Human antibodies reactive to NeuGcGM3 ganglioside have cytotoxic antitumor properties

Authors


Full correspondence Dr. Ana María Hernández, Tumor Immunology Direction, Center of Molecular Immunology, 216th Street, 15th Avenue, Atabey, Playa, Havana 11600, Cuba

Fax: +53-7-2720644

e-mail: anita@cim.sld.cu

Abstract

N-glycolylated gangliosides are not naturally expressed in healthy human tissues but are overexpressed in several tumors. We demonstrate the existence of antibodies that bind (N-glycolylneuraminyl)-lactosylceramide (NeuGcGM3) and are detectable in the sera of 65 from the 100 donors (65%) tested by ELISA. From those 65 NeuGcGM3 antibody-positive donors, 35 had antibodies that were able to recognize and kill NeuGcGM3-expressing tumor cells by a complement-mediated mechanism. After complement inactivation, 11 of the 35 positive sera showed a direct cytotoxic effect on the tumor cells. This complement-independent cytotoxicity was dependent on the presence of antigen on the membrane and resembles an oncotic necrosis cell death. Both the levels of anti-NeuGcGM3 antibodies in the sera as well as the percentage of healthy donors with this immunity decreased with the age of the donor. In contrast to age and gender-matched healthy donors, we could only detect low reactivity against NeuGcGM3 in the sera of six out of 53 non-small cell lung cancer patients. These results suggest the existence of antibodies against NeuGcGM3 with antitumor immune surveillance functions, reinforcing the importance of N-glycolylated gangliosides as antitumor targets.

Introduction

Gangliosides are sialic acid-containing glycosphingolipids that are commonly found in most vertebrate cell membranes [1]. They are considered to be important targets for tumor immunotherapy not only because of their different expression patterns in healthy and transformed human tissues, but also because of their suppressive effect on immune system functions [2, 3]. In particular, N-glycolylated gangliosides are attractive targets for tumor immunotherapy because they are not normally synthesized in human tissues. This is due to a 92 bp deletion in the gene that encodes the cytidine-monophosphate-N-acetyl-neuraminc acid hydroxylase (CMAH) enzyme that catalyzes the conversion of N-acetyl to N-glycolyl sialic acid (NeuGc) [4-6]. Although humans lack this catalytic enzyme, studies have reported the presence of NeuGc in human tumors [7-10] and, in smaller amounts, in healthy adult human tissues [11]. Since an alternative pathway for NeuGc biosynthesis has not been described, the most accepted explanation for this phenomenon is the incorporation of NeuGc from dietary sources such as red meats and milk products. This incorporation occurs preferentially in tumor cells and may be due to the high division rate characteristic of tumor cells [11]. An additional proposed mechanism is that hypoxia present in the tumor microenvironment induces the expression of a sialin transporter in tumor cells resulting in enhanced incorporation of (N-glycolylneuraminyl)-lactosylceramide (NeuGcGM3) [12, 13].

We have previously reported the induction of a high-titer antibody response against NeuGc-gangliosides in melanoma, breast, small, and non-small cell lung cancer (NSCLC) patients vaccinated with the mimetic anti-idiotypic antibody 1E10 [14-17]. One of these studies, performed in NSCLC patients, showed that the anti-NeuGcGM3 antibodies actively elicited by 1E10 vaccination were able not only to recognize NeuGcGM3-expressing tumor cells but also to induce their death by an oncotic necrosis mechanism, independent of complement activation [18]. Furthermore, there was a correlation between the induction of antibodies against NeuGcGM3 and longer survival times [17]. Surprisingly, this idiotypic vaccination also elicited a “parallel set” of antibodies that recognize NeuGcGM3 and share the cytotoxic capacity against tumor cell lines but do not recognize 1E10 mAb. This suggested that this vaccination was activating a natural response against NeuGcGM3 ganglioside [15, 17]. Taking this into account, we wondered whether this cytotoxic anti-NeuGcGM3 response was present in healthy individuals. We show here that healthy humans possess antibodies against NeuGcGM3 ganglioside able to recognize and kill tumor cells expressing this antigen. These antibodies induce tumor cell death not only by complement activation, but also by a complement independent, oncotic necrosis mechanism, similar to the one observed in cancer patients treated with 1E10 mAb. To our knowledge, this is the first report of the presence in healthy humans of anti-NeuGcGM3 antibodies with this kind of cytotoxicity against tumor cells. Of note is the fact that this natural anti-NeuGcGM3 antibody response decreases with age and is absent in most of the NSCLC patients assessed.

