Conference Scene

Abstract

A major challenge for clinicians, pharmaceutical companies and regulatory agencies is to better understand the relative contributions of ontogeny and genetic variation to observed variability in drug disposition and response across the pediatric age spectrum from preterm and term newborns, to infants, children and adolescents. Extrapolation of adult experience with pharmacogenomics and personalized medicine to pediatric patients of different ages and developmental stages, is fraught with many challenges. Compared with adults, pediatric pharmacogenetics and pharmacogenomics involves an added measure of complexity as variability owing to developmental processes, or ontogeny, is superimposed upon genetic variation. Furthermore, some pediatric diseases have no adult correlate or are more prevalent in children compared with adults, and several adverse drug reactions are unique to children, or occur at a higher frequency in children. The primary objective of this conference was to initiate an ongoing series of annual meetings on ‘Pediatric Pharmacogenomics and Personalized Medicine’ organized by the Center for Personalized Medicine and Therapeutic Innovation and Division of Clinical Pharmacology and Medical Therapeutics at Children’s Mercy Hospitals and Clinics in Kansas City, MO, USA. The primary goals of the inaugural meeting were: to bring together clinicians, basic and translational scientists and allied healthcare practitioners, and engage in a multi- and cross-disciplinary dialog aimed at implementing personalized medicine in pediatric settings; to provide a forum for the presentation and the dissemination of research related to the application of pharmacogenomic strategies to investigations of variability of drug disposition and response in children; to explore the ethical, legal and societal implications of pharmacogenomics and personalized medicine that are unique to children; and finally, to create networking opportunities for stimulating discussion, cooperation and collaboration to devise strategies to address the research needs identified.

Following the opening reception with welcoming remarks and the keynote address, there were three distinct but inter-related sessions. On the first full day of the conference, the first session explored the opportunities and barriers to implementing pharmacogenomics and personalized medicine in pediatric populations. This was followed by the afternoon session that aimed at addressing infrastructure for pediatric personalized medicine, targeting the ethical and practical issues related to biobanks, data warehouses and medical informatics. The final half-day of the conference focused on the clinical applications of pharmacogenomics in children.

Keynote address

Dr Gregory J Downing, (Program Director, Personalized Healthcare Initiative, Office of the Secretary, Department of Health and Human Services, Washington, DC, USA) set the tone for the conference in his keynote address. Dr Downing relayed that for the past 5 years he has been involved in discussions at the federal government level about how to set up opportunities for the transformation in healthcare through this science of understanding individual differences. Government, together with several other interested parties, has a very substantial interest in the success and practice of technologies that can make medical care more precise, more predictive and preemptive, more participatory, and safer. These are the characteristics that are used to define personalized medicine.

One of the more profound elements of the changing landscape of healthcare which will result from genomics, will be in the area of prevention – or more accurately ‘actionable prevention’ – the potential to design lifestyle changes or medical interventions based on current knowledge today that have the ability to affect health status 20 years from now. In the context of pediatrics, a rhetorical question to initiate discussion is: ‘how this might alter the future of newborn screening?’ However, there are very significant policy challenges that lie ahead on the pathway to personalized medicine.

One of those is in the area of pharmacogenomics and the US FDA’s role in regulating genetic tests and drug development. New clinical trial designs or ‘adaptive trial’ designs that reflect substratification of patients will be based on biomarkers. Examples of future challenges include:

• ■ Regulatory oversight on the road ahead. Clinical Laboratory Improvements Amendment certification and the emergence of FDA regulation of genetic tests continue to unfold – genetic tests are powerful technologies in their own right;

• ■ Standards of evidence: potential for closer alignment of FDA and Centers for Medicare and Medicaid Services. Payment in the future may be reliant on use of evidence-based practices. In the context of personalized medicine, many clinical studies will need to be conducted, and innovative ambulatory clinical practice research methods will be required;

• ■ Reimbursement: especially the appropriate valuation of tests and procedures that yield preventive or preemptive value;

• ■ Building translation into the scientific process;

• ■ Educating provider and patient communities – making this shift without taking a generation to move science into patient care;

• ■ Continued vigilance to maintain trust through protected privacy.

In bringing personalized medicine into clinical practice, it is really the healthcare providers who will lead it, and it will be important for the academic community, substantial leader organizations and trusted institutions to help export a personalized medicine culture to primary care physicians, in order for the promise of personalized medicine to reach the communities that will benefit from it.

