October 15, 2014 (Vol. 34, No. 18)
Richard A. A. Stein M.D., Ph.D.
From our Oct. 15 issue: One sample, one test, many results—this familiar approach is being applied ever more broadly, generating epigenetic, mutational, and signaling profiles.
For many of these assays, collecting multiple datasets over time or under distinct conditions provides distinct advantages and has become an important experimental goal. Consequently, multiplexing has emerged as a strategy to collect multiple parameters using the same sample, and has considerably expanded our ability to analyze and interpret data.
As illustrated by applications from several areas, some of the most significant advantages of multiplexing are reduced costs, increased throughput, and the potential availability of new internal controls. These attributes make multiplexing a valuable experimental resource, indispensible for complex analyses, such as systems biology studies, with the promise to reshape research and clinical applications ranging from the study of signaling pathways to disease management.
“We developed a multiplexing approach based on tandem mass spectrometry to measure the activities of selected enzymes in biological samples,” says Frantisek Turecek, Ph.D., professor of chemistry and Klaus and Mary Ann Saegebarth Endowed Chair at the University of Washington. The assay developed by Dr. Turecek and colleagues detects lysosomal storage diseases. It is performed on dried blood spots that are prepared from blood samples collected at newborn screening centers.
Lysosomal storage diseases comprise a group of metabolic disorders that have a combined prevalence of about 1 in 7,000–8,000 newborn children. These disorders stem from different DNA mutations that affect the activities of specific lysosomal enzymes. When lysosomal enzymes are defective, partially degraded molecules accumulate in the lysosomes. The lysosomes become increasingly dysfunctional, and cellular functions begin to suffer.
The clinical manifestations of this genetically heterogeneous group of diseases depend on the specific enzyme that is affected and, for the same disease, on the mutation that is present. “There are treatments available for many of these diseases, but the individuals are born asymptomatic,” notes Dr. Turecek. “It is impossible to clinically predict whether a newborn will develop any of these diseases.”
Therefore, screening the general population of newborn children is an acceptable approach to identify affected individuals. In the assays that were designed by Dr. Turecek and colleagues, synthetic substrates are converted to readily monitored products. The levels of specific ionized species is used to quantitate enzyme activity.
“In a multiplex fashion, we can determine in one or two rounds the activity of any subset of nine enzymes that we are assaying in this dry blot spot,” asserts Dr. Turecek. Multiplexing is possible because the design of the specific enzyme substrates prevents any overlap between the substrate and product masses for the nine reactions, allowing the entire set of enzymes to be analyzed simultaneously in one analytical run.
This experimental setting has two advantages, emphasizes Dr. Turecek. The first advantage is speed. Backlogs can be avoided even though large numbers of newborns need to be screened, and timely management can be implemented if a lysosomal storage disease is later confirmed.
The other advantage is a dramatic reduction in false-positive results. When enzymes are screened individually, a positive result—low enzymatic activity—may not indicate disease. It may reflect instrumental error or some other extraneous reason. Ruling out false-positives may require additional analyses, costs, and anxiety. However, a low activity concomitantly seen in several of the assay’s nine enzymes is extremely unlikely in the same individual, due to the already low prevalence of each of the individual mutations in the population.
“If several or all nine of these enzymes show low activity or no activity, we immediately suspect that it is a false-positive, and this greatly improves the quality of the analysis when we perform multiplexing,” insists Dr. Turecek. At present, multiplexing is estimated to cost one dollar per analysis to screen for all nine diseases.
Even though enzyme multiplexing is fast, it represents only the first step toward diagnosis. “In individuals with low enzyme activity, this screening step is followed by genotyping using exome sequencing,” elaborates Dr. Turecek. The nature of the mutation often predicts the course and the severity of the specific lysosomal storage disease. It also shapes therapeutic decisions, which sometimes have to be made promptly.
“This strategy provides the most efficient and robust way to screen for enzyme activities,” clarifies Dr. Turecek. “The actual diagnosis relies on different methods, such as DNA sequencing, which are performed after the screen, and these approaches cannot be used to screen large segments of the population.”
“We were interested in profiling a select number of genomic targets to examine epigenetic modifications that shape the heterogeneity of tumors,” says Michael P. Kladde, Ph.D., associate professor of biochemistry and molecular biology at the University of Florida College of Medicine. A distinguishing feature of most cancers is the presence of highly aggressive and therapeutically resistant cellular subpopulations. Capturing the biology of this subset of cells represents, therefore, a key facet of cancer management.
