HK40104269A - Use of microbiome for assessment and treatment of obesity and type 2 diabetes - Google Patents

Description

Uses of the microbiome in evaluating and treating obesity and type 2 diabetes

Related applications

This application claims priority to U.S. Provisional Patent Application No. 63/169,481, filed April 1, 2021, the contents of which are incorporated herein by reference in their entirety for all purposes.

Background of the Invention

As global living standards continue to improve, the number of overweight or even obese individuals is rapidly increasing. Because excessive weight is directly linked to serious health risks, this rising proportion of overweight individuals in the general population leads to a significant increase in the incidence of many diseases, including diabetes, heart disease, hypertension, and stroke. For example, the World Health Organization (WHO) estimates that by 2030, the number of people with diabetes worldwide will exceed 350 million. Due to the rising incidence of obesity-related diseases, their severe health impacts, and their profound economic consequences, there is an urgent need for new and effective methods to determine an individual's risk of developing obesity and type 2 diabetes (T2D), enabling individuals deemed at increased risk for obesity and T2D to receive preventative and early treatment, ultimately reducing or eliminating their risk of later developing serious conditions related to diabetes, hypertension, cardiovascular disease, etc. This invention addresses this and other related needs by providing novel methods and compositions that can effectively assess a patient's risk of obesity or T2D.

Invention Overview

This invention relates to novel methods and compositions for assessing the risk of obesity and type 2 diabetes (T2D) in subjects and for assessing the nature of obesity and T2D in humans—whether the disease state is gut microbiome-dependent. In particular, the inventors of this application have discovered that certain microbial species, especially certain bacteria, exist in the gastrointestinal tract (GI) of an individual at significantly different levels depending on whether the individual is at an increased risk of developing obesity and T2D. Therefore, in a first aspect, the invention provides methods for reducing the risk of obesity and type 2 diabetes (T2D) in subjects or for treating obesity and T2D in subjects. The method includes the step of introducing an effective amount of one or more bacterial species into the gastrointestinal tract of the subject, said bacterial species being selected from *Faecalibacterium prausnitzi*, *Bifidobacterium longum*, *Eubacterium halli*, *Bifidobacterium bifidum*, *Roseburia intestinalis*, *Eubacterium eligens*, *Lachnospiraceae bacterium 5163FAA*, *Eubacterium ventriosum*, and *Roseburia hominis*. In some embodiments, the bacterial species do not include any of the bifidobacterial species. In some embodiments, the bacterial species include no more than one bifidobacterial species. In some embodiments, the introduction step includes oral administration to the subject of a composition comprising an effective amount of said one or more bacterial species. In some embodiments, the introduction step includes delivering a composition containing an effective amount of the one or more bacterial species into the small intestine, ileum, or large intestine of the subject. In some embodiments, the introduction step includes fecal microbiota transplantation (FMT), for example, by administering a composition containing processed donor fecal material to the subject. In some embodiments, the processed donor fecal material is a mixture containing fecal material taken from at least two, possibly more, lean donors, such as those with a BMI < 23 kg/ ^m² . In some embodiments, the composition used in the method does not contain any species shown in Tables 2 or 4 that are detectable, for example, specific bacterial species that are undetectable by conventional detection methods such as nucleic acid hybridization or polymerase chain reaction (PCR). In some embodiments, the composition is administered orally. In some embodiments, the composition is deposited directly into the gastrointestinal tract of the subject. In some embodiments, the level or relative abundance of one or more bacterial species is determined in a first fecal sample obtained from the subject before the introduction step and a second fecal sample obtained from the subject after the introduction step, for example by polymerase chain reaction (PCR), preferably quantitative polymerase chain reaction (qPCR).

In a second aspect, the present invention provides a novel method for assessing the risk of obesity and type 2 diabetes (T2D) in an individual by analyzing the distribution of certain gut bacterial species. The method comprises the following steps: (1) determining the level or relative abundance of one or more bacterial species shown in Tables 1-5 in fecal samples from the subject; (2) determining the level or relative abundance of the same bacterial species in fecal samples from a reference group, the reference group including subjects with obesity and T2D and subjects without obesity and T2D; (3) generating a decision tree using the data obtained from step (2) via a random forest model, and running the level or relative abundance of one or more bacterial species from step (1) along the decision tree to generate a score; and (4) identifying subjects with a score greater than 0.5 as having an increased risk of obesity and T2D, and identifying subjects with a score less than 0.5 as not having an increased risk of obesity and T2D.

In some embodiments, the one or more bacterial species include any one, any two, or any three bacterial species shown in Tables 1-5. For example, bacterial species include (i) *Clostridium bartlettii*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA, *E. truncatella* or (ii) *Clostridium bartlettii* or (iii) *Haemophilus parainfluenzae* or (iv) *E. coli* or (v) *Trichophyceae* 5_1_63FAA or (vi) *E. truncatella* or (vii) *Clostridium bartlettii*, *Haemophilus parainfluenzae*, *E. coli*, *Trichophyceae* 5_1_63FAA or (viii) *Trichophyceae*, *Haemophilus parainfluenzae*, *E. coli*, *Trichophyceae* or (i) x) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. convex* or (x) *Clostridium pallidum*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. convex* or (xi) *Haemophilus parainfluenzae*, *E. coli*, *Trichophyton spp.* 5_1_63FAA, *E. convex* or (xii) *Clostridium pallidum*, *Trichophyton spp.* 5_1_63FAA, *E. convex* or (xiii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA or (xiv) *Haemophilus parainfluenzae*, *E. coli*, *Trichophyton spp.* 5_1_63FAA or (xv) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *E. coli* or (xvi) *Haemophilus parainfluenzae*, *E. coli*. In some embodiments, the subject was not diagnosed with obesity. In some embodiments, the subject was not diagnosed with type 2 diabetes (T2D). In some implementations, each of steps (1) and (2) includes metagenomic sequencing or polymerase chain reaction (PCR), such as quantitative PCR.

In a third aspect, the present invention provides a method for assessing whether a subject suffers from microbiome-dependent obesity and type 2 diabetes (T2D). The method comprises the following steps: (1) determining the level or relative abundance of one or more bacterial species shown in Tables 1-5 in a fecal sample from the subject; (2) determining the level or relative abundance of the same bacterial species in fecal samples from a reference group, the reference group including subjects suffering from obesity and T2D and subjects not suffering from obesity and T2D; (3) generating a decision tree using the data obtained from step (2) via a random forest model, and running the level or relative abundance of one or more bacterial species from step (1) along the decision tree to generate a score; and (4) identifying subjects with a score greater than 0.5 as suffering from microbiome-dependent obesity and T2D, and identifying subjects with a score not greater than 0.5 as suffering from microbiome-independent obesity and T2D.

In some embodiments, the one or more bacterial species include any one, any two, or any three bacterial species shown in Table 5. For example, bacterial species include (i) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* or (ii) *Clostridium pallidum* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyton spp.* 5_1_63FAA or (vi) *E. truncatella* or (vii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA or (viii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *E. truncatella* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* Bacillus or (x) Clostridium pallericum, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli or (xi) Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli or (xii) Clostridium pallericum, Trichophyton 5_1_63FAA, Escherichia coli or (xiii) Clostridium pallericum, Haemophilus parainfluenzae, Trichophyton 5_1_63FAA or (xiv) Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA or (xv) Clostridium pallericum, Haemophilus parainfluenzae, Escherichia coli or (xvi) Haemophilus parainfluenzae, Escherichia coli. In some embodiments, the subject has been diagnosed with obesity. In some embodiments, the subject has been diagnosed with T2D. In some embodiments, each of steps (1) and (2) includes metagenomic sequencing or polymerase chain reaction (PCR), such as quantitative PCR.

In a fourth aspect, the present invention provides kits for assessing the risk of obesity and type 2 diabetes (T2D) in subjects or for assessing whether subjects have microbiome-dependent obesity and T2D. The kits contain reagents for detecting one or more bacterial species shown in Tables 1-5. For example, the bacterial species include (i) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* or (ii) *Clostridium pallidum* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyton spp.* 5_1_63FAA or (vi) *E. truncatella* or (vii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA or (viii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *E. truncatella* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* Eubacterium or (x) Clostridium barlett, Escherichia coli, Trichophyceae bacteria 5_1_63FAA, Eubacterium convexum or (xi) Haemophilus parainfluenzae, Escherichia coli, Trichophyceae bacteria 5_1_63FAA, Eubacterium convexum or (xii) Clostridium barlett, Trichophyceae bacteria 5_1_63FAA, Eubacterium convexum or (xiii) Clostridium barlett, Haemophilus parainfluenzae, Trichophyceae bacteria 5_1_63FAA or (xiv) Haemophilus parainfluenzae, Escherichia coli, Trichophyceae bacteria 5_1_63FAA or (xv) Clostridium barlett, Haemophilus parainfluenzae, Escherichia coli or (xvi) Haemophilus parainfluenzae, Escherichia coli. In some embodiments, the kit comprises two or more containers, each containing a composition comprising reagents for polymerase chain reaction (PCR), such as quantitative PCR, for detecting bacterial species, such as primers and/or probes, which typically contain nucleotide sequences homologous to or complementary to polynucleotide sequences from bacterial species.

