of Frat1 in transgenic animals should have little or no effect on the incidence of spontaneous lymphomas. However, it is conceivable that such mice develop tumors faster when the initiating mutations are introduced via crosses with other oncogene-bearing transgenic mice, or by retroviral infection. We show here that this is indeed the case. The data demonstrate collaboration between Frat1 and Pim1 in lymphomagenesis, and an increased susceptibility of Frat1 transgenic mice to M-MuLV-induced T cell lymphomagenesis. We also report here on the occurrence of a nephrotic syndrome in Frat1 transgenic mice, characterized by proteinuria due to a glomerular sieving defect, and glomerular lesions featuring focal and segmental sclerosis. Results from bone marrow transplantation experiments show that overexpression of Frat1 in the hematopoietic compartment alone is insufficient for the induction of the nephrotic syndrome.


Eμ-pp-Frat1 transgenic mice constitutively express high levels of Frat1

Eμ-pp-Frat1 transgenic mice were generated that carried the mouse Frat1 gene under the control of the Pim1 promoter (pp), two copies of the immunoglobulin heavy chain enhancer (Eμ) and the Moloney murine leukemia virus (M-MuLV) long terminal repeat (LTR) (Figure 1a). This promoter/enhancer combination has been shown to give high transgene expression in both the B and T cell lineage (van Lohuizen et al., 1989; Renauld et al., 1994; Alkema et al., 1997). The Eμ-pp-Frat1 transgene lacks the Frat1 3′ untranslated region containing putative mRNA destabilizing motifs, and expression of Eμ-pp-Frat1 should result in the production of truncated Frat1 mRNA, analogous to the transcripts found in lymphomas with a provirally activated Frat1 allele. Three founders were obtained that transmitted the transgene through the germ line. Eμ-pp-Frat1 line 29 (EppF29) harbored approximately 30 copies of the transgene, whereas strains EppF5 and EppF8 contained two to three copies and one copy, respectively. Transgenic lines were established and examined for expression levels of the transgene. Northern analysis of total RNA, isolated from different organs, showed that all three Eμ-pp-Frat1 strains expressed the transgene in all organs tested (Figure 1b). A similar broad expression pattern has been reported previously for IL-9 transgenic mice that carried the murine II9 gene under the control of the same promoter and enhancer elements (Renauld et al., 1994). The level of transgene expression varied between the three different Eμ-pp-Frat1 lines, and increasing amounts of Frat1 mRNA were found in strain 29, 8 and 5, respectively.

Figure 1

Structure and expression of the Eμ-pp-Frat1 transgene (a) Structure of the Eμ-pp-Frat1 transgenic construct. The Frat1 sequences present in the transgene construct correspond to a 920 bp genomic SrfI – XhoI fragment encompassing the Frat1 coding domain, but lacking most of the 5′ and 3′ untranslated regions. The 2.6 kb Pim1 promoter fragment, represented by a thick line, contains the SP-1 and octamer recognition sites and the Cap site. The open boxes represent the duplicated immunoglobulin heavy chain enhancer (Eμ) and the M-MuLV long terminal repeat (LTR). (b) Expression of the transgene in founder lines EppF29, EppF5 and EppF8. A Northern blot containing equal amounts (20 μg) of total RNA extracted from various tissues was hybridized with the Frat1 ORF probe 0.45Pst and an actin probe as a control for integrity and transfer of the RNA. The 28S and 18S ribosomal bands are indicated as size markers. Since in most tissues the endogenous 2.5 kb Frat1 mRNA is present at very low levels, only the 1.7 kb transgenic transcript is visible. The faint actin signal in the lanes containing total RNA from liver is not due to loading differences, but to the low levels of actin mRNA in liver

Eμ-pp-Frat1 transgenic mice are not predisposed to development of spontaneous lymphomas

No significant differences in tumor incidence were observed between Eμ-pp-Frat1 and control mice. Thus far, no spontaneous lymphomas were found in EppF29 mice, even after monitoring relatively large numbers of animals (n=88) during a period of 600 days. Of 40 EppF5 females that were monitored for tumors during a 300 day period, only one animal developed a T cell lymphoma (TCRαβ+CD3+CD4+CD8) after 208 days. In line with this, histological examination revealed no obvious morphological changes in the lymphoid tissues of Eμ-pp-Frat1 transgenic mice. To determine whether specific cell populations were affected in Eμ-pp-Frat1 transgenic mice, flow cytometric (FACS) analysis of bone marrow cells, thymocytes and splenocytes was performed using standard lymphoid and myeloid markers (see Materials and methods). No significant variations could be detected in any of the hemopoietic cell populations that were tested (data not shown). Also the responses of splenocytes to the B cell mitogen lipopolysaccharide (LPS) or the T cell mitogen concanavalin A (conA), as determined by [3H]thymidine incorporation assays, were similar to responses of normal spleen cells. Finally, the responses of Eμ-pp-Frat1 thymocytes to various apoptotic stimuli (see Materials and methods), and the percentage of conA-stimulated splenocytes undergoing apoptosis upon withdrawal of IL-2, were comparable to responses of control cells (data not shown).

