Brain metastases (BM) occur in one-third of breast cancer patients during their disease course.¹ Although local therapeutic interventions, such as whole brain radiation therapy (WBRT), stereotactic irradiation (SRI), and neurosurgery, are employed to manage BMs,² patient prognosis remains poor, with a median overall survival (OS) ranging from 2 to 16 months.³ Once metastasis to the brain occurs, patients typically experience a significant decline in quality of life, culminating in organ failure and death.⁴

Traditionally, the effectiveness of drug therapy for BMs has been considered low,⁵ possibly because of the presence of the blood–brain barrier. However, in human epidermal growth factor receptor 2 (HER2)-positive (HER2+) breast cancer, new classes of drugs have been observed to challenge this notion. For instance, antibody–drug conjugates (ADCs), which comprise targeting antibodies that bind to cytotoxic payloads, can offer advantages over “naked” cytotoxic drugs. These benefits include a high therapeutic index, bystander killing through payload dispersion, and anti-tumor immune activity via antibody-dependent cytotoxicity.⁶ One notable ADC is trastuzumab deruxtecan (T-DXd), which consists of a humanized HER2-directed monoclonal antibody and DXd, a topoisomerase inhibitor. T-DXd has shown promising clinical outcomes in trials involving HER2 + metastatic breast cancer patients with BMs.^7–10 Although active/progressing BM were excluded from the three pivotal studies DESTINY-Breast01, 02 and 03, a subsequent pooled analysis of patients with baseline BM in these trials demonstrated that T-DXd is effective in patients with stable BM and in active/untreated BM. In the recent Phase IIIb DESTINY-Breast12 study, which included a cohort with 263 patient at baseline BM, intra-clinical activity of T-DXd could be confirmed across all types of BM including active/progressing BM.¹¹ In addition, although T-DXd has demonstrated clinical activity against breast cancer with HER2-low expression,¹² its effectiveness against BMs from such HER2 non-overexpressing diseases remains uncertain. Therefore, other potential target proteins expressed in BMs should be explored.

Recently, ADCs targeting cell surface proteins other than HER2, such as trophoblast surface antigen 2 (TROP2)¹³ and HER3,¹⁴ have emerged. TROP2 is a widely expressed glycoprotein belonging to the epithelial cell adhesion molecule family.¹⁵ Sacituzumab govitecan (SG) is an ADC that combines SN-38, a topoisomerase inhibitor and active metabolite of irinotecan, with an anti-TROP2 monoclonal antibody.¹⁶ SG has been approved by the Food and Drug Administration for treating advanced triple-negative breast cancer (TNBC) and hormone receptor-positive (HR+)/HER2-negative (HER2–) breast cancer. This approval was based on the results of phase III trials demonstrating the superiority of SG over physician’s choice treatments.^17,18 In these trials, SG achieved a 35% objective response rate (ORR) for extracranial lesions in patients with TNBC.¹⁷ However, the intracranial response rate was modest, at only 3% (1/32).¹⁹

HER3 is highly expressed in many solid tumors, including 30% to 50% of breast cancers.^20–22 High HER3 expression in breast cancer was reported to be associated with reduced OS.^{21, 23} Patritumab deruxtecan (HER3-DXd) is a novel, first-in-class anti-HER3 ADC currently under clinical development for multiple indications such as non-small cell lung cancer (NSCLC) and breast cancer. A phase I/II study involving the use of HER3-DXd in patients with advanced breast cancer expressing HER3 also reported promising anti-tumor activity (ORRs of 30.1%, 22.6%, and 42.9% for patients with hormone receptor (HR)+/HER2– [n = 113], TNBC [n = 53], and HER2+ [n = 14], respectively).¹⁴ Additionally, a phase II ICARUS-Breast 01 study evaluating HER3-DXd in patients with HR+/HER2– metastatic breast cancer who progressed on endocrine therapy and one line of chemotherapy also showed antitumor activity with ORR of 53.5% and a median progression free survival of 9.4 months.²⁴ In HERTHENA-Lung01, a phase II trial of HER3-DXd in advanced epidermal growth factor receptor-mutated NSCLC, the confirmed ORR was 29.8% and the intracranial ORR was 33.3% (10/30 patients; 95% CI, 17.3 to 52.8).²⁵

Despite the emerging potential of ADCs in treating BMs and their clinical applicability in targeting different proteins, evaluating target proteins for BMs is difficult in most cases owing to poor tumor accessibility. Tumors can alter their molecular profile during the development of BMs,²² making it clinically relevant to understand the correlation between the expression levels of ADC target proteins, such as HER2, HER3, and TROP2 in primary tumors (PTs) and BMs in breast cancer. However, this area remains largely unexplored. Therefore, this study was undertaken to evaluate the expression patterns of HER2, HER3, and TROP2 in paired samples of surgically resected BMs and PTs from patients with breast cancer using immunohistochemistry (IHC).

