Mycobacterium abscessus biofilms have viscoelastic properties which may contribute to their recalcitrance in chronic pulmonary infections

Mycobacterium abscessus is emerging as a cause of recalcitrant chronic pulmonary infections, particularly in people with cystic fibrosis (CF). Biofilm formation has been implicated in the pathology of this organism, however the role of biofilm formation in infection is unclear. Two colony-variants of M. abscessus are routinely isolated from CF samples, smooth (MaSm) and rough (MaRg). These two variants display distinct colony morphologies due to the presence (MaSm) or absence (MaRg) of cell wall glycopeptidolipids (GPLs). We hypothesized that MaSm and MaRg variant biofilms might have different mechanical properties. To test this hypothesis, we performed uniaxial mechanical indentation, and shear rheometry on MaSm and MaRg colony-biofilms. We identified that MaRg biofilms were significantly stiffer than MaSm under a normal force, while MaSm biofilms were more pliant compared to MaRg, under both normal and shear forces. Furthermore, using theoretical indices of mucociliary and cough clearance, we identified that M. abscessus biofilms may be more resistant to mechanical forms of clearance from the lung, compared to another common pulmonary pathogen, Pseudomonas aeruginosa. Thus, the mechanical properties of M. abscessus biofilms may contribute to the persistent nature of pulmonary infections caused by this organism.

The study of biofilm mechanics has gained interest, as it is being realized that these properties are important for understanding biofilm biology and how biofilms respond to chemical and mechanical forms of eradication 15 . Interestingly, changes in colony morphology have been correlated to differences in biofilm formation 16 , and differences in biofilm mechanics 17,18 . We therefore hypothesized that the biofilms of Ma Sm and Ma Rg may have different mechanical properties, due to their distinct colony morphologies, and that these properties may help to account for the persistence of infection caused by this organism. In the present study, we used mechanical indentation and shear rheometry to evaluate the viscoelastic properties of Ma Sm and Ma Rg biofilms. We also use known concepts of mucus viscoelasticity to predict how effectively M. abscessus biofilms may be cleared from the lung by mucociliary and cough clearance. Our findings support the hypothesis that biofilm viscoelasticity could contribute to the persistence of biofilm-associated infefctions 15,19 , and provides novel insight into why M. abscessus may be such a persistent pathogen in airway infection in people with CF.

Ma colony-biofilms maintain the distinct morphologies of Ma Sm and Ma Rg variants. Mycobac-
terium abscessus is a rapidly growing NTM, typically showing colony morphology by 4 days on agar media. To test the hypothesis that biofilms of Ma Sm and Ma Rg variants might differ in their mechanical properties, uniaxial indentation and shear rheology was performed on 4 day M. abscessus colony-biofilms (See Supplementry Methods). As this is the first time that we have used this biofilm model for M. abscessus, we examined the colony-biofilms after 4 days of growth (Fig. 1). Macroscopically, biofilm formation was evident, covering the filter. Biofilm morphology maintained the characteristic phenotypes of each variant 11 . That is, biofilms of Ma Sm , had a smooth, almost mucoid appearance (Fig. 1A), while Ma Rg biofilms showed a cauliflower-like morphology (Fig. 1B) in agreement with our previous study showing cording at the periphery of Ma Rg colonies 11 . Importantly, when these colony-biofilms were disrupted, and plated to observe single cells, the number of cells within the biofilm was similar across the two variants, and no reversion across phenotypes was observed during this time (Fig. 1C), demonstrating that during the 4 day biofilm growth period, each variant was stable.
Ma Rg biofilms are stiffer than Ma Sm biofilms in uniaxial compression. To determine the stiffness of M. abscessus biofilms under a normal force (force that is applied perpendicular to the surface), uniaxial indentation was performed on 4 day M. abscessus colony-biofilms. During this analysis, biofilms are compressed and the required force is measured. This analysis is also used to determine the biofilm thickness, which revealed that biofilms of each variant were of similar thickness ( Fig. 2A), consistent with the observation that biofilms had a similar number of CFUs (Fig. 1C). Stress-strain curves revealed that Ma Sm and Ma Rg biofilms displayed a 'J-shaped' curve, where the biofilms became progressively stiffer as they were compressed (Fig. S1). Stress-strain curves also revealed that there were significant mechanical differences between Ma Sm and Ma Rg biofilms ( Fig. 2B;  and Ma Rg biofilms were analyzed using spinning-disc rheology. For these analyses a shear force (a force that is applied parallel to a surface) is applied and the resulting stress or strain is measured (See Supplementary Methods). Oscillatory strain sweeps were performed, where the oscillatory strain was incrementally increased and the storage (G') and loss (G") moduli measured, which reflect the elastic and viscous response, respectively 15 . The measured storage and loss moduli plateaus were similar for Ma Sm and Ma Rg biofilms (Fig. 3A,B), suggesting that biofilms of each variant behaved similarly within the linear viscoelastic region. From this analysis we determined the yield strain, which represents the strain where the biofilm integrity begins to break down, due to the increasing applied strain, and transitions to fluid-like behavior 20 . Interestingly, the yield strain of Ma Sm biofilms was significantly greater than Ma Rg biofilms ( Fig. 3C; p = 0.0098). This suggests that Ma Sm biofilms are more pliant than Ma Rg biofilms, and that Ma Sm biofilms can be deformed to a greater extent before cohesive failure occurs. This is also consistent with the Young's modulus of these biofilms, which indicated that Ma Sm biofilms can be compressed more readily compared to Ma Rg biofilms (Fig. 2C).
