Sampling was conducted at Colola Beach, located in the municipality of Aquila on the Pacific coast of Michoacán, Mexico (approx. 18˚18'N, 103˚26'W; Figure 1). The geographic, physical, and magnetic substrate featured by Colola Beach have been described previously [7] [29].

Sampling was performed during the 2010 nesting season. Ten artificial relocation nests located in the central third of the beach were monitored. Following the completion of hatchling emergence, non-hatched eggs (n = 125) were excavated. Embryos were extracted using sterile plastic tools to avoid metallic contamination and thoroughly rinsed with filtered distilled water. Extracted embryos were preserved in 4% phosphate-buffered paraformaldehyde at 4˚C for one week prior to tissue dissection.

The 125 excavated, non-hatched eggs contained embryos exhibiting morphological malformations. No-normally developed embryos were recovered from these non-hatched eggs. Consequently, direct analytical comparisons between malformed and unaffected embryos from the same clutches were not possible. Gross phenotypes were photographed and classified. Field collection protocols were authorized by the Secretaría del Medio Ambiente y Recursos Naturales (SEMARNAT Permission No. SGPA/DGVS/04011/08).

Twelve deformed embryos (n = 12) were randomly selected for tissue analysis, ensuring representation from each of the ten artificial nests. Tissue dissections were performed under clean laboratory conditions. Three tissue matrixes were collected per embryo:

1) Muscle: Pooled tissue from the neck, abdominal wall, and flippers.

2) Carapace: Combined central and costal scutes.

3) Bowel: Small and large intestinal tracts.

Tissue samples were rinsed with reptile Ringer’s solution to remove extraneous debris, oven-dried at 80˚C for 12 hours, and weighed ( 100 μ g dry weight). Samples were acid-digested with ultra-pure nitric acid (HNO₃) in sealed polytetrafluoroethylene (PTFE) vessels [30] 30 .

Concentrations of fourteen trace elements ( Al , V , Cr , Mn , Fe , Co , Ni , Cu , Zn , Cd , In , Ba , Pb , and Bi ) were measured by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) [31]. Total mercury was determined by Cold Vapor Atomic Absorption Spectrometry (CVAAS). Analytical quality control was maintained using procedural blanks, standard addition calibrations, and certified reference materials from the National Institute of Standards and Technology (NIST SRM). Detection limits ranged between 0.001 and 0.05 μg/g dry weight, and matrix spike recoveries averaged between 92% and 106%.

Element concentrations were converted to parts per million on a dry weight basis ( ppm ; mg/kg d.w. ) and date were log-transformed prior to statistical analysis. Inter-individual and inter-tissue variations were evaluated using a two-way Analysis of Variance (ANOVA) incorporating embryo identity, tissue matrix, metal type, and interaction terms. Nested observational units were used to account for embryos originating from the same nest. Analyses were executed using FAUANL software (Microsoft Office Excel 2007) and plots were constructed in SigmaPlot 10.0.

Assuming an average clutch size of 70 eggs for C. mydas agassizii, an average of 12 ± 5 non-hatched eggs containing malformed embryos was recovered per artificial nest. This yields an estimated embryonic failure rate of <18% < 18 % per clutch, which aligns with malformation frequencies reported in sea turtle populations elsewhere [14] [32].

The anatomical defects varied considerably among individuals. Observed abnormalities included equal/unequal twinning, bicephaly, abdominal dysraphism, and severe kyphosis (Figure 2). Craniofacial dysmorphias included exophthalmia, anophthalmia, monophthalmia, synophthalmia, arrhinia, micrognathia, and severe jaw malformations (Figure 3). Multiple defects frequently co-occurred within single embryos.

The type of abnormalities in the non-hatched eggs varied considerably. Figure 2 illustrates the abnormal phenotypes encountered in our sample of C. mydas agassiziimalformed embryos. Various craniofacial dysmorphias, abdominal dysraphism and kyphosis were found alone but mostly combined (Figure 3). We note that the embryo malformations are comparable to those previously recorded for other marine turtle species in southern Mexico coastal zones [14].

