Microbiome methods in experimental autoimmune encephalomyelitis

Abstract

Microbiome composition studies are increasingly shedding light on animal models of disease. This paper describes a protocol for analyzing the gut microbiome composition prior to and after the induction of mice to experimental autoimmune encephalomyelitis (EAE), the principal animal model of human neuroinflammatory demyelinating disease multiple sclerosis (MS). We also address and provide data assessing the impact of mice reared in different animal facilities on EAE induction. Furthermore, we discuss potential regulators of the gut-microbiome-brain axis (GMBA) in relation to neuroinflammation and implications on demyelinating disease states. Our results suggest that mice reared in different animal facilities produce different levels of EAE induction. These results highlight the importance of accounting for consistent environmental conditions when inducing EAE and other animal models of disease.

INTRODUCTION:

Studies on gut microbiome compositions are included in an increasing number of experimental approaches in animal models of disease. Due to the critical roles of the gut microbiome on health and disease, the composition of intestinal microbes, collectively known as microbiota, is compared among experimental groups, such as disease vs. healthy animals, treated vs. untreated animals, fecal microbiota transplantations (FMT), among others. In the context of experimental autoimmune encephalomyelitis (EAE), a model for the study of central nervous system inflammatory demyelination that characterizes multiple sclerosis (MS). In rodents, the microbiome’s composition has been addressed when comparing the effects of antibiotics, monocolonization with microbial species of interest, FMT with microbes isolated from healthy and MS donors, and probiotic formulations. Different models of EAE, using different strains of mice, such as non-obese diabetic (NOD), SJL, and C57BL/6, have been employed. In addition, the time points used for the isolation of samples and analysis of the microbiome also vary and often include pre-disease induction, disease severity peak, and termination of the study.

When performing microbiome studies in the EAE model it is critical to control the environment of the animal facility and experimental housing conditions. Diet, source of drinking water, conditions of caging, and the source of the mice used in the studies can affect the results of the microbiome studies, but also the onset, disease incidence, and severity of the EAE models. In this current work, we evaluate the impact of the source of C57BL/6 mice on the composition of the microbiome and disease parameters. The protocol describes the methodology for stool isolation, storage, extraction of DNA, sequencing, and analysis of microbiome composition in C57BL/6 mice obtained from two commercial vendors (Jackson Laboratory and Envigo). The methodology listed describes the steps used in some of our previously published works (Colpitts et al., 2017; Sell et al., 2021). Specifically, it addresses the source of the mice as a potential variable that affects the study of the gut-microbiome-brain axis (GMBA) in the context of EAE.

BASIC PROTOCOL 1

Study of the composition of the gut microbiome in the neuroinflammatory model of experimental autoimmune encephalomyelitis (EAE)

We discuss and detail a protocol for studying the gut microbiome composition in wild-type mice prior to and after the induction of EAE, a widely used model for the study of MS. We provide the tools required for quantifying the microbiome and comparing the gut microbiome composition among groups. Furthermore, we evaluate the impact on the results of using mice obtained from different commercial vendors. As an illustration, we show the results on the microbiome analysis of a single experiment with naïve and EAE mice obtained from Envigo and the Jackson Laboratory.

MS is a disease that involves a complex interaction between the central nervous system (CNS) and the peripheral immune system. Mice are the primarily species used in EAE studies. EAE can be induced passively by adoptive transfer of autoreactive T cells or actively by subcutaneous injection of self-antigens obtained from neuronal myelin homogenates: proteolipid protein (PLP), myelin basic protein (MBP), myelin oligodendrocyte glycoprotein (MOG), or peptides corresponding to the encephalitogenic portions of these proteins. In mice, induction requires the emulsification of one of the self-antigens in complete Freund’s adjuvant (CFA) and additional intraperitoneal or intravenous administration of pertussis toxin (PT). Depending on the antigen and mouse strain selected, different patterns of disease that model distinct forms of human MS are observed (for a review, McCarthy et al., 2013). Our protocol describes the active EAE form induced with MOG_35-55 in C57BL/6 mice. In this model, peripheral tolerance is disrupted by the injection with a self-antigen, resulting in the proliferation of self-antigen-specific T cells that differentiate into effector, pathogenic cells within the secondary lymphoid tissues. From there, cells circulate through lymphatics to the bloodstream. The process of T cell activation also results in the acquisition of integrins. These surface molecules allow the cells to attach to endothelial cells and cross the blood-brain barrier into the CNS parenchyma. There, autoreactive, pathogenic T cells are reactivated by resident antigen-presenting cells, which results in exacerbated production of proinflammatory mediators, such as proinflammatory cytokines (IFN-γ, TNF-α, GM-CSF, IL-17). Other cells cross the blood-brain barrier, such as monocytes, neutrophils, and dendritic cells, and contribute to the pathogenesis and progression of the disease. Table 1 shows the three most common strains of mice used in EAE studies; other strains have not been included. In C57BL/6 mice, the condition is characterized by a constant progression into a chronic stage with minor, if any, periods of remission caused by sustained priming antigen-specific T cell responses.

Table 1.

Three most used models of EAE used in mice.

Peptideepitope	Strain	Adjuvant**	EAE observed
			Pattern of disease	Onset (day)	Peak(s) (day)
PLP135-155	SJL	CFA + PT	Relapsing-remitting	9-13	15-21
MOG35-55	C57BL/6	CFA + PT	Chronic	9-13	15-21
MOG35-55	NOD*	CFA + PT	Biphasic: relapsing-remitting first, chronic next	9-13	1st: 15-21 2nd: approx. 50

^*NOD: non-obese diabetic

^**Adjuvant: CFA, complete Freund’s adjuvant; PT, pertussis toxin.

Active EAE is induced by subcutaneous injection of an emulsion of self-peptide (in C57BL/6 mice, the self-peptide is myelin oligodendrocyte glycoprotein, or MOG, emulsified in CFA) and two intraperitoneal (i.p.) doses of Pertussis Toxin (resuspended in sterile saline, approximately 2 h after subcutaneous (s.c.) administration of MOG_35-55/CFA, and 24-48 h). Controlling the amount of self-peptide emulsified in CFA and PT and the effectiveness of the emulsification process will directly affect the incidence and severity of EAE induced. Since our goal is to determine the impact of the gut microbiome on the EAE model, our laboratory standardizes the disease induction process by using a commercially available EAE induction kit (Hooke laboratories, catalog number EK2110). Detailed here is the protocol used for the MOG_35-55 EAE model in C57BL/6 mice, which triggers a progressive neurological disease with extensive plaque-like demyelination, common to the manifestations of MS.

As discussed later, the results obtained using our protocol highlight the importance of the provider of the rodent strain selected for the EAE study. Stool samples obtained the day of EAE induction, 14 days post-induction (dpi), and at the end of the experiment (21 dpi) were processed for DNA extraction, sequencing, and analysis. The protocol details the experimental methodology used from the moment the animals arrived at our facilities until the microbiome analysis was performed. We also discuss some of the most important factors to consider when designing microbiome studies in the EAE model.

