Hydrocortisone treatment as a tool to study conjunctival placode induction

Abstract

Background

Conjunctival placodes are a series of placodes that develop into the conjunctival (scleral) papillae and ultimately induce a series of scleral ossicles in the eyes of many vertebrates. This study establishes a hydrocortisone injection procedure (incl. dosage) that consistently inhibits all conjunctival papillae in the embryonic chicken eye. The effects of this hydrocortisone treatment on apoptosis, vasculature, and placode‐related gene expression were assessed.

Results

Hydrocortisone treatment does not increase apoptotic cell death or have a major effect on the ciliary artery or vascular plexus in the eye. β‐catenin and Eda expression levels were not significantly altered following hydrocortisone treatment, despite the absence of conjunctival papillae. Notably, Fgf20 expression was significantly reduced following hydrocortisone treatment, and the distribution of β‐catenin was altered.

Conclusions

Our study showed that conjunctival papillae induction begins as early as HH27.5 (E5.5). Hydrocortisone treatment reduces Fgf20 expression independently of β‐catenin and Eda and may instead affect other members of the Wnt/β‐catenin or Eda/Edar pathways, or it may affect the ability of morphogens to diffuse through the extracellular matrix. This study contributes to a growing profile of gene expression data during placode development and enhances our understanding of how some vertebrate eyes develop these fascinating bones.

Key Findings

• Incrementally modulating the timing of hydrocortisone injections resulted in an increase in placode inhibition success

• The mechanism of hydrocortisone‐induced inhibition does not involve apoptotic cell death or changes in vasculogenesis.

• hydrocortisone treatment affects some placode genes but not others

1. INTRODUCTION

Placodal systems give rise to an impressive range of structures across vertebrate phylogeny. These placode‐derived structures range from specializations of the integumentary system (e.g., mammary glands, feathers, scales, hairs, and scutes) to cranial ganglia and sensory organs that characterize the vertebrate head (e.g., olfactory epithelium, inner ear, lens of the eye, ganglia of the facial, glossopharyngeal, and vagal nerves). Though diverse in terms of their appearance and function, these structures begin their development as localized thickenings of the epithelium (i.e., placodes) produced by reciprocal signaling between epithelial and mesenchymal tissues. Through a complex series of developmental and gene regulatory processes, (e.g., epithelial‐mesenchymal signaling, cell aggregation, proliferation, adhesion, migration, etc.), placodes are instructed to develop into specialized derivative structures. Those that develop into derivatives of the epithelium are referred to as cutaneous placodes, ¹ , ² , ³ and those that develop from the cranial non‐neural ectoderm are the cranial placodes. ⁴ , ⁵ , ⁶ , ⁷

Placodes are the developmental precursors of conjunctival papillae, ⁸ , ⁹ which form in a series of (typically) 14 papillae in the corneal‐scleral limbus of the chicken eye during embryonic days 6.5–8. These papillae are transient structures that begin disappearing around embryonic day 10. While present, these specialized structures of the conjunctival epithelium are responsible for inducing a ring of scleral ossicle bones. ⁸ , ¹⁰ , ¹¹ The first phase of scleral ossicle development is therefore the thickening of the conjunctival epithelium of the eye into a series of placodes, which then differentiate into raised conjunctival papillae. ¹² An entire ring of 13–15 papillae appears over the course of 1.5 days (HH30–HH34, Figure 1) in a highly conserved and intriguing spatiotemporal pattern. ¹¹ , ¹³ During the second phase of scleral ossicle development (HH35–HH38), conjunctival papillae instruct skeletogenic condensations to form in the underlying ectomesenchyme in a 1:1 ratio. ¹⁰ Should one of these papillae be manually removed or inhibited, the corresponding underlying scleral condensation will not be induced. ¹⁰ , ¹¹ , ¹⁴ , ¹⁵ , ¹⁶ After induction is complete, the papillae degenerate. Once induced, the skeletogenic condensations grow until reaching a critical size that triggers osteoblast differentiation. ¹⁷ This then leads to the deposition of a bony matrix and formation of a ring of overlapping scleral ossicle bones. ¹⁸ These bones form in the eyes of most reptiles, including all birds, but have been secondarily lost in mammals and amphibians (reviewed in ¹⁹ ).

FIGURE 1.

Conjunctival papillae of the chicken (Gallus gallus) eye develop in a distinctive pattern across 1.5 days and five developmental stages (HH30–HH34). (A) Papillae of the temporal group form first, followed by the nasal, dorsal, and ventral papillae. (B) The first conjunctival papilla forms above the ciliary artery (CA, red line) at HH30. By HH32, 6–8 distinct papillae are present in the temporal and nasal groups. By HH34, the full ring (typically 13–15 papillae) is completed with papillae from the dorsal and ventral groups, where the last papilla forms on either side of the choroid fissure (black line). The regions of epithelium between papillae are termed interpapillary (IP) regions.

The general process of how placode formation is regulated in space and time has been an engaging pursuit for many researchers and could include Turing reaction–diffusion mechanisms, molecular oscillators, chemotaxis, and/or mechanochemical patterning. ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ Common to most placode systems is the involvement of Wnt/β‐catenin, bone morphogenetic protein (Bmp), fibroblast growth factor (Fgf), ectodysplasin (Eda/Edar), and/or hedgehog (Hh) signaling families. For example, Ho et al. ²⁶ recently demonstrated a model of feather placode induction in the embryonic chicken that integrates β‐catenin, Eda, Eda receptor (Edar), Fgf20, and Bmp4 expression with mechanochemical cell aggregation. Following widespread expression of β‐catenin and Edar across the dorsal skin, a progressive front of Eda expression triggers placode formation in its wake. Placode formation is initiated by Fgf20 signaling, induced by both β‐catenin and Eda, which then promotes localized cell aggregation below the epithelium. The underlying cell aggregations lead to a positive feedback loop of Fgf20 expression, thereby sustaining placode formation, and further stimulating a BMP4 gradient that acts as a long‐range inhibitor to establish interplacodal spacing. The general integration of the key signaling pathways illustrated in this example is common to other placodal systems (i.e., taste bud, scale, hair, etc.). ³ , ³⁰ , ³¹ , ³²

Prior to genetic inhibition mechanisms commonly used in developmental biology, chemicals such as hydrocortisone (HC) were used as a treatment to inhibit placodes (e.g., scales and feathers). ³³ , ³⁴ HC is a glucocorticoid that has also been used clinically to treat a variety of skin, hormonal, and respiratory conditions (e.g., adrenocortical insufficiency, rheumatoid arthritis, dermatitis, asthma). As such, HC can affect inflammation, vasculature, extracellular matrix, apoptosis, and bone maintenance. ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ A handful of studies have explored the effect of HC treatment on development of the conjunctival papillae in the chicken eye. Johnson showed that conjunctival papillae induction is prevented after an injection of HC. ⁴¹ More recently, Hammer and Franz‐Odendaal ⁴² determined that delivering two doses of HC (1.1 mM) at separate developmental stages (i.e., HH29 and HH30) inhibited the majority of conjunctival papillae in the chicken eye. They also showed that this HC treatment had a later effect—that is, it affected the formation of the scleral ossicles, disrupted the spatial expression pattern of several extracellular matrix (ECM) components in the scleral mesenchyme, and reduced the vasculature network within the sclera at HH36 during the period of scleral condensation formation. In a few cases, complete inhibition of the conjunctival papillae was achieved but not explored further. No studies to date have addressed whether HC affects rates of apoptosis during conjunctival papillae formation.

