Focal Adhesion Kinase-dependent Apoptosis of Melanoma Induced by Tyrosine and Phenylalanine Deficiency¹

Previously, we found that restriction of tyrosine (Tyr) and phenylalanine (Phe) inhibited growth and metastasis of the highly invasive and metastatic B16BL6 murine melanoma (1). We also demonstrated that restriction of these two amino acids in this cell line blocked the cell cycle in the G₀-G₁ phase (2). Cellular proliferation depends on the rates of both mitosis and apoptosis (programmed cell death), and tumor cells usually exhibit decreased cell death and increased cell proliferation (mitosis). The control of cell death is tightly linked to the cell cycle (3). We initiated this investigation to clarify whether apoptosis is initiated during the cell cycle arrest induced by Tyr and Phe restriction. Among the regulators of the cell cycle that also regulate apoptosis are the oncogenes, c-myc and ras (4, 5), the tumor suppressor gene, p53, and the cell progression gene products, cdc2, cyclin A, and E2F-1 (6, 7). We determined the effect of Tyr and Phe restriction on p53 and c-myc proteins because they are expressed in both human and murine melanoma cells (8, 9).

Melanoma cells require adhesion through integrin receptors for their survival and growth. Attachment to the extracellular matrix suppresses apoptosis in these cells (10). Clusters of integrins, occurring in focal adhesion contact sites, interact with the matrix during cellular attachment. Focal adhesion not only is important for attachment but also is essential to subsequent cell spreading and motility (11, 12). Integrin-matrix interactions regulate cell growth and apoptosis by initiating signal transduction pathways (13, 14, 15). FAK³ is a major signaling mediator, the activation of which requires both integrin attachment and cell spreading (15, 16, 17).

FAK is required for cell survival in adhesion-dependent cells, and autophosphorylation of a major tyrosine site is needed to perform this function (18, 19). Cell attachment induces FAK autophosphorylation on Tyr-397, and this allows the SH2 domains of the Src family and other kinases to bind. Src then phosphorylates at least five Tyr residues in FAK. Phosphorylation of some of these residues leads to the activation of the mitogen-activated protein kinase cascade (18, 20, 21, 22). Therefore, FAK is thought to have various functions, ranging from the regulation of focal adhesion turnover to the prevention of apoptosis (18, 23). Moreover, inhibition of FAK expression causes apoptosis in several human tumor cell lines including melanoma (24). Intracellular calcium is also involved in induction of apoptosis in some cancer cells (25, 26). Therefore, we examined intracellular calcium levels in melanoma cells during Tyr and Phe deprivation. We found that Tyr and Phe deprivation induce apoptosis in B16BL6 murine and A375 human melanoma cells. Deprivation of these amino acids does not alter p53 or c-myc protein expression or alter calcium influx into melanoma cells. Apoptosis is correlated with decreased expression and phosphorylation of FAK protein.

B16BL6 murine melanoma cells (27) and A375 human melanoma cells (obtained from American Type Tissue Collection, Rockville, MD) were routinely cultured in DMEM (Life Technologies, Inc., Grand Island, NY) supplemented with 10% heat-inactivated FBS (Hyclone, Salt Lake City, UT), sodium pyruvate, nonessential amino acids, 2-fold vitamin solution, l-glutamine, 60 IU/ml penicillin, and 100 IU/ml streptomycin (complete DMEM). The cells were used between in vitro passage 8 and passage 12. DMEM either contained 10% FBS or was FBS-free. The term, “complete” is used to refer to cells cultured in DMEM containing 10% FBS or MEM containing 10% dialyzed FBS.

Tyr- and Phe-limited medium was prepared from MEM Selectamine kits (Life Technologies, Inc.) as described previously (1). In this medium, Tyr and Phe were restricted to 22 and 24 μm, respectfully, and contained 10% dialyzed FBS. The amino acid levels reflect a 89% reduction in Tyr and a 88% reduction in Phe as compared with the levels contained in complete DMEM. The FBS was dialyzed to remove the amino acids in serum.

