Male and female rats exhibit comparable gaping behavior but activate brain regions differently during expression of conditioned nausea

Abstract

Twenty-five to fifty percent of patients undergoing chemotherapy will develop anticipatory nausea and vomiting (ANV), in which symptoms occur in anticipation of treatment. ANV is triggered by environmental cues and shows little response to traditional antiemetic therapy, suggesting that unique neural pathways mediate this response. Understanding the underlying neural mechanisms of this disorder is critical to the development of novel therapeutic interventions. The purpose of the present study was to identify brain areas activated during ANV and characterize sex differences in both the behavior and the brain areas activated during ANV. We used a rat model of ANV by pairing a novel context with the emetic drug lithium chloride (LiCl) to produce conditioned nausea behaviors in the LiCl-paired environment. We quantitated gaping, an analogue of human vomiting, after acute or repeated LiCl in a unique environment. To identify brain regions associated with gaping, we measured c-fos activation by immunochemical staining after these same treatments. We found that acute LiCl activated multiple brain regions including the supraoptic nucleus of the hypothalamus, central nucleus of the amygdala, nucleus of the solitary tract and area postrema, none of which were activated during ANV. ANV activated c-fos expression in the frontal cortex, insula and paraventricular nucleus of the hypothalamus of males but not females. These data suggest that therapies such as ondansetron which target the area postrema are not effective in ANV because it is not activated during the ANV response. Further studies aimed at characterizing the neural circuits and cell types that are activated in the conditioned nausea response will help identify novel therapeutic targets for the treatment of this condition, improving both quality of life and outcomes for patients undergoing chemotherapy.

Introduction

Chemotherapy-induced nausea and vomiting (CINV) is one of the most common and debilitating side effects of cancer treatment ^1–5. Approximately 25–50% of patients will develop anticipatory nausea and vomiting (ANV), in which patients experience symptoms prior to administration of the toxic drug ^{2, 6–12}. ANV is a conditioned response in which contextual stimuli associated with previous chemotherapy, rather than chemotherapy agents themselves, trigger the onset of nausea and/or vomiting. The likelihood of ANV increases with the emetogenicity of the chemotherapy agent, the incidence of delayed nausea and vomiting, increasing numbers of chemotherapy treatments, and sex (women more affected) ^{2, 9–11, 13}. ANV can be refractory to treatment, reducing quality of life and, more importantly, resulting in cessation of treatment. However, little is known about the neurobiological mechanisms underlying this disorder. As such, this condition represents a critical unmet need that affects vulnerable patients, and women particularly.

The current standard of therapy for CINV consists of a cocktail of preemptive treatment with a 5HT3 antagonist, the atypical antipsychotic olanzapine (D2, 5HT2, and 5HT3 antagonist), the NK1 antagonist aprepitant, and the steroid dexamethasone ^{5, 14–18}. Blockade of dopamine (D2) and serotonin (5HT3) receptors in the area postrema (AP) and NK1 receptors in the nucleus of the solitary tract (NTS) block the ascending sensory stimuli that trigger vomiting ^19–29. These sites are especially effective at blocking incoming afferent signals from the gut (via the AP) and the vagus nerve (via the NTS), and the acute nausea triggered by toxic substances. Animal literature has also shown that completely effective blockade of acute nausea can substantially reduce or prevent expressions of conditioned nausea including gaping^{30, 31}. Conversely, similar studies show that these drugs are less effective at treating ANV once it has developed, suggesting that the mechanisms of contextual nausea/vomiting are fundamentally different than those induced by a toxin. However, the neural circuit that controls expression of ANV is poorly understood^{12, 32}. Promising studies have implicated the ability of cannabinoids and related fatty acids to blocked expression of conditioned gaping through actions in the ventral pallidum and insular cortex^{33, 34}. But beyond these interesting studies, the brain areas involved in ANV are poorly characterized.

