Novel Development of Magnetic Resonance Imaging to Quantify the Structural Anatomic Growth of Diverse Organs in Adult and Mutant Zebrafish

Abstract

Zebrafish (Danio rerio) is a widely used vertebrate animal for modeling genetic diseases by targeted editing strategies followed by gross phenotypic and biomarker characterization. While larval transparency permits microscopic detection of anatomical defects, histological adult screening for organ-level defects remains invasive, tedious, inefficient, and subject to technical artifact. Here, we describe a noninvasive magnetic resonance imaging (MRI) approach to systematically screen adult zebrafish for anatomical growth defects. An anatomical atlas of wild-type (WT) zebrafish at 5–31 months post-fertilization was created by ex vivo MRI with a 9.4 T magnet. Volumetric growth over time was measured of animals and major organs, including the brain, spinal cord, heart, eyes, optic nerve, ear, liver, kidneys, and swim bladder. Subsequently, surf1^−/−, fbxl4^−/−, and opa1^+/− mitochondrial disease mutant adult zebrafish were quantitatively studied to compare organ volumes with age-matched WT zebrafish. Results demonstrated that MRI enabled noninvasive, high-resolution, rapid screening of mutant adult zebrafish for overall and organ-specific growth abnormalities. Detailed volumetric analyses of three mitochondrial disease mutants delineated specific organ differences, including significantly increased brain growth in surf1^−/− and opa1^+/−, and marginally significant decreased heart and spinal cord volumes in surf1^−/− mutants. This is interesting as we know neurological involvement can be seen in SURF1^−/− patients with ataxia, dystonia, and lesions in basal ganglia, as well as in OPA1^+/− patients with spasticity, ataxia, and hyperreflexia indicative of neuropathology. Similarly, cardiomyopathy is a known sequelae of cardiac pathology in patients with SURF1^−/−-related disease. Future studies will define MRI signaling patterns of organ dysfunction to further delineate specific pathology.

Introduction

Zebrafish (Danio rerio) has proven to be of significant translational utility in understanding gene function, as many human genes are conserved in zebrafish¹ and efficient gene editing technologies allow for ready creation of human disease models through gene modifications.^2,3 Several attempts⁴ have been targeted toward expanding the knowledge of zebrafish anatomy, physiology, and histology to optimize their utility as a vertebrate animal model in which to further understanding of human disease genes and variants. Menke et al. described detailed histologic analyses of zebrafish,⁵ which helped to identify major zebrafish organ systems. The optical transparency of zebrafish embryos and larvae allows for facile application of fluorescence imaging techniques⁶ and newer methods such as second or third harmonic generation microscopy or spectroscopic imaging.⁷ However, as the zebrafish grow older their increasing tissue density renders imaging by light microscopy unobtainable.

Histologic analyses are required to screen organ size and morphology in adult animals, which is an inherently terminal procedure and one that is tedious and liable to artifact. Therefore, there is a strong need for developing dependable techniques for high-resolution imaging to noninvasively visualize and analyze organ growth in adult zebrafish models.

Magnetic resonance imaging (MRI) is a widely applied, noninvasive imaging technology that produces three-dimensional (3D) detailed anatomical images in diverse species including humans.⁸ Kabli et al. reported application of high-resolution magnetic resonance microscopy to study detailed anatomical structures in adult zebrafish.^9,10 While they described various anatomical details in intact adult zebrafish and attempted in vivo imaging, their studies did not provide insight into systematic organ analysis or comparative growth in adult zebrafish animals of different ages. More recent studies have used MRI to study live fish, but these methods required custom-designed and fabricated sample cells, limiting their generalized use.^11,12 In this study, we sought to demonstrate the feasibility of high-resolution ex vivo MRI as a noninvasive imaging tool to objectively assess zebrafish anatomy and discern organ-level abnormalities that may develop over time.

