Activation of the High-Mobility Group Box 1 Protein-Receptor for Advanced Glycation End-Products Signaling Pathway in Rats During Neurogenesis After Intracerebral Hemorrhage
Background and Purpose—Following intracerebral hemorrhage (ICH), high-mobility group box 1 protein (HMGB1) may promote neurogenesis that supports functional recovery. How HMGB1 regulates or participates in this process is unclear, as are the pattern recognition receptors and signaling pathways involved.
Methods—ICH was induced by injection of collagenase in Sprague–Dawley rats, which were treated 3 days later with saline, with the HMGB1 inhibitor ethyl pyruvate or with FPS-ZM1, an antagonist of the receptor for advanced glycation end-products. A Sham group was treated with saline solution instead of collagenase and then treated 3 days later with saline again or with ethyl pyruvate or N-benzyl-4-chloro-N-cyclohexylbenzamide (FPS-ZM1). Expression of the following proteins was measured by Western blot, immunohistochemistry, or immunofluorescence: HMGB1, receptor for advanced glycation end-products, toll-like receptor (TLR)-2, TLR4, brain-derived neurotrophic factor, and matrix metalloproteinase-9. The number of cells positive for 5-bromo-2-deoxyuridine or doublecortin was determined by immunohistochemistry and immunofluorescence.
Results—Levels of HMGB1, receptor for advanced glycation end-products, TLR4, TLR2, brain-derived neurotrophic factor, and matrix metalloproteinase-9 were significantly higher 14 days after ICH than at baseline, as were the numbers of 5-bromo-2-deoxyuridine- or doublecortin-positive cells. At the same time, HMGB1 moved from the nucleus into the cytoplasm. Administering ethyl pyruvate significantly reduced all these ICH-induced increases, except the increase in TLR4 and TLR2. Administering FPS-ZM1 reduced the ICH-induced increases in the expression of brain-derived neurotrophic factor and matrix metalloproteinase-9 and in the numbers of 5-bromo-2-deoxyuridine- or doublecortin-positive cells.
Conclusions—These findings suggest that HMGB1 acts via the receptor for advanced glycation end-products signaling pathway to promote neurogenesis in later phases of ICH.
- advanced glycosylation end-product receptor
- box protein 1, high mobility group
- intracerebral hemorrhage
Intracerebral hemorrhage (ICH) accounts for roughly 15% of all strokes and is associated with mortality rates approaching 50%; survivors are often left with serious disability.1 The brain responds to ICH in complex ways that involve pathophysiological responses, such as excitotoxicity, free radical damage, and inflammation. Moreover, the brain initiates remodeling processes, such as neurogenesis, angiogenesis, and synaptic plasticity, designed to restore neurological function after ICH.2–4 As part of this neurogenesis, precursor cells in the subventricular and subgranular zones of the hippocampus migrate into injured brain areas, where they differentiate into mature neurons and glia.4–6
High-mobility group box 1 (HMGB1), a highly conserved nonhistone nuclear DNA-binding protein, is passively released by necrotic cells or actively secreted by macrophages, myeloid dendritic cells, and natural killer cells.7–9 Immediately after ICH, HMGB1 serves as an alarmin or damage-associated molecular signal that mediates cross-talk between damaged and healthy cells and triggers an inflammatory response.7–9 At later times after ICH, however, HMGB1 may promote neurogenesis that supports recovery of neuronal function. HMGB1 has been shown to increase neurite outgrowth in cultures and to promote neurovascular remodeling in cerebral ischemia in mice.10 However, the involvement of HMGB1 in post-ICH neuronal remodeling has never been demonstrated directly.
HMGB1 is known to recruit stem cells and to trigger tissue remodeling after various types of brain injury, such as cerebral ischemia and damaged white matter.11,12 It does so by activating pattern recognition receptors, such as the receptor for advanced glycation end-products (RAGE), toll-like receptor (TLR)-2, and TLR4. TLR2 and TLR4 are well known for mediating innate immune responses to pathogens,13,14 whereas RAGE is a multiligand receptor of the immunoglobulin superfamily.15 It is possible that HMGB1 also triggers these pattern recognition receptors to promote neurogenesis after ICH, but previous studies have not examined this question.
