Our recent hypothesis is that targeting the P and IC regions in the ischemic brain would give information on the unique expressions of genes involved in modulating the neuroprotective effect of PACAP38 early on (i.e., 6 h post ischemia) in the ischemic insult. The 6 h time period has been specifically examined in our current analysis of gene expressions in order to know what signaling process occur at this time point, which is the acute phase following PMCAO. As generally believed, during the stroke onset (acute phase) ischemic injury causes a blockage of nutrients via deprivation of blood and there is irreversible damage to the ischemic core; and compared to the 24 h chronic phase where there may be some remodeling signaling in the penumbra (IP). Therefore, our experiment examining the effect of PACAP38 intervention at the 6 h time point is critical to see and seek out the signaling factors working in the IC and P(IP), and how the remodeling via PACAP38 is occurring to cause improvements in the brain in both the IC and the P(IP). Therefore, at present the 24 h time point (chronic phase) has not been examined in the present analysis.

Thus, this current report, first, puts forward a bioinformatics-based evidence to explain the nature of the genes expressed in specifically dissected P(IP) and IC regions in relation to recovery effect of PACAP38 treatment early on (6 h post ischemic insult), in PMCAO mouse model. Second, based on the bioinformatics analyses and gene functions and pathways unravelled during this study, targeted protein analysis was performed using immunoblotting with specific antibodies to further support the role of the previously identified collapsing response mediator protein 2 (CRMP2) [24], and other related up-stream potential target molecules for the proposed neuroprotective mechanisms behind PACAP38 action in mice brain ischemia.

MATERIALS & METHODS

DNA microarray data usage and ingenuity pathway analysis (IPA)

The present study uses raw whole gene expression data from previously conducted experiments in our laboratory and unpublished data (but deposited in the NCBI GEO database: GSE28201, GSE37565, GSE62884, GSE 67421; Figure 1) [12-16]. All the animal research (animal care and experimental procedures) were previously performed with the relevant guidelines and regulations which were approved by the Institutional Animal Care and Use Committee of Showa University (School of Medicine), Tokyo, Japan. PMCAO model mice were generated and brain regions dissected and processed for DNA microarray analysis as described previously, including PACAP38 injection/treatment [6,12-14]. Utilizing these gene expression datasets, the biological function and network analyses were performed using IPA (Ingenuity® Systems, www.ingenuity.com). Concurrent to the previous experiments, a new analysis was performed using the IPA tool [Content version: 23814503 (Release Date: 2015-03-22), QIAGEN]. The dataset from the DNA microarray which includes the differentially expressed (≧/≦ 1.5/0.75-fold) genes and their corresponding fold change values, was uploaded as an Excel spread sheet into the IPA tool. To create gene data (bio-functions, canonical pathways and upstream regulators), the genes were overlaid onto a global molecular network that was developed from information that was contained in the ingenuity knowledge base. The functional analysis identified the biological functions that were most significant to the dataset (p-value <0.05) according to a right-tailed Fisher’s exact test.

Extraction of total soluble protein

Total protein was extracted from approximately 50 mg sample powders (each P/IP and IC brain region; n=6 each, and sample powders were prepared from pooled brain regions that were extremely small in size and quantity due to the microdissection form the mentioned sampling regions of the brain) using a lysis buffer that contained Thiourea and Tris (LB-TT) for extraction of brain proteins [13, 25]. Composition of slightly modified LB-TT was as follows: 7 M (w/v) urea, 42 g; 2 M (w/v) Thiourea, 15.2 g; 4 % (w/v) CHAPS, 4.0 g; 18 mM (w/v) Tris-HCl (pH 8.0), 1.8 mL; 14 mM (w/v) Trizma base, 169.5 mg; 0.2 % (v/v) Triton X-100, 0.2 mL; 50 mM (w/v) DTT, 771.5 mg; 1 % (v/v) pH 3-10 Ampholyte, 1 mL; and two EDTA-free proteinase inhibitor (Roche) tablets in a total volume of 100 mL at room temperature (RT). Protein extraction was initiated by quickly adding 1 mL of LB-TT to the 2 mL microfuge tube containing sample powder following removal from the -80°C deep freezer, and immediately mixed by vortexing (at full speed using a Lab mixer, Iwaki, Tokyo, Japan) for 1 min at RT. Protein solution in LB-TT was incubated at RT for 30 min with mixing by vortexing (30 s) and sonication (30 s in a water-bath type sonicator) for a total of 5 times. Insoluble protein pellet and/or debris were pelleted by centrifugation at 18,500 g for 15 min at 20ºC using a high-speed refrigerated micro centrifuge (MX-150, TOMY, Tokyo, Japan). Clear supernatant (around 900 mL) was transferred to a new 1.5 mL microfuge tube and stored at -80°C as the total soluble protein. Protein concentration was determined with a Coomassie Plus^TM Protein Assay Kit (PIERCE, Rockford, IL, USA) using bovine serum albumin (BSA) as a standard and a NanoDrop 2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA).

