Effects of wildfire smoke PM_2.5 on indicators of inflammation, health, and metabolism of preweaned Holstein heifers

Abstract

Wildfires are a growing concern as large, catastrophic fires are becoming more commonplace. Wildfire smoke consists of fine particulate matter (PM_2.5), which can cause immune responses and disease in humans. However, the present knowledge of the effects of wildfire PM_2.5 on dairy cattle is sparse. The present study aimed to elucidate the effects of wildfire-PM_2.5 exposure on dairy calf health and performance. Preweaned Holstein heifers (N = 15) were assessed from birth through weaning, coinciding with the 2021 wildfire season. Respiratory rate, heart rate, rectal temperatures, and health scores were recorded and blood samples were collected weekly or twice a week for analysis of hematology, blood metabolites, and acute phase proteins. Hourly PM_2.5 concentrations and meteorological data were obtained, and temperature–humidity index (THI) was calculated. Contribution of wildfires to PM_2.5 fluxes were determined utilizing AirNowTech Navigator and HYSPLIT modeling. Mixed models were used for data analysis, with separate models for lags of up to 7 d, and fixed effects of daily average PM_2.5, THI, and PM_2.5 × THI, and calf as a random effect. THI ranged from 48 to 73, while PM_2.5 reached concentrations up to 118.8 µg/m³ during active wildfires. PM_2.5 and THI positively interacted to elevate respiratory rate, heart rate, rectal temperature, and eosinophils on lag day 0 (day of exposure; all P < 0.05). There was a negative interactive effect of PM_2.5 and THI on lymphocytes after a 2-d lag (P = 0.03), and total white blood cells, neutrophils, hemoglobin, and hematocrit after a 3-d lag (all P < 0.02), whereas there was a positive interactive effect on cough scores and eye scores on lag day 3 (all P < 0.02). Glucose and NEFA were increased as a result of combined elevated PM_2.5 and THI on lag day 1, whereas BHB was decreased (all P < 0.05). Contrarily, on lag day 3 and 6, there was a negative interactive effect of PM_2.5 and THI on glucose and NEFA, but a positive interactive effect on BHB (all P < 0.03). Serum amyloid A was decreased whereas haptoglobin was increased with elevated PM_2.5 and THI together on lag days 0 to 4 (all P < 0.05). These findings indicate that exposure to wildfire-derived PM_2.5, along with increased THI during the summer months, elicits negative effects on preweaned calf health and performance both during and following exposure.

Preweaned dairy calves exposed to wildfire smoke may be at risk of inflammation and reduced health, and performance.

Introduction

Wildfires burn land and impose poor air quality in the surrounding regions. In the year 2021 alone, there were 58,985 wildfires that burned 7,125,643 acres in the United States (NIFC, 2021). Since the mid-1980’s, there has been a notable change in wildfire trends, including a four-fold increase in the area burned by wildfires (Burke et al., 2021) and a 84-d increase in the length of the annual wildfire season (Westerling, 2016). These recent observations have largely been attributed to historical fire exclusion practices and land-use related interferences that caused fire deficits (Marlon et al., 2012; Harris and Taylor, 2015; Parks et al., 2015), as well as shifts in climatic variables, leading to changes in ecosystem dynamics that result in biomass accumulation and flammability (Westerling et al., 2006; Littell et al., 2009; Westerling, 2016; Hanan et al., 2021). Trends for greater wildfire risk and area burned are predicted to continue into the future (Abatzoglou and Williams, 2016), despite efforts of land resource managers to reduce extreme, catastrophic wildfires (Halofsky et al., 2020). The western United States disproportionately experiences large wildfires (EPA, 2016; Nagy et al., 2018) that burn a greater number of acres as compared to the eastern United States (Nowacki and Abrams, 2008; EPA, 2016). Western states also house more than 1.8 million dairy heifer calves that are retained in production systems as replacement animals (NASS, 2022).

Wildfire emissions contain air pollutants that can travel substantial distances beyond the region in which the smoke plumes originate (Damoah et al., 2004; Hung et al., 2021). These emissions include harmful gases such as carbon monoxide, hydrogen cyanide, ammonia, volatile organic compounds, polycyclic aromatic hydrocarbons, and particulate matter (PM), among others (Urbanski et al., 2008; Reisen et al., 2015; Sokolik et al., 2019). PM, particularly fine particulate matter (PM_2.5), is often considered one of the most potent and relevant components of wildfire smoke to health. PM_2.5 consists of particles less than 2.5 μm in aerodynamic diameter and wildfire-derived PM_2.5 can contribute to up to 50% of the total PM_2.5 pollution in the western United States (Kelly and Fussell, 2012; Burke et al., 2021). Furthermore, 90% of particulate mass in wildfire smoke is PM_2.5 (Vicente et al., 2013). Wildfire-derived PM_2.5 can also be more toxic than PM_2.5 originating from other sources (Wegesser et al., 2009; Aguilera et al., 2021b) and is the pollutant most often associated with health risk in humans because of its ability to access the deeper recesses of the lung (Black et al., 2017b).

In humans, a variety of respiratory issues have been associated with wildfire-PM_2.5 exposure, including exacerbations of asthma (Borchers Arriagada et al., 2019; Kiser et al., 2020), respiratory diagnoses (Liu et al., 2015) such as bronchitis and pneumonia (Hutchinson et al., 2018), mortality (Doubleday et al., 2020; Chen et al., 2021a; Liu et al., 2021; Ye et al., 2022), and cardiovascular ailments (Dennekamp et al., 2015; Kollanus et al., 2016; Navarro et al., 2019; Chen et al., 2021b). Wildfire smoke-related morbidities and mortality in humans are thought to be caused largely by inflammation, characterized by changes in biomarkers of innate immune function and the inflammatory response (Black et al., 2017b). We demonstrated for the first time that wildfire-PM_2.5 exposure, either independently, or in combination with elevated temperature–humidity index (THI), decreases milk production, alters metabolism, and induces an innate immune response involving changes in immune cell populations in the blood of dairy cows (Anderson et al., 2022). However, little information is available concerning consequences of wildfire PM_2.5 on calf health and metabolic physiology.

