The cerebellum is responsible for coordinating precision and timing of complex motor functions, cognitive and perceptual analysis, and memory and emotional processing (Middleton and Strick, 1998; Larouche and Hawkes, 2006; Sultan and Glickstein, 2007; Baumann and Mattingley, 2012). Underlying this diverse functionality is a characteristic histological organization and pattern of cellular compartmentalization. It is understood that the adult cerebellum of various mammals and birds demonstrates a conserved pattern of antigen expression by major cell types of the cerebellum, including Purkinje cells and granule cells. Although these immunoreactive compartments correspond to physiologically defined zones (Gravel et al., 1987; Ji and Hawkes, 1994; Voogd et al., 2003; Sugihara and Shionda, 2004; Voogd and Ruigrok, 2004; Pakan and Wylie, 2008; Pakan et al., 2010, 2011), little is known about their development outside of mammals. Here, we provide new information on embryonic cerebellar development in one of the most common developmental models, the domestic chicken (Gallus gallus domesticus). Domestic chickens (hereafter, chicks) have a lengthy history of use in the study of neurodevelopment (Waddington, 1930; Le Douarin, 1973; Keynes and Stern, 1984; Lumsden and Kiecker, 2004, Ma et al., 2012), offering the advantages of accessibility throughout embryogenesis (Hamburger and Hamilton, 1951; Schmutz and Grimwood, 2004) and rapid development to a fully precocial state (characterized by high levels of motor coordination and evidence of olfactory, visual, and auditory abilities; Stark, 1993). In addition, embryonic chicks are amenable to various in ovo manipulations including microsurgical tissue grafting and lineage tracing methods that form the foundation of our understanding of the commitment and determination of different neural cell types as well as their interactions during development (reviewed by Sauka-Spengler and Barembaum, 2008).
The cerebellum of birds, similar to that of mammals, has a complex and hierarchical morphology, beginning with subdivision of the cortex into three major lobes (anterior, posterior, and flocculonodular) separated by prominent fissures. Each lobe has a number of folds known as folia (= lobules in mammals) (Larsell, 1972). On the basis of molecular markers and afferent projections, the avian cerebellum is further compartmentalized into a series of transverse zones. The pattern of zonation includes the lingular zone (LZ; folium I), anterior zone (AZ; folia II–V), central zone (CZ; folia VI–VII), posterior zone (PZ; folia VIII–IX), and nodular zone (NZ; posterior folium IX and folium X) (pigeons: Pakan et al., 2007; hummingbirds: Iwaniuk et al., 2009; adult chickens: Marzban et al., 2010). These transverse zones are divided mediolaterally by species-specific and antigen-specific parasagittal striping patterns, shown as alternating subsets of immunopositive and immunonegative Purkinje cells (Wylie et al., 2011; see Apps and Hawkes, 2009 for review). In coronal and sagittal sections, the mature cerebellum is organized into three distinct layers each with a unique combination of cell types. The molecular layer includes stellate and basket cell interneurons, Purkinje cell dendrites, and parallel fibers. Deep to this, the Purkinje cell layer is a monolayer of Purkinje cell somata. Next, the granule cell layer contains primarily granule cells, Golgi cells, unipolar brush cells, and Lugaro cells (see Ambrosi et al., 2007) as well as afferent mossy fiber terminals.
Based on differential patterns of protein expression, Purkinje cells and granule cells constitute a heterogeneous neuronal population (Hawkes and Leclerc, 1987; Hawkes and Turner, 1994; Armstrong et al., 2000; reviewed in Apps and Hawkes, 2009). Previous studies in mice have demonstrated that while the pattern of protein expression by Purkinje cells and granule cells is static among adults, it is dynamic during development (Leclerc et al., 1988; Hawkes and Eisenman, 1997; Armstrong et al., 2001). In this study, we used western blot analysis and immunohistochemistry to document the onset and expression pattern of three proteins of interest at embryonic (E) days 8, 10, 12, 14, 16, 18, and 20 during chick development, focusing specifically on the antigens Calbindin (CB), Zebrin II (ZII), and Calretinin (CR).
