Influence of Growth Differentiation Factor-15 on Intraocular Pressure in Mice

Abstract

Growth differentiation factor-15 (GDF15), a stress sensitive cytokine, and a distant member of the TGF-β superfamily, has been shown to exhibit increased levels with aging, and in various age-related pathologies. Although GDF15 levels are elevated in the aqueous humor (AH) of glaucoma (optic nerve atrophy) patients, the possible role of this cytokine in modulation of intraocular pressure (IOP) or AH outflow is unknown. The current study addresses this question using transgenic mice expressing human GDF15 and GDF15 null mice, and by perfusing enucleated mouse eyes with recombinant human GDF15. Treatment of primary cultures of human trabecular meshwork (TM) cells either with a telomerase inhibitor, an ER (endoplasmic reticulum) stress inducing agent, hydrogen peroxide, or an autophagy inhibitor, resulted in significant elevation in GDF15 levels relative to the respective control cells. Recombinant human GDF15 stimulated modest but significant increases in expression of genes encoding the extracellular matrix, cell adhesion proteins, and chemokine receptor (CCR2) in human TM cells compared to controls, as deduced from differential transcriptional profiles by RNAseq analysis. There was a significant increase in IOP in transgenic mice expressing human GDF15, but not in GDF15 null mice, compared to the respective wild type control mice. AH outflow facility was decreased in enucleated wild type mouse eyes perfused with recombinant human GDF15. Light microcopy based histological examination of the conventional AH outflow pathway tissues did not reveal identifiable differences between the GDF15 targeted and control mice. Taken together, these results reveal modest elevations of IOP in mice expressing human GDF15 possibly stemming from decreased AH outflow through the trabecular pathway.

INTRODUCTION

Ocular hypertension due to elevated intraocular pressure (IOP) is a major risk factor for glaucoma (optic nerve atrophy). Glaucoma is a leading cause of irreversible blindness globally.^1,2 Although elevated IOP results from impaired drainage of aqueous humor (a clear fluid in the eye anterior chamber) through the trabecular pathway, which consists of the trabecular meshwork (TM), juxtacanalicular tissue and Schlemm’s canal (SC),³ the cellular and molecular basis underlying increases in IOP remains poorly understood.^3,4 Since lowering IOP remains the current mainstay of glaucoma treatment,⁵ there is a great deal of ongoing effort not only to identify novel molecular targets for developing efficacious IOP lowering drugs against, but also to understand the etiological mechanisms involved in the pathobiology of ocular hypertension.^6–8

We recently identified that growth differentiation factor-15 (GDF15) is secreted by human TM and associating with extracellular matrix.⁹ Importantly, we and others have also reported significant elevation in the levels of GDF15 in the aqueous humor (AH) of glaucoma patients relative to non-glaucoma controls.^10–13 GDF15 is a stress induced cytokine distantly related to the transforming growth factor-β superfamily.¹⁴ GDF15 is produced and secreted by various tissues and cell types, and its levels are elevated in disease,^14,15 and ageing.^16,17 The mature secretory form of GDF15 is a homodimer of ~ 30 kDa.¹⁴ GDF15 has been unambiguously demonstrated to play a crucial role in energy homeostasis and in influencing food intake and body weight in different species.^14,18–21 Although GDF15 has been shown to elicit various cellular and physiological responses in target cells and tissues,²² glial cell line-derived neurotrophic factor family receptor α-like (GFRAL)^18–20 was recently identified as a GDF15-specific binding receptor. Intriguingly however, GFRAL expression is known to be restricted discretely to the hindbrain area postrema and solitary tract nucleus of brain.^18,21,23,24 Upon binding to GDF15, GFRAL activates its co-receptor tyrosine kinase RET, to regulate intracellular signaling pathways.^14,18,20,24 GFRAL expression is not detectable in tissues outside the brain including human TM cells.^14,24 At large, it is unclear how GDF15 mediates the myriad effects attributed to it, in non-neuronal cells and tissues lacking the GDF15 canonical receptor GFRAL.^14,15

Since GDF15 levels are elevated in different types of glaucoma and to exhibit a correlation with the degree of degeneration or loss of retinal ganglion cells in glaucoma patients,^10,11,13 this study was undertaken to address whether changes in GDF15 expression influence IOP. The results of this study reveal elevated IOP in mice expressing human GDF15.