Results

Anti-NeuGcGM3 antibody response in healthy donors

Healthy human sera were tested by ELISA for the recognition of NeuGcGM3 and NeuAcGM3 gangliosides. In 65 out of 100 donors tested, anti-NeuGcGM3 antibodies of IgM and/or IgG isotype were detected. Only four donors showed a low reactivity against NeuAcGM3 (Fig. 1A). There were no differences between male and female anti-NeuGcGM3 antibody levels (Supporting Information Fig. 1). Previous studies about antibodies against common neuronal gangliosides showed that their levels significantly decreased with age [19]. In order to determine if the natural antibody levels against NeuGcGM3 are affected by age, the antibody response in donors of different ages was compared by ELISA. As shown in Figure 1B, there was a negative correlation between the level of the anti-NeuGcGM3 response and the increase of the donors’ age. Not only was the level of the anti-NeuGcGM3 response lower, but also the percentage of healthy donors with positive anti-NeuGcGM3 response decreased with age (Fig. 1C). Next, we determined whether the lower content of anti-NeuGcM3 anti-bodies in elderly healthy donors was a consequence of a decrease in the concentration of IgM and IgG immunoglobulins. Total IgM and IgG antibody levels did not decrease with the age of the healthy donors (Supporting Information Fig. 2).

Figure 1.

Anti-NeuGcGM3 antibodies are present in healthy humans sera and decrease with age. The reactivity against NeuAc- and NeuGcGM3 gangliosides of 100 healthy human serum samples diluted 1/50 was assessed by ELISA using biotinylated goat antihuman IgG + IgM, followed by the addition of an alkaline phosphatase-streptavidin complex. (A) Specificity of the anti-NeuGcGM3 Ab response. (B) Levels of anti-NeuGcGM3 Ab response as a function of age (***p < 0.0001, Spearman test, r = –0,5409, n = 100). (C) Percentage of anti-NeuGcGM3 positive donors per age group. (A, B) Each symbol represents the mean ± SD of three independent assays for each serum samples. Dotted line represents the minimum absorbance value from which a sample could be considered positive.

Human serum antibodies bind to NeuGcGM3-expressing tumor cell lines

Having evaluated the capacity of healthy human antibodies to bind the ganglioside NeuGcGM3 by ELISA, we tested whether these antibodies are able to recognize the ganglioside in a natural context, exposed on the cytoplasmic membrane of tumor cells. To do this, the 100 human serum samples were incubated with the murine lymphocytic leukemia cell line L1210, which expresses NeuGcGM3 ganglioside [20]. NeuGcGM3 ganglioside expression on this cell line was confirmed by TLC-immunostaining (Supporting Information Fig. 3), and the antibody binding was measured by flow cytometry. Sera from 40 of the 65 healthy donors with a positive anti-NeuGcGM3 response by ELISA showed binding to L1210 cell line. Five of the sera that did not recognize NeuGcGM3 when tested by ELISA bound to this tumor cell line, presumably by binding to a different antigen. Figure 2A shows the results obtained with sera from three representative healthy donors with different levels of recognition of L1210 cells. To confirm that human serum antibodies recognize NeuGcGM3 ganglioside on the cell surface, we compared binding to L1210 with binding to cells that do not express this ganglioside. NeuGcGM3-negative cells were healthy human PBMCs and L1210 cmah-kd cells, which do not express the enzyme that catalyzes the conversion of N-acetyl to N-glycolyl sialic acid. In addition, to confirm that the antibodies recognize the NeuGc moiety expressed on a glycolipid molecule, L1210 cells were treated with trypsin to differentially remove NeuGc-glycoproteins but not NeuGc-gangliosides. As shown in Figure 2B for three representative donors, binding of the anti-NeuGcGM3 positive responders was not affected after trypsin treatment of L1210 cell surfaces. In contrast, binding was diminished in contrast to L1210 cell binding when the sera were incubated with L1210 cmah-kd cells. However, there is some degree of recognition of the L1210 cmah-kd cell line, presumably due to binding of the serum polyclonal antibodies to non-NeuGc-related antigens. No binding was detected against normal human PBMCs. Moreover, pretreatment of the positive sera with NeuGcGM3 but not with NeuAcGM3 strongly affected the percentage of L1210 stained cells (Fig. 2C). In concordance with the results obtained by ELISA, the percentage of tumor cells recognized by the healthy donors’ sera significantly decreased with increasing donor age (Fig. 2D). Also, the number of the healthy donors with serum containing antibodies able to recognize L1210 cell line decreased with age (Fig. 2E).

Figure 2.

Healthy human sera binds to NeuGcGM3-expressing tumor cell lines and this binding decreases with age. Healthy donor (HD) sera, diluted 1/5, were incubated with different tumor cell lines or human PBMCs. Reactions were developed with PE-conjugated antihuman IgG + IgM. (A) Different levels of binding to L1210 cells by anti-NeuGcGM3-positive healthy donors’ sera. Cells incubated only with the PE-conjugated secondary antibody were the negative control (shaded histogram). (B) Binding of a negative (HD2) and two representative anti-NeuGcGM3-positive HD sera (HD4 and 13) to different cells that express or do not express NeuGcGM3 ganglioside. The control group represents cells without adding human sera. Data shown are representative of two experiments performed. (C) L1210 tumor cells were incubated with healthy humans’ sera, previously incubated with saturating amounts of NeuAcGM3 or NeuGcGM3. The numbers represent the percentage of HD sera-reacting cells after subtraction of the percentage obtained by incubating the cells only with the secondary antibody. (D) Levels of healthy human sera binding to L1210 as a function of age (***p < 0.0001, Spearman test, r = –0,382, n = 100). Each symbol represents an individual healthy donor and data shown are representative of two experiments performed. (E) Percentage of donors with positive binding capacity to L1210 in different age groups.

