Strontium, oxygen, and carbon isotope variation in modern human dental enamel

Abstract Objectives Isotopic analyses using human dental enamel provide information on the mobility and diet of individuals in forensic and archeological studies. Thus far, no study has systematically examined intraindividual coupled strontium (Sr), oxygen (O), and carbon (C) isotope variation in human enamel or the effect that caries have on the isotopic integrity of the enamel. The inadequate quantification of isotopic variation affects interpretations and may constrain sample selection of elements affected by caries. This study aims to quantify the intraindividual isotopic variation and provides recommendations for enamel sampling methods. Material and Methods This study presents the first systematic results on intraindividual variation in Sr–O–C isotope composition and Sr concentration in modern human dental enamel of third molars (affected and unaffected by caries). A multiloci sampling approach (n = 6–20) was used to analyze surface and inner enamel, employing thermal ionization mass spectrometry (TIMS) and isotope ratio mass spectrometry (IRMS). Third molars were analyzed from 47 individuals from the Netherlands, Iceland, the United States, the Caribbean, Colombia, Somalia, and South Africa. Results Intradental isotopic variation in modern Dutch dental elements was recorded for Sr, O, and C and exceeded the variation introduced by the analytical error. Single loci and bulk sampling approaches of third molars established that a single analysis is only representative of the bulk Sr isotope composition in 60% of the elements analyzed. Dental elements affected by caries showed twice the variation seen in unaffected dental elements. Caries did not consistently incorporate the isotopic composition of the geographical environment in which they developed. Discussion The isotopic variability recorded in unaffected inner enamel indicates that variations greater than 0.000200 for 87Sr/86Sr and larger than 2‰ for δ18O and δ13C are required to demonstrate changes in modern Dutch human diet or geographic location.


