Rapid nucleic acid extraction from skin biopsies using a point-of-care device

Abstract

Sample processing is often the rate-limiting step for point-of-care nucleic acid testing, especially for large, robust tissues such as skin biopsies, which can be used to diagnose a variety of dermatological diseases. Extraction of nucleic acids from these samples often relies on lengthy enzymatic digestions, increasing the time to result and reducing the potential impact of rapid molecular diagnostic approaches. To address this, we have developed BLENDER, a device for rapid nucleic acid extraction from tissue biopsies that combines bead-beating homogenization with simultaneous sample heating for enzymatic lysis. Our device can produce a complete DNA yield from a 3mm cylindrical skin biopsy with only a 15-minute extraction compared to 4 hours when using a commercially available extraction protocol. Decreasing sample-processing time for tissue biopsies could reduce time-to-result for downstream analysis, enabling faster point-of-care diagnosis of solid cancers in limited resource settings.

Graphical Abstract

BLENDER is a point-of-care device for rapid sample processing of large, robust tissue such as skin punch biopsies.

1. Introduction

Nucleic acid testing (NAT) is becoming increasingly important for disease diagnosis, especially in limited-resource settings. Point-of-care (POC) diagnostic approaches can provide improved access for patients where decentralized healthcare is common, as well as reduced time-to-result compared to traditional testing methods^1–3. A variety of dermatological diseases can be diagnosed with NAT^{4–6 7}, however, sample processing of large skin punch biopsies can add hours to an otherwise rapid molecular diagnostic approach⁸. Most recently developed POC nucleic acid extraction technologies largely focus on simpler sample matrices and have no carryover to large, robust samples like skin biopsies^9–12.

Homogenization is a mechanical lysis technique that can be used to fully disperse robust sample tissue and reduce DNA extraction times^{13, 14}. Available laboratory techniques include grinding with a pestle and mortar^{15, 16} (usually after flash-freezing with liquid nitrogen), or homogenization using a rotor stator device. However, these are not feasible for POC applications due to the equipment and safety requirements for use, including sterilization before/after each sample and open-air operation. Bead-beating is another homogenization technique that mechanically disperses the tissue sample through vigorous shaking of the tissue and a lysing bead, and has been shown to produce equivalent DNA to other homogenization techniques¹⁷. There are clear benefits for POC use, as bead-beating can be accomplished using inexpensive disposable tubes and beads, as well as being fully contained – meaning samples are safely within a closed tube during homogenization unlike alternative methods mentioned previously. However, commercially available bead-beating instruments are not suitable for POC use in limited resource settings due to high cost, power requirements, and lack of portability. Additionally, commercial bead-beaters – as well as other homogenization solutions – do not have the ability to keep samples at temperatures suitable for enzymatic digestion, requiring an additional heating device for complete DNA extraction. Innovations for POC bead-beating largely focus on pathogenic detection^18–20, such as an audio-powered bacterial lysis device used for tuberculosis²¹. However, none of these approaches is viable for larger samples, especially “tough” tissue types like skin that have increased amounts of connective proteins. A POC approach for rapid DNA extraction from skin was developed for the detection of Leishmania bacteria²². However, this approach requires manual disruption of the tissue sample during heated lysis using a toothpick, which can incur additional safety hazards during processing.

In this paper, we present the development of a device called BLENDER to perform Bead-beating Lysis with Enzymes for Nucleic aciD ExtRactions. BLENDER is a POC-accessible device for the homogenization and complete digestion of large tissue samples, such as skin biopsies, for downstream NAT and molecular analysis. Utilizing simultaneous bead-beating and enzymatic lysis methods, BLENDER is capable of rapid nucleic acid extraction from 3mm skin biopsies in only 15 minutes, whereas traditional enzyme-only digestions can take 4 hours or longer. We first discuss the mechanical features and construction of the device, as well as bead selection for skin tissue homogenization. We then demonstrate the effectiveness of BLENDER by producing complete DNA yields with a rapid 15-minute extraction, comparable to a 4-hour long extraction using a commercially available protocol. Finally, we test extracted DNA from both methods using qPCR and show equivalent amplification performance for downstream testing. Rapid sample processing could decrease the time-to-result for many NAT technologies and increase the impact of molecular diagnostics in limited resource settings.

