Adipose Tissue Dysfunction and Energy Balance Paradigms in People Living With HIV

Abstract

Over the past 4 decades, the clinical care of people living with HIV (PLWH) evolved from treatment of acute opportunistic infections to the management of chronic, noncommunicable comorbidities. Concurrently, our understanding of adipose tissue function matured to acknowledge its important endocrine contributions to energy balance. PLWH experience changes in the mass and composition of adipose tissue depots before and after initiating antiretroviral therapy, including regional loss (lipoatrophy), gain (lipohypertrophy), or mixed lipodystrophy. These conditions may coexist with generalized obesity in PLWH and reflect disturbances of energy balance regulation caused by HIV persistence and antiretroviral therapy drugs. Adipocyte hypertrophy characterizes visceral and subcutaneous adipose tissue depot expansion, as well as ectopic lipid deposition that occurs diffusely in the liver, skeletal muscle, and heart. PLWH with excess visceral adipose tissue exhibit adipokine dysregulation coupled with increased insulin resistance, heightening their risk for cardiovascular disease above that of the HIV-negative population. However, conventional therapies are ineffective for the management of cardiometabolic risk in this patient population. Although the knowledge of complex cardiometabolic comorbidities in PLWH continues to expand, significant knowledge gaps remain. Ongoing studies aimed at understanding interorgan communication and energy balance provide insights into metabolic observations in PLWH and reveal potential therapeutic targets. Our review focuses on current knowledge and recent advances in HIV-associated adipose tissue dysfunction, highlights emerging adipokine paradigms, and describes critical mechanistic and clinical insights.

Graphical Abstract

Graphical Abstract.

Essential Points.

• Antiretroviral therapy (ART) increases longevity in people living with HIV (PLWH) but heightens their risk for a severe form of insulin resistance and metabolic syndrome that increases morbidity and mortality

• PLWH manifest excessive ectopic adipose tissue that is most apparent in the liver, heart, and skeletal muscle, ultimately leading to end-organ damage from type 2 diabetes, non-alcoholic fatty liver disease, atherosclerotic cardiovascular disease, and heart failure

• The severity and progression of cardiometabolic comorbidities among PLWH are greater than those without HIV, which reflects the complex interplay between HIV persistence in reservoirs and ART effects

• HIV viral proteins and ART exposure dysregulate adipose tissue secretion of hormones and microRNAs required for inter-organ crosstalk and metabolic homeostasis

• Future efforts to identify clinically meaningful, noninvasive biomarkers of adipose tissue alterations in PLWH will facilitate clinical monitoring and provide important steps toward understanding metabolic comorbidities that currently lack therapeutic options

Adipose tissue (AT) is an endocrine organ that stores energy in lipids and regulates whole-body energy homeostasis via secretion of metabolites and adipokines. To regulate whole-body energy homeostasis, AT adjusts its size and function in response to internal and external stimuli, such as nutritional status, temperature, or changes in energy demands (eg, infection). AT stores triglycerides and, as a result, partitions lipids and lipid derivatives from the periphery. To that end, healthy AT requires the extracellular matrix to expand tissue size (1) and secrete proteins for interorgan crosstalk. In response to excess energy supply, AT expansion leads to adipocyte hypertrophy (increase in size of existing adipocytes) and, in some cases, adipocyte hyperplasia (increase in the number of adipocytes) (2).

Maladaptive AT expansion arises from chronic AT inflammation (3, 4), dysregulation of mitochondrial biogenesis and oxidative phosphorylation (5), excessive extracellular matrix accumulation, fibrosis and hypoxia (6), and altered adipokine secretion, most of which occur in people living with HIV (PLWH) on both older and newer antiretroviral therapy (ART) regimens. PLWH also harbor ectopic AT that is most apparent in the liver, heart, and skeletal muscle. The occurrence of ectopic AT in obesity may be explained by expandability theories that state that ectopic fat depots arise when the capacity to store excess carbon as lipid in subcutaneous and traditional visceral AT depots is exceeded (7). Ectopic fat deposition ultimately associates strongly with end-organ damage, manifested as non-alcoholic fatty liver disease (NAFLD) or atherosclerotic cardiovascular disease (ASCVD). The severity and progression of these metabolic comorbidities and complications are greater in PLWH compared with individuals without HIV. Furthermore, the increased severity of AT dysfunction among PLWH likely contributes to the 3-fold increase in the global burden of HIV-associated cardiovascular disease over the past 2 decades, which is responsible for 2.6 million disability-adjusted life-years per year (8).

Early in the HIV epidemic, PLWH developed severe AT dysfunction characterized by loss of peripheral AT (lipoatrophy), increased central adiposity, insulin resistance, and dyslipidemia (9). These severe metabolic alterations were observed following treatment with the early nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) and protease inhibitors (PIs) (10). With the near-universal discontinuation of toxic ART regimens, the incidence of traditional ART-induced lipodystrophy is now negligible, although lipoatrophy among patients previously exposed to early ART combinations persists as a legacy effect. PLWH treated with integrase strand transfer inhibitors (INSTIs) experience greater weight gain on ART initiation than PLWH treated with non-NRTI (NNRTI) agents. Wasting was not uncommon in the pre- and early-ART eras. The current recommendation for first-line ART among most PLWH is a combination of INSTI and contemporary NRTIs (11, 12), and in the context of this treatment regimen, some PLWH exhibit a “return-to-health” overshoot characterized by excessive AT expansion with unhealthy adipocyte hypertrophy (13, 14). The molecular mechanisms of INSTI-induced weight gain remain incompletely understood but likely derive from metabolic pathways that favor adipocyte hypertrophy, hypoxia, and fibrosis (15). This review expands on contemporary roles for AT in energy homeostasis in the context of HIV-associated metabolic comorbidities. In addition, we discuss current knowledge regarding HIV-associated AT dysfunction and identify important knowledge gaps that challenge attempts to reduce the clinical burden of metabolic disease in PLWH.

AIDS-associated Wasting and Lipodystrophy Reflect Alterations in Adipose Tissue Depots

Before the introduction of effective combination ART in the 1990s, AIDS was common among PLWH. AIDS was associated with multiple endocrine deficiencies, including nonthyroidal illness and adrenal insufficiency in patients with malignancy or opportunistic infections (16). Wasting was perhaps the most devastating metabolic manifestation of AIDS (17, 18) and was defined by weight loss >10% or <90% of ideal body weight (19, 20). Importantly, changes in body composition occurred differently in men than women, with men experiencing a disproportionate decrease in lean body mass with relative fat-sparing, whereas women experienced progressive declines in body fat (21).

Futile cycling emerged as a likely mechanism for AIDS wasting (22, 23). Studies in PLWH and lipodystrophy demonstrated high rates of AT lipolysis coupled with blunted oxidation of the released fatty acids, resulting in increased hepatic flux of the fatty acids to undergo reesterification into triglycerides, secretion from the liver, and repartitioning in AT (24, 25). Increased resting energy expenditure (REE) was an alternate hypothesis for AIDS wasting (17, 20, 26). Although increased REE can be present early in the course of HIV and further increase with progression to AIDS, most patients did not experience significant short-term weight loss despite a significant increase in REE (22). Other studies reported a poorly explained association between GH deficiency (GHD) and AIDS-associated wasting (27, 28).

After the introduction of effective ART, PLWH experienced changes in body fat distribution commonly characterized as mixed lipodystrophy, sometimes termed HIV-associated adipose redistribution syndrome or HIV-associated lipodystrophy syndrome. Mixed lipodystrophy comprised both lipohypertrophy (excessive fat deposition in visceral AT (VAT), dorsocervical region, trunk and/or breasts) and lipoatrophy (characterized by decreased subcutaneous AT [SAT] in face, limbs, and buttocks) (29, 30). Lipoatrophy was quantitatively defined by loss of ≥30% limb fat as measured by dual-energy X-ray absorptiometry, which correlated well with clinically evident lipoatrophy (31). In PLWH and clinical lipodystrophy, dual-energy X-ray absorptiometry and abdominal computed tomography (CT) demonstrate marked peripheral fat loss with moderately increased VAT and relatively preserved lean body mass (32). Lipoatrophy is associated with a cluster of metabolic defects including hypertriglyceridemia, hypercholesterolemia, low high-density lipoprotein (HDL) cholesterol, insulin resistance, lactic acidemia, and steatohepatitis (9, 32, 33).

