White adipose tissue remodeling in Little Brown Myotis (Myotis lucifugus) with white-nose syndrome

Abstract

White-nose syndrome (WNS) is a fungal wildlife disease of bats that has caused precipitous declines in certain Nearctic bat species. A key driver of mortality is premature exhaustion of fat reserves, primarily white adipose tissue (WAT), that bats rely on to meet their metabolic needs during winter. However, the pathophysiological and metabolic effects of WNS have remained ill-defined. To elucidate metabolic mechanisms associated with WNS mortality, we infected a WNS susceptible species, the Little Brown Myotis (Myotis lucifugus), with Pseudogymnoascus destructans (Pd) and collected WAT biopsies for histology and targeted lipidomics. These results were compared to the WNS-resistant Big Brown Bat (Eptesicus fuscus). A similar distribution in broad lipid class was observed in both species, with total WAT primarily consisting of triacylglycerides. Baseline differences in WAT chemical composition between species showed that higher glycerophospholipids (GPs) levels in E. fuscus were dominated by unsaturated or monounsaturated moieties and n-6 (18:2, 20:2, 20:3, 20:4) fatty acids. Conversely, higher GP levels in M. lucifugus WAT were primarily compounds containing n-3 (20:5 and 22:5) fatty acids. Following Pd-infection, we found that perturbation to WAT reserves occurs in M. lucifugus, but not in the resistant E. fuscus. A total of 66 GPs (primarily glycerophosphocholines and glycerophosphoethanolamines) were higher in Pd-infected M. lucifugus, indicating perturbation to the WAT structural component. In addition to changes in lipid chemistry, smaller adipocyte sizes and increased extracellular matrix deposition was observed in Pd-infected M. lucifugus. This is the first study to describe WAT GP composition of bats with different susceptibilities to WNS and highlights that recovery from WNS may require repair from adipose remodeling in addition to replenishing depot fat during spring emergence.

1. Introduction

White-nose syndrome (WNS) is a disease of bats caused by the psychrophilic fungus Pseudogymnoascus destructans (Pd) (Lorch et al., 2011). It is of particular concern in North America, where it has caused a precipitous decline in certain cave-dwelling species (Hoyt et al., 2021). Since its first detection in 2006 (Blehert et al., 2009), subsequent studies have shown that not all Nearctic bat species are equally susceptible to WNS mortality. In the Midwestern U.S. for example, drastic declines have been observed in Myotis species such as the Little Brown Myotis (Myotis lucifugus) (> 80% in Indiana (Pettit & O’Keefe 2017), the Great Lakes Region (Mallinger et al., 2023), and Michigan (Kurta & Smith, 2020) and Northern long-eared Myotis (Myotis septentrionalis) (> 90% in Michigan (Kurta & Smith, 2020), the Great Lakes Region (Mallinger et al., 2023), and Arkansas (Perry & Jordan, 2022). Declines in other species, including Indiana Myotis (Myotis sodalis) (Thogmartin et al., 2013) and Tricolored Bats (Perimyotis subflavus) (Smith et al., 2023), are also of concern, however, other species such as the Big Brown Bat (Eptesicus fuscus) and Virginia Big-eared Bat (Corynorhinus townsendii virginianus) have not experienced the same declines. Within species of concern, intraspecific variation in survival has been observed in populations of M. lucifugus (Langwig et al., 2017) where increased fat stores were correlated with increased survival (Cheng et al., 2019). Understanding the mechanisms and environmental traits driving this variability to WNS-associated mortality among different bat species has been of great interest, as limited fiscal resources are available for management strategies (e.g., habitat protection, hibernacula microclimate manipulation, site treatment) to mitigate the effects of WNS. In addition, elucidating mechanisms of resistance associated with natural bat populations may offer insights into the survivors and potential recovery of species currently experiencing heavy declines (Auteri & Knowles, 2020; Grimaudo et al., 2022).

To study the host mechanisms driving disease susceptibility, the effects of Pd-infection can be categorized into both cutaneous effects at the site of fungal infection and systemic effects from altered torpor patterns (Reeder et al., 2012; Lilley et al., 2016) and increased energy expenditure (Warnecke et al., 2013). As bats primarily rely on stored lipid reserves and fatty acid β oxidation to fuel their torpor and rewarming bouts (Dark, 2005), a detailed understanding of fat deposition and altered lipid metabolism is inextricably linked to understanding the WNS process that is needed to frame effective management responses. At early stages of WNS, in addition to changes in blood chemistry (Verant et al., 2014), perturbations to lipid metabolism are occurring before emaciation is evident (Pannkuk et al., 2021, 2023). In uninfected M. lucifugus and E. fuscus from northern populations (Ontario, Canada) during hibernation (mid-January), female bats showed higher hepatic triacylglyceride (TAG) levels compared to males that possibly reflects resource conservation needed for increased spring fecundity (Pannkuk et al., 2021). However, after laboratory Pd-infection, a non sex-specific increase in some glycerophospholipids (GPs) was observed in M. lucifugus, along with a decrease in female TAG levels in both species, while changes in M. lucifugus were confined to n-3 TAGs (Pannkuk et al., 2023). Although the liver may partially function in lipid storage, hepatic lipid levels are reflective of lipid “shuttling” as it is a primary site for transportation, synthesis, and catabolism (Alves-Bezerra & Cohen, 2017). Observation of these early responses to Pd-infection motivated further research on the same cohort, which included detailing splenic lipid metabolism as it is a primary site in immune modulation and T and B cell differentiation (Pannkuk et al., 2023). Rather than exhibiting immune-specific responses, WNS-susceptible bats (M. lucifugus) showed increased oxidative stress that may play a role in downstream deterioration to body condition. Upregulated cutaneous transcription of genes involved in regulating lipid metabolism are also observed in euthermic versus torpid bats and at more advanced stages of WNS (Field et al., 2018). What remains to be determined is (1) how the quality of adipose fat stores may vary among species and regions (e.g., northern vs. southern latitudes) and play a role in survival in addition to total stored fat (Cheng et al., 2019), and (2) what pathological changes may occur within WAT during increased torpor arousals.

