Hepatoma-derived growth factor is a neurotrophic factor harbored in the nucleus.

Hepatoma-derived growth factor (HDGF) is a heparin-binding proliferating factor originally isolated from conditioned medium of the hepatoma-derived cell line HuH-7. HDGF has greatest homology in an amino acid sequence with high mobility group 1 (HMG1), which has been characterized as a DNA-binding, inflammatory, and potent neurite outgrowth molecule. HDGF is reported to be widely expressed and act as a growth factor in many kinds of cells. However, it has not been investigated in the nervous system. Here, we show by Western blot analysis that HDGF is present in the mouse brain from the embryonic period until adulthood. In situ hybridization and immunohistochemical analyses revealed that HDGF was expressed mainly in neurons, and HDGF protein was localized to the nucleus. HDGF and high mobility group 1 were secreted under physiological conditions and released extracellularly in necrotic conditions. Furthermore, we showed that exogenously supplied HDGF had a neurotrophic effect and was able to partially prevent the cell death of neurons in which endogenous HDGF was suppressed. Therefore, we propose that HDGF is a novel type of neurotrophic factor, on account of its localization in the nucleus and its potential to function in an autocrine manner under both physiological and pathological conditions throughout life.

HDGF has greatest homology in amino acid sequence (32%) to high mobility group 1 (HMG1), which has been reported to have intranuclear and extracellular functions (17). HMG1 is an abundant component of mammalian nuclei (18,19). It bends the double helix on binding through the minor groove and binds with high affinity to DNA that is already sharply bent, such as linker DNA at the entry and exit of nucleosomes (20,21). HMG1, also called amphoterin, has been identified as an extracellular neurite growth factor (22), as a secreted late mediator of endotoxin shock (23), as a mediator of acute inflammation (24), and furthermore as a molecule related to tumor metastasis (25), supporting the notion that it functions in the extracellular space. In HDGF, the absence of the "HMG box," which is essential for binding DNA, also strongly suggests that HDGF has extracellular roles.
Although intensively studied in many tissues and cells, HDGF has not been investigated in the nervous system. We examined the localization of HDGF in the mouse nervous system and found that it is abundantly expressed in neurons. Since neurons do not replicate, HDGF cannot be considered a proliferation-inducing factor for neurons. In the present report, we describe the neurotrophic effects of HDGF harbored in nuclei.

MATERIALS AND METHODS
Western Blot Analysis-Western blot analysis was carried out according to the method of Sambrook et al. (26). Protein was solubilized in radioimmune precipitation assay buffer consisting of 10 mM Tris, pH 7.5, 0.1 M NaCl, 1 mM EDTA, and 0.1% Triton X-100. The concentration of the protein was determined using a Bio-Rad protein assay kit (Bio-Rad). The protein was subjected to 15% SDS-PAGE and electrotransferred to an Immobilon membrane (Millipore Corp., Bedford, MA). It was blocked by incubation for 24 h at 4°C with 5% nonfat dry milk in TBS (10 mM Tris, pH 7.5, 150 mM NaCl). The membrane was then incubated with anti-HDGF polyclonal antibody (C terminus) (14) at a dilution of 1:5000 in the blocking buffer for 1 h at room temperature. After being washed with TBS-Tween/Triton buffer (20 mM Tris, pH 7.5, 500 mM NaCl, 0.05% Tween 20, and 0.2% Triton X-100), the membrane was incubated with horseradish peroxidase-conjugated anti-rabbit IgG antibody (Cappel, Aurora, OH) at a dilution of 1:5000 in blocking buffer for 1 h at room temperature. After another wash, the membrane was developed using the ECL system (Amersham Biosciences).
