Influencing the future: interactions of skeleton, energy, protein and calcium during late gestation and early lactation

Additional keywords: calcium, health, metabolism, osteocalcin.

Lee et al. (2007) published a series of studies in mice presenting evidence of metabolic effects, orchestrated by osteocalcin (OC) produced by mature osteoblasts (OB), which influence regulation of energy metabolism. They raised the teleological question, ‘why would a bone-specific hormone regulate energy metabolism?’ We suggest in this paper, that the role of the skeleton in energy metabolism in dairy cattle reflects an essential need to integrate the homeorhetic changes that are required to upregulate metabolism in response to the demands of lactation. Homeorhetic changes are defined as the ‘coordinated changes in metabolism of body tissues necessary to support a physiological state’ (Bauman and Currie 1980). Evidence from basic and applied studies is provided to support the concept of homeorhesis. The period of most pronounced change in cattle from non-lactating to lactating occurs from 3 to 4 weeks before calving to 3 to 4 weeks after calving and is called the ‘transition’ period. Apart from examining the role of skeleton, we briefly examine other influences affecting the transition period that stimulated long-term metabolic change as indicated by increased milk production: specifically hormonal manipulation, factors that may change the nutrient demands of the conceptus, and the effects of oxidative damage and inflammation.

Evidence is provided from published papers and new data to support the hypothesis that the skeleton plays an important role in homeorhetic adaptation to lactation and that this relationship may be influenced by nutrition during the transition period. Recent developments in understanding of the role of the skeleton in the regulation of glucose homeostasis are discussed in the context of the role of calcium (Ca) and vitamin D in metabolism. The effects of pre-calving nutritional interventions are reviewed, using quantitative methods, to provide estimates of milk production responses. These positive, prolonged responses are considered in the context of the potential role of the skeleton in metabolism.

The modern dairy cow represents an extreme example of lactation, with many cows now producing more than 1.5 kg of each of milk fat, milk protein and milk lactose per day in early lactation. Cattle in early lactation need a highly integrated metabolism to ensure that production is sustained, health is optimal and that conception can occur when optimal for profit. The underlying physiology of the peri-parturient cow (Bell 1995), vulnerability to disease (Curtis et al. 1983, 1985) and methods of providing environments that result in better production and health have been extensively reviewed (Lean et al. 2003). Prospective studies of dairy cattle exposed to transition diets found marked increases in milk production (DeGaris et al. 2008; DeGroot et al. 2010) and improved reproductive performance (DeGaris et al. 2010a). The transition diets described in DeGaris et al. (2008) provided increased dry matter (DM) intake, estimated energy and metabolisable protein balances that were positive, macro- and micro-elements that were formulated to meet or exceed National Research Council (2001) requirements and a negative dietary cation anion difference (DCAD) balance. Increasing exposure to the pre-partum transition diets increased milk, milk fat and protein yield. The increase in milk production for optimal days of exposure to the transition diet (22 days for milk yield and 25 days for milk protein yield) against minimal exposure (3 days or less) was 3.75 L of 4.0% fat and 3.2% protein-corrected milk per day and 100 g of milk protein per day up to Day 150 of lactation. DeGaris et al. (2010a) also found that exposure to the transition diet increased risk of conception by 1.2% per day on the transition diet. Figure 1 shows the cumulative pregnancy rate for cows exposed to the diet for less than 10 days, to those exposed for 10–20 days and those cows exposed for more than 20 days. The cohort studies provide strong evidence that exposure to transition diets has positive effects on cattle, but does not provide evidence that any particular component of the diet caused such benefit. Notwithstanding the impressive body of information on the underlying physiology of the transition period, the longer-term responses, as expressed in milk production or fertility, require further evaluation. In particular, the roles of Ca metabolism, lipid metabolism and factors influencing gluconeogenesis, including control of insulin release, tissue sensitivity to insulin and the integration of these processes with other metabolism requires further consideration. Table 1 provides a list of transition interventions that could result in prolonged elevation of milk production and mechanisms that may influence these.

Fig. 1.  Cumulative pregnancy rate for cows exposed to a pre-partum transition diet providing a negative dietary cation anion difference for <10 days, 10–20 days and >20 days.

