Informed consent was obtained for experimentation with human subjects. This study was approved by the Ethical Committee of Xinhua Hospital affiliated with the Shanghai Jiaotong University School of Medicine (No. XHEC-D-2022-201). Urine samples from 10 patients with pathologically confirmed EOC and 10 healthy female subjects was collected. Patients with other medical or surgical disorders were excluded from the study. A total of 114 EOC primary tumor samples and 67 metastatic tissues were used for immunohistochemistry and survival analysis. All the samples were collected from January 2012 to December 2021 in the Xinhua Hospital.

Using urine samples from ovarian cancer patients and healthy female subjects, exosomes were obtained and purified via ultracentrifugation, which is considered to be the gold standard for exosome isolation [20, 21]. The procedure was performed as described below. First, urine samples were centrifuged at 300 $\times$g for 10 min and the pellets discarded. The supernatant fraction was recentrifuged at 2000 $\times$g for 10 min, followed by 10,000 $\times$g for 30 min at 4 °C. The supernatant produced from this step was subsequently ultracentrifuged at 100,000 $\times$g for 70 min to pellet exosomes. Finally, exosomes were suspended in 50 $\mu$L phosphate-buffered saline (PBS, SH30256-01, Hyclone, Logan, UT, USA) and validated via nanoparticle tracking, transmission electron microscopy, and immunoblot analyses.

After determination of protein concentrations with the BCA kit (P0012, Beyotime, Shanghai, China), 30 $\mu$g exosomal proteins were prepared for MS analysis, which was performed as previously described [22]. In the study, all the proteins from 10 ovarian cancer patients’ exosomes were mixed as one group, the rest of the exosome proteins from healthy women were mixed as another group. Gene ontology (GO) analyses was further conducted based on MS data.

Ovarian cancer cell lines SKOV3 and HEY were obtained from the American Type Culture Collection (ATCC, USA). The ovarian cancer cells were authenticated by short tandem repeat (STR) DNA profiling analysis and screened routinely for mycoplasma contamination. Cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) at 37 °C under 5% CO${}_2$. The complete culture media contained DMEM (11995065, Thermo Fisher Scientific, Waltham, MA, USA) and 10% fetal bovine serum (FBS, SV30208.02, Cytiva, Logan, UT, USA).

Silencing of LRG1 (experimental group) was achieved using shRNA technology. Lentivirus carrying LRG1 shRNA was obtained from Bio-Link biological technology company (Shanghai, China). LRG1 shRNA with the following sequences was used: 5${}^{\prime }$-GCTGCAATTAGAACGGCTACA-3${}^{\prime }$. Lentivirus was used for subsequent transduction and the results of LRG1 silencing were validated via immunoblot analysis. Scramble shRNA was used as the control group in vitro and in vivo.

For immunoblot analysis, ovarian cancer SKOV3 and HEY cells were lysed on ice in radioimmunoprecipitation assay (RIPA) lysis buffer (P0013B, Beyotime, Shanghai, China). The acquired proteins were quantified with the BCA protein assay kit (P0010, Beyotime, Shanghai, China), followed by denaturation, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transmembrane and blocking. Membranes were incubated overnight with primary antibodies (1:1000 dilution) at 4 °C and washed with 1$\times$ Tris-buffered saline with Tween 20 (TBST) three times for 10 min, followed by incubation with HRP-conjugated anti-rabbit or anti-mouse secondary antibody (1:2000 dilution; 7074, 7076; Cell Signaling Technology, Boston, MA, USA). Bands were visualized using Immobilon Western Chemiluminescence HRP substrate (WBKIS0100; Millipore, Boston, MA, USA). Primary antibodies in this study were as follows: LRG1 (ab178698, Abcam, Boston, Cambridge, UK), p-FAK (Tyr397, ab81298, Abcam), p-AKT1 (Ser473, 9018, Cell Signaling Technology) and $\beta$-actin (3700, Cell Signaling Technology). The quantification of the protein bands was performed using image J software version 1.52 (National Institutes of Health, Montgomery, MD, USA).

In order to detect the function of LRG1 in regulating ovarian cancer cell migration, transwell assays were performed. Ovarian cancer cells (5 $\times$ 10${}^4$ cells/chamber) suspended in medium without FBS were seeded into the upper chamber. Then, 500 $\mu$L complete medium (containing FBS) was added to the lower chamber. After 48 h, cells were treated with 70% methanol for 30 min and subsequently stained with 0.1% Crystal Violet for 30 min. The migration ability was assessed by the relative number of the migrated ovarian cancer cells.

The wound healing assay was conducted as follows: ovarian cancer cells were seeded into 6-well plates (1 $\times$ 10${}^6$ cells/well) at 37 °C; a scratch was cautiously created with a 20 $\mu$L pipette tip after 24 h; cells were further cultured in medium without FBS. Finally, images were obtained after 48 h.

