Bone DNA Extraction at the Netherlands Forensic Institute
Francisca E. Duijs and Titia Sijen, Netherlands Forensic Institute, Division of Biological Traces, The Hague, The Netherlands
Share this article
INTRODUCTION
At the Netherlands Forensic Institute (NFI) we use DNA bone extraction in the process of identification of human remains. Bone samples are among the more complex sample types for DNA extraction. In case of DNA identification of human remains, bone samples are generally taken as a last resort. When other (soft) tissue types are not available or do not lead to informative DNA profiles, bone tissue may result in a DNA profile as the DNA is better preserved in the calcified matrix of the bone tissue. Although some bone samples result in ample amounts of DNA, bones may be compromised as they are often retrieved under far from ideal circumstances in the forensic context. The bones may have been in contact with for example microbes, chemicals, different weather conditions, fire, soil or water (various types of soil, salt or sweet water) that affect the preservation of the DNA. These factors make DNA extraction from bone samples challenging. In some cases, the full history of the environmental conditions may be unknown (remains found on a beach for instance) which makes it difficult to develop protocols for specific preservation conditions. Also, a straightforward workflow is preferred that is suitable to the range of bones samples occurring in the forensic context; from highly compromised to relatively fresh bones. The bone samples processed at our laboratory show a large variation in decomposition stage although most are in the final skeletonization stage. One of the main difficulties we encounter during DNA extraction is the co-extraction of inhibitors leading to amongst others poor DNA yields, inhibition during PCR and artefacts in DNA profiles.
Until 2017, the most successful DNA extraction method at the NFI was to fully dissolve 1 gram of finely grinded bone powder with a demineralization buffer in an overnight incubation step. This process resulted in large volumes (up to 50mL) from which the DNA needed to be bound to a silica column. Often, this resulted in clogging of the silica column. In 2017 we started to update our bone DNA extraction protocol with main focus on lowering the bone powder starting amount and lowing the volumes to suit automated extraction platforms. Automated extraction is appreciated in the forensic workflow because of the large sample sizes and the reduced risk of sample mix-up and contamination. With bone samples, large sample sizes may occur in case of DVI (disaster victim identification) or with special projects. Ideally, the new method is applicable for all types of bone samples as well as dental elements.
THE STUDY
In our 2017 study [1] we learned that some steps in the bone extraction process are key, such as thoroughly cleaning and grinding the bone to a very fine powder, lowering the starting amount of bone powder to reduce co-extraction of inhibitors and lowering the total buffer volume so that it suits automation. In theory one would say; the more bone powder, the more DNA, the higher the DNA concentration. But in practice we noticed that lowering the amount of bone power benefits the DNA yield per gram bone powder (Figure 1). By lowering the starting amount of bone powder, the absolute volume went down from ~50mL to ~5mL. The latter still being too high of a volume for our automated platform (Microlab® STARplus, Hamilton). We noticed that by lowering the buffers, the bone powder no longer dissolved completely. But surprisingly, this resulted in an increase in the DNA yield per gram bone powder (Figure 2). The undissolved bone powder could be separated well with an AutoLys tube (Hamilton). The benefit of this tube system is that the incubation and separation take place in the same tube (after extraction, the inner tube that contains a filter is lifted) and an extra transfer step with loss of material is prevented. We performed optimization rounds in order to lower the total volume of chemicals and experimented with combining magnetic beads from different suppliers to reach the highest DNA yield.
Figure 1 (left): Yield in ng DNA per gram extracted bone powder (n=2). By lowering the amount of bone powder the DNA yield increases. Figure 2 (rechts): Reducing the volume of chemicals makes that the bone powder will not dissolve entirely (all the darker bars)
Then, we learned about Promega’s new DNA bone extraction method that could be executed on the Maxwell® FSC instrument (Promega), a small purification platform. We compared our method, although still under development, to the Promega bone extraction method and were positively surprised by the profiling information the Promega extracts yielded and the much reduced occurrence of artefact peaks (see the increased number of allele calls in Figure 3 and the artefact peaks marked with x).
