L-type amino acid transporter 1 as a potential molecular target in human astrocytic tumors

Authors


Abstract

L-type amino acid transporter 1 (LAT1) is a Na+-independent neutral amino acid transport agency and essential for the transport of large neutral amino acids. LAT1 has been identified as a light chain of the CD98 heterodimer from C6 glioma cells. LAT1 also corresponds to TA1, an oncofetal antigen that is expressed primarily in fetal tissues and cancer cells. We have investigated for the first time, the expression of the transporter in the human primary astrocytic tumor tissue from 60 patients. LAT1 is unique because it requires an additional single membrane spanning protein, the heavy chain of 4F2 cell surface antigen (4F2hc), for its functional expression. 4F2hc expression was also determined by immunohistochemistry. Kaplan-Meier analyses demonstrated that high LAT1 expression correlated with poor survival for the study group as a whole (p < 0.0001) and for those with glioblastoma multiforme in particular (p = 0.0001). Cox regression analyses demonstrated that LAT1 expression was one of significant predictors of outcome, independent of all other variables. On the basis of these findings, we also investigated the effect of the specific inhibitor to LAT1, 2-aminobicyclo-2 (2,2,1)-heptane-2-carboxylic acid (BCH), on the survival of C6 glioma cells in vitro and in vivo using a rat C6 glioma model. BCH inhibited the growth of C6 glioma cells in vitro and in vivo in a dose-dependent manner. Kaplan-Meier survival data of rats treated with BCH were significant. These findings suggest that LAT1 could be one of the molecular targets in glioma therapy. © 2006 Wiley-Liss, Inc.

Glioma, especially glioblastoma, is the most common intrinsic brain tumor and with the worst prognosis in human malignancy. The cause of death of glioma patients is not a distant metastasis but the failure of local control of the tumor. The invasive nature of glioma often prevents total surgical excision. Adjuvant therapy for the residual tumor is therefore essential, however, conventional chemotherapy, immunotherapy, and radiotherapy has been proved to be of limited value.1 Gene therapy for malignant glioma has been expected to be useful; however, no definitive clinical usefulness has been proved. Recently, several candidates for the molecular target of malignant tumors are advocated. One of them is the angiogenic and vascular proliferating factor.2 The highly proliferating malignant tumor cells may need much substrate such as sugars and amino acids. If a specific upregulation of amino acid transport system in malignant tumor cells do exist, it could be a molecular target for therapy.

L-type amino acid transporter 1 (LAT1) is a Na+-independent neutral amino acid transport agency and essential for the transport of large neutral amino acids through the plasma membrane.3, 4 LAT1 exhibits high affinity for several nutritionally essential amino acids such as leucine, isoleucine, valine, phenylalanine, tryptophan and methionine. The molecular nature of LAT1 was not characterized until 1998, when using expression cloning, a cDNA encoding a transporter subserving the LAT1 from a C6 rat glioma cell cDNA library was isolated by Kanai et al.3 Since LAT1 is highly regulated in nature and upregulated upon the isolation and cultivation of cells, it is essential to examine the in vivo expression of LAT1 in brain tumor tissues.5, 6 The specific antibody to human LAT1, which recognizes both rodent and human LAT1, was generated. Using this antibody, we have investigated for the first time, the expression of the transporter in the human astrocytic tumor tissue. LAT1 is unique because it requires an additional single membrane spanning protein, the heavy chain of 4F2 cell surface antigen (4F2hc), for its functional expression.3 When coexpressed with 4F2hc, LAT1 transports neutral amino acids with branched or aromatic side chains and does not accept basic amino acids or acidic amino acids.3 4F2hc expression was also determined by immunohistochemical staining with a polyclonal rabbit anti-human 4F2hc antibody. All astrocytic tumor cells clearly exhibited positive staining for LAT1 in variable degrees; however, we found strong expression of LAT1 in high-grade astrocytomas. On the basis of these findings, we also investigated the effect of the specific inhibitor to LAT1, 2-aminobicyclo-2(2,2,1)-heptane-2-carboxylic acid (BCH), on the growth of C6 glioma cells in vitro and in vivo using a rat C6 glioma model.7

Material and methods

Patients and tissues

All patients had primary astrocytic tumors of the brain. Patients were treated surgically for the first time between 1990 and 1999 in our hospital. Clinical data were obtained by retrospective chart review. Survival was determined from the date of initial surgery. Follow-up was available for all patients. Informed consent was obtained in all cases. Median follow-up time from resection of initial disease was 40.9 months.

