Tip:
Highlight text to annotate it
X
Jonathan Zenilman: Good morning. Thanks for inviting me and I
just want to check, am I okay with the sound and stuff like that? Okay, because we're being
taped.
What I was asked to do -- this is actually part of the NIH series and I was asked to
provide some insights on how genomics has changed infectious diseases. So what I thought
I would do this morning is take you on a little bit of an odyssey, which includes my personal
odyssey and some of the work that I've done and how my life has changed, how our diagnostics
have changed as well, and then show you a couple of case examples where the advent of
genomics technology has changed our ability to diagnose, manage, treat, and also understand
the epidemiology implications of infections.
First of all, I have no conflicts. 2003 was an interesting year. Besides Dr. Collins'
editorial in The New England Journal, it was also the year when the human genome was sequenced
and published. And it was also a year where we had the first major case where infectious
diseases were rapidly diagnosed in an outbreak setting exclusively via genomics, and I'll
show you. Now, I hope some of you probably remember this. And these still exist but,
you know, culture -- there are many -- for example, I have infectious disease fellows
now who've never seen a Gram stain. So the world has changed substantially since the
good old days.
This is gonococcal -- this is gonococcal -- this is neisseria gonorrhoeae, and here we have
the candle jar which used to be used for transporting these bugs. However, as you know, cultures
take 24 to 48 hours to process. In an era when we need rapid diagnostic information,
this is a problem. Cultures are prone to overgrowth and all kinds of problems and as we see centralization
of laboratory services, it's difficult -- there can be major transport problems. For example,
if -- I did this once, which may indicate that I'm a bit of a strange guy: If you go
to the FAA website at 4:00 in the morning and see what planes are in the air, it's FedEx
and Quest Labs because your laboratory -- your laboratory work may go to a Quest or a Labcorp
lab in Salt Lake City, depending on what specifically you're ordering. So the days when you can
go down to your local hospital microbiology lab and speak to the tech to find out what's
going on are -- in many cases, don't exist anymore.
So, in principles, the advent of the genomic era has resulted in many opportunities for
infectious disease diagnosis and management. Basically, you wanted the tech -- the ideas
of the tech species-specific DNA amplified usually by amplification technology, often
polymerase chain reaction, but there are a whole host of newer technologies which have
come online over the past few years. One of the key elements is that bacteria DNA have
unique 16S ribosome sequences which can be used to pluck bacterial DNA out of a mixture
of human, animal, or other types of DNA. This should be matched -- it's the DNAs that matched
to libraries which have the specific sequences and right now there's over 2,000 sequence
bacteria who've been fully sequenced and this number is increasing exponentially. It's linked
to a detection system and the detection system technology has been really advancing incredibly
rapidly to the point where these are now desktop machines with small footprints and I'll show
you.
And just for those of you for review, I know most of you have seen this, PCRs -- preliminary
chain reaction is the typical amplification system. This is your target DNA. The strands
are separated. There's a polymerase which actually catalyzes the attachment of another
strand and then this goes on, usually via a thermocycler, which is a heating -- which
is where the system is heated to 95 degrees Centigrade and then rapidly cooled, which
anneals the DNA. And basically within usually an hour you can get 30 amplification steps.
If you remember back from when you were a kid, if you take a penny and double it, after
a month you'll have a million dollars. So therefore this is a tenth of the eighth amplification
with 30 cycles.
What's happened since the 1990s, nucleic acid diagnostics were first commercialized in the
1990s with the early examples with wide commercialization being actually in STDs, my area, and *** viral
load. An *** viral load is actually a DNA count of *** virus in the blood. It's not
an -- and that's one of the first wide commercial uses of DNA technology to manage patients.
But there's been, since the early 1990s, many non-cultivable pace pathogens have been identified.
For example, Whipple's disease is actually caused by a non-cultivable GI pathogen.
After 2001, there were major investments in technology and microbial detection for obvious
reasons. There's been simultaneous investment in the Human Genome Project and other sequencing
projects such as the NIH's investment in TIGR, which is the sequence repository. And as I
mentioned, approximately 2,000 organisms are fully sequenced at the present time, with
large numbers increasing every year. The current trends in this area we see are the commercialization
of discovery into diagnostic tests. There's -- the development of rapid, and actually
we're going to see in the next five years point-of-care diagnostics. Genomics as clinical
management tools, I'll show you a couple of examples. Bacterial population genomics and
its impact understanding the bacteria populations and the ecology bacteria. Also, there's an
interest in host genomics. For example, there are certain host genotypes which dictate how
people will respond to infections or therapy. Hepatitis C is a very good example and I'll
show you an example of that.
Microbiome projects, which are looking at the ecology of the mucosal bacteria and we're
looking at these as communities as opposed to single organisms, and then we're seeing
through the commercialization process the development of expert and bench-top systems
-- basically turnkey systems where basically all you do is insert a sample and you get
a result for all the system behind the scenes through the black box. And there's a great
example with TB which I'll show you.
So this is -- remember SARS back in 2003. This was the index case in Hong Kong in the
Hong Kong Hotel, there was a Vietnamese physician who took care of him; who died. The patients
-- there was aircraft -- there were passengers who went to Toronto and two hospitals in Toronto
were closed down and there were large numbers of deaths due to a novel coronavirus. Well,
this virus -- this is the SARS virus and its relationship to avian other -- to avian and
other coronaviruses. This problem was solved within 30 days using sequence technology where
the pathogen was identified, a diagnostic test was developed, and control measures were
instituted. This virus was never cultured. This was solely identified from DNA. So this
is the first example where in a large-scale outbreak situation very much like the movie
-- who saw -- very much like the movie "Contagion," where this was actually solved within a month
but by the advent through completely through DNA technology.
If we go to my area in terms of sexually transmitted diseases, nucleic acid amplification tests,
which is the acronym we use, are the dominant mode of gonococcal and Chlamydial testing.
