Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients.

The development of human cancer is a multistep process characterized by the accumulation of genetic and epigenetic alterations that drive or reflect tumour progression. These changes distinguish cancer cells from their normal counterparts, allowing tumours to be recognized as foreign by the immune system. However, tumours are rarely rejected spontaneously, reflecting their ability to maintain an immunosuppressive microenvironment. Programmed death-ligand 1 (PD-L1; also called B7-H1 or CD274), which is expressed on many cancer and immune cells, plays an important part in blocking the 'cancer immunity cycle' by binding programmed death-1 (PD-1) and B7.1 (CD80), both of which are negative regulators of T-lymphocyte activation. Binding of PD-L1 to its receptors suppresses T-cell migration, proliferation and secretion of cytotoxic mediators, and restricts tumour cell killing. The PD-L1-PD-1 axis protects the host from overactive T-effector cells not only in cancer but also during microbial infections. Blocking PD-L1 should therefore enhance anticancer immunity, but little is known about predictive factors of efficacy. This study was designed to evaluate the safety, activity and biomarkers of PD-L1 inhibition using the engineered humanized antibody MPDL3280A. Here we show that across multiple cancer types, responses (as evaluated by Response Evaluation Criteria in Solid Tumours, version 1.1) were observed in patients with tumours expressing high levels of PD-L1, especially when PD-L1 was expressed by tumour-infiltrating immune cells. Furthermore, responses were associated with T-helper type 1 (TH1) gene expression, CTLA4 expression and the absence of fractalkine (CX3CL1) in baseline tumour specimens. Together, these data suggest that MPDL3280A is most effective in patients in which pre-existing immunity is suppressed by PD-L1, and is re-invigorated on antibody treatment.

Engineering human IgG1 affinity to human neonatal Fc receptor: impact of affinity improvement on pharmacokinetics in primates.

The pH-dependent binding of Igs to the neonatal FcR (FcRn) plays a critical role in the in vivo homeostasis of IgGs. Modulating the interaction between Fc and FcRn through protein engineering is one method for improving the pharmacokinetics of therapeutic Abs. Recent studies disputed the direct relationship between increasing FcRn affinity and improved pharmacokinetic properties. In this work, we studied the pharmacokinetics of two human IgG1 Fc variants in cynomolgus monkey to further clarify the affinity-pharmacokinetic relationship. First, we report a number of novel Fc point mutations and combination variants, including some with primate-specific FcRn-binding improvements. By studying these variants along with some previously described variants across a wide range of affinities, we discovered a direct correlation of pH 6 affinity improvements with neutral pH improvements, suggesting that all of the tested variants exhibit similar pH dependency in FcRn binding. We then evaluated the pharmacokinetics of variants N434A and N434W, which, respectively, gave approximately 4- and 80-fold improvements in pH 6-binding affinity to both human and nonhuman primate FcRn. Surprisingly, clearance of N434W was similar to that of wild type. N434W is the first variant studied in primates that exhibits significant binding to FcRn at pH 7.4, and its clearance substantiates the principle that too much affinity improvement, i.e., beyond that of N434W, does not yield improved pharmacokinetics. In contrast, N434A exhibited a approximately 2-fold decrease in clearance in cynomolgus monkey, supporting the notion that modest increases in pH 6 FcRn affinity can result in improved pharmacokinetics in primates.

A phase 1 study to evaluate the safety and LDL cholesterol-lowering effects of RG7652, a fully human monoclonal antibody against proprotein convertase subtilisin/kexin type 9.

