NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Lautt WW. Hepatic Circulation: Physiology and Pathophysiology. San Rafael (CA): Morgan & Claypool Life Sciences; 2009.

Cover of Hepatic Circulation

Hepatic Circulation: Physiology and Pathophysiology.

Show details

Chapter 5Resistance in the Hepatic Artery


The liver does not control portal blood flow, which is simply the outflow of the extrahepatic splanchnic organs. If the vascular resistance to portal flow is increased to a maximum level by, for example, stimulation of the hepatic sympathetic nerves, portal pressure rises but portal blood flow does not fall. Portal blood flow is determined by the net outflows of the splanchnic organs including the stomach, spleen, pancreas, intestines, and omentum. If portal blood flow changes, the hepatic arterial flow changes in the opposite direction, thus tending to maintain total hepatic blood flow constant. This mechanism does not completely compensate but rather buffers the effect of portal blood flow changes on total hepatic flow, thus accounting for the name I chose for this response, the hepatic arterial buffer response (HABR) [192].

The HABR is proposed to operate by the following mechanism (Figure 5.1). Adenosine is constantly secreted into the space of Mall, which is a very small isolated fluid compartment through which passes the portal triad, consisting of the fine terminal branches of the hepatic artery, portal vein, and bile ductule, all in intimate contact. The concentration of adenosine, a potent vasodilator, is regulated by the rate of washout from the space of Mall into the blood vessels. When portal blood flow decreases, less adenosine is washed away, and the elevated adenosine concentration leads to dilation of the hepatic artery. This is the suggested mechanism of the HABR.

FIGURE 5.1. Adenosine washout hypothesis for the integrated intrinsic flow regulators of the hepatic artery: the HABR and autoregulation.


Adenosine washout hypothesis for the integrated intrinsic flow regulators of the hepatic artery: the HABR and autoregulation. The hepatic acinus represents a cluster of parenchymal cells arranged like a 2-mm berry on a vascular stalk, the portal triad, (more...)

A second form of intrinsic hepatic arterial regulation is referred to as arterial autoregulation, which was previously believed to be myogenic in nature. If arterial pressure is decreased, this leads to a reduction in arterial blood flow, which washes away less adenosine from the space of Mall. The accumulated adenosine leads to arterial dilation. The HABR and autoregulation work through the same mechanism and operate simultaneously, thus leading to potential complex interactions. In addition, extrinsic factors such as circulating hormones, autonomic nerves, and various nutrients can interact to affect the hepatic circulation in a number of complex ways.

5.1.1. Metabolism and Hepatic Blood Flow

The first clue that hepatic metabolic demands did not control hepatic arterial flow came from a serendipitous observation made at the end of experiments designed for other purposes. In these experiments, the hepatic circulation was studied using the hepatic venous long-circuit protocol, draining blood from the liver into a reservoir, from which it was pumped back to the animal through the jugular veins (Figure 4.2, Chapter 4). The lower vena cava (below the hepatic veins) drained retrograde via two femoral venous catheters into the same warmed reservoir. In this way, we could measure total hepatic blood flow, control hepatic venous pressure, and measure liver metabolism in a preparation that had intact nerves and inflow vessels [184]. At the end of experiments, which were designed to study metabolic effects of the hepatic nerves, the blood reservoir would become depleted at a time when the biological preparation was still functional. At the point where the reservoir was almost empty, we added a solution of Dextran 75 and Ringer’s solution to produce an isovolemic hemodilution. We anticipated that the oxygen delivery to the liver would decrease and that the hepatic artery would subsequently dilate. Although the oxygen delivery decreased to 68% of control and hepatic venous and portal oxygen content declined, the hepatic artery did not show a dilation. In fact, it showed a small constriction, and total hepatic blood flow remained constant. The results were completely unexpected and quite disconcerting, and our first fear was that the surgical preparation we had used had produced an abnormal vasculature that was incapable of responding to a decreased oxygen supply. We quickly proved that the artery was capable of vasodilation by demonstrating a brisk response to isoproterenol. These procedures were repeated at the end of every experiment in several series so that we eventually had 23 animals that had been subjected to a large isovolemic hemodilution with full hepatic arterial, portal venous, hepatic venous, and systemic hemodynamic and blood gas data monitored. Before the availability of personal computers, we attempted to determine factors that correlated with changes in hepatic arterial blood flow, and after 150 hand-graphed correlations, we determined that the only unexplained correlations were with portal venous flow and hepatic arterial conductance. It appeared that in situations where the portal flow increased, hepatic arterial conductance decreased, and when portal flow decreased, hepatic arterial conductance increased. The end result was that changes in portal flow tended to produce opposite changes in hepatic arterial flow such that total hepatic blood flow was maintained at a remarkably constant level. This work had two surprising conclusions. The first was that the oxygen supply-to-demand ratio for the liver did not appear to be an appropriate stimulus to change hepatic arterial blood flow. The second was that the major factor that appeared to be controlling hepatic arterial blood flow was portal venous flow [185].

