RNA‐sequencing for transcriptional profiling of whole blood in early stage and metastatic pancreatic cancer patients

We investigated the transcriptional profile of whole blood in early and metastatic stages of pancreatic cancer (PaC) patients to identify potential diagnostic factors for early diagnosis. Blood samples from 18 participants (6 healthy individuals, 6 patients in early stage (I/II) PaC, and 6 patients in metastatic PaC) were analyzed by RNA‐sequencing. The expression levels of identified genes were subsequently compared with their expression in pancreatic tumor tissues based on TCGA data reported in UALCAN and GEPIA2 databases. Overall, 331 and 724 genes were identified as differentially expressed genes in early and metastatic stages, respectively. Of these, 146 genes were shared by early and metastatic stages. Upregulation of PTCD3 and UBA52 genes and downregulation of A2M and ARID1B genes in PaC patients were observed from early stage to metastasis. TCGA database showed increasing trend in expression levels of these genes from stage I to IV in pancreatic tumor tissue. Finally, we found that low expression of PTCD3, A2M, and ARID1B genes and high expression of UBA52 gene were positively correlated with PaC patients survival. We identified a four‐gene set (PTCD3, UBA52, A2M, and ARID1B) expressed in peripheral blood of early stage and metastatic PaC patients that may be useful for PaC early diagnosis.

pancreas malignant neoplasms involve the exocrine portion of the pancreas (duct system and acinar cells) with adenocarcinoma features (Hruban & Adsay, 2009). Surgery, chemotherapy, and radiation therapy are treatment options for PaC patients with surgery the only potentially curative treatment (Vareedayah et al., 2018). Based on cancer stage, PaC patients are usually classified into four distinct groups (stages I-IV), which determine their treatment options. Patients in stage I have resectable tumors while stage II includes patients with "borderline resectable" tumors. In stage III, patients have locally advanced unresectable disease, with no metastasis.
The final stage (stage IV) includes patients with metastatic disease, who are not eligible for surgical resection (Garcea et al., 2012;Karmazanovsky et al., 2005). Unfortunately, the absence of specific symptoms in early stage of disease and the ability of pancreatic cancer to early metastasize limit the effectiveness of surgical resection, with only 15%-20% of patients eligible for curative pancreatectomy at the time of the diagnosis (Vareedayah et al., 2018).
Early diagnosis of PaC within the potentially resectable tumor window would represent a significant increase of patients' survival and a decrease in the mortality rate. Challenges to the early PaC diagnosis include the high cost of imaging investigations-namely endoscopic ultrasound and cross-sectional imaging-and the lack of specific biomarkers. Currently, CA19-9 (carbohydrate antigen  is the only blood-based biomarker for pancreatic cancer diagnosis with 75% sensitivity and 90% specificity (Kim et al., 2004) which, due to the high rate of false-positive results (high values of CA19-9 in the absence of malignancy) (Ballehaninna & Chamberlain, 2012) it is not widely considered as an accurate test for PaC early diagnosis.
Previously, in a prospective study on 461 PaC cases in the Iranian population, we analyzed the clinical, pathological as well as therapeutic and survival features of these patients to provide reliable information of Iranian PaC patients. We observed that among 461 studied cases of PaC only 43 (9.5%) individuals were classified as stage I (with resectable tumor) at the time of diagnosis, and the 5-year survival of Iranian PaC patients was lower than the developed countries. We concluded that socioeconomic aspects alongside the stage and pathologic features of PaC may affect the probability of receiving timely treatment and, consequently, patient survival (Sheikh et al., 2020). In another study, published in 2021, we focused on the potential association between diabetes mellitus (DM) and PaC, to identify common molecular factors between these two conditions with the ultimate goal of PaC early diagnosis. In that study, we proposed the potential role of SPI1 and YY1 genes to investigate the relationship between DM and PaC and provide opportunities for PaC screening using blood tests (Kalantari et al., 2021