BACKGROUND: Proprotein convertase subtilisin/kexin type 9 (PCSK9) downregulates low-density lipoprotein (LDL) receptors, thereby leading to a rise in circulating LDL cholesterol (LDL-C). RG7652 is a fully human monoclonal antibody against PCSK9. This placebo-controlled, phase 1 ascending-dose study in healthy subjects evaluated the safety of RG7652 and its efficacy as a potential LDL-C-lowering drug. HYPOTHESIS: Anti-PCSK9 antibody therapy safely and effectively reduces LDL-C. METHODS: Subjects (N = 80) were randomized into 10 cohorts. Six sequential single-dose cohorts received 10, 40, 150, 300, 600, or 800 mg of RG7652 via subcutaneous injection. Four multiple-dose cohorts received 40 or 150 mg of RG7652 once weekly for 4 weeks, either with or without statin therapy (atorvastatin). RESULTS: Adverse events (AEs) were generally mild; the most common AEs were temporary injection-site reactions. No serious AEs, severe AEs, AEs leading to study-drug discontinuation, or dose-limiting toxicities were reported. RG7652 monotherapy reduced mean LDL-C levels by up to 64% and as much as 100 mg/dL at week 2; the effect magnitude and duration increased with dose (≥57 days following a single RG7652 dose ≥300 mg). Exploratory analyses showed reduced oxidized LDL, lipoprotein(a), and lipoprotein-associated phospholipase A2 with RG7652. Antidrug antibody against RG7652 tested positive in 2 of 60 (3.3%) RG7652-treated and in 4 of 20 (20.0%) placebo-treated subjects. Simultaneous atorvastatin administration did not appear to impact the pharmacokinetic profile or lipid-lowering effects of RG7652. CONCLUSIONS: Overall, RG7652 elicited substantial and sustained dose-related LDL-C reductions with an acceptable safety profile and minimal immunogenicity.

Transport of drugs in the kidney by the human organic cation transporter, OCT2 and its genetic variants.

The human organic cation transporter 2 (OCT2, SLC22A2) is a multispecific transporter of organic cations, including many clinically used drugs. OCT2 is primarily responsible for the uptake of organic cations across the basolateral membrane of renal tubular epithelial cells and is considered a major transporter in the active secretion of organic cations in the kidney. Uptake of organic cations by OCT2 is driven by the inside-negative membrane potential and is pH-sensitive. Regulation of OCT2 at the transcriptional level by steroid hormones and at the protein level by various protein kinases has been described. Several human genetic variants in the coding region of OCT2 have been identified and functionally characterized, including both polymorphic and rare variants. A variety of structurally diverse compounds have been shown to interact with OCT2, including endogenous compounds, drugs, and dietary supplements.

Preclinical pharmacokinetics, pharmacodynamics, tissue distribution, and tumor penetration of anti-PD-L1 monoclonal antibody, an immune checkpoint inhibitor.

MPDL3280A is a human monoclonal antibody that targets programmed cell death-1 ligand 1 (PD-L1), and exerts anti-tumor activity mainly by blocking PD-L1 interaction with programmed cell death-1 (PD-1) and B7.1. It is being investigated as a potential therapy for locally advanced or metastatic malignancies. The purpose of the study reported here was to characterize the pharmacokinetics, pharmacodynamics, tissue distribution and tumor penetration of MPDL3280A and/or a chimeric anti-PD-L1 antibody PRO304397 to help further clinical development. The pharmacokinetics of MPDL3280A in monkeys at 0.5, 5 and 20 mg · kg(-1) and the pharmacokinetics / pharmacodynamics of PRO304397 in mice at 1, 3 10 mg · kg(-1) were determined after a single intravenous dose. Tissue distribution and tumor penetration for radiolabeled PRO304397 in tumor-bearing mouse models were determined. The pharmacokinetics of MPDL3280A and PRO304397 were nonlinear in monkeys and mice, respectively. Complete saturation of PD-L1 in blood in mice was achieved at serum concentrations of PRO304397 above ∼ 0.5 µg · mL(-1). Tissue distribution and tumor penetration studies of PRO304397 in tumor-bearing mice indicated that the minimum tumor interstitial to plasma radioactivity ratio was ∼ 0.3; saturation of target-mediated uptake in non-tumor tissues and desirable exposure in tumors were achieved at higher serum concentrations, and the distribution into tumors was dose-and time-dependent. The biodistribution data indicated that the efficacious dose is mostly likely higher than that estimated based on simple pharmacokinetics/pharmacodynamics in blood. These data also allowed for estimation of the target clinical dose for further development of MPDL3280A.