Brief description of the attached diagram

Figure 1: Differential bacterial species between obese and T2D (ObT2) subjects and lean controls. Green bars represent species that are abundant in lean controls, while red bars represent species that are abundant in ObT2.

Figure 2(A): Receiver operating characteristic (ROC) curves and area under the curve (AUC) of machine learning models. AUCs using the following random forest models: all 5 markers (red) - Clostridium barlettii, Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA, and Eubacterium truncatulae.

Figure 2(B): Receiver operating characteristic (ROC) curves and area under the curve (AUC) of machine learning models. AUC of random forest models using the following individual markers: Clostridium barbarum (5-red), Haemophilus parainfluenzae (4-light blue), Escherichia coli (3-green), Trichophyton 5_1_63FAA (2-dark blue), and Eubacterium truncatum (1-orange).

Figure 3: Box plot depicting the relative abundance of biomarkers for machine learning models in ObT2 and lean controls.

Figure 4(A): Risk score of the new ObT2 subject (new subject) compared with the ObT2 and lean control. Figure 4(B): Risk score of the new lean subject (new subject) compared with the ObT2 and lean control.

Figure 5: Effects of different FMT regimens on weight loss in obese subjects. Figure 5(A) Schematic diagram of the studies. In the nFMT study, recipients received four monthly mixed-donor FMTs. Stool samples were collected from recipients at baseline, one month after the first FMT infusion, one month after the last FMT, and two to three months after the last FMT. In the iFMT study, subjects received a 3-day course of antibiotics, followed by five consecutive weeks of single-donor FMT (with two-day intervals) for four weeks. In both studies, patients' clinical parameters were followed up until week 52. Figure 5(B): Weight change after FMT. Significance between studies was calculated using repeated measures ANOVA.

Figure 6: Intensive FMT leads to an increase in the number of species derived from lean donors and resembles the donor microbiome distribution. Figure 6(A): Proportion of donor-derived species in obese subjects. Figure 6(B): Abundance of donor-derived species in obese subjects. Figure 6(C): Bray Curtis distance between the microbiota in the FMT-treated sample and the corresponding baseline sample. Figure 6(D): Bray Curtis distance between the microbiota in the recipient sample and the corresponding donor sample.

Figure 7: Mixed-donor FMT was more effective in inducing an increase in butyrate-producing bacteria compared to single-donor intensive FMT. Figure 7(A): Heatmap depicting the abundance of butyrate-producing bacteria. Figure 7(B): Depicting the Chao1 abundance and Shannon diversity index of butyrate-producing bacteria in the two studies. Figure 7(C): Depicting the aggregation abundance of butyrate-producing bacteria in the two studies. Figure 7(D): Depicting the network diagram of associations of butyrate-producing bacteria after nFMT. Figure 7(E): Depicting the Chao1 abundance and Shannon diversity index in the two FMT schemes. Significance between studies was calculated using the Wilcoxon rank-sum test. Significance within the same study was calculated using the Wilcoxon signed-rank test.

Figure 8: Implantation or replacement of butyrate-producing bacterial strains in FMT recipients. Figure 8(A): Different strain clusters based on SNP haplotype distribution. Figure 8(B): Strain replacements in FMT recipients at each time point. Strain clusters were defined at a tree height of 0.8 (80% dissimilarity).

Figure 9: The size of the LDA effect of bacterial species that changed significantly after nFMT (LDA>2, p<0.05).

Figure 10: Line plot illustrating the relative abundance of butyrate-producing bacteria in the donor and recipient after FMT and iFMT. Abundance is shown as relative abundance (%) after logarithmic transformation.

Figure 11: Correlation of abundance changes of butyrate-producing bacteria 1 month after the last FMT infusion.

Figure 12: Species present 2–3 months after baseline and the last FMT infusion.

definition

As used herein, the term "microbiome-dependent" describes the correlation between the presence and/or state of a physiological state (e.g., a person's weight) or medical condition (e.g., obesity or type 2 diabetes) and the distribution of microorganisms found in a predetermined environment (e.g., the human gastrointestinal tract) (in terms of both presence and absolute or relative quantity). Similarly, the term "bacteriome-dependent" describes the correlation between a person's physiological/pathological condition and the distribution of bacterial species present in a person (e.g., the human gastrointestinal tract).

"Relative abundance percentage," when used in the context of describing the presence of a specific bacterial species (e.g., any of those shown in any of Tables 1-5) in relation to all bacterial species present in the same environment, refers to the relative amount of that bacterial species out of the total amount of all bacterial species, expressed as a percentage. For example, the relative abundance percentage of a particular bacterial species can be determined by comparing the amount of species-specific DNA in a given sample (e.g., determined by quantitative polymerase chain reaction) with the amount of bacterial DNA in all samples of the same sample (e.g., determined by quantitative polymerase chain reaction PCR and sequencing based on the 16S rRNA sequence).

"Absolute abundance," when used in the context of describing the presence of a specific bacterial species in feces (e.g., any of those shown in Tables 1-5), refers to the amount of DNA from the bacterial species out of the total amount of DNA in a fecal sample. For example, the absolute abundance of a bacterium can be determined by comparing the amount of species-specific DNA of that bacterium in a given sample (e.g., determined by quantitative PCR) to the total amount of fecal DNA in the same sample.

As used herein, the “total bacterial load” of a fecal sample refers to the amount of all bacterial DNA in the total amount of DNA in the fecal sample. For example, the absolute abundance of bacteria can be determined by comparing the amount of bacterial-specific DNA (e.g., 16S rRNA determined by quantitative PCR) in a given sample with the total amount of all fecal DNA in the same sample.

The term “overweight” is used to describe individuals who are overweight and have a body mass index (BMI) greater than 25 (or between 23 and 24.9 in Asian populations). This term encompasses “obesity” or “obesity disorder,” which describes a condition in which a patient has a BMI greater than 30 (or greater than 25 in Asian populations).

As used in this application, the terms "treat" or "treating" describe actions that result in the elimination, reduction, alleviation, reversal, prevention, and/or delay of the onset or recurrence of any symptoms of a predetermined medical condition. In other words, "treating" a condition encompasses therapeutic and preventative interventions for that condition, including those that promote patient recovery from the condition.

The term "fecal microbiota transplantation (FMT)" or "fecal transplantation" refers to a medical procedure in which fecal material containing live fecal microorganisms (bacteria, fungi, viruses, etc.) obtained from a healthy individual is transferred to the recipient's gastrointestinal tract to restore a healthy gut microbiota that has been disrupted or destroyed by any of a variety of medical conditions, such as overweight or obesity and related conditions. Typically, fecal material from a healthy donor is first processed into a form suitable for transplantation, which can be achieved by direct delivery to the lower gastrointestinal tract, such as via colonoscopy, or via nasal intubation, or by oral ingestion of encapsulated material containing processed (e.g., dried and frozen or lyophilized) fecal material.

As used herein, the term "effective amount" refers to the amount of a substance (e.g., an antimicrobial agent) used or applied to produce a desired effect (e.g., inhibition or suppression of the growth or proliferation of one or more undesirable bacterial species). Effects include prevention, inhibition, or delay of any associated biological processes during bacterial proliferation to any detectable extent. The exact amount will depend on the nature of the substance (active agent), the manner of use/application, and the purpose of the application, and will be determined by a person skilled in the art using known techniques and those described herein. In another context, when an "effective amount" of one or more beneficial or desired bacterial species is artificially introduced into a patient's gastrointestinal tract, such as a composition intended for use in FMT, this means that the amount of the relevant bacteria introduced is sufficient to confer health benefits to the recipient, such as reduced recovery time or a reduced need for therapeutic interventions for the associated condition (e.g., overweight or obesity), including but not limited to medications (e.g., appetite suppressants) and any of a variety of treatments such as behavioral and communication therapy, educational therapy, family therapy, speech or physical therapy, etc.

