Heterogeneity Governs 3D‐Cultures of Clinically Relevant Microbial Communities

The intrinsic heterogeneity of bacterial niches should be retained in in vitro cultures to represent the complex microbial ecology. As a case study, mucin‐containing hydrogels ‐CF‐Mu3Gel ‐ are generated by diffusion‐induced gelation, bioinspired on cystic fibrosis (CF) mucus, and a microbial niche challenging current therapeutic strategies. At breathing frequency, CF‐Mu3Gel exhibits a G′ and G″ equal to 24 and 3.2 Pa, respectively. Notably, CF‐Mu3Gel exhibits structural gradients with a gradual reduction of oxygen tension across its thickness (280–194 µmol L−1). Over the culture period, a steep decline in oxygen concentration occurs just a few millimeters below the air–mucus interface in CF‐Mu3Gel, similar to those of CF airway mucus. Importantly, the distinctive features of CF‐Mu3Gel significantly influence bacterial organization and antimicrobial tolerance in mono‐ and co‐cultures of Staphylococcus aureus and Pseudomonas aeruginosa that standard cultures are unable to emulate. The antimicrobial susceptibility determined in CF‐Mu3Gel corroborates the mismatch on the efficacy of antimicrobial treatment between planktonically cultured bacteria and those in patients. With this example‐based research, new light is shed on the understanding of how the substrate influences microbial behavior, paving the way for improved fundamental microbiology studies and more effective drug testing and development.


Introduction
Structural, rheological, and compositional heterogeneity is a pervasive feature of most bacteria-hosting habitats.It has a significant impact on microbial organization, metabolic activity, and adopted pathways, and is further linked with energy consumption, gene expression and biomass production. [1]Within these complex 3D environments, individual bacteria are subjected to time-varying spatial gradients of nutrients, metabolites and oxygen (O 2 ) tension, all of which act as the driving forces of chemotaxis and aerotaxis.In multicellular communities, collective microbial dynamics contribute further to the heterogeneity of the environmental niche, generating a complex ecosystem characterized by both microbe-microbe and microbeenvironment interaction dynamics.These features must somehow be rendered in in vitro models, and they are not reproducible with cultures in intrinsically homogeneous liquid media (planktonic conditions).
Cystic fibrosis (CF) mucus is a challenging example of a heterogeneous microbial ecosystem.The presence of Staphylococcus aureus and Pseudomonas aeruginosa in CF mucus is generally associated with diminished lung function and more rapid pulmonary decline, but there is no consensus as to whether these bacteria are antagonistic or if S. aureus and P. aeruginosa can coexist in a complex microenvironment.Some studies indicate that early colonization by P. aeruginosa shows strong antagonism toward S. aureus, [2,3] through the secretion of a variety of anti-staphylococcal molecules and proteases that inhibit S. aureus growth and proliferation. [4,5]This process induces a metabolic transition of S. aureus from aerobic respiration to fermentation and eventually leads to a loss of S. aureus viability. [6]In response to this hostile environment, S. aureus may adapt to P. aeruginosa exoproducts by increasing biofilm formation and the presence of small colony variants. [5,7]Yet, other studies show that S. aureus supports colonization and pathogenicity of P. aeruginosa, [8] and that anaerobiosis is required for their coexistence. [9]he need for novel models is imperative to address the contribution of the environment in bacterial cultures and co-cultures, particularly to enhance knowledge concerning mechanisms of bacterial colonization and how bacteria-bacteria and bacteriaenvironment interaction dynamics unfold in complex microbial ecosystems.Such models can help screen existing antimicrobial agents and support the development of new, patient-specific molecules, which are possibly closer to the actual clinical infections context, tackling infections occurring in mucus. [10]In vitro models exhibiting the key characteristics of the mucus environment, which, simultaneously, enable high throughput pharmacokinetics studies, would allow for a more effective screening of the most promising compounds early on in the drug discovery phase.These will additionally reduce the number of animals used during pre-clinical analyses. [11]xisting in vitro models range from traditional methods, such as microtiter plates and dynamic systems, [12] to 3D printing of bacteria. [13,14]Co-cultures of S. aureus and P. aeruginosa have been performed using modified media in static [15] or dynamic conditions, [9,[16][17][18][19] anoxia, microtiter plates, [20,21] Calgary biofilm device, [22] or even directly over mammalian cells. [6,23]The coculture of these bacteria is of great importance, not only because of their mutual impact on behavior and metabolic activity, but also because some studies have shown that their interplay contributes to antimicrobial tolerance. [18,19,24,25]The challenge, however, is that existing models, despite shining some light on the behavior of these bacteria in 3D conditions, lack the heterogeneous environment that bacteria experience in the CF airway niche.Furthermore, it is important to acknowledge the significant role of mucus, as it contains a wide range of molecules with antimicrobial activity such as peptides, enzymes, antibodies, and various types of mucins. [26]The latter have a direct impact on bacterial behavior and their susceptibility to antimicrobial agents through the regulation of associated genes.Notably, mucins play a crucial role in the neutralization or elimination of invasive pathogenic bacteria, while also exhibiting immunomodulatory properties. [27,28]To provide a high throughput in vitro model that ensures interlaboratory reproducibility, it is important to rely on commercial reagents that are readily available in considerable amounts to offer consistent results.Now, there are mostly two sources of commercial mucins available in quantities that are compatible with these requirements, porcine stomach type III and mucin from bovine submaxillary glands.Mucin-containing structures composed of porcine gastric mucins have been recently proposed to culture microorganisms, to either culture intestinal microbiota [29] or co-culture P. aeruginosa and Staphylococcus flexneri. [30]n this paper, we exploit a new approach based on a controlled diffusion of a crosslinking agent to generate a 3D structure that has the advantage of retaining spatial gradients, permitting microscopic-restructuring dynamics of the microstructure, and allowing for spatial control of O 2 content.This enabled the development of an in vitro model bioinspired on the microbiological environment of the CF airway mucus, supporting the study of the bacterial interplay within a heterogeneous microbial ecosystem.Gastric mucin was selected as mucin source, as this is composed of MUC5AC and MUC5B, [31] which have been reported to be also the main components of CF airway mucus. [32] Results

