Carry-over effects of maternal late-gestation heat stress on granddaughter’s growth and mammary gland development

Maternal (F 0 ) exposure to late-gestation heat stress reduces their daughter’s (F 1 ) mammary gland fat pad mass (FP), parenchyma (PAR) mass, and epithelial cell proliferation when evaluated at birth and weaning, and go on to produce less milk in their first lactation. Herein, we investigated the effect of maternal late-gestation heat stress on whole-body growth and mammary development of their granddaughters (F 2 ). Multiparous F 0 cows had access to heat abatement (n = 41, shade, and active cooling via fans and water soakers) or not (n = 41, shade only) for the last 56 d of gestation during a subtropical summer. Consequently, the F 1 daughters, born to F 0 cows, were heat-stressed (HT F1 , n = 36) or cooled (CL F1 , n = 37) in utero during the last 2 mo of gestation. All F 1 heifers were raised as an identically managed cohort until first calving. The F 2 granddaugh-ters, born to HT F1 (HT F2 , n = 12) or CL F1 (CL F2 , n = 17), were raised as an identically managed cohort until 70 d of age. Dry matter intake (DMI), body weight, hip height, wither height, chest girth, head circumference, mammary gland teat length, and left-right and front-rear teat distances were measured. Average daily gain (ADG) was calculated for the pre-weaned period (0–49 d). Mammary ultrasounds were performed on d 21, 49, and 70 (n = 9/group) on the rear left and right quarters to quantify PAR and FP areas. Mammary biopsies were collected for histological evaluation of epithelial structures (H&E staining), and to quantify cells positive for ERα (estrogen receptor, α subunit), cell proliferation (Ki67), and apoptosis (TUNEL). Heifer growth from birth to d 49 was similar between CL F2 and HT F2 for all parameters evaluated. Distances between teats and teat length were not different between groups. On d 70, CL F2 tended to have a greater average PAR (right and left quarters) relative to HT F2 . Although the left FP was smaller in HT F2 relative to CL F2 , the average FP was not different. The lumenal and non-lumenal epithelial structures in the PAR of HT F2 were significantly smaller than those of CL F2. In addition, HT F2 had a reduced percentage of proliferating cells in the epithelial and stromal compartments and a greater percentage of apoptotic cells, particularly in the stroma. The percentage of ERα positive cells was significantly reduced in HT F2 . In summary, although HT F2 heifer’s DMI was similar and they grew at the same rate as CL F2 heifers throughout the pre-weaning phase, their mammary glands had smaller PAR areas with fewer epithelial structures characterized by reduced cell turn-over and lower ERα expression. These early changes in the microstructure and cellular turnover of the mammary gland may partly explain the reduction in lactation performance relative to CL F2 counterparts at maturity.


