The Importance of Stroma and Stromal Sma Expression in Pancreatic Ductal Adenocarcinoma

Objective: Pancreatic stellate cells (PSC) have been defined to be the key players in pancreatic fibrogenesis and carcinogenesis. They undergo myofibroblast-like differentiation, express α-smooth muscle actin (α-SMA), and play a crucial role in injury and inflammation sites. This study aims to evaluate the relationship between α-SMA expression and histopathological parameters of pancreatic ductal adenocarcinoma (PDAC), and investigate their association with prognosis. Material and Methods: Eighty-one consecutive pancreatectomies diagnosed as usual pancreatic ductal adenocarcinoma were included. The stromal density was scored as loose, moderate, or dense, and α-SMA expression was evaluated immunohistochemically. Results and Conclusion: Mean survival was 19.6 months. Male gender, larger tumor diameter (>3.7 cm), and older age (>64 years) were identified as independent poor prognostic factors. Perineural invasion significantly effected survival. A statistically significant correlation was found between high α-SMA expression and the presence of angioinvasion (p=0.01). Stromal α-SMA expression in PDAC may help determine the risk of angioinvasion.


INTRODUCTION
Pancreatic ductal adenocarcinoma (PDAC) ranks as the fourth leading cause of cancer-related deaths in males and the third in females (1,2).The primary reasons for its poor survival rates include low operability at the time of diagnosis and the resistance to conventional chemotherapy protocols.Hence, there is an urgent need for new targeted therapeutic approaches (1).
The stroma of cancers is a recent hot topic of research.Colorectal tumors are the first to be proven to have significant prognostic outcomes related to their stroma (3).Ueno et al. classified the desmoplastic reaction as mature, intermediate, or immature based on the presence of hyalinized keloid-like collagen and myxoid stroma at the extramural desmoplastic front.The prognostic power of this categorization in stratifying relapse-free survival was greater than any other conventional prognostic factors, including TNM stage, venous invasion, and tumor grade (4,5).Subsequently, Wang et al. applied an undefined scoring system using H&E and Masson's trichrome stains for PDACs (6).
The stroma of PDAC constitutes more than 50% of the tumor mass and includes various components such as extracellular matrix (ECM), pancreatic stellate cells (PSC), fibroblasts, macrophages, blood and lymphatic vessels, pericytes, stem cells, and inflammatory cells.Recent research has focused on developing treatment strategies targeting stromal elements, particularly PSCs, which play a central role in pancreatic fibrogenesis (7)(8)(9)(10)(11)(12).PSCs are star-shaped, vitamin A-storing cells, which comprise approximately 4% of all pancreatic cells and show a periacinar distribution in a healthy pancreas.In response to several injuries (such as cancer or inflammation), PSCs undergo an activation process via exhibiting a myofibroblastic-like phenotype, expressing α-smooth muscle actin (α-SMA).Later they become the key cells in pancreatic fibrogenesis (13,14).
In many studies, PSCs have been shown to be present since the early stages of preneoplastic transformation.PSCs are activated and exhibit myofibroblastic morphology during epithelial carcinogenesis.This observation can be explained by three hypotheses: 1) PSCs may interact with tumor cells from the early stages of carcinogenesis, 2) they may attempt to circumvent and confine tumor cells from early preneoplastic stages, or 3) PSCs may activate around genetically defective cells in an initially limiting manner, then synthesize elements of the extracellular matrix (ECM) to make the stroma more fibrotic.This fibrosis can lead to hypoxic conditions in subsequent stages, resulting in the release of reactive oxygen species and genetic instability in epithelial cells (15).
Although there are clinical (oncology, surgery) and preclinical (biochemistry, pharmacology) studies related to PDAC stroma in the literature (15)(16)(17)(18) , publications based on histopathological examinations are few and serve only to complement the gaps between these studies (6,(19)(20)(21).This study aims to evaluate the relationship between α-SMA expression and the histopathological parameters of pancreatic ductal adenocarcinoma (PDAC), as well as to investigate the effects of these parameters on the prognosis.

MATERIALS and METHODS
All pancreatic resections examined in the Department of Pathology between 2011 and 2016 were re-evaluated retrospectively.All consecutive cases with a diagnosis of usual (classical or conventional) PDAC were included in the study (n=81).PDACs with nonconventional subtypes and those associated with intraductal papillary mucinous neoplasm (IPMN) or mucinous cystic neoplasm (MCN) were excluded.A Whipple resection had been performed in 62 cases, a total pancreatectomy in 6 cases, a distal pancreatectomy in 10 cases, and partial excisions in 3 cases.Information on histological tumor subtypes, grade, location, diameter, and lymphatic, vascular, and perineural invasion, as well as surgical margin status, was retrieved from pathology reports.After revision of all tumor slides, the pattern of stromal density was classified based on all hematoxylin and eosin (H&E) stained tumor slides as loose (myxoid), moderate (keloid-like), or dense (mature) according to the relevant literature (4)(5)(6).The definition of the final stromal score in these studies was not clear enough.Therefore, the scoring of the stromal pattern was based on the dominant density (≥ 50% of the total tumor stroma) for each case.Loose stroma was defined as a loose fibroblastic and myxoid stroma containing short, faint collagen fibers (Figure 1A).Moderate density stroma was composed of haphazard bands of keloid-like collagen (Figure 1B).Dense stroma was characterized by the dominance of mature, thick collagen fibers that resembled scar tissue (Figure 1C).For statistical purposes, the stroma was grouped as mixed (loose + moderate) and dense.The best representative formalinfixed paraffin-embedded (FFPE) tumor block was selected for immunohistochemistry.This block showed the dominant type of stromal density and did not contain necrosis.

