Calreticulin is a B cell molecular target in some gastrointestinal malignancies

Abstract

Calreticulin, upon translocation to the cell surface, plays a critical role in the recognition of tumour cells and in experimentally induced cellular anti-tumour immunity. However, less is known about anti-calreticulin antibodies and their role in malignancies. Using enzyme-linked immunosorbent assay (ELISA), we found immunoglobulin (Ig)A and/or IgG anti-calreticulin antibodies in sera of approximately 63% of patients with hepatocellular carcinoma (HCC), 57% of patients with colorectal adenocarcinoma (CRA) and 47% of patients with pancreatic adenocarcinoma (PACA), while healthy controls, patients with viral hepatitis C and with chronic pancreatitis reached only 2%, 20% and 31% seropositivity, respectively. We found significantly elevated mean levels of IgA anti-calreticulin antibodies (P < 0·001) in patients with HCC (78·7 ± 52·3 AU, mean ± standard deviation), PACA (66·5 ± 30·9 AU) and CRA (61·8 ± 25·8 AU) when compared to healthy controls (41·4 ± 19·2 AU). Significantly elevated mean levels of IgG anti-calreticulin antibodies (P < 0·001) were detected in patients with HCC (121·9 ± 94·2 AU), gall bladder adenocarcinoma (118·4 ± 80·0 AU) and PACA (88·7 ± 55·6 AU) when compared to healthy controls (56·7 ± 22·9 AU). Pepscan analysis revealed a large number of antigenic epitopes of calreticulin recognized by both IgA and IgG antibodies of patients with HCC and PACA, indicating robust systemic immune response. Moreover, significantly elevated levels of antibodies against peptide KGEWKPRQIDNP (P < 0·001) in these patients, tested by ELISA, confirmed the distinct character of antibody reactivity against calreticulin. The high occurrence and specificity of serum anti-calreticulin autoantibodies in the majority of patients with some gastrointestinal malignancies provide the evidence for their possible clinical relevance.

Introduction

Extensive studies in past years have revealed a multivalent relationship between calreticulin, anti-tumour immunity and autoimmunity [1–6]. Calreticulin is a ubiquitous, phylogenetically conserved molecule that is localized primarily in the endoplasmic reticulum. Functionally, it has roles with controlling cellular calcium homeostasis, folding of nascent proteins, cellular adhesion and clearance of apoptotic cells and cell debris. The molecular mass of full-length calreticulin is approximately 42 kDa and due to its acidic isoelectric point (pI) of 4·65, it migrates as a 60–62 kDa molecule when analysed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The calreticulin molecule is divided arbitrarily into three structurally and functionally different domains, N, P and C [7]. Calreticulin is considered to be a key molecule in macrophage and dendritic cell recognition, the engulfment of starving cells and tamoxifen- or anthracycline-treated and gamma-irradiated dying tumour cells. Furthermore, anthracyclines and irradiation induce calreticulin translocation to the surface of tumour cells and elicit an anti-cancer immune response in an experimental mouse model. Such translocation is necessary for the induction of an autoimmune reaction [1–5] and it is well established that autoimmunity is often associated with cancer [8–13].

The cellular proteins that induce an antibody response in gastrointestinal malignancies vary and include both proteins encoded by mutated genes and cellular proteins overexpressed or aberrantly expressed in malignant tissue, e.g. carcinoembryonic antigen (CEA), p53, survivin, livin, mucins, Fas, dsDNA, ssDNA, HER-2/neu, cancer/testes antigens, etc. [14–20]. Interestingly, constitutive overexpression of calreticulin seems to be a common event in cells from various tumour types including colorectal adenocarcinoma, breast carcinoma and hepatocellular carcinoma [21–23]. Hence, our aim was to characterize calreticulin as a molecular target of serum immunoglobulin (Ig)A and IgG antibodies in patients suffering from gastrointestinal malignancies by using human recombinant calreticulin and its synthetic peptides.

