Research Highlights

Microfluidics and personalized cancer therapy

Evaluation of: Komen J, Wolbers F, Franke HR, Andersson H, Vermes I, van den Berg A: Viability analysis and apoptosis induction of breast cancer cells in a microfluidic device: effect of cytostatic drugs. Biomed. Microdevices 10(5), 727–737 (2008).

– Sao Jiralerspong & Ana M Gonzalez-Angulo

Cancer medicine is undergoing a modern revolution. Treatment with broadly cytotoxic chemotherapy is giving way to the use of targeted therapies that have a rational molecular target and promise increased efficacy with reduced toxicity. This has naturally given rise to the concept of personalized therapy, the selection of treatments based on the unique molecular characteristics of each individual patient and their tumor. However, at present, the full promise of personalized cancer medicine is far from a reality.

One hurdle that must be overcome is the development of diagnostic tools to identify patients who will respond to a particular therapy. Microfluidic chips are one possible answer to this obstacle. Microfluidic chips [1–3] are devices equipped with micrometer-sized channels that allow for the handling of sub-microliter volumes of liquid. The liquid streams can be fashioned serially to allow for the processing of cells or subcellular components (proteins, nucleic acids and so on) for the purposes of culture, lysis, separation, isolation and molecular analysis. Treatments can be administered to cells prior to further processing and analysis. Parallel processing of samples permits multiple analyses of similar samples, as well as the application of rigorous statistics. Advantages of miniaturization include the possibility of high-throughput analysis and automation. A potentially enormous advantage for cancer diagnostics would be the use of much smaller amounts of patient-derived tumor tissues, which could then be subjected to multiple analyses for detailed molecular characterization.

The research presented in the article by Komen et al. represents an early positive step towards the goal of using microfluidics for personalized cancer therapy [4]. They describe the design and use of microfluidic chips made of the polymer poly(dimethylsiloxane) and glass to culture a well-known breast cancer cell line (MCF-7) for up to 7 days. They then determined the sensitivity of these ‘cells on a chip‘ to the protein kinase C inhibitor and apoptosis-inducer staurosporine via fluorescence-based cell viability assays. They demonstrated that exposure of the MCF-7 cells cultured on their chip to a 30-min flow of staurosporine followed by a 24 h static incubation resulted in decreased viability compared with control cells exposed to medium alone.

As the authors acknowledge, more work remains to be done before this particular microfluidic system makes it to the clinic for routine diagnostic use. The preliminary results obtained with MCF-7 and staurosporine need to be repeated in this and other cell lines to confirm reproducibility, and more clinically relevant therapeutic agents, such as anthracyclines, taxanes and tamoxifen, need to be tested. In addition, an effort should be made to correlate the microchip results with established in vitro chemosensitivity tests utilizing conventional cell-culture techniques and cell-viability assays. Next, it would be essential to apply the system to the culture and chemosensitivity testing of cells from patient-derived tumor samples. Finally, the correlation of this ex vivo chemosensitivity test with clinical response would need to be validated in a prospective clinical trial.

In a proof-of-concept study from a different group, the authors used a microfluidic chip to isolate circulating tumor cells from nonsmall-cell lung cancer patients in sufficient quantity and purity to allow genetic determination of EGFR mutations, and to demonstrate a correlation between mutation status and clinical outcome [5]. Thus, the era of microfluidics for personalized cancer therapy may not be far off.

Critical Advancements to Linking Disease Phenotypes with Human Genomic Variations

Evaluation of: Roden D, Pulley J, Basford M et al.: Development of a large-scale de-identified DNA biobank to enable personalized medicine. Clin. Pharmacol. Ther. (2008) (Epub ahead of Print).

– Michael D Kane

The emergence of clinical genotyping as an enabling component of personalized medicine in healthcare has been hindered by many independent issues, including privacy concerns from healthcare consumers and the limited resources available to rigorously derive linkages between genomic markers and physiological traits that are relevant to human health [1]. The utilization of genomic markers or SNPs to provide guidance in drug safety and efficacy assurance has received much attention, owing to the fact that the pharmacokinetic effects of SNPs discovered in genes encoding oxidative drug-metabolizing enzymes (i.e., CYP enzymes) have been relatively simple to characterize. Furthermore, the ability to reduce the incidence and severity of adverse drug responses represents a viable value proposition for the commercialization of clinical genotyping methods, where screening for SNPs that affect the expression level or activity of drug-metabolizing enzymes can be used to provide specific guidance on drug dose and/or drug selection in the clinic. However, the ability to extend clinical genotyping beyond this pharmacogenomic application into disease predisposition requires the ability to link known SNPs with specific disease traits and clinical disorders in a large population, ultimately to discover new linkages between SNPs and the risk of disease development [2]. The derivation of large numbers of human health records, each with the respective DNA sample available for SNP screening, represents a resource-intensive undertaking, particularly when each health record must be altered to respect the privacy of each patient by removing all direct and indirect identifiers.

Roden et al. have published a report on work specifically aimed at the derivation of study samples that include both the health history of the patient and a blood sample available for SNP screening [3]. Ultimately, this work represents a critical advancement in the ability to link SNPs with specific disease traits using clinical samples. Paramount to this work are the methods for removing patient identifiers from electronic health records to facilitate patient consent for DNA sample and health record submission into their biobank system. The highlights of this work include a modified method for deriving patient consent, replication and transformation of the electronic medical record (EMR) using algorithms to remove patient identifiers, and the isolation of DNA from discarded blood samples that are linked to the derived (anonymous) EMR.

The methods utilized for patient consent are described by Roden et al. as an ‘opt-out‘ model, which is based on earlier findings from patient surveys regarding the utilization of patient-specific DNA samples for basic research [4]. In this ‘opt-out‘ model, the biobank study was described to the patient representing changes in the standard consent to treatment form, which included a permanent opt-out option for the patient, and the opt-out rate for the biobank study is reported at approximately 2.5%. Once a patient was admitted into the study, their EMR was subject to a de-identification algorithm to remove patient identifiers and create a ‘synthetic derivative‘ of the EMR. Sample collection was initiated in February 2007 and the study has enjoyed a sample accrual rate of 700–900 samples per week, with a total of 33,463 samples as of April 2008. The most common diagnoses in the sample pool include hypertension (15.7%), Type 2 diabetes (11.8%), hyperlipidemia (11.5%), coronary artery disease (7.8%) and anemia (5.9%). This biobank resource represents an important step towards the establishment of genetic linkages to these disease phenotypes.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.