Bacteriophage Cocktails in the Post-COVID Rehabilitation

Increasing evidence suggests that gut dysbiosis is associated with coronavirus disease 2019 (COVID-19) infection and may persist long after disease resolution. The excessive use of antimicrobials in patients with COVID-19 can lead to additional destruction of the microbiota, as well as to the growth and spread of antimicrobial resistance. The problem of bacterial resistance to antibiotics encourages the search for alternative methods of limiting bacterial growth and restoring the normal balance of the microbiota in the human body. Bacteriophages are promising candidates as potential regulators of the microbiota. In the present study, two complex phage cocktails targeting multiple bacterial species were used in the rehabilitation of thirty patients after COVID-19, and the effectiveness of the bacteriophages against the clinical strain of Klebsiella pneumoniae was evaluated for the first time using real-time visualization on a 3D Cell Explorer microscope. Application of phage cocktails for two weeks showed safety and the absence of adverse effects. An almost threefold statistically significant decrease in the anaerobic imbalance ratio, together with an erythrocyte sedimentation rate (ESR), was detected. This work will serve as a starting point for a broader and more detailed study of the use of phages and their effects on the microbiome.


Introduction
Gut health and gut microbiota are increasingly associated with a variety of chronic diseases. Imbalances in the gut microbiota caused by poor diet, stress, antibiotic use, and other lifestyle and environmental factors are associated with the development of intestinal inflammation and gut disorders. Scientists are finding more and more links between gut microbiota and human health, from associations with autoimmune diseases and metabolic conditions to mental health and effects on brain function [1,2].
Increasing evidence suggests that gut dysbiosis is associated with coronavirus disease 2019 (COVID- 19) infection and may persist long after disease resolution. Patients with COVID-19 had significant alterations in fecal microbiomes compared with controls, characterized by the enrichment of opportunistic pathogens and depletion of beneficial commensals at the time of hospitalization and at all time points during hospitalization. Deficiency of symbionts and gut dysbiosis persisted even after the elimination of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the resolution of respiratory symptoms [3]. It has been shown that microbiota diversity does not recover even 6 months after convalescence [4]. Moreover, the use of antimicrobials in patients with COVID-19, often excessive, can lead to additional destruction of the microbiota, as well as to the growth and spread of antimicrobial resistance (AMR), especially in intensive care units [5]. The problem of bacterial resistance to antibiotics encourages the search for alternative methods

•
Post-COVID patients of moderate severity (incl. ICU stay), discharged from the hospital more than 3 months ago; • Presence of pulmonary foci of consolidation and fibrosis on chest CT (CT 1-3); • Rehabilitation routing score-less than 3 points; • Negative polymerase chain reaction (PCR) result for SARS-CoV-2. • Exclusion criteria: • Temperature above 38 • C; • Increased dyspnea (above 30/min); • Increase in systolic blood pressure above 180 mmHg or a decrease below 90 mmHg.
The bacteriophage cocktail for oral application was administered per os 2 times a day for 14 days, in the morning and in the evening. The amount of ingested phage suspension was 10 mL.
The bacteriophage cocktail for inhalations was administered by inhalation through an ultrasonic nebulizer 2 times a day for 14 days, the duration of each inhalation was 15-20 min, and the amount of inhaled phage suspension was 5 mL.

Sample Collection
One stool sample and one venous blood sample were collected from each subject on the day of admission (before bacteriophage therapy) and 14 days after the start of therapy. Stool samples were collected into a disposable sterile container. The containers were transported to the lab and prepared immediately. The time from sample collection to sample analysis did not exceed 12 h. DNA was extracted from the supernatant: 0.1 g of the stool sample was mixed with 800 µL of the isotonic solution and vortexed until homogeneity. The resulting mixture was centrifuged at 11,300× g for 30 s (MiniSpin, Eppendorf, Hamburg, Germany). Further analysis was carried out following the protocol provided by the assay manufacturer (Colonoflor-16 by AlphaLab, St. Petersburg, Russia). Blood was collected from a venous catheter into an anticoagulant-free test tube. Serum samples were obtained by centrifuging the blood at 1500× g for 10 min (ELMI CM-6M, Riga Latvia). Serum aliquots (500 µL) were poured into disposable Eppendorf tubes, frozen, and stored at −20 • C until further use.

