mBiosphere

New method to diagnose sepsis is faster, cheaper

Schematic of the lysis centrifugation-intrinsic fluorescence (LC-IF) method.

In managing bloodstream infections, minutes count, and delays in treatment or administering the wrong antibiotic can kill a patient. In mBio today, scientists from bioMérieux, Inc. describe a new method that could cut hours off the time it takes to diagnose blood infections while also eliminating the need for complicated manual processing and expensive equipment. The lysis centrifugation-intrinsic fluorescence (LC-IF) method combines a selective lysis step in which blood cells in the sample are destroyed, a centrifugation step to collect any bacteria or fungi in the sample, and a fluorescence step that analyzes the particular fingerprint of any pathogens present in the sample. Tests show the method correctly identifies the species of bacteria or fungi in 96.5 % of positive blood culture samples, crucial information for doctors to provide the appropriate drugs for their patients.

"The primary benefit of getting a rapid identification is making sure the patient is on the right [antibiotic] therapy and to quickly make any needed adjustments to the initial therapy," says co-author John Walsh of bioMérieux in Durham, North Carolina. Patients with bloodstream infections are usually in very serious condition, says Walsh, and faster identification of the organism causing the infection can help get patients on the most effective antibiotics faster and save lives. Proper diagnosis is also important from the perspective of antibiotic stewardship: using more appropriate, targeted antibiotics reduces the risk of contributing to the spread of resistance to broad-spectrum antibiotics.

Even the "gold standard" testing is prone to errors, time-consuming

Walsh says the current standard approach to diagnosing bloodstream infections, Gram staining and overnight sub-culture followed by phenotypic ID tests, have limitations that can prevent timely treatment. Gram staining provides early, low level information about the type of microorganism present, but it sometimes takes hours to deliver a result, and technicians can make mistakes in the process that provide misleading results. Other more specific identification methods are also available for diagnosis, but they can take at least a day or two to produce results and many require expensive equipment.

New method: Intrinsic fluorescence spectroscopy

In the LC-IF technique developed by Walsh et al., a sample of positive blood culture is treated to lyse the blood cells, then transferred to a specialized optical tube. The tube is centrifuged, driving any bacteria or fungi, which are denser than the solution, down through a liquid density cushion to form a pellet at the bottom of the tube.

Then comes the intrinsic fluorescence spectroscopy (IFS): the microbial pellet is irradiated with light ranging from the deep ultraviolet to infrared, which excites certain organic molecules in the microorganisms, including tryptophan, NADH, FAD, porphyrins, and others, and causes them to fluoresce in a characteristic way depending on the identity of the microbe. The exact pattern of fluorescence is compared with a database of 37 of the most common known pathogens to identify the organism present.

"We're using intrinsic fluorescence to identify the microorganisms. It's an innate property of most living organisms. Because it’s intrinsic, no reagents are needed for the identification step," which removes many of the opportunities for mistakes and lowers test costs, says Walsh.

Testing looks promising

Testing in a controlled laboratory study shows the method can correctly identify the species in 96.5% of all test samples, and in the 2.7% of samples for which no species identity was provided, the system was able to correctly identify 67% to the family level, which is often enough information to select an effective therapy. Among over a thousand samples tested, the method never gave an incorrect result as to the family level or the Gram type.

Walsh says the research and development team in Durham is actively working on automating the LC-IF system with robotics to make it a fully hands-off process. Blood cultures grow in their own time, often producing a positive result at an inconvenient time of the day for clinical labs, he points out, so automation could speed up diagnosis significantly.

Next step: Automation

"Our vision is to have a system that will automatically identify the blood culture isolate within 15 minutes of the culture being called positive," says Walsh. If a culture is positive at 2 AM, he says, automating this method could make it possible to identify the organism by 2:15 AM and send an electronic report to a patient's physician. They hope be begin testing and evaluating the feasibility of an automated form of the system in a clinical environment within months.

How is the immune response to Giardia parasites like an onion? They've both got layers.

In mBio this week, researchers at Georgetown University use microarrays to study the immune response to Giardia in mice and show that in addition to antibodies and mast cells, the small

Mast cells, α-defensins and nitric oxide, makes a layered, redundant response to the Giardia parasite says co-author Steven Singer.

intestine also deploys matrix metalloprotease 7 and inducible nitric oxide synthase, and if one of these defenses is turned off, the other compensates for the loss.

"We know that the immune system has many weapons in its arsenal and we talk about the importance of redundancy, but there aren't many clearly demonstrated examples in the literature," says co-author Steven Singer.

Singer and his colleagues used transcription microarrays to monitor the activities of mouse genes in the small intestine 10 days after infection with Giardia parasites. Quantitative PCR was used to determine the functions of the 96 gene transcripts that were either upregulated or repressed more than 2-fold in apparent response to infection. Antibody genes, of course, were upregulated a great deal, and the most highly induced transcripts encoded mast cell proteases, known factors in the immune response to beaver fever.