In this study, we expand on a previous work that explored physiological variations in two bat species with different susceptibilities to WNS, M. lucifugus and E. fuscus, after being laboratory-infected with Pd and held in identical microclimates (Moore et al., 2018). Specifically, as premature exhaustion of stored lipid reserves is a primary contributor to bat mortality, we utilized a targeted lipidomics approach to determine (1) differences in WAT lipid composition between M. lucifugus and E. fuscus without Pd-infection, and (2) changes to the WAT lipid composition in both M. lucifugus and E. fuscus between a control group and a Pd-infection group. We predicted broad species level differences in the WAT chemical composition between M. lucifugus and E. fuscus due to dietary differences. Also, we predicted higher perturbation to lipid reserves in M. lucifugus compared to E. fuscus after Pd-infection. This is the first study to describe WAT GP composition of bats with different susceptibilities to WNS and highlights that increased perturbation to adipose metabolism is occurring in M. lucifugus compared to E. fuscus.

2. Materials and methods

2.1. Animal experiment and histology

The WAT samples used in this study were collected as part of a captive hibernation experiment performed at Bucknell University from November 2011 to March 2013 (Moore et al., 2018). As shown in Table 1 (and Supplementary Table 1), the bats were captured between 5 November 2011 and 17 November 2011 in hibernacula in Michigan or Illinois, USA (M. lucifugus) and in Iowa or Illinois, USA (E. fuscus), as previously described (Moore et al., 2018). At the time of capture, none of these populations was known to have previous exposure to WNS. The bats were transported while torpid to Bucknell University in Lewisburg, Pennsylvania, USA where they were treated with either phosphate-buffered saline with 0.5% TWEEN-20 (control) or with a suspension of 350,000 Pd conidia (Pd exposed). Control and treated bats were separately housed in environmental chambers at 4 °C and 95% relative humidity with each species in separate enclosures within the chamber. Following 3, 7, or 13 weeks of hibernation, bats were removed from the chamber and humanely euthanized by isoflurane overdose followed by decapitation. For the current study, the analysis was based on disease severity rather than hibernation time and time points were combined. The WAT was dissected from between the pelvis and the skin on both sides of the back, among other tissues, and snap-frozen in liquid nitrogen. Tissues were stored at −80 °C to minimize enzymatic activity until they were processed for analysis. All samples were stored under identical conditions, so any changes in lipid composition should be comparable across groups. Capture, handling, and sample collection protocols for this study were reviewed and approved by the Bucknell University IACUC (protocol #DMR-12), and the US Fish and Wildlife Service. In the states of Illinois and Iowa, research collections were conducted by state wildlife officials and on non-endangered bats; thus, numbered permits were not required or issued. In Michigan, research was conducted under Scientific Collector’s Permits #SC1448 to DMR.

Table 1.

Species, sex, Pd-infection status, and week euthanized post-Pd exposure for white adipose biopsies used in the current study

Species	Sex	Pd-infection status	Week
Myotis lucifugus	female	control	3
Myotis lucifugus	female	control	3
Myotis lucifugus	female	control	3
Myotis lucifugus	female	control	3
Myotis lucifugus	female	control	3
Myotis lucifugus	male	control	3
Myotis lucifugus	male	control	3
Myotis lucifugus	male	control	7
Myotis lucifugus	female	control	7
Myotis lucifugus	male	control	7
Myotis lucifugus	male	control	7
Myotis lucifugus	male	control	7
Myotis lucifugus	male	control	7
Myotis lucifugus	female	control	13
Myotis lucifugus	female	Pd-exposed	3
Myotis lucifugus	female	Pd-exposed	7
Myotis lucifugus	male	Pd-exposed	7
Myotis lucifugus	female	Pd-exposed	7
Myotis lucifugus	female	Pd-exposed	7
Myotis lucifugus	female	Pd-exposed	7
Myotis lucifugus	male	Pd-exposed	7
Myotis lucifugus	male	Pd-exposed	13
Myotis lucifugus	male	Pd-exposed	13
Myotis lucifugus	female	Pd-exposed	13
Myotis lucifugus	male	Pd-exposed	13
Eptesicus fuscus	female	control	3
Eptesicus fuscus	male	control	3
Eptesicus fuscus	female	control	7
Eptesicus fuscus	female	control	7
Eptesicus fuscus	male	control	7
Eptesicus fuscus	female	control	13
Eptesicus fuscus	male	Pd-exposed	3
Eptesicus fuscus	female	Pd-exposed	3
Eptesicus fuscus	female	Pd-exposed	3
Eptesicus fuscus	male	Pd-exposed	7
Eptesicus fuscus	female	Pd-exposed	7