Tyr- and Phe-free MEM containing 10% dialyzed FBS was used in other experiments to magnify differences between experimental groups. In some experiments, Tyr-free MEM, Phe-free MEM, or glutamine-free MEM containing 10% dialyzed FBS was used as controls.

For experiments, melanoma cells were cultured in complete DMEM until they became 30–40% confluent. In some cultures, the medium was then replaced with the appropriate amino acid-deficient MEM. The cell viability, expression of FAK protein, Tyr-phosphorylated FAK, and expression of p53 and c-myc proteins were followed up to 4 days thereafter. In the experiments shown in Fig. 4, cells were cultured in Tyr- and Phe-free MEM for 3–4 days and then refed with complete DMEM. Tyr phosphorylation and expression of FAK were followed from 30 min to 6 h after refeeding.

To clarify the importance of substrate adherence, A375 melanoma cells were cultured in a suspended (unattached) condition. A375 melanoma cells, cultured previously in complete DMEM, were harvested with 0.25% trypsin and then resuspended in MEM or in Tyr- and Phe-free MEM containing 0.5% BSA. All cells were transferred to culture plates pretreated with BSA to block attachment sites. It was confirmed that the tumor cells did not adhere to the culture dishes under these conditions. The viability of the tumor cells was assessed by annexin V and propidium iodide staining as indicated below. The viability data are presented in Fig. 3 A(d).

Two nontransformed normal cell lines, HIFFs and NME cells were used as controls in some experiments. The effects of Tyr and Phe deprivation on these cells aided in determining the selectivity of the apoptosis induced by Tyr and Phe deprivation toward melanoma. The cells were cultured under the same conditions as the melanoma cells and were used below passage 5. Data on the effect of Tyr and Phe deprivation are presented in Fig. 3, A(c) and B(b). HIFFs were a gift from Drs. Mary Sanchez-Lanier and Gwendolyn Lenk (Department of Microbiology, Washington State University).

NME cells were isolated from newborn mice according to a procedure reported previously (28). Pregnant C57BL/6 mice were purchased from Charles River Laboratories (Wilmington, MA) and maintained in the Wegner Hall Vivarium, which is accredited by the American Association for Accreditation of Laboratory Animal Care. For NME cell isolation, newborn mice were killed and washed in 70% ethanol. The limbs were amputated, and the skin was removed. The skin was minced with scissors and soaked for 20 h at 4°C in PBS containing 0.25% trypsin. The trypsin solution then was removed, and the tissue was incubated with DMEM in trypsinizing flasks for 30 min at 37°C. The suspension was filtered through Nitex to remove large clumps. Cells were plated into Falcon T75 tissue culture flasks containing complete DMEM. Cultures were incubated in a humidified atmosphere at 5% CO₂ in air at 37°C. When cell monolayers reached confluence, they were harvested using 0.1 mm ethylene glycol-bis(β-aminoethyl ether)-N,N,N′N′-tetraacetic acid in PBS and stored in liquid nitrogen for future use.

In another experiment, two tyrosine kinase inhibitors, genistein, which inhibits Src, or tyrophostin A25, which inhibits EGF receptor tyrosine kinase (Ref. 20; Calbiochem, La Jolla, CA), were added into A375 cells cultured in Tyr- and Phe-free MEM for 3 days 1 h before refeeding with complete DMEM. Viability of the cells was evaluated 24 h after refeeding by the annexin V assay. The results for genistein are shown in Fig. 5.

The effect of Tyr and Phe limitation on the cell cycle distribution of A375 melanoma cells was determined by the same procedure that we used for analysis of B16BL6 murine melanoma (2). Briefly, cells were harvested with trypsin and stained with propidium iodide. The cell cycle distribution was determined using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). For each sample, 20,000 events were collected, and the DNA content was analyzed with the S Fit Model program.

Apoptosis was determined by four methods: DNA fragmentation assay, in situ TUNEL immunostaining, electron microscopy, and annexin V flow cytometric assay, as described below.

