Antibiotics cocktail treatment causes major depletion of gut microbiota and does not alter cutaneous thermal and mechanical sensitivity in naïve animals
To explore the effects of gut microbiota on somatic pain, we first fed the mice with a combination of four different kinds of antibiotics (ABX): Vancomycin (0.5 g/L), Ampicillin (1 g/L), Neomycin (1 g/L) and Metronidazole (1 g/L), targeting different kinds of bacteria. This antibiotics cocktail was dissolved in the drinking water with free access to mice. Additional sweetener was added to prevent the aversion of mice to ABX water. The mice still showed decreased consumption of the ABX water during the first 6 days after ABX feeding but showed no preference between normal water and ABX water after 6 days (Fig. 1A). The success of ABX feeding on the gut microbiota depletion was first confirmed by the maladaptive GI tract morphology. As previously described, ABX treated mice exhibited enlarged caeca with dark-colored cecal contents as a consequence of the absence of gut microbiota (Fig. 1B). To further examine the effects of ABX feeding on gut microbiota, we collected the feces of mice and tested the fecal bacteria by plating on bacteria growth media. The colony formation unit (CFU) counting from fecal bacteria plated on tryptic soy agar (TSA) media started to decrease at day 7 and remained significantly low at day 14 and 21 after the start of ABX feeding (Fig. 1C-D). The CFUs on TSA media only captured the change of some aerobic and facultative anaerobic bacteria of the microbiota. To fully examine the change of gut microbiota, we measured the total fecal bacteria DNA concentration and performed 16S rRNA sequencing on the fecal microbiota of the mice at day 14 after the start of ABX feeding (Fig. 1E-H). The fecal bacteria DNA concentration significantly decreased at day 3 after the starting of ABX treatment and remained low until the last test at day 21 (Fig. 1E). The sequencing results showed that the composition of the microbiota was simplified, changing from a highly diverse community to a Proteobacteria phylum dominating community (Fig. 1F). The α-diversity of the microbiota was significantly decreased (Fig. 1H) after ABX feeding. The principal component analysis (PCA) showed a highly uniformized microbiota after ABX treatment and a complete division between SPF and ABX treated microbiota (Fig. 1G). These results showed that ABX feeding caused major depletion of the gut microbiota.
We then examined the painful behaviors in ABX-treated mice. ABX treated mice showed no difference in the cutaneous thermal and mechanical sensitivity compared with the control SPF mice (Fig. 1I-J). ABX treatment also showed no effect on subcutaneous injection of capsaicin-induced acute pain (Fig. 1K). Gut enterochromaffin cells (ECs) produce the majority of serotonin of our body, which is a very important neurotransmitter and neuromodulator in pain sensation. Commensal bacteria in the gut are capable of promoting the biosynthesis of serotonin by ECs. In consideration of these interactions, we also compared levels of serotonin in ABX-treated and SPF control mice. The results showed that gut microbiota depletion (14 days after ABX treatment) did not change serotonin concentrations in blood and the spinal cord (Fig. 1L-M).
Gut microbiota depletion ameliorates thermal hyperalgesia and inhibits spinal glial cell activation in animals with nerve injury.
Noxious stimulation-evoked acute pain processing mainly involves neuronal transduction. However, neuropathic and inflammatory pain requires immune activation and neuro-immune interactions. Gut microbiota has been shown to play tremendous roles in regulating the functions of immune systems. We asked whether gut microbiota depletion could alter the development of nerve injury-induced chronic pain. CCI treatment produced long-lasting thermal hyperalgesia and mechanical allodynia in mice. Gut microbiota deletion by pretreatment of ABX for two weeks resulted in significant alleviation of CCI-induced thermal hyperalgesia, but not the mechanical allodynia (Fig. 2A-C). We then wondered whether such pre-depletion of gut microbiota was needed to exert its analgesic effect and whether the gut microbiota manipulation could influence the development and persistence of neuropathic pain. We started ABX feeding simultaneously with nerve injury (at day 0 of CCI treatment) and found that the thermal hyperalgesia, but not the mechanical allodynia was also impaired starting within 24 h and lasting to the last test on day 28. The analgesic effect is significant compared with SPF control animals (Fig. 2D-F). We then had ABX feeding starting at day 7 after CCI treatment and found that the established thermal hyperalgesia was reversed and the mechanical allodynia still remained unchanged (Fig. 2G-I). Taken together, these results indicate the influence of gut microbiota on chronic neuropathic pain with several features, i.e., the gut microbiota depletion selectively impacts the pathway of the thermal hyperalgesia, but not that of the mechanical allodynia; the effects of gut microbiota on neuropathic pain could be instant; and that gut microbiota is able to influence both the early development and late maintenance of neuropathic pain.
