- Email: [email protected]
Genetic and immunological factors interact in a mouse model of CNS antiphospholipid syndrome
Behavioural Brain Research 169 (2006) 289–293
Research report
Genetic and immunological factors interact in a mouse model of CNS antiphospholipid syndrome Aviva Katzav a,∗ , Yulia Litvinjuk a,d , Chaim G. Pick b , Miri Blank c , Yehuda Shoenfeld c , Pinhas Sirota d , Joab Chapman a,e a
Department of Physiology and Pharmacology, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Israel b Department of Anatomy, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Israel c Department of Medicine B and Research Unit of Autoimmune Diseases, Sheba Medical Center, Tel Hashomer, Israel d Abarbanel Mental Health Center, Bat Yam, Israel e Department of Neurology, Sheba Medical Center, Tel Hashomer, Israel Received 30 August 2005; received in revised form 11 January 2006; accepted 17 January 2006
Abstract The antiphospholipid syndrome (APS) includes systemic and central nervous system (CNS) pathology associated with antibodies to a complex of phospholipids and 2 -glycoprotein I (2 -GPI). We have recently reported the induction of APS associated with behavioral and cognitive deficits in BALB/c female mice that developed 4–5 months after immunization with 2 -GPI. In the present study, we examined the influence of genetic factors on the ability to induce experimental APS with CNS involvement by testing several mouse strains immunized with 2 -GPI. Female mice from five strains were immunized once with 2 -GPI in complete Freund’s adjuvant (CFA) or with CFA alone (controls). Autoantibody levels were examined at 1 and 5 months after immunization. Neurological assessment in a staircase test was performed 4–5 months following the immunization. Induction of APS resulted in elevated levels of antibodies against negatively charged phospholipids and 2 -GPI in all five mouse strains. Autoantibody levels were significantly higher in Balb/c, ICR, and C57BL/6 mouse strains compared to AKR and C3H. aPL levels dropped significantly more in the C57BL/6 compared to Balb/c mice over a period of 4 months. Hyperactivity reflected by higher number of stairs climbed in 3 min, was induced by APS in the Balb/c and ICR, mouse strains. Exploratory behavior reflected by more frequent rears, was seen in the APS-Balb/c and AKR mice. Hypoactivity and less exploration were seen in the APS-C57BL/6 and C3H mice. The study supports a link between high levels of aPL and behavioral changes in a mouse APS model. Qualitative differences in behavioral patterns may be due to nervous system as well as immune genetic factors. The minimal effect of APS in C57BL/6 mice may provide a suitable background for the study of transgenes in these mice. © 2006 Elsevier B.V. All rights reserved. Keywords: Antiphospholipid syndrome; Experimental model; Antibodies; Behavior
1. Introduction The antiphospholipid syndrome (APS) is an autoimmune disorder, manifested by thromboembolic events (arterial and venous), recurrent spontaneous abortions, thrombocytopenia and elevated titers of circulating antiphosphlipid antibodies (aPL). The disorder may occur in isolation (primary APS) or as part of another autoimmune disease (such as systemic lupus erythematosus, SLE). A wide spectrum of other clinical manifestations has been reported in association with APS, including
∗
Corresponding author. Tel.: +972 36405947; fax: +972 36409113. E-mail address: [email protected] (A. Katzav).
0166-4328/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bbr.2006.01.015
valvular heart disease, dermal complications, and neurological disorders involving mainly the central nervous system (CNS) [21]. The neurological complications described in patients with APS include strokes (mainly ischemic events), seizures, ocular disturbances, dementia, migraine, transverse myelitis and chorea [7,11,14]. Of these manifestations, only stroke is definitely caused by APS while more definitive evidence is still needed for the others [5,10]. Better understanding of whether neurological deficits are an integral part of APS will determine whether they may be used in the clinical definition of the syndrome and provide better treatment for the patients. A major difficulty in assessing the causal connection between APS and its neurological manifestations is that the mechanism of neural
290
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
injury is not clear. It is well established that APS is associated with hypercoagulability and may therefore cause neural injury by occluding brain vessels [20,26]. It has also been found that aPL can interact directly with neural tissue and modulate its functions [8,9]. In this respect, experimental models of APS are valuable in elucidating the pathogenic role of aPL. A model of APS, which is of special interest, is induced by immunization with 2 -glycoprotein I (2 -GPI, also known as apolipoprotein H), which is a major cofactor in the binding of aPL [18]. Immunization of experimental animals with 2 -GPI results in raised antibody levels to phospholipids, thrombocytopenia and fetal resorbtion [3,13]. Neurological deficits have been reported in 2-GPI-immunized PL/J mice, mainly in the spinal cord [12] and MRL/++ mice [1,6] but not in a subsequent short study. We have found that a single immunization of Balb/c mice with 2 -GPI induces increased hyperactivity and exploratory behavior after a period of 4 months [16]. Further studies have extended these findings demonstrating that this behavioral change is accompanied by cognitive deficits and that both developed over a period of 4–5 months post immunization [23]. In order for the animal model to be relevant it needs to be tried in a number of mouse strains for the following reasons: To find the strain with best effect, to examine the standard strain for transgenic mice C57BL/6, to see if all strains develop antibodies and behavioral changes. Since hyperactivity in a staircase apparatus was the most sensitive to the development of brain APS in the Balb/c strain, in the present study we used this assay to assess brain involvement in one outbred strain, ICR, which is similar to Balb/c, and three inbred strains, C57BL/6, AKR, and C3H. C57BL/6 is of special interest since it is the background strain widely used for transgenic mouse models and has immunological features different from Balb/c. C57BL/6 and the Balb/c strains develop highly polarized Th1 or Th2 responses, in vitro and in vivo, respectively. These mice exhibit markedly different responses to certain pathogens as well as to self-antigens [25]. 2. Materials and methods 2.1. Animals Female Balb/c, C57BL/6 (C57/B6), ICR, AKR and C3H mice, aged 3 months, were obtained from Animal Resources, Sackler Medical School, Tel Aviv University. The mice were raised under standard conditions, 23 ± 1 ◦ C, 12-h light cycle (7 a.m.–7 p.m.) with ad libitum access to food and water. The Tel Aviv University Animal Welfare Committee approved all procedures.