After fixation with 4% paraformaldehyde (E672002; Sangon Biotech, Shanghai, China), permeabilization with 0.3% Triton X-100 (A600198, Sangon Biotech) and blocking with 1% bovine serum albumin (BSA, A2153; Sigma, Saint Louis, MO, USA), cells or tissues were incubated with LRG1 (13224-1-AP, Proteintech, Chicago, IL, USA), p-FAK (Tyr397, ab81298, Abcam), p-AKT1 (Ser473, 9018, Cell Signaling Technology) antibodies at 4 °C overnight with 1:1000 dilution with PBS. Following three washes with PBS (SH30256-01, Hyclone, Logan, UT, USA), samples were treated with the corresponding secondary antibodies. Finally, images were obtained with the aid of confocal microscopy (Leica, TCS SPS, Wetzlar, Hesssian, Germany).

Following conventional dewaxing, endogenous peroxidase elimination, microwave repair and BSA blocking, human ovarian cancer tissues were incubated in LRG1 antibody (1:100 diluted with PBS, ab178698, Abcam) at 4 °C overnight. Next, sections were treated with secondary antibody, followed by immunohistochemical staining and alcohol gradient dehydration. Stained tissue slides were finally visualized by diaminobenzidine (DAB, G1212, Servicebio, Wuhan, China). The expression level of LRG1 was assessed by histochemistry score (H-score), which was determined as multiplicity of staining intensity and percentage of positively stained cells. The quartile of the multiplicity score was set as the boundary between low and high expression. The staining intensity was classified as negative (0), weak (light yellow, 1), moderate (brownish, 2), and strong (brown, 3); the proportion of positive cells was scored as 0 to 5% (0), 5% to 25% (1), 26% to 50% (2), 51% to 75% (3), and 76% to 100% (4). LRG1, p-FAK (Tyr397) and p-AKT1 (Ser473) were analyzed by the same method in murine tumor tissues.

Animal experiments were approved by the Ethics Committee of Xinhua hospital (No. XHEC-F-2022-020). Ten four-week-old female BALB/c nude mice were obtained from the Animal Experiment Center of Xinhua hospital. The mice were housed in cages at 25 °C and 50% humidity with a 12 h light/dark cycle, and provided with sufficient food and water in the Animal Experiment Center of Xinhua hospital. The water, food, cages, and other items in contact with the mice were sterile and treated by aseptic technique within a certified biosafety cabinet. The 10 mice were divided randomly into two groups (two cages), and every cage contained 5 mice. Ovarian cancer transplanted tumor models were generated by subcutaneous injection of ovarian cancer cells into the back of nude mice. Five mice living in the same cage were injected with 5 $\times$ 10${}^6$ SKOV3/Control cells separately, and the others were treated with 5 $\times$ 10${}^6$ SKOV3/LRG1-shRNA cells. The tumors were measured twice a week. Tumor volumes were calculated using the following formula: V (volume) = L (length) $\times$ W (width) $\times$ 0.52. The mice were euthanized with carbon dioxide (CO${}_2$) after 6 weeks, and the tumors were subsequently dissected, weighed and imaged. If the mice were in a state of obvious cachexia, or the diameter of the tumor ulcer was greater than 0.5 cm, they were euthanized earlier.

Statistical analysis was conducted using the Prism8 program (GraphPad software, San Diego, CA, USA). The two sets of data were compared using Student’s t-test, with p values 0.05 (*) considered statistically significant. Data are expressed as mean $\pm$ standard deviation. Kaplan-Meier curves of progression-free survival (PFS) and overall survival (OS) were generated via the log-rank test.

Given the critical role of exosomes as sources of tumor biomarkers, urine samples were obtained from 10 EOC patients and 10 healthy females. Exosomes were extracted via conventional ultracentrifugation. Exosomes were verified using three classic methods of identification, specifically, transmission electron microscopy, nanoparticle tracking analysis and immunoblot assay (Fig. 1A–C). It indicated that exosomes were mainly distributed in the range of 50–200 nm (Fig. 1A). To find key molecular markers of ovarian cancer, exosomal proteins were extracted and subjected to MS, which led to the identification of 442 differentially expressed proteins (Fig. 1D). Interestingly, there were 56 proteins detected only in tumor samples such as LRG1, SERPINA3, C4BPA, and so on (Supplementary Table 1). Based on GO analysis of the MS results, we verified that exosome proteins were significantly enriched (Fig. 1E). Moreover, LRG1 was specifically expressed in exosomes from the ovarian cancer group with the greatest intensity. In addition, we found that increased expression of LRG1 was associated with poorer prognosis in terms of both OS and PFS (Fig. 1F,G), suggesting that LRG1 might be an oncogene in ovarian cancer. Subsequent EOC tissue microarrays showed that the metastatic lesions contained higher levels of LGR1 than the primary ovarian cancer tissue (Fig. 1H,I). Moreover, we found that LRG1 was higher expressed in the metastatic lesions compared with the corresponding primary tumors (Fig. 1J). These findings support the utility of LRG1 or exosomal LRG1 as a biomarker for ovarian cancer diagnosis and prognosis and its potential role as a mediator in the process of cell metastasis.

Fig. 1.