The overall objective of the present study was to gain insight into the genetic and cellular mechanisms that contribute to the loss of placodes after HC treatment. Specifically, we first sought to establish a reliable methodology to prevent all conjunctival papillae development using HC. Second, we wanted to determine whether the prevention of papillae development was due to increased apoptosis or disrupted vasculature. Third, we wanted to determine the effect of HC treatment on three key genes from three different regulatory pathways commonly involved in placode formation—namely β‐catenin, Eda, and Fgf20. These genes (or their receptors) are known to be expressed during conjunctival papillae formation. ⁴³ , ⁴⁴ , ⁴⁵ Furthermore, Fgf20 is disrupted in the scaleless (sc/sc) chicken mutant, which has disrupted cutaneous placodes (i.e., conjunctival, scale, and feather; see references ⁴⁶ , ⁴⁷ ).

2. RESULTS

2.1. Two HC injections inhibit the entire ring of conjunctival papillae

In order to create a phenotype that completely lacks conjunctival papillae, the dose and timing of HC injections from Hammer and Franz‐Odendaal, ⁴² namely 1.1 mM injections at HH29 and HH30, was further refined through a series of 12 injection experiments (Figure 2A). All experiments involved two HC injections. Experiments 1–4 consisted of adjusting the second injection time only while keeping the first injection time point constant at HH29 until all late‐forming papillae were inhibited. This was achieved with an injection time of 2 h prior to HH30. Subsequently, experiments 5–12 focused on adjusting only the first injection time, while keeping the second injection time point constant at 2 h before HH30, until the very first papillae in the sequence were also consistently inhibited. This was achieved with an injection time of 30 h before HH30. During these adjustments of the first HC injection (experiments 5–12), the earliest papillae persisted despite advancing the first injection time. Thus, the concentration of the HC injections was also increased from 1.1 to 2.0 mM in experiments 10–12.

FIGURE 2.

Refining hydrocortisone (HC) injection experiments. Left panel: The series of injection experiments used to refine HC injection times is shown under Methods. Injection time points are in reference to HH30 (E6.5), which is the developmental stage where the first conjunctival papilla is visible. Experiment 1 begins with the injection times and the concentration used by Hammer and Franz‐Odendaal. ⁴² Adjustments to the second injection time point were tested in experiments 2–4, followed by adjustments to the first injection throughout experiments 5–12. The concentration of HC was increased from 1 mM to 2 mM in experiments 10–12. Right panel: The number of conjunctival papillae present in ‐treated eyes across each injection experiment was recorded—the percentage of eyes containing a given range of conjunctival papillae is depicted in the corresponding bar graphs. Incrementally modulating the timing of hydrocortisone injections resulted in an increase in inhibition success. The full ring of conjunctival papillae was most consistently inhibited with 2 mM HC injections delivered either 30 or 28 h before HH30 and again 2 h before HH30 (i.e., Experiments #11 and 12). Sample sizes of HC‐treated embryos in each experiment are shown on the right of the corresponding bar graph. Group sample sizes across the 12 experiments: CNI (n = 21), CI (n = 71), and HC (n = 75).

The effectiveness of each of these injection experiments (#1–12) was quantified by the number of papillae present in HC‐treated eyes and the percentage frequency of each outcome (Figure 2B). Of experiments 1–4, experiment three resulted in eyes with as few as 0 papillae and at most 5–8 papillae, and it was observed that all late‐forming papillae (ventral group) were absent while the early‐forming papillae (temporal and nasal groups) were present (Supplementary Data 1). The first injection was then adjusted systematically in 2‐h intervals in experiments 5 through 8. Experiment 8 resulted in five embryos with 0 papillae (Figure 2B); however, this result was not consistently reproducible with further use of these injection times. Therefore, we continued adjusting the first injection time to earlier stages until we produced a complete loss of conjunctival papillae with a larger sample size (Figure 2B, n = 29, experiments 11 and 12). Through this empirical testing, we determined that delivering 2 mM HC injections 30 or 28 h before HH30 (i.e., HH27.5), and again 2 h before HH30, resulted in a reproducible complete inhibition of all the papillae (Figure 3).

FIGURE 3.

Uninjected control, injected control, and hydrocortisone‐treated phenotypes at HH34 upon completion of conjunctival papillae development. (A) A full ring of 14 conjunctival papillae is visible in the control no injection eye. (B) A full ring of 14 conjunctival papillae is also visible around the eye of control injection embryos, demonstrating that there is no effect of the injection process or solvent on conjunctival papillae development. (C) Hydrocortisone‐treated eyes lack the entire ring of conjunctival papillae. All scale bars are 1 mm.

From these experiments, we refined the injection parameters established by Hammer and Franz‐Odendaal ⁴² to more reliably produce embryos that lacked all conjunctival papillae for further analysis of the effects of HC treatment. While this treatment is robust, it is based on a very precise treatment dose and timing, which does not always yield perfect results in the context of variations in development between eggs. In continuing with these injection timings in numerous rounds of downstream tissue collection, we occasionally encountered HC‐treated eyes with 1–2 of the early papillae, which we did not include in any of our analyses. Thus, care should be taken during tissue collection to confirm that all placodes and papillae are indeed absent. From these injection times, we also determined that the developmental period from HH27.5 to HH30 is critical for conjunctival placode induction. No obvious effects on other placodal systems (such as the scales or feathers) were noted in any samples. We did however attempt to determine whether other morphological characteristics could be used to verify complete inhibition when collecting eye tissues at HH32 (discussed below).

2.2. HC treatment has mild effects on other parts of the embryo

While no obvious effects on eye size were noted, some additional effects of HC on the eye include reduced eyelids and nictitating membranes, and a protruding cornea. HC treatment also affected embryo size, closure of the body cavity, and digit curvature. These morphological effects were previously documented by Moscona and Karnofsky, ⁴⁸ Johnson, ⁴¹ and Hammer and Franz‐Odendaal. ⁴² We investigated whether we could use other embryonic features such as toe angle or body cavity closure as an indicator of a complete inhibition after HC treatment. This data and methodology are presented as Supplementary data 2 and 3 with associated figures. We found that neither of these features could effectively distinguish between HC‐treated eyes with complete inhibition (no papillae present) and those that developed a few (1–3) papillae. Therefore, these secondary features were not utilized further.

