Social rank affects the endocrine response to frequent regroupings in grazing dairy heifers

Graphical Abstract Summary Social regroupings (SR) are known social stressors that affect animal performance, while social rank may determine differences in response to repeated SR. We wanted to know if repeated SR influenced social behavior and if social rank and possible changes of social rank during SR determined differences in growth rate and in some metabolites and hormones associated with energy status in dairy heifers under grazing conditions. Fourteen “resident” heifers (remained in the same paddock; shown as black in the graph) were regrouped with 5 new heifers (“exchanged”; shown as red in the graph) every 21 days. The total number of interactions between resident heifers decreased and stabilized after the first 2 SR. Insulin-like growth factor 1 (IGF-1) concentrations in serum were greater in medium-ranked than in low- and high-ranked heifers in SR1, SR4, and SR9, and heifers that lowered social rank presented greater serum IGF-1 concentrations compared with heifers that maintained social rank across SR. Social rank influences the endocrine response to repeated SR in heifers.

S ocial relationships are involved in determining an individual's position within a group's social hierarchy (Hubbard et al., 2021a).Social position may determine changes on the productive performance (Phillips and Rhind, 2002), the metabolic status, and the response to stressors (Chebel et al., 2016;Fiol et al., 2017Fiol et al., , 2019) ) between animals within a group.Individual social positions in cattle are relatively stable over time (Kondo and Hurnik, 1990), but when animals are submitted to social regroupings (SR), the social hierarchy is challenged (Hasegawa et al., 1997;Raussi et al., 2005).The individual position in the hierarchy is established through agonistic interactions, and after stabilization, is mainly maintained through non-physical agonistic interactions (Kondo and Hurnik, 1990).Social regroupings trigger an increase in the frequency of agonistic behaviors to re-establish a new hierarchical order (Hubbard et al., 2021b), and a reduction of affiliative interactions.Repeated SR affect productive performance (Hasegawa et al., 1997;Schirmann et al., 2011), the metabolic status (Silva et al., 2013) and the behavioral pattern (von Keyserlingk et al., 2008;Moratorio et al., 2023).Although some studies reported habituation to repeated SR, reducing the number of agonistic interactions along the successive regroupings (Veissier et al., 2001;Gupta et al., 2008), other authors reported that the number of aggressive interactions increase after each SR (Raussi et al., 2005;Talebi et al., 2014).
Animals may change their position in the hierarchy after each SR, and those changes may modify their coping ability and the possible consequences of SR (Hasegawa et al., 1997;Solano et al., 2004).In regrouped lactating heifers, those who lowered their individual position show a more drastic reduction in milk production than those that increase their social rank (Hasegawa et at., 1997).In contrast, milk yield persistency was negatively related to dominance rank after SR in dairy cows (Arave and Albright, 1976), and high-ranked cows were more stressed than low-ranked ones after repeated handling (Solano et al., 2004).
Social environment may determine changes in growth rates in growing heifers (Bach et al., 2006;Fiol et al., 2017).Reduced growth rates are associated with a worse energy status, characterized by decrease of insulin-like growth factor type 1 (IGF-1) and glucose (Fiol et al., 2017;Rosadiuk et al., 2021), and an increase of nonesterified fatty acids (NEFA) concentrations (Abeni et al., 2019).In addition, greater serum NEFA concentrations are observed associated with increased serum glucose concentrations Social rank affects the endocrine response to frequent regroupings in grazing dairy heifers C Fiol,1* M Moratorio, 1 M Carriquiry, 2 and R Ungerfeld 3 in chronically stressed animals, as those frequently regrouped (Chebel et al., 2016).The individual position in the social hierarchy may influence the energy status and/or the stress response to regrouping, and thus, may determine differences in growth rates and metabolic and endocrine profile in heifers.
We hypothesized that: 1) agonistic interactions in frequently regrouped heifers increase after each SR and along the repeated SR, and 2) the growth rate and the metabolic and endocrine profiles are more affected (worse energy status -lower growth rates, glucose and IGF-1, and greater NEFA) by repeated SR in low than in high-ranked and in those heifers that lowered their social rank compared with those that maintain or raise their social rank after SR.We aimed to determine the social behavior response to SR, and if social rank and the changes of social rank affect growth rates and metabolic and endocrine profile, in "resident" replacement dairy heifers subjected to repeated SR.
