NIH CLINICAL CENTER GRAND ROUNDS
Episode 2009-018
Time: 56:49
Recorded May 27, 2009
Use of Genomic and Proteomic Tools for the Diagnosis
of Infectious Diseases
Patrick R. Murray, PhD
Chief, Microbiology Service, Department of Laboratory Medicine, CC
Genomic Analysis of the Human Skin Microbiome
Heidi H. Kong, MD
Assistant Clinical Investigator, Dermatology Branch, Center for Cancer Research, NCI
ANNOUNCER: Discussing Outstanding Science of the Past, Present and Future – this is NIH Clinical Center Grand Rounds.
(Music establishes, goes under VO)
ANNOUNCER: Greetings and welcome to NIH Clinical Center Grand Rounds, recorded May 27, 2009. We have two speakers on today's podcast. Dr. Patrck Murray, chief of the Microbiology Service in the Department of Laboratory Medicine at the Clinical Center will present "Use of Genomic and Proteomic Tools for the Diagnosis of Infectious Diseases." He'll be followed by Dr. Heidi Kong, assistant clinical investigator in the Dermatology Branch at the NCI Center for Cancer Research. She'll discuss "Genomic Analysis of the Human Skin Microbiome." We take you to the Lipsett Ampitheater in the NIH Clinical Center in Bethesda, Maryland, where Dr. Paul Plotz, chief of the Arthritis and Rheumatism Branch at the National Institute of Arthritis and Musculoskelatal and Skin Diseases at the NIH, will introduce today's speaker.
(Music fades in, under VO)
GALLIN: Good afternoon. Welcome to Clinical Center Grand Rounds. Today we have two related topics in that they both deal with microbiology of the body. But they're going to speak on different topics so I will introduce each separately.
Our first presenter is Dr. Patrick Murray, chief of the microbiology service here in the Clinical Center in the Department of Laboratory Medicine. He'll discuss the use of genomic and proteomic tools for the diagnosis of infectious diseases. Dr. Murray's research interest center on the development of novel approaches to detect and identify microorganisms. He received his B.S. degree at St. Mary's College in California, his M.D. in microbiology and immunology at UCLA. After completing his post doctoral studies -- doctoral studies he served a two year fellowship in clinical microbiology at the Mayo Clinic. He held several academic positions before coming to the Clinical Center in 2002 including being a professor of medicine pathology at Washington University in St. Louis, and a pathology at the University of Maryland. He's also served as editor in chief for the 6th through 9th editions of the Manual of Clinical Microbiology. He's a fellow of both the American Society of Microbiology and Infectious Diseases, in the Society of America and serves on the National Committee for Clinical Laboratory Standards. I can tell you having watched Dr. Murray at the -- Having watched him at the Clinical Center it's been a true treat and I have seen his activities personally for some of the work he's done as a colleague with members of my own laboratory and NIAID, discovering the new bacterium. So Pat, welcome. [Applause]
MURRAY: Thank you, John. First of all I have nothing to disclose. At least financially. Hopefully I will be able to disclose some useful information. This is just a list of the objectives of my presentation. I have to say that when I was asked to give this talk it made me stop and think about some of the good things that have happened to me over the last eight years since I have been here. I think truly it's only a place like the NIH clinical center that I would be able to do the sorts of things I'm going to present today, just the resources that available are remarkable. I say that even -- I would say that even if John wasn't sitting in the front row. Historically the way we identified organisms was by microscopy culture biochemical testing and the detection of either microbial antigens or antibodies to those microbes. These techniques are limited. The phenotypic tests are dependent on gene expression. Microbial variability and are subjective interpretation of test results are intrinsic problems with all of the identification kits that we use, accurate identification is dependent on the quality of the database which will vary tremendously from company's product to company's product. Antigen tests are organism specific and antibodies are both specific to the organism and also dependent on the host ability to amount an immune response.
So our conventional ways of identifying organisms really I think are quite limited so my goal when I came here was to try to develop alternative rapid accurate identification methods both for identifying organisms we have culture and to detect identify organisms and clinical specimens. We focused on two complimentary methods for this. The first is to detect species specific gene sequences. We do this through amplification of the gene and then sequencing the gene. The second to detect disease specific protein profiles. We use mass spectrometry to do this. So what I'm going to do is talk about some of the applications of these techniques we have introduced and now routinely use in the clinical laboratory. For the genomic approach you have to identify specific genes that you're going to target. And there's three major groups of genes that have been used historically. The first, ribosomal RNA genes, the ribosomal RNA gene in bacteria, the 18 S and 20 S ribosomal RNA genes in eukaryotic organisms like fungi and pair sites. The second is the internal transcribed spacer region. I'll show you what I mean by that in a second. The third target is housekeeping genes. That is genes that are important, produce important products that the organism requires for it to be able to function.
