Clinical Development of ADCs for Solid Tumors: 40 Years of Progress

Clinical Development of ADCs for Solid Tumors

The currently approved ADCs for solid tumors are listed in the table below:

ADCs	Target	Indications
T-DM1	HER2	HER2-positive metastatic breast cancer
SG (sacituzumab govitecan)	Trop2	Triple-negative breast cancer (TNBC) Metastatic urothelial carcinoma (mUC)
sacituzumab tirumotecan	Trop2	TNBC, NSCLC Gastric cancer/gastroesophageal junction cancer, mUC
enfortumab vedotin	Nectin-4	Locally advanced or metastatic urothelial carcinoma (la/mUC)
mirvetuximab soravtansine	FRα	Ovarian cancer, fallopian tube cancer, or primary peritoneal cancer
isotumab vedotin	TF	Refractory cervical cancer
T-DXd	HER2	HER2-positive cancers

Among these, the clinical development path of T-DXd is particularly noteworthy:

• Initially approved for third-line treatment of HER2-positive metastatic breast cancer (15%-20% of breast cancer cases)

• Subsequently demonstrated remarkable efficacy (4-fold increase in PFS) compared to T-DM1 in head-to-head trials, leading to approval for second-line treatment

• Indications expanded to include HER2 low-expressing tumors (45%-55% of breast cancer cases)

• Currently being tested in patients with ultra-low or no HER2 expression (DAISY Phase II trial shows ORR of 30%)

• In the DESTINY-Breast06 trial, T-DXd significantly improved PFS in HR+/HER2 low/ultra-low expressing (IHC>0<1+) metastatic breast cancer patients compared to standard treatment

• Indications for HER2-positive gastric cancer and HER2 mutant non-small cell lung cancer have been approved

• DESTINY-PanTumor02 Phase II trial shows significant activity of T-DXd in various HER2 overexpressing tumors, leading to the first accelerated approval of a “pan-tumor” ADC

Challenges in ADC Clinical Development and Application

Despite significant achievements, ADC clinical development still faces three major challenges: first, the toxicity profile is similar to that of traditional chemotherapy. Second, there is controversy over the sequential use of ADCs, as changes in antibody or payload targets may lead to cross-resistance. Third, identifying predictive biomarkers for ADCs is difficult, as many approved ADCs lack companion diagnostics.

ADC Toxicity

Meta-analysis shows that the overall incidence of treatment-related adverse events for ADCs exceeds 90%, with 46% being grade 3 or higher. Current strategies to improve the toxicity profile of ADCs include: optimizing dosage and regimens, developing toxicity biomarkers, implementing remote monitoring, and educating on early toxicity recognition. ADC engineering technologies (including site-specific conjugation, stable linkers, and/or antibody masking techniques) are also being used to attempt to reduce the incidence and severity of side effects, while improving tumor targeting and uptake. However, to date, these technologies have not shown significant clinical improvement, and ADC development still largely relies on empirical exploration.

Sequential Use of ADCs

As numerous ADCs are about to enter clinical use, a key question arises: will cross-resistance limit efficacy when sequentially using ADCs with overlapping targets or similar mechanism payloads?

Three ADCs have been approved for breast cancer: T-DM1, T-DXd, and SG. T-DM1 and T-DXd target the same HER2 antibody but differ in payloads, linkers, and DAR; T-DXd and SG have different antibodies and linkers but share the TOPO1i mechanism (despite significant structural differences between DXd and SN38, and differing linker stability). The sequential use of T-DM1, T-DXd, and SG suggests the impact of overlapping features (same antibody target or payload mechanism):

• Sequential use of ADCs sharing the HER2 target but with different mechanisms generally does not show significant cross-resistance

• Data on sequential use of ADCs with different targets but the same mechanism is limited, but early data and real-world experience consistently suggest potential cross-resistance

Currently, the Translational Breast Cancer Research Consortium is conducting a prospective registry study and three Phase II trials (SATEEN, TRADE-DXd, and SERIES) aimed at prospectively answering the question of sequential use of ADCs with the same payload. If cross-resistance is confirmed, the differentiation of payloads will become essential. Currently, among over 200 ADCs in development, most still rely on three payload mechanisms: microtubule inhibition, TOPO1 inhibition, and DNA alkylation. Expanding the variety of payloads to other cytotoxic classes is expected to enhance efficacy and will be a key direction for future ADC development.

