This article is mentioned in:

Introduction

During embryonic and early postnatal development, the axotomy of motorneurons or removal of their target results in significant motorneuron cell loss. In adults however, axotomy can result in either complete motorneuron survival or motorneuron death. For patients this implies an incomplete or even complete loss of function of the muscular targets of the lost motorneurons (1).

It is well-established that injuries of the spinal cord (2,3) as well as the peripheral nerves (4–6) lead to changes in gene and protein expression levels in motorneurons and glial cells, which may result in neuronal apoptosis. This is the basis for therapeutic strategies that aim to enhance axonal regeneration and functional recovery following peripheral nerve injury, including pharmacological treatments (7).

In this physiological context, minocycline has been widely used, but the advantages and disadvantages of this treatment appear equal in number. Minocycline is a semi-synthetic second generation tetracycline with broad spectrum anti-microbial activity (8). The primary applications of minocycline include treatment of pneumonia, rheumatoid arthritis, acne and infections of the skin, the genital, and urinary systems (9). There are also promising preclinical studies for the treatment of stroke (10,11), Alzheimer's disease (12), Huntington's disease (13), Parkinson's disease (14), amyotrophic lateral sclerosis (15), multiple sclerosis (15,16) and traumatic brain injury (17). Clinical trials with minocycline for the treatment of spinal cord injury have been underway since the early 2000s (18). The predominant effect of minocycline is associated with its ability to modulate microglia and immune cell activation and to reduce apoptosis (19). There have however been reports of conflicting results, with a number of previous studies demonstrating that minocycline worsened spinal cord and brain injuries (20–23).

Our previous investigations demonstrated that minocycline impairs motorneuron survival in organotypic rat spinal cord cultures (24) and inhibited the regeneration of peripheral nerves (25). The present study was undertaken to examine the effects of minocycline on the expression of selected transcriptional and translational profiles in the rat spinal cord following sciatic nerve transection and microsurgical coaptation. In addition to the spinal cord in vivo, the present study conducted in vitro experiments using NSC-34 motorneuron-like cells. NSC-34 is a hybrid cell line produced by the fusion of neuroblastoma with mouse motorneuron-enriched primary spinal cord cells (26). These cells share numerous morphological and physiological characteristics with mature primary motorneurons, and thus are an accepted model for studying the pathophysiology of motorneurons (26). Stress was induced by oxygen glucose deprivation (OGD) or lipopolysaccharide (LPS) treatment. The mRNA and protein expression levels of the following compounds were examined: i) B cell lymphoma 2 (Bcl-2)-associated X protein (Bax), which has been demonstrated to be upregulated in the spinal motorneurons of newborn rats following sciatic nerve injury (27) and in adult cats following partial dorsal root ganglion ectomy (28); ii) caspase-3, which is activated in adult spinal motorneurons during injury-induced apoptosis (29); iii) Bcl-2, which has been reported to be activated in the adult spinal motorneurons of rats in the first three weeks following sciatic nerve injury (30); iv) major histocompatibility complex of class I (MHC I), which is upregulated in the spinal motorneurons of neonatal rats following sciatic nerve injury (31); v) tumor necrosis factor (TNF-α), released from astrocytes and microglia around motorneurons in rat spinal cord in the first two weeks following sciatic nerve crush (32); vi) activating transcription factor (ATF3), which is a marker for regenerative response following nerve root injury (33), and its expression in neurons is closely associated with their survival and the regeneration of their axons following axotomy (34); vii) vascular endothelial growth factor (VEGF), which has been demonstrated to be upregulated in the spinal motorneurons of adult rats in response to neurotomy (35); viii) matrix metalloproteinase 9 (MMP9), immediately upregulated in adult mice spinal motorneurons following nerve injury (36); and ix) growth-associated protein 43 (GAP-43), which is expressed at high levels during development (37) and stressed by nerve injury adult motorneurons (38).

Materials and methods

Ethical approval

The present study was conducted in accordance with the European Commission regulations and those of the National Act on the Use of Experimental Animals of Germany, and adhered to the guidelines of the Committee for Research and Ethical Issues of the International Association for the Study of Pain.

Animal model

A total of 51 female Wistar rats (10 weeks old, 200–230 g, strain-matched, inbred) were obtained from Harlan-Winkelmann GmbH (Borchen, Germany). The rats were housed under controlled laboratory conditions with a 12-h light/dark cycle (lights on at 6 am) at 20±2°C with an air humidity of 55–60%. The animals were provided with ad libitum access to commercial rat pellets (Altromin 1324™; Altromin Spezialfutter GmbH & Co. KG, Lage, Germany) and tap water. Following intervention the rats were housed in pairs in Makrolon IIL cages (Bioscape GmbH, Castrop-Rauxel, Germany). Every effort was made to minimize the amount of suffering and the number of animals used in the experiments.

A total of 46 rats were injured and divided into four phosphate-buffered saline (PBS; Sigma-Aldrich Chemie GmbH, Munich, Germany) and four minocycline treatment groups with survival times of 3, 5, 7 and 14 days post-intervention (DPI), with five animals/group for semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR). An additional three animals from the 7-day PBS-treated and from the the 7-day minocycline-treated groups were used for immunohistochemical analysis. For semi-quantitative RT-PCR the spinal cords of five untreated animals were also prepared.

