Periodontitis is a chronic inflammatory disease characterized by an altered homeostasis between the subgingival microbiome and the host gingival inflammatory response in susceptible individuals (1). A recent report from the World Health Organization (WHO) estimates that the severe form of the disease has affected more than a billion cases worldwide and up to 1/5 of the global adult population (2). Periodontitis is characterized by the progressive destruction of hard and soft tissues supporting the tooth (gingiva, periodontal ligament, and alveolar bone) (3). If left untreated, it will inevitably result in tooth loss, masticatory impairment, and a severe impact on individuals’ quality of life (4). The global burden of this disease across the United States and Europe is estimated to be over $150 billion (5). Its treatment relies upon self-performed improved dental hygiene practices as well as professional dental biofilm removal, with a small minority of patients requiring more sophisticated procedures aimed either at resolving resistant gingival inflammation (localized gum surgeries) and/or attempting to rebuild in part the lost hard and soft tissues supporting the dentition (6). The relevance of periodontitis as a global health problem has reached greater attention over the last 30 years, confirming firm links with a substantial host response characterized by a sustained low-grade systemic inflammation (7) and an altered circulating inflammatory cell profile (8). Over decades, studies have reported that periodontitis is independently associated with numerous inflammatory disorders, such as cardiovascular disease, type-2 diabetes, and rheumatoid arthritis, amongst others (9).

For decades, researchers have tried to understand the pathogenesis of periodontitis, and its possible causal association with other common inflammatory comorbidities has been heavily researched. The latest efforts have focused on the host-pathogens interplay using genetic susceptibility and single-cell transcriptomic analyses, revealing that polymorphisms of inflammatory response genes (10) and innate immune cell dysfunction (11) confer susceptibility of the host to periodontitis. Despite periodontal tissue inflammation resolution following the periodontitis management, the augmented level of some systemic inflammation markers persists, and circulating inflammatory cells still retain their altered periodontitis-associated phenotypes, including increased counts of peripheral blood-derived immune cells (12) demonstrating enhanced inflammatory responses (13, 14). Collectively, the evidence highlights the crucial role of host inflammatory responses (immune cells and their inflammatory mediators) to periodontitis and the underappreciated potential contribution of host immunoinflammatory status to periodontitis-associated systemic inflammation.

To date, studies exploring human immune responses to a keystone pathogen Porphyromonas gingivalis (P. gingivalis) focus on ex vivo human models using tissue cultures (15–19) and local gingival tissue phenotyping (20, 21). Furthermore, animal models of periodontitis have been used to describe and understand the gingival inflammatory response to the pathogen in rodents (22, 23). However, both approaches have numerous limitations, notably that they might not translate into the complexity of the in vivo human host response to P. gingivalis due to environmental restrictions in a controlled niche of tissue cultures and different animal immune systems. For instance, neutrophil extracellular trap (NET) formation (NETosis) potentiates periodontal inflammation in mice during experimental periodontitis (24). NETosis is evident in periodontal lesions in mice and NET removal by systemic delivery of DNAse-I protects mice from in vivo periodontal bone loss. Mechanistically, NETosis is regulated by peptidylarginine deiminase-4 (PAD4)-mediated histone citrullination and this further confirms the role of NETosis in periodontitis – PAD4 KO mice displayed significant protection from periodontal bone loss compared to wild-type (WT) counterparts and in vivo pharmacologic delivery of CI-amidine to inhibit PAD4-mediated citrullination protected WT mice from ligature-induced periodontitis (25). However, the mechanism of NETosis in mice is different from that in humans, as experiments of human neutrophils have shown that the cells were still capable of performing NETosis following the PAD4 inhibitor, Cl-amidine (26). In terms of DNAse-I treatment, it failed to degrade NETs purified from neutrophils in humans (27). This conflicting evidence highlights that the mechanism shown in mice was not confirmed in humans. Therefore, the generation of new human in vivo evidence is crucial in unveiling the immunological mechanisms of periodontitis and its association with inflammatory comorbidities in humans.

