Evolution of HIV resistance during treatment interruption in experienced patients and after restarting a new therapy

Journal of Clinical Virology 34 (2005) 277–287

Evolution of HIV resistance during treatment interruption in experienced patients and after restarting a new therapy Melanie Balduin a , Saleta Sierra a , Martin P D¨aumer a , J¨urgen K Rockstroh b , Mark Oette c , Gerd F¨atkenheuer d , Bernd Kupfer e , Niko Beerenwinkel f , Daniel Hoffmann g , Joachim Selbig h , Herbert J Pfister a , Rolf Kaiser a,∗ a

c

Institute of Virology, University of Cologne, Fuerst-Pueckler Str. 56, D-50935 Cologne, Germany b Departement of Internal Medicine I, University of Bonn, Bonn, Germany Clinic for Gastroenterology, Hepatology and infectious Diseases, University of Duesseldorf, Duesseldorf, Germany d Department of Internal Medicine I, University of Cologne; Cologne, Germany e Institute of Medical Microbiology and Immunology, University of Bonn, Bonn, Germany f MPI for Informatics, Saarbruecken, Germany g Center of Advanced European Studies and Research (caesar), Bonn, Germany h MPI of Molecular Plant Physiology, Golm, Germany Received 4 July 2005; accepted 31 August 2005

Abstract Background: To analyse the evolution of resistance patterns in patients undergoing treatment interruption (TI) and reinitiating highly active anti-retroviral therapy (HAART). Methods: HIV-RT and -PR gene-sequences were analysed in 14 patients (>5 failing prior drugs) before and during TI and under a new HAART. Genotypes were interpreted using two bioinformatics systems. Additionally, virus load (VL) and CD4+ -T-cell counts were measured. Results: Six patients (42%) achieved sustained undetectable VL up to one year after TI (responders), while 8 (57%) maintained VL of more than 2000 copies/mL (non-responders). Different patterns of resistance-mutations evolution were detected. During TI loss of all mutations was observed in three patients, a reduction of mutations was detected in seven patients, and no alteration was seen in four patients. In the responders, 87.5% of protease inhibitor (PI)-resistance mutations waned during TI and remained undetectable under the new treatment. In contrast, in the non-responder group most PI-resistance mutations continued noticeable under the new therapy. Loss of primary PI-resistance mutations and the presence of one fully active PI in the new regimen significantly correlated with success of subsequent treatment (p = 0.028). In two patients new reverse transcriptase associated mutations were detected during TI, G190A (NNRTI mutation) and K70R (NRTI mutation). Appearance of K70R could be explained by a reverse direction of a previously described pathway of thymidin analogues mutation resistance development, while G190A could be due to prolonged subinhibitory drug levels after cessation of NNRTIs. Conclusion: In the evolution of HAART-resistance, different patterns were observed in responders and non-responders during but not before TI. Absence of PI-resistance associated mutations during and after TI and administration of a predicted fully active PI for the new therapy correlated with success. Newly detected mutations during TI may indicate reversibility of previously described mutational pathways. © 2005 Elsevier B.V. All rights reserved. Keywords: HIV; Therapy interruption; Resistance; Evolution

1. Introduction

∗

Corresponding author. Tel.: +49 2214787741; fax: +49 2214783904. E-mail address: [email protected] (R. Kaiser).

1386-6532/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2005.08.007

The introduction of highly active anti-retroviral therapy (HAART) has achieved in most HIV-infected subjects increases in CD4+ -T-cell counts and reductions in plasma

