Biology subjects

Mastri, M.

Publications and source records attributed to Mastri, M..

6 recordsLinked to original sources

bioRxiv2023

Machine-learning and mechanistic modeling of primary and metastatic breast cancer growth after neoadjuvant targeted therapy

Clinical trials involving systemic neoadjuvant treatments in breast cancer aim to shrink tumors prior to surgery while simultaneously allowing for controlled evaluation of biomarkers, toxicity, and suppression of distant (occult) metastatic disease. Yet such trials are rarely preceded by preclinical testing involving surgery. Here we used a mouse model of spontaneous metastasis after surgical removal to develop a predictive mathematical model of neoadjuvant treatment response to sunitinib, a receptor tyrosine kinase inhibitor (RTKI). Longitudinal data consisted of measurements of presurgical primary tumor size and postsurgical metastatic burden in 128 mice (104 for model training, 24 for validation), following variable neoadjuvant treatment schedules over a 14-day period. A nonlinear mixed-effects modeling approach was used to quantify inter-animal variability. Machine learning algorithms were applied to investigate the significance of several biomarkers at resection as predictors of individual kinetics. Biomarkers included circulating tumor- and immune-based cells (circulating tumor cells and myeloid-derived suppressor cells) as well as immunohistochemical tumor proteins (CD31 and Ki67). Our simulations showed that neoadjuvant RTKI treatment inhibits primary tumor growth but has little efficacy in preventing (micro)-metastatic disease progression after surgery. Surprisingly, machine-learning algorithms demonstrated only limited predictive power of tested biomarkers on the mathematical parameters. These results suggest that presurgical modeling might be an effective tool to screen biomarkers prior to clinical trial testing. Mathematical modeling combined with artificial intelligence techniques represent a novel platform for integrating preclinical surgical metastasis models in outcome prediction of neoadjuvant treatment. Major findingsUsing simulations from a mechanistic mathematical model compared with preclinical data from surgical metastasis models, we found anti-tumor effects of neoadjuvant RTKI treatment can differ between the primary tumor and metastases in the perioperative setting. Model simulations with variable drug doses and scheduling of neoadjuvant treatment revealed a contrasting impact on initial primary tumor debulking and metastatic outcomes long after treatment has stopped and tumor surgically removed. Using machine-learning algorithms, we identified the limited power of several circulating cellular and molecular biomarkers in predicting metastatic outcome, uncovering a potential fast-track strategy for assessing future clinical biomarkers by paring patient studies with identical studies in mice.

cancer biology↗

bioRxiv2022

Syngeneic model of carcinogen-induced tumor mimics basal/squamous, stromal-rich, and neuroendocrine molecular and immunological features of muscle-invasive bladder cancer

BackgroundBladder cancer is a heterogenous disease and the emerging knowledge on molecular classification of bladder tumors could have impact to drive treatment decisions based on molecular subtype. Pre-clinical models representing each subtype are needed to test novel therapies. Carcinogen-induced bladder cancer models represent heterogeneous, immune-competent, pre-clinical testing options with many features found in the human disease. MethodsInvasive bladder tumors were induced in C57BL/6 mice when continuously exposed to N-butyl-N-(4-hydroxbutyl nitrosamine) (BBN) in the drinking water. Tumors were excised and serially passed by subcutaneous implantation into sex-matched syngeneic C57BL/6 hosts. Eight tumor lines were developed and named BBN-induced Urothelium Roswell Park (BURP) tumor lines. The BURP lines were characterized by applying consensus molecular classification to RNA expression, histopathology, and immune profiles by CIBERSORT. Two lines were further characterized for cisplatin response. ResultsEight BURP tumor lines were established with 3 male and 3 female BURP tumor lines, having the basal/squamous (BaSq) molecular phenotype and morphology. BURP-16SR was established from a male mouse and has a stromal-rich (SR) molecular phenotype and a sarcomatoid carcinoma morphology. BURP-19NE was established from a male mouse and has a neuroendocrine (NE)-like molecular phenotype and poorly differentiated morphology. The established BURP tumor lines have unique immune profiles with fewer immune infiltrates compared to their originating BBN-induced tumors. The immune profiles of the BURP tumor lines capture some of the features observed in the molecular classifications of human bladder cancer. BURP-16SR growth was inhibited by cisplatin treatment, while BURP-24BaSq did not respond to cisplatin. ConclusionsThe BURP tumor lines represent several molecular classifications, including basal/squamous, stroma-rich, and NE-like. The stroma-rich (BURP-16SR) and NE-like (BURP-19NE) represent unique immunocompetent models that can be used to test novel treatments in these less common bladder cancer subtypes. Six basal/squamous tumor lines were established from both male and female mice. Overall, the BURP tumor lines have less heterogeneity than the carcinogen-induced tumors and can be used to evaluate treatment response without the confounding mixed response often observed in heterogeneous tumors. Additionally, basal/squamous tumor lines were established and maintained in both male and female mice, thereby allowing these tumor lines to be used to compare differential treatment responses between sexes.

