| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print August 17, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Section of Nephrology, Department of Medicine, The University of Illinois at Chicago
2 Correspondence: The University of Illinois at Chicago, Section of Nephrology, 820 South Wood Street, MC-793, Room 479/CSN, Chicago, IL 60612. E-mail: jslevine{at}uic.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) from prediseased mice of the major murine models of lupus have an identical defect in cytokine expression that is triggered by serum and/or apoptotic cells. It is striking that cytokine expression in the absence of serum and apoptotic cells is equivalent to that of nonautoimmune mice. Here, we show that m from prediseased lupus-prone MRL/MpJ (MRL/+) or MRL/MpJ-Tnfrsf6lpr (MRL/lpr) mice also have reversible abnormalities in morphology, cytoskeletal organization, and adhesive properties. In the presence of serum, MRL m adhered in increased numbers to a variety of extracellular matrix proteins compared with m from two nonautoimmune strains. However, in the absence of serum, adhesion by MRL m was similar to that of nonautoimmune m. Increased adhesion by MRL m was also observed in the presence of apoptotic, but not necrotic, cells. The morphology and actin-staining pattern of adherent MRL m were consistent with reduced activity of Rho, a cytoskeletal regulator. Indeed, MRL m cultured in the presence of serum had markedly decreased levels of active Rho compared with nonautoimmune m. It is remarkable that when cultured in the absence of serum, MRL m displayed normal Rho activity and cytoskeletal morphology. Addition of a Rho inhibitor to normal m reproduced the morphologic and cytoskeletal abnormalities observed in MRL m. Taken together, our findings support the hypothesis that m from MRL and other systemic lupus erythematosus-prone mice have an apoptotic, cell-dependent, autoimmune phenotype that affects a broad range of m functions, including cytokine gene expression and Rho-dependent cytoskeletal regulation.
Key Words: rodent • autoimmunity
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), dendritic cells, and other leukocytes, mediate the interactions of these cells with one another, extracellular matrix (ECM) proteins, or endothelial and other cells [1
2
3
4
]. For example, the recruitment of blood monocytes to the peritoneal cavity following thioglycolate injection is dependent on expression of the integrin CD11b/CD18 [5
, 6
]. Differentiation-dependent expression of various selectins and their ligands by T lymphocytes leads to the differential trafficking of naive and effector T lymphocytes to lymph nodes and inflamed peripheral tissue, respectively [1
, 3
, 4
, 7
]. The binding of integrins and other adhesion molecules such as CD44 to ECM proteins leads to the retention of antigen-stimulated T cells in lymph nodes and peripheral sites of infection or inflammation [1
, 3
, 4
, 7
]. Additionally, adhesive interactions play a critical role in stabilization of the immunologic synapse between antigen-presenting cells (APC) and T lymphocytes, thereby modulating intracellular signaling events and the overall activation of antigen-specific T cells [8
, 9
].
Given this broad role for adhesive interactions in immune homeostasis, it is not surprising that alterations in the function or expression of various adhesion molecules can affect the strength and/or target specificity of the immune response in health and disease [4
, 10
, 11
]. For example, administration of a monoclonal antibody (mAb) that blocks adhesion through the integrin CD11b/CD18, which is critical for the recruitment of m [5
], significantly decreased the development of overt autoimmune diabetes from 98% to 3% in a transfer model of diabetes using nonobese diabetic (NOD) mice [12
]. Similarly, targeted deletion of the integrin intracellular adhesion molecule-1 significantly prolonged the lifespan of lupus-prone MRL/MpJ-Tnfrsf6lpr (MRL/lpr) mice and reduced the severity of glomerular disease [13
], whereas targeted deletion of selectins exacerbated autoimmunity [14
].
We have shown that m from prediseased mice of all major murine models of systemic lupus erythematosus (SLE) have an identical defect in cytokine expression, which is triggered by serum [fetal bovine serum (FBS)] lipids and/or apoptotic cells [15
16
17
18
]. Affected strains include MRL/MpJ (MRL/+), MRL/lpr, NZB, NZW, NZB/W F1, BXSB, and LG/J, all of which express or contribute to the development of autoimmunity. No similar defect was found in 16 nonautoimmune control strains [18
, 19
]. This defect may be characteristic of autoimmunity in general, as m from prediseased, diabetes-prone NOD mice also respond abnormally to apoptotic cells, mimicking closely the apoptotic, cell-dependent abnormality that we have observed in SLE-prone strains [19
]. Elicitation of this defect by exposing m from prediseased, autoimmune-prone mice to apoptotic cells leads to the dysregulated expression of multiple cytokines [18
, 19
]. In probing the basis and consequences of this defect, we have used FBS as a surrogate for apoptotic cells, as lipid-containing FBS fully mimics the effect of apoptotic cells in eliciting all aspects of this defect [18
, 19
].
