Expression of the Lipogenic Enzyme Fatty Acid Synthase (FAS) in Retinoblastoma and Its Correlation with Tumor Aggressiveness

Abstract

purpose. Fatty acid synthase (FAS) performs the anabolic conversion of dietary carbohydrate or protein to fatty acids. Many common human cancers express high levels of FAS, and its differential expression between normal and neoplastic tissues has led to the consideration of FAS as a target for anticancer therapy. To investigate the potential of targeting FAS in the treatment of retinoblastoma, we first determined whether FAS was activated in this human tumor. Moreover, correlation of FAS expression with tumor aggressiveness was determined.

methods. FAS reactivity was evaluated by immunohistochemistry in 66 retinoblastoma specimens from 65 patients. Degree of tumor differentiation, choroid invasion, optic nerve infiltration, mitotic rate, and necrosis extension were estimated. FAS expression was correlated with all these tumor characteristics by means of parametric and nonparametric statistical analyses.

results. Eighty-two percent of tumors were FAS positive. Stronger FAS expression correlated with more advanced choroid (P < 0.001) and optic nerve (P = 0.016) invasion, high mitotic index (P < 0.001), and less differentiated histology (P = 0.047). Correlation with extension of necrosis was not statistically significant. Unaffected retina was negative.

conclusions. The data suggest that expression of FAS and fatty acid synthesis support an essential functional aspect of retinoblastoma cells, perhaps cell growth or survival. FAS activation may serve as a novel target for systemic and local antineoplastic therapy and, because it increases with tumor aggressiveness, its inhibition could represent an alternative treatment strategy in advanced and resistant retinoblastomas.

Retinoblastoma (RTB) is the most common intraocular malignant tumor in childhood, with a prevalence of 1/20,000 live births in developed countries. Although aggressive, the disease associated with this tumor has changed from nearly always fatal in the early past century to one with a survival rate of 95%. ¹ Today, treatment strategies are designed to prevent metastatic spread and to deliver optimal local therapy to avoid enucleation and unnecessary therapy. ² Cryotherapy, thermotherapy, laser photocoagulation, and plaque radiation therapy have been used in selected patients with encouraging results, ^{2 3 4 5 6} and research is directed toward novel treatment modalities ^{4 5 6} to improve effectiveness of a conservative approach. 

One therapeutic target that has not previously been considered for RTB treatment is the pathway for endogenous fatty acid synthesis. ⁷ Human fatty acid synthase (FAS), the principal enzyme in this pathway, is a dimer of two identical multifunctional proteins generating two active centers for the conversion of acetyl-coenzyme A (CoA) and malonyl-CoA to palmitate. ⁸ In normal human tissues, FAS is generally downregulated, because cells preferentially use circulating dietary fatty acids for the synthesis of new structural lipids. ^{7 8 9 10 11 12} On the other hand, FAS is highly expressed in many human tumors, ⁷ such as carcinoma of the breast, ^{13 14 15} prostate, ^{16 17 18 19} colon, ²⁰ ovary, ²¹ endometrium, ²² mesothelium, ²³ thyroid, ²⁴ lung, ²⁵ and stomach. ²⁶ In some of them, a high degree of activation correlates with poor outcome, suggesting a relationship between the expression of FAS and the tumor’s aggressiveness. ^{14 17 18 21 24} The preferential expression of FAS in cancer cells suggests that the fatty acid synthetic pathway may constitute a promising target for anti-FAS drug development. ^{23 27 28 29 30 31 32 33 34 35 36}  

The purpose of this study was to investigate the expression of FAS in RTB to assess its potential as a therapeutic target. 

Materials and Methods

Patients and Histology

Among all patients admitted to the two participating institutions between 1988 and 2001 with a diagnosis of RTB, we retrospectively identified 65 children whose treatment had included enucleation (“Bambino Gesù” Children’s Hospital contributed 47 cases, and the University of Siena 18). Data regarding sex, age, laterality, family history, treatment, and clinical course were recorded. Follow-up was calculated from date of diagnosis. 

All tumor slides were revised and staged according to the criteria of the Italian Association of Pediatric Hematology and Oncology (AIEOP) RTB-92 protocol (revised November 1999) (Table 1) . RTBs were microscopically graded in three groups according to the predominant pattern of differentiation: well-differentiated, when more than 75% of the tumor was composed of Flexner-Wintersteiner and Homer Wright rosettes; diffuse, when more than75% of the tumor showed undifferentiated cells; and mixed, when both features were present. Mitotic index was calculated as the sum of mitoses counted in 10 high-power fields (HPFs: 40×). Extension of necrosis was evaluated as a percentage of the whole tumor mass. 

