Background
One of the most impactful immune-based therapeutics for breast cancer is the humanized monoclonal antibody, trastuzumab, which specifically recognizes the HER2 protein [
1,
2]. HER2 is overexpressed in ~ 20% of invasive breast cancers, and is associated with aggressive biology and a natural history of shortened survival if not treated with targeted therapies [
3]. Trastuzumab has had such a tremendous impact on the management of patients with HER2+ advanced [
4] and early breast cancer [
2,
5‐
7], leading to the development and increased use of other HER2-targeted therapies such as pertuzumab, another IgG1 HER2-specific monoclonal antibody [
8]. The addition of pertuzumab to trastuzumab (with concurrent chemotherapy) has further improved patient survival in the metastatic setting, improved disease-free survival (DFS) in the adjuvant setting and improved pathological complete response rates in the neoadjuvant setting relative to trastuzumab and chemotherapy regimens alone [
9‐
11]. In patients with locally advanced, metastatic or inflammatory breast cancer, ~ 50% of patients receiving chemotherapy, trastuzumab and pertuzumab, did not benefit [
11,
12]. Clearly, there is a need to identify better predictors of response and long-term outcome, beyond HER2 testing. Our group has conducted several studies demonstrating a predictive immune component, such as tumor-infiltrating lymphocyte (TIL) [
13] and gene expression profiles that are associated with outcome [
14]. However, another significant area of interest is to determine whether the development of antibodies is correlated with outcome in the adjuvant setting.
Our previous work established the model that adaptive immunity is a primary mechanism of action of therapeutic efficacy when trastuzumab is added to chemotherapy regimens [
15,
16]. In those studies, in the metastatic setting, we demonstrated that development of an adaptive immune response specifically to HER2, is significantly associated with clinical response to therapy, improved progression-free survival, and overall survival. Both of these studies showed that immunity is also generated in the adjuvant setting. In our first study, four of five adjuvant patients generated anti-HER2 antibodies following treatment with chemotherapy and trastuzumab, and in our second study, ~ 50% of 26 adjuvant patients [
16] generated anti-HER2 antibodies specifically to the HER2 extracellular domain. These experiments also demonstrated that anti-HER2 antibody levels were not significantly different at early (12 weeks) and late (>20 weeks) post-treatment times [
15], demonstrating a robustness of these measures as potential biomarkers of response and some robustness in post-treatment measurement times. What these studies did not demonstrate in the adjuvant setting is whether the generation of anti-HER2 antibodies is induced by the addition of trastuzumab to the chemotherapy regimen and, specifically, whether antibody immunity is associated with improved DFS.
In the present study, we capitalized on the design and prospectively collected available biospecimen resources of the NCCTG N9831 clinical trial in early HER2+ breast cancer, which demonstrated significantly improved survival of patients treated with chemotherapy plus trastuzumab relative to patients treated with chemotherapy only [
6]. The trial was divided into three Arms (A-C). Arm A provided a control arm, of patients who received a chemotherapy regimen of doxorubicin and cyclophosphamide for four doses followed by 12 doses of weekly paclitaxel. Patients in Arm B received doxorubicin and cyclophosphamide, weekly paclitaxel, with sequential trastuzumab and patients in Arm C received doxorubicin and cyclophosphamide followed by concurrent paclitaxel and trastuzumab. Availability of serum samples pre- and post-treatment from patients in each Arm allowed us to compare the generation of patient-derived anti-HER2 antibodies specifically from chemotherapy only and from chemotherapy with the addition of sequential and concurrent trastuzumab and to determine if the generation of patient-derived anti-HER2 antibodies specifically in response to trastuzumab is associated with improved DFS.
Discussion
Our study analyzed serum samples from patients in the N9831 clinical trial of adjuvant chemotherapy and trastuzumab as an extension to our previous work in advanced HER2+ metastatic breast cancer [
15,
16] with two goals. First, to compare the generation of anti-HER2 antibodies in patients treated with chemotherapy alone to patients treated with both chemotherapy and trastuzumab. Second, to determine if the generation of anti-HER2 antibodies specifically in response to trastuzumab were associated with improved DFS.
Significantly higher post-treatment anti-HER2 humoral immune responses in patients receiving chemotherapy plus trastuzumab relative to those who received chemotherapy only is consistent with the concept that the addition of trastuzumab is the component of therapy that results in the generation of adaptive immune response, rather than to a nonspecific immune response to cytotoxic therapy. While indicating that chemotherapy alone is insufficient by itself to induce HER2 humoral immunity, our data do not rule out the possibility that the chemotherapy-mediated killing of the tumor cells contributes to other types of trastuzumab-induced anti-HER-2 immunity, such as release of neoantigens for presentation to the immune system, or by effects on immune suppressor cells. Preclinical studies demonstrate evidence for both possibilities. A study of peripheral blood monocytes isolated from 20 HER2+ breast cancer patients, demonstrated increased antibody-dependent cell-mediated cytotoxicity (ADCC) in patients who received paclitaxel plus trastuzumab compared to patients who received trastuzumab as a monotherapy [
19] and in vitro studies demonstrate that trastuzumab alone enhances class I-restricted presentation of endogenous HER-2 antigen, resulting in higher susceptibility of HER2-overexpressing tumors to lysis by the HER-2-specific cytotoxic T-lymphocytes [
20,
21].
