Of the multiple murine models of autoimmunity, the three most closely resembling human systemic lupus erythematosus (SLE) are the MRL/lpr, New Zealand Black/White F1, and male BXSB. Although these strains share many disease characteristics, no common cellular defect has previously been found in prediseased mice from all these strains. We show in this study that macrophages from prediseased mice of all three SLE-prone strains, as well as macrophages from mice whose genomes contribute to the development of SLE (MRL/+, New Zealand White, New Zealand Black, female BXSB, and LG/J), have an identical and profound defect in cytokine expression that is triggered by apoptotic cells. Strikingly, none of 13 nonautoimmune strains tested exhibited this defect. Given that apoptotic Ags have been increasingly recognized as the target of autoantibodies, a defect in cytokine expression that is triggered by apoptotic cells has broad potential to upset the balance between tolerance and immunity.

Apoptotic Ag have increasingly been recognized as the targets of autoantibodies in systemic lupus erythematosus (SLE)3 (1, 2, 3, 4). Although a number of potential abnormalities of apoptosis have been examined in SLE, nearly all have focused on the extent and regulation of apoptosis within specific lymphocyte populations. Very little is known about intracellular signaling events induced by uptake of apoptotic cells. It would seem likely that abnormalities in intracellular signaling elicited by this abundant source of self Ag might lead to a breakdown in self-tolerance and the emergence of autoimmunity. This hypothesis led us to examine the role of apoptotic cells in a defect of cytokine production observed in macrophages (Mφ) from several SLE-prone strains (5, 6, 7, 8, 9).

We have previously shown that peritoneal and bone marrow-derived Mφ from prediseased MRL/+ and MRL/lpr SLE-prone mice display a substantial defect in LPS-induced expression of several cytokines, including IL-1α, IL-1β, and IL-6 (5, 6, 7, 8, 9, 10, 11). As shown by the construction of irradiation chimeras in which bone marrow from SLE-prone MRL mice was transferred into lethally irradiated normal recipients, and vice versa, this defect is intrinsic to the Mφ and independent of the host autoimmune environment (8). Moreover, it is fully manifest in Mφ from mice as young as 1 wk of age, and neither the pattern nor the magnitude of the defect changes with age or the development of disease (8). Underexpression of IL-1 exists at both an mRNA and protein level. It is transcriptionally determined, and equally affects secreted and cell-associated IL-1 bioactivity (5, 6, 7, 8, 9, 10, 11). As assessed by bioassay and densitometry of Western or Northern blots, the defect is progressive over time, increasing in magnitude from ∼2-fold within the first 24 h of culture to as much as 50-fold by 24–48 h (5, 6, 7, 8, 9, 10, 11).

To this point, our standard protocol has been to culture Mφ in medium containing FBS. In the present study, we show that this defect is eliminated by culture in the absence of FBS. The responsible factor(s) in FBS includes lipids, as delipidation of FBS fully eliminates the defect. Moreover, only anionic lipids, as found on the surface of apoptotic cells, or apoptotic cells themselves can reproduce the defect. We show further that these results extend identically to prediseased mice of all other murine models of SLE. Importantly, none of 13 non-SLE-prone strains possesses a similar defect. Given the central role of the Mφ in apoptotic cell clearance and Ag presentation, a defect in the expression of multiple Mφ-derived cytokines that is triggered by apoptotic cells could have broad potential to alter the balance between self-tolerance and immunity.

AKR/J, B6.C3H.gld (B6/gld), B6.MRL.lpr (B6/lpr), BALB/c, BXSB, C3HeJ, C3H/HeN, C57BL/6, (C57BL/6 × New Zealand Black (NZB)) F1, CBA, LG/J, MRL/+, MRL/lpr, NZB, New Zealand White (NZW), NZB/W F1, and SWR/J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Unless otherwise indicated, these mice were used between 4 and 6 wk of age. Diseased MRL/lpr and NZB/W F1 mice were retired female breeders, greater than 8 mo of age (The Jackson Laboratory). All animal protocols were approved by the Institutional Animal Care and Use Committee at Boston Medical Center (Boston, MA).

Recombinant human plate-derived growth factor-BB (PDGF-BB) and recombinant human insulin-like growth factor-1 (IGF-1) were obtained from QCB (Hopkinton, MA). LPS (Escherichia coli derived, serotype 0111:B4) was obtained from List Biological (Campbell, CA). Dioleoyl phosphatidylserine (PS), dioleoyl phosphatidylcholine, and dioleoyl phosphatidylethanolamine (PE) were obtained from Avanti (Alabaster, AL). Murine TNF-α was obtained from Biosource International (Camarillo, CA). Lysophosphatidic acid (LPA), lipoteichoic acid (LTA), okadaic acid, zymosan, linoleic acid, and oleic acid were obtained from Sigma (St. Louis, MO). Latex beads (1.2 μm) were obtained from Seradyne (Indianapolis, IN). LPS-binding protein (LPS-BP) was the generous gift of Dr. Peter S. Tobias (Scripps Research Institute, La Jolla, CA). Neutralizing anti-murine IL-10 mAb and neutralizing anti-murine TGF-β mAb (active against TGF-β1, TGF-β2, and TGF-β3) were obtained from R&D Systems (Minneapolis, MN).

Peritoneal exudate cells were harvested by lavage 3 days after i.p. injection of 1.5 ml of 4.05% thioglycolate broth (5, 6, 7, 8). Cells were washed twice in RPMI 1640 and plated in 60 × 15-mm tissue culture dishes at 4 × 106 cells/dish in R.10 culture medium (RPMI 1640 plus 10% FBS, with 2 mM l-glutamine, 5 mM HEPES, 100 U/ml penicillin, and 100 μg/ml streptomycin). After a 4-h incubation at 37°C, nonadherent cells were removed by washing with RPMI 1640. The remaining adherent cells, >98% Mφ as determined by morphologic examination and nonspecific esterase staining, were cultured in R.10 or FBS-free R.0 medium (R.10 minus FBS) plus human rM-CSF (generous gift of Genetics Institute, Cambridge, MA). Mφ were used after overnight culture in their respective medium.

FBS was delipidated by one of two protocols. FBS was mixed 1:1 (v/v) for 2 h at 22°C with either butanol/diethyl ether (1:4, v/v) (12) or chloroform/methanol (2:1, v/v) (13), then centrifuged to separate organic and aqueous phases. Residual organic solvents were removed from the delipidated aqueous phase by applying a water-aspirator vacuum for 2 h. The resultant delipidated FBS was sterilely filtered for later use.

Apoptosis was induced in freshly isolated murine thymocytes by incubation for 4 h with 5 × 10−6 M hydrocortisone in either R.10 or R.0 medium (1). Before addition to Mφ, thymocytes were washed three times in RPMI 1640 and resuspended in R.0 medium. Viable thymocytes were defined as propidium iodide (PI)-negative cells with faint nuclear Hoechst staining. Apoptotic thymocytes were defined as PI-negative cells with bright nuclear Hoechst staining and decreased cell size. Postapoptotic thymocytes (i.e., apoptotic thymocytes that had lost cell membrane integrity) were defined as PI-positive cells with bright Hoechst staining and decreased cell size. By these criteria, ≥50% of thymocytes were apoptotic, ∼25% were viable, and ∼25% were postapoptotic. The kinetics of induction of apoptosis and the population distribution at 4 h were identical for thymocytes incubated in R.10 vs R.0 medium. Necrotic cells, as defined by increased cell size in association with uptake of PI and faint Hoechst staining, comprised <0.1% of the final cell population.