The present study took a non-pharmacologic approach to further investigate the possibility that that ANV is controlled by a distinct neural circuit from that of acute vomiting. LiCl activates the nausea circuit in rats through direct activation of the area postrema (AP). The AP activates the vomiting response through coordinated activation of brain regions, including the nucleus of the solitary tract, parabrachial nucleus, amygdala, paraventricular nucleus, supraoptic nucleus, insula, and frontal cortex^35–39. These regions are responsible for the neuromuscular and autonomic responses in the vomiting circuit in both rats and humans ^{40, 41}.

We hypothesized that brain regions activated during the acute nausea response are not activated by the conditioned nausea response, while brain regions that activated in the contextual stimulus would respond to the LiCl conditioned context. To test this hypothesis, we examined the neural activation of multiple brain regions associated with acute nausea, such as the area postrema (AP), nucleus tractus solitarius (NTS), central and basolateral nuclei of the amygdala (CeA, BLA), and supraoptic nucleus (SON) which we predicted are not activated during conditioned nausea, and regions that are responsive to contextual cues, such as the ventromedial prefrontal cortex (vmPFC), agranular and granular insula (aIC, gIC), the nucleus accumbens core (NAcC) and the paraventricular nucleus of the hypothalamus (PVN) which we predicted would respond to the LiCl-conditioned context.

In order to better understand how the conditioned nausea response differs from the acute response, we utilized a rodent model of ANV that used gaping to assess nausea and activation of the immediate early gene c-fos to define brain areas activated by the contextual stimuli associated with treatment. In this paradigm, a novel context (test cage) was repeatedly paired with the emetogenic agent LiCl. LiCl was selected for its quick onset and relatively short half-life, high emetogenicity, and established use in conditioned nausea paradigms^42–46. Rats and other rodents, like humans, express both acute and conditioned behavioral responses to emetic agents like LiCl. Rats lack anatomical and likely neurological drivers of emesis and thus do not vomit⁴⁷. However, they exhibit gaping, a behavioral response that resembles human vomiting, engaging the same muscles as vomiting and responsive to drugs that reduce nausea and vomiting in humans^48–50. Conditioned gaping (CG) is a behavior exhibited by rats after repeated pairings of an emetic agent with a contextual stimulus, such as a novel cage. After multiple pairings, rats will exhibit gaping behavior when placed in the context, even in the absence of the toxin^{51, 52} We used a high dose of LiCl that produces significant conditioned gaping ^{44, 53–55}.

We studied both male and female rats, given the female predominance of ANV and the relative paucity of data from females in the experimental literature. We monitored gaping after an acute dose of LiCl and in response to the LiCl-conditioned context. We then analyzed c-fos induction after acute LiCl and in response to the LiCl-conditioned environment. Areas analyzed included brain regions previously associated with LiCl and/or the acute vomiting reflex, including the AP, NTS, SON, and CeA, as well as brain regions that we anticipated would respond to the contextual stimulus, including the vmPFC, aIC, and PVN^{35, 56–61}. While the acute vomiting response is well-described in rats^{40, 62, 63}, there is little data regarding the neural activation during the conditioned nausea response. We hypothesized that brain regions associated with the acute vomiting response would not be involved in the conditioned vomiting response.

Materials and Methods

Animals

Male and female Sprague Dawley rats (PN60, Charles River Laboratories, Raleigh, NC) were received 1 week before behavior tests. Rats were housed by sex in ventilated polypropylene cages (41 × 25 × 18 cm) in a colony room with a temperature of 21.1 ± 1.0 °C. They were given ad libitum access to food (Purina Lab Diet for rodents #5001) and water on a 12-hour light/dark cycle with lights on from 0700–1900. Humidity ranged from 30–70%. Cages were cleaned twice a week. Temperature and humidity were monitored automatically 24 hours/day and recorded daily. Females were randomly cycling and not selected based on estrous cycle stage. Cycle was determined in females by lavage on the day of behavioral testing. All studies were conducted in accordance with the Guide for Care and Use of Laboratory Animals and were approved by the Duke University IACUC protocol number A090-19-04.

Conditioned Nausea

The experimental design, based on the model described by Parker, et. al,⁵¹ is shown in Fig. 1A. On day 0, animals received 96 mg/kg LiCl (15.1 mL/kg of 0.15M, ip) or the equivalent of isotonic NaCl and placed in a novel polypropylene cage (35 × 25 × 13 cm). Gaping behavior was recorded for the first 10 minutes on an iPad placed beneath the clear plastic cage. Animals that were collected for acute condition were then perfused and brains collected for c-fos analysis.