Session 1: Pediatric pharmacogenomics & personalized medicine: opportunities & barriers

Exploring systems medicine through translational bioinformatics

Dr Atul Butte, (Lucille Packard Children’s Hospital at Stanford, Palo Alto, CA, USA) opened the first full day of the conference with a presentation that also functioned as ‘Grand Rounds’ at Children’s Mercy Hospitals and Clinics (Kansas City, MO, USA). The presentation was structured to be relevant to the attendees of the Grand Rounds – house officers and clinical faculty members as well as conference attendees. A major hurdle to the implementation of personalized medicine in clinical settings is in translating genome-era discoveries to medicine – from bench to bedside to the community. Given that the ability to generate vast volumes of data far exceeds the capacity to annotate and interpret it, there is an acute need for robust, bidirectional information flow between basic and translational scientists. Translational bioinformatics has the potential to address this need through the development of analytic, storage and interpretive methods, and through the optimization of increasingly voluminous genomic and biological data into diagnostics and therapeutics for the clinician. To place the problem and potential solutions in context, Dr Butte described an increasingly common pediatric problem seen by primary care pediatricians, pediatric endocrinologists and other providers – a 12-year old child about to enter 7th grade presenting with acanthosis nigricans, a BMI of 32.3 kg/m², a random glucose 162 mg/dl, and elevated fasting low density lipoprotein, cholesterol and triglycerides. Genome-wide association studies have been conducted with the intent of identifying genetic markers of increased susceptibility to disease. Unfortunately, the results of these studies have been limited by inconsistent reproducibility, relatively weak effect sizes, location of strong signals in gene deserts precluding mechanistic insights, and missing heritability. Publically deposited data in National Center for Biotechnology Information’s Gene Expression Omnibus, the Genetic Association Database and Human Gene Mutation Database were mined using an integrative genomics method to systematically prioritize DNA markers and identify novel causative genes. The results of this analysis revealed that the more often a gene was differentially expressed in diseased samples relative to controls, the more likely it was that it contained disease-associated variants. Applied to Type 2 diabetes, three new disease-related candidate genes were revealed, and subsequently, two were validated in both human studies and mouse knockout models. This example illustrated the potential of genomic tools to uncover novel disease-associated genes that can serve as the targets of new drug development, and perhaps, modify the progression of disease in children with Type 2 diabetes.

The concept of ‘genomic nosology’ was introduced as another method of harnessing large volumes of publically deposited genomic data. In essence, both mRNA expression and protein interaction data were integrated to quantitatively assess the correlation between diseases. Hierarchical clustering of the integrated datasets revealed diseases that were similar on the basis of Module Response Scores incorporating gene expression and functional modules representing biological pathways or networks. Several of the significant disease correlations also shared common drugs, supporting the hypothesis that similar diseases can be treated by the same drugs, allowing predictions for new uses of existing drugs to be made [1]. Preliminary testing of the hypothesis revealed that an anti-epileptic drug had activity in a rat model of inflammatory bowel disease, and an anti-ulcer drug produced dose-dependent restriction of growth of human lung adenocarcinoma cell lines explanted in mice.

The take home message from this presentation were that the patients, samples, molecular, clinical and epidemiological data required to make an impact across medicine are publicly available. Furthermore, personalized medicine should not be restricted to a reliance on DNA, and not all systems biology is molecular. Electronic medical records are a rich source of clinical data that can also be processed into valuable information. Although new tools are necessary to convert data into knowledge, the tools must also be used to convert knowledge into practice, and this latter step requires investigators who can imagine basic questions to ask of these repositories of clinical and genomic measurements.

Translating the genome into chemical probes & new therapeutics for rare & neglected disorders

As Dr Christopher P Austin, (NIH Chemical Genomics Center, Translational Genomics Research, National Human Genome Research Institute [NHGRI], Bethesda, MD, USA) was unable to attend the meeting at the last minute, and thus his talk was presented by Dr Stephen Spielberg (meeting co-organizer; Center for Personalized Medicine and Therapeutic Innovation, Children’s Mercy Hospitals and Clinics). To develop new drugs from the genome, gene identification is only the starting point to determining function and any potential therapeutic value. ‘Validation’ is a multistep process involving definition of sequence function and role in disease, demonstration of manipulability of gene product and transformation of the gene product into a drug target. At present, only a small percentage of gene products and human diseases are being addressed for drug development. It is estimated that approximately 7000 diseases affect humans, but only a very small fraction of disease actually supports commercial development of therapeutic agents. Of the neglected diseases, there are two main types: low prevalence, or rare diseases with a prevalence under 200,000 in the USA, and high-prevalence diseases that are still neglected. In the case of the low-prevalence neglected diseases, there are approximately 6000 orphan diseases with a cumulative prevalence of 25–30 million patients consisting largely of monogenic disorders, such as cystic fibrosis, Tay-Sach’s disease, sickle cell disease, among others. Neglected high-prevalence diseases primarily affect impoverished and marginalized populations in developing nations who are unable to afford treatments and commonly are infectious diseases, such as malaria, schistosomiasis, leishmaniasis and trypanosomiasis. An alternative to private sector drug discovery pathways is the NIH Chemical Genomics Center, which was founded in 2004 to focus on novel targets and rare/neglected diseases. With over 100 collaborations with investigators worldwide, the Center produces chemical probes/leads as well as new paradigms for assay development, screening, informatics and chemistry.