Total RNA was isolated from Mφ using Trizol Reagent (Life Technologies, Grand Island, NY), according to manufacturer’s instructions. Equivalent amounts of RNA (usually 10–20 μg) were applied to 1.2% agarose gels containing 0.66 M formaldehyde, and subsequently transferred to Nytran membranes (Schleicher & Schuell, Keene, NH). The RNA blots were then prehybridized, hybridized, and washed, as recommended by the manufacturer. The IL-1α probe used in these experiments was a 1.7-kb EcoRI insert fragment from a plasmid containing the murine IL-1α cDNA. The IL-1β probe was a 1.4-kB EcoRI insert fragment from a plasmid containing the murine IL-1β cDNA. Both plasmids were kindly provided by Dr. Patrick Gray, Genentech (South San Francisco, CA). The β-actin probe was a 1.5-kb PstI insert fragment from a plasmid containing the rat brain β-actin cDNA (kindly provided by Dr. Steve Farmer, Boston University School of Medicine, Boston, MA). Insert fragments were radiolabeled by the random primer method using a kit from New England Nuclear Research Products (Boston, MA), according to manufacturer’s specifications. Equal loading of lanes was confirmed by examination of 18S and 28S ribosomal RNA bands, as well as by probing for the housekeeping gene β-actin. In all cases, the densitometric difference among lanes was ≤1.5×.

RNase protection assays were performed using a RiboQuant In Vitro Transcription Kit (PharMingen, San Diego, CA), according to manufacturer’s instructions. All probes were for murine cytokines.

Peritoneal Mφ from prediseased SLE-prone MRL/+ mice showed a profound defect in LPS-induced expression of IL-1α and β (Fig. 1, A, B, and E). This difference in IL-1 expression between nonautoimmune BALB/c and MRL/+ Mφ increased with time. In the case of IL-1α (Fig. 1, A and E), expression by MRL/+ Mφ at 8 h was ∼50% of that by BALB/c Mφ. Although BALB/c Mφ maintained expression of IL-1α through 24 and 48 h (32% and 6% of that at 8 h), expression of IL-1α by MRL/+ Mφ was virtually undetectable at these times (<0.5% of that at 8 h). The results for IL-1β followed a similar pattern (Fig. 1, A and B), with expression of IL-1β by MRL/+ Mφ at 8 h being nearly equivalent to that by BALB/c Mφ. However, while BALB/c Mφ maintained expression of IL-1β at 24 and 48 h (84% and 27% of that at 8 h), expression of IL-1β by MRL/+ Mφ was virtually undetectable at these times (<0.5% of that at 8 h). This defect is not restricted to peritoneal Mφ, but occurs also in bone marrow-derived Mφ induced to differentiate in vitro by M-CSF (5, 8). Moreover, the defect is independent of the manner of Mφ elicitation, being present in thioglycolate-elicited, proteose peptone-elicited, and resident nonelicited peritoneal Mφ (Refs. 5 and 8 , and not shown).

FIGURE 1.

Effect of FBS on LPS-induced IL-1 underexpression. A–H, Primary cultures of thioglycolate-elicited peritoneal Mφ from BALB/c, MRL/+, and C3H/HeJ mice were cultured in the presence (FBS) or absence (FBS-free) of FBS plus the indicated reagents. FBS was used at 10%, except in F, in which dose dependency was determined. All FBS-free Mφ were supplemented with 50 U/ml human rM-CSF to maintain cell viability. Rabbit LPS-BP was added at 100 ng/ml. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h (G) or the indicated times (A, D, H). As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Northern blots were probed for murine IL-1β and IL-1α. Shown are representative blots from n = 5 independent experiments. B, C, E, and F, Relative expression of IL-1α (E, F) and IL-1β message (B, C) in the presence (B, E) or absence (C, F) of FBS was determined by densitometric scanning and normalized for β-actin. The density at 8 h for BALB/c Mφ was arbitrarily set at 100%. Error bars denote mean ± SD.

FIGURE 1.

Effect of FBS on LPS-induced IL-1 underexpression. A–H, Primary cultures of thioglycolate-elicited peritoneal Mφ from BALB/c, MRL/+, and C3H/HeJ mice were cultured in the presence (FBS) or absence (FBS-free) of FBS plus the indicated reagents. FBS was used at 10%, except in F, in which dose dependency was determined. All FBS-free Mφ were supplemented with 50 U/ml human rM-CSF to maintain cell viability. Rabbit LPS-BP was added at 100 ng/ml. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h (G) or the indicated times (A, D, H). As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Northern blots were probed for murine IL-1β and IL-1α. Shown are representative blots from n = 5 independent experiments. B, C, E, and F, Relative expression of IL-1α (E, F) and IL-1β message (B, C) in the presence (B, E) or absence (C, F) of FBS was determined by densitometric scanning and normalized for β-actin. The density at 8 h for BALB/c Mφ was arbitrarily set at 100%. Error bars denote mean ± SD.

Close modal

In multiple experiments, MRL/+ Mφ only rarely revealed detectable IL-1 message at 48 h, even upon prolonged exposure of Northern blots. To highlight the magnitude of this defect, we have chosen blots whose exposure emphasizes the striking lack of IL-1 message at the 24- and 48-h time points (Fig. 1 A). A similar principle guided our choice of blots in all later figures.

Strikingly, the removal of FBS from the culture medium (FBS-free) resulted in equalization of IL-1 mRNA expression between MRL/+ and nonautoimmune BALB/c Mφ (Fig. 1, C, D, and F). In the absence of FBS, expression of both IL-1α and β by MRL/+ Mφ was virtually identical with that of BALB/c Mφ at all time points (Fig. 1, C and F). Thus, this MRL/+ defect is distinct from that of the LPS-resistant C3H/HeJ strain, which carries a mutation in the Tlr4 LPS receptor (14) and showed impaired induction of IL-1 both in the presence and in the absence of FBS (Fig. 1, A and D). The inhibitory effect of FBS on IL-1 expression by MRL/+ Mφ was dose dependent (Fig. 1 G).

As FBS-free culture of Mφ required the addition of M-CSF to maintain Mφ viability, we confirmed that addition of M-CSF was not responsible for correcting the defect. Supplementation of FBS-containing medium with M-CSF had no effect on the defect (Fig. 1,H). Moreover, M-CSF alone, in the absence of LPS, did not induce IL-1 (not shown). Another potential contributory factor in FBS-free culture is the absence of LPS-BP, which enhances the interaction of LPS with the CD14 receptor (15, 16). Consistent with its role solely as a catalyst, the addition of LPS-BP to FBS-free medium did not reconstitute the IL-1 defect (Fig. 1 H). Together, these data indicate that MRL/+ Mφ are able to respond normally to LPS, but that an unidentified factor(s) in FBS results in inhibition of LPS-induced IL-1 expression. It should be noted that FBS-free culture also increased the magnitude of IL-1 expression by BALB/c Mφ, although the degree of change was far less than that seen for MRL/+ Mφ. Thus, the response of MRL/+ Mφ to FBS may represent an exaggeration of a normal regulatory response.

