Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

MUC1 expression and anti-MUC1 serum immune response in head and neck squamous cell carcinoma (HNSCC): a multivariate analysis

MUC1 expression and anti-MUC1 serum immune response in head and neck squamous cell carcinoma... Background: HNSCC progression to adjacent tissue and nodes may be mediated by altered glycoproteins and glycolipids such as MUC1 mucin. This report constitutes a detailed statistical study about MUC1 expression and anti-MUC1 immune responses in relation to different clinical and pathological parameters which may be useful to develop new anti HNSCC therapeutic strategies. Patients and methods: Fifty three pre treatment HNSCC patients were included: 26 (49.1%) bearing oral cavity tumors, 17 (32.1%) localized in the larynx and 10 (18.8%) in the pharynx. Three patients (5.7%) were at stage I, 5 (9.4%) stage II, 15 (28.3%) stage III and 30 (56.6%) at stage IV. MUC1 tumor expression was studied by immunohistochemistry employing two anti-MUC1 antibodies: CT33, anti cytoplasmic tail MUC1 polyclonal antibody (Ab) and C595 anti-peptidic core MUC1 monoclonal antibody. Serum levels of MUC1 and free anti-MUC1 antibodies were detected by ELISA and circulating immune complexes (CIC) by precipitation in polyethylene glycol (PEG) 3.5%; MUC1 isolation from circulating immune complexes was performed by protein A-sepharose CL-4B affinity chromatography followed by SDS-PAGE and Western blot. Statistical analysis consisted in Multivariate Principal Component Analysis (PCA); ANOVA test (Tukey's test) was employed to find differences among groups; nonparametrical correlations (Kendall's Tau) were applied when necessary. Statistical significance was set to p < 0.05 in all cases. Results: MUC1 cytoplasmic tail was detected in 40/50 (80%) and MUC1 protein core in 9/50 (18%) samples while serum MUC1 levels were elevated in 8/53 (15%) patients. A significant statistical correlation was found between MUC1 serum levels and anti-MUC1 IgG free antibodies, while a negative correlation between MUC1 serum levels and anti-MUC1 IgM free antibodies was found. Circulating immune complexes were elevated in 16/53 (30%) samples and were also statistically associated with advanced tumor stage. MUC1 was identified as an antigenic component of IgG circulating immune complexes. Moreover, poorly differentiated tumors were inversely correlated with tumor and serum MUC1 detection and positively correlated with node involvement and tumor mass. Conclusion: Possibly, tumor cells produce MUC1 mucin which is liberated to the circulation and captured by IgG antibodies forming MUC1-IgG-CIC. Another interesting conclusion is that poorly differentiated tumors are inversely correlated with tumor and serum MUC1 detection. Page 1 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 In addition, circulating immune complexes levels were Background In western countries, head and neck squamous cell carci- statistically associated with tumor size, inversely associ- noma (HNSCC) occupies the 5th place in frequency and ated with MUC1 tumor expression and were not posi- also is the 5th cause of mortality due to cancer. HNSCC tively associated with serum MUC1 and free anti-MUC1 localization consists about 40% in the oral cavity, 15% in IgG. Finally, tumor size, node involvement and poor dif- the pharynx, 25% in the larynx and the rest sites such as ferentiation were positively associated. salivary glands and thyroid [1,2]. Data obtained from argentinian records show some differences compared Methods with data found in other countries since the most frequent Patients localization has been larynx (1665/3127, 53%) followed Fifty three pre treatment HNSCC patients from the "Hos- by oral cavity (1035/3127, 33%) and finally, pharynx pital General Interzonal de Agudos Gral. San Martín", La showing 427/3127, 14%. In 2002, the total number of Plata, Argentina were included in this study (Table 1); all HNSCC cases informed in Argentina was 3127 (16.9/ tumors were primary; patients with previous history of 100000) inhabitants [3]. HNSCC or who had received preoperative therapy were not included. Patients were clinically categorized accord- Worldwide, delayed diagnosis is common and frequently ing to the American Joint Cancer Committee (AJCC, Can- very scarce improvement in five-year survival over the last cer Staging Manual, 2002). Sex and age distribution of four decades has been observed [4]. patients was: 39 males (73.6%) and 14 (26.4%) females with a mean age of 60.67 (11.55 SD) years and a range HNSCC mainly progress to adjacent tissue and nodes from 29 to 98 years. Twenty six patients (49.1%) had their while distant metastasis is a late event. The ability of primary tumor located in the oral cavity, 17 (32.0%) in tumor cells to invade is an acquired and progressive phe- the larynx and 10 (18.9%) in the pharynx. Three patients nomenon mediated, in many cases, by the alteration of (5.7%) were at stage I, 5 (9.4%) at stage II, 15 (28.3%) at membrane glycoproteins such as mucins. Dabelsteen and stage III and 30 (56.6%) at stage IV; stages IVA, IVB and Gao [5] proposed that the presence of different glycosyla- IVC cases where also included. Eighteen were classified as tion patterns modulate the behavior of these membrane well differentiated (34.6%), 19 as moderately differenti- glycoproteins involved in cell signaling. In adenocarci- ated (36.5%) and 15 as poorly differentiated (28.9%). noma, particular interest has been focused on MUC1 Patient's data were obtained from clinical records. mucin; in previous publications we have extensively detected MUC1 and associated epitopes in HNSCC and Informed consent was obtained in accordance with the also we have isolated this mucin from larynx primary Hospital Ethical Commission and the World Medical squamous cell carcinoma [6-8]. MUC1 is a large het- Association Declaration of Helsinki: Ethical Principles for erodimeric glycoprotein formed by a highly glycosilated Medical Research Involving Human Subjects (Helsinki, extracellular portion associated to a small cytoplasmic tail Finland, 1964 and their modifications). [9]. From each patient, blood and tumor samples were stud- Studies developed on different carcinoma localizations ied; blood samples were obtained before surgery. Serum such as breast cancer [10,11] have proved that MUC1 aliquots were frozen at -70°C until use while tumor sam- mucin can elicit a humoral immune response; further- ples were immediately fixed in methacarn (60% metha- more, we have detected free and complexed anti-MUC1 nol, 30% chloroform and 10% acetic acid) for two hours antibodies in serum samples belonging to breast cancer and finally placed in 70% ethanol. patients [12]. Controls This report constitutes a detailed statistical study about Eighteen serum samples belonging to normal individuals MUC1 expression and anti-MUC1 immune response sex and age matched were included as a negative control related to different clinical and pathological parameters group in all assays. Ten newly diagnosed pre treatment which may be useful to increase our knowledge to patients with systemic lupus erythematosus (SLE) were develop new anti HNSCC therapeutic strategies based on included as a positive control group for circulating immunological tools. immune complex assays. Antibodies (Abs) Here, we present data that confirm a high tumor MUC1 expression in HNSCC which correlates positively with cir- The following antibodies were employed: C595, an IgG3 culating MUC1. Also, a positive correlation was found mouse monoclonal antibody directed to de APDTR between serum MUC1 versus anti-MUC1 IgG free anti- sequence in the variable number tandem repeats of bodies. MUC1 protein core, kindly provided by Dr. A Murray and Page 2 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Table 1: Patient's clinical data. Patient Age Sex Localization T stage N stage Metastasis Clinical stage Tumor Number differentiation 152 M Larynx 100I Poorly 298 M Larynx 100I Well 373 M Larynx 100I Well 457 M Oral cavity 200IIPoorly 565 F Larynx 200IIWell 668 M Oral cavity 200IIWell 764 F Oral cavity 200IIWell 863 M Pharynx 200IIWell 9 50 M Oral cavity 2 1 0 III Moderately 10 64 M Larynx 3 1 0 III Well 11 48 F Oral cavity 3 1 0 III Poorly 12 47 F Oral cavity 3 1 0 III Well 13 77 M Oral cavity 3 1 0 III Poorly 14 59 M Oral cavity 3 1 0 III Well 15 69 F Oral cavity 3 0 0 III Well 16 56 M Larynx 3 0 0 III Moderately 17 65 M Oral cavity 3 0 0 III Well 18 40 F Oral cavity 3 0 0 III Poorly 19 56 M Pharynx 3 0 0 III Poorly 20 70 M Larynx 3 0 0 III Moderately 21 75 M Pharynx 3 0 0 III Well 22 65 M Oral cavity 3 1 0 III Well 23 60 F Oral cavity 2 1 0 III Moderately 24 56 M Pharynx 400IVA Moderately 25 54 F Pharynx 400IVA Poorly 26 55 M Oral cavity 420IVA Moderately 27 65 M Pharynx 420IVA Moderately 28 58 M Larynx 400IVA Poorly 29 53 M Larynx 400IVA Moderately 30 62 M Oral cavity 420IVA ND 31 29 F Oral cavity 320IVA Moderately 32 60 F Oral cavity 320IVA Well 33 38 M Pharynx 410IVA Poorly 34 62 F Larynx 120IVA Moderately 35 68 M Larynx 320IVA Moderately 36 68 M Oral cavity 410IVA Moderately 37 50 M Oral cavity 320IVA Poorly 38 61 F Pharynx 220IVA Poorly 39 62 M Larynx 320IVA Poorly 40 53 M Larynx 420IVA Well 41 84 F Larynx 400IVA Well 42 50 M Oral cavity 400IVA Poorly 43 55 M Oral cavity 400IVA Moderately 44 73 F Oral cavity 410IVA Poorly 45 71 M Oral cavity 400IVA Moderately 46 49 M Oral cavity 400IVA Well 47 54 M Oral cavity 420IVA Poorly 48 72 M Pharynx 410IVA Moderately 49 59 M Larynx 420IVA Moderately 50 68 M Larynx 430IVB Moderately 51 54 M Larynx 330IVB Moderately 52 61 M Pharynx 430IVB Moderately 53 72 M Oral cavity 331IVC Well Site, sex, clinical stage, T and N stage and tumor differentiation are included. Page 3 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Dr. S Stevenson (Care of Prof. Alan Perkins) University of was discarded and the pellet was resuspended in 2 ml of Nottingham, UK [13] and CT33, a rabbit polyclonal anti- 0.1 N sodium hydroxide. Absorbance was read at 280 nm. body developed after immunization with a synthetic pep- A cut off point obtained from the negative control group tide corresponding to the COOH-terminal 17 amino acids was calculated at 0.229 OD (cut off = 0.171 + 2 SD) [19]. of the intracytoplasmic tail of MUC1, provided by Dr. K. Chul Kim, Lovelace Respiratory Research Institute, Albu- Isolation of circulating immune complexes by protein A sepharose-CL4B affinity chromatography querque, NM, USA [14]. One ml of patient serum was employed following the Immunohistochemistry same procedure described above. Circulating immune Immunohistochemistry was performed according to complexes were resuspended in 1 ml PBS, applied to a 15 standard procedures and as previously reported [15]. Sur- × 1 cm protein A sepharose CL4B column (Sigma, USA) gical specimens were embedded in paraffin, 4 μm sections and incubated for 30 min at 4°C. Unbound material was were cut and placed in bovine serum albumin (BSA) eluted with 15 ml of PBS and 1 ml fractions were col- (Sigma, USA) coated slides. lected. For circulating immune complexes recovery, 5 ml of 50 mM glycine chloride (pH 2.8) was added to the col- Samples were dewaxed in xylol, rehydrated and endog- umn, mixed and incubated for 30 minutes at 4°C fol- enous peroxidase activity was blocked by treatment with lowed by elution with 10 ml of glycine chloride, collecting 0.3% H O (Merck, Germany) in methanol; slides were 1 ml fractions. 