Abstract
Autism spectrum disorder, schizophrenia, and bipolar disorder are neuropsychiatric conditions that manifest early in life with a wide range of phenotypes, including repetitive behavior, agitation, and anxiety ( American Psychological Association, 2013). While the etiology of these disorders is incompletely understood, recent data implicate a role for mitochondrial dysfunction ( Norkett et al., 2017; Khaliulin et al., 2025). Mitochondria translocate to intracellular compartments to support energetics and free-radical buffering; failure to achieve this localization results in cellular dysfunction ( Picard et al., 2016). Mitochondrial Rho-GTPase 1 (Miro1) resides on the outer mitochondrial membrane and facilitates microtubule-mediated mitochondrial motility ( Fransson et al., 2003). The loss of MIRO1 is reported to contribute to the onset/progression of neurodegenerative diseases, including amyotrophic lateral sclerosis, Alzheimer's disease, and Parkinson's disease ( Kay et al., 2018). We have hypothesized that MIRO1 also has a role in nervous system development ( Lin-Hendel et al., 2016). To test this, we ablated Miro1 from cortical excitatory progenitors by crossing floxed Miro1 mice with Emx1-Cre mice and studied mice of both sex. We found that mitochondrial mislocalization in migrating excitatory neurons was associated with reduced brain weight, decreased cortical volume, and subtle cortical disorganization. Adult Miro1 conditional mutants exhibit agitative-like behaviors, including decreased nesting and abnormal home cage activity. The mice exhibited anxiety-like behavior and avoided confined spaces, features that have been linked to several human behavioral disorders. Our data link MIRO1 function with mitochondrial dynamics in the pathogenesis of several neuropsychiatric disorders and implicate intracellular mitochondrial dynamics to several anxiety-like behaviors.
Significance Statement
Neuropsychological disorders such as autism spectrum disorder, schizophrenia, and bipolar disorder have overlapping endophenotypes. While the mechanisms underlying these disorders are poorly understood, recent evidence implicates mitochondrial dysfunction and cellular mislocalization playing a role. Mitochondria support energy requirements and other physiological functions in cells. Previous research from our lab has shown distinct dynamic localization patterns within migrating excitatory and inhibitory neurons during development. To further examine the importance of mitochondrial localization, we ablated MIRO1, a protein important for coupling mitochondria to motor proteins, in excitatory neurons. The mislocalization of mitochondria in migrating excitatory neurons is associated with diminished motor skills and anxiety-like behavior in postnatal mice.
Introduction
Neuropsychiatric disorders comprise a heterogeneous group of complex conditions affecting as many as one in five individuals in the USA (National Institute of Mental Health, 2023). Their etiologies remain heterogeneous and poorly understood, though many appear to originate prenatally (Chan, 2006). These disorders can manifest early in life with a wide range of behavioral phenotypes that may include repetitive behavior, agitation, resistance to touch, psychosis, and anxiety, among a variety of other features (Miyoshi et al., 2010).
While the etiology of some neuropsychiatric disorders is believed to be multifactorial, recent data suggest mitochondrial dysfunction could play a role in both syndromic and nonsyndromic cases (Pei and Wallace, 2018). Population-based studies of individuals with a neuropsychiatric disorder indicate a prevalence of mitochondrial dysfunction as high as 80% in disorders such as autism spectrum disorder, implicating a role for mitochondria during brain development (Oliveira et al., 2005; Giulivi et al., 2010; Rossignol and Frye, 2012; Siddiqui et al., 2016). Mitochondria play roles in neuronal migration, circuit formation, synaptogenesis, and plasticity, and these processes have been implicated in the development of some neuropsychiatric disorders (Chan, 2006; Rahman, 2012; Lin-Hendel et al., 2016; Fame and Lehtinen, 2021). We have postulated that mitochondrial dysfunction would affect developmental processes, resulting in postnatal behavior phenotypes.
Mitochondria are double-membrane bound organelles essential for normal cellular physiology. They produce ATP, buffer ion homeostasis, assist with free-radical elimination, support autophagy, and assist in lipid metabolism (Nunnari and Suomalainen, 2012; Mahadevan et al., 2021). To fulfill these various roles, mitochondria must shuttle to different intracellular compartments. Mitochondrial Rho-GTPase (Miro1) is a motor adaptor protein that links mitochondria to molecular motors for transport along microtubules (Tang, 2015; López-Doménech et al., 2018). In addition to transportation, MIRO1 is known to help coordinate mitochondrial movement, fission, fusion, and mitophagy (Hsieh et al., 2016). In mature neurons, mitochondrial movement is essential for maintenance of dendritic branches, synapse formation and maintenance, and proper synaptic transmission (López-Doménech et al., 2016). The loss of Miro1 has been associated with neurodegenerative diseases such as Parkinson's disease, Alzheimer's disease, and amyotrophic lateral sclerosis (Zhang et al., 2015; Grossmann et al., 2019, 2020). Furthermore, MIRO1 has been shown to interact with PINK1, Parkin, α-synuclein, and LRRK2 to mediate mitophagy, preserving mitochondrial quality in neurons (Lin and Sheng, 2015; Shaltouki et al., 2018; Grossmann et al., 2020). Dysregulation of these protein interactions has been implicated in cerebral cortical, brainstem, and hippocampal neuronal dysfunction and degeneration resulting in spasticity, weakness, and memory loss (Nguyen et al., 2014).
In addition to its role in mature neurons and neurodegeneration, recent data implicate Miro1 in the onset of several neuropsychiatric diseases. For example, Miro1-deficient cortical neurons fail to develop normal dendritic arbors (López-Doménech et al., 2016), and MIRO1 seems to form a complex with Disrupted-in-Schizophrenia-1 (DISC1), a protein that is important for neurite outgrowth and has been associated with a wide array of neuropsychiatric disorders (Ogawa et al., 2014; Norkett et al., 2017). A sequence variant of DISC1, R37W, has been associated with individuals diagnosed with schizophrenia, depression, and anxiety in several families and was found to perturb anterograde mitochondrial transport in neurons (Ogawa et al., 2014).
