Fragmented but functional: Post-dispersion dynamics and phenotypic variation in dispersed biofilm-associated cells

TL;DR

This study explores how dispersed biofilm cells remember their development and form new, complex biofilms quickly.

Contribution

The study reveals that dispersed biofilm cells retain developmental memory and form more complex secondary biofilms.

Findings

Dispersed S. epidermidis cells retain developmental memory for rapid reattachment.

Secondary biofilms from mature dispersed cells are denser and structurally complex.

EPS matrix in secondary biofilms has increased polysaccharides.

Abstract

Schematic overview of primary biofilm formation, dispersion of primary biofilm-associated cells (PBACs), and development of secondary biofilms.Image 1

Schematic overview of primary biofilm formation, dispersion of primary biofilm-associated cells (PBACs), and development of secondary biofilms.

•Dispersed S. epidermidis cells retain developmental memory for rapid reattachment.•Mature dispersed cells show enhanced growth rate with shorter doubling time.•Secondary biofilms developed from mature PBACs are denser and structurally complex.•EPS matrix remodelling features increased polysaccharides in secondary biofilms.

Dispersed S. epidermidis cells retain developmental memory for rapid reattachment.

Mature dispersed cells show enhanced growth rate with shorter doubling time.

Secondary biofilms developed from mature PBACs are denser and structurally complex.

EPS matrix remodelling features increased polysaccharides in secondary biofilms.

Figures

Schematic of the Ibidi μ-Slide VI 0.4 flow cell setup used for *S. epidermidis* biofilm growth under continuous, dynamic flow conditions. The primary components include a syringe pump, the media reservoir, the tubing assembly, and the collection flask for the dispersed effluent. The initial step involved inoculating the channels with overnight cultures. Subsequently, the media were purged, and the flow was maintained at a regulated flow rate of 10 μL/min to facilitate biofilm formation (Created with [https://www.biorender.com/](https://www.biorender.com/)).Fig. 1

Viability of Primary Biofilm-Associated Cells (PBACs) at Defined Time Points (A) Quantification of live/dead ratios of PBACs assessed using Syto9 and propidium iodide staining followed by OD _600nm_ measurement in a microplate reader across 6-8 h, 12-15 h, and 21-24 h time points. No statistically significant differences in viability among populations were observed (B) Representative fluorescence microscopy images showing green (Syto9 positive) viable cells and small aggregates in all PBAC populations. (Scale bar represents 10 μm) Experiments performed in triplicate; data presented as mean ± S

Comparison of (A) growth kinetics and (B) doubling times across planktonic cells and biofilm-associated cells (PBACs) harvested at different biofilm maturation stages. PBACs displayed significantly shorter doubling times relative to planktonic cells. Data represent mean ± SEM from biological replicates (n = 3). Statistical significance was assessed using one-way ANOVA with Tukey's post hoc test and is indicated as follows: n.s., not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.Fig. 3

Time-lapse CLSM imaging and COMSTAT analysis showing (A) Surface area (μm^2^) and (B) Biomass (μm^3^/μm^2^) progression over 24 h duration from z-stacks captured at regular intervals. COMSTAT analysis revealed an initial decline in both biomass and surface coverage until 12 h, followed by a marked increase after 16 h, with peak values at 24 h. Data represent mean ± SD from three biological replicates. CLSM live imaging was performed using Syto9 staining. (C) Representative CLSM images illustrating structural progression of primary biofilms at early (6-8 h), developing (12-15 h), and late (21-2

Orthogonal CLSM Projections of Primary and Secondary Biofilms developed from PBACs at Different Stages. On the bottom panel of each image (XZ), horizontal cross-sections and on the right-side panel (YZ), vertical cross-sections are mentioned. Images display differences in microcolony density, vertical stratification, and distribution across biofilm types. Secondary biofilms from later-stage PBACs show denser and more uniform structures. (Scale bar represents 10 μm).Fig. 5

Quantitative (COMSTAT) comparison of key architectural features of primary and secondary biofilms developed from PBACs of different maturation stages. (A) Biomass, (B) Maximum thickness, and (C) Average thickness all increased significantly with PBAC age, indicating denser and more developed biofilm structures. (D) Roughness coefficient decreased in later-stage PBAC biofilms, suggesting more uniform architecture. (E) Surface area coverage increased, indicating improved surface colonisation in biofilms formed from late stage PBACs. (F) Surface-to-biovolume ratio declined with PBAC maturation, r

Polysaccharide composition of primary and secondary biofilms formed by PBACs of increasing maturation. (A) Maximum intensity projection of a confocal stack of Syto9 (blue signal) and WGA-stained biofilms showing enhanced polysaccharide signal (red) with PBAC age. (Scale bar represents 20 μm) (B) Quantitative analysis of percentage of polysaccharide distribution in EPS matrix, with significant increases in secondary biofilms from 12 to 15 h and 21-24 h PBACs compared to primary and early-stage PBACs. Data represent mean ± SD (n = 3). Statistical significance determined using one-way ANOVA. Stat

Full text

Introduction

1

Biofilms comprise highly organized communities of microorganisms that are attached to interfaces and encased within a matrix of extracellular polymeric substances (EPS) [1]. The biofilm matrix comprises an intricate structure consisting of diverse components, such as proteins, polysaccharides, and extracellular DNA [2,3]. These components are essential for biofilm formation, stability, and functionality. Pseudomonas species, especially Pseudomonas aeruginosa, have been thoroughly investigated for their biofilm production ability and matrix composition [4]. Biofilm matrix functions as a defense mechanism for the microbes under extreme conditions, for instance, osmotic pressure, hydrodynamic stress, temperature, pH, and chemicals. Microbiologists have generally defined biofilms as surface-attached microbial communities; however, non-attached aggregated entities are now also considered biofilms when they exhibit hallmark characteristics such as multicellular organization, matrix association, and enhanced tolerance to environmental stresses. The recognition of such non-attached aggregates is significant, as they have been shown to contribute to persistence, dissemination, and recurrence of biofilm-associated infections, particularly in clinical settings where surface attachment may be transient or heterogeneous [1,[5], [6], [7]]. The identification of non-attached biofilm-like aggregates has necessitated a re-evaluation of the role various bacterial growth stages play in infection dynamics. Recent studies have challenged the traditional perspective that acute infections are solely caused by planktonic bacterial populations, which are easily treatable with antibiotics, demonstrating that non-planktonic (biofilm-associated or aggregated) populations can also play a role in acute disease processes [8].

