|Home | About | Journals | Submit | Contact Us | Français|
Targeted deletion of the phosphatase and tensin homolog on chromosome ten (PTEN) gene in the sensorimotor cortex of neonatal mice enables robust regeneration of corticospinal tract (CST) axons following spinal cord injury as adults. Here, we assess the consequences of long-term conditional genetic PTEN deletion on cortical structure and neuronal morphology and screen for neuropathology. Mice with a LoxP-flanked exon 5 of the PTEN gene (PTENf/f mice) received AAV-Cre injections into the sensorimotor cortex at postnatal day 1 (P1) and were allowed to survive for up to 18 months. As adults, mice were assessed for exploratory activity (open field), and motor coordination using the Rotarod®. Some mice received injections of fluorogold into the spinal cord to retrogradely label the cells of origin of the CST. Brains were prepared for neurohistology and immunostained for PTEN and phospho-S6, which is a downstream marker of mammalian target of rapamycin (mTOR) activation. Immunostaining revealed a focal area of PTEN deletion affecting neurons in all cortical layers, although in some cases PTEN expression was maintained in many small-medium sized neurons in layers III-IV. Neurons lacking PTEN were robustly stained for pS6. Cortical thickness was significantly increased and cortical lamination was disrupted in the area of PTEN deletion. PTEN-negative layer V neurons that give rise to the CST, identified by retrograde labeling, were larger than neurons with maintained PTEN expression, and the relative area occupied by neuropil vs. cell bodies was increased. There was no evidence of tumor formation or other neuropathology. Mice with PTEN deletion exhibited open field activity comparable to controls and there was a trend for impaired Rotarod performance (not statistically significant). Our findings indicate that early postnatal genetic deletion of PTEN that is sufficient to enable axon regeneration by adult neurons causes neuronal hypertrophy but no other detectable neuropathology.
The phosphatase and tensin homolog (PTEN) gene has emerged as a promising target for enhancing axon regeneration in the mature central nervous system (CNS). The initial finding was that deletion of PTEN enabled regeneration of axons of retinal ganglion cells following optic nerve crush (Park et al., 2008) and reduced the retrograde degeneration of axotomized RGCs that otherwise occurred. This study used transgenic mice with a floxed PTEN gene that was deleted by local injections of AAV-Cre. A subsequent study used the same approach to show that conditional genetic deletion of PTEN in the sensorimotor cortex of neonatal mice enabled regeneration of corticospinal (CST) axons following spinal cord injury (SCI) in adulthood. CST regeneration was seen following both dorsal hemisection and complete crush injury (Liu et al., 2010). Follow-up studies using similar approaches reported CST regeneration with conditional genetic deletion of PTEN in adult mice at the time of a spinal cord injury (Danilov & Steward, 2015) as well as in the chronic post-injury period (Du et al., 2015). These results, coupled with PTEN's upstream and non-redundant negative regulation of the PI3K/AKT pathway, make it a potentially important therapeutic target following CNS injury (Don et al., 2012).
The PTEN gene has been extensively studied for its role as a tumor suppressor and mutations are seen in several human cancers and cancer syndromes (Ali et al., 1999) (Goffin et al., 2001; Shi et al., 2012). Also, the PTEN/mammalian target of rapamycin (mTOR) pathway is well known for its ability to regulate cell growth and proliferation (Laplante and Sabatini, 2012). mTOR is up-regulated by activation of phosphoinositide 3-kinase (PI3K), which converts phosphatidylinositol (4,5)-bisphosphate (PIP2) to phosphatidylinositol (3,4,5)-triphosphate (PIP3) (Georgescu, 2010). Via its action as a phosphatase, PTEN opposes PI3K activity by converting PIP3 into PIP2, thus damping activation of downstream pathway components, such as phosphorylated ribosomal protein S6 (pS6), limiting cell growth (Park et al., 2010). Deleting PTEN activates mTOR, which presumably leads to downstream changes in mRNA translation regulated by mTOR signaling. PTEN mutations in humans have been associated with neurological abnormalities, including macrocephaly and autism (Goffin et al., 2001; Waite and Eng, 2002).
Previous studies have explored the consequences of deleting PTEN using transgenic promoter-driven Cre expression in PTENf/f mice, including Ca2+/calmodulin-dependent protein kinase II (CamKII), (Sperow et al., 2012), dopamine active transporter (DAT), (Diaz-Ruiz et al., 2009), glial fibrillary acid protein (GFAP), (Backman et al., 2001; Fraser et al., 2008; Fraser et al., 2004; Kwon et al., 2001; Wen et al., 2013; Yue et al., 2005), and neuron specific enolase (NSE), (Kwon et al., 2006; Takeuchi et al., 2013). The expression of each promoter (and therefore Cre recombinase) varies, with the earliest expression beginning at embryonic day 12.5 (E12.5) for NSE (Forss-Petter et al., 1990), E13.5 for GFAP (Brenner et al., 1994), E15 for DAT (Smits and Smidt, 2006), and postnatal day 4 for CamKII (Burgin et al., 1990).
