Dissection of Genetic Mechanisms Governing the Expression of Serum Retroviral gp70 Implicated in Murine Lupus Nephritis

Abstract

The endogenous retroviral envelope glycoprotein, gp70, implicated in murine lupus nephritis is secreted by hepatocytes as an acute phase protein, and has been believed to be a product of an endogenous xenotropic virus, NZB-X1. However, since endogenous polytropic (PT) and modified polytropic (mPT) viruses encode gp70s that are closely related to xenotropic gp70, these viruses can be additional sources of serum gp70. To better understand the genetic basis of the expression of serum gp70, we analyzed the abundance of xenotropic, PT or mPT gp70 RNAs in livers and the genomic composition of corresponding proviruses in various strains of mice, including two different Sgp (serum gp70 production) congenic mice. Our results demonstrated that the expression of different viral gp70 RNAs was remarkable heterogeneous among various mouse strains and that the level of serum gp70 production was regulated by multiple structural and regulatory genes. In addition, a significant contribution of PT and mPT gp70s to serum gp70 was revealed by the detection of PT and mPT, but not xenotropic transcripts in 129 mice and by a closer correlation of serum levels of gp70 with the abundance of PT and mPT gp70 RNAs than with that of xenotropic gp70 RNA in Sgp3 congenic mice. Furthermore, the injection of lipopolysaccharides selectively up-regulated the expression of xenotropic and mPT gp70 RNAs, but not PT gp70 RNA. Our data indicate that the genetic origin of serum gp70 is more heterogeneous than previously believed, and that distinct retroviral gp70s are differentially regulated in physiological vs. inflammatory conditions.

Endogenous retroviruses are implicated in the pathogenesis of murine systemic lupus erythematosus (SLE)⁵. This relationship was first suggested when murine leukemia viral antigens were found in sera and immune deposits of diseased glomeruli in lupus-prone NZB and (NZB x NZW)F1 hybrid mice (1, 2). Subsequently, it was demonstrated that relatively large amounts of the major envelope glycoprotein, gp70, are present in the sera of lupus-prone (NZB x NZW)F1, MRL and BXSB mice, free from any association with viral particles, and that only lupus-prone mice spontaneously develop autoantibodies against serum retroviral gp70 (3–5). Indeed, gp70-anti-gp70 immune complexes (gp70 IC) were detected close to the onset of renal disease in the circulation and found within diseased glomeruli of lupus mice (3, 5). In several studies of murine lupus, levels of serum gp70 IC have been shown to closely correlate with the development of severe lupus nephritis (6–10), further supporting the pathogenic role of gp70 IC in murine SLE.

All inbred strains of mice contain numerous endogenous retroviruses as chromosomal genes. The expression of gp70, encoded by the retroviral env gene, is modulated during embryonic development and is linked to the differentiation state of the cells (11). Indeed, gp70 is a constituent of the surface of various epithelia and of thymocyte and mature peripheral lymphocyte cell membranes, and shares immunological and biochemical properties with the thymocyte differentiation antigen GIX (11–15). It has previously been demonstrated that lymphoid cells are not a major source for serum retroviral gp70, because neither thymectomy nor splenectomy affected serum levels of gp70 (16). Rather, serum gp70 behaves as an acute phase protein and is secreted by hepatic cells into the circulating blood (17, 18).

Serum concentrations of gp70 are highly variable among different strains of mice (3–5, 19). All SLE-prone strains have relatively high concentrations of gp70 in their sera (>15 μg/ml), whereas C57BL/6 (B6), C57BL/10 (B10) and BALB/c mice produce low serum levels of gp70 (<5 μg/ml). By studying the progeny of crosses of lupus-prone NZB, NZW and BXSB with non-autoimmune B6 or B10 strains, a major quantitative trait locus, called Sgp3 (serum gp70 production 3), on mid chromosome 13 was found to be strongly linked with serum levels of gp70 (10, 20–23). Markedly increased serum levels of gp70 in B6 or B10 congenic mice bearing either the NZB-, NZW- or BXSB-derived Sgp3 allele confirmed the important role of Sgp3 in serum gp70 production (21, 23, 24). In addition, a second NZB and NZW locus on distal chromosome 4 was identified to be linked to serum gp70 levels in crosses with B6 and BALB/c backgrounds (20, 25). Since no gene name has been given to this locus, we propose to designate it Sgp4.

Endogenous retroviruses are classified as ecotropic, xenotropic or polytropic according to their host range dictated by their respective gp70 proteins (26). Serological analysis clearly excluded the involvement of ecotropic gp70 as a source of serum gp70 (27), and tryptic peptide mapping analysis suggested that serum gp70 molecule resembles the envelope protein of NZB-X1 virus, one of the two distinct xenotropic viruses isolated from NZB mice (28, 29). However, NZB-X1 does not appear to be the sole source of serum gp70. The fingerprint of serum gp70 showed additional marker peptides corresponding to gp70s of other xenotropic viruses, including NZB-X2, as well as to the gp70s expressed on thymocytes and splenic lymphocytes. In addition, the analysis of a highly tumorigenic AKR SL12.3 cell line revealed the presence of four subgroups of xenotropic proviruses (Xeno-I, Xeno-II, Xeno-III and Xeno-IV), which exhibit distinct nucleotide sequences in a variable region of the 5′ portion of their env genes (30). However, it has not yet been determined which of the four xenotropic gp70 are expressed in hepatocytes of different mouse strains and might contribute to serum gp70. Furthermore, polytropic proviruses are comprised of a large group of endogenous viruses that encode gp70s closely related to xenotropic gp70 (26), as both retroviruses share a common entry receptor, XPR1 (xenotropic and polytropic retrovirus receptor) (31, 32). Polytropic proviruses have been divided into two major structural subgroups termed polytropic (PT) and modified polytropic (mPT) based on differences in their gp70 nucleotide sequences (33). Either of these two subgroups are potential sources of serum gp70.

