Microbial biofilm formation: a need to act

The interdisciplinary Biofilm Nobel conference aimed not only to cover recent progress in this research area, but also to bridge basic science and clinical disciplines. As such, the conference gathered leading investigators who collectively covered the broad spectrum of biofilm research from the molecular analysis of basic phenomena to the clinical impact and the prevention of biofilm formation (http://biofilmnobelconference.org/). Broad scoping topics were thematically organized during the five keynote lectures, nine sessions, and 67 posters. From a medical perspective, these included the clinical problem of persistent infections by biofilm-forming microorganisms in device-related as well as soft tissue infection, the molecular basis of foreign body infections; biofilms and the immune system; and biofilms and cancer. Important fundamental discoveries made in basic biofilm research with respect to structure and regulation of biofilm formation were presented, along with biofilm diagnostics, bioinformatics of biofilm formation, the resistance phenotype, and prevention of biofilm formation. Various strategies to treat biofilm infections were also discussed. Although biofilm formation was observed as early as 1708 when Antonie van Leeuwenhoek investigated tissue colonized by microbes with his new, and at that time, highly powerful microscope, research on biofilm formation only began to make significant advances in the last 15–20 years with exponential growth (Table 1). It is recognized that the developmental process of biofilm formation (the association of bacteria into multicellular consortia, consisting of still autonomous cells connected by an extracellular matrix) is not only the natural mode of life of many microorganisms 1, 2, but also an important clinical pathogenic mechanism 3, 4 strengthened by the aging population, immunocompromised and olytraumatic patients, and modern medical instrumental intervention (Fig. 1). In the USA alone, it is estimated that 1.7 million hospital-acquired infections annually involve microorganisms in the biofilm state, incurring an additional $11 billion in healthcare costs (US CDC, 2007). The total annual cost for biofilm infections in the USA is unofficially estimated to be $94 billion with more than half a million deaths 5. The dimension of community acquired biofilm infections is also huge, with, e.g., US$105 billion spent on dental care alone, making dental caries and periodontitis the most prevalent biofilm-associated diseases 3. The spectrum of biofilm diseases is wide. While catheter-associated urinary tract infection (CAUTI) is the most common device-associated biofilm infection, central line-associated blood stream infections (CLABSI) and ventilator-associated pneumonia (VAP) are also of significant concern (Fig. 1). Although historically the most prominent biofilm-associated disease might be Pseudomonas aeruginosa-mediated lung infection in cystic fibrosis (CF) patients 6, soft tissue infections, e.g., diabetic foot ulcers and mucosal infections, including chronic and recurrent urinary tract infections, are main biofilm-associated diseases of today (Fig. 1). Basic science research has demonstrated that antibiotic resistance is transferred readily in biofilms. In line with this, Escherichia coli isolated from medical devices showed the highest incidence of resistance against cefuroxime and ciprofloxacin compared to isolates obtained from other body sites 7. CAUTI and prosthetic joint infections are two of the most relevant biofilm-associated foreign body infections, with great economical significance in health care 5. CAUTI, which accounts for 15% of all infections, is the second most common healthcare-associated infection in the USA, and virtually, all healthcare UTIs are associated with indwelling devices (http://www.cdc.gov/nhsn/pdfs/pscmanual/7psccauticurrent.pdf). The design of the Foley urinary tract catheter in use today has basically not changed since its invention in the 1930s. The probability that the catheter becomes infected is 1–5% per day, and after 4 weeks, nearly all patients will have high bacterial numbers in the urine 8. The high likelihood of microbial colonization arises as the Foley catheter undermines the local defense system in the urinary tract. Although there is an urgent need for catheter amelioration, until recently, there has been no development of catheters with novel design and material 8. Biofilm infections are costly. After hip or knee joint replacement, a prosthetic joint infection (PJI) with revision of the prosthesis is estimated to be $50–100 000 per patient. The PJI incidence within the first two postoperative years is 