• Open Access

Correlation of L-type amino acid transporter 1 and CD98 expression with triple negative breast cancer prognosis

Authors


To whom correspondence should be addressed.
E-mail: mio-k@med.gunma-u.ac.jp

Abstract

Triple negative breast cancer (TNBC) is a heterogeneous, aggressive cancer for which there is no effective chemotherapy or targeted therapy. We aimed to evaluate L-type amino acid transporter (LAT) 1 and CD98 expression immunohistochemically in patients with breast cancer, especially TNBC. Out of 129 patients, LAT1 was positive in 56 patients (43.4%), and CD98 was positive in 41 patients (31.8%). The positive ratio of LAT1 expression in luminal A cases was 7.9%, 30.0% in luminal B cases, 71.4% in HER2 cases and 64.0% in TN cases. HER2 and TN subtypes expressed LAT1 and CD98 at higher levels than luminal A and B subtypes (both P < 0.001). LAT1 and CD98 expression correlated with tumor size (LAT1, P = 0.010; CD98, P = 0.007), nuclear grade (LAT1, P < 0.001; CD98, P < 0.001) and Ki67 labeling index (LAT1, P < 0.001; CD98, P = 0.001). LAT1 and CD98 expression was negatively associated with ER and PgR (both P < 0.001). In TNBC, the 5-year disease-free rate of CD98+ (63.6%) or LAT1+/CD98+ (61.9%) patients was significantly worse than that of CD98− (89.3%) patients or those with no co-expression of LAT1 and CD98 (89.7%), respectively (P = 0.014, P = 0.009). The 5-year survival rates of CD98 positive/negative patients were 77.3% and 100% (P = 0.050), respectively, whereas that of patients with LAT1+/CD98+ (76.2%) was significantly worse (100%) (P = 0.040). Multivariate analysis confirmed that CD98+ or LAT1+/CD98+ expression were risk factors for relapse in TNBC (P = 0.023, P = 0.019). Thus, in the present study we show that LAT1 and CD98 expression are prognostic factors. Inhibition of these proteins might provide a new therapeutic strategy in TNBC. (Cancer Sci 2012; 103: 382–389)

Breast cancer is a heterogeneous disease, and for many years invasive breast cancer was classified according to the extent of tumor spread, histological features and expression of the hormonal receptors, estrogen receptor (ER) and progesterone receptor (PgR). The subtypes that express hormonal receptors have been predicted to respond to endocrine therapy with tamoxifen(1) or aromatase inhibitors for two decades.(2) Recently, human epidermal growth factor receptor-2 (HER2) expression status has attracted a great deal of attention, because anti-HER2 therapies, such as trastuzumab treatment, have improved patient prognosis.(3–5) However, there are no targeted therapeutic agents for triple negative breast cancer (TNBC) (which is negative for ER, PgR and HER2) because of its heterogeneity, and the only treatment option is conventional systemic chemotherapy, despite its aggressive behavior.(6–12) Basal-like phenotype, which is classified by gene expression and low levels of ER, an absence of HER2 overexpression and positive basal-marker expression, clinically overlaps with TNBC.(12) Mutations in BRCA1 are reported to confer a high risk of development of the basal-like subtype, and loss of BRCA1 function is one of the determinants for the treatment of this phenotype.(13–15)

Carrey et al. classify breast cancer immunohistochemically into five subtypes based on ER, PgR, HER2 and basal marker status: (i) luminal A subtype (ER [+] and/or PgR [+], HER2 [−]); (ii) luminal B subtype (ER [+] and/or PgR [+], HER2 [+]); (iii) HER2 subtype (ER [−] and PgR [−], HER2 [+]); (iv) basal-like subtype (triple negative [TN], cytokeratin 5/6 [CK5/6] positive and/or epidermal growth factor receptor [EGFR] positive); and (v) unclassified subtype (TN, CK5/6 negative and EGFR negative).(16)

