Community composition shapes microbial-specific phenotypes in a cystic fibrosis polymicrobial model system

Article activity feed

1. 

   Version published to 10.7554/elife.81604 on eLife

   Jan 20, 2023
2. 

   eLife

   Dec 21, 2022

   Author Response

   Reviewer #1 (Public Review):

   The authors set out to develop an in vitro model of multiple species representing diversity in the CF airway as a platform for a range of studies on why polymicrobial communities resist therapy. The rationale for their design is sound and the methods appear justifiable and reproducible. The major strength of this work is in producing a method for a range of future work, ideally for multiple groups in the field. The primary findings are interesting but not groundbreaking. One weakness in the method of reporting interspecies interactions and another in evaluating alternative causes of lasR advantages present opportunities for a stronger research contribution beyond this terrific method.

   We thank the reviewer for this accurate summary of the data presented in our manuscript. We have …

   Author Response

   Reviewer #1 (Public Review):

   The authors set out to develop an in vitro model of multiple species representing diversity in the CF airway as a platform for a range of studies on why polymicrobial communities resist therapy. The rationale for their design is sound and the methods appear justifiable and reproducible. The major strength of this work is in producing a method for a range of future work, ideally for multiple groups in the field. The primary findings are interesting but not groundbreaking. One weakness in the method of reporting interspecies interactions and another in evaluating alternative causes of lasR advantages present opportunities for a stronger research contribution beyond this terrific method.

   We thank the reviewer for this accurate summary of the data presented in our manuscript. We have addressed the raised concerned in the revised document. The modifications and comments can be seen in the “Essential Revisions” section above.

   Reviewer #2 (Public Review):

   Differences between the infection environment and in vitro model systems likely contribute to disconnects between the antimicrobial susceptibility profile of bacterial isolates and the clinical response of patients. The authors of this paper focus on a specific aspect of the infection environment, the polymicrobial nature of some chronic infections like those in people with Cystic Fibrosis (CF), as a factor that could impact antibiotic tolerance. They first use published genomic datasets and computational techniques to identify a clinically relevant, four-member polymicrobial community composed of Pseudomonas aeruginosa, Staphylococcus aureus, Streptococcus spp., and Prevotella spp. They then develop a high throughput methodology in which this community grows and persists in a CF-like environment and in which antibiotic susceptibility can be tested. The authors determine that living as a member of this community decreases the antibiotic tolerance of some strains of biofilm-associated P. aeruginosa and increases the tolerance of most strains of planktonic and biofilm-associated S. aureus and planktonic and biofilm-associated Streptococcus. They focus on the decreased tolerance of P. aeruginosa and determine that a ΔlasR mutant of P. aeruginosa does not display increased tobramycin susceptibility in the mixed community. One of the phenotypes associated with a ΔlasR mutant is an overproduction of phenazines. The authors find that by deleting the phenazine biosynthesis genes from ΔlasR, they can restore community-acquired susceptibility. They further investigate this phenomenon by showing that a specific type of phenazine, PCA, is significantly increased in mixed communities with the ΔlasR mutant compared to WT. Finally, they demonstrate that adding a specific phenazine, pyocyanin, to mixed communities can restore the tolerance of WT P. aeruginosa.

   Strengths:

   With this study the authors address a very important problem in infectious disease microbiology - our in vitro drug susceptibility assays do a poor job of mimicking the infection environment and therefore do a poor job of predicting how effective particular drugs will be for a particular patient. By demonstrating how an infection-relevant community modifies tolerance to a clinically relevant drug, tobramycin, the authors identify specific interactions that could be targeted with therapeutics to improve our ability to treat the chronic infections associated with CF. In addition, this study provides a framework for how to effectively model polymicrobial infections in vitro.

   The experiments in the paper are very rigorous and well-controlled. Statistical analysis is appropriate. The paper is very well-written and clear.

   The authors do an admirable job of using in silico analysis to inform their in vitro studies. Specifically, they provide a comprehensive rationale for why they chose and studied the specific community they did.

   The authors provide a very robust dataset which includes determining how strain differences of each of their four community members affect community dynamics and antibiotic tolerance. These types of analyses are laborious but very important for understanding how broadly applicable any given result is.

   We appreciate the reviewer’s thorough summary of our work and their positive comments.

