Macro- and Microscale Properties of the Vitreous Humor to Inform Substitute Design and Intravitreal Biotransport

ABSTRACT

Research on the vitreous humor and development of hydrogel vitreous substitutes have gained a rapid increase in interest within the past two decades. However, the properties of the vitreous humor and vitreous substitutes have yet to be consolidated. In this paper, the mechanical properties of the vitreous humor and hydrogel vitreous substitutes were systematically reviewed. The number of publications on the vitreous humor and vitreous substitutes over the years, as well as their respective testing conditions and testing techniques were analyzed. The mechanical properties of the human vitreous were found to be most similar to the vitreous of pigs and rabbits. The storage and loss moduli of the hydrogel vitreous substitutes developed were found to be orders of magnitude higher in comparison to the native human vitreous. However, the reported modulus for human vitreous, which was most commonly tested in vitro, has been hypothesized to be different in vivo. Future studies should focus on testing the mechanical properties of the vitreous in situ or in vivo. In addition to its mechanical properties, the vitreous humor has other biotransport mechanisms and biochemical functions that establish a redox balance and maintain an oxygen gradient inside the vitreous chamber to protect intraocular tissues from oxidative damage. Biomimetic hydrogel vitreous substitutes have the potential to provide ophthalmologists with additional avenues for treating and controlling vitreoretinal diseases while preventing complications after vitrectomy. Due to the proximity and interconnectedness of the vitreous humor to other ocular tissues, particularly the lens and the retina, more interest has been placed on understanding the properties of the vitreous humor in recent years. A better understanding of the properties of the vitreous humor will aid in improving the design of biomimetic vitreous substitutes and enhancing intravitreal biotransport.

KEYWORDS:

• Vitreous humor
• vitreous substitute
• mechanical properties
• hydrogel
• viscoelasticity
• modulus
• biotransport

Introduction to the vitreous humor

The vitreous humor is a gel-like, soft, and transparent ocular tissue located between the lens and the retina, occupying 80% of the eye’s volume.¹ It is composed of 98–99% water and a framework of collagen fibers and hyaluronic acid. The hyaluronan coils intersperse and swell in the network of type II and type IX collagen, creating a hypothetical “internal tension” by Donnan swelling.^2,3 Due to its soft and viscoelastic nature, the vitreous humor serves as a mechanical damper for the eye, holding the lens and retina in place and protecting these tissues from physical insults. The vitreous also contributes to the growth processes of the ocular tissues during development.^4,5 Another vital role of the vitreous is to establish and maintain an oxygen gradient between the lens and retina, where a high concentration of oxygen is found near the metabolically active retinal-pigmented epithelial cells and a low concentration of oxygen is found near the oxygen-sensitive lens epithelial cells.^6–9 This biochemical function of the vitreous is accomplished by both a high concentration of the antioxidant ascorbic acid (also known as vitamin C), which consumes oxygen, as well as the gel-like nature of the vitreous humor, which limits transport of oxygen by convection.^10,11 The properties of the vitreous are important not only for physical protection from mechanical insults but also for biochemical protection from oxidative damage for other tissues in the eye. Thus, a better understanding of the properties of the vitreous humor may lead to better appreciation of ocular development and disease processes, improved understanding of pharmacokinetics and biotransport of drugs in the vitreous, and enhancement of the design of vitreous substitutes. This review summarizes the material properties of the vitreous humor in comparison to hydrogel vitreous substitutes, focusing particularly on the mechanical properties including storage modulus (G’) and loss modulus (G”), which describe the solid-like/elastic and liquid-like/viscous qualities of a material, respectively. The review concludes by highlighting challenges associated with current testing techniques and suggesting ways to improve vitreous substitutes and better understand intravitreal biotransport.

Studies on the mechanical properties of the vitreous humor

The earliest studies found on the mechanical properties of the vitreous humor were published in the 1970s; however, this topic garnered little attention, generating 10 publications within a 30-year span from 1975 to 2005 (Table 1 and Figure 1A). Since the publication of works by Nickerson et al,^3,38 there has been a rise in interest in the mechanical properties of the vitreous humor over the last 15 years (2005–2020), with 27 publications. The mechanical properties of the vitreous humor have commonly been studied using human (n = 13 studies), pig (n = 17 studies), or cow (n = 14 studies) eyes (Figure 1B). Sheep (n = 2 studies), goat (n = 1 study), and rabbits (n = 3 studies) have also been included. The experiments were predominately performed in vitro (n = 31 studies), followed by in situ/ex vivo (n = 15 studies), and finally in vivo (n = 6 studies) (Figure 1C). Rheology was the technique used in most of these studies (n = 28 studies), with shear rheology predominately used (n = 12 studies) (Figure 1D). Other techniques used include magnetic resonance imaging (MRI), ultrasound, and light scattering (n = 10 studies).

Table 1. Summary of publications on the mechanical studies of the vitreous humor. The species, condition, and testing technique are listed for each publication