MATERIALS AND METHODS

Mice

All mouse experiments described in this study were carried out in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and the Association for Research in Vision and Ophthalmology. The animal protocol (A213–19-10) was approved by the Institutional Animal Care and Use Committee of the Duke University School of Medicine. Well characterized GDF15 null mice and human GDF15 expressing transgenic mice maintained on C57BL/6J genetic background (males and females) along with their respective wild type control mice were used.^25,26 The GDF15 transgenic mice used in this study ubiquitously overexpress human GDF15 under control of a chicken β-actin promoter.²⁶ Mice were fed ad libitum and maintained at 21°C with a 12 hour light-dark cycle, and both GDF15 transgenic and null mice breed, grow and age normally.

Breeding pairs of GDF15 transgenic (TG) mice originally developed and characterized at National Institute of Environmental Health Sciences, (NIEHS) Research Triangle Park, NC, by Thomas Eling’s group, Ph.D.,²⁶ were used in this study and the breeding colony was developed at Duke University. The following oligonucleotide primers were used to genotype the GDF15 transgenic mice as described previously.²⁶ GDF-15 TG mice: GDF15 forward 5’-GTGCTGGTTATTGTGCTGTCTC-3’, and GDF15 reverse 5’-AGTCTT CGGAGTGCAACTCTGAGG-3.

We obtained breeding pairs of well characterized GDF15 knockout (KO) mice originally developed by the laboratory of Se-Jin Lee, M.D., Ph.D. University of Connecticut Health Center.²⁵ The following primers were used to genotype the GDF15 null mice, with SS2+SS3 combination, yielding WT allele (228bp) and SS1+SS4, yielding KO allele (598bp):

SS1: 5’-GCCTTCT TGACGAGTTCTTCTGAGGG-3,

SS2: 5’-CCTGGAGACTGTGCAGGCAACTCTTG-3,

SS3: 5’GTGACACACCACTGTCTGTCCTGTGC-3,

SS4: 5’-GCTGTCCGGATACTCAGTCCAGAGG-3,

Trabecular meshwork cell cultures and treatments

Trabecular meshwork (TM) cell cultures were isolated from male and female human donor eye corneal rims (leftover tissue from corneal transplantation surgeries performed at the Duke Ophthalmology Clinical Service). The use of human tissue has been approved by the Institutional Review Board of Duke University School of Medicine (Pro00050810), in compliance with Health Insurance Portability and Accountability Act guidelines, and the tenets of the Declaration of Helsinki.

Primary TM cells were isolated and cultured from TM tissue dissected from corneal rims (donors with no known ocular complications) as we described previously.²⁷ Experiments carried out with TM cell cultures were performed as per the consensus recommendations described in Keller et. al.²⁸ TM primary cell cultures (derived from 18y; 22y; 27y; 48y and 65y-old donors) used in this study were from passage 2 to 6. Cells were cultured in Dulbecco’s Modified Eagle complete growth medium (DMEM) containing 10% FBS (heat inactivated fetal bovine serum), penicillin (100 U/500 ml, streptomycin (100 μg/500 ml) and glutamine (4 mM) at 37°C in an aseptic incubator under 5% CO₂.

Human TM cells grown to confluence in cell culture plastic dishes were maintained under serum free conditions for 24 h prior to drug treatment. Cells were separately treated either with 20 μM Telomerase inhibitor (BIBR1532, Cat. No: HY-17353, MedChemExpress, NJ. USA) for 24 h to induce senescence, Tunicamycin (1μg/1ml media; Cat no: T7765, Millipore Sigma, St. Louis, MO. USA) for 48 h to induce ER (endoplasmic reticulum) stress, H₂O₂ (200 μM; Cat no: H1009, Millipore Sigma) for 3 h to induce oxidative stress, or with 2 μM IITZ-01, an autophagy inhibitor for 6 h (Cat. No: 1807988-47-1, Selleckchem.com Houston, TX. USA).

Following the above drug treatments, cells were washed with 1x cold phosphate buffered saline (PBS) and incubated on ice for 5 min with 10% ice-cold trichloroacetic acid (TCA) and 0.5 M dithiothreitol (DTT). Following several washes with cold deionized water, cells were scraped and transferred into Eppendorf tubes, and washed again with cold deionized water and a final wash with diethyl ether. Precipitates obtained after centrifugation at 16000xg were suspended in 8 M urea buffer containing 20 mM Tris, 23 mM glycine, 10 mM DTT and saturated sucrose, protease and phosphatase inhibitors as mentioned above, and briefly sonicated. Protein concentration was determined in the 800xg supernatants using the Micro BCA protein assay kit (Cat. No. 23235, Thermo Fisher Scientific, Waltham, MA. USA) per manufacturer’s instructions.

