Main

Obesity is a serious metabolic disorder that affects millions of people in the United States, and it has been estimated that the annual economic costs associated with obesity in the United States alone are at ∼$69 billion (1). Recently, some alarming statistics have been reported from the National Health and Nutrition Examination Surveys (NHANES) revealing that the United States prevalence of obesity in adults is 33.4% in all groups and as high as 48% in both Mexican-American and African-American women with the trend continuing to rise (2). Similar NHANES studies have reported that the prevalence of obesity is 20% and also increasing in children and adolescents (3). Although a multitude of factors including environmental and genetic influences have been implicated in this increase, early infant diet may also be important, given well controlled studies now indicating that breast-fed infants are leaner than formula-fed control subjects (4).

The mechanism involved in the decreased body fat content exhibited in breast-fed infants has been attributed to lower energy intakes in these infants compared with those infants fed formula (4). Direct inhibitory effects of human milk constituents on fat cell differentiation, however, cannot be excluded, because known inhibitors of in vitro adipocyte differentiation, such as epidermal growth factor are present in human milk (5). To date, the potential for human milk to exert direct inhibitory influences on fat cell development has not been examined and formed the basis for this study.

Currently, it has not been determined if in fact leaner breast-fed infants have fewer adipocytes as opposed to smaller adipocytes. However, in studies of childhood obesity, a significant increase in the number of fat cells has been observed (6,7). In rodents, high caloric intake in preweaning animals can increase fat cell number, whereas in primates, increased fat cell mass, not number, is seen (8,9). The effects of maternal milk on adipocyte differentiation in offspring have not been investigated.

Established cell lines, such the murine 3T3-L1, 3T3-F442A, and Ob17 cells, have proven extremely useful in studies designed to elucidate mechanisms involved in adipocyte differentiation. That these cells represent appropriate models of adipocyte differentiation was obtained from in vivo transplantation studies, in which s.c. injection into mice led to the development of normal fat pads at the injection site [reviewed in Ailhaud et al. (10)]. The murine 3T3-L1 preadipocyte cell line has been well characterized for its ability to undergo complete differentiation into mature adipocytes after induction with an appropriate 4-d hormonal-pulse treatment regimen of insulin, 1-methyl-3-isobutylmethylxanthine, and dexamethasone (collectively designated IDX) (11–13). Based on results from clinical studies that indicated breast-fed infants are leaner than formula-fed control subjects, we hypothesized that factors in human milk may inhibit adipocyte differentiation. Therefore, in this study, using the established 3T3-L1 cell line as a model of adipocyte differentiation, we set out to test the direct effects of human milk as well as standard infant formula on the differentiation of adipocytes.

METHODS

Cell culture and adipocyte differentiation. 3T3-L1 fibroblasts (American Type Culture Collection) were grown in a standard growth medium that included Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin (each from GIBCO BRL) in a 5% carbon dioxide atmosphere. Cells were grown in either two-well chamber slides for the morphologic and proliferation experiments, or in 100-mm tissue culture plates. Postconfluent cells were hormonally stimulated to differentiate by treatment with IDX (1.7 µM insulin, 0.5 µM dexamethasone, and 0.5 mM isobutylmethylxanthine) (each from Sigma Chemical Co.) as previously described (14). Briefly, 4-d postconfluent cells (d 0) were treated with standard growth medium described above supplemented with IDX for 3 d. After 3 d, this medium was replaced by medium supplemented with 1.7 µM insulin only. Typically by d 4, >95% had differentiated into adipocytes as determined by lipid accumulation with Oil Red O staining and induction of adipocyte-specific mRNAs such as aP2 and adipsin as described below.

Cell treatment with formula or human milk. Ready-to-feed formulations from multiple lots of Enfamil with Iron (Mead Johnson) and Similac with Iron (Ross) were used. Donated term human milk samples from the first 3 wk of lactation were available from mothers whose infants had expired due to severe asphyxia or multiple anomalies and from whom consent had been obtained in accordance with guidelines from the Institutional Review Board, University of Arkansas for Medical Sciences. These samples, initially collected by a low pressure electric breast pump, were frozen at -20°C. For study these samples were thawed, pooled, and tested at a concentration of 10% in routine culture medium alone or in combination with IDX. Formula was also used at a concentration of 10% in routine culture medium alone or in combination with IDX. For the skim milk studies, whole milk samples were centrifuged at 680 × g for 10 min, lipids and cells were removed, and the aqueous layer was used at a concentration of 10% in routine culture medium. Attempts to similarly centrifuge and separate formula samples for testing in a similar fashion were unsuccessful as no physical separation occurred.