2.2. Preparation of β2 -GPI Human plasma was used as a source of 2 -GPI purified by the method of Polz et al. [19]. In brief, serum proteins were precipitated by perchloric acid and the remaining supernatant was brought to 0.03 M NaCl and passed through a heparin column (Pharmacia). Fractions containing 2 -GPI were eluted with 0.35 M NaCl and examined by protein electrophoresis followed by silver stain. Fractions used for immunization contained a single major band, which cross reacts by Western blot with a commercial antibody to 2 -GPI (anti-ApoH, Behring, Marburg).
2.3. Induction of experimental APS Mice were immunized by a single intradermal injection into the hind footpads with 10 g of 2 -GPI emulsified in complete Freund’s adjuvant (CFA). Control mice were immunized similarly with CFA alone.
2.4. Serological evaluation The mice were bled by retro-orbital sinus puncture as soon as they completed their behavioral assessment. The sera were separated by centrifugation and stored at −70 ◦ C until assayed. The sera were tested by ELISA for the presence of serum-dependent (2 -GPI) autoantibodies to cardiolipin (CL) as previously described [2].
2.5. Staircase test The staircase apparatus consisted of a polyvinyl chloride (PVC) enclosure with five identical steps, 7.5 cm × 10 cm × 2.5 cm on top of each other. The inner height of the walls above the level of the stairs was constant (12.5 cm) along the whole length of the staircase. The box was placed in a room with constant lighting and isolated from external noise. Each mouse was tested individually. The animal was placed on the floor of the staircase with its back to the staircase. The number of stairs climbed and the number of rears were recorded for a 3-min period. Climbing was defined as each stair on which the mouse placed all four paws; rearing was defined as each instance the mouse rose on hind legs (to sniff the air), either on a stair or leaning against the wall. The number of stairs descended was not taken into account. Before each test, the box was cleaned with a diluted alcohol solution to eliminate smells.
2.6. Data analysis All data were expressed as mean values ± standard error of mean (S.E.M.). Antibody levels, staircase climbing and rearing was compared between the appropriate groups by means of a univarient ANOVA assessing the effects of strain and immunization. Post-hoc analysis was performed by pair wise Fisher’s LSD post-hoc tests between the groups (SPSS).
3. Results 3.1. Autoantibody levels All APS induced strains of mice developed significantly high levels of anti-2 -GPI dependent-CL antibodies compared to their respective adjuvant controls at 5 months after immunization (Fig. 1, p < 0.001 for immunization effect by ANOVA). However, as can be seen in Fig. 1, there were clear differences between the strains in the levels of antibodies present 5 months after immunization as demonstrated by the interaction of strain × immunization (p < 0.001, ANOVA). Similarly, analysis of the level of autoantibodies in only the 2 -GPI immunized animals revealed a strain difference (p < 0.001, ANOVA). As can be seen in Fig. 1, the strain differences are due to the Balb/c and ICR mice developing higher levels of anti-2 -GPI dependentCL antibodies than the C57BL/6, AKR and C3H mice. This was confirmed by post hoc analysis of antibody levels in the five strains (p < 0.01 for all comparisons). There was a significant difference in the baseline antibody level between the different mouse strains controls (p < 0.001, ANOVA). Since C57BL/6 and Balb/c mice are known to differ significantly in features of their immune systems such as relative weighting of Th1 and Th2 responses [25] we compared in detail
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
Fig. 1. Autoantibody level in the sera of the five strains of mice. Significantly high levels of anti-2 -GPI dependent-CL antibodies in all APS induced strains of mice compared to their respective adjuvant controls at 5 months after immunization. Individual autoantibody levels of each mouse and mean ± S.E.M. levels of each group of mice are presented.
the development of antibody response in these mice (Fig. 2). Both strains developed high level of anti-2 -GPI dependent-CL antibodies compared to controls at 1 month after immunization (p < 0.001 for immunization effect by repeated measures ANOVA). Four months later, this level dropped significantly more in the C57BL/6 mice than in the Balb/c mice (p < 0.001 for month × immunization × strain effect by repeated measures ANOVA). There was also significantly more variance in the levels of antibodies in the C57BL/6 immunized mice compared to the Balb/c mice 1 month after immunization (p = 0.007 by F-test). At 5 months, there was no difference in the variance between the two groups. 3.2. Behavior in the staircase test All strains of mice in both APS and control groups behaved normally in their cages. They fed normally and did not display
Fig. 2. Mean ± S.E.M. of anti-2 -GPI dependent-CL antibodies in APS induced (black bars) mice compared to controls (gray bars) at 1 month (1 m) and 5 months (5 m) post immunization (PI) in the Balb/c and C57/B6 mice.
291
Fig. 3. Activity measure in the staircase test in the 5 mouse strains. Mean ± S.E.M. stair climbing events during 3 min of test in the APS induced mice (black bars) compared to their respective adjuvant controls (gray bars) in the different mouse strains.
overt neurological deficits such as limb weakness or circling. There was no mortality in any of the groups of mice, and there was no significant difference in weights between 2 -GPIimmunized APS and adjuvant-immunized control mice in all strains. In the activity measure (stair climbing, Fig. 3) there was a significant interaction of strain × APS (p < 0.001, ANOVA), indicating a significantly different response to immunization between the strains. The Balb/c-APS mice had significantly higher number of stairs climbed compared to Balb/c controls (p = 0.002, ANOVA), whereas the C57BL/6-APS mice had significantly less number of stairs climbed compared to C57BL/6 controls (p = 0.001, ANOVA). The ICR-APS mice had significantly higher number of stairs climbed compared to ICR controls (p = 0.05, ANOVA), and no significant difference was found between the AKR-APS and C3H-APS and their respective control mice (p > 0.215, ANOVA). There was a significant difference in the baseline behavior between the different mouse strains controls (p < 0.001, ANOVA). The C57BL/6 and the ICR controls were significantly more active compared to Balb/c, AKR, and C3H controls (p < 0.045 ANOVA for all comparisons). The effect of immunization did not correlate with the baseline behavior of controls in that there was an opposite effect of immunization on the stair climbing measure in the C57BL/6 compared to ICR mice, and a significant difference in the Balb/c compared to AKR and C3H mice. Similar trends were found in the rearing exploratory behavior measure (Fig. 4). There was a significant interaction of strain × APS (p = 0.001, ANOVA). The Balb/c-APS mice had significantly more rearing movements compared to Balb/c controls (p = 0.003, ANOVA), whereas the C57BL/6-APS and C3HAPS mice had significantly less number of rearings compared to their respective controls (p < 0.04, ANOVA). 2 -GPI immunization induced increased rearing behavior in both AKR (p = 0.037, ANOVA) and ICR (p = 0.3, ANOVA) mice.
