Antibody–drug conjugates in breast cancer: mechanisms of resistance and future therapeutic perspectives

Abstract

Although antibody–drug conjugates (ADCs) revolutionized cancer treatment, especially in breast cancer, resistance remains a major challenge. It may involve target expression and distribution, linker stability, ADC intratumor penetration and intracellular internalization, payload activity but also interaction with the tumor microenvironment. Strategies to overcome resistance under investigation include combination therapies, novel payloads, and next-generation ADC designs. The identification of predictive biomarkers may also help guide treatment selection and prevent resistance.

Introduction

Antibody drug conjugates (ADCs) have revolutionized tumor treatment, but resistance often limits long-term efficacy¹. Available data on the full spectrum of resistance mechanisms remain limited, partly due ADCs structural and mechanistic complexity, which enables a wide range of potential resistance pathways. Furthermore, multiple resistance mechanisms can arise simultaneously and interact synergistically. Despite growing research efforts, effective strategies to overcome it are still under investigation.

ADC structure and main mechanisms of action

ADCs present three key components: a humanized monoclonal antibody (mAb) targeting a tumor antigen, a cytotoxic payload, and a linker connecting them². Their key mechanism of action begins with antigen binding, followed by internalization and payload release intracellularly¹. Beyond direct cytotoxicity on the antigen-expressing cells, ADCs could trigger immune responses and exert a “bystander effect”, through payload diffusion to the neighboring antigen-negative cells, particularly relevant in tumors with heterogeneous antigen expression³. Following intravenous administration, ADCs must remain stable in circulation. IgG1 antibodies are commonly used for their long half-life (~21 days), manufacturing ease, and immune effector functions⁴. The antigen-binding fragments (Fabs) region ensures antigen recognition, while the constant region (Fc) engage immune responses such as complement activation, cytotoxicity, and antibody-dependent cell-mediated cytotoxicity (ADCC)⁵. Fc modifications can modulate half-life, though may compromise antibody-dependent functions^1,6,7. Linkers are cleavable or non-cleavable: the former respond to intracellular cues (e.g., proteases, acidic pH) but may risk premature payload release; the latter require full internalization and lysosomal degradation, enhancing stability^7,8.

Payloads currently fall into two main categories: microtubule inhibitors and DNA-damaging agents like topoisomerase I inhibitors (TOP1i)⁵. The drug-to-antibody ratio (DAR) reflects the number of cytotoxic molecules per antibody, influencing pharmacokinetics, safety, and activity profile⁵.

Several ADCs are approved for breast cancer (BC) across different countries: trastuzumab emtansine (T-DM1), trastuzumab deruxtecan (T-DXd), sacituzumab govitecan (SG), datopotamab deruxtecan (Dato-DXd), and sacituzumab tirumotecan (Sac-TMT)^9,10. T-DM1 and T-DXd share the human epidermal growth factor 2 (HER2)-targeting antibody but differ in payloads, linkers, and DARs. Dato-DXd, SG, and Sac-TMT use TOP1i and target trophoblast cell surface antigen 2 (TROP2), but differ in all components^8,10,11.

Possible mechanisms of resistance to ADCs

ADC stability and penetration

ADC stability in circulation is critical for efficacy and safety, depending on both antibody choice and stable linker to prevent premature payload release, which can cause off-target toxicity and reduce tumor delivery, compromising selectivity and therapeutic benefit^8,12. Non-cleavable linkers offer greater plasma stability and therapeutic index, though newer cleavable ones also show enhanced stability¹².

Intratumoral distribution is also a critical factor in ADC efficacy¹³ (Fig. 1a). In solid tumors, poor penetration due to abnormal vasculature, hypoxia, elevated interstitial pressure, and large ADC size is an early resistance mechanism^14–16. Additional factors like high antigen expression, rapid internalization and binding site barrier (BSB), caused by antibody binding near vasculature, also impair ADC delivery^1,17.

Fig. 1. Mechanisms of resistance to ADCs.

a Poor intratumor penetration: physical barriers, blood site barrier (BSB), target-mediated drug disposition (TMDD). b Target-related resistance: reduced antigen expression and heterogeneity in antigen expression. c Altered Internalization and Trafficking: heterodimerization with human epidermal growth factor 3 (HER3) or epidermal growth factor receptor (EGFR), and lysosomal alkalinization. d Payload Resistance: overexpression of drug-efflux pumps, release via extracellular vesicles (EVs), drug target mutations. e Bypass mechanisms related to payloads: activation of alternative survival pathways involving Signal Transducer and Activator of Transcription 3 (STAT 3), Leukemia Inhibitory Factor Receptor (LIFR), Heat Shock Protein Beta-1 (HSPB1), circular RNA (crRNA), Polo-Like Kinase 1(PLK1), Cyclin B1. f Tumor microenvironment (TME): increased cancer-associated fibroblasts (CAFs), CD8 + T cells and Natural Killer (NK) cells, and enrichment of cancer stem cells (CSCs).

Some preliminary clinical evidence also pointed out the relevance of the ADC intratumoral distribution¹⁸.

In the ICARUS-BREAST01 trial with patritumab deruxtecan (HER3-DXd), exploratory analysis from on-treatment tumor biopsies suggested that the intratumoral distribution of the ADC may influence response, highlighting the need to optimize ADC delivery¹⁹.

Reduction of target expression and localization

Target expression and localization play a key role in enabling adequate ADC tumor distribution through binding to tumor cells^5,20 and therefore in determining ADC resistance (Fig. 1b).

Regarding target expression, although T-DM1 is ineffective in HER2-negative or HER2-low BC, newer-generation ADCs like T-DXd have shown activity in tumors with low HER2 expression^21,22, likely due to the bystander effect observed with T-DXd but not with T-DM1²³.

