The Role of Tnf and Soluble Tnf Receptors (TNFR1/2) as Therapeutic Targets for Inflammation Secondary to Intracerebral Stroke and Hemorrhage

1. Introduction

Cytokines are part of a network of innate and acquired immune responses to brain injury and disease. Tumor necrosis factor (TNF) is a well-studied cytokine that contributes to neuroinflammation in a range of conditions including subarachnoid hemorrhage (SAH), intracerebral hemorrhage (ICH), and ischemic stroke (IS). Despite being recognized as a pro-inflammatory cytokine, TNF signaling is linked to diverse outcomes such as cell survival and proliferation. This multifaceted impact is attributed to the distinct downstream effects of TNF receptors, where TNF receptor 1 (TNFR1) activation yields detrimental effects, while TNF receptor 2 (TNRF2) activation can result in beneficial effects. This comprehensive review aims to elucidate the complexities and crosstalk inherent in TNF signaling as well as analyze current therapeutic interventions designed to limit the threat of excessive inflammation and neuronal cell death associated with TNFR1 activation1. Since TNF can signal in its membrane-bound and soluble conformations, current therapeutic methods include neutralizing antibodies, TNFα converting enzyme (TACE) antagonists, and soluble receptor2–4. Moreover, this review demonstrates the potential of novel treatments such as Xpro1595 and evaluates certain limitations in the current research paradigm. Overall, this review highlights the potential of selectively targeting TNFR1 signaling pathways as a novel therapeutic approach. It also emphasizes the critical need for future research utilizing appropriate models to ensure the accuracy and replicability of results in human trials.

2. Materials and Methods

The literature in this review included peer-reviewed publications published between the years 1989 to 2022. The online databases utilized in this literature review included Embase, PubMed, Google Scholar, and Biogen Pipeline patent US20150239951A1. Pertinent search filters included TNF, receptors, neuroinflammation, ischemia, stroke, and cerebral hemorrhage. Selected literature consisted of peer-reviewed publications written in English, spanning preclinical and clinical case studies, cross-sectional studies, retrospective studies, meta-analyses, and systematic literature reviews. Exclusion criteria comprised non-peer-reviewed articles, conference abstracts, editorials, letters, and commentaries. Additionally, studies beyond the specified time period, not in English, solely focusing on cell lines, lacking clear methodological details, or not directly addressing TNF and its receptors in neurodegenerative diseases and injuries were excluded. These criteria were applied to ensure the inclusion of relevant and high-quality studies aligned with the review’s objectives. The collected data were tabulated into clinical and preclinical reviews. Data variables gathered from each experimental publication, when applicable, included the species tested, population size, average age, sex predominance, study duration, drug administration, anatomical outcomes, and functional outcomes.

3. TNF Receptors

The majority of known cellular responses to TNF are mediated by either TNF receptors 1 (TNFR1) or 2 (TNFR2). Existing literature has focused on TNFR1 interactions due to its higher binding affinity and variable responses compared to TNFR2. TNFR1, expressed in all cell types, can be activated by both membrane-bound TNF (tmTNF) and soluble TNF (sTNF) [1]. TNFR2 is expressed primarily on endothelial cells and predominantly binds to tmTNF. While both variants contribute to cerebral-protection, TNFR2 has demonstrated some capacity in inflammatory reactions as well [2]. Despite this, most cells express both forms of the receptor, and each lead to unique responses [3].

3.1. TNFR1

To initiate a signal transduction event, homotrimeric TNF must bind to the cysteine-rich extracellular domains of membrane-bound TNFR homodimers [4]. The binding process can be inhibited by soluble variants of TNF receptors (sTNFR), which compete with membrane-bound receptors for ligands or directly bind to the N-terminal pre-ligand binding assembly domain, preventing the formation of active receptor trimers [5]. Upon binding to its ligand, TNFR1 homotrimerizes and interacts with intracellular signaling proteins through its death domain (Figure 1). This leads to the formation of complex I, consisting of TNFR1-associated death domain protein (TRADD), receptor-interacting protein kinase-1 (RIPK1), and TNF receptor-associated factor 2 (TRAF2) or TRAF5. Complex I can activate JUN NH₂-terminal kinase (JNK), p38, and nuclear factor κB (NFκB) transcription by parallel recruitment of TAK1 and the IKK complex [4]. TAK1 activates MAP kinases (MAPK) such as JNK, p38, and ERK, as well as ceramide/sphingomyelinase pathways by recruiting cellular inhibitor apoptosis proteins 1 and 2 (cIAP1/2). The IKK complex, comprising IKKα, IKKβ, and NFκB essential modulators (NEMO), promotes Jak1, Jak2, STAT3, and STAT5 activation, freeing RelA/p50 for nuclear translocation and the transcription of NFκB inflammatory genes [6]. These pathways can induce cytoprotective or apoptotic effects depending on the duration and intensity of stimulation. However, canonical activation of NFκB via RelA/p50 is typically proinflammatory in nature [4].

