Antitumor Activities by a Humanized Cancer-Specific anti-HER2 Monoclonal Antibody (humH₂Mab-250) in Xenograft Models of Human Breast Cancer

1. Introduction

Monoclonal antibody (mAb)-based therapeutics are essential for treating various diseases. The U.S. Food and Drug Administration (FDA) approved the first therapeutic mAb, Orthoclone OKT3 (mouse anti-CD3 mAb), for kidney transplantation rejection in 1986 [1]. However, the first-generation mouse mAbs tested in clinical trials had limited effectiveness due to their immunogenicity and poor effector functions [2]. Patients developed human anti-mouse antibody responses, which caused the rapid clearance of the therapeutic mAbs from the body and restricted the number of possible treatment doses [2]. The creation of engineered chimeric, humanized, and fully human mAbs has uncovered several valuable applications for antibody-based therapies [3,4].

In mAb therapy for solid tumors, the FDA approved trastuzumab for human epidermal growth factor receptor 2 (HER2)-positive breast cancer in 1998 [5]. The HER2-positive breast cancer is defined by circumferential membrane staining that is complete, intense, and in >10% of tumor cells in immunohistochemistry (IHC 3+) and/or in situ hybridization (ISH)-positive [6]. Trastuzumab is a humanized mAb by inserting the complementary determining regions (CDRs) of mouse anti-HER2 mAb (clone 4D5) into the framework of a consensus human IgG₁ [7]. Trastuzumab exhibited antitumor efficacy against HER2-positive breast cancer xenograft in monotherapy or combination therapy with chemotherapy [8,9]. The clinical efficacy of trastuzumab is mediated by the immunologic engagement [10]. Trastuzumab exerts antibody-dependent cellular cytotoxicity (ADCC) upon the binding of Fcγ receptors on natural killer cells or macrophages [10]. The combination therapy of trastuzumab with chemotherapy improves the progression-free survival and overall survival in HER2-positive breast cancer patients with metastasis [11]. Currently, HER2 overexpression and activating mutations have been observed in gastric and gastroesophageal cancers [12,13], endometrial cancer [14,15], non-small-cell lung cancer [16,17], and ovarian cancers [18].

Trastuzumab-deruxtecan (T-DXd), a trastuzumab-based antibody–drug conjugate (ADC), has been developed and received the FDA approval [19]. T-DXd has demonstrated superior efficacy not only in HER2-positive breast cancers [20,21] but also in HER2-low (IHC 1+ or IHC 2+/ISH-non-positive) advanced breast cancers [22] and HER2-mutant non-small-cell lung cancer [23]. Given that approximately half of all breast cancers are classified as HER2-low, a substantial number of patients are expected to benefit from T-DXd therapy [24]. Although T-DXd is generally well-tolerated and rarely causes severe toxicity, studies have consistently linked it to the development of cardiac toxicity. While this issue is not clinically significant in most cases, baseline cardiac evaluation, regular monitoring, and early detection of cardiac adverse events are still crucial for T-DXd. Since HER2 plays a critical role in normal heart development and homeostasis [25,26], on-target, off-tumor toxicity in the heart would cause adverse effects. Therefore, management of the specificity of mAb to tumors will be required for further optimization.

Activation of the complement-dependent cytotoxicity (CDC) pathway has been suggested as a mechanism to enhance the therapeutic effectiveness of antitumor mAbs [27]. It is one of the reported mechanisms for B-cell targeting anti-CD20 mAbs, such as ofatumumab and rituximab [28,29,30,31]. Trastuzumab mediates antitumor effects through various mechanisms but is unable to induce CDC in HER2-positive cells in the presence of human serum [32,33]. The activation of the classical complement pathway is regulated by various factors, including the size and density of the antigen, which influence the geometry of the antigen-antibody complex needed for effective C1q binding [27]. For optimal CDC activity, the Fc domains of antibodies within antigen-antibody clusters must be organized in a hexameric structure, which creates a geometry that enhances C1q binding and complement activation [27]. Approaches to improve CDC, such as antibody hexamerization [34,35] and Fc mutations [36], have shown promise in boosting antitumor activity in preclinical studies. For example, hexamerization was used to develop an anti-CD37 biparatopic antibody with enhanced in vitro CDC activity [37].

