Fifteen healthy individuals, 12 males and 3 females, who have been following the MPD for at least one year were recruited from different urban areas across Italy. The average age of the enrolled subjects was 39.2 years (range, 26–57), and the average Body Mass Index (BMI) 22.1 kg/m² (range, 19.4–25.7) (S1 Table).

The MPD adopted by the 15 subjects is mainly based on the consumption of unprocessed foods, with high intake of vegetables, fruit, nuts and seeds, eggs, fish and lean meat, while excluding grains, dairy products, salt and refined sugar. The daily total calorie intake, as well as that of macro- and micro-nutrients, assessed through 7-day weighted food intake records (7D-WRs), are reported in S2 Table. The average daily energy intake of the enrolled cohort is 1,843.45 kcal (range, 1,563–2,186 kcal). The percentage of macronutrients is distributed as follows: fat, 51.02%; protein, 30.14%; carbohydrate, 18.84% (Fig 1A). With regard to lipids, 51.65% of total calories are from monounsaturated fatty acids (MUFAs), 30.93% from saturated fatty acids (SFAs) and 17.42% from polyunsaturated fatty acids (PUFAs) (Fig 1B). The average daily fiber intake is 14.64 g/1,000 kcal.

Based on the data collected through a questionnaire on the socio-economic status, one third of the subjects lived in highly urbanized areas, more than half in semi-urbanized areas (8/15) and only one individual in a rural setting. Two thirds lived in apartments and the remainder in independent houses. Eight out of 15 subjects declared they had pets and daily contact with nature (defined as 2 to 15 hours a week spent in a green area). According to a questionnaire on physical activity (the Global Physical Activity Questionnaire—GPAQ), 12 individuals reported practicing moderate to intense fitness activities for an average of 1 hour a day for at least 3 days a week.

The GM structure of MPD Italian adults was profiled through 16S rRNA gene sequencing of fecal DNA. A total of 864,439 high-quality reads (mean ± sd, 57,629 ± 19,752; range, 25,142–95,924) were generated and clustered in 7,483 OTUs. The phyla Firmicutes (relative abundance, mean ± sem, 65.1 ± 2.1%) and Bacteroidetes (24.6 ± 2.2%) dominate the gut microbial ecosystem, with Proteobacteria (4.4 ± 1.6%), Actinobacteria (3.4 ± 0.8%) and Verrucomicrobia (1.2 ± 0.5%) as minor components. At family level, Ruminococcaceae (26.7 ± 1.7%), Lachnospiraceae (18.7 ± 1.4%), Bacteroidaceae (13.7 ± 1.8%) and Prevotellaceae (7.4 ± 2.4%) are the dominant GM constituents. The most abundant (≥ 5%) bacterial genera are Bacteroides, Prevotella, and Faecalibacterium, while Coprococcus, Ruminococcus, Blautia, Lachnospira, Phascolarctobacterium, Streptococcus, Roseburia, Akkermansia, Oscillospira and [Eubacterium] represent minor components of the microbial ecosystem (range, 4.4 ± 0.7% - 1.0 ± 0.4%) (Fig 2; S1 Fig).

In order to investigate whether the adherence to the MPD is sufficient to promote a more diverse GM ecosystem—even in a Western urban context—we compared the GM diversity between the 15 MPD subjects and 143 urban Italians with different level of adherence to the MD, whose GM composition was described in De Filippis et al. (n = 127) [23] and Schnorr et al. (n = 16) [5]. Moreover, to extend the comparative analysis to a global level, the GM structural profiles of the following traditional hunter-gatherer populations were included: 27 Hadza from Tanzania [5], 25 Matses from Peru [6], and 21 Inuit from Canada [24]. According to our findings, significant differences in the GM biodiversity among the study groups were detected (Simpson index, P-value = 2.6 × 10⁻¹⁵; Shannon index, P-value = 2.2 × 10⁻¹⁶; Kruskal-Wallis test) (Fig 3). Interestingly, the GM diversity observed for MPD subjects far exceeds that of urban Italians adhering to the MD (Simpson index, P-value = 2.5 × 10⁻⁷; Shannon index, P-value = 6.1 × 10⁻⁹; Wilcoxon test), is comparable to that of the Hadza (P-value = 0.39; 0.26), and even greater than Matses (P-value = 0.0082; 0.0039) and Inuit (P-value = 0.00075; 0.0027).

