SEARCH

SEARCH BY CITATION

Keywords:

  • apoptosis;
  • cerebral ischemia;
  • necrosis;
  • neuroprotective DAMPs;
  • non-classical release;
  • S100A13

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Prothymosin alpha (ProTα), a nuclear protein devoid of signal sequence, has been shown to possess a number of cellular functions including cell survival. Most recently, we demonstrated that ProTα is localized in the nuclei of neurons, while it is found in both nuclei and cytoplasm in the astrocytes and microglia of adult brain. However, the cell type-specific non-classical release of ProTα under cerebral ischemia is yet unknown. In this study, we report that ProTα is non-classically released along with S100A13 from neurons in the hippocampus, striatum and somatosensory cortex at 3 h after cerebral ischemia, but amlexanox (an anti-allergic compound) reversibly blocks this neuronal ProTα release. We found that none of ProTα is released from astrocytes and microglia under ischemic stress. Indeed, ProTα intensity is increased gradually in astrocytes and microglia through 24 h after the cerebral ischemia. Interestingly, Z-Val-Ala-Asp fluoromethyl ketone, a caspase 3 inhibitor, pre-treatment induces ProTα release from astrocytes in the ischemic brain, but this release is reversibly blocked by amlexanox. However, Z-Val-Ala-Asp fluoromethyl ketone as well as amlexanox has no effect on ProTα distribution in microglia upon cerebral ischemia. Taken together, these results suggest that only neurons have machineries to release ProTα upon cerebral ischemic stress in vivo.

Abbreviations used
Amx

amlexanox

CA

cornu ammonis

DAMPs

damage-associated molecular patterns

HMGB-1

high mobility group box-1

Hip

hippocampus

i.c.v.

intracerebroventricular

GFAP glial fibrillary acidic protein

intracerebroventricular

MAP-2

Microtubule-associated protein

PBS

phosphate-buffered saline

ProTα

prothymosin alpha

Str rad

stratum radiatum

tMCAO

transient middle cerebral artery occlusion

Z-VAD-fmk

Z-Val-Ala-Asp fluoromethyl ketone

The ischemia in the central nervous system is a complex pathophysiological condition, in which neuronal necrosis in the ischemic core causes progressive secretion of cytotoxic mediators, which in turns further cause extended neuronal death (Danton and Dietrich 2003; Swanson et al. 2004; Ueda 2009; Niizuma et al. 2010; Zhao and Rampe 2010). A wide variety of intracellular molecules termed as damage-associated molecular patterns (DAMPs) are secreted into the extracellular environment upon necrotic/ischemic stress and play key roles in such deterioration of cellular damages (Rubartelli and Lotze 2007; Kono and Rock 2008; Chen and Nunez 2010; Schmidt and Tuder 2010; Zitvogel et al. 2010; Pisetsky 2011). Among these molecules, high mobility group box-1 (HMGB-1) is a representative DAMPs protein, which is extracellularly released from the nuclei of neurons upon ischemic damages (Lotze and Tracey 2005; Liu et al. 2007; Muhammad et al. 2008; Qiu et al. 2008; Sims et al. 2010; Yang et al. 2010; Zhang et al. 2011). However, there is also a case that neuroprotective molecule, such as prothymosin alpha (ProTα), is released into the extracellular milieu upon ischemic/necrotic stress in culture experiments (Ueda and Fujita 2004; Fujita and Ueda 2007; Ueda et al. 2007, 2010; Fujita et al. 2009; Ueda 2009). In this sense, ProTα may be called as a new member of cytoprotective DAMPs molecules.

ProTα is a nuclear protein and functionally implicated with cellular proliferation and survival (Pineiro et al. 2000; Jiang et al. 2003; Gomez-Marquez 2007; Ueda 2009), chromatin remodeling (Gomez-Marquez and Rodriguez 1998), DNA packaging (Diaz-Jullien et al. 1996; George and Brown 2010), and regulation of transcription (Martini et al. 2000; Karetsou et al. 2002). In addition to this, extracellular roles of ProTα have also been reported (Baxevanis et al. 1992; Garbin et al. 1997; Mosoian et al. 2006). There is an exciting report about the involvement of Toll-like receptor-4 in ProTα-induced immunoprotection against virus (Mosoian et al. 2010). The recent in vitro investigations described that ProTα is localized in nuclei of both cultured cortical neurons and embryonic astrocytes, and that is extracellularly released from these cells upon ischemic stress (Matsunaga and Ueda 2010). The mode of ischemia-induced non-classical release of ProTα was characterized in the experiments using C6 astroglioma cells in vitro (Matsunaga and Ueda 2010). This study explained that ProTα is first diffused from the nucleus to cytosol, and in turn immediately co-released to the extracellular space with S100A13 (a Ca2+-binding cargo protein) and this release is reversibly blocked by amlexanox, an anti-allergic drug.

Most recently, we demonstrated that ProTα is strictly localized in the nucleus of adult brain neurons, whereas it is expressed both in the cell body and cytosolic space of processes in the astrocytes and microglia, an indication of big difference between in vitro and in vivo studies in terms of ProTα localization in astrocytes (Matsunaga and Ueda 2010; Halder and Ueda 2012). Interestingly, nuclear ProTα intensity was drastically increased in astrocytes by diminishing cytosolic contents, but not in microglia after the pre-treatment with Z-Val-Ala-Asp fluoromethyl ketone (Z-VAD-fmk), a caspase 3 inhibitor (Halder and Ueda 2012). The existence of caspase 3 activity in astrocytes in the adult brain as well as the caspase 3-mediated ProTα fragmentation in vitro has been reported previously (Enkemann et al. 2000; Evatafieva et al. 2003; Duran-Vilaregut et al. 2010; Matsunaga and Ueda 2010). Taken together, these studies suggested that caspase 3 controls the distribution of astroglial ProTα in the brain. However, the ischemia-induced ProTα release from brain is still under investigation. In the present study, we firstly attempted to see the cell type-specific non-classical release of ProTα as well as the effect of amlexanox on ProTα distribution in brain after cerebral ischemia.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Middle cerebral artery occlusion mouse model

