This paper is dedicated to Prof. Dr. Ulrich Pfeifer on the occasion of his 70th birthday.
Potential conflict of interest: Nothing to report.
The microarchitecture of the liver is still not completely understood although various concepts of structural liver organization have been proposed. Among them, Rappaport's liver acinus stands out as one of the most accepted models. The correctness of this model, however, has also been doubted, and its applicability is hampered by the fact that the outlines of the liver acinus are disguised and nobody was ever able to give visual evidence by “unmasking” a simple liver acinus from the surrounding liver tissue. After intraportal transplantation of pancreatic islets or thyroid follicles into diabetic or thyroidectomized rats, respectively, the transplants engraft in small portal tracts and morphologically alter the downstream liver tissue due to excessive hormone secretion. Using a combined approach of perfusion fixation, stereomicroscopy, and light microscopy, we demonstrate in this study that these foci of altered liver tissue represent simple and complex liver acini, exactly as described by Rappaport. We present stereomicroscopical and histological examples of all important cut levels of altered simple and complex liver acini, including their topographical relation to the supplying and draining vessels and to the “central vein” liver lobule. Moreover, by computer-aided reconstruction of serial semi-thin sections, we were able to present the first 3-dimensional images of simple and complex liver acini. Conclusion: Our results prove the correctness of Rappaport's acinus model and confirm the simple liver acinus as the principal microcirculatory unit of the liver. (HEPATOLOGY 2007;45:705–715.)
Although the microarchitecture of the human and the rat liver has been investigated for about 200 years,1–4 it is still not completely understood and is thus a matter of debate. Various concepts of structural and functional liver units, excellently and concisely reviewed in Saxena et al.,5 have been suggested, and 2 opposing models stand out from the others at present. On the one hand, the “classic” hexagonal liver lobule with supplying portal veins in its periphery and the draining hepatic vein in its center (“central vein”) and, on the other hand, Rappaport's liver acinus (Fig. 1). While the former is based on a structural point of view derived from the microanatomy of the liver of the pig, which is subdivided into hexagonal lobules by fibrous septa, the latter is described as a hierarchal model of microcirculatory units (simple and complex liver acini) developed by Rappaport after having studied the staining patterns following dye microinjections and the microanatomy of hepatic lesions.6–8 Rappaport proposed the existence of an irregularly shaped simple liver acinus, made up by hepatocytes that drain the blood from a terminal portal vessel, the so-called terminal portal venule (TPV), which forms the central axis of the acinus, and from which a glomus of sinusoids branches off centrifugally. The venous blood is collected by draining hepatic venules (HV) in the periphery of these acini. He also claimed a metabolic heterogeneity of the simple liver acinus, leading to a subdivision into 3 metabolic zones (zones I-III) that surround the acinus axis like layers of a bulb. The 3-dimensional outlines of the acinus are thus understood as an ellipsoid. This model has been widely accepted, in particular for functional purposes, but has also been questioned by others, either in conceptual details that lead to several modifications9 or it has been rejected in general: Some investigators studying the human and rat liver microanatomy strongly oppose the existence of a simple liver acinus and the TPV.10–13
The main problem for many investigators when using the acinus concept for practical liver histology is founded on the lack of anatomical structures that help to visualize the outlines of the respective liver acini in particular when considering 3-dimensional aspects. The problem is intensified by the fact that neither in experimental animal models nor in the human liver has anybody yet succeeded in giving a direct visual evidence of a simple, model-propagated liver acinus. This is something even Rappaport himself was not able to do when using his dye injection models.6–7
Here, we report on our observations made by examining morphologically altered liver acini in animal models that were originally designed to study hormonally induced hepatocarcinogenesis.14–23 The altered morphology of the liver acini was induced by local action of hormones [such as insulin and tri-iodothyronine (T3)] on the hepatocellular metabolism after transplantation of endocrine tissues into the rat liver, resulting in considerable changes of the hepatocellular phenotype of the downstream liver acini, such as an excess in glycogen and/or fat storage in the insulin model14–18, 21–23 or increased basophilia and loss of glycogen in the T3 model.20 Therefore, we were able to depict the anatomy of a simple liver acinus for the first time, conclusively proving the correctness of Rappaport's acinus model. We also present single and complex liver acini cut at different levels and show their three-dimensional reconstruction to enhance the understanding of the liver microarchitecture for researchers working in this field of liver biology.