However, the benefit of T-DXd remains associated with HER2 expression levels, as shown in the phase 2 DAISY trial, evaluating this ADC in mBC, which reported objective response rates (ORRs) of 70.6% (95% Confidence Interval [CI], 58.3–81.0) in HER2-overexpressing, 37.5% (95% CI, 26.4–49.7) in HER2-low, and 29.7% (95% CI, 15.9–47.0) in HER2-negative disease, indicating a direct correlation between HER2 expression and efficacy²⁴.

Exploratory biomarker analyses from the ASCENT and TROPiCS-02 phase 3 trials suggested a trend toward improved ORR and median progression-free survival (mPFS) with higher TROP2 expression. In ASCENT (pretreated metastatic triple-negative [mTN]BC), SG showed longer PFS (6.9, 5.6, and 2.7 months, 95% CI) and higher ORR (44%, 38%, and 22%) in patients with high, medium, and low TROP2 expression, respectively, though benefit was observed across all subgroups²⁵.

Similarly, in the TROPiCS-02 trial (pretreated, endocrine-resistant [Hormone receptors] HR-positive/HER2-negative mBC), a post hoc analysis showed improved outcomes with SG versus control regardless of TROP2 expression (histochemical score [H-score]; range, 0–300), with a trend toward longer mPFS in patients with H-score ≥100 vs <100 (6.4 vs 5.3, 95% CI)²⁶.

ADCs can also induce resistance by downregulating antigen expression and promoting tumor cell tolerance^27,28. In vitro models of Dato-DXd resistance revealed marked TROP2 downregulation²⁹. Also in clinical settings, TROP2 mutations, like T256R, were associated with impaired membrane localization and drug binding, and ultimately with acquired resistance to SG³⁰. Similarly, the phase 1 U31402-A-U102 trial, evaluating HER3-DXd in non-small cellular lung cancer (NSCLC), identified acquired mutations in the gene codifying for human epidermal growth factor 3 (HER3); leading to reduced HER3 expression and decreased drug efficacy³¹. In the DAISY trial, 65% of patients (95% CI, 40.8–84.6) showed decreased HER2 expression upon resistance²⁴.

Furthermore, the spatial distribution of the target within the tumor also plays a critical role in shaping the response to therapy, as it can affect the interactions of the ADC with the tumor cells and therefore the drug accessibility. In the DAISY study, non-responder tumors in the HER2-positive cohort exhibited a higher proportion of cluster 6, characterized by low HER2 staining and prevalence of stromal component²⁴.

Similarly, HER2 heterogeneity was linked to resistance to T-DM1 plus pertuzumab and lower pathologic complete response (pCR) rates (55% vs 0%, P < 0.0001)³². Likewise, the KRISTINE trial showed lower pCR in tumors with focal or variable HER2-overexpression, while high HER2 clustering predicted better trastuzumab response^13,33.

Altered internalization and intracellular trafficking pathways

Once bound to their target, ADCs promote antigen clustering and internalization, mainly via clathrin-mediated endocytosis, though alternative pathways exist^15,22,34,35. Subsequently, the ADC–antigen complex is trafficked through lysosomes, where acidic pH and proteases release the payload^22,34,36 (Fig. 1c).

Resistance to T-DM1 may emerge through multiple interconnected internalization-related mechanisms in preclinical data: altered lysosomal trafficking, increased lysosomal pH, and sequestration in Caveolin-1-positive compartments with a neutral pH^37–39. Loss of Solute Carrier Family 46 Member A3 (SLC46A3, a lysosomal transporter protein), further impairs payload cytoplasmic release⁴⁰.

Preclinical studies described that epidermal growth factor receptor (EGFR) can form heterodimers with HER2, altering trafficking and reducing T-DXd internalization, while HER3/HER2 dimerization seemed to affect trastuzumab efficacy³⁵. In addition, in the EMILIA trial high EGFR expression correlated with poorer outcomes with T-DM1⁴¹.

Heregulin-induced HER2/HER3/HER4 dimerization also showed to reduce T-DM1 cytotoxicity in both preclinical and early-phase clinical studies^42,43.

Target internalization capability can also exert a key role; prior preclinical data pointed out that TROP2 cellular localization and endocytosis may vary across different tumor types^44,45.

This is in line with the recent findings of an exploratory analysis of TROPION-Lung01 trial, showing that patients with TROP2 QCS-positive tumors (reflecting higher TROP2 internalization) had improved outcomes with Dato-DXd versus docetaxel (mPFS: 6.9 vs 4.1 months; Hazard Ratio [HR] 0.57, 95% CI, 0.41–0.79)⁴⁶.

Payload-associated resistance

Several resistance mechanisms to ADC payloads have been described, involving both reduced intracellular drug concentration and impaired payload activity.

Efflux transporter upregulation

Many payload-related mechanisms of resistance mirror classic chemotherapy resistance, such as the upregulation of ATP-binding cassette (ABC) transporters^47,48 (Fig. 1d). Preclinical studies have linked Multi-Drug Resistance 1 (MDR1)-mediated efflux to T-DM1 resistance^28,49.

One potential strategy to overcome this type of resistance involves the use of ADCs less susceptible to efflux pumps: in HER2-positive resistant mouse models, disitamab vedotin (DV) outperformed T-DM1 and T-DXd, suggesting a role for efflux pumps, although mechanisms remain unclear⁵⁰. This resistance mechanism was also observed clinically: ABCC1 expression in post-T-DXd samples correlated with poorer T-DXd-associated overall survival (OS)⁵¹.

Payload target mutations

Mutations in genes involved in payload activity may also determine ADC resistance (Fig. 1d). Retinoblastoma 1 (RB1) mutations have been associated with T-DXd resistance in TP53-/HER2-mutated NSCLC⁵².

With respect to BC data, in the DAISY trial, SLX4 mutations, a gene encoding for a DNA repair protein that might have a role in resistance to TOP1i, were more frequent at progression to T-DXd than at baseline, and SLX4 knockdown in cell lines induced resistance²⁴. TOP1 mutations (e.g., E418K and frameshift) were detected in tissue samples after SG progression³⁰. Abelman et al. reported TOP1 mutations (E418K, R364H, W401C) in circulating tumor DNA (ctDNA) of patients progressing on a TOP1i-based ADC, correlating with cross-resistance and shorter duration on a second TOP1i-ADC (52 vs 455 days)⁵³.

