J Antimicrob Chemother

doi:10.1093/jac/dky383

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
Metabolic events in HIV-infected patients using abacavir are
associated with erythrocyte inosine triphosphatase activity
N. Chantal Peltenburg1,2, Jörgen Bierau3, Jolanda A. Schippers4, Selwyn H. Lowe5,6, Aimée D. C. Paulussen3,
Bianca J. C. van den Bosch3, Mathie P. G. Leers7, Eleni-Rosalina Andrinopoulou8, Jaap A. Bakker 9
and
1,2
Annelies Verbon *

1
Department of Internal Medicine, Division of Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands;
2
Department of Medical Microbiology and Infectious Diseases, Erasmus University Medical Center, Rotterdam, The Netherlands;
3
Department of Clinical Genetics, Maastricht University Medical Center, Maastricht, The Netherlands; 4Department of Integrated Care,
Maastricht University Medical Center, Maastricht, The Netherlands; 5Department of Internal Medicine, Division of Infectious Diseases,
Maastricht University Medical Center, Maastricht, The Netherlands; 6Department of Medical Microbiology, School of CAPHRI, Maastricht
University Medical Center, Maastricht, The Netherlands; 7Department of Clinical Chemistry & Hematology, Zuyderland Medical Center,
Heerlen, The Netherlands; 8Department of Biostatistics, Erasmus University Medical Center, Rotterdam, The Netherlands; 9Department
of Clinical Chemistry and Laboratory Medicine, Leiden University Medical Center, Leiden, The Netherlands

*Corresponding author. Medical Microbiology and Infectious Diseases, Erasmus University Medical Center, Postbus 2040, 3000 CA Rotterdam,
The Netherlands. Tel: !31107033511; Fax: !31107033875; E-mail: a.verbon@erasmusmc.nl

Received 23 May 2018; returned 23 July 2018; revised 2 August 2018; accepted 24 August 2018

Objectives: Abacavir use has been associated with an increased risk of cardiovascular disease (CVD) and meta-
bolic events in HIV-infected patients, although this finding was not consistently found. It is unclear whether
abacavir only increases this risk in subpopulations of HIV-infected patients. It may be hypothesized that inosine
50 -triphosphate pyrophosphohydrolase (ITPase), an enzyme involved in the metabolism of purine analogues
used in HIV treatment, plays a role in the risk of CVD and metabolic events in HIV-infected patients.
Methods: ITPase activity and ITPA genotype were determined in 393 HIV-infected patients. ITPase activity
,4 mmol IMP/mmol Hb/h was considered decreased. ITPA polymorphisms tested were: c.94C.A (rs1127354)
and c.124!21A.C (rs7270101). ORs were determined using generalized estimating equation models for devel-
oping CVD in patients who had ever been exposed to abacavir, tenofovir or didanosine and for developing meta-
bolic events in patients currently using these drugs.
Results: In patients using abacavir, metabolic events were associated with ITPase activity. No association was
demonstrated for tenofovir or didanosine. The OR for metabolic events was 3.11 in patients using abacavir with
normal ITPase activity (95% CI 1.34–7.21; P " 0.008) compared with patients with decreased ITPase activity
[adjusted for age, BMI, cumulative duration of combination ART (cART) use and the use of PI and NNRTI]. CVD
was not associated with ITPase activity or ITPA genotype.
Conclusions: This study shows, for the first time, that ITPase activity is associated with the occurrence of meta-
bolic events in patients using abacavir. Further studies are needed to confirm this association and to elucidate
the possible mechanism.

Introduction non-communicable comorbidities such as cardiovascular disease

(CVD) increase, especially among HIV-infected patients using com-
The life expectancy of patients infected with HIV has improved bination ART (cART).5–7 For CVD it is unclear what the contribution
substantially in recent decades1 and in developed countries has of the HIV infection itself is compared with the drugs that are part
been approaching that of the non-HIV-infected general popula- of the cART regimens. In a meta-analysis the pooled relative risk of
tion.2 In the Netherlands, currently, 42% of HIV patients are older CVD among HIV-infected patients without treatment was 1.61
than 50 years of age3 and in the USA, at year-end in 2015, the compared with people not infected with HIV.8 The pooled relative
group of patients aged 50–54 years made up the largest risk for HIV-infected patients on treatment was 2.00 compared
percentage of persons living with HIV (18%).4 With increasing age, with non-HIV-infected people.8 Different antiretroviral medication

C The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
V
For permissions, please email: journals.permissions@oup.com.
1 of 8
Peltenburg et al.

