Background
Leukocyte Fcγ receptors (FcγRs) are structurally related glycoproteins belonging to the immunoglobulin (Ig) superfamily, which bind the fragment C (Fc) region of IgG leading to FcγR aggregation at the cell surface and cell activation [
1]. FcγRs are divided into 3 groups, FcγRI (CD64), FcγRII (CD32) and FcγRIII (CD16) and subclasses, depending on their structural and functional properties. FcγRI and FcγRII are expressed mainly by monocytes and macrophages, whereas FcγRIII is expressed by subsets of T cells and natural killer cells [
2].
The most important function of the FcγRs is the regulation of the immune response as, upon their engagement, they can either stimulate or inhibit immune cell activation [
2], suggesting that they can have a biological significance in the outcome of antibody-based therapies in tumors.
Among the functions mediated by activatory FcγRs, the antibody-dependent cellular cytotoxicity (ADCC), mediated by cells of the innate immunity, certainly plays an important role as it can contribute to eliminate tumor cells and, in some tumors, it may even represent the predominant effector of cell killing [
3‐
6].
In the immunotherapy of HER-2 overexpressing BC, trastuzumab monoclonal antibody (mAb) may elicit ADCC upon binding to the HER-2 antigen region on tumor cells and with the Fc region to FcγRs on immune effector cells [
7,
8]. This mechanism is similar to that of other mAbs of the IgG1 isotype, such as rituximab and cetuximab, upon binding to CD20 and EGFR, respectively [
9‐
12]. Such an interaction triggers the immune cell degranulation and activation, resulting in the lysis of antibody coated target cells.
However, only 20–30 % of HER-2-overexpressing patients respond to trastuzumab and a significant proportion of them become resistant to the mAb [
13].
The extent of in vitro ADCC of BC cells mediated by trastuzumab may be influenced by several factors including single nucleotide polymorphisms (
SNPs) in the FcγR genes. It has been reported that these
SNPs can alter the FcγR binding to the therapeutic mAbs and consequently the ADCC degree. In particular, the
SNP rs396991 (G>T) corresponding to the substitution of valine (V) with phenylalanine (F) at aminoacid position 158 of FcγRIIIA (158V>F variant) and the
SNP rs1801274 (A>G) corresponding to the substitution of histidine (H) with arginine (R) at aminoacid position 131 of FcγRIIA (131H>R variant), appear to reduce the binding to the mAbs [
14‐
16].
However, the association between FcγR polymorphisms and trastuzumab efficacy in BC is controversial. Indeed, the homozygous FcγRIIIA158V/V and FcγRIIA 131H/H phenotypes (commonly identified as 158V/V and 131H/H genotypes) have been associated with ADCC, response to trastuzumab and progression-free survival in two small retrospective studies [
7,
17], whereas a larger study did not support these findings [
18].
In the present study, we have investigated the FcγRIIIA158V>F and FcγRIIA131H>R genotype frequencies in patients with BC overexpressing HER-2 and their role in the extent of in vitro trastuzumab-dependent lysis of HER2-positive BC cells. We demonstrate that PBMCs from BC patients carrying the FcγRIIIA158F genotype can induce, in some circumstances, a more efficient ADCC response than PBMCs carrying the homozygous FcγRIIIA 158V/V genotype. We also demonstrate that the ADCC associated to particular FcγRIIIA and FcγRIIA genotypes can be influenced by the HER-2 expression levels on target cells. In this context MCF-7, a BC cell line showing the lowest HER-2 expression level, allowed us to point out a correlation between genotypes and ADCC, as well as between ADCC and patient response to trastuzumab.
Methods
Patients
Women with histological diagnosis of locally advanced invasive or metastatic BC were considered eligible for the study if classified as HER-2 positive, i.e. score 3+ (by immuno-histochemical analysis: IHC) or IHC score 2+ and FISH (fluorescence in situ hybridization) amplified. Twenty-five BC patients were enrolled in the study: 15 patients in the neo-adjuvant setting (NEO) and 10 patients in the metastatic setting (MTS).
In the NEO setting, all patients (with the exclusion of 1 treated only with paclitaxel) were treated with FEC (fluorouracil, epirubicin and cyclophosphamide) for 4 cycles followed by weekly paclitaxel for 12 weeks in combination with trastuzumab. In the MTS setting, patients underwent a first line chemotherapy in combination with trastuzumab.