Extracellular vesicles (EVs) mediated resistance

Extracellular vesicles (EVs) also influence ADC response in preclinical models (Fig. 1d). HER2-positive EVs can enhance the bystander effect by delivering payloads to neighboring cells, but may also promote resistance by acting as drug decoys^54,55.

Bypass mechanisms related to payloads

A possible alternative mechanism of resistance to the payload is the activation of compensatory survival pathways (Fig. 1e). In particular, resistance to T-DM1 has been associated in preclinical studies with dysregulation of cell cycle and apoptotic signaling: loss of cyclin B1 and downregulation of Polo-Like Kinase 1 (PLK1) impair mitotic arrest, reducing T-DM1 efficacy. Notably, restoring their activity partially re-sensitizes resistant cells^56,57. Moreover, overexpression of the Leukemia inhibitory factor receptor (LIFR) activates signal transducer and activator of transcription 3 (STAT3) signaling and increases anti-apoptotic proteins, further promoting resistance to T-DM1⁵⁸. T-DM1-resistant cell lines have also shown alterations in adhesion and prostaglandin-related genes, potentially affecting microtubule binding and supporting payload-specific resistance mechanisms⁵⁹.

Regarding data on T-DXd, studies in resistant models have shown that a VDAC3-derived circular RNA (crRNA) inhibits Heat Shock Protein (HSP) Beta-1 ubiquitination, thereby contributing to resistance⁶⁰.

ADC cross-resistance

Finally, also real-world data indirectly showed resistance pattern linked to the payload, by reporting limited PFS benefit when ADCs with the same payload class (e.g., TOP1i) are used sequentially^53,61–65.

Conversely, sequencing HER2-targeted ADCs with different payloads appears more effective. In the phase 3 DESTINY-Breast02 trial, T-DXd significantly improved mPFS versus physician’s choice in HER2-positive mBC patients after T-DM1 (17.8 vs 6.9 months; HR 0.36, 95% CI, P < 0.0001)⁶⁶. However, we cannot exclude that these results are likely to be influenced also by other key differences between the two ADCs, such as the linker characteristics, the DAR, and the bystander effect. Interestingly, a recent post hoc exploratory analysis of DESTINY-Breast02 and DESTINY-Breast03 (phase 3 trial comparing T-DXd with T-DM1 in pretreated HER2-positive mBC) suggests that T-DM1 can be an effective treatment option after T-DXd, showing a median treatment duration of 7.6 months (range 0.0–42.6) in patients receiving T-DM1 after progression on T-DXd⁶⁷. These findings support payload diversification as a strategy to overcome resistance^24,49. Several phase 2 trials are ongoing to further explore these approaches (Table 1).

Table 1.

Ongoing clinical trials evaluating sequencing strategies after prior ADC exposure in BC

Clinical trial (name/NCT ID)	Phase	Type of sequencing	Setting
SATEEN/NCT06100874	2	SG + Trastuzumab after PD on T-DXd	HER2+ mBC
TRADE-DXd/NCT06533826	2	Dato-DXd or T-DXd after PD on prior ADC	HER2- (HER2-low or HER2-0) mBC
SERIES/NCT06263543	2	SG after PD on T-DXd	HR+/HER2-low mBC
ACE-Breast-03/NCT04829604	2	ARX788 after PD on T-DXd	HER2+ mBC
ICARUSBREAST02/NCT06298084	1b/2	HER3-DXd + Olaparib after PD on T-DXd	HER2-low mBC

ADC antibody–drug conjugate, mBC metastatic breast cancer, TNBC triple-negative breast cancer, HER2 human epidermal growth factor receptor 2, + positive, - negative, HR hormone receptor, SG sacituzumab govitecan, Dato-DXd datopotamab deruxtecan, T-DXd trastuzumab deruxtecan, HER3-DXd patritumab deruxtecan, PD progression disease.

This is a non-exhaustive list intended to illustrate examples.

Role of the tumor microenvironment (TME)

Beyond intracellular mechanisms, the TME plays a crucial role in ADC response and resistance (Fig. 1f). The TME can affect the ADC intratumoral distribution and intervene in the ADC degradation and payload release. On the one side, cancer-associated fibroblasts (CAFs) hinder ADC penetration by producing extracellular matrix proteins, and releasing transforming growth factor β, and also induce an immunosuppressive TME. On the other hand, recent preclinical data indicate that extracellular proteases, like cathepsin L (CTSL), in the TME contribute to T-DXd efficacy in HER2-low and HER2-negative BC, independently of HER2 expression or ADC internalization. High CTSL expression in tumor and stromal compartments enables efficient linker cleavage and payload release, promoting cytotoxicity⁶⁸.

The immune cells present in the TME may also have a major role in determining the ADC activity. Among the various forms of regulated cell death, immunogenic cell death (ICD) has emerged as a key process in promoting tumor antigen-specific immunity and enhancing responses to anticancer therapies. Comparative studies indicate that the DXd payload induces stronger hallmarks of ICD than DM1, including higher HMGB1 release and increased calreticulin surface exposure on dying tumor cells, suggesting a more potent capacity to stimulate antitumor immune activation⁶⁸. Consistently, RNA-seq analysis of co-cultured macrophages reveals marked upregulation of antigen-presentation and chemokine genes with both HER2 ADCs, with T-DXd demonstrating more potent immune-stimulatory effects than T-DM1. In HER2-positive BC cell lines, in addition, both T-DXd and T-DM1 show the ability to promote antibody-dependent cellular phagocytosis (ADCP) via Fcγ-receptor engagement. Notably, T-DXd induces markedly stronger immune activation than T-DM1 or trastuzumab, with enhanced T-cell proliferation and CD44 expression⁶⁸. However, it also upregulates tumor CD47, limiting immune activation. Combining T-DXd with CD47 blockade enhances T-DXd efficacy and induces durable CD8 + T cell memory⁶⁸.