Author(s)	Year	Species	Condition	Technique	Reference
Thakur et al.	2020	Pig	In situ	Magnetic resonance imaging	12
Rangchian et al.	2020	Pig	In situ	Creep test	13
Rangchian et al.	2019	Pig	In situ	Creep test	14
Schulz et al.	2019	Cow, Human, Pig	In vitro	Shear rheometry	15
Shafaie et al.	2018	Cow, Human, Sheep, Pig	In vitro	Shear rheometry	16
Tram et al.	2018	Human, Pig	In vitro	Creep test and shear rheometry	17
Huang et al.	2018	Cow	In vitro	Shear rheometry	18
Stein et al.	2018	Human	In situ	Magnetic resonance imaging	19
Silva et al.	2017	Rabbit	In vitro	Shear rheometry	20
Shah et al.	2016	Cow	In vitro	Creep test	21
Pokki et al.	2015	Human, Pig	In situ	Microrheometry	22
Colter et al.	2015	Sheep	In vitro	Shear rheometry	23
Abdelkawi et al.	2014	Rabbit	In vitro	Viscometry	24
Filas et al.	2014	Cow, Pig	In vitro	Shear rheometry	25
Watts et al.	2014	Rabbit	In vitro	Microrheometry	26
Zhang et al	2014	Cow	In vitro	Shear rheometry	27
Yoon et al.	2013	Pig	In situ	Acoustic radiation force	28
Rossi et al.	2012	Human	In vivo	Ultrasound	29
Piccirelli et al.	2012	Human	In vivo	Magnetic resonance imaging	30
Sharif-Kashani et al.	2011	Pig	In vitro	Shear rheometry	31
Zimberlin et al.	2010	Cow	In vitro	Cavitation rheometry	32
Swindle-Reilly et al.	2009	Pig	In vitro	Capillary rheometry	33
Swindle et al.	2008	Pig	In vitro	Capillary rheometry	34
Nickerson et al.	2008	Cow, Pig	In vitro	Shear rheometry	3
Sebag et al.	2007	Pig	In situ	Dynamic light scattering	35
Patz et al.	2007	Human	In situ	Magnetic resonance imaging	36
Suri and Banerjee	2006	Goat	In vitro	Shear rheometry	37
Nickerson et al.	2005	Cow, Pig	In vitro	Shear rheometry	38
Walton et al.	2002	Human	In vivo	Ultrasound	39
Lee et al.	1994	Cow, Pig	In vitro	Microrheometry	40
Lee et al.	1992	Human	In vitro	Microrheometry	41
Aguayo et al.	1985	Cow	In situ	Magnetic resonance imaging	42
Tokita et al.	1984	Cow	In vitro	Torsion pendulum	43
Weber et al.	1982	Cow, Human	In situ	Periodic oscillations	44
Zimmerman	1980	Human	In vivo	Light scattering	45
Bettelheim and Wang	1976	Cow	In situ	Compression chucks	46
Oksala	1975	Human	In vivo	Ultrasound	47

Figure 1. There has been an increase in publications on the mechanical properties of the vitreous humor over the last 15 years (2005–2020) (A). Human, pig, and cow are the most common species used (B). In vitro testing of the vitreous humor remains the most common testing condition (C). Rheometry is a common technique used to characterize the mechanical properties of the vitreous humor, with shear rheology being the most frequently used technique (D)

Shear rheology, commonly used to characterize the viscoelasticity of hydrogels and other non-Newtonian fluids, has become a popular technique for investigating the viscoelasticity of the vitreous humor. This technique can report various rheological properties of the vitreous humor including storage and loss modulus (G’ and G”, respectively). Nickerson et al. (2005 and 2008) were the first to use shear rheology to determine the rheological properties of the vitreous humor.^3,38 Their works arguably provided a practical methodology for later investigators to study the mechanical properties of the vitreous, particularly with the introduction of the cleated geometry to overcome slippage effects that would otherwise underestimate the mechanical properties of the vitreous in shear testing. Before Nickerson et al. 2005, common techniques used were ultrasound,³⁹ light scattering,^35,45 periodic oscillations,^43,44 compression chucks,⁴⁶ MRI,⁴² or microrheology.^20,40 Following the studies by Nickerson et al.,^3,38 there have been at least 10 studies that used shear rheology. Some other rheometric techniques used were capillary rheology,^33,34 cavitation rheology,³² microrheology,²² and creep testing.^14,21

Material properties of the vitreous can change based on the testing temperature and are also rate and strain dependent.^{3,17,31,34,43} Unfortunately, the variety of methodologies, shear frequencies, and amplitudes used in previous studies make direct comparisons difficult. Even within a single study, the reported modulus can range over a couple orders of magnitude. Additionally, some studies did not report a range of modulus but instead reported a single value for the storage and loss moduli. We attempted to consider both the testing conditions and the modulus range reported, and therefore simplified the information by reporting the average value for the storage modulus, loss modulus, and elastic modulus. If the average modulus values were not published, the average of the highest and lowest value reported was calculated. If a paper reported both a range of the values and the mean, we used the reported mean calculated by the authors. The results were presented as the average of the means ± standard error, with n signifying the number of papers that reported modulus values.

There are significant variations in the mechanical properties of the vitreous humor from different species (human, pig, cow, rabbit, and goat) (Figure 2A – note the log scale). The storage, loss, and elastic modulus of human (G’ = 78.8 ± 73.7 Pa, n = 4; G” = 26.3 ± 21.8 Pa, n = 4; and E = 1.17 ± 0.56 Pa, n = 3) were much different than those of pig (G’ = 3.57 ± 0.96 Pa, n = 10; G” = 1.04 ± 0.35 Pa, n = 10; and E = 31.2 ± 12.2 Pa, n = 4), rabbit (G’ = 2.76 ± 1.87 Pa, n = 3; G” = 0.73 ± 0.46 Pa, n = 3; E not reported), goat (G’ = 1000 Pa, n = 1; G” = 400 Pa, n = 1; E not reported), sheep (G’ = 47.1 ± 42.9 Pa, n = 2; G” = 46.4 ± 44.1 Pa, n = 2; E not reported), and cow (G’ = 72.7 ± 54.2 Pa, n = 12; G” = 4.44 ± 1.79 Pa, n = 9; and E = 1.95 Pa, n = 1) vitreous. Piccirelli et al. (2012), however, reported human storage and loss moduli both at 300 ± 100 Pa,³⁰ which were significantly larger than values reported in the other studies (ranging from 1 to 10 Pa). Thus, these data points were removed from this analysis after using Grubb’s outlier test (p < .001). The removal of these data points resulted in the human moduli (G’ = 5.09 ± 1.86 Pa, n = 3 and G” = 1.81 ± 0.98 Pa, n = 3) having a similar storage and loss modulus to pig and rabbit samples (Table 2 and Figure 2B – note the linear scale). It should be noted that the human storage and loss moduli reported by Piccirelli et al (2012) were not directly measured but calculated using MRI measurements of in vivo human vitreous.³⁰ The assumptions used in their analytical model (homogeneous and spherical vitreous), limited image resolution, and patient compliance might explain the high standard deviation (300 ± 100 Pa) and drastic difference from the moduli directly measured using shear rheology on in vitro dissected vitreous samples reported by other studies.^15–17 However, to the best of the authors’ knowledge, Piccirelli et al (2012) is the first study that reported the in vivo storage and loss moduli of the human vitreous humor. It is likely that the measured moduli from in vitro and ex vivo experiments are slightly lower than in vivo due to tissue degradation and destruction of the tissue architecture during testing.

