Thermal cycling and data acquisition were performed with a Bio-Rad iCycler IQ system (Bio-Rad Laboratories Inc., Hercules, CA). Reverse transcription and real time PCR analysis were performed in a two tube assay. Reverse transcription was performed in duplicate using a Superscript First-Strand Synthesis System for RT-PCR kit (Invitrogen, Carlsbad, CA) according to the manufacturer’s recommendations, as previously described [10] .

Real time PCR was done in duplicate using the IQ Sybr Green Supermix kit (Bio-Rad) for each duplicate of tissue. A 24 µL reaction mix was made containing 12.5 µL sybr green supermix, 1.0 µL forward primer (10 µM), 1.0 µL reverse primer (10 µM) and 9.5 µL sterile water. This reaction mix was added to each well, followed by 1.0 µL RT product (25 µL total volume).

Parameters for all reactions except cyclophilin were as follows: 1 cycle 95˚C for 15 min (PCR activation), followed by 30 cycles, 94˚C for 15 s, 58˚C for 30 s, 72˚C for 30 s, with a final extension at 72˚C for 8 min. Parameters for cyclophilin were as follows: 1 cycle 95˚C for 15 min (PCR activation), followed by 30 cycles, 94˚C for 30 s, 55˚C for 30 s, 72˚C for 30 s, with a final extension at 72˚C for 7 min. Melting curve analysis was performed on all real time PCR reactions to confirm specificity and identity of the real time PCR products. A non-template control was run for every assay. Specificity of real time PCR products was further confirmed by agarose gel electrophoresis. The two-step real time PCR reactions were optimized for linearity (exponential amplification) from >20 to <30 cycles under the conditions described above.

Birth weights did not differ between fast and slow growing piglets (P > 0.05, Table 1). Litter size varied from 8

^*Significantly different from piglets with high growth rate (P < 0.05, n = 7). Data are represented by means with standard errors in parentheses.

to 13 pigs per litter with an average of 9.9 ± 0.7. Fast growing piglets weighed 1.61 ± 0.19 kg more at day 21 than slow growing piglets (P < 0.001). The litters average daily gain was 210 ± 7 g/day, while fast growing piglets gained an additional 28 ± 11 g/day (+13.3%; P > 0.05) and slow growing piglets gained 49 ± 9 g less per day (−24.4%, P < 0.001) than the rest of the litter. The difference in average daily gain between fast and slow growing piglets was 77 ± 9 g (P < 0.001).

Among the adipokines (Figure 1), leptin was 43% lower in the SQ adipose of slow growing piglets than fast growing piglets (P < 0.05, Figure 1(a)) while PR leptin mRNA abundance was not affected by growth rate (P > 0.05). Adiponectin gene expression was reduced by 36% in SQ of the slow growing piglet versus the fast growing piglet (P < 0.01, Figure 1(b)). Perirenal and longissimus mRNA abundance were similar between slow and fast growing piglets (P > 0.05). Tumor necrosis factor α gene expression was reduced by 36% in SQ and 40% in PR of slow growing piglets relative to fast growing piglets (P < 0.05, Figure 1(c)). The mRNA abundance of MIF was reduced by 56% in PR of slow growing piglets in comparison to tissue from fast growing piglets (P < 0.02, Figure 1(d)), whereas both SQ and longissimus MIF steady state concentrations were unaffected by growth rate (P > 0.05). The mRNA abundance of IL1β was increased by 57% in longissimus of slow growing piglets relative to fast growing piglets (P < 0.02, Figure 1(e)) while IL15 gene expression was increased in longissimus by 47% (P <0.01, Figure 1(f)). Adipose tissue expression of either of these adipokines was not altered by growth rate (P > 0.05). No differences in mRNA abundance were detected for the remaining adipokines that were examined (IL6, CCL2 and IGF1) between tissues of fast and slow growing piglets (P > 0.05; data not presented).

Gene expression of the enzyme LPL was reduced by 24% in SQ (P < 0.05) while increased by 71% in longissimus (P < 0.01) of slow growing pigs, relative to fast growing piglets (Figure 2(a)). The mRNA abundance for FASN was not affected by growth rate (P > 0.05; Figure 2(b)).