CD98 (4F2hc) was discovered through screening of mRNA to identify membrane proteins that are overexpressed in invasive breast cancer.(17) Overexpression of CD98 mRNA is correlated with histological grade and the basal-like phenotype, and is negatively correlated with ER and PgR expression. Moreover, it is significantly associated with shorter overall survival (OS). One of the main functions of CD98 is to act as a heavy chain for amino acid transport systems, and it is reported that CD98 and L-type amino acid transporter 1 (LAT1) are both overexpressed on the membranes of tumor cells to function together as an Na+-independent neural amino acid transporter.(18–20) Kanai et al. analyze LAT1 function as a transporter and report that LAT1 and CD98 are essential for the transport of large neutral amino acids such as leucine, isoleucine, valine, phenylalanine, tryptophan and methionine.(18–22) Additionally, we and several other researchers have described the potential utility of LAT1 expression or LAT1 plus CD98 expression as a prognostic factor for human neoplasms.(23–28) Human breast cancer cells also express LAT1 and CD98,(29,30) and inhibition of LAT1 and CD98 reduces their proliferation in vitro.(31) Further studies have revealed that both CD98 mRNA overexpression and LAT1 protein expression in patients with breast cancer indicate poor prognosis.(17,32,33) However, to our knowledge, there is no prognostic analysis of TNBC associated with LAT1 and CD98 expression.

In the present study, we examined the association of LAT1 and CD98 expression with clinicopathological features and ER, PgR and HER2 status in patients with invasive breast cancer. We also investigated the relationship between LAT1 and CD98 expression, clinicopathological features and survival in TNBC.

Materials and Methods

Patients.  Of patients with invasive breast cancer who were surgically treated at Gunma University Hospital, Japan between 1997 and 2009, we analyzed 50 consecutive triple negative (TN) subtype (ER [−], PgR [−] and HER2 [−]) breast carcinomas, which were completely resected and well preserved with enough material; 79 cases were randomly sampled from other subtypes. In the present study, 38 cases were classified as luminal A subtype, 20 cases as luminal B subtype and 21 cases as HER2 subtype. All 129 patients had not received any neoadjuvant therapies. Tumor specimens were histologically classified according to the criteria set by the World Health Organization in 2003.(34) Of 129 cases, 119 cases were invasive ductal carcinoma, three cases were mucinous carcinoma, three cases were invasive lobular carcinoma, two cases were apocrine carcinoma, one case was an invasive micropapillary carcinoma and one case was a metaplastic carcinoma. Patient characteristics are shown in Table 1. Among four subtypes, no obvious differences were observed in age, histopathological features, pT, pN and stage. In contrast, the intraductal component was found to be less in the TN subtype than in other subtypes. The nuclear grade was higher in HER2 or TN cancers than other subtypes, and Ki67-labeling index (LI) was higher in the TN subtype than in other subtypes, as described in previous reports.(16,35) The day of surgery was considered the starting day for determining post-operative disease-free survival (DFS) and OS. The follow-up duration ranged from 1 to 163 months, and the median duration was 58.1 months.

Table 1.   Clinicopathological characteristics
 Number (%)AllLuminal ALuminal BHER2TNP-value
  1. Ki67 LI, Ki67 labeling index; NS, not significant; TN, triple negative.