Table 2. Summary of the reported storage, loss, and elastic moduli of the vitreous humor by species

Species	Storage Modulus	Loss Modulus	Elastic Modulus
Human	5.09 ± 1.86 Pa, n = 3	1.81 ± 0.98 Pa, n = 3	1.17 ± 0.56 Pa, n = 3
Pig	3.57 ± 0.96 Pa, n = 10	1.04 ± 0.35 Pa, n = 10	31.2 ± 12.2 Pa, n = 4
Rabbit	2.76 ± 1.87 Pa, n = 3	0.73 ± 0.46 Pa, n = 3	Not reported
Goat	1000 Pa, n = 1	400 Pa, n = 1	Not reported
Sheep	47.1 ± 42.9 Pa, n = 2	46.4 ± 44.1 Pa, n = 2	Not reported
Cow	72.7 ± 54.2 Pa, n = 12	4.44 ± 1.79 Pa, n = 9	1.95 Pa, n = 1

Figure 2. There are significant differences in the modulus between species (A), with human, pig, and rabbit samples having the most similar storage and loss modulus to each other (B). Note the log scale in (A) and the linear scale in (B). The wide variation in the modulus between species results in large variations in the reported modulus using shear rheology and other rheological techniques (C). Combining data for human, pig, and rabbit (excluding cow, sheep, and goat) resulted in reported moduli with a much smaller range of variation (D). Again, note the differences in scale between (C and D). The storage and loss moduli of human, pig, and rabbit vitreous are similar to each other, unlike those of cows, goats, and sheep (E and the zoomed-in plot F). *Human data analyzed includes the results reported by Piccirelli et al. 2012

The large differences in species resulted in a wide range of moduli reported using shear rheology (G’ = 51.5 ± 41.4 Pa, n = 24 and G” = 23.7 ± 17.5 Pa, n = 23) and other rheometric techniques (G’ = 112.9 ± 92.7 Pa, n = 7 and G” = 1.07 ± 0.50 Pa, n = 5) (Figure 2C). Excluding the data from sheep, goat, and cow studies, the reported moduli for human, rabbit, and pig studies using shear and other rheological techniques were much more uniform compared to data from all species (G’ = 4.16 ± 0.84 Pa, n = 13 and G” = 1.23 ± 0.34 Pa, n = 13 compared to G’ = 1.75 ± 1.23 Pa, n = 3 and G” = 0.68 ± 0.48 Pa, n = 3, respectively) (Table 3 and Figure 2D – note the difference in scale compared to Figure 2C). The studies that used shear rheology generally used similar testing conditions (strain 1–100%, frequency 0.1–100 Hz, temperature 20–37°C), which might explain the similarity in storage and loss moduli reported in these studies^3,17,20,31 (Figure 2C). The studies that used other rheology techniques are more varied (Figure 2D). The scatter plot of different species (Figure 2E) and the zoomed-in plot (Figure 2F) highlight the similarities in both storage and loss modulus between vitreous samples from humans, pigs, and rabbits. It should be noted that studies utilizing other techniques (ultrasound, light scattering, etc.) have reported other meaningful mechanical properties, but have not directly measured or reported storage and loss modulus, making comparison challenging. The data were obtained in vivo as opposed to data obtained in vitro using rheology (which could also be rate dependent as aforementioned) and, therefore, could not be directly compared to each other.

Table 3. Summary of the reported storage and loss moduli of the vitreous humor based on testing technique (excluding cow, sheep, and goat)

Testing Technique	Storage Modulus	Loss Modulus
Shear Rheometry	4.16 ± 0.84 Pa, n = 13	1.23 ± 0.34 Pa, n = 13
Other Rheometry Techniques	1.75 ± 1.23 Pa, n = 3	0.68 ± 0.48 Pa, n = 3

Aging of the vitreous humor

With increasing age, the homogeneous vitreous humor phase separates, becoming a heterogeneous mixture of a stiffened phase composed of aggregated collagen fibrils and a loosened phase composed of dissociated hyaluronans.^3,17 This occurs due to a variety of factors such as oxidative damage, digestion by enzymes, and collagen mutations.^48,49 The degradation of collagen type IX has been particularly shown to be one of the causes of vitreous degradation.² Collagen type IX normally coats the outer surface of collagen type II in the vitreous, preventing the collagen fibrils from adhering to each other and preventing collapse of the collagen network. Collagen type IX has a half-life of 11 years, with an exponential decrease in concentration in human vitreous starting at a young age (~ 5 years old).⁵⁰ With increasing age, more collagen type IX is lost on the surface of the collagen fibrils of the vitreous, leading to an increased probability that the fibrils aggregate. This aggregation of the collagen fibrils causes the collapse of the homogeneous vitreous, resulting in the creation of the solid and liquid vitreous phases due to the increase in the degree of crosslinking of the collagen network and the expulsion of the hyaluronic acid from inside the collagen-hyaluronic acid network, respectively.^3,17