CB is a 28-kDa calcium-binding protein expressed by cerebellar Purkinje cells (e.g., Wassef et al., 1985; Celio, 1990; Bastianelli and Pochet, 1993). In embryonic mice, CB is expressed in a heterogeneous pattern of immunopositive and immunonegative Purkinje cell clusters arranged on either side of the midline starting at E16 (Wassef et al., 1985). Beginning perinatally and continuing throughout postnatal development and adulthood, CB is globally expressed by all Purkinje cells in rodents and birds (rat: Celio, 1990; mouse: Ozol et al., 1999; pigeon: Pakan et al., 2007; adult chicken: Marzban et al., 2010). CB acts primarily as an intracellular Ca2+ buffer, and has been shown to play a role in Purkinje cell-dependent motor learning in CB−/− mice (reviewed in Schwaller et al., 2002).
ZII (aldolase C), is a 36-kDa glycolytic isozyme (Ahn et al., 1994) that, similar to CB, is selectively expressed by Purkinje cells within the cerebellum. Unlike CB, ZII expression in adults is discontinuous, leading to the recognition of transverse zones (Ozol et al., 1999) and parasagittal stripes (Hawkes and Leclerc, 1987; reviewed by Larouche and Hawkes, 2006) in the adult cerebellum. In mice, ZII is first expressed around postnatal day 5 (P5) in Purkinje cells of the posterior lobe (PZ and NZ: lobules VIII–X). From these posterior positions, ZII expression spreads dynamically to increasingly more anterior and lateral cells until global expression in all Purkinje cells is seen at ∼P12. As postnatal development continues, approximately half of the ZII immunopositive (ZII+) Purkinje cells stop expressing this protein (Leclerc et al., 1988; Eisenman and Hawkes, 1993; Hawkes and Herrup, 1996; reviewed in Armstrong and Hawkes, 2000). The result is a conserved striping pattern of ZII+ and ZII− Purkinje cells. In addition to mice, the pattern of ZII stripes has been mapped in numerous species including rat (Brochu et al., 1990), opossum (Doré et al., 1990), guinea pig (Larouche et al., 2003), hedgehog (Sillitoe et al., 2003), fish (Brochu et al., 1990), pigeon (Pakan et al., 2007), hummingbird (Iwaniuk et al., 2009), and adult chicken (Marzban et al., 2010). In adult avian species, ZII+ Purkinje cells are organized into a pattern of parasagittal stripes within the AZ and PZ, and Purkinje cells of the LZ, CZ, and NZ are almost uniformly ZII+ (Pakan et al., 2007; Iwaniuk et al., 2009; Marzban et al., 2010).
CR is a 29-kDa calcium-binding protein that is expressed by a variety of cells in the cerebellar cortex. In adult rodents, these include granule cells, unipolar brush cells, Golgi cells, stellate and basket cells, and Lugaro cells as well as mossy fibers and climbing fibers (Rogers, 1989; Résibois and Rogers, 1992; Floris et al., 1994; Diño et al., 1999; Nunzi et al., 2002; reviewed in Baimbridge et al., 1992).
In posthatch chicks, CR expression was seen in stellate and basket cells, mossy fibers and a subset of climbing fibers (Rogers, 1989; De Castro et al., 1998; reviewed in Bastianelli, 2003), while granule cells were CR-immunonegative (Rogers, 1989; Bastianelli and Pochet, 1993). During development, the transient expression of CR in the granule cell layer of the chick cerebellum has previously been reported (Bastianelli and Pochet, 1993). Although the physiological function of CR in the central nervous system remains poorly understood, Bastianelli and Pochet (1993) suggest that the induction and/or expression of CR, and other calcium-binding proteins, may be linked to neurodegeneration. CR-deficient mice display impaired motor coordination suggesting a role in cerebellar function (Schiffmann et al., 1999; Dove et al., 2000; Gall et al., 2003).