The transient middle cerebral artery occlusion (tMCAO) model was induced following the method as described previously (Egashira et al. 2004). Briefly, mice were anesthetized by 2% isoflurane (Mylan, Tokyo, Japan), and body temperature was monitored and maintained at 37°C during surgery. After a midline neck incision, the middle cerebral artery was occluded transiently using 8-0 in size monofilament nylon surgical suture (Natsume Co. Ltd., Tokyo, Japan) coated with silicon (Xantopren; Bayer dental, Osaka, Japan) that was inserted through the left common carotid artery and advanced into the left internal carotid artery. Following 1 h tMCAO, the animals were briefly re-anesthetized with isoflurane and the monofilament was withdrawn for reperfusion studies. As the silicon-coated nylon suture also plugs the branch from middle cerebral artery to supply blood to hippocampus in mice, due to small brain size, the ischemia-induced brain damages are also observed in the hippocampus. Cerebral blood flow was monitored by laser Doppler flowmeter (ALF21; Advance Co., Tokyo, Japan) using a probe (diameter 0.5 mm) of a laser Doppler flowmeter (ALF2100; Advance Co.) inserted into the left striatum (anterior: 20.5 mm; lateral: 1.8 mm from bregma; depth: 4.2 mm from the skull surface) through a guide cannula.

Drug treatment

Amlexanox (kindly provided by Takeda Pharmaceutical Company Ltd, Osaka, Japan) was dissolved in 0.05 N NaOH in phosphate buffered saline [K+ free phosphate-buffered saline (PBS), pH 7.4], adjusted pH 7.6 by 0.1 M H3PO4, and finally diluted in PBS. Using Hamilton syringe, amlexanox was injected intracerebroventricularly (i.c.v.) at a dose of 10 μg/5 μL in the brain 30 min before ischemia. Vehicle was treated with equal volume of solution containing 0.05 N NaOH and 0.1 M H3PO4 in PBS 30 min before ischemia in a similar manner. However, Z-VAD-fmk was purchased from Sigma-Aldrich, St Louis, MO, USA and dissolved in dimethyl sulfoxide and finally diluted in artificial CSF. Z-VAD-fmk was delivered i.c.v. at a dose of 1 μg/5 μL in the brain 30 min before ischemia (1 h tMCAO). Following similar way, vehicle was treated with equal volume of artificial CSF in the brain 30 min before cerebral ischemia.

Cell counting

Measurements of ProTα- and S100A13-positive cells in the brain were done using the BZ Image Measurement software. Briefly, cell counts were carried out in bright field images following the protocol as reported previously (Matsumoto et al. 2006). The number of ProTα- and S100A13-positive neurons, astrocytes and microglia in the somatosensory cortex of brain were stereologically counted (bregma 0.62 to −2.06) in the square fields (approximately 250 μm × 250 μm) of vehicle-treated (= 3) and amlexanox-treated (= 3) ischemic brain and were normalized to those obtained identically in the control brain (= 4). However, to determine the ProTα- and S100A13-positive cells, we carried out the counts using specific cell markers with clearly visible nuclei. The quantification was expressed as average percentage of the total number of cell type-specific ProTα- and S100A13-positive cells in the 4–7 brain sections per mouse.

Statistical analysis

All results are shown as means ± SEM. Two independent groups were compared using the Student's t-test. Multiple groups were compared using Dunnett's multiple comparison test after a one-factor ANOVA. < 0.01 was considered significant.

Other methods

Animals, tissue preparations, antigen retrieval microwave technique and proteinase K treatment, and immunohistochemical analysis are available as Appendix S1.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Cerebral ischemia-induced rapid depletion of ProTα from hippocampal neurons

Following cerebral ischemia (1 h tMCAO) in mice, we examined time-dependent changes in ProTα expression in the ipsilateral hippocampus throughout 1–24 h. The 3,3′-diaminobenzidine tetrahydrochloride immunostaining data revealed that ProTα depletion in ipsilateral CA1 pyramidal neurons starts as early as 1 h after the cerebral ischemia (Fig. 1b), completes at 3 h (Fig. 1c) and followed by a gradual recovery of ProTα signals through 24 h (Fig. 1–d–f). However, there were some cells showing intense ProTα-immunoreactivity in the ipsilateral stratum radiatum of hippocampus at 3 h (Fig. 1c). The ProTα signals in various cells in the stratum radiatum gradually increased as time goes thereafter (Fig. 1d–f). However, there was no significant change in ProTα reactivity in the respective contralateral hippocampus of ischemic brain (Fig. 1–a–f). Similar results of ProTα depletion and recovery were also observed in the regions of striatum (Fig. 2–a–f) and somatosensory cortex (Fig. 2–g–l).

image

Figure 1. Depletion of ProTα in the CA1 pyramidal cell layer of hippocampus under cerebral ischemia. (a–f) Immunostaining of ProTα in adult mice brain after ischemic stress. 3,3′-Diaminobenzidine tetrahydrochloride (DAB) immunostaining data of coronal brain sections indicate that ProTα is partially released at 1 h (b) followed by complete release at 3 h (c) in the ipsilateral CA1 pyramidal cell layer of hippocampus after ischemic stress (1 h tMCAO). Some cells show intense ProTα reactivity at 3 h in the stratum radiatum of hippocampus noted by arrow points (c). (d–f) ProTα level is recovered gradually in the ipsilateral CA1 pyramidal cell layer at the later time points that starts from 6 h (d) continuing 12 h (e) and 24 h (f) after ischemic stress. ProTα intensity is also gradually increased in the ipsilateral stratum radiatum of hippocampus through 24 h after ischemic stress noted by arrows (d–f). There is no change in ProTα staining at 0 h (a) as well as in the respective contralateral sides of hippocampus after ischemia. Insets in panels (a–f) indicate the higher magnification view of ProTα intensity noted by red squares.