In the course of these studies, we noticed that m from SLE-prone mice also displayed abnormalities in their morphology and adhesive properties, especially when cultured on bacteriologic plastic (BP) or surfaces coated with ECM proteins such as laminin (LAM), fibronectin (FN), or collagen. In the presence of FBS, m from MRL/+ and MRL/lpr mice adhered in increased numbers and with a distinct spread morphology as compared with m from two nonautoimmune strains. When cultured in the absence of FBS, the morphology and adhesive interactions of MRL m became normal. Immunofluorescent (IF) staining of adherent MRL m cultured in the presence but not the absence of FBS revealed an actin-staining pattern consistent with reduced activity of Rho, a cytoplasmic G protein and cytoskeletal regulator [20
, 21
]. In confirmation of these observations, when MRL m were cultured in the presence of FBS, the percentage of active Rho was significantly decreased compared with nonautoimmune m. In sharp contrast, culturing MRL m in the absence of FBS restored the amount of active Rho to normal levels. Addition of a Rho inhibitor to normal m reproduced the morphologic and cytoskeletal abnormalities seen in MRL m. Taken together, these results support the hypothesis that m from MRL and other SLE-prone mice have a serum- and/or apoptotic cell-dependent autoimmune phenotype that affects a broad range of m functions, including cytokine gene expression and Rho-dependent, cytoskeletal regulation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Materials and reagents
Lipopolysaccharide (LPS; Escherichia coli-derived, serotype 0111:B4) was obtained from List Biological (Campbell, CA). Bovine serum albumin (BSA) and normal goat serum were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA). Clostridium botulinum C3 exoenzyme (C3) was obtained from Cytoskeleton (Denver, CO).
Ab
The following mAb were used for flow cytometric analysis: 5C6, rat immunoglobulin G (IgG)2b 
anti-mouse CD11b (BioSource International, Camarillo, CA); MI/70.15.1, rat IgG2b anti-mouse CD11b (BioSource International); LO-DNP-11, rat IgG2b control (BioSource International); fluorescein isothiocyanate (FITC)-labeled goat F(ab')2 anti-rat Ig; phycoerythrin (PE)-labeled rat IgG2a anti-mouse CD18 (PharMingen, San Diego, CA); PE-labeled rat IgG2a anti-mouse CD49d (PharMingen); and PE-labeled rat IgG2a anti-mouse CD49e (PharMingen). The following Ab and reagents were used for IF: 1624, mouse IgG1 anti-human vinculin mAb (Chemicon International, Temecula, CA); rhodamine-conjugated sheep polyclonal anti-mouse polyclonal IgG (Cappel Research Reagents, Durham, NC); Oregon green-conjugated phalloidin (Molecular Probes, Eugene, OR); biotin-conjugated rat IgG2b anti-mouse CD11b (Caltag Laboratories, Burlingame, CA); and Cy5-conjugated streptavidin (Caltag Laboratories).
m culture
Peritoneal exudate cells (PEC) were harvested by lavage 3 days after intraperitoneal injection of 1.5 ml 4.05% thioglycolate broth [15
16
17
18
19
, 22
]. After being washed twice in RPMI-1640, PEC, destined for immediate use in an adhesion assay (see Fig. 1 ), were resuspended 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). All other PEC were cultured overnight under nonadherent conditions for use the following day. During overnight, nonadherent culture, PEC were suspended in one of three media: R.10; delipidated R.10 [R.10 with 10% delipidated FBS (dFBS) instead of FBS]; or FBS-free R.0 medium (R.10 minus FBS) supplemented with 100 U/ml human recombinant macrophage-colony stimulating factor (rM-CSF; generous gift of Genetics Institute, Cambridge, MA). For all experiments comparing control versus SLE-prone MRL strains, we used pooled m harvested from a minimum of at least five mice per strain.
|
cultures, apoptotic thymocytes were washed three times in RPMI 1640 and resuspended in R.0. Necrosis was induced by heating thymocytes to 70°C for 40 min.
Documentation of thymocyte apoptosis
Viable cells were defined as propidium iodide (PI)-negative cells with faint nuclear Hoechst staining. Apoptotic cells were defined as PI-negative cells with bright nuclear Hoechst staining and decreased cell size. Post-apoptotic cells (i.e., apoptotic cells that had lost cell-membrane integrity) were defined as PI-positive cells with bright Hoechst staining and decreased cell size. By these criteria, 
60% of cells in apoptotic cell preparations were apoptotic, 
15% were viable, and 25% were post-apoptotic [18
, 19
, 22
]. Necrotic cells, as defined by increased cell size in association with uptake of PI and faint Hoechst staining, comprised <0.1% of apoptotic cell preparations, as compared with >95% of necrotic cell preparations [18
, 19
, 22
].
Overnight, nonadherent culture
After isolation and washing, PEC were suspended in the appropriate medium (R.10, delipidated R.10, R.0 plus human rM-CSF) and added to 60 x 15 mm tissue-culture dishes at 5–10 x 106 cells per dish. Each dish had been precoated the day before with 4.0 ml 1.5% agarose (Gibco, Grand Island, NY) in phosphate-buffered saline (PBS) and then equilibrated over 24 h against three changes of the same medium in which PEC were to be resuspended. After overnight culture on agarose, PEC were collected, washed in RPMI 1640, resuspended in the same medium as they had been cultured overnight, and then used for experiments.
FBS delipidation
FBS was delipidated by one of two protocols. FBS was mixed 1:1 (v/v) for 2 h at 22°C with butanol:diethyl ether (1:4, v/v) [23] or chloroform:methanol (2:1, v/v) [24
] and 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 dFBS was sterilely filtered for later use.
