Cancer Informatics for Cancer Centers: Sharing Ideas on How to Build an Artificial Intelligence-Ready Informatics Ecosystem for Radiation Oncology.

In August 2022, the Cancer Informatics for Cancer Centers brought together cancer informatics leaders for its biannual symposium, Precision Medicine Applications in Radiation Oncology, co-chaired by Quynh-Thu Le, MD (Stanford University), and Walter J. Curran, MD (GenesisCare). Over the course of 3 days, presenters discussed a range of topics relevant to radiation oncology and the cancer informatics community more broadly, including biomarker development, decision support algorithms, novel imaging tools, theranostics, and artificial intelligence (AI) for the radiotherapy workflow. Since the symposium, there has been an impressive shift in the promise and potential for integration of AI in clinical care, accelerated in large part by major advances in generative AI. AI is now poised more than ever to revolutionize cancer care. Radiation oncology is a field that uses and generates a large amount of digital data and is therefore likely to be one of the first fields to be transformed by AI. As experts in the collection, management, and analysis of these data, the informatics community will take a leading role in ensuring that radiation oncology is prepared to take full advantage of these technological advances. In this report, we provide highlights from the symposium, which took place in Santa Barbara, California, from August 29 to 31, 2022. We discuss lessons learned from the symposium for data acquisition, management, representation, and sharing, and put these themes into context to prepare radiation oncology for the successful and safe integration of AI and informatics technologies.

MD-ALL: an integrative platform for molecular diagnosis of B-acute lymphoblastic leukemia.

B-acute lymphoblastic leukemia (B-ALL) consists of dozens of subtypes defined by distinct gene expression profiles (GEPs) and various genetic lesions. With the application of transcriptome sequencing (RNA-seq), multiple novel subtypes have been identified, which lead to an advanced B-ALL classification and risk-stratification system. However, the complexity of analyzing RNA-seq data for B-ALL classification hinders the implementation of the new B-ALL taxonomy. Here, we introduce MD-ALL (Molecular Diagnosis of ALL), an integrative platform featuring sensitive and accurate B-ALL classification based on GEPs and sentinel genetic alterations from RNA-seq data. In this study, we systematically analyzed 2,955 B-ALL RNA-seq samples and generated a reference dataset representing all the reported B-ALL subtypes. Using multiple machine learning algorithms, we identified the feature genes and then established highly sensitive and accurate models for B-ALL classification using either bulk or single-cell RNA-seq data. Importantly, this platform integrates multiple aspects of key genetic lesions acquired from RNAseq data, which include sequence mutations, large-scale copy number variations, and gene rearrangements, to perform comprehensive and definitive B-ALL classification. Through validation in a hold-out cohort of 974 samples, our models demonstrated superior performance for B-ALL classification compared with alternative tools. Moreover, to ensure accessibility and user-friendly navigation even for users with limited or no programming background, we developed an interactive graphical user interface for this MD-ALL platform, using the R Shiny package. In summary, MD-ALL is a user-friendly B-ALL classification platform designed to enable integrative, accurate, and comprehensive B-ALL subtype classification. MD-ALL is available from https://github.com/gu-lab20/MD-ALL.

MD-ALL: an Integrative Platform for Molecular Diagnosis of B-cell Acute Lymphoblastic Leukemia.

B-cell acute lymphoblastic leukemia (B-ALL) consists of dozens of subtypes defined by distinct gene expression profiles (GEPs) and various genetic lesions. With the application of transcriptome sequencing (RNA-seq), multiple novel subtypes have been identified, which lead to an advanced B-ALL classification and risk-stratification system. However, the complexity of analyzing RNA-seq data for B-ALL classification hinders the implementation of the new B-ALL taxonomy. Here, we introduce MD-ALL (Molecular Diagnosis of ALL), a user-friendly platform featuring sensitive and accurate B-ALL classification based on GEPs and sentinel genetic alterations. In this study, we systematically analyzed 2,955 B-ALL RNA-seq samples and generated a reference dataset representing all the reported B-ALL subtypes. Using multiple machine learning algorithms, we identified the feature genes and then established highly accurate models for B-ALL classification using either bulk or single-cell RNA-seq data. Importantly, this platform integrates the key genetic lesions, including sequence mutations, large-scale copy number variations, and gene rearrangements, to perform comprehensive and definitive B-ALL classification. Through validation in a hold-out cohort of 974 samples, our models demonstrated superior performance for B-ALL classification compared with alternative tools. In summary, MD-ALL is a user-friendly B-ALL classification platform designed to enable integrative, accurate, and comprehensive B-ALL subtype classification.