The skin blister model is a recognized in vivo human inflammation model that is safely and effectively used to study host-pathogen interactions (28). Using this model, researchers have unraveled important mechanisms and pathways of the human immune response to infective agents and chronic inflammatory diseases. This model involves the intradermal injection of a putative agent (infective or not) and the subsequent generation of a local skin blister (using a vacuum machine). The exudate harvested from the artificial blister offers the opportunity to study the kinetics of the onset and resolution of human inflammation, including the innate and adaptive immune responses to the challenge (29, 30). This review discusses how the skin blister model represents a novel opportunity to understand more about the human response to a common pathogen implicated in the pathogenesis of periodontitis (P. gingivalis) and its close links with other systemic inflammatory comorbidities. A brief review of the immune responses implicated in the onset of periodontitis and its systemic consequences is followed by a description of the novel P. gingivalis human challenge model, outlining its potential benefits and limitations.

Periodontitis occurs as a result of polymicrobial synergy and dysbiosis in susceptible hosts. A combination of impaired immunity and systemic and environmental factors could be a vital determinant of the shift within the gingival ecosystem to dysbiosis. The local host inflammatory response is responsible for periodontal tissue destruction (31). For the focus of this review, a prominent role will be given to the trait of immunologic susceptibility which not only affects the onset of periodontitis and its progression but also wound-healing potential during various treatment steps of periodontitis (32).

Studies on understanding the immunologic susceptibility to periodontitis have focused on population-level genetic analyses, including single nucleotide polymorphisms (SNPs) of inflammatory genes – reviewed extensively in (10) and (33). Recent genome-wide association studies (GWAS) have contributed to the identification of novel loci encoding immune response-related genes associated with periodontitis (34–39). The genome studies have expanded further using single-cell RNA sequencing (sc-RNA seq), such as in the experiments of Williams et al. demonstrating the presence of exaggerated responsiveness in stromal cells along with augmented neutrophil and leukocytes infiltration in periodontal tissue of patients suffering from periodontitis (40). Further sub-clustering of inflammatory cells, particularly myeloid cells, also exhibited the amplification of inflammatory response. Specifically, macrophages derived from gingival tissue of periodontitis individuals exhibited pro-inflammatory phenotype by highly expressing NLRP3, an inflammasome-mediating IL-1β production (41). Translating the evidence reported by SNP, GWAS, and sc-RNA seq studies ideally requires in vivo human evidence on the exact mechanisms of the pathogenesis of periodontitis.

Refined understanding of the cellular and molecular mechanisms of host-microbe interactions in the pathogenesis of periodontitis, as extensively reviewed elsewhere (42–44), has important implications for the treatment of periodontitis. The current approach of standard periodontal treatments incorporates dental biofilm reduction surrounding the teeth and periodontal tissues (non-surgical and surgical periodontal therapies) (45). However, the mechanical interventions for periodontitis have been unable to improve clinical outcomes in certain patients, not to mention those with systemic inflammatory diseases such as diabetes - uncontrolled disease is detrimental to periodontal health and can boost treatment complications (46, 47). Host modulation therapy, therefore, emerges to overcome the limitation of mechanical debridement and has been proposed as an adjunctive therapy alongside the standard treatments (48, 49). Despite the numerous studies of novel host-modulating agents for periodontitis, the majority of the agents are still in preclinical studies and have yet to move forward to clinical applications (50).