Materials and methods

Patients and serum samples

Sera of newly diagnosed patients with hepatocellular carcinoma (HCC), pancreatic adenocarcinoma (PACA), colorectal adenocarcinoma (CRC) and gall bladder adenocarcinoma (GBA), along with risk group patients with chronic pancreatitis (CP) and viral hepatitis C (VHC), were obtained from the departments of internal medicine at Charles University hospitals (Thomayer's Faculty Hospital and Hospital of the 3rd Faculty of Medicine) and the Institute of Clinical and Experimental Medicine. The baseline characteristics of patients and healthy controls tested for serum anti-calreticulin antibodies are presented in Table 1. Diagnosis of the different malignancies was made according to previously established criteria using clinical, radiological and endoscopic approaches and confirmed by histology. Patients were not treated for oncological diseases at the time of blood draw. All patients were subjected to liver function tests. Tumour markers, including alpha-fetoprotein (AFP), carbohydrate antigen (CA 19-9) and CEA, were tested for in 25–50% of patients in each cohort in an accredited hospital laboratory. Individuals with hyper- or hypogammaglobulinaemia were excluded from the study. The sera of healthy donors were used as controls. This study was approved by the local Ethics Committee.

Table 1.

The baseline characteristics of patients and healthy controls tested for serum anti-calreticulin antibodies.

Diagnosis	No	F/M ratio	Mean age, range
Hepatocellular carcinoma	43 (11†, 29‡)	13/30	70·5, 57–82
Viral hepatitis C	20 (18†)	12/8	44·9, 24–74
Pancreatic adenocarcinoma	55	23/32	71·6, 59–79
Chronic pancreatitis	16	9/7	63·7, 45–81
Colorectal adenocarcinoma	30	12/18	62·3, 51–81
Gall bladder adenocarcinoma	29	17/12	63·8, 47–82
Healthy controls	56	30/26	31·2, 23–63

^†Child–Pugh score 1;

^‡Child–Pugh score 2. F, female; M, male.

Determination of serum levels of antibodies against calreticulin, calreticulin peptides and tissue transglutaminase

An enzyme-linked immunosorbent assay (ELISA) for testing the serum levels of IgA and IgG anti-calreticulin antibodies was performed as described in our previous studies [24,25]. Briefly, human recombinant calreticulin used in a final concentration of 5 µg/ml of phosphate-buffered saline (PBS) was coated on 96-well polystyrene plates (Gama, České Budějovice, Czech Republic) overnight at 4°C. Blocking buffer [1% bovine serum albumin (BSA) in PBS; Sigma, Steinheim, Germany] was used as a negative control. Each serum sample and standard serum diluted in blocking buffer (1:20, 1:100, for IgA and 1:100, 1:500 for IgG) were analysed in triplicate. The highest optical density (OD) of negative controls was subtracted from the mean of OD measured.

An ELISA for testing the serum antibodies against calreticulin peptides was performed as described earlier [24].

The presence of serum IgA anti-tissue transglutaminase (tTG) antibodies was tested with a commercially available ELISA test, using human recombinant tTG activated by gliadin (BioSystems™, Barcelona, Spain), according to the manufacturer's instructions.

Immunoblot analysis

One µg of recombinant calreticulin was subjected to SDS-PAGE (12·5% gel) under reducing conditions [26]. The binding of serum antibodies to calreticulin was performed by immunoblot analysis as described previously [24]. The detection of secretory IgA antibodies specific for calreticulin was accomplished using sheep anti-human secretory IgA (The Binding Site, Birmingham, UK) diluted 1:1000 in a blocking solution [4% low-fat milk in PBS containing 0·1% Tween-20 (PBS-T)]. The diluted anti-human secretory IgA antibodies were incubated for 2 h with membrane strips. After washing with PBS-T, peroxidase-conjugated donkey antibodies against sheep/goat immunoglobulins (The Binding Site) diluted in the same solution and at the same concentration as anti-human secretory IgA antibodies were added to the strips for 1 h.

Pepscan analysis

Pepscan experiments were performed as described by Sánchez et al.[24,25]. Two hundred and one calreticulin decapeptides, each overlapping by eight amino acid residues and covering the entire sequence from the N- to the C-terminus, were synthesized and bound covalently to cellulose membranes (Abimed, Langenfeld, Germany). The binding of serum IgA and IgG antibodies of patients from the tested groups (HCC, n = 10; PACA, n = 10) and of healthy donors (n = 12) to calreticulin peptides was evaluated qualitatively as positive or negative by comparing various exposure times. The data are expressed as a reactivity index (RI), which is the ratio of the number of patients whose antibodies recognized an individual peptide to the number of patients in the tested group.

