Trace Analysis of Multiple Tumor Exosomal PD‐L1 Based on SERS Immunoassay Platform

Immunotherapy has received wide attention in recent years as a new avenue for the effective treatment of cancer. However, due to the lack of detection limit and sensitivity of immune checkpoint molecule (such as Programmed cell death 1 ligand 1, PD‐L1) in established clinical methods, the immunotherapy evaluation during anti‐PD‐L1/PD‐1 (Programmed cell death receptor 1) treatment is difficult to be guided accurately. In this study, a highly sensitive and maneuverable Surface‐Enhanced Raman Scattering (SERS)‐based immunoassay platform is developed for the analysis of exosomal PD‐L1, a highly potential biomarker for immunotherapy. Excellent detection of exosomal PD‐L1 with good linear fit over a wide concentration range is achieved. In addition, the detection and discrimination of exosomal PD‐L1 in the peripheral blood of cancer patients and healthy controls are successfully achieved. Moreover, the platform has also been successfully used to distinguish common diseases, cancer, and healthy control (such as liver cancer, liver cirrhosis, and normal individuals). The detection platform can be successfully used for the trace detection of PD‐L1 on exosomes, which has excellent potential for clinical development in cancer diagnosis and treatment guidelines.


Introduction
Cancer is one of the major global public health problems. Early diagnosis and treatment of cancer will give patients more chances of survival. Surgical tissue biopsy, considered the current gold standard for cancer diagnosis, only provides a snapshot of the disease and is difficult to obtain in some cases. [1] And the DOI: 10.1002/adsr.202200043 heterogeneity of tumors leads to differences in the diagnosis of patients by immune histochemistry, which results in ineffective treatment. [2] In addition, the treatment requires the patient to endure pain during multiple biopsies. Therefore, there is an urgent need to identify more efficient and less invasive diagnostic modalities to guide early diagnosis, development of treatment strategies for individual patients, and accurate estimation of prognosis. Recently, the detection of tumor markers in blood and even body fluids such as urine plays an increasingly great part in the diagnosis of cancer. These biomarkers include primary tumor-derived entities or products such as circulating tumor cells, circulating tumor DNA, and exosomes, which are regarded to play an important part in tumor development and progression. [3] Among them, exosomes as a small vesicles released into the surrounding humoral environment by cells (including tumor cells), have attracted much attention due to the following characteristics:1)The interior and surface of exosomes contain many information related to tumor cells such as proteins, lipids and mRNAs. [4] 2) Small volume, easier to enter the blood circulation through tissue barriers, therefore easier to be detected in body fluids. [5] 3) The lipid bilayer membrane structure of exosomes protects internal proteins and DNA from degradation by enzymes in blood circulation. [6] 4) Compared with the characteristic proteins secreted by tumors, exosomes are easier to be separated, purified and detected in the early stage of cancer, which is suitable for early detection of cancer. [7] Therefore, exosome-based liquid biomarker (exosomal PD-L1 or RNA) has emerged as a trending diagnostic option for non-invasive tumor progression monitoring and treatment response assessment. [8] Programmed death ligand 1 (PD-L1) as a type I transmembrane protein, can inhibit the multiplication and differentiation of T cells after interacting with programmed death receptor 1 (PD-1) on the surface of T cells, and inhibit the cellular immune response of the matrix to a certain extent. [9] Most cancer cells use the above process to achieve immune escape by upregulating the expression of PD-L1. Based on this, researchers have developed anti-PD-L1/PD-1 immunotherapy. Immune checkpoint protein inhibitors antibodies against PD-L1 and PD-1 have shown efficacy against variety of cancer such as lung cancer, and renal cancer even terminal cancer. Nevertheless, only 20-40% of patients show good response to immunotherapy, with little effect on patients like prostate cancer. [10] Recently, it has been reported that the surface and interior of exosomes contain tumor-derived PD-L1 protein, of which the expression level is significantly correlated with disease stage, indicating that the PD-L1 on exosomes may be a potential marker for monitoring disease and evaluating immunotherapy. [11] A series of analytical assay methods have been used for the test of PD-L1 on exosomes. However, traditional exosomal protein tests required super centrifugation followed by Western blotting or enzyme linked immunosorbent assay (ELISA) molecular characterization. Although the most broadly used gold standard in traditional protein detection, Western blotting is often limited by its long preprocessing time (>48 h) and relatively large number of plasma samples (>2 mL) and used for semi-quantitative analysis. For ELISA, which uses a couple of antibodies to improve specificity of test, is difficult to perform high-throughput detection at the same time. In addition, the detection limit of ng mL −1 of Western blotting and ELISA limits their usefulness in early cancer diagnosis. [12] In comparison with Western blotting and ELISA, biosensors exhibit higher sensitivity, for example integrated magnetic electrochemical exosome sensors, thermophoretic profiling and CRISPR/Cas13. [13] However, the complex and cumbersome operation and expensive equipment of these platforms limits their clinical applications. Herein, the research on the detection method of PD-L1 on exosomes still requires to be further deepened to develop suitable methods for clinical diagnosis and treatment.