As used herein, the term "inhibiting" or "inhibition" refers to any detectable negative effect on a target biological process, such as RNA/protein expression of a target gene, biological activity of a target protein, cell signal transduction, cell proliferation, etc. Typically, inhibition reflects a reduction of at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or more in the target process (e.g., the growth or proliferation of certain types of microorganisms, such as one or more bacteria shown in Table 1) or any of the downstream parameters mentioned above, compared to a control. "Inhibition" also includes 100% reduction, i.e., the complete elimination, prevention, or abolition of the target biological process or signal. Other related terms, such as “suppressing,” “reducing,” “decrease,” “decreasing,” “lower,” and “less,” are used in this disclosure in a similar manner to refer to different levels of reduction (e.g., at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or more, compared to a control level (i.e., the level before inhibition),) up to the complete elimination of the target biological process or signal. On the other hand, terms such as “activate,” “activating,” “activation,” “increase,” “increasing,” “promote,” “promoting,” “enhance,” “enhancing,” “enhancement,” “higher,” and “more” are used in this disclosure to cover positive changes at different levels of the target process or signal (e.g., at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%, 200%, or greater, such as 3-fold, 5-fold, 8-fold, 10-fold, or 20-fold increases compared to control levels (before activation), such as control levels of one or more bacterial species shown in Table 1). In contrast, the terms "substantially the same" or "substantially unchanged" mean that the amount from the comparison base (such as a standard control) has hardly changed, typically within ±10% of the comparison base, or within ±5%, 4%, 3%, 2%, 1%, or even less of the comparison base.

The term "antimicrobial agent" refers to any substance capable of inhibiting, suppressing, or preventing the growth or proliferation of bacterial species, such as any of those shown in Tables 2, 4, and 5. Known agents with antimicrobial activity include various antibiotics that generally inhibit the proliferation of a broad spectrum of bacterial species, as well as agents capable of inhibiting the proliferation of specific bacterial species, such as antisense oligonucleotides, small repressive RNAs, etc. The term "antimicrobial agent" is similarly defined to encompass agents with broad-spectrum activity that kill almost all bacterial species, as well as agents that specifically inhibit the proliferation of a target bacterial species. Such specific antimicrobial agents can be natural short polynucleotides (e.g., small repressive RNAs, microRNAs, miniRNAs, lncRNAs, or antisense oligonucleotides) that disrupt the expression of key genes in the life cycle of the target bacterial species, thus specifically inhibiting or eliminating only that species without significantly affecting other closely related bacterial species.

As used herein, the term “about” indicates a range of values that are +/- 10% of a specified value. For example, “about 10” means a range of values from 9 to 11 (10 +/- 1).

Invention Details

I. Introduction

This invention provides novel methods and compositions for assessing the risk of obesity and T2D in individuals, particularly those not diagnosed with obesity or type 2 diabetes (T2D), and for assessing whether the condition of an individual with obesity and T2D is related to and potentially caused by a certain distribution of the bacterial community or related bacterial species found in their gastrointestinal tract. At the end of such assessment, individuals deemed to have an increased risk of obesity or T2D may receive treatment to preventively reduce or eliminate such risk and prevent or delay the onset of the condition. Similarly, individuals already suffering from obesity and/or T2D, upon being identified as having one or more conditions of a microbiome-dependent nature, may receive appropriate treatment to alleviate their symptoms in terms of severity, extent, and/or duration. For example, preventative or therapeutic treatments may involve artificially altering the levels of relevant bacterial species in the human gastrointestinal tract, such as increasing the amount or level of “beneficial” bacteria or suppressing the amount or level of “harmful” bacteria through fecal microbiota transplantation (FMT) to provide health benefits to the individuals being tested and treated.

II. Treatment methods that regulate bacterial levels

The inventors' findings in this application reveal a direct correlation between medical conditions such as obesity and type 2 diabetes (T2D) and the distribution of certain bacterial species (e.g., those shown in Tables 1-5) in the patient's gut. This disclosure enables various methods for the prevention and treatment of obesity and T2D, and related symptoms, particularly for adjusting or modulating the levels of these bacterial species in the patient's gut by delivering effective amounts of one or more "beneficial" or desired bacterial species to the patient's gastrointestinal tract via, for example, an FMT procedure, or by delivering antimicrobial agents to inhibit target bacterial species to reduce the levels of one or more "harmful" or undesirable bacterial species. This helps individuals at increased risk of obesity and T2D, or obese/T2D patients, benefit from different treatment regimens, such as medications and/or various therapies. In some cases, the composition used for FMT infusion is derived from at least two donors (e.g., from two lean donors) with a desired gastrointestinal distribution of beneficial bacterial species, rather than from a mixture of fecal material from a single donor.

For example, one or more desired bacterial species, such as some shown in Tables 1 or 3, can be introduced from an exogenous source into materials prepared for FMT, such that the level of bacterial species in the transfer material reaches a desired level (e.g., at least about 0.01%, 0.02%, 0.05%, 0.10%, 0.20%, 0.40%, 0.50%, 0.60%, 0.80%, 1.0%, 2.0%, 3.0%, 4.0%, 5.0%, 6.0%, 7.0%, 8.0%, 8.5%, 9.0%, or 10% of the total bacteria in the material), and then processed for FMT to prevent or treat obesity and T2D, to reduce the risk of obesity/T2D in an individual, or to alleviate the symptoms of obesity/T2D in an individual. In some cases, sufficient quantities of beneficial bacterial species can be obtained from bacterial cultures and then formulated into suitable compositions for delivery to the recipient's gut. Similar to FMT, this composition can be administered to patients orally, nasally, or rectally.

On the other hand, the relative abundance of certain bacterial species (e.g., some shown in Tables 2, 4, and 5) has been found to increase due to the presence of obesity/T2D or an elevated risk of obesity/T2D. Therefore, patients with obesity/T2D, or those at an elevated risk of obesity/T2D, are treated to reduce the levels of these bacterial species in order to improve the patient's condition-related symptoms or the likelihood of preventing/delaying/reducing the onset of the condition. Several approaches exist to reduce the levels of these bacterial species: First, the patient can be given a specific antimicrobial agent to specifically kill or inhibit the target bacterial species, thereby reducing the levels of these bacteria. Second, the patient can first be given an antimicrobial agent, such as a broad-spectrum antibiotic to kill or inhibit all bacterial species or a specific antimicrobial agent to specifically kill or inhibit the target bacterial species; then a combination can be administered to the patient (e.g., via FMT) to introduce a well-balanced mixed bacterial culture into the patient's gastrointestinal tract.

Using a single composition containing related bacterial species in appropriate proportions (such as processed fecal material from an FMT donor), each of these protocols can be performed in a single combined step to achieve the first and second therapeutic objectives: increasing the levels of certain bacterial species and decreasing the levels of certain other bacterial species.

After completing the steps of introducing an effective amount of the desired bacterial species into the patient's gastrointestinal tract (e.g., via the FMT procedure) and/or suppressing unwanted bacterial levels, the recipient can be further monitored immediately by continuously testing the level or relative abundance of bacterial species in stool samples on a daily, weekly, or monthly basis until 6 months after the procedure. Clinical symptoms of the treated obesity/T2D and the patient's overall health status should also be monitored to assess treatment outcomes and the corresponding levels of relevant bacteria in the recipient's gastrointestinal tract. The levels of bacterial species (one or more of those shown in Tables 1-5) can be monitored in conjunction with observations of the obtained health benefits (such as improvements in weight, blood pressure, blood glucose, lipid, and cholesterol levels).

III. Assessing obesity/T2D risk and microbiome dependence

The inventors of this application have discovered that altered distribution of certain bacterial species in the human gastrointestinal tract can indicate the presence or risk of obesity/T2D, even if the person may not be diagnosed with obesity or T2D: when properly calculated using, for example, certain specific mathematical tools as described herein, the levels or relative abundance of certain bacterial species (such as one or more species shown in Table 1) have been revealed to indicate an increased risk of developing obesity/T2D later in life or that the subject’s obesity/T2D is related to the gut distribution of bacterial species (i.e., “microbiome dependence”).

Once an obesity/T2D risk assessment has been conducted, for example, if an individual is deemed to have microbiome-dependent obesity/T2D or is at increased risk of developing obesity/T2D later, appropriate treatment steps can be taken as measures to address the individual's disease or increased risk. For example, the individual may be given medications such as hypoglycemic agents, insulin sensitizers, and/or appetite suppressants, or a composition containing an effective amount of (1) one or more beneficial bacterial species or (2) an antimicrobial species that inhibits harmful bacterial species may be given via FMT or by alternative administration methods, such that the bacterial distribution in the patient's gastrointestinal tract is altered to a distribution that is conducive to weight loss and prevention of T2D or relief of T2D symptoms.

