Evolutionary modifications of human milk composition: evidence from long-chain polyunsaturated fatty acid composition of anthropoid milks

Journal of Human Evolution 55 (2008) 1086–1095

Contents lists available at ScienceDirect

Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol

Evolutionary modifications of human milk composition: evidence from long-chain polyunsaturated fatty acid composition of anthropoid milks Lauren A. Milligan a, *, Richard P. Bazinet b a b

Department of Anthropology, University of California, Santa Cruz, CA 95064, USA Department of Nutritional Sciences, Faculty of Medicine, University of Toronto, Toronto, ON, M5S 3E2, Canada

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 December 2007 Received in revised form 18 June 2008 Accepted 20 June 2008

Brain growth in mammals is associated with increased accretion of long-chain polyunsaturated fatty acids (LCPUFA) in brain phospholipids. The period of maximum accumulation is during the brain growth spurt. Humans have a perinatal brain growth spurt, selectively accumulating docosahexaenoic acid (DHA) and other LCPUFA from the third trimester through the second year of life. The emphasis on rapid postnatal brain growth and LCPUFA transfer during lactation has led to the suggestion that human milk LCPUFA composition may be unique. Our study tests this hypothesis by determining fatty acid composition for 11 species of captive anthropoids (n ¼ 53; Callithrix jacchus, Cebus apella, Gorilla gorilla, Hylobates lar, Leontopithecus rosalia, Macaca mulatta, Pan troglodytes, Pan paniscus, Pongo pygmaeus, Saimiri boliviensis, and Symphalangus syndactylus). Results are compared to previously published data on five species of wild anthropoids (n ¼ 28; Alouatta paliatta, Callithrix jacchus, Gorilla beringei, Leontopithecus rosalia, and Macaca sinica) and human milk fatty acid profiles. Milk LCPUFA profiles of captive anthropoids (consuming diets with a preformed source of DHA) are similar to milk from women on a Western diet, and those of wild anthropoids are similar to milk from vegan women. Collectively, the range of DHA percent composition values from nonhuman anthropoid milks (0.03–1.1) is nearly identical to that from a cross-cultural analysis of human milk (0.06–1.4). Humans do not appear to be unique in their ability to secrete LCPUFA in milk but may be unique in their access to dietary LCPUFA. Ó 2008 Elsevier Ltd. All rights reserved.

Keywords: LCPUFA Anthropoid primate DHA

Introduction The long-chain polyunsaturated fatty acids (LCPUFA) docosahexaenoic acid (22:6n-3; DHA) and arachidonic acid (20:4n-6; AA), are the omega-3 (n-3) and omega-6 (n-6) fatty acids found in the highest concentrations in neural membranes (Carlson, 2001). Together, they make up a third of all lipids in the brain’s grey matter (Gibson, 1997; Brenna and Diau, 2007), with DHA in particularly high concentrations in membranes surrounding neural synapses (Carlson, 2001), retinal phospholipids (Sheaff-Greiner et al., 1997), and photoreceptors (Carnielli and Sauer, 1996). Indeed, the brain appears to be selective in the incorporation of LCPUFA, preferring those with 20 and 22 fatty acids rather than their 18 carbon polyunsaturated fatty acid (PUFA) precursors (Carnielli and Sauer, 1996; Carlson, 1999, 2001; Koletzko et al., 2001; Innis, 2003). Brain growth in mammals is associated with increased incorporation of LCPUFA in brain phospholipids (Farquharson et al., 1992). The period of maximum accumulation is during the brain

* Corresponding author. E-mail address: [email protected] (L.A. Milligan). 0047-2484/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2008.07.010

growth spurt, which occurs in utero in most mammals (Huang and Brenna, 2001). Thus, most LCPUFA are transferred by the placenta to the developing fetal brain. Humans have a perinatal brain growth spurt (Huang and Brenna, 2001), selectively incorporating DHA from the third trimester (approximately 26 weeks of gestation) through the second year of postnatal life (Clandinin et al., 1980a,b; Farquharson et al., 1992; Makrides et al., 1994; Sheaff-Greiner et al., 1997). In humans, LCPUFA transfer may be important during both gestation and lactation. Further, a larger relative brain size means that humans may be unique among primates in their nutritional requirements for DHA and other LCPUFA. Alpha-linolenic acid (18:3n-3; ALA) and linoleic acid (18:2n-6; LA) cannot be synthesized de novo, and must be obtained through the diet (Lands, 1992; Huang and Brenna, 2001). Thus, ALA and LA are considered dietary essential fatty acids. Their presence in milk reflects their presence in maternal diet, past (depot fat stores) and present. DHA and AA are not considered essential as they can be synthesized from ALA and LA, respectively, by a series of reactions occurring primarily in the liver’s endoplasmic reticulum where two carbon units (elongation) and double bonds (desaturation) are added (Fig. 1). The proportion of DHA and AA in milk reflects the presence of these fatty acids in the maternal diet and maternal

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

Plant and Animal Food Sources

18:3n-3 α-Linolenic acid (ALA)

18:2n-6 Linoleic acid (LA)

1087

Plant and Animal Food Sources

Δ6 Desaturase 18:4n-3 Stearidonic acid

18:3n-6 Y-Linolenic acid Elongase

20:4n-3 Eicosatetraenoic acid

20:3n-6 Dihomo-Y-Linolenic acid Δ5 Desaturase

Animal Food Sources

20:5n-3 Eicosapentaenoic acid (EPA)

20:4n-6 Arachidonic acid (AA)

Animal Food Sources

Elongase 22:5n-3 Docosapentaenoic acid (n-3 DPA)

22:4n-6 Docosatetraenoic acid Elongase Δ6 Desaturase

Animal Food Sources

β-Oxidation

22:6n-3 Docosahexaenoic acid (n-3 DHA)

22:5n-6 Docosapentaenoic acid (n-6 DPA)

Fig. 1. The synthesis of docosahexaenoic and arachidonic acid from their 18 carbon precursors. Plant and animal sources of ALA or LA can be converted to DHA and AA, respectively, via a series of desaturation/elongation reactions occurring primarily in the liver. The efficacy of these reactions is controversial, but generally considered to be low in humans. Preformed AA is found in meats and eggs, whereas EPA and DHA are present in high quantities in marine animals.

conversion of their PUFA precursors. Studies on both humans and baboons have demonstrated that these species are inefficient in the conversion of ALA to DHA (reviewed in Sheaff-Greiner et al., 1997; Su et al., 1999, 2005; Burdge, 2006; Plourde and Cunnane, 2007). Human neonates and infants may be less efficient than adults in converting ALA to DHA. Without a preformed source of DHA in their diet, the rate of DHA formation from ALA (measured with deuterated ALA ethyl esters) in human infants may be inadequate in meeting neural requirements, especially in preterm infants who have an increased requirement for DHA (Salem et al., 1996). Martin (1983) hypothesized that human milk may be speciesspecific, having either unique nutrients or nutrients in greater quantities than seen in other precocial mammals. Further, he predicted (1995) that an investigation of primate milks would reveal the biochemical requirements necessary for human brain growth, emphasizing LCPUFA, such as DHA and AA. As noted by Sellen (2007), there has been a lack of comparative data to test this hypothesis. Robson (2004) tested Martin’s (1995) prediction by comparing LCPUFA profiles from the human milk literature with published studies on two species of macaque (Macaca mulatta and Macaca fuscata). She concluded that milk LCPUFA composition was quite similar between humans and macaques, thus refuting Martin’s hypothesis. However, comparative data from Macaca fuscata came from colostrum samples, which were likely to have higher fatty acid concentrations than samples collected from the midlactation phase (Iverson and Oftedal, 1995). Further, Robson (2004) reported that human levels of AA and DHA were quite consistent across human populations, a finding that runs counter to conclusions from a recent meta-analysis of 65 studies on human milk composition (Brenna et al., 2007). We previously reported fatty acid composition from five species of anthropoid primate living in the wild (Milligan et al., 2008). Relative to wild nonhuman primates, humans may have differential access to foods rich in preformed n-3 and n-6 LCPUFA. Captive primate diets, however, are often supplemented with fish meal, high in DHA, as a source of protein (Sheaff-Greiner et al., 1997).

