Peripheral blood mononuclear cell mitochondrial enzyme activity in calves is associated with average daily gain, reproductive outcomes, lactation performance, and survival

Mitochondria are central to metabolism and are the primary energy producers for all biosynthesis. The ob-jective of this study was to determine if the mitochondrial enzyme activity of peripheral blood mononuclear cells in heifers was associated with average daily gain, reproductive outcomes, first-lactation milk production, and survival. Twenty-three Holstein and 23 Jersey heifer calves were enrolled, and blood and body weight data were collected at 1, 2, 8, 36, 52, and 110 wk of age. Respiratory and fecal scores were recorded daily for the first 30 d of life. Milk production data were collected from herd management software through first lactation and health events were tracked to the fourth lactation on surviving animals. Mitochondrial isolation and enzyme activities for citrate synthase, complex I, complex IV, and complex V were determined using kits from Abcam. Data were analyzed using GLM and the Logistic procedure of SAS (version 9.4, SAS Institute Inc.). Multivariate regression analyses were conducted to determine if calf mitochondrial enzymatic activity and covariate health indices (fecal and respiratory scores, number of treatments, hematology) were associated with average daily gain (8, 36, 52, and 110 wk), lactation performance (milk yield, fat yield, solids yield, energy-corrected milk, 305-d mature equivalent, and relative value), and reproduction (age at first service, age at first conception, age at first calving, and number of services). For Holsteins and Jerseys, mitochondrial enzyme activities and health indices were correlated with all average daily gain and milk production outcomes (R 2 ≥ 0.63 and R 2 ≥ 0.45, respectively). Re-production outcomes were correlated with body weight gain, mitochondrial function, and red blood cell traits for Holsteins and Jerseys (R 2 ≥ 0.47 and R 2 ≥ 0.55, respectively). Logistic regression analyses were performed to determine if early-life enzymatic activity affected survival outcomes in the herd. Calves below the median for complex V enzyme activity at 1 wk were more likely to be removed from the herd compared with calves


INTRODUCTION
One opportunity for cutting expenses and maintaining profitability on dairy farms is to focus resources on heifers with high-performing mitochondria.Mitochondria are central to metabolism and health and offer a novel approach to assess cow performance.Mitochondrial traits have been shown to influence bovine BW gain and milk production (Brown et al., 1988;Niesen and Rossow, 2019;Niesen et al., 2022) and reproduction (Iwata et al., 2011;Ferreira et al., 2016;Kansaku et al., 2017).Additionally, early works by Bell et al. (1985) and Brown et al. (1988) suggested that cow cytoplasmic inheritance could indicate future milk production in progeny because mitochondria are maternally inherited.The use of peripheral blood mononuclear cells (PBMC) offers a high throughput method of assessing mitochondrial function in cattle, as the mitochondria can be obtained from blood samples (Niesen and Rossow, 2019;Niesen et al., 2022).Assays of PBMC mitochondrial enzymes of the respiratory chain complexes and citric acid cycle enzymes are minimally invasive and can identify mitochondrial impairment (Rustin et al., 1994;Hsiao and Hoppel, 2018).Dysfunction of the respiratory chain complexes can result from mutations in mitochondrial or nuclear DNA, aging, and may result in increased reactive oxygen species, cell death, and disease (DiMauro and Schon, 2003;Balaban et al., 2005;Morán et al., 2012).The mitochondrial enzymes of the respiratory chain complexes and citric acid cycle enzymes are central to the production of ATP and affect an animal's ability to produce the energy necessary to meet the demands of growth, health, and production.
If mitochondria could be screened for performance, heifer merit could be determined early in life and improve farm economic outcomes by meeting production goals with fewer heifers raised.Care and management for a replacement heifer can be as high as 20% of the total cost associated with dairy production (Fetrow, 1987;Lehenbauer and Oltjen, 1998;Gabler et al., 2000) and has been estimated to be between $1,700 and $2,400 per heifer (Overton and Dhuyvetter, 2020).Heifer culling and mortality are highest in the first 2 yr of life.Producers often battle high preweaning calf mortality, where 13% to 22% of heifers fail to reach first calving and up to 26% are culled after their first lactation (Hadley et al., 2006;Brickell and Wathes, 2011;Cooke et al., 2013).
The selection of dairy cows based on genetic milk yield traits has adversely affected their lifespan because of the increased metabolic demand (Essl, 1998;Ingvartsen et al., 2003;Oltenacu and Broom, 2010).When cows undergo negative energy balance, they are more susceptible to metabolic problems, exhibit poor physical condition, have decreased reproductive ability, and are present in the herd for a shorter period (Bauman and Currie, 1980;Rauw et al., 1998;Walsh et al., 2011).Since mitochondrial respiratory chain enzymes are central to energy production pathways, heifer selection based on mitochondrial enzyme function may select for animals that are less prone to metabolic problems.Mitochondrial function assays could be used as a screening tool to help farms make strategic breeding and culling decisions before costs associated with feed, treatments, and labor are incurred.Therefore, the objective of this study was to determine if PBMC mitochondrial enzyme activities of citrate synthase, complex I, complex IV, and complex V in Holstein and Jersey dairy cows change with time and are associated with ADG, reproductive outcomes, first-lactation milk production, and survival.

