Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
MODULATION OF THE GUT MICROBIOME BY NOVEL SYNTHETIC GLYCANS FOR THE PRODUCTION OF
PROPIONATE AND THE REDUCTION OF CARDIOMETABOLIC RISK FACTORS
Yves A. Millet1, Jeffrey Meisner1, Jie Tan1, Adarsh Jose1, Eric Humphries1, Kelsey J. Miller1, Charlie Bayne1,
Megan McComb1, Michael Giuggio1, Camille M. Konopnicki1, David B. Belanger1, Lingyao Li1, Han Yuan1,
Madeline Rosini1, Hoa Luong1, Jared Martin1, Zhengzheng Pan1, C. Ronald Kahn2 and Johan E.T. van
Hylckama Vlieg1
1Kaleido Biosciences, 65 Hayden Ave, Lexington, MA 02421, United States
2Integrative Physiology and Metabolism, Joslin Diabetes Center, Harvard Medical School, Boston, MA,
02215, United States
Short title: Novel synthetic glycan reduces CVD risk factors
Corresponding author:
Johan van Hylckama Vlieg
johan.van-hylckama-vlieg@kaleido.com
Kaleido Biosciences, 65 Hayden Ave, Lexington, MA 02421
Total word count: 10457
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
1
ABSTRACT 1
Background 2
Increasing evidence indicates that an altered gut microbiome participates in the development of 3
cardiometabolic syndrome and associated risk factors, such as insulin resistance, dyslipidemia, and 4
obesity, and that targeting the gut microbiome is a promising strategy to lower the risk for 5
cardiometabolic diseases. Part of this reduction is mediated by specific metabolites generated by the gut 6
microbiome. Propionate, a short-chain fatty acid (SCFA) produced from dietary glycans by certain gut 7
microbes is known to exert multiple beneficial metabolic effects. Here, we identify KB39, a novel gut 8
microbiome-targeting synthetic glycan selected for its strong propionate-producing capacity, and 9
demonstrate its effects in vivo to reduce cardiometabolic disease using western diet-fed LDL receptor 10
knock-out mice. 11
Methods 12
Ex vivo fermentation screening of a large library of synthetic glycan ensembles was performed 13
using gut microbiome communities from healthy subjects and overweight patients with type 2 diabetes. 14
A synthetic glycan identified for its high propionate-producing capacity (KB39) was then tested in vivo for 15
effects on systemic, blood and cecal metabolic parameters in Ldlr-/- mice fed a western diet. 16
Results 17
Ex vivo screening of ~600 synthetic glycans using human gut microbiota from healthy subjects and 18
patients with type 2 diabetes identified a novel glycan (KB39) with high propionate-producing capacity 19
that increased propionate contribution to total SCFA and propionate-producing bacterial taxa compared 20
to negative control. In western diet-fed Ldlr-/- mice, KB39 treatment resulted in an enrichment in 21
propiogenic bacteria and propionate biosynthetic genes in vivo and an increase in absolute and relative 22
amounts of propionate in the cecum. This also resulted in significant decreases in serum total cholesterol, 23
LDL-cholesterol, and insulin levels, as well as reduced hepatic triglycerides and cholesterol content 24
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
2
compared to non-treated animals. Importantly, KB39 treatment significantly reduced atherosclerosis, 25
liver steatosis and inflammation, upregulated hepatic expression of genes involved in fatty acid oxidation 26
and downregulated transcriptional markers of inflammation, fibrosis and insulin resistance with only a 27
mild lowering of body weight gain. 28
Conclusions 29
Our data show that KB39, a novel synthetic glycan supporting a high propionate-producing 30
microbiome, can reduce cardiometabolic risk factors and disease in mice and suggest this approach could 31
be of benefit for the prevention or treatment of cardiometabolic diseases in humans. 32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
3
CLINICAL PERSPECTIVE 49
What is new? 50
• A novel synthetic glycan, KB39, was selected from a library of compounds for its high propionate-51
producing capacity and beneficial effects on the human gut microbiome composition 52
• KB39 modulates the gut microbiome for high propionate production and significantly reduces 53
cardiometabolic risk factors and disease in a murine model of cardiometabolic diseases 54
What are the clinical implications? 55
• KB39, delivered orally, could be of benefit for the prevention or treatment of cardiometabolic 56
diseases in humans 57
• The efficacy of KB39 in mice compared to the clinical drug fenofibrate justifies further study in 58
humans 59
60
61
62
63
64
65
66
67
68
69
70
71
72
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
4
INTRODUCTION 73
In the past 40 years, cardiometabolic diseases, comprising of cardiovascular disease (CVD), type 2 74
diabetes mellitus (T2DM), non-alcoholic fatty liver disease (NAFLD) and other obesity-related 75
comorbidities, have become a major global health issue and are now the leading cause of death and 76
disability worldwide (WHO, World Health Statistics Report 2020). Lifestyle modifications have some effect 77
to reduce cardiometabolic risk factors (CMRFs) 1-7, however, this is not sufficient and most patients 78
ultimately require pharmacological therapies 8, 9, often with a multitude of drugs to address the many 79
individual components of disease risk 10, 11, increasing the risk of side effects, drug-drug interactions, poor 80
adherence and medication errors 12. Therefore, there is a great need to develop safe and effective drugs 81
targeting multiple CMRFs simultaneously for the prevention of cardiometabolic diseases. 82
Poor quality western diets, high in sugars, salt, saturated fats and low in fibers, play a major role 83
in the development of cardiometabolic diseases 13-15. These diets also lead to significant changes in the 84
bacterial composition and metabolic output of the gut microbiota 16, 17, and increasing evidence suggest 85
that this alteration of the gut microbiota contributes to the aggravation of multiple CMRFs, such as low-86
grade inflammation, insulin resistance, atherogenic dyslipidemia and fatty liver 18. Therefore, the gut 87
microbiome is emerging as a promising target in the prevention against cardiometabolic diseases 19. 88
Altered gut microbiome features caused by western diets and found in patients with CMRFs include an 89
enrichment in pro-inflammatory mucosa-damaging bacteria belonging to the Enterobacteriaceae family 90
and the depletion in fiber-degrading commensals leading to a reduction in the production and absorption 91
of short-chain fatty acids (SCFAs), a class of microbial fermentation products of dietary fibers, that may 92
have beneficial metabolic effects 18. Among the different SCFAs, propionate has received the most 93
attention in metabolism and obesity research due to its multiple effects against body weight gain, fatty 94
acid and cholesterol synthesis, insulin resistance, adipose tissue inflammation, atherosclerosis and 95
hypertensive cardiovascular damage in mice 20-24. 96
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
5
It is now established that glycan utilization by the gut microbes is a major determinant of the 97
composition and metabolic output of the gut microbiome 25 and the use of natural or enzymatically 98
synthesized fibers can be used to modulate the gut microbiome 26-30. However, in general, fibers as 99
therapeutics are not chosen by rational design, selection or optimized to serve a specific function. In this 100
study, we used an ex vivo fecal fermentation platform to screen a library of hundreds of novel synthetic 101
and well-characterized glycans 31 for propionate production from which we identified KB39 as a high 102
propionate-producing glycan. We also found that in vivo, KB39 promotes propionate production and the 103
enrichment of propiogenic bacterial taxa and has beneficial effects on blood lipids and insulin, and 104
preventing liver steatosis and inflammation as well as atherosclerosis in Ldlr-/- mice fed a high fat, high 105
cholesterol western diet. Our data show that glycans that promote propionate production such as KB39 106
can provide beneficial effects in lowering multiple CMRFs and preventing cardiometabolic diseases. Our 107
study also underscores the utility of screening large libraries of complex synthetic glycans for identifying 108
novel chemical entities with high therapeutic potential. 109
110
METHODS 111
Human fecal sample collection and fecal slurry preparation 112
Informed consent was obtained from all donors before fecal sample collection. T2DM subjects 113
were males and females between 18 and 70 years old with a BMI between 25 and 45 kg/m2 and currently 114
under treatment by a physician for T2DM. Fecal samples were collected into stool specimen containers, 115
immediately frozen and later stored at -80 °C. To prepare fecal slurries, samples were transferred into an 116
anaerobic chamber (Coy) and placed into filtered blender bags (Interscience). Samples were then diluted 117
in 1x PBS and glycerol to a final 20% w/v fecal slurry containing 15% glycerol. Diluted samples were 118
homogenized in a lab blender (Interscience 032230), aliquoted, flash frozen in a dry ice/ethanol bath and 119
stored at -80 °C. 120
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
6
Microbial cultivation 121
20% fecal slurry aliquots were transferred into an anaerobic chamber (Coy), thawed and further 122
diluted to a final 1% (w/v) using Clostridium minimal medium 32 supplemented with 0.1% (w/v) trypticase 123
peptone and 0.75 mM urea. 1% fecal suspensions were dispensed into the wells of 96-well deep well 124
plates containing 5 g/L (w/v) of the appropriate glycan or water used as a negative control. Each treatment 125
was tested in triplicates. Fecal microbial cultures were incubated anaerobically at 37 °C for 45 hours. Plates 126
were then centrifuged at 3,000 x g for 10 min at 4 °C. Culture supernatants were collected and stored at 127
-80 °C for acetate, butyrate and propionate quantification. The plates containing fecal bacterial cell pellets 128
were stored at -80 °C for DNA extraction and shallow shotgun sequencing. 129
Synthesis of KB39 130
D-(+)-Galactose (60.0 g, 0.333 mole), D-(+)-mannose (40.0 g, 0.222 mol), citric acid (3.27 g, 0.017 131
mol, 3 mol%), and deionized water (10 mL) were added to 1 L three-neck round-bottom flask. The flask 132
was equipped with a heating mantle configured with an overhead stirrer, a probe thermocouple, and a 133
reflux condenser in a distillation position to remove excess water throughout the course of the reaction. 134
The temperature controller was set to 130 °C and stirring was initiated as the temperature of the reaction 135
mixture was brought to 130 °C under ambient pressure. The resulting solution was maintained at 130 °C 136
for about 3.5 hours at which time most of the water was removed by distillation and the reaction was 137
determined to be complete by SEC. The heat was turned off and the crude reaction mixture was allowed 138
to cool to <80 °C while maintaining constant stirring. The crude reaction mixture was diluted with 139
deionized water (ca. 60 mL) to obtain a 40 °Bx solution. The resulting solution was passed through a 140
cationic exchange resin (Dowex® Monosphere 88H) column, an anionic exchange resin (Dowex® 141
Monosphere 77WBA) column, and a decolorizing polymer resin (Dowex® OptiPore SD-2) column. The 142