Thus, fatty acid profiles of milks from captive primates may represent the milk production capabilities of nonhuman primates when supplied with a preformed source of DHA. Here, we report on LCPUFA profiles from captive individuals representing 11 anthropoid species: Callithrix jacchus (common marmoset), Cebus apella (tufted capuchin), Gorilla gorilla (lowland gorilla), Hylobates lar (white handed gibbon), Leontopithecus rosalia (golden lion tamarin), Macaca mulatta (rhesus macaque), Pan troglodytes (chimpanzee), Pan paniscus (bonobo), Pongo pygmaeus (orangutan), Saimiri boliviensis (Bolivian squirrel monkey), and Symphalangus syndactylus (siamang). These results are compared to those from wild anthropoids: Alouatta paliatta (mantled howler monkey), Callithrix jacchus (common marmoset), Gorilla beringei (mountain gorilla), Leontopithecus rosalia (golden lion tamarin), and Macaca sinica (toque macaque); reported in (Milligan et al., 2008). This is to identify the range of variability among anthropoid n-3 and n-6 PUFA and LCPUFA, particularly those implicated in brain growth and development (ALA, LA, AA, and DHA). Finally, we compare results from captive and wild anthropoids to data on human milk fatty acid composition (taken from Gibson and Kneebone, 1981; Specker et al., 1987; Koletzko et al., 1991, 1992; Sanders and Reddy, 1992; Jensen et al., 1995; Yuhas et al., 2006; Brenna et al., 2007) to determine if human milk fatty acid profiles are species-specific or have specific LCPUFA in higher proportions than do nonhuman anthropoids (Martin, 1995). Materials and methods Milk samples This study analyzed 53 milk samples from 11 anthropoid primate species living in captivity (Callithrix jacchus, Cebus apella, Gorilla gorilla, Hylobates lar, Leontopithecus rosalia, Macaca mulatta, Pan paniscus, Pan troglodytes, Pongo pygmaeus, Saimiri boliviensis, and Symphalangus syndactylus; Table 1). Milk collections of all species were opportunistic, both in respect to species included and day of lactation. Samples from Callithrix jacchus, Gorilla gorilla,

1088

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

Table 1 Sample information for captive anthropoidsa Species

Superfamily

No. of samples analyzed

Source of milk samples

Callithrix jacchus (Common Marmoset) Leontopithecus rosalia (Golden Lion Tamarin) Cebus apella (Tufted Capuchin) Saimiri boliviensis (Bolivian Squirrel Monkey) Macaca mulatta (Rhesus Macaque) Gorilla gorilla (Lowland Gorilla) Pan paniscus (Bonobo) Pan troglodytes (Chimpanzee) Pongo pygmaeus (Bornean Orangutan) Hylobates lar (White handed Gibbon) Symphalangus syndactylus (Siamang)

Ceboidea Ceboidea Ceboidea Ceboidea Cercopithecoidea Hominoidea Hominoidea Hominoidea Hominoidea Hominoidea Hominoidea

4 1 7 8 21 4 1 4 1 1 1

ML Power, National Zoological Park National Zoological Park Alpha Genesis, Inc. Yemassee, SC SV Gibson and LE Williams, University of South Alabama KJ Hinde, California National Primate Research Center San Diego Wild Animal Park; Zoo Atlanta Milwaukee Zoo Southwest National Primate Research Center; St. Louis Zoo Zoo Atlanta Minnesota Zoo Riverbanks Zoological Park

a

Corresponding information for wild anthropoids available in Milligan et al. (2008).

Hylobates lar, Leontopithecus rosalia, Macaca mulatta, Pan paniscus, Saimiri boliviensis, and Symphalangus syndactylus were collected by other researchers prior to the development of this project. All samples were part of the mammalian milk collection at the Nutrition Laboratory, Smithsonian National Zoological Park (provided by O. Oftedal or M. Power) except for the Saimiri samples, which were provided by the University of Southern Alabama (S. Gibson and L. Williams). This project was part of a larger project aimed at understanding milk composition across anthropoids (Milligan, 2007). Thus, emphasis was placed on surveying each of the three anthropoid superfamilies (Ceboidea–New World monkeys; Cercopithecoidea–Old World monkeys; Hominoidea–apes). In addition, because of the importance of LCPUFA in brain growth and development, sample collection from captive populations of the relatively large-brained Cebus apella was initiated by one of us (LAM). As a result of the opportunistic nature of this data set, available details on milk donors and methods of milk collection for study species vary. Protocols for animal capture and milk collection were provided by individuals directly involved with each study and are detailed in Appendix 1 of the Supplementary material (Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jhevol.2008.07.010). Information on samples from wild anthropoids is described in Milligan et al. (2008). All studies received approval by the appropriate institutional authorities. Length of lactation varies widely among nonhuman primate species included in this study. Lactation in callitrichines (Callithrix and Leontopithecus) is approximately three months (Garber and Leigh, 1997; Ross, 2003), while that of Pan troglodytes is approximately five years (Goodall, 1986; Leigh and Shea, 1996; Kappeler and Pereira, 2003; Ross, 2003). It is assumed that the length of each stage (colostrum through involutional) also varies. To permit interspecific comparisons, this study followed Oftedal and Iverson’s (1995) emphasis on nutritional rather than developmental stages to describe milk samples. Lactation was divided into three stages: early lactation, midlactation, and late lactation. Midlactation is defined as the period of maximum lactation performance, when infants are completely dependent on mothers for absolute nutrition requirements (Oftedal and Iverson, 1995). Samples from early and late lactation were excluded from analyses.

level of variation in these fatty acids from several populations (Jensen et al., 1995; Yuhas et al., 2006), including the most extensive meta-analysis on human milk fatty acid profiles to date (Brenna et al., 2007). To illustrate the relationship between maternal diet and PUFA/LCPUFA composition in human milk, we also compare our results to data from populations following distinct dietary strategies (Table 2): Western (Gibson and Kneebone, 1981), non-Western (Koletzko et al., 1991), vegetarian (Specker et al., 1987), and vegan (Sanders and Reddy, 1992). These populations were selected specifically because of their difference in dietary n-3 and n-6 PUFA and LCPUFA, particularly DHA. For example, the non-Western population includes fish as a primary component of the diet, thereby illustrating the effect of preformed DHA on milk LCPUFA profiles. Data from human milk were not used in statistical analyses and were used only to explore qualitative differences in milk LCPUFA composition. Ranges in humans and nonhuman anthropoid primates represent the abilities of the mothers to secrete LCPUFA into milk under variable conditions. If human LCPUFA composition is indeed unique among anthropoid primates, the human and nonhuman anthropoid ranges should be distinct. Milk analysis All samples were maintained in a 20  C freezer at the Nutrition Laboratory until removed for subsampling or analysis. Because lipid fractions (phospholipids, triglycerides, and cholesteryl esters) may be susceptible to hydrolysis over long periods of storage, we measured total fatty acids, similar to other reports (Bazinet et al., 2003; Yuhas et al., 2006; Brenna et al., 2007; Milligan et al., 2008). Samples that were less than 1 ml in volume were allowed to thaw at room temperature, and those greater than 1 ml were placed in a warm water bath, maintained at 50  C, as per Nutrition Laboratory

Table 2 Fatty acid composition for human females from four different dietary strategies: Western, non-Western, vegetarian, and vegan Fatty acid

Australian (Western)a

Nigerian (non-Western)b

Vegetarianc

Vegand

Data on human milk LCPUFA composition

18:3n-3 (ALA) 22:6n-3 (DHA) 18:2n-6 (LA) 20:4n-6 (AA)

0.59  0.16 0.32  0.17 10.75  4.22 0.40  0.10

1.41 (0.64–5.45) 0.93 (0.70–2.16) 11.06 (5.4–13.8) 0.82 (0.38–1.48)

2.76  0.16 0.22  0.08 28.82  1.39 0.68  0.03

1.36  0.18 0.14  0.06 23.8  1.40 0.32  0.02

Fatty acids in human milk have been investigated in mothers under various dietary habits and maternal conditions, leading to the conclusion that there is no such thing as the human milk fatty acid profile (Jensen et al., 1995; Yuhas et al., 2006; Brenna et al., 2007). Thus, it was not possible to select one human population for comparison to our results on nonhuman primate milk LCPUFA profiles. Instead, we present data on studies that investigate the

a Means and standard errors from Gibson and Kneebone’s (1981) data on mature milk from Australian women (n ¼ 120). b Median values (and ranges) adapted from Koletzko et al. (1991) from women from Udo, Bendel State, Nigeria (n ¼ 10). Reported in Jensen et al. (1995). c Means and standard errors adapted from Specker et al. (1987) data on vegetarian mothers from New England (n ¼ 12). Reported in Jensen et al. (1995). d Means and standard errors adapted from Sanders and Reddy’s (1992) data on vegan mothers from England (n ¼ 19). Reported in Jensen et al. (1995).