Study Design
This prospective observational study was approved by the University of California, Davis Animal Care and Use Committee, protocol #21157.
Twenty-three Holstein and 23 Jersey heifer calves from a California commercial dairy were enrolled be-tween December 2016 and February 2017 and data were collected from 1 to 110 wk on animals that survived to each time point.This study did not interfere with farm management practices or cow culling; reasons cows were removed from the herd are shown in Table 1.A minimum sample size of 8 cows per treatment as estimated based on a 2-tailed test with a difference of 30% between electron transport chain enzyme complex activities with a power of 0.90 and an α of 0.05 using data from past studies involving mitochondrial measurements (Lancaster et al., 2014;Acetoze et al., 2015).
Cows were sampled at 5 time points throughout the study.The first samples were collected at 1 wk, as this was the earliest window that PBMC could be obtained for mitochondrial enzyme analyses due to immature cell differentiation.The second time point, 2 wk, was selected as it was near the onset of immune challenge in the form of diarrhea.The third time point was at 8 wk, before weaning, and the fourth time point was at 52 wk, before the first breeding.Last, the fifth time point was 110 wk of age in early lactation (55 to 75 DIM).

Animal Management and Housing
Detailed preweaning calf management and housing methods were presented in Niesen and Rossow (2019).In short, calves were enrolled with inclusion criteria being a respiratory score of 1, general appearance score of 1, and fecal score of 3 or less following the CalfTrack scoring system (Heinrichs et al., 2003).Calves were housed in raised individual wooden hutches with cement flush lanes and ad libitum access to water.Weaning occurred at roughly 60 d at the discretion of the calf manager and depending on heifer size.Upon leaving the hutches, postweaning heifers were grouped in mixed-breed pens according to frame size in dry lots with shade covers and fed a TMR once daily at approximately 0700 h.Heifers nearing parturition were moved to a close-up pen and remained there from approximately −21 to 0 DIM where they were fed a TMR at approximately 0530 h.Upon leaving the close-up pen, heifers were moved into milking pens sorted by stage of lactation and fed a TMR at approximately 0600 h.Both the close-up and milking pens had freestalls with attached flush lanes and were mixed by breed.

Health Events, Treatments, Milk Production, and BW Measurements
Respiratory and fecal scoring were performed daily for the first 30 d of life in preweaning calves following methods defined by Niesen and Rossow (2019) to be used as model covariates.Preweaning treatments were collected from treatment records on the hutches.Postweaning events (treatments, breeding, conception, illness, sold, died) and first-lactation milk production data were collected from DairyComp305 (Valley Ag Software).Milk production data were collected through the first lactation and events were tracked to fourth lactation on surviving animals.Production data were recorded once monthly by Tulare DHIA and analyzed for milk yield, total fat yield, total solids yield, ECM, 305-d mature equivalent (305ME), and relative value.Preweaning calves were weighed at 1, 2, and 8 wk according to Niesen and Rossow (2019).Postweaning BW measures were measured at 36, 52, and 110 wk with a Coburn breed specific weigh tape (Coburn Company Inc.).

