Abstract
The aims of this prospective study were to quantify steatosis in dogs with congenital portosystemic shunts using a fat-specific stain, to compare the amount of steatosis in different lobes of the liver, and to evaluate intra- and inter-Observer variability in lipid point counting. Computer-assisted point counting of lipid droplets was undertaken following Oil-Red-O staining in 21 dogs with congenital portosystemic shunts and 9 control dogs. Dogs with congenital portosystemic shunts had significantly more small lipid droplets (< 6 μ) than control dogs (p = 0.0013 and 0.0002, respectively). There was no significant difference in steatosis between liver lobes for either control dogs and CPS dogs. Significant differences were seen between observers for the number of large lipid droplets (> 9 μ) and lipogranulomas per tissue point (p = 0.023 and 0.01, respectively). In conclusion, computer-assisted counting of lipid droplets following Oil Red O staining of liver biopsy samples allows objective measurement and detection of significant differences between dogs with CPS and normal dogs. This method will allow future evaluation of the relationship between different presentations of CPS (anatomy, age, breed) and lipidosis, as well as the impact of hepatic lipidosis on outcomes following surgical shunt attenuation.
Keywords: steatosis, congenital portosystemic shunt, dog, liver
Introduction
Congenital portosystemic shunt (CPS) is a vascular anomaly that diverts portal blood past the hepatic sinusoids and into the systemic veins. This leads to impaired hepatic development because of reduced delivery of hepatotrophic substances such as insulin to the liver. Another, direct effect of the decreased hepatic circulation is the high level of toxins, hormones, and nutrients that bypass the liver and drain directly to the systemic circulation.4,27
Not surprisingly, a variety of histological perturbations have been described in dogs with congenital portosystemic shunts. These include hypoplasia of the hepatic portal vasculature, bile duct hyperplasia, arterial reduplication and evidence of ongoing liver damage, such as fibrosis and lipogranuloma formation.1,5,10,17,22,25 Congenital portosystemic shunt has also been reported as one of a variety of conditions leading to vacuolar hepatopathy in dogs.25 It has previously been difficult to objectively assess the degree of steatosis in individual patients as H&E staining does not permit differentiation between glycogen, lipid, or some other substance. Furthermore, there appear to be differences in the degree of histologic change between liver lobes for abnormalities such as microvascular dysplasia17 leading to potential underestimation of the severity with which an individual is affected if only one biopsy is taken, although this has not been proven for other abnormalities.11
The significance of lipogranulomas in dogs with CPS has been the subject of much speculation. They are thought to represent remnants of degenerating hepatocytes with a subsequent inflammatory reaction and are considered to provide circumstantial evidence for steatosis.3,5,10 Normal dogs develop lipogranulomas as a routine consequence of aging, but the number and size of lipogranulomas seems to be greater in age-matched dogs with CPS, although this has not previously been quantified.10 Hepatic steatosis (synonymous with “lipidosis”) may take a variety of forms. Microsteaotosis refers to a change in which lipid accumulation is in the form of small vacuoles. With macrosteatosis, the lipid vacuoles are as large as or larger than the cell nucleus and begin to distort other intracellular structures.3,30 Macrosteatosis is a recognized feature of hepatic lipidosis in cats5 and has been identified in association with congenital portosystemic shunts in dogs although not objectively assessed.1,17,22 The amount of microvacuolar change that can be attributed to lipid accumulation versus glycogen or other substances has likewise not previously been quantified.
Steatosis has been the subject of much recent attention in the human literature, as a means of predicting insulin resistance and gauging pre-existing liver injury in a variety of settings.3,19,24,26,30, Histological assessment is typically qualitative, including a description of the size of fat droplets (micro- or macrovesicular) and their location within the lobule, or semiquantitative, with a subjective multigrade scale based on the proportion of hepatocytes with steatotic droplets.12 There is an increasing amount of evidence that fatty infiltration in the human liver results in greater susceptibility to injury and impairs regeneration of the liver.7,19,24,26, It has thus become the convention to evaluate livers suspected of fatty infiltration histologically before using them for transplantation.26
In addition to the abovementioned clinical observations in humans, pre-existing liver injury as evidenced by steatosis has been shown to affect the liver’s tolerance to further injury and its ability to repair itself in experimental models.13,24,31 Hepatic regeneration as evidenced by an increase in liver volume and return of biochemical parameters to normal occurs in many dogs following attenuation of a CPS, or administration of hepatocyte growth factor.9,14,15,28,29 However, in some patients liver function never returns to normal, and portal hypertension results in development of multiple acquired shunts,9,16,20,21 leading to speculation that the liver’s ability to respond to normal trophic factors is impaired.
