HT29, CaCo-2 and HeLa cells were purchased from ATCC (US). HT-29 and Caco-2 cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM), supplemented with GlutaMAX, 4.5 g/L D-Glucose, 1% Pyruvate, 10% Fetal Bovine Serum (FBS), 1% MEM Non-Essential Ammino Acids (NEAA), and in a humidified 5% CO₂ atmosphere at 37°C. HeLa cells were grown in Eagle’s Minimal Essential Medium (EMEM), supplemented with 10% FBS. All cells were maintained at a concentration of 1x10⁶ cells/mL and passed when 80-90% confluency was reached. All culture media and supplements were obtained from Gibco (Thermo Fisher Scientific, US).

S. sonnei 53G strains used in this study were grown in Tryptic Soy Broth (TSB) liquid medium or on TSB-agar plates for determination of colony forming units (CFU) at 37°C without CO₂. Bacteria were first inoculated in 5 ml TSB with the appropriate antibiotics and incubated with shaking at 150 rpm overnight. On the following day, cultures were expanded to 40 ml TSB supplemented with 0.5% bile salts (Sigma Aldrich, Germany) to stimulate the expression of the T3SS (1) and incubated with shaking until bacteria reached the exponential phase of growth (OD₆₀₀ 0.5), which was used to perform all following experiments.

Briefly, mutants were generated in S. sonnei 53G as follows: stabilization of the virulence plasmid pINV was achieved by insertion of an antibiotic resistance cassette either into the wbg cluster (26), responsible for the synthesis of the O-Antigen (OAg- strain), or between aqpZ and RS31270 genes (OAg+ strain). Additionally, genes encoding for key components of the T3SS, namely IpaB, IpaD, Spa33, and for the adhesins VirG and IcsP, were deleted by insertion of a kanamycin resistance cassette in the corresponding genes in the OAg- background. The resistance cassette replacement construct was amplified from the pKD4 vector using forward and reverse primers composed of 50 bp homologous to the flanking regions of the gene to be deleted and approximately 20 bp ( Supplementary Table 1 ) at the 3′ end, matching the flanking region of the resistance gene. PCR products were purified and used to transform recombination-prone S. sonnei recipient cells carrying the temperature-sensitive pAJD434 plasmid. Subsequently, mutants were cured of pAJD434 plasmid by culturing them at 40-42°C. Finally, a strain that spontaneously lost pINV (pINV- strain) was propagated after selection on Congo Red, which is an artificial inducer of Shigella T3SS (27). This strain appeared as a white colony on Congo Red agar plates, indicating loss of the virulence plasmid pINV that also encodes the serotype-specific OAg biosynthesis locus in S. sonnei (28). Table 1 lists strains and antibiotics used in this study.

To visualize and analyze bacteria in vAIA experiments S. sonnei 53G strains, described in Table 1 , were transformed with the integrative plasmid pSEVA23a1 (https://seva-plasmids.com/) containing the gene encoding the fluorescent protein (FP) super folder (sf) mCherry in double copy and in tandem to achieve greater and more consistent expression levels (29). Two tandem copies of the FP were synthesized by the GeneArt company (Thermo Fisher Scientific, US) and provided in the pMK vector. The first copy was codon-optimized for K. pneumoniae (https://www.ebi.ac.uk/Tools/st/emboss_backtranseq/), while the second one was specific for E. coli to avoid recombination. The synthetic FP gene was subcloned into the HindIII and SpeI restriction sites of the plasmid Multiple Cloning Site (MCS) under control of the pTAC constitutive promoter for efficient expression in Gram-negative bacteria (30). Additionally, a hygromycin resistance cassette was inserted between BamHI and XbaI restriction enzyme sites within the MCS. The presence of the phage phi-C31 integrase and the attP site on the plasmid vector facilitated the integration of the plasmids into the bacterial genome via site-specific recombination with the attB locus on the chromosome (31).

To generate fluorescent bacterial strains, the following workflow was used. First, bacteria were grown until exponential phase. Cells were then centrifuged for 10 min at 5,000 x g and the bacterial pellets were washed 3 times in 20 ml ice-cold 10% glycerol. Finally, electrocompetent cells were resuspended in 1 ml 10% glycerol and 50 μL aliquots were made. 500 ng of plasmid DNA were added to 50 μl of bacteria and transferred to a cold electroporation cuvette. Electroporation was performed by using the Eporator (Eppendorf, Germany) at a voltage of 2500 V and a pulse time of 5-6 ms. Recovery of electroporated bacteria was obtained by adding 500 μl of fresh TSB to the bacterial suspension and incubating at 37°C for 1 h with shaking, before plating onto TSB-agar plates. Characterization of the newly generated fluorescent strains is detailed in Supplementary Figure 1 .