Moreover, internalization of the activated TNFR1 complex I results in the dissociation of some adaptor proteins, leading to the binding of the FAS-associated death domain (FADD), recruitment of pro-caspase 8, and formation of executioner caspases, ultimately resulting in cell death [7]. The structure of this new complex, denoted as complex 2, is subject to conflicting evidence, as it may retain proteins such as RIPK1, TRADD, or TRAF2, and gain FLIP-c [6,7,8]. Figure 1 depicts one proposed structure of complex II [4].

Activation of caspase 8 leads to the degradation of RIPK1 and RIPK3. However, insufficient concentrations of caspase can result in the accumulation of RIPK1 and RIPK3 in a fibrous structure that promotes programmed necrosis under certain conditions [9]. In chronic inflammatory conditions, TNFR1 has also been shown to inhibit TNFR2 signaling by activating TACE through the iRhom2 and p38 pathways [10]. As TNFR2 is only capable of activation via membrane-bound TNF, TACE cleavage of TNF results in decreased TNFR2 activation.

3.2. TNFR2

In contrast to TNFR1, TNFR2 lacks a death domain but can directly interact with TNF receptor-associated factors 1 and 2 or 3 (TRAF1 and TRAF2/3) using an amino acid motif near the C-terminus (Figure 1). The recruitment of TRAF2/3 to the activated complex results in the production of NF-κB-inducing kinase (NIK), which is typically bound to TRAF2/3 or cIAP1/2 in an inhibitory conformation [4]. NIK is capable of non-canonical NFκB activation, leading to the activation of IKKα and subsequent nuclear translocation of RelB/p52, responsible for the transcription of genes relevant to cell survival and proliferation [11,12,13]. This non-canonical activation generally has a longer duration compared to canonical NFκB activation via TNFR1.

Despite lacking a death domain, TNFR2 has also demonstrated the capability to canonically activate NFκB. The exact mechanism is unknown, but activation is still dependent upon TAK1 and IKKβ (Figure 1) [14]. Dissociation of the TRAF1/2 complex can also bind to cIAP1 and be transported to the endoplasmic reticulum, triggering stress and promoting cell death [6]. Significant crosstalk exists between TNFR2 and TNFR1, as the activation of either complex utilizes common proteins. For instance, TNFR2 activation depletes available TRAF2, inhibiting the formation of the TNFR1 complex and promoting non-canonical NFκB activation, as well as the activation of apoptotic factors [4].

4. Inflammatory Response to Stroke

Stroke ranks as the fifth leading cause of death and the primary cause of disability in the United States [15]. At a rate of 40 stroke incidents occurring every second, there are 800,000 new strokes annually in the United, resulting in prolonged cognitive and physiological dysfunction[16]. A stroke is commonly defined as a vascular neurological deficiency lasting longer than 24 hours, and can be classified as ischemic or hemorrhagic based on the mechanism, with ischemic stroke constituting 85% of all acute stroke incidents [16,17].

Ischemic stroke (IS), also referred to as a cerebrovascular accident, occurs due to the occlusion of cerebral vasculature, blocking blood flow and oxygen to brain regions and causing the breakdown of the blood-brain barrier (BBB)[18]. The lack of blood flow suspends the membrane potentials of neurons and glial cells, interrupting energy-dependent mechanisms and ion pumps in cell membranes, leading to the release of harmful extracellular components and cell death [19]. After an IS, the body’s natural response involves releasing proinflammatory cytokines such as interleukins IL-1 and IL-6, along with TNF. These cytokines play a crucial role in regulating the size of cerebral infarctions caused by IS, and their levels increase following focal ischemic injury [20,21]. It is crucial to document the modulation of these cytokines over time to understand the mechanism, targets upon their release, and identifying specific treatments at various time points. Existing evidence, utilizing reverse transcription PCR, indicates an 8-fold increase of TNF mRNA-expressing microglial cells and microglia-macrophages as early as 4 hours following neuronal ischemia [22]. TNF mRNA levels typically rise early, at the 4-hour mark post focal ischemia, compared to IL, which instead demonstrates a late increase at 96 hours [22,23].