2.3. HC treatment does not induce apoptotic cell death

In order to determine whether the absence of papillae in HC‐treated embryos is due to apoptotic cell death, we used a TUNEL assay to assess cell death at three different time points following HC treatment and in two regions of the eye—above the ciliary artery (CA) where the first papillae appear and in the adjacent interpapillary (IP) region (Figure 4, Table 1, Supplementary Data 4). Just 6 h after the first HC injection (HH28), there was no significant difference in apoptotic indices between HC‐treated eyes (0.086%) and CNI and CI controls (0.049% and 0.159%, respectively) (F(2, 11) = 1.26, p = .323, ω² = 0.035; Supplementary Figure 4A, C). A full 24 h after HC treatment (HH30), nearing the induction of the first papilla in the CA region, there was also no significant change in apoptotic indices due to HC treatment (0.029%) compared to CNI and CI controls (0.047% and 0.043%, respectively) (F(2, 11= 0.189, p = .831, ω² = −0.131); Supplementary Figure 4B, C). Upon conclusion of conjunctival papillae formation at HH34, average apoptotic indices in the CA region for CNI and CI control eyes are 1.260% and 0.862%, respectively. By comparison, HC‐treated eyes have an average apoptotic index of 0.183. The apoptotic indices for HC and control (CNI and CI) samples are statistically significant (t(15) = −2.60, p = .020), while CNI and CI controls are not significantly different from one another (t(15)= −1.02, p = 0.322) in the CA region (Supplementary Figure 4D, F). Within the IP eye region, average apoptotic indices in CNI and CI control groups are 0.290% and 0.335%, respectively. When comparing the IP region of both controls to HC‐treated eye tissue in the CA region, there is no difference in apoptotic indices [ANOVA analysis: F(2, 15) = 0.349, p = .711, ω² = −0.078; Supplementary Figure 4E, F]. In summary, HC‐treated eyes have lower apoptotic indices than control eyes in the CA region and similar indices to control eyes in the IP region. Furthermore, linear mixed models show that the apoptotic indices are significantly higher in the CA regions compared to the IP regions of CNI control eyes (t(5) = 2.94, p = .032, Supplementary Figure 4G, I). In contrast, there is no statistically significant difference in these regions of the CI control eyes (t(5) = 1.80, p = .132, Supplementary Figure 4H, I). Collectively, these results indicate that apoptotic cell death is not responsible for the absence of conjunctival papillae observed after HC treatment.

FIGURE 4.

Summary of treatment groups (CNI, CI, and HC) and eye regions (CA and IP) imaged in whole‐mount as maximum intensity projections following the TUNEL assay (green) and DAPI staining (blue). (A)–(C) Ciliary artery region of experimental control and HC‐treated tissues at HH28, 6 h after HC treatment. (D)–(F) Ciliary artery region of experimental control and HC‐treated tissues at HH30, 24 h after HC treatment (G)–(I) Ciliary artery region of experimental control and HC‐treated tissues at HH34, upon completion of the conjunctival papillae ring. Dotted white line outlines the papillae in CNI and CI control samples. (J)–(L) Interpapillary region of experimental CNI and CI controls at HH34. (L) Negative control CNI tissue treated with DNaseI and exempted from TUNEL enzyme to demonstrate absence of TUNEL signal. (M)–(O) Positive controls in each treatment group (CNI, CI, and HC) exposed to DNaseI and TUNEL enzyme to demonstrate positive TUNEL signals. All scale bars are 50 μm in size. CA, ciliary artery region; CI, control injection; CNI, control no injection; HC, hydrocortisone; IP, interpapillary region.

TABLE 1.

Summary of average apoptotic indices, expressed as percent TUNEL‐positive cells, for control (CNI and CI) and HC‐treated eyes across three time points and two regions of the eye.

	Mean apoptotic index (%) [95% CI]
Region—Stage	HC	CI	CNI
Ciliary artery—HH28	0.086 [−0.005, 0.176]	0.159 [0.048, 0.270]	0.049 [−0.062, 0.159]
Ciliary artery—HH30	0.029 [−0.015, 0.073]	0.043 [−0.011, 0.097]	0.047 [−0.007, 0.101]
Ciliary artery—HH34	0.183 [−0.403, 0.770]	0.862 [0.275, 1.448]	1.260 [0.673, 1.847]
Interpapillary—HH34	–	0.335 [0.054, 0.616]	0.290 [0.009, 0.571]

Note: Stage HH28 is 6 h after HC treatment, stage HH30 is 24 h after HC treatment, and HH34 marks the completion of conjunctival papillae formation. 95% confidence intervals are provided in brackets.

Abbreviations: CI, control injection; CNI, control no injection; HC, hydrocortisone.

2.4. HC does not significantly affect eye vasculature

Microinjections of fluorescently‐conjugated lectin coupled with confocal microscopy provided a detailed depiction of developing eye vasculature at stage HH34 when the conjunctival papillae ring is complete (Figure 5). A series of z‐stack images taken across a 4 × 4 grid revealed a dense meshwork of vasculature (the vascular plexus) developing from the temporal side of the eye. This meshwork is located more superficially than the CA and extends up to the epithelial surface of the eye. Our data shows that the boundary of this developing vascular plexus does not yet meet the ring of conjunctival papillae by HH34 (Figure 5A, C).

FIGURE 5.

Vascular analysis of lectin‐injected control and HC‐treated eyes at HH34. Methodology used to assess vasculature in maximum intensity projections of whole‐mount tissue is shown in (A) a control (CNI) eye (n = 8). (C) shows a HC‐treated eye for comparison (n = 8). Ciliary artery width was measured in micrometers (gray line), and percent area of vasculature was measured in three boxed regions of the eye (region A, region B, and region C). In (B) and (D) boxed regions are enlarged to show vasculature and masks (black and white) used to calculate percent vasculature in Fiji. (E) Graphs of the measurements for ciliary artery (CA) width and percent area of regions A, B, and C, including means (black dots) and 95% confidence intervals. CNI, control; HC, hydrocortisone.

With respect to the effect of HC treatment, the CA width in HC‐treated eyes (136.6 μm ± 27.59) was not statistically different from the width in control (CNI) eyes (112.9 μm ±15.85) (t(11.17) = −2.11; p = .058, g = −1.00 [−2.07, 0.08]; Figure 5E; Table 2). In control eyes, vasculature covered 41%–69% of the measured regions, which was not statistically different to the percentages in HC‐treated eyes, namely 41%–64% (Figure 5E; Table 2) (region A: t(14.00) = 2.04; p = .061, g = 0.96 [−0.11, 2.03]; region B: t(9.47) = 0.04; p = .971, g = 0.02 [−1.00, 1.03]; region C: t(13.32) = 1.11; p = .288, g = 0.52 [−0.51, 1.55]). Collectively, this data does not support an effect of HC treatment on CA width or the percent area covered by the developing vascular network. Despite a lack of statistical significance, the Hedges's g values for CA width (−1.00) and percent area in Region A (0.96) are both indicative of large effect sizes (i.e., effect sizes of 1 indicate a difference of approximately 1 standard deviation between control and experimental groups). Thus, further experimentation with larger samples sizes and/or additional metrics to assess the vasculature could be helpful in detecting an effect of HC treatment at this stage if it does exist.