Calves may be more susceptible to inhaled particulates than mature animals, as their organ and immune systems are immature, similar to human infants and young children (Barrington and Parish, 2001). In humans, exposure to PM is more hazardous for the young and the elderly (Ignotti et al., 2010; Le et al., 2014; Liu et al., 2015; Aguilera et al., 2021a). Studies on rhesus macaques found that exposure to wildfire smoke during infancy induced long term effects including nasal epigenome alterations (Brown et al., 2022), reduced lung function, and immune dysregulation (Black et al., 2017a). Moreover, it is well-established that a calf’s health early in life has long-term impacts on its future performance. For example, in a recent meta-analysis, Buczinski et al. (2021) found that dairy heifers diagnosed with respiratory disease at a young age have lower average daily gains, lower milk production in their first lactation, and higher mortality and culling rates. This has implications for future farm profitability as heifer calves are often maintained in the herd as replacement animals. Therefore, considering how dairy calves are affected by natural disasters, such as wildfires, is an important undertaking.

The aim of the present study was to provide a preliminary understanding of the effects of PM_2.5 from wildfire smoke on preweaned dairy calves. We hypothesized that calves naturally exposed to wildfire smoke PM_2.5 experience an inflammatory response, as evidenced by increases in acute phase proteins, including serum amyloid A (SAA) and haptoglobin (Hp), as well as hematological indices. Additionally, we hypothesized that health parameters and blood metabolites (e.g., glucose, nonesterified fatty acids [NEFA], and β-hydroxybutyrate [BHB] concentrations), would be negatively affected by inhalation of natural wildfire PM_2.5.

Materials and Methods

Preweaned calves

Animal procedures were approved by the University of Idaho Institutional Animal Care and Use Committee (protocol #IACUC-2019-39). Sixteen Holstein heifer calves at the University of Idaho Dairy Center were enrolled in the study from their births in July 2021 until weaning in September. Sample size was based on our previous study in dairy cows (Anderson et al., 2022) and by power analysis (alpha = 0.05, power set to 0.8, and effect size of 0.2). Calves were housed in individual hutches bedded with shavings in a barn that was fully enclosed by a roof and wall on three sides with the fourth side consisting of only half a wall. There were no sources of ventilation other than outside air flow directly into the barn, so calves were exposed to ambient air conditions, including wildfire smoke. Calves were given a first feeding of colostrum (3.8 liters) collected from cows on farm within 4 h of birth, and a second colostrum feeding (1.9 liters) within 12 h of birth. Calves were fed a combination of saleable and high SCC milk (2.8 liters) two times per day until weaning. Calves were provided ad libitum access to water and offered starter grain at 5 d of age and alfalfa at 45 d of age. Calves were weaned with two step-downs; the first step-down at 53 d of age to 2.8 liters of milk one time per day, and the second step-down to complete weaning at 60 d of age. During the study, one calf required treatment for gastrointestinal complications and was removed from the study, so the final sample size was 15 heifer calves. Two additional calves were treated for pinkeye with oxytetracycline following exposure to smoke, and these animals were retained in the study.

Environmental data

Hourly PM_2.5 concentrations (µg/m³), ambient temperature (°C), and relative humidity (%), were obtained from a local monitoring station maintained by the Idaho Department of Environmental Quality, which was located 5.7 km from the Dairy Center. Detailed information regarding instrumentation and sampling can be found in Idaho DEQ, 2021. Our area is characterized topographically by rolling hills and wheat fields, with sparse trees and other structures or topographical features that block winds, thus PM_2.5 levels recorded at the monitoring station are a fairly accurate representation of PM_2.5 levels on farm. THI was calculated with the equation shown below (Dikmen et al., 2008):

$${\displaystyle \begin{array}{l}{\displaystyle \begin{array}{lll}& {\displaystyle THI=[1.8xTemperature({}^{\circ }C)]-[\mathrm{0.550.0055}\ast RelativeHumidity\left(\mathrm{\%}\right)]x}& {\displaystyle [1.8xTemperature\left({}^{\circ }C\right)26].}\end{array}}\end{array}}$$

In order to ascertain if increases in PM_2.5 were derived from wildfires, maps with active wildfires and PM_2.5 concentrations were created using AirNow-Tech Navigator (airnowtech.org), and air mass trajectories were visualized using the National Oceanic and Atmospheric Administration’s (NOAA) HYSPLIT atmospheric transport and dispersion modeling system (Draxler and Hess 1997, 1998; Draxler, 1999; Stein et al., 2015). Three different atmospheric heights (50, 100, and 150 m) were used to map air mass trajectories over the previous 72 h. The modeling shows the origin of air masses, active wildfires, and trajectories of PM_2.5 dispersion.

Calf exposure to wildfire PM_2.5 was determined following the parameters previously described in Anderson et al., (2022): 1) 24 h average PM_2.5 concentrations were above 35 µg/m³, the U.S. EPA recommended threshold for human outdoor exposure to PM_2.5 (EPA, 2019), and 2) AirNow-Tech Navigator PM_2.5 and wildfire maps plus HYSPLIT wind trajectories showing wildfire smoke and PM_2.5 blowing from active wildfires to the location where calves were housed.

Health parameters

All calf measurements and blood samples were taken weekly prior to and following smoke exposure and twice weekly during smoke events (based on the exposure criteria detailed above), capturing baseline (i.e., before smoke, 3 d), during wildfire smoke exposure (3 d), and after smoke exposure (4 d), which resulted in 10 total sample days. Thus, individual animals served as their own controls. Health scoring was performed according to the University of Wisconsin-Madison Calf Health Scoring system (McGuirk, 2008). This scoring system accounts for various indicators of clinical health parameters and scores them each on a scale of zero (indicating normal status) to three (indicating clinical abnormalities). Parameters observed included nasal secretions (0 = normal or no discharge to 3 = large amount of bilateral discharge), eye secretions (0 = normal or no discharge to 3 = large amount of discharge), ear droopiness/position (0 = normal to 3 = head tilt, two ears drooping), induced and uninduced coughing (0 = neither induced nor spontaneous cough to 3 = repetitive uninduced coughing), fecal consistency (0 = normal to 3 = watery, soaks through shavings), joint (0 = normal, no inflammation to 3 = inflammation, pain, heat) and naval inflammation (0 = normal, no inflammation to 3 = inflammation, pain, heat, discharge). Respiratory rate (breaths/min) was measured at 1200 hours by counting the number of flank movements in one min. Rectal temperatures were recorded at 1230 hours using a calf rectal thermometer. Heart rates were measured at 1230 hours by placing a stethoscope cranially on the left side of the calf behind the elbow and counting the number of heart beats in one min.