Here we examined whether HMGB1 may be involved in restoring neurological function in a rat model of collagen-induced ICH and whether the protein works via the same RAGE, TLR2, and TLR4 receptors involved in recovery after other types of brain injury.
Materials and Methods
Animals and Treatment Groups
All experiments were performed in accordance with the guidelines for animal research at Sichuan University. Adult male Sprague–Dawley rats were randomly divided into the following 6 groups: ICH, ethyl pyruvate (EP), N-benzyl-4-chloro-N-cyclohexylbenzamide (FPS-ZM1), Sham, Sham+EP, and Sham+FPS-ZM1. How each group was treated is described in detail in Methods I in the online-only Data Supplement. ICH was induced by administering collagenase VII (Sigma C0773, 5.0 ul), as previously described.16 Further details of ICH induction are given in Methods II in the online-only Data Supplement.
Administration of Ethyl Pyruvate and FPS-ZM1
The HGMB1 inhibitor ethyl pyruvate (EP; Sigma) and the RAGE antagonist FPS-ZM1 (Millipore, USA) were dissolved in 0.9% saline solution and administered at respective doses of 60 and 1.5 mg/kg via intraperitoneal injection (IP) once daily starting 3 days after ICH induction.
To label S-phase cells after ICH induction, 5-bromo-2-deoxyuridine (BrdU; Sigma) in 0.9% saline was administered intraperitoneally at a dose of 100 mg/kg once daily for 14 days after ICH.
Western Blot Analysis
Western blotting was performed as previously described16 to measure levels of HMGB1, RAGE, TLR-2, TLR4, brain-derived neurotrophic factor (BDNF), matrix metalloproteinase-9 (MMP-9) in the animal groups (n=5 for each group). Complete details are provided in Methods III in the online-only Data Supplement.
Immunohistochemistry of ipsilateral striatum around the hematoma was performed as previously described16 to assess abundance of HMGB1 and RAGE, as well as determine numbers of cells positive for BrdU (1:100; Abcam) or doublecortin (DCX, 1:50; Abcam). Multiple sections were analyzed from 3 animals randomly selected from each group. Complete details are provided in Methods IV in the online-only Data Supplement.
After deparaffinization and rehydration, sections were incubated overnight at 4°C with antibody against BrdU, DCX, or RAGE, followed by secondary anti-rabbit antibody for 1 hour at room temperature. The striatum was observed by light microscopy and analyzed using Image-Pro Plus 6.0.
At 14 days after ICH or Sham induction, all animals were subjected to a battery of behavioral tests that allowed calculation of the modified neurological severity score.17,18 Eight animals were randomly selected from each group for analysis. Details of the behavioral tests are provided in Methods V in the online-only Data Supplement.
After confirming that data were normally distributed, intergroup differences were assessed for statistical significance using analysis of variance, followed by post hoc testing based on least squares differences. All data were presented as mean±SD, and differences were considered significant at the 5% level.
Expression of HMGB1 in Rat ICH
To examine whether HMGB1 may be involved in rat ICH, we used immunohistochemistry and Western blotting to probe levels of the protein in the ipsilateral striatum in the presence and absence of ICH. Brain sections were stained with hematoxylin-eosin to help ensure that we analyzed ipsilateral striatum around the site of tissue damage (Figure 1A). In the Sham group, HMGB1 was present in nuclei throughout the striatum. After ICH induction, HMGB1 moved from the nucleus into the cytoplasm (Figure 1B). Immunohistochemistry showed a significantly higher number of HMGB1-positive cells in the ipsilateral striatum of ICH rats than in the Sham group (P<0.001; Figure 1B). These results were confirmed by Western blotting (Figure 2; Table I in the online-only Data Supplement).