Deletion of PTEN in neurons via transgenic promoter-driven Cre expression in vivo leads to increased brain mass, cerebellar enlargement, disruption of cortical lamination, and neuronal hypertrophy (Kwon et al., 2001; Marino et al., 2002; van Diepen and Eickholt, 2008; Yue et al., 2005). In addition to the gross and microscopic anatomical changes following in vivo PTEN deletion, neuropathologies including seizures, hydrocephalus, and ataxia were observed (Fraser et al., 2004; Kwon et al., 2001; Pun et al., 2012).
Previous studies involving conditional postnatal PTEN deletion in neurons using transgenic promoter-driven Cre expression have limited their analyses primarily to young adult mice (up to 30 weeks old) (Kwon et al., 2006; Luikart et al., 2011; Takeuchi et al., 2013; Williams et al., 2015). The consequences of longer-term PTEN deletion have not been assessed. Also, there have not yet been assessments of the consequences of early postnatal AAV-mediated PTEN deletion in the sensorimotor cortex, in the way that enables robust regeneration of adult CST axons (Liu et al., 2010).
The goal of the present study was to determine the consequences of long-term PTEN deletion in the sensorimotor cortex, focusing on the effects on cortical structure, neuronal size, functional consequences (if any), and whether there is evidence of tumors or other neuropathology. We show that AAV-Cre injections at postnatal day 1 lead to focal deletion of PTEN but with selective maintenance of PTEN expression in some neuronal types (small-medium sized neurons). There were increases in cortical thickness in the area of PTEN deletion, and cortical lamination was altered, with increases in the ratio of neuropil to cell bodies. Cortical motoneurons lacking PTEN identified by retrograde labeling were hypertrophied. There was no evidence of tumors, necrosis, inflammation, or other evident neuropathology. Maintained PTEN expression by some neurons may be a result of selective tropism of AAV-Cre for certain neuron types in the neonatal stage of development.
All experimental procedures were approved by the Institutional Animal Care and Use Committee at the University of California, Irvine. Transgenic mice with a floxed exon 5 of the PTEN gene (PTENf/f) were from our local breeding colony that was established from founders obtained from Jackson Laboratories (Strain: C;129S4-Ptentm1Hwu/J, Stock number: 004597). The colony was maintained by intercrossing mice homozygous for the floxed allele.
PTENf/f mice were injected on postnatal day 1 with AAV-Cre (n=22) or AAV-GFP (n=12). The vectors were AAV serotype 2, and were obtained from Vector Bio Labs (1 × 1013 genome copies/mL, Catalog number 7011). The virus was diluted with sterile-filtered PBS and 5% glycerol for a final concentration of 1 × 1012 genome copies/mL. To perform the injections, 1-day-old pups were anesthetized by hypothermia (up to 5 minutes) and placed on a clay body mold attached to the bite bar of a stereotaxic apparatus. Injections were made using a Hamilton microsyringe with a pulled glass micropipette tip. 6 mice received injections as in Liu et al. with two unilateral injections of AAV-Cre (0.5 mm lateral, 0.5 mm deep to the cortical surface, and +0.5 and 0.0 mm with respect to bregma) for a total volume of 2 μl (Liu et al., 2010). 23 mice were injected using a procedure developed in Dr. Zhigang He's lab involving three unilateral injections (0.5 mm lateral, 0.5 mm deep to the cortical surface, and +0.5, 0.00, and -0.5 mm with respect to bregma) for a final total injection volume of 1.5μl of vector (n=11 AAV-Cre and 12 AAV-GFP). After injection, pups were dipped in a few drops of sesame oil and returned to their home cages. After weaning, mice were maintained under standard vivarium conditions for 12-18 months. Table I summarizes animals and experimental conditions.
Animals with PTEN deletion (n=9) and AAV-GFP injected controls (n=12) were tested for motor coordination when they were 48 to 53 weeks of age (51.6 weeks of age, on average). Mice were placed individually in each of the four chambers on the rod of a Rotarod® (San Diego Instruments) were allowed to acclimate to the stationary rod for 30 seconds before starting every trial. Initial speed of the rotorod was set to 3 RPM and acceleration was set at 1 RPM/second. Mice were allowed to stay on the rod until they fell off or up to a maximum of two minutes, for a total of 7 trials/day. The first two trials were considered training trials, and the next 5 were testing trials. The time at which the mice fell from the rod during the testing trials (if at all) was recorded. PTEN deletion was assessed by immunocytochemistry (see below) and data from 3 mice with PTEN deletion located outside of the motor cortex were excluded from the statistical analysis, leaving a final n=6 for the PTEN deletion group.