Materials:

• Mice: Age and sex affect EAE induction and severity. Use female C57BL/6 mice, 9 to 13 weeks old. The example protocol data presented in this paper illustrates the use of ten-week-old female C57BL/6 NHsd mice from Envigo RMS, Inc. (Indianapolis, IN, USA), and ten-week-old female C57BL/6J from The Jackson Laboratory (Bar Harbor, ME, USA). All mice weighed approximately 20 g when they arrived at our Eastern Washington University’s facilities.

• Hooke Kits™ for EAE Induction in C57BL/6 Mice (Hooke laboratories, catalog number EK2110).

• Sterile phosphate buffer saline (sPBS)

• 70 % ethanol

• 1ml syringes

• 27G ½” needles

• 50 ml conical tubes

• Forceps

• Eppendorf tubes

• Freezer (−80 °C)

• Multichannel pipettes (10, 200 and 1,000 μl)

Experimental procedures for EAE induction and stool collections

1. Housing conditions:

House the mice in groups of 5 to provide them with social interaction, in standard animal facility cages with wire top at a temperature of 22 +/− 1 °C, and a humidity of 23-33 % with a 12-hour light/dark cycle. The cage dimensions are 46 cm x 25 cm x 20 cm, and the mice live with 12-hour light and 12-hour dark cycles.

• 1.1.Acclimate the animals to the animal facilities and housing conditions for at least one week prior to initiating the study.
• 1.2.Administer food and water ad libitum during the experimental period. Each mouse should be fed 4g of pellet food per day, as recommended by the Humane Society of the United States.

2. EAE induction:

The protocol shown was adapted from Hooke Labs (https://hookelabs.com/protocols/eaeAI_C57BL6).

• 2.1.
  Administration of MOG_35-55/CFA emulsion contained in the Hooke Laboratories Kits (day 0). The emulsion is provided and should be directly administered s.c., at two sites, 0.1 mL/site (0.2 mL/mouse total). Anesthesia with isoflurane is optional and is dependent on specific IACUC guidelines.

  • 2.1.1.
    Inject mouse s.c. on the upper back with 0.1 mL of emulsion. The needle should be kept inserted for 15 seconds to avoid leakage of MOG_35-55/CFA emulsion.
  • 2.1.2.
    Inject s.c. the emulsion on the lower back with 0.1 mL. The needle should be kept inserted for 15 seconds to avoid leakage of MOG_35-55/CFA emulsion.
• 2.2.
  The first administration of PT (day 0): Dose of PT significantly affects EAE incidence and severity (Table 4). Follow guidelines provided by the commercial kit manufacturer for specifics regarding PT batch and disease severity. PT must be diluted in sPBS or water to reach the appropriate concentration (following vendor’s specifications). The study provided here used 140 ng of PT per mouse diluted in 0.1 ml sPBS (0.1 ml/mouse per dose):

  • 2.2.1.
    Wait between 1 to 6 h after MOG_35-55/CFA emulsion administration.
    PT solution must be prepared fresh under a biosafety cabinet at sterile conditions on the day of EAE induction and again on the day of second administration (1 dpi). Therefore, PT solution should be prepared fresh on day 0 and again prepared fresh on day 1, under sterile conditions using a biosafety cabinet.
  • 2.2.2.
    Prepare a 20% extra solution to allow loss during administration. For 30 mice used in the study (n = 30 mice): 30 * 0.12 ml sPBS = 3.6 ml sPBS in a 50 ml conical tube.
  • 2.2.3.
    Centrifuge PT vial for 10 seconds to deposit all volume in the bottom of the vial.
  • 2.2.4.
    Dilute stock of PT provided with the kit to the 50 ml conical tube with sPBS: for 30 mice at 140 ng PT per mouse, prepare: 30 mice * 140 ng PT * 1.2/200 ng/μl = 25.2 μl PT. Add the PT from the stock vial to a 50 ml conical tube with 3.6 ml sPBS.
  • 2.2.5.
    Pipette up and down several times to mix (do not vortex).
  • 2.2.6.
    Use 1 ml syringes to draw 1 ml of PT solution. Mount sterile 27G ½” needles.
  • 2.2.7.
    Inject 0.1 ml/mouse i.p.
• 2.3.
  The second administration of PT (day +1: 22 – 26 h after first administration):

  • 2.3.1.
    Prepare fresh PT solution repeating calculations and protocols indicated for first PT administration.
  • 2.3.2.
    Use 1 ml syringes to draw 1 ml of PT solution. Mount sterile 27G ½” needles
  • 2.3.3.
    Inject 0.1 ml/mouse i.p.

Table 4.

Troubleshooting guide for microbiome studies in EAE

Problem	Possible Cause	Solution
Low incidence and severity of EAE	The inadequate concentration of pertussis toxin used	Increase concentration of pertussis toxin used
		Consider using a different source for the animals used
	The young age of the animals	Consider allowing the animals to mature more before administering the pertussis toxin
The severity of EAE is too high	The inadequate concentration of pertussis toxin used	Reduce the concentration of pertussis toxin used
		Consider using a different source for the animals used

3. Disease monitorization and termination of the study:

The mice are scored as described previously (Sell et al., 2021).

• 3.1.Score 0 – no detectable signs of EAE; score 0.5 – distal limp tail; score 1.0 – complete limp tail; score 1.5 – limp tail and hind limb weakness; score 2.0 – unilateral partial hind limb paralysis; score 2.5 – bilateral partial hind limb paralysis; score 3.0 – complete bilateral hind limb paralysis; score 3.5 - complete bilateral hind limb paralysis and partial front limb paralysis; score 4.0 - quadriplegia. A score of 5 is assigned to deceased animals. A detailed description of the clinical scoring used is shown in table 2.
• 3.2.According to the specific institutional guidelines, mice suffering from clinical scores above a particular value must be euthanized. At our facilities, animals suffering from clinical scores > 3.0 for two consecutive days must be euthanized (5 is assigned).
• 3.3.To ease mouse access to water, mice are given water bottles upon disease induction/injection. To facilitate mouse access to food, when mice exhibit a score of at least 2.5 or greater, they are left food soaked with water in the bottom of the cage in the same corner where bedding being used by the mice is located. Mice maintaining partial forelimb mobility should be able to access this food source. Any indication of an inability to access this food will imply forelimb weakness, and mice must be euthanized.
• 3.4.The mice are scored daily for the duration of the experiment (Fig. 2). In addition, body weights are collected at least once weekly, always including the day of EAE induction.

Table 2.

List and description of EAE clinical scores^*.