The phase separation of the vitreous compromises its ability to function as a protective structure for the surrounding ocular tissues and causes complications including rhegmatogenous retinal detachment, macular holes, vitreous hemorrhage, and vitreous floaters.⁵¹ Vitreous floaters, which can be seen as floating shadows in one’s field of vision, are large fibrous aggregations floating in pockets of liquefied vitreous that cast shadows on the retina, interfering with vision. When the liquid pockets develop at the back of the eye, areas of high stress develop where the vitreous connects to the retina, potentially causing retinal tears or macular holes. Thereafter, the liquid component of the vitreous might leak underneath the retina layer through the tear in the retina, lifting the retina away from the choroid and causing retinal detachment. Blood from the retinal vessels may bleed into the vitreous, creating vitreous hemorrhage which might necessitate surgical removal of the tissue if the vitreous hemorrhage does not clear on its own. The degree of phase separation of the vitreous humor has been correlated to signs of cataract formation⁵² due to the diminished oxygen gradient and the ability of oxygen to travel from the retina to the lens via convectional mixing. Considering the serious complications with the phase-separated vitreous humor, it is important to understand the changing mechanical properties of the aging vitreous.

While the vitreous humor is known to degrade with age, much less is known about its age-related changes in the mechanical properties of the vitreous humor, mostly due to the difficulty in obtaining human donor tissue at younger ages. Zimmerman (1980) reported the in vivo modulus of the vitreous from patients aged 18, 26, 28, 38, 47, and 50 years old and found no discernible differences due to age, potentially due to the unreliable measurements primarily based on patient compliance.⁴⁵ Weber et al. (1982) calculated the spring constant and damping factor of human vitreous from donors aged 44 to 73 years but did not report any significant age-related changes.⁴⁴ Shafaie et al. (2018) was the first paper that used shear rheology to measure the modulus of the human vitreous of multiple patients aged 48, 61, 70, 71, and 92, but also did not report any age-related changes in the rheological properties of the vitreous.¹⁶ Tram and Swindle-Reilly (2018) was the first study that reported the age-related rheological changes of the human vitreous.¹⁷ By testing both the solid and liquid phase of the human vitreous humor with shear rheology, it was determined that the vitreous gel became stiffer while the vitreous liquid became less elastic with increasing age.

There have been two other studies to date that have reported age-related changes in the rheological properties of the vitreous. Colter et al. (2015) investigated age-related changes in the mechanical properties of the vitreous humor using sheep eyes.²³ By testing whole ovine vitreous with the hyaloid membrane, they found that the dynamic moduli of the adult vitreous was about 2 times lower than infant vitreous, albeit not statistically significant. More recently, Shultz et al. (2019) analyzed 190 whole human vitreous body samples extracted from donors aged 33 to 92 via shear rheology. The authors found the storage and loss moduli both significantly decreased with age. While these observations appeared contradictory to the findings of Tram and Swindle-Reilly (2018), it should be noted that the entire vitreous was tested in these studies. The hallmark of vitreous liquefaction is phase separation of the vitreous humor, resulting in pockets of liquid inside the vitreous. The older vitreous samples in these studies likely had those pockets of liquid which, when tested as a whole with the vitreous gel, might have lowered the dynamic moduli compared to those from the younger vitreous samples, which would have less liquid phase in the vitreous. In other words, the human vitreous becomes macroscopically softer, likely due to the increase of liquid volume with age. When the solid and liquid phases of the vitreous humor were tested separately, the heterogeneity of the aging vitreous was captured.¹⁷

The strive for improved vitreous substitutes and intraocular therapeutic delivery

As the vitreous humor is incapable of regeneration or transplantation, it must be removed and replaced with a substitute during vitrectomy. In this surgical procedure, the vitreous humor is homogenized using a rotating blade and removed via suction. The empty vitreous chamber is then backfilled with a vitreous substitute. Commonly used vitreous substitutes include silicone oil, saline solutions, perfluorocarbon liquids, and gases. These materials serve as temporary tamponades that provide tension on the ocular tissues, keeping the retina in place for healing. It should be noted that these materials are fluids and therefore provide less protection against retinal detachments than the natural viscoelastic vitreous.⁵³ The gold standard for a long-term vitreous substitute is silicone oil, despite its many known complications and drawbacks.⁵⁴ For example, since silicone oil floats on top of aqueous humor in the eye, patients receiving silicone oil as a vitreous substitute typically have to lay face down for days to weeks to position the silicone oil layer in proper contact with the damaged retina. This reduces patients’ quality of life and compliance, resulting in suboptimal retinal reattachment rates (60%-70%).⁵⁵ Additionally, silicone oil’s hydrophobic nature renders it more permeable to oxygen than the natural vitreous humor. The diffusivity of oxygen in silicone oil is 8 × 10⁻⁹ m²/s,⁵⁶ which is double its diffusivity in water (4 x 10⁻⁹ m²/s).¹⁰ This higher diffusivity allows for increased oxygen mixing in the vitreous chamber and the flattening the oxygen gradient,^10,11 exposing the sensitive lens to oxidative damage,^6,7 and potentially causing a high incidence of cataract formation after vitrectomy.⁸ Furthermore, the hydrophobicity of silicone oil can cause it to emulsify with the aqueous humor or get trapped in the trabecular meshwork, causing cytotoxicity and potentially inducing glaucoma.^55,57 There is a demonstrated clinical need for a new generation of vitreous substitutes that behave like the natural vitreous humor and reduce the side effects associated with the current vitrectomy procedure and use of silicone oil.

