Racially restricted contribution of immunoglobulin Fcγ and Fcγ receptor genotypes to humoral immunity to human epidermal growth factor receptor 2 in breast cancer
Correspondence: J. P. Pandey, Department of Microbiology and Immunology, Medical University of South Carolina, Charleston, SC 29425-2230, USA.
Tumour-associated antigen human epidermal growth factor receptor 2 (HER2) is over-expressed in 25–30% of breast cancer patients and is associated with poor prognosis. Naturally occurring anti-HER2 antibody responses have been described in patients with HER2 over-expressing tumours. There is significant interindividual variability in antibody responsiveness, but the host genetic factors responsible for this variability are poorly understood. The aim of the present investigation was to determine whether immunoglobulin genetic markers [GM (genetic determinants of γ chains)] and Fcγ receptor (FcγR) alleles contribute to the magnitude of natural antibody responsiveness to HER2 in patients with breast cancer. A total of 855 breast cancer patients from Japan and Brazil were genotyped for several GM and FcγR alleles. They were also characterized for immunoglobulin (Ig)G antibodies to HER2. In white subjects (n = 263), GM 23-carriers had higher levels of anti-HER2 antibodies than non-carriers of this allele (p = 0·004). At the GM 5/21 locus, the homozygotes for the GM 5 allele had higher levels of anti-HER2 antibodies than the other two genotypes (P = 0·0067). In black subjects (n = 42), FcγRIIa-histidine/histidine homozygotes and FcγRIIIa-phenylalanine/valine heterozygotes were associated with high antibody responses (P = 0·0071 and 0·0275, respectively). FcγR genotypes in white subjects and GM genotypes in black subjects were not associated with anti-HER2 antibody responses. No significant associations were found in other study groups. These racially restricted contributions of GM and FcγR genotypes to humoral immunity to HER2 have potential implications for immunotherapy of breast cancer.
Human epidermal growth factor receptor 2 (HER2) is over-expressed in 25–30% of breast cancer patients and is associated with poor prognosis. It is a prominent target for both active (vaccine-based) and passive (antibody-based) immunotherapy in this malignancy. Current anti-HER2-based immunotherapy (trastuzumab) has improved the quality of life of women with breast cancers that over-express HER2. However, this therapy has some drawbacks, such as the need for repeated transfusion, development of resistance and inability of the antibodies to penetrate solid tumours. A vaccine that engages the patients' own immune system would circumvent many of these hurdles and, using immunological memory, would provide sustained anti-tumour immunity. Several HER2-based vaccines are on trial in patients with breast cancer [1-3]. To evaluate clearly the efficacy of these trials, it is essential to understand the host factors that influence antibody responsiveness to HER2. Naturally occurring anti-HER2 antibody responses have been described in patients with HER2 over-expressing tumours . Although, in most cases, antibody concentration is too low to be of therapeutic benefit, some breast cancer patients with very high (>1:5000) anti-HER2 titres have been described . Thus, some people are naturally high responders to HER2, while others are low responders. This characteristic, unless taken into account, could confound the evaluation of vaccine efficacy trials.
Our knowledge of host genetic factors that influence interindividual variability in the magnitude of humoral immunity to tumour antigens is very limited. In a previous investigation of breast cancer patients from Estonia, we reported significant associations between particular γ marker (GM) alleles and endogenous anti-HER2 antibody responses . GM gene frequencies differ significantly across population groups . Also, because of almost absolute linkage disequilibrium between particular alleles within a racial group, every major race is characterized by a unique array of GM haplotypes . Therefore, it is essential to examine diverse population groups, as the GM alleles/haplotypes associated with anti-HER2 antibody responses might not be the same in all groups. In the present investigation, we aimed to determine the contribution of GM alleles to the magnitude of anti-HER2 antibody responses in breast cancer patients from Japan and Brazil. Additionally, we wished to investigate whether particular FcγR alleles, implicated in the antibody-dependent cellular cytotoxicity (ADCC) of HER2 over-expressing breast cancer cells [9, 10], also influence the magnitude of endogenous anti-HER2 antibody responses in patients with breast cancer.