Pediatric Cerebral Spetzler-Martin Grade 5 Arteriovenous Malformation.

Brain arteriovenous malformations (AVMs) are a type of intracranial high-flow vascular malformation composed of enlarged feeding arteries and draining veins. Without a capillary bed connection, there can be damage to the walls of the arteries and veins, which causes abnormally high blood flow. AVMs are rarely found in children and are thought to expand over time until they become symptomatic. We present an interesting case of a pediatric male who initially presented with seizure-like episodes and was found to have a large frontoparietal Spetzler-Martin (SM) grade 5 AVM after cerebral digital subtraction angiography. Unfortunately, given how much eloquent brain and deep cortical structures were intertwined in the SM5 AVM, risk-benefit analysis favored observation over surgical management. The patient's clinical presentation and imaging findings are described followed by a discussion of the epidemiology, grading system, and treatment of AVMs. After extensive literature review, this clinical entity has been previously reported but is relatively rare in children with prognosis and therapy correlating to the severity of the SM index.

Arterial Dissection of an Anomalous Right Vertebral Artery Arising From the Right Common Carotid Artery Associated With an Acute Right Middle Cerebral Artery Stroke.

In rare cases, the right vertebral artery can have its origin from the right proximal common carotid artery. This anatomical variant can be incidental, but if trauma to the vertebral artery occurs, there can be devastating neurological deficits. Our patient was an adult female who initially presented with new-onset left arm weakness and dysarthria. After a CT angiogram and ultrasound imaging were performed, she was found to have an acute right middle cerebral artery (MCA) stroke in association with dissection of the anomalous right vertebral artery. The patient underwent urgent mechanical thrombectomy in the right MCA and thrombolysis in cerebral infarction (TICI) grade 3 recanalization was successfully achieved. The patient made a complete recovery with no neurological sequelae. A follow-up CT angiogram after six months showed resolution of the dissection and restored patency of the right vertebral artery.

An Unusual Case of Retroperitoneal Plexiform Neurofibromas Found in a Trauma Patient.

The spectrum of presentation for patients who have neurofibromatosis type 1 (NF1), or von Recklinghausen disease, is very diverse due to a phenomenon known as variable expressivity. Patients may or may not present with cutaneous lesions or central nervous system (CNS) manifestations. However, multiple neurofibromas are the pathognomonic hallmark of NF1. The most common abdominal neoplasm is plexiform neurofibromas that affect the retroperitoneal region. We highlight the hospital course of a patient with an unknown history of NF1 who presented for head trauma with plexiform neurofibromas found incidentally on imaging. The radiographic features of neurofibromas are described in addition to the discussion of management and prognosis of NF1.

Miraculous Recovery After Penetrating Trauma to the Mediastinum.

Traumatic injury to the mediastinum can damage critical surrounding structures, including the pericardium, aorta, and bronchial tree. We highlight a miraculous case of a 13-year-old female with no past medical history who presented to the emergency department after being impaled in the chest by a metal fence post. After median sternotomy, the foreign object was removed, and the patient fortunately recovered with no permanent sequelae. The radiographic features of the injury are described, and potential unseen cardiovascular and respiratory complications are discussed.

Post-Cesarean Section Abdominal Wall Endometrioma.

A 30-year-old female with a history of multiple cesarean sections presents to the emergency department with several months of right lower quadrant abdominal pain only associated with her menstrual cycles. CT abdomen and pelvis with contrast was remarkable for an abdominal wall mass that likely represented an endometrioma, and she was subsequently discharged with pain medications and conservative treatment. However, three months later, she returned to the ED, because her pain was unbearable and refractory to medical management. Interventional radiology was consulted for percutaneous biopsy of the soft tissue mass located in her rectus abdominus muscle. Following the procedure, the patient was started on ORILISSA® (elagolix), the first FDA-approved oral treatment for the management of severe pain associated with endometriosis. We highlight an interesting case of post-cesarean section abdominal wall endometrioma implantation and describe the patient's clinical course and disease management. The radiographic features of the mass are described and proposed mechanisms for the development of an abdominal wall endometrioma following a C-section is discussed.