Statistical methods

The non-parametric Mann–Whitney U-test was used for the comparison of serum antibody levels against calreticulin and its peptides across all patient cohorts and healthy controls as tested by ELISA. Serum levels of antibodies against full-length human recombinant calreticulin were expressed as AU, which refers to the OD of an internal standard serum (taken as 100%). The cut-off value for anti-calreticulin antibodies was calculated as the mean value plus two standard deviations (s.d.) from the data pool of 56 control sera samples. Values above the cut-off were considered as positive. The levels of antibodies against calreticulin peptides were expressed in OD.

Results

Analysis of serum IgA anti-calreticulin antibodies in oncological patients and patients from risk groups

The distribution of serum IgA anti-calreticulin antibodies in patients with gastrointestinal malignancies, patients from risk groups and healthy controls is shown in Fig. 1a. The mean levels of serum IgA anti-calreticulin antibodies reached the highest values in the HCC patient cohort (78·7 ± 52·3 AU, mean ± s.d.) and the PACA cohort (66·5 ± 30·9 AU). Significantly elevated (P < 0·001) levels of anti-calreticulin antibodies were also found in patients with CRA (61·8 ± 25·8 AU) when compared to healthy controls (41·4 ± 19·2 AU). Conversely, we did not find significantly increased serum levels of these antibodies in patients with GBA (53·4 ± 30·2 AU). Similarly, neither patients with VHC (41·5 ± 30·3 AU) nor with CP (53·13 ± 21·6 AU) had significantly elevated IgA anti-calreticulin antibody levels. Interestingly, we found a statistically significant (P < 0·01) difference in IgA anti-calreticulin antibody levels between cohorts of patients with HCC and VHC. However, no significant difference between the PACA and CP patient groups was found. The number of seropositive patients exceeding the cut-off value in groups tested was 17 of 43 HCC patients (∼40%), 14 of 55 PACA patients (∼25%) and nine of 30 CRA patients (30%). Two of 16 CP patients (∼13%), three of 29 GBA patients (∼10%) and three of 20 VHC patients (∼15%) were positive for IgA anti-calreticulin antibodies. In this study none of the 56 healthy controls were positive for these antibodies.

Fig. 1.

Shown is a scatter graph of individual values of serum immunoglobulin (Ig)A (a) and IgG (b) anti-calreticulin antibodies tested by enzyme-linked immunosorbent assay (ELISA) in patients with hepatocellular carcinoma (HCC, n = 43), viral hepatitis C (VHC, n = 20), pancreatic adenocarcinoma (PACA, n = 55), chronic pancreatitis (CP, n = 16), colorectal adenocarcinoma (CRA, n = 30), gall bladder adenocarcinoma (GBA, n = 29) and healthy controls (C, n = 56). The anti-calreticulin antibody serum levels are expressed in arbitrary units (AU). Solid lines indicate the mean. Cut-off values are 80 AU for IgA and 102 AU for IgG antibodies. ***P < 0·001; **P < 0·01; n.s., not significant.

Analysis of serum IgG anti-calreticulin antibodies in oncological patients and patients from risk groups

Individual values of IgG anti-calreticulin antibodies in patients with HCC, VHC, PACA, CP, CRA and GBA are shown in Fig. 1b. Significantly elevated (P < 0·001) mean levels of these antibodies were found in the sera of patients with HCC (121·9 ± 94·2 AU) and GBA (118·4 ± 80·0 AU; negative for the IgA isotype of anti-calreticulin antibodies) and in patients with PACA (88·7 ± 55·6 AU), compared to healthy controls (56·7 ± 22·9 AU). Conversely, CRA patients (who have significantly elevated levels of IgA anti-calreticulin antibodies) did not have increased titres of IgG anti-calreticulin antibodies (77·7 ± 50·7 AU). We also did not observe significant differences in the levels of IgG anti-calreticulin antibodies between the group of healthy controls and groups of patients with CP (77·2 ± 50·7) and VHC (65·4 ± 50·4). Similar to the case of IgA anti-calreticulin antibodies, we detected significantly increased (P < 0·01) levels of IgG anti-calreticulin antibodies in the group of HCC patients compared to the VHC patient risk group, but no significant difference was found between the mean antibody levels of patients with PACA and CP. The highest ratio of seropositive patients was found within the patient groups that have significantly increased mean levels of IgG anti-calreticulin antibodies: 20 of 43 patients with HCC (∼47%), 13 of 29 patients with GBA (∼45%) and 17 of 55 patients with PACA (31%). The groups of patients with CRA, CP and VHC contained nine of 30 (30%), four of 16 (25%) and four of 20 (20%) seropositive individuals, respectively, while only one of the 56 (∼2%) healthy individuals was positive for IgG anti-calreticulin antibodies.