The same mice that were tested using the Rotarod were also tested for general locomotion and spatial learning in the open field (9 animals with PTEN deletion and 12 animals with AAV-GFP). A 24-inch by 24-inch dark plexiglass box was used for the open field test. A 3-inch by 3-inch grid was drawn onto the bottom of the plexiglass box. The grid had four different colored zones, a center square, and the corner squares marked. The four differently colored zones were determined based on their proximity to the center square. The lines that comprised the outer two zones were considered “outer”, while the lines of the inner two zones and the center square were considered “inner” lines. Individual mice were placed in the center of the box and allowed to explore freely for 5 minutes. One observer recorded total grid line crossings by forepaws, noting whether the lines crossed were “outer” or “inner” lines. The second observer recorded other events such as defecation, urination, grooming, and rearing. Open field activity was tested once a day for four consecutive days.
At approximately 1 year of age, mice that had received AAV-Cre (n=6) or AAV-GFP (n=2) received bilateral injections of the retrograde tracer Fluorogold (FG) into the spinal cord. Mice were anaesthetized with isoflurane and a laminectomy was performed to expose the spinal cord at C5. 0.2μl of 1% FG was injected with a Hamilton syringe bilaterally from the midline (0.5mm lateral, 0.5mm deep) over 1 minute. The syringe was left in place for an additional minute. After completing the injections, the skin was closed with sutures and staples, and mice were kept on a 37°C water circulating heating pad until they recovered from the anesthetic. Mice received subcutaneous injections of buprenorphine, Baytril, and lactated Ringer's solution twice a day for 3 days post surgery, and were allowed to remain in their home cages for 3 (n=7) to 4 (n=1) days before tissue collection.
At 12-18 months of age, mice received intraperitoneal injections of Euthasol (1mL/300g), and were perfused transcardially with 4% paraformaldehyde (PFA). The brain and spinal cord were dissected and post-fixed in PFA overnight, then placed in 30% sucrose solution in PBS overnight for cryoprotection. Brains were embedded in OCT compound and frozen in a container of methyl butane surrounded by ethanol/dry ice slush. 20μm thick coronal sections were collected at 200μm intervals and stored in PBS with 0.1% sodium azide at 4°C until staining.
Sections were either directly mounted for Fluorogold visualization under fluorescent microscopy, stained for H&E or cresyl violet, or immunostained (see Table I for summary). To visualize GFP signal in the control animals, we examined unstained sections for GFP fluorescence and immunostained sections for GFP.
For immunostaining for cleaved caspase-3, Ki67, pS6, and PTEN, an antigen retrieval protocol was used in which sections were placed in 1.7ml Eppendorf tubes with 1 mL 10mM citrate buffer, pH 8.80. Tubes were then immersed in boiling water for 5 minutes, followed by immediate immersion in an ice bath for 10 minutes. Sections immunostained for GFP were incubated in 1% H2O2 for 20 minutes to quench endogenous peroxidase activity. Primary antibodies and dilutions were as follows: PTEN (Cell signaling, 1:250), pS6 (Cell signaling, 1:250, #4858), GABA (1:500), GFAP (Sigma, 1:1000), GFP (Lifetech, 1:1500), Ki-67 (Thermo Scientific, 1:200), Cleaved caspase-3 (BD Biosciences, 1:250), GFP (Lifetech, 1:1000). Antibodies were diluted in Blocking Reagent for cleaved caspase-3, pS6, and PTEN (PerkinElmer), 5% secondary antibody host serum in PBS with 0.5% Triton (GABA, GFAP, Ki67), or 5% secondary host serum in TBS with 0.1% Triton (GFP).
Sections were incubated in primary antibody dilutions overnight at room temperature, with the exception of GABA, which was incubated for 36-48 hours at 4°C. Secondary antibodies were diluted in the same blocking solution as the respective primary, and applied for two hours at room temperature. For immunofluorescence, Alexa Fluor 488 or Alexa Fluor 594 (ThermoFisher Scientific, both at 1:250 dilution) were used. For sections undergoing DAB detection and fluorescent amplification, a biotinylated IgG secondary antibody was used (Vector Laboratories, 1:250 dilution). Following secondary incubation, sections were washed 3×, incubated for 1 hour with avidin and biotinylated horseradish peroxidase (Vectastain ABC kit, Vector Laboratories) and reacted with DAB (DAB Substrate Kit for Peroxidase, Vector Laboratories) or FITC (made in house, 1:250 in 0.003% H2O2 borate buffer).
To determine the effect of PTEN deletion on pyramidal neurons that give rise to the CST, cortical sections from AAV-Cre injected animals that had received intra-spinal FG injections were analyzed. Immunostaining for PTEN using a DAB reaction quenched FG fluorescence in PTEN-positive CST neurons whereas fluorescence was still prominent in CST neurons lacking PTEN. We then measured the cross sectional area of 50 PTEN-negative/FG positive CST neurons per animal. For control cells we used the nearest neighboring unstained section and visualized FG signal in the contralateral cortex in unstained sections. This allowed us to specifically measure PTEN-positive layer V pyramidal cells.