Tumor specimens were obtained by surgical resection in all cases. Formalin-fixed, paraffin-embedded sections were stained with hematoxylin and eosin, and a histological and cytological diagnosis was made.

Histological diagnosis and tumor grading were performed according to the grading system established by WHO.8 Fifteen specimens were diagnosed as Grade 2, 17 as Grade 3 and 28 as Grade 4 (glioblastoma multiforme (GBM)).

Immunohistochemistry for human LAT1 and 4F2hc in the glioma

LAT1 expression was determined by immunohistochemical staining with an affinity-purified polyclonal rabbit anti-human LAT1 antibody. Oligopeptides corresponding to amino acid residues 497–507 of human LAT1 (CQKLMQVVPQET) and amino acid residues 516–529 of human heavy chain of 4F2 cell surface antigen (4F2hc) (EPHEGLLLRFPYAAC) were synthesized. The N-terminal cystein residues were introduced for conjugation with keyhole limpet hemocyamine. Antipeptide antibodies were produced as described elsewhere.9 For immunohistochemical analysis, antisera were affinity-purified as described previously.9

Immunohistochemical staining was performed on paraffin sections using an avidin–biotinyl peroxidase complex method. Briefly, deparaffinized, rehydrated sections were treated with 0.6% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase activity. After rinsing in 0.05 M tris-buffered saline containing 0.1% tween-20, the sections were incubated with anti-LAT1antiserum (1:250) or anti-4F2hc antiserum (1:500) overnight at 4°C. Thereafter, they were incubated with Envision (+) rabbit peroxidase (DAKO, Carpinteria, CA) for 30 min. The peroxidase reaction was performed using 0.02% 3,3′-diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide in 0.05 M tris-HCl buffer, pH 7.4. Finally, nuclear counterstaining was performed with Mayer's hematoxylin. To verify the specificity of immunoreactions by absorption experiments, the tissue sections were treated with primary antibodies in the presence of antigen peptides (200 μg/ml).

Analysis of LAT1 and 4F2hc staining

Immunoreactivity was graded from – to +++ according to the percent of positive cells and the intensity of staining. The percentage of cells expressing LAT1 and 4F2hc was estimated by dividing the number of positively stained astrocytic tumor cells by the total number of tumor cells per high-power field. The cases in which all cells or more than 75% of the cells stained positively were considered diffuse staining; those in which less than 75% of the cells stained were considered patchy staining. More than 1,000 tumor cells were counted to determine the percentage of positive cells. The intensity of staining was determined and recorded as negative, weak, or strong. The cases in which weak patchy or not stained were considered (−); diffuse weak staining were considered (+); strong patchy staining were considered (++); and those strong diffuse staining were considered (+++). According to this grading protocol, 2 observers (HN and NO) of the authors, without prior knowledge of the clinical data, independently graded the staining intensity in all cases. To test the intraobserver variability, all sections were reassessed by one author (HN) after all first measurements had been completed. The time between the first and second assessments was at least 4 weeks. The interobserver variability was determined by comparing the values of the first measurements of 2 authors (HN and NO).

Proliferation rates determined by proliferating cell nuclear antigen immunostaining

The same tumor specimens were analyzed by immnohistochemistry with an anti-proliferating cell nuclear antigen (PCNA) monoclonal antibody (Novocastra Laboratories Ltd., Newcastle upon Tyne, UK). Paraffin-embedded tissue sections (3-μm thick) were employed for immunohistochemistry. After deparaffinization in xylene and blocking of endogenous peroxidase activity with 0.3% hydrogen peroxide in absolute methanol for 30 min at room temperature, the sections were incubated for 1 hr with anti-PCNA monoclonal antibody. The sections were washed with PBS, and treated with biotynyl anti-rabbit immunoglobulin for 10 min, then washed with PBS again and treated with peroxydase-labeled streptavidin for 5 min, and incubated in 3,3′-diaminobenzidine (DAB) solution, and then counterstained with methylgreen. Control study: (i) normal brain slices (negative control) (ii) adenocarcinoma (positive control) (iii) normal rabbit serum was used instead of the specific antibodies. The percentages of PCNA-positive cells were determined by counting 1,000–1,500 cells in at least 2 microphotographs of each section.

