Scale diversity in bigeye tuna (Thunnus obesus): Fat-filled trabecular scales made of cellular bone

1 INTRODUCTION

Species in the genus Thunnus are large, schooling, regionally endothermic fishes that inhabit the open ocean, where they regularly migrate long distances to utilize seasonal abundances of food in different areas (Block et al., 2011; Carey & Teal, 1966). Thunnus spp. also support some of the most economically important global fisheries (Havice & Campling, 2010). The seven species in the genus Thunnus have been the subject of study due to their migratory life history habits (Block et al., 2005; Itoh & Tsuji, 2003; Madigan et al., 2014), morphological adaptations to continuous swimming (Bernal, Dickson, Shadwick, & Graham, 2001; Graham, Koehrn, & Dickson, 1983; Kishinouye, 1923; Potthoff, 1975; Westneat, Hoese, Pell, & Wainwright, 1993), and endothermic physiology (Brill, Dewar, & Graham, 1994; Carey & Teal, 1966, 1969; Carey, Teal, Kanwisher, & Lawson, 1971; Graham & Dickson, 2004; Holland, Brill, Chang, Sibert, & Fournier, 1992). Despite this continued biological interest in Thunnus spp. and the important role that teleost fish surface structure may play in locomotion by interfacing with the fluid environment (Burdak, 1986; Lauder et al., 2016; Wainwright, Lauder, & Weaver, 2017), there is little detailed information on the structure of skin and scales in any Thunnus species.

In teleost fishes, most scales occur as single ossified plates that overlap to create an imbricating pattern on the surface of the body. These bony scales are covered with a layer of epidermis and mucus (Sire & Akimenko, 2004), both of which provide an important immune function for fishes (Shephard, 1994; Xu et al., 2013; Zaccone, Kapoor, Fasulo, & Ainis, 2001). Scales differ substantially among species (Roberts, 1993; Wainwright et al. 2017) and may show considerable variation among regions of the body within the same species (Dapar, Torres, Fabricante, & Demayo, 2012; Wainwright & Lauder, 2016). These differences can include scale shape, the presence of spines or other projections, the presence of ctenii (small interlocking spines that form as separate ossifications), and other traits such as radii (gaps in the ossification of scales) and circuli (growth rings on the surface of scales). Scales in most teleost species are composed of acellular lamellar bone (lacking osteocytes), which includes plywood-like ossified layers of connective tissue (Sire & Huysseune, 2003; Zylberberg, Bereiter-Hahn, & Sire, 1988). Scales also grow outward from a central focus and thicken as an individual fish grows (Park & Lee, 1988; Schönbörner, Boivin, & Baud, 1979). After initial scale development in teleosts (which happens in post larval stages), scale size increases with fish growth, but scale number does not increase (Taylor, 1916). Scales are replaced if they are lost, and these replacement scales quickly grow to match the size of neighboring scales (Sire & Akimenko, 2004).

Teleost scales are biologically important in a number of ways. Although specific functions have not been conclusively proven, scales likely provide protection against predators and parasites (Browning, Ortiz, & Boyce, 2013; Vernerey & Barthelat, 2014), and serve as calcium stores (Parenti, 1986). Experimental studies on fish scale function have mostly focused on possible armor-like functions and the ability of scales to resist puncture while organized into a flexible protective surface (Bergman, Lajeunesse, & Motta, 2017; Browning et al., 2013; Duro-Royo et al., 2015; Ghosh, Ebrahimi, & Vaziri, 2014; Song, Ortiz, & Boyce, 2011; Vernerey & Barthelat, 2014). However, it has also been hypothesized that scales influence the hydrodynamics of swimming fish surfaces (boundary layer flow in particular) by either directly interacting with the water, or by maintaining an epidermis and mucus layer that interacts with flow next to the fish (Burdak, 1986; Daniel, 1981; Wainwright & Lauder, 2016).