Extracting the Key Features of Mucus to Reproduce a Heterogeneous Bacterial Environment
CF mucus models, CF-Mu 3 Gel, were synthesized by slow calcium (Ca 2+ ) permeation through a membrane to slowly produce a 3D heterogeneous structure.CF-Mu 3 Gel is predominantly composed of mucin and sodium chloride, whose ranges are within those reported in the literature for CF sputum. [32,33]CF-Mu 3 Gel was produced using two common bacterial culture media, Müller Hinton broth (MHB) or Luria Bertani broth (LB).This reproduces the complex 3D mucus environment while retaining the fundamental composition of the culture media.The viscoelastic properties of the resultant CF-Mu 3 Gel hydrogels were investigated through rheological characterization at the characteristic frequencies of both breathing (0.5 Hz) and ciliary beating (≈10 Hz) [34][35][36] and further compared to those reported for CF mucus. [37]Both storage (G′) and dissipative (G″) components of the human mucus complex modulus impact its clearance.The capacity of the mucus to elastically store energy from the ciliary beating, quantified by G′, allows the mucus to slide using this energy upon recoil of the cilia.Further, independently of the medium in which CF-Mu 3 Gel was produced, the obtained G′ was always higher than G″ confirming their gel-like structure, which prevents the mucus from flowing down the airway due to gravity forces.Herein, the viscoelastic properties of the hydrogels were not strongly affected by the different culture media, although CF-Mu 3 Gel produced in MHB displayed closer characteristics to those of CF mucus reported in the literature  3 Gel was analyzed at a, breathing (≈0.5 Hz); and ciliary beating frequency (10 Hz; n ≥ 6) (b).The rheological data were analyzed using the two-way ANOVA.Significant differences were set for *p < 0.05.A minimum of six independent samples were analyzed per formulation (n ≥ 6).ns means no statistical differences were found between the viscoelastic properties of the different formulations of CF-Mu 3 Gel and those of CF sputum reported by Yuan et al. (2015). [37]c) Dependency of mesh size () of CF-Mu 3 Gel on media in which this was produced estimated by combining the Generalized Maxwell Model with the rubber elasticity theory (n ≥ 6).d) Macroscopic image of CF-Mu 3 Gel produced with 0.16% (w/v) Ca 2+ ions.e) Correlation (top) and intensity (bottom) maps c i of a region of ≈2.83 × 2.66 mm, obtained after 20 h of crosslinking reaction for a CF-Mu 3 Gel.The correlation index is evaluated for a delay of 60 s and the intensity is normalized to its value at the top.f) Schematic representation of both topographical and O 2 gradients exhibited through the 3D structure of CF-Mu 3 Gel.g) Heterogeneous distribution of O 2 tension through the structure of sterile CF-Mu 3 Gel (n = 5).
(Figure 1a,b). [37]As CF mucus characteristics vary, not only between patients, but also within a single patient in a short period of time, [38][39][40][41][42] the concentration of Ca 2+ ions was varied to produce CF-Mu 3 Gel with viscoelastic properties (Figure S1a,b, Supporting Information) tailorable to the desired stiffness according to disease progression [43] or even patient-specific to support personalized pharmacological treatment.The average mesh size, estimated for all tested CF-Mu 3 Gel formulations, decreased with increased concentration of Ca 2+ ions (Figure S1c, Supporting Information).However, the CF-Mu 3 Gel produced in MHB exhibited a closer mesh size to that reported for CF sputum than those produced in LB and NaCl (Figure 1c). [38]As a result, all further analyses were therefore performed on CF-Mu 3 Gel produced in MHB (Figure 1d), as this maximized the representation of CF mucus structural properties.
Gradients have a tremendous impact on microbial behavior and though they are present in most of the human mucus, the methods to characterize them are scarce.Photon Correlation Imaging (PCI) combines dynamic light scattering with digital imaging and allows to study slow and spatially limited spontaneous restructuring dynamics of soft disordered solids.It was previously used to study the kinetic of the formation of hydrogels. [44]In this case, this unconventional technique was adapted to study structural gradients evaluating the different microscopic dynamics of the motion of the molecular chains result-ing from the different crosslinking density (Figure S2, Supporting Information).PCI enabled the visualization of both gradients in the structure and microscopic dynamics of the CF-Mu 3 Gel to be detected via scattered intensity and correlation maps, respectively.The correlation maps evidence that diffusion-controlled gelation produced hydrogels which were stiffer and almost dynamically arrested in the regions first exposed to Ca 2+ ions (orange region, top map in Figure 1e, Figure S2, Supporting Information).The zones far from the diffusion inlet are more tenuous and display high degrees of spontaneous restructuring dynamics (green regions).Besides, CF-Mu 3 Gel retains a gradual reduction of O 2 tension from top to bottom, decreasing from 280 to 194 μmol L −1 concerning y = 0-2300 μm, respectively (Figure 1g).In this way, PCI and microsensing techniques enabled the characterization of both topographical and O 2 gradients, respectively, throughout the 3D structure of CF-Mu 3 Gel (Figure 1f).

Structural and O 2 Gradients Affect the Topographical Organization of S. aureus and P. aeruginosa
CF-Mu 3 Gel kept its integrity up to the end of the test (48 h) when cultured with S. aureus and P. aeruginosa.After 72 h of incubation in different media, sterile CF-Mu 3 Gel retained its dimensions The results obtained with CF-Mu 3 Gel (green) were further compared with planktonic cultures (white).In (a) and (b), the shadowed region depicts the range of CFU mL −1 of each bacterium reported in the literature for pathological CF mucus. [45]A minimum of five independent samples were analyzed per bacterium (n > 5).No statistical differences were detected neither between both planktonic cultures and CF-Mu 3 Gel nor between the same culturing system at different culturing periods.c) Experimental setup adopted to measure oxygen (O 2 ) tension through CF-Mu 3 Gel colonized with either d) S. aureus or e) P. aeruginosa for 12 (green), 24 (blue), and 48 h (red) (n = 5).
with variations of 20% (weight) and 40% (thickness) on average when incubated in both isotonic and hypertonic media, and with variations of 44% (weight) and 200% (thickness) when incubated in hypotonic medium alone (Figure S1d,e, Supporting Information).
Mono-cultures of S. aureus and P. aeruginosa in CF-Mu 3 Gel and planktonic conditions reached the stationary phase at 24 h with counts of ≈10 9 CFU mL −1 (Figure 2a,b).It is notable that an interlaboratory study showed a similar number for both bacteria after 24 h of culture, demonstrating that the results obtained in CF-Mu 3 Gel are laboratory-and operator-independent (Figure S3, Supporting Information).
O 2 profiles across CF-Mu 3 Gel indicated that, independently of the microorganism, O 2 content decreased in relation to the depth and extent to which the CF-Mu 3 Gel was colonized, even after 12 h of culture (Figure 2d,e).This decrease was more pronounced for longer periods of incubation, and O 2 tension was progressively reduced to a completely anoxic zone, at ≈300 μm of depth, after 48 h of culture (Figure 2d,e).
Qualitative analyses of bacterial organization indicated that during the culture period, the number of S. aureus within CF-Mu 3 Gel increased and exhibited depth-dependent differences in growth and in the presence of bacterial aggregates, these were visible after 24 h of culture (Figure 2a; Figure S4, Supporting Information) and increased in size with the progression of the time of culture (Figure S4, Supporting Information).Regardless of the observed position, a longer culture time resulted in a higher number of P. aeruginosa within the CF-Mu 3 Gel (Figure 2b, Figure S5, Supporting Information).S. aureus and P. aeruginosa were able to migrate and colonize the whole thickness of CF-Mu 3 Gel, with a higher number of bacteria at the top and bottom of CF-Mu 3 Gel.

CF-Mu 3 Gel Governs the Permeability of Antimicrobial Agents
CF-Mu 3 Gel was coupled with either Transwell supports or PAMPA (parallel artificial membrane permeability assays) membranes, widely used during permeability assessment [46] to evaluate the role of CF-Mu 3 Gel on antimicrobial retention and consequent antimicrobial susceptibility.CF-Mu 3 Gel strongly hindered the passage of the three antimicrobial agents (Figure 3), independently of the supporting system.The differences between Transwell supports and PAMPA were observable but were irrelevant when the effect of CF-Mu 3 Gel is considered (Figure 3, Figure S7, Supporting Information).