BM and PT tissues were subjected to IHC staining for estrogen receptor (ER), progesterone receptor (PgR), HER2, HER3, and TROP2 using an automated slide stainer (Ventana BenchMark ULTRA; Roche Diagnostics K.K., Basel, Switzerland). The Ventana UltraView Confirm ER (clone: SP1, Roche Diagnostics K.K.) and Ventana UltraView Confirm PGR (clone: 1E2, Roche Diagnostics K.K.) were used for ER and PgR IHC, respectively. The Ventana I-VIEW PATHWAY™ HER2 (clone: 4B5, Roche Diagnostics K.K.) and I-VIEW universal kit (Roche Diagnostics K.K.) were used for HER2 IHC. Anti-HER3 (clone: 7.3.8, Ventana Medical Systems, Inc., Oro Valley, US) and anti-TROP2 (clone: SP295, Abcam, Cambridge, UK) antibodies were used as primary antibodies for HER3 and TROP2 IHC, respectively. Heat-induced epitope retrieval for all targets was performed using ULTRA Cell Conditioning Solution #1 (Benchmark ULTRA CC1, Roche Diagnostics K.K.). The UltraView DAB IHC Detection Kit (Roche Diagnostics K.K.) was used to detect the chromogenic IHC signals of HER2, whereas the Ventana UltraView Universal DAB Detection Kit (Roche Diagnostics K.K.) was used to detect those of other targets.

IHC for ER and PgR was evaluated by a pathologist (S.F.). ER and PgR statuses were classified as positive if ≥ 1% of tumor cells expressed the proteins.²⁶ Membranous IHC staining intensity for HER2, HER3, and TROP2 in tumor cells was scored visually by pathologists (R.N. and N.M.) using a microscope. HER2 IHC staining intensity (0, 1+, 2+, and 3+) was scored according to the updated 2018 American Society of Clinical Oncology/College of American Pathologists guidelines for HER2 testing in breast cancer.²³ In this study, PTs classified as HER2 2 + were considered HER2 + if there was a confirmation of either HER2 in situ hybridization (ISH) positivity or HER2 IHC 3 + in the pathology report documented in the medical record. Two HER2 + cases in the luminal group were considered HER2– because neither of them met the above definition. Three HER2 2 + cases were considered HER2 + because of ISH positivity (luminal-HER2, n = 1; pure HER2, n = 2).

Breast cancer subtypes were classified according to the expression patterns of HR, ER, and PgR, and HER2 in PTs as follows: luminal (HR+/HER2–), TNBC (HR-/HER2–), luminal-HER2 (HR+/HER2+), and pure HER2 (HR–/HER2+). For HER3 and TROP2, IHC membranous staining intensity (0, 1+, 2+, and 3+) was scored using criteria generally applied to HER2 IHC scoring for gastric cancer:²⁴ 0, no reactivity or membranous reactivity in < 10% of tumor cells; 1+, faint/barely perceptible membranous reactivity in ≥ 10% of tumor cells; 2+, weak to moderate complete, basolateral, or lateral membranous reactivity in ≥ 10% of tumor cells; 3+, strong complete, basolateral, or lateral membranous reactivity in ≥ 10% of tumor cells.

An H-score (0–300) was also calculated based on membranous staining intensity using the following formula: H score = 3 × (percentage of strongly positive tumor cells) + 2 × (percentage of weakly to moderately positive tumor cells) + 1 × (percentage of faintly/barely perceptible positive tumor cells).²⁵ Representative expression patterns of HER2, HER3, and TROP2 are depicted in Fig. 1.

The proportions of high-level expression (IHC 2+/3+) for each protein were compared using McNemar’s test. A paired t-test was used to compare the H-scores of HER3 between PTs and BMs. OS, defined as the time from resection of BMs to death from any cause or the data cut-off date, was estimated using a log-rank test. The follow-up period was censored on December 31, 2017. OS was compared according to HER3 expression levels in BMs and changes in expression from PTs to BMs. Statistical analyses were performed using EZR software, version 1.62.

This study was approved by the Institutional Review Board (IRB) of the National Cancer Center Hospital (IRB number: 2017 − 502). The IRB waived the requirement for written informed consent from the study participants.

All BM samples exhibited a certain level of HER3 expression (≥ IHC 1+; Table 2 and Figs. 2 and 3). The proportions of high-level expression (IHC 2+/3+) were higher in BMs than in PTs for HER3 but not for HER2 (HER3: 59% vs. 91% for PT and BMs, respectively (p < 0.01); HER2: 41% vs. 43% for PT and BMs, respectively (p = 1.00); Figs. 2 and 3 and Table 2).