IV. Kits and Compositions

This invention provides kits and compositions for reducing the risk of obesity and type 2 diabetes (T2D) in subjects or for treating obesity and T2D in subjects. The kits comprise two or more containers, each containing a different composition comprising an effective amount of different bacterial species or combinations of bacterial species selected from *Bacillus predniseri*, *Bifidobacterium longum*, *Eubacterium hominis*, *Bifidobacterium bifidum*, *Rosbairiella enterica*, *Eubacterium pickyii*, *Trichophyton spp.* 5_1_63FAA*, *Eubacterium convexum*, and *Rosbairiella hominis*. The compositions are formulated and introduced into the recipient's gastrointestinal tract, for example, by oral administration or by direct delivery via suppositories. In addition to the bacterial species specified above, the compositions may also contain one or more therapeutic agents that effectively lower blood glucose, sensitize insulin response, and suppress appetite to further facilitate the management of the risk of T2D and obesity.

The present invention also provides novel kits and compositions that can be used to assess the likelihood of a patient developing obesity and type 2 diabetes later in life, or to assess whether a patient's obesity/type 2 diabetes is microbiome-dependent. Typically, the kit contains reagents for detecting one or more bacterial species shown in Tables 1-5. For example, a kit is provided comprising (1) a first container containing a first composition containing a first reagent for detecting one of the bacterial species shown in Tables 1-5, and (2) a second container containing a second and different reagent for detecting one of the bacterial species shown in Tables 1-5. Optionally, a third reagent for detecting the bacterial species in Tables 1-5 may be included in the kit. When the kit is intended for detecting two or more bacterial species in Tables 1-5, additional compositions containing additional reagents may be included in the kit to allow the user to detect and measure the presence and levels of multiple bacterial species. In some variations, the first and second (and optionally more) reagents may be included in a single composition.

In some cases, the reagent contains a set of oligonucleotide primers for amplifying polynucleotide sequences from any of the bacterial species shown in Tables 1-5. For example, the reagent may be primers and/or probes for polymerase chain reaction (PCR), such as quantitative PCR, as an amplification reaction. Typically, such reagents may contain a set of oligonucleotide primers for PCR targeting polynucleotide sequences specific to each of the relevant bacterial species (such as any one or more bacterial species selected from Tables 1-5), and preferably each of the relevant bacterial species.

As an alternative, a means of detecting one or more bacterial species shown in Tables 1-5 is metagenomic sequencing, and the kit includes a composition containing one or more reagents suitable for metagenomic sequencing of preselected bacterial species (one or more listed in Tables 1-5). For example, the kit may contain detection reagents for analyzing bacterial species including: (i) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. convexis* or (ii) *Clostridium pallidum* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyton spp.* 5_1_63FAA or (vi) *E. convexis* or (vii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA or (viii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. convexis ...ii) *Clostridium pallidum* or (iii _1_63FAA, *E. convexis* or (x) *Clostridium pallida*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xi) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xii) *Clostridium pallida*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xiii) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Trichophyceae* bacteria 5_1_63FAA or (xiv) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA or (xv) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Escherichia coli* or (xvi) *Haemophilus parainfluenzae*, *Escherichia coli*.

Example

The following embodiments are provided by way of illustration only, and not by way of limitation. Those skilled in the art will readily recognize that many non-critical parameters can be changed or modified to produce substantially the same or similar results.

background

The aim of this study was to determine how the human gut microbiome is associated with obesity and type 2 diabetes (T2D). Practical applications of this invention include assessing the risk of obesity and T2D-related diseases based on the presence and abundance of certain bacterial species in the gastrointestinal tract of test subjects, and evaluating whether obesity and T2D in test subjects are associated with the gut microbiome, particularly the bacterialbiome.

Example 1: Machine learning model for predicting the risk of obesity and type 2 diabetes

method

Group description and research subjects

The study recruited 123 Chinese adults, including 68 participants with both obesity and type 2 diabetes (ObT2) (BMI > 28 kg/ ^m² ) and 55 healthy lean participants (lean controls, BMI < 23 kg/ ^m² ). This study was approved by the Joint CUHK-NTEC CREC (CREC Ref. No: 2016.607). All participants consented to donate stool samples and completed a written informed consent questionnaire. Stool samples from the participants were stored at -80°C for downstream microbiome analysis.

Fecal DNA extraction and DNA sequencing

Fecal bacterial DNA was extracted using the RSC PureFood GMO and Authentication Kit (Promega), modified to enhance DNA yield. Approximately 100 mg of each fecal sample was pretreated: the sample was suspended in 1 ml _ddH₂O and precipitated by centrifugation at 13,000 × g for 1 min. 800 μl of TE buffer (pH 7.5), 16 μl of β-mercaptoethanol, and 250 U of lysin were added to the washed sample, mixed thoroughly, and digested at 37 °C for 90 min. The precipitate was then collected by centrifugation at 13,000 × g for 3 min.

After pretreatment, the precipitate was resuspended in 800 μl of CTAB buffer (RSC PureFoodGMO and Authentication Kit, according to the manufacturer's instructions) and thoroughly mixed. After heating the sample at 95 °C for 5 min and then cooling, nucleic acids were released from the sample by vortexing at 2850 rpm with 0.5 mm and 0.1 mm beads for 15 min. Then, 40 μl of proteinase K and 20 μl of RNase A were added, and the nucleic acids were digested at 70 °C for 10 min. Finally, the supernatant was obtained after centrifugation at 13,000 × g for 5 min and placed in an RSC instrument for DNA extraction. The extracted fecal DNA was used for ultra-deep metagenomic sequencing via an Ilumina Novaseq 6000 (Novogene, Beijing, China).

Quality control of raw sequences

First, the original sequence reads were trimmed using Trimmomatic ¹ (v0.38), and then non-human reads were separated from contaminated host reads. Several steps were taken to obtain clean reads: 1) aptamers were removed; 2) the reads were scanned with a 4-base-wide sliding window, and reads were removed when the average quality per base dropped below 20; 3) the reads were reduced to a length of less than 50 bases. The trimmed sequence reads were mapped to the human genome (reference database: GRCh38p12) using KneadData (v0.7.2) to remove host-derived reads. The two reads at paired ends were then concatenated.

Analysis of bacterial microbiome

Bacterial community composition was analyzed on metagenomically pruned reads via MetaPhlAn2 (v2.7.5) ^2. Reads were mapped to clade-specific marker genes and species pangenome annotations via Bowtie2 (v2.3.4.3) ^3. The output table contains bacterial species and their relative abundance at different levels from kingdom to species. The data were analyzed in R v3.6.1 using tidyverse (v1.2.1) ⁴ , ggpubr (v0.2, github.com/kassambara/ggpubr) and phyloseq (v1.24.2) ^5. Differential bacterial species were compared between ObT2 subjects and lean controls via linear discriminant analysis effect size (LEfSe) ^6. Another method of bacterial taxonomic annotation was used as an alternative analysis of the bacterial microbiome. In this method, species-level community composition was generated using Kraken2 (v2.0.8-beta). Reference bacterial genomes were downloaded from NCBI RefSeq on November 5, 2019, and the database was built using default parameters. Subsequently, each query was classified into taxa with the highest total hits of k-mers that matched the general taxonomic tree associated with the mapped genome. A multivariate association linear model (MaAsLin2) was used to identify the association between clinical metadata and microbial abundance, while controlling for confounding factors.

Machine learning models

Using fecal microorganisms (due to their superior performance in classifying using binary features), Random Forest (RF) was chosen to build the evaluation model. ^Random Forest is one of the most popular methods in metagenomic data analysis for identifying discriminative features and building predictive models. As a widely used ensemble learning algorithm, Random Forest consists of a series of Classification and Regression Trees (CART) to form a strong classifier. A subset of data randomly sampled from the original dataset with replacements is called bootstrap sampling and is used to build the tree. When the training dataset for the current tree is drawn using bootstrap, observations from the overall dataset are omitted. In the case of infinite N, 36.8% of the data does not appear in the training samples called out-of-bag (OOB) observations, which will not be used to construct the tree. Additionally, additional randomness is introduced into Random Forest when each decision tree splits nodes based on a random subset of features selected from the overall features. Features with the minimum Gini (Gini is used to evaluate the purity of a node) are used to split nodes in each iteration to generate the tree. The algorithm is able to train different trees for different data and feature subsets and obtain the final classification by averaging the results from the tree model. In addition to predictive models, random forests also have the ability to assess variable importance.8 Out-of-Body (OOB ⁾ observations are used to estimate the classification error for each tree in the forest. To measure the importance of a given variable, the values of the variable in the OOB data are randomly changed, and then new predictions are generated using the changed OOB data. The difference between the error rates of the changed and original OOB observations, divided by the standard error, is used to calculate the variable's importance. To classify a new sample, the sample's features are passed down to each tree to estimate the probability of classification. Random forests use the average probability across all trees to determine the final classification result.