Other prior preclinical studies consistently showed that ADCs reduce TME immunosuppression, activate dendritic cells (DC) and CD8 + T cells, and upregulate Programmed Death-Ligand 1 (PD-L1) and Major Histocompatibility Complex class I (MHC-I) expression on tumor cells^69–71.

Emerging clinical trials are starting to validate some of this evidence. In the ICARUS-BREAST01 clinical trial, response to HER3-DXd was associated with upregulation of immune-related pathways such as interferon-alpha and gamma pathways¹⁹.

Finally, ADCs can also modulate cancer stem cells (CSCs), known for their chemo-resistance and stem-like properties, by promoting receptor tyrosine kinase-like orphan receptor 1(ROR1) signaling^72–74. In BC models, T-DM1 increases CSCs via Yes-associated protein 1 (YAP1)-mediated ROR1 upregulation⁷⁵. Therefore, co-inhibition of YAP1 and ROR1 with HER2-targeted therapies has shown promise in preclinical studies^20,75.

Possible strategies to overcome ADC resistance

ADC combination

Combining ADCs with other therapies may overcome resistance and enhance efficacy, especially in heterogeneous tumors. However, careful patient selection and clinical context are crucial to balance benefits with the risk of increased adverse events (AEs) (Fig. 2a). Ongoing clinical trials are listed in Table 2.

Fig. 2. Strategies to overcome ADC resistance.

a Combination therapies. ADCs combined with immune checkpoint inhibitors (ICIs), DNA-damage response inhibitors (DDRinh), tyrosine kinase inhibitors (TKIs), other ADCs, antiangiogenic antibodies (antiVEGF Abs), Heat Shock Protein (HSP)90 inhibitors, anti-trastuzumab Abs, Signal Transducer and Activator of Transcription 3 (STAT 3) inhibitors, Phosphoinositide 3-Kinase (PI3K) and anti-epidermal growth factor Receptor (EGFR) Abs. b New-generation payloads: use of next-generation cytotoxic agents, degrader–antibody conjugates (DACs), glue–antibody conjugates (MACs), immune stimulator antibody conjugates (ISACs), antibody–oligonucleotide conjugates (AOCs), and ADCs with multiple payloads. c Innovative ADC structures: small-format drug conjugates, bispecific ADCs (BsADCs), probody–drug conjugates (PDCs), enhanced bystander effect, innovative linker technologies, and conjugation strategies.

Table 2.

Ongoing clinical trials evaluating ADCs in combination strategies for breast cancer

Clinical trial identifier	Phase	Investigational combination	Setting	Study name
ADC + ICI
NCT06508216	1b/2	Dato-DXd + durvalumab	Early-relapsing mTNBC	COMPASS-TNBC
NCT03742102	1b/2	Dato-DXd + durvalumab	PD-L1 + LA/mTNBC	BEGONIA (Arm 8)
NCT05353361	1b/2	SHR-A1811 + adebrelimab (or pyrotinib or pertuzumab)	LA/mTNBC	—
NCT01042379	2	Neoadjuvant Dato-DXd + durvalumab	HER2-/Immune-positive early BC	I-SPY2.2
NCT05629585	3	Adjuvant Dato-DXd ± durvalumab	TNBC with residual disease post-NAST	TROPION-Breast03
NCT04873362	3	Adjuvant T-DM1 + atezolizumab	Early HER2 + BC with residual disease	ASTEFANIA
NCT05633654	3	Adjuvant SG + pembrolizumab	TNBC with residual disease post-NAST	ASCENT-05/ OptimICE-RD
NCT04042701	3	Dato-DXd ± durvalumab first-line	PD-L1 + LA/mTNBC	TROPION-Breast05
NCT06112379	3	Neoadjuvant Dato-DXd + Durva (Adj. Durva)	Early-stage TNBC or HR-low/HER2- BC	TROPION-Breast04
NCT04740918	3	T-DM1 + atezolizumab	HER2+ and PD-L1 + LA/mBC	KATE 3
NCT06312176	3	Sacituzumab tirumotecan± pembrolizumab	HR + / HER2- LA/mBC	TroFuse-010
NCT06841354	3	Sacituzumab tirumotecan± pembrolizumab	PD-L1 < 10 LA/mTNBC	TroFuse-011
ADC + rDDRinh
NCT04039230	1b/2	SG + talazoparib	mBC	—
NCT06298084	1b/2	HER3-DXd + olaparib	HER2-low mBC post-T-DXd	ICARUSBREAST02
ADC + targeted therapy
NCT06331169	1	T-DXd + anlotinib	HER2-low LA/mBC	ALTER-BC-Ib-01
NCT05575804	1b/2	GQ1001 + pyrotinib	HER2+ mBC	—
NCT06157892	1b/2	DV + tucatinib	HER2 + LA/mBC or gastric/GEJC	—
NCT04983121	2	Neoadjuvant ARX788 + pyrotinib	HER2 + BC	—
NCT04539938	2	T-DXd + tucatinib	HER2+ mBC ± active brain metastasis	HER2CLIMB-04
NCT06000033	2	DV + anlotinib	HR–, HER2-low mBC	—
NCT06015113	NA	DV + pyrotinib or neratinib	HER2+ mBC + brain metastasis	—
NCT06278870	3	DV + pyrotinib first-line	HER2+ mBC	—
NCT06927180	2	Neoadjuvant SHR-A1811 + pertuzumab	HER2 + BC	—

ADC antibody–drug conjugate, ICI immune checkpoint inhibitor, rDDRinh replication DNA damage response inhibitor, mBC metastatic breast cancer, TNBC triple-negative breast cancer, LA locally advanced, HER2 human epidermal growth factor receptor 2, HR hormone receptor, PD-L1 programmed death-ligand 1, SCLC small cell lung cancer, SCNEC small cell neuroendocrine cancer, GEJC gastroesophageal junction cancer, DV disitamab vedotin, SG sacituzumab govitecan, Dato-DXd datopotamab deruxtecan, T-DM1 trastuzumab emtansine, T-DXd trastuzumab deruxtecan, HER3-DXd patritumab deruxtecan.