In Situ Hybridization-In situ hybridization was carried out according to the method of Yamamoto et al. (27) with some modifications. cDNA encoding mouse HDGF (nucleotides 356 -775) (2) was inserted into pBluescript II (Stratagene). Probes were prepared using a Digoxigenin RNA Labeling Kit (Roche Applied Science) according to the manufacturer's instructions. Corresponding sense probes were used as controls. Eight-week-old male mice (C57BL/6) were anesthetized with sodium pentobarbital (5 mg/kg) intraperitoneally. Brains were carefully removed, frozen in dry ice powder, and stored at Ϫ80°C until sectioning. Serial sections (15 m) were cut with a cryostat (Bright; Huntington, UK). After being mounted onto slides, sections were dried with a dryer and fixed in PBS containing 4% paraformaldehyde for 20 min at room temperature. After being washed with PBS, sections were treated with 10 g/ml proteinase K containing 50 mM Tris, pH 7.5, and 5 mM EDTA, and the reaction was stopped by treating with 4% paraformaldehyde in PBS for 20 min. Sections were covered overnight at 55°C with hybridization buffer (50% deionized formamide, 0.3 M NaCl, 20 mM Tris, pH 8.0, 10% dextran sulfate, 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin, 500 g/ml yeast tRNA, 0.2 mg/ml salmon testes DNA, 0.2% n-lauroyl sarcosine sodium salt) containing 0.5 g/ml antisense or sense probes. The following day, the sections were subjected to a series of stringent washes and RNase treatment before immunological detection with anti-digoxigenin-AP conjugate (1: 1000; Roche Applied Science) for 1 h at room temperature. The sections were then washed and exposed to a coloring reaction solution of 450 g/ml nitro blue tetrazolium chloride (Roche Applied Science) and 175 g/ml 5-bromo-4-chloro-3-indolyl-phosphate (Roche Applied Science).
Immunohistochemistry-Eight-week-old male mice (C57BL/6) were anesthetized with sodium pentobarbital (15 mg/kg) intraperitoneally and then perfused transcardially with 4% paraformaldehyde in PBS, pH 7.4. Brains were carefully removed and postfixed overnight at 4°C in the same fixative. Tissue blocks were rinsed for 1 h with PBS and dehydrated through a series of increasing concentrations of ethanol. After dehydration, they were cleared with chloroform and xylene and then embedded in paraffin. The blocks were sectioned 6-m-thick onto poly-L-lysine-coated slides (Matsunami, Japan). The sections were deparaffinized, and endogenous peroxidase activity was quenched by incubation with 0.3% hydrogen peroxide in methanol for 10 min at room temperature. The sections were incubated with anti-HDGF polyclonal antibody (C terminus) (14) at a dilution of 1:2000 in blocking buffer (2% bovine serum albumin in PBS) for 1 h at room temperature. Control staining was conducted without the primary antibody or with preabsorbed antibody, in which the diluted primary antibody and 2 M recombinant HDGF had been mixed and incubated at 4°C for 24 h. After being washed with PBS, the sections were incubated with secondary antibody labeled with peroxidase (Histofine simple stain kit; Nichirei, Japan) for 30 min. The sections were visualized with 3,3Јdiaminobenzidine and hydrogen peroxide and counterstained with hematoxylin.
Preparation of Mouse Primary Hippocampal Neurons-Mouse primary hippocampal neurons were prepared from embryonic day 17 C57BL/6 mouse embryos, with some modifications of the published protocol (28). Fetal hippocampi were dissected and digested with calcium/magnesium-free Hanks' balanced salt solution (Invitrogen) containing 0.025% trypsin for 8 min at 37°C. The supernatant was then removed, and the tissues were dissociated by repeated trituration with 0.1 mg/ml DNase in Hanks' balanced salt solution and layered gently on a 1-ml cushion of bovine serum albumin (4% w/v) in a 15-ml plastic tube. After centrifugation, cells were seeded onto poly-DL-ornithine (Sigma), laminin (Collaborative Biomedical Products, Bedford, MA)coated dishes at a density of 1 ϫ 10 5 /cm 2 and maintained in serum-free defined medium at 37°C in a humidified atmosphere of 5% CO 2 and 95% room air. The serum-free defined medium consisted of neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen), 0.3 g/liter glutamine, and penicillin/streptomycin. The purity of the neurons was confirmed to be more than 90% at day 1 and 95% at days 2 and 4 by staining with anti-neuronal class III-␤-tubulin (Berkeley Antibody Company, Richmond, CA), rabbit anti-cow anti-glial fibrillary acidic protein (DAKO, Denmark), and propidium iodide (4 g/ml).