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
classes have been associated with a higher risk of CVD. The relative peripheral stent or bypass, coronary event (coronary artery bypass graft,
risk of CVD was 1.41 (95% CI 1.2–1.65) for patients on PI-based percutaneous intervention or myocardial infarction) or anamnestic com-
cART compared with those on non-PI-based cART.8 An increased plaints of peripheral claudication, during the use of cART. Metabolic events
rate of myocardial infarction in patients recently treated with aba- were defined as: dyslipidaemia as the reported reason for stopping the
cART regimen, laboratory-confirmed dyslipidaemia (total cholesterol
cavir and didanosine was first reported in the D:A:D study.9 Since
.6.5 mM, triglycerides .2.3 mM, LDL .4.5 mM), use of lipid-lowering ther-
then, several cohort studies and randomized clinical trials have apy (statins and/or fibrates), DM or HT. DM was defined as fasting glucose
confirmed the association of abacavir use and increased risk for 7 mM (126 mg/dL) measured on two or more consecutive occasions with-
CVD10–12 and in a meta-analysis the pooled relative risk was 1.8 in 3 months, or glucose 11.1 mM (200 mg/dL) measured once and in
(95% CI 1.43–2.26; P , 0.001) for patients treated with abacavir.8 combination with symptoms of hyperglycaemia, or if the patient was
Currently this association is still debated, as other studies have reported as being diabetic, or if the patient had glucose-lowering therapy.21
found either no effect or no significant effect of abacavir use on HT was defined as mean systolic pressure .140 mmHg, mean diastolic
the risk of CVD.13,14 pressure .90 mmHg or if the patient used anti-hypertensive medication.22
The pathogenic mechanism of the potential association be-
tween abacavir use and CVD is not clear. Which factors predispose
ITPase activity
patients using abacavir to a higher risk of CVD, apart from the clas-
sic risk factors such as dyslipidaemia, hypertension (HT) and smok- Erythrocyte ITPase activity was determined as described previously23 and
assessed by formation of inosine 50 -monophosphate (IMP) from ITP. ITPase
ing, has not yet been elucidated. It may be hypothesized that
activity was expressed as millimoles (mmol) of IMP formed from ITP in 1 h
differences in the metabolism of abacavir predispose to a higher per mmol haemoglobin (mmol IMP/mmol Hb/h). ITPase activity of
risk for CVD. The active metabolite of abacavir is carbovir triphos- 4 mmol IMP/mmol Hb/h was considered normal, which is the lowest
phate, which is a guanosine analogue. Guanosine 50 -triphosphate value within the 95% CI for ITPA wild-type (wt/wt) carriers.19,24
(GTP) is a low-affinity substrate for the enzyme inosine 50 -triphos-
phate pyrophosphohydrolase (ITPase) compared with the natural
substrate inosine 50 -triphosphate (ITP).15 ITPase is one of the scav- ITPA genotype analysis
enger enzymes eliminating the potentially cytotoxic or genotoxic ITPA genotype was analysed as described previously by Bierau et al.19 The
non-canonical nucleoside triphosphates from the nucleotide Wizard Genomic DNA purification kit (Promega, Madison, WI, USA) was
pool16 and is encoded by the ITPA gene on chromosome 20p used to isolate genomic DNA from whole blood. The DNA was genotyped
(OMIM #147520). The SNPs c.94C.A (p.Pro32Thr, NCBI rs1127354) using Sanger sequencing for the two ITPA polymorphisms; c.94C.A
and c.124!21A.C (NCBI rs7270101) in the ITPA gene can cause a (p.Pro32Thr, rs1127354) and c.124!21A.C (rs7270101). The genotype
was considered to be wild-type (wt/wt) when neither of the polymorphisms
decrease in ITPase activity.17 In HIV-infected patients, ITPase was
was detected. All sequences were evaluated by two independent labora-
found to have decreased expression in leucocytes, which was not
tory experts.
associated with ITPA genotype,18 and decreased activity in eryth-
rocytes in patients with ITPA genotype wt/wt or c.94C.A com-
pared with a control population with the same genotype.19 Statistical analysis
The natural role of ITPase, the different ITPase activity in HIV- All analyses were performed with IBM SPSS Statistics 21 (IBM Corporation,
infected patients and the potential role of ITPase in the metabol- NY, USA) and the statistical software package R (free download from www.
ism of the purine analogues used in HIV treatment, suggest that rproject.org) version 3.4.3. Pearson’s v2 tests, Fisher’s exact test, independ-
ITPase activity may play a role in the risk of CVD in patients using ent samples t-tests and Mann–Whitney U tests were used to determine sig-
abacavir. nificant differences.
We therefore determined whether ITPase activity and ITPA The occurrence of CVD was analysed using generalized estimating
genotype are associated with the occurrence of CVD and with risk equation models to account for repeated statement and adjusted for dif-
factors such as HT, hypercholesterolaemia and diabetes mellitus ferences in CVD risk profile: age at the time of changing the cART regimen,
the last measured BMI, smoking status (current or former smoker), pre-
(DM) during abacavir use in a cohort of HIV-infected patients.
existing CVD before the start of cART, HT, DM, dyslipidaemia and gender.
Further adjustment for cumulative total duration of cART and use of a PI or
Patients and methods an NNRTI at the time of the event was performed. Numerical variables
were standardized in order to have the same scale.
Patients The occurrence of a metabolic event was analysed using generalized
HIV-infected patients at the outpatient clinic of the Maastricht University estimating equation models to account for repeated statement and
Medical Center in Maastricht, The Netherlands, who were treated with adjusted for cumulative total duration of cART, age at the time of changing
cART, were included in this study after providing written informed consent. the cART regimen, the last BMI measured, use of a PI and use of an NNRTI
The data used for this study were collected as described previously.20 The at the time of the event. Numerical variables were standardized in order to
study was performed according to the Helsinki Declaration and approved have the same scale.
by the Medical Ethics Committee of the Maastricht University Medical Since tenofovir disoproxil fumarate (further referred to as tenofovir) and
Center, Maastricht, The Netherlands. didanosine are also purine analogues, their active metabolites being aden-
ine nucleotide analogues instead of guanosine nucleotide analogues such
as carbovir, the association of ITPase activity, ITPA genotype and CVD in
patients using these drugs in their cART regimen was determined in add-
Endpoints ition. For exposure to either abacavir, tenofovir or didanosine, in the analysis
The endpoints of the study were CVD and metabolic events. CVD was of CVD we used current (which means use at the time of the event) and
defined as: reported cerebrovascular accident (CVA) (transient ischaemic past exposure. Only current exposure was used in the analysis of metabolic
attack, haemorrhagic/ischaemic/unspecified CVA), arterial occlusion, events.

