Animals and Irradiation Procedure

Six -eight weeks old mice were purchased from Charles River Laboratories and subcontracted from NCI, (Fort Dietrick, MD) and housed at the institutional animal facility. For feasibility studies, anesthetized mice were exposed to 10.4 Gy WBI (236cGy/min) using Cesium irradiator as described elsewhere [11]. 10.4 Gy dose was chosen based on our previous studies that established bone marrow and gastrointestinal injury in mice leading to mortality within 2 weeks post-WBI. Mice were returned to their cages with food and water ad libitum. For validation studies, anesthetized mice were exposed to 3, 6, 8, and 10 Gy WBI using the same irradiation protocol. All procedures were conducted by strictly adhering to the institutional IACUC-approved guidelines and protocols.

Sample Collection, Exosome isolation and LC-MS analysis

12hrs prior to sacrifice all mice were food restrained. Animals were lightly anesthetized and blood was collected by retro-orbital sinus bleeding and processed for serum isolation. Serum and Urine exosomes were isolated and analyzed by LC-MS/MS (SI). For the irradiated group, urine was collected at 24 and 72hr, representing the early and delayed responses. For total urine proteome analysis, 20ul from each urine sample (n = 5) were run on SDS-PAGE, subdivided into 6 gel pieces, and proteolytically cleaved to extract peptides for mass spectrometry-based proteomic analysis. For urine exosome proteome analysis, 50μl of urine were used to isolate total exosome (Life Technology, Inc.). 15μg of exosome lysates (n = 5) were run on SDS-PAGE, bands were isolated, digested and run on LC-MS/MS. Similar protocol was followed for serum samples. The sample preparation and MS workflow is illustrated in Figure 1A.

Open in a separate window

Figure 1

Global urine and urine exosome proteomic analysis

(A). LC-MS/MS analysis of urine and urine exosome. (B). Principle component analysis (PCA) of urine and urine exosome proteome identified by DDA.

Targeted SRM analysis

A list of precursor and product ion masses of 2 or 3 peptides from EGF, PLAU, VDAC, and HEXB was generated by the Skyline and used to perform targeted MS analysis. The following peptides were used to show quantitative MS results. EGF – INLHPSFVTPGK. PLAU – ESESDYLYP. VDAC – LTFDSSFSPNTGK. HEXB – VEPLNFEGSEK.

DDA and DIA combined approach for radiation signatures in serum and urine exosomes

One of the challenges of utilizing the label-free LC-MS workflow is the sensitivity of protein detection and quantification. Traditional LC-MS workflow based on DDA increases overall identification of proteins but suffers from reduced numbers of identifiable peptides across multiple runs [12,13]; DIA on the hand, presents an attractive solution to improve label-free quantification [14]. Additionally, fractionation of total urine and urine exosome also reduced repeated sampling of high abundance proteins and increased the depth of protein identification.

For urine exosome characterization, 274 unique proteins were identified from DDA and 178 unique proteins were identified from DIA respectively (Fig. 2). 384 proteins remained after filtering for low variance, corresponding to 263 unique urine exosome proteins used for analysis. For serum exosome characterization, 434 unique proteins were identified from DDA and 220 unique proteins identified from DIA respectively (Fig. 2). 478 proteins remained after filtering, corresponding to 317 unique serum exosome proteins used for analysis.

Open in a separate window

Figure 2

DDA and SWATH-MS combined workflow for characterizing urine and serum exosomes in response to radiation

Distinct protein signatures were found in urine and serum exosomes at 5% FDR and dendrogram clusters were found among biological replicates of three experimental groups.

GeneOntology analysis

To determine subcellular location and function of the proteins identified in our samples, we used GeneOntology (GO) analysis. Majority of urine and serum exosome proteins were found to be from three subcellular locations: plasma membrane, cytoplasm, and extracellular space (Fig. 3A, top). Urinary exosomes showed significantly increased levels of plasma membrane protein (52.4%) compared to the serum exosomes (13.3%) while, only serum exosomes showed significant increase in nuclear proteins (7.4%).