Patients 12938 (29.5)20 (15.5)21 (16.3)50 (38.8) 
AgeMean 55.25351.860.256NS
Range 28–8634–8128–7944–7631–86
Intraductal componentAbsent 40 (31.0) 3 (7.9) 7 (35.0) 6 (28.6)24 (48.0)0.001
Present 89 (69.0)35 (92.1)13 (65.0)15 (71.4)26 (52.0)
Histopathological featureInvasive ductal carcinoma119 (92.2)33 (86.8)20 (100.0)20 (95.2)46 (92.0)NS
Special subtypes 10 (7.8) 5 (13.2) 0 (0.0) 1 (4.8) 4 (8.0)
Mucinous carcinoma  3 (2.3) 3 (7.9) 0 (0.0) 0 (0.0) 0 (0.0)
Invasive lobular carcinoma  3 (2.3) 2 (5.3) 0 (0.0) 0 (0.0) 1 (2.0)
Apocrine carcinoma  2 (1.6) 0 (0.0) 0 (0.0) 0 (0.0) 2 (4.0)
Invasive micropapillary carcinoma  1 (0.8) 0 (0.0) 0 (0.0) 1 (4.8) 0 (0.0)
Metaplastic carcinoma  1 (0.8) 0 (0.0) 0 (0.0) 0 (0.0) 1 (2.0)
Nuclear grade1 or 2 49 (38.0)33 (86.8) 8 (40.0) 3 (14.3) 5 (10.0)<0.001
3 80 (62.0) 5 (13.2)12 (60.0)18 (85.7)45 (90.0)
Ki67-LIMean 39.617.839.447.153.2<0.001
Range  2.0–96.0 2.0–59.610.0–76.417.7–86.7 7.0–96.0
pT1 71 (55.0)21 (55.3)15 (75.0) 7 (33.3)28 (56.0)NS
2 52 (40.3)16 (42.1) 5 (25.0)11 (52.4)20 (40.0)
3  4 (3.1) 1 (2.6) 0 (0.0) 2 (9.5) 1 (2.0)
4  2 (1.6) 0 (0.0) 0 (0.0) 1 (4.8) 1 (2.0)
pN0 79 (61.2)23 (60.5)12 (60.0)10 (47.6)34 (68.0)NS
1 28 (21.7)10 (26.3) 5 (25.0) 5 (23.8) 8 (16.0)
2 18 (14.0) 5 (13.2) 3 (15.0) 4 (19.0) 6 (12.0)
3  4 (3.1) 0 (0.0) 0 (0.0) 2 (9.5) 2 (4.0)
pStageI 47 (36.4)14 (36.8)11 (55.0) 3 (14.3)19 (38.0)NS
II 60 (46.5)21 (55.3) 6 (30.0)10 (47.6)23 (46.0)
III 22 (17.1) 3 (7.9) 3 (15.0) 8 (38.1) 8 (16.0)
IV  0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)

Immunohistochemical staining.  Surgically resected specimens were fixed by 10% formalin and embedded in paraffin. The paraffin blocks were cut into 4-μm thick sections. We performed immunohistochemical staining for ER, PgR, HER2, Ki67, CK5/6, EGFR, LAT1 and CD98. Antibody dilutions and cut-off values are listed in Table 2.

Table 2.   Antibodies used in the present study and immunohistological assessment
AntigenCloneDilutionPositive score (%)
  1. EGFR, epidermal growth factor receptor; ER, estrogen receptor; HER2, human epidermal growth factor receptor-2; LAT1, L-type amino acid transporter 1; PgR, progesterone receptor.

ERMonoclonalRabbitSP-11:1>10
PgR1E2>10
HER24B5>10
Ki-6730-9Positive rate
CK5/6D5&16B4>5
EGFR3C6>10
LAT1MonoclonalMouse4A21:1>10
CD98PolyclonalRabbitH-300  1:200>10

Estrogen receptor, progesterone receptor, human epidermal growth factor receptor-2, Ki67, cytokeratin 5/6 and epidermal growth factor receptor.  Immunohistochemical staining for ER, PgR, HER2, Ki67, CK5/6 and EGFR was performed using BenchMark XT (Ventana Medical Systems Inc., Tucson, AZ, USA), an automatic immunohistochemical staining system.

Cases were considered positive if ER and PgR staining in the nucleus was present in >10% of tumor cells. HER2 expression was considered positive if it followed two patterns: (i) cases that showed distinct membrane staining on >30% of tumor cells; and (ii) cases that showed weak to distinct membrane staining on >10% of tumor cells and where amplifications detected by FISH were positive (HER2/CEP17 > 2.0).

