What causes body fat and metabolic changes during HAART?

Researchers are no nearer to identifying the causes

of body fat changes during HAART after the First International Workshop on

Adverse Drug Reactions and Lipodystrophy in HIV last month in San

Diego.

The meeting, organised to review data and develop a

consensus definition of the syndrome which has been labelled lipodystrophy,

heard several presentations which suggest that the causes of body fat changes

may be multi-factorial and even more complex than originally suggested. The

meeting heard little about the potential reversal or prevention of

lipodystrophy, largely because the syndrome is poorly understood. Current

discussion of the causes and treatment of lipodystrophy resembles the early

years of AIDS research.

For a detailed description of the syndrome and

background on some of the topics discussed in this report see

href="http://www.aidsmap.com/article.asp?articleID=1661&heading3=Body+fat+changes+on+HAART+%">http://www.aidsmap.com/article.asp?articleID=1661&heading3=Body+fat+changes+on+HAART+%

One purpose of the meeting was to draw up

common criteria to define what sort of body fat changes and lipid changes will

be defined as significant for the purposes of research. It is becoming apparent,

noted several speakers, that we may be dealing with several different syndromes

with different causes and different prospects of reversal, rather than a

continuum of lipid changes with one cause and one treatment. Dr Ah Garg, an

expert on lipodystrophy from the days when 200 hereditary cases was considered

`a lot', described many different forms (see below), and Dr Matthew Liang of

Harvard Medical School described the enormous complexity of defining the many

distinct rheumatic syndromes.

The Workshop also tried to introduce a little more

precision into the terms being used to describe various body fat changes.

Strictly speaking, the term lipodystrophy should be used to describe central

visceral fat accumulation. Fat loss from the arms, legs, face or buttocks should

be defined as lipoatrophy. As became clear from the cohort study reports at the

meeting, the distinction is important because some people develop one form of

body fat change but not the other, whilst others display both to some degree. If

these changes have different causes, the search for ways of preventing or

reversing the changes will be assisted by precise reporting.

Triglyceride and cholesterol

changes

Protease inhibitors may directly interfere with the

functioning of low density lipoprotein receptor-related protein. This protein

clears triglyceride-rich low density lipoprotein from the blood. Inhibition of

this protein might explain the increase in triglycerides seen in many people

after commencing PI therapy.

The current working hypothesis published by Andrew

Carr and David Cooper, suggests that PIs are responsible for the syndrome

because they block the binding of all-trans retinoic acid and hence its

subsequent metabolism into 9-cis retinoic acid, which is essential for the

production of fat cells and the proper storage of lipids. By interfering with

this process, protease inhibitors block the differentiation and apoptosis of

adipocytes (fat cells).

Stephen Sturley of Columbia University reported that

the active site of HIV protease bears a close resemblance to a string of

amino-acids on the low density lipoprotein receptor-related protein (abstract

2). A reduction in LRP activity results in a reduction in lipoprotein clearance,

and two to three fold elevations in triglycerides and LDL cholesterol, similar

to those seen in people taking protease inhibitors. Dr Sturley suggested that

high baseline LDL cholesterol levels would result in lipoprotein receptors which

are already saturated, and less able to cope with any triglyceride elevation

after commencing PI therapy, and indeed, high baseline triglyceride levels have

been associated with the greatest triglyceride increases in several cohort

studies.

However, a team from Agouron Pharmaceuticals with

considerable expertise in crystallising HIV structures compared the crystal

structures of HIV protease inhibitors with that of CRABP1, and found little

potential for binding between protease inhibitors and CRABP1 (cis-retinoic acid

binding protein) (abstract 29). Instead, the team found that protease inhibitors

directly inhibit adipocyte differentiation by mechanisms unknown, with indinavir

having the strongest effect and amprenavir the weakest.

Furthermore, the Carr/Cooper hypothesis does not

explain why triglycerides can become elevated in individuals who take PI-sparing

regimens containing some nucleosides and those containing efavirenz.