Modeling and Simulation to Support Phase 2 Dose Selection for RG7652, a Fully Human Monoclonal Antibody Against Proprotein Convertase Subtilisin/Kexin Type 9.

RG7652 is a fully humanized monoclonal antibody targeting human PCSK9, a regulator of serum low density lipoprotein cholesterol (LDLc) levels. RG7652 prevents degradation of the hepatic LDLc receptors by blocking PCSK9 binding and thereby resulting in efficient LDLc uptake by hepatocytes. The pharmacokinetics of RG7652 have been evaluated in healthy subjects after single and multiple subcutaneous doses. Pharmacokinetic (PK) and pharmacodynamic (PD) models were developed to explain the antibody PK and LDLc time course data. The PK and PD models based on data from healthy subjects were used to simulate the effects of RG7652 on LDLc levels for a range of potential dose regimens in patients with coronary heart disease. A one-compartment PK model combined with an indirect PD response model was able to adequately describe the PK and LDLc data. Simulations of 400 mg every 4 weeks or 800 mg every 8 weeks regimens show significant LDLc reduction and suggest that dosing RG7652 once every month or once every 2 months is predicted to be optimal for the treatment of hypercholesterolemia. The PK and PD model successfully described the PK and LDLc data from healthy subjects in a Phase 1 study, and the model-based simulations provided useful insights and quantitative understanding for the selection of Phase 2 study doses in patients with coronary heart disease. The approach used in the case study demonstrates the utility of modeling and simulation in designing dose-ranging studies.

Estimating the contribution of genes and environment to variation in renal drug clearance.

Renal excretion is the major pathway for elimination of many clinically used drugs and xenobiotics. We estimated the genetic component (rGC) contributing to variation in renal clearance for six compounds (amoxicillin, ampicillin, metformin, terodiline, digoxin and iohexol) using Repeated Drug Application methodology. Data were obtained from published literature. The rGC values of renal clearance of metformin, amoxicillin, and ampicillin, which undergo transporter-mediated secretion, ranged from 0.64-0.94. This finding suggests that variation in the renal clearance of these drugs has a strong genetic component. Additionally, the rGC values of renal clearance of metformin, amoxicillin, and ampicillin were similar to previously reported rGC values for metabolism. By contrast, the rGC values of renal clearance for iohexol, digoxin, and terodiline were low (0.12-0.37). Renal clearance of these compounds occurs mainly through passive processes (e.g. glomerular filtration and passive secretion/reabsorption). The low rGC values of iohexol, digoxin and terodiline suggest that environmental factors may contribute to variation in their renal clearance.

Polymorphisms in a human kidney xenobiotic transporter, OCT2, exhibit altered function.

The completion of the Human Genome Project and the development of high-throughput polymorphism identification methods have allowed researchers to carry out full genetic analyses of many clinically relevant genes. However, few studies have combined genetic analysis with in vitro phenotyping to better understand the relationship between genetic variation and protein function. Many transporters in the kidney are thought to play key roles in defense against a variety of foreign substances. The goal of this study was to understand the relationship between variation in a gene encoding a major renal xenobiotic transporter, OCT2, and transporter function. We report a comprehensive genetic analysis and functional characterization of variants of OCT2. Twenty-eight variable sites in the OCT2 gene were identified in a collection of 247 ethnically diverse DNA samples. Eight caused non-synonymous amino acid changes, of which four were present at >/= 1% in an ethnic population. All four of these altered transporter function assayed in Xenopus laevis oocytes. Analysis of nucleotide diversity (pi) revealed a higher prevalence of synonymous (pi = 22.4 x 10-4) versus non-synonymous (pi = 2.1 x 10-4) changes in OCT2 than in other genes. In addition, the non-synonymous sites had a significant tendency to exhibit more skewed allele frequencies (more negative Tajima's D-values) compared to synonymous sites. The population-genetic analysis, together with the functional characterization, suggests that selection has acted against amino acid changes in OCT2. This selection may be due to a necessary role of OCT2 in the renal elimination of endogenous amines or xenobiotics, including environmental toxins, neurotoxic amines and therapeutic drugs.