Two examples that I'm going to discuss are the SECA 1 or secretory protein gene, and the second is the RE gene found in toxoplasma. The 16S gene is illustrated in this cartoon and what you see is you have double stranded segments and you have loop segments. The loop segments will allow a great deal of variability or substitutions in those loops where the double stranded segments are more stable. This gene has been referred to as the tax nomic map of bacteria as they evolve. Bacteria that are closely related to each other will have very similar 16S ribosomal RNA genes and bacteria that are distantly related will have divergent 16S ribosomal RNA genes. This gene is used to identify bacteria at a species level. The fungi also have ribosomal RNA gene, the 18S, the 5.8S and 28S ribosomal RNA gene. Also between that, the ITS or transcribed spacer regions.
This is a more useful target for doing our sequence analysis. I'll come back to that in a couple of minutes. After we identify what the gene target will be, we assemble a collection of type and reference strains of bacteria or fungi. We use those to build a database then challenge it with well characterized clinical ice lets. So that's the general approach we have taken for doing this genomic work. For the identification of bacteria, there's a large public database in the 16S ribosomal RNA gene. That's what we have relied on for doing or sequence analysis. We have found we have been able to identify most gram positive and gram negative bacteria using this database. In fact, this is the routine method that we use now to identify the more difficult to identify organisms or the ones where we don't get an easy identification with our conventional biochemical test. So initially what we'll do is run biochemical tests. If after 4 to 24 hours we don't get an identification or if it's an identification that seems inconsistent with the clinical information on the patient we'll use sequencing 16S ribosomal RNA gene to identify that organism.
For the acid fast organisms like no cardia and microbacteria there's not a good 16S ribosomal RNA database to separate these into individual species so we have targeted a different gene, the secretory protein or SECA 1 gene, for this we had to build the database. We have a large collection of species of no cardia and microbacteria to build the database. Then we challenged it with clinical islets, we have a large number assembled over the last 20, 25 years. The advantage of this approach with bacteria or microbacteria and no cardia, we can identify organisms in one to two days. In the example of microbacteria, it can take anywhere from one to two months so it's been a remarkable change. We routinely identify all of our islets of nocardia and microbacteria using this approach. For the fungi in the last five years we have relied heavily on sequencing the ITS region and have identified a large variety of different fungi I have listed the major groups and it includes all major groups identified using this sequencing technique.
I would like to illustrate one patient in particular to show the strength of this approach. This was a 48-year-old man who was -- who presented here, this is one his initial photograph, had an undefined immunodeficiency. The infection developed in his cheeks, sinus, palate and skin. So a very extensive infection that's here. For multiple skin biopsies we were able to isolate a mold. The mold grew in one week. So not too slowly. But took nine weeks for the characteristic structures to develop and that's required to identify modes of transmission of techniques of morphologic identification. The mold was a spore but we western able to identify at the species level. Two approaches were used for molecular sequencing, the first a 28S ribosomal RNA gene and what we found is that the sequence we were able to obtain was not in the database. It turns out spore is not in the database for the 28S ribosomal RNA gene. However we were able to identify the gene as (indiscernible), the reason it wasn't in the 28S ribosomal database is this is a cucumber mold. The reason it was in the ITS region is -- database is because this database was originally developed for environmental organisms. Many models we see in our patients -- molds we see in our immunocompromised patients are these environmental islets.
So just to summarize to identify by conventional techniques took up to ten weeks. A week to grow it, another nine weeks before the characteristic structures develop by sequencing by sequencing in a couple of days. What I have talked about is taking organisms we have isolated in culture but what's more exciteing is the ability to be able to look at a clinical specimen, microscopically, see an organism and then to use molecular approaches to identify the organism. This direct approach avoids delays associated with culture and traditional identification methods. You can get a specific definitive identification again in one to two days, and we have been able to use this to identify a wide variety of different organisms.
Now, it's not without problems. There maybe relatively few organisms present that we may not be able to amplify the target organism. If the tissue is fixed in form Lynn that can degrade the DNA target so it's a larger challenge to extract the DNA in the usable form. You can have non-viable contaminating organisms or DNA from organisms present in the sample so interpretation of the results becomes important. And finally you have to select the right gene target.
