New CDC test: what is RT-PCR?
[NB: Blogger is acting erratically, eating posts; versions of this post have been up and then not up twice already today.]
Yesterday the FDA announced it had approved a new laboratory assay for avian influenza/H5 (Asian lineage). It is a complex test that can only be used by trained personnel and at the moment is limited to CDC's Laboratory Response Network (LRN), about 140 labs in the 50 states. These labs are certified to have personnel trained in the kind of rapid molecular procedures used in this test and to have the proper analytical and safety facilities to handle possibly infective viral specimens. They are also part of an established public health information network. But CDC plans to share the reagents (the primers, probes and control virus; see below for an explanation) with other national laboratories. This is a rapid test for one thing: whether there are viral proteins in a specimen that are characteristic of the kind of highly pathogenic H5 subtypes now circulating in Asia and Eurasia.
So what is this test? Technically it is a Real-Time Reverse Transcription, Polymerase Chain Reaction (RT-PCR) Primer and Probe Set for the Asian lineage of influenza viruses of the H5 subtype. So a little discussion of RT-PCR is in order.
RT-PCR is a form of PCR, so we'll start there. PCR is a means to take extraordinarily a tiny amount of a specified DNA sequence present in a complicated mixture and amplify it into billions of copies. Once amplified there is enough of it to do various things, like sequence it or just see if is there at all. Because PCR can be made to recognize very specific sequences it can be used to identify bacterial or viral strains or deceased people -- anything where knowing a genetic sequence will help identify something or someone. So you need to know what you are looking for. Actually what you need is something a little bit different than knowing the region of interest: you need to know some short but exact sequences on either side of ("flanking") the region of interest. These are the primers. You don't actually have to know everything about the region between them, even though that is the region you are interested in amplifying. But accuracy in the primers is essential.
PCR is done with DNA, not RNA, which is the genetic material of the influenza virus. So the first step is to turn the viral RNA sequence into a DNA sequence. This is done with the enzyme Reverse Transcriptase (the RT part of RT-PCR; unless it is being used, as it sometimes is, to designate Real Time PCR; context is everything. The new test is both Real Time and uses Reverse Transcriptase.) Making a matching DNA from the viral RNA can be a tricky part of the process. If you aren't successful in turning the viral RNA into DNA you are done before you've even started.
Explaining PCR can be tricky and confusing so I'm not going to give you all the details but just the basic idea.
We start with double-stranded DNA. Let's idealize it with a railroad track, each track being the sugar backbone (the deoxyriboses or D part of DNA), and the ties being the matched bases (the complementary letters of the four letter genetic code). OK, I know. The image doesn't quite work, because a railroad tie is a single piece of wood and we want the tie to be in two halves, a genetic letter and its matched pair (A - T or G - C are the match-ups). So each tie is either an A - T pair or a G - C pair, not a single piece of wood. You'll have to imagine this in your head because if I tried to draw it with a drawing program I'd be sitting here all day and the results would still be terrible.
For each side of the track we get a string of letters (AGGTCTCT . . . etc.) that represents a genetic sequence. The other side of the track has the matched pair (in this case TCCAGAGA . . ., the complementary matches). PCR works on both sides, but we want to make this simple so we'll only work with one half of the tracks. Imagine you've got a big rotary saw and you go right down the middle of the track bed and separate the two tracks by sawing through the ties. In DNA you don't need anything as powerful as a saw because the two halves of the tie aren't really a solid piece of wood. The two matching genetic letters that make up a single tie are bound together fairly weakly by forces that are not anywhere near as strong as a chemical bond. There are a lot of these weak forces so they keep the two strands pretty well stuck together. But since each of the ties is weak, if you exert some force they will fall apart. With DNA you can separate the two halves just by heating up the DNA, making the two sides of the track part from each other.
