Keywords

Introduction

The ubiquitous heat-shock protein 90 (Hsp90)-based chaperone system is an essential component of several signal transduction pathways in eukaryotic cells. Hsp90 is markedly expressed throughout neuronal subpopulations of adult rat brain, but not in non-neuronal cells [7]. The Hsp90 inhibitor geldanamycin (GA), upon binding Hsp90, releases heat-shock factor (HSF1) and induces heat-shock proteins (HSPs) [2, 5]. An injection of GA 24 h before focal ischemia in rat was neuroprotective, the effect being attributed to the induction of Hsp70 [12]. However, an earlier study found that GA protected a mouse hippocampal cell line against glutamate toxicity even when given 4 h after the insult [23]. In this paradigm, GA increased the degradation of Raf-1, the central component of a mitogen-activated protein kinase (MAPK) cascade, as well as inducing Hsp90. Thus, it is likely that various cell-signaling pathways may be altered by pharmaceutical inhibition of Hsp90 chaperone function during cerebral ischemia and reperfusion.

In this study, we tested the efficacy of a GA dose given just before ischemia or reperfusion in a rat model of focal cerebral ischemia. We also investigated the effects of these GA treatments on p38 and p44/42 MAPK signaling pathways.

Materials and Methods

Animal Model and Treatment Groups

The protocols for the animal studies were approved by the University of Michigan Committee on the Use and Care of Animals. A total of 57 male Sprague–Dawley rats (275–325 g; Charles River Laboratories, Portage, MI, USA) were used. The animals were fasted for 4 h before surgery, but had free access to water. Anesthesia was induced by inhalation of 5 % isoflurane in a 70 % nitrous oxide/30 % oxygen mixture and maintained by 1.5 % isoflurane administered through a face mask. A blood sample was collected from the tail artery to measure PaO2, PaCO2, pH, hematocrit, and glucose. Rectal temperature was maintained at 37.5 °C using a feedback-controlled heating pad. Transient focal cerebral ischemia was induced using the intraluminal suture method [13]. Ninety minutes after ischemia induction the animal was re-anesthetized and the intraluminal suture was removed. The sham operation was identical except that the intraluminal suture was not advanced beyond 15 mm.

Two GA (Sigma, St. Louis, MO, USA) treatment regimens were used: the GA ischemia regimen consisted of an intraperitoneal injection of 300 mg/kg GA 15 min before induction of ischemia, and the GA reperfusion regimen consisted of the same dose injected 15 min before reperfusion. Vehicle-treated animals received an equal volume intraperitoneal injection of 20 % dimethylsulfoxide 15 min before the induction of ischemia. Animals were sacrificed 2 or 4 h after ischemia onset (0.5 or 2.5 h of reperfusion) for Western blot analysis, or 24 h after ischemia onset for determination of tissue damage.

Morphometric Measurement of Tissue Damage

To identify the tissue damage fresh brain tissue was cut into 2-mm-thick coronal sections and incubated in 2 % 2,3,5-triphenyltetrazolium chloride (TTC). The infarct area was measured with NIH Image 1.62 software, and the volume was calculated by multiplying the distance between sections.

Western Blot Analysis

Animals were perfused with saline. Ischemic tissue in the left hemisphere was sampled using a 6-mm cork borer and divided into cortex and caudate. Western blot analysis was performed as previously described [22]. Primary antibodies were anti-Raf-1 (1:500; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and anti-phospho-p44/42 MAPK (1:1,000; Cell Signaling Technology, Boston, MA, USA).

Immunofluorescence Staining

Another set of animals that were exposed to 90-min transient MCA occlusion without any treatment were sacrificed at 2 and 4 h after ischemia onset and were perfused with saline ­followed by 4 % paraformaldehyde. The brains were removed and immersed in 4 % paraformaldehyde and 25 % sucrose for 24 h at 4 °C. The tissue was frozen, embedded in OCT compound (Sakura Finetek, Torrance, CA, USA), and sectioned at 40-μm thickness on a cryostat. Floating sections were microwaved in distilled water and single- or double-stained using anti-Hsp90 (1:400; StressGen Biotechnologies, San Diego, CA, USA), anti-phospho-p44/42 MAPK (1:200), and anti-Raf-1 (1:100) primary antibodies. Secondary antibodies were fluorescein-conjugated anti-mouse (1:100; Vector Laboratories, Burlingame, CA, USA) and rhodamine-conjugated anti-rabbit (1:100; Chemicon, Billerica, MA, USA) [22]. All solutions were buffered using Tris base. Stained sections were mounted using Vectashield (Vector Laboratories) and visualized under an Olympus FV-500 confocal microscope.

Statistical Analysis

All quantitative data were presented as mean ± standard deviation and compared using analysis of variance (ANOVA) with Student’s t test. Differences were considered significant at the P  <  0.05 level.

Results

Physiological data were measured immediately after induction of ischemia. Blood pH, PaO2, PaCO2, glucose, and hematocrit were within the normal range.

Both treatment regimens, GA ischemia and GA reperfusion, reduced infarct volume significantly compared with the vehicle group (n  =  24; 56  ±  37 mm3 and 105  ±  67 mm3 vs 267  ±  97 mm3, P  <  0.005). The protective effect of GA was more pronounced in the cortex (Fig. 1).

