Immunohistochemical Expression of Fatty Acid Synthase (FASN) is Correlated to Tumor Aggressiveness and Cellular Differentiation in Salivary Gland Carcinomas

Abstract

Fatty acid synthase (FASN) expression is closely related to cancer progression, in particular, tumor aggressiveness and poor prognosis. This study aimed to analyse the expression of FASN in carcinomas of the salivary glands and correlate it with Ki-67 expression. We analysed by immunohistochemistry the expression of FASN and Ki‐67 on tissue sections from 7 cases of adenocarcinoma, not otherwise specified (AdNOS), 6 cases of polymorphous adenocarcinoma (PAC), 16 cases of acinic cell carcinoma (AcCC), 19 cases of adenoid cystic carcinoma (AdCC), 15 cases of epithelial-myoepithelial carcinoma (EMC); 10 cases of secretory carcinoma (SC), 13 cases of mucoepidermoid carcinoma (MEC), 10 cases of salivary duct carcinoma (SDC) and 7 cases of myoepithelial carcinoma (MC). These carcinomas were classified into aggressive and indolent regarding their biological behaviour. Additionally, MEC and AdCC were also classified according to the histological grade. High expression of FASN was found in SDC (100%), SC (100%), AcCC (68.7%) and AdNOS (57.2%). No association was found between FASN and Ki-67 expression. Aggressive carcinomas showed a higher rate of Ki-67 proliferation (p < 0.001) and greater expression of FASN when compared to indolent carcinomas (p < 0.05). With regards to carcinomas categorized as indolent, FASN expression was much higher in the lesions that presented cell differentiation (SC and AcCC). Also, FASN expression was significantly higher in high-grade AdCC and MEC when compared to low-grade tumors (p < 0.05). We concluded that FASN expression was correlated to tumor aggressiveness and cellular differentiation in salivary gland carcinomas.

Introduction

Changes in the metabolism of tumor cells have been described nearly a century ago [1]. In the 1920s, Otto Warburg observed that malignant cells exhibited high glucose uptake and production of lactic acid even under aerobic conditions (Warburg effect) development [2, 3].

Tumor microenvironment is chronically hypoxic due to the disproportion between neoplastic growth and blood supply and this phenotypic alteration of the malignant cells is an essential advantage for tumor development [4]. Metabolic shift—that is, the change to a glycolytic and lipogenic phenotype—starts early in the tumor cells and increases as the cells progress to a higher degree of malignancy, sustaining their exacerbated proliferation [5–7].

Among the wide variety of genetic alterations, common phenotypes were recognized, such as the presence of metabolic alterations associated with fatty acid (FA) synthesis. In this scenario, one of the most frequent phenotypic changes in cancer cells is the overexpression of fatty acid synthase (FASN) enzyme [8, 9]. FASN is a homodimeric protein with a molecular weight of about 270 kDa, responsible for the endogenous synthesis of long-chain FA [9, 10]. It is the main enzyme acting on the anabolic conversion of carbohydrates to FA, being responsible for the NADPH-dependent condensation of the acetyl-Coa and malonyl-CoA substrates, producing palmitate—a FA with a 16-carbon chain [9, 10].

The expression of FASN is low or absent in most normal tissues, except in hepatic tissue, breast tissue during lactation, endometrial tissue in the proliferative phase, lungs tissue of newborns [11, 12], and in the lipogenic tissues, since the majority of FA used by adipose cells comes from the diet [9, 13, 14]. Several tumors have increased FASN activity including hepatocellular [8, 9], breast [11], ovary [15], prostate [16, 17], lung [18], esophagus [19], stomach [20], bladder [21], oral [22], head and neck, thyroid carcinomas, as well as in soft tissue sarcoma, cutaneous melanocytic neoplasm, including melanoma [8, 9]. Upon FASN expression, these cancers achieve a metabolic adaptation that promotes advantages in progression, survival, and metastasis [9] reflecting tumor aggressiveness and a poor prognosis [23].