| INTRODUCTION
Human dental enamel contains information regarding the geographical origin and dietary patterns of an individual. Dental enamel records the isotopic signatures of the diet consumed during enamel mineralization (Bentley, 2006;Lee-Thorp, 2008;Montgomery, 2002), which for the permanent dentition takes place between birth and circa 16 years of age (AlQahtani, Hector, & Liversidge, 2010;Nanci, 2012;Piesco & Avery, 2002). The isotopic signature of the enamel, representative of the diet and geographical location of the individual during childhood, is preserved because enamel does not remodel after mineralization.
The isotopic variation in human tissues is widely used as an indicator for diet and migration. However, the relationship between intraindividual and intrapopulation variation has not been adequately quantified in humans despite considerable geochemical variability in enamel demonstrated by previous studies (Hare, Austin, Doble, & Arora, 2011;Smith et al., 2018;Willmes et al., 2016;Wright, 2013).
Both types of variation will be related to the variability of isotopic inputs over the time of enamel formation. Intraindividual variation may, however, also be related to (a) sampling location, (b) sampling method, and (c) enamel damage, specifically caries.
This re-equilibration of the isotopic composition during the mineralization phase means that the isotopic composition of the sample taken along the incremental enamel layers would be representative of the isotopic composition during mineralization rather than incremental enamel deposition Trayler & Kohn, 2017). Moreover, sequential sampling is hampered by a poor understanding of enamel formation and mineralization. The effects of spatial and temporal controls as well as physiological factors (e.g., health, sex, diet, and physical activity) are unknown (Balasse, 2003;Blumenthal et al., 2014;Fincham et al., 1999;Reade, Stevens, Barker, & O'Connell, 2015;Simmer & Fincham, 1995;Trayler & Kohn, 2017). In addition, the effect that these physiological factors have on enamel formation differs in various populations (Tompkins, 1996). Isotopic values incorporated in the enamel are also influenced by geographical controls, especially O and Sr, which are controlled by the local precipitation (O, Lightfoot & O'Connell, 2016) and geology (Sr, Bentley, 2006). An example of temporal control is the introduction of the modern supermarket diet in the 1970s. The availability of a greater variety of products grown in different geological settings is expected to increase the isotopic variation seen in modern human dental enamel compared to archeological dental enamel (Chesson, Ehleringer, & Cerling, 2011;Valenzuela, Chesson, Bowen, Cerling, & Ehleringer, 2012;Vautour et al., 2015).
Currently, no formal guidelines have been established for enamel sampling methods used in isotopic studies of human dental elements.
The lack of a formal sampling approach may affect the intraindividual variation seen in individuals sampled as well as decrease the comparability between isotopic analyses of studies that use different sample loci. Presently, enamel sampling generally involves a single sample location of a dental element, collected across a tooth's inner enamel, indiscriminate of enamel growth phases Slovak & Paytan, 2011). This sampling approach disregards the potential influence of intraindividual isotopic variation within a single dental element, that is, intradental variation. It is therefore unknown if a single sample location is representative of the total enamel Sr-O-C isotope composition of the dental element, referred to in this study as the bulk isotopic composition. Enamel is generally sampled using a handheld dental drill with tungsten or diamond burrs or saws (e.g., Balasse, 2003;Slovak & Paytan, 2011;Trayler & Kohn, 2017).
Some studies suggest that diagenetic Sr contaminates the surface enamel (~0.1 mm) after mineralization due to diffusion of Sr in the saliva from the diet and water in the mouth (Dufour et al., 2007;Horn & Müller-Sohnius, 1999) or because of interaction with the burial environment (Kohn et al., 1999;Schoeninger, Hallin, Reeser, Valley, & Fournelle, 2003). As a result, tooth surfaces are usually mechanically cleaned by removing the outer surface layer prior to sampling (Balasse, 2002;Kootker, van Lanen, Kars, & Davies, 2016;Reade et al., 2015;Slovak & Paytan, 2011). Although detailed sampling information is rarely provided in scientific literature, researchers seem to prefer to avoid sampling surface enamel in contact with other dental elements or areas that are affected by caries or other defects (Kootker, Mbeki, et al., 2016;Montgomery, 2002). Therefore, the potential influence of carious processes, such as demineralization and remineralization (see references in Li, Wang, Joiner, and Chang (2014) and Cochrane, Saranathan, Cai, Cross, and Reynolds (2008)), on Sr-O-C isotope ratios remains unknown.
A better understanding of the intradental isotopic variation of nonmigratory individuals is therefore required to provide a baseline of intraindividual isotopic variation. This will improve the accuracy of the interpretation of mobility and dietary patterns in both archeological and forensic contexts. Therefore, this study evaluated intraindividual isotope variation of Sr-O-C isotope composition, as well as Sr concentration, within modern human dental enamel of third molars (affected and unaffected by caries) from individuals known to have lived in one location during enamel formation and mineralization.
This study aims to determine if: 1. The variation of Sr-O-C isotope composition in modern human dental enamel is in the same order of magnitude between various sample locations within the same dental element (intradental), as well as within other dental elements (interdental) of the same individual. By quantifying the intradental variation, it could be examined if there is any spatial control on the Sr-O-C isotope composition within a dental element.
2. An individual's life history (year and region of birth) has an effect on the Sr-O-C isotope variation in modern human enamel.
3. The Sr-O-C isotope values of single sample locations from the cusp match the average isotopic variation obtained from bulk sample analysis of the same dental element.
4. The presence of caries affects the variation seen in Sr-O-C isotope composition of modern human dental elements.
Finally, these data are used to recommend a sampling protocol and to establish what Sr-O-C isotope variation is required to establish dietary change or mobility in modern Dutch humans.

| Sample selection
Extracted third molars were donated to the Vrije Universiteit Amsterdam by patients of dental clinics and medical centers in the Netherlands to be used for isotopic analyses. Background information was obtained through anonymous questionnaires, providing information on an individual's geographical location at the time of enamel formation (Figure 1), as well as diet, health, smoking, and exercise habits. The isotopic analyses of the teeth were approved by the Medical Ethics Review Committee of the VU University Medical Center.
The enamel of third molars is formed between the age of 8 and 16 years (AlQahtani et al., 2010). Teeth were selected based on the mobility profiles of 38 Dutch individuals (residential stability during enamel formation and mineralization) and the presence of caries. In addition, third molars from nine individuals from the Dutch Antilles (n = 3, Curaçao, Bonaire, Aruba), the Dominical Republic (n = 1), Colombia (n = 1), the United States (n = 1), Iceland (n = 1), Somalia (n = 1), and South Africa (n = 1) were sampled to compare their isotopic variation with the Dutch isotopic variation.
The following groups were examined: 1. Three individuals born in the same decade in the late 20th century (years of birth 1989, 1995, and 1991, respectively) raised in cities in close proximity to each other in the Netherlands (T1, T2, T6), representative of inputs from a globalized supermarket diet that emerged in the 1970s. For each of these three individuals, one third molar was sampled using an ultrahigh-density approach (n = 20 samples per element), with other third molars sampled using a high-density approach (n = 6 samples per element).
2. Three individuals born in the mid-20th century (years of birth 1949, 1964, and 1942) from the Netherlands (D3, D13, D15), representative of the pre-supermarket diet. These individuals were sampled using the high-density approach (n = 6).
3. Single cusp location and bulk sampling approaches were compared for Sr isotope analysis of 35 individuals. The results from the bulk sampling approach were previously reported (Plomp et al., 2019).
The results from the Dutch individuals (n = 29) were compared with non-Dutch individuals (n = 6). These individuals were born in 4. Six individuals whose teeth developed caries. The unaffected enamel of these individuals was sampled using the high-density approach (n = 6), with caries being sampled from the surface toward the inner enamel (n = 3-4).

