Model Suggests Tumor Microenvironment Normalization Could Improve Efficacy of Antibody-Drug Conjugates
In This Article
- Antibody-drug conjugates (ADCs) and bispecific antibodies are among the fastest-growing classes of cancer drugs.
- However, the unique properties of the dysfunctional tumor microenvironment (TME) hinder the delivery and efficacy of ADCs in solid tumors.
- A new mathematical model suggests TME normalization could improve ADC delivery and efficacy while preserving sufficient vessel permeability and functional vascular density.
Antibody-drug conjugates (ADCs) and bispecific antibodies are among the fastest-growing classes of cancer drugs. Their delivery and efficacy, however, are limited by the many complexities of the dysfunctional tumor microenvironment (TME).
Subscribe to the latest updates from Oncology Advances in Motion
A new paper published in the Journal for ImmunoTherapy of Cancer presents a mathematical model designed to explore and overcome these challenges. The researchers modeled HE-S2, an experimental immunomodulatory ADC (IM-ADC) comprising an anti-programmed cell death ligand 1 (PD-L1) antibody and the bifunctional immunomodulator D18. The model was calibrated using previously published data from mouse models of colon cancer and melanoma; this was not a clinical study.
Modeling indicated that ADC treatment facilitated CD8+T cell infiltration and cancer cell–killing function, driving tumor shrinkage. This, in turn, lowered interstitial fluid pressure (IFP) and improved drug delivery, creating a positive feedback loop.
“The key message here is that repairing the TME—that is, normalizing it—can enhance both the delivery and efficacy of the drugs,” says co-corresponding author Rakesh K. Jain, PhD, director of the Edwin L. Steele Laboratories in the Department of Radiation Oncology at Massachusetts General Hospital. “This work provides a mechanistic rationale for how and when to combine TME-normalizing strategies with IM-ADCs to improve outcomes in solid tumors, a setting where newer antibody-based therapies have had mixed clinical success.”
A thoughtfully designed mathematical model
Traditionally, cancer research and treatment have focused on the “seed” (the malignant cells comprising the tumor) rather than the “soil” (the surrounding environment in which the tumor grows). During a career spanning over 50 years, Dr. Jain has focused on the soil—widely considered an unconventional approach.
“Our lab views the tumor as an organ, not just a collection of cancer cells,” he says. “In pancreatic cancer, one of the most difficult cancers to treat, just 5% of the cells are cancer cells. Leukemia, the easiest cancer to treat, is 100% cancer cells. That tells you it’s not just the cancer cells that are the problem. It’s the neighborhood around them—the microenvironment. If you don’t look at the whole picture, you won’t get the efficacy you’re hoping for.”
While mathematical modeling has proven to be useful in studying antibody-based therapeutics, few existing models account for the unique properties of the TME that can impede drug delivery. These include:
- Abnormalities (particularly leakiness) of blood vessels feeding the cancer cells
- Uniform elevation of IFP
- The dense extracellular matrix between the blood vessel wall and cancer cells
For this study, the researchers built a mathematical model consisting of three linked compartments: the blood vasculature, the tumor-draining lymph node, and the tumor-host tissue. Their approach coupled a physiologically detailed transport model of the TME with a mechanistic model of immune activation—something prior ADC and bispecific-antibody models have not done, notes Dr. Jain.
“This let us simulate not only whole-body pharmacokinetics but also spatial gradients of drug and immune cells inside the tumor itself,” he says. “It also allowed us to isolate the contribution of each mechanism—checkpoint blockade, TLR7/8 stimulation, and their synergy—to overall efficacy.”
Uncovering a synergistic effect
Dr. Jain and his colleagues calibrated their mathematical model to published mouse tumor-growth and CD8+T cell data. The resulting model reproduced the published data across both tumor models and all four treatment groups (control, D18 payload alone, anti-PD-L1 alone, and HE-S2).
Importantly, the model helped explain the synergistic effect that makes HE-S2 more effective than either component individually:
- Anti-PD-L1 alone increases immune-mediated killing.
- D18 alone boosts inflammatory signaling.
- Each component reinforces the other’s action, with D18 also raising tumor sensitivity to checkpoint blockade.
The model also offered insight into dynamics that were not measured in the underlying experiments. For example, it explained why intratumoral exposure is higher for the antibody/ADC than for D18: The antibody/ADC is large enough to exploit leaky tumor vasculature and be excluded from normal tissue. Meanwhile, the smaller D18 clears rapidly, passing freely through both tumor and normal vessels.
The findings point to the potential value of combining ADCs with existing TME-normalizing strategies. Depending on the tumor’s vascular characteristics, these could include anti-VEGF agents such as bevacizumab or stress-alleviating drugs such as losartan or ketotifen. However, achieving normalization must be done carefully. Parametric analyses identified optimal TME conditions for maximizing ADC efficacy: tumor vessels with sufficient permeability (pore radius >40 nanometers) and abundance (vascular density >70 [1/cm]).
“You have to make sure the blood vessels still have large enough holes for the drug to get in. That’s key,” Dr. Jain says. “If you make them too small, then you can seal off the tumor from the drug.”
In terms of study limitations, the model did not account for the ADC surface charge and configuration, both of which can affect the transport properties. Furthermore, the authors assumed a uniform vessel wall pore radius despite a distribution of pore radii.
Calling for experimental testing of key predictions
According to Dr. Jain, next steps in this line of inquiry include experimentally testing the model’s key predictions—particularly regarding vessel pore–size and vascular-density thresholds for optimal ADC delivery—in preclinical models combining TME-normalizing agents with ADC treatment.
In addition, he would like to extend the model to evaluate dosing and safety tradeoffs for next-generation ADCs. The authors note D18 is a potent TLR7/8 agonist—and that another drug in this class, NJH-395, “was associated with dose-limiting and other immune-related toxicities” in a phase I study.
For now, Dr. Jain believes the team has established a broad framework that could be adapted to study other solid tumors and next-generation antibody therapies. And he is pleased that his long-held views on the significance of the TME are gaining wider acceptance.
“For a long time, people thought if you knew everything about cancer cells, you could cure the cancer,” he says. “They’re finding out that’s not the case. You need to address the many barriers posed by the neighborhood as well.”