2 2 incubated in 10% horse normal serum in phosphate buff- ered saline (PBS) 1% BSA to eliminate background reac- Protein concentration was monitored by absorbance tions and antigenic recuperation was performed by measurement at 280 nm and the protein peak containing heating at 100°C in 10 mM sodium citrate buffer. fractions were thoroughly dialyzed. Slides were incubated overnight with the primary Ab Identification of tumor antigens in circulating immune together with negative controls with PBS at 4°C; after complexes by SDS-PAGE and Western blotting incubation with secondary peroxidase labeled Ab, reac- Electrophoretic separation of proteins [20] was performed with discontinuous 4% stacking-10% resolving polyacry- tion was developed with 3-3'diaminobencidine (Sigma, USA) and H O (Merck, Germany) in PBS. lamide gels in reducing conditions employing a miniVE 2 2 Vertical Electrophoresis System (Amersham Biosciences Finally, sections were counterstained with hematoxylin Corp., USA). Gels were electroblotted into nitrocellulose (Sigma, USA), dehydrated and coverslipped with mount- membranes (Schleicher & Schuell GmbH, Germany), ing media. Samples were evaluated under a light micro- which were blocked with TRIS buffered saline (TBS) 3% scope and a positive reaction was considered when more BSA at 4°C overnight and after five washes with TBS mem- than 5% of the neoplasic cells were stained. Staining branes were incubated with primary antibody at 4°C intensity was scored in a semiquantitative manner [16] overnight [21]. Nitrocellulose membranes were washed and was graded as negative (-), low (+), moderate (++) and incubated with peroxidase conjugated goat anti- and strong (+++). Patterns of reaction were classified as mouse IgG or IgM (1:400) in TBS 3% BSA at room tem- linear (plasma membrane), cytoplasmic, mixed (linear perature for 3 hours. Finally, reaction was developed with O (Merk, and cytoplasmic) and nuclear following other authors 3.3'diaminobencidine (Sigma, USA) and H 2 2 [17,18] with some modifications [15]. Germany) in 10 mM TRIS buffer. Three co-authors (AS-E, MVC and MR) performed a Serum MUC1 determination blinded analysis of all samples included; Prof. Amada Levels of serum MUC1 were measured employing a com- Segal-Eiras is Professor on Pathology (Faculty of Medical mercial enzyme-linked immunosorbent assay (ELISA) fol- Sciences, UNLP, Argentina). lowing manufacturer instructions. The Cancer Associated Serum Assay (CASA) (Medical Innovations Ltd, Australia) Precipitation of circulating immune complexes in utilizes two anti MUC1 core protein monoclonal antibod- polyethylene glycol (PEG) 3,5% ies, BC2 (IgG) as catcher and BC3 (IgM) as tracer, both One hundred μl serum sample were mixed with 900 μl of directed to MUC1 peptide epitope APDTR. Reaction was borate buffer (50 mM boric acid and 5 mM sodium detected by a peroxidase labeled anti-mouse IgM and 2,2'- borate, pH 8.5) and 1000 μl of 7% PEG 6000 (Fluka AG, azino-bis (3-ethylebenzythiazoline-6-sulfonic acid) Switzerland) in borate buffer, incubated overnight at 4°C (ABTS; Sigma, USA). Optical density (OD) was read at and centrifuged at 8000 rpm for 30 minutes at 4°C. Pre- 405 nm, MUC1 levels were extrapolated from a standard cipitates were resuspended in 2 ml of 3.5% PEG and cen- curve and the cut off employed was 4 U/ml. trifuged at 7883 g for 30 minutes at 4°C. The supernatant Page 4 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Serum anti-MUC1 antibodies detection formed with STATISTICA for Windows, StatSoft, Inc. Detection of antibodies reactive with the protein core of (1998), Tulsa, OK, USA. MUC1 was performed by an ELISA described by von Mensdorff-Pouilly [22] with some modifications [23]. Results Briefly, 100 μl/well of 1 μl/ml BSA were dispensed onto MUC1 tumor expression by immunohistochemistry 3912 Falcon III flexible PVC 96 wells plates (BD Labware, Tissue MUC1 expression from 50 HNSCC tumor samples USA) and dried overnight at 37°C. After washing, 50 μl of was studied employing two different anti MUC1 antibod- a 100mer peptide containing five tandem repeats of the ies (Table 2). CT33 Ab, reactive with MUC1 cytoplasmic MUC1 core protein (100 μ/ml) and 50 μl of carbodiimide tail, was present in 40 (80%) HNSCC specimens, 22/25 (1-cycloexil-3-[2 morpholinoethyl] carbodiimide metho- tumors localized in oral cavity, 11/16 in the larynx and 7/ p-sulphonate (Sigma, USA) were dispensed in duplicates. 9 in the pharynx. Reaction pattern was similar to previous For each sample, 2 wells were incubated with carbodiim- reports developed in breast cancer employing the CT2 ide-peptide and two other wells were employed as con- monoclonal antibody directed against the same last 17 aa trols containing 100 μl of blocking buffer (1% BSA PBS) of the MUC1 cytoplasmic tail (Croce et al, 2003). Cell at 4°C overnight. After washing, 150 μl of blocking buffer membrane staining with a linear pattern was observed in was added and incubated at room temperature for 1 hour. 34 (68%) specimens, a cytoplasmic reaction was obtained Plates were washed and diluted human serum samples in in 32 (64%) samples while 27 (54%) had a mixed pattern blocking buffer (1:80 for IgM and 1:40 for IgG determina- (Fig. 1b). Also, a nuclear staining was displayed in 4 (8%) tions) were applied and incubated overnight at 4°C. After specimens. There were no statistically significant differ- washing, 50 μl of peroxidase conjugated goat anti-human ences in MUC1 cytoplasmic tail expression related to IgM 1:400 or IgG 1:600 (Sigma, USA) in PBS Tween-BSA tumor stage, localization or differentiation, although a 0.1% were dispensed and incubated for 1 hour at room clear tendency of negative CT33 reaction was found in temperature. Finally, the reaction was developed with 50 advanced stage tumors while nuclear staining was associ- μl of freshly prepared ABTS in 0.1 M citric acid and 0.2 M ated with poorly differentiated tumor areas. Na PO H (pH 5) with H O . After 1 hour at room tem- 2 4 2 2 perature in the dark, OD was read at 405 nm. Free anti- C595 monoclonal antibody, directed to the protein core MUC1 core antibody levels were calculated as the differ- of MUC1 extracellular domain, showed a positive reac- ence between the averages of carbodiimide-peptide wells tion in 9 samples (18%), 6/25 localized in oral cavity, 2/ and controls. Cut off was 0.340 OD for anti-MUC1 IgM 17 in the larynx and 1/9 in the pharynx. Well differenti- and 0.274 for IgG Abs. ated tumors showed a strong reaction limited to the cyto- plasm and cell membrane and mainly located in areas Statistical analysis with keratin. In moderately and poorly differentiated Multivariate Principal Component Analysis was applied tumors, a reaction was observed along the cytoplasm with to standardized variables and a correlation table was a microvesicular pattern. No correlation was found obtained. Variables (vectors) were plotted against the two between C595 monoclonal antibody reactivity versus principal components extracted (Factor 1 in x axis and tumor stage, node status or tumor differentiation. An Factor 2 in y axis). ANOVA test with post hoc comparisons example of C595 staining is depicted in Fig. 1a. of means (Tukey's test) was employed to find differences MUC1 serum levels among groups; nonparametrical correlations (Kendall's Tau) were applied when necessary. Statistical significance Serum MUC1 levels were measured in 53 HNSCC patients was set to p < 0.05 in all cases; calculations were per- by means of ELISA. MUC1 serum levels were higher in Table 2: Immunohistochemical results. Antibody Positive/Total % C595 (MUC1 Extracellular Domain) 9/50 18 Mixed 4/50 8 Linear 2/50 4 Cytoplasmic 3/50 6 CT33 (MUC1 Cytoplasmic Tail) 40/50 80 Mixed 27/50 54 Linear 34/50 68 Cytoplasmic 32/50 64 Nuclear 4/50 8 Results are expressed as positive cases over total number of cases and percentage, discriminated by immunohistochemical reaction pattern. Page 5 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 a Figure 1 – Immunohistochemical detection of MUC1 extracellular epitopes a – Immunohistochemical detection of MUC1 extracellular epitopes. A stage III larynx carcinoma specimen incu- bated with anti-MUC1 C595 monoclonal antibody is shown. Cell membrane and cytoplasm are stained over all the picture with a microvesicular pattern. Staining is strong inalmost all cells as well as in high proportion of intercellular connections. b – Immunohistochemical detection of MUC1 cytoplasmic tail epitopes. Section of a stage IV larynx carcinoma incubated with CT33 antibody is shown. Different cellular patterns are identified. Intense reaction is observed along the cytoplasm with a granular pattern and linear at nuclear membrane. Plasma membrane is partially stained. Cells are polymorphic as well as nuclei, showing different size and vesicular nuclear patterns. Some intercellular bridges are also stained with this antibody. Page 6 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 cancer patients (mean = 2.46 U/mL, SD = 3.46 U/mL) those patients over MUC1 cut off levels (4 U/mL). This than controls (mean = 0.60 U/mL, SD = 0.69 U/mL); the association was also found by the multivariate Principal percentage of individuals over the cut off value (4 U/mL) Component Analysis; in Fig. 2 the vectors representing was 8/53 (15%) vs 2/18 (10.5%), respectively. Through anti-MUC1 IgG and IgM are drawn opposed, while the multivariate Principal Component Analysis, in HNSCC, a vector that represents serum MUC1 was at the same quad- positive correlation was found between immunohisto- rant of anti-MUC1 IgG. chemistry MUC1 detection and Cancer Associated Serum Assay levels (Kendall τ = 0.275, p = 0.005; Fig. 2). Circulating immune complexes in HNSCC patients Fifty three HNSCC patients presented a mean circulating Anti-MUC1 humoral immune response analysis immune complexes value of 0.224 OD (SD = 0.103 OD) Anti-MUC1 Ab detection statistically different from normal negative controls Free anti-MUC1 serum Ab levels measured in HNSCC (mean = 0.171 OD, SD = 0.029; p = 0.038) and positive were lower than normal sex and age matched controls. controls (SLE patients), mean = 0.456 OD, SD = 0.121 Anti-MUC1 IgM mean value in cancer patients was 0.085 OD; p < 0.001. Tumor stage was statistically associated OD versus 0.194 OD in controls and anti-MUC1 IgG with circulating immune complexes levels (Fig. 4); two mean was 0.097 OD vs 0.171 OD. A negative correlation way ANOVA test with post hoc analysis of means showed between IgG and IgM levels (Kendall τ = -0.309, p = differences between T2 and T3 tumors versus T4 although 0.001) was found. Both anti-MUC1 IgG and anti-MUC1 node invasion did not present any difference. Also, a sta- IgM were statistically correlated with MUC1 serum levels tistical significant difference was found between well dif- (Fig. 3), while IgG levels were detected according to ferentiated and moderately differentiated tumor samples, MUC1 levels, anti-MUC1 IgM levels were very low in which presented a large amount of circulating immune Multivariate Principal Com Figure 2 ponent Analysis of clinical and histopathological data and tissue and serum expression of MUC1 Multivariate Principal Component Analysis of clinical and histopathological