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

(Smithsonian National Zoological Park) standard protocol, in order to ensure rapid and uniform thawing. Samples were vortexed prior to dispensing approximately 100 ml into 20 ml Kimex glass centrifuge tubes using a 250 ml positive displacement pipette. Sample weights were recorded to 0.0001 g. Samples were capped and immediately frozen. Known amounts of glyceryl triheptadecanoate (Sigma) and 13, 16, 19-docosatrienoic acid methyl ester (Sigma) were added as internal standards to milk prior to extraction. Total lipids were extracted from approximately 100 mg of milk according to the method of Folch (Folch et al., 1957). Fatty acid methyl esters were formed by heating the total lipid extract in 1% H2SO4 methanol at 70  C for 3 hours (Makrides et al., 1994). The methyl esters were separated on a 30 m  0.25 mm i.d. capillary column (SP-2330, Supelco; Bellefonte, PA), using gas chromatography with a flame ionization detector (Model 6890 N, Agilent Technologies; Palo Alto, CA, USA). Runs were initiated at 80  C, with a temperature gradient to 160  C (10  C/min) and 230  C (3  C/min) in 31 minutes, and held at 230  C for 10 minutes. Peaks were identified by retention times of fatty acid methyl ester standards (Nu-Chek-Prep, Elysian, MN). Fatty acid concentrations (mg per g of milk) were calculated by proportional comparison of gas chromatography peak areas to that of the heptadecanoate methyl ester peak. The concentration of the heptadecanoate methyl ester internal standard was validated for each sample used in this study by comparison to the concentration of the second internal standard (13, 16, 19-docosatrienoic acid methyl ester). Relative concentration, given as percent composition, was calculated by dividing individual fatty acid concentration by the total concentration of all reported fatty acids (Bazinet et al., 2003). Data analysis Results are presented as total fatty acid concentration (mg/g) and percent composition, the individual fatty acid concentration as a proportion of the total concentration of all reported fatty acids. For the purpose of this report, we focus on LA, ALA, AA, and DHA. Readers interested in other fatty acids are referred to Appendix 2 of the Supplementary online material. Several species (e.g., gibbons) are represented by only one sample while others (rhesus macaques) are represented by more than 20 samples, making it problematic to investigate variation at the level of the species. Thus, phylogenetic variation among captive anthropoids in LCPUFA composition was analyzed at the level of the superfamily (i.e., cercopithecoid, ceboid, hominoid). Comparisons between milk samples from wild and captive anthropoids are also discussed at the level of the superfamily as there were not wild conspecifics for all captive samples. Data were analyzed using SPSS 13.0. Significant differences in means among species and superfamilies were determined using ANOVA and multiple pairwise comparisons between species using the Tukey-Kramer test. Two-tailed t-tests were used to test for significant differences between the wild group (all samples obtained from females living in the wild) and captive group (all samples obtained from females living in captivity) and between wild and captive females of the same superfamily. Correlations among proportions of fatty acids were explored using Spearman’s rho. Statistical significance for all tests was set at a < 0.05, and data are presented as mean  standard error.

1089

As a result of interspecific variability in total milk fat (Milligan, 2007), there was a significant difference in the mean weight (mg per gram of milk) for total fatty acids among captive species (F ¼ 2.94, p ¼ 0.006) and between the captive and wild groups (t ¼ 4.99, p < 0.001). Therefore, differences among and between superfamilies and between captive and wild groups were analyzed using percent composition to identify differences in proportions of milk fatty acids. Variation among captive anthropoids Table 5 presents means, standard errors, and test statistics for percent composition of milk ALA, DHA, LA, AA, n-3 PUFA, n-6 PUFA, and total PUFA by superfamily of all samples from captive anthropoids. Significant differences were identified at the level of the superfamily for each PUFA. Hominoid milks were significantly greater than ceboid and cercopithecoid milks in the proportion of ALA and n-3 PUFA. Milk of cercopithecoids was significantly greater than ceboids and hominoids in the proportion of DHA and significantly lower than these superfamilies in the proportion of AA. Ceboid milk was significantly greater than hominoids and cercopithecoids in the proportion of LA, n-6 PUFA, and total PUFA. Variation among captive and wild anthropoids Means and standard errors and range in values for ALA, DHA, LA, AA, n-3 PUFA, n-6 PUFA, and total PUFA by group (wild and captive) are provided in Table 6. Captive samples were significantly lower than wild samples in mean percent composition of ALA both as a group and by superfamily (p < 0.05; Fig. 2). DHA contributed a significantly higher proportion to the milks from all captive compared to wild anthropoids and was significantly higher in captive cercopithecoids, ceboids, and hominoids compared to wild samples from those superfamilies (Fig. 3). Mean percent composition of LA in captive anthropoid milks was almost three times that of wild anthropoid milks. Captive samples also were significantly higher than wild samples as a group and by superfamily (p < 0.05; Fig. 4). There was no significant difference between wild and captive anthropoid milks as a group in mean percent composition of AA. However, significant differences were identified at the superfamily level (Fig. 5). Mean percent composition of AA was significantly lower in wild monkeys as compared to captive monkeys, and was significantly higher in wild hominoids (mountain gorillas) as compared to captive hominoids. As a group, captive anthropoids were significantly higher than wild anthropoids in the proportion of n-6 PUFA and total PUFA in milk, and significantly lower than wild anthropoids in the proportion on n-3 PUFA in milk. This study identified a negative correlation between the percent composition of ALA and DHA (r ¼ 0.52, p < 0.001) among milks from captive anthropoids, which is similar to that previously identified (Milligan et al., 2008) among wild anthropoids (r ¼ 0.61, p < 0.001). Percent composition of LA was positively correlated with percent composition of AA (r ¼ 0.72, p < 0.001). This relationship between precursor and long-chain metabolite is much stronger than was seen for wild samples (r ¼ 0.32, p ¼ 0.09; Milligan et al., 2008), even when mountain gorilla samples (which were significantly higher in mean percent composition of AA) were removed from the analysis (r ¼ 0.53, p ¼ 0.009).

Results Discussion The mean percent composition of LCPUFA for captive monkey species is provided in Table 3, and in Table 4 for captive ape species. Complete fatty acid profiles for each species are provided in Appendix 2 of the Supplementary material. Comparative data from wild anthropoids come from Milligan et al. (2008).

N-3 and n-6 PUFA and LCPUFA in nonhuman anthropoid milks Variation among captive anthropoids was identified at the level of the superfamily. However, we are cautious in discussing this

1090

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

Table 3 Mean (SE) fatty acid percent composition and total fatty acids (mg/g) in ceboids and cercopithecoidsa Captive species Fatty acid

18:3n-3 (ALA)** 20:5n-3 (EPA) 22:5n-3 22:6n-3 (DHA) Sum n-3 PUFA 18:2n-6 (LA) 18:3n-6 20:3n-6 20:4n-6 (AA) 22:5n-6 Sum n-6 PUFA Total PUFA Total fatty acids (mg/g)

Callithrix jacchus (n ¼ 4) 2.21  0.28 0.04  0.003 0.19  0.01 0.15  0.002 2.59  0.27 21.67  1.70 1.20  0.16 1.23  0.11 0.49  0.02 0.08  0.002 24.76  1.72 27.35  1.96 60.65  11.04

Wild species (data from Milligan et al., 2008) Leontopithecus rosalia (n ¼ 1) 0.61 0.03 0.16 0.12 3.17 11.50 ND 0.86 0.36 0.10 25.88 29.04 15.90

Cebus apella (n ¼ 7)

Saimiri bolivienses (n ¼ 8)

Macaca mulatta (n ¼ 21)