Exosomal LRG1 is specifically expressed in urine samples of ovarian cancer patients and LRG1 was associated with the poor prognosis. (A) Urine-derived exosome concentrations and size distributions determined via nanoparticle tracking analysis. (B) Representative transmission electron microscopy micrograph of exosomes derived from urine samples. (C) Immunoblot analysis of CD63, HSC70 and Alix in exosomes from urine samples. (D) Venn diagram of mass spectrometry (MS) analysis of exosomal proteins derived from urine of ovarian cancer patients and healthy subjects. (E) Results of gene ontology (GO) analysis after MS assay. (F,G) Survival analysis of overall survival (OS) and progression-free survival (PFS) in 62 LRG1 high-expressing and 52 LRG1 low-expressing samples. (H) LRG1 expression was examined by immunochemistry assay in ovarian cancer tissue microarrays (n = 114). (I) LRG1 expression was examined by immunochemistry assay in the metastatic lesions (n = 67). (J) Paired analysis of the 67 metastatic lesions and the corresponding primary tumors. **p 0.01.

To further verify the role of LRG1 in ovarian cancer metastasis, we performed in vitro experiments. First, the expression patterns of LRG1 in different ovarian cancer cell types were examined via immunoblotting. Cell lines displaying higher LRG1 expression were selected for subsequent LRG1 silencing study (Fig. 2A). Next, LRG1 shRNA were introduced into ovarian cancer SKOV3 and HEY cells to establish LRG1-silenced cell lines (Fig. 2B). Transwell and wound healing assays were performed to identify migration changes caused by LRG1 silencing. The migration cell count was significantly inhibited by 47.83% in SKOV3/LRG1-shRNA cells and by 48.10% in HEY/LRG1-shRNA cells (Fig. 2C,D). The migration range of SKOV3 cells was reduced by 64.07% in the LRG1-shRNA group and by 37.48% in HEY cells compared with the control (Fig. 2E,F). Our results demonstrated that LRG1 modulated the process of migration in ovarian cancer cells.

Fig. 2.

LRG1 participates in migration of ovarian cancer cells. (A) Immunoblotting analysis of LRG1 in various ovarian cancer cell types. (B) The effect of LRG1 silencing was tested by immunoblotting analysis. (C) Transwell migration assay was performed to assess the migration ability induced by LRG1 silencing. (D) Statistical results of the transwell migration assay. (E) Wound healing assay was utilized to measure the migration ability of ovarian cancer cells. (F) Statistical results of the wound healing assay. The experiments were technically repeated three times. Values were expressed as mean  $\mathrm{\pm }$  SD and statistical analyses were performed using Student’s t-test, **p 0.01.

While the above results indicated that LRG1 modulated migration of ovarian cancer cells, the underlying mechanisms remained to be established. Immunoblotting and immunofluorescence assays were performed, and FAK and AKT1 were found to be expressed at low phosphorylated levels in LRG1 silencing ovarian cancer cells (Fig. 3A–C), though their total levels were seldom changed. The activated phosphorylation of FAK (Try397) and AKT (Ser473) played important roles in tumor metastasis [23, 24]. In view of these findings, we concluded that LRG1 may be involved in metastasis formation through activation of the FAK/AKT1 pathway.

Fig. 3.

LRG1 regulates FAK/AKT axis. (A) Immunoblotting assay of phosphorylation of FAK (Tyr397) and AKT1 (Ser473). The experiments were technically repeated three times and the quantification of the protein bands was performed using image J software. (B) Immunofluorescence assay of LRG1 and p-FAK (Tyr397). (C) Immunofluorescence assay of LRG1 and p-AKT1 (Ser473). 4′,6-diamidino-2-phenylindole (DAPI) was used to stain cell nuclei. **p 0.01.

To further validate the function of LRG1 in ovarian cancer, in vivo experiments were performed on 4-week-old female BALB/c nude mice. In a subcutaneous tumor model, LRG1 knockdown markedly inhibited tumor growth velocity (Fig. 4A,B) along with tumor weight (Fig. 4C). After tumors were dissected, the tissues were subjected to immunohistochemistry and immunofluorescence. Both of these two methods demonstrated that LRG1 silencing inhibited the expression of p-FAK (Try397) and p-AKT1 (Ser473) (Fig. 4D–F). The data obtained from mice tumors suggested that LRG1 participates in ovarian cancer progression by regulating the FAK/AKT axis in vivo.

Fig. 4.

LRG1 participated in ovarian cancer progression in vivo. (A) Subcutaneous tumor model of ovarian cancer. (B) Tumor growth curve indicated the changes of tumor volumes after SKOV3 cells were implanted subcutaneously. (C) Tumors were weighed after they were dissected from the mice bodies. (D) Immunohistochemistry analysis of LRG1, p-FAK (Tyr397) and p-AKT1(Ser473) in nude mice subcutaneous tumor model. (E) Immunofluorescence assay of LRG1 and p-FAK (Tyr397). (F) Immunofluorescence assay of LRG1 and p-AKT1 (Ser473). DAPI was used to stain cell nuclei. **p 0.01.