Normative Profile of Retinal Nerve Fiber Layer Thickness and Lamina Cribrosa-Related Parameters in a Healthy Non-Glaucoma Cynomolgus Monkey Colony

Abstract

Purpose: The purpose of this study was to investigate the normal range of ophthalmic parameters and the correlations between systematic and ocular parameters and retinal nerve fiber layer (RNFL) thickness among a healthy non-glaucoma cynomolgus monkey colony.

Methods: All included monkeys were given detailed ophthalmic examinations, including anterior and posterior segments. Furthermore, univariate and multivariate linear regression models were conducted to estimate the relationship between systemic and ophthalmic parameters and global RNFL thickness.

Results: A total of 349 non-glaucoma monkeys (18.69 ± 2.88 years old) were collected. The global RNFL thickness was 94.61 ± 10.13 µm, and sex-specific differences existed in all sectors. The decreasing trend of RNFL is as follows: inferotemporal, superotemporal, inferonasal, superonasal, temporal, and nasal. For lamina cribrosa (LC)-related parameters, cup depth (P < 0.01), LC thickness (P = 0.014), and Bruch's membrane opening (BMO) – minimum rim width 2 (P = 0.002) were greater in the male group. However, LC depth (P = 0.02), anterior laminar insertion depth-1 (P = 0.009), and mean anterior laminar insertion depth (P = 0.029) of female monkeys were greater than those of male monkeys. In multivariate linear regression, only older age was significantly related to reduced global RNFL thickness (P < 0.001).

Conclusions: Our findings suggest the differences in RNFL thickness distribution and sex between non-glaucoma cynomolgus monkeys and humans. Therefore, the impact of this difference on outcomes should be fully considered in laboratory animal studies. Our findings are also significant in terms of developing a normative optical coherence tomography (OCT) database in nonhuman primates (NHPs).

Translational Relevance: We found that the differences in RNFL thickness distribution and sex between non-glaucoma cynomolgus monkey colonies and humans should be thoroughly taken into account in laboratory animal studies.

Introduction

Over the past few decades, having been increasingly applied in life science studies, animal models occupy an indispensable niche in multidisciplinary clinical and basic research. In contrast to smaller animals, such as mice, nonhuman primates (NHPs) exhibit significant similarities in terms of their anatomy, physiology, and genetics with humans.^1,2 For example, macaques exhibit more than 90% similarity in DNA sequence and highly conserved protein sequences with humans, compared with 84% in mice.³ Additionally, NHP eyes possess considerable anatomic similarities with human eyes, such as corneal, iris, lens, vitreous, and retinal images, which makes these experimental animals (e.g. cynomolgus macaque and rhesus macaque) an appealing model for human visual biology and ocular diseases in preclinical studies.³ 

Optical coherence tomography (OCT) is a conventional ophthalmological examination that provides real-time, high-resolution cross-sectional imagery of the macula and optic nerve head (ONH). Spectral-domain OCT(SD-OCT), a noninvasive and convenient in vivo technique, is widely used in clinical practice and serves as an appropriate tool for measuring retinal nerve fiber layer (RNFL) thickness.⁴ The quantitative measurement of RNFL thickness using OCT represents a crucial and objective approach for assessing the early progression of glaucoma and other prevalent neurodegenerative diseases. The lamina cribrosa (LC), a collagenous meshwork situated in the ONH, potentially plays a significant role in the initial pathogenesis of glaucoma and other eye diseases.^5,6 Recently, changes in LC parameters could be assessed and measured more accurately, according to the advancements of swept-source OCT (SS-OCT) and enhanced depth imaging (EDI)-OCT with better tissue penetration.⁴ 

With only a few ocular biological characteristics being reported in NHPs, previous studies have primarily focused on macular-related parameters, the sample size of which was relatively small and unrepresentative.^7,8 Meanwhile, these studies have given little attention to the parameters associated with ONH and LC, even though the changes in these parameters are closely related to the occurrence of glaucoma and other neurodegenerative diseases. Moreover, although NHP experimental models have been utilized to study common eye diseases, such as glaucoma, a deficiency still exists in the establishment of normal distributions of ophthalmic parameters in healthy monkeys. Therefore, to establish the foundation for preclinical visual science research, it is imperative to gain a comprehensive understanding of the normal range of ophthalmic biological structure and retinal anatomic parameters in NHPs. 