We previously generated cancer-specific anti-HER2 mAbs, H₂Mab-214/H₂CasMab-1 [38] and H₂Mab-250/H₂CasMab-2 [39], selected from 278 anti-HER2 clones, using HER2 expressed by glioblastoma LN229 cells as the target antigen. Interestingly, both H₂Mab-214 and H₂Mab-250 showed no reactivity toward spontaneously immortalized normal epithelial cells, such as HaCaT and MCF 10A [38,39]. Moreover, H₂Mab-250 exhibited no binding to normal epithelial cells derived from various tissues, including the mammary gland, kidney proximal tubule, gingiva, colon, thymus, cornea, and lung bronchus [39]. In contrast, most anti-HER2 mAbs, including trastuzumab, reacted with both cancer and normal epithelial cells [39]. Furthermore, H₂Mab-250 exhibited no reactivity with the normal heart in IHC [39]. Epitope mapping identified Trp614 in HER2 extracellular domain 4 (ECD4) as a critical determinant for H₂Mab-250 recognition [39]. H₂Mab-214 was also found to target a similar epitope as H₂Mab-250, with structural analysis suggesting that H₂Mab-214 binds to a misfolded region of the β-sheet in HER2-ECD4 [32]. This suggests that localized misfolding within the cysteine-rich portion of ECD4 contributes to the cancer specificity of H₂Mab-214. Additionally, we engineered mouse IgG_2a and mouse-human chimeric versions of H₂Mab-250. Both antibodies demonstrated antitumor activity against breast cancer xenografts in vivo, performing comparably to trastuzumab despite lower binding affinity and effector function activation in vitro [40,41].

In this study, we evaluated both the in vitro and in vivo antitumor efficacy of the humanized version of H₂Mab-250 (humH₂Mab-250) compared with trastuzumab.

2. Materials and Methods

2.1. Cell Lines

CHO-K1, BT-474, and SK-BR-3 cell lines were sourced from the American Type Culture Collection (Manassas, VA, USA). CHO-K1 and CHO/HER2 (HER2-overexpressed CHO-K1) [39] were maintained in RPMI-1640 medium [Nacalai Tesque, Inc. (Nacalai), Kyoto, Japan], while BT-474 and SK-BR-3 were cultured in DMEM (Nacalai). All media were supplemented with 10% heat-inactivated fetal bovine serum [FBS; Thermo Fisher Scientific Inc. (Thermo), Waltham, MA, USA] and antibiotic-antimycotic mixed solution (Nacalai).

2.2. Recombinant mAb Production

The CDRs of H₂Mab-250 V_H, frame sequences of V_H in human Ig, and C_H of human IgG₁ were cloned into the pCAG-Neo vector to generate a humanized anti-human HER2 mAb (humH₂Mab-250). The CDRs of H₂Mab-250 V_L, frame sequences of V_L in human Ig, and C_L of human kappa chain were cloned into the pCAG-Ble vector. We transfected the antibody expression vectors of humH₂Mab-250 into BINDS-09 (fucosyltransferase 8-knockout ExpiCHO-S) cells using the ExpiCHO-S Expression System (Thermo). Trastuzumab was produced as described previously [41]. As a control human IgG₁ mAb, humCvMab-62 was produced from CvMab-62 [42] using the abovementioned method. To confirm the purity of mAbs, they were treated with sodium dodecyl sulfate sample buffer containing 2-mercaptoethanol, separated on 5%–20% polyacrylamide gel, and stained by Bio-Safe CBB G-250 (Bio-Rad Laboratories, Inc., Berkeley, CA).

2.3. Animal Experiments

To assess the antitumor effects of humH₂Mab-250, animal experiments were authorized by the Institutional Committee for Experiments at the Institute of Microbial Chemistry (approval no. 2024-059).

2.4. Flow Cytometry

CHO-K1, CHO/HER2, BT-474, and SK-BR-3 cells were harvested using 0.25% trypsin and 1 mM ethylenediaminetetraacetic acid (EDTA; Nacalai). The cells (1 × 10⁵ cells per sample) were incubated with blocking buffer (control) (0.1% BSA in PBS), trastuzumab, or humH₂Mab-250 for 30 minutes at 4 °C. Following this, the cells were treated with fluorescein isothiocyanate (FITC)-conjugated anti-human IgG (1:2000; Sigma-Aldrich Corp., St. Louis, MO, USA) for 30 minutes at 4 °C. Fluorescence data were collected using the SA3800 Cell Analyzer (Sony Corp., Tokyo, Japan) and analyzed with FlowJo software [BD Biosciences (BD), Franklin Lakes, NJ, USA].

