Open Access
Retina  |   April 2023
Fiji-Assisted Automatic Quantitative Volumetric Analysis of Choroidal Neovascularization in a Laser-Induced Choroidal Neovascularization Mouse Model
Author Affiliations & Notes
  • Dimitrios Pollalis
    USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
    USC Ginsburg Institute for Biomedical Therapeutics, University of Southern California, Los Angeles, CA, USA
  • Arjun V. Nanda
    College of Medicine, University of Oklahoma Health Science Center, Oklahoma City, OK, USA
  • Gopa Kumar Gopinadhan Nair
    USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
    USC Ginsburg Institute for Biomedical Therapeutics, University of Southern California, Los Angeles, CA, USA
  • Sun Young Lee
    USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
    Department of Physiology and Neuroscience, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
    Deptartment of Ophthalmology, Dean McGee Eye Institute, University of Oklahoma Health Science Center, Oklahoma City, OK, USA
    Department of Physiology, University of Oklahoma Health Science Center, Oklahoma City, OK, USA
  • Correspondence: Sun Young Lee, USC Roski Eye Institute, Keck School of Medicine, University of Southern California, 1450 San Pablo, Los Angeles, CA 90033, USA. e-mail: sunyoung.lee@med.usc.edu 
  • Footnotes
    *  DP and AVN contributed equally to this work.
Translational Vision Science & Technology April 2023, Vol.12, 10. doi:https://doi.org/10.1167/tvst.12.4.10
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Dimitrios Pollalis, Arjun V. Nanda, Gopa Kumar Gopinadhan Nair, Sun Young Lee; Fiji-Assisted Automatic Quantitative Volumetric Analysis of Choroidal Neovascularization in a Laser-Induced Choroidal Neovascularization Mouse Model. Trans. Vis. Sci. Tech. 2023;12(4):10. https://doi.org/10.1167/tvst.12.4.10.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: The laser-induced choroidal neovascularization (CNV) mouse model is the most frequently used animal model of CNV. To test new therapeutic agents that suppress CNV, CNV measurement in an accurate, precise, and efficient manner is important. We present the utility of Fiji-assisted automatic volumetric quantification of CNV in comparison with two-dimensional CNV analyses.

Methods: Laser-induced CNV was induced in C57BL/6J mice according to the established protocol. After CNV induction, mice were treated with intravitreal injection of either phosphate-buffered saline solution (PBS) or Aflibercept, an anti- vascular endothelial growth factor agent. One week after intravitreal injection treatment, retina pigment epithelium/choroid flat mounts were stained with rhodamine-conjugated Griffonia simplicifolia lectin B4. Z-stacks of the entire CNV lesion obtained using laser confocal microscopy were converted to binary stacks using Fiji for volumetric analysis. Data from volumetric analysis and multiple area analyses from z-stack projection, the maximum, blindly selected, and mean area were compared using Fiji.

Results: Fiji-assisted automatic quantitative volumetric analysis of CNV was useful in detecting experimental outliers in laser-induced CNV genesis and provided accurate and precise measurements of total areas of CNV with a lower coefficient of variance (63%) than in multiple area analyses, including the z-stack projection, maximum, blindly selected, and mean areas (67%, 67%, 76%, and 69%, respectively). A lower coefficient of variance in volumetric analysis than in multiple area analyses resulted in increased statistical significance when comparing CNV lesions in PBS, and Aflibercept-treated groups; P = 0.004 in volumetric analysis versus P value range between 0.03 and 0.05 in multiple area analyses.

Conclusions: Fiji-assisted automatic quantitative volumetric analysis can be useful for accurate, precise, and efficient measurements of total areas of CNV.

Translational Relevance: Volumetric measurement for CNV lesions can be advantageous in verifying the efficacy of therapeutic agents in the laser-induced CNV mouse model.