It has previously been reported that promoter-driven PTEN deletion in the forebrain results in cortical enlargement. To assess this, we measured the thickness of the cortex (distance from the edge of the corpus callosum to the pial surface) in H&E stained-sections within the area of deletion (n=5 mice). The region of PTEN deletion was verified in a nearby section that had been immunostained for PTEN. To examine changes in cortical laminar patterning, we screened H&E-stained sections from animals with PTEN deletion and compared the region of deletion to the homotopic contralateral control cortex.
To determine whether conditional postnatal PTEN deletion results in altered cell density in the motor cortex we performed point-counting analysis. H&E stained sections within the region of PTEN deletion were imaged at 20× along with images from the control cortex contralateral to the PTEN deletion. These images were then overlaid with a 50μm grid and analyzed in ImageJ for the number of grid crosshairs overlying cells vs. cell-free neuropil (154 maximum). This value was then used to calculate the relative area of neuropil vs. cell bodies.
To define the area of PTEN deletion, sections were immunostained for PTEN. On the non-injected control side, neurons in all cortical layers stained positively for PTEN. There was also diffuse staining in the neuropil layers (Figure 1B, right cortex). Higher magnification views reveal that large pyramidal neurons in layer V are robustly stained (Figure 1C, arrows). PTEN immunostains of sections from mice that received intracortical AAV-GFP injections appeared similar to the contralateral control cortex of AAV-Cre injected mice.
On the side of the AAV-Cre injection, the area of PTEN deletion is easily identified by the decrease in PTEN immunostaining (Figure 1A, left cortex). Higher magnification views reveal a characteristic immunostaining pattern in which neurons lacking PTEN stood out in negative relief as “ghost cells” (Figure 1B, arrows). Although these cells resemble holes in the tissue, visualizing the same section immunostained for PTEN using phase contrast microscopy confirms that the ghost cells are neuronal cell bodies (Figure 1E-F).
Ghost cells were especially evident in layer V (the location of the cells of origin of the CST), and large PTEN-negative dendrites could be seen emerging from the apex of the pyramid-shaped cell bodies and extending through more superficial layers of the cortex (Figure 2A). The area of deletion was defined as the total region containing ghost cells. Although neurons lacking PTEN were evident, in some cases there were many PTEN-positive cells throughout the area of deletion. In these cases, PTEN-positive neurons were especially evident in cortical layers II and IV and these layers contained fewer ghost cells than layers III and V (Figure 2A). Many of the PTEN-positive medium-sized neurons in the AAV-Cre injected cortex had the morphology typical of cortical interneurons (Figure 2A).
To assess whether neurons with preserved PTEN expression were inhibitory interneurons, two nearby sections were immunostained for PTEN or γ-Aminobutyric acid (GABA) (Figure 2B and 2C). GABA-positive neurons were present in all cortical layers within the area of deletion, but their distribution did not align with the distribution of PTEN-positive neurons within the area of deletion; thus, neurons that are PTEN-positive are not exclusively inhibitory interneurons (Figure 2B and 2C).
All mice injected with AAV-Cre had areas with ghost cells characteristic of successful PTEN deletion. The location and size of the area of PTEN deletion varied across cases, however, even when comparing cases that had used identical AAV-Cre titers, injection volumes, number of injections, and injection coordinates. By visual comparison, we found that two injections of 1.0μl of AAV-Cre resulted in more robust PTEN deletion than three injections of 0.5μl AAV-Cre. In all 6 mice that received two injections of greater volume, the rostral to caudal spread of the PTEN deletion area appeared larger and appeared to contain more ghost cells in PTEN immunostains when compared to the 11 mice that had received three injections of lower volume.
Deletion of PTEN results in the activation of AKT, which in turn activates mTOR. One downstream marker for mTOR activation is phosphorylation of ribosomal protein S6 (Ma and Blenis, 2009). Confirming previous findings (Liu et al., 2010), immunostaining with antibodies specific for S6 phosphorylated at ser235/236 revealed increased S6 phosphorylation in neurons within the area of PTEN deletion, even a year after deletion. Fig. 3A illustrates a neighboring section to the section immunostained for PTEN in Figure 1A, B, E. In the area of PTEN deletion, individual neurons exhibited elevated levels of immunostaining in comparison to the same region on the control side (Figure 3A, left vs. right cortex). Pyramidal neurons in layer V with elevated levels of pS6 appeared larger than neurons in the homologous part of the cortex on the contralateral side (compare Fig. 3B with 3C). There was also increased pS6 immunostaining of ascending apical and basal dendrites of the large layer V pyramidal neurons (Figure 3B) in comparison to the contralateral control side (Figure 3C). As a technical aside, colabeling sections for pS6 and PTEN was not attempted because both primary antibodies were generated in rabbit.