Animal care, handling, and protocols were approved by the Ethics Commission on the Use of Animals (number 465 CEUA, Universidad de la República, Uruguay).The study was performed on a private contract-rearing farm (Uruguay, 34° S, 55° W), from September (early spring) to May (autumn).The general facilities and procedures were described in a previous study, as both were performed with the same animals (Moratorio et al., 2023), but this study focused on the response of the regrouped heifers.Fourteen Holstein heifers (153.3 ± 16.1 kg; 7 to 10 mo-old) remained in the same grazing paddock with another 5 Holstein heifers that were exchanged every 21 d (total = 210 d), subjecting the basic group ("resident" heifers) to frequent SRs.On the day of each SR, between 08:00 and 10:00 h (total = 10 SR), 5 unknown heifers (with no previous contact with the resident ones) were introduced at the time that the previous 5 were retired from the group.Those 5 new heifers were of similar BW (182.3 ± 5.2 kg) and age (range = 1 mo) than the 14 resident heifers, and were selected from the general herd, totaling 64 heifers participating in the study.The heifers grazed a mixed pasture of fescue and red clover on weekly stripes (forage allowance = 8 kg DM/100 kg BW) throughout all the study and had free access to cutwaters.Pasture availability was monitored monthly (range: 810 to 2930 kg DM/ha; Haydock and Shaw, 1975).The size of the paddock ranged from 0.8 to 2 ha (minimum of 420 m 2 per heifer), depending on pasture availability.Two days after each SR (07:00 to 10:00 h), heifers BW and withers height (WH) were measured and individual average daily gains (ADG) were calculated to determine body development (Heinrichs et al., 1992(Heinrichs et al., , 2017).On SR6 it was not possible to determine BW nor ADG, thus, data for those variables are only available in 9 of the 10SR.Blood samples were collected simultaneously with body development measurements, and samples were centrifuged and stored at −20°C until measurement of NEFA, glucose and IGF-1 concentrations (Moratorio et al., 2023).Samples for IGF-1 concentrations were only analyzed on 6 of the 10 social regrouping (SR 1,3,4,5,7 and 9).
Social behavior was registered by continuous sampling 2 d before each SR, the day of SR, and 7 d later, during the last 3 h of daylight (17.00 to 20.00).During the study, a total of 4 trained observers recorded, by direct visual observation, and categorized the different social interactions between the resident heifers: head butting (pushing another with its head on any part of the body), chasing (chasing another with no physical contact), displacing (displacing another by physical contact, withdrawal of the reactor), and threat-ening (displacing another with no physical contact, the reactor may or may not withdrawal), while affiliative interactions were mounting (mounting or trying to mount another), licking (licking another, in any part of the body), and scratching (scratching any part of the body toward another animal).To assess the interobserver agreement, the 4 observers conducted a simultaneous 2-d preliminary observation assessment on the same animals, during which they recorded the different behaviors for 2 one-hour periods each day (overall agreement = 0.9 for each behavior; data not shown).A big number was painted on the back of each heifer to allow identification from long distances.Both, the animal that initiated each interaction and the one that received it, were recorded.To correct possible biases, the observers were blocked and assigned to each day.As the aim was to evaluate the effects of regrouping on the resident heifers and not on the "exchanging" ones, all responses were only measured in the 14 resident heifers.
Social status was determined by the ETlog software (Deniz, 2018), performed in R (R Core Team, 2020).The ETlog considered all agonistic interactions registered 2 d before and 7 d after each SR than an animal was involved (i.e., as actor or reactor) in relation to other herd member, of each possible dyad of the group of resident heifers.Days were chosen to minimize the acute effects of SR on social hierarchy (Kondo and Hurnik, 1990).In each SR, the dyadic dominance relationship of the ith animal relative to the j th animal (Sij) is assessed qualitatively by the sign of the difference between Xij and Xji, as proposed by Kondo and Hurnik (1990): where S is the relation between animal i and animal j; Xij is the number of winning interactions of i animal over j animal and Xji is the number of winning interactions of j animal over i animal, which always results in a value of −1 (animal i had less winning interactions than animal j), 0 (animal i had the same number of winning interactions as animal j), or +1 (animal i had more number of winning interactions than animal j) (Kondo and Hurnik, 1990).