There are very few studies investigating the expression of FASN in salivary gland tumors [24–27]. Besides, cell proliferation—a feature common to all cancers—requires FA for the synthesis of membranes and signaling molecules [28]. Considering the importance of this enzyme in tumor biology, the purpose of this study was to correlate FASN expression with the Ki-67 cell proliferation index in salivary gland carcinomas.

Materials and Methods

Tissue Samples

This study was approved by the Ethical Committee of the University of Campinas (Process number 2.115.019). A total of 103 cases of salivary gland carcinomas collected from 1990 to 2014 were analysed. The diagnoses of all cases were reviewed in paraffin sections stained with haematoxylin–eosin (HE) according to the current World Health Organization (WHO) classification of head and neck tumors [29].

Thus, the study included after classification 7 cases of adenocarcinoma, not otherwise specified (AdNOS), 6 cases of polymorphous adenocarcinoma (PAC), 16 cases of acinic cell carcinoma (AcCC), 9 cases of low-grade adenoid cystic carcinoma (AdCC), 10 cases of high-grade AdCC, 15 cases of epithelial-myoepithelial carcinoma (EMC), 10 cases of secretory carcinoma (SC), 4 cases of low-grade mucoepidermoid carcinoma (MEC), 4 cases of intermediate-grade MEC, 5 cases of high-grade of MEC, 10 cases of salivary duct carcinoma (SDC) and 7 cases of myoepithelial carcinoma (MC).

MEC and AdCC were classified microscopically as to their histopathological grade as they are the most common salivary gland carcinomas. Given that there is no widely used grading system, MEC was classified into low-grade (admixture of mucus cells, intermediate cells, and epidermoid cells, cystic component (> 20%), lack of neural invasion, necrosis or cellular anaplasia, and rare mitotic figures), intermediate-grade (predominance of intermediate cells, less prominent cystic formation (< 20%), cell atypia may or may not be observed), and high-grade (predominance of intermediate and epidermoid cells, no cystic component, frequent atypical mitotic figures, nuclear anaplasia, and necrosis) according to methods described by Armed Forces Institute of Pathology (AFIP) [28].

AdCC was classified into low-grade and high-grade according to the grading system of Szanto et al. [29]. Concerning this definition, low-grade AdCC consists of grade I (predominantly tubular, no solid) and II (predominantly cribriform, < 30% solid). High-grade AdCC consists of grade III (> 30% solid component).

Additionally, the carcinomas were classified regarding their biological behaviour into aggressive and indolent according to the current WHO classification of head and neck tumors [29]. In our casuistic, carcinomas classified as having an aggressive clinical course were: AdNOS, high-grade AdCC, high-grade MEC, and SDC. Carcinomas classified as indolent were: AcCC, PAC, low-grade AdCC, low-grade MEC, intermediate-grade MEC, EMC, MC, and SC.

Immunohistochemistry

Immunohistochemistry staining was performed in 3‐μm thick tissue sections. After deparaffinization and rehydration in ethanol solutions, antigen retrieval was performed by immersing the slides in phosphate-buffered saline (PBS) with Tris–EDTA buffer solution (pH 8.9). Endogenous peroxidase activity was blocked through 3 immersion baths of 10% hydrogen peroxide solution (3 min each). The slides were incubated with primary antibodies FASN (Sigma‐Aldrich, clone 3B3-1D6, dilution 1:200, St Louis, MO, USA) and Ki‐67 (Dako®, clone MIB‐1, dilution 1:500, Glostrup, Denmark). Detection of the primary antibody was performed using the EnVision Plus System (Dako, K4001, Glostrup, Denmark) and developed with chromogenic substrate diaminobenzidine‐hydrochloride (DAB) (Sigma‐Aldrich, St Louis, MO, USA). Lastly, the slides were counterstained with Harris hematoxylin.