There are no phase II or III trials with trastuzumab as monotherapy in which we can rigorously examine the contribution of chemotherapy to trastuzumab-induced tumor immunity, as the current standard of care is for both chemotherapy and trastuzumab to be used concurrently at the beginning of therapy to optimize clinical outcome. Of note, our study population did include both patients who received sequential trastuzumab and concomitant trastuzumab. We did observe a statistically significant increase in antibody levels from pre- to post-treatment in patients in Arm B (Fig.
1d) who received sequential chemotherapy and trastuzumab, which was not observed in patients in Arm C (Fig.
1e) who received concomitant chemotherapy and trastuzumab. One could postulate a preferential induction of anti-HER2 humoral immune responses in sequential rather than concomitant combinations but the number of patients in our study is too small to be able to reach conclusions related to this observation.
Perhaps more important is the association of the generation of anti-HER2 antibodies with improved DFS, a benefit which appears to be conferred over a long time. Our previous study showed augmented HER-2/neu-specific CD4 T-cell responses during therapy in six of ten patients [
15], leading us to hypothesize that, at least for some patients, the addition of trastuzumab induces helper T cells that foster the development of memory. Hence, generation of increased anti-HER2 antibodies and augmented T-helper response in a proportion of patients who received trastuzumab leads to the question; what can be done to improve trastuzumab-mediated immune response and does this translate to improved patient outcome?
Potential strategies to augment anti-tumor immune response in HER2+ breast cancer patients currently under development are the addition of other HER2-targeted therapies, use of HER-2 antibodies with enhanced Fc-effector ADCC functions and antigen presentation capacity (margetuximab) [
22], or additions of immune stimulatory treatments that might modulate FcγR-bearing cells in the tumor microenvironment such as innate activators (Toll-like receptor and stimulator of interferon gene agonists), bi-specific antibodies targeting HER2 and CD3, checkpoint immune blockade with PD-1 and PD-L1 inhibitors, or HER2-derived peptide or mRNA vaccines. Our reported correlation of increased generation of endogenous anti-HER2 antibodies are in line with others, who correlated loss of anti-HER2 Th1 response with lack of complete response to HER2-targeted therapy and reduced survival [
23], a defect which can potentially be restored by HER2 vaccination [
24]. Development of immune monitoring measurement such as the endogenous λ antibodies in this study could potentially provide informative pharmacodynamic data for immunomodulatory clinical trials including HER-2 vaccination, ADCC enhanced HER2 antibodies and checkpoint blockade.
Caveats of our study are small sample size (50 pairs of pre- and post-treatment serum samples across three study arms), and the fact that post-treatment samples were taken at varying time points, and indeed, some of the sera samples for patients in Arms B and C were drawn during trastuzumab treatment rather than after completion of treatment (Fig.
1a). However, there was no statistical difference in the distribution of post-treatment serum collection between arms. We also acknowledge that in the present study we did not have multiple post-serum samples to further substantiate our previous finding of stable development of the trastuzumab-induced adaptive immune response [
15]. However, in patients receiving trastuzumab, we did observe higher levels of antibody relative to baseline in serum samples taken up to 60 months on treatment, with no correlation between change from baseline and time of post-treatment serum collection. Regarding the assay used, our study documents a patient-derived differential antibody response is present after chemo-trastuzumab treatment for patients with early-stage HER2+ breast cancer by detection of the anti-HER2 IgG λ subclass. We measured the λ subclass to prevent inadvertent detection of trastuzumab. Our analyses assumed the proportion of IgG Κ to IgG λ in human serum is approximately 2:1 [
25,
26], but very little is known about light chain restriction specifically for an anti-HER2 response. Therefore, it is possible that deviation in the ratio of IgG Κ / λ in the anti-HER2 response could introduce a selection bias in our analysis, and it also remains possible that patients in which we did not detect an antibody response in IgG λ may have mounted a response in IgG Κ or indeed other antibodies such as IgA or IgM. Finally, our study does not explore if these anti-HER2 antibodies are a marker of patient immune competence, or if they are actually involved in tumor rejection and ultimate treatment effect. Prior to this work, given that patients with HER2+ cancer are routinely given both treatments, it has been difficult to dissociate the effect of trastuzumab and chemotherapy on tumor immunity. We show here that tumor immune responses, associated with improved clinical outcomes, are elicited in patients treated with trastuzumab and chemotherapy but not in patients treated with chemotherapy alone, and is the first study to rigorously establish that HER2 immunity is not a mere epiphenomenal consequence of cytotoxic therapy. Our data together with the similar finding in the same study population [
14] that the intra-tumoral expression of B cell and T cell genes is positively correlated with recurrence-free survival in trastuzumab-containing adjuvant arms but not in chemotherapy-alone arm [
14] establishes a likely causal relationship between the tumor microenvironment, trastuzumab treatment and the induction of protective tumor immunity. This mechanism was not previously considered in prior investigational regimens that took the view that trastuzumab targeted HER-2 signaling only and thus combined trastuzumab with other cytotoxic agents. Trastuzumab efficacy may now be more fully realized by combination approaches that instead targets the tumor immune contexture and enhances tumor cell immunogenicity.
Acknowledgements
We thank the patients who participated in this study and the healthcare personnel who helped to conduct the parent N9831 study.