Understanding the immunological basis of periodontitis could also be relevant when assessing the links between the disease and other common chronic inflammatory diseases. Periodontitis is not confined to the gingival tissues but has been consistently associated with increased systemic inflammation and altered circulating inflammatory cell profiles (8, 51). Clinical evidence confirms that patients with periodontitis exhibit elevated levels of pro-inflammatory mediators, including IL-1β, IL-6, TNFα, and C-reactive protein (CRP), as well as increased neutrophil counts in peripheral blood when compared to controls (52–55). This relationship seems reversible as effective treatment of periodontitis results in a normalization of some inflammatory biomarkers, including CRP (56, 57). Some cellular markers, however, do not seem to be affected by the resolution of local gingival inflammation as achieved by conventional dental cleaning sessions (13, 14, 58). Dampening the hepatic systemic response due to a reduction of local inflammatory cytokines is the most plausible mechanism explaining a reduction of CRP after treatment of periodontitis. The lack of resolution of immunologic changes could result from a trained immunity state induced by periodontitis-associated systemic inflammation (59). As previously described, this process could be defined as innate immune memory involving changes in the hematopoietic stem and progenitor cells (HSPCs) in the bone marrow after exposure to certain microorganisms (60).

Both clinical and animal experimental evidence support the notion that trained immunity could be a key mechanism explaining the systemic health implications of periodontitis. A recent clinical study confirmed a possible link between raised periodontal inflammation and femoral bone marrow activity in patients with periodontitis compared to controls (61). Ishai and colleagues provided evidence that increased bone marrow activity could mediate the relationship between periodontitis and arterial inflammation (62). Further, two key animal experiments demonstrated HSPC’s involvement when infection with pathogens linked to periodontitis was performed. In mice, continuous release (subcutaneous injection) of a keystone periodontal pathogen, P. gingivalis, caused increased osteoclast differentiation in the bone marrow that was IL-6-dependent (63). Indeed, isolated peripheral blood mononuclear cells from patients with periodontitis compared to healthy controls are predisposed to RANKL-induced osteoclastogenesis ex vivo (64). Further experimental evidence shows that periodontitis-triggered maladaptive innate immune memory in the bone marrow produces myeloid cells with enhanced inflammatory responses. Li and colleagues further confirmed that either experimental periodontitis or arthritis-induced trained immunity plays a role in the two-way relationship between periodontitis and arthritis (65). In this context, enhanced responses in inflammatory cells (e.g. increased production of pro-inflammatory mediators upon ex vivo stimulation) are evident in patients with periodontitis; the immune cells retain these phenotypes following the treatment of periodontitis – extensively discussed in ref (8). This begs the question: could trained immunity be triggered by periodontitis in humans? It remains an important question that warrants further investigation.

In vivo human inflammation models can be classified into two categories based on the delivery route of the challenge. The first group relies upon intravenous injection of bacterial endotoxins (66), cytokine administration (67), typhoid vaccine (68), and strenuous exercise (69), for example. Given the potentially robust systemic inflammatory reaction and possible side effects, most of these models require being conducted in highly controlled research centers. Alternatively, less invasive inflammation models triggered by local intervention have been developed. Our group has previously characterized a human model of treatment of periodontitis (70). A transient systemic inflammatory response of moderate magnitude has been described extensively (70–72). Local administration of inflammatory stimuli, for example, in the skin, has also been proposed as a possible human model to characterize human reaction to infective agents. This localized tissue inflammation model facilitates the investigation of leukocyte migration and accumulation of soluble inflammatory mediators at the site of injection. In a human lung infection model, different encounters (e.g. endotoxin, ozone, and rhinovirus) to induce inflammation have been utilized and can be delivered through inhalation. Sample isolation is needed to understand the temporal inflammatory profile after the challenge. However, it can be more challenging to perform and less comfortable for the participants (sputum collection) (73), limiting its applicability and reliability.