Anti-NeuGcGM3 Abs mediate cell death of NeuGcGM3-expressing tumor cells

Next, we tested whether the anti-NeuGcGM3 antibodies present in healthy human sera were able not only to recognize but also to induce the death of L1210 cells. Forty healthy donors’ samples, with positive binding to NeuGcGM3 by ELISA and to L1210 by flow cytometry, were incubated for 4 h at 37°C with L1210 cells, and cell death was detected by PI incorporation. Thirty-five of the sera tested induced complement-mediated cell death of L1210 cells (Supporting Information Fig. 4).

The anti-NeuGcGM3 mAb 14F7 and antibodies against this antigen induced in NSCLC patients treated with the 1E10 anti-idiotypic vaccine are able to kill tumor cells by a complement-independent mechanism [18, 20]. In order to test whether the anti-NeuGcGM3 antibodies present in healthy human sera share this property, the samples were heated at 56°C for 30 min to inactivate complement before evaluating their cytotoxic capacity. Interestingly, 11 out of 35 donors’ sera that induced complement-mediated cell death still showed cytotoxic capacity after complement inactivation (Fig. 3A). There was a positive correlation between the complement-independent cytotoxicity capacity and both the levels of anti-NeuGcGM3 antibodies measured by ELISA and tumor cell binding by flow cytometry (Supporting Information Fig. 5). Furthermore, ten of these 11 donors were less than 30 years of age.

Figure 3.

Healthy human sera induces cell death of NeuGcGM3-expressing tumor cells lines by complement-dependent and complement-independent mechanisms. Tumor cells were incubated for 4 h at 37°C with sera from healthy humans, HD2 (negative anti-NeuGcGM3 responder), HD4 and HD13 (positive anti-NeuGcGM3 responders), diluted 1/5. The percentage of cell death was determined by PI incorporation. (A) L1210 tumor cell death induced by healthy sera without any treatment (Active complement, dotted line histogram) or by sera in which complement was inactivated by heating (Inactive complement, black line histogram). Filled histograms: untreated cells. (B) Tumor cell lines that express or do not express NeuGcGM3 ganglioside were incubated with healthy sera that had undergone complement inactivation. Each bar represents the mean + SEM of three replicates pooled from three independent experiments. The Control group represents cells without adding human sera. (C) L1210 tumor cells were incubated with healthy sera, previously incubated with 15 μg of NeuAcGM3 or NeuGcGM3. (B, C) Values shown represent subtraction of the PI-positive untreated cells from the percentage of PI-positive cells after incubation with sera. All the cell death induction experiments were repeated at least twice.

In order to define whether the anti-NeuGcGM3 anti-bodies present in normal human sera mediate this complement-independent cytotoxic effect, we evaluated cell death in tumor cell lines that express or do not express the NeuGcGM3 ganglioside. As shown in Figure 3B for three healthy donors, sera that induced the death of L1210 cells lacked this activity against malignant cells that do not express NeuGcGM3 ganglioside. Cytotoxicity was decreased when sera were incubated with L1210 cmah-kd cells lacking NeuGc sialoconjugates or L1210 cells cultured in the presence of D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP), an inhibitor of glycosphingolipids biosynthesis. This confirmed that the antigen recognized is an N-glycolylated-glycosphingolipid. Furthermore, a competitive incubation experiment was performed demonstrating that preincubation of the positive sera with NeuGcGM3 but not with NeuAcGM3 drastically reduced the percentage of PI positive L1210 (Fig. 3C).

Next we studied the isotype of the cytotoxic anti-NeuGcGM3 antibodies present in healthy donors that showed complement-independent cytotoxicity. As shown in Figure 4A, all the positive donors had anti-NeuGcGM3 IgM antibodies when the response was measured by ELISA. Only one donor also had IgG anti-NeuGcGM3 antibodies. After incubation of the cytotoxic sera with L1210 cells we found that the binding was mediated only by IgM antibodies, even in the one donor that showed an anti-NeuGcGM3 IgG antibody response when measured by ELISA (Fig. 4B). To prove that the IgM antibodies were responsible for the cytotoxic effect detected through the PI incorporation assay by flow cytometry, IgG and IgM fractions were separated from one of the NeuGcGM3 binding healthy donors (HD 4) by protein G purification and compared with a non binding control sample (HD2). As expected, when both IgG and IgM fractions were incubated with L1210 cells only the IgM fraction showed cytotoxic capacity (Fig. 4C).

Figure 4.