EAE score	Clinical observations
0.0	No obvious changes in motor function compared to non-immunized mice. When picked up by base of tail, the tail has tension and is erect. Hind legs are usually spread apart. When the mouse is walking, there is no gait or head tilting.
0.5	Tip of tail is limp. When picked up by base of tail, the tail has tension except for the tip. Muscle straining is felt in the tail, while the tail continues to move.
1.0	Limp tail. When picked up by base of tail, instead of being erect, the whole tail drapes over finger. Hind legs are usually spread apart. No signs of tail movement are observed.
1.5	Limp tail and hind leg inhibition. When picked up by base of tail, the whole tail drapes over finger. When the mouse is dropped on a wire rack, at least one hind leg falls through consistently. Walking is very slightly wobbly.
2.0	Limp tail and weakness of hind legs. When picked up by base of tail, the legs are not spread apart, but held closer together. When the mouse is observed walking, it has a clearly apparent wobbly walk. One foot may have toes dragging, but the other leg has no apparent inhibitions of movement.
2.5	Limp tail and dragging of hind legs. Both hind legs have some movement, but both are dragging at the feet (mouse trips on hind feet). No movement in one leg/completely dragging one leg, but movement in the other leg, is also possible.
3.0	Limp tail and complete paralysis of hind legs.
3.5**	Limp tail and complete paralysis of hind legs. The mouse moves around cage, but if placed on its side, the mouse is unable to right itself. It is also possible that the hind legs are on one side, paralyzed, and together. The mouse can move, but hind legs are flat like a pancake.
4.0**	Limp tail, complete hind leg and partial front leg paralysis. Mouse is minimally moving around the cage but appears alert and feeding.
4.5**	Complete hind and partial front leg paralysis, no movement around the cage. Mouse is not alert. Mouse has minimal movement in the front legs. The mouse barely responds to contact. Euthanasia is recommended. When the mouse is euthanized because of severe paralysis, a score of 5.0 is entered for that mouse for the rest of the experiment.
5.0**	Mouse is spontaneously rolling in the cage, is found dead due to paralysis, or is euthanized due to severe paralysis.

^*The table was adapted from detailed protocol described by Hooke Laboratories, used to induce EAE in the study (https://hookelabs.com/protocols/eaeAI_C57BL6.html).

^**Per our institution’s IACUC guidelines, when a mouse reaches a score of 3.5 for two consecutive days, the mouse is euthanized and given a score of 5.0 for the rest of the experiment. When an animal is euthanized due to severe disease a score of 5 is assigned for the duration of the study.

4. Stool collection and storage:

• 4.1.Obtain fresh stool pellets (1-2 per mouse). The samples can be collected daily, if needed. For studies of the microbiome in EAE experiments, a collection before disease induction, at the peak of the disease and at the end of the experiment should be considered. However, the collection times depend on the appropriate experimental design. For the example described in this paper, we collected samples 0. 14 and 21 dpi. Rinse forceps in 70% ethanol and wipe off excess with a disposable tissue before obtaining the stool sample. Take a sample directly from the animal while excreted and store at −80 °C. Do not collect dry pellets from the cage or bench surface.
• 4.2.Although the samples can be stored in preservative buffers, the samples in the protocol were stored dry, without any buffer added. The samples used in this protocol were maintained at −80 °C for approximately two months before shipping. Different preservation and storage of stool samples and effects on data quality are available in the literature (Al et al., 2018). If samples are shipped for analysis, use dry ice and follow established guidelines for handling and shipping of frozen biological samples.

Basic Protocol 2

Experimental procedures for DNA extraction and microbiome analysis

The advances in the molecular techniques and large data management developed over the last two decades allow us now to scrutinize the composition of microbiomes in the environment and organisms. Herein, we detail a protocol for the extraction of DNA from samples obtained from EAE and healthy mice (Basic protocol 1), sequencing of the ribosomal DNA (rDNA), and analysis. For the analysis, we use free platforms made available by the National Institutes of Health (NIH): https://nephele.niaid.nih.gov.

Materials

• MagAttract PowerMicrobiome DNA/RNA Kit (formerly known as the PowerMag Microbiome RNA/DNA isolation kit) (27600-4-KF)

• Zymo DNA Clean & Concentrator (D4024).

• Qaunt-IT dsDNA Broad Range assay (Q33120)

• Kapa Hifi HotStart PCR Kit (Kapa Biosystems, KK2502).

• Nextera XT Index Kit v2 (Illumina, FC-131–2001).

• 16S rDNA V4 region primers (Table 3)

• Qubit™ dsDNA HS (high sensitivity) assay kit (Thermo Fisher Scientific, Q33230)

• Qubit™ fluorometer (Thermo Fisher Scientific)

• Agilent 2100 Bioanalyzer DNA 1000 chip

• Illumina sequencer

• PhiX 10% (Illumina)

• MiSeq Reagent Kit v3 (600-cycle; Illumina, MS-102-3003).

• Nuclease-free water (Thermo Fisher Scientific, 4387936)

• Software, or access to NIH’s free microbiome platform: https://nephele.niaid.nih.gov.

Table 3.

Primers used for amplification of 16S rDNA V3-V4 regions

Primer	Transpose sequence	Spacer
			341F 16S Forward
16S_V3V4_341F_F1	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG		CCTACGGGNGGCWGCAG
16S_V3V4_341F_F2	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	C	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F3	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	AC	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F4	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	TAC	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F5	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	GTAC	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F6	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	CGTAC	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F7	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	ACGTAC	CCTACGGGNGGCWGCAG
16S_V3V4_341F_F8	TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG	TACGTAC	CCTACGGGNGGCWGCAG
			785R 16S Reverse
16S_V3V4_785R_R8	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG		GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R7	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	G	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R6	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	TG	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R5	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	ATG	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R4	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	CATG	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R3	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	TCATG	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R2	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	ATCATG	GACTACHVGGGTATCTAATCC
16S_V3V4_785R_R1	GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG	CATCATG	GACTACHVGGGTATCTAATCC

1. DNA extraction:

• 1.1.DNA from stool samples is extracted using the MagAttract PowerMicrobiome DNA/RNA Kit (formerly known as the PowerMag Microbiome RNA/DNA isolation kit) (27600-4-KF), following manufacturer protocols on the KingFisher Flex. 250 mg of each stool samples were used in the protocol described. The DNA extraction yield and picogram concentration for each sample used in the protocol described is shown in the supplemental table 1.
• 1.2.Extracted samples are purified using the Zymo DNA Clean & Concentrator (D4024).
• 1.3.DNA is quantified using the Qaunt-IT dsDNA Broad Range assay (Q33120) following manufacturer protocols.
• 1.4.Extracted samples are normalized to 5 ng/μL in nuclease-free water for sequencing. We typically obtain 25 ul of DNA in water.

2. Microbiome sequencing:

• 2.1.Primers containing a 0 to 7 bp heterogeneity spacer and universal 16S rRNA sequence are used to amplify and sequence the V3-V4 hypervariable region of the 16S rRNA gene, using the Kapa Hifi HotStart PCR Kit (Kapa Biosystems, KK2502). Primers used in this protocol are shown in table 3.
• 2.2.Amplicons are purified using magnetic beads and barcoded using the Nextera XT Index Kit v2 (Illumina, FC-131–2001).
• 2.3.The multiplex library pool is quantified using the Qubit dsDNA HR Assay, to determine double-stranded DNA quality, diluting the samples in buffer provided with the kit. The specific concentrations and dilutions used are shown in the supplemental document 1. One to 20 μl of sample (library) can be used, and should be diluted in same volume of buffer, ready to be used at the concentration provided with the kit. One diluted 1:1 in the buffer, dsDNA is quantified with a Qubit fluorometer, following the vendor’s specifications.
• 2.4.
  Once the concentration of dsDNA is quantified, the fragment size of library pool is checked using the Agilent 2100 Bioanalyzer DNA 1000 chip. The specific protocol provided with Qubit dsDNA HR Assay should be followed, including the following steps:

  • 2.4.1.
    Prepare the Gel-Dye Mix (volumes provided are enough for 10 chips). Prior to use, the DNA dye concentrate, and the DNA gel matrix provided with the kit should be equilibrated to room temperature for 30 minutes. The DNA dye is then vortexed for 10 seconds and then spin down. The dye contains DMSO, which must be completely thawed after spinning. Once DMSO is removed, 25 μl of the DNA dye is added to the DNA gel matrix. The mix of dye and matrix is then vortexed and filtered using a spin filter (provided with kit) for 15 minutes at 6,000 x g, at room temperature. After centrifugation, the filter is discarded.
  • 2.4.2.
    Load the Gel-Dye Mix. Opening a chip (provided), pipette the gel-dye mix (9.0 μl) at the bottom of the wells, following the instructions provided with the kit.
  • 2.4.3.
    Load the marker. Add 5 μl of DNA marker at the well indicated by the vendor, and into 12 samples designated for the samples.
  • 2.4.4.
    Load ladder and samples. One μl of ladder is added to the well designated by the vendor. One μl of sample is added to the wells designated for samples by the vendor.
  • 2.4.5.
    Insert chip in the Agilent 2100 Bioanalyzer and starting the chip Run, following vendor’s instructions.
  • 2.4.6.
    Validate the results obtained. Per vendor’s instructions, a successful ladder run should contain 13 peaks, well resolved, with a flat baseline and the identification of the lower and upper markers. A successful sample run, the peaks should appear between lower and upper marker limits, there should be a flat baseline of at least 5 fluorescence units as defined by the vendor’s instructions. Samples should be at least three fluorescence units higher than the baseline, as indicated by the vendor’s instructions. Specific examples of the validation can be found at the vendor’s instructions (https://www.agilent.com/cs/library/usermanuals/public/G2938-90014_DNA1000Assay_KG.pdf). The quality of results obtained in the sequencing is shown in supplemental document 2.
• 2.5.Sample sequencing. The specific protocol to be followed depends on the sequencer. The sequencing was achieved using an Illumina MiSeq and the MiSeq Reagent Kit v3 (600-cycle; MS-102-3003).
• 2.6.The multiplexed library is spiked with 10 % PhiX. PhiX is a DNA library control used to validate sequencing in illumine MiSeq runs. The concentration used depends on the sequencer. For the sequencer used in this protocol, the recommended concentration is 10 %. PhiX 10 % is provided by Illumina as a ready-to-use reagent.

3. Data Curation and Preparation

The following steps and procedures are for individuals familiarized with sequence data and bioinformatics. Sseveral resources are available to help guide less advanced users in the bioinformatics of the microbiome. One suggested guide is provided by the “Microbiome Discovery” series on YouTube by Dan Knights (https://www.youtube.com/watch?v=6564K4-_DBI&list=PLOPiWVjg6aTzsA53N19YqJQeZpSCH9QPc&index=2).

• 3.1.Following sequencing on the Illumina MiSeq, data from the sequencing step are uploaded to Illumina’s online platform, BaseSpace. Create an account with Illumina to access your sequences on the BaseSpace platform. Near the top of the website, click “Projects” and locate the sequence run of interest.
• 3.2.Click the hyperlink on the sequence run of interest and open the FASTQ generation file. FASTQ files include the nucleotide sequences from each sample. If the same sequencing protocol as the aforementioned protocol is followed, the user will notice that a forward and reverse set of primers were used (Table 3). Each sample will contain two separate FASTQ files which correspond to the forward and reverse sequences.
• 3.3.Once inside the FASTQ generation folder, locate “download analysis” near the top of the page. Click “download analysis”. If you have not downloaded the BaseSpace Sequence Hub Downloader, you will need to click “Install the BaseSpace Sequence Hub Downloader and follow the guided prompts. Once installed, click the “All file types including VCF, BAM, & FASTQ” option followed by the “download” option.
• 3.4.Once the files are downloaded successfully, locate the files and save them to a specified folder you can access. We recommend that your save these files on a separate hard drive-in addition to your computer’s hard drive to ensure these files can be accessible in the future. BaseSpace will delete files if you do not log into BaseSpace regularly.
• 3.5.Using your internet browser, locate the Nephele site on the NIH’s NIAID site. If you have not done so, please create an account before moving onto the subsequent steps (Nephele: https://nephele.niaid.nih.gov/).
• 3.6.Click the “Analyze” step near the top of the page. This will show you all the available pipeline’s Nephele offers. If you are following the aforementioned protocol, you will recognize that we have 16S rDNA sequence data. Therefore, we are limited to the Amplicon metagenomics column on the left-hand side of the page. The options available are MOTHUR, ITS, DADA2, QIIME2 and QIIME1. MOTHUR is an excellent pipeline that is user friendly but is limited by the amount of data you have on the Nephele platform. If you used the same sequences for the sequence protocol, the data is too large and therefore MOTHUR is not appropriate for your needs. ITS is an excellent pipeline to use if you have fungal ITS gene data. This protocol uses 16S rDNA sequences, therefore ITS is not appropriate. DADA2, QIIME2, and QIIME1 are the best pipelines for this set of data. The difference between QIIME1 and QIIME2 is just the version of QIIME. QIIME1 is no longer supported and therefore QIIME2 might be the better option to run. We chose to run the QIIME1 platform for our primary needs simply due to user familiarity. Moving forward, click “Start Job” under QIIME1.
• 3.7.Click “Paired End FASTQ” if you have a forward and reverse set of data (which can be discerned by having 2 FASTQ files per sample, typically denoted as R1 for Forward and R2 for Reverse). This will give you various options for uploading your sequences to the pipeline. Since we manually downloaded the sequences onto our hard drive, click “Upload from my computer”.
• 3.8.Locate the file containing all your downloaded FASTQ files. Click and drag the folder into the dotted rectangle space. Next, click the “Select All” option near the middle of the page. This will allow you to apply the next action to all your files. Click “Start upload” and wait for each file to upload successfully. A gauge will appear with an estimated time until completion for you to track and monitor the upload progress. Once the progress is complete, click “Next” at the bottom of the page. You may need to scroll down to locate this option.
• 3.9.Next, you will need to upload a mapping excel file. Instructions and a template for the mapping file can be found here: https://nephele.niaid.nih.gov/user_guide/. Column A will be denoted as “#SampleID”. How you choose to name each sample based on the sample is entirely up to you. We tend to use clear signifiers such as the animal number or cage number and animal number. Column B will be denoted as “ForwardFastqFile”. Column B must be the exact file name for the forward FASTQ sequence file as stored on your hard drive. Column C will be denoted as “ReverseFastqFile”. Column C must be the exact file name for the forward FASTQ sequence file as stored on your hard drive and must be the corresponding reverse file to the previous forward file. Column D will be your treatment group signifier. For this experiment, we considered animal supplier as the primary variable separating Jackson from Envigo. The next few columns can be optional metadata. We created a metadata file for disease status as well as timepoint. The final column will be denoted as “Description” regardless of the number of metadata columns you wish to include. Once you are finished, save the file and upload the mapping file to nephele.
• 3.10.If there are any errors in your mapping file, nephele will alert you to the problem. Most simple problems can be corrected directly in the web browser by clicking on the cell that is highlighted and correcting the issue. Note: this will not fix the saved file, and any changes you make on the browser will not be reflected in the original mapping file.
• 3.11.The next stage is to select the parameters for your pipeline. These parameters act as filters and can drastically alter the output files from the pipeline. More seasoned bioinformaticists may alter these parameters for their needs, however, for beginners we recommend you keep all the parameters as they are. We changed the reference database to the “Greengenes 97” database. We also used the closed reference approach strategy.
• 3.12.Submit the pipeline and you will receive a conformation email from Nephele. This email will include information regarding the parameters set in the pipeline, the job name, and a link to watch the run of the pipeline in real time. You must wait for the pipeline to be completed and this step can take a variable amount of time. Once the pipeline is complete, you will receive an additional email containing a link to where you can download the output files from the pipleline.