Specifically, MRI on a 9.4 T vertical bore magnet with gradient echo imaging was optimized to permit quantitative analysis of the developmental trajectory of zebrafish organs in wild-type (WT) animals. Histologic validation of WT animals was done for organ confirmation of imaging findings. In addition, comparative growth analysis of animals and major organs was performed in three mitochondrial disease mutant adult strains, namely surf1^−/−, fbxl4^−/−, and opa1^+/−13,14 MR images were collated to serve as a reference anatomical atlas for adult zebrafish over time in both WT animals and in three primary mitochondrial disease models. This work informs the growth trajectory of organ-specific anatomical development in zebrafish and demonstrates that MRI offers the ability to screen for organ-level growth defects in mutant adult zebrafish in an efficient, low-cost, rapid, and unbiased manner.

Materials and Methods

Experimental animals

Animal procedures were conducted according to the NIH Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee (IACUC) of the Children's Hospital of Philadelphia (CHOP). WT zebrafish and mutant lines (surf1^−/−, opa1^+/−, fbxl4^−/−)^13,14 were maintained in a breeding colony at 27°C. Fifteen WT adult zebrafish, 9 surf1^−/− mutants, 3 opa1^+/− mutants, and 3 fbxl4^−/− mutant adult zebrafish, were included in each age group studied, as described in Table 1. The surf1^−/− mutants were created by the clustered regularly interspaced short palindromic repeats (CRISPR-Cas9) technique and harbored a homozygous deletion leading to a premature stop codon in mRNA predicted to result in a null state.¹³ The opa1^+/− mutants created by the transcription activator-like effector nucleases (TALEN) technique had a heterozygous deletion leading to a premature stop codon with the resultant allele unable to make protein (S. Chan, pers. comm.).

Table 1.

Age and Genotype of Adult Zebrafish Studied by Magnetic Resonance Imaging

Genotype	surf1−/−			opa1+/−	fbxl4−/−	WT
Age (months)	5	12 to 14	19	16–18	31	5	8 to 9	12 to 14	16 to 18	31
Animal number	3	3	3	3	3	3	3	3	3	3

Five distinct time points were selected to comprehensively investigate the evolution of organ growth across the adult trajectory of WT zebrafish. Cross-sectional analyses of surf1^−/− mutants were performed across the three different age groups available in the extant collection within our vivarium at the time, while also exploring the imaging patterns and characteristics of opa1^+/− and fbxl4^−/− mutants at specified time points.

WT, wild type.

The fbxl4^−/− mutant strain studied was fbxl4sa12470, created by N-ethyl-N-nitrosourea mutagenesis, carried a homozygous point mutation c.813T>A that resulted in a premature stop codon, and was previously characterized by our research group.¹⁴

Only male zebrafish were included as imaging subjects. During the pilot studies, attempts were also made to image female zebrafish, where the presence of eggs in female fish hindered the clear visualization and accurate analysis of organ volumes. Consequently, it was deemed unfavorable for the initial survey to include female zebrafish due to the potential interference caused by the obstructed views caused by crowding from eggs.

Diet of zebrafish

Until 30 days post-fertilization (dpf), zebrafish are provided with a constant food source floating on the surface of the water. Until 17 dpf they are fed Gemma 75 (Skretting), followed by Gemma 150 through 30 dpf. From 15 dpf, they are additionally fed with live brine shrimp. From 30 dpf, zebrafish are fed with Gemma 500 and/or Zeigler zebrafish diet (Zeigler Bros, Inc.). Depending on age and development, zebrafish are fed once or twice daily with Ziegler/Gemma. All zebrafish are fed once a day with brine shrimp. Tanks are provided with amounts that can be eaten by the zebrafish within 5 to 10 min.

Zebrafish preparation for MRI analysis

Animals were euthanized in an ice-water bath in accordance with a protocol approved by the CHOP IACUC following the American Veterinary Medical Association guidelines. Approximately 0.2 mL of 16% paraformaldehyde (PFA) was injected into the peritoneal cavity of each zebrafish. Subsequently, zebrafish were transferred into a 15 mL tube with 10 mL of 4% PFA and 2.5 mM of Gd-DTPA (gadolinium diethylene triamine penta-acetic acid) and incubated for 5 days.¹⁵ This MRI contrast agent was added into the media to reduce the acquisition time. Each animal's total body weight was recorded immediately before imaging following removal from PFA-Gd-solutions and wiping with a tissue (Table 2).