Introduction
Choroidal neovascularization (CNV) is a hallmark of exudative age-related macular degeneration characterized by pathologic growth of new vascular complex from the choroid. Vascular endothelial growth factor–neutralizing proteins have been the mainstay treatment for CNV. However, CNV remains a significant cause of vision impairment, and continuous efforts to improve the current treatment strategy or to develop new treatments are underway. 
The laser-induced CNV mouse model, where pathologic CNV develops after rupturing retina pigment epithelium (RPE) and Bruch's membrane by laser photocoagulation, has been the most widely accepted and frequently used animal model for studying CNV.13 The size of the CNV lesion in a laser-induced CNV model has been routinely measured and quantified to test the efficacy of therapeutic agents in preclinical studies and has been predictive of outcomes in clinical trials.1,4 Although the laser-induced CNV mouse model continues to be an important animal model for developing treatments for CNV, the methodology to measure CNV size in a laser-induced CNV mouse model often varies among investigators.511 Because laser-induced disruption of RPE and Bruch's membrane induces an irregular-shaped neovascular complex in the laser-induced CNV mouse model, quantification using an axial plane can underestimate or overestimate the real size of CNV. The volumetric analysis offers an alternative method for quantifying CNV lesions that is potentially more representative of the actual size of the lesion. However, some of the previous methods used for the volumetric analyses were regarded as more time-consuming.9,10 In this study, we demonstrate the utility of Fiji-assisted automatic quantitative volumetric analysis of CNV in comparison with various 2-dimensional (2D) area analyses of CNV to propose an accurate, precise, and efficient measurement of CNV in the laser-induced CNV mouse model in which the CNV suppressing therapeutic efficacy of tested agents can be more reliably assessed. 
Methods and Materials
Animals
Four- to 6-week-old wild-type C57BL/6 mice used for the experiments were treated according to the Association for Research in Vision and Ophthalmology guidelines on the care and use of animals in research. All protocols were approved by the Animal Care and Use Committee of the Oklahoma University Health Science Center and the University of Southern California. 
Laser-Induced Choroidal Neovascularization Mouse Model
Laser photocoagulation was performed as previously described.13 Briefly, pupils were dilated using 2.5% phenylephrine hydrochloride and 1% cyclopentolate hydrochloride. Mice were anesthetized with ketamine (100 mg/mL) and xylazine (100mg/mL) via intraperitoneal injection. Laser photocoagulation was performed using the Micron IV retinal imaging system (Phoenix Research Laboratories, Pleasanton, CA, USA) with the Meridian Merilas 532 green laser (50 µm, 70 ms, 240 mW). Three lesions were induced located approximately one disk diameter from the optic nerve, with respect to the large vessels. Laser-induced disruption of Bruch's membrane was confirmed by the appearance of a bubble at the site of photocoagulation. Induction of CNV was confirmed by fundus photography, optical coherence tomography, fluorescent angiography, and hematoxylin and eosin staining seven days after the laser photocoagulation as we previously reported.10 
Intravitreal Injection
Three days after laser photocoagulation, mice received an intravitreal injection of phosphate-buffered saline solution (PBS; 1 µL) or Aflibercept (40 µg/1 µL). For intravitreal injection, mice were anesthetized with intraperitoneal ketamine (100 mg/mL)/xylazine (100 mg/mL) and pupils were dilated with 2.5% phenylephrine/1% cyclopentolate. A small scleral incision was made posterior to the limbus using a 31-gauge (G) insulin needle. Then, a 33G blunt needle attached to a 10 µL NanoFil syringe (World Precision Instruments, Sarasota, FL, USA) was used to deliver 1 µL of solution in the vitreous cavity through the same incision. For the CNV analysis, mice were divided into two groups: PBS-treated group (n = 6-7) and Aflibercept-treated group (n = 11). 
Histological Analysis
Mice were euthanized seven days after the intravitreal injection, and eyes were harvested for flat-mount histology. For flat-mount immunohistochemistry, cornea, lens, vitreous, and retina were removed, and RPE/choroid complex was fixed in 4% PFA for three hours. Then the tissue was blocked in 10% horse serum and stained with Rhodamine-conjugated Griffonia Simplicifolia Lectin B4 (RL-1102; Vector Laboratories, Burlingame, CA, USA) overnight and flat mounted. 
Image Analysis
Images and Z-stacks of the entire CNV lesion in each laser spot were obtained by confocal microscope (Olympus, Tokyo, Japan) and converted to binary stacks using Fiji. Three-dimensional (3D) volumetric analyses for CNV were obtained using the formula: \(V = h \times \sum \nolimits_{i = 1}^N {A_i}\), V: total volume, h: interval between z-stack sections, A: threshold area. Two-dimensional (2D) area analyses for CNV were obtained from either (1) the z-stack merged projection, (2) a blindly selected slice, (3) the slice with the largest area (maximum area), or (4) the slice with the mean area within each stack (Fig. 1). For each eye, the average of the three laser spots was considered as a single experimental value. The averaged values in the Aflibercept-treated group were normalized by the PBS-treated group. The efficacy of CNV suppression in the Aflibercept-treated group compared with the PBS-treated control group (%) was calculated by Improvement (%) = (Mean Control − Aflibercept)/Mean Control × 100) and compared among the volumetric and area analyses of CNV. The coefficient of variation (CV) was calculated as CV% = standard deviation/mean × 100. 
Figure 1.
 
