Increased fat accumulation in the liver in HIV-infected patients with antiretroviral therapy-associated lipodystrophy
Reported by Jules Levin
"...Liver fat content has not been quantified previously in patients with HAART-associated lipodystrophy. We found liver fat content to be higher in the patients taking HAART and who had lipodystrophy than in patients taking HAART who did not have lipodystrophy or HIV-negative subjects. All patients were matched for age and weight..."|
The authors make the case that fat loss, not fat gain in viscera in belly, leads to fat in the liver and may be a causative factor in the development of insulin resistance in these patients. If that's true would reversing insulin resistance result in regain of fat in lipoatrophy in HIV? The authors also looked at leptin levels and discuss this at the end of the article. The relationship of leptin levels and it's association to body fat changes and it's potential use as a therapeutic intervention has been controversial. At the recent 4th Lipodystrophy Workshop leptin received attention and a number of researchers did not feel that leptin therapy would reverse fat loss, perhaps only help with lipid elevations.
editorial note: individuals coinfected with HIV and HCV or HBV appear more likely to have insulin resistance and diabetes than individuals who only have HIV as found in several small studies, most recently reported at the 4th Lipodystrophy Workshop by Hadigan, Grinspoon and Chung from Boston. It was previously reported that coinfected patients may also be more likely to develop body changes than individuals with only HIV. Taken together with the study findings discussed below, this may suggest that insulin resistance may be closely associated with lipoatrophy. This has been proposed by a number of researchers so it's not a new idea but these data may reinforce that idea. A recent small study found rosiglitazone can reverse lipoatrophy. This needs further larger studies.
Not addressed in this study discussed below are the contributions to fat in liver by HIV antiretroviral drug class or by specific drugs. Are protease inhibitors, NNRTIs or NRTIs contributing and to what extent. Are nukes contributing to the fat in the liver and is it by mitochondrial toxicity in the liver? The authors discuss at the end of this article the possibility of nukes contributing to fat in liver. Are the protease inhibitors contributing due to increased lipids? NRTIs can also raise lipids but not as much as a PI. And what is contribution from NNRTIs?
In HCV fatty liver can accelerate liver disease progression and may have a negative affect on HCV therapy outcome. We also see in recent pegylated inteferon plus ribavirin studies in coinfected patients that response rates are lower than in HCV monoinfected patients. So, are the reduced response rates to interferon plus ribavirin associated with fat in the liver? Researchers ought to examine this.
Objective: To determine liver fat content in patients with highly active antiretroviral therapy (HAART)-associated lipodystrophy.
Background: Lipodystrophy in several animal models is associated with fat accumulation in insulin-sensitive tissues, such as the liver.
This causes hyperinsulinaemia, dyslipidaemia and other features of insulin resistance.
Design: A cross-sectional study.
Subjects and methods:
Three age- and weight-matched groups were compared: 25 HIV-positive men with HAART-associated lipodystrophy (HAART+LD+), nine HIV-positive men receiving HAART, but without lipodystrophy (HAART+LD-), and 35 HIV-negative healthy men (HIV-). Liver fat content was measured using proton spectroscopy. Intra-abdominal and subcutaneous fat were determined using magnetic resonance imaging.
Liver fat content was significantly higher in the HAART+LD+ (8 ± 10%) than the HIV- (5 ± 7%; P < 0.05) or the HAART+LD- (3 ± 5%; P < 0.01) group. Liver fat content correlated with serum fasting insulin in the HAART+LD+ (r = 0.47; P < 0.05) and HIV- groups (r = 0.65; P < 0.001), but not with the amount of intra-abdominal fat. Within the HAART+LD+ group, serum insulin did not correlate with the amount of intra-abdominal fat. The HAART+LD+ group had a lower serum leptin concentration when compared to the two other groups. Features of insulin resistance, including hepatic fat accumulation, were not found in HAART+LD- group.
The severity of the insulin resistance syndrome in patients with HAART-associated lipodystrophy is related to the extent of fat accumulation in the liver rather than in the intra-abdominal region. Fat accumulation in the liver may therefore play a causative role in the development of insulin resistance in these patients.
Jussi Sutinen; Anna-Maija Hakkinena; Jukka Westerbacka; Anneli Seppala-Lindroos; Satu Vehkavaara; Juha Halavaara b; Asko Jarvinen c; Matti Ristola c; Hannele Yki-Jarvinen. From the Department of Medicine, Division of Diabetes, the aDepartment of Oncology, the bDepartment of Radiology, and the cDepartment of Medicine, Division of Infectious Diseases, Helsinki University Central Hospital, Helsinki, Finland.