Comparison between conventional AIA and vAIA is shown in Table 2 . The procedure was optimized using S. sonnei 53G OAg- with different cell lines, Multiplicity Of Infections (MOIs) and incubation times. Three different epithelial cell lines, namely CaCo-2, HT-29 and HeLa, were seeded at different concentrations, from 5,000 to 100,000 cells/100 µl, in each well of a flat-bottom 96-well plastic plate (EuroClone, Italy) for conventional AIA, or a 96-well cyclic olefin poly-D-lysine plate (Perkin Elmer, US) for vAIA. Plates were incubated for 2 days at 37°C with 5% CO₂ to form a monolayer. Cells were then detached using trypsin and counted. Shigella strains were grown until exponential phase, centrifuged at 5,000 x g for 10 min and resuspended in DMEM without FBS. Eleven Multiplicity Of Infection ratios (MOIs: MOI5, MOI10, MOI20, MOI30, MOI40, MOI50, MOI60, MOI70, MOI80, MOI90, MOI100) and 6 incubation times (15 min, 30 min, 60 min, 90 min, 120 min, 150 min) were tested ( Supplementary Figure 2 ). Fifty microliters of the bacterial suspension, previously incubated for 30 min at 37°C with serial dilutions of the mAbs to be tested, along with appropriate positive (the pooled human serum) and negative (an unrelated mAb against SARS-CoV-2) controls, were added in each well of the plate containing the cells. In parallel, a condition in which the bacteria-mAbs mixture was added to the cells without prior incubation of 30 min at 37°C was also tested ( Supplementary Figure 3 ). The plate was then covered with a porous film, centrifuged at 200 x g for 1 min and incubated at 37°C with 5% CO₂ without shaking for variable amounts of time to allow infection to occur. After infection the bacteria in excess were washed away using DMEM without FBS. From this point on, plates were treated differently.

In conventional AIA, cells were permeabilized with 1% Triton X-100 (PanReac, AppliChem, US) for 10 min at 37°C and lysates containing both “adherent” and “internal” bacteria were serially diluted, plated as 10 µl spots on TSB-agar plates and incubated at 37°C overnight for CFU determination.

In vAIA, samples were fixed in 4% paraformaldehyde (PFA - Sigma Aldrich, Germany) in PBS 1X for 15 min at room temperature (RT). PFA was then removed, followed by one wash with PBS 1X. Next, plates were incubated for 30 min at RT in 1:500 diluted commercial antiserum (Denka-Seiken, Japan) specific for OAg- Shigella strain. One wash in PBS 1X and an additional incubation for 30 min at RT with Alexa Fluor 488 anti-rabbit antibody diluted 1:2,000 in PBS 1X followed. After washing twice with PBS 1X, Cell Mask Deep Red and 4′,6-diamidino-2-phenylindole (DAPI - Thermo Fisher Scientific, US), both diluted 1:2,000, were added for 10 min at RT to stain the cell membrane and the nucleic acids, respectively. Samples were washed and kept in PBS 1X at 4°C overnight before image acquisition. Each sample was tested in triplicate and in at least 2 independent experiments.

Image acquisition was performed by means of the high-content and high-throughput Opera Phenix confocal microscope (Perkin Elmer, US) using the 63X magnification objective, numerical aperture 1.15. At least 21 fields of view and a zoom stack of 5 vertical layers with a distance of 0.5 µm were acquired in each well. The resulting images were analyzed by using the Perkin Helmer Harmony software v. 4.9. Image analysis was performed on the reconstructed maximum projection of the zoom stack and the data represented the average of all fields of view analyzed in each well.