The scales of scombrid fishes are not well studied, although it is well known that that many scombrid genera have a distinctive structure made of enlarged scales, called the corselet, on the anterior half of the body and posterior to the gill opening (see tuna diagram in Figures 1 and 2 [Collette & Nauen, 1983]). Corselet morphology and presence has been qualitatively described for different genera and species in the scombrid group (Collette, 1978; Collette & Nauen, 1983; Kishinouye, 1923) but only general descriptions of corselet scale morphology exist. Collette describes corselet scales as “enlarged” compared to scales posterior on the body and provides descriptions of scales and corselets for each of the different groups in his classification of tunas and mackerels (Collette, 1978). In Kishinouye's (1923) description of Thunnus morphology, scales are described as filled with “dentritic lumen,” although this observation is not elaborated on. Corselet scales from Thunuus spp. are also embedded under a thick epidermis, and while all fishes have an epidermal layer on their scales, tuna epidermis covering the corselet appears to be thicker compared to most other teleost fishes (Kishinouye, 1923). In a paper including a list of teleost species with acellular and cellular bone (although without any figures or other evidence), Kölliker (1857) describes corselet scales from Thunnus species as being made of cellular bone—whereas most other perciform fishes have acellular bone. These previous comments on Thunnus spp. scales describe interesting features (dentritic lumen, cellular bone) in need of elaboration and further study. Given the multiple adaptations Thunnus spp. have to their pelagic lifestyle and endothermic physiology (Block & Stevens, 2001), it is reasonable to hypothesize that their scales may show specializations compared to the typical teleost scale pattern. By studying these modifications, we may gain insight into both how the skin of tuna functions and what features of teleost scales and skin are functionally important.

In this study, we investigate the morphology of scales in bigeye tuna (Thunnus obesus, Lowe, 1839) at ten regions of the body in five individuals of similar size. We use micro computed tomography (µCT) scanning and histological methods to image and quantify the morphology of scales in this species, and we present both qualitative and quantitative results on the diversity of scales that we find in bigeye tuna. Our results document the presence of considerable scale structure variation across the body and the occurrence of remarkable thickened scales composed of cellular bone with distinctive internal morphological features (trabeculae and lipid inclusions) not known from other teleost species (Figure 1). These results contribute new information about the skin and body surface to the suite of known specializations for pelagic life that characterize Thunnus species (Graham & Dickson, 2004).

2 MATERIALS AND METHODS

2.1 Specimen acquisition and sampling

We obtained five bigeye tuna (Thunnus obesus, Lowe, 1839), mean fork length 78.7 cm (1.87 cm standard deviation), from Tropic Fish Hawaii, LLC (Honolulu, Hawaii). Based on their fork length, these fish are approaching maturity (Nootmorn, 2004). Bigeye tuna was chosen as the focus of this study due to the availability of multiple individuals of similar size. Each fish had the internal organs and gills removed at the time of measurement, and all fish were of the same age class (based on their small range of fork lengths). The skin of each fish was sampled at nine different regions of the body (Figure 2) by excising a 4 × 4 cm region of skin with scales intact from each sampling location. In two individuals, it was difficult to find intact scales on one side of the body due to damage during capture, so some samples were taken from different sides of the body. In total, 45 samples of skin and scales (9 samples from each of 5 individuals) were frozen at −20° C prior to µCT scanning. After sampling we discovered that there are two distinct scale types (termed “thick” and “thin”) at the tail region (Figure 3), and we thus treated these different types as distinct regions in our measurements and analyses below, creating a total of ten regional locations instead of the original nine we sampled (see Figure 2 for regions). For brevity, we will refer to these regions throughout the manuscript using the names we give them in Figure 2: cheek, dorsal, fairing, ventral, belly, dorsal fin, 2nd dorsal, posterior, tail thin, and tail thick regions.

The fairing is a ridge on the lateral side of bigeye tuna (and other Thunnus species) that starts at the leading-edge origin of the pectoral fin and continues down the length of the corselet down the body. The fairing ridge allows the leading edge of the pectoral fin to sit against it when the pectoral fins are held against the body—this streamlines the body by creating a ridge and pocket that the pectoral fins fit into that smooths the external body surface (Kishinouye, 1923; Walters, 1962).