292
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
Fig. 4. Exploratory behavior measure in the staircase test in the five mouse strains. Mean ± S.E.M. rearing events during 3 min of test in the APS induced mice (black bars) compared to their respective adjuvant controls (gray bars) of the different mouse strains.
3.3. Correlation between autoantibody level and behavior in the staircase test The effect of strain difference on antibody levels and behavioral measures was further analyzed. Due to difference in baseline antibody levels and behavior between the different mouse strains controls, data of APS mice were normalized in relation to control mice. Normalized antibody data were calculated by subtracting the mean control value for each strain from each individual APS mouse. Normalized behavioral data were obtained by presenting the measures of each APS mouse as percent of the mean of the controls in that group. Using these data, a regression analysis revealed a significant correlation between the normalized antibody levels and the behavioral measures of activity (stair-climbing, r = 0.318, p = 0.035) and exploratory behavior (rearing, r = 0.319, p = 0.034). To examine whether additional immunization induced factors significantly affect behavior, we performed an MANOVA with antibody levels as a cofactor. This analysis revealed a significant effect of strain on behavior (p < 0.001) and a significant interaction of strain × APS (p < 0.042). 4. Discussion The results presented support a significant behavioral effect of exposure to antiphospholipid antibodies in five strains of mice. This replicates our previous studies on Balb/c mice [16,23] and adds the interesting finding that the immunological and behavioral effects are strain-dependent. Correlation analysis indicated both a significant effect of antibodies but also other immune induced factors in the behavioral effect of APS. The most striking result from the staircase assay was that the behavioral effect in the C57BL/6 mice is opposite to what we have previously found in Balb/c mice [16,23] and to both Balb/c and ICR mice in the present study. The lack of effect in AKR and C3H mice may simply reflect the low levels of autoantibodies induced in these strains. Hyperactivity found in Balb/c mice
corroborates previous observations that when aPL are induced in Balb/c mice by a very different method (anti-ideotype immunization by monoclonal antibodies) [27] this is also associated with hyperactivity. The decreased activity of 2 GPI-immunized C57BL/6 mice has also been replicated in a series of experiments we have performed in transgenic animals in which congenic wild type mice of this strain served as controls (paper in preparation). The differences between the strains seem therefore to be well established and could be accounted for by either immunological or neurological differences or both. The data collected in the present study indicate that over a period of 4 months the antibody levels in the C57BL/6 mice did drop to lower levels than the Balb/c mice but it is difficult to envisage how this could cause an opposite effect, rather than lack of effect, indicating that there must be some interaction with the behavioral patterns in these strains. In this matter it is interesting to note that the AKR mice in the present study show a trend to hyperactivity and increased exploratory on the staircase test even though they attained lower levels of antibodies than the C57BL/6 group, whereas C3H mice show a trend to hypoactivity and decreased exploratory in the staircase test even though they attained similar levels of antibodies compared to the AKR group. The baseline (control group) levels of stair climbing and rearing by themselves do not seem to predict the effect of APS on behavior as can clearly seen in Figs. 3 and 4. It is also possible that in addition to the antibody response measured in the Balb/c and C57BL/6 strains, cellular immune responses may differ between these strains and explain the effect on the brain. There is little evidence, however, in both human and animal studies for a significant cellular immunological process affecting the brain in APS. It is likely therefore that similar processes affect the mice brains in all strains. Previous studies with MRL/lpr and NZB × NZW F1 mice have found decreased activity in genetic animal models of APS [15,22] and we propose that these differences may be due in part to speciesspecific behavior. There is much data on behavioral performance and difference of locomotor activity between different strains [4,17,24]. Low locomotor activity and exploration in Balb/c compared to C57BL/6 control mice found in the present study was similar to previous findings comparing different strains behavior in the staircase test [17]. The interaction of genetic factors with immune challenge may be relevant to findings in human SLE and APS patients. Few specific details are available regarding the genetic factors predisposing to APS and there are no data on what genetic factors predispose APS patients to develop behavioral and cognitive disturbances. Studies such as the present one may offer an experimental approach in which to study this important question. A potential problem in this approach, however, is that it is the Balb/c strain which develop the most pronounced effects in behavior and has also been shown to display cognitive deficits in a previous report [23]. This strain, however, is not widely used to generate transgenic animals in contrast to the C57BL/6 strain which has become the standard in this field. Another drawback of the Balb/c strain is its inability to swim, making this strain unsuitable for the many standardized and straightforward cognitive tests that are performed in water.