2.91  0.14 0.06  0.004 0.28  0.02 0.31  0.04 3.50  0.09 30.63  1.35 0.10  0.02 0.37  0.02 0.92  0.08 0.10  0.02 32.01  0.46 35.61  0.54 59.12  7.07

1.65  0.21 0.09  0.01 0.18  0.01 0.40  0.03 2.32  0.23 30.39  1.02 0.17  0.02 0.27  0.01 0.81  0.04 0.12  0.01 31.79  1.00 34.11  0.89 53.63  6.33

1.55  0.11 0.16  0.01 0.25  0.01 0.44  0.02 2.40  0.52 23.88  0.54 0.09  0.01 0.25  0.01 0.44  0.01 0.07  0.002 24.76  0.55 27.15  0.62 74.94  9.67

Callithrix jacchus (n ¼ 4)

Leontopithe cus rosalia (n ¼ 4)

Allouatta palliata (n ¼ 7)

Macaca sinica (n ¼ 8)

2.63  0.43 0.08  0.02 0.21  0.10 0.14  0.06 3.05  0.51 3.29  0.56 ND 0.23  0.02 0.30  0.03 0.01  0.02 3.88  0.61 6.94  1.01 14.01  5.30

1.94  0.43 0.11  0.01 0.24  0.10 0.24  0.06 2.52  0.36 5.37  1.02 ND 0.34  0.02 0.48  0.05 0.04  0.02 6.31  1.03 8.83  1.35 12.73  5.31

15.06  0.75 0.06  0.01 0.16  0.07 0.03  0.003 15.31  0.77 12.59  0.60 0.02  0.004 0.06  0.02 0.68  0.09 0.04  0.01 13.43  0.57 28.75  0.90 25.45  3.75

2.55  0.32 0.11  0.01 0.19  0.07 0.12  0.01 2.96  0.32 8.03  0.98 0.04  0.004 0.16  0.02 0.32  0.02 0.05  0.01 8.64  0.99 11.61  0.99 39.92  3.75

a ALA ¼ a-linolenic acid; EPA ¼ eicosapentaenoic acid; LA ¼ linoleic acid; AA ¼ arachidonic acid; DHA ¼ docosahexaenoic acid. ND ¼ Not detected. Percent composition was calculated by dividing individual fatty acid concentrations (includes saturates and monounsaturates; see online supplementary information) by the total fatty acid concentration. For more details on fatty acid nomenclature see Fig. 1.

variation as reflective of real phylogenetic differences. New and Old World Monkey Chows both contain soybean oil (a source of LA) and fish meal (a source of DHA), but may differ in the quantity included. Indeed, fatty acid composition of New World Monkey Chow differs by batch number (D. Robbins, Harlan Teklad, pers. comm.). Interspecific variation in the quantity of Monkey Chow (as well as fruits and vegetables) consumed could produce variation at the level of the superfamily. Milligan et al. (2008) discuss possible phylogenetic patterns in anthropoid milk fatty acids among wild anthropoids. Captive status influenced milk n-6 and n-3 PUFA and LCPUFA composition. Captive individuals were likely consuming different fatty acids or fatty acids in different proportions than their wild counterparts. Research on human milk suggests that the main dietary factors affecting milk fatty acid profiles are the amount of carbohydrates and PUFA in the diet (Koletzko et al., 1992, 2001; Sanders, 1999), two factors that are likely to differ between captive and wild individuals. Captivity also may affect maternal depot fat, an alternative source for milk fatty acids. Captive individuals are generally larger (stature, body mass) and grow faster than their wild counterparts (Zihlman et al., 1990, 2004; Leigh, 1994). As

a result of increased energy intake and decreased energy expenditure, captive individuals may be more likely to be in positive energy balance than wild individuals, possibly allowing mothers to transfer more PUFA and LCPUFA from their diet, rather than their depot stores, to milk. Samples from wild anthropoids were higher in mean percent composition of ALA. This difference between wild and captive samples is explained by higher values in wild mountain gorillas and mantled howlers. When these species are removed from the analysis, wild and captive samples are not significantly different in mean percent composition of ALA. This finding indicates that the proportion of ALA in milk is not related to captivity and may be better explained by the dietary strategy (folivory) of these two species. Like ALA, LA must be obtained from dietary sources. The highest value for wild samples was approximately 2% lower than the lowest value for captive samples. The lack of overlap suggests that captive individuals are consuming a source of this PUFA not available to individuals living in the wild. Monkey Chow, a food item fed to captive individuals in this study housed in zoos and primate research facilities, provides LA in soy, corn, and sunflower oils

Table 4 Mean ( SE) fatty acid percent composition and total fatty acids (mg/g) in hominoidsa Captive species

Wild species (data from Milligan et al., 2008)

Fatty acid

Pan paniscus (n ¼ 1)

Pan troglodytes (n ¼ 4)

Pongo pygmaeus (n ¼ 1)

Gorilla gorilla (n ¼ 4)

Hylobates lar (n ¼ 1)

Symphalangus syndactylus (n ¼ 1)

Gorilla beringei (n ¼ 5)

18:3n-3 (ALA) 20:5n-3 (EPA) 22:5n-3 22:6n-3 (DHA) Sum n-3 PUFA 18:2n-6 (LA) 18:3n-6 20:3n-6 20:4n-6 (AA) 22:5n-6 Sum n-6 PUFA Total PUFA Total fatty acids (mg/g)

2.32 0.08 0.40 0.21 3.17 24.74 ND 0.40 0.59 0.07 25.88 29.04 9.21

3.56  1.29 0.13  0.02 0.43  0.04 0.36  0.07 4.47  1.17 25.73  2.28 0.01  0.004 0.23  0.05 0.73  0.16 0.06  0.01 26.89  2.38 31.36  3.53 19.71  5.69

7.87 0.02 0.22 0.07 8.18 27.45 0.01 0.13 0.70 0.03 28.39 36.57 26.32

2.46  0.86 0.31  0.20 0.40  0.09 0.36  0.25 3.53  0.62 18.91  2.36 0.02  0.01 0.18  0.04 0.55  0.16 0.06  0.02 19.77  2.48 23.30  3.08 13.22  2.87

3.30 0.05 0.31 0.16 3.93 22.67 ND 0.13 0.52 0.03 23.42 27.35 14.54

2.02 0.10 0.38 0.25 2.89 22.76 0.04 0.21 0.44 0.05 23.59 26.48 12.54

16.31  1.19 0.22  0.04 0.69  0.08 0.09  0.04 17.13  1.19 10.56  0.54 0.004  0.01 0.13  0.02 2.08  0.15 0.12  0.01 12.90  0.46 30.03  1.51 19.82  4.34

a ALA ¼ a-linolenic acid; EPA ¼ eicosapentaenoic acid; LA ¼ linoleic acid; AA ¼ arachidonic acid; DHA ¼ docosahexaenoic acid. ND ¼ Not detected. Percent composition was calculated by dividing individual fatty acid concentrations (includes saturates and monounsaturates; see online supplementary information) by the total fatty acid concentration. For more details on fatty acid nomenclature see Fig. 1.

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

1091

Table 5 Mean (SE) percent composition of selected PUFA by superfamily and tests for significance among (ANOVA) and between (Tukey-Kramer) groups (samples from captive anthropoids only)a Fatty acid

Cercopithecoids (OW)

Ceboids (NW)

Hominoids (HM)

F statistic (p-value)

Significant differences between groups

18:3n-3 (ALA) 22:6n-3 (DHA) Sum n-3 PUFA 18:2n-6 (LA) 20:4n-6 (AA) Sum n-6 PUFA Total PUFA

1.55  0.25 0.44  0.02 2.40  0.21 23.88  0.96 0.44  0.03 24.75  0.76 27.15  0.89

2.26  0.25 0.28  0.02 2.90  0.18 28.02  0.94 0.74  0.03 30.44  0.63 33.34  0.73

3.52  0.35 0.22  0.03 4.13  0.29 23.51  1.33 0.65  0.05 24.27  1.01 28.41  1.18

10.32 (< 0.001) 16.67 (<0.001) 10.29 (<0.001) 6.18 (0.004) 23.07 (<0.001) 9.96 (<0.001) 6.83 (0.002)

HM vs. OW HM vs. NW OW vs. HM OW vs. NW HM vs. NW HM vs. OW NW vs. OW NW vs. HM OW vs. NW OW vs. HM NW vs. OW NW vs. HM NW vs. HM NW vs. OW

a

NW ¼ ceboids; HM ¼ hominoids; OW ¼ cercopithecoids. Significant differences at p < 0.05.