In the control cortex and areas where PTEN expression was maintained in the AAV-Cre injected cortex, neurons in all cortical layers stained positively for pS6 with the highest levels of staining exhibited by large pyramidal neurons in layer V (Figure 3C). Levels of immunostaining were highest in neuronal cell bodies and proximal dendrites.
An important technical point is that levels of immunostaining for pS6 depend on the animals' experience during the time just before death, varying in a way that is similar to immediate early gene (IEG) expression. For this reason, it is important to have an internal control, provided here by comparisons of immunostaining in areas with maintained PTEN expression (for example, the side contralateral to the AAV-Cre injection).
Previous studies have reported neuronal and cortical hypertrophy when PTEN is deleted in developing mice using promoter-driven CRE expression (Fraser et al., 2004; Kwon et al., 2003; Kwon et al., 2001; Sperow et al., 2012). To assess whether focal PTEN deletion led to detectable changes in the histological appearance of the cortex, 20μm sections taken every 200μm throughout the brain were stained with hematoxylin and eosin (H&E). Areas of PTEN deletion were evident in several respects. First, in most mice, the cortex was noticeably thicker, and in some mice, there was actually a bulge in the cortex corresponding to the area of PTEN deletion (Figure 3D, left vs. right cortex). Second, cortical cytoarchitecture was somewhat different in that cortical layers were less distinct and cell-packing density appeared lower with increased spacing between neuronal cell bodies (particularly in layer V, compare Figure 3E and 3F). There was no evident neuropathology, such as accumulations of basophilic (basic) immune cells. Note that the tear in the tissue in Figure 3D is a result of tissue processing damage and not damage due to the injection, as was seen in some mice (see below).
In AAV-GFP injected controls native GFP fluorescence was not detectable one year following injection. Immunostaining for GFP revealed scattered GFP positive astrocytes and a few neurons in the area of the injection (Figure 4A and 4B). In contrast to the situation with AAV-Cre-mediated PTEN deletion, cortical cytoarchitecture appeared essentially normal in cases injected with AAV-GFP (Figure 4C). As expected, PTEN immunostaining was intact in this region (Figure 4D) and there were no increases in immunostaining for pS6 in comparison to the contralateral, non-injected side (Figure 4E and 4F).
PTEN is a tumor suppressor, so it was important to screen for any evidence of tumors. No unusual cell clusters were evident in H&E stained sections. For five cases, 20μm sections taken every 200μm were immunostained with GFAP in order to detect areas of astrogliosis indicative of neurodegeneration or any astroglial tumors. GFAP stained sections within the PTEN deletion region had GFAP-positive astrocytes in the glia limitans and in white matter tracts, similar to what we found in the contralateral control cortex (Figure 5A and 5B). No abnormal accumulations of GFAP-positive cells were identified in the area of PTEN deletion.
We noticed that one animal that received AAV-Cre injections and two animals that received AAV-GFP injections had cortical defects in the region of the injection. Sections through the defect were immunostained for GFAP (Figure 5C and 5D); examination of these revealed that the defects were lined with GFAP positive stellate astrocytes. The defects are probably areas of cortical injury produced at the time of the P1 injection, although other explanations cannot be excluded.
To screen for ongoing abnormal cell proliferation in areas of PTEN deletion, sections were immunostained for Ki-67. As a positive control, Ki-67 positive cells were evident in the subventricular zone (SVZ) lining the lateral ventricles. There were no Ki-67 positive cells in the area of PTEN deletion (Figure 6B and 6C). Note that small fluorescent particles evident in Figure 5C were also present in the contralateral PTEN intact cortex immunostained for Ki-67 (Figure 6D). These particles have a different appearance than the Ki-67-positive cells in the SVZ (Figure 6E), and thus are likely immunostaining artifacts. To assess whether the decreased ratios of cell bodies relative to neuropil following PTEN deletion might be due to apoptotic cell death, we immunostained sections for cleaved caspase-3 since cleavage of caspase-3 reflects activation of pro-apoptotic pathways. We did not find any cells within the PTEN deleted or control motor cortex that were positive for cleaved caspase-3 (Figure 6F).
Previous studies have reported alterations in cortical lamination and cell packing following promoter-driven PTEN deletion during early development (Wen et al., 2013; Yue et al., 2005). Similar though less severe effects were qualitatively evident in the area of PTEN deletion. In H&E stained sections, the hexilaminar patterning of the cortex appeared disorganized (Figure 7A vs. 7B). To quantify cell-packing density in the region of PTEN deletion, we used a point-counting method to determine the relative area occupied by cell bodies vs. neuropil in the PTEN-deleted area of the motor cortex vs. the contralateral non-injected cortex. The percentage of neuropil in the control motor cortex was 71.76%±3.47% vs. 79.52%±2.73% in the area of PTEN deletion (paired t-test 12.70, p<0.0002, see Figure 7C). The percentage of neuropil is the inverse of cell-packing density.