Blood Collection, Hematology, and PBMC Isolation
Blood samples were collected at 1, 2, 8, and 52 wk via jugular venipuncture and 110 wk via coccygeal tail vein.Two sets of whole blood (30 and 4 mL) were collected into Vacutainer tubes (BD Biosciences) containing K 2 EDTA as an anticoagulant at each time point and processed within 2 h of sample collection.Samples were taken as quickly as possible to ensure minimal stress to the animals.
Well-mixed blood (4 mL) from a K 2 EDTA tube was used to determine hematocrit (%), mean corpuscular hemoglobin (pg), mean corpuscular volume (fL), and neutrophil yield (K/μL) using the Drew Scientific Hemavet 950 Hematology Analyzer System (Erba Diagnostics).Before evaluating samples, quality control samples were run to ensure that equipment was functioning within specification (Multi-Trol, Drew Scientific).
Platelet-rich plasma (PRP) and buffy coat were separated from the remaining whole blood (30 mL) by centrifugation at 2,000 × g for 20 min at 20°C.Plasma total protein was determined from the PRP using a handheld clinical ATC refractometer (Index Instruments) at 1 to 8 wk and the remaining PRP was discarded.The buffy coat was diluted (1:4) with autoMACS Rinsing Solution (PBS, pH 7.2, and 2 mM EDTA, MiltenyiBiotec) and applied to a Histopaque density gradient (specific gravity 1.077, Sigma Chemical catalog number 10771) and centrifuged without application of the brake at 2,000 × g for 20 min at 20°C.The PBMC were collected and pelleted at 300 × g for 10 min at 20°C and washed with autoMACS Rinsing Solution 3 times.Before the second wash, red cell contaminants were lysed via osmotic shock using distilled water, vortexed, and immediately diluted with autoMACS Rinsing Solution.The washed PBMC were then pelleted at 300 × g for 10 min at 4°C and the supernatant discarded.All subsequent steps used kits from Abcam and followed the manufacturer's instructions.

Mitochondrial Isolation and Protein Quantification
Mitochondria were extracted from PBMC using the Mitochondria Isolation Kit for Cultured Cells (Abcam, ab110170).Protein concentration of PBMC lysate was measured by BCA assay (Abcam, ab102536) and pellets were frozen at −80°C for 10 min to weaken cellular membranes then supplemented with 0.2 μL of universal nuclease (Fisher Scientific Co., PI88700) to reduce viscosity.Samples were resuspended to 5 mg/ mL in reagent A followed by homogenization.The homogenate was centrifuged at 1,000 × g for 10 min at 4°C, saving the supernatant and resuspending the pellet in reagent B. Homogenization and spin steps were repeated and the supernatants were combined and further centrifuged at 12,000 × g for 15 min at 4°C.The resulting supernatant was discarded, and the crude mitochondrial pellet dissolved in reagent C supplemented with protease inhibitor (Abcam, ab201111), aliquoted, and stored at −80°C.The crude mitochondrial protein concentration of one aliquot per sample was measured by bicinchoninic acid assay and used to correct the final activities of each sample (Abcam, ab102536).