(Purina Mills, LLC, pers. comm.). Indeed, LA is the most prominent n-6 PUFA in Western human diets due to high consumption of foods containing these oils (Brenna, 2002). This is in contrast to Paleolithic hunter-gatherers whose diet is argued to have included higher levels of n-3 PUFA (O’Keefe and Cordain, 2004). Monkey Chow also usually contains fish meal and dried eggs as protein sources. Fish meal (which contains fish oils) is high in DHA, and eggs are high in AA, providing captive individuals with preformed sources of these LCPUFA. It seems unlikely that DHA in milks of captive individuals resulted from elongation and desaturation of ALA. Wild anthropoids with diets high in ALA (e.g., mountain gorilla, mantled howler) had milks with the lowest proportion of DHA. The captive golden lion tamarin provides an example for the effect of captivity on milk fatty acid profiles. This female was free ranging at the National Zoological Park, Washington, DC. Her foraging activity was more similar to her wild tamarin counterparts, as was her diet, which included insects, leaves, and fruit, but very little Monkey Chow (M. Power, pers. comm.). Subsequently, her milk fatty acid composition was more similar to wild than captive callitrichines, particularly in percent composition of LA and DHA (Milligan et al., 2008). Another pattern identified in both wild and captive milk samples was the negative correlation between the percent composition of ALA and DHA in milk. The correlation among wild groups was quite similar to that of captive groups, despite a significant difference between these groups in the percent composition of ALA. Nonhuman primate females appear to be limited in the amount of ALA they can elongate and desaturate into DHA. The ALA that was not synthesized into DHA could then accumulate in milk. Humans (Farquharson et al., 1992; Agostoni et al., 2001; Huang and Brenna, 2001) and baboons (Sheaff-Greiner et al., 1997; Brenna, 2002) have demonstrated inefficiency in conversion of ALA to DHA. Experiments in both species with [13C]- labeled ALA and DHA (reviewed in Brenna, 2002) indicated low conversion rates of ALA to DHA, and the majority of milk DHA came directly from [13C]-labeled dietary DHA. When combined with data from this study and Milligan et al. (2008), it could be suggested that anthropoid primates, as a whole, are inefficient in accreting synthesized DHA into milk.

Results from several studies suggest that the most efficient way to increase the proportion of milk DHA in nonhuman anthropoids is to increase DHA in the maternal diet. Despite the wide range of variation in ALA by superfamily among wild anthropoids, the mean proportion of DHA was nearly identical among superfamilies (Fig. 3). The percent composition of DHA in the milk of wild anthropoids may, thus, represent the maximum amount of DHA that anthropoid primates are able to synthesize from ALA, but does not represent the maximum amount that anthropoid primates can put into milk. Captive groups, with a source of preformed DHA in Monkey Chow, had significantly higher values of DHA (Fig. 3). This demonstrates that milk DHA levels are not conserved, but rather, are highly sensitive to dietary intakes of DHA. In addition, that there were identical values in the wild and captive groups for the lowest percent composition of DHA (0.03%) suggests there may be conservation in the minimum contribution of DHA to anthropoid primate milk. Human milk n-6 and n-3 PUFA and LCPUFA composition in a comparative context The range of values for DHA among captive anthropoids was 0.03–1.1 percent of total fatty acids and 0.03–0.37 percent of fatty acids among wild anthropoids. The same degree of variability was identified in the percent composition of DHA in human milk (0.06– 1.4%; Koletzko et al., 1992; Gibson and Makrides, 1999; Yuhas et al., 2006; Brenna et al., 2007). The proportion of DHA in milk was the most variable component among the human populations investigated by Yuhas et al. (2006), and ranged from 0.17 to 0.99% of total fatty acids. This finding is similar to reported ranges of DHA in milk of women from populations within Africa (0.1–0.9%) and Europe (0.1–0.6%; Koletzko et al., 1992). The authors of both studies attribute the variation in milk DHA to differences in fish consumption among populations. Del Prado et al. (2001) and Mitoulas et al. (2003) suggest that the consistent finding of variability in n-3 PUFA may indicate a dependence on these fatty acids from the immediate diet rather than body depot stores. Data from this study and Milligan et al. (2008) were also compared to human females consuming four different diets:

Table 6 Mean  SE percent composition (range of values) of ALA, DHA, LA, AA, n-3 PUFA, n-6 PUFA, and total PUFA for wild and captive anthropoids and tests for significance between groups Fatty acid

Captive

Wilda

t statistic (p-value)

18:3n-3 (ALA) 22:6n-3 (DHA) Sum n-3 PUFA 18:2n-6 (LA) 20:4n-6 (AA) Sum n-6 PUFA Total PUFA

2.22  0.18 (0.61–7.87) 0.35  0.02 (0.03–0.59) 2.96  0.58 (0.93–8.18) 25.47  0.65(11.50–35.11) 0.60  0.03 (0.35–1.04) 26.58  0.63 (12.90–36.22) 29.54  0.75 (13.82–41.92)

8.06  1.29 (1.01–19.34) 0.11  0.02 (0.03–37) 8.98  0.80 (1.39–21.65) 8.56  0.71 (1.79–13.97) 0.75  0.14 (0.24–3.00) 9.80  0.87 (2.20–15.65) 18.78  1.99 (4.14–37.31)

4.48 (0.001) 7.88 (<0.001) 6.09 (<0.001) 17.54 (<0.001) 1.01 (0.32) 15.61 (<0.001) 5.07 (<0.001)

a

Data from Milligan et al. (2008).

1092

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

Fig. 2. Mean percent composition of ALA of captive and wild anthropoids by superfamily (NW ¼ ceboids; HM ¼ hominoids; OW ¼ cercopithecoids). *p  0.05 vs. captive samples. Dark bars ¼ captive; light bars ¼ wild.

Fig. 4. Mean percent composition of LA of captive and wild anthropoids by superfamily (NW ¼ ceboids; HM ¼ hominoids; OW ¼ cercopithecoids). *p  0.05 vs. captive samples. Dark bars ¼ captive; light bars ¼ wild.

a typical Western diet (higher in fat, more fat from animals, less fish); a non-Western diet that included the consumption of fish (low animal fat and total fat, high consumption of fish); vegetarians (no meat or fish, but including dairy and eggs); and vegans (no fats originating from animals; Table 2). Among the four human dietary strategies, DHA varied widely and was lowest in vegans, who lacked any preformed source of this fatty acid, and highest in Nigerian women, who had access to both dried and fresh fish (Koletzko et al., 1991). Fatty acid profiles from captive anthropoids are similar to those of human females consuming a Western diet and those consuming fish (relatively lower in ALA and higher in DHA), and those of wild anthropoids are similar to those of human females consuming a vegan or vegetarian diet (relatively higher in ALA, lower in DHA). Excluding wild mountain gorillas, the range of values for AA percent composition in milk from nonhuman anthropoids (0.32–

0.79) is similar to that from human females (Table 2; 0.32 to 0.82), suggesting that metabolism of this fatty acid also is similar between humans and nonhuman primates. Human and nonhuman anthropoid milk n-3 and n-6 PUFA and LCPUFA appear to be intimately tied to dietary supply of these fatty acids. Further, human milk does not appear to be distinct from that of anthropoids in the proportion of these fatty acids; all anthropoids may have similar metabolic capabilities in transferring dietary n-3 and n-6 PUFA to milk. Results from this project suggest that low conversion of ALA to DHA by human mothers may be a shared physiological trait of all anthropoids. This data set can only speak to the relationship between maternal diet and milk fatty acids, and it is possible that compensatory mechanisms are present in neonates and infants to convert milk ALA into LCPUFA. The inefficiency of this conversion among

Fig. 3. Mean percent composition of DHA of captive and wild anthropoids by superfamily (NW ¼ ceboids; HM ¼ hominoids; OW ¼ cercopithecoids). *p  0.05 vs. captive samples. Dark bars ¼ captive; light bars ¼ wild.