The importance of each species to the classification model was evaluated using recursive feature elimination. If a species' Pearson correlation with any existing probe in the model was <0.7, the selected species were added to the random forest model one by one, based on decreasing importance. The model's performance was re-evaluated using 10-fold cross-validation each time a new feature was added. These models were compared based on the area under the curve (AUC) of the binary classifier versus the receiver operating characteristic (ROC) curve. The final model was selected when the best accuracy and kappa were achieved. These analyses were performed using the R packages randomForest ^v4.6-14.7 and pROC ^v1.15.3.9 .

result

The gut bacteria distribution differs between lean subjects and ObT2 subjects.

Using MetaPhlAn2 and LEfSe analyses, it was found that, compared with the ObT2 subjects, the lean controls showed higher relative abundance of bacterial species including *Faecalibacterium prausnitzii*, *Bifidobacterium longum*, *Eubacterium hallii*, *Bifidobacterium bifidum*, *Rosbairiella intestinalis*, *Eubacterium tumefaciens*, *Trichophyton spp.* 5_1_63FAA, *Trichophyton spp.*, *Rosbairiella hominis*, *Clostridium barata*, *Anaerostipes hadrus*, *Gordonibacter pameaeae*, *Veillonella parvula*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 8_1_57FAA, *Streptococcus sanguinis*, *Streptococcus australis*, and *Streptococcus infantis* (Figure 1, Table 1). In contrast, compared with the lean control, the ObT2 subjects were enriched with species of Escherichia coli, Parabacteroides distasonis, Bacteroides stercoris, Clostridiales bacterium 1_4_56FAA, Clostridium bacterium 1_7_47FAA, Weissella confusa, and Actinomyces graevenitzii (Figure 1, Table 2).

Table 1: Bacterial species enriched in lean controls compared to ObT2 subjects (via MetaPhlAn2 method).

Arranged according to average relative abundance in healthy lean subjects

Table 2: Bacterial species enriched in ObT2 subjects compared to lean controls (via MetaPhlAn2 method)

Arranged according to average relative abundance in healthy lean subjects

Using Kraken2 as an alternative approach to annotating bacteriological taxonomy, a range of species showed higher relative abundance in lean controls compared to ObT2 objects (Table 3), while some species showed higher relative abundance in ObT2 objects compared to lean controls (Table 4).

Table 3: Bacterial species enriched in lean controls compared to ObT2 subjects (via Kraken2 method)

Arranged according to average relative abundance in healthy lean subjects

Table 4: Bacterial species enriched in ObT2 subjects compared to lean controls (via Kraken2 method)

Arranged according to average relative abundance in healthy lean subjects

The bacteria listed in Tables 1, 2, 3, and 4 can be used in different combinations to determine the risk of obesity and type 2 diabetes. For example, relative abundance can be determined using qPCR primer sets or metagenomic sequencing to calculate risk.

In addition, the bacteria listed in Tables 1 and 3 can be applied to subjects with obesity and type 2 diabetes (T2D) or at risk of developing T2D to improve symptoms of obesity and T2D or reduce the risk of developing T2D later. Conversely, the bacteria listed in Tables 2 and 4 can inhibit subjects with obesity and T2D or at risk of developing T2D to improve symptoms of obesity and T2D or reduce the risk of developing T2D later.

Machine learning model for predicting ObT2

Five bacterial biomarkers were used in the machine learning model, including *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton* bacteria _5_1_63FAA, and *Eubacterium buergerianum* (Table 5). The final model had a 90.3% area under the curve (AUC) in the receiver operating characteristic (ROC) curve (Figure 2A). The relative abundance of these bacteria in the ObT2 and lean controls is shown in Figure 3.

Table 5: Bacterial species included in the machine learning model used to evaluate ObT2

Bacterial species NCBI:txid Clostridium barlett 261299 Haemophilus parainfluenzae 729 E. coli 562 Bacteria of the family Trichophyceae_5_1_63FAA 658089 Protruding belly eubacterium 39496

Table 6: Relative abundance of species listed in Table 5 in ObT2 subjects (n=68) and healthy controls (n=55)

Table 7: Relative abundance of species listed in Table 2 in the new subjects

The machine learning model can be used to (1) predict the risk of ObT2 in subjects who are not obese or do not have type 2 diabetes (T2DM) at the time of testing, and (2) assess whether the subject’s ObT2 is microbiome-dependent in subjects who are already obese or have T2DM.

The following steps will be performed:

1. A set of training data was obtained by determining the relative abundance of species selected from Table 4 in obese (ObT2) subjects with type 2 diabetes and lean controls. Species selected from Table 5 should include: (ii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *Eubacterium truncatum* (all 5 species; AUC: 90.3%; Fig. 2A) or (ii) *Clostridium pallidum* (AUC: 71.2%; Fig. 2B (red)) or (iii) *Haemophilus parainfluenzae* (AUC: 73.3%; Fig. 2B (light blue)) or (iv) *Escherichia coli* (AUC: 74.4%; Fig. 2B (green)) or (v) (vi) *Clostridium pachyphyllum* 5_1_63FAA (AUC: 41.9%; Fig. 2B (dark blue)) or (vi) *Clostridium pachyphyllum* (AUC: 66.5.2%; Fig. 2B (orange)) or (vii) *Clostridium pachyphyllum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Clostridium pachyphyllum* 5_1_63FAA (AUC: 87.7%) or (viii) *Clostridium pachyphyllum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Clostridium pachyphyllum* (AUC: 86.9%) or (ix) *Clostridium pachyphyllum* Haemophilus parainfluenzae, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 86.2%) or (x) Clostridium pallidum, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 88.8%) or (xi) Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 85.0%) or (xii) Clostridium pallidum ...6.2%) or (x) Clostridium pallidum, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 88.8%) or (xi) Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 85.0%) or (xii) Clostridium pallidum, Trichophyton 5_1_63FAA, Escherichia coli (AUC: 85.0%) Bacillus (AUC: 86.7%) or (xiii) Clostridium pallidum, Haemophilus parainfluenzae, Trichophyton 5_1_63FAA (AUC: 85.0%) or (xiv) Haemophilus parainfluenzae, Escherichia coli, Trichophyton 5_1_63FAA (AUC: 84.1%) or (xv) Clostridium pallidum, Haemophilus parainfluenzae, Escherichia coli (AUC: 86.4%) or (xvi) Haemophilus parainfluenzae, Escherichia coli (AUC: 85.6%).

2. To determine the risk of ObT2 in non-obese or non-T2DM subjects, or to determine whether the pre-existing ObT2 in subjects is microbiome-related, the relative abundance of these species is determined.

3. Use a random forest model to compare the relative abundance of these species in the object with the training data.

4. Decision trees will be generated from the training data via a random forest. Relative abundance will be run along the decision trees to generate risk scores. If more than 50% of the trees in the model consider an object to be similar to the ObT2 group, the result will be "The object is considered to have an increased risk of ObT2" or "The object's existing ObT2 is microbiome-dependent." If less than 50% of the trees in the model consider an object to be similar to the ObT2 group, the result will be "The object is considered to have a low risk of ObT2" or "The object's existing ObT2 is unlikely to be microbiome-dependent."

Study 1:

The relative abundance of the five species listed in Table 5 (relative abundance listed in Table 6) in ObT2 subjects (n=68) and healthy controls (n=55) was determined by metagenomic sequencing and taxonomy specified as described in the Methods section. Decision trees were generated from the data in Table 6 using a random forest with parameters: trees=801, mtry=3.

The risk of ObT2 was determined in a 50-year-old male subject (FB002). The relative abundance of the five species listed in Table 5 in the subject's fecal samples was determined by metagenomic sequencing and taxonomy specified as described in the Methods section. The relative abundance of the five species in this subject is shown in Table 7. The relative abundance was run along a decision tree, and the relative abundance in Table 6 was used as training data to generate a risk score. The subject's score was 0.997 (Figure 4A), therefore the subject was considered likely to have ObT2. The subject had a BMI of 41.7 and was diagnosed with T2DM.

Study 2:

The relative abundance of the five species listed in Table 5 (relative abundance listed in Table 6) in ObT2 subjects (n=68) and healthy controls (n=55) was determined by metagenomic sequencing and taxonomy specified as described in the Methods section. Decision trees were generated from the data in Table 6 using a random forest with parameters: trees=801, mtry=3.