Open in a separate window

Figure 3

GO classification of urinary and serum exosomes

(A) Total urine and serum exosome unique proteins categorized based on subcellular localization and molecular functions re shown. (B) Urinary and serum exosome proteome obtained using GO classification compared to the proteomic dataset from the Vesiclepedia exosome database.

Both urinary and serum exosome hits showed diverse molecular functions, serum exosome analysis identified transporters, transcriptional regulators, phosphatases, G-protein coupled receptors, cytokines and growth factors. Although many were classified as other (function not known), three major categories emerged in urine and serum exosomes identified in this study: enzyme, peptidase, and transporter (Fig. 3A, bottom). We expanded our comparison to examine the subcellular and functional categories of proteins identified in our dataset to the proteomic dataset from the Vesiclepedia exosome database [15] and found that the urinary exosome showed over two-fold increase in plasma membrane proteins compared to the serum exosomes (Fig. 3B, top). Furthermore, both urinary and serum exosome proteins in our dataset showed two-fold or more increase in extracellular proteins than in the Vesicleopedia database (Fig. 3B, top). While serum exosome proteins had a similar distribution of molecular functions except peptidases (> 2 fold increase), striking reduction in transcriptional regulators, translational regulators, and phosphatases from urinary exosomes was observed (Fig. 3B, bottom).

Identification of radiation biomarkers from urinary exosomes

From the urinary exosome dataset, 23 proteins were altered post-WBI with 5% FDR and adjusted p-value < 0.004 (Fig. 4A). Most of radiation sensitive biomarkers in urinary exosomes showed elevated levels in irradiated mice compared to the controls. Among these, 14 showed at least 2 fold higher expression (Fig. 4A); these biomarkers also indicated alteration in radiation-responsive signaling pathways such as redox, apoptosis and inflammatory response. These included PRDX6, a non-selenium thioredoxin with phospholipase A2 activity [16]. PRDX6 is highly expressed in lungs, liver, kidney, testis, and brain, especially in the epithelium [17,18]. VDAC1, an outer mitochondrial membrane protein known to be crucial for regulation of mitochondrial function [19], was an early response protein observed to be elevated at 24 h and remained elevated at 72h (6A). VDAC1 is a key pro-apoptotic molecule in cytochrome C mediated apoptosis, and therefore is a surrogate marker for increased apoptosis in kidney at 24 and 72 hr post-WBI. [20]. PLAU, a serine kinase was significantly lower in irradiated cohorts, again indicative of increased apoptosis [21]. Others proteins included PLD3 and UROM, which were also reported as biomarkers of renal injury and graft rejection in renal transplantation patients [22]. These markers therefore represent acute and persistent renal epithelial tissue damage occurring in response to WBI. Other biomarkers of early response (only at 24 hr) included A2M [23,24], ALB, GLB, HEXB, AGT, and GALNS, routinely found in urine and lysosomes.

Open in a separate window

Figure 4

Protein signature of radiation exposure in urine and serum exosomes

(A) PCA showed a distinct separation of these two groups of mice (p<0.05). A heat map showed the differential expression of 20 urine exosome proteins in response to radiation exposure. A second heat map displaying proteins had at least a 2-fold increase in irradiated mice in response to radiation compared to the non-irradiated mice. (B) A distinct protein signature was found in serum exosomes between the first two groups of mice (non-irradiated and 10.4 Gy 24hr) and the third group of mice (10.4 Gy 72hr) (p<0.05).

Relative quantification of peptides from urinary exosome biomarkers using Skyline peak area quantification was also conducted on urine exosome proteome. The left panel in Figure 5 shows spectral counts for various proteins identified in groups I(top), II(middle), and III(bottom). As indicated in the right panel, VVFIFGPDK, derived from PRDX6 from the first group (top) and QFYSVFDR, derived from CATE from the second group (middle) confirms that these proteins indeed showed radiation-induced reduction in the spectral counts. WTEYGLTFTEK, derived from VDAC1 shown in the third group shows radiation-induced increased counts in urine exosomes.