Unfortunately, no lipids data from the 006 study of efavirenz were presented at

this meeting, even though Du Pont was able to offer 72 week data on the

incidence of body fat changes (abstract 39). Incidence of body fat changes was

remarkably small in all three arms of the study, including the indinavir arm,

and does not match the high incidence reported from much larger cohorts at the

Lipodystrophy Workshop (abstracts 12, 14-23).

A cross-sectional analysis of lipid elevations in 101

patients taking either efavirenz or nevirapine and two NRTIs was presented by Dr

Graeme Moyle of London's Chelsea and Westminster Hospital. Non-fasting

cholesterol samples were elevated in 32% of efavirenz recipients and 20% of

nevirapine recipients. Triglyceride levels were elevated in 6% of efavirenz

recipients and 2% of nevirapine rceipients. Those who had received prior PI

treatment were more likely to have elevated lipids on NNRTI treatment (abstract

54).

Carl Grunfeld of Veterans Medical Centre, San

Francisco, described the lipid changes seen in HIV infection prior to the advent

of HAART. He noted that serum triglyceride levels were high in people diagnosed

with AIDS, and that these high levels corresponded with high levels of

circulating interferon-alpha, and the clearance of triglycerides was slowed

after a high fat meal. Levels of HDL cholesterol declined to levels associated

with increased cardiovascular risk, and taken together these trends are a

typical response to acute infection.

AZT monotherapy significantly reduces triglyceride

levels and interferon levels, yet triglyceride increases have been reported in a

moderate proportion of individuals taking some dual nucleoside combinations.

These increases are modest in comparison to those seen in people taking protease

inhibitors, noted Dr Grunfeld, who went on to discuss the risk of cardiovascular

disease as a consequence of these lipid changes. These risks have been analysed

at several previous meetings and are discussed in more detail in

href="http://www.aidsmap.com/article.asp?articleID=1728&heading3=Heart+disease+and+HAART">http://www.aidsmap.com/article.asp?articleID=1728&heading3=Heart+disease+and+HAART

Although a number of reports of heart attacks and

angina putatively associated with PI treatment have appeared in the past year,

no group has yet investigated the physical condition of the arteries of people

taking PIs. At the Lipodystrophy Workshop a group from Italy reported on 13

HIV-positive patients with a median of 14 months PI experience who underwent a

variety of tests to detect early signs of atherosclerotic plaques and altered

blood flow in the common and internal carotid arteries. Six out of 13 had some

arterial dysfunction, a higher proportion than expected, although age-matched

controls not taking PIs and HIV-negative controls were not used in this study.

Four patients had thickening of the intima (commonly referred to as `furring up'

of the arteries , whilst several patients had plaques (large unstable

accumulations and blood and fat cells which if ruptured can cause thrombosis and

blockage of the artery) (abstract 42). A larger scale prospective study of this

type with properly matched controls would have considerable value in assessing

the cardiovascular significance of lipid changes on HAART.

Insulin resistance

Insulin resistance has also been widely reported in

people taking protease inhibitors, although the incidence varies between

cohorts.

Insulin resistance may be linked to the quantity of

central abdominal fat. The quantity of central abdominal fat is directly

correlated to insulin resistance in muscle tissue, but insulin resistance occurs

before central fat accumulation. How one triggers the other is still unclear,

reported Don Chisholm of St Vincent's Hospital, Sydney, but disturbances of

lipid metabolism are known to contribute to insulin resistance. Indeed, diabetes

in HIV-negative people may be treated with a group of drugs called

thiazolidinediones which work by encouraging the storage of fatty acids in

peripheral fat tissue rather than central fat tissue. This leads to a reduction

in insulin resistance. If fatty acids are being stored in central adipose tissue

and storage in peripheral tissue declines, insulin resistance would be a logical

outcome (although it will also be dependent on baseline sensitivity and diet to

some extent).

