Background
One of the most important functions that astrocytes perform is buffering the increase in potassium that occurs during neuronal firing to help restore baseline conditions [
1]. Astrocytes buffer excess potassium through different pathways in a still undefined manner: mainly via the Na
+, K
+, ATPase pump, but also using the Na
+, K
+, Cl
− co-transporter, the potassium channel Kir4.1 and through gap-junction dependent processes [
2]. It has also been suggested that the ClC-2 chloride channel may play a role in glial potassium accumulation [
3,
4]. Animal models deficient in proteins involved in this process (Kir4.1, ClC-2, Cx32/Cx47, Cx30/Cx43) show several defects in potassium clearance, increased neuronal excitability and presence of vacuoles in myelin [
5‐
8]. Since water movement is parallel to ion flow, it is possible that vacuoles are a consequence of an impaired ion uptake. Additionally, potassium and water entry into astrocytes also causes cellular swelling. A swelling-dependent chloride channel named VRAC (for Volume-Regulated Anion Channel) strongly expressed in astrocytes is then activated, releasing chloride and osmolytes from the cell, thus changing the driving force for water movement and restoring the astrocyte’s original size [
9].
A similar phenotype to what is present in knockout animals of genes involved in potassium clearance [
5‐
8] has been observed in patients affected with Megalencephalic Leukoencephalopathy with subcortical Cysts (MLC), a rare type of leukodystrophy [
10]. MLC is characterized by astrocyte and myelin vacuolization, epilepsy and early-onset macrocephaly [
11]. The epilepsy and the presence of vacuoles in MLC patients suggested a possible defect in potassium handling [
10]. MLC is caused by mutations in either
MLC1 [
12] or
GLIALCAM [
13].
MLC1 encodes for a membrane protein with eight predicted transmembrane domains (MLC1), which is specifically expressed in astrocytes at cell-cell junctions, including the Bergmann glia of the cerebellum and highly enriched in their perivascular endfeet contacting the blood brain barrier (BBB) [
14,
15]. GlialCAM is an adhesion molecule of the immunoglobulin superfamily expressed predominantly in astrocytes and oligodendrocytes [
15,
16].
The pathophysiological mechanisms leading to MLC are unclear [
17]. Apart from the phenotype of MLC patients, some experimental evidence suggest that GlialCAM/MLC1 have a role in potassium clearance: i) GlialCAM is an auxiliary subunit of the ClC-2 chloride channel [
18]. GlialCAM makes ClC-2 an ohmic channel due to a change in its gating mechanism [
19], which allow mediating chloride influx at depolarized potentials [
15], as expected for a chloride channel involved in potassium uptake; ii) in astrocyte cultures, localization of GlialCAM, MLC1 and ClC-2 at cell-cell junctions depend on extracellular potassium [
20]; iii) mice models deficient for
Mlc1 or
Glialcam display altered brain potassium dynamics [
21] and iv) astrocytes deficient in MLC1 or GlialCAM show reduced VRAC activity [
22‐
24]. Even though this experimental evidence suggested the involvement of MLC1 and GlialCAM proteins in potassium uptake, the molecular basis of these defects is unclear, as the precise functions of MLC1 of GlialCAM are still unknown.
The biochemical relationships between MLC1 and GlialCAM are also not well defined. In cultured cell lines such as HeLa cells, MLC1 cannot reach cell junctions without GlialCAM, whereas GlialCAM expressed alone is located at cell-cell junctions [
25]. In agreement with this in vitro data, mice deficient in
Glialcam show a mislocalization of Mlc1 [
15,
16]. On the other hand, MLC1 expressed alone in cell lines can reach the plasma membrane [
26‐
28], while in
Glialcam knockout mice, Mlc1 is not present at the plasma membrane and Mlc1 protein levels are reduced [
15,
16]. Considering that in primary astrocytes, GlialCAM improves the plasma membrane localization of MLC-related mutants of MLC1 that present folding defects, it has been suggested that GlialCAM has two putative roles: bringing MLC1 at cell-cell junctions and stabilizing MLC1 [
22].
Unexpectedly, both mice [
14,
15] and zebrafish [
29] deficient in MLC1 also show a mislocalization of GlialCAM in astrocytes and oligodendrocytes. However, this mislocalization is observed in Bergmann glia [
29] but not in astrocytes surrounding blood vessels [
25] in humans. Furthermore, in astrocyte cultures from
Mlc1−/− mice, GlialCAM is not mislocalized, but it loses its localization at cell-cell junctions after incubating astrocytes with a depolarizing solution [
29]. According to this, it has been suggested that the mislocalization of GlialCAM when MLC1 is not present depends on the extracellular potassium concentration by an undefined mechanism involving signal transduction processes [
20,
23,
30,
31].