Adhesion assay
PEC suspended in R.10, delipidated R.10, or R.0 plus human rM-CSF (100 U/ml) were added at 1–2 x 105 cells per well to 96-well plates in the presence or absence of LPS (100 ng/ml). Wells were untreated (BP), tissue culture-treated, or coated with FN, LAM, type I collagen (COL I), COL IV, or poly-D-lysine (P-D-LYS). After 2 or 24 h, wells were washed 3x with RPMI 1640, and the number of remaining adherent cells was quantitated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. For all strains of mice, adherent cells were 98% m, as confirmed by morphologic analysis and nonspecific esterase staining. In virtually all cases (except where adhesion was deliberately inhibited), the percentage of added m adhering to individual wells was 50% for that strain manifesting the greatest degree of adhesion.
MTT assay
The number of adherent, viable m was determined using a modification of the MTT assay [25
], whose output is proportional to the total number of viable cells present. After washing away nonadherent cells, 165 µl MTT dissolved in R.0 (1 mg/ml) was added to each well. After incubation at 37°C for 4 h, the MTT formazan was dissolved by adding 165 µl 10% sodium dodecyl sulfate in 0.01 N HCl. Aliquots from each well were read using a microELISA plate reader with a test wavelength of 570 nm and a reference wavelength of 650 nm. Data are presented as mean ± SE of the raw optical density (OD) values. OD values seen in Figures 1
2
3
may be compared, as may those seen in Figures 4
and 5
, but because of variations in experimental conditions, OD values between figures from these two groups should not be compared.
|
|
|
|
were fixed with 4% paraformaldehye in PBS for 10 min at room temperature and then washed 3x with PBS. All subsequent steps took place at room temperature. m were permeabilized with 0.1% Triton X-100 for 10 min, washed 3x with PBS, and then incubated in blocking buffer (PBS containing 5% normal goat serum, 1% BSA, and 0.1% Triton X-100) for 1 h. m were incubated in blocking buffer containing primary mAb (anti-vinculin and anti-CD11b) and Oregon green-conjugated phalloidin for 1 h, washed 3x with PBS, incubated in blocking buffer containing Cy5-conjugated streptavidin or rhodamine-conjugated sheep polyclonal anti-mouse IgG for 1 h, and washed 3x with PBS. For m adhering to BP, a coverslip was added, whereas for m adhering to LAM-coated coverslips, the coverslip was inverted onto a glass slide. Coverslips were then sealed at their edges with clear nail polish and stored overnight at 4°C for viewing the following day.
Microscopy
Cells were viewed under a Zeiss Axiovert microscope equipped with a Achrostigmat 32x/0.40 Ph1 objective lens and a Micromax digital camera (University of Chicago Integrated Microscopy Core, Cancer Center Digital Light Facility, IL). Images from double- or triple-staining were obtained at the same plane of focus. Image-capture analysis was performed using SlideBook 3.04 (Intelligent Imaging Innovations, Denver, CO), run on a Power Macintosh. Images were saved in a tagged image file format and printed on an Epson 1280 printer.
Flow cytometry
PEC were stained by one of two protocols, depending on the mAb. PEC stained with 5C6 (anti-CD11b) or MI/70 (anti-CD11b) were collected after overnight, nonadherent culture and suspended in R.10 for 30 min at 4°C to block Fc receptors for IgG. PEC were next washed 3x in RPMI 1640, incubated for 1 h at 4°C in flow cytometry buffer (R.0 containing 0.2% delipidated BSA) containing the relevant mAb, washed 3x in flow cytometry buffer, incubated for 1 h at 4°C in flow cytometry buffer containing FITC-labeled secondary mAb, washed 3x in flow cytometry buffer, and finally resuspended in flow cytometry buffer at 106 cells/ml. PEC stained with C71/16 (anti-CD18), R1-2 (anti-CD49d), or 5H10-27 (anti-CD49e) were suspended for 5 min at 4°C in flow cytometry buffer containing Fc Block® (rat anti-mouse CD16/CD32 mAb) at 1 µg per 106 cells. Fc Block® was not washed away. PE-labeled mAb were added directly to cell suspensions. PEC were then incubated for 1 h at 4°C, washed 3x in flow cytometry buffer, and finally resuspended in flow cytometry buffer at 106 cells/ml. Flow cytometric analysis of stained PEC was performed on an Epics ESP flow cytometer (Coulter Electronics, Hialeah, FL). Fluorescence was achieved with 15 mW 488 nm argon ion laser. Emitted fluorescence was detected using 530 and 575 nm bandpass optical filters for FITC and PE fluorescence, respectively. Data were analyzed by Epic Elite software (Coulter Electronics).