3 RESULTS

3.1 µCT-observations

Bigeye tuna have large, thickened scales on parts of their bodies (Table 1, Figures 1 and 2, supporting information Video 1). These thickened scales (Figure 4) are markedly different from the typical flattened scales of most other teleosts (including bigeye tuna scales from other regions of the body) which are composed of bony laminae with a solid internal structure (Figure 1). The thickened scales of T. obesus are present mostly posterior to the gill opening around the pectoral fin, and make up the corselet (Figure 2) while other parts of the body have smaller, thinner scales of lamellar bone (Table 1, Figure 2; Collette & Nauen, 1983; Kishinouye, 1923). The thickened scales of bigeye tuna are 0.7–1.2 mm thick, which is thicker than both typical teleost scales (0.06–0.2 mm thick [Bergman et al. 2017; Wainwright & Lauder, 2016]) and scales on the posterior part of the body of bigeye tuna (0.09–0.27 mm thick, Table 1). Some parts on the head of bigeye tuna are scale-less (such as the dorsal side of the head) but the body of the tuna (posterior to the opercle) is covered in scales.

Table 1. Dimensions of scales from different body regions. SE refers to one standard error of the mean. Named regions correspond to those in Figure 2

	Mean	SE	Mean	SE	Mean	SE
	length	length	height	height	thickness	thickness
Region	(mm)	(mm)	(mm)	(mm)	(mm)	(mm)
Cheek	28.21	1.51	4.20	0.15	1.19	0.15
Dorsal	10.52	1.47	8.37	0.76	0.72	0.07
Fairing	14.63	0.63	6.68	0.35	0.93	0.06
Ventral	4.54	0.94	3.36	0.47	0.27	0.07
Dorsal fin	3.89	0.46	2.57	0.36	0.24	0.04
Belly	2.84	0.21	3.31	0.19	0.14	0.02
2nd dorsal	3.46	0.46	3.20	0.2	0.12	0.02
Posterior	2.98	0.13	2.00	0.07	0.15	0.01
Tail thin	4.83	0.65	0.95	0.09	0.09	0.01
Tail thick	3.95	0.53	0.38	0.05	0.14	0.01

In bigeye tuna, thickened scales have an ossified exterior layer of lamellar bone that forms the exterior and interior sides of the scale and surrounds an internal cavity made of bony trabeculae surrounded by fat cells (Figures 1 and 4: described in more detail below). Furthermore, these scales do not have the typical growth lines (circuli) or gaps in the ossification of their anterior fields (radii) apparent in other teleost scales (visible in the bluegill scale shown in Figure 1). The thickened scales of the corselet have spindle-shape dorsoventral cross-sections, trabeculae crossing their interiors, an asymmetrical anteroposterior cross-section, and weak ossification at their edges, especially at the posterior edge (Figure 4). Thickened scales also occur on the cheek (exterior to the adductor mandibulae muscles) and tail (exterior to the hypural plate) regions of bigeye tuna (Figure 2). Scales at the cheek are elongate in anterior-posterior axis of the body and are thicker (mean thickness: 1.2 mm, Table 1) and larger (mean length: 28.2 mm, Table 1) than scales elsewhere on the body. At the tail region, however, scales of two types are present: thickened scales with an internal cavity, and thin, plate-like scales (Figure 3). Both of these scale types in this region are relatively small (mean length: 4.4 mm, Table 1), while also being elongated in a direction parallel to the anterior-posterior axis of the body. Overall, the morphology of thickened scales across the body is qualitatively diverse in both size and shape (Figure 2, Table 1). Transitions between thick and thin scales can be seen in images from the sampling regions labeled ventral and tail (Figure 3).