TABLE 2.

Tabulated results of descriptive statistics and statistical analyses performed on the vasculature data in control and HC‐treated eyes.

Measure	Group	Mean	SD	t‐test	p‐value	Hedges's g [95% CI]
Ciliary artery width (μm)	CNI HC	112.9 136.6	15.85 27.59	t(11.17) = −2.11	.058	−1.00 [−2.07, 0.08]
% Area (Region A)	CNI HC	69.4 64.3	4.90 5.18	t(14.00) = 2.04	.061	0.96 [−0.11, 2.03]
% Area (Region B)	CNI HC	41.4 41.1	8.39 19.63	t(9.47) = 0.04	.971	0.02 [−1.00, 1.03]
% Area (Region C)	CNI HC	65.7 61.0	7.55 9.51	t(13.32) = 1.11	.288	0.52 [−0.51, 1.55]

Note: In addition to the Welsh's t‐test and corresponding p‐values, Hedges's g provides an indication of effect size, corrected for small samples sizes.

Abbreviations: CNI, control no injection; HC, hydrocortisone; SD, standard deviation.

2.5. Molecular analyses of gene expression following HC treatment

To assess the potential effect of HC treatment on placode‐related gene expression, β‐catenin, Eda, and Fgf20 mRNA expression was analyzed using both qPCR and in situ hybridization. We assessed gene expression at HH32 when only half of the papillae ring has formed an active papillae induction is taking place, and again at HH34 when the full ring of conjunctival papillae has been induced.

2.5.1. β‐catenin remains ubiquitously expressed following HC treatment

At HH32, our qPCR data did not reveal any significant changes in β‐catenin expression within HC‐treated eyes compared to CNI and CI controls (F(2, 5.51) = 3.11, p = .125) (Figure 6A). In situ, hybridization staining in controls at this stage shows strong expression localized in the conjunctival papillae and faint global staining across the eye (Figure 6C, D). Cryosections show that this staining is localized to the epithelium of the papilla region and the superficial mesenchyme of the papilla and IP regions (Figure 6F, G). In HC‐treated eyes, there is a marked absence of localized β‐catenin expression in the papillae at HH32, congruent with a lack of papillae (Figure 6E). However, in these samples, a strong ubiquitous staining of β‐catenin is apparent throughout the eye (Figure 6E), which is specifically located within the superficial mesenchyme (and not in the epithelium) as shown in cryosections (Figure 6H, I). At HH34, no significant changes in β‐catenin expression levels were detected by qPCR between HC‐treated eyes compared to CNI and CI controls (F(2, 4.95) = 4.9, p = .088) (Figure 6B). In situ hybridization shows strong β‐catenin expression in all papillae regions of CNI and CI control eyes (Figure 6J, K) and diffuse expression across the eye in HC‐treated eyes (Figure 6L).

FIGURE 6.

β‐catenin gene expression data at stages HH32 and HH34. Normalized expression of β‐catenin was not significantly different in HC‐treated eyes compared to CNI and CI controls at HH32 (A) or HH34 (B). Each colored dot represents one biological replicate, black dots represent group averages, error bars represent standard error of the mean, asterisks denote statistical significance. Whole‐mount in situ hybridization expression data at HH32 (C)–(I) and HH34 (J)–(L). (C), (D) CNI (n = 6) and CI (n = 6) control eyes show localized β‐catenin expression in the conjunctival papillae and faint ubiquitous expression throughout the eye. (E) β‐catenin is ubiquitously expressed throughout HC‐treated eyes (n = 6). (F) Representative cryosection through a papilla in the temporal region of a CI control eye. The epithelium is stained in the papilla region only, while the superficial mesenchyme is stained in the papilla and interpapillary regions. (G) Enlarged region of (F). (H) Representative cryosection through the temporal region of a HC‐treated eye. The epithelium is unstained and the superficial mesenchyme is distinctly stained along the entire section (asterisk in I). (I) Enlarged region in H. (J)–(L) At HH34, β‐catenin expression is localized to the papillae regions in CNI and CI controls (n = 6) but is ubiquitously expressed across the sclera in HC‐treated eyes (n = 6). The beak is located to the right in all images. Scale bars in whole‐mount images represent 1 mm. Scale bars in cryosection images represent 100 μm. CNI, control no injection; CI, control injection; E, epithelium; HC, hydrocortisone; M, mesenchyme; NR, neural retina; RPE, retinal pigmented epithelium.

2.5.2. Eda expression is largely unaffected by HC treatment

At HH32, qPCR analyses showed that Eda expression was lower in both the CI control and HC treatment groups compared to the CNI control group (Figure 7A), however, only the CI group was significantly lower (2.8‐fold decrease; F(2, 5.40) = 12.8, p = .009; t(5.55) = 5.370, p = .005). In situ hybridization shows that Eda is similarly expressed in both CNI and CI controls, with strong expression across the eye as well as staining in the temporal and nasal papillae (Figure 7C, D). Cryosections show strong Eda expression across the epithelium, including the placode/papilla regions, with pale staining throughout the mesenchyme (Figure 7F, G). In HC‐treated eyes, strong ubiquitous Eda expression is present throughout the eye (Figure 7E). Similar to the staining in CI controls, this expression is localized throughout the epithelium with light staining within the mesenchyme (Figure 7H, I).

FIGURE 7.

Eda gene expression data at stages HH32 and HH34. Normalized expression of Eda was not significantly altered following HC treatment at HH32 (A) and was significantly higher compared only to uninjected (CNI) controls at HH34 (B). Each colored dot represents one biological replicate, black dots represent group averages, error bars represent standard error of the mean, asterisks denote statistical significance. Whole‐mount in situ hybridization expression data at HH32 (C)–(I) and HH34 (J)–(L). (C), (D) At HH32, there is strong Eda expression throughout the sclera of CNI (n = 4) and CI eyes (n = 4), as well as visible staining in the temporal and nasal papillae (n = 4). (E) Strong ubiquitous Eda expression is visible throughout the sclera of HC‐treated eyes. (F) Representative cryosection through a papilla in the temporal region of a CI control eye, showing strong Eda expression throughout the epithelium of the sclera region, including the placode/papilla regions. Light staining is also visible throughout the mesenchyme. G) Enlarged region of (F). (H) Representative cryosection through the temporal region of a HC‐treated eye, showing strong Eda expression throughout the epithelium of the sclera region and light staining throughout the mesenchyme. (I) Enlarged region of (H). (J), (K) At HH34, Eda is moderately expressed throughout the sclera of CNI (n = 4) and CI (n = 4) eyes, and visible within the temporal, nasal, and dorsal papillae. (L)There is moderate ubiquitous Eda expression throughout the sclera of HC‐treated eyes (n = 4). The beak is located to the right in all images. Scale bars in whole‐mount images represent 1 mm. Scale bars in cryosection images represent 100 μm. CI, control injection; CNI, control no injection; E, epithelium; HC, hydrocortisone; M, mesenchyme.

