December 1, 2016 (Vol. 36, No. 21)

Challenging Biobank Samples Are Evaluated in Parasite-Induced Cholangiocarcinoma Study

Circulating biomarkers have the potential to form the basis of affordable diagnostic tests for early detection and monitoring of diseases. Such diagnostic tests allow early intervention and treatment appropriate for the disease stage.

MicroRNAs (miRNAs), small noncoding RNAs of ~22 nucleotides related to post-transcriptional expression regulation, have become a significant target of biomarker research. miRNAs are found in nearly all sample matrices and their presence in accessible biofluids such as blood, urine, and saliva makes sample collection straightforward and for some matrices—noninvasive. In addition, given their small size, miRNAs are particularly stable in these biofluids, aiding storage and transportation of patient samples. 

Challenges Associated with miRNA Analysis

Sample preparation and RNA purification is a critical barrier to miRNA analysis: current methods are not only prone to human error, but have profound effects on RNA yield and sample throughput. Additionally, only a small amount of each patient sample is often available for retrospective studies, having been biobanked prior to study conception.

These limitations on miRNA yield present a major drawback for conventional techniques such as microarrays and RNA sequencing, where micrograms of RNA may be required when only nanograms of sample exist. In this respect, researchers often have to exclude cohorts or pool samples, limiting the depth of data garnered.

Although qPCR is a more sensitive technique that requires less miRNA than other techniques, another challenge is often presented for plasma samples: the presence of heparin, an anticoagulant often used for blood collection that can inhibit the reverse transcription step. Consequently, there is a practical need for high-throughput miRNA assays that can be used with small sample volumes, direct from crude biofluids.

Biomarker Potential for Parasite-Induced Cholangiocarcinoma

Cholangiocarcinoma (CCA) is a parasite-induced cancer of the bile ducts that is associated with infection by the food-borne liver fluke Opisthorchis viverrini (OV). The parasite, considered a group 1 carcinogen, is endemic to Thailand, Vietnam, and Cambodia, presenting a significant public health concern due to the consumption of raw and undercooked fish. After consumption, the parasite enters the biliary tract where it causes chronic inflammation, leading to advanced periductal fibrosis (APF) and eventually CCA.

Currently, diagnosis and monitoring is completed by ultrasound or liver resection, but the limited access to healthcare in the remote tropical regions where the disease prevails means that CCA is often diagnosed at a late state when morbidity and mortality are high. Identification of accessible and durable biomarkers in crude body fluids would greatly aid early diagnosis and treatment of CCA.

Research at The George Washington University has focused on identifying CCA biomarkers, including changes in miRNA expression at different stages of OV-induced CCA. However, this research involves working with challenging samples: the research makes use of longitudinal sample sets from large cohorts of patients with APF and CCA, but sample volume is often limited to less than 200 µL and samples consist of a range of biofluids, including serum, plasma, saliva and urine.

Testing Firefly Multiplex Assays in Biofluids and Assessing Heparin Interference

Abcam’s Firefly multiplex particle technology, which uses labeled hydrogel particles to bind target miRNAs from crude biofluids, was used for biomarker validation in this challenging sample set. After identifying preliminary miRNA signatures from both tumor tissue (formalin-fixed, paraffin-embedded) and serum, exploring the larger sample cohorts required a platform that minimized sample volume, increased throughput, and targeted multiple miRNAs.

The Firefly multiplex technology overcomes previous limitations by allowing simultaneous measurement of up to 68 miRNAs in a single well, using as little as 10 µL crude biofluid sample, with or without RNA purification.

Preliminary studies aimed to see if miRNAs could be detected directly and consistently across sample matrices, including heparin-plasma. Initially, nonstudy samples from four individuals were used to determine whether heparin interfered with assay readout. This experiment used a pre-designed miRNA focus panel consisting of 68 miRNAs to analyze 35 µL of nonstudy plasma and serum samples collected in tubes containing sodium heparin, lithium heparin, sodium citrate, and heparinase.

