N2O at a concentration of 60%, but not 30%, induces CPP rewarding behavior and activates VTA neurons

N2O is typically mixed with oxygen in clinical settings to prevent adverse effects. Therefore, we established a CPP apparatus in which the concentrations of N2O, oxygen and CO2 were continuously monitored to evaluate N2O-induced rewarding behavior (Fig. S1). C57BL/6 J mice were randomly assigned to three groups: Air, 30% N2O, and 60% N2O. Different from a previous report [29], after the pre-test phase (Day 1), mice underwent 6 consecutive acquisition days (Days 2–7), during which N2O was paired in the morning and air in the afternoon. On Day 8, a post-test phase was conducted to measure CPP scores in the absence of N2O (Fig. 1A). Mice in the Air group underwent the same procedure and were paired with air in both the morning and afternoon. CPP scores in the 30% N2O group showed no significant difference compared with those in the Air group (Fig. 1B), whereas CPP scores in the 60% N2O group showed a significant increase compared to those in the Air group during the post-test phase (Fig. 1C). These results indicated that N2O at a concentration of 60%, but not 30%, induced CPP. The expression of c-Fos in neurons represents the activation of neurons in response to behaviors. We therefore tested the expression of c-Fos in the VTA area and found that compared with that in the Air group, the number of c-Fos-positive neurons in the 30% N2O group did not significantly differ (Fig. 1D). To investigate functional changes in glutamatergic transmission onto VTA dopamine neurons, mEPSCs were recorded using the patch-clamp technique. There were no significant differences in the frequency or amplitude (Fig. 1E, F) of the mEPSCs in VTA neurons between the 30% N2O group and the Air group. However, both c-Fos expression (Fig. 1G) and the frequency and amplitude of the mEPSCs (Fig. 1H, I) were significantly greater in the 60% N2O group than in the Air group. The frequency of mEPSCs reflects the probability of presynaptic vesicular release probability, whereas the amplitude reflects postsynaptic AMPA receptor sensitivity or density. Consistent with this, these results demonstrate that 60% N2O, but not 30%, induces CPP rewarding behavior and enhances activity in both VTA neurons and their upstream afferent inputs.

Fig. 1: N2O at a concentration of 60%, but not 30%, induces CPP rewarding behavior and activates VTA neurons.The alternative text for this image may have been generated using AI.

A Schematic illustration of the experimental procedure. B Left, representative CPP traces for the Air and 30% N2O groups. Right, quantification of CPP scores. n = 12 mice (Air) and n = 13 mice (30% N2O). t23 = −1.092, p = 0.286, unpaired t test. C Left, representative CPP traces for the Air and 60% N2O groups. Right, quantification of CPP scores. n = 9 mice per group. t16 = −6.841, ***p < 0.001, unpaired t test. D Left, representative images of c-Fos expression in VTA neurons from the Air and 30% N2O groups. Right, quantification of c-Fos-positive neurons. n = 15 slices, 5 mice per group. t28 = −0.924, p = 0.363, unpaired t test. Scale bar = 200 μm. E Representative traces of mEPSCs recorded from dopamine neurons in the Air group (left) and the 30% N2O group (right). F Quantification of mEPSCs frequency (left) and amplitude (right). Frequency: 12 neurons, 3 mice (Air) and 14 neurons, 3 mice (30% N2O), t24 = −0.414, p = 0.683, unpaired t test. Amplitude: 12 neurons, 3 mice (Air) and 14 neurons, 3 mice (30% N2O), t24 = −1.252, p = 0.223, unpaired t test. G Left, representative images of c-Fos expression in the Air and 60% N2O groups. Right, quantification of c-Fos-positive neurons. n = 15 slices, 5 mice per group. Mann-Whitney U Statistic = 31.000, ***p < 0.001, a Mann-Whitney Rank Sum test. Scale bar = 200 μm. H Sample traces of mEPSCs in the Air group (left) and the 60% N2O group (right). Quantification of mEPSCs frequency (left) and amplitude (right). Frequency: 15 neurons, 3 mice (Air) and 14 neurons, 3 mice (60% N2O), t27 = −5.583, ***p < 0.001. Amplitude: 15 neurons, 3 mice (Air) and 14 neurons, 3 mice (60% N2O), t27 = −4.496, ***p < 0.001, unpaired t test.

