Trp53 deficient mice predisposed to preterm birth display region-specific lipid alterations at the embryo implantation site

Here we demonstrate that conditional deletion of mouse uterine Trp53 (p53d/d), molecularly linked to mTORC1 activation and causally linked to premature uterine senescence and preterm birth, results in aberrant lipid signatures within the heterogeneous cell types of embryo implantation sites on day 8 of pregnancy. In situ nanospray desorption electrospray ionization mass spectrometry imaging (nano-DESI MSI) was used to characterize the molecular speciation of free fatty acids, monoacylglycerol species, unmodified and oxidized phosphatidylcholine (PC/Ox-PC), and diacylglycerol (DG) species within implantation sites of p53d/d mice and floxed littermates. Implantation sites from p53d/d mice exhibited distinct spatially resolved changes demonstrating accumulation of DG species, depletion of Ox-PC species, and increase in species with more unsaturated acyl chains, including arachidonic and docosahexaenoic acid. Understanding abnormal changes in the abundance and localization of individual lipid species early in the progression to premature birth is an important step toward discovering novel targets for treatments and diagnosis.

kV for negative mode was applied to the primary capillary. The sample was mounted on a motorized XYZ translational stage operated by a custom-designed LabVIEW software and moved in z-direction according to the plane in which the sample resided (described elsewhere) (4). The stage was continuously moved under the nano-DESI probe at 20 µm/s when data was acquired at 100 000 (m/Δm) and at 40 µm/s when acquiring at 60 000 (m/Δm). Lines were spaced by 150 µm resulting in an average pixel size of approximately 12×150 µm 2 (x×y) for ion images recorded at mass resolution of 100 000 and 40×150 µm 2 (x×y) for ion images recorded at mass resolution of 60 000. To ensure independence of carry over effects between implantation sites regions, the direction of analysis was alternated, from AM-pole to M-pole or from M-pole to AM-pole, between biological replicates.
Data processing: Ion images were generated using the in-house developed software MSIQuickview (4). The presented ion images of Ox-PC species, depicted in Figure 3 in the manuscript, are normalized to the standard LPC 19:0 (0.18 µM) to account for matrix effects which could distort the ion distribution (5). Similarly, the presented ion images of DG species, depicted in Figure 2 in the manuscript, are normalized to the standard DG 14:0/14:0/0:0 (2.1 µM). The ion images of abundant PC species, depicted in Figure 1 in the manuscript, were normalized to the total ion current. Each ion image has its own intensity scale (0-100%) to increase clarity in presentation. Note that the higher ion signals at the embryonic site are interpreted to be an artifact of analysis originating from the more compact material at this site. Tables S2-S4 detail the m/z, p-values, and abundance data depicted in Figures 1, 2 & 3. For Figure 4, Comparison of molecular species within each lipid class was performed by normalizing the signal for each species within the region of interest with the signal for the sum of all species within the molecular class. Table S5 details the m/z, p-values, abundance data, and the ratio of the p53 d/d over p53 f/f which were calculated to show the differences depicted in Figure 4 in the manuscript.
Regions of interest were generated using MSIQuickview(4) using specific ions as markers for the regions. In positive mode the AM-pole region was marked by m/z 848.56 and the M-pole was marked by m/z 798.54, similarly m/z 790.53 marked the AM-pole in negative mode m/z 883.52 marked the M pole region.
Peak assignment: DG species are assigned based on 1) their m/z 2) the existence of multiple cation adducts 3) their ionization properties (similarly to the standard they do not produce negative ions) and 4) the lack of MS/MS fragments supported by the need of high collision energy for fragmentation of the standard. Phosphatic acid (PA) 17:0/17:0 (Avanti Polar Lipids) and DG 14:0/14:0/0:0 (Avanti Polar Lipids) was used to determine differences in ionization and fragmentation between PA and DG to further confirm the identity of endogenous DG; due to the close m/z range for accurate masses of DG species and PA species. Oxidized phosphatidylcholines are assigned based on 1) their m/z 2) MSMS. (See supporting information below). RNA isolation and quantitative PCR. RNA was prepared from homogenized implantation sites using TRIzol reagent (Invitrogen). RNA extraction was performed as described previously (1,6). Quantitative PCR (qPCR) was performed using StepOnePlus Real-Time PCR System (Applied Biosciences).
Statistics of Biological Experiments. Statistical analyses were performed using 2-tailed heteroscedastic Student's t-test. P-values below 0.05 were considered statistically significant. All p-values for reported results can be found in Tables S2-S5 below.

Peak assignment of DIACYLGLYCEROL species
Despite high mass resolution mass spectrometry some peaks identified as diacylglycerol by accurate mass could potentially be assigned to plasmalogen species of phosphatic acid, as shown in Table S1. Table S1 includes protonated, sodiated and potassiated species possible within 10 ppm of the experimentally obtained m/z values. In nano-DESI MSI of tissue from mouse embryo implantation sites, the majority of chemical species are found as potassium adducts. However, no potassium adducts are found for any of the suggested PA species, but there are sodium adducts found to support the assignment of potassiated DG species (marked by stars in Table S1).
The standards DG 14:0/14:0/0:0 and PA 17:0/17:0 were used at high concentrations (100 µM and 133 µM, respectively) to compare ionization and fragmentation patterns of the two lipid classes, PA and DG. It was found that while PA ionizes in both positive and negative mode, as shown in Figure S1, DG only ionizes in positive mode as shown in Figure S2. Not even at extremely high concentrations (500 µM) did DG produce ions in negative mode. This difference between the two species validates the assignment of DG since no peaks corresponding to the possible endogenous PA species were found in negative mode spectra obtained during MSI. MSMS of the standards further showed that DG requires a higher collisional energy to produce fragment ions. MSMS of the peaks of interest was performed directly from the tissue but no fragments of was found. The difficulty to produce fragment ions further supports the assignment of DG. Table S1. Diacylglycerol identification * Mark DG species supported by a sodium adduct, no potassium adduct is detected for any PA species