Dr A. McCormick, Liver Unit, St. Vincent's University Hospital, Elm Park, Dublin 4, Ireland.
Background : The pathogenesis of pruritus in cholestatic liver disease is poorly understood. Cutaneous mast cells and nerves are thought to contribute to pruritus in several dermatological diseases.
Aim : To determine if cutaneous mast cell density, neural density and mast cell–neural interaction are increased in patients with pruritus and cholestatic liver disease.
Methods : Skin biopsy specimens from (i) patients with pruritus due to cholestatic liver disease (CLDP+; n = 6), (ii) patients with chronic liver disease without pruritus (CLDP−; n = 5), and (iii) healthy controls (n = 6) were studied. Biopsies were dual stained immunohistochemically for mast cells and nerves.
Results : Mast cell density in the control group was not significantly different from that in CLDP+ group or from that in the CLDP− group. Similarly neural density was not significantly different between groups when assessed either in terms of total nerve area, or in terms of the number of neural elements seen. The frequency of mast cell–nerve contact was not significantly different between groups.
Conclusions : These findings suggest that mast cells, nerves or interaction between the two may not contribute to cholestatic pruritus. Therefore, therapies targeted at cutaneous mast cells or nerves are unlikely to be of benefit.
Pruritus is a recognized complication of cholestatic liver disease and may be an extremely distressing symptom. Furthermore, it is frequently refractory to pharmacological therapies and in extreme circumstance is accepted as a ‘quality of life’ indication for liver transplant in patients with chronic liver disease in whom transplantation would not otherwise be considered necessary.1
Despite the fact that antihistamines have been long used as a first-line treatment in the treatment of cholestatic pruritus, it is unclear whether cutaneous mast cells in fact contribute to the symptom. It is recognized that mast cells contribute to pruritus in other conditions.2–4 With regard to cholestatic liver disease, plasma histamine levels are higher in patients with pruritus than without pruritus.5 Furthermore, bile acids are known to be potent activators of mast cells.6, 7 However, it has never been established if cutaneous mast cells themselves contribute to cholestatic pruritus. Similarly, although it is known that cutaneous neural density is increased in patients with pruritic conditions8–10 it is unknown if alterations in cutaneous innervation contribute to pruritus in cholestatic liver disease.
The interaction between nerves and the immune system in general and between nerves and mast cells specifically has been the subject of much research over the last 20 years. Direct mast cell–nerve contact has been documented in the skin of patients with psoriasis,11 atopic dermatitis12 and nodular prurigo.13 Again whether mast cell–nerve contact may contribute to pruritus in cholestatic liver disease is unknown.
The aim of this study was to test the hypothesis that cutaneous mast cell density, neural density and mast cell–neural interaction are increased in patients with pruritus and cholestatic liver disease.
Patients and METHODS
Skin biopsy specimens
Three different group of patients were studied. Skin biopsies were obtained from six patients with cholestatic liver disease and pruritus (CLDP+ group) by performance of a punch biopsy (in five patients) or from the edge of the surgical wound at the time of liver transplant; secondly, five patients with chronic liver disease but without the symptom of pruritus (CLDP− group) from the edge of the surgical wound at liver transplant; thirdly, six patients without chronic liver disease or pruritus who were undergoing surgery for an unrelated reason (control group) from the edge of the surgical incision.
Performance of a punch biopsy.
After written consent had been obtained, punch biopsies were taken under aseptic conditions from normal appearing, non-lesional skin on the anterior abdominal wall, approximately 3–4 cm lateral to the umbilicus using a 4 mm diameter punch biopsy needle. Biopsies were placed in 10% formalin for subsequent histological study.
Liver biopsy specimens
For each patient liver tissue was also available for routine histological analysis, either taken from the ex-plant liver (where the patient was undergoing transplant), or from a biopsy performed for clinical purposes.