We continued examination of this first observation using dinitrophenol to increase oxygen demand or SKF525A to inhibit metabolism. The hepatic artery showed no tendency to change in response to the metabolic alterations; the hepatic oxygen demands were maintained solely by altered hepatic extraction [191]. Scattered reports had been consistent with this conclusion. As early as 1950, Myers et al. [285] reported that an increased splanchnic metabolism coexistent with a normal splanchnic blood flow provided an exception to the hypothesis that the rate of local tissue metabolism regulates the blood flow through the liver. Chronic alcohol exposure in rats, for example, led to elevated oxygen demand by the liver, but the hepatic artery actually constricted [33]. Lactic acidosis in dogs resulted in elevated portal flow but reduced oxygen delivery to the liver and reduced hepatic venous oxygen content. The hepatic artery did not respond to the decreased oxygen supply, but rather it constricted in response to the elevated portal flow [139].

Studies relating enzyme induction to blood flow and the effect of bile salts had suggested hepatic metabolic regulation of blood flow. Ohnhaus et al. [293] reported that enzyme induction with phenobarbitone increased hepatic blood flow in the rat, an observation that was generally taken to support the notion that blood flow in the liver was metabolically controlled. However, Nies et al. [287] reported that, in rats, hepatic enzyme induction with 3,4-benzpyrene and 3-methylcholanthrene caused no hemodynamic changes. In contrast, phenobarbital induction did cause elevated hepatic blood flow, but the augmented flow was via the portal vein, indicating that the increased hepatic flow was a result of an extrahepatic phenomenon unique to phenobarbital and unrelated to induction of hepatic enzymes.

In examining the effect of stimuli on metabolic activity and vascular responses, it is important to differentiate direct from indirect effects. One response that initially appeared to contradict the absence of hepatic metabolic control of arterial flow was the observation that bile salts both stimulate hepatic metabolism and cause hepatic arterial dilation. The vascular and metabolic responses of the liver to bile salts are, however, independent and occur at different doses [209]. Low doses of taurocholate (1 µM/min per kilogram of body weight) infused into the portal vein produced elevated bile flow but no arterial vasodilation, whereas higher doses produced dose-related vasodilator responses in both the superior mesenteric and hepatic arteries. The independence of vascular and metabolic effects was more clearly shown when the same dose of taurocholate infused into the hepatic artery or portal vein produced equal effects on bile flow but considerably greater arterial responses when infused into the artery. The bile salts thus produce direct arterial vaodilation unrelated to hepatic metabolic stimulation. The hepatic artery is not subservient to hepatic metabolic activity [195].

5.1.2. Portal Flow Regulation of Hepatic Arterial Flow

The earliest studies reporting an effect of changes in portal perfusion on hepatic arterial flow were credited by Child (1954) to Betz in 1863 and Gad in 1873 [87]. The observation was studied periodically by very few people over the next 70 years. Various hypotheses were proposed to account for this interaction, and it was generally concluded that a myogenic mechanism accounted for what became referred to as the reciprocal flow relationship of the hepatic artery and portal vein. This was the state of affairs in the mid-1970s and was matched with the strong statement in textbooks (based on no experimental evidence) that the hepatic artery was under metabolic control of the liver. With the discovery that the hepatic artery was not under metabolic control, we focused on determining the mechanism of vascular interactions. Very quickly we determined that there was no reciprocal relationship between the blood flows and that changes in hepatic arterial flow did not change portal venous flow or portal resistance.

The decrease in portal pressure that may be seen with complete occlusion of hepatic arterial inflow is not caused by a decreased vascular resistance of the portal vessels but rather is caused by the fact that the intrahepatic and portal pressure is largely maintained as a result of blood flowing through postsinusoidal resistance vessels located in the small hepatic veins, so that a decrease in total flow subsequent to arterial occlusion results in a reduced pressure.

Because the relationship between the portal vein and hepatic artery was clearly not a reciprocal flow relationship, we coined the term hepatic arterial buffer response (HABR) [192] to acknowledge the role of the hepatic artery in buffering changes in total hepatic blood flow that would occur secondary to changes in portal flow. The primary function of the hepatic artery was recognized as being maintenance of hepatic blood flow, per se, regardless of oxygen supply or demand.