2. Materials and Methods

2.1. BLENDER construction and characterization

BLENDER is a small, benchtop device for DNA extraction from tissue biopsies, coupling homogenization via bead-beating with simultaneous enzymatic digestion at 56–65°C [Figure 1A]. The device utilizes a 3D-printed ABS chamber with an acrylic front panel as the sample processing chamber, while the accompanying electronics and motor are kept in a separate chamber behind [Figure 1B]. The sample-processing chamber has a removable lid and front acrylic for easily inserting samples, yet completely sealing off the chamber while DNA extraction occurs [Figure 1C]. The processing chamber includes a tube holder attached to a Scotch yoke mechanism for homogenization, as well as a non-contact aluminum heat block for simultaneous enzymatic digestion of the tissue. The motor-driven Scotch yoke mechanism utilizes a 3D printed frame with inset linear bearings on both sides and an aluminum guide in the center [Figure 1D]. BLENDER utilizes a high-RPM motor with encoder for precise high- or low-RPM actuation of the Scotch yoke. The motor attaches to an aluminum wheel with a brass follower – this follower post is trapped within the aluminum guide of the sample motor, though it can spin freely and move laterally within the guide. As the motor spins, the Scotch yoke translates circular motion up to 3000RPM into linear motion up to 50Hz. Concurrent with physical lysis through homogenization, samples also undergo enzymatic lysis using the non-contact aluminum heat block at the bottom of the device [Figure 1E]. With a target range of 56–65°C for optimal enzyme activity, BLENDER reaches the minimum temperature of 56°C in only 1 minute 30 seconds and the maximum target temperature of 65°C in 3 minutes 30 seconds. With the heat block at operational temperature, the sample tube can then be inserted. Once inserted and lowered to the bottom position, 56°C is reached in 2 minutes and 65°C is reached in just over 4 minutes.

Figure 1 – BLENDER construction and operation.

(A) BLENDER is a small, portable device for rapid DNA extraction from robust tissue biopsies for downstream diagnostic testing. (B) This device features two separate compartments: a front compartment for sample processing and a rear compartment to house the motor and other electronics. This makes it relatively simple to clean when necessary and protects the electronics while keeping them accessible. (C) The sample processing unit features a Scotch yoke tube holder and a non-contact aluminum heat block for simultaneous bead-beating and enzymatic digestion. This compartment also has a removable lid and acrylic front cover for easily inserting and monitoring samples while keeping any potential biohazards contained. (D) The sample holder is 3D printed using ABS and contains linear bearings for reduced-friction movement along the two linear rails. The motor-driven aluminum wheel of the Scotch yoke mechanism has a brass follower that is trapped within an aluminum guide attached to the 3D printed tube holder. As the motor-driven wheel spins, the Scotch yoke translates the circular motion into linear motion for sufficient bead-beating performance. (E) BLENDER uses simultaneous enzymatic digestion during the rapid DNA extraction, which requires temperatures between 56–65°C. From room temperature, the minimum target temperature is reached within 1.5 minutes and the maximum is reached 2 minutes later. Once the BLENDER is at temperature, an inserted sample will reach the minimum temperature of 56°C within 2 minutes and will reach the maximum target temperature of 65°C in just over 4 minutes.

2.2. Reagents and equipment

Standard lysis of skin biopsies was performed using the QIAGEN DNeasy Blood & Tissue kit (#69504, QIAGEN). Bead-beating required an additional anti-foaming Reagent DX (#19088, QIAGEN). Bead-beating was accomplished utilizing sterile 2ml screw-cap tubes and various homogenization inserts – garnet matrix (#11079110gar, BioSpec), zirconia beads(#11079124zx, BioSpec), small (#96415K68, McMaster-Carr) and large (#96415K75, McMaster-Carr) steel beads. Electronics used for homogenization include an Arduino Uno, a Pololu G2 simple motor controller (#1363, Pololu), and a ServoCity Yellowjacket 5201 motor (#5000-0002-4008, ServoCity). The motor was mounted using M3 fasteners and motor mounts from ServoCity. Electronics used for heating and temperature control include the above-mentioned Arduino Uno, two k-type thermocouples with accompanying MAX6675 temperature sensors (SainSmart), a 2” cartridge heater (#MCH1-240W-004, Comstat Inc.), a 12V/6A power adapter (SGA60U12-P1J, DigiKey), and a 5V relay (#3-01-0340, HiLetGo). Operation of the BLENDER utilizes four 3/16” diameter linear bearings (#6489K71, McMaster-Carr) that move along two 4” long 3/16” diameter linear shaft (#1162K118, McMaster-Carr). Remaining components of the device were 3D printed in-house using a Lulzbot 2.0 printer and ABS filament.

