Advancing toward precision migraine treatment: Predicting responses to preventive medications with machine learning models based on patient and migraine features.

OBJECTIVE: To develop machine learning models using patient and migraine features that can predict treatment responses to commonly used migraine preventive medications. BACKGROUND: Currently, there is no accurate way to predict response to migraine preventive medications, and the standard trial-and-error approach is inefficient. METHODS: In this cohort study, we analyzed data from the Mayo Clinic Headache database prospectively collected from 2001 to December 2023. Adult patients with migraine completed questionnaires during their initial headache consultation to record detailed clinical features and then at each follow-up to track preventive medication changes and monthly headache days. We included patients treated with at least one of the following migraine preventive medications: topiramate, beta-blockers (propranolol, metoprolol, atenolol, nadolol, timolol), tricyclic antidepressants (amitriptyline, nortriptyline), verapamil, gabapentin, onabotulinumtoxinA, and calcitonin gene-related peptide (CGRP) monoclonal antibodies (mAbs) (erenumab, fremanezumab, galcanezumab, eptinezumab). We pre-trained a deep neural network, "TabNet," using 145 variables, then employed TabNet-embedded data to construct prediction models for each medication to predict binary outcomes (responder vs. non-responder). A treatment responder was defined as having at least a 30% reduction in monthly headache days from baseline. All model performances were evaluated, and metrics were reported in the held-out test set (train 85%, test 15%). SHapley Additive exPlanations (SHAP) were conducted to determine variable importance. RESULTS: Our final analysis included 4260 patients. The responder rate for each medication ranged from 28.7% to 34.9%, and the mean time to treatment outcome for each medication ranged from 151.3 to 209.5 days. The CGRP mAb prediction model achieved a high area under the receiver operating characteristics curve (AUC) of 0.825 (95% confidence interval [CI] 0.726, 0.920) and an accuracy of 0.80 (95% CI 0.70, 0.88). The AUCs of prediction models for beta-blockers, tricyclic antidepressants, topiramate, verapamil, gabapentin, and onabotulinumtoxinA were: 0.664 (95% CI 0.579, 0.745), 0.611 (95% CI 0.562, 0.682), 0.605 (95% CI 0.520, 0.688), 0.673 (95% CI 0.569, 0.724), 0.628 (0.533, 0.661), and 0.581 (95% CI 0.550, 0.632), respectively. Baseline monthly headache days, age, body mass index (BMI), duration of migraine attacks, responses to previous medication trials, cranial autonomic symptoms, family history of headache, and migraine attack triggers were among the most important variables across all models. A variable could have different contributions; for example, lower BMI predicts responsiveness to CGRP mAbs and beta-blockers, while higher BMI predicts responsiveness to onabotulinumtoxinA, topiramate, and gabapentin. CONCLUSION: We developed an accurate prediction model for CGRP mAbs treatment response, leveraging detailed migraine features gathered from a headache questionnaire before starting treatment. Employing the same methods, the model performances for other medications were less impressive, though similar to the machine learning models reported in the literature for other diseases. This may be due to CGRP mAbs being migraine-specific. Incorporating medical comorbidities, genomic, and imaging factors might enhance the model performance. We demonstrated that migraine characteristics are important in predicting treatment responses and identified the most crucial predictors for each of the seven types of preventive medications. Our results suggest that precision migraine treatment is feasible.

Trastuzumab-MMAU Antibody-Auristatin Conjugates: Valine-Glucoserine Linker with Stabilized Maleimide Conjugation Improves In Vivo Efficacy and Tolerability.

