Generalized Additive Mixed Modeling of Longitudinal Tumor Growth Reduces Bias and Improves Decision Making in Translational Oncology

Personalized chemotherapy selection for patients with triple-negative breast cancer using deep learning

Background

Potential uncertainties and overtreatment exist in adjuvant chemotherapy for triple-negative breast cancer (TNBC) patients.

Objectives

This study aims to explore the performance of deep learning (DL) models in personalized chemotherapy selection and quantify the impact of baseline characteristics on treatment efficacy.

Methods

Patients who received treatment recommended by models were compared to those who did not. Overall survival for treatment according to model recommendations was the primary outcome. To mitigate bias, inverse probability treatment weighting (IPTW) was employed. A mixed-effect multivariate linear regression was employed to visualize the influence of certain baseline features of patients on chemotherapy selection.

Results

A total of 10,070 female TNBC patients met the inclusion criteria. Treatment according to Self-Normalizing Balanced (SNB) individual treatment effect for survival data model recommendations was associated with a survival benefit (IPTW-adjusted hazard ratio: 0.53, 95% CI, 0.32-8.60; IPTW-adjusted risk difference: 12.90, 95% CI, 6.99-19.01; IPTW-adjusted the difference in restricted mean survival time: 5.54, 95% CI, 1.36-8.61), which surpassed other models and the National Comprehensive Cancer Network guidelines. No survival benefit for chemotherapy was seen for patients not recommended to receive this treatment. SNB predicted older patients with larger tumors and more positive lymph nodes are the optimal candidates for chemotherapy.

Conclusion

These findings suggest that the SNB model may identify patients with TNBC who could benefit from chemotherapy. This novel analytical approach may provide debiased individual survival information and treatment recommendations. Further research is required to validate these models in clinical settings with more features and outcome measurements.

The HER2-directed antibody-drug conjugate DHES0815A in advanced and/or metastatic breast cancer: preclinical characterization and phase 1 trial results

Approved antibody-drug conjugates (ADCs) for HER2-positive breast cancer include trastuzumab emtansine and trastuzumab deruxtecan. To develop a differentiated HER2 ADC, we chose an antibody that does not compete with trastuzumab or pertuzumab for binding, conjugated to a reduced potency PBD (pyrrolobenzodiazepine) dimer payload. PBDs are potent cytotoxic agents that alkylate and cross-link DNA. In our study, the PBD dimer is modified to alkylate, but not cross-link DNA. This HER2 ADC, DHES0815A, demonstrates in vivo efficacy in models of HER2-positive and HER2-low cancers and is well-tolerated in cynomolgus monkey safety studies. Mechanisms of action include induction of DNA damage and apoptosis, activity in non-dividing cells, and bystander activity. A dose-escalation study (ClinicalTrials.gov: NCT03451162) in patients with HER2-positive metastatic breast cancer, with the primary objective of evaluating the safety and tolerability of DHES0815A and secondary objectives of characterizing the pharmacokinetics, objective response rate, duration of response, and formation of anti-DHES0815A antibodies, is reported herein. Despite early signs of anti-tumor activity, patients at higher doses develop persistent, non-resolvable dermal, ocular, and pulmonary toxicities, which led to early termination of the phase 1 trial.

Leveraging homologous hypotheses for increased efficiency in tumor growth curve testing

In this note, we present an innovative approach called "homologous hypothesis tests" that focuses on cross-sectional comparisons of average tumor volumes at different time-points. By leveraging the correlation structure between time-points, our method enables highly efficient per time-point comparisons, providing inferences that are highly efficient as compared to those obtained from a standard two-sample t test. The key advantage of this approach lies in its user-friendliness and accessibility, as it can be easily employed by the broader scientific community through standard statistical software packages.

Combined PD-L1/TGFβ blockade allows expansion and differentiation of stem cell-like CD8 T cells in immune excluded tumors

TGFβ signaling is associated with non-response to immune checkpoint blockade in patients with advanced cancers, particularly in the immune-excluded phenotype. While previous work demonstrates that converting tumors from excluded to inflamed phenotypes requires attenuation of PD-L1 and TGFβ signaling, the underlying cellular mechanisms remain unclear. Here, we show that TGFβ and PD-L1 restrain intratumoral stem cell-like CD8 T cell (TSCL) expansion and replacement of progenitor-exhausted and dysfunctional CD8 T cells with non-exhausted T effector cells in the EMT6 tumor model in female mice. Upon combined TGFβ/PD-L1 blockade IFNγhi CD8 T effector cells show enhanced motility and accumulate in the tumor. Ensuing IFNγ signaling transforms myeloid, stromal, and tumor niches to yield an immune-supportive ecosystem. Blocking IFNγ abolishes the anti-PD-L1/anti-TGFβ therapy efficacy. Our data suggest that TGFβ works with PD-L1 to prevent TSCL expansion and replacement of exhausted CD8 T cells, thereby maintaining the T cell compartment in a dysfunctional state.

