Ann Surg Oncol (2013) 20:2699–2705 DOI 10.1245/s10434-013-2900-2

ORIGINAL ARTICLE – THORACIC ONCOLOGY

C4.4A Expression Is Associated with a Poor Prognosis of Esophageal Squamous Cell Carcinoma Masahisa Ohtsuka, MD, Hirofumi Yamamoto, MD, PhD, Toru Masuzawa, MD, PhD, Hidekazu Takahashi, MD, PhD, Mamoru Uemura, MD, PhD, Naotsugu Haraguchi, MD, PhD, Junichi Nishimura, MD, PhD, Taishi Hata, MD, PhD, Makoto Yamasaki, MD, PhD, Hiroshi Miyata, MD, PhD, Ichiro Takemasa, MD, PhD, Tsunekazu Mizushima, MD, PhD, Shuji Takiguchi, MD, PhD, Yuichiro Doki, MD, PhD, FACS, and Masaki Mori, MD, PhD, FACS Department of Gastroenterological Surgery, Graduated School of Medicine, Osaka University, Osaka, Japan

ABSTRACT Background. C4.4A is a glycolipid-anchored membrane protein expressed in several human malignancies. We examined clinical relevance of C4.4A expression in 111 esophageal squamous cell carcinoma (ESCC) tissue samples. Methods. Anti-human C4.4A antibody that recognizes the glycosylphosphatidyl inositol (GPI) anchor signaling sequence (C4.4A-GPI Ab) and anti-human C4.4A-119 polyclonal antibody (C4.4A-119 Ab) were used for immunohistochemistry and Western blot testing. Results. Both antibodies detected the C4.4A protein expression at the parabasal layer of normal epithelium of the esophagus. In tumor tissues, the C4.4A protein was detected in 66 (59.5 %) and 95 (85.6 %) of 111 ESCCs by the C4.4A-GPI Ab and the C4.4A-119 Ab, respectively. The C4.4A-GPI Ab mainly detected membranous C4.4A expression (83.3 %, 55 of 66 positive cases), while the C4.4A-119 Ab exclusively detected cytoplasmic C4.4A expression (100 %, 73 cytoplasm alone and 22 cytoplasm plus membrane in 95 positive cases). Western blot analysis indicated that normal epithelium expressed the band of C4.4A at 70 kDa, whereas the tumor tissues displayed the band at the lower molecular weight. Survival analysis indicated that the C4.4A-positive ESCCs had significantly worse 5-year overall survival than the C4.4A-negative ESCC samples (P = 0.021) when using the C4.4A-GPI Ab, but not when using the C4.4A-119 Ab. This difference was most evident with membranous expression of C4.4A (P = 0.005). Ó Society of Surgical Oncology 2013 First Received: 16 October 2012; Published Online: 24 February 2013 H. Yamamoto, MD, PhD e-mail: [email protected]

Conclusions. C4.4A expression was associated with a poor prognosis of ESCC when the GPI-related antibody was used. On the other hand, the C4.4A-119 Ab may be a useful diagnostic tool for ESCC because of its high detection rate.

Esophageal cancer is the eighth most common cancer worldwide and the sixth most common cause of death from cancer.1 In Asian countries, esophageal squamous cell carcinoma (ESCC) is more prevalent than adenocarcinoma and accounts for [90 % of esophageal carcinomas.2 Although the mortality rate for all cancers has declined worldwide since the mid-1980s, the prognosis of ESCC patients still remains poor.1,3,4 To improve the unfavorable outcome of ESCC, it is essential to explore the molecular basis of the underlying mechanism of this disease. The C4.4A protein was initially found in a metastasizing rat pancreatic adenocarcinoma cell line.4,5 Rat C4.4A cDNA was cloned and the glycosyl phosphatidyl inositol (GPI)-anchored membrane protein was found to have 30 % homology to the urokinase-type plasminogen activator receptor.6 The human homologue of rat C4.4A, located on chromosome 19q13.1– q13.2, was subsequently cloned.7 The C4.4A mRNA is present in normal human placental tissue, skin, esophagus tissue, and leukocytes or during the wound-healing process of migrating keratinocytes or in the urothelium.7–9 Human C4.4A mRNA has been also detected in cancer cell lines, including melanoma, breast, bladder, and renal cell carcinoma, as well as in tumor tissue samples from malignant melanoma, colorectal cancer (CRC), breast cancer, lung carcinoma, and urothelial tumors.8–13 Evidence suggests that C4.4A expression is associated with a poor prognosis in patients with non–small cell lung cancer and CRC.12,14 Furthermore, Hansen et al. reported that