AIDS 2002; 16(16):2183-2193
In mouse models of lipodystrophy, lack of subcutaneous fat is accompanied by fat deposition and insulin resistance in key insulin target tissues, i.e., the liver and skeletal muscle. In mouse models of lipodystrophy, lack of subcutaneous fat is accompanied by fat deposition and insulin resistance in key insulin target tissues, i.e., the liver and skeletal muscle. Retransplantation of fat subcutaneously reverses whole body insulin resistance and all insulin-signaling defects in insulin target tissues. Consistent with these data in mice, the features of insulin resistance increase in severity with the extent of fat loss in HIV-negative lipodystrophic patients. In non-diabetic subjects, fat accumulation in the liver and skeletal muscles has been shown to be much
more closely correlated with whole body insulin resistance than with any measure of obesity, including the amount of visceral fat. Taken together, these data imply that accumulation of fat in insulin sensitive tissues may cause
insulin resistance. Whether patients with HAART-associated lipodystrophy have increased amount of fat in the liver is unknown. The possibility that HAART causes accumulation of fat in the liver is supported by studies in mice, in which
ritonavir caused hepatic steatosis.
In the present study, we determined whether patients with HAART-associated lipodystrophy have an increased amount of fat in the liver when compared to healthy weight-matched subjects or to HIV-positive patients who have not
developed lipodystrophy during antiretroviral therapy. The latter group was studied to determine whether features of insulin resistance can be observed in patients using HAART in the absence of lipodystrophy. For this purpose, we
studied a total of 34 HIV-positive men receiving HAART and 35 healthy men, who were characterized with respect to various features of insulin resistance. Liver fat was quantified non-invasively using proton spectroscopy, and the
volumes of intra-abdominal and subcutaneous adipose tissue with magnetic resonance imaging (MRI).
Thirty-four HIV-positive men receiving antiretroviral therapy were recruited from the HIV-outpatient clinic of the Helsinki University Central Hospital. The men were subgrouped according to the presence or absence of lipodystrophy.
Lipodystrophy was defined as self-reported symptoms of subcutaneous fat loss with/ without increased girth or buffalo hump. Patients were included in the lipodystrophy group (HAART+LD+, n = 25) after the investigators confirmed the self-reported findings. HIV-positive patients without lipodystrophy were receiving HAART, but had no symptoms or signs of lipodystrophy (HAART+LD-,
n = 9). The HIV-negative healthy control group (HIV-) consisted of 35 age- and
weight-matched men who were recruited for the study from occupational health services in Helsinki. These men were healthy as judged by history and physical examination, and did not use any regular medication. None of the subjects in the HAART+LD+, the HAART+LD- or the HIV- group had serological evidence of hepatitis B or C, autoimmune hepatitis, clinical signs or symptoms of inborn errors of metabolism, or a history of use of toxins (other
than alcohol) or drugs associated with liver steatosis. In each subject, liver fat content was determined using proton spectroscopy and intra-abdominal and subcutaneous fat using MRI. In addition, various other measurements of body
composition were performed (vide infra), and a blood sample was withdrawn after an overnight fast for biochemical measurements as detailed below.
All HIV-positive patients were receiving a stable HAART regimen with no changes in antiretroviral medication during the 8 weeks before entering the study. The HIV-positive patients were currently receiving the following protease inhibitors: indinavir (32% of patients in the HAART+LD+ versus 22% in the HAART+ LD- group), nelfinavir (20% versus 11%), ritonavir (8% versus 11%), lopinavir (8% versus 0%), lopinavir + amprenavir (4% versus 0%), amprenavir (4% versus 0%), saquinavir (0% versus 11%). None of differences in the frequencies of currently used or ever used (data not shown) protease inhibitors were statistically significant between the HAART+LD+ and the HAART+LD- groups. The total duration of the use of indinavir was 32 ± 18 months in the HAART+LD+ and 33 ± 26 months in the HAART+LD- group, nelfinavir 23 ± 12 versus 21 months (only one patient in the HAART+LD- group had ever used nelfinavir), ritonavir 21 ± 21 versus 24 ± 32 months, saquinavir 11 ± 13 versus 31 ± 36 months; none of the these durations were significantly different between the groups. In addition, in the HAART+LD+ group two patients had used lopinavir for 6 ± 2 months, one amprenavir for 23 months and one patient a combination of amprenavir and lopinavir for 2 months. The duration of nucleoside analogue therapy was significantly longer in the HAART+LD+ than in the HAART+LD- group. The groups were comparable
with respect to the time since diagnosis of HIV, the most recent and nadir blood CD4 cell count, as well as HIV viral load (HIV-1 Monitor Test, Roche Diagnostics, Branchburg, New Jersey, USA). All patients were infected through
sexual contact. Alcohol consumption was quantified using a structured interview prior to participation.
Age, body mass index (BMI) and alcohol consumption were comparable between all groups. The total amount of fat in the abdominal region (by MRI) was comparable between the groups, but its distribution was different.
