Antibodies to tumor necrosis factor-a inhibit liver regeneration after partial hepatectomy PAUL AKERMAN, PIERA COTE, SHI &I YANG, CRAIG McCLAIN, STEVE NELSON, GREGORY J. BAGBY, AND ANNA MAE DIEHL Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205; Department of Medicine, University of Kentucky, Lexington, Kentucky 40506; and Department of Medicine, Louisiana State University School of Medicine, New Orleans, Louisiana 70118 Akerman, Paul, Piera Cote, Shi Qi Yang, Craig McClain, Steve Nelson, Gregory J. Bagby, and Anna Mae Diehl. Antibodies to tumor necrosis factor-a inhibit liver regeneration after partial hepatectomy. Am. J. Physiol. 263 (Gastrointest. Liver Physiol. 26): G579-G585, 1992.-Certain cytokines that are produced in liver may act as growth factors to facilitate wound healing and, hence, may influence liver regeneration. However, this hypothesis has not been directly tested. To determine whether the cytokine response evoked by partial hepatectomy (PH) modulates the process of liver regeneration, adult male rats were injected intraperitoneally with either goat polyclonal antibodies to rat tumor necrosis factor (TNF; 15 pg/g body wt) or an equal amount of goat anti-rat immunoglobulin G 1 h before PH. Animals were killed at 12, 24, 48, or 72 h postPH, 1 h after injection with [“Hlthymidine. Serum TNF levels were measured with the L929 cytotoxicity assay, titers of antibody to TNF were determined by enzyme-linked immunoabsorbent assay, and interleukin-6 (IL-6) concentrations were measured by B9 cell bioassay. Liver regeneration was assessed by [“Hlthymidine incorporation into hepatic DNA and by immunohistochemical evidence of proliferating cell nuclear antigen (PCNA) expression. Antibodies to TNF were detected in treated rats but not in controls. Titers were highest at 12 h and progressively fell. Although TNF was never detected in serum, treatment with anti-TNF pre-PH significantly inhibited increases in serum IL-6 concentration post-PH. Anti-TNF pretreatment also inhibited [3H] thymidine incorporation into DNA, as well as expression of PCNA by both hepatocytes and liver nonparenchymal cells. These data indicate that TNF positively modulates liver regeneration after PH. cytokine;
interleukin-6
is an important producer of tumor necrosis factor-a (TNF-cu), particularly after exposure to gutderived endotoxin (16). TNF-a, in turn, prompts the production of several other cytokines, including interleukin-1 (IL-l) and interleukin-6 (IL-6) (11). The liver also appears to be the organ principally responsible for the clearance of many circulating cytokines (12, 25). Indeed, the rapid hepatic clearance of circulating cytokines may constitute an important mechanism to limit their systemic toxicity (3). The balance between hepatic cytokine synthesis and cytokine “clearance” is subverted by portal systemic shunting and impairments in reticuloendothelial cell function that often occur in advanced liver disease. Indeed, it has been suggested that TNF-cu mediates several deleterious consequences of advanced liver disease, including fever, anorexia, muscle wasting, and hypoalbuminemia (18). Although increasing evidence incriminates cytokines in the pathophysiology associated with advanced hepatic dysfunction, they are not always detrimental to the liver. Some evidence suggests that cytokines may enhance the hepatic regenerative response. We have noted
THE LIVER
0193~1857/92
$2.00
Copyright
that significant increases in serum concentrations of IL-6 occur 24-48 h after partial hepatectomy (PH) (20). Cornell et al. (15) have reported that liver regeneration after PH is significantly delayed in germ-free, athymic, and lipopolysaccharide (LPS) -resistant mice, strains that may have a limited ability to release cytokines in response to gut-derived endotoxin. TNF-a administration increases DNA polymerase-cr activity as well as the incorporation of [3H]thymidine into DNA in the livers of intact adult rats (24). In the latter model, relatively large doses (25-50 pg TNF-a) of recombinant human TNF-cu were injected intravenously or intramuscularly, and immunohistochemical analysis indicated that TNFa-induced proliferation was restricted to Kupffer cells (23). However, hepatocytes rarely proliferate in healthy adult liver (32); hence an effect of TNF-a on hepatocyte DNA synthesis may have escaped detection. Furthermore, TNF-cu was administered systemically and therefore may not have had the same effect as TNF-a that is released locally in response to liver injury and/or portally delivered endotoxin. To clarify the physiological role of TNF-a in the compensatory hepatic regeneration that is triggered by PH, we treated adult rats 1 h before PH with a speciesspecific antibody to TNF-cu. Several parameters of liver regeneration in anti-TNF-a-treated rats were compared with similar endpoints in rats treated with control antirat immunoglobulin G (IgG). MATERIALS