Figure 1.

(A) Experimental protocol for conditioned nausea. (B) Normal mouth position. (C) Gaping position (D) Gaping behavior by sex on days 0 and 8. N=8 for NaCl-treated males, n=5 for NaCl-treated females, n=10 for LiCl-treated males, n=11 for LiCl-treated females.

Animals taken through the conditioning protocol were given repeat injections of 96 mg/kg LiCl or the equivalent of NaCl and placed in the same context on alternate days for 4 conditioning sessions. Animals were weighed daily to monitor for excessive weight loss. Weights are listed in Table 1 below. On the drug-free test day, animals were placed in test cage for 60 minutes. Gaping behavior was recorded for the first 10 minutes in the test cage as on day 0. Rats were then perfused, and brains collected for c-fos analysis.

Table 1.

Average weights (g) on days 0 and 8

	Males		Females
	Day 0	Day 8	Day 0	Day 8
NaCl	344±15	408±16	249±3	262±6
LiCl	405±29	443±25	233±8	244±8

Video scoring of gaping behavior

Videos collected from days 0 and 8 were analyzed for gaping behavior. The scorer was blinded to experimental condition. Gaping behavior was measured for the first 10 minutes in the test cage. The scorer counted individual gapes (Fig. 1B, C) and total number of gapes during the 10-minute period was calculated.

Drugs

LiCl was purchased from Sigma Aldrich (product number 203637). LiCl was dissolved in distilled sterile water to a concentration of 0.15M. Sterile isotonic saline (NaCl, 0.9%; 0.15M) was used as a control injection.

Transcardial perfusion and freezing of brains

Animals were deeply anesthetized with urethane (3 mg/kg). We used toe-pinch to test for responsiveness, administering additional urethane if necessary. Perfusion was performed using ice-cold 1X phosphate-buffered saline (PBS) to exsanguinate, followed by 4% paraformaldehyde (PFA) for fixation.

The brain was carefully removed and placed in a vial containing 4% PFA. Brains were stored at 4°C overnight, then cryoprotected for 3–7 days in 30% sucrose in 1X PBS solution at 4°C. Brains were rapidly frozen in 1:2 ratio of Tissue-Tek® OCT Freezing Media and 30% sucrose in EtOH/dry ice slurry and stored at −80°C until ready to be cut for c-fos analysis.

Immunohistochemistry

Brains were cut on a Leica CM3050S cryostat at 30 uM and stored in 1:1 TBS/glycerol solution at −20°C until ready to stain. To stain, slices were first washed in 0.2% Triton-X in Tris-buffered saline (TBS) solution 3 times × 10 minutes. Slices were then blocked for 1 hour in solution of 0.3% Triton-X and 5% normal goat serum (NGS) in PBS. Slices were stained overnight in 0.3% Triton-X, 5% NGS, and 1:20,000 anti-c-fos antibody (Abcam - ab190289) with gentle shaking at 4°C. Slices were then washed 3 times in solution of 0.3% Triton-X and 5% NGS in PBS (10 minutes first wash, 30 minutes second wash, 40 minutes third wash). Slices were then stained with secondary antibody (AlexaFluor 488, Invitrogen – A-11034) in 0.3% Triton-X, 5% NGS, and 1:200 secondary antibody for 2 hours. Slices were then washed in 1X PBS (10 minutes first wash, 30 minutes second wash, and 40 minutes third wash), with DAPI (R&D Systems 5748) added at a dilution of 1:10,000 for last 10 minutes of second wash. Slices were mounted on VWR Superfrost® Plus microscope slides with a drop of Vecta-Shield anti-fade medium (Vector Laboratories 101098-042).