Secretion of HDGF and HMG1-Mouse primary hippocampal neurons were prepared as described above. Neurons were seeded onto poly-DL-ornithine (Sigma), laminin (Collaborative Biomedical Products, Bedford, MA)-coated 6-cm tissue culture dishes (Falcon) at a density of 1 ϫ 10 5 /cm 2 and cultured for 24 h in neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen), 0.3 g/liter glutamine, and penicil-lin/streptomycin at 37°C in a humidified atmosphere of 5% CO 2 and 95% room air. After washing with PBS, 4 ml of Krebs-Ringer medium, consisting of 124 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 1.2 mM KH 2 PO 4 , 26 mM NaHCO 3 , and 10 mM glucose, was conditioned by incubation with the hippocampal neurons for 3 h. The conditioned medium was then collected and concentrated 100-fold using a Centricon YM-10 (Millipore). The cells were solubilized in radioimmune precipitation assay buffer consisting of 10 mM Tris, pH 7.5, 0.1 M NaCl, 1 mM EDTA, and 0.1% Triton X-100. For analyses of the mouse neuroblastoma cell line Neuro2a, 1 ϫ 10 6 cells were seeded onto a 10-cm tissue culture dish (Corning) and cultured until ϳ70% confluence in Eagle's minimal essential medium supplemented with 10% fetal calf serum at 37°C in a humidified atmosphere of 5% CO 2 and 95% room air. After washing with PBS, 4 ml of Opti-MEM I (Invitrogen) supplemented with N-2 (Invitrogen) was conditioned by incubation with the Neuro2a cells for 3 h. The conditioned medium and the cells were then treated as described above. Fifteen microliters of the concentrated conditioned medium and 1 g of the cell extract were used for Western blot analysis with anti-HDGF polyclonal antibody (C terminus) (14) or affinity-purified rabbit anti-HMG1 polyclonal antibody (Pharmingen).
Induction of Cell Damage-Three hundred thousand Neuro2a cells were seeded onto 3.5-cm tissue culture dishes (Falcon) and incubated until ϳ90% confluence in Eagles' minimal essential medium supplemented with 10% fetal calf serum at 37°C in a humidified atmosphere of 5% CO 2 and 95% room air. After washing with PBS, the medium was replaced with 2 ml of Opti-MEM I (Invitrogen), and the Neuro2a cells were either induced to undergo necrosis by treatment with 10 M ionomycin (Sigma) and 20 M carbonyl cyanide 3-chlorophenylhydrazone (Sigma) for 24 h or induced to undergo apoptosis by treatment with 300 nM etoposide (Sigma) for 24 h. Neuro2a cells without treatment served as a control. For Western blot analysis, the media from treated and untreated cells in 3.5-cm tissue culture dishes were collected and concentrated 10-fold using a Centricon YM-10 (Millipore). For staining, 5 ϫ 10 3 Neuro2a cells were seeded onto 8-well chamber slides (Nunc) and induced to undergo necrosis or apoptosis and then fixed with 4% paraformaldehyde and 0.1% glutaraldehyde in PBS for 15 min at room temperature. After being washed with PBS, cells were permeabilized with 50 mM L-lysine and 0.1% Triton X-100 in PBS for 15 min and blocked for 1 h with 10% normal goat serum in PBS. Cells were incubated with rabbit anti-HDGF polyclonal antibody (C terminus) (14) in PBS containing 5% normal goat serum for 1 h at room temperature. After washing with PBS, the bound primary antibodies were labeled by incubation with goat anti-rabbit IgG conjugated to fluorescein (1:50; Cappel) in PBS containing 5% normal goat serum. After further washing, the cells were treated with propidium iodide (4 g/ml), mounted, and observed with an Olympus microscope equipped with epifluorescent filters. Laser scanning was performed with a Zeiss LSM510, Axiovert 25 confocal imaging system.