In view of the substantial role of serum gp70 as one of the major nephritogenic autoantigens in murine SLE, it is important to define the genetic origin of serum gp70 and the genetic mechanisms responsible for its expression. To address these questions, we analyzed the abundance of xenotropic, PT and mPT gp70 RNA transcripts in liver, in relation to serum levels of gp70, and the genomic composition of corresponding proviruses in various strains of mice, including two different Sgp3 congenic strains and one Sgp4 congenic strain. Our results reveal a substantial contribution of PT and mPT gp70s, in addition to xenotropic gp70, to serum gp70 and a differential regulation of serum gp70 production in non-inflammatory vs. inflammatory conditions.

Materials and Methods

Mice

NZB, NZW, MRL, BXSB and 129 mice were purchased from the Jackson Laboratory (Bar Harbor, ME). NFS mice have been maintained at Rocky Mountain Laboratories, Veterinary Branch. B6.NZB-Sgp3 and B10.BXSB-Sgp3 congenic mice were generated by backcross procedures using marker-assisted selection, as described previously (21, 24). B6.NZB-Sgp3 mice carry an NZB-derived interval flanked by markers D13Mit13 and D13Mit26 (16 Mb interval), and B10.BXSB-Sgp3 mice carry a BXSB-derived interval encompassing markers D13Mit122 and D13Mit233 (24 Mb interval). B6.NZB-Sgp4 congenic mice were generated by backcrossing an NZB-derived Sgp4 interval flanked by markers D4Mit11 and D4Mit33 (27 Mb interval) onto the B6 background. The positions of the microsatellite markers with respect to the centromere were obtained from the Mouse Genome Database at http://www.informatics.jax.org. Blood samples were collected by orbital sinus puncture.

cDNA sequencing

RNA from NZB-X1 (Clone 35 and NZB 179) viruses (29, 34), obtained from ViroMed Biosafety Laboratories, Camden, NJ, was purified with TRIzol reagents (Invitrogen AG, Basel, Switzerland). The xenotropic gp70 cDNA clone pGP24 isolated from liver of LPS-injected NZB mice (35) was provided by Dr N. Maruyama, Tokyo Metropolitan Institute for Gerontology, Tokyo, Japan. The entire coding region of xenotropic viral gp70 cDNA was amplified with Taq DNA polymerase (MP Biomedicals Switzerland, Basel, Switzerland) using a Xeno238F forward primer (5′-TGGATACACGCCGCTCACG-3′) and a Xeno1847R reverse primer (5′-ATCTAATCCTCTCCGGTTCT-3′).

RT-PCR

The presence of viral gp70 mRNAs was detected by RT-PCR, using cDNA prepared from liver RNA. A forward primer covering the tRNA primer-binding site (5′-CATTTGGAGGYYCCASCGA-3′) and reverse primers on the env genes specific for four different subgroups of xenotropic viruses (Fig. 1A) were used to amplify gp70 mRNA, but not viral genomic RNA. In addition, two different conserved xenotropic-specific Xeno400R (5′-CTGTCACGTTGTACCGAGG-3′) and Xeno685R (5′-TTGCCACAGTAGCCCTCTCC-3′) reverse primers were used to confirm the absence of xenotropic gp70 mRNAs in 129 mice. PCR products were visualized by staining with ethidium bromide after electrophoresis on 2% agarose gels.

FIGURE 1.

Primer sequences used in RT-PCR and genomic PCR for amplification of subgroups of xenotropic and polytropic gp70s. The primer sequences in the hypervariable region A (VRA) specific for the four subgroups of xenotropic gp70s (A and B) and in a proline-rich domain (PRD) specific for PT and mPT gp70s (C) are underlined. R: A/G; S: C/G; Y: T/C.

Quantitative real-time RT-PCR

The abundance of total viral gp70 RNA (genomic RNA and mRNA) was quantified by real-time RT-PCR using cDNA prepared from liver RNA digested with DNase I (Amersham Biosciences Corp., Piscataway, NJ). For the amplification of four different subgroups of xenotropic viral gp70 cDNA, forward primers on the env genes specific for each subgroup of xenotropic viruses (Fig. 1B) and a common Xeno685R reverse primer were used. For PT and mPT viral gp70 cDNA, a common forward primer (5′-CCGCCAGGTCCTCAATATAG-3′) and reverse primers specific for PT and mPT viruses (Fig. 1C) were used. Haptoglobin mRNA levels were quantified with the following primers: forward primer (5′-TGAACACAGTCGCTGGAGAG-3′) and reverse primer (5′-GCTGCCTTTGGCATCCATAG-3′). PCR was performed using the iCycler iQ Real-Time PCR Detection System (Bio-Rad, Philadelphia, PA) and iQ SYBR green Supermix (Bio-Rad). Results were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation from NZB liver and normalized using TATA-binding protein (TBP) mRNA.

Genomic PCR

The presence of different xenotropic proviruses in the genome was determined by PCR on genomic DNA prepared from liver. The following primers were used: a Xeno277F common forward primer (5′-CCAGCCGGAACAGCATGGAAG-3′) and reverse primers specific for each xenotropic provirus (Fig. 1A).

Serological assays

Serum levels of retroviral gp70 from 2–3 mo-old male mice were determined by ELISA as described previously (36). Results are expressed as μg/ml of gp70 by referring to a standard curve obtained with a serum pool from NZB mice containing a known concentration of gp70. Serum titers of haptoglobin were measured by radial immunodiffusion in agar with goat anti-human haptoglobin antiserum (Cappel Laboratories, Cochranville, PA). Results are expressed as the percentages of the value obtained with a serum pool from 2–3 mo-old B6 male mice.

Injection of LPS

25 μg of LPS from Escherchia coli 0111:B4 (Sigma-Aldrich, Saint Louis, MO) were i.p injected into 2–3 mo-old NZB, BXSB and B6 male mice. Livers were taken 9 h after LPS injection to prepare RNA, and sera were collected 24 h after injection for the analysis of gp70 and haptoglobin.