In summary, although MLC1 and GlialCAM proteins form a complex located at cell-cell junctions, the biochemical role of each protein in this complex is not well defined. In the present work, with the aim of understanding this relationship, we have generated and analyzed zebrafish deficient in
glialcama as well as zebrafish and mice deficient in both proteins. Two orthologous genes for GlialCAM have been described in zebrafish (
glialcama and
glialcamb), although previous results suggested that glialcama is the orthologous gene of
GLIALCAM [
29]. The characterization of these models has provided new insights into the molecular basis of GlialCAM and MLC1 interactions.
Methods
Zebrafish maintenance
Zebrafish were kept at the animal facility in Bellvitge Campus, University of Barcelona, under standard conditions at 28 °C, 14 h/10 h light/dark period. AB or AB/TL strains were used in all the experiments. All experimental procedures conformed to the European Community Guidelines on Animal Care and Experimentation and were approved by animal care and use committees.
Generation of glialcama knockout zebrafish
We designed a pair of TALE nucleases to target two sequences at the beginning of glialcama exon1: CTGCTCTCAAGATGAAGGCA (where the start codon is underlined) and TGAAGGAATGGCTGTCTCT, leaving a 20 bp spacer: GAGCGGGAGGCATCATGCAA (BsrBI restriction site underlined). Plasmids containing the TALE nucleases were synthesized by GeneArt (then Life Technologies), and then cloned by Gateway into pCS2-destination vector. Plasmids were linearized with KpnI and mRNAs were synthesized with mMessage mMachine (Ambion). One hundred pg of each TALE Nuclease mRNA were injected into one cell zebrafish embryos, DNA was isolated from pooled embryos at 3dpf and the target sequence amplified with the following primers: GCCCTGAGTGGACAAATCAT and AAACTGACAACAGCGCACAC to check if the BsrBI restriction site was lost due to the action of the TALE nucleases and the subsequent mistakes made by the cellular repair mechanisms. The remaining embryos were raised to adulthood and crossed with wild-type animals. The heterozygosity of their offspring was confirmed by PCR and High Resolution melting Analysis (HRMA) on a StepOne PCR machine (Invitrogen). These F1 embryos were raised to adulthood, tail clipped and genotyped. PCR products were cloned by TA cloning into the pGEMt vector (Promega). Individual colonies were sequenced using T7 and SP6 primers to characterize the mutations generated.
Molecular biology
Plasmids used were constructed using standard molecular biology techniques employing recombinant PCR and the multisite gateway system (Life Technologies). The integrity of all cloned constructs was confirmed by DNA sequencing.
RT-PCR
Adult zebrafish were euthanized using an overdose of tricaine (MS222, Sigma). Adult tissues were quickly dissected and flash-frozen in liquid nitrogen. Total RNA was isolated with TRIzol and retrotranscribed using random hexamers with the SuperScript IV system (Life Technologies). The oligonucleotides pairs used for qPCR are the following: Rpl13a (internal control), sense: TCTGGAGGACTGTAAGAGGTATGC, anti-sense: TCTGGAGGACTGTAAGAGGTATGC; mlc1, sense: GCACGTTCAGTGGACAACTG, anti-sense: CACAATCATTGGGCCTTCAG; glialcama, sense: CCCACCCACCAAGACTAAGC, anti-sense: CATCCTCAGTCGTGCTCATCTG; glialcamb, sense: AGACCGGATCTTGGTGTTTGA, anti-sense: TAGGCTCATCCACAGTGAGATTGA.
qPCR was performed with SYBR Select reagent (Life Technologies) in a StepOne apparatus (Life Technologies). Three experiments were analyzed, with three replicate samples in each experiment. The expression levels were normalized using the comparative Ct method normalized to the internal control genes. The final results were expressed as the relative messenger RNA (mRNA) levels as indicated in the corresponding figures, taking into account the efficiency of each primer with the Pfaffl method.
Histological staining methods in zebrafish
Fish were deeply anesthetized in 0.1% tricaine methanosulfonate (Sigma, MS-222) in fresh water and fixed by vascular perfusion with 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Fish heads were post-fixed in the same fixative for at least 24 h at room temperature. Next, brains and eyes were extracted, cryopreserved in 30% sucrose in PB, frozen with liquid-nitrogen-cooled methylbutane and cut in a cryostat. Transverse sections (12–14 μm thick) were collected onto gelatinized slides.
For immunohistochemistry, sections were rinsed in saline phosphate buffer (PBS) and sequentially incubated at room temperature with: (1) normal goat serum (NGS, Sigma, 1:10 in PBS) for 1 h; (2) primary antibody or cocktail of primary antibodies, overnight (for antibodies and dilutions, see below); (3) PBS for 15 min; (4) secondary fluorescent antibody or cocktail of fluorescent antibodies for 1 h (for antibodies and dilutions, see below); (6) PBS for 15 min. Incubations with primary and secondary antibodies were made at room temperature in a humid chamber. Finally, sections were mounted using 50% glycerol in PB.