Accelerated onset of lymphomagenesis in Eμ-pp-Frat1/Eμ-Pim1 doubly transgenic mice

The Frat1 gene is frequently activated by proviral insertions in the transplanted tumors of M-MuLV-infected Eμ-Pim1 transgenic mice, suggesting cooperation between Frat1 and Pim1 in T cell lymphomagenesis. To investigate this possibility, we crossed EppF29 and Eμ-Pim1/FVB transgenic mice to produce doubly transgenic animals. The Eμ-Pim1/FVB strain used in this study was generated by microinjection of the Eμ-Pim1 transgene construct (van Lohuizen et al., 1989) into fertilized oocytes from FVB mice (J Domen, unpublished data). Eμ-Pim1/FVB transgenic mice showed a low incidence of spontaneously occurring T cell lymphomas, 26% within 1 year. This frequency is somewhat higher than the previously observed frequencies in Eμ-Pim1 transgenic lines of mixed genetic background (van Lohuizen et al., 1989; Verbeek et al., 1991). The use of inbred transgenic lines allowed us to measure the synergistic effect of both transgenes on a uniform genetic background. EppF29/Eμ-Pim1 double-transgenic mice and Eμ-Pim1 animals were monitored for spontaneous tumors during a period of 180 days. As shown in Figure 2, the incidence of spontaneously occurring tumors in Eμ-pp-Frat1/Eμ-Pim1 double-transgenic mice was significantly higher than in Eμ-Pim1 single-transgenic animals (Log-Rank P=0.0017). Lymphomas of five EppF29/Eμ-Pim1 bitransgenic mice and six Eμ-Pim1 animals were analysed by flow cytometry for lineage specific markers. Whereas all six Eμ-Pim1 lymphomas were of the immature TCRαβ+CD4+CD8+ type, the EppF29/Eμ-Pim1 tumors were either TCRαβ+CD4+CD8+ (two out of five tumors) or TCRαβ+CD4+CD8 (3/5). Northern blot analysis of the tumors showed that both transgenes were expressed at high levels in all lymphomas of doubly transgenic mice (data not shown). These results demonstrate that Frat1 and Pim1 can effectively collaborate in lymphomagenesis. Similar to what has been reported on spontaneous lymphomas of Eμ-Pim1 transgenic mice (van Lohuizen et al., 1989), all tumors of bitransgenic animals expressed high levels of Myc mRNA (data not shown).

Figure 2

Incidence of spontaneously occurring lymphomas in Eμ-Pim1 single-transgenic and Eμ-pp-Frat1/Eμ-Pim1 double-transgenic mice. The vertical axis represents the percentage of surviving mice. The horizontal axis represents the age of the mice in days. EppF29/Eμ-Pim1 double-transgenic mice and Eμ-Pim1 littermates were monitored for spontaneous tumors during a period of 180 days, after which the experiment was terminated. The incidence of spontaneous tumors in Eμ-pp-Frat1/Eμ-Pim1 double-transgenic mice was significantly higher than in Eμ-Pim1 single-transgenic animals of the same genetic background (Log-Rank P=0.0017)

High susceptibility of Eμ-pp-Frat1 transgenic mice to M-MuLV-induced T cell lymphomagenesis

In M-MuLV-induced T cell lymphomas, activating proviral insertions in Frat1 occur only in tumor cells that have already acquired two or more oncogenic mutations, suggesting that the selective advantage of Frat1 is limited to lymphoma progression (Jonkers et al., 1997). This implies that Eμ-pp-Frat1 transgenic mice, although they are not predisposed to spontaneously occurring tumors, might develop tumors with increased frequency when the initiating oncogenic lesions are induced by retorviral or chemical agents. We sought to examine the susceptibility of Eμ-pp-Frat1 transgenic mice and non-transgenic littermates to M-MuLV-induced T cell lymphomagenesis. For this purpose, newborn mice from matings between normal FVB and Eμ-pp-Frat1 (strains EppF29 and EppF8) were infected with M-MuLV. Mice were sacrificed when moribund, and the T cell origin of the tumors was confirmed by FACS analysis. Increasing levels of Frat1 expression correlated with an increasing proportion of TCRαβ+CD4+CD8 lymphomas (wild type FVB: four out of nine tumors; EppF29: 9/11; EppF8: 15/15). The remaining tumors were either TCRαβ+CD4CD8 (FVB: 2/9; EppF29: 2/11), or TCRαβ+CD4+CD8+ (FVB: 3/9). As shown in Figure 3a, Eμ-pp-Frat1 transgenic mice developed lymphomas after a significantly shorter latency period than the non-transgenic littermates (Log-Rank P=0.019 for EppF29 vs FVB; P=0.00003 for EppF8 vs FVB). In agreement with the observed difference in transgene expression, the mean latency of lymphoma onset was shorter in EppF8 mice than in EppF29 animals. To determine whether known oncogenes were activated by proviral insertion, DNA samples from tumor tissues (spleens or mesenteric lymph nodes) were subjected to Southern blot analysis with probes specific for Myc, Nmyc1, Pim1, Pim2, Pal1 and the M-MuLV LTR. A summary of the percentage of common proviral insertions is shown in Table 1. The fraction of Myc, Ncmyc1 and Pal1 rearrangements in the tumors of M-MuLV-infected Eμ-pp-Frat1 transgenic mice was comparable to those observed in the M-MuLV-induced lymphomas of non-transgenic littermates and similar to percentages previously reported for the FVB genetic background (van der Lugt et al., 1995). Surprisingly, proviral insertions near Pim1 or Pim2 were observed in 37% of the FVB tumors, in 21% of the EppF29 tumors, and in none of the EppF8 tumors (P=0.0009 for EppF8 vs FVB). Northern blot analysis of tumor RNA revealed elevated levels of normal-sized Pim1 mRNA in all EppF8 lymphomas (Figure 3b), suggesting that in these tumors Pim1 was activated in trans, or by proviral enhancer insertions located outside the 22 kb genomic Pim1 region that was examined by Southern analysis. Enhanced Pim1 expression in the absence of detectable proviral integrations in the Pim1 locus was also observed in several EppF29 tumors, as well as in tumors from FVB control littermates, showing that this phenomenon can also occur in the absence of Frat1 overexpression (Figure 3b).