2 of 8
Metabolic events in HIV associated with ITPase JAC

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
When abacavir, tenofovir or didanosine were used concomitantly in one occurred, compared with 48.5% in patients with decreased ITPase
cART regimen, this regimen was excluded from the analysis. P values ,0.05 activity (P , 0.0001; Table 3). After adjusting for age, BMI, cumula-
were considered to be statistically significant. We did not correct for mul- tive duration of cART use and the use of PI and NNRTI, the OR for
tiple testing because our analyses were hypothesis driven. developing a metabolic event was statistically significantly higher
in patients with normal versus decreased ITPase activity (3.11,
Results 95% CI 1.34–7.21; P " 0.008) in abacavir-containing regimens. In
regimens containing tenofovir or didanosine, no association be-
Patient characteristics tween ITPase activity and the occurrence of metabolic events
After excluding regimens containing a combination of didanosine, could be demonstrated (Figure 1a). Also in regimens without a pur-
tenofovir or abacavir, 1422 regimens in 393 patients were used in ine analogue but in 60% containing a PI no association was found
the analysis.20 Of these patients, 52.2% had an ITPase activity of (Figure 1a).
,4 mmol IMP/mmol Hb/h (Table 1). No statistically significant dif- In regimens prescribed in patients with wt/wt genotype, meta-
ference was found between patients with decreased ITPase activ- bolic events were more frequently found compared with the other
ity and with normal ITPase activity with regard to age, gender, genotypes (537/960 regimens versus 202/435 regimens;
race, CD4 nadir counts and PI and NNRTI use. In 60.1% of regimens P " 0.001) (Table 3). Metabolic events occurred more frequently in
containing no purine, a PI was part of the cART regimen; this was, patients using abacavir and having the wt/wt genotype
respectively, 34.5%, 22.9% and 60.2% in regimens containing (P " 0.003). This increase could not be found in patients using
tenofovir, abacavir or didanosine. Current exposure to the purine tenofovir or didanosine (Table 3). However, after adjusting for age,
analogues tenofovir, abacavir and didanosine was frequent BMI, cumulative cART use and the use of PI or NNRTI, no significant
[n " 601 (42.3%), n " 244 (17.2%) and n " 128 (9.0%) respective- difference in metabolic events between patients using abacavir
ly]. The number of regimens with current plus past exposure to a with wt/wt genotype and the other ITPA genotypes could be dem-
purine analogue was 699 (49.2%) for tenofovir, 465 (32.7%) for onstrated (Figure 1b).
abacavir and 365 (25.7%) for didanosine. CVD was present in
12.5% of the patients (n " 49) and reported to be CVA (n " 15), ar- Discussion
terial occlusion (n " 2), peripheral stent or bypass (n " 6), coronary
arterial bypass graft (n " 6), percutaneous intervention (n " 12), In this study we show, for the first time, that ITPase activity may
myocardial infarction (n " 7) or anamnestic complaints of periph- play a role in the association of abacavir use and the occurrence of
eral claudication (n " 1). In patients with current or past exposure metabolic events in HIV-infected patients using cART. In regimens
to abacavir, 24 CVD events occurred; CVA (n " 7), coronary artery containing abacavir decreased ITPase activity was associated with
bypass grafting (n " 3), percutaneous coronary intervention fewer metabolic adverse events. We could not confirm the hy-
(n " 7), myocardial infarction (n " 2), arterial occlusion (n " 2) and pothesis that decreased ITPase activity decreased the risk of CVD.
peripheral stent or bypass (n " 3). In 749 regimens (52.7%) a No association between ITPA genotype and CVD or metabolic
metabolic event was found. An overview of different components events was found after adjustment for other CVD risk factors.
of the combined endpoint metabolic event is presented in Table 2. In multiple studies it was reported that the use of abacavir
increased the risk of developing CVD or myocardial infarction.9–11
Other studies, however, could not confirm this association.14,25 We
Effect of ITPase activity and ITPA genotype on CVD found no association between abacavir use, ITPase activity and
No significant difference in CVD between the regimens used in CVD events, possibly owing to the relatively small number of CVD
patients with decreased versus normal ITPase activity was found, events found in our study. Like others, we found that metabolic
irrespective of cART regimen used (22 versus 27 events, respective- events were associated with abacavir, but not with tenofovir
ly; crude P " 0.34). ITPase activity in combination with current or use.26–30 A possible explanation may be the difference in chemical
past exposure to abacavir was not associated with a higher num- structure since tenofovir is an adenosine analogue and abacavir is
ber of CVD events compared with tenofovir or didanosine exposure a guanosine analogue.
in our cohort (Table 3). How ITPase might be involved in the occurrence of metabolic
Regardless of cART regimen, CVD more often occurred in events during abacavir use is not known. Carbovir triphosphate
patients with ITPA genotype wt/wt versus the other ITPA geno- was found not to be a direct substrate for ITPase.20 Potentially the
types [CVD: n " 35 (71.4%) versus n " 13 (26.5%) respectively; al- answer lies in cellular signal transduction. Guanosine 30 ,50 -cyclic
though this difference was not statistically significant (P " 0.72)] monophosphate (cGMP) is an important secondary messenger
(Table 3). One CVD occurred in a patient with unknown ITPA geno- that modulates multiple cellular processes, such as platelet aggre-
type. Adjusting the data for CVD risk profiles, cumulative cART use gation, neurotransmission, blood pressure, lipolysis and gut peri-
and PI and NNRTI use did not change the outcome for CVD. stalsis.31,32 cGMP is produced from GTP by the enzyme soluble
guanylate cyclase (sGC)33,34 and stimulation of sGC was found to
be protective against obesity.35 However, in the presence of nitric
Effect of ITPase activity and ITPA genotype on oxide (NO) and magnesium ions (Mg2!) sGC shifts its substrate
metabolic events specificity to ITP to form inosine 30 ,50 -cyclic monophosphate
Of all the metabolic events, 52.5% occurred during regimens pre- (cIMP) and also, in the absence of hypoxaemia, addition of ex-
scribed to patients with normal ITPase activity (crude P " 0.001) ogenous ITP to intact porcine arteries led to higher cIMP levels.36,37
(Table 3). In 72.6% of the regimens currently containing abacavir In contrast to cGMP, which induces vasodilatation, cIMP induces
used by patients with normal ITPase activity, a metabolic event vasoconstriction.36 This vasoconstriction has been well studied by

