Novel Skin- and Oral-Derived Probiotic Candidates: Functional Evaluation and Application Perspectives

TL;DR

This study identifies new probiotic candidates from skin and oral sources with potential health benefits beyond the gut.

Contribution

The study evaluates novel skin- and oral-derived probiotic isolates for functional properties relevant to probiotic applications.

Findings

Lcb. rhamnosus S3 showed the highest probiotic potential with strong antimicrobial activity and survival in simulated conditions.

Limosilactobacillus sp. (S1) exhibited the strongest antimicrobial activity against C. acnes and high adhesion to skin cells.

S. cerevisiae isolates showed strong autoaggregation and antioxidant capacity but poor freeze-drying resistance.

Abstract

The skin and oral environment represent complex microbial ecosystems that host diverse bacterial communities with potential health-promoting properties beyond the gastrointestinal tract (GIT). In this study, four bacterial and three yeast isolates were obtained from saliva (S1, S3, S5, and S6) and human skin (A1, A2, and A3) and subjected to identification and functional characterization. Phenotypic identification by API and MALDI-TOF mass spectrometry identified bacterial isolates as Limosilactobacillus sp. (S1) and Lacticaseibacillus rhamnosus (S3, S5, and S6), while the yeasts were identified as Saccharomyces cerevisiae (A1, A2, and A3). The isolates were evaluated for their functional properties, including antimicrobial activity, autoaggregation, antioxidative potential, resistance to freeze-drying, survival in simulated saliva and GIT conditions, adhesion to Caco-2 and HaCaT cell lines, and biofilm-forming ability. Lcb. rhamnosus S3 demonstrated the highest probiotic potential, characterized by strong inhibition of S. aureus, high autoaggregation capacity, substantial survival following freeze-drying, and good tolerance to simulated saliva and GIT conditions. Limosilactobacillus sp. (S1) demonstrated the strongest antimicrobial activity against C. acnes and the highest adhesion capacity to HaCaT cells, indicating its suitability for topical dermatological applications. Although S. cerevisiae isolates did not exhibit antimicrobial activity, they showed strong autoaggregation and notable antioxidant capacity. However, their low resistance to freeze-drying limits their applicability in probiotic formulation development.

Figures

Full text

1. Introduction

Probiotics are defined as live microorganisms that, when administered in adequate amounts, confer health benefits to the host by modulating the microbiome and influencing immune, metabolic, and barrier functions across various body sites [1]. While probiotics have traditionally been associated with gut health, recent years have witnessed a surge of interest in their broader therapeutic potential, revealing impacts that encompass dermatology, oral health, immunomodulation, metabolic balance, and even mental health [2,3,4]. The gut–skin and oral–skin axes are now central to current research, as scientific advances uncover complex interactions between local and systemic microbiota and their roles in disease prevention, inflammation reduction and repair mechanisms. Mechanistically, probiotics act through competitive exclusion of pathogens, production of antimicrobial substances, modulation of local and systemic immune responses, enhancement of barrier integrity and maintenance of microbial diversity [5,6]. Recent clinical and experimental studies indicate notable effects of probiotics on skin health, not only when administered orally through food or supplements but also via topical formulations that directly target the skin’s microbiota [7]. Evidence indicates that both oral and topical probiotics, particularly strains from the genera Lactobacillus and Bifidobacterium, can reduce inflammation, promote re-epithelialization, combat oxidative stress and enhance collagen synthesis and hydration, thereby benefiting conditions such as acne, atopic dermatitis, eczema and wound healing [5,6,8,9]. Furthermore, certain lactobacilli isolated from traditional dairy products demonstrated promising anti-oral cancer properties in vitro, with specific strains exhibiting significant inhibition of oral carcinoma cell lines through induction of apoptosis, while showing minimal cytotoxicity toward normal cells, suggesting therapeutic potential for oral health applications [10]. Probiotic interventions have emerged as promising alternatives or adjuncts to conventional therapies, often with improved tolerability and a lower risk of adverse effects, as topically applied probiotics exhibit minimal systemic absorption and oral intake modulates the gut–skin axis to reduce systemic inflammation [2,5]. Despite these advances, current probiotic formulations overwhelmingly rely on strains traditionally isolated from the gastrointestinal tract (GIT), fermented foods, or well-characterized commercial sources. However, the human skin and oral environment represent underexplored reservoirs of commensal bacteria, many of which demonstrate unique adaptability to local microenvironmental conditions, such as low moisture, variable pH, distinct substrates and dynamic host–microbe interactions [11,12]. Emerging research has begun to investigate the indigenous microbiota of the skin and oral surfaces as sources of novel probiotic candidates, revealing their capacity for competitive inhibition of pathogens, stimulation of local immunity, and biotransformation of topical compounds [5,6]. Harnessing these strains could enable therapies precisely tailored to the needs of individual skin and oral diseases, including inflammatory disorders, infections, impaired barrier syndromes, and mucosal pathologies. Ultimately, isolating and characterizing new probiotic strains from the skin and oral environment opens promising avenues for skin-targeted and mucosal therapeutics, suggesting that these habitats serve as relatively untapped but highly promising sources of beneficial microbes for human health applications [13,14]. Further investigation into their mechanisms, safety profiles, and therapeutic effectiveness will be vital for next-generation probiotic development and translational medicine.

Therefore, the aim of this work was to isolate and characterize potential probiotic bacteria, primarily lactic acid bacteria (LAB), and yeasts from the human microbiome, specifically from the skin surface and saliva. After isolation and phenotypic identification of the isolates, their probiotic potential was examined, including assessment of their antimicrobial and antioxidant activity, autoaggregation, ability to survive freeze-drying, simulated saliva and GIT conditions, biofilm formation, and in vitro adhesion to Caco-2 and HaCaT cells. The results of this study provide a foundation for future health research and the development of potential probiotic products for human use, particularly aimed at alleviating skin and oral diseases.

