Mengjun Cheng

We report a case of a 63-year-old female patient who developed a recurrent urinary tract infection (UTI) with extensively drug-resistant Klebsiella pneumoniae (ERKp). In the initial two rounds of phage therapy, phage resistant mutants developed within days. Although ERKp strains were completely resistant to sulfamethoxazole-trimethoprim, the combination of sulfamethoxazole-trimethoprim with the phage cocktail inhibited the emergence of phage resistant mutant in vitro, and the UTI of patient was successfully cured by this combination. Thus, we propose that non-active antibiotic and bacteriophage synergism (NABS) might be an alternative strategy in personalized phage therapy.

plaque-forming unit phages). All patients in our COVID-19-specific intensive care unit (ICU) with CRAB positive in bronchoalveolar lavage fluid or sputum samples were eligible for study inclusion if antibiotic treatment failed to eradicate their CRAB infections. While phage susceptibility testing revealed an identical profile of CRAB strains from these patients, treatment with a pre-optimized 2-phage cocktail was associated with reduced CRAB burdens. Our results suggest the potential of phages on rapid responses to secondary CRAB outbreak in COVID-19 patients.

Enterococcus faecalis is becoming an increasingly important opportunistic pathogen worldwide, especially because it can cause life-threatening nosocomial infections. Treating E. faecalis infections has become increasingly difficult because of the prevalence of multidrug-resistant E. faecalis strains. Because bacteriophages show specificity for their bacterial hosts, there has been a growth in interest in using phage therapies to combat the rising incidence of multidrug-resistant bacterial infections. In this study, we isolated a new lytic phage, EF-P29, which showed high efficiency and a broad host range against E. faecalis strains, including vancomycin-resistant strains. The EF-P29 genome contains 58,984 bp (39.97% G+C), including 101 open reading frames, and lacks known putative virulence factors, integration-related proteins or antibiotic resistance determinants. In murine experiments, the administration of a single intraperitoneal injection of EF-P29 (4×105 PFU) at 1 h after challenge was sufficient to protect all mice against bacteremia caused by infection with a vancomycin-resistant E. faecalis strain (2×109 CFU/mouse). E. faecalis colony counts were more quickly eliminated in the blood of EF-P29-protected mice than in unprotected mice. We also found that exogenous E. faecalis challenge resulted in enrichment of members of the genus Enterococcus (family Enterococcaceae) in the guts of the mice, suggesting that it can enter the gut and colonize there. The phage EF-P29 reduced the number of colonies of genus Enterococcus and alleviated the gut microbiota imbalance that was caused by E. faecalis challenge. These data indicate that the phage EF-P29 shows great potential as a therapeutic treatment for systemic VREF infection. Thus, phage therapies that are aimed at treating opportunistic pathogens are also feasible. The dose of phage should be controlled and used at the appropriate level to avoid causing imbalance in the gut microbiota.

Klebsiella pneumoniae (K. pneumoniae) spp. are important nosocomial and community-acquired opportunistic pathogens, which cause various infections. We observed that K. pneumoniae strain K7 abruptly mutates to rough-type phage-resistant phenotype upon treatment with phage GH-K3. In the present study, the rough-type phage-resistant mutant named K7RR showed much lower virulence than K7. Liquid chromatography-tandem mass spectrometry (LC-MS-MS) analysis indicated that WcaJ and two undefined glycosyltransferases (GTs)- named GT-1, GT-2- were found to be down-regulated drastically in K7RR as compared to K7 strain. GT-1, GT-2 and wcaJ are all located in the gene cluster of capsular polysaccharide (CPS). Upon deletion, even of single component, of GT-1, GT-2 and wcaJ resulted clearly in significant decline of CPS synthesis with concomitant development of GH-K3 resistance and decline of virulence of K. pneumoniae, indicating that all these three GTs are more likely involved in maintenance of phage sensitivity and bacterial virulence. Additionally, K7RR and GT-deficient strains were found sensitive to endocytosis of macrophages. Mitogen-activated protein kinase (MAPK) signaling pathway of macrophages was significantly activated by K7RR and GT-deficient strains comparing with that of K7. Interestingly, in the presence of macromolecular CPS residues (>250 KD), K7(ΔGT-1) and K7(ΔwcaJ) could still be bounded by GH-K3, though with a modest adsorption efficiency, and showed minor virulence, suggesting that the CPS residues accumulated upon deletion of GT-1 or wcaJ did retain phage binding sites as well maintain mild virulence. In brief, our study defines, for the first time, the potential roles of GT-1, GT-2 and WcaJ in K. pneumoniae in bacterial virulence and generation of rough-type mutation under the pressure of bacteriophage.

Visual appearance is considered to be the most important cue to understand images for cross-modal retrieval, while sometimes the scene text appearing in images can provide valuable information to understand the visual semantics. Most of existing cross-modal retrieval approaches ignore the usage of scene text information and directly adding this information may lead to performance degradation in scene text free scenarios. To address this issue, we propose a full transformer architecture to unify these cross-modal retrieval scenarios in a single Vision and Scene Text Aggregation framework (ViSTA). Specifically, ViSTA utilizes transformer blocks to directly encode image patches and fuse scene text embedding to learn an aggregated visual representation for cross-modal retrieval. To tackle the modality missing problem of scene text, we propose a novel fusion token based transformer aggregation approach to exchange the necessary scene text information only through the fusion token and concentrate on the most important features in each modality. To further strengthen the visual modality, we develop dual contrastive learning losses to embed both image-text pairs and fusion-text pairs into a common cross-modal space. Compared to existing methods, ViSTA enables to aggregate relevant scene text semantics with visual appearance, and hence improve results under both scene text free and scene text aware scenarios. Experimental results show that ViSTA outperforms other methods by at least 8.4% at Recall@ 1 for scene text aware retrieval task. Compared with state-of-the-art scene text free retrieval methods, ViSTA can achieve better accuracy on Flicker30K and MSCOCO while running at least three times faster during the inference stage, which validates the effectiveness of the proposed framework.