We considered a number of alternate hypotheses to account for the mechanism. Mechanisms we considered and rejected have been reviewed [192]. They included a myogenic mechanism, which proposed that a change in portal flow would produce a change in portal pressure that would be sensed by the hepatic artery. Changes in portal pressure are very minor even in the face of very large changes in portal flow. We also considered a neural mechanism, but we and others have shown that the HABR is seen in a fully denervated liver [257,331] and in transplanted human livers [131,299].

We considered the possibility of purely physical interactions that would result from the mechanical consequences of interposing a slower flowing stream (the portal blood) into the path of a faster flowing stream (the hepatic artery). The absence of a HABR in the isolated liver was the principal evidence to preclude this purely physical model. The physical model also predicted that occlusion of the hepatic artery should result in a reduction in portal conductance, which was not seen.

We also considered that the quality of the portal blood was an important factor. If portal blood contained a constrictor agent, an elevation in portal flow might deliver more constrictors to the hepatic vasculature and result in arterial constriction. We eliminated this possibility by devising a preparation that had a dual vascular shunt such that the portal outflow from the guts could be sent either to the liver or to the vena cava. A similar shunt from the vena cava allowed us to perfuse the liver with vena caval blood. Switching from portal blood to vena caval blood perfusing the liver did not result in any change in the hepatic arterial conductance and the buffer capacity was similar regardless of the source of portal flow.

We considered metabolic control, whereby metabolic by-products or oxygen content were a factor. Our earlier studies had indicated that oxygen was not a regulatory factor, and we only later realized that the vascular anatomy of the hepatic acinus precluded metabolic feedback to the hepatic arterial resistance vessels.

We were finally left with the remaining hypothesis, which we referred to as the washout hypothesis. I proposed that a dilator substance is produced within the space of Mall at a constant rate and that the concentration is dependent on the rate of washout into the portal blood (Figure 5.1). According to this theory, a decrease in portal flow would wash away less of the dilator, thus leading to an accumulation and arterial vasodilation. The challenge then was to identify the dilator. In considering possible dilators, we formulated a list of more than 20 candidates. The frustration felt at this daunting list was exemplified by a piece of doggerel that I wrote at that time [192]:

Lautt’s Lament

I think that I shall never know

the factor controlling hepatic blood flow.

In spite of our tests and our pontification,

we still don’t know what causes dilation

of the hepatic artery, which stands at the ready

keeping total blood flow to the liver quite steady.

When the portal blood flow is quickly reduced.

the artery dilates, or so we’ve deduced, .

as result of accumulated dilator.

that relaxes smooth muscle and makes the flow greater. .

But what is this magical dilator stuff ? .

We don’t know yet for we’ve not done enough. .

For a compound to be eligible as the dilator substance controlling the HABR via a washout mechanism, several criteria must be met. These include, but are not restricted to, the following: (i) the putative regulator must dilate the hepatic artery; (ii) portal blood must have access to the arterial resistance vessels so that portal flow can wash away the substance from the area of the resistance vessels; (iii) potentiators of effects of the putative regulator should also potentiate the buffer response; (iv) blockers of effects of the putative regulator should also inhibit the buffer response.

In a preliminary screening approach, based on only one to three tests per compound, the buffer response was unaltered by prior administration of atropine, propranolol, ouabain, aminophylline, theophylline, indomethacin, metiamide, and mepyramine. At that time I was fortunate to have John W. Phillis as my Department Head. He was involved with studies related to the functions of adenosine in the central nervous system and had access to some new pharmacological tools. His suggestion that adenosine was the “magical dilator stuff” initially was received with skepticism because of our previous observation that the hepatic artery was not affected by the metabolic status of the liver. At that time it was generally regarded that adenosine served as a major link between tissue metabolism and blood flow. Furthermore, aminophylline and theophylline had not blocked the HABR. The first compounds tested were isobutylmethylxanthine, as an adenosine receptor antagonist, and dipyridamole, as an adenosine uptake blocker. Serendipity again played a hand, and we showed significant blockade with isobutylmethylxanthine and potentiation with dipyridamole. Later studies indicated that these compounds had many other actions and that we were fortunate indeed to have been able to demonstrate the appropriate effects with them [220]. We were sufficiently cautious to delay publication until we had confirmed the data, using more selective compounds. Nevertheless, those experiments allowed us to fulfill the important criteria discussed above. Adenosine was shown to produce a powerful dilation of the hepatic artery. Portal flow was shown to have access to the arterial resistance vessels by the observation that intraportal adenosine produced a profound dilation of the hepatic artery. Dipyridamole potentiated the effect of infused adenosine and potentiated the buffer response. Isobutylmethylxanthine blocked the HABR and blocked the response to exogenous adenosine.