Although the type and degree of histological liver injury has been suspected to play a role in outcomes after surgery for congenital portosystemic shunts, previous studies using H&E staining have failed to identify a correlation between histologic changes and prognosis after surgery. However, these studies have focussed more on fibrosis and architectural changes, rather than the presence, degree and type of steatosis.1,17,22 The study of Baade et al showed some resolution of histological abnormalities, particularly necrosis and fibrosis.1 Lipidosis was described in some cases as “massive”, but not quantified or confirmed with special stains. Likewise, in the paper of Parker and others (2007) lipidosis was graded from 0 to 3, but not characterized in terms of size of lipid vacuoles, distribution or proportion of hepatocytes containing lipid droplets.22 The use of a stain specific for lipid should facilitate a more accurate assessment of both microvacuolar and macrovacuolar lipid accumulation within cells as well as within macrophages. Actual counting of lipid droplets using a computerized system (stereological point counting) has been shown to provide a more accurate and reproducible way of calculating hepatic lipidosis than previous visual grading systems.8
The aim of the present study was to describe the use of stereological point counting for characterizing the type and degree of steatosis occurring in the livers of dogs with congenital portosystemic shunts, and to compare this withcControl dogs.
Materials and Methods
Animals and samples
Twenty one dogs undergoing surgery for congenital portosystemic shunts between January 2009 and July 2010 at the Veterinary Medical Teaching Hospital, Davis, California. Fasting serum triglyceride levels were determined for 6 dogs immediately prior to biopsy. Samples were taken from the right medial, quadrate, or left medial liver lobe using either the wedge technique, or with a 4 or 6 mm biopsy punch, depending on the surgeon’s preference. In 13 shunt dogs, a biopsy was taken from each of the 3 lobes to enable comparison between 3 different sites within the liver.
Punch biopsies were also taken at necropsy from each of 3 lobes in 9 young, healthy dogs surrendered for euthanasia by a local shelter for non-medical issues and with laparoscopic biopsy forceps from a single lobe in three healthy young dogs presenting for laparoscopic ovariectomy.
Biopsies (52 samples from 21 CPS dogs and 9 Control dogs) were divided into at least two sections and submitted for Oil-Red-O staining.
Processing Methods
The biopsy was placed into a Tissue tek container which was then filled with Tissue tek OCT compound gel (Sakura Finetek, CA). The samples were then frozen in liquid nitrogen, prior to being cut into 7-μm slices, and then stained for lipid using a standard Oil red O protocol.
Point Counting
Each sample was assigned a random number so as to blind the observers to the identity of the patient, its shunt status and the liver division from which the biopsy was obtained. The slides were scanned with an Olympus VS110 whole slide scanner using an Olympus 20x UPlanApo N.A. 0.75 lens. The resulting images were analyzed using a commercially available system (Visiopharm, Denmark). Each digital image file underwent systematic uniform random sampling using the Visiopharm Integrator System version 4.2 (VIS), and a MicroImager module. The specimen was evaluated by counting lipid droplets and lipogranulomas which was performed in VIS with the Computer Assisted Stereological Toolbox module and a point counting grid. Each grid contained a number of coarse points (T) and a larger number of fine points (F; Figs. 1–3). The course points were used to evaluate the proportion of the scanned image that contained tissue. Each fine point was marked by a 9 μm cross that divided the point into 4 quadrants. A given feature was counted if it fell within the right upper quadrant. The fine points were used to count cell nuclei, lipogranulomas and lipid droplets. Each arm of the 9-μm cross was further divided into sections, enabling differentiation between lipid droplets of different sizes. Lipogranulomas were counted when there was an obvious aggregation of 2 or more macrophages containing lipid.
Figure 1.
Section of liver from a Control dog (40X magnification). The superimposed grids contain coarse tissue points (T) that determine whether or not a count is made for that section of the slide. For each tissue point that hits a readable portion of slide, there are a number of fine counting points (F). Features (in this case lipid droplets) are counted if they fall within the upper right quadrant of each fine point. The arms of each fine point are 9 μm in length, with colored divisions of 0–2 μm, 2–6 μm and 6–9 μm enabling easy categorization of droplet size (inset).
Figure 2.
Section of liver from a dog with CPS (40X magnification), demonstrating a variety of lipid droplet sizes.
Figure 3.
Section from a dog with CPS (10X magnification). A series of large, coalescing lipogranulomas (LG) can be seen to the left of the section, and hepatocytes containing large lipid droplets to the right of the section.
