Effect of Focal Cerebral Infarctions on Lesional RhoA and RhoB Expression FREE

MEMBERS OF the Rho family are small GTPases that act as molecular switches, transducing extracellular signals to the actin cytoskeleton by changing the inactive GDP-bound form to the active GTP-bound form,¹ and they are important regulators of the cellular actin cytoskeleton essential for the regulation of cell shape and polarity and for cell motility and adhesion.^{2 - 3} To date, the mammalian Rho GTPase family consists of at least 14 different members, of which Rho, Rac, and Cdc42 are the 3 most extensively studied in vitro.⁴ In the central nervous system (CNS), Rho GTPases are involved in neuronal morphogenesis, axonal and dendritic growth, axon guidance, dendritic remodeling and plasticity, and synapse formation.^{4 - 6} Under pathophysiologic conditions, however, the family member Rho contributes to the inability of the adult CNS to regenerate after injury by inducing neurite retraction, so far confined to the isoform RhoA. Pharmacologic inactivation of Rho using C3 transferase—an enzyme from Clostridium botulinum that selectively inhibits Rho by adenosine diphosphate–ribosylating its effector domain (Asn⁴¹)—promotes outgrowth of neurites on inhibitory substrates in cell culture experiments⁷ and facilitates regeneration after CNS injury in vivo.⁸ In addition, the gene coding for the isoform RhoB was found to be up-regulated in damaged neurons in a murine stroke model,⁹ pointing at a role for Rho proteins in cerebral ischemia. Along these lines, a reduction in infarct size after middle cerebral artery occlusion was observed in adult mice treated with C3 transferase.¹⁰

To provide evidence that the Rho signaling pathway is a potential target for intervention after focal cerebral infarction (FCI) in humans, we investigated expression patterns of RhoA and RhoB in brains of patients who died after FCI and compared them with those in brains of neuropathologically unaltered controls by immunohistochemical analysis.

Twenty-two patients with a clinical history and a neuropathologically confirmed diagnosis of FCI were included in the study. Brain specimens were derived from an updated stroke brain library reported previously¹¹ and were compared with brain specimens from 4 neuropathologically unaffected patients from a recently described brain library.¹²

RhoA immunohistochemical analysis was performed using a mouse monoclonal antibody (sc-418; Santa Cruz Biotechnology, Santa Cruz, Calif) (dilution 1:100) and a biotinylated rabbit anti–mouse IgG F(ab)₂ as secondary antibody (DAKO, Hamburg, Germany) (dilution 1:400). For immunolabeling of RhoB, a rabbit polyclonal antibody (sc-180; Santa Cruz Biotechnology) (dilution 1:100) was used, followed by incubation with a biotinylated porcine anti–rabbit IgG F(ab)₂ antibody fragment (DAKO) (dilution 1:400). Immunohistologic procedures were performed as previously described elsewhere.¹¹ Specificity of the RhoB antibody was confirmed by complete inhibition of immunostaining after preincubation with 10-fold excess of protein amount (micrograms) blocking peptide (sc-180 P; Santa Cruz Biotechnology) used initially for immunization for 3 hours on ice, whereas staining patterns were unaffected by incubation with an irrelevant control peptide (cyclooxygenase-2). Likewise, RhoA staining patterns were not altered by preincubation with RhoB blocking peptide. Negative controls consisted of sections incubated in the absence of the primary antibody. Double-labeling experiments were performed as previously described elsewhere¹¹ using monoclonal antibodies against glial fibrillary acid protein (Chemicon International Ltd, Hofheim, Germany) (dilution 1:500) and the CD68 epitope PG-M1 (DAKO) (dilution 1:500).

In all areas examined, positive cells in the parenchyma were microscopically counted per 10 high-power fields (×400, with an eyepiece grid representing 0.0625 mm²). RhoA and RhoB staining patterns were evaluated separately at the site of the lesion and in neighboring areas as previously described elsewhere (Figure 1).¹¹ In addition, to evaluate the presence of RhoA⁺ and RhoB⁺ cells in the perivascular spaces, the number of positive cells attached to the outer vessel wall and in the Virchow-Robin spaces were counted per 10 vessels within the lesion. Data were clustered into 4 groups according to infarction age (<1 day, n = 5; 2-10 days, n = 6; 2-12 weeks, n = 6; and 4-38 months, n = 5), were subsequently calculated as medians of labeled cells (MLCs) (50th percentile), and were analyzed using the Kruskal-Wallis test followed by the Dunn test. Significance was set at P<.05.