Our results demonstrate a unique sequence and pattern of expression for each of CB, ZII, and CR during cerebellar development in the chick, similar to that seen in mammals. However, the developmental expression of each antigen was consistently earlier in chick despite the fact that both species have a similar embryonic period of 19–22 days.
Animal procedures followed university regulations and the Guide to the Care and Use of Experimental Animals from the Canadian Council of Animal Care. Fertilized Barred-Rock chicken eggs were obtained from Arkell Poultry Research Station (University of Guelph), rotated 90 degrees and incubated at 40°C for 8–20 days. Chick embryos used for serial histology were staged according to morphological criteria (e.g., length of the third toe), whereas chick embryos used for western blots were staged according to morphological criteria or aged according to absolute time in ovo.
Western Blot Analysis
Cerebella and brainstems from embryonic day (E)10, 12, 14, 16, 18, and 20 chicks were dissected, then combined with triple detergent lysis buffer (200 μL/0.1 g:50 mM Tris-HCl, 150 mM NaCl, 0.2% Na azide, 0.5% Na deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1% Nonidet P-40), homogenized, and stored at −80°C. Aliquots (20 μL) were combined with sample buffer (80 μL per aliquot: 50 mM Tris Cl (pH 6.8), 2%(w/v) SDS, 0.1% bromophenol blue, 10% (V/V) glycerol, 100 mM dithiothreitol). Proteins were boiled for 10 min to 100°C. Samples (20 μg per μL) of homogenized tissue were loaded on a SDS-polyacrylamide gel (5% stacking gel, 12% resolving gel) according to the method of Laemelli (1970). Mid-range protein molecular weight markers (5 μL per lane: BioRad, Hercules, CA) were treated in the same way and loaded in parallel. Samples of homogenized mouse neocortex or brainstem (20 μg per μL) were run in parallel and used as positive controls. Samples were electrophoresed at 85 V for 90 min and then electroblotted onto an Immobilon-P membrane (Millipore, Mississauga, ON) at 90 V for 2 hr. Following the transfer, membranes were air-dried and stained with Ponceau red (Sigma, St. Louis, MO) for 30 min to confirm the presence of proteins.
Membranes were saturated with methanol, rinsed with Tris-buffered saline (TBS; pH 7.4) and then washed for 3 × 15 min in 0.2% Tween-20 + 1× TBS. Membranes were blocked for 7 hr in 3% bovine serum albumin (BSA) [IgG-Free, Protease-Free (lot#80927; Jackson ImmunoResearch Laboratories, West Grove, PA)]. Primary antibody [monoclonal mouse anti-Calbindin (1:10,000;lot# 18(F); Swant, Bellinzona, Switzerland; Code #300]; polyclonal rabbit anti-Calretinin (1:10,000, Swant; Code #7699/3H); or monoclonal mouse anti-Zebrin II (1:5,000, gift from Dr. R Hawkes, University of Calgary, AB, Canada) was added to the BSA block, and the membrane was incubated overnight on a shaker. Membranes were then washed 3 × 15 min in 0.2% Tween-20 + 1× TBS (TBST) and incubated for 1 hr at room temperature with peroxidase-labeled goat anti-mouse peroxidase secondary antibody (1:10,000; Vector Laboratories, Burlingame, CA) in TBST. Blots were washed 3 × 15 min in TBST. Peroxidase binding was detected by using the electrogenerated chemiluminescence Western blot detection kit and Hyperfilm ECL (Amersham, Arlington Heights, IL).