Blood sampling

Ten milliliters of blood was collected via jugular venipuncture into sodium-heparin plasma tubes and serum tubes with clot-activator gel. Throughout the study, blood was collected immediately following the morning feeding. Tubes were gently inverted several times immediately after collection, plasma tubes were placed on ice, and serum tubes were allowed to remain undisturbed at room temperature for approximately 30 min. A small volume of whole blood was used for hematology analysis prior to centrifugation at 3,000 × g for 17 min at 24 °C. Serum and plasma samples were aliquoted into Eppendorf tubes and stored at −20 °C until further analysis.

Hematology and acute phase proteins

A VetScan HM5C Hematology System (Abaxis, Union City, CA) was used to measure total and differential cell counts, concentrations of total white blood cell population (WBC), lymphocytes (LYMPH), monocytes (MONO), neutrophils (NEU), eosinophils (EOS), basophils (BAS), red blood cells (RBC), as well as hemoglobin (HGB) and hematocrit (HCT), expressed as a percentage. Following manufacturer procedure, approximately 2 mL of whole blood was placed in the sample receiving chamber, and samples were run using the automated bovine hematology reference intervals provided by the manufacturer.

SAA was measured in calf serum through immuno-absorbance with a SAA multispecies solid phase sandwich ELISA following the manufacturer’s protocol (no. TP-802, Tridelta Development Ltd, Kildare, Ireland). Samples were analyzed in duplicate with a Spectramax ID3 spectrophotometer (Molecular Devices, San Jose, California), and diluted from 1:500 to 1:1,500 as necessary. Inter-assay CV was 12%, and intra-assay CV was 4.8%. Serum Hp concentrations were measured through an in-house kinetic colorimetric assay modified from U.S. patent US6451550B1 and previously utilized by our group (Tsai, 2021) in which Hp is quantified through peroxidase activity as a result of the interaction with methemoglobin. Samples were run in triplicate and absorbance was read at 390 nm every min for 6 min. Intra-assay CV was 8.7%, and inter-assay CV was 2.4%.

Blood metabolites

Blood metabolites, including □-hydroxybutyrate (BHB), glucose, and NEFA, were quantified from a subset of samples from seven dates that spanned changes in PM_2.5 concentrations. Plasma BHB concentration was detected through an in-house assay that uses the enzyme 3-□-dehydrogenase to oxidize 3-BHB to acetoacetate, which concomitantly causes NAD+ to reduce to NADH, whose concentration is proportional to the concentration of 3-BHB (Tsai, 2021). The concentration of NADH was measured with a Spectramax spectrophotometer (Molecular Devices). Following addition of buffer (33 µg/mL 3-□-dehydrogenase (Roche no. 3HBDB-RO), 0.1M Tris, 0.1 M oxalic acid, 2mM EDTA, 2.25 mM NADH) to each sample or standard calibrator (no. 421-73791, FujiFilm Medical Systems, Lexington, MA), plates were incubated at 37 °C for 1 min and then read at 340 nm at 0, 1, and 2 min. NADH concentrations in samples and standard were measured in triplicate. Inter-assay CV was 8.75%, intra-assay CV was 6.32%.

Concentrations of plasma NEFA and glucose were both measured using commercial kits. Glucose concentration was measured in duplicate with a microtiter assay (Fujifilm Autokit Glucose, cat 99703001) following the manufacturer protocol. Inter-assay CV was 14.3% and intra-assay CV was 7.2%. Plasma NEFA concentration was quantified in duplicate utilizing a commercial colorimetric Wako NEFA-HR assay kit (no’s: Color reagent A: 999-34691; Solvent A: 995-34791; Color reagent B: 991-34891; Solvent B: 993-35191; NEFA Standard Solution: 276-76491; FUJIFILM Medical Systems) in accordance with the manufacturer’s protocol. Intra-assay CV was 5.8%, and inter-assay CV was 7.6%.

Statistical analysis

Statistical analysis was performed in SAS version 9.4. Correlation analysis using the CORR procedure was used to determine if PM_2.5, THI, or animal age were correlated. Age and PM_2.5 were not significantly correlated but PM_2.5 and THI were. PROC REG was used to determine bias in estimate coefficients and standard errors due to collinearity among independent variables by assessing tolerance (TOL), variance inflation factors (VIF), and condition indices (COLLIN). TOL values below 0.1 indicate multicollinearity (Schreiber-Gregory and Jackson, 2017). VIF above 10 are indicative of biases in standard errors due to multicollinearity (Myers, 1990) and COLLIN above 30 indicate strong bias whereas values between 5 and 10 indicate weak bias in coefficient estimates due to multicollinearity (Belsley et al., 1980).

General linear mixed models using the PROC MIXED procedure were used for all variables except health scores. Health scores were analyzed with logistic regression models using PROC GENMOD. In all models, fixed effects were PM_2.5, THI, and their interaction, while calf was a random effect. PM_2.5 and THI interactions that were not significant were removed from final models. Separate models with lags up to 7 d were utilized to describe both delayed and persistent effects. Plotted residuals were assessed for normality. Significance was considered when P-values were less than or equal to 0.05, whereas tendencies were considered when P-values were between 0.05 and 0.1. All data are presented as means ± SEM.

Results

Environmental data

PM_2.5 fluctuated with nearby fire conditions for the duration of the study, with a large smoke event occurring in mid-August when daily PM_2.5 concentrations surpassed 115 µg/m³ (Figures 1 and 2). The 24-h average THI ranged from 48 to 73, while PM_2.5 concentrations ranged from 2.58 to 118.8 µg/m³. During our study, the criteria for calf exposure to wildfire PM_2.5 were met on a total of three sample days and there were 10 total sample days, indicated in Figure 1. Thus, there were three calf measurements before, three calf measurements during, and four measurements taken after wildfire-PM_2.5 exposures. PM was moderately positively correlated with THI (r = 0.47, P < 0.001). TOL values were above 0.7, VIF were less than 1.4, and COLLIN were less than 3, indicating little to no bias in coefficient estimates or standard errors due to collinearity between PM_2.5 and THI.

