Chicken embryos share mammalian patterns of apoptosis in the posterior placodal area

Abstract

In the posterior placodal area (PPA) of C57BL/6N mice and primate‐related Tupaia belangeri (Scandentia), apoptosis helps to establish morphologically separated otic and epibranchial placodes. Here, we demonstrate that basically identical patterns of apoptosis pass rostrocaudally through the Pax2⁺ PPA of chicken embryos. Interplacodal apoptosis eliminates unneeded cells either between the otic anlage and the epibranchial placodes 1, 2 and/or 3, respectively (type A), or between neighbouring epibranchial placodes (type B). These observations support the idea that in chicken embryos, as in mammals, interplacodal apoptosis serves to remove vestigial lateral line placodes (Washausen & Knabe, 2018, Biol Open 7, bio031815). A special case represents the recently discovered Pax2⁻/Sox2⁺ paratympanic organ (PTO) placode that has been postulated to be molecularly distinct from and developmentally independent of the ventrally adjacent first epibranchial (or ‘geniculate’) placode (O'Neill et al. 2012, Nat Commun 3, 1041). We show that Sox2⁺ (PTO placodal) cells seem to segregate from the Pax2⁺ geniculate placode, and that absence of Pax2 in the mature PTO placode is due to secondary loss. We further report that, between Hamburger–Hamilton (HH) stages HH14 and HH26, apoptosis in the combined anlage of the first epibranchial and PTO placodes is almost exclusively found within and/or immediately adjacent to the dorsally located PTO placode. Hence, apoptosis appears to support decision‐making processes among precursor cells of the early developing PTO placode and, later, regression of the epibranchial placodes 2 and 3.

Introduction

Placodes are patch‐like thickenings in the head surface ectoderm of vertebrate embryos. Except for the lens and adenohypophysial placodes, they produce sensory cells and/or ganglionic neuroblasts. All placodes originate from the horseshoe‐shaped panplacodal primordium that surrounds the front end of the neural plate. Initially, this primordium consists of a continuous layer of thickened epithelium. Later, physically separated placodes stand out from the thinned remains of the primordium. Correspondingly, general placode markers, among others members of the Sine oculis homeobox (Six) family of transcription factors, gradually disappear from thinning parts of the panplacodal primordium but continue to be expressed in the definitive placodes (for reviews see Baker & Bronner‐Fraser, 2001; Schlosser, 2006; Park & Saint‐Jeannet, 2010).

In Xenopus laevis, zebrafish and chicken, precursor cells destined to populate different placodes are first available in mixed groups (Streit, 2002; Bhattacharyya et al. 2004; Dutta et al. 2005; Xu et al. 2008; Pieper et al. 2011). How exactly ‘rearrangement’ of these precursor cells and formation of sharp interplacodal boundaries happen is unclear in many aspects. Several scenarios have been proposed, namely: (1) changes in the expression patterns of transcription factors of (immobile) precursor cells, (2) ‘differential growth and stretching of the ectodermal layer’ or (3) changes in the positions of neighbouring precursor cells (Streit, 2002; Bhattacharyya et al. 2004; Bhat & Riley, 2011; Pieper et al. 2011; for a review see Schlosser, 2010, p. 172).

Morphogenesis of physically separated placodes also appears to depend on the occurrence of spatiotemporally regulated apoptosis. So far this phenomenon has been best studied in the posterior placodal area (PPA) of C57BL/6N mice and primate‐related Tupaia belangeri (Scandentia). In amniotes, this posterior subsection of the panplacodal primordium gives rise to the otic and epibranchial placodes, the latter contributing neurons to ganglia of the cranial nerves VII (geniculate ganglion), IX (petrosal ganglion), and X (nodose ganglion). Large‐scale apoptosis starts interplacodally between the otic and epibranchial placodes. Later, apoptosis is centred on the mature and regressing epibranchial placodes (Washausen et al. 2005; Washausen & Knabe, 2013, 2017, 2018). In T. belangeri, similar patterns of apoptosis accompany the morphogenesis of the trigeminal and lens placodes (Knabe et al. 2009).

‘Interplacodal’ apoptosis in the PPA of embryonic mice predominantly eliminates Six1⁺ placodal precursor cells (Washausen & Knabe, 2017). To better understand why this is the case, we have examined how pharmacological inhibition of apoptosis‐associated caspases affects morphogenesis of the posterior placodes. Surprisingly, it turned out that ‘those parts of the PPA which, under experimental conditions, escape apoptosis have retained the developmental potential to produce lateral line placodes and the primordia of neuromasts’ (Washausen & Knabe, 2018, p. 1). Hence, the evolutionary theory that the lateral line system which, in fish and aquatic amphibia, is responsible for the detection of movements, pressure changes or electric fields in the surrounding water (Northcutt, 1992; Schlosser, 2002, 2006; Ghysen & Dambly‐Chaudière, 2004) was completely lost in amniotes (mammals, birds, reptiles) is no longer defensible (Washausen & Knabe, 2018).