FBS-dependent IL-1 underexpression also occurred in Mφ from prediseased mice of all the other major inbred models of murine SLE (17) (Fig. 2,A). Importantly, none of six non-SLE-prone strains tested in this study (AKR/J, BALB/c, C3H/HeN, C57BL/6, CBA/J, and SWR) (Fig. 2, B and C) or seven strains evaluated previously (A/J, A.Thy, B.10, B.10A, B.10BR, C3HeB/FeJ, and DBA.2) (5, 6, 7, 8, 9, 10, 11) showed a similar defect. As opposed to control strains, SLE-prone strains are characterized by the virtual absence of detectable message at 24 and 48 h, even upon prolonged exposure of Northern blots. This property of SLE-prone Mφ is represented graphically in Fig. 2, E and F, in which the relative expression at 48 h compared with that at 8 h is shown for all strains under both FBS-containing (Fig. 2,E) and FBS-free (Fig. 2 F) conditions. Under FBS-containing conditions, relative expression at 48 h was 44.8 ± 9.9% for nonautoimmune strains and 1.7 ± 2% for SLE-prone strains (p < 0.001). In marked contrast, when Mφ were cultured in the absence of FBS, there was no difference in expression at 48 h between nonautoimmune and SLE-prone strains (79.7 ± 8.6% vs 79.9 ± 5%, p > 0.95).

FIGURE 2.

Only SLE-prone strains demonstrate FBS-dependent IL-1β underexpression. A–D, Peritoneal Mφ from the indicated strains were cultured in the presence or absence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with LPS (100 ng/ml), and Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 3 independent experiments. E and F, Relative expression of IL-1β message at 8 and 48 h was determined by densitometric scanning for each strain and normalized for β-actin. The ratio of expression at 48 h to that at 8 h was calculated and plotted for both FBS-containing (E) and FBS-free conditions (F).

FIGURE 2.

Only SLE-prone strains demonstrate FBS-dependent IL-1β underexpression. A–D, Peritoneal Mφ from the indicated strains were cultured in the presence or absence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with LPS (100 ng/ml), and Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 3 independent experiments. E and F, Relative expression of IL-1β message at 8 and 48 h was determined by densitometric scanning for each strain and normalized for β-actin. The ratio of expression at 48 h to that at 8 h was calculated and plotted for both FBS-containing (E) and FBS-free conditions (F).

Close modal

It should be stressed that, except where indicated, only young prediseased SLE-prone mice were used as a source of Mφ. This was done to minimize any effects of disease, thereby enabling us to distinguish those abnormalities that play a fundamental role in the etiology of SLE from those that are a consequence of established disease. Our results therefore suggest that FBS-dependent IL-1 underexpression is independent of disease and may represent a background trait of SLE-prone mice. Several additional features of this defect bear upon this point. First, among SLE-prone strains, despite broad variation in severity and manifestations of disease (17), the pattern and magnitude of the defect were essentially the same. Second, within individual strains, the presence of the defect was unaffected by disease and its inflammatory sequelae. Thus, FBS-dependent IL-1 underexpression was equally manifest in Mφ from both diseased and prediseased MRL/lpr and NZB/W F1 mice (Fig. 2 A). Finally, as previously shown (8), IL-1 underexpression is one of the very earliest immunologic abnormalities detectable in SLE-prone mice, being fully manifest in MRL/lpr mice as young as 1 wk of age.

Intriguingly, LG/J mice, which account for 75% of the genome of MRL mice (17) and themselves manifest systemic autoimmune disease (18), possessed a similar defect. In contrast, the other three nonautoimmune MRL parental strains, which contribute only 25% of the genome (17), showed normal IL-1 expression.

The fact that IL-1 underexpression is indistinguishable between MRL/+ and MRL/lpr mice implies that the genetic basis for this defect resides in the MRL background and is independent of such genes as lpr or gld, which exacerbate and accelerate the course of autoimmunity, but are themselves weak inducers of autoimmunity (19, 20). This was confirmed by the absence of a defect in C57BL/6 mice homozygous for either lpr or gld (Fig. 2,D). FBS-dependent IL-1 underexpression was also independent of the Y-chromosome-linked accelerant gene Yaa (17, 21), as male and female SLE-prone BXSB mice had similar defects (Fig. 2,A). Finally, no defect was seen in the nonautoimmune first generation offspring of C57BL/6 and NZB mice (22) (Fig. 2 D). This suggests that at least one of the genes contributing to this defect is recessive or, alternatively, that the defect exhibits gene dosage dependency. Together, these results suggest that FBS-dependent IL-1 underexpression represents a shared phenotype of murine SLE.

FBS-dependent IL-1 underexpression was not limited to induction by LPS, but also occurred in response to stimulation by other danger signals, including toxic shock syndrome toxin-1, LTA, and TNF-α (Fig. 3,A). In contrast, induction of IL-1 by MRL/+ Mφ in response to the phosphatase inhibitor okadaic acid or to yeast cell-derived zymosan A particles was identical with that by BALB/c Mφ despite the presence of FBS (Fig. 3 B). Importantly, induction of IL-1 by LPS vs okadaic acid (23) or zymosan A (24) occurs by distinctly different intracellular signaling pathways. These data suggest that the MRL Mφ defect is attributable to an abnormality within only one of several signaling pathways involved in IL-1 induction.

FIGURE 3.

Effect of different inducers on IL-1β underexpression. A and B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with either LPS (100 ng/ml), heat-killed Staphylococcus aureus (HKSA) (source of toxic shock syndrome toxin-1) (106/ml), LTA (5 μg/ml), murine TNF-α (1 μg/ml), okadaic acid (100 nM), or zymosan A (1 mg/ml). As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Northern blots were probed for murine IL-1β. Shown are representative blots from n = 3 independent experiments.

FIGURE 3.

Effect of different inducers on IL-1β underexpression. A and B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with either LPS (100 ng/ml), heat-killed Staphylococcus aureus (HKSA) (source of toxic shock syndrome toxin-1) (106/ml), LTA (5 μg/ml), murine TNF-α (1 μg/ml), okadaic acid (100 nM), or zymosan A (1 mg/ml). As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Northern blots were probed for murine IL-1β. Shown are representative blots from n = 3 independent experiments.

Close modal

We next sought to identify the specific FBS-dependent factor(s) that elicits this defect. Initial characterization of FBS indicated the necessity of lipid factor(s), as delipidation of FBS by two different methods (12, 13) restored IL-1 expression by MRL/+ Mφ to nearly normal levels (Fig. 4,A). Consistent with the requirement of a lipid factor(s), addition to FBS-free medium of PDGF-BB or IGF-1, two prominent cytokines in FBS, did not lead to a defect (Fig. 4,B). We next attempted to reconstitute the IL-1 defect by addition of various lipid factors to FBS-free medium. Neutral lipids, such as PE (Fig. 4,B) and dioleoyl phosphatidylcholine (not shown), did not reproduce the defect. In contrast, the addition to FBS-free medium of several anionic lipids, including LPA, linoleic acid, and PS, resulted in IL-1 underexpression by MRL/+ Mφ (Fig. 4 B). Although striking, reproduction of the defect by anionic lipids was variable and inconsistent, suggesting that they were incomplete stimuli.

FIGURE 4.