This assay was performed according to a published procedure (10). Briefly, cells were disrupted in a lysing buffer (10 mm Tris, 10 mm EDTA, 0.5% Triton X-100, and 100 μg/ml proteinase K, pH 7.4), and the DNA was extracted by sequential treatment with phenol and phenol/chloroform before being precipitated into 70% ethanol. Equal amounts of DNA from different treatment groups were electrophoresed through a 1.0% agarose gel and visualized by ethidium bromide staining. Appearance of low molecular weight DNA was used as an indication of DNA fragmentation and apoptosis.

The staining was performed with an in situ Apoptag Peroxidase kit (Oncor, Gaithersburg MD) according to a published method (29). Cells grown in chamber slides were fixed with 4% paraformaldehyde and then incubated with ethanol:acetic acid 2:1 for 5 min. Endogenous peroxide was quenched by incubation with 2% hydrogen peroxide. Then, the slides were incubated with terminal deoxyribonucleotidyl transferase enzyme working solution for the end labeling. After incubation with anti-digoxigenin-peroxidase solution, the slides were incubated with 0.02% diaminobenzidine solution containing hydrogen peroxide to develop the brown color reaction, and the slides were counterstained with methyl blue. A negative control section was used in each staining procedure by substituting the terminal deoxyribonucleotidyl transferase enzyme with water. The control cells exhibited no brown color, indicating that they were viable and not undergoing apoptosis.

The method for electron microscopic examination of apoptosis followed a published procedure (30). Melanoma cells were fixed with 2.5% glutaraldehyde and then incubated with 1% osmium tetroxide. Cells were embedded in epoxydic resin, and the ultrathin sections were stained with uranyl acetate and lead citrate. The sections were then examined with an electron microscope.

Perturbations in the cellular membrane occur during the early stages of apoptosis and lead to a redistribution of phosphatidylserine to the external side of the cell membrane. Annexin V selectively binds to phosphatidylserine, and this has enabled the use of fluorescein-labeled annexin V for identification of cells undergoing apoptosis. Cells are also stained with propidium iodide to distinguish early apoptotic cells from necrotic cells. FITC-labeled annexin V apoptosis detection kit (Trevigen, Gaithersburg, MD) was used as demonstrated by another laboratory in a solid tumor cell line and in fibroblasts (31). Ten thousand events were collected with a FACScan flow cytometer (Becton Dickinson, San Jose, CA). The percentage of live, dead, and apoptotic cells was determined as discussed in “Results” and as shown in Fig. 2. The percentage of cells that are dead is depicted in the upper left quadrant (cells single stained with propidium iodide). Cells that are in the early stages of apoptosis are depicted in the lower right quadrant (cells single stained with annexin V conjugated to FITC). Cells that are in the later stages of apoptosis are depicted in the upper right quadrant (cells doubly stained with propidium iodide and annexin V-FITC). Cells in the lower left quadrant reflect the percentage of fully viable cells.

Cells were harvested with a cell scraper and lysed in ice-cold buffer containing 50 mm Tris (pH 7.4), 50 mm NaCl, 0.5% NP40, 1 mm phenylmethylsulfonyl fluoride, 0.1 mg/ml trypsin inhibitor, 0.1 mg/ml NaF, 0.1 mg/ml β-glycerophosphate, 0.1 mg/ml Na₃VO₄, 2 μg/ml leupeptin, and 25 μg/ml aprotinin for 30 min (NP40 lysis buffer). The lysate was centrifuged, and the protein concentration of the supernatant was determined by the Bradford procedure. This method was used in all successive quantifications of protein concentration.

Immunoreactive FAK, p53, and c-myc proteins were determined by chemiluminescent immunoblotting as described previously (2). Monoclonal antibodies against FAK and phosphotyrosine proteins were obtained from Transduction Laboratories (Lexington, KY) and Upstate Biotechnology, Inc. (Lake Placid, NY). The monoclonal antibody for c-myc protein was obtained from Cambridge Research Biochemicals (Wilmington, DE). The monoclonal antibody against p53 protein was obtained from NeoMarkers, Inc. (Fremont, CA). Blots were also incubated with an anti-actin mouse monoclonal antibody (Oncogene Science, Cambridge, MA) to assess relative protein loading (2). Densitometry (Master Scan System, Scanalytics, Billerica, MA) was used to verify differences between the intensity of the bands. The ratio of the integrated absorbance of the band of interest to that of the actin band was used as an index of expression. The relative amount of protein in cells cultured in Tyr- and Phe-free MEM was expressed as a percentage of control cells.