Oil red O staining. For morphologic assessment of differentiation, cells grown in dual chamber slides were washed, fixed in 10% neutral-buffered formalin, washed in 50% isopropanol, and then stained for 10 min in a solution of Oil Red O (Sigma Chemical Co.). Slides were washed in isopropanol and water and then counterstained in Mayer's hematoxylin (Fisher).

Immunocytochemistry cell proliferation assay. Cell proliferation and DNA synthesis was assessed by monitoring incorporation of BrdU. Briefly, cells were grown in dual chamber slides and exposed to control medium, IDX, formula (10%), or human milk (10%) for 24 h. Cells were subsequently exposed to BrdU, and incorporation was assessed using immunocytochemistry with the Cell Proliferation Kit (Amersham) exactly as described by the manufacturer.

RNA isolation and Northern blot analysis. Total cellular RNA was isolated, electrophoresed, and transferred as previously described (14). Blots were hybridized using standard conditions, with cDNA probes encoding either the complete mouse adipsin mRNA or mouse aP2 mRNA that were random-primer labeled to a high specific activity with [α-32P]dATP, as previously described (15). Blots were stripped and subjected to a second round of hybridization with a murine actin cDNA probe for normalization of aP2 and adipsin signals.

RESULTS

Enfamil and whole human milk promote 3T3-L1 lipid accumulation. 3T3-L1 cells were plated in 2-well chamber slides and allowed to reach confluence. Experimental conditions included control cells left untreated in standard culture medium, or cells treated with standard culture medium supplemented 4 d with either the hormonal regimen of IDX, 10% Enfamil, 10% Similac, or 10% whole human milk as described in "Methods." As expected, and in contrast to cells grown in culture medium alone (Fig. 1A), cells grown in the presence of IDX underwent complete adipocyte differentiation by d 4 as assessed by accumulation of multilocular intracytoplasmic lipid droplets detected by lipid staining with Oil Red O (Fig. 1B). Interestingly, a large number of cells grown in the presence of 10% Enfamil also developed lipid droplets, although differentiation was not as complete as was observed with IDX treatment (Fig. 1C). In addition these lipid droplets also appeared to be larger than those seen with typical IDX-induced differentiation and were occasionally unilocular (Fig. 1C). In contrast to Enfamil-treated cells, those grown in the presence of 10% Similac did not differentiate and morphologically were very similar to cells grown in standard culture medium (Fig. 1D). In addition, exposure of cells to 10% whole human milk resulted in considerable adipocyte differentiation and developed large intracytoplasmic lipid droplets (Fig 1E). Skim human milk did not promote differentiation, suggesting that the promoting activity resides in the cream fraction. In addition, combinations of either human milk or formula with differentiation mixture had no negative effect on differentiation (data not shown). These and subsequent experiments described below were repeated three to seven times with similar results. Results were not affected by different formula lot numbers, and in addition, each of five experiments with human milk used a different random pool of samples (i.e. collected at different times of the day, from different individuals, different days of lactation, and maternal age).

Figure 1
figure 1

Effect of infant formulas and human milk on lipid accumulation on d 4 of 3T3-L1 adipocyte differentiation. 3T3-L1 cells were plated in two-well chamber slides and allowed to reach confluence. After 4 d of treatment, cells were subsequently fixed and stained with Oil Red O as described in "Methods" (Representative photomicrographs at 400× magnification.) Treatment conditions included: (A) standard culture medium alone; or standard culture media supplemented with (B) the standard hormonal IDX differentiation regimen, (C) 10% Enfamil, (D) 10% Similac, or (E) 10% whole human milk.

Full size image

Enfamil and whole human milk stimulate 3T3-L1 mitotic clonal expansion. The above results strongly suggested that both Enfamil and human milk stimulated 3T3-L1 adipogenesis; however, similar results would be expected if either of these treatments induced a nonspecific intracellular lipid uptake. After stimulation of differentiation, growth-arrested 3T3-L1 cells proceed through the G1/S boundary, reentering the cell cycle, and undergo a mitotic expansion phase within the first ∼24 h that is essential to completing terminal differentiation into mature adipocytes (16). Because reentry into the cell cycle is required for differentiation, we wanted to determine whether or not these treatments had an effect on the early mitotic clonal expansion phase. Cell proliferation and DNA synthesis were assessed in these cultures by monitoring incorporation of BrdU using immunocytochemistry with a BrdU MAb. Cells grown in dual chamber slides were exposed for 24 h to the same treatment regimens as described above; either control medium, or medium supplemented with IDX, 10% Enfamil, 10% Similac, or 10% human milk. As previously described (15,17), untreated cells are predominantly quiescent as evidenced by few BrdU-positive cells (Fig. 2A), whereas virtually all the IDX-stimulated cells are BrdU-positive (Fig. 2B). Consistent with induction of differentiation, large numbers of cells treated 24 h with both Enfamil and human milk were actively proliferating and stained BrdU-positive (Fig. 2, C and E). The number of cells actively proliferating after treatment with Similac were indistinguishable from control unstimulated cells (Fig. 2D).

