Comparison of urine proteomes from tumor-bearing mice with those from tumor-resected mice

Establishment of the MC38 tumor-bearing mouse model

Male C57BL/6 mice (18 g–20 g) were supplied by Beijing Vital River Laboratory Animal Technology Co., Ltd. All animals were maintained with free access to standard laboratory diet and water with a 12-h light–dark cycle under constant temperature (22 ± 2 °C) and humidity (65–70%). The experiment was approved by the ethics and animal welfare committee of Beijing Normal University (Approval ID: CLS-EAW-2020-0344). All experiments were performed in accordance with relevant guidelines and regulations of the National Health Commission and the Ministry of Science and Technology. MC38 cells were obtained from the Cell Culture Center of Chinese Academy of Medical Sciences (Beijing, China) and were inoculated into C57BL/6 mice.

The experimental mice were acclimated in the new environment for three days. Then, they were randomly divided into three groups: a healthy control group (n = 5), a tumor-resected group (n = 5) and a tumor-bearing group (n = 5). MC38 cells were added to complete culture medium (DMEM supplemented with, 10% fetal bovine serum and 1% penicillin/streptomycin) and placed in T75 cell culture flasks. After a sufficient cell population was established, MC38 cells were collected, centrifuged, and resuspended in phosphate-buffered saline (PBS) for the subsequent establishment of the mouse models. Cell viability was assessed by the trypan blue exclusion test. Briefly, MC38 cells were stained with 0.4% trypan blue solution and then counted using a hemocytometer; tumor cell viability was greater than 95%.

The mice were anesthetized with 4 mg/kg sodium pentobarbital and placed on disposable sterile medical pads; then, hair was removed from the injection site, which was subsequently disinfected. The experimental mice in the tumor-resected group and tumor-bearing group received a subcutaneous injection of 5 × 10⁶ viable MC38 cells in 200 µl of PBS into the right hind limb. The healthy control group was injected with PBS buffer at the same location. The mice in each of the three groups underwent resection after the establishment of the tumors. The subcutaneous tumor was completely resected from the mice in the tumor-resected group; In the tumor-bearing group, the subcutaneous tumor was preserved, and small section of the tissue and muscle were removed to ensure the same wound surface in all mice; In the healthy control group, small section of the tissue and muscle were removed in all mice to control variables in this study. After the experiment, all the animals were euthanized by an intraperitoneal injection of barbiturates.

Urine sample collection and preparation

Urine was collected from mice in the three groups at Day 7 and 30 after resection. During urine collection, all mice were placed individually in metabolic cages with no food or water. Urine was collected overnight for 12 h, and the volume of urine collected was not less than 1 ml. The urine samples were stored at −80 °C immediately after collection.

Urine samples were defrosted at 4 °C and centrifuged at 12,000 × g for 30 min at 4 °C to remove cell debris. Then, the supernatants were precipitated with three volumes of ethanol at −20 °C, followed by centrifugation at 12,000 × g for 30 min. The pellet was resuspended in lysis buffer (8 mol/L urea, 2 mol/L thiourea, 50 mmol/L Tris, and 25 mmol/L DTT). The protein concentration of each sample was measured using the Bradford assay. The protein samples were stored at −80 °C for later use.

The urinary proteins were prepared using the filter-aided sample preparation (FASP) method (Wisniewski et al., 2009).Then, peptides were collected after enzymatic digestion using trypsin (Trypsin Gold, Promega, Fitchburg, WI, USA), desalted by Oasis HLB cartridges (Waters, Milford, MA, USA) and vacuum dried. The peptides were redissolved in 0.1% formic acid water and diluted to 0.5 µg/µL. A mixed peptide sample was prepared from each sample and separated using a high pH reversed-phase peptide separation kit (Thermo Fisher Scientific, Waltham, MA, USA); the resulting peptides were dried under, a vacuum and then redissolved in 0.1% formic acid in water for subsequent library construction. The iRT standard (Biognosys, Munich, Germany) was added to all identified samples for retention time normalization.