2.2. Microscopy and histology

Flash-frozen WAT samples were immediately stored at −80 °C until fixation. Samples were kept on ice for 20 min and then placed in a tube with a 1:10 volume of fixative-1 solution (20% formaldehyde, 2% glutaraldehyde in phosphate buffer saline pH 7.4) at room temperature for 30 min. Finally, samples were transferred to a 10% formalin solution and incubated for 2–3 h at room temperature for final fixation, then transferred into 70% ethanol followed by paraffin embedding. Formalin-fixed paraffin-embedded sections were used for hematoxylin and eosin (H&E) staining using the standard protocol. The BX-63 microscope mounted with DP 28 digital camera (Olympus, MA, USA) was used to perform light microscopy on H&E-stained adult male bat WAT sections (M. lucifugus control n = 6 and Pd-infected n = 5). Digital images were acquired using cellSens Entry v1.15 (Olympus, MA, USA) software at a fixed setting for both control and infected groups. Histological assessment of adipocyte size was estimated using the Adiposoft v1.16 plug-in for ImageJ (Galarraga et al., 2012). The automated settings (min size = 5, max size = 60) were used, and then the output was checked manually. We compared the average adipocyte cell sizes using an unpaired t-test in GraphPad Prism 9.2.0 (GraphPad Software, La Jolla, CA).

2.3. Lipid extraction and profiling

Fisher Optima^™ grade reagents were used for sample preparation and analysis (Fisher Scientific, Hanover Park, IL). EquiSPLASH^® LIPIDOMIX (Avanti Polar Lipids Inc., Alabaster, AL, USA) was used for the internal standard and NIST plasma Standard Reference Material (SRM) 1950 was prepared as a quality control.

To assess the linearity of different lipids for different tissue amounts, we collected a 62.9-mg biopsy of WAT from an individual female adult parous EPFU and extracted the lipid content into methanol (300 μl) with internal standards (5 μl, EquiSPLASH^® LIPIDOMIX) according to manufacturer’s instructions. Samples were left at −80 °C overnight, and on the following day, they were placed on dry ice and homogenized. After homogenization, samples were vortexed, incubated on ice (10 min), and cold chloroform (600 μl) was added. Samples were then vortexed, incubated on ice (10 min), and water (300 μl) was added. Samples were then vortexed, incubated on ice (10 min), and centrifuged for 5 min (1000 rpm, 4 °C). The bottom layer was removed and dried via lyophilization. The sample was reconstituted in isopropanol: acetonitrile (IPA: ACN 75:25, 200 μl) and diluted 10-fold for a 5-point curve.

The optimal tissue weight for analysis fell between ~ 0.5–10 mg, at which we see a loss of linearity (primarily in TAGs) with tissue weights over 10 mg (Supplementary Fig. 1). Excellent linearity (R² > 0.99, ePE R² = 0.97) was observed between 0.62 and 6.29 mg tissue weights; however, for the broad classes that are found in naturally lower proportion in WAT (e.g., ceramides [CERs], cholesteryl esters [CEs], acylcarnitines, and minor GPs and lysoglycerophosphospholipids [LGPs]), several compounds will not be detected in smaller tissue biopsies, which needs to be considered if attempting to analyze small micro-adipose samples from nonlethal sampling (Clerc et al., 2017). Although, it is important to note that these results are highly instrument specific and newer platforms offer improved sensitivity and dynamic range for detecting lower concentration compounds in biological samples.

Samples were injected (2 μl) and analyzed using a Waters Acquity Ultra Performance Liquid Chromatography (UPLC) (CSH C18 1.7 μm, 2.1 × 100 mm column) and a Xevo^® G2 quadrupole time-of-flight (QTOF) mass spectrometry (MS) (Waters, Milford, MA, USA). Positive and negative electrospray ionization (ESI) data-independent modes were used for data acquisition. Leucine enkephalin ([M + H]⁺ = 556.2771, [M-H]⁻ = 554.2615) was used as Lock-Spray^®. Operating conditions for ESI were: capillary voltage 3.0 kV, cone voltage 30 V, desolvation temperature 500 °C, and desolvation gas flow 1000 L/Hr. Mobile phases consisted of the following: solvent A (water/0.1% formic acid/10 mM ammonium formate), solvent B (ACN/0.1% formic acid), and solvent C (IPA/0.1% formic acid). The gradient was 8.0 min (solvent A 30%, B 34%, C 36%), 0.5 min 20% (solvent A 0%, B 10%, C 90%), and 2.5 min (solvent A 30%, B 34%, C 36%) at a flow rate of 0.45 ml/min and the column temperature maintained at 65 °C.

Our targeted lipid profiling assay has been previously described with lipid classes and multiple reaction monitoring (MRM) transitions listed (Pannkuk et al., 2022, 2023). Briefly, ~ 10 mg of tissue was homogenized as above, with tissue disruption after a 12-hr period at −80 °C, lipid isolation/purification using a liquid-liquid extraction, and lyophilization (Fig. 1). The dried extract was reconstituted in 200 μL of extraction buffer (IPA) and filtered using a 0.2 μM micro-centrifuge filter. The supernatant was transferred to MS vial for LC-MS analysis. As WAT is inherently low in both water and protein content, we normalized to weight rather than protein concentration to minimize the amount of sample needed for normalization and the contribution to the weight from water should be minimal. Quality control samples of both a pooled sample and the NIST SRM 1950 plasma mix were run every 10 samples. The SIL-30 AC auto sampler was kept at 15 °C (Shimazdu, Columbia, MD, USA). A 5 μL portion was injected and analyzed (Xbridge amide column, 3.5 μm, 4.6 × 100 mm) (Waters, Milford, MA, USA) using a Sciex QTRAP 5500 Mass Spectrometer (Sciex, Framingham, MA, USA). Lipids were detected by MRM transitions in both positive and negative ionization mode. Operating conditions were: temperature = 550 °C, nebulizing gas = 50 and heater gas = 60, curtain gas = 30, CAD gas = medium, ion spray voltage = 5.5 kV in positive mode and − 4.5 kV in negative mode. Mobile phases consisted of the following: solvent D (95% acetonitrile/5% water with 10 mM ammonium acetate) and solvent E (50% acetonitrile/50% water with 10 mM ammonium acetate). The mobile phase was initially 100% solvent D, then a gradient of 3.0 min (solvent D 99.9%, E 0.01%), 3.0 min (solvent D 94%, E 6%), 4.0 min (solvent D 25%, E 75%), 6.0 min (solvent D 0%, E 100%), and then 6.0 min equilibration back to 100% solvent D at a flow rate of 0.7 ml/min and the column temperature maintained at 35 °C.

