Cisplatin Increased the Expression of PD-1 and PD-L1 by YAP1 in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) was one of the most malignant cancers in the world. Cisplatin (DDP) was one of the main chemotherapy drugs for HCC, but the mechanism of DDP treatment for HCC remains unclear. In this presentation, we found that DDP inhibited the growth of HCC cells and promoted the expression of PD-1 and its ligand PD-L1 in cancer cells. Meanwhile, ow cytometry analysis revealed that DDP enhanced PD-1 - CD8 + T cells expression and decreased PD-1 + CD8 + T cells expression. ELISA analysis suggested that DDP decreased TGF-β expression in vivo. Therefore, the study indicated that DDP enhanced PD-1 and PD-L1 expression and inhibited the growth of HCC.


Lentivirus production and infection
The plasmids that encode the lentiviruses expressing shRNA molecules were obtained from the RNAi Consortium shRNA Library. The shRNA target 21-mer sequences were: shControl, CCTAAGGTTAAGTCGCCCTCG; human shYAP1 #3CCGGGACCAATAGCTCAGATCCTTTCTCGAGAAAGGATCTGAGCTATTGGTCTTTTTG. HepG2215 cells were placed in a 24-well tissue culture dish (2 ×10 5 cells per well) and infected with 80 μL shRNA lentivirus supernatant and 4 μg/ mL polyene. The infected cells were placed in RPMI medium containing 10% fetal bovine serum and 3 μg/ mL purinomycin, and detected 72 h after infection. Western blot was used to detect shRNA knockout e ciency.

Human Protein Atlas database
Human Protein Atlas HPA (https://www.proteinatlas.org/) provided each protein in 64 cell lines, 48 kinds of normal human tissues, and 20 kinds of cancer tissues. We collected 365 patients from HPA database and plotted survival time curves.

Survival analysis of PD-1 and PD-L1
Kaplan-Meier Plotter (KM) (https://kmplot.com/analysis/) is a clinical database of survival analysis, Log rank p-value and hazard ratio (HR), and 95% con dence intervals to assess the relationship between PD-1 and PD-L1 and overall survival.

Animal experiment
All animals were fed in the SPF with constant temperature (22-24 °C) and a dark-light cycle of 12 h/12 h, and housed in plastic cages. The protocol was approved by the Ethics Committee for Animal Experiment of Hebei University of Chinese Medicine (Permit number: YXLL2018002).

Immuno uorescence assay
All uorescence images were observed under a confocal instrument. The primary antibodies were rabbit anti-CD4(GB11064, Seville Biological) and rabbit anti-CD8(GB13429, Seville Biological). The secondary antibody was FITC goat anti-rabbit IgG-HRP (GB22303, Seville Biological). Nucleus was blue by labeling with DAPI. Positive cells were green or red according to the uorescent labels used. The image was captured using a uorescence microscope (Olympus BX51, Japan).

Flow cytometry analysis
Spleens of mice were removed and placed in 5 mL PBS+1%FBS+2 mM EDTA solution. The spleen was ground and ltered to obtain spleen cells; Lyse the erythrocytes, count the cells, adjust to 1x10 6 cells/ml; Adding the antibody according to the antibody instruction, stained and tested on the machine. Finally, the results were analyzed.
Hematoxylin and eosin (H&E) staining All organs were removed quickly, and the liver tissues were xed with 4% formaldehyde for 24-48 h and then embedded with para n. The section thickness was 5 µm. The sliced sections were stained with haematoxylin and eosin (H&E), and change of histopathological was obtained using a microscope (Leica DM2500, Germany).

Immunohistochemistry (IHC)
Immunohistochemical staining was performed according to the instructions of Nakasi Jinqiao Kit. Finally, sections were determined by microscopic observation of the brown peroxidase in liver tissue at a 200× and 400× magni cation. The results of immunohistochemistry were examined by 2 senior histopathologists. The cytomembrane/cytoplasm stained with light yellow or tan were regarded as positive cells. The optical density was measured by an Image-Pro Plus v6.0 software (Media Cybemetic, USA).