The PCoA based on Bray-Curtis distances was next used to assess overall genus-level compositional differences in the GM structure between study groups. Our data show clear separation of GM profiles by provenance and, within the Italian cohort, by dietary pattern (adonis: P-value < 1 × 10⁻⁵, R² = 0.27; ANOSIM: P-value < 1 × 10⁻⁵, R = 0.48) (Fig 4A; S3 Table). Interestingly, MPD subjects show a low level of interpersonal GM variation (Bray-Curtis distances, mean ± sd, 0.42 ± 0.095), approximating that observed for the Hadza (0.36 ± 0.092) (Fig 4B). In order to identify the bacterial drivers with a statistically significant contribution (permutation correlation test, P-value < 0.001) to the sample ordination, we superimposed the genus relative abundance on the PCoA plot (S2 Fig). According to our data, the microorganisms characterizing the Italian cohort are Bacteroides, Collinsella, Coprococcus and Blautia. The genera Clostridium, Prevotella, [Prevotella], Catenibacterium and Oscillospira were found to be associated with Hadza and Matses, while Sutterella and Parabacteroides with Inuit.

Despite the small sample size, it is worth noting that the MPD microbiome shows several compositional differences with respect to the other cohorts, which well match the peculiar macronutrient intake (Fig 2; S4 Table). In particular, compared with all other populations, except for the Inuit (as expected based on available dietary information), the MPD fecal profiles are enriched in asaccharolytic genera, such as Sutterella and Odoribacter [25], in Bilophila, microorganism typically associated with animal protein and saturated fat consumption [26,27], as well as in Akkermansia, known to be associated with the consumption of unsaturated fat [28] (P-value ≤ 0.021; Wilcoxon test). Although Akkermansia has recently been identified as potential next-generation probiotics, its role in inflammatory contexts is still controversial and requires further investigation [28–31]. Moreover, when compared to hunter-gatherer populations (whose subsistence, at least during sampling, was mainly based on tubers and other plant foods), the microbiome of MPD subjects shows increased relative abundance of the bile-tolerant Bacteroides, Collinsella and Dorea (P-value ≤ 0.003). Bacteroides is indeed typically associated with Western-type animal-based diets [26], the genus Collinsella is known to comprise bacterial species capable of deconjugating bile acids and positively correlated with plasma cholesterol levels [32], and Dorea has recently been suggested to be involved in the production of the secondary bile acid, deoxycholic acid [33]. It should be remembered that secondary bile acids are generally associated with increased risk of non-infectious bowel disease and colorectal cancer [34], which stresses the need to be cautious in adhering to this dietary pattern in the long term. On the other hand, it should be noted that, compared to traditional populations, MPD profiles show greater proportions of the SCFA producers Lachnospira and Coprococcus (P-value ≤ 0.008).

We also evaluated the Prevotella ratio, i.e. the ratio of Prevotella to the sum of Prevotella and Bacteroides [35] (S3 Fig). These genera are indeed recognized as biomarkers of diet and lifestyle, with Bacteroides typically associated with high-protein high-fat Western diets and Prevotella with carbohydrate/fiber-based diets typical of more agrarian societies [35,36]. Although no detailed dietary information is available for traditional populations, Hadza and Matses diets are known to be heavily based on the consumption of highly fibrous tubers and vegetal foods [5,6]. On the other hand, the fiber intake of MPD individuals (29.47 ± 20.49 g/day) does not exceed by far that reported for urban MD Italians (range, 10.37–21.02 g/day; [23]). Consistent with this, a significantly lower Prevotella ratio was observed for MPD individuals as well as for other urban Italians compared to Hadza and Matses (P-value < 6.6 x 10⁻⁷).

Fifteen healthy individuals following a MPD for at least one year were recruited from different urban areas across Italy (Lombardia, Piemonte, Emilia-Romagna, Toscana, Umbria, Lazio, Campania, Molise, Puglia and Calabria regions). Exclusion criteria included: age below 18 and over 65 years; BMI <18.5 and >24.9 kg/m²; habitual intake of drugs and nutritional and pharmacological supplements of pre- and probiotics; taking antibiotics in the last three months; presence of intestinal and metabolic disorders (i.e. inflammatory bowel disease, bacterial contamination syndrome, irritable bowel syndrome, constipation, celiac disease, type 1 and 2 diabetes, cardio- and neurovascular diseases, rheumatoid arthritis, allergies, neurodegenerative diseases, cancer). Written informed consent was obtained from all volunteers. All work was approved by the Ethics Committee of the Sant’Orsola-Malpighi Hospital, University of Bologna (ref. number, 118/2015/U/Tess).

Each subject was asked to fill in a 7-day weighted food intake record (7D-WR), with the total food and beverage consumption (including methods of preparation whenever possible) for 7 days representing their usual intake, as previously described [42]. Daily total calorie intake as well as that of macro- and micro-nutrients were assessed through the MètaDieta software version 3.7 (METEDA). The participants were also asked to fill in two questionnaires, one regarding their socio-economic status (according to the guidelines of the Health Survey for England– 2013, http://www.hscic.gov.uk/catalogue/PUB16076) and the other on physical activity (based on the Global Physical Activity Questionnaire–GPAQ–developed by World Health Organization, http://www.who.int/chp/steps/resources/GPAQ_Analysis_Guide.pdf). A single fecal sample was self-collected by each participant after completing the 7D-WR (i.e. on day 7) and immediately frozen at -20°C. All specimens were delivered to the laboratory of the Microbial Ecology of Health Unit (Dept. Pharmacy and Biotechnology, University of Bologna, Bologna, Italy) where they were stored at -80°C until processing. Data and fecal samples were collected between March and April 2017.