PURPOSE: Antibody-drug conjugates (ADCs) have shown impressive clinical activity with approval of many agents in hematological and solid tumors. However, challenges remain with both efficacy and safety of ADCs. This study describes novel trastuzumab-auristatin conjugates with the hydrophilic MMAE prodrug MMAU, and optimization of a glycopeptide linker leading to a wider therapeutic window. EXPERIMENTAL DESIGN: Trastuzumab was conjugated with auristatin payloads via a series of linkers using a stabilized maleimide handle. The ADCs were characterized in vitro and their relative in vivo anti-tumor efficacies were assessed in HER2+ xenograft models. Relative linker stabilities and the mechanism of linker cleavage were studied using in vitro assays. Toxicity and toxicokinetics of the best performing ADC were evaluated in cynomolgus monkey (cyno). RESULTS: The trastuzumab-MMAU ADC with stabilized glycopeptide linker showed maleimide stabilization and higher resistance to cleavage by serum and lysosomal enzymes compared to a valine-citrulline conjugated trastuzumab ADC (trastuzumab-vc-MMAE). A single dose of 1 or 2 mg/kg of trastuzumab-MMAU at drug-to-antibody ratios (DAR) of 8 and 4 respectively resulted in xenograft tumor growth inhibition, with superior efficacy to trastuzumab-vc-MMAE. Trastuzumab-MMAU DAR4 was tolerated at doses up to 12 mg/kg in cyno, which represents 2- to 4-fold higher dose than that observed with vedotin ADCs, and had increased terminal half-life and exposure. CONCLUSIONS: The optimized trastuzumab-MMAU ADC showed potent antitumor activity and was well tolerated with excellent pharmacokinetics in non-human primates, leading to a superior preclinical therapeutic window. The data supports potential utility of trastuzumab-MMAU for treatment of HER2+ tumors.

Plasma pharmacokinetics and distribution of ruxolitinib into skin following oral and topical administration in minipigs.

This preclinical study compared plasma concentrations and distribution of ruxolitinib in the skin of Göttingen minipigs following twice a day oral (40 mg/kg) versus topical administration (1.5% w/w cream applied to 10% of body surface area). Following oral administration, the plasma area-under-the-curve was approximately 31-fold and maximum drug concentration was 38-fold higher than those observed following topical application. Following ruxolitinib cream application, the average plasma concentration at steady-state was 2.7 ± 1.8 nM, a concentration that is not pharmacologically relevant. The average total dermis concentration of ruxolitinib at steady-state after topical administration was 507-fold higher versus that following oral dosing, while the ratio for the total epidermal concentration following topical vs oral dosing was 1989-fold. The concentration of unbound ruxolitinib in the dermis after topical application was predicted to result in sustained and near-complete inhibition of Janus kinase/signal transducer and activator of transcription proteins (JAK/STAT) signaling in this tissue. In contrast, only partial inhibition of downstream signaling was predicted to occur after oral dosing. In conclusion, ruxolitinib cream affords an attractive disposition profile in minipigs, wherein dermis concentrations of ruxolitinib are fully effective whereas corresponding plasma concentrations are negligible. Consequently, this distribution profile should maximize the efficacy of ruxolitinib cream in the skin while minimizing the potential for deleterious systemic effects.

INCB054828 (pemigatinib), a potent and selective inhibitor of fibroblast growth factor receptors 1, 2, and 3, displays activity against genetically defined tumor models.

Alterations in fibroblast growth factor receptor (FGFR) genes have been identified as potential driver oncogenes. Pharmacological targeting of FGFRs may therefore provide therapeutic benefit to selected cancer patients, and proof-of-concept has been established in early clinical trials of FGFR inhibitors. Here, we present the molecular structure and preclinical characterization of INCB054828 (pemigatinib), a novel, selective inhibitor of FGFR 1, 2, and 3, currently in phase 2 clinical trials. INCB054828 pharmacokinetics and pharmacodynamics were investigated using cell lines and tumor models, and the antitumor effect of oral INCB054828 was investigated using xenograft tumor models with genetic alterations in FGFR1, 2, or 3. Enzymatic assays with recombinant human FGFR kinases showed potent inhibition of FGFR1, 2, and 3 by INCB054828 (half maximal inhibitory concentration [IC50] 0.4, 0.5, and 1.0 nM, respectively) with weaker activity against FGFR4 (IC50 30 nM). INCB054828 selectively inhibited growth of tumor cell lines with activation of FGFR signaling compared with cell lines lacking FGFR aberrations. The preclinical pharmacokinetic profile suggests target inhibition is achievable by INCB054828 in vivo with low oral doses. INCB054828 suppressed the growth of xenografted tumor models with FGFR1, 2, or 3 alterations as monotherapy, and the combination of INCB054828 with cisplatin provided significant benefit over either single agent, with an acceptable tolerability. The preclinical data presented for INCB054828, together with preliminary clinical observations, support continued investigation in patients with FGFR alterations, such as fusions and activating mutations.