Ultrasonic testing
At the beginning and end of the experiment, ultrasound was used to measure the number of liver tumors in the mice using an imaging system (Vevo 2100, Visual Sonics Inc., Toronto, Canada) with an MS250 ultrasound transducer. Soon, the mice were anesthetized and their abdomens shaved. M-mode recording of short and long axis views of planed abdominal wall was performed.
DEN/TCPOBOP-induced HCC model in C57BL/6 mice Yap1 LKO mice have been successfully constructed by Guangzhou Cyagen Biosciences (Guangzhou, China). The mice were identi ed by our group. The modeling method was introduced as previously described [11]. Yap1 ox/ ox were used as the control. 3-week-old male C57BL/6 mice were injected intraperitoneally with 25 mg/kg bodyweight N-nitrosodiethylamine (DEN). Then at the age of 4 weeks, the mice were injected with a dose of 3 mg/kg body weight TCPOBOP. The mice were divided into four groups after the model was successfully built. For the DDP group, the DDP dissolved in saline. The control mice were injected with saline. The Ethics Committee approved all experimental animal protocols for the Animal Experiment of Hebei University of Chinese Medicine (Permit number: YXLL2018002).

Statistical analysis
All statistical tests were performed by SPSS23.0 statistics software (SPSS, Chicago, IL). All the in vitro experiments were repeated for at least three times. Data were presented as Mean±SD. When more than two groups were enrolled, the means were compared between each two groups with one-way ANOVA.
Differences were considered statistically signi cant when P<0.05.