Assessment of Neuronal Survival and Death-Brain-derived neurotrophic factor (BDNF) was purchased from R&D Systems (Minneapolis, MA). Recombinant HDGF was purified as previously described (16). Mouse primary hippocampal neurons were prepared as described above. After a 6-h incubation with the serum-free defined medium (neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen), 0.3 g/liter glutamine, and penicillin-streptomycin), the medium was replaced with serum-free defined medium lacking B27 (Invitrogen) in order to study neuronal survival (28) but containing BDNF (10 pg/ml, 100 pg/ml, 1 ng/ml, 10 ng/ml, or 100 ng/ml) to determine the optimal BDNF concentration. Half of the medium was exchanged every other day, and after 5 days of incubation, cell viability and death were assessed using the MTT colorimetric assay (Chemicon, Temecula, CA) and a lactate dehydrogenase (LDH) Assay Kit (Wako, Japan), respectively. Neuronal survival was almost identical in the presence of BDNF at concentrations of 10 and 100 ng/ml. Finally, to examine the neurotrophic effect of HDGF, we added HDGF (1, 10, 100, or 500 ng/ml) or an optimal concentration of BDNF (10 ng/ml) to primary hippocampal neurons. Survival values were normalized, taking the survival value without factors as 0% and that in 10 ng/ml BDNF as 100%. All experiments were carried out in triplicate and repeated four times.

Assessment of Survival of Endogenous HDGF-suppressed Neurons-
For transfection, primary hippocampal neurons were seeded at a density of 1 ϫ 10 5 /cm 2 onto 3.5-cm tissue culture dishes (Falcon) and cultured for 24 h in neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen), 0.3 g/liter glutamine, and penicillin/streptomycin. In the serum-free defined medium supplemented with B27, ϳ65% of neurons were alive 5 days after transfection, compared with the number 1 day after transfection without trophic factors. The transfection of primary hippocampal neurons has proven to be inefficient, is often toxic to the cells, and does not yield reproducible results. To rectify this situation, we performed the co-transfection of two plasmids, pEGFP-C1 (Clontech) to mark the transfected cells and pSUPER-HDGF to suppress endogenous HDGF expression. The ratio of the plasmids was adjusted so that more than 98% of the GFP-positive cells were endogenous HDGF-suppressed cells. We used FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's instructions for its low toxicity and obtained reproducible results. The transfection efficiency was found to be ϳ10%. In this paradigm, although the transfection efficiency is low, one can follow endogenous HDGF-suppressed neurons using GFP and analyze the time course of the marked cells with and without exogenously supplied HDGF. To test whether exogenously supplied HDGF protects neurons against death, recombinant HDGF (100 ng/ml) was added at the time of transfection. From 1 to 5 days after transfection, GFP-positive neurons were counted every day with a Nikon DIAPHOT TMD 300 culture cell viewing microscope system. Twenty random microscopic fields at ϫ 200 magnification were counted for each dish. The transfection efficiencies for each condition were rather constant for each experiment as assessed from the GFP-positive neurons. Thus, the percentage of GFP-positive neurons after transfection was calculated in relation to the number of GFP-positive neurons at 1 day after transfection, defined as 100% for each condition. Assays were performed in triplicate for seven independent experiments.
Statistical Analysis-Statistical comparisons among groups were made using one-way analysis of variance. Student's t test was used for the post hoc comparison of individual mean values. p values less than 0.05 were considered statistically significant. Data shown represent the means Ϯ S.E.

HDGF Is Synthesized and Localized Mainly in Neurons in
the Nervous System-It has been observed that HDGF mRNA is expressed in the brain using Northern blot analysis (1,15). To examine the expression of HDGF protein in the nervous system, a developmental study was carried out by Western blot analysis of the mouse brain (Fig. 1). We found that HDGF protein was constantly expressed in the brain from embryonic day 13 until 2 years of age.
To identify cells producing HDGF in the nervous system, in situ hybridization was performed using 8-week-old male mouse brains. HDGF mRNA was highly expressed in neurons, especially in hippocampal neurons and Purkinje cells (Fig. 2, A, C, and E). No significant staining was observed with sense probes (data not shown).