   Weaknesses:

   The authors very clearly and convincingly demonstrate that WT P. aeruginosa becomes more susceptible to tobramycin in their mixed community. Our ability to turn these types of observations into therapeutic development depends on mechanistic insight. That said, it is unclear if the authors can make any solid conclusions about what specific aspects of the polymicrobial environment cause WT P. aeruginosa to become more susceptible. The authors make a compelling case that increased phenazine production by the ΔlasR mutant restores tolerance in the mixed community and that exogenous phenazine addition increases the survival of WT P. aeruginosa in the mixed community. However, it remains a plausible explanation that the effects of phenazines on tobramycin susceptibility are independent of the initial observation that WT. P. aeruginosa becomes susceptible to tobramycin in the mixed community.

   We agree with the reviewer’s comment here as it pertains to the initial observation of P. aeruginosa becoming more susceptible to tobramycin in the mixed community. However, as mentioned by the reviewer, we provide several lines of evidence that phenazines play a key role in the tolerance of the lasR mutant tobramycin, including genetic studies and feeding studies wherein exogenous addition of this molecule to WT P. aeruginosa phenocopies the lasR mutant exposed to tobramycin. Why the community impacts phenazine production of the WT strain is an open question, and the subject of future work. We have modified the abstract of the manuscript as follows at Lines 41–43:

   “Our data suggest that the molecular basis of this community-specific recalcitrance to tobramycin for the P. aeruginosa LasR mutant is increased production of phenazines.”

   Some aspects of the methodology are unclear. Specifically, the authors note that they use a specific sealed container system to grow their strains in anoxic conditions, which mimic portions of CF sputum. However, it is unclear how the authors change medium over the course of their experiments, or how they test susceptibility to tobramycin, without exposing the cells to oxygen. It is well understood that oxygen exposure impacts the susceptibility of P. aeruginosa to tobramycin, so it is very important that the methodology involving oxygen deprivation and exposure is described in detail.

   We have made the necessary modifications to the manuscript as indicated in the “Essential Revisions” section to address these concerns (see Comment #3). Furthermore, new validation experiments were performed in a controlled anoxic environmental chamber that yielded observations similar to the data presented in the original manuscript, thereby confirming that we were using anoxic conditions with the GasPak anaerobic jar system (see Figure 1 - figure supplement 2 and Figure 2 - figure supplement 7).

   Lines 198–204: “The impact of residual oxygen negatively influencing the growth of P. melaninogenica in monoculture was ruled out by performing these experiments using an anoxic environmental chamber (Figure 1 – figure supplement 2). That is, we did not detect CFU counts for either planktonic or biofilm populations of P. melaninogenica when grown in ASM in the anaerobic chamber, but as a positive control, significant growth was detected when using a medium shown previously to support growth of this microbe (10) (Prevotella Growth Medium, or PGM) (Figure 1 – figure supplement 2).”

   Lines 406–414: “Also, we ruled out the possibility of remaining oxygen in ASM negatively impacting the viability of P. melaninogenica by reproducing our results using an anoxic chamber (Figure 1 – figure supplement 2). That is, we observed that P. melaninogenica can robustly grow as a planktonic or biofilm monospecies community in a medium capable of sustaining its growth (PGM) while this microbe fails to grow in ASM (Figure 1 – figure supplement 2). Thus, we argue that the mixed-community-specific growth of Prevotella spp. we observed across several conditions (Figure 1C, Figure 1 – figure supplement 5, Figure 2 – figure supplement 6) is not due to residual oxygen.”

   Lines 290–293: “Growing and replenishing the preformed biofilm communities with fresh ASM supplemented or not with tobramycin using an anoxic environmental chamber resulted in similar phenotypes for all tested microorganisms (Figure 2 – figure supplement 7), indicating that the use of the GasPak system provides a robust anoxic environment.”

   Lines 533–540: “Plates were incubated using an AnaeroPak-Anaerobic container with a GasPak sachet (ThermoFisher) at 37 °C for 24 hours. Then, unattached cells were aspirated with a multichannel pipette and the pre-formed biofilms replenished with 100 µl of fresh ASM on the bench and incubated for an additional 24 hours at 37 °C using an AnaeroPak-Anaerobic container with a GasPak sachet (ThermoFisher). Similar experiments were performed using an anoxic environmental chamber (Whitley A55 - Don Whitley Scientific, Victoria Works, UK) with 10% CO2, 10% H2, 80% N2 mixed gas at 37 °C, yielding results identical to those observed for the GasPak system.”