The animals evaluated in this study were originally used in our previous studies in 2 different models of endocrine hepatocarcinogenesis, based on local action of insulin and T3. Details of animal treatment and transplantation procedures have been described.14, 16, 20
Briefly, in the insulin model, Lewis rats (3 months old; body weight: 250-300 g) were made diabetic by injection of a single subcutaneous dose of streptozotocin (65 mg/kg body weight). Hormone deficiency in the T3 model was reached by thyroidectomy. Thereafter, the animals received isologous islet transplants (n = 350) or thyroid tissue pieces. The amount of transplanted tissue was intended to be insufficient to fully compensate the systemic hormone deficiency. This is a decisive prerequisite for the models to function and was carefully monitored by determining glucose and hormone blood levels as described.14, 16, 20 If the systemic hormonal deficiency is secured, each transplanted endocrine tissue piece is permanently stimulated to maximally synthesize and secrete the respective hormone into the portal/sinusoidal blood. In this study we used animals between 1 and 12 weeks after transplantation, because at this time interval, the altered liver morphology was dependent only on primary effects of the grafts as previously shown14, 20 and was thus optimal for the purposes of this experiment. That portal ischemia is of no relevance was proven by the transplantation of Latex particles that did not lead to any hepatocellular alterations in the downstream liver tissue.14, 17 Animal treatment was in line with the guidelines of the Society for Laboratory Animal Service and the strict German Animal Protection Law.
Animal Sacrifices, Tissue Sampling, and Processing
Rats were anesthetized, the aorta was cannulated, the inferior caval vein was cut, and the vessels were then rinsed for 2 minutes using Ringer's solution, mixed with 0.5% procain and 4% dextran, followed by perfusion fixation using a cocktail of aqua dest containing 4% dextran, 0.5% glutaraldehyde, and 3% paraformaldehyde. After fixation, the livers were removed, cut into slices of 1 mm thickness, and examined with a stereomicroscope. A small portion of liver slices containing altered liver acini were rinsed in 50% ethanol then placed into 70% ethanol for 5 minutes and in Sudan-IV solution for 10-20 minutes, followed by another 5 minutes in 70% ethanol and rinsing in 50% ethanol. For histological analysis of macroscopically selected altered liver acini, the tissue was embedded in paraffin. Serial sections (2-3 μm) were stained with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS) reaction, and cresyl violet.
For 3-dimensional reconstruction of liver acini, appropriate tissue specimens of 1.5-2 mm thickness were selected, postfixed in OsO4, and embedded in Epon 812. The tissue block was drilled (with drill holes measuring 200 μm in diameter) in exact vertical direction at 3 locations (Supplementary Fig. 1a) using a precision mechanical driller. Thereafter, 2000 serial semithin sections, each of 750 nm thickness, were cut and stained according to Richardson.24 Every tenth section was photographed (Supplementary Fig. 1b) and the whole tissue block, orientated at the three drill holes on each section, was 3-dimensionally reconstructed and analyzed using the Kontron Videoplan Image Processing System (Zeiss, Oberkochen, Germany). The border between 2 adjacent altered liver acini was reconstructed by analyzing the direction of the sinusoids in the individual acini.