To examine the localization of HDGF protein in the nervous system, immunohistochemical experiments were carried out using 8-week-old male mouse brains. No staining was visible without the primary antibody (data not shown) or with the preabsorbed antibody (Fig. 2H). HDGF was highly stained in the nuclei of most of the hippocampal neurons and Purkinje cells (Fig. 2, B, D, and F). In cerebral cortex, most of the neurons were stained with anti-HDGF antibody. Although some glial cells were stained with anti-HDGF antibody, most of them were HDGF-negative (Fig. 2G).
These results indicate that HDGF is continuously produced in the nervous system and that in adult mice, HDGF is produced mainly in neurons and retained in their nuclei.
HDGF and HMG1 Are Secreted in Physiological Conditions-We tested whether HDGF and HMG1 were secreted in neuronal cells, using mouse primary hippocampal neurons and the mouse neuroblastoma cell line Neuro2a. The purity of the neurons was confirmed to be more than 90% (Fig. 3A). As shown in Fig. 3, B and C, both HDGF and HMG1 were detected in the conditioned media by Western blot analysis. This result was consistent with previous reports showing that HMG1 was secreted by certain cells (25,30,31). We also estimated the amount of HDGF by Western blot analysis using a series of recombinant HDGFs as standards (data not shown). The amount of HDGF in the medium of the hippocampal neuronal culture was estimated to be ϳ13 pg/24 h/10 5 cells. The amount in the hippocampal neurons in culture was ϳ1.3 ng/10 5 cells. Thus, approximately one-hundredth of the intracellular HDGF in the hippocampal neurons was secreted over 24 h in vitro.
Necrotic Cells Release HDGF-It has been reported that HMG1 is passively released from necrotic cells, whereas apoptotic cells retain HMG1 in their nuclei (32). We tested whether HDGF was released under pathological conditions, using the Neuro2a cell line. More than 50% of the cells underwent necrosis or apoptosis after 24 h of treatment at 37°C. Western blot FIG. 1. HDGF protein is constantly expressed in brain. Aliquots of 10 g of protein from the brains of mice at different ages were electrophoresed and electrotransferred onto membrane filters. Western blot analysis was performed with anti-HDGF polyclonal antibody, and specific bands of ϳ40 kDa were detected. analysis showed that with a short exposure, HDGF was not detected in the conditioned medium from apoptotic cells, whereas it was detected in the conditioned medium from untreated cells and necrotic cells (Fig. 4A, a). With a longer exposure, although the band was faint, HDGF was detected in the conditioned medium from apoptotic cells (Fig. 4A, b). Immunohistochemical analyses revealed that HDGF was localized in the nuclei of untreated cells and apoptotic cells, whereas HDGF was found neither in necrotic nuclei nor cytoplasm (Fig.  4B). Thus, HDGF is tightly associated with the nucleus, even when the cells are induced to undergo apoptosis. By contrast, HDGF was dissociated from the nuclei of necrotic cells and released from these cells. The fact that a large amount of HDGF was released by necrotic cells suggests that nuclear HDGF is more abundant than the secreted HDGF. The concentration of HDGF was higher in the medium of untreated cells than that of apoptotic cells, indicating that apoptotic cells could not secrete HDGF actively. A small amount of HDGF was detected in the medium from apoptotic cells, suggesting that apoptotic cells also release a small amount of HDGF or some intact cells involved in apoptosis secrete HDGF.