Table 2.

Genotype, Age (Months), Weight (g), and Volumes of Organs (Cubic Millimeters) of 30 Individual Adult Zebrafish Studied by Magnetic Resonance Imaging

Specimen No.	Genotype	Age (months)	Weight (g)	Brain	Spinal cord	Eyes	Optic nerve	Heart	Liver	Kidney	Swim bladder	Ears
1	opa1	16	0.621	10.3	1.61	7.27	0.279	0.986	6.94	2.83	14.2	1.17
2	AB WT	9	n/a	6.97	1.04	5.44	0.161	0.924	6.67	2.91	20.3	0.891
3	AB WT	9	0.427	6.34	0.849	4.64	0.142	0.982	4.54	2.07	17.4	0.867
4	surf1	12	0.6	7.44	1.17	5.7	0.204	0.633	5.79	2.02	16.2	1.66
6	AB WT	12	0.628	8.14	1.57	5.89	0.202	0.813	4.16	2.4	26.1	1.11
7	AB WT	12	0.561	8.09	1.53	5.79	0.191	0.968	4.82	3.03	22.8	1.25
8	AB WT	16	0.568	7.11	1.73	7.47	0.214	1.85	10.7	2.04	40.4	1.28
9	opa1	18	0.463	7.52	1.25	5.89	0.137	0.842	4.33	1.76	19.4	1.53
10	opa1	18	0.552	8.78	1.19	7	0.231	1.35	6.03	4.41	24.3	1.6
11	surf1	14	0.539	6.41	1.26	3.26	0.0696	2.44	4.67	2.24	18	1.28
12	surf1	14	0.562	7.53	1.16	3.45	0.079	0.95	4.16	3.17	26.9	1.57
13	AB WT	18	0.688	9.35	1.9	5.58	0.29	1.56	8.71	2.75	23.8	1.36
14	AB WT	18	0.726	8.22	1.86	5.96	0.156	1.05	8.63	2.8	29.1	1.21
15	AB WT	14	0.489	7.82	1.5	5.07	0.115	1.13	5.74	2.37	19.6	1.03
16	surf1	19	0.566	9.01	1.6	6.69	0.199	0.903	6.35	2.15	35.2	1.5
17	surf1	19	0.463	8.28	0.932	6.1	0.119	0.989	5.29	1.75	20.3	1.18
18	surf1	19	0.629	9.26	1.09	6.26	0.162	1.26	10	2.91	28.5	1.06
19	fbxl4	31	0.746	9.61	0.975	8.05	0.175	0.953	8.19	4.52	27.2	1.48
20	fbxl4	31	0.544	6.71	1.17	6.51	0.0991	0.807	5.5	2.27	28	1.08
21	fbxl4	31	0.423	6.47	0.965	4.78	0.082	1.01	5.89	3.1	13.7	0.975
22	AB WT	31	0.978	13.5	1.6	9.59	0.159	1.77	19.2	5.62	43.4	1.68
23	AB WT	31	0.755	11.2	1.61	10.9	0.2	1.88	9.66	4.64	39.6	1.27
24	AB WT	31	0.704	10.2	1.38	8.42	0.281	1.21	8.18	3	22.3	1.7
25	surf1	5	0.235	5.28	0.58	3.92	0.12	0.598	3.02	0.737	12.8	0.874
26	surf1	5	0.221	4.77	0.65	3.16	0.0371	0.613	3.04	0.87	10.5	0.782
27	surf1	5	0.169	4.61	0.571	2.51	0.038	0.355	1.73	0.756	9.73	0.728
28	AB WT	5	0.368	6.04	0.782	3.71	0.0531	0.643	2.22	2.3	14.8	0.93
29	AB WT	5	0.327	6.85	0.864	4.51	0.0851	0.756	3.19	1.54	12.1	1.04
30	AB WT	5	0.363	6.09	0.0946	4.43	0.0821	0.662	4.89	1.83	20.6	0.838
31	AB WT	8	0.4935	7.33	1.1	5.48	0.118	0.973	7.33	2.66	27.7	0.979

Animal no. 2 was used for comparative MRI and histologic analysis, but was not included in group analyses of normalized organ volume data as no animal weight was obtained. Animal no. 5 was not included in the study due to inadequate specimen preparation.

