A cell-level discriminative neural network model for diagnosis of blood cancers.

MOTIVATION: Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes. RESULTS: We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes sample-level training data and predicts the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations. AVAILABILITY AND IMPLEMENTATION: The source code of CSNN and datasets used in the experiments are publicly available on GitHub (http://github.com/erobl/csnn). Raw FCS files can be downloaded from FlowRepository (ID: FR-FCM-Z6YK).

cis-B7:CD28 interactions at invaginated synaptic membranes provide CD28 co-stimulation and promote CD8+ T cell function and anti-tumor immunity.

B7 ligands (CD80 and CD86), expressed by professional antigen-presenting cells (APCs), activate the main co-stimulatory receptor CD28 on T cells in trans. However, in peripheral tissues, APCs expressing B7 ligands are relatively scarce. This raises the questions of whether and how CD28 co-stimulation occurs in peripheral tissues. Here, we report that CD8+ T cells displayed B7 ligands that interacted with CD28 in cis at membrane invaginations of the immunological synapse as a result of membrane remodeling driven by phosphoinositide-3-kinase (PI3K) and sorting-nexin-9 (SNX9). cis-B7:CD28 interactions triggered CD28 signaling through protein kinase C theta (PKCθ) and promoted CD8+ T cell survival, migration, and cytokine production. In mouse tumor models, loss of T cell-intrinsic cis-B7:CD28 interactions decreased intratumoral T cells and accelerated tumor growth. Thus, B7 ligands on CD8+ T cells can evoke cell-autonomous CD28 co-stimulation in cis in peripheral tissues, suggesting cis-signaling as a general mechanism for boosting T cell functionality.

Dual receptor T cells mediate effective antitumor immune responses via increased recognition of tumor antigens.

BACKGROUND: Discovery that ~16% of T cells naturally co-express two T-cell receptor (TCR) clonotypes prompts examining the role of dual TCR cells in immune functions. METHODS: Using TCRα-reporter transgenic mice, enabling unambiguous identification of single-TCR and dual-TCR cells, we tested the role of dual TCR cells in antitumor immune responses against immune-responsive syngeneic 6727 sarcoma and immune-resistant B16F10 melanoma. RESULTS: Dual TCR cells were specifically increased among tumor-infiltrating lymphocytes (TILs) in both models, indicating selective advantage in antitumor responses. Phenotype and single-cell gene expression analyses identified dual TCR are predominant during the effective antitumor response, demonstrating selectively increased activation in the TIL compartment and skewing toward an effector memory phenotype. Absence of dual TCR cells impaired immune response to B16F10 but not 6727, suggesting that dual TCR cells may be more influential in responses against poorly immunogenic tumors. Dual TCR cells demonstrated an advantage in recognition of B16F10-derived neoantigens in vitro, providing a mechanistic basis for their antitumor reactivity. CONCLUSIONS: These results discover an unrecognized role for dual TCR cells in protective immune function and identify these cells and their TCRs as a potential resource for antitumor immunotherapy.

CTLA4 depletes T cell endogenous and trogocytosed B7 ligands via cis-endocytosis.

CD28 and CTLA4 are T cell coreceptors that competitively engage B7 ligands CD80 and CD86 to control adaptive immune responses. While the role of CTLA4 in restraining CD28 costimulatory signaling is well-established, the mechanism has remained unclear. Here, we report that human T cells acquire antigen-presenting-cell (APC)-derived B7 ligands and major histocompatibility complex (MHC) via trogocytosis through CD28:B7 binding. Acquired MHC and B7 enabled T cells to autostimulate, and this process was limited cell-intrinsically by CTLA4, which depletes B7 ligands trogocytosed or endogenously expressed by T cells through cis-endocytosis. Extending this model to the previously proposed extrinsic function of CTLA4 in human regulatory T cells (Treg), we show that blockade of either CD28 or CTLA4 attenuates Treg-mediated depletion of APC B7, indicating that trogocytosis and CTLA4-mediated cis-endocytosis work together to deplete B7 from APCs. Our study establishes CTLA4 as a cell-intrinsic molecular sink that limits B7 availability on the surface of T cells, with implications for CTLA4-targeted therapy.