Traditionally, a five-step developmental model was generally accepted that includes two initial attachment phases, two maturation phases, and a final dispersion stage [1]. The dispersion step involves both cell and matrix component escaping from the biofilm and can also result in relocalization onto a new surface [9]. The general mechanism of dispersion is divided into two categories: active and passive. Active dispersion is generally biologically mediated. External factors like shear force, abrasion, human intervention, or predator grazing lead to passive dispersion [10]. Terms such as detachment, erosion, and dispersion have been used interchangeably in the scientific literature, although they may emphasize distinct aspects such as mechanical release, matrix breakdown, or active cell escape [9]. Coenye et al. compared biofilm dispersion studies with opening Pandora's box, highlighting the complexity and potential for new therapeutic opportunities in targeting dispersion mechanisms [11]. It is also noticeable that different dispersion mechanism also plays a great role in how the cells behave after dispersion and how the antibiotic susceptibility changes with that [11]. A study conducted by Oshumi et al. showed that the residual structure of a disinfected Streptococcus mutans biofilm can facilitate secondary bacterial attachment and the relocalization of biofilm cells [12]. The bacterial dispersion event affects both secondary colonisation and the transmission of diseases, and therefore, it is important to study the process of dispersion and the outcome of it in detail [11].

The more recent mode of biofilm development with more explicit incorporation of aggregates has been proposed by Ref. [7]. However, the study of non-attached aggregates has still not received sufficient attention. Such aggregates have gained significance due to their persistence in clinical settings, resulting in chronic infections [6]. There are 3 different mechanisms that form microbial aggregates, such as restricted motility, depletion of nutrients, and bridging by bacterial extracellular polymers. Additionally, biofilms can break off and release these non-attached aggregates post-dispersion. Dispersed cells differ phenotypically from both planktonic and sessile biofilm cells, often exhibiting increased adhesion, colonisation potential, virulence, and antibiotic tolerance [[13], [14], [15]]. Understanding dispersion is clinically significant, as biofilm fragments or aggregates released into surrounding environments can seed persistent or recurrent infections [16].

Staphylococcus epidermidis is extensively used in dispersion studies due to its clinical relevance as a major nosocomial pathogen that causes device-associated infections through biofilm formation, despite being a normal skin commensal [17,18]. Their EPS matrix demonstrates strain-dependent compositional variations, with poly-N-acetylglucosamine (PNAG) as a principal component, although certain strains like S. epidermidis ATCC 12228 produce PNAG-independent biofilms [19,20]. This strain employs alternative extracellular polymeric substances including proteins, extracellular DNA (eDNA), and GlcNAc-containing cell wall polymers, resulting in a matrix architecture that differs from canonical PNAG-rich biofilms and providing a suitable model to investigate non-canonical biofilm matrix organization and dispersion mechanisms. The choice of using S. epidermidis ATCC 12228 enables the exploration of phenotypic and matrix composition variability beyond conventional PNAG-dependent pathways. Consequently, the strain serves as an ideal model for studying dispersion mechanisms under various conditions, including different surface chemistries and hydrodynamic environments in microfluidic devices [21]. Despite progress in comprehending biofilm dispersion, the subsequent fate of cells, particularly the impact of their developmental stage on reattachment and secondary biofilm formation, remains inadequately investigated. This study investigates the influence of primary biofilm-associated cells (PBACs), referring to dispersed cells originating from established biofilms that often appear as single cells or small aggregates, derived from biofilms at different maturation stages, on the secondary biofilm development under continuous-flow conditions. Using confocal laser scanning microscopy (CLSM), COMSTAT analysis, and EPS composition profiling, we evaluated structural organization, reattachment efficiency, and matrix remodelling in secondary biofilms. Our study shows the correlation between the developmental stage of PBACs and the resultant secondary biofilm architecture and EPS composition, offering insights into how dispersed cells contribute to biofilm recurrence and persistence in clinically relevant S. epidermidis systems.

Materials and methods

2

Bacterial culture and maintenance

2.1

Bacterial cultures were preserved at −80 °C in a 25% (w/v) glycerol solution. The S. epidermidis ATCC 12228 thawed aliquots were streaked onto Luria Bertani (LB) agar plates for cultivation and incubated for 24 h at 37 °C. A single bacterial colony was used to inoculate 50 mL of sterile Tryptic Soy Broth (TSB) in a 250 mL Erlenmeyer flask and incubated at 37 °C, shaking at 200 RPM overnight (16 -18 h) to achieve an approximate optical density (OD 600 nm) of 2.0.

