Pregabalin inhibits in vivo and in vitro cytokine secretion and attenuates spleen inflammation in Lipopolysaccharide/Concanavalin A -induced murine models of inflammation

Immune system alteration has been implicated in the pathogenesis of chronic pain conditions, epilepsy and generalized anxiety disorder. Targeting cytokines has recently been proposed for the management of such conditions. Pregabalin (PGB) is an antiepileptic agent used for the management of these conditions. However, little is known about its immunomodulatory effects on cytokine secretion in vivo and in vitro. Hence, a mitogen (Lipopolysaccharide [LPS] or Concanavalin A [ConA])-induced murine model of inflammation was used to investigate the effect of PGB on in vivo and in vitro IL-1β, IL-6, TNF-α and IL-2 cytokine secretion using ELISA. In addition, PGB effect on spleen histology, as a lymphoid organ, was examined. Our results revealed that PGB significantly inhibited the secretion of ConA-induced IL-6 secretion, basal and ConA-induced TNF-α and IL-2 secretion in splenocytes in vitro. In vivo, PGB inhibited basal and LPS/ConA-induced IL-6 and TNF-α secretion in addition to LPS-induced IL-1β and ConA-induced IL-2 secretion. Moreover, PGB attenuated mitogen-induced inflammatory changes in the spleen. These findings provide an evidence of the anti-inflammatory properties of PGB on cytokine secretion and lymphoid organ inflammation. This might give insights into the role of PGB in the management of the inflammatory state in PGB-indicated conditions.

PGB preparation. A stock solution of 10 mg/ml PGB was prepared in normal saline (NS) for in vivo experiments and in phosphate-buffered saline (PBS) for in vitro experiments.
Dosage and treatment protocol for in vivo studies. Three groups of mice (n = 18) were either treated with NS intraperitoneally (i.p) (control group) or with 8 or 16 mg/kg of PGB (i.p), corresponding to 1-and 2-times the maximum human therapeutic dose, respectively (Fig. 1a). Mice were incubated for 1 hour with NS or PGB before dividing each group into 3 sub-groups (n = 6), which were either treated with 0.1 mg/kg of LPS (i.p) or 10 mg/kg of ConA (i.p) or were left without additional treatment. For serum cytokines quantification, mice were then incubated for additional 2 hours in the presence or absence of the mitogen before being anesthetized with diethyl ether inhalation prior to blood collection. For the histological examination of the spleen, mice in the treatment groups continued to receive a daily injection of PGB -at the indicated doses-or NS for additional 5 days before being anesthetized with diethyl ether inhalation and scarified for spleen extraction.
Serum preparation for cytokines quantifications. Blood was collected from retro-orbital plexus into silica gel-containing plain tubes. Serum was prepared by leaving the blood to clot at room temperature (RT) for 30 minutes followed by centrifugation at 3000 rpm for 10 minutes at 4 °C. Sera were frozen at −80 °C immediately after collection. ELISA measurement of IL-6, TNF-α, IL-1β and IL-2 levels was performed as per manufacturer's protocols (ThermoFisher Scientific, USA). Ranges of detection were 4-500; 8-1000; 8-1000; 2-200 pg/mL, respectively.
Histological examination of the spleen. Collected spleens were immediately fixed in formalin. Sections of 5-µm thickness were prepared using microtome and stained with Hematoxylin & Eosin (H&E). Histologic changes were examined by a blinded pathologist using Olympus light microscope.

Preparation of Single splenocyte cell suspension (SSCS). Spleens were collected in RPMI-1640
medium supplemented with 10% FBS under aseptic conditions. Spleens were minced between the frosted edges of two microscopic slides and the cells were then pooled and centrifuged at 1200 rpm for 7 min at 4 °C. The red blood cells (RBC) within the pellets were then lysed at RT for 3 min using RBC lysis buffer as previously described 27,28 . Splenocytes were then suspended in 1 mL of RPMI-10% FBS and viable splenocyte count was determined using trypan blue dye exclusion method 29 . Isolation of peritoneal macrophages (PMs). Peritoneal cells were collected under aseptic conditions through three-time peritoneal lavage, each with fresh 5 ml of cold PBS. Cells were pooled, washed with cold PBS www.nature.com/scientificreports www.nature.com/scientificreports/ and then suspended in complete medium and plated for two hours for PMs to adhere. Non-adherent cells were washed twice with PBS 30 .
