Dynamic changes in serum HER2-peptide-specific autoantibodies predict response to neoadjuvant therapy in HER2-positive breast cancer.

BACKGROUND: Neoadjuvant chemotherapy (NAC) with anti-HER2 agents is standard for HER2-positive breast cancer, achieving pathological complete response (pCR) in 40–50% of patients; however, reliable predictors of response remain limited. HER2 is an immunogenic antigen capable of eliciting humoral responses, yet the predictive value of HER2-specific autoantibodies in the neoadjuvant setting remains unclear. We investigated whether HER2-peptide–specific antibody responses could serve as potential biomarkers of treatment response. METHODS: Paired pre- and post-NAC sera from 112 patients enrolled in the JBCRG-16 (Neo-Lath) trial, which evaluated trastuzumab- and lapatinib-containing NAC, were analyzed. IgG titers against 63 non-overlapping 20-mer HER2-derived peptides were measured using a multiplex bead array, and serum HER2 protein levels were also quantified. Data were analyzed using Wilcoxon signed-rank and rank-sum tests, Spearman’s correlation analyses, and univariate logistic regression for pCR. Predictive models were constructed using elastic net regression with nested cross-validation across pretreatment, posttreatment, and fold-change datasets, and model performance was assessed by the area under the receiver operating characteristic curve (AUC). RESULTS: HER2-derived peptides elicited heterogeneous immune responses, and IgG titers against 37 of 63 peptides (58.7%) significantly decreased after NAC. Pretreatment IgG titers were positively correlated with the Boman index and negatively correlated with the aliphatic index. In univariate analyses, neither IgG titers nor serum HER2 levels at either time point, nor their treatment-induced changes, were significantly associated with pCR. In contrast, a predictive model based on treatment-induced changes in a subset of peptide-specific IgG titers demonstrated consistent predictive performance (AUC 0.78 in training and 0.79 in validation). Six peptides distributed across the HER2 sequence contributed to pCR prediction. CONCLUSIONS: Dynamic changes in HER2-peptide-specific autoantibodies during NAC, rather than static titers alone, may provide informative biomarkers for predicting clinical response. Serum HER2 levels were not significantly associated with pCR, suggesting that autoantibody profiling may capture response-related biology not reflected by antigen levels alone. These findings provide proof-of-concept evidence that minimally invasive, longitudinal profiling of peptide-specific autoantibodies could complement established response-assessment modalities, such as imaging or molecular assays. However, these results are exploratory and require validation in larger, independent cohorts before clinical application. TRIAL REGISTRATION: UMIN000007576 (Registration date: March 26, 2012).