Immunohistochemical Analysis
The tumor blocks were sectioned at 4 µm thickness as positively charged slides.For the processes, the Leica Bond-Max automatic immunohistochemistry device was used.To block the endogenous peroxidase activity, the sections were incubated for 10 min with three steps of at room temperature.Antigen retrieval was achieved with tris-EDTA buffer (pH 8) for α-SMA.The slides incubated with α-SMA (1:200, Clone 1A4, Cell Marque) for 45 minutes at room temperature.After primary antibody, secondary antibody and poly-HRP solutions (Bond Polymer Detection Kit; DS9800, Leica) were applied in the device.Diaminobenzidine (DAB) was used as the chromogen for colored visualization of the antigens.The sections were then counterstained with hematoxylin and dehydrated with ethanol.

Statistical Analysis
Histopathologic and clinical variables were compared in terms of the SMA score.Categorical variables were compared using the Pearson Chi-Square test.Continuous variables were compared using the Kruskal-Wallis test.
Cox-regression and Kaplan-Meier estimates were used to calculate the factors effecting survival.A multivariate Cox proportional hazards model was employed to identify independent prognostic factors.Statistical significance was set at a p-value of <0.05.All analyses were performed using jamovi (version 2.3) (22).
The total scores of α-SMA did not reveal any significant correlation when compared with other clinicopathological parameters.However, α-SMA levels were found to be higher in patients with angioinvasion than in those without angioinvasion (p = 0.01) (Table V).The stromal pattern exhibited mixed stromal density in 38 cases (47%), while 43 cases (53%) had a pure mature (dense) type stroma.Among the cases with pure mature (dense) stroma, 12 had low or moderate α-SMA scores, whereas 31 had high α-SMA scores.Of the cases with mixed stromal density, 15 had low or moderate α-SMA scores, and 23 had high α-SMA scores (Table IV).
α-SMA was found to be strongly positive around the glandular tumoral component in 56 cases (Figure 3A), as well as around the normal perilobular area in 24 cases (Figure 3B).Staining was also stronger in peritumoral chronic pancreatitis in 27 cases.Additionally, the stromal components

A B C
multivariate analysis (with p-values of 0.038, 0.029, and 0.003, respectively).However, tumor diameter did not have a significant effect on the T stage of the tumor.Surgical margin status approached significance but was not an independent factor (p=0.066) (Table VI).

Oncological Follow-Up
Forty-five cases were followed up at our institution.spectively.Due to the small number of cases treated with FOLFIRINOX (only two), and combination therapies (only two), the α-SMA scores could not be compared for different therapy regimens.
Sixteen cases experienced distant metastasis, with 11 of them spreading to the liver or lungs.Additionally, 24 cases had local recurrence.All of these cases with metastasis or local recurrence had moderate or high α-SMA scores in their primary tumor.
Two cases that received neoadjuvant therapy had high α-SMA scores, with lifetimes of 37 and 42 months, respectively.Among the cases treated with adjuvant FOL-FIRINOX, two had high α-SMA scores, with lifetimes of 16 and 37 months.Forty-one cases treated with adjuvant gemcitabine had moderate or high stromal α-SMA scores, with a mean lifetime of 24.7 months (range: 3-107 months) (Table VII).Two cases treated with the adjuvant gemcitabine and capecitabine combination had moderate-high SMA scores and the lifetimes were 32 and 10 months re- The final result was a worse tumor prognosis (26).Erkan et al. proposed an "activated stromal index" combining high levels of α-SMA and low levels of collagen, and found a relationship with a worse prognosis.However, they could not reveal a significant effect of α-SMA levels alone on survival (20).Wang et al. found that the effect of α-SMA expression on the survival of 145 resected PDACs treated with adjuvant gemcitabine was significant in univariate analysis, but they did not observe any significant results in multivariate analysis (6).
In our study, there was no significant association between the immunohistochemical expression of α-SMA and overall survival (univariable/multivariable: p=0.457/p=0.414).Only 45 cases were followed up in our institution, with 43 receiving gemcitabine-based adjuvant chemotherapy and 2 receiving FOLFIRINOX.Among them, 24 experienced recurrence, and the drug types chosen for recurrence varied (gemcitabine, XELOX, FOLFOX, paclitaxel, etc.).Since most of the cases were treated with the same first-line gemcitabine-based regimen, statistical analyses to assess the relationship between the dominant stromal pattern and/ or α-SMA density with response to therapy could not be performed.
Lymph nodes were not within the scope of our study; however, in some cases (n=12), they were present on the slides selected for immunohistochemistry.We observed that the stromal components of metastatic foci in lymph nodes were also strongly positive for α-SMA, indicating that the tumor carries these stromal features wherever it spreads.We believe that examining the stroma in metastatic areas such as the liver, lung, and lymph nodes should be the next step.
The major outcome of our study was the significant correlation between higher α-SMA scores and the presence of angioinvasion (p=0.01).Previous similar studies did not report this kind of correlation (6,19).Pancreatic stellate cells are also recognized to have angiogenic or antiangiogenic effects in both the early and late stages of the disease (27).The presence of such an association may suggest that tumors with higher α-SMA scores are at a higher risk for vascular invasion and local recurrence and/or distant metastasis.This result could be valuable in predicting the metastatic capacity of the tumor.However, due to the small size