Bacteria in CF-Mu 3 Gel Resist Antimicrobial Treatment
The Minimal Inhibitory Concentration (MIC) of different antimicrobial agents (ciprofloxacin, tobramycin and colistin), which are commonly administered to CF patients with S. aureus and/or P. aeruginosa infections was assessed according to The European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines (Figure S8, Supporting Information).The determined MIC values were the starting point to further treat bacteria incubated in CF-Mu 3 Gel for 24 h.A significant antimicrobial effect was observed when S. aureus was cultured under planktonic conditions (four-log 10 and six-log 10 decline of colonies with 1 and 10 MIC, respectively; Figure 4a,b).The antimicrobial treatment was less effective toward S. aureus cultured in CF-Mu 3 Gel at short treatment times (two-log 10 and four-log 10 decline for 1 and 10 MIC, respectively; Figure 4a).When bacteria were cultured for extended periods, to simulate established infection conditions, prior to antimicrobial treatment, the number of S. aureus colonies in CF-Mu 3 Gel remained constant, even when a high concentration of ciprofloxacin was used (10 MIC; Figure 4b).In contrast, a more pronounced antimicrobial effect was observed in S. aureus cultured in planktonic conditions, with a decline in colonies of twolog 10 and three-log 10 from 1 and 10 MIC, respectively, when compared with bacteria grown without antimicrobial treatment (Figure 4a).
P. aeruginosa was more resistant to antimicrobial treatment when cultured within CF-Mu 3 Gel than in liquid medium (Figure 4c-e).In fact, in planktonic cultures, the number of CFU decreased progressively with increasing concentrations of both ciprofloxacin and tobramycin, resulting in a steep reduction of six-log 10 and nine-log 10 after 24 h of treatment with 1 and 10 MIC, respectively, when compared with the untreated control group (Figure 4c,d).In contrast, the concentration of 10 MIC tobramycin was found to be three times less effective against P. aeruginosa in CF-Mu 3 Gel than in planktonic conditions (Figure 4d).Additionally, colistin treatment was found to be ineffective against bacteria grown within CF-Mu 3 Gel, even at high concentrations, as it was not able to eradicate P. aeruginosa (Figure 4d), while bacteria cultured under planktonic conditions were only susceptible to 10 MIC of colistin.

S. aureus and P. aeruginosa Co-Exist and Compete in CF-Mu 3 Gel in Different Culture Setups
Microbial ecosystems composed of S. aureus and P. aeruginosa coexisted in vitro in CF-Mu 3 Gel, influencing each other's growth (Figure 5a-c) and topographical organization (Figure 5d-f), independent of the culture method used.As the timing of colonization may affect the way they interact, three different types of co-culture were simulated: (1) a contemporary co-culture of both S. aureus and P. aeruginosa for 24 h (Figure 5a,d); (2) a 24 h culture with S. aureus first, followed by a P. aeruginosa culture for another 24 h of incubation (Figure 5b,e); and finally (3) a 24 h culture with P. aeruginosa first, followed by S. aureus culture for another 24 h of incubation (Figure 5c,f).Contemporary co-culture provides both pathogens with equal conditions for them to thrive (Figure 5a).After 24 h of co-culture, both bacteria were present in CF-Mu 3 Gel (Figure 5a), and these were able to colonize the 3D structure (Figure 5d).This was not observed in planktonic cultures, as S. aureus was not able to grow.
To reproduce CF airways microbial colonization timeline [48] the experimental setting (2) was used (S.aureus first).After 48 h of culture, both S. aureus and P. aeruginosa co-existed either in CF-Mu 3 Gel or planktonic conditions (Figure 5b).The relative abundance of S. aureus in respect to P. aeruginosa was prevalent in planktonic conditions (Figure 5b).
When P. aeruginosa was first cultured in CF-Mu 3 Gel, experimental setting (3), both S. aureus and P. aeruginosa co-existed within CF-Mu 3 Gel, though the relative abundance of S. aureus was minor (Figure 5c).Under planktonic conditions, P. aeruginosa outcompeted S. aureus (Figure 5c).
The topographical organization in co-cultures indicated that both S. aureus and P. aeruginosa formed bacterial aggregates (Figure 5d-f).This was also observed in mono-cultures of S. aureus and P. aeruginosa within CF-Mu 3 Gel (Figures S4-S6, Supporting Information).However, the aggregates of P. aeruginosa  3 Gel colonized with P. aeruginosa for 24 h followed by treatment with: c) ciprofloxacin; d) tobramycin; and e) colistin.In the case of ciprofloxacin, 0.1, 1, and 10 MIC correspond to 0.05, 0.5, and 5 mg L −1 , respectively, while for both tobramycin and colistin 0.1, 1, and 10 MIC correspond to 0.8, 8, and 80 mg L −1 , respectively.No ATB means that no antimicrobial treatment was performed.Planktonic cultures of S. aureus and P. aeruginosa were used as controls of the experiment.0 CFU mL −1 indicates that no CFU were detected.All data were analyzed using the one-way ANOVA.Significant differences were set for *p < 0.05; and **p < 0.01.A minimum of five independent samples were analyzed per formulation (n ≥ 5).
produced in experimental setting (3) (Figure 5f) were larger than those observed in experimental setting (2) (Figure 5e), possibly due to the longer culture time of P. aeruginosa.
When S. aureus was cultured first, mimicking the microbial colonization timeline of CF airways, [38] experimental setting (2), the co-existence of the two bacteria species impacted their susceptibility to ciprofloxacin treatment.In CF-Mu 3 Gel, no differences were observed in the number of either bacteria species in the co-cultures that were treated with the tested ciprofloxacin concentrations (Figure 6b,d).Interestingly, the bacterial survival rates in CF-Mu 3 Gel of both S. aureus and P. aeruginosa after ciprofloxacin treatment were higher when in co-cultures than they were in mono-cultures.The opposite was found to be the case in S. aureus co-cultured in planktonic conditions (Figure 6a).

