- Email: [email protected]
Prognostic value of the chemokine receptor CXCR4 and epithelial-to-mesenchymal transition in patients with squamous cell carcinoma of the mobile tongue
Oral Oncology 48 (2012) 1263–1271
Contents lists available at SciVerse ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Prognostic value of the chemokine receptor CXCR4 and epithelial-to-mesenchymal transition in patients with squamous cell carcinoma of the mobile tongue q Sébastien Albert a,b, Muriel Hourseau c, Caroline Halimi a,b, Maria Serova a,d, Véronique Descatoire c, Béatrix Barry b, Anne Couvelard c, Maria Eugenia Riveiro a, Annemilaï Tijeras-Raballand d, Armand de Gramont d, Eric Raymond a,⇑, Sandrine Faivre a a
INSERM U728, RayLab, and Departments of Medical Oncology, Beaujon University Hospital (AP-HP – Paris 7 Diderot), Clichy, France Department of Head & Neck Surgery, Bichat University Hospital, Paris, France c Department of Pathology, Bichat University Hospital, Paris, France d AAREC Filia Research, 1 place Paul Verlaine, 92100 Boulogne-Billancourt, France b
a r t i c l e
i n f o
Article history: Received 17 March 2012 Received in revised form 8 June 2012 Accepted 12 June 2012 Available online 7 July 2012 Keywords: Head and neck cancer Tongue CXCR4 SDF-1 CXCL12 CA9 HIF-1 Vimentin E-cadherin Oral cancer Squamous cell carcinoma
s u m m a r y Objective: The aim of this study was to evaluate the expression and the prognostic value of chemokine receptor 4 (CXCR4), its cognate ligand the CXCL12, and markers of epithelial-to-mesenchymal transition (EMT) in squamous cell carcinoma (SCC) of the mobile tongue. Patients and methods: Patients with primary SCC of the mobile tongue who underwent surgery in our center were screened retrospectively. Patients without prior treatment, who had pre-surgery TNM staging and available tumor samples, were eligible. Protein expression of CXCL12, CXCR4, CA9, E-cadherin, and vimentin was determined by immunohistochemical staining, scored, and correlated with clinical and pathological parameters and overall survival. Multivariate and Cox proportional hazards analyses were performed. Results: Among 160 patients treated and screened, 47 were analyzed. CXCR4 and CXCL12 expression was high in tumor cells. CXCR4 expression in primary tumor samples was significantly higher in patients with high-grade tumors, lymph node metastases, and microscopic nerve invasion (p 6 0.05). There was a nonsignificant trend towards a correlation between high CXCL12 expression and pathologic tumor stage (p = 0.07). Tumors with high CXCR4 expression correlated with poor overall survival (hazard ratio = 3.6, 95% confidence interval 1.3–9.7; p = 0.011), notably in the CXCR4high/vimentin-positive subgroup. Vimentin-positive tumors, characterizing EMT, were associated with lower survival (hazard ratio = 4.5, 95% confidence interval 1.6–12.3; p = 0.0086). Multivariate analysis confirmed vimentin (but not CXCR4) expression as an independent prognostic factor of poor overall survival (p = 0.016). Conclusion: Our results suggest that CXCR4 is a marker of tumor aggressiveness and vimentin is an important and independent prognostic factor in patients with SCC of the mobile tongue. Ó 2012 Elsevier Ltd. All rights reserved.
Introduction Incidence of head and neck squamous cell carcinoma (HNSCC) is estimated to be half a million cases worldwide.1 HNSCCs are frequently invasive in the depth of the affected tissue due to the local q Financial support: This work was supported by the Foundation Nelia & Amadeo Barleta (FNAB) and the Association d’Aide à la Recherche et à l’Enseignement en Cancérologie (AAREC). ⇑ Corresponding author. Address: Department of Medical Oncology (INSERM U728 – PRES Paris 7 Diderot), Beaujon University Hospital, Assistance Publique – Hôpitaux de Paris, 100 boulevard du Général Leclerc, 92110 Clichy, France. Tel.: +33 1 4087 5614; fax: +33 1 4087 5487. E-mail address: [email protected] (E. Raymond).
1368-8375/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.oraloncology.2012.06.010
environment of soft tissues and can ultimately extend into the lymphatic and hematologic systems, with lymph node and distant metastases being observed in over 50% of patients at diagnosis.2 HNSCC often develops in hypoxic conditions either at the primary site or lymph node metastasis. Molecular mechanisms that drive metastasis in HNSCC remain poorly understood. However, hypoxia and necrosis occurring at the primary site and/or in lymph nodes have been associated with more frequent occurrence of distant metastasis.3,4 Chemokine receptors are 7-transmembrane-domain proteins which interact with specific cytokines known as chemokines involved in cellular chemotaxis.5 Chemokine receptor activation leads to cytoskeletal rearrangement and directional cell migration toward
1264
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
the chemokine source.6 Chemokines play a role in the progression of a large number of human malignancies by stimulating proliferation, invasion, and metastasis, and attract cancer cells toward lymphatic tissues, the liver, lung, and bone marrow, all of which are primary sources of chemokines.7–9 CXCR4 is a chemokine receptor specific for CXCL12 (also known as SDF-1, stromal-derived-factor-1) involved in chemotaxis of lymphocytes to inflammatory sites which in turn can promote chemotaxis of cancer cells.10 It has been shown that the CXCL12/CXCR4 interaction plays a critical role in carcinogenesis by activating angiogenesis, invasion, and cancer cell migration in various carcinomas.7 CXCR4 was shown to play a role in the development of metastases by regulating the directional migration of cancer cells expressing the receptor CXCR4 to target tissues that release CXCL12.11 Furthermore, hypoxic conditions activate expression of hypoxia-inducible factor-1 (HIF-1a), a transcription factor known to trigger CXCR4 expression in cancer cells.12 Although the chemokine receptor CXCR4 and its ligand, CXCL12, are known to mediate invasiveness and the process of metastasis in a number of solid cancers,13 their role in HNSCC remains incompletely elucidated. Epithelial-to-mesenchymal transition (EMT) has been recognized as one of several cellular processes facilitating the dedifferentiation of potentially cancerous epithelial cells into mesenchymal cells, thus acquiring migration and invasive capacities.14 EMT is primarily associated with both the loss of cellular adhesion, recognized by repression of E-cadherin expression, and also with changes in the cell cytoskeleton which can be identified by vimentin expression.15 EMT has previously been shown to be a key factor in the development of metastases for several tumor types including HNSCC.16,17 The aim of this study was to evaluate the expression of CXCR4 and CXCL12, as well as of the hypoxia markers carbonic anhydrase 9 (CA9) and EMT markers (E-cadherin and vimentin) in an homogeneous subgroup of HNSCC patients with resected oral squamous cell carcinoma (SCC) of the mobile tongue, and to assess correlations between these biomarkers and clinical tumor staging and prognosis. Patients and methods Patients Tissue samples were collected retrospectively from patients with SCC of the tongue treated at the University center Paris Nord Val de Seine (Bichat and Beaujon University Hospitals) between 2005 and 2008. All patients underwent surgery at the Department of Head and Neck Surgery at Bichat University Hospital (Paris, France). To be eligible for the study, patients had to provide written consent for the use of their tumor samples, have received no prior therapy, with the primary tumor site located in the mobile tongue,
have full TNM staging available prior to surgery, adequate followup in our institution including a record of date of death when appropriate, and pathological specimens for CXCR4, CXCL12, carbonic anhydrase 9 (CA9), E-cadherin and vimentin immunostaining. Patients were excluded if they had any history of other alcohol or tobacco-related tumors including other HNSCC, lung, esophageal or bladder carcinomas. Tumor staging was performed according to the American Joint Committee on Cancer (AJCC) TNM classification. Clinical data were recorded in an anonymized database. All patients provided informed consent and the study was performed in accordance with the Declaration of Helsinki. Histology A representative tissue block containing tumor and non-neoplastic tissues was selected for each patient for immunohistochemical analysis. Immunohistochemistry Tissues were fixed in 4% formalin and embedded in paraffin. Paraffin blocks of 4-lm thickness were cut, and sections were deparaffinized with xylene and rehydrated in graded alcohol. Immunohistochemistry was performed using an automated stainer, and according to the manufacturer’s guidelines (incubation 30 min at room temperature, after antigen retrieval using citrate buffer pH = 7.7, revelation with streptavidin-peroxidase kit and DAB staining) (Leica Bond™, Wetzlar, Germany). Phosphate-buffered saline was used as a negative control. The section were incubated with an anti-human CXCR4 antibody (1:100 dilution; fusin/ sc-53534, Santa Cruz Biotechnology, California, USA), an anti-human CXCL12 antibody (1:1000 dilution; 5393-100, BioVision, California, USA), an anti-human CA9 antibody (1:800 dilution; NB 100417, Novus Biological, Littleton, California, USA), an anti-human Ecadherin antibody (1:100 dilution; NCH-38, Dako, Glostrup, Denmark) and an anti-human vimentin antibody (1:400 dilution; M0725, Dako, Glostrup, Denmark) were used. The sections were then incubated with a secondary antibody according to the kit manufacture’s instructions (Leica): post primary rabbit anti-mouse IgG (<10%g/ml) in 10%(V/V) animal serum in tris-buffered saline/ 0.09% (Proclin 950, Sigma–Aldrich, Saint Louis, USA). Immunohistochemical evaluation All sections were evaluated independently by two investigators who were blinded to the clinical data. We made the choice to not use semi-quantitative method because, in our experience, the image system analyzed all the tissue without distinguishing tumor
Figure 1a Representative CXCR4 and CXCL12 immunostaining used for scoring SCC of the mobile tongue (100). The intensity of immunostaining was classified as one of four levels: (a)1 negative, (a)2 weak, (a)3 moderate, (a)4 strong.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
1265
Figure 1b Representative CXCR4 and CXCL12 immunostaining (100). Location of CXCR4 and CXCL12 staining within a tumor cell: (b)1 on the membrane (b)2 in the cytoplasm and (b)3 in the nucleus.
and normal tissue in resected tumors, contrary to the evaluation made by trained pathologists. For the CXCR4, CXCL12 and CA9 antibodies, staining intensity (including membrane, cytoplasm and nucleus localization) was scored as 0 = no detectable signal, 1 = weak staining, 2 = moderate staining, 3 = strong staining (Fig. 1a). A score was calculated according to the percentage of cancer cells with an immunostaining intensity of 0, 1, 2, and 3 using the following formula: ‘‘x score = (percentage of cancer cells with an immunostaining intensity of 0) x0 + (percentage of cancer cells with an immunostaining intensity of 1) x1 + (percentage of cancer cells with an immunostaining intensity of 2) x2 + (percentage of cancer cells with an immunostaining intensity of 3) x3’’. When the two reviewers’ scores were discordant, a joint review was performed to obtain a consensus. For each specimen, the staining location (membrane, cytoplasm or nucleus) was recorded. The overall median and mean scores were determined and patients were then allocated to one of two groups according to whether the CXCR4, CXCL12 and CA9 score; samples with a score lower than the median were classified as low (CXCR4low, CXCL12low and CA9low) and samples with a score higher than the median were classified as high (CXCR4high, CXCL12high and CA9high). Scoring the immunohistochemical evaluation of E-cadherin and vimentin was not technically feasible, notably for vimentin which was frequently not expressed and when detectable was primarily located at the rim of the tumor. Statistics The correlation of CXCR4 staining intensity with clinicopathological patterns was assessed with Chi-square test, or Fisher’s exact Table 1 Tumor staging at diagnosis according to the American Joint Committee on Cancer TNM system for carcinoma of oral cavity.