Quantification of area or volume of CNV lesions. Images and Z-stacks of the entire rhodamine-conjugated Griffonia Simplicifolia Lectin B4 (GSL)-stained CNV lesion in each laser spot were obtained via confocal microscopy and converted to binary stacks using Fiji. The 3D volumetric analyses for CNV were obtained using the formula: V: total volume, h: the interval between z-stack sections, and A: threshold area. The 2D area analyses for CNV were obtained from either (1) the z-stack projection, (2) the maximum, (3) a blindly selected, or (4) the mean area within a stack.
Figure 1.
 
Quantification of area or volume of CNV lesions. Images and Z-stacks of the entire rhodamine-conjugated Griffonia Simplicifolia Lectin B4 (GSL)-stained CNV lesion in each laser spot were obtained via confocal microscopy and converted to binary stacks using Fiji. The 3D volumetric analyses for CNV were obtained using the formula: V: total volume, h: the interval between z-stack sections, and A: threshold area. The 2D area analyses for CNV were obtained from either (1) the z-stack projection, (2) the maximum, (3) a blindly selected, or (4) the mean area within a stack.
Statistical Analysis
All values were reported as mean ± standard deviation. GraphPad Prism was used for the statistical analysis and graph plotting. Statistical differences were measured with Student's t-test for comparison between the two groups. P values <0.05 were considered significant. 
Results
In all animals, the size of CNV was automatedly measured in either 2D or 3D. Intravitreal Aflibercept treatment effectively suppressed CNV in the laser-induced CNV mouse model. The mean efficacy of CNV suppression in the Aflibercept-treated group was 70.56%, 58.23%, 46.8%, 56.08 and 55.69% from volumetric, z-stack projection area, maximum area, blindly chosen area, mean area analyses, respectively. P values of the efficacy of CNV suppression in the Aflibercept-treated group were 0.004, 0.05, 0.04, 0.03, and 0.03 using volumetric analysis, and analyses of the z-stack projection, the maximum, a blindly chosen, and the mean areas, respectively (Fig. 2). The CV of each measurement method, including volumetric analysis, and analyses of the z-stack projection, maximum, blindly chosen, and mean areas, were 63%, 67%, 67%, 76%, and 69% in the PBS-treated group and 27%, 28%, 56%, 37%, and 36% in the Aflibercept treated group, respectively (Fig. 3). 
Figure 2.
 
Comparison of different CNV measurement methods on the efficacy of CNV suppression by Aflibercept. The quantitative measurements of area or volume in each group are shown (A). Averaged three CNV measurements in one mouse retina was regarded as single n. The efficacy of CNV suppression in the Aflibercept treated group compared with the untreated group (%) was calculated by Improvement (%) = (Mean Control – Aflibercept-treated group)/Mean Control × 100 and compared among the volumetric and area analyses of CNV (B).
Figure 2.
 