Ki67 expression was considered positive if nuclear staining was present. A highly cellular area of the immunostained sections was evaluated. We counted approximately 1000 cancer cells on each slide and assessed the proliferative activity of Ki67-LI as the percentage of immunopositive nuclei.

L-type amino acid transporter 1 and CD98.  Sections were deparaffinized with xylene and rehydrated with ethanol, then treated with 0.3% hydrogen peroxidase in methanol for 30 min to block endogenous peroxidase activity. Antigen retrieval was performed for LAT1 staining by placing the specimen in 0.01 mol/L of citrate buffer at pH 6.0 and exposed to microwave heating for 20 min. After treatment with 5% normal horse serum, sections were incubated for 2 h at room temperature with LAT1 antibody (2 μg/mL, anti-human monoclonal mouse antibody, 4A2, provided by Dr H. Endou [J-Pharma, Tokyo, Japan]). Sections were incubated with Simple Stain MAX PO (Nichirei Bioscience, Tokyo, Japan) secondary antibody for 20 min. The peroxidase reaction was performed using 0.02% 3-3′ diaminobenzidine tetrahydrochloride (Nichirei Bioscience). Finally, nuclear counterstaining was performed with Mayer’s hematoxylin. Immunohistochemical staining for CD98 was performed similarly to LAT1 but without antigen retrieval. After treatment with 10% normal goat serum, sections were incubated with CD98 antibody (1:200, anti-human polyclonal rabbit antibody, H-300; Santa Cruz Biotechnology, Santa Cruz, CA, USA).

L-type amino acid transporter 1 and CD98 expression were considered positive only when membrane staining was present, regardless of staining of the intraductal or invasive components. Staining intensity was scored as follows: 0, no tumor cells stained; 1, ≥0% and <10% of tumor area stained; 2, ≥10% and <30% stained; 3, ≥30% and <50% stained; 4, >50% stained. Scores from 2 to 4 were considered positive.

Statistical analysis.  Statistical analysis was carried out using SPSS 17.0 software (SPSS Japan, Tokyo, Japan). Spearman’s rank test was performed to examine the association between two variables, the Mann–Whitney and the t-test, to examine the differences between two variables, and one way analysis of variance and the Kruskal–Wallis test were performed to verify the differences among plural variables. The chi-square-test and Fisher’s exact test were used to examine plural categorical variables.

Disease-free survival time and OS time were determined as the time from tumor resection to recurrence or death caused by breast cancer. For survivors, the time from resection to the last date that the patients were confirmed to be alive was counted. Cox’s proportional hazard method was performed for multivariate analysis to identify the independent prognostic factors. A P-value <0.05 was considered to be statistically significant.

Results

L-type amino acid transporter 1 and CD98 expression in breast cancer.  L-type amino acid transporter 1 and CD98 immunoreactivity was detected in the membrane of tumor cells (Fig. 1). No expression was observed in normal mammary ducts. LAT1 and CD98 immunohistochemical scores for all 129 cases of breast cancer are shown in Table 3. LAT1 was positive in 43.4% (56 of 129 patients) and CD98 was positive in 31.8% (41 of 129 patients).

Figure 1.

 Immunohistochemical analysis of resected tumor. The immunostaining pattern of L-type amino acid transporter 1 (LAT1, A) and CD98 (B) was localized to the cell membrane.

Table 3.   LAT1 and CD98 expression in all cases
SubtypenLAT1 scoreLAT1+ (%)CD98 scoreCD98+ (%)
0123401234
  1. Staining intensity was scored as follows: 0, no tumor cells stained; 1, ≥0 and <10% of tumor area stained; 2, ≥10% and <30%; 3, ≥30% and <50%; 4, >50%. LAT1, L-type amino acid transporter 1.