Biofilm growth

2.2

For primary biofilm

2.2.1

An Ibidi μ-Slide VI 0.4 (Ibidi GmbH, Gräfelfing, Germany) with six channels was employed for growing S.epidermidis biofilms under dynamic flow conditions [22]. The overnight culture was diluted to an optical density at 600 nm starting (OD 600 nm) of 0.02 using fresh 0.1x TSB media. The empty flow cell channel was inoculated with 200 μL of the diluted overnight culture and incubated at 37 °C for 1 h to facilitate the attachment of planktonic cells. Tubing, elbow joints, luer locks, and bubble traps were sequentially connected to the system, and a sterile syringe was filled with fresh 0.1x TSB media for use in the InfusionONE™ NE-300 syringe pump (New Era Pump Systems, Inc., Farmingdale, NY, USA). Diluted TSB (0.1x) was used to establish nutrient-limited conditions effective for biofilm formation under flow, better reflecting environmental and clinical setups with restricted nutrient availability [21,23]. The inlet of the flow cell channel was connected to the syringe containing 0.1x TSB media. After 1 h of attachment, fresh sterile 0.1x TSB media was flushed through the channel at a flow rate of 100 μL/min using the syringe pump to remove unattached cells. The outlet was connected to another tube that was collecting the dispersed effluents (biofilm-associated cells) in a sterile flask (Fig. 1). After rinsing the channel, the flow system was incubated at 37 °C for 24 h with continuous media circulation at a flow rate of 10 μL/min. To ensure that effluent samples contained only PBACs released within a specific time interval, the flow was briefly paused (<1 min) at the beginning of each collection window, the collection flask connected to the outlet tubing was replaced with a fresh sterile flask, and flow was immediately resumed. Effluent was then collected continuously over the designated intervals (6-8 h, 12-15 h, and 21-24 h). The same procedure was followed for each time window to avoid carryover between collection periods. Collected effluent samples were gently mixed and processed immediately for optical-density normalization and downstream analyses. All experiments were conducted in triplicate.Fig. 1. Schematic of the Ibidi μ-Slide VI 0.4 flow cell setup used for S. epidermidis biofilm growth under continuous, dynamic flow conditions. The primary components include a syringe pump, the media reservoir, the tubing assembly, and the collection flask for the dispersed effluent. The initial step involved inoculating the channels with overnight cultures. Subsequently, the media were purged, and the flow was maintained at a regulated flow rate of 10 μL/min to facilitate biofilm formation (Created with https://www.biorender.com/).Fig. 1

For secondary biofilms

2.2.2

The dispersed biofilm-associated cells mentioned above were collected at three distinct time points during the primary biofilm development. The three temporal phases were early stage (6-8 h), developing stage (12-15 h), and late stage (21-24 h). The durations for PBAC collection were determined based on practical factors derived from experimental observations. In the first stages of biofilm growth (6-8 h), detached cells were easily retrieved, facilitating the collection of representative PBAC populations within a reduced timeframe. In subsequent stages (12-15 h and 21-24 h), the detached entities exhibited greater heterogeneity and often contained smaller aggregates, necessitating extended collection durations to get adequate material for subsequent analysis. These three distinct cell populations were collected in sterile flasks and diluted to an optical density (600 nm) of 0.02 using fresh 0.1x TSB media. The identical setup and conditions outlined in section 2.2.1 were employed for the secondary biofilm growth within the Ibidi flow cells, utilising the syringe pumps. All experiments were conducted in triplicate.

Characterisation of inocula

2.3

The dispersed biofilm cells, collected from the primary biofilm, were diluted with fresh 0.1x TSB media, and the OD was maintained at 0.02. For live cell inoculum characterisation, the diluted cell suspension was stained with SYTO 9 (Thermo Fisher Scientific, Waltham, MA, USA). SYTO 9 was prepared as a working stock in sterile distilled water according to the manufacturer's instructions, and 10 μL of the working stock was added to 1 mL of the cell suspension. The stained suspension was incubated in the dark for 15 min prior to inoculation into an empty Ibidi flow cell for imaging. Z-stacks were acquired at five random positions per biological replicate (n = 3) for each condition within the flow cells using a Confocal laser scanning microscope, LSM 900 (Zeiss, Germany). The Zeiss Zen 3.10 software (Zeiss, Germany) was utilised to convert the z-stack into its maximum intensity projection for qualitative representation.

Cell viability assay

2.4

To evaluate the cell viability of dispersed biofilm cells harvested from the primary biofilm at various time intervals, the cells were subsequently maintained at an optical density (600 nm) of 0.02. The live cells were stained with Syto9 and propidium iodide (PI) as a counterstain for dead cells following the Live/Dead BacLight viability kit (Thermo Fisher Scientific, Waltham, MA, USA) manufacturer's protocol, with minor adaptations for microplate readout. The stain mixture was prepared in sterile distilled water using a 1:1000 dilution of each stain (v/v). 10 μL of Syto9/PI stain was added per 1 mL of each sample. The cells were incubated in the dark for 15 min. 200 μL of each sample was transferred to an empty 96-well plate, and using a microplate reader (SpectraMax iD3, Molecular Devices, USA), the OD was measured at two different wavelengths. To account for any unbound dye signal, dye only blanks were included, and fluorescence values were blank corrected prior to analysis. Unstained cell suspensions were included to assess background autofluorescence [24,25]. An excitation wavelength of 485 nm was used for both, while emission wavelengths of 530 nm and 630 nm were utilised for Syto9 and PI, respectively. All experiments were conducted in triplicate.