Not everyone who takes a protease inhibitor or other

anti-retroviral therapy develops insulin resistance. Dr Frank Goebel of Munich's

Ludwig-Maximilian University reported on changes in insulin sensitivity in a

cohort of patients studied in 1998. 36 untreated patients, 25 patients taking

NRTIs alone, 5 taking NRTIs and an NNRTI and 171 taking at least two NRTIs and

PI were compared for insulin resistance (abstract 5). Sensitivity was highest in

the untreated patients and lowest in those taking a protease inhibitor. 55% of

those taking a PI exhibited some degree of insulin resistance, with the highest

level of resistance occurring in those taking indinavir. 27% of those taking an

NRTI developed insulin resistance. There was no difference in insulin

sensitivity between those taking AZT/3TC and those taking d4T/ddI as part of

their combination. Muscle biopsies revealed that people on anti-retroviral

therapy who developed insulin resistance had fewer insulin receptors than

uninfected type II diabetics!

A team from the University of Southern California

School of Medicine (Bube, abstract 028) demonstrated that insulin sensitivity

declined by 30% in ten individuals within eight weeks of commencing indinavir

therapy, and that detectable changes in pancreatic beta cells also occurred

within eight weeks. Pancreatic beta cells regulate insulin secretion, which in

turns modulates blood levels of glucose and the development of hyperglycemia.

Within eight weeks of commencing indinavir, beta cell production of insulin had

declined when tested by infusing glucose into an individual's bloodstream.

Several studies in HIV-negative people have shown

that a small reduction in glucose tolerance is associated with a twofold

increase in the risk of cardiovascular disease. However, research in diabetics

and other groups has demonstrated that exercise alone can restore insulin

sensitivity. Don Chisholm cited two studies, in Australian aboriginals and

Japanese sumo wrestlers. The Australian aboriginal population has a very high

incidence of diabetes, whilst retired sumo wrestlers have the highest incidence

of type II diabetes and heart attack in the Japanese population. In both groups

low levels of physical activity have been blamed. In the Australian study,

aboriginals with impaired glucose tolerance who went back to hunter/gatherer

lifestyles lost 1kg a week and reverted to almost normal glucose tolerance over

the course of several months. In the Japanese study, sub-cutaneous and

intra-abdominal fat of active and retired sumo wrestlers were compared. Despite

consuming more than 7,000 calories a day, the active wrestlers had little

intra-abdominal fat and near-normal insulin sensitivity. Their vast bulk was

comprised almost entirely of sub-cutaneous fat. Retired wrestlers had very large

amounts of intra-abdominal fat and impaired glucose tolerance.

How exactly exercise improves insulin sensitivity was

not discussed, and no data was presented at the Workshop regarding the impact of

exercise on insulin sensitivity in HIV infection. Several speakers noted that

exercise has been shown to have a greater impact on sub-cutaneous fat than on

visceral fat in studies of obesity.

Rare forms of lipodystrophy are inheritable, and Dr

Ambhimanyy Garg noted that lipodystrophy is associated with insulin resistance

in some of these syndromes. He also pointed out that many of the defects

proposed as mechanisms in HIV-associated lipodystophy have already been

investigated in cases of inherited and/or genetic lipodystrophy to estasblish

whether particular genetic mutations are associated with the syndromes. In the

case of familial partial lipodystrophy (Dunnigans type), the gene associated

with the syndrome is located very close to CRABP-2, at chromosome IQ21. CRABP-2

is the lipoprotein receptor-related protein which bears a close resemblence to

HIV protease.

Acquired forms of lipodystrophy which resemble those

reported in HIV have also been reported in uninfected individuals, and may be

associated with an auto-immune response. The syndrome is 2.6 times more common

in women than in men, conforming to the prevalence of auto-immune diseases in

the population. In 90% of cases auto-antibody to C3 nephritic factor has been

detected. C3neF lyses adipocytes in vitro. Anything which interferes with

adipocyte lipolysis is likely to interfere with peripheral and central fat

storage.