Oscillatory frequency sweeps were also performed, where the oscillatory frequency was incremented under a consistent applied strain, and the storage and loss moduli measured. Across the range of frequencies analyzed, the storage and loss moduli were relatively independent of frequency, with the storage modulus greater than the loss modulus for both M. abscessus variant biofilms (Fig. 4A,B). This indicates that both Ma Sm and Ma Rg biofilms displayed elastic behavior that was dominant across the analyzed conditions. There were no significant differences between the mechanical behavior of Ma Sm and Ma Rg biofilms using this analysis (Fig. 4C,D). However, compared to Ma Sm biofilms, Ma Rg biofilms trended toward a reduced loss modulus across the analyzed frequencies (Fig. 4D).
Theoretical indices suggest that M. abscessus biofilms, independent of morphotype, may resist clearance from the lung. To determine if the mechanical properties of M. abscessus biofilms may    (2) and (3), using values from the frequency sweep analysis of these biofilms (Fig. 4). The MCI and CCI were developed to correlate sputum viscoelasticity to predicted levels of clearance from the lung via either mechanism 21,22 . The MCI and CCI of Ma Sm and Ma Rg biofilms were similar. However, due to the highly viscoelastic behavior of both colony variant biofilms (Fig. 4A,B), the indices were less than the MCI and CCI reported for sputum collected from people with CF ( Fig. 5; line).  www.nature.com/scientificreports/ To compare M. abscessus biofilms to the biofilms of another common pathogen causing chronic pulmonary infections in people with CF, we analyzed 4 day wild type Pseudomonas aeruginosa colony-biofilms 23 . P. aeruginosa is considered a model organism for biofilm formation, and we have previously determined the MCI and CCI for this organism 17 . Using the same testing conditions for M. abscessus biofilms described here, we performed oscillatory frequency sweeps on 4 day wild type P. aeruginosa colony-biofilms (See Supplmentary Results; Fig. S2), and similarly determined the MCI and CCI for comparison (Fig. 5). The indices of Ma Sm and Ma Rg biofilms were significantly less than that determined for wild type P. aeruginosa biofilms (Fig. 5). This suggests that M. abscessus biofilms may be more resistant to mechanical clearance from the lung, compared to other common pulmonary pathogens, such as P. aeruginosa.

Discussion
Compared to other biofilm forming organisms, such as P. aeruginosa, the understanding of biofilm formation and the role of biofilms in infection of NTMs is limited. It is important to understand biofilm mechanics from a biofilm biology stand-point, as these properties dictate the stability of the three-dimensional community structure 15 . We therefore set out to analyze the mechanical properties of M. abscessus in vitro biofilms to gain a better understanding of the biofilm biology of this organism. We found that Ma Rg colony-biofilms were stiffer under normal forces (Fig. 2C), while Ma Sm colony-biofilms were more pliable under both normal and shear forces (Figs. 2C, 3C). In constrast to the uniaxial indentation analysis (Fig. 2), oscillatory frequency sweeps indicated that there were no significant difference in the viscoelasticity of Ma Sm and Ma Rg biofilms under shear forces (Fig. 4). This indicates that M. abscessus biofilms were anisotropic, in that they have different mechanical properties when exposed to either a normal or shear force.
Earlier research by others suggested that only Ma Sm variants formed biofilms and speculated that GPL expression enhanced sliding motility in CF mucus, making it better adapted to behave as a colonizing, biofilm-forming phenotype 24,25 . However, we have previously shown that the rough morphotype is hyper-aggregative 11 , indicating that each colony morphology variant is capable of forming biofilm aggregates. The finding here that Ma Sm colony-biofilms were more pliable under both normal and shear forces (Figs. 2C, 3C) suggests that GPL likely affects the mechanical properties of M. abscessus biofilms. Further studies are needed to better understand the relationship of M. abscessus cell wall determinants, such as lipids, GPLs, and extracellular polyermic substance (EPS) with biofilm function and is the focus of future studies.