Fig. 1.

(A) General schematic for tissue sampling and analytical workflow. (B) H&E staining of WAT sections from M. lucifugus showing decreased adipocyte size in the Pd-infected group. Bats used in the current study also had a higher number of inflammatory foci in the Pd-infected group. (*P ≤ 0.05; lines from top to bottom represent max value, 75th percentile, median, 25th percentile, min value)

2.4. Data analysis

Untargeted raw data files were imported into MassLynx v.4.1 (Waters Corporation, Milford, MA) and aligned used Progenesis QI (Nonlinear Dynamics, Newcastle, UK). Normalization was performed using a software chosen QC chromatogram as an alignment reference and normalizing to the “normalize to all compounds function”. For lipids of interest that were identified from the below targeted dataset, we visually inspected the m/z and MSe fragmentation patterns to confirm identities, and then standard curves were plotted in GraphPad Prism 9.2.0 the obtain R² values (GraphPad Software, La Jolla, CA). For the targeted lipidomics dataset, the qlm was imported into MultiQuant v 2.0 (Sciex, Framingham, MA, USA), peak areas were visually inspected, and the dataset was exporting to excel. We first removed lipids in the pooled QC sample with a coefficient of variation > 30%. The remaining lipids were normalized to weight of adipose biopsies and respective internal standards, (ESI⁺ dataset normalized to TAG.IS [15:0–18:1(d7)-15:0 TAG], for the ESI⁻ dataset PEs and LPEs were normalized to PE.IS [15:0–18:1(d7) PE] and the remaining dataset was normalized to [15:0–18:1(d7) PC]). We compared the control groups for M. lucifugus (n = 14) and E. fuscus (n = 6) (false-discovery rate [FDR] corrected P < 0.10 using a standard Benjamini-Hochberg procedure based on R p.adjust(raw.pvals.”fdr”) for species level differences. To determine if sex differences between uninfected M. lucifugus adult females vs. males (n = 5 for both groups) could influence the results of infection, we screened for statistically significant lipids that were below a FDR corrected P < 0.10. The sample size in the current study did not permit sex comparisons within E. fuscus. To determine disease effects for M. lucifugus (n = 14 control; n = 11 Pd-infected) and for E. fuscus (n = 6 control; n = 5 Pd-infected), we ranked lipids with the highest fold changes corresponding to the lowest P values (P < 0. 025, FDR corrected P ≤ 0.128). Lipid profiles were screened in SAS 9.4 (SAS, Cary, NC) and by using the t test feature in MetaboAnalyst 5.0 (Chong et al., 2018; Pang et al., 2021). MetaboAnalyst 5.0 was also used to construct heatmaps and box-and-whisker plots were generated in GraphPad Prism 9.2.0 (GraphPad Software, La Jolla, CA).

3. Results

Histological examination of WAT showed that there was a significant decrease in adipocyte size of Pd-infected M. lucifugus (control n = 6 and Pd-infected n = 5) that correlated with increased inflammatory foci previously reported (Fig. 1) (Moore et al., 2018). We then performed lipidomic analysis of WAT tissues of M. lucifugus and E. fuscus, with and without Pd exposure, predicting that we would find differences in the WAT chemical composition that depended on species and WNS-susceptibility.

Overall, the WAT composition in both M. lucifugus and E. fuscus is dominated by TAGs (> 90%) and a smaller proportion of glycerolipids, followed by PCs, PEs, FFAs, SMs, and other GPs. Importantly, although total proportion between M. lucifugus and E. fuscus are similar, the WAT chemical composition between control groups showed some interesting differences. We found 130 statistically significant different lipids between M. lucifugus and E. fuscus (Fig. 2, Supplementary Table 2). There were 26 TAGs in higher concentration in M. lucifugus compared to E. fuscus, with nearly half containing 20:4 or 22:5 acyl chains. A total of 12 TAGs were higher in E. fuscus. Higher GP levels in E. fuscus WAT were dominated by compounds containing unsaturated or monounsaturated moieties along with potential n-6 (18:2, 20:2, 20:3, 20:4) fatty acids. Higher GP levels in M. lucifugus WAT were primarily compounds containing potential n-3 (20:5 and 22:5) fatty acids. As mammals cannot synthesize these lipids, differences between M. lucifugus and E. fuscus were likely due to diet.

Fig. 2.