2. Materials and Methods

2.1. Isolation of Microorganisms from Skin and Saliva Samples

Samples were collected from five healthy adult female volunteers aged 18–60, after obtaining informed consent. Participants were in good health, with no history of antibiotic therapy, skin injuries, or diseases. Participants were advised not to shower for at least 48 h before sampling to preserve the natural skin microbiota [15]. Samples were collected from the forearm, lower back, oral cavity and saliva. Sterile cotton-tipped plastic swabs moistened with sterile saline were used to sample an area of approximately 4 cm^2^ by applying 10 horizontal and 10 vertical strokes. After sampling, the swab tips were aseptically cut with sterilized scissors and placed into test tubes containing MRS (De Man–Rogosa–Sharpe; Biolife, Milan, Italy) or malt extract (Biolife, Milan, Italy) broth, and incubated under anaerobic conditions for 24 h at 37 °C or under aerobic conditions for 24 h at 28 °C, respectively. Following this initial enrichment, aliquots from the incubated broths were streaked onto solid MRS (Biolife, Milan, Italy) and malt extract (Biolife, Milan, Italy) agar plates, which were then incubated under the same conditions. Distinct individual colonies that developed on the plates were subcultured into fresh MRS and malt extract broths. To obtain pure isolates, the cultures were again streaked onto MRS and malt extract agar plates. Bacterial and yeast isolates were stored at −80 °C in MRS broth or malt extract broth, respectively, supplemented with 15% glycerol.

Approval for the collection and handling of the aforementioned strains was obtained from the Ethics Committee of the University of Zagreb School of Medicine (Ref. No.: 251-59-10106-25-111/44, Class: 641-01/25-02/04, date of approval: 17 April 2025).

2.2. Identification of Isolates

2.2.1. Phenotypic Identification of Isolates

Four bacterial isolates were characterized using Gram staining, KOH, catalase, and API 50 CHL assays (BioMérieux, Marcy l’Etoile, France), while three yeast isolates were characterized using simple staining and the API 20 C AUX assay (BioMérieux, Marcy l’Etoile, France).

2.2.2. MALDI-TOF Mass Spectrometry Identification

Following preliminary phenotypic identification using the API system, bacterial and yeast isolates were further identified by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry on a Microflex Smart LS MALDI-TOF MS system (Bruker Daltonics, Bremen, Germany) according to the manufacturer’s standard identification protocol. Briefly, mass spectra were generated based on the mass-to-charge (m/z) ratios of predominantly ribosomal proteins and subsequently compared with reference spectra contained in the Bruker Biotyper database(MBT Compass 4.1. software). Identification was assigned by calculating log(score) values reflecting the similarity between sample spectra and reference entries [16]. Interpretation of identification results followed the criteria recommended by the manufacturer: log(score) values < 1.70 were considered unreliable for identification (no identification possible); values between 1.70 and 1.99 indicated identification at the genus level (low-confidence identification); and values ≥ 2.00 indicated highly probable genus-level identification and probable species-level identification (high-confidence identification).

2.3. Antibiotic Susceptibility Testing

Susceptibility to nine antibiotics was assessed using the disk diffusion method described by Matuschek et al. [17]. Briefly, 100 μL of overnight bacterial isolates were inoculated onto MRS agar. BD BBL Sensi-Disc filter paper discs (6 mm diameter; BD Diagnostics, Franklin Lakes, NJ, USA) containing defined concentrations of antibiotics (ampicillin 10 μg, erythromycin 15 μg, gentamicin 10 μg, clindamycin 2 μg, chloramphenicol 30 μg, kanamycin 30 μg, streptomycin 10 μg, tetracycline 30 μg, and vancomycin 30 μg) were placed on the agar surface. After overnight incubation at 37 °C, isolates were classified as susceptible or resistant based on the presence or absence of a visible inhibition zone.

2.4. Antimicrobial Activity

For the test, microorganisms used to assess the antimicrobial activity of the isolates were obtained from the DSMZ (The Leibniz Institute DSMZ German Collection of Microorganisms and Cell Cultures GmbH, Braunschweig, Germany), ATCC (American Type Culture Collection), and JCM (Japan Collection of Microorganisms) microbial repositories and are part of the microbial collection of the Laboratory for General Microbiology and Food Microbiology at the University of Zagreb Faculty of Food Technology and Biotechnology. The test microorganisms included Escherichia coli ATCC^®^25922^TM^, Staphylococcus aureus ATCC^®^25923^TM^, Salmonella typhimurium ATCC^®^29631^TM^, Listeria monocytogenes ATCC^®^23074^TM^, Pseudomonas aeruginosa ATCC 27853 and Cutibacterium (formerly Propionibacterium) acnes JCM 6425. All bacterial cultures were stored at −80 °C in nutrient broth supplemented with 15% glycerol.

In this method, agar plates were inoculated with 100 µL of test microorganism (approximately 10^10^ CFU/mL) and perforated using a borer (7 mm diameter), after which 50 µL of overnight cultures of isolates from skin and saliva samples were placed into each well. A control well containing sterile MRS broth was included and did not produce a zone of inhibition against any of the tested microorganisms. All test microorganisms were incubated aerobically at 37 °C for 24 h, except for C. acnes JCM 6425, which was incubated under anaerobic conditions. After incubation, the diameters of the inhibition zones were measured. All experiments were performed in three independent biological replicates (n = 3), and results are reported as mean ± SD.