In chicken embryos, the PPA is also referred to as the otic‐epibranchial progenitor domain (OEPD; Freter et al. 2008), which expresses the Paired box transcription factor 2 (Pax2; Chen & Streit, 2013). As virtually nothing is known about whether apoptosis contributes to the subdivision of the chicken PPA into individual placodes, we decided to examine this issue in detail. We further aimed to investigate whether apoptosis might be involved in the development of the recently discovered PTO placode, which later forms the mechanosensory paratympanic organ. Using the transcription factor Sox2 (Sex determining region Y‐box 2) as a further marker, it was found that this novel Pax2⁻/Sox2⁺ placode resides dorsally adjacent to the Pax2⁺/Sox2⁻ ‘geniculate’, or first epibranchial placode (O'Neill et al. 2012). Hence, whether the PTO placode also is derived from the Pax2⁺ PPA is at present unclear (Chen & Streit, 2013) and this, therefore, has been addressed here as well.

Materials and methods

Embryos

Chicken embryos (Gallus domesticus) were collected in accordance with German animal care guidelines and in agreement with the responsible animal welfare officer of the University Hospital Münster, Germany. Fertilized Brown Leghorn eggs, purchased from a local hatchery, were incubated at 38.3 °C in a temperature‐controlled breeder (BRUJA 400‐D‐H, Brutmaschinen‐Janeschitz, Hammelburg, Germany) to Hamburger–Hamilton (HH; Hamburger & Hamilton, 1951) stages HH8 to HH26 (n = at least 3 embryos per stage). From day 3 of incubation onwards, the eggs were turned (three times per day) and allowed to cool (10 min per day). Eggs were opened using the method described by Korn & Cramer (2007). In brief, the egg was placed horizontally on a mould of modelling clay. Stripes of plastic tape (764i, 3M, Neuss, Germany) were stuck to the blunt end and onto the top of the egg, respectively. To lower the level of the embryo, 3 mL of albumen were withdrawn by inserting an 18‐gauge needle with an attached 10‐mL syringe at the blunt end. Thereafter, a window was cut on the top of the egg using a pair of curved spring scissors (91461‐11, Fine Science Tools, Heidelberg, Germany). With the help of a stereomicroscope (S8 APO, Leica, Wetzlar, Germany), the extra‐embryonic membranes were cut at a safe distance circularly around the embryo using a pair of fine Bonn scissors (14084‐08, Fine Science Tools). Depending on its size, the embryo was either aspirated with a transfer pipette or taken up with a perforated spoon (10370‐18, Fine Science Tools). Embryos were briefly rinsed in phosphate‐buffered saline (PBS: 137 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, 2 mm KH₂PO₄, pH 7.4) and immediately transferred to 4% paraformaldehyde in PBS. Depending on the developmental stage, immersion fixation was performed for 24–48 h at room temperature (rt). Thereafter, specimens were washed in PBS for 30 min twice and dehydrated in a series of ascending ethanols. To avoid damage to delicate specimens during paraffin embedding, embryos belonging to stages HH8 to HH16 were pre‐embedded in 1% low gelling‐point agarose (Seakem LE, 50001, Lonza, Cologne, Germany) at the 50% ethanol level. Finally, dehydrated specimens were cleared in chloroform, embedded in Surgipath Formula ‘R’ paraffin (3801450, Leica), and serially sectioned at 5 μm with a rotary microtome (RM2245, RM2265, Leica). As serial sections were alternately mounted on two sets of slides (Knabe et al. 2002), consecutive sections were available for immunostaining experiments with different primary antibodies or routine staining with Mayer's haematoxylin (Romeis, 1948).