Table 1.  Potential interventions pre-calving that have a prolonged effect on milk production and reproduction in lactation

Detailed examination of evidence pertaining to particular aspects of diet is warranted, particularly in the context of recent understandings of the role of bone in the regulation of metabolism in mice models that should equally apply to the cow (McNeill and Anderson 2012). In order to investigate the mechanisms by which an intervention, limited to the period immediately before calving can influence future production and reproduction, we evaluated the effects of Ca and vitamin D on bone and the integration of energy metabolism; examined the effects of pre-calving DCAD and macro-mineral content of the diet on milk production and undertook further analysis of data collected by DeGaris et al. (2010b).

The possibility that the skeleton was involved in energy homeostasis and homeorhesis was first postulated when it was observed that obesity reduced the risk of osteoporosis in humans (Felson et al. 1993). Ducy et al. (2000) proposed that bone and energy metabolism may be regulated by the same hormones. An extensive series of Murine studies identified actions of OC produced by mature OB that completed a negative feedback loop between bone and energy metabolism, that is the hallmark of homeostatic regulation (Lee et al. 2007). The carboxylated form of OC has a high affinity for bone and is considered biologically inactive, whereas the uncarboxylated form (uOC) is biologically active and promotes β-cell proliferation, insulin secretion and independently increases peripheral tissue insulin sensitivity and stimulates adiponectin secretion by adipose cells (Lee et al. 2007). Adiponectin increases OB proliferation and differentiation (Berner et al. 2004) and increases bone deposition (Kanazawa et al. 2007). Further, adiponectin increases glucose uptake by skeletal muscle and may suppress hepatic gluconeogenesis (Yamauchi et al. 2002). Figure 2 shows understandings of interactions between skeleton, energy and protein metabolism and the influence of these effects on metabolism in general.

Fig. 2.  This figure modified from Wolf (2008) shows the relationships between skeleton and energy metabolism. The potential role for acidifying or negative dietary cation anion difference diets has been included and the effects of skeleton on reproductive metabolism, identified to date, are outlined.

Insulin acts to directly inhibit OB activity, thereby enhancing bone resorption (Lee et al. 2007). As a result of these actions, the extracellular environment within the absorption lacunae is acidified and increases the vitamin K-dependent decarboxylation of OC released from the bone matrix, thus although OC is produced by OB, decarboxylation of OC to the active form is determined by osteoclast activity. This may indicate that bone resorption is the key link between bone and energy metabolism. Changes in the concentrations of markers of bone metabolism in the plasma and urine, suggest that the most likely time to replenish bone is mid–late lactation, and to mobilise bone is late pregnancy and early lactation (McNeill et al. 2002; Bhanugopan et al. 2010).

Adipose tissue also influences bone metabolism (Fig. 2). Leptin, produced by adipocytes, acts to inhibit bone mass accrual through the action of the sympathetic nervous system on osteoblast β2-AR receptors after central processing involving neuropeptide Y (NPY) and neuromedian U to inhibit osteoblast activity (Takeda et al. 2002). NPY acts directly to inhibit OB and stimulate adipogenesis through non-hypothalamic Y1 receptors (Baldock et al. 2007). Leptin also has a potent anorexigenic effect and enhances reproduction possibly through increased expression of Kiss1 and resultant increase in gonadotrophin releasing hormone (GnRH) release (Popa et al. 2008). Leptin stimulates LH release in fasted cows and ewes (Nagatani et al. 1998; Williams et al. 2002). In contrast NPY is orexigenic and decreases release of GnRH and luteinising hormone (LH) and interestingly, oestrogen has been shown to both increase NYP production, but also be anorexogenic (Bonavera et al. 1994).

There are limited data to evaluate the role of bone in energy metabolism in cattle; however, there are findings that support the hypothesis that these are linked. Heuer et al. (1999) found that obese cows [body condition score (BCS) >4.5/5] were at greater risk of milk fever and DeGaris et al. (2010b) found a positive relationship between BCS and area under the curve for blood Ca concentrations after calving when BCS at calving was less than 3.0/5. Martinez et al. (2014) induced sub-clinical hypocalcaemia (SCH) by infusion of 5% ethylene glycol tetra-acetic acid, a selective ionised Ca chelator in non-lactating Holstein cows. These SCH cattle had lower DM intake, higher blood glucose concentrations and higher plasma non-esterified fatty acids (NEFA) than the cows infused with saline and fed 43 g of Ca as sulfate and chlorides. Physiological concentrations of Ca in blood may be necessary for glucose stimulation of insulin secretion by β-cells (Capen and Rosol 1989); Martinez et al. (2014) found lower insulin in the infused SCH cows.