It's very hard to get a culture anymore. This can be used -- the advantage of this, which
we identified very early and is widely used, is that this can be used for general and non-general
samples but in field settings such as schools. So if you're in Africa, sub-Saharan Africa,
doing an NIH-funded study, and there's a lot of them, or doing a control program, or if
you're in Philadelphia or New Orleans, where they have large school-based screening programs,
basically these are done by urine or self-administered vaginal swabs where there's no transport problems,
the material is stable, and you -- it can be transported to the lab and you can get
diagnostic accuracy.
In fact, this resulted -- the nucleic acid amplification resulted in a conundrum for
the FDA because it turns out that the sensitivity of nucleic acid testing for STDs and most
other infections is actually 30 to 40 percent higher than culture because you don't have
death during transport, or you don't have fastidious organisms. So, therefore, from
a regulatory standpoint, there was a lot of discussions because using traditional reference
standards, i.e., culture, the specificity was low. However, in fact, the problem was
that the reference standard which was used had a low sensitivity. So has caused a re-thinking
from a regulatory standpoint of how you define the gold standard. These also could be turned
around in 24 hours in commercial labs when in fact the test takes six hours to run. It
can be multiplexed, but, and this is a big problem now with gonorrhea and with other
organisms, you cannot identify resistant bugs unless you know specifically what you're looking
for. So, you can't -- you won't know if it's resistant to tetracycline or penicillin because
you're not growing the bug. Now, you can -- if you know what the genes are, you can identify
them, but this is a challenge.
So this is a great example of how we adapted gonococal and Chlamydial testing into a field
setting. This was a household survey in Baltimore which we did in the late 1990s, which was
a population-based field survey. This was published in JAMA. And what we did is a household
prevalence study where we did a population-based sample, knocked on people's doors, took a
behavioral survey and asked, "By the way, would you pee in a cup for us?"
[laughter]
And what you find is that -- [laughs] -- among -- this is for Chlamydia and gonorrhea. We
had 15 percent of black women and 6.4 percent of black men were Chlamydia positive, 9 and
3 percent were GC positive. So there -- sorry, so combined, between 15 percent of women and
6 percent of men were either positive for Chlamydia or GC and similarly for non-black,
2.8 and 1.3 percent. So basically this is prevalent gonococcal and Chlamydial infection
in Baltimore in the late 1990s. Interestingly, the study was repeated three years ago and
found substantially lower prevalences. And this study been done -- the same study's been
done in a number of different areas. It's been done in job corps applicants and military
recruits. It's part of the -- it's part of the NHANES. And because you're not dealing
with transfer problems it's easy to do.
C. difficile, Clostridium difficile, which is a major pathogen hospital-acquired infections
represent major, major advances in diagnostics. The clinicians among you will remember that
the old way was taking a C. diff culture, which we recommend not to do because culture
may be false positive because the bug may be there but not causing diarrhea. The toxin
assay, which was the most common one, required a stool filtrate then being laid on tissue-cultured
cells in the laboratory. This takes about a day or two to do. Meanwhile, the patient
has got lots of diarrhea. It's also technically complicated and expensive. About eight years
ago there were ELISA assays but the sensitivity is 80 percent. But right now the current standard
is PCR of the toxin A and toxin B genes, which has a -- which can have a six hour turnaround
time. So, this has all happened, again, within 10 years.
Now, but what happens when you start using the new tests? Well, Dr. Rothstein is that
equality and he probably faced -- he'll appreciate this problem which we faced -- he's, sorry,
Bob's the VPMA [spelled phonetically] [unintelligible], he has quality under his portfolio. We actually
-- this is our C. diff rate at Bayview through the 2000s, and in 2008 we instituted an aggressive
antibiotic management program, which reduced our C. diff rates by about 50 percent, and
that's that dip here. And then it went up in late 2009. But what happened? What happened
was, we changed the testing algorithm in the lab and went from the EIA to the PCR. The
PCR has about a 40 percent increased sensitivity, so when you institute these tests, you're
going to increase your rate -- your reported rate. Your actual rate stays the same. But
this has to be explained to people -- you know, you have to understand this because
otherwise your quality people or administrators may actually have a heart attack when they
see this.
[laughter]
Now, similarly, this is what happened. This is actually a series of clinics in Germany.
This was published in Sexually Transmitted Infections in 2006. This is their Chlamydia
positivity rate. And then what they did at this point in time, they introduced nucleic
acid testing and their Chlamydia positivity rate almost doubled. This does not represent
a change in *** behavior. This does not represent a change in clinical -- in any kind
of clinical protocol. This is solely a result of testing. What this is saying, actually
during this period of time they were underestimating the amount of tests.
Male Speaker: What is PCR again?
Jonathan Zenilman: PCR is polymerase chain reaction. It's a nucleic
-- it's one of the nucleic acid amplification technologies. So we're looking at DNA amplification
here.
Second, what about detecting undetectable or hard-to-detect organisms? I mention the
causative agent of Whipple's disease. There's PCRs, helicobacter in GI ulcers, although
actually those were identified by culture and histology using -- are highly fastidious.
Bartonella and other fastidious bacteria are identified often through PCR -- through nucleic
acid amplification. TB is increasingly being diagnosed by nucleic acid amplification, and
I'll show you a very exciting technology for that. HPV, the famous HPV that causes cervical
cancer, and that's in the vaccine, this is a non-cultivable virus. So, here we have a
vaccine which was developed on a non-cultivable virus and is identified, if you're doing diagnostic
assays, for example, probe test for a -- you know, for -- in women, this is completely
DNA-based.
In the STD area, Treponema pallidum and lymphogranuloma venereum are completely DNA-based now because
this is -- at least when you're looking for the direct organism, because this is non-cultivable,
even though people have been trying for 100 years. This is a great example of one of the
most interesting things that I've seen in this area. This is a study in France in which
they look at 20 patients with brain abscess. Traditional cultures found 22 strains of organisms.