Due to its similarities to the natural vitreous, hydrogels have garnered much attention as potential vitreous substitutes. Since hydrogels are hydrophilic, they eliminate the emulsification problem seen with silicone oil vitreous substitutes. The rise in popularity of hydrogel vitreous substitutes is similar to the rise in published research of the mechanical properties of the vitreous humor (Figure 3A). A big discrepancy in the number of publications on hydrogel vitreous substitutes and vitreous humor in the years 1995–2000 (9 publications and 0 publications, respectively) should be noted, considering the approval of silicone oil for use as a vitreous substitute by the Food and Drug Administration (FDA) in 1994.⁵⁸ Synthetic^{33,34,59–91} and semisynthetic^92–108 hydrogels have been more common materials for experimental vitreous substitutes in recent years, while natural hydrogels^37,109–123 were much more prevalent in the early years (Figure 3B). The foldable capsular vitreous body (FCVB) is a different polymeric design for a vitreous substitute that is currently undergoing clinical trials in China.^124–143 The FCVB comprises a polymeric bag that is surgically inserted into the vitreous body and subsequently inflated with the use of filling materials (saline or silicone oil). Overall, a synthetic hydrogel (n = 36) is the most commonly reported type of experimental vitreous substitute (Figure 3C), followed by FCVB (n = 20), semisynthetic hydrogel (n = 17), and natural hydrogel (n = 16) vitreous substitutes. Direct injection (n = 41) is the most common mechanism for instilling the vitreous substitute into the eye (Figure 3D), followed by FCVB (n = 20), in situ crosslinking (n = 16), thermogelling (n = 6), and thermogelling/in situ crosslinking (n = 5). The different types of hydrogel vitreous substitutes and mechanisms of injection are listed in Supplemental Table 1. More details about these vitreous substitutes can be found in several review articles.^53,144–154

Figure 3. Number of publications on the mechanical properties of the vitreous humor (from Figure 1a) compared to the number of publications on hydrogel vitreous substitutes (A). There has been an increase in the publications on hydrogel vitreous substitutes concurrent with publications on the mechanical properties of the vitreous humor. Natural hydrogels attracted the most attention in the early publications, while synthetic and semisynthetic hydrogels have become more prevalent in vitreous substitute research in recent years (B). Foldable capsular vitreous body (FCVB) is another prominent candidate for vitreous substitution. Synthetic hydrogels remain the most common type of material for experimental vitreous substitutes (C). Direct injection of the hydrogel is the most frequently reported mechanism for delivering the vitreous substitute into the eye (D)

Most current experimental hydrogel vitreous substitutes have moduli that are orders of magnitude higher than those of the human vitreous (Table 4 and Figure 4A). However, stiff hydrogels can be difficult to inject through a small gauge needle into the vitreous chamber. The storage and loss moduli of synthetic hydrogels^{34,59–62,69,73,74,76,81–83} (G’ = 1203 ± 755 Pa, n = 18 and G” = 373 ± 330 Pa, n = 12) and semisynthetic hydrogels^{92–94,97,98,100–102,104} (G’ = 1330 ± 1090 Pa, n = 9 and G” = 291 ± 223 Pa, n = 9) are approximately two orders of magnitude higher, while natural hydrogels’^37,110 (G’ = 99.4 ± 53.4 Pa, n = 3 and G” = 51.5 ± 48.5 Pa, n = 2) are about one order of magnitude higher compared to the human vitreous^15–17 (G’ = 5.09 ± 1.86 Pa, n = 3 and G” = 1.81 ± 0.98 Pa, n = 3). Hydrogels with different mechanisms of injection into the eye also have a wide variation in storage and loss moduli (Table 5 and Figure 4B). Hydrogel vitreous substitutes with direct injection mechanism^{59,62,73,76,81–83,94,110} (G’ = 49.9 ± 31.3 Pa, n = 9 and G” = 26.3 ± 21.8 Pa, n = 9) have the closest modulus to the human vitreous, followed by thermogelling/in situ crosslinking hydrogels^{37,96,100,101} (G’ = 136 ± 33.1 Pa, n = 4 and G” = 28.9 ± 23.8 Pa, n = 4). The moduli of hydrogels with in situ crosslinking^{34,69,93,97,98,104} (G’ = 1626 ± 1031 Pa, n = 13 and G” = 338 ± 332 Pa, n = 6) or thermogelling^60,61,74,102 (G’ = 2946 ± 2360 Pa, n = 4 and G” = 1206 ± 937 Pa, n = 4) mechanisms are, on average, more than two orders of magnitude higher than those of the natural human vitreous. Scatter plots of different hydrogel types (Figure 4C) and injection mechanisms (Figure 4D) highlight the widespread differences in both storage and loss modulus for current experimental hydrogel vitreous substitutes.

Table 4. Summary of the reported storage and loss moduli of the hydrogel vitreous substitutes based on hydrogel type

Hydrogel Type	Storage Modulus	Loss Modulus
Human vitreous humor	5.09 ± 1.86 Pa, n = 3	1.81 ± 0.98 Pa, n = 3
Synthetic	1203 ± 755 Pa, n = 18	373 ± 330 Pa, n = 12
Semisynthetic	1330 ± 1090 Pa, n = 9	291 ± 223 Pa, n = 9
Natural	99.4 ± 53.4 Pa, n = 3	51.5 ± 48.5 Pa, n = 2

Table 5. Summary of the reported storage and loss moduli of the hydrogel vitreous substitutes based on delivery mechanism