4. Analysis of microbiome’s composition:

• 4.1.Upon completion of the pipeline, various output files are generated based on the pipeline script. The operational taxonomical unit (OTU) table is an Excel file containing a curated list of each observed taxon listed on each row and the count of each taxon in each sample in each column.
• 4.2.The OTU table is transformed, so each row represents the samples, and each column represents the taxon. This new OTU table is saved as a comma-separated value file.
• 4.3.Using R, an in-house script is generated to visualize the compositional heterogeneity using non-metric multidimensional scaling and the metaMDS function in the vegan package. The specific code used in this protocol is shown in supplemental documents 3 (alpha diversity) and 4 (beta diversity, or compositional heterogeneity).
• 4.4.Dissimilarity in the heterogeneity is defined using the Bray-Curtis Dissimilarity index. To statistically evaluate the heterogeneity of the samples, the permutational ANOVA strategy can be used via the ADONIS function in vegan and visualized by NMDS analysis. Since ADONIS is a permutational strategy, 1,000 permutations are performed and the average p-values (adjusted) are reported.
• 4.5.Additional file outputs from the pipeline include files containing the α-diversity measurements of each sample reported using the Shannon and Chao1 diversity calculations. These data are analyzed and visualized using the built-in box plot function and statistically evaluated via one-way ANOVA, followed by Tukey’s multiple comparisons test.
• 4.6.The .biom file generated from the pipeline is uploaded to the MicrobiomeDB web-based platform.
• 4.7.Using the relative abundance investigational tool, the samples are separated by treatment and timepoint. The median count of relative abundance for each phylogenetic level and each taxon is compared and statistically evaluated using the Wilcoxon Rank Sum Test.

COMMENTARY:

Background Information:

There is potential crosstalk between the GMBA, neurotransmitters, and neuroinflammation. Despite extensive research and efforts developing novel and safer therapeutics, CNS diseases involving intrinsic and peripherally induced inflammatory pathways remain a clinically unmet need. In recent years, much attention has been put on the GMBA and dysbiosis observed in patients and experimental models of CNS inflammatory diseases. Among them, demyelinating conditions have been explored in detail. MS is an autoimmune disease of the CNS. During MS, inflammatory pathways trigger a process of demyelination, axonal damage and loss, and neuronal death that result in primarily three forms of devastating diseases: relapsing-remitting MS (RRMS), primary progressive MS (PPMS), and secondary progressive MS (SPMS) (Baranzini & Oksenberg, 2017). MS affects individuals from both sexes, although with increased frequency in women than men (FRCP et al., 2018). Although the etiology of MS remains unknown, both genetic and environmental factors have been proposed (Baranzini, 2011; Baranzini & Oksenberg, 2017; Olsson et al., 2017). Among the environmental factors, the composition of the gut microbiome has been proposed as a potential mechanism of disruption before or during disease onset (Ochoa-Repáraz & Kasper, 2014, p.). Recent advances in molecular techniques allow us to evaluate the impact of the microbiome on the disease (Baranzini, 2018a, 2018b). The most recent studies indicate that dysbiosis, the altered composition of the microbiome that could lead to inflammation and disease, is observed in stool samples of MS patients when compared to healthy donors (Berer et al., 2017; Cekanaviciute et al., 2017; Chen et al., 2016; Jangi et al., 2016; Miyake et al., 2015; Takewaki et al., 2020; Tremlett, Fadrosh, Faruqi, Hart, et al., 2016; Tremlett, Fadrosh, Faruqi, Zhu, et al., 2016).

To date, there are more than 14,000 peer-review published works listed in PubMed linked to EAE. The number of studies exploring the interactions between the microbiome and the disease has grown since 2008. Early studies showed that the oral administration of antibiotics resulted in reduced severity of EAE (Ochoa-Reparaz et al., 2009; Yokote et al., 2008). The effects of gut microbiome alterations with antibiotics in EAE were soon confirmed in germ-free models (Colpitts et al., 2017; Ochoa-Reparaz & Kasper, 2018, p.). Since then, the EAE model has been used in numerous microbiome studies. The use of broad-spectrum antibiotics (Carrillo-Salinas et al., 2017; Colpitts et al., 2017; Ochoa-Reparaz et al., 2009; Yokote et al., 2008), germ-free rodent models (Berer et al., 2011; Lee et al., 2011), fecal microbiota transplantation (FMT), (Berer et al., 2017; Cekanaviciute et al., 2017), monocolonization studies (Berer et al., 2011; Lee et al., 2011), and probiotic approaches (Calvo-Barreiro et al., 2020; Kwon et al., 2013; Lavasani et al., 2010; Maassen & Claassen, 2008; Mangalam et al., 2017; Secher et al., 2017; Shahi et al., 2019; Tankou et al., 2018), and symbiotic factors purified from gut bacteria (Erturk-Hasdemir et al., 2019; Ochoa-Repáraz et al., 2010, p.; Wang et al., 2014) have been used to discern the association between the microbiome and the CNS diseases mechanistically.

A bidirectional association between the gut microbiome and CNS inflammatory demyelination has been hypothesized (Colpitts et al., 2017; Ochoa-Reparaz & Kasper, 2018, p.). Previous work, done in our laboratory and others, indicates the induction of EAE alters the gut microbiota composition (Colpitts et al., 2017), intestinal barrier permeability (Nouri et al., 2014), and gut-associated immune responses (Nouri et al., 2014). Our previous work also indicates that among those microbes affected by the induction of EAE are bacteria capable of producing neurotransmitters (Strandwitz et al., 2019).

