In situ SEM/EDS compositional characterization of osteocytes and blood vessels in fossil and extant turtles on untreated bone surfaces; different preservational pathways microns away

Introduction

Bone is a complex biological tissue that characterizes extant and fossil vertebrates, and consists of a mineralized (calcium, phosphorus) and a non-mineralized (collagen and non-collagenous proteins) extracellular matrix, plus water and some lipids (Boskey & Gehron, 2013; Rey et al., 2009). Cells involved in bone tissue are osteoclasts, osteoblasts, and the most abundant of them osteocytes (Bonewald, 2011). Osteocytes are embedded within the hard-mineralized component of bone throughout life (exceptions being when released by fracture or during remodeling) (Robling & Bonewald, 2020), providing them high preservation potential within fossil bones, which has been extensively documented in different clades of vertebrates (e.g., Bailleul, O’Connor & Schweitzer, 2019; Enlow & Brown, 1956; Pawlicki & Nowogrodzka-Zagorska, 1998; Schweitzer, 2011; Schweitzer et al., 2013; Surmik et al., 2019). Similar preservation of osteocytes- and blood vessels-like has also been documented in fossil turtles, showing that their preservation is independent of geologic time, paleoenvironment, lithology, lineages, and latitude (Cadena, 2016; Cadena, Ksepka & Norell, 2013; Cadena & Schweitzer, 2012; Cadena & Schweitzer, 2014).

Something in common to all aforementioned studies are the analytical tools used to study and characterize these fossil bone microstructures, which include principally: (1) ground sections and observation under transmitted and polarized microscopy (Cadena & Schweitzer, 2012; Surmik et al., 2019); (2) bone demineralization using ethylenediaminetetraacetic acid (EDTA) as a chelating agent (0.5 M, pH 8.0), facilitating release the osteocytes-, blood vessels-, and any other cells- or soft-tissue fibers-like from the bone matrix for their posterior study by transmitted and/or polarized light, scanning and/or transmission electron microscopy and any coupled elemental analyzer, Raman spectroscopy, Fourier-transform infrared spectroscopy (FTIR), immunological and antibody studies (e.g., Alfonso-Rojas & Cadena, 2020; Bailleul, O’Connor & Schweitzer, 2019; Bailleul et al., 2020; Cadena, 2016; Saitta et al., 2019; Schweitzer et al., 2013; Surmik et al., 2019; Wiemann et al., 2018).

The preservation of these soft-tissue microstructures (osteocytes and blood vessels) and their potential original constituents (proteins and DNA) has been questioned and considered a consequence of microbial interactions within fossil bone and its microenvironment or even as a result of cross-contamination in the laboratory (Buckley et al., 2017; Kaye, Gaugler & Sawlowicz, 2008; Saitta et al., 2019). The ‘biofilm hypothesis’ as a source for soft-tissue preservation in dinosaur bones has been rigorously tested, which identified fundamental morphological, chemical and textural differences between the resultant biofilm structures and those derived from dinosaur bone, demonstrating that the recovered microstructures in the reports cited above are endogenous in origin and that the ‘biofilm hypothesis’ should therefore be rejected (Schweitzer, Moyer & Zheng, 2016). Issues concerning cross-contamination and replications, timing of sample collections, and reagents have also been addressed by Schweitzer et al. (2019).

Compositionally, the osteocytes- and blood vessels-like from different clades of fossil vertebrates have been shown to commonly be enriched in iron (Cadena, 2016; Schweitzer et al., 2014; Surmik et al., 2019; Ullmann, Pandya & Nellermoe, 2019), an element that has been suggested to play a key role in preserving and even masking identification of proteins in fossil tissues via Fenton reactions (Schweitzer et al., 2014). Other elements typically found in these fossil bone microstructures are carbon, calcium, and silicon (Cadena, 2016; Ullmann, Pandya & Nellermoe, 2019). At present, all these studies of elemental characterization have been conducted using SEM/EDS on isolated (post-demineralization) osteocytes- and blood vessels-like, or from polished ground sections, which implies some degree of manipulation or contact with reagents or preparation tools, potentially raising skepticism on the elemental results.

Here, I explore the in situ (directly on fresh and untreated surfaces) preservation and elemental composition of bone microstructural elements (cells and blood vessels) of fossil turtle bones from three localities which have completely different geological settings (lithological, taphonomic, and fossil diagenesis), including: (1) Gobi Desert, Mongolia, from the Late Cretaceous (late Campanian–early Maastrichtian); (2) Messel Pit, Germany, from the Eocene; and (3) La Venta fauna, Colombia, from the Miocene. Comparison samples include bone from two extant turtles and a domesticated chicken. I discuss herein the results of these analyses and the advantages of using in situ SEM/EDS for understanding preservation of cells/tissues in fossils.

