Article Text

Download PDFPDF

Increases in intrahepatic CD68 positive cells, MAC387 positive cells, and proinflammatory cytokines (particularly interleukin 18) in chronic hepatitis C infection
  1. P H McGuinnessa,
  2. D Painterb,
  3. S Daviesb,
  4. G W McCaughana
  1. aThe A W Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Camperdown, Australia, bAnatomical Pathology Department
  1. Dr P H McGuinness, The A W Morrow Gastroenterology and Liver Centre, Royal Prince Alfred Hospital, Missenden Road, Camperdown NSW 2050, Australia.


BACKGROUND Upregulation of Th1 associated intrahepatic cytokines in chronic hepatitis C virus (HCV) infection should lead to a significant non-specific cellular immune response, a prerequisite for viral clearance. However, to date, the role of this non-specific response in HCV has been understudied.

AIMS To analyse the intrahepatic macrophage activity in chronic HCV infection by immunostaining and by quantitation of cytokine mRNA.

METHODS HCV positive liver tissues (chronic hepatitis, n=10; cirrhosis, n=5) were immunostained for CD68, MAC387, and semiquantitated by polymerase chain reaction for intrahepatic cytokine mRNAs (interferon γ (IFNγ), interleukin 1β (IL-1β), IL-6, IL-18, tumour necrosis factor α (TNFα), and macrophage inflammatory protein 1β (MIP1β)). HCV negative normal liver tissues (for cytokines, n=6; for immunostaining, n=5) were included as controls.

RESULTS MAC387+ cells were focally increased in areas of erosion at the limiting plate while lobular staining was minimal. CD68+ staining was diffuse in both portal (increased in HCV) and lobular areas. The portal tract (mean) density of CD68+ and MAC387+ cells was significantly increased in patients with HCV compared with normal tissue. IFNγ and IL-18 mRNA levels were highly correlated and significantly upregulated in chronic hepatitis and cirrhotic tissue versus controls. TNFα mRNA was upregulated in chronic hepatitis without cirrhosis, while IL-6 mRNA was significantly downregulated. IL-1β, IL-6, and MIP1β mRNA levels were significantly correlated with portal tract MAC387+ cell density.

CONCLUSIONS The significant upregulation of IFNγ and IL-18 mRNA and significant correlations between IFNγ and other proinflammatory cytokines, suggest a Th1/cell mediated intrahepatic immune response in chronic HCV infection. However, further clarification of the cellular sources of these cytokines is required.

  • hepatitis C
  • macrophage
  • cytokine
  • interleukin
  • MAC387

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

It is well established that hepatitis C virus (HCV) infection becomes chronic in greater than 80% of cases.1-3 The reason why HCV is not cleared from these subjects is not known. Intense research into the complex interaction between viral factors such as genotype,4-6 viral load,6-9 quasispecies formation,5 10 11 high viral turnover,12and host immune factors such as antigen presentation,13 14 cell mediated immune response,15-17 and the humoral immune responses18-20 has yet to reveal any clear reason for chronicity.

Experimental evidence suggests that the immune response to HCV is site specific (compartmentalised).21-24 In this model, the peripheral immune response maybe a T helper type 2 (Th2) response—that is, HCV induces Th2 cells to proliferate, which in turn triggers a humoral immune response. In contrast, a cell mediated immune T helper type 1 (Th1) response appears to exist in the liver.15 Our laboratory has previously reported a significant intrahepatic upregulation of Th1 associated cytokines (interleukin 2 (IL-2) and interferon γ (IFNγ)) and a significant downregulation of the Th2 cytokine IL-10.23 Dumoulinet al likewise found that intrahepatic IL-2 and IFNγ mRNA levels were raised in hepatitis C, but they did not find IL-10 mRNA downregulation.16

