Expression of PD-L1 and YWHAZ in Patients with Diffuse Large B Cell Lymphoma: A Possible Association with the Prognosis of Lymphoma

Diffuse large B cell lymphoma (DLBCL) is the most common pathological subtype of non-Hodgkin lymphoma (NHL) and is the most common type of adult lymphoma. Due to the poor prognosis of relapsed/refractory DLBCL, new drug targets and therapeutic methods are urgently needed. We investigated the expression of programmed death ligand 1 (PD-L1) and 3-monooxygenase/tryptophan 5-monooxygenase activating protein zeta (14-3-3ζ or YWHAZ) in patients with DLBCL. The purpose was to verify the expression levels of YWHAZ and PD-L1 and their relationships with the prognosis of DLBCL and to lay a foundation for further study on the role of YWHAZ and PD-L1 in DLBCL. Immunohistochemistry was used in 140 patients with DLBCL to test protein expression levels of YWHAZ and PD-L1. All patients were followed up in the hospital or by telephone or via WeChat. The positive expression rate of YWHAZ was 62.14% (87/140). The expression was negatively correlated with the positive expression of BAD (r = −0.177, P = 0.036) and positively correlated with the positive expression of BCL-2 (r = 0.180, P = 0.033). When the cut-off value for PD-L1 was established at 5%, 10%, 15%, and 20%, the corresponding positive expression rates of PD-L1 were 79.66% (94/118), 51.69% (61/118), 40.68% (48/118), and 36.44% (43/118). YWHAZ significantly affected the OS of DLBCL (P ≤ 0.001). The prognosis of the patients was related to the positive expression of PD-L1 when the cut-off value of PD-L1 was 5% (P = 0.033). However, positive expression of PD-L1 was not associated with the prognosis when the cut-off values of PD-L1 were 10% (P = 0.404), 15% (P = 0.208), and 20% (P = 0.408). The positive expression of YWHAZ (hazard ratio 6.215; 95% confidence interval 3.214-12.017; P < 0.05) was an independent adverse prognostic factor for OS. YWHAZ may be an important oncogene in the occurrence and development of DLBCL and may be used as a therapeutic target. PD-L1 may be an oncogene or tumor suppressor gene in the occurrence and development of DLBCL. Different cut-off values of PD-L1 may affect the prognosis of DLBCL.


Introduction
According to the Global Cancer Statistics 2018 [1], the annual growth rate of lymphoma incidence in China is 3%-5%, with approximately 100,000 new cases per year. The age of morbidity tends to be younger. The overall 5year survival rate of lymphoma patients in China is only 37.2%-38.4% [2,3], which is much lower than that of the United States (68.1%) and Japan (57.3%). Only 22.1% of patients with diffuse large B cell lymphoma (DLBCL) received more than eight standard treatment courses in China in 2019 [4]. The general poor efficacy and prognosis of lymphoma have become an unbearable burden to patients, individuals, families, and society. Clinical studies have shown that the survival time of lymphoma patients can be significantly prolonged and even cured after standard treatment and comprehensive management [5]. The standardized diagnosis and treatment of lymphoma require cooperation from different disciplines. New therapeutic options, such as monoclonal antibodies, small molecule targeted drugs, and immunotherapy, have significantly improved the short-term efficacy and long-term survival of patients with non-Hodgkin lymphoma (NHL). However, the prognosis of patients with relapsed/refractory (R/R) NHL is still poor, which severely impacts the quality of life. DLBCL is the most common pathological subtype of NHL and the most common type of adult lymphoma. Due to the poor prognosis of R/R DLBCL, new drug targets and therapeutic methods are urgently needed.
In recent years, immune checkpoint inhibitors (ICIs) have been the focus of tumor immunotherapy, and tumor therapy has entered a new era of immunotherapy. Programmed death ligand 1 (PD-L1, also named B7-H1 or CD274), which is overexpressed on tumor cells and tumorassociated macrophages (TAMs), binds to PD-1. PD-1 is expressed in T cells and tumor-infiltrating lymphocytes (TILs), which can inhibit T cell activation via PD-1 during tumor development. T cells appear to function under exhaustion, thus achieving tumor immune escape. Kiyasu et al. [6] show that PD-L1 expression in DLBCL is associated with a poor prognosis. However, McCord et al. [7] obtained the opposite results. Kim et al. [8] suggest that the PD-L1 expression in the primary central nervous system DLBCL (PCNS-DLBCL) is associated with an adverse prognosis. The cut-off values of PD-L1 are controversial. Therefore, the expression level of PD-L1 should be further studied to predict the prognosis of DLBCL.
Our previous study found that the tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activating protein zeta (also named 14-3-3ζ or YWHAZ) was differentially expressed according to the isobaric tags for relative and absolute quantification (iTRAQ). Parallel reaction monitoring (PRM) confirmed that YWHAZ was upregulated in DLBCL tissues. The high-level expression of YWHAZ was associated with an unfavorable prognosis of DLBCL [9]. This study was aimed at verifying the expression levels of YWHAZ and PD-L1 and their relationships with the prognosis of DLBCL. The study will lay the foundation for further research on the role of YWHAZ and PD-L1 in DLBCL.