Fig. 5. Mean percent composition of AA of captive and wild anthropoids by superfamily (NW ¼ ceboids; HM ¼ hominoids; OW ¼ cercopithecoids). *p  0.05 vs. captive samples. Dark bars ¼ captive; light bars ¼ wild.

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

adults and infants in both humans (Farquharson et al., 1992; Salem et al., 1996; Agostoni et al., 2001; Huang and Brenna, 2001; Plourde and Cunnane, 2007) and baboons (Sheaff-Greiner et al., 1997; Brenna, 2002) does not support this position, however. Among anthropoids, the most efficient way to increase DHA in milk and in tissues, including neural tissues, seems to be with a preformed source in the diet. Nonhuman anthropoids have the capacity to produce milk with comparable DHA levels to humans, if given a dietary supply of this fatty acid. Selection does not seem to have provided humans with a unique mechanism for supplying DHA during lactation, a finding that seems surprising in light of the supposed increased requirements for DHA in brain and central nervous system tissues among human infants relative to nonhuman primates (Gibson and Kneebone, 1981; Gibson and Makrides, 1999, 2000; Agostoni et al., 2001; Carlson, 2001). Our results do not speak to how much LCPUFA the infant was consuming nor do they indicate how LCPUFA are utilized by the developing anthropoid infants. Data on milk composition reflect what the mother is physiologically capable of manufacturing, which is not equivalent to what the infant is able to extract over the course of lactation. For example, human mothers with a high body mass index (BMI) produced milks of higher fat concentration than women of low BMI, but produced less milk (Barbosa et al., 1997). The overall fat consumption by infants of both populations was, thus, not significantly different. Conversely, if human infants ingest more milk than nonhuman primates, their intake of DHA will be higher despite similar proportions of this fatty acid in milk. Humans may also differ from nonhuman primates in PUFA metabolism. Selection may have acted on human infant metabolism, such that any DHA that was ingested was preferentially used for brain growth and development. One possible mechanism was suggested by Chamberlain (1996). He argued that the enzyme ethanolamine phosphotransferase, which sequesters n-3 LCPUFA, might have been under selection over the course of human evolution. It is also possible that selection operated on maternal physiology, to ensure the transfer of necessary building blocks for fetal and infant brain growth and development. Among humans, the amount of DHA and AA increases markedly in the central nervous system during the last trimester and the first year of postnatal life (Carlson, 2001). This pattern is different than that observed in rats, who obtain most of their DHA after birth (Carlson, 2001). It may be possible that larger relative brain size, and the pattern of rapid postnatal brain growth, selected for preferential placental transfer of LCPUFA among humans (and their ancestors who shared this brain growth trajectory; Crawford et al., 2003). Maternal plasma concentrations of DHA have been demonstrated to increase over the course of pregnancy (Burdge, 2004; Giltay et al., 2004; Stark et al., 2005), and the placenta is believed to selectively incorporate DHA and AA rather than elongate and desaturate their n-3 and n-6 PUFA precursors, respectively (Cunnane, 2005). Tentative support for preferential transfer of LCPUFA comes from examination of body fat composition of human neonates. At birth, human baby fat contains relatively low amounts of LA and ALA, despite their inclusion in maternal fat stores, and has higher levels of AA and DHA than maternal fat (Cunnane, 2005).

Implications for human evolution Martin (1995) predicted that an investigation of primate milks would reveal the biochemical requirements necessary for human brain growth, emphasizing LCPUFA, such as AA and DHA. His prediction was based on the larger relative brain size of humans and the unique trajectory of postnatal brain growth that involved increased requirements for AA and DHA (Innis, 2003). This comparative study of

1093

anthropoid milk fatty acid composition suggests that human milk fatty acid profiles fit within the larger anthropoid pattern. What may be unique in humans, compared to nonhuman anthropoids, is the inclusion of preformed sources of LCPUFA, particularly DHA, in the diet. There is currently much debate surrounding which foods would have provided DHA, or its precursor ALA, and when these foods would have been incorporated into the human diet (Crawford, 1992; Broadhurst et al., 1998, 2002; Crawford et al., 1999; Cordain et al., 2001; Cunnane 2005; Carlson and Kingston, 2006). A direct link between increased meat from terrestrial animals in the diet and milk composition is tenuous. Animal based fatty acids, particularly AA, would be predicted to increase in concentration and thus, relative proportion, in milk as more animal tissues were incorporated in the diet. However, increased milk AA does not influence infant tissue concentration (including the brain) of AA. Formula fed human and baboon infants with no source of AA had similar brain concentrations to those receiving AA through breast milk (Farquharson et al., 1992; Makrides et al., 1994; Sarkadi-Nagy et al., 2003, 2004). This suggests that a compensatory mechanism for AA synthesis in brain tissues may be part of our anthropoid, or at least Old World anthropoid, legacy. Thus, increased dietary AA is an unlikely prime releaser (Robson, 2004) for encephalization. Unlike AA, there is a strong correlation between consumption of foods rich in DHA, levels of milk DHA, and infant tissue concentrations of DHA in humans and nonhuman primates. Brain and marrow are terrestrial sources of this fatty acid (Cordain et al., 2001), and access to these parts of animal carcasses may characterize the dietary niche of the earliest members of the genus Homo (Blumenschine and Madrigal, 1993; Blumenschine, 1995; Brantingham, 1998; Madrigal and Blumenschine, 2000). In contrast, Broadhurst et al. (1998, 2002) and Crawford (1992; Crawford et al., 1999) argue that growth and maintenance of a larger brain required a diet based on marine and lacustrine fish and shellfish. Unlike terrestrial plants and animals, these foods are high in the brain specific nutrients DHA, iodine, zinc, copper, iron, and selenium of which DHA has received the most attention in the literature (Cunnane, 2005; Milligan and Bazinet, 2007). The aquatic diet hypothesis requires that encephalization (as well as optimal growth, development, and maintenance of a larger relative brain) would necessitate an increase in milk DHA, made possible by a consistent, predictable, and generous supply of DHA. In contrast to this model, Carlson and Kingston (2006) argue that a more consistent, albeit small, source of DHA in a terrestrial diet would result from elongation and desaturation of ALA, found in a wide variety of foods available to hominins. They propose that although the biosynthesis of ALA to DHA is inefficient among humans, it is sufficient for normal brain growth, development, and maintenance in modern humans (authors’ emphasis). Infants who receive milk without a source of DHA (e.g., infants fed formula that was not supplemented with DHA) – or infants consuming breast milk from vegan mothers – have to rely on conversion from ALA (either by the mother via breast milk or in vivo). Carlson and Kingston (2006) review a number of compensatory mechanisms that ensure a sufficient supply of DHA in maternal depot stores (e.g., the influence of estrogen on synthesis of ALA). Unlike Martin’s (1983) predictions regarding human milk composition, neither Cordain et al. (2001) nor Broadhurst et al. (1998, 2002; and see Cunnane, 2005) predict that milk of extant humans will diverge from that of nonhuman primates. Higher levels of milk DHA are predicted in any population that consumes foods with preformed sources of DHA. Evidence from captive nonhuman primates and human populations with a high consumption of marine based foods (including fish oil) supports this prediction. The finding of a negative correlation between the percent composition of ALA and DHA in nonhuman anthropoid