This approach to characterizing cancer, however, is obstructed. Instead of gathering subpopulation-specific data, many studies obtain averaged data from heterogeneous cancer cell populations. “Powerful assays are able to survey histone acetylation and DNA methylation, but it is challenging to tell whether marks that are captured with different assays occur in the same cell or in different cells,” explains Dr. Kladde. “This is what makes it so important to integrate and multiplex different epigenetic features.”
To address this challenge, Dr. Kladde and colleagues developed a high-throughput method for the multiplex detection of DNA methylation and chromatin structure at the level of single molecules. The method, called MAPit-patch, uses multiplexed amplification of targeted sequences from submicrogram quantities of genomic DNA followed by next-generation bisulfite sequencing.
“Multiplexing with MAPit-patch allows the simultaneous determination of two epigenetic modifications at multiple genes of interest from the same biological sample,” says Nancy H. Nabilsi, Ph.D., postdoctoral scientist and first author of the study that introduced this strategy. “It enables researchers to study the effect of an experimental manipulation or drug on an entire biological pathway or network of genes.”
The possibility of integrating detection of multiple epigenetic features at high resolution in a single experiment provides a significant advantage over previously existing approaches. It is, remarks Dr. Kladde, “akin to a multiplexing of multiplexing.”
By analyzing 71 promoters of genes that are frequently mutated in cancer, investigators in Dr. Kladde’s lab identified several promoters that are differentially methylated and/or differentially accessible between glioblastoma and neural stem cell populations, changes that are associated with altered gene expression. The heterogeneity in epigenetic changes across the molecules indicates the existence of epigenetically distinct cellular populations.
MAPit-patch takes advantage of Bisulfite Patch PCR, a technique developed by Rob Mitra and colleagues that allows the highly multiplexed PCR and bisulfite sequencing of many genomic targets. “With relatively small amounts of input material per reaction, such as 100 ng, we were able to process at least 100 targets per tube,” asserts Dr. Kladde.
Chromatin analyses make use of an enzyme that methylates DNA at GpC sites, which are distinct from the CpG dinucleotides that are recognized by endogenous DNA methyltransferases. A key advantage of this approach is that it uses deep sequencing.
“We were able to detect chromatin states that may occur only in small subpopulations of cells, for example, in drug-resistant cancer cells,” details Dr. Nabilsi. “This also enabled us to detect otherwise elusive intermediate chromatin states that can give us insight into the temporospatial coordination of multiple epigenetic events during a biological process.”
One of the shortcomings of the approach is the need to rely on restriction endonucleases. This requirement creates several technical constraints, some of which stem from the existence of repetitive elements and genomic regions in which restriction site targets of interest are sparse or absent. Others stem from differences in the efficiency of various restriction enzymes.
“The most urgent need for improving this technology is overcoming the dependence on restriction enzymes,” advises Dr. Nabilsi. “[Doing so] would allow us to mitigate length bias and increase the diversity and number of genomic regions that can be targeted. It would also give us the flexibility to multiplex a third layer of epigenetic information.”
Researchers have shown great interest in applying targeted pathway inhibitors to treat solid tumors. This approach, however, is fraught with unknowns. “We simply do not know which signaling pathways are activated in a specific patient, how to adequately select patients for these therapies, and how to monitor these patients on these specific inhibitors,” says John A. Sandoval, M.D., a pediatric surgeon at St. Jude Children’s Research Hospital.
In spite of such difficulties, researchers have demonstrated a particular interest in studying deregulated PI3K signaling in malignant tumors. The hope is to design new therapeutic agents against this pathway in solid neoplasms. One such initiative, undertaken by Dr. Sandoval and colleagues, involves the development of a multiplexing-based approach to profile the PI3K pathway in neuroblastoma, a malignancy that is common in infants and children.
Dr. Sandoval and colleagues have reported using the Luminex xMAP bead array technology in a panel of neuroblastoma cell lines to test four PI3K inhibitors. By pharmacologically disturbing PI3K signaling at specific regulatory nodes, the investigators proposed to characterize therapeutically relevant PI3K phosphoprotein signatures. Ultimately, Dr. Sandoval and colleagues integrated the phosphoprotein profiles to create biomarkers that can be used to interrogate the pathways.
“This allowed us to generate profiles of the signaling that occurs in response to inhibitors of interest, and we were able to quickly look at single agents and also combine several inhibitors to study how they are able to further disturb the pathway,” asserts Dr. Sandoval. Additional studies confirmed that the results correlate well with the ones obtained using previously established approaches that are technically more laborious, such as Western blotting.