MRI, magnetic resonance imaging; n/a, not available.

Ex vivo MRI

Zebrafish (n = 30) were individually placed in a 10 mm nuclear magnetic resonance tube with 4 mL of Fomblin (Ausimount, Thorofare, NJ), a proton-free liquid used in MRI to provide a dark background signal. MRI experiments were performed with a 9.4 T vertical bore magnet interfaced with a Bruker console. 3D gradient echo imaging of each animal was performed with these acquisition parameters: repetition time 100 ms, echo time 6 ms, 1 scan, field of view 2.56 × 1 × 1 cm³, 840 × 330 × 330 matrix, leading to an isotropic resolution of 30 μm and an acquisition time of ∼3 h.

Histology

Following MRI of a 9-month-old WT zebrafish, the animal was submitted to the CHOP Pathology Core Facility for histology. Tissue underwent an extended processing cycle on an automated processor, and then was embedded in paraffin. A series of 4 μm sections (sagittal through the body or coronal to examine eyes) were stained by standard hematoxylin and eosin methods, and then digitally scanned on an Aperio ScanScope CS whole-slide scanner. Histologic and MR images were manually compared side by side to improve reliable MRI identification of each organ.

Image analysis

Acquired 3D images were displayed using “ImageJ”¹⁶ and reformatted into a stack of two-dimensional images in Analyze 7.5 format. Files were reviewed with ITK-SNAP (3.8.0 version).^17,18 Organ delineation was performed manually with the Active Counter (“snake”) tool. A 3D mask of the brain, spinal cord, liver, kidneys, heart, ears, eyes, optic nerve, and swim bladder was created, and the volume of each organ was calculated (Table 2). Vertebrae 1 to 17 (4 Weberian, 10 abdominal, 3 caudal) were used as anatomical landmarks to permit consistent delineation of the full spinal cord length, as the caudal end of some animals was not well visualized on MRI.

Statistical analyses

R version 4.1.0 (May 18, 2021) was used to conduct the data analysis. Organ volumes were converted from voxels, which are equivalent to 30 μm³, to mm³ by multiplying with a conversion factor of 2.7 × 10⁻⁵. As the weight and age of WT zebrafish were found to be highly correlated (Pearson correlation coefficient 0.89, p < 0.0001), the volume of each organ was normalized by animal weight. One WT fish (animal no. 2) did not have a weight recorded and was excluded from MRI analysis but was used for histologic study. In addition, one WT fish (animal no. 5) was excluded from analysis due to excessive contrast injection that prohibited reliable image analysis due to organ bulging.

Volume of the normalized organ size for each mutant strain was compared with age-matched WT fish at the same time point using a paired Student's t-test. Plots were generated using the “ggplot2” and “tidyverse” R packages.^19,20 A two-sided p-value cutoff was set to be 0.05 for statistical significance.

Results

Comparative MRI and histologic analysis of a WT adult zebrafish

MRI of 9-month-old WT zebrafish animal no. 2 (Table 2) was compared with histologic sections obtained at multiple anatomic levels to identify different organs. Figures 1 and 2, respectively, depict this animal's organs as identified on MRI and histologic analyses. Figure 1 includes sagittal (Fig. 1A), axial (Fig. 1B), and coronal (Fig. 1C) sections on MRI as well as a 3D reconstruction image (Fig. 1D) to depict relative organ structure and position. This comparative analysis enabled creation of a radiological anatomical atlas for MR-only organ comparison of the other zebrafish studied.

FIG. 1.

MRIs of 9-month-old WT zebrafish with (1A) sagittal, (1B) axial, (1C) coronal sections, and (1D) 3D reconstruction. Organs analyzed include (a) brain, (b) kidneys, (c) liver, (d) spinal cord, (e) heart, (f) swim bladder, (g) ears/otolith organ, (h) eyes, and (i) optic nerve. 3D, three-dimensional; MRI, magnetic resonance imaging; WT, wild type.

FIG. 2.

