Infection of diabetic foot ulcer (DFU) presents a significant social and economic burden to our society. Diabetes and its associated infection affect an estimated 9.3% of people in the United States, accounting for more than 20% of all hospitalization (Kim and Boye, 2009; Skrepnek et al., 2017; Schmidt, 2022; McDermott et al., 2023). They frequently result in prolonged hospitalization, poor clinical outcomes, readmission, and, in extreme cases, amputation (Herber et al., 2007; Phillips et al., 2016; Armstrong et al., 2017; Skrepnek et al., 2017; Smith-Strom et al., 2017; Polikandrioti et al., 2020; Raffetto et al., 2020; Zhao et al., 2020; Tai et al., 2021; Guo et al., 2023). Infection or re-current infection of DFU can have an impact on quality of life and present social challenges and a declined in psychological wellbeing (Phillips et al., 2016; Egan et al., 2020; Guo et al., 2023). Staphylococcus aureus is the most common pathogen (70%–80%) found in patients with DFU infection (Macdonald et al., 2020). The main challenge in treating musculoskeletal infection (MSKI) is the identification of main pathogen (Lipsky et al., 2012; Heravi et al., 2020; Lipof et al., 2021). False-negative or false-positive cultures in polymicrobial MSKI may result in inappropriate antibiotic coverage plan, subsequent recurrent infection, and loss of the limb. Although inflammatory markers, such as white blood cell count (WBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP), have been utilized to diagnose and monitor treatment response, they are non-specific (Masters et al., 2019; Grondman et al., 2020; Lipof et al., 2021).

Previous investigations have demonstrated an increased host production of S. aureus– specific antibodies in patients with DFU. Specifically, antibodies against iron- regulated surface determinant (IsdB and IsdH) protein, recombinant staphylococcal complement inhibitor, and the recombinant clumping factors (ClfA and ClfB) were suggested as biomarker candidates to diagnose and monitor treatment response (Masters et al., 2019; Heravi et al., 2020; Choi et al., 2021; Hao et al., 2021; Lipof et al., 2021; Wang et al., 2021; Fritz et al., 2022). These endeavors to characterize humoral responses to S. aureus infection with the use of immunoassay techniques have shown some potential for clinical application, but there are certain limitations (Masters et al., 2019; Choi et al., 2021; Hao et al., 2021; Fritz et al., 2022). Because of the historical abundance of these antibodies produced by host’s previous exposure to infection, the elevated serum titers may not accurately reflect the state of the current infection.

Advancements in bioinformatics have suggested using RNA- seq technology to accelerate the knowledge on transcriptomic analysis. By profiling the transcriptome landscape of cells in heterogeneous tissues, it provides a deep insight into cell function and disease pathophysiology (Jiang et al., 2015). Transcriptome research focuses on mRNA and non-coding RNA acting as a regulator to numerous proteins and gene function/structure and reveals specific molecular mechanisms of disease occurrence (Jiang et al., 2015; Stubbington et al., 2017). The new technology paradigm provides a greater specificity for studying the immune response. The shift from the use of convectional culture-dependent methodologies and serum antibody levels to transcriptome analysis may help us better understand DFU infection, patient outcomes, and wound progression. Few research studies have recently elucidated the role of RNA-seq/transcriptomes in DFU (Theocharidis et al., 2020; Wang et al., 2020; Li Y, et al., 2022; Theocharidis et al., 2022; Sandoval-Schaefer et al., 2023). Transcriptome data analysis, for example, has been utilized to identify gene expression profiles, biological signal pathways, and biomarkers in various diseases such as cancer, tuberculosis, neuronal dysfunction, autoimmune diseases, and COVID-19 prognosis (Baine et al., 2011; Islam et al., 2017; Bellaver et al., 2019; Wang et al., 2019; Xiong et al., 2020; Zheng et al., 2020; Fang et al., 2021; Han et al., 2021; Ascoli et al., 2022; Ghosh et al., 2022; Maleknia et al., 2022; Mendelsohn et al., 2022; Zimmer et al., 2022). Theocharidis et al. recently reported the use of RNA-seq RNA-Seq to elucidate gene expression associated with wound healing in the single-cell transcriptome landscape of DFU in human and mice. The above study focuses on wound healing in DFU. However, in our study, we sought to investigate whether the transcriptome could be used to diagnose infected DFU. Our objective is to develop a more accurate transcriptome-based test specific for infection and its application for diagnosis and prognostication potential. Probing how host immune system responds to infection in DFU through the variability in gene expression will increase our knowledge in diagnosis and aid clinical decision in management of MSKI.