Materials & Methods

Results

Allaeochelys crassesculpta, Eocene of Germany

The most abundant bone microstructures preserved in this sample are blood vessels-like and the walls that formed the Haversian-Volkmann-like (H-V) canals; also, in some, there is evidence of very small (2.5 µm diameter) structures with a striated margin which resemble the morphology of osteoblast cells (Figs. 4A–4D; Fig. S2). The blood vessels-like exhibit a width of 1–3 µm, with an average wall thickness of 0.2 µm (Fig. 4D). Compositionally, the blood vessels-like are mainly composed of carbon and nitrogen, with minor amounts of calcium, phosphorus and iron (Fig. 2B; 4E–4G; 4J –4N; Fig. S2; Data S2). The bone matrix surrounding them lacks nitrogen and carbon, and it is exclusively characterized by calcium, phosphorus, and iron (Figs. 4E, 4I). A bone sample with rock matrix attached shows that the bone is composed of calcium, phosphorus, iron, and nitrogen, and, in contrast, the rock matrix is rich in aluminum and silicon (Figs. 2C–2D; Figs. 4O–4Q; Data S2).

Podocnemidid indet, Miocene of Colombia

The sample of the side-necked turtle from La Venta, Colombia, shows on the bone external cortex preservation of walls that formed the H-V-canals, blood vessels- and osteocytes-like tightly embedded in the very homogenous bone matrix (Figs. 5A–5B; 5F–5G). Elementally, the blood vessels-like and H-V-canal walls are rich in carbon, nitrogen, and calcium, with minor amounts of phosphorous and silicon (Figs. 2B; 5C–5D; 5H–5I; Data S2). In contrast, the osteocytes-like are composed of iron, calcium, aluminum, manganese, phosphorus, and minor amounts of silicon (Figs. 2B; 5H–5J; Data S2). The bone matrix lacks carbon and nitrogen, and it is constituted by calcium and phosphorus mainly (Figs. 2C; 5C, 5E 5H). An isolated bone fragment (post-demineralization) shows some of the osteocytes-like still embedded in the matrix, varying in color from orange to black, the darker ones located closer to black, dendritic mats (Figs. 5K–5M).

In situ extant turtle and chicken bone microstructures

The carapace bone fragment of the extant side-necked turtle Podocnemis lewyana shows osteocytes within lacunae (Figs. 6A–6B). Their composition is rich in carbon, nitrogen, calcium, and phosphorus (Fig. 2A; 6C–6D; Data S2). The bone matrix is relatively richer in calcium (Fig. 2C; 6C, 6E). The H-V canals exhibit a distinct wall and a high concentration of blood vessels and red blood cells, which are rich in carbon and nitrogen (Fig. 2B; 6F–6G; Figs. S3; S4; Data S2). Similar spatial patterns and composition are shared by the bone of the extant marine turtle Lepidochelys olivacea (Fig. 2; 6H–6M; Data S2), and the bone of Gallus gallus (chicken) (Fig. 2; 6N–6P; Data S2).

Discussion

As previously shown (Cadena, 2016; Schweitzer et al., 2014; Surmik et al., 2019; Ullmann, Pandya & Nellermoe, 2019), the in situ analyses presented here, concur with that iron is a very common constituent of fossil osteocytes-like, such as those found in the Late Cretaceous Mongolemys elegans and the Miocene podocnemidid indet. bone samples studied herein (Figs. 1 and 5). However, this composition is not always homogenous and may vary between the external and the internal layer of osteocytes-like, as shown in a broken and folded osteocyte-like from M. elegans, which exhibits richer content of manganese internally and iron externally (Figs. 1O–1P). High levels of manganese were also detected in osteocytes-like from the Miocene side-necked turtle from Colombia, indicating that besides iron as initially suggested by Schweitzer et al. (2014), manganese may also be involved in the preservation of these bone microstructures in deep time. The source for this rich content of manganese seems to be from manganese oxides such as pyrolusite penetrating bone microfractures, which were found herein in some fragments of the Miocene podocnemidid indet. from Colombia (Figs. 5K–5M), and also has been characterized to occur in dinosaur fossil bones from the same Nemegt Formation, from which the M. elegans studied herein was collected (Owocki et al., 2016). The color variation exhibited by the fossil osteocytes-like of M. elegans and podocnemidid indet. seems to be related to enrichment of manganese, higher their manganese content darker their color. In contrast to the osteocytes of the extant turtle and chicken bone, which are rich in carbon and nitrogen (Figs. 2 and 6), these elements only appear in minor amounts in fossil osteocytes-like. However, these cells exhibited a very distinct composition when compared to the surrounding bone matrix and even the wall surfaces of their osteocytes-lacunae, indicating that their mineralized preservation occurred at micro-scale inside the bone, a hypothesis that should be tested by future studies using additional tools (e.g., Raman and FTIR spectroscopy).