This is a non-exhaustive list intended to illustrate examples.

ADC+ immune checkpoint inhibitor (ICI)

The rationale for this combination is based on data showing that combining ADCs with ICIs enhances CD8 + T cell infiltration and activity^70,76,77. In the phase 2 KATE2 trial, T-DM1 plus atezolizumab did not improve PFS and showed more serious AEs than T-DM1 alone in HER2-positive pretreated mBC. However, a potential benefit in PD-L1–positive tumors prompted the ongoing KATE3 trial^78,79. In the BEGONIA platform trial (first-line mTNBC), Dato-DXd plus durvalumab (Arm 7) showed a 79% ORR (95% CI, 67–88) and 13.8-month mPFS (95% CI, 11–not calculable [NC])⁸⁰; T-DXd plus durvalumab (Arm 6) achieved a 57% ORR (95% CI, 41–71) and 12.6 months PFS (95% CI, 8–not reached)^71,81. Responses occurred regardless of PD-L1 status. Arm 8 is focusing on PD-L1–positive patients receiving Dato-DXd +durvalumab⁸² (Table 2).

In DS8201-A-U105 phase 1b trial, T-DXd plus nivolumab yielded ORRs of 65.6% (95% CI, 46.8–81.4%) and 50.0% (95% CI, 24.7–75.3%) for pretreated HER2-positive and HER2-low mBC, regardless of PD-L1 status, though limited by small sample sizes⁸³.

The MORPHEUS-panBC umbrella study in untreated PD-L1-positive mTNBC reported a 76.7% ORR for SG plus atezolizumab vs 66.7% in the control arm, with responses linked to TROP2 and stromal Tumor-Infiltrating Lymphocytes (TILs)⁸⁴.

In the Phase 3 ASCENT-04/KEYNOTE-D19 trial (treatment-naïve, PD-L1–positive locally advanced [LA]/mTNBC), SG plus pembrolizumab improved PFS over chemo-pembrolizumab (HR 0.65; 95% CI 0.51–0.84, P = 0.0009), with a median duration of response of 16.5 versus 9.2 months and an early positive trend in OS⁸⁵.

Encouraging response rates were also reported in the early BC setting; in early TNBC, Arm A2 of the phase 2 NeoSTAR trial tested neoadjuvant SG plus pembrolizumab, achieving a protocol-defined pCR of 34% (16/50; 95% CI, 19.5–46.7)⁸⁶. Both Dato-DXd as a single agent and in combination with durvalumab have been evaluated in the I-SPY2.2, a phase 2 neoadjuvant study, as part of sequential treatments in patients with HER2-negative stage II-III BC. Particularly, in the subtype defined as immune-positive according to the I-SPY classification system, the combination of Dato-DXd + durvalumab for 4 cycles showed a pCR of 54% that reached 92% after the sequential treatment with taxanes⁸⁷ (Table 2).

ADC + DNA-damage response inhibitor (DDRinh)

Combining ADC with a DDRinh offers a promising strategy to overcome resistance to TOP1i payloads through “synthetic lethality”. TOP1i cause DNA damage, which is normally repaired via Poly(ADP-ribose) Polymerase (PARP)-dependent pathways. PARP inhibition blocks this repair, determining DNA breaks accumulation and enhanced cell death^1,34,88,89. Preclinical models showed synergy between SG and talazoparib or olaparib, with significant tumor growth delay (P < 0.0017)⁸⁹.

However, tolerability, especially myelosuppression, remains a key limitation^34,90,91. In the phase 1b SEASTAR trial, rucaparib plus SG, showed activity, even post-DDRinh, but with high hematologic toxicity⁹¹.

In the NCT04039230 phase 1b study, concurrent SG-talazoparib caused high dose-limiting toxicities (DLTs-71.4%), while sequential administration reduced toxicity and maintained efficacy, enabling phase 2 dose selection⁹⁰. Trials are ongoing to refine scheduling (Table 2).

ADC + targeted therapy

Incorporating tyrosine kinase inhibitors (TKIs) into dual-targeted approaches may enhance ADC efficacy and selectivity, though data on resistance are still limited¹. In EGFR-mutant NSCLC models, osimertinib plus T-DM1 showed synergy, with T-DM1 delaying osimertinib resistance⁹². Furthermore, EGFR inhibition may upregulate HER3, potentially enhancing HER3-DXd uptake in preclinical models⁹³.

Similarly, HER2-targeting TKIs also enhance T-DM1 efficacy: lapatinib increases HER2 expression, neratinib promotes internalization, while tucatinib’s impact is less defined^1,94–97. The phase 2 TEAL study found improved responses with T-DM1+ lapatinib + nab-paclitaxel as neoadjuvant treatment in HER2-positive BC⁹⁸. In pretreated HER2-positive mBC, the phase 1b NCT05575804 trial of GQ1001 plus pyrotinib showed a 71.4% ORR but high-grade 3–4 AEs (73.3%)⁹⁹. Another phase 2 trial combining pyrotinib with DV determined a 31.6% ORR and mPFS of 7.1 vs 4.6 months with DV alone, without new safety signals¹⁰⁰.

In HER2CLIMB-02 phase 3 trial, tucatinib plus T-DM1 improved PFS (HR 0.76, 95% CI, P = 0.0163) in pretreated HER2-positive LA/mBC, with OS data pending^101,102.

