Architectural changes during regenerative and ontogenic liver growth in the rat



Although liver architecture has a major impact on function, morphological aspects of liver growth are relatively neglected. In our recent experiments, the architectural changes of the rat liver were compared during 2 basic processes: ontogeny and regenerative liver growth. The hepatic tissue is constructed as structural/functional units, and probably the most established and well-defined such unit is the classic lobule. The extent and orientation of the lobules are variable in the liver, and this renders their accurate size determination more difficult. The filling of the liver vasculature by a colored resin nicely outlined the surface lobules, enabling an analysis of the alterations of these structures during liver growth. There are 3 structural components of postnatal physiological liver development: enlargement of the hepatocytes and expansion and multiplication of the liver lobules. However, the enlargement of the lobules is exclusively responsible for the regenerative liver growth following partial hepatectomy. The number of hepatic lobules does not change during this latter reaction, but they gain a more complex, irregular structure. Liver Transpl 15:177–183, 2009. © 2009 AASLD.

Partial hepatectomy (Phx) along with the subsequent regenerative response is one of the most studied nontumorous growth processes in mammalians. There are several excellent reviews summarizing the available information about the regulation of this well-synchronized hyperplastic process.1–3 The dynamics of cell proliferation/apoptosis are also well documented; therefore, that was not the object of this study. Unfortunately, our knowledge is much more limited regarding the structural changes,4 so this aspect of the growth was analyzed in this work. Although this process is mostly called liver regeneration, in a biological sense it should be referred to as compensatory hyperplasia because the resected lobes do not grow back. Instead, the remaining lobes enlarge as a result of hepatocyte proliferation. The liver size also increases several-fold during postnatal life. This growth process has been studied in several species,5–8 but it has not been compared with compensatory hyperplasia. The liver has a modular architecture.9, 10 Although the liver lobules may not be the true functional units, they are well-defined compartments that build up the liver. Therefore, hepatic growth can be achieved by an increase in the size and/or number of lobules. This question was addressed in our recent experiment in rats. The filling of the hepatic and portal venous system outlined the lobules on the liver surface very nicely. Morphometric analysis of the right lateral liver lobe revealed that the number of liver lobules did not change, but their size increased during compensatory growth. Conversely, during the studied postnatal period, the liver grew by enlargement and multiplication of the lobules. In summary, the adult liver lobes seem to be constructed with a standard number of lobules, and the fully developed liver is not able to form new lobules upon a further proliferative trigger.


CTR, control; nd, not determined; Phx, partial hepatectomy; RLL, right lateral lobe.


Experimental Design

Several parameters of the hepatic lobules were measured in 4-week-old (50 g), 8-week-old (160 g), and 12-week-old (250 g) male F344 rats; in rats 4 weeks (250 g) after Phx at 8 weeks of age (160 g); and at certain interim experimental points. The schema of the measurements and the number of the investigated rats are shown on Fig. 1. The animal study protocols were conducted according to the National Institutes of Health guidelines for animal care.

Figure 1.

Schematic representation of the experimental design. The upper line illustrates the time points of experiments after partial hepatectomy; the lower illustrates those during ontogeny. Numbers in parentheses represent the number of investigated rats. Abbreviations: C, CYP450 II E1 expression; H, hepatocyte size determination; N, total number of hepatic lobules on the convex surface of the right lateral lobe; P, portal vein number around the central vein; S, size determination (surface area and circumference) of hepatic lobules.

Liver Regeneration

Traditional, two-thirds Phx11 was performed on rats weighing 160 g under ether anesthesia. The size of the liver lobules was measured 1, 2, and 4 days and 1, 2, 3, and 4 weeks after the operation. The surface lobules of the right lateral lobe were counted on control rats (160-g bodyweight) and 2 and 4 weeks after Phx. The size of the hepatocytes was determined 4 weeks after the operation.