Measurement of Mitochondrial Complex I, Complex IV, Complex V, and Citrate Synthase Enzyme Activities
All mitochondrial enzyme activities were measured in duplicate using crude mitochondrial extracts.Microplates were incubated for 3 h before the collection of absorbance data using a VersaMax tunable microplate reader (Molecular Devices) in kinetic mode.Before evaluating samples, a calibration test plate (Bio-Tek Instruments Inc.) was used to ensure the spectropho-tometer was within specification.All enzymatic assays were performed the day after blood sample collection and mitochondria isolation.All assay kits were bovine species reactive and the intraassay coefficient of variation (CV) for controls and samples was <5%, and the interassay CV for all kits was <15%.Assay sensitivity data, where appropriate, can be found in the manufacturer's protocol.Spontaneous product conversion (background) was determined for each kit by measuring the slope of blank wells containing only the reaction solution.This activity was determined for each plate and subtracted from the activity of each sample run per plate.Each enzymatic activity was determined with the following assay kits.
Complex I (EC 1.6.5.3)Enzyme Activity Microplate Assay Kit (Abcam, ab109721) was used to determine the activity of complex I via immunocapture and spectrophotometric analysis.In short, activity was determined by an increase in absorbance at 450 nm following the oxidation of NADH to NAD + and the simultaneous reduction of dye.Kinetic readings were measured at room temperature, 450 nm, and 20-s intervals for 30 min with shaking between readings.
Complex IV (EC 1.9.3.1)activity was measured using the Complex IV Human Enzyme Activity Microplate Assay Kit (Abcam, ab109909).Complex IV was immunocaptured and activity was determined by decreased absorbance at 550 nm resulting from the oxidation of reduced cytochrome c.Kinetic readings were measured at 30°C at 3-min intervals for 60 min without shaking between readings.
Complex V (EC 3.6.3.14)enzyme activity was determined using the ATP Synthase Enzyme Activity Microplate Assay Kit (Abcam, ab109714).The hydrolysis of ATP to ADP facilitated by immunocaptured ATP Synthase was coupled to the oxidation of NADH to NAD + resulting in reduced absorbance at 340 nm.Kinetic readings were measured at 30°C at 1-min intervals for 60 min without shaking between readings.

Statistical Analysis
Cow was the experimental unit of interest and enzyme activity was defined as the linear rate of change of the absorbance per minute per microgram of crude mitochondrial protein loaded into the well.Only pre-steady-state kinetics were evaluated.The slope for each sample was determined using the GLM procedure of SAS (version 9.4, SAS Institute Inc.) to regress absorbance on time with outlier removal set at 2 standard deviations and final activities corrected by crude mitochondrial protein.The model was Y OD = β 0 + β 1 Time + ε OD , in which Y OD = optical density, β 0 = y intercept, β 1 = regression coefficient of time, and ε OD = the error.
Enzymatic activity and hematological variables were modeled 2 ways, the first as a single time point and the second as the difference between 2 time points.This allowed mitochondrial and hematological outcomes to be evaluated at a given stage of life and also explored how they changed in response to age.Variables that represent a difference between 2 time points (in wk) are noted with the delta symbol (△) between the time points (e.g., variable_2△1, the difference in the variable from 2 wk to 1 wk), whereas single time point variables are expressed with a single time point following the variable (e.g., variable_1, the variable at 1 wk).For data analysis, respiratory score covariates were defined as days with a score ≥3 and fecal score covariates were defined as days with a score >3.The number of preweaning treatments covariate was a count of all individual treatments administered to a calf (lactated ringers, electrolytes, and antibiotics).Calculations of ADG were determined using the BW measurement from 1 wk as the starting weight for all subsequent ADG calculations.Only covariates with P ≤ 0.05 were included in the models.All models were visually assessed for fit and residual uniformity, and covariates were assessed for collinearity and removed from the models if they had a variance inflation factor greater than 5.
Logistic regression analyses were conducted to evaluate if mitochondrial function and preweaning health indices affected survivability of calves using the Logistic procedure of SAS (version 9.4).Survivability was defined as 0 = removed from the herd, or 1 = survived to lactation (lactation 1, 2, 3, and 4).Removal from the herd was determined by farm records and only cows that died or were culled for disease or reproductive failure were included in the analysis.Single time point mitochondrial enzyme activity, difference in mitochondrial enzyme activities, respiratory scores, number of preweaning treatments, fecal scores, single time point hematological values, and differences in hematological values were assessed as risk factors by splitting each variable into halves (above and below the median) and assessing if calves below the median had increased odds of being removed from the herd when compared with calves above the median.The model was Logit(p) = β 0 + β 1 ENZ + β 2 RESP + β 3 TRT + β 4 FEC + β 5 HEM, where p is the probability of being removed from the herd, β 0 = y intercept, β 1 = regression coefficient of enzyme activity (ENZ), β 2 = regression coefficient of respiratory score (RESP), β 3 = regression coefficient of number of preweaning treatment (TRT), β 4 = regression coefficient of fecal score (FEC), and β 5 = regression coefficient of hematological value (HEM).