Discussion
By replicating key features of the microbial ecology of a specific environment, a 3D-substrate able to support the growth of mono-and poly-microbial communities was developed.The challenging environment of CF mucus airway was taken as in-spiration due to its complex ecology.The key features governing the microbial behavior were identified in the 3D-structural gradients, physical and chemical composition, O 2 distribution, and its ability to function as a barrier to antimicrobial diffusion, antagonistic phenomena and bacteria susceptibility to antimicrobials.
The intention was to model the airway CF mucus by replicating the viscoelastic properties and mucin content that were previously reported. [33,37,49]The viscoelastic properties of CF-Mu 3 Gel increased with frequency (Figure 1a,b).This has been linked to limited mucus clearance creating an environment for infections to thrive. [3,24,50]The production of gradients within hydrogels is not a trivial issue, and so is their characterization. [51]Specifically challenging to study are the gradients at the molecular level, which are crucial for the organization of both microbial niches and biological tissues.Here, we employed PCI to gain information about the molecular structure of the produced hydrogels, and it was possible to evaluate the change in restructuring dynamics related to the gradient of crosslinking.The gradient structure of CF-Mu 3 Gel (Figure 1e,f) is similar to the structure of CF airway mucus, where stagnating mucins form a higher number of entanglements at the lower strata of the mucus than at the  3 Gel with both S. aureus and P. aeruginosa following three different co-culture settings: a,d-1) contemporary culture of both S. aureus (in light red) and P. aeruginosa (in light green) for 24 h; b,e-2) first culture with S. aureus for 24 h followed by P. aeruginosa culture for other 24 h; and finally, c,f-3) culture with P. aeruginosa first for 24 h, after which S. aureus is added and incubated for further 24 h.Bacteria numbers determined in CF-Mu 3 Gel (green) were further compared with planktonic cultures (white) (a-c).The shadowed region depicts the range of CFU mL −1 of each bacterium reported in the literature for pathological CF mucus. [45,47]All data were analyzed using two-way ANOVA.Significant differences were set for *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.A minimum of five independent samples were analyzed per formulation (n ≥ 5).The relative abundance of S. aureus in respect to P. aeruginosa per each experimental setting is depicted in the pie charts in the upper part of (a-c).Confocal microscopy analyses of CF-Mu 3 Gel co-colonized with both DsRed-fluorescent S. aureus (fluorescent red) and GFP-fluorescent P. aeruginosa (fluorescent green) (d-f).Co-colonized CF-Mu 3 Gel was further stained with Hoechst 33342 to depict total bacteria (in blue) and 3D reconstruction are also provided.The scale bar corresponds to 10 μm.
airway-mucus interface.The gelation, induced by diffusion of the crosslinking agent, results in reproducible gradients of structural stiffness, spontaneous restructuring power, and receptiveness to O 2 .
O 2 gradients within native microbial ecosystems are known to affect bacterial metabolic activity and influence their susceptibility to antimicrobial agents.The O 2 gradient is a unique feature of CF-Mu 3 Gel, with a lower O 2 tension at the bottom and a higher O 2 tension at the top that is linked to low crosslinking density (Figure 1f,g).This profile is related to the diffusion of O 2 in hydrogels, dependent on bound, free, and interfacial water content, mesh size, and crosslinking density. [52]Given that CF-Mu 3 Gel exhibits a gradient structure with a tighter mesh at the bottom than at the top (Figure 1e), higher water-to-polymer ratios are found at the top than at the bottom (Figure 1g).Consequently, higher levels of O 2 are available at the top for bacteria metabolism.
CF-Mu 3 Gel sustained the growth of S. aureus and P. aeruginosa (Figure 2a,b, Figure 4a-e "No ATB").In CF-Mu 3 Gel, the P. aeruginosa number was comparable to the numbers found in the lungs of CF patients (10 8 -10 10 bacteria mL −1 ), while the number of S. aureus was higher. [45]S. aureus aggregates were found to increase in number and size over time in CF-Mu 3 Gel (Figure 2a, Figure S4, Supporting Information).A similar number and type of aggregates of S. aureus formed in a more complex system containing agarose gels. [53]Though, agarose gels have a uniform structure and rely on bacteria O 2 consumption to produce the gradient. [53]Whenever bacteria are introduced in CF-Mu 3 Gel, they are able to sense, self-organize and exacerbate the O 2 profile (Figure 2d,e), with a higher number of bacteria at the top (Figures S4 and S5, Supporting Information).The heterogeneous structure of CF mucus is not only a natural consequent of mucus build-up, but microbial colonization further contributes to this phenomenon, as, during disease progression, airway CF mucus gradually acquires a heterogeneous structure with detectable O 2 depletion at a few millimeters below the air-mucus interface.Indeed, the resultant O 2 profile reproduces that determined in CF mucus colonized by S. aureus and P. aeruginosa. [54,55]gar or alginate 3D systems such as beads, [56][57][58][59][60] and, more recently, a hydrogel composed of alginate/mucin, [30] have been proposed as suitable means to culture P. aeruginosa.The latter is similar to our previous mucus models [61,62] and all of them lack the presence of 3D gradients.Indeed, our results show that bacteria can sense the intrinsic gradients of CF-Mu 3 Gel.In fact, P. aeruginosa forms multicellular aggregates, similar in morphology and size to those seen in CF mucus and chronic wounds (Figure 2b; Figures S5 and S6, Supporting Information). [63,64]The formation of these aggregates is associated with O 2 gradients, [54,[65][66][67] with higher concentration of aggregates at higher O 2 concentrations.This organization leads to increased production of alginate by P. aeruginosa, [66,68] a process which is believed to be the underlying cause of chronic infections. [69]This level of organization was also observed in vitro in alginate beads, in which P. aeruginosa aggregates were mostly present at the surface. [57,58]upplementation with NO 3 − promoted the formation of aggregates in deeper layers.In the CF-Mu 3 Gel, however, there was no need for NO 3 − supplementation to achieve deeper layer colonization and aggregate formation (Figure S5, Supporting Information).Steep O 2 gradients were present in CF-Mu 3 Gel and anoxia was found after 48 h of culture at 200-300 μm depth with a close similarity to airway CF mucus (Figure 2d,e). [54]hese features may be related to the presence of aggregates of both S. aureus and P. aeruginosa, suggesting possible biofilm formation. [70,71]icrobial ecology seems to be highly dependent on the presence of a 3D structure and on the culture period.19][20][21][22][72][73][74] Some studies suggest that P. aeruginosa outcompetes S. aureus, [2,3,6] others refer that S. aureus supports colonization and pathogenicity of P. aeruginosa [8] or that anaerobiosis is required for their coexistence. [9]In CF-Mu 3 Gel, both S. aureus and P. aeruginosa co-existed independently of the experimental condition adopted (Figure 5), while in planktonic conditions, these only co-existed when S. aureus was cultured first followed by the addition of P. aeruginosa (Figure 5b).Co-cultures under planktonic conditions, by definition, lack a 3D support, meaning that the exoproducts secreted by P. aeruginosa and other molecules freely and easily spread in the medium, thus negatively influencing the survival of S. aureus.Yet, within a 3D substrate, bacteria by-products slowly diffuse due to interactive and steric phenomena characteristic of 3D environments.This supports the results with CF-Mu 3 Gel (Figure 5a-c), where both S. aureus and P. aeruginosa co-exist with a similar number as in the case with CF mucus. [47]he presence of O 2 gradients may also be a key feature to enable simultaneous growth.It is described that, under anoxia, S. aureus growth did not change when co-cultured with either strains or clinical isolates of P. aeruginosa. [9]This is further corroborated by our results of mono-cultured S. aureus and P. aeruginosa that exacerbated the intrinsic O 2 gradients of CF-Mu 3 Gel with a steep decrease in O 2 tension throughout its thickness (Figure 2d,e).Given that, in CF-Mu 3 Gel, it was possible to co-culture both bacteria without external apparatus to achieve anaerobic conditions, the proposed 3D substrate provides a further advantage to co-culture these pathogens, and possibly, other complex microbial communities, i.e., polymicrobial infections or microbiota, where it is needed to co-culture anaerobic, microaerophilic and aerobic microorganisms together.In CF-Mu 3 Gel, both bacteria were able to colonize and form aggregates across the 3D structure, which appear in contact with each other (Figure 5d-f, Figure S9, Supporting Information).When the coculture period was prolonged from 48 h (Figure 5) to 72 h (Figure S10c, Supporting Information "No ATB"), P. aeruginosa progressively overtook S. aureus.Indeed, the size of P. aeruginosa aggregates increased over longer incubation times, as can be seen in Figure 5e,f, in which P. aeruginosa was allowed to grow 24 h in co-culture with pre-cultured of S. aureus (S. aureus first, followed by another 24 h of co-culture with P. aeruginosa), while in the second P. aeruginosa was allowed to grow for full 48 h (P.aeruginosa first, followed by another 24 h of co-culture with S. aureus).The organization of P. aeruginosa aggregates (Figure 5e,f) negatively influenced S. aureus growth for longer periods of culture (Figure 5b; Figure S10c, Supporting Information "No ATB").This result is supported by previous studies on the P. aeruginosa/S.aureus antagonistic phenomenon, in which co-culture induced the secretion of anti-staphylococcal molecules and proteases. [4,5]t is clearly understood that antimicrobial treatment drastically reduces CF patient's morbidity and increases life expectancy. [75,76]he key therapeutic challenge, however, as addressed in this paper, is that efficacy testing on planktonically cultured bacteria is misleading.Indeed, it is estimated that the antimicrobial tolerance of bacteria in biofilms is 100-1000 times greater than those determined planktonically. [77]With our approach, we re-produced what occurs in CF antimicrobial treatment conditions.This is evidenced by the fact that in the CF-Mu 3 Gel the number of bacteria after exposure to the minimal inhibitory concentration (MIC) was three orders of magnitude higher than those found under planktonic conditions.Overall, even at antimicrobial concentrations which are lower or higher than the MIC, S. aureus and P. aeruginosa were more resistant to antimicrobial treatment when cultured within CF-Mu 3 Gel, independently of the antimicrobial agent (Figure 4a-e).S. aureus exhibited high tolerance toward ciprofloxacin, even when shorter incubation periods were adopted, an effect previously linked to reduced ATP levels, associated with inhibition of S. aureus respiration by ciprofloxacin. [78]F-Mu 3 Gel, as CF mucus, provides a 3D steric and interactive barrier to the diffusion of antimicrobial agents, a phenomenon which may explain the lack of efficiency of antimicrobial treatment (Figure 3).The CF-Mu 3 Gel, as CF mucus together with the exopolysaccharidic matrix produced by bacteria, offered a significant barrier to the diffusion of all antimicrobial agents (Figure 3 and Figure S7, Supporting Information).The diffusion of colistin sulphate, tobramycin, and ciprofloxacin was strongly reduced by CF-Mu 3 Gel, possibly due to the interaction of these drugs with mucin, [79] its main component.Colistin and tobramycin are highly positively charged, and as previously demonstrated, these interact with mucins.This binding decreases the MIC potency of antimicrobials.[79] Low permeability through the mucus may limit the amount of drug that encounters both S. aureus and P. aeruginosa, thus affecting its efficacy.In this sense, it is significant to note that the MIC of these antimicrobial agents against S. aureus and P. aeruginosa was shown to be one to three orders of magnitude higher in the presence of mucin.[79,80] The lack of antimicrobial efficiency can be also linked to the formation of persister populations.[81] In CF-Mu 3 Gel, both S. aureus and P. aeruginosa were found to be more resistant to ciprofloxacin treatment when co-cultured than when in mono-cultures (Figure 6b,d).This is supported by the literature, as S. aureus co-cultured with P. aeruginosa revealed to be more resistant to ciprofloxacin treatment when in the presence of one another.[82,83] All of this underscores and emphasizes the need for significantly more reliable screening tools.In the wide library of antimicrobial agents that can be tested, ciprofloxacin, tobramycin and colistin were selected because these are the most administered treatments for CF therapy and because they have distinctive modes of action.Although this suggests that the platform is versatile, we cannot confirm that the observed effects can be extrapolated to other antimicrobial agents. Atimicrobial treatments were performed selecting, as reference concentration, the MIC determined following EUCAST guidelines.However, local concentrations may greatly vary according to the method of administration and the pharmacokinetic of each drug.For example, the local peak concentrations of aerosolized drugs could be different from those determined systematically after oral or intravenous administration.[84][85][86] The localized presence of mucins may locally negatively affect the antimicrobial efficacy and thus counteract the benefits of local delivery.[79] In fact, by providing the mucin content and the characteristic steric and interactive layers of mucus together with the presence of bacterial aggregates, CF-Mu 3 Gel closely reproduces the obstacles that antimicrobials must overcome to efficiently treat infections.The obtained results may also suggest that improvement of therapeutic strategies against CF airway infections may lay in favoring the diffusion of antimicrobial agents through the CF mucus either by designing antimicrobial drugs that are non-binding to the mucus or by incorporating antimicrobials in micro-to nanoparticles that are able to surpass the mucus barrier.We have in this paper considered the two main pathogens colonizing CF airways and have demonstrated the efficacy of this approach.Further studies are needed to investigate and confirm the ability of CF-Mu 3 Gel to sustain the growth of multiple pathogens, and possibly the CF airway microbiota retrieved from different patients.Our work has the potential to create opportunities for the development of personalized and effective antimicrobial treatments.

