A temporary cholesterol-rich diet and bacterial extracellular matrix factors favor Salmonella spp. biofilm formation in the cecum

Asymptomatic chronic carriers occur in approximately 5% of humans infected with Salmonella enterica serovar Typhi (S. Typhi) and represent a critical reservoir for bacterial dissemination. While chronic carriage primarily occurs in the gallbladder through biofilms on gallstones, additional anatomic sites have been suggested that could also harbor Salmonella. S. Typhimurium, orally-infected 129X1/SvJ mice were pre-treated with a cholesterol-rich diet as a gallstone model for chronic carriage. We observed S. Typhimurium in feces and the cecum during early and persistent infection. Furthermore, bacterial biofilm-like aggregates were associated with the cecum epithelium at 7- and 21-day post-infection (DPI) in mice on a lithogenic diet (Ld) and correlated with an increase in cecal cholesterol at 21 DPI. Salmonella’s extracellular matrix (ECM) was demonstrated as important in colonizing the cecum, as survival and aggregate formation significantly decreased when mice were infected with a quadruple ECM mutant strain. Gallbladder Salmonella counts were low at 36 DPI while cecal Salmonella were high, suggesting that gallbladder colonization was likely not responsible for the high cecal burden. All cecum phenotypes were significantly diminished in mice fed a normal diet (Nd). Finally, we examined the capability of S. Typhi to colonize the cecum and showed S. Typhi in feces and in aggregates in the cecum up to 7 DPI, with slightly higher counts in mice fed a Ld compared to Nd. Our findings suggest that the cecum, particularly under cholesterol-rich conditions, serves as an adaptive niche for Salmonella spp. aggregates/biofilms and is a putative site for long-term infection

Methods
Ethics statement. The mice used in this study were housed in strict adherence to the guidelines established by the Abigail Wexner Research Institute (AWRI) Institutional Animal Care and Use Committee (IACUC). All experimental procedures involving mice were conducted with the appropriate authorization obtained through AWRI IACUC protocol AR18-00080, which complied with the statutory requirements outlined in the Guide for the Care and Use of Laboratory Animals.

Bacterial Strains. Previously, we described the constructions of Salmonella mutant strains deficient in biofilm formation capacity (35). The following bacterial strains were utilized in this study: S. Tm 14028 (S. Tmwt; JSG210), S. Tm curli mutant (S. TmΔcsgA; JSG3540), S. Tm colanic acid mutant (S. TmΔwcaM; JSG3712), S. Tm O antigen capsule mutant (S. TmΔyihO; JSG3672), S. Tm cellulose mutant (S. TmΔbcsE; JSG3838), the doble ECM mutant (S. TmΔcsAΔbcsE: JSG3977), S. Tm quadruple ECM mutant ΔwcaM, ΔcsgA, ΔyihO, ΔbcsE (S. Tm ECMmut; JSG3841), S. Tm promoter vector for curli (S. Tmlux:csgDEFG; JSG3977), and S. Typhi rpoS+ Ty2 (S. Typhiwt; JSG698). Planktonic cultures were cultivated for 16 hours at 37°C in tryptic soy broth (TSB). All strains were regrown to an OD600 of 0.6 prior to infection. Mice were infected intragastrically with a dose of 1x107 for S. Tm or 5x107 for S. Typhi.

In vitro bacterial growth curves. Overnight (ON) cultures of Salmonella WT and mutant strains were normalized to an OD600 of 1. The cultures were then diluted 1:1000 in fresh TSB medium and incubated at 37°C. Bacterial growth was monitored over a 15-hour period, with OD600 readings taken every 15 min.

Luciferase assay: O/N cultures of S. Tmlux:csgDEFG were normalized to an OD600 of 1. The cultures were subsequently diluted 1:1000 in fresh Luria-Bertani (LB) medium and incubated at either 26°C or 37°C in microtiter plates in the presence or absence of cholesterol (well coating). The strain was grown over a 15-hour period, with luminescence (OD490) readings taken every 15 min.

Bacterial shedding and fecal collection. Fresh fecal samples were obtained by individually placing mice in isolation containers until approximately 100mg of feces were collected. The collected fecal samples were homogenized in 1 mL of sterile PBS and plated on selective agar XLD (refer to the bacterial isolation section), followed by incubation for 24h (S. Tm) or 48h (S. Typhi) at 37°C.