Western blot analysis

SDS-PAGE (mini-gel) was carried out on the separated proteins as per standard protocols proteins [13, 25]. The well separated proteins were electro transferred onto a polyvinyldifluoride [PVDF (Trans-Blot Turbo Midi PVDF, 0.2 mM, Transfer Packs kit; Cat. no. 170-4157)] using the Trans-Blot Turbo Transfer System (Bio-Rad; StandardSD protocol; 25V, 1.0 A, 30 min). Following the transfer of proteins on the PVDF membrane (confirmed by visualizing all the 10 colored molecular mass standards), the membrane was incubated in 25 mL of 5 % blocking solution (Block-Ace powder, Cat. no. UK-B80, DS Pharma, Osaka, Japan; Yukizirushi, Sapporo, Hokkaido, Japan) for 1 h under constant slow shaking at RT. Blocking solution was prepared by dissolving 4 g powder in 80 mL 1X TTBS [10X TTBS: NaCl, 80 g; 1 M Tris-HCl, pH 7.5, 200 mL; Tween-20, 5 mL]. Western blotting and detection was carried out using the Immun-Star WesternC Chemiluminescent Kit (Cat. no. 170-5070, Bio-Rad) following the manufacturer’s instructions. Blocking solution was decanted and the membrane was washed once in 1X TTBS (5 min), followed by incubation in 25 mL of primary antibody solution [PAS; 1 mL rabbit anti-CRMP2 phophoT514 protein antibody (Cat. no. ab62478; 100 mg, 2 mg/mL; Abcam, www.abcam.co.jp; anti-CRMP2 phophoSer522 and anti-CRMP2 phophoThr555] for 1 h, as above. The membrane was then washed with 25 mL of 1X TTBS for five times. The last TTBS wash was decanted and the membrane was incubated in 25 mL of secondary antibody solution [SAS; 0.5 mL of Amersham, ECL anti-rabbit IgG, HRP linked species-specific whole antibody (from Donkey); Cat. no. NA 934; GE Healthcare] for 1 h with slow shaking at RT. The 1X TTBS wash step was repeated five times. For blot development, the luminol/enhancer and peroxide buffer solutions were mixed in a 1:1 ratio (1 mL: 1 mL; one membrane volume) and spread over the membrane and incubated at RT for 5 min. Excess solution was drained by touching one end of the membrane on a KimWipe paper towel and the signal (cross-reacting protein bands) was visualized using the ChemiDoc XRS+imaging system (Bio-Rad). The Western blot analysis was repeated at least three times, and representative data from one blot is shown.

RESULTS AND DISCUSSION

Specifically up- and down-regulated genes in ischemic core and penumbra following PACAP38 treatment and their IPA

IPA-based bioinformatics analysis provided a number of IC-specific genes: 241 (up-regulated)/229 (down-regulated) and also P(IP)-specific genes: 181 (up-regulated)/236 (down-regulated) at 6 h post-ischemia following the injection of PACAP38 (Figure 1). The 6 h early time point was targeted because of early visualization of the brain regions and ischemia (also by TTC staining which shows the ischemic region in the brain) in the brain along with numerous gene expression changes [12-16] and the very early treatments for stroke given within 4 to 6 h of the stroke. Keeping in mind the hypothesis that PACAP38 should be effective by its nature as a neuroprotective peptide/agent on the cells (or neurons) by direct action or indirect action through modulating the cells through activating the release of neuroprotective factors. Thus the authors wished to see if the narrowed down microarray-based transcriptome contains these hints and whether they can be delineated in the IC and P(IP) regions, on the very early time point (acute phase) of 6 h post-ischemia. Remarkable differences were seen upon performing IPA and generating the bio-functions, canonical pathways and upstream regulators for both IC (Figures 2, 3 and 4) and P(IP) (Figures 5, 6 and 7).