Immunohistochemistry

Sixty-six globular specimens were analyzed. Four-micrometer sections from formalin-fixed paraffin-embedded tissue were dewaxed in xylene, rehydrated in ethanol dilution series, and incubated for 20 minutes in 3% hydrogen peroxide to block endogenous peroxidase. After washes in TBS, the sections were treated with 20% normal bovine serum and then incubated at room temperature for 60 minutes with an affinity-purified rabbit polyclonal antiserum raised against human FAS (anti-FAS antibody was a gift of Francis P. Kuhajda, MD, (Department of Pathology, The Johns Hopkins University School of Medicine, Baltimore, MD; concentrated serum dilution 1:3000). After they were washed, the slides were reincubated with biotinylated anti-rabbit IgG (Dako A/S, Glostrup, Denmark) at room temperature for 30 minutes, followed by incubation with an avidin and biotinylated horseradish peroxidase complex (Dako) at room temperature for 30 minutes. Chromogenic development was obtained by using either 3,3′-diaminobenzidine tetrahydrochloride (DAB) with 0.03% hydrogen peroxidase (Dako) or 3-amino-9-ethylcarbazole (Dako). Finally, sections were counterstained with hematoxylin, cleared, and mounted. Sections without primary antibodies served as the negative control, whereas periocular fat tissue served as the positive control. FAS expression was evaluated independently and semiquantitatively by two pathologists, without knowledge of the patients’ characteristics. The staining was scored on a four-tiered scale: 0, negative (0%–5% positive cells); 1+, weak to moderate (6%–30%); 2+, intense (31%–50%); and 3+, very intense (>50%). Because FAS reaction was not uniformly distributed, 15 vital tumor fields were randomly selected, and a final mean score for each tumor was obtained. Periocular fatty tissue was arbitrarily considered as very intensely stained (3+). 

Statistical Analysis

For statistical analysis, choroidal invasion, optic nerve infiltration, tumor differentiation, and prior chemotherapy were considered to be categorical variables (C0–C4, N0–N4, 1–3, and yes/no, respectively) and were analyzed separately. Mitotic index and percentage of necrosis were considered to be continuous values. The correlation between FAS expression and these parameters was evaluated through the Spearman ρ index for categorical variables and through the Pearson ρ index for continuous variables. 

Results

Patients

Clinical features are summarized in Table 2 . Leukocoria and/or strabismus were observed as initial symptoms in 51 (78%) of the children, with the remaining patients having other clinical features. 

In 50 unilateral cases, treatment approach varied according to time of recruitment. The 22 earliest patients underwent surgery at diagnosis. Eighteen children were referred after surgery had been performed elsewhere. The most recent 10 patients had undergone preoperative chemotherapy (vincristine, etoposide, and carboplatin), laser photocoagulation, and cryotherapy in an attempt at eye-sparing. Surgical enucleation was performed in all children and, according to pathologic staging, postoperative therapy was administered. Two patients with stage I and 44 early stage II (C0–C2 and N0–N2) disease did not receive any treatment, and three with advanced stage II (C3 or N3) disease and one with stage III disease underwent chemotherapy (carboplatin and etoposide) and radiotherapy (40 Gy). 

In 15 children with bilateral tumors, enucleation of the more involved eye was followed by conservative treatment of the less involved one (chemotherapy, laser photocoagulation, and radiotherapy). One child underwent bilateral enucleation at different time points during treatment. In four orbital recurrences, cyclophosphamide, etoposide, carboplatin, and thiotepa (CECAT protocol), along with external beam radiation (40 Gy), were administered. In three metastatic RTBs, the CECAT protocol, intrathecal therapy (methotrexate and dexamethasone), and craniospinal radiation were used. 

Close follow-up during and after treatment included ophthalmic examinations and blood counts and chemistry; no extrahematologic toxicity was recorded. Mean duration of follow-up was 51 months (range, 6–195); 3 (4.6%) of 65 patients died of the disease. 

Histology and Immunohistochemistry

Of 66 RTBs, 2 tumors developed in microphthalmic eyes and could not be properly staged, because rudimentary ocular structures did not allow evaluation of local invasion (they were classified as V1, Cx, Nx). Two RTBs were stage I, 61 were stage II (with variable degrees of local infiltration), and 1 that had invaded extraocular structures was stage III. The majority of tumors (44%) showed a mixed pattern of differentiation, whereas diffuse and well-differentiated RTBs represented 39% and 17%, respectively. In five RTBs there was no evidence of necrosis, whereas in the remaining tumors (87%) a variable percentage of necrosis (from 5% to 80%) was present. Mean mitotic index per 10 HPFs ranged from 14 to 90. 