Preclinical characterization of itacitinib (INCB039110), a novel selective inhibitor of JAK1, for the treatment of inflammatory diseases.

Pharmacological modulation of the Janus kinase (JAK) family has achieved clinically meaningful therapeutic outcomes for the treatment of inflammatory and hematopoietic diseases. Several JAK1 selective compounds are being investigated clinically to determine their anti-inflammatory potential. We used recombinant enzymes and primary human lymphocytes to assess the JAK1 specificity of itacitinib (INCB039110) and study inhibition of signal transducers and activators of transcription (STAT) signaling. Rodent models of arthritis and inflammatory bowel disease were subsequently explored to elucidate the efficacy of orally administered itacitinib on inflammatory pathogenesis. Itacitinib is a potent and selective JAK1 inhibitor when profiled against the other JAK family members. Upon oral administration in rodents, itacitinib achieved dose-dependent pharmacokinetic exposures that highly correlated with STAT3 pharmacodynamic pathway inhibition. Itacitinib ameliorated symptoms and pathology of established experimentally-induced arthritis in a dose-dependent manner. Furthermore, itacitinib effectively delayed disease onset, reduced symptom severity, and accelerated recovery in three distinct mouse models of inflammatory bowel disease. Low dose itacitinib administered via cannula directly into the colon was highly efficacious in TNBS-induced colitis but with minimal systemic drug exposure, suggesting localized JAK1 inhibition is sufficient for disease amelioration. Itacitinib treatment in an acute graft-versus-host disease (GvHD) model rapidly reduced inflammatory markers within lymphocytes and target tissue, resulting in a marked improvement in disease symptoms. This is the first manuscript describing itacitinib as a potent and selective JAK1 inhibitor with anti-inflammatory activity across multiple preclinical disease models. These data support the scientific rationale for ongoing clinical trials studying itacitinib in select GvHD patient populations.

Preclinical characterization of INCB053914, a novel pan-PIM kinase inhibitor, alone and in combination with anticancer agents, in models of hematologic malignancies.

The Proviral Integration site of Moloney murine leukemia virus (PIM) serine/threonine protein kinases are overexpressed in many hematologic and solid tumor malignancies and play central roles in intracellular signaling networks important in tumorigenesis, including the Janus kinase-signal transducer and activator of transcription (JAK/STAT) and phosphatidylinositol 3-kinase (PI3K)/AKT pathways. The three PIM kinase isozymes (PIM1, PIM2, and PIM3) share similar downstream substrates with other key oncogenic kinases and have differing but mutually compensatory functions across tumors. This supports the therapeutic potential of pan-PIM kinase inhibitors, especially in combination with other anticancer agents chosen based on their role in overlapping signaling networks. Reported here is a preclinical characterization of INCB053914, a novel, potent, and selective adenosine triphosphate-competitive pan-PIM kinase inhibitor. In vitro, INCB053914 inhibited proliferation and the phosphorylation of downstream substrates in cell lines from multiple hematologic malignancies. Effects were confirmed in primary bone marrow blasts from patients with acute myeloid leukemia treated ex vivo and in blood samples from patients receiving INCB053914 in an ongoing phase 1 dose-escalation study. In vivo, single-agent INCB053914 inhibited Bcl-2-associated death promoter protein phosphorylation and dose-dependently inhibited tumor growth in acute myeloid leukemia and multiple myeloma xenografts. Additive or synergistic inhibition of tumor growth was observed when INCB053914 was combined with selective PI3Kδ inhibition, selective JAK1 or JAK1/2 inhibition, or cytarabine. Based on these data, pan-PIM kinase inhibitors, including INCB053914, may have therapeutic utility in hematologic malignancies when combined with other inhibitors of oncogenic kinases or standard chemotherapeutics.