resulting aqueous solution of KB39 was adjusted to ~20 °Bx and then freeze-dried to afford the title 143
compound that could be used without additional purification. The KB39 samples used for ex vivo 144
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
7
experiments were further purified by chromatography using ISCO RediSep Gold Amine column on a 145
Biotage Isolera equipped with an ELSD detector and water/acetonitrile as mobile phase. 146
Glycan linkage analysis 147
Permethylation was performed as previously described with slight modification 33. The glycan 148
sample (500 µg) was dissolved in DMSO for 30 min with gentle stirring. A freshly prepared sodium 149
hydroxide suspension in DMSO was added, followed by a 10 min incubation. Iodomethane (100 µL) was 150
added, followed by a 20 min incubation. A repeated round of sodium hydroxide and iodomethane 151
treatment was performed for complete permethylation. The permethylated sample was extracted, 152
washed with dichloromethane (DCM), and blow dried with nitrogen gas. The sample was hydrolyzed (2M 153
TFA for 2h), reduced with sodium borodeuteride (10 mg mL-1 in 1M ammonia for overnight), and 154
acetylated using acetic anhydride/TFA. The derivatized material was extracted, washed with DCM, and 155
concentrated to 200 µL. Glycosyl linkage analysis was performed on an Agilent 7890A GC equipped with 156
a 5975C MSD detector (EI mode with 70 eV), using a 30-meter RESTEK RTX®-2330 capillary column. The 157
GC temperature program: 80 °C for 2 min, a ramp of 30 °C min-1 to 170 °C, a ramp of 4 °C min-1 to 245 158
°C, and a final holding time of 5 min. The helium flow rate was 1 mL/min, and the sample injection was 1 159
µL with a split ratio of 10:1. 160
DNA sequencing of fecal bacterial cell pellets, mouse fecal samples and microbiome analysis 161
The DNA from the fecal bacterial pellets and mouse fecal samples were characterized using a 162
shallow shotgun sequencing approach like the one previously described 34. Briefly, fecal samples’ DNA was 163
extracted with Qiagen’s DNeasy PowerSoil, quantified using the Quant-iT PicoGreen dsDNA assay (Thermo 164
Fisher) and libraries prepared using the NexteraXT kit and sequenced using HiSeq 1 × 150-cycle v3 kit 165
(Illumina). The operational taxonomic unit (OTU) count tables were generated from the DNA sequences 166
using the SHOGUN pipeline by aligning the filtered reads to a curated database containing all 167
representative genomes in RefSeq for bacteria with additional manually curated strains (Venti) using fully 168
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
8
gapped alignment with BURST (https://github.com/knights-lab/BURST) 34. The generated OTUs that could 169
be resolved to the species level and the samples that were retained after rarefying to 10,000 reads without 170
replacement were used for downstream analyses. For functional profiling, metagenomic samples were 171
profiled using HUMAnN2 v2.8.2 with default settings 35. Gene family files were regrouped using the 172
humann2_regroup_table command with 'uniref90_ko' and renormalized to relative abundance using 173
humann2_renorm_table including reads unmapped to strains and unintegrated to pathways. Taxa with 174
relative abundances greater than or equal to 0.1% in either treatment group were retained for 175
downstream analysis. KO (KEGG Orthology) families identified were retained for further analysis if (1) they 176
were among the top 100 terms in at least one sample and (2) they were associated with microbial 177
synthesis of propionate 36. To evaluate the treatment effect on the overall microbiome structure, we 178
calculated pair-wise Bray-Curtis dissimilarities based on square-root transformed relative abundances of 179
bacterial species. We then performed principal coordinate analysis on the dissimilarity matrix. To identify 180
taxa and KO families responding to KB39 treatment ex vivo, we performed paired Wilcoxon rank sum test 181
between the KB39 treated samples and water treated samples across different communities. For the 182
mouse in vivo study, we assessed whether the abundance of a taxon/KO term significantly differed 183
between the KB39 treatment group and the no treatment group at the end of the study using unpaired 184
Wilcoxon rank sum test. The test results were corrected using FDR. 185
Short-chain fatty acids quantification 186
Acetate, butyrate and propionate in microbial culture supernatants were quantified by gas 187
chromatography with flame-ionization detection (GC-FID) (7890A, Agilent). For each sample and 188
calibration standards, 50 μL was mixed with 20 μL of a 400mM 2-ethylbutyric acid (used as an internal 189
standard) solution prepared in HPLC water. Immediately before injection, each sample was acidified with 190
20 µL of 6% formic acid and 1 μL injected. The GC-FID conditions used were as follows: 15 m x 0.53 mm x 191
0.50 µm DB-FFAP column (Agilent) - carrier gas: helium at 28.819 mL/min - inlet conditions: 250 °C, 5 192
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
9
mL/min purge flow, 4:1 split ratio – oven temperature gradient: 70 °C, increase at a rate of 70 °C for 1 193
min, increase at a rate of 100 °C for 1 min, 1.8 min hold. Acetate, butyrate, propionate and succinate in 194
mouse cecal content samples were quantified by liquid chromatography with tandem mass spectrometry 195
(LC-MS/MS). Briefly, 4% (w/v) mouse cecal content homogenates were prepared in 1:1 (v/v) acetonitrile 196
(ACN):water by vortex mixing for 5 min and sonicating for 5 min to extract analytes. Homogenates were 197
then further diluted to 1% (w/v) in 1:1 (v/v) ACN:water, centrifuged and supernatants collected. 198
Supernatants were derivatized in 40 mM 3-nitrophenylhydrazine (3NPH), 37.5 mM N-(3-199
dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride, and 1.5% (v/v) pyridine in 1:1 v/v ACN:water 200
at 40 °C for 90 min. The derivatized samples were then further diluted with 10% ACN in water and spiked 201
with labeled internal standards (reacted similarly) before LC-MS/MS analysis. All standards were 202
purchased from Millipore-Sigma. For the HPLC, a Kinetex 2.6 μm C18 50 x 2.1 mm column was used and 203
0.01% formic acid in water and 0.01% formic acid in methanol were used as mobile phase A and mobile 204
phase B respectively. The MS/MS was run in positive mode and the transitions used were as follows: acetic 205
acid-3NPH: 196.33>136.99 – butyric acid-3NPH: 224.35>137.38 – propionic acid-3NPH: 210.34>137.29 – 206
succinic acid-3NPH: 389.35>137.11. The total amount of each SCFA was calculated based on the weight 207
of the cecum content for each animal. 208
Animals and treatment 209
Nine- to 11-week-old male C57BL/6J Ldlr-/- mice (Jackson Laboratory, B6.129S7-Ldlrtm1Her/J, Stock 210
No: 002207) were split into one of three treatment groups (12 animals per group) where they received 211
treatment for 16 weeks: 1) no treatment, western diet only; 2) western diet plus KB39; 3) western diet 212
plus fenofibrate. The western diet consisted of 40 kcal% fat and 1.5% cholesterol (Research Diets 213
D12079B). Mice in the KB39 group were fed the western diet supplemented with KB39 at 7.5% w/w. The 214
KB39 treatment diets were modified to provide the same caloric value to the mice as those mice in the no 215
treatment group. Mice in the fenofibrate group were fed the western diet supplemented with fenofibrate 216
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
10
(100 mg/kg/day). Mice in each treatment group received one week of normal chow before the study 217
began. The body weight for each animal was measured before the treatment period and weekly over the 218
treatment period. Four-day food intake for each cage was measured weekly over the treatment period 219
and averaged as food intake per mouse per day. 220
Mouse stool, blood and tissue 221
Fresh fecal samples were collected within a week before treatment initiation and during week 15 222
post treatment initiation. For each fecal collection period, one fresh fecal pellet was collected on three 223
separate days for each animal, frozen on dry ice and stored at -80 °C. At week 12, an oral glucose tolerance 224
test was performed. Blood was collected from a tail cut after a 6h fast and baseline glucose measured 225
using a glucometer (Lifescan, Johnson & Johnson). The mice then received 2 g/kg of a 100 mg/mL glucose 226
solution in sterile water delivered by oral gavage and blood was collected from the tail wound at 15, 30, 227
60 and 120 min for glucose measurement. On the day of termination, and an overnight fast, retroorbital 228
blood was collected under anesthesia for the quantification of serum total cholesterol (colorimetric assay, 229
WAKO Diagnostics), serum very low-density lipoprotein (VLDL), serum low-density lipoprotein (LDL) and 230
serum high-density lipoprotein (HDL) (Lipoprint, Qantimetrix), serum non-esterified fatty acids (NEFA) 231
(colorimetric assay, WAKO Diagnostics), serum insulin (MA2400 Mouse/Rat insulin kit K152BZC, Meso 232
Scale Discovery) and whole blood triglycerides (Cardiochek, PTS Diagnostics). The animals were then 233
euthanized, their cecum content collected, and their whole liver dissected and weighed. Liver sections 234
(~100 mg each) were collected and frozen for the quantification of hepatic triglycerides and hepatic total 235
cholesterol. One liver lobe was also preserved in 10% neutral buffered formalin (NBF) for embedding, 236
sectioning and staining with H&E before being scored for severity of steatosis, lobular inflammation and 237
hepatocellular ballooning 37. The NAFLD activity score (NAS) was calculated as the sum of liver steatosis, 238
lobular inflammation, and hepatocellular ballooning. The aortic arch was perfused with PBS, dissected, 239
fixed in 10% NBF and later surface stained with oil red O for the quantification of atherosclerotic plaque 240
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
11
surface area using the ImagePro Plus software. The heart was dissected, fixed in 10% NBF for embedding 241
and sectioning. Sections of the aortic sinus were stained with elastin trichrome and the atherosclerotic 242
plaque area quantified by morphometric analysis using the ImagePro Plus software. In addition, the foamy 243
macrophages, extracellular lipids and plaque fibrosis in the aortic sinus sections were scored blindly (0: 244
no change; 1: minimal change; 2: mild change, distinct but not prominent; 3: moderate change, prominent 245
tissue feature not effacing pre-existing structures; 4: marked to severe change, overwhelming tissue 246
feature that may efface pre-existing tissue features) and the overall plaque severity score calculated as 247
the sum of the three. 248
Hepatic gene expression analysis 249
Gene expression in liver tissue was analyzed by RNAseq. Total RNA was extracted and RNA 250
samples quantified using Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, CA, USA) and RNA integrity 251
was checked using Agilent TapeStation 4200 (Agilent Technologies, Palo Alto, CA, USA). RNA sequencing 252
libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina following 253
manufacturer’s instructions (NEB, Ipswich, MA, USA). The sequencing libraries were validated on the 254
Agilent TapeStation (Agilent Technologies, Palo Alto, CA, USA), and quantified by using Qubit 2.0 255