The Ig that recognizes and kills NeuGcGM3-positive tumor cells are of the IgM isotype. (A) The isotype of the anti-NeuGcGM3 antibodies was determined by ELISA for ten cytotoxic sera diluted 1/50 using biotinylated goat antihuman IgG or antihuman-IgM, followed by the addition of alkaline phosphatase-streptavidin complex. (B) The isotype of the antibodies that recognize L1210 cells, visualized by flow cytometry using FITC-conjugated antihuman IgG or antihuman IgM conjugates. (C) The isolated IgG and IgM fractions from healthy donors HD2 and HD4 were incubated for 4 h with L1210 cells. The lytic capacity of each fraction was detected by flow cytometry. Data shown are representative of two experiments performed.

Cytotoxic anti-NeuGcGM3 Abs induce tumor cell death by an oncotic necrosis mechanism

Having identified anti-NeuGcGM3 antibodies in healthy human sera with the potential to induce tumor cell death independent of complement cascade activation, we further characterized this death mechanism.

First, we studied the kinetics of the cell death induction and the effect of temperature on the cytotoxic effect. L1210 cells were incubated with heat-inactivated donors’ sera at 37 or 4°C for 30 min, 2 and 4 h, respectively. After 30 min of incubation, PI positive cells were already detectable, showing the rapid nature of this cytotoxic mechanism (Fig. 5A). Furthermore, there were no differences in the percentage of dead cells when the incubation took place at 4° or 37°C (Fig. 5B). This result suggests an energy-independent mechanism, differing in this regard from apoptosis [18]. One of the major hallmarks of apoptosis induction is the activation of caspases. Among these proteins, caspase 3 converges in the two main pathways of apoptosis [21]. No significant caspase-3 activation was detected in the L1210 cells after incubation with cytotoxic healthy human sera for 4 h, the time at which approximately 40% of the cells already incorporated PI (Supporting Information Fig. 6).

Figure 5.

Induced complement-independent cell death exhibits features of an oncotic necrosis. (A) L1210 cells were incubated with healthy donor sera for 30 min at 37°C. The cells were stained with FITC-conjugated goat antihuman Ig and visualized by fluorescence microscopy (40×). PI was added to visualize dead cells. (B) L1210 cells were incubated with healthy donor sera at 4°C or 37°C, for 30 min (blue line histogram), 2 h (red line histogram), or 4 h (green line histogram).The percentage of cell death was determined by the PI incorporation assay. The numbers represent subtraction of the percentage of PI-positive untreated cells (filled histogram) from the percentage of PI-positive cells after incubation with sera. (C) Forward scatter plots of L1210 cells incubated with healthy donor sera for 30 min at 37°C. (D) Morphologic changes in L1210 cells after 4 h incubation at 37°C with healthy donor sera visualized by hematoxylin and eosin staining: oncotic necrotic cells with swelling, membrane disruption, and karyolysis (black arrows) (100× magnification). (C and D) Control represents cells without adding human sera. 14F7 mAb and CIGB 300 were used as positive controls of oncotic necrosis and apoptosis, respectively. Data shown are representative of three independent experiments.

Figure 6.

The anti-NeuGcGM3 antibody levels in untreated NSCLC patients are lower than in healthy humans. (A) Reactivity against NeuGcGM3 gangliosides of 100 healthy donors and 53 NSCLC patients’ sera samples diluted 1/50 was assessed on ELISA plates coated with NeuGcGM3 using biotinylated goat antihuman IgG + IgM, followed by the addition of an alkaline phosphatase-streptavidin complex. For statistical comparison the sample was divided in three age groups attending to the age of the NSCLC patients. **p ≤ 0.01, Mann–Whitney U-test, one-tailed. (B) Total IgM and IgG concentration in the sera of the cancer patients measured by ELISA using biotinylated goat antihuman IgG and anti-human IgM. ***p ≤ 0.001, Mann–Whitney U-test, one-tailed. Each symbol represents an individual patient and the bars represent the means. Data shown are pooled from three experiments performed.

Then, we studied the morphological changes of the affected cells. Forward scatter plots showed that the size of the cells increased after the incubation with the cytotoxic sera, suggesting that recognition by anti-NeuGcGM3 antibodies induced cell swelling (Fig. 5C). This effect is comparable to the morphological changes induced by 14F7 mAb and is quite different from the cell shrinkage observed for CIGB300, used as apoptosis inducer control.

Finally, after incubation with sera, the L1210 cells were stained with hematoxylin and eosin (H&E) and visualized by light microscopy. This examination revealed that after 4 h incubation, cells treated with cytotoxic sera had the morphology of oncotic necrotic cells with cellular swelling, membrane disruption, and karyolysis (Fig. 5D). No chromatin condensation or apoptotic body formation, hallmarks of apoptosis, were detected in the stained cell nuclei after incubation with the cytotoxic sera.