2700

C4.4A was expressed in 14 ESCC samples tested, when using the C4.4A antibody against domain III (the GPI anchor sequence plus the C-terminus).21 However, we are not aware of any report that examined C4.4A for the prognostic value of C4.4A in ESCC. Therefore, we examined C4.4A expression with two anti-human C4.4A polyclonal antibodies in 111 ESCC samples from patients who had not received preoperative chemotherapy or radiotherapy. MATERIALS AND METHODS Clinical Tissue Samples Esophageal tissue samples (n = 111) were collected during surgery (1998–2007) at the Department of Surgery, Osaka University, Osaka, Japan. None of the patients had preoperative chemotherapy or irradiation. Samples were fixed in buffered formalin at 4 °C overnight, processed through graded ethanol solutions, and embedded in paraffin. A piece of tissue sample was frozen in liquid nitrogen and stored at -80 °C until protein extraction. The specimens were used appropriately and under the approval of the ethics committee at the Graduate School of Medicine, Osaka University. Antibodies A rabbit anti-human C4.4A GPI-related polyclonal antibody (C4.4A-GPI Ab), and a rabbit anti-human C4.4A119 polyclonal antibody (C4.4A-119 Ab) were used. Both antibodies were previously developed in our laboratory.14–16 The rabbit anti-human actin antibody was purchased from Sigma-Aldrich (St. Louis, MO).

M. Ohtsuka et al. FIG. 1 a IHC for C4.4A with the anti-C4.4A-GPI Ab in ESCC tissue c samples. a-I Normal esophageal squamous epithelium and a C4.4AGPI Ab–negative tumor. a-II C4.4A staining on the plasma membrane and in the cytoplasm. a-III C4.4A staining in the cytoplasm. The black scale bar indicates 250 lm. b Western blot analysis. Two cases (cases 1 and 2) where the ESCC samples were positive for C4.4A by IHC, were examined by Western blot analysis using the C4.4A polyclonal antibody (C4.4A-GPI). Actin bands served as a loading control. c Kaplan–Meier plots showing survival curves stratified by C4.4A expression in ESCC tissues. c–I Using the anti-C4.4A-GPI Ab, a significant difference in 5-year OS rate was noted between C4.4A-positive and C4.4A-negative samples (P = 0.021). c-II A significant difference in 5-year OS was found between membrane staining positive and negative tumors (P = 0.036). c-III There was not a significant difference in 5-year OS between cytoplasm staining positive and negative tumors. c-IV A significant difference in 5-year OS was found between membrane staining alone and the remaining groups (P = 0.005)

onto a PVDF membrane. The membrane was incubated with the primary antibodies at the appropriate concentrations (1:4000 for C4.4A-GPI Ab, 1:400 for C4.4A-119 Ab, and 1:2000 for actin) for 1 h. The protein bands were detected with the Amersham ECL Detection System (Amersham Biosciences, Piscataway, NJ). Statistical Analysis Statistical analysis was carried out by JMP8 software (SAS Institute, Cary, NC). The Kaplan–Meier method was used to estimate tumor recurrence from CRC, and the logrank test was used to determine the statistical significance. Associations between discrete variables were assessed by the Chi square test. Mean values were compared by the Mann–Whitney U-test. P values of \0.05 were considered statistically significant. RESULTS

Immunohistochemistry Tissue sections (4 lm thick) were prepared from paraffinembedded blocks. After antigen retrieval treatment in 10 mM citrate buffer (pH 6.0) at 95 °C for 40 min, immunostaining was carried out with the Vectastain ABC peroxidase kit (Vector Laboratories, Burlingame, CA), as we described previously.17,18 The slides were incubated overnight at 4 °C at the following dilutions: C4.4A-GPI Ab, 1:300, and antiC4.4A-119 Ab. Nonimmunized rabbit IgG or mouse IgG (Vector Laboratories) was used as a negative control. Western Blot Analysis Western blot analysis was carried out as described previously.19,20 Briefly, the protein samples (20 lg) were separated by 12.5 % PAGE followed by electroblotting