Given the relationship between neuropsychiatric disorders and cerebral cortical neuronal dysfunction (McManus and Golden, 2005; Rubenstein, 2011), we hypothesized that disruption of Miro1 in developing excitatory neurons would result in behavioral disorders. The conditional abrogation of Miro1 from projection neurons resulted in agitative-like behaviors including resistance to touch, decreased nesting behavior, repetitive behaviors, and decreased interactions with littermates. The mice also demonstrated abnormal behavior during open field testing and other anxiety-like testing. Together, these data suggest Miro1 is required for the normal development of excitatory neurons and provide novel insights into the underlying pathogenesis of neuropsychiatric diseases that include anxiety-like behaviors.
Materials and Methods
Mice
Mice were housed and sustained in the Animal Vivarium at Hamilton College (studies were initiated at Brigham and Women's Hospital) and were given food and water ad libitum. All experiments were approved by the Institutional Care and Use Committee at Hamilton College or Brigham and Women's Hospital, respectively. Miro1(f/f) mice (Strain #031126, The Jackson Laboratory) and Emx1-Cre(+/+) mice (Strain #005628, The Jackson Laboratory) were maintained on a C57/Bl6 background. Miro1(+/+), Miro1(f/+), Miro1(f/f), Miro1(+/+); Emx1-Cre(+/+), Miro1(f/+); Emx1-Cre(+/−), and Miro1(+/+); Emx1-Cre(+/−) mice were littermates of the Miro1 conditional mutant mice and used as controls for experiments. Mice of either sex were used for experiments.
Weight measurements
Mice were weaned on postnatal day 21 (P21) and fed a lab diet (5058, LabDiet). All mice were genotyped and weighed every 7 d following weaning for 5 weeks. The percentage difference in weight between controls and conditional knock-out (CKO) animals was calculated by dividing the average weight of the Miro1CKO animals by the average weight of the control animals during that week. Adult mice (control and CKO) were weighed between 6 and 10 months of age to examine weight differences in adult animals.
Food intake
During cage changes, food was weighed out for individual mice for 1 week and added to the food trough. At the end of the week prior to the next cage change, the remaining food in the trough was weighed to establish food intake.
Daily cage recordings
Control and Miro1CKO mice were singly housed in clear (18.5 cm × 21 cm) cages. Red light was used to visualize the mice without disrupting their light/dark cycles. Cameras (WiFi Indoor Camera G7, Galayou Smart Home Security Cameras) were positioned on the top of each cage to monitor movements. Each cage was recorded during the dark cycle when mice are most active. ANY-maze software (Stoelting) was used to track each prerecorded movie for 2 h between 12 A.M. and 2 A.M. during the dark cycle.
Hindlimb footprint pattern test
Receipt paper was placed on and secured to the floor in a narrow hallway (110 cm × 10 cm). The hindlimb feet of adult mice were brushed with India Ink (Speedball) using a paintbrush. The mice were placed on the receipt paper in the narrow hallway and allowed to walk one length of the hallway. A series of three consecutive footprints for each hindlimb were measured (adopted from Crawley, 2000).
Forelimb grip strength test
A T-shaped pull bar was attached to a force meter (Ametek Chatillon DFS II, 10 N) fixed to a table. Mice were held by the tail and allowed to grasp the pull bar with their forelimbs. Once grip was established, the mouse was gently pulled by the tail until they released the bar (adopted from Crawley, 2000). The force was recorded (in Newtons) on three consecutive trials with each mouse.
Vertical pole test
A wooden cylindrical beam of 92.2 cm in length and 1.8 cm in diameter was used to assess gross motor control of body muscles utilized in hanging onto the thick beam as opposed to grip strength alone (adopted from Crawley, 2000). The wooden beam was fixed at one end in the middle of a large 1.02 × 1.02 m rodent testing container. The mouse was placed upon the wooden beam which was then slowly raised from one end until an angle of 90° was reached. Upon reaching 90°, the time to fall off the pole was measured (maximum 60 s). Mice that were unable to stay on the beam during the elevation phase were given a 2 min rest and then retested. If a mouse was unable to reach 90° of elevation during the retest, the final degree of elevation was recorded.
Hanging wire test
Mice were placed on top of a 26 cm by 36 cm wire grid that had 1.2 cm2 square holes. The wire grid was lightly shaken horizontally for the mice to grasp the wire with all paws, then flipped 180°, and held 20.0 cm high from a testing area. The mice were then timed while inverted with a maximum threshold of 60 s (adopted from Crawley, 2000).
Wire grid test
The wire grid used for the hanging wire test was placed on top of a box and a camera placed on the floor of the box to record the limb and paw placement throughout the test. The mouse was placed on the wire grid, and then a clear plastic container was used to restrict the mouse to the wire grid area. Each mouse was tested for 5 min on the wire grid (adopted from Crawley, 2000). The video was analyzed by counting the number of times a paw lost grip and fell through the wire grid.
Tests for anxiety-like behaviors
Before each anxiety-like behavior test, cages were transported from the animal facility to the behavior room and allowed to acclimate to the new environment for 15 min. Tests were completed under normal room lighting and with general room background noise. At the conclusion of each trial, the apparatus was cleaned with a 70% ethanol solution.
Open field test
Adult mice were placed in the center of a (30 cm × 40 cm) arena and allowed to travel freely for 5 min (adopted from Crawley, 2000). The arena was divided into a central region and an outer region to distinguish locations within the box. Movements around the box were tracked using the ANY-maze software package (Stoelting). Distance, speed, time mobile, freezing episodes, center entries, and distance traveled in the center were exported as measurements from the ANY-maze software. A freezing episode was defined as no movement except for respiration for a period of 250 ms.
Elevated plus maze
Adult mice were placed in the center of the elevated plus maze at the start of the trial and allowed to explore the maze for 5 min (adopted from Crawley, 2000). The ANY-maze software package (Stoelting) was used to label each wing of the elevated plus maze as an “open” or “closed.” The time and distance traveled in the open and closed arms were saved from the software.