N0 N1 N2a N2b N2c N3 Total
T1
T2
T3
T4
Total
14 0 0 0 0 0 14
7 2 1 2 1 1 14
2 3 2 6 2 1 16
0 1 0 1 1 0 3
23 6 3 9 4 2 47
T: size or direct extent of the primary tumor from T1 to T4; N: degree of spread to regional lymph nodes from N0 to N3.
test if values were lower than five. Survival rates were determined using the Kaplan–Meier method, and statistical analysis was carried out using the log-rank test. For multivariate analyses, independent prognostic factors were determined using a Cox proportional hazards model. p-Values <0.05 were considered significant in all statistical analyses. Results Patient characteristics Among 160 patients treated for HNSCC of the oral cavity during the study period, 47 were evaluable for analysis. The other 113 patients were excluded due to lack of informed consent, tumor localization other than the mobile tongue, prior history of HNSCC or other tumors, lost to follow-up, or a therapeutic decision other than surgery as first-line treatment. Included patients had a median age at diagnosis of 61 years (range 34–92), with a 4.3–1 ratio of males to females. Thirty-six patients had a tumor confined to the mobile tongue while for 11 patients the tumor extended to the floor of the oral cavity. Seventy-five percent of patients were current or ex-tobacco smokers, 62% were current or ex-alcohol drinkers and 61% were both. Clinical stage was defined according to the classification of malignant tumors (AJCC-TNM) in 47 patients including 14 (30%) with stage I, 7 (15%) with stage II, 7 (15%) with stage III and 19 (40%) with stage IV disease (details of TNM are shown in Table 1). Pathological evaluation of the tumor resection showed that 67% of primary tumors were well differentiated; nerve and vascular invasion were observed in 55% and 46% of cases, respectively. Cancerous cells were identified in surgical resection margins of the primary tumor in 19% of cases (R1). Tumor lymph node involvement was observed in 49% of cases, 79% of which had extra-capsular invasion. Median follow-up was 48 months (range 22–75 months) and the 5-year overall survival rate was 56%. Eighteen patients died during the duration of follow-up, which included 13 patients who died of local recurrences, 3 patients who died of distant metastases, and 2 patients who died of other diseases. CXCR4, CXCL12, and CA9 immunostaining CXCR4, CXCL12 and CA9 were frequently highly expressed in tumor cells (87%, 95% and 83% of tumors, respectively) compared
1266
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
Figure 2 Comparison of CXCR4 immunostaining in normal and tumor tissues (25). (a) CXCR4 immunostaining in normal tissues of the tongue: 1: negative control of squamous epithelium, 2: positive control of muscle tissue. (b) CXCR4 immunostaining in SCC of the tongue showing strong CXCR4 expression.
Table 2 Median and mean scores of CXCR4, SDF-1 and CA9 immunostaining.
CXCR4 SDF-1 CA9
Median score [range] (n = 47)
Mean score ± SD (n = 47)
Patients with low IHC scorea Mean ± SD
Patients with high IHC scoreb Mean ± SD
120 [20; 250] 110 [20; 230] 110 [10; 280]
111.0 ± 54.9 111.6 ± 52.1 116.3 ± 65.3
67.9 ± 31.6 70.4 ± 24.4 67.1 ± 36.8
156.1 ± 33.1 152.8 ± 37.6 161.3 ± 51.9
IHCS: immunohistochemical score. a Low score = score less than the median. b High score = score greater than the median.
to normal epithelial cells adjacent to the tumor tissues (Fig. 2a and b). The median immunohistochemical scores for CXCR4, CXCL12 and CA9 were 120, 110, and 110 respectively (Table 2). Staining for CXCR4 was predominantly located in the cytoplasm (98% of cases) and the nucleus (98% of cases) of tumor cells and on the membrane in 55% of cases. CXCL12 was mainly expressed in cytoplasm (93% of cases) and also on the membrane (59% of cases) and in the nucleus (54% of cases) of tumor cells (Fig. 1b). Lymphocytes infiltrating the normal epithelium and muscle cells were shown to have high CXCR4 expression (Fig. 2a and b). CA9 immunostaining was mainly observed in the cytoplasm and membrane but never in the nucleus of tumor cells. In 17 patients with paired primary and lymph node tumor biopsies, we found no difference in the level of CXCR4 and CXCL12 expression between primary and lymph node tumor cells. Correlation of CXCR4, CXCL12 and CA9 expression with clinical and pathological features An analysis of CXCR4, CXCL12 and CA9 expression according to primary tumor stage is shown in Fig. 3, and statistical analyses of the correlation between expression of these markers and various histopathological and clinical parameters is shown in Table 3. Expression of CXCR4 in primary tumor samples was significantly higher in patients with advanced cancers (stages III and IV; p = 0.005), and correlated with both tumor size (p = 0.002) and nerve invasion (p = 0.05). CXCR4 expression was also significantly higher in tumors with cervical lymph node metastasis than those without (p = 0.01). In addition, high CXCR4 expression in the primary tumor was reported in 67% of patients with lymph node metastasis vs. 29% of patients without. However, the degree of differentiation and vascular invasion were not associated with CXCR4 expression levels, nor gender or smoking history, although patients with a history of alcohol consumption showed borderline significance (p = 0.047).
CXCL12 expression was not correlated with any of the clinical and pathological variables evaluated, although there was a trend towards a correlation between higher pathological stage and stronger CXCL12 expression (p = 0.072). In primary tumor samples, CA9 expression was significantly higher in patients with advanced stage III and IV diseases (p = 0.037) and was higher in primary tumors with cervical lymph node metastasis than in those without (p = 0.019). When CXCR4 and CA9 were both over-expressed, the positive predictive value for patients to have lymph node metastasis was 91%. In terms of histopathology parameters, CA9 expression was higher in poorly differentiated tumors (p = 0.01) and those with vascular invasion (p = 0.018). No significant inter-marker correlation was observed between CA9 and CXCR4/CXCL12 immunostaining scores. No survival difference was observed between patients with CXCL12low and CXCL12high values (data not shown). Conversely, high expression of CXCR4 was highly correlated with poor prognosis. Overall survival was statistically higher in patients with CXCR4low compared to those with CXCR4high expressions (Hazard ratio: 3.6, 95% confidence interval: 1.3–9.7, p-value = 0.011; Fig. 4).
CXCR4 expression in tumors undergoing EMT In primary tumor samples, expression of E-cadherin and vimentin was observed in 36 of 43 patients (83.7%) and 14 of 43 patients (32.5%), respectively (samples were not evaluable for 4 patients). Among the 14 vimentin-positive tumors, 7 (50%) also expressed E-cadherin (Fig. 5A upper panel). The 14 vimentin-positive tumors, characterizing EMT, were associated with a poorer overall survival outcome (Fig. 5A lower panel) compared to vimentin-negative tumors (hazard ratio: 4.5; 95% confidence interval: 1.6–12.3, p = 0.0086). No significant correlations between vimentin and expression of CXCR4 (Fig. 5B upper panel) or CXCL12 (data not shown) were found. Interestingly, patients whose tumors were
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
1267
Figure 3 Correlation between CXCR4/CXCL12 and tumor stage. Expression of (a1) CXCR4 and (a2) CXCL12, according to clinical tumor stage (two groups: stage I–II, stage III–IV); (a3) mean CXCR4/CXCL12 scores for the two groups. Expression of (b1) CXCR4 and (b2) CXCL12, according to histopathological pT staging (two groups: pT1–2, pT3–4); (b3) mean CXCR4/CXCL12 scores for the two groups. The y-axis of the 4 upper panel figures (a1, a2, b1 and b2) represents the number of patients. The Y-axis of the 2 lower panel graphs (a3 and b3) represents the mean value, standard deviation was then determined.
vimentin-negative and CXCR4low had significantly higher survival compared to other combinations (Fig. 5B lower panel, p = 0.0043). Univariate and multivariate analyses of prognostic factors A univariate analysis showed a significant correlation for histopathological lymph node invasion (p < 0.001), high histopathological tumor stage (pT3 or pT4; p = 0.011), histopathological nerve invasion (p = 0.04), CXCR4 expression (p = 0.046), and vimentin expression (p = 0.0034) with a poor survival prognosis. However, no correlation was observed for the degree of tumor differentiation (p = 0.37) or histopathological vascular invasion (p = 0.7). In a multivariate analysis, vimentin (p = 0.016) but not CXCR4 (p = 0.24) expression was shown to be an independent marker of poor prognosis. Discussion HNSCCs are frequently locally invasive at diagnosis and many develop lymph node invasion followed by distant metastasis.18 Lymph node invasion and metastases are often responsible for
relapse after complete surgery or high dose local radiotherapy.19 Low oxygen tissue pressure, typically less than 5 mm Hg, has been widely reported in both primary HNSCC tumors and lymph node metastases.20 Intratumoral hypoxia has been recognized as a major factor that contributes to angiogenesis, invasiveness, metastasis, alteration of tumor metabolism, and genomic instability, and is also suspected to play a role in resistance to chemotherapy and radiotherapy.21 In this study, we investigated tumor hypoxia using carbonic anhydrase-9 (CA9). CA9 is a transmembrane protein which is involved in the catalytic hydration of carbon dioxide to carbonic acid. Transcription of this gene is known to be regulated by the Von Hippel–Lindau tumor suppressor gene, responsible for targeting HIF-1a for oxygen-dependent proteolysis.22 When low levels of oxygen occur in tumors, HIF-1a over-expression is induced as well as the expression of CA9.23,24 Recent studies showed that CA9 was overexpressed in a wide variety of tumor types and was induced by hypoxia.22 In addition, CA9 is a stable marker of hypoxia in opposition to HIF-1, which is very labile and difficult to use in IHC, explaining our choice for CA9 in agreement to many other authors. We have shown that CXCR4, CXCL12 and CA9 were commonly expressed, and at different levels, in patients with SCC of the mobile
1268
Table 3 Correlations between clinicopathological features and immunostaining results by patient and staining intensity. SDF-1
CXCR4
CA-9
E-cadherin
Vimentin
SDF1high
P value
CXCR4low
CXCR4high
P value
CA9low
CA9high
P value
E-cadh. negative
E-cadh. positive
P value
Vimentin negative
Vimentin positive
P value
21 2
16 7
NS (.13)
19 5
19 4
NS (.94)
18 5
19 4
NS (1)
7 0
29 8
NS (.17)
23 7
13 1
NS (.19)
16 7
18 5
NS (.73)
17 7
18 5
NS (.8)
18 5
17 6
NS (1)
6 1
27 10
NS (.48)
23 7
11 3
NS (.88)
16 7
12 11
NS (.36)
11 13
18 5
.047
18 5
11 12
NS (.54)
6 1
22 15
NS (.19)
16 14
12 2
0.038
16 7
17 6
NS (1)
19 5
15 8
NS (.45)
18 5
15 8
NS (.51)
5 2
26 11
NS (.95)
22 8
9 5
NS (.54)
12
9
NS (.65)
16
5
.005
14
6
.037
2
17
NS (.39)
15
3
NS (.073)
Stage III–IV
12
14
8
18
9
17
5
20
15
11
Histopathology Tumor staging pT1 or pT2
17
10
20
8
16
11
4
22
20
5
pT3 or pT4 6 Lymph node invasion Negative 10
13
4
15
7
12
3
15
10
9
7
16
7
16
3
19
16
5
Positive Differentiation Well
13
10
17
7
16
7
4
18
14
9
18
13
19
13
20
11
6
23
17
12
5
10
5
10
3
12
1
14
13
2
13 10
10 9
NS (1)
10 14
17 6
.05
11 12
16 7
NS (.23)
4 3
23 14
NS (0.8)
17 13
9 5
NS (.63)
9 14
12 11
NS (.55)
10 14
11 12
NS (.9)
6 17
15 8
.018
2 5
18 19
NS (0.33)
15 15
5 9
NS (.38)
Clinical parameter Gender Male Female Smoker Yes No Alcohol Yes No Tumor description B or UB U Stage Stage I–II
Moderate/ poor Nerve invasion Present Absent Vasc. invasion Present Absent
13
NS (.072)
NS (.56)
NS (.2)
.002
.01
NS (.18)
NS (.23)
.019
.01
NS (0.91)
NS (0.68)
NS (0.23)
NS (.053)
NS (.28)
NS (.058)
B: budding; UB: ulcerating and budding; U: ulcerating; SDF-1low: number of patients with an IHC score lesser than 110 (median value of IHC SDF-1 scores); SDF-1high: number of patients with an IHC score higher than the median 110; CXCR4low: number of patients with an IHC score lesser than 120 (corresponding to the median value of IHC CXCR4 scores); CXCR4high: number of patients with an IHC score higher than 120; CA-9low: number of patients with an IHC score lesser than 110 (median value of IHC CA-9 scores); CA-9high: number of patients with an IHC score higher than 110; Vasc.: vascular; NS not significant.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
SDF1low
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
A
Percent survival
100 75 50 25 0 0
10
20
30
40
50
60
70
80
Months 100
Percent survival
B
75 50 25
p-value: 0.011 Hazard ratio: 3.6 (95%CI: 1.33-9.72)
0 0
10
20
30
40
50
60
70
80
Months CXCR4high (Median: 41 months) CXCR4low (Median: not reach) Figure 4 Overall survival of patients with resected SCC of the mobile tongue (A) and prognostic value of CXCR4 expression (B). Patients were grouped as CXCR4low and CXCR4high according to their score values as lower or higher than the median value of CXCR4 score in the entire population respectively.
