Serum response factor modulates neuron survival during peripheral axon injury

Background The transcription factor SRF (serum response factor) mediates neuronal survival in vitro. However, data available so far suggest that SRF is largely dispensable for neuron survival during physiological brain function. Findings Here, we demonstrate that upon neuronal injury, that is facial nerve transection, constitutively-active SRF-VP16 enhances motorneuron survival. SRF-VP16 suppressed active caspase 3 abundance in vitro and enhanced neuron survival upon camptothecin induced apoptosis. Following nerve fiber injury in vitro, SRF-VP16 improved survival of neurons and re-growth of severed neurites. Further, SRF-VP16 enhanced immune responses (that is microglia and T cell activation) associated with neuronal injury in vivo. Genome-wide transcriptomics identified target genes associated with axonal injury and modulated by SRF-VP16. Conclusion In sum, this is a first report describing a neuronal injury-related survival function for SRF.


Background
The gene regulator SRF modulates multiple aspects of neuronal motility. In SRF-deficient mice, cell migration, neurite outgrowth, branching, growth cone shape and axon guidance are impaired. In turn, constitutively-active SRF-VP16, a fusion protein of SRF and the viral VP16 transactivation domain, enhances neuronal motility [1]. Thus, SRF's impact on physiological neuronal motility might proof beneficial also during axonal regeneration, that is the stimulation of regrowth of severed nerve fibers.
In addition to cell differentiation, SRF has been implicated in cell survival of various cell types including hepatocytes [2], thymocytes [3], heart cells [4], and during embryogenesis [5]. In embryonic stem cells lacking SRF apoptosis was strongly upregulated [4]. The latter result is in line with downregulation of the antiapoptotic protein Bcl-2 upon SRF-deficiency. Bcl-2 was identified as SRF target gene in the same study [5]. In primary cortical neurons, SRF overexpression mediates BDNFdependent cell survival in various paradigms of neuronal injury [6]. Also, SRF conveys expression of the immediate early gene (IEG) Cyr61 during neuronal cell death [7].
SRF operates through interaction with co-factors of the MRTF (myocardin-related transcription factors) and TCF (ternary complex factors) family. Through interaction with TCFs SRF can mediate an IEG response of for example c-fos, Egr1 and Arc. IEGs are well-established molecular switches of cell survival vs. cell death [8]. Further, while interacting with MRTFs SRF directs expression of actin isoforms (Acta, Actb, Actc) or actinbinding proteins (for example tropomyosin, calponin and gelsolin) thereby regulating cytoskeletal dynamics [1,9].
Similar to SRF, MRTF-A and the TCF Elk-1 enhance cell survival of primary neurons [6,[10][11][12][13] and non-neuronal cells [14]. In opposite to primary neurons, cell survival and apoptosis are not overtly altered during physiological nervous system development as revealed by SRF-deficient mice [15][16][17]. Indeed, apoptosis was only elevated in the subventricular zone of SRF-deficient mice [15] but not documented in for example cortical, hippocampal, striatal and peripheral neurons [15][16][17]. This suggests that SRF is not a major neuronal survival regulator during physiological brain development.
As mentioned above, SRF and co-factors mediate injury-related neuronal survival in vitro. Thus, an in vivo function of SRF in neuronal survival (which has not been demonstrated so far) might become apparent during application of neuronal injury.
Here we applied facial nerve injury in adult mice to investigate a role of SRF-VP16 in survival of facial motorneurons in vivo. In mice, the bilateral facial nerve innervates muscles regulating whisker pad and eyelid movements, for example [18]. Facial nerve axotomy is a model system for studying motorneuron survival, axonal regeneration as well as neuron and immune cell interactions during neuronal injury. We observed an SRF-VP16 dependent increase in motorneuron survival in vivo. In addition, SRF-VP16 enhanced outgrowth and survival of transected primary neurons in vitro. Mechanistically this SRF-VP16 function involves suppression of active caspase 3 expression in vitro and increased microglia and T cell activation around transected motorneurons in vivo. Finally, using transcriptomics, we provide axonal injury-induced and SRF-VP16 modulated target genes potentially associated with neuronal survival.