Effect of lipids on IL-1β underexpression. Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS, either untreated or delipidated by one of two methods (dFBS1 or dFBS2) (1920 ) (A), or in FBS-free medium supplemented with either human rPDGF-BB (50 ng/ml), human rIGF-I (50 ng/ml), PE (7 μM), LPA (7 μM), linoleic acid (7 μM), or PS (7 μM) (B). Total cellular RNA was harvested at the indicated times after continuous stimulation with LPS (100 ng/ml), and Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 4 independent experiments.

FIGURE 4.

Effect of lipids on IL-1β underexpression. Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS, either untreated or delipidated by one of two methods (dFBS1 or dFBS2) (1920 ) (A), or in FBS-free medium supplemented with either human rPDGF-BB (50 ng/ml), human rIGF-I (50 ng/ml), PE (7 μM), LPA (7 μM), linoleic acid (7 μM), or PS (7 μM) (B). Total cellular RNA was harvested at the indicated times after continuous stimulation with LPS (100 ng/ml), and Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 4 independent experiments.

Close modal

Because apoptotic, but not viable, cells express the anionic lipid PS on their outer cell membrane (25), we considered the possibility that apoptotic cells might be a more complete stimulus for eliciting the defect (Fig. 5,A). Apoptotic thymocytes alone failed to elicit a defect. However, when thymocytes were induced to undergo apoptosis in the presence of FBS and then extensively washed, they now yielded a significant defect by MRL/+ Mφ. Importantly, while neither delipidated FBS (dFBS) nor unopsonized apoptotic thymocytes alone could reproduce the defect, dFBS-opsonized apoptotic thymocytes yielded a powerful defect. This result indicates that interaction of a nonlipid FBS factor, presumably a protein, with the surface of apoptotic cells creates a ligand capable of eliciting the defect. As a final control, viable thymocytes (alone or washed after preincubation with FBS) were unable to produce a defect. Because viable thymocytes eventually undergo apoptosis during prolonged culture, this last result suggests that the necessary nonlipid factor in FBS interacts specifically with the surface of apoptotic, but not viable, thymocytes and, therefore, is lost during the washing of viable cells. Although we have previously shown that the abundant FBS protein β2-glycoprotein I (β2GPI) selectively interacts with apoptotic cells (1), β2GPI-opsonized apoptotic thymocytes failed to elicit a defect (not shown). This implies that proteins other than, or in addition to, β2GPI may be necessary. The effect of dFBS-opsonized apoptotic cells on IL-1 expression by MRL/+ Mφ was dose dependent over a range of apoptotic cell to Mφ ratios extending from 1:1 to 10:1 (Fig. 5 C).

FIGURE 5.

Effect of apoptotic cells on IL-1β underexpression. Peritoneal Mφ (4 × 106) from BALB/c and MRL/+ mice were cultured in FBS-free medium containing 10% FBS or 40 × 106 thymocytes (A), latex beads (1.2 μm diameter) (B), or the indicated number of syngeneic or heterogeneic thymocytes (C). Apoptotic thymocytes were produced by a 4-h incubation with 5 × 10−6 M hydrocortisone in FBS-free medium alone (Ø) or containing 10% FBS, dFBS1, or dFBS2. Viable thymocytes and latex beads were similarly handled, but in the absence of hydrocortisone. Before addition to Mφ, thymocytes and latex beads were extensively washed with FBS-free medium. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h (C) or the indicated times (A, B). Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 3 independent experiments.

FIGURE 5.

Effect of apoptotic cells on IL-1β underexpression. Peritoneal Mφ (4 × 106) from BALB/c and MRL/+ mice were cultured in FBS-free medium containing 10% FBS or 40 × 106 thymocytes (A), latex beads (1.2 μm diameter) (B), or the indicated number of syngeneic or heterogeneic thymocytes (C). Apoptotic thymocytes were produced by a 4-h incubation with 5 × 10−6 M hydrocortisone in FBS-free medium alone (Ø) or containing 10% FBS, dFBS1, or dFBS2. Viable thymocytes and latex beads were similarly handled, but in the absence of hydrocortisone. Before addition to Mφ, thymocytes and latex beads were extensively washed with FBS-free medium. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h (C) or the indicated times (A, B). Northern blots were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 3 independent experiments.

Close modal

To confirm that underexpression of IL-1 occurs specifically in response to apoptotic cells and is not a nonspecific response to phagocytosis, we added latex particles to FBS-free cultures of BALB/c and MRL/+ Mφ. Latex particles were precultured in either FBS-free medium or 10% FBS and then extensively washed in a manner identical with that used for apoptotic thymocytes. Phagocytic uptake of latex particles by MRL/+ Mφ did not elicit a defect in IL-1 expression (Fig. 5 B).

To this point, Mφ had been exposed only to syngeneic apoptotic thymocytes. We next sought to ensure that the defect was intrinsic to the MRL/+ Mφ, and not the result of a difference between BALB/c and MRL/+ thymocytes. We therefore exposed MRL/+ and BALB/c Mφ to syngeneic or heterogeneic FBS-opsonized apoptotic thymocytes (Fig. 5 C). Irrespective of the source of thymocytes, BALB/c Mφ displayed no defect in IL-1 expression, whereas MRL/+ Mφ showed a dose-dependent defect. This result confirms that the defect resides within the MRL/+ Mφ.

Consistent with the idea that FBS-dependent IL-1 underexpression represents a shared phenotype of murine SLE, dFBS-opsonized apoptotic thymocytes mimicked the effect of FBS and elicited a dose-dependent defect in IL-1 expression by Mφ from all the major murine inbred SLE models, as well as the MRL-parental autoimmune LG/J strain (Fig. 6). IL-1 expression by Mφ from five other representative nonautoimmune strains (AKR/J, C3H/HeN, C57BL/6, CBA/J, and SWR) resembled that of BALB/c Mφ and was not down-regulated in response to dFBS-opsonized apoptotic thymocytes (not shown). Thus, there was full concordance between FBS alone and dFBS-opsonized apoptotic cells in producing the defect. Furthermore, delipidation of FBS normalized IL-1 expression in all these same strains (Fig. 6).

FIGURE 6.

Dose dependence of IL-1β underexpression for apoptotic cells by multiple SLE-prone strains. Peritoneal Mφ (4 × 106) from the indicated strains were cultured in FBS-free medium containing either 10% dFBS or the indicated number of syngeneic dFBS-opsonized apoptotic thymocytes. Apoptosis was induced by 4-h incubation with 5 × 10−6 M hydrocortisone in FBS-free medium containing 10% dFBS. Before addition to Mφ, thymocytes were extensively washed with FBS-free medium. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h. Northern blots of total cellular RNA were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 4 independent experiments.

FIGURE 6.

Dose dependence of IL-1β underexpression for apoptotic cells by multiple SLE-prone strains. Peritoneal Mφ (4 × 106) from the indicated strains were cultured in FBS-free medium containing either 10% dFBS or the indicated number of syngeneic dFBS-opsonized apoptotic thymocytes. Apoptosis was induced by 4-h incubation with 5 × 10−6 M hydrocortisone in FBS-free medium containing 10% dFBS. Before addition to Mφ, thymocytes were extensively washed with FBS-free medium. After overnight culture, Mφ were stimulated with LPS (100 ng/ml) for 24 h. Northern blots of total cellular RNA were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×. Shown are representative blots from n = 4 independent experiments.