Cell cultures. Two million cell/ ml of splenocytes or peritoneal cells were cultured in complete RPMI-1640 medium or complete high-glucose DMEM medium, respectively. Both media were supplemented with 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin. Cultures were maintained at 5% CO 2 / 37 °C. For cell proliferation assay and in vitro cytokine secretion assessment, splenocytes were treated immediately while peritoneal cells were left for 2 hours to allow the adherence of PMs. Cultures of both cell types were pre-treated for 1 hour with PGB at a final concentration of 1.5, 3, 6, 10, 30 or 60 µg/ml, or were left untreated (Fig. 1b). Splenocytes were then co-treated with 1 µg/ml ConA while PMs were co-treated with 100 ng/ml LPS, or both cultures were left without further treatment, for 24, 48 or 72 hours. PGB (3-30 µg/ml) concentration range was chosen according to previous reports in mice and on therapeutic concentrations in human 20,31 . Concentrations lower and higher than the previously reported concentrations were also tested (1.5 µg/ml; 60 µg/ml, respectively).
In vitro cell proliferation assay. In flat-bottomed 96-well microplates, 4 × 10 5 cells/200 µl/well of splenocytes or peritoneal cells were cultured as described above. Two hours prior to the end of each incubation interval, MTT assay was performed 32 . Optical density (OD) was measured at 570 nm using a microplate reader (Biotek, USA).

Assessment of cytokine secretion in vitro.
Splenocytes or PMs were cultured in 24-well plates and treated as described under "Cell culture" above. At the end of each incubation interval, culture supernatants were collected and immediately frozen at −80 °C. Levels of IL-6, TNF-α, IL-1β and IL-2 were measured by ELISA as per manufacturer's protocols (ThermoFisher Scientific, USA).
Statistical analysis. One-way ANOVA followed by Dunnett's post-hoc analysis were used for data analysis using Prism 5 software (GraphPad, USA). All data are presented as means ± standard error of the means (SEM). P-values are indicated with asterisks as follows: *P < 0.05, **P < 0.01, ***P < 0.001.  Table S1).

Effect of PGB on basal and mitogen-induced cytokine secretion in vitro.
Effect on isolated splenocytes. The effect of PGB on cytokine secretion in vitro was first investigated in isolated splenocytes, in the presence or absence of ConA. As presented in Fig. 2(a,b), treatment of splenocytes with PGB resulted in a significant  www.nature.com/scientificreports www.nature.com/scientificreports/ Treatment of splenocytes with PGB, resulted in a significant inhibition of basal and ConA-induced TNF-α secretion after 72 hours of incubation only (p < 0.0001; p = 0.0232, respectively). Whilst the basal TNF-α secretion was inhibited across all the concentrations, the ConA-induced secretion was only inhibited at 30-60 µg/ ml, as compared to PGB-untreated control group (Fig. 2c,d, respectively). PGB, across the whole concentration range, had no significant effect on ConA-induced IL-1β secretion at the three time points, as compared to PGB-untreated control group (Fig. 2e). The effect of PGB on basal IL-1β secretion could not be reliably determined as the levels were below the detection limit of the assay ( < 8 pg/ml, data not shown).
Effect on isolated PMs. Our results showed that PGB inhibited the secretion of a range of cytokines in ConA-treated or untreated splenocytes. Therefore, we next investigated the effect of PGB on LPS-induced cytokines secretion in PMs. Treatment of PMs with PGB had no effect on both basal and LPS-induced IL-6 and TNF-α secretion at the three time points (Supplementary Fig. S1a-d). IL-1β was undetectable at basal conditions and was only induced after treatment of PMs with LPS (100 ng/ml) for 72 hours where PGB had no effect on its secretion ( Supplementary Fig. S1e). The effect of PGB on LPS-induced IL-2 was not examined as LPS does not induce IL-2 secretion 33 .