However, when they PCR'd the samples, they did nucleic acid amplification and found the
bacterial DNA and matched it up against the libraries, they got 72 different strains in
these 20 patients including 27 species that were not previously seen in brain abscess.
About five were not even -- were not even known. These were newly discovered DNA, bacterial
DNA sequences. One subject had 16 strains.
So, to give you a sense of, from -- from hard to reach areas in small samples, this could
be very helpful in diagnostics. Now, one of the problems implementing this type of stuff
clinically besides the fact that it's technically -- at least at this level it's still technically
very challenging, specificity false positives are a major problem. I'll give you a very
good example. About 10 years ago when I was still running my lab for STD pathogens at
the main Hopkins campus, one of the techs had actually sonicated gonococcal organisms
on one side of the lab, and our PCRs for the next three months were positive, which means
that it kind of makes you wonder, what was I breathing during that period of time, either?
[laughter]
All right. What about antimicrobial resistance? Genomics can rapidly detect antibiotic resistance
when you know what you're looking for. So if you know specifically what you're looking
for, this can be a great tool. It can be used as rapid screens and this is increasingly
being thought of for screening in areas where we know we have resistance problems such as
multiple-drug resistant Gram negatives; in the hospital setting, screening for MRSA and
so forth in the appropriate setting. I would posit to you that the exact way that this
is used -- these are going to be used in practice -- has not been worked out yet, but I think
that we are going to be seeing a lot of interest in this area because of the rapidity in desktop
settings of doing this. And it has a very high utility in tracking outbreaks.
So, for example, if we have a multi-drug resistant pseudomonas in the hospital, if we get the
genetic profile and we have half a dozen cases in an ICU, if the genetic profile of all those
cases match, then that means we have an outbreak. On the other hand, if they're divergent that
means they're independent cases and we have to get people -- you know, the hand-washing
intervention is different for both in terms of both how we approach that.
Now, this is an interesting story and it's actually again from our own experience dealing
with gonococcal disease. Michael Don [spelled phonetically] here is a collaborator of mine
in Tel Aviv who had a collection of about a hundred gonococcal organisms which were
obtained from commercial sex workers who were originally from the former Soviet Union. There
had been a large problem there around the central bus station in the early 2000s, late
1900s -- late 1990s, early 2000s -- of trafficked women who were commercial sex workers with
high levels of antibiotic resistance. We were interested in looking at the organisms.
The problem was, he was going to send them to us and then -- then 9/11 happened. And
the antibiotic -- the major infectious disease meetings were postponed until December. And
he told me he was able -- and he was going to come to Chicago and he was going to bring
the plates with him. And I told him, you know, "Israeli security is pretty tight." He said,
"No, I got that one covered. I can get it. I can bring -- I can get it onto the plane."
And I told him, "If you come from the Middle East in December of 2001 with 50 or 60 bacterial
plates in your bag --"
[laughter]
"-- you're going to be -- you're going to be remanded -- we're not going to see you."
[laughter]
So, we came up with an alternate plan which he actually took the organisms and actually
absorbed them onto filter paper and shipped them to us by DHL. And we had no problem except
for one problem at Kennedy Airport -- we got a call from the agriculture inspector asking
us -- we needed an affidavit that it was not a pathogen of domestic animals.
[laughter]
Otherwise, they had no problem.
[laughter]
I couldn't make this up.
[laughter]
But what we have -- here we have -- this is -- these are the bugs that we see. And so
what we did was he had actually run the MICs, the minimum inhibitory concentrations, and
what we did was we had probes for quinolone resistance, which I actually had some very
well characterized probes of the gyrase and part C which are sequences -- and here you
see, these are the amino acid changes which occur which cause quinolone resistance. So
this is called the QRDR -- quinolone resistance determining regions in the organism. And what
we were actually able to do was take the organisms which he had -- he had this part of the -- he
had the left part, and we determined the right part and match and you can see that there
are specific sequences which are associated with different levels in the MICs. And in
this way we were able to characterize these organisms.
We also recently did this similarly with the study of a former graduate student of mine
who had studied -- who has a large gonococcal collection in Kisumo, Kenya and we did the
same work for her. So here you have situation where you have well characterized mutations,
you can develop probes to specifically look for them.
This is an example of syphilis, where two pallidum cannot be cultured but genomics has
facilitated the understanding of the epidemiology of resistance. So, syphilis has been around
for obviously a long time. If you want to culture it, you have to culture it in rabbit
testicles -- a live animal model. The rabbits are not happy nor are the people who take
care of them.
[laughter]
But, what's happened is, is that there is -- sorry -- there is -- there's been a lot
of interest in traditional therapy of syphilis with benzathine penicillin. There's been a
lot of interest in using azithromycin and macrolides because especially, it's much easier
to do, especially in settings where you're not dealing with the formal health care system.
The problem is, is that there's a 23S RNA gene mutation which actually causes resistance
to azithromycin. It's been associated with the resistance to azithromycin. So studies
-- so what Sheila Lukehart, who published his paper in New England Journal in 2004,
is they did prevalence studies.
So how common is this? And this is -- you can see there were changes in San Francisco
where there was a lot of syphilis at that time. In Dublin, in Ireland, for some reason,
there was also a lot of resistance, and they tend to segregate more in gay communities.
So here we have a situation where syphilis in gay men was much more resistant than syphilis
in straight heterosexuals. What's interesting is that since that time this has changed almost
completely resistant in San Francisco, almost completely resistant in most of the areas
where you have large groups of gay men. So, azithromycin is not a good therapeutic option.
This is important because NIH funded a large study three or four years ago in sub-Saharan
Africa which looked at azithromycin for syphilis control and found that it's very effective.