3 of 8
Peltenburg et al.

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
Table 1. Demographic and clinical characteristics of the patients (n " 393) with ITPase activity ,4 and 4 mmol IMP/mmol Hb/h

ITPase activity (mmol IMP/mmol Hb/h)

Characteristic ,4 (n " 205) 4 (n " 188) P value

a
Age, years, mean (SD) 50.1 (11.1) 49.7 (11.9) 0.78
Male, n (%)a 164 (80.0) 155 (82.4) 0.53
Race, n (%)a 0.81
white non-Hispanic 164 (80.0) 147 (78.2)
white Hispanic 5 (2.4) 4 (2.1)
black 22 (10.7) 27 (14.4)
Asian or other 14 (6.9) 10 (5.3)
ITPase activity, mean (SD)a 2.44+1.12 5.24+ 1.09 ,0.001
ITPA genotype, n (%)a ,0.001
wt/wt 90 (43.9) 175 (93.1)
wt/c.124 ! 21A.C 59 (28.8) 9 (4.8)
wt/c.94C.A or otherb 53 (25.9) —
unknown 3 (1.5) 4 (2.1)
Alcohol (IU/day)a 0.51
,2 157 (76.6) 133 (70.7)
2 35 (17.1) 36 (19.1)
unknown 13 (6.3) 19 (10.1)
Smoking, n (%)a 0.65
never 18 (8.8) 21 (11.2)
current or former 130 (63.4) 116 (61.7)
not current, unknown if ever 55 (26.8) 46 (24.5)
unknown 2 (1.0) 5 (2.7)
BMI, kg/m2, n (%)a 0.09
underweight (,18.5) 14 (6.8) 3 (1.6)
normal (18.5 and ,25) 102 (49.8) 95 (50.5)
mild/moderate obesity (25 and ,30) 59 (28.8) 52 (27.7)
severe obesity (30) 10 (4.9) 15 (8.0)
unknown 20 (9.8) 23 (12.2)
CD4 nadir, cells/mm3, median (min–max)a 207 (1–1022) 209 (3–612) 0.40
Year of start, median (min–max)a 2006 (1987–2013)c 2006 (1987–2013)d 0.25
Total number of cART regimensa 734 688
Type of cART, n (%)a 0.09
NNRTI 342 (46.6) 295 (42.9)
PI 239 (32.6) 267 (38.8)
NNRTI and PI 59 (8.0) 45 (6.5)
other 94 (12.8) 81 (11.8)

a
Previously published data.
b
Other indicates homozygous c.124!21A.C or homozygous c.94C.A or heterozygous c.124!21A.C/c.94C.A.
c
Year of start unknown for one patient.
d
Year of start unknown for two patients.