Expression of HMGB1 Receptors in Rat ICH
Next we examined whether HMGB1 receptors may be involved in how the brain responds to ICH in later stages. Using Western blotting, we measured expression levels of RAGE, TLR2, and TLR4 in ipsilateral striatum on day 14 after ICH induction. Compared with Sham animals, ICH animals showed significant increases in levels of RAGE, TLR4, and TLR2 (Figure 3; Table I in the online-only Data Supplement).
Neurogenic Effects of HMGB1 After ICH
Compared with Sham animals, ICH animals showed higher levels of HMGB1, BDNF, and MMP-9 in the ipsilateral striatum on day 14 after ICH induction, as well as higher numbers of BrdU-positive cells and DCX-positive cells. EP administration led to levels of BDNF in the ipsilateral striatum that were significantly lower than those in the ICH group, but still significantly higher than those in the Sham group on day 14 (Figure 4; Table I in the online-only Data Supplement). EP also led to levels of MMP-9 that were significantly lower than those in the ICH group on day 14, but still significantly higher than those in the Sham group (Figure 4; Table I in the online-only Data Supplement). Similarly, EP administration led to significantly lower numbers of BrdU-positive cells and DCX-positive cells in ipsilateral striatum than in ICH animals on day 14 (both P<0.001; Figure 5).
Both immunohistochemistry and immunofluorescence showed much higher numbers of BrdU-positive cells and DCX-positive cells in the ipsilateral striatum and subventricular zone in ICH animals than in Sham animals on day 14 (Figure 5). To confirm whether neuroblasts were being newly generated in our rat model of collagen-induced ICH, we performed immunofluorescence double staining to determine the overlap in BrdU-positive and DCX-positive cell populations. Our results indicated that BrdU-positive cells were also positive for DCX.
HMGB1-RAGE Axis and Neurogenesis After ICH
Our analysis of HMGB1 expression in ICH and Sham animals suggested that HGMB1 promotes neurogenesis after ICH. This raised the question of which receptor(s) may mediate this HMGB1-induced neurogenesis. To test this hypothesis, we compared post-ICH expression of RAGE and TLR4 in the presence and absence of EP. Administration of this HGMB1 inhibitor led to RAGE levels on Western blots that were much lower than those in the ICH group on day 14, but still significantly higher than those in the Sham group (Figure 3; Table I in the online-only Data Supplement). Similar results were observed by immunohistochemistry (Figure 1C, P<0.001). However, TLR4 and TLR2 expressions were similar in the presence or absence of EP based on Western blotting and immunohistochemistry (data not shown). Moreover, RAGE was easily detectable in the cytoplasm after ICH (Figure 1D).
We performed analogous experiments to compare post-ICH RAGE expression in the presence or absence of the RAGE antagonist FPS-ZM1. FPS-ZM1 administration led to BDNF and MMP-9 levels in the ipsilateral striatum that were significantly lower than those in the ICH group on day 14, but still significantly higher than those in the Sham group (Figure 4; Table I in the online-only Data Supplement). FPS-ZM1 administration led to numbers of BrdU-positive cells and DCX-positive cells that were significantly lower than those in the ICH group, but still significantly higher than those in the Sham group (Figure 5). These data suggest that HMGB1 can act through RAGE to promote neurogenesis after ICH.
Controlling for Effects of EP and FPS-ZM1 not Mediated by HMGB1 or RAGE
To control for the possibility that some or all of the observed effects of EP or FPS-ZM1 on our animals reflected processes not specifically mediated by HMGB1 or RAGE, including general toxic effects, we performed all experiments with Sham animals that we treated with EP or FPS-ZM1 at 3 days after Sham induction of ICH. Protein levels in these 2 groups were similar to those in the Sham group that was given saline 3 days after Sham induction (Table I in the online-only Data Supplement). These results suggest that the observed effects of EP or FPS-ZM1 on protein levels were mostly or entirely mediated by HMGB1 or RAGE.