Analysis of Gut Microbiota Taxonomic Abundance
The composition of the gut microbiota was analyzed using Colonoflor-16 kits (Al-phaLab; Russia), which include reagents for DNA extraction, PCR primers specific for all bacterial DNA (total bacterial mass), and species-specific primers for 16 microbial species. Measurements were performed using a CFX 96 Real-Time PCR Detection System (BioRad, Hercules, CA, USA).

CT Scan Analysis
To analyze the results of lung CT scans, the method of automatic calculation of the volume of damaged lung tissue according to the type of ground glass using the software "Ground glass" (InfoRad 3.0 DICOM Vievew, Moscow, Russia) was used. Segmentation of the right and left lung and trachea was performed with a threshold of −250 HU, and areas of damage within the lung with density in the custom range (from −785 HU to 150 HU) were identified.

Experimental Part In Vitro Real-Time Phage Lysis Assay
In the growth phase of the culture, before adding the bacteriophage complex, an increase in the number of bacteria in the microscope field of view was observed ( Figure 1). The first stages of biofilm formation were observed: adhesion on glass and the beginning of microcolony formation. A full video of the growth phase is provided in Supplementary Video S1.

CT Scan Analysis
To analyze the results of lung CT scans, the method of automatic calculation of the volume of damaged lung tissue according to the type of ground glass using the software "Ground glass" (InfoRad 3.0 DICOM Vievew, Moscow, Russia) was used. Segmentation of the right and left lung and trachea was performed with a threshold of −250 HU, and areas of damage within the lung with density in the custom range (from −785 HU to 150 HU) were identified.

Experimental Part
In Vitro Real-Time Phage Lysis Assay In the growth phase of the culture, before adding the bacteriophage complex, an increase in the number of bacteria in the microscope field of view was observed ( Figure  1). The first stages of biofilm formation were observed: adhesion on glass and the beginning of microcolony formation. A full video of the growth phase is provided in Supplementary Video S1. The bacteriophage cocktail was added 1 h 40 min after the start of the incubation and filming. Within 2 h 30 min after the addition of phages, there was a slight decrease in the number of bacteria in the field of view but formed bacterial microcolonies were still visible. Active reduction in the number of bacteria in the field of view started 2 h 30 min The bacteriophage cocktail was added 1 h 40 min after the start of the incubation and filming. Within 2 h 30 min after the addition of phages, there was a slight decrease in the number of bacteria in the field of view but formed bacterial microcolonies were still visible. Active reduction in the number of bacteria in the field of view started 2 h 30 min after the addition of the bacteriophage complex, and within 30 min, all microcolonies were disrupted ( Figure 2). A full video of the bacterial lysis phase is provided in Supplementary Video S2. Additionally, a lysis of the bacterium adhered to the glass in the center of the microscope field of view was captured (Supplementary Video S3).
after the addition of the bacteriophage complex, and within 30 min, all microcolonies were disrupted ( Figure 2). A full video of the bacterial lysis phase is provided in Supplementary Video S2. Additionally, a lysis of the bacterium adhered to the glass in the center of the microscope field of view was captured (Supplementary Video S3). There was a decrease in the area covered by bacterial cells from 27.9% to 10.5% ( Figure 3).  There was a decrease in the area covered by bacterial cells from 27.9% to 10.5% ( Figure 3). after the addition of the bacteriophage complex, and within 30 min, all microcolonies were disrupted ( Figure 2). A full video of the bacterial lysis phase is provided in Supplementary Video S2. Additionally, a lysis of the bacterium adhered to the glass in the center of the microscope field of view was captured (Supplementary Video S3). There was a decrease in the area covered by bacterial cells from 27.9% to 10.5% ( Figure 3). Further observation showed no continuation of lysis or elimination of the bacterial culture. Further observation showed no continuation of lysis or elimination of the bacterial culture.  37). All subjects had evidence of lung involvement on CT scans ranging from 10% to 75% (Stage I-50%, II-31%, III-12, IV-7%) during acute COVID-19. Antibacterial therapy was administered to more than 25% of subjects during the underlying illness (thirdgeneration cephalosporins, amoxicillin in combination with beta-lactamase inhibitors). The comparison group consisted of subjects (n = 8) also who had undergone COVID-19 in the past and received the complex program of rehabilitation in August 2022 but did not receive bacteriophage cocktails. The age range was from 21 to 77 years (median: 53 (44; 62) years); 7 study participants were female (87%), and 1 was male (13%).