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
In order to improve our APS mouse model we are currently attempting to find transgenic components in which we would predispose C57BL/6 mice to develop both behavioral and cognitive changes. Such factors include Factor V Leiden mutation and the amyloid precursor protein (APP) model of Alzheimer’s disease. The specific pathological findings, which lead to the behavioral changes described in this study, are not known. We have preliminary histological data from these mice which do not show large infarcts, which is compatible with the normal neurological function of the mice. We hope that when definite histopathological changes are found to differentiate Balb/c APS mice from their controls these measures will be examined in the brains of other strains in order to verify their pathogenic importance. It is assumed that such findings in the mouse model will be immediately applicable to human disease since they will shed light on the importance of immune mediated processes affecting blood vessels and neuronal structures. Acknowledgments This study was performed in partial fulfillment of the requirements for a PhD degree of Aviva Katzav, Sackler Faculty of Medicine, Tel Aviv University, Israel. We thank Dr. Naam Kariv for invaluable advice and help. Supported by the Shreiber Fund for Medical Research. References [1] Aron AL, Cuellar ML, Brey RL, McKeown S, Espinoza LR, Shoenfeld Y, et al. Early onset of autoimmunity in MRL/++ mice following immunization with beta 2 glycoprotein I. Clin Exp Immunol 1995;101:78– 81. [2] Bakimer R, Fishman P, Blank M, Hohmman A, Sredni B, Djaldetti M, et al. Induction of primary antiphospholipid syndrome in mice by immunization with a human monoclonal anticardiolipin antibody (H-3). J Clin Invest 1992;89:1558–63. [3] Blank M, Faden D, Tincani A, Kopolovic J, Goldberg I, Gilburd B, et al. Immunization with anticardiolipin cofactor (beta-2-glycoprotein I) induces experimental antiphospholipid syndrome in naive mice. J Autoimmun 1994;7:441–55. [4] Bothe GW, Bolivar VJ, Vedder MJ, Geistfeld JG. Behavioral differences among fourteen inbred mouse strains commonly used as disease models. Comp Med 2005;55:326–34. [5] Brey RL, Chapman J, Levine SR, Ruiz-Irastorza G, Derksen RH, Khamashta M, et al. Stroke and the antiphospholipid syndrome: consensus meeting Taormina 2002. Lupus 2003;12:508–13. [6] Brey RL, Cote SA, Teale JM. Autoimmune disease is accelerated in MRL/++ mice after apolipoprotein-H immunization. Neurology 1994;44(Suppl 2):A316. [7] Brey RL, Escalante A. Neurological manifestations of antiphospholipid antibody syndrome. Lupus 1998;7:S67–74. [8] Caronti B, Pittoni V, Palladini G, Valesini G. Anti-beta 2-glycoprotein I antibodies bind to central nervous system. J Neurol Sci 1998;156: 211–9.
293
[9] Chapman J, Cohen-Armon M, Shoenfeld Y, Korczyn AD. Antiphospholipid antibodies permeabilize and depolarize brain synaptoneurosomes. Lupus 1999;8:127–33. [10] Chapman J, Rand JH, Brey RL, Levine SR, Blatt I, Khamashta MA, et al. Non-stroke neurological syndromes associated with antiphospholipid antibodies: evaluation of clinical and experimental studies. Lupus 2003;12:514–7. [11] Chapman J, Shoenfeld Y. Neurological and neuroendocrine-cytokine inter-relationship in the antiphospholipid syndrome. Ann N Y Acad Sci 2002;966:415–24. [12] Garcia CO, Kanbour-Shakir A, Tang H, Molina JF, Espinoza LR, Gharavi AE. Induction of experimental antiphospholipid antibody syndrome in PL/J mice following immunization with beta 2 GPI. Am J Reprod Immunol 1997;37:118–24. [13] Gharavi AE, Sammaritano LR, Wen J, Elkon KB. Induction of antiphospholipid autoantibodies by immunization with beta 2 glycoprotein I (apolipoprotein H). J Clin Invest 1992;90:1105–9. [14] Katzav A, Chapman J, Shoenfeld Y. CNS dysfunction in the antiphospholipid syndrome. Lupus 2003;12:903–7. [15] Katzav A, Kloog Y, Korczyn AD, Niv H, Karussis DM, Wang N, et al. Treatment of MRL/lpr mice, a genetic autoimmune model, with the Ras inhibitor, farnesylthiosalicylate (FTS). Clin Exp Immunol 2001;126:570–7. [16] Katzav A, Pick CG, Korczyn AD, Oest E, Blank M, Shoenfeld Y, et al. Hyperactivity in a mouse model of the antiphospholipid syndrome. Lupus 2001;10:496–9. [17] Lepicard EM, Joubert C, Hagneau I, Perez-Diaz F, Chapouthier G. Differences in anxiety-related behavior and response to diazepam in BALB/cByJ and C57BL/6J strains of mice. Pharmacol Biochem Behav 2000;67:739–48. [18] McNeil HP, Simpson RJ, Chesterman CN, Krilis SA. Anti-phospholipid antibodies are directed against a complex antigen that includes a lipidbinding inhibitor of coagulation: beta 2-glycoprotein I (apolipoprotein H). Proc Natl Acad Sci USA 1990;87:4120–4. [19] Polz E, Wurm H, Kostner GM. Investigations on beta 2-glycoprotein-I in the rat: isolation from serum and demonstration in lipoprotein density fractions. Int J Biochem 1980;11:265–70. [20] Roubey RA, Hoffman M. From antiphospholipid syndrome to antibodymediated thrombosis. Lancet 1997;350:1491–3. [21] Ruiz-Irastorza G, Khamashta MA, Hughes GR. Hughes syndrome crosses boundaries. Autoimmun Rev 2002;1:43–8. [22] Schrott LM, Crnic LS. Anxiety behavior, exploratory behavior, and activity in NZB × NZW F1 hybrid mice: role of genotype and autoimmune disease progression. Brain Behav Immun 1996;10:260–74. [23] Shrot S, Katzav A, Korczyn AD, Litvinju Y, Hershenson R, Pick CG, et al. Behavioral and cognitive deficits occur only after prolonged exposure of mice to antiphospholipid antibodies. Lupus 2002;11:736–43. [24] Weizman R, Paz L, Backer MM, Amiri Z, Modai I, Pick CG. Mouse strains differ in their sensitivity to alprazolam effect in the staircase test. Brain Res 1999;839:58–65. [25] Zhang F, Liang Z, Matsuki N, Van Kaer L, Joyce S, Wakeland EK, et al. A murine locus on chromosome 18 controls NKT cell homeostasis and Th cell differentiation. J Immunol 2003;171:4613–20. [26] Ziporen L, Polak-Charcon S, Korczyn DA, Goldberg I, Afek A, Kopolovic J, et al. Neurological dysfunction associated with antiphospholipid syndrome: histopathological brain findings of thrombotic changes in a mouse model. Clin Dev Immunol 2004;11:67–75. [27] Ziporen L, Shoenfeld Y, Levy Y, Korczyn AD. Neurological dysfunction and hyperactive behavior associated with antiphospholipid antibodies. A mouse model. J Clin Invest 1997;100:613–9.