3.3 Histological results

The hematoxylin and eosin stained sections show the presence of adipocytes inside the thickened scales as transparent globular structures (Figure 6a). The presence of fat inside the thickened scales was confirmed with an osmium tetroxide fixation and stain of lipids (Figure 6b). A layer of fat cells is also present beneath the dermis in body regions with thin scales (Figure 5). The presence of subdermal fat cannot be confirmed beneath the corselet because our sections are not deep enough.

Histological sections also reveal both arterioles and venules among the fat cells (Figure 7), which might serve to deposit or mobilize lipids in the scale interior. These arterioles and venules are consistent with the size and morphology of the same structures in other vertebrate taxa (Bloom & Fawcett, 1975; Patt & Patt, 1969). Also, lacunae (30–50 µm in diameter) are present in the external bone of the thickened scales (Figure 7e,f), and likely correspond to locations where the circulatory system penetrates the bony scale exterior. Within the lamellar bone covering each thickened scale (as well as the trabeculae: Supplementary Figure 1), smaller lacunae (about 10 µm in diameter) contain osteocytes and reflect the presence of bone (Figure 7e,f). These small lacunae and osteocytes are consistent with those seen in other vertebrate taxa (Atkins et al., 2014; Bloom & Fawcett, 1975; Patt & Patt, 1969). Thin scales occur on body regions other than the corselet (Figure 2) and are plate-like with a lamellar structure. These scales also have small lacunae containing osteocytes at scale margins, indicating the presence of cellular bone (Figure 7g,h).

4 DISCUSSION

Analysis of scale morphology in bigeye tuna shows that this species has modified scales compared to typical flattened ctenoid and cycloid scales in many other teleosts (Figure 1, supporting information Video 1). These thickened scales are composed of bony outer shells filled with fat and braced by trabeculae. In addition, the bone of both thick and thin scales is made of cellular bone. Arterioles and venules are also visible among the fat cells inside the thick scales. These thickened, fat-filled scales occur on the cheek, at the base of the caudal fin, and as part of the corselet (posterior to the gill opening).

On much of the posterior regions of the body, bigeye tuna have thin plate-like scales that are superficially similar to the scales of most perciform fishes (Parenti, 1986; Sire & Akimenko, 2004), and these scales are also composed of cellular bone at their margins. Bigeye tuna have remarkably diverse scales among the ten sampled regions of their body. Skin samples from different regions differ with respect to their scale volume, total scale layer thickness, and scale aspect ratio, indicating that they vary in size, shape, and the extent of overlap between scales (Figure 2). In general, larger scales are located in more anterior body regions, while smaller scales are found on the posterior half of bigeye tuna.

The thickened, fat-filled scales of the corselet and cheek are remarkable in their morphology, but no data on scale function in tuna are available. Hypotheses of the function of these modified thickened scales and of variation in scale morphology across the body are thus necessarily somewhat speculative. Despite this, if we restrict ourselves to first principles of the composition and structure of the modified scales, we can confidently state a few functional implications using hypothetical comparisons. First, thickened scales will protect the body surface from damage more so than smaller flattened scales or no scales. Second, fat-filled scales will increase buoyancy compared to scales of the same size but made entirely of bone. Third, fat-filled scales will provide better thermal insulation by decreasing conductance across the skin compared to similar-sized bone scales, or no scales. Subdermal fat (Figure 2) would provide additional thermal insulation—another endothermic fish, the opah, is described as having an exceptionally thick subdermal layer of fat around their body (Wegner, Snodgrass, Dewar, & Hyde, 2015). Although we can be confident in the action of these physical mechanisms, we do not know if these mechanisms have large enough effects on bigeye tuna to be functionally important. Further studies including experimental measurements of scale thermal and material properties, and studies comparing scale morphology across taxa in resolved scombrid trees (Miya et al., 2013; Santini, Carnevale, & Sorenson, 2013) would inform potential functions of thickened scales in protection, buoyancy, and thermal insulation. Although Thunnus spp. have the largest and most well-developed corselets (where we see the modified scales), non-endothermic scombrid genera such as Sarda are also described as having corselets (Collette & Nauen, 1983; Kishinouye, 1923), although there are no publications discussing details of their scale morphology.