RNA extraction, library construction and RNA-seq analysis

To investigate the effects of GDF15 on gene expression in TM cells, human TM cells (derived from 25, 29 and 65 year old donors; passage 3) were grown to 80% confluence in complete growth medium, then serum starved for 24 h prior to treatment with 20 ng/ml recombinant human GDF15 (rhGDF-15; Cat. no: 957-GD/CF, Lot no: EHF2320121; R&D Systems, Minneapolis, MN. USA) for 24 h. The following day, after a medium change (with serum-free media) cells were again treated with 20 ng/ml of rh GDF-15 for an additional 24 h. At the end of 48 h (total) of treatment with rh GDF15, cells were harvested for RNA extraction using the RNeasy Micro kit (Qiagen, Inc., Valencia, CA, USA). RNA was extracted per manufacturer’s protocol.

RNA was quantified using a Qubit instrument (Thermo Fisher Scientific, Waltham, MA USA). RNA quality was analyzed on an Agilent TapeStation Model 2200 (Agilent Technologies, Santa Clara, CA, USA). Stranded mRNA-seq libraries were constructed using the Kapa kit (Kapa Biosystems, Inc, Wilmington, MA, USA) following the manufacturer’s instructions with the help of the Duke Center for Genomic and Computational Biology core service. Briefly, RNA-Seq libraries were indexed, normalized to a concentration of 10 nM and pooled in equimolar ratio. The final pool of libraries was sequenced in one lane of an Illumina HiSeq 4000 sequencing flow cell (Illumina, Inc, San Diego, CA. USA) with a 50 bp single end read length. The resultant raw bcl files were converted into fastq files and sequences demultiplexed using Illumina bcl2fastq conversion software version 2 (Illumina, Inc).

RNA-seq data were processed using the TrimGalore toolkit which employs Cutadapt to trim low-quality bases and Illumina sequencing adapters from the 3’ end of the reads. Only reads that were 20nt or longer after trimming were utilized for further analysis. Reads were mapped to the GRCh38v93 version of the human genome and transcriptome using the STAR RNA-seq alignment tool. Reads were used for subsequent analysis if they mapped to a single genomic location. Gene counts were compiled using the HTSeq tool. Only genes that had at least 10 reads in any given library were used in subsequent analysis. Normalization and differential expression analysis was carried out using the DESeq26 Bioconductor package with the R statistical programming environment. The false discovery rate was calculated to control for multiple hypothesis testing. Gene set enrichment analysis was performed to identify gene ontology (GO) terms and pathways associated with altered gene expression for each of the comparisons performed.

Immunoblot analysis

Equal amounts of protein (10 μg) were mixed with Laemmli sample buffer, separated by sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose membranes. Membranes were blocked for 2 h at room temperature in Tris buffered saline (TBS) containing 5% (wt/vol) nonfat dry milk and 0.1% Tween 20 and subsequently probed overnight at 4°C with human GDF15 mouse monoclonal antibody (Cat. no: sc-377195, Santa Cruz Biotechnology, Inc. Dallas, TX. USA). Membranes were washed with TBS buffer containing 1% Tween-20 and incubated with a secondary antibody (goat anti-mouse IgG secondary antibody, HRP; Cat no: 31430, Thermo Fisher Scientific) for 2 h at room temperature. Immunoblots were developed by enhanced chemiluminescence (Millipore Sigma), followed by scanning and analysis using ChemiDoc Touch imaging and Image Lab^™ Touch Software (Bio-Rad Laboratories, Hercules, CA. USA), respectively. For loading controls, GAPDH (glyceraldehyde 3-phosphate dehydrogenase) was immunoblotted for normalization using GAPDH mouse monoclonal antibody (Cat no: 6004–1-Ig Proteintech, Rosemont, IL. USA).

ELISA for analysis of serum GDF15

Blood droplets were collected into plastic tubes after puncturing the submandibular vein on either sides of the cheeks, using a 6mm animal lancet (Goldenrod Medipoint Inc, Mineola, NY. USA). Blood samples were held at room temperature for 30 min and then centrifuged using a benchtop Eppendorf centrifuge for 15 min at 1000xg. Serum samples were collected into low binding Eppendorf tubes and stored at −80 °C until ELISA quantification. A human GDF15 enzyme-linked immunosorbent assay kit (human GDF15 DuoSet ELISA, R&D Systems, Inc.) was used to determine the levels of GDF15 in serum samples in accordance with the manufacturer’s protocol, which included appropriate standards and background controls, and using a SpectraMax M3 plate reader (Molecular Devices, San Jose, CA, USA). Serum samples were diluted 10-fold with diluent buffer provided in the kit, and 10 μl of the diluted sample was analyzed in duplicate. Results are expressed as picograms of GDF15/ml.