Figure 3: NGM profiling results of the protocol that was still under development (left) and the Promega DNA bone extraction method (right). Peaks marked with X are classified as artefacts.
The bone extraction method of Promega starts with 0.1 gram bone powder mixed with bone lysis cocktail A and an incubation for 2.5 hour at 56˚C while shaking at 1,000rpm. The undissolved bone powder is separated from the lysate by pelleting and transferring the supernatant to a new tube. Bone lysis cocktail B is added to the tube with the supernatant, mixed, and the complete volume is transferred to the Maxwell cartridge. The Maxwell run is prepared by placing up to 16 cartridges with plungers and elution tubes with elution buffer in the apparatus and the protocol is started. This is an easy and short protocol that is partly automated.
At this stage we decided that further optimizing the DNA bone extraction method from Promega on the Maxwell® FSC instrument would be the most opportune for our lab. With the knowledge that we obtained from our research we quickly introduced the AutoLys tube to the Promega sample preparation. Next, we started investigating if we could make the method even faster by reducing the incubation time. Our research had shown that full dissolvement of the bone powder was not crucial for a high DNA yield and since dissolving the bone powder is the most time-consuming step, shortening the incubation time could greatly reduce the total time needed for DNA extraction. We performed DNA bone extractions with several incubation times (Figure 4), and observed little to no difference in DNA yields.
Figure 4: Reducing incubation time (n=2)
Then, we proceeded to a more extensive validation in which we compared 20 bone samples with incubation times of 2.5 hour and 2 minutes (Figure 5).
Results show that for many samples lowing the incubation time to 2 minutes results in similar DNA yields (samples 5, 6, 10, 11, 12, 13, 14 and 15 have overlapping error bars); for some samples (sample 1, 3, 4, 7, 8, 9, 16, 17, 18 and 20) longer incubation resulted in higher DNA yields; while others (sample 2 and 19, especially sample 2) benefited from a shorter incubation time.
Figure 5: DNA yield for 20 different bones (n=2) with 2 minutes and 2.5 hour incubation time.
We assume the reduced incubation time prevents the co-extraction of inhibitors in some samples (for example sample 2 in Figure 5). We confirmed this with a bone sample for which we knew it contained inhibitors. A slightly higher amount of bone powder was used to increase the amount of inhibitors. For this bone, we used 150mg finely grinded powder and compared 2.5 hour with 2 minutes incubation time (n=2). The results show a 29-times higher DNA yield (ng/gram bone powder) for the 2 minute incubation time over the 2.5 hour incubation time. The reduced inhibition for some samples and the reduced waiting time for all samples made us favor a 2 minute incubation time.
Next, we investigated whether the bone grinding time and fineness of the bone powder could have a positive effect on the DNA yield. Prior to grinding, cleaning the bone sample is an important step to prevent co-extraction of inhibitors and contaminants. Our standard procedure is cleaning the samples with nuclease free water and removing soil, dirt and soft tissue (including bone marrow) when present. Often, we clean with chlorine, rinse with water followed by 70% alcohol to speed up the drying process. With more porous samples we only clean with water and not with chlorine to prevent chlorine residues in the bone powder upon grinding. The samples are grinded with the Freezer/Mill® 6875 High capacity Cryogenic Grinder (SPEX® SamplePrep) cooled by liquid nitrogen.
With a 2 minute incubation time, the bone powder must be grinded to a real fine grain size to maximize the bone powder surface. Apparently, the fine grinding promotes accessibility of the DNA in Bone lysis cocktail A.
Grinding only a few minutes longer gives a much finer bone powder as illustrated with the microscopically created images in Figure 6, and resulted in higher DNA yields (data not shown). In practice, not all bones need the same time to be grinded to a fine powder. After 4 minutes samples are visually checked, when the powder still contains bigger parts the sample is submitted to further grinding till a finely granulation is obtained.
Figure 6: Grinding 4 minutes (left) versus grinding 6 minutes (right).