Physiological Liver Growth

The size of the surface lobules was measured in normal rats with the following bodyweights: 20, 50, 100, 160, 200, and 250 g. The number of surface lobules was counted on the right lateral lobe of rats with a bodyweight of 50 or 160 g. The average size of the hepatocytes was also measured for rats with a total bodyweight of 50, 160, 200, or 250 g.

Determination of the Size of the Lobules

For outlining the surface liver lobules, a cannula, washed with heparin, was inserted into the inferior vena cava. The blood was washed out from the liver with phosphate-buffered saline through the left ventricle of the heart. When the flush fluid became colorless, the vena cava was ligated above the level of the hepatic veins, and the portal vein was opened. The hepatic veins and the sinusoids were filled up through the vena cava cannula by a fluorescent dye containing polystirol resin and were monitored by eye with a stereomicroscope. The filling was stopped when the resin filled up the central veins and partially filled the hepatic sinusoids on the surface of the liver. At this stage, the negatives of the interlobular borders were outlined by the resin. The right lateral liver lobe was removed from the rats, its weight was recorded, and all further measurements were performed on this lobe.

The lobe was placed onto a wet slide and examined with a Nikon TE200 inverted microscope. The surface image of the lobe was captured with a Bio-Rad (Richmond, CA) MRC1024 confocal system (Ex488/Em520 ± 16 nm). The periportal zone appeared on the pictures as black areas. Interlobular borders were highlighted by a line drawn halfway between the central veins (along the vascular septa). The approximate center of the black areas determined the corners of the polygon representing the lobules. In cases of neighboring overfilled central veins, the border was determined by an analysis of the direction of the sinusoids. At the border zone, the sinusoids changed their direction from running toward one another other to running along the border. The circumference and surface area of the lobules were determined with the Image J program.

Counting of the Surface Liver Lobules

In another set of rats, the portal venous system was filled with a blue-stained resin in addition to the red resin–outlined hepatic veins. The blood was removed from the liver as described previously. The portal vein was filled with blue resin until it just entered the sinusoids, and this was followed by filling of the central veins. This method could not be applied to the determination of the size of lobules because of the slight enlargement of the liver. These animals were used to count the absolute number of liver lobules on the convex surface of the right lateral lobe. Counting of the lobules was performed by the placement of a mark on each lobule with an Indian ink pen with a 0.2-mm line width. The number of portal vein branches around a central vein was also counted on these specimens.

Hepatocyte Size Determination

Frozen sections from the liver were stained with fluorescein isothiocyanate–labeled pancytokeratin antibody (catalog number F0859, Dako, Glostrup, Denmark; dilution, 1:10), which outlined the cell membranes of the hepatocytes. The nuclei were stained with propidium iodide. The circumference and surface of 50 pericentral hepatocytes were measured in each liver lobule on 3 randomly selected fields. The size of hepatocytes is dependent on their zonal position,12 so data from different animals are more comparable if they are referring to a confined compartment, such as the pericentral zone. Furthermore, the radial orientation of the pericentral sinusoids makes possible a more exact determination of the outline of the hepatocyte borders versus the more irregular arrangement of the periportal hepatocytes.

Zonality of the Liver Lobules

The right lateral liver lobe was frozen under slight pressure to produce a flat surface for cutting. Frozen sections were made from the livers of control rats and the livers of rats 4, 8, and 12 weeks after Phx in a plane parallel with the surface. The section was fixed in methanol (−20°C) and stained for CYP450 II E1 (catalog number BV-3084-3, MBL, Woburn, MA; dilution, 1:100) for an hour at room temperature, and the reaction was visualized by a fluorescein isothiocyanate– labeled secondary antibody (30 minutes at room temperature). The sections were counterstained with propidium iodide.

Statistical Analysis

Statistical analysis was performed with the Student t test.