Close modal

We next used an RNase protection assay to determine whether Mφ-derived cytokines other than IL-1 show a similar FBS-dependent defect in LPS-induced expression (Fig. 7). As compared with BALB/c Mφ, FBS-dependent cytokine expression by MRL/+ Mφ followed one of three patterns: decreased duration and magnitude of expression (IL-1β, IL-6, IL-12 p35, IL-12 p40, GM-CSF, Mφ-inflammatory protein-1β, RANTES, TNF-α); increased expression (IL-10); and equivalent expression (M-CSF, Mφ migration inhibition factor, TGF-β1, TGF-β2, TGF-β3). Northern analysis showed that expression of IL-1α is also decreased (Fig. 1). In all cases, FBS-free culture led to equalization of LPS-induced expression between MRL/+ and BALB/c Mφ. As previously noted in the cases of IL-1α and IL-β (Fig. 1, A and D), FBS-free culture also increased the magnitude of cytokine expression by BALB/c Mφ, although the magnitude of change was far less than that seen for MRL/+ Mφ (Fig. 7). Whereas BALB/c Mφ expressed most cytokines at 24 and 48 h under both FBS-free and FBS-containing conditions, MRL/+ Mφ expressed these same cytokines at 48 h only under FBS-free conditions. These data suggest that the response of MRL/+ Mφ to FBS may represent an exaggeration of a normal regulatory response.

FIGURE 7.

Differential expression of cytokine mRNA in response to FBS. A and B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence or absence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with LPS (100 ng/ml), and RNase protection assays (Phar-Mingen) were performed for the indicated murine cytokines. Murine GAPDH and ribosomal L32 were used as housekeeping genes. Breaks in the blot represent different exposure times for optimal presentation of data. No IL-4 and IFN-γ were detected, even after prolonged exposure, confirming the purity of Mφ cultures. Shown are representative blots from n = 4 independent experiments.

FIGURE 7.

Differential expression of cytokine mRNA in response to FBS. A and B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence or absence of 10% FBS. Total cellular RNA was harvested at the indicated times after stimulation with LPS (100 ng/ml), and RNase protection assays (Phar-Mingen) were performed for the indicated murine cytokines. Murine GAPDH and ribosomal L32 were used as housekeeping genes. Breaks in the blot represent different exposure times for optimal presentation of data. No IL-4 and IFN-γ were detected, even after prolonged exposure, confirming the purity of Mφ cultures. Shown are representative blots from n = 4 independent experiments.

Close modal

Finally, recent reports have shown that phagocytosis of apoptotic cells by human Mφ inhibits LPS-induced expression of several cytokines, including IL-1 (26, 27). Inhibition appears to be mediated by autocrine release of anti-inflammatory factors such as IL-10 or TGF-β1. Because the inhibitory effect of FBS and dFBS-opsonized apoptotic cells on cytokine production by MRL/+ Mφ resembles that induced by IL-10 or TGF-β1, we determined whether IL-1 underexpression by SLE-prone Mφ could be attributable to exaggerated release of inhibitory autocrine factor(s). We addressed this question through three independent sets of studies (Fig. 8).

FIGURE 8.

Secreted factors alone cannot account for IL-1β underexpression. A, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS alone or 10% FBS supplemented with neutralizing anti-IL-10 mAb (10 μg/ml), neutralizing anti-TGF-β (anti-TGF-β1, β2, β3) mAb (20 μg/ml), or both mAb together. These concentrations are 10× greater than those recommended by the company to achieve 100% neutralization of cytokine activity. B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in either FBS-containing medium (−) or freshly obtained FBS-containing conditioned medium (CM) produced by 48-h stimulation with LPS (100 ng/ml) of separate cultures of BALB/c (B) or MRL/+ (M) peritoneal Mφ. C, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in either FBS-free medium (−) or freshly obtained FBS-free conditioned medium (CM) produced by 48-h stimulation with LPS (100 ng/ml) of separate cultures of BALB/c or MRL/+ peritoneal Mφ exposed to 40 × 106 dFBS-opsonized apoptotic syngeneic thymocytes. Control BALB/c and MRL/+ Mφ were exposed to 40 × 106 dFBS-opsonized apoptotic syngeneic thymocytes (Apo [dFBS]) and illustrate the magnitude of the defect in BALB/c vs MRL/+ Mφ used to generate CM. A–C, Mφ cultured in the presence of neutralizing mAb or CM were stimulated with LPS (100 ng/ml) for the indicated times, and Northern blots of total cellular RNA were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×, and there was no differential effect on IL-1 expression by BALB/c vs MRL/+ CM. Shown are representative blots from n = 3 independent experiments.

FIGURE 8.

Secreted factors alone cannot account for IL-1β underexpression. A, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in the presence of 10% FBS alone or 10% FBS supplemented with neutralizing anti-IL-10 mAb (10 μg/ml), neutralizing anti-TGF-β (anti-TGF-β1, β2, β3) mAb (20 μg/ml), or both mAb together. These concentrations are 10× greater than those recommended by the company to achieve 100% neutralization of cytokine activity. B, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in either FBS-containing medium (−) or freshly obtained FBS-containing conditioned medium (CM) produced by 48-h stimulation with LPS (100 ng/ml) of separate cultures of BALB/c (B) or MRL/+ (M) peritoneal Mφ. C, Peritoneal Mφ from BALB/c and MRL/+ mice were cultured in either FBS-free medium (−) or freshly obtained FBS-free conditioned medium (CM) produced by 48-h stimulation with LPS (100 ng/ml) of separate cultures of BALB/c or MRL/+ peritoneal Mφ exposed to 40 × 106 dFBS-opsonized apoptotic syngeneic thymocytes. Control BALB/c and MRL/+ Mφ were exposed to 40 × 106 dFBS-opsonized apoptotic syngeneic thymocytes (Apo [dFBS]) and illustrate the magnitude of the defect in BALB/c vs MRL/+ Mφ used to generate CM. A–C, Mφ cultured in the presence of neutralizing mAb or CM were stimulated with LPS (100 ng/ml) for the indicated times, and Northern blots of total cellular RNA were probed for murine IL-1β. As assessed by densitometric analysis of 18S and 28S ribosomal bands and β-actin expression, variance in loading among lanes was ≤1.5×, and there was no differential effect on IL-1 expression by BALB/c vs MRL/+ CM. Shown are representative blots from n = 3 independent experiments.

Close modal

In the first, we added neutralizing anti-IL-10 mAb, neutralizing anti-TGF-β (TGF-β1, TGF-β2, and TGF-β3) mAb, or both anti-IL-10 and anti-TGF-β mAb to LPS-stimulated BALB/c and MRL/+ Mφ cultured in the presence of FBS. As shown in Fig. 8 A, addition of these neutralizing mAb failed to correct IL-1 underexpression by MRL/+ Mφ. Although these data definitively rule out a significant role for IL-10 and TGF-β in mediating the defect, they do not rule out a role for other potential autocrine inhibitory factors such as platelet activating factor or PGE2 (26).