One such paradigm is the quantitative high-throughput screening (qHTS) process that involves multiple (7–15 concentrations over a wide range [2–100 μM]) in a 1536-well format, coupled with an informatics pipeline for data processing, curve fitting and classification, and extraction of structure-activity relationships. This system has been applied to repurpose an existing drug for Niemann–Pick Disease Type C, and has been utilized to identify oxadiazoles as new drug leads for the control of schistosomiasis. The Chemical Genomics Center, Roadmap Molecular Libraries Program and Therapeutics for Rare and Neglected Diseases Program are all new NIH programs developed to translate genes into drugs for rare and neglected diseases, many of which affect children.

Whole genome approaches to pediatric disease

Dr Hakon Hakonarson, (Center for Applied Genomics, Children’s Hospital of Philadelphia, Philadelphia, PA, USA) began his presentation by emphasizing the scope of the healthcare burden represented by childhood disease. Approximately 15% (10 million) of US children 0–18 years have special healthcare needs and account for nearly 50% of total medical care costs in the US. Furthermore, conditions unique to children (congenital malformations and pediatric cancers) are the most common causes of mortality and disabilities in childhood. One example of the cost associated with the treatment of chronic disease in children is illustrated by the rising costs associated with the treatment of cancer. A New England Journal of Medicine article published in 2006 [2], reported that among 10,397 of childhood cancer survivors, 62.3% had at least one chronic condition, 27.5% had a severe or life-threatening condition, and among survivors, the cumulative incidence of chronic health condition reached 73%, 30 years after cancer diagnosis. Given that genetic variations underlie both disease susceptibility and drug response, the need for genomic approaches to the diagnosis and treatment of pediatric disease is critical. The Center for Applied Genomics at the Children’s Hospital of Philadelphia was founded in June 2006, to address this challenge. The current staff of 45 individuals is involved in over 30 active disease-related projects with collaborators at Children’s Hospital of Philadelphia and the University of Pennsylvania, PA, USA, and has genotyped over 100,000 samples to date, with a genotyping capacity exceeding 1500 samples/week. In addition, over 60 collaborations have been established world-wide for replication purposes. The genomic data are accompanied by extensive phenotypic information derived from electronic medical records for each child. This rich resource has been utilized for genome-wide association studies in autism spectrum disorders, asthma, neuroblastoma, Crohn’s disease, ulcerative colitis, Type 1 diabetes and eosinophilic enteritis. The results of these studies have led to subsequent functional genomic investigations to identify and characterize pathways involved in disease pathogenesis. Ultimately, the goal is to utilize this information to individualize therapy based on genetic subsets, and to both predict response to treatment and optimize medications to avoid toxicity, and enhance treatment as well as to identify high-risk patients through screening of siblings and offspring.

Venture philanthropy as a derisking strategy for pediatric research & development

Lesa Mitchell (Ewing Marion Kauffman Foundation, Kansas City, MO, USA) described the role that venture philanthropy is playing to encourage collaborations with the intent of fostering research into orphan diseases affecting children, and the development of new therapeutic options to treat them. Although the traditional role of nonprofit foundations is to provide basic research grants to increase scientific knowledge in their disease sectors, some are now moving towards activities that represent investments in their areas of specific interest. Such initiatives include early-stage funding for proof-of-concept and target validation, as well as project management support and access to their network of scientific experts and research clinics critical for translating discoveries into clinical applications and implementation strategies. In particular, access to patient populations represents, in essence, multisite clinical trial capabilities that create eff iciencies critical in de-risking these early-stage proof-of-concept and target-validation activities. A more detailed discussion of venture philanthropy and derisking strategies can be found at [101].

Session 2: Infrastructure for personalized medicine: ethical issues related to biobanks, data warehouses & medical informatics