Our study aimed to describe the normative profile of peripapillary RNFL thickness and other fundamental ocular parameters of a healthy cynomolgus monkey colony in southern China, which might be beneficial in providing criteria for RNFL-related disease research utilizing NHP experimental models, as well as in further investigating the pathogenesis of glaucoma. 

Methods

Animals and Ethical Statements

All NHPs used in this study were obtained from Huazhen Biosciences, a full Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) – an accredited facility authorized to perform laboratory animal research in China. The monkeys included in this study had not been used for any scientific research during their growth period, thereby avoiding any significant impact on the measurement of ophthalmic parameters. These animals had previously been utilized for breeding purposes only. The group of monkeys used for scientific research is housed separately and subjected to strict experimental procedures, with each monkey being used for only one invasive experiment. To ensure the welfare of the animals, all veterinarians monitoring the healthy status of the cynomolgus have more than 20 years of experience, and their condition is continuously monitored for 24 hours. All cynomolgus monkeys were housed in a climate-controlled room maintained at a temperature of 16°C to 26°C, relative humidity of 40% to 70%, and a 12-hour light/12-hour dark cycle. The monkeys were fed a daily commercial primate diet, drinking water, fresh fruit, and additional supplements. Moreover, toys and music were provided to enhance their welfare. 

All procedures complied with the Association for Research in Vision and Ophthalmology's Statement for the Use of Animals in Ophthalmic and Vision Research guidelines. The study received approval from both the Ethical Committee of the Guangzhou Huazhen Biosciences Company (Ethics Number: 2020-168) and Zhongshan Ophthalmic Center (Permit Number: SYXK (YUE) 2018-0189). 

Standardized Examination Process

All tests were conducted during a single session within approximately 60 to 120 minutes while the monkeys were under general anesthesia. Following standardized protocols, all monkeys underwent basic systemic examinations and multimodal ophthalmic examinations, encompassing weights, autorefraction, intraocular pressure (IOP) measurements at the beginning, middle, and end of anesthesia, slit lamp examination, anterior segment OCT (AS-OCT) images, AL-100 examination, SD-OCT images, and fundus photography (FP). The applied anesthetic doses were within safe thresholds. Moreover, the veterinarian was present on-site to oversee the entire procedure. 

IOP Measurements

IOP was measured by an Icare tonometer (TA01i; Icare). After anesthetizing the monkeys with intramuscular injections of tiletamine/zolazepam 4 mg/kg (Zoletil 50; Virbac) mixed with xylazine/ketamine 0.2 mg/kg (Sumianxin; Shengda Animal Medicine), the technician stabilized the monkey in a sitting position, keeping the monkey’s head in a primary gaze state. Afterward, the instrument automatically gathered and exhibited the mean of six measurements. We obtained six measurements in three series, specifically, measuring IOP at the beginning, middle, and end of anesthesia for both eyes of monkeys, and recording the values. 

Ocular Examination

After anesthesia, an experienced ophthalmologist conducted a comprehensive slit-lamp biomicroscopy (TOPCON Slit-lamp SL-D701) to carefully examine the overall structure of the monkey's anterior segment. Therewith, refraction and corneal curvature were measured without pupil dilation using an autorefractor (FKR-710, fario). The anterior chamber and anterior chamber angle were scanned with AS-OCT (Heidelberg Engineering GmbH, Heidelberg) in the undilated state to determine the degree of opening of the anterior chamber angle and central corneal thickness. Then, AL-100 (TOMEY AL-4000) was used to measure the anterior chamber depth (ACD), lens thickness, and axial length (AXL).⁹ After completing the above examinations, both eyes were dilated with tropicamide phenylephrine (Mydrin; Santen). Skilled ophthalmic technicians then utilized fundoscopy (APS-BER Fundus Camera & FFA model, AITOMU) to capture images of monkeys, with a focus on the macula and optic disc, for each eye separately. 