data and tissue and serum expres- sion of MUC1. Each variable is represented by vectors, which longitude correspond to the weight of the variable in the model, and their projection to both axes corresponds to the weight of the two main factors extracted (Factor 1 and Factor 2). These two factors explain 39% of the variations observedin the model. This percentage is remarkable taking into account that variables belong to patients. CIC = Circulating immune complexes levels; T = Tumor stage; N = Node status; DIF = tumor dif- ferentiation; CASA = Serum MUC1 levels; IgM Ab = free anti-MUC1 IgM antibodies; IgG Ab = free anti-MUC1 IgG antibodies; C595 = MUC1 extracellular domain expression; CT33 = MUC1 Cytoplasmic tail expression. Page 7 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Correlation Figure 3 between serum MUC1 levels measured by Cancer Associated Serum Assay (U/mL) and free anti-MUC1 antibodies Correlation between serum MUC1 levels measured by Cancer Associated Serum Assay (U/mL) and free anti- MUC1 antibodies. Serum MUC1 levels were plotted in abscises against anti-MUC1 IgG Abs (left coordinates) and IgM Abs (right coordinates). A negative correlation between serum MUC1 levels versus anti-MUC1 IgM Abs (τ = -0.266; p < 0.05) and a positive correlation between serum MUC1 levels and IgG Abs are observed (τ = 0.194; p < 0.05). complexes (p < 0.05) (Fig. 5). On the other hand, heavy chains. No statistical correlation was found although undifferentiated tumors had circulating between MUC1-CIC and tumor or sera MUC1 expression. immune complexes levels similar to well differentiated tumors, no statistical difference was found between poor Principal component analysis correlation studies versus moderately differentiated specimens. Elevated cir- Multivariate analysis of MUC1 expression, anti-MUC1 culating immune complexes levels were found in 16/53 immune response and clinical and pathological parame- patients (30.2%) which belonged to tumor stage I 1/3 ters in HNSCC showed that histological detection with (33%), stage III 3/15 (20%) and stage IV 12/30 (40%). CT33 and C595 Abs were associated with serum MUC1 levels. Moreover, serum MUC1 levels were positively cor- Identification of tumor antigens isolated from circulating related with anti-MUC1 IgG free antibodies and inversely immune complexes correlated with anti-MUC1 IgM. This suggested an associ- Eight out of 16 (50%) serum samples contained MUC1- ation between the presence of MUC1 in serum and the CIC; results are shown in Table 3. In each case, clinical type of antibody response. data have been included. In Fig. 6, results found in two patients are depicted, Western blot employing C595 mon- Vectors representing circulating immune complexes were oclonal antibody showed a double band at about 200 associated with advanced tumor T stage which may indi- kDa. Also, bands at the molecular weight ranging from 25 cate that a large tumor mass produces the MUC1 mucin to 60 kD were observed which correspond to Ig light and that reacts with antibodies, forming circulating immune complexes. Page 8 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Compar Figure 4 ative analysis of circulating immune complexes' levels (OD 280 nm) among different primary tumor T stages Comparative analysis of circulating immune complexes' levels (OD 280 nm) among different primary tumor T stages. Circulating immune complexes' levels in HNSCC increased with tumor T stage (F = 5.375; p < 0.05). Finally, as expected, we found an association between These observations clearly demonstrate the transforming advanced T stage, advanced node status and poor tumor properties of MUC1 overexpression in carcinoma and differentiation. Interestingly, the vectors that represent emphasize the importance of MUC1 cytoplasmic tail sub- these variables were opposed to serum MUC1 and anti- cellular localization. In HNSCC we have found extensive MUC1 IgG antibody levels. expression and cellular localization of MUC1 cytoplasmic tail; large and poorly differentiated tumors were likely to be less reactive with CT33 and C595 Abs. Discussion In our knowledge this is the first report that employs an anti MUC1 cytoplasmic tail Ab in HNSCC; it revealed a When C595 (anti-extracellular MUC1) monoclonal anti- high cellular expression of MUC1 (80% of malignant body was assayed, a lower percentage of reactivity with samples). respect to anti-MUC1 cytoplasmic tail was found; we have previously reported a high reactivity to MUC1 protein In previous studies developed in colon and breast cancer core epitopes in HNSCC [7,8]. The difference among [13,14], we have found similar results. The MUC1 cyto- results was expected since previous reports included a plasmic tail amino acid sequence contains seven tyrosine larger number of tumors localized in oral cavity (58%) in residues constituting intracellular signaling motifs [22] comparison with 49.1% in the present report; we found which play important roles in cell response to external that this localization is reactive with anti-protein core stimuli and is related to different oncogenes and genes monoclonal antibodies [7]. On the other hand, results controlling the cell cycle [23,24][25,26] and β-catenin presented here corresponded to 18.9% of pharyngeal [27]. Furthermore, it has been reported that in vivo, tumors which did not usually react with these mono- MUC1 cytoplasmic tail regulates erbB receptors signaling clonal antibodies; in past reports, only one patient with a through the activation of MAPK pathways [28]. pharyngeal tumor was included [7]. MUC1 cytoplasmic Page 9 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Wester Figure 5 n Blot analysis of circulating immune complexes' fractions eluted and isolated from affinity chromatography Western Blot analysis of circulating immune complexes' fractions eluted and isolated from affinity chromatog- raphy. Circulating immune complexes were obtained from a patient with a stage IV larynx tumor. Lane 1: 3.5% PEG precipi- tate; lane 2: PBS eluate from protein A Sepharose CL-4B column and lane 3: glycine HCl eluate. In lanes 1 and 3 a double band at about 200 kD is observed. tail reactivity is not affected by MUC1 glycosylation and Furthermore, we proved that MUC1-CIC detected by sialylation and, consequently, represents a better indica- Western blot were of the IgG isotype since isolation was tor of cell membrane associated MUC1 [15]. Other performed by protein A Sepharose CL-4B chromatogra- authors [29-31] have described MUC1 RNA splice variants phy, which is known to bind with high affinity to 1, 2 and which do not express the protein core region. Our results 4 IgG classes. also showed that serum MUC1 levels were elevated in 15% of HNSCC samples; interestingly, we found a statis- Presence of high amounts of circulating immune com- tically significant correlation between serum and tumor plexes in cancer patients is not very well understood. Anti- MUC1 detection indicating that this mucin may possibly body formation may represent the onset of an anti tumor be released by the tumor. immune response but, on the other hand, may be the con- sequence of tumor immune evasion related to a defective In this research we found anti-MUC1 antibodies; also, we cellular immune response. Nonetheless, circulating observed that serum MUC1 levels were positively corre- immune complexes have been associated with tumor pro- lated with free anti-MUC1 IgG, which may indicate that gression and also, their formation clearly affects the cor- an anti-MUC1 immune response is mounted. Tumor rect evaluation of several tumor markers. derived MUC1 is known to induce cellular and humoral specific immune responses in cancer patients [32-34]. In Only a few reports in the literature have described the this context, it is possible to speculate that a specific anti- antigenic component of circulating immune complexes in MUC1 immune response is induced in HNSCC patients. HNSCC. Vlock DR and others [36] have exhaustively stud- ied the reactivity of serum Abs against autologous tumor In another report [11], we pointed out that low MUC1 cell lines. These authors found that some patients pre- serum levels in stage I breast cancer patients were associ- sented high Ab titers which occasionally cross reacted ated with the presence of free and complexed anti-MUC1 with tumor cell lines derived from other histogenesis. antibodies. Similarly, circulating anti-MUC1-IgG anti- body levels were found predictive of survival in breast [35] In accordance with other reports [37,38], we found higher and pancreatic cancer patients [34]. We have found that circulating immune complexes levels in HNSCC patients anti-MUC1 IgG Ab correlated negatively with poor in comparison with normal subjects; furthermore, we also HNSCC differentiation and disease stage which are the agree with these authors that elevated circulating immune main predictors for patient outcome in this localization. complexes levels were detected in advanced tumor stage. They pointed out that high circulating immune complexes levels in HNSCC patients is the result of increased tumor Page 10 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Table 3: Data from HNSCC patients' serum employed to isolate circulating immune complexes. Sample Localization Stage Sex Age Differentiation MUC1 In MUC1 in MUC1-CT in CIC levels MUC1 serum levels (U/mL) CIC IHQ IHQ (OD) 1 Larynx I M 52 Poorly + - - 0.306 3.9 2 Larynx III M 64 Well + - - 0.371 1.4 3 Larynx III M 56 Moderately + - + 0.320 1.2 4 Larynx III M 70 Moderately + - + 0.229 0.9 5 Larynx IVA M 58 Poorly - - + 0.411 2.2 6 Larynx IVA M 53 Moderately - - + 0.312 1.7 7 Oral Cavity IVA M 55 Moderately - + + 0.322 4.7 8 Oral Cavity IVA F 29 Moderately + - + 0.258 1.7 9 Oral Cavity IVA M 68 Moderately + - + 0.323 1.1 10 Oral Cavity IVA M 55 Moderately - - - 0.477 0.6 11 Oral Cavity IVA M 49 Well - - + 0.403 0.7 12 Pharynx IVA M 56 Moderately - - - 0.497 5.3 13 Pharynx IVA M 65 Moderately + - + 0.335 1.9 14 Pharynx IVA M 38 Poorly + + + 0.231 1.8 15 Pharynx IVA M 72 Moderately - - - 0.473 0.4 16 Pharynx IVB M 61 Moderately - - + 0.349 1.5 Positive/Total or 8/16 2/16 11/16 0.351 2.0 mean value Clinical data with immunohistochemical detection and serum determination of MUC1, anti-MUC1 Abs and circulating immune complexes are depicted. mass which would mediate changes in anti tumor immu- ated with advanced tumor stage; furthermore, the vectors nity. Moreover, Das TK et al [39] found that circulating that represent these variables did not positively correlate immune complexes persist after surgical excision of the to serum MUC1 and anti-MUC1 IgG antibody levels, and primary tumor due to the presence of remaining tumor also were inversely correlated with MUC1 tumor expres- tissue or occult metastasis. sion. Considering that we identified MUC1 as an anti- genic component of circulating immune complexes, taken MUC1 is a mucin that has been detected both in the cyto- together, these results indicate that a large tumor mass plasm, plasma membrane and nucleus in different tumor produces high MUC1 mucin that is liberated to circula- localizations. MUC1 is synthesized in the ER and glyco- tion and captured by IgG antibodies forming MUC1-CIC. sylated in the Golgi apparatus but it has been described that it suffers several glycosylation cycles being expressed Another interesting