Anti-NeuGcGM3 antibody levels in healthy donors and cancer patients

Due to the antitumor potential of the detected anti-NeuGcGM3 antibodies, we evaluated their presence in cancer patients. We compared 53 NSCLC patients with gender- and age-matched healthy donors. Analysis of antibody levels in the sera from these patients by ELISA revealed statistically significant lower anti-NeuGcGM3 responses in NSCLC patients less than 60 years of age than in healthy donors (Fig. 6A). We detected low levels of anti-NeuGcGM3 antibodies only in six patients, two of which also reacted with NeuAcGM3 ganglioside (Supporting Information Fig. 7). These six NSCLC patients were not able to recognize the L1210 tumor cell line (data not shown). When we measured the total IgM and IgG concentration in the sera of the cancer patients, although the levels of total IgM and IgG antibodies did not change with age (data not shown), there was a significantly lower total IgM level in cancer patients’ sera when compared with that of healthy donors. In contrast, the total levels of IgG in the NSCLC patients were similar to the levels observed for healthy donors (Fig. 6B).

Discussion

Natural antibodies have been considered to be important in the primary defense against invading pathogens [22], the clearance of damaged structures, dying cells and oxidized epitopes [23], and the modulation of cell functions [24]. But also, naturally occurring antibodies could play a role in the protection against neoplastic transformation [25-29]. In this study, we describe the presence of antibodies against NeuGcGM3 ganglioside, circulating in the sera of healthy adult individuals. NeuGcGM3 ganglioside is not only overexpressed on tumor cell membranes, but are also important for tumor development due to its suppressive effect on immune system function [2]. Sixty-five healthy donors’ sera out of 100 tested bound to NeuGcGM3 by ELISA, and did not recognize the acetylated form of this ganglioside. This result is in concordance with a previous result about reactivity against different N-glycolylated compounds of 16 healthy donors, reported by Padler-Karavani et al. [30]. Previous reports have shown the existence of a naturally occurring immunity against glycolipidic antigens, specifically gangliosides. Some of these reactivities have been associated with the induction of pathological alterations, as is the case for the antibodies against ganglioside complexes, such as GD1a and GD1b, or GM1 and GD1a in Guillian–Barre syndrome [31]. However, other studies suggest that naturally occurring antiganglioside antibodies may play an important role in immune surveillance against tumors in humans [25, 27]. Different studies about the antibody response against Neu5Gc containing molecules have shown opposite findings regarding its impact on tumor growth. In a mouse model of human-like Neu5Gc deficiency, transferred polyclonal syngeneic mouse anti-Neu5Gc antibodies interacted with Neu5Gc-positive tumors generating chronic inflammation and facilitating tumor progression [32]. On the other hand, the same group later reported a reduction in tumor growth in mice passively treated with higher amounts of human anti-Neu5Gc antibodies, arguing that the effect on tumor progression or suppression depends on the dose of the anti-Neu5Gc antibodies [33]. Another explanation for the contrasting results could reside in the fact that Neu5Gc-containing glycans are diverse and presented on many different glycoconjugates, with further structural diversity due to different possible Neu5Gc modifications and linkage differences [34]. Thus, in a polyclonal anti-Neu5Gc pool there can be antibodies with different fine specificities and properties. In fact, the anti-Neu5Gc antibodies purified in the previous reports [30] had minimal reaction with NeuGcGM3 ganglioside, the Neu5Gc-containing antigen recognized by the healthy donors’ sera evaluated in our study.