2.5. Antioxidant Activity

The antioxidant activity of bacterial and yeast isolates from saliva and skin samples of participants was assessed using the method described by Cizeikiene and Jagelaviciute [18]. To prepare samples of intact cells, overnight cultures of the isolates were centrifuged for 5 min at 2200× g (Hermle, Gosheim, Germany). The cell pellets were washed and resuspended in PBS (pH 7.4). Intact cells were then mixed with 0.2 mM DPPH (1,1-diphenyl-2-picryl-hydrazyl; Merck, Darmstadt, Germany) solution, and prepared in ethanol, in a 1:1 ratio and incubated at room temperature in the dark for 30 min. After incubation, the samples were centrifuged for 5 min at 2200× g, and the absorbance of the supernatant was measured at 517 nm using a microplate reader (Tecan, Männedorf, Switzerland). As a control, DPPH in ethanol and PBS were used, while ethanol and the prepared cell suspension served as the blank. The DPPH radical scavenging activity (%) was calculated using the following equation:

$$[eqn]$$

where A_s_ represents sample and A_c_ control absorbance. All experiments were performed in three independent biological replicates (n = 3), and results are reported as mean ± SD.

2.6. Autoaggregation Capacity

The autoaggregation assay was carried out following the protocol described by Tuo et al. [19]. Bacterial and yeast isolates were harvested, washed twice with PBS (pH 7.4), and resuspended in PBS to achieve OD_620_ = 1. Absorbance at 620 nm was measured at 0 h and after 5 h using a microplate reader (PerkinElmer, Waltham, MA, USA). Autoaggregation (%) was calculated as follows:

$$[eqn]$$

where A_t_ represents absorbance after 5 h and A_0_ the absorbance at 0 h. All experiments were performed in three independent biological replicates (n = 3), and results are reported as mean ± SD.

2.7. Freeze-Drying Survival

Overnight cultures of bacterial and yeast isolates were centrifuged at 2200× g for 5 min. The cell pellets were washed twice and resuspended in 10% (w/v) skim milk (Biolife, Italy), which served as a cryoprotectant. Suspensions were first frozen at −80 °C (Eppendorf, Hamburg, Germany), and freeze-dried using a Christ Alpha 1-2 LD plus freeze-dryer (Martin Christ, Osterode am Harz, Germany) for 24 h at a temperature of −64 °C and a vacuum of 0.016 mbar. Cell viability was determined before and after freeze-drying by counting colony-forming units (CFU) on MRS or malt extract agar after incubation at 37 °C and 28 °C, respectively. The results are expressed as the change in log(CFU/mL) [(Δlog CFU/mL)] between pre- and post-freeze-drying. All experiments were performed in three independent biological replicates (n = 3) for each isolate.

2.8. Survival of Isolates in Simulated Saliva

Survival of isolates in simulated saliva was assessed following the procedure described by Kostelac et al. [20], with minor modifications. After overnight cultivation in MRS broth at 37 °C for bacteria and malt extract broth at 28 °C for yeast, the isolates were harvested by centrifugation at 2200× g for 5 min, washed twice with sterile saline, and resuspended in simulated saliva. Simulated saliva was prepared according to Marques et al. [21] and designated as simulated saliva 1 (SS1). It was composed of the following components (g/L): potassium chloride (0.72; Merck, Germany), calcium chloride dihydrate (0.22; Gram Mol, Zagreb, Croatia), sodium chloride (0.6; J.T. Baker, Wroclaw, Poland), potassium phosphate monobasic (0.68; Fisher Chemicals, Pittsburgh, PA, USA), sodium phosphate dibasic dodecahydrate (0.866; Sigma-Aldrich, St. Louis, USA), potassium bicarbonate (1.5; Carlo Erba Reagents, Milan, Italy), potassium thiocyanate (0.06; Carlo Erba Reagents, Milan, Italy), and citric acid (0.03; Lach-Ner, Neratovice, Czech Republic). The pH was adjusted to 6.5. After 1 h and 24 h of incubation in simulated saliva, and viable cell counts were determined by inoculating serial dilutions onto MRS and malt extract agar plates, respectively, followed by colony enumeration. The saliva survival assay was performed in three independent biological experiments (n = 3) for each isolate and each time point. Results are expressed as mean ± SD of log(CFU/mL).

2.9. Survival of Isolates Under Simulated GIT Conditions

To evaluate cell survival under simulated GIT conditions, simulated gastric and small intestine juices were prepared according to the procedure described by Čuljak et al. [22]. Overnight cultures of the isolates were harvested by centrifugation (2200× g for 5 min), washed twice in PBS, and resuspended in simulated gastric juice (pH 2), then incubated at 37 °C for 2 h. The samples were centrifuged again (2200× g for 5 min), and the cells were resuspended in small intestine juice (pH 8) and incubated for an additional 4 h at 37 °C. Cell viability was determined before and after simulated GIT conditions by CFU counting on MRS and malt extract agar plates, respectively. The GIT survival assay was conducted in three independent biological experiments (n = 3) for each isolate and each experimental stage. Results are expressed as mean ± SD of log(CFU/mL).

2.10. Biofilm Formation of Isolates

The biofilm-forming ability of isolates was assessed according to Kostelac et al. [20], with slight modifications. Briefly, 1 mL of MRS or malt extract broth was added to a 24-well polystyrene microtiter plate (Falcon, Eschau, Germany) and inoculated with 100 µL of bacterial or yeast isolates adjusted to an optical density (OD) of 0.5. Wells containing only MRS or malt extract broth without microbial inoculation were used as negative controls. After 48 h of incubation, the wells were washed with sterile deionized water to remove non-adherent cells, and adherent cells were fixed by adding methanol (Kemika, Zagreb, Croatia). The cells were then stained by incubating for 15 min with 0.1% crystal violet (Kemika, Croatia). To remove excess stain, the wells were washed with sterile deionized water, and the stain bound to adherent cells was solubilized using 33% citric acid (Kemika, Croatia). The OD was measured at 595 nm using a spectrophotometer (Tecan, Männedorf, Switzerland). OD values were compared with the mean OD value of the negative control (ODC) and classified according to Borges et al. [23] as non-biofilm producers (OD ≤ ODC), weak (ODC < OD ≤ 2 × ODC), moderate (2 × ODC < OD ≤ 4 × ODC), or strong producers (4 × ODC < OD). All experiments were performed in three independent biological replicates (n = 3), and results are reported as mean ± SD.

