Foscarnet salvage therapy efficacy is associated with the presence of thymidine-associated mutations (TAMs) in HIV-infected patients

Journal of Clinical Virology 43 (2008) 212–215

Contents lists available at ScienceDirect

Journal of Clinical Virology

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j c v

Foscarnet salvage therapy efficacy is associated with the presence of thymidine-associated mutations (TAMs) in HIV-infected patients ^夽

Charlotte Charpentier ^a,b , Didier Laureillard ^c , Mustapha Sodqi ^c , Ali Si-Mohamed ^a , Marina Karmochkine ^c , Laurent Bélec ^{a , b} , Laurence Weiss ^{b , c} , Christophe Piketty ^{b , c ,∗}

aLaboratoire de Virologie, AP-HP, Hôpital Européen Georges Pompidou, Paris, France

bUniversité Paris-Descartes, Faculté de Médecine, Paris, France

cService d’Immunologie Clinique, AP-HP, Hôpital Européen Georges Pompidou, Paris, France

a r t i c l e i n f o

Article history:

Received 22 September 2007 Received in revised form 1 July 2008 Accepted 1 July 2008

Keywords:

HIV Foscarnet Salvage therapy

Thymidine-associated mutations

a b s t r a c t

Background:

Salvage therapy based on foscarnet plus a thymidine analog is effective in patients with advanced-stage HIV disease and viruses harbouring multiple drug-resistance mutations.

Objective:To identify viral genetic determinants associated with the virological efficacy of foscarnet salvage

therapy.

Study design:

Thirteen patients received foscarnet at a fixed dose of 80 mg/kg twice daily for 14 days, in combination with zidovudine or stavudine.

Results:

The baseline median HIV viral load and CD4 cell count were 5.10 log

copies/ml and 23 cells/mm

, respectively. Following foscarnet therapy, viral load fell by a median of 1.84 log

copies/ml (range:

to

copies/ml in 11 patients, all of whom harboured viruses with at least three thymidine-associated mutations (TAMs). The two patients with smaller declines in viral load (<0.50 log

copies/ml) harboured viruses with only one or zero TAMs.

Conclusions:

These findings corroborate,

The virological response to foscarnet salvage therapy in multiclass-experienced patients may thus differ according to the number of TAMs.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Foscarnet is a pyrophosphate (PP

) analog with activity on various viral DNA polymerases in vitro, including HIV reverse tran- scriptase (RT). Foscarnet is mainly used to treat cytomegalovirus (CMV) infection. Ten years ago, HIV plasma viral load was observed to fall during foscarnet treatment of CMV retinitis.

However, the advent of highly active antiretroviral therapy (HAART), together with the need for intravenous administration and frequent nephro- toxicity, restricted the evaluation of foscarnet in the treatment of HIV infection.

Foscarnet has been used as salvage therapy for patients with late-stage HIV infection and few other therapeutic options. Two studies involved patients with low CD4 cell counts, virologi-

夽Presented in part: XVI International HIV Drug Resistance Workshop, Los Barba- dos, West Indies, June 2007; Abstract 116.

∗Corresponding author at: Département d’Immunologie, Hôpital Européen Georges Pompidou, 20 rue Leblanc, 75015 Paris, France. Tel.: +33 1 56 09 27 01;

fax: +33 1 56 09 28 59.

E-mail address:christophe.piketty@egp.aphp.fr(C. Piketty).

cal failure, and multidrug-resistant (MDR) viruses.

Mathiesen et al. reported that HIV plasma viral load fell by a median of 1.80 log

copies/ml (range: − 1.20 to − 3.20) after 2 weeks (W2) of foscarnet therapy in seven patients.

More recently, Canestri et al.

obtained similar results in 11 patients, with a median decrease of 1.99 log

copies/ml (range: − 0.50 to − 2.49) at week 2 of foscarnet- zidovudine (AZT) combination therapy.

Being a PP

analog, foscarnet acts as a substrate for phosphoroly- sis and is able to block the excision of chain-terminating nucleotides catalysed by HIV RT.

Excision of chain-terminating nucleotides is the main mechanism of resistance to thymidine analogs (AZT and d4T). TAMs increase the capacity of RT to remove thymidine analogs from growing DNA chains. A negative correlation has been observed between the degree of HIV resistance to AZT and to foscarnet.

TAMs confer resistance to AZT but also hypersusceptibility to fos- carnet in vitro. In addition, mutations associated with foscarnet resistance can reduce or even overcome phenotypic resistance to AZT among viruses harbouring TAMs.