Minocycline hydrochloride (Sigma-Aldrich, St. Louis, MO, USA) was administered once daily for ≥7 consecutive days by intraperitoneal injection at a dosage of 50 mg/kg body weight (~10 times the usual human dose), starting at 30 min following nerve reconstruction. The drug was dissolved in saline (pH 7.2, freshly prepared daily) at 37°C. A dosage of >20 mg/kg was selected to induce the maximal anti-hyperalgesic effect, as lower doses are unable to affect gene expression in a sufficiently stable manner (39). Control rats were injected with PBS (pH 7.2) using an identical treatment regime.

The surgical procedure protocol for nerve reconstruction was the same for all groups, and consisted of exposing the right sciatic nerve through a dorsal incision under general anesthesia (60 mg/kg pentobarbital, intraperitoneal; Sigma-Aldrich) and aseptic conditions using an SV8 operating microscope (Zeiss GmbH, Jena, Germany). The nerve was transected at the proximal origin of the gracilis muscle and immediately microsurgically coaptated with respect to intraneuronal topography using epineural sutures (Ethilon 11×0; Johnson & Johnson, New Brunswick, NJ, USA) followed by closure of the dorsal incision.

Following the respective survival times (3, 5, 7 and 14 days), the animals were sacrificed by an excess of anesthesia (pentobarbital) via intraperitoneal injection. L3-L6 sections of the spinal cord, divided into ipsilateral and contralateral sites were harvested and homogenized in peqGOLD TriFast total RNA isolation reagent (cat. no. 30–2030; PeqLab Biotechnologie GmbH, Erlangen, Germany) using an Ultra-Turrax Homogenizer (IKA® Werke GmbH & Co. KG, Staufen im Breisgau, Germany). Total RNA was prepared according to the manufacturer's instructions. Potentially contaminating DNA was removed by treating 5 µg total cell RNA with Turbo DNA-free (Ambion; Thermo Fisher Scientific, Inc., Waltham, MA, USA). RNA (4 µl; 2 µg input RNA) was reverse transcribed using a RevertAid™ H Minus First Strand cDNA Synthesis kit primed with Oligo(dT)18 primers (cat. no. K1631; Thermo Fisher Scientific, Inc.; primers listed in Table I). cDNA (1 µl) was then amplified by PCR using Taq DNA polymerase (PeqLab Biotechnologie GmbH), as previously described (40). One-tenth of each reaction product was electrophoresed on a 1% agarose gel (Serva Electrophoresis GmbH, Heidelberg, Germany) (excluding TNF-α, which required a 2% agarose gel). The PCR product bands were quantified by densitometric analysis using a GeneGenius bio-imaging system (Syngene, Cambridge, UK) and the ratio of their expression levels to those of the GAPDH reference gene were calculated. Each experiment was repeated in triplicate.

Table I. Sequences of primers used for semi-quantitative reverse transcription-polymerase chain reaction.

Table I.

Sequences of primers used for semi-quantitative reverse transcription-polymerase chain reaction.

Gene	Sequence	Primer location	Product size (base pairs)	Cycle no.	Annealing temperature (°C)	Accession number
ATF3	5′-TTCCGAGAGTTTGGGGGTCTGCC-3′	1,285–1,307	320	36	60	NM 012912
	5′-CCAGTCTCCACGGGCTGTGGTT-3′	1,604–1,583
Bax	5′-GGATGGTTGCTGATGTGGATAC-3′	86–107	342	40	58	XM_001081479
	5′-CCATCTTCTTCCAGATGGTGAG-3′	427–406
Bcl-2	5′-CGACTTTGCAGAGATGTCCAG-3′	555–575	392	40	58	NM_016993
	5′-CTCACTTGTGGCCCAGGTATG-3′	946–926
Caspase-3	5′-CCTCAGAGAGACATTCATGGC-3′	275–295	247	40	56	NM_012922
	5′-TCGGCTTTCCAGTCAGACTC-3′	522–503
GAP-43	5′-AGCGCAGCCTCCAACGGAGAC-3′	759–779	247	36	60	NM 017195
	5′-GCTCACACACGTGAGCAGGACA-3′	1,005–984
MHC I	5′-GGCTCACACTCGCTGCGGTATTT-3′	82–104	626	36	60	M 31018
	5′-TAGAAGCCCAGGGCCCAGCA-3′	707–688
MMP9	5′-AGTTTGGTGTCGCGGAGCAC-3′	509–528	754	40	60	NM_031055
	5′-TACATGAGCGCTTCCGGCAC-3′	1,262–1,243
TNF-α	5′-CCCAGACCCTCACACTCAGAT-3′	389–413	214	38	56	NM 012675
	5′-TTTTCCCTTGAAGAGAACCTG-3′	563–541
VEGF	5′-GAAGTTCATGGACGTCTACC-3′	121–139	237	40	55	NM_031836
	5′-CATCTCTCCTATGTGCTGGC-3′	357–338
GAPDH	5′-TTAGCACCCCTGGCCAAGG-3′	802–820	531	24	55	XM_228411
	5′-CTTACTCCTTGGAGGCCATG-3′	1,332–1,314

[i] ATF3, activating transcription factor 3; Bax, Bcl-2-associated X protein; Bcl-2, B cell lymphoma 2; GAP-43, growth associated protein-43; MHC I, major histocompatibility complex I; MMP9, matrix metalloproteinase 9; TNF-α, tumor necrosis factor-α; VEGF, vascular endothelial growth factor; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Dunn's multiple comparison test was used as a post-hoc test. For statistical analysis of the groups within one survival time, analysis of variance with Tukey's post-hoc test was performed. Graph Pad Prism 4 software (GraphPad Software Inc., La Jolla, CA, USA) was used to conduct the statistical analyses. P<0.05 was considered to indicate a statistically significant result.