2.11. Adhesion of Isolates to Caco-2 and HaCaT Cell Lines

Colorectal adenocarcinoma Caco-2 (ATCC HTB-37) and human immortalized keratinocyte HaCaT cell lines were preserved in liquid nitrogen in DMEM (Dulbecco’s Modified Eagle Medium; Capricorn Scientific, Ebsdorfergrund, Germany) supplemented with 10% FBS (Capricorn Scientific, Germany) and 10% dimethyl sulfoxide (DMSO; Kemika, Croatia). For routine culture, cells were grown in 24-well plates until they reached 80–90% confluence. Prior to the adhesion experiment, monolayers were gently washed three times with PBS (pH 7.4).

The adhesion assay was performed as described in Čuljak et al. [22], with minor modifications. A total of 1 mL of the prepared bacterial or yeast suspension (OD adjusted to 1) was added to each well containing Caco-2 or HaCaT cells. After 1 h of incubation at 37 °C, monolayers were washed three times with PBS to remove non-adherent bacterial or yeast cells. Adherent cells were lysed with 0.05% Triton X-100 (AppliChem, Darmstadt, Germany) at 37 °C for 10 min, and the total number of adhered bacterial and yeast cells in each well was determined by inoculation onto appropriate agar plates followed by colony enumeration.

All adhesion experiments were performed in three independent biological replicates (n = 3) for each isolate and cell line. The results are visualized as log(CFU/mL) values before and after adhesion, and the corresponding percentage of adhesion is provided in Table S1.

2.12. Statistical Analysis

The results are presented as mean values from three independent experiments ± standard deviation (SD). Statistical significance was determined using one-way analysis of variance (ANOVA). Pairwise differences between group means were assessed using Tukey’s honestly significant difference test for pairwise comparisons after analysis of variance (https://www.statskingdom.com/index.html, accessed on 8 October 2025) [24]. Differences between samples were considered significant at p < 0.05.

3. Results

3.1. Isolation and Identification of Bacteria and Yeasts from Skin and Saliva Samples

After repeated subculturing on MRS and malt extract agar and broth to obtain pure cultures, a total of seven isolates were successfully recovered and purified from forearm skin swabs and saliva samples, while no isolates were obtained from oral-cavity or lower-back swabs. These included four bacterial isolates from saliva and three yeast isolates from forearm swabs. Gram staining revealed that four bacterial isolates had a Gram-positive cell wall structure and exhibited rod-shaped morphology. Consistent results were obtained with the KOH test, as adding the bacterial suspension to a drop of KOH did not result in the formation of a gel-like suspension, confirming that all isolates were Gram-positive. Furthermore, the catalase test indicated the absence of catalase activity, as no gas vesicle formation was observed following the addition of hydrogen peroxide. Microscopic analysis of native preparations of the remaining isolates (A1, A2, and A3) revealed morphology typical of yeasts, which was further confirmed by simple methylene blue staining.

After preliminary morphological and phenotypic tests, identification of the isolates was performed using API test systems. Based on the analysis of fermentation profiles obtained from the API tests, the bacterial and yeast isolates were identified to the species level. The final identifications, including isolate labels, species names, and reliability percentages, are presented in Table 1. Isolate S1 was identified as L. fermentum with a 99.7% identification percentage, whereas the other bacterial isolates S3, S5, and S6, were identified as L. rhamnosus with a high identification percentage (99.9%). All three yeast isolates A1, A2, and A3 were identified as Saccharomyces cerevisiae with an identification percentage of 86.8%.

The results obtained by the API assay were further confirmed using MALDI-TOF, which is a more reliable method of identification than the API (Table 2). All tested microorganisms were identified with high confidence, except for isolate S1.

3.2. Antibiotic Susceptibility Testing

The susceptibility of the isolates to nine antibiotics was assessed using the disk diffusion method, with results shown in Table 3. Among the bacterial isolates, S1, S3, S5, and S6 showed high sensitivity to ampicillin, clindamycin, chloramphenicol, erythromycin and tetracycline, but were resistant to vancomycin, kanamycin and gentamicin. Notably, isolates S1, S3, and S5 were also resistant to streptomycin, while isolate S6 was the only bacterial isolate susceptible to streptomycin.

3.3. Antimicrobial Activity of Isolates

The antimicrobial activity of the bacterial (S1, S3, S5, and S6) and yeast (A1, A2, and A3) isolates was evaluated against six test-microorganisms using the well-diffusion method. The results showed isolate-specific differences in inhibition zones (Table 4). Bacterial isolates S3, S5, and S6 exhibited the strongest antimicrobial activity, inhibiting the growth of all tested pathogens with inhibition zones larger than 13.5 mm. Isolates S3 and S6 were particularly notable, showing the highest inhibition zone diameters against four test-microorganisms—E. coli (19.75 ± 0.50 mm), S. aureus (17.00 ± 0.82 mm), S. typhimurium (17.00 ± 0.82 mm), and L. monocytogenes (14.75 ± 1.26 mm) for S3, and E. coli (19.25 ± 0.50 mm), S. typhimurium (16.50 ± 0.58 mm), P. aeruginosa (21.25 ± 0.96 mm), and L. monocytogenes (16.25 ± 0.50 mm) for S6. Isolate S1 displayed the weakest antimicrobial activity among the bacterial isolates, producing the smallest inhibition zones for nearly all test-microorganisms, except for C. acnes where the largest inhibition zone (19.22 ± 0.80 mm) was observed. In contrast, yeast isolates A1, A2, and A3 showed no antimicrobial activity against any of the selected test-microorganisms.