A cell-level discriminative neural network model for diagnosis of blood cancers.

Motivation: Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes. Results: We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes the available sample-level training data and predicts both the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations. Availability: The source code of CSNN and datasets used in the experiments are publicly available on GitHub and FlowRepository. Contact: Edgar E. Robles: roblesee@uci.edu and Yu Qian: mqian@jcvi.org. Supplementary information: Supplementary data are available on GitHub and at Bioinformatics online.

SnapShot: Cancer immunoediting.

In the tumor microenvironment, immune cells and tumor cells interact in a process called cancer immunoediting, giving rise to changes in gene expression, metabolism, mutational burden, and cellularity in the tumor. This SnapShot compares endogenous versus therapy-induced cancer immunoediting and outlines the molecular and cellular characteristics of interactions that result in complete tumor regression versus tumor escape and progression. To view this SnapShot, open or download the PDF.

Monomethyl auristatin antibody and peptide drug conjugates for trimodal cancer chemo-radio-immunotherapy.

Locally advanced cancers remain therapeutically challenging to eradicate. The most successful treatments continue to combine decades old non-targeted chemotherapies with radiotherapy that unfortunately increase normal tissue damage in the irradiated field and have systemic toxicities precluding further treatment intensification. Therefore, alternative molecularly guided systemic therapies are needed to improve patient outcomes when applied with radiotherapy. In this work, we report a trimodal precision cytotoxic chemo-radio-immunotherapy paradigm using spatially targeted auristatin warheads. Tumor-directed antibodies and peptides conjugated to radiosensitizing monomethyl auristatin E (MMAE) specifically produce CD8 T cell dependent durable tumor control of irradiated tumors and immunologic memory. In combination with ionizing radiation, MMAE sculpts the tumor immune infiltrate to potentiate immune checkpoint inhibition. Here, we report therapeutic synergies of targeted cytotoxic auristatin radiosensitization to stimulate anti-tumor immune responses providing a rationale for clinical translational of auristatin antibody drug conjugates with radio-immunotherapy combinations to improve tumor control.

Lymphatic-preserving treatment sequencing with immune checkpoint inhibition unleashes cDC1-dependent antitumor immunity in HNSCC.

Despite the promise of immune checkpoint inhibition (ICI), therapeutic responses remain limited. This raises the possibility that standard of care treatments delivered in concert may compromise the tumor response. To address this, we employ tobacco-signature head and neck squamous cell carcinoma murine models in which we map tumor-draining lymphatics and develop models for regional lymphablation with surgery or radiation. We find that lymphablation eliminates the tumor ICI response, worsening overall survival and repolarizing the tumor- and peripheral-immune compartments. Mechanistically, within tumor-draining lymphatics, we observe an upregulation of conventional type I dendritic cells and type I interferon signaling and show that both are necessary for the ICI response and lost with lymphablation. Ultimately, we provide a mechanistic understanding of how standard oncologic therapies targeting regional lymphatics impact the tumor response to immune-oncology therapy in order to define rational, lymphatic-preserving treatment sequences that mobilize systemic antitumor immunity, achieve optimal tumor responses, control regional metastatic disease, and confer durable antitumor immunity.

Single-cell characterization of step-wise acquisition of carboplatin resistance in ovarian cancer.