Three sagittal histologic sections of 9-month-old WT zebrafish, with annotation of selected tissues.

Volumetric analysis of WT zebrafish organs from 5 to 31 months post-fertilization

To assess organ growth over the course of normal zebrafish development, we imaged three WT zebrafish per age group across five age groups, including the following: (1) 5 months, (2) 8–9 months, (3) 12–14 months, (4) 16–18 months, and (5) 31 months post-fertilization. Nine organs were evaluated in each zebrafish animal, with organ volumes calculated in cubic millimeters using the ITK-SNAP software (3.8 version). These data were compiled to discern the organ-specific growth rates in WT animals (Fig. 3). Some artifact was noted in the area of the swim bladder due to the presence of air and water. However, as depicted in Supplementary Figure S2, this artifact did not prohibit delineating boundaries of the swim bladder or any mass effect on adjacent organs.

FIG. 3.

Comparative volumetric analysis from 5 to 31 months post-fertilization of organ volumes in adult WT zebrafish. Each dot indicates a single animal.

These data demonstrate that, as predicted, organ volume appears to increase with age in all zebrafish organs, with perhaps the exception of the spinal cord that, based on MRI, appears to have minimal, if any, increased growth after approximately age 1 year post-fertilization.

Comparative volumetric analysis of organ volumes in mitochondrial disease mutants from 5 to 31 months post-fertilization

surf1^−/− mutants were imaged by studying three animals per age group in three different age groups, including (1) 5 months, (2) 12–14 months, and (3) 19 months (Table 2). These data were compiled to quantify organ-specific growth rates in the surf1^−/− mutants relative to age-matched WT controls (Fig. 4). These data suggest that surf1^−/− mutants demonstrated a fairly consistent degree of growth delay in nearly all organs, with apparent catch up by 19 months in the brain, heart, and swim bladder.

FIG. 4.

Comparative volumetric analysis in surf1^−/− mutants from 5 to 19 months post-fertilization relative to age-matched WT controls. Each dot indicates a single animal. Black line and dots are WT, gray line and triangles are surf1^−/−.

A single time point analysis was also performed in two other mitochondrial disease mutant zebrafish strains to allow cross-sectional screening analysis for organ-specific growth defects in mitochondrial disease models. Specifically, fbxl4^−/− mutants (n = 3) were studied at age 31 months post-fertilization and opa1^+/− mutants (n = 3) were imaged at 16–18 months post-fertilization (Table 2). A comparative volumetric analysis of all organs in the three mutant strains relative to age-matched WT controls was performed (Fig. 5). While Figure 5 demonstrates raw organ volumes for clear visual representation, statistical analyses were performed using organ volumes normalized by animal weight as an internal control.

FIG. 5.

Comparison of organ volumes of three mitochondrial disease mutants (triangles) versus age-matched WT adult zebrafish (circles). From left to right, 5-month-old WT, 5-month-old surf1^−/− mutants; 12–14-month-old WT, 12–14-month-old surf1^−/− mutants; 16–18-month-old WT, 16–18-month-old opa1^+/− mutants, 19-month-old surf1^−/− mutants; 31-month-old WT, 31-month-old fbxl4^−/− mutant zebrafish. ^#0.10 < p > 0.05, *p < 0.05, **p < 0.01.

Since organ volumes were normalized to animal weight, analysis was performed of animal weights, which demonstrated that 5-month-old surf1^−/− mutants (p < 0.05) and 31-month-old fbxl4^−/− (p < 0.01) mitochondrial disease mutants were significantly smaller than age-matched WT controls, but no significant weight differences were seen at the other mutant age studies (Fig. 6). We have also included Supplementary Figure S1 to depict comparison of normalized organ volume of mitochondrial mutants versus age-matched WT controls.

FIG. 6.

Comparison of animal weight in three mitochondrial disease mutants (triangles) versus age-matched WT (circles) adult zebrafish. From left to right, 5-month-old WT, 5-month-old surf1^−/− mutants; 12–14-month-old WT, 12–14-month-old surf1^−/− mutants; 16–18-month-old WT, 16–18-month-old opa1^+/− mutants, 19-month-old surf1^−/− mutants; 31-month-old WT, 31-month-old fbxl4^−/− mutant zebrafish. *p < 0.05, **p < 0.01.