AND METHODS
Materids. All chemicals used were purchased from Sigma (St. Louis, MO), with the following exceptions. Phenol was purchased from Fluka (Ronkonkoma, NY). [“Hlthymidine (sp act 20 Ci/mmol) came from Amersham (Arlington Heights, IL). Reagents for determination of hepatic protein content were purchased from Bio-Rad (Richmond, CA). Monoclonal antibody to proliferating cell nuclear antigen (PCNA) was obtained from Boehringer Mannheim (Indianapolis, IN). Recombinant murine TNF-a was a gift from Genentech (San Francisco, CA). As previously reported (5; 33), polyclonal antibody against this recombinant protein was produced in a goat using the Ribi adjuvant system containing 0.5 mg each of monophosphoryl lipid A, trehalose dimycolate, and cell wall skeleton in 0.2% Tween 80 (Ribi Immunochem Research, Hamilton, MT). The serum IgG fraction was obtained by polyethylene glycol 4000 precipitation and column chromatography with DEAE Bio-Gel A (Bio-Rad). The neutralizing capacity of the anti-TNF-cu IgG fraction was determined by mixing equal volumes of murine recombinant TNF-cu (400 U/ml) with serial dilutions of the anti-TNF-cu IgG. This mixture was incubated for 1 h at 37°C and then tested for residual TNF-cu activity. Under these conditions, the antibody was determined to contain 6.5 and 9.0 x lo5 50% neutralizing units/mg of IgG protein against murine
0 1992 the American
Physiological
Society
G579
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.130.211.199) on July 29, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
G580
CYTOKINE
REGULATION
recombinant TNF-cu and TNF-a-containing rat serum, respectively. Normal goat IgG prepared in the same way had no detectable TNF-cu neutralizing activity. The binding properties of anti-TNF-ar IgG and normal goat IgG were also tested in microtiter plates precoated with LPS or murine recombinant IL-l, interferon-y, or TNF-cu. Neither IgG preparation demonstrated effective binding to LPS, IL-l, or interferon-y. Only the antiTNF-ar IgG bound TNF-cu (5, 33). Methods. Seventy-two adult male Sprague-Dawley rats (mean wt ~300 g) were housed with a 12:12 light-dark cycle and permitted ad libitum consumption of a standard rat pellet diet. After a I-wk equilibration period, the animals underwent a 70% PH with light ether anesthesia in midmorning (27). PH has been shown to substantially increase the proportion of hepatocytes in S phase and is a standard technique used to assess the response to various effecters of regeneration (32). Sham-operated controls were not included in these experiments because it has been shown that the effects of PH are not reproduced by sham laparotomy and liver manipulation (19). One hour before PH, half the rats (the anti-TNF-a-treated group) were injected intraperitoneally with goat polyclonal antibodies to rat TNF-cr (15 pg/g body wt). The remaining rats (the controls) were similarly treated with an identical amount of goat anti-rat IgG. To estimate DNA synthesis, animals that were killed at 12, 24, 48, or 72 h post-PH were injected intraperitoneally with [3H]thymidine (0.05 &i/g body wt) 1 h before they were killed. Rats were killed by exsanguination at various times (0, 12, 24, 48, or 72 h) post-PH. The resected livers were immediately weighed. Aliquots of liver tissue were fixed in methanol for subsequent immunohistochemical analysis. The remaining liver tissues were freeze clamped in liquid nitrogen. Sera and liver tissue from each rat were saved at -70°C for subsequent analysis. DNA synthesis was estimated by two techniques: [“HIthymidine incorporation into hepatic DNA and nuclear expression of PCNA, a marker of