Skin biopsies fixed in 10% formalin were processed and embedded in paraffin blocks according to routine histological procedures. Tryptase-containing mast cells were identified in 6 μm-thick tissue sections with AA1, a monoclonal mouse antibody raised against human mast cell tryptase (Dako, Ely, UK). Endogenous peroxidase activity was quenched using 3% H2O2 for 30 min. Non-specific binding was blocked with 5% normal goat serum/5% bovine serum albumin in phosphate-buffered saline (PBS) for 1 h. Endogenous biotin activity was blocked using an Avidin/Biotin blocking kit (Vector Labs, Peterborough, UK). The primary AA1 antibody was applied and detected using a biotinylated horse antimouse IgG antibody and avidin-biotin conjugated horseradish peroxidase (Vectastain Elite ABC detection kit; Vector Labs). Colour development was achieved using the chromogen VIP (Vector VIP peroxidase substrate kit; Vector Labs), which rendered mast cells purple in colour.
After mast cells had been stained, a procedure to identify nerves in the same section was carried out. Blocking of non-specific binding was performed by 1-h incubation with 5% normal goat serum/5% bovine serum albumin in PBS. The primary PGP9.5 antibody was applied and detected using an avidin-biotin peroxidase method as above. A brown colour (to contrast with the purple colour of mast cells) was achieved using 3′,3′-diaminobenzidine (DAB; Vector DAB peroxidase substrate kit; Vector Labs). The tissue was then lightly counterstained with haematoxylin. Human tonsil served as a positive control for mast cells. Human appendix served as a positive control for nerves and for dual staining. Skin tissue incubated with 5% normal goat serum instead of primary antibody served as negative control. Separate sections from each biopsy were stained with haematoxylin and eosin (H&E) for routine histological analysis.
Quantitative histological analysis
Sections were examined for quantification of mast cell and nerve density. Biopsy and nerve areas were measured using the public domain NIH Image programme (http://rsb.info.nih.gov/nih-image/). For each section 15 fields were examined at 40× magnification as follows: five random fields (sampling the epidermis and superficial level of the dermis) were examined. The five fields immediately adjacent to each of these fields but deeper into the dermis were also sampled and finally five fields deeper into the reticular layer of the dermis were examined. In this way, it was possible to ensure representative sampling from all layers of the dermis and epidermis and to examine differentially the region of the superficial dermis and epidermis where it is thought that free nerve endings mediate the sensation of pruritus. The measured area of each field was 0.03 mm2 giving a total area of 0.45 mm2 for 15 fields.
Digital images of each field were stored on computer using image acquisition software (Adobe Photoshop 3.0; Adobe, San Jose, CA, USA). The total number of mast cells in each section was counted. The relationship of mast cells to hair follicles, sebaceous glands, sweat glands and blood vessels was noted. For each section mast cell density was calculated and expressed per mm2 tissue.
For each section neural density was measured in two ways. First, using the NIH Image programme the area of each individual neural element was measured. Total nerve area per section was then calculated and expressed per mm2 tissue. Secondly, the total number of individual nerve elements in each section was counted and expressed as the total number of nerve elements per mm2 tissue.
All sections were examined to determine if there was evidence of mast cell–nerve contact. This was defined as apparent direct membrane–membrane contact of mast cells and nerves. The mast cell number that was seen in contact with nerves in each section was expressed as a percentage of the total number of mast cells in that section.
For each biopsy (and therefore each patient) two separate sections cut at a minimum distance of 50 μ apart were examined. The value of each of the parameters measured was taken as the mean of the two values calculated for each of the sections.
The study was carried out with the approval of the Ethics Committee, St Vincent's University Hospital, Dublin.
Results are expressed as the mean ± S.E.M. Comparison between groups was made using either Student's unpaired, two-tailed t-test (if data had a normal distribution) or Mann–Whitney unpaired, two-tailed test (if data were not normally distributed). Differences between groups were considered to be statistically different if P < 0.05.
The patient profiles of the groups studied are shown in Table 1. All six patients in the CLDP+ group had chronic refractory pruritus. Two had chronic refractory itch as a result of primary biliary cirrhosis necessitating treatment with plasmapheresis. A third patient with primary biliary cirrhosis also required plasmapheresis and subsequently liver transplant for symptomatic relief of pruritus. The fourth patient developed severe pruritus as a consequence of ischaemic cholangiopathy because of late hepatic artery thrombosis after liver transplant. The fifth patient had chronic refractory pruritus as a result of chronic rejection/autoimmune hepatitis in her transplanted liver, and had multiple sessions of plasmapheresis. The sixth patient developed severe refractory pruritus in association with the development of a drug-induced vanishing bile duct syndrome. All patients at the time of biopsy were taking standard antipruritic medicines, or had a previous history of failed therapeutic trials of standard antipruritic therapies. All five patients in the CLDP− group had end-stage chronic liver disease necessitating liver transplant. None reported the presence of pruritus.