Numerous studies have described the expression of adipokines in adult animals and humans [4] [21] . Hausman et al. [9] have previous reported that 5 - 7 day old pig adipose tissue expresses a variety of adipokines. However, the present study is the first attempt to examine the relationship of growth rate and adipokine expression within adipose tissue in vivo during the preweaning period. The neonatal period is characterized by rapid growth rates

that are paralleled by rapid accumulation of lipid within newly formed adipocytes and the differentiation of preadipocytes into adipocytes [1] [12] [22] [23] . Slower growing and subsequently smaller piglets have less fat deposition during the preweaning period [12] [13] . This slower growth rate may be the consequence a difference in relative fetal maturity, health status, nutrition, response to stress or a combination of these factors.

The majority of adipose tissue deposition in swine is subcutaneous and this is reflected in the accumulation of SQ LPL mRNA in the present study, which is a marker for adipocyte differentiation and lipid metabolism [16] [24] [25] . Slower growing piglets had less SQ LPL mRNA abundance than fast growing piglets. Leptin and adiponectin mRNA abundance were also lower in the SQ of the slower growing piglets. Both genes are expressed almost exclusively within the adipocyte in swine [26] [27] . Thus the smaller abundance of SQ leptin, adiponectin and LPL mRNA by the SQ adipose tissue of the slow growing piglet during the first 21 days of life may reflect lower rates of lipid accumulation. Dual x-ray absorptiometry of additional pairs of pigs from other litters within this herd has confirmed that slower growing piglets have lower rates of lipid deposition than faster growing piglets [13] . This lower rate of fat deposition and change in gene expression could be the consequence of a reduced nutrient intake by the slow growing pig. Dietary restriction has been demonstrated to reduce adipose tissue expression of leptin [28] [29] and LPL [29] [30] . However, adiponectin mRNA abundance is increased with dietary restriction [31] [32] , while it was depressed in this study, suggesting intake was not reduced in slow growing pigs. In addition, the runt pig has a poor growth rate and suffers from an inability to compete for nutrition with its much larger littermates, yet no differences in the expression of any of these three genes was observed at 21 days of age between runt and normal growing littermates [11] . Unfortunately, dietary intake was not measured in the present study.

These changes in the SQ were not accompanied by similar changes in the PR. This may reflect the later development of the depot or unique differences between the two sites of deposition. Previous studies have reported differential gene expression by the SQ and PR in the neonatal pig [10] [11] . Numerous studies have documented the intrinsic differences in the development, physiology and metabolism between internal and external sites of adipose tissue deposition of swine [2] [33] -[35] . The physiological and metabolic demands of growth may produce developmental changes that cause internal and external sites of adipose tissue deposition to respond differently.

Tumor necrosis factor α has been the most thoroughly evaluated adipokine; TNFα can promote lipolysis, suppress lipogenesis and inhibit lipid uptake by adipose tissue [36] [37] . Differences in growth rate had an impact on TNFα mRNA abundance in both sites of adipose tissue deposition unlike LPL, leptin or adiponectin. Pigs with lower growth rates had lower TNFα mRNA abundance than pigs with higher growth rates. This parallels the reported effects of dietary restriction on adipose TNFα mRNA abundance [38] and would imply the slower growing piglet cannot compete for nutrition from the sow. However, no difference in adipose TNFα mRNA abundance was observed in adipose tissues of slow growing, poorly competitive runt pigs relative to normal growing littermates [11] . The observed differences in TNFα gene expression in the present study suggest that TNFα may be counter regulatory to the metabolic activity of the tissue. Previously Mitchell et al. [13] , using comparable pigs, demonstrated that the slow growing pigs had reduced lipid accumulation; thus suggesting that the consequences of lowering TNFα expression in those slower growing pigs may be to alter lipid accumulation and the partitioning of energy between adipose tissue and muscle.