cellular proliferation. To determine [“HIthymidine incorporation into liver DNA during the hour before death, liver tissues (0.2 g) were homogenized in buffer containing RNAases and proteases to degrade contaminating RNA and protein. DNA was extracted from this solution with phenolchloroform as described by Blin and Stafford (8). The ethanolprecipitated DNA was resuspended in TAE buffer (pH 8), and the concentration of DNA in an aliquot was estimated by the relative absorbance at 260-to-280-nm wavelength (8). The radioactivity of a similar volume was measured and expressed as counts per minute per microgram hepatic DNA. Although the incorporation of [3H]thymidine into DNA has been shown to be an accurate measure of DNA synthesis in liver regeneration (2 l), immunocytochemical analysis was done for additional confirmation. Methanol-fixed tissues were embedded in paraffin and sectioned. Glass-mounted sections were stained with an antibody to PCNA (29). The antibody was used at a final dilution of l:200. Because the prevalence of liver cells in S and/or M phase in adult liver has been documented to be
interleukin-6
is an important producer of tumor necrosis factor-a (TNF-cu), particularly after exposure to gutderived endotoxin (16). TNF-a, in turn, prompts the production of several other cytokines, including interleukin-1 (IL-l) and interleukin-6 (IL-6) (11). The liver also appears to be the organ principally responsible for the clearance of many circulating cytokines (12, 25). Indeed, the rapid hepatic clearance of circulating cytokines may constitute an important mechanism to limit their systemic toxicity (3). The balance between hepatic cytokine synthesis and cytokine “clearance” is subverted by portal systemic shunting and impairments in reticuloendothelial cell function that often occur in advanced liver disease. Indeed, it has been suggested that TNF-cu mediates several deleterious consequences of advanced liver disease, including fever, anorexia, muscle wasting, and hypoalbuminemia (18). Although increasing evidence incriminates cytokines in the pathophysiology associated with advanced hepatic dysfunction, they are not always detrimental to the liver. Some evidence suggests that cytokines may enhance the hepatic regenerative response. We have noted
THE LIVER
0193~1857/92
$2.00
Copyright
that significant increases in serum concentrations of IL-6 occur 24-48 h after partial hepatectomy (PH) (20). Cornell et al. (15) have reported that liver regeneration after PH is significantly delayed in germ-free, athymic, and lipopolysaccharide (LPS) -resistant mice, strains that may have a limited ability to release cytokines in response to gut-derived endotoxin. TNF-a administration increases DNA polymerase-cr activity as well as the incorporation of [3H]thymidine into DNA in the livers of intact adult rats (24). In the latter model, relatively large doses (25-50 pg TNF-a) of recombinant human TNF-cu were injected intravenously or intramuscularly, and immunohistochemical analysis indicated that TNFa-induced proliferation was restricted to Kupffer cells (23). However, hepatocytes rarely proliferate in healthy adult liver (32); hence an effect of TNF-a on hepatocyte DNA synthesis may have escaped detection. Furthermore, TNF-cu was administered systemically and therefore may not have had the same effect as TNF-a that is released locally in response to liver injury and/or portally delivered endotoxin. To clarify the physiological role of TNF-a in the compensatory hepatic regeneration that is triggered by PH, we treated adult rats 1 h before PH with a speciesspecific antibody to TNF-cu. Several parameters of liver regeneration in anti-TNF-a-treated rats were compared with similar endpoints in rats treated with control antirat immunoglobulin G (IgG). MATERIALS
AND METHODS