Several neurotransmitters have been implicated in contributing to the modulation of the GMBA, including dopamine (DA), norepinephrine (NE), serotonin (5-HT), glutamate (GLUT), and gamma-aminobutyric acid (GABA). Three common experimental strategies used to investigate neurotransmitter involvement of the GMBA are: 1.) germ-free models, utilizing animals with reduced intestinal flora; 2.) dysbiosis, where the microbial population is disrupted; and 3.) probiotic treatment, where the microbial population is enhanced. Germ-free models have been employed to explore the neurotransmitter systems implicated in GMBA regulation. Germ-free mice have increased DA receptor mRNA expression in the hippocampus, decreased DA receptor mRNA expression in the striatum, and increased turnover rates of DA, NE, 5-HT in striatum (Heijtz et al., 2011), compared to controls. Germ-free mice have decreased serum levels of 5-HT (Yano et al., 2015), DA (Velagapudi et al., 2010), and GABA (Matsumoto et al., 2012), relative to conventionally raised mice. Furthermore, germ-free mice have altered gene expression related to monoamine neurotransmitters (Arvidsson et al., 2012; Pan et al., 2019). Studies using dysbiosis also suggest that neurotransmitters are involved in GMBA regulation. Mouse gut dysbiosis increases the levels of DA metabolites (i.e., L-3,4-dihydroxyphenylalanine (L-DOPA) and homovanillic acid (HVA)) in the amygdala (Desbonnet et al., 2015). Hoban et al. (2016) demonstrated that gut dysbiosis altered 5-HT, NE, L-DOPA, and HVA levels. Disruption of the microbial population in a mouse model of depression decreases NE and 5-HT (Wu et al., 2020). Additionally, Wu et al. (2020) revealed correlations between neurotransmitter levels and differential gut bacterial taxa. Probiotics alter neurotransmitter systems (Sandhu et al., 2017). Furthermore, specific gut bacteria have been implicated in the modulation of host neurotransmitters or related pathways. Probiotics containing Lactobacillus rhamnosus JB-1 alter GABA receptor mRNA expression in many brain areas, including the hippocampus and amygdala (Bravo et al., 2011), as well as increase GABA levels in the brain (Janik et al., 2016). Probiotics containing Lactobacillus plantarum PS128 increase DA and 5-HT in the striatum (Liu et al., 2016), and those containing Lactobacillus casei 54-2-33 increase 5-HT receptor mRNA expression in the hippocampus (Barrera-Bugueño et al., 2017). Studies suggesting an association between the gut microbiome and neurotransmitter systems highlight the importance of further investigating the potential crosstalk between the gut microbiome, neurotransmitters, and neuroinflammation.

Our results, and others, suggest that environmental influences affect the composition of the gut microbiome, which can complicate the results of experimental microbiome manipulations. The gut microbiome varies markedly in humans (Clemente et al., 2012; Falony et al., 2016; Ley et al., 2008; Ursell et al., 2012) and rodents (Stecher et al., 2010; Thiemann et al., 2017). Furthermore, different animal facilities can produce varied results with treatments to alter the microbiota (Eberl et al., 2020; Hildebrand et al., 2013; Macpherson & McCoy, 2015; Rausch et al., 2016; Sadler et al., 2017). Thus, making the reproducibility of experimental microbiome altering procedures challenging. Therefore, it is crucial to consider the standardization of confounding factors to minimize these effects (Mooser et al., 2018). Our protocol considers the source of the animals as a fundamental parameter to consider when designing EAE and microbiome studies. The results shown below highlight the source of animals and harbored microbiome as critical parameters to consider in EAE studies.

Critical Parameters:

Aging, sex, food and water sources, animal grouping, and other environmental factors should be carefully considered in microbiota studies, also in the context of EAE. It is also essential to consider that the disease itself modifies the microbiota composition (Colpitts et al., 2017). This observation is particularly relevant when animals present different severity scores and could potentially harbor microbiotas of distinct composition but share the same cage. Since it is difficult to evaluate all critical parameters at once, we focused on the impact of the source of mice (Envigo vs. Jackson Laboratory). After acclimating the animals to our facilities for seven days, Envigo and Jackson Laboratory mice were separated into groups. One cage of 5 Envigo mice and one cage of 5 Jackson Laboratory mice were maintained “naïve.” At the same time, EAE was induced in 9 and 10 Envigo and Jackson Laboratory mice, respectively, separated into 4 or 5 mice per cage. Daily EAE clinical scores and weekly body weights were monitored (Figs. 1–2), and stool samples from all mice were isolated weekly and stored at −80 °C. At the end of the experiment, DNA was extracted from samples obtained 0-, 14-, and 21-days post-induction (dpi). After 16S rRNA sequencing, the alpha diversity and beta diversity of the microbiota of Envigo and Jackson Laboratories mice and experimental groups (naïve and EAE) were compared (Figs. 3, 4, and 5).

Figure 1.

Graphical representation of the in vivo experimental design. Fourteen Envigo and 15 Jackson Laboratory C57BL/6 mice were used in the study. EAE was induced in 9 and 10 Envigo and Jackson Laboratory mice, respectively, while 5 mice from each cohort served as naïve controls. EAE was induced on day 0, and the disease was monitored daily until 21 days post-induction (21 dpi). Fresh stool samples from each mouse were collected at 0, 14, and 21 dpi. The microbiota composition (alpha diversity, beta diversity, and taxonomy) was compared within groups.

Figure 2.

The source of C57BL/6 mice affects disease severity in EAE mice (Envigo EAE, n = 9; Jackson EAE, n = 10; Envigo Naïve, n = 5; Jackson Naïve, n = 5). A) Clinical severity score from induction (0 dpi) to end of treatment (21 dpi). B) EAE onset. C) Clinical severity score at 21 dpi. D) EAE incidence. E) Percent survival. Data are presented as mean ± SEM. For EAE clinical scores and disease incidence, group differences were estimated using repeated measures and mixed-effect ANOVA, followed by Tukey’s multiple comparison post-hoc test. Group differences in disease onset and severity scores were evaluated using non-parametric Kruskal-Wallis followed by Dunn’s multiple comparisons tests.

Figure 3.

Alpha diversity of gut microbiota of Envigo and Jackson Laboratory mice, and relative abundances of phyla identified. Stool samples were harvested from mice on day 0 (immediately before EAE induction). DNA extracted from stool samples was analyzed by amplifying the V4 region of the 16S rRNA. Shannon index (A) and Chao1 (B) index were compared between samples from Envigo (n = 14) and Jackson Laboratory (n = 15) mice. The relative abundances at the phylum level are shown (C). Statistical analysis of the differences in the alpha diversity indexes was performed by one-way ANOVA, followed by Tukey’s multiple comparisons test. The statistical analysis of the differences in the relative abundances of phyla found in the study was performed with the Wilcoxon Rank Sum Test, by testing for differences in median relative abundances. Only taxa with differences with p < 0.05 are shown (Table 5).

Figure 4.

Alpha diversity of gut microbiota of Envigo and Jackson Laboratory mice. Stool samples were harvested from mice 0, 14, and 21 dpi. On day 21, the mice that survived were compared. Shannon index (A) and Chao1 index (B) were compared between all groups (Envigo EAE, n = 9; Jackson EAE, n = 10; Envigo Naïve, n = 5; Jackson Naïve, n = 5). Statistical analysis of the differences in the alpha diversity indexes was performed by one-way ANOVA, followed by Tukey’s multiple comparisons test. P-values are shown in tables 6 and 7.

Figure 5.