At HH34, qPCR results showed higher Eda expression in both the CI control and HC treatment groups compared to the CNI control group (Figure 7B), however, the only significant difference occurred between the HC treatment and CNI control groups (2.6‐fold increase; F(2, 5.45) = 9.15, p = .0.18); t(5.91) = −4.518, (p = .0.10). In situ hybridization analyses at this stage show moderate Eda expression throughout the sclera in all eyes as well as staining in the temporal, nasal, and dorsal papillae of CNI and CI eyes (Figure 7J–L).

2.5.3. HC treatment results in decreased Fgf20 expression at HH32

At HH32, qPCR gene expression analyses showed that Fgf20 expression is significantly lower following HC treatment compared to both the CI and CNI control groups (Figure 8A; one‐way ANOVA (F(2, 5.42) = 18.5, p = .004)). The differences in expression between the HC treatment group and both the CNI (6.0‐fold decrease, t(4.22) = 5.760, p = .008) and CI (2.9 fold decrease, t(4.92) =4.070, p = .023) control groups were striking. In situ hybridization analyses at HH32 show some moderate Fgf20 staining within the temporal and nasal papillae, and faint staining in the placodes of later papillae that have yet to fully form in the CNI and CI eyes (Figure 8C, D). Cryosections show that this Fgf20 expression is localized to the epithelium, basement membrane, and the innermost contiguous region of the papilla (region immediately adjacent to the papilla), as well as very light staining throughout the rest of the epithelium (Figure 8F, G). In HC‐treated eyes, there is pale ubiquitous FGF20 staining across the eye (Figure 8E) which is localized throughout the epithelium of the eye and possibly the deep mesenchyme (Figure 8H, I).

FIGURE 8.

Fgf20 gene expression data at stages HH32 and HH34. Normalized expression of Fgf20 was significantly reduced in HC‐treated eyes compared to both CNI and CI controls at HH32 (A). A similar trend is present but not statistically present and HH34 (B). Each colored dot represents one biological replicate, black dots represent group averages, error bars represent standard error of the mean, and asterisks denote statistical significance. Whole‐mount in situ hybridization expression data at HH32 (C)–(I) and HH34 (J)–(L). (C), (D) At HH32, there is light Fgf20 expression throughout the sclera of CNI (n = 4) and CI (n = 6) eyes, and moderate staining in the temporal and nasal papillae. (E) In HC‐treated eyes, there is pale ubiquitous Fgf20 staining across the sclera (n = 6). F) Representative cryosection through a papilla in the temporal region of a CI control eye, showing Fgf20 expression in the epithelium (specifically the upper part of the papilla), in the innermost parts of the contiguous region of the papilla, and very light staining throughout the rest of the epithelium. The basement membrane in the region of the papilla appears to also be stained, and the deepest layer of the mesenchyme could have faint background staining (asterisk in G). (G) Enlarged region of (F). (H) Representative cryosection through the temporal region of an HC‐treated eye, showing very faint Fgf20 expression throughout the epithelium of the sclera and potential background staining in the deep mesenchyme. (I) Enlarged region of (H). (J), (K) At HH34, there is faint Fgf20 expression throughout the sclera of CNI (n = 4) and CI (n = 4) eyes, as well as staining in the temporal, dorsal, and nasal papillae. (L) There is pale ubiquitous Fgf20 staining across the sclera of HC‐treated eyes (n = 4). The beak is located to the right in all images. Scale bars in whole‐mount images represent 1 mm. Scale bars in cryosection images represent 100 μm. CI, control injection; CNI, control no injection; E, epithelium; HC, hydrocortisone; M, mesenchyme.

At HH34, qPCR analyses showed lower Fgf20 expression in both the CI control and HC treatment groups compared to the CNI control group, however, these were not statistically significant differences (F(2, 4.81) = 4.9, p = .069; Figure 8B). In situ hybridization shows faint Fgf20 expression throughout all eyes as well as staining in the temporal, dorsal, and nasal papillae of CNI and CI eyes (Figure 8J–L).

3. DISCUSSION

3.1. Apoptosis and vasculature during normal conjunctival papillae development and after HC treatment

Apoptosis has been detected during the normal development of some cranial placodes, such as the lens, ⁴⁹ , ⁵⁰ olfactory, ⁵¹ and trigeminal placodes, ⁵² as well as in the late stages of conjunctival papillae development, ¹¹ specifically between stages HH34.5 and HH37. Congruent with these results, we show the presence of TUNEL‐positive apoptotic cells at stage HH34 within the first papilla (above the CA) of control eye tissue. This papilla first appeared at HH30 and is at an advanced stage of development by HH34; it begins to degenerate beginning at HH36. ¹¹ Apoptosis thus appears to be a normal part of papillae maintenance.

Vasculature has been mapped in the chicken eye from stages HH28–HH36, ⁵³ and later from HH35–HH38. ⁵⁴ These studies showed the development of the vascular network at HH34, after conjunctival papillae development. However, these studies used non‐specific dyes and standard fluorescent microscopy to label the blood vessels. Our analyses using confocal microscopy, coupled with lectin‐specific binding of endothelial cells, provide a higher resolution of the complex vascular plexus developing on the edge of the eye at HH34 and confirm that this plexus develops some distance from the complete ring of conjunctival papillae. These data re‐confirm that this developing vasculature is unlikely to direct the placement and development of the conjunctival placodes.

The combined effects of two HC injections span the duration of conjunctival papillae formation (i.e., HH30–HH34) allowing us to analyze the cumulative effects of HC treatment on apoptosis and vascular development at HH34 for the first time. Our TUNEL analyses did not reveal any differences in the amount of apoptotic cell death occurring shortly (6 h) after the first HC injection, nor a full day later (24 h) in the region where the first conjunctival papillae is being induced (Supplementary Figure 4A–C). Later at HH34, contrary to the expectation that increased apoptotic cell death could be responsible for the lack of conjunctival papillae development following HC treatment, less apoptotic cell death was detected in HC‐treated eyes compared to the region where the first papilla develops in controls (Supplementary Figure 4D, F). However, the presence of apoptosis is inherent in fully developed papillae structures, ¹¹ and our analyses between regions with and without a papilla in controls show a higher rate of apoptosis in the CA region compared to IP region of CNI controls (Supplementary Figure 4G, I). Therefore, the finding of less apoptotic death in HC‐treated eyes is significantly influenced by the lack of a papilla structure. Based on the comparison of HC‐treated eyes to the IP regions of CNI and CI controls (Supplementary Figure 4E, F), we conclude that there is no change in rates of apoptotic cell death following HC treatment. Thus, apoptotic cell death is not the mechanism through which HC treatment prevents conjunctival papillae development.