BioVU: the future of pediatric biobanking at Vanderbilt University

Dr Louis J Muglia, (Monroe Carell Jr Children’s Hospital, and Department of Pediatrics, Vanderbilt University School of Medicine, Nashville, TN, USA) provided an overview of the extension of Vanderbilt Medical Center’s BioVU program to include pediatric patients. The BioVU program combines genomic DNA isolated from discarded blood samples with a de-identified ‘mirror’ image of the electronic health record referred to as the ‘synthetic derivative’ [3]. Extensive preparatory work, including focus groups and community consultation, and evaluation by the institutional review board, multiple ethics boards, Vanderbilt’s legal department, and the federal Office for Human Research Protections (OHRP), was undertaken prior to implementation of the program. A key element enabling the development of the BioVU program was a change in the ‘consent-to-treat’ form that patients sign every year. The form now includes a prominently positioned box that allows patients/families to ‘opt-out’ of the program. At present the pediatric program is being implemented in ambulatory clinics where children are being seen for routine care; extension to inpatients is being planned for the future. Pediatric-specific concerns were explored with a targeted approach aimed at raising awareness of the program among parents, caregivers and the community through press announcements, local print materials, posters and brochures and newsletters. Based on interviews with 66 English-speaking parents, grandparents and guardians involving a variety of ethnicities and national origins, the vast majority of parents interviewed (91%) would permit their own child’s information to be included in BioVU. A demonstration project involving genotyping ‘high-value’ SNPs in the first 10,000 adult samples accrued in the BioVU program replicated 21 SNPs that had previously been associated by replicated genome-wide experiments with common diseases and traits, including atrial fibrillation, Crohn’s disease, rheumatoid arthritis, Type 2 diabetes and multiple sclerosis. Thus, the pediatric BioVU resource has considerable potential to accelerate discovery research for pediatric diseases.

Ethical & legal barriers to pediatric biobanking initiatives

Dr Kyle B Brothers, (Center for Biomedical Ethics and Society, Vanderbilt University Medical Center, Nashville, TN, USA) addressed some of the ethical and legal issues that were considered for the inclusion of pediatric samples in the BioVu program. According to 45 CFR, Part 46, Section 102 [102,103]: “Human subject means a living individual about whom an investigator (whether professional or student) conducting research obtains data through intervention or interaction with the individual, or identifiable private information”. Private information includes information which is provided: “…by an individual and which the individual can reasonably expect will not be made public, such as an electronic medical record”, and “individually identifiable (i.e., the identity of the subject is or may readily be ascertained by the investigator or associated with the information)”. Differentiating between ‘human subjects research’ and ‘nonhuman subjects research’ is a key consideration in developing a resource like BioVU. Specifically, the October 16, 2008 OHRP guidance indicates that “under certain limited conditions, research involving only coded private information or specimens is not considered human subjects research”. With respect to nonhuman subjects research involving children, several additional issues need to be addressed, including the scope of informed consent, adolescent assent and re-consent of subjects at 18 years of age. The decision to pursue an opt-out model was based on a population-based study conducted in the Nashville, TN, USA, area that involved 677 respondents and oversampled minority groups. The results indicated that 93.5% of respondents would approve of a biobank if the opt-out option were available whereas only 39.1% would approve of a biobank if conducted without getting written permission. Overall, the Vanderbilt experience with the development of the BioVU program revealed that the spectrum of responsibility in research requires more granularity than current policy provides, and ‘best practices’ in human nonsubjects biobanking involves an optout option and long-term monitoring for re-identification risk.

Critical issues in pediatric ethics: consent, assent & distributive justice

Dr John Lantos (Children’s Mercy Bioethics Center, Children’s Mercy Hospitals and Clinics) addressed theoretical and practical issues in assent and consent. He began by reviewing the recent history of the development of research regulations in pediatrics. He showed how these began with three assumptions. First, research is seen as riskier than standard therapy. Second, because of this, research subjects are seen as being in need of special protections. Finally, children are seen as particularly vulnerable to exploitation in research because they cannot give valid informed consent. He raised questions about each of these assumptions. Research, as defined today, is not inherently risky. Some research projects are very low risk, others have risks but those risks are comparable to or lower than the risks of standard therapy. Finally, children are only particularly vulnerable if their parents are not appropriate protectors and decision makers. Usually, they are. He then showed the implications of these views for research on biobanking. He argued that the Vanderbilt/BioVU model seeks a compromise between the true goals of the research and the restraints of current regulation that demands rigorous de-identification of research subjects as a trade-off for waiving (or minimizing) the requirement for parental consent or child assent. The result of this compromise is to protect research subjects from any risk that might arise from a breach of confidentiality. However, it also insures that no research subject can possibly benefit from the research. This, he concluded, is a necessarily tentative and temporary compromise.

Ethical implications of pediatric genomes: a regulatory perspective

Dr Robert ‘Skip’ Nelson (Office of Pediatric Therapeutics, FDA, MD, USA) reviewed the current state of pharmacogenomics and then explored selected ethical and regulatory issues in the application of pharmacogenomics to drug development. To place his presentation in perspective, Dr Roberts summarized the promise of pharmacogenomics in terms of improving the productivity of new drug discovery and development pipelines, reducing adverse drug reactions, and improving disease prevention and management, thus reducing healthcare costs. Contrasting the promise with what has been delivered thus far was a sobering experience. Reviewing the findings of a recently published assessment [4], the current state of pharmacogenomic research can be described as an excess of reviews and commentaries over primary research articles. Original research employs primarily a candidate gene approach using common alleles and small sample sizes with no trend towards larger sample sizes apparent between 1987 and 2007. The consequence of these limitations is the risk of under-powered studies and a preponderance of nominally significant associations (defined as p < 0.05). Dr Roberts then focused the remainder of his presentation of key elements of recommendations for future research outlined by Holmes et al. [4]:

• ■ Biospecimen collection and storage;

• ■ The impact of genetic stratification on clinical trial design;

• ■ Linkage of diagnostic testing to drug indications;

• ■ Risks of genetic profiling and discrimination.

