The majority of cases of age-related vision loss in the world today are associated with the lens pathologies presbyopia and cataract (Frick et al., 2015). Presbyopia is the loss of the ability of the lens to dynamically change its shape (accommodate) to focus on near objects (Glasser and Kaufman, 1999), while cataract is the loss of the transparent properties of the lens (Asbell et al., 2005). Although it is believed these two lens pathologies are linked by oxidative damage to lens proteins (Kupfer, 1985), antioxidant-based therapies have to date proven ineffective in slowing the onset of either presbyopia or cataract (Braakhuis et al., 2019). Confirmation of the existence of a lens microcirculation system (MCS) (Vaghefi et al., 2011; Candia et al., 2012; Vaghefi and Donaldson, 2018), which generates a circulating flux of water, provides an alternative view not only to understand the onset of presbyopia and cataract, but also opens new therapeutic pathways to treat these lens pathologies.

It has been shown in animal lenses that movement of water through the lens generates a substantial and highly regulated pressure gradient (Gao et al., 2015), delivers nutrients and antioxidants to the lens center (Vaghefi and Donaldson, 2018), controls lens water content and volume (Donaldson et al., 2009), and maintains overall lens optics (Vaghefi et al., 2015). It is envisaged that in the human lens, these processes become dysfunctional with advancing age, and manifest first during middle age as presbyopia and then as cataract in the elderly. Consistent with this view, it has been shown that the free water content of the human lens increases with age (Lie et al., 2021) and is significantly increased in cataratous lenses (Bettelheim et al., 1986; Heys et al., 2008). Motivated by this growing recognition of the importance of water transport in the maintenance of the transparency and refractive properties of the lens, in this review we have focused on the role of water channels from the aquaporin (AQP) protein family in mediating, directing and regulating water flow in different regions of the lens. Since AQPs mediate the passive diffusion of water across cell membranes, the direction and magnitude of which is determined by osmotic gradients established across the membrane (Knepper, 1994; Verkman et al., 1996), we will first provide a review of the microcirculation system that generates the ion fluxes that drive water transport in the lens, before concentrating on how the expression of different lens AQPs and regional differences in their subcellular distribution, post-translational modification and regulation all contribute to lens water transport. We hypothesize that these regional differences, specifically the tissue and cellular locations of AQPs and their modifications facilitate water movement throughout the lens via the MCS. Further, we hypothesize that age-related changes in AQP structure and function lead to cataract formation. Finally, we will discuss therapeutic strategies being developed in other tissues that are targeting AQPs with the view to inform efforts to develop novel therapies on the lens.

To compensate for the lack of a blood supply, the lens operates an internal MCS (Figure 1) that delivers nutrients to deeper fiber cells, maintains fiber cell ionic homeostasis, and actively preserves the transparent and optical properties of the lens (Donaldson et al., 2017; Mathias et al., 2007; Mathias et al., 1997). The MCS is generated by a circulating current of Na⁺ ions, which primarily enters the lens at the poles and travels into the lens via the extracellular spaces between fiber cells (Figure 1, Influx). Na⁺ ions cross fiber cell membranes and return towards the surface via an intercellular outflow pathway mediated by gap junction (GJ) channels (Cx46 & Cx50) (Figure 1, Outflow), where Na⁺ is actively removed by Na⁺ pumps concentrated at the lens equator (Figure 1, Efflux). This circulating ion current generates a near isotonic water flow that enters the lens at both poles and exits at the equator (Candia et al., 2012). This flow of water has two consequences. First, the extracellular flow of water convects nutrients and antioxidants towards the deeper lying fiber cells where multiple membrane transporters enable cellular uptake (Donaldson and Lim, 2008). Mapping this extracellular delivery of solutes to the lens cores using MRI (Vaghefi et al., 2012) and confocal microscopy (Grey et al., 2003; Lim et al., 2009) has revealed that a barrier to extracellular diffusion forms in the inner cortex of the lens where mature fiber cells have lost their organelles. Through the use of spatially-resolved quantitative proteomics it has been shown that this closing of the extracellular space is accompanied with increased expression of gap junction, adherens junction and tight junction related proteins and by increases in the expression of AQP0 and its interaction partners, ezrin-radixin-moesin (ERM) proteins (Wang et al., 2021). It has been proposed that this barrier, by restricting radial extracellular space diffusion in the inner cortex, serves to preferentially direct water and solute fluxes into the lens nucleus via the sutures (Vaghefi et al., 2012).

