Understanding of Thirst in Medical Science

ABSTRACT

Thirst is fundamentally considered as a physiological function developed to maintain the homeostasis of body fluids. However, we occasionally experience thirst in situations that are not necessarily related to the maintenance of bodily fluid homeostasis. Because the only method available had been to quantify the degree of thirst using psychological indices, thirst research had made little progress until very recently. To quantitatively analyze thirst, it is necessary to elucidate the nature of the brain’s “thirst center,” which is believed to become active in response to thirst. Textbooks of physiology often refer to the “thirst center” in the hypothalamus, which is considered to sense the osmotic pressure of body fluids. However, they did not specify the location of this center in the hypothalamus. Furthermore, the existence of the so-called “osmotic pressure sensors” has yet to be confirmed. However, recently, a series of findings have been published that delve into the true nature of the thirst center. These advancements have been achieved using several new techniques, including the real-time monitoring of neural activities related to thirst regulation within the brains of experimental animals. At least at the animal level, recent advancements in experimental techniques have made it possible to objectively quantify the intensity of thirst as a physiological response. In this review article, the history of our research is presented and latest developments in thirst research are presented.

THE WATER ENVIRONMENT IN THE BODY AND THIRST

Homeostasis of water in the body

The human body contains large amounts of water. The percentage of body weight made up of water varies depending on sex and fat content, but it is typically 50–70%, and approximately 80% for newborns. For example, a 60-kg adult has approximately 60% of body weight in water, which equates to approximately 36 L. However, two-thirds of this water is found inside the cells; therefore, only approximately 12 L is considered body fluid. Of these, approximately 4.8 L is present in the blood, which is approximately 8% of the body weight.¹

We lose approximately 1.6 L of water daily through excretion, and an additional 0.9 L through breathing and sweating. In total, we need to replace 2.5 L of water each day. Typically, we gain approximately 1 L from food and 0.3 L through metabolic processes inside the body, leaving the remaining 1.2 L to be consumed through beverages. However, because humans are terrestrial, they do not always have access to fresh water, and a lack of water leads to dehydration. Various abnormalities appear in the body when dehydration becomes severe (Table 1). To avoid this and maintain homeostasis, water must be consumed while minimizing loss.

Table 1.  Fluid Loss Percentage and Physical Symptoms.

2%	Thirst
3%	Strong thirst, loss of appetite, dizziness
4%	Irritability, increased body temperature, fatigue, darker urine with reduced volume, reddened skin
5%	Headache, flushing
8–10%	Seizures, trembling, dizziness
20%	No urine output, cessation of life functions

However, because the regulation of body temperature is essential, it is difficult to suppress sweating. To counteract dehydration, the body first triggers the sensation of thirst, which prompts us to seek water. Simultaneously, the antidiuretic hormone (vasopressin) is released to reduce the amount of water excreted by the kidneys.

Sensation of thirst and saliva

As mentioned above, thirst is fundamentally considered as a physiological function developed to maintain the homeostasis of body fluids. However, we occasionally experience thirst in situations that are not necessarily related to the maintenance of bodily fluid homeostasis (Fig. 1). To distinguish this from dehydration, we refer to it as the “sensation of thirst.” Some of these sensations are thought to be influenced by salivary secretions. Salivary glands, which produce saliva, are regulated by opposing signals from the sympathetic and parasympathetic nervous systems. Signals from the parasympathetic nervous system promote the secretion of watery saliva, whereas those from the sympathetic nervous system induce the secretion of thicker and more viscous saliva (Fig. 2). Adults secrete approximately 1–1.5 L of saliva daily. Generally, the sensation of “saliva flow” is associated with the secretion of the watery type of saliva.

Fig. 1.

 Signals that elicit thirst sensation.

Fig. 2.

 Neural circuit for saliva controls.

For instance, there is a traditional teaching in Japan that, when thirsty, one can imagine pickled plums “Umeboshi” to stimulate saliva and endure the thirst. Memory or smell of pickled plums triggers salivary secretion, temporarily alleviating the sensation of thirst. Similarly, chewing gum or engaging in chewing motions can stimulate saliva production, reducing the sensation of thirst.² However, age-related muscle weakness can lead to oral dryness (dry mouth), resulting in a sensation of thirst. Additionally, when one is nervous, such as when speaking in front of an audience, the sympathetic nervous system activation suppresses the secretion of watery saliva, causing a dry throat. Prolonged activation of the sympathetic nervous system due to chronic mental stress can lead to dry mouth.