Figure 1.

Average daily PM_2.5 concentrations and THI from July 1 to October 1, 2021, which encompassed the duration of the study period. There were multiple wildfire events over the course of the study period, resulting in spikes in daily average PM_2.5. Purple squares indicate THI and green circles represent PM_2.5. Yellow stars below the x-axis indicate dates in which samples were collected.

Figure 2.

Maps of active wildfires, PM_2.5, and air mass trajectories on two separate days. (A) A day with wildfire smoke present (August 13, 2021), resulting in a 24-h average PM_2.5 concentration of 118.75 μg/m³. (B) A day without wildfire smoke present (August 29, 2021), resulting in 24-h average PM_2.5 concentration of 3.42 μg/m. AirNow-Tech Navigator (https://www.airnowtech.org) mapping system was used to produce the maps, with NOAA HYSPLIT atmospheric transport and dispersion modeling system utilized to visualize air mass trajectories. Air masses are shown as 72-h backward trajectories, at atmospheric heights of 50, 100, and 150 m (green, blue, and red lines, respectively). Colored dots indicate ranges of PM_2.5 concentrations at monitoring stations. HMS indicates Hazard Mapping System, with smoke plumes and wildfire locations displayed. The yellow star depicts the location of the University of Idaho Dairy Center where the research was conducted.

Health parameters

Wildfire smoke PM_2.5 and THI had interacting effects on several measures of health in calves (Figure 3, Table 1). There was a positive interacting effect of PM_2.5 and THI on rectal temperature on lag day 0 (P = 0.04), lag day 6 (P = 0.03), and lag day 7 (P < 0.0001). Elevated PM_2.5 and THI together also increased heart rate on lag day 0 (P < 0.001) and lag day 1 (P < 0.0001), and decreased heart rate on lag day 3 (P < 0.0001), lag day 6 (P < 0.01), and lag day 7 (P < 0.001). Respiratory rate was altered as a result of increased PM_2.5 and THI, increasing on lag day 0 (P = 0.04) but decreasing on lag day 2 (P < 0.01), lag day 3 (P < 0.0001), and lag day 5 (P = 0.05). Additionally, there was a tendency for higher THI and PM_2.5 together to increase respiratory rate on lag day 1 (P = 0.06), heart rate on lag day 4 (P = 0.1), and rectal temperature on lag day 3 (P = 0.09). Higher THI alone increased respiratory rate on lag day 4 (P = 0.0001) and on lag day 6 (P = 0.09). Health scores were affected by the interaction of elevated of PM_2.5 and THI (Table 2). Cough score (P < 0.01) and eye score (P = 0.01) were increased on lag day 3, but eye score decreased on lag day 4 (P = 0.04) with combined elevated PM_2.5 and THI. Additionally, the interaction of increased PM_2.5 and THI had a tendency to increase cough scores on lag day 7 (P = 0.10).

Figure 3.

Interactions between fine particulate matter (PM_2.5) and temperature–humidity index (THI) at lag day 0 on (A) calf respiratory rate (RR), (B) heart rate (HR), and (C) rectal temperature (RT). Preweaned Holstein heifer calves (N = 15) born in July of 2021 were naturally exposed to wildfire smoke multiple times throughout the summer. Respiratory rate was increased as a result of PM_2.5 and THI interactions (P = 0.04). PM_2.5 and THI had a positive interaction on heart rate (P < 0.001). Rectal temperature was increased due to the positive interaction of PM_2.5 and THI (P = 0.04).

Table 1.

Physiological measurements on lag day 0 as related to atmospheric PM_2.5 and THI

Model predictors
Variable 1	Intercept			PM2.5, µg/m3			THI			PM2.5 × THI
	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value
RRa	9.12	15.3	–	−2.17	1	0.03	0.62	0.24	0.01	0.03	0.02	0.04
HRb	157.1	33.9	–	−8.4	2.2	<0.001	−0.38	0.52	0.46	0.12	0.03	<0.001
RTc	39.6	0.37	–	−0.05	0.02	0.06	−0.01	< =0.01	0.18	0.0007	0.0004	0.04

¹RR = Respiratory rate (breaths per min); HR = heart rate (beats/min); RT = rectal temperature (°C).

^aElevated PM₂.₅ and THI also interacted to decrease RR on lag day 2 (P < 0.01), lag day 3 (P < 0.0001), and lag day 5 (P = 0.05). Higher THI alone increased RR on lag day 4 (P = 0.0001).

^bElevated PM₂.5 and THI also interacted to increase HR on lag day 1 and decrease HR on lag day 3 (both P < 0.0001), lag day 6 (P < 0.01), and lag day 7 (P < 0.001).

^cElevated PM₂.₅ and THI also interacted to increase RT on lag day 6 (P = 0.03) and lag day 7 (P < 0.0001).

Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to several wildfire smoke events across the summer. Data were analyzed using general linear mixed models with lags of up to 7 d, with PM_2.5, THI, and their interaction as fixed effects, and calf as a random effect. Coefficients, standard error of the means (SEM), and P-values from statistical models are provided.

Table 2.

Calf health scores on lag day 3 in relation to atmospheric PM_2.5 and THI

Model predictors
Variable 1	Intercept 2			PM2.5,µg/m3			THI			PM2.5 × THI
	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value
Cough	2.84, 5,73, 7,38	3.78, 3.63, 3.60	–	−0.62	0.19	0.001	−0.12	0.06	0.001	0.01	0.003	0.001
Eye	14.72, 17.1	4.16, 4.14	–	−0.72	0.30	0.02	−0.29	0.06	<0.0001	0.01	0.004	0.01

¹Cough = cough score; Eye = eye score.

²For cough scores, there were three intercepts, and for eye scores there were two intercepts.

Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to several wildfire smoke events across the summer. Data were analyzed using general linear mixed models with lags of up to 7 d, with PM_2.5, THI, and their interaction as fixed effects, and calf as a random effect. Coefficients, standard error of the means (SEM), and P-values from statistical models are provided.