At 5 DPI, anesthetized rats administered an excess of intraperitoneal pentobarbital were transcardially perfused with 4% paraformaldehyde (PFA; Sigma-Aldrich). L3-L6 sections of the spinal cord were removed and postfixed for 24 h, cryo-protected in 30% sucrose (in 0.4% buffered PFA) for 24 h, rapidly frozen, and sectioned using a cryostat (Jung Frigocut 2800 E; Leica Microsystems GmbH, Wetzlar, Germany; 20 µm). The tissue sections were subsequently immunostained for mouse monoclonal anti-pan-neuronal neurofilament marker (neuronal marker; pan-NF SMI311; non-phosphoneurofilament specific; 1:1,000; Covance Inc., Princeton, NJ, USA) combined with the following antibodies: i) Rabbit polyclonal anti-glial fibrillary acidic protein (GFAP; astroglial marker; 1:1,000; cat. no. 10555; Progen Biotechnik GmbH, Heidelberg, Germany); ii) goat polyclonal anti-ionized calcium binding adaptor molecule 1 (IBA1; microglia marker; 1:1,000; cat. no. ab5076; Abcam, Cambridge, UK); iii) rabbit monoclonal anti-Bax (1:200; cat. no. ab32503; Abcam); iv) rabbit polyclonal anti-caspase-3 (1:100; cat. no. AB3623; EMD Millipore, Billerica, MA, USA); v) rabbit polyclonal anti-Bcl-2 (1:1,000; cat. no. AB1722; EMD Millipore); vi) rat monoclonal anti-MHC I (1:100; cat. no. ab15680; Abcam); vii) rabbit polyclonal anti-TNF-α (1:500; cat. no. ab9755; Abcam), and viii) rabbit polyclonal anti-MMP9 (1:100; cat. no. ab7299; Abcam) or ix) rabbit polyclonal anti-GAP-43 (1:500; cat. no. AB5220;EMD Millipore). Co-staining of mouse monoclonal anti-VEGF (1:500; cat. no. 05–1116; EMD Millipore) and mouse monoclonal anti-ATF3 (1:200; cat. no. ab191513; Abcam) was performed with rabbit polyclonal anti-β-III-tubulin (1:1,000; cat. no. 802001; Biolegend, San Diego, CA, USA) as a neuronal marker. All antibodies were diluted in 1% normal goat serum and 0.3% Triton X-100 (Sigma-Aldrich) in PBS. The tissue sections were incubated overnight at 7°C. PBS washing was conducted prior to secondary antibody incubation for 3 h with goat anti-mouse Alexa 488 (1:500; cat. no. A-11001; Invitrogen; Thermo Fisher Scientific, Inc.) and donkey anti-rabbit Cy3 (1:250; cat. no. 711-165-152; Dianova Vertriebs-Gesellschaft mbH, Hamburg, Germany) diluted in 1% normal goat serum and 0.3% Triton in PBS. Cryosections were mounted on slides, embedded with Immu-Mount (Thermo Fisher Scientific, Inc.) and examined under a fluorescent AxioImager M1 microscope (Zeiss GmbH, Jena, Germany) with rhodamine and fluorescein isothiocyanate filters and a Plan-Neofluar objective (20/0.5).

Cell culture model

NSC-34 cells (cat. no. CLU140; Cedarlane, Burlington, Canada) were stored at −80°C in cryo tubes. In preparation for experiments, 106 cells were pre-cultured in 15 ml pyruvate-free Dulbecco's modified Eagle's Medium (DMEM) supplemented with 4.5 g/l glucose, 10% fetal calf serum (FCS) and 0.2% Ciprobay (all Gibco®; Thermo Fisher Scientific, Inc.) for 7 days in vitro (DIV) in 75-cm2 flasks at 37°C in a humidified atmosphere containing 5% CO2 (herein referred to as ‘normal conditions’). Thereafter, the cells were harvested by scraping from the bottom, centrifuged for 10 min at 360 × g at room temperature, resuspended in 10 ml DMEM (constituents as above) and seeded (0 DIV) in 96-well plates (1×104 cells/100 µl DMEM/well) for assessment of cell proliferation/survival by MTT assay, and in 25-cm2 flasks (2×105 cells/5 ml DMEM/flask) for semi-quantitative RT-PCR, or in φ 35-mm culture dishes (5×104 cells/2 ml DMEM/dish) for immunohistochemistry.

OGD was induced following 4 DIV. Briefly, the medium was removed and replaced with normal medium under normal conditions or OGD medium (glucose-free DMEM supplemented with 10% FCS and 0.2% Ciprobay) under OGD conditions. OGD conditions were reached by exposing the cultures to an atmosphere composed of 5% CO2 and 1% O2, using nitrogen gas to displace ambient air in a C200 incubator (Labotect Technik-Göttingen GmbH, Rosdorf, Germany) at 37°C for 6 h. For reoxygenation the incubator atmosphere was reestablished to 5% CO2 and 21% O2 and 4.5 mg/ml glucose was added.

Addition of LPS (Escherichia coli; Sigma-Aldrich) was also performed at 4 DIV. Regarding OGD, the medium was replaced with normal medium supplemented with 2 mg/ml LPS for 24 h.

Minocycline hydrochloride (molecular weight, 493.9 g/mol; Sigma-Aldrich) was dissolved in sterile PBS to obtain a stock solution of 5 mg/ml (pH 6.5). From this stock solution, 1 µl/well, 20 µl/dish and 50 µl/flask were added to the respective groups (control, OGD, LPS) following medium replacement in order to start the stress induction (final minocycline concentration 100 µM, final pH 7.0).