Comparison of different CNV measurement methods on the efficacy of CNV suppression by Aflibercept. The quantitative measurements of area or volume in each group are shown (A). Averaged three CNV measurements in one mouse retina was regarded as single n. The efficacy of CNV suppression in the Aflibercept treated group compared with the untreated group (%) was calculated by Improvement (%) = (Mean Control – Aflibercept-treated group)/Mean Control × 100 and compared among the volumetric and area analyses of CNV (B).
Figure 3.
 
Comparison of different CNV measurement methods on the coefficient of variation (CV) efficacy of CNV suppression by Aflibercept. The coefficient of variation (CV, %) for each CNV measurement method was calculated by CV (%) = standard deviation (SD)/mean × 100. AVG, average; aVEGF, antivascular endothelial growth factor treatment; SD, standard deviation; CV, coefficient of variance.
Figure 3.
 
Comparison of different CNV measurement methods on the coefficient of variation (CV) efficacy of CNV suppression by Aflibercept. The coefficient of variation (CV, %) for each CNV measurement method was calculated by CV (%) = standard deviation (SD)/mean × 100. AVG, average; aVEGF, antivascular endothelial growth factor treatment; SD, standard deviation; CV, coefficient of variance.
Discussion
The murine model of laser-induced CNV, introduced by the Campochiaro group in 1998, is the most widely accepted and most frequently used experimental murine CNV model.13 Although CNV in the laser-induced mouse model has a fundamentally different pathophysiology than in exudative AMD in humans, this model has proven to be suitable for testing the efficacy of new drugs through systemic or intraocular administration and has shown predictive value for drug effects in patients with CNV. For CNV assays, image analysis with quantification of CNV is a major parameter to evaluate the impact of drug therapies on CNV regression.4 Because mice do not have macula or accumulate prominent fluids in CNV as humans do, where the central retinal thickness or fluid volumes are important parameters for CNV activity, quantification of the abnormal vascular complex has been the main strategy to determine CNV activity in the laser-induced CNV model.512 However, the methodology to quantify CNV has been variable among the investigators. 
CNV lesions can be quantified from the serial tissue sections throughout the extent of each burn. However, this process is time consuming and is not automated. CNV lesions can also be quantified from REP/choroid flat-mount using either 2D or 3D analyses. For 2D analysis, images obtained by fluorescence microscopy are binarized and thresholded according to their background and the masked area (in micrometers squared) of CNV is measured using image analysis software, such as Image J. In our study, we demonstrated that the CNV-suppressing benefit of Aflibercept could be recapitulated differently depending on the method of CNV measurement used. We found that the raw values of CNV measurement varied from the different methods of CNV measurement. Thus the efficacy of CNV suppression and statistical significance varied. Our result shows that volumetric analysis (in µm3) leads to a lower coefficient of variance (CV) compared with the multiple area analyses. The great uniformity observed with the volumetric approach suggests that the 3D volumetric analysis may offer more accurate and precise measurements of total areas of CNV than the 2D area analysis method. Despite this finding, 2D area analysis, using axial planes to measure CNV lesions, remains a frequently used measurement method.5,6 This method poses potential pitfalls that may introduce biased results because it ignores the heterogeneity of vascular complex, stains with different thresholds and the thickness of abnormal vascular complex in the laser-induced CNV mouse model. Without capturing the whole lesion, lesion size and extent cannot be fully appreciated. Furthermore, investigators may be inclined to select areas with the strongest stain, which may not necessarily be the most representative of the lesion. This can introduce investigator bias and potentially overestimate or underestimate the therapeutic efficacy of CNV-suppressing study drugs. Therefore the volumetric approach, which captures the entire lesion, offers a more representative measurement of true lesion size and is advantageous in determining the efficacy of therapeutic agents in the laser-induced CNV mouse model. 