So I would like to do is illustrate three examples of where we use this approach in the last couple of months, these are all patients that we receive specimens from the last two months where this work was done.
The first is a 45-year-old HIV infected man seen in April at T.W. med center and they suspected the patient had a microbacteria tuberculosis based on epidemiologic grounds. A needle biopsy was collected and stain was demonstrated to show acid fast organisms though only a few were seen. The biopsy however was not sent for culture, it was just all sent to pathology where it had been fixed in formalin so we received a portion of this block and surgical pathology who developed an extraction procedure was able to extrack the DNA -- extract the DNA. Then the SECA 1 was targeted and in my laboratory, the organism we were able to identify was microbacterium AVIUM so not tuberculosis as suspected.
The second patient had endocarditis here in D. C. Cxfc conventional blood were negative and aortic and microvalves were replaced. Unfixed tissues from these valves were sent to us to do the DNA extraction. Sequencing targeting the 16S ribosomal RNA gene, and the amplified product then was sequenced. The diagnosis was bartanela QUINTANA. What I have shown to the side and the top field is bartanella by silver stain and the bottom is what the colonies would look like if cultured. The problem is this organism takes about a month before it will grow in vitro culture. And that was not the way it was cultured at the VA med center. So this really was the only way this diagnosis was made.
Then the last example was a patient, another HIV patient seen at Washington Hospital Center with multiple brain lesions. There is a high suspicion of malignancy based on the MRI but the brain biopsy was negative. I'm showing the picture of an organism -- whoops. Showing a picture of an organism here, unfortunately this organism was not seen, if it had been seen the diagnosis would have been much simpler, but we received formalin fixed brain tissue. The DNA was extracted in surgical pathology. The product was amplified in the microbiology laboratory in this case because toxoplasma was suspected. We targeted the repetitive elm gene. This was a target that was developed for PCR assay we developed in our laboratory for toxoplasma. And again, we were able to make the diagnosis of this patient's disease as toxoplasmic ONDIA.
I could have cited other examples of where we were able to directly analyze clinical specimens in very rapidly make the diagnosis using the genomic approach. Now, I think that's neat but actually the next portion is even more exciting. What I'm going to do now is talk about the use of proteomic tools and specifically I'm going to talk about MALDITOF. That is a acronym for matrix assisted laser desorption ionization time of flight mass spectrometry. Which is why we say MALDITOF. It's a simple procedure. What you do is you take a organism isolated in culture, you take a small loop of the the organism and wash in ethanol. You mix it in formicsc acid and acetyl nitryl. You spot it on a target plate. These are reusable plates and overlay with a matrix. The whole process takes about five to ten minutes to do this. It is then inserted into an instrument, it's ionized with a laser beam and the proteins are separated as they move through an electric field that's the time of flight component of this. Until it goes to a detector. They're separated based on their mass and charge. So what you end up with then after again, minutes is a profile.
This is a profile of E-Coli and you see there's a whole series of peaks. These profiles are very complex so it does take sophisticated software to do the analysis but all that is done now. This is an E-Coli. Here is an example of four other bacteria. Just grossly four other organisms two fungi, two bacteria. Grossly you can see they have a variety of different profiles. These profiles turn out to be quite unique to the individual organisms. In addition they're reproducible. This is an experiment with under completely different conditions and you see the same profile. So essentially what you have is a protein signature for each organism and a signature you could on tape in minutes using this procedure. So the limitations. You need a large number of organisms so that generally restricts the identification to organisms isolated in culture. The organisms should be in pure culture but the different mix of organisms could be identified by this procedure.
So the last six months Steve Drake in Critical Care Medicine and Lindsey Stevenson one of my fellows developed and evaluated databases for yeast. Databases didn't exist for these organisms. They evaluated and supplemented a commercial bacteria database and they demonstrated this technique is used to identify bacteria in yeast that are present in positive blood culture. So just to summarize the data very quickly, for the yeast the database was developed using types and reference strains for 40 species. They were identified as the 40 clinically important species of yeast so it's a comprehensive database. There was then challenged with more than 100 clinical islets from our NIH clinical collection and 95% of the organisms were correctly identified as the species level.
This identification the ones we were not able the identify organisms we saw profiles but they didn't have a high enough score to be accepted. So we didn't get any incorrect identifications and the 95% that were correctly identified were identified within a few hours after isolating them. For nocardia we did the same aprove, we built the database with 43 -- approach, we bill the database with 43 reference strain, 74% of the islets were correctly identified at the the species level T. Misidentification, the non-identified organisms were not identified at a species level but were lumped in groups of organize. S closely related nocardia strains.