We therefore designed the following studies using conditioned media (CM). Our reasoning was as follows. First, if an increase in secreted anti-inflammatory mediators by SLE-prone Mφ is responsible for IL-1 underexpression at 24 and 48 h, then culturing nonautoimmune Mφ in CM derived from SLE-prone Mφ stimulated with LPS for 48 h should result in sharply decreased levels of IL-1 production by nonautoimmune Mφ. Correspondingly, culturing SLE-prone Mφ in CM derived from nonautoimmune Mφ should result in increased levels of IL-1 production by SLE-prone Mφ. Second, if IL-1 underexpression by SLE-prone Mφ at 24 or 48 h is the result solely of heightened sensitivity to otherwise normal levels of secreted autocrine mediators, then culturing SLE-prone Mφ in CM derived from nonautoimmune Mφ stimulated with LPS for 48 h should result in comparably decreased levels of IL-1 production by the SLE-prone Mφ.

We generated CM by two means. In the first, BALB/c and MRL/+ Mφ were cultured in the presence of FBS and stimulated with LPS for 48 h. MRL/+ CM generated in this manner had no effect on BALB/c IL-1 production (Fig. 8,B). Similarly, BALB/c CM had no effect on MRL/+ IL-1 production (Fig. 8,B). Analogous results were obtained comparing CM from BALB/c vs NZW Mφ (not shown). To eliminate any potential confounding effect of FBS in CM on IL-1 expression, CM was also generated by exposure to dFBS-opsonized apoptotic thymocytes. BALB/c and MRL/+ Mφ were cultured in FBS-free medium and exposed to dFBS-opsonized apoptotic thymocytes, then stimulated with LPS for 48 h, as done in generating Figs. 5 and 6. Before use, CM was subjected to ultracentrifugation at 13,000 × g for 30 min so as to remove all residual apoptotic cells and bodies. BALB/c and MRL/+ CM generated in this manner also lacked a differential effect on IL-1 expression by freshly cultured Mφ (Fig. 8 C).

Taken together, these results indicate that FBS-dependent IL-1 underexpression is unlikely to represent an exaggeration of the anti-inflammatory effect of apoptotic cell uptake previously described in normal Mφ. These results also indicate that IL-1 underexpression cannot be attributed to a heightened sensitivity of SLE-prone Mφ to an otherwise normal pattern of secreted autocrine factors. Rather, IL-1 underexpression appears to be the consequence of an abnormality within a specific signaling pathway that is directly modulated by serum lipids and/or apoptotic cells.

We have shown that Mφ from prediseased mice from all the major murine models of SLE have a defect in cytokine expression that is triggered by FBS and/or apoptotic cells. Affected strains include MRL/+, MRL/lpr, NZB, NZW, NZB/W F1, and BXSB, all of which express or contribute to the expression of autoimmunity (17). In addition, the MRL parental strain LG/J, which has recently been shown to develop autoimmunity (18), also manifests the defect. No similar defect can be found in 13 nonautoimmune strains, including the remaining three nonautoimmune MRL parental strains. In addition, the defect is independent of genes such as lpr, gld, and Yaa, which accelerate and exacerbate the development of autoimmunity, but in and of themselves are weak inducers of autoimmunity.

Elicitation of this defect in Mφ exposed to FBS or apoptotic cells leads to the dysregulated expression of multiple cytokines. Of 15 cytokines examined, the expression of 9 was down-regulated (IL-1α, IL-1β, IL-6, IL-12 p35, IL-12 p40, GM-CSF, Mφ-inflammatory protein-1β, RANTES, TNF-α), the expression of one was up-regulated (IL-10), and the expression of the remaining 5 was unaffected (M-CSF, Mφ migration inhibition factor, TGF-β1, TGF-β2, TGF-β3). In all cases, culture under FBS-free conditions led to equivalent expression of cytokines. Such broad dysregulation of the pattern of cytokines produced by Mφ has the potential to perturb the overall cytokine network and upset the balance among regulatory and effector B and T cell subsets.

Accumulating evidence suggests that apoptotic cells and their products are the target of autoantibodies across a broad spectrum of autoimmune diseases (1, 2, 3, 4). Many of the most prominent autoantigens in SLE have been shown to localize on the surface of apoptotic cells. These autoantigens include nucleosomal DNA-histone complexes, small nuclear ribonucleoproteins (including the Smith and U1-A Ag), cytoplasmic ribonucleoproteins (SS-A/Ro and SS-A/La), and the target of anti-phospholipid autoantibodies (1, 2, 3, 4). Importantly, self Ag expressed on the surface of apoptotic cells are conformationally intact, as autoantibodies from patients with SLE and/or the anti-phospholipid syndrome are able to recognize and bind to these self Ag on the surface of apoptotic cells. Indeed, the paradigm that autoantigens are expressed on the surface of apoptotic cells may extend to autoimmune diseases other than SLE, as we have shown that myeloperoxidase and proteinase 3, two targets of anti-neutrophil cytoplasmic autoantibodies, are expressed on the surface of apoptotic neutrophils (4).

In addition to surface expression of conformationally intact self Ag, apoptotic cells may also provide an abundant source of self Ag in the form of processed peptide fragments that can be recognized by T cells in the context of MHC molecules. Each day, billions of cells throughout the body die by apoptosis. A major proportion of these apoptotic cells is cleared by professional phagocytes, including the Mφ. At least in the case of MHC class I, peptides from phagocytosed apoptotic cells have been shown to end up in the peptide groove of MHC molecules (28, 29).

Thus, apoptotic cells provide a more or less continuous source of self Ag in two distinct forms, each recognizable by one or the other of the two major Ag receptors of the immune system. When present as intact molecules on the surface of apoptotic cells, apoptotic self Ag will be recognizable by Ab in the form of secreted Ig or membrane-bound Ig within the B cell Ag receptor. In contrast, when presented as linear peptide fragments in the context of MHC molecules by APC such as the Mφ, apoptotic self Ag will be recognizable by TCRs.

The central challenge for an immune system is the discrimination of self from nonself. Although central tolerance mechanisms within the thymus and bone marrow provide an efficient first line of defense against autoimmunity, a certain percentage of potentially autoreactive T and B cells regularly escape deletion and emerge into the periphery. Such escape is most likely a consequence of the close similarity between many foreign and self Ag, and reflects the fact that elimination of all lymphocytes whose receptors bear any affinity for self Ag would leave dangerous holes in the immune repertoire. These potentially autoreactive T and B cells require ongoing monitoring and containment through peripheral tolerance mechanisms (30, 31, 32, 33). We have suggested that presentation of apoptotic Ag may play an active role in maintaining peripheral self-tolerance and that defects in the handling of apoptotic cells have the potential to predispose to autoimmunity (1, 2, 4). It is an axiom of discrimination theory that correct classification, as for example between self and nonself, is considerably enhanced when a template exists for comparison. We propose that apoptotic Ag fulfills the role of template for the immune system, and acts as a renewable, continuously updated source of self Ag against which the immune system can check and reeducate itself.