Download figure to PowerPoint

image

Figure 2. ProTα is released from striatum and somatosensory cortex of ischemic brain. 3,3′-Diaminobenzidine tetrahydrochloride (DAB) immunostaining of coronal brain sections is performed using antibody against ProTα. (a–f) ProTα signal is partially lost at 1 h (b), followed by complete lost at 3 h (c) and recovery gradually at 6 h (d), 12 h (e) and 24 h (f) in the ipsilateral striatum after ischemic stress (1 h tMCAO) in adult mice, retaining the normal staining at 0 h (a) and also in the respective contralateral sides. (g–l) DAB immunostaining data show the time–course ProTα expression in the ipsilateral as well as contralateral somatosensory cortex of brain from 0 to 24 h after ischemic stress. (c, i) Some non-neuronal-like cells show intense ProTα reactivity at 3 h in the ipsilateral striatum (c) and somatosensory cortex (i) noted by arrows.

Download figure to PowerPoint

Ischemia-induced depletion of neuronal ProTα

To identify the cell type specificity for ProTα release after cerebral ischemia (1 h tMCAO) and reperfusion, coronal brain sections were co-stained with anti-ProTα IgG and antibody against Microtubule-associated protein (MAP-2), a cytoplasmic neuronal marker. Our double fluorescence immunohistochemical data explained that ProTα immunoreactivity is strictly localized in nuclei of MAP-2-positive CA1 pyramidal neurons of hippocampus in the control mice (Fig. 3a). In confocal microscopy observation, ProTα signals were also found in nuclei in the MAP-2-positive CA1 pyramidal neurons of control hippocampus (Fig. 3g). As early as 3 h after the cerebral ischemia and reperfusion, ProTα signals in CA1 pyramidal neurons were completely lost, whereas the signals were significantly enhanced in some non-neuronal cells in the stratum radiatum of hippocampus (Fig. 3b). ProTα in the nuclei of pyramidal neurons was recovered largely to the control levels, and the signals were also localized in the nuclei at 24 h (Fig. 3c). Like control, there was also no change in nuclear ProTα levels observed in the contralateral side of brain (data are not shown). Similar patterns of ProTα release at 3 h and recovery at 24 h were observed in MAP-2-positive neurons in the striatum (Fig. 3i) and somatosensory cortex (Fig. 3j and k) of ischemic brain.

image

Figure 3. Ischemia-induced ProTα depletion in neurons is blocked by amlexanox. Amlexanox (Amx) is injected (10 μg/5 μL; i.c.v.) in the mice brain 30 min before cerebral ischemia (1 h tMCAO). (a–j) Coronal brain sections are co-stained with antibodies against ProTα and MAP-2. (a, b) Double fluorescence immunostaining shows that ProTα signal is completely lost in MAP-2-positive neurons in the ipsilateral CA1 pyramidal cell layer of hippocampus (MAP-2, red; ProTα, green) at 3 h after cerebral ischemia (1 h tMCAO), compared with the nuclear ProTα staining in the MAP-2-positive neurons of control brain. (b) Some non-neuronal cells shows higher ProTα immunoreactivity in the ipsilateral stratum radiatum of hippocampus indicated at 3 h by arrow points. (c) ProTα is recovered in nuclei in the ipsilateral CA1 pyramidal neuronal cells at 24 h, but some non-neuronal cells shows the higher ProTα intensity in the stratum radiatum noted by arrows. (d) ProTα staining is completely lost in the ipsilateral MAP2-positive CA1 pyramidal neurons of PBS-pre–treated (vehicle) ischemic mice at 3 h. (e) Following Amx pre-treatment in ischemic mice brain, ProTα release is blocked in the ipsilateral MAP-2-positive CA1 pyramidal neurons and consequently translocated in the neuronal cytoplasmic spaces at 3 h after stress, compared with the contralateral side (f). (g, h) Conforcal microscopy observation indicates ProTα signals in the MAP-2-positive CA1 pyramidal neurons of hippocampus. A higher magnification view is indicated as doted square in (g) and (j), respectively. Arrowheads indicate the 3D imaged line (thickness: 10 μm), as shown in upper (x-axis) and right panels (y-axis). (i, j) Double fluorescence immunostaining shows that ProTα signal is completely lost in MAP-2-positive neurons in the ipsilateral striatum and somatosensory cortex at 3 h after 1 h tMCAO, compared with the normal nuclear ProTα staining in the control. ProTα signal is recovered in nuclei in the ipsilateral striatum and somatosensory cortex at 24 h after ischemia. Amx injection 30 min before ischemia inhibits the release of neuronal ProTα in the ipsilateral striatum and somatosensory cortex at 3 h after ischemic stress. Insets indicate the higher magnification view of ProTα localization in CA1 pyramidal neurons noted by dotted squares. (k) Quantitative analysis of ProTα -positive neurons in the somatosensory cortex. Data represent the means ± SEM (*,#< 0.01, vs. the control: 0 h and the Amx: 3 h, respectively).