In our study, we also automatedly quantified CNV volume using Fiji. Fiji is an open-source image processing package included in ImageJ2, a next-generation version of ImageJ.13 Fiji is easy to install and benefits from bundling many plugins that offer comprehensive documentation. Plugins are much faster and more powerful, whereas Macros, which also allows that automatic 3D analyses are slow, waste memory for more complicated tasks, cannot use the full functionality of ImageJ, and always run in the foreground, thus blocking the computer. Other commercial software such as IMARIS (Oxford Instrument, Abingdon, UK) or MATLAB can provide 3D volumetric analyses. However, it may take extra steps of converting images, an additional cost to purchase, or knowledge of the script language. As other groups and we have demonstrated, freely accessed Fiji easily and reliably quantifies CNV volumes from automated volumetric measurements.11,14 The 3D reconstruction using Fiji can also be used for colocalization analyses within CNV, because it allows for visualization of tissue structures stained with multiple markers.11 
In conclusion, volumetric analysis of CNV lesions offers consistent results that more accurately characterize the true lesion size. Fiji efficiently allows automatic quantification of CNV volume in a laser-induced CNV mouse model. Provided the widespread use of the laser-induced CNV mouse model in pre-clinical studies, using a volumetric approach to CNV lesion quantification can maximize the accuracy and precision of results. 
Acknowledgments
The authors thank Feng Li and Mark Dittmar (OUHSC P30 Live Animal Imaging Core, Dean A. McGee Eye Institute, Oklahoma City, OK, USA). 
Supported by Oklahoma Shared Clinical and Translational Resources (OSCTR), Oklahoma City, OK (Pilot Program no. U54 GM104938) (to L.S.Y.); NIH/NEI (K12EY028873 and 1R56EY034193-01) (to L.S.Y.); Unrestricted Grant to the Department of Ophthalmology from Research to Prevent Blindness, New York, NY; NIH/NEI P30EY029220; NIH/NEI P30EY021725. 
Disclosure: D. Pollalis, None; A.V. Nanda, None; G.K.G. Nair, None; S.Y. Lee, None 
References
Tobe T, Ortega S, Luna JD, et al. Targeted disruption of the FGF2 gene does not prevent choroidal neovascularization in a murine model. Am J Pathol. 1998; 153: 1641–1646. [CrossRef] [PubMed]
Lambert V, Lecomte J, Hansen S, et al. Laser induced choroidal neovascularization model to study age related macular degeneration in mice. Nat Protocols. 2013; 8: 2197–2211. [CrossRef] [PubMed]
Shah RS, Soetikno BT, Lajko M, Fawzi AA. A mouse model for laser-induced choroidal neovascularization. J Vis Exp. 2015;(106): e53502.
Heier JS, Brown DM, Chong V, et al. Intravitreal aflibercept (VEGF trap-eye) in wet age-related macular degeneration. Ophthalmology. 2012; 119: 2537–2548 [CrossRef] [PubMed]
Takahashi K, Nakamura S, Otsu W, Shimazawa M, Hara H. Progranulin deficiency in Iba-1+ myeloid cells exacerbates choroidal neovascularization by perturbation of lysosomal function and abnormal inflammation. J Neuroinflammation. 2021; 18: 164. [CrossRef] [PubMed]
Bucher F, Aguilar E, Marra KV, et al. CNTF prevents development of outer retinal neovascularization through upregulation of CxCl10. Invest Ophthalmol Vis Sci. 2020; 61(10): 20. [CrossRef] [PubMed]
Lv W, Chen W, Huang S, et al. Reduction of laser-induced choroidal neovascularization in mice with erythropoietin RNA interference. Transl Vis Sci Technol. 2022; 11(8): 1. [CrossRef] [PubMed]
Yang TJ, Yao MD, Sun YN, Li XM, Jiang Q, Yan B. Suppression of choroidal neovascularization by silencing of long non-coding RNA IPW. Aging. 2021; 13: 10584–10602. [CrossRef] [PubMed]
Martinez-Camarillo JC, Spee CK, Trujillo-Sanchez GP, et al. Blocking ocular sympathetic activity inhibits choroidal neovascularization. Front Neurosci. 2022; 15: 780841. [CrossRef] [PubMed]
Sakurai E, Anand A, Ambati BK, van Rooijen N, Ambati J. Macrophage depletion inhibits experimental choroidal neovascularization. Invest Ophthalmol Vis Sci. 2003; 44: 3578–3585. [CrossRef] [PubMed]
Pollalis D, Kim D, Nair GKG, Kang C, Nanda AV, Lee SY. Intraocular RGD-engineered exosomes and active targeting of choroidal neovascularization (CNV). Cells. 2022; 11: 2573. [CrossRef] [PubMed]
Jager RD, Mieler WF, Miller JW. Age-related macular degeneration. N Engl J Med. 2008; 358: 2606–2617. [CrossRef] [PubMed]
Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: An open-source platform for biological-image analysis. Nat Methods. 2012; 9: 676–682. [CrossRef] [PubMed]
Berger A, Cavallero S, Dominguez E, et al. Spectral-domain optical coherence tomography of the rodent eye: Highlighting layers of the outer retina using signal averaging and comparison with histology. PLoS One. 2014; 9(5): e96494. [CrossRef] [PubMed]
Figure 1.
 