For the bacteria, virtually all the gram positive and gram negative bacteria we tested in this system can be identified at a species level. There are selected organisms that are not correctly identified and that's what we're working on is to define exactly which organisms those are. But again, I would say probably greater than 90 to 95% are easily identified at a species level. Finally, using our positive blood culture we found 75% of the organisms that are isolated in blood culture broths could be identified using this technique. And again, the ones that weren't identified generally had an unacceptable profile because there weren't enough organisms present in the blood culture broth.
So turns out in our hand to be a very powerful tool for identifying organisms. Now, in the last minute what I would like to do is show you how you can combine the genomic approach that I started with with a proteomic approach. And identify even small numbers of organisms present in specimens.
What I'm going to do is refer to an instrument that was developed by IBYSS 5,000 biosensor. It illustrates how a tool like this can be used. What they do is they target genes, ribosomal RNA genes, housekeeping genes, these target genes are amplified directly in the clinical specimen. The way the test works is you can amplify a series of different genes in that target specimen. You then separate the DNA strands and then you separate -- you denature and then separate them by electron spray ionization mass spectroscopy. So the same as the MALDITOF approach where you separate target products and you run them through this mass spec system in order to get a series of peaks. These peaks are peaks of DNA strands. And the DNA strands are going to be separated based on the length of the strand as well as the concentration of A, T, C and G in those strands. Theoretically using this approach you can identify multiple organisms directly in a clinical specimen so you don't even have to culture the organisms.
To illustrate how this can be used, what I would like to do is present one last patient. This is a 16-year-old boy with (indiscernible) to develop recurrent painful leagues on the arms and legs. The biopsy shows vasculitis but no -- others have been seen in Brewton's patients here at the NIH and were demonstrated by a unusual organize. , it was called flexispiral. Now it's HELICO BACTER. So this patient was referred to NIH where a histology was performed. It then revealed thin curved rods. The arrow is pointing at them, I can barely see them from here, I know you can't see them from where you're sitting so just illustrates the fact these are very difficult to see organisms but that morphology is typical of a HELICAB BACTER. Culture grew a group B streptococcus. This grew in one day. We didn't think this organism was important, we thought it was colonizing the wound. So we continued to incubate the specimen and after about six days a thin growth of film of organisms was seen on the culture plate around the group B strep colonies.
When we did a gram stain and orange stain of that material, we could see very thin curved gram negative rods. Unfortunately when we sub cultured it, it failed to grow so we weren't able to maintain the viability. What we did is sent the plate out to California where IVIS bioscience is located. They amplified a variety of gene targets, denatured the DNA and separated the strands by their mass spectroscopy system. What they saw then was this: We have three peaks, three pairs of peaks. First is group B strep, CALALITICA. Second is a bacterium contaminating the culture because of the various handling, and the third was another pair of peak which is corresponding to where we thought the peak would be. That's how we confirmed that film of growth was HELICOBACTER isolateed from the patient.
So in summary in the last few years we have refined our ability to identify organisms by gene sequencing and mass spectrometry. Sequencing of single or multiple target genes can identify most microbes definitively. The down side of the sequencing approach is the technique of labor intensive somewhat costly and can take one to two days to get a result. In contrast, the use of mass spectrometry offers the potential of rapid inexpensive and highly accurate identification of both bacteria and fungi and large numbers of organisms need to be present, we maybe able to get around that by a second approach using gene sequencing combined with mass spectrometry like we did in the last example. So I think that the sort of biochemical approach that we've taken historically -- I know clearly in my laboratory it's replaced using these techniques and I'm just enthusiastic too sew what the future is going to bring because I think there's really remarkable work here. Thank you. [Applause]
GALLIN: That was terrific. Thank you. Is there any hope that you'll be able to use these techniques to detect antibiotic sensitivitys?
MURRAY: It's interesting. There are certain markers that would be clear like for meth zone resistant stafforiuos. I think many of the problems we see in drug resistance is related to beta LACTOMASE production. Hundreds have been produced. So I'm not sure it's going to be terribly helpful there but for targeted antibiotic resistance, it might be.
GALLIN: Any other questions?
QUESTION: What I understand is in most cases here you had a candidate organism that you were going after in the clinical -- with the clinical specimen. Have you had cases where the diagnosis that you've made molecularly has completely changed your perspective on how to treat the individual?