Growth kinetics

2.5

To compare the growth kinetics of biofilm-associated cells and planktonic overnight cells, the cell fractions were collected, and their OD was measured at 600 nm. The bacterial cultures were diluted to an OD of 0.02 with fresh 0.1x TSB media and were vortexed thoroughly for 15 min to break up small aggregates. 150 μL of fresh 0.1x TSB media was transferred to empty wells of a 96-well plate, and to each well, 10 μL of diluted cultures was added. The plate was incubated at 37 °C, shaking at 200 RPM for 24 h. OD at 600 nm was measured in each 1 h interval up to 24 h. Bacterial growth curve parameters were calculated from microplate reader data using the ipolygrowth R package [26]. Curves were fitted with an isotonic polynomial regression to model microbial growth dynamics without assuming a predefined parametric growth function. For each condition, biological replicates were fitted separately using the ipg_multisample function, with time and raw OD_600_ values as exported from the plate reader specified as input variables. The package extracted key kinetic parameters, including peak growth rate (maximum specific growth rate during exponential phase), peak growth time, doubling time (time required for bacterial density to double at peak growth rate), lag time (duration from inoculation to onset of exponential growth), and maximum OD (max y). These parameters were used to compare growth characteristics between planktonic controls and PBACs from biofilms of varying maturity. (Growth curve data were analysed using custom scripts, which are publicly available on GitHub https://github.com/JacobusBrink/fragmented_but_functional).

Live cell imaging for tracking the biofilm development over time

2.6

Live imaging was performed by placing the flow cell setup (mentioned in 2.2.1.) inside the CLSM 900 incubation chamber, where the flow cell system was connected to a syringe pump delivering 0.1x TSB medium at a constant flow rate of 10 μL/min for 24 h. Biofilms were stained with Syto9 added to the growth medium for live cell visualization. SYTO 9 (Thermo Fisher Scientific, Waltham, MA, USA) was prepared as a working stock in sterile distilled water according to the manufacturer's instructions, and 5% (v/v) of that stock solution was added to the media for continuous flow during imaging. The microscope's imaging parameters were configured to capture z-stacks at 15-min intervals. The laser power was set at the lower optimum to avoid photobleaching over time. The resulting time-lapse images were processed using ImageJ [27], Zeiss Zen 3.10, and COMSTAT software [28,29] for quantitative analysis of biomass and surface area coverage over time, providing insights into biofilm growth and dynamics under flow conditions.

Biofilm characterisation using confocal laser scanning microscopy

2.7

Morphological variations and COMSTAT image analysis

2.7.1

To observe the possible morphological variations between primary and secondary biofilms, the biofilms after 24 h were rinsed once with sterile distilled water to eliminate unattached planktonic cells at a slightly higher flow rate of 15 μL/min. After rinsing, a 5 mL syringe was employed to administer the Syto9/PI stain through the channel at a flow rate of 10 μL/min. The Syto9/PI staining solution was prepared in sterile distilled water using a 1:1000 dilution of each dye, consistent with the manufacturer's recommendations. Based on the tubing length and internal diameter of the flow-cell system, the stain retention time within the channel was calculated, and flow was maintained for this duration to ensure complete exposure of the biofilm to the stains. The biofilms were rinsed again post-staining with sterile distilled water to remove excess or unbound stains. Using the CLSM 900, images and Z-stacks were taken at both 10 × and 40 × magnification with appropriate acquisition settings. The stained samples were imaged using excitation/emission filters suitable for Syto9 (excitation 483 nm ∼ emission 503 nm) and PI (excitation 493 nm ∼ emission 636 nm). Detector gain and laser power were optimised to avoid signal saturation, and all imaging parameters, including pinhole size and detector settings, were kept constant across samples. Image analysis and visualization were performed using Fiji/ImageJ, Imaris 10.1.1, and Zeiss Zen 3.10 software. All experiments were conducted in triplicate.

EPS matrix characterisation with staining

2.7.2

To stain the multiple components of the EPS matrix of biofilms, various stains alone and in combination were used. SYTO 9 was employed to visualize cell-associated biomass, serving as a reference for bacterial distribution within the biofilm, whereas WGA was utilised to label polysaccharide components of the EPS matrix. The biofilms grown inside a flow cell were stained with a combination of Syto9 and Wheat germ agglutinin (WGA) Alexa flour 647 (Thermo Fisher Scientific, USA) was prepared in sterile distilled water. Both stains were prepared in sterile distilled water according to the manufacturer's instructions and introduced into the flow cell using a 5 mL syringe and syringe pump as described in Section 2.7.1. The biofilms were exposed to the stains at a final concentration of 10 μg/mL for each stain and were rinsed with sterile distilled water to remove excess and unbound stain prior to imaging. For visualization of bacterial cells and polysaccharide in the biofilm, stained samples were imaged using CLSM 900 with excitation/emission filters suitable for Syto9 (excitation 483 nm ∼ emission 503 nm) and WGA (excitation 650 nm ∼ emission 665 nm). Z-stacks were taken at both 10 × and 40 × magnification with appropriate acquisition settings. Percentage EPS distribution was calculated from 3D CLSM z-stacks by normalising the WGA-labelled polysaccharide biovolume to the total biofilm biovolume obtained using COMSTAT analysis and expressing the value as a percentage. Image analysis and visualization were performed using Fiji/ImageJ, Imaris 10.1.1 (Oxford Instruments), and Zeiss Zen 3.10 software. All experiments were conducted in triplicate.

Statistical analysis

2.8

All experiments were conducted with a minimum of three independent biological replicates, each comprising triplicates unless specified otherwise. Statistical analyses were conducted using GraphPad Prism version 10.6.0. One-way analysis of variance (ANOVA) was performed to evaluate statistical significance (p < 0.05). Tukey's multiple comparison post-hoc test was applied to determine pairwise significance. To ensure the reproducibility of image-based analyses (CLSM/COMSTAT), all images were processed using consistent automatic thresholding parameters applied across the dataset to ensure comparability. Images were analysed independently, and results were averaged across biological replicates. These clarifications ensure transparency and reproducibility of our imaging analysis.