Effect of PGB on basal and mitogen (ConA or LPS)-induced cytokine secretion in vivo.
As PGB inhibited cytokine secretion in splenocytes in vitro with no effects on cytokines secretion in PMs, we next investigated whether it has similar effect on cytokines secretion in vivo in the presence or absence of ConA/LPS. PGB significantly inhibited basal, ConA-and LPS-induced IL-6 secretion at 16 mg/kg dose (p = 0.0199; 0.0489; 0.048, respectively) ( Fig. 3a) while it inhibited basal, ConA-and LPS-induced TNF-α secretion at both doses tested; 8 and 16 mg/kg (p = 0.0217; 0.0249; 0.0232, respectively) (Fig. 3b). PGB significantly inhibited LPS-induced IL-1β secretion at 16 mg/kg (p = 0.0266), while it had no significant effect on ConA-induced IL-1β secretion under the experimental conditions used (Fig. 3c). IL-1β was undetectable at basal conditions, and therefore, the effect of PGB on IL-1β level could not be examined (<8 pg/ml, data not shown). With regards to IL-2 secretion, PGB had no effect on basal IL-2 secretion at 8 and 16 mg/kg doses while it significantly inhibited ConA-induced IL-2 secretion at the same doses (p = 0.0103) (Fig. 3d). All comparisons were made against PGB-untreated control group in the presence or absence of ConA/LPS.

Effect of PGB on spleen histology in the presence or absence of mitogens (ConA or LPS). As
PGB inhibited the secretion of several cytokines in vivo and in splenocytes in vitro. We next examined whether PGB could modulate mitogen-induced histologic changes in the spleen.
Representative images of the effect of PGB on spleen histology and on ConA-or LPS-induced inflammatory changes in the spleen are shown in Fig. 4(a-f). Both ConA and LPS induced only moderate changes in the spleens at the doses used in this study (Fig. 4b-d), as compared to the normal untreated spleen (PGB-untreated control group) (Fig. 4a). The white pulp was increased in the form of expanded splenic nodules with germinal centre formation and expanded periarteriolar lymphoid sheaths. In addition, there was an increase in the number of macrophages in the sinuses as well as the number of megakaryocytes that showed compacted clustering. On the other hand, PGB, particularly at 16 mg/kg, ameliorated both ConA-and LPS-induced histologic changes in the spleens, where the spleens of mice treated with 16 mg/kg PGB in the presence of ConA or LPS (Fig. 4e,f) showed comparable number of macrophages and megakaryocytes to those seen in the PGB-untreated control group (Fig. 4a). In addition, the morphology of the white pulp was less prominent with no germinal centre formation and with a reduction in the periarteriolar lymphoid sheath. PGB, at 8 µg/kg, did not exert any significant effects on ConA-or LPS-induced histologic changes in the spleens (data not shown). Furthermore, PGB alone had no effect on spleen histology (data not shown). In conclusion, 16 mg/kg PGB attenuated ConA-and LPS-induced changes in murine spleen. Chronic treatment protocols are required to further investigate these effects.

Discussion
In this short-term exposure study, we show for the first time that PGB has significant inhibitory effect on LPS/ ConA-induced cytokines secretion both in vivo, and in vitro in splenocytes, and on the basal secretion of some of the cytokines tested. Besides, PGB ameliorated mitogen-induced inflammatory changes in the spleen. The results of this study might give some clues to the effect of PGB on the inflammation observed in PGB-indicated conditions (neuropathic pain conditions, fibromyalgia, epilepsy and generalized anxiety disorders).