The molecular underpinnings of acquired resistance to carboplatin are poorly understood and often inconsistent between in vitro modeling studies. After sequential treatment cycles, multiple isogenic clones reached similar levels of resistance, but significant transcriptional heterogeneity. Gene-expression based virtual synchronization of 26,772 single cells from 2 treatment steps and 4 resistant clones was used to evaluate the activity of Hallmark gene sets in proliferative (P) and quiescent (Q) phases. Two behaviors were associated with resistance: (1) broad repression in the P phase observed in all clones in early resistant steps and (2) prevalent induction in Q phase observed in the late treatment step of one clone. Furthermore, the induction of IFNα response in P phase or Wnt-signaling in Q phase were observed in distinct resistant clones. These observations suggest a model of resistance hysteresis, where functional alterations of the P and Q phase states affect the dynamics of the successive transitions between drug exposure and recovery, and prompts for a precise monitoring of single-cell states to develop more effective schedules for, or combination of, chemotherapy treatments.

Cancer-cell-secreted extracellular vesicles suppress insulin secretion through miR-122 to impair systemic glucose homeostasis and contribute to tumour growth.

Epidemiological studies demonstrate an association between breast cancer (BC) and systemic dysregulation of glucose metabolism. However, how BC influences glucose homeostasis remains unknown. We show that BC-derived extracellular vesicles (EVs) suppress pancreatic insulin secretion to impair glucose homeostasis. EV-encapsulated miR-122 targets PKM in ß-cells to suppress glycolysis and ATP-dependent insulin exocytosis. Mice receiving high-miR-122 EVs or bearing BC tumours exhibit suppressed insulin secretion, enhanced endogenous glucose production, impaired glucose tolerance and fasting hyperglycaemia. These effects contribute to tumour growth and are abolished by inhibiting EV secretion or miR-122, restoring PKM in ß-cells or supplementing insulin. Compared with non-cancer controls, patients with BC have higher levels of circulating EV-encapsulated miR-122 and fasting glucose concentrations but lower fasting insulin; miR-122 levels are positively associated with glucose and negatively associated with insulin. Therefore, EV-mediated impairment of whole-body glycaemic control may contribute to tumour progression and incidence of type 2 diabetes in some patients with BC.

AMC-070: Lenalidomide Is Safe and Effective in HIV-Associated Kaposi Sarcoma.

PURPOSE: Kaposi sarcoma (KS), an endothelial cell tumor associated with KS herpesvirus (KSHV), remains among the most common malignancies occurring with HIV infection (HIV-KS). As an oral anti-inflammatory, antiangiogenic, and immunomodulatory agent, lenalidomide is potentially an attractive alternative to standard chemotherapy for KS. EXPERIMENTAL DESIGN: The primary objectives of this phase I/II trial were to determine the maximum tolerated dose (MTD) and response rates for lenalidomide in HIV-KS. Secondary objectives included correlating response with natural killer (NK) and T-cell subsets, plasma cytokines, viral copy number, and KSHV gene expression in biopsies. Four dose levels of oral lenalidomide taken 21 consecutive days of 28-day cycles were evaluated in adults with HIV-KS on antiretroviral therapy with controlled viremia. RESULTS: Fifteen and 23 participants enrolled in phases I and II, respectively, 76% of whom had received prior KS therapy. The MTD was not reached, declaring 25 mg as the recommended phase II dose (RP2D). The most frequent adverse events were neutropenia, fatigue, leukopenia, and diarrhea. Of the 25 evaluable participants receiving RP2D, 60% responded. Correlative studies performed in a subset of participants demonstrated a significant increase in proportions of blood T cells with T-regulatory phenotype, and plasma cytokines trended toward a less inflammatory pattern. Clinical response was associated with loss of KSHV transcription. CONCLUSIONS: Lenalidomide is active in HIV-KS. The most common adverse events were manageable. With 60% of participants receiving RP2D obtaining a partial response and <10% discontinuing due to adverse events, the response and tolerability to lenalidomide support its use in HIV-KS. See related commentary by Henry and Maki, p. 2485.

Genetically engineered and enucleated human mesenchymal stromal cells for the targeted delivery of therapeutics to diseased tissue.