INTRODUCTION
Climate change trends are leading to an increase of heat stress days regardless of geographical location.Consequently, more cattle are exposed to environmental heat stress, regardless of the physiological stage or age of the animals.Exposure of pregnant cows (F 0 ) to heat stress during the last trimester of gestation negatively impacts their first-generation daughters (F 1 ), who experienced heat stress in utero.Numerous studies have investigated the detrimental effects of in utero heat stress during late gestation on the F 1 progeny.Even when F 1 heifers are managed as a cohort after birth, those experiencing in utero heat stress showed lower survivability rates, decreased passive transfer of immunity, shorter gestation times, and lower birth weights compared with those born to cows kept cool in late gestation (Dado-Senn et al., 2020;Ouellet et al., 2020;Cattaneo et al., 2022).The negative impacts of intrauterine hyperthermia on growth and development are long-lasting.Indeed, F 1 daughters born to heatstress cows remain smaller and shorter up to 1 year of age (Monteiro et al., 2016).Further, when these animals reach maturity and enter the lactating herd, they produce significantly less milk in their first lactation, despite having similar mature body weight at calving relative to their in-utero cooled counterparts (Monteiro et al., 2016;Laporta et al., 2020).One possible explanation for this reduction in first lactation milk yield could be attributed to alterations in mammary gland size, microstructure, and cellular turnover observed in early life (Dado-Senn et al., 2022), some of which persist until their first lactation, 2 years after the prenatal insult occurred (Skibiel et al., 2018).More recently, through retrospective analysis of production records, our group has shown that the reduction in milk yield of F 1 heifers exposed to intrauterine hyperthermia persists through their 2nd and 3rd lactations, indicating a permanent carry-over effect on mammary performance (Laporta et al., 2020).Indeed, a higher culling rate manifests in the second-generation granddaughters (F 2 ) (Laporta et al., 2020).
Although multigenerational effects are well-studied and extensively documented in other species (Shiota and Kayamura, 1989;Ross et al., 2015;Truong et al., 2023), there's still much to discover and document in dairy cattle.In recent years, there has been a research focus aimed at identifying and characterizing phenotypic differences and understanding the mechanistic link between early-life exposure to environmental heat stress and the long-lasting phenotypes.During the last 2 mo of pregnancy, fetal growth increases exponentially, and maternal exposure to heat stress during this developmental period can "directly" impact the growth and development of F 1 fetal daughters.However, the impact on the F 2 granddaughters is different.The oocytes (female gametes, germline) begin developing in the F 1 fetal daughter, and her entire reserve is present and readily available in the ovaries when she is born (Erickson, 1966).It is thought that late gestation inutero heat stress could impact the F 2 generation "indirectly" through the fetal daughter germline (Hedhly et al., 2020;Rogers and Phillips, 2021).To date, it is unclear if the reduced milk production at maturity of the F 2 granddaughters (Laporta et al., 2020) is caused by disparities in body size, as a consequence of stunted mammary gland development as observed in F 1 daughters, or both.
To ensure optimal production and herd longevity in the face of climate change, it is becoming more evident that the experiences of previous generations might impact future ones and, therefore, should be considered when selecting replacement animals.This study seeks to characterize the body growth parameters and mammary gland growth trajectories of F 2 granddaughters of cows that were either heat-stressed or cooled during late gestation in a subtropical summer.We hypothesize that F 0 heat stress exposure in late gestation will negatively impact the F 2 generation's growth and mammary development, consistent with our previous observations in the F 1 daughters.Developing effective strategies to reduce the production losses arising from late-gestation heat stress in dairy cattle requires a deeper understanding of its multigenerational phenotypes and the molecular mechanisms driving it.

Maternal (F 0 ) treatments and generation of daughters (F 1 ) and granddaughters (F 2 )
In the summer of 2020, at a commercial dairy farm in Trenton, Florida, pregnant Holstein dams (F 0 , granddams, n = 82 blocked by parity, mature-equivalent milk and offspring sire) were exposed to naturally occurring environmental heat stress (HT F0 , shade of a free stall barn, n = 41) or exposed to the same environment but with access to active cooling (CL F0 , fans and water soakers, n = 41) during the last 56 ± 5 d of gestation (i.e., the entire dry period) (Dado-Senn et al., 2021).The first generation of female daughters (F 1 , n = 73) born to heat stressed or cooled F 0 dams experienced the treatments via the intrauterine environment (HT F1 , n = 36; CL F1 , n = 37).These daughters, relocated to Arlington Wisconsin, were managed as a cohort from birth until puberty, when they were artificially inseminated using sexed semen from 5 sires balanced between the 2 groups.The F 1 daughters continued to be managed as a cohort during pregnancy until their first calving (Dado-Senn et al., 2021;Davidson et al., 2022) and gave birth to the second generation of female granddaughters (F 2 , n = 30, from August to October 2022).The F 2 heifers were exposed to their granddam's heat-stressed or cooling treatments in utero indirectly through the F 1 fetal daughter germline.No treatment was applied to the F 2 generation.These groups will be referred to as HT F2 (n = 12) and CL F2 (n = 18).