Thus, a final count for a biopsy sample contains the following information:
Number of coarse points that fell upon tissue: to enable calculation of the absolute amount of tissue analyzed, the density of lipid droplets per tissue point and comparison of features between biopsies of different sizes.
Number of fine points that fell upon nuclei.
Number of fine points that fell upon lipid droplets of the following sizes: ≤ 2 micron, 2–6 micron, 6–9 micron, >9 micron.
Number of fine points that fell upon lipogranulomas (classified as more than one cell containing pigments and/or fat characteristic of lipogranuloma).
Evaluation of intra-observer repeatability
Biopsy samples from 8 CPS dogs and 2 control dogs were analyzed by observer 1 on 2 separate occasions, and the counts compared.
Evaluation of inter-Observer repeatability
Biopsy samples from 9 CPS dogs and 6 control dogs were analyzed by two different observers and the counts compared.
Effect of liver lobe
Biopsies were taken from 3 different liver lobes in 16 CPS dogs and 9 control dogs. All counts were performed by a single observer.
Effect of shunt status, age, sex and triglyceride level
Point counts from a biopsy sample from the left division of the liver, performed by a single observer, were compared between 18 CPS dogs and 9 control dogs.
Statistical Analysis
Variability associated with different liver divisions was evaluated using Repeated Measures Analysis of Variance with a Huynh-Feldt correction for samples with different variances. Intra and interobserver repeatability was evaluated using the nonparametric Wilcoxon signed rank test. Differences between control dogs and dogs with CPS, and between dogs demonstrating a particular trait (eg, lipid droplets > 9 μ) were evaluated using Analysis of Variance. Correlation between age and point counts was evaluated using regression analysis. P values less than 0.05 were considered significant.
Results
Animals
Signalment and clinical signs
All CPS dogs suffered signs of poor hepatic function resulting from a single, congenital, portosystemic shunt. Details of signalment, shunt anatomy and age at time of biopsy are presented in Table 1. All dogs displayed serum biochemical evidence of hepatic dysfunction, including elevation of the liver enzymes alanine aminotransferase, alkaline phosphatase, and reduced levels of hepatic metabolites such as blood urea nitrogen, albumin and cholesterol.
Table 1.
Case details for patients undergoing Oil-Red-O staining and point counting, and electron microscopy.
| Case | Breed | Sex | Age (months) | Shunt type |
|---|---|---|---|---|
| Oil Red O | ||||
| 1 | Schnauzer | MC | 108 | E |
| 2 | Yorkshire | MC | 36 | E |
| 3 | Chiahuahua | MC | 15 | E |
| 4 | Yorkshire | F | 6 | E |
| 5 | Pekinese | M | 6 | E |
| 6 | Min Schnauzer | MC | 9 | E |
| 7 | Pug | FS | 24 | E |
| 8 | Soft coated wheaten | FS | 12 | E |
| 9 | Beagle | MC | 5 | E |
| 10 | Shiba Inu | FS | 60 | E |
| 11 | Chiahuahua | MC | 18 | E |
| 12 | Yorkshire | F | 5 | E |
| 13 | Yorkshire | FS | 132 | E |
| 14 | Shih Tzu | F | 7 | E |
| 15 | Yorkshire | F | 4 | E |
| 16 | Jack Russell Terrier | FS | 24 | E |
| 17 | Yorkshire | MC | 18 | E |
| 18 | Labrador | MC | 96 | I |
| 19 | Labrador | MC | 9 | I |
| 20 | Poodle Lab cross | F | 4 | I |
| 21 | Samoyed | M | 24 | I |
E = extrahepatic, F = female, FS = spayed female I = intrahepatic, M = male, MC = castrated male.
Seventeen dogs had an extrahepatic shunt and four dogs had an intrahepatic shunt. The age of the dogs ranged between 4 and 132 months at the time of surgery (median 13.5 months). Eleven dogs were male and 10 were female. The age was unknown for most of the control dogs, but the clean and undamaged state of their dentition suggested that they were less than 2 years of age.
Point Counts using Oil-Red-O samples
Intra-observer repeatability
There was a high level of correlation for counts performed by the same observer on different occasions, with no parameters showing a significant difference (Table 2).
Table 2.