tongue. Activation of HIF-1 under hypoxic conditions may result in the transcription of several growth factors and receptors associated with tumor cell migration and angiogenesis, such as VEGF, VEGFR, CXCL12,25 and CXCR4.12,26–29 Expression of CXCR4 has been associated with a poor outcome in head and neck carcinomas.4,30–34 Ishikawa et al. also demonstrated that HIF-1a regulates hypoxia-induced CXCR4 expression in oral HNSCC.35 Tumor cells expressing a high level of CXCR4 in an hypoxic environment had enhanced motility and invasive capacity to enter the blood and the lymphatic microcirculation. In our study, over-expression of CXCR4 or CA9 was observed in patients with metastatic lymph nodes. When CXCR4 and CA9 were both over-expressed, the positive predictive value for patients to have lymph node metastasis was 91%. These findings suggest that the over-expression of CXCR4 in hypoxic conditions may correlate with a higher risk for developing lymph node metastasis. CXCR4 and CXCL12 may play a key role in invasion and migration processes in HNSCC via EMT and activation of metalloproteases. Ou et al.36 have studied the role of TWIST a transcription factor involved in EMT. The authors suggest that TWIST regulates CXCR4 expression in SCCs, which in turn might be associated with lymph node metastasis.36 In addition, Taki et al.37 demonstrated that CXCR4 was up-regulated by CXCL12 and TGF-b1 in oral SCCs, enhancing migratory activity. They also showed that SNAIL, another transcription factor promoting EMT, induced over-expression of CXCL12 giving a mesenchymal phenotype to the epithelial tumor cell. Onoue et al.38 reported in vitro that EMT markers (vimentin and snail) were induced by CXCL12/CXCR4 in oral SCC cells through the PI3K-Akt pathway. These results suggest that up-regulation of
1269
CXCR4 may be a consequence of EMT and that EMT is a consequence of CXCR4 activation. In our study, tumors expressing vimentin (33% of patients) were associated with high level of CXCR4 in 71% of cases. Furthermore, recent studies suggested that CXCL12/CXCR4 enhanced invasion and migration mechanisms in HNSCC through the regulation of MMP-9 and MMP-13.39,40 Therefore, EMT and matrix metalloproteinase activation might promote invasion and motility of tumor cells and metastatic processes in HNSCC.41–43 In our study, the size of the primary tumor and its degree of local invasiveness (as reflected by the T-stage), pathological nerve invasion and lymph node metastasis both of which represent clinical aggressiveness, were strongly correlated with high CXCR4 expression. We also expected to observe correlations between CXCR4 expressions and vascular invasion in tumors, but no statistical correlation was detectable in this study. This can be partly explained by the fact that vascular invasion is often found in histopathological analysis in HNSCC. Considering the number of patients in this study, the power to detect statistical correlation is likely to be too low to reach conclusion in this study. However, our results are consistent with recent studies showing that CXCR4 activation plays a major role in lymph node metastasis in head and neck squamous cell carcinoma.35,44–48 CXCR4 may be sensitive to the chemo-attractive action of CXCL12 located in distant tissues, which would then be homing targets for metastasizing tumor cells expressing high CXCR4 levels. Consistent with this hypothesis, relatively high levels of CXCL12 are found in the lung and the lymph nodes of patients with HNSCC.11,49–51 To date, very few groups have investigated CXCR4 as a prognostic factor for survival in patients with HNSCC. The median follow-up in our study was 48 months, which despite being somehow limited could be considered sufficient to use overall survival as a primary endpoint using Kaplan–Meier method, as most recurrence are likely to occur within this time frame. In a previous study, the 5-year survival rate in patients with CXCR4-negative tumors was significantly higher (71.4%) than those with CXCR4-positive tumors (39.1%).52 In another trial, CXCR4 was found to be an independent prognostic factor for poor survival in patients with HNSCC.53 These data are consistent with our results which also showed that high CXCR4 expression in tumor cells was associated with poorer survival. Interestingly, both vimentin and CXCR4 expression were associated with poor prognosis in our study, patients with the better outcome being those with tumors which were vimentin-negative with low CXCR4 expression. However, vimentin appears to be a stronger marker than CXCR4 to predict poor outcome as demonstrated by multivariate analysis in our patient population. Several studies have evaluated CXCL12 expression by immunohistochemistry in biopsies from HNSCC patients and various associated clinicopathological factors. Uchida et al.52 described that the 5-year survival rate was 25.0% for the CXCL12-positive group and 71.4% for the CXCL12-negative group in patients with oral SCC. The results obtained by Almolfi et al.45 found no clear correlation between the expression of CXCL12 and other prognostic factors in oral HNSCC patients. In contrast, a recent study demonstrated that low CXCL12 level correlated with occurrence of metastases and poor prognosis.54 In this latter trial, the authors argued that a low level of CXCL12 in the area surrounding tumor cells could facilitate tumor cell detachment from the primary tumor and migration to metastatic sites. However, CXCL12 levels were assayed by quantitative RT-PCR, which reflects transcription but may not correlate with protein expression. Discrepancies between studies may be explained by regional differences of Caucasian population or Asiatic population. In our study most patients had alcohol and tobacco habits. The influence of HPV infection on HNSCC may account for these differences but would require larger cohorts to reach conclu-
1270
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
A
B
Figure 5 Correlations and prognostic value of epithelial-to-mesenchymal transition (EMT) and CXCR4 expression in tumors of patients with SCC of the mobile tongue. (A) Number of patients with E-cadherin and/or vimentin expression (upper panel), and overall survival according to EMT status (lower panel); (B) Number of patients with EMT and/or CXCR4 expression (upper panel), and overall survival according to CXCR4 and EMT status (lower panel).
sions. Furthermore, discrepancies between studies may also be attributed to the fact that most include a global analysis without considering different sub-locations of the head and neck carcinoma (oral cavity, larynx, oropharynx, hypopharynx). The anatomical location is an essential consideration in clinical practice influencing the therapeutic strategy and prognostic. For this reason we chose to restrict our study to patients with SCC of the mobile tongue thus evaluating a homogeneous cohort and avoiding misinterpretation of results. CXCR4 was homogeneously detected in the cytoplasm and nucleus and at the cell membrane of tumor cells. Previous immunohistochemical studies also demonstrated that CXCR4 was primarily located in the cytoplasm49,55 but was also found in the nucleus in head and neck carcinomas,56 as well as in hepatocellular carcinomas,57 in non-small-cell lung cancers,58 and in nasopharyngeal carcinomas.59 It is possible that when translocated into the nucleus, CXCR4 acquires different functions and may regulate transcription of genes involved in cellular invasion, migration and EMT. This hypothesis was supported by Wang et al.60 who showed that nuclear localization of CXCR4 might result in a migratory or metastatic phenotype in renal carcinoma cells in vitro. Further studies are necessary to better understand the mechanism of action of CXCR4 as a transcriptional regulator. In summary, our results indicate that CXCR4 and vimentin could be used as potential prognostic factors of aggressiveness in completely resected SCC of the mobile tongue. Conflict of interest statement None declared.