For determination of Tyr phosphorylation of FAK, nonradiolabeled immunoprecipitation was performed by using anti-FAK antibody, followed by Western blotting with anti-phosphotyrosine-protein monoclonal antibody. All steps in the immunoprecipitation procedure were carried out in the cold room. Cells were washed twice with cold PBS, scraped into NP40 lysis buffer, and then lysed by rotation for 15–30 min. Lysates were centrifuged, and the protein concentration was determined. Lysates were precleared by incubating with either preimmune serum or normal rabbit or mouse serum, and immunoglobulin G sorb (Pharmacia Piscataway, NJ) for 20 to 30 min, followed by a 10-min centrifugation at 15,000 × g. The lysates were incubated with antibody-coated protein A-Sepharose 4B or protein G-Sepharose 4B (Pharmacia) for 6 to 8 h. The beads were washed with lysis buffer three to four times, resuspended in Lammeli buffer, and then separated on an SDS-polyacrylamide gel. After determination of phosphorylated FAK, the blot was stripped with buffer containing 2% SDS, 2.5 mm Tris (pH 6.8), and 100 mm β-mercaptoethanol for 30 min at 55°C. The blot was washed several times with PBS and reprobed with anti-FAK antibody. The amount of protein in the blot was determined by densitometry. The ratio of the integrated absorbance of the total immunoreactive FAK to that of the Tyr-phosphorylated FAK band was used as an index of Tyr phosphorylation.

The measurement of free calcium in melanoma cells was performed following published methods (32). One h before cell harvesting, 1.5 μg/ml of Fluo-3 (Molecular Probes, Inc., Eugene, OR), a membrane permeable long-wavelength calcium indicator, was added into the medium. Cells were trypsinized, washed with calcium-free PBS, and then analyzed by flow cytometry. The intensities of Fluo-3 fluorescence represent the relative concentration of free calcium. Ten thousand events were analyzed for each sample.

The cell survival curves were compared between two groups at the same time point by using Student’s t test within ANOVA (33).

We demonstrated previously that restriction of Tyr and Phe arrested B16BL6 melanoma cells in the G₀-G₁ phase of cell cycle in vitro and in vivo (2). The results in Fig. 1 A show that Tyr and Phe restriction in vitro similarly arrest A375 human melanoma cells in the G₀-G₁ phase. Refeeding the cells with fresh DMEM significantly increased the progression of arrested cells from the G₁ through S phase within 16–24 h (data not shown). This is consistent with the results in B16BL6 murine melanoma cells (2).

By using different methods, we demonstrated that melanoma cells are especially sensitive to induction of apoptosis in response to Tyr and Phe deprivation. This effect is specific for the amino acid deprivation and is not induced by serum starvation. Restriction of these two amino acids does not induce apoptosis in NME cells and HIFFs. Additionally, deprivation of Tyr and Phe also decreases viability of melanoma cells in suspension.

Fig. 1 B shows the DNA fragmentation induced by Tyr and Phe deprivation in A375 melanoma cells. DNA fragmentation is evident 1 day after culturing cells in Tyr- and Phe-free MEM. Degradation induced by Tyr and Phe does not result in a typical DNA ladder formation in A375 melanoma cells, and DNA is degraded with a predominant band at 500 bp. A similar tendency was observed in B16BL6 murine melanoma cells (not shown). Other investigators also have shown that the typical DNA ladder is not always associated with apoptosis (10, 34). DNA degradation is now thought to be a sign of late-stage apoptosis (degradation stage) rather than an essential component of the death process at the effector stage (35). Therefore, we also performed different experiments to confirm apoptosis in melanoma cells as described below.

Histological evidence of apoptosis in melanoma cells induced by Tyr and Phe deprivation was demonstrated by TUNEL staining and electron microscopy (not shown). The annexin V assay provides the advantage of quantifying the rate of induction and progression of apoptotic cells. We used this method to follow the kinetic changes in melanoma cells under Tyr and Phe deprivation and compared them to changes in the control cell lines under the same conditions. Representative sets of data are shown in Fig. 2,A (A375 melanoma) and Fig. 2 B (HIFF).