However, what clearly has to happen is that the organism needs to be monitored from a
surveillance perspective, and here you can see these are data from all over the world
where if you have access to the biological material -- and the biological material is
just a swab from a lesion. This is -- so all you've got to do is take a swab from a lesion,
dry it out and ship it, and you can determine the prevalence and this can be very useful
as a surveillance tool.
This is a really exciting area, and many of you may know about this. This is where the
same technology, same applications apply to the diagnosis of tuberculosis and the diagnosis
of resistance in INH and rifapentine resistance specifically to determine therapy at initiation
-- at initial diagnosis. The sensitivity and specificity for both detection of TB and detection
of resistance is 98 percent, and this was gone in two hours from a sputum sample. So
patients coughing in front of you -- or hopefully not in front of you -- to the side of you
-- you take a sputum sample. It's inserted into this machine which is called the Gene
Expert. So this is the first -- basically it’s entered into a cassette inside the
box. Everything is taken care of and you get a read-out of whether this is TB or not and
whether it's resistant to INH rifapentine, and this is critically important in places
like South Africa where there are major problems with TB resistance.
Domestically, for example, we do see some TB in Baltimore. You see some here a lot,
but you know, we're not that concerned about resistance in our population. But in South
Africa and other parts where there is actually major -- in Southeast Asia where there's major
problems with resistance, this is important because if you start somebody on the incorrect
therapy, you're going to not know about it for six weeks for TB and another six weeks
for the susceptibility tests, up to 12 weeks to get a full panel. All of this has been
compressed into two weeks.
This is my favorite bug. It’s gonorrhea again. This is actually -- there's been an
increased problem in susceptibility and now we're seeing -- you may have seen this in
the press -- major problems in Cefriaxone and Cefixime resistance. This is actually
the current issue of antimicrobial agents chemotherapy where there’s a high-level
treatment failure, Ceftriaxone-resistant GC, which was diagnosed in Japan. This is actually
surveillance from a phenotypic standpoint in terms of MICs over the past 10 years and
you see that the numbers have actually been going up. When I was at CDC, this number was
actually one log lower so there's been a gradual creep over the time.
But this is a hard slide to read, but the bad bug is this one -- is the one at the bottom
-- this one, 5042 -- and this is changes in sequences from the wild type and what you
see here in the middle are intermediates. So, for example, this is a -- either the sequences
of the bugs which would have an MIC of one and it changes in a -- in what's called the
panelocus [spelled phonetically]. And then you have an MIC of two, an MIC of four. The
one that's eight is the treatment failure one. And you can see that this guy is accumulating
different mutations and eventually these are cumulative. So this is to show you how this
is actually done. In a situation where in quinolones you have a single step where there's
one change which causes a resistance which causes a substantial increase. Here you have
accumulation of different mutations and if you have your data lined up right, you can
actually see that.
I was in France in 2010, and, you know, and I don't speak French. I speak Spanish. I don't
speak French but I can understand this. And this was actually when they had the MTM [spelled
phonetically] metallo-beta-lactamase organism which was the pan-resistant bug and in fact
there was one of those which was seen in the fall of 2000 in Pennin [spelled phonetically]
Howard County. The reason I show this is that the French just have a much nicer way of saying
bad bugs than we do. It just -- there's an elegance -- there's an elegance about it which
we don't have.
[laughter]
Okay. How does understanding genomics facilitate epidemiology? Well, let me show you a couple
of examples, which again are from the STD area but which I think really demonstrate
this nicely. The old model was that 90 percent of persons with genital *** shed virus
asymptomatically 1 to10 percent of the time. And if you take studies where people were
cultured -- this is from 2006 from a study by Anna Wald -- sorry, 1995-‘96, these are
serial cultures. She had people culture themselves every day for about a -- for months at a time,
and the [unintelligible] cultures here, perianal [unintelligible] here, perianal -- these are
-- this is -- and the only time that the pink is when she was symptomatic.
So you see here, it's kind of episodic and it's about 1 percent of the time, usually
associated with symptoms. Now, what happens when you do genomics and see what's going
on in terms of how often they're shedding DNA and what you see is this is the percentage
of days for detection of HSV on genital skin mucosa. So here you see basically about 20
percent of people have – 25 percent of people have -- are shedding. Here you can see so
basically what the -- the point being here is that people are shedding virus with ***
about a third to half of the time. It may not be detectable by culture but it's sub-clinical
so therefore this changes the way you think about the disease process. The disease process
is not something which turns on and turns off, but something which is actually on all
the time, but just modulating itself.
Swine flu. Many of you remember that from a couple of years ago. This is an interesting
one because what it actually represented was a recombinant event between a human virus,
a pig virus, and a bird virus. So the typical flu is -- usually it's the avian -- I mean
here we said – here -- and basically what happened here is that different people have
different organisms but this represented a recombinant event between three different
zoonoses of influenza which then combined and was diagnosed by genomic sequencing.
***. Had genomics guide epidemiological investigation, understanding transmission, interventions,
and therapy. Well this is actually from my colleague Tom Quinn, who is also an investigator
across the street, the NIH. The important thing is, this is the transmission rate in
study -- in a large study -- and this has been replicated a number of times between
dichotomous couples. So, *** transmission between couples. One partner has the virus,
one partner does not: and you see as the viral load goes up, the transmission rate goes up.
So, essentially below 400 there's almost no transmission. Between 400 and 4,000, there's
transmission approximately 5 percent per year and once you get up to over 50,000, it's 25
percent a year. This has profound implications. Furthermore, what happens when you put circumcision
into the mix. If you have circumcised people you see the transmission rate goes down to
zero as well. So, circumcision and viral load are synergistic in *** transmission.
Next point. What about acute *** viral sero-conversion syndrome? ***, as you know -- acute *** sero-conversion
syndrome, within a week or two after you get the virus, you have your viral load may be
as high as a million and the transmission rate at that time may be as high as 20 percent
before it settles down. So, here we have viral load is describing to us the natural history
of the disease as well as also giving us some insights into transmission and also how to
intervene. So the intervention's based on the knowledge afforded us by the genomic testing
in ***, help us -- result in interventions to detect acute *** cases, circumcision, and
transmission to prevent infection.