Adiponectin concentrations were relatively low in early lactation and pre-partum and decreased in response to diets designed to induce a negative energy balance, whereas there was little change in adiponectin concentrations induced in mid or late lactation by diets creating a similar energy deficit (Singh et al. 2014). Singh et al. (2014) also observed a negative correlation between BCS and adiponectin and a positive correlation between plasma insulin and adiponectin. The insensitivity of dairy cattle in early lactation to insulin is well recognised. Bigner et al. (1996) found an increase in insulin resistance in cows fed low DCAD rations with a difference of >500 meq/kg to a high DCAD ration group after calving, although Grunberg et al. (2011) found no such association in a study conducted in groups of cows with a difference in DCAD of ~200 meq/kg and –100 meq/kg in the negative DCAD group. The role of bone and Ca metabolism in mediating the effects of body tissue reserves, lipid and protein, and energy and protein fluxes obtained from feed in the peri-parturient period clearly needs further exploration in cattle.

Support for a possible role for Ca in the increased productive and reproductive performance can be derived from several studies. Clinical milk fever was associated with increased days to first service, conception and number of services per conception (Borsberry and Dobson 1989) and increased risk of reproductive disorders (Erb et al. 1985). SCH was associated with increases in risk of uterine diseases in dairy cows (Martinez et al. 2012). In New Zealand, McKay (1994) supplemented pasture-fed cows with Ca-enriched molasses and found a 12% increase in first service conception rate for cows of parity 3 and greater, compared with controls and Stevenson et al. (1999) found a marked reduction in time from planned start of mating to conception for cows treated with Ca chloride in one of three herds studied. In the US, it has been reported that Wang et al. (1991) found a 17% increase in conception rates in 510 cows in a randomised controlled study of DCAD intervention before calving. The results of inducing SCH extend beyond the reduced DM intake and impaired energy metabolism to altered immune function as indicated by reduced phagocytic and killing activities of neutrophils (Martinez et al. 2014).

In addition to elucidating the homeostatic regulatory pathways acting between bone and energy metabolism, Lee et al. (2007) demonstrated a homeorhetic relationship between energy and bone metabolism; β-cell proliferation increased by between 60% and 300% over controls by 30 days of age in mice selected for a high level of OC endocrine activity. While the concept of homeorhesis is well established (Bauman and Currie 1980), it has rarely been specifically tested in cattle. The time series cross-correlation analysis of Lean et al. (1992) demonstrated for the first time temporal relationships between 3-hydroxybutyrate (BHB), glucose, cholesterol and free fatty acids (FFA), over a periods as long as 9 days, that were consistent with homeorhetic change in response to increasing milk production and decreasing estimated net energy balance.

Negative DCAD diets have been used to reduce the risk of hypocalcaemia in periparturient cattle (Block 1984; Lean et al. 2006). Apart from the effects in reducing the risk of hypocalcaemia, negative DCAD diets increase mobilisation of bone as indicated by bone markers (Block 1984; Goff and Horst 1997; Leclerc and Block 1989). Other effects of lowering the DCAD of pre-calving diets include induction of metabolic acidosis in goats (Fredeen et al. 1988a, 1988b) and cattle (Gaynor et al. 1989), an action consistent with the strong ion theory of acid-base balance. There is decreased renal sensitivity to parathyroid hormone (PTH) in cows fed a strongly positive DCAD pre-calving diet (Gaynor et al. 1989; Goff et al. 1991) and evidence of increased Ca homeostatic responses to PTH injection in cows fed a negative DCAD diet (Goff et al. 2014) leading those authors to conclude that hypocalcaemia and milk fever are, in part, a diet induced pseudo-hypoparathyroidism. Further, there is enhanced renal production of 1,25(OH) vitamin D₃ in response to a low DCAD pre-calving diet (Gaynor et al. 1989; Goff et al. 1991), increased responsiveness of target tissues to 1,25(OH) vitamin D₃ associated with increased Ca absorption from the intestinal tract (Allsop and Pauli 1985) and increased plasma ionised Ca concentrations (Oetzel et al. 1991; Phillippo et al. 1994). Critically, the overall effect of negative DCAD diets is to increase Ca turnover through increased gastrointestinal tract (GIT) absorption and increased sensitivity of target tissues to homeostatic signals such as PTH, rather than an improvement in overall Ca balance. The effects of a negative DCAD diet may well include an enhanced sensitivity to the actions of skeleton in integrated metabolism (Fig. 2), and allow increased milk production in treated cattle.

