Association between Neu5Gc carbohydrate and serum antibodies against it provides the molecular link to cancer: French NutriNet-Santé study

Abstract

Background: High consumption of red and processed meat is commonly associated with increased cancer risk, particularly colorectal cancer. Antibodies against the red meat-derived carbohydrate N-glycolylneuraminic acid (Neu5Gc) exacerbate cancer in "human-like" mice. Human anti-Neu5Gc IgG and red meat are both independently proposed to increase cancer risk, yet how diet affects these antibodies is largely unknown.

Methods: We used world global data to demonstrate that colorectal cancer incidence and mortality are associated with increased national meat consumption. In a well-defined large cohort, we used glycomics to measure daily Neu5Gc intake from red meat and dairy, and investigated serum as well as affinity-purified anti-Neu5Gc antibodies. Based on 24-h dietary records, daily Neu5Gc intake was calculated for 19,621 subjects aged ≥ 18 years of the NutriNet-Santé study. Serum and affinity-purified anti-Neu5Gc antibodies were evaluated by ELISA and glycan microarrays in representative 120 individuals, each with at least eighteen 24-h dietary records (aged 45-60, Q1-Q4; aged > 60, Q1 and Q4; 10 men/women per quartile).

Results: We found that high-Neu5Gc diet, gender, and age affect the specificity, levels, and repertoires of anti-Neu5Gc IgG immune responses, but not their affinity. Men consumed more Neu5Gc than women, mostly from red meat (p = 0.0015), and exhibited higher overall serum anti-Neu5Gc IgG levels by ELISA (3.94 ng/μl versus 2.22 ng/μl, respectively; p = 0.039). Detailed glycan microarray analysis against 56 different glycans revealed high Neu5Gc-specificity with increased anti-Neu5Gc IgG and altered repertoires, associated with higher consumption of Neu5Gc from red meat and cow dairy. Affinity purification of serum anti-Neu5Gc antibodies revealed increased levels and biased array repertoire patterns, without an increase in antibody affinity, in individuals consuming higher Neu5Gc levels. Furthermore, in a high-meat diet, antibody diversity patterns on glycan microarrays shifted towards Neu5Gcα3-linked glycans, increasing the α3/α6-glycans ratio score.

Conclusions: We found a clear link between the levels and repertoire of serum anti-Neu5Gc IgG and Neu5Gc intake from red meat and dairy. These precise rational methodologies allowed to develop a Gcemic index to simplify the assessment of Neu5Gc in foods that could potentially be adapted for dietary recommendations to reduce cancer risk.

Keywords: Antibodies; Cancer; Red meat; Sialic acid.

Fig. 1

Global world data show the association between colorectal cancer (CRC) and red meat intake in different nations. a CRC incidence (Pearson r = 0.7352) and mortality (Pearson r = 0.5624) in different world nations (n = 152) strongly correlate with meat intake (both p < 0.0001). International CRC age-standardized incidence and mortality rates (ASR per 100,000 person-years, including colon, rectum, anus cancers) in individuals aged 45–69 from GLOBOCAN [34] and international per capita meat intake from FAOSTAT [33] (including bovine, mutton, goat and pig; excluding poultry and aquatic mammals; Additional file 2: Data file S1). b Distribution of CRC incidence (Pearson r = 0.8482) and mortality (Pearson r = 0.7249) per nation of the highest and lowest meat intake quartiles (n = 38 each) strongly correlates (both p < 0.0001). c CRC incidence and mortality per nation of the highest and lowest quartiles of meat consumption (n = 38 each) divided by gender show a strong correlation in nations with high levels of meat intake (Kruskal-Wallis test, **p < 0.0049 and ****p < 0.0001, respectively), but not in nations with low levels of meat intake

Fig. 2

Daily Neu5Gc intake in the NutriNet-Santé study cohort. a Flow chart describing the selection of study cohort. b Distribution of the NutriNet-Santé study participants (May 2009 through May 2015) according to daily Neu5Gc intake calculated from the total mean Neu5Gc of 24-h dietary records for each individual. c Ten men and 10 women were selected per Neu5Gc intake quartile by gender (age 45–60, Q1–Q4; age > 60, Q1 and Q4), each with at least 18 dietary records. d Diversity of daily Neu5Gc intake in the selected 120 individuals (of 16,149 examined)

Fig. 3

Distribution of daily Neu5Gc intake and anti-Neu5Gc IgG by age and gender. a Significantly higher total daily Neu5Gc intake in men compared to women (age 45–60; n = 40 per gender) mostly contributed from higher consumption of red meat. Similar trend in the group aged > 60 (n = 20 per gender; median and whiskers of min-max; two-way ANOVA with Bonferroni posttest; **p = 0.0015). b Overall anti-Neu5Gc IgG (by EIA) were significantly higher in men compared to women aged 45–60, with a similar trend in the group aged > 60 (median with 95% CI, Mann-Whitney test; *p = 0.0397; ns, p = 0.0822)