Clinical
As a result of the rehabilitation program involving phage therapy, the patients' saturation increased, respiratory rate normalized, and heart rate decreased within the reference values (Table 1). A statistically significant increase in hemoglobin level and decrease in erythrocyte sedimentation rate (ESR) ( Table 1) were registered in comparison with the control group (Table S1). A trend toward normalization of the leukocyte level was observed. X-ray imaging assessment comparing the volume of lung tissue lesion "retrospectively/now" revealed the following results: residual and limited pneumofibrosis in 55% of patients; resolution of pneumonia from 5% to 20% (previously 60%) in 30% of cases, and resolution of pneumonia up to 40% (previously 80%) in 15% (Table S2). An example of CT scan analysis using the "Ground Glass" software is presented in Figure S1.  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (∆ 0.12 after versus ∆ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant. A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).
A trend towards a decrease in the total bacterial mass was revealed after the bacteriophage therapy, as well as a decrease in the concentrations of proinflammatory microorganisms (lg CFU/g): S. aureus, Enterococcus spp., Clostridium spp., Proteus spp., Citrobacter spp.
After 14 days of application of the bacteriophage cocktail, a statistically significant decrease in the inflammatory ratio was revealed ( Figure 4).

0.951
Bifidobacterium spp. 10 9 -10 10 3 × 10 10 (4 × 10 9 ; 7 × 10 10 ) 2 × 10 10 (3 × 10 9 ; 6 × 10 10 ) A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant. A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant. Bacteroides spp. 10 9 -10 12 4 × 10 13 (9 × 10 12 ; 9 × 10 13 ) 2 × 10 13 (10 13 ; 6 × 10 13 ) A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). The composition of the gut microbiota included Clostridium difficile and Clostridium perfringens, Fusobacterium nucleatum, and Parvimonas micra, which are not found in significant concentration in a healthy gut. It also included a high ratio of Bacteroides fragilis group/Faecalibacterium prausnitzii, a sign of anaerobic imbalance, a condition characteristic of inflammatory bowel diseases such as ulcerative colitis and Crohn's disease, and autoimmune pathology as an indicator of a disturbed state of local intestinal immunity.
A trend towards a decrease in the total bacterial mass was revealed after the bacteriophage therapy, as well as a decrease in the concentrations of proinflammatory microorganisms (lg CFU/g): S. aureus, Enterococcus spp., Clostridium spp., Proteus spp., Citrobacter spp.
After 14 days of application of the bacteriophage cocktail, a statistically significant decrease in the inflammatory ratio was revealed ( Figure 4).

0.122
Bacteroides thetaiotaomicron 10 9 -10 12 5 × 10 9 (4 × 10 8 ; 2 × 10 10 ) 2 × 10 9 (7 × 10 8 ; 10 10 ) A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3).  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant.  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant.  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant.  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant.  A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant. The composition of the gut microbiota included Clostridium difficile and Clostridium perfringens, Fusobacterium nucleatum, and Parvimonas micra, which are not found in significant concentration in a healthy gut. It also included a high ratio of Bacteroides fragilis group/Faecalibacterium prausnitzii, a sign of anaerobic imbalance, a condition characteristic of inflammatory bowel diseases such as ulcerative colitis and Crohn's disease, and autoimmune pathology as an indicator of a disturbed state of local intestinal immunity.