The anti-NeuGcGM3 antibodies present in the healthy donors’ sera were not only able to recognize NeuGcGM3 coated on ELISA plates, but also when NeuGcGM3 was expressed on tumor cell membranes. We confirmed that the binding to L1210 cells was dependent on the presence of NeuGcGM3. First, we demonstrated that the sera detected an N-glycolylated molecule, by showing that the antibodies in the sera did not recognize L1210-cmah-kd cells. Next, we demonstrated that the detected glycolylated molecule was not a glycoprotein, since the binding was not affected by trypsin treatment. Finally, we blocked cell line recognition by preincubation of the sera with NeuGcGM3. This binding was not inhibited by NeuAcGM3, a ganglioside that differs only in the presence of a hydroxyl group in the N-glycolylated variant. Furthermore, we demonstrated that these antibodies were able not only to recognize but also to induce the death of NeuGcGM3-expressing tumor cells by complement cascade activation, and also by a complement-independent mechanism. This cell death mechanism is different from apoptosis, since it was temperature independent, did not induce caspase activation, and chromatin condensation or apoptotic body formation were not detectable. The incubation of the cells with sera increased the size of the cells and disrupted cell membranes. These characteristics resemble the oncotic cell death reported for anti-NeuGcGM3 mAb 14F7, and for anti-NeuGcGM3 antibodies induced in NSCLC patients treated with 1E10 anti-idiotypic vaccine [18, 20]. This cell death depended on the expression of antigen on the cell membrane, since the cytotoxicity was completely abrogated on tumor cells that do not express NeuGcGM3, or by preincubation of the donors’ sera with saturating amounts of this ganglioside. Nguyen et al. reported the capacity of healthy donors’ sera to bind and kill human leukemic cells and activated T cells that were exogenously fed with Neu5Gc, but in these studies the detected cell death was mediated only by a complement-mediated mechanism [12]. The antibodies that recognized NeuGcGM3-expressing cells were of the IgM isotype. The IgM fraction isolated from one of the healthy donor's sera retained the capacity to induce complement-independent death of the tumor cells. To our knowledge, this is the first report of anti-NeuGcGM3 antibodies that are able to induce the oncotic cell death of antigen-expressing tumor cells without the necessity of any other immune component. These results suggest the existence of antibodies with antitumor potential, which could contribute to tumor immune surveillance. It is interesting to observe that the levels of anti-NeuGcGM3 antibodies decreased as the age of the donors increased. Not only is the level of anti-NeuGcGM3 antibodies lower in elderly donors, but also the percentage of responding donors decreases with age. An age-associated decrease in antibody levels against foreign antigens was first reported more than 70 years ago [35], supporting the idea of an immune deficiency state in the elderly. However, this seems to be a phenomenon dependent on the nature of the antigen and the cells involved in the different responses, since other studies have shown that the concentration of serum antibodies against a variety of self-antigens such as thyroglobulin, DNA, and IgG, increases with age [36]. In fact our results demonstrate that the total amount of IgG and IgM did not decrease with age, suggesting that it is not the amount of antibodies but the antibody repertoire that changes with age. One possible explanation for the decrease in antibody levels with increasing age involves an impaired capacity of T cells to facilitate the maturation of B cells in the periphery and the generation of a diverse B-cell repertoire from precursors within the bone marrow [37]. According to this theory, the response against T-independent antigens should not be affected by age [38]. However, the antibody response against not only NeuGcGM3 but also against other tumor related gangliosides (T-independent antigens), significantly decrease with increasing donor age [19]. Another possibility could be a reduction in the B-cell population responsible for the production of naturally occurring antibodies. Recently, Griffin et al. described a human B-cell population equivalent to mouse B1 cells [39], the main source of murine natural antibodies [40]. These researchers showed that human B1 cells decline with age. The reduction of B cells secreting antibodies with immune surveillance properties could explain, at least in part, the increased susceptibility of aged individuals to neoplastic transformation. In fact, it is interesting to observe that in NSCLC patients, who had not been exposed to any antitumor treatment (including radio or chemotherapy), we could not detect cytotoxic anti-NeuGcGM3 antibodies in the conditions used for our study. This behavior was observed even in those patients less than 60 years of age. Only six of the 53 NSCLC patients studied had a low response against NeuGcGM3, and their sera were not able to bind to tumor cells expressing the antigen. The levels of IgG and IgM antibodies did not decrease with the age of the cancer patients, however, we did detect a significantly lower total IgM concentration in the cancer patients’ sera when compared with healthy donors’. In contrast, the IgG concentrations were similar, suggesting that the IgM reduction is not due to a general state of immunosuppression in these patients. The reduced level of anti-NeuGcGM3 antibodies detected in these patients could be a consequence of the low total IgM levels, the isotype of the antibodies that recognize NeuGcGM3. But this specificity could be particularly affected, resembling what we observed for elderly healthy donors. In the case of these cancer patients, the observed behavior could be due to the anti-NeuGcGM3 antibody-secreting B-cell population being affected, or to the capacity of this B-cell population to secrete antibodies with this specificity being inhibited. By idiotypic vaccination, however, we have been able to boost this kind of immune response in cancer patients, which suggests that these cells are not completely deleted [17]. Another possibility is that, in NSCLC patients, anti-NeuGcGM3 antibodies form immune complexes with gangliosides released from the tumor cells, which might affect their detection. This phenomenon could also result from the recruitment of such antibodies to the tumors since the presence of NeuGcGM3 in NSCLC tumor samples has been reported [41-43]. To our knowledge this is the first report showing that the levels of anti-NeuGcGM3 antibodies are lower in cancer patients in comparison with healthy donors. Previous work reported that, depending on the ganglioside and the kind of tumor, higher or lower concentrations of antibodies against gangliosides in the sera of cancer patients with respect to healthy donors, could have a prognostic value [25, 44]. Further studies are needed to evaluate whether this is also the case for the antibody response against NeuGcGM3.

Currently, we are carrying out experiments to elucidate the cause of the reduced levels of anti-NeuGcGM3 antibodies in NSCLC patients and extending these determinations to other kinds of tumors. In particular, we are trying to understand if the absence of this kind of response is a consequence of disease, or one of the causes increasing susceptibility to malignant transformation.