Open in a separate window

Figure 5

Skyline displayed peak area view of 4 peptides with distinct patterns

Top Row - Down-regulated with radiation: VVFIFGPDK, derived from PRDX6 and extracted by precursor ion at m/z 511.2895 (2+) and QFYSVFDR, derived from CATE and extracted by precursor ion at m/z 531.2562 (2+). Bottom Left - Acute-phase response: LQPALWPFPR, derived from HEXB and extracted by precursor ion at m/z 612.8480 (2+). Bottom Right - Late-phase response: WTEYGLTFTEK, derived from VDAC1 and extracted by precursor ion at m/z 687.8324 (2+).

Identification of radiation biomarkers from serum exosomes

Exosome protein profiles of serum samples with 5%FDR and adjusted p-value < 0.009 are shown in Fig. 4B. Figure 6B shows key proteins that were identified in serum exosome analysis. A2M, which was down-regulated in urine exosomes at 72 hour post-WBI, is also an APP known to have beneficiary effects on sepsis in terms of recruitment of neutrophil at the site of infection or on the surface of endothelial cells [23]. A2M levels were correlated with surviving sepsis patients [25]. It is interesting that A2M was the only APP marker that was identified in urine exosomes; whereas, several APP markers such as SAA1 and 2, SPA3N, APOA2 were identified in serum exosomes [23,24]. Also, the kinetics of APP changes detected in serum versus urine was different. Most APP in serum exosomes were early responsive and were highly elevated at 24 hr post-WBI, and returned to normalcy by 72 hr post-WBI, the only APP found in urine (A2M) was slow responsive and was down-regulated at 72 h, interestingly it was not identified in serum exosomes. Additionally, progressive reduction of fibrinolysis (KLKB1 and PROS1), cell motility/adhesion factors (EGFR, HRG, IGFALS, and TF), and alternative complement pathway components (C8A, C8B, C8G, and C9) was observed in serum exosome biomarkers after radiation exposure (Fig. 6B).

Open in a separate window

Open in a separate window

Open in a separate window

Figure 6

Categorization and validation of of urinary and serum exosome biomarkers

Five functional categories of urinary exosome biomarkers (A) and four functional categories of serum exosome biomarkers (B) are shown. Examples of proteins in each functional category from the biomarker panel are shown. Semi-quantitative measurements of protein abundance (spectra counts) and variations among biological replicates are plotted for each protein in each group (control, 10.4 Gy 24hr, and 10.4 Gy 72hr). Nice groups of samples were collected for targeted MS analysis (C). Group 1: urine from non-irradiated mice (Control). Group 2 & 3: urine from 3 Gy WBI mice collected at 24 hours (3 Gy 24hr) or collected at 72 hours (3 Gy 72hr). Group 4 & 5: urine from 6 Gy WBI mice collected at 24 hours (6 Gy 24hr) or collected at 72 hours (6 Gy 72hr). Group 6 & 7: urine from 8 Gy WBI mice collected at 24 hours (8 Gy 24hr) or collected at 72 hours (8 Gy 72hr), Group 8 & 9: urine from 10 Gy WBI mice collected at 24 hours (10 Gy 24hr) or collected at 72 hours (10 Gy 72hr). Urine exosome was isolated from urine collected from each group of mice (method described in the Methods). Peptides from four proteins (EGF, PLAU, VDAC, HEXB) were used to acquire targeted measurements from urine exosome. The Skyline software generated the peak area of each peptide and the log value of peak area was used in the boxplot to show the variability and expression differences among these 9 groups of mice. Student t-test was used to determine if the expression level of each protein from irradiated urine exosome comparing to urine exosome from the non-irradiated controls. * above the boxplot indicates significantly altered peak area (p<0.05) in the irradiated group comparing to the control group.

The research was supported by NIH grants U19 AI091175 to CG.