HDGF Is a Neurotrophic Factor for Neurons-Many proliferating and/or differentiating factors have been reported to affect the survival of neurons (33)(34)(35). We examined the survival-promoting effect of HDGF using mouse primary hippocampal neurons. Mouse primary hippocampal neurons were prepared, and the medium was replaced with the serum-free defined medium without B27 (Invitrogen) in order to study neuronal survival. After confirmation that the optimal concentration of BDNF for neuronal survival was 10 ng/ml (data not shown), neurons were cultured for 5 days in the serum-free defined medium with HDGF (1-500 ng/ml) or BDNF (10 ng/ml) or medium alone. HDGF protein expression was observed in all neurons prepared from embryonic day 17 hippocampus (Fig.  5A). The HDGF concentration in the medium of the hippocampal neuronal culture without trophic factors was estimated to be less than 1.0 ng/ml by Western blot analysis using recombinant HDGF as a standard (data not shown). Cell viability and cell death were assessed using the MTT colorimetric assay and by measuring lactate dehydrogenase activity, respectively. Since the results of these two assays were reciprocal, only those of the MTT assay are shown in Fig. 5B. The absorbance values at 540 nm were 0.443 Ϯ 0.012 (control), 0.444 Ϯ 0.008 (HDGF, 1 ng/ml), 0.611 Ϯ 0.021 (HDGF, 10 ng/ml), 0.747 Ϯ 0.006 (HDGF, 100 ng/ml), 0.717 Ϯ 0.016 (HDGF, 500 ng/ml), and 0.780 Ϯ 0.015 (BDNF, 10 ng/ml). Then, the data were standardized to BDNF (100%) and no growth factors (0%). HDGF had a dose-dependent neurotrophic effect on neurons, although compared with BDNF, a well defined neurotrophic factor, a 10 times higher concentration of HDGF was required for similar activity.
These results combined with the data on the expression of HDGF in neurons and on secretion of HDGF from neurons suggest that HDGF is an autocrine neurotrophic factor.
Exogenously Supplied HDGF Partially Protected Endogenous HDGF-suppressed Neurons against Death-We tested the effect of endogenous HDGF suppression using siRNA technol- FIG. 3. HDGF and HMG1 are secreted. A, the purity of neurons was confirmed to be more than 90% by staining with anti-neuronal class III-␤-tubulin monoclonal antibody. B and C, the serum-free defined medium was prepared and conditioned as described under "Materials and Methods," and the 100-fold concentrated conditioned medium (CM) (15 l) and the cell extracts (1 g) from mouse primary hippocampal neurons (B) and from Neuro2a cells (C) were analyzed by Western blotting using anti-HDGF antibody and anti-HMG1 antibody. Anti-HDGF antibody and anti-HMG1 antibody showed specific bands of ϳ40 and ϳ29 kDa, respectively.

FIG. 4. HDGF is retained in nuclei of apoptotic cells, whereas HDGF is released from necrotic cells.
Necrosis was induced by treatment with 10 M ionomycin and 20 M carbonyl cyanide 3-chlorophenylhydrazone, or apoptosis was induced by treatment with 300 nM etoposide. A, the conditioned media from untreated cells, apoptotic cells, and necrotic cells were analyzed by Western blotting with anti-HDGF polyclonal antibody. With a short exposure, HDGF was not detected in the conditioned medium from apoptotic cells, whereas it was detected in the conditioned medium from untreated cells and necrotic cells (a). With a longer exposure, HDGF was significantly detected in the conditioned medium from apoptotic cells (b). B, double labeling for HDGF (green) and nuclei with propidium iodide (PI; red) showed that HDGF was localized to the nuclei of untreated cells and apoptotic cells, whereas HDGF was not found in the necrotic nucleus or cytoplasm. ogy (29). A plasmid, pSUPER-HDGF, was constructed to suppress endogenous HDGF expression. Primary hippocampal neurons were transiently transfected with either pEGFP-C1 alone, pEGFP-C1 and pSUPER, or pEGFP-C1 and pSUPER-HDGF. The plasmid pEGFP-C1 was used to mark the transfected cells as well as assess the transfection efficiency. The transfection efficiency was found to be ϳ10%. A reproducible reduction in HDGF expression to the background level was confirmed by immunohistochemistry for the neurons co-transfected with pEGFP-C1 and pSUPER-HDGF at 5 days after transfection (Fig. 6A). The time courses of the GFP-positive neurons for each condition are shown in Fig. 6B. In the control where pEGFP-C1 and pSUPER were co-transfected, GFP-positive neurons decreased to 57% after 5 days in culture, whereas in the endogenous HDGF-suppressed neurons, GFP-positive neurons decreased to 32% (Fig. 6C). These results indicate that nuclear HDGF might be important for survival.