Materials and methods

Gangliosides and cells

Gangliosides NeuAcGM3 and NeuGcGM3, purified from dog and horse erythrocytes (at least 98% purity), respectively, as described earlier [45], were provided by Dr. A. Carr (Center of Molecular Immunology, Havana, Cuba). L1210 murine lymphocytic leukemia cell line was obtained from the American Tissue Type Culture Collection. L1210 cmah-kd cell line was generated in our institution as previously described [46] by CMP-Neu5Ac-neuraminic acid hydroxylase gene knock-down using a specific shRNA. Cells were grown in DMEM (Gibco-BRL, Paisley, UK) supplemented with 10% heat-inactivated FCS (Invitrogen, USA), 2 mM L-glutamine, 25 mM HEPES, 100 U/mL penicillin, 100 μg/mL streptomycin, and maintained at 37°C with 5% CO2. L1210 cells were treated with trypsin (Gibco) 0.05% for 5 min at 37°C for testing the importance of the gangliosides in binding and cytotoxicity experiments. Fresh blood from healthy volunteers was centrifuged over a Ficoll-Hypaque density gradient to obtain PBMCs as described earlier [47].

Human serum samples

One hundred normal serum samples were obtained from healthy adults of both genders and various ethnic backgrounds. None of the donors presented the evidence of infectious disease, cancer, atherosclerosis, or autoimmune diseases at the time of blood collection. Cancer patients’ sera were obtained from 53 advanced NSCLC patients who had not been exposed to any antitumor treatment, with approval from the Institutional Review Board of the Hermanos Ameijeiras Hospital. The cancer patients were gender- and age-matched with the healthy donors. Written informed consent was obtained in advance from all the volunteers. The serum samples were decomplemented by heat inactivation for 30 min at 56°C. All sera were stored at –20°C until use.

Detection of anti-NeuGcGM3 antibodies in human sera

Anti-NeuGcGM3 antibodies present in human sera were detected by ELISA with some modifications as previously described [48]. Briefly, 96-well polystyrene plates (PolySorp, Nunc, Denmark) were coated with NeuGcGM3 or NeuAcGM3 at a saturating concentration of 200 ng/well in methanol. Plates were allowed to dry for 2 h at 37°C and then blocked with 4% human serum albumin in PBS for 2 h at 4°C. Control wells, coated only with methanol, were equally treated with blocking solution. Diluted human sera (1/50 in PBS-0.4% human serum albumin) were added to the wells and incubated overnight at 4°C. The plates were washed six times with PBS containing 0.1% Tween 20 (PBST) and then incubated with biotin-conjugated goat antihuman IgG + IgM (Jackson ImmunoResearch Laboratories, Inc, West Grove, PA, USA) for 1.5 h at RT. After washing in the same conditions, alkaline phosphatase conjugated streptavidin (Jackson ImmunoResearch Laboratories) was added and incubated for an additional 1.5 h at RT. Finally, a substrate solution consisting of 1 mg/mL p-nitrophenylphosphate in diethanolamine buffer, pH 9.8, was added to the plates and the absorbance was measured at 405 nm in an ELISA reader (Organon Teknika, Salsburg, Austria). To consider that a serum sample had a positive reaction to a particular ganglioside, values of absorbance had to be ≥0.25 and at least three times the absorbance value obtained by incubating the serum in wells containing no ganglioside (only methanol dried on the wells). Assays were performed in triplicate for each sample. The optical densities (ODs) of the blanks were less than 0.1.

Quantification of serum Igs

The levels of serum IgM and IgG were determined by ELISA. Microtitre plates (MaxiSorp, Nunc) were coated with 50 μL of antihuman IgG or antihuman IgM, at 2 or 5 μg/mL, respectively. Serum Igs (IgM and IgG) levels were determined using alkaline phosphatase-coupled goat antihuman IgM or anti-human IgG (Sigma-Aldrich). The absorbance was measured at 405 nm in an ELISA reader (Organon Teknia). Absorbance values were quantified into milligrams per millilitter using the standard dilution curves of the corresponding purified human Igs (Sigma-Aldrich).

Extraction, isolation, separation of acidic glycosphingolipid by HPTLC and immunostaining

Glycosphingolipid extraction from L1210 tumor cells was performed as reported previously [49]. The acidic glycosphingolipid fraction was desiccated and then dissolved in chloroform/methanol (2:1; v/v) for developing on high-performance thin-layer chromatography (HPTLC) on precoated thin-layer plates (Merck, Darmstadt, Germany) in the solvent system consisting of chloroform/methanol/0.25% KCL and 2.5 M NH3 (5:4:1; v/v). Gangliosides were visualized with orcinol stain [50]. Immunostaining with 14F7 mAb on HPTLC plates was performed as previously reported [50]. The plates were incubated with biotinylated goat antimouse IgG (Jackson Immunoresearch Laboratories) and strepdavidin-alkaline phosphatase (Jackson Immunoresearch Laboratories). Color was developed with an alkaline-phosphatase (AP)-conjugated substrate kit (Biorad, CA, US).