Similarly, we did not find conclusive evidence to support an effect of HC on the developing vasculature at HH34 using measurements of CA width and percent area of vasculature. When taking into consideration the large effect sizes that were calculated for both CA width and percent area in Region A (Table 2), it is possible that a larger sample size could be beneficial in detecting an effect of HC treatment on these metrics if one exists. Additional aspects of the developing vasculature (such as number of branching points, and VEGF expression) could also be assessed to verify that no effect on vasculogenesis is present, along with comparisons to injected controls. However, given our confirmation that the vascular plexus is not overlapping with the region where the ring of conjunctival papillae are being induced, we suspect it is unlikely to be a factor contributing to their inhibition following HC treatment.

3.2. HC treatment as a tool to study conjunctival placode induction

Here, we show that two HC injections administered at 30 and 2 h before HH30 is a very effective method to inhibit conjunctival placode development. This methodology has improved upon the outcome of Hammer and Franz‐Odendaal, ⁴² who showed complete inhibition in only 11.1% of eyes (n = 18 embryos) compared to 100% of eyes across a larger sample size (n >30 embryos) in this study. Further, we did not observe any raised epithelial areas indicative of placodes at HH32 following HC treatment (normal eyes would have 6–8 papillae at this stage), ⁹ suggesting that HC treatment prevented conjunctival placode induction, as opposed to maintenance.

In the morphogenesis of specialized epidermal structures, pre‐patterning establishes epithelial competency, followed by placode induction and subsequent development into papillae or bud structures. ² , ⁵⁵ Based on our results, we conclude that the pre‐patterning period for conjunctival papillae likely begins as early as HH27.5. Our finding that an injection at approximately HH27.5 (E5.5), a full day before the first papilla is distinct (HH30, E6.5), is required to prevent its development is not unexpected; it has previously been shown that neither HC injections at HH29 ⁴² nor surgical ablation at HH29 ¹⁶ could inhibit the earliest conjunctival papillae from forming. Furthermore, injections of Ecto‐D2, an Eda blocking antibody, at E5.5 similarly show absence of all conjunctival papillae. ²⁶ We also show that the late‐forming papillae, those that appear from HH33–34 (E7.5–8) are likely induced at approximately HH30 (E6.5). The finding that induction of a papilla takes 1–1.5 days is consistent with the estimate provided by Jourdeuil and Franz‐Odendaal ¹⁶ based on the time required to regenerate a single conjunctival papilla following ablation. Why these papillae have such a lengthy induction period is not yet understood.

3.3. Gene expression patterning during conjunctival placode formation

Previous research has shown that β‐catenin is broadly expressed prior to the appearance of the conjunctival papillae and then localizes to the papillae. ⁴³ Specifically, it is strongly expressed at HH30 in the first papillae that develop (but not earlier), and its expression in the epithelium precedes the development of the placodes. This led to the hypothesis that this factor likely pre‐patterns the epithelium prior to conjunctival placode induction, similar to other placodal systems (e.g., mammary gland, taste bud, hair, feather, etc.). ²⁶ , ⁵⁶ , ⁵⁷ , ⁵⁸ In our study, we show that HC injections were needed as early as HH27.5, indicating that other (unknown) gene(s) function upstream of β‐catenin.

We further show that after HC treatment, qPCR expression of this gene is relatively stable across the three treatment groups (CI, CNI, and HC) at both stages (HH32 and HH34) despite the loss of the localized epithelial signaling in papillae following HC treatment. Thus, this gene is not likely part of the mechanism by which HC prevents conjunctival papillae formation. Our results also suggest that upstream signals that induce β‐catenin are present in the eye after HC treatment (because mesenchymal expression was still present), but that the signals involved in localizing it into discrete placodes or to the epithelial tissue were affected by the treatment. Previous research has shown that placode localization could include downstream epithelial‐mesenchymal signaling targets (e.g., Edar and Fgf20) and/or could be the result of changes in tissue density, as in feather placodes. ²⁶

β‐catenin functions together with Edar to pre‐pattern the epithelium for feather placode formation and both factors are globally expressed across the skin initially. ²⁶ , ⁵⁹ , ⁶⁰ Furthermore, Ho et al. showed that inhibition of the Eda/Edar pathway leads to widespread absence of the conjunctival papillae, demonstrating that this pathway is important in this system. ²⁶ In addition to demonstrating the expression patterns of Eda within the developing conjunctival papillae for the first time, our data show that Eda expression levels are not affected above and beyond any effect of injection following HC treatment at HH32 or HH34 and that Eda remains ubiquitously expressed across the epithelium (Figure 7). Thus, Eda expression does not appear to be directly involved in the mechanism by which HC prevents these placodes from forming. Furthermore, based on our analysis at HH32 (when half the conjunctival placodes are present and half are still being induced) we did not observe a wave of Eda expression in control eyes, as observed in the feather placode system. ²⁶ Given the recent finding by Ho et al. that blocking Eda/Edar signaling activity leads to a similar loss of conjunctival papillae, ²⁶ investigating the expression of this pathway, as well as the upstream and downstream signaling interactions with the Wnt/β‐catenin pathway would provide a more comprehensive picture of the pre‐patterning and induction phases of conjunctival papillae development.

This study demonstrated that the spatial expression pattern of Fgf20 in control eyes overlaps with the expression pattern of its corresponding receptors, Fgfr1c and Fgfr2c. ⁴⁴ Loss of Fgf20 expression leads to the absence of conjunctival placodes in the sc/sc chicken, ⁴⁷ therefore it is possible that the mechanism of HC‐induced conjunctival placode loss could involve diminished Fgf20 expression. Our qPCR expression analyses revealed that Fgf20 expression is reduced in HC‐treated eyes compared to controls at HH32 (Figure 8); however, it is unclear if this reduction is the result of the loss of expression from conjunctival placode regions that fail to form or from a more direct effect of HC on Fgf20.

Beyond the description of gene expression presented in this paper, further experimentation is needed to reveal the role of these signaling molecules and their respective pathways during the induction of the conjunctival papillae. It is likely that β‐catenin and Eda expression induce Fgf20 expression in conjunctival placodes, as they do in feather placodes. ²⁶ , ⁶¹ , ⁶² It is also possible that additional members of the Wnt/β‐catenin and Eda/Edar pathways could have been affected by HC treatment, such as Lef1 or Edar (reviewed in References [63, 64, 65]). Alternatively, Fgf20 could play a different role in conjunctival placodes, such as regulation of placode size or boundaries.

Houghton et al. suggested that impaired Edar signaling could be responsible for failure to stabilize feather placode fate in sc/sc skin, ⁶⁶ which lacks functional FGF20. Here, we show that β‐catenin and Eda remain ubiquitously expressed following HC treatment and never become localized within the epithelium. If Edar expression is also disrupted following HC treatment, this could explain why normal levels of Eda expression do not result in activation of this pathway and the subsequent induction of Fgf20 expression needed for placode localization. Information regarding Edar expression in HC‐treated eyes could be a valuable addition to our understanding of how conjunctival papillae are inhibited after HC treatment.