3.4. Antioxidant Activity

The antioxidant activity of the tested isolates ranged from 45 to 90% (Figure 1). Among the orally derived isolates, S6 showed the highest activity (70.21 ± 2.90%), followed by S1 (66.97 ± 2.83%), while S3 and S5 (58.22 ± 3.28% and 51.95 ± 6.27%, respectively) exhibited lower values. In comparison, the skin-derived isolates A1, A2, and A3 demonstrated statistically higher antioxidant activity—86.57 ± 1.05%, 82.01 ± 4.77% and 74.15 ± 2.07%, respectively.

3.5. Autoaggregation Capacity of Isolates

Autoaggregation capacity among the tested isolates varied considerably, ranging from below 20% to nearly 100% (Figure 2). Isolates S5 and S6 showed the lowest levels of autoaggregation (18.92 ± 4.76% and 23.95 ± 8.26%, respectively), while S1 exhibited moderate activity (42.17 ± 1.77%), and S3 showed higher values (79.33 ± 0.7%). The yeast isolates A1, A2, and A3 displayed the strongest autoaggregation ability, with values close to 100%.

3.6. Survival of Isolates After Freeze-Drying

Figure 3 shows the changes in the number of viable cells [log(CFU/mL)] for the isolates before and after freeze-drying. The lowest cell loss after freeze-drying, expressed as Δlog(CFU/mL), was recorded for isolates S1 and S3 (Δlog < 0.2), while S5 and S6 showed a moderate reduction in viable cell counts (Δlog ≈ 0.3–0.4). A significantly greater loss of viability was observed in yeast isolates A1, A2, and A3 (Δlog > 1), indicating higher sensitivity to the freeze-drying process. These results demonstrate that freeze-drying had a variable impact on the tested isolates, with bacterial isolates showing greater resistance compared to the yeasts, which exhibited a more pronounced decrease in viable cell counts.

3.7. In Vitro Survival of Isolates in Simulated Saliva and Gastrointestinal Tract (GIT) Conditions

The survival results of the isolates in simulated saliva are presented as log(CFU/mL) values (Figure 4). Comparing values before exposure and after 1 h of incubation in simulated saliva, no statistically significant loss of cell viability was observed in any isolate. In contrast, all isolates exhibited a significant decrease in viability after 24 h of exposure. The greatest decline was observed for bacterial isolate S1, with a reduction of 3.07 ± 0.14 Δlog(CFU/mL). The remaining isolates showed reductions in Δlog(CFU/mL) ranging from 0.40 to 0.98, with isolates S3 and S5 displaying the highest resistance to the simulated saliva conditions.

Figure 5 shows the survival of bacterial and yeast isolates after sequential exposure to simulated GIT conditions, including simulated gastric and small intestine juices. Initial counts for all bacterial isolates ranged from approximately 8 to 9.5 log(CFU/mL), and for yeast isolates from 7 to 8 log(CFU/mL). After exposure to simulated gastric conditions, isolates S3, S5, A1, and A3 did not exhibit a statistically significant reduction in log(CFU/mL). However, after exposure to simulated small intestine conditions, only isolate A1 remained the most resilient, again showing no significant reduction. The other isolates exhibited statistically significant decreases in log(CFU/mL), ranging from 0.27 to 1.57 Δlog(CFU/mL). Nevertheless, all isolates maintained viability above 7 log(CFU/mL), except for A3 (6.66 ± 0.10), indicating that they can withstand simulated passage through the GIT, although their survival varies depending on the isolate at each stage.

3.8. Biofilm Formation Ability of Isolates

The results of the biofilm formation ability of the isolates are presented in Table 5. All yeast isolates (A1, A2, and A3) and one bacterial isolate (S1) were classified as weak biofilm producers, bacterial isolate S5 as a moderate producer, and bacterial isolates S3 and S6 as strong biofilm producers, according to the classification by Borges et al. [23].

3.9. Adhesion of Isolates to Caco-2 and HaCaT Cell Lines In Vitro

The results of the adhesion ability of the isolates to Caco-2 (Figure 6a) and HaCaT (Figure 6b) cell lines are presented as log(CFU/mL) values before and after adhesion, where greater difference between pre- and post-adhesion values (Δlog) indicates lower adhesive capacity of the tested isolates. For adhesion to Caco-2 cells, isolates S6 and A1 exhibited the strongest adhesion, with Δlog(CFU/mL) values of 1.53 ± 0.07 and 1.67 ± 0.08, respectively, while isolate A3 showed the weakest adhesion (Δlog(CFU/mL) of 3.37 ± 0.15). For adhesion to HaCaT cells, isolate S1 demonstrated the strongest adhesion among all tested isolates, with a Δlog(CFU/mL) value of 1.87 ± 0.12. The corresponding percentage of adhesion relative to the initial inoculum is provided in Table S1.

4. Discussion

Studies on the isolation and characterization of new probiotic bacteria have primarily focused on LAB from the human GIT, breast milk, feces, or vaginal microbiota [25,26,27,28]. This paper provides insights into the isolation, identification and functional characterization of both bacterial and yeast isolates from human skin and salivary microbiota, aiming to explore and expand potential reservoirs of beneficial probiotics and identify new opportunities for their application. Novel potential probiotic strains were isolated, emphasizing the need to move beyond traditional intestinal sources and explore innovative approaches for identifying valuable new probiotic habitats. The antibiotic susceptibility, antioxidant activity, autoaggregation, and antimicrobial potential of the isolates may serve as a basis for developing diverse probiotic formulations, either oral or topical, enabling reintroduction to their site of origin. Additionally, the survival outcomes following freeze-drying and in vitro digestion were promising, further broadening the possibilities for practical application and formulation design.