Figure 3

Acceleration of T cell lymphomagenesis in M-MuLV infected Eμ-pp-Frat1 transgenic mice. (a) Survival curves for neonatally infected FVB mice and for transgenic mice from line EppF29 and line EppF8. The percentage of surviving mice is plotted against the age of the animals. Mice were killed when moribund and lymphomas were collected. Eμ-pp-Frat1 transgenic mice developed M-MuLV induced lymphomas after a significantly shorter latency period than the non-transgenic littermates (Log-Rank P=0.019 for EppF29 vs FVB; P=0.00003 for EppF8 vs FVB). (b) Expression of Frat1, Myc, Nmyc1, Pim1, and Pim2 in M-MuLV-induced lymphomas of FVB, EppF29, and EppF8 mice. A Northern blot containing 20 μg of total RNA extracted from tumor tissue, was hybridized subsequently with Frat1, Myc, Nmyc1, Pim1, and Pim2 probes. The same Northern blot was also hybridized with an actin probe as a control for loading and transfer. Ribosomal bands are indicated as size markers. In tumors X26, X18 and X30, proviral insertions in Pim1 were detected

Table 1 Proviral integration sites in M-MuLV-induced T cell lymphomas

Nephrotic syndrome in Frat1 transgenic mice

While Eμ-pp-Frat1 transgenic mice were not found to be tumor-prone, all male animals of strains EppF5 and EppF8 succumbed to renal failure after a variable latency period (Figure 4a). A mild nephrotic syndrome was also observed in EppF29 males (Table 2). The occurrence of a nephrotic syndrome in all three Eμ-pp-Frat1 transgenic lines, and the correlation of the disease phenotype with the expression level of the transgene demonstrate that the mutant phenotype is caused by deregulated Frat1 expression, and not by the interruption of an endogenous gene at the transgene integration site. The nephrotic syndrome in Frat1 transgenic mice was characterized by proteinuria and generalized edema. In male mice of strains EppF5 and EppF8, proteinuria was first observed between day 70 and 110 (Figure 4b). In these mice, massive proteinuria associated with a full blown nephrotic syndrome had developed by day 150. In contrast female EppF5 and EppF8 mice had markedly milder lesions, starting at a later age with pathology at day 400 resembling the changes observed in male mice at day 150 (Figure 4a and Table 2). In most mice with end-stage disease, macroscopically visible protein aggregates were present in the bladder. Blood serum analysis of animals with end-stage disease revealed elevated levels of cholesterol and low-density lipoproteins (hyperlipidemia) and hypoalbuminemia due to proteinuria. These abnormalities are together with generalized edema indicative of nephrotic syndrome in man.

Figure 4

Nephrotic syndrome in Eμ-pp-Frat1 transgenic mice (a) Survival of Eμ-pp-Frat1 transgenic mice. The vertical axis represents the percentage of surviving mice. The horizontal axis represents the age of the mice in days. Mice with end-stage renal disease were sacrificed and kidneys, urine, and blood serum samples were collected. The survival curves for EppF5 males, EppF5 females and EppF8 males are shown. (b) Progressive proteinuria in EppF5 transgenic males. The vertical axis represents the total amount of albumin present in urine secreted during a period of 48 h. The age of the mice at the time of urine-sampling is indicated at the horizontal axis. Each bar represents a single animal of the specified genotype. The absence of a bar at a given time point indicates that the animal has died of nephrosis. At the age of 50 days, no differences in the amounts of urine-secreted albumin were found between EppF5 transgenic males and wild type littermates (not shown)