2.3. Skin biopsy sample collection

A single, large piece of skin tissue (~15in²) was provided by a collaborator at Weill Cornell Medical College in New York, NY. This singular sample tissue was used for all extraction experiments herein, in an attempt to keep DNA yields between biopsies as consistent as possible. The skin tissue was kept at −20°C except for when removing additional biopsies for analysis. All biopsies used in this paper were taken using a standard 3mm punch biopsy tool and weighed immediately. 3mm punch biopsies were chosen as they are roughly half the volume of a standard clinical 5mm punch biopsy, where a biopsy could be bisected with one half for continuing standard-of-care and one half for molecular analysis. Experiments were performed over multiple months, so ~8 biopsies were taken at a time to minimize potential freeze-thaw effects on the sample DNA.

2.4. DNA extraction using the DNeasy kit

DNA extraction was accomplished using the QIAGEN DNeasy Blood & Tissue kit. For non-homogenized samples, pre-weighed 3mm biopsies were added to a screw-cap tube. Lysis reagents from the QIAGEN kit were then added, including 180 μL of Buffer ATL and 20 μL of Proteinase K. For samples that would undergo bead-beating homogenization, a single 1/4” corrosion-resistant stainless steel bearing was inserted into the tube, and 10 μL of Reagent DX was added to prevent excessive foaming. BLENDER was turned on and allowed to reach operational temperature prior to sample-tube insertion, where a 30-second homogenization step at 3000 RPM was followed by 14.5 minutes of enzymatic digestion, for a total of 15 minutes. For non-homogenized samples, tubes were added to a heat block kept at 56°C and vortexed every 30 minutes, per QIAGEN’s recommendations.

2.5. DNA purification using the DNeasy kit

For all samples, extracted DNA was purified using the QIAGEN DNeasy Blood & Tissue kit. Lysate was allowed to cool to room temperature following enzymatic digestion. 200 μL of Buffer AL and 200 μL of 200-proof Ethanol were added to the lysate and mixed by vortexing before being pipetted into a DNeasy spin-column. Samples were centrifuged at 6,000xG for 1 minute to remove unwanted cellular debris. Spin columns were moved to a new collection tube and flow-through was discarded. 500 μL of Buffer AW1 wash buffer was added, samples were centrifuged for 6,000xG for 1 minute, spin-columns were moved to a new collection tube, and flow-through was discarded. 500 μL of Buffer AW2 wash buffer was added, samples were centrifuged for 17,000xG for 3 minutes, spin-columns were moved to a sterile microcentrifuge tube, and flow-through was discarded. For the elution step, 75 μL of Buffer AE was added and let sit on the membrane for 1 minute. Samples were then centrifuged at 6,000xG for 1 minute to elute the purified DNA. Eluted DNA samples were stored at −20°C when not undergoing additional testing.

2.6. DNA analysis using spectrophotometer and qPCR

DNA analysis was done using a SpectraMax QuickDrop micro-volume spectrophotometer. The device was cleaned using DI water and blanked using Buffer AE as a reference. For each sample, 2.0 μL of sample was added, measured for DNA yield, and cleaned off using a Kimtech wipe. Each DNA sample was measured in triplicate using 3 separate 2.0 μL drops and the DNA concentration was averaged. Additional analysis was performed on some samples using the QuantStudio 7 commercial qPCR device. A GAPDH assay was performed with a total reaction volume of 10 μL, including: 5 μL of TaqMan Genotyping Master Mix (ThermoScientific, 4371355), 0.5 μL of a 20x GAPDH TaqMan Copy Number Assay (ThermoScientific, 4400292-Hs00483111_cn), and 4.5 μL of sample. All sample DNA analyzed with qPCR was diluted to a standard 9ng total DNA for each reaction. The reaction cycle proceeded with an initial hold at 50°C for 2 minutes, a hold at 95°C for 10 minutes, and then 40 cycles between 95°C for 15 seconds and 60°C for 60 seconds. All samples were run in duplicate against a standard plasmid curve.