C4.4A Expression in ESCC by the C4.4A-GPI Ab C4.4A protein expression was shown in the normal epithelium of the esophagus by immunohistochemistry (IHC) using the C4.4A-GPI Ab. Expression was on the plasma membrane, mainly at the parabasal layer (Fig. 1a-I). In contrast, tumor cells in the same tissue sample did not always express the C4.4A protein (Fig. 1a-I). We defined the esophageal carcinoma tissues as C4.4A-negative if staining was not noted at all in the tumor cells on the tissue sections. There were 45 C4.4A-negative esophageal tumors in the 111 cases tested (40.5 %); while, 66 esophageal tumors (59.5 %) provided clear C4.4A staining on the plasma membrane (Fig. 1a-II) or cytoplasm (Fig. 1a-III). Sixty-six positive cases were classified according to intracellular localization, i.e., 29 membrane staining alone, 11 cytoplasmic staining alone, and 26 both membrane and cytoplasmic staining.

C4.4A Expression in ESCC

A

B

C

2701

2702

M. Ohtsuka et al.

TABLE 1 C4.4A expression and clinicopathological parameters in ESCC by C4.4A-GPI Ab Characteristic

C4.4A positive (n = 66)

C4.4A negative (n = 45)

P

Age, y, median (range)

64.8 (38–84)

63.4 (49–79)

NS

59 7

41 4

NS

58

26

\0.001

8

19

T1, T7

29

25

T3,T4

37

20

Positive

56

30

Negative

10

15

Positive

51

12

Negative

35

33

47 19

28 17

NS

Positive

6

4

NS

Negative

60

41

Sex Male Female Histologya Well, moderate Poor T stage NS

Lymphatic invasion 0.036

Venous invasion 0.046

Lymph node metastasis Positive Negative Distant metastasis

ESCC esophageal squamous cell carcinoma, GPI Ab glycosylphosphatidyl inositol antibody, NS not significant a

gave prominent bands around 70 kDa in the normal squamous epithelium, while the band appeared at a slightly lower molecular weight around 60 kDa in tumor tissues (Fig. 1b). When C4.4A-positive cases (n = 66) and C4.4A-negative cases (n = 45) were compared for the various clinical and pathological parameters including age, sex, histology, T stage, lymphatic invasion, venous invasion, lymph node metastasis, and distant metastasis, C4.4A expression was associated with differentiated histology, lymphatic invasion, and venous invasion (Table 1, P \ 0.001, P = 0.036, and P = 0.046, respectively). Survival analysis indicated that the C4.4A-positive group had significantly poorer prognoses than the C4.4A negative group in 5-year overall survival (OS) (Fig. 1c-I, P = 0.021). Sub-analysis based on localization of C4.4A revealed that membranous C4.4A contributed to poor prognosis (Fig. 1c-II, P = 0.036) but cytoplasmic C4.4A did not (Fig. 1c-III). Furthermore, this difference was most evident with membrane alone expression of C4.4A (Fig. 1c-IV, P = 0.005). The univariate analysis indicated that C4.4A expression, tumor stage, lymph node metastasis, and venous invasion were predictors of a poor 5-year OS (Table 2A). We then carried out a multivariate analysis to further determine the most significant prognostic factors. The C4.4A expression, tumor stage, and lymph node metastasis were identified as independent prognostic factors (P = 0.012, P = 0.012, and P = 0.039, respectively) (Table 2B). C4.4A Expression in ESCC by the C4.4A-119 Ab

Well, moderately, and poorly differentiated squamous cell carcinoma

To validate the staining of the C4.4A protein with the C4.4AGPI Ab, Western blot analysis was performed. This antibody TABLE 2 Five-year overall survival in 111 ESCC patients

By IHC, with the C4.4A-119 Ab, 95 of 111 (85.6 %) ESCCs expressed the C4.4A protein. Thus, there were 16 C4.4A negative ESCCs (Fig. 2a-I), 73 ESCCs with

Analysis

Risk ratio

95 % CI

P

A. Univariate analysis

ESCC esophageal squamous cell carcinoma, GPI Ab glycosylphosphatidyl inositol antibody, CI confidence interval, NS not significant a

Well, moderately, and poorly differentiated squamous cell carcinoma

C4.4A-GPI Ab (positive vs. negative)

2.092

1.126–4.127

0.019

C4.4A-119 Ab (positive vs. negative)