   Reviewer #3 (Public Review) :

   This manuscript by Jean-Pierre et al. describes the creation and experimentation with a model CF lung community in an artificial sputum medium. The group uses data from 16S rRNA sequencing studies to select organisms for creating the model and then performs experiments to determine outcomes of growth competition and antibiotic tolerance in a community context. The main finding of the manuscript is that P. aeruginosa, notorious for its antimicrobial resistance phenotypes, is more susceptible to tobramycin in the community context than when grown alone. The manuscript is well prepared and follow-up experiments with mutant strains and phenazines greatly strengthen the project overall. The initial results paragraph where the authors go through the rationale for selecting the different organisms is perhaps a bit overkill, the organisms selected make sense based on their prevalence in CF airways, which in and of itself is a strong enough rationale. This aspect of the manuscript could be minimized to focus more on the exciting culture experiments in the latter parts of the results. Overall, this is a strong and well-crafted manuscript that will have a broad interest in the CF and microbial ecology fields.

   We thank the reviewer for this thoughtful review of our manuscript. We have not minimized the “front-end” of the paper because we believe the rationale for selecting the community and its members, and the validation of the model system are key for placing the resulting observations in a robust context, and for providing the underlying rationale to support the relevance of the findings.

   Major Critiques. I have two major critiques of this study.

   (1) Prevotella growth in monoculture. After reading the methods section it appears that the cultures were extensively washed and prepped prior to the inoculation into ASM. Prevotella did not grow alone, is this due to oxygen penetration of the cells during preparation? Perhaps oxygen is present in ASM prior to placement in an anaerobic bag? It is interesting, and perhaps worth exploring, whether the mixed community draws down oxygen from the media explaining the ability of Prevotella to grow. I suspect this is the case, but more detail is needed in the methods and this experiment would help us understand this interesting result.

   As presented in the “Essential Revisions” section (Comment #3), we have repeated the experiment using fully anoxic conditions (i.e., using an anoxic environmental chamber where the cultures were grown, washed and mixed before incubation) and still observed absence of growth of Prevotella cultivated in ASM in both biofilm and planktonic populations. Moreover, including a positive control, Prevotella Growth Medium, resulted in robust growth of this microbe. Taken together, our data suggest that residual oxygen in ASM is not the driver of the community-specific growth of P. melaninogenica.

   (2) Dilution of the community reproducing toby tolerance of P. aeruginosa. In supplemental figures, the replication of the 1:1000 dilution of the mixed community with P. aeruginosa shows poor replication and very large error bars. This experiment should be repeated to ensure it is reproducible.

   The diluted mixed community experiment was repeated a fourth time, yielding the same statistical conclusions. An updated “Figure 2 – figure supplement 1” was added to the paper. The highest (1:1000) dilution still yielded high variation which could perhaps be explained by low (i.e., ~103 CFU/mL) inoculum for S. aureus, S. sanguinis and P. melaninogenica used in these experiments; see updated “Microbial assays” paragraph of the “Materials and Methods” section). Thus, the variation at low inoculum is robust and reproducible. The Materials and Methods section was also updated to clarify the CFU counts used for those experiments. We have added modifications to the text as follows to address this critique:

   Lines 526–532: “The optical density (OD600) was then measured for each bacterial suspension and diluted to an OD600 of 0.2 in ASM. Monocultures and co-culture conditions were prepared from the OD600 = 0.2 suspension and diluted to a final OD600 of 0.01 for each microbial species in ASM corresponding to final bacterial concentrations of 1x107 CFU/mL, 3.5x106 CFU/mL, 1.2x106 CFU/mL and 4.6x106 CFU/mL of P. aeruginosa, S. aureus, Streptococcus spp. and Prevotella spp. respectively. A volume of 100 µl of bacterial suspension all at a final OD600 of 0.01 each in the mix was added to three wells.”

   Lines 558–570: “For experiments with varying concentrations of S. aureus, S. sanguinis and P. melaninogenica in monocultures and co-cultures, the organisms were grown from bacterial suspensions adjusted to an OD600 = 0.8 in ASM. Suspensions were further diluted in ASM to an OD600 of either 0.1, 0.001, 0.0001 or 0.00001 while maintaining P. aeruginosa at OD600 = 0.01 (approximating 1x107 CFU/mL) in all conditions. The OD600 = 0.1 dilution factor resulted in CFU/mL count average of 3.8x108 CFU/mL for S. aureus, 1.6x108 CFU/mL for S. sanguinis and 1.0x108 CFU/mL for P. melaninogenica. The OD600 = 0.001 dilution factor resulted in a CFU/mL count average of 6.7x105 CFU/mL for S. aureus, 1.1x105 CFU/mL for S. sanguinis and 1.4x105 CFU/mL for P. melaninogenica. The OD600 = 0.0001 dilution factor resulted in a CFU/mL count average of 4.2x104 CFU/mL for S. aureus, 3.3x104 CFU/mL for S. sanguinis and 4.6x104 CFU/mL for P. melaninogenica. The OD600 = 0.00001 dilution factor resulted in a CFU/mL count average of 5.6x103 CFU/mL for S. aureus, 4.4x103 CFU/mL for S. sanguinis and 6.2x103 CFU/mL for P. melaninogenica.”