Combined administration of RG7652, a recombinant human monoclonal antibody against PCSK9, and atorvastatin does not result in reduction of immune function.

RG7652 is a human IgG1 monoclonal antibody designed to inhibit proprotein convertase subtilisin/kexin type 9 (PCSK9) binding to hepatic low density lipoprotein receptor (LDL-r), thereby blocking PCSK9-mediated degradation of LDL-r. This therapeutic candidate is under development for the prevention of cardiovascular mortality and morbidity in dyslipidemic patients. The primary objective of this study was to evaluate the potential immunotoxicological effects of RG7652 when given to cynomolgus monkeys either alone or in combination with a daily oral dose of atorvastatin. Administration of RG7652 via subcutaneous injection every other week for 12 weeks (a total of seven doses), daily oral doses of atorvastatin (total of 85 doses), and combinations of each up to 15 and 20 mg/kg/dose, respectively, were well tolerated and there was no evidence of alteration in immune function. Administration of pharmacologically relevant doses of RG7652 in combination with atorvastatin to healthy monkeys does not result in clinically meaningful immunosuppression as measured by T-cell dependent antibody responses, natural killer cell activity, immunophenotype, or delayed type hypersensitivity. The only pharmacologically mediated changes observed during the dosing period were the anticipated changes in circulating cholesterol.

Critical role of bioanalytical strategies in investigation of clinical PK observations, a Phase I case study.

RG7652 is a human immunoglobulin 1 (IgG1) monoclonal antibody (mAb) targeting proprotein convertase subtilisin/kexin type 9 (PCSK9) and is designed for the treatment of hypercholesterolemia. A target-binding enzyme-linked immunosorbent assay (ELISA) was developed to measure RG7652 levels in human serum in a Phase I study. Although target-binding assay formats are generally used to quantify free therapeutic, the actual therapeutic species being measured are affected by assay conditions, such as sample dilution and incubation time, and levels of soluble target in the samples. Therefore, in the presence of high concentrations of circulating target, the choice of reagents and assay conditions can have a significant effect on the observed pharmacokinetic (PK) profiles. Phase I RG7652 PK analysis using the ELISA data resulted in a nonlinear dose normalized exposure. An investigation was conducted to characterize the ELISA to determine whether the assay format and reagents may have contributed to the PK observation. In addition, to confirm the ELISA results, a second orthogonal method, liquid chromatography tandem mass spectrometry (LC-MS/MS) using a signature peptide as surrogate, was developed and implemented. A subset of PK samples, randomly selected from half of the subjects in the 6 single ascending dose (SAD) cohorts in the Phase I clinical study, was analyzed with the LC-MS/MS assay, and the data were found to be comparable to the ELISA data. This paper illustrates the importance of reagent characterization, as well as the benefits of using an orthogonal approach to eliminate bioanalytical contributions when encountering unexpected observations.

Effects of altered FcγR binding on antibody pharmacokinetics in cynomolgus monkeys.