Alternatively, there could be effects of HC treatment on tissue morphogenesis, such as changes in cell density or changes in the ECM environment. Ho et al. showed that cell density gradients and mesenchymal cell aggregation are both critical for feather placode formation. ²⁶ Fgf20 is initially induced in the epithelium by Eda and β‐catenin, and then sustained by mechanochemical stimulation via underlying cell aggregations. In the conjunctival placode system, we also see localized Fgf20 expression in the epithelium of the conjunctival placode/papilla region and expression of its corresponding receptors in the superficial mesenchyme below the developing conjunctival placode. Although there is evidence of densely arranged mesenchymal cells directly below the conjunctival papillae and above the region of scleral ossicle induction, ¹¹ , ⁶⁷ , ⁶⁸ these regions are only present after placode formation has taken place (M stages 2–5) ⁶⁷ and do not persist as cell aggregations. With respect to the ECM, Stuart et al. ³⁴ and Hammer and Franz‐Odendaal ⁴² both reported observations of a more densely arranged ECM following HC treatment, suggesting that inductive signals may be unable to diffuse through the ectomesenchyme. Further attention directed towards changes in the ECM following HC treatment could help clarify our understanding of how HC inhibits the formation of these placodes. Using HC treatment to prevent conjunctival papillae development provides a convenient framework in which to investigate the importance of the ECM during placode formation.

4. CONCLUSIONS

A refined HC injection protocol resulting in loss of the entire conjunctival papillae ring has provided a means to investigate the processes important for conjunctival papillae induction in the chicken eye. In a normal developing chicken eye, our results confirm that apoptosis is present at later stages of conjunctival papillae development, and the developing vascular plexus is not associated with the region where conjunctival papillae induction takes place. In the HC phenotype, neither apoptosis nor vasculature appears to be responsible for conjunctival placode inhibition. The developing vascular plexus could be investigated further to better understand how vasculogenesis at HH34 is affected by HC treatment. This study also showed that Eda is globally expressed across the epithelium and Fgf20 expression is localized to the epithelium within placode/papillae regions. Following HC treatment, β‐catenin and Eda remain expressed in the mesenchyme and conjunctival epithelium, respectively. Furthermore, we show that HC‐treated eyes have decreased Fgf20 expression. Our finding that conjunctival placode induction begins as early as HH27.5, 1.5 days before the first papillae form and before the first detection of β‐catenin in the anterior eye, suggests that interactions with other gene pathways should be investigated at earlier stages in order to expand our understanding of conjunctival placode induction. The effect of HC treatment on the ECM and Edar gene expression, as well as downstream targets of these pathways, remain to be explored as potential mechanisms through which conjunctival placode development is prevented.

5. EXPERIMENTAL PROCEDURES

5.1. Chicken embryo incubation

Fertilized chicken eggs were obtained from Atlantic Poultry Inc. (New Minas, N.S.) and incubated at 37°C ± 1°C and 40% relative humidity in an Ova‐Easy 380 Advance Series II Cabinet Incubator (Brinsea Products Inc.) until the desired stages of interest (i.e., between HH32 and HH34).

5.2. HC injection experiments

To optimize the HC treatment such that it results in complete conjunctival papillae inhibition, a series of 12 HC injection experiments were conducted. Each experiment consisted of three treatment groups: control no injection (CNI), control injection (CI), and HC injection. Injected eggs received either 100 μL of 2 mM HC (Sigma‐Aldrich) dissolved in 3% ethanol in Howard's Ringer solution (HC group) or 100 μL of the solvent alone (CI group). Injections were delivered through a small hole made in the shell at one end of the egg, as in Hammer and Franz‐Odendaal. ⁴² Adjustments to HC injection timings were made in 2‐h intervals. The timeframe of injections spanned from HH27.5 to HH30—HC treatment thus starts at a period prior to when the first papillae appear at HH30 and endures throughout papillae formation from HH30–HH34.

Following injections, embryos were incubated to the desired stage for the downstream application and fixed as described below. For the optimization experiments, embryos were incubated to HH34‐35 (stages when all papillae are distinct), decapitated, and fixed overnight in 10% neutral buffered formalin (Fisher Scientific) followed by storage in 1X PBS. In each of the 12 experiments, the number of conjunctival papillae was compared between control (CNI and CI) and HC‐treated samples (CNI: n = 21; CI: n = 71; HC: n = 75). Once the optimal HC treatment was determined, it was used for all subsequent downstream experiments.

5.3. TUNEL apoptotic cell death detection

Embryos were fixed overnight in 4% paraformaldehyde (Sigma‐Aldrich) at 4°C, rinsed and stored in 1X PBS, and then dissected. The neural and pigmented retinas, vitreous humor, lens, nictitating membrane, and eyelids were removed. Right eyes were used for experimental samples and left eyes were used for negative and positive controls. The eye tissue from untreated (CNI), control‐injected (CI), and HC‐treated embryos were analyzed for apoptotic cell death at three time points by fluorescently labeling DNA degradation via the TUNEL reaction (TdT‐mediated dUTP‐X nick end labeling kit; Roche). CNI control (n = 4), CI control (n = 4), and HC‐treated (n = 6) eyes were fixed at 6 and 24 h following the first HC injection time point, at approximately HH28 and HH30. The 6‐h time point addresses the immediate effects of HC treatment and the 24‐h time point assesses whether any effect of HC is sustained throughout the developmental period of the first papilla before it becomes visible at HH30. HH34 was also selected for analysis because at this stage the conjunctival papillae ring is completely formed in untreated samples, and the cumulative effects of both HC injections are apparent (CNI n = 6; CI n = 6; HC n = 6). Positive and negative control tissues were both treated with DNase I (New England Biolabs), however only the positive control tissues were exposed to the TUNEL enzyme. Vectashield with DAPI (Vector Laboratories) was applied to stain nuclei. Tissues were imaged as a z‐stack in whole‐mount using a Zeiss LSM‐900 confocal microscope and compiled as a maximum intensity projection for analysis.

For quantitative analyses at all three time points, the papilla above the CA was imaged within each piece of tissue. Because HC‐treated tissue at HH34 lacks papillae, control tissues collected at this time point were also imaged in the adjacent epithelium (IP region). Within each imaged region, the total number of DAPI‐stained cells (blue) and TUNEL‐positive cells (green) were counted and used to calculate the ratio of apoptotic: DAPI‐stained cells (apoptotic index). Total DAPI‐positive cells were estimated in Fiji using a combination of the Isodata threshold algorithm and binary processing options. TUNEL‐positive cells were counted manually using the multipoint function in Fiji ⁶⁹ based on the characterizations of apoptotic cells outlined by Moore et al. ⁷⁰

Comparisons of apoptotic indices were performed across treatment groups and tested for statistical significance in Jamovi ⁷¹ using the GAMLj package. ⁷² For all three time points, ANOVA models with planned Helmert contrasts were used to test for effects of group (HC vs. CI vs. CNI) on apoptotic cell death. For HH34 analyses, the effect of eye region (CA vs. IP) was also tested, and linear mixed models fitted with random intercepts were used to assess the effect of eye region (CA vs. IP) within each control group (CNI and CI).