Previous studies have also reported cortical hypertrophy following neonatal PTEN deletion (Fraser et al., 2004; Kwon et al., 2006). Since AAV-CRE injections were unilateral, we compared cortical thickness in the area of maximal PTEN deletion vs. the contralateral control. In the area of deletion, cortical thickness ranged from 1171.1 μm to 1527.9 μm (1332.7 μm average, ±139.6), vs. 980.2 μm to 1290.9 μm (1177.6 μm average, ±123.1) on the contralateral un-injected side (two-way ANOVA: F=12.69 for vector type, p<0.002, Figure 8). In mice that received injections of AAV-GFP, the injected region of the cortex was somewhat thinner than the contralateral cortex. In the area of AAV-GFP injection, cortical thickness ranged from 913.4 μm to 1242.2 μm (1087.5 μm average, ±105.9), vs. 967.2 μm to 1160.1 μm (1104.6 μm average, ±67.1) on the contralateral un-injected side (two-way ANOVA: F=0.6280, p=0.4413).
To identify PTEN-negative cells of origin of the CST, 6 mice that had received unilateral injections of AAV-Cre at P1 received bilateral injections of Fluorogold (FG) at C5 as adults. In 5 mice with sufficient FG labeling, FG-positive neurons were evident bilaterally in layer V of the sensorimotor cortex (Figure 9A and 9B). FG-positive dendrites could be seen extending from the labeled cell bodies, and in some cases their dendritic processes and axonal projections were also labeled with FG.
FG-labeled neurons in layer V within the region of PTEN deletion (Figure 9A) appeared larger than FG-labeled layer V neurons in the homologous part of the contralateral cortex (Figure 9B). To identify retrogradely-labeled neurons lacking PTEN, sections from mice that had received FG injections were immunostained for PTEN using DAB for detection. The DAB reaction quenched the FG signal in any neurons that stained positively for PTEN, leaving FG fluorescence intact in neurons that did not stain for PTEN (Figure 9C). For comparison, retrogradely-labeled neurons in the cortex contralateral to the PTEN deletion were identified in near-adjacent sections that were not immunostained for PTEN. The cross-sectional areas of PTEN negative vs. PTEN positive cell bodies were measured in NIH ImageJ. Figure 9E-I illustrate the cell body size distributions in 5 mice; values of t-tests for cell size data in each individual animal are illustrated in each graph; two way AVOVA of data of individual cell sizes from all 5 cases: F=178.9, p<0.0001, df=392. In Figure 9J, data points are the average size of neurons in the 5 individual animals. The overall average cross-sectional area for FG-positive, PTEN negative neurons was 484.0 ±125.7μm2 vs. 291.3 ±35.1μm2 for FG-positive neurons in the contralateral cortex (paired t-test=5.47, n=5 pairs, p=0.005).
We also prepared tissue from 2 mice that had received vector control injections of AAV-GFP and intra-spinal Fluorogold injections. Examination of GFP immunostained sections from these two cases revealed only a few GFP labeled neurons consistent with the findings above, which precluded meaningful cell size measurements. Thus, we were not able to compare the size of neurons that had been transfected with AAV-GFP with controls.
Casual observation revealed no noticeable behavioral differences between mice with PTEN deletion in the cortex vs. control mice. Mice grew to normal size, appeared as healthy as controls even up to 18 months of age, were active, seizures were not observed, and there was no premature death.
To screen for major deficits in overall motor function of mice with PTEN deletion, motor coordination was assessed by Rotorod on 4 non-consecutive days in mice that had received AAV-Cre vs. AAV-GFP (Figure 10A). Behavioral testing was carried out when the mice were 48 to 53 weeks old. For both groups, the average amount of time spent on the rod increased over the testing period, reflecting motor learning. Mice with PTEN deletion did not achieve the same level of performance as controls, but the difference was not statistically significant (p=0.3764, df= 16, repeated measures two way ANOVA).
Activity in an open field is considered to be a measure of overall locomotor activity, exploratory activity, spatial memory, and anxiety. Decreases in exploration with repeated exposure is considered to be a measure of familiarity with the open field (and thus spatial memory). During the initial exposure to a novel environment, mice typically remain close to the walls of the enclosure (thigmotaxis) (Simon et al., 1994) and with repeated exposure, tend to increase time spent in the center of the open field. To examine these behaviors in our mice, we tested them over four consecutive days. Open field performance was measured in the same mice used for Rotorod assessment, and was carried out when the mice were around 51.6 weeks of age.