conclusion is that poorly differenti- in the plasma membrane several times [40]. Furthermore, ated tumors are inversely correlated with tumor and different splicing variants have been described that can be serum MUC1 detection. This finding suggests the possibil- present in the cytoplasm and in the plasma membrane, as ity to take into account its consideration regarding prog- well [31,41-43]. Wen et al [44] investigated intracellular nosis and follow up of this disease. trafficking of MUC1 cytoplasmic tail in human pancreatic cancer cell lines S2-013 and Panc-1 and detected this frac- Competing interests tion at the inner cell surface, in the cytosol and in the The author(s) declare that they have no competing inter- nucleus. They hypothesized that the association between ests. β-catenin and fragments of the MUC1 cytoplasmic tail facilitated the cytosol-to-nuclear translocation of β-cat- Authors' contributions enin and contributed to its nuclear accumulation. We MER carried out the studies and drafted the manuscript. have found [14] MUC1 cytoplasmic tail and protein core MVC conceived the study, participated in the design and expression in the plasma membrane, cytoplasm and coordination and wrote the manuscript. AP obtained the nucleus in breast and colorectal tumor samples. samples and clinical records and also participated in the design of the manuscript. ASE participated in the design and coordination of the study and helped to draft the Conclusion In HNSCC, Principal Component Analysis (Fig. 2) manuscript. All authors read and approved the final man- showed that circulating immune complexes were associ- uscript. Page 11 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 17. Renkonen J, Paavonen T, Renkonen R: Endothelial and epithelial Acknowledgements expression of sialyl Lewis(x) and sialyl Lewis(a) in lesions of The authors whish to thank Dr. A Murray and Dr. S Stevenson for provid- breast carcinoma. Int J Cancer 1997, 74:296-300. ing C595 monoclonal antibody (care of Prof. Alan Perkins, Nottingham, 18. Luna-More S, Rius F, Weil B, Jimenez A, Bautista MD, Perez-Mellado UK), Dr. K. Chul Kim (Lovelace Respiratory Research Institute, Albu- A: EMA: a differentiation antigen related to node metastatic capacity of breast carcinomas. Pathol Res Pract 2001, 197:419-25. querque, NM, USA) for generously providing CT33 polyclonal antibody. 19. Croce MV, Price MR, Segal-Eiras A: Expression of monoclonal- The authors are also extremely grateful to Lic. S Demichelis for performing antibody-defined antigens in fractions isolated from human statistical analysis, Dr. J Carri for encourage in photographical techniques breast carcinomas and patients' serum. Cancer Immunol Immu- and JC Molina for skillful technical assistance. nother 1995, 40:132-7. 20. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, Prof. Amada Segal-Eiras and Prof. María Virginia Croce are members of the 227:680-5. research career of the Consejo Nacional de Investigaciones Científicas y 21. Towbin H, Staehelin T, Gordon J: Electrophoretic transfer of Técnicas (CONICET) and the Comisión de Investigaciones Científicas de la proteins from polyacrylamide gels to nitrocellulose sheets: Provincia de Buenos Aires (CIC/PBA), respectively. Dr. Martin E. Rabassa procedure and some applications. Proc Natl Acad Sci USA 1979, 76:4350-4. is a fellow of CONICET. Financial support from FONCYT, CONICET, 22. Von Mensdorff-Pouilly S, Gouretvich MM, Kenemans P: An enzyme- CICPBA and University of La Plata (Argentina) is very much appreciated. linked immunosorbent assay for the measurement of circu- lating antibodies to polymorphic epithelial mucin (MUC1). References Tumor Biol 1989:186-195. 23. Croce MV, Isla-Larrain MT, Capafons A, Price MR, Segal-Eiras A: 1. Döbrössy L: Epidemiology of head and neck cancer: magni- Humoral immune response induced by the protein core of tude of the problem. Cancer and Metastasis Reviews 2005, 24:9-17. MUC1 mucin in pregnant and healthy women. Breast Cancer 2. Shermann CD Jr: Cancer of the head and neck. In Manual of Clin- Res Treat 2001, 69:1-11. ical Oncology UICC, Springer; 1990. 24. Wang H, Lillehoj EP, Kim KC: Identification of four sites of stim- 3. International Agency for research in Cancer, GLOBOCAN ulated tyrosine phosphorylation in the MUC1 cytoplasmic 2002 [http://www-dep.iarc.fr] tail. Biochem Biophys Res Commun 2003, 310:341-6. 4. Stewart B, Kleihues P, editors: World cancer report. Lyon: IARC 25. Zrihan-Licht S, Baruch A, Elroy-Stein O, Keydar I, Wreschner DH: Press; 2003. Tyrosine phosphorylation of the MUC1 breast cancer mem- 5. Dabelsteen E, Gao S: ABO blood-group antigens in oral cancer. brane proteins. Cytokine receptor-like molecules. FEBS Lett J Dent Res 2005, 84:21-8. 1994, 356:130-6. 6. Croce MV, Price MR, Segal-Eiras A: Detection and isolation of 26. Pandey P, Kharbanda S, Kufe D: Association of the DF3/MUC1 MUC1 mucin from larynx squamous cell carcinoma. Pathol breast cancer antigen with Grb2 and the Sos/Ras exchange Oncol Res 2000, 6:93-9. protein. Cancer Res 1995, 55:4000-3. 7. Croce MV, Rabassa ME, Price MR, Segal-Eiras A: MUC1 mucin and 27. Yamamoto M, Bharti A, Li Y, Kufe D: Interaction of the DF3/ carbohydrate associated antigens as tumor markers in head MUC1 breast carcinoma-associated antigen and beta-cat- and neck squamous cell carcinoma. Pathol Oncol Res 2001, enin in cell adhesion. J Biol Chem 1997, 272:12492-4. 7:284-91. 28. Schroeder JA, Thompson MC, Gardner MM, Gendler SJ: Transgenic 8. Croce MV, Rabassa ME, Pereyra A, Segal-Eiras A: Influence of sialic MUC1 interacts with epidermal growth factor receptor and acid removal on MUC1 antigenic reactivity in head and neck correlates with mitogen-activated protein kinase activation carcinoma. Pathol Oncol Res 2005, 11:74-81. in the mouse mammary gland. J Biol Chem 2001, 276:13057-64. 9. Taylor-Papadimitriou J, Burchell JM, Plunkett T, Graham R, Correa I, 29. Wreschner DH, Hareuveni M, Tsarfaty I, Smorodinsky N, Horev J, Miles D, Smith M: MUC1 and the immunobiology of cancer. J Zaretsky I, Kotkes P, Weiss M, Lathe R, Dion AS, Keydar I: Human Mammary Gland Biol Neoplasia 2002, 7:209-21. epithelial tumor antigen cDNA sequences – differential splic- 10. Kotera Y, Fontenot JD, Pecher G, Metzgar RS, Finn OJ: Humoral ing may generate multiple protein forms. Eur J Biochem 1990, immunity against a tandem repeat epitope of human mucin 189:463-73. MUC-1 in sera from breast, pancreatic, and colon cancer 30. Zrihan-Licht S, Vos HL, Baruch A, Elroy-Stein O, Sagiv D, Keydar I, patients. Cancer Res 1994, 54:2856-60. Hilkens J, Wreschner DH: Characterization and molecular 11. von Mensdorff-Pouilly S, Gourevitch MM, Kenemans P, Verstraeten cloning of a novel MUC1 protein, devoid of tandem repeats, AA, Litvinov SV, van Kamp GJ, Meijer S, Vermorken J, Hilgers J: expressed in human breast cancer tissue. Eur J Biochem 1994, Humoral immune response to polymorphic epithelial mucin 224:787-95. (MUC-1) in patients with benign and malignant breast 31. Obermair A, Schmid BC, Stimpfl M, Fasching B, Preyer O, Leodolter tumours. Eur J Cancer 1996, 32A:1325-31. S, Crandon AJ, Zeillinger R: Novel MUC1 splice variants are 12. Croce MV, Isla-Larrain MT, Demichelis SO, Gori JR, Price MR, Segal- expressed in cervical carcinoma. Gynecol Oncol 2001, 83:343-7. Eiras A: Tissue and serum MUC1 mucin detection in breast 32. Agrawal B, Reddish MA, Longenecker BM: In vitro induction of cancer patients. Breast Cancer Res Treat 2003, 81:195-207. MUC-1 peptide-specific type 1 T lymphocyte and cytotoxic 13. Price MR, Pugh JA, Hudecz F, Griffiths W, Jacobs E, Symonds IM, T lymphocyte responses from healthy multiparous donors. J Clarke AJ, Chan WC, Baldwin RW: C595 – a monoclonal anti- Immunol 1996, 157:2089-95. body against the protein core of human urinary epithelial 33. Petrarca C, Rughetti A, Rahimi H, D'Agostini F, Turchi V, Apollonj mucin commonly expressed in breast carcinomas. Br J Cancer Ghetti C, Scambia G, Frati L, Nuti M: Human antibodies against 1990, 61:681-6. the polymorphic epithelial mucin in ovarian cancer patients 14. Croce MV, Isla-Larrain MT, Remes-Lenicov F, Colussi AG, Lacunza E, recognise a novel sequence in the tandem repeat region. Eur Kim KC, Gendler SJ, Segal-Eiras A: MUC1 cytoplasmic tail detec- J Cancer 1996, 32A:2155-63. tion using CT33 polyclonal and CT2 monoclonal antibodies 34. Hamanaka Y, Suehiro Y, Fukui M, Shikichi K, Imai K, Hinoda Y: Cir- in breast and colorectal tissue. Histol Histopathol 2006, culating anti-MUC1 IgG antibodies as a favorable prognostic 21:849-55. factor for pancreatic cancer. Int J Cancer 2003, 103:97-100. 15. Croce MV, Isla-Larrain MT, Rua CE, Rabassa ME, Gendler SJ, Segal- 35. von Mensdorff-Pouilly S, Verstraeten AA, Kenemans P, Snijdewint FG, Eiras A: Patterns of MUC1 tissue expression defined by an Kok A, Van Kamp GJ, Paul MA, Van Diest PJ, Meijer S, Hilgers J: Sur- anti-MUC1 cytoplasmic tail monoclonal antibody in breast vival in early breast cancer patients is favorably influenced by cancer. J Histochem Cytochem 2003, 51:781-8. a natural humoral immune response to polymorphic epithe- 16. Feickert H, Anger B, Cordón-Cardo C, Lloyd K: Cell-surface anti- lial mucin. J Clin Oncol 2000, 18:574-83. gens of human tumors detected by mouse monoclonal anti- 36. Vlock DR, Scalise D, Schwartz DR, Richter DE, Krause CJ, Baker SR, bodies: definition of blood-group- and non-blood-group- Carey TE: Incidence of serum antibody reactivity to autolo- related antigenic systems. Int J Cancer 1990, 46:1007-1013. gous head and neck cancer cell lines and augmentation of Page 12 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 antibody reactivity following acid dissociation and ultrafiltra- tion. Cancer Res 1989, 49:1361-5. 37. Yamanaka N, Himi T, Harabuchi Y, Hoki K, Kataura A: Soluble immune complexes and squamous cell carcinoma-related antigens in patients with head and neck cancer. Cancer 1988, 62:1932-8. 38. Wolf GT, Wolfe RA: Circulating immune complexes in patients with nasopharyngeal carcinoma. Laryngoscope 1990, 100:302-8. 39. Das TK, Aziz M, Rattan A, Sherwani R: Prognostic significance of circulating immune complexes in malignant tumours of head and neck. J Indian Med Assoc 1995, 93:3-7. 40. Litvinov SV, Hilkens J: The epithelial sialomucin, episialin, is sia- lylated during recycling. J Biol Chem 1993, 268:21364-71. 41. Ligtenberg MJ, Vos HL, Gennissen AM, Hilkens J: Episialin, a carci- noma-associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini. J Biol Chem 1990, 265:5573-8. 42. Williams CJ, Wreschner DH, Tanaka A, Tsarfaty I, Keydar I, Dion AS: Multiple protein forms of the human breast tumor-associ- ated epithelial membrane antigen (EMA) are generated by alternative splicing and induced by hormonal stimulation. Biochem Biophys Res Commun 1990, 170:1331-8. 43. Levitin F, Stern O, Weiss M, Gil-Henn C, Ziv R, Prokocimer Z, Smorodinsky NI, Rubinstein DB, Wreschner DH: The MUC1 SEA module is a self-cleaving domain. J Biol Chem 2005, 280:33374-86. 44. Wen Y, Caffrey TC, Wheelock MJ, Johnson KR, Hollingsworth MA: Nuclear association of the cytoplasmic tail of MUC1 and beta-catenin. J Biol Chem 2003, 278:38029-39. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2407/6/253/pre pub Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 13 of 13 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Cancer Springer Journals