A persistent Th1 response may cause a gradual accumulation of liver injury primarily induced by cytotoxic T lymphocytes (CTLs), natural killer (NK) cells, and macrophages. Although many studies have found evidence that suggests persistent intrahepatic anti-HCV CTL activity in chronic HCV, only a few reports (mostly abstracts) have investigated intrahepatic macrophage activation.25-33 The high expression of IFNγ present in HCV infected livers should activate both infiltrating monocytes/macrophages and resident macrophages (Kupffer cells).34 These macrophages should then express a range of cytokines such as tumour necrosis factor α (TNFα), tumour growth factor β (TGFβ), IL-1β, IL-6, IL-10, IL-12, IL-18, and chemokines such as macrophage inflammatory proteins (MIP), which would strengthen the continuation of a T1 response within the liver and further encourage the migration of macrophages/monocytes and T cells to the inflammatory site. Also, activated intrahepatic macrophages may take on a cytotoxic profile and destroy infected hepatocytes directly.35

Unfortunately, data on intrahepatic macrophage activity in HCV infection are sparse and contradictory. Two studies have detected the upregulation of CD14 expression on macrophages (marker of macrophage activation) in active hepatitis.32 35 36 This upregulation was shown to be directly related to the degree of liver injury.33 In addition, an increase in CD68 (universal marker for monocyte/macrophage lineage35) positive macrophage density has been found to be associated with increased liver damage in HCV infection.28 35 In contrast, a study by Marrogi and colleagues28 failed to find any association between another macrophage marker, MAC387 (marker for a subset of reactive/infiltrating monocytes/macrophages,37-39 but not Kupffer cells37 40-43) and HCV associated liver injury.

Evidence from two studies on intrahepatic macrophage cytokine levels26 27 found reduced expression of IL-1β, IL-6, and TNFα in chronic HCV, which would appear to be inconsistent with macrophage activation. Similarly, a study by Khakooet al found no significant difference between two different macrophage subsets (antigen presenting versus phagocytic) when three different stages of disease (mild hepatitis, active hepatitis, and advanced cirrhosis) were compared.25

The aim of this study was therefore to investigate macrophage activity within the HCV infected liver by jointly analysing the macrophage infiltrate and state of macrophage activation (immunostaining for macrophage markers and measuring the intrahepatic mRNA expression of cytokines associated with macrophage activation, including the chemokine MIP1β and the recently recognised interferon inducing factor IL-18).



For the cytokine mRNA and immunostaining analysis liver tissues were collected from 15 patients with chronic hepatitis C infection (biopsy, n=12; resected tissue collected at the time of hepatectomy for liver transplantation, n=3). All liver tissues used for cytokine mRNA analysis were stored immediately in liquid nitrogen on collection and no patient had ever been treated with interferon. The control group for cytokine analysis consisted of HCV negative (anti-HCV antibody negative and reverse transcriptase polymerase chain reaction (RT-PCR) negative) histologically normal liver tissues (n=6): three were from resected carcinoma (unaffected areas for secondary hepatic malignancy) and the other three were cadaver donors (at the time of hepatectomy). The control groups for immunostaining consisted of group A—three cadaver donor tissues (also used in cytokine estimation); and group B—five randomly selected formalin fixed liver biopsy specimens that were histologically normal by haematoxylin and eosin (H&E) staining. Note that carcinoma tissues (as used in cytokine analysis) were unavailable for immunostaining.


A portion of each specimen was fixed in formalin and mounted in paraffin wax. Sections were stained by H&E and for the expression of the following antigens: CD68 (clone KP1, Dako Corporation, Carpinteria, California, USA) and MAC387 (clone MAC387, Dako Corp.). For CD68 staining, heat induced epitope retrieval (HIER) in Tris-citrate EDTA (pH 8.0) was performed prior to application of the primary antibody. For MAC387 staining enzymatic predigestion with pronase (bacterial protease type XXIV) was used to enhance sensitivity. Following incubation in primary antibodies immunoperoxidase staining was completed using labelled strepavidin-biotin (LSAB kit, Dako Corp.) and diaminobenzidine (DAB) as chromogen.

