Nonpathological inflammation drives the development of an avian flight adaptation

Abstract

The development of modern birds provides a window into the biology of their dinosaur ancestors. We investigated avian postnatal development and found that sterile inflammation drives formation of the pygostyle, a compound structure resulting from bone fusion in the tail. Inflammation is generally induced by compromised tissue integrity, but here is involved in normal bone development. Transcriptome profiling and immuno/histochemistry reveal a robust inflammatory response that resembles bone fracture healing. The data suggest the involvement of necroptosis and multiple immune cell types, notably heterophils (the avian equivalent of neutrophils). Additionally, nucleus pulposus structures, heretofore unknown in birds, are involved in disc remodeling. Anti-inflammatory corticosteroid treatment inhibited vertebral fusion, substantiating the crucial role of inflammation in the ankylosis process. This study shows that inflammation can drive developmental skeletogenesis, in this case leading to the formation of a flight-adapted tail structure on the evolutionary path to modern avians.

Keywords: corticosteroid; heterophil; intervertebral disc; necroptosis; nucleus pulposus.

Fig. 1.

Tail morphology and vertebral fusion in extant and fossil birds. (A) G. gallus tail. Free, or unfused, caudal vertebrae are at the proximal end, and the distal caudal vertebrae are fused into the pygostyle structure. Solid red areas are IVDs, and dotted red lines indicate fusion planes within the pygostyle. (B) Left, nucleus pulposus (NP) formation diagram, sagittal view, from the embryonic notochord (NC) (Left) to postnatal NP (Right); SC: spinal cord. IVDs are red. Right, postnatal IVD diagram. The annulus fibrosus (AF) surrounds the NP, and these are encased in the cartilage end plates (CE) of the disc between vertebral centra (VC). Bone end plates (BE) are outside of the IVD. (C) Top, vertebral fusion diagram. Neural arch (NA) and spinous process (SP) elements fuse to form the neural arch, followed by ankylosis of the fused neural arch to the centrum (C). Middle, diagram of the Archaeopteryx spinal column (adapted from ref. 9); anterior to the Left. Archaeopteryx had 4 to 6 fused vertebrae in its sacrum. Bottom, diagram of the G. gallus spinal column, anterior to the Left. The chicken has 23 to 24 combined fused vertebrae in its axial column in the notarium, synsacrum, and pygostyle compound structures.

Fig. 2.

NP formation and IVD architecture in the avian tail. All panels are sagittal views, as noted in the schematic in (A), except for (B), middle panel, transverse view. (A) NP development in chicken from E19 to D50 (ABPR staining). Notochord (NC) partitioning occurs at E19; OC are ossification centers in the centra; scale bar, 250 µm. At D8, cellular NPs are evident in both pygostyle discs (Left) and free vertebrae caudal discs (Right). Arrows indicate NPs; scale bar, 200 µm. At D50, NPs are acellular (scale bar, 2 mm) (magnified view of the NP to the Right of the panel, scale bar, 100 µm) compared to the cellular adult mouse NP (scale bar, 1 mm), right magnified panel, scale bar, 100 µm. NPs (white arrows), AF (yellow arrowhead), and CE (green arrowhead) structures. (B) Lamellar architecture of chicken and mouse IVDs by plane polarized light. Left, D50 chicken free tail, scale bar, 500 µm. Middle, D50 chicken, scale bar, 300 µm. Right, adult mouse, scale bar, 100 µm. (C) Pygostyle IVDs in the emu and chicken, ABPRH staining. In the emu hatchling pygostyle, IVDs with nascent NPs are observed (cryosection; scale bar, 250 µm). At 4.5 mo, proteoglycan-rich NPs in the pygostyle are evident in the full pygostyle IVD and in magnified view; scale bar, 250 µm. In chicken pygostyle IVDs, NPs and fibrous connective tissue (FCT) are evident; D50; scale bar, 250 µm.

Fig. 3.