Fig. 4

Glycan microarray analysis shows high anti-Neu5Gc IgG specificity and increased levels and diversity with higher Neu5Gc intake. Human serum IgG (n = 120; 1/100 dilution) detected with Cy3-anti-human IgG by glycan microarrays (24-pairs Neu5Gc-/Neu5Ac-glycans, αGal; Additional file 1: Table S3, Additional file 3: Data file S2). Relative fluorescence units (RFU) normalized to IgG (ng/μl) against printed standard curve/array [30]. a Sum serum IgG response/individual against Neu5Gc-/Neu5Ac-glycans, per Neu5Gc intake quartile, showed specific anti-Neu5Gc IgG (mean ± sem; Friedman ANOVA, p < 0.0001), a trend of elevated anti-Neu5Gc IgG at higher Neu5Gc intake. b Serum IgG/individual against αGal showed no change in levels/Neu5Gc intake quartile. c Men 45–60 stratified according to total Neu5Gc intake (Q1–Q4), contributing dietary sources plotted, showing increased Neu5Gc intake between quartiles, red meat dominant. d Men 45–60, antibodies/quartile show increased anti-Neu5Gc IgG levels between Q1 and Q2–Q4 (sum mean IgG response/glycan across individuals; different colors/specific Neu5Gc-glycan). e Pie charts of sum anti-Neu5Gc IgG response (d) divided per quartile according to reactivity against Neu5Gc-glycans with different Sia-linkages (Siaα2–3/6/8 linkages: α3, α6, α8, respectively) or underlying glycans [Lac (lactose; Galβ3Glc), Gal (galactose), type 1 (Galβ3GlcNAc), GalNAc (N-acetylgalactoseamine), LacNAc (N-acetyllactoseamine; Galβ4GlcNAc), core 1 (Galβ3GalNAcα)]. Differences in diversity at higher Neu5Gc intake, characterized by increased levels of α3-linked-Neu5Gc and Lac underlying glycans. f Women aged 45–60 stratified according to total Neu5Gc intake (Q1–Q4), increase between quartiles, similar contributions of Neu5Gc intake from red meat and dairy cow, dominance for dairy cow. g Women 45–60, anti-Neu5Gc IgG reactivity per quartile increase in levels between Q1 and Q2–Q4. h Women 45–60, pie charts of sum anti-Neu5Gc IgG response (g) divided by characteristic Neu5Gc-glycans linkage/skeleton, differences in diversity, as in men

Fig. 5

Characteristics of anti-Neu5Gc IgG and Neu5Gc in food. a Ratio between sum anti-Neu5Gc IgG against α3-linked-/α6-linked-Neu5Gc-glycans calculated in men 45–60 stratified based on meat Neu5Gc intake (Fig. S3). The α3/α6 linkage ratio in Q1 (n = 10) is lower than that in Q2–Q4 (n = 30) (Mann-Whitney test; p = 0.08; Q1 0.6080 ± 0.2806, Q2–Q4 1.472 ± 0.3722; mean ± sem). Similar comparing Q1 and Q4 (p = 0.09; Q1 0.6080 ± 0.2806, Q4 1.384 ± 0.3961). b ROC curve of α3/α6 linkage ratio score in Q1 (n = 10) compared to Q2–Q4 (n = 30) showed AUC 0.687 ± 0.098 (p = 0.08). Similar between Q1 and Q4 (p = 0.08, AUC 0.73 ± 0.1162). c Anti-Neu5Gc antibodies were affinity-purified from pooled sera of men 45–60 consuming low/high Neu5Gc from meat (Q1 5.39 ml and Q4 6.8 ml sera; n = 10 per group). Antibody yield was higher in Q4 than Q1 (8.01 versus 4.01 μg/ml serum, respectively). d, e IgG reactivity on sialoglycan microarrays (2 μg/block; detected with Cy3-anti-human IgG) revealed high specificity against Neu5Gc-glycans, no reactivity against Neu5Ac-glycans (d; each dot is IgG response/glycan), with altered diversity of glycans in Q4 over Q1 (e; Pearson r = 0.22). The α3/α6 linkage ratios: Q1 1.34, Q4 2.04. f Affinity (K_D) per glycan calculated from anti-Neu5Gc IgG on microarrays at 16 serial dilutions (40–4.9 × 10⁻³ ng/μl; 266.7–0.033 nM; non-linear fit with one-site specific binding), showing no change in affinities with higher Neu5Gc intake (mean ± sem; t test). g Gram of food to consume to reach daily nmol Neu5Gc per quartile based on Neu5Gc content (measured by DMB-HPLC). Q1 max value based on gr food to reach men Q1 average of 9443 nmol/day. Q4 min value is based on gr food to reach women Q4 average of 19,627 nmol/day. Gcemic index is the Neu5Gc content (nmol/gr) in each food item relative to the amount measured in beef (163 nmol/gr)