Statistical analysis

Survival was analyzed using the Kaplan-Meier method, and prognostic factors were assessed by log-rank analysis. Univariate and multivariate analyses were made of disease-specific survival (based on the number of patients who did not die from glioma). LAT1 and 4F2hc staining score, and other putative prognostic factors (age, gender, tumor histology, PCNA staining index) were used to stratify patients. A stepwise multivariate Cox regression analysis was also performed to further test the independence of LAT1 expression from other parameters. The distribution of the LAT1 score in relation to tumor and patient characteristics was investigated using the χ2-test. Correlations between variables were obtained using Spearman's rank correlation. All tests were two-sided, and p < 0.05 was considered significant.

Cell line and culture condition

C6 rat glioma cell line was purchased from Dainippon Pharmaceutical Company (Osaka, Japan). Cells were maintained in a CO2 incubator at 37°C by in vitro passage at 3–4-day intervals in Ham's F10 medium with 2 mM L-glutamine (GIBCO BRL, Grand Island, NY) supplemented with 15% horse serum (GIBCO BRL) and 2.5% fetal bovine serum (HyClone®). After the cells reached subconfluence, single-cell suspension was obtained by trypsinization and the numbers of cells were counted with a particle counter (Model PC-607, Erma, Tokyo, Japan). Then cell suspensions of desired concentrations were prepared and used for the following experiments.

Evaluation of cell survival in vitro (colorimetric MTT assay)

Reagents required for the colorimetric MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrasolium bromide) assay were purchased from Chemicon International (Temecula, CA). Cells at a desired concentration were plated in 96-well flat-bottomed tissue culture trays (Falcon) and placed gently until they completely adhered to the bottom. Then the culture medium was replaced by a new one containing a desired concentration of BCH, and its effect on the cell growth was evaluated. The vehicle was culture medium used to introduce the BCH in cell cultures. At the end of the assay, 10 μl of MTT solution was added to each well and cells were incubated at 37°C for 4 hr. Then 100 μl of isopropanol/HCl solution was added to each well to dissolve the MTT formazan. Within an hour, the absorbance at a test wavelength of 595 nm and a reference wavelength of 655 nm (OD595–655) was measured on an ELISA plate reader (Bio-Rad, Model 3550 microplate reader).

A rat glioma model

Adult male Wistar rats weighing 200 g were used for this study. Animals were maintained and experiments conducted according to guidelines established by the Institutional Animal Care and Use Committee of National Defense Medical College.

The rats were anesthetized with isoflurane in 30% oxygen and 70% nitrous oxide gas mixture through a facemask. The rats were fixed in a stereotactic head holder in a flat-skull position. A burr hole was made using a 1.4-mm diamond-tipped burr at the following coordinates: 1.5 mm anterior to bregma and 3.0 mm lateral of midline. A needle was inserted into the right caudate nucleus, depth 5.0 mm from the top of skull. C6 cells were prepared fresh from culture to ensure optimal viability of cells during tumor inoculation. Each rat was injected with 1.0 × 106 C6 cells in 10 μl phosphate-buffered saline-glucose medium. After injection, injector remained for 5 min to allow the injected cell suspension to come to equilibrium inside the brain.

For continuous infusion of the tumor inoculation site, each rat was implanted with an osmotic minipump-brain infusion assembly 1 day or 8 days after tumor inoculation. The minipump (average infusion rate, 5.0 l/hr; Alzet model 2ML2; Alza, Palo Alto, CA) was filled with either vehicle alone (saline, control), 50 mM D-mannitol, or a desired concentration (50 mM or 230 mM) of BCH. The brain infusion assembly consisted of a catheter tubing and a stainless steel cannula with two-depth adjustment spacers to obtain stereotaxically correct depth. The tubing-cannula assembly was also filled with the appropriate solution and joined to the pump. One day or 8 days after tumor inoculation, the cannula was inserted at the inoculation site and secured in position with dental cement. The osmotic pump was housed in a subcutaneous pocket in the midscapular area of the back of the rat for 28 days. Animals recovered from anesthesia and resumed their previous activity in cages. The animals were housed individually to prevent dislodging of the brain infusion assembly.