In addition to exposure to the early PIs and NRTIs (34), other factors contributing to HIV-associated lipoatrophy include age, direct effects of HIV proteins, and a complex but unique dysregulation of lipid kinetics (34). Although the prevalence of lipoatrophy has decreased over time because of improved knowledge of risk factors and the availability of less toxic ART, exposure to some of the early agents (particularly the thymidine analog NRTIs; Table 1) is associated with persistent lipoatrophy after drug cessation, known as a legacy effect (35). Two studies documented a biphasic trend in thymidine analog NRTI-associated lipoatrophy in PLWH, with initial fat gain during the initial 24 to 32 weeks (the “return to health” that was regarded as normal) (35, 36), followed by limb fat loss that varied from 1.7% per year for zidovudine (ZDV)/lamivudine to 19% per year for stavudine/didanosine (35). Of note, thymidine analog NRTI-induced lipoatrophy, although persistent, is partially reversible after switching to a nonthymidine analog NRTI (37). Currently, both the World Health Organization and the US Department of Health and Human Services recommend an INSTI in combination with nonthymidine analog NRTIs as the first-line ART regimen for most PLWH (11, 12).

Table 1.

Adipose tissue redistribution and metabolic endpoints by ART group

Drug class	Generation	Drug
Nucleoside reverse transcriptase inhibitors (NRTIs)	Early ART era	Stavudine (D4T)ab	Strongest independent factor associated with lipoatrophy (19%/year fat limb loss when combined with DDI) (35). Associated with type 2 diabetes incidence after adjusting for traditional risk factors and lipodystrophy (38).
		Zidovudine (ZDV/AZT) and lamivudine (3TC)a	ZDV/3TC combo led to 1.7% limb fat loss (35). Type 2 diabetes incidence persisted after adjusting for traditional risk factors and lipodystrophy (38). Combination ZDV/3TC also decreased insulin-mediated glucose disposal whereas fasting lipolysis was increased (39).
		Didanosine (DDI)b	Incidence of type 2 diabetes incidence remained after adjusting for traditional risk factors and lipodystrophy (38).
	Contemporary ART era	Tenofovir disoproxil fumarate (TDF)	Weight-suppressive effects (40).
		Tenofovir alafenamide (TAF)	TDF to TAF switch induced weight gain of 1.8 to 4.47 kg/year regardless of INSTI use (41). TDF to TAF switch is associated with increased LDL and TG levels during the first 9 to 16 months of transition (42).
		Abacavir (ABC)	Higher myocardial infarction risk (43-45).
		Emtricitabine (FTC)	FTC exposure was associated with the lowest T2D risk (adjusted hazard ratio-AHR: 0.30) compared with ABC (AHR: 0.58) and tenofovir (AHR: 0.48) (46).
Nonnucleoside reverse transcriptase inhibitors (NNRTIs)	Contemporary ART era	Efavirenz (EFV)	EFV increased triglyceride levels (47) coupled with weight-suppressive effects (40). Mean total cholesterol, HDL, LDL, and triglyceride levels increased by 38, 10, 21, and 40 mg/dL in the EFV group after 96 weeks of ART initiation in a prespecified subanalysis of the STARTMRK study of treatment-naïve PLWH (47).
		Rilpivirine (RPV)	EFV to RPV switch led to decreased serum lipid profiles at weeks 24 and 48 and a significant increase in LDL/HDL ratio at week 48 compared with baseline (48).
		Doravirine (DOR)	ACTG A5391 (phase 4, 48-week, open-label, randomized clinical trial) will evaluate the effect of INSTI/TAF to DOR (and NNRTI) switch plus TAF or TDF in obese adults with HIV, to reduce or prevent further weight gain (49).
		Nevirapine (NVP)	Switching from the indinavir (PI) to NVP decreased insulin levels but did not alter subcutaneous adipocyte apoptosis markers in people with ART-associated lipodystrophy (50).
Protease inhibitors (PIs)	Early ART era	Indinavir (IDV)b	IDV to NVP switch decreased insulin levels but did not alter subcutaneous adipocyte apoptosis in patients with ART-associated lipodystrophy (50).
		Ritonavir (RTV)	In men without HIV, 4 weeks of lopinavir (LPV)/RTV increased triglyceride levels, VLDL cholesterol, and FFA levels. 2-hour glucose tolerance was reduced, but no change in insulin-mediated glucose disposal rate was observed (51).
		Saquinavir (SQV)b	SQV and ATV, in combo with TDF/FTC, had comparable modest effects on lipids and little effect on glucose metabolism (52).
	Contemporary ART era	Atazanavir (ATV)	Risk of diabetes development was 27% higher among people who started ART on PIs vs NNRTIs, and similar to INSTI (53).
		Darunavir (DRV)	Relative to other EFV-based regimens, DRV/TDF/FTC was associated with an adjusted mean weight gain of 3.68 kg (54).
Integrase inhibitors (INSTIs)	Contemporary ART era	Dolutegravir (DTG)	Greater weight gain compared with EFV (55, 56). ADVANCE controlled trial suggested a synergistic effect between DTG and TAF on weight gain (55, 57). NA-ACCORD observed a 7.2-kg mean weight gain after 24 months and comparatively beneficial total cholesterol-to-HDL ratio, suggestive of minimal lipid-associated cardiovascular risk among PLWH on INSTIs (58). Although INSTI use was associated with an obesity odds ratio of 1.63, the REPRIEVE cohort reported INSTI use was not associated with differences in glucose, LDL cholesterol, or higher odds of metabolic syndrome or hypertension (59).
		Bictegravir (BIC)	4- to 6-kg weight gain vs non-INSTI ART (57, 58).
		Raltegravir (RAL)	Risk of insulin resistance with RAL was similar to that of RTV-boosted ATV or DRV (60). Mean total cholesterol, HDL, and LDL levels increased by 10, 3, and 7 mg/dL in the RAL group, whereas the mean change in total cholesterol, HDL, and LDL levels was 38, 10, and 21 mg/dL in the EFV group after 96 weeks of ART initiation (47).
		Elvitegravir (EVG)	Mean weight gain of 2.72 kg after 96 weeks (57). The mean weight gain was 4.1 kg after 24 months in NA-ACCORD (58).
		Cabotegravir (CAB)	Among PLWH switching to CAB + RPV, changes in weight, BMI, and body composition measurements were minor and similar between treatment arms through months 11-12. There were no clinically relevant changes in the proportion of participants with metabolic syndrome, abdominal obesity, or insulin resistance between arms at months 11-12 (61).

“Early ART era” denotes ART that were Food and Drug Administration-approved between 1987 and 1997.

Abbreviation: ART, antiretroviral therapy.

^a Thymidine analogs.

^b No longer available in the United States, Canada, and/or Europe.

Uptake of ¹⁸F-fluorodeoxyglucose demonstrated higher metabolic activity of SAT in the extremities and abdomen of ART-treated adult PLWH with lipoatrophy compared with people without HIV. In contrast, ¹⁸F-fluorodeoxyglucose uptake in VAT remains unchanged despite increased depot size (62). Additionally, abdominal SAT biopsies from individuals with HIV-associated lipoatrophy demonstrate reduced expression of genes encoding mitochondrial proteins and adipocyte-specific markers compared with persons without HIV, regardless of ART (63). In addition, mRNA expression levels of adiponectin and leptin are downregulated in PLWH compared with persons without HIV, regardless of ART treatment. Taken together, these findings suggest gene expression changes that predict AT competence are present in PLWH before ART initiation (63).

Circulating levels of leptin were noted to be low in PLWH with lipoatrophy and lipohypertrophy (64). This finding motivated a pilot study of treatment with recombinant methionyl leptin (metreleptin), which improved insulin sensitivity and HDL cholesterol levels (65) and decreased VAT (66). Another study reported that metreleptin improved glucose metabolism and non-HDL cholesterol levels but failed to improve abnormal fasting lipid kinetics, triglyceride, or HDL-cholesterol levels in PLWH with lipoatrophy and low fasting leptin levels (≤4 ng/mL) (Table 2) (67). The modest response to metreleptin might be explained by the fact that PLWH exhibit low-normal leptin levels, as opposed to overt leptin deficiency. Despite some beneficial responses in PLWH, metreleptin is not approved for treating HIV lipodystrophy. In fact, metreleptin is only available through a restricted program under a risk evaluation and mitigation strategy (Myalept REMS Program) for the management of congenital or acquired generalized lipodystrophy because of the risks of developing anti-metreleptin antibodies and lymphoma (68).