Research report
Genetic and immunological factors interact in a mouse model of CNS antiphospholipid syndrome Aviva Katzav a,∗ , Yulia Litvinjuk a,d , Chaim G. Pick b , Miri Blank c , Yehuda Shoenfeld c , Pinhas Sirota d , Joab Chapman a,e a
Department of Physiology and Pharmacology, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Israel b Department of Anatomy, Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Israel c Department of Medicine B and Research Unit of Autoimmune Diseases, Sheba Medical Center, Tel Hashomer, Israel d Abarbanel Mental Health Center, Bat Yam, Israel e Department of Neurology, Sheba Medical Center, Tel Hashomer, Israel Received 30 August 2005; received in revised form 11 January 2006; accepted 17 January 2006
Abstract The antiphospholipid syndrome (APS) includes systemic and central nervous system (CNS) pathology associated with antibodies to a complex of phospholipids and 2 -glycoprotein I (2 -GPI). We have recently reported the induction of APS associated with behavioral and cognitive deficits in BALB/c female mice that developed 4–5 months after immunization with 2 -GPI. In the present study, we examined the influence of genetic factors on the ability to induce experimental APS with CNS involvement by testing several mouse strains immunized with 2 -GPI. Female mice from five strains were immunized once with 2 -GPI in complete Freund’s adjuvant (CFA) or with CFA alone (controls). Autoantibody levels were examined at 1 and 5 months after immunization. Neurological assessment in a staircase test was performed 4–5 months following the immunization. Induction of APS resulted in elevated levels of antibodies against negatively charged phospholipids and 2 -GPI in all five mouse strains. Autoantibody levels were significantly higher in Balb/c, ICR, and C57BL/6 mouse strains compared to AKR and C3H. aPL levels dropped significantly more in the C57BL/6 compared to Balb/c mice over a period of 4 months. Hyperactivity reflected by higher number of stairs climbed in 3 min, was induced by APS in the Balb/c and ICR, mouse strains. Exploratory behavior reflected by more frequent rears, was seen in the APS-Balb/c and AKR mice. Hypoactivity and less exploration were seen in the APS-C57BL/6 and C3H mice. The study supports a link between high levels of aPL and behavioral changes in a mouse APS model. Qualitative differences in behavioral patterns may be due to nervous system as well as immune genetic factors. The minimal effect of APS in C57BL/6 mice may provide a suitable background for the study of transgenes in these mice. © 2006 Elsevier B.V. All rights reserved. Keywords: Antiphospholipid syndrome; Experimental model; Antibodies; Behavior
1. Introduction The antiphospholipid syndrome (APS) is an autoimmune disorder, manifested by thromboembolic events (arterial and venous), recurrent spontaneous abortions, thrombocytopenia and elevated titers of circulating antiphosphlipid antibodies (aPL). The disorder may occur in isolation (primary APS) or as part of another autoimmune disease (such as systemic lupus erythematosus, SLE). A wide spectrum of other clinical manifestations has been reported in association with APS, including
∗
Corresponding author. Tel.: +972 36405947; fax: +972 36409113. E-mail address: [email protected] (A. Katzav).
0166-4328/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bbr.2006.01.015
valvular heart disease, dermal complications, and neurological disorders involving mainly the central nervous system (CNS) [21]. The neurological complications described in patients with APS include strokes (mainly ischemic events), seizures, ocular disturbances, dementia, migraine, transverse myelitis and chorea [7,11,14]. Of these manifestations, only stroke is definitely caused by APS while more definitive evidence is still needed for the others [5,10]. Better understanding of whether neurological deficits are an integral part of APS will determine whether they may be used in the clinical definition of the syndrome and provide better treatment for the patients. A major difficulty in assessing the causal connection between APS and its neurological manifestations is that the mechanism of neural
290
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
injury is not clear. It is well established that APS is associated with hypercoagulability and may therefore cause neural injury by occluding brain vessels [20,26]. It has also been found that aPL can interact directly with neural tissue and modulate its functions [8,9]. In this respect, experimental models of APS are valuable in elucidating the pathogenic role of aPL. A model of APS, which is of special interest, is induced by immunization with 2 -glycoprotein I (2 -GPI, also known as apolipoprotein H), which is a major cofactor in the binding of aPL [18]. Immunization of experimental animals with 2 -GPI results in raised antibody levels to phospholipids, thrombocytopenia and fetal resorbtion [3,13]. Neurological deficits have been reported in 2-GPI-immunized PL/J mice, mainly in the spinal cord [12] and MRL/++ mice [1,6] but not in a subsequent short study. We have found that a single immunization of Balb/c mice with 2 -GPI induces increased hyperactivity and exploratory behavior after a period of 4 months [16]. Further studies have extended these findings demonstrating that this behavioral change is accompanied by cognitive deficits and that both developed over a period of 4–5 months post immunization [23]. In order for the animal model to be relevant it needs to be tried in a number of mouse strains for the following reasons: To find the strain with best effect, to examine the standard strain for transgenic mice C57BL/6, to see if all strains develop antibodies and behavioral changes. Since hyperactivity in a staircase apparatus was the most sensitive to the development of brain APS in the Balb/c strain, in the present study we used this assay to assess brain involvement in one outbred strain, ICR, which is similar to Balb/c, and three inbred strains, C57BL/6, AKR, and C3H. C57BL/6 is of special interest since it is the background strain widely used for transgenic mouse models and has immunological features different from Balb/c. C57BL/6 and the Balb/c strains develop highly polarized Th1 or Th2 responses, in vitro and in vivo, respectively. These mice exhibit markedly different responses to certain pathogens as well as to self-antigens [25]. 2. Materials and methods 2.1. Animals Female Balb/c, C57BL/6 (C57/B6), ICR, AKR and C3H mice, aged 3 months, were obtained from Animal Resources, Sackler Medical School, Tel Aviv University. The mice were raised under standard conditions, 23 ± 1 ◦ C, 12-h light cycle (7 a.m.–7 p.m.) with ad libitum access to food and water. The Tel Aviv University Animal Welfare Committee approved all procedures.