Fifty-one rats were divided into 6 experimental groups. Group 1 (n = 9) were given 230 mM of BCH 1 day after tumor inoculation. Group 2 (n = 9) were given the saline 1 day after tumor inoculation. In Group 3 (n = 7), animals were given 230 mM of BCH 8 days after tumor inoculation. Group 4 (n = 12) received 50 mM of BCH 8 days after tumor inoculation. This dose was selected because cell survival was disturbed in vitro when BCH was added at a concentration of more than 25 mM. Group 5 (n = 7) were given the saline 8 days after tumor inoculation. Group 6 (n = 7) were given 50 mM D-mannitol 8 days after tumor inoculation.

After tumor inoculation, the body weight of rats was measured and recorded every day. All survivors were sacrificed 22 days after implantation. They were deeply anesthetized with intraperitoneal pentobarbital (100 mg/kg). Animals were perfused transcardially with normal saline followed by 4% buffered paraformaldehyde. The brains were removed and embedded in paraffin after fixation in 4% buffered paraformaldehyde followed by 0.1 mmol/l PBS (pH 7.4) for 24 hr at 4°C. Serial coronal sections (5-μm-thick) were prepared. The serial sections were mounted onto silanated slides and were used for histology and histochemistry. The sections were stained with hematoxylin and eosin to confirm the tumor. The same tumor specimens were analyzed by immnohistochemistry with an anti-PCNA monoclonal antibody as described above. The extent of the tumor at 20 predetermined levels was measured with the computer-assisted image analyzing system (NIH image 1.57). The tumor volume was calculated by taking the sum of the tumor areas of the different brain slices times the thickness of the slices. Tumor volumes and the percentages of PCNA-positive cells in each animals were analyzed by analysis of variance (ANOVA) followed by the Bonferroni/Dunn test to ascertain significance between groups. The body weight of each animal was analyzed by repeated-measures ANOVA. Statistical significance was set at p < 0.05.

Results

Sixty specimens were obtained from 60 patients with primary astrocytic gliomas. Twenty-nine patients were male and 31 were female. They ranged from 11 to 88 yr in age (mean, 45.1 years ± 19.3). All patients had Karnofsky performance scores of at least 70 at diagnosis. Patients were treated with surgery and adjuvant chemoradiation therapy (consisting of ACNU, interferon β, local external beam radiation) for all patients with GBM or anaplastic astrocytoma and adjuvant radiation therapy for most patients with low-grade astrocytoma. There were 28 GBMs, 17 anaplastic astrocytomas and 15 low-grade fibrillary astrocytomas.

Qualitative immunohistochemical analysis for LAT1 and 4F2hc

LAT1 and 4F2hc immunoreactivity was observed in all the tumor specimens examined. LAT1 immunostaining was observed predominantly on the plasma membrane and astrocytic process (Figs. 1c and 1f). In cases of strong diffuse LAT1 staining in tumor cells, intense cytoplasmic staining was also evident (Figs. 1g and 1h). In sections containing areas of normal cortex adjacent to the tumor, infiltrating tumor cells showed more intense LAT1 staining (Fig. 2b). Examples of LAT1 and 4F2hc immunostaining of the same specimen are shown in Figures 1d and 1e. In the absorption experiments in which tissue sections were treated with primary antibodies in the presence of antigen peptides, the immunostaining was drastically decreased, confirming the specificity of the immunoreaction (Fig. 1i). The nuclear staining for LAT1 might be artifactual; however, the nuclear and perinuclear staining was also decreased in the absorption experiments. LAT1 and 4F2hc immunostaining on normal uninvolved brain are shown in Figures 2a and 2c. No significant immunostaining in neurons or astrocytes was seen in the uninvolved normal brain. The intraobserver reproducibility of scoring was high (correlation coefficient, 0.93; p < 0.0001; coefficient of variance, 29.7%). The interobserver reproducibility of scoring was also high (correlation coefficient, 0.9; p < 0.001; coefficient of variance, 31.1%).

Figure 1.