Pygostyle fusion events. All panels are sagittal views. (A) Progression of fusion. By D50, in the proximal disc, blood vessels (arrows) accumulate in the FCT below the NP; PRH staining; scale bar, 250 µm. In a magnified view, blood cells enclosed in a blood vessel are seen in the FCT before extravasation (eosin/SYTOX green staining; scale bar, 50 µm). Also, at D50 but in the distal disc, at a slightly later stage, blood cell extravasation into the FCT is observed (eosin/SYTOX green staining; the white asterisk indicates the center of a deteriorated blood vessel; scale bar, 50 µm). At D50, cells morphologically consistent with heterophils (8 to 10 µm cell diameter, bilobed nuclei) are observed (the arrow indicates a cluster of heterophils in the FCT; Giemsa–Wright stain; scale bar, 50 µm). By D105, chondrocyte differentiation to hypertrophic chondrocytes (arrows) is observed in the AF of the proximal pygostyle disc, ABPR staining; scale bar, 100 µm). Also, at D105 in the proximal disc, osteogenesis at the IVD center, where the NP had resided, is seen (noted by the arrow; scale bar, 2 mm). The distal disc has completely remodeled to bone. Similarly, ankylosis begins at the NP site in the emu; distal disc, 4.5 mo. (B) TUNEL assays; scale bar’s 100 µm. At D21, sparse cell death is detected (note background cytoplasmic FCT stain). By D50, in both discs, there is considerable cell death in the FCT and surrounding AF cells; TUNEL-stained nuclei are evident in the magnified boxed area. At D105, cell death is observed in the AF, below the NP, in hypertrophic chondrocytes.

Fig. 4.

RNAseq data, supported by RT-PCR and protein analyses. (A) Left, all top 10 GO biological processes from the RNAseq profiling are immune-specific (SI Appendix, Table S3). Right, Profiler² RT-PCR screen of chicken cytokines and chemokines, up-regulated genes. All RT-PCR differentially expressed genes were also differentially expressed in the RNAseq data. (B) Corroboration of the RNAseq data at the protein level. Top, IHC of MPO (myeloperoxidase; scale bar, 50 µm) and LYZ (lysozyme; scale bar, 50 µm) immunostaining of D50 chicken pygostyle IVD cryosections. DAPI nuclei stain with LYZ staining emphasizes detection of LYZ within blood vessels. Bottom: Left, AP activity stain of a D50 fusing pygostyle disc cryosection (scale bar, 200 µm); Right, anti-RIPK3 western blot and corresponding bar graph, showing that RIPK3 is more highly expressed in fusing pygostyle discs than in free discs, and its expression is inhibited upon prednisolone treatment.

Fig. 5.

Prednisolone inhibits pygostyle fusion. Chicken, panels in sagittal view, distal to the Right. (A) Ankylosis inhibition. Top, microCT of 8-wk-old control (Left) and prednisolone-treated (Right) pygostyles. Red arrows indicate IVDs. Below, corresponding ABPRH-staining of the same respective conditions; the asterisk indicates the fused distalmost IVD region; scale bar’s, 2 mm. Note that pygostyle discs are thinner than free vertebrae discs. (B) Prednisolone inhibition of cell death and inflammatory events. Left, TUNEL analysis of (Left to Right) 3-wk-old controls, 7- to 8-wk controls, and 7- to 8-wk prednisolone-treated IVDs. In the next three panels (all at 8 wk of age), compared to untreated controls (Left, Scale bar, 100 µm), prednisolone treatment (for 5 wk) inhibits FCT expansion (Middle, Scale bar, 100 µm) and blood cell extravasation into the FCT (Right, Scale bar, 50 µm); eosin/SYTOX green staining; images show red-stained FCT below the NP and green/yellow-stained nuclei. (C) Prednisolone effects on NP viability and debris clearance. Prednisolone treatment preserves NP cells in both pygostyle and free discs in the 8-wk-old pygostyle (Left 2 panels; compare to the acellular chicken NP in Fig. 2A; scale bar’s 100 µm and 50 µm, respectively); ABPRH staining. Prednisolone treatment (7 wk of age/4 wk prednisolone) also causes accumulation of cell debris in free caudal IVDs in AF tissue compared to 7-wk-old controls (Right 2 panels); ABPRH staining; scale bar’s 100 µm.