With the obtained results of the previous equation, the dominance value for each individual was calculated according to the following equation: Si is the sum of all the interactions in which i animal was involved; n is the number of possible interactions of one animal of the group with the others.Dominance position was then assigned according to Si values, from highest (α) to lowest (ω).When 2 animals had the same Si value, the tiebreaker was the result of the dominance relationship among the dyad.With the sociometric matrix, the ETlog software calculated the linearity index proposed by Landau (1951) and applied the improved test of linearity (h') due to the unknown relationships as described by De Vries (1995).Finally, a dominance scale was constructed in each SR, based on the difference between the maximum and minimum dominance values plus 1 (corresponds to the dominance value zero), and heifers were categorized in 3 social ranks: heifers in the first tertile were clas- sified as low-ranked heifers (LRA), the ones in the second tertile as medium-ranked heifers (MRA), and heifers in the third tertile (higher positive dominance values) of the dominance scale were classified as High-ranked heifers (HRA) (Table 1).
All analyses were performed with SAS on Demand for Academics.The normal distribution of all data and residues was checked with the UNIVARIATE procedure.The mean number of social interactions (not normally distributed) was analyzed with a generalized linear mixed-effects model using the GLIMMIX procedure (Poisson distribution): where: Yij was the dependent variable (agonistic, affiliative, and total interactions), μ the general mean, Ti was the fixed effect of the SR (i = SR 1 to SR 10), Hj the fixed effect of the day of social behavior determination (j = d 0 vs d 7), (TxH)ij the interaction between T and H, Bi the random effect of the animal and Cj the random effect the farm of origin, and eijk the residual error.
The effect of the individual position in the hierarchy on growth rate and metabolic and endocrine profiles was analyzed as a repeated measure according to social rank in each SR (LRA, MRA, and HRA heifers).In addition, to determine the effects of social rank changes across the SRs, the social rank in the first and the last 5 SRs were compared, resulting in 3 categories: heifers that remained in the same social rank (Mai), heifers that lowered their social rank (Low) and heifers that raised their social rank (Rai) (Table 1).For this, as the social rank was stable along the first 5 and along the last 5 SRs, the social rank of the first 5 and last 5 SR was considered as that in which the heifer was categorized more times (4 of 5 or 3 of 5 SR).For example, heifer n°1 in the first 5 SR was classified as: LRA, LRA, LRA, MRA, LRA; thus, the social rank of the heifer on the first 5 SR was LRA.For the last 5 SR she was classified as LRA, MRA, LRA, MRA, LRA; thus, the social rank for the last 5 SR was LRA.Then, the heifer n°1 was defined as Mai.
The effect of social rank in each SR and social rank changes across the first 5 and last 5 SR on body growth parameters and metabolites' concentrations were analyzed by a linear mixed-effects model, as follows: Yijk = μ + Ti + Aj + Hk + (TxH)ik + (AxH)jk + Bi + Cj + eijkl, where: Yijk was the dependent variable (BW, ADG, WH, serum NEFA, glucose and IGF-1 concentrations), μ the general mean, Ti was the fixed effect of the social rank (i = HRA, MRA and LRA), Aj the fixed effect of social rank change (j = Low, Mai, and Rai), Hk the fixed effect of the SR (repeated measure; k = 1 to 10), (TxH) ik the interaction between social rank and SR, (TxA)jk the interaction between social rank changes and SR, Bi the random effect of the animal, Cj the random effect of the farm of origin, and eijkl the residual error.The covariance structure AR (1) (for constant periods) or SP (pow) (for nonconstant periods) were used.Body weight, ADG, WH, serum NEFA and glucose were analyzed by the MIXED procedure, while IGF-1 concentrations (Poisson distribution) were analyzed using the GLIMMIX procedure.Baseline BW and concentrations of NEFAs, glucose and IGF-1 were included as covariates, in the corresponding analyses.In both GLIMMIX and MIXED procedures, the Kenward-Rogers method was used to adjust denominator of degrees of freedom, and Tukey Kramer tests were conducted to analyze differences between groups.Differences were considered significant when P ≤ 0.05 and as tendencies when 0.05 < P ≤ 0.1.Results are presented as least squares means ± standard error of the mean.