Wide/narrow box test
Developed as a measure of “claustrophobia,” a 60 cm × 60 cm box was constructed and divided into two 30 cm sections: a wide section (60 cm width × 30 cm length) and a narrow section (5 cm width × 30 cm length; adopted from El-Kordi et al., 2013). The wide section of the box was lit to 300 lumens (lx) and the narrow section of the box lit to 150 lx. Mice were placed in the wide section of the box facing the narrow section to begin the trial and were allowed to roam freely in the box for 10 min. The ANY-maze software (Stoelting) was used to track the path of the mice in the box and measure the time spent in the narrow and wide sections of the box.
Plasmid construction
Separate fragments of EGFP, mito-DsRed2 (TaKaRa, catalog #632421), and P2A derived from porcine teschovirus-1 2A (Kim et al., 2011) were amplified by PCR. The fragments were subcloned into pCAG (Addgene) by Geneart (Thermo Fisher Scientific). The mito-targeting sequence “SVLTPLLLRGLTGSARRLPVPRAKIHSL” was derived from the precursor of subunit VIII of human cytochrome C oxidase (Rizzuto et al., 1989).
Electroporation and slice culture experiments
Mouse embryos were harvested on embryonic day (E) 14.5. Brains were immediately removed and placed into cold complete HBSS bath (Tucker et al., 2006; Lysko et al., 2014). The pCAG-GFP-P2A-mitoDSred plasmid was micropipetted into both lateral ventricles and electroporated into the ventricular zone of the cerebral cortex (Nepa Gene CUY21 electroporator; 45 V; pulse interval, 100 ms; pulse duration, 100 ms; number of pulses, 4). Brains were embedded in 3% low melting point agarose (Fisher Scientific, catalog #BP165-25) and cut into 300 μm sections (Leica VT 1000S). Slices were placed on transwell inserts (BD Biosciences, catalog #353102) coated with laminin/poly-ʟ-lysine and cultured for 3 d (Polleux and Ghosh, 2002). Slices were fixed on the third day with 4% paraformaldehyde.
Immunohistochemistry
Adult mice were anesthetized and perfused with 1× PBS followed by 4% paraformaldehyde. Brains were harvested, postfixed in 4% paraformaldehyde for 24 h, and then placed in 30% sucrose to prepare for cryosectioning. Sixty-micrometer brain sections were cut on a freezing microtome (Leica SM 2000R). Embryonic brains were harvested at E13.5 or E15.5, fixed in 4% paraformaldehyde, and then placed into 10% sucrose overnight, followed by 30% sucrose the next day for 24 h for cryosectioning. Brains were sectioned at 15 μm on a cryostat (Leica CM1950 or Dakewe 6250). Sections were blocked with 5% normal goat serum. Primary antibodies included Rabbit anti-CUX1 (1:250, Proteintech, catalog #11733-1-AP), Rat anti-CTIP2 (1:250, Abcam, catalog #ab18465), Rat anti-KI67 (1:500, Thermo Fisher Scientific, catalog #14-5698-82), Rabbit anti-activated Cleaved Caspase 3 (1:300, Cell Signaling, catalog #9664), Chicken anti-GFAP (1:1,000, Invitrogen, catalog #PA1-10004), and Rabbit anti-TOMM20 (1:500, Invitrogen, catalog # PA5-52843). Alexa Fluor secondary antibodies (Invitrogen, Goat anti-Chicken Alexa Fluor 488, catalog #A-11039; Goat anti-Rabbit Alexa Fluor 488, catalog #A-11008; and Goat anti-Rat Alexa Fluor 555, catalog #A-21434) were used for detection along with DAPI labeling. Labeled sections were imaged on a Leica SP5, Zeiss 910 Confocal Microscope, or Leica DM6 B Microscope and counted using Adobe Photoshop. Cortical areas were measured using ImageJ.
To detect oligodendrocytes, 60 μm microtome sections underwent antigen retrieval for 20 min using L.A.B. Solution (Polysciences; 24310-500), followed by permeabilization with 0.3% Triton X-100 for 30 min. Sections were blocked for 1 h at RT in 10% normal goat serum and then incubated overnight with Rabbit anti-ASPA (GeneTex; GTX110699; 1:500) in 2% normal goat serum for 24 h. After rinsing in 1× PBS, sections were incubated with FITC goat anti-rabbit secondary antibody (Jackson ImmunoResearch; 111-095-003; 1:200) for 1 h at RT. Sections were rinsed in 1× PBS prior to being counter stained with DAPI (Thermo Fisher Scientific; D1306; 1:10,000) for 5 min and mounted in ProLong Gold anti-fade reagent (Thermo Fisher Scientific, P36930). Images were acquired on a DMi8 Leica inverted confocal microscope (Leica Microsystems). Quantification of ASPA+ mature oligodendrocytes was performed using five 300 × 300 μm regions of interest from the cortex across two to three sections from each animal. ASPA+/DAPI+ and ASPA−/DAPI+ were counted using ImageJ and the percentage of mature oligodendrocytes was calculated.
All immunohistochemical analyses were conducted blind to the treatment groups.
Analysis of mitochondria location within cell body during corticogenesis
The somas of migrating excitatory neurons were segmented, and vectors were calculated with the Aivia software (Leica Microsystems). A spherical coordinate system with Cartesian coordinates was used to measure where the mitochondria labeled with MitoDSred were localized within the cell body.
The following scripts in the measurement component of the Aivia software were used to measure the angular location of the mitochondria within the soma using the azimuth, the angle within the x–y plane:
X Displacement: ValueAtFrame([Centroid X (Mitochondria)], 2) - ValueAtFrame([Centroid X (Mitochondria)], 1), Y Displacement: ValueAtFrame([Centroid Y (Mitochondria)], 2) - ValueAtFrame([Centroid Y (Mitochondria)], 1), Z Displacement: MaxOverTime([Centroid Z (Mitochondria)]) - MinOverTime([Centroid Z (Mitochondria)]), Azimuth: Atan2([Y Displacement (Mitochondria)], [X Displacement (Mitochondria)]) * (180/PI()).