Table 2 Nephrotic syndrome in EppF29 males and EppF8 females

Histological examination of the kidneys of diseased mice revealed focal and segmental glomerulosclerosis with hyalinosis. As shown in Figure 5c and d, these lesions are characterized by mesangial proliferation, increase of extracellular matrix, and segmental collapse of capillaries associated with cellular and fibrous adhesions to Bowman's capsule. Extensive interstitital fibrosis and inflammation was present in areas of tubular atrophy characterized by pseudo-thyroid degeneration. Immunofluorescence microscopy revealed no abnormalities, in particular no immune deposits. Electron microscopy revealed widespread, but not diffuse obliteration of visceral epithelial cell foot processes with microvillous transformation and vacuolization of the podocytes (Figure 5e and f). Endothelial cells were normal. The glomerular basement membrane (GBM) showed irregular thickness with local spike-like protrusions of the GBM. No electrondense deposits that are characteristic for immune complexes were found.

Figure 5

Renal histological features of Eμ-pp-Frat1 transgenic mice with nephrotic syndrome. (a – d) Light microscopic characteristics of representative sections of control kidneys (a and b) and of diseased Eμ-pp-Frat1 kidneys (c and d). (c) Overview of corticomedullary border of severely affected kidney, showing focal segmental hyalinosis and glomerulosclerosis, and dilated tubules with proteinaceous casts. (d) Glomerulus showing segmental capillary collapse together with mesangial expansion and adhesion to Bowman's capsule. (e – f) Renal ultrastructure of Eμ-pp-Frat1 transgenic mice with nephrotic syndrome, showing widespread, but not diffuse obliteration of visceral epithelial cell foot processes with microvillous transformation of the podocytes. (a and c: PAS, ×100; b and d: PAS, ×630; e: ×1000; f: ×3000)

The immune system, and disturbed T lymphocyte function in particular, has previously been implicated in the pathogenesis of several kidney disorders (Klein, 1993; Couser, 1994). The presence of protein deposits in the tubules of diseased kidneys suggested that the nephrotic syndrome in Eμ-pp-Frat1 transgenic mice might be caused by a systemic autoimmune disease resulting in increased serum levels of specific auto-antibodies. To test this hypothesis, serum was collected from 4 months old EppF5 males (n=8) and from animals with end-stage disease (n=19). All sera were screened by indirect immunofluorescence for anti-nuclear antibodies (ANA) on Hep2 cells, and for anti ds-DNA antibodies on Crithidia luciliae, but in no instance positive fluorescence was observed. Likewise, no antibodies against components of the glomerular basement membrane (GBM) or the glomerular epithelial cell (GEC) membrane could be detected in sera from nephrotic Eμ-pp-Frat1 mice.

To determine whether overexpression of Frat1 in the hematopoietic compartment is sufficient for induction of the nephrotic syndrome, long term bone marrow transplantation experiments with bone marrow cells derived from EppF5 transgenic males were performed. The levels of urine-secreted albumin from the resulting bone marrow chimeras were monitored over a period of 10 months following transplantation, but no differences were found between animals reconstituted with EppF5 bone marrow (n=5) and mice reconstituted with FVB/N bone marrow cells (n=5).


Lymphomagenesis in Frat1 transgenic mice

To explore the oncogenic potential of Frat1 and to study cooperative interactions with other oncogenes, we have generated three independent transgenic mouse lines in which Frat1 is expressed at different levels in several organs, including hematopoietic tissues. Thus far, no mice of the EppF29 or EppF8 lines have developed any spontaneous lymphomas. One out of 40 females of the EppF5 line succumbed from lymphoma at the age of 208 days, but it is not clear whether this was due to the transgene expression, or to the low incidence of spontaneously occurring lymphomas in FVB mice. The observation that Eμ-pp-Frat1 transgenic mice are not markedly predisposed to the development of spontaneous T cell lymphomas underscores the notion that the role of Frat1 in lymphomagenesis is restricted to tumor progression. In line with this, the absolute and relative sizes of the major hematopoietic cell compartments of Eμ-pp-Frat1 animals were found to be normal, and also no abnormalities have been recognized in the responses of various lymphocyte populations to mitogenic or apoptotic stimuli.

Direct evidence for a role of Frat1 in lymphomagenesis was obtained by intercrossing Eμ-pp-Frat1 and Eμ-Pim1 transgenic mice. The offspring developed spontaneous lymphomas significantly faster than Eμ-Pim1 single-transgenic animals. The elevated Myc expression in all tumors from bitransgenic animals is in concordance with the concerted proviral activation of Myc and Frat1 in several transplanted lymphomas from M-MuLV-infected Eμ-Pim1 transgenic mice (Jonkers et al., 1997), and provides further evidence for the collaborative action of Pim1, Myc and Frat1 in lymphomagenesis.

