Osmosis as nature's method for establishing optical alignment

Abstract

For eyes to maintain optimal focus, precise coordination is required between lens optics and retina position, a mechanism that in vertebrates is governed by genetics, visual feedback, and possibly intraocular pressure (IOP).¹ While the underlying processes have been intensely studied in vertebrates, they remain elusive in arthropods, though visual feedback may be unimportant.² How do arthropod eyes remain functional while undergoing substantial growth? Here, we test whether a common physiological process, osmoregulation,³ could regulate growth in the sophisticated camera-type eyes of the predatory larvae of Thermonectus marmoratus diving beetles. Upon molting, their eye tubes elongate in less than an hour, and osmotic pressure measurements reveal that this growth is preceded by a transient increase in hemolymph osmotic pressure. Histological evaluation of support cells that determine the lens-to-retina spacing reveals swelling rather than the addition of new cells. In addition, as expected, treating larvae with hyperosmotic media post-molt leads to far-sighted (hyperopic) eyes due to a failure of proper lengthening of the eye tube and results in impaired hunting success. This study suggests that osmoregulation could be of ubiquitous importance for properly focused eyes.

Keywords: emmetropia; eye development; invertebrates; microfluidics; optical alignment; optics; osmosis; refractive error; visual systems; visually guided behavior.

Figure 1.. T. marmoratus larvae, often referred to as water tigers, are voracious aquatic hunters with two principal eyes (Eye 1 and Eye 2) on each side of the head.

A. Depicted here by schematic diagrams of orthogonal (a) and sagittal planes (b), each principal eye has a cuticular lens, a group of support cells (SCs, green), and distal (DR) and proximal (PR) retinae (purple). Insert b also schematically illustrates the organization of a support cell, and the square highlights the approximate position of light (B) and electron microscopic (C) images. The SCs construct the tubular portion of these eyes and form both the transparent core and the pigmented (P) boundary with the hemolymph (H). Scale bars are 10μm in B and 5μm in C. See also Figure S1.

Figure 2.. Uniform support cell counts, and variable osmotic pressure are consistent with the possibility of osmotically regulated eye growth early post-molt.

DAPI-stained sections of one-day-old (A) second (day1S) and (B) third (day1T) larvae. C. Use of depth coding to accurately identify overlapping nuclei in support cells (arrows). D. Cell counts in both eyes are not significantly different between second and third instar individuals (n = 5 for all samples, p_E1 = 1; p_E2 = 0.94, unpaired T-test). E. Measurements of hemolymph osmolality in T. marmoratus larvae reveal a significant transient increase in osmolarity preceding molting followed by a significant decrease in osmolality post molt (n_d1S = 17, n_d2S = 12, n_d3S = 22, n_FM = 11, n_d1T = 18, n_d2T = 18; p_d2S-d3S = 6.3 × 10⁻⁵, p_d3S-FM = 0.00044, p_d3S-d1T = 2.2 × 10⁻⁶, Wilcoxon’s rank sum test). Purple and blue lines indicate the osmotic pressure of hyperosmotic and isosmotic solutions that were used for experimental manipulations. FM= Freshly Molted F. Post-molt treatment with hyperosmotic solution leads to significantly higher post-molt osmotic pressure in the hemolymph. (n = 10, p = 4.3 × 10⁻⁵, Wilcoxon’s rank sum test). Scale bars = 100 μm. Boxplot whiskers show upper and lower bounds with the bottom of the boxes representing the bottom 25th percentile, the top being the 75th percentile, and the middle line is median. See also Figure S2.

Figure 3.. Treating freshly molted larvae with a hyperosmotic solution leads to hyperopia.

A. T. marmoratus larvae have four principal camera eyes (RE1, RE2, LE1, and LE2, as indicated by colored labels). B. Visualization of the proximal retina with a micro-ophthalmoscope was used to assess refractive states. C. Treatment with a hyperosmotic solution early post molt causes a significant hyperopic shift in all eyes compared to control larvae (n_Control = 36, n_Hyperosmotic = 22; p_LE1 = 0.0044, p_LE2 = 0.0044, p_RE1 = 6.0 × 10⁻⁴, p_RE2 = 0.0012, Wilcoxon’s rank sum test). D. In contrast, similar treatment with an isosmotic solution does not result in significant hyperopia (n_Control = 12, n_Isosmotic = 14; p_LE1 = 0.17, p_LE2 = 0.35, p_RE1 = 0.87, p_RE2 = 0.86, Wilcoxon’s rank sum test). Scale bar = 500 μm. Boxplot whiskers show upper and lower bounds with the bottom of the boxes representing the bottom 25th percentile, the top being the 75th percentile, and the middle line is median. See also Figure S3.

Figure 4.. Hyperosmotic treatment does not affect larval activity and larval prey detection but results in significant deficiencies in hunting ability.

A. In contrast to controls, some hyperosmotically treated individuals died, showed anatomical deformities (ND = no deformities), or postural defects (TP = typical posture) and thus were excluded from further analysis. B. Treatment with hyperosmotic solution does not influence mobility, as indicated by comparable distances traveled (n_Control= 14, n_Hyperosmotic = 11; p = 0.1, unpaired T-test). C) Additionally, treatment with hyperosmotic solution does not impair the ability to respond to prey (n_Control = 14, n_Hyperosmotic = 11; p = 0.43, unpaired T-test). Notably, all treated larvae had at least one visual response to a prey item. D & H. The number of strikes per individual are comparable between the groups in the horizontal arena (n_Control = 14, n_Hyperosmotic = 11; p = 0.51, unpaired T-test) but are significantly lower for the treated group in the vertical arena (n_Control = 14, n_Hyperosmotic = 11; p = 0.01, unpaired T-test). E & I. In both arenas, treatment leads to a significant reduction in hunting success (nControl = 14, nHyperosmotic = 11; pHorizontal = 0.0074, pVertical = 0.00087, Wilcoxon’s rank sum test). F & J. In both arenas, treatment also results in a significantly increased time to capture the first prey item (n_Control = 12, n_Hyperosmotic = 7, p_Horizontal = 0.006; nControl = 14, nHyperosmotic = 5, pVertical = 0.014, Wilcoxon’s rank sum test). G & K. Despite these deficits, the strike distance is similar for the two groups in both arenas (n_Control = 12, n_Hyperosmotic = 6, pHorizontal = 0.86; nControl = 13, nHyperosmotic = 5, pVertical = 0.443, Wilcoxon’s rank sum test). Boxplot whiskers show upper and lower bounds with the bottom of the boxes representing the bottom 25th percentile, the top being the 75th percentile, and the middle line is median. See also Figure S4 and Supplementary Movies 1–3..