Instilling Mechanism	Storage Modulus	Loss Modulus
Human vitreous humor	5.09 ± 1.86 Pa, n = 3	1.81 ± 0.98 Pa, n = 3
Direct injection	49.9 ± 31.3 Pa, n = 9	26.3 ± 21.8 Pa, n = 9
In situ crosslinking	1626 ± 1031 Pa, n = 13	338 ± 332 Pa, n = 6
Thermogelling	2946 ± 2360 Pa, n = 4	1206 ± 937 Pa, n = 4
Thermogelling/in situ crosslinking	136 ± 33.1 Pa, n = 4	28.9 ± 23.8 Pa, n = 4

Figure 4. Hydrogel vitreous substitutes have higher modulus compared to the human vitreous (A), regardless of what type of hydrogel was used. Natural hydrogels are, on average, softer than synthetic and semisynthetic hydrogels, but are still much stiffer than the human vitreous. Different mechanisms for instilling the hydrogel vitreous substitute into the eye also have a wide range of moduli that are also larger than the moduli of the human vitreous. Hydrogels that employ direct injection and thermogelling/in situ crosslinking mechanisms have the most similar modulus compared to the human vitreous, as opposed to hydrogels employing in situ crosslinking or thermogelling mechanisms (B). The storage and loss modulus of different hydrogel types (C) and injection mechanisms (D) differed by orders of magnitude (note that the axes are in logarithmic scale)

Foldable Capsular Vitreous Body (FCVB) is another prominent candidate for vitreous substitutes. The FCVB, designed to treat severe retinal detachment, consists of a vitreous-shaped capsule with a tube-valve system made of silicone rubber that can be triple folded and implanted into the vitreous cavity of the eye. The mechanical properties of the FCVB depend on the filling materials (saline, silicone oil, or polyvinyl alcohol (PVA)). Since 2008, there have been at least 20 publications on the FCVB, ranging from original design and in vitro mechanical testing to animal testing and human clinical trials.^124–143 While radically different from the natural vitreous humor as well as other hydrogel vitreous substitute designs, the FCVB has been shown to be a safe and effective vitreous substitute in humans over a 1- and 3-year observation period. With clinical trials currently underway in China, the FCVB is arguably closest to routine clinical use in humans compared to other experimental hydrogel vitreous substitutes. However, due to its radical design and a more invasive implantation approach (a transscleral port is made for the tube-valve system postoperatively), it remains unknown whether the FCVB will be readily accepted by clinicians and patients.

Loading therapeutics into vitreous substitutes also has the potential to improve their utility. Notably, due to the presence of nano-scale apertures on its capsule surface, the FCVB has been used to sustain release of various ophthalmic therapeutics such as levofloxacin, 5-fluorouracil, siRNA-PKCα, and dexamethasone sodium phosphate.^{130–132,136,137,141} More recently, a PVA/chitosan hydrogel loaded with fluorouracil-containing poly(lactic-co-glycolic acid) microspheres was proposed as a new vitreous substitute for proliferative vitreoretinopathy.⁹³ Besides providing treatments for vitreoretinal diseases, it may be worthwhile to consider vitreous substitutes with the ability to prevent cataract formation post-vitrectomy. Hydrogel vitreous substitutes loaded with antioxidants such as ascorbic acid have recently been reported to reduce reactive oxygen species activity for lens cells, potentially preventing or reducing the incidence of cataract formation.⁵⁹ Glutathione, another antioxidant found at high concentrations in the ocular lens,^155–157 has recently been used in conjunction with ascorbate in a hydrogel vitreous substitute and was shown to curtail the potential cytotoxicity of ascorbate, prolong its antioxidant activity, and improve its stability in an aqueous environment.¹⁵⁸ These developments provide a novel avenue for ophthalmologists to better treat and control vitreoretinal and other ocular diseases while reducing the frequency of therapeutic administration for patients.

The need for in vivo measurements of the vitreous humor

The vitreous connects to specific locations inside the eye in vivo, creating a complex mechanical response that depends not only on the material properties of the vitreous humor but also on the properties of the corneoscleral shell and other intraocular tissues, particularly the lens and retina. A common problem with shear rheology and other rheological techniques is that they are destructive, particularly for irreversibly shear thinning materials like the vitreous.³ The vitreous samples were usually dissected from the eye, disrupting and damaging the internal connection of the vitreous to the eye. Nickerson et al. and various other authors have pointed out that the properties of the vitreous rapidly change following dissection, going from a homogeneous vitreous to a heterogeneous gel surrounded by a pool of liquid.^3,17,20,38 This expulsion of liquid from the vitreous humor occurs rapidly, within 10 minutes of dissection. It is likely that the measured properties of the dissected in vitro vitreous samples are different from the properties of intact in vivo vitreous in the eye. Therefore, future studies should focus on the in vivo or in situ mechanical properties of the intact vitreous humor.

Recent attempts to measure the undissected vitreous utilize volume-controlled cavity expansion, cavitation rheology, and creep testing with a submerged rotating rheometric head or microprobes.^{13,14,22,26,32,159} Volume-controlled cavity expansion is a needle-based technique that employs a large deformation viscoelastic model to capture a material response measured after several cycles of expansion-relaxation at controlled stretch rates in a cavity expansion setting.¹⁵⁹ Similarly, cavitation rheology involves inserting a syringe needle into a vitreous sample and inducing an elastic instability via slow pressurization that correlates with the local mechanical properties of the sample.³² In situ rheological creep tests, in which a rotating rheometric head is submerged into the vitreous chamber, were recently used to measure the changes in the mechanical behavior of the vitreous after degradation using collagenase.^13,14 Minimally invasive rheological tests of the vitreous humor have also been attempted using microprobes controlled magnetically or optically.^22,26 While still relatively invasive (an opening has to be made in the sclera for the insertion of cavitating needle, rheometric head, silica or magnetic beads, etc.), these techniques ensure that the vitreous remains mostly intact, thereby preserving the internal structure of the vitreous humor inside the eye.