Targeting the delivery of therapeutics specifically to diseased tissue enhances their efficacy and decreases their side effects. Here we show that mesenchymal stromal cells with their nuclei removed by density-gradient centrifugation following the genetic modification of the cells for their display of chemoattractant receptors and endothelial-cell-binding molecules are effective vehicles for the targeted delivery of therapeutics. The enucleated cells neither proliferate nor permanently engraft in the host, yet retain the organelles for energy and protein production, undergo integrin-regulated adhesion to inflamed endothelial cells, and actively home to chemokine gradients established by diseased tissues. In mouse models of acute inflammation and of pancreatitis, systemically administered enucleated cells expressing two types of chemokine receptor and an endothelial adhesion molecule enhanced the delivery of an anti-inflammatory cytokine to diseased tissue (with respect to unmodified stromal cells and to exosomes derived from bone-marrow-derived stromal cells), attenuating inflammation and ameliorating disease pathology. Enucleated cells retain most of the cells' functionality, yet acquire the cargo-carrying characteristics of cell-free delivery systems, and hence represent a versatile delivery vehicle and therapeutic system.

Intratumoral immunotherapy using platelet-cloaked nanoparticles enhances antitumor immunity in solid tumors.

Intratumoral immunotherapy is an emerging modality for the treatment of solid tumors. Toll-like receptor (TLR) agonists have shown promise for eliciting immune responses, but systemic administration often results in the development of adverse side effects. Herein, we investigate whether localized delivery of the TLR agonist, resiquimod (R848), via platelet membrane-coated nanoparticles (PNP-R848) elicits antitumor responses. The membrane coating provides a means of enhancing interactions with the tumor microenvironment, thereby maximizing the activity of R848. Intratumoral administration of PNP-R848 strongly enhances local immune activation and leads to complete tumor regression in a colorectal tumor model, while providing protection against repeated tumor re-challenges. Moreover, treatment of an aggressive breast cancer model with intratumoral PNP-R848 delays tumor growth and inhibits lung metastasis. Our findings highlight the promise of locally delivering immunostimulatory payloads using biomimetic nanocarriers, which possess advantages such as enhanced biocompatibility and natural targeting affinities.

Elongated neutrophil-derived structures are blood-borne microparticles formed by rolling neutrophils during sepsis.

Rolling neutrophils form tethers with submicron diameters. Here, we report that these tethers detach, forming elongated neutrophil-derived structures (ENDS) in the vessel lumen. We studied ENDS formation in mice and humans in vitro and in vivo. ENDS do not contain mitochondria, endoplasmic reticulum, or DNA, but are enriched for S100A8, S100A9, and 57 other proteins. Within hours of formation, ENDS round up, and some of them begin to present phosphatidylserine on their surface (detected by annexin-5 binding) and release S100A8-S100A9 complex, a damage-associated molecular pattern protein that is a known biomarker of neutrophilic inflammation. ENDS appear in blood plasma of mice upon induction of septic shock. Compared with healthy donors, ENDS are 10-100-fold elevated in blood plasma of septic patients. Unlike neutrophil-derived extracellular vesicles, most ENDS are negative for the tetraspanins CD9, CD63, and CD81. We conclude that ENDS are a new class of bloodborne submicron particles with a formation mechanism linked to neutrophil rolling on the vessel wall.

Evaluation of IL-17D in Host Immunity to Group A Streptococcus Infection.

IL-17D is a cytokine that belongs to the IL-17 family and is conserved in vertebrates and invertebrates. In contrast to IL-17A and IL-17F, which are expressed in Th17 cells, IL-17D is expressed broadly in nonimmune cells. IL-17D can promote immune responses to cancer and viruses in part by inducing chemokines and recruiting innate immune cells such as NK cells. Although bacterial infection can induce IL-17D in fish and invertebrates, the role of mammalian IL-17D in antibacterial immunity has not been established. To determine whether IL-17D has a role in mediating host defense against bacterial infections, we studied i.p. infection by group A Streptococcus (GAS) in wild-type (WT) and Il17d -/- mice. Compared with WT animals, mice deficient in IL-17D experienced decreased survival, had greater weight loss, and showed increased bacterial burden in the kidney and peritoneal cavity following GAS challenge. In WT animals, IL-17D transcript was induced by GAS infection and correlated to increased levels of chemokine CCL2 and greater neutrophil recruitment. Of note, GAS-mediated IL-17D induction in nonimmune cells required live bacteria, suggesting that processes beyond recognition of pathogen-associated molecular patterns were required for IL-17D induction. Based on our results, we propose a model in which nonimmune cells can discriminate between nonviable and viable GAS cells, responding only to the latter by inducing IL-17D.