Results

3

Primary biofilm-associated cells maintain consistent viability across developmental stages

3.1

To evaluate the impact of age on the viability of primary biofilm-associated cells (PBACs), live/dead assays were conducted at three specific time intervals: 6-8 h, 12-15 h, and 21-24 h. Quantitative analysis via a microplate reader revealed no significant differences in normalised live/dead cell ratios among the three cell populations (Fig. 2A). This suggests that the inoculum used to initiate secondary biofilms originates from PBACs with similar viability profiles, providing a consistent baseline for assessing differences in structural and functional biofilm outcomes. Qualitative assessment from fluorescence microscopy corroborated these findings, with Syto9-stained (green) viable cells and small aggregates visible across all conditions (Fig. 2B), suggesting that PBACs maintain viability irrespective of biofilm age. These results confirm that the inoculum used to initiate secondary biofilms maintains consistent viability across all maturation stages, eliminating cell viability as a confounding factor.Fig. 2. Viability of Primary Biofilm-Associated Cells (PBACs) at Defined Time Points (A) Quantification of live/dead ratios of PBACs assessed using Syto9 and propidium iodide staining followed by OD 600nm measurement in a microplate reader across 6-8 h, 12-15 h, and 21-24 h time points. No statistically significant differences in viability among populations were observed (B) Representative fluorescence microscopy images showing green (Syto9 positive) viable cells and small aggregates in all PBAC populations. (Scale bar represents 10 μm) Experiments performed in triplicate; data presented as mean ± SD. Statistical significance is indicated as follows: n.s., not significant. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 2

PBACs exhibit enhanced growth kinetics and reduced lag phase compared to planktonic cells

3.2

To evaluate the post-dispersion growth potential of biofilm-associated cells, PBACs collected at early (6-8 h), developing (12-15 h), and late (21-24 h) stages were cultured with planktonic cells and observed over 24 h. Growth kinetic profiles demonstrated notable disparities in growth patterns between PBACs and planktonic populations (Table 1). Planktonic cells demonstrated the longest doubling time of 36.5 min and a notable lag phase of approximately 2.2 h. Conversely, PBACs exhibited markedly reduced doubling times, declining from 24.5 min (6-8 h) to 21.3 min (21-24 h) (Supplementary Table 1). Significantly, PBACs from the 12-15 h and 21-24 h biofilms displayed no detectable lag phase, signifying a prompt shift into exponential growth following inoculation. The statistical analysis demonstrated that all PBAC groups exhibited significantly shorter doubling times in comparison to planktonic cells (p < 0.05) (Fig. 3). Peak growth rates in PBACs (0.028-0.033 min^−1^) exceeded those in planktonic cultures (0.019 min^−1^), indicating improved metabolic priming. A notable difference in the maximum OD values at 600 nm was observed across all groups. This suggests differences in nutrient utilisation dynamics or stress adaptation strategies associated with biofilm dispersal history. Collectively, these results indicate that PBACs, especially those originating from developing-to-late-stage biofilms, exhibit enhanced metabolic ease and a distinct growth advantage following dispersion, compared to planktonic cells. Altogether, these findings highlight that PBACs, especially from developing-to-late-stage biofilms, exhibit enhanced metabolic priming and rapid regrowth potential compared to planktonic cells.Table 1. Growth kinetics profile of planktonic cells and biofilm-associated cells (PBACs) harvested at different biofilm maturation stages.Table 1. ConditionsMax specific growth rate (min^−1^)Lag time (hours)Max OD_600nm_Planktonic cells0.0192.170.1606-8H PBACs0.0282.420.21512-15H PBACs0.0310.000.02621-24H PBACs0.0330.000.210Note: Growth parameters derived from OD_600nm_ measurements over 24 h for planktonic cells and PBACs from early (6-8 h), developing (12-15 h), and late stage (21-24 h) biofilms. Growth parameters, including max specific growth rate, lag time, and maximum OD 600nm, are presented in tabular format (Refer to Supplementary Figs. 1, 2, 3, & 4 for individual growth curves).Fig. 3. Comparison of (A) growth kinetics and (B) doubling times across planktonic cells and biofilm-associated cells (PBACs) harvested at different biofilm maturation stages. PBACs displayed significantly shorter doubling times relative to planktonic cells. Data represent mean ± SEM from biological replicates (n = 3). Statistical significance was assessed using one-way ANOVA with Tukey's post hoc test and is indicated as follows: n.s., not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.Fig. 3

Biofilm characterisation using confocal laser scanning microscopy

3.3

Overtime tracking of primary biofilm development using CLSM live imaging and COMSTAT analysis

3.3.1

To understand the dispersion dynamics in a biofilm, time-lapse CLSM imaging and COMSTAT quantification were used to track primary biofilm development over 24 h under continuous flow. Total surface area and biomass values were calculated from z-stacks using COMSTAT at regular intervals. As shown in Fig. 4, both biomass and surface area showed an initial decrease during the first 12 h, followed by a sharp increase after 16 h with maximum values recorded at 24 h, suggesting active biofilm maturation. A transient reduction of biomass and surface area was observed around 12 h (Supplementary video S1), which served as the basis for selecting the three time points in the developmental stage (early 6-8 h, developing 12-15 h, and late stage 21-24 h). This transition suggests a critical shift towards structural consolidation and maturation during the late growth phase and supports the selected time points for PBAC harvesting. CLSM images (Fig. 4C) revealed a clear transition from sparsely distributed microcolonies at 6-8 h to dense, multilayered biofilm structures at 21-24 h, indicating progressive biofilm maturation.Fig. 4. Time-lapse CLSM imaging and COMSTAT analysis showing (A) Surface area (μm^2^) and (B) Biomass (μm^3^/μm^2^) progression over 24 h duration from z-stacks captured at regular intervals. COMSTAT analysis revealed an initial decline in both biomass and surface coverage until 12 h, followed by a marked increase after 16 h, with peak values at 24 h. Data represent mean ± SD from three biological replicates. CLSM live imaging was performed using Syto9 staining. (C) Representative CLSM images illustrating structural progression of primary biofilms at early (6-8 h), developing (12-15 h), and late (21-24 h) stages. Orthogonal XZ and YZ projections demonstrate vertical biofilm architecture. Biofilms were stained using Syto9. Scale bars represent 20 μm.Fig. 4