Perfusion and Tissue Sectioning
Embryonic chicks were deeply anaesthetized with sodium pentobarbital (100 mg/kg, i.p.) and transcardially perfused with 0.9% NaCl in 0.1 M phosphate buffer (pH 7.4) followed by 4% paraformaldehyde with 0.02% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were then removed, post fixed in 4% paraformaldehyde fixative at 4°C for 24 hr, and stored in Millonig's solution (pH 7.6). After being embedded in 3% agar, a series of 50 μm thick transverse and sagittal sections were cut through the extent of the cerebellum on a Vibratome.
Immunohistochemistry was carried out as described previously (Sloviter and Nilaver, 1987; Armstrong et al., 2000, 2001). Briefly, tissue sections were washed thoroughly, blocked with 10% normal goat serum (Jackson Immunoresearch Laboratories), and then incubated in 0.1 M PBS buffer containing 0.1% Triton-X, 0.005% PBS (PBS B), and the primary antibody [monoclonal mouse anti-Calbindin: diluted 1:10,000, lot #18(F), Swant; Code #300], or monoclonal anti-Zebrin II (diluted 1:500, gift from Dr. R Hawkes), or monoclonal mouse anti-Calretinin (diluted 1:10,000, Swant; Code #6B3) for 16–18 hr at 4°C. Primary antibody was diluted until specific staining was seen and omitted in some cases to serve as controls (Saper and Sawchenko, 2003). Secondary incubation in biotinylated goat anti-mouse (115-066-003) or biotinylated goat anti-rabbit (111-066-003) antibody (both at 1:1,000 in PBS B; Jackson Immunoresearch Laboratories) lasted 45 min at 4°C. The Vectastain ABC Staining Kit (1:1,000, Vector Laboratories) and diaminobenzidine were used to visualize the reaction product. Sections were dehydrated through an alcohol series, cleared in xylene, and coverslipped with Entellan mounting medium (BDH Chemicals, Toronto, ON).
For double-labeling, E18 coronal sections were washed and blocked as described above and then incubated in polyclonal rabbit anti-Calbindin [diluted 1:10,000, lot #18(F), Swant; Code #CB38] and monoclonal mouse anti-Zebrin II (diluted 1:200; gift from Dr. R Hawkes) for 16–18 hr at 4°C. Sections were rinsed in PBS, incubated in Cy3-goat anti-rabbit (111-165-003), and Cy2-goat anti-mouse (115-225-003) secondary antibodies (both at 1:1,000 in PBS; Jackson Immunoresearch Laboratories) overnight before mounting with FluorSave (Calbiochem, La Jolla, CA).
Cresyl Violet Histochemical Staining
Embryonic cerebellar sections processed for histochemistry were mounted on slides and dried overnight. Slides were dipped in deionized water and dehydrated through an alcohol series of 70%, 95%, and 100% ethyl alcohol. Slides were dipped in xylene and then rehydrated through the alcohol series and deionized water. Slides were placed in 1% Cresyl violet solution (Sigma-Alderich, Saint Louis, MO) for 30 sec, rinsed in deionized water, and dehydrated to 100% ethyl alcohol. Slides were transferred to 100% ethyl alcohol with acetic acid followed by 100% ethyl alcohol and then cleared in xylene before being coverslipped with Entellan mounting medium (BDH Chemicals).
Photomicrographs were captured with an AxioCam MRc5 digital camera (Carl Zeiss Canada), and all images were assembled in Adobe Photoshop 10.0 (San Jose, CA). Images were cropped and corrected for brightness and contrast but were not otherwise manipulated. Cerebellar lobules were labeled according to Puelles et al. (2007).
Western Blot Analysis
An ∼28-kDa immunoreactive band of protein was seen following western blot analysis of embryonic chick cerebella homogenates at E10, 12, 14, 16, 18, and 20 consistent with the established molecular weight for CB (Bastianelli and Pochet, 1993; Marzban et al., 2010) (Fig. 1A). Similarly, a ∼36-kDa immunoreactive protein band was seen in homogenates of E12, 14, 16, 18, and 20 embryonic chick cerebella consistent with the established molecular weight for ZII (e.g., Sillitoe et al., 2005; Fig. 1B). Finally, a ∼30-kDa immunoreactive band of protein was seen in homogenates of E10, 12, 14, and 18 embryonic chick cerebella and brainstem, consistent with the established molecular weight for CR (e.g., Villa et al., 1994; Fig. 1C).