Table 2.

Randomized controlled trials for HIV-associated lipoatrophy and lipohypertrophy

Reference	Study population	Intervention	Duration	Outcome
Supplements
Sutinen et al (69)	PLWH (85% men) with lipoatrophy on ART (n = 10) and people without HIV (n = 10)	Uridine (36 g 3 times daily)	3 months	Increased total limb fat, intraabdominal fat, and total body fat from baseline.
Benedini et al (70)	PLWH on PI/NRTI (n = 9) and people without HIV (n = 9)	Acetyl-L-carnitine (2 g/day)	8 months	Higher lipid oxidation, decreased intramyocellular triglyceride content, redistribution of fat depots.
Calmy et al (71)	Men living with HIV and lipoatrophy	Uridine (36 g 3 times daily)	24 weeks	No significant change in limb mass.
Pharmacologic therapies
Grinspoon et al (72)	Eugonadal men living with HIV-associated wasting (n = 54)	Testosterone enanthate (200 mg/week) or placebo injections and progressive resistance training (3 times weekly) or no training	12 weeks	Muscle mass in limbs improved after testosterone. HDL was significantly decreased. The authors concluded supervised exercise effectively increased muscle mass without the detrimental metabolic effects of testosterone.
Falutz et al (73)	PLWH (86% men) with accumulation of abdominal fat (n = 412)	Tesamorelin (2 mg subcutaneous daily) or placebo	26 weeks	VAT assessed by CT decreased 15.2% in the tesamorelin group and increased by 5.0% in the placebo group. Triglycerides decreased by 50 mg/dL and increased by 9 mg/dL in the placebo group. The ratio of total cholesterol to HDL cholesterol decreased by 0.31 in the tesamorelin group and increased by 0.21 in the placebo group.
Slama et al – ANRS 113 (74)	PLWH with lipoatrophy (n = 64) and people without HIV (n = 66)	Pioglitazone 30 mg daily or placebo	48 weeks	Improved limb fat atrophy in PLWH receiving ART, but clinical benefits were not perceived by the patients. Treatment did lead to a favorable lipid profile.
Sekhar et al (67)	Men living with HIV, lipodystrophy, and fasting plasma leptin concentration ≤4 ng/mL (n = 17)	Metreleptin (0.02 mg/kg/day for 2 months, followed by 0.04 mg/kg/day from month 2 to 4) or matching placebo subcutaneously twice daily	4 months	No differences in BMI, total fat mass, or lean body mass between the study groups at baseline, or after 2 or 4 months of treatment. Two months of treatment with 0.02 or dose escalation to 0.04 mg/kg/day for 2 additional months of metreleptin did not significantly alter rates of total lipolysis, net lipolysis, plasma fatty acid oxidation, or intrahepatic re-esterification.

Abbreviations: ART, antiretroviral therapy; BMI, body mass index; CT, computed tomography; HDL, high-density lipoprotein; NRTI, nucleoside/nucleotide reverse transcriptase inhibitor; PI, protease inhibitor; PLWH, people living with HIV; VAT, visceral adipose tissue.

Excessive VAT in lipodystrophy impairs GH secretion through increased serum free fatty acids (FFAs) and insulin resistance (75). In contrast, decreasing FFA levels with a potent niacin-derived agent improved GHRH-stimulated secretion of GH in PLWH and lipodystrophy, but not in PLWH without lipodystrophy (75). GHD in PLWH is best discriminated by a serum GH peak <7.5 mg/L after provocative testing with GHRH plus arginine (76). Based on this cutoff point, the prevalence of GHD among PLWH is estimated to be approximately 36.9 per 100 000 adults (76). By comparison, the incidence of adult-onset GHD among individuals without HIV ranges between 1.42 (females) to 1.9 (males) per 100 000 adults (77). The factors influencing abnormal basal and stimulated GH secretion patterns in PLWH include age, gender, body composition, VAT mass, and intrinsic HIV factors, and have been reviewed elsewhere (78).

PLWH Exhibit a “Return to Health” Overshoot in the Contemporary ART Era

Weight gain is a complex physiological phenomenon modified by age, gender, ethnicity, diet, and activity. In the contemporary ART era, up to 65% of PLWH may be classified as overweight or obese in resource-limited and resource-abundant settings (79-81). Weight gain following ART initiation is accompanied by both beneficial and detrimental outcomes. For example, the Veterans Aging Cohort Study reported improved survival with weight gain within the first year of ART in PLWH who started treatment when normal-weight or underweight, whereas overweight or obese individuals lacked a clinical benefit (82), consistent with a return-to-health effect. although a return to normal weight may be desired in underweight persons and a small amount of weight gain in overweight or obese persons may be tolerable after ART initiation, some PLWH demonstrate weight gain that exceeds what could reasonably reflect a “return to health.” Excessive weight gain beyond a “return to health” state has been observed more frequently in PLWH treated with INSTIs and the NRTI tenofovir alafenamide (TAF) (Table 1) (57). One hypothesis suggests that a favorable gastrointestinal side effect profile might partly explain greater weight gain with newer ART (57).

Although PIs are not part of the recommended first-line ART regimen, it is worth noting the effects of PIs on weight gain. In the Centers for AIDS Research Network of Integrated Clinical Systems cohort, 6-month weight gain after ART initiation was evaluated in 3232 treatment-naïve participants (84% men). Darunavir (DRV) in combination with tenofovir disoproxil fumarate (TDF)/emtricitabine (FTC) induced a mean weight gain of 3.68 kg during the initial 6 months of ART in treatment-naïve individuals when compared with efavirenz (EFV)/TDF/FTC (54). This weight gain was only surpassed by dolutegravir (DTG) (mean weight gain, 4.37 kg) and bictegravir (BIC) (mean weight gain, 3.86 kg) in combination with TAF/FTC and persisted after adjusting for multiple factors (54). Of note, both TDF and EFV may have weight-suppressive effects (40), complicating data interpretation of cumulative effects of ART on body composition and weight.

The ADVANCE trial, a phase 3, open-label, randomized clinical trial compared the combination of FTC and DTG plus TAF or TDF against the standard-of-care regimen EFV/TDF/FTC. After 48 weeks, persons randomized to DTG-based ART gained significantly more weight than persons randomized to EFV (mean increase, 6.4 kg in the DTG/TAF group, 3.2 kg in the DTG/TDF group, and 1.7 kg in the standard-of-care group), with the greatest weight gain seen in participants treated with DTG/TAF. These trends in weight gain continued to week 96 without a plateau in DTG-treated persons. Weight gain was significantly higher in females in all 3 groups (55). As previously mentioned, TDF and EFV may also have weight-suppressive effects (40). However, in most studies, the difference between TDF and TAF is only about 2 to 3 kg (41, 83, 84), making the clinical relevance of differences between the 2 tenofovir prodrugs unclear. Numerous other ART start-and-switch studies support the findings of the ADVANCE trial (58, 85). Results from a large US observational database (OPERA) demonstrated accelerated weight gain from 1.8 to 4.47 kg/year following TDF to TAF switch, regardless of whether PLWH switched from a boosted PI or NNRTI to an INSTI (Table 1), suggesting an independent effect of TAF on weight gain (41). There were differences among the different INSTIs. Combinations of BIC/TAF, DTG/TAF, or elvitegravir/cobicistat/TAF induced 4.47 kg/year, 3.09 kg/year, and 2.55 kg/year weight gain, respectively, within the first 9 months following the switch from non-INSTI to INSTI. However, after the initial 9 months, there was no significant additional weight change attributable to TAF (41).

In the Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE) trial, current INSTI use was associated with higher odds of obesity (odds ratio = 1.63), mean body mass index (BMI) change (+1.5 kg/m²), and larger waist circumference (+3.6 cm) compared with non-INSTI drugs. Differences in weight-related to INSTI use were greater among women and non-White participants, with sex and race additively affecting BMI (59). Importantly, INSTI use was not associated with differences in fasting blood glucose, low-density lipoprotein (LDL) cholesterol, or odds of metabolic syndrome or hypertension (59). In contrast, the large North American HIV cohort (NA-ACCORD) reported a type 2 diabetes hazard ratio of 1.17 in the INSTI group that was attenuated to 1.03 after adjusting for 12-month weight gain (53).