Download figure to PowerPoint

Amlexanox-reversible blockade of neuronal ProTα release

To understand the phenomenon whether amlexanox blocks ischemia-induced non-classical release of ProTα in vivo, mice was treated with amlexanox (10 μg/5 μL; i.c.v.) 30 min before cerebral ischemia (1 h tMCAO). Our immunohistochemical results revealed that ProTα signals are completely lost in ipsilateral MAP-2-positive CA1 pyramidal neuronal cells in the hippocampus of PBS-pre–treated (vehicle) brain at 3 h after ischemia and reperfusion (Fig. 3d). However, ProTα was diffused to the cytoplasm from the nucleus in ipsilateral MAP-2-positive CA1 neurons in the amlexanox-pre–treated ischemic brain (Fig. 3e), preserving normal nuclear staining in the contralateral hippocampus (Fig. 3f). In confocal microscopy observation, we also found that ProTα is diffused to MAP-2-positive neuronal cytoplasm from nucleus at 3 h after ischemic stress in the amlexanox-pre–treated brain (Fig. 3h). Similar results were also observed in MAP-2-positive neuronal cells of ipsilateral striatum (Fig. 3i) and somatosensory cortex (Fig. 3j and k) in the ischemic brain.

Caspase 3 inhibition causes ProTα release from astrocytes

To investigate the clue whether ProTα is released from astrocytes in vivo, coronal brain sections were co-stained with anti-ProTα IgG and antibody against an astroglial marker, glial fibrillary acidic protein (GFAP). ProTα immunoreactivity was observed both in nucleus and cytoplasm in the GFAP-positive astrocytes located in the stratum radiatum of control hippocampus (Fig. 4a and g). In the presence of ischemic stress in brain, the findings clarified that the ProTα immunoreactivity is still observed with higher intensity in ipsilateral GFAP-positive astrocytes in the stratum radiatum of hippocampus at 3 h after cerebral ischemia and reperfusion (Fig. 4b), compared with the control brain (Fig. 4a). Indeed, the ProTα signals were increased gradually in astrocytes through 24 h (Fig. 4c). Similar results were also observed in astrocytes in the ipsilateral striatum (Fig. 4h) and somatosensory cortex (Fig. 4i and j) after ischemic stress in brain. Most recently, we demonstrated that ProTα is distributed in both cell body and cytosolic space of processes in adult astrocytes of mouse brain, and that ProTα signal in the astroglial nuclei is drastically increased by diminishing cytosolic levels when pre-treatment with Z-VAD-fmk, a caspase 3 inhibitor (Halder and Ueda 2012). In the present study, we confirmed that ProTα is also localized in both cell body and processes in immature astrocytes in the neonatal mice brain including stratum radiatum of hippocampus (Fig. 4k). Our in vivo experiments revealed that ProTα signal is significantly lost in mature astrocytes in the stratum radiatum of hippocampus as early as 3 h after cerebral ischemic stress in adult mice brain pre-treated with Z-VAD-fmk (Fig. 4e and g), whereas higher reactivity was observed in astrocytes in the PBS-pre–treated (vehicle) ischemic brain (Fig. 4d and g), an indication of ProTα release from astrocytes. Following the combined pre-treatment with Z-VAD-fmk, amlexanox and subsequent ischemia, the signals in astrocytes were recovered as the level in vehicle at 3 h, indicates that Z-VAD-fmk-induced ProTα release from astroglia is blocked by amlexanox (Fig. 4f and g). We also found the similar effects of Z-VAD-fmk and amlexanox on GFAP-positive astroglial ProTα in the ipsilateral the striatum (data are not shown) and somatosensory cortex (Fig. 4j).

image

Figure 4. Initiation of ProTα release from astrocytes by caspase 3 inhibition. Z-VAD-fmk (Z-VAD), a caspase 3 inhibitor (1 μg/5 μL) as well as Amx (1 μg/5 μL) is injected (i.c.v.) in the brain 30 min before cerebral ischemia (1 h tMCAO). Coronal brain sections are co-stained with antibodies against ProTα and GFAP. (a) Immunohistochemical analysis shows the expression of ProTα in GFAP-positive astrocytes in the stratum radiatum of hippocampus in control brain (GFAP, red; ProTα, green). (b) Higher ProTα signal is found in GFAP-positive astrocytes in the ipsilateral stratum radiatum at 3 h after ischemia. (c) The signal is gradually increased in astrocytes through 24 h after ischemia. (d) Intense ProTα signal is observed in GFAP-positive astrocytes in the ipsilateral stratum radiatum at 3 h after ischemia in the in PBS-pre–treated (vehicle) mice. (e) Following Z-VAD-fmk pre-treatment and ischemic stress, ProTα signal is significantly decreased in GFAP-positive astrocytes in the ipsilateral stratum radiatum at 3 h. (f) Z-VAD-fmk-induced ProTα release is blocked from GFAP-positive astrocytes in the ipsilateral stratum radiatum at 3 h after ischemia in Amx pre-treated brain. (g) Higher magnification views of ProTα in astrocytes in the control brain (upper left panel), the vehicle pre-treated ischemic brain (upper right panel), the Z-VAD-fmk pre-treated ischemic brain (lower left panel), and Amx + Z-VAD-fmk pre-treated ischemic brain (lower right panel) at 3 h after stress. (h, i) ProTα is not released from astrocytes in the striatum and somatosensory cortex after cerebral ischemia. ProTα signal is found with higher intensity in GFAP-positive astrocytes in the ipsilateral striatum and somatosensory cortex at 3 h after ischemic stress, compared with the ProTα signals in the control brain. ProTα intensity is gradually increased in GFAP-positive astrocytes in the ipsilateral striatum and somatosensory cortex through 24 h after ischemic stress. (j) Quantitative analysis of ProTα localization of astrocytes in the stratum radiatum of hippocampus (left panel) and somatosensory cortex (right panel). Data represent the means ± SEM. (*,#,†< 0.01, vs. the control: 0 h, the Veh: 3 h, and the Z-VAD: 3 h, respectively). (k) ProTα is localized both in cell body and processes in the astrocytes of post-natal (P1) mice brain (GFAP, red; ProTα, green; Nuclei, blue). Insets indicate the higher magnification view of ProTα expression in astrocytes noted by dotted squares.