The specific turnover of MTT (6 mg/ml; Sigma-Aldrich) to formazan by viable cells was analyzed 24 h following OGD induction using photometry. Briefly, 8 µl MTT (6 mg/ml) was added to each well and incubated for 3 h prior to complete removal of the medium. A total of 100 µl dimethyl sulfoxide (DMSO; Merck Millipore, Darmstadt, Germany) was then added to each well, and extinction coefficients in each well were determined using an Infinite® M200 (Tecan GmbH, Crailsheim, Germany) and calculated by subtracting the reference absorbance at 690 nm from the absorbance at 570 nm. The absorbance of the empty wells filled only with DMSO was subtracted. Subsequently, the mean values of the respective treatment groups were calculated and associated with the norm medium control. Each experiment was performed with 12 repeats/treatment group.

As the MTT assay is a general assay for cell viability and proliferation, the mitotic indices were additionally determined using BrdU (Roche Diagnostics GmbH, Mannheim, Germany), which was added at the same time as stress induction, and 24 h prior to fixation with 4% PFA, as previously described (41). Fixed cell cultures were washed with PBS, incubated with 2 N HCl at 37°C for 1 h, washed repeatedly with borate buffer (pH 8.5) and PBS, and finally incubated at 7°C for 24 h with monoclonal rat anti-BrdU antibody (1:100; cat. no. OBT0030; AbD Serotec; Bio-Rad Laboratories, Inc., Hercules, CA, USA) combined with mouse monoclonal anti-pan-NF (cat. no. 837802; Biolegend). Subsequently, the washed cultures were then incubated for 3 h with secondary antibodies goat anti-rat Alexa 546 (1:500; cat. no. A11081; Thermo Fisher Scientific) and anti-mouse Alexa 488 (1:500; Invitrogen; Thermo Fisher Scientific, Inc.) prior to examination using an AxioImager M1 fluorescence microscope with a 20× objective lens. Each treatment group consisted of three culture dishes in which the BrdU-positive NSC-34 cells in three different fields of view were counted. The three values/dish were combined, and the percentage of BrdU-positive cells relative to the total number of NSC-34 cells was calculated.

Cell viability was assessed by double-labeling with fluorescein diacetate and propidium iodide (PI) (42). The assay is based on the ability of living cells to hydrolyze fluorescein diacetate (10 µg/ml PBS, 5 min; Sigma-Aldrich) using intracellular esterases, resulting in a green/yellow-colored fluorescence. Dead cells were labeled with PI (5 µg/ml PBS, 5 min; Sigma-Aldrich), which interacts with DNA to produce a red fluorescence of cell nuclei. The analysis procedure was the same as described above for the BrdU assay.

For all assays, the respective mean values were analyzed using a non-parametric Kruskal-Wallis test and Dunn's multiple comparison test as post-hoc tests using Graph Pad Prism 4 software. P≤0.05 was considered to indicate a statistically significant result. All experiments were independently performed in triplicate.

Cells were harvested 24 h prior to OGD induction or LPS treatment. The mRNA expression levels were determined as described above for the spinal cord tissue sections. Five flasks were prepared for each treatment group (control, control + minocycline, OGD, OGD + minocycline, LPS and LPS + minocycline). The experiment was repeated in duplicate. Statistical analysis was performed with a non-parametric Kruskal-Wallis test and Dunn's multiple comparison test as post-hoc test using Graph Pad Prism 4 software. P<0.05 was considered to indicate a statistically significant result.

At 5 DIV (24 h after OGD induction), the cell cultures were fixed for 30 min in 4% buffered PFA, and unspecific binding sites were blocked with 10% bovine serum albumin (BSA; Sigma-Aldrich)/0.3% Triton X-100 in PBS for 1 h. Subsequently, the cultures were incubated with the aforementioned primary antibodies: Anti-Bax (1:200), anti-caspase-3 (1:100), anti-Bcl-2 (1:1,000), anti-MHC I (1:100), anti-TNF-α (1:500), anti-MMP9 (1:100), anti-GAP-43 (1:500), anti-VEGF (1:500), and anti-ATF3 (1:200) co-stained with mouse monoclonal pan-NF (1:1,000) or with rabbit polyclonal anti-β-III-tubulin (1:1,000) at 7°C overnight, followed by a wash with PBS, then incubation with the secondary antibodies goat anti-mouse Alexa 488 (1:500) and donkey anti-rabbit Cy3 (1:250) at room temperature for 3 h. All antibodies were diluted in 10% BSA/0.3% Triton in PBS. The specificity of the immunoreaction was controlled by the application of buffer instead of primary antibodies. Cell cultures were examined using a fluorescence microscope (AxioImager M1; Plan-Neofluar objective; 20/0.5). For each treatment group (control, control + minocycline, OGD, OGD + minocycline, LPS, LPS + minocycline) and staining type two dishes were examined, in total 108 dishes. The experiment was performed in duplicate.

Results

Animal model

The surgical procedure was well tolerated and wounds healed well. The treatments were fatal to none of the animals. In the first two post-operative weeks, clinical signs of hyperalgesia or discomfort were observed. Compared with injured PBS-treated rats, the minocycline-treated animals demonstrated diminished peripheral nerve regeneration, as indicated by significantly lower axon counts in the distal stump when compared with PBS-treated animals. Functional outcome (response of animals to thermal stimuli and muscle weight ratio of the gastrocnemius muscle) has been previously described (25).