Fluorometer (Invitrogen, Carlsbad, CA) as well as by quantitative PCR (KAPA Biosystems, Wilmington, MA, 256
USA). The sequencing libraries were pooled and sequenced using a 2x150bp Paired End (PE) configuration 257
using a HiSeq 4000. Raw sequence data files were converted into fastq files and de-multiplexed using 258
Illumina's bcl2fastq 2.17 software. One mismatch was allowed for index sequence identification. After 259
investigating the quality of the raw data, sequence reads were trimmed to remove possible adapter 260
sequences and nucleotides with poor quality using Trimmomatic v.0.36. The trimmed reads were mapped 261
to the murine reference genome available on ENSEMBL using the STAR aligner v.2.5.2b. Unique gene hit 262
counts were calculated by using feature Counts from the Subread package v.1.5.2. Unique reads that fall 263
within exon regions were counted. After extraction of gene hit counts, the gene hit counts table were 264
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
12
used for downstream differential expression analysis. Using DESeq2 38, a comparison of gene expression 265
between the groups of samples was performed. The Wald test was used to generate P-values and Log2 266
fold changes. A principal component analysis (PCA) was performed on log transformed count per million 267
values (logCPM) to compare the treatment impact of KB39 or fenofibrate compared to no treatment on 268
the transcriptomics profile. 269
270
RESULTS 271
Screening of a library of synthetic glycans and identification of KB39 as a high propionate-producing 272
glycan 273
A library of hundreds of synthetic glycans with diverse monosaccharide composition, bond type, 274
branching, and degrees of polymerization was produced as previously described 31, of which 593 275
compounds (Figure S1) were screened for their ability to promote propionate production in a semi-276
automated ex vivo fermentation assay using the fecal microbiota from a healthy human subject. While 277
propionate could be identified in the culture supernatants, there was a wide-range of concentrations, 278
with some glycans supporting 10-fold higher propionate production than others (Figure 1A). This ex vivo 279
screen revealed one synthetic glycan, KB39, to be among the glycans enabling the highest propionate 280
production, increasing propionate by 14.5 mM compared to the negative control (Figure 1A, red bar vs 281
blue bar). KB39 (average MW = 2260 g/mol) is composed of D-galactose and D-mannose (60/40 ratio) 282
with complex distribution of glycosidic bonds linking the monosaccharides (Figure S2). The linkage and 283
extensive branching contrast with known naturally occurring dietary fibers 31. 284
The composition and metabolic function of gut microbiota can vary across individuals. To evaluate 285
the reproducibility of propionate production by different gut microbiota, we tested KB39 in our ex vivo 286
assay system with fecal microbiota samples from ten healthy subjects and 31 overweight subjects with 287
diagnosed T2DM, a population reported to have altered gut microbiomes 18. High propionate production 288
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
13
was observed for most fecal samples incubated with KB39 and was similar between healthy and T2DM 289
subjects (Figure 1B vs. 1C, center panels). The fermentation of KB39 with these fecal samples also resulted 290
in a similar increase in acetate production, and a more moderate increase in butyrate (Figure 1B, 1C). 291
Importantly, KB39 increased propionate fraction of total SCFAs, unlike acetate and butyrate (Figure S3). 292
This fermentation product selectivity suggests that KB39 was disproportionately metabolized by 293
propionate-producing bacteria. 294
KB39 enriches the propionate-producing phylum Bacteroidetes in fecal microbiota from overweight-295
T2DM subjects ex vivo 296
The increase and disproportionate production of propionate by fecal microbiota incubated with 297
KB39 suggested that it preferentially supported the growth of propiogenic commensal bacteria. Although 298
the ability to produce propionate is shared by numerous bacterial species in the gut, it is particularly 299
widespread among the species in the phylum Bacteroidetes. To identify the commensal bacteria primarily 300
responsible for KB39 fermentation in samples from overweight-T2DM patients, we subjected the ex vivo 301
fecal microbiota cultures to metagenomics sequencing reasoning that the bacteria responsible for KB39 302
fermentation would increase in relative abundance compared to the rest of the bacteria in the cultures. 303
We observed a significant enrichment of the phylum Bacteroidetes (7-fold increase), which was driven 304
primarily by an increase in the relative abundance of several species belonging to the genus 305
Parabacteroides (51-fold increase) in the family Tannerellaceae (Figure 2). We also observed a lower 306
magnitude enrichment of Bacteroides fragilis and Bacteroides thetaiotaomicron species. Thus, KB39 307
fermentation and the production of propionate likely derives from the growth and metabolic activity of 308
species of Parabacteroides and Bacteroides. In addition, metagenomic sequencing also revealed an 309
increase in the relative abundance of Eisenbergiella tayi, a species in the family Lachnospiraceae known 310
to produce butyrate as its main fermentation product, in addition to lactate, acetate and succinate 39. 311
Furthermore, significant increases in the relative abundance of the family Erysipelotrichaceae and the 312
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
14
species Blautia producta were also detected. The family Erysipelotrichaceae is commonly found in the 313
human gut microbiome and contains the genera Holdemania and Turicibacter known to produce acetate 314
and lactate, respectively 40, 41. Blautia producta produces lactate and acetate as its main products of 315
fermentation 42. By contrast, KB39 decreased the relative abundance of Citrobacter, Escherichia coli, 316
Enterococcus faecalis and Klebsiella pneumoniae, bacterial species generally considered as opportunistic 317
pathogens and pro-inflammatory. 318
KB39 increases cecal propionate and propionate-producing Bacteroidetes in vivo in western diet-fed 319
Ldlr-/- mice 320
The effect of KB39 on propionate production in vivo was determined in LDL receptor knockout 321
(Ldlr-/-) mice fed a western diet (high fat, high cholesterol), a model known to develop multiple 322
cardiometabolic risk factors (CMRFs) such as obesity, hypercholesterolemia, hyperinsulinemia, mild 323
hyperglycemia and insulin resistance resulting in fatty liver and atherosclerosis 43-45. Briefly, Ldlr-/- mice 324
were fed a western diet supplemented with KB39 or fenofibrate for 16 weeks. Fenofibrate, a peroxisome 325
proliferator receptor alpha (PPARα) agonist, was used as a positive control for its documented anti-326
obesity, lipid and glucose lowering effects in the western diet-fed Ldlr-/- mouse model 46, 47. Similar to what 327
was observed ex vivo in fecal communities from humans, treatment with KB39 increased the levels of 328
total cecal acetate and propionate compared to no treatment (Figure 3A). A small increase in cecal 329
butyrate was also observed, although non-significant. In addition, KB39 decreased the proportion of 330
acetate and butyrate and increased the proportion of cecal propionate compared to no treatment, similar 331
to what was observed ex vivo (Figure 3B) and suggesting an enrichment in propiogenic bacteria in vivo. 332
KB39 also increased the total amount and proportion of succinate (Figure 3A, 3B). The “succinate 333
pathway” of propionate formation is the major propionate biosynthetic pathway in the large intestine for 334
taxa of the Bacteroidetes phylum 36. This result suggests that KB39 is converted to propionate via the 335
succinate pathway in vivo and that succinate might accumulate due to limiting factors such as vitamin B12 336
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
15
48. As expected, fenofibrate did not modify the total amount or proportion of SCFAs or succinate (Figure 337
3A, 3B). 338
The effect of KB39 on the composition of the gut microbiome was determined by shotgun 339
sequencing of stool samples collected before and at the end of the treatment period. The principal 340
coordinate analysis (PCoA) plot of the microbiome composition data shown in Figure S4 shows that the 341
western diet had a profound impact on the fecal microbiome composition of Ldlr-/- mice. Western diet 342
feeding led to a decrease in major SCFA producing taxa, including Bacteroidetes and Ruminococcaceae, 343
and the reduction in Akkermansia muciniphila, a species associated with reduced obesity and type 2 344
diabetes 49, in non-treated animals (Figure S5). Unlike fenofibrate, treatment with KB39 led to a widely 345
distinct microbiome structure (Figure S4). Compared to no treatment, KB39 enriched propiogenic families 346
and genera belonging to the Bacteroidetes phylum, including Rickenellaceae, Tannerellaceae, Alistipes 347
and Parabacteroides (Figure 4). As noted above, this enrichment in propionate producing Bacteroidetes, 348
which are known to utilize the “succinate pathway” for propionate formation, is consistent with the ex 349
vivo data and the observed increase in cecal succinate in KB39-treated animals compared to no treatment. 350
This was corroborated by the increase, both ex vivo and in vivo, in the abundance of sequencing reads 351
mapping to genes encoding the methylmalonyl-CoA epimerase enzyme, essential for propionate 352
production via the “succinate pathway” and used by Bacteroidetes bacterial species (Figure S6) and no 353
response of genes associated with the two other (acrylate and propanediol) propionate synthesis 354
pathways 36. KB39 also promoted expansion of the acetate producer Blautia producta and depletion of 355
Enterococcus faecalis, consistent with what was observed ex vivo with human fecal gut microbiota. 356
Notably, KB39 also enriched Akkermansia muciniphila mentioned previously, and Parasutterella, a genus 357
belonging to the Proteobacteria phylum and recently shown to be associated with reduced LDL-C in 358
humans 50. 359
KB39 attenuates hypercholesterolemia, hyperinsulinemia in WD-fed Ldlr-/- mice 360
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
16
Following on the confirmation that KB39 increases propionate and propionate-producing gut 361
bacteria in vivo, we evaluated the effect of KB39 on the metabolic parameters of western diet fed Ldlr-/- 362
mice. Unlike fenofibrate that dramatically prevented body weight gain, KB39 only had a modest effect on 363
relative body weight gain during the treatment period (Figure 5A), and no effect on food intake (Figure
364
5B). However, KB39 treatment greatly lowered serum total cholesterol (43%), LDL-C (39%), VLDL-C (58%), 365
and HDL-C (47%) and provided a lowering trend for non-esterified fatty acids (NEFA) (16%) compared with 366
no treatment (Table 1). By contrast, treatment with fenofibrate provided non-significant lipid lowering 367
trends except for a decrease in serum LDL-C (23%) and NEFA (41%) (Table 1). In addition to the large 368
decrease in serum cholesterol, KB39 also significantly reduced hepatic total cholesterol (39%) and hepatic 369