Peripheral inflammation is observed in the primary PGB-indicated conditions which all share elevated levels of IL-6, TNF-α, IL-1β and IL-2 cytokines 7-17 . Mitogen (LPS or ConA)-induced models of inflammation has long been used to assess the effects of drugs on macrophages-and T-cells-mediated proinflammatory cytokines production 34 and were also used to study the inflammation in PGB-indicated conditions [21][22][23][24]26 . Therefore, we used these models of inflammation to investigate the effect of PGB on the production of the above-mentioned cytokines. LPS is a potent stimulator of TLR4 in macrophages, resulting in the production of IL-6, TNF-α and IL-1β 33 . Hence, LPS was used in this study to stimulate PMs in vitro and to induce these cytokines in vivo. ConA is a potent stimulator of T-cells, resulting in enhanced IL-2 secretion, as well as TNF-α and IL-6 secretion 33,35 . As IL-2 is induced after ConA stimulation, but not after LPS stimulation 33,36 , we assessed the effect of PGB on IL-2 secretion in ConA-model only, both in vivo and in splenocytes in vitro. Thus, employment of both models allowed the investigation of the four cytokines tested. Besides, the investigation on peritoneal macrophages is supported by the emerging role of monocytes in promoting brain inflammation and exacerbation of neuronal damage in Scientific RepoRtS | (2020) 10:4007 | https://doi.org/10.1038/s41598-020-61006-1 www.nature.com/scientificreports www.nature.com/scientificreports/ neuroinflammatory conditions such as status epilepticus 37 . The doses of the mitogens, and the time points for cytokine measurements were selected based on previous studies and on optimization experiments [21][22][23][24][25]28,30 . In LPS-model of neuroinflammatory conditions, seizure and anxiety, LPS was used at doses of 0.01-3 mg/kg [21][22][23][24]28 . In this study, a dose of 0.1 mg/kg was used to induce inflammation in mice while keeping minimal lethality, and to allow the moderate anti-inflammatory effect of PGB to be significantly observed. In the ConA-model of inflammation, the dose and the concentration used were based on our previous studies 28, 30 . The 2-hour time point for cytokines measurement after LPS administration in vivo was previously shown to significantly increase IL-6, TNF-α, IL-1β and IL-2 levels in LPS-induced inflammation in mice 25,30 . In vitro, cytokines were previously measured after 24, 48 and 72 hours of mitogen administration in BALB/c mice 28,30 .
The data presented here show that PGB had no effect on splenocytes proliferation in vitro neither in the presence nor in the absence of ConA (Supplementary Table S1). This is consistent with a previous report where PGB had no effect on the proliferation of phytohemagglutinin (a mitogen)-stimulated splenocytes 20 . However, we demonstrated a significant inhibitory effect of PGB on cytokines secretion in vivo, and in splenocytes but www.nature.com/scientificreports www.nature.com/scientificreports/ not in PMs in vitro. Whilst our data were similar to previous reports in the inhibition of IL-6, IL-1β and TNF-α secretion 19,[38][39][40][41] , careful comparisons should be applied as those reports used different mitogens (carrageenan) 19 , different doses of PGB (at least 30 mg/kg of PGB) 19 , different inflammatory models (hyperglycemic stroke 39 or chronic constriction injury models 40,41 ), different sample for the assessment of cytokines levels (brain tissue 39 or nerve tissue 40,41 ) or even different animal species (rat) 19,[38][39][40][41] . Noteworthy, species, strain-and sex-associated differences in the regulation of some immune responses have been previously reported and similarity in responses is dependent on the outcomes measured [42][43][44][45] . Consequently, our study was the only report that investigated the anti-inflammatory effect of PGB in LPS/ConA-models of inflammation in BALB/c, employing comparable doses www.nature.com/scientificreports www.nature.com/scientificreports/ of PGB to the approved human therapeutic doses, and utilizing the mitogens which were previously used to study PGB-indicated conditions [21][22][23][24]26 .