Although not discussed in detail here, the reduction in salivary secretion caused by this mental thirst has a substantial impact on taste perception, chewing, swallowing, digestion, speech, and even the body’s ability to defend against infections. Therefore, fundamental treatments for this condition are urgently required.

Preabsorptive satiety

Stimuli, such as placing something cold, like ice, in the mouth can reduce the sensation of thirst. Even placing a chilled metal sphere in the mouth lowers thirst. The sensation of cold is perceived by a sensor molecule called TRPM8, which is present on the surface of cell membranes.³ TRPM8 is activated not only by cold temperatures, but also by menthol, which can create a false sensation of coldness and subsequently reduce thirst. Therefore, the addition of menthol to a refreshing beverage can alleviate thirst when menthol reaches the mouth. This process occurs much faster than when the body absorbs water and resolves dehydration.

As a result, some consumers may prefer mentholated beverages to plain water, because the former provides immediate relief from thirst. This phenomenon, in which satisfaction arises before water or nutrients are actually absorbed and the internal environment is corrected, is known as “preabsorptive satiety.” This process involves sensory molecules in the mouth and intestines. For example, glucose can be detected in the intestines, which induces a sense of satisfaction. In contrast, artificial sweeteners create a sweet taste in the tongue, but do not induce satisfaction originating from the intestines.⁴

Similarly, administering water to the stomach or intestines can suppress thirst.⁵ This “water” sensation is known to be transmitted to the brain via the afferent parasympathetic nerves.⁶ Moreover, studies of neural activity in the brain’s thirst center have revealed that the neural activity activated by dehydration can be suppressed merely by seeing water.⁵

Other factors regulating thirst

However, this type of pre-absorptive satiety is transient. Even after drinking begins, it takes time for the osmotic pressure of the body fluids to return to normal.⁷ The sensation of water in the mouth is believed to be mediated by the acid-sensing protein, PKD2L1. When the drinking spout of a water dispenser does not release water, but is designed to artificially stimulate PKD2L1-expressing water-sensing cells in the mouth, dehydrated mice will continue licking the spout as though water were being dispensed.⁸ The mice persistently lick until actual water is provided. Ultimately, merely sensing “water” in the mouth does not alleviate the sensation of thirst.

When the water consumption of mice was measured over a 24-hour period, a peak in the amount of water consumed was observed immediately after waking (for nocturnal mice, this occurred just after nightfall), and another peak occurred before sleeping. The mechanism underlying this pre-sleep drinking behavior has been studied, and it has been reported that thirst is induced by neural activity originating from the suprachiasmatic nucleus, which regulates the 24-hour circadian rhythm of the brain.⁹ However, there are no data to confirm whether humans also feel thirst before going to sleep.

Thirst as a complex sensation

As described, what we refer to as thirst is a complex sensation influenced by various external and internal factors, including body fluid homeostasis, psychological stress, and visceral sensations. To develop beverages that can effectively alleviate thirst, it is necessary to obtain objective data while considering the effects of these complex factors. Traditionally, there is no method to objectively measure the sensation of thirst. However, recent academic and technological advancements have revealed the nature of the thirst center of the brain, making it increasingly possible to quantitatively evaluate its activity. By studying the neural activity of the thirst center, we can identify the components that effectively reduce the sensation of thirst.

In the following sections, the process of our research that led to the discovery of the nature of the thirst center is outline and the latest findings in this field are presented.