Clinical assessment of mice: Mice were evaluated at 7 and 21 DPI to assess phenotypic signs of the severity of the infection. The evaluation included measurements of body weight, assessment of general health (fur, eyes, posture, clinical complications), and observations of motility.

Mouse infection model and tissue CFUs. Six-week-old female 129X1/SvJ mice (The Jackson Laboratory, ME) were utilized in this study. Prior to infection, the mice were fed either a normal diet or a lithogenic diet for 6 weeks. The LD consisted of mouse chow supplemented with 0.5% cholesterol and 0.5% cholic acid (Envigo, TD 140673). Two weeks after the diet treatment, the mice were intragastrically infected with S. Tm or S. Typhi in 200µL of PBS containing the bacteria.

CFUs from feces and tissues: Bacterial isolation. Mice were euthanized at 7 and 21 DPI. The liver, spleen, mesenteric lymph nodes, small intestine, large intestine, and cecum were isolated. Approximately 150mg of tissues or feces were homogenized in 1 mL of sterile 1X PBS. The samples were briefly centrifuged for 30s, and the supernatant was filtered through a 15 mm strainer. The filtrate was transferred to a U-bottom 96-well plate and subjected to centrifugation at 1500 rpm for 8 min at 5°C. The supernatants were transferred to new wells (x3), followed by centrifugation at 3700 rpm (8 min, 5°C) to obtain a bacterial pellet. The pellet was resuspended and plated on selective XLD agar to determine the bacterial load of each organ (Fig. S3, a). The remaining tissues were preserved for cholesterol analysis, histology, or flow cytometry.

Bacterial flow cytometry: Following the isolation of bacteria (as described above), the pellet was blocked with 5% BSA for 10 min and washed with 3% BSA in PBS at 3700 rpm (8 min, 5°C). Primary antibodies (5 μg ml-1) and secondary (2.5 μg ml-1) antibodies were incubated with the pellet at 37°C for 20 min. Each antibody incubation was followed by three washes. The pellets were resuspended in 300µL of PBS, and the samples were analyzed using flow cytometry (BD LSRII), and the data were analyzed in FlowJo v10. The following antibodies were used: goat ⍺-S. Tm/Typhi CSA-1 (BacTrace, 5310-0322), mouse IgG2a ⍺-S Typhi LPS (Meridian Bio, C01362M), rabbit IgG ⍺-S. Tm (Thermo, PA-17244), ⍺-mouse IgG2a FITC (Invitrogen, 11421082) and ⍺-goat AF647 (Invitrogen, A21469), and DAPI (Invitrogen, D1306).

Cholesterol quantification: After the initial centrifugation of bacteria at 3700 rpm, 50 µL of supernatant from homogenized cecum was collected, and the cholesterol concentration was measured using the Amplex™ Red Cholesterol Assay Kit following the manufacturer’s instructions.

Histology: All tissues were fixed in 10% neutral-buffered formalin (Fisher Scientific, MA) for 24 hours. The morphology core at Abigail Wexner Research Institute embedded the tissues in paraffin. Deparaffinization and rehydration of samples were carried out before staining. Permeabilization was performed by incubating the samples in PBS containing 1% Triton X-100 for 1h at 37°C. Histological sections of 4μm were used for H&E staining, IHC, or IF microscopy. For IHC and IF, the primary antibody (Ab) incubation was performed by overnight incubation at 4°C in antibody dilution buffer (3% BSA and 0.35% Triton X-100 in PBS) with 0.5-1 μg ml-1 for each Ab. The antibodies used included a goat IgG ⍺-S. Tm/Typhi CSA-1, mouse IgG2a ⍺-S Typhi LPS, rabbit IgG ⍺-S. Tm, rabbit ⍺-Villin 1 (Abclonal, A11650), rabbit IgG ⍺-Vi Antigen (Abcam, ab79002), Human ⍺-amyloid APP 3H3 (Creative Bio, TAB-0801CLV), and pAb rabbit ⍺- S. Tm CsgA (gift from Dr. Cagla Tükel). For IHC, after primary Ab incubation, the VECTA Elite ABC Kit-HRP (Vector Lab, PK-6101 and PK6105), UltraTek Bio-HRP ⍺-goat IgG (ScyTek Labs, ABL/AGL015) were used, and staining was revealed using the metal enhancer DAB Kit (Thermo, PI34065). The slides were then counterstained with hematoxylin for 30 seconds, and washed in dH2O for 1 min. Finally, coverslips were applied using ProLong mounting media (Thermo, P36930).