MUC1 expression and anti-MUC1 serum immune response in head and neck squamous cell carcinoma (HNSCC): a multivariate analysis

Download PDF
13 pages

Loading next page...
 
/lp/springer-journals/muc1-expression-and-anti-muc1-serum-immune-response-in-head-and-neck-W0kB6mIxgN

References (52)

Publisher
Springer Journals
Copyright
2006 Rabassa et al; licensee BioMed Central Ltd.
ISSN
1471-2407
eISSN
1471-2407
DOI
10.1186/1471-2407-6-253
Publisher site
See Article on Publisher Site

Abstract

Background: HNSCC progression to adjacent tissue and nodes may be mediated by altered glycoproteins and glycolipids such as MUC1 mucin. This report constitutes a detailed statistical study about MUC1 expression and anti-MUC1 immune responses in relation to different clinical and pathological parameters which may be useful to develop new anti HNSCC therapeutic strategies. Patients and methods: Fifty three pre treatment HNSCC patients were included: 26 (49.1%) bearing oral cavity tumors, 17 (32.1%) localized in the larynx and 10 (18.8%) in the pharynx. Three patients (5.7%) were at stage I, 5 (9.4%) stage II, 15 (28.3%) stage III and 30 (56.6%) at stage IV. MUC1 tumor expression was studied by immunohistochemistry employing two anti-MUC1 antibodies: CT33, anti cytoplasmic tail MUC1 polyclonal antibody (Ab) and C595 anti-peptidic core MUC1 monoclonal antibody. Serum levels of MUC1 and free anti-MUC1 antibodies were detected by ELISA and circulating immune complexes (CIC) by precipitation in polyethylene glycol (PEG) 3.5%; MUC1 isolation from circulating immune complexes was performed by protein A-sepharose CL-4B affinity chromatography followed by SDS-PAGE and Western blot. Statistical analysis consisted in Multivariate Principal Component Analysis (PCA); ANOVA test (Tukey's test) was employed to find differences among groups; nonparametrical correlations (Kendall's Tau) were applied when necessary. Statistical significance was set to p < 0.05 in all cases. Results: MUC1 cytoplasmic tail was detected in 40/50 (80%) and MUC1 protein core in 9/50 (18%) samples while serum MUC1 levels were elevated in 8/53 (15%) patients. A significant statistical correlation was found between MUC1 serum levels and anti-MUC1 IgG free antibodies, while a negative correlation between MUC1 serum levels and anti-MUC1 IgM free antibodies was found. Circulating immune complexes were elevated in 16/53 (30%) samples and were also statistically associated with advanced tumor stage. MUC1 was identified as an antigenic component of IgG circulating immune complexes. Moreover, poorly differentiated tumors were inversely correlated with tumor and serum MUC1 detection and positively correlated with node involvement and tumor mass. Conclusion: Possibly, tumor cells produce MUC1 mucin which is liberated to the circulation and captured by IgG antibodies forming MUC1-IgG-CIC. Another interesting conclusion is that poorly differentiated tumors are inversely correlated with tumor and serum MUC1 detection. Page 1 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 In addition, circulating immune complexes levels were Background In western countries, head and neck squamous cell carci- statistically associated with tumor size, inversely associ- noma (HNSCC) occupies the 5th place in frequency and ated with MUC1 tumor expression and were not posi- also is the 5th cause of mortality due to cancer. HNSCC tively associated with serum MUC1 and free anti-MUC1 localization consists about 40% in the oral cavity, 15% in IgG. Finally, tumor size, node involvement and poor dif- the pharynx, 25% in the larynx and the rest sites such as ferentiation were positively associated. salivary glands and thyroid [1,2]. Data obtained from argentinian records show some differences compared Methods with data found in other countries since the most frequent Patients localization has been larynx (1665/3127, 53%) followed Fifty three pre treatment HNSCC patients from the "Hos- by oral cavity (1035/3127, 33%) and finally, pharynx pital General Interzonal de Agudos Gral. San Martín", La showing 427/3127, 14%. In 2002, the total number of Plata, Argentina were included in this study (Table 1); all HNSCC cases informed in Argentina was 3127 (16.9/ tumors were primary; patients with previous history of 100000) inhabitants [3]. HNSCC or who had received preoperative therapy were not included. Patients were clinically categorized accord- Worldwide, delayed diagnosis is common and frequently ing to the American Joint Cancer Committee (AJCC, Can- very scarce improvement in five-year survival over the last cer Staging Manual, 2002). Sex and age distribution of four decades has been observed [4]. patients was: 39 males (73.6%) and 14 (26.4%) females with a mean age of 60.67 (11.55 SD) years and a range HNSCC mainly progress to adjacent tissue and nodes from 29 to 98 years. Twenty six patients (49.1%) had their while distant metastasis is a late event. The ability of primary tumor located in the oral cavity, 17 (32.0%) in tumor cells to invade is an acquired and progressive phe- the larynx and 10 (18.9%) in the pharynx. Three patients nomenon mediated, in many cases, by the alteration of (5.7%) were at stage I, 5 (9.4%) at stage II, 15 (28.3%) at membrane glycoproteins such as mucins. Dabelsteen and stage III and 30 (56.6%) at stage IV; stages IVA, IVB and Gao [5] proposed that the presence of different glycosyla- IVC cases where also included. Eighteen were classified as tion patterns modulate the behavior of these membrane well differentiated (34.6%), 19 as moderately differenti- glycoproteins involved in cell signaling. In adenocarci- ated (36.5%) and 15 as poorly differentiated (28.9%). noma, particular interest has been focused on MUC1 Patient's data were obtained from clinical records. mucin; in previous publications we have extensively detected MUC1 and associated epitopes in HNSCC and Informed consent was obtained in accordance with the also we have isolated this mucin from larynx primary Hospital Ethical Commission and the World Medical squamous cell carcinoma [6-8]. MUC1 is a large het- Association Declaration of Helsinki: Ethical Principles for erodimeric glycoprotein formed by a highly glycosilated Medical Research Involving Human Subjects (Helsinki, extracellular portion associated to a small cytoplasmic tail Finland, 1964 and their modifications). [9]. From each patient, blood and tumor samples were stud- Studies developed on different carcinoma localizations ied; blood samples were obtained before surgery. Serum such as breast cancer [10,11] have proved that MUC1 aliquots were frozen at -70°C until use while tumor sam- mucin can elicit a humoral immune response; further- ples were immediately fixed in methacarn (60% metha- more, we have detected free and complexed anti-MUC1 nol, 30% chloroform and 10% acetic acid) for two hours antibodies in serum samples belonging to breast cancer and finally placed in 70% ethanol. patients [12]. Controls This report constitutes a detailed statistical study about Eighteen serum samples belonging to normal individuals MUC1 expression and anti-MUC1 immune response sex and age matched were included as a negative control related to different clinical and pathological parameters group in all assays. Ten newly diagnosed pre treatment which may be useful to increase our knowledge to patients with systemic lupus erythematosus (SLE) were develop new anti HNSCC therapeutic strategies based on included as a positive control group for circulating immunological tools. immune complex assays. Antibodies (Abs) Here, we present data that confirm a high tumor MUC1 expression in HNSCC which correlates positively with cir- The following antibodies were employed: C595, an IgG3 culating MUC1. Also, a positive correlation was found mouse monoclonal antibody directed to de APDTR between serum MUC1 versus anti-MUC1 IgG free anti- sequence in the variable number tandem repeats of bodies. MUC1 protein core, kindly provided by Dr. A Murray and Page 2 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Table 1: Patient's clinical data. Patient Age Sex Localization T stage N stage Metastasis Clinical stage Tumor Number differentiation 152 M Larynx 100I Poorly 298 M Larynx 100I Well 373 M Larynx 100I Well 457 M Oral cavity 200IIPoorly 565 F Larynx 200IIWell 668 M Oral cavity 200IIWell 764 F Oral cavity 200IIWell 863 M Pharynx 200IIWell 9 50 M Oral cavity 2 1 0 III Moderately 10 64 M Larynx 3 1 0 III Well 11 48 F Oral cavity 3 1 0 III Poorly 12 47 F Oral cavity 3 1 0 III Well 13 77 M Oral cavity 3 1 0 III Poorly 14 59 M Oral cavity 3 1 0 III Well 15 69 F Oral cavity 3 0 0 III Well 16 56 M Larynx 3 0 0 III Moderately 17 65 M Oral cavity 3 0 0 III Well 18 40 F Oral cavity 3 0 0 III Poorly 19 56 M Pharynx 3 0 0 III Poorly 20 70 M Larynx 3 0 0 III Moderately 21 75 M Pharynx 3 0 0 III Well 22 65 M Oral cavity 3 1 0 III Well 23 60 F Oral cavity 2 1 0 III Moderately 24 56 M Pharynx 400IVA Moderately 25 54 F Pharynx 400IVA Poorly 26 55 M Oral cavity 420IVA Moderately 27 65 M Pharynx 420IVA Moderately 28 58 M Larynx 400IVA Poorly 29 53 M Larynx 400IVA Moderately 30 62 M Oral cavity 420IVA ND 31 29 F Oral cavity 320IVA Moderately 32 60 F Oral cavity 320IVA Well 33 38 M Pharynx 410IVA Poorly 34 62 F Larynx 120IVA Moderately 35 68 M Larynx 320IVA Moderately 36 68 M Oral cavity 410IVA Moderately 37 50 M Oral cavity 320IVA Poorly 38 61 F Pharynx 220IVA Poorly 39 62 M Larynx 320IVA Poorly 40 53 M Larynx 420IVA Well 41 84 F Larynx 400IVA Well 42 50 M Oral cavity 400IVA Poorly 43 55 M Oral cavity 400IVA Moderately 44 73 F Oral cavity 410IVA Poorly 45 71 M Oral cavity 400IVA Moderately 46 49 M Oral cavity 400IVA Well 47 54 M Oral cavity 420IVA Poorly 48 72 M Pharynx 410IVA Moderately 49 59 M Larynx 420IVA Moderately 50 68 M Larynx 430IVB Moderately 51 54 M Larynx 330IVB Moderately 52 61 M Pharynx 430IVB Moderately 53 72 M Oral cavity 331IVC Well Site, sex, clinical stage, T and N stage and tumor differentiation are included. Page 3 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Dr. S Stevenson (Care of Prof. Alan Perkins) University of was discarded and the pellet was resuspended in 2 ml of Nottingham, UK [13] and CT33, a rabbit polyclonal anti- 0.1 N sodium hydroxide. Absorbance was read at 280 nm. body developed after immunization with a synthetic pep- A cut off point obtained from the negative control group tide corresponding to the COOH-terminal 17 amino acids was calculated at 0.229 OD (cut off = 0.171 + 2 SD) [19]. of the intracytoplasmic tail of MUC1, provided by Dr. K. Chul Kim, Lovelace Respiratory Research Institute, Albu- Isolation of circulating immune complexes by protein A sepharose-CL4B affinity chromatography querque, NM, USA [14]. One ml of patient serum was employed following the Immunohistochemistry same procedure described above. Circulating immune Immunohistochemistry was performed according to complexes were resuspended in 1 ml PBS, applied to a 15 standard procedures and as previously reported [15]. Sur- × 1 cm protein A sepharose CL4B column (Sigma, USA) gical specimens were embedded in paraffin, 4 μm sections and incubated for 30 min at 4°C. Unbound material was were cut and placed in bovine serum albumin (BSA) eluted with 15 ml of PBS and 1 ml fractions were col- (Sigma, USA) coated slides. lected. For circulating immune complexes recovery, 5 ml of 50 mM glycine chloride (pH 2.8) was added to the col- Samples were dewaxed in xylol, rehydrated and endog- umn, mixed and incubated for 30 minutes at 4°C fol- enous peroxidase activity was blocked by treatment with lowed by elution with 10 ml of glycine chloride, collecting 0.3% H O (Merck, Germany) in methanol; slides were 1 ml fractions. 