Variation in mean point counts between two counts performed by the same observer.
| TP | Nuclei/TP | <2μm/TP | 2–6μm/TP | 6–9μm/TP | >9μm/TP | Lipo/TP | |
|---|---|---|---|---|---|---|---|
| Count 1 | 583.8 | 0.699 | 3.76 | 0.199 | 0.087 | 0.042 | 0.03 |
| Count 2 | 583.6 | 0.71 | 3.62 | 0.198 | 0.085 | 0.042 | 0.03 |
| SE | 3.35 | 0.01 | 0.12 | 0.004 | 0.006 | 0 | 0.009 |
| Correlation | 0.99 | 0.99 | 0.76 | 0.99 | 0.99 | 1 | 0.99 |
| P value | 0.95 | 0.39 | 0.26 | 0.8 | 0.77 | 0.69 |
Lipo = Lipogranuloma
TP = tissue points
SE = standard error
Inter-observer repeatability
There was no significant difference between observers for the most frequent variables (lipid droplets < 6 μm) although counts for individual patients varied by up to 20% (Table 3). Significant differences were observed for the less common variables such as lipid droplets 6–9 μm and number of lipogranulomas per tissue point (p = 0.023 and 0.01, respectively). There was also a significant difference in the number of nuclei per tissue point (p = 0.02). There were major differences in counts for lipid droplets > 9 μm, but this did not reach statistical significance.
Table 3.
Variation in mean point counts between two observers.
| TP | Nuclei/TP | <2um/TP | 2–6μm/TP | 6–9 um/TP | >9um/TP | Lipo/TP | |
|---|---|---|---|---|---|---|---|
| Observer 1 | 653 | 0.7 | 3.89 | 0.13 | 0.025 | 0.013 | 0.011 |
| Observer 2 | 698 | 0.65* | 3.71 | 0.15 | 0.063* | 0.019 | 0.003* |
| SE | 50.8 | 0.02 | 0.10 | 0.015 | 0.014 | 0.005 | 0.003 |
| Correlation | 0.85 | 0.95 | 0.1 | 0.94 | 0.85 | 0.98 | 0.79 |
| P value | 0.39 | 0.02 | 0.11 | 0.21 | 0.023 | 0.24 | 0.01 |
Lipo = Lipogranuloma
TP = tissue points
SE = standard error
Differences between groups
There was no significant effect of age, sex, shunt type or triglyceride level on point counts for CPS dogs. Counts per cell and per tissue point are shown for the different study groups in (Table 4).
Table 4.
Point counting results (mean ± S) for dogs with CPS and control dogs.
| TP | Nuclei/TP | <2μm/TP | 2–6μm/TP | 6– μm/TP | >9μm/TP | Lipo/TP | |
|---|---|---|---|---|---|---|---|
| CPS | 742+93 | 0.7+0.03 | 3.86+0.08 | 0.3+0.036 | 0.022+0.007 | 0.067+0.036 | 0.068+0.02 |
| Control | 898+155 | 0.499+0.03 | 3.66+0.19 | 0.013+0.024 | 0.0004+0.0009 | 0 | 0.004+0.01 |
| P value | 0.35 | 0.0002 | 0.0013 | 0.0002 | 0.083 | 0.31 | 0.089 |
TP = tissue point.
Control dogs had significantly fewer nuclei per tissue point than CPS dogs (p = 0.0002, Table 4). They also had significantly fewer lipid droplets < 2 μm and 2–6 μm per tissue point (p = 0.0013 and 0.0002, respectively). The number of droplets of 6–9 μm, > 9 μm and lipogranulomas per tissue point was subjectively lower for controls than CPS dogs, but did not reach statistical significance. Notably, all control dogs displayed lipogranulomas (mean 0.007, range 0 – 0.029). In contrast, CPS dogs had a mean of 0.068 (range 0 – 0.43, with 10/21 being above 0.029. When dogs in which no lipogranulomas were seen were excluded, the number of lipogranulomas was significantly higher for CPS dogs than for control dogs (0.106 ± 0.027 versus 0.0074 ± 0.034, p = 0.034).
Effect of liver lobe
Most variables evaluated were similar between the different lobes of the liver in either control dogs or CPS dogs, but there were some exceptions (Table 5). In control dogs, there were significantly fewer nuclei per tissue point (p = 0.007), and more lipid droplets < 2μm/cell in samples from the left division of the liver, but this difference was not sustained when the number of droplets < 2μm per tissue point was analyzed. Both control dogs and CPS dogs showed significantly fewer lipid droplets between 2 and 6 μm per tissue point in the left lateral liver lobe versus the other liver lobes (p = 0.02 and 0.012, respectively).
Table 5.