Materials and methods
The experimental design, recruitment criteria and the demographics of the study population have been described in detail elsewhere . Briefly, blood samples from histologically verified breast cancer patients were obtained at four hospitals in Nagano, Japan and at eight hospitals in São Paulo, Brazil. The study protocol was approved by the Medical Research Ethics Committee/Institutional Review Board of the respective institutions. All subjects provided informed consent. The study population consisted of the following: 263 patients of Caucasian descent (Brazil), 42 patients of African descent (Brazil), 80 patients of Japanese descent (Brazil), 80 patients from the Brazilian mixed race population and 397 patients from Nagano, Japan.
Anti-HER2 antibody determination
Serum/plasma samples were stored at −80°C until required. Using a recombinant protein (ProSpec-Tany TechnoGene Ltd, Rehovot, Israel) containing the extracellular domain (Met 1-Thr 652) of HER2 as antigen, immunoglobulin (Ig)G antibody levels were determined by a previously described enzyme-linked immunosorbent assay (ELISA) .
For the determination of IgG1 markers GM 3 and 17 (arginine→lysine, a G→A substitution in the CH1 region of the γ1 gene), we used a predesigned TaqMan® genotyping assay from Applied Biosystems Inc. (Foster City, CA, USA), employing the following primers and probes: forward primer 5′-CCCAGACCTACATCTGCAACGTGA-3′, reverse primer 5′-CTGCCCTGGACTGGGACTGCAT-3′; and reporter 1 (GM 17-specific) VIC-CTCTCACCAACTTTCTTGT-NFQ, reporter 2 (GM 3-specific) FAM-CTCTCACCAACTCTCTTGT-NFQ.
GM 23 – valine to methionine, a G to A substitution in the CH2 region of the γ2 gene – was determined by a nested polymerase chain reaction–restriction fragment length polymorphism (PCR–RFLP) method. In brief, a 915 base pairs (bp) region of the γ2 gene that incorporates the sites for the allelic substitutions was amplified as described by Brusco et al. , using the following primers: 5′-AAATGTTGTGTCGAGTGCCC-3′ and 5′-GGCTTGCCGGCCGTGGCAC-3′. A 197 bp segment was amplified further from this 915 bp fragment using the following primers: 5′-GCACCACCTGTGGCAGGACC-3′ and 5′-TTGAACTGCTCCTCCCGTGG-3′. After digestion of the amplified product with the restriction enzyme NlaIII, the following products corresponding to the three genotypes were obtained: GM 23+, 90 bp, 63 bp and 44 bp; GM 23−, 134 bp and 63 bp; GM 23+, 23−, 134 bp, 90 bp, 63 bp and 44 bp.
IgG3 markers GM 5 and 21 were determined by a previously described PCR–RFLP method .
The activating receptors FcγRIIa and FcγRIIIa are genetically polymorphic: a change in the nucleotide at position 497 of FcγRIIa gene from A to G results in a change of amino acid histidine→arginine (H/R131); a change in the nucleotide at position 559 of the FcγRIIIa gene from T to G results in phenylalanine→valine substitution (F/V158). The single nucleotide polymorphisms responsible for the allelic variation were determined previously by the TaqMan® genotyping assays .
Genotyping was performed blinded to the racial status of the subjects. Due to technical reasons (usually lack of amplification), certain samples were not typed for certain alleles at both GM and FcγR loci, causing slight variations in the sample number for each genotype.
Genotype frequencies were in Hardy–Weinberg equilibrium in all groups except the mixed race population, which could be due to population admixture. This group was excluded from further analyses. Linear regression models were constructed within population groups. Tests of genotype models – 2 df tests with no assumptions about inheritance models – as well as 1 df tests of additive, dominant and recessive effects of the minor allele were constructed. The phenotype of interest, anti-HER2 antibody level (μg/ml), was log-transformed to avoid violating model assumptions. All tests were two-sided with α = 0·05.
The distribution of GM and FcγR genotypes among white study subjects in relation to the mean levels of IgG antibodies (μg/ml) to HER2 is given in Table 1. The association between GM 23 genotypes and the level of anti-HER2 antibody responses was significant for the genotype model as well as for additive and dominant models, but not for the recessive model of inheritance. The anti-HER2 antibody levels associated with GM 23+/GM 23+ homozygotes and GM 23+/GM 23− heterozygotes were similar (0·26 versus 0·32 μg/ml) and significantly higher than those associated with GM 23−/GM 23− homozygotes (0·04 μg/ml; P = 0·004). The genotypes at the GM 5/21 locus were also associated with anti-HER2 antibody responses at the genotype, additive and dominant models of inheritance. Subjects homozygous for the GM 5 allele, which is in linkage disequilibrium with GM 23, had significantly higher levels of anti-HER2 antibodies than GM 5/GM 21 heterozygotes and GM 21/GM 21 homozygotes (0·32 versus 0·06 μg/ml; P = 0·0067).
