Abstract
BaxΔ2 is a pro-apoptotic protein originally discovered in colon cancer patients with high microsatellite instability. Unlike most pro-apoptotic Bax family members, BaxΔ2 mediates cell death through a non-mitochondrial caspase 8-dependent pathway. In the scope of analyzing the distribution of BaxΔ2 expression in human tissues, we examined a panel of human brain samples. Here, we report 4 cerebellar cases in which the subjects had no neurological disorder or disease documented. We found BaxΔ2 positive cells scattered in all areas of the cerebellum, but most strikingly concentrated in Purkinje cell bodies and dendrites. Two out the four subjects tested had strong BaxΔ2- positive staining in nearly all Purkinje cells; one was mainly negative; and one had various levels of positive staining within the same sample. Further genetic analysis of the Purkinje cell layer, collected by microdissection from two subjects, showed that the samples contained G7 and G9 Bax microsatellite mutations. Both subjects were young and had no diseases reported at the time of death. As the distribution of BaxΔ2 is consistent with that known for Baxa, but in a less ubiquitous manner, these results may imply a potential function of BaxΔ2 in Purkinje cells.
Introduction
The cerebellum is the area of the brain responsible for the modification of motor commands and is involved in balance, posture, voluntary movements, motor learning and more (Koziol et al. 2014; Bernard and Seidler 2014). The cerebellar cortex is composed of three layers, the granular layer, the Purkinje cell layer, and the molecular layer. The dendritic trees from Purkinje cells extend into the molecular layer (Casoni et al. 2017). Recent studies showed that Purkinje cells play a role in information storage for learned response sequences in coordination of motor behaviors (Jirenhed et al. 2017). It also has been shown that Purkinje cells may be involved in the pathophysiology of schizophrenia (Picard et al. 2007; Mothersill et al. 2016).
Bax is a pro-apoptotic Bcl-2 family member, that is ubiquitously distributed throughout all human tissues (Krajewski et al. 1994; Penault-Llorca et al. 1998). Bax plays a critical role in development and in tissue homeostasis, and its dysregulation can lead to many diseases. Due its involvement in the development and progression of neurodegenerative diseases and ischemic insults, expression and distribution of Bax in the brain has been widely studied (Hara et al. 1996; Fan et al. 2001; Vogel 2002; Dorszewska et al. 2004; Jung et al. 2008; Didonna et al. 2012; Garcia et al. 2013). Bax is known to have several functional isoforms, but only the distribution of the canonical isoform, Baxa, has been fully studied. Baxa is present in all the different areas of the brain, mainly in neuronal bodies, with very low to no presence in glial cells (Hara et al. 1996; Vogel 2002; Didonna et al. 2012). Cerebellar Purkinje cells and hippocampal neurons have the highest levels of Baxa (Hara et al. 1996; Vogel 2002; Casoni et al. 2017), which is believed to be the reason why these cells are so vulnerable to ischemic insults (Krajewski et al. 1995). Apart from its apoptotic role, other physiological functions of Baxa in the brain are unknown.
BaxΔ2 is a unique isoform of the Bax subfamily originally discovered in colorectal cancer patients with high microsatellite instability (MSI-H) (Haferkamp et al. 2012; Zhang et al. 2015). It is generally believed that generation of BaxΔ2 requires a microsatellite frameshift mutation in combination with an alternative splicing event that restores the reading frame. Like Baxa, BaxΔ2 is pro-apoptotic and has similar characteristics, such as binding with Bcl-2; however, BaxΔ2 does not target mitochondria, and instead activates a caspase 8-dependent death pathway (Haferkamp et al. 2012; Zhang et al. 2014; Mañas et al. 2017). In the process of analyzing the expression and distribution of BaxΔ2 throughout the human body, we examined several tissue sections of human brain. Here, we report the expression and distribution of BaxΔ2 in the cerebellum of four young and healthy human subjects.
Materials and methods
Materials
All tissue sections and tissue microarray slides were commercially obtained from Biomax. All samples were de-identified and assigned with codes. Antibody against BaxΔ2 was generated previously (Haferkamp et al. 2012).
Immunohistochemistry and tissue analysis
Tissue slides were de-waxed using xylene and rehydrated using graded ethanol solutions. Endogenous peroxidase activity was blocked using 0.3% hydrogen peroxide. Slides were incubated in sodium citrate buffer (0.01 M, pH 6.0) at 95°C for epitope retrieval. After blocking with 5% BSA, CoverWell® incubation humidity chambers were used to incubate the slides with anti-BaxΔ2 antibody (1:100 in blocking buffer) at 4°C overnight and then with Biotin-conjugated Goat anti-mouse secondary antibody (1:200 in PBS) at room temperature for 2 hours. Vectrastain® ABC Kit (Vector Laboratories) and ImmPACT™ DAB Peroxidase Substrate Kit (Vector Laboratories) were used for visualization, and Hematoxolin QS (Vector Laboratories) was used for nuclear staining. Finally, slides were dehydrated and fixed using xylene based mounting media Poly-Mount® (Polysciences Inc.). Fluorescence staining was performed as described for the DAB staining until the step of the primary antibody incubation. Slides were then incubated with Alexa Fluor® 488 donkey anti-mouse IgG (Invitrogen) [1:200] at room temperature for 2 hours. Slides were fixed using ProLong® Gold antifade reagent (Invitrogen). Slides were scanned at the Integrated Light Microscopy Core Facility at the University of Chicago and visualized using Pannoramic Viewer 1.15.2. Each slide was analyzed by three independent viewers and samples with debated evaluation were analyzed by a fourth individual. Each sample was assigned an H-score value based on the intensity and number of positive cells.