In order to further evaluate the positive milk production responses observed to negative DCAD diets in some studies (Block 1984; DeGaris et al. 2008; DeGroot et al. 2010), the studies included in Lean et al. (2006) were examined to identify studies that included the effects of pre-calving diet on milk production. Further literature searches were conducted using PubMed, Google Scholar and Thompson ISI to ensure that recent studies were identified using the following keywords DCAD or DCAB or CAB or cation anion difference or cation anion balance and milk production or yield. A total of 15 studies and 34 comparisons were found and analysed (Stata 13.0, Statacorp, TX, USA) using a random effects model (DerSimonian and Laird 1986), with the Hartung and Knapp (2001) method used to provide estimates of effect size and confidence intervals for cows and heifers. A weighted mean difference between treated and control, with the weighting reflecting the inverse of the variance of the studies was calculated according to the nostandard method (Stata 13.0, Statacorp College Station Texas, USA).

Figure 3 shows the effect size estimates for all studies for which milk production and measures of dispersion were available as a Forest plot. Information on the parity of the cattle in the studies and diets obtained from the papers were entered into CPMDairy (Version 3.08, Miner Institute, Cornell, PA, USA) and meta-regression analysis was conducted to identify whether parity of the cattle, difference in DCAD between treatment and the control diet and estimated energy density, crude protein, crude fat, neutral detergent fibre (NDF), non-fibre carbohydrate, DCAD, Ca, magnesium, phosphorus, or potassium content of the control diet influenced responses. Only the parity (P = 0.003) and NDF of the control diet (P = 0.02) significantly influenced milk production responses, which increased in multiparous cows (P = 0.01), but not in heifers (P = 0.11). The effect size response was lower with higher NDF diets when tested on all studies.

Fig. 3.  A Forest plot of the effect size or standardised mean difference (standardised using the z-statistic)and 95% confidence interval of pre-calving dietary cation anion difference on milk production (milk, fat-corrected milk or 305 days milk) for the multiparous cow studies only. Estimates were made of the standardised mean difference using a random effects method (DerSimonian and Laird 1986). The weights that each study contributed are in the right-hand column and are indicated by the size of the box. The larger the box, the greater the study contribution to the overall estimate. The solid vertical grey line represents a mean difference of zero or no effect. Points to the left of the line represent a reduction in milk yield, while points to the right of the line indicate an increase. The upper and lower limit of the line connected to the square represents the upper and lower 95% confidence interval for the effect size. The overall pooled effects size and 95% confidence interval is indicated by the diamond at the bottom. This effect was heterogenous as indicated by the I² of 77.1%.

The average ± s.e. days of exposure to a transition diet for cows was 29 ± 3 and the DCAD of controls and treated cows 232 ± 37 and –27 ± 33 meq/kg, respectively. The effect size estimate for milk production of studies in cows estimated using the Hartung and Knapp (2001) method was positive 0.671 (95% confidence interval 0.178–1.164). The I² was 77.1 indicating considerable variability in responses. In studies that estimated responses in milk or fat-corrected milk over the first 65 ± 14 days of lactation, the weighted mean difference was 1.153 (95% confidence interval 0.335 to 1.971) L per day. In heifers milk production responses were negative with an effect size –1.225 (95% confidence interval –2.807 to 0.357) and I² 87.2 indicating considerable variability in responses. The weighted mean difference was –1.482 (95% confidence interval –1.872 to –1.093) L per day. The responses in heifers were strongly influenced by two trials in which DCAD changes were confounded with Ca concentrations in the diet. The lower response to DCAD in high NDF studies may indicate a role for ruminal outputs to influence acid–base status. Critically, other sources of variation, as indicated by the high I², were not identified, despite the large number of covariates tested.

These studies provide strong evidence that the physiological effects of negative DCAD diets, potentially mediated, in part, through mechanisms described above and by others in mice and man (Lee et al. 2007; Wolf 2008), are to increase milk production independent of most other dietary factors characterised using a contemporary feeding evaluation system, CPMDairy. These findings derived from randomised controlled studies are supported by epidemiological studies indicating that SCH before and after calving was associated with lower milk production (Chapinal et al. 2012). It is very likely that responses to DCAD diets differ between heifers and multiparous cows, suggesting that prior lactation or age substantially influence responses to DCAD diets day and indicates a need to understand differences in Ca metabolism between heifers and multiparous cows, that might include sensitivity to PTH, differences in labile Ca reserves and the absorptive surface area of bone.