1.357

0.621–3.568

NS

Age ([ 65 y vs. \ 65 y)

1.285

0.717–2.300

NS

Sex (male vs. female)

1.690

0.616–6.977

NS

T stage (T1/T2 vs. T3/T4)

2.769

1.540–5.080

\0.001

Differentiation (well/moderate vs. poor)a

1.696

0.875–3.119

NS

Lymph node metastasis (positive vs. negative) Distant metastasis (positive vs. negative)

3.351 1.857

1.640–7.770 0.759–3.906

\0.001 NS

Lymph invasion (positive vs. negative)

2.060

0.980–5.043

NS

Venous invasion (positive vs. negative)

2.104

1.175–3.787

0.013

C4.4A expression (positive vs. negative)

2.287

1.198–4.615

0.012

T stage (T1/T2 vs. T3/T4)

2.277

1.201–4.420

0.012

Lymph node metastasis (positive vs. negative)

2.226

1.038–5.353

0.039

Venous invasion (positive vs. negative)

1.294

0.707–2.385

NS

B. Multivariate analysis

C4.4A Expression in ESCC

2703

A

B

C

FIG. 2 a IHC for C4.4A with the anti C4.4A-119 Ab in ESCC tissue samples. a-I Normal esophageal squamous epithelium and a C4.4A119-negative tumor. a-II C4.4A staining in the cytoplasm. a-III C4.4A staining in the cytoplasm and on the plasma membrane. The black scale bar indicates 250 lm. B Western blot analysis. Two cases (cases 1 and 2) where the ESCC samples were positive for C4.4A were examined by Western blot analysis with the C4.4A-119 Ab. Actin bands served as a loading control. c Kaplan–Meier plots

showing survival curves stratified by C4.4A expression in ESCC tissues. c-I Using the anti-C4.4A-119 Ab, there was no difference in 5-year OS rate between C4.4A-positive and C4.4A-negative samples. c-II No difference in 5-year OS was found between membrane staining positive and negative tumors. c-III There was no difference in 5-year OS between the group with cytoplasm staining alone and others

cytoplasmic staining alone (Fig. 2a-II), and 22 ESCCs with membrane and cytoplasmic staining (Fig. 2a-III), and there were no ESCCs with membrane staining alone. Western blot analysis showed prominent bands for C4.4A (Fig. 2b).

There was no difference in 5-year OS between C4.4Apositive and C4.4A-negative cases (Table 2A, Fig. 2c-I), irrespective of its intracellular localization (Fig. 2c-II and III).

2704

M. Ohtsuka et al.

TABLE 3 C4.4A expression and clinicopathological parameters in ESCC by C4.4A-119 antibody Characteristic

C4.4A positive (n = 95)

C4.4A negative (n = 16)

P

Age, y, median (range)

64.0 (38–84)

64.6 (49–79)

NS

84 11

16 0

NS

Well, moderate

77

7

0.003

Poor

IS

9

T1, T2

53

9

T3, T4

42

7

Positive

76

10

Negative

19

6

Positive

39

4

Negative

56

12

66 29

9 7

NS

Positive

7

3

NS

Negative

88

13

Sex Male Female Histologya

T classification NS

Lymphatic invasion NS

Venous invasion NS

Lymph node metastasis Positive Negative Distant metastasis

ESCC esophageal squamous cell carcinoma, NS not significant a

Well, moderately, and poorly differentiated squamous cell carcinoma

Clinical and pathological survey showed that C4.4A expression was associated with differentiated histology (P = 0.003, Table 3). DISCUSSION Hansen et al. previously reported that C4.4A was observed at the invasive front of 14 ESCCs when using the C4.4A antibody that recognizes the Domain III portion of the C4.4A molecule (GPI anchor plus C-terminus).21 In the current study, we performed an extended immunohistochemical study and examined C4.4A expression in 111 ESCCs to elucidate its prognostic value. Chemo-radiotherapy for advanced esophageal cancer is commonly used in Western countries and neoadjuvant chemotherapy is a standard therapy before surgery in stage II and III ESCC patients in Japan.22–24 Therefore, we collected samples from previous ESCC patients (1998–2007) who did not receive preoperative radiotherapy and/or chemotherapy because our institute usually provides preoperative