To confirm binding of mAbs to bacteria at single-cell level in the presence of cells, a vBA was performed as follows. First, cells and bacteria were prepared as described in the previous sections. Cells were then infected with 50 µl of the bacterial suspension at MOI100 and the plate was covered with a porous film, centrifuged at 200 x g for 1 min and incubated at 37°C with 5% CO_2, without shaking, for 1.5 h. After infection, cells were washed in DMEM without FBS and then fixed in 4% PFA for 15 min at RT. PFA was then removed and cells were further washed in PBS 1X without touching the well bottom. Mild permeabilization of the bacterial membrane in 0.1% of Triton X-100 for 10 min at RT to enhance the accessibility of mAbs to the bacterial antigens was performed (32). After an additional wash in PBS 1X, the mAbs to be tested and appropriate positive and negative controls were added in serial dilutions and the plate was incubated for 30 min at RT. Samples were then washed once with PBS 1X and anti-human or anti-mouse antibodies conjugated to Alexa Fluor 488 (Thermo Fisher Scientific, US) were added at a 1:2,000 dilution for 30 min at RT, followed by two washes with PBS 1X. Finally, the plate was incubated for 10 min at RT in a freshly made solution containing Cell Mask Deep Red and DAPI, both diluted 1:2,000 in PBS 1X. Samples were washed and kept in PBS 1X at 4°C overnight and images were acquired and processed as described in the section above.

All graphs and statistics were generated using GraphPad Prism v. 8.4.3. One-way ANOVA and the unpaired t test were used to compare the infection rates of Shigella mutants to those of the parental strain and the effect of the pooled human serum and the anti-IpaD C12-5 mAb to the no-treatment condition. The Pearson r correlation test was used to correlate the results obtained from conventional AIA to the ones of vAIA. All conditions were run in duplicate, and each experiment was repeated at least two times.

To test mAbs in immunoblot analyses, Shigella cytosolic and secreted proteins were obtained using the following procedure. Bacterial strains were grown until exponential phase as previously described. Suspensions were further incubated with 10 µM Congo Red at 37°C without shaking for 30 min and finally pelleted at 5,000 x g for 10 min.

Pellets were weighted and resuspended in a 10-fold volume of lysis buffer from Lipopolysaccharide Isolation Kit (Sigma Aldrich, Germany). Lysates were then filtered (0.22-micron), incubated on ice for 10 min and centrifuged at 16,000 x g for 10 min at 4°C to remove unbroken cells. Aliquots of the resulting supernatants were treated with 0.1 mg/ml proteinase K (Qiagen, Germany) and incubated at 60°C for 1 h, followed by a final step of centrifugation at 16,000 x g for 10 min at 4°C. Supernatants were transferred to a clean tube and quantified by using the Pierce bicinchoninic acid (BCA) protein assay (Thermo Fisher Scientific, US).

Supernatants obtained from the centrifugation of the original cultures were filtered (0.22-micron) and concentrated by centrifugation at 5,000 x g for approximately 30 min using 3 kDa Amicon ultrafiltration devices (Merck Millipore, Ireland). The resulting secretomes were transferred to a clean tube and quantified as described for the lysates.

Ten or twenty micrograms of lysates and secretomes were loaded on a NuPAGE Bis-Tris 4-12% protein gel (Thermo Fisher Scientific, US) and run for 1 h at 150 V. The gel was then transferred onto a nitrocellulose membrane with the iBlot2 device (Thermo Fisher Scientific, US) and the membrane was blocked for 2 h at RT in TBST-milk buffer (Tris-Buffered saline (TBS) 1X, containing 0.1% Tween20 and 5% milk). Incubation with 1 µg/ml mAbs prepared in TBST-milk was performed at 4°C rocking overnight. The pooled human serum (diluted 1:1,000 in TBST-milk) and TBST-milk without any mAb were used as a positive and negative control, respectively. The membrane was then washed 3 times with TBST for 3 min each and incubated for 1 h at RT with anti-human or anti-mouse antibodies conjugated to horseradish peroxidase (HRP - Sigma Aldrich, Germany) and diluted 1:75,000 in TBST-milk. A final step of washing with TBST (3 times, 5 min each) followed, before developing the blots by detecting chemiluminescence using the SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific, US) and the iBright 1500 Imagining System (Thermo Fisher Scientific, US).

The pooled human serum utilized as a positive control in this work derives from the sera collected 28 days post-challenge with S. sonnei from individuals enrolled in the NCT03527173 clinical trial (25). The study was conducted in accordance with all applicable regulatory requirements, International Conference on Harmonisation - Good Clinical Practice guidelines, and the Declaration of Helsinki. The protocol and study-related documents had been reviewed and approved by the Cincinnati Children’s Hospital Medical Center (CCHMC) institutional review board on 10 May 2018. Written informed consent was obtained from all participants.