Fulfilling the promise of personalized medicine will require wider incorporation of pharmacogenomics into product development through, for example:

• ■ Genomic stratification;

• ■ Pharmacogenomic-aided pharmacokinetic studies;

• ■ Dose/exposure/response study designs incorporating dosing adjustments as well as individual exposure/response relationships using validated biomarkers, and large-scale clinical trials to determine, for example, the positive and negative predictive value of genome-based testing.

Finally, the pace and scope of pharmacogenomic discovery will track the technological advances in DNA sequencing technology, with the extension of study designs beyond candidate gene approaches towards increased utilization of genome-wide association studies and targeted resequencing to detect both common and rare variants in pharmacogenomic studies and, ultimately, whole-genome sequencing. Future challenges include establishing a common international standard for biospecimen collection, storage and use, reflecting the shared values of transparency, community engagement and benefit-sharing, reframing genomic data based on functional use, delinking it from racial and ethnic categories.

Ethical & social challenges facing pharmacogenomic research & implementation of personalized medicine in children: a Japanese perspective

Dr Hidefumi Nakamura, (National Children’s Medical Center. National Center for Child Health and Development, Setagaya-ku, Tokyo, Japan) reported that attitudes towards studying medications in children have undergone dramatic change in Japan over the past 10 years. For example, in 1999 many pediatricians were reluctant to conduct pharmacokinetic studies in children, and clinical trial infrastructure was very weak. By contrast today, full pharmacokinetic and/or population pharmacokinetic studies are ongoing in many areas including neonatology, and discussions regarding possible collaboration in pharmacogenomic/pharmacokinetic research in children has been initiated. To place this change in the proper context, Dr Nakamura outlined the history of pediatric pharmacology in Japan from its inception in 1974 to the formation of the Japan Society of Developmental Pharmacology and Therapeutics (JSDPT) in 1996 and its current membership of 366 individuals. Currently, emphasis is being placed on the development of the necessary infrastructure to conduct clinical studies in children. An example of recent developments in this area has been the opening of the government supported National Center for Child Health and Development, (public organization, Tokyo, Japan) devoted to pediatrics/obstetrics and related medical fields. Facilities include the National Research Institute for Child Health and Development, and the 450-bed National Medical Center for Children and Mothers. A 5-year Clinical Trial Promotion Plan has resulted in establishment of core Clinical Research Centers and additional major clinical trial institutions that serve as sites for performing trials. The pediatric clinical trial capabilities consist of a network of 27 major children’s hospitals and a continually updated database of clinical trial capacity that interacts with disease-focused networks. Thus, the recent development of clinical trials infrastructure for pediatric studies is one difference between Japan and the USA, for example. Another important distinction relates to cultural differences in the approach to ethical considerations. Asian approaches have always been based on relationships rather than on individual rights. More specifically, consideration of individual rights has been a relatively recent development in that there really was no discussion in non-Western countries before interaction with Westerners. Asian approaches reject self-interest, individualism and contractualism; the heart of Confucian ethics is love and care for others. At present, the majority of Japanese people polled regarding this issue, appear to accept the prospects of pharmacogenetic testing, but limitations are similar to those existing in North America and Europe. Although significant challenges exist, advances are being made on multiple fronts such that prospects for the future are bright.

Session 3: Clinical applications

Pediatric pharmacogenomics: what makes children different?