Heterogeneity and clonal relationships of adaptive immune cells in ulcerative colitis revealed by single-cell analyses.

Inflammatory bowel disease (IBD) encompasses a spectrum of gastrointestinal disorders driven by dysregulated immune responses against gut microbiota. We integrated single-cell RNA and antigen receptor sequencing to elucidate key components, cellular states, and clonal relationships of the peripheral and gastrointestinal mucosal immune systems in health and ulcerative colitis (UC). UC was associated with an increase in IgG1+ plasma cells in colonic tissue, increased colonic regulatory T cells characterized by elevated expression of the transcription factor ZEB2, and an enrichment of a γδ T cell subset in the peripheral blood. Moreover, we observed heterogeneity in CD8+ tissue-resident memory T (TRM) cells in colonic tissue, with four transcriptionally distinct states of differentiation observed across health and disease. In the setting of UC, there was a marked shift of clonally related CD8+ TRM cells toward an inflammatory state, mediated, in part, by increased expression of the T-box transcription factor Eomesodermin. Together, these results provide a detailed atlas of transcriptional changes occurring in adaptive immune cells in the context of UC and suggest a role for CD8+ TRM cells in IBD.

A Salt-Templated Synthesis Method for Porous Platinum-based Macrobeams and Macrotubes.

The synthesis of high surface area porous noble metal nanomaterials generally relies on time consuming coalescence of pre-formed nanoparticles, followed by rinsing and supercritical drying steps, often resulting in mechanically fragile materials. Here, a method to synthesize nanostructured porous platinum-based macrotubes and macrobeams with a square cross section from insoluble salt needle templates is presented. The combination of oppositely charged platinum, palladium, and copper square planar ions results in the rapid formation of insoluble salt needles. Depending on the stoichiometric ratio of metal ions present in the salt-template and the choice of chemical reducing agent, either macrotubes or macrobeams form with a porous nanostructure comprised of either fused nanoparticles or nanofibrils. Elemental composition of the macrotubes and macrobeams, determined with x-ray diffractometry and x-ray photoelectron spectroscopy, is controlled by the stoichiometric ratio of metal ions present in the salt-template. Macrotubes and macrobeams may be pressed into free standing films, and the electrochemically active surface area is determined with electrochemical impedance spectroscopy and cyclic voltammetry. This synthesis method demonstrates a simple, relatively fast approach to achieve high-surface area platinum-based macrotubes and macrobeams with tunable nanostructure and elemental composition that may be pressed into free-standing films with no required binding materials.

Machine Learning of Discriminative Gate Locations for Clinical Diagnosis.

High-throughput single-cell cytometry technologies have significantly improved our understanding of cellular phenotypes to support translational research and the clinical diagnosis of hematological and immunological diseases. However, subjective and ad hoc manual gating analysis does not adequately handle the increasing volume and heterogeneity of cytometry data for optimal diagnosis. Prior work has shown that machine learning can be applied to classify cytometry samples effectively. However, many of the machine learning classification results are either difficult to interpret without using characteristics of cell populations to make the classification, or suboptimal due to the use of inaccurate cell population characteristics derived from gating boundaries. To date, little has been done to optimize both the gating boundaries and the diagnostic accuracy simultaneously. In this work, we describe a fully discriminative machine learning approach that can simultaneously learn feature representations (e.g., combinations of coordinates of gating boundaries) and classifier parameters for optimizing clinical diagnosis from cytometry measurements. The approach starts from an initial gating position and then refines the position of the gating boundaries by gradient descent until a set of globally-optimized gates across different samples are achieved. The learning procedure is constrained by regularization terms encoding domain knowledge that encourage the algorithm to seek interpretable results. We evaluate the proposed approach using both simulated and real data, producing classification results on par with those generated via human expertise, in terms of both the positions of the gating boundaries and the diagnostic accuracy. © 2019 The Authors. Cytometry Part A published by Wiley Periodicals, Inc. on behalf of International Society for Advancement of Cytometry.