Surface-Enhanced Raman Spectroscopy (SERS) is widely used in disease diagnosis, microbial detection and soil heavy metal detection due to its advantages of easy operation, short timeconsuming, high detection sensitivity, low detection limit, and the ability to provide fingerprint information of analytes. [14] SERS nanotags which the signal of Raman molecular with large Raman scattering cross sections is greatly enhanced by noble metal nanoparticles such as gold and silver is proved to be good label candidates. [15] Compared with fluorescent tags, SERS tags whose signals originate from the vibration of molecular chemical bonds are more photostable, and have a narrow bandwidth of the characteristic peaks (usually less than 2 nm). [16] Currently, many SERS tags have developed as powerful tools in biological applications. Muhammad and co-workers utilized ultrasensitive aptamer-functionalized SERS tags to monitoring of circulating exosomal immune checkpoint, and this method reached a detection limit of 4.31 ag mL −1 for the detection of PD-L1 protein on exosomes. [17] Pang reported a simple method to detect exosomal PD-L1 from serum, which can capture 96.5% of exosomes in the sample by Fe 3 O 4 @TiO 2 nanoparticles and achieve a detection limit of one PD-L1 exosome μL −1 by Au@Ag@MBA SERS tags. [18] These study demonstrate the superior performance of SERS tags in detecting exosomal PD-L1, providing a feasible and effective method for evaluating the efficacy of anti-PD-L1/PD-1 treatment.
Herein, we present a SERS-based immunoassay platform for monitoring of PD-L1 protein on exosomes. By testing a series of standard concentrations of PD-L1 and exosomal PD-L1, it is verified that the Au NPs@DTNB@anti-PD-L1 SERS tags have good detection and semiquantitative analysis capabilities. The LOD of purified PD-L1 was 0.1 ng mL −1 , and exosomes based on PD-L1 was 4.8 × 10 6 particles mL −1 lower than that of traditional standard detection methods like Western blotting and ELISA. Further detect of exosomal PD-L1 from the serum of liver cancer patients and healthy controls indicated the practicability of clinical cancer diagnosis by detecting the Raman intensity at the 1328 cm −1 from tags. This platform has also been successfully applied to the diagnosis of other types of cancer: lung cancer, breast cancer, rectal cancer, colon cancer, and renal cancer. In addition, we also apply this method to detection of exosomal PD-L1 in liver cirrhosis patients, and successfully distinguish normal persons, liver cirrhosis patients and liver cancer patients. Overall, a simple, sensitive and high specificity platform has been developed to diagnose several kinds of cancer and liver cirrhosis through Au NPs@DTNB@anti-PD-L1 SERS tags monitoring the exosomal PD-L1, which has great application potentials in cancer clinical diagnosis and immunotherapy guidance.