To determine the pattern of distribution of CB immunopositive (CB+), Purkinje cells we used immunohistochemistry on serial parasagittal and coronal sections of the chick cerebellum at each of E8, E10, E12, E14, E16, E18, and E20 (Figs. 2 and 3). CB+ Purkinje cells demonstrate a dynamic and asynchronous pattern of distribution and immunoreactivity within the cerebellar cortex. As revealed in parasagittal sections through the cerebellar cortex at E10 (Fig. 2A–C), clustered CB+ Purkinje cells were seen in folia II, III/IV, VII, VIII, and IX, interspersed with CB-regions. Immunostaining is most intense in folia VII, VIII, and IX (Fig. 2A–C). Transverse boundaries marking a sharp interface between CB+/− Purkinje cells were seen in folia II, V, and at the dorsal edge of folium VIII (Fig. 2A–C). The location of this posterior transverse boundary appears to align with the CZ-PZ boundary seen in mice (Ozol et al., 1999; Armstrong et al., 2000). Differential interference contrast (DIC: Fig. 2C) imaging confirmed that Purkinje cells were present beyond this boundary, but that the majority of these cells do not express CB (bracket). By E12, the majority of Purkinje cells in the anterior cerebellum were CB+ while transverse boundaries were still seen in posterior regions of the cerebellum (Fig. 2D). For example, the distinct transverse boundary between folia VII and VIII is still apparent (Fig. 2D, E, and F). From E14 on, the majority of Purkinje cells were CB+ (Fig. 2G) with well-defined somata and distinct dendritic outgrowths (Fig. 2H and I), although a distinct transverse boundary between immunopositive and immunonegative cells persists in folium X. By E18, the trilaminar organization of the mature cerebellar cortex is established with a complete monolayer of CB+ Purkinje cells forming the Purkinje cell layer, as well as immunopositive dendritic trees in the molecular layer (Fig. 2J and K).
In coronal sections through the E10 chick cerebellum, CB+ Purkinje cells were arranged in mediolateral clusters on either side of the midline separated by raphes at the midline and bilateral positions (Fig. 3A). DIC imaging confirmed that immunonegative raphes are cell-sparse regions (Fig. 3B). These developmental raphes were clearly visible at E14 (Fig. 3C and D) and E16 (Fig. 3E and F). Narrow raphes persist in lateral positions of the anterior vermis at E20 (Fig. 3G and H).
The morphology of Purkinje cells in the embryonic chick cerebellum changes dramatically during embryonic development (Fig. 4). CB+ Purkinje cells throughout the E10–E14 cerebellum remain morphologically immature (Fig 4A). Tight clusters of round Purkinje cell bodies appear to lack the complex dendritic arborization that usually define this cell type (Wassef et al, 1985; Celio, 1990; Bastianelli and Pochet, 1993; Larouche and Hawkes, 2006; Marzban et al., 2010). By E16, the Purkinje cell bodies are somewhat angular, but remain in clusters rather than forming a monolayer (Fig. 4B). CB+ axons were first seen at E18 at the same time as the nearly complete formation of the monolayer (Fig. 4C). By E20, the Purkinje cells are morphologically mature with teardrop-shaped cell bodies, complex dendritic trees in the molecular layer, and axonal projections into the granule cell layer into the white matter of the cerebellum (Fig. 4D). By this age, Purkinje cells are consistently aligned into a monolayer.