Table 1. Tests of associations between γ markers (GM) and FcγR variants and anti-human epidermal growth factor receptor 2 (HER2) antibody levels (μg/ml) in white breast cancer patients (n = 263)
|GM3/17||3/3||92||0·19 ± 0·66||0·1933||0·0991||0·2976||0·0807|
|3/17||125||0·26 ± 1·12|| || || || |
|17/17||43||0·05 ± 0·08|| || || || |
|GM23+/− ||23+/23+||48||0·26 ± 0·89||0·0160||0·0120||0·0040||0·2745|
|23+/23−||112||0·32 ± 1·19|| || || || |
|23−/23−||100||0·04 ± 0·08|| || || || |
|GM5/21||5/5||143||0·32 ± 1·17||0·0198||0·0276||0·0067||0·8483|
|5/21||95||0·06 ± 0·16|| || || || |
|21/21||20||0·06 ± 0·09|| || || || |
|FcγRIIa||H/H||52||0·30 ± 1·51||0·9282||0·8902||0·8681||0·7001|
|H/R||140||0·18 ± 0·66|| || || || |
|R/R||70||0·17 ± 0·58|| || || || |
|FcγRIIIa||F/F||123||0·29 ± 1·22||0·7372||0·8928||0·8098||0·7474|
|F/V||109||0·12 ± 0·35|| || || || |
|V/V||25||0·13 ± 0·33|| || || || |
FcγRIIa and FcγRIIIa genotypes – neither individually nor epistatically with GM genotypes – were associated with antibody responsiveness to HER2 in this group of patients.
The distribution of GM and FcγR genotypes among black study subjects in relation to the mean levels of IgG antibodies (μg/ml) to HER2 is given in Table 2. In contrast to the results in white subjects, none of the GM genotypes were associated with anti-HER2 antibody responsiveness in this group. However, genotypes at both FcγR loci contributed significantly to humoral immunity to HER2 in black subjects with breast cancer. FcγRIIa-H/H homozygotes had significantly higher levels of anti-HER2 antibodies than FcγRIIa-H/R heterozygotes and FcγRIIa-R/R homozygotes (0·45 versus 0·12 μg/ml; P = 0·0071). The associations were significant for the genotype and recessive models, but not for additive and dominant models of inheritance. At the FcγRIIIa locus, the F/V heterozygotes had significantly higher anti-HER2 antibody levels than the two homozygotes (0·32 versus 0·08 and 0·02 μg/ml; P = 0·0275). These associations were significant for the genotype and dominant models, but not for additive and recessive models of inheritance. No significant associations (P > 0·2) were found in the Japanese subjects living in Japan or Brazil (data not shown).
Table 2. Tests of associations between γ markers (GM) and FcγR variants and anti-human epidermal growth factor receptor 2 (HER2) antibody levels (μg/ml) in black breast cancer patients (n = 42)
|GM3/17||3/3||2||0·32 ± 0·44||0·7947||0·9886||0·8170||0·5738|
|3/17||17||0·14 ± 0·36|| || || || |
|17/17||21||0·21 ± 0·41|| || || || |
|23+/23−||12||0·23 ± 0·45|| || || || |
|23−/23−||28||0·17 ± 0·36|| || || || |
|GM5/21||5/5||24||0·27 ± 0·47||0·2346||0·3055||0·1474||0·6689|
|5/21||13||0·03 ± 0·03|| || || || |
|21/21||2||0·27 ± 0·37|| || || || |
|FcγRIIa||H/H||8||0·45 ± 0·50||0·0118||0·2208||0·7044||0·0071|
|H/R||22||0·08 ± 0·25|| || || || |
|R/R||12||0·18 ± 0·42|| || || || |
|FcγRIIIa||F/F||21||0·08 ± 0·25||0·0275||0·2688||0·0463||0·2963|
|F/V||18||0·32 ± 0·48|| || || || |
|V/V||3||0·02 ± 0·02|| || || || |
The results presented here show that breast cancer patients carrying the GM 23+ allele of IgG2 and those homozygous for the GM 5 allele of IgG3 have higher anti-HER2 IgG antibody concentration than other genotypes at these loci. A close linkage of these alleles to more efficient enhancer DNA sequences on chromosome 14, where IgG genes are located, might explain their association with high antibody responsiveness . Another mechanism underlying these associations might involve GM allotypes being part of the recognition structures for the HER2 antigen. Perhaps membrane-bound IgG molecules with the GM 23 and GM 5 allotypes are more efficient in the uptake, processing and subsequent presentation of HER2 epitopes to the collaborating T cells, resulting in strong humoral immunity. It is also possible that the associations we have observed are due to linkage disequilibrium between particular GM alleles determined in this investigation and those not determined here (e.g. GM 10, 11, 13) or alleles of another locus, as yet unidentified, for humoral immune responsiveness to HER2.