RESULTS AND DISCUSSION
This study explored how the mitochondrial enzymatic activities of citrate synthase, complex I, complex IV, and complex V in Holstein and Jersey dairy cows change with time and are associated with ADG, reproductive outcomes, first-lactation milk production, and survival.

Mitochondrial Enzyme Activity and Changes With Age
To evaluate how PBMC mitochondrial enzyme activities changed from birth to first lactation, the least squares means of citrate synthase, complex I, complex IV, and complex V from each time point were plotted for Holstein (Figure 1) and Jersey cows (Figure 2).For both breeds, there was a trend of increased enzymatic activity from weaning (8 wk) to first lactation (110 wk), where each enzyme has maximal activity at 110 wk.The activity of citrate synthase has been associated with mitochondrial number (Holloszy et al., 1970;Williams et al., 1986) and complexes I and IV are 2 of the 3 enzymes in the electron transport chain that form the electrochemical gradient that produces ATP through complex V.The maximal activity observed at 110 wk for all enzymes likely resulted from the increased metabolic pressure the cows faced, as this time point was between 55 and 75 DIM in their first lactation.These results agree with Niesen et al. (2022), where differences in mitochondrial enzymatic activity were observed between high-and low-producing lactating cows (55-75 DIM), indicating that metabolic pressure can affect mitochondrial response.Similarly, Brown et al. (1988) observed a positive association between lactation performance and mitochondrial respiration activities.In addition to lactational pressure, these heifers were still growing, and increased activity of enzymes interrelated to ATP output may help them meet their energy requirements during this metabolically demanding time.
At 52 wk there was a decrease in activity of citrate synthase, complex IV, and complex V compared with 8 wk for both Holsteins (Figure 1A, 1C, 1D) and Jerseys (Figure 2A, 2C, 2D).Complex I increased from 8 to 110 wk for both Holsteins (Figure 1B) and Jerseys (Figure 2B).Because citrate synthase, complex IV, and complex V had a decrease in activity at 52 wk, selected hematological values were plotted to determine if the cows experienced shifts in blood cell traits near this time (Figure 3).For both Holstein and Jerseys, lymphocyte number increased and neutrophil number decreased at 52 wk.Increased lymphocytes can result from viral, bacterial, or parasitic pressure and decreased neutrophils limit the ability to fight infection.There were no health events in farm records that explained the shifts in white blood cell populations.However, nutritional deficiencies can affect neutrophil differentiation (Robertson et al., 1992;Tsai and Collins, 1993) and negatively affect mitochondrial homeostasis (Acin-Perez et al., 2010).The increase in lymphocyte number and decrease in neutrophil number were within the equipment's normal ranges for adult cows (2.5-7.5 K/μL and 0.6-0.41K/μL, respectively) and agree with adult reference ranges observed in Roland et al. (2014), but these heifers were not fully grown.The shift in cell populations seen at this time could indicate that the cows were experiencing immunological or nutritional stress and explain the decreased mitochondrial activity of citrate synthase, complex IV, and complex V at 52 wk.Conversely, it is possible that heifers were minimally challenged at this time, as they were past the immune challenge events faced in the hutches and are not yet experiencing the pressures of pregnancy and lactation.Further research is needed to explain whether a decrease in mitochondrial activity at this time point is normal and explore shifts in blood cell parameters near breeding.Complex I activity was not affected by this perturbance that was reflected in lymphocyte and neutrophil populations.