Antibodies

The rabbit anti‐cleaved caspase‐3 antibody (9661; Cell Signaling Technology, Frankfurt/Main, Germany; lot 37; RRID: AB_2341188), raised against residues adjacent to Asp175 of human caspase‐3 (CRGTELDCGIETD), specifically detects the large fragment of cleaved caspase‐3 but not the inactive full‐length form or other cleaved caspases (manufacturer's technical information). As demonstrated for gentamicin‐induced cell death in the avian cochlea (Kaiser et al. 2008), the antibody specifically detects cleaved caspase‐3‐labelled apoptotic cells in the chicken. The rabbit anti‐Pax2 antibody (71‐6000; Thermo Fisher Scientific, Schwerte, Germany; lot 1117672A; RRID: AB_2533990), raised against amino acids 188–385 of mouse Pax2 (Dressler & Douglass, 1992), recognizes major epitopes between residues 204 and 373 of murine Pax2 (manufacturer's technical information). The antibody has been widely used to label Pax2 in the PPA and its derivatives in mouse, zebrafish and chicken embryos (for references see Washausen & Knabe, 2017; O'Neill et al. 2012). The mouse anti‐Sox2 antibody (clone 9‐9‐3, ab79351; Abcam, Cambridge, UK; lot GR323583‐1; RRID:AB_10710406) was raised against a recombinant protein derived from within amino acids 300 to the C‐terminus of human Sox2 (manufacturer's technical information). The antibody recognizes the chicken PTO placode (O'Neill et al. 2012) as well as Sox2⁺ stem cells in the mouse olfactory epithelium (Engel et al. 2016) and in the chicken inner ear (Mulvaney et al. 2015). For immunohistochemical experiments, biotinylated goat secondary antibodies from Vector Laboratories (Burlingame, CA, USA) were applied (anti‐rabbit, BA‐1000, RRID:AB_2313606; anti‐mouse, BA‐9200, RRID:AB_2336171). For immunofluorescence, donkey secondary antibodies from Jackson Immunoresearch Europe (Cambridgeshire, UK) were used (Cy3‐conjugated anti‐mouse, C715‐165‐150, RRID:AB_2340813; biotinylated anti‐rabbit, 711‐065‐152, RRID:AB_2340593).

Immunohistochemistry and immunofluorescence

Sections were deparaffinized in xylene, rehydrated in a series of descending ethanol concentrations and washed in Tris‐buffered saline (TBS: 0.05 m Tris, 0.15 m NaCl, pH 7.4). For antigen unmasking, slides were boiled in citrate buffer (0.01 m, pH 6) in a high‐pressure cooker (Moulinex CE4000; Krups, Offenbach/Main, Germany). Next, sections were allowed to cool to rt, rinsed in distilled water, and permeabilized in 0.3% Triton X‐100 in TBS for 30 min. For immunohistochemistry, this solution additionally contained 1% H₂O₂ to block possible endogenous peroxidases. Thereafter (and generally following further incubation steps), slides were washed in TBS (three times 5 min). Incubation with the primary antibody was performed in background‐reducing Dako REAL antibody diluent (S202230‐2, Agilent Technologies, Waldbronn, Germany) at 37 °C for 1 h. For immunohistochemistry, primary antibodies were applied either at 1 : 1000 (anti‐cleaved caspase‐3, anti‐Pax2) or 1 : 4000 (anti‐Sox2). Following incubation steps (1 h at rt) with the appropriate biotinylated secondary antibody, diluted 1 : 400 in TBS with 2% normal goat serum (S‐1000, Vector Laboratories), and with the avidin‐biotin peroxidase complex (Elite ABC reagent, PK‐7100, Vector Laboratories), immunoreactions were developed with 0.06% 3,3’‐diaminobenzidine (DAB; D5637, Sigma‐Aldrich) and 0.007% H₂O₂ in Tris‐HCl buffer (0.1 m, pH 7.6). DAB‐reacted sections were counterstained with Mayer's haematoxylin and embedded with DePeX mounting medium (18243, Serva, Heidelberg, Germany).

For double immunofluorescence stainings, sections were first incubated (1 h, 37 °C) with the anti‐Sox2 antibody at a dilution of 1 : 25 and later with the Cy3‐conjugated anti‐mouse antibody (1 h, rt), diluted 1 : 100 in TBS with 2% normal donkey serum (017‐000‐001, Jackson Immunoresearch Europe). Thereafter, the second primary antibody (anti‐Pax2) was applied at 1 : 50 dilution (1 h, 37 °C). It was detected with the biotinylated donkey anti‐rabbit antibody (1 : 200 in TBS with 2% normal donkey serum; 1 h, rt) and later with Cy2‐conjugated streptavidin (1 : 100 in TBS; 1 h, rt; 016‐220‐084, Jackson Immunoresearch Europe; RRID:AB_2337246). Finally, sections were stained with 4’,6‐diamidine‐2’‐phenylindole dihydrochloride (DAPI, 1 μg mL^–1; 10236276001, Roche, Mannheim, Germany), dehydrated, cleared and embedded (DePeX).

In negative controls where the primary antibody was omitted, no immunoreactivity was detectable. To further verify the specificity of our immunohistochemical experiments, well established expression sites of Pax2 (optic vesicle, otic anlage, nephric duct, hindbrain interneurons: Hurtado & Mikawa, 2006) and Sox2 (olfactory placode, lens, otic anlage, pharyngeal pouches, neural stem cells of the developing central nervous system: Matsumata et al. 2005; O'Neill et al. 2012) were examined.