Species level differences in lipid compounds in white adipose tissue (WAT) between the Myotis lucifugus (n = 14) and Eptesicus fuscus (n = 6) control groups. Although broad lipid class proportions are similar between the species, there are distinct differences in the WAT chemical composition, which may have important downstream implications for over winter energy reserves. Box and whisker plots of some select lipids show a trend for increased concentration of lipids with n-6 constituents in E. fuscus converse to more highly unsaturated n-3 compounds in M. lucifugus. (*FDR corrected P ≤ 0.10; ** FDR corrected P ≤ 0.01, lines from top to bottom represent max value, 75th percentile, median, 25th percentile, min value., P value and fold change are listed in Table S2)

We screened for sex differences between adult male and female M. lucifugus as we had a sufficient sample size of adult animals to examine potential differences. Upon finding no significant differences in the WAT chemical composition between adult male and female M. lucifugus control groups (Supplementary Table 3), we lumped the sexes for further analysis due to Pd-infection.

Comparing the uninfected and Pd-infected groups within each species, we did not find any significant differences in E. fuscus or in the glycerolipids or low abundance compounds in M. lucifugus (Supplementary Table 4). The remaining lipids in M. lucifugus were ranked by P value, with the top 66 lipids having a raw P value < 0.025 (FDR corrected P value = 0.128) (Table 2). These 66 compounds correlated to primarily PCs and PEs that are higher in Pd-infected M. lucifugus compared to uninfected M. lucifugus (Fig. 3). Functionally, these compounds are structural in nature as they comprise cellular membranes, with PEs also being found in high levels in mitochondria (Lee et al., 2019). These observations on the increasing concentration of GPs correlates with the histological observation that WAT remodeling is occurring along with increased ECM deposition in M. lucifugus during Pd-infection but not in E. fuscus.

Table 2.

Top ranked lipids found in white adipose tissue of Myotis lucifugus comparing control group vs. Pd-infection group (normalized peak area Mean ± Std dev)

	M. lucifugus	M. lucifugus	Fold	P value	FDR
	Control	Pd-infected	Change
LPE(22:6)	54.0 ± 56.1	148.2 ± 126.4	2.7	0.022	0.128
PC(14:0/20:1)	60.2 ± 36.7	149.9 ± 99.7	2.5	0.006	0.128
PC(14:0/20:3)	57.5 ± 29.9	105.3 ± 59.9	1.8	0.018	0.128
PC(16:0/20:2)	175.5 ± 86.7	329.7 ± 168.8	1.9	0.008	0.128
PC(16:0/20:3)	1747.8 ± 758.8	2761.6 ± 1174.3	1.6	0.018	0.128
PC(16:0/22:6)	1860.6 ± 1260.8	3490.4 ± 1712.5	1.9	0.013	0.128
PC(17:0/16:1)	1093.9 ± 400.1	1667.7 ± 603.8	1.5	0.010	0.128
PC(17:0/18:1)	1819.2 ± 801.9	2985.3 ± 1227.9	1.6	0.010	0.128
PC(17:0/18:2)	2480.4 ± 1112.7	4001.3 ± 1736.5	1.6	0.016	0.128
PC(17:0/20:3)	140.7 ± 69.7	262.3 ± 102.4	1.9	0.002	0.128
PC(17:0/20:4)	1152.0 ± 582.2	1774.3 ± 657.0	1.5	0.022	0.128
PC(17:0/22:4)	58.6 ± 26.4	112.8 ± 58.1	1.9	0.006	0.128
PC(17:0/22:6)	159.6 ± 77.2	315.5 ± 212.4	2.0	0.020	0.128
PC(18:0/16:1)	5749.6 ± 1611.1	10220.4 ± 5180.1	1.8	0.007	0.128
PC(18:0/20:0)	423.3 ± 250.4	920.9 ± 537.4	2.2	0.006	0.128
PC(18:0/20:1)	133.1 ± 63.1	287.7 ± 151.1	2.2	0.002	0.128
PC(18:0/20:2)	98.9 ± 70.1	263.1 ± 158.7	2.7	0.002	0.128
PC(18:0/20:3)	2044.2 ± 1246.3	4536.2 ± 2644.2	2.2	0.006	0.128
PC(18:0/22:4)	629.5 ± 391.0	1232.4 ± 738.2	2.0	0.017	0.128
PC(18:0/22:6)	1148.6 ± 820.3	2817.2 ± 1685.2	2.5	0.004	0.128
PC(18:1/20:1)	482.5 ± 291.9	947.8 ± 605.5	2.0	0.021	0.128
PC(18:1/20:2)	197.2 ± 122.4	436.1 ± 246.3	2.2	0.005	0.128
PC(18:1/20:3)	1551.5 ± 968.4	2877.0 ± 1319.5	1.9	0.009	0.128
PC(18:1/22:4)	443.4 ± 237.4	748.1 ± 342.0	1.7	0.017	0.128
PC(18:1/22:6)	762.8 ± 566.8	1409.9 ± 676.9	1.8	0.018	0.128
PC(18:2/20:3)	216.5 ± 135.8	426.3 ± 252.3	2.0	0.016	0.128
PC(20:0/22:6)	58.4 ± 31.5	128.5 ± 76.4	2.2	0.006	0.128
PE(14:0/22:2)	36.9 ± 25.6	73.8 ± 42.4	2.0	0.015	0.128
PE(16:0/16:0)	844.2 ± 391.1	1790.3 ± 1259.2	2.1	0.016	0.128
PE(16:0/20:2)	30.5 ± 24.7	94.9 ± 79.3	3.1	0.010	0.128
PE(16:0/20:3)	282.7 ± 251.9	658.6 ± 465.1	2.3	0.019	0.128
PE(16:0/22:6)	659.2 ± 600.9	1853.3 ± 1427.7	2.8	0.011	0.128
PE(17:0/18:1)	669.8 ± 554.1	1483.3 ± 842.5	2.2	0.009	0.128
PE(17:0/22:4)	259.6 ± 124.8	414.8 ± 184.9	1.6	0.022	0.128
PE(18:0/16:0)	152.1 ± 130.0	343.6 ± 247.1	2.3	0.023	0.128
PE(18:0/16:1)	543.6 ± 495.9	1373.1 ± 1068.6	2.5	0.019	0.128
PE(18:0/18:2)	34137.0 ± 25058.0	68948.2 ± 43331.5	2.0	0.021	0.128
PE(18:0/20:1)	113.1 ± 68.2	235.4 ± 141.3	2.1	0.010	0.128
PE(18:0/20:2)	124.4 ± 66.5	261.1 ± 148.5	2.1	0.006	0.128
PE(18:0/20:3)	1659.0 ± 1400.7	4389.3 ± 3055.3	2.6	0.008	0.128
PE(18:0/20:4)	40427.8 ± 34733.4	80906.8 ± 47369.1	2.0	0.024	0.128
PE(18:0/22:4)	612.7 ± 533.9	1604.4 ± 1168.0	2.6	0.011	0.128
PE(18:0/22:5)	6307.4 ± 4606.8	12490.6 ± 7679.3	2.0	0.023	0.128
PE(18:0/22:6)	1931.3 ± 1724.6	5912.7 ± 5667.8	3.1	0.023	0.128
PE(18:1/16:1)	6961.9 ± 5955.6	15763.2 ± 10835.3	2.3	0.019	0.128
PE(18:1/18:2)	74301.5 ± 52436.4	131510.8 ± 64160.0	1.8	0.025	0.128
PE(18:1/20:1)	328.3 ± 272.4	717.5 ± 496.6	2.2	0.023	0.128
PE(18:1/20:2)	168.9 ± 147.2	398.8 ± 279.7	2.4	0.016	0.128
PE(18:1/22:4)	1429.3 ± 1241.2	3053.2 ± 1924.8	2.1	0.020	0.128
PE(18:1/22:6)	3541.1 ± 3373.6	9421.3 ± 7578.0	2.7	0.018	0.128
PE(O-16:0/16:0)	60.9 ± 37.9	132.5 ± 88.1	2.2	0.013	0.128
PE(O-16:0/18:2)	271.7 ± 245.7	615.6 ± 445.9	2.3	0.025	0.128
PE(O-18:0/18:1)	1421.5 ± 1016.5	2908.1 ± 1772.5	2.0	0.017	0.128
PE(O-18:0/18:2)	3899.8 ± 3113.7	8373.6 ± 5502.8	2.1	0.019	0.128
PE(O-18:0/20:3)	237.6 ± 204.9	518.2 ± 353.3	2.2	0.023	0.128
PE(P-18:1/18:2)	4785.8 ± 4139.4	13352.1 ± 12000.9	2.8	0.023	0.128
PE(P-18:1/20:3)	1561.8 ± 1466.4	3786.7 ± 2688.5	2.4	0.017	0.128
PE(P-18:1/22:4)	2287.3 ± 2206.4	6622.8 ± 5765.6	2.9	0.019	0.128
PG(16:0/20:2)	137.9 ± 179.2	549.1 ± 587.7	4.0	0.023	0.128
PG(16:0/20:2)_2	415.9 ± 520.0	1598.3 ± 1694.0	3.8	0.024	0.128
PS(18:0/18:1)	1560.3 ± 1459.5	3804.4 ± 2883.5	2.4	0.021	0.128
FFA(20:3)	35013.5 ± 15203.6	57342.6 ± 26179.7	1.6	0.015	0.128