The density of cells within the portal tract and liver lobule (macrophages/monocytes), which stained positive to the various monoclonal antibodies was determined by counting the number of cells in an area 0.125 mm2 at a magnification of 400× under a light microscope. Where possible serial sections and mirror sites were used. In each case, for the portal tract and lobular region, at least three fields and 10 fields, respectively, were examined with the mean expressed as cell number per mm2. The cell counting was done in an independent fashion by two of the investigators (PHM and DP) with good correlation: portal CD68 (r=0.58, p=0.003), lobular CD68 (r=0.5, p<0.02), portal MAC387 (r=0.55, p=0.006), and lobular MAC387 (r=0.6, p=0.003). All specimens were histologically assessed by a single pathologist (DP) and graded according to the Scheuer44 classification. HCV positive liver tissue was classified as either chronic hepatitis (n=10) or cirrhotic (n=5).


Liver tissue (approximately 5–10 mg) was homogenised in a liquid nitrogen cooled mortar and pestle; the RNA was extracted as described previously.45 The RNA pellet was washed in 75% ethanol and resuspended in 20 μl of diethylpyrocarbonate (DEPC) treated autoclaved H2O. The extracted RNA was stored in 1 μg aliquots (determined by spectrophotometry at 260 nm) in liquid nitrogen. All liver tissues (patient and control) were tested for hepatitis C RNA by nested RT-PCR.7


For each sample, a 1 μg aliquot of liver RNA was primed with oligo-dT (Boehringer Mannheim, Germany) and reverse transcribed in a 20 μl reaction using Superscript RNase H-enzyme kit (Gibco BRL, Gaithersburg, Maryland, USA) as described previously.23 46 47 Two separate cDNA syntheses were performed using two different aliquots of RNA from the same test subject. The two separate cDNA products produced were aliquoted and stored in liquid nitrogen until required.

Next, two separate PCR reactions were performed as described previously.23 47 Briefly, each 50 μl reaction contained 2 μl cDNA, 0.125 mmol/l dNTP, 15 pmol forward/reverse primers, 1.5 mmol/l MgCl2 (1.25 mmol/l for TNFα, IL-18), and 1 unitTaq polymerase mixed in 1× enzyme buffer. After an initial denaturation at 94°C all tubes were amplified (94oC, 30 seconds; 55°C, 30 seconds; 72°C, 60 seconds) for either 26 cycles (TNFα, IFNγ), 24 cycles (IL-1β, IL-6, IL-18, MIP1β), or 14 cycles (aldolase B). A 4.5 μl aliquot of PCR product was removed from each tube and transferred to a 96 well plate. Eight aliquots were collected at two cycle intervals. All aliquots were diluted with 100 μl of 0.4 M NaOH, 10 mM EDTA and blotted onto Hybond N+ membrane (Amersham, Buckinghamshire, UK) using a dot blot apparatus (Bio-Rad, Richmond, California, USA). The membranes were hybridised at 42°C with an internal probe end labelled with33P (DNA 5′ end labelling kit, Promega, Madison, Wisconsin, USA) in Amersham rapid hyde buffer (Amersham, Buckinghamshire, UK) and washed in 5× saline sodium citrate (SSC)/0.1% sodium dodecyl sulphate (SDS) followed by two washes in 1× SSC/0.1% SDS. Membranes were exposed on a phosphor imaging plate (Fiji Photo Film Co., Tokyo, Japan) for approximately 1–3 hours, which in turn was scanned on a Fujix Bioimaging analyser, BAS 1000 (Fuji) to quantify the radioactive intensity.

The concentration (molecule number) of cytokine cDNA for each sample was calculated from a standard curve constructed from standards containing a known concentration of cytokine PCR product (DNA molecules) that was included in each PCR run as described previously.23 47 Note that two separate assays per cytokine were performed; the average of these two assays was used in the analysis. Table 1 displays primer and probe sequences used in this assay.

Table 1

Nucleotide sequence of primers


In every RNA extraction, cDNA synthesis, and PCR assay, several reactions with only DEPC treated H2O were included as a contamination control. In addition, the omission of reverse transcriptase from the cDNA synthesis was included as a negative control for each sample; to control for run to run variation, duplicate PCR and hybridisation reactions were performed for each sample. Contamination prevention measures of Kwok and Higuchi were adhered to at all times.48