2.2. Preparation of β2 -GPI Human plasma was used as a source of 2 -GPI purified by the method of Polz et al. [19]. In brief, serum proteins were precipitated by perchloric acid and the remaining supernatant was brought to 0.03 M NaCl and passed through a heparin column (Pharmacia). Fractions containing 2 -GPI were eluted with 0.35 M NaCl and examined by protein electrophoresis followed by silver stain. Fractions used for immunization contained a single major band, which cross reacts by Western blot with a commercial antibody to 2 -GPI (anti-ApoH, Behring, Marburg).
2.3. Induction of experimental APS Mice were immunized by a single intradermal injection into the hind footpads with 10 g of 2 -GPI emulsified in complete Freund’s adjuvant (CFA). Control mice were immunized similarly with CFA alone.
2.4. Serological evaluation The mice were bled by retro-orbital sinus puncture as soon as they completed their behavioral assessment. The sera were separated by centrifugation and stored at −70 ◦ C until assayed. The sera were tested by ELISA for the presence of serum-dependent (2 -GPI) autoantibodies to cardiolipin (CL) as previously described [2].
2.5. Staircase test The staircase apparatus consisted of a polyvinyl chloride (PVC) enclosure with five identical steps, 7.5 cm × 10 cm × 2.5 cm on top of each other. The inner height of the walls above the level of the stairs was constant (12.5 cm) along the whole length of the staircase. The box was placed in a room with constant lighting and isolated from external noise. Each mouse was tested individually. The animal was placed on the floor of the staircase with its back to the staircase. The number of stairs climbed and the number of rears were recorded for a 3-min period. Climbing was defined as each stair on which the mouse placed all four paws; rearing was defined as each instance the mouse rose on hind legs (to sniff the air), either on a stair or leaning against the wall. The number of stairs descended was not taken into account. Before each test, the box was cleaned with a diluted alcohol solution to eliminate smells.
2.6. Data analysis All data were expressed as mean values ± standard error of mean (S.E.M.). Antibody levels, staircase climbing and rearing was compared between the appropriate groups by means of a univarient ANOVA assessing the effects of strain and immunization. Post-hoc analysis was performed by pair wise Fisher’s LSD post-hoc tests between the groups (SPSS).
3. Results 3.1. Autoantibody levels All APS induced strains of mice developed significantly high levels of anti-2 -GPI dependent-CL antibodies compared to their respective adjuvant controls at 5 months after immunization (Fig. 1, p < 0.001 for immunization effect by ANOVA). However, as can be seen in Fig. 1, there were clear differences between the strains in the levels of antibodies present 5 months after immunization as demonstrated by the interaction of strain × immunization (p < 0.001, ANOVA). Similarly, analysis of the level of autoantibodies in only the 2 -GPI immunized animals revealed a strain difference (p < 0.001, ANOVA). As can be seen in Fig. 1, the strain differences are due to the Balb/c and ICR mice developing higher levels of anti-2 -GPI dependentCL antibodies than the C57BL/6, AKR and C3H mice. This was confirmed by post hoc analysis of antibody levels in the five strains (p < 0.01 for all comparisons). There was a significant difference in the baseline antibody level between the different mouse strains controls (p < 0.001, ANOVA). Since C57BL/6 and Balb/c mice are known to differ significantly in features of their immune systems such as relative weighting of Th1 and Th2 responses [25] we compared in detail
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
Fig. 1. Autoantibody level in the sera of the five strains of mice. Significantly high levels of anti-2 -GPI dependent-CL antibodies in all APS induced strains of mice compared to their respective adjuvant controls at 5 months after immunization. Individual autoantibody levels of each mouse and mean ± S.E.M. levels of each group of mice are presented.
the development of antibody response in these mice (Fig. 2). Both strains developed high level of anti-2 -GPI dependent-CL antibodies compared to controls at 1 month after immunization (p < 0.001 for immunization effect by repeated measures ANOVA). Four months later, this level dropped significantly more in the C57BL/6 mice than in the Balb/c mice (p < 0.001 for month × immunization × strain effect by repeated measures ANOVA). There was also significantly more variance in the levels of antibodies in the C57BL/6 immunized mice compared to the Balb/c mice 1 month after immunization (p = 0.007 by F-test). At 5 months, there was no difference in the variance between the two groups. 3.2. Behavior in the staircase test All strains of mice in both APS and control groups behaved normally in their cages. They fed normally and did not display
Fig. 2. Mean ± S.E.M. of anti-2 -GPI dependent-CL antibodies in APS induced (black bars) mice compared to controls (gray bars) at 1 month (1 m) and 5 months (5 m) post immunization (PI) in the Balb/c and C57/B6 mice.
291
Fig. 3. Activity measure in the staircase test in the 5 mouse strains. Mean ± S.E.M. stair climbing events during 3 min of test in the APS induced mice (black bars) compared to their respective adjuvant controls (gray bars) in the different mouse strains.
overt neurological deficits such as limb weakness or circling. There was no mortality in any of the groups of mice, and there was no significant difference in weights between 2 -GPIimmunized APS and adjuvant-immunized control mice in all strains. In the activity measure (stair climbing, Fig. 3) there was a significant interaction of strain × APS (p < 0.001, ANOVA), indicating a significantly different response to immunization between the strains. The Balb/c-APS mice had significantly higher number of stairs climbed compared to Balb/c controls (p = 0.002, ANOVA), whereas the C57BL/6-APS mice had significantly less number of stairs climbed compared to C57BL/6 controls (p = 0.001, ANOVA). The ICR-APS mice had significantly higher number of stairs climbed compared to ICR controls (p = 0.05, ANOVA), and no significant difference was found between the AKR-APS and C3H-APS and their respective control mice (p > 0.215, ANOVA). There was a significant difference in the baseline behavior between the different mouse strains controls (p < 0.001, ANOVA). The C57BL/6 and the ICR controls were significantly more active compared to Balb/c, AKR, and C3H controls (p < 0.045 ANOVA for all comparisons). The effect of immunization did not correlate with the baseline behavior of controls in that there was an opposite effect of immunization on the stair climbing measure in the C57BL/6 compared to ICR mice, and a significant difference in the Balb/c compared to AKR and C3H mice. Similar trends were found in the rearing exploratory behavior measure (Fig. 4). There was a significant interaction of strain × APS (p = 0.001, ANOVA). The Balb/c-APS mice had significantly more rearing movements compared to Balb/c controls (p = 0.003, ANOVA), whereas the C57BL/6-APS and C3HAPS mice had significantly less number of rearings compared to their respective controls (p < 0.04, ANOVA). 2 -GPI immunization induced increased rearing behavior in both AKR (p = 0.037, ANOVA) and ICR (p = 0.3, ANOVA) mice.
