The present study provided evidence for the link between MOH and leaky gut by demonstrating an elevated occludin and VE-cadherin levels in the systemic circulation, increased LBP levels indicating LPS leak into the bloodstream and a significant inflammatory response in piroxicam induced MOH model. Five-week exposure to oral piroxicam elicited pain related behavior consistent with MOH. Reduced periorbital mechanical withdrawal thresholds, increased head and face grooming, freezing, head shake behavior and increased serum CGRP levels were compatible with trigeminal pain in the presented MOH model in female rats.

Similar to our study, spontaneous and evoked pain behavior have been shown in MOH animal models. Four-week mefenamic acid exposure has been shown to result in decreased periorbital mechanical withdrawal thresholds, elevated number of head shakes, freezing and grooming in rats [13]. Eleven day sumatriptan injection triggered mechanical allodynia in both periorbital regions and paws in mice [27]. Repeated sumatriptan injection induced activation of microglial cells and increased purinergic receptor P2X7 (P2X7R) expression and activation of NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome in trigeminal nucleus caudalis [27] and inhibition of P2X7R and NLRP3 inflammasome attenuated sumatriptan induced mechanical allodynia [27]. Six doses of lasmitidan resulted in bright light stress or nitric oxide induced cutaneous allodynia in the periorbital and hindpaw regions [28].

Elevation of LBP levels which is a reliable biomarker of increased LPS in the bloodstream along with tight junction protein occludin and adherens junction component VE-cadherin suggests a disruption in the intestinal integrity and permeability change leading to leaky gut. The increased levels of junctional proteins can be detected in vascular permeability changes associated with BBB disruption and parenchymal damage [29, 30]. VE-cadherin also serves as a biomarker for gut-vascular barrier integrity, that limits passage of intestinal toxins into bloodstream [31]. In our study, LBP, occludin and VE-cadherin levels were positively correlated with each other. Also, serum VE-cadherin and occludin levels showed strong correlations with inflammatory cytokines, serum CGRP levels and pain behaviors. These findings indicated that trigeminal nociceptive responses and headache related behavior were correlated with intestinal barrier disruption and inflammatory responses.

The crucial role of gut-brain axis in neurological disorders including migraine headache has been acknowledged. There is a bidirectional relationship between gut permeability and inflammation [11]. Increased gut permeability stimulates inflammatory response in the host by leakage of LPS into the bloodstream. The induction of inflammatory cascade and release of pro-inflammatory cytokines upon LPS exposure further increases gut permeability [11]. Several pro-inflammatory cytokines such as IL-6 and IL-1β are also nociceptive in the trigeminovascular system and take a part in the headache development [32]. Disrupted intestinal barrier and associated inflammation has not been studied in MOH. However, in a translational NTG migraine model, antibiotic administration induced prolonged pain was blocked by tumor necrosis factor alpha (TNF-α) targeted interventions [33]. Increased NTG provoked pain in germ free mice suggested a role of microbiome in headache disorders [33].

Gut barrier integrity has gained importance in diverse systemic and neurological diseases [25, 26, 31]. Diagnostic tests to measure intestinal mucosal injury are challenging for the patients, therefore, potential blood biomarkers are used for indirect assessment of intestinal barrier function [34]. Bacterial endotoxins must be strictly kept in intestinal lumen, and the detection of LPS in the bloodstream indicates intestinal hyperpermeability and leaky gut. The latter is supported by the detection of elevated serum levels of tight/adherens junction proteins that are components of the intestinal barrier. In line with this notion, increased blood level of LPS was detected in patients with irritable bowel syndrome (IBS) [32].

LPS in the systemic circulation is recognized and bound by a soluble acute phase protein called LBP [35, 36]. A significant increase in serum LBP levels in MOH group was consistent with the presence of LPS in the blood circulation. LBP has been shown to be a reliable biomarker for the presence of LPS in the circulation and the assessment of intestinal permeability with its longer half-life and low intraindividual variability, independent of age, sex, and body-mass index [34, 37].

There is also a sexual dimorphism in intestinal permeability and women are more vulnerable to perturbations in intestinal permeability by NSAIDs [15]. The disturbance in intestinal permeability usually recovers 4–6 weeks after discontinuation of NSAID [15]. The latter may be one of the mechanisms underlying the clinical observation of headache improvement starting 1–2 months after the cessation of the analgesics in MOH patients [1, 2].