Interestingly, significantly increased brain volume was noted relative to age-matched WT controls both in 16–18-month-old opa1^+/− mutants (p = 0.022) and in 19-month-old surf1^−/− mutants (p = 0.006). In addition, marginally statistically significant decreases were evident in volumes of multiple organs of the surf1^−/− mutants relative to age-matched WT controls, including heart volume at 5 months (p = 0.079), and spinal cord volume at 12 and 19 months (p = 0.07 and p = 0.09, respectively). The fbxl4^−/− mutant fish showed a trend toward small organ volume in nearly all organs, although this difference did not reach statistical significance, likely due to the small sample size. opa1^+/− zebrafish trended toward smaller volumes in the heart, liver, spinal cord, and swim bladder, but these comparative differences did not reach statistical significance when normalized to relative animal weight.

Supplementary Table S1 presents detailed information on the percentage differences and corresponding p-values of organ volumes in the zebrafish mutants compared with their age-matched WT controls.

Discussion

We have developed a novel protocol that utilizes a 9.4 T vertical bore magnet with gradient echo acquisition sequence to enable high-resolution, ex vivo, MRI of adult zebrafish with volumetric screening of eight different organs, including the brain, spinal cord, heart, eyes, optic nerve, ear/otolith, liver, kidneys, and swim bladder. This imaging approach was applied to establish the first reference MRI atlas of comparative organ volumes at various stages of adult growth and development from 5 to 31 months post-fertilization in WT zebrafish, with organ delineation aided by comparative analysis to histologic analysis. We further demonstrated the efficacy of MRI as an effective tool for phenotyping organ-level differences in human disease models.

Specifically, comparative MRI relative to age-matched WT controls was performed in adult zebrafish models harboring mutations in nuclear genes that impair mitochondrial function, including surf1^−/− at three age ranges from 5 to 19 months post-fertilization and opa1^+/− and fbxl4^−/− each at single ages. Interestingly, surf1^−/− animals showed an overall growth delay in nearly all organs relative to age-matched WT controls, with apparent catch-up growth by 19 months post-fertilization in the heart and swim bladder, but a marginally significant decrease in heart volume at 5 months and in spinal cord volume at 12 and 19 months, and surprisingly, a significantly increased brain volume at 19 months.

Applying noninvasive MRI technology represents an important advance in zebrafish studies, as it is the first technology to reliably quantify linear growth volumes of essential organs in WT animals between 5 and 31 months of age (Fig. 3). Indeed, MRI offers an accurate 3D representation of tissues without gaps while tissues remain in an aqueous environment. By contrast, histological methods of organ evaluation used in the past significantly underestimated organ volumes due to the elaborate tissue preparation techniques required that result in tissue shrinkage and artifactual gaps between tissue slices. Also, the tomographic nature of MRI, which is not possible with traditional histological staining, allows observation of intact zebrafish animals in three orthogonal orientations, which is beneficial for visualization and evaluation of internal organs.

The study conducted by Ding et al.²¹ provided compelling evidence that synchrotron-based X-ray microcomputed tomography (micro-CT) surpasses the resolution capabilities of histological tests as well as our own MRI zebrafish data, enabling the acquisition of more detailed information regarding tissue and cell structures. It is important to note that the images presented in the Ding article were generated using a national high-energy beam as would be required for micro-CT, making it much less widely accessible. In contrast, the zebrafish MR images in our study were obtained utilizing a 9.4 T narrow magnet, a technology readily available in many research institutions. In addition, the micro-CT methodology limited the size of zebrafish that could be studied and included data for only a single 33 dpf juvenile zebrafish along with five smaller larval-stage zebrafish.

These distinctions highlight the improved practicality and widespread applicability of our chosen imaging modality, MRI, in comparison with the specialized equipment and limited samples characterized in the aforementioned study.