NMR analysis of free and lipid nanodisc anchored CEACAM1 membrane proximal peptides with Ca2+/CaM.

CEACAM1, a homotypic transmembrane receptor with 12 or 72 amino acid cytosolic domain isoforms, is converted from inactive cis-dimers to active trans-dimers by calcium-calmodulin (Ca2+/CaM). Previously, the weak binding of Ca2+/CaM to the human 12 AA cytosolic domain was studied using C-terminal anchored peptides. We now show the binding of 15N labeled Phe-454 cytosolic domain peptides in solution or membrane anchored using NMR demonstrates a significant role for the lipid bilayer. Although binding is increased by the mutation Phe454Ala, this mutation was previously shown to abrogate actin binding. On the other hand, Ca2+/CaM binding is abrogated by phosphorylation of nearby Thr-457, a post-translation modification required for actin binding and subsequent in vitro lumen formation. Binding of Ca2+/CaM to a membrane proximal peptide from the long 72 AA cytosolic domain anchored to lipid nanodiscs was very weak compared to lipid free conditions, suggesting membrane specific effects between the two isoforms. NMR analysis of 15N labeled Ca2+/CaM with unlabeled peptides showed the C-lobe of Ca2+/CaM is involved in peptide interactions, and hydrophobic residues such as Met-109, Val-142 and Met-144 play important roles in binding peptide. This information was incorporated into transmembrane models of CEACAM1 binding to Ca2+/CaM. The lack of Ca2+/CaM binding to phosphorylated Thr-457, a residue we have previously shown to be phosphorylated by CaMK2D, also dependent on Ca2+/CaM, suggests stepwise binding of the cytosolic domain first to Ca2+/CaM and then to actin.

SNPs in inflammatory genes CCL11, CCL4 and MEFV in a fibromyalgia family study.

BACKGROUND: Fibromyalgia (FM) is a chronic pain syndrome with a high incidence in females that may involve activation of the immune system. We performed exome sequencing on chemokine genes in a region of chromosome 17 identified in a genome-wide family association study. METHODS AND FINDINGS: Exome sequence analysis of 100 FM probands was performed at 17p13.3-q25 followed by functional analysis of SNPs found in the chemokine gene locus. Missense SNPs (413) in 17p13.3-q25 were observed in at least 10 probands. SNPs rs1129844 in CCL11 and rs1719152 in CCL4 were associated with elevated plasma chemokine levels in FM. In a transmission disequilibrium test (TDT), rs1129844 was unequally transmitted from parents to their affected children (p< 0.0074), while the CCL4 SNP was not. The amino acid change (Ala23Thr), resulting from rs1129844 in CCL11, predicted to alter processing of the signal peptide, led to reduced expression of CCL11. The variant protein from CCL4 rs1719152 exhibited protein aggregation and a potent down-regulation of its cognate receptor CCR5, a receptor associated with hypotensive effects. Treatment of skeletal muscle cells with CCL11 produced high levels of CCL4 suggesting CCL11 regulates CCL4 in muscle. The immune association of FM with SNPs in MEFV, a chromosome 16 gene associated with recurrent fevers, had a p< 0.008 TDT for a combined 220 trios. CONCLUSIONS: SNPs with significant TDTs were found in 36% of the cohort for CCL11 and 12% for MEFV, along with a protein variant in CCL4 (41%) that affects CCR5 down-regulation, supporting an immune involvement for FM.

Characterization of 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine-N-[Methoxy(polyethylene glycerol)-2000] and Its Complex with Doxorubicin Using Nuclear Magnetic Resonance Spectroscopy and Molecular Dynamics.