Fig. 3.

Differences in lipid compounds in white adipose tissue (WAT) between the Myotis lucifugus control group (n = 14) and Pd-infected group (n = 11). Values for Eptesicus fuscus groups (n = 6 control, n = 5 Pd-infected) are provided as a comparison. Disease level changes were only observed in glycerophospholipid levels, primarily PCs and PEs, indicating a greater contribution from structural membranes or increased mitochondrial numbers. (*FDR corrected P = 0.128, lines from top to bottom represent max value, 75th percentile, median, 25th percentile, min value, P value and fold change are listed in Table 2)

4. Discussion

During Pd-infection, WNS sensitive bats undergo increased torpor arousals during winter months and prematurely exhaust fat stores and die due to emaciation (Reeder et al., 2012; Lilley et al., 2016). As not all bats are equally susceptible to WNS associated mortality, a key research priority has been deciphering the traits driving this differential susceptibility among species. While total fat stores have been the focus of previous studies (Haase et al., 2021), another factor that has been less explored is potential differences in prehibernation WAT quality and pathological changes during WNS. Here, we noted that while total WAT is comprised largely of TAGs and other glycerolipids associated with energy storage, several other lipid species, such as GPs and SMs, are present and aid in structural integrity. For GPs, significantly higher compounds found in E. fuscus WAT had n-6 fatty acid moieties, whereas n-3 dominated in M. lucifugus. We did not find sex differences in the WAT from adult M. lucifugus, which has also been found in the Common pipistrelle (Pipistrellus pipistrellus) (Arévalo et al., 1990). Compared to hibernating bats without infection, M. lucifugus infected with Pd showed WAT remodeling including increased presence of ECM, decreased adipocyte size, and increased GP levels.