Analysis of mitochondria in adult cortices
Control and Miro1CKO cortices were regionally imaged in layer 5/6 in a rostral and caudal location. Images were thresholded in ImageJ, and areas (in pixels) of all particles were analyzed from the thresholded image. Particles were organized by the frequency of areas.
Statistical analysis
A Student's t test was used to compare control and Miro1CKO groups in our brain and behavior analyses. Two-way analysis of variance (ANOVA) followed by Fisher’s LSD post hoc analysis was used to analyze the bin distributions of cells in the cortex. The Watson–Wheeler test was used to compare circular data for mitochondrial localization within the cell body.
The longitudinal data of behavior components from the elevated plus maze and open field tests were analyzed using a rank-based nonparametric method (Noguchi et al., 2012) to examine treatment group (Miro1CKO vs control), time (10–300 s), and their interaction effects on individual behavior outcomes. For speed data from the elevated plus maze, missing data were imputed by the mean value of non-missing data at given timepoint in each group. Analyses were performed using R version 4.2.3 (R Core Team, 2023; circular and nparLD packages) with two-sided tests at a significance level of 0.05.
Results
Miro1CKO mice have a reduced body and brain size
Miro1f/f mice were crossed with Emx1-Cre+/+ to generate Miro1+/−;Emx1-Cre+/− that were then crossed to generated Miro1−/−;Emx1-Cre+/− (subsequently referred to as Miro1CKO or CKO) and littermates with the genotypes listed in the Materials and Methods. By weaning, Miro1CKO mice were noticeably smaller than littermate control mice, weighing 22.6% less (Fig. 1A). By 8 postnatal weeks, the difference was less pronounced (11.9%) with a 5% reduction in average weight for adult Miro1CKO mice (Fig. 1B). To determine if food intake accounted for the weight reduction, the average daily weight of food consumed was measured. Over 7 d, there was no statistical difference in food consumptions (31.2 vs 32.8 g for control and Miro1CKO mice, respectively). Miro1CKO mice were also found to live a full lifespan of 1.5 years or longer.
Miro1CKO body and brain size are small. A, Miro1CKO mice (right) are smaller than their littermate controls (left). B, Quantification of difference in weight between controls (weaning age: n = 9; 4 females, 5 males, adult: n = 14; 8 females, 6 males) and Miro1CKO mice (weaning age: n = 12; 5 females, 7 males, adult: n = 11; 3 females, 8 males). C, Size differences between control and Miro1CKO brains. D, Quantification of brain weight (Control n = 5; 2 females, 3 males and Miro1CKO n = 5; 3 females, 2 males, t(8) = 2.824, p = 0.0224, Student's t test). E, Quantification of cortical area at rostral, middle, and caudal locations (n = 5, rostral: t(8) = 7.728, p = 0.0002, middle: t(8) = 12, p < 0.0001, caudal: t(8) = 6.801, p = 0.0005, Student's t test). F, G, Example of cortical, ventricular, and fornix size differences between control and Miro1CKO mouse. H, Expression of Ai14 reporter line indicating CRE expression and hippocampal size differences in control and Miro1CKO. I, Quantification of rostral hippocampal area (Control n = 4; 1 female, 3 males and Miro1CKO n = 4; 2 females, 2 males, t(6) = 4.937, p = 0.0026, Student's t test). n.s. p > 0.05, *p ≤ 0.05, **p ≤ 0.01.
In addition to a reduced body size, Miro1CKO brains were also smaller and weighed less than littermate control brains (Fig. 1C,D). The cortical volume was found to be significantly reduced, and the ventricles enlarged in adult Miro1CKO brains (Fig. 1E–G). The hippocampi also appeared smaller, and after confirming that CRE is expressed in the hippocampus of Emx1-Cre mice (Fig. 1H), we found a significant reduction in the hippocampal volume of Miro1CKO mice compared with control mice (Fig. 1H,I).
Mitochondria are mislocalized in migrating and mature Miro1CKO mice
To establish the intracellular mitochondrial distribution, pCAG-EGFP-P2A-mitoDSred was electroporated into control and Miro1CKO mouse forebrains on E14.5, and the brains were harvested and slice cultures established. After 3 d in culture, the mitochondria localization was assayed in migrating excitatory neurons. Control neurons displayed their expected mitochondria localization, proximal to the nucleus in the direction of the leading process as we previously described (Fig. 2A,C, top graph; Lin-Hendel et al., 2016). In contrast, the mitochondria in Miro1CKO neurons were predominately mislocalized to the rear of the migrating cells (Fig. 2B,C, bottom graph).
Emx1-Cre mediated ablation of Miro1 leads to mislocalization of mitochondria in migrating excitatory neurons and adult cortices. A, B, Migrating excitatory neurons (GFP) and labeled mitochondria (MitoDSred) in E14.5 control and Miro1CKO cortices. C, Quantification of red intensity (MitoDSred) location in the cell bodies of migrating excitatory neurons (Control, top graph, n = 265 cells, Miro1CKO, bottom graph, n = 211 cells; p < 0.001; Watson–Wheeler test for significance). D, E, Mitochondria labeled with anti-TOMM20 in layer 5/6 of adult control and Miro1CKO cortices.
To corroborate that mitochondria continue to be mislocalized in adult cortices, an anti-TOMM20 antibody was used to label mitochondria. Control cortices revealed small, distinct mitochondria that were dispersed in cell bodies (Fig. 2D). In contrast, Miro1CKO mitochondria were clustered (Fig. 2E, arrows).