Neurological function was lower in ICH animals than in Sham animals based on behavioral tests. This function gradually recovered in ICH animals that did not receive any additional treatments; in contrast, the recovery was significantly inhibited by EP and FPS-ZM1 (Figure IA in the online-only Data Supplement). Animals treated with either substance showed significantly lower recovery on day 14 than ICH animals or Sham animals (both P<0.001; Figure IB in the online-only Data Supplement). All 3 Sham groups showed similar results on behavioral tests, suggesting that the observed effects of EP and FPS-ZM1 on functional recovery were mostly or entirely mediated by HMGB1 or RAGE.
In this study, we observed a stroke-induced increase in HMGB1 levels that was associated with increases in BDNF and MMP-9 levels, numbers of BrdU-positive cells and DCX-positive cells, and recovery of neurological function. All these effects were suppressed by the HMGB1 inhibitor EP. EP also inhibited the expression of RAGE, but not of TLR4 or TLR2, all 3 of which can be activated by HGMB1 to trigger downstream signaling pathways. Similar to EP, the RAGE antagonist FPS-ZM1 suppressed ICH-induced increases in BDNF and MMP-9 levels, numbers of BrdU- and DCX-positive cells, and recovery of neurological function. Together, these findings suggest that HMGB1 acts via RAGE-dependent pathways to trigger neurogenesis after ICH.
The nuclear transcription factor HMGB1 is a crucial regulatory molecule that stimulates inflammation after traumatic brain injury and acute stroke.7,18,19 In previous work, we showed that HMGB1 is released into the cytoplasm during the acute phase after ICH, which resulted in neuronal apoptosis, cerebral edema, and neurological impairment.7 Taken together, these results suggest that HMGB1 acts as an early proinflammatory cytokine within the neurovascular unit to mediate inflammation during the acute phase of ICH.
Subsequent to this detrimental proinflammatory activity, the same HMGB1 has been proposed to promote restorative tissue remodeling in later stages of ICH. A study in mice showed that HMGB1 helps drive neurovascular remodeling and recovery of neurological function after cerebral ischemia.11 Another study found that HMGB1 from reactive astrocytes attracts endothelial progenitor cells to sites of white matter injury in order to promote recovery.12 The present study extends these findings by showing that the HMGB1-RAGE axis plays a crucial role in neurogenesis in later stages after ICH.
The present work also extends previous findings that RAGE plays a critical role in determining pathological outcomes in trauma.20 RAGE is also expressed in neurons and glial cells of the brain, and its expression can be unregulated by activated astrocytes and microglia cells.8 In a mouse model of ischemia/reperfusion injury in heart, HMGB1 binds to RAGE soon after injury and thereby activates proinflammatory pathways and exacerbates myocardial injury.21 In a human pancreatic tumor cell culture model, RAGE and HMGB1 coordinately enhance mitochondrial complex I activity, ATP production, tumor cell proliferation, and migration.22 In fact, extracellular HMGB1, which acts as a strong macrophage-activating factor, binds to RAGE, activates endothelial cell proliferation,23 and induces endothelial cell migration and sprouting, thereby promoting tumor progression and propagation.24 RAGE signaling modulates neurotrophin-dependent neurite outgrowth in cultured adult sensory neurons25 and promotes differentiation of neuronal cells.26 The present study adds to the number of pathways, both physiological and pathophysiological, that the HMGB1-RAGE axis helps to regulate.
