Fourth Annual DNA Grantees' Workshop
Wednesday, June 25, 2003
MORNING SESSION
Research Update Briefings: Ongoing Projects (Continued From Day 2)Angelo Della Manna, Moderator
Biography
MR. DELLA MANNA: Good morning everybody. Welcome to day 3. For those of you who don't know me, I'm Angelo Della Manna. I'm with the Alabama Department of Forensic Sciences in Birmingham, the home of Ruben Studdard, of American Idol fame, and Eric Rudolph, one of America's most wanted. The former is not in the national database, and one hopes the latter will be in the national database.
Method to Determine the Age of a Biological Sample
Stacey Anderson
MR. DELLA MANNA: This morning we have a great series of three very interesting talks. The first is going to be given by Stacey Anderson, who is a fifth-year graduate student in Dr. Cliff Bishop's lab.
Dr. Bishop is the program director of the Forensic and Investigative Science Program at West Virginia University (WVU). He received his Ph.D. in biology from the University of Virginia and is now at WVU. He is one of the few people that I know who's actually gone from the Atlantic Coast Conference to the Big East Conference.
Stacey will be speaking today on his behalf—because she does all the work—on a method to determine the age of a biological sample.
Please join me in welcoming Stacey Anderson.
MS. ANDERSON: Thank you and good morning. Today I'm going to speak about a method to determine the age of a biological sample. First, I want to talk about the specific need for a method to determine the age of a biological sample. Over the past several years, traditional PCR (polymerase chain reaction) has revolutionized the forensic community because it gives us the ability to place a person at the scene of a crime. But problematically, traditional PCR provides no information on when a sample was actually deposited. Anderson: Slides 1 and 2
That is why we've developed temporal PCR, which will help to estimate the time that the sample was deposited. In combination, traditional PCR and temporal PCR will provide both a spatial and a temporal link to a crime scene.
These methods have potential applications for both military and law enforcement use, specifically in tracking suspects. If traditional PCR can link Osama bin Laden and identify what cave he had occupied, then temporal PCR can help approximate how long ago he had occupied that cave and, if given in a relevant amount of time, we could approximate how far he could have traveled in a given time frame, potentially aiding in identifying his current location. Anderson: Slide 3
Temporal PCR can have an advantage when the suspect and the victim have close ties. I'm going the take everybody back to the O.J. Simpson trial. They found Nicole's blood in O.J.'s Bronco, but one problem with this piece of evidence was that the prosecution wanted to say that the blood was 2 weeks old, dating from when the murder had occurred, while the defense was saying that the blood was 6 months old (when Nicole had cut herself in the Bronco). This piece of evidence had to be thrown out of court because they couldn't determine how old the sample really was. But temporal PCR would allow for the identification of an approximate age for this sample.
This is a very important concept for the forensic community, and for nearly 100 years, numerous researchers have tried to develop methods to determine the age of blood samples. They have looked at changes in color and solubility, the transformation of hemoglobin into its derivatives, and serum protein profiles. Anderson: Slide 4
These methods, however, have several limitations. Because most of these are enzymatic assays, the limitation would be sample size. The strength of the signal is directly proportional to the amount of starting material. A large sample size could give misleading results. None of them are human specific, and contamination could give false positives. They're all limited to blood samples because most of them deal with hemoglobin in some way or the other, thus they have a narrow window of usefulness.
So for our purposes, we have chosen to examine RNA (ribonucleic acid), specifically looking at its stability over time. There are several mechanisms that will contribute to RNA breakdown. With RNAs occurring both endogenously and exogenously, the most prevalent is enzymatic breakdown. There's also the potential for breakdown caused by high or basic pH (potential of hydrogen) and breakdown caused by ultraviolet light. Anderson: Slide 5
Here, we're looking at RNA decay based on a relative ratio. This is very similar to carbon-14 dating. We will identify two different species of RNA: RNA 1, the less stable RNA species, which would be the equivalent of carbon 14; and RNA 2, the more stable RNA species, which would be the equivalent to carbon 12. Anderson: Slide 6
Each of these have a set amount of molecules at day 0, but after 10 days, the more stable RNA species is only going to lose about 5 of its molecules. On the other hand, the less stable RNA species is going to lose more than half of its molecules. Over time, the ratio of RNA 1 molecules to RNA 2 molecules changes from 40:50 to 10:45.
There are several advantages to looking at the relative ratio:
- The isolation is identical for all RNA. So even though we're looking at two different RNAs, they will both still be extracted through the same isolation procedure.
- The relative ratio is independent of sample size. This was a limitation of the previous techniques. The relative ratio should remain invariant regardless of the amount of starting material.
- Multiplexing the two RNAs will serve as an internal control. It will control for enzymatic efficiency, pipetting errors, and the amount of template starting material.