Response to trastuzumab was evaluated on the basis of clinical, pathological and radiologic examination of the tumor before and after treatment. In details, for the NEO patients, pathological complete response (pCR) was used to evaluate the treatment response. pCR was assigned in absence of invasive residual carcinoma in the breast and/or at axillary lymph node level after surgery. In the presence of residual invasive carcinoma the response was considered partial (pPR). For the MTS patients, the revised RECIST criteria (version 1.1) were used to evaluate the treatment response which was classified as stable disease (SD), partial response (PR), complete response (CR) and disease progression (PD).
This study was approved by the Ethics Committee of IRCCS AOU San Martino-IST, Genoa, Italy and written informed consent was obtained from each patient.
Thirty-three unrelated healthy Italian women (Transfusion Service, Galliera Hospital, and IRCCS AOU San Martino-IST, Genoa, Italy), matched for patient’s age, were also included as a control population, upon written informed consent. In addition, a cohort of 64 unselected healthy donors were analyzed for FcγRIIIA and FcγRIIA genotyping and ADCC.
Genotyping of FcγRIIIA 158V>F and FcγRIIA 131H>R polymorphisms
Genotyping of FcγRIIIA 158V>F (rs396991) and FcγRIIA 131H>R (rs1801274) variants was performed on genomic DNA by multiplex Tetra-Primer Amplification Refractory Mutation System (T-ARMS) PCR, direct sequencing (SBT) and pyrosequencing (PSQ). Genomic DNA was extracted from peripheral blood mononuclear cells (PBMCs) using a conventional proteinase K protocol, as previously described [
19].
The FcγRIIIA was analyzed by PSQ and due to the high homology of FcγRIIIA with FcγRIIIB in the genomic region of interest, DNA samples were first amplified by primer pair specifically designed for it. To this aim, the PCR reaction was performed in a final volume of 20 μl containing ≈25 ng genomic DNA, 1.5 mM Mg
2+ (5x buffer A and buffer B), 200 μmoli/L dNTPs, 0.8 μL Elongase
® Enzyme Mix (Life Technologies) and 0.5 μmol/L of each primers (Table
1).
Table 1
Primers and conditions for the analysis of FcγRIIIA gene polymorphism
FcγRIIIA-F | AAGTCATTTGGGGTCAATTTC | 2416d
| |
FcγRIIIA-R | CAGAATAGTTTCATCTCGTATATCb
| | |
FcγRIIIA-R-S | CAGGAATAAGGTGACGGTGG | 598 | |
Py-FcγRIIIA-F |
Bio-AAAGCCACACTCAAAGACAGCc
| 122e
|
MAAGCCCCCTGCAGAAGTAGf
|
Py-FcγRIIIA-R | ATTCCAGGGTGGCACATGTC | | |
Py-FcγRIIIA-S | ACACATTTTTACTCCCAA | | |
PCR products were analyzed on 1 % agarose gel containing Eurosafe Nucleic Acid Stain (EuroClone, Milan, Italy) and the remaining amount was purified and diluted 1:20; then, to confirm specificity of FcγRIIIA amplification, a short amplicon was analyzed by direct sequencing (using the FcγRIIIA-F and FcγRIIIA-R-S primers (Table
1).
For the PSQ assay, a nested-PCR was setup by amplifying 1 μl of diluted PCR products with the primer pair designed using the Pyrosequencing Assay Design software (Biotage, Uppsala, Sweden) (Table
1).
The PSQ-PCR reactions were assembled with the EpMotion5070 liquid handling station (Eppendorf, Milan, Italy) in a final volume of 50 μl containing 200 µmol/l dNTPs, 1× GeneAmp buffer (1.5 mM MgCl2), 1.25U of Immolase Hot Start polymerase (Bioline, Milan, Italy) and 0.3 μM of the PCR primer pairs specific for the FcγRIIIA 158V>F SNP. The Pyro Gold reagent kit PSQ 96MA was used for the sequencing reactions, according to the manufacturer instructions. The PSQ assay was performed using a PSQ96MA instrument (Qiagen, Milan, Italy) and the sequencing analysis was conducted with the PSQTM 96MA (version 2.02) software.
The FcγRIIA 131H>R
SNP was analyzed by T-ARMS PCR [
20] and by PCR amplification followed by SBT [
7] in both forward and reverse directions using previously described methods.