Antibody interactions with Fcγ receptors (FcγRs), like FcγRIIIA, play a critical role in mediating antibody effector functions and thereby contribute significantly to the biologic and therapeutic activity of antibodies. Over the past decade, considerable work has been directed towards production of antibodies with altered binding affinity to FcγRs and evaluation of how the alterations modulate their therapeutic activity. This has been achieved by altering glycosylation status at N297 or by engineering modifications in the crystallizable fragment (Fc) region. While the effects of these modifications on biologic activity and efficacy have been examined, few studies have been conducted to understand their effect on antibody pharmacokinetics (PK). We present here a retrospective analysis in which we characterize the PK of three antibody variants with decreased FcγR binding affinity caused by amino acid substitutions in the Fc region (N297A, N297G, and L234A/L235A) and three antibody variants with increased FcγRIIIA binding affinity caused by afucosylation at N297, and compare their PK to corresponding wild type antibody PK in cynomolgus monkeys. For all antibodies, PK was examined at a dose that was known to be in the linear range. Since production of the N297A and N297G variants in Chinese hamster ovary cells results in aglycosylated antibodies that do not bind to FcγRs, we also examined the effect of expression of an aglycosylated antibody, without sequence change(s), in E. coli. All the variants demonstrated similar PK compared with that of the wild type antibodies, suggesting that, for the six antibodies presented here, altered FcγR binding affinity does not affect PK.

Amelioration of type 2 diabetes by antibody-mediated activation of fibroblast growth factor receptor 1.

Clinical use of recombinant fibroblast growth factor 21 (FGF21) for the treatment of type 2 diabetes and other disorders linked to obesity has been proposed; however, its clinical development has been challenging owing to its poor pharmacokinetics. Here, we describe an alternative antidiabetic strategy using agonistic anti-FGFR1 (FGF receptor 1) antibodies (R1MAbs) that mimic the metabolic effects of FGF21. A single injection of R1MAb into obese diabetic mice induced acute and sustained amelioration of hyperglycemia, along with marked improvement in hyperinsulinemia, hyperlipidemia, and hepatosteatosis. R1MAb activated the mitogen-activated protein kinase pathway in adipose tissues, but not in liver, and neither FGF21 nor R1MAb improved glucose clearance in lipoatrophic mice, which suggests that adipose tissues played a central role in the observed metabolic effects. In brown adipose tissues, both FGF21 and R1MAb induced phosphorylation of CREB (cyclic adenosine 5'-monophosphate response element-binding protein), and mRNA expression of PGC-1α (peroxisome proliferator-activated receptor-γ coactivator 1α) and the downstream genes associated with oxidative metabolism. Collectively, we propose FGFR1 in adipose tissues as a major functional receptor for FGF21, as an upstream regulator of PGC-1α, and as a compelling target for antibody-based therapy for type 2 diabetes and other obesity-associated disorders.

Effect of genetic variation in the organic cation transporter 2 on the renal elimination of metformin.

OBJECTIVE: The goal of this study was to determine the effect of a genetic variant in the organic cation transporter 2 (OCT2), OCT2-808G/T, which results in an amino acid change, A270S, on the pharmacokinetics of the antidiabetic drug, metformin. METHODS: The uptake of metformin was performed in stably transfected HEK-293 cells expressing the empty vector (MOCK), the reference OCT2-808G, and the variant OCT2-808T. Healthy individuals with known OCT2 genotypes [14 homozygous for the OCT2 reference allele (808G/G) and nine heterozygous for the variant allele (808G/T, *3D)] were recruited to this study. Metformin concentrations in plasma and urine were measured by liquid chromatography-tandem mass spectrometry method. Creatinine levels were also measured in plasma and urine. Pharmacokinetic parameters were evaluated for both the groups. RESULTS: We observed that in HEK-293 stably transfected cells, OCT2-808T had a greater capacity to transport metformin than did the reference OCT2. Metformin pharmacokinetics was characterized in 23 healthy volunteers of Caucasian and African-American ancestries. We observed that the renal clearance (CL(R)) and the net secretion (SrCL(R)) of metformin were significantly different between the volunteers heterozygous for the variant allele (808G/T), and the volunteers homozygous for the reference allele (808G/G) (P<0.005). Multivariate analysis revealed that OCT2 genotype was a significant predictor of CL(R) and SrCL(R) of metformin (P<0.01). CONCLUSION: We conclude that genetic variation in OCT2 plays an important role in the CL(R) and SrCL(R) of metformin in healthy volunteers.