Modulating tumor vasculature through antiangiogenic agents may improve ADC delivery¹. In platinum-resistant epithelial ovarian cancer models, bevacizumab enhanced tumor necrosis and vascular disruption when combined with mirvetuximab soravtansine or anetumab ravtansine^103,104. Clinical data support the safety and efficacy of the mirvetuximab–bevacizumab, though optimal dosing remains under investigation¹⁰⁵.

Combining T-DXd with EGFR mAbs (cetuximab or panitumumab) may overcome resistance by enhancing endocytosis of HER2–EGFR heterodimers and target-bound T-DXd, thereby improving its efficacy³⁵.

Dual HER2 inhibition, like T-DM1 plus pertuzumab, may improve ADC tumor distribution^43,106,107.

In the phase 3 MARIANNE trial, T-DM1 ± pertuzumab was noninferior to the control arm (mOS 51.8 vs 50.9, HR 0.86, 97.5% CI, 0.67–1.11) in first-line HER2-positive mBC¹⁰⁸. KRISTINE phase 3 trial, evaluating neoadjuvant therapy with TDM1 plus pertuzumab in early HER2-positive BC, showed lower pCR with the combination, but similar DFS (HR 1.11; 95% CI, 0.52–2.40)¹⁰⁹. In the DESTINY-Breast09 phase 3 trial, T-DXd plus pertuzumab improved mPFS versus the control arm (HR 0.56; 95% CI; P < 0.00001), in first-line HER2-positive mBC¹¹⁰.

Targeting resistance pathways is promising in preclinical studies. STAT3 inhibition (napabucasin), HSP90 inhibition (geldanamycin), and Phosphoinositide 3-Kinase ([PI3K] CDC-0941) enhanced T-DM1 efficacy in HER2-positive models, though toxicity remains a concern^58,111–113. Several trials are ongoing (Table 2).

ADC + ADC

Combining ADCs with non-overlapping toxicities may enhance efficacy. In the phase 1 DAD trial, SG plus enfortumab vedotin showed a 70% ORR (95% CI, 47–87%) in metastatic urothelial carcinoma¹¹⁴. Currently, no clinical trials are assessing ADC combinations in BC, their feasibility would depend on manageable toxicity profiles.

New-generation payloads

Cytotoxic payloads

To overcome ADC resistance, one strategy is using payloads less susceptible to efflux or resistance mechanisms (Fig. 2b). PNU-159682, a highly potent anthracycline (~100× doxorubicin) not affected by efflux pumps, showed preclinical efficacy where monomethyl auristatin (MMA)E or DM1 failed¹¹⁵. Another approach uses payloads like eribulin, which penetrates cells efficiently due to its balanced hydrophilicity/lipophilicity¹¹⁶. BB-1701, an eribulin-based HER2-targeting ADC, outperformed T-DM1 and T-DXd in HER2-expressing and HER2-low models, with strong bystander and immune-modulating effects (e.g., TILs recruitment and induction of immunogenic cell death)¹¹⁷. Preliminary phase 1 data suggest antitumor activity and manageable safety in HER2-low BC, including patients pretreated with anti-HER2 ADCs¹¹⁸ (Table 3).

Table 3.

Active clinical trials of innovative ADC platforms in breast cancer

Clinical trial identifier	Phase	Investigational agent	Setting	Platform type
NCT05511844	1	ORM-5029	HER2+ advanced solid tumors	MAC
NCT05399654	2	TAC-001	Advanced solid tumor	ISAC
NCT05070247	1/2	TAK-500 ± Pembrolizumab	LA/m solid tumor	ISAC
NCT05514717	1	XMT-2056	HER2+ advanced/recurrent Solid tumors	ISAC
NCT04561362	1/2	BT8009 ± Pembrolizumab	Nectin-4+ advanced malignancies	Small format
NCT04180371	1/2	BT5528 ± Nivolumab	EphA2+ advanced solid tumors	Small format
NCT06042894	2	SI-B003 ± BL-B01D1	HER2– LA/mBC	BsADC
NCT06382142	3	BL-B01D1 + Chemotherapy	LA/mTNBC	BsADC
NCT05470348	1	BL-B01D1	LA/mBC and solid tumors	BsADC
NCT06471205	2	BL-B01D1 + anti-PD-1	LA/mTNBC	BsADC
NCT06343948	3	BL-B01D1 + Chemotherapy	HR+/ HER2– LA/mBC	BsADC
NCT06846437	3	JSKN003	Pretreated HER2 + LA/mBC	BsADC
NCT06079983	3	JSKN003	Pretreated LA/mBC	BsADC
NCT06592417	1	JSKN016	Advanced solid tumors	BsADC
NCT05735496	1	TQB2102	Advanced solid tumors	BsADC
NCT06452706	2	TQB2102	HER2– mBC	BsADC
NCT07008976	3	TQB2102	HER2+ mBC	BsADC
NCT06198751	2	TQB2102 neoadjuvant	Early HER2 + BC	BsADC
NCT06561607	3	TQB2102	HER2-low BC	BsADC
NCT07003074	3	TQB2102	HER2+ recurrent/mBC	BsADC
NCT06942234	1/2	JSKN016 + combination therapy	HER2– LA/mBC	BsADC
NCT06349408	1	IBI3001	Solid tumors	BsADC
NCT06418061	1	IBI3005	Solid tumors	BsADC
NCT06493864	1	BL-B16D1	BC and other solid tumors	BsADC
NCT06475937	1	DM001	Advanced solid tumors	BsADC
NCT06554795	1/2	DB-1419	Advanced/m solid tumors	BsADC
NCT06725381	1	SKB571	Advanced solid tumors	BsADC
NCT06707610	1	ALK202	Advanced solid tumors	BsADC
NCT06685068	1/2	GEN1286	Advanced solid tumors	BsADC
NCT06751329	1	DM002	Advanced solid tumors	BsADC
NCT05785039	2	BL-B01D1	Urinary system and other solid tumors	New Linker/Conj.
NCT04829604	2	ARX788	HER2+ mBC	New Linker/Conj.
NCT05377996	1	XMT-1660	Solid tumors	New Linker/Conj.
NCT 04257110	1	BB-1701	LA/m HER2+ Solid Tumors	Bystander effect
NCT06188559	2	BB-1701	HER2 + / HER2-low mBC	Bystander effect

ADC Antibody–drug Conjugate, MAC Glue–antibody Conjugate, ISAC immune-stimulating antibody conjugate, LA locally advance, m metastatic, BC breast cancer, Small format small format antibody–drug conjugate, BsADC bispecific antibody–drug conjugate, TNBC triple-negative breast cancer, HR+ hormone receptor-positive, HER2 human epidermal growth factor receptor 2, PD-1 programmed cell death protein 1, HNSCC head and neck squamous cell carcinoma, Conj. conjugation.