There were several species level differences observed in the WAT composition between M. lucifugus and E. fuscus. This is not surprising as these are different species and they consume different types of insects (Whitaker Jr, 2004; Wray et al., 2018). The ratio of depot fat to lean mass of hibernators peaks in the fall before they reduce their metabolic rate and begin torpor bouts, so the “snapshot” of WAT composition presented in this work represents the metabolic fuel that would be available for periodic rewarming during winter and sustain them for spring foraging and female reproduction (Kunz et al., 1998). In addition to species-level differences in adipocyte biology being well recognized (Börgeson et al., 2022), fat composition and lipid content have also been linked to hibernation strategies between different bat species (Bachorec et al., 2023). Interestingly, the many of the significantly higher TAG species found in M. lucifugus compared to E. fuscus contained 22:5 (docosapentaenoic acid) n-3 acyl species, which is the pathway we previously found perturbed in hepatic and splenic tissues in a different cohort of Pd-infected M. lucifugus (Pannkuk et al., 2021, 2023). Similarly for the GP fraction, M. lucifugus WAT contained higher n-3 species while E. fuscus had predominantly higher levels of n-6 GPs. Several studies have highlighted that not all fats are equal, in that during hibernation there is a selective retention of lipids based on saturation and that the n-6/n-3 ratio affects torpor quality (see (Siutz et al., 2022) for review). Even in animals with dramatically different hibernation patterns (captive garden dormice [Eliomys quercinus] vs. free-ranging brown bears [Ursus arctos]), elevated n-3 plasma GPs were observed suggesting a common hibernation lipid phenotype (Strandvik et al., 2023). A potential mechanism for these observations is that increased membrane fluidity may be needed to maintain function, such as sarcoendoplasmic reticulum calcium ATPase activity (Ruf & Arnold, 2008). Further research may elucidate if perturbation to n-6/n-3 membrane ratios may play a role in differential torpor patterns between M. lucifugus and E. fuscus during WNS.

The basic component of WAT is the adipocyte, which is a unilocular structure consisting of stored TAGs (and other glycerolipids and cholesteryl esters) surrounded by a thin rim containing the cell membrane, cytoplasm, and organelles. The adipocytes are surrounded by a structural ECM that together forms a type of connective tissue. Within WAT, the adipocytes are metabolically active, whereas the ECM is composed of noncellular components, including collagens, laminins, fibronectin, and proteoglycans and not generally considered metabolically active (Mariman & Wang, 2010). During periods of torpor and low prey availability, adipocytes play a vital role as they release FFAs from stored TAGs providing fuel through fatty acid β oxidation (Choe et al., 2016), and during the hibernation period and periodic arousals ~ 85% of total WAT is consumed (Thomas et al., 1990). The chemical composition of WAT can change during these periods of differential energy output (e.g., exercise (May et al., 2017) or winter torpor) and also other factors, such as region of the body (Al-Sari et al., 2020; Hou et al., 2020) or diet (Caesar et al., 2010). In addition, WAT remodeling can be associated with an aging or pathogenic state. Patients with cancer-induced cachexia show decreased adipocyte size along with increased cell atrophy, fibrosis, and ECM deposition (Alves et al., 2017). Pre-hibernation WAT remodeling has also been described in a food-storing hibernator (Syrian hamster, Mesocricetus auratus, (Chayama et al., 2019). In torpor and early arousal, the thirteen-lined ground squirrel (Ictidomys tridecemlineatus) shows suppressed inflammatory cytokine and matrix metalloproteinase activity in WAT suggesting that tissue remodeling is not typical during this period in a nonpathogenic state (Logan & Storey, 2021). A hallmark of WNS is the premature exhaustion of fat stores, so a reduction in adipocyte size or number comprising WAT would be expected. In this regard, our observation of a decrease in adipocyte cell size and an increase in ECM indicates both histological and subsequent metabolic remodeling of WAT, reflected through increased PC and PE concentration in the WAT of Pd-infected M. lucifugus. There is a dynamic balance between adipocytes and the composition of the ECM within WAT that is essential for its normal metabolic function and torpor (Mariman & Wang, 2010). Further studies are required to characterize the ECM changes and their implications in adipocyte metabolism between species during WNS.

Inherent to understanding the WNS process is determining metabolic changes that are occurring in resistant vs. susceptible bat species during Pd-infection (McGuire et al., 2017; Meierhofer et al., 2018). These changes will occur in tandem with the metabolic changes associated with normal torpor arousal, which has presented some difficulty in teasing out the effects from infection and interbout arousals. One approach used an RNA-Seq method to analyze transcript levels in wing tissue with/without Pd-infection during periods of torpor or euthermy in M. lucifugus (Field et al., 2015; Lilley et al., 2019). Interestingly, two of the most consistently upregulated genes due to local Pd-infection were involved in lipid metabolism, including FFAR2 (free fatty acid receptor 2) and PTGS2 (cyclooxygenase). FFAR2 was highly upregulated due to both euthermy and Pd-infection. It is activated by short chain fatty acids (SCFAs) acetic acid, propionic acid (3 carbon fatty acid), and butyric acid (4 carbon fatty acid) that are primarily microbially produced and play a major role in adipocyte metabolism (May & den Hartigh, 2023). In addition to being directly metabolized for energy, SCFAs can influence adipocyte lipolysis, adipogenesis, and browning of adipose tissue. SCFAs can may also modulate inflammation by attenuating cytokine levels; however, much of these data are derived from in vitro experiments and their in vivo response in wild bat populations remains to be determined. A more direct correlation to inflammation was identified as increased transcription of PTGS2, which encodes for cyclooxygenase-2 (COX-2: EC 1.14.99.1) and is involved in the oxygenation of arachidonic acid (FFA 20:4) to the eicosanoid, prostaglandin H₂ (PGH₂). These downstream metabolites of arachidonic acid are primary drivers in inflammation, which has made them interesting candidates to study in WNS. PGH₂ can also be converted to PGI₂ (prostacyclin), an unstable lipid that is measured in the body as 6-keto-PGF1α. Levels of 6-keto-PGF1α were significantly increased in the spleen tissue of M. lucifugus at early stages of WNS, however, it plays a more important role in vasodilation and platelet formation rather than inflammation (Pannkuk et al., 2023). Analysis of the metabolites involved in the eicosanoid pathway at more advanced stages of WNS and the presence/severity of inflammation between species and torpor vs. arousals remains an area where additional data are needed.