Similar to previous reports 46, 47 , we showed undetectable basal IL-1β levels both in vivo and in vitro. In vivo, PGB led to a significant inhibition of LPS-induced IL-1β secretion. However, ConA-induced IL-1β secretion was only slightly inhibited. In vitro, IL-1β secretion was only induced in PMs after 72-hour incubation with 100 ng/ml LPS, and was not inhibited by PGB (Supplementary Fig. S1).
We demonstrated herein that PGB had no effect on the proliferation or cytokine secretion in PMs in vitro (Supplementary Table S1 and Supplementary Fig. S1). However, using the LPS-induced model of inflammation, other antiepileptic drugs such as lamotrigine significantly inhibited basal and LPS-induced IL-6 and TNF-α secretion in PMs in vitro 30 . This can be attributed to the structural differences between the two drugs.
Interestingly, our results showed that PGB not only significantly reduced mitogen-induced cytokines secretion but also the basal secretion. This further supports the anti-inflammatory effect of PGB and might suggest the possible utilization of similar molecular mechanisms for the reduction of both the basal and mitogen-induced cytokines secretion in BALB/mice. The exact mechanism of this anti-inflammatory effect is still not clear. It could be due to a direct effect in modulating the activity of some transcription factors, particularly nuclear factor-kappaB (NF-κB), as reported before 48 . Besides, there is an increasing body of evidence for the expression of voltage-dependent calcium channels in immune cells 49 , thus it might be possible that as PGB alter the function of the voltage-dependent calcium channels in the brain, it might also act on the voltage-dependent calcium channels in the immune cells. However, this theory requires further investigation. Similar basal anti-inflammatory effect was observed before by antiepileptic drugs that inhibits neuronal sodium channels like vinpocetine, carbamazepine and lamotrigine 30,50 . This basal inhibition might be clinically relevant to the current PGB-indicated conditions where elevated basal levels of these cytokines were clinically observed [7][8][9][10][11][12][13][14][15][16][17] .
In both models tested here, and as reported before 33,51 , we showed that LPS/ConA resulted in the elevation of the proinflammatory cytokines IL-6, TNF-α and IL-1β, and that IL-2 was specifically induced with ConA treatment. IL-6 is a pleiotropic cytokine which has multiple effects on immune and non-immune cells, including the nervous system 52 . In the nervous system, IL-6 has a key role in the normal homeostasis of neuronal tissue, while its overexpression in the brain results in neurodegeneration 52 . Both TNF-α and IL-1β are proinflammatory cytokines that induce the inflammatory response, TNF-α additionally regulates the inflammatory response by limiting the extent and the duration of inflammation as demanded 53 . Interestingly, IL-1β and TNF-alpha are inducers for the production of IL-6 through activation of transcription factors 53 . Therefore, PGB inhibitory effect on IL-6 might be a direct effect, or indirect effect through inhibition of IL-1β and/or TNF-α expression. IL-2, among other effects, enhances the proliferation of effector T cells and B cells 53 . Thus, PGB inhibitory effect on IL-2 production might alter the adaptive immune response. The effect of PGB on this complex interaction among the tested proinflammatory cytokines requires further investigation.
The spleen is a lymphoid organ that is composed of a red pulp and a white pulp. The white pulp contains a large reserve of lymphocytes and monocytes 54 . We showed here that both LPS and ConA resulted in expansion of the white pulp and an increase in the number of macrophages in the sinuses as well as the number of megakaryocytes. These effects were all attenuated with 16 µg/kg PGB treatment to almost the levels observed in the PGB-untreated control group (Normal saline). This supports our results of the inhibition of cytokines secretion by isolated splenocytes in vitro. Similarly, PGB was reported to ameliorate the inflammatory damage caused by LPS, however, in the hippocampus and cerebellum in rats 18 . Of note, although in our study PGB attenuated mitogen-induced inflammatory changes in the spleen, a tumorigenic effect of PGB was demonstrated before, where chronic administration of 1000-5000 mg/kg of PGB for up to 1 year induced hemangiosarcoma in mice and resulted an increase in the number of macrophages and megakaryocytes in mouse spleen 55 . Therefore, the immunomodulatory effects of PGB seems to be dose and time dependent.