I think the circumcision trials have been well described. This is one of the most exciting
things that happened over the past year and Mike Cohen led this group from the University
of North Carolina. And what happened in group one was early *** intervention, so therefore,
again, we have dichotomous couples, people where one partner had the infection, the other
partner did not. The partners were -- the infected partners were treated, even if they
did not meet criteria yet for initiation of treatment, and the control group received
standard of care. This is *** transmission in the intervention group. *** transmission
in the control group. I think basically this tells the story very clearly: *** treatment
reduces viral load below 400, reduces transmission, and it's probably -- you're going to see a
lot of focus on prevention from that standpoint.
Similarly with hepatitis C, but with other -- with other, you know -- with other slants,
hepatitis C is an infection which again is another virus which has not -- which cannot
be cultured in vitro where most of the work has been done genomically. Genomics guide
detection -- for example, these are used viral load measures are used to determine whether
somebody needs therapy or not. They define therapy and therapy outcomes. The viral load
is a therapeutic outcome. Resistance is genomically defined, similar to ***. And ***, for example,
you -- when we have initial patient showing for treatment, an *** viral load is obtained
and also the genomic profile of the virus is obtained which will tell us which resistance
mutations they have so we -- very similar to what you saw on the TB situation, so we
don't treat them with drugs to which they are resistant.
Therapy strategies are based on genomic testing. Here we have an interesting thing where in
hepatitis C, the host’s susceptibility can be genomically defined based on genes in the
individual patient. So, just for an example, for example, codon, changes in the codon in
the HCV protease will render, and this is the wild type, the protease will render the
individual resistant to a series of protease inhibitor drugs and therefore understanding
this can help guide therapy. This is actually slides provided me by Dave Thomas, who's the
head of infectious disease at Hopkins and hepatitis C expert. Here we have the drug
classes for hepatitis C and again this is not to be expert, just to give you an example,
and we have the mutations, and what you can do is based on you can actually define how
you can get the genomic outcome of an individual virus and the person, develop his grid and
see what drugs they're going to be sensitive to and define therapy based on actual knowledge
of the resistance profile of the drugs of the virus without ever having to grow up the
virus. This is all done based on genetic sequencing.
And this does not take long to do, either. This is also could be done in a single run.
This is -- for example, you have a good example of somebody who's -- somebody with a resistant
virus -- the virus was diagnosed and here we have instituting the peginterferon ribavirin
regimen, which it was sensitive to. And this is viral load here and what happens after
initiation of appropriate therapy. Now this is Ashwin Balagopal’s work from our group,
again at the main Hopkins campus, showing that there are these specific SNPs which are
sequences in the human genome, which defines susceptibility for interferon -- for interferon
-- and these determine whether persons are going to be susceptible to therapy if they're
infected with Type 1 HCV virus, and this is again taking that to a larger scale from a
surveillance standpoint and looking for the prevalence of this mutation worldwide. You
can see that they are predominantly seen in Africa. Very little in Asia. And this is clinically
important because this explains why Type 1 HCV does not respond to anti-HCV therapy without
the new protease inhibitors. And Type 1 is most commonly seen in African Americans.
So, with that very rapid overview, I'll spend the next 10 minutes kind of going over some
microbiome projects and some of the things that we've done and leave it open for questions.
I think I mentioned that the human microbiome is a direction where the field is actually
-- it's not going -- it's here -- but this does not look at single organisms. This looks
at organism communities. So think of bacteria as bacterial communities. This is especially
relevant on mucosal surfaces. Without understanding the interactions between our human microbial
genomes, it is impossible to obtain a complete picture of our biology.
And the important part here is that each part of our body has a different suite of organisms
which are present, live there. We have the gut. We have the oral mucosal. We have the
-- there's the ***. There's the skin. And even different parts of the skin have different
organisms. Now we were traditionally taught, for example, that skin organisms are staph,
strep, propriana bacterium and perhaps carini [spelled phonetically]. But it turns out to
be much more complicated than that. C-diff is a problem in the gut because it is a -- because
the normal flora have been eradicated by antibiotic therapy. Furthermore, there's some non-infectious
diseases which are associated with microbiome changes. For example, inflammatory bowel disease.
Some people are thinking some other autoimmune diseases are associated with microbiome abnormalities
as well.
So, the critical questions are: how do we acquire or maintain our microbic communities;
how does it respond to stress; can we use this information to intentionally -- on a
-- from a therapeutic standpoint; and how do genotype environmental exposures and physiological
status affect microbiome composition? So, the reason I show this, the pyrosequencing,
is because the pyrosequencing is basically DNA -- DNA analysis on steroids. Here we're
not looking for a specific organism. And there's actually other, other methods now as well,
including something called the Illumina, the Titanium, and there's recently something called
Ion Torrent.
They're all variants on the same theme. Essentially what these technologies allow you to do is
sequence the entire DNA from a -- from a specimen. Not just looking for one organism but do the
complete bacterial sequencing on a hosted DNA. And that's going to result in a large
number of different organisms. So, for example, if I take a swab of my mouth and put it into
a pyrosequencing machine, I'm going to get 50, 60 different bacterial sequences and it'll
tell me how many millions of each one's were there. So, this is actually an enormous proposition
and, you know, it actually -- one of the things -- that's where the cardinal rule in this
is that sequencing gets 20 percent cheaper every year. So, as this happens, what cost
$50,000 five years ago is -- now costs about $5,000.
I'm going to show you two examples where we, you know -- from -- with -- where we've had
clinical interest in this area and then sum up. So, bacterial vaginosis is the most common
cause of inflammatory vaginal disorders and it's actual an ecological disturbance of the
vaginal flora. It's not an STD in the traditional standpoint and diagnosis is based on clinical
criteria of Gram stain.