Total bacterial DNA was extracted from each stool sample using the DNeasy Blood and Tissue kit (QIAGEN) with the modifications previously described by Biagi et al. [43]. In brief, 250 mg of fecal samples were suspended in 1 ml of lysis buffer (500 mM NaCl, 50 mM Tris-HCl pH 8, 50 mM EDTA, 4% (w/v) SDS), added with four 3-mm glass beads and 0.5 g of 0.1-mm zirconia beads (BioSpec Products) and homogenized using a FastPrep instrument (MP Biomedicals) with three bead-beating steps at 5.5 movements/sec for 1 min, and 5-min incubation in ice between treatments. After incubation at 95°C for 15 min, stool particles were pelleted by centrifugation at 14,000 rpm for 5 min. Nucleic acids were precipitated by adding 260 μl of 10 M ammonium acetate and one volume of isopropanol. The pellets were then washed with 70% ethanol and suspended in TE buffer. RNA was removed by treatment with 2 μl of DNase-free RNase (10 mg/ml) at 37°C for 15 min. Protein removal and column-based DNA purification were performed following the manufacturer’s instructions (QIAGEN). DNA was quantified with the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies).

For each sample, the V3-V4 region of the 16S rRNA gene was amplified using the S-D-Bact-0341-b-S-17/S-D-Bact-0785-a-A-21 primers [44] with Illumina overhang adapter sequences. PCR reactions were performed in a final volume of 25 μl, containing 12.5 ng of genomic DNA, 200 nM of each primer, and 2X KAPA HiFi HotStart ReadyMix (Kapa Biosystems, Roche), in a Thermal Cycler T (Biometra GmbH) with the following gradient: 3 min at 95°C for the initial denaturation, 25 cycles of denaturation at 95°C for 30 sec, annealing at 55°C for 30 sec and extension at 72°C for 30 sec, and a final extension step at 72°C for 5 min. PCR products of around 460 bp were purified using a magnetic bead-based system (Agencourt AMPure XP; Beckman Coulter) and sequenced on Illumina MiSeq platform with the 2 × 250 bp paired-end protocol, according to the manufacturer’s guidelines (Illumina). Briefly, each indexed library was prepared by limited-cycle PCR using Nextera technology, and further purified as described above. The libraries were subsequently pooled at equimolar concentrations, denatured with 0.2 N NaOH, and diluted to 6 pM before loading onto the MiSeq flow cell. Sequencing reads were deposited in MG-RAST (project ID, mgp89161).

Raw sequences were processed using a pipeline that combines PANDAseq [45] and QIIME [46]. The UCLUST software [47] was used to bin high-quality reads into operational taxonomic units (OTUs) at 0.97 similarity threshold through an open-reference strategy. Taxonomy was assigned through the RDP classifier, using the Greengenes database as a reference (release May 2013). Chimera filtering was performed by using ChimeraSlayer [48]. All singleton OTUs were discarded.

16S rRNA gene sequencing data of our cohort were compared with publicly available data from the following previous studies: De Filippis et al. [23] (127 Italians; NCBI Sequence Read Archive (SRA) accession number: SRP042234), Schnorr et al. [5] (16 Italians and 27 Hadza hunter-gatherers from Tanzania; MG-RAST ID: 7058), Obregon-Tito et al. [6] (25 Matses hunter-gatherers from Peru; NCBI SRA: PRJNA268964), and Girard et al. [24] (21 Inuit from the Canadian Arctic; Qiita ID: 10439). Genus-level community composition was generated for all cohorts combined. Alpha diversity was assessed using the Shannon and Simpson indices. Beta diversity was evaluated using the Bray-Curtis dissimilarity measure. All statistical analysis was performed in R 3.3.2, using R Studio 1.0.44 and the libraries vegan, made4 and stats. The significance of data separation in the Principal Coordinates Analysis (PCoA) of Bray-Curtis distances was tested using a permutation test with pseudo-F ratios (function adonis of vegan package) and ANOSIM test. Superimposition of bacterial genera on the PCoA plot was performed using the envfit function of vegan. Wilcoxon test was used to assess significant differences between groups (for intra- and inter-individual diversity), while Kruskal–Wallis test was used for multiple comparisons. P-values were corrected for false discovery rate (FDR, Benjamini-Hochberg) and P-values ≤ 0.05 were considered statistically significant.