Reconstructions

Schematic reconstructions of the chicken PPA and its derivatives were performed as described in Washausen & Knabe (2013, 2017). Initially, frameworks of the reconstructions were established in coreldraw X4 (Corel, Unterschleißheim, Germany) by comparison with images of unsectioned embryos taken from the chicken embryo gene expression database Gallus Expression in Situ Hybridization Analysis (GEISHA; Bell et al. 2004; Darnell et al. 2007). First, the plane of sectioning was determined using relevant topographical landmarks (e.g. optic and otic anlagen). Secondly, the positions of the pharyngeal pouches and membranes as well as of the high‐grade thickened placodes were transferred from serial sections to the schematic maps. Further, pit‐like invaginations of the otic and epibranchial placodes, the detachment site of the otic vesicle and, later, the position of the internalised otic vesicle were included. For the diagnosis of apoptotic cells in tissue sections, at least two independent detection techniques have to be applied (Stadelmann & Lassmann, 2000; Taatjes et al. 2008). Consequently, we combined immunohistochemical experiments with antibodies against cleaved caspase‐3 and structural analysis (Sanders & Wride, 1995; Häcker, 2000). Our schematic reconstructions were based on every second 5‐μm‐thick section of completely sectioned embryos (interval = 10 μm, 60–150 analysed sections per embryo) and thus faithfully represent the patterns of apoptotic cells/bodies, Pax2⁺ nuclei and/or Sox2⁺ nuclei. Intensity of the immunostaining was semi‐quantitatively classified as absent, weak or strong.

Photomicrographs

Unless otherwise specified, all equipment used for photomicrography was obtained from Carl Zeiss (Göttingen, Germany). Micrographs of immunoreacted sections were acquired with an Axioskop 2 MOT microscope, an Axiocam HR digital camera, and the KS400 software (v3.0) that was also used to perform a shading correction of the images. Micrographs of double‐labelled immunofluorescent sections were captured with an Axiocam HR digital camera, the axiovision software (v4.7.1) and an Axioskop 2 microscope equipped with a mercury vapour lamp (HBO 100), as well as blue, green and red fluorescence filter sets (02, 09, 15, respectively). Finally, acquired images were cropped, resized and processed in photo‐paint X4 (Corel). All adjustments (brightness, colour balance and sharpness) were applied all over the whole images, and no specific details within them were modified, removed or inserted.

Results

According to Chen & Streit (2013), Pax2 mRNA expression is present in the PPA of the 5‐somite stage chicken embryo. Correspondingly, on HH8 (4‐somite stage), weakly Pax2⁺ cells are scattered in the developing PPA, which mostly consists of at least one row of columnar epithelium and which extends from the level of the Pax2⁺ isthmic organizer to the position of the first somite (Fig. 1A,A’). On HH9, Pax2 is expressed throughout almost all the PPA (Fig. 1B,B’). The Pax2⁺ otic placode consists of three rows of columnar epithelium and thus stands out from all other parts of the PPA (Fig. 1B). On HH10, strongly Pax2⁺ central parts of the thickened PPA are ensheathed by a narrow mantle of weakly Pax2⁺ placodal precursor cells (Fig. 1C,C’).

Figure 1.

Pax2 protein expression and apoptosis in the posterior placodal area (ppa) of chicken embryos, Hamburger–Hamilton (HH) stages HH8 to HH10. (A‐C) Schematic reconstructions [top views, not drawn to scale; orientation mark indicating medial (m), lateral (l), rostral (r) and caudal (c) positions; paired maps, lettered left and mirrored right body sides] demonstrate strong or weak Pax2 immunopositivity (magenta, dark or light shades, respectively), different grades of thickened surface ectoderm (specified below) and apoptotic cells (red with black margin). Neuroepithelium (grey), cranialmost somites (yellow) and caudal parts of the Pax2⁺ isthmic organizer (io) were also included. Sectioning planes indicate the positions of anti‐Pax2 (A’, B’, C’) or anti‐cleaved caspase‐3 (C’’) stained sections (A’: mirror‐imaged left body side). (A, A’) In HH8 chicken embryos, Pax2⁺ placodal precursor cells are scattered within or adjacent to the thickened ppa, which mostly consists of at least one row of columnar epithelium (circumscribed by thin black lines). (B‐C’’) From HH9 to HH10, Pax2 expression intensifies within the thickened ppa (thin black lines), including the otic placode (ot), which consists of at least three rows of columnar epithelium (circumscribed by thick black lines in B, C; thick dashed line in C’). From HH8 to HH10, apoptosis is virtually absent from the developing ppa (A, B, C, C’’). nf, neural fold; np, neural plate; nt, neural tube; s1/s2, first/second somite. Scale bar: 20 μm.