- Because it's RNA, the primers and probes for PCR can be made human specific, based on polymorphic regions of the coding region.
- Because RNA exists in all tissues, we're not limited to blood samples as in the previous techniques. So we can apply this to any tissue that is found at a crime scene.
- We also can isolate RNA and DNA from the same sample, which is important for identification purposes through traditional PCR. Anderson: Slide 7
So in isolating the RNA and the DNA from the same sample, I used TRI Reagent BD (Molecular Research Corporation) for blood derivatives for the RNA isolation. It's an organic extraction, so the RNA is going to separate into the upper aqueous level, and the DNA is going to separate into the lower organic level.
We also use DNAzol, another product from Molecular Research Corporation. It's actually based on a modification of a protocol for DNA extraction from mouse tails.
Here, I've isolated RNA and DNA from the same blood sample. The RNA was isolated from the upper aqueous phase, and the DNA was isolated from the lower organic phase. Also, a portion of the RNA was converted to cDNA. Anderson: Slide 8
I've designed RNA-specific probes and DNA-specific probes. The RNA-specific probes were based on an axon-axon boundary in the beta-actin gene, and the DNA-specific probes were based on a non-transcribed region of the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene. This is a multiplex reaction for real-time PCR. So what we're seeing is that the RNA-specific probes are recognizing only the cDNA. It's not detecting a signal from the DNA, suggesting that this is RNA specific. With DNA, we're only getting a signal for the DNA sample, we're not getting amplification for the cDNA or the RNA, suggesting that RNA only recognizes RNA and DNA only recognizes DNA.
The group of RNAs that we examined for this purpose are termed housekeeping genes. Housekeeping genes are critical for cell survival, and they're expressed in all types of tissues. We have chosen two specific species: 18S ribosomal RNA (rRNA) and beta-actin messenger RNA (mRNA). Anderson: Slide 9
We believe that 18S rRNA is the more stable of the two species because it exists in a cellular environment in a ribosomal protein complex. We believe that the protein complex makes it more stable because it serves as a protective environment, shielding it from agents that can contribute to RNA breakdown. Relatively few regions of the RNA are left exposed in the cell.
On the other hand, although also incorporated by a polyribosome complex, the beta-actin mRNA has significant gaps that are left exposed in the cell, as demonstrated by the electron micrograph. With these gaps being exposed, we believe it is more susceptible to agents contributing to RNA breakdown, making it the less stable of the two RNA species.
For our experimental approach, we collected blood without anticoagulants from human subjects. This is important because we wanted to mimic a real-world situation. The blood was collected (day 0) and aliquotted onto either cotton fabric or Epindorf tubes and stored at a constant temperature and humidity until the sample was aged to an approximate date. Anderson: Slide 10
At this point, we isolated total RNA, converted the RNA to cDNA through a reverse transcriptase reaction, and used real-time quantitative PCR and a multiplex reaction between 18S and beta-actin, using an ABI (Applied Biosystems) 7700 real-time PCR machine.
What we're showing in this amplification plot are the amplification of 18S rRNA based on the blood collected at day 0, which is in the red amplification blot, and the 6-day-old blood, which is in the blue amplification blot. In real-time PCR, the threshold cycle (TC) value is the measurement of interest and it's the cycle at which exponential amplification occurs. We see relatively little difference between these two time points, so there's little difference in the amount of 18S rRNA present. One TC value is equivalent to a two-fold difference in starting material. Anderson: Slide 11
On the other hand, we see a significant change in the amount of starting material of the beta-actin mRNA, with significant decreases in the amount of beta-actin present at day 6 versus day 0, indicating that this is the less stable of the two species.
Last year Dr. Bishop spoke about a proof of concept. Well, we looked at the relative ratio over time of RNA decay for one individual. Recently, we've expanded this to a small population study (four males and four females), specifically looking at the areas where variability occurs; this will demonstrate person-to-person variability, if any. For each of these individuals, we performed three different blood draws on three different dates; this will demonstrate within-person variability. For each of these blood draws, we aliquotted the samples onto cotton fabric and let them dry. So for each blood draw for each identification for each time point, we processed three samples for that time point. We isolated RNA from three samples of the same draw; this would demonstrate technique variability, specifically for the RNA isolation technique. For each of these three samples, we did replicate assays when we set up the real-time PCR reaction; this would demonstrate any variability contributed by the real-time PCR machine. Anderson: Slide 12
We've pooled the results for the eight individuals. Each of these time points represents 144 assays. The blood samples were aliquotted onto cotton fabric, dried, kept at a constant temperature of 25ºC and 50-percent humidity, and aged for the indicated time points: day 0, 30, 60, 90, 120, and 150. At that point, the RNA was isolated, and the TC values for 18S and beta-actin were determined. Anderson: Slide 13
In looking at the difference in TC values between 18S and beta-actin, we see over time that there is a decreased amount of beta-actin relative to 18S.