Trace element concentrations in muscle, carapace, and bowel tissues are presented in Table 1 and Figure 4. The metals exhibiting the highest overall concentrations across all tissue matrixes were Fe , Zn , Al , and Ba . These were followed by Mn , Cu , Cr , V , Ni , and Pb . The lowest concentrations were consistently recorded for Co , Cd , In , Bi , and Hg . More specifically, Fe (195 ppm),

Zn (143 ppm), Ba (37 ppm), and Al (33 ppm) were the elements with the maximum concentration levels in the muscle. Similarly, Fe (246 ppm), Zn (187 ppm), Ba (36 ppm), and Al (25 ppm) with the highest concentrations in the carapace. Lastly, Fe (397 ppm), Zn (127 ppm), Al (21 ppm), and Ba (13 ppm) were those elements with the greater concentration in the bowel.

The two-way ANOVA revealed statistically significant differences in total metal

Table 1. Trace element concentrations in tissues of marine turtles.

concentrations among individual embryos ( F = 22.6367 , df = 14 , p = 0.0001 ), confirming substantial inter-individual variability. Metals accumulated non-uniformly across tissues, demonstrating matrix-specific affinity; iron was preferentially concentrated in bowel tissue, whereas zinc and barium showed higher affinities for carapace scutes (Figure 4).

Heavy metals did not accumulate uniformly in the different tissues. Differences between the muscle, carapace, and bowel tissues suggest that tissue types influence the distribution and retention of heavy metals within the body. Tissues show different capacities for absorbing, storing, transporting, and eliminating contaminants. Concentrations for metals across tissue samples were similar, although they accumulated differently in the muscle, carapace and bowel (Figure 4).

In our survey of C. mydas agassizii embryos exhibiting major anatomical abnormalities, heavy metals accumulated non-uniformly across the analyzed tissue matrixes (Figure 4). Concentrations of Fe , Zn , Al , and Ba were highest across muscle, carapace, and bowel tissues, followed by Mn , Cu , Cr , V , Ni , and Pb , whereas Co , Cd , In , Bi , and Hg were present at lower concentrations.

The differential accumulation across tissue matrixes reflects organ-specific differences in transport mechanisms, buffering capacity, metabolic clearance, and storage dynamics [34]. Muscle tissue reflects active metabolic accumulation; carapace scutes serve as long-term deposition sinks for divalent cations; and the bowel reflects mucosal absorption and embryonic excretion pathways.

On the other hand, individual embryos exhibited marked variations in overall heavy metal burdens. Inter-individual variability may stem from differences in maternal exposure history, foraging ecology, diet composition, trophic position, or maternal physiological condition [25] [27] [34] [36] [37]. Also, because heavy metals enter marine food webs and accumulate over time, foraging adult females encounter heterogeneous contamination levels depending on their specific migratory corridors and feeding grounds [25] [27] [37].

While trace elements like Fe , Zn , Cu , and Mn perform essential metabolic and structural functions during embryogenesis, excess concentrations and exposure to non-essential metals ( Pb , Cd , As , Hg ) disrupt critical developmental signaling cascades. Accordingly, the craniofacial dysmorphias, ocular defects, axial skeletal deformities, and abdominal wall dysraphism observed in Colola Beach embryos coincide with critical organogenesis windows where specific heavy metals disrupt embryonic patterning through well-defined molecular mechanisms [38]-[42]. For instance, embryonic exposure to Cd inhibits metallothionein gene expression, causing severe head and craniofacial malformations [41]. Pb linked to microcephaly, jaw hypoplasia, and facial dysmorphias inhibiting the canonical Wnt / β -catenin pathway, which downregulates facial identity factors ( Pax3 , Sox9 , Sox10 ) in cranial neural crest cells (CNCs). Pb also disrupts retinoic acid ( RA ) receptor signaling across the pharyngeal arches and mimics Ca 2 + in the Ca 2 + / Calmodulin ( CaMKII ) pathway, causing premature neural progenitor cell loss in the rostral neural tube [43]. As , Co , Cu , Fe , and Mn promote hydroxyl radical production through Fenton reactions, inhibiting PAX2 (optic stalk specification) and MITF (retinal pigment epithelium development), which restricts final ocular volume (microphthalmia) [44] [45]. Hg induces anophthalmia and microphthalmia by downregulating the master ocular gene PAX6 and suppressing the Sonic Hedgehog signaling cascade emanating from the precordal plate. Furthermore, Hg binds to sulfhydryl ( − SH ) groups on tubulin, causing microtubule depolymerization that impairs neuroectodermal optic vesicle evagination and triggers p53 -mediated apoptosis in neural progenitor cells [46].