| Sample preparation
The enamel was sampled, chemically processed, and analyzed at the Faculty of Science, Vrije Universiteit Amsterdam. Sample preparation and procedures are described in detail in Plomp et al. (2017) and Plomp, Smeets, & Davies (2020). The enamel was sampled perpendicular to the enamel dentine junction (EDJ) using a dental microdrill fitted with an acid cleaned diamond-tipped rotary burr and blade (Minilor Perceuse). Enamel at the EDJ was not sampled as the thin enamel layer at the EDJ is difficult to isolate (Reade et al., 2015). Occlusal fissures were avoided as they (a) are difficult to mechanically clean and sample and (b) have been reported to be less mineralized (He, Huang, Jing, & Hao, 2010;Montgomery, 2002), making them potentially more prone to diagenesis.
To indicate the sample locations on each molar, a coding system was developed (Table 1 and Figure 2). A distinction was made between the occlusal surface (cusp, or cuspal enamel, Dean, 2000) and the sides of the dental element (wall, or lateral enamel, Dean, 2000) (Figure 2c), where lateral enamel is secreted in a lateral direction and does not contribute to increase the tooth height (Dean, 2000). The tooth wall was sampled on the wall surface (S) and the inner enamel wall (W). Similarly, the tooth cusp was sampled on the surface (CS) and from the inner enamel cusp (C). Sampling at increasing depth for the cusps is indicated by suffix 0.1 to 0.3 (in sample location numbers), representative of 0.2-0.5 mm. To indicate the difference between the four walls, directional terms were used: lingual, buccal, medial, and distal (see Table 1 and Figure 2a).
The lingual tooth wall (WL) is next to the tongue, the buccal tooth wall (WB) is opposite the lingual tooth wall, toward the cheek. The mesial tooth wall (WM) is in contact with the second molar toward the midpoint of the dental arch. The distal tooth wall (WD) is opposite the mesial tooth wall toward the back of the dental arch. To indicate the difference between the three to four cusps on the third molar, cusps were classified based on the largest cusp (protocone/ protoconid, with names ending in -cone indicative for the upper dentition and names ending with -id for the lower dentition, see Table 1,     Table 2. Cusp samples (C1.1) and bulk samples were taken from the same third molar for Sr isotope analysis, as described earlier and in Table 1.
The bulk enamel samples represent~90% of the enamel of a dental element (excluding the surface enamel and the enamel from C1.1).
The bulk samples ranged from 273 to 1,310 mg, of which 1-2% aliquots were taken after sample dissolution (Plomp et al., 2019).
The error in the TSTD value is taken as the analytical error of the study.

| Oxygen and carbon isotope analysis
Oxygen and carbon isotope analysis was performed on powdered enamel. The sample (0.3-0.8 mg) was weighed into an exetainer vial.   Note: Information is provided on the geographical location (city), sample ID, dental element (using the FDI World Dental Federation notation-ISO 3950), sample location (described in Figure 2), location of caries, strontium isotope ratio (n = 148) and concentration (n = 146), and oxygen and carbon isotope values (n = 132).

| Statistical analyses
(Section 3.2), and single cusp location versus bulk sampling approaches (Section 3.3). Finally, the high-density sampling results from dental elements affected by caries are outlined (Section 3.4).