Another upregulated gene was matrix metalloprotease 7 (Mmp7), the enzyme responsible for production of mature α-defensins in mice, microbicidal peptides. However, infecting mice that were Mmp7-deficient revealed only a small increase in parasite numbers, indicating that Mmp7 is not strictly necessary for fending off the pathogen.

Previous work had turned up another immune response that was not strictly necessary: the enzyme nitric oxide synthase (Nos2). Nitric oxide produced by Nos2 works against parasites in the lab, but when the gene for the enzyme is knocked out in mice, the animals can, nonetheless, mount an effective control response. The authors reasoned that perhaps these mechanisms - Mmp7 and Nos2 - act in a redundant fashion, as overlapping means of Giardia control in the gut.

Take one out, the other steps up

Testing with knockout mice proved them right. They crossed Nos2- and Mmp7-deficient mice for two generations to get mice homozygous and heterozygous for each of these traits. When they infected and genotyped animals, they found that deleting either Nos2 or Mmp7 has a minimal effect on the ability of a mouse to fend off Giardia. However, eliminating both at the same time resulted in significant increases in the numbers of trophozoites recovered from infected mice, indicating that both α-defensins and nitric oxide are redundant mechanisms for control of Giardia infections in vivo. When you take one defense out, the other steps up to do the job.

This layered, redundant approach to fending off Giardia with antibodies, mast cells, α-defensins and nitric oxide, makes a difficult path for the parasite in the body, says Singer.

"The immune system can make several different types of patterned responses," says Singer. "Most of the time the system 'chooses' the right pattern to eradicate the pathogen which triggered the response.  Having multiple effector mechanisms which can substitute for each other, increases the likelihood that an individual can successfully control their infection."

Gut microbiome helps determine risk of tumors

The gut microbiome plays a role in a number of phenomena, including immunity, metabolism, and disease, but it might also play a role in tumorigenesis. According to the results of a study in mBio

Germ-free mice that consumed the feces of tumor-bearing mice developed 2x more tumors than control mice.

this week, transferring the gut microbes from a mouse with colon tumors to germ-free mice makes those mice prone to getting tumors as well. Scientists have known for years that inflammation plays a role in the development of colorectal cancer, but this new information indicates that interactions between inflammation and subsequent changes in the gut microbiota create the conditions that result in colon tumors.

Surprisingly clear results

Co-author Patrick Schloss of the University of Michigan was somewhat surprised by the clarity of the results. "We saw more than two times the number of tumors in mice that received the cancerous community," than in mice that received a "healthy" gut community, says Schloss. "That convinced us that it is the community that is driving tumorigenesis. It's not just the microbiome, it's not just the inflammation, it's both."

Known risk factors for developing colorectal cancer include consuming a diet rich in red meat, alcohol consumption, and chronic inflammation in the gastrointestinal tract (patients with inflammatory bowel diseases, such as ulcerative colitis, are at a greater risk of developing colorectal cancer, for instance). Cancer patients also exhibit shifts in the composition of their gut microbiota - a phenomenon called dysbiosis - but it's unclear whether changes in the microbiome drive the development of cancer or the cancer drives changes in the microbiome. It's a question of the chicken and the egg, says co-author Joseph Zackular. "Is this the microbiome of someone with cancer or is the microbiome driving tumorigenesis?" asks  Zackular.

Representative images of tumors in the distal colon of mice conventionalized with a healthy microbiome (n = 10) or the microbiome of tumor-bearing mice (n = 9).

Schloss, Zackular, and their colleagues reasoned that the composition, structure, and functional capacity of the gut microbiome all directly affect tumor development in the colon, so they set out to address this chicken-and-egg conundrum with mice. Using a tumor-inducing regimen, they induced the formation of colorectal tumors in a set of mice, then collected feces and bedding from those tumor-bearing mice and gave them over to germ-free mice. (Mice, are coprophagic, so inoculating germ-free mice with a new gut microbiome is as easy as that.) They then administered the regimen to these new mice.

Passing colon tumors to germ-free mice

The results were stark: mice given the microbiota of the tumor-bearing mice had more than two times as many colon tumors as the mice given a healthy microbiota. What's more, normal mice that were given antibiotics before and after inoculation had significantly fewer tumors than the mice that got no antibiotics, and tumors that were present in these antibiotic-treated mice were significantly smaller than tumors in untreated mice. This suggests that specific populations of microorganisms were essential for the formation of tumors, so the researchers then drilled down into which groups of bacteria were present in the test animals and controls.

Looking at the microorganisms, they found that tumor-bearing mice harbored greater numbers of bacteria within the Bacteroides, Odoribacter, and Akkermansia genera, and decreased numbers of bacteria affiliated with members of the Prevotellaceae and Porphyromonadaceae families. Three weeks after they were inoculated with the communities from the tumor-bearing mice, the germ-free mice had a gut microbiome that was very similar to the tumor-bearing mice, and they had a greater abundance of the same bacterial groups associated with tumor-formation.