Management of granddaughters (F 2 ) from birth to weaning
The F 2 granddaughters were raised in individual sand-bedded polyethylene calf hutches (Calf-Tel, L. T. Hampel Corp.) at the University of Wisconsin-Madison Arlington Agricultural Research Station.All heifers received colostrum (average BRIX 26%) within 4 h of birth.Afterward, all heifers received whole pasteurized milk in 2 daily meals (2 quarts/meal the first 2 d of life, 3 quarts/meal from 3 to 14 d, and 4 quarts/meal from 15 to 42 d).At 42 d, milk was decreased incrementally and ceased at 49 d.Health was monitored daily by the farm staff, and all heifers were treated according to farm standard operating protocols by trained veterinarians.Starter (VitaPlus medicated 18% crude protein, Larsen and Laporta: PRENATAL HEAT STRESS ON F2'S MAMMARY GROWTH 5.2% fat, Table 1) was offered ad libitum beginning at birth, and intake was monitored from 28 d through 56 d.At 56 d, all heifers were moved to a 4-calf superhutch (Caft-Tel, L. T. Hampel Corp.) until 12 wk of age.All heifers were transitioned to a grower grain at this stage, and individual intake was not monitored.

Whole-body growth and mammary gland measurements
At 1, 7, 21, 35, 49, and 70 d old, hip height, withers height, chest girth, head circumference, and body weight were measured.Hip height (i.e., ground to hook) and wither height (i.e., ground to top of wither) were measured using a tape measure.Chest (heart) girth was measured with a soft tape measure directly behind the front legs and around the body, and head circumference was measured around the head in front of the ears with a soft tape measure.Body weight was measured using a calf cart with a built-in scale.At 21, 49, and 70 d old, mammary ultrasounds (Mindray Z60, Shenzhen, China) were performed using a convex probe.Heifers were in a standing position at 21 and 49 d, where all 4 quarters were imaged.At 70 d, a subset of heifers' (n = 8-10 per treatment) left and right rear quarters were scanned with a convex probe in the supine position where visible parenchyma and fat pad were captured.The probe depth was set at 2.8 for all calves, and the visible parenchyma and fat pad were captured in a 4-s video for each quarter.The probe was placed directly behind the teat, the gland cistern was located, and the video was captured.At d 21, 35, 39, and 70, measures of teat length (base of the teat to the tip) and distance between teats (center of teat base to center of teat base) were recorded using calipers.Individual frames from each video were isolated, and the frame with the greatest visible parenchyma area was chosen for each animal at each time point evaluated.The surface areas of parenchyma and fat pad tissue were measured using Q-Path version 0.4.3 (Bankhead et al., 2017).Pixels 2 were converted to mm 2 in Excel (Microsoft).