Engraftment of Pancreatic Islets and Thyroid Follicles
Isolated pancreatic islets and thyroid tissue fragments measured between 50-200 μm in diameter. After their embolization into the portal vein, they streamed within the portal-venous blood into the liver and engrafted within portal tracts (Fig. 2 and Supplementary Fig. 2). Rarely, small transplant fragments, consisting of approximately 10 cells also engrafted within the beginning of the TPV but never in its middle or distal part. The number of liver acini that were actually under the influence of insulin or T3 depended on the diameter of the respective graft (see below).
Hepatocellular Alterations in the Insulin Model
The hyperinsulinemia in the downstream portal-venous blood led to several morphological alterations in the hepatocytes. Most conspicuous was a strong increase in glycogen and/or lipid storage that resulted in a clear-cell morphology in H&E stain and in a considerable enlargement of the individual cell. These alterations are typical insulin effects that have been studied in detail earlier, and were shown to be the results of insulin-induced alterations of enzymes of the carbohydrate and lipid metabolism.14–16, 21–23 Thus, the altered liver acini sharply contrasted to the hormonally unaffected surrounding liver tissue not only histologically but also even on macroscopic evaluation of unstained liver slices (see below).
Hepatocellular Alterations in the T3 Model
Owing to entirely different effects of T3 on the hepatocellular metabolism, the morphological alterations in this model were different. The hepatocytes showed an increase in cytoplasmic basophilia, a loss of glycogen, and an enlargement and hyperchromasia of their nuclei (Fig. 2E-H). This was previously shown to be the result of a peculiar pattern of alterations in enzyme activities in the T3 model different from that in the insulin model.20 In contrast to the insulin model, the overall cell size of the affected hepatocytes was not increased, which is important for the dimensions of the entire liver acinus (see below).
General Aspects of Altered Liver Acini
Owing to the combined technique of perfusion fixation and stereomicroscopy, the islet and thyroid grafts were easily recognizable on unstained liver slices. The accompanying liver acini, however, were macroscopically much better visible in the insulin model, which is why the macroscopic data and figures in this study stemmed from this model.
It should be noted that the formation of altered liver acini was a consistent finding in each individual animal and occurred in each liver acinus that was located downstream to a transplanted graft.
Morphology of the Simple Liver Acinus Depending on the Cut Level
Longitudinal Section of the TPV and Cross-Section of the HVs.
This cut level represents the idealized schematic drawing by Rappaport (Fig. 1A), and our results corroborate his concept. When the acinus is cut at its equator or only slightly transposed, as depicted in Fig. 2A-H, the longitudinal axis of the acinus harbors the TPV, and the blood streams from the portal tract through the TPV to a “dead end”. The sinusoids are supplied to both sides (2-dimensionally) of the TPV, while the HVs drain the blood at the end of the orthogonal acinus axis. The outer form of the acinus varies with regard to its cellular composition, depending in particular on the size of the hepatocytes. When the individual hepatocytes are enlarged, as it is the case in the insulin model (see above), the whole acinus is also enlarged and the outlines were convex (Fig. 2A-D). By contrast, when the hepatocytes are of normal size or are even slightly reduced owing to the glycogen loss in the T3 model, the acinus border approaches the HV much more directly without that sort of curvilinear extension seen in the insulin model, then giving the acinus the form of a rhombus (Fig. 2E-H).
Longitudinal Section of the TPV and the HV.
This situation is depicted in different examples in Fig. 3. This cut level is particularly often seen in liver slices that were made perpendicular to the liver surface, as shown in Fig. 3E, because TPV and HV often transport the blood in a perpendicular direction towards or from the liver surface.
Cross-Section of the TPV and Cross-Section or Longitudinal Section of the HV.
A cross section with both types of transversally cut vessels presents as an oval-shaped acinar outline containing the TPV in the center and the HVs at the periphery at both ends of the transversal axis (Fig. 4A). It is also possible to obtain longitudinal sections from the HVs when the TPV is cross-cut (Fig. 4B). This picture is also valuable, because it again demonstrates that the outlines of an acinus are irregular and depend on certain circumstances, such as the anatomic localization. In addition to the influence of the hepatocytes' size, as already mentioned, we can see in this figure that the liver surface modifies the acinus outlines to a more cubic rather than rugby ball-like form of the ellipsoid.