Facial nerve transection
The facial nerve transection was performed as described in [19]. Adult wild-type mice (>2 month) were anaesthetized, a skin incision was made behind the left ear and the facial nerve was exposed. In experiments with no virus application, the nerve was transected with small microscissors about 2 mm posterior to the foramen stylomastoideum. For viral infection, 1 μl virus was injected into the facial nerve using a 26 G Hamilton syringe. Afterwards, the nerve was transected and another 1 μl of virus was injected into the nerve stump. Of note, this virus injection with a syringe causes already a facial nerve lesion. Therefore it is only possible to delineate SRF-VP16 specific effects on the basis of experiments employing control virus, SRF-ΔMADS-VP16. Cesium-chloride purified SRF-VP16 (4.6 × 10 12 PFU/mL) and SRF-ΔMADS-VP16 (4.9 × 10 12 PFU/mL) adenoviral particles were purchased from Vector Biolabs. Both viruses drive GFP expression via a second CMV promoter. Absence of eyelid closure and whisker movement ensured successful nerve transection. All experiments are in accordance with institutional regulations by the local animal ethical committee (Regierungspräsidium Tübingen).
For neuronal injury experiments in vitro, hippocampal neurons were grown on poly-L-lysine and laminin coated video dishes. One neurite/neuron was cut with a microscalpel driven by an InjectMan® NI 2 Micromanipulator (Eppendorf). The cell reaction was monitored in a life cell imaging set-up (37°C, 5% CO 2 ; Zeiss, Axiovert 200 M) every 5 min for a total of 6 h. Ten neurons/condition in 13 independent experiments were evaluated.

Microarrays
The facial nuclei were dissected from 300 μm brainstem sections prepared with a tissue chopper using tungsten needles. Facial nuclei of four mice/ condition were pooled and resulted on average between 0.5 and 1 μg RNA. Total RNA was isolated with the RNeasy kit (Qiagen). RNA of 0.1 μg was processed on Affymetrix GeneChips (Mouse Gene 1.0 ST array) according to protocols of the Microarray Facility Tübingen (http://www.microarray-facility.com/cms/index. php). Raw data normalized to the control sample were analyzed in such way that only genes with a fold-change of ≥ 1.5 (up-or down-regulated) were carried forward. Genes were considered SRF-VP16 specific if their fold-change differed two-fold from the respective factor obtained for SRF-ΔMADS-VP16.

Quantitative real-time PCR (qPCR)
Total RNA derived from facial nuclei of four animals was isolated with the RNeasy kit (Qiagen). Reverse transcription was performed with 0.5 to 1 μg RNA using reverse transcriptase (Promega) and random hexamers. qPCR was performed on ABI PRISM 7700 Sequence Detector with the Power PCR SYBR green PCR master mix (Applied Biosystems). Expression was determined in relation to Gapdh RNA levels. Mouse primers used were as follows: Cnn1

Statistics and quantification
Numbers of independent experiments or animals are indicated in figure bars. For all cell culture experiments at least three independent cultures derived from different animals were analyzed. For quantification of neuron numbers in facial nerve injury experiments, all sections (that is 10 to 15 sections/animal) were evaluated. Neurons were scored as non-degenerated, if they protruded at least one neurite and if the cell body showed a typical angled shape. For microglia and T cell numbers 4 to 6 sections/ animal were analyzed. Statistical significance was calculated using two-tailed t test or, where appropriate, a one-way analysis of variance (ANOVA) with a Bonferroni post hoc test.

SRF-VP16 enhances motorneuron survival in vivo
To investigate a role of SRF in neuron survival, we employed a well-established model system of neuronal injury, that is unilateral de-afferentiation of facial motorneurons in mice ( Figure 1A). The transected facial nerve was infected with viral particles expressing GFP in addition to SRF-ΔMADS-VP16 or SRF-VP16. SRF-VP16 consists of SRF fused to the viral VP16 transactivation domain. To control for VP16 off-target effects, SRF-ΔMADS-VP16, lacking DNA binding activity, was used as control [20]. SRF expression commenced 1 day after infection. Around the virus injection site of the facial nerve, SRF was also found in fibroblasts and glial cells, whereas in the facial nucleusafter retrograde viral transport -SRF expression was motorneuron restricted (Figure 1, Additional file 1: Movie S1 and Additional file 2: Movie S2 and data not shown).
Survival was quantified by analyzing number and morphology of GFP-positive motorneurons in the facial nucleus. SRF-VP16 increased the number of surviving motorneurons compared to SRF-ΔMADS-VP16 ( This suggests that SRF plays a role in survival of axotomized facial motorneurons.