FAS immunostaining was positive in 54 (82%) of 66 tumors (Table 3 ,Fig. 1 ): 31 with moderate, 16 with strong, and 7 with very intense reactions. Undifferentiated areas stained more intensely than differentiated ones (Fig. 1A) , even though Flexner-Wintersteiner rosettes were often positive (Fig. 1D) . Anti-FAS positivity was more intense in neoplastic cells around vessels and weaker or even absent in the peripheral regions, close to areas of necrosis (Fig. C) . The correlation between more intense FAS immunostaining and higher mitotic rate (discussed later) and its prevalent perivascular distribution suggests that the activation of FAS was probably a consequence of increased fatty acid demand by intensely proliferating cells. In contrast, perinecrotic neoplastic cells did not express FAS probably because of slowing cellular metabolism before cell death. 

Normal retina (Figs. 1A 1C 1D) and conjunctival epithelium (Fig. 1B) were negative in all examined tumors. 

Statistical Findings

Spearman ρ correlation indexes between increasing FAS expression and choroidal invasion and between FAS expression and optic nerve infiltration were estimated as 0.431 (P < 0.001) and 0.299 (P = 0.016), respectively. There was a slight trend of less differentiated RTBs toward more intense anti-FAS reaction (Spearman ρ = 0.245, P = 0.047). Correlation between increasing expression of FAS and mitotic index as estimated through the Pearson ρ index was 0.0539 (P < 0.001). No statistically significant correlation was observed between tumor necrosis extension and prior chemotherapy administration. 

Discussion

The evidence of FAS overexpression in RTB parallels other observations of increased FAS expression in a variety of human cancers. ^{7 13 14 15 16 17 18 19 20 21 22 23 24 25 26} Although the mechanism underlying its upregulation in neoplastic cells is not completely elucidated, recent studies have revealed that other enzymes of the same pathway are also activated. ^{14 18} These findings suggest that changes in regulation of the lipogenic program, rather than in the FAS locus itself, are likely to be responsible for increased fatty acid synthesis in cancers. Growth factors, hormones, and cytokines are physiologically capable of regulating lipogenesis through the mitogen-activated protein (MAP) kinase or the phosphatidylinositol (PI) 3-kinase signaling cascades. In cancers, a variety of oncogenic changes involving growth factors and transduction pathways may act through the same signaling cascades. The activation of these transduction pathways increases the levels of active sterol regulatory element-binding protein (SREBP)-1, a transcription factor that upregulates the expression of genes involved in the synthesis of fatty acids, including FAS. ³⁷ Recent studies have demonstrated a direct relationship between activation of the MAP and PI 3-kinase pathways, increased SREBP-1 levels, and FAS overexpression in neoplastic cell lines, ^{37 38 39} although the effective extent to which this mechanism contributes in vivo is probably more complex than in experimental models. 

The biological advantage of endogenous fatty acid synthesis is probably the supply of structural constituents for membrane synthesis by actively proliferating cells. ^{7 11}  

Many studies have shown that pharmacological inhibition at the FAS step leads to a slowing down of cellular growth, ultimately triggering apoptosis. ^{23 27 28 29 30 31 33 34 35 36} In vitro tests have shown that FAS-blocked neoplastic cells did not grow, even when physiologic amounts of fatty acids were available. ^{27 28 29 34 35 36} In vivo studies have revealed significant antitumor activity of FAS inhibitors against human mesothelioma, ²³ human breast, ²⁹ and human ovary cancer mice ³³ xenografts expressing high levels of this lipogenic enzyme. The cytostatic effects of FAS inhibition may derive from the product depletion, depriving highly proliferating cells of their energetic support. In parallel, the link with apoptosis is probably due to intracytoplasmic accumulation of the FAS substrate malonyl-CoA, ^{29 30} whose synthesis is continued by acetyl-CoA carboxylase and driven in part by the decreasing fatty acid synthesis itself. ⁷  

Along with its potential therapeutic role, FAS also has been a reliable prognostic marker in many malignancies, in that increased expression correlates significantly with tumor aggressiveness. ^{14 17 18 21 24}  

The importance of identifying tumors that overexpress FAS, has led some investigators to develop an enzyme-linked immunosorbent assay (ELISA) method to measure FAS concentration in serum. ⁴⁰ This indicator gives an indirect estimate of FAS activation in the tissue and can provide important information about prognosis and the therapeutic setting of a tumor, without the need to obtain a surgical specimen. 