Table 1. Patient details of those with pruritus (CLDP+, a), with chronic liver disease without pruritus (CLDP−, b) and the control group (c). All patients in the control group had normal bilirubin and alkaline phosphatase levels
A Phos (unit/L)
A Phos, alkaline phosphatase; HAT, hepatic artery thrombosis; PBC, primary biliary cirrhosis; CR, chronic rejection; VBDS, vanishing bile duct syndrome (in this case idiopathic); ALD, alcoholic liver disease; PSC, primary sclerosing cholangitis; AIH, autoimmune hepatitis.
Alk Phos (unit/L)
Reason for surgery
Laparaotomy for colon cancer
Inguinal hernia repair
Inguinal hernia repair
Incisional hernia repair
Removal axillary lipoma
In all patient groups H&E staining revealed some mild histological abnormalities. Minimal or mild perivenular inflammation with mononuclear cell infiltration was seen commonly in the CLDP+ and CLDP− groups. In six biopsy specimens (two from the control group, one from the CLDP+ group, three from the CLDP− group) areas of moderate to severe perivenular neutrophilic inflammation were seen, with occasionally the presence of red cell extravastion. These findings were present only in some of the samples taken from the edge of a surgical wound and were never present in samples obtained from punch biopsies, and may reflect surgical damage to the skin.
Histological information regarding the liver tissue of the patients is shown in Table 1. In the CLDP+ group, four of six had evidence of canalicular cholestasis; one of six had stages two to three fibrosis and one of six had cirrhosis. In the CLDP− group, two of five had evidence of significant canalicular cholestasis and all had cirrhosis.
Mast cell analysis
Staining with AA1 resulted in clean and specific staining of mast cells. Mast cells were seen in both the papillary and reticular layers of the dermis but rarely in the epidermis (Figure 1). They were frequently seen in association with blood vessels, sweat glands, sweat ducts, sebaceous glands and hair follicles as well as in the tissue stroma (Figure 2). They were often seen close to nerves and apparent direct mast cell–nerve contact was seen in all experimental groups (see below). Mast cell distribution was not uniform. Clusters of cells were seen around blood vessels and the various skin appendages. Morphologically they exhibited either ovoid or elongated dendritic forms.
Mast cell distribution and morphology were the same in all experimental groups. Mast cell density measured in the epidermal and dermal areas was not significantly different between the CLDP+ group (46 ± 9 per mm2 tissue), the CLDP− group (41 ± 15 per mm2 tissue) and the control group (43 ± 12 per mm2 tissue; Table 2a). Furthermore, when analysis was performed focusing on the epidermal and superficial dermal areas, no significant difference in mast cell density was found between groups (Table 2b).
Table 2. (a) Mast cell (MC) and neural density in human skin (epidermis and dermis). (b) Mast cell and neural density in human skin (epidermis and papillary dermis). Mast cell density and the degree of dermal innervation were not significantly different between groups. Mast cell-nerve contact was seen in all groups but its frequency was not different between groups
Control (n = 6)
CLDP+ (n = 6)
CLDP− (n = 5)
(a) MC density (per mm2 tissue)
43 ± 12
46 ± 9
41 ± 15
Nerve area (μ2/mm2 tissue)
535 ± 197
799 ± 224
1252 ± 685
Number of neural elements (per mm2 tissue)
25 ± 9
32 ± 9
42 ± 14
Observed MC–nerve contact (%)
2.2 ± 1.4
2.0 ± 1.0
2.6 ± 2.6
(b) MC density (per mm2 tissue)
32 ± 14
37 ± 11
37 ± 14
Nerve area (μ2/mm2 tissue)
829 ± 432
1108 ± 396
1772 ± 991
Number of neural elements (per mm2 tissue)
44 ± 23
60 ± 17
73 ± 23
Observed MC–nerve contact (%)
4.2 ± 4.2
2.1 ± 2.1
5 ± 5
Analysis of nerves
Staining with PGP9.5 resulted in satisfactory staining of cutaneous nerves. Nerves were seen both in the dermal and epidermal layers (Figure 3). Larger trunks and nerve plexuses were seen in the deeper dermal layers. Smaller fibres were seen superficially in the dermis. Small fibres were also seen within the epidermis. Small fibres in the papillary dermis and epidermis were sometimes close to mast cells and sometimes in direct contact with them (see below). As with mast cells, nerves were sometimes seen in association with blood vessels, sweat glands, sweat ducts, sebaceous glands and hair follicles, especially in the deeper dermal layers. Their distribution throughout sections was non-uniform. Where large nerve plexuses were seen to occur around hair follicles, sweat glands, and sebaceous glands in the deep dermal layers, they were excluded from quantitative analysis, as inclusion of these predominantly autonomic nerves may have distorted the quantification of the population of sensory nerves involved in the mediation of the sensation of itch.