This is a non-exhaustive list intended to illustrate examples.

Trastuzumab rezetecan (SHR-A1811), an HER2-targeting ADC carrying a TOP1i payload, showed promising results in a phase 1 trial with ORRs in HER2-positive BC (76.3%), in HER2-low BC (60.4%), and non-breast tumors (45.9%)¹¹⁹. Notably, responses persisted in patients previously treated with T-DM1 (82.4%) and other HER2 ADCs, including T-DXd¹²⁰.

ADC with multi-payloads

Dual-drug ADCs aim to overcome tumor heterogeneity by enhancing antitumor activity and accumulation, while avoiding binding competition seen with co-administered single-drug ADCs¹²¹ (Fig. 2b). However, ensuring consistent duo-payload conjugation with uniform DARs remains a challenge. Advances in linker technology, such as cysteine/selenocysteine engineering and hetero-trifunctional linkers, enable site-specific, stable attachment of dual payloads like PNU-159682 and MMAF, enhancing cytotoxicity in HER2-positive models^122–124. However, all related clinical trials have been discontinued to date.

Degrader–antibody conjugates (DACs)

Proteolysis-targeting chimeras (PROTACs) are heterobifunctional molecules that induce targeted protein degradation by recruiting E3 ligases to specific intracellular proteins¹²⁵. Though limited cell permeability, their conjugation to mAbs enables targeted delivery, leading to the development of DACs^115,126,127 (Fig. 2b).

Structurally similar to ADCs, DACs carry higher DARs due to PROTACs’ lower potency. After receptor-mediated endocytosis, PROTACs are released into the cytoplasm to trigger degradation¹²⁸. Maneiro et al. developed a trastuzumab-PROTAC conjugate to selectively degrade bromodomain-containing protein 4 (BRD4) in HER2-positive BC cell lines, without affecting HER2-negative cells¹²⁹. To date, no clinical trials are ongoing.

Molecular glues and glue–antibody conjugates (MACs)

MACs represent a novel ADC class using monovalent degraders (molecular glues) as payloads that recruit E3 ligases like cereblon to trigger proteasomal degradation of neo substrates proteins, such as G1 to S phase transition 1 (GSPT1), often overexpressed in cancer, including BC^130,131 (Fig. 2b). ORM-5029, an AnDC™ (Antibody neoDegrader Conjugate) combining the GSPT1 degrader SMol006 with pertuzumab via a Val-Cit PABc linker, shows greater cytotoxicity than other GSPT1 degraders and comparable efficacy to T-DXd in HER2-positive models¹³². A phase 1 trial is ongoing¹³³ (Table 3).

Immune-stimulating antibody conjugates (ISACs)

ISACs combine a tumor-targeting antibody, non-cleavable linker, and immunostimulatory payload to boost anti-tumor immunity by activating innate and adaptive immune responses within TME, overcoming some limitations of conventional ADCs (Fig. 2b). However, systemic immune activation and cytokine release syndrome remain safety concerns. Development focuses on Toll-Like Receptors (TLRs), particularly TLR7/8/9 and Stimulator of Interferon Genes (STING) agonists¹³⁴. STING activation triggers type I interferon responses^134,135. IMSA1729, an EGFR-targeting STING-agonist ADC, showed strong tumor control and synergy with anti-PD-L1 in preclinical melanoma models¹³⁶. XMT-2056, a HER2-targeting STING-agonist ISAC, demonstrated broad efficacy, including in HER2-low models, with manageable cytokine induction¹³⁷; a phase 1 trial is ongoing (Table 3).

TLRs recognize pathogen-associated patterns to initiate immune responses¹³⁴. SBT6050 (HER2-targeting TLR8-agonist) showed localized immune activation preclinically and entered early-phase trials, alone or with ICIs and HER2-targeted therapies (NCT04460456, NCT05091528), but both were discontinued^138,139, other trials are ongoing (Table 3).

Antibody–oligonucleotide conjugates (AOCs)

AOCs combine the gene-silencing ability of oligonucleotides, like small interfering RNA (siRNA) and anti-sense oligonucleotides, with the target deliverability of antibodies, with the goal of improving serum stability, membrane permeability, and tissue specificity (Fig. 2b). However, challenges remain, particularly in the conjugation of oligonucleotides to nanoparticles, due to their size and charge¹⁴⁰. A preclinical HER2-targeted AOCs using PLK1 siRNAs and HER2-ScFv-protamine fusion protein (F5-P) showed effective PLK1 inhibition, apoptosis induction, and tumor suppression in HER2-positive BC models¹⁴¹.

Innovative ADC structures

Small-format drug conjugates

Smaller antibody formats, like single-chain variable fragment (scFvs) and Fabs, improve tumor penetration and internalization, due to their reduced size compared to full-length immunoglobulins¹⁴² (Fig. 2c). Moreover, they present a lower off-target toxicity from lacking an Fc domain¹³. However, these formats also face limitations, including rapid systemic clearance¹⁴². Bicycle peptides offer a promising alternative, combining small size with efficient uptake¹¹⁵ (Table 3).