Secondly, Mathias et al. have shown that the removal of water from the centre of the lens across the extracellular diffusion barrier, through an intercellular outflow pathway thought to be mediated by gap junction channels, generates a substantial hydrostatic pressure gradient, which ranges from 0 mmHg in the periphery to 335 mmHg in the lens centre in all lenses studied to date (Gao et al., 2013; Gao et al., 2011). This lens pressure gradient is maintained through a dual feedback system that utilizes the mechanosensitive Transient Receptor Potential Vanilloid channels, TRPV1 and TRPV4, to sense changes in pressure at the surface of the lens (Gao et al., 2015). TRPV1 and TRPV4 channels sense decreases and increases, respectively, in lens pressure and utilize distinct signalling pathways to modulate Na⁺ pump (Gao et al., 2015) and NKCC1 (Shahidullah et al., 2018) activity to ensure that a constant hydrostatic pressure and therefore water content is maintained (Figure 2). Altering the tension applied to the mouse lens via the zonules, through pharmacological modulation of ciliary muscle contractility, alters lens surface pressure and this change in pressure is mediated by TRPV1 and TRPV4 channels (Chen et al., 2019). This finding suggests that the hydrostatic pressure gradient, and therefore steady state water transport, in the non-accommodating mouse lens can be altered by contraction of the ciliary muscle.

In this scheme, the active transport of Na⁺ drives a directed isotonic flow of fluid through the lens. Near isotonic fluid flow, in turn, requires a high membrane permeability to water that is known to be mediated by the aquaporin family of water channels (Knepper, 1994; Verkman et al., 1996). However, what is not known is how different lens AQPs, which exhibit distinctly different functional properties, regulation mechanisms, and expression patterns, contribute to the generation of the microcirculation system. To address this knowledge gap we first discuss general properties and functions of AQPs before comparing the expression patterns of three lens AQPs and their modified forms. Finally, we detail lens AQP functional properties and how these properties are regulated in order to develop a working model on how the different AQP isoforms contribute to the generation of the lens microcirculation system which is so central to the maintenance of lens homeostasis.

Aquaporins are a class of transmembrane proteins that function as water channels. These water channels exist in 13 known isoforms (AQP0-AQP12) in mammals and act to move water bidirectionally across biological membranes through osmotically driven passive diffusion (Ishibashi et al., 2011; Rojek et al., 2008). Clearly, water transport is an essential cellular function as evidenced by the wide and diverse tissue-specific expression of aquaporins. For example, AQPs show wide-ranging distribution throughout tissues such as the eye, brain, kidney, liver, and heart with varying physiological functions in each tissue (Azad et al., 2021). Aquaporin isoforms display many structural similarities including six transmembrane domains and two conserved asparagine-proline-alanine (NPA) motifs that are associated with water transport (Maurel et al., 2015) (Figure 3). However, there is a subfamily of aquaporins, coined the superaquaporins, that show sequence deviation from the greater family, particularly in the NPA motif (Ishibashi et al., 2014). Early crystallographic studies confirmed previous reports that physiologically, aquaporins adopt a tetrameric structure (Mitra et al., 1994; Daniels et al., 1999). Recent native mass spectrometry results have confirmed this tetrameric structure of AQP0 in solution (Harvey et al., 2022).

The water permeability rate varies between aquaporin isoforms; for instance, AQP0 water permeability is approximately 20-fold lower than AQP1 and AQP5 (Chandy et al., 1997; Yang and Verkman, 1997). In addition to transporting water, some aquaporins can transport other molecules such as glycerol (AQPs 3, 7, 9, and 10), ammonia (AQPs 3, 7, 8 and 9), urea (AQPs 7, 9 and possibly 3), and hydrogen peroxide (AQPs 0, 1, 3, 5, 8 and 9) (Bienert et al., 2008; Litman et al., 2009; Ishibashi et al., 2014; Varadaraj and Kumari, 2020). Aquaporins may play other roles in addition to their roles as membrane channels. For example, AQP0 has also been shown to possess cell adhesive properties that are critical to establishing lens transparency and that may be involved in the development the refractive index gradient, that is a key aspect of the refractive properties of the lens (Kumari and Varadaraj, 2009).