0.5–2%. As the number of arthroplasties in the USA is extrapolated to four million in 2030 from the one million in 2009 9, the incidence of PJIs will increase accordingly. Diagnosis of PJIs frequently requires approaches beyond standard microbiological tests 10. Biofilm infections are typically difficult to diagnose and recalcitrant to antimicrobial treatment. Suggested diagnostic guidelines to identify biofilm infections, therefore, consider several clinical criteria even in the presence of culture-negative samples 11, 12. The European Society for Clinical Microbiology and Infectious Diseases (ESCMID) is currently developing practical guidelines for clinical microbiologists and infectious disease specialists to diagnose some of the most common and prominent biofilm infections 13. Therefore, the identification of biological markers, which discriminate between biofilm and acute infections, will facilitate the choice of appropriate treatment including antibiotics. For example, transient lipopolysaccharide modifications are currently investigated as potential biofilm markers 14. Systematic analysis of the molecular mechanism of biofilm formation started as late as in the last 5 years of the nineteenth century (Table 1 gives a non-comprehensive overview over the milestones in biofilm research). Biofilm formation of microorganisms is now being recognized as an ancient developmental process towards multicellularity with common structural principles and regulation (Figs 2 and 3; 15, 16). With this recognition, facilitated by computer algorithms for aiding large-scale single-cell tracking and nano-resolution microscopy, biofilm formation can now be elucidated with previously unavailable temporal and spatial resolution 17, 18. The hallmarks of multicellularity, ubiquitous throughout the tree of life, include coordinated cellular behavior, division of labor between cells, and the protection from environmental stressors. Biofilm formation is characterized by the expression of an extracellular matrix, which has structural and physiological functions (Figs 2 and 3). Although differences in biofilm formation may exist even within a single strain grown under different environmental conditions, the structure and composition of the extracellular matrix follow common principles. Typically, a biofilm is composed of amyloid and adhesive fimbriae, non-fimbrial large surface proteins, exopolysaccharides, and extracellular DNA (eDNA) 19-21. Poly N-acetylglucosamine (PAG) and cellulose are common extracellular matrix polysaccharides in many phylogenetically diverse bacteria. Cellulose expression is common for gastrointestinal commensals and pathogens such as E. coli and Salmonella typhimurium, but cellulose is also a major matrix component in cyanobacterial mats and a cell wall component of some fungi 22. Identification of matrix polysaccharide composition, structure, and function is challenging. The structure of the Vibrio cholerae polysaccharide was only recently solved 22, 23, and the composition of P. aeruginosa's main matrix exopolysaccharides, Psl and Pel, is poorly understood and unknown, respectively. Structural and compositional information on biofilm matrix polysaccharides, once elucidated, can serve as targets for vaccination and biofilm infection diagnosis 25. Imprinting of the surface with exopolysaccharides before the establishment of microcolonies, the seeding units of mature biofilm structures 18, and cell differentiation, e.g., the development of hyper biofilm variants 26, 27 make biofilm formation analogous to tissue formation in higher organisms. The extracellular matrix of a biofilm features signaling properties similar to those of the multifunctional extracellular matrix of eukaryotic tissues, with extracellular molecules for communication (e.g., quorum sensing) and redox biology 20. These signaling events are continuous, bidirectional, and mediate interactions between cells in response to their immediate environment (a phenomenon known as dynamic reciprocity), just as tissue differentiation in higher organisms 28. Biofilm formation is regulated by the codon usage of biofilm genes 29, genetic factors including small RNAs 30, short secreted peptides 31, quorum sensing molecules 32, and by many environmental conditions. Although still controversial in the context of bacterial signaling 14, the presence of distinct genetic programs is required to coordinate the multicellular behavior. The most widespread globally occurring coordinated developmental program regulating biofilm formation in bacteria is the signaling network that involves the secondary messenger cyclic di-GMP 33. Cyclic di-GMP regulates the lifestyle transition between sessility and