cancer biology↗

bioRxiv2021

Glucocorticoid signaling delays castration-induced regression in murine models of prostate cancer

Androgen deprivation therapy (ADT) induces regression of recurrent and advanced prostate cancer (PrCa), but many tumors recur. To understand the response to ADT, changes in tumor volume were imaged after castration of murine PrCa models. While mouse (non-tumor) prostate begins to regress within two days of castration, murine PrCa regresses after a delay of 3-14 days in two distinct mouse models. Intra-tumoral androgens are undetectable after castration, but tumor cells proliferate during this period. Intratumoral glucocorticoids and glucocorticoid receptor (GR) protein increase, as does GR mRNA and a set of GR-regulated genes specifically in tumor epithelial cells identified using scRNAseq. A selective GR antagonist (CORT125281, relacorilant), in clinical trials for late-state PrCa, eliminates the delayed regression phenotype in both models. Thus, activated GR signaling and murine tumor proliferation following castration resembles the GR-dependent escape mechanism of castrate resistant PrCa. These results suggest simultaneous inhibition of GR and androgen receptor signaling could improve PrCa therapy. In briefAndrogen deprivation therapy for high risk and recurrent prostate cancers is initially effective, but ultimately fails; better understanding the mechanisms should improve therapy. In two murine prostate cancer models, GR signaling is activated immediately following castration, substituting for the acute reduction in AR signaling, and allowing for continued tumor growth. This continued growth is blocked by relacorilant, selective GR antagonist in clinical trials for late-state PrCa. HighlightsO_LIAndrogen deprivation therapy induces regression of prostate cancer, but tumors recur C_LIO_LIMurine PrCa continues to proliferate for 3-14 days in two distinct mouse prostate cancer models C_LIO_LITumor cells proliferate during this period, and intratumoral glucocorticoids and glucocorticoid receptor (GR) protein increase, as does GR mRNA and a set of GR-regulated genes C_LIO_LIRelacorilant, a selective GR antagonist in clinical trials for late-state PrCa, eliminates the delayed regression C_LI

cancer biology↗

bioRxiv2021

Acquired resistance to PD-L1 inhibition is associated with an enhanced type I IFN-stimulated secretory program in tumor cells

BackgroundInterferon (IFN) pathway activation in tumors can have dual, sometimes opposing, influences on immune responses. Therapeutic inhibition of programmed cell death ligand (PD-L1) - a treatment that reverses PD-1-mediated suppression of tumor-killing T-cells - is linked to alterations in IFN signaling; however, less is known about the role of IFNs after treatment resistance. Since IFN-regulated intracellular signaling can control extracellular secretory programs in tumors to modulate immunity, we examined the consequences of PD-L1 blockade on IFN-related secretory changes in preclinical models of acquired resistance. MethodsTherapy-resistant cell variants were derived from orthotopically grown mouse tumors initially sensitive or insensitive to PD-L1 antibody treatment. Cells representing acquired resistance were analyzed for changes to IFN-regulated secretory machinery that could impact tumor progression. ResultsWe identified a PD-L1 treatment-induced secretome (PTIS) that was enriched for several IFN-stimulated genes (ISGs) and significantly enhanced when stimulated by type I IFNs (IFN or IFN{beta}). Secretory changes were specific to treatment-sensitive tumor models and found to suppress activation of T cells ex vivo while diminishing tumor cell cytotoxicity, revealing a tumor-intrinsic treatment adaptation with potentially broad tumor-extrinsic effects. When reimplanted in vivo, resistant tumor growth was slowed by the blockade of individual secreted PTIS components (such as IL6) and stopped altogether by a more generalized disruption of type I IFN signaling. In vitro, genetic or therapeutic methods to target PD-L1 could only partially recapitulate the IFN-enhanced PTIS phenotype, showing that in vivo-based systems with intact tumor:immune cell interactions are needed to faithfully mimic acquired resistance as it occurs in patients. ConclusionsThese results suggest that prolonged in vivo PD-L1 inhibition can rewire type I IFN signaling to drive secretory programs that help protect tumors from immune cell attack and represent a targetable vulnerability to overcome acquired resistance in patients.