Rho assay
After overnight, nonadherent culture in R.10 or R.0 plus human rM-CSF, 8 x 106 PEC were assayed for active Rho • guanosine 5'-triphosphate (GTP). Active Rho • GTP constitutes only 10% of total Rho in cells. As active Rho • GTP is rapidly hydrolyzed to inactive Rho • guanosine 5'-diphosphate (GDP), a pull-down assay was used to separate active Rho • GTP. We used a kit from Cytoskeleton containing a fusion protein between glutathione S-transferase and the Rho-binding domain of the rhotekin protein coupled to glutathione agarose beads. Kit directions were followed, and all samples were maintained on ice throughout. Individual lysates, subjected to the pull-down assay (active Rho•GTP), were then separated by polyacrylamide gel electrophoresis on a 4–20% tris-glycine gradient gel (Invitrogen, Carlsbad, CA) and transferred to a nitrocellulose membrane at 100 V for 1 h. After blocking the membrane overnight at room temperature in 5% nonfat milk in 100 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween 20 (TBST buffer), the membrane was incubated for 1.5 h at room temperature in a 1:500 dilution in TBST of mouse anti-RhoA IgG1 mAb, supplied with the kit. Blots were washed three times with TBST and then incubated with a 1:5000 dilution in TBST of an alkaline phosphatase-conjugated goat anti-mouse Ig, supplied as part of a detection kit and used according to the manufacturer’s instructions (chemiluminescent immunoblot detection system, Applied Biosystems, Bedford, MA).
Statistics
Ten replicate wells per experiment were examined in all adhesion assays, and the results were averaged. A minimum of three experiments was performed for all data points. Data are expressed as mean ± SEM of the averaged values obtained from each experiment. Statistical significance was determined by a two-tailed Student’s t-test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
from SLE-prone MRL/+ and MRL/lpr mice adhere in increased numbers to a variety of surfaces and ECM proteins from SLE-prone MRL/+ and MRL/lpr mice adhered to a variety of surfaces when m were plated in the presence of FBS. To quantify this effect, we used a MTT-based adhesion assay, in which the relative number of m that remained adherent after vigorous washing was assessed by mitochondrial conversion of pale yellow MTT to dark blue crystals of MTT formazan. Linearity of the assay was established for cell numbers ranging from 2000 to 200,000 cells per well by adding known numbers of PEC followed immediately by MTT without intervening washing steps (not shown). Moreover, graphs of MTT conversion (OD650) versus cell number were roughly superimposable with equivalent slopes, irrespective of strain, in the presence and absence of FBS, thereby ensuring that mitochondrial activity per cell did not vary as a function of genetic susceptibility to SLE or experimental conditions (not shown).
As shown in Figure 1A
, in the presence of FBS, an increased proportion of m from SLE-prone MRL/+ and MRL/lpr versus nonautoimmune-prone, control BALB/c and C57BL/6 mice adhered to a variety of surfaces. These surfaces included TCP, non-TCP, non-tissue-culture-treated, or bacteriologic, plastic (BP), and plastic coated with COL I, FN, or LAM. The relative increase in the number of adherent MRL m, compared with BALB/c and C57BL/6 m, ranged from 1.5-fold (TCP, COL I, FN) up to threefold (BP). Adhesion to LAM was marginal for all strains, as it has been shown previously that thioglycollate-elicited peritoneal m require activation with phorbol 12-myristate 13-acetate or LPS before they will adhere to LAM [26
]. Although the results depicted in Figure 1A
were obtained after 24 h of adhesion, an identical pattern of increased adhesion by MRL/+ and MRL/lpr m was observed after 1 and 4 h of adhesion (not shown). Assay results were independent of time, as elicited peritoneal m are terminally differentiated and nonproliferative [22
]. No consistent, statistically significant differences in relative adhesion were seen for MRL/+ versus MRL/lpr m or for BALB/c versus C57BL/6 m.
It should be stressed that we used only young, prediseased MRL mice, less than 6 weeks of age, as a source of m. We did this to minimize any effects of disease or inflammation, 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. Further evidence against an effect of disease derives from the absence of a difference between MRL/+ and MRL/lpr m, as the presence of the lpr mutation accelerates and exacerbates autoimmunity. Moreover, when we compared m from prediseased (<6 weeks) versus overtly diseased (>6 months) MRL/lpr mice, the overall pattern and magnitude of increased adhesion did not differ (not shown). Nonetheless, the possibility still formally existed that subclinical disease or inflammation may have contributed to the observed differences. We therefore repeated these studies in the presence of LPS (100 ng/ml), an extremely potent activator of m, as a means of achieving a similar degree of activation across all m. As shown in Figure 1B
, the presence of LPS had essentially no effect on the pattern of enhanced adhesion seen with MRL m. The sole exception was for LAM-coated surfaces, to which activated m were now able to adhere, as has been previously reported [26
]. MRL m displayed more than threefold greater adhesion to LAM. As in the case of unactivated m, the results in Figure 1B
reflect 24 h of adhesion. An identical pattern of increased adhesion by MRL/+ and MRL/lpr m was also observed after 1 and 4 h of adhesion (not shown). Taken together, these results suggest that increased adhesion by MRL m is independent of disease and may therefore represent an inherent abnormality in these mice.
As opposed to TCP, BP, COL I, FN, and LAM, adhesion to plastic coated with COL IV or P-D-LYS did not differ between MRL and control m, for unactivated and for LPS-activated m (Fig. 1)
. As m adhere to P-D-LYS nonspecifically, in a receptor-independent manner, the absence of an abnormality on P-D-LYS-coated surfaces suggests that increased adhesion by MRL m is likely to be receptor-dependent. Also, as adhesion to TCP, BP, FN, COL I, LAM, and COL IV occurs via a wide range of different receptors, integrin and nonintegrin in nature, our results further suggest that multiple, but not all, adhesion receptors contribute to this abnormality.