Finally, from a more technical standpoint, there has been interest in developing non-lethal sampling techniques to elucidate the WNS process to minimize both the number of animals removed from the wild and possible confounding effects from laboratory infection (Reeder et al., 2016). Non-lethal sampling methods from wild bat species have also been proposed for other types of studies, including viral monitoring and discovery (Young & Olival, 2016). Sample collection has typically included wing biopsies/wipes (Paassilta et al., 1998; Pannkuk et al., 2014), blood (Eshar & Weinberg, 2010), saliva (Amman et al., 2015), and fecal samples (Ingala et al., 2021). In addition, adipose tissue biopsies have been used to measure changes in lipid content by fatty acid profiling via chemical derivatization to fatty acid methyl ester (FAME) analysis (Caesar et al., 2010; Clerc et al., 2017). Although derivatization can increase sensitivity that is useful for low tissue masses, a limitation inherent to FAME analysis is that structural information is lost after fatty acid cleavage from their headgroup (e.g., glycerol, sphingosine, cholesterol), and fractionation of lipid subclasses prior to analysis is a tedious process (Bermúdez-Cardona & Velásquez-Rodríguez, 2016). While restricting the analysis to carbon chains can be beneficial when analyzing the dietary component of WAT, such as TAGs that will be used for fatty acid β oxidation (Frank et al., 2012, 2014; Ingala et al., 2017), or homogenous chemical mixtures like hydrocarbons used in energy combustion, complications in biological interpretation may arise when dealing with heterogeneous mixtures with varied biological function. We highlight (in addition to others (Al-Sari et al., 2020) that the WAT has several GP and sphingolipid species involved in cellular structure and ECM maintenance (Qu et al., 2023) with differing n-6/n-3 ratios that would not been detected in the current study using fatty acid profiling. Here, we found that total lipid profiling (i.e., lipidomics) was capable of analyzing small tissue masses that could be collected in a nonlethal manner and offer a level of sensitivity to detect disease severity in wild populations. Although lipid structural integrity is maintained during this analysis, the complexity of data interpretation does increase as a higher number of analytes are detected. During a typical lipidomics experiment, ~ 1,000–10,000 compounds are detected (depending on targeted vs. untargeted analysis) vs. ~5–20 features measured during fatty acid profiling (assuming mammalian samples).

5. Conclusions

We highlighted differences in WAT lipid composition between M. lucifugus and E. fuscus that may have implications on pre-hibernation strategies between these species. As expected, histological and chemical changes are only detected in the WNS sensitive M. lucifugus following Pd-infection. Perturbation to the WAT in Pd-infected M. lucifugus manifested as smaller adipocyte size, increased ECM deposition, and increased in GP levels. We did not observe significantly decreased levels of TAGs in Pd-infected M. lucifugus that would likely be associated with free-ranging populations, which is in line with the lack of change in body condition (Moore et al., 2018), and may be an artifact of study duration. As previous studies have focused solely on total fat loss during WNS, these studies have shown that additional metabolic stress is occurring in bats at early stages of disease (Pannkuk et al., 2021, 2023) and that the effects from Pd-infection to adipose stores includes ECM deposition that may impact recovery periods beyond simply replenishing fat stores during spring emergence.

Supplementary Material

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s11306-024-02165-4.

Abbreviations

WNS

White-nose syndrome

Pd

Pseudogymnoascus destructans

MYLU

Myotis lucifugus

EPFU

Eptesicus fuscus

GL

Glycerolipid

FFA

Free fatty acid

TAG

Triacylglyceride

Cer

Ceramide

GP

Glycerophospholipid

PC

Phosphatidylcholine

LysoPC

Lysophosphatidylcholine

PG

Phosphatidylglycerol

DAG

Diacylglyceride

MAG

Monoacylglyceride

CE

Cholesteryl ester

SM

Sphingomyelins

ePC

Ether-linked phosphatidylcholines

PE

Phosphatidylethanolamines

ePE

Ether-linked phosphatidylethanolamines

PI

Phosphatidylinositols

PS

Phosphatidylserine

LGP

Lysoglycerophospholipid

LPI

Lysophosphatidylinositols

LPE

Lysophosphatidylethanolamines

SCFA

Short-chain fatty acid

ECM

Extracellular matrix

WAT

White adipose tissue

MRM

Multiple reaction monitoring

ESI

Electrospray ionization

UPLC

Ultra-performance liquid chromatography

QTOF

Quadrupole time-of-flight

MS

Mass spectrometry

IPA

Isopropanol

ACN

Acetonitrile

Data availability

Data for targeted lipidomics have been entered into Metabolomics Workbench with DataTrack ID 5132 with an embargo of 1/1/25 (https://www.metabolomicsworkbench.org/).