60% N2O specifically induces homocysteine accumulation in the VTA and its upstream DCN

Since 60% N2O altered presynaptic transmission to the VTA, we hypothesized that N2O may affect its upstream regions. To investigate this connection between the VTA and its upstream regions, the retrograde tracer cholera toxin subunit B (CTB)-555 was injected into the VTA to trace the projections of the VTA in mice (Fig. 2A). Immunofluorescence staining revealed five primary brain regions connected to the VTA, including the DCN, central lateral nucleus (CLA), NAc, primary motor cortex (M1), agranular insular cortex (AIC) (Fig. 2B), as well as other regions (Fig. S2). These results confirm that these brain regions are connected with VTA neurons. However, a question has been raised concerning which brain area is mainly affected by N2O. Since homocysteine accumulation is a characteristic activity of N2O, we measured the homocysteine levels in the VTA and five upstream regions. Compared with the Air group, homocysteine levels in the VTA region in the 60% N2O group weresignificantly elevated (Fig. 2C). Furthermore, compared with the Air group, the homocysteine levels in the DCN area, but not in the CLA, NAc, M1, or AIC, were significantly greater (Fig. 2D–H). To further confirm that 60% N₂O, but not 30% N₂O, effectively modulates the VTA and DCN, we examined c-Fos expression and homocysteine levels in the 30% group. We found no significant changes in either c-Fos or homocysteine levels following 30% N₂O exposure compared to the Air group (Fig. S3). Taken together, these findings suggest that 60% N₂O promotes both homocysteine accumulation and neuronal activity in the VTA and its upstream DCN.

Fig. 2: N2O specifically induces homocysteine accumulation in the VTA and its upstream DCN.The alternative text for this image may have been generated using AI.

A Injection of CTB into the VTA. B Immunofluorescence mapping of VTA-projecting regions. Representative images showing CTB labeling in the DCN, CLA, NAc, M1, and AIC. n = 3 mice. C Immunohistochemical staining and quantification of homocysteine levels in the VTA in the Air and 60% N2O groups. n = 15 slices, 5 mice per group. Mann-Whitney U Statistic= 17.500, ***p < 0.001, a Mann-Whitney Rank Sum test. D Homocysteine levels in the CLA in the Air and 60% N2O groups. n = 19 slices, 5 mice (Air) and n = 24 slices, 5 mice (60% N2O). t41 = 1.584, p = 0.121, unpaired t test. E Homocysteine levels in the DCN in the Air and 60% N2O groups. n = 15 slices, 5 mice per group. Mann-Whitney U Statistic= 24.000, ***p < 0.001, a Mann-Whitney Rank Sum test. F Homocysteine levels in the NAc in the Air and 60%N2O groups. n = 23 slices, 5 mice (Air) and n = 28 slices, 5 mice (60% N2O). Mann-Whitney U Statistic= 256.500, p = 0.218, a Mann-Whitney Rank Sum test. G Homocysteine levels in the M1 in the Air and 60% N2O groups. n = 24 slices, 5 mice (Air) and n = 23 slices, 5 mice (60% N2O). Mann-Whitney U Statistic= 209.500, p = 0.160, a Mann-Whitney Rank Sum test. H Homocysteine levels in the AIC in the Air and 60% N2O groups. n = 15 slices, 5 mice (Air) and n = 23 slices, 5 mice (60% N2O). t36 = 1.271, p = 0.212, unpaired t test.

Intervention with a homocysteine inhibitor does not reduce N2O-induced rewarding behavior