Polyethylene glycol (PEG) lipid nanoparticles (LNPs) spontaneously assemble in water, forming uniformly sized nanoparticles incorporating drugs with prolonged blood clearance compared to drugs alone. Previously, 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycerol)-2000] (DSPE-PEG2000) and several drug adducts, including doxorubicin, were analyzed by a combination of physical and molecular dynamic (MD) studies. In this study, a complete chemical shift assignment of DSPE-PEG2000 plus or minus doxorubicin was achieved using nuclear magnetic resonance (NMR), one-dimensional selective nuclear Overhauser spectroscopy (1D-selNOESY), NOESY, correlation spectroscopy (COSY), total correlated spectroscopy (TOCSY), heteronuclear single quantum coherence (HSQC), and HSQC-TOCSY. Chemical shift perturbation, titration, relaxation enhancement, and NOESY analysis combined with MD reveal detailed structural information at the atomic level, including the location of doxorubicin in the micelle, its binding constant, the hydrophilic shell organization, and the mobility of the PEG2000 tail, demonstrating that NMR spectroscopy can characterize drug-DSPE-PEG2000 micelles with molecular weights above 180 kDa. The MD study revealed that an initial spherical organization led to a more-disorganized oblate structure in an aqueous environment and agreed with the NMR study in the details of the fine structure, in which methyl group(s) of the stearic acid in the hydrophobic core of the micelle are in contact with the phosphate headgroup of the lipid. Although the molecular size of the LNP drug complex is about 180 kDa, atomic resolution can be achieved by NMR-based methods that reveal distinct features of the drug-lipid interactions. Because many drugs have unfavorable blood clearance that may benefit from incorporation into LNPs, a thorough knowledge of their physical and chemical properties is essential to moving them into a clinical setting. This study provides an advanced basic approach that can be used to study a wide range of drug-LNP interactions.

How Do Short Chain Nonionic Detergents Destabilize G-Protein-Coupled Receptors?

Stability of detergent-solubilized G-protein-coupled receptors (GPCRs) is crucial for their purification in a biologically relevant state, and it is well-known that short chain detergents such as octylglucoside are more denaturing than long chain detergents such as dodecylmaltoside. However, the molecular basis for this phenomenon is poorly understood. To gain insights into the mechanism of detergent destabilization of GPCRs, we used atomistic molecular dynamics simulations of thermostabilized adenosine receptor (A2AR) mutants embedded in either a lipid bilayer or detergent micelles of alkylmaltosides and alkylglucosides. A2AR mutants in dodecylmaltoside or phospholipid showed low flexibility and good interhelical packing. In contrast, A2AR mutants in either octylglucoside or nonylglucoside showed decreased α-helicity in transmembrane regions, decreased α-helical packing, and the interpenetration of detergent molecules between transmembrane α-helices. This was not observed in octylglucoside containing phospholipid. Cholesteryl hemisuccinate in dodecylmaltoside increased the energetic stability of the receptor by wedging into crevices on the hydrophobic surface of A2AR, increasing packing interactions within the receptor and stiffening the detergent micelle. The data suggest a three-stage process for the initial events in the destabilization of GPCRs by octylglucoside: (i) highly mobile detergent molecules form small micelles around the receptor; (ii) loss of α-helicity and decreased interhelical packing interactions in transmembrane regions are promoted by increased receptor thermal motion; (iii) transient separation of transmembrane helices allowed penetration of detergent molecules into the core of the receptor. The relative hydration of the headgroup and alkyl chain correlates with detergent harshness and suggests new avenues to develop milder versions of octylglucoside for receptor crystallization.

Discovery of potential new gene variants and inflammatory cytokine associations with fibromyalgia syndrome by whole exome sequencing.

Fibromyalgia syndrome (FMS) is a chronic musculoskeletal pain disorder affecting 2% to 5% of the general population. Both genetic and environmental factors may be involved. To ascertain in an unbiased manner which genes play a role in the disorder, we performed complete exome sequencing on a subset of FMS patients. Out of 150 nuclear families (trios) DNA from 19 probands was subjected to complete exome sequencing. Since >80,000 SNPs were found per proband, the data were further filtered, including analysis of those with stop codons, a rare frequency (<2.5%) in the 1000 Genomes database, and presence in at least 2/19 probands sequenced. Two nonsense mutations, W32X in C11orf40 and Q100X in ZNF77 among 150 FMS trios had a significantly elevated frequency of transmission to affected probands (p = 0.026 and p = 0.032, respectively) and were present in a subset of 13% and 11% of FMS patients, respectively. Among 9 patients bearing more than one of the variants we have described, 4 had onset of symptoms between the ages of 10 and 18. The subset with the C11orf40 mutation had elevated plasma levels of the inflammatory cytokines, MCP-1 and IP-10, compared with unaffected controls or FMS patients with the wild-type allele. Similarly, patients with the ZNF77 mutation have elevated levels of the inflammatory cytokine, IL-12, compared with controls or patients with the wild type allele. Our results strongly implicate an inflammatory basis for FMS, as well as specific cytokine dysregulation, in at least 35% of our FMS cohort.