| Intradental variation indicated by ultrahighdensity sampling
The Sr-O-C isotope results for the three individuals (T1, T2, T6) sampled using the ultrahigh-density approach are shown in Figure 3. Differences in strontium isotope ratios (Δ 87 Sr/ 86 Sr maxmin ) ranged from 0.000091 to 0.000193 (up to 10 times larger than the analytical error) (

| Population variation indicated by highdensity sampling
To compare the variation seen in Sr-O-C isotopes between individuals (interindividual variation), the high-density sampling approach was applied to six individuals from the Netherlands (T1, T2, T6, D3, D13, D15; 10 teeth in total) ( Figure 5,  (48). The x-axis indicates sampling location (see Figure 2 and Table 1) T A B L E 3 Strontium, oxygen, and carbon isotope results for the individuals from the Netherlands, Johannesburg, Dominican Republic, and Somalia (see Table 2 Table 1 for abbreviations) to match the locations with the other sampled teeth. Dental elements F6, F7, M6, J, DR, and S were affected by caries. Individuals J, DR, and S are non-Dutch.

| High-density sampling across the Netherlands
The difference in 87 Sr/ 86 Sr (Δ max−min ) within the six Dutch individuals using the high-density-sampling approach on inner enamel ranged from 0.000021 to 0.000161 (2 SD ± 16-115, avg_isovar ± 62,

F I G U R E 4
Interdental Sr-O-C isotope variation of T1, T2 and T6. The x-axis represents the sample locations as described in Figure 2

| The effect of caries on isotopic composition
To evaluate the effect of caries on the isotopic composition of dental enamel, both unaffected and affected locations were sampled from the same dental element (Table 3,

| DISCUSSION
The maximum intraindividual variation in inner dental enamel seen in this study ( 87 Sr/ 86 Sr = 0.000161, 44 ppm Sr, δ 18 O and δ 13 C~1.4‰) highlights the importance of quantifying intraindividual isotopic variation before interpretations on mobility and diet are made.

| Intra-and interindividual variation
Previous studies proposed possible diagenetic Sr contributions to surface enamel (Dufour et al., 2007;Horn & Müller-Sohnius, 1999    the dental elements analyzed ( Figure 6). The two samples with the largest differences in 87 Sr/ 86 Sr in cusp and bulk samples (0.000078/9) belong to individuals M2.1 and R16, raised in a single location (Maastricht and Dordrecht, respectively), with individual R16 indicating that they did not consume fish. These data demonstrate that even where individuals are sedentary, the dental Sr isotope ratios are heterogeneous. Although the current study did not contrast single versus bulk samples for oxygen and carbon isotopes, results for oxygen and carbon isotopes in caprid (Ammotragus lervia, Reade et al., 2015) similarly suggest that single location samples may not represent the average isotopic value of the dental element, and that bulk sampling condenses the full isotopic variation within a dental element.
Future studies are encouraged select their sampling strategies based on the type of isotopic variation they would like to evaluate. In cases where intraindividual variation should be assessed, or when the life history of an individual is examined, the multiloci or high-density approach is more effective than single loci sampling. If interindividual variation is determined and only an estimation of the isotopic signature is required, single loci sampling may be sufficient. As bulk sampling averages the total isotopic variation and requires large sample sizes, this method may not be applicable. Compared with bulk analyses, single-loci sampling approaches provide similar information, are more efficient, and are less destructive. In situ sampling analysis using laser ablation inductively coupled mass spectrometry (LA-ICP-MS) analysis may be better suited to assess the intraindividual variation in Sr concentration and composition in a dental element (Smith et al., 2018;Willmes et al., 2016). The (ultra)high-density sampling approach showed considerable intraindividual isotopic variation within modern Dutch individuals. Significant interdental variation was seen in strontium and oxygen isotopes in two out of three individuals (T1 and T6), indicating that a single third molar is not always representative of the isotopic results of other third molars and possibly other dental elements. Using the high-density sampling approach, the maximum differences in isotopic results from inner enamel of a single dental element reached 0.000161 for strontium, 1.4‰ for oxygen, and up to 1.3‰ for carbon. Increased variation is seen using the ultrahigh-density sampling approach (including results from additional molars from the same individual), with maximum intraindividual variation ( 87 Sr/ 86 Sr = 0.000192, 2‰ for oxygen and carbon, Table 3) approaching levels of variability in the Dutch population for oxygen (δ 18 O = 3.1‰, Font et al., 2015) and carbon (δ 13 C = 2.3‰). Intraindividual Sr variation did not reach interindividual/population variation, which was 10 times higher (Δ =~0.002). The analytical precision ( 87 Sr/ 86 Sr ± 0.000019, δ 18 O ± 0.17‰ and δ 13 C ± 0.04‰) is therefore significantly less than the intraindividual and population variability recorded here. The estimated variation based on the average Dutch isotopic intraindividual variation (avg_isovar) of the inner enamel of single dental elements in this study ( 87 Sr/ 86 Sr ± 0.000062, Sr ppm ± 15, δ 18 O and δ 13 C ± 0.7‰) provides F I G U R E 6 The difference in Sr isotope ratio recorded between bulk and cusp samples from an individual. Note that the analytical error does not accurately capture the intraindividual variation seen in Sr isotope composition of Dutch individuals; see Table 4 for all ΔCusp-Bulk values estimations for the expected variation in modern Dutch third molars.
The maximum intraindividual differences seen in this study for inner enamel ( 87 Sr/ 86 Sr = 0.000161, Sr ppm = 44, δ 18 O = 1.4‰ δ 13 C = 1.3‰) can be used as an indication for the maximal expected variation within inner enamel of a dental element and may increase in future studies when more individuals are analyzed using a highdensity sampling approach. Third molars are the most variable dental element in the human dentition in terms of enamel formation and eruption (AlQahtani et al., 2010;Reid & Dean, 2006), and they can therefore be expected to record the largest isotopic variation in the dentition of habitual diet and residence. Consequently, studies using other dental elements may be expected to exhibit less isotopic variation as time frames of dental development are shorter and more constrained. In order to establish how representative this study is of the isotopic variability recorded in human dentition, it would be of particular interest to study first/second molars and pre-molars, as these are most often sampled in archeological/forensic provenance studies. A comparative study of archeological inner enamel of dental elements would also be of interest, keeping in mind the effects that diagenesis may have on the isotopic composition of archeological dental elements.