MURRAY: Again, if you looked at the examples that I gave, for the toxoplasma we suspected toxoplasma so we targeted a gene specific for toxoplasma. The avian we suspected a microbacterium based on what we saw by the acid test stain. But the organism that we identified was avian, not tuberculosis so the treatment was different for that organism. For the bartanella isolated in the heart valve tissue, it was suspected but the technique that we used would have identified any bacterium that was present there. So it was a broader approach. For the IVIS system that I presented at the HELICOBACTER, that can identify any organism so my goal would be to use techniques that can be broadly applicable rather than focused like we have historically done with antigen testing or antibody testing.
GALLIN: Thank you very much. We'll move on to our second topic, the microbiome has become a major area of emphasis for the NIH. And today our second speaker, Dr. Heidi Kong, assistant investigator in the Dermatology Branch of the NCI Center for Cancer Research is going to present the genomic analysis of the human skin microbiome.
Dr. Kong's research focuses on the interactions of skin and external conditions including how systemic medications affect the skin. The effects of microbes on skin disorders and protocol driven therapeutic trials for dermatologic diseases. She received her BS degree from Stanford and the M.D. from Baylor College of Medicine. After interning at Baylor and completing a residency in dermatology at Duke she came to the NIH in 2005 as post-doctoral clinical research fellow at NCI's Dermatology branch. In June Dr. Kong will graduate from the NIH Clinical Center in Duke University's training program in clinical research earning a master of health science and clinical research at Duke. Dr. Kong's membership includes Society of Investigative Dermatology, the American Academy of Dermatology and the Women's Dermatologic Society. Welcome.
KONG: Thank you, Dr. Gallin. It's a pleasure to be here to discuss with you our research using genome analysis of the human cutaneous biome.
The work I present here today is a collaboration spearheaded by investigators from the dermatology branch of the NCI and investigators from NHGRI and we are also doing other research in collaboration with Dr. Patrick Murray as well as investigators from NIAID.
At this time I want to bring your attention to the image on the top right corner. This is a scanning electron micrograph of microbes on the skin underneath the toe nail. You can see a lot of microbes here. It's this abundance of skin microbes that we are investigating. My disclosure statement is I have no relevant conflicts of interest.
So why is it that dermatologists someone like myself is interested in examining the microbes opt skin? So I have depicted here three dermatologic disorders fairly well known. Each of these three acne atopic dermatitis and psoriasis we know as being inflammatory disorders, not truly infections of the skin but in all of these three cases there are associations with specific microbes that we have found based on prior research. But also all three of these respond to anti-microbial approaches.
So there appears to be a role for microbes in this -- in these specific diseases. I'll go through each one, acne on the left side is very common. Often affects individuals in their teenage years and bacterium acne is implicated in the pathogenesis and most of these studies have been culture based. And dermatologists treat patients with topical antibiotics and systemic antibiotics with good efficacy.
Moving on to the middle picture, this is a child with atopic dermatitis, it affects 15% of U.S. children and we commonly see for those who manage patients with atopic dermatitis that the disease flairs in association with stafforeus that we can see on culture. Over 90% of AD patients are culture positive for stafforueos. On lesional, non-lesional skin. When we use anti-staff approaches we can actually have good efficacy in controlling these patients. Even if they don't have an infection. On the last picture to the right is psoriasis. It affects 2% of the U.S. population. And there is a subtype called drop like psoriasis where we see flairs of disease or the eruptive -- the onset of the disease with streptococcal infection. So we can see here that dermatologists are quite interested in the role of microbes but we don't specifically understand what the role is in skin disease and that's why we elected to use some genomic approaches to studying microbes on the skin. So after my talk today participants should be able to differentiate between culture based and genomic approaches for studying bacteria.
To summarize findings of the human cutaneous microbiome surveys and recognize potential medical research plaincations of microbial genomic analysis.
So I have divided my talk first into microbes in the skin and then I'll discuss our research that we have done using these genomic techniques to study the human cutaneous microbiome. Specifically what we're doing in atopic dermatitis. First microbes in the skin. So when we talk about the human microbiome what are we referring to? The reason this is important in that is that based on some of these genomic studies it's been observed that the microbial cells in and on humans outnumber human cells by a factor of ten. And many of these microbes are not cultivatable and Dr. Murray, the next segue, he is already described many reasons why it's difficult to cultivate certain microbes. And when we talk about microbes, the human genome in this regard we talk about the interaction between human genes and the genes of our microbial partners.
So why do I say microbial partners? Because microbes with perform metabolic functions that are important for the maintenance of human health. They synthesize vitamins, they are important in the renewal of gut epithelial cells and they are also vital in the development and activity of the immune system.