An allosteric pan-TEAD inhibitor blocks oncogenic YAP/TAZ signaling and overcomes KRAS G12C inhibitor resistance

The Hippo pathway is a key growth control pathway that is conserved across species. The downstream effectors of the Hippo pathway, YAP (Yes-associated protein) and TAZ (transcriptional coactivator with PDZ-binding motif), are frequently activated in cancers to drive proliferation and survival. Based on the premise that sustained interactions between YAP/TAZ and TEADs (transcriptional enhanced associate domain) are central to their transcriptional activities, we discovered a potent small-molecule inhibitor (SMI), GNE-7883, that allosterically blocks the interactions between YAP/TAZ and all human TEAD paralogs through binding to the TEAD lipid pocket. GNE-7883 effectively reduces chromatin accessibility specifically at TEAD motifs, suppresses cell proliferation in a variety of cell line models and achieves strong antitumor efficacy in vivo. Furthermore, we uncovered that GNE-7883 effectively overcomes both intrinsic and acquired resistance to KRAS (Kirsten rat sarcoma viral oncogene homolog) G12C inhibitors in diverse preclinical models through the inhibition of YAP/TAZ activation. Taken together, this work demonstrates the activities of TEAD SMIs in YAP/TAZ-dependent cancers and highlights their potential broad applications in precision oncology and therapy resistance.

Systematic pharmacological analysis of agonistic and antagonistic fibroblast growth factor receptor 1 MAbs reveals a similar unique mode of action

Fibroblast growth factor receptor 1 (FGFR1) is a receptor tyrosine kinase that plays a major role in developmental processes and metabolism. The dysregulation of FGFR1 through genetic aberrations leads to skeletal and metabolic diseases as well as cancer. For this reason, FGFR1 is a promising therapeutic target, yet a very challenging one due to potential on-target toxicity. More puzzling is that both agonistic and antagonistic FGFR1 antibodies are reported to exhibit similar toxicity profiles in vivo, namely weight loss. In this study, we aimed to assess and compare the mechanism of action of these molecules to better understand this apparent contradiction. By systematically comparing the binding of these antibodies and the activation or the inhibition of the major FGFR1 signaling events, we demonstrated that the molecules displayed similar properties and can behave either as an agonist or antagonist depending on the presence or the absence of the endogenous ligand. We further demonstrated that these findings translated in xenografts mice models. In addition, using time-resolved FRET and mass spectrometry analysis, we showed a functionally distinct FGFR1 active conformation in the presence of an antibody that preferentially activates the FGFR substrate 2 (FRS2)-dependent signaling pathway, demonstrating that modulating the geometry of a FGFR1 dimer can effectively change the signaling outputs and ultimately the activity of the molecule in preclinical studies. Altogether, our results highlighted how bivalent antibodies can exhibit both agonistic and antagonistic activities and have implications for targeting other receptor tyrosine kinases with antibodies.

Selective PROTAC-mediated degradation of SMARCA2 is efficacious in SMARCA4 mutant cancers

The mammalian SWItch/Sucrose Non-Fermentable (SWI/SNF) helicase SMARCA4 is frequently mutated in cancer and inactivation results in a cellular dependence on its paralog, SMARCA2, thus making SMARCA2 an attractive synthetic lethal target. However, published data indicates that achieving a high degree of selective SMARCA2 inhibition is likely essential to afford an acceptable therapeutic index, and realizing this objective is challenging due to the homology with the SMARCA4 paralog. Herein we report the discovery of a potent and selective SMARCA2 proteolysis-targeting chimera molecule (PROTAC), A947. Selective SMARCA2 degradation is achieved in the absence of selective SMARCA2/4 PROTAC binding and translates to potent in vitro growth inhibition and in vivo efficacy in SMARCA4 mutant models, compared to wild type models. Global ubiquitin mapping and proteome profiling reveal no unexpected off-target degradation related to A947 treatment. Our study thus highlights the ability to transform a non-selective SMARCA2/4-binding ligand into a selective and efficacious in vivo SMARCA2-targeting PROTAC, and thereby provides a potential new therapeutic opportunity for patients whose tumors contain SMARCA4 mutations.