The high activation of basal PI3K pathway may be able to predict the patients who would preferentially respond to specific compounds or combinations, as compared to those with low activation of the pathway, who would not be good therapeutic candidates for the same compounds. “We were able to show that we can establish a foundation from where we can move this forward to study this pathway in preclinical trials,” explains Dr. Sandoval.
MicroRNAs in Context
“A main focus of the lab is to determine if specific microRNAs can be used as tissue biomarkers in breast cancer and other cancer types,” says Lorenzo F. Sempere, Ph.D., assistant professor and head of the laboratory of microRNA diagnostics and therapeutics at Van Andel Research Institute. In recent years, research has increasingly elucidated the role of microRNAs during cancer initiation, development, invasion, and metastasis. MicroRNAs, it has been found, help shape disease course and the response to treatment.
Multiple cell types are involved in the architecture of solid tumors. Accordingly, researchers are trying to identify the specific cell types from the tumor microenvironment where microRNA alterations occur.
In the years following the discovery of microRNAs, biological material containing a mixture of multiple cell types has been used to profile microRNAs from solid tumors. Although this reveals perturbations that are present at the level of the entire sample, it is not informative about the specific cell types that are affected.
To circumvent this problem, Dr. Sempere and colleagues took advantage of fluorescence-based protocols to detect microRNAs and, at the same time, co-detect reference RNA and protein markers. “We examined samples from almost 900 patients with breast cancer and showed that in triple negative breast cancer—an aggressive subtype that lacks targeted therapy—there was higher expression of miR-21 in the stroma, but not in the cancer cells, which was correlated with a much poorer outcome,” notes Dr. Sempere.
The strategy can be applied for other microRNAs and other malignancies; however, interpreting the imaging data can be a challenge. “Multispectral imaging platforms enable us to separate the different fluorophores and scan the whole tissue section in the research setting,” explains Dr. Sempere. Still, it is a challenge knowing “how to quantitate and weigh in contextual information from different tumor regions to generate a prognostic score.”
Screening Multiple Suspended Objects
“Having a true screening system for suspended cells, beads, and microbes—and mixtures of any and all such objects—is a tool that has really been missing from the drug discovery toolbox,” says Joe Zock, senior director, product management, IntelliCyt. Even though flow cytometry has been available for several years, its use in imaging has outpaced its use in high-content screening. Similarly, microscopy did not serve as a screening system until, in Zock’s words, “a group of us transformed it, and ‘high content’ was born.”
IntelliCyt developed a multiplexable screening platform by joining a flow cytometry detection engine with a rapid sampling technology licensed from the University of New Mexico. The platform, known as the iQue® Screener, can acquire and analyze very small volumes, such as 1 μL samples that can be retrieved from microtiter plate wells.
Cells, beads, microbes, or other materials to be sampled are continuously transferred to the detector in air-gap-delimited “packets,” and multiple readouts are collected from every object. “Six measurements are collected simultaneously per object, at up to 10,000 objects per second, allowing similar subpopulations to be identified and multiple targets to be assessed in each subpopulation,” details Zock.
The iQue Screener is able to detect objects based on size and granularity, separate populations of objects, and quantitate internal or surface fluorescent tags or dyes. “We can complete 96-well plates in as little as 3 minutes and 384-well plates in 12 minutes,” notes Zock. “The iQue Screener HD for high density can process 1,536-well plates in 1 to 1.5 hours.”
The iQue Screener is suitable for antibody screening, phenotypic drug discovery, and in vitro toxicity testing. Historically, antibody screening has been performed using ELISA, which is often not multiplexed, or flow cytometry, which is slow and complex to set up for routine use. By overcoming these limitations, IntelliCyt provides a way to screen with conformationally intact antigens presented on cell surfaces, which improves epitope presentation, and to screen multiple populations per well, to access specificity.
The platform, Zock says can multiplex in three dimensions: “First, we use label-free parameters to isolate (group) objects by relative size and density. We can also multiplex at specific points (channels) across the visible spectrum for each object. However, because of the sensitivity we get from interrogating objects one at a time in the stream, we can also perform vertical or concentration-based multiplexing. We can take a single channel and dye cells or objects with multiple different concentrations to isolate them.”
The iQue Screener can provide physiologically relevant results because it can perform measurements in the natural environment of the objects that are being studied.
“It can measure primary immune cell proliferation, isolate specific subtypes by immunophenotyping, and capture a secreted cytokine profile with encoded capture beads … all in the same well at the same time,” asserts Zock. “Interrogating a cellular microenviroment at this level can help to uncover the next generation of immunomodulation therapies to stop cancer or control neural inflammation.”