Detection Principle and Material Characterization
Here, an exosomes capturing platform composed of CD63 antibody binding to Au film was constructed that can recognize and combine the exosomes in blood by immunoreaction between antibody and CD63 on exosomes, which is among the most frequently expressed and common proteins on the vast majority of exosomes. [17] In order to identify PD-L1 protein on exosomes captured in Au film, a SERS tag was constructed by functionalizing PD-L1 antibody on 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) Raman reporter modified Au NPs (Au NPs@DTNB@anti-PD-L1). As shown in Figure 1a, the exosomes were captured in Au film by CD63 antibody, and then the Au NPs@DTNB@anti-PD-L1 SERS tags were introduced that could recognize and combine the PD-L1 on exosomes via antigen-antibody reaction. This resulted in a sandwich-like structure. Therefore, the semiquantitative detection of exosomal PD-L1 was possible via analyzing this sandwich structure produced strong Raman signals of DTNB. The peak assignments of DTNB are shown in Table S1 (Supporting Information).
Considering that the exosomes are very small extracellular vesicles, typically 30-120 nm in diameter, it is difficult to achieve the detection of the exosomal PD-L1 by golden nanoparticles with a large particle size. [19] Therefore, ultra-small gold crystal were used as golden nanoparticles SERS substrates. As shown in the TEM and the size distribution of Au NPs (Figure 1b, Figure S1, Supporting Information), the dimensions of Au NPs were 4.72 ± 1.06 nm. As shown in Figure 1c,d, a significant red-shift of the absorption peak of Au NPs changed from 510 to 568 nm after modified, and zeta potentials of Au NPs also shown noticeable changes from −8.8 to −25.4 mV of Au NPs and Au NPs@DTNB, which indicate that DTNB molecule was successfully combined the Au NPs by Au─S bond. Moreover, zeta potentials of Au NPs@DTNB changed from −25.4 to −15.6 mV after the PD-L1 antibody modified, indicated that PD-L1 antibody was successfully combined with Au NPs@DTNB. To verify the feasibility of the detection platform, the exosomal PD-L1 were firstly detected. As shown in Figure S2 (Supporting Information), the sandwich structure formed when the exosomes are present. At this time, there was an obvious Raman signal at 1328 cm −1 , which was assignment to symmetric stretch of the N-O nitro group of DTNB.

Detection of PD-L1 Protein
To verify the quantity of Au NPs@DTNB@anti-PD-L1 SERS tags for PD-L1 detection, this platform was employed to detect a series of PD-L1 with different concentrations into PBS buffer. Figure 2b shows the Raman spectra for various concentrations of PD-L1. Their concentration varied from 0 to 10 −4 g mL −1 . In the absence of PD-L1, weak Raman signals at 1328 cm −1 were observed owing to the formation of few sandwich structures like Figure 2a.
Hence, most Au NPs@DTNB@anti-PD-L1 SERS tags in the supernatant solution were removed during washing. However, with the increase of PD-L1 concentration, the number of sandwich structures increased and the Raman intensity of 1328 cm −1 corresponding increased ( Figure 2c). As shown in Figure S3 (Supporting Information), the western blotting analysis results of different concentrations of PD-L1. The detection limit of PD-L1 on the SERS platform was 0.1 ng mL −1 , much lower than that of western blotting. The quantity of PD-L1 results in Figure 2d indicate a good linearity between the peak intensity of 1328 cm −1 and the concentration of PD-L1 (R 2 = 0.96713), which provides the feasibility for the following semiquantitative assay of PD-L1 in clinical serum exosomes.

Detection of PD-L1 on Exosomes
Additionally, to evaluate the performance of detection of the exosomal PD-L1, this platform was utilized for the detection of a series of exosomes with different concentrations. As shown in Figures S4 and S5 (Supporting Information), the size of exosomes we collected was 127.7 nm and the original concentration was 4.8 × 10 11 particles mL −1 . After gradient dilution in PBS buffer, a series of exosome solutions of different concentrations are finally obtained, varied from 4.8 × 10 6 to 4.8 × 10 11 particles mL −1 . Figure 3b showed the Raman spectra for various concentrations of exosome. Also, the Raman signals at 1328 cm −1 gradually decreased with the decrease of exosomes, owing to the formation of few sandwich structures like Figure 3a. As shown in Figure 3c, the LOD of exosomes was 4.8 × 10 6 particles mL −1 . Figure 3d shows the good linearity between the peak intensity at 1328 cm −1 and the concentrations of exosomes (R 2 = 0.96262). All these results demonstrate that the sandwich platform has excellent performance in the detection of PD-L1 on exosomes.