Figure 2
figure 2

Effect of infant formulas and human milk on mitotic clonal expansion on d 1 of 3T3-L1 adipocyte differentiation. 3T3-L1 cells were plated on two-well chamber slides and after reaching confluence, were treated for 24 h with: (A) standard culture medium alone; or standard culture media supplemented with (B) the standard hormonal IDX differentiation regimen, (C) 10% Enfamil, (D) 10% Similac, or (E) 10% whole human milk. After a 30-min incubation to allow incorporation of BrdU into replicating DNA, nuclear BrdU was detected by immunocytochemistry as described in "Methods" (magnification, 200×).

Full size image

Enfamil and Whole Human Milk Stimulate Expression of Differentiation-Dependent Adipocyte mRNAs. Activation of 3T3-L1 adipocyte differentiation is accompanied by dramatic increases in the expression of many differentiation-specific mRNAs (16,18). Northern blots were used to further characterize the effect of human milk and infant formulas on adipocyte 3T3-L1 differentiation. These experiments used complete cDNA probes encoding two well characterized mRNAs known to be up-regulated during 3T3-L1 adipocyte differentiation, the lipid-binding protein, aP2, and adipsin, a serine protease associated with the complement system (16,18). Total cellular RNA was isolated from cells after 4 d of either no treatment, or treatment with either IDX, 10% Enfamil, or 10% Similac. Northern blots from three separate experiments are shown in the three panels of Figure 3, the upper two panels probed with the aP2 cDNA and the lower panel probed with adipsin. As expected, when compared with untreated cells, IDX stimulation resulted in activation of both aP2 and adipsin mRNAs (Fig. 3, lanes 3 and 4 in each panel). Again, and consistent with stimulation of adipogenesis, Enfamil, and not Similac, induced expression of the differentiation-specific aP2 and adipsin mRNAs (Fig. 3, lanes 1 and 2 in each panel).

Figure 3
figure 3

Effect of infant formulas on expression of adipocyte-specific mRNAs. Postconfluent 3T3-L1 cells were treated 4 d in standard culture medium supplemented with either: (E) 10% Enfamil, (S) 10% Similac, (I) IDX, or (NA) standard culture medium alone. Total cellular RNA was isolated at d 4, and 20 µg were subjected to Northern blot analysis. Results shown represent three separate experiments. Filters were hybridized to [α-32P]dATP-labeled probes encoding either the complete aP2 cDNA (upper two panels, or the complete adipsin cDNA (lower panel). After hybridization with either aP2 or adipsin cDNAs, blots were stripped and reprobed with a cDNA encoding murine actin to control for sample loading.

Full size image

Similar Northern blots using total cellular RNA isolated from cells treated 4 d with 10% whole human milk also resulted in increased expression of the aP2 mRNA (Fig. 4, right panel). In contrast to whole human milk, defatted skim milk completely abolished the capacity of whole human milk to stimulate 3T3-L1 aP2 mRNA expression (Fig. 4, left panel).

Figure 4
figure 4

Effect of whole vs skim human milk on expression of adipocyte-specific mRNAs. Postconfluent 3T3-L1 cells were treated 4 d in: (NA) standard culture medium; or standard culture medium supplemented with (I) IDX; or (H) human milk (left panel), 10% skim milk or (H, right panel) 10% human whole milk. Total cellular RNA was isolated at d 4, and 20 µg were subjected to Northern blot analysis. Results shown represent two separate experiments (skim milk in left panel and whole milk on right panel. Filters were hybridized to [α-32P]dATP-labeled probes encoding the complete aP2 cDNA. After hybridization with the aP2 cDNA, blots were stripped and reprobed with a cDNA encoding murine actin to control for sample loading.