The results revealed that heparin had no effect on the assay, with well-correlated data obtained from tubes containing heparin and those not containing heparin (Figure 1). This removed the need to purify RNAs or pre-treat samples, and allowed full use of the extensive cohort (composed of both plasma and sera) of available samples available for this study.

Data from the initial study also showed a strong correlation between different sample matrices taken from the same individual, meaning that data from serum could be directly compared to plasma. This meant that the entire comprehensive biorepository of study samples could be used.

Figure 1. (A) Heat map output from sample matrices using a predesigned Firefly multiplex panel. Samples from four individuals (B, G, P, R) were assessed on the panel, including sera and plasma derived from sodium citrate (Na-C), lithium heparin (Li-Hep), and sodium heparin (Na-Hep). (B) Correlation between sample types used within the assay. Pearson’s correlation values of 0.90–0.99 were obtained across different sample matrices (e.g. sera vs. lithium heparin) from the same individual.

Using the Assay in CCA Cancer Patient Samples

Having established that the protocol and technology could successfully detect miRNAs in serum and plasma samples, including those containing heparin, the next step was to use the assays to assess patient samples from the repository. A custom miRNA panel was developed containing 50 candidate miRNAs identified from previous work on OV-induced CCA, as well as additional control and normalization miRNAs.

Paired serum, urine, and saliva samples from a cohort of approximately 50 individuals were tested and analyzed amongst three groups at different clinical stages of the disease: 1) OV endemic (individuals who are OV positive, but do not have any signs of cancer), 2) advanced periductal fibrosis (individuals with middle clinical progression or APF), and 3) CCA. In addition, a further plasma cohort of more than 200 individual samples was analyzed, comprising controls and individuals diagnosed with CCA classified according to well-differentiated, moderately differentiated, or papillary carcinoma.

Out of the 50 target miRNAs included in the custom panel, 48 were detected above the limit of detection in all study groups and sample types (Figure 2), highlighting the suitability of the assay for quantifying miRNAs in saliva, urine, serum, and plasma samples from the biorepository. The assay also detected miRNAs that were differentially regulated between different stages of the disease (Figure 3), demonstrating the potential for developing biomarker signatures to aid stage-specific diagnosis and treatment of OV-induced CCA.

Figure 2. Visualization of miRNAs detected across the CCA study sample. A customized Firefly panel composed of ~50 target miRNAs previously determined to be associated with CCA and additional controls for normalization was used to analyze study samples of various matrices and defined study groups. Green-highlighted miRNAs were detected above threshold values for all study samples, and the number of differentially expressed miRNAs between disease stages is shown.


Although miRNAs have the potential to be ideal biomarkers for diseases such as OV-induced CCA, their analyses can be made challenging by complex signatures, limited sample volumes, and errors introduced during sample processing. An advanced multiplex particle technology was used to successfully facilitate miRNA analysis direct from biofluids with sample volumes of as little as 10 µL.

Applying the technology to CCA patient samples demonstrated that a multiplex miRNA signature could be quantified from a range of biofluids, and that the presence of heparin in plasma samples did not interfere with the assay.

Specific miRNAs that are differentially expressed between disease stages were identified, highlighting the potential of miRNA biomarkers for stage-specific clinical diagnosis of OV-induced CCA and the focus of ongoing work and analysis.

Given the prominence of many of the miRNAs investigated here associated with other cancers, it is understandable that a signature of multiple miRNAs may be necessary to specifically identify CCA, emphasizing the importance of carrying multiplex signatures into large sample cohorts for verification.

Figure 3. Box-plot visualization of selected miRNAs from saliva (left panel) and sera (right panel). OV-positive endemic sera (non-CCA), followed by intermediary disease progression (APF), and two cohorts of CCA positive patients were compared for differential miRNA expression. In the analysis of sera, two additional groups of nonendemic, OV-negative sera were used as additional controls.

Jordan Plieskatt is senior research associate at The George Washington University, and Daniel Pregibon ([email protected]) is general manager of platform innovation at Abcam.

Previous articleSmelly Camel Urine Finding Could Lead to Novel Therapies for Sleeping Sickness
Next articleModified PEG Drug Delivery Method Evades Immune System Attack