Most research in the human microbiome focused a lot on the gut. What about the skin? The skin is also very important in the development and activity of the immune system. Are some of the other functions of the skin preventing moisture loss, temperature regulation. But skin being the largest organ in the human the first line of defense in protecting us against the environment. It's important in protecting us from chemicals, from the damaging effects of ultra violent light but also a lot of pathogen, pathogenic bacteria. The media likes to talk about flesh eating staff but we are beginning to realize that microbes also play a helpful role in maintaining some of the health of humans. But the skin is very complex. There are many microenvironments of the skin.
I have divided it here opt left side being microscopic level of cutaneous microenvironments and on the right side more a macroscopic level looking at the different types of cutaneous microenvironments. On the left side you can see in this picture that it's quite complicated the hair follicles, there are sebaceous glands, so is the topography of the skin is not uniform. On the right hand side the surface of the human body is also not uniform. There's a lot of variability. You have sebaceous site, the forehead, central face, the upper chest and back. These are areas with the highest density of sebacius glands. There are moist areas and dry areas such as the buttocks areas without creases that are otherwise smooth and exposed. But unfortunately most of dermatologic literature when we look at bacteria composition of the skin it's mostly based on culture data. And if you talk to any dermatologist, the dogma they teach us is the commencial bacteria on the skin are primarily staph.
From this pie chart you can see when you do culture data that is we see, the orange is staff species and it's over 50% is this orange but there are other islets that can be found some organize. S that Dr. Mary mentioned -- Dr. Murray mentioned but there are other approaches. So our projects, we used 16S ribosomal RNA gene analysis and fortunately Dr. Murray has explained this but it's universal and prokaryoteses. By sequencing the microbes we can compare their conserve regions and their variable regions and it can be used to identify bacteria. Often these species are depending on who you talk to classified by those who have 97% or 99% sequence homology.
On the right here you can see an image of the e-Coli, 16S small sub unit ribosomal RNA and highlighted area are areas of primers so these are some of the areas you can use to help sequence these organisms. So we use this technique, we first set off to do a pilot study examining the microscopic level of the skin microenvironments. The left-hand side. So we had five healthy volunteers we examine, individuals who had no chronic medical conditions, no obvious chronic infections, and no obvious chronic dermatologic conditions. We sampled the anti-cube tall FOSSA on the right and left of all five individuals after a standardized skin preparation regimen. And we use three different techniques because we wanted the as certain whether there was a difference in -- if we sample them in the data that we obtained.
We performed swabs essentially using something like a moistened Q tip. We also performed scrapes using a sterile scalpel to scrape off visibly scrape off dead sites and perform full thicken buy I don't know schism there was ooh question whether or not we would miss something without a full thickness skin biopsy because of the skin with the hair follicles and the ducts. So we wanted to see first if any -- if any of these methods were better than others and -- methods were better and to see if there was a core microbuy yes, ma'am that appeared.
Here in this data you can see this large bar graph on the bottom here you can see the subject 1, 2, 3, 4 and 5 and the different sampling techniques we use. Left-hand side the relative abundance here to 100%. Mostly they're very similar. The bottom is the legend blue being predominantly proteobacteria. And some of us may not be familiar with the different organisms in these FYLA, I have included examples, Actin bacteria and fermicutes such as staff lowcoccus in yellow and bacteria DIDES. And there are many other phyla out there and they did not comprise a significant amount of the sequences that we found in these five subjects. And fortunately for us you can see that they're very similar in comparing the different sampling methods that -- whether it was a swab, scrape, or a biopsy we found similar results. In fact they were identical when we assess the different species that we were examining as well as the relative abundance.
So fortunately for us swabs are -- were okay to perform on future studies. But also you can see that they are very similar so there appears to be a core microbiome on the skin of these patients accept for this individual, what happened to subject 2. This individual was colonized with stafforeus and you can see the change in their samples being colonized, what it does to the rest of their microbes. Then we looked with the pilot data sampling with the swabs was effective and there appeared to be a core microbiome of the skin so we moved to the macroscopic level.
We wanted to validate what we found from our pilot study. So we selected over 20 sites on the skin for one, we wanted to make sure that we found diverse microenvironments such as the -- inside the nose versus the AXILLA. We also selected this is sites because these are sites that have -- that there are certain dermatologic diseases that have predilection for these sites so we carefully selected sites for many dermatologic reasons but if you look on these color coded blue, green and red. Blue sebaceous areas, green moist areas an red being dry. On the left-hand side you can see GLUABELUM, external auditory canal and on the right side. Green or moist areas are NARES, anti-cube tall FOSA, umbilicus. And the plantar heel. Initially we categorize plantar heel as dry but I will get back to that in a second.