The data in Fig. 2,A(a) indicate the background level of annexin V and propidium iodide staining for untreated A375 melanoma cells. By the staining criteria, ∼95% of the cells are viable, 2% are dead, 2% are at an intermediate stage of apoptosis, and 0.8% represent early apoptotic cells. Thus, we conclude that the A375 melanoma cells normally undergo some basal level of apoptosis or that some minimal degree of apoptosis and cell death result from the isolation procedure. Apoptosis is clearly present in A375 melanoma cells cultured for 24 h in Tyr- and Phe-free MEM [Fig. 2,A(b)]. The number of cells undergoing late-stage apoptosis is dramatically increased after 48 h [Fig. 2,A(c)]. Tyr and Phe deprivation increase the early-state apoptosis by 4.5–5-fold (lower right quadrant) [Fig. 2, A(b) and A(c)]. The percentage of cells undergoing the late stages of apoptosis (upper right quadrant) increases by 2-fold at 24 h and by >10-fold at 48 h in response to Tyr and Phe deprivation. The percentage of cells staining with propidium iodide did not increase at 24 h after deprivation of Tyr and Phe, but by 48 h was increased by 4-fold. These data show that apoptosis is a dynamic process that is initiated within 24 h after converting A375 melanoma cells to a Tyr- and Phe-free environment. A similar pattern of apoptosis induced by Tyr and Phe deprivation was observed with B16BL6 melanoma cells (not shown). Fig. 2 B shows that non-transformed HIFF are not susceptible to apoptosis induced by Tyr and Phe deprivation. Similar results were observed in NME cells (not shown).

The kinetic changes in viability of A375 melanoma cells under Tyr and Phe deprivation in comparison to HIFF are shown in Fig. 3,A. B16BL6 murine melanoma cells and NME cells (Fig. 3 B) show similar changes in response to Tyr and Phe limitation as compared with A375 melanoma cells and HIFFs. The results also indicate that HIFFs and NME cells are resistant to the effects of Tyr and Phe deprivation. In contrast, A375 melanoma and B16BL6 melanoma cells begin a gradual decline in viability within 1 day after culture of the cells in the amino acid-restricted MEM.

The effect of Tyr and Phe deprivation on A375 melanoma cells cultured in a suspended condition is shown in Fig. 3 A(d). Melanoma cells were ≈80% viable for 36 h when cultured in suspension in complete DMEM; however, viability dropped to ∼45% at 48 h. Viability progressively decreased within 24 h after culture in Tyr- and Phe-free MEM. Less than 10% were viable at 48 h.

Fig. 4,A shows that the total amount of immunoreactive FAK is decreased to ∼25% of control in A375 melanoma cells after 4 days of Tyr and Phe deprivation. FAK expression fully recovers within 60 min after refeeding with complete DMEM. Fig. 4 B shows that the Tyr phosphorylation of FAK in A375 melanoma cells is reduced to nondetectable levels within 3 days after Tyr and Phe deprivation. Serum starvation did not affect FAK expression or Tyr phosphorylation of FAK. Tyr phosphorylation of FAK recovered rapidly. FAK phosphorylation was enhanced from 30- to 51-fold within 3 h after refeeding cells with complete DMEM. Phosphorylation of FAK began to return to normal baseline levels at 6 h after refeeding. B16BL6 melanoma cells responded similarly to A375 melanoma under all conditions of this experiment (data not shown). Tyr and Phe deprivation did not inhibit the expression and Tyr phosphorylation of FAK in HIFFs and NME cells (data not shown).