Alkylation of cysteine 468 in Stat3 defines a novel site for therapeutic development.

Stat3 is a latent transcription factor that promotes cell survival and proliferation and is often constitutively active in multiple cancers. Inhibition of Stat3 signaling pathways suppresses cell survival signals and leads to apoptosis in cancer cells, suggesting direct inhibition of Stat3 function is a viable therapeutic approach. Herein, we identify a small molecule, C48, as a selective Stat3-family member inhibitor. To determine its mechanism of action, we used site-directed mutagenesis and multiple biochemical techniques to show that C48 alkylates Cys468 in Stat3, a residue at the DNA-binding interface. We further demonstrate that C48 blocks accumulation of activated Stat3 in the nucleus in tumor cell lines that overexpress active Stat3, leading to impressive inhibition of tumor growth in mouse models. Collectively, these findings suggest Cys468 in Stat3 represents a novel site for therapeutic intervention and demonstrates the promise of alkylation as a potentially effective chemical approach for Stat3-dependent cancers.

Importance of receptor flexibility in binding of cyclam compounds to the chemokine receptor CXCR4.

We have elucidated the binding sites of four moncyclam and one bicyclam antagonist AMD3100, in the human chemokine receptor CXCR4. Using the predicted structural models of CXCR4, we have further predicted the binding sites of these cyclam compounds. We used the computational method LITiCon to map the differences in receptor structure stabilized by the mono and bicyclam compounds. Accounting for the receptor flexibility lead to a single binding mode for the cyclam compounds, that has not been possible previously using a single receptor structural model and fixed receptor docking algorithms. There are several notable differences in the receptor conformations stabilized by monocyclam antagonist compared to a bicylam antagonist. The loading of the Cu(2+) ions in the cyclam compounds, shrinks the size of the cyclam rings and the residue D262(6.58) plays an important role in bonding to the copper ion in the monocylam compounds while residue E288(7.39) is important for the bicyclam compound.

Elucidation of binding sites of dual antagonists in the human chemokine receptors CCR2 and CCR5.

Design of dual antagonists for the chemokine receptors CCR2 and CCR5 will be greatly facilitated by knowledge of the structural differences of their binding sites. Thus, we computationally predicted the binding site of the dual CCR2/CCR5 antagonist N-dimethyl-N-[4-[[[2-(4-methylphenyl)-6,7-dihydro-5H-benzohepten-8-yl] carbonyl]amino]benzyl]tetrahydro-2H-pyran-4-aminium (TAK-779), and a CCR2-specific antagonist N-(carbamoylmethyl)-3-trifluoromethyl benzamido-parachlorobenzyl 3-aminopyrrolidine (Teijin compound 1) in an ensemble of predicted structures of human CCR2 and CCR5. Based on our predictions of the protein-ligand interactions, we examined the activity of the antagonists for cells expressing thirteen mutants of CCR2 and five mutants of CCR5. The results show that residues Trp98(2.60) and Thr292(7.40) contribute significantly to the efficacy of both TAK-779 and Teijin compound 1, whereas His121(3.33) and Ile263(6.55) contribute significantly only to the antagonistic effect of Teijin compound 1 at CCR2. Mutation of residues Trp86(2.60) and Tyr108(3.32) adversely affected the efficacy of TAK-779 in antagonizing CCR5-mediated chemotaxis. Y49A(1.39) and E291A(7.39) mutants of CCR2 showed a complete loss of CCL2 binding and chemotaxis, despite robust cell surface expression, suggesting that these residues are critical in maintaining the correct receptor architecture. Modeling studies support the hypothesis that the residues Tyr49(1.39), Trp98(2.60), Tyr120(3.32), and Glu291(7.39) of CCR2 form a tight network of aromatic cluster and polar contacts between transmembrane helices 1, 2, 3, and 7.

High plasma levels of MCP-1 and eotaxin provide evidence for an immunological basis of fibromyalgia.