Versatility of Cancer Diagnostic through SERS Immunoassay
The content of PD-L1 protein in tumor-derived exosomes is different from normal cell derived exosomes. Therefore, diagnosis of cancer is possible by analyzing PD-L1 on exosomes. To assess the feasibility of cancer diagnosis, exosomal PD-L1 was detected from serum of normal individuals and different cancer patients by Au NPs@DTNB@anti-PD-L1 SERS tags. Here, several cancers with high global cancer incidence and high mortality rate were selected for exosomal PD-L1 detection, including liver cancer, lung cancer, breast cancer, rectal cancer, colon cancer, and renal cancer. [20] Figure 4a shows the typical Raman spectra for normal individuals and different cancer patients. It was obvious that all types of cancer have obvious signals from SERS tags compared with normal person, attributed to tumor cells secreting exosomes with higher PD-L1 expression than normal cells. To prove the reliability of the experiment, exosomal PD-L1 extracted from 10 healthy controls (number, n = 10) and 6 patients with liver can-cer (n = 6) randomly selected from Zhongnan Hospital (Wuhan, China) were further detected. Figure 4b,c shows the peak intensity at 1328 cm −1 of exosomal PD-L1, respectively. It was obvious that all Raman spectra measured from healthy controls have lower peak intensity at 1328 cm −1 . Meanwhile, the Raman spectrum of exosomal PD-L1 of liver cancer patients had a high level at 1328 cm −1 , which were much higher than that of normal people. It is consistent with the conclusion that tumor-derived exosomes have high content of PD-L1 protein. It showed that detection of PD-L1 on exosomes by Au NPs@DTNB@anti-PD-L1 SERS tags can successfully distinguish liver cancer patients with normal individuals. In order to further prove that the SERS platform had higher sensitivity than the clinical detection methods, exosomal PD-L1 extracted from 6 patients with liver cancer (n = 6) and 10 healthy controls (number, n = 10) randomly selected from Zhongnan Hospital were further detected by Western blot. As shown in Figure S6 (Supporting Information), Western blotting cannot always detect exosomal PD-L1 in the serum of patients with liver cancer, and there are certain false negatives. However, this situation was almost absent in the SERS platform, which showed that it had a higher sensitivity. To further verification of universality of diagnose cancer by detecting PD-L1 on exosomes using Au NPs@DTNB@anti-PD-L1 SERS tags, this platform was used to detect the PD-L1 on exosomes extracted from 8 lung cancer patients (n = 8), 7 breast cancer patients (n = 7), 6 rectal cancer patients (n = 6), 2 colon cancer patients (n = 2) and 2 renal cancer patients (n = 2) randomly chosen from Zhongnan Hospital. Figure S7a-e (Supporting Information) shows the peak intensity at 1328 cm −1 of exosomal PD-L1 of these cancer patients, respectively. In summary, as shown in Figure 4d, the peak intensity at 1328 cm −1 of all types of cancer tested were higher than normal, and significant thresholds existed between cancer samples and healthy controls. The results showed that diagnose cancer by detecting exosomal PD-L1 using Au NPs@DTNB@anti-PD-L1 SERS tags have excellent performance in cancer diagnosis.

Detection of Exosomal PD-L1 of Liver Cirrhosis Patients
Liver cirrhosis is the final state of chronic liver disease and one of the main causes of death of patients. Recently, studies have shown that liver cirrhosis is characterized by a decline in the amount of liver macrophages and a remarkable impairment of macrophages, the PD-L1 plays an important role in this process. [21] PD-L1 was considered as a novel marker for predicting the degree of liver cirrhosis. Therefore, Au NPs@DTNB@anti-PD-L1 SERS tags were further used to detect the PD-L1 on exosomes extracted from liver cirrhosis patients (n = 6) randomly chosen from Zhongnan Hospital. Figure 5a shows the Raman spectra for normal individuals, liver cirrhosis, and liver cancer. It was obvious that the Raman intensity at 1328 cm −1 of liver cancer patients was the highest, followed by liver cirrhosis patients, and that of normal individuals was the lowest. Figure 5b showed the peak intensity at 1328 cm −1 of 6 liver cirrhosis patients. Then, the peak intensity at 1328 cm −1 measured by normal individuals, liver cirrhosis patients and liver cancer patients were counted and presented in form of box plot (Figure 5c). The results showed that the content of PD-L1 on exosomes in patients with liver cirrhosis was higher than that of normal people, but lower than that of liver cancer patients. In addition, it shows that the test of PD-L1 on exosomes by Au NPs@DTNB@anti-PD-L1 SERS tags is feasible to distinguish liver cancer, liver cirrhosis and normal individuals. It is important to note that effective differentiation between common diseases and cancer is essential for the early and effective treatment of cancer, which plays a decisive role in reducing cancer mortality.