Immunohistochemical assessment was performed at 5 DPI as at this DPI the PCR revealed the most marked alterations (see below). At 5 DPI, the population of SMI311-expressing motorneurons of the contralateral and ipsilateral ventral horns (VH) was equal in number, form and staining intensity (Fig. 1). Astroglia-specific GFAP (Fig. 1A) and microglia-specific IBA1 (Fig. 1B) were expressed in the contralateral VH. In the ipsilateral side, a marked induction of both markers was evident (Fig. 1A and B). Microglia activation could also be demonstrated by cell morphology. The cells were altered from their ramified form, and became thicker and retracted their branches. Glia activation indicated ongoing neurodegenerative processes at the nerve fiber level, which was not yet evident from SMI311 staining. With regards to GFAP, treatment with minocycline was ineffective (Fig. 1A). However, microglia activity was decreased by minocycline, and this effect was more marked in the ipsilateral VH (Fig. 1B).

Figure 1. Representative fluorescence images of rat L4 and L5 ventral spinal cord sections at day 5 post-intervention. SMI311 immunofluorescence (green) shows a normal pattern of motorneurons in the contralateral and ipsilateral ventral horns. (A) GFAP-expressing astroglia (red). (B) IBA1-positive microglia cells (red). Ø mino, not treated with minocycline; + mino, treated with minocycline; GFAP, glial fibrillary acidic protein; IBA1, ionized calcium binding adaptor molecule 1.

In the spinal cord of untreated animals, all experimental genes were constitutively expressed. GAP-43 possessed the highest expression levels, which were increased at ≥7 DPI. The expression levels of all other genes were significantly increased by sciatic nerve injury at 3 DPI. This increase in expression levels was evident >5 DPI. In the case of MMP9 a marked increase in expression levels was observed at 3 and 5 DPI. At 7 DPI, the expression levels of caspase-3, Bcl-2, ATF3, TNF-α and VEGF returned to levels similar to those of the control, and the expression levels of VEGF were already significantly reduced at 5 DPI. The expression levels of Bax, MHC I and MMP9 remained high at ≥14 DPI. With the exception of MMP9, no significant differences were observed between ipsilateral and contralateral effects following nerve injury. At 5 DPI, MMP9 was expressed at significantly higher levels on the ipsilateral side (Fig. 2).

Figure 2. Semi-quantitative reverse transcription-polymerase chain reaction analysis of the mRNA expression levels of various genes in rat L3-L6 spinal cord sections in the naive control, contralateral (Ø and + mino) and ipsilateral (Ø and + mino) groups. *,#,+P<0.05; **,++P<0.01; ***,###P<0.001. Bax, Bcl-2-associated X protein; MHC I, major histocompatibility complex I; Bcl-2, B cell lymphoma 2; ATF3, activating transcription factor 3; TNF-α, tumor necrosis factor-α; MMP9, matrix metalloproteinase 9; VEGF, vascular endothelial growth factor; GAP-43, growth associated protein-43; D, days; Ø mino, not treated with minocycline; + mino, treated with minocycline.

Only the expression levels of certain genes were affected by minocycline. Treatment with minocycline reduced the ipsilateral expression levels of MHC I at 3 DPI. The expression levels of TNF-α were reduced ipsilaterally at 3 and 5 DPI. At 5 DPI, the contralateral expression levels of TNF-α were also diminished. Minocycline reduced the ipsilateral expression levels of MMP9 at 3 DPI and the contralateral expression levels at 5 DPI. Treatment with minocycline decreased the ipsilateral and contralateral expression levels of VEGF at 3 DPI. Furthermore, the nerve injury-induced expression of GAP-43 was significantly suppressed (Fig. 2).

The protein expression levels of the genes examined by PCR were evaluated by fluorescence immunohistochemistry at 5 DPI (Fig. 3). For caspase-3, no fluorescence signal was detected in the contralateral VH (Fig. 3A). Bax, Bcl-2, TNF-α, MHC I, MMP9 (Fig. 3A), ATF3 and VEGF (Fig. 3B) were expressed in motorneurons of the contralateral VH, although at relatively low levels. Only GAP-43 was markedly expressed, indicating synaptic contacts (Fig. 3A).

Figure 3. Representative fluorescence images of rat L4 and L5 ventral spinal cord sections at 5 DPI. (A) Tissue sections were double-immunostained with SMI311 (green staining, neuron-specific) and various antibodies (mentioned at the left, red staining), resulting in a combined yellow immunosignal. (B) Tissue sections were double-immunostained with β-III-tubulin (red staining, neuron-specific) and ATF3 or VEGF (green staining), resulting in a combined yellow immunosignal. Arrows indicate nuclear expression, arrowheads indicate cytoplasmic expression. Ø mino, not treated with minocycline; + mino, treated with minocycline; Bax, Bcl-2-associated X protein; Bcl-2, B cell lymphoma 2; TNF-α, tumor necrosis factor-α; MHC I, major histocompatibility complex I; MMP9, matrix metalloproteinase 9; GAP-43, growth associated protein-43; DPI, days post-intervention; ATF3, activating transcription factor 3; VEGF, vascular endothelial growth factor.