So the dry area forearm, palm and buttock. Those are usually flat areas without occlusion. So when we initially looked at heel we said that's a dry area. When analyzing sequences it fell in more in line with the moist areas. And that would be related to that most people wear shoes. So it is an occluded area. And that would make sense with why it fell into the moist category. Here is the data the bar graph for all those ten subjects. When we separated them into sebacous sites, moist and dry sites you can see patterns emerge. For these ten individuals the left axis is the relative abundance that bacterium predominates. That makes sense because P acne itself is very lipophilic.
When you look at moist sites more green starts to develop in these bar graphs and green being proteobacteria or something like pseudomonis. That's other bacteria that starts to be visible in this sample and when you look at the dry sites there's more variability. But there's more data within you look at the individual sites and here this is the first four healthy volunteers that we had in our study and looking at the pace you can see most of them at the SEBACIUOS site have bacterium except for healthy volunteer number 4. This individual was also colonized with stafforeus. You can see at the site of the reticular crease in somebody colonized with stafforeus. You can see this very site specific even though healthy volunteer number 4 was colonized by stafforeus the back was still bacterium for all four subjects here.
This is a moist area, a folded area. There's a much more green. This was a bit of a surprise for the clinicians in that we don't typically think of it being everywhere. Especially when we talk about patients such as atopic dermatitis when we culture the anti-cube tall FOSA we don't normally see it. We're looking at stafforeus so we start to see maybe there's a change in the microbiome in subjects who are healthy, otherwise healthy or those who have atopic dermatitis or other skin diseases. And the NARES you see more bacterium in this site in addition to the appropriate bacterium and to the UMBILICUS there's more variability so in this study we performed over 20 different skin sites we saw there was a core that we saw among the subjects but it depended -- the actual sequences that we found depended on the actual site. So important as we move on to looking to disease states we have a baseline of what we consider a normal microbiome in those skin sites.
How do we compare the 16S data and culture data? Most dermatologists think okay we're going to see staffhominus. Sure enough when you look at the cultures on -- in the bar graphs here they see mostly orange in the culture data from the ALAR crease and UMBILICUS. Survey data opt left-hand side, the orange area diminishes, it decreases so that there are other microbes we are able to detect using sequencing genomic technology that are not being captured when we culture them due to -- most likely due to media conditions or culture conditions that are not as amenable to these other bacteria to grow. Moving on. Now that we have done our wide survey of healthy subjects we have now moved on to looking at a disease state atopic dermatitis. So atopic dermatitis is a reminder, affects 15% of U.S. children. Most children outgrow it by the time they reach adulthood. However there are some patients who persist and have severe disease through adulthood. Stafforeus is very important in these patients. Here are some clinical images of patients we take care of. You can see in the anti-cube tall FOSA and poplitel FASA that's predilection towards these sites.
Why? We don't know. But you wonder if these are moist site, does that have something to do with it but it still needs to be investigated. This is a chronic pure etic skin that's difficult to manage. It's important in our healthcare system in that over a billion dollars is spent a year in direct cost in managing and caring for atopic dermatitis patients. So it's a very costly disease as well. So that's why we are looking at atopic dermatitis. And it is a very complex disease in that it is -- works together with the skin barrier, the immune system and microbes in this disease. We know that they have a defective immune system in that there's a atopic dermatitis, asthma and allergic rye nightis run together and they have elevated iges to stafforeus. But a defected skin barrier is very important. In the last three years there's been a lot of research done looking at the flagrin mutation. Flagrin is an epidermal structural protein. When that is mutated there is -- it's associated with atopic dermatitis and inherented in a semidominant fashion certain populations so we know a defective skin barrier is important. This is a complexin play among these three difference area bus we are focused primarily on microbes and the role that microbes play in this disease. We are evaluated moderate to severe atopic dermatitis patients and again, focusing on pediatrics because this is population mostly affected. We are performing skin swabs of a classically affected the anti-cube tall increase the poplitele FASA three different time points at baseline or quiescent disease states after they have been treated or post flair. Our goal is to analyze whether or not -- whether there is a shift in the microbiome.