The anti-IpaD mAb named C12-5 was obtained from OF1 mice immunized with the purified IpaDHis recombinant protein from Shigella flexneri M90T strain (33) and produced according to Guillotte et al. (34). Briefly, 10 µg/dose of the recombinant protein were injected subcutaneously at 3-week intervals. Mice with high specific IgG titers against the immunogen received an intraperitoneal boost immunization 4 days before being sacrificed for splenic B cell fusion. Culture supernatants were first screened by ELISA on the recombinant IpaDHis protein and positive hybridomas were then cloned by limiting dilution and selected based on VarO iRBC surface reactivity by flow cytometry. Monoclonal IgGs were precipitated with 50% ammonium sulfate from ascitic fluid and purified using the Melon™ gel monoclonal IgG purification kit (Thermo Fisher Scientific, US). IgG sub-classes were determined by ELISA using the isotrip mouse mAb isotyping kit (Roche). C12-5 IgG1 was biotinylated using EZ-link sulfo-NHS-biotin (Thermo Fisher Scientific, US), centrifuged through Zeba™ spin desalting columns (Thermo Fisher Scientific, US) and quantified by the Pierce BCA protein assay (Thermo Fisher Scientific, US). This study was carried out in strict accordance with the recommendations in the guide for the care and use of laboratory animals of the Pasteur Institute and complied with the French [ref: 35. Décret n° 2013-118 du 1er février 2013 relatif à la protection des animaux utilisés à des fins scientifiques. NOR: AGRG1231951D] and European Union guidelines for the handling of laboratory animals [ref: Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. Document 32010L0063]. Animal care and handling was approved by the Ministère de l’Agriculture et de la Pêche (Ref 107503056792, issued to OMP) and the protocols and procedures approved by the Direction Départementale des Services Vétérinaires du Préfet de Police de Paris (clearance number C75-273 issued to OMP). All animal experiments were planned and executed to minimize animal suffering.

To properly establish a high-content and high-throughput methodology, the choice of the cell line to be used in the assay is pivotal. To this end, 3 different cell lines were analyzed: HeLa, as they are extensively described in the literature for adhesion and invasion assays with Gram-negative bacteria (35–37); CaCo2 and HT29, as they were isolated from intestinal epithelium and are often utilized in adhesion assays for Shigella species (38, 39). Each cell line was plated at different densities and incubated at 37°C for 24, 36 and 48 h. At the end of each time point cells were stained with DAPI and Cell Mask Deep Red to visualize nuclei and cell membranes, respectively ( Figure 1 ). After microscopy inspection, a cell density of 45,000 cells/well was selected since the desired confluency was reached after 48 h incubation. HT29 was the cell line of choice as these cells displayed a more conventional nuclear state than CaCo2, where polynucleated cells were often observed. In addition, HT29 cells presented the most uniform staining of the cell membrane among the 3 selected cell lines.

Initial experiments, where HT29 cells were infected with S. sonnei and then stained as described in Materials and Methods to identify cells and bacteria, aimed at designing the image analysis pipeline using the Harmony software associated with Opera Phenix. This consisted of a series of consecutive modules ( Figure 2 ; Supplementary Table 2 ). Cells were initially identified and counted by segmenting their nuclei after DAPI staining ( Figure 2A , panel 2 “Cell nuclei” and Figure 2B , panel 9 “Cellular DNA”); the corresponding membranes were then localized by staining the cells with Cell Mask Deep Red ( Figure 2A , panel 3 “Cell cytoplasm” and Figure 2B , panel 10 “Plasma membrane”). Next, bacteria were recognized and segmented thanks to the sfmCherry fluorescence signal they expressed ( Figure 2B , panel 11 “Bacteria”). Expression of sfmCherry is crucial to visualize bacteria, which would not be distinguishable from the cell nuclei based only on DAPI staining ( Figure 2B , panel 9 “Cellular DNA”). Unspecific signals were excluded based on Shigella morphological properties (i.e., area, length, and width). In this module, the objects that the software considers as bacteria are automatically depicted in green, while those excluded from the analysis are depicted in red ( Figure 2A , panel 4 “Bacteria”). Finally, the number of bacteria and their position with respect to cells was determined by the outcome of the immunostaining. Specifically, bacteria were grouped into 3 different populations: external, adherent and internal bacteria. External bacteria were those bound by the Denka antiserum specific for the OAg- strain but not in contact with the host cells, displaying a green signal decorating the bacterial surface due to the incubation with the Alexa Fluor 488 anti-rabbit antibody ( Figure 2B , panel 12 “Adherent and external bacteria”). In this module, the software selects only those bacteria that are in contact with the host cells (the infecting bacteria) and depicts them automatically in green, while those excluded from the analysis (the external bacteria) are depicted in red ( Figure 2A , panel 7 “Infecting bacteria”). Bacteria that exhibited the green signal due to the incubation with the Alexa Fluor 488 anti-rabbit antibody while being near the cells were considered adherent ( Figure 2B , panel 12 “Adherent and external bacteria”), while those appearing only yellow due to the expression of sfmCherry were considered internal. The immunostaining is crucial to distinguish the adherent from the internal bacteria, considering that the latter could not be reached by the Denka antiserum since cells were not permeabilized in the vAIA procedure. In this module, the software selects only the internal bacteria and depicts them automatically in green, while those excluded from the analysis (the adherent bacteria) are depicted in red ( Figure 2A , panel 8 “Internal and adherent bacteria”).

