Intraperitoneal prophylaxis with CpG oligodeoxynucleotides protects neutropenic mice against intracerebral Escherichia coli K1 infection

Immunocompromised individuals, also the elderly, have an increased risk of experiencing systemic and central nervous system (CNS) infections [1, 2]. In neutropenic patients, the causative agents of meningitis often are enteric Gram-negative bacilli such as Escherichia coli that live in the patient’s own digestive tract [3]. The ability of bacteria to escape the host defense and achieve the threshold of bacteremia necessary for subsequent invasion of the brain is probably higher in immunocompromised individuals than in immunocompetent adults thus explaining the differences in the occurrence of E. coli K1 meningitis [4]. Head trauma, neurosurgical interventions, or sepsis are other risk factors for the development of E. coli meningitis in adults either as a consequence of the impairment of the local host defense or subsequent to direct inoculation of bacteria into the CNS [5, 6]. In immunodeficient and older persons the efficacy of current vaccines is low [7]. Moreover, immunization efficacy probably decreases with complex vaccination regimes against multiple pathogens. Vaccination against the majority of pathogens which may cause an infection in immunocompromised patients is an unrealistic goal. Thus, it seems rational to pursue a concept of pattern-specific stimulation of the innate immune system with the goal of increasing resistance to infections by several pathogens in the immunocompromised host.

Bacterial DNA containing unmethylated cytosine-guanidine motifs linked by a phosphodiester (p) group (CpG) activates mammalian lymphocytes and macrophages to produce cytokines including tumor necrosis factor (TNF)-α, interleukin (IL)-6, IL-12, and interferon gamma (IFN-γ), which are crucial for the immune response to bacterial infections [8]. CpG oligodeoxynucleotides (ODN) are short single-stranded DNA molecules which contain unmethylated CpG motifs and mimic bacterial DNA with immunostimulatory properties [9]. CpG-containing motifs are considered pathogen-associated molecular patterns (PAMPs) and are recognized by the pattern recognition receptor (PRR) Toll-like receptor 9 (TLR9) [10]. We recently showed that stimulation of primary murine microglial cells with CpG ODN 1668 increases phagocytosis and intracellular killing of E. coli K1, an important pathogen for meningitis and meningoencephalitis [11, 12]. In previous studies with experimental animals, CpG ODN pre-treatment conferred protection against a variety of bloodstream and other extracerebral bacterial infections [13‐20].

In this study, we investigated the protective properties of CpG ODN 1668 pre-treatment in immunocompetent mice as well as immunocompromised animals which were depleted of granulocytes. To mimic infections after cerebral/spinal trauma or surgery, murine meningoencephalitis was induced by direct injection of E. coli K1 into the CNS. Here, we report for the first time that CpG ODN induces protection against a primary bacterial CNS infection in neutropenic mice in a TLR9-dependent manner but not in immunocompetent animals. CpG prophylaxis promoted bacterial clearance which correlated with enhanced production of IL-12/IL-23p40, IFN-γ, and MIP-1α, and increased recruitment of Ly6C^highCCR2⁺ monocytes at early infection.

The concept of increasing the resistance against several bacterial infections upon systemic stimulation with one immune activator has been pursued for decades, with mixed success. In rodent models, CpG ODN convey protection against a variety of bloodstream and other extracerebral infections induced by intraperitoneal injection of Francisella tularensis[14], Listeria monocytogenes[17] and Burkholderia pseudomallei[20], oral infection of neonates with Cryptosporidium parvum[13], intranasal/intraperitoneal injection of Yersinia pestis[16], intratracheal challenge with Cryptococcus neoformans[18], against experimental malaria elicited by intravenous injection of Plasmodium yoelii[15] and polymicrobial sepsis induced by the colon ascendens stent peritonitis procedure [19].