Here we evaluated foscarnet, in combination with a thymi- dine analog, as salvage therapy in 13 HIV-infected patients with severe immunodeficiency, multiple virological failure, and MDR viruses.

1386-6532/$ – see front matter © 2008 Elsevier B.V. All rights reserved.

doi:10.1016/j.jcv.2008.07.001

C. Charpentier et al. / Journal of Clinical Virology 43 (2008) 212–215 213

2. Patients and methods

The 13 HIV-infected patients selected for this study had been heavily treated and had few remaining antiretroviral (ARV) treat- ment options. They received foscarnet intravenously at a fixed dose of 80 mg/kg twice daily for 14 days, plus a thymidine analog (usu- ally AZT) in order to enhance foscarnet activity by maintaining antiretroviral drug pressure on TAMs. The aim was to reduce plasma HIV viral load, just before introducing an optimized ARV regimen comprising only one or two new, fully active drugs (enfuvirtide, darunavir, etravirine or raltegravir).

All the patients also received AZT (300 mg twice daily), either alone (patients A and C) or together with the previous failing ARV regimen. All the patients were hospitalized and received close clinical and biological monitoring. Plasma HIV-RNA load was mea- sured with a commercial assay (COBAS

Taqman HIV-1 Test, Roche Molecular Systems, Branchburg, NJ), twice weekly and at the end of treatment ( ± 2 days). A virological response was defined as a fall of at least 1 log

in viral load on day 14 of foscarnet therapy.

The protease and RT (codons 1–330) regions were sequenced in all the patients at baseline, using the ViroSeq HIV-1 genotyping system (Celera Diagnostics, Alameda, CA).

3. Results

The 13 patients (10 men, 3 women) were enrolled between November 2005 and July 2007. Their median age was 40 years (range: 34–52). All had been extensively treated, with a median of 14 ARV (range: 8–17) over a median of 11 years (range: 6–14).

Nine patients had AIDS. Eleven patients were infected by HIV sub- type B strains. Patient A had subtype CRF02 AG infection, and patient H was infected by a complex recombinant between sub- types CRF15 01B and A1. At baseline, all the patients harboured MDR viruses, exhibiting a median of 22 genotypic resistance muta- tions (range: 17–31), including five nucleoside reverse transcriptase inhibitor (NRTI) resistance-associated mutations (range: 3–8), two non-nucleoside reverse transcriptase inhibitor (NNRTI) resistance mutations (range: 0–4), and 15 protease inhibitor major and minor resistance mutations (range: 13–22). Two patients (D and M) had no NNRTI resistance mutations at baseline, but both had a history of such mutations and of virological failure during NNRTI treatment.

The patients’ characteristics are listed in Table 1. The median baseline plasma HIV-RNA level was 5.10 log

copies/ml (range:

4.52–5.95) and the median baseline CD4 cell count was 23/mm

(range: 0–109).

The median change in the HIV-RNA level at the end of the study treatment (day 14) was − 1.84 log

copies/ml (range: − 0.29 to − 2.82).

Two patterns of virological response to foscarnet combina- tion therapy were observed (Fig. 1). Patients A–K (group 1) had marked responses with a median decrease in HIV viral load of 1.97 log

copies/ml (range: − 1 to − 2.82) on day 14.

Patients L and M (group 2) had only weak responses ( − 0.29 and − 0.49 log

copies/ml, respectively). In patient L, plasma viral load rebounded during foscarnet therapy after an initial fall of 0.88 log

copies/ml on day 10. A significant difference in the decrease in HIV viral load between the two groups was observed (p = 0.029, Mann–Whitney test).

The results of direct RT gene sequencing in the 13 patients are shown in Table 1. Interestingly, plasma viruses from the eleven responder patients harboured three to five TAMs, whereas the two non-responders had no TAMs (patient L) or only one TAM (T215F, patient M). Plasma viruses from patient L harboured two resistance mutations in the RT gene, including K65R and the MDR mutation

Fig. 1.HIV plasma viral load during foscarnet therapy. Changes in HIV viral load from baseline are shown in the 13 patients receiving foscarnet therapy. A decline of 1 log10copies/ml is symbolized by a dashed line.

Q151M. Patient M’s plasma viruses bore the 69 insertion complex in addition to the T215F TAM.