Parameters
A trend towards a decrease in the total bacterial mass was revealed after the bacteriophage therapy, as well as a decrease in the concentrations of proinflammatory    A statistically significant increase in the level of total protein was also revealed, which may indirectly indicate an improvement in absorption processes in the intestine against the background of the normalization of the state of the gut biocenosis. A decrease in the level of C-reactive protein (CRP) as a marker of inflammation (Δ 0.12 after versus Δ 0.6 before) should be noted. C-reactive protein levels were elevated in 4 of 24 (17%) patients at admission and in 2 (8%) at discharge (p = 0.894).

Gut Dysbiosis
The microbiota was characterized by various markers of dysbiosis upon admission to rehabilitation in 100 % of the subjects: excessive bacterial growth (92%), increased concentration (lg CFU/g) of proinflammatory types of microorganisms (Bacteroides fragilis group, Candida spp., S. aureus, Proteus spp., Enterococcus spp., Enterobacter spp., Citrobacter spp.) exceeding the reference values by 1.5-2 times, and low levels of Bacteroides thetaiotimicron, Akkermansia muciniphila compared with the reference values (Table 3). Table 3. Taxonomic abundance of the gut microbiota in subjects before and after the bacteriophage therapy (BPh). * Correlation is statistically significant. The composition of the gut microbiota included Clostridium difficile and Clostridium perfringens, Fusobacterium nucleatum, and Parvimonas micra, which are not found in significant concentration in a healthy gut. It also included a high ratio of Bacteroides fragilis group/Faecalibacterium prausnitzii, a sign of anaerobic imbalance, a condition characteris-Viruses 2022, 14, 2614 9 of 15 tic of inflammatory bowel diseases such as ulcerative colitis and Crohn's disease, and autoimmune pathology as an indicator of a disturbed state of local intestinal immunity.

Parameters
A trend towards a decrease in the total bacterial mass was revealed after the bacteriophage therapy, as well as a decrease in the concentrations of proinflammatory microorganisms (lg CFU/g): S. aureus, Enterococcus spp., Clostridium spp., Proteus spp., Citrobacter spp.
After 14 days of application of the bacteriophage cocktail, a statistically significant decrease in the inflammatory ratio was revealed (Figure 4).

Clinical Part
Our results demonstrate the potential of the bacteriophage cocktail to modulate the gut microbiome.
Previously, phages were thought to cause minimal changes in the composition and diversity of the gut microbiota in experimental studies compared with antibiotics [27]. Other studies using gnotobiotic mice have shown that exposure to phages leads to compositional changes in the murine gut microbiota [28,29]. The conducted study of the application of the bacteriophage cocktail targeting E. coli showed a decrease in the number of fecal E. coli loads. No significant changes in alpha and beta diversity parameters were observed, suggesting that the phages consumed did not disrupt the microbiota. However, an increase in the number of representatives of the butyrate-producing genus Eubacterium and a decrease in the proportion of taxa most closely related to Clostridium perfringens was observed. Short-chain fatty acid production, inflammatory markers, and lipid metabolism were virtually unchanged, but there was a small but significant decrease in circulating interleukin-4 (Il-4). Taken together, these data demonstrated the potential of bacteriophages to selectively reduce the number of target organisms without disrupting the gut community [23].
However, a clinical study demonstrated a virtually unchanged microbiota profile over 4 weeks of phage therapy when added to systemic antibiotics in one patient with S. aureus infection [30]. There are several publications describing the use of inhaled [31] or