Fibromyalgia (FMS), a predominantly female (85%) syndrome, affects an estimated 2% of the US population with skeletal muscle ache, fatigue, headache, and sleep disorder. The pathogenesis of FMS is unknown and there is no laboratory test for diagnosis. In this study, plasma levels of 25 cytokines and chemokines in 92 female patients with FMS and 69 family members were measured compared to 77 controls. Trans-endothelial migration of normal leukocytes in response to FMS plasma and the cytokine profile of human myoblasts were analyzed. High levels of MCP-1 (P<0.001) and eotaxin (P<0.01) were found in patients and family members compared to controls. Patients (56/92) treated with the single agent guaifenesin (>3 months) had higher levels of eotaxin than those not treated (P<0.01). Diluted plasma from patients increased the migration of normal eosinophils and monocytes, but not neutrophils, through an endothelial/Matrigel barrier only when mast cells are included in the lower wells (P<0.05). Furthermore, myoblasts can secrete MCP-1, eotaxin, and IP-10, while treatment with MCP-1 caused secretion of IL-1beta, eotaxin and IP-10. FMS is associated with inflammatory chemokines, that MCP-1 and eotaxin may contribute to the symptoms of FMS, and that similar cytokine profiles found in family members support the idea that FMS has a genetic component. Furthermore, the chemokine profile associated with FMS has direct effects on the migration of eosinophils and monocytes in the presence of mast cells, and skeletal muscle itself may secrete.

Estradiol acts directly on bone marrow myeloid progenitors to differentially regulate GM-CSF or Flt3 ligand-mediated dendritic cell differentiation.

Estrogen receptor (ER) ligands modulate hemopoiesis and immunity in the normal state, during autoimmunity, and after infection or trauma. Dendritic cells (DC) are critical for initiation of innate and adaptive immune responses. We demonstrate, using cytokine-driven culture models of DC differentiation, that 17-beta-estradiol exerts opposing effects on differentiation mediated by GM-CSF and Flt3 ligand, the two cytokines that regulate DC differentiation in vivo. We also show that estradiol acts on the same highly purified Flt3+ myeloid progenitors (MP) to differentially regulate the DC differentiation in each model. In GM-CSF-supplemented cultures initiated from MP, physiological amounts of estradiol promoted differentiation of Langerhans-like DC. Conversely, in Flt3 ligand-supplemented cultures initiated from the same MP, estradiol inhibited cell survival in a dose-dependent manner, thereby decreasing the yield of plasmacytoid and conventional myeloid and lymphoid DC. Experiments with bone marrow cells from ER-deficient mice and the ER antagonist ICI182,780 showed that estradiol acted primarily via ERalpha to regulate DC differentiation. Thus, depending on the cytokine environment, pathways of ER signaling and cytokine receptor signaling can differentially interact in the same Flt3+ MP to regulate DC development. Because the Flt3 ligand-mediated differentiation pathway is important during homeostasis, and GM-CSF-mediated pathways are increased by inflammation, our data suggest that endogenous or pharmacological ER ligands may differentially affect DC development during homeostasis and disease, with consequent effects on DC-mediated immunity.

Estrogen selectively promotes the differentiation of dendritic cells with characteristics of Langerhans cells.

The steroid hormone estrogen regulates the differentiation, survival, or function of diverse immune cells. Previously, we found that physiological amounts of 17beta-estradiol act via estrogen receptors (ER) to promote the GM-CSF-mediated differentiation of dendritic cells (DC) from murine bone marrow progenitors in ex vivo cultures. Of the two major subsets of CD11c(+) DC that develop in these cultures, estrogen is preferentially required for the differentiation of a CD11b(int)Ly6C(-) population, although it also promotes increased numbers of a CD11b(high)Ly6C(+) population. Although both DC subsets express ERalpha, only the CD11b(high)Ly6C(+) DC express ERbeta, perhaps providing a foundation for the differential regulation of these two DC types by estrogen. The two DC populations exhibit distinct phenotypes in terms of capacity for costimulatory molecule and MHC expression, and Ag internalization, which predict functional differences. The CD11b(int)Ly6C(-) population shows the greatest increase in MHC and CD86 expression after LPS activation. Most notably, the estrogen-dependent CD11b(int)Ly6C(-) DC express langerin (CD207) and contain Birbeck granules characteristic of Langerhans cells. These data show that estrogen promotes a DC population with the unique features of epidermal Langerhans cells and suggest that differentiation of Langerhans cells in vivo will be dependent upon local estrogen levels and ER-mediated signaling events in skin.