So now you've got the two sides of the track with their halves of a tie sticking off them like little hairs, each representing a genetic letter, its matched pair on the other half of the tie on the other track, which we are going to forget about for clarity. This is where our primer sequences come in. Each primer is a sequence of genetic letters flanking a section of track we want to amplify. It might look something like the sequence AAGGCCGT, although usually primers are a bit longer. Anyway, we walk along the entire length of the track until we find a section of (half) ties that matches it, if there is one. If we find such a section we stick it onto the matching section and then go back to the railyard and bring out the heavy machinery. This is a machine (called DNA polymerase) starts by clamping onto the newly placed primer and moves to the right laying down both new metal track and inserting, from warehouse stock, the right matching halves of new ties as it comes to each spot. We needed the primer because this machine won't work without that little piece to grab onto to get it started. And it moves in only one direction, say to the right.
After we run the machine to the end we haven't quite recreated the railroad track because to the left of the primer there is no match. Oh, well. Not to worry. Get out your saw (or your heater) and separate these two tracks again (really one full length half track and one half track that begins at the primer sequence). Let's take the shortened side and work with that (I'm leaving out some stuff here, but I just want you to get the idea). Get out your other (flanking) primer sequence, the one that is on the other side of the section of track you want to amplify. Now walk along the track and try to find a match. If you find one, line it up and get out your machine and run it back in the other direction (you are on the other side of the track now, so "to the right" means to go back in the direction you came from, toward the first primer). This time it stops when it gets to the first primer's position (because there isn't anything beyond that now). Now when you separate them one half is a section of railroad track that is just between the two primers. Run the machine on that side and it is now a section of complete two-sided railroad track between the two primers. When you separate them and run the machine on each side you double the number. The PCR machine does this automatically by cycling the temperatures up and down, separating the sides and supplying the cross-tie pieces to the polymerase molecular machine. Do that twenty or thirty times (doubling the piece to be amplified each time) until you have a whole shit house of track section. Voila. PCR.
If you wanted to use it to see if a virus with a specific sequence between the primers is present (say one characteristic of the H5 subtype) you need to have a way to see if the PCR process actually amplified anything. After all, if there were no virus there, the primers wouldn't have anything to match with and then there wouldn't be any product. Sometimes this requires time consuming and complicated processes which take a few days. But newer methods allow it to be done in "real time" (meaning here a couple of hours). Hence Real Time PCR. Usually this involves fluorescent "probes" that match certain portions of the amplified sequence. The changing intensity of fluorescence is then used to watch how much of the probe finds the desired sequence as the process moves forward.
It sounds simple in outline, but it can be fairly tricky in practice. The quality of the specimen may be poor or the RNA in it degraded. Maybe the primers aren't exactly right, because they were designed for a virus whose primer section has now changed. Maybe the probe isn't specific enough for the sequence you are trying to detect because that part has changed, too. Maybe the primers also fit elsewhere. Maybe the PCR process just didn't work right. It is good practice to run a positive control, i.e., a parallel sample you know should come up positive. Then if both the control and the sample are negative you know something might be wrong with the PCR assay. The new kit includes primers, probes and a dead control virus. But if you are wrong about the primers or the probe, you'll get a false negative.
This test is "is designed to detect highly pathogenic influenza A/H5 viruses from the Asian lineage associated with recent laboratory-confirmed infections of avian influenza in humans in east Asia and, most recently, in Turkey and Iraq." (CDC) If it suggests that an Asian lineage H5 is present, additional tests need to be done to see if it is N1 subtype. The announcement comes with this important qualification
This is an incremental advance in testing for H5 virus and we can expect similar developments as time passes. It is faster and can be done at lower biosafety levels than some other testing. No information was released as yet about its sensitivity and specificity so we will have to see from experience how well this works in the real world.