MECHANISMS FOR MAINTAINING THE BALANCE OF WATER AND SALT IN BODY FLUIDS

Body fluids and salt concentration

Body fluids are liquids with an osmotic pressure of approximately 300 mOsm. Saline solutions commonly used in medical settings are composed of ~0.9% sodium chloride (NaCl). Sodium ions (Na⁺), a major component of body fluids, are maintained at 135–145 mEq (mM) in most organisms. Body fluids are extracellular fluids, therefore, Na⁺ levels within cells are maintained at only a few mEqs. Organisms utilize this Na⁺ concentration gradient between the inside and outside of the cells for various purposes. For example, nerve cells generate electrical signals by rapidly allowing Na⁺ (a cation) to flow into the cells. Individual cells use numerous transporters (proteins embedded in the lipid cell membrane) to selectively transport or expel the necessary substances across the cell membrane. This process often harnesses energy derived from Na⁺ moving across the membrane to transport substances into or out of the cell. Additionally, because Na⁺ constitutes a substantial proportion of the solutes in body fluids (extracellular fluids), fluctuations in Na⁺ concentration can substantially affect osmotic pressure. For example, in cases of low osmotic pressure, water flows into the cells through the semipermeable cell membrane, causing cells to swell and potentially sustain damage. Therefore, organisms maintain high Na⁺ concentrations in body fluids with high precision. Notably, although the range of Na⁺ concentrations in body fluids (135–145 mEq) may vary among individuals; however this concentration is tightly regulated within each individual.

Figure 3 shows the state of body fluids, with a focus on the balance between water and Na⁺. For example, during dehydration, the loss of water from body fluids leads to a relative increase in the Na⁺ concentration. Because water is scarce, the body cannot excrete Na⁺ through the urine, which triggers thirst and encourages water intake. To prevent further increases in the Na⁺ concentration, the preference for salty solutions decreases, which discourages salt intake. Typically, animals, including humans, consume saline solutions with a concentration similar to that of body fluids (approximately 0.15 M), which are the most palatable. This holds true regardless of the type of salty liquid, such as miso soup or broth, because the acceptable salt range remains largely consistent.¹⁰

Fig. 3.

 State of body fluids and physiological responses.

In a two-bottle test, in which mice were presented with pure water and saline, mice in environments without other sources of salt consumed almost equal amounts of water and 0.3 M saline. Combined, this intake matches the body fluid level of approximately 0.15 M, indicating that mice balance their intake well. However, when the saline concentration exceeds 0.3 M, mice avoid the saline and consume more water. Notably, in a dehydrated state, their preference shifts to avoid even 0.3 M saline entirely. This suggests the existence of neural circuits that monitor body fluid conditions and accordingly adjust preferences.

Discovery of the brain’s Na⁺ sensor, Na_x

Whether it is possible for a protein to serve as a highly sensitive Na⁺ sensor capable of detecting fluctuations as small as a few millimolars was unclear. In 2002, we discovered a protein with an Na⁺ sensor function.¹¹ It has a structure similar to that of the protein responsible for generating electrical signals in neurons (voltage-dependent Na channels). However, because a critical part of the structure required for the voltage sensor was defective, it was regarded as a mysterious molecule with unknown function. This molecule is subsequently referred to as Na_x.

Interestingly, when Dr. Eiji Watanabe, an associate professor of National Instititute for Basic Biology, analyzed mice lacking Na_x (Na_x-knockout mice), he found that while these mice exhibited no abnormalities under normal conditions, they showed unusual salt intake behavior during dehydration.¹² As described earlier, 0.3 M saline should be avoided in a dehydrated state; however, Na_x-knockout mice did not avoid it. Moreover, examinations of electrical signals from their tongues revealed no abnormalities in their salt taste response. This raised the question: where and how does Na_x function? This became the focus of our next research endeavor.

Sensory circumventricular organs (sCVOs): specialized neural nuclei for monitoring body fluids

Na_x is present in only a few small regions of the brain. These include the subfornical organ (SFO) and organum vasculosum of the lamina terminalis (OVLT), which are part of a specialized group of brain regions known as sCVOs (Fig. 4).¹² The brain and spinal cord are protected by cerebrospinal fluid (CSF), a body fluid that is more strictly controlled than fluids in other parts of the body. Brain cells use various trace substances for signal transmission, therefore, the unrestricted entry of such substances from the bloodstream could have serious consequences. To prevent this, most brain regions have a unique barrier structure in their capillaries called the blood-brain barrier, which restricts the free passage of blood components. However, sCVOs lack this barrier, allowing substances in the blood to freely reach close to the nerve cells. Therefore, the sCVOs are specialized brain regions designed to monitor blood-borne substances.

Fig. 4.

 The sCVOs and SFO. Horizontal (upper left) and sagittal (upper right) planes of human brain showing the location of SFO (lower) and the other sCVOs.