| Temporal and spatial effects on Sr-O-C isotope variation
Previous studies (Chesson et al., 2011;Valenzuela et al., 2012;Vautour et al., 2015) have highlighted the potential influence of the modern global supermarket diet on the isotopic results of modern human tissues. Assessing the impact of globalization on the individuals in this study is complicated as the individuals originate from various areas in the Netherlands, which will likely result in some degree of isotopic variation due to different sources of potable water and variation in local food availability. In addition, variation is expected due to  Figure 6). Strontium concentration in caries is indistinguishable from unaffected enamel in the same dental element (see also Little & Steadman, 1966), with the exception of one caries (F6-G2, Figure 7). moving and it appears likely that the three caries were active at the same time.

| Increased Sr-O-C isotope variation and elevated oxygen values in carious enamel
The non-Dutch data strongly suggest that the de-and remineralization processes involved in the development of caries (Piesco & Simmelink, 2002;Wazen & Nanci, 2012) are not likely to incorporate Sr from the diet consumed when the caries were active. Although in most cases caries do not significantly affect the Sr-C isotope composition, the data presented here indicate that sampling caries is not recommended for modern dental elements. Oxygen isotope values are altered and the elemental incorporation process in caries remains poorly understood, as well as the time frame involved in the development of the caries.

| CONCLUSION
Assessing the intradental, interdental, and population isotopic variability is a vital step in providing a framework for provenance and dietary interpretations. This work establishes that a single sample location is not representative for the total intradental enamel isotopic variation and that bulk analyses average the total variation present in the modern third molars.
This study indicates that drilled samples should be taken from the inner enamel, with no preference for a particular cusp/wall region as these locations offer comparable isotopic results. Sampling approaches should avoid carious enamel as this study indicated that caries produce inconsistent results. The unaffected enamel of carious dental elements seems to be isotopically unaltered and can be used for isotopic analyses.
Further studies are required to quantitatively evaluate the intraindividual variability in modern and archeological enamel in dental elements other than third molars, as well as the effect of caries on the isotopic composition of enamel. The intraindividual isotopic variation is expected to be controlled by a combination of the geological area in which food is grown and personal diet preferences of individuals. The resulting isotopic variation needs to be quantified for other modern or archeological populations living in regions with larger topographical and geological variation by analyzing the enamel of multiple individuals (>20) to provide a baseline to which intraindividual isotopic variation can be compared.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available in Tables 2-4 as well as openly available at the 4TU.Centre for Research