All cases129462722171756 (43.4)42462214541 (31.8)
Luminal A383141113 (7.9)25111102 (5.3)
Luminal B20953216 (30.0)883104 (20.0)
HER2212426715 (71.4)1785013 (61.9)
TN50414168832 (64.0)820107522 (44.0)

Co-expression of LAT1 with CD98 was observed in 38 patients. LAT1+ and CD98-expression was found in 18 patients, and LAT1− and CD98+ expression was identified in two patients. Neither was expressed in 71 patients. The rate of coincident expression between LAT1 and CD98 was 84.5%. LAT1 and CD98 scores were significantly correlated (γ = 0.772, P < 0.001) (Fig. 2). The average score of CD98 corresponded to LAT score 0 was 0.30; LAT1 score 1, 0.78; LAT1 score 2, 1.68; LAT1 score 3, 2.24; and LAT1 score 4, 2.47.

Figure 2.

 Correlation between L-type amino acid transporter 1 (LAT1) and CD98 scores. LAT1 and CD98 scores were significantly correlated (γ = 0.772, < 0.001).

There were no obvious differences between non-invasive and invasive components for LAT1 and CD98 scoring. In one case, both LAT1 and CD98 scored 3 in the total area of all tumor cells, whereas both were of score 1 in the invasive area.

Correlation between L-type amino acid transporter 1 and CD98 and clinicopathological features.  Both LAT1 and CD98 expression were significantly correlated with tumor size (LAT1, P = 0.010; CD98, P = 0.007), nuclear grade (LAT1, P < 0.001; CD98, P < 0.001) and Ki67-LI (LAT1, P < 0.001; CD98, P = 0.001), respectively (Table 4). Significant correlations were observed between Ki67-LI and LAT1 or CD98 score, independently (LAT1, γ = 0.519, P < 0.001; CD98, γ = 0.419, P < 0.001) (Fig. 3).

Table 4.   Correlation between LAT1, CD98 and clinicopathological featuresThumbnail image of
Figure 3.

 Correlation between L-type amino acid transporter 1 (LAT1), CD98 classification and Ki67 labeling index. Correlation of Ki67 labeling index (A) with L-type amino acid transporter 1 (LAT1) (γ = 0.519, < 0.001) and (B) with CD98 (γ = 0.419, < 0.001).

L-type amino acid transporter 1 and CD98 expression in breast cancer subtypes and association with estrogen receptor, progesterone receptor and human epidermal growth factor receptor-2.  L-type amino acid transporter 1 and CD98 scores in each subtype are listed in Table 3. HER2 and TN subtypes expressed LAT1 and CD98 at a higher frequency than luminal A and luminal B subtypes (both P < 0.001). Moreover, LAT1 and CD98 expression in the TN subtype and other subtype was significant (LAT1, P < 0.001; CD98, P = 0.018).

Expression of LAT1 and CD98 according to hormonal receptors or tyrosine kinase is shown in Table 5. Both LAT1 and CD98 expression was significantly associated with negative expression of ER or PgR (both P < 0.001). HER2 expression showed no obvious association with either LAT1 or CD98 expression (LAT1, P = 0.222; CD98, P = 0.107). The TN subtype was associated with LAT1 or CD98 expression (LAT1, P < 0.001; CD98, P = 0.001).

Table 5.   Correlation of LAT1 and CD98 expression with ER, PgR and HER2
 nLAT1+ (%) CD98+ (%) 
  1. *Statistically significant, < 0.05. ER, estrogen receptor; HER2, human epidermal growth factor receptor-2; LAT1, L-type amino acid transporter 1; PgR, progesterone receptor; TN, triple negative.

