Human granulocytic anaplasmosis, caused by the tick-transmitted Anaplasma phagocytophilum, is the third most common human vector-borne infection in the U.S., where 1% die and 7% require ICU admission (1–3). A. phagocytophilum is not controlled by innate immunity, but induction of a proinflammatory disease state occurs with innate immune cell activation via TLR2 and the inflammasome to achieve STAT1-mediated IFNγ and NF-κB-mediated proinflammatory gene transcription (4–8). As a result, infection in humans and in animal models leads to a macrophage activation syndrome (MAS) where severity is related to high serum levels of IFNγ, IL-10, IL-12, and ferritin (9–12).

In A. phagocytophilum murine infection models, inflammatory hepatic histopathologic injury is abrogated in Ifng^−/− and enhanced in Il10^−/− animals, confirming an important role for IFNγ in driving tissue injury (5, 6, 9, 10, 13). Key candidates for production of IFNγ include the innate immune NK and NKT cells, but also include CD8 T lymphocytes as adaptive immune responses mature. NK, NKT, and CD8 T lymphocytes react to fundamentally distinct ligands, microbe-associated molecules, or other signals (14–16). TLR2-activation implies a role for A. phagocytophilum lipoproteins as a ligand, and induction of the inflammasome via NLRC4 relates to endogenous host cell eicosanoid production following infection. However, the specific A. phagocytophilum ligands that drive inflammation and disease are not known (4, 17). Regardless, these observations secure an inflammatory basis for A. phagocytophilum-induced septic or toxic shock-like manifestations and imply that severity has its pathophysiologic basis in MAS and hemophagocytic syndrome (HPS) (11, 18).

MAS and HPS are related disorders that have either a genetic basis or an infectious trigger. Both are cytokine-driven diseases characterized by progressive fever, shock, organ failure, pancytopenia, liver dysfunction, and coagulopathy, and usually attributed to excessive IFNγ production (19, 20). Genetic forms of HPS result from mutations of genes encoding proteins involved in signaling perforin or granzyme delivery for cytolysis of target cells resulting in impaired cytotoxic lymphocyte (CTL) function, often among NK cells. Similarly, infection-associated HPS is characterized by defects in CTLs, either by NK cell lymphopenia, or NK cell defects in perforin delivery, although HPS has been observed in humans and animal models with defects in CD8 T cells as well (21). The explanation for the relentless progression with HPS and MAS is that the antigen-presenting cell (APC)-cytotoxic cell synapse through TCR–MHC class I interaction results in activation of the CTL and production of IL-12, IL-18, IL-15, and IL-2, resulting in lymphoproliferation and IFNγ generation (19, 20). IFNγ activates macrophage effector production (nitric oxide, reactive oxygen species, TNFα, phagocytosis). However, the inability of CTLs to deliver perforin to the APC presenting a cognate ligand frees the cascade from regulation, exacerbating disease due to unremitting cytokine stimulation (22).

Recent studies provide evidence that the signaling pathways generating the key functions for this response are distinct, providing a framework for understanding the dichotomy of hypercytokinemia without cytotoxicity (20, 23). To better understand the nature of MAS/HPS induced by A. phagocytophilum infection, we studied in vivo and in vitro responses of NK, NKT, and CD8 T lymphocytes from infected animals to determine if their interactions with A. phagocytophilum-loaded APCs results in delivery of cytotoxic cargo dissociated from intracellular production of IFNγ.

Obligate intracellular bacteria such as A. phagocytophilum use their host cells as part of an extended environment and have evolved a capacity to interact with and manipulate host cells to improve microbial fitness, sometimes at the cost of damage to the host in the form of disease (18, 29, 30). Studies of A. phagocytophilum, therefore, focus on two important aspects of the unique bacterium, how it survives and propagates within the key host defense cell, the neutrophil, and how it causes disease, which increasingly is documented to involve inflammatory and immune induction (18, 31–33). Much has been studied regarding mechanisms by which the immune response resolves A. phagocytophilum infection, largely predicated on the development of CD4 T lymphocyte responses (5, 34). During investigations of immune pathways involved in resolving infection, it was discovered that although many innate and adaptive immune pathways are activated, few impact microbial burden (5). Of great interest was the discrepancy between bacterial load and inflammatory tissue injury or disease manifestations in humans and animal models (9, 10). It has increasingly become clear that infection of the neutrophil likely contributes to disease by induction of proinflammatory responses and lack of antimicrobial killing (18, 33), but that key protective mechanisms are initiated via macrophages and other APCs that do not sustain infection, but contribute to the proinflammatory disease process and severity of infection (17, 35–37). In fact, aspects of granulocytic anaplasmosis in humans and animals mimic MASs and HPSs (10–12). We sought to discern whether A. phagocytophilum induces MAS or HPS driven by defective CTL delivery of perforin and granzyme while promoting hypercytokinemia primarily focused on IFNγ to further activate macrophages. Thus, we designed in vivo pilot studies and in vitro studies to examine the roles of immune CD8 T lymphocytes, NK cells, and NKT lymphocytes to interrogate both cytotoxicity responses and promotion of IFNγ production.