SRF-VP16 suppressed cell death and enhanced neurite regrowth in vitro
As shown above (Figure 1), SRF-VP16 protects from motorneuron loss upon facial nerve lesion in vivo. To investigate whether SRF-VP16 has also an impact on neuronal survival of primary neurons we employed an in vitro assay of axonal injury ( Figure 2). Here, neurites of primary neurons were transected using a micro-scalpel followed by recording the neuronal response with timelapse video-microscopy (Figure 2A-C). VP16 expressing neurons were identified via GFP-expression. After lesion, neurites of an SRF-ΔMADS-VP16 expressing neuron did not re-grow and neurons frequently died (Figure 2A). In contrast, neurites of an SRF-VP16 expressing neuron were capable of re-growth, often protruded dynamic growth cones and survived neurite transection ( Figure 2B; quantification in 2C). Thus, SRF-VP16 enhances neuronal survival and re-growth of severed neurites in vitro.
In a next step, we investigated potential mechanisms of SRF's function in neuronal survival. SRF-VP16 might modulate motorneuron survival via blocking apoptosis. To investigate this further, we employed primary neurons assessing protein levels of the pro-apoptotic protein active caspase-3 upon SRF-VP16 expression ( Figure 2D, 2E). In SRF-ΔMADS-VP16 expressing neurons, active caspase 3 levels were strongly induced. In contrast SRF-VP16 suppressed this myelin-induced activation of active caspase-3 ( Figure 2D, 2E). Notably, this effect was more obvious in neurons lacking SRF compared to wild-type neurons ( Figure 2D and quantification in 2E).
In a further set of experiments we analyzed whether SRF-VP16 might enhance neuronal survival upon camptothecin induced DNA damage ( Figure 2F, 2G). In control infected primary neurons either expressing GFP alone or SRF-ΔMADS-VP16, camptothecin induced apoptosis in a concentration dependent manner. This was quantified by either counting numbers of active caspase 3 positive ( Figure 2F) or numbers of surviving GFP-positive neurons ( Figure 2G). In contrast, SRF-VP16 expression reduced this camptothecin induced neuronal cell death compared to control constructs ( Figure 2F, 2G). Thus, in cultures infected with adenoviral particles expressing SRF-VP16 (along with GFP) the number of active caspase 3 positive neurons was reduced ( Figure 2F) whereas more GFP-positive neurons survived camptothecin treatment ( Figure 2G). This finding is in agreement with previous observations made with wild-type SRF in cortical neurons [6].
Thus, results from primary neurons suggest that SRF-VP16 might down-regulate expression of pro-apoptotic proteins to enhance neuronal survival.

SRF-VP16 modulates injury associated immune responses
Neuronal injury is accompanied by immune responses, for example astrocyte, microglia, and T cell activation and their subsequent infiltration of lesioned neuronal tissue. Injury-related immune responses might dampen as well as exacerbate neuronal loss [18,22]. Regarding the facial nerve lesion model, peri-neuronal accumulation of microglia cells facilitates axonal regeneration [23]. Also T cells were assigned important roles for the immune surveillance during facial nerve injury [24]. Given this important link between an immune response and neuronal injury, we asked whether SRF-VP16 might modulate immune responses associated with axon injury (Figure 3). Firstly, we inspected microglia activation in the de-afferented facial nucleus. Microglia were expectedly elevated at the lesion side compared to the control side at 5 and 25 d.p.i ( Figure 3A, 3E and 3C, 3G; 3M). SRF-VP16 augmented microglia activation in the lesioned facial nucleus at both time-points compared to SRF-ΔMADS-VP16 ( Figure 3B, 3 and 3D, 3H). Notably, SRF-VP16 also enhanced microglia occupancy along the facial nerve axons and the axon exit point (arrows and arrowheads in Figure 3H, respectively; Figure 3N).
Secondly, we investigated T cells labeled with an anti-CD3 directed antibody. T cells did not enter the facial nucleus 5 d.p.i. regardless of virus type ( Figure 3O). In contrast, at 25 d.p.i., we observed T cell infiltration in lesioned SRF-ΔMADS-VP16 expressing neurons but not on the uninfected control side ( Figure 3I, K, and 3O). Similar to results obtained on microglia ( Figure 3A-H), SRF-VP16 also enhanced T cell infiltration around motorneurons and axons ( Figure 3L and 3O).
Taken together, microglia and T cell responses are augmented upon SRF-VP16 expression.

Microarray analysis of lesion and SRF-VP16 induced transcripts
SRF might enhance neuronal survival through various mechanisms including regulation of survival/apoptosis related ( Figure 2) and immune regulatory genes (Figure 3). To identify genes modulated by facial nerve injury per se and by SRF-VP16, we performed transcriptomics after three days of facial nerve lesion (Figures 4  and 5, Table 1 and Additional file 3: Table S1). For this, facial nuclei of four mice were pooled for each condition.
To corroborate microarray data (Figure 4), qPCR analysis of selected genes employing independent cDNA samples was performed. Indeed fold changes obtained in this qPCR analysis were comparable to the microarray data ( Figure 5).