OCT Examination

OCT images with uniform illumination and good focus were taken using the Spectralis SD-OCT (Heidelberg Engineering GmbH, Heidelberg, Germany). We verified the quality of all OCT images to exclude those with inadequate quality, a signal strength index of less than 15 decibel (dB), and decentration or segmentation errors. A default corneal curvature (7.7 mm) was adopted, the operators adjusted the position of the aiming circle to match the ONH, and then performed a well-conducted 15 degree, 24-line radial scan covering a 3.4-mm diameter circle, including the structure surrounding the ONH.^9,10 Subsequently, cup disc ratio (CDR) and RNFL thickness were evaluated using the automatic measurement software included with the Heidelberg OCT. The RNFL was partitioned automatically into six sectors, superonasal (NS), superotemporal (TS), temporal (T), inferotemporal (TI), inferonasal (NI), and nasal (N) divisions. When the RNFL boundary delineated by the automatic OCT segmentation algorithm was different from the actual anterior and posterior RNFL borders, the segmentation error was corrected manually. The instrument software automatically identifies the innermost end of the Bruch membrane as a Bruch's membrane opening (BMO), and the operators then confirm or correct the position of the BMO. Other scanning modes were also used, such as the retinal thickness measurement consisting of 49 horizontal b-scans, and the specific protocol was described in an already published paper.⁹ 

Inclusion and Exclusion

OCT images of 398 cynomolgus monkeys (765 eyes) were available, whereas 83 eyes that were diagnosed with glaucoma suspect, high myopia, optic nerve diseases, or other retinal diseases were excluded from the analysis. Glaucoma suspects were diagnosed with at least 1 eye that detected at least 2 of the following criteria when independently scored by 2 experienced ophthalmologists: symmetric CDR > 0.5 or disc hemorrhage based on inspection of fundus photo, diffuse or focal narrowing of the RNFL or the neuroretina rim (especially in the superior or inferior sector of the optic disc) based on slit-lamp biomicroscopic optic nerve examination. High myopia was defined as SE ≤ −6 diopters (D), and myopic retinopathy was classified using the International Photographic Classification and Grading System.^9,11 Finally, 349 cynomolgus monkeys (682 eyes) were included in our analysis (Fig. 1). 

Figure 1.

 

Flow diagram showing inclusion criteria for cynomolgus monkeys and the reasons OCT scans were excluded. IOP, intraocular pressure; NAION, non-arteritic ischemic optic neuropathy; RNFL, retinal nerve fiber layer; SD-OCT, spectral-domain optical coherence tomography.

Statistical Analysis

All final LC parameters of the vertical scan were measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA), including cup depth (CD), lamina cribrosa depth (LCD), posterior lamina cribrosa surface depth (PLCSD), prelaminar tissue thickness (PTT), lamina cribrosa thickness (LCT), anterior laminar insertion depth (ALID), mean ALID (mALID), and lamina cribrosa curvature index (LCCI).¹² The specific measurement method is shown in Supplementary Figure S1. All statistical analyses were conducted utilizing the Statistical Package for the Social Sciences version 25.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism software version 9.0. The quantitative data were presented as mean ± standard deviation (SD), whereas qualitative data were characterized in the form of counts and percentages. 

To ensure data consistency and repeatability, we conducted a pilot study with approximately 100 subjects to determine the repeatability of the examination and diagnosis by different examiners. Intraclass correlation coefficient (ICC) and coefficient of variation (CV) were calculated. The results showed that the ICC values were between 0.68 and 0.862, and the CV values were between 9.39% and 22.73%, indicating good consistency and repeatability of our measurements (Supplementary Table S1).¹³ Therefore, we averaged these repeated measurements. The variables were compared using an independent t-test (for continuous variables) and Pearson’s χ² test (for categorical variables). Univariate and multivariate linear regression models were conducted to assess the association between demographic and ophthalmic features and average RNFL, with the statistical significance determined at P < 0.05. 