Having shown that N2O induced homocysteine accumulation in the VTA and DCN, we investigated whether reducing homocysteine levels could prevent N2O-induced rewarding behavior. Betaine (BET) facilitates the metabolism of homocysteine, thereby reducing its levels. Mice were randomly assigned to three groups: the Air group, 60% N2O group, and BET-60% N2O group. Both the Air and 60% N2O groups received intraperitoneal injections of an equivalent volume of saline, whereas the BET-60% N2O group was administered a BET solution (100 mg/kg) 1 h before the acquisition phase test (Fig. 3A). BET itself did not change the CPP scores in the Air group of mice (Fig. S4). Furthermore, the CPP behavioral results showed that BET did not significantly reduce the N2O-induced CPP scores compared with the 60% N2O group (Fig. 3B). Homocysteine levels in the DCN and VTA were then measured in the three groups. Compared with 60% N2O-exposed mice, homocysteine levels in the DCN and VTA were markedly lower in BET-treated mice (Fig. 3C, D), indicating that BET effectively reduced homocysteine levels following N2O exposure. To further examine the effect of BET on excitatory synaptic transmission, mEPSCs were recorded in VTA neurons. Representative traces of mEPSCs for the Air group, 60% N2O group, and BET-60% N2O group are shown (Fig. 3E). We found that the frequency and amplitude of mEPSCs in VTA neurons from mice in the BET-60% N2O group did not differ from those in the 60% N2O group (Fig. 3F, G), suggesting that BET does not affect glutamatergic transmission in VTA dopamine neurons. Taken together, these findings suggest that reducing homocysteine levels does not reverse N2O-induced rewarding behavior.

Fig. 3: Intervention with a homocysteine inhibitor does not decrease N2O-induced rewarding behavior.The alternative text for this image may have been generated using AI.

A Schematic illustration of the experimental procedures. B Left, representative traces of CPP behavior in the Air, 60% N2O and BET-60% N2O groups. Right, quantification of CPP scores in the three groups. n = 8 mice (Air), n = 8 mice (60% N2O) and n = 9 mice (BET-60% N2O). F2, 22 = 17.089, ***p < 0.001, one-way ANOVA. ***p < 0.001 for 60% N2O vs Air, ***p < 0.001 for BET-60% N2O vs Air, p = 0.216 for BET-60% N2O vs the 60% N2O group. C Left, representative images of homocysteine levels in the DCN across the three groups. Right, quantification of homocysteine levels. n = 10 slices, 3 mice (Air), n = 12 slices, 3 mice (60% N2O), n = 12 slices, 3 mice (BET-60% N2O). F2, 31 = 7.619, **p = 0.002, one-way ANOVA. **p = 0.002 for 60% N2O vs Air, *p = 0.031 for 60% N2O vs BET-60% N2O. D Left, representative images of homocysteine levels in the VTA. Right, quantification of homocysteine levels. n = 14 slices, 3 mice (Air), n = 16 slices, 3 mice (60% N2O), n = 12 slices, 3 mice (BET-60% N2O). F2, 39 = 16.710, ***p < 0.001, one-way ANOVA. ***p < 0.001 for 60% N2O vs Air, ***p < 0.001 for 60% N2O BET-60% vs N2O. E Sample traces of mEPSCs. F Quantification of mEPSCs frequency in VTA neurons. n = 15 neurons, 3 mice (Air), 14 neurons, 3 mice (60% N2O), and 12 neurons, 3 mice (BET-60% N2O). F2, 38 = 17.756, ***p < 0.001, one-way ANOVA. ***p < 0.001 for 60% N2O vs Air, p = 0.847 for 60% N2O vs BET-60% N2O. G Quantification of mEPSC amplitude in VTA neurons. n = 15 neurons, 3 mice (Air), 14 neurons, 3 mice (60% N2O), and 12 neurons, 3 mice (BET-60% N2O). F2, 38 = 7.808, ***p < 0.001, one-way ANOVA. **p = 0.003 for 60% N2O vs Air, p = 0.759 for 60% N2O vs BET-60% N2O.

N2O potentiates the DCN-VTA circuit by enhancing AMPAR-mediated responses in both DCN and VTA dopamine neurons