To study the buffer response, we had to develop a surgical protocol that allowed repeated buffer responses to be quantified. A standard preparation was used to do mechanistic studies of the buffer response (Figure 8.5, Chapter 8). Using this system, all of the portal blood flow is supplied by the superior mesenteric artery, which allows measurement and control of portal flow at the arterial inflow side. Hepatic arterial and superior mesenteric arterial blood flow are measured using flow probes. An occlusion of the portal inflow results in a rapid dilation of the hepatic artery (Figure 5.2, left panel). At that point it is important to control hepatic arterial blood pressure at a constant level so that any change in arterial flow is secondary only to the buffer response and not to changes in systemic arterial pressure.

FIGURE 5.2. The response to adenosine and the HABR and blockade by adenosine receptor antagonist.


The response to adenosine and the HABR and blockade by adenosine receptor antagonist. Stepwise dose–response relationship for intraportal infusion of adenosine followed by the HABR elicited by complete occlusion of portal inflow vessels. Note (more...)

The availability of good pharmacological tools is often a primary limitation in physiological studies. With the development of 8-phenyltheophylline, another gift from J.W. Phillis, we had access to a much improved adenosine receptor antagonist. This compound is certainly not perfect because it has to be administered in a very alkaline solution and is poorly soluble. Nevertheless, it has proven to be a useful tool (Figure 5.3) that quantified dose-related competitive inhibition of exogenous adenosine and parallel inhibition of the buffer response [216]. The inhibition of the response to adenosine and the buffer response are shown from one cat (Figure 5.2).

FIGURE 5.3. The progressive competitive antagonism of the hepatic arterial vasodilator effect of complete reduction of portal blood flow (the hepatic arterial buffer response) and the dilator effect of intraportal adenosine in the presence of progressive increases in intra-arterial doses of adenosine receptor antagonist (8-phenyltheophylline).


The progressive competitive antagonism of the hepatic arterial vasodilator effect of complete reduction of portal blood flow (the hepatic arterial buffer response) and the dilator effect of intraportal adenosine in the presence of progressive increases (more...)

5.1.3. Autoregulation

It occurred to us that if adenosine was being washed away from the space of Mall by portal blood flow, it should equally be able to be washed away by arterial flow. We thus suggested that what was previously considered to be a myogenic mechanism might, in fact, be mediated by the adenosine washout mechanism. According to this hypothesis, an increase in arterial pressure would lead to an increase in arterial flow and a washout of the dilator. The resultant decrease in basal level of adenosine would lead to constriction of the artery. Conversely, a decrease in arterial perfusion pressure would lead to reduced flow and reduced washout of adenosine and subsequent arterial dilation. This hypothesis was confirmed, and we demonstrated that the buffer response and autoregulation were blocked in parallel with the response to exogenous adenosine by adenosine receptor antagonists [77].

The observation that autoregulation could be demonstrated to be consistent with the mechanism of the buffer response was crucial to our own acceptance of the adenosine washout hypothesis.

5.1.4. Quantitative Aspects of the HABR

If portal blood flow is severely reduced, the buffer response results in the hepatic artery dilating maximally as demonstrated by the inability to produce additional dilation in response to intra-arterial infusion of adenosine. Conversely, when portal flow is doubled, the hepatic artery constricts to a maximal extent, as demonstrated by the inability of intra-arterial norepinephrine to produce further constriction [222]. Thus, the HABR is sufficiently powerful to regulate the vascular tone in the hepatic artery over the full range from maximal vasodilation to maximal vasoconstriction.

To quantitate certain aspects of the HABR, a number of special techniques had to be developed. As previously discussed, it is important to be able to produce a known change in portal flow and to measure the change in hepatic arterial flow in the absence of a change in hepatic arterial perfusion pressure. We developed a method [220] to quantify the HABR for subsequent mechanistic studies (Figure 8.5, Chapter 8). Using this method, the entire portal blood flow is supplied through the superior mesenteric artery. Anastomotic connections to other arteries that were ligated (inferior mesenteric artery, gastric artery) provide adequate blood flow to those areas. The spleen is removed. This, unfortunately, leads to a reduced portal flow at the outset of the experiment, which results in a partially activated buffer response. This complication leads, in many experiments, to attempting to study the buffer response in an artery that is already nearly maximally dilated. Although such a method was necessary to carry out studies related to the mechanism of the HABR, the impact that this method has on the estimation of buffer capacity is unknown.