To determine the expression pattern of ZII immunopositive (ZII+) Purkinje cells, we used immunohistochemistry on serial sections (parasagittal and coronal) of chick cerebella at each of E8, E10, E12, E14, E16, E18, and E20. ZII immunoreactive Purkinje cells were not seen until E12, at which point weakly immunostained clusters were seen along the dorsal border of folia VII (Fig. 5A, arrowheads in inset). Before this, ZII expression was restricted to brainstem nuclei and midline tracts (data not shown). At E14, there is a rostrocaudal gradient of ZII immunoreactivity, with the majority of ZII+ Purkinje cells appearing in clusters positioned medially within folium I (lingual zone), and folia IX and X (NZ: Fig. 5B). At E16–E18, ZII is expressed in Purkinje cells throughout the developing chick cerebellum (Fig. 5C–F). Following the establishment of the trilaminar cerebellar cortex at E18, subsets of Purkinje cells begin to turn off their ZII expression (Fig. 5G–H) initiating a pattern of ZII+ and ZII− parasagittal stripes. Double-labeling with ZII and CaBP at E18 reveals the emergence of these parasagittal stripes in folium VIII (Fig. 6A–C).
To determine the expression pattern of CR, we used serial parasagittal and coronal sections through the chick cerebellum at E8, E10, E12, E14, E16, E18, and E20. Similar to CB and ZII expression, the timing of initial immunoreactivity and pattern of distribution of CR immunopositive (CR+) cells is dynamic and asynchronous throughout development. From E8–E10, CR+ cells were seen in the deep cerebellar nuclei and developing Purkinje cells clustered in posterior folia (Fig. 7A–C). By E12, CR immunoreactivity was seen in the granule cell layer from folia II to IX (Fig. 7D) as well as scattered interneurons throughout the molecular layer and the granule cell layer of folium I (LZ) and Purkinje cells in the ventral border of folia IXc and X (NZ) (Fig. 7F). A similar, but more intense pattern of immunostaining is observed at E16 (Fig. 7G–I). By E18, CR immunoreactivity was confined primarily to the granule cell layer in folia VI–VII and the dorsal border of folium IXc (Fig. 7J and K), with scattered immunopositive interneurons in the molecular layer (Fig. 7K and L). In the E20 cerebellum, CR immunoreactivity is confined to the molecular layer and granule cell layer: CR+ Purkinje cells are no longer seen in folia IX, X (Fig. 7M), or elsewhere in the embryonic chick cerebellum.
Our results demonstrate that CB, ZII, and CR are asynchronously and dynamically expressed during cerebellar development in the embryonic chick. Of the markers investigated, CB and CR have the earliest onset at ∼E8, immunostaining Purkinje cells within medial positions of the posterior folia (IX, X), compared to ZII which was not seen until ∼E12–E14 in the chick cerebellum.
CB expression in developing Purkinje cells is conserved across many species, including mice (Wassef et al., 1985, Larouche and Hawkes, 2006), rats (Wierzba-Bobrowicz et al., 2011), kittens (Résibois and Poncelet, 2004), humans (Milosevic and Zecevic, 1998), and chicks (Jeffrey et al., 2003; Pires et al., 2006; this study), and thus is a useful marker of this cell type in the cerebellum. In embryonic chicks, CB expression is seen in Purkinje cells beginning at E10. Immunostaining is most pronounced in the medial posterior cerebellum, and Purkinje cells appear in clusters lacking the adult Purkinje cell morphology (Fig. 4A). During embryogenesis, these Purkinje cells remain CB+, become organized into a monolayer and, by E18–20, adopt a distinctive morphology with a prominent cell body and elaborate dendritic arbor (Fig. 4C and D). In embryonic mice, CB+ Purkinje cell clusters are seen in medial posterior regions of the cerebellum beginning at E16 (Wassef et al., 1985, Larouche and Hawkes, 2006). Just before birth, these predominantly CB+ Purkinje cells have dispersed to form a monolayer (Wassef et al., 1985, Pires et al., 2006). A similar expression pattern has been seen with the closely related calcium-binding protein parvalbumin which is expressed in Purkinje cells in the embryonic chick cerebellum from E12 on (Pires et al., 2006).