5.4. Vasculature analysis with lectin injections

Embryos were transferred to ex‐ovo cultures at E3 (HH 19–20) as described by Cloney and Franz‐Odendaal ⁷³ ; this method provides easier access to the vasculature and was shown not to alter the patterning or timing of papillae development. Based on previous lectin injection experiments in chicken embryos, ⁷⁴ lens culinaris agglutinin (LCA) lectin (Vector Laboratories) was diluted to 0.42 μg/μL with sterile 1X PBS and filtered through a 0.2 μm syringe filter (VWR) prior to injection. At HH34, CNI (n = 8) and HC‐treated (n = 8) embryos were injected with 7.5 μL of LCA lectin warmed to 37°C. To circulate the lectin throughout the vasculature, injected embryos were incubated for 13 min followed by fixation in 4% paraformaldehyde overnight at 4°C. Fronts of eyes were dissected in 1X PBS and the vitreous humor, lens, nictitating membrane, and eyelids were removed. The samples were mounted in glycerol in 1X PBS (9:1) with 1 M propyl gallate. Images of the vasculature in CNI (n = 8) and HC‐treated (n = 8) eyes was captured as a z‐stack in a 4 × 4 tile arrangement using a Zeiss LSM‐710 confocal microscope and compiled as maximum intensity projections.

Using Fiji software, ⁶⁹ the vasculature was characterized and compared between control and HC‐treated tissues. We measured the CA width before its main branch point and the percentage of area covered by the vascular meshwork. The percent area of vascular meshwork was measured in three different regions of the eye and calculated using a Fiji script developed by Rust et al. ⁷⁵ Briefly, because the structure of the vasculature was not uniform across the field of view, the percentage of area containing vasculature was measured in three different regions of the eye. To standardize these three regions across each eye, boxes 600 × 600 pixels in size were placed at the end of 2 mm lines extending from the main branch point of the CA (Figure 5A and C). The first line extended towards the edge of the eye to a point along the CA that is 2 mm away from the branch point. A second and third line was drawn 25° above and below the first line, leading to two additional boxed regions. Using the script developed by Rust et al., ⁷⁵ each image was screened based on a brightness threshold, producing a black‐and‐white mask that was then used to estimate the vasculature coverage and calculate the percentage of area taken up by the vascular meshwork. No outliers were found when data was subjected to a Grubbs's outlier test at the 1% significance level for eight samples. ⁷⁶ CA width and the percent area of vasculature meshwork in control and HC‐treated tissues at HH34 were plotted with means and 95% confidence intervals, tested for statistical significance using a Welsh's t‐test, and the effect size expressed as Hedges's g values.

5.5. Quantitative PCR gene expression analysis

Tissue from all three treatment groups (CNI, CI, and HC) was collected at HH32 (the midpoint of papillae formation) and HH34 (the endpoint of papillae formation) for quantitative PCR (qPCR) analysis. The dissected anterior eyes (with vitreous humor, lens, nictitating membrane, and eyelids removed) were preserved in 200 μL RNAlater (Qiagen) and stored at—80°C. CNI and CI eye tissue included between 6 and 8 papillae at stage HH32 and 13–14 papillae at HH34, while HC‐treated eyes had no papillae at either stage. Each biological replicate contained eye tissue from 3 to 4 different embryos. Total RNA was isolated for four biological replicates of each treatment group using Qiazol Lysis Reagent (Qiagen) and RNeasy Plus Universal kit (Qiagen) and quantified using a BioDrop spectrophotometer. A total of 1 μg of total RNA was reverse transcribed into cDNA using iScript Reverse Transcription Supermix (Bio‐Rad Laboratories) in an Eppendorf Mastercycler.

Amplification of reference (Rpl19 and Hprt1) and target (β‐catenin, Eda, and Fgf20) genes was performed in triplicate using SYBR Green (Bio‐Rad Laboratories) in a CFX384 Touch Real‐Time PCR Detection System (Bio‐Rad Laboratories) using standard amplification parameters with a 60°C annealing temperature. Primer specificity was confirmed using gel electrophoresis and confirmation of one melt curve peak for all samples. Primer sequences for β‐catenin, Eda, Fgf20, and reference genes (Hprt1 and Rpl19) are from Giffin and Franz‐Odendaal. ⁴⁵ Bio‐Rad CFX Maestro software was used to measure gene amplification (Cq values) and calculate normalized expression (ΔΔCt method) of target genes with respect to the two validated reference genes (Rpl19 and Hprt1). When Cq values for each gene were subjected to a Grubbs's outlier test at the 1% significance level for 12 samples, one outlier was removed from the FGF20 data set at HH32. ⁷⁶ Normalized expression values were analyzed in Jamovi ⁷¹ using a one‐way ANOVA and Games‐Howell post‐hoc tests.

5.6. Whole‐mount in situ hybridization

HC‐treated and control (CNI and CI) eyes were collected at HH32 and HH34. The tissue was fixed overnight in 4% paraformaldehyde at 4°C, rinsed in 1X PBST, dissected to isolate the anterior eye, then dehydrated and stored in methanol at −20°C until further use. Antisense RNA probes for β‐catenin were generated from plasmids kindly donated by Dr. Randall Widelitz (University of Southern California, USA). ⁷⁷ Antisense probes for Eda and Fgf20 were generated from plasmids generously provided by Dr. Denis Headon (University of Edinburgh, UK). ²⁶ β‐catenin, Eda, and Fgf20 mRNA probes were created using a DIG‐RNA labeling kit (Roche) and SP6/T7/T3 restriction enzymes, respectively. The in situ hybridization protocol was carried out according to Jourdeuil & Franz‐Odendaal. ⁴³ Each group (CNI, CI, and HC) contained four to six eye tissue samples from different embryos.

To assess the distribution of β‐catenin, Eda, and Fgf20 staining specifically within the epithelium and mesenchymal layers of control and HC‐treated eyes at stage HH32, the temporal region of CI and HC‐treated eyes following in situ hybridization was cryosectioned and observed with a compound microscope. The dissected tissues were embedded in 1% agar and placed in a 30% sucrose solution overnight at 4°C. Samples were mounted the following day on cold chucks using Frozen Sectioning Compound (VWR) and frozen solid for 10 min. The tissue was sectioned at 16–20 μm thickness at −20°C and placed on Superfrost Plus microscope slides (Fisher Scientific). Slides were stored at −80°C until the addition of mounting media (9:1 glycerol in 1X PBS with 1 M propyl gallate) and coverslips (Globe Scientific) in preparation for imaging using NIS‐Elements software. ⁷⁸

Supporting information

Data S1. Supporting Information.

Drake PM, Franz‐Odendaal TA. Hydrocortisone treatment as a tool to study conjunctival placode induction. Developmental Dynamics. 2025;254(1):74‐93. doi: 10.1002/dvdy.729