Intercellular tight junctions and adherens junctions are important structures for building the epithelial barrier [38, 39] and leaky gut syndrome occurs with loosening of these structures in the intestinal wall [39, 40]. Occludin is a key tight junction protein and serum occludin level can be a surrogate biomarker for intestinal permeability and gut disorders as well as BBB disruption in neurological disorders [29, 30]. VE-cadherin, an adherens junction molecule connecting cells is involved in the regulation of paracellular permeability [41]. The breakdown of endothelial adherens junction proteins is associated with disruption of vascular integrity in various organs [42]. Serum VE-cadherin was shown to be significantly elevated in severe sepsis [43]. Thereby we have proof that not only LPS is present in the blood circulation but also both tight junction and adherens junction structures are disrupted and occludin and VE-cadherin levels are increased in the blood circulation in MOH. All these findings indicate the presence of leaky gut in our MOH model.

LPS when present in the circulation in low dose, called metabolic endotoxemia is associated with metabolic disorders such as diabetes mellitus and obesity [44]. LPS elicits inflammatory response by activating TLR-4 receptors. Interestingly the same receptor can also be activated by HMGB1 [20]. Both LPS and HMGB1 are critical actors in innate immune response [45]. Upon cell activation or cell damage, HMGB1 translocates from the nucleus to the cytoplasm and is released into extracellular space. HMGB1 is a potent pro-inflammatory molecule and induces nociception through its TLR4 or RAGE receptors in the trigeminal neurons [46]. LPS and HMGB1 have a mutual interaction that augments the inflammatory response including IL-6 [20, 45]. HMGB1 has been proposed to mediate inflammation and nociception in preclinical models or in secondary headaches such as in coronavirus disease 2019 (COVID-19) [21, 22]. HMGB1, NLRP3 and IL-6 levels were positively correlated with headache in patients with COVID-19. Brain HMGB1 level was positively correlated with serum CGRP level and trigeminal pain behavior in our study. Therefore, MOH seems to be accompanied by innate immune response as brain HMGB1 and serum LBP levels were increased.

An association of inflammatory cytokines and migraine was revealed by a recent meta-analysis, that concluded higher levels of pro-inflammatory cytokines IL-6, IL-8 and TNF-α were detected in migraine patients compared to control subjects [18]. Elevated IL-6 levels were shown to be associated with headache chronification in migraine patients [47]. Increased serum IL-6 levels were not only related to headache chronification but also with headache severity and drug unresponsiveness in COVID-19 headache which is also a secondary headache [21, 48]. The significant role of IL-6 as a both pro-inflammatory cytokine and nociceptive molecule in trigeminal system is implicated in preclinical NTG models [17]. IL-17 is a key pro-inflammatory cytokine involved in neuroinflammation, cognitive dysfunction and gastrointestinal disorders such as IBS. IBS symptoms were shown to be higher in MOH patients and subjective cognitive complaints were increased in MOH patients with IBS [49]. Both serum and brain levels of IL-17 were significantly increased in MOH group and were all correlated with biomarkers of leaky gut and pain behavior. Higher levels of cerebral cortical HMGB1 and IL-17 cortex indicate that the inflammatory response in the central nervous system is associated with MOH and may contribute to the maintenance of chronic pain. Besides NTG model [19], NSAID induced MOH is also related to pro-inflammatory cytokine IL-17 increase in the brain. Chronic paracetamol administration has been also shown to increase inflammatory cytokine expressions (IL1-α and TNF-α) in hippocampus however no association with pain was demonstrated in the same study [50].

CGRP is a prime molecule involved in migraine headache signaling and has a role in initiating the local inflammatory response [51]. CGRP levels are found to be high in the systemic circulation in migraine attacks and antagonism of circulating CGRP is a successful treatment in migraine patients [52, 53]. Regarding the above facts, CGRP levels were correlated with trigeminal pain behavior in this MOH model. CGRP also correlated positively with HMGB1 and IL-17 levels in the brain and LBP, VE-cadherin, occludin and IL-17 levels in the serum.

The present study has several limitations. One of the limitations was that an extensive inflammatory response panel including TNF-α and anti-inflammatory cytokines was not studied. Additionally, male rats were not used in the experiments to decide on the sexual dimorphism in piroxicam induced MOH model. The peripheral and central mechanisms of trigeminal nociceptive system involvement remain to be investigated. Another limitation is that the disruption of intestinal barrier in the GI tract was not studied. Translating these findings to humans is necessary and the results remain to be confirmed in patients with MOH.

We showed for the first time that MOH induced by chronic use of oral piroxicam, is linked to leaky gut as LBP, tight junction and adherence junction protein levels in the blood circulation were elevated. We used only female rats to model MOH for clinical relevance as both MOH and chronic migraine patients are predominantly female and NSAIDs have strong perturbations in intestinal permeability in women. All animals were tested outside the estrus and pro-estrus phase, eliminating their confounding effect on trigeminal pain response. Moreover, other strengths of our study are the use of a NSAID for MOH induction since they are the most common analgesics overused in MOH patients and the oral use of piroxicam to increase clinical congruency.