With the introduction of better designed chemotherapy protocols and more sophisticated local therapies, the treatment of RTB has shifted from a predominantly surgical approach to one incorporating both systemic chemotherapy and local treatment. The combination of different therapeutic modalities has been very effective, ^{2 3 4 5 6} and today the goals of clinicians have shifted to preserving the eye (and, it is hoped, vision), avoiding external beam radiation, reducing systemic therapy in favor of locally administered chemotherapy, and identifying new anticancer targets. In this context, the finding that the majority of RTBs (82%) expressed FAS hints of a possible role of FAS inhibition as a new therapeutic tool for several reasons. First, because normal retina cells do not express FAS, collateral damage to normal tissue could be avoided. Second, because FAS is more expressed in advanced RTBs, it may become a target for salvage treatment strategies in aggressive cases. Third, because some RTBs express the multidrug-resistance phenotype, ^{3 41 42 43} anti-FAS agent application could contribute to overcoming resistance. To assess FAS activation, the same ELISA method tested in human serum ⁴⁰ could be assayed in aqueous fluid and vitreous humor. This diagnostic procedure could be performed at the onset, avoiding the need to perform a biopsy. 

In conclusion, we believe that RTB should be included among tumors potentially sensitive to anti-FAS agents and that it would be advisable to perform additional experiments in animals. Of course, clinical investigations in the form of phase I and II studies are needed to evaluate the feasibility and efficacy of such an approach in children affected by this tumor. 

 Submitted for publication September 11, 2002; revised December 24, 2002; accepted January 8, 2003.

 Disclosure: F. Diomedi Camassei, None; R. Cozza, None; A. Acquaviva, None; A. Jenkner, None; L. Ravà, None; R. Gareri, None; A. Donfrancesco, None; C. Bosman, None; P. Vadalà, None; T. Hadjistilianou, None; R. Boldrini, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Table 1.

 

 

Synthetic AIEOP/RTB-92 Pathologic Classification (Revised 1999)

Table 1.

 

Synthetic AIEOP/RTB-92 Pathologic Classification (Revised 1999)

	Mass	Choroidal Invasion	Optic Nerve Invasion
Stage I	R1: unique	C0: none	N0: none
Intraocular	R2: multiple
RI < 50%	V0: no mass
RD minimum	V1: seedlings
	V2: mass
Stage II	V0: no mass	C0: none	N0: none
Intraocular	V1: seedlings	C1: inner half	N1: within lc
RI > 50%	V2: mass	C2: whole, CB, I	N2: involving lc
RD extensive		C3: sclera, Vv	N3: beyond lc
Stage III		C4: extrascleral	N4: whole length
Extraocular within orbit
Stage IV	CNS extension			N5: N sheets (L−)
Beyond orbit			N6: N sheets (L+)
			N7: chiasm
Stage V		Metastasis
Diffuse

Table 2.

 

 

Clinical Characteristics of RTB Patients

Table 2.

 

Clinical Characteristics of RTB Patients

	n	Sex (M:F)	Age (mo)	Family History	Associated Malformations	Deaths
Bilateral	15	1.14	Mean: 11.5	1 (brother)	2 microphthalmy	1
			Median: 7.5
			Range: 0–52
Unilateral	50	0.85	Mean: 21.5	1 (father)	—	2
			Median: 17.5
			Range: 1–109

Table 3.

 

 

Immunohistochemical Results Correlated with Choroid Infiltration and Optic Nerve Invasion

Table 3.

 

Immunohistochemical Results Correlated with Choroid Infiltration and Optic Nerve Invasion

Choroid (64*)
FAS	C0 (32)	C1 (22)	C2 (7)	C3 (2)	C4 (1)
0	9	3	—	—	—	12
+	16	11	2	—	—	29
++	6	6	4	—	—	16
+++	1	2	1	2	1	7

Optic Nerve (64*)
FAS	N0 (30)	N1 (23)	N2 (6)	N3 (4)	N4 (1)
0	6	5	1	—	—	12
+	17	10	1	1	—	29
++	6	5	4	1	—	16
+++	1	3	—	2	1	7

Figure 1.

 

Anti-FAS immunostaining (DAB chromogen-hematoxylin) (A) RTB graded 3+. The tumor growth was prevalently endophytic with partial retinal invasion through the inner nuclear layer (arrow) to the photoreceptor cell layer (arrowhead). Undifferentiated areas (left) showed a stronger positivity in comparison with the more differentiated rosettelike part of the tumor (right). Normal retina was negative (top). (B) Anterior chamber extension of a stage III RTB (bottom) graded 3+. Substantia propria of the cornea (center) and overlying corneal epithelium (top) were negative. (C) RTB scored 3+. Immunohistochemistry revealed the perivascular distribution of positive neoplastic cells. RTB cells close to necrotic areas (✶) became less positive or even negative. Normal retina was negative. (D) RTB scored 2+. Well-differentiated tumor showed positivity of Flexner-Wintersteiner rosettes (bottom). A few cells in the ganglion cell layer were positive (center). Normal retina was negative (top). Magnification: (A–C) ×20; (D) ×40.