ADG and Milk Production
Since ADG and milk production can be influenced by a variety of factors, multivariate regression models were developed to identify variables that correlate with ADG and first-lactation milk production in Holstein and Jersey cows (Tables 2 and 3).For Holsteins, mitochondrial enzymes, preweaning health indices, and red blood cell hematological traits were present in the models for 8, 36, 52, and 110 wk ADG (R 2 = 0.63, R 2 = 0.72, R 2 = 0.70, and R 2 = 0.99, respectively, Table 2).Jersey models were similar, and mitochondrial enzymes, prewean- ing health indices, red blood cell hematological traits, and neutrophils were correlated with 8, 36, 52, and 110 wk ADG (R 2 = 0.64, R 2 = 0.77, R 2 = 0.70, and R 2 = 0.76, respectively).Complex I, fecal scores, and mean corpuscular hemoglobin appeared more frequently in the models of Holstein ADG compared with Jerseys.In contrast, neutrophils and respiratory scores were more frequently included in Jersey growth models compared with Holsteins.
For both breeds, complex I_2△1 was the covariate that had the greatest effect on milk production models, as indicated by the greatest model variable sum of squares error.The model for Holstein milk yield was the only exception, where complex V_2△1 had the largest model variable sum of squares error.
The breeds differed by the early-life variables that were correlated with their first-lactation milk production.Numbers of preweaning treatments, fecal score, and complex V activity were included in Holstein production models more frequently than Jerseys, and citrate synthase activity was present in Jersey production models and not Holstein (Table 3).Jerseys had increased citrate synthase activity from 1 to 2 wk (Figure 2A) and Holsteins did not (Figure 1A).This could indicate differences in mitochondrial number (Kirby et al., 2007) and may explain why different mitochondrial enzymes are associated with future milk production across breeds.The repeated inclusion of complex I and complex V in the ADG and milk production models is likely the result of their role in the production of ATP.These results agree with previous works that found complex I is correlated with BW gain in heifers and complex I and V are associated with high milk production (Niesen and Rossow, 2019;Niesen et al., 2022).The numbers of preweaning treatments, neutrophil number, hematocrit, mean corpuscular hemoglobin, and fecal and respiratory scores in the models implicate the importance of calf health and nutrition metrics.Combined, these model variables could indicate the health, nutrition, and energy status of the heifers, which affect production outcomes such as treatments, mortality risk, ADG, increased age at first calving, and reduced first-lactation milk yield (Bach, 2011;Heinrichs and Heinrichs, 2011;Buczinski et al., 2021).