Clinical Part
Our results demonstrate the potential of the bacteriophage cocktail to modulate the gut microbiome.
Previously, phages were thought to cause minimal changes in the composition and diversity of the gut microbiota in experimental studies compared with antibiotics [27]. Other studies using gnotobiotic mice have shown that exposure to phages leads to compositional changes in the murine gut microbiota [28,29]. The conducted study of the application of the bacteriophage cocktail targeting E. coli showed a decrease in the number of fecal E. coli loads. No significant changes in alpha and beta diversity parameters were observed, suggesting that the phages consumed did not disrupt the microbiota. However, an increase in the number of representatives of the butyrate-producing genus Eubacterium and a decrease in the proportion of taxa most closely related to Clostridium perfringens was observed. Short-chain fatty acid production, inflammatory markers, and lipid metabolism were virtually unchanged, but there was a small but significant decrease in circulating interleukin-4 (Il-4). Taken together, these data demonstrated the potential of bacteriophages to selectively reduce the number of target organisms without disrupting the gut community [23].
However, a clinical study demonstrated a virtually unchanged microbiota profile over 4 weeks of phage therapy when added to systemic antibiotics in one patient with S. aureus infection [30]. There are several publications describing the use of inhaled [31] or intravenous [32] bacteriophages in patients with pneumonia, bronchitis, and infectious endocarditis. All studies confirm the safety and declare the absence of adverse reactions when using bacteriophages for therapeutic purposes. Previously, in our pilot clinical trial, the safety of bacteriophages was evaluated for the first time in a group of patients with a chronic critical illness. In addition to the absence of local and general adverse events, an important aspect to note is that the effectiveness of the technique was confirmed by the treatment outcomes seen in the phage therapy group, which were not inferior to those in the group receiving conventional antibiotic therapy [26].
In the present study, a cocktail of bacteriophages for ingestion was used. Almost all subjects had levels of total bacterial mass above the reference values on admission (92%, 22/24) that decreased slightly by the time of discharge. The microbiota of the subjects on admission was characterized by the presence of proinflammatory types of microorganisms and low levels of anti-inflammatory microorganisms, as in patients with the post-COVID syndrome. These findings are consistent with previously identified microbiota disturbances after COVID-19 [33][34][35]. Dysbiosis was also confirmed by an increased ratio of the Bacteroides fragilis/Faecalibacterium prausnitzii group in 67 % of the subjects, which is a sign of anaerobic imbalance. Previously, similar changes in this ratio were observed in chronic critical illnesses, reaching maximum values during massive antibiotic therapy [36]. Application of the phage cocktails for two weeks did not cause adverse effects among the subjects, with normalization of stools and a reduction in the severity of gastrointestinal symptoms.
After phage therapy, a statistically significant threefold decrease in the anaerobic imbalance ratio was detected. Although no correlation between this coefficient and Creactive protein was found, there was a normalization of this biomarker and a statistically significant decrease in the ESR, a parameter that may also indicate inflammation. Reducing the intensity of the inflammatory response associated with bacterial infection, in particular, reducing the level of CRP, may be one of the effects of the bacteriophage application [37].
Bacteriophages in nature play an important role as regulators of bacterial populations, as well as in regulating the number of bacteria in the human gut microbiome. Understanding this role of bacteriophages became the basis for the development of the concept of immunity based on the adhesion of phages to mucus (bacteriophage adherence to mucus immunity) [38]. Bacteriophages on mucous membranes are considered the first line of counteraction to the development of dysbiosis; they respond to the bacterial proliferation process even before the reaction of cellular and humoral mechanisms of the immune system and contribute to faster achievement of a new point of balance by the microbiome in case of dysbiosis development [39].
We consider a significant clinical effect of bacteriophage inhalation to be a statistically significant increase in saturation and normalization of respiratory rate. The effectiveness of the inhaled form was confirmed by the results in patients with nosocomial pneumonia, where the treatment results were comparable in the conventional antibiotic therapy group [26].
For the in vitro study, we chose K. pneumoniae, a microorganism that frequently causes nosocomial pneumonia; moreover, patients often become infected with an intestinal colonization strain [40]. Because of the short incubation time, it was not possible to observe biofilm formation under the microscope. However, the first stages of biofilm formation-adhesion on glass and formation of microcolonies-were recorded. The characteristics of the bacteriophages studied in vitro have been previously examined: they encode polysaccharide-depolymerases, and the latent period is 30-35 min for all bacteriophages included in the cocktail. The delay in lysis on the video was due to the fact that the phage cocktail was added to the Petri dish laterally through a micro syringe in order to avoid significant mixing of the bacterial culture, and the filming took place in the central part of the Petri dish. Thus, the lysis of the bacteria began unevenly in the studied volume and reached the microscope field of view 2 h and 30 min after adding the bacteriophage complex.
Active lysis lasted for 30 min, and no decrease in the number of bacteria in the field of view was observed during further incubation. This may be due to a complex population relationship between bacteriophages and bacteria. It is known that bacteriophages reproduce most actively in exponentially growing bacterial populations, while in a poor culture, lysis may stop, and bacteriophages may go into the so-called hibernation phase [41]. Additionally, bacteriophages usually actively lyse populations with high density. Thus, bacteriophages act as a regulator of bacterial populations in nature. This was observed in the present experiment when active lysis stopped after a significant reduction in the number of bacteria. The method allows us to visualize the growth of the bacterial population and the lysis of bacteria after the addition of bacteriophages. However, the closed type of the system does not allow sampling from it, so it is not possible to estimate lytic efficacy by cultural methods correctly. We propose to consider the analysis using the 3D Cell Explorer microscope as a method to visualize lysis in real time, which is rarely performed with phages because the more common method of imaging is electron microscopy, which involves fixation of the sample and staining.
This study had certain limitations. One of the limitations of this work is the use of real-time PCR to characterize the intestinal microbiota instead of 16s rRNA sequencing, which limits the taxa that can be detected and does not allow the evaluation of minor bacterial species. At the same time, this method provides the ability to track changes and identify major nosocomial pathogens virtually at the patient's bedside in routine clinical practice. Another limitation was the subjects' insufficient completion of questionnaires such as the Gastrointestinal Symptom Rating Scale (GSRS) [42] and the Bristol Stool Scale. Questionnaires were received from only four people, which does not allow us to discuss the significance of the obtained results. However, the four questionnaires received indicated a decrease in gastrointestinal symptoms according to GSRS and normalization of stool according to the Bristol Stool Scale. In the in vitro study, the main limitation was the choice of a clinical strain of K. pneumoniae for visualization. We chose K. pneumoniae because higher titers of this bacterium in the gut increase the risk of bacterial infections, including bacterial pneumonia. However, we did not find Klebsiella spp. exceeding the reference values in the intestines of patients, as was assumed when planning this work. Nevertheless, the objective of the in vitro experiment was not to show the efficacy of phages on a specific bacterium but to test a new method for visualizing the action of bacteriophages. Usually, methods used to visualize the action of bacteriophages do not allow for recording lysis in real time. In this study, we first imaged bacteriophage lysis in real time using a 3D Cell Explorer microscope.