The blood vessels- and H-V canal walls-like preserved in the Eocene Allaeochelys crassesculpta from the Messel Pit and the Miocene podocnemidid indet. specimen from the La Venta not only exhibited a similar morphology, but also exhibited the same elemental composition as their corresponding tissues in extant turtle and chicken bone. (Figs. 2, 4 and 5, Figs. S2– Figs. S4; Data S2). In both cases (extant and fossils) being rich in carbon and nitrogen, and differing from the surrounding bone matrix which is richer in calcium and phosphorus, or the rock matrix which is rich in silicon and aluminum (without any traces of carbon, calcium, or nitrogen) which suggests that carbonates or nitrates were absent in the surrounding microenvironment. The in situ measurements performed on some of the preserved blood vessels-like from A. crassesculpta, exhibiting uniform fabric and thin walls of 0.2 µm thickness (Fig. 4D) suggest that they are not consistent with the characteristics of biofilms, which tend to be amorphous and larger in diameter (Schweitzer, Moyer & Zheng, 2016). Blood vessels constitute one of the most promising microstructures preserved in fossil turtles for molecular paleontology studies, and future studies should focus on their molecular in situ characterization using ToF-SIMS mass spectrometry, similarly as it has been used in dinosaurs and other fossil vertebrates (Alfonso-Rojas & Cadena, 2020; Henss et al., 2013; Lindgren et al., 2018; Schweitzer et al., 2019).

For the first time, I herein report the preservation of osteoblasts-like in fossil vertebrates, particularly in the Messel Pit turtle A. crassesculpta. They occur as oval objects with striated margins that are attached to the H-V canals (Figs. 5B–5D; Fig. S2), and thus resemble the morphology and size of osteblasts observed in electron micrographs of human bone (Nakamura, 2007; Schmidt et al., 2002). At the same time, evidence here provided from the Miocene podocnemidid indet. turtle from Colombia (Figs. 5F–5J) shows that, in the same bone specimen, osteocytes- and blood vessels-like that are only 20 µm away from each other are compositionally different. This indicates that each microstructure went through a different preservational pathway. Osteocytes-like seem to be more mineralized than blood vessels-like in these fossil samples, with high amount of iron and manganese, and less organic components than blood vessels-like (Figs. 2 and 3). In the extant bone of turtles and chicken, osteocytes and blood vessels exhibit similar elemental composition under SEM/EDS, both being rich in carbon and nitrogen, which are typically present in abundance within proteins (Torabizadeh, 2011) (Figs. 2 and 6). A similar composition was detected herein in fossil blood vessels from A. crassesculpta from Germany and the podocnemidid indet., from Colombia (Figs. 4 and 5).

Traditionally, it has been suggested that SEM/EDS has to be performed on homogenous or polished surfaces to avoid topographic effects on EDS analyses (Goldstein et al., 2003). However, as I showed here, such effects were negligible for the analyzed samples with composition and signal intensities being very similar in both the fossil and extant samples (Fig. 2). I therefore suggest that a more critical condition for EDS analysis on untreated samples is acquire the highest maximum count rate possible; above 1 million counts is ideal.

Conclusion

This study provided evidence that in situ analyses using a conventional technique, SEM/EDS, on untreated fresh surfaces of fossil and extant bones constitutes a protocol that should be added to the rigorous plethora of proxies and tools (e.g., those recently reviewed and summarized by Schweitzer et al. (2019) to support and demonstrate the preservation of cells, soft-tissues and their original constituents in deep time. Furthermore, in situ analyses of fossil and extant bone samples may also help eliminate any potential skepticism of results obtained by molecular paleontology studies, because, as demonstrate here, it requires minimal sample preparation/manipulation, use of reagents, or contact with lab tools that could cause possible contamination.

Supplemental Information