Morphometric Analysis of the Surface Liver Lobules During Liver Regeneration

The partial filling of the central veins and liver sinusoids through the hepatic veins outlined the lobules on the liver surface. This image was captured with a microscope, and the lobule borders were drawn on the digitalized image (Fig. 2A-C). Phx was performed on young, adult rats (8 weeks old, 160-g bodyweight), which are optimal for this procedure. All the investigated parameters of the livers were compared to the values measured on the right lateral lobe of unoperated rats (controls, 160 g). Both the circumference and surface area of the lobules grew gradually during the regenerative growth of the right lateral liver lobe. Both parameters increased rapidly in the first 7 days, and this was followed by a slight but not significant increase in the upcoming 3 weeks. Finally, the average lobular circumference increased 1.51-fold and the surface area increased 2.3-fold during our observation period. The weight of the studied lobe was also recorded, and it grew 3.29-fold (Fig. 3A). The number of liver lobules on the convex surface of the right lateral lobe was also counted in 160-g unoperated control rats and at 2 selected time points during regeneration. It did not show any significant alterations (Table 1). In brief, the liver grew by the enlargement of the liver lobules, whereas their number did not change.

Figure 2.

Retrograde filling of the hepatic sinusoids through the terminal (central) veins with green fluorescent resin: (A-C) 2, 4, and 28 days after partial hepatectomy and (D-F) livers of rats with bodyweights of 50, 160, and 250 g. The lobular borders (red lines) are shown in parts B and E. Note the enlargement of the lobules with increasing time (after partial hepatectomy) and age. Scale bar: 1 mm.

Figure 3.

The numerical dimensions of the growing liver during regeneration following (A) Phx and (B) ontogeny: (□) the circumference and (▵) the surface of the hepatic lobules. The bars represent the standard error. Abbreviations: CTR, control; Phx, partial hepatectomy; RLL, right lateral lobe.

Table 1. Number and Complexity of Surface Lobules in the Right Lateral Lobe of Growing Livers
 Control (Body Weight)Days After Phx
50 g160 g1428
  • Abbreviations: nd, not determined; Phx, partial hepatectomy.

  • *

    P < 0.05 between the 160-g control and the 50-g control.

  • #

    P < 0.05 between 28 days after Phx and the 160-g control.

Number586 ± 10.29794.5 ± 37.1*730.75 ± 54.15719.33 ± 31.18
Number of portal vein branches around the central veinnd6.04 ± 0.2nd8.13 ± 0.38#

However, the architecture of the enlarged lobules showed some characteristic changes. The images of the central veins on the liver surface became elongated and more branched. The shape of the lobules became more variable and polygonal. This observation was supported by the increased number of portal vein branches around the central veins (Fig. 4A-D and Table 1). The CYP II E1 enzyme showed a typical zonal distribution in the liver lobule.13 It was preferentially expressed by the pericentral and midzonal hepatocytes. The positive cells were surrounded by an evenly broad band of the negative cells in the control liver. The immunostaining of this enzyme showed a peculiar, arborescent distribution in the regenerated liver throughout our 3-month observation period, demonstrating the permanent functional modification of the lobular structure (Fig. 4E,F).

Figure 4.

Architectural changes of the liver lobules during regeneration. (A-D) Double filling of the liver vasculature by blue (portal veins) and red (central veins) resins: (A,B) control liver and (C,D) 28 days after partial hepatectomy. Note the enlargement of the hepatic lobules and the more complex branching pattern of the central veins in the regenerated livers. (E,F) Immunohistochemical staining for CYP450 II E1 (green) on (E) a control liver and (F) a regenerated liver 28 days after partial hepatectomy. Note the arborescent distribution of CYP450 II E1 on the regenerated liver. Nuclei were stained with propidium iodide. Scale bar: 1 mm.