The homocysteine inhibitor did not attenuate reward behavior, likely because it failed to decrease glutamatergic transmission in the VTA [24, 25]. Therefore, we hypothesized that N2O may act directly on both AMPAR- and NMDAR-mediated glutamatergic transmission. Since N2O is a known NMDAR antagonist, we first measured NMDAR-EPSCs in DCN neurons using in vitro brain slices. Compared with baseline, N2O significantly decreased the amplitude of NMDAR-EPSCs (Fig. 4A), confirming its inhibitory effect on NMDAR function. Next, we tested its effect on AMPAR-EPSCs. Interestingly, N2O significantly increased the amplitude of AMPAR-EPSCs (Fig. 4B). Addiction-related processes are associated with synaptic plasticity changes, characterized by a rapid and relatively sustained increase in the phosphorylation of GluA1 at both synaptic and extrasynaptic sites under conditions of GluA2 receptor deficiency [30, 31]. We therefore examined p-GluA2- and p-GluA1-AMPAR expression after in vivo N2O exposure. 60% N2O decreased the expression of p-GluA2-containing AMPARs (Fig. 4C), but increased the expression of p-GluA1-containing AMPARs (Fig. 4D), while total GluA1- and GluA2-AMPARs levels remained unchanged (Fig. S5), indicating that N2O directly affects the AMPAR-mediated glutamatergic transmission in DCN neurons. In vivo monitoring with glutamatergic probes revealed that after 60% N2O exposure, DCN excitatory activity increased and was accompanied by motivated behavior (Fig. 4E and video 1). Consistent with these findings, in vitro slice recordings demonstrated that AMPAR-mediated transmission increased the firing frequency of DCN neurons (Fig. 4F). Together, these results suggest that N2O potentiates the DCN-VTA circuit by enhancing AMPAR-mediated glutamatergic transmission, thereby increasing DCN neuronal activity.

Fig. 4: N2O potentiates the DCN-VTA circuit by enhancing the AMPAR-mediated response in both DCN and VTA dopamine neurons.The alternative text for this image may have been generated using AI.

A NMDAR-EPSCs in DCN neurons after in vitro exposure to N2O. Sample traces (left) and statistical analysis (right). n = 7 neurons, 6 mice. F2, 12 = 8.279, **p = 0.006, one-way RM ANOVA. **p = 0.004 for BL vs N2O. The blue arrow indicates electrical stimulus (100 µs). B AMPAR-EPSCs in DCN neurons. Sample traces (left) and statistical analysis (right). n = 7 neurons, 6 mice. F2, 12 = 7.537, **p = 0.008, one-way RM ANOVA. **p = 0.008 for N2O vs BL, *p = 0.014 for N2O vs Washout. C Western blot samples (left) and statistical analysis (right) of p-GluA2 in the Air and 60% N2O groups. n = 6 repetitions, 5 mice per group. t10 = 3.020, *p = 0.013, unpaired t test. D Western blot samples (left) and statistical analysis (right) of the p-GluA1 levels in the two groups. n = 6 repetitions, 5 mice per group. t10 = −2.343, *p = 0.041, unpaired t test. E Calcium imaging of VGluT2⁺ neurons during N2O exposure. Left, Representative ΔF/F traces showing changes in VGluT2⁺ neuronal Ca²⁺ activity during the Air–N2O–Air recording period. The lower panel shows the corresponding heatmap of VGluT2⁺ neuronal activity. Middle, enlarged traces illustrating Ca²⁺ responses during N2O exposure, with red arrows indicating representative Ca²⁺ transients. The lower panel shows the corresponding heatmap. Right, quantification of the area under the curve (AUC) of VGluT2⁺ neuronal Ca²⁺ signals during Air (before), N2O exposure, and Air (after). n = 3 mice. F2, 6 = 21.76, **p = 0.0018, one-way ANOVA. **p = 0.003 for Air (before) vs N2O, **p = 0.003 for Air (after) vs N2O. F Left, sample traces showing evoked firing of DCN neurons during baseline (BL), N2O exposure, and after application of the AMPAR antagonist CNQX (10 µM). Right, quantification of the frequency across the three conditions. n = 5 neurons, 4 mice. F2, 8 = 5.463, *p = 0.032, one-way RM ANOVA. *p = 0.030 for N2O vs BL. G Sample traces (left) and statistics (right) of spontaneous firing in VTA dopamine neurons in the Air and 60% N2O groups. n = 19 neurons, 4 mice (Air) and n = 25 neurons, 4 mice (60% N2O). t42 = −4.243, ***p < 0.001, unpaired t test. H Sample traces (left) and statistics (right) of evoked firing of VTA dopamine neurons. n = 17 neurons, 4 mice (Air) and n = 16 neurons, 4 mice (60% N2O). F1, 161 = 8.41,##p = 0.007, two-way RM ANOVA. I Injection sites and optical fiber placements for DA3m viruses in the NAc. Scale bars = 200 µm. J Dopamine signaling during N2O exposure. Left: representative ΔF/F traces showing changes in dopamine signals across the recording period, including BL, Air, and N2O exposure (0–60%). Right: quantification of the AUC of dopamine signals. n = 3 mice. t4 = −4.277, *p = 0.013, unpaired t test. K Sample traces (left) and statistics (right) of the evoked firing during BL, N2O, and CNQX. n = 5 neurons, 4 mice. F2, 8 = 6.518, *p = 0.021, one-way RM ANOVA. *p = 0.029 for BL vs N2O, *p = 0.017 for CNQX vs N2O. L Sample images showing ChrimsonR-tdTomato expression in DCN neurons and their terminals in the VTA. M Representative traces (left) and statistical analysis (right) of the paired-pulse ratio (PPR) of optogenetically evoked EPSCs. n = 14 neurons, 4 mice (Air), n = 19 neurons, 4 mice (60% N2O). t31 = 2.585, *p = 0.015, unpaired t test. N Representative traces (left) and statistical analysis (right) of AMPAR/NMDAR ratios in VTA neurons. n = 18 neurons, 4 mice (Air) and n = 14 neurons, 4 mice (60% N2O). t30 = −3.981, ***p < 0.001, unpaired t test.