We explored the interactions between metabolites over time, in order to further understand the potential for integration of metabolism of peri-parturient cattle exposed to pre-calving diets to extend beyond homeostatic responses. In brief, data from blood samples collected every 3.5 ± 0.5 days from ~25 days before calving to 30 days after calving were extracted for a range of metabolites from 32 individual cattle (DeGaris et al. 2010b); the data series for each metabolite were de-trended to produce approximately stationary series and cross-correlated with other metabolites. Cross-correlations for each cow were treated as separate studies and a random effects, pooled effect of estimate was produced using meta-analytic methods described by Hedges and Vevea (1998) after Fisher’s transformation of each cross-correlation coefficient. Only the results relating to Ca are provided here (Table 2). The results show the random effects average response as effect size, 95% confidence interval, and measures of heterogeneity or variance among cows as assessed by I² and τ². Calcium concentrations 3.5 days before were positively correlated with NEFA and negatively correlated on the same day, but there was significant heterogeneity in these results. There were homogeneous correlations between NEFA concentrations 17.5 and 21 days before and Ca. For 3 hydroxybutyrate and glucose, blood concentrations were negatively, and positively associated respectively, with Ca concentrations 7 days before and these results were homogenous. Calcium concentrations 7 days before were positively associated with cholesterol and Ca concentrations 3.5 days before were more strongly negatively associated with cholesterol concentrations. Both findings were homogenous, indicating that results were consistent among the cattle. Calcium concentrations on the same day were positively associated with cholesterol, but results were significantly heterogenous indicating more variability in these results. The pattern of positive and negative associations over time, particularly between Ca and NEFA, indicates negative feedback mechanisms and this feedback pattern was very evident in individual cows (Fig. 4).

Table 2.  Time series cross-correlations between calcium and other metabolites measured a lag of 3.5 days
Each cow is treated as an individual study and results are pooled using meta-analytical methods. One lag = 3.5 ± 0.5 days. I² is the variation in effect size due to heterogeneity. T² is the estimate of between-cow effect size variance. BHB, 3-hydroxybutyrate; NEFA, non-esterified fatty acids

Fig. 4.  Cross-correlations between calcium and non-esterified fatty acids for an individual cow showing evidence of feedback mechanisms.

These observations, indicating strong associations between serum Ca and NEFA, BHB, glucose and cholesterol in 32 peri-parturient cows, can be considered in the context of relationships between bone function and energy metabolism. They support a positive homeorhetic and homeostatic effect of Ca on energy metabolism with higher NEFA, glucose, and lower BHB subsequent to increased Ca concentrations. In most cases, Ca concentrations up to 7 days before and on the same day were associated with these metabolites, indicating that changes in Ca concentrations preceded those in NEFA, glucose and BHB. The relationships with NEFA and Ca 17–21 days later raise the possibility of lipid mobilisation and body lipid reserves influencing Ca metabolism and the negative correlation between Ca and NEFA on the same day is consistent with the findings of Martinez et al. (2014). The relationships identified using the time series methods are consistent with the relationships identified in mouse models and man (Lee et al. 2007; Wolf 2008; Confavreux 2011), given that acidifying conditions increase Ca concentrations and uOC production. The strong association between Ca and cholesterol may also be consistent with an overall positive effect of Ca metabolism on energy balance, given that energy balance and cholesterol concentrations are positively associated (Lean et al. 1992). Blood cholesterol concentrations have positive associations with reproductive performance in dairy cattle in studies (Westwood et al. 2002).

A potential role for the skeleton in reproduction has also been postulated based on increased cases of osteoporosis in hypogonadal human patients (Confavreux 2011). Recently, uOC has been shown to enhance male fertility via increased production of testosterone in Leydig cells (Oury et al. 2011) but not the ovary. However, a porcine study found OC levels increased with increasing oestradiol and markers of bone resorption showed a cyclical fluctuation during the oestrus cycle with highest concentrations found around oestrus and the lowest concentrations during the luteal phase (Babel 2007). Adipose tissue has also recently been identified as a site of both OC and uOC production under the influence of androgens (Foresta et al. 2010).