The HAART+LD+ group had significantly more intra-abdominal fat than the HAART+LD- or the HIV- group, and also significantly less subcutaneous fat than the HIV- group. The ratio of intra-abdominal to subcutaneous fat was
4.4-fold higher in the HAART+LD+ group (3.1 ± 2.7) than in the HAART+LD- group (0.7 ± 0.4; P < 0.001) and 6.2-fold higher than in the HIV- group (0.5 ± 0.2; P < 0.001). The waist : hip ratio was significantly higher in the HAART+LD+ group (1.00 ± 0.07) than in the HIV- (0.94 ± 0.05; P < 0.01) or the HAART+LD- (0.93 ± 0.09; P < 0.05) group.
Serum insulin and C-peptide concentrations were significantly higher in the HAART+LD+ group than either in the HAART+LD- or the HIV- group. Serum insulin concentrations did not correlate with the amount of
intra-abdominal fat within HAART+LD+ group (r = 0.26, not significant). The HAART+LD+ patients had significantly lower concentrations of serum HDL cholesterol and higher concentrations of serum triglycerides than the HAART+LD- patients or the HIV-negative subjects. Serum total cholesterol was significantly higher in the HAART+LD+ group (5.9 ± 1.1 mmol/l) than in the HAART+LD- group (4.8 ± 0.8 mmol/l; P < 0.01) or in the HIV- group (5.1 ± 1.0 mmol/l; P < 0.01). Serum ALT and AST concentrations were significantly higher in the HAART+LD+ group than in the HAART+LD- or the HIV- group, and serum GGT was significantly higher in the HAART+LD+ group than in the HIV-negative subjects. Blood lactate concentrations were similar in both HIV-positive groups and none of the patients had acidosis.
Liver fat content in the HAART+ LD+ patients (8.1 ± 9.9%) was 53% higher than in the normal subjects (5.3 ± 6.6%; P < 0.05) and 179% higher than in the HAART+ LD- patients (2.9 ± 4.7%; P < 0.01). Liver fat content did not differ between the HAART+LD- and HIV-negative groups (P = 0.31). Liver fat content was highly significantly correlated with fasting serum insulin concentration in the HIV-negative subjects (r, 0.65; P < 0.001) and in the HAART+LD+ group (r, 0.47; P < 0.05). Similar highly significant relationships were observed between liver fat and C-peptide concentrations.
The slopes of the regression lines relating liver fat and fasting insulin concentration were similar in HAART+LD+ and HIV- groups. The intercepts of the regression lines were, however, significantly different between the HAART+LD+ group and the HIV-negative subjects (P < 0.001) implying that for a given percentage of liver fat, serum fasting insulin concentrations were significantly higher in the HAART+LD+ group than in the HIV-negative subjects.
Liver fat did not correlate with the amount of intra-abdominal fat or waist : hip ratio in the HAART+LD+, HIV- group or the HAART+LD- group (data not shown).
Relationship between measures of body composition and serum leptin concentrations
The HAART+LD+ group had significantly lower leptin concentrations than the two other groups. Serum leptin concentrations were closely correlated with the amount of subcutaneous fat in both the HAART+LD+ and the HIV- groups. Within the groups of HAART+LD+ and HIV-, serum leptin was correlated with BMI, but the slopes of these relationships were different. For the same BMI above approximately 20 kg/m2, the HAART+LD+ group had a lower leptin concentration than the HIV- group.
Discussion by Authors
The correlation of serum fasting insulin with liver fat content, but not with the amount of intra-abdominal fat in HAART+LD+ patients challenges the idea that intra-abdominal fat, at least alone, is responsible for features of insulin
resistance in HAART-associated lipodystrophy. The only significant difference in the treatment history was a longer duration of nucleoside analogue therapy in the HAART+LD+ than in the HAART+LD- group.
Liver fat content has not been quantified previously in patients with HAART-associated lipodystrophy. We found liver fat content to be higher in the HAART+ LD+ patients than in the age- and weight-matched HIV-negative subjects or the HAART+LD- patients. Within the HAART+LD+ group, serum insulin concentration correlated with the percentage of liver fat but did not correlate with the amount of intra-abdominal fat. This finding suggests that fat accumulation in insulin sensitive tissues, such as the liver, rather than accumulation of intra-abdominal fat is critical for the development of features of insulin resistance in HAART-associated lipodystrophy.
Common non-invasive methods for evaluating liver fat content, such as ultrasound are not specific and are at best only semi-quantitative. In the present study, we used magnetic resonance proton spectroscopy, which allows
non-invasive quantification of liver fat without radiation exposure. Liver fat content measured using proton spectroscopy correlates closely with that determined histologically from liver biopsies and with liver density measurements calculated by computed tomography. Furthermore, by spectroscopic determination of liver fat a larger volume of liver
tissue can be evaluated than by liver biopsies. However, it is important to emphasize that quantification of liver fat with proton spectroscopy does not allow distinction between micro- and macrovesicular steatosis or evaluation of the presence or absence of fibrotic or inflammatory changes or mitochondrial abnormalities. Such information would be important for the understanding of the aetiology of steatohepatosis but could not be obtained in the present study because liver biopsies were not clinically indicated.