Primary antibodies and dilutions used in the study were: rabbit anti-zebrafish mlc1 (1:100) and rabbit anti-zebrafish glialcama (1:100). The secondary antibody used was goat anti rabbit- Alexa Fluor 488 (Invitrogen, 1:500). All dilutions were done in 10% NGS in PBS. Negative controls omitting incubation with primary antibody were performed, showing no unspecific immunoreactivity.
Sections were first observed in a Nikon Eclipse Fluoresencent microscope and then selected sections of were imaged in a Nikon A1R confocal microscope. Confocal and fluorescent data was processed and analysed using ImageJ software.
MRI imaging in zebrafish
Magnetic resonance microimaging (μMRI) of Zebrafish was performed on a vertical wide-bore 7 T Bruker Avance 300WB spectrometer, with a 1000 mT·m− 1 actively shielded imaging gradient insert (Bruker Biospin GmbH, Germany). The system was interfaced to a Linux PC running Topspin 2.0 and ParaVision 3.2 software (Bruker Biospin GmbH, Germany). For RF excitation and detection, a birdcage radio-frequency (RF) coil with an inner diameter 10 mm was used. For μMRI, adult zebrafish were euthanized and fixed in 4% buffered paraformaldehyde (Zinc Formal-Fixx, ThermoShandon, UK) for 7 days and subsequently embedded in Fomblin (Solvay Solexis, Inc.) to avoid any artefacts that may arise due to magnetic susceptibility differences at air–tissue boundaries. The magnetic field homogeneity was optimized by shimming before each μMRI measurement. For position determination and selection of the desired region, each session of measurements began with a multislice orthogonal gradient-echo sequence. Subsequently, high resolution T2 weighted images were acquired by using a rapid acquisition with relaxation enhancement (RARE) sequences with repetition time (TR) = 3000 ms; effective echo time (TE) = 18 ms; RARE factor = 4; slice thickness 0.2 mm; field of view 1.2 × 1.2 mm; image matrix of 256 × 256 pixels, resulting in a spatial resolution of 47 μm.
For transverse relaxation time (T2) measurement, a standard multi-slice multi-echo (MSME) sequence was used. This sequence is based on the Carr-Purcell Meiboom-Gill (CPMG) sequence, where transverse magnetization of a 90° pulse is refocused by a train of 180° pulses generating a series of echoes. The following imaging parameters were used: nominal flip angles = 90° and 180°, and a train of 12 echoes with TEs ranging from 8.17 ms to 98 ms with 8.17 ms echo-spacing; TR = 2 s, slice thickness 0.5 mm; number of slices 8 and a matrix size 256 × 256 pixels.
For calculation of T2 relaxation time, regions of interest (ROIs) were drawn at various locations within the zebrafish brain using an image sequence analysis (ISA) tool package (Paravision 5, Bruker). Another ROI in the muscle was used as an internal control. Monoexponential fitting was then used to calculate T2 using a monoexponential fit function [y = A+ C*exp. (−t/T2)], where A = Absolute bias, C = signal intensity, T2 = transverse relaxation time. Means and standard deviation for T2 relaxation times for each ROI were calculated.
For measurement of brain areas, the desired telencephalone and whole brain regions were drawn on the image and areas were computed using an image sequence analysis (ISA) tool package (Paravision 5, Bruker). The data were exported to OriginPro v. 8 (OriginLab, Northampton, MA, USA) for further analysis and percentage of Telencephalon with respect to whole brain area was calculated. One-way ANOVA (Bonferroni’s post-test) for comparison of mean between each group was performed. Levene’s test was performed for homogeneity of variance analysis.
Mouse studies
The generation of
Glialcam−/− and
Mlc1−/− mice has been previously described [
15]. For histological analyses of brains, mice were perfused with 4% PFA/PBS and organs were postfixed overnight. Haematoxylin–eosin staining was performed on 6 μm paraffin sections of brains.