Five aquaporins have been reported in the lens (AQP0, AQP1, AQP5, AQP7, AQP8) with each protein displaying unique localization and abundance patterns, discussed in detail below (Figure 4A). Since the rate of movement of water across cell membranes, given by the membrane permeability to water (P_H2O), is enhanced by the presence of AQPs in cell membranes, we can expect that differences in AQP expression, subcellular distribution, lipid and protein interactions, and function will all contribute to the directed movement of water into and out of the lens. In addition, other AQP properties, such as cell adhesion, could also be important in establishing the lens MCS and will be discussed below. There are few reports on lenticular AQP7 and AQP8 expression and both appear to be expressed exclusively in lens epithelial cells (Tran et al., 2013; Hayashi et al., 2017; Varadaraj and Kumari, 2020). Although AQP8 is a known peroxiporin (Bienert et al., 2008), the specific function of AQP7 and AQP8 in the lens has yet to be determined. Thus, our discussion will focus on AQPs 0, 1, and 5.

As fiber cells differentiate from lens epithelial cells at the lens equator, we expect normal differentiation-related processes, such as protein synthesis/degradation, protein trafficking and post-translational modification, to govern protein expression and function. However, fully mature lens fiber cells are organelle-free and therefore lack the ability to synthesize new protein (Bassnett and Costello, 2017). Thus, regulation of protein function in mature fiber cells is accomplished by post-translational modifications, protein-protein interactions, or by changes in the local environment (Schmid and Hugel, 2020). Layered on top of these regulatory processes is a set of additional changes that occur to lens AQPs as we age. Because the lens continues to grow throughout life, an age gradient is established where the oldest cells are found in the lens core and the youngest cells are found in the outer cortex. Accumulation of age-related lens protein modifications have been studied for multiple decades (Takemoto and Takehana, 1986; Garland et al., 1996; Lampi et al., 1998; Ueda et al., 2002; Schey et al., 2020). Interestingly, modifications, as well as protein interactions and lipid environment, change throughout the lens suggesting that the functionality and regulation of lens AQPs also changes as a function of their location (i.e., age) in the lens. Such age-related changes could lead to changes in the local properties of AQPs that may contribute to the development of age-related cataract. In the next two sections we first highlight mechanisms known to regulate lens AQP functionality before discussing how age-related changes can alter this regulation in different areas of the lens.

Experimental confirmation of the existence of water fluxes (Mathias et al., 1997; Mathias et al., 2007; Donaldson et al., 2010; Gao et al., 2011; Candia et al., 2012; Vaghefi et al., 2012) along with evidence of their dynamic regulation (Gao et al., 2015) led to the development of an initial model of how regional differences in lens aquaporin expression, localization, and regulation combine to produce regional differences in fiber cell P_H2O that facilitate the outflow and efflux of water from the lens driven by the local osmotic gradients generated by the ion fluxes that drive the MCS (Schey et al., 2017). However, this model was based solely on regional differences in lens AQP functionality and did not consider that fiber cells within a specific lens region also exhibit differences in the distribution of AQP0 and AQP5 that occur in the different membrane domains observed along the length of individual fiber cells (Figure 4B) (Zampighi et al., 2002; Grey et al., 2009; Petrova et al., 2020). Thus, AQP channels in the anterior and posterior tips of fiber cells located at the anterior and posterior poles of the lens will mediate water influx, while in the same cell, water flow across the lateral membranes will be directed out of the lens via an intracellular pathway mediated by gap junctions where at the lens surface AQP channels mediate the efflux of water from the lens. In this section, we now present an updated model of regional differences in AQP functionality which includes recently acquired data (Petrova et al., 2020) that suggest differences in localization and regulation of AQP5 in the anterior and posterior sutures differentially modulates the influx of water at the anterior and posterior poles. While this model attempts to show how regional differences in the molecular and cellular structure and function of lens AQPs contributes to overall function of the lens at the whole tissue level, like all models it is at best an approximation of the real situation and will require further experimental validation.