After nerve injury, activation of microglia and astrocyte in the spinal cord is critical for the chronification of pain hypersensitivity. Gut microbiota has been found to shape the maturation of microglia in the spinal cord and the astrocyte functions in the brain. Thus, the microbiota may contribute to the nerve injury-induced thermal hyperalgesia through influencing the glia cell function. We continued to examine the glia cell activation in CCI animals with ABX treatment. Immunofluorescence staining showed that CCI-induced increased activation of microglial cells (Iba1) and astrocytes (GFAP) in the spinal cord was greatly suppressed in animals with ABX treatment. Tissues of the spinal cord was taken on day 7 after CCI treatment (Fig. 3A-D). These results suggest that microbiota may contribute to the nerve injury-induced glial cell activation, which contributes to the development of neuropathic pain.
Gut microbiota depletion ameliorates mechanical allodynia and inhibits cytokines production in DRG neurons in animals with chemotherapy
A recent study shows that gut microbiota depletion by antibiotics or germ-free condition can prevent the development of mechanical pain in chemotherapy induced-neuropathic pain. It remains unknown how gut microbiota may influence different modalities of pain behaviors and different phases over the time course of the disease in chemotherapy-induced painful hypersensitivity. Therefore, parallel to that we have tested in mice with CCI treatment, we performed similar investigations in mice with chemotherapy-induced neuropathic pain. In those animals with repetitive administration of chemotherapy drug oxaliplatin (OXA)(i.p., daily for 5 consecutive days), ABX feeding was started prior to (from 2 weeks before), simultaneously at the time point of, or 7 days after OXA treatment. Unlike CCI, OXA treatment induced long-lasting behaviorally expressed mechanical allodynia without thermal hypersensitivity (Fig. 4A-I). Pretreatment of ABX feeding starting 2 weeks prior to OXA completely inhibited the development of mechanical allodynia (Fig. 4B), which is consistent to the previous report . What’s more, our results further showed that ABX feeding starting simultaneously with OXA treatment also completely prevented the development of mechanical allodynia (Fig. 4E). However, ABX feeding after OXA-induced mechanical allodynia had established did not produce significant analgesic effect on the mechanical allodynia, but a temporary and slight inhibitory effect 3 days after the ABX treatment (Fig. 4H). ABX treatment did not show any additional effect on the thermal sensitivity (Fig. 4C, F and I). These results indicate that gut microbiota depletion can prevent development and persistence of OXA-induced mechanical allodynia, suggesting that microbiota may contribute to the chemotherapy-induced chronic pain.
Different from nerve injury, studies have shown that OXA chemotherapy does not cause glia cell activation and glial cell activation dose not contribute to OXA chemotherapy-induced neuropathic pain. We here confirmed that both microglial cells and astrocytes in the spinal cord were not activated following OXA treatment with or without ABX treatment (Fig. 5A-B). These findings may help to exclude the possible roles of glia cell activation in chemotherapy-induced neuropathic pain and microbiota-depletion-caused alterations of sensory sensitivity. It was previously reported that gut microbiota facilitated recruitment of cytokines producing macrophage into DRG after chemotherapy. We thus examined some macrophage producing cytokines in DRG by western blot. The results showed OXA treatment resulted in greatly increased expression of IL-6 and TNF-α and such increased cytokines were significantly reversed following ABX treatment (Fig. 5C). These findings demonstrate that in chemotherapy-induced neuropathic pain the gut microbiota facilitates the cytokine production and immune functions in DRG and that gut microbiota may influence chemotherapy- and nerve injury-induced neuropathic pain in distinct mechanisms.
Gut microbiota depletion prevents development of STZ-induced diabetes and mechanical allodynia and inhibits glia cell activation in the spinal cord
Diabetic neuropathic pain (DNP) is another common form of neuropathic pain. We investigated whether gut microbiota could be involved in development of DNP. Repetitive administration of STZ (i.p., 40 mg/kg, one dose per day for five consecutive days) (Fig. 6A) produced significant and persistent increase of blood glucose level (Fig. 6B) accompanied with mechanical allodynia, but not thermal hypersensitivity (Fig. 6C-D). Surprisingly, pretreatment of ABX feeding for two weeks completely prevented STZ-induced increase of blood glucose (Fig. 6B) and the mechanical allodynia or thermal hypersensitivity (Fig. 6C-D). These results show that pre-depletion of gut microbiota can completely prevent the development of high blood glucose and DNP, suggesting that gut microbiota may play a critical role in the development of STZ-induced diabetes and DNP.