Mouse primary astrocyte cultures were prepared from cortex and hippocampus, which were removed from newborn mice. Astrocyte cultures were prepared from 0 to 1 day old OF1 mice. Cerebral cortices were dissected and the meninges were carefully removed in cold sterile 0.3% BSA, 0.6% glucose in PBS. The tissue was trypsinized for 10 min at 37 °C and mechanically dissociated through a small bore fire-polished Pasteur pipette in complete DMEM medium (Dulbecco’s Modified Eagle’s Medium with 10% heat-inactivated fetal bovine serum (Biological Industries), 1% penicillin/streptomycin (Invitrogen) and 1% glutamine (Invitrogen) plus 40 U/ml DNase I (Sigma)). The cell suspension was pelleted and re-suspended in fresh complete DMEM, filtered through a 100-μm nylon membrane (BD Falcon) and plated into 75 cm2 cell culture flasks (TPP). When the mixed glial cells reached confluence, contaminating microglia, oligodendrocytes and precursor cells were dislodged by mechanical agitation and removed. Astrocytes were plated in 6-well plates, at density of 4·105 cells per well, or in poly-D-lysine-coated cover slips at 7.5·104cells in 24-well plates. Medium was changed every 3 days. In order to obtain astrocyte cultures arrested in the cell cycle, medium was replaced and cytosine β-D-arabinofuranoside (AraC, Sigma) (2 μM) was added. Cultured astrocytes were identified by their positive GFAP (Glial Fibrillary acid protein) staining (Dako), being > 95% of cells GFAP positive.
For Western blot studies, astrocyte lysates were prepared by homogenization of cells in PBS containing 1% Triton X-100 and protease inhibitors: 1 μM Pepstatin and Leupeptin, 1 mM Aprotinin and PMSF, incubated for 1 h at 4 °C and centrifugated. Supernatants were quantified using BCA kit (Pierce) and mixed with SDS loading sample buffer. After SDS PAGE, membranes were incubated with primary antibodies: anti-MLC1 (1:100), anti-GlialCAM (1:100) and anti-β-Actin (1:10000, Sigma) and secondary antibodies: HRP-conjugated anti-rabbit and anti-mouse (1:10000; Jackson). Quantification of Western blots was performed by ImageJ at different exposition times to ensure linearity.
Discussion
In this work, we have obtained and characterized a
glialcama knockout in zebrafish. The knockout displays megalencephaly and fluid accummulation, indicating that glialcama and not glialcamb, is the functional ortholog gene of GlialCAM in zebrafish. We do not know which could be the role of glialcamb in zebrafish. However, in vitro studies suggest the possibility that it may act as a negative regulator of MLC1 and ClC-2 [
29,
32]. Taking into account that overexpression of MLC1 has been reported to be toxic in mice [
33], there could be regulatory mechanisms inhibiting MLC1 function, such as interaction with glialcamb in zebrafish, although experimental evidence to support this hypothesis is lacking.
We also show that additional disruption of mlc1 in
glialcama knockout zebrafish or in
Glialcam knockout mice does not potentiate the vacuolating phenotype characteristic of MLC disease, indicating that loss-of-function mutations in these genes cause leukodystrophy through a common pathway. Previous [
13] and recent [
11] reports indicate that the phenotype of patients with mutations in
MLC1 is the same to those with recessive mutations in
GLIALCAM. Thus, this genetic evidence in humans, together with biochemical studies in mice and zebrafish models of the disease and in vitro studies that indicated GlialCAM and MLC1 interaction, indicate that these proteins need to form a complex to carry out their physiological role. The situation is completely different for the ClC-2 protein. First, genetic evidence indicates that defects in
MLC1 or
CLCN2 lead to different diseases [
34]. Second, the vacuolating phenotype of
Clcn2−/− mice increased after additional disruption of
Glialcam [
15]. Thus, we proposed that defects in ClC-2 might contribute partially to the MLC phenotype, but it is not the only reason to explain the phenotype of MLC patients.
The fact that the MLC1/GlialCAM complex is a functional unit is evident in the zebrafish knockout for
glialcama, in which mlc1 protein is neither reduced nor mislocalized but yet it displays an MLC-like phenotype. In clear contrast, lack of
Mlc1 in mice or
mlc1 in zebrafish causes GlialCAM and glialcama mislocalization, respectively. Surprisingly, this localization defect could only be observed in primary cultured astrocytes from mouse after incubation with a depolarizing solution [
29,
30]. Possibly, the mislocalization of GlialCAM when MLC1 is absent is a consequence of an unknown depolarization-dependent regulatory mechanism.
We speculate that mlc1 protein levels and localization in zebrafish are unaltered in the glialcama−/−, because in the zebrafish knockout there is an up-regulation of mlc1 mRNA, which does not occur in the Glialcam knockout mice. In agreement with this hypothesis, in primary Glialcam−/− astrocytes, where endogenous MLC1 is mislocalized, zebrafish or human MLC1 overexpressed are located at cell-cell junctions, suggesting that perhaps MLC1 overexpression compensates for lack of GlialCAM stabilizing effect.
Unlike in astrocytes, however, MLC1 overexpressed in cell lines without GlialCAM is never located at cell-cell junctions [
25]. Possibly, in astrocytes, MLC1 may reach cell junctions not only by its interaction with GlialCAM, but also with the help of other proteins that may not be present in non-astrocyte cell lines.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.