Over the testing period the overall activity of both groups declined (measured as a decrease in total number of lines crossed during the testing period) (Figure 10B). Both groups also demonstrated a preference for the perimeter of the arena throughout the testing period (determined by perimeter line crossings), and a progressive decline in the total number of perimeter lines crossed over the four days of testing (Figure 10D). For both PTEN deleted and AAV-GFP injected controls, the change in total lines crossed/day, inner lines crossed/day, and outer lines crossed/day were only significant between day 1 and day 2 (repeated measures two way ANOVA, p=0.0007, df=16 for total lines, p=0.0006, df=16 for inner lines, p=0.0001, df=16 for outer lines). From day 2 to day 4 both PTEN deleted and control animals crossed a stable number of inner lines, illustrating an increase in exploratory behavior across testing sessions (Figure 10C). Comparing the PTEN deleted and AAV-GFP injected controls, there were no significant differences on any of the measures (repeated measured two way ANOVA, p=0.7380, df=16 for total lines crossed/day over four days, p=0.8455, df=16 for perimeter lines crossed/day over four days, p=0.8569, df=16 for inner lines crossed/day over four days).
The present study was undertaken to define the long-term consequences of postnatal deletion of PTEN in the sensorimotor cortex, which previous studies have shown enables cortical motoneurons to regenerate their axons after an adult spinal cord injury (Liu et al., 2010; Zukor et al., 2013; Lewandowski and Steward, 2014; Danilov and Steward, 2015). These findings, together with previous findings in other systems (Park et al., 2008), establish PTEN deletion as an effective way to enable axon regeneration following injury. We show here that conditional deletion of PTEN in the motor cortex of early postnatal mice results in robust hypertrophy of layer V cortical motoneurons and of the cortex itself, decreased cell density in the motor cortex, and maintained mTOR pathway activation a year after PTEN deletion. There was no evident neuropathology, and behavioral assessment revealed normal open field activity and somewhat impaired motor coordination on the Rotarod task that was not statistically significant.
Our results extend previous studies reporting structural and functional phenotypes resulting from promoter-driven genetic deletion of PTEN in early development. Other groups have observed a variety of neuropathologies following the loss of PTEN in multiple brain regions during development including macrocephaly, alterations in cortical lamination, hydrocephalus, seizures, premature death, and autism phenotypes (Backman et al., 2001; Fraser et al., 2008; Fraser et al., 2004; Kwon et al., 2006; Kwon et al., 2001; Sperow et al., 2012; Wen et al., 2013; Yue et al., 2005). Most of these previous studies examined the effects of PTEN deletion after only a few months. Indeed, morbidity and premature death would have precluded longer-term analyses. It is noteworthy that we did not observe any of the major neuropathologies, perhaps because PTEN deletion was postnatal and focal.
Immunostaining for PTEN in mice that received AAV-Cre intracortical injections confirmed PTEN deletion in neurons in layer V that give rise to the CST as well as neurons in other cortical layers. PTEN expression was maintained in some neurons in layers II and IV that had the morphological features of cortical interneurons. Immunostaining for GABA revealed that GABAergic neurons did not account for the PTEN positive neurons within the deletion region.
It is noteworthy that intracortical injections of AAV-PTEN shRNA (shPTEN) in adult rats create a focal region in which there is a complete loss of PTEN immunostaining in all cell types (Lewandowski and Steward, 2014). The same is true of AAV-Cre injections into the cerebral cortex of adult PTENf/f mice (Danilov et al., 2015). Thus, selective preservation of expression in some neuron types following injections at P1 is likely to be due to the stage of development and not differential promoter-driven expression of the AAV in different neuron types.
Selective preservation of PTEN expression in some neurons of mice with neonatal deletion of PTEN may be due to neuron type selectivity of AAV transfection at the age of deletion or a developmentally time sensitive expression of the necessary receptors and co-receptors for AAV transfection. AAV-2 is known to use heparan sulfate proteoglycan, αVβ5 integrin, and the laminin receptor as receptors for transfection (Akache et al., 2006; Summerford et al., 1999; Summerford and Samulski, 1998). Onset of expression of these necessary receptors may vary by cell type so that a subpopulation of neurons are preferentially transfected by AAV-Cre in neonatal mice.
In control mice that received AAV-GFP injections, we did not detect native GFP fluorescence in animals one year after viral injection, but immunostaining for GFP revealed GFP-positive astrocytes and a few neurons. Preferential preservation of expression in astrocytes up to 1 year is noteworthy; previous studies have reported maintained AAV-GFP mediated GFP fluorescence up to 3 months following injection (Chamberlin et al., 1998), but to our knowledge this is the first time that AAV-GFP expression has been found up to 12 months after injection,
While PTEN deletion was successfully achieved in all experimental animals, the size of the deletion region varied between animals. The inconsistency in size may have implications for SCI research projects that utilize conditional PTEN deletion as a therapeutic intervention. If the number of layer V upper motor neurons that have PTEN deleted varies between animals, the regenerative capacity of injured CST axons will vary similarly and appear less robust when averaging data across multiple animals. It may be useful in future experiments to determine the relationship between the area of PTEN deletion and functional recovery measures.