triglycerides (37%) compared with no treatment. Fenofibrate appeared to lead to a similar decrease in 370
hepatic cholesterol (33%) as KB39 but provided a larger decrease in hepatic triglycerides (73%). The effect 371
of KB39 on liver fat was accompanied by lower liver weights (Table 1). By contrast, consistent with 372
previous reports of hepatomegaly and peroxisome proliferation in response to fibrates in mice 51, liver 373
weights in mice treated with fenofibrate significantly increased, despite reductions in hepatic cholesterol 374
and triglyceride levels. Both KB39 and fenofibrate decreased fasting serum insulin levels by ~66%, 375
consistent with reduced insulin resistance, but only fenofibrate significantly reduced fasting blood glucose 376
levels or improved glucose tolerance during an OGTT (Table 1, Figure 5C). Taken together, these data 377
show that KB39 reduces multiple CMRFs with, in the cases of cholesterol and insulin, effect sizes 378
comparable to fenofibrate, a compound used to treat high cholesterol and triglycerides. 379
KB39 attenuates liver steatosis, liver inflammation and atherosclerosis in western diet-fed Ldlr-/- mice 380
The beneficial metabolic effects of KB39 in western diet-fed Ldlr-/- mice reported in Table 1 were 381
associated with improvement in liver histopathology, with decreases in liver steatosis (-1 point), 382
inflammation (-1 point), and overall NAFLD activity score (NAS) (-2.5 points) (Figure 5D, 5E, 5F). Treatment 383
with fenofibrate demonstrated significant decreases in liver steatosis and NAS but had no significant effect 384
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
17
on liver inflammation. In addition, consistent with its LDL-C lowering effect, KB39 also had very significant 385
effects on atherosclerosis in this mouse model, with 64% reduction in atherosclerotic plaque formation in 386
the aortic arch (Figure 5I) and a 38% reduction of plaque in the aortic sinus (Figure 5H). KB39 also led to a 387
significant reduction in the overall plaque severity score in the aortic sinus (-2 point) (Figure 5G). Thus, 388
KB39 treatment reduced biochemical markers of cardiometabolic disease risk, reduced fatty liver disease, 389
and produced a major reduction in atherosclerosis development in Ldlr-/- mice on western diet. 390
KB39 upregulates genes involved in fatty acid oxidation and repressed markers of inflammation and 391
insulin resistance in the liver of western diet-fed Ldlr-/- mice 392
Given the cholesterol and insulin lowering effects of KB39 described herein and the central role 393
played by the liver in lipid and glucose metabolism, we compared the transcriptional profile of the non-394
treated, KB39 and fenofibrate-treated western-diet fed Ldlr-/- mice using RNA-seq. The PCA of liver RNA-395
seq data presented in Figure S7 shows a distinct liver transcriptional profile between the different groups, 396
confirming the differences observed in liver histology and biochemistry. As shown in Table 2, KB39 397
significantly increased the expression of PPARα, a central regulator of fatty acid oxidation 52, compared to 398
no treatment. Like fenofibrate, a known PPARα agonist, KB39 also increased the expression of genes 399
involved in fatty acid oxidation, including mitochondrial and microsomal oxidation. On the other hand, 400
KB39 significantly repressed the expression of PPARγ and SREBP1 (encoded by Pparg and Srebf1 genes), 401
two major transcriptional regulators involved in hepatic lipogenesis 52, 53. Accordingly, Scd1, encoding for 402
the stearoyl-coenzyme A desaturase 1, a critical enzyme for the synthesis of monounsaturated fatty acids 403
and triglycerides and positively regulated by PPARγ and SREBP1 54-56, was repressed by KB39 treatment. 404
By contrast, fenofibrate induced the expression of Srebf1, Scd1 and other major genes involved in fatty 405
acid synthesis and elongation controlled by SREBP1 (e.g., Acaca, Acacb, Fasn, Elovl6). These results are 406
consistent with previous studies that showed that PPARα agonists activate both fatty acid oxidation and 407
fatty acid synthesis simultaneously in the liver 57. In addition to fatty acid synthesis, KB39 and fenofibrate 408
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
18
also differed in their effect on bile acid synthesis and transport genes. KB39 increased the expression of 409
bile acid synthesis genes and the bile acid transporter genes Abcd11 (Ntcp) and Slc10a1 (Bsep). 410
Importantly, KB39 and fenofibrate increased the expression of the master regulator of cholesterol 411
synthesis SREBP2 (encoded by Srebf2) and numerous SREBP2-regulated cholesterol synthetic genes 412
compared to no treatment. These results are consistent with the observed decrease in hepatic 413
cholesterol, what is observed with other cholesterol lowering agents such as statins, and the negative 414
transcriptional regulation cholesterol exerts on its own synthesis in the liver 58, 59. Notably, and similar to 415
what is observed with statins 60, KB39 also increased the expression of the LDL receptor (LDLR). 416
Importantly, in addition to its effect on genes involved in lipid metabolism, KB39 decreased the expression 417
of markers of inflammation, oxidative stress, apoptosis and liver damage. This decrease was associated 418
with a reduction in markers of fibrosis, cell adhesion and macrophage activation, and were consistent with 419
the reduction in hepatic inflammation observed histologically. A downregulation of the markers of insulin 420
resistance Socs3 and Trib3 61, 62 was also observed for KB39 treated animals compared to no treatment, 421
supporting an increase in hepatic insulin sensitivity corroborated with the observed decrease in fasting 422
blood insulin. Interestingly, fenofibrate increased the expression of Trib3, a known PPARα target with 423
potential negative impacts on hepatic insulin signaling 63. KB39 also upregulated Ceacam1, an important 424
gene mediating hepatic insulin clearance and the maintenance of insulin sensitivity, known to be 425
downregulated in rodent models of obesity, insulin resistance and NAFLD as well as obese subjects with 426
insulin resistance 64. Finally, KB39 increased the expression of multiple markers of autophagy, an 427
important protection mechanism against liver damage and lipotoxicity 65. 428
429
DISCUSSION 430
In our study, we sought to identify novel glycans that might serve as optimal substrates for 431
propionate production by gut microbes and enrich the gut microbiota in propiogenic bacterial taxa, to 432
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
19
leverage the therapeutic potential of gut-derived propionate against CMRFs such as obesity, atherogenic 433
dyslipidemia and insulin resistance. To do so, we screened a large library of synthetic glycans for 434
propionate production using an ex vivo fermentation platform, and selected KB39 as glycan capable of 435
supporting high propionate production. We showed that KB39 greatly increased the SCFAs acetate and 436
propionate and increased the ratio of propionate over acetate and butyrate, both ex vivo, in fecal 437
communities from healthy subjects and overweight patients with T2DM and in vivo, in western diet-fed 438
Ldlr-/- mice. In the latter, KB39 demonstrated multiple beneficial effects, including a large decrease in 439
blood total cholesterol and LDL-cholesterol, and a strong reduction in hepatic cholesterol and 440
triglycerides, with only a mild attenuation in body weight gain. Those effects translated to a significant 441
reduction in liver steatosis and inflammation, and a marked reduction in atherosclerosis, which were 442
similar or superior to fenofibrate. In addition to its effect on lipids, KB39 strongly decreased fasting blood 443
insulin, suggesting that KB39 increased insulin sensitivity. This is of particular importance considering the 444
crucial role played by insulin resistance in the onset of T2DM and NAFLD and the increased risk of CVD 66. 445
To our knowledge, KB39 is the first synthetic glycan specifically selected for high propionate 446
production by the gut microbiota and offers significant advantages versus oral propionate. Unlike 447
propionate that has poor taste and smell and is readily absorbed in the upper gastrointestinal tract, KB39 448
is tasteless and odorless, non-absorbable, non-digestible and will be metabolized in the lower small 449
intestine and colon, the natural sites of propionate absorption and signaling through its molecular 450
receptors. In addition, and unlike oral propionate, KB39 has the potential to elicit beneficial changes in 451
the composition of the gut microbiome of patients. KB39 enriched propiogenic bacterial taxa belonging 452
to the Bacteroidetes phylum, including Bacteroides, Parabacteroides and Alistipes. Consistent with 453
Bacteroidetes using the succinate pathway to generate propionate from complex polysaccharides, KB39 454
treatment also led to an increase in total and relative cecal succinate in the Ldlr-/-mice in vivo. Succinate 455
itself is emerging as another important microbial metabolite improving metabolic dysfunctions 67, 68. In 456
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
20
addition to Bacteroidetes, KB39 also increased the relative abundance of Akkermansia muciniphila in vivo, 457
a microbe which has been shown to improve metabolic disorders and obesity through the production of 458
immunomodulatory proteins (such as Amuc_1100) and improve the gastrointestinal barrier function 69, 70. 459
KB39 also decreased bacterial species considered pro-inflammatory such as Escherichia coli, Enterococcus 460
faecalis and Klebsiella pneumoniae. This, along with the beneficial effects of propionate, could play a role 461
in KB39 anti-inflammatory properties observed in the liver. One potential mechanism of action for KB39 462
on lipid and glucose metabolism is the activation of the AMP-activated protein kinase (AMPK) as it is now 463
established that AMPK is phosphorylated in response to propionate and other SCFAs in muscle, liver and 464
intestinal tissue and AMPK is emerging as an important component in SCFA-conferred metabolic benefits 465
20, 71, 72. 466
In conclusion, we have shown that it is possible to identify synthetic glycans that support high 467
propionate production through ex vivo screening and that administration of such glycans to western diet-468
fed Ldlr-/- mice results in a similar induction of high propionate producing gut microbiota in vivo. This has 469
multiple beneficial effects, in lowering circulating LDL-cholesterol, reduction in fat accumulation in the 470
liver and reduction in development of atherosclerotic lesions in multiple vessels throughout the body. 471
These data highlight the potential of propionate producing glycans such as KB39 in improving major 472
CMRFs and the prevention of cardiometabolic diseases. 473
474
ACKNOWLEDGMENTS 475
We thank the Kaleido Biosciences’ Discovery and Technical Operations teams for their hard work and 476
expertise and Peter Turnbaugh (UCSF) for critical reading of the manuscript. 477
478
SOURCES OF FUNDING 479
This study was funded by Kaleido Biosciences. 480
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
21
DISCLOSURES 481
C. Ronald Kahn is on the Kaleido Biosciences Scientific Advisory Board. All other authors either are or were 482