292
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
Fig. 4. Exploratory behavior measure in the staircase test in the five mouse strains. Mean ± S.E.M. rearing events during 3 min of test in the APS induced mice (black bars) compared to their respective adjuvant controls (gray bars) of the different mouse strains.
3.3. Correlation between autoantibody level and behavior in the staircase test The effect of strain difference on antibody levels and behavioral measures was further analyzed. Due to difference in baseline antibody levels and behavior between the different mouse strains controls, data of APS mice were normalized in relation to control mice. Normalized antibody data were calculated by subtracting the mean control value for each strain from each individual APS mouse. Normalized behavioral data were obtained by presenting the measures of each APS mouse as percent of the mean of the controls in that group. Using these data, a regression analysis revealed a significant correlation between the normalized antibody levels and the behavioral measures of activity (stair-climbing, r = 0.318, p = 0.035) and exploratory behavior (rearing, r = 0.319, p = 0.034). To examine whether additional immunization induced factors significantly affect behavior, we performed an MANOVA with antibody levels as a cofactor. This analysis revealed a significant effect of strain on behavior (p < 0.001) and a significant interaction of strain × APS (p < 0.042). 4. Discussion The results presented support a significant behavioral effect of exposure to antiphospholipid antibodies in five strains of mice. This replicates our previous studies on Balb/c mice [16,23] and adds the interesting finding that the immunological and behavioral effects are strain-dependent. Correlation analysis indicated both a significant effect of antibodies but also other immune induced factors in the behavioral effect of APS. The most striking result from the staircase assay was that the behavioral effect in the C57BL/6 mice is opposite to what we have previously found in Balb/c mice [16,23] and to both Balb/c and ICR mice in the present study. The lack of effect in AKR and C3H mice may simply reflect the low levels of autoantibodies induced in these strains. Hyperactivity found in Balb/c mice
corroborates previous observations that when aPL are induced in Balb/c mice by a very different method (anti-ideotype immunization by monoclonal antibodies) [27] this is also associated with hyperactivity. The decreased activity of 2 GPI-immunized C57BL/6 mice has also been replicated in a series of experiments we have performed in transgenic animals in which congenic wild type mice of this strain served as controls (paper in preparation). The differences between the strains seem therefore to be well established and could be accounted for by either immunological or neurological differences or both. The data collected in the present study indicate that over a period of 4 months the antibody levels in the C57BL/6 mice did drop to lower levels than the Balb/c mice but it is difficult to envisage how this could cause an opposite effect, rather than lack of effect, indicating that there must be some interaction with the behavioral patterns in these strains. In this matter it is interesting to note that the AKR mice in the present study show a trend to hyperactivity and increased exploratory on the staircase test even though they attained lower levels of antibodies than the C57BL/6 group, whereas C3H mice show a trend to hypoactivity and decreased exploratory in the staircase test even though they attained similar levels of antibodies compared to the AKR group. The baseline (control group) levels of stair climbing and rearing by themselves do not seem to predict the effect of APS on behavior as can clearly seen in Figs. 3 and 4. It is also possible that in addition to the antibody response measured in the Balb/c and C57BL/6 strains, cellular immune responses may differ between these strains and explain the effect on the brain. There is little evidence, however, in both human and animal studies for a significant cellular immunological process affecting the brain in APS. It is likely therefore that similar processes affect the mice brains in all strains. Previous studies with MRL/lpr and NZB × NZW F1 mice have found decreased activity in genetic animal models of APS [15,22] and we propose that these differences may be due in part to speciesspecific behavior. There is much data on behavioral performance and difference of locomotor activity between different strains [4,17,24]. Low locomotor activity and exploration in Balb/c compared to C57BL/6 control mice found in the present study was similar to previous findings comparing different strains behavior in the staircase test [17]. The interaction of genetic factors with immune challenge may be relevant to findings in human SLE and APS patients. Few specific details are available regarding the genetic factors predisposing to APS and there are no data on what genetic factors predispose APS patients to develop behavioral and cognitive disturbances. Studies such as the present one may offer an experimental approach in which to study this important question. A potential problem in this approach, however, is that it is the Balb/c strain which develop the most pronounced effects in behavior and has also been shown to display cognitive deficits in a previous report [23]. This strain, however, is not widely used to generate transgenic animals in contrast to the C57BL/6 strain which has become the standard in this field. Another drawback of the Balb/c strain is its inability to swim, making this strain unsuitable for the many standardized and straightforward cognitive tests that are performed in water.