The preparation of specimens, along with meticulous imaging and analysis of zebrafish images using MRI, demanded a considerable investment of time and labor. Manual analysis of MR images for each fish was imperative since no automated software exists to reliably delineate organ boundaries and accurately calculate volumes, which represents a considerable limitation. Indeed, the future development of automated software tailored for zebrafish MRI organ analyses would undoubtedly be advantageous and time-saving. Such advancements would streamline the image analysis process, reducing the reliance on manual efforts and potentially enhancing efficiency. The implementation of automated software further holds the promise of increased accuracy, consistency, and reproducibility in the analysis of zebrafish MRI data. Moreover, it would facilitate the broader adoption of this imaging technique by researchers, as it would eliminate the need for specialized expertise in manual image analysis.

Human SURF1 encodes an assembly factor of mitochondrial complex IV (cytochrome-c oxidase, COX), the terminal component of the mitochondrial respiratory chain. SURF1 deficiency causes a range of severe, multisystemic problems, including Leigh syndrome, sensorineural hearing loss, and Charcot–Marie–Tooth disease type 4K.²² In zebrafish, surf1 affects COX activity and is involved in embryonic organ development.²³ Baden et al. described other zebrafish models with COX deficiency that were noted to have developmental defects in endodermal tissue, cardiac function, and swimming behavior.²³ Given our MR findings of growth delay and organ-specific volumetric differences relative to age-matched WT controls, future studies will be interesting to study the heart, brain, and spinal cord pathology in surf1^−/− mutants. Indeed, we have demonstrated that these animals have larval stress sensitivity that induces brain death reminiscent of Leigh syndrome, as well as reduced exercise capacity at baseline in adults¹³; however, their spinal cord has not previously been studied to date.

Human OPA1 encodes a protein that localizes to the inner mitochondrial membrane, where it plays a key role in mitochondrial fusion, stabilizing complex I and IV subunits and sequestering proapoptotic cytochrome-c molecules within the mitochondrial cristae.²⁴ OPA1 heterozygous loss-of-function mutations are a common cause of autosomal dominant optic atrophy, predominantly manifesting as vision loss although “plus” features can be seen including sensorineural hearing loss, peripheral neuropathy, and multisystem dysfunction, while recessive OPA1 pathogenic variants have been shown to cause autosomal recessive Behr syndrome and autosomal recessive mitochondrial DNA depletion syndrome 14 with neurological involvement.²⁵ In zebrafish, opa1 has been shown to be involved in chordate embryonic development, mitochondrion morphogenesis, and ventricular cardiac muscle cell development.^26–28 Interestingly, MR screening of the opa1^+/− heterozygous mutant adult zebrafish at 16–18 months post-fertilization revealed they had a significantly larger brain volume relative to WT controls (Fig. 5).

Based on these findings, subsequent studies can be directed toward unraveling the underlying pathology responsible for increased brain volume in surf1^−/− and opa1^+/− zebrafish mutants.

Human FBXL4 encodes a mitochondria-localized protein of uncertain functions, but likely involved in mitochondrial ubiquitylation quality control²⁹ and mitophagy.³⁰ Human FBXL4 deficiency due to autosomal recessive pathogenic variants causes mitochondrial respiratory chain deficiency, variable degrees of mitochondrial depletion, and a range of multisystem organ dysfunction in the Leigh syndrome spectrum.²⁹ The fbxl4^−/− mutant fish showed a trend toward small-organ volume in nearly all organs, although these differences did not reach statistical significance, perhaps due to the small sample size with only three animals studied. Our research group recently reported extensive phenotypic characterization of the fbxl4^−/− mutant fish, which surprisingly showed no major gross anomalies under basal growth conditions, but acute brain death in fbxl4^−/− mutant larvae exposed to acute stress.¹⁴

It would be intriguing in future studies to compare MRI of adult fbxl4^−/− mutant zebrafish following a similar acute stress exposure, as this is a common cause of neurodevelopmental regression and metabolic stroke in human mitochondrial disease.

Surprisingly, many apparent differences in comparative organ volumes of mutant mitochondrial disease zebrafish models relative to WT fish disappeared when organ volumes were normalized for fish weight. Due to high variability in animal weight and size, normalization of organ volumes by a specific parameter for comparison is prudent. However, as growth delay may be a feature of a mutant animal's phenotype, it would be valuable to explore additional parameters, such as animal length, for organ normalization in future studies.