Results
DDP inhibited the liver cancer growth in DEN/TCPOBOP-induced HCC model in C57BL/6 mice.
DDP was one of the most commonly used chemotherapy drugs for the treatment of HCC. To further evaluate the effect of DDP in vivo, the mice model of liver carcinoma in situ was established successfully. We found that the mice's left and right liver lobes had tumors in the saline (NS) group. But there were no apparent tumors on the liver surface of the DDP group (Fig.1A). Further statistics found that DDP decreased the number of tumors (P<0.05). DDP also reduced the volume of tumor, but not signi cantly (P>0.05) (Fig.1A). Moreover, the development of tumors could be monitored by small animal ultrasound imaging. The results showed that on the 1st day, in the transverse section, tumor area in the NS group was 19.5 mm 2 , the long diameter was 2.431 mm, and the short diameter was 2.38 mm; in the vertical section, tumor area in the NS group was 5.686 mm 2 , the long diameter was 2.094 mm, and the short diameter was 1.525 mm. The transverse section area of the tumor in the DDP group was 22.093 mm 2 , the long diameter was 3.97 mm, and the short diameter was 2.302 mm; the vertical section area of the tumor in the DDP group was 10.217 mm 2 , the long diameter was 3.05 mm, and the short diameter was 2.28 mm. On the 25th day, in the transverse section, the area of the tumor in the NS group was 13.132 mm 2 , the long diameter was 4.841 mm, and the short diameter was 3.837 mm; in the vertical section, the area of the tumor in the NS group was 19.916 mm 2 , the long diameter was 4.887 mm, and the short diameter was 3.805 mm. The transverse section area of the tumor in the DDP group was 8.502 mm 2 , the long diameter was 1.093 mm, and the short diameter was 0.567 mm; the vertical section area of the tumor in the DDP group was 7.67 mm 2 , the long diameter was 1.29 mm, and the short diameter was 0.459 mm (Fig. 1B). These results showed that DDP signi cantly inhibited the growth of liver tumor in C57BL/6 mice.
However, DDP also showed adverse reactions in tumor inhibition. Compared with NS group, DDP decreased the weight of the thymus (P=0.04) and kidney P=0.04). It had little effect on the weight of liver and spleen. These results suggested that DDP decreased the weight of thymus and kidney, did not affect organ index (Fig. 1C).
As the component of the immune system, CD8 + T cells played an important role in anti-cancer [12,13]. To further study the expression of CD8 + T cells after the DDP treated hepatic carcinoma in situ C57BL/6 mice, immuno uorescence assay and ow cytometry were used for detection. The immuno uorescence assay results suggested that in the NS group, the number of CD8 + T cells decreased, while the DDP group had highly CD8 + T cells expression ( Fig. 2A). Consistent with this, the previous study showed that compared with the control group, the DDP group increased CD8 + T-cell in ltration into tumor tissues [14]. Moreover, DDP or DDP combined with anti-PD-1 failed to inhibited the growth of tumor in the depletion of CD8 + T cells in vivo [14]. These ndings indicated that CD8 + T cells played a vital role in the anti-tumor effect of DDP. Previous studies showed that high levels of PD-1 + CD8 + T cells were associated with a poor prognosis for a range of cancers, including liver, pancreatic, and head and neck cancers [15][16][17][18]. In our study, further study showed that DDP also signi cantly reduced exhausted CD8 + T lymphocytes (PD-1 + CD8 + T cell) in the blood (Fig. 2C) and spleen (P<0.05) (Fig. 2B). It also increased non -draining CD8 + T lymphocytes (PD-1 -CD8 + T cells) in blood (P<0.05) (Fig. 2C) and spleen (P<0.05) (Fig. 2B). These results showed that DDP inhibited the growth of HCC by increasing the number of PD-1 -CD8 + T cells and decreasing the number of PD-1 + CD8 + T cells. In general, Th1 cells secrete Interleukin-2 (IL-2) and Interferon-γ (IFN-γ) to inhibit tumor growth, while Th2 cells release immunosuppressive factors Interleukin-4 (IL-4) and Interleukin-10 (IL-10) to produce immunosuppression. Chemokines recruited bone marrow-derived suppressor cells and promoted cancers. Interleukin-1 beta (IL-1β), as proin ammatory cytokines, was elevated in cancers [19]. In our study, we found that DDP had almost no effect on IL-2, IFNγ, IL-4 and IL-10 (Fig. 2D). Transforming growth factor-β (TGF-β) played crucial role in liver cancer [20]. A study showed that overactivation of TGF-β signaling led to progression of tumor [21]. In our study, ELISA results showed that DDP decreased the expression of TGF-β in liver tumor tissues. TGF-β was signi cantly lower in the DDP group than that in the NS group (Fig. 2D). The results suggested that DDP decreased the expression of TGF-β. In summary, these results showed that DDP reduced the expression of TGF-β, but did not affect the expression of Th1-type cytokines and Th2-type cytokines, IL-1β, and chemokines in the liver cancer.
DDP signi cantly increased the expression of PD-1 in HepG2 and HepG2215 cells.
PD-1 receptor on the surface of immune cells was an immune checkpoint molecule that mediated immune escape from tumor cells. The study revealed that PD-1 expressed in a wide range of cancer cells [7]. To further verify the effect of DDP on PD-1 on hepatocellular carcinoma cells, immuno uorescence assay and Western blot were performed. Immuno uorescence assay results showed that in the NS group, PD-1 had less expression. DDP group signi cantly improved the expression of PD-1 compared with the NS group (Fig. 3A). The results of Western blot and statistical analysis also showed that DDP increased the PD-1 expression in HepG2 and HepG2215 cells (P<0.05) (Fig. 3B-C). PD-L1 was a major ligand [22]. PD-L1 and PD-1 had a similar effect in tumors. A previous study showed that PD-L1 was expressed on these cancer cell lines [7]. Indeed, in our previous study, we found that DDP promoted the PD-L1 expression on tumor cells by promoting YAP1 expression [23]. In summary, these results showed that DDP increased the PD-1 and PD-L1 expression in HepG2 and HepG2215 cells. YAP1 is a critical molecule in the Hippo pathway. The Hippo pathway was primarily known for its role in cell proliferation and differentiation, but some studies suggested its importance in tumor immunity.
Studies showed that YAP1 was positively correlated with tumor-in ltrating cells, including CD8 + T cells and CD4 + T cells, in various cancers, including liver cancers [10]. This result indicated that the elevation of YAP1 was related to the in ltration of some immune cells and had an anti-tumor function. Our previous research showed that compared with the NS group, DDP improved the expression of YAP1 in HepG2215 cells in a dose-dependent manner and promoted the expression of YAP1 in BALB/c mice liver H22 cells [23]. We did immunohistochemical staining. We found that DDP promoted the high expression of YAP1. In this study, the result of immunohistochemical staining suggested that compared with the NS group, DDP improved the presentation of YAP1 in C57BL/6 mice (Fig. 5A).
However, it was unclear whether DDP induced PD-1 and PD-L1 through YAP1 in HepG2215 cells. Previous studies also showed that YAP1 directly bond to the PD-L1 primer to promote the transcription of PD-L1 in PC9 cells of lung cancer. [24]. The study also showed that PD-1 and PD-L1 had a similar effect on the cancer. We further explored the impact of YAP1 on PD-1 and PD-L1. The HepG2215 shYAP1 cells were constructed. Meanwhile, in our study, the results of Western blot indicated that compared with HepG2215 shCont cells, knockdown of YAP1 decreased the expression of PD-L1 and PD-1 in HepG2215 cells (Fig.  5B). In summary, DDP promoted the expression of PD-L1 and PD-1 through encouraged YAP1 expression.
To explore the DDP effect on PD-L1 and PD-1, we performed an experiment in C57BL/6 mice. Liver tissuespeci c Yap1 knockout (Yap1 LKO ) mice successfully modeled for HCC. In Yap1 ox/ ox group, DDP reduced the number of tumors (Fig. 6A). In Yap1 LKO group, DDP reduced the number of tumors. PCR was used to verify the expression of ox and cre (Fig. 6B) (Fig. 6C). These results showed that Yap1 affected the expression of PD-L1. Besides, we found that Yap1 knockout reduced PD-L1 and PD-1 expression. However, DDP had no signi cant effect on PD-1 in C57BL/6 mice (Fig. 6C). The reason could be that PD-1 was expressed in a variety of tissues, including cancer tissues.