Giredestrant reverses progesterone hypersensitivity driven by estrogen receptor mutations in breast cancer

ESR1 (estrogen receptor 1) hotspot mutations are major contributors to therapeutic resistance in estrogen receptor-positive (ER⁺) breast cancer. Such mutations confer estrogen independence to ERα, providing a selective advantage in the presence of estrogen-depleting aromatase inhibitors. In addition, ESR1 mutations reduce the potency of tamoxifen and fulvestrant, therapies that bind ERα directly. These limitations, together with additional liabilities, inspired the development of the next generation of ERα-targeted therapeutics, of which giredestrant is a high-potential candidate. Here, we generated Esr1 mutant-expressing mammary gland models and leveraged patient-derived xenografts (PDXs) to investigate the biological properties of the ESR1 mutations and their sensitivity to giredestrant in vivo. In the mouse mammary gland, Esr1 mutations promote hypersensitivity to progesterone, triggering pregnancy-like tissue remodeling and profoundly elevated proliferation. These effects were driven by an altered progesterone transcriptional response and underpinned by gained sites of ERα-PR (progesterone receptor) cobinding at the promoter regions of pro-proliferation genes. PDX experiments showed that the mutant ERα-PR proliferative program is also relevant in human cancer cells. Giredestrant suppressed the mutant ERα-PR proliferation in the mammary gland more so than the standard-of-care agents, tamoxifen and fulvestrant. Giredestrant was also efficacious against the progesterone-stimulated growth of ESR1 mutant PDX models. In addition, giredestrant demonstrated activity against a molecularly characterized ESR1 mutant tumor from a patient enrolled in a phase 1 clinical trial. Together, these data suggest that mutant ERα can collaborate with PR to drive protumorigenic proliferation but remain sensitive to inhibition by giredestrant.

Growth‐rate model predicts in vivo tumor response from in vitro data

A major challenge in oncology drug development is to elucidate why drugs that show promising results in cancer cell lines in vitro fail in mouse studies or human trials. One of the fundamental steps toward solving this problem is to better predict how in vitro potency translates into in vivo efficacy. A common approach to infer whether a model will respond in vivo is based on in vitro half‐maximal inhibitory concentration values (IC₅₀), but yields limited quantitative comparison between cell lines and drugs, potentially because cell division and death rates differ between cell lines and in vivo models. Other methods based either on mechanistic modeling or machine learning require molecular insights or extensive training data, limiting their use for early drug development. To address these challenges, we propose a mathematical model integrating in vitro growth rate inhibition values with pharmacokinetic parameters to estimate in vivo drug response. Upon calibration with a drug‐specific factor, our model yields precise estimates of tumor growth rate inhibition for in vivo studies based on in vitro data. We then demonstrate how our model can be used to study dosing schedules and perform sensitivity analyses. In addition, it provides meaningful metrics to assess association with genotypes and guide clinical trial design. By relying on commonly collected data, our approach shows great promise for optimizing drug development, better characterizing the efficacy of novel molecules targeting proliferation, and identifying more robust biomarkers of sensitivity while limiting the number of in vivo experiments.

Distinct resistance mechanisms arise to allosteric vs. ATP-competitive AKT inhibitors

The AKT kinases have emerged as promising therapeutic targets in oncology and both allosteric and ATP-competitive AKT inhibitors have entered clinical investigation. However, long-term efficacy of such inhibitors will likely be challenged by the development of resistance. We have established prostate cancer models of acquired resistance to the allosteric inhibitor MK-2206 or the ATP-competitive inhibitor ipatasertib following prolonged exposure. While alterations in AKT are associated with acquired resistance to MK-2206, ipatasertib resistance is driven by rewired compensatory activity of parallel signaling pathways. Importantly, MK-2206 resistance can be overcome by treatment with ipatasertib, while ipatasertib resistance can be reversed by co-treatment with inhibitors of pathways including PIM signaling. These findings demonstrate that distinct resistance mechanisms arise to the two classes of AKT inhibitors and that combination approaches may reverse resistance to ATP-competitive inhibition.

How resistance to different classes of AKT inhibitors can emerge is unclear. Here, the authors show that resistance to allosteric inhibitors is mainly due to mutation of AKT1 while the ATP competitive resistance is driven by activation of PIM kinases in prostate cancer models.