Anti-BR3 antibodies: a new class of B-cell immunotherapy combining cellular depletion and survival blockade.

Removal of pathogenic B lymphocytes by depletion of monoclonal antibodies (mAbs) or deprivation of B-cell survival factors has demonstrated clinical benefit in both oncologic and immunologic diseases. Partial clinical responses and emerging data demonstrating incomplete B-cell depletion after immunotherapy fuels the need for improved therapeutic modalities. Lessons from the first generation of therapeutics directed against B-cell-specific antigens (CD20, CD22) are being applied to develop novel antibodies with additional functional attributes. We describe the generation of a novel class of B-cell-directed therapy (anti-BR3 mAbs) that combines the depleting capacity of a therapeutic mAb and blockade of B-cell-activating factor (BAFF)-BR3 B-cell survival. In mice, treatment with antagonistic anti-BR3 antibodies results in quantitatively greater reduction in some B-cell subsets and qualitatively different effects on bone marrow plasma cells compared with BR3-Fc BAFF blockade or with anti-CD20 treatment. Comparative analysis of BR3-Fc and anti-BR3 mAb reveals a lower B-cell dependence for BAFF-mediated survival in nonhuman primates than in mice. This novel class of B-cell-targeted therapies shows species characteristics in mice and primates that will guide translation to treatment of human disease.

Functional genomics of membrane transporters in human populations.

Although considerable progress has been made toward characterizing human DNA sequence variation, there remains a deficiency in information on human phenotypic variation at the single-gene level. We systematically analyzed the function of all protein-altering variants of eleven membrane transporters in heterologous expression systems. Coding-region variants were identified by screening DNA from a large sample (n = 247-276) of ethnically diverse subjects. In total, we functionally analyzed 88 protein-altering variants. Fourteen percent of the polymorphic variants (defined as variants with allele frequencies > or =1% in at least one major ethnic group) had no activity or significantly reduced function. Decreased function variants had significantly lower allele frequencies and were more likely to alter evolutionarily conserved amino acid residues. However, variants at evolutionarily conserved positions with approximately normal activity in cellular assays were also at significantly lower allele frequencies, suggesting that some variants with apparently normal activity in biochemical assays may influence occult functions or quantitative degrees of function that are important in human fitness but not measured in these assays. For example, eight (14%) of the 58 variants for which we had measured the transport of at least two substrates showed substrate-specific defects in transport. These variants and the reduced function variants provide plausible candidates for disease susceptibility or variation in clinical drug response.

Functional analysis of polymorphisms in the organic anion transporter, SLC22A6 (OAT1).

OBJECTIVES: The organic anion transporter, OAT1 (SLC22A6), plays a role in the renal elimination of many drugs and environmental toxins. The goal of this study was to identify and functionally characterize OAT1 variants as a first step towards understanding whether genetic variation in OAT1 may contribute to interindividual differences in renal elimination of xenobiotics. METHODS: As part of a larger study, 276 DNA samples from an ethnically diverse population were screened and 12 coding region variants of OAT1 were identified. The non-synonymous variants were then constructed and characterized in Xenopus laevis oocytes. A small family-based clinical study was conducted to determine the renal elimination of a model OAT1 substrate, adefovir (an antiviral agent) in human subjects who possessed a non-functional variant, OAT1-R454Q. RESULTS: Six non-synonymous variants were identified; two (OAT1-R50 H and OAT1-R293W) were present at > or = 1% in at least one ethnic population. These two variants exhibited normal uptake of p-aminohippurate, ochratoxin A and methotrexate assayed in X. laevis oocytes. One variant, OAT1-R454Q, was non-functional with respect to the above substrates. In the clinical study, there was no significant decrease in the renal secretory clearance of adefovir in family members heterozygous for OAT1-454Q in comparison to those with the reference transporter, OAT1-454R. CONCLUSIONS: These data indicate that the coding region of OAT1 has low genetic and functional diversity and suggest that coding region variants of OAT1 may not contribute substantially to interindividual differences in renal elimination of xenobiotics.