Mammary gland biopsies and histological analysis
At 70 ± 5 d of age, post-weaning mammary biopsies were performed on a subset of heifers (n = 6 per group).Briefly, all heifers were given intravenous Xylazine (AnaSed, Lloyd Inc., Iowa) at 0.02 mg/kg and placed in the supine position.The mammary gland parenchyma was palpated, and 2 mL of lidocaine (Clipper Distributing Company, Missouri) was administered into the gland.All biopsies were taken from the rear right gland quarter.Using a #15-scalpel blade, a small incision was made, and a 2 mm biopsy punch (Integra Miltex #33-31) was inserted and twisted in a clockwise motion to cut and collect the tissue.After the procedure, Flunixin Meglumine (Flunazine, Bimeda, Illinois) was administered intravenously at a 1.6 mg/ kg dose.The PAR tissue (~20-60 mg) was washed in cold PBS, placed in tissue cassettes, and stored in 10% NBF (Neutral Buffered Formalin) for 18-20 h.The tissue was paraffin-embedded, sectioned at 5 µm, and stained with H&E to visualize the architecture of the epithelial structures.Additional sections were used to assess cellular turnover and estrogen receptor α (ERα) via immunohistochemistry and immunoperoxidase procedures.Briefly, Ki-67 antibody (Mouse anti-Human Ki67, SAKO #M7240, clone MIB-1) was used to quantify cellular proliferation.To evaluate cell death, the TUNEL method was used with the ApopTag Peroxidase In Situ Apoptosis Detection Kit (Millipore Sigma #s7100) according to the manufacturer's protocol.To evaluate ERα, sections were incubated with mouse monoclonal ERα primary antibody (1:100, Santa Cruz Biotechnology #C-311sc787) for one hour with stain visualization using Mach 2 Mouse HRP polymer (Biocare Medical, #MHRP520L).
Tissue images were captured using a brightfield microscope (Keyence BX800, Keyence Corporation of America, Osaka, Japan).Five random photomicrographs were taken per animal per stain.The H&Estained mammary sections were used to determine the area of epithelial structures which were classified into either lumenal (presence of a hollow lumen) or nonlumenal structures (no visible lumen).Using Image J, the areas were quantified by tracing closely around the  mammary epithelial structures.The Ki-67 antibody was used to determine the percent of proliferating cells, and the TUNEL assay was used to determine the percent of apoptotic cells.The total number of cells per photomicrograph were counted using BX-800 Keyence Analyzer software (Keyence Corporation of America, Osaka, Japan), and positive cells were hand-counted in image J.The number of cells by epithelial and stroma cell compartments was counted and analyzed separately.

Statistical analyses
All statistical analysis was performed using the statistical software SAS (SAS Institute Inc., Cary, NC).Repeated measures of calf body and mammary growth, feed intake, and efficiency were analyzed through ANO-VA using the MIXED procedure.The model for growth (BW, HH, WH, CG, HC) and ultrasonography at 21 and 49 d included fixed effects of maternal F 0 treatment (TRT), time (d), and their interaction.Animal (ID) within TRT was used as a random effect.Ultrasonography and epithelial structure data collected at 70 d were analyzed with TRT as a fixed effect and ID as a random effect.The GLIMMIX procedure was used to analyze cell proliferation, apoptosis, and ERα count data.The model evaluated the ratio of total positively stained cells to total cells for each tissue cell type, with TRT as the fixed effect and ID within replicate as a random effect.All residuals were tested for normality.Box-Cox was used to determine which transformation was necessary if normality was unmet.The mammary lumenal epithelial areas were log-transformed based on Box-Cox lambda.The Least Square Means ± standard error is presented unless noted otherwise.A P-value ≤0.05 was considered statistically significant, and Pvalues >0.05 and ≤0.10 were considered tendencies.

RESULTS
Whole-body growth and mammary gland macrostructural outcomes are presented in Table 2.There were no significant effects of the F 0 treatment or interactions for the whole-body and macrostructural mammary growth parameters evaluated in the F 2 generation (P > 0.12).A main effect of day was observed for all variables (P < 0.01), reflecting a similar age-related growth trajectory for all heifers through the first 10 weeks of life (d 1-70).
Mammary ultrasound outcomes evaluated at d 21 and 49 are presented in Table 3.For all the parameters analyzed via ultrasound, there were no significant effects of the maternal treatment or significant interactions (P > 0.25).However, an effect of the day was observed (P < 0.04), whereby the visible parenchyma and fat pad areas increased from d 21 to 49.
Outcomes related to PAR epithelial structures and ERα cells are shown in Figure 1.The PAR tissue of CL F2 had greater non-lumenal and lumenal structure areas than HT F2 (Figure 1b; non-lumenal: 23,386 ± 1,738 vs. 16,789 ± 1,570 um 2 P = 0.005; lumenal: 71,125.41± 8,443 vs 48,484.53± 3,556 um 2 P = 0.006 for CL F2 and HT F2, respectively).The percent of ERα positive cells was greater in the PAR epithelial compartment of CL F2 when compared with that of HT F2 (Figure 1d; 53.5% ± 3.3 vs. 40.0%± 2.7; P = 0.0033).The percent of ERα positive cells did not differ between groups in the PAR stromal compartment.
Outcomes related to mammary gland ultrasonography collected at d 70 are presented in Figure 3. Visible parenchyma area was greater in CL F2 heifers compared with HT F2 for the right rear quarter (Figure 3c; Right: 2,771.03+ 253.41 vs 1,885.98 + 283.32 mm 2 P = 0.045) as well as the average of the right and left rear quarters (average: 2793.9 ± 222.6 vs 2097.6 ± 236.1 P = 0.05).There were no differences in the visible parenchyma areas of the left rear quarters (P > 0.42).There were no differences in the visible fat pad of the right rear quarters nor the average of the right and left rear quarters combined (P > 0.12).However, CL F2 heifers had a greater visible fat pad area of the left rear quarter than HT F2 (4593.6 ± 383.3 vs. 3366.0± 428.51 P = 0.05).