Acknowledgements We thank Jessica Sabatier, Christophe Assailly and Gilles Arcangeli for their skilful technical assistance and Fabien Reyal for statistical analysis. References 1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009;59:225–49. 2. Gourin CG, Podolsky RH. Racial disparities in patients with head and neck squamous cell carcinoma. Laryngoscope 2006;116:1093–106. 3. Brizel DM, Schroeder T, Scher RL, et al. Elevated tumor lactate concentrations predict for an increased risk of metastases in head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:349–53. 4. Brizel DM, Sibley GS, Prosnitz LR, Scher RL, Dewhirst MW. Tumor hypoxia adversely affects the prognosis of carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 1997;38:285–9. 5. Laing KJ, Secombes CJ. Chemokines. Dev Comp Immunol 2004;28:443–60. 6. Takabayashi T, Takahashi N, Okamoto M, Yagi H, Sato M, Fujieda S. Lipopolysaccharides increase the amount of CXCR4, and modulate the morphology and invasive activity of oral cancer cells in a CXCL12-dependent manner. Oral Oncol 2009;45:968–73. 7. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer 2004;4:540–50. 8. Schimanski CC, Bahre R, Gockel I, et al. Dissemination of hepatocellular carcinoma is mediated via chemokine receptor CXCR4. Br J Cancer 2006;95:210–7. 9. Sun X, Cheng G, Hao M, et al. CXCL12/CXCR4/CXCR7 chemokine axis and cancer progression. Cancer Metastasis Rev 2010;29:709–22. 10. Bleul CC, Farzan M, Choe H, et al. The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry. Nature 1996;382:829–33. 11. Muller A, Homey B, Soto H, et al. Involvement of chemokine receptors in breast cancer metastasis. Nature 2001;410:50–6. 12. Staller P, Sulitkova J, Lisztwan J, Moch H, Oakeley EJ, Krek W. Chemokine receptor CXCR4 downregulated by von Hippel–Lindau tumour suppressor pVHL. Nature 2003;425:307–11. 13. Balkwill F. The significance of cancer cell expression of the chemokine receptor CXCR4. Semin Cancer Biol 2004;14:171–9.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271 14. Sabbah M, Emami S, Redeuilh G, et al. Molecular signature and therapeutic perspective of the epithelial-to-mesenchymal transitions in epithelial cancers. Drug Resist Updat 2008;11:123–51. 15. Satelli A, Li S. Vimentin in cancer and its potential as a molecular target for cancer therapy. Cell Mol Life Sci 2011;68:3033–46. 16. Ghoul A, Serova M, Astorgues-Xerri L, et al. Epithelial-to-mesenchymal transition and resistance to ingenol 3-angelate, a novel protein kinase C modulator, in colon cancer cells. Cancer Res 2009;69:4260–9. 17. Nijkamp MM, Span PN, Hoogsteen IJ, van der Kogel AJ, Kaanders JH, Bussink J. Expression of E-cadherin and vimentin correlates with metastasis formation in head and neck squamous cell carcinoma patients. Radiother Oncol 2011;99:344–8. 18. Coca-Pelaz A, Rodrigo JP, Suarez C. Clinicopathologic analysis and predictive factors for distant metastases in patients with head and neck squamous cell carcinomas. Head Neck 2012;34:771–5. 19. Hauswald H, Simon C, Hecht S, Debus J, Lindel K. Long-term outcome and patterns of failure in patients with advanced head and neck cancer. Radiat Oncol 2011;6:70. 20. Lartigau E, Le Ridant AM, Lambin P, et al. Oxygenation of head and neck tumors. Cancer 1993;71:2319–25. 21. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer 2011;11:393–410. 22. Ivanov SV, Kuzmin I, Wei MH, et al. Down-regulation of transmembrane carbonic anhydrases in renal cell carcinoma cell lines by wild-type von Hippel– Lindau transgenes. Proc Natl Acad Sci USA 1998;95:12596–601. 23. Beasley NJ, Wykoff CC, Watson PH, et al. Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 2001;61:5262–7. 24. Wykoff CC, Beasley NJ, Watson PH, et al. Hypoxia-inducible expression of tumor-associated carbonic anhydrases. Cancer Res 2000;60:7075–83. 25. Ceradini DJ, Kulkarni AR, Callaghan MJ, et al. Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nat Med 2004;10:858–64. 26. Schioppa T, Uranchimeg B, Saccani A, et al. Regulation of the chemokine receptor CXCR4 by hypoxia. J Exp Med 2003;198:1391–402. 27. Zagzag D, Krishnamachary B, Yee H, et al. Stromal cell-derived factor-1alpha and CXCR4 expression in hemangioblastoma and clear cell-renal cell carcinoma: von Hippel–Lindau loss-of-function induces expression of a ligand and its receptor. Cancer Res 2005;65:6178–88. 28. Zagzag D, Lukyanov Y, Lan L, et al. Hypoxia-inducible factor 1 and VEGF upregulate CXCR4 in glioblastoma: implications for angiogenesis and glioma cell invasion. Lab Invest 2006;86:1221–32. 29. Carver BS, Tran J, Gopalan A, et al. Aberrant ERG expression cooperates with loss of PTEN to promote cancer progression in the prostate. Nat Genet 2009;41:619–24. 30. Tilakaratne WM, Nissanka-Jayasuriya EH. Value of HIF-1alpha as an independent prognostic indicator in oral squamous cell carcinoma. Expert Rev Mol Diagn 2011;11:145–7. 31. Eckert AW, Schutze A, Lautner MH, Taubert H, Schubert J, Bilkenroth U. HIF1alpha is a prognostic marker in oral squamous cell carcinomas. Int J Biol Markers 2010;25:87–92. 32. Hoogsteen IJ, Marres HA, Bussink J, van der Kogel AJ, Kaanders JH. Tumor microenvironment in head and neck squamous cell carcinomas: predictive value and clinical relevance of hypoxic markers. A review. Head Neck 2007;29:591–604. 33. Harris AL. Hypoxia–a key regulatory factor in tumour growth. Nat Rev Cancer 2002;2:38–47. 34. Le QT, Denko NC, Giaccia AJ. Hypoxic gene expression and metastasis. Cancer Metastasis Rev 2004;23:293–310. 35. Ishikawa T, Nakashiro K, Hara S, et al. CXCR4 expression is associated with lymph-node metastasis of oral squamous cell carcinoma. Int J Oncol 2006;28:61–6. 36. Ou DL, Chien HF, Chen CL, Lin TC, Lin LI. Role of twist in head and neck carcinoma with lymph node metastasis. Anticancer Res 2008;28:1355–9. 37. Taki M, Higashikawa K, Yoneda S, et al. Up-regulation of stromal cell-derived factor-1alpha and its receptor CXCR4 expression accompanied with epithelial– mesenchymal transition in human oral squamous cell carcinoma. Oncol Rep 2008;19:993–8.
1271
38. Onoue T, Uchida D, Begum NM, Tomizuka Y, Yoshida H, Sato M. Epithelial– mesenchymal transition induced by the stromal cell-derived factor-1/CXCR4 system in oral squamous cell carcinoma cells. Int J Oncol 2006;29:1133–8. 39. Yu T, Wu Y, Helman JI, Wen Y, Wang C, Li L. CXCR4 promotes oral squamous cell carcinoma migration and invasion through inducing expression of MMP-9 and MMP-13 via the ERK signaling pathway. Mol Cancer Res 2012;9:161–72. 40. Tan CT, Chu CY, Lu YC, et al. CXCL12/CXCR4 promotes laryngeal and hypopharyngeal squamous cell carcinoma metastasis through MMP-13dependent invasion via the ERK1/2/AP-1 pathway. Carcinogenesis 2008;29:1519–27. 41. Hartmann TN, Burger JA, Glodek A, Fujii N, Burger M. CXCR4 chemokine receptor and integrin signaling co-operate in mediating adhesion and chemoresistance in small cell lung cancer (SCLC) cells. Oncogene 2005;24:4462–71. 42. Janowska-Wieczorek A, Marquez LA, Dobrowsky A, Ratajczak MZ, Cabuhat ML. Differential MMP and TIMP production by human marrow and peripheral blood CD34(+) cells in response to chemokines. Exp Hematol 2000;28:1274–85. 43. Samara GJ, Lawrence DM, Chiarelli CJ, et al. CXCR4-mediated adhesion and MMP-9 secretion in head and neck squamous cell carcinoma. Cancer Lett 2004;214:231–41. 44. Ueda M, Shimada T, Goto Y, et al. Expression of CC-chemokine receptor 7 (CCR7) and CXC-chemokine receptor 4 (CXCR4) in head and neck squamous cell carcinoma. Auris Nasus Larynx 2010;37:488–95. 45. Almofti A, Uchida D, Begum NM, et al. The clinicopathological significance of the expression of CXCR4 protein in oral squamous cell carcinoma. Int J Oncol 2004;25:65–71. 46. Zheng C, Ge R, Ge M, Sun X. The expression and relationship of CXCR4 and microvessel density in laryngeal squamous cell carcinoma. Lin Chung Er Bi Yan Hou Tou Jing Wai Ke Za Zhi 2011;25:123–6. 47. Oliveira-Neto HH, Silva ET, Leles CR, et al. Involvement of CXCL12 and CXCR4 in lymph node metastases and development of oral squamous cell carcinomas. Tumour Biol 2008;29:262–71. 48. Yin D, Gao Z. Expression and clinical significance of chemokine receptor 4 in oral squamous cell carcinoma. Hua Xi Kou Qiang Yi Xue Za Zhi 2007;25:392–5. 49. Meng X, Wuyi L, Yuhong X, Xinming C. Expression of CXCR4 in oral squamous cell carcinoma: correlations with clinicopathology and pivotal role of proliferation. J Oral Pathol Med 2010;39:63–8. 50. Sun YX, Wang J, Shelburne CE, et al. Expression of CXCR4 and CXCL12 (SDF-1) in human prostate cancers (PCa) in vivo. J Cell Biochem 2003;89:462–73. 51. Phillips RJ, Burdick MD, Lutz M, Belperio JA, Keane MP, Strieter RM. The stromal derived factor-1/CXCL12-CXC chemokine receptor 4 biological axis in nonsmall cell lung cancer metastases. Am J Respir Crit Care Med 2003;167:1676–86. 52. Uchida D, Onoue T, Tomizuka Y, et al. Involvement of an autocrine stromal cell derived factor-1/CXCR4 system on the distant metastasis of human oral squamous cell carcinoma. Mol Cancer Res 2007;5:685–94. 53. Lee JI, Jin BH, Kim MA, Yoon HJ, Hong SP, Hong SD. Prognostic significance of CXCR-4 expression in oral squamous cell carcinoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;107:678–84. 54. Clatot F, Picquenot JM, Choussy O, et al. Intratumoural level of SDF-1 correlates with survival in head and neck squamous cell carcinoma. Oral Oncol 2011;47:1062–8. 55. Delilbasi CB, Okura M, Iida S, Kogo M. Investigation of CXCR4 in squamous cell carcinoma of the tongue. Oral Oncol 2004;40:154–7. 56. Spano JP, Andre F, Morat L, et al. Chemokine receptor CXCR4 and early-stage non-small cell lung cancer: pattern of expression and correlation with outcome. Ann Oncol 2004;15:613–7. 57. Shibuta K, Mori M, Shimoda K, Inoue H, Mitra P, Barnard GF. Regional expression of CXCL12/CXCR4 in liver and hepatocellular carcinoma and cellcycle variation during in vitro differentiation. Jpn J Cancer Res 2002;93:789–97. 58. Kato M, Kitayama J, Kazama S, Nagawa H. Expression pattern of CXC chemokine receptor-4 is correlated with lymph node metastasis in human invasive ductal carcinoma. Breast Cancer Res 2003;5:R144–50. 59. Wang N, Wu QL, Fang Y, et al. Expression of chemokine receptor CXCR4 in nasopharyngeal carcinoma: pattern of expression and correlation with clinical outcome. J Transl Med 2005;3:26. 60. Wang L, Wang Z, Yang B, Yang Q, Sun Y. CXCR4 nuclear localization follows binding of its ligand SDF-1 and occurs in metastatic but not primary renal cell carcinoma. Oncol Rep 2009;22:1333–9.