Evolutionary conservation predicts function of variants of the human organic cation transporter, OCT1.

The organic cation transporter, OCT1, is a major hepatic transporter that mediates the uptake of many organic cations from the blood into the liver where the compounds may be metabolized or secreted into the bile. Because OCT1 interacts with a variety of structurally diverse organic cations, including clinically used drugs as well as toxic substances (e.g., N-methylpyridinium, MPP(+)), it is an important determinant of systemic exposure to many xenobiotics. To understand the genetic basis of extensive interindividual differences in xenobiotic disposition, we functionally characterized 15 protein-altering variants of the human liver organic cation transporter, OCT1, in Xenopus oocytes. All variants that reduced or eliminated function (OCT1-R61C, OCT1-P341L, OCT1-G220V, OCT1-G401S, and OCT1-G465R) altered evolutionarily conserved amino acid residues. In general, variants with decreased function had amino acid substitutions that resulted in more radical chemical changes (higher Grantham values) and were less evolutionarily favorable (lower blosum62 values) than variants that maintained function. A variant with increased function (OCT1-S14F) changed an amino acid residue such that the human protein matched the consensus of the OCT1 mammalian orthologs. Our results indicate that changes at evolutionarily conserved positions of OCT1 are strong predictors of decreased function and suggest that a combination of evolutionary conservation and chemical change might be a stronger predictor of function.

SNP analysis and presentation in the Pharmacogenetics of Membrane Transporters Project.

The multidisciplinary UCSF Pharmacogenetics of Membrane Transporters project seeks to systematically identify sequence variants in transporters and to determine the functional significance of these variants through evaluation of relevant cellular and clinical phenotypes. The project is structured around four interacting cores: genomics, cellular phenotyping, clinical phenotyping, and bioinformatics. The bioinformatics core is responsible for collecting, storing, and analyzing the information obtained by the other cores and for presenting the results, in particular, for the genomic data. Most of this process is automated using locally developed software written in Python, an open source language well suited for rapid, modular development that meets requirements that are themselves constantly evolving. Here we present the details of transforming ABI trace file data into useful information for project investigators and a description of the types of data analysis and display that we have developed.

Natural variation in human membrane transporter genes reveals evolutionary and functional constraints.

Membrane transporters maintain cellular and organismal homeostasis by importing nutrients and exporting toxic compounds. Transporters also play a crucial role in drug response, serving as drug targets and setting drug levels. As part of a pharmacogenetics project, we screened exons and flanking intronic regions for variation in a set of 24 membrane transporter genes (96 kb; 57% coding) in 247 DNA samples from ethnically diverse populations. We identified 680 single nucleotide polymorphisms (SNPs), of which 175 were synonymous and 155 caused amino acid changes, and 29 small insertions and deletions. Amino acid diversity (pi(NS)) in transmembrane domains (TMDs) was significantly lower than in loop domains, suggesting that TMDs have special functional constraints. This difference was especially striking in the ATP-binding cassette superfamily and did not parallel evolutionary conservation: there was little variation in the TMDs, even in evolutionarily unconserved residues. We used allele frequency distribution to evaluate different scoring systems (Grantham, blosum62, SIFT, and evolutionarily conservedevolutionarily unconserved) for their ability to predict which SNPs affect function. Our underlying assumption was that alleles that are functionally deleterious will be selected against and thus under represented at high frequencies and over represented at low frequencies. We found that evolutionary conservation of orthologous sequences, as assessed by evolutionarily conservedevolutionarily unconserved and SIFT, was the best predictor of allele frequency distribution and hence of function. European Americans had an excess of high frequency alleles in comparison to African Americans, consistent with a historic bottleneck. In addition, African Americans exhibited a much higher frequency of population specific medium-frequency alleles than did European Americans.