Finally, using this paradigm, we examined whether exogenously supplied HDGF rescued endogenous HDGF-suppressed neurons from death. Recombinant HDGF was supplied exogenously to neurons co-transfected with pEGFP-C1 and pSUPER-HDGF. As shown in Fig. 6C, exogenously supplied HDGF partially prevented the cell death that occurred in endogenous HDGF-suppressed neurons. DISCUSSION HDGF has been intensively studied as a mitogenic factor (1, 10 -12). In the present study, we found another role for HDGF as a neurotrophic factor. Peculiar to HDGF among neurotrophic factors is that it is localized in the nucleus and that it has the potential to function in physiological and pathological conditions from the embryonic period through to adulthood.
These features are considered to be derived from the inherent domains of the amino acid sequence of HDGF. First, HDGF does not have a signal peptide-like hydrophobic region (1). Likewise, no signal peptide has been reported in the primary amino acid sequence of acidic or basic fibroblast growth factors (36). These molecules are thought to be secreted via a pathway other than the classical secretory pathway with a signal peptide. Other features are that HDGF possesses nuclear localization signals and heparin binding capacity. We speculate that HDGF protein synthesized in the cytoplasm preferentially moves to the nucleus due to its nuclear localization signals or is secreted via the nonclassical pathway and functions in an autocrine manner via a mechanism involving its heparin-binding capacity. The amount of HDGF in the medium and in the neurons in the hippocampal neuronal culture was estimated to be ϳ13 pg/24 h/10 5 cells and ϳ1.3 ng/10 5 cells, respectively.

FIG. 5. Effects of HDGF on the survival of mouse hippocampal neurons.
A, phase micrograph of the hippocampal neurons after 1 day of culture. All neurons were stained with anti-HDGF antibody. B, neurons were cultured for 5 days in the serum-free defined medium with HDGF (1-500 ng/ml) or with an optimal concentration of BDNF (10 ng/ml) or medium alone. The cell viability was assessed by an MTT colorimetric assay. Values (mean Ϯ S.E.; n ϭ 3) were normalized, with survival in the serum-free defined medium alone taken as 0% and that in 10 ng/ml BDNF taken as 100%.
FIG. 6. Exogenously supplied HDGF partially rescues endogenous HDGF-suppressed neurons. Primary hippocampal neurons were transfected with either pEGFP-C1 alone, pEGFP-C1 and pSU-PER, or pEGFP-C1 and pSUPER-HDGF. HDGF (100 ng/ml) was added at the time of transfection. A, a reduction in HDGF expression to the background level was confirmed by immunohistochemistry for the neurons co-transfected with pEGFP-C1 and pSUPER-HDGF. A representative neuron is shown. B, the time courses of GFP-positive neurons for each condition are shown. The percentage of GFP-positive neurons after transfection was calculated in relation to the number of GFP-positive neurons at 1 day after transfection defined as 100% for each condition. C, in the control where pEGFP-C1 and pSUPER were co-transfected, GFP-positive neurons decreased to 57% after 5 days in culture, whereas in the endogenous HDGF-suppressed neurons, GFP-positive neurons decreased to 32%. In the endogenous HDGF-suppressed neurons with exogenously supplied HDGF, GFP-positive cells decreased to 51%. Thus, exogenously supplied HDGF partially rescued endogenous HDGF-suppressed neurons.