Between HH8 and HH10, apoptosis is virtually absent from the PPA of chicken embryos (Fig. 1A–C,C’’). However, on HH11, apoptosis centres on the ventral margin of the otic placode that demonstrates, at most, a shallow depression (Figs 2A and 3A). Apoptosis persists in this ventral position during progressive formation of the otic pit (HH12 to HH14) and later shifts to the detachment site of the otic vesicle (Figs 2B and 3B‐F). Formation of the otic pit is accompanied by the initial development of the epibranchial placodes 1 and 2 (Fig. 3B–D). HH14 also sees the anlage of the epibranchial placode 3, which is still in continuity with the developing epibranchial placode 2 (Fig. 3D). Epibranchial placode 3 becomes morphologically distinct on HH15 (Fig. 3E), and around HH17/HH18 extends caudally beyond the level of the pharyngeal pouch 4 (Figs 3F and 4A). Between HH19 and HH23, epibranchial placode 3 segregates into the epibranchial placodes 3₁ and 3₂ (Fig. 4B‐D).

Figure 2.

Apoptosis in the posterior placodal area of chicken embryos as revealed by anti‐cleaved caspase‐3 immunohistochemistry. (A) Hamburger–Hamilton (HH) stage HH11 shows apoptosis at the ventral margin of the otic placode (ot), enlarged in the inset (arrowheads). (B) HH17 demonstrates apoptosis (arrow) at the detachment site of the otic vesicle (ov), interplacodal apoptosis type A (black arrowheads) between the developing otic vesicle and the combined anlage of the first epibranchial and paratympanic organ (PTO) placodes (e1/pt), and apoptosis (white arrowheads) in dorsal parts of this anlage. Asterisk, geniculate ganglion; hb, hindbrain; p1, pharyngeal pouch 1. Scale bars: 20 μm.

Figure 3.

Apoptosis in the posterior placodal area of chicken embryos, Hamburger–Hamilton (HH) stages HH11 to HH17. Schematic reconstructions [lateral views, not drawn to scale; orientation mark indicating dorsal (d), ventral (v), rostral (r) and caudal (c) positions; paired maps, lettered left and mirrored right body sides] demonstrate high‐grade thickenings (grey) revealing the otic placode (ot), the combined anlage of the first epibranchial and paratympanic organ (PTO) placodes (e1/pt), and the epibranchial placodes 2 and 3 (e2, e3). Patterns of apoptosis (red), pharyngeal membranes/pouches (dotted black lines), borders of the branchial arches (thin black lines in B‐F) and cranialmost somites (yellow) are also shown. (A‐D) Between HH11 and HH14, apoptosis in the thickened posterior placodal area (at least one row of columnar epithelium, circumscribed by a thin black line in (A) is centred on the ventral margin of the invaginating otic placode (circumscribed by thick black line). Thereafter, apoptosis shifts to the detachment site (‘otic pore’, again circumscribed by a thick black line) of the developing otic vesicle (thin dashed black line in E, F). Interplacodal apoptosis type A between the otic anlage and the combined anlage of the first epibranchial and PTO placodes, and/or between the otic anlage and the epibranchial placodes 2 and 3 starts around HH14 (D) and peaks around HH17/HH18 (F). Apoptosis in dorsal parts of the combined anlage of the first epibranchial and PTO placodes and/or in dorsal parts of the epibranchial placode 2 arises between HH15 (E) and HH17 (F). b2, branchial arch 2.

Figure 4.

Apoptosis in the posterior placodal area of chicken embryos, Hamburger–Hamilton (HH) stages HH18 to HH23. Schematic reconstructions [lateral views, not drawn to scale; orientation mark indicating dorsal (d), ventral (v), rostral (r) and caudal (c) positions; paired maps, lettered left and mirrored right body sides] demonstrate high‐grade thickenings (grey) revealing the combined anlage of the first epibranchial (e1) and paratympanic organ (PTO) placodes (pt), and the epibranchial placodes 2 and 3 (e2, e3 or e3₁ and e3₂, respectively). Patterns of apoptosis (red), immunopositivity for Sox2 in the PTO placode (green, with dark or light shades indicating strong or weak immunopositivity, respectively), pharyngeal membranes/pouches (dotted black lines) and borders of the branchial arches (thin black lines) are additionally shown. Interplacodal apoptosis type A between the position of the internalized otic vesicle (thin dashed black lines in A‐D, also see solid otic stalk: thick dashed black lines in A, B) and the combined anlage of the first epibranchial and PTO placodes, and/or between the otic anlage and other epibranchial placodes, peaks around HH17/HH18 (A) and regresses from HH19 onwards (B–D). Interplacodal apoptosis type B between neighbouring epibranchial placodes starts around HH21 (C) and peaks on HH23 between the epibranchial placodes 2 and 3₁ (D). (A–C) Between HH18 and HH21, apoptosis is present within and/or adjacent to the developing (Sox2⁺) PTO placode and/or in dorsal parts of the epibranchial placode 2. (D) From HH23 onwards, apoptosis slightly decreases within the PTO placode, and centres on dorsal and/or ventral parts of the epibranchial placodes 2, 3₁ and 3₂, the latter two demonstrating a pit‐like invagination (circumscribed by thick black lines). b1/b2, branchial arches 1 and 2 that, from HH23 onwards, overgrow e1 and e2, respectively (D).