Imaging and analysis

Slides were imaged with an Axio Imager upright microscope at 20x (16 z-stacks 2 uM apart) for IHC or 20x (8 z-stacks 1 uM apart) for RNAScope. Images were z-projected for max intensity using FIJI (ImageJ) software and Fos+ neurons were manually counted. 2–6 slices per region were imaged and counted per animal. Counts were averaged per animal. Brain region coordinates are described below:

Brain region	Coordinates (mm from Bregma)
vmPFC	+3.54 to +2.28
aIC, gIC, NAcC	+2.28 to +1.44
SON	−1.52 to −1.20
PVN	−1.80 to −2.04
CeA, BLA	−2.40 to −2.92
NTS	−13.20 to −13.80
AP	−13.68 to −14.16

Statistics

All results were analyzed by ANOVA with posthoc Fisher’s exact test corrected for multiple comparisons using the statistical package NCSS. Behavioral results were analyzed by 3-way ANOVA (sex × treatment × day) followed by lower order ANOVA for significant interactions. C-fos responses were analyzed by 3-way repeated measures ANOVA (sex and condition as between measures and brain area as repeated measure). Individual ANOVAs were also run for each brain area independently as area was a strong main effect in the 3-way ANOVA. N for all behavior and c-fos experiments was estimated by power analysis and estimated variance based on previous experience with CTA behavior and c-fos analysis of multiple brain regions ^64–67.

Results

LiCl produces robust conditioned gaping

We performed 3-way RMANOVA of sex × treatment × day for gaping on Day 0 and day 8 (Fig. 1D). We found a main effect of treatment [F(1, 29)=10.06, p=0.004], a main effect of day [F(1, 29)=9.87, p=0.004], and an interaction of treatment × day [F(1, 29)=10.28, p=0.003]. These results showed that LiCl-treated animals gaped day 8 but not on day 0, and that NaCl-treated controls did not gape significantly on either day. There was no effect of treatment × sex or sex × day.

On day 0, we found no effect of either treatment or sex. On day 8, we found a main effect of treatment [F(1,28)=14.37, p=<0.005], and no effect of sex or interaction of sex × treatment. Males and females did not differ significantly in their gaping behavior. The LiCl group gaped significantly more than NaCl control by Fisher’s posthoc.

The neural activation of acute and conditioned nausea

We performed immunohistochemistry (IHC) for c-fos in brain regions after acute NaCl or LiCl, and after conditioning with NaCl or LiCl (Table 2). We performed a repeated measures ANOVA (sex × treatment, area as repeated measure) on all conditions. We found a main effect of treatment [F(3, 38)=25.13, p=<0.001], a main effect of area [F(6, 37)=836.07, p=<0.001], an interaction of sex × area [F(10, 136)=28.12, p=<0.001], an interaction of treatment × area [F(18, 136)=43.37, p=<0.001], and a three-way interaction of treatment × sex × area [F(18, 136)=43.07, p=<0.001]. We then performed lower-level ANOVAs by brain areas individually.

Table 2.

c-fos±SEM in all brain regions

			vmPFC	aIC	gIC	NacC	SON	PVN	BLA	CeA	NTS	AP
Acute	Male	NaCl	320±17	211±14	114±19	96±22	11±4	266±92	40±14	45±9	39±3	26±2
		LiCl	249±8	187±28	132±15	101±18	215±8	358±35	49±10	244±45	89±7	75±10
	Female	NaCl	340±10	288±35	100±31	200±33	17±4	217±69	40±12	66±21	25±7	15±7
		LiCl	311±8	289±28	134±15	185±18	174±8	323±35	62±10	198±45	107±7	68±10
Conditioned	Male	NaCl	143±13	145±9	78±36	90±14	8±5	110±31	15±2	36±9	11±4	13±5
		LiCl	241±9	210±10	75±18	100±17	14±7	160±12	32±5	42±11	16±3	10±1
	Female	NaCl	255±26	207±14	87±6	121±14	12±2	121±31	30±4	17±1	20±7	19±7
		LiCl	231±16	221±10	60±17	126±17	15±5	93±16	38±6	27±8	5±1	3±1

Fig. 2 shows c-fos expression in the parts of the brain previously associated with acute nausea. The area postrema (AP) showed an effect of treatment [F(3, 22)=16.03, p=<0.001], with the acute LiCl condition showing increased expression compared to all other conditions. There was no increased expression in the conditioned LiCl rats (Fig. 2A–C). We found a similar pattern in other parts of the brain associated with acute nausea. The nucleus tractus solitarius (NTS) showed an effect of treatment [F(3, 26)=63.66, p=<0.001], with only acute LiCl differing from all other conditions, and no increased expression the conditioned LiCl group (Fig. 2D–F).