This is the good stuff. This is a swab from a healthy ***. Nice lactobacilli, healthy-looking
epithelial cells. This is a clue cell, which is one of the diagnostic criteria. This is
a clue cell from a woman with bacterial vaginosis and there's a ground glass appearance and
what you have here are anaerobes which are attaching to the outer surface. And this is
a Gram stain of the same -- of an individual with the same condition. Now this one looks
a lot different than this. So I would say this looks sicker than the first one.
And what happens is, is that you have lactobacillus, which is the normal flora in the *** being
replaced by Gardnerella, anaerobes and mycoplasmas. and this is what we knew phenotypically, you
know, up to about 10 years ago. Rebecca Brotman, who's a former graduate student of mine and
who is now a professor at the University of Maryland Institute for Genomic Sciences, has
been studying the microbiome of the ***, and what you have here -- the way to look
at this is the pH over here in the Newton square, both of which are indirect measures
of BV.
Here we have, that you so, as you go, this is going from healthy to BV, healthy to BV.
And this is the microbiome of individuals in each category. When you see here we have
here there's mostly is lactobacillus. Here we have lots of organisms which -- many of
which were non-cultivable prior to using this technology, lots of things like ureaplasma,
mycoplasma, Finegoldia, which is an anaerobe. Most of these are anaerobes. There's -- Atopium
is an anaerobe which was first discovered studying women with bacterial vaginosis. Prevotella
is an anaerobe which is also found in the mouth. So what we see here is a shift from
a normal flora towards anaerobes and a lot of anaerobes with a lot of anaerobic diversity
and a lot of anaerobes which were formerly non-cultivable.
I'm going to show you the same story with chronic wounds. We've had a major interest
in our group in studying chronic wounds, chronic lower extremity wounds. These are the bane
of an internist's existence. One pundit said the safest place to hide a $20 bill on a medical
ward is under a large dressing.
[laughter]
Wounds have direct medical cost impact. They're $25 billion a year. My brother will take the
dressing off.
[laughter]
Wounds have substantial indirect cost benefit. This is -- there's little research being done
in this area and this is a great opportunity because nobody knows what's really going on
here. From the infection standpoint, there's no good definition of infection in chronic
wounds. How do you define colonization versus infection? People give antibiotics because
bugs are there but they have no idea really what's going on there. People have proposed
quantitative culture as one modality but this is -- actually, when you look at the data
to support this, it's really not good and we don't do this anymore.
And this is -- I say welcome to my world. This is typical person in my clinic, chronic
venous stasis ulcer, non-healing, and uses compression or slowly healing. And does this
wound require treatment or not? And what is the bacterial community of that wound? And
how is that different from what we see elsewhere? I think we saw that. So our research program
objectives were to describe the prevalence of bacterial species in chronic wounds; to
assess the microbial burden by different modalities, and I'll show you the bacterial culture and
the DNA; and compare the microbial populations at two different sites, which I want to show
you. But we demonstrated that the wounds are pretty homogeneous across the whole wound.
So, if you do a traditional culture you find the usual players: MRSA, Staph aureus, Pseudomonas,
which actually turns out to be present only in people who have been treated with antibiotics,
Group strep B and a whole host of other things. MRSA is present in 45 percent, Pseudomonas
in 28 percent, and Group B strep in 28 percent. I would argue part of this is because we're
a reference center -- we're going to get people who've already failed therapy elsewhere. So
97 percent of our wounds have at least one oreism [spelled phonetically], Pseudomonas
as I mentioned here -- very high bacterial load in those with MRSA and so forth.
Now when we do the quantitative microbiology -- I'm going to show you two different methods.
I'm going to show you that there's a theme here. This is an older way of looking at the
quantitative -- sorry, at the, at the microbial DNA. Here we took the microbial DNA of the
wound and typed it out and speciated it and see what was there. And we find that these
are skin bugs -- valcalgonese [spelled phonetically] and bacterias or anaerobes. Here we have a
Pseudomonas. This is a Gram negative. This is Group B strep. This is mobiluncus. Now
what's interesting if you -- you probably don't remember, but mobiluncus and Atopobium
are the BV bugs -- the bacterial vaginosis bugs. So we're seeing -- so we have again
chronic mucosal infection characterized by large numbers of anaerobes including some
which are non-cultivable.
This is what's called a heat map in which each vertical column is one of our specimens
and this is the horizontal, which are the different organisms that are seen. Again you
see there's large diversity of organisms. Again, if we went through -- a lot of these
are Pseudomonas because we have a group which was treated with antibiotics previously. But
the important role is that there's a large diversity of anaerobic bugs so we're not just
seeing MRSA in there. We're not just seeing strep. We're seeing a whole host of other
things. The question that we're interested in now is that if you -- how does this impact
on wound healing and can these patterns -- can we develop patterns -- are there ecological
patterns which will predict whether a wound will heal or not? And almost -- and are there
ways of correcting the flora that they need to be?
So the conclusion from metagenomics from the wounds are that microbial diversity was significantly
lower in patients treated with antimicrobials. There's a high proportion of anaerobes in
non-cultivables. Genomics data suggests that anaerobes are critically important and this
may represent synergistic infections. However, and here we get into the next step in genomics
studies, this is DNA only. So dead bugs are going to be there. RNA indicates what's active.
So, for example, in the STD area, we don't recommend people do traditional tests of cures
because you'll shed bacterial DNA from dead bugs for two weeks. So, therefore, interpretation
of these things has to recognize that DNA is also present in dead bugs and may be present
for a while.