Despite coadministration of tenofovir to four patients, renal tol- erance was good. Median creatinine clearance values at baseline and on day 14 were 95 ml/min (range: 67–125) and 88 ml/min (range: 68–111), respectively. One patient (A) developed renal tubu- lopathy at the end of foscarnet therapy but rapidly improved after drug discontinuation. Patient F was switched to stavudine (d4T) because of AZT intolerance.

Following foscarnet therapy, all the patients received optimized treatment including at least one new, active ARV drug (median 1, range: 0–3) (Table 1). Raltegravir and darunavir were prescribed to seven patients, etravirine to six patients, and enfuvirtide to four patients. Darunavir was considered as a new drug in patients B and M, even though genotyping of their plasma virus protease genes revealed mutations theoretically conferring resistance to this drug.

Raltegravir and etravirine were both given to four patients. Maravi- roc was not available at time of the study.

At week 12 of optimized treatment, the median decrease in plasma HIV RNA was 3.04 log

copies/ml (range: − 4.14 to +0.33) (Fig. 2) and the proportions of patients who had values below 400 copies/ml and 40 copies/ml were 69% (n = 9) and 38% (n = 5), respectively.

At week 24 the median decrease in plasma HIV RNA was 3.03 log

copies/ml (range: − 4.36 to +0.01) and the proportions of patients with values below 400 copies/ml and 40 copies/ml were both 61% (n = 8).

The median increases in the CD4 cell count at weeks 12 and 24 were 89 cells/mm

(range: 6–129) and 59 cells/mm

(range:

5–181), respectively.

Fig. 2.Virological and immunological changes during and after foscarnet therapy.

The median HIV viral load and CD4 cell count are shown at baseline, at the end of foscarnet therapy, and at weeks 12 and 24 after foscarnet therapy. The HIV RNA detection limit is symbolized by a dashed line.

214C.Charpentieretal./JournalofClinicalVirology43(2008)212–215

Table 1

Characteristics of the patients at baseline and on day 14 (D14) of foscarnet therapy

Baseline D14

Patients ARV treatment at baseline RT mutations^a Major protease mutations^b

CD4 (cells/mm³)

HIV RNA (log10copies/ml)

HIV RNA (log10copies/ml)