278

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

viraemia to undetectable levels, which can be maintained for years. As a consequence, a dramatic reversal in disease progression and the incidence of AIDS-associated complications has been observed within the last years (Hogg et al., 1997; Palella et al., 1998). However, long-term virological success of HAART is limited by resistance development, severe toxicity, pharmacokinetic interactions or patients’ personal choice (d’Arminio-Monforte et al., 2000; Duran et al., 2001; 2Mocroft et al., 2001; Tuldr´a et al., 2001; Carrieri et al., 2003). In highly experienced HIV-infected patients under HAART the virus accumulates numerous resistanceassociated mutations. Indeed, almost 50% of the HIVinfected patients receiving HAART have now already developed at least one mutation against one compound of their present HAART (Bartlett et al., 2001; Richman et al., 2001, 2004). A treatment interruption (TI) frequently causes a shift in the circulating virus to a more sensitive strain (Devereux et al., 1999; Verhofstede et al., 1999; Izopet et al., 2000; Miller et al., 2000; Deeks et al., 2001; Delaugerre et al., 2001; Hance et al., 2001; Falkensammer et al., 2002; Kijak et al., 2002; Halfon et al., 2003). In many cases, a virus identical to plasma virus before TI or to proviruses detected in latent reservoirs rebounds when a therapy is reinitiated (Chun et al., 2000; Delaugerre et al., 2001; Imamichi et al., 2001; Kijak et al., 2002; Deeks et al., 2003). So far, detailed models of resistance development and virus evolution under changing therapy conditions have been missing. The usefulness of a TI for the success of a subsequent therapy is currently a matter of controversial debate. Some studies observed better virological and immunological responses to the reinitiated therapy after a TI (Deeks et al., 2001; Izopet et al., 2002; Katlama et al., 2003; Benson et al., 2004; Gianotti et al., 2004), while others detected no significant improvement in the response and even development of clinical complications (Chun et al., 1999; Bonhoeffer et al., 2000; Dorman et al., 2000; Ru´ız et al., 2000, 2003; Benson, 2001; Delaugerre et al., 2001; Girard et al., 2001; Ort´ız et al., 2001; Oxenius et al., 2002; Yozviak et al., 2002; Metzner et al., 2002; Deeks et al., 2003; Lawrence et al., 2003; Dieterich, 2003). Data about the evolution of viral strains during TI are rare but valuable because they could enable a rational choice of a successful subsequent therapy. In this study, we have analysed the evolution of resistance patterns during and after TI and the consequences for therapy success in fourteen patients with multiple previous therapy failures.

2. Materials and methods 2.1. Viral samples Plasma samples derived from fourteen patients from the University Hospitals of Cologne, Bonn and Duesseldorf (Germany) were analysed in this study. All patients

were included in the “Arevir project” (www.genafor.org). Within the scope of this project authorization statements from patients to use their medical data and samples for investigation were signed. HIV-pol region was sequenced at least three time-points: G0 : genotype at initiation of the TI (median of 2.5 days in TI, range 54 before TI to 23 in TI), G1 : genotype within the TI (median of 67 days after starting TI, range 21–159), G2 : genotype after the TI and under a median of 39 days under the new treatment (range 26–188). Viral load and CD4+ -T-cell counts at these time points and during the follow up under the new treatment (median 433 days, range 376–582) were determined. Samples were collected between March 2001 and November 2002. 2.2. Genotypic resistance testing Viral RNA was isolated from patients plasma using QIAamp viral RNA mini-kit (Qiagen, Hilden, Germany). Reverse transcription and polymerase chain reaction were performed using OneStep RT-PCR kit (Qiagen, Hilden, Germany) and second round PCR was carried out with HotStarTaq kit (Qiagen, Hilden, Germany). PCR products were purified using the QIAquick spin PCR purification kit (Qiagen, Hilden, Germany) and extended using the ViroSeq HIV-1 Genotyping System sequencing module (Applied Biosystems, Foster City, CA, USA). Extension products were purified using MultiScreen purification plates (Millipore, Bedford, MA, USA) and Sephadex G-50 superfine (Amersham Biosciences, Uppsala, Sweden) and were run on an ABI Prism 310 capillary sequencer. Sequence data were generated by using Sequencing Analysis v3.4 (Applied Biosystems, Foster City, CA, USA). The sequences were edited using the ViroSeq HIV-1 Genotyping software v2.5 (ABBOTT, Chicago, USA). Mutations were classified according to the International AIDS Society USA (IASUSA). 2.3. Genotypic resistance data interpretation The geno2pheno system (g2p) and the Stanford HIV drug resistance database prediction tool (Stanford) were used to predict the susceptibility to 17 of the currently approved drugs (NRTIs, NNRTIs and PIs) from viral genotypes. The g2p (http://www.genafor.org) provides different interpretation tools based on machine learning techniques (Beerenwinkel et al., 2003). The Probability Score was used in this study. Probablility scores <0.05 were considered as “fully susceptible virus” or “fully active drug”. Values between 0.05 and 1.0 were considered as “not-fully susceptible virus” or “not-fully active drug”. The Stanford HIV drug resistance database (http://hivdb. stanford.edu/) classifies each drug into one out of five categories. To avoid underestimation of drug resistance, only drugs evaluated as “susceptible” were considered as “fully susceptible virus” or “fully active drug” in our study. “potential low level resistant”, “low level resistant”, “intermediate

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

279

resistant” and “high level resistant” drugs were categorised as “not-fully susceptible virus” or “not-fully active drug”. 2.4. Statistics p-Values were calculated by Fisher’s exact test (http:// faculty.vassar.edu/lowry/tab2x2.html).