Dr J Steven Leeder, (Division of Clinical Pharmacology and Medical Toxicology, Department of Pediatrics, Children’s Mercy Hospitals and Clinics) began his presentation by outlining several reasons why pharmacogenomic principles should be broadly integrated into pediatric therapeutics. First, most drugs are developed for ‘adult’ diseases, but that does not necessarily mean that they are equally appropriate for pediatric disease, even if the diseases are nominally similar. For example, the clinical characteristics of juvenile idiopathic arthritis (JIA) are quite different than those found in adult rheumatoid arthritis patients; the rheumatoid factor is rarely found in children less than 10 years of age, and is present in only approximately 5% of JIA patients, whereas uveitis is much more frequently a complication in JIA compared with adults. Furthermore, some pediatric diseases have no adult correlate. Examples include patent ductus arteriosus and persistent pulmonary hypertension of the newborn where there are no adult treatment paradigms that can be extrapolated to patients in the newborn intensive care unit. Likewise, acute lymphoblastic anemia (ALL), Kawasaki’s disease, neuroblastoma and Wilm’s tumor are encountered in children, and have no close correlate in adults. Some adverse drug reactions (ADRs) are unique to, or more prevalent in, children compared with adults. Finally, children with asthma, autism, attention deficit hyperactivity disorder and epilepsy, for example, become adults with asthma, autism, attention deficit hyperactivity disorder or epilepsy, and appropriate early intervention during childhood has the potential to alter the disease presentation at older ages. Given this context, there are several important concepts unique to the application of pharmacogenomics in children. Specifically, developmental changes in the expression of genes involved in drug disposition or drug activation is superimposed upon variability owing to genetic variation. Therefore, it is essential that the relative contributions of ontogeny and genetic variation be well characterized to facilitate a better understanding of the dose–exposure–response relationship for medications in various pediatric subpopulations – a key step in determining dosing strategies that are required to optimize safety and efficacy at a given age or developmental stage. A systematic approach [5] to gather information relevant to designing studies to explore the relative contributions of ontogeny and genetic variation in children was presented, and its application was illustrated using CYP2D6 as an example. More recent investigations reveal that it is also important to characterize the extent to which medications administered to children perturb biochemical pathways undergoing change during growth and development, interindividual variability in the extent of perturbation, the contribution of developmental, genetic and environmental factors to the observed variability in perturbation, and ultimately, to develop biomarkers predictive of risk for lack of efficacy or toxicity in children.

Pharmacogenomics & advances in treatment of ALL

Dr William E Evans, (CEO, St Jude Children’s Research Hospital, Memphis, TN, USA) addressed recent developments related to the importance of genetic polymorphisms in drug metabolism, disposition and drug targets in determining the response to chemotherapy for the treatment of ALL. Highly variable clearance of chemotherapeutic agents in children results in highly variable systemic exposure to those same agents. Variable systemic exposure to ALL chemotherapy, in turn, can influence efficacy. Although many factors influence drug disposition and response, it is likely that genome variation is a major determinant. Initial investigations of the role of genetic variation and variability in the response to chemotherapeutic drugs focused on thiopurine S-methyltransferase (TPMT) pharmacogenetics and the use of 6-mercaptopurine (6MP) in ALL. The risk of hematological toxicity is determined by TPMT genotype with approximately one in 300 patients homozygous for nonfunctional TPMT alleles. These individuals have dramatically reduced TPMT activity, and more drug is available to be converted to thioguanine nucleotides, thus increasing the risk of bone marrow toxicity. Investigations have expanded beyond TPMT to identify additional genes that modulate the response to other ALL treatment modalities. One strategy has been to characterize gene-expression patterns in ALL to elucidate genomic determinants of drug effects. Such studies have addressed issues such as disease classification and prognosis in ALL, the role of SNPs and chromosomal abnormalities in ALL, treatment-specific changes in gene expression, genomic determinants of drug pharmacokinetics in ALL cells, and finding new genes related to drug resistance and treatment outcome. For the latter investigations, the sensitivity (concentration of drug that is lethal to 50% of cells) of primary ALL cells to four antiluekemic agents, prednisone, vincristine, daunorubicin and asparaginase was determined in vitro, and gene-expression patterns were compared in sensitive and resistant cells. Across the four drugs there were 124 genes that were differentially expressed between the sensitive and resistant cells; only three of these genes have been previously associated with drug resistance or prognosis of ALL indicating that there is much to be learned about mechanisms of drug resistance. Another strategy being applied to find new genes influencing ALL therapy is genome-wide SNP analysis. Using this approach, the common 521C>T SNP in SLCO1B1 was associated with the clearance of high-dose methotrexate (MTX) and gastrointestinal toxicity was reduced in ALL patients with CC genotypes [6]. A similar approach identified five SNPs in the IL15 gene that were associated with increased risk of minimal residual disease after initial treatment, which is associated with treatment outcome in ALL. Furthermore, 21 of 102 additional SNPs identified in this analysis could be associated with anticancer drug pharmacokinetics, 20 of these 21 SNPs have been plausibly linked to low-minimal residual disease via greater drug exposure. Finally, whole-genome approaches have also provided novel insights into genes associated with risk of childhood ALL, such as IKZF1 and ARID5B.