Both FcγRIIIA 158V>F and FcγRIIA 131H>R polymorphisms were independently re-tested in a different laboratory by T-ARMS PCR or SBT, as previously described, and results where fully concordant.
Cell lines and culture conditions
The BC cell lines SKBR3, BT474 and MCF-7 (American Type Culture Collection; ATCC, Rockville, MD, USA) were maintained in monolayer culture in RPMI 1640 medium supplemented with 10 % heat-inactivated FBS, 5 mg/ml penicillin and 5 mg/ml streptomycin, 2 mM l-glutamine, at 37 °C in a humidified 5 % CO2 atmosphere and subcultured every 3–7 days. The leukaemic cell line K562 (ATCC) was grown in complete RPMI 1640 culture medium. All reagents were purchased from Biochrom KG (Berlin, Germany).
Analysis of trastuzumab reactivity by flow cytometry
For HER-2 cell surface staining with trastuzumab antibody (Herceptin; Roche, Basel, Switzerland), titration assays were performed by indirect immunofluorescence using as targets SKBR3, BT474 and MCF-7 cell lines. In particular, 2 × 105 cells/sample were incubated with different concentrations (2, 0.2, 0.02, 2 × 10−3 and 2 × 10−4 µg/ml) of the mAb for 30 min at 4 °C. After washing in FACS buffer, an Alexafluor 647-conjugated goat anti-human IgG secondary antibody (Molecular Probes, Inc. Eugene, OR, USA) was added and incubated for 30 min at 4 °C. After 2 washes, the cells were analyzed by flow cytometry on a CyAn-ADP-Flow-Cytometer (Beckman-Coulter); 5000 cellular events were analyzed per sample. Data were analyzed using Summit version 4.3 Software (Beckman-Coulter).
Results were expressed as mean ratio of relative fluorescence intensity (MRFI), calculated as follows: mean fluorescence intensity (MFI) of HER-2 staining/MFI of secondary antibody staining.
PBMC separation and antibody-dependent cellular cytotoxicity (ADCC) assay
PBMCs were obtained after Ficoll-Hypaque density centrifugation of blood samples derived from BC patients before starting trastuzumab treatment. In particular, the NEO PBMCs included 4 V/V carriers, 9 V/F or F/F carriers for the FcγRIIIA SNP and 3 H/H carriers, 10 H/R or R/R carriers for the FcγRIIA SNP. The MTS PBMCs included 3 V/V carriers, 6 V/F or F/F carriers for the FcγRIIIA SNP and 4 H/H carriers, 5 H/R or R/R carriers for the FcγRIIA SNP.
Ex-vivo isolated PBMCs were used as effector cells in the ADCC assay in the presence of trastuzumab. As in preliminary experiments we found that ADCC triggering could be effective in a wide range of trastuzumab concentrations (from 2 to 2 × 10
−4 µg/ml), we chose to perform all the cytolytic assays with trastuzumab at 2 µg/ml (concentration at which the differences in trastuzumab reactivity among target cell lines were more evident by flow cytometry). SKBR3, BT474 and MCF-7 target cells were labelled with sodium
51Chromate (
51Cr) and seeded in 96-V-bottomed microplates with effector cells (E)/target (T) at different E:T ratios in 200 µl of volume. Triplicate wells were set up for each E:T ratio and the percentage of lysis was calculated as previously described [
21]. In details, the percent of specific cell lysis was calculated as [(experimental − spontaneous release)/(maximum − spontaneous release)] × 100, in which spontaneous release refers to the
51Cr release of target cells incubated for 4 h without effector cells. The maximum release is the
51Cr release of each target cell in the presence of 1 N HCl. The maximum lysis was represented by the maximum − spontaneous release.
K562 cells were also used as HER-2 negative target cells. Basal cell lysis and ADCC were scored as present (≥5 % of lysis) or absent (<5 % of lysis). We estimated variability among the PBMC lytic activities by the coefficient of variation (CV %), calculated as the standard deviation of cell lysis values among all the PBMC samples tested divided by the mean and multiplied by 100.