In this updated model (Figure 5), AQP1 is a constitutively active water channel exclusively localized to lens epithelial cells, where it mediates water influx and efflux in the central and equatorial regions of the lens, respectively. In contrast, fiber cell P_H2O is dependent on both AQP0 and AQP5, which undergo distinctly different differentiation-dependent changes in their subcellular location and post-translational modification which by altering local P_H2O contribute to the overall magnitude and directionality of water fluxes that circulate through the lens. Based on our current understanding of these changes to AQP0 and AQP5 expression we propose that full length AQP0 provides a basal level of water permeability in the outer lens cortex and this permeability can be altered by AQP0 phosphorylation/calmodulin binding (Lindsey Rose et al., 2006; Fields et al., 2017). In this outer region of the lens, AQP5 appears to operate as a regulated water channel, where its membrane location, and hence contribution to fiber cell P_H2O, can be modulated by either mechanically or pharmacologically altering zonular tension via a process that is mediated by TRP channels (Petrova et al., 2018; Petrova et al., 2020). In the lens inner cortex, AQP0 forms cell-cell junctions, either AQP0-AQP0 junctions or AQP0-plasma membrane junctions (Zampighi et al., 1989; Gonen et al., 2004), that may enable the formation of an extracellular diffusion barrier (Grey et al., 2009). Recent proteomics analysis of the barrier region indicates that adhesion proteins also play a role in barrier formation (Wang et al., 2021). A major consequence of this barrier formation is that delivery of water and nutrients to the lens core occurs via the lens sutures (Vaghefi and Donaldson, 2018). Further, AQP modifications that reduces the P_H2O of the plasma membrane relative to the P_H2O of the gap junctions would tend to facilitate the cell-to-cell removal of water via the intracellular outflow pathway mediated by gap junctions. In this regard it has been shown that AQP0 becomes associated with the periphery of gap junction plaques in this region (Grey et al., 2009). In contrast, we have preliminary data for AQP5 that shows in the inner cortex AQP5 accumulates in plaque-like structures on the broadsides of fiber cells (unpublished data) that resemble the gap junction plaques known to form in this region of the lens (Jacobs et al., 2004). Whether AQP5 also contributes to the intracellular outflow of water from the core that is thought to be mediated by gap junctions (Gao et al., 2011) remains to be determined. In the lens core, extensive age-related modifications such as truncation of the C-terminus of AQP0 will change the regulation of AQP0 P_H2O permeability by phosphorylation/calmodulin binding and interaction with binding partners, while the altered lipid environment in the lens is predicted to reduce the P_H2O of AQP0 based on in vitro studies (Tong et al., 2013). Since it appears that the majority of C-terminus of AQP5 is largely intact in the lens core (Grey et al., 2013; Petrova et al., 2015), we propose that the bulk of water transport occurs via full length AQP5 in the core rather than via truncated AQP0 water channels.

Since the diffusion of water through water channels is driven by the local osmotic gradient established by ion transport we will now propose how these observed spatial differences in AQP0 and AQP5 location and functionality combine to drive water transport through the lens (Figure 5). In parallel to ion movement, water enters the lens at both poles via an extracellular route that is associated with the sutures (Candia et al., 2012; Vaghefi et al., 2011). This influx zone spans from the lens surface to the core of the lens (Figure 5, influx pathway) and crosses the extracellular diffusion barrier formed in the inner cortex (Vaghefi and Donaldson, 2018; Vaghefi et al., 2012). Since in the outer cortical region of this influx zone AQP5 is not associated with the apical or basal tips of fiber cells (Petrova et al., 2020), we envisage that water will only enter fiber cells via AQP0 water channels (Figure 5, anterior influx—outer cortex). In addition, it is possible water will remain in the extracellular space between fiber cells as it moves away from the sutures towards the equatorial regions of the lens before crossing fiber cell membranes via AQP0 and/or AQP5 channels located in the lateral membranes of fiber cells. In the inner cortex the restriction of extracellular space would reduce this lateral extracellular movement of water away from the sutural influx zone and direct water flow towards the lens core. In this inner cortical region movement of water from the suture into the fiber cells via AQP5 can be dynamically modulated by changes in zonular tension in the anterior (Figure 5, anterior influx–inner cortex), but not the posterior sutural influx zone (Petrova et al., 2020). We speculate that this ability to differentially change water influx in this region of the lens may be a mechanism via which the lens can change the curvature of its anterior surface and hence the optical properties of the lens. In the core of the lens the water delivered to this zone would then be taken up into fiber cells via both AQP0 and AQP5 (Figure 5, influx/outflow—core), however as mentioned above, we are currently unsure about the relative contributions of truncated AQP0 and non-truncated AQP5 to fiber cell P_H2O in this region of the lens.