Statistical analysis

Analysis for serum levels of gp70 in Sgp3 and Sgp4 congenic mice was performed with the Mann-Whitney test. Unpaired comparison for gp70 RNA expression and paired comparison for serum levels of gp70 and haptoglobin before and after injection of LPS were analyzed by Student s t test. Probability values <5% were considered significant.

Results

Presence of four different subgroups of xenotropic proviruses in mice

The presence of four different xenotropic proviruses was previously reported in an AKR lymphoma SL12.3 cell line (30). We confirmed by BLAST search analysis the presence in the mouse genome of these four subgroups which are distinguishable by differences within the hypervariable region A (VRA) of gp70. These differences include amino acid insertions or deletions as well as substitutions (Fig. 2). Based on tryptic peptide fingerprint analysis, it was suggested that the structure of serum gp70 was closer to that of NZB-X1 gp70 than to that of NZB-X2 gp70 (28, 29). Therefore, the nucleotide sequence of the coding region of gp70 derived from two NZB-X1 (Clone 35 and NZB 179) virus isolates was determined to assign this virus to its xenotropic subgroup, and compared with that of NZB-X2 gp70 (37). The deduced VRA amino-acid sequence revealed that both NZB-X1 and NZB-X2 gp70s belong to the Xeno-I subgroup, and that their amino-acid sequences differ by only three amino-acid residues near the C-terminus (Fig. 2). Although the nucleotide sequence of the xenotropic gp70 cDNA clone (pGP24), isolated from liver of LPS-injected NZB mice, was previously reported to be slightly different from the NZB-X1 and NZB-X2 gp70 sequences (35), our sequence analysis indicated that the gp70 sequence of pGP24 is identical to that of NZB-X1 gp70.

FIGURE 2.

Predicted amino-acid sequences of different retroviral gp70s. Nucleotide sequence analysis of cDNA prepared from two different NZB-X1 (Clone 35 and NZB 179) viruses revealed that both NZB-X1 and NZB-X2 gp70s belong to the Xeno-I subgroup, but their predicted amino acid sequences differ by three amino-acid residues at the C-terminus, which are highlighted in bold. The GenBank accession number for NZB-X1 gp70 cDNA sequence is EU334447. NZB-X2, PT and mPT gp70 sequences are derived from NZB-9-1, MX27 and MX33 proviruses, respectively (26, 37), and Xeno-II, Xeno-III and Xeno-IV gp70 sequences from RP24-240L12, RP24-114A21 and RP23-110C17 BAC clones, respectively. The hypervariable regions A and B (VRA and VRB) and a proline-rich domain (PRD) are shaded. Identities are indicated by dashes. Numbers indicate amino-acid positions.

Differential expression of four different xenotropic gp70 mRNAs in liver of various strains of mice with high serum gp70

To assess the expression of the four different subgroups of xenotropic viral gp70 mRNAs in liver, we developed RT-PCRs with a common forward primer covering the tRNA primer-binding site and subgroup-specific reverse primers (Fig. 1A), which specifically amplify gp70 mRNA but not viral genomic RNA. Since the reverse primer designed for the amplification of Xeno-III gp70 cDNA is also able to amplify Xeno-II gp70 cDNA, the RT-PCR products corresponding to Xeno-III could contain both Xeno-III and Xeno-II replicons. Therefore, the presence of Xeno-III gp70 mRNA was confirmed by sequence analysis of the amplified fragments.

When a panel of murine strains having high serum levels of gp70 (>10 μg/ml) was analyzed for the presence of different xenotropic gp70 mRNAs, we found considerable heterogeneity among them (Fig. 3A). Accordingly, at least five different groups of mice could be defined: 1) NZB-type expressing all four xenotropic gp70 (NZB and NZW); 2) MRL-type expressing Xeno-I, Xeno-III and Xeno-IV; 3) BXSB-type expressing Xeno-II, Xeno-III and Xeno-IV; 4) NFS-type expressing only Xeno-III; and 5) 129-type expressing no xenotropic mRNA. Notably, we observed no differences in the expression patterns of xenotropic gp70 mRNAs between male and female mice (data not shown). The absence of xenotropic gp70 transcripts in 129 mice was further confirmed by additional RT-PCRs using primers for conserved xenotropic-specific gp70-encoding sequences: these primers, which are capable of amplifying all four types of xenotropic mRNAs, did not yield RT-PCR products with liver RNA from 129 mice (data not shown).

FIGURE 3.

RT-PCR and genomic PCR analyses for four subgroups of xenotropic gp70 mRNAs and proviruses in different strains of mice. (A) The presence of Xeno-I, Xeno-II, Xeno-III and Xeno-IV gp70 mRNAs in liver from male mice of different strains was determined by RT-PCR with a forward primer covering the tRNA primer-binding site and reverse primers specific for the four different subgroups of xenotropic viruses (Fig. 1A). Since the reverse primer designed for the amplification of Xeno-III gp70 also amplifies Xeno-II gp70, the results obtained with this primer are indicated as Xeno-II/III. (B) Genomic DNA from different strains of female (F) and male (M) mice was analyzed for the presence of Xeno-I, Xeno-II, Xeno-III and Xeno-IV gp70 sequences by PCR with a common forward primer (Xeno277F) and the same reverse primers specific for the four different subgroups of xenotropic viruses used for the detection of gp70 mRNAs.