Discussion
DDP was one of the most widely used drugs in the treatment of cancers. It was used in many cancers, such as liver cancers [25] and lung cancers [26]. However, the mechanism of DDP treatment for HCC in the microenvironment remains unclear. In this study, we showed that YAP1 expression promoted the expression of PD-1 and PD-L1 in liver cancer cells. Interestingly, DDP increased the expression level of PD-1 and PD-L1 by YAP1 expression. DDP also decreased the percentage of PD-1 + CD8 + T cells in blood and spleen, and it decreased TGF-β expression level in vivo. This study provides a mechanism for the treatment of liver cancer with cisplatin.
PD-1 promoted cancer growth by activation of mTOR in the cancer [27]. Anti-PD-1 treatment reduced the growth of bladder and ovarian cancer cells [28]. These results revealed that PD-1, inherent in cancer cells, was a pro-tumor factor. A study showed that silencing PD-1 or PD-L1 promoted cell proliferation and colony formation in vitro, In contrast, overexpression of PD-1 and PD-1 inhibited the proliferation of cancer and colony formation [7]. Consistently, our studies found that DDP increased the expression of PD-1 and PD-L1, and inhibited liver cancer growth in mice. Anti-PD-1 and anti-PD-L1 have been applied in many cancers by activating T cells. However, a considerable number of patients failed to block the response to the PD-1/PD-L1 axis or relapsed after response. Paradoxical progressive disease(PPD) and hyperprogressive disease (HPD) were gradually recognized in people [7,29]. However, if the antibody-activated T cell level is insu cient, cancer grows faster by activating the PD-1/PD-L1 function inherent to the tumor, and then degenerates after T cell overactivation (PPD). In contrast, antibody-mediated therapy enhanced cancer cell growth and suppressed anti-cancer immunity, leading to the development of HPD in the case of immunocompromised or antibody-activated T cells, and appropriately elevated PD-1/PD-L1 expression on cancer cells [30]. A previous study also indicated that the subpopulation of tumor cells expressing both PD-1 and PD-L1 reduced tumor growth [7]. A similar study found that transarterial chemoembolization (TACE) increased the expression of PD-1 and PD-L1 in HCC, optimizing tumor response in selected cases, which used TACE in combination with immunotherapy [31]. At the same time, we also found that DDP increased the expression of PD-1 and PD-L1 on HCC cells and high expression of PD-1 and PD-L1 had a longer overall survival time.
YAP1, as a transcriptional co-activator, was a key downstream effector of the Hippo pathway [32]. YAP1 promoted the recruitment of immune cells [33]. Pancancer analysis found that YAP1 activity was strongly positively correlated with many types of tumor-in ltrating cells in various cancers [10]. Interestingly, this our study also found that DDP promoted YAP1 expression and increased CD8 + T cells in C57BL/6 mice. Our previous study showed that DDP promoted YAP1 expression in HepG2215 cells, and YAP1 expression was restored with DDP in shYAP1 HepG2215 cells [23]. YAP1 also directly bond to the PD-L1 primer promote the transcription of PD-L1 [24]. Our previous study suggested that overexpression YAP1 promoted PD-L1 expression. Consistently, a study showed that DDP induced the over-expression of PD-L1 in H22 hepatoma cells [34]. In this study, we also found that YAP1 knockdown and YAP1 knockout also reduced the expression of PD-1 and PD-L1in liver cancer cells in vivo and in vitro.
CD8 + T cells played a vital role in the anti-cancer effect. Accordingly, other studies showed that CD8 + T cell in ltration into cancer tissues increased during treatment, including in the DDP alone group, compared to the control group, and the therapeutic effect diminished as the number of CD8 + T cells decreased [14]. Consistently, one study showed that DDP exhibited CD8 + T cell-mediated therapeutic effect on tumor-bearing mouse models [35]. Studies suggested that PD-1 expression was related to T cell exhaustion. In general, without immunotherapy, high expression of PD-1 + CD8 + T cell was associated with poor prognosis in various tumors, including liver cancer [36]. In our study, we found that DDP signi cantly reduced exhausted CD8 + T lymphocytes in blood and in spleen. It also increased non-draining CD8 + T lymphocytes in the blood and in spleen. Previous studies showed that TGF-β and IL-2 and IFN-γ expressing CD8 + T cells contributed to an anti-cancer immune reaction [37]. TGF-β played a crucial role in liver cancer. The overactivation of TGF-β signaling may lead to the progression of the cancer. In general, Th1 cells secreted IL-2 and IFN-γ to inhibit cancer growth, while Th2 cells released immunosuppressive factors IL-4 and IL-10 to produce immunosuppression. Furthermore, CD8 + T cells expressed by IL-2 played an essential role in the response of anti-cancer immune. The study showed that the treatment of IL-2 reversed the exhaustion of CD8 + T cells and signi cantly increased the IFN-γ in malignant pleural effusion [38]. IL-2 promoted CD8 + T cells and natural killer cells responding to antigens [39]. Treatment of IL-2 reversed the exhaustion of CD8 + T cells and signi cantly increased the IFN-γ in malignant pleural effusion. IFN-γ had direct anti-proliferation, pro-apoptotic and anti-angiogenesis effects on cancer cells [40]. Studies showed that chemokines recruited bone marrow-derived suppressor cells and promoted cancers [41]. IL-1β was a pro-in ammatory cytokine that promoted the growth of breast cancer cells through the P38/MAPK and PI3K/Akt pathways [19]. In our study, DDP decreased TGF-β expression, but it did not affect the expression of Th1-type cytokines and Th2-type cytokines, IL-1β, and chemokines.
In conclusion, our study found that DDP inhibited the growth of HCC. It increased the expression of PD-1 and PD-L1 by YAP1 expression in the cancer. DDP increased the CD8 + T cell number in tumor microenvironment. Meanwhile, DDP decreased the percentage of PD-1 + CD8 + T cells in blood and spleen. DDP decreased TGF-β expression level in liver cancer. Therefore, this study provided the mechanism for the liver cancer microenvironment with DDP. Figure 1 Cisplatin inhibited the growth of HCC. A. The organ of C57BL/6mice in the NS group and DDP group. Blue and red arrowheads indicate liver and spleen. The white circles are tumors. B. When the mice developed tumors, they were enrolled as individuals. Two successive ultrasonic measurements were carried out for detection. Ultrasonography was performed on the rst day and 25 days after medication intervention. Ultrasonography was performed on the transverse and longitudinal sections of the lesion. The white dotted line represents the tumor. C. Effect of cisplatin on weight and index of liver, spleen, kidney, and thymus. D. * indicates P<0.05, **P<0.01, ***P<0.001 compared with the saline group.