measurement of flow-mediated dilatation (FMD) of the brachial competed with GTP or a shift in substrate specificity of sGC
artery and is strongly associated with the risk of CVD.38–40 Patients occurred. New research is warranted to investigate under which
using abacavir were found to have lower FMD than patients not circumstances ITP is a substrate for sGC, what the consequences
using abacavir,41 whereas didanosine and tenofovir use were not of this shift in substrate specificity are and what the impact of car-
associated with lower FMD. Support for the hypothesis that abaca- bovir triphosphate is on sGC activity in human cells.
vir shifts sGC to use ITP instead of GTP, leading to higher cIMP lev- In 2010 Fellay et al.17 showed that the SNPs 94C.A and
els, can be found in the study by Baum et al.42 Formation of cGMP 124!21A.C in the ITPA genotype were associated with the pro-
was decreased in human platelets incubated with carbovir triphos- tection against haemolytic anaemia during treatment with riba-
phate (the active metabolite of abacavir) and the authors con- virin for hepatitis C infection. We found ITPase might be a more
cluded that carbovir triphosphate inhibited sGC. The question accurate predictor of the development of haemolytic anaemia in
remains whether sGC was inhibited, carbovir triphosphate these patients than ITPA genotype.43 In HIV-infected patients

4 of 8
Metabolic events in HIV associated with ITPase JAC

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
Table 2. DM, dyslipidaemia and use of blood pressure medication or cholesterol-lowering therapy for regimens used in patients with decreased and
normal ITPase activity

ITPase activity (mmol IMP/mmol Hb/h)

all regimens regimens currently containing abacavir

,4 (734 regimens) 4 (688 regimens) P ,4 (131 regimens) 4 (113 regimens) P

DM
yes 28 (3.8) 45 (6.5) 0.03 3 (2.3) 11 (9.7) 0.04
unknown — 18 (2.6) — —
HT
yes 146 (19.9) 171 (24.9) 0.03 26 (19.8) 31 (27.4) 0.16
unknown — — — —
Dyslipidaemia
yes 195 (26.6) 226 (32.8) 0.01 40 (30.5) 44 (38.9) 0.17
unknown — — — —
Cholesterol-lowering therapy
yes 86 (11.7) 117 (17.1) 0.014 20 (15.3) 23 (20.4) 0.68
unknown 10 (1.4) 9 (1.3) 1 (0.8) 1 (0.9)

Results are shown as n (%).

Table 3. CVD and metabolic events for regimens used in patients with decreased and normal ITPase activity and different ITPA genotypes

ITPase activity (mmol

IMP/mmol Hb/h) P value ITPA genotype P value
a
,4 4 crude adjusted other wt/wt crude adjusted

CVD
total (N " 1422)b 22/734 27/688 0.34 0.82 13/435 35/960 0.72 0.86
tenofovir (n " 699)c 12/354 17/345 0.31 0.50 7/211 21/478 0.35 0.77
abacavir (n " 465)d 11/245 13/220 0.49 0.40 6/132 18/328 0.86 0.41
didanosine (n " 365)e 2/165 6/200 0.30 0.45 0/91 8/272 0.24 NA
Metabolic events
total (N " 1422)b 356/734 393/688 0.001 0.12 202/435 537/960 0.001 0.14
tenofovir (n " 601)f 155/306 168/295 0.12 0.55 89/186 231/409 0.13 0.53
abacavir (n " 244)d 63/131 82/113 ,0.0001 0.008 39/75 106/164 0.003 0.24
didanosine (n " 128)g 25/51 37/77 0.92 0.75 12/31 49/96 0.26 0.61

Results are shown as number of events/number of regimens. NA, not applicable.

a
Heterozygous wt/c.124!21A.C or wt/c.94C.A or homozygous c.124!21A.C or homozygous c.94C.A or compound heterozygous
c.124!21A.C/c.94C.A.
b
Genotype unknown in 27 regimens.
c
Genotype unknown in 10 regimens.
d
Genotype unknown in five regimens.
e
Genotype unknown in two regimens.
f
Genotype unknown in six regimens.
g
Genotype unknown in one regimen.

both ITPase activity and ITPA genotype were associated with ad- ITPase activity versus ITPA genotype in HIV-infected patients may
verse events; however, ITPA genotype correlated less well with ad- be explained by the finding that in these patients the wt/wt and
verse events compared with ITPase activity.20 In the present study the c.94C.A carriers have decreased erythrocyte ITPase activity
again, we found ITPA genotype and ITPase activity were both compared with non-HIV-infected patients carrying the same
crudely associated with adverse events; however, after adjusting genotypes.19
with the logistic mixed effects model, only ITPase activity Here, we report a lower risk of metabolic events in HIV-infected
remained statistically significant. The difference in the effect of patients that use abacavir and have decreased ITPase activity. In a

5 of 8
Peltenburg et al.