Why are we looking for selective shifts in the different populations of microbes that we see? Well, some of you maybe aware that the gut microbiome has been evaluated and examined extensively. There's data that demonstrates there are select active shifts between the gut microbiota in lean versus obese individuals. So we wonder if there's dermatologic reasons where we see the shift in microbiota. Ideally if we're able to identify the shifts or if there are unusual bacteria that are isolated from atopic dermatitis but not healthy subjects we can identify possible therapeutic options to rationally develop treatments to assist these -- help these patients. Our goal is to analyze these subjects and to eventually determine treatments for this costly in this disableling disease. I want to acknowledge the individuals who have spearheaded this collaboration, Dr. Maria Turner and Dr. Julie SEGRAY and individuals in our lab, especially Dr. Betsy Grice.
There's a lot of supportive players in this project including Dr. Mark UDI and NHGRI and Sean con Lynn, Joey Davis nurse practitioner now with NIAID as well as LORENA GOVAR important in sampling these patients. But we are working with this and other projects on the skin microbiome with Dr. Murray as well as investigators from NIAID, Steven Holland, Alexander freeman and we're especially indebted to the NIH sequencing center for sequencing our samples. But I do want to mention that this -- the AD human microbiome project is part of the HMP which is the NIH road map initiative to assess microbuild diversity of healthy individuals at five sites. The gut, nasal mucosea. The oral, the vaginal and the skin. And in addition to sequencing the bacteria, one thing that the -- this initiative is looking at is to correlate changes in the microbial communities with disease states of which atopic dermatitis is one. So this is a larger project ongoing across the country. If you have children with moderate to severe atopic disease dermatitis or know of others who have moderate to severe atopic dermatitis we welcome them for asking more questions and possibly participating in our study. Thank you very much for your time. [Applause]
QUESTION: Dr. Kong, one of the questions I have is whether you have any inclination or suggestion how these microbes do something worthwhile to the skin in maintaining body health?
KONG: We know that a lot of the microbes we -- I guess one of the things to think about is we hope to see whether or not the microbes that help develop our immune system may play a role in determining -- responding to antigens. There is data on epicutaneous sensitization to antigens and whether or not that may help develop our immune system and help fighting off other bacterial pathogens so we think that is beneficial in a sense in developing our immune system but it's possible that the presence of certain bacteria may help proliferate oils on our skin to help maintain skin health so there's a lot -- it's an open area at this point and not a lot of research is done on the microbes on the skin and how it benefits us. But that is definitely something we look forward to investigating. (off mic)
KONG: So the question was that there have been studies to look at we look at the atopic dermatitis for example as a predilection for skin sites that maybe moist skin folds and whether or not there are studies that alter the microenvironment to see if that makes a difference. And there is in the pediatrics this year there was a study looking at bleach baths for atopic dermatitis. Literally decreasing the bacterial burden. And there was efficacy with believe bleach baths in managing children with atopic dermatitis. So there is some role that appears there does appear to be if you challenge or change the something, whether it's micro-- microbial burden or whether it's baths themselves, whether or not that would play a role. The question of changing the moisture. Well, that's where the studies have been done looking at emollients on the skin. Because we know that the defective skin barrier plays a role in atopic dermatitis. So increasing the emollients and repairing that skin barrier is important in controlling the disease as well. But looking at whether or not moisture changing the moisture, that has not been done.
GALLIN: Okay. I would like to thank both speakers for a terrific Grand Rounds. Thank you. [Applause]
ANNOUNCER: You’ve been listening to NIH Clinical Center Grand Rounds, recorded May 27, 2009. Our speakers were Dr. Patrick Murray, chief of the Microbiology Service in the Department of Laboratory Medicine at the Clinical Center who discussed "Use of Genomic and Proteomic Tools for the Diagnosis of Infectious Diseases." He'll was followed by Dr. Heidi Kong, assistant clinical investigator in the Dermatology Branch at the NCI Center for Cancer Research who presented "Genomic Analysis of the Human Skin Microbiome." You can see a closed-captioned videocast of this lecture by logging onto http://videocast.nih.gov -- click the "Past Events" link -- or by clicking the "View Videocast" link on the podcast homepage at www.cc.nih.gov/podcast. The NIH CLINICAL CENTER GRAND ROUNDS podcast is a presentation of the NIH Clinical Center, Office of Communications, Patient Recruitment and Public Liaison. For more information about clinical research going on every day at the NIH Clinical Center, log on to http://clinicalcenter.nih.gov. From America’s Clinical Research Hospital, this has been NIH CLINICAL CENTER GRAND ROUNDS. In Bethesda, Maryland, I’m Bill Schmalfeldt at the National Institutes of Health, an agency of the United States Department of Health and Human Services.