Ideally, the mechanical properties of the intact vitreous humor should be tested in vivo. Non-invasive techniques used to study the material properties of the vitreous include MRI,^{12,19,30,36,42} light scattering,^35,45 microbubble-based acoustic radiation force²⁸, ultrasound,^29,39,47 and Brillouin spectroscopy.¹⁶⁰ With the advantageous ability of non-destructive in vivo vitreous assessment, the disadvantage of these methods is that they do not provide data (storage modulus and loss modulus) that are comparable to the current rheological data on the vitreous humor or reported experimental hydrogel vitreous substitutes. For example, using Brillouin microscopy, the storage and loss moduli for the porcine vitreous humor were calculated to be 2.3 GPa and 0.7 GPa, which are six orders of magnitude higher than what have been reported in other papers using shear rheology.¹⁶¹

Alternatively, MRI was used to predict the storage and loss moduli of in vitro vitreous-mimicking gels and ex vivo porcine vitreous humor using transverse relaxation time (T2). Correlation between MRI-predicted values and the actual values measured using shear rheology still has room for improvement.¹² It should be noted that the T2 relaxation time measurements follow the local motion of water. Since the vitreous is composed of 99% water, T2 relaxation time has great potential to be utilized in future investigations to study the rates of diffusion in the vitreous. This can potentially have a broad impact on our understanding of intravitreal biotransport phenomena including drug delivery, oxygen gradient formation, and antioxidant sequestration. Studying the high-frequency harmonics or vibrational modes of the vitreous humor may also provide some information noninvasively, particularly on the upper bound of the elasticity of the vitreous without the influence of relaxation, deformation, or flow on longer timescales of milliseconds to minutes.^162,163 While these techniques are promising and should continue to be investigated in future studies, more work is needed to improve our understanding of the in vivo properties of the vitreous humor.

In vivo mechanical properties of the vitreous humor are useful for improved biomimetic vitreous substitutes. Currently, there is minimal agreement between the known rheological properties of the vitreous humor and the hydrogel vitreous substitutes being developed. Ideally, the material properties of the vitreous substitute should be measured using the same methods as the vitreous humor would be measured.^17,34,59 Only a handful of hydrogel vitreous substitutes (n = 13) have similar storage and loss moduli to those reported for the human vitreous humor (less than 20 Pa).^{34,59,62–64,69,73,76,81,84,110,128,129} In contrast, a majority of the previously proposed vitreous substitutes have modulus values that are orders of magnitude higher than those of the reported human vitreous humor. While the in vivo results for some of these hydrogel vitreous substitutes were promising for short- and intermediate-term applications (less than 1 or 2 years), it is unknown whether the high modulus of these stiff gels would have long-term adverse effects on ocular tissues. However, the modulus of the human vitreous has been hypothesized to be underestimated by current in vitro techniques,^3,38 so it may warrant investigation to develop vitreous substitutes with moduli slightly higher than the currently reported moduli for the vitreous humor.

Microscale factors to consider

This review primarily focuses on the macroscale properties of the vitreous humor and hydrogel vitreous substitutes. The microscale properties of the vitreous humor and substitutes are also important factors that need to be considered for the restoration of the natural oxygen gradient and establishment of redox homeostasis in the vitreous humor after vitrectomy. The viscosity, solute concentration, interaction between solute and diffusion medium, mesh size and tortuosity of the diffusion network all contribute to a solute’s diffusion through a medium.^164,165 Age-related changes in the vitreous structure adds complexity to the diffusion problem by influencing the hyaluronic acid and collagen network through vitreous liquefaction and fiber aggregation.¹⁶⁶ The solute’s surface charge has also been shown to affect the diffusion of a solute in the vitreous humor.¹⁶⁷ Shui et al (2009) showed that the ascorbate concentration and oxygen consumption decrease with increasing vitreous liquefaction,¹¹ and Filas et al (2013) showed that the ascorbate content and the structure of the vitreous gel are critical determinants of lens oxygen exposure.¹⁰ The phase separation of the vitreous into collagen-rich and hyaluronic-acid-rich vitreous phases¹⁷ further complicates the intravitreal biotransport and pharmacokinetics of therapeutics. The interactions between the vitreous phases (intact or liquefied) and the solutes (oxygen, antioxidants, or other therapeutics) remains largely unknown and warrant further investigation in the future.

A better understanding of the micro- and macrostructures of the vitreous humor can be used to inform designs of hydrogels with appropriate properties, both macroscopically and microscopically, that can better serve as a biomimetic vitreous substitute. Shafaie et al (2018) showed that, even with similar rheological properties between the human, porcine, and bovine vitreous, the steady-state flux of fluorescein (a model drug) through the human vitreous was significantly greater than through both porcine and bovine.¹⁶ It should be noted, however, that the human vitreous samples tested were from older adults (aged 48–92 years old) with presumably liquefied vitreous. The ages of bovine and porcine samples were not reported, but the tissues were by-products collected from abattoirs and derived from healthy animals with presumably intact vitreous humor. Therefore, although the macroscopic rheological properties of the human, porcine, and bovine vitreous were similar, the microscale properties of the liquefied human vitreous might be different from those of the intact bovine and porcine vitreous. This might explain why the rate of diffusion and flux in human vitreous was significantly higher than in bovine or porcine vitreous as reported. Future investigations should focus more on how mechanical properties, particularly microscale properties, relate to the mass transfer, antioxidant balance, and establishment of the oxygen gradient in the vitreous humor.