Uniaxial indentation, revealed that Ma Sm and Ma Rg biofilms each displayed a 'J-shaped' curve in response to increasing compression (Fig. 2B). This response is typical of biological materials, and has previously been observed for a number of different types of bacterial biofilms 17,26-28 . The Young's modulus of Ma Sm and Ma Rg colony-biofilms, determined here, is similar to that reported for Streptococcus mutans hydrated biofilms 29 , P. aeruginosa colony-biofilms 17 , and surface adhered Staphylococcus aureus, Staphylococcus epidermidis and Streptococcus salivarius cells 30 .
Interestingly, Ma Sm and Ma Rg colony-biofilms have a storage modulus, which describes the elastic solidlike behavior 31 , approximately tenfold greater than that previously observed for other bacterial biofilms (Fig. S2A) 17,18,20,[32][33][34][35][36][37][38][39][40] . This suggests that M. abscessus biofilms, regardless of the morphotype, are highly viscoelastic, and much stiffer than biofilms formed by other bacterial pathogens. It would, therefore, be of interest to compare the mechanical properties of other NTM biofilms, as well as other biofilms that have a lipid dominant EPS, to determine the contributions of lipid content to the EPS and to the biofilm mechanical properties. It is interesting to speculate that different host niches in the lung (or other infection sites) may favor stiffer or softer biofilms depending on different shear and chemical micro environments. However, this requires further clinical work.
Viscoelasticity is a property common to biofilms, suggesting that it is an adaptive trait of these mircobial communites 15 . It has therefore been predicted that biofilm viscoelasticity, not only is important from a biofilm biology stand-point, but also when considering microbial survival 41 . Understanding how biofilm mechanics impacts microbial survival in an infection is still unclear. However, there is a growing consensus that these properties are important for resisting both mechanical and chemical methods of eradication 15,19 . To gain further insight into how the mechanics of M. abscessus biofilms may influence the recalictrance of these infections, we used known, well-established concepts of mucus viscoelasticity to draw new hypotheses that may be relevant to M. abscessus biofilms in vivo, in the context of mechanical clearance from the lung. Changes in mucus viscoelasticity in CF are attributed to mucus hyper-secretion and changes to mucus composition, impeding mucus clearance by both mucociliary and cough clearance 42 . MCI and CCI are theoretical indices that were developed from in vitro lung clearance models to correlate viscoelastic properties of expectorated mucus with predicted levels of clearance from the lung 21,22 . These indices have been used to assess the efficiency of mucolytics for therapy regimes for people with CF 43 . MCI predicts that mucus elasticity correlates to improved mucociliary clearance by promoting efficient cilia beating 21,44 , while CCI predicts that mucus viscosity correlates to improved cough clearance by promoting mucus-airflow interactions 44,45 . We previously used these known relationships between mucus viscoelasticity and clearance, and applied these concepts to P. aeruginosa colony-biofilms, to determine if the viscoelastic properties of bacterial biofilms could likewise theoretically impact mechanical clearance from the lung 17 . We identified that elastic non-mucoid P. aeruginosa biofilms had a reduced CCI, while mucoid P. aeruginosa biofilms that had a low viscosity, had both a reduced MCI and CCI 17 . Here we determined the MCI and CCI of M. abscessus biofilms, and observed that they had an even lower or negative CCI compared to P. aeruginosa, attributed to the high viscoelasticity of these biofilms (Fig. 5B). This indicates that M. abscessus biofilms of each morphotype may resist clearance from the lung by cough. These findings provide novel insight into possible reasons why M. abscessus is a persistent and difficult to treat pulmonary pathogen. www.nature.com/scientificreports/ A longitudinal rheological study of sputum from CF patients indicated that both viscosity and elasticity increase during pulmonary exacerbations, suggesting that CF sputum viscoelasticity is linked to disease state and airflow obstruction 46 . By applying known concepts of mucus rheology to biofilm mechanics and infection, our findings provide a novel way to understand M. abscessus biofilms and their recalcitrance. Moving forward, it will be important to develop in vitro and in vivo models of biofilm-mucus interactions to explore the hypotheses drawn from this work. More broadly, the impact of evaluating bacterial biofilm mechanical properties offers new avenues for the treatment of intractable pulmonary infections. Disruption of M. abscessus aggregates increased susceptibility to several antibiotics 13 . Moreover, liposomes, lipid nanoparticles and nanoparticle lipid carriers are being evaluated for pulmonary use 47 . This approach has been used on biofilms of Burkholderia cepacia complex, as well as S. aureus and P. aeruginosa biofilms in vitro 48,49 . These products not only have the potential to reduce exposure to high concentrations of antimicrobial agents following aerosol administration by reducing toxicity, but particularly relevant to NTM antibiotic therapy, might ameliorate the severe side effects often associated with antimycobacterial treatment. This strategy could further be used to target NTM, which have a high lipid content, to facilitate cohesive failure of biofilm aggregates in the lung in adddtion to delivering targeted antibiotics.