Miro1CKO cortices have increased cell death during corticogenesis
To investigate if the reduction in cortical volume was due to a cell cycle defect during development or increased cell death, Miro1CKO cortices were labeled with KI67, phosphohistone H3 (pH3), and Activated Cleaved Caspase 3 (CC3) antibodies. At E13.5 and E15.5, KI67 and pH3 labeling revealed no significant difference in the percentage of cycling cells between controls and Miro1CKO (Fig. 3A–D,E–H). CC3-labeled cells were evaluated at P4 during a period of programmed apoptosis of excitatory neurons (Wong and Marín, 2025). The number of CC3-labeled cells was trending toward significance in the rostral location and significantly increased in the caudal location (Fig. 3I).
Miro1CKO's have increased cell death at P4 and decreased number of neurons in adult cortices. A, KI67-labeled cells at E13.5 (scale bar, 50 μm). B, pH3 labeling at E13.5. C, Quantification of KI67 labeling in E13.5 cortices (n = 5, of either sex, t(8) = 0.076, p = 0.9402, Student's t test). D, Quantification of pH3-labeled cells in E13.5 cortices (n = 5, of either sex, t(8) = 0.3156, p = 0.7604, Student's t test). E, KI67 labeling at E15.5 (scale bar, 50 μm). F, pH3 labeling at E15.5. G, Quantification of KI67-labeled cells in E15.5 cortices (n = 5, of either sex, t(8) = 0.3016, p = 0.7652, Student's t test). H, Quantification of pH3-labeled cells in E15.5 cortices (n = 5, of either sex, t(8) = 1.339, p = 0.2101, Student's t test). I, Quantification of activated cleaved caspase-labeled neurons (n = 4, of either sex, t(8) = 3.563, p = 0.0119, Student's t test). J, Distribution of cells in layers 2/3 (CUX1) and layers 5/6 (CTIP2) of middle coronal sections. K, Quantification of total CUX1-labeled neurons in rostral, middle, and caudal locations (Control and Miro1CKO n = 5; 2 females, 3 males each, rostral: t(8) = 0.3923, p = 0.7051, middle: t(8) = 0.1088, p = 0.9160, and caudal: t(8) = 0.2466, p = 0.8115, Student's t test). L, Quantification of Total CTIP2-labeled neurons in rostral, middle, and caudal locations (Control and Miro1CKO n = 5; 2 females, 3 males each, rostral: t(8) = 1.057, p = 0.3212, middle: t(8) = 0.9721, p = 0.3595, caudal: t(8) = 0.6034, p = 0.5692, Student's t test). n.s. p > 0.05, *p ≤ 0.05, **p ≤ 0.01.
To determine if cortical neurons were properly positioned in the cerebral cortex, rostral, middle, and caudal coronal sections were immunolabeled with antibodies to CUX1 (labeling layers 2/3) and CTIP2 (labeling layers 5/6). Although the overall distribution of the cells was not overtly disrupted, minor distribution changes were noted (Fig. 3J). CUX1-labeled cells were found in regions outside of layers 2/3 in many Miro1CKO cortices including layer 1 (Fig. 3J, down pointing arrows) and layer 4 (Fig. 3J, side pointing arrows). A significant shift of CUX1-labeled neurons toward the lower portion of layers 2/3 was also detected in caudal sections, with a similar trend observed in mid-level sections (Table 1). CTIP2-positive cell counts did not differ significantly between control and Miro1CKO tissue. However, total numbers of CUX1- and CTIP2-labeled cells were significantly reduced in Miro1CKO cortices (Fig. 3K,L). When normalized to total DAPI counts, these reductions were no longer statistically significant (Tables 2, 3).
CUX1 neuronal counts normalized to DAPI
CTIP2 neuronal counts normalized to DAPI
CUX1 and CTIP2 total neuronal counts normalized to DAPI
Miro1CKO mice were found to have astrogliosis and a reduction in oligodendrocytes
Since the Miro1CKO cortices were significantly smaller than controls (Fig. 1E) and had fewer CUX1- and CTIP2-labeled neurons (Fig. 3K,L), we also wanted to examine if there were changes to glia. Astrocytes and oligodendrocytes were labeled with GFAP and ASPA, respectively. Miro1CKO mice had a significant infiltration of astrocytes in adult cortices when compared with controls (n = 4, Rostral: F(9,60) = 19.66, p < 0.0001, data not shown; Middle: Fig. 4A–C; Caudal: F(9,50) = 19.33, p < 0.0001, data not shown). ASPA-labeled oligodendrocytes were significantly reduced in adult Miro1CKO cortices (n = 6, t(10) = 2.991, p = 0.0135, Student's t test; Averages - Control: 100.0; Miro1CKO: 72.77; SEM: Control: 7.831 Miro1CKO: 4.643).
Miro1CKO cortices have increased GFAP-labeled astrocytes. A, B, Control and Miro1CKO GFAP-labeled cortices (green). C, Quantification of GFAP-labeled astrocytes (Control n = 4; 1 female, 3 males and Miro1CKO n = 4; 2 females, 2 males, F(9,60) = 25.29, p < 0.0001, two-way ANOVA) n.s. p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ****p < 0.0001; Bin 1, pial surface; Bin 10, above corpus callosum.
Abnormal home cage behaviors
Miro1CKO mice were observed to be hyperactive and to engage in repetitive activities, including circling the periphery of the cage (Movie 1). They also displayed poor nesting (Fig. 5D–G) and resistance to being handled by animal care and research staff (biting, squealing, and extreme writhing). To determine if the hyperactive behaviors were only in response to workers being present or if they were ever-present, control and Miro1CKO mice were tracked during their dark cycle (see Materials and Methods). Traces from the overnight recordings showed control mice traveled throughout their cage and spent time in their nest (Fig. 5A, left). In contrast, Miro1CKO mice traveled in a circular track around the cage perimeter throughout the 2 h window, never returning to a nest (Fig. 5A, right). The distance and rates of travel found for Miro1CKO mice were significantly increased when compared with controls (Fig. 5B,C).