In the present study, we used mice lacking CD11b⁺Ly6G⁺Ly6C^int neutrophils as a model of the immunocompromised host to investigate the capacity of CpG ODN to strengthen the local immune response for a successful elimination of bacteria injected directly into the brain. We were able to show for the first time that systemic immunostimulation with CpG ODN not only protects against systemic bacterial infections but also against an intracerebral challenge with E. coli. The CpG motif evaluated in this work stimulated microglial cells and increased their ability to phagoytose and destroy E. coli strains intracellularly [11]. Prophylactic intraperitoneal administration of 100 μg of this CpG motif 3 days before intracerebral inoculation of E. coli increased the number of neutropenic mice surviving the infection. By observing the time course of the clinical symptoms we assumed that CpG administration increased the local resistance of the brain against infection: only animals which did not develop clinical symptoms (clinical score >1, severe lethargy), survived. Once the infection had spread and caused clinical symptoms, the ultimate outcome was fatal. Prophylaxis with CpG ODN did not prolong the interval from infection to death (median: 43 h for CpG- and buffer-treated neutropenic mice) suggesting a moderate contribution of the systemic inflammatory response which mainly influences the later course of the infection. The lack of protection in TLR9-deficient neutropenic mice strongly suggests that CpG ODN mediated protection by stimulation of TLR9 and not by any other non-specific immunomodulatory effect. Among the animals that survived 14 days after infection, 45% of the CpG-treated mice showed positive bacterial cultures in cerebellum compared to 10% of buffer-treated animals (P = 0.08; Mann-Whitney U test). No mouse died later than 6 days after infection. To confirm that CpG-treated animals which were alive after 14 days were able to completely control bacterial growth, one of the survival experiments was extended for a period of 2 months. After 2 months, all surviving CpG pre-conditioned animals had cleared bacteria from brain and spleen. In 2-week survival studies using immunocompetent wt mice, CpG ODN pre-treatment reduced the mortality of the buffer group by only 15% (P = 0.2) with no positive cultures in homogenates at the end of the experiment.

In correlation with the enhanced survival, CpG ODN administration significantly reduced bacterial loads in spleen homogenates, an indicator of bacteremia, 42 h after infection. Similarly, bacterial burdens in cerebellum were signicantly lower in mice pre-conditioned with CpG ODN compared to buffer-treated neutropenic animals. However, the meningeal inflammation scores were comparable in both groups. To identify the possible mechanisms behind the CpG-induced protection in our neutropenic mice, we investigated the pattern of cytokine/chemokine production. Because resistance to infections upon CpG ODN pre-conditioning has been related to the promotion of T helper 1 (Th1) responses [13‐15, 18‐20], we focused on IL12 and IFN-γ. IL-12p40 is produced primarily by activated inflammatory cells including monocytes/macrophages, neutrophils, microglia, and dendritic cells (DCs) in response to pathogens mediated by PRRs such as TLRs [28]. IL-12p40 associates with the p35 chain to form IL-12p70, but also with a p19 chain to form IL-23 [29]. In the present study, administration of CpG ODN caused a six-fold increase in serum IL-12/IL-23p40 levels at 52 h before infection. The elevated levels of IL-12/IL-23p40 in serum persisted for at least 17 days after CpG ODN prophylaxis (median: 313 pg/mL in CpG ODN vs. 192 pg/mL in buffer-treated mice at 14 days after infection). Correspondingly, CpG-treated animals showed an increased concentration of IL-12/IL-23p40 in spleen homogenates at 42 h, and these high levels were maintained until 14 days after infection. CpG-preconditioned animals showed increased levels of IFN-γ in spleen homogenates compared to the buffer group at 42 h after infection. CpG DNA induces the production of IL-12 that promotes IFN-γ secretion by NK cells [30]. We did not observe differences in the number of NK cells between the CpG and buffer-treated groups at 42 h after infection.

