Animal & Veterinary
Antimicrobial Resistance Among Non-Typhi Salmonella by Maria Karlsson, Ph.D.
DR. KARLSSON: Thank you Heather. So I will be presenting on antimicrobial resistance among non-typhi Salmonella enterica, mainly isolates collected from humans.
As we all know antimicrobial resistance is a clinically relevant problem that contributes to the burden of infectious disease by limiting treatment options, increasing health care costs, and increasing the duration and severity of illness.
NARMS-CDC has been monitoring antimicrobial resistance in non-typhi Salmonella since 1996. The NARMS-CDC Applied Research Unit mainly focuses on molecular characterization of isolates resistant to extended spectrum cephalosporins and fluoroquinolones.
This slide here shows you the prevalence of ceftiofur resistance in NARMS non-typhi Salmonella collected from humans from 1996 to 2007. As you can see, we have observed an increase over the years. In 1996, 0.2 percent were ceftiofur resistant, whereas 2.9 percent were resistant in 2005, and 3.3 percent in 2007.
This slide shows you the ceftiofur resistance within the serotypes Newport, Typhimurium, and Heidelberg which have been the main players when it comes to ceftiofur resistance.
From 1996 to 1998, ceftiofur resistance was mainly seen in isolates of Typhimurium and Typhimuriums carrying the AmpC type beta-lactamase CMY. However, in 1998 and 1999, we started observing a sharp increase in Newport with this CMY gene. These Newport strains were also resistant to other drugs including chloramphenicol, tetracycline, streptomycin, and sulfamethoxazole. These are what we call the multidrug resistant MC Newport, the MDR Amp phenotype.
These phenotype in Newport rose from constituting 1 percent of our submissions in 1999 to over 25 percent in 2001. Since then the prevalence of MDR MC Newport has somewhat declined but are still present. We have lately also been seeing an increase in Typhimurium and Heidelberg being ceftiofur resistant.
In order to characterize cephalosporin-resistant isolates better, we have performed a couple of beta-lactamase surveys. Now I am going to share some data with you on the 2005 and 2006 NARMS beta-lactamase survey. The objectives were to determine the prevalence of non-typhi Salmonella displaying decreased susceptibility defined here as MIC greater than or equal to 2 mg/L to either ceftriaxone or ceftiofur.
We also wanted to characterize resistance mechanisms by PCR screening for six different beta-lactamase genes including blaTEM, blaOXA, blaSHV, blaCTX-M, blaPSE, and blaCMY.
In 2005 and 2006, a bit over 4,000 isolates of non-typhi Salmonella were submitted to NARMS, CDC. Of these, 175 displayed decreased susceptibility to ceftriaxone or ceftiofur. Among these 175, 30 different serotypes were represented.
You can see the top seven serotypes listed in this table here. Newport, Typhimurium, and Heidelberg were the top three.
By PCR, we could detect one or more beta-lactamase gene in 139 of the isolates and a blaCMY beta-lactamase was found in 133 of these 139 isolates. We were also able to detect 11 TEM genes, 3 PSE, 2 OXA, and one CTX-M gene. But as you can see, the majority were blaCMY-positive.
So we decided to look for at these isolates more closely and this graph here shows you the serotypes among the blaCMY-containing isolates. Newport, Typhimurium, and Agona were the top three serotypes. Again, you can see Newport constituted 37 percent of the isolates so we decided to look at these isolates in more detail.
This slide shows you PFGE patterns of nine of the CMY-containing Newports. By doing PFGE we were able to identify seven different PFGE patterns with primary enzyme Xba1.
By performing PFGE with a second enzyme Bln1, we showed that two isolates had indistinguishable patterns with both enzymes. Three of the isolates displayed Bln patterns not previously seen in the PulseNet national database. All of the Xba1 patterns had been reported to the PusleNet database previously.
This confirms what other studies have shown before, that cephalosporin-resistant Salmonella serotype Newport strains showed some genetic diversity and that there is not only a single clone responsible or a single clone circulating.
We also wanted to determine the location of the CMY gene and to do this, we extracted plasmids from nine Newport isolates and we transformed them into E. coli DH10B. This table here shows you the resistance genotypes of the original isolates as well as of the transformants.
On the very first line or row marked in orange, you can see the susceptibility patterns for the E. coli DH10B. The dark pink rows shows you the resistance phenotype of the original isolates and the light pink rows show you the susceptibility pattern or resistance phenotype of the transformants.
I know there are a lot of data on this slide, but there are two things that I would like to point out. First, all of the original isolates showed resistance to additional drugs than beta-lactams, including chloramphenicol, sulfisoxazole, streptomycin, tetracycline, and kanamycin to mention some.
The second thing is that these resistances transferred to the transformants and this confirms that CMY is located on a plasmid that also harbors other resistance determinants. So by doing this we could confirm that CMY is located on multidrug resistant plasmids and by using Inc typing, we were able to show that all of them were of the same type, called the Inc A/C type.