So, in conclusion, genomics has impacted our ability to discover new pathogens, our ability
to detect pathogens, understanding the epidemiology of these pathogens, guiding therapy interventions,
understanding resistance, and understanding susceptibility with the example of hepatitis
C. They're rapidly replacing traditional microbiology and I call it the cell phone paradigm in the
appropriate settings because, for example, they can be -- we're seeing benchtop in expert
systems that are being downsized to cell phones. For example, the military has bio-threat agent
detection systems which are the size of a cell phone which they carry with them in the
field. But also, if you're in a developing country and you're looking for specific pathogens
-- for example, if you're fielding TB diagnostics, you get the capital funding to basically fund
the box which actually does the whole thing, you avoid having to set up the whole culture
systems. This is very similar to what's seen in developing countries where they -- it's
much cheaper now to install a cell phone system than to install a landline system because
you don't need all that infrastructure.
So, I think that's it. So the microbiome as an ecological concept that is leading to new
understanding of infectious diseases based on microbial community concepts. So, thank
you.
[applause]
Yeah.
Male Speaker: So there are few groups like Julio Montaner
in British Columbia and so forth, shouting from the rooftops that reduction in viral
load actually leads to substantial health benefits. But what we have is a --
Jonathan Zenilman: This is for ***?
Male Speaker: For ***, yes.
Jonathan Zenilman: [affirmative]
Male Speaker: What we have is a policy problem. We have
a policy problem that governmental systems are not generally willing to invest in prevention
as treatment, and so -- and that's actually applicable to a lot of potential health benefits
in genomic medicine. So, do you have ideas in terms of the policy sphere, how we consider
-- how we encourage policy?
Jonathan Zenilman: Okay. I think -- it's a good question and
I think until -- it's an important question -- until the question -- until the data came
out, there was a very valid argument to be made and the question is reducing viral load
as prevention, should that, you know, has obviously benefits should be -- cannot be
generalized and should we be moving in a policy direction? If I'm hearing you right.
Until the data was published last year, there was a valid question to be -- a valid question
is that, first of all, does treating people who are not meeting thresholds for treatment
yet, result in unintended consequences in terms of anti-retroviral resistance down the
line, drug side effects, and so forth? I think with the data now this is very compelling.
And I think any program, from my standpoint, this is a no-brainer. You know, for example,
but we do see your issue, there's this policy disconnect. Because, for example -- and we've
seen this over the past year -- that in states where they run short of funds, that the AIDS
drug assistance programs have long waiting lists and things like that and I think what
we now have is a business case to make that is a very strong business case that this results
in further -- in reduction of further infection.
Now this all may change because part of the problem is that health care has been funded
-- the prevention side of health care is funded from a different pot than the treatment side.
With the integration of health care that's going to happen over the next five years -- this
is going to partially resolve itself and I think people may be more responsive. But I
think these data form -- especially convincing. Now, from a policy standpoint, if policymakers
don't want to make that decision after being presented with the data, then the only solution
is to replace the policymaker. I don't have any --
[laughter]
You know, I think despite comments, you know, I'm being taped so I can't let my political
feelings be known, but --
[laughter]
Female Speaker: What is the current understanding as to why
some bugs remain non-cultivable?
Jonathan Zenilman: Oh, okay. Well, actually -- the question is
what's the current understanding of why some bugs remain non-cultivable. I think -- I would
not say remain -- I think a lot of these were undiscovered until the DNA technology became
available, and they have growth requirements which are extremely fastidious. So, for example,
they just, you know, and either they're biochemical leads, nutritional leads or they're oxygen,
you know. For example, oxygen may be toxic to them. There have been a -- tremendous amount
of work on this in environmental microbiology because if you go into the environmental side
you see, you know, you take, you know, river water and grow what's in there -- there's
an enormous number of things that are in there. The microbiome is enormous. So I think this
-- it's more of a technical problem because we're using -- actually, one of my colleagues
says we're using technology now that was developed by Pasteur 130 years ago. Yeah.
Male Speaker: When [inaudible] methods identify organisms
that can't be identified by other tests like culture or toxin [inaudible]. Is there good
evidence to show that those actually represent infections that need to be treated?
Jonathan Zenilman: Ah. Okay. Well, you know, the question -- you're
actually asking the million dollar question. And basically, what constitutes colonization,
what constitutes disease? So I think what happens is that -- just find -- and this is
my whole issue with wounds in a lot of wound care -- just finding the organism there doesn't
mean it's doing anything. And it's just, you know, I think if you look at the papers where
they've identified organisms, for example, Whipple's, which is a beautiful paper, what
they do is they take a series of people with disease and then basically, because you can't
fulfill Koch's postulates in this, but what they do is -- and they demonstrate that there
is a consistent, and that the organism is found consistently. And then there's a lot
-- in each case with the disease that it's -- in terms of where it is in the organism
it's there, when you treat them it goes away so I think you have to establish a case that
this is actually causing invasive infection. The brain abscess is a very, you know, is
a very easy one because this is a brain abscess. It's a sterile space. We're not supposed to
have stuff in our brain and you're finding 16 bugs there in somebody with an abscess.
That's easy. That's a slam dunk. Yeah.
Female Speaker: In febrile neutropenics we often have fever,
low white count, and bacterial blood cultures are negative. Has anybody looked with these
DNA technologies to see if there's --.
Jonathan Zenilman: For -- yeah. Blood is actually -- we've actually,
we actually looked at blood a while. We were looking at blood not in febrile neutropenics
but in burn patients. Blood is very difficult. So if you look at the blood literature you'll
see that there's a lot of work done on DNA technology once the blood cultures turn positive.
So, once if it's in the machine and it turns positive, there's a lot of work being done
to identify what's growing in there with DNA technology. At admission a time -- I'm interested
in what happens at the time of the initial blood culture. The problem is, is what is
the concentration of bacteria in blood cultures? There's two problems. The concentration of
bacteria in blood culture ranges between one to 10 CFUs per mil. So, it's, know you, there
are people with high level Staph bacterium -- you may have a thousand, you know, but
it's not like we what we see in viruses. So, you have to have high volumes. And second
is the background of human DNA is enormous.