Buffer capacity is calculated as the change in hepatic arterial flow expressed as a percentage of the change in portal venous flow, where a 100% capacity would provide for full compensation. In anesthetized, splenectomized cats, the buffer capacity is only approximately 25% [220]. Others found a buffer capacity in the same range in anesthetized dogs [255] and pigs. Many studies report changes in portal and hepatic arterial flow but, without a steady hepatic arterial pressure, it is not appropriate to assume that calculated buffer capacity actually represents a true buffer response. This is well demonstrated by the following example (Figure 5.2). If perfusion pressure is not held constant, occlusion of the superior mesenteric artery produces a decrease in portal flow but an elevation of arterial pressure. The HABR tends to raise arterial flow, but autoregulation tends to counter this rise. After adenosine receptor blockade, both intrinsic mechanisms are blocked, so that the decrease in portal flow does not activate the HABR, but the rise in arterial pressure now causes a rise in hepatic arterial flow unimpeded by autoregulation [216]. Thus, without controlling for hepatic arterial perfusion pressure, one would conclude that adenosine receptor blockade had not affected the buffer response. With this precaution in mind, there are a number of publications that report compensatory changes in hepatic arterial flow that suggest a buffer capacity close to 100%. In our early studies, hemodilution [187] and metabolic inhibitors and stimulators [191] resulted in altered portal flow, with the hepatic arterial flow compensating to hold total hepatic blood flow within 4% of basal flow. In a cirrhotic rat model, without significant portacaval shunts, portal blood flow was significantly elevated and the hepatic arterial flow was reduced sufficiently to hold total hepatic blood flow at similar levels in the control and cirrhotic animals [79]. From the mean flow data, the buffer capacity was 72%. The constriction of the hepatic artery was unique in that all other arteries showed vasodilation, implying that the buffer overcame some general systemic dilator influence and yet produced an impressive compensation. Lactic acidosis in dogs [139] resulted in an increase in portal blood flow and a decrease in hepatic arterial flow resulting in a 100% compensation.

To carry out in vivo quantitative pharmacodynamic studies, it was also necessary to develop certain techniques. A key point is that changes in arterial vascular tone normally result in changes in flow, with minor changes occurring in systemic arterial pressure. In this situation it is imperative that changes in vascular tone be expressed as vascular conductance rather than resistance. Conductance is simply the inverse of resistance, and it is essential that the parameter that changes to the greater extent be in the numerator, that is, flow divided by pressure gradient rather than pressure gradient divided by flow [198]. One dramatic example will suffice for this chapter. Consider if a severe vasoconstriction is produced by sympathetic nerve stimulation or infusion of norepinephrine, with flow decreasing to approach zero. Conductance will also approach zero, but resistance will approach infinity. Clearly, resistance cannot be used to estimate an ED50 for drug effect. Resistance cannot be used even to calculate a simple arithmetic average [198] because resistance is nonlinearly related to flow. With the use of vascular conductance and nonlinear regression analysis, maximum responses (Rmax) and the dose of drug that produces 50% of maximal vasoconstriction (ED50), or the frequency of nerve stimulation that produces 50% of the maximal response (Hz50), can be calculated. These parameters are important for determining the effects of pharmacological antagonists and modulators. Using this approach we were able to demonstrate that 8-phenyltheophylline is capable of blocking 100% of the buffer response. Furthermore, this approach allowed a clear demonstration that an exogenously administered receptor antagonist is capable of blocking exogenously administered agonists more effectively than the endogenous agonists. The ID50 for 8-phenyltheophylline to produce a 50% blockade of the effect of exogenous adenosine was 0.33 ± 0.03 mg/kg and the ID50 for blockade of the buffer was 1.31 ± 0.47 mg/kg. These data were recalculated from the data shown in Figure 5.3. Lack of recognition of the differential ability to block endogenous versus exogenous agonists can lead to serious error. For example, Mathie and Alexander [254] used one dose of 8-phenyltheophylline and demonstrated that that dose blocked the response to exogenous adenosine but did not completely block the buffer response. They concluded that factors other than adenosine must, therefore, be involved. Without demonstrating that a higher dose of blocker also left the buffer response incompletely blocked, this conclusion is inappropriate. It can be seen from Figure 5.3 that some doses of blocker could eliminate the response to adenosine but still leave a sizeable HABR.

Two other precautions must be considered that apply to in vivo pharmacology in general and to the use of adenosine receptor antagonists in particular. The first is that, because of the high variability in blocking effectiveness, we found it necessary to test for effective blockade in every animal and to double the blocking dose until full blockade of the response to infused adenosine is produced. The usual dose of 8-phenyltheophylline administered is 8 mg/kg. The second precaution relates to selectivity. As important as it is to show effective blockade, it must also be selective. That is, dilation to other agents, such as isoproterenol, should not be impaired.

Although the HABR has been shown in a wide variety of species, it has not been quantitated in terms of buffer capacity in conscious animals or humans. Clearly, to fully appreciate the homeostatic role of the HABR, it is essential to know the buffer capacity of the hepatic artery.