Results

The comparisons of systemic and ocular characteristics of the included and excluded groups are presented in Table 1. Six hundred eighty-two (682) eyes of 349 monkeys were included in the analysis (224 female and 125 male monkeys). The mean age and weight of included monkeys were 18.69 ± 2.88 years old and 4.54 ± 1.19 kg, respectively. In comparison to the ocular characteristics observed in the excluded group, the included monkeys exhibited lower IOP at the beginning of anesthesia (P = 0.002), AXL (P < 0.001), and SE (P < 0.001). Furthermore, the RNFL thicknesses in the included group were thicker in the global (G; P = 0.003), NS (P = 0.013), TS (P = 0.005), TI (P < 0.001), and NI (P = 0.001) sectors. 

Table 1.

 

Comparisons of Baseline Characteristics Between the Included and Excluded Groups

 

Figure 2 shows the distribution of basic ocular parameters in the included group. The average ACD, AXL, and IOP at the end of anesthesia and RNFL thickness of included monkeys were 3.25 ± 0.3 mm (1.93 to 4.08), 18.57 ± 0.65 mm (18 to 22.24), 19.42 ± 4.06 mm Hg (11 to 38), and 94.61 ± 10.13 (71 to 155), respectively. 

Figure 2.

 

Distribution of basic ophthalmic parameters in healthy monkeys. (A) ACD. (B) AXL. (C) IOP at the end of anesthesia. (D) RNFL thickness. ACD, anterior chamber depth; AXL, axial length; IOP, intraocular pressure; RNFL, retinal nerve fiber layer.

Systematic and biological differences between male and female groups are shown in Supplementary Table S2. The average age of the male group and female group was 17.93 years (5–23) and 18.86 years (6–26), respectively. There were statistical differences in weight, IOP, AXL, and SE between male and female groups, and the female group had lower weight (P < 0.001), IOP (P = 0.009), AXL (P < 0.001), and SE (P = 0.032). 

The average RNFL thickness for all the included monkeys was 94.61 ± 10.13 µm. The average RNFL thickness was 3.85 µm thinner in female monkeys than in male monkeys (P = 0.019). Sex-specific differences existed in all sectors. However, statistically significant differences were only observed in the NS (P = 0.016) and T (P = 0.031) sectors (Table 2). Based on the general distribution trend of RNFL, the RNFL was thickest in the TI quadrant, followed by TS, NI, NS, and T, and thinnest in the N quadrant (Fig. 3). Then, we compared the RNFL thickness of the young (age ≤ 18 years) and old (age > 18 years) groups (Supplementary Table S3). The RNFL thickness of the young group was thicker than that of the old group in the G, TS, T, TI, and N sectors, and there was a statistical difference, except for the N sector (P = 0.626). However, the RNFL was thinner in the young group than in the old group in the NI region (P = 0.521). 

Table 2.

 

Distribution and Comparison of RNFL and LC-Related Parameters in Healthy Eyes by Sex

 

Table 3.

 

Associations Between Mean Global RNFL Thickness as Measured by Spectral-Domain Optical Coherence Tomography and Systemic and Ophthalmic Parameters

 

Figure 3.

 

Distribution of RNFL thickness in different sectors by gender. N, nasal; NI, inferonasal; NS, superonasal; RNFL, retinal nerve fiber layer; T, temporal; TI, inferotemporal; TS, superotemporal.

The comparison of LC-related parameters between female and male groups is shown in Table 2. The CD in the male group was significantly deeper than that in the female group (P < 0.001). Nevertheless, the two groups’ LCD was opposite (P = 0.02). The ALID1 and mALID were significantly deeper in the female group than in the male group (P = 0.009 and P = 0.029, respectively). The LCT and BRO – minimum rim width 2 (BMO-MRW2) of the male group (169.69 µm and 336.22 µm, respectively) were significantly thicker than those of the female group (155.16 µm, P = 0.014 and 320.37 µm, P = 0.002). 

Both univariate and multivariate linear regression analyses were conducted to analyze the relationship between systemic and ophthalmic parameters and mean global RNFL thickness (Table 3). In the univariate regression analysis, global RNFL thickness showed a significant negative linear correlation with age (P < 0.001), sex (P = 0.02), IOP in the middle of anesthesia (P = 0.03), AXL (P = 0.01), ALID2 (P = 0.04) and mALID (P = 0.03), and a positive correlation with BMO-MRW2 (P = 0.04). Then, these related parameters were incorporated into the multivariate analysis, the results demonstrated that, except for the negative correlation between age and global RNFL thickness (P < 0.001), the other parameters did not show any significant correlation with RNFL thickness. 