(a) Glioblastoma with diffuse weak (+) immunoreactivity for LAT1. (b) Low grade glioma with diffuse weak (+) immunoreactivity for LAT1. (c) Anaplastic astrocytoma with strong patchy (++) staining for LAT1. (d) Anaplastic astrocytoma with strong patchy (++) staining for 4F2hc. (e) Anaplastic astrocytoma (the same specimen as shown in d) with diffuse strong (+++) staining for LAT1. (f) Anaplastic astrocytoma with intense immunoreactivity for LAT1 observed predominantly on the plasma membrane (arrow heads). Inset at higher magnification. (g) Glioblastoma with strong diffuse (+++) cytoplasmic staining (arrows) for LAT1. Inset at higher magnification. (h) Glioblastoma with strong diffuse (+++) cytoplasmic staining for LAT1. (i) In the absorption experiments, the LAT1 immunostainings were diminished, confirming the specificity of the immunoreaction (the same specimen as shown in H). Immunoreactions were visualized with diaminobenzidine and nuclear counterstaining was performed with Mayer's hematoxylin. All bars = 20 μm.

Figure 2.

(a) Control LAT1 staining on normal uninvolved brain. (b) Glioblastoma cells of the same patient in the border zone with strong staining for LAT1 showing perineuronal satellitosis (arrow heads). (c) Control 4F2hc staining on normal uninvolved brain in the same patient. (d) Glioblastoma cells of the same patient in the border zone with weak staining for 4F2hc showing perineuronal satellitosis (arrows). All bars = 20 μm.

Correlation of LAT1 and 4F2hc staining with clinical and histopathological features

LAT1 immunostaining did not correlate with patient age or gender (Table I). However, the intensity of LAT1 staining was greater in GBMs than in low-grade astrocytomas (Table I). The grade of LAT1 staining increased with glioma grade, and this finding was statistically significant. The grade of LAT1 staining correlated statistically with PCNA index (p = 0.0075) and with 4F2hc staining (p = 0.0098) (Table I). The grade of 4F2hc staining also correlated statistically with PCNA index (p = 0.0264).

Table I. Correlation Of LAT1 and 4F2hc Staining with Clinical and Histopathological Features of Astrocytic Tumors
 LAT1p4F2hcp
(+) (n = 21)(++) (n = 18)(+++) (n = 21)(+) (n = 27)(++) (n = 22)(+++) (n = 11)
Age (yr)
 0–196200.06584400.731
 20–39543642
 40–597881184
 60+3410665
Tumor histology
 Low-grade astrocytoma10410.00598700.1074
 AA4851043
 GBM76159118
Gender
 Male96140.095315770.1354
 Female1212712154
PCNA index (%)
 <57400.00756500.0264
 5–30101081693
 >304413588
4F2hc
 (+)12960.0098    
 (++)886    
 (+++)119    

Correlation with patient survival

Kaplan-Meier survival plots for all patients showed a statistically significant association between high grade of LAT1 staining and poor outcome (p < 0.0001; Fig. 3a, Table II). Because survival of patients with glioma has been associated with several clinicopathological variables, we attempted to define the relative contribution of LAT1 immunostaining to survival by using multivariate Cox regression analyses with 6 variables (age, tumor histology, gender, PCNA index, LAT1 staining and 4F2hc staining). In the initial univariate analysis, age (p = 0.0006), tumor histology (p < 0.0001), 4F2hc staining (p = 0.0183), gender (p = 0.0227), PCNA index (p = 0.0392), and LAT1 staining (p < 0.0001) were all significant (Table II). For the multivariate analysis, we used the backward stepwise (Wald) method, in which variables were removed at each step, based on a 0.05 level of significance. At the final step, the last 3 variables, tumor histology (p < 0.0001), LAT1 staining (p = 0.0004), and age of patients (p = 0.0244) were found to be significant and independent of one another (Table III).

Figure 3.

(a) Actuarial survival (Kaplan-Meier method) of patients with astrocytoma whose tumors had LAT1 immunostaining of (+), (++) or (+++). (b) Actuarial survival (Kaplan-Meier method) of patients with GBM whose tumors had LAT1 immunostaining of (+), (++) or (+++).