cancer biology↗

bioRxiv2021

Patient derived models of bladder cancer amplify tumor specific gene expression compared to surgical specimen while maintaining gene expression of molecular subtype and epithelial mesenchymal transition markers

Patient derived models (PDMs) are a powerful tool to study preclinical responses. However, the benefits of each model have not been compared head-to-head when models are derived from the same surgical specimen. PDMs derived from surgical specimens were established as xenografts (PDX), organoids (PDO), and spheroids (PDS). PDMs were molecularly characterized by RNA sequencing. Differential gene expression was determined between the PDMs and surgical specimens. Surgical specimens had the most differentially expressed genes reflecting loss of immune and stromal compartments in PDMs. PDMs and surgical specimens were clustered using the Euclidian distance analysis to test model fidelity. PDMs upregulated a clear, patient-specific bladder cancer signal. Overall, the molecular profiles of PDXs were the most similar to the matching patient surgical specimen than the PDO and PDS from that patient. The epithelial mesenchymal transition (EMT) gene expression profile is maintained in the PDMs showing the persistence of EMT in both in vivo and in vitro model setting. The consensus molecular subtype was determined in order to compare PDMs to each other and their matching surgical specimen, and only surgical specimens with Basal/Squamous or Luminal Papillary molecular subtype established PDMs. Patient derived models reduce tumor heterogeneity and allow analysis of specific tumor compartments while maintaining the gene expression profile representative of the original tumor.

cancer biology↗

bioRxiv2019

A reduced Gompertz model for predicting tumor age using a population approach

Tumor growth curves are classically modeled by ordinary differential equations. In analyzing the Gompertz model several studies have reported a striking correlation between the two parameters of the model.\n\nWe analyzed tumor growth kinetics within the statistical framework of nonlinear mixed-effects (population approach). This allowed for the simultaneous modeling of tumor dynamics and interanimal variability. Experimental data comprised three animal models of breast and lung cancers, with 843 measurements in 94 animals. Candidate models of tumor growth included the Exponential, Logistic and Gompertz. The Exponential and - more notably - Logistic models failed to describe the experimental data whereas the Gompertz model generated very good fits. The population-level correlation between the Gompertz parameters was further confirmed in our analysis (R2 > 0.96 in all groups). Combining this structural correlation with rigorous population parameter estimation, we propose a novel reduced Gompertz function consisting of a single individual parameter. Leveraging the population approach using bayesian inference, we estimated the time of tumor initiation using three late measurement timepoints. The reduced Gompertz model was found to exhibit the best results, with drastic improvements when using bayesian inference as compared to likelihood maximization alone, for both accuracy and precision. Specifically, mean accuracy was 12.1% versus 74.1% and mean precision was 15.2 days versus 186 days, for the breast cancer cell line.\n\nThese results offer promising clinical perspectives for the personalized prediction of tumor age from limited data at diagnosis. In turn, such predictions could be helpful for assessing the extent of invisible metastasis at the time of diagnosis.\n\nAuthor summaryMathematical models for tumor growth kinetics have been widely used since several decades but mostly fitted to individual or average growth curves. Here we compared three classical models (Exponential, Logistic and Gompertz) using a population approach, which accounts for inter-animal variability. The Exponential and the Logistic models failed to fit the experimental data while the Gompertz model showed excellent descriptive power. Moreover, the strong correlation between the two parameters of the Gompertz equation motivated a simplification of the model, the reduced Gompertz model, with a single individual parameter and equal descriptive power. Combining the mixed-effects approach with Bayesian inference, we predicted the age of individual tumors with only few late measurements. Thanks to its simplicity, the reduced Gompertz model showed superior predictive power. Although our method remains to be extended to clinical data, these results are promising for the personalized estimation of the age of a tumor from limited measurements at diagnosis. Such predictions could contribute to the development of computational models for metastasis.

cancer biology↗