Interestingly, 27 of 43 patients with HCC (∼63%), 17 of 30 patients with CRA (∼57%) and 26 of 55 patients with PACA (∼47%) were seropositive for IgA and/or for IgG anti-calreticulin antibodies, while healthy controls, patients with VHC and CP reached only ∼2%, 20% and ∼31% seropositivity, respectively.

Immunoblot analysis of the reactivity of patient sera with calreticulin

The binding of antibodies from patients with HCC and PACA to human recombinant calreticulin was confirmed by immunoblot. Figure 2 shows the reactivity of serum IgA and IgG antibodies of oncological patients with calreticulin, while the binding of antibodies from healthy controls was not observed. Moreover, we also demonstrated the presence of a secretory form of IgA antibodies specific to calreticulin in the serum of a HCC patient.

Fig. 2.

Using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting, serum immunoglobulin (Ig)A and IgG antibody reactivity with human recombinant calreticulin (CRT, ∼60 kDa) in patients with hepatocellular carcinoma (HCC) and pancreatic adenocarcinoma (PACA) was analysed. Lane 1 represents the marker of molecular weights, and lane 2 is SDS-PAGE-subjected calreticulin stained by Coomassie brilliant blue. Nitrocellulose-blotted recombinant calreticulin was incubated with sera of HCC (lanes 3–5) and PACA (lanes 6–8) patients, and for control purposes, with specific rabbit anti-calreticulin antibodies (lane 12) and sera of healthy individuals (lanes 14, 15). The binding of anti-calreticulin antibodies was detected by peroxidase conjugated anti-human IgA (lanes 3, 4, 6, 7, 14), anti-human IgG (lanes 5, 8, 15) or anti-rabbit secondary antibodies (lane 12). Lane 9 shows the reactivity of secretory IgA antibodies of patient with HCC. As a control, secondary antibodies were used (lanes 10, 11, 13).

Fine specificity analysis of anti-calreticulin antibodies

Antigenic epitopes recognized by IgA antibodies in patients with hepatocellular carcinoma and pancreatic adenocarcinoma

Fine specificity of IgA and IgG anti-calreticulin antibodies was analysed by Pepscan in groups of patients with HCC and PACA, in whom we found significantly increased levels of these antibodies of both isotypes. Figure 3 represents examples of the reactivity of anti-calreticulin antibodies with individual calreticulin decapeptides. Figure 4 maps cumulatively the IgA and IgG antibody reactivity of patient and control sera with calreticulin peptides. Both IgA and IgG anti-calreticulin antibodies of HCC and PACA patients recognized a number of epitopes distributed over the entire calreticulin molecule. IgA antibodies of HCC and PACA patients recognized predominantly amino acid sequences VQFTVKHEQNID (nine of 10 HCC patients demonstrated recognition, seven of 10 PACA), FAEAPFSNKGQT (nine of 10 HCC), LSASFAEAPF (10 of 10 PACA) in the N-domain, EVKIDNSQVE (10 of 10 HCC, 10 of 10 PACA) and its overlapping peptide KIDNSQVESG (10 of 10 HCC, eight of 10 PACA) at the boundary of the N- and P-domains and KGEWKPRQID (eight of 10 HCC) and its overlapping peptide EWKPRQIDNP (10 of 10 HCC) in the P-domain. The C-domain of calreticulin contained the largest number of immunodominant epitopes recognized by the patients' IgA antibodies tested. The C-domain contained epitopes IFDNFLITND, KDKQDEEQRLKE, EQRLKEEEED and EEEDKKRKEEEEAEDKED, which were recognized by IgA antibodies of all patients with HCC (10 of 10), and also RKEEEEAEDKEDDA and EDDAEDKDEDEE, which were recognized by IgA antibodies of all patients with PACA (10 of 10). IgA antibodies of healthy controls also recognized, although with low frequency (four of 12), calreticulin peptides DEEKDKGLQTSQ and KGKNVLINKD (in the N-domain) and EDKKRKEEEE (in the C-domain).