Regarding vertebral, axial, and limb deformities As , Fe and Mn disrupt the “somitogenesis clock” and morphogen gradients of FGF8 and Wnt3a in the presomitic mesoderm [39] [42] [43]. This asymmetrical segmentation of cervical somites causes hemivertebral fusions, resulting in severe torticollis and axial skeletal twisting (twisted neck) As , Co and Cu induce micromelia (shortened limbs) [44] [45] by disrupting reciprocal SHH / FGF signaling between the apical ectodermal ridge ( AER , source of FGF4/8 ) and the zone of polarizing activity ( ZPA , source of SHH ). These metals inhibit focal adhesion kinase ( FAK ) and suppress the BMP/Runx2 cascade, arresting precartilaginous mesenchymal condensation [44] [45]. Pb , Cd and Al interfere with Notch-dependent signaling and sclerotome proliferation during axial patterning leading to extensive vertebral arch malformations, kyphosis, and scoliosis [39] [42] [43].

Although the precise molecular signaling cascades mediating ventral wall and visceral defects remain obscure, it has been shown that excess Zn and the combined metal exposure impair gut rotation and intestinal looping mechanics, resulting in incomplete gut coiling, as well as the disruption of ventral body wall closure during early lateral fold fusion and extracellular matrix remodeling, producing severe everted viscera, gastroschisis, and schistosomus reflexus syndromes [40] [45].

Exposure to heavy metals stimulates the overproduction of reactive oxygen species, inducing oxidative stress and damaging proteins, membrane lipids, and DNA [33]. Embryonic tissues are exceptionally vulnerable to oxidative injury because their endogenous antioxidant defense systems are not yet fully mature. Metals such as Hg, Cd, and Pb damage DNA, disrupt chromatin architecture, and alter developmental gene expression [41] [46] [47]. Furthermore, heavy metals interfere with calcium metabolism during skeletal formation, disrupting cartilage mineralization and resulting in severe bone and shell dysmorphias (see below).

Heavy metals are recognized as teratogenic agents across diverse metazoan taxa, including crustaceans [48], echinoderms [49], anurans [21] [50]-[52], fish [53]-[55], birds [44] [56], and mammals [57] [58]. Even essential trace elements with vital physiological roles can become toxic and teratogenic when systemic concentrations exceed physiological thresholds. The magnitude of fetal deformities is greatest, and overall embryonic viability is lowest, when metal overexposure occurs during critical early windows of fertilization and initial organogenesis [55] [59]-[61].

The transport of heavy metals to developing C. mydas agassizii embryos likely involves a two-stage environmental mechanism [60] [61]. The first stage would be defined by regional transport and substrate absorption. Heavy metals discharged from inland mining operations (e.g., the Aquila iron district) [22] may be transported via river networks and ocean currents to nesting beaches. Sand substrates inundated with contaminated seawater or dust fall allow trace elements to penetrate eggshells via passive diffusion and absorptive processes during incubation [60]-[62]. Transgenerational maternal transfer would then take place. Foraging females exposed to heavy metals in polluted feeding grounds transfer circulating toxicants across the uterine epithelium into the egg yolk during vitellogenesis, directly exposing embryos to toxic loads from early cleavage [25] [27] [60] [62]. Finally, in situ nest substrate absorption would add an additional source of pollutants. Colola beach receives hydrological run-off carrying eroded tailings from the nearby Aquila iron mining district via regional rivers [22]. Trace elements deposited in beach sand can cross the porous eggshell during incubation, altering eggshell gas/moisture permeability and directly exposing embryonic membranes [60]-[62].