Casework
The optimized Promega protocol was introduced in casework. Since many forensic bone samples are of low quality, we aimed to increase the DNA yields to allow obtaining informative profiles. Therefore, we start with 3 portions of 0.1 gram finely grinded bone powder and extract each of these portions separately in the Maxwell® resulting in three extracts of 33.3µL DNA that we combine to our standard DNA extraction volume of 100µL. Figure 7 shows the effect of the 3 combined extracts in profiling with the PowerPlex® Fusion 6C system (Promega).
Even then, some inferior samples showed too low DNA yields and in those cases the process is repeated 4 times. This means that in total 12 portions of 0.1 gram are used. Since each portion is eluted with 33.3µL, the total volume of DNA extract is 400µL. After applying an ethanol (EtOH) precipitation using Glycoblue™ (Thermo Fisher Scientific) as a coprecipitant, this total volume is reduced to a volume suitable for one PowerPlex® Fusion 6C profile. This procedure can in some cases lead to an informative DNA profile.
Figure 7: PowerPlex® Fusion 6C profiles, 0.1 gram bone powder (left), 3x 0.1 gram bone powder (right).
When relevant in a case, other types of profiling systems may be applied such as Y-chromosomal analysis (for example PowerPlex® Y23 System (Promega)), mtDNA analysis or autosomal STR analysis with system especially suited for degraded DNA such as MiniFiler™ (Thermo Fisher Scientific) or massively parallel sequencing methods (such as IDseek® OmniSTR™, NimaGen.). When the standard procedure (3 x 0.1g eluted into 100µL) is used, sufficient extract remains; when the method including ethanol precipitation is applied new extraction from the bone powder needs to be started.
Since the introduction of the method in 2018 through the beginning of 2023 a total of 536 different bone and dental element samples were processed with the DNA bone extraction method on the Maxwell (so around 100 samples year). These represent a broad range of bone sample quality, which is reflected in the range of yields (in ng DNA per g bone powder) that is obtained: 33 samples yield more than 100ng/g, 42 samples between 10-100ng/g; 176 between 1-10 ng/g; 161 samples between 0.1-1ng/g, 59 between 0.01-0.1ng/g and 65 below 0.01ng/g.
For a subset of the samples we monitored how effective EtOH precipitation was by performing quantification before and after concentrating. In 45 of these samples an increased DNA concentration was obtained after concentrating, which was on average a 5.1 times increase in concentration (this fits with the amount of DNA extract used since generally ~60µL of the extract is precipitated and the remaining ~40µL extract is kept as contra extract). This shows that the quantification reaction is not inhibited when extracts are EtOH-precipitated indicating that the method results in limited co-extraction of inhibitors.
The success of the method is illustrated by the outcome in a specific casework project on the identification of remains collected in and around the North Sea: 62 samples were submitted to the extraction protocol resulting in 30 different profiles of unknown persons that could be added to the Interpol missing person database.
The current method is suited for a limited number of samples and semi-automated. In case of a disaster where disaster victim identification (DVI) with large amounts of bone samples is necessary, a system for larger sample numbers or a fully-automated method would be opportune. When changing to Promega’s Maxwell® RSC 48 platform, 48 samples can be processed simultaneously, which is three times more than the Maxwell® FSC instrument that is currently used at the NFI. Moving to a fully-automated method requires further development.
Before introducing the DNA bone extraction method on the Maxwell® FSC, DNA extraction was the most time-consuming step of the process. This is currently one of the fastest parts in obtaining a DNA profile from a bone sample. Now, the bone powder preparation is the most time-consuming step, which inspires new research.
REFERENCES:
- Duijs, F. E., & Sijen, T. (2020). A rapid and efficient method for DNA extraction from bone powder. Forensic Science International: Reports, 2, 100099
Image Header: Adobe Stock/IHX
Francisca E. Duijs Netherlands Forensic Institute, Division of Biological Traces, The Hague, The Netherlands
Francisca Duijs graduated from Hogeschool Leiden, the Netherlands in 2015. She did her final thesis and internship at the Netherlands Forensics Institute (NFI) and has been working there, in different teams, ever since. In 2016 she joined the Research Team of the Biological Trace department where she specializes in topics such as human identification using bone tissue, the automation of manual methods, optimizing and validating DNA profiling technologies and performing specialized casework using various DNA and RNA profiling technologies.