Neural distribution and morphology were the same in all experimental groups. Neural density was not significantly different between groups: when either assessed in terms of total nerve area (799 ± 224 μ2/mm2 tissue in the CLDP+ group; 1252 ± 685 μ2/mm2 tissue in the CLDP− group; 535 ± 197 μ2/mm2 tissue in the control group); or when assessed in terms of the number of neural elements seen per section (32 ± 9 per mm2 tissue in the CLDP+ group; 42 ± 14 per mm2 tissue in the CLDP− group; 25 ± 9 per mm2 tissue in the control group; Table 2a). Furthermore, when analysis was performed focusing on the epidermal and superficial dermal areas, no significant difference in neural density was found between groups (Table 2b).
Mast cell–nerve contact
Mast cell–nerve contact was a feature of all three group of patients (Figure 4): in two of six control patients, three of six patients in the CLDP+ group and one of five patients in the CLDP− group. The mean data for each group are shown in Table 2. On average in the control and CLDP+ groups 2% of mast cells were in contact with nerves, with 3% of those in the CLDP− group being in contact with nerves. All mast cell–nerve contact was seen in the dermis (mast cells were rarely seen in the epidermis). There were no significant differences between the frequency of mast cell–nerve contacts between groups.
Mast cell and nerve density in those with canalicular cholestasis
To determine if there was a correlation between histological findings in the liver and histological findings in the skin, further analyses were performed comparing mast cell density, neural density and mast cell–nerve interaction: (i) in those with and without canalicular cholestasis (irrespective of the presence of pruritus); and (ii) in those with or without cirrhosis (irrespective of the presence of pruritus). No significant difference in mast cell density, neural density or mast cell–nerve contact was found in either of these cases (data not shown).
To the best of our knowledge this is the first study to examine whether cutaneous mast cell, nerves or interaction between the two may have a role to play in the pathogenesis of cholestatic pruritus. We have demonstrated that mast cell density and neural density in the skin of patients with cholestatic pruritus are not significantly different from control. Furthermore, whereas we demonstrated the presence of mast cell–nerve contact in skin of controls as well as in patients with pruritic cholestatic liver disease, the frequency of its occurrence was not different between groups. These results suggest that cutaneous mast cells, nerves, or interaction between the two may not contribute significantly to the symptom of cholestatic pruritus.
It was the aim of this study to investigate the symptom of pruritus in cholestatic liver disease in general rather than in any specific liver disease. Accordingly patients were allocated on the basis of their symptoms. The CLDP+ group consisted of patients with severe intractable pruritus because of liver diseases known to be cholestatic, demonstrated by biochemical cholestasis, with or without histological evidence of canalicuar cholestasis. If mast cells or nerves are involved in the pathogenesis of cholestatic liver disease, significant changes should be seen in this group. Patients with chronic liver disease without pruritus made up the disease control group (CLDP−). The non-disease controls were normal subjects without liver disease. We assumed that the mechanism of pruritus is the same in all cholestatic liver diseases. This is logical and in line with clinical experience, but we cannot exclude the possibility that mechanisms differ in different cholestatic diseases.
Of note the presence of pruritus did not correlate with the presence of jaundice or canalicular cholestasis. Some of the pruritic patients had neither jaundice nor canalicular cholestasis. On the contrary, some of the patients with cholestatic disease, but without pruritus, had evidence of canalicular cholestasis. The presence of canalicular cholestasis did not correlate with altered mast cell density, neural density or mast cell–nerve contact in the skin.