The risk of ObT2 was determined in a 45-year-old male subject (H45). The relative abundance of the five species listed in Table 5 in the subject's fecal samples was determined by metagenomic sequencing and taxonomy specified as described in the Methods section. The relative abundance of the five species in this subject is shown in Table 7. The relative abundance was run along a decision tree, and the relative abundance in Table 6 was used as training data to generate a risk score. The subject's score was 0.137 (Figure 4B), therefore the subject was considered to have a low risk of ObT2. The subject had a BMI of 21.09 and did not have T2DM.

Example 2: Mixed donor FMT induces increased abundance and diversity of butyrate-producing bacteria

background

The aim of this study was to determine how the human gut microbiome is associated with obesity and type 2 diabetes (T2D). Practical applications of this invention include assessing the risk of obesity and T2D-related diseases based on the presence and abundance of certain bacterial species in the gastrointestinal tract of test subjects, and evaluating whether obesity and T2D in test subjects are associated with the gut microbiome, particularly the bacterialbiome.

method

Research subjects and research design

Two FMT studies, a non-intensive randomized controlled trial (nFMT) and an intensive FMT study (iFMT), were conducted and compared. In both studies, obese participants aged 18–70 years with a body mass index (BMI) ≥28 kg/ ^m² and ≤45 kg/ ^m² were recruited from tertiary referral centers in Hong Kong, China (ClinicalTrials.gov NCT03789461, NCT03127696). Patients who had used weight-loss medications in the past year, or who had immunodeficiency syndromes, contraindications to oral esophagogastric endoscopy (OGD), a history of food allergies, severe organ failure including decompensated cirrhosis, inflammatory bowel disease, renal failure, epilepsy, malignancies known within the past two years, and active sepsis were excluded. Participants were excluded if they took antibiotics or probiotics during the 12-week screening period, or if they took sodium-glucose cotransporter-2 inhibitors, glucagon-like peptide-1 (GLP-1) receptor agonists, or proton pump inhibitors at randomization. Antibiotics, probiotics, or prebiotics were prohibited during the study. Patients maintained the same doses of oral hypoglycemic and lipid-lowering medications throughout the study. All patients provided written informed consent. This study was approved by the Clinical Research Ethics Committee of the New Territories East Cluster of the Chinese University of Hong Kong (2016.136-T and 2018.444).

FMT solution

In the nFMT study, recipients received a non-intensive FMT infusion once a month for 4 months (4 FMTs in total). The FMT infusions were derived from a mixture of at least two lean donors. In the iFMT study, each subject received a 3-day course of antibiotics (vancomycin, metronidazole, and amoxicillin, 500 mg each, three times daily), followed by 5 consecutive days of single-donor FMT per week for 4 weeks (20 FMTs in total). Serial stool samples were collected from all recipients at four time points (baseline, one month after the first FMT, one month after the last FMT, and 2–3 months after the last FMT) (Figure 5A).

FMT donor

According to a set of stringent criteria ¹⁰ as previously described, the FMT solution was derived from lean donors with a BMI < 23 kg/ ^m² . Fecal samples from eligible donors were collected within 4 hours of defecation and visually inspected for suitability (formed stool, absence of blood or mucus). The donor feces were homogenized with isotonic saline and glycerol, filtered, and then stored at -80°C. Under conscious sedation, the FMT solution (50 g/m³ of feces in 100–200 ml saline) from a single donor or a collection of donor feces was infused into the distal duodenum via esophagogastric-duodenoscopy (OGD).

Shotgun metagenomic sequencing and fecal microbiota analysis

Fecal DNA for bacterial metagenomic sequencing was isolated using the RSC PureFood GMO and Authentication Kit, following the manufacturer's instructions. DNA libraries were constructed through end repair, purification, and PCR amplification, and sequenced using a 150bp sequencing strategy with paired ends via an Illumina Novaseq 6000 (Novogene, Beijing, China), generating 93.5 million ± 15.2 million (mean ± standard deviation, SD) raw reads per sample. Metagenomic reads were quality filtered and pruned using Trimmomatic ¹¹ (v0.38) and purified against the human genome (reference: hg38) using Kneaddata (v0.7.2, bitbucket.org/biobakery/kneaddata/wiki/Home). Species-level metagenomic analysis was generated using MetaPhlAn2 ¹² (v2.6.0). Strain-level metagenomic analysis was generated using StrainPhlAn ¹³ (v3). Abundance tables were processed using R v3.6.0 and the tidyverse ¹⁴ (v1.2.1), ggpubr ¹⁵ (v0.2), and phyloseq ¹⁶ (v1.24.2) R packages. Raw sequencing data were available on NCBI using Bioproject PRJNA644456 and RJNA633456.

Related variations in bacterial species

As described above, section ¹⁷ identified relevant variations in bacterial species after FMT. In short, the relative changes in bacterial species were calculated as the difference in relative abundance between each post-FMT sample and the baseline sample. The correlations of the relative changes in bacterial species at each post-FMT time point were then tabulated. Variations were considered relevant when the correlation was significant (p<0.05, Pearson correlation) one month after the last FMT and 2–3 months after the last FMT.

Statistical analysis

Continuous variables are expressed as mean ± SD or median (25th to 75th percentile, P25–P75) (as applicable), while categorical variables are expressed as numbers (percentages). Repeated measures ANOVA (logarithmically transformed for skewed variables) was applied for inter-group comparisons. The Wilcoxon rank-sum test was used to investigate significant differences in inter-group comparisons. The Wilcoxon signed-rank test was used to compare data between different time points within the same treatment. ^Linear discriminant analysis was used to determine the effect size (LEfSe) between the two groups. The Bray curtis distance after central log-ratio ^{transformation} was used to assess differences between samples before and after FMT. All statistical tests were two-sided. Statistical significance was considered as P < 0.05.

result

Research subjects

In the nFMT study, a total of 38 obese subjects with a BMI ranging from 28.0 to 44.9 kg/ ^m² received fecal infusions from 2–5 lean donors. In the iFMT study, 9 obese subjects with a BMI ranging from 31.9 to 41.5 kg/ ^m² received fecal infusions from a single lean donor. Recipient baseline characteristics are summarized in Table 8.

Table 8: Baseline characteristics of subjects in each study

Data are presented as the number of subjects (%) or the median (P25–P75). Abbreviations: BMI, Body Mass Index; LDL, Low-Density Lipoprotein; HDL, High-Density Lipoprotein; ALT, Alanine Aminotransferase.

The effect of intensive versus non-intensive FMT on weight loss in obese subjects

Following FMT intervention, obese participants in both studies showed heterogeneous weight loss compared to baseline (nFMT 3.1% ± 4.8% vs. iFMT 4.8% ± 1.7%, mean ± SD, maximum weight loss during 52 weeks of follow-up). No significant improvement in weight loss was observed in iFMT recipients compared to nFMT (repeated measures ANOVA, p = 0.403, Figure 5B). None of the nine participants receiving iFMT achieved ≥10% weight loss, while 13.2% (5 out of 38) of the participants receiving nFMT achieved ≥10% weight loss (maximum weight loss during 52 weeks of follow-up, Figure 5B).

Dense FMT leads to an increased number of lean donor-sourced species and a distribution similar to that of the donor microbiome.

Compared with obese subjects receiving nFMT, obese subjects receiving iFMT had significantly more donor-derived bacterial species one month after the first FMT and 2–3 months after the last FMT infusion (p = 0.03 and p < 0.01, Wilcoxon rank-sum test, Figure 6A). In subjects receiving iFMT, the aggregation abundance of donor-derived bacterial species was significantly higher one month after the first FMT than in subjects receiving nFMT (p = 0.02, Wilcoxon rank-sum test, Figure 6B). The Brecurtis distance between baseline and post-FMT samples in subjects receiving iFMT was significantly greater than that in subjects receiving nFMT, indicating that iFMT conferred greater variation in overall bacterial composition (p < 0.001, Wilcoxon rank-sum test, Figure 6C). One month after the last FMT, the Brectis distance between the post-FMT samples of iFMT recipients and the corresponding donor samples was significantly smaller than that of nFMT recipients (p = 0.06, Wilcoxon rank-sum test, Figure 6D), indicating that the post-FMT bacterial distribution showed a more similarity to the bacterial distribution of its corresponding donor compared to the post-nFMT bacterial distribution.

Mixed donor FMT was associated with increased abundance and diversity of butyrate-producing bacteria in obese individuals.