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
(a) Effect of ITPase activity on metabolic events
OR (95% CI) P values

Total 1.34 (0.93–1.94) 0.12

Tenofovir 1.17 (0.70–1.95) 0.55

Abacavir 3.11 (1.34–7.21) 0.008

Didanosine 0.86 (0.36–2.07) 0.75

No purine 1.14 (0.64–2.03) 0.66

10
1
0.

m o
More adverse events More adverse events
in ITPase activity 4 in ITPase activity t4

(b)
Effect of ITPA genotype on metabolic events
OR (95% CI) P values

Total 1.35 (0.90–2.01) 0.14

Tenofovir 1.18 (0.70–2.01) 0.53

Abacavir 1.68 (0.71–3.99) 0.24

Didanosine 1.27 (0.51–3.18) 0.61

No purine 1.42 (0.75–2.69) 0.29

1

10
1
0.

m o
More adverse events More adverse events
in other ITPA genotypes in ITPA genotype wt/wt

Figure 1. Effect of ITPase activity (a) and ITPA genotype (b) on metabolic events. The effect of decreased versus normal ITPase activity (a) and of
other ITPA genotypes versus ITPA genotype wt/wt (b) on the occurrence of metabolic events are plotted for all regimens (total), for regimens contain-
ing tenofovir, abacavir or didanosine and for regimens without tenofovir, abacavir or didanosine (no purine). OR with 95% CI and matching P values
are displayed.

previous study, in tenofovir-containing regimens, decreased decrease in adverse events in the other study. In other studies the
ITPase activity was also associated with a lower risk of adverse association between decreased ITPase activity or an SNP in the
events.20 In contrast, the use of abacavir in this previous study was ITPA genotype and adverse events was shown to be dependent on
crudely associated with an increase in adverse events in patients the kind of adverse event and the drug that was used, leading to,
having decreased ITPase activity. The adverse events in that for instance, an increase in hepatic toxicity (during azathioprine or
study were different from the current study and included gastro- 6-mercaptopurine use) but a decrease in haemolytic anaemia
intestinal, neurological, renal, skin and liver-related adverse (during ribavirin use).44–47 We hypothesize that the effect of
events, and potentially this explains why decreased ITPase activity ITPase activity in patients using abacavir is also dependent on the
may lead to an increase in adverse events in one study, but a nature of the adverse events.