The occlusion of blood flow to regions in the brain leads to a measurable area of lethally damaged and hypoxic tissue known as the infarct core, surrounded by a reversibly damaged tissue zone that is known as the penumbra [20,24]. Clinical aspects of infarction examined in experimentation and clinical studies include infarct volume, location, and quantity at variable locations. Many studies have explored the measured expression and concentration of TNF mRNA and cytokines concerning these factors to implement therapies controlling the concentration of these highly regulated and pleiotropic ligands. Higher levels of TNF following IS have been suggested to correlate with larger infarct volumes. In a study involving transgenic rats overexpressing the murine TNF gene and subjected to Middle Cerebral Artery Occlusion (MCAO), the transgenic rats exhibited a higher infarct volume at 24 hours (p≤0.05) and 3 days (p≤0.01). Within the first 10 min of MCAO, cortical perfusion was reduced in the transgenic rats (p≤0.05), leading to increased cellular death and DNA denaturation [25]. Preclinical studies have also demonstrated an increase in the concentration of microglia surrounding the penumbra in a time-dependent manner, starting at 12 hours up to 1 day (p≤0.01) and 3 days (p≤0.005) following transient MCAO [18]. NFκB was also evaluated in this study and was observed to be activated 12 hours after global ischemia in both wild-type and TNF-deficient mice [18]. With increasing levels of TNF under most experimental conditions correlated with cell death and larger infarct volumes, potential therapies for limiting the long-term cognitive and physiological effects of IS involve introducing specific receptors that will bind to and reduce the toxic inflammatory effects [1,26]. Table 1 and Table 2 present pre-clinical studies analyzed regarding the effects of TNF stimulation and current clinical studies regarding TNF activation in the context of stroke, respectively.

4.1. sTNFRs and Stroke

Soluble receptors, investigated for their potential to prevent cytokines from binding to membrane receptors and inhibit their biological activity, offer advantages such as high ligand specificity and low immunogenicity [58]. However, they may exhibit lower affinities and half-lives as compared to membrane receptors, making the therapeutic response of sTNFR dependent on TNF concentration as well as TNFR1 and TNFR2 [59,60].

In stroke therapy, soluble cytokine receptors serve as natural antagonists. The increase in sTNFR is achieved through multiple pathways triggered by TNF, interleukins, transforming growth factors, and interferons. Proteolytic release of TNFR1 is regulated mainly by metalloprotease ADAM8, while TNFR2 release is mediated by TACE in a process known as ectodomain shedding [1]. Studies monitoring TNF regulation in CNS diseases, including stroke, have shown that ADAM8 mRNA expression increases in affected regions proportionally to TNF levels. To measure the effect of sTNFR1 levels on neuronal survival, a comparative analysis of ADAM8 deficient neurons and wild-type neurons was conducted in the presence of exogenous TNF. ADAM8 deficiency increases neuronal sensitivity to varying TNF concentrations, suggesting that shedding of TNFR1 has a protective effect on primary neuron survival [39].

Experimenters have explored alternative methods of anti-cytokine therapy to protect against ischemic damage, including TNF binding protein (TNF-BP) and monoclonal antibodies (mAb). TNF-BP, formed by the proteolytic cleavage of membrane TNF receptors, demonstrated therapeutic capability to modulate TNF toxicity [61,62]. TNF itself induces the release of TNF-BP, self-modulating the toxic effects related to increased concentrations. In experiments examining TNF upregulation in the brain following ischemic injury, TNF-BP administration immediately or 60 minutes after MCAO led to smaller infarct volumes, with animals receiving TNF-BP showing a 25% reduction in infarct volume up to 2 weeks following treatment compared to control groups [33,63,64,65]. Studies have investigated the protective role of mAb in the context of TNF upregulation in the brain following ischemic injury. Rats were administered exogenous TNF before transient or permanent MCAO. Despite TNF injection leading to increased infarct size, prior mAb injection significantly reduced hemispheric infarct size by 8%. Treatment with mAb prior to exogenous TNF also alleviated cognitive deterioration and mitigated neurological scores in the TNF groups [32].