Our finding that HMGB1 may promote the expression of MMP-9 is consistent with previous studies implicating this enzyme in tissue recovery after injury. In a mouse model of cerebral ischemia, HMGB1 upregulated MMP-9 expression in neurons and astrocytes.27 Upregulation of MMP-9 is believed to promote tissue recovery because the high levels of enzyme degrade extracellular matrix and allow new neuronal cells to migrate toward damaged tissue, where they differentiate into mature neurons that compensate for the loss of neuronal function.28,29 This implies that upregulation of MMP-9 should be associated with neuronal proliferation, which can be measured using the S-phase marker BrdU, and with the appearance of new neurons, which can be measured using DCX.30 We previously reported that ICH induction in the same rat model used here led to significant increases in the levels of MMP-9 and in the numbers of BrdU- and DCX-positive cells in the ipsilateral brain during the later phase after ICH.16 Intracerebroventricular injection of MMP-9 siRNA reduced these ICH-induced increases. In the present study, we found that ICH induced increases in MMP-9 levels and in the numbers of BrdU- and DCX-positive cells in the subventricular zone and ipsilateral striatum on day 14. Furthermore,. we found essentially complete overlap between BrdU- and DCX-positive cells, suggesting that they are the same cells. These results lead us to propose that after ICH, MMP-9 promotes the migration of BrdU- and DCX-positive cells from the subventricular zone toward the ipsilateral striatum.
An advantage of our study is that we correlate changes in protein levels in the ipsilateral striatum with functional recovery using the modified neurological severity score. Although this scoring system has proven effective for assessing functional recovery,17,18 it does not capture all aspects of behavioral function. Another limitation of our study is that we measured neurogenesis using only DCX staining, which is an indirect method because DCX-positive cells may not be viable. Nevertheless, we did find that rat groups with greater numbers of DCX-positive cells also showed greater functional recovery. This suggests that the DCX-positive cells in our study were indeed functional neurons. Future work should examine neurogenesis using more direct methods.
Neurogenesis after brain injury is highly complex and presumably plays a key role in ameliorating the damage and providing at least partial recovery of neurological function. Understanding the pathways responsible for this postinjury recovery will help us understand basic brain biology, and they may also provide clues to molecules and processes that we can manipulate in the clinic to improve and accelerate recovery. The current study provides strong evidence that HMGB1, RAGE, and RAGE-dependent pathways deserve further investigation for their role in promoting neurogenesis after ICH.
Sources of Funding
This research was supported by the National Natural Science Foundation of China (81371283, 81371282).
Guest Editor for this article was Malcolm R. Macleod, PhD, FRCP.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.006825/-/DC1.
- Received July 23, 2014.
- Revision received October 29, 2014.
- Accepted November 18, 2014.
- © 2014 American Heart Association, Inc.
- Zurolo E,
- Iyer A,
- Maroso M,
- Carbonell C,
- Anink JJ,
- Ravizza T,
- et al
- Hayakawa K,
- Pham LD,
- Katusic ZS,
- Arai K,
- Lo EH
- Hoshino K,
- Takeuchi O,
- Kawai T,
- Sanjo H,
- Ogawa T,
- Takeda Y,
- et al
- Dobrovolskaia MA,
- Medvedev AE,
- Thomas KE,
- Cuesta N,
- Toshchakov V,
- Ren T,
- et al
- Herold K,
- Moser B,
- Chen Y,
- Zeng S,
- Yan SF,
- Ramasamy R,
- et al
- Kim JB,
- Sig Choi J,
- Yu YM,
- Nam K,
- Piao CS,
- Kim SW,
- et al
- Andrassy M,
- Volz HC,
- Igwe JC,
- Funke B,
- Eichberger SN,
- Kaya Z,
- et al
- Andersson U,
- Erlandsson-Harris H,
- Yang H,
- Tracey KJ
- Saleh A,
- Smith DR,
- Tessler L,
- Mateo AR,
- Martens C,
- Schartner E,
- et al
- Qiu J,
- Xu J,
- Zheng Y,
- Wei Y,
- Zhu X,
- Lo EH,
- et al
- Ergul A,
- Alhusban A,
- Fagan SC
- Galis ZS,
- Khatri JJ
- Yang S,
- Song S,
- Hua Y,
- Nakamura T,
- Keep RF,
- Xi G