Conclusion
In this study, we demonstrate, with this example-based research, the enormous advantage of switching from planktonically cultured bacteria to heterogenous 3D structures that mimic key characteristics of the natural microbial niche, including the presence of O 2 and structural gradients.Specifically, the complex heterogeneity of CF-Mu 3 Gel enabled single-and dual-species survival and formation of clinically witnessed microcolony aggregates which are characteristic of the CF mucus.These distinctive outcomes were obtained under normal culture conditions, without further use of anaerobic conditions, by simply controlling the materials production, which involved the formation of structural gradients at the molecular level and their characterization.These features required unconventional tools for their characterization, such as PCI.This technique spotted structural and dynamical heterogeneities that can be associated with the gradient crosslinking produced by the diffusion method.
Considering the possibility to provide a model with a clinical relevance, the proposed mucin-containing hydrogel exhibits a chemical composition, viscoelastic properties and both structural and O 2 gradients that reproduce key features of those of CF airway mucus, but, most importantly, these characteristics drove bacterial organization in size and format similar to those clinically seen in CF mucus.Additionally, the mismatch of 1-to-3 orders of magnitude on the efficacy of antimicrobial treatment between bacteria cultured in traditional methods (planktonic cultures) and those in patients was also depicted in cultures in CF-Mu 3 Gel.We are aware of the possible limits of the use of commercially available mucins, which are themselves a "model" of the different native mucins within a model.However, the flexibility of the production process does not preclude further tuning of the hydrogel composition to include specific mucins or other components that are relevant to control bacterial behavior.
From a technological point of view, the developed 3D-substrate can be further tuned to closely and accurately model different environmental conditions.This opens the possibility to produce standardized in vitro platforms for basic microbiology research and drug development.