Miro1CKO mice display abnormal home cage behavior. A, Example traces from home cage tracking. B, Quantification of distance traveled in home cage (Controls n = 12, Miro1CKO n = 9). C, Quantification of speed in home cage (Control n = 12; 7 females, 5 males; and Miro1CKO n = 9; 3 females, 6 males, Distance: t(19) = 2.423, p = 0.0256, Student's t test; Speed: t(19) = 2.282, p = 0.0342, Student's t test). D–G, Spectrum of Miro1CKO nesting in home cage. *p ≤ 0.05.
Miro1CKO mice display hyperactive and repetitive behaviors compared with their control littermates. [View online]
Impaired motor skills
Previous data suggested that Miro1CKO mice ablated with Eno2-Cre develop motor neuron disease (amyotrophic lateral sclerosis)-like phenotypes (Nguyen et al., 2014). Although our mice moved freely throughout their cage without noticeable impairment, to establish if they had any subtle motor impairments, gross motor skills were assessed using the hindlimb footprint pattern test (Fig. 6A). Hind-base width and stride length measurements showed no difference between the control and Miro1CKO mice (Fig. 6B,C). The vertical pole test, hanging wire, wire grid, and forelimb grip strength were used to further investigate motor coordination and strength in the Miro1CKO mice. Although most of the control and Miro1CKO mice were able to complete the initial vertical pole incline (with the exception of one control and one Miro1CKO falling off at 63 and 37°, respectively), the Miro1CKO mice were unable to stay on the rod as long as the controls at 90° (Fig. 6G). Additionally, the Miro1CKO mice showed statistically significant decreases in time on the hanging wire, reductions in forelimb grip strength, and increased numbers of foot slips on the wire grid (Fig. 6D–F). Together these data suggest Miro1CKO mice have subtle motor skill deficits, although less severe than the Eno2-Cre;Miro1 conditional mutant mice (Nguyen et al., 2014).
Miro1CKO mice have impaired motor strength and coordination. A, Examples of control and Miro1CKO footprint patterns and measurements. B, C, Quantification of hind-base width and stride length (Control n = 13 and Miro1CKO n = 12, of either sex; Hind-Base Width: t(23) = 0.2952, p = 0.7705, Student's t test, Stride Length: t(23) = 1.198, p = 0.2433, Student's t test) from the footprint pattern test. D–G, Quantification of grip strength, vertical pole, and wire grid data (Grip Strength: Control n = 25; 11 females, 14 males and Miro1CKO n = 11; 4 females, 7 males, t(34) = 3.420, p = 0.0016, Student's t test; Hanging Wire: Control n = 15; 8 females, 7 males and Miro1CKO n = 15; 8 females, 7 males, t(28) = 6.821, p < 0.0001, Student's t test; Vertical Pole: Control n = 15; 8 females, 7 males and Miro1CKO n = 15; 8 females, 7 males, t(28) = 2.071, p = 0.0477, Student's t test; Wire Grid: Control n = 15; 8 females, 7 males and Miro1CKO n = 15; 8 females, 7 males, t(28) = 5.570, p < 0.0001, Student's t test). *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001.
Anxiety-like behavior in Miro1CKO mice
To determine if the hyperactive and agitative behavior was due to anxiety, mice were tested with two anxiety-like behavior paradigms: the open field test and the elevated plus maze. We examined components of behavior over time and by averaging. In the open field test, the average time mobile for the Miro1CKO mice was significantly lower than the controls (Fig. 7B, right graph). During the experiment, the control group was mobile for longer times when compared with the CKO group (p < 0.001). The time mobile for the control group decreased over time (p = 0.001) while there was no change in the time mobile for the CKO group (Fig. 7B, left graph; p = 0.674).
Miro1CKO mice exhibit anxiety-like behavior during the open field test. A, Example traces from the controls and Miro1CKO mice during 5 min open field test. B–F, Quantification of time mobile, distance traveled, maximum speed, freezing episodes, and entries into the center (Controls n = 14; 8 females, 6 males, Miro1CKO n = 13; 5 females, 8 males; Time mobile: Time mobile: t(25) = 4.12, p = 0.0004, Student's t test, Distance: t(25) = 1.37, p = 0.183, Student's t test, Max speed: t(25) = 0.955, p = 0.349, Student's t test, Freezing episodes: t(25) = −2.63, p = 0.0145, Student's t test, Entries into center: t(25) = 7.25, p ≤ 0.0001, Student's t test). n.s. p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
The average maximum speed and total distance traveled were not significantly different between the control and the Miro1CKO mice (Fig. 7E,F, right graphs). However, the time traveling was significantly different (Fig. 7B, right graph; p < 0.001). The distance traveled decreased over time for the control group (p < 0.001) while increasing over time for the CKO group (p = 0.045; Fig. 7E, left graph). There was no difference in the max speed recorded between the control and CKO groups (p = 0.229), but the maximum speed decreased over time for the control group (p < 0.001) while there was no change for the CKO group (Fig. 7F, left graph; p = 0.167).
Throughout the experiment, Miro1CKO mice displayed anxiety-like behaviors such as freezing, circling the perimeter, and avoiding the center of the open field (Fig. 7C,D). The control group had a smaller number of freezing episodes when compared with the CKO group (p = 0.009). The traces of the Miro1CKO mice consistently showed the mice circling the outside of the arena, distinct from control mice traces (Fig. 7A).
To further evaluate anxiety-like behavior in the Miro1CKO mice, we employed the elevated plus maze test. Traces indicated that the Miro1CKO mice spent more time in the open arms of the elevated plus maze (Fig. 8A). Interestingly, the Miro1CKO mice also showed greater variability between mice (Fig. 8B). The data reveal that during the first half of the experiment, both groups spent the same average amount of time in the closed and open arms (Table 4). During the second half of the experiment, the controls spent more time in the closed while the Miro1CKO mice spent more time in the open arms (Fig. 8C,D). Additionally, the Miro1CKO mice had longer average visits to the open arms than controls, and the Miro1CKO mice covered less distance than the controls in the closed arms (Table 4). The overall means in the first and second half of the experiments for the controls and Miro1CKO mice were not statistically different for speed in the closed and open arms. These data were an unexpected contrast to the open field test. This led us to consider an alternative explanation that the Miro1CKO mice manifest a different kind of anxiety related disorder, seeking to avoid closed spaces, analogous to claustrophobia.