Irrespective of the mechanism(s) involved, these findings could aid in identifying subjects (GM 23+, GM 5+) who are more likely to benefit from HER2-based vaccines. For the carriers of GM 23− and GM 21 alleles, HER2 could be fused with appropriate adjuvants, such as heat shock proteins or flagellin, to circumvent the allotypic restriction in immune responsiveness. It is interesting to note that, like HER2, IgG antibody levels to certain heat shock proteins and flagellin are also influenced by GM genotypes [15, 16], making it conceivable to formulate a fusion HER2 heat shock protein/flagellin vaccine that could potentially generate high antibody responses in the majority of the population.
The results involving the white subjects presented here are in broad agreement with the earlier results , although the two studies involved different study populations (Estonian whites versus Brazilian whites) and employed different methods of GM allotyping (serological versus molecular). In both populations, GM 5 and GM 23 alleles were associated with high anti-HER2 antibody responses. We did not find a significant association between GM alleles and anti-HER2 antibody responses in black subjects with breast cancer. It is possible that, with only 42 subjects, this study was underpowered to detect an association. Also, certain alleles present in populations with African ancestry (e.g. GM 6) were not typed in this study, which could have contributed to the observed racial differences in associations between GM and anti-HER2 antibody responses. More GM determinants can be typed serologically than at the DNA level, but serological reagents are either extremely scarce or not available at all. Nucleotide substitutions responsible for most of the 18 serologically detectable GM specificities have not yet been identified.
Genotypes at both FcγR loci were associated with anti-HER2 antibody responses in black but not in white subjects with breast cancer. Breast cancer subjects with the FcγRIIa-H/H genotype, which is associated with high anti-HER2 antibody responses in this study, tended to have higher response rate to trastuzumab therapy and contributed significantly to the ADCC of breast cancer cells in a clinical efficacy study . The V allele of FcγRIIIa is considered a high-affinity allele and its homozygosity is associated with a favourable outcome of immunotherapy in many cancers. The V-carriers in the present study had higher anti-HER2 responses than F/F homozygotes (0·32 versus 0·08 μg/ml). There were too few V/V homozygotes to draw any firm conclusions. Thus, FcγR genotypes associated with humoral immunity to HER2 in the present investigation are known to contribute to anti-HER2-mediated effector functions, such as ADCC [9, 10], but the exact mechanism(s) underlying their association with endogenous anti-HER2 antibody responses is not understood.
The reasons for racial differences in associations between GM and FcγR alleles and anti-HER2 antibody responses are not clear. One contributory factor could be the divergent allele frequencies at these loci among the racial groups examined in this investigation. A potential limitation of our study is that data on HER2 expression are not available, but it is hoped that results presented here would inspire further large-scale studies that would include the measurement of all confounding variables and would be powered to detect the possible epistatic contribution of GM and FcγR alleles to humoral immunity to HER2.
This work was supported in part by a grant from the US Department of Defense (W81XWH-08-1-0373) and by a Grant-in-Aid for Research on Risk of Chemical Substances from the Ministry of Health, Labor and Welfare of Japan, and Grants-in-Aid for Scientific Research on Priority Areas (17015049). We thank Shizhong Bu, Laurel Black and Ray Deepe for assistance in GM genotyping.
The authors declare no financial conflicts of interest.