Fig. 1
figure 00061

A dose of geldanamycin (GA) given 15 min before the onset of ischemia (GA ischemia) or reperfusion (GA reperfusion) reduced infarct volume (detected 24 h after ischemia) compared with the vehicle-injected group by 79 and 61 % respectively. n  =  8, *P  <  0.05 vs vehicle

Full size image

The phosphorylated active forms of p38 (p-p38) and p44/42 (p-p44/42) MAPKs detected by Western blotting were increased both in the ischemic caudate and cortex at 2 and 4 h after ischemia onset. The average increase in p-p38 MAPK was 1.6-fold, whereas p-p44/42 MAPK increased an average of 24-fold. Raf-1 was unchanged at 2 h, but decreased at 4 h after ischemia onset compared with the pre-ischemic levels (data not shown). GA ischemia and GA reperfusion significantly attenuated p-p44/42 MAPK in ischemic caudate (n  =  12: 386  ±  60 and 1,091  ±  187 vs 2,333  ±  314 %, P  <  0.005), and even more so in the ischemic cortex (103  ±  34 % and 163  ±  31 % vs 2,468  ±  306 %, P  <  0.0005) at 4 h after ischemia compared with the vehicle. Both treatment regimens of GA also decreased Raf-1 levels in ischemic regions at 2 h after ischemia compared with vehicle. Although GA-treated animals showed slightly reduced p38 MAPK activation, the effect was not significant (Fig. 2).

Fig. 2
figure 00062

The GA treatment regimens did not significantly alter p38 MAPK activation (a), but effectively blocked p44/42 MAPK activation (b) in the ischemic cortex 4 h after ischemia onset. Premature depletion of Raf-1 by GA treatment in the same region 2 h after ischemia onset (c) may explain the downstream blockage of p44/42 MAPK phosphorylation. n  =  4, *P  <  0.05 vs vehicle

Full size image

Hsp90 was predominantly located in the nucleus of neurons in normal rat brain, with little cytoplasmic staining detected by immunofluorescence. The cytoplasm of these neurons also stained positive for Raf-1 in double-labeled sections. Normal rat brain parenchyma showed no p-p44/42 MAPK staining. Two hours after ischemia onset (0.5 h after reperfusion), Hsp90 was increased in the cytoplasm of ­neurons in the ischemic core and co-localized with Raf-1 and p-p44/42 MAPK. Four hours after ischemia onset (2.5 h after reperfusion), Hsp90 co-localized with Raf-1 and p-p44/42 MAPK in the cytoplasm of neurons in the penumbral regions. In contrast, there was decreased Hsp90 and Raf-1 staining in the ischemic core at this time point. However, neurons in the ischemic core were still positive for p-p44/42 MAPK.

Conclusion

The present study shows that inhibition of Hsp90 with GA during ischemia or early reperfusion protects adult rat brain against focal ischemic damage. This treatment is also associated with a rapid decrease in Raf-1 protein levels and attenuation of p44/42 MAPK activation, whereas the p38 MAPK activation is not altered significantly.

The MAPKs comprise a group of signaling proteins that play a prominent role in regulating cell proliferation, differentiation, and adaptation. Both p38 and p44/42 MAPKs have been implicated in neuronal injury and disease (reviewed in Chu et al. [4], Harper and Wilkie [8], and Irving and Bamford [10]). Following ischemia many factors are released, including glutamate, free radicals, growth factors, cytokines, and thrombin, all of which are known to stimulate MAPK pathways. Ischemia rapidly induces p38 and p44/42 MAPK activation in rodent neurons and inhibition of these kinase systems reduces ischemic brain damage [1, 3, 6, 9, 11, 14, 19–21]. Our Western blot and immunofluorescence staining data also showed increased levels of activated p38 and p44/42 MAPKs in affected neurons at 0.5 and 2.5 h following 90 min of MCA occlusion. GA treatment during ischemia or reperfusion, while reducing brain injury, blocked p44/42 MAPK activation, but did not significantly affect p38 MAPK activation.

The Raf-MEK-p44/42 MAPK signaling module is the best characterized of the three main MAPK cascades and is emerging as an important regulator of neuronal responses to both functional (learning and memory) and pathological (regulated cell death) stimuli. Raf-1 is targeted to the cell membrane by Ras and integrates extracellular signals by phosphorylating the dual specificity kinase MEK (MAPK kinase), which in turn phosphorylates p44/42 MAPK. Activation of p44/42 MAPK leads to activation of other downstream kinases, as well as several transcription factors. Raf-1 is primarily cytosolic in location and exists in a native heterocomplex with Hsp90 [18]. Treatment by GA disrupts Raf-1 complex formation with Hsp90, leading to aberrant intracellular trafficking and increased degradation of Raf-1 [15, 16]. It is also shown that Raf-1 is the only component of the p44/42 MAPK signaling pathway that is depleted by GA and Raf-1 depletion by GA is sufficient to interdict signaling through this pathway [17]. In our experiments, GA treatment during ischemia/reperfusion decreased Raf-1 protein levels at 2 h after ischemia onset, which, in turn, may account for the reduced p44/42 MAPK activation observed at 4 h. Our confocal microscopic data demonstrated temporal co-localization of Hsp90, Raf-1, and p-p44/42 in the cytoplasm of neurons located in the ischemic core and penumbra.

The small lipophilic drug GA readily crosses the blood–brain barrier and is a promising option for the treatment of clinical stroke. GA treatment induces HSPs and blocks the activation of p44/42 MAPK in experimental stroke, which may in part account for the resulting neuroprotection. However, alterations in HSPs and MAPKs may also be the result of alterations in ischemic damage, complicating definitive conclusions. More detailed investigations of the involvement of HSP90 in ischemic neurodegeneration and the specific consequences of MAPK activation in this paradigm are necessary to fully reveal the mechanism of GA-induced neuroprotection.