Contents lists available at SciVerse ScienceDirect
Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology
Prognostic value of the chemokine receptor CXCR4 and epithelial-to-mesenchymal transition in patients with squamous cell carcinoma of the mobile tongue q Sébastien Albert a,b, Muriel Hourseau c, Caroline Halimi a,b, Maria Serova a,d, Véronique Descatoire c, Béatrix Barry b, Anne Couvelard c, Maria Eugenia Riveiro a, Annemilaï Tijeras-Raballand d, Armand de Gramont d, Eric Raymond a,⇑, Sandrine Faivre a a
INSERM U728, RayLab, and Departments of Medical Oncology, Beaujon University Hospital (AP-HP – Paris 7 Diderot), Clichy, France Department of Head & Neck Surgery, Bichat University Hospital, Paris, France c Department of Pathology, Bichat University Hospital, Paris, France d AAREC Filia Research, 1 place Paul Verlaine, 92100 Boulogne-Billancourt, France b
a r t i c l e
i n f o
Article history: Received 17 March 2012 Received in revised form 8 June 2012 Accepted 12 June 2012 Available online 7 July 2012 Keywords: Head and neck cancer Tongue CXCR4 SDF-1 CXCL12 CA9 HIF-1 Vimentin E-cadherin Oral cancer Squamous cell carcinoma
s u m m a r y Objective: The aim of this study was to evaluate the expression and the prognostic value of chemokine receptor 4 (CXCR4), its cognate ligand the CXCL12, and markers of epithelial-to-mesenchymal transition (EMT) in squamous cell carcinoma (SCC) of the mobile tongue. Patients and methods: Patients with primary SCC of the mobile tongue who underwent surgery in our center were screened retrospectively. Patients without prior treatment, who had pre-surgery TNM staging and available tumor samples, were eligible. Protein expression of CXCL12, CXCR4, CA9, E-cadherin, and vimentin was determined by immunohistochemical staining, scored, and correlated with clinical and pathological parameters and overall survival. Multivariate and Cox proportional hazards analyses were performed. Results: Among 160 patients treated and screened, 47 were analyzed. CXCR4 and CXCL12 expression was high in tumor cells. CXCR4 expression in primary tumor samples was significantly higher in patients with high-grade tumors, lymph node metastases, and microscopic nerve invasion (p 6 0.05). There was a nonsignificant trend towards a correlation between high CXCL12 expression and pathologic tumor stage (p = 0.07). Tumors with high CXCR4 expression correlated with poor overall survival (hazard ratio = 3.6, 95% confidence interval 1.3–9.7; p = 0.011), notably in the CXCR4high/vimentin-positive subgroup. Vimentin-positive tumors, characterizing EMT, were associated with lower survival (hazard ratio = 4.5, 95% confidence interval 1.6–12.3; p = 0.0086). Multivariate analysis confirmed vimentin (but not CXCR4) expression as an independent prognostic factor of poor overall survival (p = 0.016). Conclusion: Our results suggest that CXCR4 is a marker of tumor aggressiveness and vimentin is an important and independent prognostic factor in patients with SCC of the mobile tongue. Ó 2012 Elsevier Ltd. All rights reserved.
Introduction Incidence of head and neck squamous cell carcinoma (HNSCC) is estimated to be half a million cases worldwide.1 HNSCCs are frequently invasive in the depth of the affected tissue due to the local q Financial support: This work was supported by the Foundation Nelia & Amadeo Barleta (FNAB) and the Association d’Aide à la Recherche et à l’Enseignement en Cancérologie (AAREC). ⇑ Corresponding author. Address: Department of Medical Oncology (INSERM U728 – PRES Paris 7 Diderot), Beaujon University Hospital, Assistance Publique – Hôpitaux de Paris, 100 boulevard du Général Leclerc, 92110 Clichy, France. Tel.: +33 1 4087 5614; fax: +33 1 4087 5487. E-mail address: [email protected] (E. Raymond).
1368-8375/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.oraloncology.2012.06.010
environment of soft tissues and can ultimately extend into the lymphatic and hematologic systems, with lymph node and distant metastases being observed in over 50% of patients at diagnosis.2 HNSCC often develops in hypoxic conditions either at the primary site or lymph node metastasis. Molecular mechanisms that drive metastasis in HNSCC remain poorly understood. However, hypoxia and necrosis occurring at the primary site and/or in lymph nodes have been associated with more frequent occurrence of distant metastasis.3,4 Chemokine receptors are 7-transmembrane-domain proteins which interact with specific cytokines known as chemokines involved in cellular chemotaxis.5 Chemokine receptor activation leads to cytoskeletal rearrangement and directional cell migration toward
1264
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
the chemokine source.6 Chemokines play a role in the progression of a large number of human malignancies by stimulating proliferation, invasion, and metastasis, and attract cancer cells toward lymphatic tissues, the liver, lung, and bone marrow, all of which are primary sources of chemokines.7–9 CXCR4 is a chemokine receptor specific for CXCL12 (also known as SDF-1, stromal-derived-factor-1) involved in chemotaxis of lymphocytes to inflammatory sites which in turn can promote chemotaxis of cancer cells.10 It has been shown that the CXCL12/CXCR4 interaction plays a critical role in carcinogenesis by activating angiogenesis, invasion, and cancer cell migration in various carcinomas.7 CXCR4 was shown to play a role in the development of metastases by regulating the directional migration of cancer cells expressing the receptor CXCR4 to target tissues that release CXCL12.11 Furthermore, hypoxic conditions activate expression of hypoxia-inducible factor-1 (HIF-1a), a transcription factor known to trigger CXCR4 expression in cancer cells.12 Although the chemokine receptor CXCR4 and its ligand, CXCL12, are known to mediate invasiveness and the process of metastasis in a number of solid cancers,13 their role in HNSCC remains incompletely elucidated. Epithelial-to-mesenchymal transition (EMT) has been recognized as one of several cellular processes facilitating the dedifferentiation of potentially cancerous epithelial cells into mesenchymal cells, thus acquiring migration and invasive capacities.14 EMT is primarily associated with both the loss of cellular adhesion, recognized by repression of E-cadherin expression, and also with changes in the cell cytoskeleton which can be identified by vimentin expression.15 EMT has previously been shown to be a key factor in the development of metastases for several tumor types including HNSCC.16,17 The aim of this study was to evaluate the expression of CXCR4 and CXCL12, as well as of the hypoxia markers carbonic anhydrase 9 (CA9) and EMT markers (E-cadherin and vimentin) in an homogeneous subgroup of HNSCC patients with resected oral squamous cell carcinoma (SCC) of the mobile tongue, and to assess correlations between these biomarkers and clinical tumor staging and prognosis. Patients and methods Patients Tissue samples were collected retrospectively from patients with SCC of the tongue treated at the University center Paris Nord Val de Seine (Bichat and Beaujon University Hospitals) between 2005 and 2008. All patients underwent surgery at the Department of Head and Neck Surgery at Bichat University Hospital (Paris, France). To be eligible for the study, patients had to provide written consent for the use of their tumor samples, have received no prior therapy, with the primary tumor site located in the mobile tongue,
have full TNM staging available prior to surgery, adequate followup in our institution including a record of date of death when appropriate, and pathological specimens for CXCR4, CXCL12, carbonic anhydrase 9 (CA9), E-cadherin and vimentin immunostaining. Patients were excluded if they had any history of other alcohol or tobacco-related tumors including other HNSCC, lung, esophageal or bladder carcinomas. Tumor staging was performed according to the American Joint Committee on Cancer (AJCC) TNM classification. Clinical data were recorded in an anonymized database. All patients provided informed consent and the study was performed in accordance with the Declaration of Helsinki. Histology A representative tissue block containing tumor and non-neoplastic tissues was selected for each patient for immunohistochemical analysis. Immunohistochemistry Tissues were fixed in 4% formalin and embedded in paraffin. Paraffin blocks of 4-lm thickness were cut, and sections were deparaffinized with xylene and rehydrated in graded alcohol. Immunohistochemistry was performed using an automated stainer, and according to the manufacturer’s guidelines (incubation 30 min at room temperature, after antigen retrieval using citrate buffer pH = 7.7, revelation with streptavidin-peroxidase kit and DAB staining) (Leica Bond™, Wetzlar, Germany). Phosphate-buffered saline was used as a negative control. The section were incubated with an anti-human CXCR4 antibody (1:100 dilution; fusin/ sc-53534, Santa Cruz Biotechnology, California, USA), an anti-human CXCL12 antibody (1:1000 dilution; 5393-100, BioVision, California, USA), an anti-human CA9 antibody (1:800 dilution; NB 100417, Novus Biological, Littleton, California, USA), an anti-human Ecadherin antibody (1:100 dilution; NCH-38, Dako, Glostrup, Denmark) and an anti-human vimentin antibody (1:400 dilution; M0725, Dako, Glostrup, Denmark) were used. The sections were then incubated with a secondary antibody according to the kit manufacture’s instructions (Leica): post primary rabbit anti-mouse IgG (<10%g/ml) in 10%(V/V) animal serum in tris-buffered saline/ 0.09% (Proclin 950, Sigma–Aldrich, Saint Louis, USA). Immunohistochemical evaluation All sections were evaluated independently by two investigators who were blinded to the clinical data. We made the choice to not use semi-quantitative method because, in our experience, the image system analyzed all the tissue without distinguishing tumor
Figure 1a Representative CXCR4 and CXCL12 immunostaining used for scoring SCC of the mobile tongue (100). The intensity of immunostaining was classified as one of four levels: (a)1 negative, (a)2 weak, (a)3 moderate, (a)4 strong.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
1265
Figure 1b Representative CXCR4 and CXCL12 immunostaining (100). Location of CXCR4 and CXCL12 staining within a tumor cell: (b)1 on the membrane (b)2 in the cytoplasm and (b)3 in the nucleus.
and normal tissue in resected tumors, contrary to the evaluation made by trained pathologists. For the CXCR4, CXCL12 and CA9 antibodies, staining intensity (including membrane, cytoplasm and nucleus localization) was scored as 0 = no detectable signal, 1 = weak staining, 2 = moderate staining, 3 = strong staining (Fig. 1a). A score was calculated according to the percentage of cancer cells with an immunostaining intensity of 0, 1, 2, and 3 using the following formula: ‘‘x score = (percentage of cancer cells with an immunostaining intensity of 0) x0 + (percentage of cancer cells with an immunostaining intensity of 1) x1 + (percentage of cancer cells with an immunostaining intensity of 2) x2 + (percentage of cancer cells with an immunostaining intensity of 3) x3’’. When the two reviewers’ scores were discordant, a joint review was performed to obtain a consensus. For each specimen, the staining location (membrane, cytoplasm or nucleus) was recorded. The overall median and mean scores were determined and patients were then allocated to one of two groups according to whether the CXCR4, CXCL12 and CA9 score; samples with a score lower than the median were classified as low (CXCR4low, CXCL12low and CA9low) and samples with a score higher than the median were classified as high (CXCR4high, CXCL12high and CA9high). Scoring the immunohistochemical evaluation of E-cadherin and vimentin was not technically feasible, notably for vimentin which was frequently not expressed and when detectable was primarily located at the rim of the tumor. Statistics The correlation of CXCR4 staining intensity with clinicopathological patterns was assessed with Chi-square test, or Fisher’s exact Table 1 Tumor staging at diagnosis according to the American Joint Committee on Cancer TNM system for carcinoma of oral cavity.