The dot blot (DB) PCR assay has been validated previously.7 23 47 In this study, the interassay reproducibility between the separate cDNA/DB-PCR assays was found to be highly significant: aldolase B (r=0.89, p<0.0001), IFNγ (r=0.91, p<0.0001), TNFα (r=0.79, p<0.0001), MIP1β (r=0.97, p<0.0001), IL-1β (r=0.94, p<0.0001), IL-6 (r=0.95, p<0.0001), and IL-18 (r=0.88, p<0.0001). Secondly, the quality of mRNA extracted from the different liver tissues was compared by measuring aldolase B mRNA expression (internal control—that is, constitutively expressed by hepatocytes) and the amount measured was similar for all samples (2.2 (0.5) × 106 aldolase B cDNA molecules/μg total liver RNA, n=21). No statistical difference was noted in the cytokine mRNA levels between the control carcinoma and cadaver liver tissues; therefore, for statistical analysis they were combined. Finally, negative controls (omitted reverse transcriptase and H2O substituted for hepatic cDNA) failed to amplify any product.


Statistical analysis was performed using Statview version 4.5 software (Abacus Concepts, Berkeley, California, USA). Mean values were compared by the Mann-Whitney test. Individual results were compared by linear regression analysis. Results are expressed as mean (SD) and p values less than 0.05 were considered significant.


The age of HCV positive patients ranged from 32 to 61 years (mean 41.9 years); 10 patients were male and five were female (table 2). No statistical association was found between the patients' demographic details and any of the cytokine or immunostaining results.

Table 2

Demographic and histological profile of patients with chronic hepatitis C infection


CD68 and MAC387 antigens were both expressed in the portal and lobular regions of normal and diseased livers. The anti-CD68 monoclonal highlighted both infiltrating monocytes/macrophages and resident Kupffer cells in normal and diseased liver tissues. The morphology of cells expressing CD68 antigen appeared to be site specific—the portal tract contained predominantly (CD68+) round/oval shaped cells (presumably infiltrating macrophages/monocytes) while the lobular region contained predominantly (CD68+) star shaped or elongated shaped cells (presumably resident Kupffer cells) (see fig 1A, C, E, G). In contrast, the MAC387 antigen positive cells (infiltrating “activated” myelomonocytic cells40 42) were primarily located in the portal region, especially in areas of “interface hepatitis” and were round or oval in shape (see fig 1D, F, H).

Figure 1

Immunostaining of HCV positive and negative liver tissues with CD68 and MAC387 monoclonal antibodies (original magnification × 200). (A) Normal tissue stained with CD68; (B) normal tissue stained with MAC387; (C) chronic hepatitis tissue stained with CD68; (D) chronic hepatitis tissue stained with MAC387; (E) chronic hepatitis tissue stained with CD68; (F) chronic hepatitis tissue stained with MAC387; (G) cirrhotic tissue stained with CD68; (H) cirrhotic tissue stained with MAC387.

The density of CD68+ staining cells (fig 2A) within the portal tract was significantly higher in the cohort of HCV positive livers with chronic hepatitis and/or cirrhosis compared with the normal liver control groups: group A (cadaver donors, p=0.008) and group B (randomly selected normal liver tissues, p=0.001). A similar pattern was observed for the MAC387+ cells (HCV versus normal group A, p=0.02; and normal group B, p=0.001; fig 2C). These observations contrast with the liver lobule where there was no significant change in CD68+ density between chronic hepatitis, with or without cirrhotic tissue, compared with the normal controls (fig 2B). Finally, the mean density of lobular MAC387+ stained cells was greater than both the normal control groups, but this difference only reached statistical significance in group B normal liver control (p=0.006; fig 2D). Furthermore, there was no significant correlation or trend in either the portal tract or liver lobule between CD68+ or MAC387+ cell density in HCV infected tissue, and the total Scheuer score or any of its components.

Figure 2

Comparison of CD68+ and MAC387+ macrophage cell densities. (A) Portal tract CD68 staining; (B) liver lobule CD68 staining; (C) portal tract MAC387 staining; (D) liver lobule MAC387 staining; p values denote the statistical difference between normal liver (groups A or B) versus chronic hepatitis.

Finally, it should be noted that: (1) the chronic hepatitis and cirrhotic groups were combined in this analysis as cell numbers were not statistically different between these two groups; (2) we only measured cell density and not total cell population and therefore, at a descriptive level there were obviously more CD68+ and MAC387+ macrophages within the fibrous bands of cirrhotic tissue when compared with chronic hepatitis or normal liver tissue (fig1).