The data in Fig. 5 confirm that genistein at a dose of 50 μm, which does not induce apoptosis (36), did not affect the viability of A375 melanoma cells cultured in complete DMEM. Genistein, a Tyr kinase inhibitor of Src that inhibits Tyr phosphorylation of FAK (20), also did not enhance cell death due to deprivation of Tyr and Phe for 3 days (also shown in Fig. 5). However, genistein blocked the recovery of cell survival in Tyr- and Phe-deprived cells after refeeding with complete DMEM (Fig. 5). DMSO (1.7 μl/ml), the solvent of genistein, did not affect cell survival. The experiment was repeated with another tyrosine kinase inhibitor, tyrophostin A25, which inhibits EGF receptor tyrosine kinase (20). This inhibitor did not alter the viability of A375 melanoma cells and did not block the recovery of cell survival in Tyr- and Phe-deprived cells after refeeding with complete DMEM (data not shown).

The single effects of deprivation of Tyr, Phe, and glutamine on viability of A375 melanoma cells are shown in Fig. 6. Deprivation of glutamine did not induce apoptosis, and single deprivation of Tyr and Phe only slightly induced apoptosis. This is in contrast to the magnitude of apoptosis induced by combined Tyr and Phe deprivation.

The expression of p53 and c-myc proteins in A375 and B16BL6 melanoma cells was demonstrated by Western blot analysis in the lysates from cells cultured in either complete DMEM or Tyr- and Phe-free MEM. No significant differences in the intensities of the bands were observed after normalization of the bands to actin by densitometry (not shown).

In the experiments to measure intracellular free calcium, we used hydrogen peroxide treatment as a positive control group because it is a nonselective inducer of apoptosis that results in elevation of intracellular free calcium (26). Fig. 7,A shows that apoptosis in A375 cells induced by hydrogen peroxide is associated with a significant elevation of intracellular free calcium. No elevation was observed in cells under Tyr and Phe deprivation (Fig. 7, B and C). Apoptosis induced by hydrogen peroxide was confirmed by annexin V staining. The same results were obtained in B16BL6 melanoma cells (data not shown).

Induction of apoptosis is a viable and important new therapeutic approach to control cancer (37). Previously, we showed that restriction of Tyr and Phe blocked the cell cycle in the G₀-G₁ phase in B16BL6 melanoma cells (2). Here we also showed a similar cell cycle arrest in human A375 melanoma cells. Control of apoptosis is tightly linked to the cell cycle (3), and amino acids control cell cycle progression through regulation of gene expression (2, 38, 39). Thus, we initiated this investigation to test the hypothesis that Tyr and Phe restriction induces apoptosis in melanoma cells and that it is relatively specific. We found that the restriction of Tyr and Phe arrests murine B16BL6 and human A375 melanomas in the G₀-G₁ phase of the cell cycle. Moreover, Tyr and Phe restriction also induce apoptosis in these cells. Although the aggressive metastatic phenotype of B16BL6 melanoma is linked to its resistance against apoptosis (34), induction of apoptosis in these cells in vitro may partially explain how the restriction of Tyr and Phe inhibits metastasis of melanoma and other tumors in mice (40, 41).

The results of this study showed that melanoma cells are especially sensitive to induction of apoptosis in response to combined Tyr and Phe deprivation but not to serum or glutamine deprivation, although serum or glutamine deprivation also inhibit melanoma growth (1, 2). Moreover, the growth inhibition induced by serum or glutamine deprivation is not reflected in inhibition of melanoma metastasis (1). Thus, the effects of Tyr and Phe deprivation on melanoma cells are not due to a general starvation of the cells for growth factors because deprivation of these amino acids induces apoptosis in the presence of serum (10% FBS). Melanoma cells are especially sensitive to the availability of Tyr and Phe, and these two amino acids may be regarded as survival factors for this tumor. Deprivation of Tyr and Phe does not induce apoptosis in HIFFs and NME cells. Thus, Tyr and Phe restriction would be expected to exhibit little host toxicity. This has been confirmed in our previous in vivo studies in mice and humans (40, 41, 42).

Intracellular calcium is involved in apoptosis. Thapsigargin and ionomycin induce calcium-dependent apoptosis in some cancer cells (25, 43). However, the present study shows that apoptosis of melanoma in response to Tyr and Phe deprivation is not associated with a significant elevation of intracellular free calcium. This suggests that apoptosis induced by deprivation of these amino acids is independent of capacitative entry of extracellular calcium.