ER
 +589 (15.5)P < 0.001*6 (10.3)P < 0.001
 −7147 (66.2)35 (49.3)
PgR
 +467 (15.2)P < 0.0015 (10.9)P < 0.001
 −8349 (59.0)36 (43.4)
HER2
 +4121 (51.2)P = 0.22217 (41.5)P = 0.107
 −8835 (39.8)24 (27.3)
Non-TN7924 (30.4) 19 (24.1) 
TN5032 (64.0)P < 0.001*22 (44.0)P = 0.018*

L-type amino acid transporter 1 and CD98 expression in triple negative subtype.

Correlation between L-type amino acid transporter 1 or CD98 expression and clinicopathological features in TN subtype.  In the TN subtype, there was no correlation between LAT1/CD98 expression and clinicopathological features such as age, tumor size, lymph node metastasis and Ki67-LI, with the exception of nuclear grade (LAT1+/−, 96.9/77.8%, P = 0.031; CD98+/−, 100.0/82.1%, P = 0.037). Significant correlation with Ki67-LI was observed with LAT1 score only (LAT1, γ = 0.289, P < 0.042; CD98, γ = 0.069, P = 0.635).

Triple negative subtype was divided into two categories, basal-like subtype and unclassified subtype:(16) the basal-like subtype expresses CK5/6 and/or EGFR, and the unclassified subtype expresses neither CK5/6 nor EGFR. The ratio of LAT1 expression was significantly higher in the basal-like subtype than the unclassified subtype (P = 0.022). Moreover, the ratio of LAT1 or CD98 expression in EGFR-positive cases was higher than in EGFR-negative cases (LAT1, P = 0.012; CD98, P = 0.005) (Table 6).

Table 6.   Basal marker expression in triple negative subtype
  LAT1+ (%) CD98+ (%) 
  1. *Statistically significant, < 0.05. EGFR, epidermal growth factor receptor; LAT1, L-type amino acid transporter 1.

Basal like subtype3023 (76.7)P = 0.022*15 (50.0)P = 0.295
Unclassified subtype209 (45.0)7 (35.0)
CK5/6
 +2215 (68.2)P = 0.5859 (40.9)P = 0.696
 −2817 (60.7)13 (46.4)
EGFR
 +1312 (92.3)P = 0.012*10 (76.9)P = 0.005*
 −3720 (54.1)12 (32.4)

L-type amino acid transporter 1 and CD98 expression and correlation with survival in triple negative subtype.  The follow-up period for 50 patients with the TN subtype ranged from 1 to 163 months (mean, 68.6 months). Tumors reoccurred in 13 patients and seven patients died as a result of breast cancer during the course of the study. For all 50 patients, the 5-year DFS rate was 22.0% (n = 11) and the 5-year survival rate was 10.0% (n = 5).

The 5-year DFS rates of LAT1 positive and negative patients, and CD98 positive and negative patients, were 75.0% and 83.3% (P = 0.398, Fig. 4A), and 63.6% and 89.3% (P = 0.014, Fig. 4B), respectively. In turn, 5-year DFS rates for patients exhibiting co-expression of LAT1/CD98 and others, or with no expression of either LAT1 nor CD98 and others, were 61.9% and 89.7% (P = 0.009, Fig. 4C), and 84.2% and 74.2% (P = 0.860), respectively.

Figure 4.

 Survival rate of TNBC patients was associated with L-type amino acid transporter 1 (LAT1) and CD98 expression. Post-operative disease-free survival (DFS) and overall survival (OS) of patients with completely resected TNBC according to LAT1 and CD98 expression status. (A) DFS correlated with LAT1 expression (= 0.398); (B) DFS correlated with CD98 expression (P = 0.014); (C) DFS correlated with LAT1/CD98 co-expression and others (P = 0.009); (D) OS correlated with LAT1 expression (P = 0.338); (E) OS correlated with CD98 expression (P = 0.050); (F) OS correlated with LAT1/CD98 co-expression and others (P = 0.040).