Previous studies have suggested that corselets reduce drag in tuna swimming by tripping the boundary layer from laminar to turbulent to delay boundary layer separation (Blake, 2004; Fierstine & Walters, 1968; Walters, 1962; Webb, 1975). Although no study has experimentally examined flow over this region during swimming, we believe this is an unlikely function of the corselet because both the enlarged scales of the corselet and the smaller scales posterior to the corselet are coated in both epidermis and mucus. Covering scales with epidermis and mucus has an effect of smoothing the topography of the scales in fishes (Wainwright & Lauder, in press; Wainwright et al., 2017), which would make it difficult to trip the boundary layer from laminar to turbulent. Furthermore, the end of the corselet (Figure 2: fairing) does not create the type of roughness element (a sharp drop or bump) that is required to trip a boundary layer from laminar to turbulent. To support this, the cross-section of skin from the ventral region (Figure 2) illustrates the gradient of scale morphology at the edge of the corselet – this smooth transition is unlikely to trip the boundary layer.

The thickened scales of the corselet also give form to the fairing, which is a ridge that is dorsal to the pectoral fin and runs posteriorly down the body to the edge of the corselet—it allows the pectoral fin to be adducted against the body with its thickened leading edge against the ridge of the fairing, streamlining the contours of the body and pectoral fin (Kishinouye, 1923; Walters, 1962). This fairing is constructed using the thickened scales of the corselet (fairing region, Figure 2) where more scales are layered on top one another, and where scales are thinner in the depression for the pectoral fin, ventral to the fairing ridge. In this way, the corselet scales may contribute indirectly to drag reduction by providing streamlining for adducted pectoral fins. However, scales are not required for creating fairing-like structures, as similar structures are seen near the pectoral fins of beaked whales (Dalebout et al., 2003; Mead, Walker, & Houck, 1982). Therefore, it is difficult to assign a purely drag reductive role to the modified scales we have described, because as beaked whale morphology shows, construction of a fairing does not require scales.

Thunnus spp. have often been described as having a specialized undulatory swimming mode, called thunniform swimming (Breder, 1926; Webb, 1975) which is characterized by low amplitude bending in the head and body and higher amplitude bending in the peduncle and tail. Although thunniform swimming is often cited as a specialized case of swimming kinematics where lateral head amplitude (yaw) during swimming is exceptionally low, one study shows that the lateral head amplitude of thunniform swimmers is no different than carangiform swimmers (Donley & Dickson, 2000). Similarly, another study comparing midline kinematics of tunas to other species shows little difference in the ratio of head-to-tail amplitude in lateral bending between tunas and other fishes (Lauder & Tytell, 2006). Although one may propose that the thickened scale layer of the corselet creates higher stiffness in the anterior body reflective of thunniform swimming, kinematic data do not show any reduced lateral movement in the anterior body for swimming tuna, and thus do not support either the idea that thickened scales contribute to higher body stiffness at the head, or that thunniform swimming has smaller head oscillation amplitude than other swimming categories.

The function of the cellular bone and circulatory system elements (Figure 7) within thickened scales is unknown, but both anatomical features suggest a higher metabolic rate in these scales (compared to scales without either feature) and that calcium and lipids may be mobilized for metabolic needs. Although a circulatory system is known to be present on the surface of fish skin and on the surface of scales in teleost fishes (Farrell, Eliason, Clark, & Steinhausen, 2014; Rummer, Wang, Steffensen, & Randall, 2014), the circulatory elements identified inside bigeye tuna thick scales are, to our knowledge, unique among teleost fishes. Circulatory control of deposition and mobilization inside thickened scales would allow for control of lipid use with growth or during migratory periods with limited feeding where additional energy may be needed (as in salmonids [e.g., Penney & Moffitt, 2015]). Fat deposition in the scales and below the skin in tuna could also function as a reservoir for gamete production. Analyzing the fat content of thick scales in bigeye tunas under stress, starvation, or spawning conditions could determine if interior lipids are mobilized to meet metabolic needs, and determining what lipids are being stored would inform potential metabolic or buoyancy-related functions (Phleger, 1998).