employees of Kaleido Biosciences. 483
484
SUPPLEMENTAL MATERIAL 485
Figures S1-S7 486
487
REFERENCES 488
1. Shai I, Schwarzfuchs D, Henkin Y, Shahar DR, Witkow S, Greenberg I, Golan R, Fraser D, Bolotin A, 489
Vardi H, Tangi-Rozental O, Zuk-Ramot R, Sarusi B, Brickner D, Schwartz Z, Sheiner E, Marko R, Katorza E, 490
Thiery J, Fiedler GM, Bluher M, Stumvoll M, Stampfer MJ and Dietary Intervention Randomized 491
Controlled Trial G. Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet. N Engl J Med. 492
2008;359:229-41. 493
2. Estruch R, Ros E, Salas-Salvado J, Covas MI, Corella D, Aros F, Gomez-Gracia E, Ruiz-Gutierrez V, 494
Fiol M, Lapetra J, Lamuela-Raventos RM, Serra-Majem L, Pinto X, Basora J, Munoz MA, Sorli JV, Martinez 495
JA, Martinez-Gonzalez MA and Investigators PS. Primary prevention of cardiovascular disease with a 496
Mediterranean diet. N Engl J Med. 2013;368:1279-90. 497
3. Lasa A, Miranda J, Bullo M, Casas R, Salas-Salvado J, Larretxi I, Estruch R, Ruiz-Gutierrez V and 498
Portillo MP. Comparative effect of two Mediterranean diets versus a low-fat diet on glycaemic control in 499
individuals with type 2 diabetes. Eur J Clin Nutr. 2014;68:767-72. 500
4. Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P, Keinanen-501
Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, Salminen V, Uusitupa M and Finnish Diabetes 502
Prevention Study G. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with 503
impaired glucose tolerance. N Engl J Med. 2001;344:1343-50. 504
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
22
5. Paffenbarger RS, Jr., Hyde RT, Wing AL, Lee IM, Jung DL and Kampert JB. The association of 505
changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J 506
Med. 1993;328:538-45. 507
6. Bravata DM, Smith-Spangler C, Sundaram V, Gienger AL, Lin N, Lewis R, Stave CD, Olkin I and 508
Sirard JR. Using pedometers to increase physical activity and improve health: a systematic review. JAMA. 509
2007;298:2296-304. 510
7. Critchley JA and Capewell S. Mortality risk reduction associated with smoking cessation in 511
patients with coronary heart disease: a systematic review. JAMA. 2003;290:86-97. 512
8. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, Himmelfarb CD, 513
Khera A, Lloyd-Jones D, McEvoy JW, Michos ED, Miedema MD, Munoz D, Smith SC, Jr., Virani SS, 514
Williams KA, Sr., Yeboah J and Ziaeian B. 2019 ACC/AHA Guideline on the Primary Prevention of 515
Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association 516
Task Force on Clinical Practice Guidelines. J Am Coll Cardiol. 2019;74:e177-e232. 517
9. Chiuve SE, McCullough ML, Sacks FM and Rimm EB. Healthy lifestyle factors in the primary 518
prevention of coronary heart disease among men: benefits among users and nonusers of lipid-lowering 519
and antihypertensive medications. Circulation. 2006;114:160-7. 520
10. Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, Himmelfarb CD, 521
Khera A, Lloyd-Jones D, McEvoy JW, Michos ED, Miedema MD, Munoz D, Smith SC, Jr., Virani SS, 522
Williams KA, Sr., Yeboah J and Ziaeian B. 2019 ACC/AHA Guideline on the Primary Prevention of 523
Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association 524
Task Force on Clinical Practice Guidelines. Circulation. 2019;140:e596-e646. 525
11. Eckel RH, Kahn R, Robertson RM and Rizza RA. Preventing cardiovascular disease and diabetes: a 526
call to action from the American Diabetes Association and the American Heart Association. Diabetes 527
Care. 2006;29:1697-9. 528
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
23
12. Grundy SM. Drug therapy of the metabolic syndrome: minimizing the emerging crisis in 529
polypharmacy. Nat Rev Drug Discov. 2006;5:295-309. 530
13. Kris-Etherton PM, Petersen KS, Velarde G, Barnard ND, Miller M, Ros E, O'Keefe JH, Williams K, 531
Sr., Horn LV, Na M, Shay C, Douglass P, Katz DL and Freeman AM. Barriers, Opportunities, and 532
Challenges in Addressing Disparities in Diet-Related Cardiovascular Disease in the United States. J Am 533
Heart Assoc. 2020;9:e014433. 534
14. Mozaffarian D, Wilson PW and Kannel WB. Beyond established and novel risk factors: lifestyle 535
risk factors for cardiovascular disease. Circulation. 2008;117:3031-8. 536
15. Drake I, Sonestedt E, Ericson U, Wallstrom P and Orho-Melander M. A Western dietary pattern is 537
prospectively associated with cardio-metabolic traits and incidence of the metabolic syndrome. Br J 538
Nutr. 2018;119:1168-1176. 539
16. Aron-Wisnewsky J and Clement K. The gut microbiome, diet, and links to cardiometabolic and 540
chronic disorders. Nat Rev Nephrol. 2016;12:169-81. 541
17. Brown K, DeCoffe D, Molcan E and Gibson DL. Diet-induced dysbiosis of the intestinal microbiota 542
and the effects on immunity and disease. Nutrients. 2012;4:1095-119. 543
18. Fan Y and Pedersen O. Gut microbiota in human metabolic health and disease. Nat Rev 544
Microbiol. 2021;19:55-71. 545
19. Attaye I, Pinto-Sietsma SJ, Herrema H and Nieuwdorp M. A Crucial Role for Diet in the 546
Relationship Between Gut Microbiota and Cardiometabolic Disease. Annu Rev Med. 2020;71:149-161. 547
20. Canfora EE, Jocken JW and Blaak EE. Short-chain fatty acids in control of body weight and insulin 548
sensitivity. Nat Rev Endocrinol. 2015;11:577-91. 549
21. Canfora EE, Meex RCR, Venema K and Blaak EE. Gut microbial metabolites in obesity, NAFLD and 550
T2DM. Nat Rev Endocrinol. 2019;15:261-273. 551
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
24
22. Villette R, Kc P, Beliard S, Salas Tapia MF, Rainteau D, Guerin M and Lesnik P. Unraveling Host-552
Gut Microbiota Dialogue and Its Impact on Cholesterol Levels. Front Pharmacol. 2020;11:278. 553
23. Bartolomaeus H, Balogh A, Yakoub M, Homann S, Marko L, Hoges S, Tsvetkov D, Krannich A, 554
Wundersitz S, Avery EG, Haase N, Kraker K, Hering L, Maase M, Kusche-Vihrog K, Grandoch M, Fielitz J, 555
Kempa S, Gollasch M, Zhumadilov Z, Kozhakhmetov S, Kushugulova A, Eckardt KU, Dechend R, Rump LC, 556
Forslund SK, Muller DN, Stegbauer J and Wilck N. Short-Chain Fatty Acid Propionate Protects From 557
Hypertensive Cardiovascular Damage. Circulation. 2019;139:1407-1421. 558
24. Haghikia A, Zimmermann F, Schumann P, Jasina A, Roessler J, Schmidt D, Heinze P, Kaisler J, 559
Nageswaran V, Aigner A, Ceglarek U, Cineus R, Hegazy AN, van der Vorst EPC, Doring Y, Strauch CM, 560
Nemet I, Tremaroli V, Dwibedi C, Krankel N, Leistner DM, Heimesaat MM, Bereswill S, Rauch G, Seeland 561
U, Soehnlein O, Muller DN, Gold R, Backhed F, Hazen SL, Haghikia A and Landmesser U. Propionate 562
attenuates atherosclerosis by immune-dependent regulation of intestinal cholesterol metabolism. Eur 563
Heart J. 2022;43:518-533. 564
25. Deehan EC, Yang C, Perez-Munoz ME, Nguyen NK, Cheng CC, Triador L, Zhang Z, Bakal JA and 565
Walter J. Precision Microbiome Modulation with Discrete Dietary Fiber Structures Directs Short-Chain 566
Fatty Acid Production. Cell Host Microbe. 2020;27:389-404 e6. 567
26. Le Bastard Q, Chapelet G, Javaudin F, Lepelletier D, Batard E and Montassier E. The effects of 568
inulin on gut microbial composition: a systematic review of evidence from human studies. Eur J Clin 569
Microbiol Infect Dis. 2020;39:403-413. 570
27. Reider SJ, Moosmang S, Tragust J, Trgovec-Greif L, Tragust S, Perschy L, Przysiecki N, Sturm S, 571
Tilg H, Stuppner H, Rattei T and Moschen AR. Prebiotic Effects of Partially Hydrolyzed Guar Gum on the 572
Composition and Function of the Human Microbiota-Results from the PAGODA Trial. Nutrients. 2020;12. 573
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
25
28. Jalanka J, Major G, Murray K, Singh G, Nowak A, Kurtz C, Silos-Santiago I, Johnston JM, de Vos 574
WM and Spiller R. The Effect of Psyllium Husk on Intestinal Microbiota in Constipated Patients and 575
Healthy Controls. Int J Mol Sci. 2019;20. 576
29. Liu F, Li P, Chen M, Luo Y, Prabhakar M, Zheng H, He Y, Qi Q, Long H, Zhang Y, Sheng H and Zhou 577
H. Fructooligosaccharide (FOS) and Galactooligosaccharide (GOS) Increase Bifidobacterium but Reduce 578
Butyrate Producing Bacteria with Adverse Glycemic Metabolism in healthy young population. Sci Rep. 579
2017;7:11789. 580
30. Canfora EE, van der Beek CM, Hermes GDA, Goossens GH, Jocken JWE, Holst JJ, van Eijk HM, 581
Venema K, Smidt H, Zoetendal EG, Dejong CHC, Lenaerts K and Blaak EE. Supplementation of Diet With 582
Galacto-oligosaccharides Increases Bifidobacteria, but Not Insulin Sensitivity, in Obese Prediabetic 583
Individuals. Gastroenterology. 2017;153:87-97 e3. 584
31. Tolonen AC, Beauchemin N, Bayne C, Li L, Tan J, Lee J, Meehan BM, Meisner J, Millet Y, LeBlanc 585
G, Kottler R, Rapp E, Murphy C, Turnbaugh PJ, von Maltzahn G, Liu CM and van Hylckama Vlieg JET. 586
Synthetic glycans control gut microbiome structure and mitigate colitis in mice. Nat Commun. 587
2022;13:1244. 588
32. Theriot CM, Koenigsknecht MJ, Carlson PE, Jr., Hatton GE, Nelson AM, Li B, Huffnagle GB, J ZL 589
and Young VB. Antibiotic-induced shifts in the mouse gut microbiome and metabolome increase 590
susceptibility to Clostridium difficile infection. Nat Commun. 2014;5:3114. 591
33. Hakomori S. A Rapid Permethylation of Glycolipid, and Polysaccharide Catalyzed by 592
Methylsulfinyl Carbanion in Dimethyl Sulfoxide. J Biochem. 1964;55:205-8. 593
34. Hillmann B, Al-Ghalith GA, Shields-Cutler RR, Zhu Q, Gohl DM, Beckman KB, Knight R and Knights 594
D. Evaluating the Information Content of Shallow Shotgun Metagenomics. mSystems. 2018;3. 595
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
26
35. Franzosa EA, McIver LJ, Rahnavard G, Thompson LR, Schirmer M, Weingart G, Lipson KS, Knight 596
R, Caporaso JG, Segata N and Huttenhower C. Species-level functional profiling of metagenomes and 597
metatranscriptomes. Nat Methods. 2018;15:962-968. 598
36. Reichardt N, Duncan SH, Young P, Belenguer A, McWilliam Leitch C, Scott KP, Flint HJ and Louis 599
P. Phylogenetic distribution of three pathways for propionate production within the human gut 600
microbiota. ISME J. 2014;8:1323-35. 601
37. Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, Ferrell LD, Liu YC, 602
Torbenson MS, Unalp-Arida A, Yeh M, McCullough AJ, Sanyal AJ and Nonalcoholic Steatohepatitis 603
Clinical Research N. Design and validation of a histological scoring system for nonalcoholic fatty liver 604
disease. Hepatology. 2005;41:1313-21. 605
38. Love MI, Huber W and Anders S. Moderated estimation of fold change and dispersion for RNA-606
seq data with DESeq2. Genome Biol. 2014;15:550. 607
39. Amir I, Bouvet P, Legeay C, Gophna U and Weinberger A. Eisenbergiella tayi gen. nov., sp. nov., 608
isolated from human blood. Int J Syst Evol Microbiol. 2014;64:907-914. 609
40. Willems A, Moore WE, Weiss N and Collins MD. Phenotypic and phylogenetic characterization of 610
some Eubacterium-like isolates containing a novel type B wall murein from human feces: description of 611
Holdemania filiformis gen. nov., sp. nov. Int J Syst Bacteriol. 1997;47:1201-4. 612
41. Bosshard PP, Zbinden R and Altwegg M. Turicibacter sanguinis gen. nov., sp. nov., a novel 613
anaerobic, Gram-positive bacterium. Int J Syst Evol Microbiol. 2002;52:1263-1266. 614
42. Liu C, Finegold SM, Song Y and Lawson PA. Reclassification of Clostridium coccoides, 615
Ruminococcus hansenii, Ruminococcus hydrogenotrophicus, Ruminococcus luti, Ruminococcus 616