Bispecific antibody–drug conjugates (BsADCs)

BsADCs are bispecific antibody conjugates targeting two distinct epitopes on the same or different antigens, improving tumor selectivity and internalization in heterogeneous or resistant tumors¹⁴³ (Fig. 2c). Biparatopic BsADCs, bind two non-overlapping epitopes on the same antigen like HER2, enhancing receptor clustering and degradation, showing activity in T-DM1-resistant and HER2-low settings^144,145.

Andreev et al. developed a HER2/ Prolactin Receptor (PRLR) BsADC that enhanced HER2 degradation and outperformed HER2-ADCs in dual-expressing tumors¹⁴⁶. Other preclinical BsADCs, including HER2/HER3, TROP2/HER2, demonstrated superior antitumor activity versus monospecific ADCs^13,147.

BL-B01D1, targeting EGFR/HER3 with a camptothecin payload, showed promising activity in early trials^148,149 (Table 3). In HER2-negative or HER2-low mBC, it showed an ORR of 42.1% and a mPFS of 6.9 months¹⁴⁹.

TQB2102, a BsADC targeting HER2 extracellular domains II and IV, showed preliminary antitumor activity in a phase 1 study (NCT05735496) in advanced solid tumors, including mBC, with an ORR of 41.2% and good tolerability. Notably, 31% of HER2-expressing mBC patients had received prior anti-HER2 ADCs, including T-DM1 and T-DXd¹⁵⁰.

JSKN016, a TROP2/HER3-targeting BsADC, showed preliminary antitumor activity (ORR 80%) and a manageable safety profile in pretreated mTNBC in an ongoing phase 1 trial (NCT06592417)¹⁵¹.

A list of ongoing trials is provided in Table 3.

Prodrug-conjugated antibodies (PDCs)

Emerging ADCs aim to mitigate on-target, off-tumor toxicity, especially when targeting antigens like EGFR or TROP2, also found on normal tissues^7,8. PDCs use inactive payloads or self-masking domains that are selectively activated in the tumor TME by proteases or low pH, enhancing tumor specificity and minimizing systemic damage (Fig. 2c). However, precise tumor profiling is needed to match masking strategies effectively^7,152.

New linkers and novel conjugation strategies

Increasing linker hydrophilicity is a key strategy to enhance ADC efficacy and overcome resistance (Fig. 2c).

Hydrophilic linkers, like polyethylene glycol-4 Maleimide (PEG4Mal), improve antitumor activity and reduce clearance, especially in MDR-expressing tumors, outperforming hydrophobic linkers⁴⁷. Other hydrophilic linkers (sulfonate, PEG, polysarcosine, DNA-based linkers) improve stability, pharmacokinetics, and reduce toxicity¹¹⁵. Traditional stochastic conjugation yields heterogeneous ADCs with variable DARs⁷, while site-specific conjugation (e.g., engineered cysteines, unnatural amino acids [UAAs], enzymatic ligation) improve homogeneity and therapeutic index^115,153.

GQ1001, a trastuzumab–DM1 ADC using ligase-dependent conjugation, reduces free DM1 toxicity and showed 40% ORR with manageable side effects in a phase 1 trial^154,155. ARX788, a site-specific HER2-targeted ADC using UAA-MMAF conjugation, demonstrated activity in HER2-positive and HER2-low mBC, including pretreated T-DXd patients, a phase 2 trial is ongoing¹⁵⁶ (Table 3). In ACE-Breast-02 phase 3 trial, ARX788 improved PFS (HR 0.64, 95% CI P = 0.0006) versus lapatinib-capecitabine with low discontinuation¹⁵⁷.

Emlitatug ledadotin, a site-specific B7-H4 ADC, showed promising activity (ORR 23%; 95% CI, 5–54%) in heavily pretreated B7-H4–high mTNBC¹⁵⁸.

In addition, branched linkers, such as Kumar’s hetero-trifunctional linker, enable higher DARs or dual-payload designs, further enhancing ADC potency¹²⁴. Some clinical trials are ongoing (Table 3).

Conclusion

Understanding resistance mechanisms to ADCs is essential for developing new agents capable of overcoming them. Although many have been identified in preclinical studies, further research is needed to clarify their clinical relevance. To address resistance in clinical practice, several strategies are currently under development. These include combination therapies with other agents, such as ICI, DDRinh, and TKI, as well as the use of new-generation payloads and alternative formats, such as DACs, MACs, ISACs, AOCs, or multi-payload ADCs. Another promising approach involves moving beyond the traditional ADC structure, with the development of small-format drug conjugates, BsADCs, PDCs, and optimization of the bystander effect, linker technologies, and conjugation strategies. Multiple clinical trials are currently ongoing to assess the efficacy and safety of these approaches. In addition, the identification of predictive dynamic biomarkers, such as ctDNA, plasma copy number variations, and artificial intelligence-based analyses, may enable real-time monitoring of ADC efficacy and resistance development.

Author contributions

I.V. and T.G.: conceptualization, design, and initial drafting of the manuscript. B.P.: conceptualization, design, and critical revision of the manuscript. T.G., L.A., and B.P.: critical revision of the manuscript. All authors reviewed and approved the final version of the manuscript.

Data availability

No datasets were generated or analyzed during the current study.

Competing interests

I.V. declares no competing interests. T.G. reports having received travel support from AstraZeneca, Gilead, and Pfizer; has served in a consulting/advisory role for AstraZeneca and Cancerologie-Pratique; and has received research funding from Amgen. L.A. reports relationships with Amgen, Roche, AstraZeneca, and Novartis involving royalties, and has served in an advisory role for Merck Serono. B.P. has received consulting fees (institutional) from AstraZeneca, Seagen, Gilead, Novartis, Lilly, MSD, and Pierre Fabre; personal consulting fees from Pierre Fabre and Daiichi-Sankyo; research funding (institutional) from AstraZeneca, Daiichi-Sankyo, Gilead, Seagen, and MSD; and travel support from AstraZeneca, Pierre Fabre, Lilly, Daiichi-Sankyo, and MSD.