The presence of cellular bone in both thickened and thin bigeye tuna scales warrants further investigation, as current comparative data suggest that “basal” teleosts and teleosts with ganoid scales have cellular bone in their scales, while most perciform fishes possess acellular bone (Meunier, 1981; Meunier & Brito, 2004; Parenti, 1986). Previous work indicates that tunas have cellular bone (our histological work confirms this for bigeye tuna), while the closely related mackerels (Scomber spp.) have acellular bone (Kölliker, 1857; Meunier, 2011). Most teleosts have acellular bone not only in their scales, but also in their body skeleton, yet we know little about the functional implications of having cellular versus acellular bone in fishes (Cohen et al., 2012; Moss, 1961; Shahar & Dean, 2013). It has long been thought that acellular bone is a derived condition in fishes, with euteleosts having a skeleton of acellular bone compared to cellular bone in more basal taxa (Meunier, 2011; Parenti, 1986). However it has also been shown that many euteleosts have cellular bone in parts of their skeleton (Hughes, Bassett, & Moffat, 1994) and that there are many transitions from acellular to cellular bone throughout teleosts (Kölliker, 1857; Kranenbarg, van Cleynenbreugel, Schipper, & van Leeuwen, 2005; Meunier & Huysseune, 1992; Moss, 1961, 1965; Sire, Huysseune, & Meunier, 1990).

In fishes, the function of acellular versus cellular bone has not been determined, although some hypotheses have been tested and found to have little support. For example, studies have found that acellular bone has similar material properties to cellular fish bone (Cohen et al., 2012; Dean & Shahar, 2012; Horton & Summers, 2009), acellular bone can remodel itself despite lacking enclosed cells (Atkins et al., 2014; Kranenbarg, et al., 2005), and cellular bone does not appear to be more prevalent in freshwater fish species, which may need to draw on calcium stores more often than saltwater fishes (Moss, 1965). Perhaps the presence of cellular bone in the large scales of bigeye tuna assists in rapid scale growth or remodeling, as well as rapid mobilization and deposition of calcium. Large, pelagic, fast-growing fishes may have higher needs for calcium and bone remodeling, especially during long migrations, necessitating cellular bony scales. One study has demonstrated bone resorption in fin spines of Atlantic bluefin tuna (Thunnus thynnus) (Santamaria et al., 2015) and more work on different bones of other scombrid species could determine if tunas have specializations for increased bone remodeling or resorption, compared to Scomber spp. (close relatives with acellular bone). Curiously, the osteocytes in the bone of the thin scales of Thunnus appear to be concentrated near scale margins, where remodeling and growth would also be concentrated (Figure 7g,h). However, there is no apparent pattern to osteocyte distribution in the thickened scales (Figures 6 and 7). Leatherback turtle (Dermochelys coriacea, Vandelli 1761) bones have a similarly modified growth pattern, perhaps due to high growth rates, large adult sizes, and endothermy (Rhodin, 1985) and the modified bone structure seen in bigeye tuna could be a reflection of some of these characteristics as well.

We have discovered considerable diversity in simple measures of size and shape among scales from different regions of the body of bigeye tuna and a remarkable trabecular structure in thickened anterior scales. Generally, larger scales occur on the cheek and corselet of bigeye tuna, while smaller scales occur posteriorly on the body. The larger scales of the cheek and corselet, along with some smaller scales at the base of the tail, have a distinct bony shell made of cellular bone and an internal structure of fat cells, trabeculae, and circulatory elements. These scales represent an interesting departure from normal teleost scales which consist of flat sheets of acellular bone, and the function of these modified scales represents an intriguing area for future research.

AUTHOR CONTRIBUTIONS

DKW and GVL conceived the research. DKW and SI led data collection and DKW led the writing of the manuscript. All authors contributed to the manuscript and gave approval for publication.