So you have to sort that out. So I think there's -- there are people looking at this. I think
there's a lot of technical problems. I think there are -- there are some very promising
technologies including some microbium fluorescent amplification technologies which some folks
at the University of Maryland, Baltimore County are even doing and some other types of things.
But it's not there yet. I mean, ideally you would love to have somebody in the -- somebody
would just say -- draw a tube of blood, send it to lab and know whether you're bactorimic,
you know, it's not there yet. Yeah.
Male Speaker: Wouldn't some suggestion that there's a different
val flora [spelled phonetically] in obese patients?
Jonathan Zenilman: Yeah. I think the question was a suggestion
that there's different val flora in obese patients. There are -- I'm not an expert in
this because my area is infection but I think there are absolutely differences in obese
patients. There's differences in animal models for that. And I think the question is and
how that regulates food intake and I think there's been a lot of interest in can you
change the flora -- you know, what is it the chicken or the egg problem? Besides that,
I can't answer that because that's not my area of expertise. Yeah.
Male Speaker: When you clean the skin with alcohol, what
are you doing actually before [unintelligible], what bacteria are you killing?
Jonathan Zenilman: You're reducing the bacterial load -- well,
it's skin flora that you're worried about. Skin flora in healthy people is Staph and
strep and sick people may be also Gram negatives that are acquired from the hospital setting.
You're basically -- both mechanically removing and killing at least 10 to the 3 of the bacteria
-- of the load so you're reducing the risk of infection. It's both mechanical and biochemical.
Yeah.
Male Speaker: After your lecture, how can you live with
dialysis catheters now?
Jonathan Zenilman: [laughs] The question was, after my lecture,
how can you live with dialysis catheters? Well, you know, again we run into the question
what constitutes infection and what constitutes colonization, and a topic which I did not
discuss is the whole issue of biofilm. Because, actually, organisms on surfaces do not live
in a planktonic state like they do in the lab. They're actually in film situations where
they may be relatively inactive. We actually surveyed inpatients at Bayview -- had a medical
student survey about six months ago, what proportion of patients -- it's not only dialysis
catheters, what proportion of patients who are on a general medicine surface have stuff
in them that they weren't born with. And it's including -- including catheters, ICDs pacemakers,
prosthetic hips, things like -- it's about 60 percent, and that does not include bullets.
[laughter]
So, which we have, so I think -- so this is a big -- actually, what scares the living
daylights out of me is prosthetic joints. Actually prosthetic joint infections is one
of the areas also where they've been using PCR. There's been a lot of research looking
at PCRs to identify pathogens. But this is a big -- you know, this is our world. And
I don't have a good answer but it's an enormous problem, yeah.
Male Speaker: I'm a genobiologist and thank you for a great
lecture.
Jonathan Zenilman: I'm the clinician so I'm just a translating
guy.
[laughter]
Male Speaker: But my question is clinical, so, we've definitely
made a significant amount of progress in the detection. But if you think of it as an arms
Jonathan Zenilman: Okay. Good question. The question is how do
I see this going through the treatment standpoint. I see a couple of directions. One is pathogen-directed
therapy. So probably the most important -- the TB example is a very good one. We know exactly
what the bug is, the characteristics. So right now the approach to infectious disease in
most hospital settings is, you know, is broad spectrum, which actually is the worst thing
we can do for our flora. So pathogen-directed. Second, focusing on drilling down on that
and pathogen-directed with antimicrobic susceptibility directed. Third is, from an epidemiological
standpoint, if you're worried about specific outbreak situations, being able to identify
specific outbreak bugs and deal with them in a rapid way.
The other question which you asked which is interesting is dealing with flora and, for
example, there's been a lot of interest for years on correcting the bacterial flora in
bacterial vaginosis with lactobacilli, with yogurt which actually has lactobacilli, with
lactobacillus crispatus which is actually the specific one, and all the studies have
fallen flat on their face. They haven't worked. On the other hand, probably the most successful
one of reflora-ing a mucosal surface has been in the case of C-diff, where, as my boss,
John Bartlett, says, the therapy has -- is incredibly effective but has aesthetic complications,
and that is, in people with severe recurrent C. diff, and there are these people, you probably
know some of them who cannot get off of vancomycin or, you know, other therapies. They recur
all the time. The therapy which works is a stool transplant and it fixes it within 24
hours. And there are people who are doing this now, and you can imagine -- there is
obvious regulatory problems with -- how do you do this with the FDA and safe and things
like that, but it works. And, you know, it has, you know, it's actually -- the way it's
done is actually, stool's taken from a relative and it's either instilled from below by somebody
-- there's actually -- there are job opportunities for the people who used to do barium enemas.
[laughter]
Because you want to get it all the way up or through a long canter tube. And there's
another -- I heard that there's a formulation being developed that were actually going to
be in a gel cap but it's sort of like it's going to absorb -- you know, what you take
but the capsule does not dissolve until it's down into the gut. So, that's very -- because
you know I think but clearly you're raising -- for example we'd see this in urinary tract,
you know, we're interested in the wound, you know, can we repopulate the flora with what
we want -- there's a lot of interest in that. Last question.
Male Speaker: When you have a [unintelligible] of diarrhea
and a lot of them say well, it's a viral type of technology, but then you've got identified
Jonathan Zenilman: Okay. It's hard -- on an individual basis
you can't differentiate clinically after you've done the settings. So for example, C. diff
will be clinically obvious, for example, anybody in a long-term care facility or who's been
on antibiotics recently -- you know, has new onset diarrhea -- generally if they have leukocytosis,
a white cell count, generally C. diff until proven otherwise. So what we would do is we
would get the test for C. diff and then treat empirically. But in terms of viral outbreaks,
one of the problems with viral outbreaks is that they don't -- the viruses don't grow.
But, so from a public -- and they're usually self-limited, so the utility here of the diagnostic
testing is understand the epidemiology as opposed to guide therapy. Okay? Thank you.
[applause]