The majority of immunofluorescence signals were activated by unilateral nerve injury in the ipsilateral VH. In the case of Bax, in addition to marked cytoplasmic staining of motorneurons, marked nuclear fluorescence was visible (Fig. 3A). Such injury/hypoxia-induced translocation of Bax to the nucleus has previously been described for neonatal neurons of the spinal cord (27) and brain (43). The enhanced motorneuronal signals of caspase-3 and Bcl-2 (Fig. 3A) were located in the cytoplasm, and Bcl-2 was also markedly expressed in the contralateral VH motorneurons (Fig. 3A). Intense TNF-α staining was observed in the motorneuronal cytoplasm with compaction/concentration around and inside the nucleus (Fig. 3A). The markedly intense immunosignals of MHC I, MMP9 (Fig. 3A) and VEGF (Fig. 3B) were evenly distributed in the cytoplasm of the motorneurons. In addition, the expression of the ATF3 transcription factor was predominantly upregulated in the cytoplasm (Fig. 3B). A similar pattern with the majority of neurons exhibiting marked cytoplasmic staining and only a minority also exhibiting nuclear translocation was reported by Seijffers et al (44). GAP-43 immunostaining demonstrated a reduction in synaptic contacts (Fig. 3A).

The effect of minocycline was marginal. MHC I appeared to be upregulated by minocycline in the contralateral and ipsilateral VH motorneurons (Fig. 3A). Conversely, the injury-induced upregulation of VEGF was reversed by minocycline (Fig. 3B).

Cell culture model

MTT, bromodeoxyuridine (BrdU) and vital staining were used to assess cellular survival and proliferation (Figs. 4 and 5). The MTT assay is based on the specific turnover of MTT to formazan, requiring viable cells. An increased extinction coefficient indicates an enhanced MTT turnover rate and thus a greater number of viable cells. The MTT assay demonstrated a significant (P<0.05) OGD-induced reduction of the metabolic activity of NSC-34 cells, whereas LPS had no effect on metabolic activity levels (Fig. 5A). Independent of treatment, minocycline did not alter the treatment group-specific MTT turnover (Fig. 5A).

Figure 4. (A) Assessment of NSC-34 cell integrity by double-labeling with PI (red staining) and FD (green staining). (B) Assessment of NSC-34 cell mitotic index by BrdU incorporation and SMI311 (green staining, neuron-specific). Ø mino, not treated with minocycline; + mino, treated with minocycline; OGD, oxygen glucose deprivation; LPS, lipopolysaccharide; PI, propidium iodide; FD, fluorescein diacetate; BrdU, bromodeoxyuridine.

Figure 5. Respective quantitative analysis of (A) MTT, (B) BRdU and (C) PI assays. Data are presented as the mean ± standard deviation; n=9 cultures/group; for each culture, a mean of three vision fields/dish was used for calculations. *P<0.05; **P<0.01; ***P<0.001. Con, control; OGD, oxygen glucose deprivation; LPS, lipopolysaccharide; PI, propidium iodide; BrdU, bromodeoxyuridine; Ø mino, not treated with minocycline; + mino, treated with minocycline.

In contrast to post-mitotic motorneurons, NSC-34 cells are able to proliferate as they are neuroblastoma-spinal cord hybrids. The basic mitotic index, determined by BrdU incorporation in the control group, was 48±5% (Figs. 4B and 5B). OGD significantly reduced the proliferation of NSC-34 cells (P<0.01); however, LPS was ineffective (Figs. 4B and 5B). Minocycline had no effect on NSC-34 cell proliferation, in either the control or the LPS group. In addition, minocycline was not able to reverse the inhibitory effect of OGD (Figs. 4B and 5B).

Vital staining of control cultures revealed only scattered dead (PI-positive) cells, which were not affected by minocycline (Figs. 4A and 5C). OGD induced significant neurotoxicity (P<0.001), while minocycline was marginally able to reverse this neurotoxicity (P<0.1; Fig. 4A and 5C). LPS alone or in combination with minocycline had no effect on NSC-34 cell viability (Fig. 4A and 5C).

In untreated control NSC-34 cell cultures, all experimental genes were constitutively expressed, with low expression levels observed for ATF3, caspase-3 and VEGF. OGD was able to significantly increase the expression levels of Bcl-2 (P<0.05), TNF-α (P<0.01) and MMP9 (P<0.05). TNF-α and MMP9 were also significantly upregulated by LPS (P<0.05). Similarly to the results observed following the analysis of the tissue samples, the OGD stress-induced expression was significantly (P<0.05) suppressed by minocycline. Furthermore, minocycline also significantly reduced VEGF expression (P<0.05; Fig. 6).

Figure 6. Semi-quantitative reverse transcription-polymerase chain reaction analysis of mRNA expression levels in NSC-34 cells in the con, OGD or LPS groups treated with or without minocycline. Data are presented as means ± standard deviation; n=10 flask/group. *P<0.05; **P<0.01. Bax, Bcl-2-associated X protein; Con, control group; OGD, oral glucose deprivation; LPS, lipopolysaccharide; MHC I, major histocompatibility complex I; Bcl-2, B cell lymphoma 2; ATF3, activating transcription factor 3; TNF-α, tumor necrosis factor-α; MMP9, matrix metalloproteinase 9; VEGF, vascular endothelial growth factor; GAP-43, growth associated protein-43.