Beta diversity of gut microbiota of Envigo and Jackson Laboratory mice. Stool samples were harvested from mice 0, 14, and 21 dpi. On day 21, the mice that survived were compared. Bray-Curtis Dissimilarity index was compared between all groups (Envigo EAE, n = 9; Jackson EAE, n = 10; Envigo Naïve, n = 5; Jackson Naïve, n = 5). Statistical analysis was performed by permutational ANOVA by ADONIS, with 10,000 permutations. Average p-values are shown in table 8.

Troubleshooting:

A common problem in EAE studies is the difficulty of replicating the incidence and severity of the disease through subsequent experiments. The source and concentration of pertussis toxin are critical factors to consider in all EAE studies, as they have been sufficiently established (Aharoni et al., 2021; Stromnes & Goverman, 2006). We hypothesize that the source of animals considered the same strain is also a factor that can influence EAE induction and therefore should be controlled (Table 4). Possibly due to genetic differences of C57BL/6 mice, or the environmental conditions of the vendor’s facilities naïve mice from two different vendors show microbiotas with different alpha and beta diversity microbiota composition (Figs. 3–5). There are other factors subject to troubleshooting that are relevant to control when isolating stool samples for microbiome studies; however, our protocol does not focus on those factors, as they have been previously investigated (Bittner et al., 2014; Constantinescu, 2005).

Understanding the results:

The EAE model induced in mice is a widely used experimental approach based on quantifying clinical scores using variable scales, cumulative or severity indexes, the onset of disease (days), the incidence of the disease, and survival (mortality). Despite having the same C57BL/6 background, mice obtained from two different commercial vendors showed a dissimilar EAE profiles (Fig. 2). Overall, the clinical scores observed in Jackson Lab mice induced with the disease were higher than those in Envigo mice (Fig. 2A). Although the mixed-effect statistical analysis, including repeated measures, showed no statistical differences when considering scores only, combined time and score parameters showed a statistical difference among groups (Fig. 2A; p < 0.001). However, no statistical differences were observed when comparing individual time points between Envigo and Jackson Laboratory EAE mice. No significant differences were observed in the disease onset of Envigo vs. Jackson Laboratory EAE mice (Fig. 2B). However, when comparing the severity index, reflecting the sum of clinical scores for each mouse divided by the number of days the animals showed signs of the disease, a significant increase in Jackson Laboratory mice was observed compared to Envigo mice (Fig. 2C; p < 0.01). In terms of disease incidence, no significant differences were observed between Envigo and Jackson Laboratory mice, and in both cases, the percentage of EAE incidence was above 90 % (Fig. 2D). However, as denoted by an increased severity index, Jackson Laboratory mice had a more reduced survival rate by the end of the experiment (20 %) than Envigo mice (85 %). It is important to note that animals suffering scores above 3 for two consecutive days were euthanized and none of the animals that succumbed to the disease were found expired in their cages.

Overall, our EAE study indicates that although the same self-peptide and adjuvants were used for the induction of disease, we observed a significant variation in the disease parameters quantified. Since the only experimental variable was the source of the animals, and due to the reported differences in the microbiota carried by mice when arriving from separate facilities, we next compared the microbiota composition in naïve Envigo and Jackson Laboratory mice vs. EAE mice (14 and 21 dpi).

Alpha diversity is a measure of the richness or evenness of the composition of a population. Several different alpha diversity measurements are used in microbiome studies. We selected two different indexes that quantify alpha diversity accounting for richness and evenness: the Shannon Index and Chao1. Due to differences in sensitivity of the statistical methodology used in alpha diversity parameters, it is important to quantify it by at least two methods. The initial comparison of alpha diversity was made between all Envigo mice and Jackson Laboratory mice, after one week of acclimation to our animal facilities, immediately before EAE induction. As shown in figure 3, both Shannon and Chao1 indexes reported differences between Envigo and Jackson Laboratory mice, with a significantly increased alpha diversity in Envigo vs. Jackson Laboratory mice. Increased Shannon Index (p < 0.01) and Chao1 (p < 0.001) were observed in Envigo vs. Jackson Labs mice (Fig. 3). These results suggest that, under our same housing conditions, the diversity of gut microbiota directly depends on the source of the animals.

Taxonomically, the analysis done in samples obtained at 0 dpi confirms the significant differences in the microbiota composition of Envigo vs. Jackson Laboratory mice. Figure 3C shows marked differences in the relative abundances of phyla between Envigo and Jackson Laboratories mice. At the genus level, the analysis reports significant differences between mice from both vendors (Table 5, and supplemental figure 1). The gut microbiota of Jackson Laboratory mice showed increased relative abundances of Bacteroidetes, Verrucomicrobia, unclassified Eukaryota, and reduced Firmicutes, Proteobacteria, Patescibacteria, Deferribacteres, and Cyanobacteria phyla than gut microbiota of Envigo mice. The differences are also shown at the Class, Order, Family, and Genus levels. Interestingly, relative abundances of bacterial genera linked with differences in EAE severity were found to be statistically different in the gut microbiota of Jackson Lab mice vs. Envigo mice. For example, Akkermansia was increased in Jackson Lab mice vs. Envigo mice, while Prevotella was significantly reduced in Jackson Laboratory mice vs. Envigo mice (Table 5). The taxonomical analysis also revealed statistically significant differences between Envigo and Jackson Laboratory mice 14 and 21 dpi (Supplemental Table 2 and 3, respectively) and between EAE and naïve Envigo and Jackson Laboratory mice 0, 14 and 21 dpi (Supplemental Tables 4, 5 and 6).

Table 5.

Relative abundance of taxa in Jackson Laboratory mice versus Envigo at time of EAE induction (0 dpi)