The levels of xenotropic viral gp70 RNA transcripts in liver from male mice were quantified by real-time RT-PCR using subgroup-specific forward primers (Fig. 1B) and a common reverse primer, which amplifies cDNA derived from both viral genomic RNA and mRNA. We elected to measure total gp70 RNA since we could not design subgroup-specific primers for gp70 mRNA suitable for real-time RT-PCR. As was the case with the mRNA reverse primers above, the Xeno-III forward primer also amplifies Xeno-II RNA, while the Xeno-II primer does not amplify Xeno-III RNA. Thus, the abundance of Xeno-III gp70 RNA could be estimated by subtracting the levels obtained with the Xeno-II forward primer from those obtained with the Xeno-III forward primer. In all cases, the expression patterns of the four xenotropic viral gp70 mRNAs observed (Fig. 3A) were corroborated by the real-time RT-PCR experiments (Table I). In addition to the qualitative differences in the expression patterns, we also observed remarkable quantitative differences among the four xenotropic viral gp70 RNAs and between the six strains of mice tested (Table I). Both PT and mPT gp70 transcripts were detectable in all of the strains, although mPT RNA was expressed at lower levels in NFS and 129 mice (Table I).

Table I.

Levels of serum gp70 and hepatic retroviral gp70 RNAs in different strains of mice

				gp70 RNAb
Mice	Serum gp70a	Xeno-I	Xeno-II	Xeno-II/III	Xeno-IV	PT	mPT
NZB	66.4 ± 5.7	5.58 ± 0.54	3.46 ± 1.44	4.58 ± 0.40	0.20 ± 0.04	6.49 ± 0.30	6.92 ± 1.14
NZW	49.0 ± 7.0	4.18 ± 0.43	0.01 ± 0.01	0.83 ± 0.11	0.05 ± 0.01	4.10 ± 0.77	2.54 ± 0.34
MRL	17.7 ± 1.8	0.15 ± 0.09	NDc	1.53 ± 0.58	0.09 ± 0.01	2.49 ± 0.71	5.09 ± 1.21
BXSB	26.7 ± 5.3	ND	0.06 ± 0.02	0.80 ± 0.22	0.06 ± 0.02	3.82 ± 0.91	5.47 ± 0.95
NFS	32.6 ± 7.4	ND	ND	2.69± 0.91	ND	5.79 ± 0.69	0.18 ± 0.06
129	11.1 ± 3.4	ND	ND	ND	ND	2.31 ± 0.62	0.08 ± 0.04

^aSerum levels of gp70 (μg/ml; mean ± SD of 7–10 male mice at 2–3 months of age).

^bLevels of each gp70 RNA (mean ± SD of 3 male mice at 2–3 months of age) were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation and normalized using TBP mRNA. Since the forward primer designed for the amplification of Xeno-III gp70 also amplifies Xeno-II gp70, the results obtained with this primer are indicated as Xeno-II/III.

^cND: Not detectable

Correlation of xenotropic gp70 RNA expression in liver of different mouse strains with the presence of structural genes encoding xenotropic gp70

The lack of expression of some of the xenotropic gp70 RNAs in BXSB, MRL, NFS and 129 mice could be due to the absence of structural genes encoding xenotropic gp70 or, alternatively, to transcriptional control mechanisms. To examine these possibilities, we performed PCR analyses to determine the identity of the xenotropic gp70 structural genes in the genomic DNA of the different mouse strains and compared the results with the gp70 RNA expression patterns in the livers of the respective strains. PCRs were performed using a combination of a common forward primer and the reverse primers specific for the four different xenotropic gp70s (Fig. 1A). Since the expression patterns of xenotropic gp70 mRNAs between male and female mice were not different among the strains of mice tested, the initial PCR analyses were performed on DNA from females of these strains. Additional PCR analyses were carried out on DNA from male mice of those strains which lacked one or more xenotropic proviruses in the female DNA (MRL, NFS and 129) to determine if xenotropic proviruses were present on the Y chromosome. In four of the six strains tested there was complete correlation between the presence of xenotropic gp70 genes and the expression of gp70 RNAs in the liver (Fig. 3A and 3B). These included NZB and NZW mice (Xeno-I, -II, -III and -IV), MRL mice (Xeno-I, -III and -IV) and NFS mice (Xeno-III). Thus, the lack of expression of particular xenotropic RNA transcripts in the livers of MRL and NFS mice is due to the absence of the respective proviruses. In contrast, the analysis of the genomic DNA from BXSB mice showed the presence of all four xenotropic proviruses (Fig. 3B), despite the lack of expression of Xeno-I gp70 transcripts in their livers (Fig. 3A). In addition, no xenotropic gp70 sequences were transcribed in the livers of 129 mice, although Xeno-I and Xeno-IV gp70 sequences were detectable in male mice of this strain. Our results indicated that the absence of expression of certain xenotropic proviruses in BXSB and 129 mice was due to transcriptional suppression rather than the absence of the respective proviruses.

Enhanced hepatic expression of xenotropic, PT and mPT gp70 RNAs in B6 or B10 congenic mice bearing the NZB- or BXSB-Sgp3 allele

Our recent analyses on B6 and B10 congenic mice bearing the NZB-Sgp3 or BXSB-Sgp3 allele, respectively, revealed a substantial role of the Sgp3 locus in the increase of serum gp70 (21, 24). Indeed, serum levels of gp70 in B6.NZB-Sgp3 and B10.BXSB-Sgp3 mice were 6- and 18-fold higher than those in wild-type B6 and B10 mice, respectively (p<0.0001), and their gp70 levels were almost as high as those in BXSB mice (Table II). The Sgp3 gene from either source (NZB or BXSB) did not increase gp70 levels to the very high levels observed in NZB mice, suggesting that additional genes influence the serum gp70 levels in this strain.

Table II.