Figure 2.

(A) c-fos activation of the AP in male and female rats in acute and conditioned context, n=3–6. (B) Atlas image of AP. (C) Representative images of the AP (D) c-fos activation of the NTS in male and female rats in acute and conditioned context, n=3–7. (E) Atlas image of NTS. (F) Representative images of the NTS. (G) c-fos activation of the CeA in male and female rats in acute and conditioned context, n=3–6. (H) Atlas image of CeA. (J) Representative images of the CeA (K) c-fos activation of the SON in male and female rats in acute and conditioned context, n=3–5. (L) Atlas image of SON (M) Representative images of the SON. Data shown as mean ± SEM.

The central nucleus of the amygdala (CeA) showed an effect of treatment [F(3, 24)=30.52, p=<0.001], with acute LiCl increased compared to all other groups, and no increased expression in the conditioned LiCl group (Fig. 2G–J).

The supraoptic nucleus (SON) showed an effect of treatment [F(3, 19)=530.05, p=<0.001], as well as an interaction of sex × treatment [F(3, 19)=7.82, p=0.001]. Males showed greater expression in the acute LiCl condition than females. There was no increase in c-fos in the conditioned LiCl group (Fig. 2K–M).

Unlike the brain regions associated with the acute nausea circuit, the vmPFC, aIC, and PVN showed higher levels of c-fos expression in the conditioned LiCl context compared to the conditioned NaCl context in male rats, as shown in Figure 3. The vmPFC showed a main effect of sex [F(1, 30)=15.28, p=<0.001], a main effect of treatment [F(3, 30)=21.58, p=<0.001], and an interaction of treatment × sex [F(3, 30)=5.30, p=0.005]. Overall, c-fos in females was higher than males in the vmPFC. Males showed greater expression in the acute NaCl context compared to the acute LiCl context. Unlike areas profiled above, c-fos expression was higher in LiCl-conditioned males than NaCl-conditioned males. C-fos in females, in contrast, did not differ between the acute conditions (Fig. 3A–C).

Figure 3.

(A) c-fos expression in the vmPFC in male and female rats in acute and conditioned context, n=4–6. (B) Atlas image of vmPFC. (C) Representative images of the vmPFC. (D) c-fos expression in the aIC in male and female rats in acute and conditioned context, n=4–6. (E) Atlas image of aIC. (F) Representative images of the aIC. (G) c-fos expression in the PVN in male and female rats in acute and conditioned context, n=3–6. (H) Atlas image of PVN. (J) Representative images. Data shown as mean ± SEM.

The agranular insula (aIC) showed a main effect of sex [F(1, 32)=20.88, p=<0.005], and a main effect of treatment [F(3, 32)=5.46, p=0.004]. C-fos in females was increased compared to males. There was no difference between acute NaCl and acute LiCl conditions in either sex. C-fos in the Conditioned LiCl group was increased compared to conditioned NaCl in males, but not females (Fig. 3D–F).

The paraventricular nucleus (PVN) showed a main effect of treatment [F(3, 20)=9.83, p=<0.001]. There was no overall effect of sex. We then considered the acute and conditioned groups separately. The acute groups showed no differences between treatments. The conditioned groups showed an interaction of sex × treatment [F(3, 20)=9.83, p=<0.001]. C-fos in males showed an increase in the conditioned LiCl group compared to the conditioned NaCl group. C-fos in females in each group did not differ (Fig. 3G–J).

The nucleus accumbens core (NAc_C), granular insula (gIC), and basolateral nucleus of the amygdala (BLA) did not show a significant response to the acute or conditioned context, and so will not be discussed further.