Discussion
So far, SRF signaling was not assigned a major role in neuronal survival in vivo [15,16] in contrast to injury-related survival in vitro [6,10,12,26]. This suggests that SRF regulates neuron survival primarily in an injury related situation rather than in physiological brain development. In accordance, we here demonstrate a neuroprotective SRF-VP16 function in vivo, that is preventing motorneuron degeneration upon facial nucleus deafferentiation (Figure 1). Further, SRF-VP16 prevented expression of proapoptotic active caspase 3, enhanced regrowth of severed neurites in vitro and reduced camptothecin induced apoptosis ( Figure 2). The latter might be directly linked to SRF-VP16 induced cytoskeletal genes such as actin isoforms (Actc1, Acta2), calponin (Cnn1), and actinin (Actn3; Figure 4 and Figure 5).

How might SRF-VP16 enhance facial motorneuron survival?
SRF-VP16 suppressed active caspase 3 in vitro and reduced camptothecin-induced neuronal cell death ( Figure 2). This SRF-VP16 mediated reduction of active caspase 3 was stronger in primary neurons lacking SRF compared to wild-type neurons (Figure 2). Such a reduction in proapoptotic protein levels by SRF-VP16 might enhance neuronal survival also upon facial nerve injury in vivo. To modulate expression of apoptosis related proteins, SRF-VP16 might recruit IEGs, known regulators of neuronal survival [8], such as Egr-1 and Egr-2 [27] which were induced by SRF-VP16 during facial nerve lesion (Figures 4 and 5).
In contrast to primary neurons (Figure 2), we did not observe any major effect of SRF-VP16 compared to SRF-ΔMADS-VP16 on active caspase 3 and BAX expression upon facial nerve lesion of wild-type mice in vivo (data not shown). In addition SRF-VP16 did not alter Ki-67 expression, a proliferation marker. Ki-67 was strongly induced in lesioned facial motorneurons compared to unlesioned neurons at 7 days but notably not anymore at 21 days after lesion (data not shown).
Thus, similar to primary neurons (Figure 2), in vivo SRF-VP16's potential to enhance neuronal survival might be more pronounced and only become visible in the absence of endogenous wild-type SRF. Indeed it is known that wild-type SRF competes with SRF-VP16 for access to certain SRF target gene promoters such as Bcl-2 [8].
In sum, using SRF-deficient primary neurons we demonstrate that SRF-VP16 modulates apoptosis in vitro. Thus it will be useful to employ SRF-deficient mice to unmask SRF-VP16's impact on apoptosis also in vivo.
SRF-VP16 enhanced injury associated immune responses including microglia and T cell activation ( Figure 3). SRF-VP16 enhanced microglia occupancy at facial nerve axons (Figure 3). In axonal injury, immune cells such as microglia remove myelin debris and have neuroprotective potential which might enhance neuronal survival [22]. Thus, by stimulating microglia and T-cell number and infiltration in the lesioned FMN and thereby increasing motorneuron cell body and facial nerve axon occupancy with these potentially neuroprotective immune cells, SRF-VP16 might enhance neuronal survival.
Of note, SRF-VP16 expression in the facial nucleus was confined to motorneurons ( Figure 1 and Additional file 1: Movie S1 and Additional file 2: Movie S2). Thus, SRF-VP16 expression in neurons might influence immune cells such as microglia and T cells via a paracrine mechanism. Such a paracrine mechanism whereby neuronal SRF affects neighboring cells via regulation of secreted molecules has been described before, for example in oligodendrocytes [28,29]. With regard to immune responses initiated upon facial nerve injury such a paracrine mechanism might involve, for example, cytokine/hormone secretion by neurons. Indeed,    Table 1, and Additional file 3: Table S1). In microarray results presented in this study (Figures 4 and 5, Table 1, and Additional file 3:  Table S1), facial nuclei of four animals were collected in a single biological sample. Thus, although we confirmed some results with independent cDNAs in qPCR ( Figure 5), interpretation of microarray data is limited by a lack of statistical evaluation. Taking this into account, SRF target genes associated with up-regulated immune responses might include Il1r2, Galp, Nms, Il1f9, and Timp1. For instance Timp1, a regulator of matrix metalloproteases activity and thereby modulator of, for example, microglia migration [30] is more than 12-fold induced by SRF-VP16 (Table 1).