2.5. ADCC

The effector splenocytes were obtained from female BALB/c nude mice (Jackson Laboratory; Kanagawa, Japan). Target cells (CHO/HER2, BT-474, and SK-BR-3) were labeled with 10 µg/mL of Calcein AM (Thermo). The target cells were plated in 96-well plates at a density of 1 × 10⁴ cells/well and combined with effector splenocytes (effector-to-target ratio, 50:1) and 100 μg/mL of either control human IgG₁, trastuzumab, or humH₂Mab-250. After incubating for 4.5 hours, the calcein released into the supernatant was measured as described previously [40].

2.6. CDC

The target cells labeled with Calcein AM (CHO/HER2, BT-474, and SK-BR-3) were seeded and combined with rabbit complement (final concentration 15%, Low-Tox-M Rabbit Complement; Cedarlane Laboratories, Hornby, ON, Canada) along with 100 μg/mL of either control human IgG₁, trastuzumab, or humH₂Mab-250. After a 4.5-hour incubation at 37 °C, the amount of calcein released into the medium was measured.

2.7. Antitumor Activity of humH₂Mab-250 in Xenografts of CHO-K1, CHO/HER2, BT-474, and SK-BR-3

BALB/c nude mice were subcutaneously inoculated with 5 × 10⁶ cells suspended in BD Matrigel Matrix Growth Factor Reduced (BD). On day 7 post-injection, the mice were treated with 100 μg of either control human IgG₁ (n = 8), trastuzumab (n = 8), or humH₂Mab-250 (n = 8) via intraperitoneal injection. The treatment was repeated on days 14 and 21. Tumor size was monitored on days 7, 10, 14, 21, 23, and 27, and tumor volume was calculated as described previously [40].

2.8. Statistical Analyses

Data are presented as the mean ± standard error of the mean (SEM). Statistical analyses for ADCC, CDC, and tumor weight were performed using one-way ANOVA followed by Tukey's multiple comparisons test. Two-way ANOVA with Tukey's multiple comparisons test was applied to measure tumor volume and mouse weight. A p-value of less than 0.05 was considered statistically significant.

3. Results

3.1. Humanized anti-HER2 mAb, humH₂Mab-250

We previously established an anti-HER2 mAb (H₂Mab-250; mouse IgG₁, kappa) by immunization with the HER2 ectodomain produced by glioblastoma LN229 cells [39]. H₂Mab-250 was shown to be useful for flow cytometry [39]. In this study, we engineered a humanized H₂Mab-250 (humH₂Mab-250) by fusing the V_H and V_L CDRs of H₂Mab-250 with the C_H and C_L chains of human IgG₁, respectively (Figure 1A). We also produced trastuzumab and humCvMab-62 from CvMab-62 (mouse IgG₁) [42]. In reduced conditions, we confirmed the purity of mAbs by SDS-PAGE (Figure 1B).

As shown in Figure 2, humH₂Mab-250 and trastuzumab were detected in CHO/HER2 cells in a dose-dependent manner (Figure 2A) but not in parental CHO-K1 cells (Figure 2B). Furthermore, humH₂Mab-250 and trastuzumab reacted with HER2-positive breast cancer BT-474 (Figure 2C) and SK-BR-3 (Figure 2D). The reactivity of humH₂Mab-250 to HER2-positive cells was similar compared to that of parental mAb, H₂Mab-250 [39]. We also confirm that humCvMab-62 did not react with CHO-K1, CHO/HER2, BT-474, and SK-BR-3 at 10 µg/ml (Supplementary Figure S1). We used humCvMab-62 as a control human IgG₁.