motility and, accordingly, also the transition between acute and chronic infections. This secondary messenger signaling system is very complex, involving a variety of cyclic di-GMP metabolizing enzymes and a diversity of physiological processes and reactions. These are regulated by cyclic di-GMP signaling directly and/or coupled through sophisticated molecular mechanisms that are mediated by a diversity of receptors to provide a fine-tuned output response 33-35. Cyclic di-GMP extensively crosstalks with other signaling systems such as the two-component systems 36. Cyclic di-AMP is another recently discovered novel secondary messenger molecule 37. It is responsible for coordinating stress responses and the peptidoglycan cell wall homeostasis and thereby affects biofilm formation and antibiotic resistance 38. Cyclic di-AMP has a phyletic distribution different from the cyclic di-GMP signaling system. Surprisingly, the presence of cyclic di-AMP is essential for some bacteria, although the molecular basis of cyclic di-AMP sensing molecules, proteins and RNA, is only beginning to be understood 39, 40. In summary, experimental studies of biofilms of various bacteria have revealed a great variety of signal transduction systems that control biofilm formation. At the same time, the availability of complete genome sequences allows prediction of all regulatory components in a given organism, and the subsequent identification of components that affect biofilm formation. An unexpected finding from such studies is that microbial signaling systems are far more diverse than previously thought and, furthermore, include a network of diverse sensors that transmit signals to a variety of regulatory mechanisms. One urgent task is to develop a system of genome-derived parameters that could serve as predictors of the organism's propensity to either stay motile or form a biofilm. Michael Galperin and colleagues have developed several metrics for genome comparisons, including signal protein family profiles, which reflect the abundance of each type of signal transduction protein and can be used to trace their evolution in the course of adaptation to specific ecological niches 41, 42. Besides being secondary messengers, the cyclic di-GMP and cyclic di-AMP secondary messenger molecules serve as (almost) unique microbial-associated molecular patterns (MAMPs) manipulating host responses and physiology. Listeria monocytogenes secretes cyclic di-AMP into the cytosol, triggering a type 1 interferon response that inhibits cell-mediated immunity 43. In eukaryotic cells, bacterial cyclic di-GMP and cyclic di-AMP are both sensed by the immune adaptor STING, which is not only a receptor for bacterial nucleotides, but also the intrinsic sensor of the non-canonical cyclic di-AGMP, the first and recently discovered eukaryotic cyclic di-nucleotide second messenger involved in innate immunity sensing 44, 45. Indeed, due to their unique immunostimulatory features, cyclic di-nucleotides are effective, non-toxic adjuvants 46. Cyclic di-GMP has been reported also to retard the growth of cancer cells 47. Thus, there is an immediate potential impact of biofilm research on seemingly remote research fields such as immunology and cancer biology. Biofilm formation is strongly dependent on environmental conditions. Historically, the relevance of structural and regulatory factors in biofilm formation during the infection processes has been derived from in vitro biofilm models such as adhesion to abiotic surfaces, colony morphology biofilm, pellicle formation and the widely used flow-cell system (Fig. 3). These factors are now being validated by relevant in vivo animal models, such as different foreign body infection models 48, 49. Certain bacteria selectively colonize cancer tissues 50. A surprising finding was the requirement of biofilm formation for the effective bacterial colonization of cancer tissue 51. Although the details of biofilm components required for effective tissue colonization are just emerging 52, the cancer biofilm model has the potential to serve as an in vivo model for routine screening of novel antimicrobial therapies. In addition, a determinative role of the immune response to bacterial biofilm formation in vivo has been observed 52. Current infectious diseases therapy is confronted by several challenges. Antibiotic resistance is rising rapidly, and there are relatively few novel compounds or strategies under development or under clinical testing. The intrinsic resistance of biofilm infections to antimicrobial treatment and even a competent immune system pose additional challenges. While some antibiotic agents such as rifampin have a reasonable activity against biofilm-forming