The 5-year survival rates of LAT1 positive and negative patients, and CD98 positive and negative patients, were 84.4% and 100% (P = 0.338, Fig. 4D), and 77.3% and 100% (P = 0.050, Fig. 4E), respectively. For patients with co-expression of LAT1/CD98 and others, or with no expression of either LAT1 or CD98 and others, they were 76.2% and 100% (P = 0.040, Fig. 4F), and 100% and 83.9% (P = 0.383), respectively.

Univariate analysis was performed to identify the factors that were significantly associated with relapse. These were found to be lymph node metastasis (P = 0.017), CD98 expression (P = 0.025) and LAT1/CD98 co-expression (P = 0.018). However, death was not found to be a significant factor. We performed multivariate analysis for DFS using these data and found that CD98 expression and lymph node metastasis (N) were independent and significant risk factors for relapse (CD98, hazard ratio 0.224, 95% CI 0.061–0.816, P = 0.023; N, hazard ratio 0.250, 95% CI 0.082–0.767, P = 0.015). Likewise, co-expression of LAT1/CD98 and lymph node metastasis were independent risk factors for relapse (LAT1/CD98, hazard ratio 0.213, 95% CI 0.058–0.779, P = 0.019; N, hazard ratio 0.261, 95% CI 0.085–0.800, P = 0.019).

Discussion

CD98 and LAT1 are both overexpressed on the membrane of tumor cells to act as an Na+-independent neural amino acid transporter.(18–20) The heavy chain of CD98 is encoded by SLC3A2 and binds with the LAT1 light chain protein product encoded by SLC7A5. They play an important role in the regulation of cellular proliferation; however, the precise mechanism by which they promote tumor cell proliferation is still unknown. An in vitro study first hypothesized a link between amino acid transporter expression and mammalian target-of-rapamycin (mTOR) function.(21) CD98 was discovered through mRNA screening to identify membrane proteins overexpressed in invasive breast cancer.(17) In addition to its role in amino acid transport, CD98 mediates integrin-dependent signals and promotes tumorigenesis.(36,37) Therefore, the LAT1 and CD98 complex can function as molecular machinery to regulate amino acid transport and cell adhesion. The present study evaluated the association between LAT1 and CD98 immunohistochemical expression with clinicopathological significance in invasive breast cancer, to determine their clinical features, especially in TNBC.

In our study, the expression of LAT1 and CD98 was not observed in normal mammary ducts, and the scores of LAT1 and CD98 were significantly correlated with each other. A previous study demonstrated that the grade of LAT1 and CD98 expression is correlated in glioma,(23) and in another study, LAT1 protein co-expression with CD98 was observed in 71.7% in transitional cell carcinoma of the upper urinary tract cancers and in 84.9% of stage I pulmonary cancers.(24) These results demonstrated that LAT1 and CD98 protein are co-expressed at a high rate, and confirmed that LAT1 requires the CD98 heavy chain for its functional expression.(18–20)

We showed that the rate of LAT1 (43.4%) positive expression was higher than that of CD98 (31.8%) in invasive breast cancer. Earlier reports revealed that the LAT1 expression rate was lower than CD98,(24) whereas others report that LAT1 is expressed more highly than CD98.(23,26) This might be caused by differences in sensitivity of the LAT1 or CD98 antibodies. Alternatively, the method of assessment for immunostaining, such as if the cytoplasmic staining is considered to be positive, or differences in the cutoff point for positive staining might be factors.

Both LAT1 and CD98 expression were significantly correlated with tumor size, nuclear grade and Ki67-LI. Significant correlation was observed between Ki67-LI and LAT1 or CD98 score, independently. We analyzed the association of Ki67-LI with LAT1 and CD98 expression in invasive breast cancer. Previous reports of other organs also show correlation between LAT1 and CD98 expression and tumor proliferative index. In glioma, the grade of LAT1 and CD98 staining was significantly correlated with anti-proliferating cell nuclear antigen (PCNA) index (LAT1, P = 0.0075; CD98, P = 0.0264).(23) In stage I pulmonary cancer, LAT1 and CD98 expression correlated significantly with Ki67-LI (P < 0.0001).(26) Therefore, LAT1 and CD98 were confirmed to act as markers of proliferative activity in the present study.