1094

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095

milk taken together with research on humans and baboons (SheaffGreiner et al., 1997; Su et al., 1999, 2005; Brenna, 2002; Burdge, 2006) suggests that inefficient conversion of ALA to DHA is a pleisomorphic trait of anthropoids. This finding neither supports nor refutes Carlson and Kingston’s (2006) hypothesis; the data do not speak to infant metabolic capabilities or other compensatory methods to ensure a sufficient supply of DHA. However, if high levels of DHA are found to have been important for hominin encephalization, our data would suggest that infants who suckled from mothers with a preformed dietary source of this fatty acid would be most likely to incur this benefit. Acknowledgements We thank Katherine Hinde, Olav Oftedal, Michael Power, Susan Gibson, Larry Williams, Tara Stoinski, Mark Edwards, Kathy Brasky, the Milwaukee Zoo, Zoo Atlanta, the Minnesota Zoo, Riverbanks Zoological Park, the St. Louis Zoo, the San Diego Wild Animal Park, and the National Zoological Park for providing milk samples for this project. We also thank anonymous reviewers whose comments greatly improved this manuscript. Financial support for this project was provided by L.S.B. Leakey Foundation Research Grant #7360, Wenner-Gren Foundation Dissertation Fieldwork Grant #1965, and the Natural Sciences and Engineering Research Council of Canada. This project was reviewed and approved by The University of Arizona IACUC, Central Animal Facility. Appendix. Supplementary data Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.jhevol.2008.07.010. References Agostoni, C., Marangoni, F., Lammardo, A.M., Galli, C., Giovannini, M., Riva, E., 2001. Long-chain polyunsaturated fatty acid concentrations in human hindmilk are constant throughout twelve months of lactation. In: Newburg, D.S. (Ed.), Bioactive Components of Human Milk. Plenum Press, New York, pp. 157–161. Barbosa, L., Butte, N.F., Villalpando, S., Wong, W.W., Smith, E.O., 1997. Maternal energy balance and lactation performance of Mesoamerindians as a function of body mass index. Am. J. Clin. Nutr. 66, 575–583. Bazinet, R.P., McMillan, E.G., Cunnane, S.C., 2003. Dietary a-linolenic acid increases the n-3 PUFA content of sow’s milk and the tissues of the suckling piglet. Lipids 38, 1045–1049. Blumenschine, R.J., 1995. Percussion marks, tooth marks, and experimental determinations of the timing of hominid and carnivore access to long bones at Flk Zinjanthropus, Olduvai Gorge, Tanzania. J. Hum. Evol. 29, 21–51. Blumenschine, R.J., Madrigal, T.C., 1993. Variability in long-bone marrow yields of East-African ungulates and its zooarchaeological implications. J. Archaeol. Sci. 20, 555–587. Brantingham, P.J., 1998. Hominid carnivore coevolution and the invasion of the predatory guild. J. Anthropol. Archaeol. 17, 327–353. Brenna, J.T., 2002. Efficiency of conversion of a -linolenic acid to long-chain n-3 fatty acids in man. Curr. Op. Clin. Nutr. Metab. Care 5, 127–132. Brenna, J.T., Diau, G.Y., 2007. The influence of dietary docosahexaenoic acid and arachidonic acid on central nervous system polyunsaturated fatty acid composition. Prostaglandins Leukot. Essent. Fatty Acids 77, 247–250. Brenna, J.T., Varamini, B., Jensen, R.G., Diersen-Schade, D.A., Boettcher, J.A., Arterburn, L.M., 2007. Docosahexaenoic acid and arachidonic acid concentrations in human breast milk worldwide. Am. J. Clin. Nutr. 85, 1457–1464. Broadhurst, C.L., Cunnane, S.C., Crawford, M.A., 1998. Rift Valley lake fish and shellfish provided brain-specific nutrition for early Homo. Br. J. Nutr. 79, 3–21. Broadhurst, C.L., Wang, Y.Q., Crawford, M.A., Cunnane, S.C., Parkington, J.E., Schmidt, W.F., 2002. Brain-specific lipids from marine, lacustrine, or terrestrial food resources: potential impact on early African Homo sapiens. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 131, 653–673. Burdge, G.C., 2004. Alpha-linolenic acid metabolism in men and women: nutritional and biological implications. Curr. Opin. Clin. Nutr. Metab. Care 7, 137–144. Burdge, G.C., 2006. Metabolism of alpha-linolenic acid in humans. Prostaglandins Leukot. Essent. Fatty Acids 75, 161–168. Carlson, S.E., 1999. Long-chain polyunsaturated fatty acids and development of human infants. Acta Paediatr. 88, 72–77. Carlson, S.E., 2001. Docosahexaenoic acid and arachidonic acid in infant development. Semin. Neonatol. 6, 437–449.

Carlson, B.A., Kingston, J.D., 2006. Docosahexaenoic acid, the aquatic diet, and hominin encephalization: difficulties in establishing evolutionary links. Am. J. Hum. Biol. 19, 132–141. Carnielli, V.P., Sauer, P.J.J., 1996. Essential fatty acid metabolism in neonates. In: Bindels, J.G., Goedhart, A.C., Visser, H.K.A. (Eds.), Recent Developments in Infant Nutrition. Kluwer Academic Publishers, London, pp. 173–181. Chamberlain, J.G., 1996. The possible role of long-chain, omega-3 fatty acids in human brain phylogeny. Perspect. Biol. Med. 39, 436–445. Clandinin, M.T., Chappell, J.E., Leong, S., Heim, T., Swyer, P.R., Chance, G.W., 1980a. Intrauterine fatty acid accretion rates in human brain: implications for fatty acid requirements. Early Hum. Dev. 4, 121–129. Clandinin, M.T., Chappell, J.E., Leong, S., Heim, T., Swyer, P.R., Chance, G.W., 1980b. Extrauterine fatty acid accretion rates in human brain: implications for fatty acid requirements. Early Hum. Dev. 4, 131–138. Cordain, L., Watkins, B.A., Mann, N.J., 2001. Fatty acid composition and energy density of foods available to African hominids: evolutionary implications for human brain development. In: Simopoulos, A.P. (Ed.), Nutrition and Fitness: Metabolic Studies in Health and Disease. Karger, Basel, pp. 144–161. Crawford, M.A., 1992. The role of dietary fatty acids in biology: their place in the evolution of the human brain. Nutr. Rev. 50, 3–11. Crawford, M.A., Bloom, M., Broadhurst, C.L., Schmidt, W.F., Cunnane, S.C., Galli, C., Gehbremeskel, K., Linseisen, F., Lloyd-Smith, J., Parkington, J., 1999. Evidence for the unique function of docosahexaenoic acid during the evolution of the modern hominid brain. Lipids 34, S39–S47. Crawford, M.A., Golfetto, I., Ghebremeskel, K., Min, Y., Moodley, T., Poston, L., Phylactos, A., Cunnane, S., Schmidt, W., 2003. The potential role for arachidonic and docosahexaenoic acids in protection against some central nervous system injuries in preterm infants. Lipids 38, 303–315. Cunnane, S.C., 2005. Survival of the Fattest: The Key to Human Brain Evolution. World Scientific, Hackensack. Del Prado, M., Villalpando, S., Elizondo, A., Rodriguez, M., Demmelmair, H., Koletzko, B., 2001. Contribution of dietary and newly formed arachidonic acid to human milk lipids in women eating a low-fat diet. Am. J. Clin. Nutr. 74, 242–247. Farquharson, J., Cockburn, F., Patrick, W.J.A., Jamieson, E.C., Logan, R.W., 1992. Infant cerebral cortex phospholipid fatty-acid composition and diet. Lancet 340, 810–813. Folch, J., Lees, M., Stanley, G.H.S., 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226, 497–509. Garber, P.A., Leigh, S.R., 1997. Ontogenetic variation in small-bodied new world primates: implications for patterns of reproduction and infant care. Folia Primatol. 68, 1–22. Gibson, R.A., 1997. Long-chain polyunsaturated fatty acids and infant development. Lancet 354, 1919–1920. Gibson, R.A., Kneebone, G.M., 1981. Fatty acid composition of human colostrum and mature breast milk. Am. J. Clin. Nutr. 34, 252–257. Gibson, R.A., Makrides, M., 1999. Long-chain polyunsaturated fatty acids in breast milk: are they essential? In: Newburg, D.S. (Ed.), Bioactive Components of Human Milk. Plenum Press, New York, pp. 375–383. Gibson, R.A., Makrides, M., 2000. n-3 polyunsaturated fatty acid requirements of term infants. Am. J. Clin. Nutr. 71, 251S–255S. Giltay, E.J., Gooren, L.J., Toorians, A.W., Katan, M.B., Zock, P.L., 2004. Docosahexaenoic acid concentrations are higher in women than in men because of estrogenic effects. Am. J. Clin. Nutr. 80, 1167–1174. Goodall, J., 1986. The Chimpanzees of Gombe: Patterns of Behavior. Harvard University Press, Cambridge. Huang, M.C., Brenna, J.T., 2001. On the relative efficacy of alinolenic acid and preformed docosahexaenoic acid as substrates for tissue docosahexaenoate during perinatal development. In: Mostofsky, D., Yehuda, S., Salem, N. (Eds.), Fatty Acids: Physiological and Behavioral Functions. Humana Press Inc., Totowa, pp. 99–113. Innis, S.M., 2003. Perinatal biochemistry and physiology of long-chain polyunsaturated fatty acids. J. Pediatr. 143, 1–8. Iverson, S.J., Oftedal, O.T., 1995. Phylogenetic and ecological variation in the fatty acid composition of milks. In: Jensen, R.G. (Ed.), Handbook of Milk Composition. Academic Press, San Diego, pp. 790–827. Jensen, R.G., Bitman, J., Carlson, S.E., Couch, S.C., Hamosh, M., Newburg, D.S., 1995. Human milk lipids. In: Jensen, R.G. (Ed.), Handbook of Milk Composition. Academic Press, San Diego, pp. 495–542. Kappeler, P.M., Pereira, M.E., 2003. Primate Life Histories and Socioecology. University of Chicago, Chicago. Koletzko, B., Agostoni, C., Carlson, S.E., Clandinin, T., Hornstra, G., Neuringer, M., Uauy, R., Yamashiro, Y., Willatts, P., 2001. Long chain polyunsaturated fatty acids (LC-PUFA) and perinatal development. Acta Paediatr. 90, 460–464. Koletzko, B., Mrotzek, M., Bremer, H.J., 1991. Fatty acid composition of mature human milk. Zeitschrift fu¨r ernahrungswiss – Evr. J. Nutr. 30, 289–297. Koletzko, B., Thiel, I., Abiodum, P.O., 1992. The fatty acid composition of human milk in Europe and Africa. J. Pediatr. 120, S62–S70. Lands, W.E., 1992. Biochemistry and physiology of n-3 fatty acids. FASEB J. 6, 2530–2536. Leigh, S.R., 1994. Relations between captive and noncaptive weights in anthropoid primates. Zool. Biol. 13, 21–43. Leigh, S.R., Shea, B.T., 1996. Ontogeny of body size variation in African apes. Am. J. Phys. Anthropol. 99, 43–65. Madrigal, T.C., Blumenschine, R.J., 2000. Preferential processing of high return rate marrow bones by Oldowan hominids: a comment on Lupo. J. Archaeol. Sci. 27, 739–741. Makrides, M., Neumann, M.A., Byard, R.W., Simmer, K., Gibson, R.A., 1994. Fatty acid composition of brain, retina, and erythrocytes in breast fed and formula fed infants. Am. J. Clin. Nutr. 60, 189–194.