Levels of serum gp70 and hepatic retroviral gp70 RNAs in Sgp3and Sgp4 congenic mice

				gp70 RNAb
Mice	Serum gp70a	Xeno-I	Xeno-II	Xeno-II/III	Xeno-IV	PT	mPT
B6	3.1 ± 0.6	0.05 ± 0.02	0.03 ± 0.004	0.03 ± 0.01	0.06 ± 0.02	1.13 ± 0.18	0.13 ± 0.03
B6.NZB-Sgp3	19.7 ± 4.5 (6.4)c	0.18 ± 0.02 (3.7)	0.06 ± 0.02 (2.4)	0.14 ± 0.04 (4.5)	0.04 ± 0.01 (0.7)	2.70 ± 0.49 (2.4)	1.54 ± 0.52 (12.0)
B6.NZB-Sgp4	9.9 ± 1.6 (3.2)	0.22 ± 0.03 (4.5)	<0.01	<0.01	0.09 ± 0.04 (1.5)	1.36 ± 0.05 (1.2)	0.13 ± 0.01 (1.0)
NZB	66.4 ± 5.7	5.58 ± 0.54	3.46 ± 1.44	4.58 ± 0.40	0.20 ± 0.04	6.49 ± 0.30	6.92 ± 1.14
B10	1.1 ± 0.2	0.01 ± 0.001	0.02 ± 0.003	0.01 ± 0.002	0.04 ± 0.01	0.11 ± 0.04	0.14 ± 0.02
B10.BXSB-Sgp3	19.5 ± 6.6 (17.7)	0.10 ± 0.07 (14.9)	0.05 ± 0.02 (2.5)	0.16 ± 0.10 (14.7)	0.04 ± 0.004 (1.2)	3.80 ± 0.59 (35.5)	0.83 ± 0.0 3 (5.8)
BXSB	26.7 ± 5.3	NDd	0.06 ± 0.02	0.80 ± 0.22	0.06 ± 0.02	3.82 ± 0.91	5.47 ± 0.95

^aSerum levels of gp70 (μg/ml; mean ± SD of 7–12 male mice at 2–3 months of age).

^bLevels of each gp70 RNA (mean ± SD of 3 male mice at 2–3 months of age) were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation and normalized using TBP mRNA.

^cFold increase in serum gp70 and hepatic gp70 RNA levels in Sgp congenic vs. wild-type mice.

^dND: Not detectable

To investigate the genetic origin of serum gp70, the abundance of different retroviral gp70 RNA transcripts in liver was compared between B6, B6.NZB-Sgp3, and NZB mice. Quantification of gp70 RNAs revealed small but significant (2- to 5-fold) increases in Xeno-I (p<0.02), Xeno-II (p<0.01), Xeno-II/III (p<0.005), and PT (p<0.005) gp70 RNAs and a larger (12-fold) increase in mPT (p=0.0005) gp70 RNA in liver of B6.NZB-Sgp3 mice, as compared with B6 mice (Table II). No significant increase in Xeno-IV gp70 transcripts was observed. Notably, the proportion of xenotropic gp70 RNAs in B6.NZB-Sgp3 mice was still very limited, accounting for only <5% of xenotropic gp70 RNAs detectable in NZB mice. In contrast, the levels of PT and mPT gp70 RNAs in B6.NZB-Sgp3 mice mounted to 41.6% and 22.3%, respectively, of those found in NZB mice.

We also compared the expression levels of different gp70 RNAs among B10, B10.BXSB-Sgp3 and BXSB mice. As in the case of B6.NZB-Sgp3 congenic mice, the presence of the BXSB-Sgp3 allele promoted the expression of all gp70 RNAs, except Xeno-IV gp70 RNA (Xeno-I: p<0.002; Xeno-II: p<0.01; Xeno-II/III: p<0.005; PT: p<0.0001; mPT: p<0.0001; Table II). The most striking up-regulation was obtained with PT gp70 RNA (36-fold higher than that of B10 mice), the level of which became comparable to that of BXSB mice. Notably, B10.BXSB-Sgp3 mice expressed Xeno-I gp70 RNA at a level more than 10-fold higher than that of B10 mice, indicating that in addition to Xeno-II and Xeno-III proviruses, the BXSB-Sgp3 allele is able to enhance the transcription of Xeno-I provirus, although the latter is not expressed in BXSB mice.

Increases in serum levels of gp70 in association with a selective increase in Xeno-I gp70 RNA in the liver of B6 congenic mice bearing the NZB-Sgp4 allele

Linkage of the Sgp4 locus to serum gp70 levels was previously revealed in genetic crosses of NZB and NZW with B6 and BALB/c strains (20, 25). To confirm the contribution of the Sgp4 locus to the production of serum gp70, B6.NZB-Sgp4 congenic mice homozygous for an NZB-derived interval flanked by markers D4Mit11 and D4Mit33 (27 Mb interval) on distal chromosome 4 were produced and analyzed for serum levels of gp70. B6.NZB-Sgp4 male mice had approximately 3-fold higher levels of serum gp70 as compared with B6 male mice (p<0.0001; Table II). The analysis of the abundance of retroviral gp70 RNA transcripts in the liver of B6.NZB-Sgp4 mice revealed that the observed increase in serum gp70 was associated with an approximately 5-fold selective up-regulation of Xeno-I gp70 RNA in B6.NZB-Sgp4 mice (p=0.01; Table II). No significant up-regulation was found for any other xenotropic gp70 RNA or for PT and mPT gp70 RNAs. Instead, we noted that the presence of the NZB-Sgp4 locus markedly suppressed the levels of Xeno-II and Xeno-III gp70 RNAs. These observations suggested that the enhanced serum levels of gp70 attributable to Sgp4 were the result of an up-regulation of Xeno-I RNA transcripts, and that this locus may differentially influence the expression of multiple xenotropic proviruses.