2 2 incubated in 10% horse normal serum in phosphate buff- ered saline (PBS) 1% BSA to eliminate background reac- Protein concentration was monitored by absorbance tions and antigenic recuperation was performed by measurement at 280 nm and the protein peak containing heating at 100°C in 10 mM sodium citrate buffer. fractions were thoroughly dialyzed. Slides were incubated overnight with the primary Ab Identification of tumor antigens in circulating immune together with negative controls with PBS at 4°C; after complexes by SDS-PAGE and Western blotting incubation with secondary peroxidase labeled Ab, reac- Electrophoretic separation of proteins [20] was performed with discontinuous 4% stacking-10% resolving polyacry- tion was developed with 3-3'diaminobencidine (Sigma, USA) and H O (Merck, Germany) in PBS. lamide gels in reducing conditions employing a miniVE 2 2 Vertical Electrophoresis System (Amersham Biosciences Finally, sections were counterstained with hematoxylin Corp., USA). Gels were electroblotted into nitrocellulose (Sigma, USA), dehydrated and coverslipped with mount- membranes (Schleicher & Schuell GmbH, Germany), ing media. Samples were evaluated under a light micro- which were blocked with TRIS buffered saline (TBS) 3% scope and a positive reaction was considered when more BSA at 4°C overnight and after five washes with TBS mem- than 5% of the neoplasic cells were stained. Staining branes were incubated with primary antibody at 4°C intensity was scored in a semiquantitative manner [16] overnight [21]. Nitrocellulose membranes were washed and was graded as negative (-), low (+), moderate (++) and incubated with peroxidase conjugated goat anti- and strong (+++). Patterns of reaction were classified as mouse IgG or IgM (1:400) in TBS 3% BSA at room tem- linear (plasma membrane), cytoplasmic, mixed (linear perature for 3 hours. Finally, reaction was developed with O (Merk, and cytoplasmic) and nuclear following other authors 3.3'diaminobencidine (Sigma, USA) and H 2 2 [17,18] with some modifications [15]. Germany) in 10 mM TRIS buffer. Three co-authors (AS-E, MVC and MR) performed a Serum MUC1 determination blinded analysis of all samples included; Prof. Amada Levels of serum MUC1 were measured employing a com- Segal-Eiras is Professor on Pathology (Faculty of Medical mercial enzyme-linked immunosorbent assay (ELISA) fol- Sciences, UNLP, Argentina). lowing manufacturer instructions. The Cancer Associated Serum Assay (CASA) (Medical Innovations Ltd, Australia) Precipitation of circulating immune complexes in utilizes two anti MUC1 core protein monoclonal antibod- polyethylene glycol (PEG) 3,5% ies, BC2 (IgG) as catcher and BC3 (IgM) as tracer, both One hundred μl serum sample were mixed with 900 μl of directed to MUC1 peptide epitope APDTR. Reaction was borate buffer (50 mM boric acid and 5 mM sodium detected by a peroxidase labeled anti-mouse IgM and 2,2'- borate, pH 8.5) and 1000 μl of 7% PEG 6000 (Fluka AG, azino-bis (3-ethylebenzythiazoline-6-sulfonic acid) Switzerland) in borate buffer, incubated overnight at 4°C (ABTS; Sigma, USA). Optical density (OD) was read at and centrifuged at 8000 rpm for 30 minutes at 4°C. Pre- 405 nm, MUC1 levels were extrapolated from a standard cipitates were resuspended in 2 ml of 3.5% PEG and cen- curve and the cut off employed was 4 U/ml. trifuged at 7883 g for 30 minutes at 4°C. The supernatant Page 4 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Serum anti-MUC1 antibodies detection formed with STATISTICA for Windows, StatSoft, Inc. Detection of antibodies reactive with the protein core of (1998), Tulsa, OK, USA. MUC1 was performed by an ELISA described by von Mensdorff-Pouilly [22] with some modifications [23]. Results Briefly, 100 μl/well of 1 μl/ml BSA were dispensed onto MUC1 tumor expression by immunohistochemistry 3912 Falcon III flexible PVC 96 wells plates (BD Labware, Tissue MUC1 expression from 50 HNSCC tumor samples USA) and dried overnight at 37°C. After washing, 50 μl of was studied employing two different anti MUC1 antibod- a 100mer peptide containing five tandem repeats of the ies (Table 2). CT33 Ab, reactive with MUC1 cytoplasmic MUC1 core protein (100 μ/ml) and 50 μl of carbodiimide tail, was present in 40 (80%) HNSCC specimens, 22/25 (1-cycloexil-3-[2 morpholinoethyl] carbodiimide metho- tumors localized in oral cavity, 11/16 in the larynx and 7/ p-sulphonate (Sigma, USA) were dispensed in duplicates. 9 in the pharynx. Reaction pattern was similar to previous For each sample, 2 wells were incubated with carbodiim- reports developed in breast cancer employing the CT2 ide-peptide and two other wells were employed as con- monoclonal antibody directed against the same last 17 aa trols containing 100 μl of blocking buffer (1% BSA PBS) of the MUC1 cytoplasmic tail (Croce et al, 2003). Cell at 4°C overnight. After washing, 150 μl of blocking buffer membrane staining with a linear pattern was observed in was added and incubated at room temperature for 1 hour. 34 (68%) specimens, a cytoplasmic reaction was obtained Plates were washed and diluted human serum samples in in 32 (64%) samples while 27 (54%) had a mixed pattern blocking buffer (1:80 for IgM and 1:40 for IgG determina- (Fig. 1b). Also, a nuclear staining was displayed in 4 (8%) tions) were applied and incubated overnight at 4°C. After specimens. There were no statistically significant differ- washing, 50 μl of peroxidase conjugated goat anti-human ences in MUC1 cytoplasmic tail expression related to IgM 1:400 or IgG 1:600 (Sigma, USA) in PBS Tween-BSA tumor stage, localization or differentiation, although a 0.1% were dispensed and incubated for 1 hour at room clear tendency of negative CT33 reaction was found in temperature. Finally, the reaction was developed with 50 advanced stage tumors while nuclear staining was associ- μl of freshly prepared ABTS in 0.1 M citric acid and 0.2 M ated with poorly differentiated tumor areas. Na PO H (pH 5) with H O . After 1 hour at room tem- 2 4 2 2 perature in the dark, OD was read at 405 nm. Free anti- C595 monoclonal antibody, directed to the protein core MUC1 core antibody levels were calculated as the differ- of MUC1 extracellular domain, showed a positive reac- ence between the averages of carbodiimide-peptide wells tion in 9 samples (18%), 6/25 localized in oral cavity, 2/ and controls. Cut off was 0.340 OD for anti-MUC1 IgM 17 in the larynx and 1/9 in the pharynx. Well differenti- and 0.274 for IgG Abs. ated tumors showed a strong reaction limited to the cyto- plasm and cell membrane and mainly located in areas Statistical analysis with keratin. In moderately and poorly differentiated Multivariate Principal Component Analysis was applied tumors, a reaction was observed along the cytoplasm with to standardized variables and a correlation table was a microvesicular pattern. No correlation was found obtained. Variables (vectors) were plotted against the two between C595 monoclonal antibody reactivity versus principal components extracted (Factor 1 in x axis and tumor stage, node status or tumor differentiation. An Factor 2 in y axis). ANOVA test with post hoc comparisons example of C595 staining is depicted in Fig. 1a. of means (Tukey's test) was employed to find differences MUC1 serum levels among groups; nonparametrical correlations (Kendall's Tau) were applied when necessary. Statistical significance Serum MUC1 levels were measured in 53 HNSCC patients was set to p < 0.05 in all cases; calculations were per- by means of ELISA. MUC1 serum levels were higher in Table 2: Immunohistochemical results. Antibody Positive/Total % C595 (MUC1 Extracellular Domain) 9/50 18 Mixed 4/50 8 Linear 2/50 4 Cytoplasmic 3/50 6 CT33 (MUC1 Cytoplasmic Tail) 40/50 80 Mixed 27/50 54 Linear 34/50 68 Cytoplasmic 32/50 64 Nuclear 4/50 8 Results are expressed as positive cases over total number of cases and percentage, discriminated by immunohistochemical reaction pattern. Page 5 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 a Figure 1 – Immunohistochemical detection of MUC1 extracellular epitopes a – Immunohistochemical detection of MUC1 extracellular epitopes. A stage III larynx carcinoma specimen incu- bated with anti-MUC1 C595 monoclonal antibody is shown. Cell membrane and cytoplasm are stained over all the picture with a microvesicular pattern. Staining is strong inalmost all cells as well as in high proportion of intercellular connections. b – Immunohistochemical detection of MUC1 cytoplasmic tail epitopes. Section of a stage IV larynx carcinoma incubated with CT33 antibody is shown. Different cellular patterns are identified. Intense reaction is observed along the cytoplasm with a granular pattern and linear at nuclear membrane. Plasma membrane is partially stained. Cells are polymorphic as well as nuclei, showing different size and vesicular nuclear patterns. Some intercellular bridges are also stained with this antibody. Page 6 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 cancer patients (mean = 2.46 U/mL, SD = 3.46 U/mL) those patients over MUC1 cut off levels (4 U/mL). This than controls (mean = 0.60 U/mL, SD = 0.69 U/mL); the association was also found by the multivariate Principal percentage of individuals over the cut off value (4 U/mL) Component Analysis; in Fig. 2 the vectors representing was 8/53 (15%) vs 2/18 (10.5%), respectively. Through anti-MUC1 IgG and IgM are drawn opposed, while the multivariate Principal Component Analysis, in HNSCC, a vector that represents serum MUC1 was at the same quad- positive correlation was found between immunohisto- rant of anti-MUC1 IgG. chemistry MUC1 detection and Cancer Associated Serum Assay levels (Kendall τ = 0.275, p = 0.005; Fig. 2). Circulating immune complexes in HNSCC patients Fifty three HNSCC patients presented a mean circulating Anti-MUC1 humoral immune response analysis immune complexes value of 0.224 OD (SD = 0.103 OD) Anti-MUC1 Ab detection statistically different from normal negative controls Free anti-MUC1 serum Ab levels measured in HNSCC (mean = 0.171 OD, SD = 0.029; p = 0.038) and positive were lower than normal sex and age matched controls. controls (SLE patients), mean = 0.456 OD, SD = 0.121 Anti-MUC1 IgM mean value in cancer patients was 0.085 OD; p < 0.001. Tumor stage was statistically associated OD versus 0.194 OD in controls and anti-MUC1 IgG with circulating immune complexes levels (Fig. 4); two mean was 0.097 OD vs 0.171 OD. A negative correlation way ANOVA test with post hoc analysis of means showed between IgG and IgM levels (Kendall τ = -0.309, p = differences between T2 and T3 tumors versus T4 although 0.001) was found. Both anti-MUC1 IgG and anti-MUC1 node invasion did not present any difference. Also, a sta- IgM were statistically correlated with MUC1 serum levels tistical significant difference was found between well dif- (Fig. 3), while IgG levels were detected according to ferentiated and moderately differentiated tumor samples, MUC1 levels, anti-MUC1 IgM levels were very low in which presented a large amount of circulating immune Multivariate Principal Com Figure 2 ponent Analysis of clinical and histopathological data and tissue and serum expression of MUC1 Multivariate Principal Component Analysis of clinical and histopathological data and tissue and serum expres- sion