Quantification of area or volume of CNV lesions. Images and Z-stacks of the entire rhodamine-conjugated Griffonia Simplicifolia Lectin B4 (GSL)-stained CNV lesion in each laser spot were obtained via confocal microscopy and converted to binary stacks using Fiji. The 3D volumetric analyses for CNV were obtained using the formula: V: total volume, h: the interval between z-stack sections, and A: threshold area. The 2D area analyses for CNV were obtained from either (1) the z-stack projection, (2) the maximum, (3) a blindly selected, or (4) the mean area within a stack.
Figure 1.
 
Quantification of area or volume of CNV lesions. Images and Z-stacks of the entire rhodamine-conjugated Griffonia Simplicifolia Lectin B4 (GSL)-stained CNV lesion in each laser spot were obtained via confocal microscopy and converted to binary stacks using Fiji. The 3D volumetric analyses for CNV were obtained using the formula: V: total volume, h: the interval between z-stack sections, and A: threshold area. The 2D area analyses for CNV were obtained from either (1) the z-stack projection, (2) the maximum, (3) a blindly selected, or (4) the mean area within a stack.
Figure 2.
 
Comparison of different CNV measurement methods on the efficacy of CNV suppression by Aflibercept. The quantitative measurements of area or volume in each group are shown (A). Averaged three CNV measurements in one mouse retina was regarded as single n. The efficacy of CNV suppression in the Aflibercept treated group compared with the untreated group (%) was calculated by Improvement (%) = (Mean Control – Aflibercept-treated group)/Mean Control × 100 and compared among the volumetric and area analyses of CNV (B).
Figure 2.
 
Comparison of different CNV measurement methods on the efficacy of CNV suppression by Aflibercept. The quantitative measurements of area or volume in each group are shown (A). Averaged three CNV measurements in one mouse retina was regarded as single n. The efficacy of CNV suppression in the Aflibercept treated group compared with the untreated group (%) was calculated by Improvement (%) = (Mean Control – Aflibercept-treated group)/Mean Control × 100 and compared among the volumetric and area analyses of CNV (B).
Figure 3.
 
Comparison of different CNV measurement methods on the coefficient of variation (CV) efficacy of CNV suppression by Aflibercept. The coefficient of variation (CV, %) for each CNV measurement method was calculated by CV (%) = standard deviation (SD)/mean × 100. AVG, average; aVEGF, antivascular endothelial growth factor treatment; SD, standard deviation; CV, coefficient of variance.
Figure 3.
 
Comparison of different CNV measurement methods on the coefficient of variation (CV) efficacy of CNV suppression by Aflibercept. The coefficient of variation (CV, %) for each CNV measurement method was calculated by CV (%) = standard deviation (SD)/mean × 100. AVG, average; aVEGF, antivascular endothelial growth factor treatment; SD, standard deviation; CV, coefficient of variance.
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×