Increased adhesion by MRL m is corrected by removal or delipidation of FBS
Aberrant cytokine expression by MRL m is dependent on the presence of FBS lipids and is fully corrected by culturing MRL m under FBS-free conditions or in the presence of dFBS [18
, 19
]. We therefore determined whether increased adhesion by MRL m displayed a similar dependence on FBS. To minimize the effects of prior in vivo exposure to FBS, m were cultured overnight under nonadherent conditions in the presence of FBS (R.10), in the presence of dFBS (delipidated R.10), or in the absence of FBS (R.0 plus rM-CSF). After overnight culture, m were collected, resuspended in fresh medium identical to that used for overnight culture, and plated for adhesion assays. It should be noted that this protocol differs slightly from that used in generating Figure 1 , for which m were plated in R.10 directly after isolation. Nevertheless, as long as MRL m were cultured in R.10, they displayed an identical pattern of increased adhesion on all surfaces, irrespective of whether m were plated directly after isolation or subjected to overnight, nonadherent culture (not shown).
Remarkably, when cultured under FBS-free conditions (Fig. 2
) or in the presence of dFBS (Fig. 3
), MRL m no longer adhered in increased numbers to TCP, BP, FN, COL I, and LAM. Equivalent adhesion of m from all four strains was independent of the state of m activation, occurring in the absence and the presence of LPS (Figs. 2
and 3)
. MRL m adhered normally to COL IV and P-D-LYS when cultured in R.0 or delipidated R.10, as they did when cultured in R.10.
As FBS-free culture in R.0 required the addition of human rM-CSF to maintain m viability, we confirmed that addition of human rM-CSF was not responsible for correcting the abnormality. Supplementation of R.10 with human rM-CSF had no effect on increased adhesion of MRL m (not shown). An effect of rM-CSF is rendered further unlikely by the fact that culture in delipidated R.10, which required no supplementation by rM-CSF, also fully corrected the abnormal adhesion of MRL m (Fig. 3)
.
The receptors contributing to increased adhesion by MRL m have characteristics similar to those of integrins
At least five distinct mechanisms, not mutually exclusive, could contribute to FBS-dependent, increased adhesion. First, MRL m could adhere through a type or family of adhesion receptors not used by nonautoimmune m. Second, MRL m may be a heterogeneous population with increased adhesion attributable to one or more subpopulations not found among control m. Finally, if MRL m use the same receptors as do nonautoimmune m, then MRL m might possess an increased number of one or several adhesion receptors, the affinity of individual adhesion receptors for their ligands might be increased, or cytoskeletal alterations might lead to an overall increase in the avidity of the interaction between adhesion receptors in aggregate and their cognate surface.
As a first step toward distinguishing among these mechanisms, we examined the divalent cation dependency of the abnormality. In these studies, we focused on two surfaces, BP and TCP. We chose BP as an example of a surface to which m adhere predominantly through a single receptor the integrin CD11b/CD18 (also known as Mac-1, CR3, or 
Mß2) [6
]. Adhesion to BP via CD11b/CD18 requires Mg2+ but not Ca2+ [6
]. We chose TCP as an example of a surface to which m adhere through a wide variety of receptors, integrin and nonintegrin [27
]. Nonintegrin receptors include the type A scavenger receptor, which mediates adhesion in a Ca2+- and Mg2+-independent manner [27
].
Chelation of Mg2+ and Ca2+ by EDTA (5 mM) completely inhibited FBS-dependent adhesion to BP, not only by control m as expected but also by MRL m (Fig. 4A
). In contrast, in the presence of EGTA (5 mM), which chelates Ca2+ but not Mg2+, adhesion to BP by control and MRL m occurred normally, presumably through CD11b/CD18, and FBS-dependent, increased adhesion by MRL m was fully manifest (Fig. 5A
). Thus, in the case of BP, the cation dependency of FBS-dependent, increased adhesion by MRL m paralleled that of CD11b/CD18-mediated adhesion [6
] in that it was abrogated by removal of Mg2+ but was unaffected by removal of Ca2+.
In the case of TCP, chelation of Ca2+ and Mg2+ by EDTA led to equivalent adhesion by MRL and control m, despite the presence of FBS (Fig. 4B)
. This implies that FBS-dependent, increased adhesion by MRL m does not extend to the type A scavenger or other nonintegrin receptors that mediate adhesion to TCP in a Ca2+- and Mg2+-independent manner [27
]. In contrast, the abnormality was not corrected by chelation of Ca+2 only with EGTA (Fig. 5B)
. Taken together, these data suggest that only a subset of the many adhesion receptors on the m cell surface contributes to FBS-dependent, increased adhesion by MRL m. Like integrins, the responsible receptors have a requirement for Mg2+ but not Ca2+.
Under FBS-free culture conditions, adhesion to BP and TCP was equal for control versus MRL m, irrespective of the presence of EDTA (Fig. 4)
or EGTA (Fig. 5)
. Adhesion to BP in the presence of EDTA under FBS-free conditions may appear surprising, but, as previously reported [6
, 27
], adhesion to BP in the absence of FBS does not occur through CD11b/CD18 but rather through an unknown receptor(s) in a Ca2+- and Mg2+-independent manner.