Yesterday the FDA announced it had approved a new laboratory assay for avian influenza/H5 (Asian lineage). It is a complex test that can only be used by trained personnel and at the moment is limited to CDC's Laboratory Response Network (LRN), about 140 labs in the 50 states. These labs are certified to have personnel trained in the kind of rapid molecular procedures used in this test and to have the proper analytical and safety facilities to handle possibly infective viral specimens. They are also part of an established public health information network. But CDC plans to share the reagents (the primers, probes and control virus; see below for an explanation) with other national laboratories. This is a rapid test for one thing: whether there are viral proteins in a specimen that are characteristic of the kind of highly pathogenic H5 subtypes now circulating in Asia and Eurasia.
So what is this test? Technically it is a Real-Time Reverse Transcription, Polymerase Chain Reaction (RT-PCR) Primer and Probe Set for the Asian lineage of influenza viruses of the H5 subtype. So a little discussion of RT-PCR is in order.
RT-PCR is a form of PCR, so we'll start there. PCR is a means to take extraordinarily a tiny amount of a specified DNA sequence present in a complicated mixture and amplify it into billions of copies. Once amplified there is enough of it to do various things, like sequence it or just see if is there at all. Because PCR can be made to recognize very specific sequences it can be used to identify bacterial or viral strains or deceased people -- anything where knowing a genetic sequence will help identify something or someone. So you need to know what you are looking for. Actually what you need is something a little bit different than knowing the region of interest: you need to know some short but exact sequences on either side of ("flanking") the region of interest. These are the primers. You don't actually have to know everything about the region between them, even though that is the region you are interested in amplifying. But accuracy in the primers is essential.
PCR is done with DNA, not RNA, which is the genetic material of the influenza virus. So the first step is to turn the viral RNA sequence into a DNA sequence. This is done with the enzyme Reverse Transcriptase (the RT part of RT-PCR; unless it is being used, as it sometimes is, to designate Real Time PCR; context is everything. The new test is both Real Time and uses Reverse Transcriptase.) Making a matching DNA from the viral RNA can be a tricky part of the process. If you aren't successful in turning the viral RNA into DNA you are done before you've even started.
Explaining PCR can be tricky and confusing so I'm not going to give you all the details but just the basic idea.
We start with double-stranded DNA. Let's idealize it with a railroad track, each track being the sugar backbone (the deoxyriboses or D part of DNA), and the ties being the matched bases (the complementary letters of the four letter genetic code). OK, I know. The image doesn't quite work, because a railroad tie is a single piece of wood and we want the tie to be in two halves, a genetic letter and its matched pair (A - T or G - C are the match-ups). So each tie is either an A - T pair or a G - C pair, not a single piece of wood. You'll have to imagine this in your head because if I tried to draw it with a drawing program I'd be sitting here all day and the results would still be terrible.
For each side of the track we get a string of letters (AGGTCTCT . . . etc.) that represents a genetic sequence. The other side of the track has the matched pair (in this case TCCAGAGA . . ., the complementary matches). PCR works on both sides, but we want to make this simple so we'll only work with one half of the tracks. Imagine you've got a big rotary saw and you go right down the middle of the track bed and separate the two tracks by sawing through the ties. In DNA you don't need anything as powerful as a saw because the two halves of the tie aren't really a solid piece of wood. The two matching genetic letters that make up a single tie are bound together fairly weakly by forces that are not anywhere near as strong as a chemical bond. There are a lot of these weak forces so they keep the two strands pretty well stuck together. But since each of the ties is weak, if you exert some force they will fall apart. With DNA you can separate the two halves just by heating up the DNA, making the two sides of the track part from each other.
So now you've got the two sides of the track with their halves of a tie sticking off them like little hairs, each representing a genetic letter, its matched pair on the other half of the tie on the other track, which we are going to forget about for clarity. This is where our primer sequences come in. Each primer is a sequence of genetic letters flanking a section of track we want to amplify. It might look something like the sequence AAGGCCGT, although usually primers are a bit longer. Anyway, we walk along the entire length of the track until we find a section of (half) ties that matches it, if there is one. If we find such a section we stick it onto the matching section and then go back to the railyard and bring out the heavy machinery. This is a machine (called DNA polymerase) starts by clamping onto the newly placed primer and moves to the right laying down both new metal track and inserting, from warehouse stock, the right matching halves of new ties as it comes to each spot. We needed the primer because this machine won't work without that little piece to grab onto to get it started. And it moves in only one direction, say to the right.