Although some efforts have been undertaken to explore the mechanisms of ART-induced weight gain, the causal effects of ART and HIV on body composition remain incompletely understood. In mice, a high-fat diet admixed with ART (EFV/FTC/TDF) additively increased pro-inflammatory genes in VAT and impaired glucose metabolism and energy balance (86). In vitro exposure of adipose stem cells to DTG was associated with greater lipid accumulation via elevated expression of pro-adipogenic and lipogenic transcription factors (87). Moreover, DTG and raltegravir (RAL) decreased expression and secretion of adiponectin and leptin (87). In the same study, larger adipocytes were observed in the SAT and VAT of macaques treated with DTG (87). Other in vitro studies showed that DTG significantly reduced brown adipogenic markers, such as uncoupling protein-1 (UCP1) (88). Moreover, DTG reduced UCP1 and levels of mitochondrial complex IV components, which contributed to depleted mitochondrial respiratory capacity and slowed insulin-stimulated glucose uptake (88). DTG and BIC also decreased expression of the brown adipocyte markers including Cbp/P300-interacting-transactivator-1 (CITED1), PR domain containing 16 (PRDM16), and iodothyronine deiodinase-2 (DIO2) (15). Mechanistically, DTG induced hypoxia-inducible factor 1 subunit alpha (HIF1A) expression and elevated fibrosis markers (15), which strongly correlated with repressed PRDM16, UCP1, and DIO2. Altogether, a few animal and cell-based studies suggest INSTI-induced weight gain occurs because of cellular metabolic programs that favor adipocyte hypertrophy, hypoxia, and fibrosis (15).

AT Dysfunction Contributes to Ectopic Fat Accumulation in PLWH

HIV-associated Nonalcoholic Fatty Liver Disease

NAFLD is an independent risk factor for cardiovascular disease, type 2 diabetes, and all-cause mortality (89). NAFLD occurs on a well-defined disease spectrum ranging from steatosis (increased hepatic triglyceride content) to non-alcoholic steatohepatitis (NASH) with or without fibrosis to advanced irreversible scarring (cirrhosis), hepatocellular carcinoma, and end-stage liver disease (90, 91). NAFLD affects 25% of adults worldwide (92). In contrast, the pooled prevalence of HIV-associated NAFLD and significant fibrosis are estimated at 38% (95% CI, 31-45) and 13% (95% CI, 8-18), respectively (93).

In a case-control study comparing age- and sex-matched patients with primary NAFLD (n = 33) and patients with HIV-associated NAFLD (n = 33), patients with HIV-associated NAFLD had significantly higher mean aspartate aminotransferase, alanine aminotransferase, alkaline phosphatase, and serum triglycerides (94). Similarly, compared with age and sex-matched primary NAFLD, patients with HIV-associated NAFLD had significantly higher rates of steatohepatitis (37% vs 63%, P = .04), and features of liver injury, including lobular inflammation and acidophil bodies (94). Although obesity is associated with NAFLD in the general population, NAFLD is noted in PLWH at significantly lower mean BMI, suggesting PLWH manifest higher prevalence of “lean NAFLD” (Fig. 1) (95).

Figure 1.

Characteristics of primary NAFLD vs HIV-associated NAFLD. Primary NAFLD is associated with obesity in the general population, whereas HIV-associated NAFLD is noted at a significantly lower mean BMI. In a head-to-head comparison between age- and sex-matched people with primary and HIV-associated NAFLD, people with HIV-associated NAFLD showed significantly higher mean liver enzymes, alkaline phosphatase, and serum triglycerides. Similarly, patients with HIV-associated NAFLD had significantly higher rates of biopsy-proven steatohepatitis and features of liver injury, including lobular inflammation and acidophil bodies.

The natural progression of HIV-associated NAFLD was assessed in a subanalysis of a double-blind trial that evaluated the effects of tesamorelin vs placebo on NAFLD (96). In the placebo group, fibrosis progressed in 38% (n = 9) of participants with HIV-associated NAFLD over a 12-month period (97). Furthermore, 5 of 9 participants with fibrosis progression had no evidence of fibrosis at baseline. Baseline VAT was higher among volunteers with fibrosis progression than without progression (306 ± 119 cm² vs 212 ± 89 cm²; P = .04). In multivariable regression modeling, a 25 cm² higher VAT at baseline was associated with a 37% increased odds of fibrosis progression on adjusting for baseline NAFLD Activity Score, liver fat content, and BMI (odds ratio, 1.37; 95% CI, 1.03-2.07; P = .03). Of note, 50% (n = 12) of subjects had no change in fibrosis and 13% (n = 3) experienced fibrosis regression (97).

The high prevalence of NAFLD in PLWH was previously ascribed to coinfection with hepatitis C virus (HCV) or adverse ART effects, particularly those of stavudine (98-103). However, several epidemiological studies reported that only 50% of HIV-associated liver diseases were attributable to hepatitis B virus and HCV coinfection or alcohol abuse (104-106). Moreover, new data demonstrate that PLWH develop NAFLD and accelerated liver disease without HCV or hepatitis B virus coinfection (98). The factors that promote development and progression of a specific, HIV-associated form of NAFLD are likely multifactorial and closely tied to AT dysfunction. Greater VAT accumulation predicts risk for NAFLD and metabolic syndrome in PLWH (97, 107). Kinetic studies using stable isotope infusions and mass spectrometry identified accelerated AT lipolysis and defective whole-body oxidation of fatty acids as an important element in the pathophysiology of NAFLD in PLWH (24, 25), associated with a mixed hyperlipidemia in which triglycerides, very low LDL and LDL are increased (108). Additionally, HIV-associated hormonal imbalances such as hypogonadism or increased renin-angiotensin system activation appear to contribute to energy balance regulation and possibly NAFLD risk in PLWH (109, 110).

As previously mentioned, early ART agents (NRTIs and PIs) were associated with a lipodystrophic syndrome of which NAFLD is a component (9, 111). Older NRTIs may contribute to NAFLD pathogenesis through altered mitochondrial toxicity, insulin resistance, and redistribution of AT to the visceral space (112, 113). Moreover, both NRTIs and NNRTIs affect adipogenesis and adipocyte differentiation, leading to mitochondrial injury, adipocyte death, and FFA and triglyceride accumulation (113), which correlate with hepatic steatosis. Contributions of specific drugs to NAFLD risk in the current ART era are under investigation but likely mediate risk in part through their effects on weight gain and body composition.

Because there are no US Food and Drug Administration-approved medications for NAFLD treatment, management includes lifestyle changes that promote weight loss of at least 5% (preferably ≥10% weight loss) because weight loss is associated with greater liver and cardiometabolic benefit (114). Consideration of weight loss and insulin-sensitizing (eg, pioglitazone) medications are recommended in the general population (114). Few NAFLD/NASH treatment trials have included PLWH. Therapy options for HIV-associated NAFLD may include pioglitazone because this medication was well-tolerated and reduced hepatic fat by magnetic resonance spectroscopy in individuals with HCV and HIV coinfection (115). An open-label clinical trial recently showed vitamin E decreases markers of liver fat content and hepatocyte apoptosis in PLWH with biopsy-proven NASH (116). The GHRH analog tesamorelin improves liver enzymes, hepatic steatosis, and angiogenic, fibrogenic, and pro-inflammatory mediators compared with placebo (96, 117, 118). The effectiveness of glucagon-like peptide-1 receptor agonists for management of NAFLD in PLWH is under study (119).

Skeletal Muscle Fat Accumulation

Intermuscular adipose tissue (IMAT) accumulates beneath the muscle fascia and between bundles of myocytes. Both VAT and IMAT increase with age, whereas SAT and skeletal muscle mass decrease (120, 121). The Fat Redistribution and Metabolic Change in HIV Infection study was the first large cross-sectional study to assess the total IMAT in 425 PLWH and 211 individuals without HIV from diverse ethnic groups (122). Contrary to predictions, the Fat Redistribution and Metabolic Change in HIV Infection researchers found the total IMAT assessed by magnetic resonance imaging was 51% lower in participants with HIV compared with participants without HIV (P = .003), a finding that was attenuated but persisted in multivariate analyses (122). When stratified by sex, IMAT was significantly lower in men but not in women. The authors attributed the lower IMAT to prior exposure to NRTIs such as stavudine, didanosine, and nelfinavir (122). A recent study by Adrian et al demonstrated that the PI DRV/ritonavir was associated with greater fat area and lower density of both fat and muscle, and RAL with less intermuscular fat in the psoas (123). The clinical relevance of IMAT relies on its resemblance to VAT, which constitutes a marker of increased metabolic and cardiovascular disease risk in the general population (124, 125), and metabolic syndrome among PLWH (126). The relationship between metabolic and cardiovascular health and improvements in VAT or other fat depots is not straightforward. PLWH exhibited improvements in VAT after 1 year of lifestyle modifications but experienced an increase in LDL and significant progression of carotid intima-media thickness (127), a surrogate marker for cardiovascular disease. These findings underscore the complex relationships between AT maladaptation and cardiometabolic risk in PLWH.