Importantly, fluorescence associated with each biological component involved in the experiment was easily detectable when analyzed separately and as a merged image ( Figures 2A, B ). Not only was bacterial and cell fluorescence well defined, but most importantly it was possible to observe a homogenously distributed green signal decorating the surface of adherent and external bacteria. Appropriate negative controls, represented by uninfected cells, were used to evaluate cell autofluorescence in the green channel. Of note, there was no or very low background.

The metrics collected in Harmony were used to quantify the readouts of the assay in terms of both the percentage of infected cells and the total number of infecting bacteria (both depicted in green in the corresponding panels of Figure 2A ).

S. sonnei was chosen for assay establishment and validation because serum samples of individuals who had encountered the pathogen in a controlled human infection model CHIM (25) were available in the laboratory.

The infection rates of S. sonnei 53G OAg- and OAg+ strains at MOI100 were compared to test whether the absence of the OAg compromised the capacity of bacteria to infect cells, given that the LPS is an important virulence factor of Shigella (24), but could also shield virulence proteins. S. sonnei 53 OAg+ strain infected HT29 cells approximately 10-fold less than the OAg- mutant ( Figure 3A , Supplementary Figure 4 ). This statistically significant different infection rate was comparable to that of the non-virulent pINV- strain, which lacks the T3SS ( Figure 3B , Supplementary Figure 4 ). The same trend was observed for the total number of infecting bacteria (internal + adherent bacteria) and when considering separately the percentages of adherent and internal bacteria ( Figures 3C, D , Supplementary Figure 4 ). Results obtained at MOI30 (see Supplementary Figure 5 ) reflected what was seen at MOI100, even though differences between strains were less pronounced. Taken together, these observations confirmed the importance of T3SS for the virulence of Shigella and the masking effect that the LPS has on the T3SS in S. sonnei, which has been already reported in the literature and is further discussed in the Discussion section. As we were interested in maximizing the ability of vAIA to detect different infection rates after treatment with sera and mAbs, we decided to focus on the OAg- strain, since it infected HT29 cells more proficiently.

To further evaluate the sensitivity of vAIA in discriminating the adhesion and virulence properties of different S. sonnei OAg- strains, mutants in single components of the T3SS (IpaB, IpaD and Spa33) and in the 2 proteins involved in Shigella adhesion and invasion of host cells (VirG and IcsP) were assessed.

A statistically significant difference between the infection rate of S. sonnei 53G OAg- and the corresponding mutants was observed ( Figure 4A ). The percentage of infected cells for the control strain (OAg-) was almost 5-fold higher compared to that of OAg- IpaB-, OAg- IpaD- and OAg-Spa33- mutant strains. The difference between the OAg- strain and the OAg- VirG- and OAg- IcsP- mutants was less pronounced. These results were mirrored by the total number of infecting bacteria (internal + adherent bacteria – Figure 4B ). Interestingly, the behavior of the non-virulent pINV- strain was similar to that of the OAg- IpaB-, OAg- IpaD- and OAg- Spa33- strains, presenting approximately 4% of infected cells, thus confirming the important role played by IpaB, IpaD and Spa33 in Shigella pathogenesis. When considering separately the percentage of adherent and internal bacteria, no internal bacteria were counted for mutants of IpaB, IpaD, Spa33 and the whole pINV when compared to the OAg- control strain ( Figures 4C, D ). In the case of OAg- VirG- and OAg- IcsP- strains, the percentage of adherent bacteria was comparable to the OAg- strain, as well as the percentage of internal bacteria ( Figures 4C, D ), suggesting that the roles of VirG and IcsP might be less critical in the in vitro conditions tested.