Thus, approximately one-hundredth of the intracellular HDGF in the hippocampal neurons was secreted over 24 h in vitro. Although we have not examined the turnover of HDGF in vivo, these protein movements might be finely controlled under physiological conditions. Some growth factors including nerve growth factor, epidermal growth factor, and platelet-derived growth factor, have been reported to exert their effects by binding to their receptor, being internalized, and being conveyed to the nucleus associated with their receptor or even without a receptor by targeting to the nucleus (37)(38)(39). Exogenously supplied HDGF is reported to induce proliferation of fibroblasts (1), endothelial cells (10), vascular smooth muscle cells (11), and some hepatoma cells (12). Transfection of deletion mutants of HDGF clearly showed that nuclear targeting was required for mitogenesis (13,14). Exogenously supplied GFP-HDGF was reported to translocate to the nucleus in 293 cells (14). Therefore, extracellular HDGF is considered to enter into the nucleus and show mitogenic effects. In our experiments on neurons, we showed that exogenously supplied HDGF has a neurotrophic effect on hippocampal neurons (Fig. 5). The death of endogenous HDGF-suppressed neurons in Fig. 6 suggests that nuclear HDGF contributes to the survival. Besides, we showed that the reduction in the survival of endogenous HDGF-suppressed neurons was attenuated by exogenously supplied HDGF (100 ng/ml). As for neurons, the mechanism behind the effectiveness of exogenously supplied HDGF remains to be elucidated; however, our results support the notion that extracellular HDGF enters the nucleus and shows neurotrophic effects in neurons.
HDGF was shown here to share some characteristics with HMG1. Both molecules are localized to the nucleus, secreted under physiological conditions, and released from necrotic cells, while being retained within apoptotic nuclei (32). We demonstrated that HDGF has neurotrophic activity, whereas HMG1 has been reported to be an extracellular neurite growth factor (22). HMG1 enhances the activity of several transcription factors, including the glucocorticoid receptor, as well as the activity of RAG recombinase (19,40). One member of the HDGF family, LEDGF/p75/p52, was reported to function as a general transcriptional coactivator that enhanced activated transcription through direct interaction with the general transcription factor TFIIF and RNA polymerase II and also as a modulator of pre-mRNA splicing through interaction with alternative/essential splicing factor 2 (7,8). As for HDGF, although nuclear targeting of HDGF has been reported to be required for mitogenesis (13,14) and we showed that HDGF in the nucleus contributed to the survival of neurons, the molecules controlled by HDGF in the nucleus have not been identified. Whether HDGF in the nucleus functions in the regulation of general transcription or specific transcription and whether HDGF modulates pre-mRNA splicing are yet to be explored.
An extracellular role of HMG1 in neurite outgrowth is exerted via its binding to the receptor for advanced glycation end products (RAGE), a multiligand transmembrane receptor belonging to the immunoglobulin superfamily (41). The homology between HDGF and HMG1 implies that RAGE is a reasonable candidate receptor for HDGF. However, the recent report that a COOH-terminal motif in HMG1 is responsible for RAGE binding (42) does not support this possibility, because this motif is absent in HDGF.
Since HDGF is mitogenic to many kinds of cells (1, 10 -12), it is possible that HDGF has mitogenic and/or inflammatory effects on glial cells, which surround neurons and communicate with them. We showed that although some glial cells were stained with anti-HDGF antibody, most of them were HDGF-negative. HDGF mRNA expression was reported to be upregulated in astrocytes after exposure to neural cell adhesion molecule (43). Thus, glial cells may be induced to express HDGF when activated and/or stimulated by certain signals and then might allow HDGF to work as a mitogenic factor on glial cells or as a neurotrophic factor on neighboring neurons.
We propose that HDGF is a novel type of neurotrophic factor based on the fact that it is localized to the nucleus and has the potential to function in an autocrine manner in physiological and pathological conditions. Numerous lines of evidence have established that neurotrophic factors govern neuronal development and survival. More recent evidence indicates that some neurotrophic factors are involved in synaptic modification, neurotransmitter release, and long term potentiation (44 -48). Among neurotrophic factors, BDNF has been intensively studied and has been reported to modulate hippocampal plasticity and memory in cell models and in animals (49 -51). Furthermore, the literature on the subject of neurotrophic factors and neurodegenerative diseases, including Alzheimer's disease and Parkinson's disease, is expanding prodigiously (52). Further investigations on HDGF in the nervous system may throw light on the physiological mechanism of aging and memory and its relationship with the pathogenesis of neurodegenerative diseases.