DISCUSSION
This is the first study to assess macrostructural and microstructural phenotypes in the second generation (F 2, granddaughters) of dairy cows undergoing heat Larsen and Laporta: PRENATAL HEAT STRESS ON F2'S MAMMARY GROWTH stress during late gestation.A key finding is that the F 2 granddaughters born to heat-stressed granddams do not manifest the classic growth retardation hall-marks observed in utero heat-stressed F 1 daughters (i.e., reduced birth weight and stature; Monteiro et al., 2016).Additionally, phenotypic mammary gland   4 Ultrasound videos were taken from the front and rear right and left mammary gland quarters.macrostructural differences were not observed between HT and CL F 2 granddaughters, as previously reported for CL and HT F 1 daughters (i.e., shorter teat length and distance; Dado-Senn et al., 2022).However, despite their unimpacted whole-body and macrostructural mammary growth trajectory, the mammary gland epithelial microstructure and cellular turnover of HT F2 's are significantly stunted.The findings of this study are relevant to the dairy industry, as farmers often rely on pre-established body weight and stature benchmarks to make culling and replacement decisions.Despite having lower production potential, HT F2 faces the same risk of being culled as CL F2 .The lack of phenotypic differences between F 2 heifers born to in utero HT F1 or CL F1 's highlights a critical pitfall in the management of multigenerationally impacted heat-stressed animals, as HT F2 heifers will pass through herd checkpoints while their mammary glands are programmed to perform at suboptimal levels, ultimately affecting milk production.
The first-generation CL F1 daughters are heavier and taller than HT F1 daughters through one year of age, around the time of breeding (Monteiro et al., 2016).This disparate growth rate is not observed in the second generation of HT F2 granddaughters, where growth rates are identical to CL F2 counterparts.This could mean that the HT F2 granddaughters were able to overcome the growth challenge their mothers (in utero exposed F 1 ) faced due to more direct exposure to hyperthermia as a developing fetus.Conversely, the resulting F 2 heifers evaluated herein were exposed as a germ cell developing inside the fetal F 1 daughters at the time of maternal F 0 heat stress exposure.Therefore, the effect of hyperthermia might have been less exacerbated for the developing egg relative to the whole fetus.Yet, CL F2 granddaughters have longer, thinner hair coats and fewer but larger sebaceous glands than HT F2 granddaughters (Davidson et al., 2023).These differences may indicate that some phenotypic differ- ences exist between these animals, potentially indicating different thermoregulatory capacities arising from germline exposure to hyperthermia.
By evaluating mammary gland parenchyma tissue size via ultrasonography combined with histological assessments, we begin to unravel hidden carry-over effects of late-gestation heat stress in the F 2 generation.These alterations in mammary microstructure and cellular turnover, along with the reduced ERα, may lead to a slower mammary growth rate in response to estrogen at puberty, resulting in reduced milk synthetic capacity and lactation outcomes at maturity.Ultrasonography allows the visualization of the growing mammary gland parenchyma tissue without requiring terminal studies to harvest the glandular structures and dissect specific types of tissues post-harvest.Although there were no differences between groups in visible parenchyma or fat pad surface areas in the pre-weaned period (assessed at 21 and 49 d of age), these differences arise as the heifers begin to develop their mammary glands allometrically.It has previously been thought that the onset of al-lometric growth did not occur until puberty (Tucker, 1987), however, recent studies have shown a shift in this dogma, highlighting the importance of assessing prepubertal mammary growth while also introducing this period as a time of allometric growth (Sørensen et al., 1987;Geiger et al., 2016;Soberon and Van Amburgh, 2017) Additionally, our group has shown that PAR mass increases about 25-fold from birth to weaning (Dado-Senn et al., 2022) The lack of differences at earlier time points could be attributed to the sensitivity of the ultrasound or the standing position at which these measures were taken compared with the supine position which may increase visible surface area of the gland.Regardless, the ongoing development of more refined quantification methods and ultrasound technologies is expected to allow for an