6 of 8
Metabolic events in HIV associated with ITPase JAC

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
There are limitations that need to be mentioned. Our study was patients enrolled in the D:A:D study: a multi-cohort collaboration. Lancet
retrospective with all the possible caveats of retrospective studies. 2008; 371: 1417–26.
For example, factors other than the cART regimen could have con- 10 Marcus JL, Neugebauer RS, Leyden WA et al. Use of abacavir and risk of
tributed to alterations in metabolic events. This is, however, not dif- cardiovascular disease among HIV-infected individuals. J Acquir Immune
ferent for patients with decreased or patients with normal ITPase Defic Syndr 2016; 71: 413–9.
activities, and thus the influence of this factor seems limited. 11 Young J, Xiao Y, Moodie EE et al. Effect of cumulating exposure to abacavir
Although our results could be mechanistically made plausible, fur- on the risk of cardiovascular disease events in patients from the Swiss HIV co-
hort study. J Acquir Immune Defic Syndr 2015; 69: 413–21.
ther prospective studies are still warranted to confirm our results.
Owing to the number of patients included, a low frequency of car- 12 Rotger M, Glass TR, Junier T et al. Contribution of genetic background,
traditional risk factors, and HIV-related factors to coronary artery disease
diovascular events was found. This may explain the lack of associ-
events in HIV-positive persons. Clin Infect Dis 2013; 57: 112–21.
ation of CVD with ITPase activity. Also, the clinical relevance of our
13 Lang S, Mary-Krause M, Cotte L et al. Impact of individual antiretro-
findings is to be further established by larger prospective trials.
viral drugs on the risk of myocardial infarction in human immunodefi-
In conclusion, we showed that ITPase activity may be associ-
ciency virus-infected patients: a case-control study nested within the
ated with a risk of metabolic events in HIV-infected patients using French Hospital Database on HIV ANRS cohort CO4. Arch Intern Med
abacavir. Using ITPase activity as a potential biomarker to predict 2010; 170: 1228–38.
adverse metabolic events may be a further step towards more 14 Ding X, Andraca-Carrera E, Cooper C et al. No association of abacavir use
patient-tailored medicine in the future. However, elucidation of with myocardial infarction: findings of an FDA meta-analysis. J Acquir
the pathogenic mechanism needs further studies and our results Immune Defic Syndr 2012; 61: 441–7.
need to be confirmed in a prospective trial. 15 Lin S, McLennan AG, Ying K et al. Cloning, expression, and characterization
of a human inosine triphosphate pyrophosphatase encoded by the ITPA
gene. J Biol Chem 2001; 276: 18695–701.
Funding 16 Galperin MY, Moroz OV, Wilson KS et al. House cleaning, a part of good
housekeeping. Mol Microbiol 2006; 59: 5–19.
This work was supported by an unrestricted scientific grant by Janssen
pharmaceutical companies of Johnson & Johnson to A. V. 17 Fellay J, Thompson AJ, Ge D et al. ITPA gene variants protect against an-
aemia in patients treated for chronic hepatitis C. Nature 2010; 464: 405–8.
18 Peltenburg NC, Leers MP, Bakker JA et al. Inosine triphosphate pyrophos-
phohydrolase expression: decreased in leukocytes of HIV-infected patients
Transparency declarations using combination antiretroviral therapy. J Acquir Immune Defic Syndr 2016;
None to declare. 73: 390–5.
19 Bierau J, Bakker JA, Schippers JA et al. Erythrocyte inosine triphosphatase
activity is decreased in HIV-seropositive individuals. PLoS One 2012; 7: e30175.
References 20 Peltenburg NC, Bierau J, Bakker JA et al. Erythrocyte inosine triphospha-
1 GBD 2015 Mortality and Causes of Death Collaborators. Global, regional, tase activity: a potential biomarker for adverse events during combination
and national life expectancy, all-cause mortality, and cause-specific mortal- antiretroviral therapy for HIV. PLoS One 2018; 13: e0191069.
ity for 249 causes of death, 1980–2015: a systematic analysis for the Global 21 WHO. Definition and Diagnosis of Diabetes Mellitus and Intermediate
Burden of Disease Study 2015. Lancet 2016; 388: 1459–544. Hyperglycemia: Report of a WHO/IDF Consultation. Geneva: WHO, 2006.
2 Nakagawa F, May M, Phillips A. Life expectancy living with HIV: recent esti- 22 Whitworth JA, World Health Organization, International Society
mates and future implications. Curr Opin Infect Dis 2013; 26: 17–25. ofHypertension Writing Group. 2003 World Health Organization (WHO)/
3 Sighem A, Gras L, Smit C et al. Monitoring Report 2015. Monitoring of International Society of Hypertension (ISH) statement on management of
Human Immunodeficiency Virus (HIV) Infection in The Netherlands. 2015. hypertension. J Hypertens 2003; 21: 1983–92.
https://www.hiv-monitoring.nl/files/1014/4791/9393/Monitoring_Report_ 23 Bierau J, Bakker JA, Lindhout M et al. Determination of ITPase activity in
2015_19_nov.pdf. erythrocyte lysates obtained for determination of TPMT activity. Nucleosides
4 Centers for Disease Control and Prevention. HIV Surveillance Report, 2016. Nucleotides Nucleic Acids 2006; 25: 1129–32.
http://www.cdc.gov/hiv/library/reports/hiv-surveillance.html. 24 Shipkova M, Lorenz K, Oellerich M et al. Measurement of erythrocyte in-
5 Schouten J, Wit FW, Stolte IG et al. Cross-sectional comparison of the osine triphosphate pyrophosphohydrolase (ITPA) activity by HPLC and correl-
prevalence of age-associated comorbidities and their risk factors between ation of ITPA genotype-phenotype in a Caucasian population. Clin Chem
HIV-infected and uninfected individuals: the AGEhIV cohort study. Clin Infect 2006; 52: 240–7.
Dis 2014; 59: 1787–97. 25 Obel N, Farkas DK, Kronborg G et al. Abacavir and risk of myocardial in-
6 Shahmanesh M, Schultze A, Burns F et al. The cardiovascular risk manage- farction in HIV-infected patients on highly active antiretroviral therapy: a
ment for people living with HIV in Europe: how well are we doing? AIDS 2016; population-based nationwide cohort study. HIV Med 2010; 11: 130–6.
30: 2505–18. 26 Hill A, Sawyer W, Gazzard B. Effects of first-line use of nucleoside ana-
7 Friis MN, Sabin CA, Weber R et al; The Data Collection on Adverse Events of logues, efavirenz, and ritonavir-boosted protease inhibitors on lipid levels. HIV
Anti-HIV Drugs (DAD) Study Group. Combination antiretroviral therapy and Clin Trials 2009; 10: 1–12.
the risk of myocardial infarction. N Engl J Med 2003; 349: 1993–2003. 27 Gagliardini R, Fabbiani M, Colafigli M et al. Lipid-lowering effect and
8 Islam FM, Wu J, Jansson J et al. Relative risk of cardiovascular disease changes in estimated cardiovascular risk after switching to a tenofovir-
among people living with HIV: a systematic review and meta-analysis. HIV containing regimen for the treatment of HIV-infected patients. J Chemother
Med 2012; 13: 453–68. 2016; 29: 1–9.
9 Sabin CA, Worm SW et al; D:A:D Study Group. Use of nucleoside reverse 28 Arae H, Tateyama M, Nakamura H et al. Evaluation of the lipid concentra-
transcriptase inhibitors and risk of myocardial infarction in HIV-infected tions after switching from antiretroviral drug tenofovir disoproxil fumarate/

7 of 8
Peltenburg et al.