5.1.5. Roles of the HABR

If the hepatic arterial flow is not regulated by hepatic oxygen supply or demand, does this imply a serious pathophysiological problem for the liver with regard to hypoxic damage? The liver has often been referred to as an organ verging on hypoxia. This, in fact, is not the case. When oxygen delivery was reduced to the liver to 68% in our hemodilution studies, oxygen uptake was able to be maintained at completely normal levels simply by increasing the oxygen extraction from the available supply. Because of the unique microvascular anatomy, the liver is able to extremely efficiently extract compounds from the blood.

In many conditions, the HABR, despite being insensitive to reduced hepatic oxygen, serves to protect oxygen delivery coincidentally. For example, the hepatic arterial flow is preserved during hemorrhage. This relative sparing of the hepatic artery is secondary to the HABR responding to the decrease in portal blood flow [227]. Thus, the buffer response in this instance serves to preserve the oxygen supply of the liver without the oxygen supply being directly regulated.

In considering the teleological purpose for evolving a flow-regulatory system that tends to maintain blood flow, per se, constant to the liver, it was fortunate that we had recently concluded studies relating blood flow to hepatic drug clearance. The liver is involved with clearance of an extremely wide range of endogenous compounds including hormones, such as aldosterone and corticosterone [186,265], in a blood flow-dependent manner. It was proposed [186] that a major function of the liver might be considered to be an endocrine function according to the following logic. Traditionally, when one thinks of regulation of plasma hormone levels, only the endocrine gland that produces the hormone is considered. However, for the endocrine gland to be able to increase or decrease hormone levels in the blood, it is important that there be a reasonably rapid and quite constant clearance of the hormone to serve as a background against which secretion can lead to fine tuning. If hepatic blood flow was not prevented from rapid, transient changes secondary to similar changes in the portal venous flow, endocrine homeostasis would be imperiled. It thus seems highly likely that general endocrine and metabolic homeostasis is subserved by the function of the HABR.

A second consequence of a functional buffer response is related to the impact of altered hepatic blood volume on cardiovascular status. Hepatic blood volume is passively altered in response to changes in total hepatic blood flow [204]. The decrease in volume per unit decrease in blood flow is similar regardless of whether the reduction was in portal or arterial flow. Most resistance to portal blood flow through the liver is at the hepatic venous outflow [215]. Intrahepatic and portal pressures are altered by changes in blood flow through this resistance site. Changes in intrahepatic pressure lead to changes in hepatic stressed blood volume. Stressed volume is the product of compliance and intrahepatic pressure. The liver is extremely compliant and changes in intrahepatic pressure are also reflected back to the portal vein, which is also a very complaint system. The splanchnic venous system can account for two thirds of the blood volume response to hemorrhage and represents the largest blood volume reservoir in the body [112]. By tending to maintain hepatic blood flow constant and hence portal and intrahepatic pressures constant, transient alterations in hepatic blood volume and, therefore, in venous return are minimized by the HABR.

The HABR is capable of being activated at the acinar level and thereby plays a role in maintaining perfusion homogeneity and prevention of regional stagnation secondary to small local perturbations in regional pressure such as caused by external mechanical forces or shifts of abdominal contents. Maintenance of homogeneity at sinusoidal levels, smaller than the regions served by one terminal hepatic arteriole branch, are probably maintained through stellate cell activity (Chapter 2).

A further major homeostatic role for the adenosine washout mechanism is related to adenosine acting as a neurotransmitter. Reductions in portal flow, leading to elevated adenosine concentrations in the space of Mall, result in activation of hepatic sensory nerves, which leads to a reflex fluid retention and thus increased venous return. This mechanism is important for both physiological homeostasis and the pathology of salt and water retention in chronic liver disease (see Chapter 13).

5.1.6. Clinical Relevance

The existence of the HABR is reported to be the best prognostic indicator of a successful outcome for patient survival of a portacaval shunt to treat portal hypertension [37]. Similarly, it has been found that a brisk buffer response to a brief occlusion of portal flow is seen in patients after transplant [131] and is a useful indicator of suitable vascular reconnection.

That the liver cannot regulate its blood flow in accordance with the metabolic activity of the parenchymal cells is not disadvantageous to the liver under normal physiological conditions because of the excess oxygen delivered and the hepatic capacity to increase oxygen extraction. However, a relatively hypoxic liver shows increased toxic effects of alcohol [145], carbon tetrachloride [342], and halothane [262]. The formation of active toxic metabolites may occur to a greater extent and the ability of glutathione to detoxify such metabolites may be reduced when the NAD+/NADH ratio is altered by hypoxia (Chapter 12). Although the HABR is seen in cirrhotic livers, the buffer capacity may be insufficient to maintain a normal oxygen supply [332], and there is suggestion that severely diseased livers lose the HABR [329].