Material and Methods
2.1. Clinical Data. Patient information sex, age, race, and contact information were collected. The following laboratory tests were collected: white blood cell (WBC) count, platelet count, red blood cell (RBC) count, neutrophil granulocyte, monocyte count, hemoglobin, lactate dehydrogenase (LDH), and serum creatinine (Scr). Imaging (CT/MRI/ultrasound) was recorded. Clinical notes included B symptoms (fever, night sweats, and weight loss), first symptoms and locations of the disease, the Ann Arbor Staging classification, Eastern Cooperative Oncology Group performance status (ECOG PS), infectious diseases (HIV, hepatitis B or C, tuberculosis, and syphilis), other systemic diseases (hypertension, diabetes, and heart disease), biopsy (histological/bone marrow), and survival results. The age-adjusted International Prognostic Index (aaIPI) was categorized according to age (≤60 years old and >60 years old). Pathological parameters such as protein expression levels, Ki67 ≥ 70 %, and positivity for Epstein-Barr virus (EBV) were also collected.

Pathological Data
2.2.1. Tissue Samples. Informed consent before biopsies was obtained from patients or family members. All samples were obtained from formalin-fixed paraffin-embedded (FFPE) tissues. No treatments were given before this research. The characteristics of all patient samples are recorded in Table 1.

Tissue
Microarray and Immunohistochemistry. Tissue chip technology was used for this experiment. One hundred and thirty-seven cases of FFPE tissues were from chips, and three were from three separate slices. The method of testing YWHAZ and PD-L1 was similar to the previous study [10]. However, PD-L1 was detected using an automated immunohistochemical assay. The main antibodies chosen are listed in Table 1. The cut-off values for the overexpression of C-MYC and BCL-2 were ≥40% and ≥70%, respectively. The positive standards for CD10, BCL-6, and MUM-1 were ≥30% [11]. YWHAZ was examined following the manufacturer's protocols. No standard cut-off values of PD-L1 are available. The cut-off values for PD-L1 in this study were tested at ≥5%, ≥10%, ≥15%, and ≥20%, respectively. The diagnostic criteria for double-expressor lymphoma (DEL) or tripleexpressor lymphoma (TEL) are the same as in the literature [12]. DLBCL was classified according to the Han's algorithm, consistent with the previous study [12].