chemotherapy with FAP (5-FU, Adriamycin, CDDP) or DCF (Docetaxel, CDDP, 5-FU) in current ESCC cases. The C4.4A antibody provided concordant signal intensities when used in IHC and Western blot analysis (Figs. 1b, 2b), suggesting that this antibody reacts well with the C4.4A protein on the ESCC tissue samples. Western blot analysis indicated that the C4.4A protein in carcinoma tissues was smaller than the protein found in normal squamous epithelium, which had a MW of 70 kDa. This is larger than the estimated MW of 32 kDa, possibly due to modification by N- and O-glycosylation.9 The C4.4A protein is cleaved by trypsin treatment or by ADAM10, suggesting it might be subjected to proteolytic cleavage in ESCC tissues.13,25 By IHC using the GPI-related antibody we found membranous staining was often detected in tumor cells, with the (55 of 111: 49.5 %; membrane alone: 29, membrane plus cytoplasm: 26). Survival analysis indicated that C4.4A expression was related to a poor prognosis of ESCC when it was detected by the GPI-related antibody. When the staining patterns were further analyzed, in terms of intracellular localization, the difference in 5-year survival rate was the most evident in samples with membrane staining alone; while, cytoplasmic staining had no impact on prognosis. It is therefore suggested that membranous C4.4A may play a crucial role in the unfavorable outcome of ESCC. These results are consistent with our previous report that membranous C4.4A in CRC samples was associated with a poor prognosis.14 Furthermore, we recently found that membranous C4.4A was tightly linked to tumor budding, a putative hallmark of invasion of colon cancer cells, and it was related to epithelial-mesenchymal transition (EMT) in CRC cells and tissue samples.15 Taken together, it is likely that C4.4A on the plasma membrane could be functionally important. We hypothesized that the C4.4A antibody recognizing the GPI anchor signaling sequence may have an essential role in detecting membranous C4.4A because the C4.4A protein is supposed to be linked to the plasma membrane via the GPI anchor at the C-terminus end.9 To explore this hypothesis, we generated a GPI-unrelated C4.4A antibody, C4.4A-119 Ab, that reacts with the amino acid residues 119 to 140, which are far from the GPI signaling sequence.16 As expected, C4.4A-119 Ab resulted in cytoplasmic staining for C4.4A in the majority of ESCCs (95 of 111: 85.6 %) and no tumors expressed membrane staining alone. In contrast to GPI-related antibody, no significant difference was noted in 5-year OS. These findings suggest membranous C4.4A may be of functional importance rather than cytoplasmic C4.4A and that the GPI-related antibody is essential to detect membranous C4.4A. However, we should emphasize that the C4.4A-119 Ab was able to detect the C4.4A protein in the majority of ESCCs. This

C4.4A Expression in ESCC

suggests that the C4.4A-119 Ab may be a useful diagnostic tool for ESCC. In conclusion, we found that C4.4A expression, especially the membranous type, was associated with a poor prognosis of ESCC when stained with the GPI-related antibody. Using the C4.4A-119 Ab, on the other hand, C4.4A could be a useful tumor marker for ESCC.

2705

12.

13.

14. ACKNOWLEDGMENT This work was supported by a Grant-in Aid for Cancer Research from the Ministry of Education, Science, Sports and Culture Technology, Japan (grant 21390360). DISCLOSURE

15.

The authors declare no conflict of interest. 16.

REFERENCES 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer. 2010;127:2893–917. 2. Holmes RS, Vaughan TL. Epidemiology and pathogenesis of esophageal cancer. Semin Radiat Oncol. 2007;17:2–9. 3. Taccioli C, Chen H, Jiang Y, et al. Dietary zinc deficiency fuels esophageal cancer development by inducing a distinct inflammatory signature. Oncogene. 2012;31:4550–8. 4. Matzku S, Wenzel A, Liu S, Zoller M. Antigenic differences between metastatic and nonmetastatic BSp73 rat tumor variants characterized by monoclonal antibodies. Cancer Res. 1989;49: 1294–9. 5. Claas C, Herrmann K, Matzku S, Moller P, Zoller M. Developmentally regulated expression of metastasis-associated antigens in the rat. Cell Growth Differ. 1996;7:663–78. 6. Rosel M, Claas C, Seiter S, Herlevsen M, Zoller M. Cloning and functional characterization of a new phosphatidyl-inositol anchored molecule of a metastasizing rat pancreatic tumor. Oncogene. 1998; 17:1989–2002. 7. Wurfel J, Seiter S, Stassar M, et al. Cloning of the human homologue of the metastasis-associated rat C4.4A. Gene. 2001;262: 35–41. 8. Smith BA, Kennedy WJ, Harnden P, Selby PJ, Trejdosiewicz LK, Southgate J. Identification of genes involved in human urothelial cell–matrix interactions: implications for the progression pathways of malignant urothelium. Cancer Res. 2001;61:1678–85. 9. Hansen LV, Gardsvoll H, Nielsen BS, et al. Structural analysis and tissue localization of human C4.4A: a protein homologue of the urokinase receptor. Biochem J. 2004;380(Pt 3):845–57. 10. Seiter S, Stassar M, Rappl G, Reinhold U, Tilgen W, Zoller M. Upregulation of C4.4A expression during progression of melanoma. J Invest Dermatol. 2001;116:344–7. 11. Fletcher GC, Patel S, Tyson K, et al. hAG-2 and hAG-3, human homologues of genes involved in differentiation, are associated with oestrogen receptor-positive breast tumours and interact with