cells 53, the majority of conventional antimicrobial treatment regimens are not effective against biofilms. Elucidation of the molecular mechanisms of antibiotic killing, on the one hand, and the developmental process of biofilm formation on the other have led to the identification of multiple novel points for targeted intervention strategies 121, 54. Strategies to prevent early onset of biofilm formation include manipulation of abiotic and biotic surfaces through rational surface design 55 and stimulation of the innate immune response 56, 57. Intensive screening activity for antibiofilm agents has identified compounds effective in diminishing the biofilm build-up or eradicating established biofilms. Promising antibiofilm agents include natural compounds that interfere with P. aeruginosa quorum sensing signaling 32. Pillicides and curlicides are novel compounds strategically developed against fimbriae involved in biofilm formation by E. coli causing UTIs 58. The latter compounds also reduce the formation of intracellular bacterial communities in the superficial epithelium 59 associated with chronic cystitis and recurrent infections. Lessons learned from those approaches and compounds identified might be useful also to fight biofilm formation in agricultural and technical settings, not the least derived from the fact that biofilm formation in medical settings, on plants and in technical domains shares, to some extent, common principles of adhesion and regulation [1, 33, 60, 61 ]. Systems analyses and network biology approaches have been used to study the actions of antibiotics and the mechanisms underlying the emergence of antibiotic resistance. Subsequently, different classes of bactericidal antibiotics were shown to induce alterations to central metabolism, cellular respiration, and iron homeostasis in Gram-negative and Gram-positive bacteria 62-64. These effects lead to the production of reactive oxygen species (ROS) and other damaging molecules, which ultimately contribute to antibiotic-induced cell death. Antibiotic-induced oxidative stress also causes DNA damage and can be enhanced by deletion of recA and disabling the SOS response, both involved in DNA repair. Further, engineered bacteriophages that overexpress the lexA3 allele, a non-cleavable variant of the repressor of the SOS response, enhance antibiotic-mediated killing by several orders of magnitude 65. The engineered bacteriophage can be used to restore antibiotic susceptibility to resistant strains and can be readily modified to target different gene networks. The contribution of the ROS pathway to antibiotic killing has been challenged 66; however, overall redox homeostasis is required for antibiotic resistance and cell viability 67, 68. The human innate immune system has an intrinsic capability to restrict the outgrowth of commensal and pathogenic microorganisms and, as such, is a rich source of antibiofilm agents 69. Based on observations that a cationic human host defense peptide LL-37 is able to inhibit biofilm formation and dissolve existing biofilms of P. aeruginosa 70, Hancock screened for and obtained small peptides 71, including protease-resistant variants, that have potent broad-spectrum antibiofilm properties. Importantly, antibiofilm activity was not correlative with antimicrobial activity versus planktonic (free swimming) cells 71. These novel peptides can kill multiple species of bacteria in biofilms (minimal biofilm inhibitory concentrations of >1–4 μg mL−1), including the so-called ESKAPE pathogens (multidrug-resistant strains of Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, P. aeruginosa, and Enterobacter spp.). Flow-cell biofilm observation demonstrated that these peptides can both prevent biofilm formation when added at the start of the biofilm development cycle and cause dispersal and/or death of bacteria in 2-day pre-formed biofilms. The novel peptides are able to work synergistically with several different antibiotics. Hancock proposed that these peptides work by a common mechanism in diverse bacteria and demonstrated that they are suppressing a stress response 72. While NO is a component of the innate immune system, it has also been demonstrated to act as a biofilm dispersal signal and is active against a wide range of biofilm-forming bacteria 73. NO is produced endogenously in mature biofilms and functions through the intracellular secondary messenger cyclic di-GMP, which in turn activates a range of effectors that result in biofilm dispersal 74, 75. Recently, the first cyclic di-GMP-specific phosphodiesterase was identified to be specifically involved in NO-induced dispersal in P. aeruginosa 76. Because NO induces dispersal at low, non-toxic