Exploration of Predictive Biomarkers for ADCs

Theoretically, given the targeted mechanism of ADCs, high expression of antibody targets should predict better efficacy. Indeed, certain ADCs have shown superior ORR, PFS, and OS in patients with high target expression. This applies to T-DM1, which is approved only for HER2-positive breast cancer, or mirvetuximab soravtansine, which is approved only for platinum-resistant ovarian cancer with high expression of folate receptor α. T-DXd shows better efficacy in HER2-positive (IHC 3+ or IHC 2+/FISH+) patients compared to those with low HER2 expression, although the benefits for the latter remain clinically significant.

However, for most ADCs, there is no clear correlation between target expression and response rate. For example: SG (targeting TROP2) in metastatic TNBC, HR+/HER2- breast cancer, endometrial cancer, and urothelial carcinoma; tisotumab vedotin (targeting tissue factor) in cervical cancer; enfortumab vedotin (targeting nectin-4) in urothelial carcinoma, HR+/HER2- breast cancer, and TNBC; brentuximab vedotin (targeting CD30) in T-cell and B-cell non-Hodgkin lymphoma; polatuzumab vedotin (targeting CD79b) in diffuse large B-cell lymphoma; loncastuximab tesirine (targeting CD19) in B-cell non-Hodgkin lymphoma, etc.

The relationship between specific target characteristics of ADCs and efficacy remains unclear, as several ADCs using similar linker-drug technologies to approved ADCs have failed to demonstrate clinically meaningful efficacy. Possible reasons include: tumor type sensitivity to payload classes; differences and heterogeneity in target expression between tumor and normal tissues; and inherent characteristics of ADCs (binding affinity, tumor penetration, internalization, hydrophobicity, PK, etc.). For instance, preclinical studies show that ADCs with lower binding site barrier effects penetrate better in solid tumors (through co-administration of naked antibodies or antibody engineering techniques), which correlates with better efficacy. This has not been fully validated clinically, but preliminary evidence suggests that co-administration of unconjugated antibodies can improve the intratumoral distribution of antibody-dye conjugates (panitumumab-IRDye800CW).

Target expression IHC testing may not be the best predictor of efficacy for certain ADCs. Currently, various improvement strategies are being explored:

• New detection technologies to enhance quantitative accuracy of target expression: quantitative immunofluorescence, mass spectrometry, reverse-phase protein arrays, computational pathology-based IHC analysis, etc.

• Target expression and dynamic changes in circulating tumor cells (CTCs) as predictive indicators of ADC efficacy

• 89Zr-trastuzumab PET imaging to predict T-DM1 activity

• Specific genomic alterations (e.g., ERBB2 amplification) guiding ADC treatment: the HERALD trial demonstrated that plasma free DNA detection of HER2 amplification in advanced solid tumor patients receiving T-DXd resulted in an ORR of 56%

Researchers are actively exploring the relationship between tumor genomic and transcriptomic features and ADC activity:

• In terms of genomic alterations: for example, patients receiving SG treatment detected TROP2 (TACSTD2) mutations; CD30 target loss occurred in patients with relapsed treated with brentuximab vedotin; ERBB2 loss is a known resistance mechanism for trastuzumab and T-DM1; ERBB2 heterozygous deletion is associated with lower response rates and shorter PFS trends for T-DXd; ERBB2 activating mutations may enhance ADC internalization or tumor sensitivity to payloads, correlating with T-DXd efficacy; tumors overexpressing RAB5a (involved in receptor-mediated endocytosis) respond better to T-DM1.