Canadian pharmacogenomics network for drug safety: genomic approaches to adverse drug reactions in children

Dr Bruce C Carleton (Director, Pharmaceutical Outcomes Programme, Children’s and Women’s Health Centre of British Columbia, Vancouver, BC, Canada) described the development of the Canadian Pharmacogenomics Network for Drug Safety (CPNDS) and reported on some of its early successes related to the pharmacogenomics of ADRs in children. Originally funded by Genome Canada as the ‘Genotype-specific Approaches to Therapy in Children (GATC)’ network, the program has evolved into the CPNDS with CAD$20M of additional funding (2009–2013) from the Canadian Institutes of Health Research and the Canada Foundation for Innovation. This Canada-wide network represents a hospital-based active surveillance network consisting of several surveillance clinicians in 13 tertiary care teaching hospitals accounting for approximately 90% of hospitalized pediatric patients; new collaborations have been established to accelerate active recruitment of ADR cases and controls in the Canadian pediatric oncology network. At each hospital site, the surveillance clinician is responsible for identifying severe ADRs using standardized ascertainment algorithms, entering clinical data in a remote terminal that sends the data to the centralized clinical database, identifying age- and sex-matched controls, obtaining informed consent from the ADR case and parents and the controls, and then sending the collected blood or saliva samples to the central repository [7]. Pharmacogenomic analyses are prioritized using an 11-item prioritization process and initiated once a sufficient number of cases have been collected.

As of September 2009, a total of 2614 severe ADR cases and 19,957 drug-matched controls have been enrolled in the study. Genotyping is conducted using a custom Illumina panel including 3072 SNPs in 248 genes involved in drug absorption, distribution, metabolism and excretion (ADME). Prioritized drug-ADR pairs include cisplatin ototoxicity, anthracycline-induced cardiotoxicity and codeine-induced infant mortality. Dr Carleton presented recent results related to cisplatin ototoxicity in which this genotyping panel was applied to DNA samples obtained from an initial cohort of 54 children treated in pediatric oncology units in Vancouver and a second replication cohort of 112 children recruited from the rest of Canada. Genotype analysis identified genetic variants in TPMT (rs12201199, p = 0.00022, OR: 17.0, 95% CI: 2.3 –125.9) and COMT (rs9332377, p = 0.00018, OR: 5.5 , 95% CI: 1.9–15.9) associated with cisplatin-induced otoxicity in this pediatric cohort. Furthermore, increasing numbers of risk alleles were associated with increased severity, frequency and earlier onset of hearing loss in this study [8]. These results illustrate the value of multicenter active surveillance networks, such as CPNDS, as they are well-equipped to identify ADRs in pediatric patients, and to capture the clinical and genomic data that can then be used to investigate the etiology of these drug responses in children.

Making a difference: challenges of incorporating pharmacogenomics into the treatment of juvenile idiopathic arthritis

Dr Mara L Becker (Division of Clinical Pharmacology and Medical Toxicology, Department of Pediatrics, Children’s Mercy Hospitals and Clinics) described her work investigating the pharmacogenomics of variability in the response of patients with Juvenile Idiopathic Arthritis to MTX, the first choice second line agent used to treat JIA worldwide. Use of the drug is characterized by considerable variability in response and toxicity that currently is unpredictable. Intracellular MTX polyglutamate concentrations, in conjunction with folate pathway gene polymorphisms, have been investigated as predictors of response in JIA and rheumatoid arthritis and have produced conf licting results. With the onset of action of MTX being several weeks to months after initiation of treatment, the risk:benefit ratio early in treatment is shifted more towards ‘risk’. Therefore, there is a critical need for early identification of responders and/or those at increased risk for toxicity to promote the safest and most effective therapeutic management.

Dr Becker presented the results of a study investigating MTX polyglutamates (MTXPGs) in 104 JIA patients maintained on stable doses of the drug. Concentrations and relative proportions of MTXPGs ranging in chain length from native MTX containing one glutamate residue (MTXPG₁) to the heptaglutamate form (MTXPG₇) were determined in erythrocyte lysates utilizing a newly developed ion-pair chromatographic procedure with mass spectrometric detection (LC-MS/MS). Intracellular MTXPG_1–7 concentrations were found to be extremely variable in this cohort of JIA patients, with a 40–100-fold difference in individual subtypes between patients. Clinical factors associated with this variability included dose of MTX, duration of treatment, and route of administration with higher concentrations of long-chain MTXGlu_3–5 in patients who were dosed via the subcutaneous route, and higher concentrations of short-chain MTXGlu₁₊₂ in patients dosed via the oral route after correction for MTX dose administered. Genetic predictors of MTXPG variability were also evaluated, and several SNPs within the purine synthesis, adenosine, and mitochondrial folate pathway were associated with differential MTXGlu patterns with some of these patterns associated with increased risk of hepatotoxicity. In addition, long chain polyglutamate concentrations corrected for dose administered (cMTX-Glu_3–5) were higher in subjects who had elevated liver function tests at the time of their visit and in subjects who reported GI side-effects. However, minimal to no association between MTXPG_n concentrations and clinical response to MTX was observed. Rather, the consequences of the extent to which MTX perturbs the folate pathway may be a function of the folate pathway phenotype at the time that MTX treatment is initiated, and this hypothesis is the focus of ongoing investigation.