Point counting results (mean ± SD) for the different divisions of the liver in CPS and control dogs.
| TP | Nuclei/TP | <2μm/TP | 2–6μm/TP | 6–9 μm/TP | >9μm/TP | Lipo/TP | |
|---|---|---|---|---|---|---|---|
| CPS L | 709+83 | 0.79+0.05 | 3.93+0.03 | 0.29+0.1* | 0.018+0.015 | 0.077+0.092 | 0.02+0.006 |
| CPS C | 713+86 | 0.85+0.05 | 3.9+0.03 | 0.56+0.1 | 0.034+0.016 | 0.154+0.095 | 0.022+0.007 |
| CPS R | 766+83 | 0.79+0.05 | 3.9+0.03 | 0.43+0.1 | 0.015+0.015 | 0.061+0.092 | 0.022+0.0064 |
| P Value | ND | 0.33 | 0.85 | 0.012 | 0.41 | 0.08 | 0.49 |
| Control L | 898+155 | 0.499+0.03* | 3.66+0.19 | 0.013+0.024* | 0.0004+0.0009 | 0 | 0.004+0.001 |
| Control C | 786+155 | 0.54+0.03 | 3.30+0.19 | 0.036+0.024 | 0.0008+0.0009 | 0 | 0.003+0.001 |
| Control R | 913+155 | 0.59+0.03 | 3.8+0.19 | 0.053+0.024 | 0.0015+0.0009 | 0 | 0.003+0.001 |
| P Value | ND | 0.0007 | 0.18 | 0.02 | 0.2 | 0.5 | 0.83 |
C = central division, CPS = congenital portosystemic shunt, L = left division, Lipo = lipogranuloma, ND = not done because total number of tissue points dependent on random sampling by computer, R = right division, TP = tissue point.
Effect of triglyceride level
Triglyceride levels were within the normal range for all dogs in which they were measured (median 47, range 26 – 66 mg/dl (normal range 19–103 mg/dl). Point counts were no different for the sub-group of dogs in which triglyceride levels were measured and hence they can be considered representative of the whole group.
Correlation between number and size of lipid droplets and lipogranuloma formation There was a highly significant association between the number of lipid droplets and the number of lipogranulomas for all sizes of lipid droplet (p < 0.0001 in all instances). Dogs that demonstrated macrosteatosis had a significantly higher number of lipogranulomas than dogs that did not have macrosteatosis (0.027 ± 0.016 versus 0.139 ± 0.032, p = 0.004).
Discussion
The use of a digital system for counting lipid droplets following Oil-red-O staining confirmed previous reports of significant hepatic lipid accumulation in dogs with CPS.1,10,17,22,25–8 Despite attempts to objectify this method of analysis by creating “rules” for inclusion of lipid droplets in the count, it still relies on discretion by the operator as to the size of the lipid droplet, whether it falls within the appropriate quadrant or whether it actually touches the grid point and overlaps quadrants. This could explain the substantial and significant variation between observers. However, we demonstrated an extremely high correlation between counts performed by the same observer on the same samples at different times, suggesting that although the objective criteria are still subject to interpretation by individuals, once the observer has decided on their preferred conventions, they are highly reproducible. These results suggest that a single observer should always be used when attempting to interpret differences between study groups.
The present study was able to quantify the difference in number of lipid droplets within the livers of CPS dogs versus control dogs. Steatosis has been proven in experimental models to be a marker of liver injury13, and this also seems likely in dogs with CPS. The objective measurement tool described in the present paper thus provides a mechanism for future comparison of steatosis in dogs with different types of portosystemic shunt, dogs that recover normal liver function following surgery versus those that do not, assessment of changes that may occur following surgery in patients for which postoperative biopsies can be obtained and comparison of dogs with CPS against experimental models of liver injury. The results of the present study also confirm that CPS livers show significant evidence of steatosis even in the absence of the lipogranulomas and large intracellular vacuoles necessary to make the diagnosis using H&E staining. Failure of the counting system used in the present study to identify a significant difference between the number of lipogranulomas in control versus CPS dogs until all dogs that did not display lipogranulomas were excluded is probably counfounded by the fact that the control and CPS populations were not matched for age, which has a significant effect on development of lipogranulomas.10 Furthermore, the counting protocol used in this study did not take into account the size of lipogranuloma. Therefore, a dog with multiple microscopic lipogranulomas could return the same result as a dog with large lipogranulomas, despite the feature being far more prominent histologically in the second case. Further investigation is warranted into the size and location of lipogranulomas that represent normal “wear-and-tear” versus those that represent a pathological process specific to the livers of CPS dogs. The impact of the number and size of lipogranuloma, and their zonal distribution within the liver, on liver function and response to surgical attenuation in dogs with CPS is worthy of further investigation.