Hematology and acute phase proteins

Multiple hematological variables were affected by exposure to increased PM_2.5 and THI, most frequently on lag day 3 (Figure 4, Table 3). PM_2.5 and THI together decreased WBC (P = 0.01) and NEU counts (P < 0.01) on lag day 3. Together, PM_2.5 and THI interacted to decrease LYMPH on lag day 2 (P = 0.03), and increased EOS on lag day 0 (P = 0.02). We detected no effect of THI, PM_2.5, or their interaction on BAS or MONO. Elevated THI and PM_2.5 interacted to increase HCT on lag day 1 (P < 0.01), and to decrease HCT on lag day 3 (P < 0.0001) and lag day 5 (P = 0.03). HGB was also decreased on lag day 3 (P < 0.0001) and lag day 5 (P < 0.01) as a result of the interaction of elevated PM_2.5 and THI. Together, elevated PM₂.₅ and THI also slightly increased NEU (P = 0.06) and RBC (P = 0.07) on lag day 1 and decreased EOS on lag day 3 (P = 0.08).

Figure 4.

Interactions between PM_2.5 and THI at lag day 3 on (A) total white blood cell count, (B) neutrophil count, (C) hemoglobin concentration, and (D) hematocrit. Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to wildfire smoke multiple times throughout the summer. Total white blood cell count was reduced after exposure to elevated THI and PM_2.5 together (P = 0.01). PM_2.5 and THI interacted to decrease neutrophil count (P < 0.01). Hemoglobin concentration was decreased as a result of the interaction of PM_2.5 and THI on lag day 3 (P < 0.0001). Combined increased PM_2.5 and THI resulted in lower hematocrit (P < 0.0001).

Table 3.

Calf blood hematology and inflammatory markers on lag day 3 in relation to atmospheric PM_2.5 and THI

Model predictors
Variable 1	Intercept			PM2.5, µg/m3			THI			PM2.5 × THI
	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value
WBC	−3.08	4.85	–	0.71	0.28	0.01	0.2	0.08	0.01	−0.01	0.004	0.01
EOSa	−0.1	0.11	–	0.01	0.01	0.1	0.003	0.002	0.12	−0.0002	0.0001	0.08
BAS	0.02	0.05	–	−0.0002	0.0002	0.24	0.0004	0.0008	0.06	–	–	–
LYMPHb	6.25	2.98	–	0.24	0.17	0.17	0.002	0.05	0.96	−0.003	0.003	0.17
NEUc	−10.68	3.09	–	0.49	0.18	0.01	0.21	0.05	< 0.0001	−0.01	0.003	0.005
MONO	0.99	0.34	–	−0.0003	0.001	0.74	−0.004	0.01	0.49	–	–	–
RBCd	3.23	2.69	–	0.19	0.15	0.22	0.09	0.04	0.03	−0.003	0.002	0.19
HCT, %e	−46.6	7.12	–	3.47	0.41	<0.0001	1.17	0.11	<0.0001	−0.05	0.01	< 0.0001
HGBf	−9.45	1.57	–	0.86	0.09	< 0.0001	0.28	0.02	< 0.0001	−0.01	0.001	< 0.0001

¹. Units for WBC, EOS, BAS, LYMPH, NEU, MONO are 10⁹ cells/L; RBC is 10¹² cells/L; HCT is %, and HGB is g/dL.

^aElevated PM_2.5 and THI also interacted to increase EOS on lag day 0 (P = 0.02). Higher THI alone decreased EOS on lag day 1 (P = 0.04).

^bElevated PM₂.5 and THI also interacted to decrease LYMPH on lag day 2 (P = 0.03).

^cElevated THI alone also increased NEU on lag day 0 (P < 0.01).

^dElevated THI alone also increased RBC on lag day 3 (P = 0.03).

^eElevated PM₂.5 and THI also interacted to decrease HCT on lag day 1 (P < 0.01) and lag day 5 (P = 0.03). THI alone increased HCT on lag day 0, lag day 2 (both P < 0.0001), lag day 4 (P = 0.03), and lag day 6 (P < 0.001).

^fElevated PM₂.5 and THI also interacted to decrease HGB on lag day 5 (P < 0.01). THI alone increased HGB on lag day 0 and lag day 2 (both P < 0.0001), lag day 1 (P < 0.01), lag day 4 (P = 0.01), and lag day 6 (P < 0.01).

Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to several wildfire smoke events across the summer. Data were analyzed using general linear mixed models with lags of up to 7 d, with PM_2.5, THI, and their interaction as fixed effects, and calf as a random effect. Coefficients, standard error of the means (SEM), and P-values from statistical models are provided.

While not a main objective of the present study, we observed THI alone had an effect on several hematological variables. Neutrophil counts were increased by higher THI on lag day 0 (P < 0.01), while EOS were decreased on lag day 1 (P = 0.04). THI had a positive effect on HCT and HGB on lag days 0, 2, 4, and 6 (all P < 0.04). Hemoglobin concentration was also increased by higher THI on lag day 1 (P < 0.01). Red blood cells were increased by elevated THI on lag day 3 (P = 0.03). Additionally, THI tended to increase NEU on lag day 2 (P = 0.07) and lag day 5 (P = 0.06) and decrease EOS on lag day 6 (P = 0.06).

Higher PM_2.5 and THI together changed circulating SAA and Hp concentrations (Table 4). On lag days 0, 2, 3, and 4, SAA was decreased while Hp was increased (all P < 0.05) with higher combined PM_2.5 and THI. However, on lag day 1, SAA was increased while Hp was decreased (both P < 0.01) with higher combined PM_2.5 and THI. Additionally, elevated THI and PM_2.5 had a positive interactive effect on SAA on lag day 7 (P < 0.001), and a negative interactive effect on Hp on lag day 5 (P < 0.01). Hp was decreased with increasing THI alone on lag day 7 (P = 0.04).

Table 4.