After isolating and cultivating pure cultures, phenotypic and molecular identification were performed. Preliminary identification of the isolates was carried out using the API system (Table 1), while additional and more accurate identification was subsequently performed using MALDI-TOF (Table 2). Both methods confirmed the same species, with MALDI-TOF providing more reliable and up-to-date identification compared to API [29]. Although the API identification of bacterial isolates was generally correct, the nomenclature was not up-to-date as reclassification of genus Lactobacillus was conducted in 2020 [30]. Isolate S1 was identified as Limosilactobacillus sp., isolates S3, S5 and S6 as Lacticaseibacillus rhamnosus, and A1, A2 and A3 as Saccharomyces cerevisiae. These findings are consistent with previous reports highlighting the prevalence of LAB in the oral environment and the occasional occurrence of S. cerevisiae as a commensal or transient skin species [2,31,32]. The recovery of Limosilactobacillus sp. and Lcb. rhamnosus is particularly relevant, as both taxa are widely recognized for their probiotic properties, including antimicrobial metabolite production, immunomodulation, and resilience to environmental stress [5,6].

When selecting potential probiotic candidates for human use, it is essential to assess their susceptibility to antibiotics designated by the European Food Safety Authority to prevent the horizontal transfer of resistance genes to pathogenic microorganisms in the host microbiota. Therefore, antibiotic susceptibility testing of the bacterial isolates against nine antibiotics was performed as a preliminary qualitative screening using disc-diffusion method (Table 3). Limosilactobacillus sp. (S1) and Lcb. rhamnosus S3, S5, and S6 showed resistance to vancomycin, kanamycin, and gentamicin, while only S6 was susceptible to streptomycin. These results do not raise immediate safety concerns, as resistance to certain antibiotics is intrinsic to many Lactobacillus species [33,34,35]. Nevertheless, continuous monitoring of lactic acid bacteria for the emergence of acquired and potentially transferable resistance genes using molecular approaches, such as PCR-based assays and whole-genome sequencing (WGS), is essential prior to their application in probiotic formulations.

The antimicrobial or antagonistic activity of probiotics is an important characteristic, encompassing the production of antimicrobial compounds such as bacteriocins, enzymes, hydrogen peroxide and organic acids [36]. One of the main metabolites responsible for the antimicrobial activity of LAB is lactic acid, which lowers the pH and results in strong inhibitory activity against both Gram-positive and Gram-negative bacteria [37]. Using the agar well-diffusion method (Table 4), Lcb. rhamnosus S3, S5, and S6 exhibited antimicrobial effects against all tested pathogens, while Limosilactobacillus sp. (S1) inhibited E. coli, S. typhimurium, P. aeruginosa, and C. acnes. These results suggest the potential production of antimicrobial metabolites such as organic acids or bacteriocins, consistent with previous studies on LAB [38], indicating potential application either orally as a probiotic supplement or in the food industry. However, the exact mechanism of antimicrobial activity of the isolates should be further examined. Notably, Lcb. rhamnosus S3 had the strongest antimicrobial activity against S. aureus, a bacterium that commonly colonizes the skin of atopic dermatitis (AD) patients and contributes to the development and exacerbation of AD [39], making it the most promising candidate for topical formulation. However, application options are not limited to topical formulations. Vougiouklaki et al. [40] demonstrated that Lcb. rhamnosus has antimicrobial effects against food pathogens, suggesting its use in controlling the proliferation of undesirable microorganisms in various plant- or animal-derived food products. In conclusion, Limosilactobacillus sp. (S1) showed more promising antimicrobial activity against common food pathogens, while Lcb. rhamnosus S3, S5, and S6 exhibited broader antimicrobial activity and a wider range of potential applications. Notably, Limosilactobacillus sp. (S1) demonstrated the strongest inhibitory effect against C. acnes, a bacterium acting as an opportunistic pathogen in inflammatory skin disease acne vulgaris [41], suggesting its potential use in topical formulations for the management of acne. In contrast, yeast isolates did not display any antimicrobial activity against selected test-microorganisms, limiting their role in direct pathogen inhibition. Unlike many LAB, S. cerevisiae does not efficiently produce significant levels of acids or antimicrobial peptides under standard conditions, which are critical for broad-spectrum inhibition of pathogens [42,43]. Its ability to produce killer toxin proteins with antifungal effects is strain-dependent and not universal, meaning many wild-type or commercial strains lack potent killer activity [44].

Although S. cerevisiae did not demonstrate pathogen-inhibitory activity, it exhibited significant antioxidant effects (Figure 1), as reported in various studies [45,46,47]. Yeast isolates A1, A2, and A3 showed the highest antioxidant activity, while bacterial isolates displayed comparatively lower activity. As a eukaryotic microorganism, S. cerevisiae owes much of its antioxidant potential to the structure of its cell wall, primarily β-glucans and mannans, which possess strong radical scavenging properties and contribute to redox homeostasis [48,49,50,51]. Precisely through this mechanism—via its cell wall structure—S. cerevisiae exhibits probiotic and immunomodulatory potential by supporting gut barrier integrity and modulating immune responses. Therefore, its potential as a probiotic should not be considered negligible [52]. However, S. boulardii, a specific strain of S. cerevisiae, is better clinically validated for the prevention of antibiotic-associated and Clostridioides difficile-related diarrhea due to its anti-toxin activity and antibiotic resistance [53,54]. Further clinical studies are required to confirm the efficacy and safety of S. cerevisiae (A1, A2, and A3) in human applications. In contrast, LAB lack mannans, which may explain differences in their antioxidant profiles. Nevertheless, LAB exert antioxidant effects through multiple mechanisms, including metal–ion chelation, production of antioxidant enzymes, and metabolites such as glutathione and folate. They also modulate cellular signaling pathways like Nrf2 and NF-κB, down-regulate ROS-producing enzymes, and influence gut microbiota composition, thereby reducing oxidative stress [55]. Nevertheless, isolates of both yeasts and LAB have demonstrated significant antioxidant activity, highlighting their potential for ROS scavenging and their suitability for inclusion in next-generation probiotic formulations.