| Pancancer view and survival map of shared DEGs
Using GEPIA2 data set analysis, we compared the transcription levels    Nanashima et al., 1999), autoantibodies (Capello et al., 2013;Nagayoshi et al., 2014), and exosomes (Capello et al., 2017;Slater et al., 2013). Some studies have also used the microarray technique to investigate the gene expression profile of peripheral blood mononuclear cells (PBMCs) in PaC patients (Baine et al., 2011;Caba et al., 2014;Huang et al., 2010). In contact with tumor tissue PBMCs can alter their gene expression profile and induce a complex transcriptional response (Liotta et al., 2003;Twine et al., 2003;Whitney et al., 2003). Therefore, studies on blood gene expression profiling may contribute to identify reliable cancer biomarkers.
In the current project, a transcriptome study of differentially  Low expression levels of A2M were also observed in blood samples of PaC cases compared to normal samples with decreased pattern from early stage to metastatic PaC (logFc: early stage = −3.13408, metastasis = −4.01882). Alpha-2-macroglobulin (A2M) plays a key role as a universal proteinase inhibitor (Granger et al., 2005) by covalently binding to different proteinase, such as carboxyl, thiol, serine, and metalloprotease (Zhu et al., 2021). This glycoprotein can also bind to various growth factors, cytokines and other nonproteolytic molecules (Burgess et al., 2008). Due to its unique molecular structure, A2M can bind to proteinases and plays its anti inflammatory function, especially for proteinase released from granulocytes and other inflammation-related cells (Zhu et al., 2021).
Overexpression matrix metalloproteinase 9 (MMP9) in blood has been correlated with cell-growth, invasion, and tumor metastasis in breast, colorectal, and pancreatic cancers (Hurst et al., 2007;Kapischke et al., 2008;Somiari et al., 2006). In contrast, inhibition of the MMP9 activity has been associated to primary tumor mass shrinkage and metastasis probability (Kapischke et al., 2008) tissue (e.g., colorectal cancer, as reported by Zhou et al. (2019). We can therefore speculate that UBA52 plays a key role in tumor inhibition, although its upregulationin in our study was in contrast with this report. Hence, we suppose that UBA52, when expressed in blood cells, may be involved in other pathways resulting in its upregulation during tumor progression.
Lastly, we compared the expression levels of PTCD3, UBA52, A2M, and ARID1B in blood and pancreatic tumor tissue according to individual cancer stages, pancreatic tumor grade, and nodal metastasis, reported in TCGA database. Similarly, we observed low expression of PTCD3, A2M, and ARID1B genes in blood and pancreas tumor tissue compared to normal individuals. But UBA52 showed high F I G U R E 6 The stage plot represents the expression level of four shared DEGs in pancreatic cancer based on nodal metastasis status compared with normal samples. Normal (n = 4), N0: No metastatic regional lymph nodes (n = 49), and N1: metastatic regional lymph nodes (n = 124) (by UALCAN data set analysis). DEGs, differentially expressed genes; PAAD, pancreatic adenocarcinoma.

C Cell ell B Biology iology I International nternational
| 245 expression in blood samples although its expression was not significant in tumor tissue compared to normal control (see Figure 3 and Online Supplementary File S3). Similar result was observed for grades 1-4 of panctratic tomours and in metastatic and no-metastatic regional lymph nodes (see Figures 5 and 6). The stage plot illustrated sinusoidal changes in expression levels of PTCD3, A2M, and ARID1B genes from stage I to IV although their increasing trend was observed during cancer progression. While UBA52 expression showed a slight increase from stage I to IV (see Figure 4). The sinusoidal changes in expression levels of PTCD3, A2M, and ARID1B were also observed in Grade 1-4 of panctratic tomours. The PTCD3 and ARID1B genes showed a decreasing trend, but A2M gene showed an increasing trend and UBA52 showed no significant expression during tumor progression (see Figure 5).
Additionally, we observed that UBA52 expression levels were not significant in metastatic and no-metastatic regional lymph nodes tumors compared with normal control. While expression levels of PTCD3, A2M, and ARID1B genes in metastatic and no-metastatic regional lymph nodes were lower than normal samples, although these expression levels were almost equal in metastatic and no-metastatic regional lymph node tumors (see Figure 6). Using Kaplan-Meier plot we finally found that low expression of PTCD3, A2M, and ARID1B genes in PaC patients has positive effect on PaC patients overall survival while high expression level of UBA52 gene was in positively correlated with survival of PaC patients (see Figure 7).

| Conclusion and future perspective
We hypothesized that the identification of genes differentially expressed from early stage to metastatic PaC patients in peripheral blood could be helpful to design noninvasive and cost-effective diagnostic tests for early diagnosis of pancreatic tumor in respectable stages. Using RNA-sequencing analysis, we identified four DEGs in the blood transcriptome profile of pancreatic cancer patients that were shared between early and metastatic stages and we further investigated their expression trend during cancer progression.
Overall, 146 DEGs were identified although we subsequently focused on those statistically significant DEGs according to jointly agreed F I G U R E 7 The Kaplan-Meier plot represents the impact of four shared DEGs on overall survival rate of patients with pancreatic cancer. In all graphs, blue and red lines demonstrate low and high expression levels of transcript, respectively. High expression level (n = 45), low expression level (n = 132) (by UALCAN data set analysis). DEGs, differentially expressed genes; PAAD, pancreatic adenocarcinoma.