There is controversy in the field of lipidology as to

whether adipocytes can be regenerated if they are lost after adolescence. If

some factor in HIV infection or HIV treatment is destroying adipocytes, a

pessimistic interpretation is that fat loss, or lipoatrophy, is irreversible.

Whilst anecdotal reports have disputed this, none of the cohort or switching

studies presented at the San Diego meeting reported substantial reversal of

lipoatrophy except for a French group. Dr Thierry St Marc and colleagues at the

Hopital Herriot in Lyons discontinued d4T (stavudine) therapy in a controlled

non-randomised study in which 14 individuals with lipoatrophy on stable NRTI

therapy discontinued d4T, and 15 individuals with lipoatrophy on stable PI-based

therapy also discontinued. These individuals were compared with 15 and 16

individuals respectively on stable therapy with no evidence of body fat changes.

Subcutaneous fat in the mid-thigh and abdomen improved significantly after 6

months, although it was not restored to the levels seen in those without

lipoatrophy.

Findings of this sort have led some to suggest that

nucleoside analogues, and d4T in particular, may contribute to the development

of body fat changes.

Mitochondrial toxicity

Thomas Kakuda of University of Minnesota Antiviral

Pharmacology Laboratory has proposed that the inhibition of mitochondrial

polymerase gamma by NRTIs may eventually lead to the disruption of lipid

storage, transport and glucose uptake by adipocytes. Mitochondrial DNA is easily

damaged, and cannot be repaired if polymerase gamma is inhibited. As

mitochondrial DNA becomes more damaged, oxidative phosphorylation declines,

leading to cell damage, toxicity and cell death.

If mitochondrial DNA in central adipocytes is

impaired, it is suggested that intracellular lipids could accumulate. In

peripheral sub-cutaneous fat, Kakuda speculates that mitochondrial toxicity

could lead to apoptosis of adipocytes, leading to fat loss in the legs, arms,

buttocks and face. Whilst at first sight this may appear contradictory (like so

many other aspects of the syndromes), Kakuda and Kees Brinkman of the University

of Amsterdam both argue that different NRTIs are likely to have different

effects in different tissues, depending on the distribution and penetration of

NRTIs into particular cells, the distribution of adipocytes and the type of fat

in different locations, the cellular capacity to phosphorylate NRTIs, and the

extent to which various NRTIs are incorporated by mitochondrial polymerase

gamma.

Differential effects of protease inhibitors on

adipocytes were also reported at the Workshop by Jim Lenhard of Glaxo Wellcome,

based on experiments using human adipocytes in the laboratory and mouse models

(abstract 1).

In vitro, indinavir has been shown to block adipocyte

differentiation in the presence of retinoic acid and a retinoic acid receptor

antagonist may prevent the negative effects of indinavir on adipogenesis.

However, the effects of other protease inhibitors seem far more contradictory,

and any marketing statements about the relative effects of different PIs on body

fat changes must be treated with scepticism at this time.

For example, whilst indinavir blocks adipogenesis in

vitro, indinavir treatment was associated with 48% increase in epididymal fat in

a mouse model when fed on a low fat diet. There was no change in fat levels when

mice receiving indinavir received a high fat diet, although the researchers did

not demonstrate how they could tell that indinavir was being absorbed in these

animals (fat interferes with the absorption of indinavir in the human digestive

system). However, a similar phenomenon was seen in nelfinavir-treated mice.

Conversely, triglycerides rose in mice treated with a high fat diet and either

nelfinavir or saquinavir, but remained stable in those which received a low fat

diet. A low fat diet on any drug was associated with an increase in serum

glucose.

New adverse events on HAART

A number of new phenomena were reported at the

Workshop. Small lipomas, or irregular and localised fat deposits, have not been

reported previously in the medical literature, and do not usually occur on the

hands, feet or face. They are seen most frequently on the back, neck and trunk.