Nuclear images analysis for Ki-67

The slides immunolabelled with the Ki-67 antigen were digitized (AperioScanscopeCS® Slide Scanner—Aperio Technologies, Inc., Vista, CA, USA) and visualized in the ImageScope program (Aperio Technologies). Three areas of the sample (“hot spots”) were selected using the Nuclear V9® algorithm program (Aperio Technologies Inc., Vista, CA, USA). The program performed the expression index calculation of the Ki-67 antigen, obtained from the relationship between the number of positive nuclei and the total number of tumor cell nuclei. To avoid possible false-positives the pen tool was used to isolate the tumor cells from the adjacent inflammatory infiltrate. The positive pen tool was used to demarcate the tumor area to be quantified, and the negative pen tool to demarcate the area not to be included in the quantification. The results were provided in the percentage of positive nuclei.

Semiquantitative Analysis for FASN

The evaluation of the positivity and the intensity of the FASN expression were analysed by three researchers with recognized experience in the field. The quantify and the intensity of FASN expression were classified in accordance with other studies [27, 30] as score 0: absent staining or positivity less than 10% of cells examined, with regards to positive cases; score 1: positivity less than 50% of cells examined; and score 2: positivity greater than or equal to 50% of the cells examined. The intensity of the staining was subjectively evaluated as weak/moderate (1 +) or intense (2 + +).

Statistical Analysis

Data were statistically analyzed using Wilcoxon’s, Mann–Whitney, and Kruskal–Wallis non-parametric tests. Fisher's test was used to compare positive cases for FASN between aggressive and indolent carcinomas. Spearman’s rank correlation coefficient was applied between continuous variables. Values less than 0.05 were considered significant. The power of the tests used in the data analysis was equivalent to Power (1–β) = 0.8, software G*Power 3.1.9.4.

Results

FASN and Ki-67 Expression in Salivary Gland Carcinomas

In general, FASN expression was positive in all cases of SDC and SC, 68.7% of AcCC cases, and 57.2% of AdNOS cases. FASN expression was intense (2 + +) in all cases of SC and most cases of SDC and AdNOS. Only 20% of the EMC expressed FASN, with weak/moderate expression intensity (1 +) in 13.3% of the cases. Besides, FASN expression was negative in all cases of PAC and MC. Regarding Ki-67, its expression was higher in SDC (15.1%), followed by MEC (11.8%) and AdCC (10.9%). MC, PAC, AcCC, and SC were the tumors with the lowest cell proliferation rates. Statistical analysis did not reveal a correlation between the values of FASN expression and proliferation rate (Ki-67) across tumors (r = 0.28; p = 0.15) (Table 1).

Table 1.

Immunohistochemical expression of FASN and Ki-67 in salivary gland carcinomas

		FASN					Ki-67	Spearman correlation test
		Amount of positive cells			Intensity of the staining		Average
Histologic type	Total cases	Score 0	Score 1	Score 2	1 +	2 +  +	%
AdNOS	7	3 (42.8%)	0	4 (57.2%)	1 (14.4%)	3 (42.8%)	3.3	r = 0.28; p = 0.15
PAC	6	6 (100%)	0	0	0	0	0.5
AcCC	16	5 (31.3%)	2 (12.5%)	9 (56.2%)	11 (68.7%)	0	1.2
AdCC	19	16 (84.2%)	2 (10.5%)	1 (5.2%)	2 (10.5%)	1 (5.2%)	10.9
EMC	15	12 (80%)	1 (6.7%)	2 (13.3%)	2 (13.3%)	1 (6.7%)	3.4
SC	10	0	0	10 (100%)	0	10 (100%)	2.5
MEC	13	10 (46.1%)	1 (7.7%)	2 (15.4%)	1 (7.7)	2 (15.4%)	11.8
SDC	10	0	1 (10%)	9 (90%)	2 (20%)	8 (80%)	15.1
MC	7	7 (100%)	0	0	0	0	0

AdNOS adenocarcinoma not otherwise specified, PAC polymorphous adenocarcinoma, AcCC acinar cell carcinoma, AdCC adenoid-cystic carcinoma, EMC epithelial-myoepithelial carcinoma, SC secretory carcinoma, MEC mucoepidermoid carcinoma, SDC salivary duct carcinoma, MC myoepithelial carcinoma, FASN fatty acid synthase