Discussion

Our research assessed the normative distribution of RNFL thickness around the optic disc, LC-related parameters, and other basic ocular parameters in a healthy non-glaucoma suspect cynomolgus monkey colony. A total of 398 adult cynomolgus monkeys were included in our study, of which 33 (8.29%) were diagnosed with high myopia, 3 (0.75%) as glaucoma suspects, 1 (0.25%) as non-arteritic ischemic optic neuropathy (NAION), and 12 (3.02%) as other RNFL-related diseases (3.02%). Excluded monkeys tended to have higher initial IOP, AXL, and SE. Equally importantly, their RNFL thickness was thinner than healthy monkeys in most quadrants, except for the temporal side.¹⁴ 

In our study, the mean RNFL thickness for a healthy cynomolgus monkey colony was 94.61 ± 10.13 µm. Compared with other studies, our results showed that the RNFL thickness in each quadrant was thinner, which may be partially explained by the different instruments and the average order age of our study (Table 4). However, it still followed the (inferior ≥ superior ≥ nasal ≥ temporal (ISNT) pattern) rules, as observed in previous studies.⁸ Afterward, we further analyzed the gender-specific differences in ocular parameters in the included groups. In terms of RNFL thickness, our results indicated that female monkeys showed a thinner trend in the measurement of RNFL thickness, which has statistical significance in the G, T, and NS quadrants (P < 0.05). Although Pasquale et al. found a similar gender-specific trend in a healthy rhesus macaque colony, their results were not statistically significant, possibly due to the smaller sample size included in their study.⁸ Contrary to our findings, the population-based studies found that female monkeys had thicker RNFL thickness in most quadrants than male monkeys (Table 5).¹⁵ Hence, further analysis of gender differences in RNFL thickness between different races is required. 

Table 4.

 

Comparison of RNFL Thickness With Other Non-Human Primate Studies

 

Table 5.

 

Comparison of RNFL Thickness With Other Population-Based Studies

 

There is a growing recognition that alterations in RNFL thickness are promising early biomarkers of glaucoma and other retinal diseases. Hence, we sought to find systemic and ocular parameters associated with modifications in RNFL thickness in healthy cynomolgus monkey colonies. However, the univariate and multivariate linear regression analysis revealed that only age and RNFL thickness exhibited a negative linear correlation. Age-related reduction in RNFL thickness may be attributed to reduced blood supply, optic nerve aging, and apoptosis,²² which indicates that further studies of the distribution of RNFL thickness among different age groups will be needed in the future. However, no correlations were observed between other ocular or systemic parameters and RNFL thickness. This may be due to the slow rate of global RNFL loss (<1 µm/year) in POAG occurrence and progression.²³ Although univariate linear regression analysis showed a negative linear relationship between high IOP and longer AXL and RNFL, these factors were not statistically significant in multivariate regression analysis. Yoo et al. revealed that the effect of increased AXL on RNFL varied across different quadrants.²⁴ However, we did not further evaluate the relationship between various parameters and RNFL in separate quadrants. 

Earlier investigations have showcased the impact of high IOP on LC, which not only deforms but also reshapes due to biomechanical forces. Being closely related to the progression of RNFL, the LC curvature index may be a pivotal marker for evaluating the advancement of glaucoma.¹² Therefore, it is imperative to assess the average values of LC-related parameters and the relationship between them and the RNFL. Specific LC-related parameters were statistically different between male and female monkeys. CD, LCT, and BMO-MRW2 were more remarkable in the male group. However, LCD, ALID1, and mALID of female monkeys were more considerable than those of male monkeys. Previous findings showed that the measurement of ONH parameters and RNFL thickness by OCT in healthy monkey eyes have high repeatability and reproducibility, similar to those reported in human eyes.²⁵ However, our study failed to notice the apparent correlation between LC-related parameters and RNFL, possibly because only healthy subjects were included and the sample size was relatively small. Hence, a larger sample size and further analyses are required to assess the optic nerve architecture in monkey populations. 