New York doctors reported two cases in which small lipomas appeared on the

hands, feet or face, during periods of subcutaneous fat depletion, one of which

resolved after human growth hormone treatment commenced. The authors speculated

that these pockets of fat might be protected from lipoatrophy due to some

genetic or molecular feature of this protected compartment of fat, and urged

further investigation of the characteristics of fat cells within lipomas of this

sort in order to shed further light on the pathophysiology of body fat changes

on HAART.

Andrew Carr and David Cooper reported a syndrome of

body fat changes, lactic acid elevation and liver dysfunction in people taking

NRTIs without protease inhibitors. 14 people on NRTIs were compared with 32

drug-naïve patients without body fat changes, 28 NRTI patients without body fat

changes, 44 PI recipients without body fat changes and 102 PI recipients who had

experienced body fat changes. The body fat changes in people on NRTIs were

either peripheral fat loss or abdominal weight gain, but in most cases the

abdominal weight gain was reported to be hepatomegaly (swelling of the liver).

Although the body fat changes looked very similar to those seen in PI

recipients, they were accompanied by elevated lactate, recent onset nausea,

fatigue and lower levels of lipids, glucose and insulin (abstract

11).

Lactic acidosis has been reported to be rare in NRTI

recipients, and this syndrome is a disturbing new feature.(see

href="http://www.aidsmap.com/article.asp?articleID=1707&heading3=Lactic+acidosis&search=true">http://www.aidsmap.com/article.asp?articleID=1707&heading3=Lactic+acidosis&search=true

For further information). Kees Brinkman reported on

five patients in whom lactic acidosis on NRTI treatment appeared to have been

triggered by a period of severe metabolic stress, such as a severe infection or

malnutrition. Riboflavin or L-carnitine have been proposed as treatments for

lactic acidosis ( and L-carnitine has also been proposed as a treatment for

peripheral neuropathy, another potential mitochondrial toxicity). A team from St

Louis in the US reported 37 HIV-positive hospital admissions out of 1261 during

1996-98 has elevated lactic acid, but only 6.1% of these cases could be

definitively pinned on anti-retroviral therapy (sepsis was the most common cause

- 42.4% of cases) (abstract 69).

Cohorts offer no clear links between metabolic

and body fat changes

Despite intensive testing of all the metabolic

parameters reviewed at the meeting, cohort studies are unable to offer a

template for understanding how metabolic changes might be linked to body fat

changes, and in what order any of the metabolic events might be expected to

occur.

Although triglyceride changes generally occur before

reductions in insulin sensitivity and before changes in body fat, none of the

cohorts found a clear relationship. Not everyone with elevated triglycerides

subsequently developed body fat changes or insulin resistance, for example, and

lipoatrophy (peripheral fat loss) was not differentiated from lipodystrophy

(central fat accumulation) in terms of the metabolic changes which

occurred.

Lipodystrophy was associated with:

• Older age

• AIDS diagnosis prior to commencement of PI

  treatment

• Protease inhibitor treatment

• Undetectable viral load

• Duration of treatment (the only drugs significantly

  implicated were d4T and indinavir, but this may be a marker for the frequency

  with which these drugs have been used in successful regimens that suppress viral

  load below the limits of detection)

Several of these risk factors hint at the

contribution of pre-existing lipid profiles (age, AIDS diagnosis), but cohort

findings regarding the contribution of body mass index were contradictory (even

within cohorts), with the previously reported association between higher body

mass index (BMI) and body fat accumulation failing to recur.

What is clear from the cohort studies is that body

fat and metabolic changes are becoming increasingly common as the duration of

therapy lengthens, and a larger proportion of people than originally expected is

developing metabolic and physical abnormalities which cannot be explained. An

International Workshop on this subject was long overdue, and will hopefully

serve as a stimulus for better research, more discussion of the subject and a

greater sense of urgency about finding out why just three years after the

optimism of the Vancouver conference, the public face of AIDS once again has

hollow cheeks and a wasted appearance.

References

First International Workshop on Adverse Drug

Reactions and Lipodystrophy in HIV, San Diego, 1999

Abstracts will be available from

href="http://www.intmedpress.com/">http://www.intmedpress.com