We further examined the firing activity of VTA dopamine neurons in response to N2O both in vitro and in vivo. First, we employed the patch-clamp technique to measure the spontaneous and evoked firing of VTA dopamine neurons after the behavioral experiments. Compared with the Air group, the 60% N2O group showed significant increases in both spontaneous (Fig. 4G) and evoked firing (Fig. 4H). Next, in vivo dopaminergic (DA3m) sensors were used to assess the effects of N2O on the activity of dopamine neurons (Fig. 4I). We found that compared with exposure to air, 60% N2O increased the dopamine release (Fig. 4J). Finally, to test whether this enhancement resulted from upstream glutamatergic transmission, we performed in vitro recordings of VTA dopamine neuron firing during N2O application. N2O increased the firing frequency of VTA dopamine neurons, an effect that was prevented by AMPAR antagonists (Fig. 4K). These data suggest that N2O enhances upstream AMPAR-mediated glutamatergic transmission, thereby increasing the excitability of VTA dopamine neurons.

Having demonstrated that upstream brain regions may drive the activation of dopamine neurons during chronic N2O exposure, we next evaluated whether the DCN is critical for this response. To investigate the specific enhancing effects of the DCN on the activity of VTA dopamine neurons, we examined synaptic transmission within this circuit using optogenetic and patch-clamp techniques in brain slices. We injected a ChrimsonR-tdTomato virus into the DCN and confirmed its expression in the VTA (Fig. 4L). After the behavioral test, we recorded the VTA dopamine neurons and stimulated DCN axon terminals in the VTA with 590-nm light. We found that the PPR values in VTA neurons in the 60% N2O group were significantly lower than in the Air group (Fig. 4M), indicating an increased probability of glutamate release from DCN terminals after N2O exposure. Moreover, the AMPAR/NMDAR ratios in VTA neurons were significantly higher in the 60% N2O group than in the Air group (Fig. 4N), suggesting strengthened glutamatergic transmission from the DCN to the VTA. These results suggest that N2O activates DCN neurons and enhances the activity of VTA dopamine neurons during rewarding behavior.

Optogenetic LTD induction reduces glutamatergic transmission in the DCN-VTA neural circuit