• Identifying ADC payload sensitivity and resistance biomarkers, screening known and novel biomarkers: for example, high expression of ATP-binding cassette efflux transporters (inherent or acquired) is associated with chemotherapy resistance (including ADC payloads); expression of lysosomal membrane transport protein SLC46A3 may relate to sensitivity to DM1 and PBD; SLFN11 is a known predictive factor for DNA damage chemotherapy sensitivity, recently shown to potentially predict TOPO1i ADC sensitivity; TOP1 mutations lead to resistance to small molecule camptothecin, recently found in patients treated with TOPO1i ADCs; TUBB3 is a taxane resistance biomarker, potentially indicating resistance to various microtubule-inhibiting ADC payloads; UDP-glucuronosyltransferase (encoded by UGT1A1 gene) is responsible for the glucuronidation of SN38 payload of SG. High levels of SN38 (due to impaired glucuronidation) can lead to severe side effects—this phenomenon has been well established for irinotecan. Among 495 patients in the SG basket trial, those with the UGT1A1*28/*28 genotype (~10%) had a significantly increased incidence of adverse effects, with a doubled risk of severe anemia and neutropenia.

Future Directions

More ADCs are expected to be approved in the coming years: more ADCs carrying TOPO1i, microtubule inhibitors, and DNA damaging agents; exploration of approved ADCs in new indications; and combination assessments with ICI and other anticancer drugs.

Development of next-generation ADCs: aiming to minimize toxicity while maintaining (or enhancing) the efficacy of existing ADCs. The modular structure of ADCs provides a unique opportunity to achieve these goals through fine-tuning and innovation of various components (antibodies, linkers, payloads).

Exploration of groundbreaking structural innovations: innovations can be achieved by replacing one or more components of ADCs:

• Bispecific or dual-target ADCs enhance selectivity, improve internalization, or expand the beneficiary population by targeting multiple epitopes or proteins

• Introducing peptide masking techniques to prevent antibody binding to targets (until the masking is removed)

• Engineering antibodies to achieve pH-dependent binding or modulate Fcγ receptor binding to improve selectivity, tolerability, and/or PK characteristics

• Replacing antibodies entirely with smaller molecules (such as peptide-drug conjugates or antibody fragment-drug conjugates), which may improve tumor penetration despite faster clearance

• Site-specific conjugation techniques can design ADCs with higher uniformity and stability; currently, dozens are in clinical testing

• Recognizing the significant impact of shifting from microtubule inhibitors to TOPO1i payloads has rekindled interest in payload differentiation—including linking multiple chemotherapy payloads and exploring entirely new payload classes (immune stimulators, protein degraders, radioactive isotopes, etc.)

It should be noted that no innovative technology has yet demonstrated superior activity or safety signals compared to standard ADCs. For example:

• Praluzatamab ravtansine (CX-2009) using peptide masking technology showed low activity (ORR<10%) and high ocular toxicity (43%, ≥grade 3 in 11%) and other serious adverse events

• The doxorubicin peptide-drug conjugate AVA6000 (designed for selective activation in the tumor microenvironment) had an ORR<5%, with high rates of alopecia (52%), fatigue (50%), and nausea (33%)

• Several site-specific conjugated ADCs with enhanced linker stability and uniformity have exhibited unexpected toxicity: ARX788 (keratitis 46%, interstitial lung disease 34%, liver toxicity up to 68%); A166 (keratitis 84%, peripheral neuropathy 53%); DP303c (keratitis ~95%)

• While some innovative payload ADCs show activity signals, they have not yet reached the levels established by approved payload ADCs, and some are associated with significant toxicity. For example, the phase I trial of immune-stimulating antibody conjugate NJH395 showed no tumor responses, with high incidences of cytokine release syndrome (55%), fever (44%), and nausea (44%).

In summary, embracing the pharmacological characteristics of such complex drugs is a key step toward developing safer, more effective next-generation ADCs that do not exhibit cross-resistance with existing ADCs.

References:

Raffaele Colombo, Paolo Tarantino, Jamie R. Rich, Patricia M. LoRusso, Elisabeth G.E. de Vries; The Journey of Antibody–Drug Conjugates: Lessons Learned from 40 Years of Development. Cancer Discov 1 November 2024; 14 (11): 2089–2108. https://doi.org/10.1158/2159-8290.CD-24-0708