Intramyocardial and Epicardial Fat Accumulation

HIV intramyocardial steatosis is associated with diastolic dysfunction (128, 129), a precursor of heart failure with both preserved ejection fraction and reduced ejection fraction (130), both of which are prevalent among PLWH (131, 132). In healthy conditions, cardiomyocytes populating the myocardial structural space contain approximately 0.4% to 0.5% triglycerides (133, 134). The REPRIEVE trial reported a median intramyocardial triglyceride content of 0.59% (Q1, Q3: 0.28%, 1.15%). Intramyocardial triglyceride content (IMTC) was increased (>0.5%) among 52% and markedly increased (>1.5%) among 22% of participants (135). Parameters associated with increased myocardial steatosis included age, BMI ≥ 25 kg/m², history of IV drug use, and CD4⁺ T-cell count < 350 cells/mm³ (135). Importantly, 67% of individuals with increased IMTC (>0.5%), 71% with nonmarkedly increased IMTC (≤1.5%), and 72% with markedly increased IMTC (>1.5%) had a normal waist (<102 cm in men, <88 cm in women) (135). These results suggest that intramyocardial steatosis in PLWH occurs even in individuals without abdominal obesity.

Men living with HIV also accumulate epicardial fat, which is associated with a higher risk of cardiovascular events compared with men without HIV with similar cardiovascular and metabolic disease risk profiles (136). Ectopic epicardial fat volume in HIV is associated with VAT area, fasting glucose, and insulin. The associations with elevated blood glucose and insulin remained significant after controlling for age, race, BMI, adiponectin, VAT, and ART (137). Recently, epicardial fat volume was shown to correlate with abnormal cardiac structure and function as evidenced by left ventricular mass index and left ventricular global longitudinal strain, respectively (138). Furthermore, Guaraldi et al showed an association between epicardial fat and traditional risk factors for atherosclerosis, suggesting epicardial fat may be a useful marker of cardiovascular risk in PLWH (139). Because the sequelae of expanding epicardial and intramyocardial fat depots resemble those of NAFLD and its increased cardiometabolic risk, intrinsic (viral proteins) and extrinsic (ART) HIV-specific factors likely generate parallel effects in driving these ectopic fat depots, including triggering proinflammatory and profibrotic changes at the cellular level. These findings underscore the importance of ectopic fat accumulation as a target for the management of cardiometabolic alterations in PLWH.

Comorbidities of Obesity Contribute to Additional Disease Burden in PLWH

Lipotoxicity, Insulin Resistance, and Type 2 Diabetes in PLWH

In 2019, the crude incidence of type 2 diabetes (T2D) was estimated at 5.9 per 1000 among diagnosed US adults aged 18 years or older (140). By comparison, the incidence of T2D and prediabetes among PLWH has been estimated to be 13.7 per 1000 person-years and 125 per 1000 person-years, respectively (141). The increased incidence of T2D among PLWH is likely related to increased circulating levels of lipids that contribute to lipotoxicity in the peripheral organs, which in turn leads to insulin resistance in the liver and skeletal muscle (142). Lipotoxicity is characterized by the accumulation of lipolysis mediators (eg, palmitate derivatives), which may account for the hypertriglyceridemia seen in lipodystrophic PLWH (24, 25). To date, no published studies have addressed the role of lipotoxicity on insulin resistance or T2D in the contemporary ART era. In contrast, most efforts have been directed toward addressing the role of AT inflammation as a mechanism for insulin resistance and T2D in PLWH. Indeed, several studies reported that HIV and simian immunodeficiency virus (SIV) are associated with a substantial increase in the proportion of adipose CD8⁺ T cells relative to CD4⁺ T cells, which is remarkably similar to enrichments of T cells described in obesity (143-145).

A study that compared 8 non- and prediabetic HIV-negative persons and 9 (of 26) PLWH paired by similar age, glycated hemoglobin, and BMI reported that, despite a much higher proportion of CD8⁺ T cells in the SAT of PLWH, there was no significant difference in the overall proportions of SAT CD4⁺ and CD8⁺ memory T-cell subsets between PLWH and HIV-negative persons (146). Of note, there were significant gender (10% vs 63% females, P = .02) and race (40% vs 100% Caucasian, P < .01) differences between the HIV subgroup and non-HIV controls that could potentially explain the lack of differences between groups (146). Thus, the contribution of AT inflammation to insulin resistance and T2D in PLWH remains uncertain. In addition, these results could also explain why treatment with medications that diminish vascular inflammation, such as statins, show conflicting results in PLWH (147, 148).

Significant efforts were also devoted to assessing the contribution of different ART regimens on the risk of T2D or prediabetes, as this was a modifiable factor compared with aging, family history of diabetes, or Black or Hispanic origin (141). In the early ART era, the main T2D risk factor in PLWH was PI use. Indeed, incidence rates of hyperglycemia as high as 5-fold were reported in the setting of PI use (149). Mechanisms that account for these effects include skeletal muscle and AT insulin resistance, and β-cell impairment in PLWH exposed to 12 weeks of PI (150). However, the effects of PIs on glucose homeostasis vary widely among the different PIs. For example, an in vitro study demonstrated that an oligopeptide found in indinavir inhibited glucose transport in primary rat adipocytes via reversible noncompetitive inhibition of GLUT4 (151), whereas a clinical study of PLWH showed atazanavir in combination with ritonavir had a modest effect on insulin sensitivity compared with lopinavir/ritonavir, as measured by glucose disposal rate during hyperinsulinemic-euglycemic clamps (152). Moreover, at therapeutic doses, atazanavir does not inhibit glucose uptake in vitro or in vivo (153, 154).

In addition to PIs, older-generation NRTIs have been associated with T2D (Table 1) (38). In contrast to PIs, the mechanism for NRTI-induced T2D was attributed to mitochondrial dysfunction. One study found that 1-month treatment with stavudine was associated with decreased mitochondrial DNA from muscle biopsies and insulin sensitivity, as measured by the glucose infusion rate during hyperinsulinemic-euglycemic clamps (155). As previously mentioned, thymidine analog NRTIs were the main culprit of lipoatrophy; hence, lipolysis was increased by 22% after 3 months of combination lamivudine/ZDV (39). The risk of T2D for stavudine, ZDV, and DDI persisted after adjusting for traditional risk factors and lipodystrophy (38). Similar to the newer generation of PIs, the use of newer NRTIs, including FTC, abacavir, and tenofovir, is associated with a lower risk of T2D (46).

Case reports of new-onset diabetes have been reported in conjunction with INSTI use (156). However, this finding has not been replicated in larger studies. Despite its class effect on weight gain, INSTIs are not associated with differences in fasting blood glucose, LDL cholesterol, or odds of metabolic syndrome or hypertension (59). Furthermore, the NA-ACCORD reported a T2D hazard ratio of 1.17 in the INSTI group that was attenuated to 1.03 after adjusting for 12-month weight gain (Table 1) (53). One clinical trial compared RAL with 2 boosted PI regimens and found that the increases in homeostasis model assessment-insulin resistance in all 3 arms were not significantly different from one another and appeared to be independent of changes in VAT (60).

Multiple observational studies shed light on important associations between different risk factors, cellular mediators and markers, and metabolic phenotypes in PLWH. Further mechanistic studies will be necessary to dissect individual contributions of HIV viral proteins, ART regimens, and chronic immune activation to the pathophysiology of insulin resistance in the contemporary ART era. Only then will we be able to devise new therapeutic strategies for managing these HIV-associated metabolic comorbidities and limit their impact on the quality of life of PLWH.