earlier assessment and detection of PAR growth differences among animals.In the present study, mammary gland ultrasounds performed at 10 weeks old revealed that CL F2 granddaughters had greater average parenchyma surface area in the rear quarters.This could indicate a greater rate of growth and development compared with that of HT F2 granddaughters.Notably, the images extracted from the ultrasound videos had similar developmental characteristics as those reported in a recent study, where calves also at 10 weeks of age had visible parenchyma in an oval shaped structure with some apparent developing ductal structures (Seibt et al., 2023).The lack of difference in PAR between groups evaluated before 6 weeks of age may have been due to the younger age of the animal and the overall smaller PAR mass due to growth rates at this time.This observation is consistent with Seibt et al. (2023), who found that most heifers had little to no visible parenchyma area until they reached 8 weeks old.
To further evaluate the development of the mammary glands in these heifers, a mammary biopsy was collected at 70 d of life to assess the microstructure and cellular makeup.We quantified the area of the ductal epithelial structures via H&E staining, revealing that CL F2 granddaughters had larger lumenal (structures with a single layer of mammary epithelial cells with a hollow lumen formed) and non-lumenal (structures with mammary epithelial cells but no hollow lumen) epithelial structures.This is consistent with data from their mothers (CL F1 ) PAR harvested at a similar age, where lumenal structures were significantly larger than their HT F1 counterparts (Dado-Senn et al., 2022).An interesting observation of the present study is the larger total area of the non-lumenal structures of HT F2 heifers.This observation, along with smaller luminal-epithelial structures manifesting lower cell proliferation, possibly reflects a delayed developmental progression of the ru- dimentary ductal tree within the smaller PAR tissue of the mammary gland of HT F2 heifers.Interestingly, the stunted development of ductal-epithelial structures in the PAR tissue is conserved through at least 2 generations of heifers, daughters, and granddaughters.This is the first study to report a multigenerational perpetuation of mammary gland microstructural alterations arising from late-gestation heat exposure.
Estrogen receptor α (ERα) plays a large role in the formation of ductal structures, especially during the early phases of allometric mammary gland growth and development.In the present study, although ERα is present and is expressed in both groups of heifers, there is significantly less expression in the HT F2 mammary epithelial cell compartment relative to CL F2 .This finding, combined with the overall smaller luminal epithelial structures, may reflect a delayed developmental progression of ductal structures as a result of the heat stress insult experienced by their F 0 grandmothers impacting their F 1 mothers and their germline in utero (Li and Capuco, 2008).Other groups have identified ERα as a regulator of estrogen-responsive genes in the bovine.One study has reported that the amount of proliferating cells decreased due to lack of ERα, however no differences in apoptosis were reported (Mallepell et al., 2006).In this study, CL F2 granddaughters had a greater percentage of proliferating cells, and a lower percentage of cells undergoing apoptosis compared with the HT F2 granddaughters.Exposure to high temperatures can cause cell death in the bovine mammary epithelial gland cells (Wohlgemuth et al., 2016;Chen et al., 2020).This suggests that the HT F2 granddaughters' impaired proliferation could be a consequence of decreased ERα expression, presumably from the heat stress insult, and the increased apoptosis may be a more direct programming effect from the heat-stressed germ cell in the fetal F 1 daughter or via another unknown mechanism not evaluated herein.Programming for higher cell death due to heat stress is a disadvantage because there may be less capacity for mammary and ductal development.Further research is necessary to understand the underlying molecular mechanisms driving the observed changes and whether heat stress during other gestation periods (i.e., early and mid-gestation) would trigger similar multigenerational effects in dairy cattle.It is imperative that we continue to investigate how heat stress impacts the fetal germline to better understand how to prevent and rescue the negative hidden effects of heat stress in these animals before there is a loss of production.