Table II. Univariate Analysis of Prognostic Factors for Survival
 No. of patients3-yr survival (%)p (log rank)
Age (yr)
 0–19875.00.0006
 20–391241.7
 40–592343.5
 60+175.9
Tumor histology
 Low-grade astrocytoma1580.0<0.0001
 AA1735.3
 GBM2810.7
Gender
 Male2920.70.0227
 Female3154.8
PCNA index (%)
 <51163.60.0392
 5–302839.3
 >302119.0
LAT1
 (+)2161.9<0.0001
 (++)1844.4
 (+++)210.0
4F2hc
 (+)2740.70.0183
 (++)2245.5
 (+++)119.1
Table III. Multivariate Cox Regression Analysis of The Factors Associated with Survival
StepVariableRelative risk (95% CI)p (log-rank)
  • CI, confidence interval.

  • 1

    Reference category.

  • 2Potential prognostic factors selected from Table II were used.

The firstAge (yr) 0.0499
  0–191.040 (0.226–4.790)0.9600
  20–39  
  40–5911.049 (0.404–2.719)0.9220
  60+3.073 (1.173–8.048)0.0223
 Tumor histology <0.0001
  Low-grade astrocytoma1  
  AA3.673 (0.636–21.207)0.0208
  GBM13.894 (4.287–45.028)<0.0001
 LAT1 0.0059
  (+)0.179 (0.062–0.516)0.0014
  (++)0.405 (0.167–0.982)0.0455
  (+++)1  
 4F2hc 0.4627
  (+)1.455 (0.614–3.451)0.3945
  (++)0.925 (0.358–2.391)0.8714
  (+++)1  
 Gender  
  Male1.578 (0.764–3.260)0.2180
  Female1  
 PCNA (%) 0.7026
  <50.376 (0.248–2.186)0.5813
  5–300.755 (0.384–1.483)0.4139
  >301  
The finalAge (yr) 0.0244
  0–190.824 (0.192–3.538)0.7948
  20–391  
  40–590.951 (0.377–2.402)0.9159
  60+2.990 (1.162–7.697)0.0231
 Tumor histology <0.0001
  Low-grade astrocytoma1  
  AA4.153 (1.305–13.218)0.0159
  GBM11.310 (3.819–33.495)<0.0001
 LAT1 0.0004
  (+)0.167 (0.065–0.430)0.0002
  (++)0.303 (0.139–0.660)0.0026
  (+++)1  

To evaluate the effect of high LAT1 staining grade within tumor grades, we analyzed the GBM subgroup for an association between LAT1 staining and survival. We found that GBM patients with tumors of high LAT1 staining grade had a statistically significant poorer prognosis than did those with tumors of low LAT1 staining grade (p = 0.0001 log-rank) (Fig. 3b). We also found that patients with low grade atrocytomas of high LAT1 staining had a statistically poor prognosis (p = 0.0035 log-rank). In the anaplastic astrocytoma group, no difference in survival was found, but the numbers of patients in this group was too small for accurate statistical sampling.

In vitro effect of BCH on the survival of C6 glioma cells

First, to ascertain the linearity of the MTT assay in C6 glioma cells, we performed serial dilution of the cells, and effect of cell number on the colorimeter reading was observed. The OD595–655 well correlated with the actual number of the viable cells in the tested range. Accordingly, the cell number to give OD595–655 value of 0.5 (namely 25,000 cells/well) was used for the following experiment.

Effect of the various concentration of BCH (from 1 to 100 mM) on the survival of C6 glioma cells was serially observed by MTT assay (Fig. 4). In the control (without BCH), cells continuously increased in number and the OD595–655 value at day 5 showed more than 1.0. In contrast, almost complete suppression of the cell growth was observed when 25 mM BCH was added to the culture medium. When BCH was added at a concentration of more than 25 mM, cell survival was disturbed and a decrease in the OD595–655 value was observed. BCH at a concentration of 1–10 mM revealed a limited effect. We confirmed that this growth inhibitory effect of the BCH on C6 glioma cells was not due to high osmolarity of the BCH-added culture medium since addition of the equivalent molarity of D-mannitol showed no remarkable effect on the growth of C6 glioma cells (data not shown).

Figure 4.

Effect of the various concentration of BCH (from 1 to 100 mM) on the survival of C6 glioma cells was serially observed by MTT assay.