In fig 3, intrahepatic mRNA cytokine levels grouped according to histological classification are compared. The IL-18 mRNA expression was significantly upregulated in HCV associated chronic hepatitis (p=0.02) versus the controls (fig 3C). IL-18 mRNA upregulation was accompanied by a significant increase in IFNγ mRNA expression (chronic hepatitis, p=0.002; cirrhosis, p=0.006; fig 3D). The mRNAs of the three other inflammatory associated cytokines (IL-1β, MIP1β, and TNFα) were also upregulated in HCV induced chronic hepatitis versus normal controls (fig 3A, B, E); however, only TNFα reached statistical significance (p=0.02). In contrast, IL-6 mRNA expression was significantly downregulated in chronic hepatitis (p=0.04). In HCV cirrhotic liver tissue, the cytokine mRNAs of all these cytokines were upregulated in relation to the normal controls; however, only IL-1β, IL-18, and MIP1β reached statistical significance (p<0.05, p=0.006, p<0.05, respectively). In addition, it was observed that IL-1β, IL-6, and MIP1β mRNA upregulation was greater in cirrhotic liver tissue than in chronic hepatitis versus normal controls. Finally, no significant association (by linear regression analysis) was found between any cytokine mRNA level and the total Scheuer score or any of its components (data not shown). The only exception was a weak, but statistically significant negative association between the portal component of the Scheuer score and TNFα (p=0.04).

Figure 3

Cytokine expression (molecules per 107molecules of aldolase B) between hepatitis C patients with cirrhosis or chronic hepatitis versus normal controls. (A) IL-1β; (B) MIP1β; (C) IL-18; (D) IFNγ; (E) TNFα; (F) IL-6; p values denote the statistical difference between histological groups versus normal controls.


Relations between different liver cytokine mRNA levels (HCV positive patients only) were investigated by univariant regression analysis. Statistically significant positive correlations were found between IFNγ and four of the inflammation associated cytokines (IL-1β, IL-18, MIP1β, and TNFα) and between MIP1β and IL-1β (table 3). Interestingly, TNFα became positively correlated with IL-1β (p=0.004, n=10) and MIP1β (p=0.0002, n=10) when the cirrhotic patients were removed from the analysis.

Table 3

Comparison of individual cytokines in hepatitis C infected liver by regression analysis (n=15)


Finally, inflammatory cytokine mRNA levels and CD68 and MAC387 cell densities within the portal tract and lobular regions of HCV infected liver tissues were statistically assessed (table 4). Statistically significant positive correlations were found between CD68, MAC387, and IL-1β mRNA (p=0.025, p=0.003, respectively) and MIP1β mRNA (p=0.009, p=0.01, respectively). IL-6 mRNA was significantly correlated to the activated macrophage population (MAC387+) cell density. Although not significant, TNFα and IFNγ mRNAs showed a strong positive association with CD68 cell density (p=0.1, p=0.06, respectively) and likewise IL-18 mRNA with MAC387 cell density (p=0.09). In contrast, neither lobular CD68 nor MAC387 cell densities were significantly related to any of the cytokines examined.

Table 4

Comparison of immunostained cell density within the portal tract with cytokine mRNA levels in hepatitis C virus infected liver tissue by regression analysis (n=15)


Although we have previously documented the increased production of intrahepatic Th1 cytokines in chronic HCV,23 the role of non-specific inflammatory cells and non-specific inflammatory cytokines required further investigation. We have addressed this by two separate approaches: (1) measurement of CD68+ and MAC387+ cell densities within the portal tract by immunostaining; and (2) measurement of intrahepatic mRNA from non-specific proinflammatory cytokines.