Mitochondrial Enzyme Activity Reproduction and Survival
To determine the effect of growth, mitochondrial enzyme activity, and hematological parameters on reproductive outcomes, multivariate regression models were developed (Table 4).For Holsteins, ADG, mitochondrial enzymes, and red blood cell hematological traits were present in the reproduction models for age at first service, age at first conception, age at first calving, and number of services (R 2 = 0.91, R 2 = 0.93, R 2 = 0.89, and R 2 = 0.47, respectively).For the majority of the Holstein models, ADG_8 was the growth covariate correlated with reproductive outcomes, with the exception of age at first service, which was correlated with ADG_36.These findings signal that preweaning growth rather than postweaning growth was better at predicting reproductive success in Holstein heifers and agree with previous work showing growth rates are associated with reproductive outcomes (Gardner et al., 1977;Cooke et al., 2013).For mitochondrial enzymes, citrate synthase_8△1 was correlated with all reproductive outcomes, and complex IV_52△8 and complex V_8△2 were included in 3 out of the 4. Ge et al. (2012) reported that mitochondrial metabolism affected oocyte development and subsequent embryo development in mice, which could explain why mitochondrial enzymes linked to energy production correlate with reproductive outcomes in cattle.Last, all Holstein reproduction models included one or both red blood cell hematological covariates, mean corpuscular volume_8, and mean corpuscular hemoglobin_52△8.For Jerseys, ADG, mitochondrial enzymes, and red blood cell hematological parameters were also correlated with age at first service, age at first conception, age at first calving, and number of services (R 2 = 0.55, R 2 = 0.73, R 2 = 0.55, and R 2 = 0.70, respectively).Jerseys differed from Holsteins in that ADG_36 was the growth covariate included in the models, indicating that postweaning growth better predicts Jersey reproductive outcomes.Similar to Holsteins, Jersey models included citrate synthase_8△1, complex IV_52△8, complex V_2△1, and complex V_8△2.These results differ from the models of ADG and milk production, in that mitochondrial changes later in life (8△1, 8△2, 52△8) were correlated with reproductive outcomes and early-life mitochondrial changes (2△1) were correlated with ADG and milk production (Tables 2 and 3).For red cell variables, Jerseys differed from Holsteins in that mean corpuscular volume_8 was included in all reproductive models rather than mean corpuscular hemoglobin (Table 4).Mean corpuscular volume estimates the average size of red blood cells, and mean corpuscular hemoglobin is an estimate of the average hemoglobin held per red cell.The covariates in these models signify the importance of BW gain, mitochondrial function, and oxygen-carrying capacity to reproductive outcomes in Holstein and Jersey heifers.They are similar to the models of ADG, and milk production, in that they include mitochondrial covari-   The least squares means of the item. 3 The difference in complex I enzyme activity from 2 to 1 wk (units are milli-optical density/min per μg of mitochondrial protein). 4 The difference in complex V enzyme activity from 2 to 1 wk (units are milli-optical density/min per μg of mitochondrial protein). 5 The number of days with a respiratory score ≥3 during the first month of life. 6 The number of treatments administered by farm staff during the preweaning period. 7 The number of days with a fecal score >3 during the first month of life.
8 Hematocrit at 2 wk (units are %).9 Mean corpuscular hemoglobin at 8 wk (units are pg). 10 The difference in mean corpuscular hemoglobin from 2 to 1 wk (units are pg).
ates integral to energy production, and hematological variables that are linked to oxygen-carrying capacity.Logistic regression analyses were performed to determine if being below the median for a particular variable increased risk of dying or being culled across lactations.Being below the median for complex V_1, complex V_8△2, mean corpuscular hemoglobin_2△1, and he-matocrit_2△1 were correlated with removal from the herd (Figure 4).For complex V_1, calves below the median were more likely to be removed from the herd compared with calves above the median by lactation 1, 2, 3, and 4 (Figure 4A, odds ratio = 4.7, 7.7, 7.0, and 6.9, respectively).For complex V_8△2, the majority of calves below the median showed no change or a decrease in activity during this time range (Figure 4B).These calves were more likely to be removed from the herd compared with calves above the median by lactation 2 and 3 (odds ratio = 6.9 and 5.2, respectively).These findings indicate that increased complex V activity near birth and the calves' ability to increase their complex V activity across the preweaning period are protective against early culling or death.The majority of calves below the median for mean corpuscular hemo-globin_2△1 showed no change or a decrease in mean corpuscular hemoglobin during this time range (Figure 4C).These calves were more likely to be removed from the herd compared with calves above the median by lactation 1, 2, 3, and 4 (Figure 4C, odds ratio = 4.7, 5.0, 4.2, and 4.1, respectively).Panousis et al. (2018) reported decreases in mean corpuscular hemoglobin from 1 to 9 d that align with what was observed in calves below the median (Figure 4C).However, in this study, increasing mean corpuscular hemoglobin per red cell was protective when compared with calves that show no change.Last, for hematocrit_2△1, calves below the median were more likely to be removed from the herd compared with calves above the median by lactation 1, 2, 3, and 4 (Figure 4D, odds ratio = 13, 10, 5.2, and 4.7, respectively).Similar to mean corpuscular hemoglobin, these findings indicate that calves that are able to increase their hematocrit percentage early in life are protected against early removal from the herd.To our knowledge, no previous research has explored the relationship between calf complex V activity, mean  ), or relative value, where β 0 = y intercept, β 1 = regression coefficient of enzyme activity for citrate synthase (Enz1), β 2 = regression coefficient of enzyme activity for complex I (Enz2), β 3 = regression coefficient of enzyme activity for complex V (Enz3), β 4 = regression coefficient of respiratory score (RESP), β 5 = regression coefficient of number of preweaning treatments (TRT), β 6 = regression coefficient of fecal score (FEC), β 7 = regression coefficient of hematocrit (HCT), β 8 = regression coefficient of mean corpuscular hemoglobin (MCH), β 9 = regression coefficient of neutrophils (NE), and ε = the error, with the criteria for inclusion being P ≤ 0.05. 2 The least squares means of the item.
corpuscular hemoglobin, and hematocrit to survival outcomes.Hematocrit and mean corpuscular hemoglobin are linked to cellular oxygen, the final electron acceptor in the electron transport chain, and complex V is the site of ATP production.Therefore, it is logical to conclude that reduced performance of these variables would result in energetic stress to the cow and affect her health and survivability.