Overall, 763 (509 agonistic and 254 affiliative) social interactions were registered.Heifers performed more agonistic (2.6 vs 1.8 ± 0.6, day of SR and 7 d later, respectively) and total (3.8 ± 1.2 vs 2.5 ± 0.7, day of SR and 7 d later, respectively) interactions on the day of SR than 7 d later (P = 0.01 for both comparisons), with no differences in affiliative interactions (2.1 ± 0.6).In addition, social interactions changed along the different SRs: in general, the number of interactions decreased throughout the study in relation to the first 2 SRs (Figure 1).
Concerning social rank changes, 7 animals maintained their rank, 5 lowered and 2 raised their social rank.Considering the low  Deniz, 2018), and heifers were classified as Low (LRA), Medium (MRA) and High (HRA) rank heifers.In addition, according to the change in social rank of heifers comparing the social rank in the first 5 and in the last 5 SR, resulted in three categories: heifers that maintained (Mai), lowered (Low) or raised (Rai) social rank in the first 5 SR compared with the last 5 SR.At the end of the study, 7 animals maintained their rank, 5 lowered and 2 raised their social rank.
number of animals that raised their social rank, growth rates, metabolites and IGF-1 analysis were performed comparing only Mai and Low heifers.Mean ADGs were similar between Low and Mai heifers throughout the study (0.800 ± 0.04 kg/d), but there was an interaction between the changes in social rank and SR for ADG, as in SR4 Low heifers gained more BW than Mai heifers (0.99 ± 0.13 vs 0.61 ± 0.11 kg/d, Low and Mai heifers, respectively; P = 0.03).
In addition, IGF-1 concentrations were greater in Low than Mai heifers along the SR (132.5 ± 17.1 vs 97.8 ± 11.2 ng/mL, Low and Mai heifers, respectively; P = 0.04).There was also an interaction between the change in social rank and SR for IGF-1 concentrations (P = 0.08): Low heifers had greater IGF-1 concentrations at SR1, SR3, and SR9, and tended to be greater at SR 4, than Mai heifers (Figure 2B).There were no effects of change of social rank nor interaction social rank by SR in any other variable analyzed (data not shown).
As hypothesized, social regrouping triggered an increase in the display of agonistic interactions, but the frequency of agonistic interactions decreased after the first 2 SRs and remained stable along the rest of the study.This suggest that at least, even when exposed to a chronic repeated stressor, heifers could habituate and cope with social instability.In addition, not all heifers could adapt similarly, as the evolution of IGF-1 concentrations were related to hierarchical positions and to the changes in individual social positions that heifers experienced along the repeated SR.Age is a factor related with social hierarchy (Hubbard et al., 2021a), so it would have been important to have the exact age of each animal, but as they came from different commercial farms, it was not possible to obtain all the exact data, limiting the consideration of these data in the study.
As previously reported (Raussi et al., 2005;von Keyserlingk et al., 2008), regrouping disturbs social behavior, as heifers display more agonistic interactions immediately after each SR than  one week later.However, as the number of SR progressed, the frequency of agonistic interactions remained stable.In contrast, the number of agonistic interactions of the introduced animals increased with the successive regroupings, which was related to the lack of habituation of the animals to frequent SR (Raussi et al., 2005).Introduced individuals, as opposed to resident ones, must deal with SR simultaneously with the challenge of a new allocation and general facilities.According to our knowledge, there are scarce studies that evaluated the effects of SR on the animals that remain as residents (Bach et al., 2006;Rocha et al., 2020).Rocha et al. (2020) reported that social bonds in a group of resident animals get stronger over time and less aggressive interactions are necessary to reestablish hierarchy after regrouping with new animals.In that sense, resident heifers in this study could develop coping strategies that imply no increases in their aggressiveness, probably redirecting the aggressive requirements to the group of introduced heifers.Finally, it should also be considered that heifers remained grazing extensively, with a low stocking density (minimum of 420 m 2 per heifer), reducing the negative effects of SR as animals have enough space to avoid contact between them, and thus, reducing the number of possible aggressive encounters (Talebi et al., 2014).