Closer analysis of the ectodermal thickening, which appears at first sight to represent solely epibranchial placode 1, confirms that dorsal parts of this thickening generate the Sox2⁺ PTO placode (O'Neill et al. 2012; Figs 4B‐D and 5). Using schematic reconstructions (Fig. 5A,B) and double immunostainings (Fig. 5C‐F), we now show that both the first epibranchial and the PTO placodes develop from a common Pax2⁺ patch of ectoderm. Between HH19 and HH21, this Pax2⁺ patch almost perfectly matches the dorsal margin of the corresponding thickening but slightly exceeds this morphological ‘border’ in the ventral direction (Fig. 5A). At the same time, the Pax2⁺/Sox2⁺ PTO placode constitutes the approximate dorsal third of the thickening, which ventrally bears the Pax2⁺/Sox2⁻ epibranchial placode 1 (Fig. 5A,C‐F). From HH23 onwards, Pax2 largely disappears from the Sox2⁺ PTO placode, but persists in the remnants of the epibranchial placode 1 (Fig. 5B).

Figure 5.

Apoptosis contributes to the development of the paratympanic organ (PTO) placode, which appears to segregate from the Pax2⁺ chicken first epibranchial placode. (A, B) Schematic reconstructions, lateral views, not drawn to scale; orientation mark indicating dorsal (d), ventral (v), rostral (r) and caudal (c) positions; paired maps, left and mirrored right body sides. Strong or weak Pax2 (magenta) and Sox2 (green) immunopositivity is indicated by dark or light shades, respectively. (C‐F) Double immunofluorescence stainings for Pax2 (magenta) and Sox2 (green), each row showing an overview (merged channels) and three images of the magnified placode (merged and individual channels) with relative z‐positions indicated. (A, B) Apoptosis (red) is centred on the PTO placode (pt), but is almost absent from the first epibranchial placode (e1). (A, C‐F) During Hamburger–Hamilton (HH) stages HH19 and HH21, both the dorsally located Sox2⁺ PTO placode and the ventrally located Sox2⁻ epibranchial placode 1 originate from a combined Pax2⁺ anlage. Later (HH23, HH25), downregulation of Pax2 in the PTO placode helps to distinguish between the Pax2⁻/Sox2⁺ PTO placode and the Pax2⁺/Sox2⁻ epibranchial placode 1 (B). Arrows, migrating neuroblasts; dotted black lines (A, B) or p1 (C–F), pharyngeal membrane/pouch 1; gg, geniculate ganglion; ov, otic vesicle. Scale bars in (C–F): 50 μm (overviews) or 20 μm (details).

From HH11 to the end of the investigation period (HH26), a rostrocaudal wave of apoptosis passes through the PPA. Except for the above‐mentioned sites, which were more directly associated with the invaginating otic placode, apoptosis operates in three fundamental positions: (1) between the otic anlage and the epibranchial placodes 1, 2, and/or 3, respectively (‘interplacodal apoptosis type A’), (2) between neighbouring epibranchial placodes (‘interplacodal apoptosis type B’) and (3) within distinct PTO and epibranchial placodes (‘intraplacodal apoptosis’).

Interplacodal apoptosis type A starts on HH14/HH15 (Fig. 3D,E) and peaks around HH17/HH18 (Figs 3F and 4A). From HH19 onwards, type A apoptosis disappears between the otic anlage and the epibranchial placodes 1 and 2, and, with a delay, between the otic anlage and the epibranchial placodes 3 or 3₁ and 3_2, respectively (Figs 4B‐D and 6). Interplacodal apoptosis type B between the epibranchial placodes 1 and 2, as well as between the epibranchial placodes 2 and 3, starts around HH21 (Fig. 4C) and disappears on HH23 between the epibranchial placodes 1 and 2 (Fig. 4D). Again reflecting the caudad progression of the apoptotic wave, interplacodal apoptosis type B peaks between the epibranchial placodes 2 and 3₁ on HH23 (Figs 4D, 7 and 8) and, around HH26, between the epibranchial placodes 3₁ and 3₂ (Figs 6 and 9).

Figure 6.