Once delivered to the central lens, water moves towards the surface of the lens via an intracellular outflow pathway that delivers water to an equatorial efflux zone where water leaves the lens. In this outflow pathway (Figure 5, outflow—inner cortex) the gradients for water movement are generated by both the movement of ions through gap junctions and the hydrostatic pressure gradient generated by the outflow of water. In this area of the lens we envisage that limiting the flow of water across fiber cell membranes into the extracellular space by altering the functionality of AQP0/AQP5 via post-translational modifications will promote the intracellular passage of water through gap junction channels. Finally, once water reaches the equatorial efflux zone it can leave the lens through AQP1 or AQP5 channels located in equatorial epithelial cells or by AQP0 or AQP5 channels in peripheral fiber cells (Figure 5, outflow—equator). Since the subcellular distribution of AQP5 can be dynamically altered by mechanically or pharmacologically altering zonular tension via a process that is mediated by TRP channels (Petrova et al., 2020), P_H2O of fiber cells in the efflux zone can be regulated in parallel to the ion fluxes that drive the transport of water throughout the lens. Taken together, it is clear that the expression level, the subcellular localization, the extent of modification, the local lipid environment, and the extent of protein-protein interactions all play roles in the regulation lens water transport which has been shown to be so critical for the maintenance of the transparent and refractive properties of the lens (Donaldson et al., 2017).

Based on our knowledge that AQP mutations or AQP deletion leads to cataract formation (Shiels and Bassnett, 1996; Shiels et al., 2000; Okamura et al., 2003; Tang et al., 2021) and that AQPs are an integral component of lens water transport, we propose that age-related modifications to AQP functionality are involved in cataract formation. Consistent with this view, glycemic stress in AQP5 knockout mice leads to cataract, suggesting a protective role of AQP5 under conditions of stress (Sindhu Kumari and Varadaraj, 2013). Thus, if AQP age-related modifications and diabetic stress lead to altered AQP function, then we surmise that age-related opacification is a possible consequence. It is also conceivable that age-related changes in water transport could affect lens stiffness and, as such, AQPs could represent a novel target for presbyopia treatment.

In other tissues where AQP dysfunction has been linked to numerous diseases, AQPs have been promoted as therapeutic drug targets (Verkman et al., 2014; Soveral and Casini, 2017; Salman et al., 2022). For example, the involvement of AQPs in cancer initiation, cell migration and tumor angiogenesis, has made AQPs attractive targets for novel anticancer therapies (Wang et al., 2015; Elkhider et al., 2020). AQP1 inhibitors have also been considered as a therapeutic target for the treatment of intraocular hypertension in glaucoma (Patil et al., 2018). Thus, therapeutic targeting of aquaporins is an active area of research (Villandre et al., 2022), including in ocular tissues, and we can expect continued discoveries of new AQP modulators. In the context of lens water transport and cataractogenesis, the lens microcirculation system needs to be tightly regulated and it is worth considering lens AQPs as a potential anti-cataract targets (Pierscionek, 2021). In the search for natural products for cataract treatment, Heliotropium indicum extract was found to alleviate selenite-induced cataract and also increase AQP0 levels (Kyei et al., 2015). Considering the increased water content in the aged lens nucleus (Siebinga et al., 1991), the accumulation of water and sodium, and the swelling of fiber cells in AQP5 deficient lenses under hyperglycemic stress (Sindhu Kumari and Varadaraj, 2013), therapeutics designed to increase efflux of water could be used to alleviate cataractous conditions.

There are many features of our working model for lens water transport that require further investigation. Given the known effects of post-translational modifications, lipid environment and expected effects on local water permeability, spatial definition of these lens properties is necessary to define the molecular basis of lens water transport. It is important to recognize that many previous studies have used either homogenized whole lenses or separation methods with low spatial resolution (Ball et al., 2004; Gutierrez et al., 2016) resulting in limited information on spatial regulation of lens water transport. Furthermore, most spatially-resolved lens studies have examined equatorial sections (Wenke et al., 2015; Wang et al., 2021); however, there may be axial protein differences where expression or modifications occur along the length of individual fiber cells (Zampighi et al., 2002; Kuszak et al., 2004). Thus, spatially-resolved methods of analysis are necessary to increase our understanding of AQP structure and function in the different regions of the lens. In addition, as described above, AQP structure and function can be affected by age. We predict, therefore, that the combined local effects of protein modification, protein-protein interactions, lipid environment, and age-related alterations will affect global lens water transport; a property that has yet to be measured as a function of age. With a better understanding of AQP function as a function of age, we expect that new directions in the treatment of presbyopia and cataract will emerge.