Multiple interacting factors include heavy metal contamination, organic pollutants, environmental factors, microbial infections, genetic abnormalities, nutritional deficiencies and habitat degradation [63]-[67]. Controlled experimental studies remain scarce because marine turtles are protected species, making field observations and metal contamination analyses important sources of evidence.

How heavy metals reach the developing C. mydas agassizii embryos might be related to a two-stage mechanism, with regional long-range transport from their sources by rivers and ocean currents and at nesting beaches bathed by contaminated water, heavy metals may pass through the egg shells by adsorptive/absorptive processes [60] [61] [68]. A possibility is that fertile females and/or males exposed to these pollutants while navigating through contaminated waters could transfer blood borne heavy metal to their gametes and/or fertilized eggs [60] [68]. This mechanism focuses on developmental trajectories though genetic/epigenetic factors [69]. Both alternatives are not mutually exclusive and could be experimentally evaluated.

A paradox in our observations is that embryos from the same clutch displayed different types and severities of malformations, while most eggs in these clutches hatched successfully without gross physical defects. If all eggs within a clutch are exposed to similar maternal or environmental contaminant loads, why are only certain embryos affected, and why do phenotypes vary so widely? This idiosyncratic embryonic response suggests that heavy metal teratogenesis in sea turtles does not depend solely on direct gene mutations, but rather on complex gene-environment interactions and epigenetic chromatin modifications [69] [70]. Differences in micro-location within the nest cavity, variations in microvascular eggshell contact, differential yolk metal partitioning, and individual sensitivity to epigenetic remodeling (such as altered DNA methylation or histone acetylation) could influence whether an embryo suffers developmental arrest, specific malformations, or normal growth [67] [69] [70]. Embryonic developmental heterochrony may be another source of variation, since intra-clutch development across individuals is hardly synchronous.

To properly contextualize these findings, several study limitations must be acknowledged explicitly.

The first one relates to the absence of a normal control baseline. The 125 non-hatched eggs recovered from the sampled artificial nests contained malformed embryos. No-normally developed embryos were available to serve as analytical controls. As a result, this study cannot determine whether trace metal burdens in malformed embryos differ significantly from those in healthy hatchlings from the same cohort. In addition, there is no single, standardized regulatory threshold (such as an EPA or WHO limit) that defines a specific metal concentration as universally teratogenic for sea turtle eggs. Because sea turtles are protected species, controlled toxicity experiments (which determine exact dose-response curves like LD 50 or EC 50 ) are ethically restricted, making established threshold values scarce. Consequently, field observations and tissue bioaccumulation analyses remain indispensable for ecotoxicological assessment [36] [37].

Samples were obtained from artificial relocation hatcheries. Relocation introduces altered incubation microclimates, including thermal spikes, gas exchange alterations, and hydric fluctuations, that independently may induce physical deformities [66]. Isolating the relative contributions of microclimatic stress, maternal transfer, and heavy metal exposure is a labor that must be considered for future field work to reduce confounding factors associated with artificial nesting stressors.

The embryos were stored in 4% paraformaldehyde prior to tissue dissection. Formalin fixation can leach highly soluble, unbound ions ( K + , Ca 2 + , Mg 2 + ) by over 50% - 80% and induce protein cross-linking that alters sample homogeneity and redistributes metals like Fe , Zn , and Cu . Absolute tissue concentrations should therefore be interpreted with appropriate caution.

Relationships between genetics and environment in marine turtles remain complex to unravel, with genetic factors influencing embryo’s vulnerability and environmental pollutants affecting development. Further studies are needed to address these questions on the biological bases of marine turtle fetal malformations and to constrain impacts of genetic mutations and environmental pollutants on marine turtle’s conservation.