PD-L1:CD80 Cis-Heterodimer Triggers the Co-stimulatory Receptor CD28 While Repressing the Inhibitory PD-1 and CTLA-4 Pathways.

Combined immunotherapy targeting the immune checkpoint receptors cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) and programmed cell death 1 (PD-1), or CTLA-4 and the PD-1 ligand (PD-L1) exhibits superior anti-tumor responses compared with single-agent therapy. Here, we examined the molecular basis for this synergy. Using reconstitution assays with fluorescence readouts, we found that PD-L1 and the CTLA-4 ligand CD80 heterodimerize in cis but not trans. Quantitative biochemistry and cell biology assays revealed that PD-L1:CD80 cis-heterodimerization inhibited both PD-L1:PD-1 and CD80:CTLA-4 interactions through distinct mechanisms but preserved the ability of CD80 to activate the T cell co-stimulatory receptor CD28. Furthermore, PD-L1 expression on antigen-presenting cells (APCs) prevented CTLA-4-mediated trans-endocytosis of CD80. Atezolizumab (anti-PD-L1), but not anti-PD-1, reduced cell surface expression of CD80 on APCs, and this effect was negated by co-blockade of CTLA-4 with ipilimumab (anti-CTLA-4). Thus, PD-L1 exerts an immunostimulatory effect by repressing the CTLA-4 axis; this has implications to the synergy of anti-PD-L1 and anti-CTLA-4 combination therapy.

Immuno-oncological Efficacy of RXDX-106, a Novel TAM (TYRO3, AXL, MER) Family Small-Molecule Kinase Inhibitor.

Expression of the TAM (TYRO3, AXL, MER) family of receptor tyrosine kinases (RTK) has been associated with cancer progression, metastasis, and drug resistance. In immune cells, TAM RTKs can dampen inflammation in favor of homeostatic wound-healing responses, thus potentially contributing to the evasion of cancer cells from immune surveillance. Here we characterize the small-molecule RXDX-106 as a selective and potent pan-TAM RTK inhibitor with slow dissociation kinetics and significant antitumor activity in multiple syngeneic tumor models. Expression of AXL and MER on both immune and tumor cells increased during tumor progression. Tumor growth inhibition (TGI) following treatment with RXDX-106 was observed in wild-type mice and was abrogated in immunodeficient mice, suggesting that the antitumor activity of RXDX-106 is, in part, due to the presence of immune cells. RXDX-106-mediated TGI was associated with increased tumor-infiltrating leukocytes, M1-polarized intratumoral macrophages, and activation of natural killer cells. RXDX-106 proportionally increased intratumoral CD8+ T cells and T-cell function as indicated by both IFNγ production and LCK phosphorylation (pY393). RXDX-106 exhibited its effects via direct actions on TAM RTKs expressed on intratumoral macrophages and dendritic cells, leading to indirect activation of other immune cells in the tumor. RXDX-106 also potentiated the effects of an immune checkpoint inhibitor, α-PD-1 Ab, resulting in enhanced antitumor efficacy and survival. Collectively, these results demonstrate the capacity of RXDX-106 to inhibit tumor growth and progression and suggest it may serve as an effective therapy against multiple tumor types. SIGNIFICANCE: The pan-TAM small-molecule kinase inhibitor RXDX-106 activates both innate and adaptive immunity to inhibit tumor growth and progression, indicating its clinical potential to treat a wide variety of cancers.