IOP recording

To evaluate IOP changes in GDF15 transgenic and null mice, either 6 and 10 month-old male and female GDF15 transgenic animals were used along with their respective littermate controls, or 10 month-old male and female GDF15 KO mice along with their age matched wild type (WT) control were used. IOP was monitored under mild anesthesia (Isoflurane), using a rebound Tonolab tonometer (iCare Laboratory, Espoo, Finland). Mice were positioned with the visual axis horizontal to the probe and 5 to 6 readings were recorded and averaged.

Aqueous humor collection

After injecting animals with a lethal dose of Euthasol, a 33-gauge needle with 50 μl Hamilton syringe was inserted through the cornea and gently aspirated AH from both eyes. AH samples collected in Eppendorf tubes were centrifuged (800xg) to obtain the supernatant which was stored at −20 °C. AH pooled from three mice (6 eyes) was used for immunoblotting analysis using an anti-human GDF15 mouse monoclonal antibody.

Aqueous humor outflow facility

To measure and analyze the direct effects of human rGDF15 on AH outflow in enucleated mouse eyes, perfusion studies were performed using an iPerfusion system as described previously.^29,30 The outflow measurements were performed on C57/BL6 WT mice (8–9 months old, males/females). Mouse eyes were enucleated after sacrificing animals, and contralateral eyes were fixed to the platform inside the warmed bath with cyanoacrylate glue (Loctite). An XYZ micromanipulator (World Precision Instruments, Sarasota, FL. USA) was used to cannulate both the eyes via the anterior chamber, using a 33 gauge needle as we described earlier.³⁰ The control perfusate (Dulbecco’s PBS containing divalent cations), supplemented with 5.5 mM glucose was passed through a 0.2 μm sterile filter. After cannulation, the bath was filled with PBS to fully submerge the eye and the temperature was raised to 35°C. The applied pressure was held at 8 mmHg for a period of 30–45 min to allow the eye to acclimatize to the pressure and temperature, after which eyes were perfused with media containing either rh GDF15 (4 ng/ml) or vehicle (PBS), with continuous monitoring of IOP for 90 min with nine sequential pressure steps (10 min each) of 4.5, 6.0, 7.5, 9.0, 10.5, 12.0, 15.0, 18.0, and 21.0 mm Hg. Data were analyzed as described previously.³⁰ A non-linear flow-pressure model was used to account for the pressure dependence of outflow facility in mice, and the reference facility was analyzed at the reference pressure of 8 mmHg as described by Sherwood et.al.²⁹

Histological analysis

Enucleated GDF15 transgenic (TG) and null (KO) mouse eyes along with their respective WT control mouse eyes were fixed with 2% paraformaldehyde and 2% glutaraldehyde in PBS for 24 h at 4°C (tiny incisions were made to the cornea and retina of the enucleated eyes to enable entry of the fixative). Fixative was removed and the tissues were washed with PBS before dissecting fixed eyes. The anterior segments were cut into four quadrants and embedded in epoxy resin prior to generating semi-thin sections (0.5 μm). These sections were subsequently stained with toluidine blue to enable visualization under light microscopy and imaging.

Gene ontology enrichment analysis

Genes were annotated with gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) terms and subjected to enrichment analysis using the ShinyGO 0.77 tool via Fisher’s exact test (FDR = 0.05).

Statistical analysis

All data derived from cell culture studies, are presented as mean ± SEM (standard error of the mean) values, and are based on the use of at least two independent cell strains. GraphPad prism software (version 9.3.1; GraphPad Software) was used for statistical analyses, and comparisons were analyzed using the Student’s t-test (for the biochemical data), with P < 0.05 being considered statistically significant. The nonparametric Wilcoxon rank sum test of differences between medians was performed to compare IOP and AH facility between WT vs TG or KO animals.