ASCVD Burden in PLWH

Metabolic abnormalities are tightly associated with ASCVD risk factors in adults with HIV (157). The disproportionate incidence of cardiometabolic comorbidities and complications negatively affects the quality of life of this population. Large cohort studies demonstrated a high prevalence of hypertension, dyslipidemia, and insulin resistance resulting from greater VAT among PLWH (126). The high prevalence of known cardiovascular risk factors only partially accounts for the 3-fold ASCVD risk in PLWH compared with individuals without HIV (8). Furthermore, conventional therapies (eg, statins) exert modest effects for primary or secondary prevention of ASCVD in PLWH (32).

ASCVD risk prevention is based on disease prediction using individual risk factors. Traditional ASCVD risk factors such as age, T2D, current smoking, hypertension, and dyslipidemia are associated with elevated ASCVD risk among PLWH, as in the general population (158, 159). HIV constitutes an independent ASCVD risk factor (158), especially in individuals with low current or nadir CD4⁺ T-cell counts or a history of sustained untreated HIV (160). Although no general population risk assessment models focused on PLWH, the latest American Heart Association scientific statement on management of ASCVD in HIV recommends applying the ACC/ASCVD risk estimator, Framingham cardiovascular disease risk estimator, or D:A:D plus risk enhancer factors, such as history of prolonged HIV viremia or delayed ART initiation, low current or nadir CD4⁺ T-cell count (<350 cell/mm³), HIV treatment failure or nonadherence, metabolic syndrome/lipoatrophy/lipohypertrophy, NAFLD, or hepatitis C co-infection (161).

Associations between intra-abdominal VAT and increased metabolic and cardiovascular disease risk are well described in cross-sectional and longitudinal studies (124, 125). Large cohort studies demonstrate a higher prevalence of hypertension, dyslipidemia, and insulin resistance resulting from greater VAT among PLWH (126). Despite a reduction in VAT after 1 year of lifestyle modifications (diet, exercise, smoking cessation, and antidiabetic/dyslipidemic therapy), PLWH with and without lipodystrophy, exhibited an increase in LDL and had significant progression of carotid intima-media thickness (127), a surrogate marker for cardiovascular disease in the general population.

Taken together, the role of AT dysfunction in ASCVD risk in PLWH cannot be predicted from changes in immune cells or fat distribution. Intricate interactions between diverse cellular mediators, including circulating metabolites from AT, may play a more prominent role in the ASCVD risk in PLWH.

PLWH Exhibit Intestinal Dysbiosis That Corresponds With AT Inflammation

The gastrointestinal tract is colonized extensively with commensal microorganisms collectively called the microbiota, which play an important role in metabolism and immune defense as part of a symbiotic relationship with the host. Dysbiosis depicts an imbalance in the microbiota composition that frequently accompanies the chronic inflammation observed in generalized obesity (162) and PLWH (163). A few studies reported intestinal dysbiosis and reduction of bacterial alpha diversity in PLWH compared with individuals without HIV (164, 165). Not surprisingly, PLWH accumulate markers of bacterial translocation and immune activation that correspond with inflammation in AT during acute and chronic infections. Bacterial alpha diversity correlates with VAT and systemic inflammation (166), strengthening the notion that intestinal microbiota contribute to ectopic fat accumulation in PLWH.

Bacteria within the gut break down dietary fibers and starch through anaerobic fermentation, producing short-chain fatty acids as metabolite byproducts. Although short-chain fatty acids maintain intestinal homeostasis, they quickly enter the bloodstream and can be metabolized by peripheral tissues including AT (167). In addition to loss of gut microbial diversity, obesity and HIV are associated with lower butyrate-producing bacteria (168). The potential relationships between visceral obesity, gut microbiota, short-chain fatty acid abundance, and immune activation in people living with HIV are an active area of research, including the potential roles of dysbiosis in altering ART outcomes.

AT Stress Alters Secretion of Metabolic Hormones That Change Metabolism in PLWH

Efficient interorgan communication facilitates metabolic adaptation to energy demands. In addition to its role in nutrient sensing and energy homeostasis, AT secretes a large number of bioactive molecules (eg, adipokines, microRNAs [miRNAs]) that participate in immune responses and metabolic regulation through their paracrine and/or endocrine actions (169). In addition, adipose-derived extracellular vesicles carry miRNAs and proteins, which act as cellular messengers for interorgan crosstalk from fat cells. The communication between AT and other organs is altered in many pathological conditions, such as obesity, diabetes, and HIV. However, only a few candidate molecules that perform interorgan communication have been rigorously studied in PLWH.

Leptin

The identification of the adipocyte-specific protein hormone leptin defined AT as a major endocrine organ (170) that signals energy deficit to the central nervous system (171). Leptin modulates energy homeostasis (172) and counteracts metabolic dysfunction during overnutrition by increasing lipolysis and glucose uptake and inhibiting hepatic glucose output (173). Leptin physiology in PLWH is variably disrupted. Some PLWH and lipoatrophy exhibit low leptin levels because of reduced adipocyte mass, whereas others exhibit high leptin levels, suggesting leptin resistance (174). Conversely, in a larger clinical trial (AIDS Clinical Trials Group A5260s), mean leptin levels increased 22% after 96 months of ART initiation, whereas mean adiponectin changed 1% (175). Furthermore, higher circulating leptin was associated with greater homeostasis model assessment-insulin resistance and high-sensitivity C-reactive protein independent of fat depot size, suggesting that greater adipocyte lipid content contributed to impaired glucose tolerance and systemic inflammation among PLWH starting ART (175). Possibly because of the variable leptin levels among PLWH, treatment with metreleptin yielded modest results. PLWH exhibit low-normal leptin levels, as opposed to overt leptin deficiency, which may explain modest effects of metreleptin in clinical trials (65-67). However, as noted previously, metreleptin is not approved for management of lipodystrophy in PLWH.

Adiponectin

Adiponectin targets diverse tissues and generates insulin-sensitizing, antiatherogenic, and anti-inflammatory properties (176, 177). In addition, adiponectin enhances hepatic fatty acid oxidation, suppresses hepatic de novo lipogenesis (178), and increases ceramidase activity, thus preventing or reversing diet-induced steatosis, insulin resistance, and glucose intolerance (179). The role of adiponectin as a marker of metabolic health in PLWH is not clear. In a cardiovascular substudy of the Multicenter AIDS Cohort Study, higher adiponectin levels were associated with metabolic health in men with HIV (180). Conversely, in the same cohort, lower adiponectin levels were observed in men with and without HIV and were associated with coronary stenosis by CT angiography (181). Furthermore, low adiponectin levels in men living with HIV correlate with the severity of subclinical atherosclerosis, independent of traditional cardiovascular disease risk factors (181). In contrast, another study reported a correlation between higher adiponectin levels and worse neuropsychological test scores in men with HIV, suggesting that adiponectin excess may contribute to cognitive decline in men living with HIV. Adiponectin levels in this study were inversely correlated with VAT (r = −0.37, P < .0001) and positively correlated with thigh SAT (r = 0.19, P < .001) area in men with HIV, suggesting a disequilibrium of traditional relationships between AT, adipokines, and end-organ damage (182). Moreover, lean ART-treated men with HIV displayed low adiponectin levels that mirror those of obese, insulin-resistant men without HIV (183).

Adiponectin changes following ART vary (184, 185). For example, among treatment-naïve PLWH, adiponectin increased by 9%, 8%, and 1% in persons treated with ATV/r, DRV/r, and RAL, respectively, after 48 weeks (186). Conversely, a decline in adiponectin from baseline to 48 weeks was associated with a rise in high-sensitivity C-reactive protein and IL-6 (175). Some studies have tried to clarify the role of leptin and adiponectin in PLWH by estimating the leptin/adiponectin ratio, a surrogate biomarker of metabolic competence in the general population (187). Interestingly, a small single-center study in Japan reported a significant increase in leptin/adiponectin ratio and serum leptin levels 12 months after initiation of DTG- or BIC-based ART (188). These results may reflect the increased size of AT depots observed on INSTI therapy initiation.