In the nFMT study, where each subject received fecal infusions from 2–5 lean donors, a significant increase in butyrate-producing bacteria was observed, including eubacterial species, *Rosbyrus human*, *Anaerostipeshadrus ^{* 20} , *Proteus vulgaris ^{* 21} , and *Collinsella* species ²² (Fig. 7A, Fig. 9, LDA>2, p<0.05). Compared to baseline samples, samples after FMT showed significantly higher Chao1 abundance, Shannon diversity, and aggregation abundance of butyrate-producing species (p<0.01 and p<0.05, Wilcoxon signed-rank test, Fig. 7B, C). In contrast, in the iFMT study, where recipients received single-donor FMT, there was no significant increase in Chao1 abundance, Shannon diversity, or aggregation abundance of butyrate-producing species (Fig. 7A–C, Fig. 10). The abundance of *Bifidobacterium bifidum* (which has been shown to interact with butyrate-producing bacteria via mutualistic interactions^²³) was significantly increased in samples following nFMT, but not in samples following iFMT (Fig. 9, LDA>2, p<0.05). Changes in major butyrate-producing species remained consistent one month and 2–3 months after the last FMT (Fig. 7D, Fig. 11), indicating that these species maintained relevant variation despite significant perturbation after FMT. The richness of the overall bacterial community was also significantly increased in nFMT recipients one month after the first FMT and 2–3 months after the last FMT (p<0.01 and p<0.05, Wilcoxon signed-rank test, Fig. 7E) ^¹⁰ , while no significant changes were observed in iFMT recipients. These results suggest that mixed-donor FMT, rather than single-donor FMT, is associated with increased abundance and diversity of butyrate-producing bacteria in obese subjects.

Compared to non-dense FMT, dense FMT leads to increased replacement of butyrate-producing bacterial strains.

The inventors then sought strain implantation or replacement in recipients of the primary butyrate-producing bacteria. *E. hominis*, *Bacillus pretzel*, and *Anaerostipes hadrus* were present in over 50% of FMT recipients, forming distinct clusters based on SNP haplotype distribution (Fig. 8A, Fig. 12). Recipients receiving intensive FMT had a higher proportion of strain implantation or replacement compared to those receiving non-intensive FMT at the second follow-up (*E. hominis*: 77.8% vs. 26.3%, *Bacillus pretzel*: 66.7% vs. 57.9%, *Anaerostipes hadrus*: 88.9% vs. 52.6%, intensive vs. non-intensive FMT, Fig. 8B). This indicates that intensive FMT is more effective in replacing the original strain with a donor-derived strain.

discuss

This study aimed to investigate whether intensive fat-transfer microbiome therapy (FMT) could improve donor implantation and induce weight loss, and to evaluate factors influencing FMT outcomes in obese subjects. It was found that intensive FMT did not induce more weight loss compared to monthly FMT with mixed donors. Although intensive FMT induced significantly higher numbers of lean donor-derived species, the aggregation abundance of donor-derived species only transiently increased compared to monthly FMT with mixed donors. Conversely, monthly FMT with mixed donors was more effective in inducing an increase in butyrate-producing bacteria and induced significant weight loss (≥10%) in the obese recipient subgroup. Among all baseline factors, the baseline microbiome composition of recipients showed the strongest ability to predict weight change. High baseline levels of *B. doreei* were associated with greater weight loss after FMT.

Mixed-donor intensive FMT has shown increased FMT efficacy in diseases such as ulcerative colitis^²⁴. Changes in microbiome distribution in obese patients were compared after single-donor intensive FMT versus mixed-donor monthly FMT. The hypothesis is that frequent intensive FMT following antibiotic administration can enhance microbiome alteration and improve clinical outcomes in obese subjects. As expected, intensive FMT resulted in an increased abundance of donor-derived species compared to mixed-donor FMT. However, the abundance of donor-derived species aggregation only showed a significant difference one month after the first FMT compared to mixed-donor FMT, but no significant difference was observed during follow-up visits. Similarly, in subjects receiving intensive FMT, changes in overall composition and similarity to donor microbiome distribution peaked during FMT and one month after the last FMT, but decreased to levels similar to mixed-donor FMT two to three months after the last FMT infusion. These data suggest that single-donor intensive FMT only leads to a transient improvement in microbiome distribution compared to mixed-donor monthly FMT.

Butyrate-producing bacteria are a group of symbiotic bacteria capable of converting poorly digestible carbohydrates into butyrate, ²⁵ which has shown a reduction in circulating cholesterol ^levels26,27 . Following monthly fecal microbiota transplantation (FMT) with mixed donors, a widespread increase in multiple butyrate-producing species was observed, as indicated by increased abundance and increased diversity. In subjects receiving intensive single-donor FMT, the increase in butyrate-producing species showed high inter-person variability. This contrasts with previous single-donor FMT studies, where an increase in only a few butyrate-producing species was ^{observed28-30} . Previous studies have reported associations with licorisoflavan A, pyrrole, and p-cresol sulfate, and the levels of these metabolites were significantly associated with transfer of *F. preibryophyte* ^strains31 . However, no significant association between butyrate-producing strains and clinical outcomes has been observed, which may be related to factors such as limited sample size or insufficient strain variation to influence clinical outcomes. By interpreting relevant variation patterns in the recipient's fecal microbiota, the inventors of this application demonstrate that several butyrate-producing species act as covariant units that remain positively correlated with each other after FMT. Despite the same selection criteria, the distribution of the microbiome of lean donors varied greatly in terms of the presence and abundance of butyrate-producing species. Therefore, co-infusing butyrate-producing species with fecal samples from multiple donors can increase the colonization of butyrate-producing species in the recipient's gut.

All patents, patent applications and other publications (including GenBank accession numbers, etc.) cited in this application are incorporated in their entirety by reference for all purposes.

Reference List

1.Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 2014; 30: 2114-20.

2.Truong DT, Franzosa EA, Tickle TL, et al. MetaPhlAn2 for enhancedmetagenomic taxonomic profiling. Nat Methods 2015;12:902-3.

3.Langmead B,Salzberg SL.Fast gapped-read alignment with Bowtie 2.NatMethods 2012;9:357-9.

4.Hadley W, Mara A, Jennifer B, et al. Welcome to the Tidyverse. Journal of Open Source Software 2019; 4:1686.

5. McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data.PLoS One 2013;8:e61217.

6. Segata N, Izard J, Waldron L, et al. Metagenomic biomarker discovery and explanation. Genome Biol 2011; 12: R60.

7.Breiman L.Random Forests.Machine Learning 2001;45:5-32.

8. Cutler DR, Edwards Jr TC, Beard KH, et al. Random forests for classification in ecology. Ecology 2007; 88:2783-2792.

9. Robin

10.Ng SC, Xu Z, Mak JWY, et al. Microbiota engraftment after faecalmicrobiota transplantation in obese subjects with type 2diabetes: a 24-week, double-blind, randomised controlled trial. Gut 2021.

11.Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 2014; 30: 2114-20.

12.Truong DT, Franzosa EA, Tickle TL, et al. MetaPhlAn2 for enhancedmetagenomic taxonomic profiling. Nat Methods 2015;12:902-3.

13.Beghini F, McIver LJ, Blanco-Miguez A, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife 2021; 10.

14.Wickham H AM,Bryan J,et al. "Welcome to the tidyverse.".ournal ofOpen Source Software 2019;4(43),1686.

15.Wickham H.ggpubr:‘ggplot2’Based Publication Ready Plots.

16. McMurdie PJ, Holmes S. Phyloseq: a bioconductor package for handling and analysis of high-throughput phylogenetic sequence data. Pac Symp Biocomput2012:235-46.

17. Raman AS, Gehrig JL, Venkatesh S, et al. A sparse covarying unit that describes healthy and impaired human gut microbiota development. Science 2019;365:eaau4735.

18. Segata N, Izard J, Waldron L, et al. Metagenomic biomarker discovery and explanation. Genome Biol 2011; 12: R60.

19.Gloor GB, Macklaim JM, Pawlowsky-Glahn V, et al. Microbiome DatasetsAre Compositional: And This Is Not Optional. Front Microbiol 2017;8:2224.

20. Kant R, Rasinkangas P, Satokari R, et al. Genome Sequence of the Butyrate-Producing Anaerobic Bacterium Anaerostipes hadrus PEL85. Microbiology Resource Announcements 2015; 3.

21. Vital M, Karch A, Pieper DH. Colonic Butyrate-Producing Communities in Humans: an Overview Using Omics Data.mSystems 2017; 2:e00130-17,/msystems/2/6/msys.00130-17.atom.

22.Qin PP, Zou YQ, Dai Y, et al.Characterization a Novel Butyric Acid-Producing Bacterium Collinsella aerofaciens Subsp.ShenzhenensisSubsp.Nov.Microorganisms 2019; 7.

23. Rivière A, Selak M, Lantin D, et al. Bifidobacteria and Butyrate-Producing Colon Bacteria: Importance and Strategies for Their Stimulation in the Human Gut. Frontiers in Microbiology 2016; 7.