FASN and Ki-67 Expression According to Clinical Behaviour

Bearing in mind that the behaviour and clinical course of salivary gland carcinomas can be variable, we divided our sample into indolent and aggressive carcinomas (Table 2). Aggressive tumors (Fig. 1) showed a higher rate of Ki-67 proliferation (p < 0.001) and greater expression of FASN when compared to indolent carcinomas (p < 0.05) (Fig. 2). With regards to carcinomas categorized as indolent, FASN expression was much higher in the lesions that presented cell differentiation: lactation-like secretory differentiation (SC) and serous acinar cell differentiation (AcCC).

Table 2.

Correlation between FASN and Ki-67 expression according to histological grade of MEC and AdCC

Histological grade	MEC	AdCC	FASN + cases (%)	Ki-67 index#
Low-grade	4	9	0 (0.0)a	7.89 (0.0–14.1)c
Intermediate-grade	4	–	1 (25.0)	0.69 (0.2–1.7)c
High-grade	5	10	5 (33.3)b	16.26 (10.2–19.4)d

AdCC adenoid cystic carcinoma, MEC mucoepidermoid carcinoma, FASN fatty acid synthase

P < 0.05 (^a versus ^b)

P < 0.005 (^c versus ^d)

^#Median (Q25-Q75)

Fig. 1.

Salivary gland carcinomas with aggressive clinical behaviour: a–d Histopathological features (HE). a High-grade MEC (original magnification 10×) inset: malignant and pleomorphic cells arranged in a solid pattern with the absence of cystic component (original magnification 40×). b High-grade AdCC (original magnification 10×) inset: Epithelial and myoepithelial neoplastic cells arranged in a solid pattern (> 30%) original magnification 40×) c AdNOS (original magnification 10×) inset: tumoral islands with atypic and pleomorphic cells (original magnification 40×) d SDC (original magnification 10×) inset: large ducts with cribriform and Roman bridge-like features (original magnification 40×). e–h Ki-67 expression (original magnification 40×). Positivity in high-grade MEC (e), high-grade AdCC (f), AdNOS (g), and SDC (h). i–l FASN expression (original magnification 40×). Strong cytoplasmic staining in neoplastic cells of high-grade MEC (i), high-grade AdCC (j), AdNOS (k), and SDC (l)

Fig. 2.

Salivary gland carcinomas with indolent clinical behaviour. a–d Histopathological features (HE). a SC (original magnification 10×) inset: tumor cells arranged in a ductal pattern with bubbly secretions, creating an open appearance (original magnification 40×) b AcCC (original magnification 10×) inset: serous acinar cells—large, polygonal cells with abundant lightly basophilic, granular cytoplasm (original magnification 40×) c intermediate-grade MEC (original magnification 10×) inset: an admixture of mucous cells, intermediate cells, epidermoid cells, and cystic component (> 20%) (original magnification 40×) d EMC (original magnification 10x) inset: biphasic arrangement of malignant cells in luminal ductal structures with dense eosinophilic cytoplasm and abluminal polygonal cells with clear cytoplasm (original magnification 40×). e–h Ki-67 expression (original magnification 40×). Low Ki-67 expression in SC (e), AcCC (f), intermediate-grade MEC (g), and EMC (h). i–l FASN expression (original magnification 40×). Strong cytoplasmic staining in SC (i), and AcCC (j), positivity in MEC intermediate-grade (k), and some cases of EMC (l)

Correlation Between FASN and Ki-67 Expression According to Histological Grade of MEC and AdCC

As MEC and AdCC are commonly classified microscopically as to their histopathological grade, in our study, we subdivided our sample of MEC and AdCC by histological grade. Only high-grade AdCC were positive for FASN. With regards to MEC, only two cases of high-grade and one of intermediate-grade were positive for FASN. We noticed, therefore, that FASN expression was significantly higher in high-grade AdCC and MEC when compared to low-grade tumors (p < 0.05). They also had higher rates of cell proliferation (Ki-67) than those of low/intermediate-grade (p < 0.001) (Table 3).