DISCUSSION
The stroma of colorectal tumors has been the first to be proven to significantly influence the prognostic outcome.(23).Although it appeared to be an accurate method for calculation, this was reported as an expensive and non-reproducible method for scoring the stroma of PDAC.
Afterward, Wang et al. applied a modified method using hematoxylin-eosin with Masson's trichrome, and α-SMA for PDACs (6).They found that high stromal density was associated with a significantly better clinical outcome compared to patients with intermediate or low stromal density in multivariate analysis.The stromal pattern was defined as an independent poor prognostic factor (6).
Methods for examining stromal density vary widely in the literature, ranging from eyeballing to computer-based image analysis.We believe that a specific, reproducible, easy, and cost-effective method for assessing stromal density needs to be established through larger studies.
Pancreatic stellate cells (PSCs) become activated in situations such as inflammation and carcinogenesis, leading them to express α-SMA (13,14,24).We also observed stronger α-SMA staining in peritumoral chronic pancreatitis, which supports the thesis that pancreatic stellate cells (PSCs) are triggered by chronic inflammation (not shown here).
Fujita et al. studied 109 resection specimens of PDACs and discovered that high levels of α-SMA mRNA were associated with a worse prognosis.However, it is worth noting that the adjuvant chemotherapy regimens in this study were heterogeneous (19).Similar results were obtained in the CONKO-001 study (25).However, a "tissue microarray" assay was used for immunohistochemical staining in this study, which we believe may not adequately demonstrate the heterogeneous distribution of stromal α-SMA.The therapy regimens were classified as treated with or of our cohort, this preliminary finding did not reach statistical significance in the multivariate analysis (p=0.136).
We believe that the disparate results regarding the effects of α-SMA on survival in the literature are primarily attributable to the retrospective nature of the studies, as well as variations in tissue types, tissue sizes, antibody clones, and the quality of the cohorts.To mitigate these discrepancies, prospective studies and clinical trials should be designed based on analyses and scoring of stromal α-SMA.Additionally, our study primarily focuses on α-SMA expression at the protein level.Integrating molecular analyses, such as gene expression profiling or mutation status, could provide deeper insights into the underlying biological mechanisms driving the observed associations.

Figure 1 :
Figure 1: A) Myxoid or loose stroma of pancreatic ductal carcinoma.The periglandular area is pale, has less myofibroblasts, and looks edematous (Hematoxylin & Eosin, x10).B) Moderate or keloid-like stroma of pancreatic ductal carcinoma.Interrupted bands of dense eosinophilic keloid-like collagen are seen in the picture.This collagen is the product of activated myofibroblasts (Hematoxylin & Eosin, x40).C) Strong or mature stroma of pancreatic ductal carcinoma.Collagen replaced everything between and around the tumor cells.There are a few tumor cells at the top center of the figure (Hematoxylin & Eosin, x40).

Figure 3 :Figure 2 :
Figure 3: A) α-SMA staining was stronger at the immediate borders of the glandular tumoral component, and the interglandular area was less dense (α-SMA antibody, x10).B) Activated pancreatic stellate cells stained with α-SMA highlighting the normal perilobular area (α-SMA antibody, x40).C) Lymph node metastasis of a pancreatic ductal carcinoma.The same strong staining pattern was also seen in metastatic areas (α-SMA antibody, x10).A B C

Table I :
Scoring method of immunohistochemical staining for stromal α-SMA* * α-SMA: alpha smooth muscle actin

Table II :
The baseline characteristics of the patients and histopathological findings (n = 81) *AJCC: American Joint Committee on Cancer (T: Tumor stage, N: Lymph node stage)

Table III :
Distribution of immunohistochemical staining for stromal α-SMA in the cohort

Table IV :
Distribution of stromal pattern and α-SMA scores

Table V :
Relation of α-SMA and other pathological parameters

Table VI :
Multivariate survival analysis