N0 N1 N2a N2b N2c N3 Total
T1
T2
T3
T4
Total
14 0 0 0 0 0 14
7 2 1 2 1 1 14
2 3 2 6 2 1 16
0 1 0 1 1 0 3
23 6 3 9 4 2 47
T: size or direct extent of the primary tumor from T1 to T4; N: degree of spread to regional lymph nodes from N0 to N3.
test if values were lower than five. Survival rates were determined using the Kaplan–Meier method, and statistical analysis was carried out using the log-rank test. For multivariate analyses, independent prognostic factors were determined using a Cox proportional hazards model. p-Values <0.05 were considered significant in all statistical analyses. Results Patient characteristics Among 160 patients treated for HNSCC of the oral cavity during the study period, 47 were evaluable for analysis. The other 113 patients were excluded due to lack of informed consent, tumor localization other than the mobile tongue, prior history of HNSCC or other tumors, lost to follow-up, or a therapeutic decision other than surgery as first-line treatment. Included patients had a median age at diagnosis of 61 years (range 34–92), with a 4.3–1 ratio of males to females. Thirty-six patients had a tumor confined to the mobile tongue while for 11 patients the tumor extended to the floor of the oral cavity. Seventy-five percent of patients were current or ex-tobacco smokers, 62% were current or ex-alcohol drinkers and 61% were both. Clinical stage was defined according to the classification of malignant tumors (AJCC-TNM) in 47 patients including 14 (30%) with stage I, 7 (15%) with stage II, 7 (15%) with stage III and 19 (40%) with stage IV disease (details of TNM are shown in Table 1). Pathological evaluation of the tumor resection showed that 67% of primary tumors were well differentiated; nerve and vascular invasion were observed in 55% and 46% of cases, respectively. Cancerous cells were identified in surgical resection margins of the primary tumor in 19% of cases (R1). Tumor lymph node involvement was observed in 49% of cases, 79% of which had extra-capsular invasion. Median follow-up was 48 months (range 22–75 months) and the 5-year overall survival rate was 56%. Eighteen patients died during the duration of follow-up, which included 13 patients who died of local recurrences, 3 patients who died of distant metastases, and 2 patients who died of other diseases. CXCR4, CXCL12, and CA9 immunostaining CXCR4, CXCL12 and CA9 were frequently highly expressed in tumor cells (87%, 95% and 83% of tumors, respectively) compared
1266
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
Figure 2 Comparison of CXCR4 immunostaining in normal and tumor tissues (25). (a) CXCR4 immunostaining in normal tissues of the tongue: 1: negative control of squamous epithelium, 2: positive control of muscle tissue. (b) CXCR4 immunostaining in SCC of the tongue showing strong CXCR4 expression.
Table 2 Median and mean scores of CXCR4, SDF-1 and CA9 immunostaining.
CXCR4 SDF-1 CA9
Median score [range] (n = 47)
Mean score ± SD (n = 47)
Patients with low IHC scorea Mean ± SD
Patients with high IHC scoreb Mean ± SD
120 [20; 250] 110 [20; 230] 110 [10; 280]
111.0 ± 54.9 111.6 ± 52.1 116.3 ± 65.3
67.9 ± 31.6 70.4 ± 24.4 67.1 ± 36.8
156.1 ± 33.1 152.8 ± 37.6 161.3 ± 51.9
IHCS: immunohistochemical score. a Low score = score less than the median. b High score = score greater than the median.
to normal epithelial cells adjacent to the tumor tissues (Fig. 2a and b). The median immunohistochemical scores for CXCR4, CXCL12 and CA9 were 120, 110, and 110 respectively (Table 2). Staining for CXCR4 was predominantly located in the cytoplasm (98% of cases) and the nucleus (98% of cases) of tumor cells and on the membrane in 55% of cases. CXCL12 was mainly expressed in cytoplasm (93% of cases) and also on the membrane (59% of cases) and in the nucleus (54% of cases) of tumor cells (Fig. 1b). Lymphocytes infiltrating the normal epithelium and muscle cells were shown to have high CXCR4 expression (Fig. 2a and b). CA9 immunostaining was mainly observed in the cytoplasm and membrane but never in the nucleus of tumor cells. In 17 patients with paired primary and lymph node tumor biopsies, we found no difference in the level of CXCR4 and CXCL12 expression between primary and lymph node tumor cells. Correlation of CXCR4, CXCL12 and CA9 expression with clinical and pathological features An analysis of CXCR4, CXCL12 and CA9 expression according to primary tumor stage is shown in Fig. 3, and statistical analyses of the correlation between expression of these markers and various histopathological and clinical parameters is shown in Table 3. Expression of CXCR4 in primary tumor samples was significantly higher in patients with advanced cancers (stages III and IV; p = 0.005), and correlated with both tumor size (p = 0.002) and nerve invasion (p = 0.05). CXCR4 expression was also significantly higher in tumors with cervical lymph node metastasis than those without (p = 0.01). In addition, high CXCR4 expression in the primary tumor was reported in 67% of patients with lymph node metastasis vs. 29% of patients without. However, the degree of differentiation and vascular invasion were not associated with CXCR4 expression levels, nor gender or smoking history, although patients with a history of alcohol consumption showed borderline significance (p = 0.047).
CXCL12 expression was not correlated with any of the clinical and pathological variables evaluated, although there was a trend towards a correlation between higher pathological stage and stronger CXCL12 expression (p = 0.072). In primary tumor samples, CA9 expression was significantly higher in patients with advanced stage III and IV diseases (p = 0.037) and was higher in primary tumors with cervical lymph node metastasis than in those without (p = 0.019). When CXCR4 and CA9 were both over-expressed, the positive predictive value for patients to have lymph node metastasis was 91%. In terms of histopathology parameters, CA9 expression was higher in poorly differentiated tumors (p = 0.01) and those with vascular invasion (p = 0.018). No significant inter-marker correlation was observed between CA9 and CXCR4/CXCL12 immunostaining scores. No survival difference was observed between patients with CXCL12low and CXCL12high values (data not shown). Conversely, high expression of CXCR4 was highly correlated with poor prognosis. Overall survival was statistically higher in patients with CXCR4low compared to those with CXCR4high expressions (Hazard ratio: 3.6, 95% confidence interval: 1.3–9.7, p-value = 0.011; Fig. 4).
CXCR4 expression in tumors undergoing EMT In primary tumor samples, expression of E-cadherin and vimentin was observed in 36 of 43 patients (83.7%) and 14 of 43 patients (32.5%), respectively (samples were not evaluable for 4 patients). Among the 14 vimentin-positive tumors, 7 (50%) also expressed E-cadherin (Fig. 5A upper panel). The 14 vimentin-positive tumors, characterizing EMT, were associated with a poorer overall survival outcome (Fig. 5A lower panel) compared to vimentin-negative tumors (hazard ratio: 4.5; 95% confidence interval: 1.6–12.3, p = 0.0086). No significant correlations between vimentin and expression of CXCR4 (Fig. 5B upper panel) or CXCL12 (data not shown) were found. Interestingly, patients whose tumors were
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
1267
Figure 3 Correlation between CXCR4/CXCL12 and tumor stage. Expression of (a1) CXCR4 and (a2) CXCL12, according to clinical tumor stage (two groups: stage I–II, stage III–IV); (a3) mean CXCR4/CXCL12 scores for the two groups. Expression of (b1) CXCR4 and (b2) CXCL12, according to histopathological pT staging (two groups: pT1–2, pT3–4); (b3) mean CXCR4/CXCL12 scores for the two groups. The y-axis of the 4 upper panel figures (a1, a2, b1 and b2) represents the number of patients. The Y-axis of the 2 lower panel graphs (a3 and b3) represents the mean value, standard deviation was then determined.
vimentin-negative and CXCR4low had significantly higher survival compared to other combinations (Fig. 5B lower panel, p = 0.0043). Univariate and multivariate analyses of prognostic factors A univariate analysis showed a significant correlation for histopathological lymph node invasion (p < 0.001), high histopathological tumor stage (pT3 or pT4; p = 0.011), histopathological nerve invasion (p = 0.04), CXCR4 expression (p = 0.046), and vimentin expression (p = 0.0034) with a poor survival prognosis. However, no correlation was observed for the degree of tumor differentiation (p = 0.37) or histopathological vascular invasion (p = 0.7). In a multivariate analysis, vimentin (p = 0.016) but not CXCR4 (p = 0.24) expression was shown to be an independent marker of poor prognosis. Discussion HNSCCs are frequently locally invasive at diagnosis and many develop lymph node invasion followed by distant metastasis.18 Lymph node invasion and metastases are often responsible for
relapse after complete surgery or high dose local radiotherapy.19 Low oxygen tissue pressure, typically less than 5 mm Hg, has been widely reported in both primary HNSCC tumors and lymph node metastases.20 Intratumoral hypoxia has been recognized as a major factor that contributes to angiogenesis, invasiveness, metastasis, alteration of tumor metabolism, and genomic instability, and is also suspected to play a role in resistance to chemotherapy and radiotherapy.21 In this study, we investigated tumor hypoxia using carbonic anhydrase-9 (CA9). CA9 is a transmembrane protein which is involved in the catalytic hydration of carbon dioxide to carbonic acid. Transcription of this gene is known to be regulated by the Von Hippel–Lindau tumor suppressor gene, responsible for targeting HIF-1a for oxygen-dependent proteolysis.22 When low levels of oxygen occur in tumors, HIF-1a over-expression is induced as well as the expression of CA9.23,24 Recent studies showed that CA9 was overexpressed in a wide variety of tumor types and was induced by hypoxia.22 In addition, CA9 is a stable marker of hypoxia in opposition to HIF-1, which is very labile and difficult to use in IHC, explaining our choice for CA9 in agreement to many other authors. We have shown that CXCR4, CXCL12 and CA9 were commonly expressed, and at different levels, in patients with SCC of the mobile
1268
Table 3 Correlations between clinicopathological features and immunostaining results by patient and staining intensity. SDF-1
CXCR4
CA-9
E-cadherin
Vimentin
SDF1high
P value
CXCR4low
CXCR4high
P value
CA9low
CA9high
P value
E-cadh. negative
E-cadh. positive
P value
Vimentin negative
Vimentin positive
P value
21 2
16 7
NS (.13)
19 5
19 4
NS (.94)
18 5
19 4
NS (1)
7 0
29 8
NS (.17)
23 7
13 1
NS (.19)
16 7
18 5
NS (.73)
17 7
18 5
NS (.8)
18 5
17 6
NS (1)
6 1
27 10
NS (.48)
23 7
11 3
NS (.88)
16 7
12 11
NS (.36)
11 13
18 5
.047
18 5
11 12
NS (.54)
6 1
22 15
NS (.19)
16 14
12 2
0.038
16 7
17 6
NS (1)
19 5
15 8
NS (.45)
18 5
15 8
NS (.51)
5 2
26 11
NS (.95)
22 8
9 5
NS (.54)
12
9
NS (.65)
16
5
.005
14
6
.037
2
17
NS (.39)
15
3
NS (.073)
Stage III–IV
12
14
8
18
9
17
5
20
15
11
Histopathology Tumor staging pT1 or pT2
17
10
20
8
16
11
4
22
20
5
pT3 or pT4 6 Lymph node invasion Negative 10
13
4
15
7
12
3
15
10
9
7
16
7
16
3
19
16
5
Positive Differentiation Well
13
10
17
7
16
7
4
18
14
9
18
13
19
13
20
11
6
23
17
12
5
10
5
10
3
12
1
14
13
2
13 10
10 9
NS (1)
10 14
17 6
.05
11 12
16 7
NS (.23)
4 3
23 14
NS (0.8)
17 13
9 5
NS (.63)
9 14
12 11
NS (.55)
10 14
11 12
NS (.9)
6 17
15 8
.018
2 5
18 19
NS (0.33)
15 15
5 9
NS (.38)
Clinical parameter Gender Male Female Smoker Yes No Alcohol Yes No Tumor description B or UB U Stage Stage I–II
Moderate/ poor Nerve invasion Present Absent Vasc. invasion Present Absent
13
NS (.072)
NS (.56)
NS (.2)
.002
.01
NS (.18)
NS (.23)
.019
.01
NS (0.91)
NS (0.68)
NS (0.23)
NS (.053)
NS (.28)
NS (.058)
B: budding; UB: ulcerating and budding; U: ulcerating; SDF-1low: number of patients with an IHC score lesser than 110 (median value of IHC SDF-1 scores); SDF-1high: number of patients with an IHC score higher than the median 110; CXCR4low: number of patients with an IHC score lesser than 120 (corresponding to the median value of IHC CXCR4 scores); CXCR4high: number of patients with an IHC score higher than 120; CA-9low: number of patients with an IHC score lesser than 110 (median value of IHC CA-9 scores); CA-9high: number of patients with an IHC score higher than 110; Vasc.: vascular; NS not significant.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
SDF1low
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
A
Percent survival
100 75 50 25 0 0
10
20
30
40
50
60
70
80
Months 100
Percent survival
B
75 50 25
p-value: 0.011 Hazard ratio: 3.6 (95%CI: 1.33-9.72)
0 0
10
20
30
40
50
60
70
80
Months CXCR4high (Median: 41 months) CXCR4low (Median: not reach) Figure 4 Overall survival of patients with resected SCC of the mobile tongue (A) and prognostic value of CXCR4 expression (B). Patients were grouped as CXCR4low and CXCR4high according to their score values as lower or higher than the median value of CXCR4 score in the entire population respectively.