These data suggest that MRL m use the same adhesion receptors as do control m. The fact that control m adhere to BP in the presence of FBS predominantly via the integrin CD11b/CD18 [6
] allowed us to examine this question more directly. Saturating concentrations of the blocking anti-CD11b mAb 5C6 [6
] decreased adhesion to BP by control and MRL m to a similar extent (Fig. 6A
). If MRL m used another receptor(s) in addition to CD11b/CD18, then blockade of CD11b/CD18 should have inhibited adhesion by MRL m to a lesser degree than adhesion by control m. Higher concentrations of 5C6 mAb did not increase the percentage of inhibition for control or MRL m (not shown).
|
(Fig. 6B)
. As expected, neither 5C6 nor M1/70 affected adhesion to TCP (Fig. 6A
and 6B)
. Although we did not formally examine the cation dependency of adhesion to surfaces other than BP and TCP, our data suggest that MRL m use the same receptors as do control m. Assuming that FBS-dependent, increased adhesion does not occur through different mechanisms on different surfaces, then a common mechanism must involve abnormalities of receptor number, affinity, and/or avidity.
Increased adhesion by MRL m is not attributable to heterogeneity in m cell populations or to differences in surface expression of integrins
We directly evaluated the surface expression of several integrins on control and MRL m that had undergone overnight, nonadherent culture in the presence or absence of FBS. Integrins examined included CD11b (M), CD18 (ß2), CD49d [very late antigen-4 (VLA-4), 4], and CD49e (VLA-5, 5). No differences in the surface expression of these integrins were seen among nonadherent control and MRL m, irrespective of the presence or absence of FBS (Fig. 7
). Moreover, in all cases, control and MRL m comprised a single, uniform population, as assessed by forward- and side-scatter (not shown) and surface-staining for integrins (Fig. 7)
. We conclude that differences in adhesion receptor number and/or the presence of a heterogeneous MRL m cell subpopulation are unlikely to account fully for FBS-dependent, increased adhesion by MRL m.
|
is associated with FBS-dependent changes in the actin cytoskeleton (Fig. 4B)
further implicates cytoskeleton-mediated changes in avidity, as cells adhering through this receptor attach to surfaces but do not spread, a process requiring cytoskeletal reorganization [20
, 21
, 31
].
We chose BP and LAM as representative surfaces for two reasons. First, the magnitude of increased adhesion by MRL m was greatest for BP and LAM (Fig. 1)
, and second, FBS-dependent, morphologic changes by MRL m were most apparent on these two surfaces. Results for BP and LAM are discussed together, as all observed FBS-dependent changes were equally evident for MRL m adherent to either surface. It is important that the morphologic and cytoskeletal changes demonstrated by MRL m cultured in the presence of FBS are all characteristic of m having reduced or absent activity of the cytoplasmic G protein Rho, a regulator of the cytoskeleton [32
, 33
].
After overnight, nonadherent culture, MRL/lpr and BALB/c m were allowed to adhere to LAM or BP in the presence or absence of FBS. In the presence of FBS, MRL/lpr m assumed a more spread morphology than did BALB/c m, occupying, on average, a larger surface area per cell (Figs. 8
, phase, and 9
). In addition, MRL/lpr m often assumed a highly elongated, dendritic morphology, with very long, delicate, cytoplasmic extensions (Fig. 9
, arrows). In marked contrast, under FBS-free culture, MRL/lpr m were less spread, in many cases, covering an area only slightly greater than that of unspread cells (Figs. 8
, phase, and 9
). BALB/c m, irrespective of the presence or absence of FBS, resembled MRL/lpr m cultured under FBS-free conditions (Figs. 8
, phase, and 9
). These differences were seen irrespective of the surface and irrespective of whether m were cultured under conditions of high or low density (Figs. 8
and 9
, see Fig. 12
, and not shown).
|
|
|
, when cultured in the presence of FBS, reflects the fact that these cells spread more and therefore take up a larger surface area. MRL/lpr and BALB/c m in suspension, irrespective of whether they are cultured in the presence or absence of FBS, are all of the same size and have the same diameter, as assessed by forward-scatter on flow cytometry (not shown).
|
differs in several important ways from that of fibroblasts, the cell most commonly used to study the effect of cytoskeletal organization on morphology and adhesion. As opposed to fibroblasts, cultured m do not form focal adhesion complexes [20
, 32
, 33
]. In keeping with the absence of focal adhesion complexes, Rho-activated m also do not assemble actin into polymerized bundles of stress fibers, which, in fibroblasts, converge on focal adhesion complexes. Instead, Rho-activated m possess a finer contractile network of actin filaments [20
, 32
, 33
]. Without the anchoring provided by focal adhesion complexes, contraction of this actin filament network in m leads to rounding up of m, as opposed to the spread-out morphology seen in Rho-activated fibroblasts [20
, 32
, 33
].
Staining for the focal adhesion protein vinculin confirmed the absence of focal adhesion complexes in BALB/c and MRL/lpr m, irrespective of the presence or absence of FBS (not shown). In accord with the absence of focal adhesion complexes, vinculin was distributed uniformly over the m cell body, without any differences in the staining pattern between BALB/c and MRL/lpr m (not shown).