Download figure to PowerPoint

No ProTα release from microglia

To see cerebral ischemia-induced ProTα expression in microglia, coronal brain sections were co-stained with anti-ProTα IgG and antibody against Iba-1 (a microglia marker). Recently, we reported that ProTα is localized in both cell body and cytosolic space of processes in microglial in the adult mice brain (Halder and Ueda 2012). Our immunohistochemical analysis suggested that ProTα is also distributed in whole cell in microglia in the adult and neonatal mice brain (Fig. 5a and f). Using the adult brain, the ProTα reactivity was still observed with higher intensity in Iba-1-positive microglia in the ipsilateral stratum radiatum of hippocampus as early as 3 h after ischemia (1 h tMCAO) and reperfusion (Fig. 5b), compared with the control brain (Fig. 5a). However, the ProTα signals were increased gradually in microglia through 24 h (Fig. 5c). On the other hand, there was no effect on microglial ProTα levels by pre-treatment with Z-VAD-fmk in ischemic brain (Fig. 5e), compared with the intensity in the PBS-pre–treated (vehicle) ischemic brain (Fig. 5d). Similar results of ischemia-induced ProTα expression in Iba-1-positive adult microglia were also observed at 3 h and 24 h after ischemia in the regions of striatum (Fig. 5g) and somatosensory cortex (Fig. 5h and i), the regions from where the ProTα signals were completely lost in neurons.

image

Figure 5. Microglia have no tendency to release ProTα. (a–f) Coronal brain sections are co-stained with antibodies against ProTα and Iba-1, a microglial marker. (a) Immunohistochemical data shows the expression of ProTα in Iba-1-positive microglia in the stratum radiatum of hippocampus in control brain (Iba-1, red; ProTα, green). (b) ProTα staining is still found with higher intensity in Iba-1-positive microglia in the ipsilateral stratum radiatum at 3 h after ischemic stress. (c) ProTα intensity is increased gradually in Iba-1-positive microglia through 24 h after ischemic stress. (d, e) PBS (5 μL) and Z-VAD-fmk (Z-VAD) at a dose of 1 μg/5 μL (i.c.v.) are injected in the brain 30 min before cerebral ischemia (1 h tMCAO). (d) ProTα reactivity is still found with higher intensity in Iba-1-positive microglia in the ipsilateral stratum radiatum at 3 h after ischemic stress in the PBS-pre–treated (vehicle) brain. (e) Following Z-VAD-fmk pre-treatment, intense ProTα signal is also observed in Iba-1-positive microglia in the stratum radiatum at 3 h after ischemic stress, indicates that ProTα is not released from microglia. (f) ProTα is expressed both in whole Iba-1-positive microglia in the stratum radiatum of hippocampus of post-natal (P1) brain. (g, h) ProTα is not released from micrglia in the striatum and somatosensory cortex after cerebral ischemia. ProTα signal is found with higher intensity in Iba-1-positive micrglia in the ipsilateral striatum and somatosensory cortex at 3 h after ischemic stress, compared with the ProTα signals in the control brain. ProTα intensity is gradually increased in Iba-1-positive micrglia in the ipsilateral striatum and somatosensory cortex through 24 h after ischemic stress. Insets indicate the higher magnification view of ProTα expression in microglia noted by dotted squares. (i) Quantitative analysis of ProTα-positive microglias in the somatosensory cortex. Data represent the means ± SEM.

Download figure to PowerPoint

Neuronal depletion of S100A13, a cargo protein for non-classical release

Recently, in vitro experiments demonstrated that the non-classical release of ProTα requires interaction with C-terminal sequence of S100A13, a cargo protein (Matsunaga and Ueda 2010). To examine the phenomenon whether S100A13 is released from neurons of adult brain under cerebral ischemic stress (1 h tMCAO), coronal brain sections were co-stained with anti-S100A13 and anti-NeuN antibodies. Our immunostaining data revealed that S100A13 is expressed in NeuN-positive neurons in the CA1 pyramidal cell layer of hippocampus (Fig. 6a), and also in neurons in the striatum (Fig. 6i) and somatosensory cortex (Fig. 6j) of mice brain. As early as 3 h after ischemia and reperfusion, S100A13 was released completely from ipsilateral NeuN-positive CA1 pyramidal neurons (Fig. 6b), compared with the control (Fig. 6a). However, the dot-like signals were observed in some non-neuronal cells in the stratum radiatum of hippocampus (Fig. 6b). S100A13 in pyramidal neurons was recovered in a lesser level at 24 h after ischemia, whereas non-neuronal cells in stratum radiatum completely released S100A13 at this time point (Fig. 6c). Similar results of S100A13 release were observed in ipsilateral neurons of striatum (Fig. 6i) and somatosensory cortex at 3 h (Fig. 6j and k) as well as recovery in neurons at 24 h (Fig 6k) after cerebral ischemic stress.