Pre-treatment with CpG ODN induced the recruitment of leukocytes to lung tissue and gastric mucosa in healthy and infected mice [16, 27]. In the absence of functional CD11b⁺Ly6G⁺Ly6C^int neutrophils, our CpG-treated mice controlled bacterial growth by recruiting higher percentages of Ly6C^highCCR2⁺ monocytes in both spleen and brain than buffer-treated animals. CpG pre-conditioning also decreased the percentage of CD45⁺CD4⁺CD3⁺CD25⁺FoxP3⁺ regulatory T cells in spleen and brain 42 h after infection. Similarly, during acute infection with Toxoplasma gondii and Listeria monocytogenes a transient loss of Treg cells caused by IL-2 insufficiency was essential for initiation of Th1 responses and host protection against infection [31]. Furthermore, CpG ODN 2006 downregulated the proportion of Treg cells in peripheral blood mononuclear cells from patients with non-small cell lung cancer [32].

Chemokines are key players in leukocyte recruitment by endothelial cells upon infection. In patients with bacterial meningitis, CSF levels of chemokines including MIP-1α are elevated [33]. Chemokines can be released into the CSF by different cell populations, including microglial cells, resident macrophages and migrating leukocytes depending on the stage of disease [34]. MIP-1α release in human brain microvessel endothelial cells is negligible under physiological conditions but can be upregulated after stimulation with IL-1β or LPS [35]. We found higher levels of MIP-1α in both cerebellum and spleen homogenates of CpG-treated infected animals compared to infected control mice. Peak MIP-1α spleen concentrations were found at 42 h but the levels were elevated until 14 days after infection. We recently reported a strong correlation between E. coli K1 loads and increased levels of IL-1β, IL-6, KC (rodent homologue of growth-related oncogene-α/CXCL1), and macrophage inflammatory protein 2 (MIP-2/CXCL2) in cerebellum homogenates of untreated neutropenic wt animals [22]. In the current work, we did not find a significant correlation between the bacterial burdens and the levels of MIP-1α and IL-12/IL-23p40 in cerebellum homogenates of CpG-treated mice that survived 2 weeks after E. coli challenge.

The administration of CpG ODN caused weight loss even in uninfected mice. This weight loss was absent in neutropenic TLR9-deficient mice, suggesting that it corresponds to a CpG ODN-induced sickness behavior [25] which has been observed in animal and clinical studies [36‐38]. In vitro, exposure of co-cultures of neurons with microglial cells to CpG ODN caused neuronal injury, which often started with the attack of microglial cells on axons [39]. In this study, CpG ODN strongly increased survival in spite of the initial weight loss induced by the CpG ODN treatment.

An intraperitoneal injection of CpG ODN protected neutropenic mice against intracerebral infection with E. coli K1 by recruiting higher amounts of inflammatory monocytes into the CNS, decreasing bacterial burdens in the cerebellum and reducing the degree of bacteremia. Therefore, CpG ODN activated the innate immune response not only by influencing the activity of phagocytes of the peritoneum and with direct contact to the circulation, but also those situated in deep compartments such as the CNS. By this way, the systemic administration of CpG ODN may help to prevent CNS infections in immunocompromised individuals even after direct inoculation of bacteria into the intracranial compartments which can occur with open head trauma and after surgery, including placement of an external ventricular drain. Supporting our findings, Marabelle et al. recently reported that the addition of CpG to low doses of anti-CTLA-4 and anti-OX40 (two antibodies directed against surface markers of tumor-specific Treg cells) injected locally in a peripheral tumor site in the rat expanded anti-tumor responses to distant tumor sites including the brain [40]. The use of CpG DNA in vaccinations offers several advantages [41]: (1) it can be synthesized with high purity; (2) CpG motifs can exert their immunostimulatory effects either as part of a vaccine or when delivered alone; (3) the presence of CpG motifs contributes to DNA vaccination success rates [42, 43]; and (4) Th1 adjuvant properties of CpG DNA enable successful immunization in neonates who are difficult to immunize because of their immature innate immune system [44]. In accordance with this concept, Soogard and collaborators reported the improvement of the immunogenicity of the 7-valent pneumococcal conjugated vaccine in HIV-infected adults by the addition of a variant of CpG (CPG 7909, Coley Pharmaceutical Group) [45]. For these reasons, we suggest conducting a clinical trial with CpG ODN in immunocompromised patients with a high risk of CNS infections.