Materids. All chemicals used were purchased from Sigma (St. Louis, MO), with the following exceptions. Phenol was purchased from Fluka (Ronkonkoma, NY). [“Hlthymidine (sp act 20 Ci/mmol) came from Amersham (Arlington Heights, IL). Reagents for determination of hepatic protein content were purchased from Bio-Rad (Richmond, CA). Monoclonal antibody to proliferating cell nuclear antigen (PCNA) was obtained from Boehringer Mannheim (Indianapolis, IN). Recombinant murine TNF-a was a gift from Genentech (San Francisco, CA). As previously reported (5; 33), polyclonal antibody against this recombinant protein was produced in a goat using the Ribi adjuvant system containing 0.5 mg each of monophosphoryl lipid A, trehalose dimycolate, and cell wall skeleton in 0.2% Tween 80 (Ribi Immunochem Research, Hamilton, MT). The serum IgG fraction was obtained by polyethylene glycol 4000 precipitation and column chromatography with DEAE Bio-Gel A (Bio-Rad). The neutralizing capacity of the anti-TNF-cu IgG fraction was determined by mixing equal volumes of murine recombinant TNF-cu (400 U/ml) with serial dilutions of the anti-TNF-cu IgG. This mixture was incubated for 1 h at 37°C and then tested for residual TNF-cu activity. Under these conditions, the antibody was determined to contain 6.5 and 9.0 x lo5 50% neutralizing units/mg of IgG protein against murine
0 1992 the American
Physiological
Society
G579
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (130.130.211.199) on July 29, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
G580
CYTOKINE
REGULATION
recombinant TNF-cu and TNF-a-containing rat serum, respectively. Normal goat IgG prepared in the same way had no detectable TNF-cu neutralizing activity. The binding properties of anti-TNF-ar IgG and normal goat IgG were also tested in microtiter plates precoated with LPS or murine recombinant IL-l, interferon-y, or TNF-cu. Neither IgG preparation demonstrated effective binding to LPS, IL-l, or interferon-y. Only the antiTNF-ar IgG bound TNF-cu (5, 33). Methods. Seventy-two adult male Sprague-Dawley rats (mean wt ~300 g) were housed with a 12:12 light-dark cycle and permitted ad libitum consumption of a standard rat pellet diet. After a I-wk equilibration period, the animals underwent a 70% PH with light ether anesthesia in midmorning (27). PH has been shown to substantially increase the proportion of hepatocytes in S phase and is a standard technique used to assess the response to various effecters of regeneration (32). Sham-operated controls were not included in these experiments because it has been shown that the effects of PH are not reproduced by sham laparotomy and liver manipulation (19). One hour before PH, half the rats (the anti-TNF-a-treated group) were injected intraperitoneally with goat polyclonal antibodies to rat TNF-cr (15 pg/g body wt). The remaining rats (the controls) were similarly treated with an identical amount of goat anti-rat IgG. To estimate DNA synthesis, animals that were killed at 12, 24, 48, or 72 h post-PH were injected intraperitoneally with [3H]thymidine (0.05 &i/g body wt) 1 h before they were killed. Rats were killed by exsanguination at various times (0, 12, 24, 48, or 72 h) post-PH. The resected livers were immediately weighed. Aliquots of liver tissue were fixed in methanol for subsequent immunohistochemical analysis. The remaining liver tissues were freeze clamped in liquid nitrogen. Sera and liver tissue from each rat were saved at -70°C for subsequent analysis. DNA synthesis was estimated by two techniques: [“HIthymidine incorporation into hepatic DNA and nuclear expression of PCNA, a marker of cellular proliferation. To determine [“HIthymidine incorporation into liver DNA during the hour before death, liver tissues (0.2 g) were homogenized in buffer containing RNAases and proteases to degrade contaminating RNA and protein. DNA was extracted from this solution with phenolchloroform as described by Blin and Stafford (8). The ethanolprecipitated DNA was resuspended in TAE buffer (pH 8), and the concentration of DNA in an aliquot was estimated by the relative absorbance at 260-to-280-nm wavelength (8). The radioactivity of a similar volume was measured and expressed as counts per minute per microgram hepatic DNA. Although the incorporation of [3H]thymidine into DNA has been shown to be an accurate measure of DNA synthesis in liver regeneration (2 l), immunocytochemical analysis was done for additional confirmation. Methanol-fixed tissues were embedded in paraffin and sectioned. Glass-mounted sections were stained with an antibody to PCNA (29). The antibody was used at a final dilution of l:200. Because the prevalence of liver cells in S and/or M phase in adult liver has been documented to be