Conclusion

We have developed a novel, high-resolution MRI methodology to reliably quantify zebrafish organ volumes and anatomy in fixed adult animals. This imaging approach can now be readily used as a first-line, 3D test to noninvasively screen for global organ involvement in novel mutant zebrafish disease lines. This capacity represents a major advance over traditional histological staining methods enabling maintenance of organ structure and 3D topography. We believe that the ability to identify organs and study their evolution via MRI will prove advantageous, as this is exponentially less time-consuming, cumbersome, labor-intensive, and prone to artifacts compared with histologic analyses. In particular, paraffin embedding often causes organ shrinkage, which makes volumetric analysis difficult. Faster and unbiased analyses will also allow for incorporation of a larger number of animals in screening studies, which is essential to accurately identify organ-level anomalies.

As a novel MRI technique for zebrafish applications, this methodology requires an expert radiologist familiar with ex vivo imaging techniques. Applying our knowledge gained from ex vivo MRI analysis reported here to evolving imaging techniques, such as in vivo MRI, will enable novel experiments comparing organ-level effects over time in the same animal. Based on detailed volumetric analyses performed in three primary mitochondrial disease mutant zebrafish strains with specific organ volume differences identified, further functional assessment is planned to evaluate for cardiac function in surf1^−/− mutants, as well as brain and spinal cord pathology in surf1^−/− and opa1^+/− mutants. Future research will also focus on defining MRI signaling patterns of dysfunction within diverse organs, as well as developing live animal biochemical imaging methods that may permit intermediary metabolite quantification in living zebrafish models of primary mitochondrial disease.

Authors' Contributions

M.J.F. conceived of and designed the study. S.M. developed MRI experimental and NMR-based analytic methods. S.H., M.L., and S.C. assisted with zebrafish mutant strain characterization, husbandry, and animal selection for imaging. S.S. and C.S. prepared and fixed zebrafish for MRI analysis. S.S. performed image analysis, with the assistance of S.M. and M.S. ER performed statistical analyses with guidance from R.X. S.S., B.J.W., and D.M. performed histologic analyses and interpretation. S.S., S.M., E.R., and M.J.F. wrote the article. All authors approved the final version.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Disclosure Statement

S.C. is CEO of Neuroene Therapeutics. M.J.F. is engaged with several companies involved in mitochondrial disease therapeutic preclinical and/or clinical-stage development. M.J.F. is cofounder of Rarefy Therapeutics; advisory board member with equity interest in RiboNova, Inc.; scientific board member and paid consultant with Khondrion, Larimar Therapeutics, and MiMo Therapeutics; paid consultant for Astellas (formerly Mitobridge), Casma Therapeutics, Cyclerion Therapeutics, Epirium Bio (formerly Cardero Therapeutics), HealthCap VII Advisor AB, Imel Therapeutics, Mayflower, Inc/Primera, Minovia Therapeutics, Mission Therapeutics, NeuroVive Pharmaceutical AB, Reneo Therapeutics, Stealth BioTherapeutics, Vincere Bio, and Zogenix; and/or sponsored research collaborator for Aadi Bioscience, Astellas, Cyclerion Therapeutics, Epirium Bio, Imel Therapeutics, Khondrion, Merck, Minovia Therapeutics, Mission Therapeutics, NeuroVive Pharmaceutical AB, PTC Therapeutics, Raptor Pharmaceuticals, REATA, Inc., Reneo Therapeutics, RiboNova, Standigm, and Stealth BioTherapeutics.

M.J.F. also has received royalties from Elsevier and speaker fees from Agios Pharmaceuticals, GenoMind, and educational honorarium from PlatformQ. None of the other authors has relevant conflicts of interest to declare.

Funding Information

This work was funded, in part, by the Children's Hospital of Philadelphia Mitochondrial Medicine Fellowship Program and the National Institutes of Health (R35-GM134863 to M.J.F.).

Supplementary Material

Supplementary Table S1

Supplementary Figure S1

Supplementary Figure S2