Both IC and the P(IP) regions are affected by the PACAP38 injection in various gene bio-functions and pathways and upstream targets; a greater number of gene categories are being affected in the IC by PACAP38. In the IC gene categories of bio function: lipid metabolism, small molecule biochemistry, gene expression, cell to cell signaling and interaction, others (down-regulated) to up-regulated genes in cellular development to embryonic development and cell morphology are seen (Figure 2); In the IC, gene functions can be sub-divided mainly into antithrombotic, ant-inflammation and which are expected if PACAP38 is acting to suppress the spread of the ischemia. This is not the case in the P(IP) region. In the case of bio-functions, PACAP38 appears to show a general decrease in the categories of the nervous system development and function (sensation and nociception functions), cell cycle, tissue development and cellular growth and proliferation; all these are related to reduction in cellular proliferation function in the P(IP) region (Figure 5). This can be expected as the penumbra (IP) is supposed to be the site or boundary between active recovery (IC, under PACAP38) and the presumably healthy regions. However, we find that an increase in the gene category of endocrine system disorders with benign thyroid disease as the annotated function is seen in the P(IP) (Figure 5). For example, one of the prominent genes is CTLA4 which suppresses the T-cells sending them turn-off signals in case of chronic infections [26]. This anti-inflammation function of CTLA4 would be required for the P(IP) region and which should not be actively producing a defensive response by the T cells. Similarly, the TTR gene is being up-regulated and which along with BDNF has been shown to have neuroprotective effects [15, 27-29]. However, it is to be noted that TTR and especially BNDF may have different functions depending on the cells and conditions and that is why we see it (BDNF) being reduced in the annotated function of sensation.

Second, in the canonical pathways, IL1 signaling, dopamine receptor signaling, acute phase response signaling, CDK5 signaling, and cAMP-mediated signaling show a decrease in the P(IP) region (Figure 6) compared to the opposite gene functions in the IC (Figure 3). Anti-inflammation and blood coagulation are the main events occurring in the P(IP) region following PACAP38 treatment, especially with the cAMP-mediated pathway having an effect on the blood coagulation. Considering the fact that the region around the core is healthy, it would not be desirable to have a higher level of blood flow in response to the ischemic insult. In the case of the IC region, we see that Ga12/13 signaling function,leukocyte extravasation signaling, LXR/AXR signaling is being decreased. These decreased functions can be sub-divided into anticoagulation, anti-inflammation. The G12/13 family is well known for its involvement in cell proliferation and morphology, and PACAP38 might be function to prevent cell proliferation till the recovery process is best suited for new cell/neuron growth. Interestingly, the anti-inflammation response function might require a reduction in the fatty acid synthesis, and this is what we find (IL1RN) in the IC region. On the other hand, Wnt/β-catenin signaling may be more pronounced in the P(IP) region as compared to the IC region where it may both be induced and suppressed.

Thirdly, in the case of identified upstream regulators (Figures 4 and 7), we can see that proliferation (STAT5B) is reduced in the P(IP) region compared to the growth factor (WISP2) which is activated. These play active roles in the central nervous system and their modulation by PACAP38 suggests PACAP38 recruitment of these molecular factors and pathways for aiding the recovery process in the P(IP) region. Taken together, these different gene functions being differentially expressed under PACAP38 injection in the IC and P(IP) regions of the ischemic brain are evidences of a specific and different PACAP38 action in the IC and P(IP), in light of its neuroprotective role in the ischemic brain. Further, in the IC region, as upstream regulators, MKL1 and MKL2 can be seen along with their target molecules, and where both are inhibited (Figure 4). The MKL/myocardin-like protein 1 or myocardin-related transcription factor A (MRTF-A) produces a protein which is a key regulator of smooth muscle cell differentiation. Interestingly, the protein has other functions, including in the intracerebral hemorrhagic stroke and a role in maintaining cerebral small vessel (blood) integrity [30]. This makes for an important finding for inhibited MLK1/MRTF-A function also in neuroprotective functions under PACAP38 treatment. In the P(IP), two upstream regulators, STAT5B (signal transducer and activator of transcription-5, inhibited) and WISP2 (Wnt-1 inducible signaling pathway protein-2, activated) were identified (Figure 7). Both these are now know to be involved in the ischemic insult signaling cascades. For example, the STAT5B has been shown to contribute to the erythropoietin-mediated neuroprotection against hippocampal neuronal death after transient global cerebral ischemia [31], whereas, the WISP2, a member of the connective tissue growth factor and nephroblastoma overexpressed gene family of matricellular proteins is critical in growth factor mediated cell proliferation [32,33]. In summary, the identification of these three (MKL1, STAT5B and WISP2) factors in the ischemic brain regions supports the proposed PACAP-mediated neuroprotective function.

Six-hour post ischemic response in the penumbra might involve cell cycle suppression in conjunction with CRMP2 function