Brain tissue samples from 22 patients with FCI and from 4 neuropathologically unaltered controls were investigated for localization of RhoA and RhoB expression.

In the brain parenchyma of neuropathologically unaltered controls, no expression of RhoA was detected (MLC = 0) (Figure 2A), except in 1 control in whom positive glial cells mostly resembling oligodendrocytes were found. No expression of RhoA was observed in neuronal cells. A positive staining result for RhoB was detected in the cytoplasm of individual glial cells mainly located in the white matter of control brains (MLC = 0) (Figure 2B). One control showed additional staining of several neurons in the cortical layer. No extracellular staining pattern was observed with RhoA and RhoB antibodies in any experiments performed.

A significant increase in the number of RhoA⁺ cells and RhoB⁺ cells was observed in patients with FCI at the lesion site (MLC = 10 and 5.5, respectively; P<.001 for both) (Figure 2A and B). Evaluation of the time course of RhoA and RhoB expression revealed a comparable pattern for both molecules. During the first day after infarction, a slight but nonsignificant increase in RhoA⁺ cells (MLC = 1; P>.05) (Figure 3A) and RhoB⁺ cells (MLC = 2; P>.05) (Figure 3B) was observed. Thereafter, significant up-regulation of RhoA and RhoB was detected that was most prominent 2 to 12 weeks after FCI (MLC = 23.5 and 13.5, respectively; P<.001 for both) (Figure 3). RhoA and RhoB remained up-regulated 4 to 38 months after the ischemic event (MLC = 16 and 9, respectively; P<.001 for both) (Figure 3). Expression of RhoA was restricted to the lesion site, as no significant difference was detected in neighboring areas vs controls (MLC = 1; P>.05) (Figure 2A). In contrast, the number of RhoB⁺ cells in neighboring areas differed significantly from that in controls (MLC = 1; P<.05) (Figure 2B), although they were significantly lower than the number of positive cells within the lesion (P<.001). Thus, up-regulation of RhoB was observed predominantly but not exclusively at the site of the lesion.

The first cells to show relatively weak immunoreactivity for RhoA were polymorphonuclear leukocytes, clearly distinguishable by morphologic features (Figure 4B). These cells appeared after 2 days and contributed to the pool of RhoA⁺ cells until 4 days after FCI. The main cellular sources of RhoA, however, were monocytes and macrophages (Figure 4C), in which expression of RhoA was detected from day 6 until the late stages, that is, 4 to 38 months after FCI. This finding was confirmed by double-labeling experiments in which a co-localization of RhoA and the CD68 epitope PG-M1 was observed (Figure 4C, inset). Furthermore, few reactive astrocytes were found to weakly express RhoA predominantly at later stages (weeks and months after FCI) (Figure 4D). A distinct immunoreactivity of neurons for RhoA was not observed in this series. A slight increase in the number of RhoA⁺ perivascular cells was noticed only after 2 to 10 days (MLC = 1.5, P<.05) (data not shown).

The main cellular sources of RhoB after FCI were macrophages and reactive astrocytes. Numerous RhoB⁺ macrophages were observed after 2 to 12 weeks (Figure 4H). However, immunoreactivity of these cells for RhoB was clearly weaker than that for RhoA (compare Figure 4C with Figure 4H). In contrast, reactive astrocytes strongly expressed RhoB weeks and months after FCI (Figure 4I). In the very early stages (<1 day), expression of RhoB was detected in neurons in 3 of 5 cases (Figure 4F). Furthermore, RhoB was expressed by neurons in the neighboring areas of some of the later cases weeks and months after FCI (Figure 4J and K). Only a few polymorphonuclear leukocytes were found to express RhoB several days after FCI (Figure 4G). No significant up-regulation of RhoB⁺ cells was observed in the perivascular spaces (data not shown).

Rho activity and its signaling pathway have been shown to contribute to the inability of the adult CNS to regenerate after injury by inducing neurite retraction, so far confined to the isoform RhoA. Several studies have revealed that inactivation of Rho or of its downstream target Rho-associated kinase promotes growth of neurites and injured axons on inhibitory substrates and facilitates recovery after spinal cord injury in vivo,^{7 - 8} a process that induces traumatic and ischemic neuronal damage. In addition, the gene coding for the isoform RhoB is up-regulated in ischemia-damaged neurons in a murine stroke model, and apoptosis in hippocampal neurons is associated with an early increase in RhoB protein, suggesting that RhoB plays a role in neuronal signal transduction after brain injury.⁹ Furthermore, it has been demonstrated that thrombin, a serine protease that is released during CNS injury, induces apoptosis in cultured neurons and astrocytes through activation of RhoA, whereas inhibition of RhoA abolished thrombin-induced cell death.¹³ Another important mediator of CNS injury, tumor necrosis factor, has been shown to affect the formation of neurites in neurons cocultured on astrocytic cells by activation of RhoA.¹⁴ Taken together, these data provide strong evidence that RhoA and RhoB contribute to a major extent to the inhibition of neuronal survival and regeneration after CNS injury.