"In all these [mouse] models the inflammation is critical, but so is the change in the communities," says Schloss. "We liken it to a feed-forward type mechanism where the inflammation is changing the community and the community is inducing inflammation. They make each other worse to the point that you have higher rates of tumor formation."

Implications for humans?

The work has implications for human health, Zackular points out, because it indicates the risk of colorectal cancer may well have a microbial component. "We know that humans have a number of different community structures in the gut. When you think about it, maybe different people - independent of their genetics - might be predisposed," says Zackular.

To follow up on the work, Schloss and Zackular are now studying the functions of the groups that are and are not associated with tumor formation. "If you can better understand what functions in the microbial community are important for protecting against tumor formation or making it worse, we can hopefully translate those results to human to understand why people do or do not get colorectal cancer, to help develop therapeutics or dietary manipulations to reduce people's risk," says Schloss.

Microbiome in gut, mouth, and skin of low birth weight infants differentiate over first weeks after birth

Low birth weight infants are host to numerous microorganisms immediately after birth, and the microbiomes of their mouths and gut start out very similar but differentiate significantly by day 15 according to a study in mBio this week. Researchers from Stanford University and the University of Pittsburgh School of Medicine examined changes in the oral, skin, and gut microbiomes of low birth weight infants over the course of the first three weeks after birth and found that although the microbiomes in each of these sites start our markedly similar, they gradually differentiate over time. This is the first time the differentiation of the microbiota in multiple body sites in newborn infants has been investigated.

"We could watch this differentiation over time. With each passing day, two body sites [mouth and distal gut] became more and more differentiated from each other. It was a consistent pattern," says co-author Elizabeth K. Costello of Stanford University.

Does developing microbiome play a role in susceptibility?

Low birth weight infants, who are often born premature, are more susceptible than normal weight infants to invasive infections like necrotizing enterocolitis, a vulnerability that may be related to colonization by bacteria from their surroundings. Unlike adults, the microbiomes of the mouth, skin, and gut of infants right after birth are undifferentiated, says Costello - the microbiomes look more or less similar at each of these body sites. The researchers sought to find out how rapidly the communities of microbes in these different sites take on a character of their own.

"We chose to look at premature infants between the ages of eight and 21 days old and asked, over this time period, what is going on with their oral, gut and skin communities," says Costello. The period from 8 to 21 days after birth marks a critical window for colonization of an infant, and it's also the period of onset for necrotizing enterocolitis (although none of the infants in this study were struck by the disease).

A divergence in microbiotas

Figure 2: Unweighted UniFrac-based principal coordinate analysis (PCoA) of LBW infant-associated bacterial communities.

The researchers collected stool, saliva, and skin swabs from six low birth weight infants (five of whom were born premature) that ranged in weight from 1.65 - 4.01 lbs on postnatal days 8, 10, 12, 15, 18, and 21. They amplified, pyrosequenced, and analyzed the bacterial 16S genes present in each sample and compared them with analogous data from normal-birthweight infants and healthy adults.

In the 8 - 21 day age range, there was a subtle but important divergence in the composition of the oral and gut microbiotas, a differentiation that was mostly driven by changes that evolved in the composition of the gut microbiome. The babies' microbiomes were also dominated at times by bacterial types that have been associated with newborn infections and necrotizing enterocolitis, including Staphylococcus, C. perfringens, P. aeruginosa and others.

The scientists also tracked the effects of antibiotic treatment in one infant in the study, noting the rise of a type of Mycoplasma in the mouth that has previously been associated with vaginal infection.

Of the three sites studied, neonatal skin was the most adult-like in its microbiota composition. And like healthy adults, the microbiota of the different body sites in the infants was apparently determined mostly by body site and by host individual.

Co-author Michael J. Morowitz of the University of Pittsburgh School of Medicine says understanding the vulnerability of preemies from a microbial perspective can provide insights into how to better care for them.

"Premature infants are unique because they can spend several months in the hospital, where they're exposed to virulent bacteria, they also frequently have antibiotics exposure, and they're dietary intake is irregular - sometimes they're not able to eat anything by mouth. That probably effects colonization patterns," says Morowitz. "The first step [in understanding this] is to define what's normal for these infants and what's abnormal."

Moving forward to a predictive understanding of infant susceptibility

Morowitz, Costello and others are currently working on a larger study to expand on these results. By studying more babies, and by monitoring microbial colonization of low birth weight infants who eventually develop infections like necrotizing enterocolitis, they hope to better pinpoint the microbial profile of babies who are susceptible to disease, hopefully leading to more informative surveillance techniques and better interventions to help keep these most vulnerable infants on the path to becoming healthy children and adults.