The HABR shows considerable variability in diseased livers. The HABR is fully maintained in transplanted human livers [25,131] and appears to be maintained in liver diseases of considerable severity [11,127,152,282,316,330]. Portal flow remains high for as long as 2 years after liver transplantation, mainly as a result of elevated splenic blood flow, and is associated with reduced hepatic arterial flow [25].

If an HABR response to brief portal occlusion does not occur before establishment of a portacaval shunt to reduce portal hypertension, those patients will show the greatest reduction in portal pressure [389]. Unfortunately, those are the same patients with the poorest prognosis for survival [37]. The demonstration of an intact HABR, by observing an elevated portal and decreased arterial flow after a balanced liquid meal [64], has been suggested as a tool to assess the severity of liver disease because this response is decreased in very severely diseased livers [179].

5.1.7. Unresolved Issues

A number of aspects of the adenosine washout hypothesis have not been resolved or have not been tested. Although the anatomy is consistent with the space of Mall serving as the isolated fluid compartment into which adenosine is secreted to affect the hepatic artery and from which adenosine can be washed away by portal or arterial blood supply, no direct evidence exists to localize the anatomical site.

Another unresolved aspect of this hypothesis is the pathway of adenosine production. Hepatic arterial regulation is independent of liver metabolism. Adenosine production can occur by breakdown of adenine nucleotides or cyclic AMP; however, these sources are directly linked to the energy status of the cells. The liver generates adenosine from these sources, but the cells of production are parenchymal cells that are downstream from the hepatic arterial inflow resistance vessels. The unique microvascular anatomy of the hepatic acinus precludes metabolites released from the parenchymal cells from diffusing upstream to directly affect the hepatic artery. I suggest that the adenosine involved with the HABR and autoregulation is produced at a constant rate and secreted into the space of Mall and is most likely derived from demethylation of S-adenosylhomocysteine, a reaction that is oxygen independent and is proposed to account for basal adenosine production in the heart [243].

It is also unclear whether the rate of adenosine secretion can be modulated by some metabolic or hormonal responses that can lead to altered baseline secretion, thereby modulating the magnitude of the buffer response. In at least one third of animals tested under anesthesia and using the intensive surgical preparation required to study the mechanism of the HABR, the intrinsic hepatic arterial regulation and the response to exogenous adenosine are absent. We have eliminated several possible mechanisms for this interference with intrinsic regulation but currently cannot account for it. In a series of unpublished studies we demonstrated, to our satisfaction, that increasing the basal tone of the hepatic artery using infusion of vasoconstrictors does not return the HABR in animals where it is absent. Furthermore, we have not been able to link the absence to alterations in blood gases, pH, or lactic acid levels. Nitric oxide is also not involved with the HABR or autoregulation, in contrast to the superior mesenteric artery in which nitric oxide is shown to antagonize autoregulation [249]. It was our impression that in cats that had intestinal worms, or that had recently been treated for intestinal worms, the HABR is weak or absent. Furthermore, we have witnessed the HABR disappearing during an experimental protocol only to recover to normal levels after several hours. These observations suggest that some quite powerful modulation of the HABR is possible, but the mechanism remains completely unknown (see discussion on hydrogen sulfide below).

Although we have shown that adenosine causes postjunctional inhibition of nerve-, norepinephrine-, angiotensin-, and vasopressin-induced constriction of the hepatic artery (but not portal vein), this effect likely occurs only at pharmacological doses of adenosine, with no evidence that such inhibition occurs under physiological conditions [217].

Another puzzling aspect of this work is that the competitive adenosine receptor antagonist, 8-phenyltheophylline, results in dose-related suppression of autoregulation, the HABR, and the response to exogenous adenosine but does not alter the basal tone of the hepatic artery until the dose of antagonist reaches a concentration that produces full blockade, whereupon higher doses result in massive and prolonged constriction [216]. A similar effect is seen in the superior mesenteric artery [197]. Because our hypothesis involves hepatic arterial tone being affected by increases or decreases in adenosine concentration, the absence of effect of small doses of adenosine receptor antagonists on basal tone is without explanation.


Although the HABR is the primary regulator of hepatic arterial blood flow, HA flow is also affected by a number of extrinsic factors, drugs, hormones, and nerves.