Calf acute phase proteins on lag days 0, 1, and 2 in relation to atmospheric PM_2.5 and THI

Model predictors
Variable	Intercept			PM2.5, µg/m3			THI			PM2.5 × THI
	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value
Lag day 0
Hp a	7.65	1.98	–	−1.94	0.61	<0.01	−0.08	0.03	<0.01	0.03	0.01	<0.01
SAA	−2,441,606	810,682	–	567,303	249,373	0.03	37,914	11,028	0.001	−8,104.37	3,541.01	0.03
Lag day 1
Hp	−37.37	14.06	–	4.09	1.47	0.01	0.58	0.21	<0.01	−0.06	0.02	<0.01
SAA	15,249,213	5,764,383	–	−1,644,146	604,162	0.01	−222,767	85,609	0.01	23,988	8,821.01	0.01
Lag dAY 2
Hp	41.71	13.44	–	−8.91	2.99	<0.01	−0.59	0.20	<0.01	0.132	0.04	<0.01
SAA	−1.2E + 07	5,493,039	–	2,561,002	1,223,081	0.04	183,978	80,870	0.03	−37,976	18,129	0.04

^aElevated PM_2.5 and THI together also increased Hp on lag day 3 (P < 0.01) and lag day 4 (P < 0.03), and decreased Hp on lag day 5 (P < 0.01). THI alone decreased Hp on lag day 7 (P < 0.040).

^bHigher PM₂.₅ and THI together also decreased SAA on lag day 3 (P = 0.001) and lag day 4 (P < 0.0001), and increased SAA on lag day 7 (P < 0.001).

Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to several wildfire smoke events across the summer. Data were analyzed using general linear mixed models with lags of up to 7 d, with PM_2.5, THI, and their interaction as fixed effects, and calf as a random effect. Coefficients, standard error of the means (SEM), and P-values from statistical models are provided.

Blood metabolites

Elevated PM_2.5 and THI had an interactive effect on blood glucose, BHB, and NEFA concentrations, largely on lag days 1, 3, and 6 (Table 5). The effect on each metabolite on lag day 1 was opposite of that observed on subsequent lag days. Glucose was increased on lag day 1 (P < 0.01) and decreased on lag day 3 (P = 0.02) and lag day 6 (P < 0.001) with combined elevated PM_2.5 and THI. Also, there was a negative interacting effect of PM_2.5 and THI on BHB on lag day 1 (P < 0.001) and a positive interacting effect on lag day 3 (P < 0.001), day 6 (P < 0.0001), and day 7 (P < 0.01). Additionally, NEFA was increased on lag day 1 (P = 0.05) and decreased on lag day 3 (P < 0.001), day 6 (P < 0.0001), and day 7 (P = 0.03) when PM_2.5 and THI together were increased.

Table 5.

Blood metabolites on lag days 1, 3, and 6 in relation to atmospheric fine particulate matter (PM_2.5) and temperature–humidity index (THI)

Model predictors
Variable 1	Intercept			PM2.5, µg/m3			THI			PM2.5 × THI
	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value	𝛽	SEM	P-value
Lag day 1
Glucosea	140.6	61.4	–	−16.3	5.23	0.002	−0.38	0.91	0.68	0.24	0.076	<0.01
BHBb	−54.3	137.7	–	40.9	11.7	<0.001	3.67	2.03	0.07	−0.6	0.17	<0.001
NEFAc	0.29	0.19	–	−0.033	0.02	0.04	−0.002	0.003	0.44	0.0005	0.0002	0.048
Lag day 3
Glucose	−338.2	97.2	–	10.9	4.67	0.02	6.9	1.5	<0.0001	−0.163	0.07	0.02
BHB	1140.7	211.4	–	−39.5	10.1	<0.001	−14.4	3.26	<0.0001	0.58	0.15	<0.001
NEFA	−0.88	0.29	–	0.049	0.014	<0.001	0.015	0.004	<0.001	–0.0007	0.0002	<0.001
Lag day 6
Glucose	−282.6	92.3	–	42.7	12.2	< 0.001	5.65	1.41	0.0001	−0.59	0.17	< 0.001
BHB	1284.3	187.2	–	−132.8	24.7	< 0.0001	−15.7	2.85	< 0.0001	1.87	0.34	< 0.0001
NEFA	−1.16	0.26	–	0.16	0.034	< 0.0001	0.019	0.004	< 0.0001	−0.002	0.0005	< 0.0001

¹Glucose, mg/dL; BHB, □–hydroxybutyrate, µM; NEFA, nonesterified fatty acids, mEq/L.

^aHigher THI alone increased glucose on lag day 0 (P < 0.001), lag day 2 (P < 0.0001), lag day 5 (P < 0.01), and lag day 7 (P < 0.001).

^bHigher PM₂.5 and THI also interacted to increase BHB on lag day 7 (P < 0.01). Higher THI alone decreased BHB on lag day 2 (P < 0.01) and lag day 5 (P = 0.04).

^cHigher PM₂.₅ and THI also interacted to decrease NEFA on lag day 7 (P = 0.03). Higher THI alone also increased NEFA on lag day 2 (P < 0.01) and lag day 5 (P = 0.04).

Preweaned Holstein heifer calves (N = 15) born in July 2021 were naturally exposed to several wildfire smoke events across the summer. Data were analyzed using general linear mixed models with lags of up to 7 d, with PM_2.5, THI, and their interaction as fixed effects, and calf as a random effect. Coefficients, standard error of the means (SEM), and P-values from statistical models are provided.

Higher THI alone affected blood metabolites on several days. Glucose increased with THI on lag day 0, (P < 0.001), day 2 (P < 0.0001), day 5 (P < 0.01), and day 7 (P < 0.001). BHB was decreased as a result of increased THI on lag day 2 (P < 0.01) and lag day 5 (P = 0.04), while NEFA was increased with higher THI on lag day 2 (P < 0.01) and on lag day 5 (P = 0.04). There was also a tendency for THI alone to increase glucose on lag day 4 (P = 0.09) and decrease BHB on lag day 0 (P = 0.08) and lag day 4 (P = 0.1).

Discussion

Wildfires cause escalations in atmospheric PM_2.5 that far exceed the threshold for hazardous exposure in humans. Further, projection models show that wildfires and as a consequence, air pollution associated with wildfire smoke, will continue to worsen over the next several decades (Littell et al., 2010). Wildfire-derived PM_2.5 inhalation in humans is linked to respiratory morbidities (Liu et al., 2015; Hutchinson et al., 2018; Borchers Arriagada et al., 2019; Kiser et al., 2020), cardiovascular disease (Dennekamp et al., 2015; Kollanus et al., 2016; Navarro et al., 2019; Chen et al., 2021b), and mortality (Doubleday et al., 2020; Liu et al., 2021; Chen et al., 2021a; Ye et al., 2022). Recently, we reported that lactating dairy cows naturally exposed to wildfire smoke, which produced daily average PM_2.5 concentrations reaching roughly 300 µg/m³, experienced changes in milk production and metabolism and had an innate immune response (Anderson et al., 2022). However, impacts of wildfire smoke on calves had not yet been investigated. Herein, we evaluated preweaned Holstein heifer calves naturally exposed to wildfire smoke to explore effects on calf health and performance.