Materials and methods
M. abscessus culture and biofilm model. The colony-biofilm model was used here due to the established utility of this model for rheological analysis of biofilms 17,50 (See Supplementary Methods). Colony-biofilms were grown as previously described with modifications 17 . Briefly, M. abscessus cultures were prepared by inoculating a 10 μL loop of isolated Ma Sm or Ma Rg from 7H10 agar plates into 7H9 without tween media and a single cell suspension prepared as described 11 . Cultures were normalized to an OD 600 of 0.15-0.2 and 100 μL was pipetted onto sterile nitrocellulose filter membranes (25 mm, 0.45 μm pore size; Millipore) and incubated at 37 °C, 5% CO 2 under humidified conditions for 4 days. This time point was selected as it provided sufficent biomass necessary for rheological analysis, while maintaining the characteristic morphology of Ma Sm or Ma Rg variants. Colony-biofilms of PAO1 P. aeruginosa were grown as previously described 17 .
Colony-biofilms were imaged using a Stereo Microscope (AmScope) fitted with a Microscope Digital Color CMOS camera (AmScope). Images were processed in FIJI 51 . Figure was complied in Biorender.com.
To assess the number of CFU/cm 2 on sterile membranes after 4 days, colony-biofilms were scraped and transferred to 5 mL of 7H9 + tween 80 and vortexed with glass beads to disperse cells from the biofilm 11 . Cell suspensions were serially diluted and enumerated for CFUs in triplicate on 7H10 agar plates. To determine if either variant converted/reverted to another colony variant during biofilm growth, colony morphology was also assessed and verified by sub-culturing three colonies per replicate to ensure that the colony morphology was stable. CFUs were plated to 10 8 dilution and colonies were streaked to isolation to validate colony variants.

Rheometry apparatus.
A Discovery Hybrid Rheometer-2 with a heat exchanger attached to the Peltier plate (TA Instruments) was used for all rheological measurements. For uniaxial indentation and spinning disc measurements (See Supplementary Methods), the rheometer was fitted with an 8 mm and 25 mm sand-blasted Smart Swap geometry, respectively. All measurements were performed at 37 °C, with the Peltier plate covered with a moist Kimwipe to prevent dehydration of the colony-biofilm. TRIOS v4 (TA instruments) software was used.
Uniaxial indentation measurements. Uniaxial indentation measurements were performed under compression using an approach rate of 1 μm/s. Two colony-biofilms were analyzed per biological replicate (total of 4 biofilms analyzed), with two measurements per biofilm. The point of contact with the biofilm was determined to be where the force began to increase after the initial point of pull-on adhesion. Force-displacement curves were converted to stress-strain curves, for ease of comparison, as previously described 17 . The Young's modulus was calculated according to Eq. (1) 52 : where v is the assumed Poisson's ratio of a biofilm (v = 0.5) 27 and r is the radius of the geometry. The slope is of the force-displacement curve, and was taken at the lower linear region corresponding to 0-30% strain where R 2 ≥ 0.9.
Spinning-disc rheology. To normalize for differences in biofilm thickness across replicates, prior to analysis, colony-biofilms were compressed to a normal force of 0.01 N. For all analyses a total of 4 biofilms were analyzed, 2 per biological replicate.
Strain sweeps were performed by incrementing the oscillatory strain from 0.01 to 100% at a frequency of 1 Hz. The yield strain was taken where the storage (G') and loss (G") modulus intersected. Frequency sweeps were performed by incrementing the oscillating frequency from 0.1 to 100 rad/s at a constant strain of 0.1%. This strain was determined to be within the linear viscoelastic region of the strain sweeps for both Ma Sm and Ma Rg colony-biofilms.
Theoretical mucociliary clearance index (MCI) and cough clearance index (CCI), developed from in vitro models of mucus clearance 32,33 , were calculated according to Eqs. (2) and (3), respectively, using the relationship between the complex modulus (G*) and tanδ values determined from frequency sweep measurements 21,22,53 . The MCI and CCI was calculated from values determined at an angular frequency of 1 rad/s and 100 rad/s, respectively. (1)