As their names suggest, they are located along the walls of brain ventricles. The ventricles are cavities within the brain. Every day, approximately 0.5 L of fresh CSF is produced from filtered blood, and the entire CSF volume is replaced three to four times a day. The CSF flows from the lateral ventricles, located on either side of the brain, toward the centrally located third ventricle. Among the sCVOs, the SFO is located precisely at the opening where the CSF flows from the lateral ventricles into the third ventricle (Monroe foramen). Positioned at this frontline of exposure to the CSF, the SFO is believed to monitor also the components of the CSF (Fig. 4).

The SFO as the central regulator of salt appetite

Gene therapy experiments were conducted to restore Na_x specifically to localized regions of the brain in Na_x-deficient mice using a viral vector. Notably, only when Na_x was restored to the SFO, among the sCVOs, did the salt intake behavior return to normal.¹³ This finding suggested that Na_x might be involved in detecting dehydration states in the SFO. To investigate this, cells from SFO were dispersed, and the extracellular sodium concentration was altered. The cells responded only when the sodium concentration exceeded the physiological levels found in body fluids. Notably, cells from Na_x-deficient mice did not exhibit such responses. Furthermore, the response of Na_x positive cells was not influenced by changes in osmotic pressure.¹¹ When Na_x expression was induced in the cells that naturally lacking Na_x, the response reappeared.^{11, 14,15,16} These results confirm that Na_x functions as a Na sensor.

It was hypothesized that when dehydration causes the Na level in body fluids to increase, the brain uses Na_x in the SFO to detect this condition, triggering a change in preference to avoid salt. Initially, it was believed that Na_x, as a sensor molecule, is located in neurons capable of generating electrical signals. However, further investigation revealed that Na_x is present in glial cells.¹⁴ Glial cells are traditionally considered to play a supportive role in neuronal activity and are not known to generate electrical signals in neurons. The discovery that the sensor was located in glial cells was entirely unexpected, leaving us uncertain about how to proceed with our investigations.

Lactate signaling via the Na_x complex

To gain insights into these research prospects, a comprehensive screening of molecules that bind to Na_x was conducted. The screening was carried out by Dr. Hidetada Shimizu, a graduate student of my group at the time.¹⁵ Ion channels are molecules that can function as pores to allow ions to pass, and there were few examples of screening the binding molecules for such channels at that time. By exploring five intracellular regions of Na_x, multiple binding molecules were identified (Fig. 5).^{15, 17} Many of these molecules are thought to be involved in Na_x transport and anchoring. However, some functionally interesting binding molecules have also been identified. Notably, Na⁺/K⁺-ATPase (Na/K pump) and triosephosphate isomerase (TPI), which are enzymes involved in key glycolytic reactions, were identified.

Fig. 5.

 Na_x and its binding molecules.

Figure 6 illustrates the model I proposed to explain what happens when Na_x, Na/K pump, and TPI work together.^{15, 18} When dehydration increases the concentration of Na⁺ in the body fluids, the extracellular Na⁺ concentration in the SFO of the brain also increases. Na_x detects this increase and opens it. The Na⁺ ions that flow through Na_x are then expelled by the Na/K pump. The Na/K pump uses energy from ATP hydrolysis to expel three Na⁺ ions from the cell while bringing two K⁺ ions into the cell against their concentration gradients. As long as the extracellular Na⁺ concentration remains high, the Na_x remains open.¹¹ Na⁺ continues to flow in, and the Na/K pump continues to expel it. Glycolysis is activated to replenish the rapidly consumed ATP. Glycolysis breaks down glucose to produce ATP and its byproduct is lactate. Lactate is expelled from cells into the extracellular space. Measurements of lactate secretion from the SFO tissues revealed that lactate secretion increased in response to Na⁺-level elevation, supporting our model. Consistently, lactate secretion did not increase in the SFO of Na_x-deficient mice.

Fig. 6.

 Na_x complex and its function.