Selective increases in xenotropic and mPT gp70 RNAs in liver of LPS-injected mice

We have previously shown that the expression of serum gp70 is markedly enhanced following the injection of various stimuli, such as LPS, which up-regulate the production of acute phase proteins (17). This response was strain-dependent, since only mice with high serum levels of gp70 responded strongly to the injection of LPS (38, 39). We quantified the changes in abundance of the different retroviral gp70 RNAs in the liver after injection of LPS in responding (NZB and BXSB) and non-responding (B6) mice. Strong (10- to 17-fold) up-regulation of Xeno-I (p<0.0001), Xeno-II (p<0.0001) and Xeno-II/III (p<0.0005) gp70 RNA levels were observed in NZB mice, while Xeno-IV and mPT gp70 RNAs were moderately (5- to 6-fold) up-regulated (Xeno-IV: p<0.001; mPT: p<0.0005; Table III). A similar pattern of up-regulation was observed with BXSB mice (Xeno-II: p<0.0001; Xeno-II/III: p<0.0001; Xeno-IV: p<0.02; mPT: p<0.0005), except that these mice still failed to express Xeno-I RNA even after injection of LPS. The level of PT gp70 RNA was hardly up-regulated in NZB and BXSB mice (Table III). In contrast to NZB and BXSB mice, the abundance of the gp70 transcripts were only marginally, if any, increased in B6 mice after injection of LPS, in agreement with the lack of serum gp70 responses in this strain. It should be noted, however, that B6 mice did exhibit up-regulation of the acute phase protein haptoglobin in response to LPS (Table IV). Indeed, the injection of LPS induced approximately 5-fold increases in both serum haptoglobin concentrations (p<0.005) and hepatic haptoglobin mRNA levels (p<0.0001), comparable to those obtained with LPS-treated NZB and BXSB mice (Table IV).

Table III.

Levels of serum gp70 and hepatic retroviral gp70 RNAs in NZB, BXSB and B6 mice injected with LPS

				gp70 RNAb
Mice	Serum gp70a	Xeno-I	Xeno-II	Xeno-II/III	Xeno-IV	PT	mPT
NZB	66.0 ± 6.2	5.58 ± 0.54	3.46 ± 1.44	4.58 ± 0.40	0.20 ± 0.04	6.49 ± 0.30	6.92 ± 1.14
NZB + LPS	634.3 ± 156.2 (9.6)c	54.22 ± 3.64 (9.7)	57.54 ± 4.31 (16.6)	76.44 ± 9.10 (16.7)	1.17 ± 0.26 (5.9)	8.46 ± 1.46 (1.3)	32.27 ± 5.05 (4.7)
BXSB	26.3 ± 2.1	NDd	0.06 ± 0.02	0.80 ± 0.22	0.06 ± 0.02	3.82 ± 0.91	5.47 ± 0.95
BXSB + LPS	189.2 ± 22.1 (7.2)	ND	20.50 ± 4.25 (347.5)	38.15 ± 8.69 (47.7)	0.17 ± 0.04 (3.0)	4.10 ± 0.62 (1.1)	28.56 ± 4.74 (5.2)
B6	2.5 ± 0.5	0.05 ± 0.02	0.03 ± 0.004	0.03 ± 0.01	0.06 ± 0.02	1.13 ± 0.18	0.13 ± 0.03
B6 + LPS	3.3 ± 1.2 (1.3)	0.08 ± 0.04 (1.6)	<0.01	0.07 ± 0.03 (2.2)	0.10 ± 0.01 (2.0)	1.31 ± 0.24 (1.2)	0.15 ± 0.02 (1.2)

^aSerum levels of gp70 (μg/ml; mean ± SD of 5 male mice at 2–3 months of age) before and 24 h after an i.p injection of 25 μg LPS. Serum gp70 levels after injection of LPS were significantly increased in NZB and BXSB mice (p<0.001), but not in B6 mice.

^bLevels of each gp70 RNA (mean ± SD of 3 male mice at 2–3 months of age) were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation and normalized using TBP mRNA.

^cFold increase in serum gp70 and hepatic gp70 RNA levels after injection of LPS.

^dND: Not detectable

Table IV.

Levels of serum haptoglobin and hepatic haptoglobin mRNA in NZB, BXSB and B6 mice injected with LPS

Mice	Serum haptoglobina	Haptoglobin mRNAb
NZB	101 ± 17	0.17 ± 0.05
NZB + LPS	532 ± 142	0.66 ± 0.06
BXSB	89 ± 18	0.12 ± 0.05
BXSB + LPS	640 ± 78	0.70 ± 0.03
B6	119 ± 32	0.17 ± 0.03
B6 + LPS	503 ± 137	0.73 ± 0.07

^aSerum levels of haptoglobin (mean ± SD of 5 male mice at 2–3 months of age) before and 24 h after an i.p. injection of 25 μg LPS are expressed as the percentages of the value obtained with a serum pool from 2–3 mo-old B6 male mice.

^bLevels of haptoglobin mRNA (mean ± SD of 3 male mice at 2–3 months of age) were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation and normalized using TBP mRNA.

We next examined whether Sgp3 and Sgp4 could contribute to the enhanced gp70 production observed after LPS stimulation. As shown in Table V, both B6.NZB-Sgp3 and B6.NZB-Sgp4 congenic mice displayed moderate but significant 2.2- and 1.8-fold increases of serum gp70 (p<0.001 and p<0.01, respectively) in response to LPS (Table V). These increases paralleled the substantial up-regulation (4- to 12-fold) of Xeno-I (p<0.02), Xeno-II (p<0.05), Xeno-II/III (p<0.05) and mPT (p<0.05) gp70 RNA levels in B6.NZB-Sgp3 mice, and of Xeno-I (p<0.005) and Xeno-II/III (p<0.05) gp70 RNA levels in B6.NZB-Sgp4 mice (Table V). As in the case of NZB and BXSB mice, levels of PT gp70 RNA were not increased in both congenic mice injected with LPS.

Table V.