Because the morphology of ONH and pathophysiological changes of experimental glaucoma of cynomolgus and rhesus monkeys are similar to those of human patients with glaucoma, monkey eyes are a helpful model for studying glaucoma and other RNFL-related diseases.²⁶ In recent years, OCT, which has been extensively used to quantitatively and objectively measure the ONH structure of both humans and monkeys, has emerged as a crucial diagnostic for glaucoma and other retinal diseases.²⁵ However, the existing OCT platform does not possess standardized data for monkeys. It has been demonstrated that RNFL thickness is influenced by race,^7,27 implying that the standard database of human RNFL thickness may have errors when measuring monkeys via OCT. With an increasing number of animal experiments being conducted, it is necessary to investigate the normal distribution range and gender differences of RNFL thickness and LC-related parameters in healthy monkeys. Such research provides essential control and baseline information for experimental glaucoma studies.²⁸ Moreover, it is crucial to establish normal ranges of OCT parameters in monkeys. Our results provide an essential reference value for the establishment of such a database. 

Our investigation presents several noteworthy advantages. First, the study encompassed a significant sample size of 398 monkeys, and comprehensive ophthalmic evaluations were conducted, yielding detailed, reliable, and reproducible parameters. Normative databases for ocular parameters specific to the monkey colony were established. Second, our results provide fundamental biological and OCT data on the retinal anatomy of cynomolgus monkeys, after excluding various confounding diseases that could affect RNFL thickness. Third, to furnish a reference for future research endeavors investigating glaucoma pathogenesis, we concurrently evaluated several potential risk factors associated with global RNFL thickness. 

Our study has limitations. First, it included middle-to-old monkeys with a relatively narrow age span, making it difficult to observe and interpret age-related changes in ophthalmic characteristics. Second, data inclusion were limited. Specifically, our analyses did not include macular region OCT data, which was indispensable in the assessment of early glaucoma and other fundus diseases.²⁹ Third, LC-related parameters were subjectively measured by clinicians, which may be partially biased. Fourth, we did not consider the relationship among RNFL thickness, ONH parameters, and AXL, so the corrections accounting for AXL were applied to the measured values, and a default corneal curvature was adopted. In fact, due to the difference between NHP and human eyes, factors such as refractive error, corneal curvature, and AXL need to be taken into account in the measurement process to increase the reliability of the results. 

Conclusions

In conclusion, we measured the distribution of RNFL and ONH parameters in cynomolgus monkeys. Our findings demonstrated that the RNFL thickness in monkeys was thinner than in the human OCT database, underscoring the necessity to establish an RNFL database in experimental animals. In addition, we provide an unprecedented reference for upcoming studies to understand the normal condition of the monkey eye and the pathogenesis of eye diseases such as glaucoma. 

Acknowledgments

The authors thank the Nonhuman Primate Eye Study Group, the Guangzhou Huazhen Biosciences Company, and all the staff at Huazhen Laboratory Animal Breeding Centre for their valuable contributions to this study. 

Supported by the National Natural Science Foundation of China (82130029 to N.W. and 81901202 to C.Y.), the Fundamental Research Funds for the Central Universities (PKU2021LCXQ007 to C.Y.). 

Author Contributions: J.W. and R.L. completed the design, data collection, data analysis, and manuscript writing of this study, they also contributed equally as co-first authors. S.Z., C.K., L.C., T.J., P.L., L.H., J.X., Y.Z., L.Z., Z.Y., L.W., and Y.C. participated in the data collection and method design of this study. J.T. and C.L. participated in the diagnosis of the disease. N.W. and Z.Y. were the corresponding authors of this study, are responsible for the overall content as guarantors, accept full responsibility for the finished work and the conduct of the study, have access to the data, and control the decision to publish. 

Disclosure: J. Wu, None; R. Li, None; S. Zhu, None; K. Chen, None; C. Lin, None; J. Tian, None; L. Pan, None; H. Liu, None; X. Jia, None; Z. Yu, None; Z. Li, None; Y. Zhu, None; W. Liu, None; C. Yang, None; C. Wong, None; N. Wang, None; Y. Zhuo, None