5. TNF as a Therapeutic Target

The release of TNF by microglia and a subsequent pattern of increased sTNFR suggests an anti-inflammatory role of these receptors as they scavenge and decrease the serum concentration of this cytokine. Currently, two commonly used therapeutic options targeting TNF are etanercept and infliximab. These drugs share similar mechanisms of action, aiming to decrease inflammation caused by the pro-inflammatory release of TNF. Etanercept, marketed as Enbrel^®, functions as an artificial sTNFR and consists of two TNFR2 receptors joined to the Fc portion of human IgG as demonstrated in Figure 2 [66]. It exhibits significantly greater binding efficiency and half-life compared to sTNFR2 (Table 3). The protective role of soluble receptors in physiological and neurological outcomes has been examined through controlled preclinical experiments. Rats subjected to fluid percussion injury (FPI) were treated with etanercept every 12 hours for 3 days. In the etanercept-administered groups, there was a significant decrease in cerebral ischemia area (p<.01), microglial production of TNF (p<.01), and neurological and motor dysfunction (p<.05) [67]. Additionally, etanercept therapy led to a significant reduction in glial scar formation, a major inhibitor of neuronal regeneration. This study suggests etanercept attenuates microglia-associated TNF expression without changes to astrocyte or neuron expression following TBI [67,68]. In a retrospective study of patients with chronic neurological deficits post-TBI, etanercept demonstrated clinical improvements in motor function, spasticity, walking, cognition, sensation, aphasia, and pain, even when administered 115 to 160 months after traumatic brain injury⁸⁴.

Additional TNF-binding drugs include infliximab, a chimeric monoclonal anti-TNF antibody that has been used to reduce inflammation associated with a variety of chronic diseases and conditions [69,70]. Marketed as Remicade^®, infliximab functions to target TNF, attenuating inflammation by binding to Fc receptors to decrease the activation of monocyte and caspase-3-induced oxidative stress [71,72]. While infliximab has been widely used for chronic diseases, its potential therapeutic role in neuroinflammation, like etanercept, warrants exploration. However, limitations associated with TNF inhibitors, such as etanercept or infliximab, have prompted research for novel TNF blockers. While Etanercept offers therapeutic benefits, unintended side effects have been observed, leading to decreased long-term functional outcomes. The drug’s inability to differentiate between sTNF and tmTNF poses challenges, inhibiting baseline activity and causing adverse effects [69]. Exogenous TNF blockers, including etanercept, may increase the risk of serious infections and cancer according to a meta-analysis of clinical trials comparing TNF-binding medications like etanercept and infliximab [73]. Etanercept, in placebo-controlled trials, demonstrated a 35% infection rate over the therapy course, associated with serious bacterial, viral, and fungal infections [69]. Infliximab has similar associations and a close link to tuberculosis development, often reflecting latent infection reactivation. Give the limitations and potential side effects of current TNF blockers like etanercept and infliximab, a second-generation drug, XPro1595, has been developed to address these issues and enhance patient outcomes.

5.1. Proposed Second-Generation Therapies

A novel sTNF inhibitor, XPro1595, has demonstrated success in inflammatory disease trials, notably for its ability to penetrate the BBB [74,75]. Also known as INB03, XPro1595 inhibits TNFR1 activation by forming a non-viable heterotrimer with sTNF as shown in Figure 2 [76]. Xpro1595 mimics the monomeric proteins in sTNF homotrimers and freely exchanges native proteins in a 1:2 or 2:1 ratio. XPro1595 is dominant-negative mutant of TNF, with specific mutations (A145R and Y87H) at the TNFR binding site, selectively inhibiting sTNF activity at TNFR1 receptors [77]. As mentioned earlier, TNFR1 is responsible for downstream neuroinflammatory effects and subsequent cognitive impairments. Unlike other TNF inhibitors, XPro1595 selectively inhibits the effects of TNFR1, ensuring continued TNFR2 function, crucial for neurogenesis and myelination. The newly formed XPro1595 heterotrimer does not interfere with the affinity of transmembrane TNF for TNFR1 or TNFR2. Allowing the continued function of TNFR2 while blocking TNFR1 differentiates XPro1595 from other TNF inhibitors and contributes to the drug’s increased effectiveness. To enhance its properties, XPro1595 has been modified via PEGylation, resulting in XENP1595, with increased half-life and reduced immunogenicity [77].

In preclinical studies, XPro1595 administration in mice subcutaneously 1-hour post-TBI reduced decreased glial activity, prevented dendritic degeneration, and improved spatial learning and memory [75]. Similar beneficial results have been demonstrated in focal cerebral ischemia models where topical Xpro1595 decreased infarct volumes 1 and 3 days post MCAO, outperforming Etanercept [40]. Additionally, XPro1595 treatment has shown promise in mitigating the effects of obesogenic diets in a preclinical model of Alzheimer’s disease and increased BBB permeability, attenuating inflammation and improving cognitive function [78].

