NEUROSCIENCE PERSPECTIVES Application and Optimization of Microarray Technologies for Human Postmortem Brain Studies Margaret M. Ryan and Stephen J. Huffaker, Maree J. Webster, Matt Wayland, Tom Freeman, and Sabine Bahn A number of microarray investigations using human postmortem brain tissue have been published recently, exploring a multitude of human brain disorders with the aim of unraveling the underlying pathologies. Although the technology is still developing and lacks sufficient sensitivity with regard to detecting splice variants and low abundance transcripts, microarrays are becoming the prominent method for candidate gene screening in complex neuropsychiatric disorders. The use of postmortem tissue harbors a variety of potential pitfalls, however, which could result in unreliable or, at worst, meaningless results. During the course of our large-scale gene expression study on 150 human postmortem brain samples, using more than 200 Affymetrix GeneChips, we have identified several aspects within microarray experimental procedure that allows for the early identification of potentially unreliable samples. The general application of the guidelines and technical tips described here increase the efficiency, reliability, and amount of data generated by this powerful screening technology while reducing superfluous consumption of time and resources. Key Words: Microarray, Affymetrix, postmortem, human brain, quality control M icroarray investigations have rapidly become a popular method for global gene expression studies, already proving invaluable in profiling expression patterns in a multitude of organisms and human diseases. Additionally, a number of microarray experiments using RNA from postmortem human brain tissue have been published including studies exploring expression profiles of complex neuropsychiatric dis- orders such as schizophrenia and bipolar affective disorder (Hakak et al 2001; Mirnics et al 2000; Pongrac et al 2002; Vawter et al 2001). It is now known that there are far fewer genes in the human genome than was originally thought, with current estimates ranging between 30,000 – 40,000 (Lander et al 2001). Even though a functional understanding for many of these genes is still lacking, few will doubt that the completion of the Human Genome Sequencing Project (HGP) marks the beginning of a new era in molecular research. The HGP has already had an immense impact on the development of global profiling technol- ogies and is already transforming the field of medicine with respect to disease understanding, diagnosis, and therapeutic interventions. Here we present some of the technical approaches we have adapted for microarray use on human postmortem brain tissue and discuss the potential pitfalls and experimental variations that can occur when using postmortem RNA in conjunction with GeneChips. Through the course of our extensive gene expres- sion analysis on 150 postmortem brain samples, using more than 200 GeneChips in the process, we have generated a quality control procedure with which it is possible to identify samples as outliers early on in the experimental process, before further time and resources are wasted on unsuitable samples. This article first concentrates on general technical observations and advice that affect the efficiency and variability of sample preparations. Second, we present explanations of quality control (QC) steps and screens for several stages of sample preparation to obtain high-quality and meaningful results from microarray investiga- tions. Finally, we briefly discuss the limitations of current Gene- Chip technology and highlight some future applications of microarrays for research into the pathomechanisms of neuropsy- chiatric disorders. Although our investigations and optimizations focus on the use of Affymetrix (Santa Clara, California) Gene- Chips, these QC methods are broadly applicable to all microarray technologies. It is our hope that other research groups using Affymetrix GeneChips, microarrays, and perhaps even other high-throughput transcriptome screening techniques for post- mortem studies can use these findings and techniques as a means of reducing expenditures while increasing the reliability and comparability of expression profiling data. The Basics of Affymetrix GeneChip Methodology The Affymetrix Human Genome U133 (HG-U133) set repre- sents approximately 39,000 transcripts including 33,000 fully annotated genes represented on two GeneChips, the HG-U133A and B chips. Probe-sets consist of 11 probe-pairs typically designed with a bias toward the 3'-end of each gene. These probe-pairs are distributed across the array and consist of a perfect match and a mismatch oligonucleotide (each averaging 25 base pairs in length). The mismatch oligonucleotide contains a single base pair mismatch in the center of the probe and is used to quantify and subtract nonspecific hybridization and back- ground signals. Each chip also includes Escherichia coli internal spike control probes (bioB, bioC, bioD, cre) to determine hybridization efficiency. Affymetrix also offers a Test3 chip that contains probe-sets for genes from 23 organisms including eubacteria, plants, and mammals (www.Affymetrix.com). These probes are constructed in the same manner as those on the HG-U133 set and allow for relatively inexpensive sample screening for various aspects of chip quality to estimate their performance on a GeneChip expression array. Our standard operating procedure was adapted from the Affymetrix recommended protocol (Figure 1). In brief, total RNA was isolated from 50 –100 mg of prefrontal cortex tissue from human postmortem brains using Trizol (Invitrogen, Carlsbad, From the Department of Neurobiology (MMR, SJH, SB), Babraham Institute, Cambridge, United Kingdom; Department of Psychiatry (MJW), Stanley Laboratory of Brain Research, Uniformed Services University for the Health Sciences, Bethesda, Maryland; Medical Research Council (MV, TF), Rosalind Franklin Centre for Genomics Research, Wellcome Trust Ge- nome Campus, Hinxton, Cambridge, United Kingdom. Address reprint requests to Dr. Sabine Bahn, Department of Neurobiology, Babraham Institute, Babraham, Cambridge CB2 4AT, United Kingdom. Received March 28, 2003; revised August 26, 2003; accepted October 24, 2003. BIOL PSYCHIATRY 2004;55:329 –336 0006-3223/04/$30.00 doi:10.1016/j.biopsych.2003.10.016 © 2004 Society of Biological Psychiatry