While studies investigating the relationship between uOC and energy metabolism are lacking in dairy cattle, there is a good body of evidence that OC is a reliable marker of bone formation (Liesegang et al. 1998). In the cow, OC is elevated in the plasma of younger animals, especially fetuses, declines with age, before the first lactation, is lower in multiparous than primiparous cows, declines in late pregnancy and at the time of calving and rises in the days to weeks thereafter (Sato et al. 2011). Taylor et al. (2008) similarly found lower OC concentrations in cows >2 lactations than second-lactation cows. Kim et al. (2011) administered an exogenous injection of 1,25-hydroxy-vitamin D₃ to stimulate bone formation in non-pregnant non-lactating cows, and found that the injection elevated plasma Ca, ‘total OC and uOC within days’. However, an earlier study (Taylor et al. 2008) found no increase in total OC concentrations following treatment with 25-hydroxy-vitamin D₃ or vitamin D₃.

The role of vitamin D in metabolism requires consideration in the context of the role of skeleton in metabolism. While the role of vitamin D in Ca homeostasis has been well established in monogastric species, including the human (Fraser 1995), there are questions remaining about ruminant metabolism. When there is an increased demand for Ca as in growth or lactation, the ionised concentration of Ca (Ca²⁺) in blood tends to fall. This initiates an increased secretion of PTH, which acts in both the kidney and bone to enhance the supply of Ca²⁺ to the extracellular fluid. In the kidney, PTH induces an increased secretion of the vitamin D hormone, 1,25(OH)₂D, which then induces increased active Ca transport in the small intestine. The molecular basis for this active transport involves an entry ion channel across the microvillous membrane (the transient receptor potential vanilloid channel type 6 or TRPV6), which facilitates Ca entry from the intestinal lumen into the mucosal cell (Peng et al. 1999). The internalised Ca²⁺ is bound to a cytoplasmic specific Ca-binding protein, calbindin-D_9k (Bronner 1987). A basolateral membrane Ca pump then transfers the absorbed Ca²⁺ into the submucosal extracellular fluid.

In ruminants, however, there is also considerable evidence that Ca²⁺ is absorbed across the rumen mucosa (Höller et al. 1988; Beardsworth et al. 1989; Wadhwa and Care 2000) and that this is mediated by a regulated active transport process. The rumen is the main site for magnesium absorption in ruminants (Tomas and Potter 1976; Care et al. 1984) and it is possible that the rumen is a major site also for the absorption of Ca. In sheep, when the blood concentration of 1,25(OH)₂D is raised, there is not only an increase in the absorption capacity for Ca²⁺ in the small intestine but also an increase in the transport capacity for Ca²⁺ across the ruminal mucosa in vivo (Hyde and Fraser in press). Therefore, in comparison to monogastrics, ruminants have two distinct regions of the alimentary tract where there are regulated, active absorption processes, which can increase the supply of Ca²⁺ when there is an increased demand, as at the onset of lactation.

The transport process for Ca²⁺ from the rumen in sheep is enhanced when the level of 1,25(OH)₂D in the circulation is raised, thus suggesting that the rumen mucosa has the same Ca transport mechanism, regulated by the vitamin D hormone as in the small intestine. However, no effect of vitamin D was detected in vitro in sheep on the active transport of Ca across the ruminal mucosa of sheep (Schröder et al. 2001). Furthermore, Wilkens et al. (2009, 2011) found that the components of the Ca transport mechanism that were readily demonstrated in the small intestinal mucosa, were either undetectable or were expressed in very small amounts in the rumen mucosa of sheep. Hence, there is a paradox in understanding the Ca homeostasis processes of the alimentary tract in ruminants. Studies in vivo have found that increased circulating levels of 1,25(OH)₂D are associated with increased absorption capacities for Ca in both the rumen and small intestine (Hyde and Fraser in press), yet the rumen mucosa does not have the molecular machinery, present in the small intestinal mucosal cells, for the active transport of Ca²⁺. An additional puzzle in this paradox is that the receptor protein for 1,25(OH)₂D that mediates the specific action of this hormone on the expression of the components of Ca transport, is at a very low concentration in the rumen mucosa cells compared with those of the small intestine. Therefore, although it appears that increased secretion of 1,25(OH)₂D by the kidney would lead to increased absorption of Ca in both the small intestine and the rumen, this is unlikely to be by direct action of 1,25(OH)₂D on the rumen mucosa cells. This suggests that some other signal, activated by 1,25(OH)₂D, is upregulating the absorption capacity for Ca in the rumen. The identification of this postulated secondary regulating factor might help to explain part of the mechanism for the development of clinical parturient hypocalcaemia in some dairy cows at the start of lactation.