After we run the machine to the end we haven't quite recreated the railroad track because to the left of the primer there is no match. Oh, well. Not to worry. Get out your saw (or your heater) and separate these two tracks again (really one full length half track and one half track that begins at the primer sequence). Let's take the shortened side and work with that (I'm leaving out some stuff here, but I just want you to get the idea). Get out your other (flanking) primer sequence, the one that is on the other side of the section of track you want to amplify. Now walk along the track and try to find a match. If you find one, line it up and get out your machine and run it back in the other direction (you are on the other side of the track now, so "to the right" means to go back in the direction you came from, toward the first primer). This time it stops when it gets to the first primer's position (because there isn't anything beyond that now). Now when you separate them one half is a section of railroad track that is just between the two primers. Run the machine on that side and it is now a section of complete two-sided railroad track between the two primers. When you separate them and run the machine on each side you double the number. The PCR machine does this automatically by cycling the temperatures up and down, separating the sides and supplying the cross-tie pieces to the polymerase molecular machine. Do that twenty or thirty times (doubling the piece to be amplified each time) until you have a whole shit house of track section. Voila. PCR.
If you wanted to use it to see if a virus with a specific sequence between the primers is present (say one characteristic of the H5 subtype) you need to have a way to see if the PCR process actually amplified anything. After all, if there were no virus there, the primers wouldn't have anything to match with and then there wouldn't be any product. Sometimes this requires time consuming and complicated processes which take a few days. But newer methods allow it to be done in "real time" (meaning here a couple of hours). Hence Real Time PCR. Usually this involves fluorescent "probes" that match certain portions of the amplified sequence. The changing intensity of fluorescence is then used to watch how much of the probe finds the desired sequence as the process moves forward.
It sounds simple in outline, but it can be fairly tricky in practice. The quality of the specimen may be poor or the RNA in it degraded. Maybe the primers aren't exactly right, because they were designed for a virus whose primer section has now changed. Maybe the probe isn't specific enough for the sequence you are trying to detect because that part has changed, too. Maybe the primers also fit elsewhere. Maybe the PCR process just didn't work right. It is good practice to run a positive control, i.e., a parallel sample you know should come up positive. Then if both the control and the sample are negative you know something might be wrong with the PCR assay. The new kit includes primers, probes and a dead control virus. But if you are wrong about the primers or the probe, you'll get a false negative.
This test is "is designed to detect highly pathogenic influenza A/H5 viruses from the Asian lineage associated with recent laboratory-confirmed infections of avian influenza in humans in east Asia and, most recently, in Turkey and Iraq." (CDC) If it suggests that an Asian lineage H5 is present, additional tests need to be done to see if it is N1 subtype. The announcement comes with this important qualification
Testing with the FDA-cleared laboratory RT-PCR assay should be conducted in conjunction with other laboratory testing and clinical observations to help diagnose influenza in patients who might be infected with influenza A/H5 (Asian lineage) viruses and to provide epidemiologic information for surveillance purposes. The test also will help to identify influenza A/H5 (Asian lineage) viruses in laboratory viral cultures. Definitive diagnosis of influenza A/H5 (Asian lineage), either directly from patient specimens or from viral culture, might require additional laboratory testing and clinical and epidemiologic assessment in consultation with national influenza surveillance experts. Negative results do not preclude influenza virus infection and should not be used as the sole basis for treatment or other patient management decisions.That's my "toy" version of PCR. Feel free to send in comments, corrections or clarifications.
This is an incremental advance in testing for H5 virus and we can expect similar developments as time passes. It is faster and can be done at lower biosafety levels than some other testing. No information was released as yet about its sensitivity and specificity so we will have to see from experience how well this works in the real world.
posted by Revere at
2/04/2006 11:25:00 PM
<< Home