Overall, these data suggested that the visual assay could be exploited for evaluating the reduction in bacterial adhesion/invasion caused by genetic or functional inactivation of virulence factors.

After defining the best conditions for vAIA, the capacity of a pooled human serum derived from subjects challenged with S. sonnei in a CHIM (25) to inhibit adhesion to and/or invasion of epithelial cells was tested. Comparison with the standard assay, represented by the conventional AIA, was used to validate results.

The outcomes of both vAIA and conventional AIA showed that the absolute number of internal and adherent S. sonnei 53G OAg- bacteria significantly decreased in a dose-depended manner with the increasing concentration of the pooled human serum ( Figure 5A ). Pearson r test confirmed a statistically significant correlation between the two procedures (r=0.8631; P<0.05), thus highlighting the accuracy of the newly developed protocol in reproducing results obtained using the standard assay. Moreover, in both assays it was possible to observe a statistically significant reduction (P<0.05) in the total number of internal and adherent bacteria upon comparing the serum at the highest concentration used with the no-serum condition. By looking at the relative percentages of adherent and internal bacteria, a significant 40% increase in the number of adherent bacteria was noted, together with a 46.5% decrease of the internal ones ( Figure 5B ).

Images in Figure 5C exemplify the above: in the wells where no serum was added the overall number of bacteria was higher compared to those wells where samples were treated with increasing concentrations of serum. Furthermore, in the no-serum condition most bacteria were internal (depicted in green in the segmented panels), while upon addition of serum almost all bacteria were located outside the host cells (in red in the segmented panels). A homogenous binding of the serum around the bacterial surface was observed at the highest dose used (dilution factor 1:10) in a vBA assay ( Figure 5D ). Immunoblot analysis further confirmed binding of the pooled human serum to different antigens of the S. sonnei 53G OAg- total protein extract ( Figure 5D ), with an observed enrichment of the binding signal to proteins at around 70 kDa mass size in both the whole lysate and the secretome of bacteria. Taken together, these data indicate that a polyclonal human serum can inhibit invasion of HT-29 cells by the S. sonnei 53G OAg- strain.

Given the promising results obtained with the pooled human serum, vAIA was employed to evaluate the efficacy of a specific mAb against a key component of the T3SS in inhibiting the invasion of Shigella. To do so, the murine mAb targeting IpaD, named C12-5, was tested.

The C12-5 anti-IpaD mAb was able to bind homogeneously S. sonnei 53G OAg- in vBA ( Figure 6A ) and immunoblot analysis revealed that C12-5 identified a protein of 40 kDa in both lysate and secretome of S. sonnei 53G OAg- ( Figure 6A ). No signal was detected in the lysate and secretome preparation from the negative control strain lacking IpaD ( Figure 6A ), thus confirming its specificity. Image analysis of vBA showed more than 40% of bound bacteria at the highest dose (40 µg/mL) of C12-5, whilst the percentage decreased in a dose-dependent manner with the decreasing concentration of the mAb ( Figure 6B ). No, or very low, binding was detected in the negative controls (C12-5 against the IpaD-deficient strain and by using an unrelated mAb against SARS-CoV-2). When tested in vAIA and conventional AIA, C12-5 reduced the total number of internal and adherent bacteria in a dose-dependent fashion ( Figure 6C ), indicating that C12-5 inhibited invasion of S. sonnei 53G OAg- strain in HT-29 cells. Pearson r test confirmed a statistically significant correlation between the two procedures (r=0.9307; P<0.01), thus highlighting the accuracy of the newly developed protocol in reproducing results obtained using the standard assay also with a mAb. Additionally, a statistically significant reduction (P<0.05) in the total number of internal and adherent bacteria was measured in both assays upon comparing the mAb at the highest concentration used with the no-mAb condition. No effect was observed on the mutant S. sonnei OAg- lacking IpaD and with an unrelated mAb against SARS-CoV-2, used as negative controls. Images in Figure 6D are representative of the above: in the wells where no mAb was added the overall number of infecting bacteria (in green in the segmented panels) was higher compared to those wells where samples were treated with increasing concentrations of C12-5. Notably, at the highest mAb dose (40 µg/mL) the number of internal and adherent bacteria measured in vAIA ( Figure 6C ) was almost the same as that of the negative control used in the assay (OAg- IpaD-), suggesting that C12-5 effectively neutralized S. sonnei 53G OAg- ability to invade host cells.