RESULTS

Increased levels of GDF15 in TM cells exposed to various stresses

In our previous study, GDF15 levels were found to be upregulated in response to growth factor, glucocorticoid, and cytokine treatment of human TM cells.⁹ Since GDF15 has been characterized as a stress inducible cytokine,^21,31,32 we further investigated the effect of various stressors on the levels of GDF15 in human TM cells. Interestingly, endoplasmic reticulum stress (Tunicamycin; 1μg/ml for 48 h), oxidative stress (200 μM H₂O₂ for 3 h), senescence (telomerase inhibitor; 20 μM for 24 h), and inhibition of autophagy (IITZ-01, 2 μM for 6 h) were each found to elicit significant (n=5 to 6; minimum at P<0.05) increases in the levels of GDF15 in 2 different strains of human TM cells, based on immunoblot analysis compared to control TM cells (Fig. 1). Cells were found to be attached to cell culture dishes and did not exhibit noticeable changes in cell morphology under any of the listed drug treatments, excluding contribution from cell death to the observed changes in GDF15 levels.

Figure. 1. Stress-induced increase in the levels of growth differentiation factor-15 in human trabecular meshwork cells.

Human trabecular meshwork (TM) cells treated with agents inducing senescence (20 μM BIBR1532, telomerase inhibitor for 24 h, ER stress (1 μg/ml Tunicamycin, 48 h), oxidative stress (200 μM H₂O_2, 3 h) and autophagy inhibitor (2 μM IITZ-01, 6 h) significantly (** P<0.001) elevated the levels of growth differentiation factor-15 (GDF15) compared to control cells based on immunoblot analysis. N=5 to 6.

Recombinant GDF15 induces expression of ECM encoding genes in human TM cells

To gain unbiased insight regarding the effects of GDF15 on TM cell biology, we evaluated the effects of recombinant human (rh) GDF15 on the transcriptome profile of human TM cells by differential RNAseq analysis in comparison to untreated control TM cells. For this purpose, we used three strains of human TM cells (primary cultures, passage 3, derived from 25-, 29- and 65-year-old donors). Cells maintained in serum free culture medium for 24 h were treated with 20 ng/mL rh GDF15 for 48 h (media changed out with fresh GDF15 containing medium after 24 h) prior to evaluating changes in gene expression profile by RNAseq analysis in comparison with control (untreated) cells. A total of nearly 20, 000 genes were found to be expressed in control and GDF15 treated cells. Principal Component Analysis (PCA) however, revealed that the largest contributor to variance of individual gene counts was not GDF15 treatment, but rather the cell strain per se, as shown in Fig. S1. The variation in individual gene counts between control and treatment samples from the same cell strain was found to be much smaller for all three cell strains tested (Fig. S1). Interestingly, heat map construction (z-score normalized) and hierarchical clustering of the differentially expressed genes (DEGs; defined as genes exhibiting a fold-change ≥ 1.5 and an adjusted p-value ≤ 0.05) revealed that only a small number of genes (16 genes) exhibit differential expression between control and GDF15 treated TM cells (Fig. 2A). The volcano plot shows the results of the comparison between GDF15 treated and control cells (Fig. 2B), where the log2-fold changes calculated based on GDF15/control cells, are plotted on the x-axis, and the – log10 p-values are plotted on the y axis, and each dot represents a gene. Table 1 lists the individual genes which exhibited significant differential expression in GDF15 treated TM cells. Of the 16 genes which exhibited differential expression, only one gene (WSB1 encoding WD repeat and SOCS box-containing protein 1) appeared to be downregulated under GDF15 treatment. Interestingly, most of the 15 upregulated genes including FBN1, VCAN, COL6A3, SVEP1, TOP2A, COL12A1, PRKDC, FAT1, CENPF, ADAMTS1, IGF2R, STC2, LAMC1, LAMA2, and CCL2 were found to be either extracellular matrix, cell adhesion, chemokine or migration related genes. GO enrichment analysis of the differentially expressed genes in GDF15 treated cells for molecular function and cellular components revealed enrichment of predominantly ECM proteins (Fig. S2A & S2B).

Figure. 2. Upregulation of expression of genes encoding extracellular matrix proteins and other proteins in human TM cells treated with recombinant human GDF15.

Serum starved human TM cells (three strains; n=3) treated with rh human GDF15 (20 ng/ml for 48 h) are evaluated for differential gene expression based on RNAseq analysis. A). Heat map of expression values of differentially expressed genes in rh GDF15 treated TM cells compared to control TM cells. Only differentially expressed genes (≥ 1.5 fold difference in expression, adjusted P value of ≤ 0.05) were used for creating the heat map. Three individual TM cell strains (treated and control, side by side columns) are shown. Each box within each column represents the expression of indicated gene. The genes are hierarchically clustered using a correlation distance with complete linkage and show the similarity of their expression profiles between rh GDF15 treated and control TM cells. B). The volcano plot depicting the results of the comparison between rhGDF15 treated and control TM cells. The log2-fold changes which were calculated based on rhGDF15 treated/control were plotted on the x-axis. The – log10 p-values are plotted on the y axis. Each dot represents a gene.