Table 3.

Immunohistochemical expression of FASN and Ki-67 according to clinical behaviour of salivary gland carcinomas

Clinical behaviour	Number of cases	FASN + cases (%)	Ki-67 index#
Agressive*1	32	18 (56.3)a	13.91 (2.8–18.7)c
Indolent*2	71	25 (35.2)b	0.52 (0.07–2.8)d

^*1Aggressive: AdNOS, high-grade AdCC, high-grade MEC, and SDC

^*2Indolent: AcCC, PAC, low-grade AdCC, low-grade MEC, intermediate-grade MEC, EMC, MC and SC

FASN fatty acid synthase

P < 0.05 (^a versus ^b, Fisher’s exact test)

P < 0.001 (^c versus ^d, Mann–Whitney Test)

^#Median (Q25-Q75)

Discussion

FASN is related to lipid metabolism, signal transduction pathways, regulation, proliferation, and survival of malignant cells [8–10]. Recently, our group showed in a series of pleomorphic adenomas (PA) and carcinomas ex-pleomorphic adenomas (CXPA) that the amount of FASN was higher in malignant tissues [27]. In the present study, in an attempt to better understand the expression pattern of FASN protein in different malignant salivary gland tissues, we selected a series of 103 salivary gland carcinomas and we demonstrated that de novo FA synthesis is increased in aggressive carcinomas (AdNOS, SDC, high-grade AdCC, and high-grade MEC). These findings corroborate with other studies [8, 21, 31–33], where the most aggressive cancers had a greater energy demand and higher demand for FA synthesis for their growth, increasing expression of FASN.

To the best of our knowledge, we are the first to evaluate the expression of FASN in AdNOS. The expression of FASN in SDC was reported recently by Hirai et al. (2020) who showed that almost all their 147 cases of SDC expressed FASN, which were greater in SDC ex-PA than in SDC de novo [34]. In fact, in CXPA, we recently showed that a greater FASN gene expression was found in the SDC subtype [27].

Regarding AdCC and MEC, we found that FASN expression was higher in high-grade carcinomas (p < 0.05). Other studies have shown median FASN expression values in AdCC when compared to other tumours [24, 25] and increased expression of FASN in high-grade MEC has been reported in the past [24]. Of the 30 MEC analyzed by immunohistochemistry, Ito and colleagues showed that the expression of FASN protein was higher in high grade tumors. In contrast, do Prado et al. [25] in 2011 had slightly different results, but still consistent with our findings. Even though FASN expression was statistically lower in malignant tumors, when we analyzed the 14 cases of MEC analyzed in their study, no mucous cells (most common in low and intermediate-grade MEC) were positive for FASN. On the other hand, a variable expression of FASN was found in intermediate and epithelial cells (much more common in high-grade MEC). These data corroborate with our results which showed FASN expression only in high grade and intermediate-grade of MEC.

It is important to emphasize a finding published by our group [26, 27], where FASN protein was found to be more expressed in epithelial cells as compared to myoepithelial cells in CXPA. In fact, in our series, tumors that showed the highest expression of FASN were originating in the epithelial cell, namely: SC (100%), SDC (90%), and AdNOS (57.2%). We believe that FASN expression in these structures is related to protein affinity for the epithelial cell and, possibly, to the metabolic shift in transformed neoplastic cells. On the other hand, MC, originating purely from the myoepithelial cells, showed no expression for the FASN protein.

Curiously, concerning the carcinomas categorized as indolent, the increase in FASN expression occurred only in the lesions that presented cell differentiation: lactation-like secretory differentiation (SC) and serous acinar cell differentiation (AcCC). In SC, this result may reflect an inherent characteristic of these tumors that share morphological similarities with milk-secreting mammary epithelial cells (lactation-like secretory differentiation) [30]. The lactating mammary gland is one of the few adult tissues that strongly stimulate de novo FA synthesis after physiological stimulation, and the expression of lipogenic genes (such as FASN) is strongly induced during this period [35].