Based on the enhancement of AMPAR-mediated glutamatergic transmission by N2O, we used in vivo optogenetic LTD to reduce this transmission in the DCN-VTA circuit and evaluated its effect on N2O-induced CPP behavior (Fig. 5A). Mice from all three groups were injected with ChrimsonR-tdTomato into the DCN (Fig. 5B). After viral expression, optical fibers were implanted in the VTA, and an optogenetics-induced LTD protocol (Fig. 5C) was applied during the behavioral test. Previous studies have shown that the LTD protocol at 1 Hz and 900 repeated pulses successfully decreased glutamatergic transmission [32]. We found that the CPP scores in the LTD group were significantly lower than those in the N2O group (Fig. 5D). The results from paired and unpaired LTD induction in the CPP chamber using Air control mice (Fig. S6) further confirmed that LTD effectively altered conditioned learning behavior. In addition, we assessed the total distance traveled and movement speed of the three groups of mice within the open field box. We found that, compared with the Air group, the 60% N2O and LTD groups did not significantly differ in total distances traveled or movement speed (Fig. S7). Taken together, these findings suggest that the optogenetic induction of LTD in the DCN-VTA circuit decreases N2O-induced rewarding behavior. Since optogenetic LTD induction prevented this behavior, specific changes in the DCN-VTA were evaluated. Following the behavioral experiments, the PPR scores, AMPAR/NMDAR ratios, AMPA receptor-mediated EPSCs, and mEPSCs of VTA neurons were recorded. PPR analysis revealed that the PPR scores in VTA neurons in the 60% N2O group were significantly lower than in the Air group. Furthermore, PPR scores in the LTD group were significantly greater than those in the 60% N2O group (Fig. 5E). The AMPAR/NMDAR ratio in the 60% N2O group showed a significant increase, while the LTD group exhibited a decrease in this ratio compared to the 60% N2O group (Fig. 5F). Additionally, the AMPAR-EPSCs curves of VTA neurons were recorded under optical stimulation at five different intensities. Compared with the N2O group, the LTD group showed markedly lower AMPAR-EPSCs amplitude (Fig. 5G). These findings suggest that optogenetic induction of LTD decreased both presynaptic and postsynaptic glutamatergic transmission between the DCN and VTA. Moreover, compared with those in the 60% N2O group, the frequency and amplitude of the mEPSCs in the LTD group significantly decreased (Fig. 5H, I). Collectively, these findings indicate that optogenetic LTD induction reduces glutamatergic transmission in the DCN-VTA neural circuit.

Fig. 5: Optogenetic LTD induction reduces glutamatergic synaptic transmission in the DCN-VTA neural circuit.The alternative text for this image may have been generated using AI.

A Schematic diagram showing optogenetic LTD induction during CPP procedures. B Virus injection in the DCN. C Application of the LTD protocol. D Left, representative traces in all three groups. Right, statistical analysis of the CPP scores. n = 8 mice per group. F2, 21 = 12.644, ***p < 0.001, one-way ANOVA. ***p < 0.001 for Air vs 60% N2O, **p = 0.007 for LTD vs 60% N2O. E Sample traces (left) and statistics (right) of PPR. n = 14 neurons, 4 mice (Air), n = 19 neurons, 4 mice (60% N2O), n = 17 neurons, 4 mice (LTD). F2, 47 = 4.923, *p = 0.011, one-way ANOVA. *p = 0.022 for Air vs 60% N2O, *p = 0.014 for LTD vs 60% N2O. F Sample traces (left) and statistics (right) of the AMPAR/NMDAR ratios. n = 17 neurons, 4 mice (Air), n = 14 neurons, 4 mice (60% N2O), n = 10 neurons, 4 mice (LTD). F2, 38 = 8.775, ***p < 0.001, one-way ANOVA. ***p < 0.001 for Air vs 60% N2O, *p = 0.012 for LTD vs 60% N2O. G AMPAR-EPSCs curves for the three groups of mice. Left, sample traces; right, statistical analysis. n = 17 neurons, 4 mice (Air), n = 17 neurons, 4 mice (60% N2O), n = 22 neurons, 4 mice (LTD). F2, 279 = 5.463, **p = 0.007, two-way RM ANOVA. **p = 0.006 for Air vs 60% N2O, #p = 0.025 for LTD vs 60% N2O. H Sample traces of mEPSCs in VTA neurons from each group. I Quantification of mEPSCs frequency and amplitude. Frequency: 25 neurons, 4 mice (Air), 13 neurons, 4 mice (60% N2O), and 17 neurons, 4 mice (LTD). F2, 52 = 9.143, ***p < 0.001, one-way ANOVA. ***p < 0.001 for Air vs 60% N2O, *p = 0.016 for LTD vs 60% N2O. Amplitude: 25 neurons, 4 mice (Air), 13 neurons, 4 mice (60% N2O), and 17 neurons, 4 mice (LTD). F2,52 = 6.309, **p = 0.004, one-way ANOVA. **p = 0.002 for Air vs 60% N2O, *p = 0.028 for LTD vs 60% N2O.