Steatosis has received a great deal of attention in the literature recently due to its presence in a variety of liver diseases in man, and in particular the conditions grouped as non-alcoholic steatohepatitis (NASH).2,7,12,30, It is as yet unclear whether steatosis precedes, or follows, liver injury in many patients. Hepatic lipid accumulation can also be related to diet.18 All CPS dogs of the present report were receiving the same diet, specially formulated for protein intolerance secondary to hepatic dysfunction, prior to surgery. Unfortunately, measurement of serum triglyceride levels is not standard practice and hence we were not able to retrospectively determine triglyceride levels for all the patients of this report. Nevertheless, triglyceride levels were within normal limits for the small number of patients in which they were analyzed, and no association between triglyceride level and lipid counts was observed.
Three previous veterinary studies attempted to determine the degree of histologic change before surgery and correlate changes to the dogs’ response to surgery. In the study of Parker et al (2008) severity of preoperative histologic changes in CPS liver samples failed to correlate with the prognosis after surgery.22 Baade et al (2006) compared biopsy samples of canine livers with CPSS before and after partial ligation of the shunt. Arteriolar and ductular proliferation, hypoplasia of the portal veins, and atrophy and steatosis were present preoperatively, and these signs resolved to some extent after surgery.1 Lee et al (2008) reported similar findings in that dogs with small numbers of portal veins were less likely to tolerate complete shunt ligation, and that vacuolation of hepatocytes resolved to some extent following surgery.17 However, none of these studies specifically assessed steatosis using special stains and it thus is possible that they failed to clearly correlate liver injury with postoperative outcome due to an underestimation of the injury present in each of their patients. Standard grading systems cannot differentiate between microvacolar changes due to glycogen, lipid or other substances, 2,12,23 whereas the present study clearly shows that most CPS patients have substantial amounts of lipid in their hepatocytes even if they only have microvacuolar changes.
Macrosteatosis has negative prognostic significance for liver regeneration in humans patients undergoing liver transplantation.19 Further investigation of the type and nature of steatosis in dogs with congenital portosystemic shunts and objective postoperative measures of surgical success such as contrast-enhanced CT is warranted to determine whether the type of steatosis has any prognostic significance.
The larger numbers of nuclei per tissue point between CPS and Control dogs can probably be explained by the combined phenomena that have previously been report in association with CPS; namely hepatic lobular atrophy, bile duct proliferation and arteriolar reduplication.3,5 This results in larger numbers of cells in a given portion of liver and artificially “dilutes” the number of lipid droplets if they were expressed as droplets per cell. Furthermore, in this study we did not attempt to differentiate between hepatocytes and other cells such as biliary epithelium. For this reason, we chose to report the data as lipid droplets per tissue point, as this is an expression of the proportion of lipid per quantity of liver, which is independent of the size or architecture of the liver, or the proportions of different cell types present.
The reason for significantly lower nuclei counts for the left division of the liver in Control dogs is not immediately clear. It may represent a previously undescribed difference between liver lobes, or it may be an artefact due to relatively small sample size. There is some evidence for physiological differences in blood flow to the different liver lobes due to streaming within the portal vein.6 Hence, blood from the spleen and/or the intestine may preferentially flow to a particular division of the liver, stimulating different patterns of cell growth. The present study also showed a significant difference in the number of lipid droplets in the left liver lobes of both Control and CPS dogs. A previous report found no difference between liver lobes in the histologic features of CPS assessed using H&E staining.11 Differential growth of different liver lobes has been observed following CPS ligation, but the reasons have not been clarified.25 Point counting using a larger sample size, and correlating it with clinical outcome and patterns of liver growth following surgical attenuation in dogs with CPS is indicated to explore this possible difference further.
In conclusion, computer-assisted counting of lipid droplets following Oil Red O staining of liver biopsy samples allows objective measurement and detection of significant differences between dogs with CPS and normal dogs. This method will allow future evaluation of the relationship between different presentations of CPS (anatomy, age, breed) and lipidosis, as well as the impact of hepatic lipidosis on outcomes following surgical shunt attenuation.
Acknowledgments
Barry Puget was invaluable in processing the Oil-red-O sections. We also thank the Canine Research Foundation (2007) for funding some of the research reported herein.