Supplementary video related to this article can be found at https://doi.org/10.1016/j.bioflm.2026.100353

The following is/are the supplementary data related to this article:Multimedia component 2Multimedia component 2

Biofilm architecture and developmental dynamics differ between primary and secondary biofilms

3.3.2

To understand the morphological variations, a confocal laser scanning microscope (CLSM) was used to visualize primary and secondary biofilms (Fig. 5). Primary biofilm exhibited relatively uniform vertical development, with cells densely packed into a structured matrix. These biofilms showed consistent microcolony formation, characteristic of early colonisation and stable attachment under continuous flow. In contrast, secondary biofilms developed from PBACs of increasing age displayed notable structural differences. CLSM acquired Z-stacks, and orthogonal projection showed that secondary biofilms formed from early stage (6-8 h) PBACs were thinner, sparsely distributed, lacked vertical stratification, and had less clustering. Biofilms developed from 12 to 15 h PBACs have moderate cell aggregation. The most compact structure was observed in biofilms developed from 21 to 24 h PBACs, exhibiting dense, cohesive microcolonies, robust vertical growth, and a uniform biomass distribution that exceeded that of primary biofilms. The visual observations were subsequently corroborated by quantitative image analysis.Fig. 5. Orthogonal CLSM Projections of Primary and Secondary Biofilms developed from PBACs at Different Stages. On the bottom panel of each image (XZ), horizontal cross-sections and on the right-side panel (YZ), vertical cross-sections are mentioned. Images display differences in microcolony density, vertical stratification, and distribution across biofilm types. Secondary biofilms from later-stage PBACs show denser and more uniform structures. (Scale bar represents 10 μm).Fig. 5

Quantitative image analysis using COMSTAT supported the observed structural variations between the primary and secondary biofilms (Fig. 6). Secondary biofilms from older PBACs exhibited significantly increased biovolume, mean thickness, and average thickness (p < 0.0001). Roughness coefficient decreased with PBAC age, indicating formation of smoother, more uniformly distributed structures. Additionally, the surface-to-biovolume ratio decreased in later-stage PBACs, suggesting compactness inside the biofilms. Nonetheless, surface area coverage markedly increased in older PBAC-derived biofilms, indicating enhanced reattachment efficiency and spatial colonisation. The progressive enhancements in structural complexity among secondary biofilms indicate that PBACs from late stage biofilms possess an enhanced capacity to rapidly establish robust structures upon reattachment, allowing them to reinitiate structurally competent biofilms more efficiently than their early-stage counterparts. Our findings demonstrate the development of biofilm structural complexity; however, the molecular or cellular mechanisms underlying this remain unidentified.Fig. 6. Quantitative (COMSTAT) comparison of key architectural features of primary and secondary biofilms developed from PBACs of different maturation stages. (A) Biomass, (B) Maximum thickness, and (C) Average thickness all increased significantly with PBAC age, indicating denser and more developed biofilm structures. (D) Roughness coefficient decreased in later-stage PBAC biofilms, suggesting more uniform architecture. (E) Surface area coverage increased, indicating improved surface colonisation in biofilms formed from late stage PBACs. (F) Surface-to-biovolume ratio declined with PBAC maturation, reflecting compactness. Data represent mean ± SD (n = 3). Statistical significance was determined by one-way ANOVA and is indicated as follows: n.s., not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.Fig. 6

PBAC maturation stage influences polysaccharide accumulation in the secondary biofilm matrix

3.3.3

To assess the EPS matrix distribution of primary and secondary biofilms, polysaccharide-specific staining was conducted utilising wheat germ agglutinin (WGA). The maximum intensity projection of acquired z-stacks demonstrated a progressive increase in polysaccharide signal intensity correlating with PBAC maturation (Fig. 7A), signifying improved matrix development in biofilms derived from older PBACs. Quantitative analysis corroborated this trend, revealing a substantial increase in polysaccharide distribution within secondary biofilms formed from 12 to 15 h and 21-24 h PBACs (p < 0.0001), in contrast to early-stage PBACs and primary biofilms (Fig. 7B). The findings indicate that the enhancement of EPS polysaccharides is associated with PBAC age, potentially leading to increased reattachment efficiency and improved biofilm structure.Fig. 7. Polysaccharide composition of primary and secondary biofilms formed by PBACs of increasing maturation. (A) Maximum intensity projection of a confocal stack of Syto9 (blue signal) and WGA-stained biofilms showing enhanced polysaccharide signal (red) with PBAC age. (Scale bar represents 20 μm) (B) Quantitative analysis of percentage of polysaccharide distribution in EPS matrix, with significant increases in secondary biofilms from 12 to 15 h and 21-24 h PBACs compared to primary and early-stage PBACs. Data represent mean ± SD (n = 3). Statistical significance determined using one-way ANOVA. Statistical significance is indicated as follows: n.s., not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)Fig. 7

Discussion

4

Dispersion is mediated by a combination of biological, physical, and chemical factors [9,11,[30], [31], [32]]. Dispersed cells have unique behaviour relative to their planktonic equivalents, characterized by increased adhesion, colonisation, and pathogenicity [15,33]. These cells can more efficiently colonise surfaces and evade host immune responses, providing advantages during the early stages of infection [15]. The mode of dispersion, whether active or passive, influences the properties of dispersed cells, particularly their antibiotic susceptibility [34]. Comprehending the mechanisms and catalysts of biofilm dispersion is essential for developing strategies to mitigate biofilm relocalization onto novel surfaces [35].