It is remarkable that, in accord with the increased adhesion shown by MRL/lpr m (Fig. 1)
, we observed dramatic FBS-dependent differences in cytoskeletal organization between MRL/lpr and BALB/c m. In the presence of FBS, the overall intensity of actin staining was consistently diminished in MRL/lpr compared with BALB/c m (Fig. 8
, ACTIN-FITC, and Fig. 10
). In addition, BALB/c m usually contained a subcortical actin band, most visible in a higher plane of focus (Fig. 10
, arrows). MRL/lpr m, in the presence of FBS, consistently lacked this subcortical actin band. Instead of a subcortical actin band, MRL/lpr m frequently displayed a concentrated staining of subcortical actin within dense, plaque-like structures near the cell periphery (Fig. 10) . Notably, in the absence of FBS, the pattern of actin staining did not differ between MRL/lpr and BALB/c m and closely resembled that seen in FBS-cultured BALB/c m (Figs. 8
and 10)
.
The micrographs shown in Figures 8
9
10
were obtained after 16 h of adhesion. Identical, although less marked, FBS-dependent differences in cytoskeletal organization were observed in MRL/lpr m after 6 h of adhesion (not shown). The morphology and staining of C57BL/6 m paralleled that of BALB/c m (not shown). Similarly, MRL/+ and MRL/lpr m were indistinguishable in morphology and staining (not shown). Finally, to confirm our results for vinculin as well as to rule out differences in membrane clustering of integrins, we also stained BP-adherent MRL/lpr and BALB/c m for CD11b. Staining for CD11b was distributed uniformly over the m cell body and closely resembled that for vinculin, with no differences seen between MRL and control m (not shown).
MRL m cultured in the presence of FBS have decreased activated Rho
The morphology and cytoskeletal organization of MRL m, cultured in the presence but not the absence of FBS, was consistent with reduced Rho activity compared with that of BALB/c m [20, 32, 33] (Figs. 8
9
10)
. To determine whether FBS-dependent dysregulation of Rho activity may underlie FBS-dependent, increased adhesion by MRL m, we measured Rho activity directly.
After overnight, nonadherent culture in the presence or absence of FBS, lysates prepared from equal numbers of MRL/lpr and BALB/c m were examined for active Rho, according to assay specifications [34
]. Consistent with the results of cytoskeletal staining, active Rho • GTP was reduced in FBS-cultured MRL/lpr m compared with BALB/c m (Fig. 11
). Under FBS-free conditions, active Rho did not differ between MRL/lpr and BALB/c m and exceeded by several-fold that found in FBS-cultured MRL/lpr m (Fig. 11)
. Thus, the level of active Rho correlated with the observed adhesion abnormality of MRL m. In the presence of FBS, Rho activity of MRL m was reduced compared with that of control m, and MRL m showed relatively increased spreading and adhesion. In the absence of FBS, Rho activity was normalized, and MRL m behaved comparably with control m.
|
reproduces the morphologic and actin cytoskeletal changes seen in MRL m were allowed to adhere to LAM in the presence or absence of the C3 toxin. As shown in Figure 12
, nonautoimmune BALB/c m, in which Rho was inhibited by incubation in the presence of C3, displayed many of the same features as did FBS-cultured MRL m. C3-treated BALB/c m had a more spread, often highly elongated, dendritic morphology, with very long, thin, cytoplasmic extensions (indicated by arrows). The overall intensity of actin staining was diminished in C3-treated BALB/c m. As opposed to non-C3-treated BALB/c m, which consistently contained a subcortical ring of actin, C3-treated BALB/c m lacked this subcortical ring of actin. It should be noted that C3 enters cells via nonspecific pinocytosis [35
] so that only 60% of m showed evidence of Rho inhibition. C3-treated BALB/c m also showed increased adhesion to BP and LAM, as compared with untreated BALB/c m (not shown). However, because of the limited entry of C3, the difference in adhesion between C3-treated versus untreated BALB/c m was less than that between untreated BALB/c and MRL/lpr m (not shown).
Increased adhesion by MRL macrophages can be reproduced by exposure to apoptotic, but not necrotic, thymocytes
To this point, we have used FBS as a surrogate for apoptotic cells. To test the effect of apoptotic cells directly, BALB/c and MRL/lpr m underwent overnight, nonadherent culture in the presence of apoptotic or necrotic thymocytes but in the absence of FBS. MRL/lpr m, exposed to apoptotic thymocytes in the absence (Fig. 13A
) or presence (Fig. 13B)
of LPS, demonstrated approximately twofold greater adhesion to BP than did BALB/c m similarly exposed to apoptotic thymocytes. In contrast, after overnight exposure to necrotic thymocytes, adhesion by MRL/lpr m to BP was no greater than that of BALB/c m.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
from prediseased, SLE-prone, MRL mice have an abnormality in adhesion that is triggered by serum lipids. In the presence of FBS, m from MRL/+ and MRL/lpr mice adhered in increased numbers to a variety of surfaces and ECM proteins, including TCP, BP, COL I, FN, and LAM, as compared with m from two nonautoimmune strains. When MRL m were cultured under FBS-free conditions or in the presence of FBS that had been delipidated, their adhesion was similar to that of m from nonautoimmune mice. Microscopic examination of adherent MRL m cultured in the presence but not the absence of FBS revealed a dendritic spread morphology and an actin-staining pattern consistent with reduced activity of Rho, a cytoplasmic G protein and cytoskeletal regulator [32
, 33
]. This was confirmed by the finding that MRL m, cultured in the presence of FBS, showed decreased levels of active Rho when compared with nonautoimmune m. In sharp contrast, when MRL m were cultured in the absence of FBS, levels of active Rho were increased to normal levels. Thus, Rho activity in MRL m paralleled exactly the FBS-dependent abnormalities that we observed in adhesion and cytoskeletal organization: normal in the absence of FBS and abnormal in the presence of FBS. As in our studies on cytokine gene expression, apoptotic cells and FBS were equivalent in eliciting the defect. In the presence of apoptotic but not necrotic cells, MRL m adhered in increased numbers to BP, similarly to that observed in the presence of FBS. Finally, inhibition of Rho with C3 exoenzyme [35
] in nonautoimmune m reproduced the morphologic and cytoskeletal abnormalities seen in MRL m.