tongue. Activation of HIF-1 under hypoxic conditions may result in the transcription of several growth factors and receptors associated with tumor cell migration and angiogenesis, such as VEGF, VEGFR, CXCL12,25 and CXCR4.12,26–29 Expression of CXCR4 has been associated with a poor outcome in head and neck carcinomas.4,30–34 Ishikawa et al. also demonstrated that HIF-1a regulates hypoxia-induced CXCR4 expression in oral HNSCC.35 Tumor cells expressing a high level of CXCR4 in an hypoxic environment had enhanced motility and invasive capacity to enter the blood and the lymphatic microcirculation. In our study, over-expression of CXCR4 or CA9 was observed in patients with metastatic lymph nodes. When CXCR4 and CA9 were both over-expressed, the positive predictive value for patients to have lymph node metastasis was 91%. These findings suggest that the over-expression of CXCR4 in hypoxic conditions may correlate with a higher risk for developing lymph node metastasis. CXCR4 and CXCL12 may play a key role in invasion and migration processes in HNSCC via EMT and activation of metalloproteases. Ou et al.36 have studied the role of TWIST a transcription factor involved in EMT. The authors suggest that TWIST regulates CXCR4 expression in SCCs, which in turn might be associated with lymph node metastasis.36 In addition, Taki et al.37 demonstrated that CXCR4 was up-regulated by CXCL12 and TGF-b1 in oral SCCs, enhancing migratory activity. They also showed that SNAIL, another transcription factor promoting EMT, induced over-expression of CXCL12 giving a mesenchymal phenotype to the epithelial tumor cell. Onoue et al.38 reported in vitro that EMT markers (vimentin and snail) were induced by CXCL12/CXCR4 in oral SCC cells through the PI3K-Akt pathway. These results suggest that up-regulation of
1269
CXCR4 may be a consequence of EMT and that EMT is a consequence of CXCR4 activation. In our study, tumors expressing vimentin (33% of patients) were associated with high level of CXCR4 in 71% of cases. Furthermore, recent studies suggested that CXCL12/CXCR4 enhanced invasion and migration mechanisms in HNSCC through the regulation of MMP-9 and MMP-13.39,40 Therefore, EMT and matrix metalloproteinase activation might promote invasion and motility of tumor cells and metastatic processes in HNSCC.41–43 In our study, the size of the primary tumor and its degree of local invasiveness (as reflected by the T-stage), pathological nerve invasion and lymph node metastasis both of which represent clinical aggressiveness, were strongly correlated with high CXCR4 expression. We also expected to observe correlations between CXCR4 expressions and vascular invasion in tumors, but no statistical correlation was detectable in this study. This can be partly explained by the fact that vascular invasion is often found in histopathological analysis in HNSCC. Considering the number of patients in this study, the power to detect statistical correlation is likely to be too low to reach conclusion in this study. However, our results are consistent with recent studies showing that CXCR4 activation plays a major role in lymph node metastasis in head and neck squamous cell carcinoma.35,44–48 CXCR4 may be sensitive to the chemo-attractive action of CXCL12 located in distant tissues, which would then be homing targets for metastasizing tumor cells expressing high CXCR4 levels. Consistent with this hypothesis, relatively high levels of CXCL12 are found in the lung and the lymph nodes of patients with HNSCC.11,49–51 To date, very few groups have investigated CXCR4 as a prognostic factor for survival in patients with HNSCC. The median follow-up in our study was 48 months, which despite being somehow limited could be considered sufficient to use overall survival as a primary endpoint using Kaplan–Meier method, as most recurrence are likely to occur within this time frame. In a previous study, the 5-year survival rate in patients with CXCR4-negative tumors was significantly higher (71.4%) than those with CXCR4-positive tumors (39.1%).52 In another trial, CXCR4 was found to be an independent prognostic factor for poor survival in patients with HNSCC.53 These data are consistent with our results which also showed that high CXCR4 expression in tumor cells was associated with poorer survival. Interestingly, both vimentin and CXCR4 expression were associated with poor prognosis in our study, patients with the better outcome being those with tumors which were vimentin-negative with low CXCR4 expression. However, vimentin appears to be a stronger marker than CXCR4 to predict poor outcome as demonstrated by multivariate analysis in our patient population. Several studies have evaluated CXCL12 expression by immunohistochemistry in biopsies from HNSCC patients and various associated clinicopathological factors. Uchida et al.52 described that the 5-year survival rate was 25.0% for the CXCL12-positive group and 71.4% for the CXCL12-negative group in patients with oral SCC. The results obtained by Almolfi et al.45 found no clear correlation between the expression of CXCL12 and other prognostic factors in oral HNSCC patients. In contrast, a recent study demonstrated that low CXCL12 level correlated with occurrence of metastases and poor prognosis.54 In this latter trial, the authors argued that a low level of CXCL12 in the area surrounding tumor cells could facilitate tumor cell detachment from the primary tumor and migration to metastatic sites. However, CXCL12 levels were assayed by quantitative RT-PCR, which reflects transcription but may not correlate with protein expression. Discrepancies between studies may be explained by regional differences of Caucasian population or Asiatic population. In our study most patients had alcohol and tobacco habits. The influence of HPV infection on HNSCC may account for these differences but would require larger cohorts to reach conclu-
1270
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271
A
B
Figure 5 Correlations and prognostic value of epithelial-to-mesenchymal transition (EMT) and CXCR4 expression in tumors of patients with SCC of the mobile tongue. (A) Number of patients with E-cadherin and/or vimentin expression (upper panel), and overall survival according to EMT status (lower panel); (B) Number of patients with EMT and/or CXCR4 expression (upper panel), and overall survival according to CXCR4 and EMT status (lower panel).
sions. Furthermore, discrepancies between studies may also be attributed to the fact that most include a global analysis without considering different sub-locations of the head and neck carcinoma (oral cavity, larynx, oropharynx, hypopharynx). The anatomical location is an essential consideration in clinical practice influencing the therapeutic strategy and prognostic. For this reason we chose to restrict our study to patients with SCC of the mobile tongue thus evaluating a homogeneous cohort and avoiding misinterpretation of results. CXCR4 was homogeneously detected in the cytoplasm and nucleus and at the cell membrane of tumor cells. Previous immunohistochemical studies also demonstrated that CXCR4 was primarily located in the cytoplasm49,55 but was also found in the nucleus in head and neck carcinomas,56 as well as in hepatocellular carcinomas,57 in non-small-cell lung cancers,58 and in nasopharyngeal carcinomas.59 It is possible that when translocated into the nucleus, CXCR4 acquires different functions and may regulate transcription of genes involved in cellular invasion, migration and EMT. This hypothesis was supported by Wang et al.60 who showed that nuclear localization of CXCR4 might result in a migratory or metastatic phenotype in renal carcinoma cells in vitro. Further studies are necessary to better understand the mechanism of action of CXCR4 as a transcriptional regulator. In summary, our results indicate that CXCR4 and vimentin could be used as potential prognostic factors of aggressiveness in completely resected SCC of the mobile tongue. Conflict of interest statement None declared.
Acknowledgements We thank Jessica Sabatier, Christophe Assailly and Gilles Arcangeli for their skilful technical assistance and Fabien Reyal for statistical analysis. References 1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009;59:225–49. 2. Gourin CG, Podolsky RH. Racial disparities in patients with head and neck squamous cell carcinoma. Laryngoscope 2006;116:1093–106. 3. Brizel DM, Schroeder T, Scher RL, et al. Elevated tumor lactate concentrations predict for an increased risk of metastases in head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:349–53. 4. Brizel DM, Sibley GS, Prosnitz LR, Scher RL, Dewhirst MW. Tumor hypoxia adversely affects the prognosis of carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 1997;38:285–9. 5. Laing KJ, Secombes CJ. Chemokines. Dev Comp Immunol 2004;28:443–60. 6. Takabayashi T, Takahashi N, Okamoto M, Yagi H, Sato M, Fujieda S. Lipopolysaccharides increase the amount of CXCR4, and modulate the morphology and invasive activity of oral cancer cells in a CXCL12-dependent manner. Oral Oncol 2009;45:968–73. 7. Balkwill F. Cancer and the chemokine network. Nat Rev Cancer 2004;4:540–50. 8. Schimanski CC, Bahre R, Gockel I, et al. Dissemination of hepatocellular carcinoma is mediated via chemokine receptor CXCR4. Br J Cancer 2006;95:210–7. 9. Sun X, Cheng G, Hao M, et al. CXCL12/CXCR4/CXCR7 chemokine axis and cancer progression. Cancer Metastasis Rev 2010;29:709–22. 10. Bleul CC, Farzan M, Choe H, et al. The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry. Nature 1996;382:829–33. 11. Muller A, Homey B, Soto H, et al. Involvement of chemokine receptors in breast cancer metastasis. Nature 2001;410:50–6. 12. Staller P, Sulitkova J, Lisztwan J, Moch H, Oakeley EJ, Krek W. Chemokine receptor CXCR4 downregulated by von Hippel–Lindau tumour suppressor pVHL. Nature 2003;425:307–11. 13. Balkwill F. The significance of cancer cell expression of the chemokine receptor CXCR4. Semin Cancer Biol 2004;14:171–9.