Conclusion
In summary, SERS immunoassay detection platform was successfully developed for the rapid detection of a wide range of clinical cancers and effective screening for common diseases/cancer based on exosomal PD-L1. To verify the feasibility and quantitative analysis performance of the detection platform, it was used  Raman spectra of exosomal PD-L1 from serum of healthy controls, liver cirrhosis and liver cancer. a) Typical Raman spectra of exosomal PD-L1 separated from serum of healthy controls, liver cirrhosis patients and liver cancer patients. b) Peak intensity at 1328 cm −1 from plasma of liver cirrhosis (n = 6), black dashed line represents the average of peak intensity at 1328 cm −1 from plasma of 6 liver cirrhosis. c) Box plot illustrating the Raman intensity at 1328 cm −1 of normal individual, liver cirrhosis, and liver cancer. The error bars represent the standard deviations from four measurements. *p < 0.05; **p < 0.01; ***p < 0.001. The error bars represent the standard deviations from four measurements.
to detect a series of standard concentration of PD-L1 and exosomal PD-L1. The results showed that the detection limit of PD-L1 was 0.1 ng mL −1 and exosomes based on PD-L1 was 4.8 × 10 6 particles mL −1 , which lower than that of traditional standard detection methods like Western blotting. And good linear fit between peak intensity at 1328 cm −1 and accordingly concentration. R 2 for test of PD-L1 and exosomal PD-L1 were 0.96713 and 0.96262, respectively. The diagnosis of healthy controls and patients with different cancer was further achieved by sandwich SERS platform. And the peak intensity at 1328 cm −1 of cancer patients (including 6 liver cancer patients, 8 lung cancer patients, 7 breast cancer patients, 6 rectal cancer patients, 2 colon cancer patients, and 2 renal cancer patients) tested were great higher than normal individuals. Interestingly, the platform could also be used for effective screening of common diseases. The expression of exosomal PD-L1 in serum of liver cirrhosis patients were obvious differences from serum of normal individuals, and liver cancer. Therefore, the diagnostic platform proposed in this study has good semiquantitative analysis of exosomal PD-L1, which can be effectively used for dynamic monitoring of PD-L1 during immunotherapy and provide valuable guidance for clinical cancer treatment.
Preparation of Gold Nanoparticles: Gold nanoparticles were prepared according to the previous procedure. [22] Briefly, tetrachloroaurate (III) trihydrate (HAuCl 4 ·3H 2 O) (0.05 m) was dissolved in ultrapure Millipore water (18.2 MΩ) and the HAuCl 4 solution (25 μL) weas then mixed with deionized water (9.7 mL) under vigorous stirring. Then trisodium citrate solution (0.25 mL, 0.01 m) and sodium borohydride solution (0.05 mL, 0.01 m) were added in sequence under stirring. The change in the color of the solution from light yellow to wine red was observed immediately. Finally, the solution was incubated overnight at room temperature.
Preparation of Antibody-Conjugated SERS Nanotags: Dithiobis(succinimidylpropio nate) solution (0.1 mL, 100 mM) and 5,5′-Dithiobis-(2-nitrobenzoic acid) solution (0.5 mL, 100 mm) in DMSO were added in Au NPs solution in sequence under stirring. The color of the solution changed from wine red to pale purple and then to dark purple was observed immediately. The solution was then incubated one day at room temperature. Next, the Au NPs@DTNB colloid was centrifuged twice at 13 000 rpm for 15 min to remove excess DSP and DTNB molecules. It was then dispersed in PBS buffer (2 mL). Subsequently, anti-PD-L1 antibodies (20 μg) were mixed to the solution with 1 h incubation time. Then the solution was centrifuged at 13 000 rpm for 15 min to remove the residual antibody and washed four times with PBS buffer. After that, add 0.1% BSA solution to mix well in the colloid and incubate for 1 h at room temperature. Finally, the anti-PD-L1 antibody-conjugated SERS nanotags were dispersed in PBS buffer and kept at 4°C for the future use.