In this study we used AA1, a monoclonal antibody to human mast cell tryptase, known to stain mast cells with excellent specificity and sensitivity,14 to visualize mast cells. Human mast cells exhibit phenotypic heterogeneity on the basis of their protease content, being classified as either MCT or MCTC depending on whether they contain either tryptase alone (MCT) or both tryptase and chymase (MCTC).15 In human skin the mast cell population is predominantly made up of MCTC cells, with up to 95% of the total being both tryptase and chymase positive.16, 17 We were therefore satisfied that use of AA1 ensured detection of the vast majority of mast cells present. Furthermore, the mast cell density reported here is similar that reported in other studies.18, 19
The PGP9.5 is a cytoplasmic protein that is widely expressed in the neurones of the central and peripheral nervous system in vertebrates,20 and its antibody detects neural elements well. Its advantage over certain other neural markers is its ability to stain all neural elements and we have found it useful in the visualization of both enteric and hepatic nerves. However, its strength is also its weakness and its disadvantage in the current study is that it is not specific for nerves that may mediate the sensation of itch. We attempted to overcome this potential problem first, by excluding from analysis large plexuses of nerves surrounding sweat glands, sebaceous glands and hair follicles, which are more likely to be autonomic in function;21 and secondly by performing a second analysis restricted to the epidermis and immediately adjacent superficial dermis, where free nerve endings are thought to serve as itch receptors.22
When these selective analyses were performed we still found no increase in the density of neural innervation in those with the pruritus of cholestasis (CLDP+). Furthermore, our concerns regarding the specificity of anti-PGP9.5 antibody were mitigated somewhat by the findings of other investigators that PGP9.5 density is altered in other dermatological conditions characterized by pruritus.9, 23 Our results, therefore, suggest that the symptom of pruritus in cholestatic liver disease is not due to alterations in the density of cutaneous innervation.
It is possible that antipruritic therapy in the CLDP+ group may have altered the histological findings in that group. However, we feel this is unlikely, and a literature review revealed no evidence that the standard antipruritic agents used might regulate mast cell populations.
This study supports the findings of other studies, which have demonstrated that both structural contact and functional interaction occurs between mast cells and nerves in skin tissue.11, 24, 25 However, whereas others have found increased frequency of mast cell–nerve contact in both non-pruritic11 and pruritic12, 13 conditions, we did not find this to be the case in the current study (activation of a normally developed, resident populations of mast cells and/or nerves could still account for the symptoms of pruritis – all we have seen here is lack of cellular changes).
The lack of alteration in mast cell density, neural density, or mast cell–nerve contact in cholestatic pruritus, therefore, stands in contrast to several dermatological conditions where there is evidence of mast cell hyperplasia, increased neural density and enhanced mast cell–nerve contact. If, as has been suggested, the cholestatic liver is producing one or more substances capable in some way of producing pruritus,26 the results of this study suggest that the mechanism is other than mast cell recruitment or neural hyperplasia in the skin. Furthermore, although bile acids are known both to be potent activators of mast cells6, 7 and to accumulate systemically in cholestatic liver disease, this study suggests that they are not recruiting cutaneous mast cells to release histamine as a mediator of itch.
As mast cells and nerves appear to be important players in the pathogenesis of several diseases of the skin in which pruritus is an important symptom, we hypothesize that because alteration in mast cell and neural density does not occur in the skin of cholestatic patients then it is likely that the pruritus of cholestasis is not a disease of the skin. This indirectly supports the hypothesis that cholestatic pruritus comes about as a result of biological alterations elsewhere than in the skin, such as is suggested in the hypothesis of opioidergic up-regulation.27
This study suggests, therefore, that in patients with chronic cholestatic disease cutaneous mast cells, nerves, and mast cell–nerve interaction do not have a significant role to play in the pathogenesis of pruritus. This is in contrast to certain dermatological diseases where mast cells, nerves and mast cell–nerve interaction may contribute to pruritus. In terms of treatment of this symptom, the current study, plus the lack of any significant evidence that histamine is involved, suggest that the use of antihistamines is unlikely to be of benefit, or is the use of topically applied agents (such as capsaicin) which are capable of modifying neural activity locally.
Special thanks goes to Eamonn Fitzpatrick and Margot Coady in the Department of Anatomy, Faculty of Veterinary Medicine, UCD, for their advice and assistance with regard to the processing of tissue samples.