Fig. 3.

This figure shows the reactivity of serum immunoglobulin (Ig)A (a,b) and IgG (c,d) antibodies from patients with hepatocellular carcinoma (a,c), pancreatic adenocarcinoma (b,d), and healthy controls [e (IgA), f (IgG)] with calreticulin decapeptides using Pepscan (no. = number of calreticulin peptides).

Fig. 4.

Recognition profiles (—) of serum immunoglobulin (Ig)A and IgG antibodies against calreticulin decapeptides (each overlapping by eight amino acids) in patients with hepatocellular carcinoma (a, n = 10), pancreatic adenocarcinoma (b, n = 10) and healthy controls (----, n = 12). Peaks of antibody reactivity [expressed as index of reactivity (RI)] which exceeded the horizontally dotted line (*) represent calreticulin peptides that are recognized frequently by antibodies of patients in the tested group. The synthetic peptides 1–88, 89–143 and 144–201 correspond to the N-, P- and C-domains (distinguished by vertically dotted lines) of the calreticulin molecule, respectively.

Antigenic epitopes recognized by IgG antibodies in patients with hepatocellular carcinoma and pancreatic adenocarcinoma

The recognition profile of patients' IgG antibodies was similar to that of IgA antibodies. IgG antibodies from patients with HCC and PACA mainly recognized amino acid sequences LSASFAEAPFSNKGQT (six of 10 PACA) and FAEAPFSNKGQT (10 of 10 HCC) in the N-domain, EVKIDNSQVE (eight of 10 HCC, eight of 10 PACA) and its overlapping peptide KIDNSQVESG (eight of 10 HCC) at the boundary of the N- and P-domains and ERAKIDDPTD (10 of 10 HCC, six of 10 PACA) and DKAPEHIPDPDA (eight of 10 PACA) in the P-domain of calreticulin. The most immunodominant epitopes of calreticulin were localized to its C-domain, as was the case for the IgA antibodies. IgG antibodies of patients with HCC recognized epitopes EQRLKEEEED (10 of 10) and EDKKRKEEEEAEDKED (10 of 10). IgG antibodies of PACA patients bound with high frequency peptides IFDNFLITND (eight of 10), EDKKRKEEEEAEDK (eight of 10), EEEEAEDKED (10 of 10) and EDKDEDEEDEEDKEED (eight of 10). Epitopes GTIFDNFLITNDEA and EDKKRKEEEEAE were recognized with low frequency (four of 12) by IgG antibodies of healthy controls.

Levels of serum IgA antibodies against calreticulin peptides in patients with hepatocellular carcinoma and pancreatic adenocarcinoma

Using ELISA, we quantified the serum levels of IgA antibodies against selected calreticulin peptides KGEWKPRQIDNP and VQFTVKHEQNID, which were found to be immunodominant for patients with HCC and/or patients with PACA. Peptide GVTKAAEKQMKD, characterized previously as immunodominant for patients with refractory coeliac disease (RCD) [24], was employed as a control peptide. The levels of IgA antibodies against calreticulin peptides in sera of 16 patients with HCC, PACA and healthy controls are shown in Fig. 5. Although the mean levels of antibodies against each tested calreticulin peptide were higher in the patient cohorts compared to the control group, only the mean levels of antibodies against peptide KGEWKPRQIDNP were elevated significantly (P < 0·001), both in patients with HCC (318·4 ± 120·6 OD, mean ± s.d.) and PACA (330·3 ± 107·5 OD) when compared with healthy controls (197·4 ± 38·8 OD).

Fig. 5.

The serum levels of immunoglobulin (Ig)A antibodies against peptides KGEWKPRQIDNP (a), VQFTVKHEQNID (b) and GVTKAAEKQMKD (c) were tested by enzyme-linked immunosorbent assay (ELISA) in patients with hepatocellular carcinoma (HCC, n = 16), pancreatic adenocarcinoma (PACA, n = 16) and healthy controls (C, n = 16). Data are expressed as optical density (OD). Solid lines represent the mean values. ***P < 0·001; n.s., not significant.

IgA antibodies against tissue transglutaminase in patients with oncological diseases

In addition to testing serum anti-calreticulin antibodies in oncological patients, serum levels of IgA antibodies against tTG, another autologous antigen related to cancer [27], were evaluated for comparison. Only three of 43 patients with HCC and one of the 55 patients with PACA were seropositive for IgA anti-tTG antibodies. Control sera were negative for these antibodies.