Discussion

This behavioral paradigm produced robust CG in both male and female rats. LiCl induced similar levels of CG in male and female rats after 4 pairings with the contextual stimulus, while demonstrating minimal gaping behavior after the first exposure to LiCl. -fos experiments showed that the brain areas that are activated by acute LiCl are not activated during the conditioned response. Meanwhile, in male rats, brain regions that are responsive to novel context (i.e., activated in the acute condition by both NaCl and LiCl) continue to show high levels of c-fos activation on test day in the LiCl-paired context. Other studies support the role of the insula in this conditioned response ^{51, 68–70}. The present study expands these brain areas to include the ventromedial prefrontal cortex and paraventricular nucleus in male rats. C-fos in female rats likewise did not respond in brain regions associated with acute nausea (AP, NTS, SON, and CeA) in the conditioned task relative to the acute challenge but demonstrated habituation of the response to the cage (responses to both NaCl and LiCl) instead in the vmPFC, aIC, or PVN.

We speculate that the context previously associated with nausea triggers the gaping behavior without activating the acute nausea circuit. The contextual stimulus produced robust gaping behavior, while the brain regions normally responsible for initiating the vomiting reflex were silent. The novel context produced high levels of c-fos expression in the vmPFC, aIC, and PVN in both the NaCl- and LiCl-treated groups, but conditioned diminution of c-fos expression in the NaCl-associated context and conditioned activation in the LiCl treated male rats.

These results lay potential groundwork for further defining the neural substrates of conditioned nausea. Clarification of these substrates and cell-specific analyses will be useful in deepening our understanding of the neuronal subtypes that are activated by the contextual stimulus and promote the conditioned gaping response.

While males and females elicited similar conditioned gaping, the c-fos responses to the conditioned contexts differed by sex. Females showed neither an exaggerated c-fos response in brain regions after exposure to the conditioned environment nor habituation in the NaCl conditioned environment. In contrast, in males, c-fos expression in these brain regions (vmPFC, aIC, and PVN) showed habituation in the NaCl-associated conditioned context relative to the acute NaCl condition and higher c-fos expression was maintained in the LiCl-conditioned context. Further studies that assess a broader range of brain areas are necessary to determine whether this may be due to variances in c-fos expression due to estrous cycle or if other brain regions are responsible for the conditioned response in females.

An important area of further study is whether other long-term transcription factors may be present, such as ΔFosB, a member of the fos family of transcription factors that accumulates over time ^71–74. It is possible that longer term changes in neural activity may be reflected in different transcription factors.

The current study was limited by the subset of brain regions that are associated with caudal processing of the acute stimulus. The course of afferent acute stimuli is well characterized: it flows from the area postrema that we sampled, rostrally through the bed nucleus of the stria terminalis to the parabrachial nucleus and on to the visceral (granular) cortex ⁶⁹. The present study was not a comprehensive study of enough areas to integrate the current findings with what is known about the neural circuit for the conditioned nausea response but it provides insight into core areas involved. Further studies would also benefit from cell-specific analysis of the insula in particular to determine whether the same cell types are responding to both the acute and conditioned stimuli. Studies show that the insula has a role in assessing toxic inputs ^{75, 76}. Better characterization of the specific cell populations that respond to the acute and conditioned LiCl contexts could provide valuable insight into how the insula assesses visceral versus contextual stimuli. An additional caveat is that, since there is no control of animals receiving LiCl in their home cages, it is difficult to conclude that behavior and c-fos activation is due to the contextual stimulus.

Our studies showed that the neural activation occurring during conditioned gaping differs dramatically from that occurring during the acute stimulus with an emetogenic agent. Conditioning to contextual stimuli may contribute to the pathogenesis of ANV, which suggests that quite different pharmacotherapies might prove effective. It is known that aggressive anti-emetic therapy at the initiation of treatment is extremely important to prevent the development of ANV and has diminishing returns with repeated treatments ^{12, 32, 77, 78}. Our study supports the finding that pharmacotherapy targeting brain regions associated with the acute vomiting stimulus is likely to be less effective for ANV. Further clarification of the circuit controlling the conditioned stimulus is important for laying a groundwork to better our understanding of this difficult disorder, and to has the potential to identify novel targets to improve pharmacotherapy.