S. Albert et al. / Oral Oncology 48 (2012) 1263–1271 14. Sabbah M, Emami S, Redeuilh G, et al. Molecular signature and therapeutic perspective of the epithelial-to-mesenchymal transitions in epithelial cancers. Drug Resist Updat 2008;11:123–51. 15. Satelli A, Li S. Vimentin in cancer and its potential as a molecular target for cancer therapy. Cell Mol Life Sci 2011;68:3033–46. 16. Ghoul A, Serova M, Astorgues-Xerri L, et al. Epithelial-to-mesenchymal transition and resistance to ingenol 3-angelate, a novel protein kinase C modulator, in colon cancer cells. Cancer Res 2009;69:4260–9. 17. Nijkamp MM, Span PN, Hoogsteen IJ, van der Kogel AJ, Kaanders JH, Bussink J. Expression of E-cadherin and vimentin correlates with metastasis formation in head and neck squamous cell carcinoma patients. Radiother Oncol 2011;99:344–8. 18. Coca-Pelaz A, Rodrigo JP, Suarez C. Clinicopathologic analysis and predictive factors for distant metastases in patients with head and neck squamous cell carcinomas. Head Neck 2012;34:771–5. 19. Hauswald H, Simon C, Hecht S, Debus J, Lindel K. Long-term outcome and patterns of failure in patients with advanced head and neck cancer. Radiat Oncol 2011;6:70. 20. Lartigau E, Le Ridant AM, Lambin P, et al. Oxygenation of head and neck tumors. Cancer 1993;71:2319–25. 21. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer 2011;11:393–410. 22. Ivanov SV, Kuzmin I, Wei MH, et al. Down-regulation of transmembrane carbonic anhydrases in renal cell carcinoma cell lines by wild-type von Hippel– Lindau transgenes. Proc Natl Acad Sci USA 1998;95:12596–601. 23. Beasley NJ, Wykoff CC, Watson PH, et al. Carbonic anhydrase IX, an endogenous hypoxia marker, expression in head and neck squamous cell carcinoma and its relationship to hypoxia, necrosis, and microvessel density. Cancer Res 2001;61:5262–7. 24. Wykoff CC, Beasley NJ, Watson PH, et al. Hypoxia-inducible expression of tumor-associated carbonic anhydrases. Cancer Res 2000;60:7075–83. 25. Ceradini DJ, Kulkarni AR, Callaghan MJ, et al. Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nat Med 2004;10:858–64. 26. Schioppa T, Uranchimeg B, Saccani A, et al. Regulation of the chemokine receptor CXCR4 by hypoxia. J Exp Med 2003;198:1391–402. 27. Zagzag D, Krishnamachary B, Yee H, et al. Stromal cell-derived factor-1alpha and CXCR4 expression in hemangioblastoma and clear cell-renal cell carcinoma: von Hippel–Lindau loss-of-function induces expression of a ligand and its receptor. Cancer Res 2005;65:6178–88. 28. Zagzag D, Lukyanov Y, Lan L, et al. Hypoxia-inducible factor 1 and VEGF upregulate CXCR4 in glioblastoma: implications for angiogenesis and glioma cell invasion. Lab Invest 2006;86:1221–32. 29. Carver BS, Tran J, Gopalan A, et al. Aberrant ERG expression cooperates with loss of PTEN to promote cancer progression in the prostate. Nat Genet 2009;41:619–24. 30. Tilakaratne WM, Nissanka-Jayasuriya EH. Value of HIF-1alpha as an independent prognostic indicator in oral squamous cell carcinoma. Expert Rev Mol Diagn 2011;11:145–7. 31. Eckert AW, Schutze A, Lautner MH, Taubert H, Schubert J, Bilkenroth U. HIF1alpha is a prognostic marker in oral squamous cell carcinomas. Int J Biol Markers 2010;25:87–92. 32. Hoogsteen IJ, Marres HA, Bussink J, van der Kogel AJ, Kaanders JH. Tumor microenvironment in head and neck squamous cell carcinomas: predictive value and clinical relevance of hypoxic markers. A review. Head Neck 2007;29:591–604. 33. Harris AL. Hypoxia–a key regulatory factor in tumour growth. Nat Rev Cancer 2002;2:38–47. 34. Le QT, Denko NC, Giaccia AJ. Hypoxic gene expression and metastasis. Cancer Metastasis Rev 2004;23:293–310. 35. Ishikawa T, Nakashiro K, Hara S, et al. CXCR4 expression is associated with lymph-node metastasis of oral squamous cell carcinoma. Int J Oncol 2006;28:61–6. 36. Ou DL, Chien HF, Chen CL, Lin TC, Lin LI. Role of twist in head and neck carcinoma with lymph node metastasis. Anticancer Res 2008;28:1355–9. 37. Taki M, Higashikawa K, Yoneda S, et al. Up-regulation of stromal cell-derived factor-1alpha and its receptor CXCR4 expression accompanied with epithelial– mesenchymal transition in human oral squamous cell carcinoma. Oncol Rep 2008;19:993–8.
1271
38. Onoue T, Uchida D, Begum NM, Tomizuka Y, Yoshida H, Sato M. Epithelial– mesenchymal transition induced by the stromal cell-derived factor-1/CXCR4 system in oral squamous cell carcinoma cells. Int J Oncol 2006;29:1133–8. 39. Yu T, Wu Y, Helman JI, Wen Y, Wang C, Li L. CXCR4 promotes oral squamous cell carcinoma migration and invasion through inducing expression of MMP-9 and MMP-13 via the ERK signaling pathway. Mol Cancer Res 2012;9:161–72. 40. Tan CT, Chu CY, Lu YC, et al. CXCL12/CXCR4 promotes laryngeal and hypopharyngeal squamous cell carcinoma metastasis through MMP-13dependent invasion via the ERK1/2/AP-1 pathway. Carcinogenesis 2008;29:1519–27. 41. Hartmann TN, Burger JA, Glodek A, Fujii N, Burger M. CXCR4 chemokine receptor and integrin signaling co-operate in mediating adhesion and chemoresistance in small cell lung cancer (SCLC) cells. Oncogene 2005;24:4462–71. 42. Janowska-Wieczorek A, Marquez LA, Dobrowsky A, Ratajczak MZ, Cabuhat ML. Differential MMP and TIMP production by human marrow and peripheral blood CD34(+) cells in response to chemokines. Exp Hematol 2000;28:1274–85. 43. Samara GJ, Lawrence DM, Chiarelli CJ, et al. CXCR4-mediated adhesion and MMP-9 secretion in head and neck squamous cell carcinoma. Cancer Lett 2004;214:231–41. 44. Ueda M, Shimada T, Goto Y, et al. Expression of CC-chemokine receptor 7 (CCR7) and CXC-chemokine receptor 4 (CXCR4) in head and neck squamous cell carcinoma. Auris Nasus Larynx 2010;37:488–95. 45. Almofti A, Uchida D, Begum NM, et al. The clinicopathological significance of the expression of CXCR4 protein in oral squamous cell carcinoma. Int J Oncol 2004;25:65–71. 46. Zheng C, Ge R, Ge M, Sun X. The expression and relationship of CXCR4 and microvessel density in laryngeal squamous cell carcinoma. Lin Chung Er Bi Yan Hou Tou Jing Wai Ke Za Zhi 2011;25:123–6. 47. Oliveira-Neto HH, Silva ET, Leles CR, et al. Involvement of CXCL12 and CXCR4 in lymph node metastases and development of oral squamous cell carcinomas. Tumour Biol 2008;29:262–71. 48. Yin D, Gao Z. Expression and clinical significance of chemokine receptor 4 in oral squamous cell carcinoma. Hua Xi Kou Qiang Yi Xue Za Zhi 2007;25:392–5. 49. Meng X, Wuyi L, Yuhong X, Xinming C. Expression of CXCR4 in oral squamous cell carcinoma: correlations with clinicopathology and pivotal role of proliferation. J Oral Pathol Med 2010;39:63–8. 50. Sun YX, Wang J, Shelburne CE, et al. Expression of CXCR4 and CXCL12 (SDF-1) in human prostate cancers (PCa) in vivo. J Cell Biochem 2003;89:462–73. 51. Phillips RJ, Burdick MD, Lutz M, Belperio JA, Keane MP, Strieter RM. The stromal derived factor-1/CXCL12-CXC chemokine receptor 4 biological axis in nonsmall cell lung cancer metastases. Am J Respir Crit Care Med 2003;167:1676–86. 52. Uchida D, Onoue T, Tomizuka Y, et al. Involvement of an autocrine stromal cell derived factor-1/CXCR4 system on the distant metastasis of human oral squamous cell carcinoma. Mol Cancer Res 2007;5:685–94. 53. Lee JI, Jin BH, Kim MA, Yoon HJ, Hong SP, Hong SD. Prognostic significance of CXCR-4 expression in oral squamous cell carcinoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;107:678–84. 54. Clatot F, Picquenot JM, Choussy O, et al. Intratumoural level of SDF-1 correlates with survival in head and neck squamous cell carcinoma. Oral Oncol 2011;47:1062–8. 55. Delilbasi CB, Okura M, Iida S, Kogo M. Investigation of CXCR4 in squamous cell carcinoma of the tongue. Oral Oncol 2004;40:154–7. 56. Spano JP, Andre F, Morat L, et al. Chemokine receptor CXCR4 and early-stage non-small cell lung cancer: pattern of expression and correlation with outcome. Ann Oncol 2004;15:613–7. 57. Shibuta K, Mori M, Shimoda K, Inoue H, Mitra P, Barnard GF. Regional expression of CXCL12/CXCR4 in liver and hepatocellular carcinoma and cellcycle variation during in vitro differentiation. Jpn J Cancer Res 2002;93:789–97. 58. Kato M, Kitayama J, Kazama S, Nagawa H. Expression pattern of CXC chemokine receptor-4 is correlated with lymph node metastasis in human invasive ductal carcinoma. Breast Cancer Res 2003;5:R144–50. 59. Wang N, Wu QL, Fang Y, et al. Expression of chemokine receptor CXCR4 in nasopharyngeal carcinoma: pattern of expression and correlation with clinical outcome. J Transl Med 2005;3:26. 60. Wang L, Wang Z, Yang B, Yang Q, Sun Y. CXCR4 nuclear localization follows binding of its ligand SDF-1 and occurs in metastatic but not primary renal cell carcinoma. Oncol Rep 2009;22:1333–9.