Discussion

In this study, we estimated serum levels of antibodies against calreticulin in sera from patients with newly diagnosed gastrointestinal malignancies and from patients from risk groups. The highest percentage of seropositive patients (40% and more) and the highest levels of anti-calreticulin antibodies were detected in cohorts of patients with HCC (IgA, IgG) and GBA (IgG). The ratio of anti-calreticulin antibody-positive patients was lower when compared to patients with active coeliac disease (CD, IgA, IgG isotype), alcoholic liver cirrhosis (IgA) and autoimmune hepatitis type I (AIH, IgA), but higher when compared to patients with other autoimmune and gastrointestinal diseases [25,28]. Until now, it was described that serum antibodies of some patients with HCC and pancreatic cancer recognized calreticulin in lysates of tumour cell lines or liver and pancreatic tumours analysed by combination of two-dimensional electrophoresis and immunoblot [29,30].

The existence of a number of antigenic epitopes recognized by both IgA and IgG antibodies from patients with HCC and PACA suggests a robust systemic immune response, which includes determinant spreading, affinity maturation and isotype switching. Moreover, the specificity of antibodies to calreticulin peptides of the N- and P-domains in oncological patients is different from that demonstrated in patients with AIH and primary biliary cirrhosis, as described in our previous study [25]. The distinct antibody response in oncological patients is also demonstrated by ELISA using selected calreticulin peptides. Patients with HCC and PACA had significantly elevated mean levels of serum IgA antibodies against peptide KGEWKPRQIDNP, while significantly elevated mean levels of IgA antibodies against peptide GVTKAAEKQMKD (characteristic for RCD patients) were not detected [24].

In our study, anti-calreticulin antibody positivity did not correspond fully with tumour markers (AFP, CEA, Ca 19-9), serum biochemical markers (albumin, bilirubin, aspartate aminotransferase/alanine aminotransferase, alpha-amylase) or with the presence of metastasis, age, gender or body mass index. Similar observations have been made in the case of anti-p53 antibodies [20].

Generally, overexpression and translocation of calreticulin onto cell surface or its release to the extracellular spaces is a prerequisite for development of autoimmune response against this molecule in autoimmune diseases [5]. The reason for induction of anti-calreticulin autoantibodies in cancer is, however, unknown. In cancer, we can speculate that induction of antibody response against calreticulin could be caused by surface expression of this molecule on highly proliferating tumour cells released into circulation and entrapped in adjacent lymphoid tissues (nodules) in which immune response is probably primed. Equally enigmatic is the role of anti-calreticulin antibodies in cancer. It is suggested that antibody response against autologous antigens may play a rather negative role in anti-tumour defence [31,32]. We can speculate that anti-calreticulin antibodies can interact with calreticulin in a complex with tumour peptides exposed on the surface of tumour cells, and decrease the immunogenicity of the cells. Moreover, anti-calreticulin antibodies can bind calreticulin peptides presented by major histocompatibility complex (MHC) on antigen-presenting cells and prevent T cell specific response.

The distribution of serum antibodies against calreticulin in the tested oncological patients showed high variance, suggesting a rather individual specific character of antibody response of patients. This variance is caused probably by different genetic, biological and pathological backgrounds as well as the immune status of the patients. Despite this fact, the incidence of serum anti-calreticulin antibodies in patients with HCC (IgA, IgG) and GBA (IgG) is relatively high and is comparable with autoantibodies against carcinoembryonic antigen in gastrointestinal malignancies [20]. Interestingly, 47–63% of patients with gastrointestinal malignancies (PACA, CRA, HCC) were seropositive for IgA and/or IgG anti-calreticulin antibodies. For this reason, the occurrence of anti-calreticulin antibodies in oncological patients should be evaluated when considering potential anti-cancer immunotherapy based on calreticulin.

In conclusion, our results show that calreticulin could also be a B cell molecular target in some gastrointestinal malignancies. The existence of antibodies to immunodominant epitopes of calreticulin suggest their diagnostic potential and may thereby contribute to the development of patient-specific approaches to anti-cancer immunotherapy based on calreticulin.

Disclosure

The authors have no financial conflict of interest.