The requirement for transforming events other than overexpression of Frat1 was indicated by the absence of spontaneous lymphomas in Eμ-pp-Frat1 transgenic mice, and was further confirmed by the increased incidence of spontaneous lymphomas in Eμ-pp-Frat1/Eμ-Pim1 double-transgenic animals. To probe this point further, we employed proviral tagging to determine whether Eμ-pp-Frat1 transgenic mice are more susceptible to M-MuLV-induced lymphomagenesis, and to establish which genes collaborate with Frat1 in virus-induced malignancy. We found that the average latency period of lymphoma development was reduced from 81 days for M-MuLV-infected wild type FVB mice to 66 days for EppF29 mice, and to 45 days for EppF8 animals. This reduction in latency correlates well with higher expression levels of Frat1 in the EppF8 line, compared to the EppF29 line.

The percentage of proviral insertions near Myc/Nmyc1 and Pal1 in the virus-induced Eμ-pp-Frat1 tumors was found to be comparable to the frequencies observed in control littermates, indicating that Frat1 can effectively cooperate with these genes in lymphoma development. However, the fraction of Pim1/Pim2 rearrangements was markedly reduced, and no integrations in these loci were detected in tumors from the EppF8 line. All EppF8 lymphomas expressed nevertheless high levels of normal-sized Pim1 mRNA, suggesting that activation in trans or cis-activation by distant proviral enhancer insertions had occurred. Although the correlation between this type of Pim1 activation and Frat1 overexpression is absolute within the EppF8 tumor panel, some EppF29 tumors with high Frat1 expression showed no augmented expression of Pim1. It is therefore unlikely that overexpression of Frat1 induces Pim1 in a direct fashion. An alternative explanation for this phenomenon might be that overexpression of Frat1 predisposes to a specific class of lymphomas in which Pim1 is induced by epigenetic events, since increasing levels of Frat1 expression were found to result in an increasing proportion of TCRαβ+CD4+CD8 lymphomas of the mature helper T cell class. Also this correlation is however imperfect, since most but not all TCRαβ+CD4+CD8 lymphomas showed augmented Pim1 expression, and conversely, two lymphomas of a different class expressed Pim1 mRNA in the absence of detectable Pim1 integrations.

Nephrotic syndrome in Frat1 transgenic mice

With aging, all Eμ-pp-Frat1 transgenic mice developed a nephrotic syndrome, characterized by proteinuria, hypoalbuminemia, hyperlipidemia, and ultimately generalized edema. From a morphological point, diseased mice developed glomerular lesions characteristic of focal segmental glomerulosclerosis with hyalinosis and extensive tubular degeneration. Based on these histopathological features, the nephropathy in Eμ-pp-Frat1 transgenic mice can be classified as focal and segmental glomerular sclerosis (FGS) (Weening et al., 1986). The occurrence of a nephrotic syndrome was observed in all three transgenic lines, and its progression-rate correlated with the level of transgene expression, demonstrating that this disease is a direct consequence of the transgene expression, and not due to interruption of an endogenous gene at the transgene integration site. The average latency period was significantly shorter for transgenic males than for female littermates. This difference might be caused by the presence of a modifier-gene on the Y chromosome of FVB mice, analogous to the Y chromosome-linked Yaa gene that is known to accelerate autoimmune disease in systemic lupus erythematosus (SLE) (Izui et al., 1994).

In a number of kidney disorders, immune mechanisms have been implicated as the initiating factors or the proximate causes of injury. These include glomerulopathies associated with multisystem diseases, such as lupus nephritis in SLE (Kotzin, 1996), and nephropathies associated with a nephrotic syndrome, such as FGS, minimal change nephrotic syndrome (MCNS), membranous nephropathy (MN) or IgM mesangial nephropathy (IgMN) (Klein, 1993). We have tested extensively the possibility that the nephrotic syndrome in Eμ-pp-Frat1 transgenic mice is caused by a systemic autoimmune disease, but we have thus far obtained no evidence for this notion. Eμ-pp-Frat1 animals do not display the accumulation of immune deposits in the kidney, nor do they show the hypergammaglobinemia, splenomegaly, and self-reactive T and B cell populations as observed in lupus nephritis or Fli1-induced renal disease (Zhang et al., 1995; Kotzin, 1996). Moreover, the responses of Eμ-pp-Frat1 thymocytes to various apoptotic stimuli (including α-Fas), and the apoptotic response of activated peripheral T cells to IL-2 withdrawal were found to be normal. Likewise, no autoantibodies against nuclear or glomerular constituents have been detected in blood sera from healthy Eμ-pp-Frat1 transgenics and from nephrotic mice. Finally, long-term bone marrow transplantation experiments showed that overexpression of Frat1 in the hematopoietic compartment alone does not induce a nephrotic syndrome. Given the fact that all three founder lines used in this study show transgene expression in kidney, the combined results from our studies favor the hypothesis that the primary lesion leading to nephropathy in Eμ-pp-Frat1 transgenic mice is kidney-borne instead of blood-borne. In this respect our findings are in contrast with the observations of Nishimura et al. (1994), who reported that transplantation of bone marrow cells or purified hematopoietic stem cells from FGS mice in which FGS appears to be controlled by two pairs of autosomal recessive genes, can induce FGS in normal mice. The recent finding that Frat can activate the Wnt signal transduction pathway through inhibition of GSK3 kinase activity, raises the possibility that the nephropathy in Eμ-pp-Frat1 transgenic mice is caused by ectopic Wnt signaling activity in the (developing) kidney (Yost et al., 1998). In this respect it is noteworthy that several Wnt family members (Wnt-4, 7b and 11) are expressed in the metanephric kidney, and that expression of Wnt-4 correlates with and is required for, kidney tubulogenesis (Stark et al., 1994; Kispert et al., 1996, 1998).