We considered a number of different mechanisms that might contribute to FBS-dependent, increased adhesion by MRL m. These included heterogeneities in the population of MRL m, use of novel or atypical adhesion receptors, increased numbers of receptors, increased affinity of individual receptors, and/or increased avidity of the overall interaction between adhesion receptors in aggregate and their cognate surface. In light of a FBS-dependent decrease in Rho activity, the most likely mechanism is a cytoskeleton-mediated increase in avidity. As m lack focal adhesion complexes (refs. [20
, 32
, 33
], and not shown), Rho-mediated contraction of the actin network leads to cell-rounding, with resultant decreased adhesion and spreading. In contrast, diminished Rho activity permits greater adhesion and spreading of m. Although we present evidence that the first three mechanisms do not contribute substantially to the abnormality, we cannot exclude the possibility that alterations in the affinity of individual adhesion receptors also play a role, perhaps through Rho-mediated signaling events.
Rho is a member of the Ras superfamily of small GTP-binding proteins [20 , 21 ]. Members of this superfamily cycle between an active, GTP-bound state and an inactive, GDP-bound state, thereby functioning as molecular switches or timers. Rho is active only as long as it remains in a GTP-bound form. Upon hydrolysis of GTP to GDP, Rho loses its activity. In its active, GTP-bound form, Rho interacts with effector or target molecules to initiate downstream responses. The proportion of active Rho is determined by two major reactions. GTPase-activating proteins (GAP) augment the low, intrinsic GTPase activity of Ras superfamily members and therefore, decrease the proportion of active Rho. Guanine nucleotide exchange factors (GNEF) catalyze the dissociation of GDP from Ras superfamily members. As the intracellular concentration of GTP is much greater than that of GDP, the spontaneous reassociation of Rho with guanine nucleotides leads to an increased proportion of active Rho.
Regulation of Rho activity is extremely complex, with over 10 GAP and over 30 GNEF identified to date [21
]. As Rho activity in MRL m is normalized under FBS-free conditions, it is unlikely that the Rho gene itself is mutated in MRL mice. Rather, the responsible abnormality most likely involves one or more of the GAP and GNEF, which interact directly with Rho, or other pathways further upstream that regulate these GAP and GNEF.
We have also examined the role of apoptotic cells in triggering the observed abnormalities in adhesion and demonstrated that increased numbers of MRL m adhered in the presence of apoptotic but not necrotic cells. We have previously shown that apoptotic cells fully mimic the effect of FBS in eliciting dysregulated cytokine expression by m from MRL and other SLE-prone strains [18
, 19
]. The simplest hypothesis to explain our data is that MRL mice have a mutation or allelic difference affecting signal transduction within a specific Rho-regulatory pathway, which is triggered by the recognition and/or uptake of apoptotic cells. Although the identity of the responsible lipids in FBS that mimic the effects of apoptotic cells remains undefined, the most likely candidates are oxidatively modified lipids, lipoproteins, or lipid-protein adducts, whose clearance by m depends on many of the same receptors used to clear apoptotic cells [36
37
38
].
The consequences of dysregulated Rho activity may not be limited to the adhesive properties of MRL m. The number of downstream targets and effector functions known to be influenced by Rho is continually increasing [21
]. Many of these are dependent on the actin cytoskeleton and include cell migration [39
], complement-dependent phagocytosis [40
], and cytokinesis [41
]. It is important that activated Rho also regulates a number of other signal transduction pathways that are independent of the cytoskeleton. These include activation of the transcription factors serum response factor [42
] and nuclear factor-B [43
], cell-cycle progression [44
], and cadherin-mediated, cell-cell contact [45
]. Hence, an abnormality in the regulation of Rho could have wide-reaching effects on immune homeostasis, affecting such m functions as migration, trafficking, antigen uptake, and interaction with other immune cells. All of these functions are crucial to the balance between tolerance and immunity. Indeed, alterations in the function or expression of various adhesion molecules have been shown to modulate the course of disease in several murine models of autoimmunity [4
, 10
11
12
13
14
].
Moreover, recent studies have shown that cytoskeletal events within the APC play a critical role, not only in formation of the immunologic synapse but also in T cell activation [46
, 47
], so that the Rho-dependent cytoskeletal abnormalities we have described in MRL m may affect the manner in which T cells are activated. Given that dysregulated Rho activity is dependent on the presence of FBS (a surrogate for apoptotic cells), abnormalities in T cell activation may be limited to those antigens that are presented by APC in the context of apoptotic cells. The potential dependence of dysregulated Rh