17.

18.

19.

20.

21.

22.

23.

24.

25.

metastasis gene C4.4A and dystroglycan. Br J Cancer. 2003;88: 579–85. Hansen LV, Skov BG, Ploug M, Pappot H. Tumour cell expression of C4.4A, a structural homologue of the urokinase receptor, correlates with poor prognosis in non–small cell lung cancer. Lung Cancer. 2007;58:260–6. Paret C, Hildebrand D, Weitz J, et al. C4.4A as a candidate marker in the diagnosis of colorectal cancer. Br J Cancer. 2007; 97:1146–56. Konishi K, Yamamoto H, Mimori K, et al. Expression of C4.4A at the invasive front is a novel prognostic marker for disease recurrence of colorectal cancer. Cancer Sci. 2010;101:2269–77. Oshiro R, Yamamoto H, Takahashi H, et al. C4.4A is associated with tumor budding and epithelial–mesenchymal transition of colorectal cancer. Cancer Sci. 2012;103:1155–64. Yamamoto H, Oshiro R, Ohtsuka M, et al. Distinct expression of C4.4A in colorectal cancer detected by different antibodies. Int J Oncol. 2013;42:197–201. Hayashi N, Yamamoto H, Hiraoka N, et al. Differential expression of cyclooxygenase-2 (COX-2) in human bile duct epithelial cells and bile duct neoplasm. Hepatology. 2001;34(4 Pt 1): 638–50. Noura S, Yamamoto H, Ohnishi T, et al. Comparative detection of lymph node micrometastases of stage II colorectal cancer by reverse transcriptase polymerase chain reaction and immunohistochemistry. J Clin Oncol. 2002;20:4232–41. Takemasa I, Yamamoto H, Sekimoto M, et al. Overexpression of CDC25B phosphatase as a novel marker of poor prognosis of human colorectal carcinoma. Cancer Res. 2000;60:3043–50. Yamamoto H, Soh JW, Shirin H, et al. Comparative effects of overexpression of p27Kip1 and p21Cip1/Waf1 on growth and differentiation in human colon carcinoma cells. Oncogene. 1999; 18:103–15. Hansen LV, Laerum OD, Illemann M, Nielsen BS, Ploug M. Altered expression of the urokinase receptor homologue, C4.4A, in invasive areas of human esophageal squamous cell carcinoma. Int J Cancer. 2008;122:734–41. Sjoquist KM, Burmeister BH, Smithers BM, et al. Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated meta-analysis. Lancet Oncol. 2011;12:681–92. Stahl M, Walz MK, Stuschke M, et al. Phase III comparison of preoperative chemotherapy compared with chemoradiotherapy in patients with locally advanced adenocarcinoma of the esophagogastric junction. J Clin Oncol. 2009;27:851–6. Ando N, Kato H, Igaki H, et al. A randomized trial comparing postoperative adjuvant chemotherapy with cisplatin and 5-fluorouracil versus preoperative chemotherapy for localized advanced squamous cell carcinoma of the thoracic esophagus (jcog9907). Ann Surg Oncol. 2012;19:68–74. Esselens CW, Malapeira J, Colome N, Moss M, Canals F, Arribas J. Metastasis-associated C4.4A, a GPI-anchored protein cleaved by ADAM10 and ADAM17. Biol Chem. 2008;389:1075–84.