concentrations, and its mechanism of action is broadly conserved across species, the use of NO signals represents a potentially promising strategy for biofilm control 77. NO can be delivered to biofilms by a range of means including the use of NO-donor compounds 78. NO can also be directly applied as a gas to infection sites exposed to air. Recently, the first clinical trial was conducted to evaluate the use of low-dose inhaled NO gas combined with standard intravenous antibiotic therapy for the disruption of P. aeruginosa biofilms in patients with CF. Patients who received NO gas at 10 ppm showed significant reductions in Pseudomonas biofilm aggregates and marginal improvement in lung function compared to patients who received a placebo. These results suggest that using NO as adjunctive therapy may be beneficial for the treatment of CF-related biofilm infections (Webb, Faust et al. in preparation). Due to the short half-life in biological systems, a preferred approach may be to deliver NO directly to the biofilm. An innovative new class of NO-donor prodrugs was described that can liberate NO upon specific activation by bacterial β-lactamase enzymes. These prodrugs enhance the efficacy of standard antibiotic therapies against in vitro biofilms and thus have significant clinical potential 79. Biofilms contain a high fraction of antibiotic tolerant quasi-dormant persister cells that are shielded from the immune system through the biofilm 80. Pathways leading to dormancy are highly redundant in E. coli and, potentially, in other bacteria. Induction of dormancy depends largely on the action of toxin/antitoxin modules 81, 82 where toxins function as protein synthesis inhibitors 83 and mRNA endonucleases 81, 82, 84. Further, damage of DNA induces expression of the TisB toxin 85, an endogenous antimicrobial peptide that opens an ion channel 86 causing decrease in the proton motive force (PMF) and ATP, resulting in antimicrobial target shutdown and consequently a dormant, drug-tolerant state. The multiplicity of dormancy pathways seemingly precludes development of drugs that could prevent persister formation 87, 88. However, Lewis reasoned that a compound capable of a unspecific corrupting action on a target in dormant, energy-deprived cells would kill persisters. The ClpP protease targeted by acyldepsipeptide (ADEP4) leads to non-specific activation of the enzyme independently of ATP in both growing and non-growing cells 89-91. Null clpP mutants are highly susceptible to killing by a variety of antibiotics, but resistant to ADEP4. As a consequence, the combination of ADEP4 with rifampicin eradicated persister cells in growing and biofilm populations of S. aureus in vitro. Importantly, a deep-seated biofilm infection in mice, not treatable with conventional antibiotics, could be cured. ADEP4 points to a general principle of killing by activation of a target with bacterial proteases such as Lon as additional candidate targets. Even in developing therapeutics to treat fungal infections and cancer this general principle of targeting might prove effective. Another global approach to eradicate persister cells is to target the aminoglycoside resistance of these cells. Aminoglycoside antibiotics require a threshold PMF and energy to be taken up by bacterial cells. The Collins group showed, using a network biology approach, that specific metabolic stimuli (e.g., sugars) enable the killing of Gram-negative and Gram-positive persisters with aminoglycosides 92. These metabolites act by restoration of the metabolism of persister cells, thereby causing persisters to take up aminoglycosides. This approach was demonstrated to be effective against bacterial biofilms on a medical implant as well as improving the treatment of chronic urinary tract infections. This work established an inexpensive, clinically viable means for treating persistent bacterial infections. The 60th Nobel Conference on Biofilm Formation held at the Nobel Forum, Karolinska Institutet from August 28-30, 2013 provided a forum to stimulate the exchange of ideas and cross-disciplinary discussions. As the borders between different biofilm research disciplines are virtual, with similar principles of biofilm formation and research approaches in the medical, technical and environmental biofilm research fields 93, this conference was a step toward synergistic integrative biofilm research. No conflicts of interest to declare. We appreciate the contribution of James J. Collins, Michael Galperin, Bob Hancock, and Kim Lewis to this conference summary. The conference was sponsored by the Nobel Assembly of the Karolinska Institutet, the Swedish Research Council, the Karolinska Institutet, and Afa Insurance.