Table 1:

GDF15 induced differential gene expression in human TM cells

Gene Name	Log2 Fold Change (GDF15/Control)	Wald test	pvalue	FDR corrected p-Value
FBN1	0.333	8.74	2E-18	4.54135E-14
VCAN	0.396	7.92	2E-15	2.34483E-11
COL6A3	0.263	6.44	1E-10	7.82264E-07
SVEP1	0.225	6.09	1E-09	5.51834E-06
TOP2A	0.496	5.94	3E-09	1.12276E-05
COL12A1	0.181	5.85	5E-09	1.64762E-05
PRKDC	0.223	5.7	1E-08	3.40437E-05
FAT1	0.194	5.21	2E-07	0.000484052
CENPF	0.605	5.08	4E-07	0.000840682
ADAMTS1	0.235	5.06	4E-07	0.000850758
IGF2R	0.204	4.93	8E-07	0.001528115
STC2	0.196	4.87	1E-06	0.001887728
LAMC1	0.171	4.71	2E-06	0.003778881
LAMA2	0.185	4.38	1E-05	0.017226863
WSB1	−0.200	−4.2	2E-05	0.030038683
CCL2	0.175	4.13	4E-05	0.045001574

Elevated IOP in transgenic mice expressing human GDF15

We used a well characterized transgenic mouse model expressing human GDF15 under the chicken β-actin promotor and confirmed the expression of human GDF15 in the serum of only transgenic mice (n=12) by ELISA analysis, as shown in Fig. 3A. Additionally, as has been reported previously,³³ the body weights of mice expressing transgenic human GDF15 were found to be significantly reduced in both 6- (n=8, 21%, P<0.005) and 10-month-old (n=10, 32%, P<0.0007) mice compared to littermate wild type mice (n=7 and 8, respectively, Fig. S3A). We also confirmed the expression of human GDF15 in the AH of transgenic mice based on immunoblot analysis of pooled samples (from 3 mice) as shown in Fig. 3B.

Figure 3: Elevated intraocular pressure in transgenic mice expressing human GDF15.

To determine the effects of GDF15 on intraocular pressure (IOP), changes in IOP were evaluated in the transgenic mice (in 6 and 10 month-old) expressing human GDF15 under a control of chicken β-actin promoter in comparison of littermate wild type mice. A). Confirming the expressing of human GDF15 in the serum of only the transgenic mice (n=12) as determined by enzyme-linked immunosorbent assay. Human GDF15 levels are robustly (**** P<0.0001; n=12) detected in the serum of only the transgenic mice compared to control mice (n=8). B). Human GDF15 was also detected in the AH of only the transgenic mice based on immunoblot analysis of the pooled sample (n=6 eyes) of AH compared to control mice. C & E). IOP in 6 (n=14 eyes) and 10-month old (n=27 eyes) transgenic mice was significantly elevated (* P<0.011 and P<0.02, respectively) compared to the littermate wild type mice (n=12 & 22 eyes, respectively. D & F). The transgenic (6 and 10 month-old, respectively) and littermate wild type mice exhibit a comparable histological integrity of the AH outflow pathway based on light microscopy imaging. TM; trabecular meshwork, SC; Schlemm’s canal, TG; transgenic, WT; wild type, LC; loading control. Bars indicate image magnification.

Changes in IOP were evaluated in 6- and 10-month-old transgenic and littermate wild type mice under mild anesthesia (isoflurane), using a rebound tonometer. IOP in transgenic 6- month-old (16.26 mmHg, median, n=14 eyes) and 10-month-old (17.50 mmHg; median, n=27 eyes) mice was higher compared to the corresponding values measured in the respective control mice (13.25 mmHg, n=12, and 15.90 mmHg and n=22, respectively). In both the 6- and 10-month-old GDF15 transgenic mice, IOP was significantly elevated by ~16% (P<0.011) and ~10% (P<0.028), respectively, relative to the respective control mice (Fig. 3 C & E). Tables S1A and S1B provide additional details of IOP analyses in the GDF15 transgenic and control mice.

The elevated IOP recorded in transgenic mice was not associated with noticeable differences in the histological integrity of the conventional outflow pathway (including TM and SC) based on light microscopic examination as shown in Fig. 3D & F. The images shown are representative of tissue sections from the three eyes derived from three mice per group.