productus and Ruminococcus schinkii as Blautia coccoides gen. nov., comb. nov., Blautia hansenii comb. 617
nov., Blautia hydrogenotrophica comb. nov., Blautia luti comb. nov., Blautia producta comb. nov., 618
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
27
Blautia schinkii comb. nov. and description of Blautia wexlerae sp. nov., isolated from human faeces. Int 619
J Syst Evol Microbiol. 2008;58:1896-902. 620
43. Wouters K, van Gorp PJ, Bieghs V, Gijbels MJ, Duimel H, Lutjohann D, Kerksiek A, van Kruchten R, 621
Maeda N, Staels B, van Bilsen M, Shiri-Sverdlov R and Hofker MH. Dietary cholesterol, rather than liver 622
steatosis, leads to hepatic inflammation in hyperlipidemic mouse models of nonalcoholic 623
steatohepatitis. Hepatology. 2008;48:474-86. 624
44. Bieghs V, Van Gorp PJ, Wouters K, Hendrikx T, Gijbels MJ, van Bilsen M, Bakker J, Binder CJ, 625
Lutjohann D, Staels B, Hofker MH and Shiri-Sverdlov R. LDL receptor knock-out mice are a physiological 626
model particularly vulnerable to study the onset of inflammation in non-alcoholic fatty liver disease. 627
PLoS One. 2012;7:e30668. 628
45. Wu L, Vikramadithyan R, Yu S, Pau C, Hu Y, Goldberg IJ and Dansky HM. Addition of dietary fat to 629
cholesterol in the diets of LDL receptor knockout mice: effects on plasma insulin, lipoproteins, and 630
atherosclerosis. J Lipid Res. 2006;47:2215-22. 631
46. Jeong S, Kim M, Han M, Lee H, Ahn J, Kim M, Song YH, Shin C, Nam KH, Kim TW, Oh GT and Yoon 632
M. Fenofibrate prevents obesity and hypertriglyceridemia in low-density lipoprotein receptor-null mice. 633
Metabolism. 2004;53:607-13. 634
47. Srivastava RA, Jahagirdar R, Azhar S, Sharma S and Bisgaier CL. Peroxisome proliferator-activated 635
receptor-alpha selective ligand reduces adiposity, improves insulin sensitivity and inhibits 636
atherosclerosis in LDL receptor-deficient mice. Mol Cell Biochem. 2006;285:35-50. 637
48. Chen M and Wolin MJ. Influence of heme and vitamin B12 on growth and fermentations of 638
Bacteroides species. J Bacteriol. 1981;145:466-71. 639
49. Everard A, Belzer C, Geurts L, Ouwerkerk JP, Druart C, Bindels LB, Guiot Y, Derrien M, Muccioli 640
GG, Delzenne NM, de Vos WM and Cani PD. Cross-talk between Akkermansia muciniphila and intestinal 641
epithelium controls diet-induced obesity. Proc Natl Acad Sci U S A. 2013;110:9066-71. 642
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
28
50. Bush JR and Alfa MJ. Increasing levels of Parasutterella in the gut microbiome correlate with 643
improving low-density lipoprotein levels in healthy adults consuming resistant potato starch during a 644
randomised trial. BMC Nutr. 2020;6:72. 645
51. Bentley P, Calder I, Elcombe C, Grasso P, Stringer D and Wiegand HJ. Hepatic peroxisome 646
proliferation in rodents and its significance for humans. Food Chem Toxicol. 1993;31:857-907. 647
52. Wang YX. PPARs: diverse regulators in energy metabolism and metabolic diseases. Cell Res. 648
2010;20:124-37. 649
53. Shimano H and Sato R. SREBP-regulated lipid metabolism: convergent physiology - divergent 650
pathophysiology. Nat Rev Endocrinol. 2017;13:710-730. 651
54. Moran-Salvador E, Lopez-Parra M, Garcia-Alonso V, Titos E, Martinez-Clemente M, Gonzalez-652
Periz A, Lopez-Vicario C, Barak Y, Arroyo V and Claria J. Role for PPARgamma in obesity-induced hepatic 653
steatosis as determined by hepatocyte- and macrophage-specific conditional knockouts. FASEB J. 654
2011;25:2538-50. 655
55. Lee YK, Park JE, Lee M and Hardwick JP. Hepatic lipid homeostasis by peroxisome proliferator-656
activated receptor gamma 2. Liver Res. 2018;2:209-215. 657
56. Horton JD, Goldstein JL and Brown MS. SREBPs: activators of the complete program of 658
cholesterol and fatty acid synthesis in the liver. J Clin Invest. 2002;109:1125-31. 659
57. Pawlak M, Lefebvre P and Staels B. Molecular mechanism of PPARalpha action and its impact on 660
lipid metabolism, inflammation and fibrosis in non-alcoholic fatty liver disease. J Hepatol. 2015;62:720-661
33. 662
58. Luo J, Yang H and Song BL. Mechanisms and regulation of cholesterol homeostasis. Nat Rev Mol 663
Cell Biol. 2020;21:225-245. 664
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
29
59. Schonewille M, de Boer JF, Mele L, Wolters H, Bloks VW, Wolters JC, Kuivenhoven JA, Tietge UJ, 665
Brufau G and Groen AK. Statins increase hepatic cholesterol synthesis and stimulate fecal cholesterol 666
elimination in mice. J Lipid Res. 2016;57:1455-64. 667
60. Taylor BA and Thompson PD. Statins and Their Effect on PCSK9-Impact and Clinical Relevance. 668
Curr Atheroscler Rep. 2016;18:46. 669
61. Du K, Herzig S, Kulkarni RN and Montminy M. TRB3: a tribbles homolog that inhibits Akt/PKB 670
activation by insulin in liver. Science. 2003;300:1574-7. 671
62. Emanuelli T, Pagel FW, Alves LB, Regner A and Souza DO. Inhibition of adenylate cyclase activity 672
by 5-aminolevulinic acid in rat and human brain. Neurochem Int. 2001;38:213-8. 673
63. Koo SH, Satoh H, Herzig S, Lee CH, Hedrick S, Kulkarni R, Evans RM, Olefsky J and Montminy M. 674
PGC-1 promotes insulin resistance in liver through PPAR-alpha-dependent induction of TRB-3. Nat Med. 675
2004;10:530-4. 676
64. Heinrich G, Ghadieh HE, Ghanem SS, Muturi HT, Rezaei K, Al-Share QY, Bowman TA, Zhang D, 677
Garofalo RS, Yin L and Najjar SM. Loss of Hepatic CEACAM1: A Unifying Mechanism Linking Insulin 678
Resistance to Obesity and Non-Alcoholic Fatty Liver Disease. Front Endocrinol (Lausanne). 2017;8:8. 679
65. Ueno T and Komatsu M. Autophagy in the liver: functions in health and disease. Nat Rev 680
Gastroenterol Hepatol. 2017;14:170-184. 681
66. Ormazabal V, Nair S, Elfeky O, Aguayo C, Salomon C and Zuniga FA. Association between insulin 682
resistance and the development of cardiovascular disease. Cardiovasc Diabetol. 2018;17:122. 683
67. Fernandez-Veledo S and Vendrell J. Gut microbiota-derived succinate: Friend or foe in human 684
metabolic diseases? Rev Endocr Metab Disord. 2019;20:439-447. 685
68. De Vadder F, Kovatcheva-Datchary P, Zitoun C, Duchampt A, Backhed F and Mithieux G. 686
Microbiota-Produced Succinate Improves Glucose Homeostasis via Intestinal Gluconeogenesis. Cell 687
Metab. 2016;24:151-7. 688
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
30
69. Plovier H, Everard A, Druart C, Depommier C, Van Hul M, Geurts L, Chilloux J, Ottman N, Duparc 689
T, Lichtenstein L, Myridakis A, Delzenne NM, Klievink J, Bhattacharjee A, van der Ark KC, Aalvink S, 690
Martinez LO, Dumas ME, Maiter D, Loumaye A, Hermans MP, Thissen JP, Belzer C, de Vos WM and Cani 691
PD. A purified membrane protein from Akkermansia muciniphila or the pasteurized bacterium improves 692
metabolism in obese and diabetic mice. Nat Med. 2017;23:107-113. 693
70. Ottman N, Reunanen J, Meijerink M, Pietila TE, Kainulainen V, Klievink J, Huuskonen L, Aalvink S, 694
Skurnik M, Boeren S, Satokari R, Mercenier A, Palva A, Smidt H, de Vos WM and Belzer C. Pili-like 695
proteins of Akkermansia muciniphila modulate host immune responses and gut barrier function. PLoS 696
One. 2017;12:e0173004. 697
71. den Besten G, Bleeker A, Gerding A, van Eunen K, Havinga R, van Dijk TH, Oosterveer MH, Jonker 698
JW, Groen AK, Reijngoud DJ and Bakker BM. Short-Chain Fatty Acids Protect Against High-Fat Diet-699
Induced Obesity via a PPARgamma-Dependent Switch From Lipogenesis to Fat Oxidation. Diabetes. 700
2015;64:2398-408. 701
72. Yoshida H, Ishii M and Akagawa M. Propionate suppresses hepatic gluconeogenesis via 702
GPR43/AMPK signaling pathway. Arch Biochem Biophys. 2019;672:108057. 703
704
FIGURE LEGENDS 705
Figure 1. Identification of high propionate-producing synthetic glycan KB39. (A) High-throughput ex vivo 706
screen of synthetic glycan library with fecal microbiota from a healthy subject. Fecal microbiota cultures 707
were incubated with synthetic glycans under anaerobic conditions and propionate was measured in 708
culture supernatants by gas chromatography with flame ionization detection (GC-FID). Data represent the 709
distribution of propionate production. The blue and red bars indicate where the no added carbon control 710
and KB39 are on the distribution, respectively. The curve represents the normal distribution and the 711
dotted lines indicate the 95% confidence interval. (B, C) Production of each SCFA by KB39 with fecal 712
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
31
microbiota from 10 healthy subjects (B) and 31 overweight-T2DM patients (C). Fecal microbiota cultures 713
were incubated without (negative control) or with KB39 and SCFA were measured in culture supernatants. 714
* P<0.05, *** P<0.001**** P<0.0001, paired t-test. 715
716
Figure 2. KB39 enriches propionate-producing commensal bacteria ex vivo in overweight-T2DM 717
subjects. Fecal microbiota cultures from 31 overweight-T2DM subjects were incubated ex vivo without 718
(negative control) or with KB39. Significantly enriched and depleted taxa (fdr < 0.1, Wilcox rank sum test 719
with FDR correction) at phylum, family, genus, and species level in KB39 treated group in comparison to 720
water. Data represent the log2 fold change. 721
722
Figure 3. KB39 increases propionate, acetate and succinate production and increases the proportion of 723
cecal propionate and succinate in the cecum of western diet-fed Ldlr-/- mice. Male Ldlr-/- mice were fed 724
a western diet, western diet supplemented with KB39 (7.5% w/w) or fenofibrate (100 mg/kg/day) for 16 725
weeks. (A) Total cecal acetate, butyrate, propionate and succinate. (B) Acetate, butyrate, propionate and 726
succinate relative fractions. NT: no treatment; Feno: fenofibrate. *** P<0.001, **** P<0.0001, 1-way 727
ANOVA with Dunnett pairwise comparison to NT group. 728
729
Figure 4. KB39 shifts the gut microbiome in western diet-fed Ldlr-/- mice and promotes propionate-730
producing bacterial taxa. Male Ldlr-/- mice were fed a western diet, western diet supplemented with KB39 731
(7.5% w/w) or fenofibrate (100 mg/kg/day) for 16 weeks. Fecal samples were collected before treatment 732
during the last week of the treatment period. Significantly enriched and depleted taxa (fdr < 0.1, Wilcox 733
rank sum test with FDR correction) at phylum, family, genus, and species levels in KB39 treated group in 734
comparison to no treatment group at the end of the study. 735
736
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
32
Figure 5. KB39 attenuates liver steatosis, liver inflammation and atherosclerosis in western diet-fed Ldlr-
737
/- mice. Male Ldlr-/- mice were fed a western diet, western diet supplemented with KB39 (7.5% w/w) or 738
fenofibrate (100mg/kg/day) for 16 weeks. An OGTT was performed at week 15. (A-C) Relative body weight 739
(BW) compared to BW at treatment initiation (A), daily food intake (B) and OGTT blood glucose 740
measurements (C); mean ± SEM, 2-way ANOVA, Dunnett. (D-F) Liver steatosis score (D), inflammation (E) 741
score and NAFLD activity score (NAS) (F) at termination; median, Kruskal-Wallis, Dunn. (G) Aortic sinus 742
plaque severity score; median; Kruskal-Wallis, Dunn. (H, I) Plaque area in the aortic sinus quantified by 743
morphometric analysis of elastin trichrome-stained sections (H) and in the aortic arch quantified by oil 744
red O staining “en face” analysis (I), 1-way ANOVA, Dunnett. * P<0.05, ** P<0.01, *** P<0.001, **** 745
P<0.0001. NT: no treatment; Feno: fenofibrate.