These are some recently pooled results that have been submitted to our statistician for statistical analysis, but in some preliminary results, we have shown a sexual dimorphism between males and females, indicating that the ratios are different from one sex to the other. You can't see it in this graph because the results for both male and female were pulled, but further statistical analysis of our results will validate this theory.
One problem with this experiment was that our primers and probes for 18S and beta-actin were not human specific. Our 18S probe—Universal 18S—was commercially available from ABI. It was universal in that it should recognize RNA from all species. The beta-actin, also available from ABI, was designated to be human specific, but when we tested it on various types of RNA, we found that it was not human specific. In response, we designed 18S primers and probes. They were based on polymorphic regions of the 18S gene and were designated 18S4. We tested them on RNA from cat, dog, mouse, human, drosophila, rat, pig, and yeast; and we have successful amplification of 18S for a human and a pig. You can draw your own conclusions from that. Anderson: Slide 14
We'll continue to resolve this problem by trying to make our primers and probes more specific for humans and to test them on an increased amount of RNAs from other species. If we can't pick up the signal for a pig, we will design pig-specific primers and probes so we can test the sample. We also can test a sample to make sure that it's not contaminated by RNA from a pig.
So you might think that you're also seeing a signal from some of the other RNAs. Putting this into a mathematical perspective, one TC value is equivalent to a two-fold difference in starting material. 18S is a prevalent gene; we're getting amplification at cycle 14. Our average for other RNAs is about cycle 33. This is a difference in TC value of 219 or more than a 500,000-fold difference, contributing 0.00019 percent. So something amplifying at cycle 33 is going to contribute 0.00019 percent to something amplifying at cycle 14, and this would represent background levels of detection.
On the other hand, beta-actin is conserved less than 18S, so it was easier to design human-specific primers for beta-actin. Human RNA is the only RNA recognized by beta-actin multiplexed with the Universal 18S primer and probes available from ABI.
We recently performed a DNA microarray. We wanted to generate a shorter time line, something that will allow for detection during the first several days. We looked at RNA from human blood that had aged for 0 hours and compared it with RNA from blood that had aged for 4 hours. The microarray examined 16,000 genes, and of these 16,000 genes, 1,100 were significantly altered, either up-regulated or down-regulated. The majority of them were up-regulated. This is not surprising because when a cell is placed in a stressful environment, it will induce gene expression to try to survive. I would think that when blood is placed outside the body, it would elicit a stress response. Anderson: Slide 15
So this is the amplification plot showing the results of the DNA microarray. It compares the RNA from the 0-hour blood with the RNA from the 4-hour blood. The blue represents the genes that remain relatively unchanged between the two samples; the red indicates the genes for which expression increases from 0 hours to 4 hours; and the green shows the genes for which expression decreases from 0 hours to 4 hours.
As you can see by the red, the majority of genes were up-regulated. The three most altered genes were all increased. It was a G0 to a G1 switch gene, 96-fold increase; the early growth response gene, 66-fold increase; and a regulator of G protein signaling, 49-fold increase. At this point I have designed primers and probes for real-time analysis of these three genes, and I will continue to validate these results by real-time PCR and try to establish a shorter time line.
In the future, we want to try to reduce variability. There are several techniques that we have thought of hopefully to establish a more precise time line for our samples. We want to look at additional tissues, including hair and saliva. At this point we have successfully isolated RNA from a single follicle cell, a single hair strand, and saliva and actively detected both 18S and beta-actin in these samples. Anderson: Slide 16
We also want to look at variable humidity and temperature. As of now, our conditions have been kept constant at 25ºC and 50-percent humidity. We have ordered two new environmental chambers that will allow us to alter these environmental conditions. We're also going to look at the effect that full-spectrum light has on the samples.
We want to look at variable source and an increase of population size, specifically looking at ethnicity, sex, and age. This will help to determine whether there is sexual dimorphism in our relative ratio between males and females, and we want to continue our early time line with the DNA microarray results.
We also want to continue with our extended time line. We want to see how far past 150 days we can take our RNA, and we want to look at introducing contaminants to our samples, because most often when a blood sample will be found, especially outdoors, different microorganisms and/or different types of blood are often present. We want to make sure these won't contribute to our results.
With that, I want to thank Cliff Bishop, my adviser and principal investigator; Brandi Howard and Dana Kubinski, two research technicians; Gerry Hobbs, statistician; Rachel Wakeman, phlebotomist; and NIJ for their funding. Anderson: Slide 17
MR. DELLA MANNA: If we could, we're going to take the questions at the end for the panel.