5.2.1. Caffeine

Considering that caffeine is the drug with largest worldwide consumption and that it is reported to block adenosine receptors, we tested the possibility that caffeine was capable of modifying the buffer response [199]. We found that at doses compatible with extremely heavy consumption, caffeine was without effect on hepatic or superior mesenteric flow or vascular conductance. Caffeine did not antagonize the buffer response even at doses that produced cardiac arrhythmias, whereas it did produce a noncompetitive antagonism of the vasodilation induced by exogenous adenosine. Interestingly, the maximum effect of caffeine was to suppress 60% of the vasodilation induced by adenosine with a remaining 40% being unsuppressable at any dose. This suggested that adenosine caused dilation by two different mechanisms. One, perhaps cyclic AMP dependent, was not affected by caffeine. Caffeine was without effect on isoproterenol-induced vasodilation. The portion of dilation blocked by caffeine may be related to a calcium channel or cyclic GMP effect, but there are no data to directly support this speculation. We concluded from this study that caffeine was unlikely to affect endogenous hepatic or splanchnic blood flow or intrinsic regulatory parameters in response to acute exposure. Effects of chronic exposure have not been evaluated. Adenosine produces dilation through action on A2 receptors; caffeine is quite selective for A1 receptors (see Chapter 13).

5.2.2. Vasodilators

Isoproterenol (β2 receptor), adenosine (A2 receptor), and glucagon are all dilators of the hepatic artery when the drugs are administered directly to the artery. However, if administration is intravenous, direct actions on the splanchnic arteries will lead to increased portal blood flow that, in turn, will activate the buffer response and can actually result in a decrease in flow, depending on the dose [208]. Blocking adenosine receptors in the liver leads to blockade of autoregulation, the buffer response, and the dilator response to adenosine but does not alter the vasodilator responses to other compounds such as isoproterenol.

The hepatic arterial resistance vessels constrict in response to activation of hepatic sympathetic nerves, administration of norepinephrine, vasopressin, and angiotensin. Adenosine is not only a direct-acting vasodilator but it also results in dose-related ability to completely inhibit the vasoconstrictions induced by these stimuli (Figure 5.4).

FIGURE 5.4. Trace from one animal showing complete abolition of norepinephrine- and angiotensin-induced vasoconstriction to a high intra-arterial (ia) dose of adenosine.


Trace from one animal showing complete abolition of norepinephrine- and angiotensin-induced vasoconstriction to a high intra-arterial (ia) dose of adenosine. The small rise in hepatic arterial blood flow produced by the constrictions during adenosine (more...)

Adenosine modulation of the nerve-induced constriction of the hepatic artery is likely to be postsynaptic because the constrictor response in the hepatic artery is eliminated by adenosine, but the portal constriction is not simultaneously significantly affected [217]. The importance of the ability of adenosine to interfere with general vasoconstrictors is seen by the protective effect that the buffer response has during hemorrhage where portal blood flow is reduced substantially with the resultant accumulation of adenosine serving to dilate the hepatic artery and, at least partially, counteract the effects of blood-borne and nerve-induced vasoconstriction [227].

5.2.3. Carbon Monoxide

Superoxide anions, H2O2, nitric oxide (Chapter 9), hydrogen sulfide, and carbon monoxide (CO) are active oxygen species that are produced endogenously and exert their biological actions in the liver. CO, a product of heme oxygenase, upregulates cyclic GMP via activation of guanylyl cyclase and thereby shares several biological actions with NO. Although there has not yet been a proposed specific role for CO in the regulation of the hepatic circulation, suppression of endogenous CO generation results in an increase in the vascular resistance of the liver [354]. Although it has been suggested that the endogeous level of CO generation is the major determinant of steady-state vascular resistance in the liver, the evidence is rather sparse for such a claim. Vascular effects of CO appear to be primarily through effects at the sinusoidal level acting on stellate cells. However, it must be noted that these studies are carried out in the perfused liver, and Greenway and Stark [122] have previously warned against the use of isolated or arterial perfused vascular circuits to imply functional roles in vivo because the responses in these preparations are considerably different.

5.2.4. Hydrogen Sulfide

A third gaseous mediator, hydrogen sulfide (H2S), is synthesized through degradation of cysteine by cystathionine-γ-lyase (CSE) or cystathionine-β synthase (CBS), both of which are found in the liver. In the vascular system, including the hepatic vasculature, CSE appears to be the only H2S-generating enzyme. Although H2S does not appear to regulate basal vascular tone, H2S donors potentiate the buffer response, whereas blockade of H2S formation reduced the buffer capacity. Blockade of ATP-sensitive potassium channels, using glibenclamide, reversed the H2S-induced increase of buffer capacity to the control level [344]. What other neurovascular modulating effects H2S may have in the liver has yet to be explored.

Image fig4.2
Image fig8.5
Copyright © 2010 by Morgan & Claypool Life Sciences.
Bookshelf ID: NBK53066


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...