Skin models of inflammation offer instead a contained and practical alternative to these systemic challenge models. Indeed, dermal models are characterized by minimal invasiveness, multiple sites accessible to monitor during the induction of inflammation, and easier access for sample/fluid collection. Several types of skin inflammation models have been established in humans based on the categorization of agents used, including tuberculin purified protein derivative (PPD) (74), BCG (75), cantharidin (28), Candida (candin) antigen (76), varicella-zoster virus (VZV) (77), Escherichia coli (E. coli) (78), and Streptococcus pneumoniae (S. pneumoniae) (79) and lipopolysaccharides (LPS) (80). An artificial blister is created at the inoculation site using a vacuum machine secured on the skin at selected time points after intradermal injection. The negative pressure generated by the machine separates the epidermis from the dermis ( Figure 1 ), allowing confinement, collection and drainage of an inflammatory exudate (81). The intradermal injection of the pathogen resulted in localized inflammatory responses with complete local clearance of the pathogen and no systemic dissemination of the pathogen (30, 82). As a self-resolving inflammatory model in humans, these experiments have been replicated by several research groups and have shown good tolerability for participants recruited (including healthy individuals of different ages and patients with chronic diseases) (83–85).

For example, in human immunobiology the cantharidin-induced skin blister model helped to identify impaired efferocytosis in the elderly. Furthermore, using this model researchers identified oral intake of losmapimod – a selective p38 mitogen-activated protein kinase (MAPK) inhibitor – as a therapeutic intervention capable of reversing this inflammatory resolution defect (84). A similar model has also been validated to evaluate the anti-inflammatory properties of anti-TNF (adalimumab) and corticosteroids in reducing innate inflammatory cell recruitment to the injection site (86).

Amplified acute inflammation onset with delayed resolution was evaluated in patients with ulcerative colitis when challenged with UV-killed E. coli and S. pneumoniae prior to suction skin blister formation (83). Using the same model of UV-killed E. coli injection, another study reported prolonged immune alterations following the resolution of inflammation, challenging the notion that homeostasis is achieved after acute inflammation resolves (78). Several endogenous specialized pro-resolving mediators (SPMs) and their receptors have been identified as key regulators of the initial phase of inflammation resolution, and furthermore the administration of exogenous SPMs has been shown to boost this process in humans (87).

The skin blister model has also contributed to a greater understanding of the dynamics of adaptive immune cells in humans. When a tuberculin PPD injection (skin blister model) was utilized to study human T cell recall responses, numerous regulatory mechanisms for maintaining human memory T cells in vivo were observed (88). The isolation of antigen-specific memory T cells from the suction blister followed by in vitro stimulation to become regulatory T cells (Tregs) was attainable, and this was demonstrated to have a suppressing role in inflammation (89). Synchronized production of IL-2 by CD4+ T cells in blister fluid and skin biopsy was showed in the model, enabling the non-invasive characterization of potential defects in adaptive immune response in the elderly. Blister fluid analyses became a valid alternative to skin biopsy to characterize the dynamic of host response to various triggers (74).

A series of further experiments involving patients with active, latent, or cured tuberculosis (TB) was able to describe the role of IL-17A and Th17 responses in patients with active TB, when compared to those with latent TB (85). When TB treatment was administered in these patients, the same immune changes resolved, with a decrease in IL-17A and Th17 responses. In this case, it was skin biopsy samples following the intradermal injection of tuberculin PPD that were analyzed (85). Subsequent experiments using this model enabled the assessment of the safety and specificity of novel monoclonal antibodies aimed at reducing the recruitment of Th17 during adaptive immunity (90). When candin was used as an inflammatory insult, the skin blister model revealed that restricted immunosurveillance by memory T cells in the ageing population is dependent on the reduction of macrophage-derived TNFα production (76). Finally, when using VZV as an insult, researchers showed how regulatory T cells may be derived from memory T cells during localized inflammatory responses (91).

The inflammation-induced skin blister model is a well-established human model of inflammation that successfully captures the immunoinflammatory status of healthy and diseased individuals. Specifically, the model can unveil the complex mechanisms of human immune responses in chronic inflammatory diseases and/or in response to specific inflammatory stimuli/infections. The host laboratory developed for the first time a safe and reproducible UV-killed P. gingivalis intradermal injection model followed by artificial suction blister formation and collection of cellular and soluble inflammatory infiltrates.

Periodontal research could use this novel model to enhance our understanding of which immunological mechanisms occur in humans with periodontitis and shed greater light on the potential link between this common chronic disease and other systemic inflammatory disorders.