Purkinje Cell Development and Dispersal
In the chick, Purkinje cells are born between E3 and E5 and are initially organized into clusters of morphologically immature cells (Yurkewicz et al., 1981; Kanemitsu and Kobayashi, 1988). Consistent with previous findings (Feirabend and Voogd, 1979; Karam et al., 2000), we observed a bilateral distribution of CB+ Purkinje cell clusters early during development (e.g., E10), with adjacent cell clusters separated by developmental raphes. These cell sparse raphes persist until the later stages of embryonic development (Fig. 3; see also Bastianelli and Pochet, 1993, Pires et al., 2006), but appear to be absent in the adult chicken (Marzban et al., 2010). Developmental raphes have also been described in postnatal rat pups (Wassef et al., 1985). Although it remains unclear what role, if any, raphes play in Purkinje cell dispersal and/or cerebellar pattern formation, it has been suggested that these raphes may function as pathways for the migration of granule cells (Mecha et al., 2001).
As described in the developing rodent cerebellum (e.g. Wassef et al., 1985), the formation of a Purkinje cell monolayer began with clusters of morphologically immature cells (E10) which then disperse (E14) and gradually adopt the distinctive mature phenotype complete with an elaborate dendritic arbor (by E18–20; Figs. 3 and 4). Future studies should aim to further characterize dendritic development and correlative physiological maturation with Purkinje cell morphology, as has been done in the postnatal rat (Armengol and Sotelo, 1991; McKay and Turner, 2005).
Initiation of ZII expression in the chick cerebellum corresponds to global expression of CB and precedes the dispersal of Purkinje cells into a monolayer. As in rat (Leclerc et al., 1988), cerebellum development in the chick includes a transient period of near global expression of ZII, where the majority of Purkinje cells are ZII+ before the formation of alternating ZII+ and ZII− stripes (∼E14–E16; Fig. 5). It is worth noting that ZII expression by Purkinje cells is not uniform, with more intense immunostaining observed in folia I and X compared with folia II–IX. These data are consistent with the prediction that folia I and X are embryologically similar (Pakan et al., 2007). In the adult chicken, Purkinje cells in folia I and X remain ZII+ while in folia II-V and VIII–IX subsets of Purkinje cells downregulate ZII expression during late embryogenesis (E18–20), revealing a distinctive parasagittal pattern of alternating immunopositive and immunonegative stripes (Marzban et al., 2010).
Overall, the expression pattern of ZII in late staged chick embryos closely matches that of adult birds and, to some extent, mammals. For example, the cerebellum of an E20 chick just before hatch and an adult chicken cerebellum demonstrate a comparable pattern of ZII: global expression of ZII in the lingular (I), central (VI–VIII) and nodular (X) zones, and parasagittal stripes in the anterior (II–V) and posterior (VIII–IXcd) zones (Marzban et al., 2010). A nearly identical pattern of ZII expression has also been described for the adult pigeon (Pakan et al., 2007) and hummingbird (Iwaniuk et al., 2009). A similar striped pattern of ZII expression by Purkinje cells is also characteristic of many mammalian species (reviewed by Sillitoe et al., 2005).
The pattern of CR expression during cerebellar development is complex. Previous studies in mice describe CR expression at E14 in cells predicted to become unipolar brush cells (Abbott and Jacobowitz, 1995). In this study, we observed a distinctive subset of cells in the NZ that transiently express CR from E8–E12 that we believe are Purkinje cells based on morphology and location. Although other studies have failed to identify CR+ Purkinje cells in the developing chick cerebellum (e.g., Bastianelli and Pochet, 1993), CR+ Purkinje cells have been observed in the developing cerebellum of humans (Yew et al., 1997), in various mutant mouse strains both in developing and adult cerebella (Résibois et al., 1997), and in adult cerebella of squirrel monkeys (Fortin et al., 1998). The discrepancy between these results cannot currently be explained, although it may reflect differences in the immunohistochemical or tissue processing protocol, and will be investigated further.