Experimental Section
CF-Mu 3 Gel Properties: Gradient-CF-Mu 3 Gel (for short, it was termed CF-Mu 3 Gel) is composed of 2.5% (w/v) porcine stomach type III mucin (Sigma-Aldrich, M1778; lot# SLBQ7188V, Germany) and 0.71% (w/v) sodium chloride that falls within the ranges previously determined in CF sputum. [32,33]G-CF-Mu 3 Gel was produced in MHB, LB or in 7.07 mg mL −1 NaCl.The method of production is now patented (IT102018000020242A). [43] Briefly, G-CF-Mu 3 Gel was produced using a controlled diffusion system.The apical compartment contained a mucin/alginate solution, while the basolateral compartment comprises the crosslinking media that in this study was composed of different concentrations of Ca 2+ ions ranging from 0.12-1.20%(w/v).In this way, 3D structures with spatial gradients were generated.
Rheological Characterization: The viscoelastic properties of CF-Mu 3 Gel produced in different media (7.07 mg mL −1 NaCl, MHB, and LB) were evaluated using an Anton Paar MCR502 Rheometer (Austria) with a 25 mm diameter plate geometry (serial number 52530/19910) at 25 °C.The linear viscoelastic region (LVR) was determined through strain sweep analyses employing a strain logarithmic ramp varying from 0.1% to 1000% at a frequency of 1 Hz.Oscillatory frequency sweeps were further performed to evaluate both storage, G', and dissipative, G″, moduli, as well as complex shear modulus, G*, at 0.5% (at strain amplitudes within the linear regime) with frequencies changing logarithmically in the 0.1-20 Hz range.The viscoelastic properties of CF-Mu 3 Gel were further compared to those reported for CF sputum.[36] The rheological data was further exploited to estimate the mesh size of different formulations of CF-Mu 3 Gel by exploiting both the generalized Maxwell model and the rubber elasticity theory, as previously reported. [43,61]hoton Correlation Imaging (PCI): PCI is an optical technique that blends the power of Dynamic Light Scattering (DLS) with that of digital imaging.The setup employed for the characterization of CF-Mu 3 Gel is reported in Figure S2 (Supporting Information), adapted from a previous setup. [87]A classical DLS experiment allows one to investigate the microscopic dynamics of a sample by measuring the degree of correlation c(t, t+) between the intensity of the light scattered by the sample at two different times t and t+.In the simplest case of the Brownian motion of a dispersed nanoparticle, for instance, this quantity is simply related to particle diffusion over distances of the order of the inverse of the scattered wave vector q = 4/ sin (/2), where  is the angle at which the scattered light is measured (in the experiment,  = 90°) and is the wavelength of the scattered light.PCI yields similar information but, by operating on an image of the sample produced on a CMOS multi-pixel camera by a suitable "stopped-down" optics, [88] retains at the same time the spatial resolution of an imaging system.This allows the spatial distribution of both the scattered intensity and of the degree of correlation at the given delay time t to be mapped (respectively dubbed "intensity" and "correlation" maps).To the aims, PCI has two main advantages with respect to standard DLS: (a) it provides spatial distribution of both the scattered intensity and of the degree of correlation, thus allowing to spot and quantify heterogeneity both in the structure and in the microscopic dynamics of the sample.
(b) standard DLS measurements require the medium to be "ergodic," namely that the scatterers are free to move over distances that are not much smaller than the wavelength.This condition, which is required for the averages in time measured in a DLS experiment to coincide with true statistical averages ("ensemble" averages), is satisfied by Brownian particle dispersions, but not by a hydrogel, where the microscopic motion is almost completely frozen.This limitation is overcome by PCI, where a correct ensemble average is easily obtained from a spatial average over a suitable but still "region of interest."Thus, PCI provides a unique method to investigate the slow and spatially limited spontaneous restructuring dynamics of soft disordered solids. [88]cterial Strains and Culture Conditions: The microorganisms used were Staphylococcus aureus ATCC 25923 (S. aureus) and Pseudomonas aeruginosa ATCC 15692 (PA01 strain, P. aeruginosa), kindly supplied by R. Migliavacca (Department of Clinical Surgical, Diagnostic and Paediatric Sciences, University of Pavia, Italy).Brain heart infusion broth (BHI; Sigma-Aldrich, 53286, Germany) and Luria Bertani broth (LB; Formedium, LMM0102, United Kingdom) were used to inoculate S. aureus and P. aeruginosa, respectively.Bacteria were grown overnight in their appropriate medium, under aerobic conditions at 37 °C using a shaker incubator (VDRL Stirrer 711/CT, Asal Srl, Italy).Both cultures were prepared at the final density of 1 × 10 9 bacteria mL −1 as determined by comparing the optical density at 600 nm (OD600) of the sample with a standard curve relating OD600 to bacterial number by using a UV-VIS spectrophotometer (Aurogene S.r.l., HJ1908003, Italy). [89]Müller Hinton broth (MHB; Sigma-Aldrich, 70192, Germany) was used to prepare the final inoculum of both bacterial mono-and co-cultures to be inoculated in CF-Mu 3 Gel or used as planktonic culture controls.The total number of colony forming units (CFU) was counted by plating the serial dilutions (prepared in 0.9% NaCl) of bacterial cultures on Müller Hinton agar (MH agar; Sigma-Aldrich, 70191, Germany) plates, while cetrimide agar (Sigma-Aldrich, 22470, Germany) and mannitol salt (Sigma-Aldrich, 63567, Germany) were used as selective media to count the CFU of P. aeruginosa and S. aureus after co-culture experiments, respectively.All media and plates of MH, mannitol salt and cetrimide agar were prepared following supplier instructions.Sodium citrate tribasic dihydrate (NaCitrate; Sigma-Aldrich, 71404, Germany) was used as a dissolving agent of CF-Mu 3 Gel to retrieve bacteria that grew within its 3D structure.Three antimicrobial agents were used, including colistin (Sigma-Aldrich, C4461, Germany), ciprofloxacin (Fresenius Kabi, lot# 15LA518P2, Italy), and tobramycin (Ibi, lot# 118B, Italy).
Mono-Cultures in CF-Mu 3 Gel: S. aureus or P. aeruginosa were cultured in CF-Mu 3 Gel for 24 and 48 h incubation.After overnight culture, the bacterial suspension was subsequently diluted in fresh medium, until reaching the final concentration of 10 4 bacterial cells mL −1 .Cultures in CF-Mu 3 Gel were achieved by inoculating on the top 100 μL of each separated bacterial suspension and incubating for 24 and 48 h at 37 °C.Planktonic cultures of S. aureus or P. aeruginosa (i.e., bacteria cultured in liquid medium without hydrogels) in 96 multiwell flat tissue culture plates were used as controls of the experiment.
Co-Cultures in CF-Mu 3 Gel: Co-cultures of both S. aureus and P. aeruginosa were induced in CF-Mu 3 Gel in a proportion of 1:1 following three different co-culture settings: a) simultaneous culture of both P. aeruginosa and S. aureus for 24 h; b) culturing first S. aureus for 24 h followed by P. aeruginosa cultured for another 24 h; and finally, c) colonizing CF-Mu 3 Gel with P. aeruginosa first for 24 h, after which S. aureus was added and incubated for further 24 h.Before adding the second bacterial strain, the excess medium was removed and substituted with 100 μL of 10 4 bacterial cells mL −1 of the secondly introduced bacterium.Co-cultures of S. aureus and P. aeruginosa in planktonic conditions were also carried out as the controls of the experiment for the three different co-culture settings.
Bacterial Viability within CF-Mu 3 Gel: CF-Mu 3 Gel was dissolved with 150 μL of 50 mm NaCitrate, pH 7.4, for bacterial viability determination.After 2 min contact time, the suspension was diluted to 10 −6 -10 −10 in 0.9% NaCl solution, pH 7.4, to perform serial dilution and CFU counting.MH agar plates were incubated overnight at 37 °C.When cocultures were carried out, the total number of CFU was counted using MH agar plates, while cetrimide or mannitol salt agar plates were used as selective media to count the CFU of P. aeruginosa and S. aureus, respectively.CFUs were then calculated taking into consideration a dilution factor of 2.5, resulting from the addition of 50 mm sodium citrate.Planktonic cultures of S. aureus or P. aeruginosa were used as controls of the experiment.
Oxygen Tension Measurements: O 2 tension of colonized CF-Mu 3 Gel was measured using a Clark-type O 2 sensor (OX-25; Unisense, Aarhus N, Denmark), connected to the Unisense microsensor multimeter S/N 8678 (Unisense, Denmark), a high sensitivity pico-ampere four-channel amplifier.Before each measurement, the reference anode and the guard cathode were polarized overnight and calibrated with either water saturated with air or with an anoxic solution of 2% (w/w) sodium hydrosulfite.Once the calibration was carefully made, a low melting point agarose consisting of 2% (w/v) agarose in 0.071% (w/v) NaCl (same concentration as in CF-Mu 3 Gel) was cast into a petri dish.CF-Mu 3 Gel colonized with either S. aureus or P. aeruginosa were placed over the agarose layer.Microsensors with a tip diameter of 50 μm (OX-50) were positioned at the air-CF-Mu 3 Gel interface ("depth zero") using a motorized micromanipulator (Unisense, Denmark).Measurements were performed at the center of the hydrogels starting at their surface (0 mm) through their thickness, every 100 μm in triplicate, until the tip completely penetrated the whole structure.The maximum depth reached by the tip, termed end depth, was 2300 μm.Sterile CF-Mu 3 Gel was used as control of the experiment.The Unisense software SensorTrace automatically converts the signal from partial pressure (O 2 tension) to the equivalent O 2 concentration in μmol L −1 .
Mono-Cultures in CF-Mu 3 Gel: The viability of S. aureus ATCC 25923 or P. aeruginosa PA01 within CF-Mu 3 Gel was assessed using the Live/Dead dual staining BacLight Kit (Live/Dead Bacterial Viability Kit, L-7007, Molecular Probes) according to manufacturer specifications.The BacLight kit is composed of two fluorophores SYTO9 and propidium iodide (PI) that stain total and dead bacteria, respectively.After 10 min incubation with SYTO9, the colonized CF-Mu 3 Gel was washed three times, followed by a 3 min staining with PI.Viable (green) and non-viable (red) bacteria within CF-Mu 3 Gel were differentiated by using the dual-channel option of the Leica TCS SP8 confocal microscope.Since the P. aeruginosa strain fluoresces green, only the PI or Hoechst 33342 (Hoechst; Invitrogen, Eugene, Code) staining were used, whereas for S. aureus (fluoresces in red) only the SYTO9 was used.
Co-Cultures in CF-Mu 3 Gel: The experiment was performed using DsRed-S.aureus ATCC 25923 (red fluorescence) and GFP-P.aeruginosa PA01 (green fluorescence) to discriminate between the two bacterial strains.The general total number of these bacteria after co-culture in CF-Mu 3 Gel was evaluated using the Hoechst 33342 staining.Hoechst 33342 can readily cross cell membranes to stain the DNA (blue) of total bacteria (blue).50 μL of a mixture of Hoechst 33342 with a final concentration of 1 μg mL −1 was added to each colonized CF-Mu 3 Gel.After Hoechst staining, DsRed-S.aureus ATCC 25923 appeared purple and the GFP-P.aeruginosa PA01 were blue-green.
Confocal Laser Scanning Microscopy: CF-Mu 3 Gel were seeded with DsRed-expressing S. aureus and/or GFP-expressing P. aeruginosa to analyze bacterial colonization in mono-and co-cultures using Confocal Laser Scanning Microscopy.In this way, cultures were performed in CF-Mu 3 Gel following the aforementioned method to induce mono-and co-cultures for 12, 24, and 48 h.After each incubation period, colonized CF-Mu 3 Gel were observed under a high-resolution spectral confocal-laser microscope (Leica TCS SP8 SMD, Leica Microsystems CMS GmbH, Germany).Colonized CF-Mu 3 Gel were analyzed from the top, bottom, and middle.Analyses were performed at different wavelengths according to fluorochrome and staining, namely 550 (570/650) nm excitation for DsRed-expressing S. aureus, 488 (500/550) nm excitation for both GFP-expressing P. aeruginosa and SYTO9, 535 (570/670) nm excitation for PI, and 405 (420/500) nm excitation for Hoechst 33342.The 3D images obtained were analyzed using the LasX 3.7.5 software provided by Leica.In the case of monocultures by DsRed-S.aureus ATCC 25923, CF-Mu 3 Gel was supplemented with 10 μg mL −1 CP, whereas in the case of mono-cultures by GFPexpressing P. aeruginosa PA01, CF-Mu 3 Gel containing 300 μg mL −1 CB was used.As a control, CF-Mu 3 Gel without bacteria was observed under confocal microscopy at the aforementioned wavelengths to evaluate the absence of possible autofluorescence effects.
Antimicrobial Susceptibility: Determination of Minimal Inhibitory Concentration: Ciprofloxacin, tobramycin, and colistin were selected as antimicrobial agents to which P. aeruginosa is sensitive to determine antimicrobial susceptibility within CF-Mu 3 Gel, while ciprofloxacin chosen as antimicrobial agents against S. aureus.The minimal inhibitory concentration (MIC) was determined by the broth microdilution method using MHB inoculated with a standard inoculum according to The European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines.In this way, 10 3 bacterial cells in MHB were incubated with serial dilutions of 1:2 of each antimicrobial agent (0, 0.0625, 0.125, 0.25, 0.5, 1, 2, 4, 8, 16, 32, 64 mg L −1 ) for 24 h at 37 °C.OD600 was measured to determine the values of MIC using the 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) colorimetric test (Sigma-Aldrich, Germany) according to supplier instructions, which transforms the metabolic products of bacteria into a violet observable precipitate that can be read using a spectrophotometer.The intensity of the violet color is proportional to the number of viable bacteria.The plate was then read at a wavelength of 570 nm and the background read at 630 nm was subtracted (Clariostar microplate reader, BMG Labtech Clariostar, Germany).MIC was established as the lowest concentration of antimicrobial agents that inhibited bacterial growth.
Antimicrobial Susceptibility within CF-Mu 3 Gel: Mono-cultures of either S. aureus or P. aeruginosa in CF-Mu 3 Gel were induced by culturing 10 3 bacteria to achieve 10 8 bacteria after 24 h of culture.After this period, the supernatant over the hydrogels was removed and these were washed twice with fresh MHB medium.Then, 100 μL of antimicrobial agent at three different concentrations was added, namely 0.1, 1, and 10 MIC.Antimicrobial treatment was carried out for 24 h under static incubation, at 37 °C, after which CFU count was performed.Similarly, cocultures of P. aeruginosa and S. aureus in CF-Mu 3 Gel were treated with ciprofloxacin, given that both bacteria are sensitive to this antimicrobial agent.Both S. aureus and P. aeruginosa were cultured in CF-Mu 3 Gel in a proportion of 1:1 following the co-culture setting b: culturing S. aureus first followed by P. aeruginosa cultured for another 24 h.Several controls were performed: I) planktonic cultures after antimicrobial treatment; II) planktonic cultures without antimicrobial treatment; and III) nontreated but colonized CF-Mu 3 Gel.Both MHB medium and CF-Mu 3 Gel without bacteria and antimicrobial treatment were used as controls of sterility.
Drug Diffusion Studies: Drug diffusion studies were conducted using either polycarbonate Transwell supports (Corning Transwell, CLS3413, Merck) with a porosity of 0.4 μm and inner diameter equal to 6.5 mm or a 96-well plate permeable support system with a filter plate pre-coated with structured layers of phospholipids (PAMPA; Corning Gentest Pre-coated PAMPA, 353015, USA) with a porosity of 0.45 μm and inner diameter equal to 6.2 mm.Drug diffusion through empty Transwell inserts and empty PAMPA were performed as controls of the experiment.Both ciprofloxacin, tobramycin, and colistin were tested at a concentration of 500 μm.CF-Mu 3 Gel was introduced on both platforms with a thickness of ≈500 μm.Afterward, the donor compartment was filled with the relevant drug solution (200 μL per well) and 300 μL of phosphate saline buffer was added to the acceptor compartment.The filter plate was then coupled to the receiver plate and incubated at room temperature without agitation for 5 and 24 h.At the end of the incubation period, the plates were separated and the volume of both the donor and the receiver plate was collected for further quantification.Liquid chromatography-mass spectrometry was further carried out to quantify the permeated amount of drugs using their relevant calibration curves.The percentage of drug permeated after 5 and 24 h was calculated as follows: where C t denotes the concentration of drug released at time t, and C 0 represents the initial concentration introduced on the donor compartment.
The apparent permeability coefficient (P app ) was expressed using Equation (1) derived from Fick's law [92] for steady state conditions: where dQ is the quantity of drug expressed as moles permeated into the acceptor compartment at time t (18 000 s), C 0 is the initial concentration in the donor well, and A is the area of the well membrane (0.3 cm 2 ).The P app was used as an average of all the measures.Liquid Chromatography-Mass Spectroscopy: An HPLC-MS using a Varian HPLC equipped with a 410 autosampler and an Ascentis C18 column (10 cm × 2.1 mm, 3 μm).Gradient mobile phases composed of acetonitrile and water with 0.1% formic acid as organic and aqueous phases, respectively, were pumped at a flow rate of 200 μL min −1 .A flow of 200 μL min −1 and an injection volume of 10 μL were used.Compounds were detected on a Varian 320 MS TQ Mass Spectrometer equipped with an electrospray ionization (ESI) source operating in positive mode.The detector was used in multiple reaction monitoring (MRM) mode and the transitions of each drug are reported in Table S1 (Supporting Information).
Statistical Analysis: The results of at least three independent experiments are presented as mean ± standard deviation (SD).Statistical analysis was performed using the t-test student and ANOVA using GraphPad Prism version 8 (GraphPad Software, USA).Significant differences were set for *p < 0.05.n corresponds to the number of distinct samples in which different measurements were performed.supported by S.V. and L.V., which equally contributed to the manuscript.D.P.P. and N.S.V. developed the production method of CF-Mu 3 Gel, under the supervision of P.P. D.P.P., F.B., N.S.V., G.G. and A.Z. established and conducted the microbiological and antimicrobial susceptibility analyses under the supervision of LV.D.P.P., A.Z., G.G., and S.v.U.carried out confocal microscopic investigations and analyses.C.S.B. determined the permeability of the antimicrobial agents under the supervision of S.V. D.P.P. performed the rheological characterization and further analyses of all variations of CF-Mu 3 Gel under the supervision of F.B.V. D.P.P. and E.C. established and performed the analyses of O 2 tension.D.P.P., S.v.U.prepared the samples for PCI analysis.V.R., S.B analyzed PCI data under the supervision of R.P. D.P.P. wrote the original draft of the manuscript and took care of the revisions of the draft, P.P. contributed to the writing and revisions.All authors discussed the results, reviewed the manuscript, and contributed to its final version.