LC–MS/MS analysis

The peptides were separated by using an EASY-nLC 1200 UPLC system (Thermo Fisher Scientific, Waltham, MA, USA). Peptides were dissolved in 0.1% formic acid in water, and 1 µg of each peptide sample was loaded on a PepMap column (75 µm × 2 cm, 3 µm, C18; Thermo Fisher). The eluate was loaded onto a reversed-phase analytical column (50 µm × 250 mm, 2 µm, C18; Thermo Fisher) with an elution gradient of 4–35% mobile phase B (80% acetonitrile + 0.1% formic acid + 20% water at a flow rate of 300 nL/min) for 90 min. For fully automated and sensitive signal processing, a calibration kit (iRT kit, Biognosys, Switzerland) was used with all samples at a concentration of 1:20 v/v. Then, fractions were analyzed by data-dependent acquisition (DDA)-MS with the following settings: 2.4 kV spray voltage, 60,000 Orbitrap primary resolution, 350–1,550 m/z scan range, 200–2,000 m/z secondary scan range, 30,000 resolution, 2 Da screening window, 30% HCD collision energy. The AGC target was 5e4, and the maximum injection time was 30 ms. The raw files were used to build a dataset and were analyzed by PD software (Proteome Discoverer 2.1; Thermo Fisher Scientific, Inc.).

The PD results were used to establish the data-independent acquisition (DIA) mode, and the m/z distribution density was used to calculate the window width and number. A single peptide sample was subjected to DIA mode, and each sample was repeated twice. Thirty samples were analyzed by DIA-MS. The liquid phase parameters were acquired in the same way as for the DDA database. The MS parameters of the mass spectrometry were set as follows: first level full scan at 60,000 resolution with a 350–1,550 m/z scan range, followed by a second level scan at 30,000 resolution with 39 screening windows, 30% HCD collision energy, AGC target of 1e6, and 50 ms maximum injection time. For the window calculation: the DDA results from the library acquisition were sorted into 39 groups based on m/z. The m/z range of each group is the window width of the collected DIA data. During the sample analysis, a mixture from each sample was analyzed after every seven samples for quality control (QC).

Data analysis

MS data were processed and analyzed using Spectronaut software. The raw files acquired by DIA for each sample were imported for the library search. The false discovery rate (FDR) of the proteins was less than 1%. A highly credible protein identification was indicated by q < 0.01, and all fragment ion peak areas of secondary peptides were used for protein quantification. The k-nearest neighbor (K-NN) method was used to fill missing values of protein abundance. Comparisons between two groups were performed by Student’s t test. The differentially expressed proteins at Day 7 and 30 were screened by the following criteria: fold change ≥ 1.5 or fold change ≤ 0.67. Group differences at p < 0.05 were identified as statistically significant. Considering that omics datasets are large but the sample size is limited, the differences between two groups may be random (Meng & Gao, 2020). To confirm whether the differentially expressed proteins were indeed due to tumor resection, we randomly allocated the proteomic data for 10 samples at each time point. To determine that the differential proteins identified were due to random allocation, the same criteria for screening differential proteins were applied: fold change ≥ 1.5 or fold change ≤ 0.67 and p < 0.05. We used the Wu Kong platform (https://www.omicsolution.com/wkomics/main/) relative orthogonal signal-corrected partial least squares discriminant analysis (OPLS-DA). The differentially expressed proteins at different time points were functionally annotated using DAVID (https://david.ncifcrf.gov/), and p < 0.05 was considered to indicate statistical significance. Venn diagrams indicating the overlap of differentially expressed proteins were plotted using software PyCharm 2018.2 (Edu).

Establishment of animal models

All three groups underwent resection 7 or 8 days after tumor cell inoculation, when the tumors were palpable, and urine was collected on Days 7 and 30 after resection. The procedure is shown in Fig. 2. The animal model was successfully established, and tumor size on Day 7 or 8 is shown in Table 1. No tumor recurrence was seen 90 days after resection in the tumor-resected group; Therefore, the resection was considered successful. No abnormalities were seen in the healthy control group.

Analysis of urine proteome changes

Analysis of the results of the urine proteome at two time points for the three groups led to the identification of a total of 405 proteins in all samples. The MS proteomics data are available at iProX project PXD034552.