Cooperation, collaboration & the success of pediatric pharmacogenomic initiatives

Dr Bertram H Lubin (President and CEO, Children’s Hospital and Research Center Oakland, Oakland, CA, USA) summarized the current state of knowledge regarding pharmacogenomics and its application to personalized medicine, and then presented several hypotheses underlying successful implementation in a pediatric context:

• ■ Genomic information will be useful to predict, prevent and treat illness in children;

• ■ Genomic information will contribute to cost effective and evidence-based medicine;

• ■ Genomic information will be reimbursed;

• ■ National health insurance will embrace personalized medicine;

• ■ Laboratories will provide appropriate tests at a reasonable cost;

• ■ Clinical trials will generate data to support genomic, proteomic, metabolomic and pharmacogenetic approaches.

Testing these hypotheses is not without challenges. For example, compared with the situation in adult medicine, clinical trials with pediatric populations are much more limited. As a consequence, chronic illness results in use of multiple drugs, requires understanding interactions, and often is not based on clinical trial results. The absence of systematically obtained data means that the dose for a particular medication used in children often is not supported by the results of a clinical trial. Furthermore, clinical variability of monogenetic diseases is poorly understood, and little attention is paid to the issue of ontogeny. Finally, there is a general lack of understanding, education and training for pediatricians in pharmacology. Dr Lubin went on to suggest several potential building blocks for success such as patient and parent advocacy groups for orphan drug development, increased representation of minority groups, more pediatric representation in the Personalized Medicine Coalition, increased commitment to pediatric issues from the NIH (e.g., Better Pharmaceuticals for Children Act, or BPCA), and international involvement. As no single institution will have sufficient numbers of patients to adequately power genomic studies requiring thousands of patients, networks and other forms of collaborative enterprises are necessary to move the field forward. In Dr Lubin’s experience with research networks he observed that friends work best together, especially when relationships are built on mutual respect, interest and are noncompetitive in nature such that teamwork is valued. Successful enterprises also require adequate resources, such as infrastructure, patient numbers and technical capabilities. Finally, the overall mission must contribute to academic interests directed toward improving the care of pediatric patients. Dr Lubin noted that pediatricians like to work together, and there are several examples of successful multicenter networks in diseases such as leukemia and sickle cell anemia (with which he has personal experience), as well as other networks, such as the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Network of Pediatric Pharmacology Research Units. To be successful, though, participants will need to reach consensus on common institutional review board approval processes, and be willing to contribute to centralized resources like biological sample repositories and data warehouses into which clinical data from electronic medical records are deposited. To move forward, major challenges include overcoming the relative dearth of pediatric representation in existing pharmacogenomic and personalized medicine initiatives, and there are also several educational issues that will need to be addressed if genomic medicine is to reach practitioners and the patients/families they serve.

Personalized medicine for children: where are we & where are we headed?

Stephen P Spielberg summarized the meeting and articulated some of the opportunities and challenges to be expected in the future. He proposed that the challenge is how to define in real time for real patients the correct etiologic diagnosis and select the best medicine at the best dose with the highest probability of benefit and the lowest risk of side effects. He proposed a series of questions to consider when assessing the consequences of diversity in drug targets. For example, does the target have genetic variants in the population? Are the variants likely to alter the effectiveness of the drug? What percentage of the population expresses the variants, and how might this impact the percentage of the population likely to benefit? What is the relevance of the target in the metabolic ‘economy’ of complex pathways we are trying to impact? And related to this last question, are there alternative pathways that may lead to ‘escape’ from impacting the target, and how do these pathways vary in the population? However, even though application of genomic technologies may be foremost in the minds of many, those at the interface of patient care face a different set of realities that differentiate ‘efficacy’ from ‘effectiveness’. For example, it is important to take into consideration patient preferences, understanding and adherence to prescribed treatments as factors determining individual risk and benefit as even ‘minor’ side effects may lead to patients discontinuing therapy. In the context of pediatrics, availability of age-appropriate (e.g., oral liquid) formulations with suitable tastes and textures is essential as the medication will not be effective no matter how well the drug and dose have been individualized using genomic data if the child refuses to swallow it. Dr Spielberg concluded by stating that as genomic medicine becomes a reality, it is important that children are not left behind. Integrating a new understanding of disease pathogenesis, and of judicious use of interventions hold enormous promise to prevent morbidity and mortality, and rationally manage chronic illness with enhanced effectiveness and safety for each child, and thus, for children in general.

The next meeting in the series will be held March 30 to April 1, 2011 in Kansas City, MO, USA.