A. Katzav et al. / Behavioural Brain Research 169 (2006) 289–293
In order to improve our APS mouse model we are currently attempting to find transgenic components in which we would predispose C57BL/6 mice to develop both behavioral and cognitive changes. Such factors include Factor V Leiden mutation and the amyloid precursor protein (APP) model of Alzheimer’s disease. The specific pathological findings, which lead to the behavioral changes described in this study, are not known. We have preliminary histological data from these mice which do not show large infarcts, which is compatible with the normal neurological function of the mice. We hope that when definite histopathological changes are found to differentiate Balb/c APS mice from their controls these measures will be examined in the brains of other strains in order to verify their pathogenic importance. It is assumed that such findings in the mouse model will be immediately applicable to human disease since they will shed light on the importance of immune mediated processes affecting blood vessels and neuronal structures. Acknowledgments This study was performed in partial fulfillment of the requirements for a PhD degree of Aviva Katzav, Sackler Faculty of Medicine, Tel Aviv University, Israel. We thank Dr. Naam Kariv for invaluable advice and help. Supported by the Shreiber Fund for Medical Research. References [1] Aron AL, Cuellar ML, Brey RL, McKeown S, Espinoza LR, Shoenfeld Y, et al. Early onset of autoimmunity in MRL/++ mice following immunization with beta 2 glycoprotein I. Clin Exp Immunol 1995;101:78– 81. [2] Bakimer R, Fishman P, Blank M, Hohmman A, Sredni B, Djaldetti M, et al. Induction of primary antiphospholipid syndrome in mice by immunization with a human monoclonal anticardiolipin antibody (H-3). J Clin Invest 1992;89:1558–63. [3] Blank M, Faden D, Tincani A, Kopolovic J, Goldberg I, Gilburd B, et al. Immunization with anticardiolipin cofactor (beta-2-glycoprotein I) induces experimental antiphospholipid syndrome in naive mice. J Autoimmun 1994;7:441–55. [4] Bothe GW, Bolivar VJ, Vedder MJ, Geistfeld JG. Behavioral differences among fourteen inbred mouse strains commonly used as disease models. Comp Med 2005;55:326–34. [5] Brey RL, Chapman J, Levine SR, Ruiz-Irastorza G, Derksen RH, Khamashta M, et al. Stroke and the antiphospholipid syndrome: consensus meeting Taormina 2002. Lupus 2003;12:508–13. [6] Brey RL, Cote SA, Teale JM. Autoimmune disease is accelerated in MRL/++ mice after apolipoprotein-H immunization. Neurology 1994;44(Suppl 2):A316. [7] Brey RL, Escalante A. Neurological manifestations of antiphospholipid antibody syndrome. Lupus 1998;7:S67–74. [8] Caronti B, Pittoni V, Palladini G, Valesini G. Anti-beta 2-glycoprotein I antibodies bind to central nervous system. J Neurol Sci 1998;156: 211–9.
293
[9] Chapman J, Cohen-Armon M, Shoenfeld Y, Korczyn AD. Antiphospholipid antibodies permeabilize and depolarize brain synaptoneurosomes. Lupus 1999;8:127–33. [10] Chapman J, Rand JH, Brey RL, Levine SR, Blatt I, Khamashta MA, et al. Non-stroke neurological syndromes associated with antiphospholipid antibodies: evaluation of clinical and experimental studies. Lupus 2003;12:514–7. [11] Chapman J, Shoenfeld Y. Neurological and neuroendocrine-cytokine inter-relationship in the antiphospholipid syndrome. Ann N Y Acad Sci 2002;966:415–24. [12] Garcia CO, Kanbour-Shakir A, Tang H, Molina JF, Espinoza LR, Gharavi AE. Induction of experimental antiphospholipid antibody syndrome in PL/J mice following immunization with beta 2 GPI. Am J Reprod Immunol 1997;37:118–24. [13] Gharavi AE, Sammaritano LR, Wen J, Elkon KB. Induction of antiphospholipid autoantibodies by immunization with beta 2 glycoprotein I (apolipoprotein H). J Clin Invest 1992;90:1105–9. [14] Katzav A, Chapman J, Shoenfeld Y. CNS dysfunction in the antiphospholipid syndrome. Lupus 2003;12:903–7. [15] Katzav A, Kloog Y, Korczyn AD, Niv H, Karussis DM, Wang N, et al. Treatment of MRL/lpr mice, a genetic autoimmune model, with the Ras inhibitor, farnesylthiosalicylate (FTS). Clin Exp Immunol 2001;126:570–7. [16] Katzav A, Pick CG, Korczyn AD, Oest E, Blank M, Shoenfeld Y, et al. Hyperactivity in a mouse model of the antiphospholipid syndrome. Lupus 2001;10:496–9. [17] Lepicard EM, Joubert C, Hagneau I, Perez-Diaz F, Chapouthier G. Differences in anxiety-related behavior and response to diazepam in BALB/cByJ and C57BL/6J strains of mice. Pharmacol Biochem Behav 2000;67:739–48. [18] McNeil HP, Simpson RJ, Chesterman CN, Krilis SA. Anti-phospholipid antibodies are directed against a complex antigen that includes a lipidbinding inhibitor of coagulation: beta 2-glycoprotein I (apolipoprotein H). Proc Natl Acad Sci USA 1990;87:4120–4. [19] Polz E, Wurm H, Kostner GM. Investigations on beta 2-glycoprotein-I in the rat: isolation from serum and demonstration in lipoprotein density fractions. Int J Biochem 1980;11:265–70. [20] Roubey RA, Hoffman M. From antiphospholipid syndrome to antibodymediated thrombosis. Lancet 1997;350:1491–3. [21] Ruiz-Irastorza G, Khamashta MA, Hughes GR. Hughes syndrome crosses boundaries. Autoimmun Rev 2002;1:43–8. [22] Schrott LM, Crnic LS. Anxiety behavior, exploratory behavior, and activity in NZB × NZW F1 hybrid mice: role of genotype and autoimmune disease progression. Brain Behav Immun 1996;10:260–74. [23] Shrot S, Katzav A, Korczyn AD, Litvinju Y, Hershenson R, Pick CG, et al. Behavioral and cognitive deficits occur only after prolonged exposure of mice to antiphospholipid antibodies. Lupus 2002;11:736–43. [24] Weizman R, Paz L, Backer MM, Amiri Z, Modai I, Pick CG. Mouse strains differ in their sensitivity to alprazolam effect in the staircase test. Brain Res 1999;839:58–65. [25] Zhang F, Liang Z, Matsuki N, Van Kaer L, Joyce S, Wakeland EK, et al. A murine locus on chromosome 18 controls NKT cell homeostasis and Th cell differentiation. J Immunol 2003;171:4613–20. [26] Ziporen L, Polak-Charcon S, Korczyn DA, Goldberg I, Afek A, Kopolovic J, et al. Neurological dysfunction associated with antiphospholipid syndrome: histopathological brain findings of thrombotic changes in a mouse model. Clin Dev Immunol 2004;11:67–75. [27] Ziporen L, Shoenfeld Y, Levy Y, Korczyn AD. Neurological dysfunction and hyperactive behavior associated with antiphospholipid antibodies. A mouse model. J Clin Invest 1997;100:613–9.