Full size image

DISCUSSION

The exact influence of early infant diet on fat cell development in humans remains to be clarified. Numerous animal studies have revealed differences in fat depot mass between breast- and formula-fed offspring, as well as alterations in cholesterol metabolism (9,19,20). In recent human studies, breast-fed infants were leaner than formula-fed control infants throughout the second year of life (4). These results suggested that there are components present in human milk that may exert a negative influence on adipocyte development. For example, epidermal growth factor, known to either stimulate or inhibit 3T3-L1 adipocyte differentiation depending on the time of exposure, has been demonstrated to be present in human milk in levels ranging from 25 nM in colostrum to 10 nM in mature human milk (21). In addition, TNF-α has been measured in mature human milk at concentrations of 620 pg/mL (22). Numerous in vitro studies have established that TNF-α exerts a potent inhibition on adipocyte differentiation (23–25). TNF-α has also been demonstrated to be important in several models of apoptosis (26,27) and may have contributed to the induction of apoptosis reported in various cell lines after exposure to human milk (28). Given the presence of these known inhibitory agents (and potentially others, such as hormones, cytokines, or growth factors) in human milk, we hypothesized that human milk might directly inhibit adipocyte differentiation and could, therefore, contribute to a reduced body fat content in breast-fed infants. Using the well characterized 3T3-L1 preadipocyte cell line, an in vitro model of adipocyte differentiation, we tested the direct effects that whole and skim milk fractions of human milk as well as two standard infant formulas have on 3T3-L1 adipogenesis. Contrary to our initial hypothesis, we observed a differentiation-stimulating effect in whole milk that was lost when most of the fat fraction was removed. To our surprise, an equally potent differentiation-promoting effect was identified in Enfamil that was not present in Similac.

The ability of human milk and Enfamil to independently stimulate adipocyte differentiation in vitro was characterized at several levels, both morphologically and biochemically. To our knowledge, these results represent the first demonstration of such an effect of infant dietary supplements on adipocyte differentiation. Similar in vitro studies using other cell lines such as fetal intestinal cells have assessed the potential growth-promoting activity of human milk or formula (29). In these studies the mitogenic potential of human milk was attributed to the presence of epidermal growth factor (29). Similar studies examining the effects of human milk on lymphocytes have also been performed and have revealed that factors in human milk can induce proliferation of lymphocytes as well as affect their function (30,31).

The mechanisms responsible for the promotion of 3T3-L1 adipocyte differentiation by formula or human milk remain to be fully elucidated, but potentially reside in their respective lipid components, given some similarities in the fatty acid composition between Enfamil and human milk. This hypothesis is supported by results presented here demonstrating the diminished capacity of skim milk to induce differentiation. Furthermore, although there are numerous factors in human milk capable of stimulating adipocyte differentiation, many of these components such as arachidonic acid, insulin, and thyroid hormone would not be found in infant formula (5). In addition, during the course of 3T3-L1 adipogenesis, dramatic alterations occur in gene expression (32), providing a multitude of possible sites of activity for formula or human milk-derived lipid components. For example, the PPARs have recently been demonstrated to be involved in regulating 3T3-L1 differentiation and are expressed early during the course of differentiation (33–35). When activated by lipid-like compounds such as fatty acids or arachidonic acid analogs, PPARs promote the terminal differentiation of preadipocytes (36–38). Therefore, a possible mechanism for the differentiation-promoting effect of Enfamil and human milk could reside in their fatty acid effects on PPAR expression. Studies have also demonstrated the capability of certain fatty acids to directly affect the expression of adipocyte-specific genes such as aP2, a lipid binding protein involved in maintenance of the mature adipocyte phenotype (39,40). Palmitic acid, another important lipid found in various quantities in both human milk and infant formula (including Enfamil and Similac), has been shown to independently induce adipocyte differentiation in similar in vitro studies (41). In this study, the potential influence of lipid or lipid-like substances during in vitro differentiation was highlighted by the loss of adipogenic activity when human milk was centrifuged and the lipid and cellular components were removed. Similar attempts to remove the lipid component of formula were unsuccessful and were therefore not tested.

The relevance of these studies to the intact individual remain to be determined. Clearly these in vitro studies indicating that human milk and Enfamil are similar in their differentiation-promoting activity are not in agreement with clinical studies comparing body fat in infants fed formula versus human milk. This discrepancy may likely reside in the fact that although some components of human milk are absorbed intact, modifications during absorption may lead to components that exert inhibitory actions in vivo. The model as used in these studies would not be able to examine the components that are modified by intermediate events within the gastrointestinal tract and periphery. Other factors present in human milk and not present in formula may be more important in affecting infant fat development. An example of such a factor is leptin, a protein produced by adipocytes with potent satiety effects, which has been recently identified in human milk (42–44). This protein may be important in regulating energy intake if absorbed and biologically active. Regardless of the role of leptin or other factors in human milk yet to be identified, studies such as these described herein suggest that more information is needed to fully understand the effect of early infant diet on adipose tissue development.