The tendency for higher HER3 expression in BMs than in PTs was consistent across all breast cancer subtypes except the pure HER2 subtype. The proportions of IHC 2+/3 + in BM vs. PT were 75% vs. 50%, 90% vs. 60%, and 100% vs. 50% for TNBC (HR–/HER2–), luminal (HR+/HER2–), and luminal-HER2 (HR+/HER2+) subtypes, respectively (Table 2) (Fig. 3). For the pure HER2 subtype (HR–/HER2+), all patients showed IHC 2+/3 + HER3 expression in both BMs and PTs. In the HER2 + subtypes (luminal and pure HER2 subtypes), all 16 cases showed high HER3 expression in BMs (Table 2) (Fig. 3). On an individual patient basis, 80% (35/44) had higher membranous HER3 H-scores in BMs than in PTs, and the difference between BMs and PTs was statistically significant (Supplementary Fig. 1).

For TROP2, all but one BM sample exhibited a certain level of expression (≥ IHC 1+, Table 2 and Fig. 3). TROP2 was highly expressed (IHC 2+/3+) in both PTs and BMs in the majority of cases (PTs 70% and BMs 86%, p = 0.11).

The impact of HER3 expression in BMs on patient survival was explored. No statistically significant difference in survival was observed between HER3 IHC scores in BMs (Supplementary Fig. 2). Similarly, no statistically significant difference in survival was observed between populations with decreased, equal, and increased BM HER3 expression compared with those with PT HER3 expression (Supplementary Fig. 2).

Our finding of more frequent high-level HER3 expression in BMs than in PTs is consistent with those of previous studies. A breast cancer study demonstrated HER3 positivity by IHC in 59% (22/37) of matched BMs and 30% (11/37) of PTs.²⁷ Another recent study indicated that HER3 was frequently expressed in BMs from breast cancer, with HER2 + and HER2-low BMs exhibiting significantly higher rates of HER3 co-expression than HER2-null BMs.²⁸ A study on NSCLC also demonstrated that HER3 was more abundantly expressed in BMs than in matched extracranial samples.²⁸ Considering that neuregulin 1, the ligand for HER3, is abundantly expressed in the brain and is released via various mechanisms, including hypoxia,²⁷ HER3 may play a role in the formation of BMs. In addition, HER3 overexpression has been suggested as a mechanism that confers resistance to anti-HER2 therapy. Therefore, the more frequent high-level HER3 expression in BMs than in PTs may result from clonal selection by anti-HER2 systemic treatment. Our finding that all HER2 + subtypes exhibited IHC 2+/3 + HER3 expression in BMs supports this hypothesis and is consistent with previous findings.²⁸

Although BMs have generally been considered resistant to drug treatment, ADCs, particularly those targeting HER2, have been observed to challenge this notion. Over the past decade, multiple ADCs have been developed for various types of cancer, including breast cancer, with promising clinical efficacy for BMs. Currently, the membranous expression level of the target protein is the most reliable predictor of response to ADCs.⁷ The DAISY trial demonstrated that response rates to T-DXd decreased hierarchically with HER2 expression.²⁹ Additionally, a decrease in HER2 expression level has been suggested as one of the possible resistance mechanisms to T-DXd. These findings emphasize the importance of target protein levels for the efficacy of ADCs. Our finding that virtually all BMs expressed HER3 and TROP2 suggests that these proteins could be optimal targets for ADCs in patients with BMs from breast cancer. However, the intracranial response rate of SG has been reported to be only 3%.¹⁹ It is currently unclear whether factors beyond target expression affect the intracranial activity of ADCs or if it depends more on the properties of the drugs. The TUXEDO-2 study is ongoing to evaluate the intracranial activity of datopotamab deruxutecan, another anti-TROP2 ADC, in patients with metastatic TNBC with active brain metastases (NCT05866432). Results from the first stage exploratory cohort were encouraging, three intracranial responses of seven evaluable patients, and the study has progressed to the second stage.³⁰ The DATO-BASE study (NCT06176261) is also ongoing to evaluate intracranial responses of Dato-DXd in patients with HER2– metastatic breast cancer.³¹ Notably, a clinical trial specifically evaluating the clinical activity of HER3-DXd in breast cancer BMs is ongoing.³²

This study had several limitations. First, the sample size was relatively small, which may have limited the statistical power. Second, there were significant time gaps, up to 22 years, between the collection of PTs and BMs, raising concerns about differences in antigen preservation. However, considering that HER2 and TROP2 were expressed at similar levels and frequencies in both PTs and BMs (Table 2), the observed tendency for higher HER3 expression in BMs than in PTs is unlikely to be solely due to better antigen preservation in BMs. Third, only surgically removed BMs were included in this study, suggesting that these patients were diagnosed with oligometastases or oligoprogression of breast cancer. Patients with more advanced diseases, such as multiple BMs, might have different biological characteristics, potentially introducing bias in patient selection. In addition, this may explain why we did not observe a significant difference in survival by HER3 IHC scores in BM (Supplementary Fig. 2).

In conclusion, this study demonstrated that HER3 was more frequently expressed at high levels in BMs than in PTs, and that TROP2 was highly expressed in both PTs and BMs. As multiple ADCs with various targets and payloads continue to be developed, HER3- and TROP2- ADCs may emerge as key drugs for BMs, regardless of the breast cancer subtype in the PT; however, this hypothesis warrants validation through clinical trials.