Conclusions
In the present study, complex phage cocktails targeting several bacterial species were used for the first time. The results showed the safety of this application and the absence of side effects. The phage therapy group showed statistically significant improvements in saturation and respiratory rate as well as a decrease in inflammatory markers, such as ESR and Bacteroides fragilis group/Faecalibacterium prausnitzii ratio, according to PCR analysis of gut microbiota. The results demonstrate the relevance of bacteriophages in the rehabilitation of the microbiota of patients who have undergone COVID-19 and received antibiotic treatment. This work will serve as a starting point for a broader and more detailed study of the effects of phages on the human microbiome and the whole organism.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/v14122614/s1, Video S1: 3D Cell Explorer visualization of K. pneumoniae (Kl 315) culture growth during 1 h 40 min of incubation. The field of view size is 100 × 100 µm; Video S2: 3D Cell Explorer visualization of K. pneumoniae (Kl 315) culture lysis by bacteriophage cocktail. Magnification ×58, scale bar 20 µm; Video S3: 3D Cell Explorer lysis visualization of the bacterium adhered to the glass in the center of the microscope field of view; Table S1: clinical and laboratory parameters in the control group. Data are presented as the median and interquartile range (IQR); Table S2: CT lung volume assessment in subjects on bacteriophage therapy (BPh); Figure S1: lung CT data visualization using the "Ground glass" software (patient ID 20). Lung capacity 3220 cm 3 , lung damage 800 cm 3 (26%), left lung damage 436 cm 3 (22%), right lung damage 364 cm 3 (31%).