The influence of age on the phenotypic traits of dispersed cells and their subsequent effects on reattachment and secondary biofilm formation is inadequately defined. This study examines the phenotypic and structural characteristics of primary biofilm-associated cells (PBACs) at different maturation stages and their role in the reestablishment of secondary biofilms. Our findings demonstrate that released PBACs from S. epidermidis ATCC 12228 retain a form of developmental memory that influences their ability to establish secondary biofilms. In the context of this study, “biofilm developmental memory” refers to the persistence of biofilm-associated phenotypic traits in released PBACs, whereby the physiological state acquired during residence within a biofilm influences subsequent growth behaviour and secondary biofilm formation. PBACs, unlike planktonic cells, retain physiological and structural traits from the primary biofilm phase, leading to developmental stage-specific variations in reattachment, growth kinetics, and extracellular matrix composition. Time-lapse CLSM microscopy combined with COMSTAT analysis of primary biofilms under continuous flow revealed a unique developmental trajectory. A fluctuation in biomass and surface area was observed with a transient decrease around 12 h followed by a noticeable rise after 16 h, reaching peak values at 24 h, indicating a dispersion-like event. This trajectory facilitated the selection of three developmental time points (early, developing, and late stage) for the harvesting of PBACs and subsequent analysis. This phase likely indicates a dynamic reorganization within the primary biofilm, characterised by the active shedding of microcolonies as part of early biofilm restructuring, a strategy to facilitate surface reattachment and spatial expansion. According to Ref. [36], biofilm dispersion is a regulated biological process that cannot be deduced purely from overall reductions in biomass, which may also result from physical detachment or structural reorganization. Dispersion frequently occurs as limited microcolony relocation instead of significant biomass depletion. The observed change is thus regarded as exhibiting dispersion-like or detachment-associated behaviour, in alignment with existing biofilm lifecycle models. Biofilm dispersion has been demonstrated to be induced in Pseudomonas putida, P. aeruginosa, Vibrio cholerae, and Staphylococcus aureus by nutrient depletion, such as glucose or carbon source exhaustion [37,38]. This results in significant decreases in biofilm biomass and an increase in cell detachment. Furthermore, changes in pH, especially those brought on by metabolic byproducts building up within the biofilm, can cause biofilms to spread, as higher acidity in biofilms destabilizes matrix integrity and encourages cell release [39,40]. Overall, the accumulation of quorum-sensing molecules, pH shift, and nutrient depletion are some of the well-established environmental triggers that can be related to the 12 h dispersion like event observed here in the S. epidermidis biofilms. The temporal progression and flow-based biofilm development model utilised by Ref. [33] revealed that in Candida albicans biofilms, a comparable pattern was noted, with the highest cell dispersion occurring between the initial 5-12 h of growth, consistent with our findings. Ultimately, due to the re-establishment of the matrix during the 20-24 h growth period, the quantity of dispersed cells diminished. This further substantiates our hypothesis that dispersion is synchronised with biofilm developmental stages, and that dispersal events during the early, developing, and late stages may produce functionally and structurally distinct populations.

Cell viability assays confirmed that PBAC inocula were standardised across developmental stages, with no significant differences in live/dead ratios, indicating that variations in secondary biofilm formation were not driven by viability differences but by stage-dependent physiological traits. Morphological analysis showed that early-stage PBACs were mainly single cells, whereas developing and late stage PBACs formed small aggregates, which may enhance surface reattachment and microcolony development. These findings corroborate earlier research [41], indicating that dispersed biofilm cells from Candida albicans maintain a biofilm-primed phenotype, thereby promoting rapid growth upon reattachment. Growth kinetics analysis revealed stage-dependent differences among PBACs. Cells derived from developing and late-stage biofilms exhibited shorter doubling times, higher growth rates, and an absence of a detectable lag phases compared to planktonic cells. The lack of a detectable lag phase in these PBACs indicates heightened metabolic priming or physiological preconditioning, presumably inherited from their biofilm origin. In contrast, planktonic cells and early-stage PBACs displayed slower growth initiation. These observations are consistent with previous studies on Pseudomonas species showing that the use of different inocula can significantly influence biofilm growth dynamics, including lag phase duration in flow cell systems [42]. The extended stationary phase observed in PBACs may reflect improved resilience or altered nutrient utilisation. Similar observations in dispersed cells of P. aeruginosa and Candida albicans have demonstrated enhanced motility, nutrient acquisition, and adaptive capabilities linked to distinct gene expression patterns [41,43]. While this study focused on phenotypic and growth characteristics, previous work indicates that biofilm dispersal is regulated by transcriptional networks involving quorum sensing, cyclic-di-GMP signalling, and regulatory RNAs, which enable dispersed cells to adapt rapidly to new environments [41,44,45]. Future studies integrating transcriptomic or proteomic analyses would help validate these regulatory processes.