Zonal Organization of the Liver Acinus.
We found many altered simple liver acini that displayed a heterogeneity both macroscopically and histologically. In these cases, the periphery of the acinus, i.e., the area that Rappaport designated as zone III, showed a stronger glycogen and/or fat storage than the area that lies nearer to the TPV, particularly visible in animals that were less hyperglycemic. This pattern is depicted in Fig. 5.
Morphology of Complex Liver Acini
In cases when the transplant was so large that it engrafted in a portal tract from which several TPV branched off, the hormonal influence affected several adjacent simple liver acini, thus forming an altered complex acinus. One such example has already been shown in Fig. 3A, others are depicted in Fig. 6. With the knowledge of the morphology of a simple acinus, as described above, it is easier to understand the different cut levels of a complex acinus, in which several single acini are usually cut tangentially in their periphery. Figure 6A depicts the stereomicroscopic aspect of several adjacent simple acini in Sudan stain. The HVs were always located at the periphery of the lesions and particularly visible at 1, 3, 6, and 9 o'clock. The right part of this lesion offers a nearly horizontal view on two simple liver acini, as the cut level lies parallel to the respective TPV. It is noteworthy, that this example also displays the zonal heterogeneity within the simple liver acini as zones II and III of the acinus stain more strongly than zone I, which is due to the different lipid content of the hepatocytes. Another example is shown in Fig. 6B, which depicts two separate complex acini located beneath the liver capsule. The right one, in particular, is noteworthy, because several radially branching TPVs and the accompanying altered acini can be recognized in a horizontal section. Figure 6C displays a complex acinus at the histological level. This picture is of particular high value, because it shows the TPV, which is branching off toward 2 o'clock from the islet-containing portal tract, cut longitudinally in its entire length. One-half of the accompanying simple liver acinus is easy to recognize, as it drains to the HV at the lower right and the surrounding liver tissue is unaltered, thus not supplied with blood coming from the portal tract containing the islet graft. The other half drains to the tangentially cut HV at the top of the picture. This area, however, directly borders another altered simple liver acinus on the left. The TPV that supplies this second altered simple liver acinus lies above or below the cut level of this picture and moves from the islet to the upper left (approximately to 10 o'clock). Parts of at least a third altered liver acinus complete the complex altered liver acinus at the lower left.
Liver Acinus Versus Liver Lobule
Figure 7 is ideally suited to demonstrate the topographical relationship of the liver acinus and the hexagonal liver lobule. On the right side, a part of a complex liver acinus is depicted with an islet graft located in its center. The plane of 2 TPVs, branching off in an angle of 120°, is horizontally cut. These TPVs move toward 11 o'clock and 7 o'clock. Thus, these TPV represent 2 edges of a hexagonal lobule. The accompanying simple liver acini, supplied by these 2 TPV, lie with one half in this hexagonal lobule, representing 2 of its 6 segments, and their draining HV is the “central” vein. The area of the 2 altered segments is slightly larger than one-third of the lobule's overall area as a consequence of the increased size of the altered hepatocytes. The outlines of the (not exactly) hexagonal lobule bordered by another portal vein at its upper left part are indicated in the figure.
Three-Dimensional Reconstruction of Liver Acini
Computer-aided 3-dimensional reconstruction allows for the illustration of altered liver acini in their third dimension, including rotations around all axes and virtual sectioning of liver acini at any desired plane. Figures 8, 9 and Supplementary Fig. 3 depict the same 2 adjacent altered simple liver acini. They are shown at different angles and are rotated in different directions in Fig. 8 and Supplementary Fig. 3.