In the present study, macrophages are one main cellular source of RhoA and RhoB after FCI. Rho proteins are known to play an important role in the recruitment and activation of macrophages. Another cell type observed to express Rho after traumatic brain injury was the reactive astrocytes—the major cellular component of the CNS glial scar. It has been demonstrated (1) that the Rho signaling pathway mediates the neurite growth–inhibitory activity of the glial scar and (2) that this inhibition can be antagonized by C3 transferase–operated Rho inactivation.¹⁵ A weak expression of RhoA⁺ and RhoB⁺ was also detected in polymorphonuclear leukocytes several days after FCI. The Rho signaling pathway has been described to be involved in the detachment, migration, and aggregation of granulocytes¹⁶ and in the production of superoxide by neutrophils,¹⁶ which is known to be involved in the mediation of secondary damage after FCI.

We also observed neuronal expression of RhoB within the lesion during the very early stages as in the neighboring areas during some of the later stages. After FCI, a variety of inhibitory proteins are produced by oligodendrocytes (ie, components of the myelin sheath such as Nogo-A, myelin-associated glycoprotein, and oligodendrocyte myelin glycoprotein) and by astrocytes (ie, tenascin-R, chondroitin sulfate proteoglycans, Ephrin B3), contributing to the hostile environment and to the inability of neurons to regenerate. It was proposed that the effects of most of these inhibitory proteins are mediated by the Rho signaling pathway in neurons.^{17 - 18} Thus, it can be expected that Rho activation, being downstream of several inhibitory proteins, is a promising target for therapeutic intervention.

Rho is accessible to pharmacologic inactivation, that is, by the C3 exoenzyme from C botulinum, which specifically adenosine diphosphate–ribosylates Rho at its effector domain (Asn⁴¹), thereby inactivating it.^{19 - 20} The lesional up-regulation of RhoA and RhoB in FCI described herein displays the substrate for a pharmacologic intervention after FCI. RhoA and RhoB were detected by immunohistochemical analysis irrespective of their activation status. This is of interest because total RhoA and RhoB—active and inactive forms—are accessible to therapeutic intervention with C3 transferase by either inactivating the active GTP-bound forms or preventing inactive GDP-bound forms from becoming active. Thus, total RhoA and RhoB expression patterns represent a substrate for pharmacologic intervention targeting the Rho signaling pathway.

In the present study, we observed prolonged up-regulation of RhoA and RhoB for months after FCI. Thus, these data suggest that (1) targeting the Rho signaling pathway after FCI is conclusive and (2) owing to prolonged RhoA and RhoB expression, the therapeutic window for treatment with Rho inhibitors may be wider than expected.

Corresponding author and reprints: Jan M. Schwab, MD, PhD, Institute of Brain Research, University of Tuebingen, Calwer Str 3, D-72076 Tuebingen, Germany (e-mail: jmschwab@med.uni-tuebingen.de).

Accepted for publication March 25, 2003.

Author contributions: Study concept and design (Drs Brabeck, Meyermann, Schluesener, and Schwab); acquisition of data (Drs Brabeck, Mittelbronn, and Schwab and Mr Bekure); analysis and interpretation of data (Drs Brabeck and Schwab); drafting of the manuscript (Drs Brabeck and Schwab); critical revision of the manuscript for important intellectual content (Drs Mittelbronn, Meyermann, Schluesener, and Schwab and Mr Bekure); statistical expertise (Drs Brabeck and Schwab); obtained funding (Dr Schwab); administrative, technical, and material support (Drs Mittelbronn, Meyermann, and Schluesener and Mr Bekure); study supervision (Drs Meyermann, Schluesener, and Schwab). Drs Brabeck and Schwab contributed equally to this article.

This study is supported by grant G-2024-1065.1/2000 from the Young Scientists Program of the German Israeli Foundation for Scientific Research and Development (Dr Schwab) and the Herti-foundation (Dr Schluesener).