Bovine somatotropin (rBST) has been used to increase milk production in lactation and studies with rBST administered in the pre-calving period have resulted in increased milk production in lactation. Putnam et al. (1999) treated cows with rBST every 14 days from 28 days before anticipated calving until parturition in a factorial study with protein treatments. Milk production responses to the rBST intervention were substantial, 3.3 kg/day more milk to Day 42 of lactation. Gulay et al. (2004a, 2004b) similarly, found increases in milk production to Day 70 of lactation in treated cattle, but these did not persist to Day 100, suggesting the influence of rBST alone administered before calving can be moderately effective in triggering longer-term milk responses and raising the possibility of the somatotropic axis being involved in responses to pre-calving diets. Studies to identify cattle treated with rBST have found that OC concentrations are elevated following treatment (Ludwig et al. 2012), as are OC concentrations and bone turnover indicators in humans treated with human growth hormone (Wallace et al. 2000), thereby providing a link between the somatotropic axis and bone. Some caution should be applied to these findings as the links identified may not indicate a potential to influence milk production through the skeletal axis. Studies in rats (Ammann et al. 2010) highlight an interaction between protein intake and bone responses to bovine growth hormone, because rats with low protein intake increased osteoclast surface area, but had lower bone mineral density and strength. The loss of bone mass in humans associated with low protein diets (Ammann et al. 2010) may indicate the potential for protein loss in cattle to be a factor in the susceptibility of older cattle to hypocalcemia.

Wolfenson et al. (1988) found that increased calf birthweight was associated with higher milk production, suggesting that effects of female progeny of increasing milk production in the next lactation identified by Hinde et al. (2014) may not be mediated through calf weight alone. This result may be consistent with the lack of significant effect in a meta-analysis of increased pre-calving crude protein intake on milk production (Lean et al. 2013). The latter finding is unsurprising because estimated metabolisable protein balance is a more meaningful measure of protein nutrition and stores of labile protein will vary among cattle (Belyea et al. 1978) and could influence responses to diet. Greater DM intake has increased milk production significantly after calving as demonstrated by forced feeding periparturient cows through a ruminal fistula (Bertics et al. 1992). Cows that received more feed had less hepatic lipid accumulation. The higher milk production after calving, resulted from greater post-calving feed intake and a highly significant positive correlation between pre- and post-calving feed intake was identified (Bertics et al. 1992). However, more work is needed to examine the effects of mineral, carbohydrate, lipid and protein nutrition in the pre-calving period, sire of calf and calf birthweight on milk production in a subsequent lactation to understand, isolate and potentially manipulate the magnitude of effects arising from the metabolic demands of the fetus for energy, amino acids and other nutrients (Bell 1995) including those required for skeletal development as opposed to those of hormonal or other influences of the fetus on future performance.

As consequence of rapid fetal growth during the last trimester of pregnancy and production of large amounts of colostrum and milk at the beginning of lactation, both maternal and fetal metabolism is increased due to augmented mitochondrial activity in maternal tissues and the conceptus (Aurousseau et al. 2006). This activity results in an increase in the production of reactive oxygen species during late gestation (Castillo et al. 2005) and early lactation (Pedernera et al. 2010) and increased requirements for micronutrients, including antioxidants. A pro-inflammatory state may be an essential part of the homeorhetic adaptations to lactation (Farney et al. 2013a). The potential to control the effects of oxidative damage through manipulation of micronutrients, other antioxidants or mediators of inflammation should be considered in the context of the exposure of dairy cows to oxidative stress and inflammatory events (Miller et al. 1993; Sordillo and Raphael 2013). Moreover the cow, and the calf, may compete for nutrients (energy, protein, minerals and antioxidants) during the transition period (Curtis 1997). A lowered antioxidant status has been associated with disease and metabolic disorders including mastitis, metritis, retained placenta, acidosis, ketosis and milk fever (Miller et al. 1993; Celi 2011). The most compelling support for the importance of mediation of inflammation in the peri-parturient period comes from the intervention studies with anti-inflammatory therapeutics of Bertoni et al. (2004), Shwartz et al. (2009) and Farney et al. (2013b) who either injected acetyl-salicylic acid for 5 days after calving (Bertoni et al. 2004) or fed sodium salicylate in water for 7 days (Farney et al. 2013b) or injected flunixin (Shwartz et al. 2009). In both the salicylate studies, milk production increased significantly, but did not increase in the flunixin study.