There are several key findings from this study. Firstly, the non-specific cytokine IL-18 (well characterised IFNγ inducing factor49) and IFNγ mRNA expression were significantly correlated with each other and both were significantly upregulated in chronic hepatitis with or without cirrhosis. These data confirmed our earlier finding that IFNγ mRNA expression is upregulated in chronic hepatitis C23; likewise, Fukuda et al noted that increased liver injury was significantly associated with increased IFNγ mRNA load.50 In addition, the strong correlation we observed between the presumed macrophage produced IL-18 mRNA and the presumed T cell produced IFNγ mRNA might suggest that a positive feedback loop exists that is consistent with hepatitis C driving a Th1 type/cell mediated immune response within the liver. The relation between IL-18 and IFNγ mRNA has also been detected by Gaweco et al (abstract only) using a non-quantitative PCR method—that is, they found that intrahepatic IL-18 and IFNγ mRNA were both correlated and upregulated in 24 chronic HCV infected patients,51 and they confirmed upregulation of IFNγ protein expression by immunostaining. In addition, the positive correlation we found between IFNγ and other inflammatory cytokines (IL-1β, MIP1β, and TNFα) is consistent with a Th1/non-specific cell mediated immune response. Furthermore, support for a Th1 type intrahepatic response comes from the recent identification of CXCR3 expression on all liver infiltrating T cells in chronic HCV.52 The CXCR3 α chemokine receptor is thought to be important in selective recruitment of Th1 type cells to sites of tissue inflammation.

As expected, there was a significant increase in the density of portal tract CD68+ stained cells in chronic HCV, a phenomenon also observed by Tomita et al.35Intrahepatic CD68+ cells in chronic HCV most likely represent cells of the macrophage/Kupffer/monocyte cell lineage,40 although CD68 may also be present on other cell types such as neutrophil granulocytes and plasmacytoid T cells.53-55 Leicester et al(abstract) used CD14 as a marker for macrophage activation and found that the cell density of activated macrophages in relation to the total macrophage density was increased in HCV chronic hepatitis compared with normal tissue controls.33 We used the MAC387 antigen as a marker for (a subset of) infiltrating activated macrophages38 41 55 and we observed the infiltration of MAC387+ cells within the portal tracts. However, the number of MAC387+ cells was significantly less (approximately 45% of CD68 cell numbers); they were primarily sited at the limiting plate in patients with chronic hepatitis. In this study, we did not detect extensive MAC387 expression by Kupffer cells as previously reported by Flavell et al,38 and our results are therefore consistent with those of other studies.37 40-43 However, the potential of MAC387 to stain resident Kupffer cells and neutrophil granulocytes38 43 may potentially invalidate results, if care is not taken during the cell counting procedure.

The trophism of MAC387+ cells for the limiting plate is not unique, as cells positive for other potential macrophage markers (CD11b, CD11c, CD14, and CD16) have also been sited primarily at the periphery of HCV positive inflamed portal tracts.30 31This trophism may be partially cytokine driven: in a recent study, Narumi and colleagues56 found that in HCV infected liver tissue, IFN inducible protein 10 (IP-10) mRNA (α chemokine that attracts mononuclear cells and activated T cells57) was expressed by hepatocytes specifically in the region around periportal piecemeal necrosis. Furthermore, Shields et al noted increased expression of hepatic IP-10 expression in chronic HCV.52 It is unclear, however, whether these attracted periportal mononuclear cells have cytotoxic capabilities or whether they are present in this region as a reaction to the cellular damage, and therefore may play a phagocytic role. Interestingly, MAC387+ cells also line the outer margin of the portal tract in chronic HBV infection43 and like HCV infection, the exact function of these cells is not clear. However, there are several possible explanations—for example, evidence exists for both antibody dependent monocyte mediated cytotoxicity with neighbouring hepatocytes (periportal)58 and for an immunoregulatory role with neighbouring CD8+ suppressor/cytotoxic T cells.43

In this study, we are assuming that monoclonal antibodies to CD68 and MAC387 are staining the monocyte/macrophage population within the liver. Ideally, double immunostaining for CD68/MAC387 or if fresh tissue is available, macrophage isolation by collagenase digestion, Percoll density centrifugation followed by immunostaining/in situ hybridisation/RT-PCR, would not only allow more specific association of these markers to the macrophage population, but would also facilitate cytokine characterisation within this group. Unfortunately, in this study, these methods were unable to be performed and therefore, our results relating their expression to specific locations/cell types and to each other are limited.