image

Figure 6. Amlexanox reversibly blocks the ischemia-induced release of S100A13 from neurons. PBS (5 μL) as well as amlexanox (Amx) at a dose of 10 μg/5 μL (i.c.v.) is injected in the mice brain 30 min before cerebral ischemia (1 h MCAO). (a–j) Double fluorescence immunohistochemical analysis of coronal brain sections is performed using antibodies against S100A13 and NeuN, a neuronal marker. (a) S100A13 is expressed in NeuN-positive CA1 pyramidal neurons of hippocampus in the control brain (NeuN, green; S100A13, red). (b) S100A13 signal is completely lost in NeuN-positive neurons in the ipsilateral CA1 pyramidal cell layer of hippocampus at 3 h after cerebral ischemia. Whereas, the dot-like signals are found in some non-neuronal cells in the stratum radiatum of hippocampus at 3 h indicated by arrows (b). (c) S100A13 is recovered with a lesser intensity in the ipsilateral NeuN-positive CA1 pyramidal neuronal cells at 24 h after ischemic stress, but the signal is completely lost in non-neuronal cells in the stratum radiatum at that time point. (d) S100A13 signal is found in NeuN-positive CA1 pyramidal neurons in the control brain. (e) S100A13 signal is completely lost at 3 h in NeuN-positive CA1 pyramidal neurons in the ipsilateral hippocampus of PBS-pre–treated (vehicle) ischemic brain. (f) Amx pre-treatment blocks S100A13 release from ipsilateral CA1 pyramidal neurons at 3 h after ischemic stress. (g, h) Conforcal microscopy observation in CA1 pyramidal neurons. A higher magnification view is indicated as doted square in panels (g) and (h), respectively. Arrowheads indicate the 3D imaged line (thickness: 10 μm), as shown in upper (x-axis) and right panels (y-axis). (i, j) Amlexanox inhibits S100A13 release from striatal and somatosensory cortical neurons. S100A13 is expressed in NeuN-positive neurons in the striatum and somatosensory cortex at 0 h as control. S100A13 signal is completely lost in NeuN-positive neurons in the striatum and cortex at 3 h after ischemia. Amx pre-treatment blocks S100A13 release from striatal and cortical neurons at 3 h after ischemic stress respectively. Insets indicate the higher magnification view of S100A13 expression neurons noted by squares. (k) Quantitative analysis of S100A13-positive neuons in the somatosensory cortex. Data represent the means ± SEM (*,#< 0.01, vs. the control: 0 h and the Amx: 3 h, respectively).

Download figure to PowerPoint

Blockade of neuronal S100A13 release by amlexanox

It has been described previously that the non-classical release of S100A13 from C6 glioma cells is blocked by amlexanox upon serum-deprivation stress (Matsunaga and Ueda 2010). To investigate the in vivo effect of amlexanox on the stress-induced non-classical release of S100A13 from adult brain, mice were treated with amlexanox (10 μg/5 μL; i.c.v.) 30 min before cerebral ischemia (1 h tMCAO) and reperfusion. Using anti-S100A13 and anti-NeuN antibodies, our immunohistochemical findings suggested that S100A13 signals are completely lost in ipsilateral NeuN-positive neurons in the CA1 pyramidal cell layer of hippocampus at 3 h after ischemia in PBS-pre–treated (vehicle) brain (Fig. 6e), retaining the normal staining in the control brain (Fig. 6d). Whereas, S100A13 reactivity was rescued in ipsilateral NeuN-positive CA1 neurons in the ischemic brain pre-treated with amlexanox, an indicative of non-classical blockade of neuronal S100A13 release by amlexanox (Fig. 6f). In confocal microscopy observation, we found that S100A13 immunoreactivity was increased in NeuN-positive CA1 pyramidal neurons of hippocampus due to the blockade of its release by amlexanox (Fig. 6h), compared with the control (Fig. 6g). This result suggests that ischemic stress might cause the up-regulation of S100A13, as shown in Fig. 6c. We also observed the similar results of amlexanox effect on the release of neuronal S100A13 in the striatum (Fig. 6i) and somatosensory cortex (Fig. 6j and k) after the onset of cerebral ischemic stress.

S100A13 is released from astrocytes, but not from microglia

To find out whether S100A13 is released from non-neuronal astrocytes and microglia in the adult brain under cerebral ischemic stress (1 h tMCAO), coronal brain sections were co-stained with anti-S100A13 and antibodies against GFAP and Iba-1. Our double immunostaining data showed that S100A13 is expressed in GFAP-positive astrocytes in the adult mice brain including stratum radiatum of hippocampus (Fig. 7a and c). Following cerebral ischemia and reperfusion, S100A13 signals were partially lost at 3 h in the GFAP-positive astrocytes of ipsilateral stratum radiatum (data are not shown), followed by completely lost at 24 h after cerebral ischemia (Fig. 7b and d). Similar results of astroglial S100A13 release were observed in the striatum (data are not shown) and somatosensory cortex (Fig. 7e).

image

Figure 7. Ischemia-induced release of S100A13 from astrocytes, but not from microglia. Coronal brain sections are co-stained with antibodies against S100A13, GFAP and Iba-1. (a) Double immunofluorescence staining indicates that S100A13 is expressed in GFAP-positive astrocytes in the stratum radiatum of hippocampus in the control brain (GFAP, green; S100A13, red). (b) Following cerebral ischemia and reperfusion (1 h tMCAO), S100A13 signal is lost completely at 24 h in GFAP-positive astrocytes in the stratum radiatum. (c, d) A higher magnification view of S100A13 in astrocytes. Images were collected by a conforcal microscopy. (e) Quantitative analysis of S100A13-positive astrocytes in the somatosensory cortex. Data represent the means ± SEM. (*< 0.01, vs. the control: 0 h). (f) S100A13 reactivity is absent in Iba-1-positive microglia in the stratum radiatum of normal adult brain (Iba-1, green; S100A13, red). (g) S100A13 is not expressed in Iba-1-positive microglia in the stratum radiatum through 24 h after ischemic stress. (h, i) Antigen retrieval microwave technique (h) as well as proteinase K method (i) followed by double fluorescence immunostaining indicates that S100A13 is not expressed in Iba-1-positive microglia in the stratum radiatum of hippocampus in the adult control brain. (j) Immunostaining data show the lack of S100A13 expression in Iba-1-positive microglia in the stratum radiatum of post-natal (P1) control brain. Insets indicate the high-magnification view of S100A13 expression in astrocyte and microglia noted by dotted squares.