The mRNA expression profile of NSC-34 cells was also investigated by fluorescence immunohistochemical evaluation of protein expression (Fig. 7). In untreated control cell cultures (Fig. 7A and B), all proteins were expressed endogenously. This result was expected as primary cell cultures are usually characterized by a low and constant cell death rate. The stressors OGD (Fig. 7A and B) and LPS (Fig. 7A and B) induced the activation of all proteins. Minocycline had no effect on control cultures (Fig. 7A and B), but was able to reduce the stress-induced upregulation of TNF-α and MMP9 (Fig. 7A) expression. Combined with LPS, minocycline was able to inhibit VEGF expression (Fig. 7B). Conversely to in vivo experiments, the expression of Bax was predominantly located in the cytoplasm (Fig. 7A). Only in stressed and minocycline-treated cells was nuclear expression visible (Fig. 7A). The expression of transcription factor ATF3 was upregulated following stress and located in the nucleus (Fig. 7B), and this expression was inhibited by minocycline (Fig. 7B). Furthermore, the expression pattern of GAP-43 was different to those observed in vivo. Under all experimental conditions, fluorescence signals were present in the cytoplasm of NSC-34 cells, although not in the fibers (Fig. 7A).

Figure 7. Representative fluorescence images of NSC-34 cells under various experimental conditions. (A) Cell cultures were double-immunostained with SMI311 (green staining) and various antibodies (red staining), often resulting in a combined yellow immunosignal. (B) Cell cultures were double-immunostained with β-III-tubulin (red staining) and ATF3 or VEGF (green staining), resulting in a combined yellow immunosignal. OGD, oxygen glucose deprivation; LPS, lipopolysaccharide; Ø mino, not treated with minocycline; + mino, treated with minocycline; Bax, Bcl-2-associated X protein; Bcl-2, B cell lymphoma 2; TNF-α, tumor necrosis factor-α; MHC I, major histocompatibility complex I; MMP9, matrix metalloproteinase 9; GAP-43, growth associated protein-43; ATF3, activating transcription factor 3; VEGF, vascular endothelial growth factor.

Discussion

Transection of peripheral nerves induces a complex cascade of reactions, including retrograde processes targeting the axotomized spinal motorneurons. In neonatal rats, axotomized motorneurons often die (45,46). However, in adult animals, degeneration of spinal motorneurons following peripheral nerve axotomy rarely occurs (47,48). Only severe nerve ventral root avulsion has more severe effects and induces a significant loss of axotomized motorneurons in the respective spinal cord segments (49). In addition, less severely injured adult ipsilateral motorneurons develop classical post-traumatic signs, including synaptic terminal retraction (48). Furthermore, microglia activation is regularly observed (50,51).

In the rat model of sciatic nerve reconstruction in the present study, a post-traumatic pattern in the VH was also observed. Motorneurons did not exhibit visible signs of neurodegeneration. Aside from that the sciatic nerve was reconstructed immediately following transection, which is neuroprotective (52). However, microglia activation and synaptic terminal retraction were detectable, and mRNA expression levels correlated with microglial activation. In the untreated spinal cord, all experimental genes were constitutively expressed at mRNA and protein levels. GAP-43, as a crucial component of the axon and presynaptic terminals exhibited, as expected, the highest expression levels. The genes and proteins involved in inflammation (MHC I, MMP9, TNF-α), apoptosis (Bax, Bcl-2, caspase-3), or stress response (ATF3, VEGF) were expressed at basic levels. Sciatic nerve transection and reconstruction significantly increased the expression levels of these genes, although only in the first week. Similar time courses have been described previously, including those for Bax (53), ATF3 (54) and MHC I (55). The coincident but temporary upregulation of pro-apoptotic genes and proteins (Bax and caspase-3) or anti-apoptotic genes and proteins (Bcl-2), inflammation or stress demonstrated the presence of a self-defensive response to injury, and a conflict between injury-induced neurodegenerative signaling cascades and neuroprotective mechanisms during the acute phase following injury. In certain cases, this results in the survival and recovery of stressed motorneurons, as was observed in the present model of sciatic nerve reconstruction. In severe traumatic injuries, such as nerve avulsion, significant motorneuronal death occurred, accompanied by mitochondrial accumulation of Bax, cytochrome c redistribution and activation of caspase-3 (29). Schwartz et al (56) termed this process a detrimental cost-benefit ratio; inflammation, being primarily a positive self-response eliminating or neutralizing injurious stimuli and restoring tissue integrity, exceeds the threshold of tolerability, and contributes to neuropathology. In this regard it appeared that the immune response to nerve injury in neonatal rats is reversed (57), thus offering one possible explanation for the aforementioned enhanced neuro-vulnerability in young animals.

Only MMP9 and GAP-43 demonstrated significantly increased ipsilateral induction compared with the contralateral side. All other genes reported no significant differences when the ipsilateral and contralateral sides were compared. The absence of ipsilateral vs. contralateral differences is in contrast to findings of Tang et al (58), which demonstrated that unilateral root-avulsion resulted in significant alterations to microRNA expression only in the ipsilateral spinal cord. However, the present results are in agreement with Rotshenker and Tal (59), who revealed that sprouting and synapse formation is enhanced by contralateral axotomy. Furthermore, there is evidence for transneuronal correspondence between ipsilateral and contralateral motorneurons. Transneuronal labeling of the L4 and L5 VH neurons following pseudorabies virus injection in the rat medial gastrocnemius muscle has been previously described (60). Neuropeptides, including peptide histidine isoleucine (61) and calcitonin gene-related peptide (CGRP) (62), were also induced bilaterally in rat spinal motor neurons following unilateral sciatic nerve transection. CGRP has been proposed to be involved in pain transmission and inflammation (63), as well as in repair mechanisms for neural regeneration following brachial plexus (2) or sciatic nerve (5) injury, in which its anti-apoptotic properties (64) were essential. These results are concordant with the hypothesis that unilateral sciatic nerve injury is able to induce bilateral stress and self-defense.