Figure 9 demonstrates that virtual sectioning of the 3-dimensionally reconstructed acini indeed yields cut profiles that exactly match the microanatomic pictures seen in the paraffin sections before. Figure 10 shows the 3-dimensional reconstruction of 2 adjacent complex acini receiving blood from 2 islet grafts. The blue color highlights the draining HVs only in those segments in which they border the altered acini. This figure is particularly suited to study the topographical relationship of the portal and hepatic veins.
It can be concluded from the 3-dimensional reconstruction that each actual and idealized section of altered liver acini in our model can also be achieved by adequate virtual sectioning of any 3-dimensionally reconstructed acinus.
Detailed studies on the microarchitecture of the liver date back to the first half of the nineteenth century. One of the first reports on a subdivision of the hepatic parenchyma was given by Kiernan, who investigated the pig liver and found hexagonal lobules of hepatic tissue separated by fibrous septa and surrounding the efferent hepatic venules, the so-called central veins.1
In 1954, Rappaport developed the model of the liver acinus as the primary hepatic unit after having studied the staining pattern of the hepatic parenchyma following dye microinjections.6 The results of this and subsequent studies have suggested that the TPV is the central axis of the smallest microcirculatory hepatic unit, the simple liver acinus, and the “central” veins, i.e., the efferent hepatic venules, lie at the acinus periphery, bordering adjacent liver acini. The main point of this concept is the fact that the irregularly shaped and irregularly sized simple liver acinus is supplied by only one afferent vessel and thus represents an area of individual blood supply that does not communicate with the neighboring tissue.
Although he was able to show in his model that the dye staining pattern resulted in sharp borders of stained to adjacent unstained tissue, this method was not able to depict a simple liver acinus.6–8 Thus, the correctness of his model could not be proven. In the following years several attempts have been made to visualize this smallest anatomical unit of the liver, often in a critical approach.
These studies can be categorized into 2 major groups: (1) The ones that primarily investigated the microangioarchitecture to subsequently interpret the tissue organization,12, 13 a way Rappaport employed himself. (2) Studies that investigated a peculiar metabolic function of hepatocytes organized in a specific manner within the parenchyma. The hepatic unit organization is then evaluated as a consequence of the distribution pattern of this function, such as the activity of a certain enzyme,10, 11, 25, 26 or, among others, the selectivity of cell permeabilization following digitonin-pulse perfusion of the liver27 or colloidal gold uptake.28
These studies yielded interesting but very contradictory data, and none was able to visualize Rappaport's liver acinus. Instead, different types of parenchymal units were newly introduced or revisited, such as the portal lobule,29 so called “cone-shaped primary lobules”, 6-8 of which form a classical (central-vein) but secondary lobule according to Matsumoto,12, 13 or “primary lobular units” and “liver modules” with differences to, but basically the same organization as the classic (central vein) liver lobule.10, 11
Using a specific experimental model based on the low-number intrahepatic portal-embolic transplantation of endocrine tissues it took about 50 years to prove the correctness of Rappaport′s acinus model since his first description in 1954. The following principles can now be concluded from our results:
(a) The Rat Liver is Organized in Liver Acini Which Constitute the Principal Microcirculatory Hepatic Unit.
Our results corroborate Rappaport's hypothesis that there is an area of hepatic tissue that is individually supplied by portal blood. This does not rule out that there may be some variations in this general concept in peculiar anatomic situations, for instance in direct vicinity to a very large portal tract, which we did not investigate in this study. However, we do not agree with other authors, who claim that there is a common inflow front originating from several portal tracts to supply a primary hepatic unit, as is the case with the “central vein” lobule and variations of it. The transplantation of pancreatic islets and thyroid follicles into the portal triads obviously changed the vascular microanatomy within the portal triads. Nevertheless, the principle acinar architecture of the microcirculatory hepatic units downstream of the portal triads was preserved and the altered acini presented exactly as they were described by Rappaport.