3. Results and Discussion

3.1. Bead-beating optimization for skin tissue biopsies

Bead-beating is a homogenization approach to fully disperse a tissue sample prior to enzymatic digestion. A bead or other insert is placed into a bead-beating tube containing lysis buffer and the tissue sample, which then undergoes vigorous shaking. The ideal protocol will produce a highly-dispersed sample throughout the lysis buffer, greatly increasing the surface area available to the digestion enzymes. There are two main factors to consider when developing a bead-beating approach: insert (bead) selection and RPM level/duration. Insert selection can vary depending on the sample type used, with a variety of commercially available options. For robust tissue samples such as skin biopsies, many of the beads used for bacterial lysis are completely ineffective and will just bounce off the tissue. When considering RPM speed and duration, too slow/short will produce little effect on the tissue, whereas too fast/long may eventually induce DNA damage and reduce the viability of downstream analysis.

We tried four different insert types that were suggested for large tissue pieces – small and large diameter steel bearings, cubic zirconia beads, and angular garnet pieces – at four different RPM speed/duration combinations [Figure 2]. At only 500 RPM, with a duration of 3 minutes, none of the inserts had very good performance and all of the skin biopsies were largely intact. Similar results were found at 1000 and 1500 RPM, where there were still large portions of tissue remaining. Both sizes of steel beads started to disperse the tissue, but were unable to reach complete homogenization. A final test of 3000 RPM was attempted for each insert type, with a reduced homogenization time of only 30 seconds to prevent unwanted DNA damage. Both sizes of steel bearings were able to produce complete homogenization of the skin biopsies, indicated by the lack of large pieces remaining and consistent small particles throughout the lysis buffer. However, after additional testing, homogenization with multiple small diameter bearings would sometimes result in cracked/broken tubes and lost samples, where the large diameter bearings had no breakage in any runs. With comparable homogenization performance, the risk of breakage and subsequent sample loss/contamination was too high with small diameter bearings and the single large diameter bearing was used moving forward.

Figure 2 – Parameter optimization for homogenization of tissue biopsies.

Various inserts – small and large diameter steel bearings, cubic zirconia beads, and angular garnet pieces – were tested for homogenization ability of skin tissue. Each insert type was tested using a range of RPM speed and duration from 500–1500 RPM for 3 minutes and 3000 RPM for 30 seconds. With the goal of a fully-dispersed tissue with no large intact pieces remaining, only the small and large diameter steel bearings at 3000 RPM were successful. All other combinations left biopsies largely intact, which would inhibit the rapid DNA extraction. Of the two possible choices, the singular large steel bead was selected due to a high incidence of tube-breakage when using the smaller diameter steel bearings.

With equal masses used for all bead selection trials, the cubic zirconia bearings seemed to lack the density to sufficiently penetrate and disperse the tissue biopsies. Angular garnet was indicated for skin tissue bead-beating by a commercial supplier; however, it did not seem to work well on the skin and instead just degraded the other pieces of garnet into a fine powder that coated the tissue. Steel beads were the clear choice for ideal skin biopsy homogenization, and the singular steel bead makes it very simple for future users – adding one bearing instead of counting or weighing multiple smaller ones. The only other consideration not shown here for bead selection was the switch to corrosion-resistant stainless steel beads, as some of the initial beads tested would start to visibly rust during the experiment.

3.2. DNA yields from standard and shortened extraction steps

With optimized insert, RPM level, and bead-beating duration, a complete tissue homogenization and DNA extraction was attempted using BLENDER. Three different DNA extraction methods were performed: (1) a control using the standard extraction protocol from QIAGEN with a 4-hour extraction time; (2) a “shortened control” using the same standard extraction protocol as in (1) but limited to only a 15-minute extraction time; and (3) a rapid DNA extraction using the BLENDER with a 30-second homogenization step at 3000RPM and a 15-minute total extraction time [Figure 3A]. Total sample processing time – the time required from start of patient visit to the time at which purified DNA is available for downstream NAT – includes the tissue biopsy removal and preparation, a DNA extraction step, and finally a DNA purification step.

Figure 3 – Sample processing time and DNA yield comparison.

(A) Total processing time, including biopsy removal, DNA extraction, and DNA purification, is shown for three separate protocols. The control protocol utilizes a standard commercial enzyme-only digestion with a 4-hour extraction time. The shortened control and BLENDER protocols were limited to a 15-minute extraction step. (B) BLENDER is capable of producing comparable DNA yields to the control protocol in a much shorter time. Whereas the shortened control that does not utilize homogenization produced very little DNA comparatively.