IOP changes in GDF15 null mice

In contrast to the elevated IOP observed in human GDF15 expressing transgenic mice, the IOP in 10-month-old GDF15 null mice (14.47 mmHg, median, n=16) was not different from that of age-matched control (WT) mice (15.1 mmHg, median, n=20) (Fig. 4A). Table S2 shows additional details of IOP values and analyses for the GDF15 null and control mice. As documented for the transgenic mice, the histology of the conventional outflow pathway based on the light microscopy evaluation was not different between GDF15 null and wild type mouse eyes (Fig. 4B). Additionally, the body weight (grams/mouse) of the GDF15 null mice (31.71 g; median of n=8) was comparable to that of age-matched wild type control mice (30.30g; median of n=10) (Fig. S3B).

Figure 4. IOP changes in GDF15 null mice.

A). To determine the effects of absence of GDF15 expression on IOP, IOP was evaluated in GDF15 null mice (10-month-old) using a rebound tonometer. GDF15 null mice (KO; n=16 eyes) showed no change in IOP compared to the age-matched wild type mice (n=20 eyes). B). Light microscopy based histological examination showed no recordable differences in the conventional AH pathway, and in ciliary body (CB)) between GDF15 null and wild-type mice. TM; trabecular meshwork, SC; Schlemm’s canal, KO; knockout, WT; wild type, ns; not significant. Bar indicates image magnification.

Decreased AH outflow facility in mouse eyes perfused with recombinant human GDF15

Mice expressing transgenic human GDF15 showed an elevation in IOP, as illustrated in Figure 3. Moreover, elevated levels of GDF15 have been observed in the aqueous humor (AH) of glaucoma patients.^10,11 Notably, recombinant GDF15 induced contractile activity and the expression of ECM proteins in TM cells.⁹ These findings prompted our exploration into whether GDF15 affects the outflow facility of the AH through the trabecular pathway. To test this possibility, we evaluated changes in AH outflow facility (nl/min/mmHg) in enucleated eyes (from 8- to 9-month-old, male and female wild type mice) perfused with recombinant human GDF15 (4 ng/ml). Contralateral eyes from the same mouse were used for the test (GDF15) and control (vehicle) treatments as described in the Methods section. As shown in Fig. 5A, AH outflow facility was found to be significantly decreased (by ~30%, P<0.013; median 3.35; n=12) in GDF15 perfused eyes compared to vehicle perfused control eyes (median 4.65, n=12). Table S3, provides additional details for the rh GDF15 induced AH outflow facility changes. The AH outflow facility changes recorded in GDF15 perfused eyes were not associated with noticeable differences in histology of the conventional outflow pathway (relative to vehicle perfused eyes) based on light microscopy evaluation as shown in Fig. 5B.

Figure 5. Decreased aqueous humor outflow facility in recombinant human GDF15 perfused enucleated mouse eyes.

A). To test whether GDF15 influences aqueous humor (AH) outflow facility, enucleated eyes derived from the 8 to 9 month-old (male and female) mice were perfused with rh GDF15 (4 ng/ml) and determined AH outflow facility in comparison with vehicle perfused control eyes (contralateral) as described in the Methods section. The AH outflow facility (nl/min/mmHg) in rh GDF15 perfused eyes (n=12) was found to be significantly (*P<0.01; by ~30%) decreased compared to vehicle perfused contralateral eyes (n=12). B). Light microscope-based histological examination of the conventional AH outflow pathway showed a similar morphology integrity between the GDF15 perfused and vehicle perfused eyes. TM; trabecular meshwork, SC; Schlemm’s canal. Bar indicates image magnification.

DISCUSSION

The main goal for this study was to gain insights into whether altered expression of GDF15, whose levels are reported to be elevated in the AH of glaucoma patients, influences IOP. Elevated IOP is a major risk factor for glaucoma.⁶ To address this aspect, herein we have used two well characterized GDF15 gene targeted mouse models including a human GDF15 expressing transgenic mouse and a GDF15 null mouse, to evaluate changes in IOP in comparison to the respective WT mice. The findings of this study reveal that while mice expressing human GDF15 show elevated IOP, the GDF15 null mice appear to maintain normal IOP. Additionally, perfusion of enucleated eyes with recombinant human GDF15 revealed a decrease in AH outflow facility. Furthermore, treatment of human TM cells with recombinant GDF15 resulted in increased expression of genes encoding various ECM proteins. Taken together, these results indicate that increased levels of GDF15 potentially affect the homeostasis of IOP and AH outflow, and lead to IOP elevation in mice. Whether the observed modest increase in IOP in GDF15 transgenic mice was partly caused by increased levels of ECM by GDF15 in the tissues of the conventional (trabecular) AH outflow pathway requires additional mechanistic studies.