One essential characteristic to validate when developing a topical probiotic formulation is autoaggregation, a crucial feature that enhances the ability of microorganisms to adhere to skin surfaces, enabling effective colonization and competitive exclusion of pathogens [56]. Strains with strong autoaggregation capacity, such as certain Lactobacillus species, can form dense cellular clusters that reinforce both cell-to-cell and cell-to-surface interactions [57]. These aggregated microcolonies facilitate extended persistence on the epidermis and promote the expression or exposure of surface adhesins, which mediate strong binding to keratinocytes [58,59]. This dual mechanism, combining physical aggregation and molecular adhesion, underpins the probiotic’s ability to occupy skin niches, protect against environmental stress, and effectively compete with opportunistic pathogens like S. aureus, making strains with strong autoaggregation and adhesive properties promising candidates for topical applications aimed at maintaining or restoring skin microbial balance [60]. Autoaggregation assays revealed that Lcb. rhamnosus S3 exhibited the highest aggregation capacity among bacterial isolates, while yeast isolates A1, A2, and A3 displayed nearly complete aggregation (Figure 2), consistent with findings by Kil et al. [61], who reported pronounced autoaggregation ability in S. cerevisiae strains. Although isolates S3, S5, and S6 belong to the same bacterial species (Lcb. rhamnosus), their differing aggregation results may reflect strain-level variability, as phenotypic properties such as autoaggregation are known to vary substantially among strains of the same species. Tuo et al. [19] demonstrated variability among Lcb. rhamnosus strains, where strain 110 showed an autoaggregation value of 26.48 ± 17.46%, while the well-known probiotic strain Lcb. rhamnosus GG reached 41.39 ± 3.30%. The S3 isolate demonstrated substantially higher aggregation ability compared to both strains from Tuo et al. [19], whereas isolates S5 and S6 exhibited comparable or slightly lower values than Lcb. rhamnosus 110. These findings suggest that specific strains, such as Lcb. rhamnosus S3, may possess enhanced probiotic potential due to superior cell-to-cell adhesion capacity, an advantageous trait when selecting candidates for functional or topical probiotic formulations.

Freeze-drying is one of the most suitable methods for drying sensitive materials such as microorganisms. Despite its advantages, significant loss of probiotic viability may occur during the drying process and subsequent storage, due to alterations in membrane lipids and sensitive cellular proteins caused by low temperatures during the freezing phase, the formation of ice crystals, and the rate of their sublimation [62,63]. According to the criteria outlined by Frece [64], for a product to be classified and marketed as a probiotic, in addition to further necessary analysis, it must contain a minimum of 10^6^ CFU of viable probiotic microorganisms per gram or milliliter of product at the time of consumption. Therefore, the ability of isolates to survive freeze-drying was evaluated using skim milk as a lyoprotective agent, as it has been shown to stabilize the cell membrane and provide a protective coating around the cell [65]. The results showed that all isolates experienced some loss of viability after freeze-drying; however, Limosilactobacillus sp. (S1) and Lcb. rhamnosus S3 exhibited the highest survival rates, whereas yeast isolates generally showed lower survival ability (Figure 3). This difference could be related to the surface-to-volume ratio (S/V), which influences heat and water transfer during the freezing process. Bacterial cells are smaller than yeast cells and possess approximately five times higher S/V ratios. As a result, water and heat are released more rapidly from bacterial cells, preventing intracellular crystallization and leading to higher survival rates during freezing. In addition, differences in the composition of the cell wall and plasma membrane between bacteria and yeasts may also contribute to their varying resistance to stress induced by freeze-drying [66].

For probiotics intended for oral use, the ability to withstand the challenges of the local environment, including maintaining viability in the oral cavity, is of critical importance. In the mouth, bacteria must overcome various salivary defense mechanisms, including antimicrobial peptides, enzymes, fluctuating pH, and mechanical clearance by the continuous flow of saliva [67]. A 1 h exposure in simulated saliva was used to reflect typical oral residence times, while a 24 h exposure served as a test for assessing resilience in potential applications targeting oral pathogens or dental biofilms. In this study, isolates S3 and S5 showed the highest resistance to simulated saliva conditions (Figure 4), indicating their enhanced ability to survive the initial oral phase and potentially colonize the oral cavity. Similarly, Kostelac et al. [20] reported the ability of potential probiotics isolated from equid milk to withstand simulated saliva intended for oral protection.

One of the key characteristics of probiotic microorganisms is their ability to survive passage through the upper part of the GIT, as it is essential that probiotics reach their site of action in a viable state to exert their beneficial effects [68]. Therefore, the ability of isolates to survive simulated GIT conditions was tested, and the results showed satisfactory survival for all the isolates (Figure 5). After incubation under simulated small intestine conditions, the log(CFU/mL) values for all isolates were greater than 7 (except for A3), indicating that the number of viable cells after exposure to adverse conditions exceeded 10^7^. This meets one of the technological criteria for the selection of probiotic strains, suggesting their potential to remain viable and active upon oral administration. However, although the tested in vitro model can credibly represent and simulate conditions in the digestive system, future research may need to employ more advanced models that include microbial communities such as SHIME^®^ or other similar models.