of MUC1. Each variable is represented by vectors, which longitude correspond to the weight of the variable in the model, and their projection to both axes corresponds to the weight of the two main factors extracted (Factor 1 and Factor 2). These two factors explain 39% of the variations observedin the model. This percentage is remarkable taking into account that variables belong to patients. CIC = Circulating immune complexes levels; T = Tumor stage; N = Node status; DIF = tumor dif- ferentiation; CASA = Serum MUC1 levels; IgM Ab = free anti-MUC1 IgM antibodies; IgG Ab = free anti-MUC1 IgG antibodies; C595 = MUC1 extracellular domain expression; CT33 = MUC1 Cytoplasmic tail expression. Page 7 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Correlation Figure 3 between serum MUC1 levels measured by Cancer Associated Serum Assay (U/mL) and free anti-MUC1 antibodies Correlation between serum MUC1 levels measured by Cancer Associated Serum Assay (U/mL) and free anti- MUC1 antibodies. Serum MUC1 levels were plotted in abscises against anti-MUC1 IgG Abs (left coordinates) and IgM Abs (right coordinates). A negative correlation between serum MUC1 levels versus anti-MUC1 IgM Abs (τ = -0.266; p < 0.05) and a positive correlation between serum MUC1 levels and IgG Abs are observed (τ = 0.194; p < 0.05). complexes (p < 0.05) (Fig. 5). On the other hand, heavy chains. No statistical correlation was found although undifferentiated tumors had circulating between MUC1-CIC and tumor or sera MUC1 expression. immune complexes levels similar to well differentiated tumors, no statistical difference was found between poor Principal component analysis correlation studies versus moderately differentiated specimens. Elevated cir- Multivariate analysis of MUC1 expression, anti-MUC1 culating immune complexes levels were found in 16/53 immune response and clinical and pathological parame- patients (30.2%) which belonged to tumor stage I 1/3 ters in HNSCC showed that histological detection with (33%), stage III 3/15 (20%) and stage IV 12/30 (40%). CT33 and C595 Abs were associated with serum MUC1 levels. Moreover, serum MUC1 levels were positively cor- Identification of tumor antigens isolated from circulating related with anti-MUC1 IgG free antibodies and inversely immune complexes correlated with anti-MUC1 IgM. This suggested an associ- Eight out of 16 (50%) serum samples contained MUC1- ation between the presence of MUC1 in serum and the CIC; results are shown in Table 3. In each case, clinical type of antibody response. data have been included. In Fig. 6, results found in two patients are depicted, Western blot employing C595 mon- Vectors representing circulating immune complexes were oclonal antibody showed a double band at about 200 associated with advanced tumor T stage which may indi- kDa. Also, bands at the molecular weight ranging from 25 cate that a large tumor mass produces the MUC1 mucin to 60 kD were observed which correspond to Ig light and that reacts with antibodies, forming circulating immune complexes. Page 8 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Compar Figure 4 ative analysis of circulating immune complexes' levels (OD 280 nm) among different primary tumor T stages Comparative analysis of circulating immune complexes' levels (OD 280 nm) among different primary tumor T stages. Circulating immune complexes' levels in HNSCC increased with tumor T stage (F = 5.375; p < 0.05). Finally, as expected, we found an association between These observations clearly demonstrate the transforming advanced T stage, advanced node status and poor tumor properties of MUC1 overexpression in carcinoma and differentiation. Interestingly, the vectors that represent emphasize the importance of MUC1 cytoplasmic tail sub- these variables were opposed to serum MUC1 and anti- cellular localization. In HNSCC we have found extensive MUC1 IgG antibody levels. expression and cellular localization of MUC1 cytoplasmic tail; large and poorly differentiated tumors were likely to be less reactive with CT33 and C595 Abs. Discussion In our knowledge this is the first report that employs an anti MUC1 cytoplasmic tail Ab in HNSCC; it revealed a When C595 (anti-extracellular MUC1) monoclonal anti- high cellular expression of MUC1 (80% of malignant body was assayed, a lower percentage of reactivity with samples). respect to anti-MUC1 cytoplasmic tail was found; we have previously reported a high reactivity to MUC1 protein In previous studies developed in colon and breast cancer core epitopes in HNSCC [7,8]. The difference among [13,14], we have found similar results. The MUC1 cyto- results was expected since previous reports included a plasmic tail amino acid sequence contains seven tyrosine larger number of tumors localized in oral cavity (58%) in residues constituting intracellular signaling motifs [22] comparison with 49.1% in the present report; we found which play important roles in cell response to external that this localization is reactive with anti-protein core stimuli and is related to different oncogenes and genes monoclonal antibodies [7]. On the other hand, results controlling the cell cycle [23,24][25,26] and β-catenin presented here corresponded to 18.9% of pharyngeal [27]. Furthermore, it has been reported that in vivo, tumors which did not usually react with these mono- MUC1 cytoplasmic tail regulates erbB receptors signaling clonal antibodies; in past reports, only one patient with a through the activation of MAPK pathways [28]. pharyngeal tumor was included [7]. MUC1 cytoplasmic Page 9 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Wester Figure 5 n Blot analysis of circulating immune complexes' fractions eluted and isolated from affinity chromatography Western Blot analysis of circulating immune complexes' fractions eluted and isolated from affinity chromatog- raphy. Circulating immune complexes were obtained from a patient with a stage IV larynx tumor. Lane 1: 3.5% PEG precipi- tate; lane 2: PBS eluate from protein A Sepharose CL-4B column and lane 3: glycine HCl eluate. In lanes 1 and 3 a double band at about 200 kD is observed. tail reactivity is not affected by MUC1 glycosylation and Furthermore, we proved that MUC1-CIC detected by sialylation and, consequently, represents a better indica- Western blot were of the IgG isotype since isolation was tor of cell membrane associated MUC1 [15]. Other performed by protein A Sepharose CL-4B chromatogra- authors [29-31] have described MUC1 RNA splice variants phy, which is known to bind with high affinity to 1, 2 and which do not express the protein core region. Our results 4 IgG classes. also showed that serum MUC1 levels were elevated in 15% of HNSCC samples; interestingly, we found a statis- Presence of high amounts of circulating immune com- tically significant correlation between serum and tumor plexes in cancer patients is not very well understood. Anti- MUC1 detection indicating that this mucin may possibly body formation may represent the onset of an anti tumor be released by the tumor. immune response but, on the other hand, may be the con- sequence of tumor immune evasion related to a defective In this research we found anti-MUC1 antibodies; also, we cellular immune response. Nonetheless, circulating observed that serum MUC1 levels were positively corre- immune complexes have been associated with tumor pro- lated with free anti-MUC1 IgG, which may indicate that gression and also, their formation clearly affects the cor- an anti-MUC1 immune response is mounted. Tumor rect evaluation of several tumor markers. derived MUC1 is known to induce cellular and humoral specific immune responses in cancer patients [32-34]. In Only a few reports in the literature have described the this context, it is possible to speculate that a specific anti- antigenic component of circulating immune complexes in MUC1 immune response is induced in HNSCC patients. HNSCC. Vlock DR and others [36] have exhaustively stud- ied the reactivity of serum Abs against autologous tumor In another report [11], we pointed out that low MUC1 cell lines. These authors found that some patients pre- serum levels in stage I breast cancer patients were associ- sented high Ab titers which occasionally cross reacted ated with the presence of free and complexed anti-MUC1 with tumor cell lines derived from other histogenesis. antibodies. Similarly, circulating anti-MUC1-IgG anti- body levels were found predictive of survival in breast [35] In accordance with other reports [37,38], we found higher and pancreatic cancer patients [34]. We have found that circulating immune complexes levels in HNSCC patients anti-MUC1 IgG Ab correlated negatively with poor in comparison with normal subjects; furthermore, we also HNSCC differentiation and disease stage which are the agree with these authors that elevated circulating immune main predictors for patient outcome in this localization. complexes levels were detected in advanced tumor stage. They pointed out that high circulating immune complexes levels in HNSCC patients is the result of increased tumor Page 10 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 Table 3: Data from HNSCC patients' serum employed to isolate circulating immune complexes. Sample Localization Stage Sex Age Differentiation MUC1 In MUC1 in MUC1-CT in CIC levels MUC1 serum levels (U/mL) CIC IHQ IHQ (OD) 1 Larynx I M 52 Poorly + - - 0.306 3.9 2 Larynx III M 64 Well + - - 0.371 1.4 3 Larynx III M 56 Moderately + - + 0.320 1.2 4 Larynx III M 70 Moderately + - + 0.229 0.9 5 Larynx IVA M 58 Poorly - - + 0.411 2.2 6 Larynx IVA M 53 Moderately - - + 0.312 1.7 7 Oral Cavity IVA M 55 Moderately - + + 0.322 4.7 8 Oral Cavity IVA F 29 Moderately + - + 0.258 1.7 9 Oral Cavity IVA M 68 Moderately + - + 0.323 1.1 10 Oral Cavity IVA M 55 Moderately - - - 0.477 0.6 11 Oral Cavity IVA M 49 Well - - + 0.403 0.7 12 Pharynx IVA M 56 Moderately - - - 0.497 5.3 13 Pharynx IVA M 65 Moderately + - + 0.335 1.9 14 Pharynx IVA M 38 Poorly + + + 0.231 1.8 15 Pharynx IVA M 72 Moderately - - - 0.473 0.4 16 Pharynx IVB M 61 Moderately - - + 0.349 1.5 Positive/Total or 8/16 2/16 11/16 0.351 2.0 mean value Clinical data with immunohistochemical detection and serum determination of MUC1, anti-MUC1 Abs and circulating immune complexes are depicted. mass which would mediate changes in anti tumor immu- ated with advanced tumor stage; furthermore, the vectors nity. Moreover, Das TK et al [39] found that circulating that represent these variables did not positively correlate immune complexes persist after surgical excision of the to serum MUC1 and anti-MUC1 IgG antibody levels, and primary tumor due to the presence of remaining tumor also were inversely correlated with MUC1 tumor expres- tissue or occult metastasis. sion. Considering that we identified MUC1 as an anti- genic component of circulating immune complexes, taken MUC1 is a mucin that has been detected both in the cyto- together, these results indicate that a large tumor mass plasm, plasma membrane and nucleus in different tumor produces high MUC1 mucin that is liberated to circula- localizations. MUC1 is synthesized in the ER and glyco- tion and captured by IgG antibodies forming MUC1-CIC. sylated in the Golgi apparatus but it has been described that it suffers several glycosylation cycles being expressed Another interesting conclusion is that poorly differenti- in the plasma membrane several times [40]. Furthermore, ated tumors are inversely correlated with tumor and different splicing