Glaucoma is one of leading causes of blindness globally and elevated levels of GDF15 in the AH of primary open-angle glaucoma and Pseudoexfoliative glaucoma patients have been well documented.^10,11,13 Moreover, IOP fluctuation has also been reported to be associated with changes in the levels of GDF15 in glaucoma patients.³⁴ In our previous study, we reported elevated levels of GDF15 in human TM cells treated with various physiological agents known to influence IOP including transforming growth factor-β2 (TGF-β2), LPA (lysophosphatidic acid), glucocorticoid (dexamethasone) and endothelin-1.^7,9,30,35,36 Additionally, GDF15 has been reported to induce the TM cell contractile response and modulate cell adhesive interactions,⁹ cellular attributes known to impact AH outflow and IOP.⁷ Consistent with the well-recognized characteristics of GDF15 as a stress inducible protein, senescence, oxidation, impaired autophagy and ER stress all increased levels of GDF15 in TM cells. However, whether GDF15 directly has any influence on IOP has not been previously addressed.

While the role of GDF15 in energy metabolism, body weight and food intake is well recognized and mechanistically understood to be mediated through GFRAL/Ret kinase regulated signaling,^14,18,19,24 GDF15 has also been shown to influence various pathological changes including metastasis, fibrosis, inflammation, and senescence.^{14,15,37–39} In this study, treatment of TM cells with rh GDF15 induced gene expression predominantly of various ECM and cell adhesion proteins and the inflammatory chemokine CCL2. Consistent with this, our previous study using cDNA microarray-based analysis also recorded increased levels of certain ECM proteins and other proteins.⁹ However, caution may be warranted here regarding findings derived from the use of recombinant GDF15, since concerns of possible contamination with TGF-β during the purification of recombinant GDF15 preparations have been raised in connection with the discovery of GFRAL.¹⁴

With respect to IOP changes recorded in human GDF15 expressing transgenic mice, while we believe that these changes are directly related to changes in the levels of GDF15, the data also raise a question regarding the mechanism(s) underlying this observation. Although, as stated above, we cannot completely eliminate the possible contamination of rhGDF15 obtained from commercial source with TGF-β in our studies, perfusion of enucleated mouse eyes with rhGDF15 led to decreases in AH outflow facility. This finding together with rhGDF15 induced gene expression of ECM proteins including fibronectin, various collagens and laminins recorded in TM cells could be partly responsible for the elevation of IOP in GDF15 transgenic mice. Interestingly, increased expression of and perfusion with fibronectin and other ECM proteins has been shown to elevate IOP in various experimental models.^8,40,41 Thus, it is possible that the increased levels of ECM proteins and their interactions with integrins and cell surface proteoglycans can potentially influence the actin cytoskeletal organization and cell adhesion, attributes known to influence AH outflow and IOP.^7,42,43 In support of this possibility, our previous study documents rhGDF15 - induced increases in actin stress fibers and myosin II activity in human TM cells.⁹ Additionally, in some studies, GDF15 has been reported to induce fibrosis in different tissues.^37,44,45 Current literature on GDF15 induced biological or pathological changes which associate with and independent of the participation of GFRAL/Ret signaling is somewhat inconsistent and context dependent. Identification of non-canonical signaling pathway(s) which might mediate the effects of GDF15 in cell and tissue types that do not express GFRAL is required to better understand the role of GDF15 in normal TM physiology and the pathobiology of different diseases including impaired AH outflow and ocular hypertension in glaucoma.

In conclusion, this study reports a modest elevation in IOP in transgenic mice which express human GDF15 and exhibit increased levels of GDF15 in the AH. In light of the consistently elevated levels of GDF15 reported in the AH of glaucoma patients, the in vivo and in vitro experimental findings described in this study suggest that GDF15 has an influence on IOP homeostasis and potentially plays a role in the pathobiology of ocular hypertension.

Data Availability:

All raw RNA-seq data and normalized expression values are available in the Gene Expression Omnibus with accession number- PRJNA1055941. All other data from this study are included in the article and supplementary information files. Additionally, data supporting the findings of this study can be made available by the corresponding author upon request.