Effects of BCH on tumor sizes in vivo and on survival of rats after tumor inoculation

The volume of tumor averaged 77.9 ± 16.7 mm3 in Group 1 (n = 9) (BCH230/1 day), 146.8 ± 21.4 mm3 in Group 2 (n = 9) (saline/1 day), 95.3 ± 13.6 mm3 in Group 3 (n = 7) (BCH230/8 days), 109.9 ± 12.1 mm3 in Group 4 (n = 12) (BCH50/8 days), 144.7 ± 19.1 mm3 in Group 5 (n = 7) (saline/8 days), and 130.1 ± 21.8 mm3 in Group 6 (n = 7) (mannitol50/8 days), respectively (mean ± SE) (Fig. 5). The volume of tumor in Group 1 was significantly smaller than that in Group 2 (p = 0.022). The volume of tumor in Group 3 was also smaller than that in Group 5, but the difference was not significant (p = 0.057). All animals lost weight within 6–17 days after tumor inoculation and continuously lost it thereafter. The time point of maximal body weight was significantly prolonged in the animals treated with 230 mM of BCH (Group 1) when compared with that of Group 2 (p = 0.026). Kaplan-Meier survival data of rats in Group 1 were significant (Fig. 6), compared to that of rats in Group 2 (p = 0.016 log-rank). BCH, saline or D-mannitol administration to a concentration of 50–230 mM was not associated with occurrence of seizures, or changes in behavior (such as sluggishness or inability to eat) in rats without tumor cells. The percentage of PCNA-positive cells in Group 1 was significantly smaller than that in Group 2 (p = 0.0182).

Figure 5.

Average tumor volumes ± SE in each group are shown. Fifty-one rats were divided into 6 experimental groups. Group 1 (BCH230/1 day) (n = 9) were given 230 mM of BCH 1 day after tumor inoculation. Group 2 (saline/1 day) (n = 9) were given the saline 1 day after tumor inoculation. In group 3 (BCH230/8 days) (n = 7), animals were given 230 mM of BCH 8 days after tumor inoculation. Group 4 (BCH50/8 days) (n = 12) received 50 mM of BCH 8 days after tumor inoculation. Group 5 (saline/8 days) (n = 7) were given the saline 8 days after tumor inoculation. Group 6 (mannitol50/8 days) (n = 7) were given 50 mM D-mannitol 8 days after tumor inoculation. Statistical significance was determined by ANOVA followed by Bonferroni/Dunn test.

Figure 6.

Kaplan-Meier survival curves of C6 glioma-bearing rats after implantation of tumor cells.

Discussion

We have demonstrated for the first time that high LAT1 immunostaining predicts a poor prognosis in patients with astrocytic brain tumors in general and in patients with GBM or low grade astrocytoma in particular. However, LAT1 was the second strongest predictor of outcome in general. It is speculated that LAT1 expression is upregulated so as to provide cells with essential amino acids for high levels of protein synthesis associated with cell activation and also to support rapid growth or continuous proliferation. Indeed, we found that high LAT1 expression correlated with high proliferating potential of the tumor estimated by PCNA immunohistochemistry. LAT1 also corresponds to TA1, an oncofetal antigen that is expressed primarily in fetal tissues and cancer cells.10 A high level of LAT1 expression was also detected in human tumor cell lines such as stomach signet ring cell carcinoma (KATOIII), malignant melanoma (G-361), and lung small-cell carcinoma (RERF-LC-MA) by Northern blot analysis.3 The database search indicated that partial or incomplete sequences of LAT1 (E16, TA1 and ASUR4b) were already reported.5, 6, 11 E16 and ASUR4b were identified to be up-regulated upon the mitogenic stimulation of lymphocytes and the stimulation of A6 epithelial cell line by aldosterone, respectively,5, 11 suggesting highly regulated nature of LAT1 gene expression. TA1 was identified as a tumor-associated sequence with the oncofetal pattern of expression in rat liver.6 TA1 immunoreactivity was abundant in human colon cancer in vivo but barely detected in surrounding normal colon tissue,12 confirming the high level of expression of LAT1 protein in tumor cells. The 4F2hc is thought to be involved in the trafficking and regulation of system L neutral amino acid transport in mammalian cells as mentioned previously. Because 4F2hc is essential for LAT1 to be functional, the level of 4F2hc expression would greatly affect the formation of functional system L transporters in the plasma membrane. We found that high 4F2hc immunoreactivity also correlated with high LAT1 expression; however, LAT1 staining had a major impact on survival rate. Involvement of LAT1 in tumor progression is also strongly suggested by a recent study that showed that up-regulation of the CD98 complex, but not the CD98 heavy chain (4F2hc) alone, in Balb3T3 cells resulted in tumorigenicity in nude mice.13 LAT1 has been shown to be a transiently expressed membrane protein with the rapid degradation signal AUUUA.5 Nakamura et al. demonstrated that LAT1 is expressed minimally at the plasma membrane in cancer cells, remaining mostly in the Golgi area, and requires 4F2hc to be sorted to the cell surface.14 The immunoreactivity of LAT1 in the plasma membrane may represent its function. We did not differentiate the immunoreactivity of both cytosol and plasma membrane when estimating the grade of immunoreactivity of LAT1. However, the overall immunoreactivity for LAT1 did correlate well with the prognosis of patients with astrocytic tumors. Cytoplasmic LAT1 immunoreactivity may represent an intracellular pool of LAT1, and may correlate with the biological activity of cells.