The increased amount of hepatic fat in the HAART+ LD+ group could be explained neither by alcohol consumption, which was comparable between the study groups, nor by autoimmune causes. Co-infection with hepatitis C virus has recently been shown to increase the risk for severe liver damage during HAART treatment. This association cannot account for our results as none of the subjects in this study were carriers of hepatitis C or B virus. Mitochondrial toxicity induced by nucleoside analogues has been suggested to cause lactic acidosis and hepatic steatosis in some patients using HAART. However, this mechanism for liver abnormalities is unlikely to explain our findings, as the lactate concentrations were similar between the HAART+LD+ and the HAART+LD- groups and none of the patients had acidosis. On the other hand, in the absence of liver biopsies, we cannot exclude the possibility that at least some patients had mitochondrial abnormalities typical of those described in patients with non-alcoholic steatohepatitis, who also are known to be insulin resistant.
LEPTIN: Although liver fat has not been quantified previously in patients with HAART-associated lipodystrophy, Ariogolu et al. found fatty infiltration of the liver in 12 out of 20 HIV-negative patients with various forms of lipodystrophy. In addition, scattered case reports have described hepatosplenomegaly in patients with various forms of acquired or congentital lipodystrophies and liver fat content was also increased in a patient with acquired generalized
lipoatrophy (Lawrence syndrome). These human data resemble those of genetically modified mice. Mice in which adipose tissue has been genetically ablated develop insulin resistant diabetes, hypertriglyceridaemia, fatty liver and have reduced levels of leptin.
In the present study, the HAART+LD+ group had significantly lower leptin concentrations than the other two groups. Serum leptin was similarly correlated with subctaneous fat in both HAART+LD+ and HIV- groups. Serum leptin concentrations also correlated with BMI significantly within both groups, but the slopes of the these relationships were significantly different. For a given BMI, approximately above 20 kg/m2, the HAART+LD+ group had a lower leptin concentration than the HIV- group. This result suggests that subcutaneous fat is the major, if not the exclusive source of circulating leptin, and also that the quantity of leptin produced per unit of subcutaneous fat mass is unaltered in HAART+LD+ patients. These data are consistent with
previous data demonstrating low plasma leptin concentrations in patients with HAART-associated lipodystrophy and with in vitro studies of human adipose tissue, which have shown that subcutaneous adipose tissue produces two- to
threefold more leptin than visceral fat. In the lipodystrophic SREBP-1c (sterol regulatory element binding protein-1c) overexpressing mouse, a low dose leptin infusion completely normalizes insulin sensitivity and depletes the
liver of its massive fat deposits. In another lipodystrophic mouse model characterized by severely reduced peroxisome proliferator activator receptor-[gamma] activity, leptin alone has been insufficient to completely normalize
insulin sensitivity. In the latter model, a combination of leptin and adiponectin, which is another hormone produced exclusively by adipose tissue, completely reversed insulin resistance and normalized the fat content of the liver. It is not
known whether similar manoeuvres would be helpful in HAART-associated lipodystrophy.
For a given amount of liver fat, serum-free insulin concentrations were higher in the HAART+LD+ than the HIV- group implying that liver fat alone was insufficient to explain all of the variation in serum insulin concentrations.
cannot be deposited only in the liver but also intramyocellularly in skeletal muscles in humans. Intramyocellular fat correlates negatively with insulin sensitivity in the muscle. It is thus possible that insulin resistance in
skeletal muscles might explain some of the variation in fasting insulin concentrations. As we did not perform euglycemic insulin clamp studies combined with infusion of glucose tracers in the present study, it is not possible to determine the contribution of individual tissues to the increase in fasting insulin concentrations. On the other hand, after an overnight
fast, the liver rather than skeletal muscle is the key target for insulin action [49, 50]. At least in mice, selective deletion of the insulin receptor from skeletal muscle does not lead to hyperinsulinaemia or abnormal glucose tolerance, while tissue-specific deletion of the insulin receptor from the liver induces dramatic fasting and post-prandial hyperinsulinaemia and severe glucose intolerance as well as fatty degeneration of the liver. An alternative potential explanation for the disproportionate hyperinsulinaemia is that liver fat content is not a perfect measure of hepatic insulin sensitivity. Information of the rates of free fatty acid (FFA) influx into the liver and their metabolism, or of other processes possibly interfering with insulin action in the liver, would be helpful in this regard