Monocarboxylate transporter (MCT) is responsible for transporting lactate across the cell membrane. Glial cells and neurons possess different types of MCTs, and owing to the differences in their affinities, lactate is transported from glial cells to neurons (Fig. 7). Neurons metabolize lactate to produce ATP. As the intracellular ATP concentration increases, the K_ATP channels constitutively open and close. In neurons, K⁺ ions constantly flow out of the cell through K channels, maintaining a membrane potential near the resting membrane potential (approximately -60 mV). When K_ATP channels close in response to ATP, the outflow of K⁺ ions decreases, leading to an increase in membrane potential (depolarization). When the membrane potential is sufficiently depolarized to surpass the threshold for voltage-dependent Na channels, these channels open simultaneously, generating an action potential, a pulse-like electrical signal (also referred to as firing activity).

Fig. 7.

 The mechanism of lactate transport.

In the SFO, glial cells surround the inhibitory neurons.¹⁴ In an experiment using acutely prepared brain slices containing the SFO, electrical activity from the neurons within the SFO was measured.¹⁵ As expected, the activity increased in an Na⁺-concentration-dependent manner. However, no activity was observed in Na_x-deficient mice. Inhibitory neurons suppress the activity of connected neurons when transmitting neural signals. During dehydration, when the inhibitory neurons in the SFO become active, the preference for salt decreases. Therefore, it is hypothesized that the target neurons of these inhibitory neurons are the ones that increase salt preference (salt neurons). The mechanism described above is summarized in Fig. 8.

Fig. 8.

 Signaling during dehydration. In dehydrated animals, (1) when the Na⁺ concentration in extracellular fluid increases, Na_x channels on the glia cell membrane of SFO sense this and open. (2) The threshold of Na_x is regulated by endothelin (ET-3).¹⁸ When Na_x channels open, the intracellular Na⁺ concentration rises. (3) This immediately activates the Na/K pump. The Na/K pump consumes a large amount of ATP to expel Na⁺, and to compensate for this, (4) glucose metabolism in glial cells is activated. (5) As a result, lactate is produced and secreted, which increases the firing frequency of adjacent GABA neurons (inhibitory neurons).

Water neurons and salt neurons

Approximately half a century ago, it was reported that the administration of angiotensin II (Ang II) increased both water and salt intake.¹⁹ We measured the blood AngII levels in mice that had undergone conditions expected to increase water or salt intake, such as “reduced body fluid volume,” “dehydration,” and “salt deficiency,” and found that AngII levels were elevated in all of these conditions.²⁰ Analysis of AT1a-deficient mice (which lack the AngII receptor AT1a) revealed that salt intake almost disappeared and water intake decreased. This raises the question of how the mechanisms of thirst and salt appetite are independently controlled according to the body fluid status.

The following experiments were conducted with Dr. Takashi Matsuda, a graduate student in my group at that time (he recently became an associate professor of Gifu University in 2024).²⁰ Similar to Na_x, AT1a is highly expressed in sCVOs. We deleted AT1a-genes specifically in the SFO or OVLT of mice and investigated their water and saltwater intake. We found that the amount of water consumed decreased regardless of whether AT1a was deleted in the SFO or the OVLT. However, saltwater intake decreased only when AT1a was deleted specifically in SFO. Based on this, we hypothesized that there are two types of neurons in the SFO with different neural connections.

First, we labeled the neurons in the SFO to examine the brain regions to which they projected. We then selected one of these brain regions and created a mouse model with selective AT1a deletion from the cells that project to that region to assess its impact on drinking and salt consumption behaviors. The results of behavioral studies introduced in the next paragraph revealed that the neurons projecting to the ventral bed nucleus of the stria terminals (vBNST) were responsible for salt appetite (“salt neurons”), while those projecting to the OVLT were responsible for water appetite (thirst) (“water neurons”).

We used optogenetics, a method that allows the artificial control of specific neuronal activity using light. First, we generated animals in which the neurons of interest expressed light-sensitive molecules derived from microorganisms. When the targeted brain region was illuminated with light, only the neurons possessing the light sensors became active. The behavioral changes corresponding to this activity can then be interpreted as the physiological functions of these neurons. When water neurons were stimulated with light, the animals selectively drank water, whereas when salt neurons were stimulated, they selectively consumed saltwater (Fig. 9). Through Na_x activation, inhibitory neurons activated during dehydration inhibit the activity of salt neurons. In contrast, in the state of salt deficiency, which suppresses water intake, another inhibitory neuron group, activated by cholecystokinin (CCK), suppresses the activity of water neurons (Fig. 10).