Levels of serum gp70 and hepatic retroviral gp70 RNAs in B6 Sgp3 and Sgp4 congenic mice injected with LPS

				gp70 RNAb
Mice	Serum gp70a	Xeno-I	Xeno-II	Xeno-II/III	Xeno-IV	PT	mPT
Sgp3	18.2 ± 5.7	0.13 ± 0.02	0.07 ± 0.02	0.13 ± 0.06	0.06 ± 0.02	2.32 ± 0.48	1.42 ± 0.48
Sgp3 + LPS	40.1 ± 10.7 (2.2)c	0.82 ± 0.44 (6.1)	0.78 ± 0.34 (12.1)	1.12 ± 0.49 (8.6)	0.14 ± 0.05 (2.3)	1.92 ± 0.47 (0.9)	4.96 ± 1.95 (3.6)
Sgp4	10.4 ± 2.1	0.25 ± 0.12	<0.001	0.004 ± 0.002	0.09 ± 0.03	1.20 ± 0.20	0.14 ± 0.05
Sgp4 + LPS	17.3 ± 3.0 (1.8)	1.16 ± 0.21 (5.9)	0.003 ± 0.003 (~3.0)	0.033 ± 0.016 (7.7)	0.11 ± 0.03 (1.3)	0.77 ± 0.14 (0.7)	0.13 ± 0.02 (1.0)

^aSerum levels of gp70 (μg/ml; mean ± SD of 5 male mice at 2–3 months of age) before and 24 h after an i.p injection of 25 μg LPS.

^bLevels of each gp70 RNA (mean ± SD of 3 male mice at 2–3 months of age) were quantified using a standard curve generated with serial dilutions of a reference cDNA preparation and normalized using TBP mRNA.

^cFold increase in serum gp70 and hepatic gp70 RNA levels after injection of LPS.

Discussion

gp70, the major envelope glycoprotein of endogenous retroviruses secreted by hepatocytes, is expressed in virtually all strains of mice, but serum gp70 concentrations are highly variable among the strains. The pathogenic role of gp70 in murine lupus nephritis has been underscored through the presence of circulating gp70 IC near the onset of lupus nephritis and of gp70 in immune deposits of diseased glomeruli in lupus-prone mice. To better understand the genetic origin of serum gp70 and the genetic mechanisms responsible for its expression, we analyzed the abundance of different hepatic gp70 RNAs and the genomic composition of corresponding proviruses in various strains of mice, including congenic mice bearing either of the recently identified Sgp loci (Sgp3 and Sgp4). Our results demonstrate that 1) the level of serum gp70 production is regulated by multiple structural and regulatory genes, 2) PT and mPT gp70s, in addition to xenotropic gp70, significantly contribute to basal levels of serum gp70, and 3) a selective up-regulation of the expression of xenotropic and mPT gp70s takes place during an acute phase response.

RT-PCR analyses in livers from different strains of mice revealed the transcription of four structural subgroups of xenotropic viral gp70 RNAs, differing in sequence in the N-terminal variable region, the presence of which was initially reported in an AKR tumorigenic cell line (30). However, gp70 RNA expression patterns were highly variable among different strains of mice, which is partly due to the lack of some of the xenotropic proviruses in the respective genomes. This is the case with MRL, 129 and NFS mice which harbor in their genomes only three, two or one of the four subgroups of xenotropic proviruses, respectively. However, some strains appear to differentially transcribe their xenotropic gp70 genomic sequences, perhaps as a result of differences in integration sites (13, 40–42) and/or in transcriptional regulation. For example, BXSB mice are unable to express Xeno-I gp70 RNA, even after the injection of LPS, despite the presence of Xeno-I proviruses in their genome and despite the fact that they carry the Sgp3 allele which can up-regulate the abundance of Xeno-I gp70 RNA in B10 mice. In addition, 129 male mice encode two types of xenotropic proviruses on their Y chromosome, yet do not express transcripts of these proviruses in their livers.

Both B6 and B10 mice congenic for Sgp3 derived from two different strains of mice (NZB and BXSB) display comparably increased levels of serum gp70 (21, 24). However, the analysis of gp70 RNA abundance in their livers revealed that the effect of the Sgp3 locus derived from the NZB strain in the B6 background was not identical to that of the Sgp3 locus derived from the BXSB strain in the B10 background: NZB-Sgp3 and BXSB-Sgp3 most prominently enhanced mPT and PT gp70 RNA levels in B6 and B10 mice, respectively. These differences could be due to allelic variations of Sgp3 between NZB and BXSB mice and/or to differences in the regulation of PT and mPT gp70 structural genes between B6 and B10 mice. In addition, we observed that despite the lack of expression of Xeno-I gp70 RNA in the liver of BXSB mice, B10 mice bearing the BXSB-Sgp3 allele displayed substantial increases in Xeno-I gp70 RNA. This supports the view that Sgp3 does not encode serum gp70 structural genes, but rather regulates the expression of multiple endogenous retroviral transcripts in trans. In this regard, it is worth noting that the Gv1 (Gross virus antigen 1) gene that directly overlaps with the Sgp3 locus (43) regulates the expression of thymic GIX gp70 antigen (13), the expression of which is closely correlated to serum levels of gp70 (38, 44). As Gv1 controls in trans the expression of multiple endogenous retroviral transcripts in different tissues, including the liver (45), it is reasonable to assume that Gv1 and Sgp3 are identical or related genes regulating the transcription of retroviral sequences.

The analysis of B6.NZB-Sgp4 congenic mice confirmed the contribution of the Sgp4 locus to the production of serum gp70, although its effect was more modest than that of Sgp3. BXSB mice contain the Sgp3 locus but the Sgp4 locus has not been identified in the linkage analyses of the genetic crosses between BXSB and B10 mice (10). Since NZB mice contain both Sgp3 and Sgp4, it seems likely that Sgp4 contributes to the elevated serum gp70 levels observed in NZB mice compared with BXSB mice. Notably, the action of the Sgp4 locus is different from that of Sgp3, since the presence of the Sgp4 locus resulted in a selective up-regulation of Xeno-I gp70 RNA and a suppression of Xeno-II and Xeno-III gp70 RNA expression. At present it is difficult to determine whether the observed effect of Sgp4 is due to the presence of one or more different regulatory elements in the NZB Sgp4 interval. However, the finding that the injection of LPS led to increases in not only Xeno-I gp70 RNA but also Xeno-II/III gp70 RNAs in B6.NZB-Sgp4 mice rather suggests that the Sgp4 locus may carry regulatory elements, which independently control the transcription of xenotropic gp70 RNAs in non-inflammatory vs. inflammatory conditions. Further analysis of B6 mice bearing both Sgp3 and Sgp4 loci should help to define the molecular mechanism responsible for the action of Sgp4.