The mechanisms for the increased milk production were not elucidated; however, both antioxidant defence and reactions catalysed by steroidogenic enzymes require reducing equivalents provided by nicotinamide adenine dinucleotide (NADPH) (Liebler 1993). Excessive consumption of reducing equivalents by oxidative stress can lower NADPH₂ and increase NADP concentrations despite elevated activity of the monophosphate shunt, which generates the reduced form (Kehrer and Lund 1994). Consumption of reducing equivalents by oxidative stress reactions can diminish the supply of NADPH available for important physiological processes. The induction of the monophosphate shunt by an impaired oxidant/antioxidant balance would divert glucose from other pathways. This possibility assumes greater importance when the requirement for glucose and the quantity available in the ruminant are considered and may be a factor influencing the production responses to the interventions (Bertoni et al. 2004; Farney et al. 2013a, 2013b). The positive milk production responses in these studies (Bertoni et al. 2004; Farney et al. 2013b) appeared to be independent of effects of treatment on disease, although this hypothesis was not specifically tested in either study.

While adequate antioxidant concentrations, including selenium and vitamin E, are important for the maintenance of health and performance of cows and calves (Lacetera et al. 1996), responses to antioxidant supplementation have resulted in conflicting reports. For example, vitamin E and selenium supplementation at doses greater than the National Research Council (2001) recommendations did not result in improved daily gain or morbidity in feedlot cattle (Cusack et al. 2009), nor did use of an organic selenium improve reproductive performance (Cerri et al. 2009). Such variable results are to be anticipated from the extensive interactions between antioxidants (Curtis 1997) and the pro-oxidative stressors of the metabolism of cattle per se and exerted through the environment. There is considerable capacity for interactions with other dietary substrates and environmental conditions including heat stress (Dunshea et al. 2013). It is possible that more sophisticated understandings of the optimal means of controlling oxidative stress and inflammation around calving will improve reproduction and health.

There are several major pathways through which improved production, reproduction and health may be achieved by manipulation of the transition environment and probably more to be identified. Lee et al. (2007) demonstrated that OC, produced by mature OB, can regulate fat metabolism in mice, thus completing the homeostatic loop between the neuro-endocrine regulation of energy and bone metabolism. They also demonstrated homeorhetic responses in glucose metabolism mediated by OC and characterised by pancreatic β-cell proliferation extending over 45–60 days.

Negative DCAD diets stimulated milk production in cows, but not in heifers and these effects were not modified by dietary factors evaluated, other than the NDF of the diet. Negative DCAD diets stimulate bone turnover, acidify the blood, thereby potentially enhancing the vitamin K-mediated decarboxylation of OC to uOC and increasing hormone release from the skeleton. This could enhance the effect of uOC on energy and protein metabolism and possibly explain some of the very long-term effects of negative DCAD diets on metabolism.

We found that nutritional interventions, including a negative DCAD diet, applied only in the pre-calving period markedly increase milk production, improved reproduction and reduce the risks of metabolic disease. These interventions associated with negative DCAD diets fed before calving also influenced concentrations of glucose and BHB. When examined by time series methods, correlations were found for blood Ca concentrations with blood concentrations of glucose, FFA, BHB and cholesterol taken on the same day and on up to 21 days later. This provides evidence to support a homeorhetic role for Ca metabolism in cattle. Treatment with rBST had a moderate, but positive effect on production and has links to bone and protein metabolism. Studies have found prolonged differences in production with weight and sex of the conceptus indicating the importance of considering the role of conceptus in nutrient demand and hormonal signalling. Inflammation plays an important role in lactation and health. Two studies showed benefits of anti-inflammatory treatments administered for a few days near calving to increase milk production.

While we have provided information to support such a hypothesis, it remains to be specifically determined if the homeostatic and homeorhetic pathways demonstrated in murine models exist in cattle. The potential for manipulation of both energy metabolism and bone metabolism in a coordinated fashion increases production, and may have the potential to modify the risk of disease in early lactation. If these effects are demonstrated to be clinically significant, it may require a reconsideration of current recommendations for dietary Ca, phosphorus and magnesium balance, particularly during the transition period and early lactation. Such recommendations may be modified according to body tissue reserves of the cattle. We have provided background on Ca and vitamin D metabolism highlighting new information on the absorption of Ca that may lead to better dietary recommendations for intake. The quantitative importance of bone resorption to health, production and reproductive performance needs to be determined, however, there is sufficient evidence to merit such studies.