References
- 1.Baade S, Aupperle H, Grevel V, et al. Histopathological and immunohistochemical investigations of hepatic lesions associated with congenital portosystemic shunt in dogs. J Comp Path. 2004;134:80–90. doi: 10.1016/j.jcpa.2005.07.003. [DOI] [PubMed] [Google Scholar]
- 2.Brunt EM. Grading and staging the histopathological lesions of chronic hepatitis: the Knodell histology activity index and beyond. Hepatol. 2000;31:241–6. doi: 10.1002/hep.510310136. [DOI] [PubMed] [Google Scholar]
- 3.Burt AD, Mutton A, Day CP. Diagnosis and interpretation of steatosis and steatohepatitis. Semin Diagn Pathol. 1998;15:246–58. [PubMed] [Google Scholar]
- 4.Center SA, Magne ML. Historical, physical examination, and clinicopathologic features of portosystemic vascular anomalies in the dog and cat. Semin Vet Med Surg. 1990;5:83–93. [PubMed] [Google Scholar]
- 5.Cullen JM, van den Ingh TS, Bunch SE, et al. WSAVA standards for clinical and histological diagnosis of canine and feline liver disease: WSAVA Liver Standardization Group. Edinburgh: Saunders Elsevier; 2006. Morphological classification of circulatory disorders of the canine and feline liver; pp. 41–59. [Google Scholar]
- 6.Daniel GB, DeNovo RC, Sharp DS, et al. Portal streamlining as a cause of nonuniform hepatic distribution of sodium pertechnetate during per-rectal portal scintigraphy in the dog. Vet Radiol Ultrasound. 2004;45:78–84. doi: 10.1111/j.1740-8261.2004.04013.x. [DOI] [PubMed] [Google Scholar]
- 7.Farrell GC, Teoh NC, McCuskey RS. Hepatic microcirculation in fatty liver disease. Anat Rec. 2008;291:684–692. doi: 10.1002/ar.20715. [DOI] [PubMed] [Google Scholar]
- 8.Franzen LE, Ekstedt E, Kechagias S, et al. Semiquantitative evaluation overestimates the degree of steatosis in liver biopsies: a comparison to stereological point counting. Mod Pathol. 2005;18:912–916. doi: 10.1038/modpathol.3800370. [DOI] [PubMed] [Google Scholar]
- 9.Hunt GB, Kummeling A, Tisdall PL, et al. Outcomes of cellophane banding for congenital portosystemic shunts in 106 dogs and 5 cats. Vet Surg. 2004;33:25–31. doi: 10.1111/j.1532-950x.2004.04011.x. [DOI] [PubMed] [Google Scholar]
- 10.Isobe K, Matsunaga S, Nakayama H, et al. Histopathological characteristics of hepatic lipogranulomas with portosystemic shunt in dogs. J Vet Med Sci. 2008;70:133–8. doi: 10.1292/jvms.70.133. [DOI] [PubMed] [Google Scholar]
- 11.Isobe K, Matsunaga S, Nakayama H, et al. Hepatic lesions of a standard poodle dog with an intrahepatic portosystemic shunt. J Vet Med Sci. 2008;70:1125–1128. doi: 10.1292/jvms.70.1125. [DOI] [PubMed] [Google Scholar]
- 12.Kleiner DE, Brunt EM, Van Natta M, et al. Nonalcoholic Steatohepatitis Clinical Research Network. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatol. 2005;41:1313–21. doi: 10.1002/hep.20701. [DOI] [PubMed] [Google Scholar]
- 13.Koteish A, Diehl AM. Animal models of steatosis. Semin Liver Dis. 2001;21:81–94. doi: 10.1055/s-2001-12932. [DOI] [PubMed] [Google Scholar]
- 14.Kruitwagen HS, Arends B, Spee B, et al. Recombinant hepatocyte growth factor treatment in a canine model of congenital liver hypoplasia. Liver Int. 2011;31:940–949. doi: 10.1111/j.1478-3231.2011.02513.x. [DOI] [PubMed] [Google Scholar]
- 15.Kummeling A, Vrakking DJ, Rothuizen J, et al. Hepatic volume measurements in dogs with extrahepatic congenital portosystemic shunts before and after surgical attenuation. J Vet Intern Med. 2010;24:114–9. doi: 10.1111/j.1939-1676.2009.0439.x. [DOI] [PubMed] [Google Scholar]
- 16.Landon BP, Abraham LA, Charles JA. Use of transcolonic portal scintigraphy to evaluate efficacy of cellophane banding of congenital extrahepatic portosystemic shunts in 16 dogs. Aust Vet J. 2008;86:169–179. doi: 10.1111/j.1751-0813.2008.00278.x. [DOI] [PubMed] [Google Scholar]