Plasminogen-activator Inhibitor 1

Plasminogen-activator inhibitor 1 (PAI-1) is another protein that can be secreted in abundance by VAT in obese people and predicts insulin resistance (189) and NAFLD (190, 191). In large epidemiological studies including the Framingham study, serum PAI-1 levels have been associated with increased fasting glucose (192), systolic blood pressure, and triglyceride levels (193). Similarly, PAI-1 appears to be an independent predictor of impaired insulin sensitivity (194) and fatty liver disease in PLWH with lipodystrophy (195). However, conflicting results have also been reported in PLWH. One study reported that PAI-1 expression was significantly reduced at the mRNA and protein level in PLWH compared with matched persons without HIV, a finding that appeared to be related to NNRTI exposure (196). In contrast, in lipodystrophic PLWH, plasma PAI-1 levels positively correlated with BMI (r = 0.74, P < .01), TNF-α level (r = 0.64, P < .01), and VAT quantity (r = 0.67, P < .01) (197). Cytokines such as TNF-α, IL-6, and IL-8 are determinants of adiponectin and PAI-1 abundance in plasma, underscoring the role of inflammation on adipokines and cardiometabolic risk in PLWH (196, 197).

Fibroblast Growth Factor 21

Secretion of the protein hormone fibroblast growth factor 21 (FGF21) from the liver improves glucose clearance and insulin sensitivity (198, 199). In healthy mice and people, the most prompt and consistent effects of an acute rise in serum FGF21 levels are increased fat oxidation and lipolysis suppression in concert with insulin (200-202). Along these lines, FGF21 is required for acute insulin sensitization of AT, manifested by enhanced glucose uptake, blunted lipolysis, and reduced plasma FFAs (203). Although FGF21 physiology is well-documented, we do not yet understand how FGF21 contributes to the metabolic phenotype of PLWH.

Despite observations that FGF21 is secreted from AT and exerts paracrine effects (204), FGF21 is primarily secreted from the liver (205). Circulating FGF21 levels are increased in PLWH compared with controls without HIV, regardless of ART (206-208), but this is associated with an unhealthy metabolic profile and precursor lesions for liver cancer. FGF21 levels in PLWH strongly and positively correlate with lipodystrophy, dyslipidemia, and NAFLD (206). FGF21 levels also correlate with the predictors of hepatocellular carcinoma, hepatic steatosis, and NASH in other conditions of altered energy metabolism, such as obesity (209). Identifying clinically meaningful relationships between FGF21 and chronic liver disease will enhance our understanding of the pathogenesis of hepatocellular cancer risk in PLWH, facilitate clinical monitoring, and potentially provide a noninvasive endpoint for future therapeutic studies.

FGF21 appearance in the context of fatty liver disease likely reflects energetic stress on peripheral tissues and consequently lacks physiological impacts on glucose metabolism (210, 211). FGF21 increases thermogenic gene expression, including UCP1, in specific white adipose tissue depots (212). Srinivasa et al noted that FGF21 in PLWH is associated with increased expression of brown adipose tissue marker genes in subcutaneous dorsocervical fat biopsies (208). These observations argue that FGF21 resistance occurs in specific tissues, such as AT, that renders elevated levels of this hormone unable to reverse metabolic defects.

Growth Differentiating Factor 15

Growth differentiation factor 15 (GDF15) is a recently acknowledged circulating peptide that exerts weight-lowering effects when administered to rodents and nonhuman primates (213, 214). Intracellular stress drives the expression of GDF15 in the AT, kidney, liver, skeletal muscle, and some immune cells. Other cellular factors that elevate GDF15 include mitochondrial dysfunction (215-219), the integrated stress response (216, 220), and AMP-associated protein kinase activation (221). Consequently, GDF15 expression is upregulated in chronic inflammatory diseases (222), cardiovascular disease (218, 223), NAFLD (224), and diabetes (218, 223, 225).

As with FGF21, the activation of transcription downstream of the integrated stress response drives expression of GDF15. Specific HIV-related risk factors (eg, viral proteins such as viral protein R [Vpr], immune activation, ART) may result in concerted stress that alters the course of liver disease and sustained GDF15 secretion. Men living with HIV have higher plasma GDF15 levels than HIV-negative men (both healthy and with metabolic syndrome) (226). There was a positive correlation between GDF15 levels and age in healthy subjects and subjects with metabolic syndrome, but not in PLWH, which may reflect premature aging in PLWH.

Glial cell line-derived neurotrophic factor receptor family receptor alpha-like, a transmembrane receptor localized to the hindbrain (227) binds GDF15 and mediates its satiety effects in rodents (220). In PLWH, higher GDF15 levels are associated with greater weight loss after 12-week treatment with metformin (228). However, it remains unknown whether GDF15 induces appetite suppression effects in this population. Additional studies will be needed to test whether GDF15 levels in PLWH are associated with fatty liver disease and resistance to the central nervous effects of GDF15, similar to other hormone resistance paradigms in PLWH (229). Nonetheless, the satiety and weight loss effects of GDF15 represent important biological endpoints to slow weight gain in PLWH and manage HIV-associated metabolic comorbidities.

microRNAs

microRNAs (miRNAs) are noncoding RNAs of 20 to 25 nucleotides that bind target mRNAs in the 3′ untranslated region to induce mRNA degradation and inhibit protein translation (230). DICER is an integral protein in the miRNA processing required for adipocyte functions (231). Mice lacking Dicer in adipocytes (Dicer a-KO) display lipodystrophy, insulin resistance, fatty liver, and premature mortality (231). Similarly, PLWH and lipodystrophy exhibit reductions in AT Dicer expression (232). These observations strongly support a rate-limiting role for miRNAs in the functions of adipose tissue in PLWH.

AT-derived exosomal miRNAs likely represent a new form of adipokine. In addition to their endogenous actions, miRNAs can be transported into the extracellular space within nanoparticles (50-150 nm in size) termed exosomes. Exosomes transport most of the miRNA secretome and allow paracrine and endocrine transfer of miRNA silencing into circulation. Dicer a-KO depletes circulating miRNAs that communicate with peripheral tissues (233). Along these lines, transplantation of wild-type white adipose tissue into Dicer a-KO restored plasma miRNAs and improved glucose tolerance. Dicer a-KO also showed reduced miR-99b in circulating exosomes coupled with higher liver Fgf21 mRNA, which could be partially corrected with the administration of miR-99b loaded into exosomes. These mechanisms of Fgf21 regulation derived from AT exosomes may be relevant for interpreting the hyperelevation of plasma FGF21 observed in PLWH.

Furthermore, expression of many circulating miRNAs correlated significantly to adipose Dicer expression among all participants (persons without HIV; PLWH with and without lipodystrophy). Moreover, the lipodystrophic miRNA signature strongly correlated with dorsocervical AT area (a clinical sign of lipodystrophy) and decreased expression of the brown adipocyte genes, UCP1 and PRM16 (234). Taken together, these data suggest that alterations at the level of circulating miRNAs from AT serve as markers of lipodystrophy among PLWH. More studies will be needed to understand the endocrine functions and predictive power of miRNAs in AT exosomes in PLWH.

Preclinical Models Define Adipose Tissue as a Reservoir for Latent HIV and Viral Proteins

Preclinical Models for HIV-associated Metabolic Dysfunction

Animal models have provided critical insights into HIV latency and reservoirs (Fig. 2). Although many models have been developed to study HIV/AIDS (235), very few proved to be effective for the study of cardiometabolic consequences. Whereas humanized mice make it possible to model HIV infection of human cells in vivo and design studies using genetically identical animals, these models are limited in their ability to replicate the effects of HIV and ART on nonhematopoietic tissues and to recreate basic features of HIV disease in humans (235).

Figure 2.

Adipose tissue is an important anatomic HIV reservoir. The main cellular reservoirs for HIV are CD4⁺ T cells and macrophages. After acute HIV infection, surviving HIV-infected CD4⁺ T cells and macrophages continue to harbor and sustain the proliferation of provirus and viral proteins. ART initiation (dashed line) decreases new viral replication and cellular infection but does not eliminate the survival of these latently infected cells, which can persist in anatomical reservoirs, such as AT. In aviremic PLWH on ART, CD4⁺ T cells of AT harbor replication-competent HIV with a median latent HIV proviral DNA copy number equivalent to that of circulating and mesenteric lymph node CD4⁺ T cells. Ultimately, HIV viral proteins and ART exposure contribute to chronic AT inflammation and excessive extracellular matrix accumulation.