variants have been described that can be serum MUC1 detection. This finding suggests the possibil- present in the cytoplasm and in the plasma membrane, as ity to take into account its consideration regarding prog- well [31,41-43]. Wen et al [44] investigated intracellular nosis and follow up of this disease. trafficking of MUC1 cytoplasmic tail in human pancreatic cancer cell lines S2-013 and Panc-1 and detected this frac- Competing interests tion at the inner cell surface, in the cytosol and in the The author(s) declare that they have no competing inter- nucleus. They hypothesized that the association between ests. β-catenin and fragments of the MUC1 cytoplasmic tail facilitated the cytosol-to-nuclear translocation of β-cat- Authors' contributions enin and contributed to its nuclear accumulation. We MER carried out the studies and drafted the manuscript. have found [14] MUC1 cytoplasmic tail and protein core MVC conceived the study, participated in the design and expression in the plasma membrane, cytoplasm and coordination and wrote the manuscript. AP obtained the nucleus in breast and colorectal tumor samples. samples and clinical records and also participated in the design of the manuscript. ASE participated in the design and coordination of the study and helped to draft the Conclusion In HNSCC, Principal Component Analysis (Fig. 2) manuscript. All authors read and approved the final man- showed that circulating immune complexes were associ- uscript. Page 11 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 17. Renkonen J, Paavonen T, Renkonen R: Endothelial and epithelial Acknowledgements expression of sialyl Lewis(x) and sialyl Lewis(a) in lesions of The authors whish to thank Dr. A Murray and Dr. S Stevenson for provid- breast carcinoma. Int J Cancer 1997, 74:296-300. ing C595 monoclonal antibody (care of Prof. Alan Perkins, Nottingham, 18. Luna-More S, Rius F, Weil B, Jimenez A, Bautista MD, Perez-Mellado UK), Dr. K. Chul Kim (Lovelace Respiratory Research Institute, Albu- A: EMA: a differentiation antigen related to node metastatic capacity of breast carcinomas. Pathol Res Pract 2001, 197:419-25. querque, NM, USA) for generously providing CT33 polyclonal antibody. 19. Croce MV, Price MR, Segal-Eiras A: Expression of monoclonal- The authors are also extremely grateful to Lic. S Demichelis for performing antibody-defined antigens in fractions isolated from human statistical analysis, Dr. J Carri for encourage in photographical techniques breast carcinomas and patients' serum. Cancer Immunol Immu- and JC Molina for skillful technical assistance. nother 1995, 40:132-7. 20. Laemmli UK: Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, Prof. Amada Segal-Eiras and Prof. María Virginia Croce are members of the 227:680-5. research career of the Consejo Nacional de Investigaciones Científicas y 21. Towbin H, Staehelin T, Gordon J: Electrophoretic transfer of Técnicas (CONICET) and the Comisión de Investigaciones Científicas de la proteins from polyacrylamide gels to nitrocellulose sheets: Provincia de Buenos Aires (CIC/PBA), respectively. Dr. Martin E. Rabassa procedure and some applications. Proc Natl Acad Sci USA 1979, 76:4350-4. is a fellow of CONICET. Financial support from FONCYT, CONICET, 22. Von Mensdorff-Pouilly S, Gouretvich MM, Kenemans P: An enzyme- CICPBA and University of La Plata (Argentina) is very much appreciated. linked immunosorbent assay for the measurement of circu- lating antibodies to polymorphic epithelial mucin (MUC1). References Tumor Biol 1989:186-195. 23. Croce MV, Isla-Larrain MT, Capafons A, Price MR, Segal-Eiras A: 1. Döbrössy L: Epidemiology of head and neck cancer: magni- Humoral immune response induced by the protein core of tude of the problem. Cancer and Metastasis Reviews 2005, 24:9-17. MUC1 mucin in pregnant and healthy women. Breast Cancer 2. Shermann CD Jr: Cancer of the head and neck. In Manual of Clin- Res Treat 2001, 69:1-11. ical Oncology UICC, Springer; 1990. 24. Wang H, Lillehoj EP, Kim KC: Identification of four sites of stim- 3. International Agency for research in Cancer, GLOBOCAN ulated tyrosine phosphorylation in the MUC1 cytoplasmic 2002 [http://www-dep.iarc.fr] tail. Biochem Biophys Res Commun 2003, 310:341-6. 4. Stewart B, Kleihues P, editors: World cancer report. Lyon: IARC 25. Zrihan-Licht S, Baruch A, Elroy-Stein O, Keydar I, Wreschner DH: Press; 2003. Tyrosine phosphorylation of the MUC1 breast cancer mem- 5. Dabelsteen E, Gao S: ABO blood-group antigens in oral cancer. brane proteins. Cytokine receptor-like molecules. FEBS Lett J Dent Res 2005, 84:21-8. 1994, 356:130-6. 6. Croce MV, Price MR, Segal-Eiras A: Detection and isolation of 26. Pandey P, Kharbanda S, Kufe D: Association of the DF3/MUC1 MUC1 mucin from larynx squamous cell carcinoma. Pathol breast cancer antigen with Grb2 and the Sos/Ras exchange Oncol Res 2000, 6:93-9. protein. Cancer Res 1995, 55:4000-3. 7. Croce MV, Rabassa ME, Price MR, Segal-Eiras A: MUC1 mucin and 27. Yamamoto M, Bharti A, Li Y, Kufe D: Interaction of the DF3/ carbohydrate associated antigens as tumor markers in head MUC1 breast carcinoma-associated antigen and beta-cat- and neck squamous cell carcinoma. Pathol Oncol Res 2001, enin in cell adhesion. J Biol Chem 1997, 272:12492-4. 7:284-91. 28. Schroeder JA, Thompson MC, Gardner MM, Gendler SJ: Transgenic 8. Croce MV, Rabassa ME, Pereyra A, Segal-Eiras A: Influence of sialic MUC1 interacts with epidermal growth factor receptor and acid removal on MUC1 antigenic reactivity in head and neck correlates with mitogen-activated protein kinase activation carcinoma. Pathol Oncol Res 2005, 11:74-81. in the mouse mammary gland. J Biol Chem 2001, 276:13057-64. 9. Taylor-Papadimitriou J, Burchell JM, Plunkett T, Graham R, Correa I, 29. Wreschner DH, Hareuveni M, Tsarfaty I, Smorodinsky N, Horev J, Miles D, Smith M: MUC1 and the immunobiology of cancer. J Zaretsky I, Kotkes P, Weiss M, Lathe R, Dion AS, Keydar I: Human Mammary Gland Biol Neoplasia 2002, 7:209-21. epithelial tumor antigen cDNA sequences – differential splic- 10. Kotera Y, Fontenot JD, Pecher G, Metzgar RS, Finn OJ: Humoral ing may generate multiple protein forms. Eur J Biochem 1990, immunity against a tandem repeat epitope of human mucin 189:463-73. MUC-1 in sera from breast, pancreatic, and colon cancer 30. Zrihan-Licht S, Vos HL, Baruch A, Elroy-Stein O, Sagiv D, Keydar I, patients. Cancer Res 1994, 54:2856-60. Hilkens J, Wreschner DH: Characterization and molecular 11. von Mensdorff-Pouilly S, Gourevitch MM, Kenemans P, Verstraeten cloning of a novel MUC1 protein, devoid of tandem repeats, AA, Litvinov SV, van Kamp GJ, Meijer S, Vermorken J, Hilgers J: expressed in human breast cancer tissue. Eur J Biochem 1994, Humoral immune response to polymorphic epithelial mucin 224:787-95. (MUC-1) in patients with benign and malignant breast 31. Obermair A, Schmid BC, Stimpfl M, Fasching B, Preyer O, Leodolter tumours. Eur J Cancer 1996, 32A:1325-31. S, Crandon AJ, Zeillinger R: Novel MUC1 splice variants are 12. Croce MV, Isla-Larrain MT, Demichelis SO, Gori JR, Price MR, Segal- expressed in cervical carcinoma. Gynecol Oncol 2001, 83:343-7. Eiras A: Tissue and serum MUC1 mucin detection in breast 32. Agrawal B, Reddish MA, Longenecker BM: In vitro induction of cancer patients. Breast Cancer Res Treat 2003, 81:195-207. MUC-1 peptide-specific type 1 T lymphocyte and cytotoxic 13. Price MR, Pugh JA, Hudecz F, Griffiths W, Jacobs E, Symonds IM, T lymphocyte responses from healthy multiparous donors. J Clarke AJ, Chan WC, Baldwin RW: C595 – a monoclonal anti- Immunol 1996, 157:2089-95. body against the protein core of human urinary epithelial 33. Petrarca C, Rughetti A, Rahimi H, D'Agostini F, Turchi V, Apollonj mucin commonly expressed in breast carcinomas. Br J Cancer Ghetti C, Scambia G, Frati L, Nuti M: Human antibodies against 1990, 61:681-6. the polymorphic epithelial mucin in ovarian cancer patients 14. Croce MV, Isla-Larrain MT, Remes-Lenicov F, Colussi AG, Lacunza E, recognise a novel sequence in the tandem repeat region. Eur Kim KC, Gendler SJ, Segal-Eiras A: MUC1 cytoplasmic tail detec- J Cancer 1996, 32A:2155-63. tion using CT33 polyclonal and CT2 monoclonal antibodies 34. Hamanaka Y, Suehiro Y, Fukui M, Shikichi K, Imai K, Hinoda Y: Cir- in breast and colorectal tissue. Histol Histopathol 2006, culating anti-MUC1 IgG antibodies as a favorable prognostic 21:849-55. factor for pancreatic cancer. Int J Cancer 2003, 103:97-100. 15. Croce MV, Isla-Larrain MT, Rua CE, Rabassa ME, Gendler SJ, Segal- 35. von Mensdorff-Pouilly S, Verstraeten AA, Kenemans P, Snijdewint FG, Eiras A: Patterns of MUC1 tissue expression defined by an Kok A, Van Kamp GJ, Paul MA, Van Diest PJ, Meijer S, Hilgers J: Sur- anti-MUC1 cytoplasmic tail monoclonal antibody in breast vival in early breast cancer patients is favorably influenced by cancer. J Histochem Cytochem 2003, 51:781-8. a natural humoral immune response to polymorphic epithe- 16. Feickert H, Anger B, Cordón-Cardo C, Lloyd K: Cell-surface anti- lial mucin. J Clin Oncol 2000, 18:574-83. gens of human tumors detected by mouse monoclonal anti- 36. Vlock DR, Scalise D, Schwartz DR, Richter DE, Krause CJ, Baker SR, bodies: definition of blood-group- and non-blood-group- Carey TE: Incidence of serum antibody reactivity to autolo- related antigenic systems. Int J Cancer 1990, 46:1007-1013. gous head and neck cancer cell lines and augmentation of Page 12 of 13 (page number not for citation purposes) BMC Cancer 2006, 6:253 http://www.biomedcentral.com/1471-2407/6/253 antibody reactivity following acid dissociation and ultrafiltra- tion. Cancer Res 1989, 49:1361-5. 37. Yamanaka N, Himi T, Harabuchi Y, Hoki K, Kataura A: Soluble immune complexes and squamous cell carcinoma-related antigens in patients with head and neck cancer. Cancer 1988, 62:1932-8. 38. Wolf GT, Wolfe RA: Circulating immune complexes in patients with nasopharyngeal carcinoma. Laryngoscope 1990, 100:302-8. 39. Das TK, Aziz M, Rattan A, Sherwani R: Prognostic significance of circulating immune complexes in malignant tumours of head and neck. J Indian Med Assoc 1995, 93:3-7. 40. Litvinov SV, Hilkens J: The epithelial sialomucin, episialin, is sia- lylated during recycling. J Biol Chem 1993, 268:21364-71. 41. Ligtenberg MJ, Vos HL, Gennissen AM, Hilkens J: Episialin, a carci- noma-associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini. J Biol Chem 1990, 265:5573-8. 42. Williams CJ, Wreschner DH, Tanaka A, Tsarfaty I, Keydar I, Dion AS: Multiple protein forms of the human breast tumor-associ- ated epithelial membrane antigen (EMA) are generated by alternative splicing and induced by hormonal stimulation. Biochem Biophys Res Commun 1990, 170:1331-8. 43. Levitin F, Stern O, Weiss M, Gil-Henn C, Ziv R, Prokocimer Z, Smorodinsky NI, Rubinstein DB, Wreschner DH: The MUC1 SEA module is a self-cleaving domain. J Biol Chem 2005, 280:33374-86. 44. Wen Y, Caffrey TC, Wheelock MJ, Johnson KR, Hollingsworth MA: Nuclear association of the cytoplasmic tail of MUC1 and beta-catenin. J Biol Chem 2003, 278:38029-39. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2407/6/253/pre pub Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 13 of 13 (page number not for citation purposes)

Journal

BMC Cancer – Springer Journals

Published: Dec 1, 2006

Keywords: cancer research; oncology; surgical oncology; health promotion and disease prevention; biomedicine, general; medicine/public health, general

There are no references for this article.