In certain contexts, biofilm formation may represent an undesirable phenomenon [69]; however, many exogenously introduced microorganism that are not biofilm producers in a new environment are likely to be eliminated, even when administered in large quantities, which largely explains the low efficiency of many probiotic strains [70]. In contrast, biofilm-forming probiotics exhibit enhanced resistance to environmental conditions, offering a novel strategy to improve probiotic colonization and maintenance of their population, thereby strengthening their health-promoting effects [71]. Several probiotics with notable biofilm-forming abilities, such as Lpb. plantarum, Llb. fermentum, Lcb. rhamnosus, and S. cerevisiae, have been isolated [72,73]. The results of this study revealed Lcb. rhamnosus S3 and S6 to be strong biofilm producers, according to the classification by Borges et al. [23], while all yeast isolates (S. cerevisiae A1, A2, and A3) and the bacterial isolate Limosilactobacillus sp. (S1) were classified as weak biofilm producers (Table 5). Interestingly, both Lcb. rhamnosus S3 and S6 were identified as strong biofilm producers (Table 5) and also exhibited pronounced antimicrobial activity (Table 4). This concurrent ability to form biofilms and inhibit the growth of pathogenic microorganisms is particularly noteworthy. However, the specific mechanisms underlying the antibacterial activity of LAB biofilms remain insufficiently understood [74]. It is generally suggested that LAB biofilms may suppress competing microorganisms either through competitive exclusion or via the production and accumulation of antimicrobial compounds within the biofilm matrix, such as bacteriocins or bacteriocin-like substances.

Adhesion of potential probiotics to their target site is a crucial initial step for successful colonization, enabling them to exert beneficial effects such as stimulation of the immune system and protection against pathogens through several mechanisms, such as the production of antimicrobial compounds and competition for adhesion sites, thereby reducing pathogen colonization [22,75]. Colonic adenocarcinoma Caco-2 cells were used as an in vitro model of the intestinal epithelial barrier, while the immortalized human keratinocyte line HaCaT served as an in vitro model for studying skin (Figure 6). Although adhesion is often reported as the percentage of adhered cells relative to the initial inoculum, the absolute adhesion percentages observed in this study were relatively low. To enable clearer comparison of adhesion performance among isolates and cell lines, adhesion results were therefore primarily visualized as log(CFU/mL) values before and after adhesion. Isolates S6 and A1 demonstrated the strongest adhesion to Caco-2 cells, making them the most promising candidates for oral application. Numerous studies have reported the ability of probiotics to adhere to the Caco-2 cell line, with some also demonstrating their capacity for competitive exclusion of pathogens [76,77,78]. Regarding adhesion to HaCaT cells, isolate S1 exhibited the highest adhesive capacity, indicating its potential suitability for incorporation into topical formulations. Similarly, Lizardo et al. [79] demonstrated the pathogenic inhibition ability of Lcb. rhamnosus, showing that it can bind to HaCaT cells and inhibit the adhesion of S. aureus, one of the most relevant skin pathogens.

5. Conclusions

The human salivary and skin microbiota have proven to be valuable reservoirs for isolating new probiotic candidates, particularly Lcb. rhamnosus, Limosilactobacillus sp. and S. cerevisiae, thereby expanding research beyond traditional gastrointestinal sources. The combined findings of this study—including the survival rates after freeze-drying, tolerance to simulated gastrointestinal and salivary conditions, antimicrobial activity, antioxidant potential, autoaggregation, adhesion capacity and biofilm-forming ability—demonstrate that several of the isolates possess strong functional and technological properties suitable for both topical and oral probiotic applications. Among the bacterial isolates, Lcb. rhamnosus S3 and Limosilactobacillus sp. (S1) emerged as the most promising candidates. Lcb. rhamnosus S3 showed strong inhibition of S. aureus, high autoaggregation capacity, substantial survival following freeze-drying and good resilience under simulated saliva and GIT conditions, highlighting its potential for incorporation into formulations targeting skin disorders such as atopic dermatitis, as well as its applicability in oral or food-related probiotic products due to its ability to inhibit foodborne pathogens. Conversely, Limosilactobacillus sp. (S1) exhibited the strongest antimicrobial activity against C. acnes and demonstrated the highest adhesion to HaCaT cells among the tested strains; however, the observed adhesion was insufficient in absolute terms to support direct topical application, indicating that these findings are preliminary and that further optimization and validation are required before considering inclusion in topical formulations aimed at managing acne. Although S. cerevisiae isolates did not display antimicrobial activity, they exhibited marked antioxidant capacity attributable to the presence of β-glucans and mannans, indicating potential utility in formulations designed to protect against oxidative stress. However, due to their low survival after freeze-drying, yeast isolates were not considered suitable candidates for further development of probiotic formulations.

While these findings highlight the potential of skin and salivary isolates as multifunctional probiotics, it is important to note that the study involved a small cohort of healthy adult female volunteers with limited sampling sites. Additionally, species-level identification of isolate S1 relied on biochemical profiling and low-confidence MALDI-TOF MS results. Further studies with larger, more diverse cohorts and molecular validation of isolates are warranted to confirm and extend these exploratory findings.

Future studies should aim to elucidate the molecular mechanisms underlying these functional properties, evaluate host interactions in vitro and in vivo, and assess the feasibility of incorporating such strains into dermatological and oral care formulations. In summary, the findings indicate that the skin and salivary microbiota represent promising, underexplored reservoirs of multifunctional probiotics with potential for both topical and oral applications, justifying continued research integrating genomic safety validation and mechanistic characterization.