Apoptosis in the posterior placodal area of chicken embryos, Hamburger–Hamilton (HH) stage HH26. Schematic reconstructions [lateral views, not drawn to scale; orientation mark indicating dorsal (d), ventral (v), rostral (r) and caudal (c) positions; paired map, lettered left and mirrored right body sides] demonstrate the high‐grade thickened (grey) combined anlage of the first epibranchial (e1) and paratympanic organ (PTO) placodes (pt), the epibranchial placode 2 (e2), the invaginated epibranchial placodes 3₁ and 3₂ (e3₁, e3₂, circumscribed by a thick black line), and interplacodal remnants of the posterior placodal area. Patterns of apoptosis (red), pharyngeal membranes/pouches (dotted black lines), and borders of the branchial arches (thin black lines) are also shown. Interplacodal apoptosis type B peaks between the epibranchial placodes 2 and 3₁ as well as between the epibranchial placodes 3₁ and 3₂. Large‐scale apoptosis is also present in the regressing epibranchial placode 2 and in the invaginated epibranchial placodes 3₁ and 3₂. Residual apoptosis is present within and/or adjacent to the Sox2⁺ (green, with dark or light shades indicating strong or weak immunopositivity, respectively) PTO placode. Thin dashed black lines, otic vesicle; b1/b2, branchial arches 1 and 2 that overgrow e1 and e2, respectively.

Figure 7.

Apoptosis in the posterior placodal area of chicken embryos as revealed by anti‐cleaved caspase‐3 immunohistochemistry. Hamburger–Hamilton (HH) stage HH23, relative z‐positions indicated. Overviews (A, C, E) with boxed areas enlarged in (B, D, F) demonstrate intraplacodal apoptosis (white arrowheads) in the epibranchial placode 2 (e2: A‐D) and interplacodal apoptosis type B (black arrowheads) between the epibranchial placodes 2 and 3₁ (A‐F). Branchial arch 2 (b2) has overgrown epibranchial placode 2, and pharyngeal pouch 3 (p3) opens to the amniotic cavity (asterisk in E). Arrow (in A), migrating neuroblasts; b3, b4, branchial arches 3 and 4, respectively; da, dorsal aorta; pg, petrosal ganglion; ph, pharynx; III, IV, branchial arteries 3 and 4, respectively. Scale bars: 100 μm (overviews) or 20 μm (details).

Figure 8.

Apoptosis in the posterior placodal area of chicken embryos as revealed by anti‐cleaved caspase‐3 immunohistochemistry. Hamburger–Hamilton (HH) stage HH23, relative z‐positions indicated (continued from Fig. 7). Overviews (A, C) with boxed areas enlarged in (B, D) demonstrate interplacodal apoptosis type B (black arrowheads) between the epibranchial placode 2 and the deeply invaginated epibranchial placode 3₁ (e3₁). Arrow (in C), migrating neuroblasts; b4, branchial arch 4; cv, cardinal vein; da, dorsal aorta; ng, nodose ganglion; ph, pharynx; p4, pharyngeal pouch 4; va, ventral aorta; white arrowheads, apoptosis in epibranchial placode 3₁; IV, branchial artery 4. Scale bars: 100 μm (overviews) or 20 μm (details).

Figure 9.

Apoptosis in the posterior placodal area of chicken embryos as revealed by anti‐cleaved caspase‐3 immunohistochemistry. Hamburger–Hamilton (HH) stage HH26, relative z‐positions indicated. Sections demonstrate interplacodal apoptosis type B (black arrowheads) between the invaginated epibranchial placodes 3₁ (e3₁) and 3₂ (e3₂, A–C) as well as intraplacodal apoptosis (white arrowheads) in these two placodes (A, B, D). Arrows (in D), migrating neuroblasts; b4, b6, branchial arches 4 and 6, respectively; ng, nodose ganglion; p3, p4, pharyngeal pouches 3 and 4, respectively; IV, branchial artery 4. Scale bar: 20 μm.

Apoptosis in dorsal parts of the combined anlage of the first epibranchial and PTO placodes starts around HH14 (Figs 3D–F and 4A). From at least HH19 onwards, it becomes obvious that apoptotic events in this anlage predominantly eliminate precursor cells within or adjacent to the developing Sox2⁺ PTO placode, but almost never within the ventrally adjacent Sox2⁻ epibranchial placode 1 (Figs 4B‐D, 5A,B and 6). Apoptosis in dorsal parts of the epibranchial placode 2 also starts around HH14 (Fig. 3D) and, with few exceptions, persists or even intensifies from HH15 onwards (Figs 3E,F and 4A‐C). From HH23 to at least HH26, massive apoptosis is present in dorsal and ventral parts of the epibranchial placode 2, as well as in the epibranchial placodes 3₁ and/or 3₂ (Figs 4D, 6, 7A‐D and 9A,B,D). Hence, intraplacodal apoptosis contributes to the regression of these caudal epibranchial placodes, which cease neurogenesis between HH24 and HH25 (Blentic et al. 2011).