   Read the original source
3. 

   eLife

   Nov 15, 2022

   eLife assessment

   This article establishes a model experimental bacterial community to represent the microbiome found in ~1/3 of patients with cystic fibrosis (CF) with the goal of understanding why these infections do not respond to treatments that are effective in single-species infections. The authors show that susceptibility to the most common antibiotic used against the dominant pathogen P. aeruginosa is different when grown in this mixed community, and a mutant of this pathogen (lasR) that frequently occurs during infections alters this sensitivity. This study is significant for producing an experimental resource for the microbiology of CF, and it could be strengthened by more detailed measures of interactions between species and how the phenotypes produced by lasR alter species interactions.

   Read the original source
4. 

   eLife

   Nov 15, 2022

   Reviewer #1 (Public Review):

   The authors set out to develop an in vitro model of multiple species representing diversity in the CF airway as a platform for a range of studies on why polymicrobial communities resist therapy. The rationale for their design is sound and the methods appear justifiable and reproducible. The major strength of this work is in producing a method for a range of future work, ideally for multiple groups in the field. The primary findings are interesting but not groundbreaking. One weakness in the method of reporting interspecies interactions and another in evaluating alternative causes of lasR advantages present opportunities for a stronger research contribution beyond this terrific method.

   Read the original source
5. 

   eLife

   Nov 15, 2022

   Reviewer #2 (Public Review):

   Differences between the infection environment and in vitro model systems likely contribute to disconnects between the antimicrobial susceptibility profile of bacterial isolates and the clinical response of patients. The authors of this paper focus on a specific aspect of the infection environment, the polymicrobial nature of some chronic infections like those in people with Cystic Fibrosis (CF), as a factor that could impact antibiotic tolerance. They first use published genomic datasets and computational techniques to identify a clinically relevant, four-member polymicrobial community composed of Pseudomonas aeruginosa, Staphylococcus aureus, Streptococcus spp., and Prevotella spp. They then develop a high throughput methodology in which this community grows and persists in a CF-like environment and in …

   Reviewer #2 (Public Review):

   Differences between the infection environment and in vitro model systems likely contribute to disconnects between the antimicrobial susceptibility profile of bacterial isolates and the clinical response of patients. The authors of this paper focus on a specific aspect of the infection environment, the polymicrobial nature of some chronic infections like those in people with Cystic Fibrosis (CF), as a factor that could impact antibiotic tolerance. They first use published genomic datasets and computational techniques to identify a clinically relevant, four-member polymicrobial community composed of Pseudomonas aeruginosa, Staphylococcus aureus, Streptococcus spp., and Prevotella spp. They then develop a high throughput methodology in which this community grows and persists in a CF-like environment and in which antibiotic susceptibility can be tested. The authors determine that living as a member of this community decreases the antibiotic tolerance of some strains of biofilm-associated P. aeruginosa and increases the tolerance of most strains of planktonic and biofilm-associated S. aureus and planktonic and biofilm-associated Streptococcus. They focus on the decreased tolerance of P. aeruginosa and determine that a ΔlasR mutant of P. aeruginosa does not display increased tobramycin susceptibility in the mixed community. One of the phenotypes associated with a ΔlasR mutant is an overproduction of phenazines. The authors find that by deleting the phenazine biosynthesis genes from ΔlasR, they can restore community-acquired susceptibility. They further investigate this phenomenon by showing that a specific type of phenazine, PCA, is significantly increased in mixed communities with the ΔlasR mutant compared to WT. Finally, they demonstrate that adding a specific phenazine, pyocyanin, to mixed communities can restore the tolerance of WT P. aeruginosa.

   Strengths:

   With this study the authors address a very important problem in infectious disease microbiology - our in vitro drug susceptibility assays do a poor job of mimicking the infection environment and therefore do a poor job of predicting how effective particular drugs will be for a particular patient. By demonstrating how an infection-relevant community modifies tolerance to a clinically relevant drug, tobramycin, the authors identify specific interactions that could be targeted with therapeutics to improve our ability to treat the chronic infections associated with CF. In addition, this study provides a framework for how to effectively model polymicrobial infections in vitro.