Physiological measurements and health scores

In the present study, we found that when both wildfire PM_2.5 and THI increased, there was a concomitant increase in calf heart rate, respiratory rate, and rectal temperature. This supports a wide body of literature on the effects of wildfire smoke exposure on human cardiac and pulmonary systems. For example, heart rhythm disturbances and higher heart rate in humans have been previously associated with PM_2.5 from wood smoke or wildfire smoke (Unosson et al., 2013; Le et al., 2014). Additionally, heat stressed calves have higher heart rates, respiratory rates, and rectal temperatures (Kovács et al., 2018; Dado-Senn et al., 2020). The thermoregulatory role of respiration may pose increased risk to calves that are subject to both higher temperatures and pollution from smoke as there may be an increase in rate of inhalation of PM. A similar observation has been made in rats exposed to pollutants while exercising (Mautz, 2003). Thus, animals concomitantly exposed to high THI resulting in elevated body temperature and higher respiratory rate, could be more vulnerable to toxic compounds in wildfire smoke.

Both eye scores and cough scores were increased 3 d after calf exposure to elevated wildfire PM_2.5 and THI together. Ocular discharge following exposure could indicate irritation of the eye, respiratory tract, or both. Indeed, eye irritation has previously been reported in children exposed to wildfire smoke (Künzli et al., 2006). Coughing assists in the removal of particulates from the airways (Ackermann et al., 2010). Additionally, in the respiratory tract, mucociliary clearance and alveolar macrophages, both of which are involved in pulmonary particulate clearance (Ackermann et al., 2010), may be damaged by various smoke components, including PM (Ferreira-Ceccato et al., 2011; Franzi et al., 2011; Hamon et al., 2018). Thus, animals may cough in effort to assist compromised clearance mechanisms in removal of particulates and products of inflammation from their airways. Environmental pollutants, such as PM from dust on farms, are also related to incidence of BRD and pulmonary consolidation in calves (Dubrovsky et al., 2019; van Leenen et al., 2021). Although not evaluated in the present study, inflammation and compromised integrity of pulmonary tissues from PM could alter the pleural environment in a manner such that opportunistic pathogens could cause secondary bacterial infections. This etiology could also result in the clinical signs we observed such as coughing, ocular discharge, and elevated rectal temperatures.

Hematology

In the present study, we observed delayed reductions in several white blood cell populations in systemic circulation. There was a negative interacting effect of elevated wildfire PM_2.5 and THI on LYMPH after 2 d, and on WBC and NEU after 3 d. These findings are similar to those described in lactating dairy cows, which had lower lymphocytes, neutrophils, and total white blood cell counts following exposure to higher PM_2.5 and THI (Anderson et al., 2022). Previous studies using other animal models have reported similar reductions in leukocytes in blood including lower total white blood cells and lymphocytes of mice exposed to smoke-derived PM (Kim et al., 2018), as well as delayed reductions in lymphocytes in the lung (Ramos et al., 2021) following exposure to PM.

Leukocyte recruitment to the respiratory tract could explain the overall reduced white blood cell counts in circulation. For example, in cattle, cells such as alveolar macrophages release chemokines after being activated by cytokines, which allows them to recruit leukocytes to the lung in response to pulmonary insult (Ackermann et al., 2010). Inflammatory responses related to secondary infections by opportunistic pathogens, or even pathogens carried by inhaled particulates, could also explain the delayed reduction in lymphocytes and neutrophils. In addition, stress caused by exposure to smoke may induce an elevation in circulating glucocorticoids, which can result in lymphocyte depletion as well as impairment of neutrophil function (Roth and Kaeberle, 1982; Latimer, 2011); however, circulating glucocorticoid concentration was not measured in our study.

We also observed an increase in eosinophils on the day of exposure to higher PM_2.5 and THI together. Bovine eosinophilia is typically associated with parasitic and hypersensitivity responses, such as in Th2 allergic response (Ackermann et al., 2010). Because a parasite control program was employed on farm and the response we observed was relatively immediate and transient, it is likely that the increase in circulating eosinophils was a result of a hypersensitivity response. Elevated eosinophils following natural exposure to wildfire smoke has been previously described in dolphins (Venn-Watson et al., 2013) and in lactating dairy cows (Anderson et al., 2022).

As a result of exposure to combined higher PM_2.5 and THI, we found a transient increase of HCT and subsequently, a reduction in both HCT and HGB. It is unclear whether the initial polycythemia is relative or absolute, but could suggest reduced hydration in the calves, where there is loss of fluid volume resulting in relative polycythemia (Latimer, 2011; Roland et al., 2014). Reduced hydration impacts optimal mucociliary clearance, which can leave calves susceptible to amassing both pathogens and PM (Ackermann et al., 2010). However, higher hematocrit also could indicate increased oxygenation capacity of the blood, which could be a response to greater oxygenation demands of extravascular tissue or hypoxia. In this case, absolute polycythemia could be caused by higher metabolic rate, immune demands, or compromised respiratory function (Jones and Allison, 2007).

The subsequent reduction in HGB and HCT could be a reflection of the lower red blood cells per fluid volume proportionally decreasing hemoglobin concentrations; however, in the present study red blood cell counts were not reduced in calves during wildfire smoke exposure. Lower HGB and HCT could be indicative of greater water consumption, which could result in increased plasma volume (Roland et al., 2014). Alternatively, reduced HGB and HCT could indicate toxin-induced red blood cell lysis, which can cause hemolytic anemia, characterized by a reduction in hemoglobin and red blood cells in cattle (Roland et al., 2014). However, in the present study, clinical diagnosis of anemia was not confirmed. Anemia and reduced hemoglobin have been previously described in relation to ambient air pollution exposures including PM of various sizes (PM₁₀, PM_2.5, and PM_1.0) in humans (Nikolić et al., 2008; Morales-Ancajima et al., 2019; Elbarbary et al., 2020). Another study discovered a connection between anemia and pneumonia in children exposed to air pollution, indicating a possible role of anemia in respiratory outcomes during pollution exposure (Harris et al., 2011). Regardless of the etiology, decreased hemoglobin could indicate reduced oxygenation capacity to extravascular tissues.