Previously, we had shown that the CRPM2 [24] protein is accumulated during the ischemia following PACAP38 treatment, especially at 6 h post-ischemia, and we also showed that the additional band observed in our experiments might be not a degradative product but having some role in neuroprotection following PACAP38 treatment [14,15]. Recently, using brain-specific Crmp2 knockout (cKO) mice it was shown that these mice not only display molecular, cellular, structural and behavioral deficit but also that a loss of Crmp2 in the hippocampus leads to reduced long-term potentiation, abnormal NMDA receptor composition, aberrant dendrite development and defective synapse formation in CA1 neurons, and a knockdown of crmp2 specifically in newborn neurons results in stage-dependent defects in their development during adult hippocampal neurogenesis [34]. This recent paper [34] provides the strongest evidence to date on the importance of CRMP2 in axonal growth/neurogenesis, and we attempted to link our gene data with CRMP2 action, and which is presented as a model in Figure 8 (upper panel). The identified (DNA microarray-based gene expressions) up-stream up-regulated (in red arrowheads) and down-regulated (in green arrowheads) molecules are marked in the diagram, which is linked to the phosphorylation-dephosphorylation of the CRMP2 protein, in our model (Figure 8, lower panel). PACAP38 influences the regulation of other gene functions as indicated, especially those acting to suppress the GSK3β-pathway; it may be possible that CRMP2 action is enhanced in the penumbra (IP).

 Our combined data also focuses on the short form of CRMP2 which could play an important role in nerve regeneration upon treatment by PACAP38. A study by Rogemond et al. [35] looked at a new short isoform of 58 kDa, expressed during brain development, and derived from C-terminal processing of the CRMP2B subtype. Those authors demonstrated its presence in the nuclear fraction of brain extract and showed that this short CRMP2 induces neurite outgrowth inhibition in neuroblastoma cells and suppressed axonal growth in cultured cortical neurons [35]. They went on to suggest that the post-transcriptionally processed and nuclear localized CRMP2 may be key in the regulation of neurite outgrowth in brain development. Rogemond et al. [35] study shows a link with our finding of CRMP2 involvement and its post-translational modified state among the signaling pathways identified via the IPA analysis in IC and IP regions. Dephosphorylation of CRMP2 by PACAP38 was confirmed by Western blotting. As a result of using the CRMP2 antibodies of Ser522, Thr555 and Thr514, it was confirmed that the addition of PACAP38 promoted the dephosphorylation (Figure 8, upper panel-A, B, C). RhoA signaling, CDK5 signaling pathways are involved in CRMP2 dephosphorylation (Ser522, Thr555, Thr514), and gsk3β [36]. The phosphorylation of Gsk3β by PACAP was confirmed by Western blotting. Gsk3β is known to be inactivated by phosphorylation (GSK3β-Ser9). As a result of Western blotting, it was confirmed that the expression level of Gsk3β was reduced by PACAP38, but it was not confirmed that phosphorylation (GSK3β-Ser9) was promoted. As a result of the IPA pathway analysis, PACAP38 suppressed RhoA Signaling and its upstream Gα12/13 Signaling in the 6 h IC, and suppressed CDK5 Signaling in 6h IP, as well as CXCR4 signaling, ILK signaling, ERK/MAPK signaling, Wnt/β-activation of catenin signaling was predicted. As shown in Figure 8 (lower panel) from these results, in the IC 6 h scenario RhoA signaling and its upstream Gα12/13 signaling dephosphorylates CRMP2-Thr555 and suppresses GSK3β, and in the IP 6 h scenario the CDK5 signaling dephosphorylated CRMP2-Ser522. It was suggested that CXCR4 Signaling, ILK Signaling, ERK/MAPK Signaling, and Wnt/βcatenin signaling are involved in GSK3β suppression and CRMP2-Thr514 dephosphorylation.

CONCLUSION

Using the unique datasets of specifically up and down-regulated genes in the IC and P(IP) regions post PACAP38 treatment, we have provided new insight into how PACAP38 might act in the ischemic brain of mice. The datasets have been obtained from the highly credible and reproducible Agilent Gene Expression Microarray platform wherein we utilized the two-color and dye-swap approach (Agilent Technologies). Our results suggest how the PACAP38 might be affecting the recovery of the ischemic region by targeting the adjacent P(IP) region signalling pathways and this neurogenesis or new nerve growth would be critical for the recovery from the ischemic insult and therein to recovery. However, the above discussion based on the obtained bioinformatics gene analysis (IPA) will have to be proved further in wet-lab experiments and downstream analyses in new studies. It is our view that a bioinformatics strategy to dissect out genes specifically to the ischemic and adjoining regions will be most helpful in not only pinpointing gene candidates potentially involved in mediating the neuroprotective effect of PACAP38 but also support the reasons behind the current therapy for stroke in human patients. A recently published paper utilized metabolomics, the third high-throughput technique in the omics arsenal, to suggest the involvement of small molecule metabolites in brain injury and functional recovery. Their study [37] combined with transcriptomic and proteomic data (recently reviewed by [18]) imply that continued research and interlinking big data (omics) will be critical to truly understand the brain stroke and mechanisms therein, which would then lead to not only diagnostic markers but also therapeutic drugs.

ACKNOWLEDGMENT

 This work is supported by a grant KAKENHI Grant Number JP16H02684 (SS). RR also acknowledges the support of KAKENHI (C): 26462745 in this research.