In Situ
Hybridization. An in situ hybridization kit for EBER (EBV-encoded RNA) was purchased from ZSBIO (ISH-5022, ZSBIO, China). The procedure followed the manufacturer's instructions. Nasopharyngeal carcinoma tissues with positive EBER were used as a positive control.
2.4. Follow-Up Visits. The first diagnosis was the start of the follow-up visits. Follow-ups ended on July 16, 2021. Followup visits were made at the hospital, by telephone, or by WeChat. The reasons for ending follow-up were recorded. Progression-free survival (PFS) is the period between the start of treatment and the observed disease progression or
(2) COX multivariate survival analysis Combined therapy was more effective than surgery alone or chemotherapy alone. The risk of death was lower than that of surgery alone or chemotherapy alone (HR = 0:436;        Table 4).

Discussion
DLBCL can occur at any age. Most of the patients are middle-aged and older adults. Extranodal diseases account for about 30%-40%, consistent with this study [13,14]. The prognosis of patients with Ann Arbor staging in the early stage was not different from that in the advanced stage, consistent with a previous study [15]. Because some patients with an advanced stage had been treated in other hospitals before coming to this hospital for treatment, the survival time was longer. The Ann Arbor stage was not related to the prognosis of patients with DLBCL.
In recent years, the morbidity and mortality of primary extranodal DLBCL have risen. Annual OS is also gradually increasing, closely related to the continuous improvement of the lymphoma evaluation system, increased patient awareness of the disease, the development of genetics and molecular biology, and the use of rituximab [16]. The onset of extranodal DLBCL first appears in the central nervous system, followed by the gastrointestinal tract, possibly due to information bias and recall bias. Finally, the prognosis for patients with extranodal DLBCL is better than that for intranodal patients, which is inconsistent with a previous report [17]. The different locations of the primary extranodal DLBCL reflect different clinical features and prognostic effects. Therefore, a new risk stratification feature that involves the origin of the disease is needed to guide the treatment of extranodal DLBCL [18].
Five cases were complicated with the hemophagocytic syndrome and died at the end of follow-up. Among these 13 Journal of Immunology Research patients, three were caused by chemotherapy. These patients were critically ill with a poor prognosis and high mortality rate. Clinicians and pathologists should pay enough attention to this type of patient. Many NHL patients have a history of autoimmune diseases, such as Sjögren syndrome, rheumatoid arthritis, or Helicobacter pylori gastritis [19]. Klein et al. [20] hypothesized that continuous disease activity and immune stimulation were the most vital factors for developing DLBCL in patients with rheumatoid arthritis. Consistent with the report [21], we found that DLBCL was a disease diagnosed after Sjögren syndrome.
Compared to sporadic cases, patients with an impaired immune system are more likely to develop EBV-positive DLBCL [22]. This study had two cases of EBV-positive DLBCL: one had silicosis, and the other was diagnosed without complications. The findings are consistent with a previous report [22]. A meta-analysis showed that OS and PFS of EBV-positive DLBCL were significantly worse [23]. These patients should receive a complete course of treatment, especially those with respiratory diseases. There was an autoimmune disorder in patients with silicosis, but no correlation was found [24]. The two cases of EBV-positive DLBCL in this study were sporadic. Because the IPI score is related to performance status, it is difficult to determine whether poor performance status is caused by rheumatic disease or DLBCL in patients with DLBCL with rheumatic diseases.
The prognosis of patients with DLBCL is closely related to hepatitis B virus infection and replication [25,26]. In this study, due to the small number of samples and cases of hepatitis B combined with other chronic diseases, it is impossible to accurately assess its effect on the prognosis of DLBCL. Therefore, more research is needed. At the same time, paying more attention to the patients with DLBCL complicated by hepatitis B infection is necessary. Clinicians should pay attention to patients with systemic diseases, especially hepatitis B and Sjögren syndrome. It is essential to follow these patients closely. The level of virus replication before chemotherapy should be assessed to avoid reactivating the virus, causing virus DNA replication, increasing the difficulty of treatment, and accelerating patient death.