In addition to PM_2.5, wildfire smoke contains PM₁₀ that may also contribute to inflammatory responses in calves, including alterations in immune cell populations. For example, non-wildfire sources of PM₁₀ in calf barns is associated with changes in neutrophil and epithelial cell percentages in broncho-alveolar fluid from dairy and beef calves (van Leenen et al., 2021). Although most of the PM fraction in wildfire smoke consists of PM_2.5 (Vicente et al., 2013), exploration of other components of wildfire smoke in relation to calf health would be a worthy avenue of future research.

Blood metabolites

We observed a shift in blood metabolites following wildfire smoke and elevated THI exposure, characterized by an initial increase in circulating glucose and NEFA concentrations concomitant with decreased BHB, followed by a reduction in glucose and NEFA, and higher BHB concentrations across subsequent days. This shift in blood metabolites indicates changes in utilization of available energy sources when exposed to elevated THI and PM_2.5 together. Similarly, NEFA was increased in lactating dairy cows exposed to combined elevated PM_2.5 and THI (Anderson et al., 2022). Changes in metabolism may result from reduced feed intake, however, we did not observe a reduction in milk intake during exposure. Thus, the metabolic responses observed are likely a reflection of increased energy utilization as a result of physiological demands, such as mounting an immune response.

Lower blood glucose following exposure to smoke sources has been previously described in the literature. Rats exposed to peat smoke, and subsequently given a high-fat oral gavage, had reduced blood glucose concentration, suggesting that inhalation of smoke could induce dysregulation of glucose metabolism (Martin et al., 2018). Additionally, a variety of stressors, such as environmental changes, increased energy demands, or inflammation could directly influence nutrient metabolism. For instance, catecholamines and glucocorticoids, which can be elevated in response to environmental stressors such as heat stress (Farooq et al., 2010; W.S. Kim et al., 2018), respectively can induce lipolysis and thereby NEFA release, as well as gluconeogenesis, in cattle (Herdt, 2000; Hammon et al., 2012). Higher circulating pro-inflammatory cytokines are also linked to lipid metabolism, including elevated circulating NEFA concentrations (Kushibiki et al., 2002). Further research is needed to parse out the direct effects of exposure to concomitantly higher PM_2.5 and THI on feed intake and nutrient metabolism.

Acute phase proteins

Circulating Hp and SAA concentrations changed in opposite directions in response to calf exposure to combined elevated THI and PM_2.5. Hp and SAA are proteins involved in the acute phase response, which is part of the inflammatory response (Alsemgeest et al., 1994). An acute phase response following wildfire smoke exposure has been previously reported in humans and rats (Ferguson et al., 2016; Martin et al., 2018), as well as in rhesus macaques exposed in utero (Capitanio et al., 2022). While both Hp and SAA are positive acute phase proteins (Tothova et al., 2014), the oppositional changes in circulating Hp and SAA in calves in the present study suggest different roles in their responses to elevated PM_2.5 and THI. For example, Hp released from the liver and extrahepatic tissues, including the lung, binds to free HGB followed by clearance of Hp/HGB complexes (Yang et al., 1995), which may explain, in part, the reduced HGB detected in calves after exposure to elevated PM_2.5 and THI. This response can stymie oxidative damage produced by free hemoglobin (Yang et al., 2003).

SAA is involved in inflammatory responses by binding to lipids (Uhlar and Whitehead, 1999). Its involvement in lipid metabolism is particularly noteworthy here as we found that SAA was elevated simultaneously with higher NEFA on lag day 1. Guzelbektes et al. (2010) reported similar parallel elevations in both SAA and NEFA, which has been linked to the inflammatory mediator, TNF-α, response to lipid mobilization from adipocytes associated with negative energy balance in fresh cows (Ametaj et al., 2005). Additionally, because SAA can be used as an indicator of pneumonia (Ackermann et al., 2010; Joshi et al., 2018) and particularly in acute inflammation (Alsemgeest et al., 1994), the delayed elevation in SAA on lag day 7 could be the result of a secondary inflammatory response, which is also supported by our observations of delayed rectal temperature elevation that occurred at the same time as the increase in circulating SAA concentrations.

Conclusion

With the threat of wildfires continuing into the foreseeable future, the effects of hazardous emissions in smoke on production animal health are a growing welfare concern. Unlike humans, cattle are less likely to have access to shelter with ventilation that allows for protection from inhalation of smoke. Thus, many of the calves in the top states in dairy production may be at increased risk of exposure to wildfire smoke for the duration of a wildfire event. To our knowledge, this is the first study to investigate the effects wildfire smoke exposure has on preweaned calves. Our data indicate that calves experienced symptoms of respiratory irritation and clinical respiratory signs such as coughing and ocular discharge. This was combined with alterations in leukocyte counts and an acute phase response, indicative of a systemic innate immune response. Additionally, calf blood metabolites shifted, suggesting changes in energy metabolism that may be associated with feed intake and/or increased energy requirements to support an immune response. Although we found delayed and persistent effects of wildfire smoke exposure on metabolism and the immune system, long-term effects and links between these physiological changes and calf morbidity and mortality have yet to be investigated..

Glossary

Abbreviations:

BAS

basophils

BHB

β-hydroxybutyrate

EOS

eosinophils

HCT

hematocrit

HGB

hemoglobin

Hp

β-hydroxybutyrate

LYMPH

lymphocytes

MONO

monocytes

NEFA

β-hydroxybutyrate

NEU

neutrophils

NOAA

National Oceanic and Atmospheric Administration

PM

particulate matter

PM_2.5

fine particulate matter

RBC

red blood cells

SAA

β-hydroxybutyrate

THI

β-hydroxybutyrate

WBC

white blood cells

Conflict of interest statement

The authors declare no real or perceived conflicts of interest.