24.Paramsothy S, Kamm MA, Kaakoush NO, et al. Multidonor intensive faecalmicrobiota transplantation for active ulcerative colitis: a randomized placebo-controlled trial. The Lancet 2017;389:1218-1228.25.Fu

26. Canfora EE, Jocken JW, Blaak EE. Short-chain fatty acids in control of body weight and insulin sensitivity. Nat Rev Endocrinol 2015; 11:577-91.

27. Kenny DJ, Plichta DR, Shungin D, et al. Cholesterol Metabolism by Uncultured Human Gut Bacteria Influences Host Cholesterol Level. Cell Host Microbe 2020; 28:245-257 e6.

28.Kootte RS,Levin E,J,et al.Improvement of InsulinSensitivity after Lean Donor Feces in Metabolic Syndrome Is Driven byBaseline Intestinal Microbiota Composition.Cell Metabolism 2017;26:611-619.e6.

29. Dao MC, Everard A, Aron-Wisnewsky J, et al. Akkermansia muciniphila and improved metabolic health during a dietary intervention in obesity:relationship with gut microbiome richness and ecology. Gut 2016;65:426-436.

30.Forslund K, Hildebrand F, Nielsen T, et al. Disentangling type 2diabetes and metformin treatment signatures in the human gutmicrobiota. Nature 2015;528:262-+.

31.Chen L, Wang D, Garmaeva S, et al. The long-term genetic stability and individual specificity of the human gut microbiome. Cell 2021; 184:2302-2315e12.

Claims (31)

1. A method for reducing the risk of obesity and type 2 diabetes (T2D) in a subject or for treating obesity and T2D in a subject, comprising introducing an effective amount of one or more bacterial species into the gastrointestinal tract of the subject, said bacterial species being selected from Faecalibacterium prausnitzi, Bifidobacterium longum, Eubacterium halli, Bifidobacterium bifidum, Roseburia intestinalis, Eubacterium eligens, Lachnospiraceae bacterium 5163FAA, Eubacterium ventriosum, and Roseburia hominis. 2. The method of claim 1, wherein the introduction step comprises orally administering to the subject a composition comprising an effective amount of the one or more bacterial species. 3. The method of claim 1, wherein the introduction step comprises delivering a composition comprising an effective amount of the one or more bacterial species to the small intestine, ileum, or large intestine of the subject. 4. The method of claim 1, wherein the introduction step comprises fecal microbiota transplantation (FMT). 5. The method of claim 4, wherein the FMT comprises applying to the object a composition comprising processed donor fecal material. 6. The method according to any one of claims 5, wherein the processed donor fecal material is derived from at least two lean donors. 7. The method according to any one of claims 1 to 6, wherein the composition does not contain any species in detectable amounts of Table 2 or 4. 8. The method of claim 2, wherein the composition is administered orally. 9. The method of claim 2, wherein the composition is delivered directly to the gastrointestinal tract of the subject. 10. The method of claim 1, wherein the level or relative abundance of the one or more bacterial species is determined in a first fecal sample obtained from the object before the introduction step and a second fecal sample obtained from the object after the introduction step. 11. The method of claim 10, wherein the level of the one or more bacterial species is determined by polymerase chain reaction (PCR), preferably quantitative polymerase chain reaction (qPCR). 12. A method for assessing the risk of obesity and type 2 diabetes (T2D) in subjects, comprising: (1) Determine the level or relative abundance of one or more bacterial species shown in Tables 1-5 in fecal samples from the subject; (2) Determine the level or relative abundance of the same bacterial species in fecal samples from a reference group, which includes subjects with obesity and T2D and subjects without obesity and T2D; (3) Using the data obtained from step (2), generate a decision tree through a random forest model, and run the level or relative abundance of one or more bacterial species from step (1) along the decision tree to generate a score; and (4) Subjects with a score greater than 0.5 were identified as having an increased risk of obesity and T2D, and subjects with a score not greater than 0.5 were identified as not having an increased risk of obesity and T2D. 13. The method of claim 12, wherein the one or more bacterial species comprises any two or three bacterial species shown in Tables 1-5. 14. The method of claim 12, wherein the subject has not been diagnosed with obesity. 15. The method of claim 12, wherein the subject has not been diagnosed with T2D. 16. The method of claim 12, wherein each of steps (1) and (2) comprises metagenomic sequencing. 17. The method of claim 12, wherein each of steps (1) and (2) comprises polymerase chain reaction (PCR). 18. The method of claim 17, wherein the PCR is quantitative PCR. 19. The method according to claim 12, wherein the bacterial species is (i) *Clostridium bartlettii*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA, *Escherichia coli* or (ii) *Clostridium bartlettii* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyceae* 5_1_63FAA or (vi) *Escherichia coli* or (vii) *Clostridium bartlettii*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA or (viii) *Clostridium bartlettii*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA, *Escherichia coli*, *E. protuberans* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyceae* 5_1_63FAA, *E. protuberans* or (x) *Clostridium pallidum*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA, *E. protuberans* or (xi) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* 5_1_63FAA, *E. protuberans* or (xii) *Clostridium pallidum*, *Trichophyceae* 5_1_63FAA, *E. protuberans* or (xiii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyceae* 5_1_63FAA or (xiv) *Haemophilus parainfluenzae*, *Escherichia coli* or (xvi) *Haemophilus parainfluenzae*, *Escherichia coli*. 20. A method for assessing whether a subject has microbiome-dependent obesity and type 2 diabetes mellitus (T2D), comprising: (1) Determine the level or relative abundance of one or more bacterial species shown in Tables 1-5 in fecal samples from the subject; (2) Determine the level or relative abundance of the same bacterial species in fecal samples from a reference group, which includes subjects with obesity and T2D and subjects without obesity and T2D; (3) Using the data obtained from step (2), generate a decision tree through a random forest model, and run the level or relative abundance of one or more bacterial species from step (1) along the decision tree to generate a score; and (4) Subjects with a score greater than 0.5 were identified as having microbiome-dependent obesity and T2D, and subjects with a score less than 0.5 were identified as having microbiome-independent obesity and T2D. 21. The method of claim 20, wherein the subject has been diagnosed with obesity. 22. The method of claim 20, wherein the subject has been diagnosed with T2D. 23. The method of claim 20, wherein each of steps (1) and (2) comprises metagenomic sequencing. 24. The method of claim 20, wherein each of steps (1) and (2) comprises polymerase chain reaction (PCR). 25. The method of claim 24, wherein the PCR is quantitative PCR (qPCR). 26. The method of claim 20, wherein the bacterial species is (i) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* or (ii) *Clostridium pallidum* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyton spp.* 5_1_63FAA or (vi) *E. truncatella* or (vii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA or (viii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *E. truncatella* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella ... 63FAA, *E. convexis* or (x) *Clostridium pallida*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xi) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xii) *Clostridium pallida*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xiii) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Trichophyceae* bacteria 5_1_63FAA or (xiv) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA or (xv) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Escherichia coli* or (xvi) *Haemophilus parainfluenzae*, *Escherichia coli*. 27. A kit for assessing the risk of obesity and type 2 diabetes (T2D) in a subject or for assessing whether a subject has microbiome-dependent obesity and T2D, comprising reagents for detecting one or more bacterial species shown in Tables 1-5. 28. The kit of claim 27, wherein the reagent comprises a set of oligonucleotide primers for amplifying polynucleotide sequences from any of the bacterial species shown in Tables 1-5. 29. The kit according to claim 28, wherein the amplification is PCR. 30. The kit according to claim 29, wherein the PCR is quantitative PCR. 31. The kit according to claim 27, wherein the bacterial species is (i) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella* or (ii) *Clostridium pallidum* or (iii) *Haemophilus parainfluenzae* or (iv) *Escherichia coli* or (v) *Trichophyton spp.* 5_1_63FAA or (vi) *E. truncatella* or (vii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyton spp.* 5_1_63FAA or (viii) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Escherichia coli*, *E. truncatella* or (ix) *Clostridium pallidum*, *Haemophilus parainfluenzae*, *Trichophyton spp.* 5_1_63FAA, *E. truncatella ... 63FAA, *E. convexis* or (x) *Clostridium pallida*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xi) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xii) *Clostridium pallida*, *Trichophyceae* bacteria 5_1_63FAA, *E. convexis* or (xiii) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Trichophyceae* bacteria 5_1_63FAA or (xiv) *Haemophilus parainfluenzae*, *Escherichia coli*, *Trichophyceae* bacteria 5_1_63FAA or (xv) *Clostridium pallida*, *Haemophilus parainfluenzae*, *Escherichia coli* or (xvi) *Haemophilus parainfluenzae*, *Escherichia coli*.