The development from primary to secondary biofilms signifies an essential phase in the biofilm lifecycle, during which dispersal and reattachment capabilities are modified [46,47]. Confocal microscopy and COMSTAT analysis revealed morphological and structural distinctions between primary and secondary biofilms. Qualitative imaging revealed that the primary biofilms exhibited stable early-stage adhesion and consistent microcolony formation. The images indicated that the primary biofilm resembled those formed by late stage PBACs, exhibiting denser, thicker, and more spatially organised structures in contrast to early and developing PBACs. Distinct structural differences were observed in primary and secondary biofilms. The COMSTAT analysis corroborated the qualitative image analysis, revealing distinctions between the primary and secondary biofilms. The developmental stage of the PBACs significantly influenced the structure of secondary biofilms. In comparison to the primary biofilm, secondary biofilms derived from 21 to 24 h PBACs demonstrated similar or improved structural integrity, especially regarding thickness and biomass. Quantitative parameters, including biomass, surface area, average thickness, and maximum thickness, progressively increased with PBAC maturation (6-8 h < 12-15 h < 21-24 h). Early-stage PBACs, although functional, produced thinner and more spatially dispersed biofilms with reduced surface area coverage. The roughness coefficient, reflecting structural heterogeneity, decreased with the advancement of PBAC age, indicating more uniform and cohesive architectures in biofilms formed from late stage PBACs. Correspondingly, the surface-to-biovolume ratio diminished throughout the PBAC maturation stages, aligning with a shift from loosely arranged monolayers to dense, multilayered microcolonies. The surface area coverage, indicative of attachment efficiency, was significantly higher in biofilms derived from older PBACs, thereby emphasising their enhanced reattachment potential. These findings collectively indicate that dispersed cells signify a distinct developmental phase, likely prepared for surface reattachment and biomass accumulation, exhibiting unique phenotypic characteristics relative to their planktonic equivalents. A study by Kalia et al. [43] corroborates this concept, demonstrating that in P. aeruginosa, dispersed cells exhibit unique gene expression patterns within minutes of detaching from biofilms, including an increase in virulence proteins and swimming motility, which may facilitate rapid re-attachment.

The analysis of EPS matrix composition provided additional mechanistic understanding of the structural distinctions identified between primary and secondary biofilms. CLSM imaging with WGA staining showed stage-dependent increases in GlcNAc-associated polysaccharide content, with the strongest signal observed in biofilms derived from late stage PBACs. A study by Ref. [48] demonstrated that the surface polysaccharide Poly-N-acetylglucosamine (PNAG) is crucial for the formation of the biofilm extracellular polymeric substance matrix in S. epidermidis, in conjunction with extracellular DNA (eDNA) and surface proteins. While WGA is commonly associated with PNAG detection, it can also bind other GlcNAc-rich polymers; therefore, in S. epidermidis ATCC 12228, which lacks the icaADBC operon required for PNAG synthesis, the signal likely reflects alternative cell wall glycopolymers [[48], [49], [50]]. The findings were quantitatively consistent, indicating improved matrix development during reattachment. Secondary biofilms formed from early PBACs showed weaker and more dispersed staining, corresponding to thinner and less cohesive structures. Increased polysaccharide accumulation in late stage PBAC-derived biofilms likely contributes to enhanced structural stability, cell adhesion, and stress resistance, consistent with previous findings linking EPS composition to biofilm integrity [48,51].

Collectively, these findings highlight that the developmental state of PBACs influences both the structural characteristics of secondary biofilms and alterations in polysaccharide distribution within the biofilm matrix. Late stage PBACs demonstrate a preconditioned capacity for rapid and dense matrix production, facilitating the development of more stable, intricate biofilms that are comparable to or surpass their primary counterparts. In contrast, early PBACs produce less organized biofilms with weaker matrix scaffolds, which restricts their architectural development following reattachment. Comprehending these structural variations is crucial for addressing biofilm resilience and inhibiting recolonization in clinical and industrial settings. In an ecological and evolutionary framework, biofilm dispersion can be interpreted as a source-sink dynamic, where late stage biofilms function as stable “source” populations that release phenotypically diverse spreaders capable of colonising transient or challenging “sink” habitats. The observed dispersal heterogeneity across biofilm maturation stages can also be related to the bet-hedging strategy, a recognised microbial survival mechanism. This strategy enables the population to diversify its phenotypic outputs, thereby optimising fitness in unpredictable environments [44].

Limitations

4.1

While the novel insights are noteworthy, it is essential to recognize several limitations of this study. This study was conducted using a single clinical isolate, Staphylococcus epidermidis ATCC 12228; therefore, the observed phenotypic and structural characteristics represent strain-specific behaviour and may not be directly generalizable across other S. epidermidis isolates. In addition, extracellular matrix characterization was focused on polysaccharide-associated components, and other matrix constituents such as proteins and extracellular DNA were not quantitatively resolved. The strain utilised in this study, S. epidermidis ATCC 12228, may demonstrate strain-specific EPS compositions and may lack canonical PNAG synthesis pathways, potentially restricting direct comparisons to clinical isolates. Incorporating PNAG-producing strains, such as clinical isolates possessing the ica operon, into future studies could further substantiate the functional significance of polysaccharides in secondary biofilm architecture and antibiotic tolerance.

Concluding remarks

5

This study shows that primary biofilm-associated cells (PBACs) actively influence the development of secondary biofilms by retaining a developmental “memory” of the primary biofilm. This memory appears to encode both structural and metabolic traits, including rapid regrowth capacity and matrix remodelling. Notably, these traits distinguish PBACs from naive planktonic cells and may underlie their ability to rapidly establish stable secondary biofilms. The findings indicate that dispersal is not a neutral reset of planktonic physiology; instead, it represents a transitional state with significant functional implications for subsequent recolonization. The structural maturation observed in secondary biofilms seeded with late-stage PBACs may resemble clinical situations in which biofilm fragments disperse from chronic infection sites and reattach in different locations, thereby contributing to persistent or recurrent infections. Strategies aimed at biofilm dispersal should account for both the release of cells and their stage-dependent readiness for reattachment and regrowth. Future research will focus on investigating the transcriptomic and proteomic profiling of PBACs at various developmental stages, alongside enhanced real-time imaging of reattachment dynamics.

CRediT authorship contribution statement

Dibyangshee Singh: Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Jacobus Brink: Writing – review & editing, Formal analysis, Data curation. Dishon Wayne Hiebner: Writing – review & editing, Supervision, Methodology, Formal analysis, Conceptualization. Eoin Casey: Writing – review & editing, Visualization, Supervision, Resources, Project administration, Methodology, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.