A daily examination of bone marrow involvement should be performed to adjust the treatment scheme. The Ki67 index of the non-GCB subtype is higher than that of the GCB subtype, which contradicted the results of this study [27]. Some studies showed a reverse relationship between the Ki67 index and clinical prognosis, while others did not [28]. The importance of early immunohistochemical and FISH detection in predicting the prognosis of DLBCL should be noted. Under standard treatment, patients with the non-GCB subtype tend to have a poorer prognosis than those with the GCB type [29], which is confirmed in this study. Ki67 is associated with the prognosis of patients with primary intestinal DLBCL [30], unlike the results in this study. The findings may be related to the range of Ki67 cut-off values and the selection of research objects.  The positive expression of YWHAZ was correlated with BAD and BCL-2, suggesting that YWHAZ may be related to apoptosis and YWHAZ may restore sensitivity to CHOPinduced apoptosis [31]. The protein expression level of YWHAZ was not correlated with those of p-AKT and AKT. This could be due to the low positive rates of AKT and p-AKT proteins in immunohistochemistry, which were affected by various factors, such as the preservation time of the selected samples, antibodies without high sensitivity, and the technique of the experimenter. These factors could lead to false negative results. The correlation can be further analyzed by expanding the sample size or performing experiments using Western blot or qRT-PCR methods. Positive expressions of the YWHAZ, AKT, and p-AKT proteins might promote the development and proliferation of DLBCL, related to the poor prognosis of DLBCL. BAD and BAX were associated with a good prognosis of DLBCL in this study. BAX might be related to DLBCL proliferation and inhibiting tumor cell proliferation. The prognosis can be preliminarily evaluated according to the expression level of BAD. Consistent with the previous report, the positive expression of YWHAZ is related to the prognosis of DLBCL [9] with the short OS. It is suggested that YWHAZ may be an oncogene in DLBCL, which is worthy of further study.
Although the cut-off value of PD-L1 was different, C-MYC was consistently positively correlated with the positive expression of PD-L1 in our research. The positive expression of C-MYC protein in DLBCL is related to poor prognosis [32], suggesting that PD-L1 may be an oncogene in DLBCL. When the cut-off value of PD-L1 was 10%, 15%, and 20%, BCL-6 was always negatively correlated with prognosis. However, BCL-6 was found to be an oncogene in DLBCL [33]. PD-L1 may be a tumor suppressor gene in DLBCL. It can be seen that PD-L1 with different cut-off values has different effects on the prognosis of DLBCL. When the cut-off value of PD-L1 was 5%, the positive expression of PD-L1 was related to the poor prognosis of DLBCL. The result is consistent with some reports [6,8] but inconsistent with this report [7]. The expression of PD-L1 is significantly correlated with the prognosis of melanoma patients when 5% is used as the cut-off value, which is consistent with this study [34]. PD-L1 may be an oncogene or tumor suppressor gene in DLBCL, which provides a new target for the treatment of R/R DLBCL. Macrophages are the main source of PD-L1 expression in the tumor microenvironment (TME) of DLBCL [35]. Based on this study, our team intends to explore the relationship between the expression of TAM and PD-L1 expression in TME and their effects on the prognosis of DLBCL.
In this study, the OS of patients with DLBCL treated with comprehensive treatment was higher than with simple surgical resection or chemotherapy, and the prognosis was better. Therefore, clinicians should first recommend comprehensive treatment when formulating the treatment plan to improve OS and quality of life. We also found that the ECOG score, hyphemia, complicated systemic diseases (such as hepatitis B and Sjögren syndrome), LDH, aaIPI, and BCL-2 were related to the prognosis of DLBCL. However, sex was not associated with the prognosis of DLBCL. These results were consistent with some studies [36][37][38][39][40] but inconsistent with others [37,38,41].

Conclusions
YWHAZ may be an important oncogene in the occurrence and development of DLBCL and may be used as a therapeutic target. PD-L1 may be an oncogene or tumor suppressor gene in the occurrence and development of DLBCL. Different cut-off values of PD-L1 may have different effects on the prognosis of DLBCL. These findings may suggest novel targets for the diagnosis and therapy of DLBCL.

Data Availability
All data generated or analyzed during this study are included in this published article.