AcCC, whose expression of FASN was expressive (68.7%), shows serous acinar cell differentiation [36]. Corroborating with our previous findings, we showed that 100% and 85.7% of SC presented large lipid droplets and AcCC presented small lipid droplets of adipophilin expression [30]. As the adipophilin staining reflects the accumulation of lipids [37] and as FASN is known to be a central enzyme in de novo lipogenesis, we believe that in the carcinomas, FASN is involved in the induction of lipids accumulation in these cells. This result leads us to believe that FASN may be related not only to proliferative activity but also to the process of cell differentiation. More studies are needed to elucidate the role of FASN expression in these carcinomas.

In the context of cancer, neoplastic cell proliferation requires FA for the synthesis of membranes and signaling molecules [28]. However, when associating FASN expression and the cell proliferation rate (Ki-67) of all carcinomas, we did not find a correlation between these markers. Do Prado and collaborators [25] only found a correlation between FASN and Ki-67 in the PA. In all other tumors studied (which included MEC and AdCC), no correlation was found among tumors. Similarly, Díaz et al. [25] also did not find a correlation between FASN and Ki-67 when analyzed a series of PA and CXPA. However, in our study, aggressive tumors, which showed significantly higher expression of FASN (p < 0.05), were also the ones with the highest rate of proliferation (p < 0.001).

In summary, despite the lack of available data to correlate the expression of FASN protein and the patient’s clinical information, we believe that the expression of FASN in aggressive salivary gland carcinomas could be suggest used as a potential therapeutic target for the treatment of these tumors. Inhibition of FASN activity by drugs that block its function leads to inhibition of cell cycle progression and consequent apoptosis of malignant cells in various cancers [10, 38, 39]. Thus, the alteration of the lipid metabolism, presented by these salivary carcinomas is interesting from the therapeutic point of view and deserves attention. However, further studies are needed to assess the therapeutic applicability of FASN inhibitors in salivary gland carcinomas.

Conclusions

Our findings indicate that FASN expression in malignant neoplasms of the salivary gland is correlated to tumor aggressiveness and cellular differentiation. We also believe that, due to the scarcity of studies correlating FASN and salivary gland tumors, our results may contribute to a better understanding of the knowledge about the metabolism of neoplastic cells in these carcinomas. Lastly, this work may be helpful to the understanding of salivary gland carcinomas metabolism and the potential use of FASN inhibitors to treat patients.

Author Contributions

Conception and design of the study: Camila Matsunaga de Angelis, Reydson Alcides de Lima-Souza, João Figueira Scarini, Albina Altemani and Fernanda Viviane Mariano. Selection of cases, histological classification, and immunohistochemical reactions: Camila Matsunaga de Angelis, Reydson Alcides de Lima-Souza, and João Figueira Scarini. Cells counting: Gleyson Kleber do Amaral-Silva. Statistical Analysis: Rogério de Oliveira Gondak. Analysis and interpretation of results, and drafting the manuscript: Camila Matsunaga de Angelis, Reydson Alcides de Lima-Souza, João Figueira Scarini, and Erika Said Abu Egal. Revising the manuscript critically for important intellectual content: Oslei Paes de Almeida, Carlos Takahiro Chone, Luiz Paulo Kowalski, Albina Altemani, Fernanda Viviane Mariano. All authors read and approved the final manuscript.

Funding

This study was supported by São Paulo Research Foundation (FAPESP), grants numbers 2015/07304-0, 2019/0698-2, 2019/09419-0, and by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001.

Declarations

Conflict of Interest

Neither of the authors have any conflict of interest to declare.

Ethical Approval

All procedures performed in this study were in accordance with the ethical standards of the Ethical Committee of the University of Campinas (Process number 2.115.019).