Although c-myc and p53 proteins are involved in the induction of apoptosis (5), their expression in melanoma cells is not altered by Tyr and Phe deprivation. Therefore, it is unlikely that deprivation of these amino acids plays an important role in the induction of apoptosis in these cells.

It is known that melanoma cells require adhesion through integrin receptors for their survival. Growth and attachment to the extracellular matrix suppresses apoptosis in these cells (10). Integrin-matrix interactions regulate cell growth and apoptosis by initiating signal transduction (13, 14, 15). FAK is a signaling protein induced during integrin-matrix interactions, and autophosphorylation of this protein is required for cell survival in adhesion-dependent cells (18, 19).

In this study, we show that FAK expression and Tyr phosphorylation of FAK were reduced during apoptosis induced by Tyr and Phe deprivation. The decrease in FAK protein may also contribute to the lower Tyr phosphorylation of FAK during Tyr and Phe deprivation. The Tyr phosphorylation of FAK is rapidly elevated after refeeding the cells with complete DMEM. FAK expression and cell viability recover sequentially. This suggests that the apoptosis induced by Tyr and Phe deprivation is FAK dependent. Other researchers found that FAK is overexpressed in human invasive and metastatic tumors. Using different approaches to inhibit either the expression or the function of this signaling molecule, they showed that inhibition of FAK causes apoptosis of tumor cells including melanoma (23, 24).

Genistein, a Src Tyr kinase inhibitor (20), when added to culture medium at a dose that does not induce apoptosis in melanoma (36), did not affect viability of A375 melanoma cultured in DMEM with 10% FBS. However, it blocked the recovery of cell survival in Tyr- and Phe-deprived cells after refeeding with complete DMEM. By contrast, another EGF receptor tyrosine kinase inhibitor, tyrophostin A25 (20), did not affect the viability of A375 cells. It did not block the recovery of cell survival in Tyr- and Phe-deprived cells after refeeding with DMEM. Moreover, apoptosis of human melanoma cells is induced by treatment with antisense oligonucleotides to FAK (24).⁴ Collectively, these data provide further evidence that Tyr and Phe deprivation mediate apoptosis in melanoma cells through a FAK-dependent pathway. Although the present study reveals a connection between amino acid deprivation, induction of apoptosis, and inhibition of FAK, the exact molecular mechanism(s) is unclear. Ongoing research in our laboratory is further defining the specific mechanism(s) associated with this finding.

For solid tumors that have invaded into the blood stream either directly or through lymphatics, survival in the circulation is a critical step to the formation of distal metastases. Because melanoma cells are adhesion dependent, it is possible that these cells would be more sensitive to the effects of Tyr and Phe deprivation while unattached. The present study shows that Tyr and Phe deprivation in vitro reduce the survival of melanoma cells in suspension [Fig. 3 A(d)]. If these data can be extrapolated to an in vivo situation, it is reasonable to speculate that melanoma cells undergoing hematogenous metastasis are especially sensitive to Tyr and Phe restriction. This may explain, at least in part, why spontaneous metastasis is inhibited in rodents fed diets restricted in Tyr and Phe (40).

Healthy individuals can be maintained in an outpatient setting for at least 4 weeks on a formula diet restricted in Tyr and Phe (42). Dietary restriction of Tyr and Phe enhances the experimental chemotherapy of l-dopa methylester against B16 melanoma in vivo and may even prevent the emergence of drug resistance during chemotherapy (44). Recently, Yoshioka et al. (45) found that multiple doses of a recombinant methioninase enzyme were effective against human tumors in nude mice without adverse side effects, including the lack of acute immune hypersensitivity. Methionine depletion due to this enzyme potentiated the antitumor efficacy of 5-fluorouracil. Thus, the use of amino acid depletion to enhance chemotherapy and retard metastasis is a novel approach to the control of cancer.

In summary, the present study reveals that Tyr and Phe deprivation preferentially induce apoptosis in melanoma cells but not in normal cells and that apoptosis is FAK dependent. Because normal cells are resistant to apoptosis, and because melanoma cells are induced to undergo apoptosis in response to Tyr and Phe restriction, it should be possible to exploit this difference therapeutically with minimal host toxicity.