Structural Characterization of Postnatal Liver Growth

In order to compare the regenerative/hyperplastic growth response of the liver lobules with the physiological, ontogenic liver growth, similar measurements were performed on the livers of rats of various ages. All these studies were also confined to the right lateral liver lobe. Similarly to the regenerative growth, the enlargement of the hepatic lobules was recorded with age (Fig. 2D-F). Although the bodyweight of the rats increased from 20 to 250 g (12.5-fold), the weight of the studied lobe increased 11.27-fold. The average surface area of the lobules increased 5.16-fold, and the circumference increased 2.68-fold. Additionally, the number of surface lobules also increased by approximately 30% (Table 1) while the bodyweight changed from 50 to 160 g (Fig. 3B). It is important to emphasize that the absolute values of these parameters are not comparable to those of the hepatectomized rats because the studied lobe represents a much smaller portion of the whole liver than that in the other experimental model.

Alteration of Hepatocyte Size During Liver Growth

Expansion of the lobules can be a result of hepatocyte enlargement. Therefore, the size of these cells was also measured. Indeed, obvious hepatocyte enlargement could be observed in the rats while the bodyweight increased from 50 to 160 g (Fig. 5). The average hepatocyte surface area grew 1.66-fold, and the circumference grew 1.29-fold. The weight of the investigated lobe increased 2.9-fold during this period. The size of the hepatocytes slightly but statistically not significantly increased further during ontogeny. Although the investigated lobe grew more than 3-fold during the regeneration, the size increase of the hepatocytes did not reach a significant level (Figs. 5 and 6) in that model either.

Figure 5.

(A) Surface and (B) circumference of the pericentral hepatocytes in rats with different liver weights and 28 days after partial hepatectomy. The bars represent the standard error.

Figure 6.

Immunohistochemical staining for cytokeratin (green) outlines the cell membranes of the hepatocytes; the nuclei are decorated by propidium iodide (red). Representative images of pericentral hepatocytes in the livers of rats with different bodyweights are shown: (A) 50 g, (B) 160 g, (C) 250 g, and (D) 28 days after Phx. Note the increasing cell size from part A to part C. (D) The examined liver lobe grew further during regeneration without a change in the size of the hepatocytes. Scale bar: 50 μm.


We have investigated the size of the surface liver lobules in 2 different liver growth models: during the regenerative response following Phx and during physiological postnatal liver growth. The liver grew exclusively by enlargement of the hepatic lobules during regeneration. Conversely, an increase in the number and size of the lobules contributed to the postnatal liver growth of ontogeny.

The liver tissue has a modular architecture.9, 10 There is relatively little information available about the behavior of these modules during hepatic growth. There are still ongoing debates about the real functional unit of the liver.14 This issue was not addressed by these experiments. We chose to study the classic liver lobule15, 16 because it is widely used and can be defined even in species (eg, in rat) in which it is not surrounded by interlobular connective tissue septa. However, the examination of these lobules is not an easy task, and this may be the explanation for the fact that the model of liver regeneration was described almost a century ago, but what happens with the lobules during this process is still not known.1 There are observations indicating the enlargement of the hepatic lobules during regeneration,17, 18 but this process has not been analyzed in detail. The 3-dimensional construction of the rat liver was carefully analyzed by Teutsch et al.9 According to this study, the size and shape of Teutsch et al.'s primary hepatic units (which correspond to the classic liver lobule) are variable. They are arranged in cone-shaped secondary units. The size of the primary units decreases toward the top of the secondary unit (the liver surface). Such morphogenetic plasticity explains why a traditional histological section does not provide reliable and comparable information regarding the size of the lobules: it will cut lobules at different positions in the hierarchy at different angles. In our present work, we took advantage of the fact that superficial lobules (in contact with the Glisson capsule) have a uniform apicobasal arrangement (perpendicular to the liver surface) and occupy an identical position in the liver hierarchy,5 contrary to the variable orientation, size, and situation of the lobules that are located deep in the parenchyma and are visible on a traditional histological section. The terminal branch of the hepatic vein is in the center of these lobules, generally oriented perpendicularly to and opening at the surface. Because the terminal portal venules terminate about 0.2 mm below the liver surface,9 only filling the liver retrogradely through the hepatic vein allowed us to determine the size of the surface lobules precisely. Thus, when sinuses of the surface lobules were filled up retrogradely through the hepatic veins, the outlines of the lobules provided reliable information about the 2-dimensional extension of these structural units. Enlargement of the liver lobules during the regenerative growth could be clearly demonstrated in this way. Because the number of subcapsular liver lobules did not change, the enlargement of the lobules was alone responsible for restoration of the liver mass. The measurements were confined to the 2-dimensional extension of the subcapsular hepatic lobules, and so they do not provide direct information about the changes deep in the liver tissue. However, considering the strictly defined cone shape of the primary and secondary units in Teutsch et al.'s model, we must assume an increase in the axial dimension of the subcapsular lobules as well as similar alterations of the deep hepatic lobules. Furthermore, Wagenaar et al.18 measured increased portocentral distances on traditional histological sections in regenerating livers and concluded that the lobules were enlarged; this experiment provided rough data about the deep lobules.