(b) The Liver Acinus Has a Variable Size and Shape.
This is one of the major problems for understanding the acinus concept. One has not only to accept that there are no anatomical structures in the normal liver that would help to outline the borders of the acinus but also that the acinus outlines are not uniform but depend on several factors, such as the hepatocytes' size (Fig. 2), functional gene activation (e.g., fatty acid synthase, see Evert et al.23) or the anatomical location (Fig. 4).
(c) The TPV is the Terminal Vessel that Supplies the Liver Acinus in the Rat Liver.
As Rappaport suggested, the TPV runs longitudinally and terminates in a dead end, supplying the acinus by forming a central axis. There is no communication of a TPV with TPVs from neighboring acini. Thus, the liver acinus is an autonomous circulatory unit. Recently, studies were published that deny the existence of a TPV in the rat and human liver.10, 11 Regarding the rat liver, these results are now refuted by our study as seen in several of our figures (e.g., Fig. 2C,E; Fig. 3A,C; or Fig. 6E). This is particularly important as the same authors conclude that “…the concept of the liver acinus cannot be applied to the liver of the rat”, a wrong assumption which we believe is the result of the method applied in that study. The cryostat sections of 15 μm thickness investigated by them where much too coarse to visualize such delicate structures as the TPV. Moreover, perfusion fixation, allowing for an optimal preservation of morphological details and also the removal of blood cells, as done in our model, is also important for identifying these tiny vessels. However, the existence of a further subdivision of the liver acinus, as proposed by McCuskey and some other authors (reviewed in McCuskey30), cannot be ruled out.
(d) Complex Liver Acini are Composed of Several Simple Acini.
Complex liver acini were already depicted by Rappaport in his dye injection studies.6–8 Our study corroborates his results. Small portal tracts represent the corresponding structure of the liver vasculature that supply complex acini. The size and shape of the complex acinus depends on the number and directions of TPVs originating from the respective portal tract. Hence, in our model, the size and shape of the complex acini depended on the size of the engrafted transplants, which measured between 50-200 μm in diameter.
(e) The Draining HVs Border the Simple and Complex Acini at the Circulatory Periphery.
While the TPVs follow a strict longitudinal direction, the HVs run more variably around the acinus surface (Figs. 8-10 and Supplementary Fig. 3). Thus, despite the idealized sections already given by Rappaport to enhance the intelligibility of the model (Fig. 1A), it is clear that depending on the cut direction, one can face a variety of longitudinal, tangential, or cross-cut sections of HVs, regardless of whether the TPV is longitudinally or cross-cut. This holds true in particular for complex acini, because the HVs form a complex meshwork of draining vessels around their surface as shown in the three-dimensional reconstruction of Fig. 10.
(f) Zonal Heterogeneity.
Rappaport suggested 3 metabolic zones (I-III) that envelope the central axis like layers of the bulb.6–8 He also claimed that zone III of both acinus halves of the idealized section plane communicate at the terminal end of the TPV (Fig. 1A). Other authors, investigating the activities of single or a combination of certain enzymes, were critical to this approach regarding the exact boundaries or overall number of the different zones or opposed the acinus concept in general.9–11 Our model is not designed to study the subdivision of the liver acinus into metabolic zones in detail. However, we found examples (Fig. 5) that match the situation as described by Rappaport. This may be the result of the metabolic situation (degree of hyperglycemia) and stronger insulin effects in zone III, owing to the predominant expression of the insulin receptor in the circulatory periphery of the acinus (perivenous zone) as shown by Jungermann and co-workers.31
In summary, from these results, we strongly believe that it is indeed Rappaport's liver acinus that constitutes the basic microcirculatory unit of the rat liver.
We are grateful to Jürgen Bedorf for technical assistance, Annelie Rowlin, Gerrit Klemm, and Thomas Jonczyk-Weber for graphical and photographical work, and Bernd Wüsthoff for critically reading the manuscript.