Macrophage migration inhibitory factor was the only gene that demonstrated a difference in transcript level between slow and fast growing piglets exclusively in PR. The MIF mRNA abundance was much lower in the PR from slow than fast growing piglets. This is the opposite of what was observed in runt pigs at day 21 of age relative to average weight littermates. Expression of MIF is responsive to glucose and insulin status of the adipocyte [39] [40] . The MIF protein has been shown to both impair glucose uptake by the adipocyte and attenuate insulin sensitivity [41] . The consequences of the lower MIF mRNA abundance in the slow growing piglet would imply a paracrine counter regulatory attempt to increase the lower lipid accretion rate of the slow growing pig [12] [13] . However, MIF has numerous other functions that have yet to be evaluated within the context of adipose tissue biology [42] .

Contrary to the reduction in mRNA abundance for LPL in adipose tissue, gene expression was elevated in the longissimus muscle of the slow growing piglet. Lipoprotein lipase is an essential enzyme for the uptake and metabolism of fatty acids by skeletal muscle [43] . The increased LPL mRNA abundance in the slower growing piglet may imply an increased partitioning of lipid substrates to the skeletal muscle of the slower growing piglet, potentially in response to a lower feed intake. Restricted feed intake has been shown to elevate skeletal muscle LPL activity [44] . Slower growing piglets have been reported to have higher rates of lean deposition [12] . Dual x-ray absorptiometry of additional slow and fast growing pairs of pigs from other litters at our laboratory has confirmed that slower growing piglets have higher rates of lean deposition than faster growing piglets [13] . Thus the energy derived from lipid uptake by the skeletal muscle through the actions of LPL may be used for protein accretion.

Interleukin 1β has been reported to reduce muscle protein synthesis [45] and to reduce α-amino isobutyrate uptake [46] , indicative of a reduction in amino acid uptake, when injected into rats. However, these effects appear to be indirect and it has been proposed that IL1β only serves as a mediator [46] . The direct and specific actions of IL1β on skeletal muscle metabolism or growth have not been defined. The increase in IL1β mRNA abundance in the longissimus of the slower growing piglet relative to the fast growing piglet would imply an increase in protein breakdown, but since the slower growing pig has a higher rate of protein accretion [12] [13] , that would imply that overall protein turnover is increased in the smaller piglet versus the larger piglet. This would require validation with in vivo protein turnover experiments.

Interleukin 15 is highly expressed by skeletal muscle cells [47] and secreted [48] . The protein has been demonstrated to reduce protein breakdown [49] [50] and to stimulate muscle glucose uptake and lipid oxidation ([51] [52] . The greater IL15 gene expression in the slower growing piglet relative to the faster growing piglet suggests that this protein may be functioning to enhance energy utilization and protein deposition rates in the slower growing piglet. Dunshea et al. [12] and Mitchell et al. [2012] have reported that slower growing piglets have higher protein accretion rates in support of this hypothesis.

Previous studies have reported that slower growing piglets have reduced lipid accretion rates relative to their faster growing littermates, when comparing pigs born at similar birth weights. The present study supports those studies with the observation of reduced mRNA abundance for LPL, leptin and adiponectin expression in SQ adipose tissue from slow growing pigs relative to their faster growing littermates, suggestive of lower lipid accumulation within the adipose tissue. These changes seen in the SQ adipose tissue were not observed in the PR adipose tissue, indicative of developmental differences between the two sites of fat deposition. Limited effects of differential growth rates were seen for several of the adipokines analyzed; however, a reduction in TNFα mRNA abundance in the SQ adipose tissue may be a compensatory mechanism in response to the lower lipid accumulation rates reported for slower growing pigs. While an increase in muscle, IL-1β may be suggestive of an increase in protein turnover in the slow growing pig relative to the faster growing sibling. The increase in IL15 gene expression in association with an increase in LPL mRNA suggests skeletal muscle energy utilization and protein depositions are elevated in slow growing pigs; in agreement with the reported higher protein accretion rates of slower growing pigs.

The authors thank A. Shannon for her assistance with animal collection. The authors also thank the swine herd staff, USDA-Beltsville, for their work in animal care.