CONCLUSIONS
Models including mitochondrial enzyme activities of citrate synthase, complex I, complex IV, and complex V as well as early-life health indices and hematological values were associated with ADG, reproductive outcomes, future milk production, and survival across breeds.When considering the models of ADG, milk production, reproduction, and survival together, all include variables indicative of health, nutrition, and energy status of the heifers.By monitoring mitochondrial function, early-life health traits, and hematological parameters, farms could identify high-risk animals and make informed and strategic breeding and culling decisions about their youngstock.Focusing financial resources on long-living, high-producing heifers would maintain profitability and reduce environmental expenses such as manure and methane.

ACKNOWLEDGMENTS
This work was supported by the USDA National Institute of Food and Agriculture (Washington, DC), Box and whisker plots of variables correlated with calf survival with odds ratios of calves below the median being removed from the herd by lactation number.Only significant odds ratios (P ≤ 0.05) are presented.Complex V activity at 1 wk (A), the difference in complex V activity from 8 to 2 wk (B), the difference in mean corpuscular hemoglobin from 2 to 1 wk (C), and the difference in hematocrit from 2 to 1 wk (D).mOd = milli-optical density.The upper and lower whiskers represent the maximum and minimum, respectively; upper and lower edges of the boxes represent the third and first quartiles, respectively; the midlines represent the median (second quartile); the diamond represents the mean.
Hatch/Multistate project.Graduate student funding was provided by the California Dairy Research Foundation (Davis, CA).The authors have not stated any conflicts of interest.

Figure 1 .
Figure 1.Enzymatic activity of peripheral blood mononuclear cells in Holstein cows from birth to first lactation.Citrate synthase activity versus time (A), complex I activity versus time (B), complex IV activity versus time (C), and complex V activity versus time (D), where n = 23, 23, 22, 20, and 10 at 1, 2, 8, 52, and 110 wk, respectively.mOd = milli-optical density.Error bars represent the SE.

Figure 2 .
Figure 2. Enzymatic activity of peripheral blood mononuclear cells in Jersey cows from birth to first lactation.Citrate synthase activity versus time (A), complex I activity versus time (B), complex IV activity versus time (C), and complex V activity versus time (D), where n = 23, 23, 19, 19, and 17 at 1, 2, 8, 52, and 110 wk, respectively.mOd = milli-optical density.Error bars represent the SE.

Figure 3 .
Figure 3. Lymphocyte and neutrophil yields of Holstein and Jersey cows from birth to first lactation.Lymphocyte number versus time (A) and neutrophil number versus time (B) for Holstein (○, solid line) and Jersey (□, dashed line) cows.Error bars represent the SE.
respiratory score (RESP), β 5 = regression coefficient of number of preweaning treatments (TRT), β 6 = regression coefficient of fecal score (FEC), β 7 = regression coefficient of hematocrit (HCT), β 8 = regression coefficient of mean corpuscular hemoglobin (MCH), β 9 = regression coefficient of neutrophils (NE), and ε = the error, with the criteria for inclusion being P ≤ 0.05. 2 Figure 4. Box and whisker plots of variables correlated with calf survival with odds ratios of calves below the median being removed from the herd by lactation number.Only significant odds ratios (P ≤ 0.05) are presented.Complex V activity at 1 wk (A), the difference in complex V activity from 8 to 2 wk (B), the difference in mean corpuscular hemoglobin from 2 to 1 wk (C), and the difference in hematocrit from 2 to 1 wk (D).mOd = milli-optical density.The upper and lower whiskers represent the maximum and minimum, respectively; upper and lower edges of the boxes represent the third and first quartiles, respectively; the midlines represent the median (second quartile); the diamond represents the mean.

Table 1 .
Niesen and Rossow: MITOCHONDRIAL ENZYME ACTIVITY IN CALVES The number of Holstein and Jersey cows removed at each time point and reasons for exiting the herd 2Farm records indicate sold to another dairy.

Table 2 .
Multivariate regression of mitochondrial enzyme activities and health indices that contribute to ADG (kg/d) in Holsteins and Jerseys Item