viral load log10from baseline

Optimized treatment following foscarnet therapy

Number of active ARV drugs A AZT M41L, L74V, L100I, K103N, M184V,

L210W,T215Y

M46I, I47V, I84V, L90M

0 5.51 3.21 −2.30 TDF + 3TC + DRV/r +ENF 1

B AZT + ABC + TDF + fAPV/r M41L,D67N, L74V, Y181C, M184V,T215F V32I, L33F, M46I, I47V, I54M, L90M

36 4.52 2.17 −2.35 AZT + 3TC + TDF + DRV/r 0

C AZT M41L, E44D,D67N, T69D, K101E, Y181C, M184V, G190A,L210W,T215Y

M46L, I54M, I84V 8 5.84 3.77 −2.07 AZT + 3TC + TDF + DRV/r +ENF 1

D AZT + 3TC + TDF + DRV/r M41L,D67N,L210W,T215Y,K219E D30N, M46I, V82T, N88D, L90M

32 5.50 2.68 −2.82 3TC + TDF + DRV/r +ENF 1

E AZT + 3TC + ABC + DRV/r D67N, T69N,K70R, L100I, K103N, M184V, T215F,K219E

V32I, L33F, M46I, I54L, V82T, I84V, L90M

58 4.63 2.18 −2.45 AZT + 3TC +ABC+ETV+RAL 2

F d4T + TDF + FTC M41L, E44D,D67N, T69D, V75M, M184V, G190A,L210W,T215Y

V32I, M46I, I47V, I50V, I84V, L90M

3 4.58 3.58 −1.00 AZT + TDF + FTC +ETV+RAL 2

G AZT + TDF + FTC + TPV/r M41L, L74V, Y181C, M184V,L210W, T215Y

L33F, M46L, V82F 109 5.95 3.15 −2.80 AZT + 3TC +DRV/r+RAL+ENF 3

H AZT + 3TC + ddI + DRV/r D67N, T69N,K70R, L74I, A98G, M184V, T215F,K219E

V32I, L33F, M46I, I47V, I54L, I84V

2 4.62 3.02 −1.60 AZT + 3TC +ddI+ ABC +ETV+RAL 3

I AZT + ddI + TPV/r M41L,D67N, Y181C,L210W,T215Y, Q219E

M46I, I47V, I84V, L90M

28 4.79 2.92 −1.86 AZT +3TC+ ABC +ETV+ LPV/r +RAL 2

J AZT + 3TC + TPV/r M41L, L74V, V90I, A98G, M184V,L210W, T215Y

V32I, L33F, M46I, I47V, L90M

63 5.03 3.67 −1.36 AZT + 3TC +ETV+RAL 2

K AZT + 3TC + ABC + DRV/r M41L, INS69^c, A98G, V106I, Y181C, G190S,L210W,T215Y

M46I, I47V, I54M, I84V, L90M

15 5.68 3.86 −1.82 AZT + 3TC + ETV + LPV/r + ATZ/r +RAL 1

L AZT + 3TC + LPV/r K65R, Y115F, Q151M, Y188L L33F, M46L, V82A,

I84V,L90M

9 5.51 5.22 −0.29 AZT + 3TC + ABC + TDF +DRV/r 1

M AZT + 3TC + SQV/r + fAPV/r T69S, INS69^c, M184V,T215F V32I, L33F, M46I, I54L, I84V, L90M

23 5.09 4.60 −0.49 AZT + 3TC + EFV + DRV/r 0

Thymidine-associated mutations (TAMs) are underlined and in boldface type. Antiretroviral drugs active according to the baseline genotype are in italicface type. 3TC, lamivudine; ABC, abacavir; ATZ, atazanavir; AZT, zidovudine; d4T, stavudine; ddI, didanosine; DRV, darunavir; ENF, enfuvirtide, ETV, etravirine; fAPV, fosamprenavir; FTC, emtricitabine; LPV, lopinavir; RAL, raltegravir; SQV, saquinavir; TDF, tenofovir; TPV, tipranavir ;X/r, boosted with low-dose ritonavir.

aResistance mutations were reported as listed by ANRS 2007 algorithm, version 16.

b According to the IAS-USA list.

c The 69 insertion complex (insertion of two amino-acids between codons 69 and 70).

C. Charpentier et al. / Journal of Clinical Virology 43 (2008) 212–215 215

4. Discussion

We tested foscarnet, in combination with AZT, as salvage ther- apy prior to an optimized ARV regimen comprising only one fully potent drug from a novel class (enfuvirtide or raltegravir). Thirteen patients with advanced HIV disease were enrolled. Foscarnet ther- apy was well tolerated; in particular, no significant change in renal function was noted.

Eleven of the 13 patients had a virological response to the 14-day foscarnet course, with a median viral load reduction of 1.84 log

copies/ml. The likelihood of a virological response was dependent on the baseline number of TAMs in plasma viruses: the eleven responders had at least three TAMs, whereas the two non- responders (patients L and M) had one and zero TAMs. Patient L had a viral load rebound during foscarnet therapy, possibly owing to the K65R mutation, which is known to impact foscarnet activity.

Patient M had the 69 insertion complex, which does not affect foscarnet susceptibility in vitro.

It is noteworthy that patient K, who had the same insertion complex, had a significant response to foscarnet therapy ( − 1.82 log

copies/ml from baseline).

Ten of the responder patients had a substantial fall in viral load, while the eleventh responder patient (patient F) had a decline of only 1 log

copies/ml, despite the presence of four TAMs at baseline. None of the previously described foscarnet resis- tance mutations (K65R, W88G/S, E89K, L92I, S156A, Q161L, H208Y, K219R, L228R)

was detected at baseline by direct sequencing of this patient’s plasma viruses, although this does not rule out the presence of minority variants bearing such mutations. In addi- tion, this patient was switched to d4T during foscarnet combination therapy, because of AZT-related anemia. Little is known of the antiviral activity of foscarnet in combination with d4T, although a mechanism similar to that seen with AZT is likely to occur. Impor- tantly, none of the ARV-based regimens previously received by the patients in this study led to a viral load decline as strong as that obtained with foscarnet. The failing regimen was always maintained during foscarnet therapy, and the antiviral activity of foscarnet was at least twice that of AZT alone in naive patients.

Thus, it is very likely that the observed virological responses were mainly due to foscarnet itself.

Previous studies of foscarnet in this setting have involved selected patients with at least three TAMs, and the authors could not therefore examine the impact of the number of TAMs on foscar- net susceptibility.

The relationship observed here between the virological response to foscarnet in vivo and the number of TAMs is compatible with a recent study showing that foscarnet interferes with phosphorolysis reactions in vitro.