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
33
TABLES 761
Table 1. Effects of KB39 and fenofibrate in western diet-fed Ldlr-/- mice. Male Ldlr-/- mice were fed a 762
western diet, western diet supplemented with KB39 (7.5% w/w) or western diet supplemented with 763
fenofibrate (100 mg/kg/day) for 16 weeks. An OGTT was performed at week 15. Measurements were 764
taken in fasted animals; see material and methods for details. Data represent the mean ± SEM; statistics: 765
1-way ANOVA with Dunnett pairwise comparison to no treatment group. 766
Mean ± SEM
P-value
No Treatment
KB39
Fenofibrate
KB39
Fenofibrate
Serum total cholesterol (mg/dL)
1924.3±140.4
1094.7±132.4
1564.3±59.6
<0.0001
0.0660
Serum VLDL (mg/dL)
449.9±63.0
190.9±39.6
344.4±46.6
0.0019
0.2457
Serum LDL (mg/dL)
949.5±76.8
581.1±66.1
728.4±36.2
0.0004
0.0298
Serum HDL (mg/dL)
601.5±73.4
321.3±47.4
484.9±36.9
0.0016
0.2230
Blood triglycerides (mg/dL)
354.9±28.3
293.8±24.1
317.3±29.1
0.2105
0.5243
Serum NEFA (mM)
1.11±0.07
0.93±0.05
0.65±0.05
0.0683
<0.0001
Hepatic total cholesterol (mg/g)
12.61±0.84
7.66±0.48
8.50±0.31
<0.0001
<0.0001
Hepatic triglycerides (mg/g)
156.8±13.5
98.6±10.2
42.4±5.4
0.0006
<0.0001
Blood glucose (mg/dL)
142.1±5.7
139.4±6.8
113.1±5.1
0.9284
0.0028
Blood glucose AUC (g/dL.min) (OGTT)
10.9±3.5
12.0±4.1
8.1±2.0
0.5909
0.0882
Serum insulin (ng/mL)
1.01±0.18
0.35±0.08
0.33±0.10
0.0022
0.0015
Final body weight (g)
39.3±1.4
38.5±1.4
29.2±0.6
0.8285
<0.0001
Body weight change (%)
58.2±3.8
52.5±5.3
14.0±1.2
0.4800
<0.0001
Liver weight (mg)
1690±112.7
1338±81.1
2442±81.8
0.0218
<0.0001
Relative liver weight (% body weight)
4.27±0.20
3.46±0.10
8.37±0.27
0.0128
<0.0001
767
768
769
770
771
Table 2. Genes differentially expressed in the liver of western diet-fed Ldlr-/- mice treated with KB39 or 772
fenofibrate. Male Ldlr-/- mice were fed a western diet, western diet supplemented with KB39 (7.5% w/w) 773
or fenofibrate (100 mg/kg/day) for 16 weeks. Livers were collected at termination and hepatic gene 774
expression was analyzed by RNAseq. Deseq2 38 was used to determine the log2FC and adjusted P-value 775
(padj) in gene expression compared to the no treatment group. 776
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
34
KB39
Fenofibrate
Gene
log2FC
padj
log2FC
padj
description
Fatty acid oxidation
Cpt1a
0.43
1.45e-03
0.05
6.84e-01
carnitine palmitoyltransferase 1a
Cpt2
0.37
6.55e-05
0.32
1.89e-04
carnitine palmitoyltransferase 2
Hadha
0.54
3.68e-10
0.91
3.01e-39
hydroxyacyl-CoA dehydrogenase trifunctional subunit alpha
Acads
0.21
9.44E-03
0.53
2.43E-14
acyl-Coenzyme A dehydrogenase, short chain
Acadm
0.22
1.02E-03
0.60
7.13E-17
acyl-Coenzyme A dehydrogenase, medium chain
Acadl
0.22
1.04E-03
0.87
2.17E-31
acyl-Coenzyme A dehydrogenase, long-chain
Hadha
0.54
3.68e-10
0.91
3.01e-39
hydroxyacyl-CoA dehydrogenase trifunctional subunit alpha
Hadhb
0.40
3.02E-07
0.92
4.39E-44
hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex subunit beta
Acaa2
0.39
1.81e-06
0.57
6.98e-13
acetyl-Coenzyme A acyltransferase 2
Abcd3
0.38
6.03e-06
0.71
1.27e-10
ATP-binding cassette, sub-family D (ALD), member 3
Acox1
0.32
1.91e-03
1.54
1.24e-49
acyl-Coenzyme A oxidase 1, palmitoyl
Acaa1a
0.32
7.77E-05
0.53
9.39E-10
acetyl-Coenzyme A acyltransferase 1A
Acaa1b
0.43
2.97e-05
1.15
4.43e-31
acetyl-Coenzyme A acyltransferase 1B
Ehhadh
0.58
6.02e-02
2.98
7.56e-105
enoyl-CoA, hydratase/3-hydroxyacyl Coenzyme A dehydrogenase
Cyp4a14
1.38
2.08e-04
2.72
1.77e-27
cytochrome P450, family 4, subfamily a, polypeptide 14
Regulators
Ppara
0.33
5.83e-03
-0.13
3.15e-01
peroxisome proliferator activated receptor alpha
Pparg
-0.60
7.49e-04
-0.63
5.73e-06
peroxisome proliferator activated receptor gamma
Srebf1
-0.43
7.04e-03
0.98
1.02e-25
sterol regulatory element binding transcription factor 1
Srebf2
0.46
3.30e-03
0.69
1.63e-17
sterol regulatory element binding factor 2
Fatty acid synthesis
Acaca
-0.05
7.62e-01
1.16
7.94e-20
acetyl-Coenzyme A carboxylase alpha
Acacb
0.37
9.00e-02
0.96
5.46e-09
acetyl-Coenzyme A carboxylase beta
Fasn
0.11
6.24e-01
1.98
7.25e-27
fatty acid synthase
Elovl6
0.11
4.42e-01
0.47
1.27e-04
ELOVL family member 6, elongation of long chain fatty acids
Scd1
-0.79
2.03e-02
0.44
3.94e-04
stearoyl-Coenzyme A desaturase 1
Bile acids
Abcb11 (Ntcp)
0.29
1.35e-03
-0.02
8.82e-01
ATP-binding cassette, sub-family B (MDR/TAP), member 11
Cyp7a1
0.69
1.61e-03
-0.13
7.00e-01
cytochrome P450, family 7, subfamily a, polypeptide 1
Cyp7b1
1.26
4.82e-03
-0.40
2.91e-01
cytochrome P450, family 7, subfamily b, polypeptide 1
Cyp8b1
0.85
4.66e-03
0.76
1.08e-03
cytochrome P450, family 8, subfamily b, polypeptide 1
Slc10a1 (Bsep)
0.38
8.06e-04
0.33
4.23e-03
solute carrier family 10 (sodium/bile acid cotransporter family), member 1
Cholesterol synthesis
Ldlr
0.25
1.35e-02
0.13
9.70e-02
low density lipoprotein receptor
Hmgcs1
0.57
1.59e-02
1.80
1.29e-22
3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1
Hmgcr
0.36
8.02e-02
0.55
2.13e-04
3-hydroxy-3-methylglutaryl-Coenzyme A reductase
Mvk
0.46
4.23e-02
0.85
3.72e-08
mevalonate kinase
Mvd
0.82
3.39e-02
1.88
4.83e-17
mevalonate (diphospho) decarboxylase
Idi1
2.22
1.59e-04
3.04
7.36e-34
isopentenyl-diphosphate delta isomerase
Fdps
0.98
2.16e-02
2.80
8.26e-68
farnesyl diphosphate synthetase
Sqle
0.64
1.93e-01
1.92
2.89e-10
squalene epoxidase
Cyp51
1.53
2.14e-04
1.84
6.71e-26
cytochrome P450, family 51
Inflammation
Ccl2 (Mcp1)
-2.33
1.20e-11
-1.71
1.87e-14
chemokine (C-C motif) ligand 2
Ccl5
-1.84
2.82e-15
-1.11
3.20e-06
chemokine (C-C motif) ligand 5
Il1a
-1.03
1.67e-08
-0.50
1.00e-03
interleukin 1 alpha
Il1b
-0.64
1.21e-02
-0.38
5.06e-02
interleukin 1 beta
Lcn2
-1.45
7.94e-04
-2.21
1.28e-08
lipocalin 2
Tnf
-2.41
1.24e-07
-1.52
3.94e-05
tumor necrosis factor
Oxidative stress
Cybb
-0.99
4.04e-10
-0.64
2.54e-06
cytochrome b-245, beta polypeptide
Hmox1
-1.22
1.15e-19
-0.33
1.68e-02
heme oxygenase 1
Ncf2
-1.10
1.07e-09
-0.86
1.76e-09
neutrophil cytosolic factor 2
Apoptosis
Bax
-0.33
1.09e-02
-0.20
6.06e-02
BCL2-associated X protein
Bcl2
-0.88
2.60e-05
-0.74
6.81e-05
B cell leukemia/lymphoma 2
Bcl2a1a
-1.25
3.03e-06
-1.10
1.35e-06
B cell leukemia/lymphoma 2 related protein A1a
Traf1
-1.43
6.37e-06
-0.98
1.60e-04
TNF receptor-associated factor 1
Fibrosis
Acta2 (α-Sma)
-1.10
3.16e-03
-0.10
8.10e-01
actin alpha 2, smooth muscle
Col1a1
-2.75
1.24e-10
-1.62
6.00e-05
collagen, type I, alpha 1
Mmp13
-3.67
3.33e-12
-2.56
4.29e-08
matrix metallopeptidase 13
Tgfb1
-0.52
4.58e-04
-0.38
6.88e-03
transforming growth factor, beta 1
Timp1
-3.87
1.33e-11
-2.87
6.58e-11
tissue inhibitor of metalloproteinase 1
Cell adhesion
Icam1
-0.77
1.25e-06
-0.64
3.66e-07
intercellular adhesion molecule 1
Vcam1
-1.45
1.41e-14
-0.77
5.79e-07
vascular cell adhesion molecule 1
Kupffer Cell/macrophage
Adgre1 (F4/80)
-0.34
7.39e-02
-0.45
8.14e-03
adhesion G protein-coupled receptor E1
Cd14
-1.18
3.70e-06
-0.48
3.79e-02
CD14 antigen
Cd68
-1.38
1.61e-13
-0.91
5.64e-08
CD68 antigen
Itgam (Cd11b)
-0.93
2.73e-03
-1.09
5.38e-05
integrin alpha M
Itgax (Cd11c)
-2.61
5.72e-13
-1.37
2.02e-09
integrin alpha X
Insulin signaling
Ceacam1
0.48
1.78e-05
0.08
4.46e-01
carcinoembryonic antigen-related cell adhesion molecule 1
Socs3
-0.48
5.18e-03
-0.94
6.73e-10
suppressor of cytokine signaling 3
Trib3
-0.96
4.72e-12
0.90
9.57e-12
tribbles pseudokinase 3
Autophagy
Hspa8 (Hsp70)
0.53
5.09e-08
-0.32
5.52e-05
heat shock protein 8
Map1lc3a (Lc3a)
0.40
6.99e-04
1.02
1.97e-31
microtubule-associated protein 1 light chain 3 alpha
Ulk1
0.41
3.77e-03
-0.06
6.18e-01
unc-51 like kinase 1
Ulk2
0.26
3.97e-03
-0.10
2.56e-01
unc-51 like kinase 2
777
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
Control KB39
0
5
10
15
20
25
[Acetate] (mM)
****
Control KB39
0
5
10
15
20
[Propionate] (mM)
****
Control KB39
0
2
4
6
[Butyrate] (mM)
*
Control KB39
0
10
20
30
40
[Acetate] (mM)
****
Control KB39
0
5
10
15
20
25
[Propionate] (mM)
****
Control KB39
0
5
10
15
[Butyrate] (mM)
***
0 5 10 15 20
0
20
40
60
80
[Propionate] (mM)
Number of compounds
A
B
C
Figure 1. Identification of high propionate-producing synthetic glycan KB39. (A) High-throughput ex vivo screen of
synthetic glycan library with fecal microbiota from a healthy subject. Fecal microbiota cultures were incubated with
synthetic glycans under anaerobic conditions and propionate was measured in culture supernatants by gas
chromatography with flame ionization detection (GC-FID). Data represent the distribution of propionate production.