Figure 2
Cisplatin increased PD-1-CD8+ T cells and reduced PD-1+ CD8+ T cell. A. Cells were treated with DDP and 0.1% DMSO (NC), respectively, for 24 h. CD8+cells were stained with green by immuno uorescence, and nuclei were counter-stained with DAPI (blue). The upper and bottom panels were respectively 100 ×.

Figure 3
Cisplatin increased PD-1 expression in HCC. A. Cells were treated with DDP and 0.1% DMSO (NC), respectively, for 24 h. PD-1 was stained with red by immuno uorescence, and nuclei were counter-stained with DAPI (blue). The upper and bottom panels were respectively 400 ×. B. Western blot was detected the expression of PD-1 in HepG2 cells. * indicates P<0.05, **P<0.01, ***P<0.001compared with the saline group. C. Western blot was detected the expression of PD-1 in HepG2215 cells.  of YAP1 expression levels after cisplatin treatment was detected by Image J. In the cisplatin-treated group, the expression of YAP1 was signi cantly higher than that in the saline group. B. Western blot analysis of PD-1 and PD-L1 expression in YAP1 knockdown in HepG2215 cells. β-actin was used as a loading control. The expression of YAP1 in C57BL/6 mice was examined by immunohistochemical staining. And * indicates P<0.05, **P<0.01, ***P<0.001compared with the saline group.