Separation of IgM and IgG antibody fractions from human sera

Serum IgM and IgG fractions were isolated using a protein G mini column (Pro-Chem Inc., MA, USA) following the manufacturer's instructions. Purity and reactivity against gangliosides of the eluted (IgG) and unbound (IgM) fractions were tested by ELISA as described above. The column fractions were screened both for binding and cytotoxic activity against L1210 tumor cells (see below).

Flow cytometry analysis

To assess the binding of anti-NeuGcGM3 Abs present in human sera, the cells were blocked in PBS containing 1% FCS for 20 min on ice. Human serum samples, diluted 1/5, were incubated with 105 cells for 30 min on ice. After washing with cold PBS, cells were incubated with PE-conjugated goat antihuman Igs (IgM + IgG), FITC-conjugated goat antihuman IgG or FITC-conjugated goat antihuman IgM (Jackson ImmunoResearch Laboratories), for 30 min on ice. The percentage of positive stained cells was determined in a FACScan flow cytometer (Becton Dickinson, San Jose, CA, USA). The WinMDI 2.9 program was used to analyze a total of 104 cells acquired on every assay. To be considered positive, a serum sample percentage of binding had to be ≥15% and at least two times the percentage obtained by incubating the cells only with the secondary antibody.

Serum cytotoxicity assay

Human sera, diluted 1/5, were incubated with 105 tumor cells or PBMCs in DMEM supplemented with 1% FCS at 37 or 4°C, for the indicated times. Cell death induction was detected by the addition of propidium iodide (PI; Sigma-Aldrich, St. Louis, MO, USA) at a final concentration of 10 μg/mL and analyzed by flow cytometry. Similar experiments were performed with serum samples previously heated at 56°C for 30 min to inactivate complement and with both IgG and IgM fractions isolated from the serum of healthy donors HD2 and HD4.

We considered a serum sample to be positive when the percentage of dead cells was ≥20% and at least two times the percentage observed for the untreated cells. To determine if the cytotoxic effect of serum samples was mediated by the anti-NeuGcGM3 antibodies, L1210 cells were cultured for 3 days with 10 μmol/L of D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (Matreya, LLC, PA, USA), an inhibitor of glucosylceramide synthetase that affects glycosphingolipids biosynthesis. With this same objective, before cell death induction, serum samples were incubated with 15 μg of NeuGcGM3, previously air dried and sonicated in PBS, in order to block the anti-NeuGcGM3 antibodies. As a control for apoptosis induction, L1210 cells were treated with 10 μM CIGB 300 for 20 min at 37°C [51], an apoptosis inducer kindly provided by Dr. Perea from the Centre of Genetic Engineering and Biotechnology.

Optical microscopy

To determine the nuclear and membrane morphology, after incubation with serum samples during the indicated times, L1210 cells were dried on microscope slides, fixed with 4% formaldehyde and stained with H&E. Apoptotic or oncotic necrotic cells were identified by morphological criteria. Cell death with chromatin condensation, cell shrinkage and formation of apoptotic bodies was regarded as apoptosis. Morphologic criteria such as karyolysis, cell membrane disruption and cellular swelling were used to determine oncotic necrosis [52, 53].

To visualize antibody binding to the cell membrane and incorporation of PI after 30 min of treatment with the sera, cells were washed and blocked with PBS containing 1% FCS, and incubated with FITC-conjugated goat antihuman Igs (IgM + IgG) (Jackson ImmunoResearch Laboratories) for 30 min at room temperature in the dark and with PI for 10 min at a final concentration of 10 μg/mL. After washing with PBS, cells were immediately visualized on a fluorescence microscope (OLYMPUS BH-2, Tokyo, Japan).

Activated caspase-3 detection assay

The involvement of caspase-3 in induced cell death was studied after 2 or 4 h of incubation of L1210 cells with the serum samples. Next, the cells were stained with FLICA (SR-DEVD-FMK; Immunochemistry Technologies, Bloomington, IN, USA), following the manufacturer's instructions. The cells were visualized on a fluorescence microscope (OLYMPUS BH-2).

Statistical analysis

Data analyses were performed using Graph-Pad Prism 5.03 Software. Each experiment was repeated at least twice. Unless specified otherwise, data is described as mean ± SD. Mann–Whitney U test was used as a nonparametric test for pair-wise comparisons. Correlation analysis was done using Spearman Rank Analysis. p-values below 0.05 were considered significant.

Acknowledgments

The authors want to thank all the subjects who volunteered to participate in this study. The work of Zuyen Gonzáles, Esperanza Hechevarría, and Belkys Gómez collecting the blood samples is greatly acknowledged. The authors also want to thank Dr. Thomas Rothstein and Dr. Daniel Griffin for critical reading of the manuscript. This work was supported by the Center of Molecular Immunology.

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
CMAH

cytidine-monophosphate-N-acetyl-neuraminic acid hydroxylase

HD

healthy donor

HPTLC

high performance thin layer chromatography

NeuGc

N-glycolyl sialic acid

NeuGcGM3

(N-glycolylneuraminyl)-lactosylceramide

NSCLC

non-small cell lung cancer

Ancillary