The blue and red bars indicate where the no added carbon control and KB39 are on the distribution, respectively.
The curve represents the normal distribution and the dotted lines indicate the 95% confidence interval. (B, C)
Production of each SCFA by KB39 with fecal microbiota from 10 healthy subjects (B) and 31 overweight-T2DM
patients (C). Fecal microbiota cultures were incubated without (negative control) or with KB39 and SCFA were
measured in culture supernatants. * P<0.05, *** P<0.001**** P<0.0001, paired t-test.
Figure 1
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8
g_Klebsiella__s_Klebsiella_pneumoniae
f_Peptoniphilaceae__g_Peptoniphilus
g_Citrobacter__s_Citrobacter_freundii
g_Emergencia__s_Emergencia_timonensis
o_Tissierellales__f_Peptoniphilaceae
g_Enterococcus__s_Enterococcus_faecalis
g_Klebsiella__s_Klebsiella_oxytoca
g_Citrobacter__s_Citrobacter_braakii
g_Eggerthella__s_Eggerthella_lenta
g_Bariatricus__s_Bariatricus_massiliensis
f_Clostridiales_Family_XIII._Incertae_Sedis__g_Emergencia
g_Sutterella__s_Sutterella_wadsworthensis
f_Enterobacteriaceae__g_Klebsiella
f_Eggerthellaceae__g_Eggerthella
g_Roseburia__s_Roseburia_intestinalis
g_Coprococcus__s_Coprococcus_comes
o_Eggerthellales__f_Eggerthellaceae
f_Enterobacteriaceae__g_Citrobacter
g_Escherichia__s_Escherichia_coli
g_Flavonifractor__s_Flavonifractor_plautii
g_Lachnoclostridium__s_[Clostridium]_hylemonae
g_Alistipes__s_Alistipes_shahii
d_Bacteria__p_Actinobacteria
d_Bacteria__p_Proteobacteria
o_Fusobacteriales__f_Fusobacteriaceae
g_Proteus__s_Proteus_mirabilis
f_Fusobacteriaceae__g_Fusobacterium
f_Enterobacteriaceae__g_Escherichia
f_Bifidobacteriaceae__g_Bifidobacterium
g_Enterococcus__s_Enterococcus_hirae
f_Lachnospiraceae__g_Lachnoclostridium
f_Rikenellaceae__g_Alistipes
o_Bacteroidales__f_Rikenellaceae
g_Bacteroides__s_Bacteroides_cellulosilyticus
f_Lachnospiraceae__g_Roseburia
o_Erysipelotrichales__f_Erysipelotrichaceae
g_Bacteroides__s_Bacteroides_thetaiotaomicron
g_Blautia__s_Blautia_producta
g_Bacteroides__s_Bacteroides_fragilis
d_Bacteria__p_Bacteroidetes
g_Eisenbergiella__s_Eisenbergiella_tayi
f_Lachnospiraceae__g_Eisenbergiella
g_Parabacteroides__s_Parabacteroides_distasonis
f_Tannerellaceae__g_Parabacteroides
o_Bacteroidales__f_Tannerellaceae
g_Parabacteroides__s_Parabacteroides_goldsteinii
g_Parabacteroides__s_Parabacteroides_johnsonii
g_Parabacteroides__s_Parabacteroides_timonensis
g_Parabacteroides__s_Parabacteroides_merdae
log2FC
Figure 2. KB39 enriches propionate-producing commensal bacteria ex vivo in overweight-T2DM subjects. Fecal
microbiota cultures from 31 overweight-T2DM subjects were incubated ex vivo without (negative control) or with
KB39. Significantly enriched and depleted taxa (fdr < 0.1, Wilcox rank sum test with FDR correction) at phylum,
family, genus, and species level in KB39 treated group in comparison to water. Data represent the log2 fold change.
Figure 2
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
NT KB39 Feno.
0
5
10
15
20
Acetate (μmol)
ns
✱✱✱
NT KB39 Feno.
0
2
4
6
Propionate (μmol)
ns
✱✱✱✱
NT KB39 Feno.
0.0
0.5
1.0
1.5
Butyrate (μmol)
ns
ns
NT KB39 Feno.
0.0
0.5
1.0
1.5
2.0
Succinate (μmol)
ns
✱✱✱✱
NT KB39 Feno.
0.5
0.6
0.7
0.8
0.9
Acetate fraction
ns
✱✱✱
NT KB39 Feno.
0.0
0.1
0.2
0.3
0.4
Propionate fraction
ns
✱✱✱✱
NT KB39 Feno.
0.00
0.05
0.10
0.15
Butyrate fraction
ns
✱✱✱
NT KB39 Feno.
0.00
0.05
0.10
0.15
Succinate fraction
ns
✱✱✱✱
A
B
Figure 3. KB39 increases propionate, acetate and succinate production and increases the proportion of cecal
propionate and succinate in the cecum of western diet-fed Ldlr-/- mice. Male Ldlr-/- mice were fed a western
diet, western diet supplemented with KB39 (7.5% w/w) or fenofibrate (100 mg/kg/day) for 16 weeks. (A) Total
cecal acetate, butyrate, propionate and succinate. (B) Acetate, butyrate, propionate and succinate relative
fractions. NT: no treatment; Feno: fenofibrate. *** P<0.001, **** P<0.0001, 1-way ANOVA with Dunnett
pairwise comparison to NT group.
Figure 3
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7
o_Clostridiales__f_Peptostreptococcaceae
f_Lachnospiraceae__g_Dorea
f_Lachnospiraceae__g_Roseburia
g_Eubacterium__s_Eubacterium_plexicaudatum
f_Chlamydiaceae__g_Chlamydia
o_Clostridiales__f_Lachnospiraceae
g_Chlamydia__s_Chlamydia_abortus
f_Oscillospiraceae__g_Oscillibacter
g_Acetatifactor__s_Acetatifactor_muris
f_Lachnospiraceae__g_Acetatifactor
d_Chlamydiae__p_Chlamydiae
o_Chlamydiales__f_Chlamydiaceae
g_Enterococcus__s_Enterococcus_faecalis
o_Clostridiales__f_Oscillospiraceae
f_Enterococcaceae__g_Enterococcus
f_Ruminococcaceae__g_Anaerotruncus
f_Staphylococcaceae__g_Staphylococcus
g_Staphylococcus__s_Staphylococcus_xylosus
f_Eggerthellaceae__g_Adlercreutzia
g_Adlercreutzia__s_Adlercreutzia_equolifaciens
f_Ruminococcaceae__g_Acutalibacter
f_Streptococcaceae__g_Lactococcus
g_Lactococcus__s_Lactococcus_lactis
o_Bacillales__f_Staphylococcaceae
o_Lactobacillales__f_Enterococcaceae
g_Ruminococcus__s_Ruminococcus_torques
o_Bacteroidales__f_Muribaculaceae
o_Lactobacillales__f_Streptococcaceae
o_Eggerthellales__f_Eggerthellaceae
o_Clostridiales__f_Ruminococcaceae
g_Erysipelatoclostridium__s_Clostridium_cocleatum
f_Erysipelotrichaceae__g_Faecalibaculum
g_Faecalibaculum__s_Faecalibaculum_rodentium
o_Bacteroidales__f_Porphyromonadaceae
d_Firmicutes__p_Firmicutes
f_Ruminococcaceae__g_Ruminococcus
g_Akkermansia__s_Akkermansia_muciniphila
f_Akkermansiaceae__g_Akkermansia
g_Parasutterella__s_Parasutterella_excrementihominis
f_Sutterellaceae__g_Parasutterella
d_Proteobacteria__p_Proteobacteria
o_Burkholderiales__f_Sutterellaceae
d_Verrucomicrobia__p_Verrucomicrobia
d_Bacteroidetes__p_Bacteroidetes
o_Verrucomicrobiales__f_Akkermansiaceae
f_Tannerellaceae__g_Parabacteroides
f_Lachnospiraceae__g_Blautia
o_Bacteroidales__f_Tannerellaceae
f_Eubacteriaceae__g_Eubacterium
g_Clostridium__s_Clostridium_hathewayi
g_Alistipes__s_Alistipes_finegoldii
f_Rikenellaceae__g_Alistipes
f_Erysipelotrichaceae__g_Erysipelatoclostridium
o_Clostridiales__f_Eubacteriaceae
o_Bacteroidales__f_Rikenellaceae
g_Clostridium__s_Clostridium_bolteae
g_Erysipelatoclostridium__s_Erysipelatoclostridium_ramosum
log2FC
Figure 4
Figure 4. KB39 shifts the gut microbiome in western diet-fed Ldlr-/- mice and promotes propionate-
producing bacterial taxa. Male Ldlr-/- mice were fed a western diet, western diet supplemented with KB39
(7.5% w/w) or fenofibrate (100 mg/kg/day) for 16 weeks. Fecal samples were collected before treatment
during the last week of the treatment period. Significantly enriched and depleted taxa (fdr < 0.1, Wilcox rank
sum test with FDR correction) at phylum, family, genus, and species levels in KB39 treated group in
comparison to no treatment group at the end of the study.
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
0 4 8 12 16
80
100
120
140
160
180
Weeks
Relative BW (%)
***
****
NT KB39 Feno
*
0 4 8 12 16
0
1
2
3
4
5
Weeks
Daily food intake (g)
0 30 60 90 120
0
100
200
300
400
500
Time (Min)
[Glucose] (mg/dL)
**
** ***
**
NT KB39 Feno.
0
1
2
3
Liver steatosis score
✱✱
✱✱
NT KB39 Feno.
0
1
2
3
Liver inflammation
ns
✱✱✱
NT KB39 Feno.
0
1
2
3
4
5
6
7
8
NAS
✱
✱✱✱
NT KB39 Feno.
0
1
2
3
4
5
6
7
8
9
10
11
12
Plaque severity score
✱
✱
NT KB39 Feno.
0.0
0.2
0.4
0.6
0.8
Aortic sinus
plaque area (mm2)
✱
✱✱
NT KB39 Feno.
0
5
10
15
20
25
Aortic arch
plaque area (%)
ns
✱✱
A B C
D E F
G H I
Figure 5. KB39 attenuates liver steatosis, liver inflammation and atherosclerosis in western diet-fed Ldlr-/- mice.
Male Ldlr-/- mice were fed a western diet, western diet supplemented with KB39 (7.5% w/w) or fenofibrate
(100 mg/kg/day) for 16 weeks. An OGTT was performed at week 15. (A-C) Relative body weight (BW) compared to
BW at treatment initiation (A), daily food intake (B) and OGTT blood glucose measurements (C); mean ± SEM, 2-way
ANOVA, Dunnett. (D-F) Liver steatosis score (D), inflammation (E) score and NAFLD activity score (NAS) (F) at
termination; median, Kruskal-Wallis, Dunn. (G) Aortic sinus plaque severity score; median; Kruskal-Wallis, Dunn. (H,
I) Plaque area in the aortic sinus quantified by morphometric analysis of elastin trichrome-stained sections (H) and
in the aortic arch quantified by oil red O staining “en face” analysis (I), 1-way ANOVA, Dunnett. * P<0.05, ** P<0.01,
*** P<0.001, **** P<0.0001. NT: no treatment; Feno: fenofibrate.
Figure 5
.CC-BY-NC-ND 4.0 International licenseavailable under a
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
The copyright holder for this preprint (whichthis version posted April 4, 2022. ; https://doi.org/10.1101/2022.04.04.487010doi: bioRxiv preprint