Materials and methods

Generation of Eμ-pp-Frat1 and Eμ-Pim1/FVB transgenic mice

A 920-bp SrfI – XhoI fragment encompassing the Frat1 coding domain was cloned into a cassette vector (van Lohuizen et al., 1989), linking the Frat1 gene to the Pim1 promoter, a duplicated Eμ enhancer, and a M-MuLV LTR. A purified 6.4 kb SalI fragment was microinjected into fertilized oocytes from FVB/NA mice (Taketo et al., 1991). The recovery of oocytes, conditions for microinjection, and transfer of the 2-cell stage embryos have been described previously (van Lohuizen et al., 1989). Genotyping was performed by Southern analysis of tail-tip DNA, according to (Laird et al., 1991). The Eμ-Pim1/FVB transgenic mice used in this study were obtained by microinjection of the previously described Eμ-Pim1 transgene (van Lohuizen et al., 1989) into the pronuclei of FVB/NA zygotes (J Domen, unpublished results).

Mice and lymphoma induction

Mice were fed with a standard diet (AM-II from Hope Farms, Woerden, The Netherlands) and water of pH 2.7 – 2.8. For lymphoma induction, newborn mice from crosses between normal FVB mice and Eμ-pp-Frat1 transgenic mice were infected with 104 – 105 plaque-forming units of M-MuLV clone 1A as described previously (Jaenisch et al., 1975). Mice were sacrificed when moribund and tumor tissues (spleen, thymus, mesenteric/peripheral lymph nodes, liver) were frozen at −80°C. FACS analysis was performed on single-cell suspensions from mesenteric lymph node or spleen.

DNA and RNA analysis

High molecular weight DNA was isolated from lymphoma tissues as described previously (van der Putten et al., 1979). Southern analysis was performed with 10 μg of genomic DNA, digested with the appropriate restriction enzymes. For Northern analysis, 20 μg of total RNA, isolated according to (Sambrook et al., 1989), was used. Blotting and hybridization procedures were as described previously (Sambrook et al., 1989). The following probes were used: Myc, 3 kb XbaI – HindIII fragment (Shen-Ong et al., 1982); Nmyc1, 3.5 kb PstI fragment (Taya et al., 1986); Pim1 probe A, 0.9 kb BamHI fragment (Cuypers et al., 1984); Pim2 5′ probe, 0.5 kb BamHI fragment (van der Lugt et al., 1995); Pim2 probe C, 0.7 kb SacI fragment (van der Lugt et al., 1995); Pal1 probe A, 1.0 kb BglII fragment (van Lohuizen et al., 1991); Frat1 probe 0.45Pst, 0.45 kb PstI fragment (Jonkers et al., 1997) and M-MuLV U3LTR, 180 bp HpaII fragment (Cuypers et al., 1984).

Flow cytometric analysis

Approximately 106 cells were incubated in 20 μl PBA (phosphate-buffered saline with 1% BSA and 0.1% sodium azide) and saturating amounts of monoclonal antibody, in 96-well plates. After incubation, 30 min at 4°C, the cells were washed twice with PBA and incubated with streptavidin-phycoerythrin for biotinylated antibodies, or with PBA. The following antibodies were used: CD45R/B220 (6B2), CD3 (145-2C11), CD4 (RM4-5), CD8 (53-6.7), TCRαβ (H57-597), CD43 (S7), all from Pharmingen (San Diego, CA, USA), and slg/goat anti mouse immunoglobulin (GαM-FITC) from Tago (Burlingame, CA, USA). Heat Stable Antigen (30F1) was kindly provided by K Rajewsky (Cologne, Germany).

Histological analysis

Tissues were isolated from mice and directly fixed in Harrison's fixative (4% v/v formol, 40% v/v ethanol, 0.43% w/v NaCl and 5% v/v acetic acid) for at least 24 h (Harrison, 1984). Fixed tissues were dehydrated, embedded in Histowax, cut at 5μm sections and stained with periodic acid-Schiff (PAS) (Sheehan and Hrapchak, 1980). Immune deposits in diseased kidneys of Eμ-pp-Frat1 mice were studied by direct immunofluorescence on cryosections with GαM-FITC (Tago, Burlingame, CA, USA). For electron microscopy, tissue was processed according to standard procedures.

Proliferation assays

Splenocytes were seeded at various densities in 96-well plates in 100 μl medium (RPMI 1640 (Gibco – BRL) with 10% FCS (Gibco – BRL) and 50 μM 8-mercaptoethanol), supplemented with 2.5 μg/ml conA or 30 μg/ml LPS. After 24 h incubation at 37°C, [3H]thymidine was added to a final concentration of 4 μCi/ml. The cultures were incubated for an additional 8 h at 37°C. The cells were collected on filters, washed, and [3H]thymidine incorporation was determined by scintillation counting.