In addition to IL-18, a wide range of other non-specific proinflammatory associated intrahepatic cytokines were examined. These cytokines may be produced by many cell types: for example, IL-1β by fibroblasts and epithelial cells59; TNFα by neutrophils, T cells, and NK cells60; and MIP1β by T cells, B cells, fibroblasts, and endothelial cells.61 However, the observed upregulation of these cytokines (except IL-6) in HCV infected liver tissue is consistent with macrophage activation. Furthermore, the levels of increase were marginal compared with the very significant levels of IL-18 and IFNγ—for example, only TNFα mRNA was significantly increased in non-cirrhotic HCV, but we have noted that these marginal increases maybe influenced by our controls. Specifically, histological examination of our control liver tissues revealed low level inflammation within the portal tracts, a probable consequence of being taken from cadaver donors (inadvertent macrophage activation during ischaemia and organ preservation). Consequently, some cytokine levels within these controls may be marginally upregulated and this may mask the true difference in cytokine expression between HCV infected and uninfected “normal” liver.

The monokine IL-6 warrants a special mention, because it was the only cytokine mRNA to be significantly downregulated in chronic hepatitis without cirrhosis; a similar observation was reported by Gawecoet al.26 Fukudaet al also observed that IL-6 mRNA expression was inversely proportional to the severity of liver injury.50 Additional indirect support for HCV associated intrahepatic IL-6 downregulation was the detection of C reactive protein (IL-6 upregulates its production) downregulation in chronic HCV liver injury.62 However, IL-6 downregulation in chronic hepatitis may only reflect a Th1 dominated intrahepatic microenviroment—that is, IL-6 maybe naturally suppressed because it is functionally anti-inflammatory and associated with Th2 cell activity; alternatively, HCV may directly effect IL-6 mRNA expression. Further investigation is required to address these issues—for example, the observed upregulation of IL-6 mRNA in cirrhosis may be due to its production by activated hepatic stellate cells in addition to other IL-6 sources such as Kupffer, mononuclear, Th2, and endothelial cells.63

Finally, two important associations were found between cytokine mRNA expression, and CD68 and MAC387 cell density within the portal tract. Firstly, a statistically significant association was found between MAC387+ cell density and the cytokines IL-1β, MIP1β, and IL-6. Secondly, a positive association (not significant) was noted between MAC387+ cell density and cytokines IL-18 (p=0.09) and IFNγ (p=0.1). To understand these observations better—that is, to clarify the exact role of intrahepatic macrophages in chronic HCV, requires the direct identification of the cellular sources of these cytokines. For example, the observed upregulation of intrahepatic cytokines may be due to an increase in the number of macrophages attracted to the liver (possibly IL-1β and MIP1β) and/or it may be due to an increase in the activation state of a subpopulation of intrahepatic macrophages (possibly IL-18). To explore this issue further, future studies could include in situ hybridisation and/or cellular isolation and FACS analysis for macrophage markers in addition to cytokine estimation.

To conclude, this study showed a close association between IL-18 and IFNγ mRNA production, which suggests that the intrahepatic immune response involves a partnership between macrophages and Th type 1 cells. In addition, intrahepatic cytokine mRNA production varied greatly from a notable upregulation in IL-18, to a strong increase in TNFα, to a weak upregulation in IL-1β and MIP1β, and to a downregulation in IL-6. Furthermore, some cytokine mRNA levels (MIP1β and IL-1β) correlated with MAC387+ and CD68+cell densities within the portal tracts. Finally, we showed a significant increase in intrahepatic macrophage numbers in persistent HCV infection. In summary, these data further support the persistent intrahepatic cell mediated immune response model in chronic HCV, but do not explain why HCV persists in the presence of this response.


The authors wish to acknowledge Ms Robyne Soper from the Anatomical Pathology Department, for her assistance in the preparation of paraffin wax sections for immunostaining; and Drs Alex Bishop and Mark Gorrell for their helpful discussions.



  • Abbreviations used in this paper:
    dot blot PCR
    cytotoxic T lymphocyte
    hepatitis C virus
    macrophage inflammatory protein
    natural killer cell
    polymerase chain reaction
    tumour growth factor
    tumour necrosis factor