Fig. 9.

 Optogenetics experiment. A mouse with light stimulation of water neurons (left) and a mouse with light stimulation of salt neurons (right). Optical fibers are inserted into their skulls.

Fig. 10.

 Regulatory mechanisms of thirst and salt appetite according to body fluid status. In cases of fluid loss, dehydration, or salt deficiency, the concentration of AngII in circulating blood increases, activating all neurons in the SFO with AT1a receptors. However, in dehydration, lactate signaling mediated by Na_x, and in salt deficiency, CCK secreted in the SFO, activate specific inhibitory neurons (GABA neurons). The thirst sensation and salt appetite are independently regulated.

Why, then, are fundamental physiological functions such as thirst and salt intake so complexly regulated? My hypothesis is as follows. Life originated in the ocean, and later, the atmosphere emerged through photosynthesis by cyanobacteria, forming the ozone layer, which enabled the conquest of land. In the ocean, both water and minerals are abundant. In this environment, the mechanism for both water and salt intake in response to fluid loss, such as hemorrhage (as shown on the left side of Fig. 10), was established by Ang II signaling. Later, when organisms migrated to the land, dehydration and salt deficiency became substantial challenges in inland environments. We believe that the mechanisms shown in the middle and right side of Fig. 10 evolved to regulate the original AngII system that was already in place. Notably, Na_x has not been found in the fish.

Real-time analysis of thirst-regulating neurons

Recently, the miniaturization and performance enhancement of imaging sensors used in cameras in smartphones and other devices have progressed, allowing them to be mounted on the head of a mouse. By combining these with endoscopic lenses, it has become possible to observe the activity of neurons in the brains of freely moving mice in real time. First, we created a mouse expressing GCaMP, a protein that changes its fluorescence intensity in response to neural activity, and implanted a cylindrical lens (gradient index lenses) into its head. We then attached a combined fluorescence illumination system and camera above the lens (Fig. 11).²¹

Fig. 11.

 Activated CCK neurons (white cells in the bottom panel) in the SFO of a drinking mouse (Top). A white cable for the camera is connected to the miniaturized head-mounted microscope on the mouse’s head. For the movie, see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7655816/ Adapted from the supplemental movie of reference ²¹. CC BY 4.0.

Using this method, we observed the activity of individual CCK neurons in the SFO for the first time in the world. We found that neurons could be classified into three groups. The first group consisted of cells that remained continuously active under salt-deficient conditions (water intake suppression), which was the most common. However, the second group exhibited minimal activity under salt-deficient conditions and instead became transiently active immediately after drinking behavior was initiated, triggered by thirst. The third group did not exhibit any characteristic activity patterns.

These neurons were activated almost simultaneously with drinking behavior, but their activity gradually subsided while the mice were drinking. This suggests that these neurons do not control drinking behavior itself. Rather, it is hypothesized that the neural activities, triggered by oral sensations, prevent excessive drinking. Much like when driving a car, fully accelerating with a green light could lead to loss of control and accidents. To effectively control the behavior, the acceleration must be balanced with braking in response. Immediately before drinking began, the sensation of thirst peaked. Drinking water and feeling it in the mouth induces transient preabsorptive satiety, which may lead to the transient suppression of thirst through the activity of CCK neurons.

These results indicate that the SFO is not the center for controlling drinking behavior, but rather the center for regulating thirst. This is further supported by another group’s experiment, in which neural activity in the SFO of dehydrated mice was suppressed when a cold metal rod was inserted into their mouths.⁷ Additionally, we discovered that adipsic hypernatremia (a condition in which blood sodium and osmolarity levels are abnormally elevated, but thirst sensation does not occur) is due to autoimmune inflammation of the SFO, leading to its dysfunction.^{22, 23} This suggests that the SFO is likely the center of the thirst sensation in humans as well.

SFO contains several cell types. As mentioned earlier, even CCK neurons, which were originally thought to be of one type, were divided into three groups with distinct activity patterns. When observing collective activity, one may overlook the fact that these neurons are mixed together, which could lead to incorrect conclusions. The technique of analyzing individual neuron activity in the SFO, as shown in Fig. 11, is expected to become the standard and most effective method for precisely quantifying “thirst sensation.”