To clarify the role of LAT1 in glioma cells, we tested a relatively specific inhibitor to LAT1, BCH, and found that BCH suppressed strongly C6 glioma cell growth in vitro. In addition, BCH also inhibited mortality of rats treated with C6 glioma cells. BCH at a concentration of 25 mM showed no significant effect on normal human astrocytes in vitro (data not shown). BCH is a nonmetabolizable artificial amino acid and a transportable inhibitor for LAT1. Yanagida et al. showed that a high-affinity substrate, leucine, and a low-affinity substrate, glutamine, but not a nonsubstrate, alanine, were effluxed via LAT1 by the application of leucine in the medium, confirming that LAT1 is an amino acid exchanger.15 This, furthermore, suggests that the substrate selectivity of the intracellular substrate binding site of LAT1 is similar to that of the extracellular substrate binding site. Amino acids are released via LAT1 in exchange for the influx of amino acids; thus, no net amino acid influx should be observed. Glutamine, which is abundantly present in cells and generated intracellularly, is transported by LAT1 albeit with low affinity, consistent with a previous report for Xenopus LAT1.4 Yanagida et al. further demonstrated that intracellularly loaded glutamine is effluxed in exchange for extracellularly applied leucine.15 Therefore, they propose that extracellular high-affinity LAT1 substrates, most of which are essential amino acids, are taken up by cells via LAT1 driven by the exchange for intracellular-glutamine, which results in the net influx of essential amino acids.15 An interesting finding of the Northern blot analysis of the tumor cell line is that the expression level of 4F2hc is quite varied among tumor cell lines, particularly in leukemia cell lines.15 Yanagida et al. found 3 leukemia cell lines in which 4F2hc messages were not detected.15 In those cell lines that lack 4F2hc expression, LAT1 was still expressed, suggesting different mechanisms of regulation in LAT1 and 4F2hc gene expression. Consistent with this, it was shown that LAT1 and 4F2hc respond differently to amino acid availability in rat hepatic cells.16

Several clinical investigations demonstrated the significant relation of the uptake of radiolabeled amino acid in gliomas and proliferation, biological aggressiveness or histological grading of these tumors.17, 18 A significant correlation of iodine-123–methyltyrosine (IMT) uptake in gliomas and the expression of the proliferation marker Ki-67 has been reported.18 Recent studies also demonstrated significant longer survival times in patients with cerebral gliomas with low amino acid uptake than in gliomas with high amino acid uptake.19 The results of this study support the hypothesis that the uptake of radiolabeled amino acids such as IMT is dependent on the proliferative activity of human gliomas. It is noteworthy that in cultured human glioma cells, membrane transport of IMT is dominated by BCH-sensitive transport system, LAT1.20, 21 LAT1-mediated IMT transport and 4F2 antigen expression are dependent on proliferation rate of human glioma cells in vitro and are significantly correlated to each other.20 These data give further support to the involvement of the LAT1 in cell proliferation.20, 22 Thompson and coworkers recently reported the role of LAT1 as a potential therapeutic target in hepatic tumor cells in vitro.23

The present study suggests that LAT1 may play an important role in human high-grade gliomas. In addition, inhibitors to LAT1 may be an effective therapeutic option for high-grade gliomas.

Ancillary