Preparation of Au Film and Antibody Modification: The gold film was prepared by coating 25 nm Cr and 50 nm Au on a clean silicon wafer by thermal evaporation. Cutting the gold film into the size of 2 × 2 mm 2 and then wash with ethanol and water. After that, these Au films were soaked in 1 mmol DSP in DMSO solution for 1-2 h. Then, immerse the gold film in 10 μg mL −1 CD63 antibody solution and incubate it on shaking table. Washing four times with PBS buffer after 1 h. Add 1% BSA solution and seal it at room temperature for 1 h to obtain CD63 antibody modified fold film.
Separation of Exosomes from Plasma: All of the serum samples were from Zhongnan Hospital of Wuhan University, China. The specific operation of exosomes extraction is as below. Place about 1.5 mL collected peripheral blood in an enzyme-free EP tubes and centrifuge 2000 g at 4°C for 20 min. The upper plasma was transferred to another EP tubes and centrifuge 10 000 g at 4°C for 20 min. Take 300 μL of its supernatant, then adding PBS buffer (150 μL) and proteinase K (30 μL) into supernatant in turn, mix them well and place them in 37°C incubator for 10 min. Further add 90 μL exosome precipitation to the above mixture, mix well and place in refrigerator at 4°C for 30 min. Then the solution was centrifuged at 10 000 g for 5 min at 4°C to obtain exosomes. Finally, the exosomes were suspended in PBS buffer (100 μL) for use or stored at −80°C.
Raman Detection of Clinical Plasma Exosomal PD-L1 by SERS Tags: Add 50 μL exosomes to a microplate with gold film and incubate for 1 h at room temperature on shaker. Then washed with PBS buffer three times, added 50 μL Au NPs@DTNB@anti-PD-L1 SERS tags and incubated on shaker for 1 h. Finally washed three with PBS buffer to remove excess SERS tags and subsequently for Raman detection. The Raman spectra were collected under a He-Ne laser at 633 nm excitation (LabRAM HR Evolution) through a 100× objective with a power of 1.9 mW. The integration time for Raman spectra was set to 20 s, and the integration times were once. The spectral resolution was 1 cm −1 . Calibration of the instrument was done by recording the Raman spectrum of the standard silicon wafer.
Statistical Analysis: The correlation between Raman intensity and clinicopathology feature was performed by SPSS software. The associations between intensity and the clinicopathological characteristics of patients with cancer were assessed with t-test analysis. *p < 0.05; **p < 0.01; ***p < 0.001. The error bars represent the standard deviations from four measurements.
Western Blotting Analysis: The exosomes were lysed with RIPA lysis buffer. And then separated by 12% gel electrophoresis. Further blocked with 5% skimmed milk powder for 2 h and analyzed by western blot. The primary antibodies include anti-TSG101, anti-CD63, and anti-PD-L1 (according to 1:2000 dilution), and the immune response area was visualized.
UV-vis Spectroscopy Measurement: The UV-vis spectroscopy of Au NPs and Au NPs@DTNB were characterized with UV-vis spectrophotometer (UV-2500).
Zeta Potential Measurement: The Zeta potential of Au NPs, Au NPs@DTNB and Au NPs@DTNB@anti-PD-L1 was done by dynamic light scattering (DLS) analysis technique (Malvern Zatasizer Nano ZS).
TEM Measurement: The morphology and size of the Au NPs were characterized with transmission electron microscopic (JEM-F200).
Institutional Review Board Statement: The study was conducted in accordance with the Decla-ration of Helsinki, and approved by the Ethics Committee of Zhongnan Hospital of Wuhan University (protocol code 2022003 and date of approval 18 January 2022).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.