Teutsch et al.'s model9, 10 also maintains that rat and human livers are constructed of several layers of lobules. The weight/volume of the right lateral liver lobe expanded more than 3-fold during regeneration. This could have been accomplished, according to Teutsch et al.'s model, either by enlargement of the lobules or by the formation of an additional layer of lobules on the liver surface. Our results clearly support the first option. Highly developed mammalian organisms have a quite limited capacity for regeneration of complex structures.19 Therefore, it is reasonable that the lost liver parenchyma is regenerated exclusively by the enlargement of the preexisting lobules.

The enlarged liver lobules had a more complex structure throughout our observation period. The projection of the central vein on the liver surface was elongated and frequently divided. The number of portal vein branches at the periphery of the lobules also increased. This more complex structure of the lobules was reflected very well in the arborescent shape of the zonally distributed CYP II E1 expression, which was the result of the adaptive changes of hepatocytes.20 The distribution of 2 other zonally expressed enzymes (glutamine synthase and carbamoyl phosphate synthase) showed similar alterations (data not shown). This modified expression pattern was present throughout our observation period (3 months), indicating a permanent or at least long-lasting architectural change. The cause of the altered lobular structure is not clear, but it probably helps to maintain or approach the normal portocentral distance.

Although the size of the lobules grew during the investigated postnatal period of ontogeny, their number also increased. This is in agreement with previous observations. It is well known that the maturation of the biliary system and thus that of the lobular arrangement are completed only postnatally.21 The hilar-peripheral orientation of this process has been clearly demonstrated by the characterization of the liver of Alagille's patients.22 That is, a new layer of lobules might be formed postnatally. Ekataksin also reported the elongation of the portocentral distance13, 15 and multiplication of the liver lobules in the developing human liver. The size of lobular units increasing with age has been described in the rat liver.7 Studies of pigs have also shown that the average diameter, as well as the numbers of lobules, increases during normal growth of the liver.8 This growth involves an increase in the number and size of the hepatocytes. In fact, we have also observed the enlargement of the hepatocytes during the earlier period of the postnatal liver growth, as reported before.23 The periphery of the hepatocytes increased 1.29-fold during this period, and this predicted growth in the liver volume of 1.293 (2.15-fold). However, the liver weight (which is related to the liver volume) increased 2.9-fold, and this indicated that an increase in the cell number also contributed to liver growth. Later, during ontogeny and liver regeneration, the size of the hepatocytes did not grow significantly, probably because they approached the limit beyond which they could not function properly.

In conclusion, we observed 3 different structural mechanisms—enlargement of hepatocytes and multiplication and expansion of the hepatic lobules—contributing to postnatal liver growth during ontogeny. It seems that 2 of these parameters, the size of the hepatocytes and the number of lobules in a given liver lobe, are fixed by adulthood, and the liver is able to adapt by changing the size of the lobules. Under extreme conditions, the hepatocytes are able to further enlarge,24 but this state of the liver is not stable. The potential functional consequences and limitations of lobular enlargement remain to be studied.


The authors thank Noemi Baffy for correcting their use of the English language in this article.