In the current study, PGB demonstrated some degree of differential effects at the different experimental setting employed. In splenocytes, all the cytokines tested were significantly inhibited by PGB treatment with the exception of IL-1β. Similarly, previous reports showed differential inhibitory effects of anti-inflammatory compounds on cytokines secretion in splenocytes 56,57 . Such effects varied based on the cytokine investigated, the mitogen used to stimulate splenocytes, and the concentration of the inhibitors 56,57 . In addition, PGB demonstrated here a cell-based differential effect on cytokine secretion, unlike in splenocytes, PGB did not inhibit cytokine secretion in macrophages at the experimental setting used in our study (Supplementary Fig. S1). This could be attributed to the different receptor expression in the different cell population, and therefore, their responsiveness to PGB. Such differential responsiveness of immune cells to anti-inflammatory agents was observed before 58 . Our data showed that both IL-6 and TNF-α were inhibited by PGB in vivo. However, PGB inhibited LPS-induced but not ConA-induced IL-1β secretion, and inhibited ConA-induced but not LPS-induced IL-2 secretion in vivo. These in vivo responses to PGB resulted from the complex interaction of the immune system elements in the in vivo environment, which resulted in the observed mitogen-specific differences in PGB effects. Other antiepileptic agents had also shown differential effects on cytokines secretion, where lamotrigine has been shown to inhibit IL-2, IL-1β and TNF-α secretion, but not IL-6 secretion, in stimulated-human blood in vitro 4 .
Of note, in the current report we have focused on the effects of PGB on peripheral inflammatory responses. Although neuroinflammation per se was not investigated in this study, our results might give some clues to the effect of PGB on peripheral inflammation observed in PGB-indicated conditions (neuropathic pain conditions, fibromyalgia, epilepsy and generalized anxiety disorders) where inflammatory mechanisms and elevation of proinflammatory cytokines, such as IL-6, TNF-α, IL-1β and IL-2 have been documented [7][8][9][10][11]16,17 . Besides, compared to our study, similar model of inflammation and comparable doses of LPS were employed before to study these conditions [21][22][23][24] . Additionally, it was previously shown that LPS-induced systemic inflammation lead to neuroinflammation in mice 25 . Importantly, targeting IL-6, TNF-α and IL-1β proinflammatory cytokines has been previously proposed for the management of neuropathic pain conditions and epilepsy 8 www.nature.com/scientificreports www.nature.com/scientificreports/ the therapeutic benefit of PGB in the management of these conditions might involve the anti-inflammatory mechanisms demonstrated in this study, in combination with the mechanism involving binding to the voltage-dependent calcium channels in the brain 6 . However, careful interpretation of the data in clinical setting is required as this study represents a short-term exposure protocol while PGB is clinically used for long-term treatment. Development of chronic exposure protocols and chronic models of inflammation is demanded to further clarify the long-term anti-inflammatory effects of PGB. Besides, the effects of PGB on inflammatory responses within the CNS could be investigated in future studies using CSF samples or microglial cell cultures.

conclusion
The data presented in the current study demonstrated for the first time a significant peripheral anti-inflammatory effect of PGB on murine LPS/ConA-induced cytokines secretion both in vivo, and in vitro in splenocytes. Besides, PGB attenuated mitogen-induced inflammatory changes in murine spleen. PGB, however, had no effect on cytokine secretion in PMs in vitro. Our findings may extend the current state of knowledge regarding the role of PGB in modulating the immune system in neuroinflammatory conditions such as neuropathic pain conditions, fibromyalgia, epilepsy and generalized anxiety disorders. This anti-inflammatory effect might be adjunctive to the documented mechanism of action of PGB (binding to the voltage-dependent calcium channels in the brain) in treating these conditions. Further studies, however, are necessary to investigate the effects of PGB on cytokine levels within the CNS and to investigate these effects during chronic exposure protocols.

Data availability
The dataset analyzed during the current study is available from the corresponding author on reasonable request.