It is instructive to list the requirements of an ideal template and to examine how well apoptotic self Ag meets these requirements. First, the template should be present wherever needed. Because apoptosis occurs in all organs and tissues throughout the body, apoptotic self Ag exists everywhere that the immune system performs its surveillance. Second, the template should be plentiful. With billions of cells dying each day by apoptosis, apoptotic self Ag should always exist in surplus. Third, the template must be properly tagged as template so it will not be confused with prospective Ag. Although the precise ligands on apoptotic cells that are recognized by phagocytes remain to be identified (34, 35), there are a number of surface changes, such as redistribution of the anionic phospholipid PS from the cytoplasmic to the external leaflet of the cell membrane followed by capture of circulating β2GPI, that appear to definitely mark cells as apoptotic (1, 23, 36, 37, 38). In addition, as a corollary to the “danger model” (39, 40), apoptotic cell death and clearance occur in the complete absence of inflammation, so there are no danger signals available to induce costimulatory molecules necessary for immune activation (26, 27). Fourth, the template must be easily read by the immune system. As discussed above, apoptotic self Ag not only exists in a conformationally intact form on the surface of apoptotic cells (1, 2, 3, 4), but is also processed by APC and presented as linear peptide fragments in the context of MHC (28, 29). Finally, a mechanism must exist for updating the self-template. In other words, reading of apoptotic self Ag should induce signaling cascades that continually update and alert the immune system to potential changes in the definition of self, as for example may occur during pregnancy or puberty (39, 40).

In terms of this model, one may envision multiple abnormalities that could predispose an individual to autoimmunity. Examples include the following: 1) errors in the generation of self-template, perhaps due to abnormalities in the executionary phase of apoptosis; 2) errors in the tagging and/or recognition of apoptotic cells as self-template, involving either the ligands specific to apoptotic cells or the receptors for apoptotic cells on phagocytes; 3) errors in the presentation of self-template to the immune system, perhaps due to abnormalities in processing or presentation of self-Ag by those phagocytes that also function as APC; and 4) errors in the reading, interpretation, and/or updating of self-template in response to apoptotic self Ag. A Mφ abnormality triggered by apoptotic cells, such as we describe, would act to disrupt peripheral tolerance through either of the last two mechanisms. This model would also explain the emergence over time of autoimmunity, as errors accumulated within the template and eventually exceeded a certain critical threshold.

In summary, FBS- and apoptotic cell-dependent IL-1 underexpression appears to represent a shared phenotype for murine SLE. This phenotype may even represent a more generalized feature of autoimmunity, as a similar defect exists in Mφ from nonobese diabetic mice (41), which spontaneously develop autoimmune diabetes mellitus, but not in Mφ from nonautoimmune diabetic strains (J. S. Koh and J. S. Levine, manuscript in preparation). We hypothesize that the basis for this defect lies within a specific signaling pathway that is triggered by the recognition and/or uptake of apoptotic cells. Although the identity of the responsible lipid components in FBS remains undetermined, the most likely candidates are oxidatively modified lipids, lipoproteins, or lipid-protein adducts, whose uptake by Mφ relies upon many of the same receptors as do apoptotic cells (42, 43, 44). Although we have focused on cytokine expression, FBS-dependent abnormalities may affect a broader range of Mφ functions. For example, in the presence of FBS, MRL Mφ showed increased adhesion to a variety of extracellular matrix proteins, whereas, in the absence of FBS, adhesion of MRL Mφ was identical with that of control Mφ (J. S. Koh and J. S. Levine, manuscript in preparation). The presence of a shared phenotype triggered by FBS and/or apoptotic cells among all the major murine SLE models raises the possibility of a common genetic defect (or separate defects within a common signaling pathway) linking these SLE-prone strains.

We dedicate this manuscript to the memory of Dr. Beldon Idelson, a clinician of rare devotion and compassion, whose helping hand reached beyond the clinical realm to nearly all in need, including us.

1

This work was supported by National Institutes of Health Grants AR/AI42732 and a Clinical Scientist Award from the National Kidney Foundation.

3

Abbreviations used in this paper: SLE, systemic lupus erythematosus; β2GPI, β2 glycoprotein I; CM, conditioned medium; dFBS, delipidated FBS; IGF-1, insulin-like growth factor-1; LPA, lysophosphatidic acid; LPS-BP, LPS-binding protein; LTA, lipoteichoic acid; Mφ, macrophage; NZB, New Zealand Black; NZW, New Zealand White; PDGF, platelet-derived growth factor; PE, dioleoyl phosphatidylethanolamine; PI, propidium iodide; PS, dioleoyl phosphatidylserine; R.0, serum-free RPMI-based medium; R.10, RPMI-based medium supplemented with 10% FBS.