Synapse stripping is a regular result of peripheral axotomy, in which the extent of synaptic terminal retraction depends on the distance between motorneuron and lesion [it is lessened when the lesion side is further from the cell soma (65)], and on the severity of the lesion. This process occurs when neuronal cell death is not obvious (66). This remodeling has been suggested to be an adaptive mechanism of self-defense underlying enhanced neuronal viability (67,68). Although the microglia activation demonstrated in the present study may be associated with synaptic stripping, a previous study suggested that the activation of glia is not correlated with the degree of synaptic stripping (65).

Numerous studies have demonstrated minocycline-induced neuroprotection (69–71). For axotomized motorneurons it influenced both the ipsilateral as well as the contralateral site (72). In the present investigation the effects of minocycline were relatively low, but did induce marginal inhibition with a reduction in the expression levels of MMP9, TNF-α, MHC I, VEGF and GAP-43. The inhibitory effects of minocycline have previously been described for MMP9 (73), MHC I (74), VEGF (75,76), TNF-α (77,78), and GAP-43 (79). One target of minocycline appears to be the transcription factor NF-κB. Minocycline has been demonstrated to inhibit the activation of NF-κB (80), as well as its translocation into the cell nucleus (81). NF-κB however induces the expression of MMP9 (82), MHC I (83) and VEGF (84). Furthermore, the activation of NF-κB culminates in the release of TNF-α (85), which is a potent activator of NF-κB (86), thus an escalation of the minocycline effects can be assumed.

A late induction of motorneuronal GAP-43 expression following sciatic nerve injury has been previously demonstrated (87). GAP-43 is widely used as a marker for the growth/regeneration state of motorneurons, including synapse reconstruction, referred to as the ‘cell body response’ (88). The expression of GAP-43 requires acetylated p53 (89). However, minocycline is able to downregulate the expression of p53 (90) and to inhibit acetylation (91), which results in the minocycline-induced downregulation of GAP-43 expression demonstrated by the results of the present study.

Minocycline is able to downregulate or upregulate Bax and Bcl-2, respectively, thereby resulting in an anti-apoptotic ratio (76,92–94). However, the present study did not demonstrate any significant minocycline effects on Bax or Bcl-2. These non-concordant results may be due to the heterogeneity of the models and minocycline treatment regimes. Matsukawa et al (95) demonstrated that minocycline attenuates experimentally-induced ischemic cell death by upregulating Bcl-2 expression at low doses. However, high minocycline doses exacerbated ischemic injury and reduced the number of Bcl-2-expressing neurons. Furthermore, minocycline targeted neurons alone, not astrocytes (95). The present results of the PCR analysis on the spinal cord tissue samples revealed the expression pattern of all spinal cord cell types including glial cells. Experiments were conducted using the NSC-34 motorneuronal-like cell line.

OGD, but not LPS, was highly toxic for NSC-34 cells and minocycline reduced the OGD-induced cell death rate, although these results were not significant (P>0.05). Notably, stress-induced changes in apoptosis-associated Bax and caspase-3 expression were not observed. These results were concordant with an in vivo study that demonstrated that dying lumbar motorneurons did not always exhibit apoptotic morphology (96). There is also evidence that NSC-34 cells expressed the apoptotic markers only under specific conditions (97). NSC-34 cell death could be induced by various apoptotic agents, and when intracellular protein inclusions containing mutant SOD1 existed, dispersed SOD1 prevented NSC-34 cells from apoptotic cell death (98). This may explain the observed apoptotic death of NSC-34 cells following H2O2-induced oxidative stress (99), as oxidative stress induced SOD1 aggregation (100). The absence of Bax, caspase-3, MHC I and ATF3 activation in NSC-34 cells suggested that motorneurons were not preferentially or even solely responsible for the nerve injury-mediated upregulation of these genes. PCR analysis demonstrated the expression pattern of neurons and glial cells, and the expression and stress-mediated regulation of these genes has been well-described: Astroglial Bax and caspase-3 (101); MHC I (102); ATF3 (103); microglial Bax and caspase-3 (104); MHC I (105); ATF3 (34); oligodendroglial Bax and caspase-3 (106); MHC I (107); and ATF3 (108). The absence of GAP-43 activation in NSC-34 cells may be a result of the in vitro absence of retrograde signals, which in vivo originate from the distal nerve stump and the disconnected nerve targets to initiate and support axonal regeneration (109).

TNF-α and VEGF induced the expression of MMP9; however, only OGD induced the expression of Bcl-2, and was able to parallelize the activating potency of sciatic nerve reconstruction. These results suggested that the expression level changes observed in vivo may be induced by motorneurons. However, the involvement of glia cells cannot be excluded. The glial expression of the four genes has previously been described: MMP9 (110,111); TNF-α (112); VEGF (113,114); and Bcl-2 (92,115).

In NSC-34 cells minocycline exhibits inhibitory effects, whereby the above-mentioned NF-κB signaling pathway may be accepted due to the fact that NF-κB is expressed by NSC-34 cells, and activated and translocated into the nucleus as a result of cell stress (116,117).

The present study demonstrated a massive but temporary SNR-mediated upregulation of all studied genes in L3-L6 sections of the spinal cord that was moderately affected by minocycline. The results observed within NSC-34 cells indicate that motorneurons are not significantly or solely responsible for these SNR-mediated changes in gene expression. To further clarify the cell-specific gene profiles, a more complex model of organotypic cell cultures may be a helpful alternative. This model could mimic tissue architecture of the spinal cord, which would allow an understanding of cellular etiology of these processes.

Acknowledgements

The authors of the present study are grateful to Ms. Leona Bück for technical assistance.

References