Statistical analysis
The comparison of genotype and allele frequencies between groups, as well as the association between genotype frequency, clinical-pathological parameters and response to trastuzumab were performed by the Pearson Chi square test or the Fisher’s exact test, as appropriate. For the correlation test between response to trastuzumab and cytotoxicity (both basal and trastuzumab-mediated), the Bonferroni correction for three comparisons (indicated as P
c
), that would require a P value <0.017 to reject the null hypothesis, was applied. The comparison between basal cell lysis and trastuzumab-mediated lysis was performed using a paired Student’s t test.
All statistical tests were two-sided and they were carried out using the SPSS package (version 19.0 for Windows). Statistical significance was accepted for any P value <0.05.
Discussion
Trastuzumab is a human mAb that specifically binds the extracellular domain of HER-2 receptor expressed by BC cells. Trastuzumab is currently used as a single agent [
26] and in combination with chemotherapy [
27‐
29] for the treatment of women with either early [
30] or advanced HER-2 overexpressing BC [
27,
28]. The principal mechanisms of action of trastuzumab include direct growth inhibition leading to apoptosis of tumor cells, complement-dependent cytotoxicity and ADCC by NK cells, monocytes and granulocytes [
31].
It has been previously shown that functional FcγR
SNPs can regulate the trastuzumab-mediated ADCC and predict the clinical outcome of BC patients treated with trastuzumab-based therapy [
7,
17,
32].
In this study, we investigated in vitro the ADCC elicited by PBMCs from NEO and MTS BC patients and correlated its extent with FcγRIIIA 158V>F and FcγRIIA 131H>R genotypes and response to trastuzumab.
We found that the FcγRIIIA 158F and/or the FcγRIIA 131R genotypes, that are commonly reported as unfavourable genotypes not only in BC [
7,
32] but also in other tumours [
10,
11], may actually behave as favourable genotypes. Indeed, the 158F and 131R genotypes were able to elicit a stronger ADCC, as compared to the 158V/V and 131H/H genotypes, which remained significant also at low E:T ratios. In addition, in some instances, only the 158F and 131R genotypes showed significant ADCC. The MCF-7 cell line has previously been found weakly- or not-expressing HER-2 [
24,
25,
33,
34] so that it has been used as a negative control for trastuzumab-mediated ADCC [
34]. In our experimental system, MCF-7 displayed low, but detectable, reactivity with trastuzumab by flow cytometry, as compared to SKBR3 and BT474 cell lines that showed high reactivity. Therefore, our ADCC results with MCF-7 cell line may have implications in trastuzumab-based therapy because they suggest that also tumor cells which are not HER-2-amplified and display low levels of HER-2 expression can be efficiently targeted with trastuzumab.
In the NEO patients, the FcγRIIIA 158V/V genotype correlated with a statistically significant ADCC in all cell lines tested, at high E:T ratios, whereas this correlation was not observed in the MTS patients at any E:T ratio. The FcγRIIA131H/H genotype correlated with a statistically significant ADCC at high E:T ratios with two or three cell lines depending on the setting, NEO or MTS, analyzed.
Although different effector cell types within PBMCs can engage trastuzumab, we here chose to evaluate the overall lytic activity of PBMCs isolated before starting trastuzumab therapy in order to mimic the ADCC condition that may occur in vivo. For this reason, we evaluated the FcγRIIIA and FcγRIIA gene polymorphisms that are respectively involved in the activation of the NK cell population and other accessory cells of the immune system. We found a broad interindividual variability of ADCC levels among patients which might be related to the FcγR genotypes, as well as to the proportion and/or lytic efficiency of effector cells present in the PBMCs samples.
The ADCC variability was indicated by the high CV values that we observed among all the PBMC samples tested (CVs ranging from 72 to 103 %, depending on the BC setting and target cell line). Remarkably, we noticed that the variability was more pronounced in the spontaneous (basal) cytotoxicity (CVs ranging from 108 to 239 %), as compared to ADCC, condition that was also observed among healthy individuals (Additional file
4 ). To explain the lower CV values in trastuzumab-mediated cell lysis, it is conceivable that the addition of trastuzumab to the cytolytic assay provides an optimal stimulus to trigger the activation of cytolytic effector cells. Indeed, the FcγRIIIA engagement with trastuzumab resulted in a homogeneous triggering of FcγRIIIA + PBMCs abrogating the individual specific feature of spontaneous lysis of each donor and pointing out their maximal cytolysis. Indeed, only using trastuzumab we could find a biologically significant degree of target cell lysis.