Download figure to PowerPoint

However, S100A13 immunoreactivity was absent in Iba-1-positive microglia in the stratum radiatum of normal adult brain (Fig. 7f). We found the similar results of S100A13 absence in the adult microglia of non-ischemic (control) brain using antigen retrieval microwave technique and proteinase K treatment (Fig. 7h and i, respectively). Our findings also suggested that S100A13 is not expressed in Iba-1-positive microglia in the stratum radiatum of post-natal (P1) brain (Fig. 7j). Interestingly, S100A13 signals were absent in Iba-1-positive microglia in the stratum radiatum through 24 h after ischemic stress (Fig. 7g). We found the similar lack of S100A13 expression in Iba-1-positive microglia in the striatum and somatosensory cortex of non-ischemic adult and neonatal brain (data are not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

ProTα, a signal peptide-deficient nuclear protein, has been identified as a unique cell death regulatory molecule in that it converts the intractable necrosis into the controllable apoptosis (Ueda and Fujita 2004; Ueda et al. 2007). This apoptosis is inhibited by brain-derived neurotrophic factor (Ueda 2008). In addition, ProTα potentially inhibits cerebral and retinal ischemia-induced necrosis as well as apoptosis (Fujita and Ueda 2007; Fujita et al. 2009). Ischemia-specific and ProTα-induced up-regulation of brain-derived neurotrophic factor or erythropoietin is found to contribute to this apoptosis inhibition (Fujita et al. 2009; Ueda 2009). Taken together the exclusive findings that the pre-treatments with antisense oligodeoxynucleotide or antibody against ProTα deteriorated the retinal ischemic damages (Fujita et al. 2009; Ueda et al. 2010), it is evident that ProTα is a key neuroprotective molecule against ischemic damages.

In the present study, the novel in vivo findings include the followings: (i) ProTα is non-classically released along with S100A13 from neuronal cells in the CA1 pyramidal cell layer of hippocampus, striatum and somatosensory cortex of adult brain at 3 h after cerebral ischemia; (ii) amlexanox reversibly blocks this non-classical neuronal ProTα release as well as S100A13; (iii) there is no ProTα release from adult astrocytes and microglia after ischemic stress in brain, followed by gradual up-regulation of ProTα signals in these non-neuronal cells through 24 h; (iv) caspase 3 inhibition by Z-VAD-fmk pre-treatment induces ProTα release from astrocytes as early as 3 h after cerebral ischemia, but this release is reversibly blocked by amlexanox; and (v) ischemia-induced ProTα distribution in microglia is not affected either by Z-VAD-fmk or amlexanox.

The recent in vitro study explained that ProTα is localized in the nuclei of cultured cortical neurons and embryonic astrocytes, and that is released from these cells into the extracellular space upon serum-deprivation stress (Matsunaga and Ueda 2010). This study also suggested that the non-vesicular release of ProTα is initiated through the interaction with S100A13 in the serum-deprived C6 glioma cells in vitro. Most recently, we demonstrated that ProTα is strictly localized in nuclei of neuronal cells in adult brain, while it is found in both nuclei and cytosilic space of processes in the astrocytes and microglia (Halder and Ueda 2012). In the present study, we confirmed that ProTα is also localized both in cell body and processes in astrocytes and microglia in the neonatal mice brain, an indication of difference between the brain and culture cell experiments in terms of ProTα expression in astrocytes.

The current study suggested that ProTα is completely released from only neurons in the brain and followed by recovery the signal strictly in the neuronal nuclei after cerebral ischemia, suggesting that the active mechanism may be involved in the epigenetic regulation of ProTα gene expression in neurons of brain under stress condition. It has been reported previously that amlexanox, a potent inhibitor of S100A13, blocks the non-vesicular release of ProTα as well as S100A13 from C6 glioma cells under serum-deprivation stress in vitro, due to loss of interaction of S100A13 with C-terminal sequence of ProTα (Matsunaga and Ueda 2010). This in vivo study demonstrated that ischemia-induced neuronal ProTα release is reversibly blocked by amlexanox and the signals are diffused in the cytosol of neurons. However, S100A13 was distributed in both cell body and cytosol, and released at the same time point as like ProTα from neurons after the onset of cerebral ischemia. Interestingly, this S100A13 release was in turn blocked in neurons, as the same time point as ProTα released was inhibited, in the amlexanox pre-treated ischemic brain. Together, it can be hypothesized that ProTα is non-classically co-released with S100A13 from neuronal cells in the adult brain under ischemic stress in vivo.