After completing this survey we have moved on to look at our isolates collected in 2008. In this beta-lactamase survey, we are not only looking at isolates collected from humans, but also isolates collected from animals and retail meats.
I should also say that we are also collaborating with CIPARS in this project and they are performing a similar study and we will be comparing our data with them.
This table here shows you the total number of non-typhi Salmonella collected by each agency in 2008. It also shows you the number of isolates showing decreased susceptibility to either ceftriaxone or ceftiofur.
We used the same approach in screens for five different beta-lacatamase genes. As you can see, CMY predominated and was found among isolates from all sample sources. We were also able to find TEM genes among isolates from all sample sources. However, CTX-M genes, SHV and OXA genes could only be found among isolates from humans. We are currently working on characterizing these genes and these isolates further.
I would also like to say a few words about fluoroquinolones resistance in non-typhi Salmonella. As you know, quinolone resistance is most commonly caused by mutations in the quinolone resistance determining region, the QRDR, or the topoisomerase genes gyrA, gyrB, parC, and parE.
Resistance develops in a stepwise fashion and one mutation confers resistance to the quinolone nalidixic acid, whereas two or more mutations are required for fluoroquinolones resistance.
In addition to these mutations, plasmid mediated mechanisms have also been described, including the qnr-genes, the aac(6’)-1b-cr variant, and the efflux mechanism qepA. These do not confer clinical resistance to ciprofloxacin but they do confer decreased susceptibility with MICs greater than or equal to 0.125 mg/L.
This slide shows you the MIC distribution for ciprofloxacin on non-typhi Salmonella submitted to NARMS from 1996 to 2008. As you can see, the majority of the isolates are susceptible. The distribution peaks at 0.016 and 0.032. Clinical resistance is defined as greater than or equal to 4 and we have 24 isolates with this phenotype.
As you can see, we also have a little bit over 500 isolates showing decreased susceptibility with MICs greater than or equal to 0.125 mg/L. If we look at these isolates a bit closer, the majority of them are resistant to nalidixic acid and this suggests that presence of QRDR mutations.
However, you can see that we have some isolates here that are susceptible to nalidixic acid. Among these isolates we have been able to identify plasmid-mediated mechanisms.
So far we have performed three studies on plasmid-mediated quinolone resistance in NARMS. The first study was performed by Katie Gay and colleagues and was published in CID in 2006. The second study was published last year in AAC, and a third study is underway and is currently under review.
If we look at the first study, Katie and her colleagues looked at Salmonella collected in 1996 to 2003. Among these isolates, ten were found to be carrying plasmid-mediated quinolone resistance mechanisms. All of these were qnr positive and most common was qnrB5 that was found in seven isolates.
In this table you can also see the MIC range for nalidixic acid and ciprofloxacin in these isolates.
These qnr genes were found in four different serotypes including Berta, Anatum, Bovismorbificans, and Mbandaka.
If you look at data from the second and the third study, here we included Salmonella collected in 2004 and 2007. By looking at these isolates we were able to identify 24 additional strains with plasmid-mediated mechanisms. Again, the majority were qnr positive. The most common gene found was qnrS1 that was found in 14 isolates.
In addition to the qnrs we were also able to identify two isolates with an aac(6’)-1b-cr gene.
I should also point out that 20 of these 24 were isolated in 2006 and 2007.
The serotypes that these were found in are listed here. Most common was Typhimurium followed by Corvalis, Enteritidis, Montevideo, and Saintpaul. The other serotypes found are listed below.
So to summarize, CMY beta-lactamases are the predominant cause of extended-spectrum cephalosporin resistance among non-typhi Salmonella isolated not only from humans, but also from animals and retail meats.
The blaCMY gene is commonly located on multidrug resistant plasmids.
Serotype Newport and Typhimurium continue to be important sources of cephalosporin resistance among isolates collected from humans.
Among over 22,000 isolates of Salmonella submitted to NARMS, CDC in 1996 to 2008, 97.5 percent show ciprofloxacin MICs less than or equal to 0.064 mg/L.
Among the isolates that display decreased susceptibility to ciprofloxacin, 90.6 percent are resistant to nalidixic acid.
The first NARMS isolate with a plasmid-mediated quinolone resistant mechanism was detected in 1997 and this was a qnrB5 in a serotype Berta.
Since then, 33 additional isolates with a PMQR determinant have been identified. Of these, 20 were isolated in 2006 and 2007. Thank you for your attention.
DR. HARBOTTLE: Thank you very much Dr. Karlsson. I think we are going to hold questions until the end. We are going to move right along. My co-moderator here can introduce me.
DR. FRYE: Hi. I am Jonathan Frye with the USDA and I would like to introduce Dr. Heather Harbottle. She is with the FDA and she is going to be talking today about antimicrobial resistance mechanisms identified by high-density microarrays.