   The experiments in the paper are very rigorous and well-controlled. Statistical analysis is appropriate. The paper is very well-written and clear.

   The authors do an admirable job of using in silico analysis to inform their in vitro studies. Specifically, they provide a comprehensive rationale for why they chose and studied the specific community they did.

   The authors provide a very robust dataset which includes determining how strain differences of each of their four community members affect community dynamics and antibiotic tolerance. These types of analyses are laborious but very important for understanding how broadly applicable any given result is.

   Weaknesses:

   The authors very clearly and convincingly demonstrate that WT P. aeruginosa becomes more susceptible to tobramycin in their mixed community. Our ability to turn these types of observations into therapeutic development depends on mechanistic insight. That said, it is unclear if the authors can make any solid conclusions about what specific aspects of the polymicrobial environment cause WT P. aeruginosa to become more susceptible. The authors make a compelling case that increased phenazine production by the ΔlasR mutant restores tolerance in the mixed community and that exogenous phenazine addition increases the survival of WT P. aeruginosa in the mixed community. However, it remains a plausible explanation that the effects of phenazines on tobramycin susceptibility are independent of the initial observation that WT. P. aeruginosa becomes susceptible to tobramycin in the mixed community.

   Some aspects of the methodology are unclear. Specifically, the authors note that they use a specific sealed container system to grow their strains in anoxic conditions, which mimic portions of CF sputum. However, it is unclear how the authors change medium over the course of their experiments, or how they test susceptibility to tobramycin, without exposing the cells to oxygen. It is well understood that oxygen exposure impacts the susceptibility of P. aeruginosa to tobramycin, so it is very important that the methodology involving oxygen deprivation and exposure is described in detail.

   Read the original source
6. 

   eLife

   Nov 15, 2022

   Reviewer #3 (Public Review):

   This manuscript by Jean-Pierre et al. describes the creation and experimentation with a model CF lung community in an artificial sputum medium. The group uses data from 16S rRNA sequencing studies to select organisms for creating the model and then performs experiments to determine outcomes of growth competition and antibiotic tolerance in a community context. The main finding of the manuscript is that P. aeruginosa, notorious for its antimicrobial resistance phenotypes, is more susceptible to tobramycin in the community context than when grown alone. The manuscript is well prepared and follow-up experiments with mutant strains and phenazines greatly strengthen the project overall. The initial results paragraph where the authors go through the rationale for selecting the different organisms is perhaps a bit …

   Reviewer #3 (Public Review):

   This manuscript by Jean-Pierre et al. describes the creation and experimentation with a model CF lung community in an artificial sputum medium. The group uses data from 16S rRNA sequencing studies to select organisms for creating the model and then performs experiments to determine outcomes of growth competition and antibiotic tolerance in a community context. The main finding of the manuscript is that P. aeruginosa, notorious for its antimicrobial resistance phenotypes, is more susceptible to tobramycin in the community context than when grown alone. The manuscript is well prepared and follow-up experiments with mutant strains and phenazines greatly strengthen the project overall. The initial results paragraph where the authors go through the rationale for selecting the different organisms is perhaps a bit overkill, the organisms selected make sense based on their prevalence in CF airways, which in and of itself is a strong enough rationale. This aspect of the manuscript could be minimized to focus more on the exciting culture experiments in the latter parts of the results. Overall, this is a strong and well-crafted manuscript that will have a broad interest in the CF and microbial ecology fields.

   Major Critiques:

   I have two major critiques of this study.

   1. Prevotella growth in monoculture. After reading the methods section it appears that the cultures were extensively washed and prepped prior to the inoculation into ASM. Prevotella did not grow alone, is this due to oxygen penetration of the cells during preparation? Perhaps oxygen is present in ASM prior to placement in an anaerobic bag? It is interesting, and perhaps worth exploring, whether the mixed community draws down oxygen from the media explaining the ability of Prevotella to grow. I suspect this is the case, but more detail is needed in the methods and this experiment would help us understand this interesting result.

   2. Dilution of the community reproducing toby tolerance of P. aeruginosa. In supplemental figures, the replication of the 1:1000 dilution of the mixed community with P. aeruginosa shows poor replication and very large error bars. This experiment should be repeated to ensure it is reproducible.

   Read the original source
7. 

   Version published to 10.1101/2022.06.23.497319v1 on bioRxiv

   Jun 24, 2022