By acting on RT-catalysed nucleotide excision, foscarnet would enhance the effect of AZT, par- ticularly on AZT-resistant viruses harbouring TAMs. Thus, TAMs can be beneficial in two ways: (i) by inducing hypersusceptibility to fos- carnet, and (ii) by preventing the selection of foscarnet resistance mutations.

The presence of TAMs in plasma HIV is probably not the only predictor of the antiviral activity of foscarnet, as the virological responses observed in our study did not correlate strictly with the number of TAMs. Indeed, patient F, who had four TAMs at baseline, had only 1 log

viral load decline. In the study by Canestri et al.,

one patient, despite having three TAMs, had no significant response to foscarnet (viral load change of 0.11 log

copies/ml).

In our study, at W24 of optimized ARV treatment after foscarnet therapy, 61% of patients had viral loads below the detection limit of 40 copies/ml.

Our results thus confirm the value of salvage therapy based on foscarnet plus AZT,

in patients selected for their overall resistance profile and not for the number of TAMs, and also corroborate in vivo the impact of the number of TAMs on the virological response to foscarnet. We are currently examining whether foscarnet “induc- tion” therapy before an optimized antiretroviral regimen has any long-term benefits.

References

1. Kaiser L, Perrin L, Hirschel B, Furrer H, Von Overbeck J, Olmari M, et al. Foscarnet decreases human immunodeficiency virus RNA.J Infect Dis1995;172:225–7.

2. Devianne-Garrigue I, Pellegrin I, Denisi R, Dupon M, Ragnaud JM, Barbeau P, et al. Foscarnet decreases HIV-1 plasma load.J Acquir Immune Defic Syndr 1998;18:46–50.

3. Bergdahl S, Jacobsson B, Moberg L, Sonnerborg A. Pronounced anti-HIV-1 activ- ity of foscarnet in patients without cytomegalovirus infection.J Acquir Immune Defic Syndr1998;18:51–3.

4. Mathiesen S, Roge BT, Weis N, Lundgren JD, Obel N, Gerstoft J. Foscarnet used in salvage therapy of HIV-1 patients harbouring multiple nucleotide excision mutations.AIDS2004;18:1076–8.

5. Canestri A, Ghosn J, Wirden M, Marguet F, Ktorza N, Boubezari I, et al. Foscar- net salvage therapy for patients with late-stage HIV disease and multiple drug resistance.Antivir Ther2006;11:561–6.

6. Cruchaga C, Anso E, Rouzaut A, Martinez-Irujo JJ. Selective excision of chain-terminating nucleotides by HIV-1 reverse transcriptase with phospho- noformate as substrate.J Biol Chem2006;281:27744–52.

7. Tachedjian G, Mellors J, Bazmi H, Birch C, Mills J. Zidovudine resistance is sup- pressed by mutations conferring resistance of human immunodeficiency virus type 1 to foscarnet.J Virol1996;70:7171–81.

8. Meyer PR, Matsuura SE, Zonarich D, Chopra RR, Pendarvis E, Bazmi HZ, et al. Relationship between 3-azido-3-deoxythymidine resistance and primer unblocking activity in foscarnet-resistant mutants of human immunodefi- ciency virus type 1 reverse transcriptase.J Virol2003;77:6127–37.

9. Hammond JL, Parikh UM, Koontz DL, Schlueter-Wirtz S, Chu CK, Bazmi HZ, et al. Alkylglycerol prodrugs of phosphonoformate are potent in vitro inhibitors of nucleoside-resistant human immunodeficiency virus type 1 and select for resistance mutations that suppress zidovudine resistance.Antimicrob Agents Chemother2001;45:1621–8.

10. Mellors JW, Bazmi HZ, Schinazi RF, Roy BM, Hsiou Y, Arnold E, et al. Novel muta- tions in reverse transcriptase of human immunodeficiency virus type 1 reduce susceptibility to foscarnet in laboratory and clinical isolates.Antimicrob Agents Chemother1995;39:1087–92.

11. Tachedjian G, Hooker DJ, Gurusinghe AD, Bazmi H, Deacon NJ, Mellors J, et al. Characterisation of foscarnet-resistant strains of human immunodeficiency virus type 1.Virology1995;212:58–68.

12. Eron JJ, Benoit SL, Jemsek J, MacArthur RD, Santana J, Quinn JB, et al. Treatment with lamivudine, zidovudine, or both in HIV-positive patients with 200 to 500 CD4+ cells per cubic millimetre. North American HIV Working Party.N Engl J Med1995;333:1662–9.