Altogether, the MTS patients showed higher levels of ADCC against all BC cell lines used as targets, as compared to the NEO patients, thus confirming that ADCC can be efficient also in the advanced stage of BC, as reported in previous studies [
7,
35]. A possible explanation for the higher ADCC in MTS patients may be the different immunologic context under which immune cells become activated in the two patient settings. In particular, the immune response elicited by BC in the NEO patients might be confined to the tumour microenvironment whereas that elicited in the MTS patients might be systemic resulting from a generalized and more potent immune cell stimulation due to the spread of BC to other parts of the body. Indeed, we found that the ADCC observed in MTS patients was similar, in intensity and variability, to that observed in healthy donors suggesting that MTS patients are in a different activation state as compared to NEO patients. Thus, the different efficacy of MTS and NEO cytolytic activity may be due to intrinsic biological characteristics of the immune cells (activation state, lytic efficiency, different cell number) rather than to the effects of clinical-pathological parameters. In both BC subgroups, the entity of ADCC reflected the HER-2 expression levels as the highest absolute values were observed with the SKBR3 cell line.
In NEO and MTS subgroups the enhancement of basal lysis induced by the FcγRIIIA158F carriers was dependent on HER-2 expression, whereas the one induced by the 158V/V carriers was dependent on HER-2 expression only in the MTS subgroup. Concerning the FcγRIIA, the enhancement of basal lysis induced by the H/H carriers was independent from HER-2 expression in both MTS and NEO subgroups and the one induced by the 131R carrier was dependent on HER-2 expression in NEO, but not in MTS patients.
In summary, these data demonstrate a different behaviour, in respect to the FcγR genotypes and ADCC extent, of BC cell lines expressing different HER-2 levels. These levels might be representative of the tumour and intratumoral heterogeneity of HER-2 expression that has recently been reported in a proportion of breast cancers (
36). This heterogeneity is clinically relevant for its association with the response to HER-2-targeted therapy.
To date, there are still conflicting reports in the literature concerning the role of FcγR polymorphisms in the clinical outcome of trastuzumab-treated BC patients. Even the largest retrospective analysis performed by Hurwitz et al. in the adjuvant and metastatic settings does not seem to completely exclude a contribution of FcγR polymorphisms and ADCC/FcγR engagement to the outcome of trastuzumab-treated patients [
18]. Thus, the conflicting results may derive from different factors including ethnical differences in the
SNP frequency, genotyping problems due to the high homology between FcγR members, differences in the modalities of therapeutic use of trastuzumab (i.e. NEO, MTS or adjuvant setting) or in the chemotherapy regimen.
In addition, only few reports analyzed the association of FcγR polymorphisms with ADCC intensity and with clinical outcome [
7,
8]; other reports analyzed only the association of FcγR polymorphisms with clinical response without taking into account the ADCC [
17,
18,
32], or the association between ADCC and response to trastuzumab without considering the FcγR gene polymorphisms [
34,
35]. Although performed in a quite small number of patients, our study correlates FcγR gene polymorphisms to the ADCC extent in combination with the HER-2 expression levels on tumor target cells. Similarly, this study also correlates FcγR polymorphisms to the basal lysis extent. Indeed, in both NEO and MTS patients the 158 V/V carriers showed, in some instances, a more pronounced lysis as compared to that of the 158F carriers, although in MTS patients with BT474 cell line we found that 158F carrier showed a more efficient basal lysis than 158 V/V. However, this difference was not statistically significant and may depend on intrinsic features of each patient.
Our results demonstrate a statistically significant correlation of the ADCC degree with the complete response of NEO patients to trastuzumab. In particular, the frequency of patients with positive ADCC was significantly increased using the MCF-7 cell line as target, as compared to patients without ADCC; this suggests that also small variations in the percent of cell lysis observed in vitro might contribute to the induction of a complete response in vivo.
It is of note that the increase of patients showing cytotoxicity was observed with the MCF-7 cell line suggesting that the use of low-HER-2-expressing cells might be useful for testing of ADCC.
Even if further investigations in larger cohort studies are required to confirm our findings, these results provide additional information for a better understanding of the role of FcγR gene polymorphisms/ADCC in trastuzumab-treated BC; this study might improve the testing of patient ADCC in order to identify subgroups of patients who might benefit from trastuzumab treatment.