We aimed to analyze evolving human peripheral blood mononuclear cell (PBMC) transcriptome associated with treatment of clinically infected DFU to identify the genes involved in immunologic response to invasive S. aureus DFU infection. We sought to identify up- and downregulated genes among healed patients at the time of active infection (0 weeks) vs. 8 weeks after initiation of therapy. Secondary outcome was to investigate whether certain gene expression (singly or in conjunction with others) could serve as a prognostic indicator for resolution of infection and eventual wound healing.

In this study, we performed RNA- seq transcriptome analysis of PBMCs to accurately and systematically profile gene expression in PBMCs of patients with S. aureus– infected DFU at the time of active infection (0 weeks) and after antibiotic therapy (8 weeks). Developing a robust system for earlier diagnosis of S. aureus infection is essential for diagnosis, treatment, and prognostication. Unlike the report by Theocharidis et al. and other authors on the genes profile of PBMC single-cell RNA that focused primarily on fibroblast- and keratinocyte-associated genes (wound healing) between patients with DFU whose ulcers heal (DFU healers) and those who do not heal (DFU non-healers), we focus primarily on profiling the genes that were differentially expressed in patients with infected DFU at different time points (0 vs. 8 weeks) to improve our understanding of earlier diagnosis, prognostication, and monitoring treatment response. We have identified various genes encoding different immunoglobulin and chemokines that are involved in the progression and healing of wounds in the PBMC samples of patients with S. aureus– infected DFU. Our findings suggest a complex interplay of host immune response via neutrophile and immunoglobulins.

We identified several members of histone genes that are highly regulated during S. aureus– infected DFU wound healing. Among the patients that were healed, we found that core histones such as HIST1H2AJ, HIST1H2AL, HIST1H2BM, HIST1H3B, and HIST1H3G were highly expressed during active infection at 0 weeks (and among the three unhealed patients) but became downregulated at 8 weeks after treatment follow-up, indicating that histone may play a significant role in host immune response to infection resolution. Histones are classified as lysine-rich (H1, H2A, and H2B) or arginine-rich (H3 and H4); among these two groups, H2A, H2B, H3, and H4 are known as “core histones”, whereas H1 is known as “linker histones” (Felsenfeld and Groudine, 2003; Li X, et al., 2022). Histones are the main components of nucleosomes, and they play a role in the packaging and arranging of DNA into functional units. However, there is growing evidence that histones can be found outside the nucleus and become extranuclear histones. These “extranuclear” histones have reported to enhance host defense functions and contribute to inflammatory responses (Urban et al., 2009; Venkatesh and Workman, 2015; Frydman et al., 2020).

Parallel to upregulation of core histones, the CD177 gene (a neutrophil-specific marker), MYBL2 and RRM2 were highly expressed during the active phase of infection (0 weeks) and the expressed genes were reduced at 8 weeks after therapy. These data are consistent with reported positive correlation of RRM2 with most chemokines and chemokine receptors (Zhou et al., 2022). Histones can act as antimicrobial peptides and directly kill bacteria and other pathogens in a variety of animal hosts. In addition, histones can trigger inflammatory responses, in some cases, acting through Toll-like receptors (Morita et al., 2013; Hoeksema et al., 2016). In addition, phagocytes are essential components of innate immune defense against infectious pathogens, and there is mounting evidence that histones and phagocytes have interactions (Li X, et al., 2022). Once phagocytosed, S. aureus is exposed to a variety of toxic products (NO, H₂O₂, superoxide and hydroxyl radicals, hydrolases, and proteolytic enzymes) that kill and degrade the engulfed bacteria.

Neutrophil extracellular traps (NETs) are fibrous structures with decondensed chromatin networks released from activated neutrophils during infection, which can capture and eliminate pathogens (Papayannopoulos, 2018; Song et al., 2019; Tan et al., 2021). NETs contain all core histones (H2A, H2B, H3, and H4), accounting for 65%–70% of all NET-related proteins; although their functions have not been fully elucidated, the presence of histones in NET structure suggested that histones could act as antimicrobial agents in NETs. Histones and their fragments play a vital role in host defense by binding to bacterial nucleic acid and lipopolysaccharide and by changing the permeability of bacterial cell membrane, and the antibacterial role of histones in different species is becoming more widely recognized (Howell et al., 2003; Patat et al., 2004; Lemaire et al., 2008; Urban et al., 2009; Mehta and Jeffrey, 2015; Shrestha et al., 2019; Doolin et al., 2020; Duong et al., 2020). High expression of CD177 and histones at 0 weeks may suggest a strong association with the severity of S. aureus infection.