3.2. ADCC and CDC by humH₂Mab-250 Against HER2-Positive Cells

We next examined whether humH₂Mab-250 exerted ADCC activity against CHO/HER2 cells. As shown in Figure 3A, humH₂Mab-250 and trastuzumab induced ADCC in the presence of effector splenocytes against CHO/HER2 (19.9 and 23.0% cytotoxicity, respectively) more effectively than the control human IgG₁ (5.9% cytotoxicity; p < 0.05). Furthermore, both humH₂Mab-250 and trastuzumab induced ADCC against BT-474 (8.8 and 9.9% cytotoxicity, respectively) more effectively than the control human IgG₁ (1.7% cytotoxicity; p < 0.05, Figure 3B). Both humH₂Mab-250 and trastuzumab also induced ADCC against SK-BR-3 (5.5 and 5.8% cytotoxicity, respectively) more effectively than the control human IgG₁ (1.4% cytotoxicity; p < 0.05, Figure 3C).

We investigated CDC by humH₂Mab-250 and trastuzumab against CHO/HER2. As shown in Figure 4A, humH₂Mab-250 showed a significant CDC in the presence of complements against CHO/HER2 (14.4% cytotoxicity) more effectively than the control human IgG₁ (3.5% cytotoxicity; p < 0.05). In contrast, trastuzumab did not (Figure 4A). Furthermore, both humH₂Mab-250 and trastuzumab induced CDC against BT-474 (9.7 and 7.7% cytotoxicity, respectively) more effectively than the control human IgG₁ (1.8% cytotoxicity; p < 0.05 [trastuzumab], p < 0.01 [humH₂Mab-250], Figure 4B). In contrast, both humH₂Mab-250 and trastuzumab did not induce CDC against SK-BR-3 significantly (Figure 4C).

3.3. Antitumor Effects of humH₂Mab-250 Against BT-474 and SK-BR-3 Xenografts

In the BT-474 and SK-BR-3 xenograft tumor-bearing mice, humH₂Mab-250, trastuzumab, or control human IgG₁ was intraperitoneally administered on days 7, 14, and 21. The humH₂Mab-250 treatment significantly reduced the volume of BT-474 xenografts on days 10 (p < 0.05), 14 (p < 0.01), 21 (p < 0.01), 23 (p < 0.05), and 27 (p < 0.01) compared with that induced by the control human IgG₁ (Figure 5A). The humH₂Mab-250 treatment also caused a significant reduction in SK-BR-3 xenograft on days 21 (p < 0.01), 23 (p < 0.01), and 27 (p < 0.01) compared with that induced by the control human IgG₁ (Figure 5B). Trastuzumab exhibited almost the same antitumor efficacy against BT-474 and SK-BR-3 xenografts with humH₂Mab-250 (Figure 5A and 5B, respectively). The humH₂Mab-250 and trastuzumab treatments resulted in similar decreases (57%) in BT-474 xenograft weight compared with that induced by the control human IgG₁ on day 27 (Figure 5C). The humH₂Mab-250 and trastuzumab treatments also resulted in 54% and 55% decrease in SK-BR-3 xenograft weight compared with that induced by the control human IgG₁ on day 27 (Figure 5D).

Figures 5E and 5F demonstrate the BT-474 and SK-BR-3 xenografts resected on day 27. Body weight loss was rarely observed in BT-474 and SK-BR-3 xenograft-bearing mice treated with humH₂Mab-250, trastuzumab, or control human IgG₁ (Figure 5G and 5H). Supplementary Figure S2 presents the body appearance of BT-474 and SK-BR-3 xenograft-inoculated mice treated with humH₂Mab-250, trastuzumab, or control human IgG₁ on day 27.

4. Discussion

In the development of mAbs for cancer treatment, identifying and validating suitable antigenic targets is crucial [4]. To achieve a favorable therapeutic index and minimize on-target toxicity, the ideal target antigens should be highly expressed in tumors with minimal or no presence in normal tissues. However, finding such optimal targets remains a significant challenge. Technologies like bispecific antibodies, defucosylated antibodies, and ADCs have improved antibody efficacy and advanced cancer therapy. However, the issue of on-target toxicity—caused by antigen recognition in normal cells—still persists. For instance, 97 ADCs have been evaluated in clinical trials since 2000. However, 81 trials were terminated. The reason for the termination was a lack of efficacy (32 agents) and safety issues (32 agents) [43]. On-target, off-tumor toxicity is thought to be a cause of adverse effects when the target antigen is expressed in normal cells. Therefore, selecting mAbs that specifically recognize cancer-related epitopes is critical to reducing unwanted side effects.