Finally, there has been increased interest in the biotransport of the vitreous humor and vitreous substitutes, particularly due to increased clinical use of intravitreal injection to deliver therapeutics.^{10,16,18,168–183} The motion of fluid inside the vitreous (intact or liquefied) due to eye rotation or saccadic eye movement has been investigated by many groups,^{168,169,171,180,183} with particular implications shown for drug delivery and pharmacokinetics in the vitreous.^{170,172,173,178,184} A model of enzymatic degradation of the vitreous was created and used to study the mobility of intravitreal nanoparticles, which could better simulate the biotransport of a therapeutic injected into aged and liquefied vitreous.¹⁸ An in vitro ocular flow model called the PK-Eye was proposed for use in preclinical drug development,¹⁷⁵ which can allow for testing of therapeutics in vitreous substitutes.^174,181,182 These advances will enable significant progress in determining pharmacokinetics and therapeutic feasibility at earlier stages. However, these advancements are still mostly based on the in vitro properties of the vitreous humor, which likely limits their usefulness and accuracy in in vivo applications. Understanding the in vivo properties of the vitreous will provide insights on the biotransport of intact or degraded vitreous humor and hydrogel vitreous substitutes, particularly on the intravitreal injection of therapeutics or incorporation of therapeutics inside a vitreous substitute.

Future directions

The degradation of the vitreous humor causes complications that could benefit from further investigations. The heterogenization of vitreous humor can disrupt the internal structure of the vitreous humor and unbalance the stress distribution in the eye. As the vitreous phase-separates, the retina experiences increased traction at locations with strong adhesion to the vitreous humor. Interestingly, the adhesion between the vitreous and the retina has been shown to decrease with age.¹⁸⁵ This reduction in vitreoretinal adhesion coincides with the timeline of vitreous liquefaction and has been hypothesized to facilitate posterior vitreous detachment, which might protect the retina from tearing or detaching. Induction of posterior vitreous detachment was recently reported, with particular focuses on both vitreous liquefaction and dehiscence of vitreoretinal adhesion.¹⁸⁶ A mechanical model of posterior vitreous detachment and generation of vitreoretinal tractions has also been published.¹⁸⁷ These are all interesting biomechanical problems that should be further investigated in future studies.

Additionally, the phase-separated vitreous humor alters the diffusion and transport of various nutrient and waste molecules, contributing to complications including cataract formation and proliferative vitreoretinopathy. For example, the gel quality of the vitreous humor has been shown to negatively correlate with the concentration of ascorbic acid in the vitreous,¹¹ and that a liquefied vitreous allows for more mixing of oxygen via convection,¹⁰ thereby exposing the lens to more oxidative damage and increased incidence of cataract formation.^8,52,188 Advanced glycation end-products (AGEs) formation, a reactive-oxygen-species-dependent process that increases the accumulation of collagen crosslinks, has been implicated in vision loss associated with cataract formation, diabetic retinopathy, and glaucoma.¹⁸⁹ AGEs has been shown to increase in the vitreous with age, especially in patients with diabetes¹⁹⁰ and/or rhegmatogenous retinal detachment.¹⁹¹,192 AGE accumulation has also been shown to reduce vitreous permeability,¹⁹² which might explain the disruption to the antioxidant balance and the oxygen gradient in liquefied vitreous. The connection between a healthy oxygen gradient, homogeneous vitreous, rate of AGE formation, and biotransport of molecules (ascorbate, glutathione, other therapeutics, etc.) provides an opportunity to tie together multiple threads in this review and hopefully brings together interdisciplinary teams of clinicians, engineers, and scientists to answer these multifaceted problems in future investigations.

Future studies can expand upon limitations described above. In this review, the reported moduli were simplified and do not reflect the diversity in the testing conditions. A closer investigation of the testing methods, strain, rate, and temperature might help identify the source of variability between the studies. This review also mostly focused on the macroscopic properties of the vitreous humor and hydrogel vitreous substitutes. However, microscale factors are also important to the structure and function of the vitreous and will need to be further investigated. A better understanding of the macro- and microscale properties and age-related changes of the human vitreous humor will aid in the design of biomimetic vitreous substitutes, enhancement in analyzing transport of therapeutics in the vitreous, and comprehension of the pathological conditions of the vitreous humor.

Conclusions

Studies on the vitreous humor and vitreous substitutes have increased in recent years. Samples from several species have been evaluated, with the mechanical properties of pig and rabbit vitreous found to be most similar to the native human vitreous, warranting the preferred use of pig and rabbit models for future studies. The vitreous has mostly been tested in vitro, despite known limitations relating to the rapid changes of the vitreous humor once removed from the eye. With increasing age, the vitreous phase-separates and the stiffness of the whole vitreous was found to decrease, while there was localized stiffening of the solid-like components of the aged vitreous. Future work should focus on determining the in vivo properties of the human vitreous humor and the biotransport of nutrients, waste, and injected therapeutics in the intact and degraded vitreous humor and hydrogel vitreous substitutes. The recent advancements in hydrogel vitreous substitutes are promising; however, more focus should be placed on ensuring the biomimicry of the substitutes, especially relating to the rheological properties. Finally, drug-eluting hydrogel vitreous substitutes are a promising new generation of vitreous substitutes that have the potential to shift the design paradigm of current vitreous substitutes to a more holistic strategy that considers not only the mechanical roles but also the biochemical functions of the natural vitreous humor.

Patent Applications

Reilly KE, Reilly MA, Tram NK. Antioxidant-Releasing Vitreous Substitutes and Uses Thereof. US Patent Application PCT/US2020/017525, Feb. 10, 2020.

Acknowledgments

We would like to acknowledge the past and present members of the Swindle-Reilly Lab for Biomimetic Polymeric Biomaterials for help and encouragement, and Drs. Matthew Reilly, Cynthia Roberts, and Jun Liu for advice and research guidance.

Data availability statement

The data that support the findings of this study are available from the corresponding author, KESR, upon request.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

We would like to acknowledge The Ohio State University College of Engineering for funding.