We continued to repeat the experimental protocols in STZ model as those used in CCI and OXA models. ABX feeding simultaneously with STZ treatment (Fig. 6E) slightly decreased STZ-induced high blood glucose (Fig. 6F) and permanently completely diminished the mechanical allodynia (Fig. 6G) without affecting the thermal sensitivity (Fig. 6H). ABX feeding starting 42 days after STZ treatment (Fig. 6I) did not affect the established high blood glucose (Fig. 6J) but did successfully suppressed the established mechanical allodynia (Fig. 6K) and had no effect on the thermal sensitivity (Fig. 6L). Glia cell activation has been reported to play roles in diabetic neuropathic pain. We continued to examine the glia cell activation in the spinal cord in STZ-treated animals with high blood glucose and mechanical allodynia with or without ABX treatment. The immunostaining showed that expression of Iba1 and GFAP in the spinal cord was significantly increased 2 weeks after STZ treatment. Simultaneous feeding of ABX with STZ treatment partially but significantly decreased STZ-induced increased expression of Iba1 and GFAP (Fig. 7A-D). Taken together, these results demonstrate that the gut microbiota is required for the production, but not the persistence of STZ-induced diabetes, which was characterized as the high blood glucose. Meanwhile, the gut microbiota is required for both induction and maintenance of STZ-induced mechanical allodynia.
Fecal bacteria transplantation partially reverses the gut microbiota colonization and fully rescues gut microbiota depletion-induced behaviorally expressed pain in CCI, OXA, and STZ models
Given that gut microbiota depletion can greatly reduce or completely diminish the painful thermal hyperalgesia or mechanical allodynia caused by nerve injury, chemotherapy, and diabetes, we asked whether restoration of gut microbiota in ABX-treated animals could reverse the painful behaviors in CCI, OXA and STZ animals. ABX feeding was started from 14 days prior to CCI, OXA, or STZ treatment and then was withdrawn by replacing ABX with normal water at day 7 after CCI, day 14 after OXA, and day 35 after STZ treatment followed immediately by oral gavage with fresh fecal bacteria solution taken from SPF mice (a dose per day for 3 continuous days). Feces of the mice was collected at the end time point of the behavioral tests and the fecal bacteria DNA concentration was measured. The results showed that fecal bacteria transplantation did recover ABX treatment-induced depletion of gut microbiota colonization and alterations of sensory sensitization in animals with CCI, OXA or STZ treatment (Fig. 8A-F). In CCI mice, ABX treatment depleted gut microbiota and produced limited but significant inhibitory effect on the thermal hyperalgesia. Such gut microbiota depletion and thermal hypersensitivity reduction were completely reversed following fresh fecal bacteria transplantation (Fig. 8A-B). In mice with OXA or STZ treatment, ABX treatment depleted gut microbiota and prevented or completely suppressed the mechanical allodynia. Again, the treatment with fresh fecal bacteria transplantation quickly and completely restored the gut microbiota and recovered the mechanical hypersensitivity (Fig. 8C-F).
To compare the detailed composition of gut microbiota in fecal bacteria transplanted mice with that from SPF mice, we performed 16S rRNA sequencing analysis in mice with CCI treatment. The treatment protocol is shown in Fig. 9A. The sequencing results showed that fecal bacteria transplantation induced a significant restoration of the microbiota from ABX treated mice (Fig. 9B-E). The restoration of the microbiota community was indicated by the increase of the α-diversity of the composition of the microbiota (Fig. 9B), separation of transplanted mice feces from ABX treated mice in PCA analysis (Fig. 9C), and increased OTU counting (Fig. 9D). Nevertheless, the composition of the microbiota of the mice that received fecal bacteria transplantation was still significantly different from that from SPF mice (Fig. 9F). The transplanted mice microbiota showed lower α-diversity, separation from SPF mice in the PCA analysis and lower OTU counting than that in SPF mice (Fig. 9B-D). In order to link the possible relationship of some specific bacteria families with nerve injury-induced painful behaviors, we further examined the abundance of Akkermansia, Bacteroides, and Desulfovibrionaceae phylus in the feces of mice with nerve injury, CCI treatment. The results showed that the abundance of these three bacteria was restored or even over expressed in fecal bacteria transplanted mice and positively related to the extent of painful behaviors (Fig. 9F), indicating their possible specific roles in pain regulation.
Taken together, these results indicate that the effects of gut microbiota on neuropathic pain behaviors are reversible simply by microbiota transplantation and that the restoration of behaviorally expressed painful syndromes doesn’t require fully restoration of gut microbiota, suggesting that certain bacteria or sub-community of the microbiota may be specifically responsible for the pain behavior regulation. Akkermansia, Bacteroides, and Desulfovibrionaceae phylus may belong to the important gut microbiota, which are important to the development of painful manifestation in the nerve injured-mice.