- 17.Lee KCL, Winstanley A, House JV, et al. Association between hepatic histopathologic lesions and clinical findings in dogs undergoing surgical attenuation of a congenital portosystemic shunt: 38 cases (2000–2004) J Am Vet Med Assoc. 2011;239:638–645. doi: 10.2460/javma.239.5.638. [DOI] [PubMed] [Google Scholar]
- 18.Lombardi B. Considerations on the pathogenesis of fatty liver. Lab Invest. 1966;15:1–20. [PubMed] [Google Scholar]
- 19.Marsman WA, Wiesner RH, Rodriguez L, et al. Use of fatty donor liver is associated with diminished early patient and graft survival. Transplantation. 1996;62:1246–1251. doi: 10.1097/00007890-199611150-00011. [DOI] [PubMed] [Google Scholar]
- 20.Mehl ML, Kyles AE, Hardie EM, et al. Evaluation of ameroid ring constrictors for single extrahepatic portosystemic shunts in dogs: 168 cases (1995–2001) J Am Vet Med Assoc. 2005;226:2020–2030. doi: 10.2460/javma.2005.226.2020. [DOI] [PubMed] [Google Scholar]
- 21.Nelson NC, Nelson LL. Anatomy of extrahepatic portosystemic shunts in dogs as determined by computed tomography angiography. Vet Radiol Ultrasound. 2011;52:498–506. doi: 10.1111/j.1740-8261.2011.01827.x. [DOI] [PubMed] [Google Scholar]
- 22.Parker JS, Monnet E, Powers BE, et al. Histologic examination of hepatic biopsy samples as a prognostic indicator in dogs undergoing surgical correction of congenital portosystemic shunts: 64 cases (1997–2005) J Am Vet Med Assoc. 2008;233:1511–1514. doi: 10.2460/javma.232.10.1511. [DOI] [PubMed] [Google Scholar]
- 23.Schermerhorn T, Center SA, Dykes NL, et al. Characterization of hepatic microvascular dysplasia in a kindred of Cairn terriers. J Vet Int Med. 1996;10:219–230. doi: 10.1111/j.1939-1676.1996.tb02053.x. [DOI] [PubMed] [Google Scholar]
- 24.Selzner N, Selzner M, Jochum W, et al. Mouse livers with macrosteatosis are more susceptible to normothermic ischemic injury than those with microsteatosis. J Hepatol. 2006;44:694–701. doi: 10.1016/j.jhep.2005.07.032. [DOI] [PubMed] [Google Scholar]
- 25.Sepesy LM, Center SA, Randolph JF, et al. Vacuolar hepatopathy in dogs: 336 cases (1993–2005) J Am Vet Med Assoc. 2006;229:246–252. doi: 10.2460/javma.229.2.246. [DOI] [PubMed] [Google Scholar]
- 26.Spitzer AL, Lao OB, Dick AA, et al. The biopsied donor liver: incorporating macrosteatosis into high-risk donor assessment. Liver Transpl. 2010;16:874–84. doi: 10.1002/lt.22085. [DOI] [PubMed] [Google Scholar]
- 27.Stalker MA, Hayes MJ. Responses of the liver to injury. In: Maxie JE, editor. Jubb, Kennedy and Palmer’s Pathology of Domestic Animals. 5. Philadelphia, PA: Saunders; 2007. p. 309. [Google Scholar]
- 28.Stieger SM, Zwingenberger A, Pollard RE, et al. Hepatic volume estimation using quantitative computed tomography in dogs with portosystemic shunts. Vet Radiol Ultrasound. 2007;48:409–13. doi: 10.1111/j.1740-8261.2007.00268.x. [DOI] [PubMed] [Google Scholar]
- 29.Tivers MS, Lipscomb VJ, Scase TJ, et al. Vascular endothelial growth factor (VEGF) and VEGF receptor expression in biopsy samples of liver from dogs with congenital portosystemic shunts. J Comp Pathol. 2012;147:55–61. doi: 10.1016/j.jcpa.2011.09.001. [DOI] [PubMed] [Google Scholar]
- 30.Turlin B, Ramm GA, Purdie DM, et al. Assessment of hepatic steatosis: comparison of quantitative and semiquantitative methods in 108 liver biopsies. Liver Int. 2009:530–535. doi: 10.1111/j.1478-3231.2008.01874.x. [DOI] [PubMed] [Google Scholar]
- 31.Vetelainen R, van Vliet AK, van Gulik TM. Severe steatosis increases hepatocellular injury and impairs liver regeneration in a rat model of partial hepatectomy. Ann Surg. 2007;245:44–50. doi: 10.1097/01.sla.0000225253.84501.0e. [DOI] [PMC free article] [PubMed] [Google Scholar]