Quantifying the purified DNA yield from each of the three methods (n=4), we see no significant difference between the standard protocol control taking 4 hours and the BLENDER rapid extraction taking only 15 minutes [Figure 3B]. However, there is a very significant difference in the DNA yield when using a non-homogenized, shortened control with only a 15-minute digestion. This indicates that the complete homogenization and simultaneous enzymatic digestion can produce a complete DNA yield in only ~6% of the time of a standard digestion. With no significant difference between BLENDER and the 4-hour control, this indicates that optimal temperature is maintained within the BLENDER despite homogenization and stays well below the 95°C requires for enzyme inactivation.

With some downstream NAT approaches capable of short turnaround times^{6, 23, 24}, often within hours, compared to more time-consuming traditional protocols like pathology or microbial culture, a 15-minute extraction step could enable same-day results rather than patients returning for a later follow-up. Additionally, 4-hour long extractions in this study were performed with an attentive operator, thoroughly mixing the sample by vortexing every 30 minutes throughout the extraction. In limited-resource settings, it may be difficult to demand as much attention from healthcare workers, meaning the extraction step would likely take even longer than 4 hours. A related approach produced amplifiable DNA after a 10-minute digestion period with a manual homogenization step using a toothpick after 5 minutes²². However, this open-air homogenization increases safety risks when using hazardous samples in POC settings. Additionally, they utilized 2mm biopsies and note that there was still tissue remaining after the 10-minute cutoff, indicating an incomplete digestion of tissue. BLENDER is able to produce a completely clear lysate in 15 minutes using a 3mm skin biopsy within a completely contained system for maximum efficacy and safety in limited-resource settings.

3.3. DNA yield and amplification performance using qPCR

With equivalent performance between BLENDER’s 15-minute extraction and a standard 4-hour control, the only remaining question of the viability of this approach is the downstream testing using the extracted DNA. Since the end goal of this approach is to produce DNA more quickly for later quantitative or qualitative testing, the DNA needs to be usable in common NAT modalities. A previous study suggested that bead-beating could reduce quantifiable DNA through excessive fragmentation that is not obvious in spectrophotometric measurements²⁵. A qPCR assay for GAPDH, a housekeeping gene, was used to quantify the amplifiable DNA produced by each method to determine whether bead-beating homogenization would have a detrimental effect on DNA quality. The standard enzymatic digestion control was compared to a homogenized-tissue extraction (n=8), with comparable DNA yield between the two [Figure 4A]. Using a standard curve, GAPDH copy numbers for each sample was determined, with no significant different between the standard control extraction and the extraction using bead-beating [Figure 4B]. This indicates that the chosen bead-beating parameters – single steel bead, 3000 RPM, 30 seconds – was optimal for fully dispersing the robust tissue while being gentle enough to keep the DNA viable for downstream amplification.

Figure 4 – DNA yield and qPCR amplification comparison.

(A) The control protocol utilizing a 4-hour DNA extraction produces similar DNA yields to homogenized tissue. (B) A qPCR assay was performed to assess the viability of homogenized DNA for downstream amplification. GAPDH copy numbers were equivalent between both methods, indicating no detrimental effects of bead-beating to the extracted DNA.

4. Conclusions and Discussion

Bead-beating homogenization using the BLENDER is a fully-contained, POC-accessible way to quickly disperse tissue within a lysis buffer. Simultaneous enzymatic digestion completely releases DNA as well as removes DNA-degrading extracellular proteins and qPCR inhibitors. Utilizing the BLENDER for a rapid DNA extraction, a complete DNA yield that is viable for downstream amplification can be produced in only 15 minutes compared to an average time of 4 hours using a standard commercial extraction protocol. Greatly reducing the sample processing time required for downstream nucleic acid diagnostics could remove some of the barriers to rapid point of care diagnosis in limited-resource settings.

Future improvements to the BLENDER can also be made to further increase the POC-potential, such as transitioning to battery power or hand-crank actuation of the Scotch yoke. An internal battery could reduce the reliance on stable electricity, however, still requires charging when electricity is available. However, since the Scotch-yoke is driven by circular motion, a hand-crank mechanism could be used to actuate the homogenizer, with an added reverse-worm-gear setup to function as an RPM increaser in order to achieve ~3000RPM reliably. The heating system could also be re-designed to accept alternative power sources, such as a battery, or potentially a small flame (such as from a Bunsen burner) to accept and store heat energy. The BLENDER could utilize a phase change material within the aluminum heating block to absorb and store excess heat, keeping the system at 65°C and preventing unwanted overheating which could denature the enzymes. In addition to these mechanical changes, future assay improvements such as a direct-to-LAMP approach²⁶ could decrease the need for additional processing post-extraction and allow operators to test extracted DNA directly, though more investigation is needed.