Figure 1 .
Figure 1.Topographical, chemical, and physical features of CF-Mu 3 Gel.ab) Viscoelastic properties of CF-Mu 3 Gel in function of production media: storage modulus, Gʹ (Pa), dissipative modulus, Gʹʹ (Pa), and complex dynamic modulus, G* (Pa) of CF-Mu3 Gel was analyzed at a, breathing (≈0.5 Hz); and ciliary beating frequency (10 Hz; n ≥ 6) (b).The rheological data were analyzed using the two-way ANOVA.Significant differences were set for *p < 0.05.A minimum of six independent samples were analyzed per formulation (n ≥ 6).ns means no statistical differences were found between the viscoelastic properties of the different formulations of CF-Mu 3 Gel and those of CF sputum reported byYuan et al. (2015).[37]c) Dependency of mesh size () of CF-Mu 3 Gel on media in which this was produced estimated by combining the Generalized Maxwell Model with the rubber elasticity theory (n ≥ 6).d) Macroscopic image of CF-Mu3 Gel produced with 0.16% (w/v) Ca 2+ ions.e) Correlation (top) and intensity (bottom) maps c i of a region of ≈2.83 × 2.66 mm, obtained after 20 h of crosslinking reaction for a CF-Mu 3 Gel.The correlation index is evaluated for a delay of 60 s and the intensity is normalized to its value at the top.f) Schematic representation of both topographical and O 2 gradients exhibited through the 3D structure of CF-Mu 3 Gel.g) Heterogeneous distribution of O 2 tension through the structure of sterile CF-Mu 3 Gel (n = 5).