Downloaded from https://academic.oup.com/jac/advance-article-abstract/doi/10.1093/jac/dky383/5124335 by The University of British Columbia Library user on 14 October 2018
emtricitabine to abacavir sulfate/lamivudine in virologically-suppressed human 39 Celermajer DS, Sorensen KE, Gooch VM et al. Non-invasive detection of
immunodeficiency virus-infected patients. Intern Med 2016; 55: 3435–40. endothelial dysfunction in children and adults at risk of atherosclerosis.
29 Moyle GJ, Orkin C, Fisher M et al. A randomized comparative trial of con- Lancet 1992; 340: 1111–5.
tinued abacavir/lamivudine plus efavirenz or replacement with efavirenz/ 40 Gokce N, Keaney JF Jr, Hunter LM et al. Predictive value of noninvasively
emtricitabine/tenofovir DF in hypercholesterolemic HIV-1 infected individu- determined endothelial dysfunction for long-term cardiovascular events in
als. PLoS One 2015; 10: e0116297. patients with peripheral vascular disease. J Am Coll Cardiol 2003; 41:
30 Campo R, DeJesus E, Bredeek UF et al. SWIFT: prospective 48-week study 1769–75.
to evaluate efficacy and safety of switching to emtricitabine/tenofovir from 41 Hsue PY, Hunt PW, Wu Y et al. Association of abacavir and impaired endo-
lamivudine/abacavir in virologically suppressed HIV-1 infected patients on a thelial function in treated and suppressed HIV-infected patients. AIDS 2009;
boosted protease inhibitor containing antiretroviral regimen. Clin Infect Dis 23: 2021–7.
2013; 56: 1637–45. 42 Baum PD, Sullam PM, Stoddart CA et al. Abacavir increases platelet re-
31 Friebe A, Sandner P, Seifert R. From bedside to bench—meeting report of activity via competitive inhibition of soluble guanylyl cyclase. AIDS 2011; 25:
the 7th International Conference on cGMP “cGMP: generators, effectors and 2243–8.
therapeutic implications” in Trier, Germany, from June 19th to 21st 2015. 43 Peltenburg NC, Bakker JA, Vroemen WH et al. Inosine triphosphate pyro-
Naunyn Schmiedebergs Arch Pharmacol 2015; 388: 1237–46. phosphohydrolase activity: more accurate predictor for ribavirin-induced an-
32 Potter LR. Guanylyl cyclase structure, function and regulation. Cell Signal emia in hepatitis C infected patients than ITPA genotype. Clin Chem Lab Med
2011; 23: 1921–6. 2015; 53: 2021–9.
33 Katsuki S, Arnold W, Mittal C et al. Stimulation of guanylate cyclase by so- 44 Shipkova M, Franz J, Abe M et al. Association between adverse effects
dium nitroprusside, nitroglycerin and nitric oxide in various tissue prepara- under azathioprine therapy and inosine triphosphate pyrophosphatase activ-
tions and comparison to the effects of sodium azide and hydroxylamine. J ity in patients with chronic inflammatory bowel disease. Ther Drug Monit
Cyclic Nucleotide Res 1977; 3: 23–35. 2011; 33: 321–8.
34 Hardman JG, Robison GA, Sutherland EW. Cyclic nucleotides. Annu Rev 45 Buster EH, van Vuuren HJ, Zondervan PE et al. Thiopurine-methyl-
Physiol 1971; 33: 311–36. transferase and inosine triphosphate pyrophosphatase polymorphism
35 Hoffmann LS, Etzrodt J, Willkomm L et al. Stimulation of soluble guanylyl in a liver transplant recipient developing nodular regenerative hyperpla-
cyclase protects against obesity by recruiting brown adipose tissue. Nat sia on low-dose azathioprine. Eur J Gastroenterol Hepatol 2008; 20:
Commun 2015; 6: 7235. 68–72.
36 Chen Z, Zhang X, Ying L et al. cIMP synthesized by sGC as a mediator of 46 Wan Rosalina WR, Teh LK, Mohamad N et al. Polymorphism of ITPA
hypoxic contraction of coronary arteries. Am J Physiol Heart Circ Physiol 2014; 94C.A and risk of adverse effects among patients with acute lymphoblastic
307: H328–36. leukaemia treated with 6-mercaptopurine. J Clin Pharm Ther 2012; 37:
237–41.
37 Beste KY, Burhenne H, Kaever V et al. Nucleotidyl cyclase activity of sol-
uble guanylyl cyclase a1b1. Biochemistry 2012; 51: 194–204. 47 Tanaka Y, Manabe A, Nakadate H et al. The activity of the inosine triphos-
phate pyrophosphatase affects toxicity of 6-mercaptopurine during main-
38 Neunteufl T, Heher S, Katzenschlager R et al. Late prognostic value of
flow-mediated dilation in the brachial artery of patients with chest pain. Am J tenance therapy for acute lymphoblastic leukemia in Japanese children. Leuk
Cardiol 2000; 86: 207–10. Res 2012; 36: 560–4.

8 of 8