Microdissection and Genotyping
The stained tissue sections of cerebellum from Subjects 1 and 3 were uncovered using xylene and dehydrated with graded ethanol solutions. The Purkinje cells layer was microdissected under a microscope using a sterile needle. The tissue was collected and lysed with Proteinase K. DNA was then isolated using AMPure XP magnetic beads (Beckman Coulter). Sanger sequencing was performed at the DNA Sequencing & Genotyping Facility at the University of Chicago Comprehensive Cancer Center.
Results and Discussion
The monomeric form of Baxa is ubiquitously distributed throughout different human tissues, including the brain (Hara et al. 1996). However, distribution of other Bax isoforms is unknown. In the process of screening the tissue distribution of BaxΔ2, we found 4 interesting cases involving brain cerebellar tissues. The patient information about these 4 subjects was summarized in Figure 1A. There were two males and two females, all between the ages of 15 and 35. All subjects were healthy at the time of death and had no tumors or neuronal diseases reported. All tissues were immunohistochemically stained with an anti-BaxΔ2 antibody, which has been shown previously to be very specific for BaxΔ2 and has no cross-staining with parental Baxa. We detected significant amounts of BaxΔ2-positive cells in three out of the four subjects. As shown in Figure 1B, the most significant BaxΔ2-positive cells were found in the Purkinje cell layer, between the granular and molecular layers. Subjects 3 and 4 had strong BaxΔ2 staining in almost all Purkinje cells; Subject 2, on the other hand, had almost no positive staining observed; and Subject 1 had mixed populations of both positive and negative BaxΔ2 cells in close regions.
Further analysis of the positive samples showed that most BaxΔ2-positive staining was detected in the Purkinje cell bodies (Fig. 2A), as well as in the dendrites extending into the molecular layer (Fig. 2D, 2E and 2F). In contrast with the strong positive-stained Purkinje cells, we also observed some weaker positive cells scattered throughout the granular layer (Fig. 2B) and the molecular layer (Fig. 2C). The nature of these cells remains to be determined; however, morphologically, they appeared to be granule cells and Golgi cells in the granular layer, and basket cells (large nuclei and close to the Purkinje cell layer) and stellate cells (smaller nuclei and body size) in the molecular layer. Overall, BaxΔ2 seemed evenly distributed in the cerebellum at low levels, with significantly higher BaxΔ2-positive staining in the Purkinje cells. These results are consistent with published data for the distribution of Baxα, which also presents higher levels in Purkinje cells. However, the amount and distribution of BaxΔ2- positive cells, outside of the Purkinje cell layer, was lower and less ubiquitous than for Baxa (Hara et al. 1996).
BaxΔ2 was previously detected in MSI-H tumors which involve mutations in the Bax microsatellite (Haferkamp et al. 2012). As none of these subjects were reported to have any tumors or neurological conditions, we wondered whether they could have a Bax microsatellite mutation. Due to limited sample size and variation of sample quality, we were only able to genetically analyze two of the four subjects. The tissue slides were first immunohistochemically stained with anti-BaxΔ2 antibody, then the Purkinje cell layer was carefully harvested through microdissection under a microscope, as shown in Figure 3A, and finally genomic DNA was isolated for sequence analysis. A 200 base-pair segment of the Bax gene containing the microsatellite region, was amplified by PCR and subjected to Sanger sequence analysis. As shown in Fig 3B, in comparison with the wild type Bax microsatellite, which contains a stretch of eight guanines (G8), the sample from Subject 1 contains a G9 mutation and the sample from Subject 3 contains a G7 mutation.
It is generally believed that expression of BaxΔ2 requires a Bax microsatellite G7 mutation and an alternative splicing event, which salvages the frameshift caused by the guanine deletion (Haferkamp et al. 2012). The result from Subject 3 seemed consistent with the BaxΔ2-positive staining, as the G7 mutation was detected. However, a G9 mutation was detected for Subject 1, who also had a strong BaxΔ2 staining in the area dissected, similar to that in Subject 3. This cannot be explained by the general rule for BaxΔ2 expression. Although we have previously shown that the G9 mutation can generate another BaxΔ2 sub-isoform, BaxΔ2(G9), which has a similar behavior as BaxΔ2 (Haferkamp et al. 2013), the antibody used here is specific for BaxΔ2 and does not cross react with BaxΔ2(G9). One possibility is that the area collected by microdissection also contained a significant amount of unstained cells with predominant G9 genotype. Another possibility is that the G8-to-G7 deletion could occur at the transcriptional level by transcriptional slippage, or at the translational level by ribosomal frameshift (Ketteler 2012; Atkins et al. 2016). Both transcriptional slippage and ribosomal frameshift are currently under investigation in our laboratory.
Nevertheless, we report the discovery of pro-apoptotic BaxΔ2 proteins in cerebellar neuron cells of young healthy individuals. The similar distribution pattern as canonical Baxa, especially in Purkinje cells, implies a potential physiological function of BaxΔ2 in neurons, independently or in compensation of Baxa.
Compliance with Ethical Standards
The authors declare they have no conflict of interest.
Acknowledgements and Funding
This work was supported by the National Institutes of Health [R15 CA195526].