Apoptosis assays

The assays measuring response to apoptosis were performed as described previously (Brady et al., 1996). Thymocytes were exposed to one of the following apoptotic stimuli: 2 μM dexamethasone for 6 h, 1 μg/ml ionomycin for 14 h, 5 ng/ml phorbol myristate acetate (PMA) for 14 h, α-Fas antibody (Jo-2; Pharmingen) for 20 h, or α-Fas/cycloheximide (30 μg/ml) for 20 h. Alternatively, thymocytes were exposed to γ radiation (100 rads) from a 137Cs source (Von Gahler Nederland BV) and cultured for an additional 16 h. Activated T cells, obtained by incubating splenocytes with ConA for 72 h, were cultured for another 24 h in medium without exogenous IL-2, and the percentage of apoptotic cells was determined at subsequent time points. The percentage of cells undergoing apoptosis was estimated by flow cytometric quantitation of hypodiploid thymocytes, essentially as described (Nicoletti et al., 1991). The percentage apoptosis in the treated thymocytes was normalized to that found in untreated cultures derived from the same animal. All analyses were performed in triplicate, and consistent results were obtained for duplicate sets of mice tested in two independent experiments.

Quantitation of albumin in urine

Urinary albumin secretion was measured in urine collected by housing the mice for 48 h in metabolic cages with free access to water and food. Total amounts of albumin were determined using radial immunodiffusion according to (Mancini et al., 1965). For routine analysis, a semi-quantitative determination of protein in urine samples was obtained by reflectance photometric analysis of Combur9 urine test strips with the Urotron RL 9 photometer (Boehringer Mannheim).

Determination of autoantibodies

Sera from Eμ-pp-Frat1 transgenic mice and from control littermates were screened at a dilution of 1/10 in PBS on Crithidia luciliae, as described (Aarden et al., 1979). For detection of ANA, spots of Hep2 cells, cultured on slides, were incubated for 45 min with sera diluted 1/40 in PBS. After washing in PBS (3×5 min) cells were incubated with the secondary antibody for 45 min, washed again, and stained with 0.01% Evans Blue. GαM-FITC (M 1204 from CLB, Amsterdam) was used as secondary antibody. For detection of antibodies against components of the GBM or the GEC membrane, cryosections of kidneys from healthy mice were incubated with different dilutions of sera from nephrotic Eμ-pp-Frat1 mice, followed by incubation with GαM-FITC (Tago, Burlingame, CA, USA).

Long-term bone marrow transplantations

Bone marrow cells were isolated from femurs and tibiae of 22-week-old Eμ-pp-Frat1 transgenic mice and wild type littermates. As bone marrow recipient animals, 8-week-old sex-matched FVB/N mice were used. These animals were subjected to a single dose of X-ray irradiation (6.8 Gy). Four hours post-irradiation, approximately 5×106 cells were injected intravenously into each recipient. For 2 weeks following irradiation and bone marrow transplantation, the recipients were maintained on aqueous antibiotics by supplementing the water source with 2 mg/ml neomycin-sulfate. The percentage of transgenic peripheral blood lymphocytes (PBL) in the bone marrow recipients exceeded 75%, as determined by semi-quantitative PCR analysis of PBL DNA.


  1. Aarden LA, de Groot ER and Feltkamp TEW. . 1979 NY Acad. Sci. USA 254: 505–515.

  2. Adams JM, Harris AW, Pinkert CA, Corcoran LM, Alexander WS, Cory S, Palmiter RD and Brinster RL. . 1985 Nature 318: 533–538.

  3. Alkema MJ, Jacobs H, van Lohuizen M and Berns A. . 1997 Oncogene 15: 899–910.

  4. Berns A, van der Lugt N, Alkema M, van Lohuizen M, Domen J, Acton D, Allen J, Laird PW and Jonkers J. . 1994 Cold Spring Harbor Symposia on Quantitative Biology, Volume LIX. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY pp. 435–447.

  5. Bishop JM. . 1991 Cell 64: 235–248.

  6. Brady HJ, Salomons GS, Bobeldijk RC and Berns AJ. . 1996 EMBO J. 15: 1221–1230.

  7. Couser WG. . 1994 West. J. Med. 160: 440–446.

  8. Cuypers HT, Selten G, Quint W, Zijlstra M, Robanus-Maandag E, Boelens W, van Wezenbeek P, Melief C and Berns A. . 1984 Cell 37: 141–150.

