Time for a Culture Change

At a Glance

• Lung infections in patients with cystic fibrosis are commonly diagnosed using microbial culture
• The traditional method is far from perfect though; it’s subject to bias, can miss disease-causing bacteria, and may take several days to return a diagnosis
• Molecular methods are frequently used in research settings and demonstrate the diversity of bacteria found in lung infections, but many hospital labs currently lack the resources
• A rapid and accurate PCR method described here, carried out in tandem with viral analysis, could bring affordable molecular testing to hospital labs

Frequent lung infections are a hallmark of cystic fibrosis (CF), can lead to severe clinical decline, and contribute to the cause of death in the vast majority of cases. Traditionally, lung infection is diagnosed by plating sputum from patients onto different microbial media, and looking for the growth of major CF pathogens such as the bacterium Pseudomonas. But this culture-based approach has not substantially changed in half a century, even though it’s well known that CF sputum contains many different bacterial species, not all of which are easy to grow in medium (see Figure 1). Getting results can take days, and a culture of the dominant or disease-causing pathogen is not guaranteed. The problems we observed motivated my research team, and our clinical collaborators at the Manchester Adult Cystic Fibrosis Centre, UK, to tackle the challenge of developing a straightforward molecular test that could be used by clinical microbiology labs to better diagnose CF infections in public health settings.

Culture is a tried and tested method for identifying and quantifying the bacterial species present in a sample, but its known to be open to multiple flaws and biases. The approach has been fine-tuned to allow the capture and identification of key pathogens, but it frequently ignores other bacteria (1). It is known that numerous infections are polymicrobial, including lung and other respiratory infections, urinary tract infections, wounds, ulcers, abscesses and many others. Molecular analyses have also demonstrated that anaerobic bacteria are present in large numbers in many infections, but most standard aerobic diagnostic methods cannot account for them. In contrast, molecular tests can detect all bacteria (and other microorganisms) present in these infections from their characteristic DNA signatures, and can return results much faster. I believe their applicability, accuracy and speed makes them an obvious candidate for more widespread use, and could bring many advantages to clinical bacteriology, but with the complexity and expense of most molecular technologies, smaller labs often feel these genetic tools are out of their reach.

Moving to molecular, inexpensively

In the last 10 years, multiple research studies have already adopted molecular methods. But DNA sequencing-based techniques, although the “gold standard”, are currently beyond the capability of many diagnostic microbiology labs.

We evaluated a simple diagnostic PCR method, known as ribosomal intergenic spacer analysis (RISA) PCR, to see if it could be used as an alternative to expensive sequencing methods to rapidly (and inexpensively) profile the mixture of bacteria, and detect problematic pathogens, in CF sputum. RISA PCR amplifies the space between two ribosomal genes that are present in all bacteria, allowing universal detection, but crucially the size of this spacer region varies between different species – for example, Pseudomonas has a 753 base spacer, in contrast Achromobacter has an 887 base spacer. So, if we analyze DNA from a sample containing these two bacteria, two PCR bands will be seen, giving a simple profile that captures the bacterial diversity in the sample (see Figure 2).

We collected 200 sputum samples from 93 adult CF patients, and each sample was split for use in culture, and molecular analysis. In 11 cases, a CF pathogen was not found in culture, but was identified using the RISA profile. Some of these samples were characterized as containing “normal flora”, despite known CF pathogens being detected using the molecular profiling tests (2).

We simplified the process further by using readily-available DNA from the sputum samples, which had been robotically extracted in order to test for viruses, such as influenza – molecular testing for viral infections is much more routine, as they are extremely difficult to grow in culture. Making use of this readily-available DNA would allow for a two-pronged diagnostic approach to simultaneously and rapidly diagnose bacterial and viral lung infection, using resources already available to many hospital labs.

Spotting missing microbes

When we put the RISA technique to the test, we were successful in quickly identifying CF patients with lung infections dominated by bacteria such as Pseudomonas and other multidrug resistant microbes known as emerging, non-fermenting, Gram-negative bacteria. Worryingly, we also found that in 11 percent of cases, these pathogens, which are associated with severe infection, were missed by routine growth-based bacteriology. The clinical condition of people with lung infections dominated with these bacteria will deteriorate unless the infection is treated appropriately, and we believe our research provides a means of identifying and addressing these infections, something that isn’t always possible with non-molecular methods.

Profiling bacterial diversity in CF has other uses, too – when carrying out ribosomal ribonucleic acid (rRNA) gene pyrosequencing in a subset of 59 patients, we found that poor lung function significantly correlated to less microbial diversity, including a low abundance of Streptococcus in patients with poor lung function. rRNA pyrosequencing is currently considered to be a state-of-the-art method for analysis of lung microbiota (3), but is currently beyond the resources of many labs. We believe that RISA PCR is a useful alternative that can detect both dominant pathogens and a loss of microbial diversity.

Trials and translations

Molecular diagnosis offers multiple advantages – for the microbiologist, the most important are speed and accuracy. The former has already been demonstrated in molecular tests used routinely for notifiable pathogens such as Neisseria meningitis (where speed is crucial to diagnosing meningitis) and Mycobacterium tuberculosis (where its slow growth prevents timely diagnosis). Greater application of molecular testing to polymicrobial infections could also offer greater insight to the pathogens driving disease: it will proportionately identify the bacteria with the greatest abundance, rather than those that are easiest to grow. This may improve the therapies patients receive, by allowing more targeted antibiotics to be prescribed, and avoiding the use of broad spectrum agents which may be driving antibiotic resistance.

Following our promising initial results, we are now seeking translational funding in order to test the approach we have developed in real-time, and to compare it to routine practices for CF infection diagnosis. If our research can demonstrate the clear benefits we have seen so far, then we hope that they could soon be delivered as a new standard of care in CF microbiology. We’re also working with the pharmaceutical industry to examine the bacterial diversity in CF sputum samples before and after the application of novel therapeutics; this will be one of the first clinical trials to compare molecular methods with standard bacterial culture as an outcome measure.

The strategy of using the same sample for bacterial and viral testing could also be explored in the future as a means to further investigate the respiratory microbiome in a clinical setting, which could have uses not just in CF, but other respiratory diseases such as chronic obstructive pulmonary lung disease. We believe that molecular testing could potentially bring both improved care and decreased costs to clinical microbiology laboratories, and ultimately improve our understanding of lung infections, both in CF and beyond.