Taxonomical Level	Taxon	p-value	Jackson relative to Envigo*
Phylum	Firmicutes	0.007937	down
	Bacteroidetes	0.007937	up
	Verrucomicrobia	0.03175	up
	Unclassified Eukaryota	0.03175	up
	Proteobacteria	0.007937	down
	Patescibacteria	0.007495	down
	Deferribacteres	0.007495	down
	Cyanobacteria	0.02537	down
Class	Bacteroidia	0.007937	up
	Verrucomicrobia	0.03175	up
	Erysipelotrichia	0.007937	up
	Unclassified Eukaryota	0.03175	up
	Saccharimondia	0.007495	down
	Deferribacteres	0.007495	down
	Deltaproteobacteria	0.007495	down
	Unclassified Firmicutes	0.007495	down
	Melainabacteria	0.02537	down
Order	Bacteroidales	0.007937	up
	Verrucomicrobiales	0.03175	up
	Erysipelotrichales	0.007937	up
	Unclassified Eukaryota	0.03175	up
	Saccharimondales	0.007495	down
	Anaeroplasmatales	0.01116	up
	Deferribacterales	0.007495	down
	Desulfovibrionales	0.007495	down
	Unclassified Firmicutes	0.007495	down
	Gastroanaerophilales	0.02537	down
	Rhodospirallales	0.025374	down
Family	Muribaculaceae	0.007937	up
	Akkermansiaceae	0.03175	up
	Erysipelotrichaceae	0.007937	up
	Prevotellaceae	0.007937	down
	Unclassified Eukaryota	0.03175	up
	Saccharimondaceae	0.007495	down
	Anaeroplasmataceae	0.01116	up
	Peptococcacea	0.03175	down
	Deferribacteraceae	0.007495	down
	Tannerellaceae	0.007495	down
	Enterococcaceae	0.007495	down
	Desulfovibrionaceae	0.007495	down
	Unclassified Bacteroidales	0.007495	down
	Clostridiaceae_1	0.007937	down
	Marinifilaceae	0.007495	down
	Christensenellaceae	0.01116	down
	Atopobiaceae	0.009701	up
	Streptococcaceae	0.007495	down
	Unclassified Firmicutes	0.007495	down
	Coriobacteriales_Incertae_Sedis	0.03114	down
	Unclassified Gastranaerophilales	0.02537	down
	Unclassified Rhodospirillales	0.02537	down
	Tyzzerella	0.007495	down
Genus	Unclassified Muribaculaceae	0.007937	up
	Akkermansia	0.03175	up
	Alistipes	0.007937	up
	Ruminococcacea_UCG_014	0.007937	down
	Rosburia	0.007397	down
	Muribaculum	0.007937	down
	Unclassified Prevotellaceae	0.007495	down
	Mucispirillum	0.007495	down
	Parabacteroides	0.007495	down
	Enterococcus	0.007495	down
	Unclassified Bacteroidales	0.007495	down
	Odoribacter	0.007495	down
	Desulfovibrio	0.007495	down
	Turcibacter	0.01116	up
	Rikenella	0.007495	down
	Candidatus_Arthromitus	0.007495	down
	Butyricicoccus	0.02001	down
	Bilophila	0.007495	down
	Ruminococcaceae_UCG_009	0.03445	down
	Tyzzerella	0.009701	down
	Coriobacteriaceae_UCG_009	0.009701	up
	Family_XIII_UCG_001	0.007495	down
	Christensenellaceae_R-7_group	0.007495	down
	Streptococcus	0.007495	down
	Ruminoclostridium_6	0.007495	down
	Unclassified Firmicutes	0.007495	down
	Faecalibacterium	0.02537	down
	Lachnospiraceae_NK4B4_group	0.02537	down
	Parvibacter	0.02537	down
	Pygmaiobacter	0.02537	down
	Ruminococcaceae_UCG_003	0.02537	down

^*Wilcoxon Rank Sum Test: Testing for Differences in Median Relative Abundance. Only taxa with differences with p < 0.05 are shown

We also evaluated whether the progression of the disease could affect alpha diversity in Envigo vs. Jackson Laboratory mice. When comparing naïve vs. EAE on day 0, no significant differences were observed between Envigo mice (Fig. 4A for Shannon index and Fig. 4B for Chao1). Similarly, no significant differences were observed in the Shannon index and Chao1 between naïve and EAE Jackson Laboratory mice (Fig. 4A and Fig. 4B). As indicated above, on day 0, alpha diversity of Envigo and Jackson Laboratory mice differed significantly by Chao 1 (Fig. 3B); although, separating the mice into naïve and EAE groups resulted in the loss of the statistical difference determined by Shannon index, indicating that Chao1 was more sensitive in accounting for differences than the Shannon index (Fig. 3).

As the disease progressed, the differences observed in alpha diversity between naïve and EAE groups were more evident in Jackson Laboratory mice than in Envigo mice. By Shannon index, no significant differences between Envigo naïve and EAE mice were detected 14 dpi and 21 dpi (Fig. 4A). By contrast, the Shannon index of Jackson Laboratory naïve and EAE mice was significantly different 14 dpi and 21 dpi (Fig. 4A). A similar pattern was observed when comparing Chao1 (Fig. 4B). These results suggest that the effects of EAE in alpha diversity are more pronounced in Jackson Laboratory mice than in Envigo mice. It is worth noting that disease severity was also found to increase in Jackson Laboratory mice compared to Envigo mice (Fig. 2). The analysis of the microbial diversity as quantified by richness and evenness suggest that the parameter is affected by the source of the animals used and that the differences can be aggravated with disease progression.

To further explore the differences in the microbiotas of Envigo and Jackson Laboratory mice before and during the EAE induction, we compared the beta diversity of the mice used in our study. Interestingly, the differences in the microbiota composition not only appeared as alpha diversity changes, but the overall composition of the microbiotas at the genus level was also different in the context of EAE disease (Fig. 5). Beta diversity, a parameter that determines the level of change in the composition of the populations between distinct samples, was significantly different at days 0-, 14-, and 21-dpi when Envigo and Jackson Laboratory EAE mice were compared (Fig. 5). The statistical analysis showed a lower p-value (p < 0.001) at day 14 when the progression of the disease was more pronounced. When comparing the beta diversity of Envigo vs. Jackson Laboratory mice prior to EAE induction (0 dpi), the difference was already statistically significant (p < 0.001; data not shown).

In summary, the results obtained when applying the protocol detailed here suggest the critical importance of the microbiome harbored by mice used in EAE studies and the need to control housing conditions that affect the microbiome when addressing the severity of disease.

Time considerations:

The duration of the EAE experiment directly depends on the model of the specific disease selected. For the MOG_35-55 model in C57BL/6 mice and the microbiome analysis, the duration also depends on the experimental question to be addressed. In our study, the in vivo section was performed in four weeks. The DNA extraction and sequencing were done in three weeks, while the analysis and figures generation required two more weeks. In total, the experiment was finalized in 9 weeks from the time of arrival of the animals.

Table 6.

p-values^* obtained from Shannon index comparisons shown in figure 4A

	0 dpi	14 dpi	21 dpi
Overall	0.3117	3.59 x 10−7	0.02658
Envigo vs Jackson (Naïve)	0.1019	1.80 x 10−5	0.05268
Envigo vs. Jackson (EAE)	0.5591	0.1999	0.5016
Naïve vs EAE (Envigo)	0.5201	0.05045	0.5648
Naïve vs. EAE (Jackson)	0.3416	3.59 x 10−5	0.005159

^*,p-values were obtained by one-way ANOVA, followed by Tukey’s multiple comparisons test.

Table 7.

p-values^* obtained from Shannon index comparisons shown in figure 4B

	0 dpi	14 dpi	21 dpi
Overall	3.84 x 10−11	1.42 x 10−10	5.12 x 10−5
Envigo vs Jackson (Naïve)	8.2 x 10−7	3.24 x 10−6	2.34 x 10−5
Envigo vs Jackson (EAE)	7.74 x 10−8	3.93 x 10−5	0.2769
Naïve vs EAE (Envigo)	0.8757	0.6291	0.6661
Naïve vs. EAE (Jackson)	0.4889	1.96 x 10−5	0.00224

^*,p-values were obtained by one-way ANOVA, followed by Tukey’s multiple comparisons test.

Table 8.

ADONIS analysis of beta diversity comparisons between Jackson and Envigo shown in figure 5.

	0 dpi	14 dpi	21 dpi
Jackson Laboratory vs. Envigo	p ≤ 0.01	p ≤ 0.001	p ≤ 0.05

Permutational ANOVA by ADONIS: 10,000 permutations performed. Average p-values reported.

DATA AVAILABILITY STATEMENT:

The data, tools and material (or their source) that support the protocol are available from the corresponding author upon reasonable request.