Miro1CKO mice prefer the open arm of the elevated plus maze. A, Coordinates from each mouse's trace are overlayed to display where control and Miro1CKO mice spent time in the elevated plus maze. B, Sample traces from control and Miro1CKO mice. C, D, Quantification of time spent in the closed and open arm of the elevated plus maze during the second half of the experiment (Controls n = 19; 10 females, 9 males; Miro1CKO n = 9; 4 female, 5 males; 150–300 s; Closed Arm: t(26) = 2.420, p = 0.0228, Student's t test, Open Arm: t(26) = 3.013, p = 0.0057, Student's t test). *p ≤ 0.05, **p ≤ 0.01.
Elevated plus maze data from the first and second half of the experiment
Miro1CKO mice avoid confined spaces
Since the Miro1CKO mice spent more time in the open arms than the closed arms of the elevated plus maze, we hypothesized they avoided confined spaces. To further investigate this possibility, we utilized a test where mice had access to a wide space and a narrow space (Fig. 9A; El-Kordi et al., 2013). Miro1CKO mice had a longer latency to enter and fewer entries into the narrow portion of the box when compared with controls (Fig. 9B,C). Additionally, Miro1CKO mice traveled significantly less distance in the narrow part of the box than the controls (Fig. 9D). This data, together with the elevated plus maze data, support the hypothesis that the Miro1CKO mice avoid confined spaces.
Miro1 ablated mice avoid narrow spaces. A, Dimensions of the wide and narrow spaces in the box (adopted from El-Kordi et al. (2013)). B–D, Quantification of the number of entries, latency to enter, and distance in the narrow space (Controls n = 15; 10 females, 5 males, Miro1CKO n = 7; 3 females, 4 males, Latency to enter: t(20) = 2.301, p = 0.0323, Student's t test; Entries into narrow space: t(20) = 3.935, p = 0.0008, Student's t test; Distance in narrow space: t(20) = 3.034, p = 0.0065, Student's t test). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
Discussion
We have identified a unique role for Miro1 during early cortical development and an association with an anxiety-like behavior. The abrogation of Miro1 from excitatory neural progenitors results in a mislocalization of mitochondria to the rear of radially migrating neurons and clustering of mitochondria in the cell bodies of excitatory neurons in adults. Although no cell cycle defects were found at E13.5 or E15.5, increased apoptosis was observed in cortical progenitors at P4. As a result, adult Miro1CKO mice have decreased overall brain size as determined by a reduced weight, decreased cortical and hippocampal area, and enlarged ventricles. Miro1CKO cortices show a global reduction of CUX1- and CTIP2-labeled neurons; however, the overall organization remains largely preserved. A relative astrocytosis was present in Miro1CKO cortices while the number of oligodendrocytes was decreased.
We observed selective behavioral abnormalities in the Miro1CKO mice. In contrast to control mice, Miro1CKO mice demonstrate abnormal home cage behavior and poor self-care, including neglecting to build a nest, along with an aversion to handling. Motor testing revealed Miro1CKO mice to have mild deficits in grip strength, vertical pole, hanging wire, and wire grid; all less severe than those seen in Eno2-Cre;Miro1CKO mice. Unlike the Eno2-Cre;Miro1CKO mice, Emx1-Cre;Miro1CKO mice remain able to normally move within their cage (Nguyen et al., 2014).
Open field testing found Miro1CKO mice displayed anxiety-like behaviors, spending more time on the outer edge of the arena than in the center and exhibiting increased freezing. They also have an aversion to closed spaces, as demonstrated on the elevated plus maze and the wide/narrow box test.
This is the first study to examine a role for Miro1 in migrating excitatory neurons. Previous studies have found decreases in cortical and hippocampal volume and increased astrocytosis using a CAMKII-Cre to conditionally delete Miro1 postnatally (López-Doménech et al., 2016). Additionally, previous research investigating a role for Miro1 in the developing nervous system have found profound neurodegenerative symptoms including rigidity, spasticity, and death during early postnatal development; however, neuronal migration was not evaluated (Nguyen et al., 2014). These studies have shown histological changes in neurite extension and degeneration, mitochondrial localization, and bunina-like body formation during development (Nguyen et al., 2014; López-Doménech et al., 2016). Research from our lab suggests that mitochondrial dynamics and their ATP-producing pathways are important early in development during neuronal migration (Lin-Hendel et al., 2016). These processes appear to be neuronal subtype specific. Excitatory neurons utilize glycolysis and/or oxidative phosphorylation for energy production interchangeably during migration whereas cortical inhibitory neurons are fully dependent on oxidative phosphorylation (Lin-Hendel et al., 2016). In addition, mitochondrial dynamics differ in developing neural subtype. Mitochondria in inhibitory neurons are highly dynamic (moving around the cell) through migration whereas they remain primarily adjacent to the nucleus in the direction of the leading process in migrating excitatory neurons (Lin-Hendel et al., 2016). Here we find that perturbing mitochondrial dynamics in migrating excitatory neurons results in subtle changes in the mature cerebral cortical architecture.