Esseghir et al.(17) report that CD98 mRNA shows negative correlation with ER and PgR. Moreover, we revealed that CD98 expression and LAT1 are negatively associated with ER and PgR. Although there is no obvious explanation for the mechanism, mammary cell differentiation and tumorigenesis might be related to the reverse relationship according to the newly proposed relationships between human breast epithelial hierarchy and breast cancer subtypes.(14,38)

The TN subtype is divided into two categories, basal-like subtype and unclassified subtype, depending upon the expression of basal markers, as stated above.(16) In the present study, we revealed the correlation between LAT1 expression and the basal-like subtype; however, we found no obvious correlation between CD98 and the basal-like subtype. Moreover, LAT1 and CD98 was significantly associated with EGFR; however, neither showed obvious correlation with CK5/6. Esseghir et al.(17) reveal that CD98 mRNA showed correlation with CK 5/6 and EGFR (P < 0.0001). However, both LAT1/CD98 and EGFR stimulate mTOR signaling and regulate cell proliferation, but the direct mechanism is unknown at present. EGFR provides the signal for mTOR in the PI3K/Akt pathway; however, LAT1 and CD98 are reported to stimulate mTOR via a pathway other than PI3K/Akt.(21,39)

We analyzed the relationship between LAT1 and CD98 expression and prognosis in TNBC. Previous studies revealed LAT1 or co-expression of LAT1 and CD98 to be prognostic factors of lung cancer, prostate cancer, urothelial carcinoma of the upper urinary tract and astrocytoma.(23–28) The present study revealed that TNBC patients with CD98 expression or LAT1 and CD98 co-expression had unfavorable DFS, whereas patients with LAT1 and CD98 expression had unfavorable OS. Multivariate analysis revealed that positive expression of CD98 or co-expression of LAT1 and CD98 were independent, and were significant risk factors for relapse.

This result is generally compatible with past studies of breast cancer examining all subtypes;(17,32,33) however, there was no obvious relation between LAT1 expression only and the prognosis in TNBC. Because there was still no significant relationship, even though we raised the cutoff value for LAT1 expression in TNBC, the other roles of CD98, such as controlling integrin signals, might be more significant. In any case, co-expression of LAT1 and CD98 is a prognostic factor in DFS and OS, and, furthermore, it is an independent relapse factor in patients with TNBC. TNBC patients receive no targeted therapy, with the only option being conventional systemic chemotherapy.(6–12) In some patients with TNBC, chemotherapy is more successful than for other subtypes; however, in others, recurrence at the early stage is problematic and the OS for TNBC patients is poorer.(40) Because TNBC is a highly heterogeneous subtype,(12) it was extremely important that co-expression of LAT1 and CD98 was included as a significant prognostic factor in the multivariate analysis.

In conclusion, we revealed that LAT1 and CD98 expression are positively correlated with tumor proliferation and negatively correlated with ER and PgR status, respectively. Although the present study was limited by the small sample size, and analysis was restricted to TNBC with respect to prognosis, CD98, especially CD98 plus LAT1 expression, is found to be a prognostic factor for DFS and OS in TNBC. Moreover, CD98 plus LAT1 expression is an independent prognostic factor for DFS. Although further investigation is necessary, inhibition of LAT1 and CD98 function might aid in diagnosis and provide therapeutic targets for hormonal receptor-negative cancer.

Acknowledgments

The authors thank Dr Hitoshi Endou (J-Pharma) for the gift of the LAT1 antibody. We also express our gratitude to Mr T. Hikino and Ms M. Saito for their excellent technical assistance.

Disclosure Statement

The authors have no conflict of interest.

Ancillary