The findings from this investigation support the concept of impaired cytotoxicity of innate and adaptive immune CTLs that are, thus, unable to exert homeostatic control of APCs following antigen processing and MHC class I expression to these CTLs. While IFNγ is readily detected during in vivo infections, the relative lack of its expression in CTLs after exposure to APCs that should otherwise present microbial targets for activation suggests that it originates from other sources and that the functional defect could instead reside within the APC. While these data do not support the classical paradigm toward development of MAS (19), the ongoing production of IFNγ by other cells coupled with the lack of effective feedback cytotoxicity by classical CTLs would be equivalent.

Anaplasma phagocytophilum can stimulate proinflammatory responses in macrophages in vitro through TLR2, presumably via a lipoprotein (4, 38). However, the recognition that intracellular pathogens are also sensed and controlled via intracellular pattern recognition receptors has triggered a line of investigations into nucleotide-binding domain and leucine-rich repeat containing proteins (NLRs) (37, 39). In fact, recent studies showed that A. phagocytophilum activates the inflammasome in macrophages by a novel process that involves NLRC4 and production of endogenous eicosanoids (17). Moreover, studies of inflammasome activation by this and other obligate intracellular bacteria that lack classical pathogen-associated molecular patterns (e.g., type III secretion system components or cell wall products) are now just being initiated.

Anaplasma phagocytophilum does not productively infect macrophages or presumably other APCs that could be involved in CTL engagement and activation of cytotoxicity (40). Therefore, how could its presence in APCs lead to dysfunctional recognition by immune CD8 T cells, NKT cells, and NK cells? The bacterium’s lifestyle is obligatory intracellular, and it likely stimulates APCs in delivery of secreted effector proteins or molecules via type IV secretion or other methods to translocate such products into the cell’s cytosol for MHC class I processing (41). By contrast, its ability to engage surfaces of myeloid and monocytic cells suggests effective endocytosis but ineffective remodeling of the newly formed vacuole for prolonged survival within monocyte-differentiated cells (32). It is likely that these vacuoles fuse with the degradative complexes that precede antigen processing and presentation, typically the domain of MHC class II presentation. Given the lack of the ability to generate effective cytotoxicity across three major CTL types (CD8, NKT, and NK cells) suggest a defect in processing for MHC class I presentation within APCs. Permutations for the MHC class I processing route could occur at delivery into or with proteasome degradation within the cytosol, delivery through transporter associated with antigen processing protein into the endoplasmic reticulum, proteolytic processing within the endoplasmic reticulum, loading of cargo onto MHC class I molecules, or delivery of the loaded MHC class I to the cell surface. While there are potential interactions that could be experimentally tested (32, 42, 43), little data currently exist to support most of these potential points of subversion.

Could A. phagocytophilum products derived via the processing for presentation by APCs impact its ability to present and activate CTLs? Among annotated NLRs, recent studies performed implicate NLRC5 as the key regulator of MHC class I gene expression (44–46). Regulation of NLRC5 expression is unclear, but likely linked in part to stimulation by type I and II interferons (47), Poly I:C (a mimic of double stranded RNA) (48), and possibly by bacterial porins, of which A. phagocytophilum expresses large quantities (49). Is it plausible that A. phagocytophilum impairs MHC class I expression via NLRC5, resulting in the observed failure to activate cytotoxicity and IFNγ generation in A. phagocytophilum-specific responses? We recently completed RNAseq transcriptional profiling of A. phagocytophilum infection in ATRA-differentiated HL-60 promyelocytic leukemia cells and examined splice variant transcripts (unpublished data). Although the studies were conducted in a cell differentiated toward myeloid maturity, within the transcriptome were identified 17 distinct alternative isoforms of NLRC5, all but one which were not differentially regulated by the infection; however, one nonsense-mediated decay isoform, NLRC5-010, was expressed at least ninefold greater with infection. Nonsense-mediated decay forms are now recognized as regulators of transcription in cell differentiation, in response to stress, and in development of disease (50). We suggest that additional studies to investigate whether this isoform transcript could affect NLRC5 and MHC class I expression should be conducted.

The results of this study will drive future investigations into the cellular and molecular mechanisms for the CTL functional dissociation of MASs. These data also provide another avenue to explore acquired or infectious MASs unrelated to this pathogen through the conduct of in vivo cytotoxicity studies. Such investigations should confirm in vitro studies and illustrate which cells possess in vivo cytotoxicity, and whether impaired cytotoxicity leads to an enhanced proinflammatory phenotype which would recapitulate MAS and HPS disease states, and whether approaches directed at MHC class I expression might be fruitful avenues for control of disease manifestations.

All animal studies were reviewed and approved by the Johns Hopkins University Institutional Animal Care and Use committee. All mice were housed and cared for following the “Guide for the Care and Use of Laboratory Animals” (25).

DS helped conceive the work, conduct experimentation, interpret results, and write the manuscript. KSC helped to conceive the work, conducted most of the research, interpreted results, and helped to edit the manuscript. JD helped conceived the work, interpret the results, and write the manuscript.

The opinions expressed herein are those of the author(s) and are not necessarily representative of those of the Uniformed Services University of the Health Sciences (USUHS), the Department of Defense (DOD); or, the United States Army, Navy, or Air Force.

The research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.