By E12, we observed additional CR+ cells in the deep cerebellar nuclei, granule cell layer (folia II–IX; most likely unipolar brush cells—see Braak and Braak, 1993; Floris et al., 1994; Diño et al., 1999), and interneurons of the molecular layer (folium I). Gradually, CR immunoreactivity becomes restricted to the granule cell layer (folia VI, VII, and dorsal border of IXc, with only isolated immunopositive cells in the molecular layer (E18–E20).
The zonal boundaries described in mice (Ozol et al., 1999; Marzban and Hawkes, 2011), and adult avian species (Pakan et al, 2007; Iwaniuk et al., 2009) are revealed by the expression patterns of CB and CR during embryonic development of the chick cerebellum. At E10, we observed an overall anterior–posterior gradient of CB expression as well as specific boundaries of expression between folia I and II (LZ-AZ boundary), folia VII and VIII (CZ-PZ boundary), and folia IX and X (PZ-NZ boundary). From E12 on, the majority of Purkinje cells expressed CB. These expression boundaries were apparent in lateral but not medial sections through the cerebellum, as previously reported in the embryonic rodent cerebellum (Wassef et al., 1985). The expression of CR also supports the presence of transverse zonal boundaries between folia I and II (LZ-AZ), V and VI (AZ-CZ), VII and VIII (CZ-PZ), and within lobule IX (PZ-NZ) in both the Purkinje cell and granule cell layers (Fig. 7).
Comparative Neurodevelopment: Chicks and Mice
Cerebellum development follows a conserved sequence of developmental events in both chicks and mice. Interestingly, while both species demonstrate a similar timeframe of in ovo/ in utero development (19–22 days), the timing of neurodevelopmental events is consistently earlier in chicks than mice. For example, we have shown that CB is expressed by Purkinje cells in chicks as early as E8, but is not seen in mice until E16 (Wassef et al., 1985; Fig. 8). Similarly, ZII is expressed by Purkinje cells in chicks at E12 and is not seen until postnatal day 5 (P5) in mice (Eisenman and Hawkes, 1993; Fig. 5). The time required to establish the parasagittal pattern of ZII+/− stripes in chicks is also accelerated compared to mice (∼4 days in chick vs. ∼10 days in rodents: Rivkin and Herrup, 2003). Although presently untested, we suggest that the observed differences in timing of initial onset and duration of neurogenic events may reflect the spectrum of different behaviors, sensory abilities, and levels of motor coordination seen at hatch/birth in chicks and mice. Precocial species, such as chicks, demonstrate comparatively high levels of motor coordination at hatch (Rogers, 1995) matching with the earlier embryonic expression of the cerebellum proteins CB, ZII, and CR. In contrast, altricial species, such as mice, have limited sensory-motor functions at birth, rendering them immobile and thus reliant on parental care (Starck and Ricklefs, 1998). These functional and behavioral features may correspond with a comparatively delayed onset of cerebellum protein expression and hence a less mature cerebellum (Starck and Ricklefs, 1998; Charvet and Striedter, 2011). Future investigations targeting additional precocial/altricial mammalian/avian species (e.g., guinea pig and zebra finch), combined with functional studies, will provide valuable insight into the morphological, physiological, and biochemical changes that characterize brain development in precocial or altricial species and the possible role of neurodevelopmental heterochrony as a source of structural, functional, and evolutionary variation.
The authors are grateful to Dr Richard Hawkes (Hotchkiss Brain Institute, Dept of Cell Biology and Anatomy, University of Calgary, Calgary AB) for his generous donation of the Zebrin II antibody, for the technical assistance of Sharmin Chowdhury, and to Dr. Ann Hahnel (BioMedical Sciences, University of Guelph) as well as Samantha Payne for excellent discussion.