At the later time, we characterized the phenomenon whether ProTα is released from non-neuronal cells after cerebral ischemia. We found that none of ProTα is released from astrocytes and microglia after the ischemic stress in brain. Indeed, ProTα intensity was gradually increased in astrocytes and microglia localized in the ischemic regions of brain through 24 h after ischemia. Interestingly, Z-VAD-fmk (caspase 3 inhibitor) pre-treatment induced ProTα release from astrocytes in the ischemic brain, but this release was reversibly blocked by amlexanox. However, S100A13 was expressed in astrocytes of normal brain and released partially at 3 h, followed by complete release at 24 h after brain ischemia. Several in vitro studies suggested that the fragmentation of ProTα is mediated by active caspase 3 at the C-terminal sequence located within the spacer region bipartite nuclear localization signal of ProTα (Rubtsov et al. 1997; Enkemann et al. 2000; Evatafieva et al. 2003; Matsunaga and Ueda 2010). The presence of active caspase 3 in the nuclei of astrocytes in adult brain has also been reported in vivo (Duran-Vilaregut et al. 2010). Recently, it has been demonstrated that S100A13 interacts with C-terminal sequence of ProTα through the C-terminal 11 amino acid peptide sequence of S100A13 in Ca2+-sensitive manner in vitro, and that the expression of ∆88–98 mutant of S100A13 selectively inhibits the stress-induced non-classical release of ProTα, but the release of S100A13 mutant itself occurs from C6 glioma cells (Matsunaga and Ueda 2010). This study explains the crucial role of C-terminal peptide sequence of ProTα for non-classical releasing itself. Most recently, we explained that nuclear ProTα level is drastically increased in the astrocytes of non-ischemic brain by Z-VAD-fmk pre-treatment, an indication of active caspase 3-mediated cleavage of C-terminal part possessing nuclear localization signal of ProTα (Halder and Ueda 2012). Therefore, we can explain the possible mechanisms in the following way: (i) astroglial ProTα in the adult brain might loose the capacity to interact with S100A13 due to the cleavage of C-terminal amino acid sequence of ProTα by activated caspase 3 so that no further release is occurred upon brain ischemia; and (ii) the full-length ProTα is redistributed from the cytosol into nuclei of astrocytes in the brain pre-treated with Z-VAD-fmk, followed by consequent release from astrocytes after cerebral ischemic stress in vivo. Our findings also suggested that Z-VAD-fmk as well as amlexanox has no effect on the distribution of ProTα in microglia in the ischemic brain. The present study explained the lack of S100A13 expression in microglia of non-ischemic brain, even in brain microglia under the cerebral ischemic stress. Therefore, we can explain one possible mechanism is that microglia loses its capacity to release ProTα from ischemic brain due to the absence of S100A13. However, these findings encourage us to investigate the possible intracellular roles of cytosolic ProTα in astrocytes as well as in microglia. Although there is a close interaction between neurons and non-neuronal cells, astrocytes and microglia are more resistant than neurons to most of ischemic stress (Chen and Swanson 2003; Giffard and Swanson 2005; Trendelenburg and Dirnagl 2005; Rossi et al. 2007; Oshiro et al. 2008; Lambertsen et al. 2009; Faustino et al. 2011). There is an interesting report about the ProTα-mediated cellular protection against oxidative stress through the dissociation of the intranuclear Nrf2–Keap1 complex and subsequently facilitation of oxidative stress-protecting genes expression (Karapetian et al. 2005). It has also been described that ProTα prevents cells from apoptosis through the inhibition of apoptosome formation (Jiang et al. 2003; Letsas and Frangou-Lazaridis 2006). Taken together, our findings indicate the possible in vivo role of cytosolic ProTα in astrocytes and microglia in the inhibition of apoptosis.

In the present study, we performed pharmacological inhibitor study against ischemic stress-induced ProTα release. To confirm our hypothesis of intracellular and extracellular roles of ProTα, we need to perform the study using double knockdown or knockout strategies for S100A13, caspase 3, and caspase 7. As the knockdown strategy using an intracerebroventricular injection is presumed to only partially decrease in the levels of these proteins, the conclusion would not be clear. Although double or triple knockout mice would be perfect to discuss this issue, such mice are not available at present. So, detailed mechanisms would be the next subjects.

Several intracellular proteins lack of conventional signal peptides are released from varieties of cells through non-classical endoplasmic reticulum-Golgi-independent pathways under necrotic/ischemic stress (Gardella et al. 2002; Nickel 2005; Prudovsky et al. 2008; Nickel and Rabouille 2009; Matsunaga and Ueda 2010). Such a mode for necrotic/ischemic stress-induced extracellular release from neuronal nuclei seems to be similar to the case with HMGB-1, a popular member of DAMPs (Nickel 2005; Faraco et al. 2007; Foell et al. 2007; Rubartelli and Lotze 2007; Qiu et al. 2008). The reciprocal relation from ProTα would be found in the nature that HMGB-1 induces cytotoxic effects in vitro and in vivo (Scaffidi et al. 2002; Lotze and Tracey 2005; Bianchi 2007; Liu et al. 2007; Qui et al. 2008; Yang et al. 2010; Zitvogel et al. 2010). However, the pattern of ProTα release from neuronal nuclei is as similar as HMGB-1 release, but dissimilar in the case that ProTα induces robust neuroprotection (Fujita and Ueda 2007; Fujita et al. 2009; Ueda et al. 2010). Although ProTα-mediated cell survival activity against viral infection through Toll-like receptor-4 has been reported (Mosoian et al. 2010), the exact receptor for ProTα signaling is yet unknown. Considering the case, ProTα may be referred as a novel neuroprotective molecule of DAMPs family.

In conclusion, the present study demonstrated that neurons, but not astrocytes and microglias, is the main store of endogenous ProTα, which is released through non-classical pathway upon cerebral ischemia, due to the presence of releasing machineries in neuronal cells in brain. Therefore, the discovery cell type-specific mechanisms of ProTα signaling in the brain may provide a novel solution to protect chronic cellular damages in stroke.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank J. Sugimoto and T. Eihara for technical assistance. We also thank H. Kurosu for helpful suggestions. We acknowledge Takeda Pharmaceutical Company Ltd. for providing amlexanox. We also acknowledge T. Maciag for supplying the rabbit anti-S100A13 antibody. Parts of this study were supported by Grants-in-Aid for Scientific Research (to HU) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and Health and Labor Sciences Research Grants (to HU) on Research from the Ministry of Health, Labor and Welfare. We have no conflict interest to report.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
jnc7897-sup-0001-MethodsS1.docWord document46KAppendix S1. Materials and methods.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.