Rhesus macaques with SIV infection are the most prevalent animal models used to study HIV pathogenesis. These models provided critical insights into the physiopathology of HIV-associated metabolic complications. The AT of SIV-infected rhesus macaques harbors infected immune cells early during infection, with subsequent defects in AT metabolism (236). Moreover, SIV DNA and RNA are present in AT stromal vascular fraction cells, CD4⁺ T cells, and CD14⁺ macrophages in SAT and VAT from SIV-infected macaques (237). Furthermore, in vitro cultures of AT stromal vascular cells from PLWH and macaques demonstrated the limited antiviral effectiveness of some NRTIs, particularly tenofovir, whereas the INSTI DTG penetrates AT and effectively inhibits viral replication (Fig. 2) (238). The main limitation of rhesus macaque models is the high costs of colony maintenance.

Mouse models helped formalize the notion that factors released by HIV provoke a confluence of chronic maladaptive responses in AT that result in dyslipidemia and insulin resistance. Transgenic mouse models for HIV accessory Vpr (239) and negative factor (Nef) (240) provided critical knowledge about virus-driven mechanisms of metabolic disease in PLWH. Subsequently, recombinant Vpr and Nef were developed, allowing in vivo studies via continuous intraperitoneal infusions (241, 242). The use of adenovirus-associated virus to expand the nature of viral protein expression may provide new tools for expressing HIV-specific factors in tissue-specific ways and allow diet, strain, and ART effects to be explored in vivo.

The phenotype of the Vpr mouse models recapitulates the critical aspects of HIV-associated metabolic disease in PLWH (241, 243). Enforced Vpr expression, via transgenic expression or continuous infusion of recombinant Vpr, in the skeletal muscle, liver, brain, and AT upregulates AT lipolysis and impairs adipocyte differentiation, leading to lipoatrophy with AT inflammation (241). In the liver, Vpr expression causes accelerated lipogenesis, impaired fatty acid oxidation, slowed fatty acid export, and steatohepatitis (243). As a result, the mice exhibit glucose intolerance and hypertriglyceridemia. Although mechanisms of Vpr-associated AT dysfunction are incompletely understood, Vpr animal models remain the most viable way to study metabolic complications related to HIV and ART.

The in vivo roles of Tat and Nef are less clear, with studies suggesting both inhibition and facilitation of adipocyte differentiation (244-246). Although the contributions of specific HIV viral proteins to metabolic derangement in PLWH remain to be fully delineated, the mechanistic data from experimental animal models and the presence of Vpr, Tat, and Nef in the plasma of PLWH on suppressive ART strongly suggest a causal role for HIV viral proteins in AT dysfunction.

AT is an Anatomic HIV Reservoir

Data from animal models set the foundation for experiments that defined HIV persistence in anatomic reservoirs of PLWH. ART inhibits HIV viral replication and partially restores immune function but does not fully eradicate HIV virions, leading to persistence in cellular and anatomic reservoirs. AT is an important reservoir for HIV harboring activated memory CD4⁺ T cells and HIV DNA, even in the setting of suppressive ART (Fig. 2) (247). In aviremic PLWH on ART, replication-competent HIV is present in CD4⁺ T cells of both SAT and VAT with a median latent HIV proviral DNA copy number equivalent to that of circulating and mesenteric lymph node CD4⁺ T cells (237). These data indicate that AT supports viral persistence and chronic immune activation and inflammation in ART-treated PLWH (Fig. 2).

Conclusions

PLWH manifest excessive ectopic AT that is most apparent in the liver, heart, and skeletal muscle and ultimately leads to end-organ damage from T2D, NAFLD, ASCVD, and heart failure. The increased severity of AT dysfunction among PLWH likely contributes to the 3-fold increase in the global burden of HIV-associated cardiovascular disease over the past 2 decades (8).

We lack a fundamental understanding of why ART and viral factors alter metabolism in PLWH because animal models are limited. In the transgenic Vpr mouse model, expression of the HIV accessory protein Vpr is sufficient to cause all the cardinal manifestations of HIV-associated metabolic disease (226, 241, 243, 248). Thus, the Vpr animal models are the most viable way to study metabolic complications related to HIV and ART. Animal models remain the most important preclinical tools for testing medical interventions and acquiring the basic scientific knowledge that will ultimately be needed to develop safe therapies to manage the long-term metabolic impact of HIV and ART. Thus, additional animal models and methods using contemporary gene editing and transgenic approaches (249) may enable the study of HIV-related risk factors that contribute to metabolic diseases. However, a critical limitation of rodent models in the study of immune activation in HIV resides in the biological differences between human and rodent gut barrier integrity, one of the underlying factors that mediate chronic inflammation even in virologically suppressed PLWH.

HIV-related factors and ART likely contribute to metabolic deterioration in PLWH, but the endocrine arms of these responses remain unclear. Although many biomarkers have been explored, the limitations of study design do not draw a clear picture, and interpretations are limited to purely correlative observations. Reports of inflammatory markers that go up or down in the serum of PLWH also abound. Emergent biomarkers derived from AT may also be relevant for predicting metabolic disease outcomes in PLWH. AT is a major contributor to the circulating small RNAs carried by small extracellular vesicles, which have regulatory functions in distant tissues (250). Although AT secretes extracellular vesicles that uniquely reflect lipodystrophy in PLWH (233, 234) and likely mediate responses in recipient cells and tissues, more studies will be needed to understand their endocrine functions and predictive power.

Significant pitfalls must be considered to identify valid mechanistic biomarkers in PLWH, including uniformity in measurement techniques and confounding factors related to age, BMI, and sex differences. For example, insulin resistance has been associated with lower serum adiponectin and higher triglycerides among men living with HIV and with hyperleptinemia among women living with HIV, respectively (251). In persons without HIV, NAFLD prevalence is higher in men than in premenopausal women, whereas NAFLD tends to become more common in women after menopause (252). Estrogen (253) may engender the differences in metabolic syndrome prevalence and liver cancer risk between men and women living with HIV. Unfortunately, most of the available data in PLWH comes disproportionately from studies of men living with HIV (254). Hence, the impact of sex hormones on HIV comorbidities might be underestimated. Considering that women and minorities living with HIV have a greater prevalence of obesity and are at greater risk of experiencing poorer outcomes from HIV treatment (255, 256), there is a clear need for biomarker studies of NAFLD and disease risk in these communities. Analysis of large cohort data, such as the MACS/WIHS Combined Cohort Study, with contemporary methodologies (eg, high-throughput sequencing and multi-omics approaches), will be needed to define and clarify biomarkers that predict metabolic disease in PLWH.

Abbreviations

ASCVD

atherosclerotic cardiovascular disease

ART

antiretroviral therapy

AT

adipose tissue

BIC

bictegravir

BMI

body mass index

CT

computed tomography

DICER a-KO

lacking Dicer in adipocytes

DRV

darunavir

DTG

dolutegravir

EFV

efavirenz

FFA

free fatty acid

FGF21

fibroblast growth factor 21

FTC

emtricitabine

GDF15

growth differentiation factor 15

GHD

GH deficiency

HCV

hepatitis C virus

HDL

high-density lipoprotein

IMAT

intermuscular adipose tissue

IMTC

intramyocardial triglyceride content

INSTI

integrase strand transfer inhibitor

LDL

low-density lipoprotein

miRNA

microRNA

NAFLD

non-alcoholic fatty liver disease

NASH

non-alcoholic steatohepatitis

Nef

negative factor

NNRTI

non-nucleoside/nucleotide reverse transcriptase inhibitor

NRTI

nucleoside/nucleotide reverse transcriptase inhibitor

PAI-1

plasminogen activator inhibitor-1

PI

protease inhibitor

PLWH

people living with HIV

RAL

raltegravir

REE

resting energy expenditure

SAT

subcutaneous adipose tissue

SIV

simian immunodeficiency virus

T2D

type 2 diabetes

TAF

tenofovir alafenamide

TDF

tenofovir disoproxil fumarate

VAT

visceral adipose tissue

Vpr

viral protein R

ZDV

zidovudine

Funding

This work was funded by the National Institutes of Health grants 5T32AI055413 (C.E.R.B), R01DK126042 (J.E.L), and R01DK114356 (S.M.H). C.E.R.B is also supported by the Baylor St. Luke's Roderick D. MacDonald Fellowship. A.B. is supported by the Rutherford Fund for Diabetes Research.

Disclosures

J.E.L. serves as a consultant to Theratechnologies. The remaining authors declared no conflicts of interest.