OPLS-DA showed a clear distinction between the tumor-resected group and tumor-bearing group at 7 and 30 days after surgery. The R2 (overall goodness of fit) and Q2 (overall goodness of prediction) values were calculated, and the results are shown in Fig. 3. On Day 7 after surgery, the R $\stackrel{ˆ}{}$X, R $\stackrel{ˆ}{}$Y and Q $\stackrel{ˆ}{}$Y values were 0.369, 0.993 and 0.788, respectively, indicating good model fit and accuracy. The variable importance in projection (VIP) score was calculated for the differentially expressed proteins, and all 20 differential expressed proteins met the criteria of VIP > 1.0. Similarly, on Day 30 after surgery, the R $\stackrel{ˆ}{}$X, R $\stackrel{ˆ}{}$Y and Q $\stackrel{ˆ}{}$Y values were 0.435, 0.968 and 0.766, respectively, indicating good model fit and accuracy. The VIP scores for all 33 the differentially expressed proteins met the criteria of VIP > 1.0. This model demonstrates the possibility of distinguishing whether MC38 tumors have been resected.

In total, the levels of 20 proteins (Table 2) were significantly different between the tumor-resected group and the tumor-bearing group at Day 7 after resection, of which nine were upregulated and 11 were downregulated. There were 125 random allocations, and the average number of differentially expressed proteins in all random combinations was 3.44. indicating that the false-positive rates were 17.2%. Six biological processes were enriched by the differentially expressed proteins on Day 7 (Fig. 4), including circadian rhythm, Notch signaling pathway, leukocyte cell–cell adhesion, and heterophilic cell–cell adhesion via plasma membrane cell adhesion molecules. In the cellular component category (Fig. 4B), all the differentially expressed proteins were enriched in the secreted and lysosome terms. In the molecular function category (Fig. 4C), serine protease and hydrolase were overrepresented.

Thirty-three differentially expressed proteins were identified between the tumor-resected group and the tumor-bearing group at Day 30 after surgery (Table 3), of which 13 were upregulated and 20 were downregulated, enriched in 19 biological processes (Fig. 5). There were 125 random allocations, and the average number of differential proteins in all random combinations was 3.76. indicating that the false-positive rates were 11.4%. Based on the data for Day 30, the biological processes on Day 30 included cell adhesion, regulation of protein localization to the cell surface, complement activation, alternative pathway, immune system processes, positive regulation of glucose metabolic processes, mitochondrion morphogenesis, and angiogenesis. In the cellular component category (Fig. 5B), the majority of the differentially expressed proteins were secreted. In the molecular function category (Fig. 5C), growth factor binding was overrepresented.

Compared to the healthy control group, there were 22 and 8 differentially expressed proteins (Tables 4 and 5) in the tumor-resected group on Days 7 and 30 after surgery. Eight differentially expressed proteins on Day 30 were not sufficient for enrichment analysis of biological processes. There were 125 random allocations, and the average number of differentially expressed proteins in all random combinations was 3.77, indicating a false-positive rate of 47%. These data suggest that the difference in the physiological condition of mice in the tumor-resected group compared to that of mice in the healthy control group was smaller after 30 days of recovery.

The overlap of these differentially expressed proteins is shown in the Venn diagram in Fig. 6. The six proteins were significantly different between the tumor-resected group and the tumor-bearing group at both Day 7 and Day 30, including vascular cell adhesion protein 1, basement membrane-specific heparan sulfate proteoglycan core protein, resistin-like alpha, guanylate cyclase activator 2B, lymphocyte antigen 6C1, and alpha-amylase 1. The changes of these six proteins were consistent at two time points. The vascular cell adhesion protein 1, basement membrane-specific heparan sulfate proteoglycan core protein, resistin-like alpha, guanylate cyclase activator 2B, and lymphocyte antigen 6C1 protein in the urine of the tumor-resected group decreased, while the alpha-amylase 1 protein increased. Five proteins were at identical levels in the tumor-resected group and tumor-bearing group and in the tumor-resected group and health control group at Day 30, including CD59A glycoprotein, lymphocyte antigen 6C1, fibronectin, basement membrane-specific heparan sulfate proteoglycan core protein, keratin, type II cytoskeletal 5.