Phosphorylation of ribosomal protein S6 (pS6) is a biomarker of mTOR activation, and immunostains for pS6 following PTEN deletion in adult neurons show increased S6 phosphorylation even a year following deletion. The regenerative effect of PTEN deletion is thought to be due to increased activity of mTOR, and specifically rapamycin-sensitive mTOR complex 1 (mTORC1) (Park et al., 2008). mTORC1 regulates the translation of a subset of mRNAs that contain either a long unstructured 5′ un-translated region (5′UTR), or a 5′UTR with a specific structure (Ma and Blenis, 2009). mTOR regulates the transcription of these transcripts by phosphorylating three downstream targets- the S6 kinases, the inhibitory eIF4E- binding proteins (4E-BPs), and eIF4G initiation factors (Thoreen et al., 2012). It is not yet established how PTEN deletion and activation of s6 phosphorylation actually affects mRNA translation in neurons.
Histological assessment revealed no necrosis, abnormal growths, or tumor formation in areas of PTEN deletion. There was evidence of cortical injury in some mice, with GFAP positive astrocytes bordering the lesion, but this is most likely a result of damage caused at the time of the injections. Since astrocytes are responsible for repairing damage in the central nervous system, this finding is not surprising.
Focal increases in cortical thickness could be related to the increased cell body size of PTEN deleted neurons. Other groups have found marked cortical hypertrophy in mice with PTEN deletion (Fraser et al., 2008; Kwon et al., 2006; Kwon et al., 2001), and in some of our cases, cortical hypertrophy was evident. The laminar disruptions that we observed were more qualitative in nature, in that it was difficult to identify the 6 cortical layers in H&E stains in the PTEN deletion region.
Layer V cells of CST origin with PTEN deletion exhibited considerable cell body hypertrophy. Promoter regulated PTEN deletion has been associated with cell body enlargement of dentate and cerebellar granule cells, dopaminergic cells in the substantia nigra, and cortical neurons (Diaz-Ruiz et al., 2009; Fraser et al., 2008; Kwon et al., 2001). Assessment of the size of PTEN deleted layer V output neurons in the motor cortex had not been previously assessed, and here we demonstrate that layer V neurons are capable of robust growth. Within the motor cortex we found a statistically significant decrease in cell density (the ratio of neuropil to cell bodies) following PTEN deletion. It remains to be determined how neuronal hypertrophy and decreased cortical cell density affect the signaling properties of these neurons or the motor cortex.
Despite the presence of markedly hypertrophied layer V pyramidal cells, and in some cases hypertrophy of the cortex itself, we did not observe any adverse behavioral effects of AAV-Cre injection into the sensorimotor cortex of PTENf/f mice. Casual observations and our limited testing of exploratory activity did not reveal obvious behavioral differences in mice with PTEN deletion, although time spent on the Rotarod was shorter than for AAV-GFP-injected controls. Previous studies of mice in which PTEN deletion was regulated by neuron specific enolase-Cre (NSE-Cre) revealed no difference in initial Rotarod performance and an eventual increase in time spent on the rod later in testing compared to controls (Kwon et al., 2006). The lack of major behavioral consequences seen here may be due to the localized deletion of PTEN in the motor cortex rather than deletion in widespread brain regions that occurs with promoter-driven Cre expression.
Although we did not monitor animals for seizure activity, the animals were handled throughout the duration of the experiment and never exhibited visible signs of motor seizures. Due to previous findings of seizure activity in animals with conditional transgenic promoter-driver PTEN deletion (Backman et al., 2001; Kwon et al., 2001; Ogawa et al., 2007; Pun et al., 2012), the possibility of sub-clinical seizures cannot be excluded without additional studies with electrophysiological monitoring.
The results of the present study provide provisional insights into the potential risks of complete and long-term deletion of PTEN in a subpopulation of neurons, as might occur with a therapeutic intervention to enable axon regeneration. We focused our analyses on the long-term effects of PTEN deletion induced early in postnatal life in order to closely mimic the method used to promote axon regeneration after injury (Liu et al., 2010; Park et al., 2008). The lack of major negative consequences is noteworthy, although it remains possible that some negative consequences were not detected.
It remains to be determined if PTEN deletion performed similarly in adult animals, which has also been shown to enhance axon regeneration and recovery following SCI, will be as benign histologically and behaviorally (Danilov and Steward, 2015; Du et al., 2015). It will also be of interest to explore interactions between the alterations in cortical morphology seen here and responses to axotomy due to spinal cord injury.
Thanks to Sara Jahangiri, Kelly Matsudaira Yee, and Jamie Mizufuka for superb technical assistance. Supported by NIH grant NS047718 (O.S.). We gratefully acknowledge generous donations from Cure Medical, Research for Cure, Stonepole LTD, and Vinstar Development LTD.
Author contributions: OS, EG and MB designed research, EG and MB performed research, EG analyzed data, OS and EG contributed to the writing of the paper.
Conflict of interest: OS is one of the co-founders of the company “Axonis”, which holds options on patents relating to PTEN deletion and axon regeneration.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.