A clinical trial evaluating XPro1595 treatment in patients with Alzheimer’s disease was completed in 2021. The open-label dose-identification study of XPro1595 in Alzheimer’s disease patient aimed to evaluate the drug’s safety, efficacy in reducing neuroinflammation, and impact on cognitive function. Study results are yet to be published but are available in online videos through the study sponsor’s website [79]. Results of the phase 1b study showed that weekly treatment with XPro1595 for 3 months in patients with signs of neuroinflammation demonstrated decreased white-matter free water and CSF inflammatory proteins as well as increased axonal fiber density and remyelination [79]. Prompted by these initial results, two phase II studies are currently recruiting with the same primary outcomes in a larger patient population to assess the long-term safety, tolerability, and efficacy of XPro1595.

Proof of concept trials for novel antibody-drug conjugates such as ABBV-3373 may be a potential alternative to current-generation anti-inflammatory therapies such as adalimumab. Composed of a glucocorticoid receptor modulator conjugated with an anti-TNF monoclonal antibody, ABBV-3373 was able to significantly reduce DAS28 (CRP) scores in rheumatoid arthritis patients with greater efficacy when compared to adalimumab (p<.022) [80]. It should be noted that the study lacked sufficient sample size to draw statistical conclusions past 12 weeks and that patients treated with ABBV-3373 received a placebo past 12 weeks but demonstrated similar responses to the 24-week adalimumab regimen. These results warrant continuing research and the potential application of antibody-drug conjugates in a cerebroprotective role.

6. Conclusions

This review investigates the role of TNF and potential therapeutic targets in cerebrovascular and neurodegenerative disease. TNF, a major inflammatory cytokine, plays a crucial role in neuroinflammation associated with these conditions. It can signal through its membrane-bound form or the homotrimeric sTNF, formed by cleaving tmTNF from the cell membrane via the proteolytic enzyme TACE. TNF activates two receptors, TNFR1 and TNFR2, with promoting detrimental effects, including cell death and increased inflammation, and TNFR2 demonstrating beneficial effects, such as enhanced neuronal tissue growth and CNS autoimmunity. Selectively targeting the sTNF and TNFR1 signaling pathway may facilitate new mechanisms to combat neuroinflammation.

The generation of sTNF and sTNFRs involved the proteolytic cleavage of tmTNF and TNFR1/2, respectively. Metalloproteinases, particularly ADAM8, preferentially cleave TNFR1 without affecting TNFR2, providing a potential therapeutic avenue to modulate TNFR1 signaling. Studies suggest a negative feedback mechanism in which TNF signaling leads to ADAM8 activation and therefore TNFR1 shedding. TNFR2 appears to be mainly cleaved through TACE, and further research is required to explore similar feedback mechanisms in sTNFR2 production.

The myriad of chemical and biological reactions that TNF can activate makes it an ideal target for therapeutic interventions. Several drugs, including etanercept and infliximab, have been developed to target TNF signaling. However, these drugs can cause unwanted side effects due to the indiscriminate inhibition of TNFR2 activation. The second-generation TNF inhibitor XPro1595 has been designed to selectively neutralize sTNF without affecting tmTNF or TNFR2, aiming to maintain pathways of neurogenesis and innate immunity [81]. Early results in mice with TBI demonstrated positive effects on dendritic plasticity, reduced neurodegeneration, and glial reactivity [75]. Clinical trials of the drug have demonstrated decreased neuroinflammation and neurodegeneration in Alzheimer’s Disease patients and the ability of XPro1595 to cross the BBB has aided in attenuating inflammatory biomarkers in other neurodegenerative diseases [79].

Limitations in using mice models for TNF inhibitors include the unresponsiveness of mouse TNF to human TNF inhibitors [82]. Humanized mouse models expressing human TNF (hTNF) or human TNFR (hTNFR) are utilized to study the molecular mechanisms of these cytokines [83]. However, overexpression of hTNF can result in spontaneous autoimmune diseases, prompting the use low-copy or tissue-specific models [83]. Humanized mice with hTNFR2 LoxP sites have been developed to address these challenges [84]. Recently, there has been a rise in various humanized mouse models used to study the TNF/TNFR1/2 mechanism [85,86]. Future research should leverage these humanized mouse models to test TNF inhibitor drugs, ensuring accuracy and replicability in human trials.