1
Price, B. E., J. Rauch, M. A. Shia, M. T. Walsh, W. Lieberthal, H. M. Gilligan, T. O’Laughlin, J. S. Koh, J. S. Levine.
1996
. Antiphospholipid autoantibodies bind to apoptotic, but not nonapoptotic, thymocytes in a β2-glycoprotein I-dependent manner.
J. Immunol.
157
:
2201
2
Koh, J. S., J. S. Levine.
1997
. Apoptosis and autoimmunity.
Curr. Opin. Nephrol. Hypertens.
6
:
259
3
Casciola-Rosen, L. A., G. Anhalt, A. Rosen.
1994
. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface structures on apoptotic keratinocytes.
J. Exp. Med.
179
:
1317
4
Gilligan, H. M., B. Bredy, H. M. Brady, M.-J. Hébert, H. S. Slayter, Y. Xu, J. Rauch, M. A. Shia, J. S. Koh, J. S. Levine.
1996
. Antineutrophil cytoplasmic autoantibodies interact with primary granule constituents on the surface of apoptotic neutrophils in the absence of neutrophil priming.
J. Exp. Med.
184
:
2231
5
Donnelly, R. P., J. Levine, D. W. Hartwell, G. Frendl, M. J. Fenton, D. I. Beller.
1990
. Aberrant regulation of interleukin 1 expression in macrophages from young autoimmune-prone mice.
J. Immunol.
145
:
3231
6
Hartwell, D. W., M. J. Fenton, J. S. Levine, D. I. Beller.
1995
. Aberrant cytokine regulation in macrophages from young autoimmune-prone mice: evidence that the intrinsic defect in MRL macrophage IL-1 expression is transcriptionally controlled.
Mol. Immunol.
10
:
743
7
Levine, J. S., J. S. Koh, D. Hartwell, D. I. Beller.
1998
. Adhesion elicits an intrinsic macrophage defect in interleukin-1 expression by autoimmune-prone MRL mice.
J. Autoimmun.
11
:
141
8
Levine, J. S., B. J. Pugh, D. Hartwell, J. M. Fitzpatrick, A. Marshak-Rothstein, D. I. Beller.
1993
. IL-1 dysregulation is an intrinsic defect in macrophages from MRL autoimmune-prone mice.
Eur. J. Immunol.
23
:
2951
9
Fitzpatrick, J. M., J. S. Koh, D. Hartwell, D. I. Beller, J. S. Levine.
1996
. Dysregulated cytokine expression in vivo in prediseased and diseased autoimmune-prone MRL mice.
Autoimmunity
23
:
217
10
Alleva, D. G., S. B. Kaser, D. I. Beller.
1997
. Aberrant cytokine expression and autocrine regulation characterize macrophages from young MRL+/+ and NZB/W F1 lupus-prone mice.
J. Immunol.
159
:
5610
11
Alleva, D. G., S. B. Kaser, D. I. Beller.
1998
. Intrinsic defect in IL-12 production associated with immune dysfunction the MRL+/+ and New Zealand Black/White F1 lupus-prone mice and the Leishmania major susceptible BALB/c strain.
J. Immunol.
161
:
6878
12
Hardy, R. W. 1986. Purification and characterization of monoclonal antibodies. In Handbook of Experimental Immunology. D. M. Weir, L. A. Herzenberg, C. Blackwell, and L. A. Herzenberg, eds. Blackwell Scientific Publications, Oxford, pp. 13.1–13.13.
13
Goldstein, J. L., S. K. Basu, M. S. Brown.
1983
. Receptor-mediated endocytosis of low-density lipoprotein in cultured cells.
Methods Enzymol.
98
:
241
14
Poltorak, A., X. He, I. Smirnova, M.-Y. Liu, C. Van Huffel, X. Du, D. Birdwell, E. Alejos, M. Silva, C. Galanos, et al
1998
. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene.
Science
282
:
2085
15
Wright, S. D., R. A. Ramos, P. A. Tobias, R. J. Ulevitch, J. C. Mathison.
1990
. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein.
Science
249
:
1431
16
Hailman, E., H. S. Lichenstein, M. M. Wurfel, D. S. Miller, D. A. Johnson, M. Kelley, L. A. Busse, M. M. Zukowski, S. D. Wright.
1994
. Lipopolysaccharide (LPS)-binding protein accelerates the binding of LPS to CD14.
J. Exp. Med.
179
:
269
17
Theofilopoulos, A. N., F. J. Dixon.
1985
. Murine models of systemic lupus erythematosus.
Adv. Immunol.
37
:
269
18
Peng, S. L., M. P. Madaio, J. Craft.
1996
. Systemic autoimmunity in LG/J mice.
Immunol. Lett.
53
:
153
19
Cohen, P. L., R. A. Eisenberg.
1991
. Lpr and gld: single gene models of systemic autoimmunity and lymphoproliferative disease.
Annu. Rev. Immunol.
9
:
243
20
Nagata, S., P. Golstein.
1995
. The Fas death factor.
Science
267
:
1449
21
Izui, S., M. Iwamoto, L. Fossati, R. Merino, S. Takahashi, N. Ibnou-Zerki.
1995
. The Yaa gene model of systemic lupus erythematosus.
Immunol. Rev.
144
:
137
22
Drake, G. C., S. J. Rozzo, T. J. Vyse, E. Palmer, B. L. Kotzin.
1995
. Genetic contributions to lupus-like disease in (NZB × NZW)F1 mice.
Immunol. Rev.
144
:
51
23
Jiang, C., A. T. Ting, B. Seed.
1997
. PPAR-γ agonists inhibit production of monocyte inflammatory cytokines.
Nature
391
:
82
24
Bondeson, J., K. A. Browne, F. M. Brennan, B. M. J. Foxwell, F. Feldmann.
1999
. Selective regulation of cytokine induction by adenoviral gene transfer of IκBα into human macrophages: lipopolysaccharide-induced, but not zymosan-induced, proinflammatory cytokines are inhibited, but IL-10 is nuclear factor-κB independent.
J. Immunol.
162
:
2939
25
Fadok, V. A., D. R. Voelker, P. A. Campbell, J. J. Cohen, D. L. Bratton, P. M. Henson.
1992
. Exposure of phosphatidylserine on the surface of apoptotic lymphocytes triggers specific recognition and removal by macrophages.
J. Immunol.
148
:
2207
26
Fadok, V. A., D. L. Bratton, A. Konowal, P. W. Freed, J. Y. Westcott, P. M. Henson.
1998
. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-β, PGE2, and PAF.
J. Clin. Invest.
101
:
890
27
Voll, R. E., M. Hermann, E. A. Roth, C. Stach, J. R. Kalden.
1997
. Immunosuppressive effects of apoptotic cells.
Nature
390
:
350
28
Albert, M. L., B. Sauter, N. Bhardwaj.
1998
. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs.
Nature
392
:
86
29
Bellone, M., G. Iezzi, P. Rovere, G. Galati, A. Ronchetti, M. P. Protti, J. Davoust, C. Rugarli, A. A. Manfredi.
1997
. Processing of engulfed apoptotic bodies yields T cell epitopes.
J. Immunol.
159
:
5391
30
Parijs, L. V., A. K. Abbas.
1998
. Homeostasis and self-tolerance in the immune system: turning lymphocytes off.
Science
280
:
243
31
Hannum, L. G., D. Ni, A. M. Haberman, M. G. Weigert, M. J. Shlomchik.
1996
. A disease-related rheumatoid factor autoantibody is not tolerized in a normal mouse: implications for the origins of autoantibodies in autoimmune disease.
J. Exp. Med.
184
:
1269
32
Ray, S. K., C. Putterman, B. Diamond.
1996
. Pathogenic autoantibodies are routinely generated during the response to foreign antigen: a paradigm for autoimmune disease.
Proc. Natl. Acad. Sci. USA
93
:
2019
33
Lou, Y.-H., M. F. McElveen, K. M. Garza, K. S. K. Tung.
1996
. Rapid induction of autoantibodies by endogenous ovarian antigens and activated T cells: implication in autoimmune disease pathogenesis and B cell tolerance.
J. Immunol.
156
:
3535
34
Savill, J., V. Fadok, P. Henson, C. Haslett.
1993
. Phagocyte recognition of cells undergoing apoptosis.
Immunol. Today
14
:
131
35
Platt, N., R. P. da Silva, S. Gordon.
1998
. Recognizing death: the phagocytosis of apoptotic cells.
Trends Cell Biol.
8
:
365
36
Koopman, G., C. P. M. Reutelingsperger, G. A. M. Kuijten, R. M. J. Keehnen, S. T. Pals, M. H. J. van Oers.
1994
. Annexin V for flow cytometric detection of phosphatidylserine expression on B cells undergoing apoptosis.
Blood
84
:
1415
37
Morris, R. G., A. D. Hargreaves, E. Duvall, A. H. Wyllie.
1984
. Hormone-induced cell death: surface changes in thymocytes undergoing apoptosis.
Am. J. Pathol.
115
:
426
38
Duvall, E., A. H. Wyllie, R. G. Morris.
1985
. Macrophage recognition of cells undergoing programmed cell death (apoptosis).
Immunology
56
:
351
39
Matzinger, P..
1994
. Tolerance, danger, and the extended family.
Annu. Rev. Immunol.
12
:
991
40
Matzinger, P., E. J. Fuchs.
1996
. Beyond “self” and “non-self”: immunity is a conversation, not a war.
J. NIH Res.
8
:
35
41
Wicker, L. S., J. A. Todd, L. B. Peterson.
1995
. Genetic control of autoimmune diabetes in the NOD mouse.
Annu. Rev. Immunol.
13
:
179
42
Terpstra, V., N. Kondratenko, D. Steinberg.
1997
. Macrophages lacking scavenger receptor A show a decrease in binding and uptake of acetylated low-density lipoprotein and of apoptotic thymocytes, but not of oxidatively damaged red blood cells.
Proc. Natl. Acad. Sci. USA
94
:
8127
43
Fukasawa, M., H. Adachi, K. Hirota, M. Tsujimoto, H. Arai, K. Inoue.
1996
. SRB1, a class B scavenger receptor, recognizes both negatively charged liposomes and apoptotic cells.
Exp. Cell Res.
222
:
246
44
Chang, M.-K., C. Bergmark, J. L. Witztum.
1999
. Monoclonal antibodies against oxidized low-density lipoprotein bind to apoptotic cells and inhibit their phagocytosis by elicited macrophages: evidence that oxidation-specific epitopes mediate macrophage recognition.
Proc. Natl. Acad. Sci. USA
96
:
6353