We noted a high level expression of RRM2, MYBL2, and some other genes such as GALNT14, CENPU, ACOXL, and U2 at 0 weeks during active phase of infection, which were downregulated at 8 weeks after therapy ( Table 1 ). We also observed a low level expression of ARL2-SNX15, MT-RNR1, and AL591806.3 at 0 weeks during active infection, which were upregulated at 8 weeks after therapy ( Figures 4A–C ; Table 1 ). The role of these genes in S. aureus– infected DFU wound healing and resolving infection is not completely understood. Among others, RRM2 has been reported as a member of the ferric iron-binding ferritin superfamily and that RRM2 engaged in the iron metabolism in hepatocellular carcinoma (Key et al., 2020; Shen et al., 2021). Although we did not know whether RRM2 is important for S. aureus pathogenicity, we observed a high expression of this seemly important gene among the three unhealed patients. Given that iron (heme) acquisition is a pathognomonic finding of invasive S. aureus infection, it would be prudent to further investigate factors associated with iron metabolism, which may potentially involve RRM2.

We reported a high expression of IGHG family (IGHG1, IGHG2, and IGHG3) and IGLV family (IGLV3-21 and IGLV6-57) at 0 weeks, which are clinically correlate with active infection. Conversely, IGHG and IGLV expression was reduced at 8 weeks after therapy but remain higher in unresolved infection. Because neutrophil- associated histones and CD117 were also reduced at 8 weeks, these findings reflect a reduction of S. aureus infection burden at 8 weeks after therapy. Our finding suggests that diverse IGHG and IGLV family may contribute to a protective host immune response against S. aureus infection in DFU and may serve as a potential biomarker for diagnosis and prognostication of infection and wound healing.

In addition, we noted a significant elevation in HSP levels in unresolved infection (not healed n = 3) compared with healed patients (n = 17) at 8 week after antibiotic therapy ( Table 2 ). The presence of high levels of inducible HSPs in unhealed patients may be due to the continuous release of virulence machineries by S. aureus and the host immune response to unresolved infection. In line with our observation, HSPs have been reported in cellular response to heat shock/fever and oxidative stress signals and has been found significantly higher during bacterial infection (Borges et al., 2012; Varano Della Vergiliana et al., 2013; Tulapurkar et al., 2015). Further study is needed to investigate clinical utilization of HSPs as a potential transcriptome biomarker for early detection of recurrent or persistent infection after antibiotic therapy.

We found upregulated IGHG1, IGHG2, IGHG3, GLV3-21, IGLV6-57, CD177, HIST1H2AJ, HIST1H2AL, HIST1H2BM, HIST1H3B, and HIST1H3G genes from PBMC during active phase of S. aureus infection in patients with DFU at 0 weeks; these genes became downregulated at 8 weeks among healed patients, and these transcriptomes may reflect infectivity before and after antibiotic therapy. In line with the report by Sandoval-Schaefer et al. (2023) on transcriptional heterogeneity in human diabetic foot wounds, it was reported that profiling chronic foot ulcers from non-diabetic and diabetics patients using single-cell RNA sequencing display transcription changes in gene expression that may provide therapeutic baseline for treatment of DFU. Our findings further add evidence to the existing knowledge that transcriptome profiling of RNA-seq can delineate genes signatures and could be a potential valuable technique for diagnosing infection, determining severity, and measuring host immune response to therapies.

The data presented in the study are deposited in the Gene Expression Omnibus (GEO) repository, accession number “GSE230426”.

The studies involving human participants were reviewed and approved by Human Research Protection and Ethics Committee (Reference No. 2000030908), Yale university. The patients/participants provided their written informed consent to participate in this study.

Conceptualization and supervision: IO; Methodology: IO, JK, DZ; Validation and interpretation of results: IO, FL, TA, H-KK, JK and DZ; writing of original draft: TA; review and editing: TA and IO, funding acquisition: IO. All authors contributed to the article and approved the submitted version.