Design and use of mouse control DNA for DNA biomarker extraction and PCR detection from urine: Application for transrenal Mycobacterium tuberculosis DNA detection

Highlights

• Designed/tested spike DNA control for extraction/removal of PCR inhibitors for acellular samples

• Extraction/PCR control DNA performed well and was detectable in 97% of urine samples.

• No differences in transrenal DNA between TB + and TB − retrospective samples were observed.

• TB −/HIV + and TB +/HIV + were statistically different as reported by others.

• Control DNA may improve transrenal DNA detection in clinical samples when included at collection.

Abstract

Urine samples are increasingly used for diagnosing infections including Escherichia coli, Ebola virus, and Zika virus. However, extraction and concentration of nucleic acid biomarkers from urine is necessary for many molecular detection strategies such as polymerase chain reaction (PCR). Since urine samples typically have large volumes with dilute biomarker concentrations making them prone to false negatives, another impediment for urine-based diagnostics is the establishment of appropriate controls particularly to rule out false negatives. In this study, a mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) DNA target was added to retrospectively collected urine samples from tuberculosis (TB)-infected and TB-uninfected patients to indicate extraction of intact DNA and removal of PCR inhibitors from urine samples. We tested this design on surrogate urine samples, retrospective 1 milliliter (mL) urine samples from patients in Lima, Peru and retrospective 5 mL urine samples from patients in Cape Town, South Africa. Extraction/PCR control DNA was detectable in 97% of clinical samples with no statistically significant differences among groups. Despite the inclusion of this control, there was no difference in the amount of TB IS6110 Tr-DNA detected between TB-infected and TB-uninfected groups except for samples from known HIV-infected patients. We found an increase in TB IS6110 Tr-DNA between TB/HIV co-infected patients compared to TB-uninfected/HIV-infected patients (N = 18, p = 0.037). The inclusion of an extraction/PCR control DNA to indicate successful DNA extraction and removal of PCR inhibitors should be easily adaptable as a sample preparation control for other acellular sample types.

Introduction

Biomarkers for infectious diseases may be obtained from several types of patient samples including blood, sputum, stool, and urine. Many factors affect the sample choice, including characteristics of the infection, concentration of the biomarker, volume of sample available, biomarker stability within the sample, and patient willingness to provide the sample. Patient willingness to provide samples and the improved sensitivity of molecular reagents has led to the detection of biomarkers from urine samples being increasingly used for diagnosis of infectious diseases, including Escherichia coli (van der Zee et al., 2016), leptospirosis (Iwasaki et al., 2016), Mycobacterium tuberculosis (Green et al., 2009), Dengue virus (Andries et al., 2015), Zaire Ebola virus (Southern et al., 2015), and Zika virus (Gourinat et al., 2015). Urine samples are particularly advantageous for diagnosing infectious diseases such as tuberculosis (TB), because sample collection is non-invasive, sample volumes are relatively large, and samples are more easily obtained from patients than blood or sputum samples (Green et al., 2009). However, even in the case of DNA excreted in urine, often referred to as trans-renal DNA (Tr-DNA), DNA biomarkers must still be extracted and polymerase chain reaction (PCR) inhibitors must be removed prior to analysis. Green and colleagues identified Tr-DNA extraction as a critical barrier to make TB diagnosis using urine a feasible option (Green et al., 2009). A number of factors contribute to the successful completion of the extraction step. First, most commercially available extraction kits require additional laboratory equipment and trained laboratory personnel, and second, most commercial DNA extraction kits can only accommodate small sample volumes, usually less than 1 milliliter (mL) and variability in the extraction method used affects diagnostic performance (Sarhan et al., 2015). As pointed out by Green et al. (2009), since the presence of TB DNA biomarkers in urine remains controversial, new designs that include indicators of successful sample preparation are needed. Many other types of patient samples typically contain intact cells which serve as a reservoir for endogenous DNA controls. Unfortunately, urine samples contain few intact cells so that the use of endogenous controls from this source is unavailable (Green et al., 2009). We previously reported a relatively simple method to extract and concentrate nucleic acid biomarkers from larger volume patient samples using magnetic beads (Bordelon et al., 2013, Bitting et al., 2016). The use of magnetic bead nucleic acid extraction is advantageous because the processing is self-contained and relies on simple magnetically-induced movement of beads without the need for expensive laboratory equipment. In this report, we describe an approach to address the need for endogenous controls for urine samples based on modifications to our previously reported laboratory studies using surrogate urine samples to detect Tr-DNA sequences of Mycobacterium tuberculosis IS6110 (Aceti et al., 1999, Sechi et al., 1997, Cannas et al., 2008, Englen and Kelley, 2000), a sequence that has been reported to be repeated up to 25 times within the Mycobacterium tuberculosis genome (Green et al., 2009). We describe an improvement to diagnostic design aimed at reducing false negatives due to failure in DNA extraction or PCR failure. To achieve this we add a 120 base pair segment of the mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) DNA to the lyophilized reagents used in our previous method (Bordelon et al., 2013, Bitting et al., 2016) that rehydrates in the patient sample. Detection of this DNA fragment is used to reduce false negatives in two ways: 1) to indicate that DNA was successfully extracted and 2) to indicate that PCR inhibitors have been reduced to a level that they no longer prevent DNA amplification. We report on the performance on this extraction/PCR control and on its application to classify retrospectively collected samples of urine from TB-infected and uninfected patients from Peru and South Africa.