  9. Harrison PTC. . 1984 Lab. Anim. 18: 325–331.

  10. Haupt Y, Alexander WS, Barri G, Klinken SP and Adams JM. . 1991 Cell 65: 753–763.

  11. Hunter T. . 1991 Cell 64: 249–270.

  12. Izui S, Merino R, Fossati L and Iwamoto M. . 1994 Int. Rev. Immunol. 11: 211–230.

  13. Jaenisch R, Fan H and Croker B. . 1975 Proc. Natl. Acad. Sci. USA 72: 4008–4012.

  14. Jonkers J and Berns A. . 1996 Biochim. Biophys. Acta 1287: 29–57.

  15. Jonkers J, Korswagen HC, Acton D, Breuer M and Berns A. . 1997 EMBO J. 16: 441–450.

  16. Kispert A, Vainio S and McMahon AP. . 1998 Development 125: 4225–4234.

  17. Kispert A, Vainio S, Shen L, Rowitch DH and McMahon AP. . 1996 Development 122: 3627–3637.

  18. Klein JB. . 1993 Curr. Opin. Nephrol. Hypertens. 2: 225–230.

  19. Kotzin BL. . 1996 Cell 85: 303–306.

  20. Laird PW, Zijderveld A, Linders K, Rudnicki MA, Jaenisch R and Berns A. . 1991 Nucl. Acids Res. 19: 4293.

  21. Mancini G, Carbonara AO and Heremans JF. . 1965 Immunochemistry 2: 235–254.

  22. Nicoletti I, Migliorati G, Pagliacci MC and Grigani F. . 1991 J. Immunol. Meth. 139: 271–279.

  23. Nishimura M, Toki J, Sugiura K, Hashimoto F, Tomita T, Fujishima H, Hiramatsu Y, Nishioka N, Nagata N, Takahashi Y and Ikehara S. . 1994 J. Exp. Med. 179: 1053–1058.

  24. Renauld J-C, van der Lugt N, Vink A, van Roon M, Godfraind C, Warnier G, Merz H, Feller A, Berns A and Van Snick J. . 1994 Oncogene 9: 1327–1332.

  25. Sambrook J, Fritsch EF and Maniatis T. . 1989 Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY.

  26. Sheehan DC and Hrapchak BB. . 1980 Theory and Practice of Histochemistry. Mosby: St. Louis, MO.

  27. Shen-Ong GLC, Keath EJ, Piccoli SP and Cole MD. . 1982 Cell 31: 443–452.

  28. Stark K, Vainio S, Vassileva G and McMahon AP. . 1994 Nature 372: 679–683.

  29. Taketo M, Schroeder AC, Mobraaten LE, Gunning KB, Hanten G, Fox RR, Roderick TH, Stewart CL, Lilly F, Hansen CT and Overbeek PA. . 1991 Proc. Natl. Acad. Sci. USA 88: 2065–2069.

  30. Taya Y, Mizusawa S and Nishimura S. . 1986 EMBO J. 5: 1215–1219.

  31. van der Lugt NM, Domen J, Verhoeven E, Linders K, van der Gulden H, Allen J and Berns A. . 1995 EMBO J. 14: 2536–2544.

  32. van der Putten H, Terwindt E, Berns A and Jaenisch R. . 1979 Cell 18: 109–116.

  33. van Lohuizen M, Verbeek S, Krimpenfort P, Domen J, Saris C, Radaszkiewicz T and Berns A. . 1989 Cell 56: 673–682.

  34. van Lohuizen M, Verbeek S, Scheijen B, Wientjens E, van der Gulden H and Berns A. . 1991 Cell 65: 737–752.

  35. Verbeek S, van Lohuizen M, van der Valk M, Domen J, Kraal G and Berns A. . 1991 Mol. Cell. Biol. 11: 1176–1179.

  36. Weening JJ, Beukers JJB, Grond J and Elema JD. . 1986 Kidney Int. 29: 789–798.

  37. Yost C, Farr GH 3rd, Pierce SB, Ferkey DM, Chen MM and Kimelman D. . 1998 Cell 93: 1031–1041.

  38. Zhang L, Eddy A, Teng YT, Fritzler M, Kluppel M, Melet F and Bernstein A. . 1995 Mol. Cell. Biol. 15: 6961–6970.

Download references


We are indebted to Jan de Jong and Dr Ruud Smeenk (CLB, Amsterdam) for excellent analysis of the mouse sera. We wish to thank Dr KJM Assmann for quantitation of proteinuria, and Eric Noteboom for his help and advice on the FACS analysis. Thanks also to Hugh Brady and Gabriel Gil-Gomez for sharing both knowledge and reagents required for the apoptosis assays, and to Nikos Tripodis for statistical analysis. We also thank Dr Paul Krimpenfort for assistance in microinjection, Rein Regnerus for genotyping the mice, and the technical staff of the NKI animal facility for expert assistance in animal care, M-MuLV-infection of newborn mice, and collection of urine and blood samples: Fina van der Ahé, Nel Bosnie, Halfdan Raaso, Loes Rijswijk and Auke Zwerver. This work was supported by the Dutch Cancer Society (J Jonkers).

Author information

Correspondence to Anton Berns.

Rights and permissions

Reprints and Permissions

About this article


  • insertional mutagenesis
  • Moloney leukemia virus
  • transgenic mice
  • T cell lymphoma
  • oncogenes
  • glomerulosclerosis

Further reading