Discussion

In mammals, spatiotemporally regulated apoptosis accompanies placode morphogenesis in the PPA as well as in the anterior subsection of the panplacodal primordium (Washausen et al. 2005; Knabe et al. 2009; Washausen & Knabe, 2013, 2017, 2018). However, the nature of its functional contribution has been elusive for many years, and only recently has experimental evidence been provided that, in the PPA of mice, ‘phylogenetic cell death’ (Glücksmann, 1951) eliminates vestigial lateral line placodes with an astonishingly high developmental potential (Washausen & Knabe, 2018). The present work demonstrates that spatiotemporally regulated apoptosis also occurs in the PPA of chicken embryos. Important common features between mammalian and chicken embryos include: (1) the wave‐like rostrocaudal progression of PPA apoptosis, (2) the presence of apoptosis between the otic anlage and the epibranchial placodes 1, 2 and/or 3, respectively (‘interplacodal apoptosis type A’), (3) the additional occurrence of interplacodal apoptosis between neighbouring epibranchial placodes (‘interplacodal apoptosis type B’) and (4) the final emergence of apoptosis within morphologically distinct placodes.

Our findings strongly suggest, but do not yet conclusively prove, the hypothesis that apoptosis in the PPA of chicken embryos, as in mammals (Washausen & Knabe, 2018), eliminates vestigial lateral line placodes. This hypothesis is not inconsistent with our observation that differences of detail exist between the patterns of PPA apoptosis in chicken, mouse and T. belangeri. To name just two examples. Interplacodal apoptosis type A appears to be stronger in mice (Washausen & Knabe, 2013) than in chicken (present results) or T. belangeri (Washausen et al. 2005). In contrast, interplacodal apoptosis type B, especially between the caudal epibranchial placodes, tends to be more massive in chicken (present results) and T. belangeri (Washausen et al. 2005) than in mice (Washausen & Knabe, 2013, 2017). Such interspecies differences are reminiscent of the fact that, in different gnathostome lineages, subsets of lateral line placodes (primitive condition: three preotic and three postotic placodes) have to different extents been lost (Northcutt, 1992; Schlosser, 2002, 2006). Overall, a greater level of similarity in PPA apoptosis appears to exist between chicken and primate‐related T. belangeri as opposed to C57BL/6N mice. In a like fashion, development of the parathyroid glands in chicken embryos differs from that observed in mice but concurs with that in humans (Okabe & Graham, 2004). Hence, chicken and human/primate‐related embryos appear to display similar ancestral features, and mouse embryos display derived features (A. Graham, pers. comm.; present results).

Using identical primary antibodies against Sox2, onset of the formation of the PTO placode was observed between HH14 (O'Neill et al. 2012) and HH19 (present results). Irrespective of this discrepancy, which may depend on strain and/or methodological differences, apoptosis in the combined anlage of the first epibranchial and PTO placodes, from HH14 to HH26, almost exclusively operates within or immediately adjacent to the dorsally located PTO placode (Figs 2B, 3D‐F, 4, 5A,B and 6). These findings strongly suggest that apoptosis is required for decision‐making processes among precursor cells of the early developing PTO placode. From HH29 onwards, apoptosis in the pharyngeal pouch epithelium medial to the developing PTO appears to contribute to the separation of the PTO from the endoderm (O'Neill et al. 2012). Our results also shed new light on the question of whether the PTO placode originates from the Pax2⁺ PPA. Using whole‐mount in situ hybridization, O'Neill et al. (2012; p. 3) observed that, on HH18, ‘a patch of Sox2‐positive ectoderm’ (…) resides ‘dorsal to the geniculate placode, which we identified by Pax2, Sox3 and Delta1 expression’. Correspondingly, on HH24, double immunostaining for Sox2 and Pax2 revealed ‘a strongly Sox2‐positive patch of thickened ectoderm immediately dorsal to the Pax2‐positive, Sox2‐negative geniculate placode remnant’ (O'Neill et al. 2012, p. 5). Hence, Pax2 mRNA and/or Pax2 protein expression appear to be absent from the paratympanic placode (Fig. 3b, c, k in O'Neill et al. 2012). Our study complements these previous observations by showing (1) that, at least between HH19 and HH21, a common anlage of the geniculate and PTO placodes can be defined morphologically (‘thickened placodal ectoderm’) as well as by immunostaining for Pax2 (Fig. 5A,C‐F); (2) that Sox2⁺ (PTO placodal) cells seem to segregate from the Pax2⁺ geniculate placode and (3) that absence of Pax2 immunostaining from the PTO placode from HH23 onwards is due to secondary loss (Fig. 5A,B). These findings suggest that the PTO placode derives from the Pax2⁺ PPA. Correspondingly, Ladher et al. (2010) assumed that the PTO placode forms from non‐otic regions of the PPA.

Author contributions

Both authors were involved in every aspect of the work.