L.A. Milligan, R.P. Bazinet / Journal of Human Evolution 55 (2008) 1086–1095 Martin, R.D., 1983. Human Brain Evolution in an Ecological Context. American Museum of Natural History, New York. Martin, R.D., 1995. Phylogenetic aspects of primate reproduction: the context of advanced maternal care. In: Pryce, C.R., Martin, R.D., Skuse, D. (Eds.), Motherhood in Human and Nonhuman Primates, 3rd Schultz-Biegert Symposium, Kartause Ittingen, 1994. Karger, Basel, pp. 16–26. Milligan, L.A., 2007. Nonhuman primate milk composition: relationship to phylogeny, ontogeny, and ecology. Ph.D. Dissertation, University of Arizona. Milligan, L.A., Bazinet, R.P., 2007. Is docosahexaenoic acid a red herring for the aquatic diet? Br. J. Nutr. 97, 1023–1024. Milligan, L.A., Rapoport, S.I., Cranfield, M.R., Dittus, W., Glander, K.E., Oftedal, O.T., Power, M.L., Whittier, C.A., Bazinet, R.P., 2008. Fatty acid composition of wild anthropoid primate milks. Comp. Biochem. Physiol., B. 149, 74–82. Mitoulas, L.R., Gurrin, L.C., Doherty, D.A., Sherriff, J.L., Hartmann, P.E., 2003. Infant intake of fatty acids from human milk over the first year of lactation. Br. J. Nutr. 90, 979–986. Oftedal, O.T., Iverson, S.J., 1995. Comparative analysis of nonhuman milks: phylogenetic variation in the gross composition of milks. In: Jensen, R.G. (Ed.), Handbook of Milk Composition. Academic Press, San Diego, pp. 749– 788. O’Keefe, J.H., Cordain, L., 2004. Cardiovascular disease resulting from a diet and lifestyle at odds with our Paleolithic genome: how to become a 21st-century hunter-gatherer. Mayo Clin. Proc. 79, 101–108. Plourde, M., Cunnane, S.C., 2007. Extremely limited synthesis of long-chain polyunsaturated fatty acids in adults: implications for dietary essentiality and use as supplements. Appl. Physiol. Nutr. Metab. 32, 619–634. Robson, S.L., 2004. Breast milk, diet, and large human brains. Curr. Anthropol. 45, 419–425. Ross, C., 2003. Life history, infant care strategies, and brain size in primates. In: Kappeler, P.M., Pereira, M.E. (Eds.), Primate Life Histories and Socioecology. University of Chicago Press, Chicago, pp. 266–284. Salem Jr., N., Wegher, B., Mena, P., Uauy, R., 1996. Arachidonic and docosahexaenoic acids are biosynthesized from their 18-carbon precursors in human infants. Proc. Natl. Acad. Sci. 93, 49–54. Sanders, T.A.B., 1999. Essential fatty acid requirements of vegetarians in pregnancy, lactation, and infancy. Am. J. Clin. Nutr. 70, 555S–559S.

1095

Sanders, T.A.B., Reddy, S., 1992. The influence of a vegetarian diet on the fatty acid composition of human milk and the essential fatty acid status of the infant. J. Pediatr. 120, S71–S77. Sarkadi-Nagy, E., Wijendran, V., Diau, G.Y., Chao, A.C., Hsieh, A.T., Turpeinen, A., Nathanielsz, P.W., Brenna, J.T., 2003. The influence of prematurity and long chain polyunsaturate supplementation in 4-week adjusted age baboon neonate brain and related tissues. Pediatr. Res. 54, 244–252. Sarkadi-Nagy, E., Wijendran, V., Diau, G.Y., Chao, A.C., Hsieh, A.T., Turpeinen, A., Lawrence, P., Nathanielsz, P.W., Brenna, J.T., 2004. Formula feeding potentiates docosahexaenoic and arachidonic acid biosynthesis in term and preterm baboon neonates. J. Lipid Res. 45, 71–80. Sellen, D.W., 2007. Evolution of infant and young child feeding: implications for contemporary public health. Ann. Rev. Nutr. 27, 123–148. Sheaff-Greiner, R.C., Winter, J., Nathanielsz, P.W., Brenna, J.T., 1997. Brain docosahexaenoate accretion in fetal baboons: bioequivalence of dietary a-linolenic and docosahexaenoic acids. Pediatr. Res. 42, 826–834. Specker, B.L., Wey, H.E., Miller, D., 1987. Differences in fatty acid composition of human milk in vegetarian and nonvegetarian women: long-term effect of diet. J. Pediatr. Gastroenterol. Nutr. 6, 764–768. Stark, K.D., Beblo, S., Murthy, M., Buda-Abela, M., Janisse, J., Rockett, H., Whitty, J.E., Martier, S.S., Sokol, R.J., Hannigan, J.H., Salem Jr., N., 2005. Comparison of bloodstream fatty acid composition from African-American women at gestation, delivery, and postpartum. J. Lipid Res. 46, 516–525. Su, H.-M., Corso, T.N., Nathanielsz, P.W., Brenna, J.T., 1999. Linoleic acid kinetics and conversion to arachidonic acid in the pregnant and fetal baboon. J. Lipid Res. 40, 1304–1312. Su, H.-M., Huang, M.-C., Saad, N.M.R., Nathanielsz, P.W., Brenna, J.T., 2005. Fetal baboons convert 18:3n-3 to 22:6n-3 in vivo: a stable isotope tracer study. J. Lipid Res. 42, 581–586. Yuhas, R., Pramuk, K., Lien, E.L., 2006. Milk fatty acid composition from nine countries varies most in DHA. Lipids 41, 851–858. Zihlman, A., Bolter, D., Boesch, C., 2004. Wild chimpanzee dentition and its implications for assessing life history in immature hominin fossils. Proc. Natl. Acad. Sci. U.S.A. 101, 10541–10543. Zihlman, A.L., Morbeck, M.E., Goodall, J., 1990. Skeletal biology and individual life history of Gombe chimpanzees. J. Zool. 221, 37–61.