The nature of osmoreceptors involved in thirst

In the research introduced in this article, we analyzed thirst in response to Ang II treatment. AngII is produced downstream of renin secretion from the kidneys, therefore, it can be considered as a drinking behavior directed by signals from the kidneys. However, in the case of dehydration-induced thirst, there was little effect, even when AT1a was deleted.²⁰ During dehydration, a redundant system involving sensors other than Ang II maybe involved. Among them, osmoreceptors are the most promising candidates. Over half a century ago, experiments were conducted in which hypertonic saline or sugar solutions were injected into the ventricles of large animals, \such as goats, and drinking behavior was observed. The presence of osmoreceptors was inferred from the experimental results.

The first osmoreceptor in the brain to be reported was the ion channel, TRPV1.^24,25 Neurons in Trpv1 knockout mice lose osmoreceptive sensitivity. However, while TRPV1 is known to respond to capsaicin, an active component of chili peppers, and to high temperatures, its response to osmolality has not been confirmed. I speculated that TRPV1 may become more easily activated at body temperature, enabling it to respond to slight changes in osmolality. The analysis was conducted with Dr. Eri Nishihara, a postdoc researcher in my group at that time.²⁶ As expected, TRPV1 could function as an osmoreceptor near body temperature. However, subsequent analyses by Dr. Hraki Sakuta, an assistant professor of National Institute for Basic Biology, revealed that drinking behavior in Trpv1-knockout mice showed no abnormalities, and there is no evidence that TRPV1 was involved in thirst.²⁷

Another ion channel, TRPV4, has been reported to be a sensor that detects a decrease in osmolality rather than an increase.²⁸ However, there have also been reports that drinking behavior in TRPV4 knockout mice is reduced with an increase in osmolality, leading to conflicting conclusions.²⁹ We hypothesized that the discrepancy in these experimental results might be due to the systemic stimulation caused by dehydration, which could obscure the function of the brain’s osmoreceptors owing to the influence of various hormones and sensory inputs from the body. To test this hypothesis, Dr. Sakuta conducted experiments in which hypertonic solutions were injected into the brain ventricles and the resulting drinking behavior was measured.²⁷ Drinking amount was reduced in Trpv4-knockout mice when hypertonic saline was administered. However, no difference was observed when sugars were used alone to change osmolality. Moreover, when hypertonic saline was administered to Na_x-knockout mice, their drinking behavior was reduced. A detailed analysis revealed that activation of Na_x caused the production of epoxyeicosatrienoic acids (EETs) in glial cells in the OVLT, which in turn activated TRPV4 in neurons, inducing drinking behavior. Thus, TRPV4 appears to function not as a sensor, but as a downstream molecule in the mechanism by which Na_x, as a Na⁺ sensor, triggers thirst. This suggests that TRPV4 is involved in a mechanism of Na⁺-dependent thirst rather than osmoregulatory thirst. In the OVLT, Na_x is also involved in the Na-dependent blood pressure control (This study was mainly conducted by Dr. Kengo Nomura, a post doc researcher of my group at that time; he is now an assisitant professor in Kyoto Prefectural University of Medicine).³⁰

Furthermore, we discovered another Na⁺ sensor that is involved in thirst. This research was conducted with Dr. Sakuta and Chia-Hao Lin, a post doc researcher, in my group at that time.³¹ We conducted a comprehensive expression analysis to search for new sensory molecules related to thirst, targeting molecules specifically expressed in the SFO and OVLT. Among the molecules specifically expressed in these regions, we focused on membrane proteins, conducted experiments to express them in cultured cells, and investigated their properties. These results revealed that the Na⁺/H⁺ exchanger Slc9a4 could function as a sensor. Slc9a4 was expressed in OVLT, and we confirmed that Slc9A4-expressing cells were activated during dehydration. Subsequently, we conducted experiments in which hypertonic saline was injected into the brain ventricles of mice with reduced Slc9a4 expression in the OVLT and found that their drinking behavior was reduced compared to that in normal mice. In contrast, no differences were observed when sugars were administered.

Based on our analyses, we determined that Na_x and Slc9a4 are the Na⁺ sensors involved in thirst. However, the true molecular entities of osmoreceptors remain unclear, and further analysis is needed in the future.