It has long been believed that gp70 derived from the NZB-X1 xenotropic virus is the predominant gp70 present in serum, based on the results obtained by tryptic peptide fingerprint analysis on serum gp70 and various retroviral gp70s (27–29). Since we have shown that the NZB-X1 virus belongs to the Xeno-I subgroup, the selective up-regulation of Xeno-I gp70 RNA in association with a moderate increase in serum gp70 in B6.NZB-Sgp4 congenic mice supports the contribution of Xeno-I gp70 to serum gp70. However, the lack of expression of Xeno-I gp70 RNA in BXSB, NFS and 129 mice, which have relatively high serum levels of gp70 (10–30 μg/ml), indicates that Xeno-I gp70 is not uniformly the major gp70 expressed in serum. In addition, the expression of PT and mPT, but not xenotropic gp70 RNAs in 129 mice clearly indicates that PT- and mPT-derived gp70s are important additional sources of serum gp70. The contribution of polytropic gp70s to serum gp70 in NZB and BXSB mice was further supported by the findings in Sgp3 congenic mice. First, the levels of PT or mPT gp70 RNAs in Sgp3 congenic mice were highly elevated, while the amount of xenotropic gp70 RNA remained very low. Second, the production of serum gp70 in B10.BXSB-Sgp3 mice reached levels close to that observed for BXSB mice, in correlation with a prominent increase of PT gp70 RNA in B10.BXSB-Sgp3 mice to the same level as in BXSB mice. This suggests that PT gp70 may be a substantial component of serum gp70 in the BXSB strain. Taken together, our present data indicate that PT and mPT gp70s are the predominant sources of serum gp70 in at least some strains of mice, and suggest that they are important sources for serum gp70 in general.

It is striking to see that the injection of LPS selectively up-regulated xenotropic and mPT gp70 RNAs, but not PT gp70 RNA, in the liver of NZB, BXSB and B6 Sgp3 mice and xenotropic gp70 RNA in the liver of B6 Spg4 congenic mice, all of which exhibited an increased production of serum gp70 in response to LPS. This selective effect of LPS on xenotropic and mPT gp70 RNA abundance is likely due to the remarkable heterogeneity of the U3 regulatory regions of the long terminal repeat among different classes of endogenous retroviruses. Indeed, the presence of an NF-kB-binding motif (GGAAAGTCCC) has been described in the U3 region of xenotropic viruses (46). Furthermore, we also noted the presence of an IL-6-responsive element (IL6-RE) with the consensus sequence CCGGGAA common to genes encoding other acute phase proteins (47) in the U3 region of some, but not all, xenotropic and mPT viruses. In contrast, these sequences were not present in the U3 region of PT viruses. LPS induces acute phase responses by stimulating the production of cytokines, such as IL-6 and IL-1, and these cytokines induce the synthesis of various acute phase proteins through several distinct signaling pathways, which involve different transcription factors such as NF-kB, C/EBP (CAAT/enhancer binding protein) and the STAT family (48–50). Thus, the presence of both NF-kB motif and IL6-RE in xenotropic viruses, but only IL6-RE in mPT viruses, is compatible with our finding that LPS-induced up-regulation of xenotropic gp70 RNA was much stronger than that of mPT gp70 in NZB, BXSB and B6 Sgp congenic mice.

In agreement with the lack of LPS-induced gp70 responses, the levels of gp70 RNAs were only marginally up-regulated in LPS-injected B6 mice. However, we observed, in parallel to approximately 2-fold increases in serum gp70, substantial up-regulation of xenotropic and mPT gp70 transcripts in the liver of B6.NZB-Sgp3 and of xenotropic gp70 transcripts in the liver of B6.NZB-Sgp4 congenic mice. This indicated the contribution of Sgp3 and Sgp4 to an enhanced production of serum gp70 in response to LPS, thereby at least partially explaining the failure of B6 mice to produce more serum gp70 after injection of LPS. Notably, the extent of increases in serum gp70 in B6 Sgp congenic mice after injection of LPS was still limited (~2 fold), as compared with those (7–10 fold) observed in NZB and BXSB mice. It is possible that the Sgp3 and Sgp4 loci could act additively or synergistically to promote the production of gp70 in response to LPS. Moreover, the LPS-induced gp70 production could be regulated by other genetic factors. In this regard, it has previously been claimed that the Sgp2 locus present on distal chromosome 7, close to the Gv2 locus, regulates the enhanced production of gp70 in response to LPS (39). Again, further analysis in B6 Sgp3 and Sgp4 double congenic mice should help clarify this issue.

The present analysis of the abundance of group- and subgroup-specific retroviral gp70 RNAs and the presence of corresponding proviruses disclosed that the genetic origin of serum gp70 is more heterogeneous than previously believed, and that serum levels of gp70 are under the control of multiple structural and regulatory genes. Moreover, in view of the emerging roles of TLR7 and TLR9 in the development of murine SLE (51, 52), it would also be of interest to explore the possible contributions of TLR7 and TLR9 to the expression of serum gp70 and endogenous retroviruses. In addition to the role of retroviral gp70 as a nephritogenic autoantigen in murine lupus, the possible importance of endogenous retroviruses as a triggering factor for autoimmune responses in SLE has also been highlighted by the production of anti-nuclear autoantibodies in mice infected or immunized with retroviruses (53, 54) and in Sgp3 and GIX congenic mice (23, 24, 55). However, it is still unclear how retroviruses might promote anti-nuclear autoantibody production. Clearly, the eventual identification of mouse genes regulating the production of endogenous retroviruses in general and serum retroviral gp70 in particular will enable us to address the relevance of their human counterparts, and has obvious and promising implications for diagnostic, prognostic and therapeutic approaches to SLE and related autoimmune diseases.