We found differences on the endocrine profile according to social rank: IGF-1 concentrations were greater in medium than in high and low ranked heifers at SR1, 4 and 9.In contrast, we expected to find increase IGF-1 concentrations in heifers of higher social ranks, possibly associated to lower level of social stress in heifers on the upper stratum of the hierarchy compared with the ones in the medium and low stratums.In that sense, previous studies reported that low-ranked cows presented greater cortisol concentrations after SR (Mench et al., 1990), and more negative effects on milk production in subordinate dairy cows after SR (Hasegawa et al., 1997;Soonberg et al., 2021), compared with animals of high and medium social rank.It is necessary to consider that the low number of animals evaluated may limit detecting some other possible effects.In agreement with our findings, Miranda de la Lama et al. (2013) reported that medium-ranked males had greater ADG, lower cortisol concentrations, and better meat quality than high and low ranked bulls.Possible explanations to the contrasting results are not obvious; it may be argued that high-ranking heifers are less successful in coping with SR in comparison to medium ranking ones (Arave and Albright, 1976;Chebel et al., 2016), and/or that they are more sensitive to the negative consequences of social stress than medium ranked ones (Miranda de la Lama et al., 2013).
The effects of changes in social position along the SRs were unexpected, as heifers that lowered their social rank had greater serum IGF-1 concentrations along the study and at SR 1, 3, and 9 than those heifers that maintained their social rank.High growth rates were associated to an increase of serum IGF-1 (Rosadiuk et al., 2021), while changes in social rank status were expected to be linked to the social stress response to SR (Chebel et al., 2016).Thus, the greater IGF-1 concentrations in heifers that lowered their social position after SR may be related to reduced levels of social stress and/or changes in feeding utilization compared with those heifers that successfully maintained their social rank across SR.As previously mentioned, the low number of animals evaluated may limit detecting some possible effects.
In conclusion, agonistic interactions increased on the day of each SR, but this response was decreasing along the successive SR.Social rank and social rank changes were associated to the endo-crine profile in heifers subjected to repeated social regrouping, as serum IGF-1 concentrations were greater in medium-rank heifers than in high and low-ranked, and in those heifers that lowered their social rank compared with those that maintained their social rank.Thus, individual position in the social hierarchy may influence the endocrine response to social regroupings in the farm.

FiolFigure 1 .
Figure 1.Mean number of agonistic (A), affiliative (B), and total (agonistic + affiliative; C) interactions in dairy replacement heifers (n = 14) regrouped every 21 d with 5 new heifers for 205 d, according to the number of social regrouping (SR).Data were analyzed according to the number of social regrouping (SR 1 to SR 10) using the GLIMMIX procedure.Time (number of SR) was considered the main effect, and the animal and farm of origin were considered the random effects.Tukey Kramer tests were conducted to analyze differences between SR.Different letters in each column represent significant differences (P < 0.05) between SR.Error bars represents SEM.

Figure 2 .
Figure 2. Serum IGF-1 concentrations in dairy replacement heifers (n = 14) regrouped every 21 d with 5 new heifers according to the social rank (panel A) and to changes in social rank (panel B) in each social regrouping (SR).Social rank was determined according to agonistic interactions in each SR (ETlog software; Deniz, 2018), and heifers were categorized as: low (LRA), medium (MRA) and high (HRA) social rank.In addition, heifers were categorized as those that maintained (Mai) or lowered (Low) their social rank in the last 5 SR compared with the first 5 SR.Panel A: different letters represent significant differences (P < 0.05) between heifers in each SR.Panel B: different letters represent significant differences between Low and Mai heifers in each SR: a vs b = P < 0.05; c vs d = P ≥ 0.05 ≤ 0.1.Error bars represents SEM.

Table 1 .
Fiol et al. | Social rank affects… Social rank during the first five and last five social regroupings (SR) and social rank changes in heifers regrouped every 21 d with heifers for 205 d (total = 10 SR)