Figure 2 .
Figure 2. Mono-cultures of CF-Mu 3 Gel with either S. aureus or P. aeruginosa.a,b) Bacteria viability, determined through Colony Forming Units (CFU), after incubating either S. aureus (a) or P. aeruginosa (b) with 10 3 bacteria for 24 and 48 h in CF-Mu 3 Gel (n = 9) with the confocal microscopy images of colonized CF-Mu 3 Gel for 24 and 48 h (the scale bar corresponds to 10 μm).The results obtained with CF-Mu 3 Gel (green) were further compared with planktonic cultures (white).In (a) and (b), the shadowed region depicts the range of CFU mL −1 of each bacterium reported in the literature for pathological CF mucus.[45]A minimum of five independent samples were analyzed per bacterium (n > 5).No statistical differences were detected neither between both planktonic cultures and CF-Mu 3 Gel nor between the same culturing system at different culturing periods.c) Experimental setup adopted to measure oxygen (O 2 ) tension through CF-Mu 3 Gel colonized with either d) S. aureus or e) P. aeruginosa for 12 (green), 24 (blue), and 48 h (red) (n = 5).

Figure 4 .
Figure 4. Antimicrobial treatment of CF-Mu3 Gel colonized with either S. aureus or P. aeruginosa for 24 h.Ciprofloxacin treatment of CF-Mu3 Gel colonized with S. aureus for a) 6; and b) 24 h.0.1, 1, and 10 MIC correspond to 0.05, 0.5, and 5 mg L −1 , respectively.Antimicrobial treatment of CF-Mu3 Gel colonized with P. aeruginosa for 24 h followed by treatment with: c) ciprofloxacin; d) tobramycin; and e) colistin.In the case of ciprofloxacin, 0.1, 1, and 10 MIC correspond to 0.05, 0.5, and 5 mg L −1 , respectively, while for both tobramycin and colistin 0.1, 1, and 10 MIC correspond to 0.8, 8, and 80 mg L −1 , respectively.No ATB means that no antimicrobial treatment was performed.Planktonic cultures of S. aureus and P. aeruginosa were used as controls of the experiment.0 CFU mL −1 indicates that no CFU were detected.All data were analyzed using the one-way ANOVA.Significant differences were set for *p < 0.05; and **p < 0.01.A minimum of five independent samples were analyzed per formulation (n ≥ 5).

Figure 5 .
Figure 5. Co-cultures of CF-Mu3 Gel with both S. aureus and P. aeruginosa following three different co-culture settings: a,d-1) contemporary culture of both S. aureus (in light red) and P. aeruginosa (in light green) for 24 h; b,e-2) first culture with S. aureus for 24 h followed by P. aeruginosa culture for other 24 h; and finally, c,f-3) culture with P. aeruginosa first for 24 h, after which S. aureus is added and incubated for further 24 h.Bacteria numbers determined in CF-Mu 3 Gel (green) were further compared with planktonic cultures (white) (a-c).The shadowed region depicts the range of CFU mL −1 of each bacterium reported in the literature for pathological CF mucus.[45,47]All data were analyzed using two-way ANOVA.Significant differences were set for *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.A minimum of five independent samples were analyzed per formulation (n ≥ 5).The relative abundance of S. aureus in respect to P. aeruginosa per each experimental setting is depicted in the pie charts in the upper part of (a-c).Confocal microscopy analyses of CF-Mu 3 Gel co-colonized with both DsRed-fluorescent S. aureus (fluorescent red) and GFP-fluorescent P. aeruginosa (fluorescent green) (d-f).Co-colonized CF-Mu 3 Gel was further stained with Hoechst 33342 to depict total bacteria (in blue) and 3D reconstruction are also provided.The scale bar corresponds to 10 μm.

Figure 6 .
Figure 6.Antimicrobial treatment of CF-Mu 3 Gel co-colonized with both S. aureus and P. aeruginosa followed by antimicrobial treatment with ciprofloxacin.0.1, 1, and 10 MIC correspond to 0.05, 0.5, and 5 mg L −1 , respectively.Bacterial survival after ciprofloxacin treatment of mono-(white) and co-cultures (colored) of: a,b) S. aureus; and c,d) P. aeruginosa either in planktonic conditions (a, c); or CF-Mu 3 Gel (b, d).Bacterial survival was calculated by dividing the CFU number of S. aureus and P. aeruginosa after antibiotic treatment by the CFU number of S. aureus and P. aeruginosa without antibiotic treatment (No ATB).All data were analyzed using two-way ANOVA.Significant differences were set for *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.A minimum of five independent samples were analyzed per bacterium (n ≥ 5).