Inhibiting VTA dopamine neurons prevents N2O-induced rewarding behavior

Given that N2O activates the DCN-VTA circuit in rewarding behavior, inhibiting VTA dopamine neurons may therefore prevent this behavior. A chemogenetic approach was employed to inhibit the excitability of VTA dopamine neurons, to examine its effects on N2O-induced CPP behavior (Fig. 6A). A mixture of TH-cre and DIO-hM4Di viruses was injected into the VTA region (Fig. 6B, C). This approach specifically inhibited dopamine neuron activity after CNO treatment before the post-test phase. Mice were divided into Air, 60% N2O, and hM4Di groups. We found that the CPP scores in the hM4Di group were significantly lower than those in the 60% N2O group (Fig. 6D). After the behavioral test, we examined the excitability and glutamatergic transmission of VTA dopamine neurons. Analysis of spontaneous firing revealed a significant increase in dopamine neuron activity in the 60% N2O group compared with the Air group. Furthermore, compared with mice in the 60% N2O group, mice in the hM4Di group exhibited significantly reduced spontaneous firing of dopamine neurons (Fig. 6E, F). In addition, a series of step current injections was used to evoke firing, which reflects neuronal excitability. We found that compared with that in the 60% N2O group, the number of firing events in the hM4Di group significantly decreased (Fig. 6G). Finally, the frequency and amplitude of mEPSCs in dopamine neurons were significantly greater in the 60% N2O group than in the Air group. Conversely, compared with the 60% N2O group, the hM4Di group showed significant decreases in the frequency and amplitude of mEPSCs (Fig. 6H–J). Together, these findings suggest that N2O induces rewarding behavior by activating VTA dopamine neurons.

Fig. 6: Inhibiting the function of VTA dopamine neurons prevents N2O-induced rewarding behavior.The alternative text for this image may have been generated using AI.

A Schematic illustration of the experimental procedure for the CPP behavioral test and CNO treatment. B AAV-TH-cre mixed with AAV-DIO-hM4Di was injected into the VTA. C Sample images showing viral expressions in the VTA. D Statistical analysis showing the CPP scores of the mice. n = 8 mice per group. F2, 21 = 9.569, ***p = 0.001, one-way ANOVA. ***p = 0.001 for 60% N2O vs Air, **p = 0.008 for hM4Di vs 60% N2O. E Images showing the VTA dopamine neurons used for recording. F Sample traces (left) and statistical analysis (right) of the spontaneous firing of VTA dopamine neurons in each group. n = 17 neurons, 3 mice (Air), n = 20 neurons, 4 mice (60% N2O), n = 22 neurons, 4 mice (hM4Di). F2, 56 = 4.948, *p = 0.011, one-way ANOVA. *p = 0.017 for Air vs 60% N2O, *p = 0.013 for hM4Di vs 60% N2O. G Sample traces (left) and statistical analysis (right) of evoked firing in VTA dopamine neurons. n = 12 neurons, 3 mice (Air), n = 14 neurons, 4 mice (60% N2O), and n = 19 neurons, 4 mice (hM4Di). F2, 224 = 4.934, *p = 0.012, two-way RM ANOVA. *p = 0.025 for Air vs 60% N2O, ##p = 0.009 for hM4Di vs 60% N2O. H Sample traces showing mEPSCs in VTA neurons. I Statistical analysis of mEPSCs frequency in VTA neurons. n = 11 neurons, 3 mice (Air), n = 13 neurons, 4 mice (60% N2O), n = 13 neurons, 4 mice (hM4Di). F2, 34 = 4.264, *p = 0.022, one-way ANOVA. *p = 0.035 for Air vs 60% N2O, *p = 0.022 for hM4Di vs 60% N2O. J Statistical analysis of the mEPSCs amplitude in VTA neurons. n = 11 neurons, 3 mice (Air), n = 13 neurons, 4 mice (60% N2O), n = 13 neurons, 4 mice (hM4Di). F2, 34 = 3.800, *p = 0.032, one-way ANOVA. *p = 0.041 for Air vs 60% N2O, *p = 0.038 for hM4Di vs 60% N2O.