In this study, humH₂Mab-250 exhibited antitumor efficacy in mouse xenograft models (Figure 5). The humH₂Mab-250 demonstrated enhanced CDC activity in the presence of complement (Figure 4). Therefore, the formation of the MAC (membrane attack complex) is thought to form efficiently on the cell surface. Various factors, such as antigen size and density, influence the activation of the classical complement pathway [44]. Moreover, the geometry of the antigen–mAb complex facilitates efficient binding of C1q, which initiates the classical complement activation pathway [27]. Since IgG antibodies can form ordered hexamers upon binding to their target antigen on cell surfaces [34,45], the structure of the humH₂Mab-250-HER2 complex may allow sufficient access for complement proteins to trigger CDC. Further research is needed to understand better the mechanisms by which humH₂Mab-250 induces CDC.

CasMabs targeting HER2 (clones H₂Mab-214 [38] and H₂Mab-250 [39]) were identified through screening for reactive with cancer and non-reactive with normal cells in flow cytometry. Both CasMabs demonstrated antitumor effects in mouse xenograft models with their recombinant mouse IgG_2a or mouse-human chimeric IgG₁ mAbs [38,40,41]. The recognition mechanism of H₂Mab-214 was elucidated by X-ray crystallography, revealing that it binds to a locally misfolded structure in the extracellular domain IV of HER2, which typically forms a β-sheet [38]. Structural analysis of the H₂Mab-250 and tumor-derived HER2 complex will be critical for further understanding the mechanism of cancer-specific recognition.

H₂Mab-250 was also converted to a single chain variable fragment (scFv), developed to chimeric antigen receptor (CAR)-T cell therapy. A phase I clinical trial for patients with HER2-positive advanced solid tumors is underway in the US (NCT06241456). In CD19-positive relapsed/refractory B-cell leukemia patients who have previously been treated with CD19 CAR-T possessing mouse-derived scFv (mCD19 CAR-T), the reinfusion of mCD19 CAR-T cells may not be practical due to the development of antibodies against the anti-mouse scFv [46,47]. To address the immunogenicity, humanized CD19 CAR-T cell therapy was developed and showed a clinical benefit for the patients who had received mCD19 CAR-T therapy [48]. The scFv from humH₂Mab-250 could be another option for CAR-T therapy targeting cancer-specific HER2.

Both trastuzumab and H₂Mab-250 recognize the domain IV of HER2. Trastuzumab recognizes a wider epitope of HER2 (residues 579-625) [49]. In contrast, H₂Mab-250 recognizes a narrow and membrane-proximal epitope of HER2 (residues 613-617) [39]. Significantly, the reactivity wholly disappeared in a HER2 (W614A) mutant [39]. Furthermore, H₂Mab-250 showed a lower binding affinity (~10⁻⁹ M) than trastuzumab (~10⁻¹⁰ M) to HER2 ectodomain [41]. Several studies have shown that lower affinity CARs against CD19, glypican-3, and disialoganglioside (GD2) avoid excessive stimulation and exhaustion in the presence of low antigen burden, which leads to durable antitumor responses [50,51,52] Furthermore, a novel anti-CD19 mAb, h1218 that possesses a membrane-proximal epitope and exhibits faster on/off rates compared to clinically approved FMC63, was developed [53]. The h1218-CAR-T showed increased killing of B-cell malignancies compared to FMC63-CAR-T. Mechanistically, the h1218-CAR-T has reduced activation-induced cell death compared to FMC63-CAR-T owing to faster on/off rates [53]. These results support that the low affinity and membrane-proximal epitope possessing H₂Mab-250 CAR-T exhibited effectiveness in a preclinical study [54]. Furthermore, the formation of the MAC at the membrane-proximal region is considered essential to attack the plasma membrane of tumor cells. Among mAbs targeting CD20, ofatumumab has been shown to possess potent CDC activity compared to rituximab [28,55], which might be due to the membrane-proximal epitope and kinetics of binding to CD20 for C1q binding[29,55]. Therefore, further studies are essential to reveal the relationship between the epitope and CDC in HER2-targeting mAbs.