In addition to structural and molecular changes, these are the first data implicating Miro1 in the pathogenesis of behavior phenotypes observed in neurodevelopmental disorders. While MIRO1 has been linked to neurodegenerative disorders, its role in neurodevelopmental diseases such as autism spectrum disorders or schizophrenia has been incompletely studied (Nguyen et al., 2014; Norkett et al., 2020; Kontou et al., 2021). Interestingly, the Miro1CKO mice display similar behaviors to Neurogranin and GPM6a knock-out mice. Neurogranin has been implicated in neurodevelopmental disorders such as ADHD, autism, and schizophrenia and is a postsynaptic protein kinase that binds to calmodulin in the absence of calcium. Its expression begins in the first 3 weeks after birth and is known to be expressed in hippocampus pyramidal and granular neurons. Previous research suggests that the absence of neurogranin causes changes in synaptic plasticity, including paired-pulse depression, synaptic fatigue, and long-term potentiation induction (Pak et al., 2000). Neurogranin knock-out mice show decreased nesting, hyperactive behavior in their home cage, decreased time spent in the center of the open field arena, and increased time in the open arm of the elevated plus maze (Nakajima et al., 2021). Given the phenotypic similarities, studying interactions between mitochondria and neurogranin seem appropriate or possibly neurogranin and MIRO1.
Similarly, GPM6a has been linked to autism and schizophrenia and is suggested to act as a nerve growth factor-gated calcium channel (Mukobata et al., 2002). GPM6a is known to play a role in developing neurons and is suggested to participate in neuronal migration, neuronal differentiation, and synapse development, including neurite outgrowth and spine formation (Mukobata et al., 2002; Alfonso et al., 2005; Michibata et al., 2008; Zhao et al., 2008; Mita et al., 2015; Formoso et al., 2016; Aparicio et al., 2023). Studies examining GPM6a knock-out mice have found similar behavioral changes, including increased time spent in the open arm of the elevated plus maze and decreased time spent in the narrow portion of the wide/narrow box (El-Kordi et al., 2013).
Calcium is an important modulator of neurodevelopmental processes including neuronal migration and synaptic transmission. Previous research suggests that calcium regulates leading process extension and branching as well as organization during neuronal migration, acting as a “stop” and “go” signal in the developing cortex and is important for neurotransmitter release in the mature cortex (Horigane et al., 2019, review). MIRO1 has EF-1 and EF-2 calcium sensing domains that play a role in directing calcium shuttling mitochondria to regions in need of calcium buffering. Although previous research suggests Miro1 loss does not affect the cytosolic or mitochondria calcium concentrations in mouse embryonic fibroblasts during development, ablation of Miro1 has been found to decrease endoplasmic reticulum-mitochondrial tethering which alters calcium buffering and contributes to increasing autophagy and mitophagy leading to neuronal death in neurodegenerative diseases such as Parkinson's and Alzheimer's diseases (Berenguer-Escuder et al., 2020; Kam et al., 2020).
One limitation of our current study is that calcium levels were not established in the developing excitatory neurons. Future studies will address the molecular and cellular underpinnings including the role of calcium dynamics in the anxiety-like behaviors that Miro1CKO mice display. Similarities to other mouse models that have indicated potential calcium signaling dysregulation make this an intriguing direction for future study.
MIRO1 is not the only mitochondrial motor adaptor protein; TRAK and MYO19 can bind to and locate mitochondria on microtubules and actin, respectively, potentially accounting for different localization patterns within a cell (MacAskill et al., 2009; López-Doménech et al., 2018; Kontou et al., 2021). It is currently not known whether these proteins are involved in certain types of localization within developing neurons and how this affects organization of, and communication between, mature neurons. Future studies are required to establish the involvement of these other mitochondrial motor adaptor proteins to further characterize their respective roles in developing neurons.
The underlying mechanisms of neuropsychiatric disorders are not well understood. This study suggests that mitochondrial location in migrating excitatory neurons could play a role in the development and onset of behavioral phenotypes in neuropsychiatric diseases such as autism spectrum disorder and schizophrenia. Our mouse model, using an Emx1-Cre to ablate Miro1, displays distinct anxiety-like behavior patterns that are distinct from those observed in previous models such as when Miro1 is removed using the Eno2-Cre. The latter model showed classic characteristics of neurodegeneration including muscle spasticity and weakness, hindlimb clasping, premature death around postnatal day 40, and pathologic features such as bunina bodies (Nguyen et al., 2014). Unfortunately, their analysis of P30 cerebral cortex was limited and developmental changes were not considered. One possible explanation for the difference between our model and the Eno2-Cre;Miro1 mutant is that each disrupts specific networks of neurons that leads to different disease phenotypes; some that have an early onset and others that have a later onset. Distinguishing the networks of neurons that are impacted by the ablation of Miro1 in each of these cases could provide a deeper understanding of the underlying circuitry dysfunction that plagues individuals with neurodevelopmental and neurodegenerative diseases. This may also lead to novel therapies that could improve the outcomes and quality of life for afflicted individuals. Finally, our study provides a new model to investigate the neurobiology underlying behavioral phenotypes related to anxiety-like disorders.
Evidence of mitochondrial dysfunction has been recognized in neuropsychiatric conditions such as autism spectrum disorder and schizophrenia (Poling et al., 2006; DiMauro and Schon, 2008; Weissman et al., 2008). Our study suggests that early loss of Miro1 in migrating excitatory neurons could contribute to and sustain behavioral symptoms that accompany these disorders. A continued focus and understanding of the cellular and molecular mechanisms of Miro1 could lead to a deeper knowledge of the origins of neuropsychiatric disorders and the identification of potential treatment targets.
Footnotes
The authors declare no competing financial interests.
This work was supported by the National Institutes of Neurologic Disorders and Stroke (NS100007 to J.A.G.) and financial support from the Hamilton College Biology and Psychology Departments and Dean’s Office. We thank Tom Freeland and Walt Zarnoch for building the elevated plus maze and the wide-narrow box, Trevor Lancon for helping to compute mitochondria localization, and Russell Hardesty for help constructing traces, heat maps, and graphs. The Microscope Imaging Facility at Hamilton College, Siobhan Robinson, the Blatt BioImaging Center at Syracuse University (especially Dr. Abrar Aljiboury), and the Masonic Medical Research Institute (especially Dr. Chase Kessinger) kindly shared equipment. We also thank members of the Golden Lab for their many discussions, suggestions, and support. Finally, we thank Brenna Stallings for managing the mouse colony while at Brigham and Women's Hospital.
This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.
References
In this issue