CONCLUSIONS
The naturally occurring environmental conditions of the subtropical climates in the southeastern U.S, characterized by chronic periods of high temperature and humidity, can lead to chronic heat stress in dairy cattle resulting in long-lasting damage to the production potential of multiple generations.Providing active cooling can effectively dissipate accumulated heat under these conditions, allowing the pregnant F 0 dam to maintain thermoneutrality while her F 1 daughter is developing in utero.In addition, the thermal environment in which the female germ line (that will give rise to the F 2 granddaughters) undergoes the last stages of development is critical.Despite CL F2 and HT F2 having identical growth trajectories, dry matter intakes, and phenotypic anatomy of the mammary gland, CL F2 heifers develop larger parenchyma mass, with more developed ductal-epithelial structures, greater expression of ERα, and higher rates of proliferation.There seem to be hidden multigenerational effects of late-gestation heat stress that contribute to HTF2 heifers "passing under the radar," leading to cows entering the lactating herd at a disadvantage, ultimately leaving less milk in the bulk tank.
Larsen and Laporta: PRENATAL HEAT STRESS ON F2'S MAMMARY GROWTH Larsen and Laporta: PRENATAL HEAT STRESS ON F2'S MAMMARY GROWTH

Figure 3 .
Figure 3. Ultrasonographic analysis of heifer's mammary glands.Second-generation heifers (granddaughters, F 2 ) were born to first-generation (daughters, F 1 ) cows that were heat-stressed or cooled in-utero (HT F1 and CL F1 , respectively) during late gestation (last 56 ± 5 d).Mammary ultrasounds were performed in CL F2 (n = 9) and HT F2 (n = 8) granddaughters at 70 d of age.(A) Representative photo of the position of the ultrasound probe in the rear quarter relative to the teat at the time of video imaging.(B) Ultrasonographic photos of visible mammary parenchyma (PAR, dotted highlighted areas) and fat pad (FP, solid highlighted areas) for CL F2 (green) and HT F2 (purple).(C) Quantification of parenchyma and fat pad area in left, right, and average of left and right quarters calculated using Q-path Software.Asterisk (*) represents P < 0.05.

Table 1 .
Nutritional content of the starter grain concentrate offered at libitum to Holstein dairy heifers from birth to 70 d. of age 1 Averaged (Mean ± Standard Deviation) chemical analysis of grain samples taken during experimental period.

Table 2 .
Heifer's growth trajectory, feed intake, and mammary gland macrostructural teat measures, during the preweaning period 2Average of SEM from both treatments. 3Days = 1, 7, 21, 35, 49, 70 d of age for calf growth; 21, 35, 49, and 70 for mammary growth measures.0 exposed to heat stress with access to shade, fans, and water soakers) for the last 56 d of gestation of the F 1 (first generation), HT F2 = granddaughter of heat-stressed granddams (heat stress with access to shade only). 2 SEM average from both treatments. 3Analysis of ultrasound images taken at d 21 and d 49.