3. Results Fourteen chronically HIV-infected patients with multiple therapy failures undergoing a therapy interruption and restarting a new regimen were included in this retrospective analysis. Evolution of patients’ viral resistance profiles before, during and after TI was investigated. Other parameters related to resistance profile development and their correlation with response of subsequent therapy were also studied. Genotype (G), therapy (T), Viral load (VL) and CD4+ -Tcell counts (CD4) data collected before TI are marked with subscript 0, data obtained during TI are marked with subscript 1, and data from samples after TI are marked with subscript 2. Follow up of VL and CD4 data after TI are chronologically numbered with subindex 3, 4, 5 and 6. 3.1. Study design and patients’ characteristics Thirteen of the patients were male and patient #1 was female. The viruses from all patients were classified as HIV1 subtype B, with the exception of patient #3 who harboured a virus with subtype C. With the exception of patient #3, who was NNRTI-naive, all patients were experienced with all three major drug classes; median 4 PIs (range 1–7), median 6 NRTIs (range 4–7), and median 1 NNRTI (range 1–2). None of them had received fusion inhibitors during time of observation. The patients were at least 4 years under HAART and had taken a minimum of five different drugs from at least two drug classes. All patients had a minimum of two recorded failing therapies, including at least one failing PI-containing regimen, before they started the TI. At baseline VL were higher than 2000 copies/mL. The reasons to stop the failing therapy were viral resistance, severe toxicity, opportunistic infections and patient’s personal choice. Detailed patients’ characteristics are shown in Table 1. 3.2. Virologic and clinical parameters The focus of our study was on the evolution of HIV resistance mutations during and after TI although for a complete analysis we also evaluated clinical and virological data. The median duration of TI was 80.5 days (range 23–175). VL and CD4 were measured throughout the study (Fig. 1). Patients were classified as “non-responders” or “responders” according to their viral load values during the first year after re-implementation of a treatment. Non-responders where those patients with sustained detectable viraemia in at least

Fig. 1. (A) Patients’ viral load values (VL) before, during and after therapy interruption (TI). VL0 (bars with pattern): median of 2.5 days in TI (range 54 before TI - 23 in TI); VL1 (bars in black): median of 65 days in TI (range 21–159); VL2 (bars in white): median of 39 days under the new treatment (range 26-188); VL3 (bars in white): median of 17 weeks under the new treatment (range 12–34); VL4 (bars in white): median of 36 weeks under the new treatment (range 23–45); VL5 (bars in white): median of 50 weeks under the new treatment (range 34–70); VL6 (bars in white): median of 66 weeks under the new treatment (range 54–109). The values under the limit of detection (VL<50 copies/mL) have been represented as 49 copies/mL. The responder patients are underlined. (B) CD4+ -T-cell-counts before, within, and after TI. The figure shows the CD4+ -cell-counts (cells/␮L) along the study. CD40 (bars with pattern): median of 2.5 days in TI (range 54 before TI-23 in TI); CD41 (bars in black): median of 65 days in TI (range 21–159); CD42 (bars in white): median of 39 days under the new treatment (range 26–188); CD43 (bars in white): median of 17 weeks under the new treatment (range 12–34); CD44 (bars in white): median of 36 weeks under the new treatment (range 23–45); CD45 (bars in white): median of 50 weeks under the new treatment (range 34–70); CD46 (bars in white): median of 66 weeks under the new treatment (range 54–109).

two consecutive time-points (patients #1, #3, #4, #6, #7, #8, #9, and #12). Responders (patients #2, #5, #10, #11, #13 and #14) achieved VL3 to VL6 ≤50 (although two patients had an temporary increase in viral load higher than level of detection so called blips). Patients started TI with a median VL0 of 30820 copies/mL and CD40 of 212 cells/␮L. During TI, the viral load in the responder-group increased a median of 0.81 log10 and in the non-responder-group a median of 0.11 log10 . CD4+ -T-cell-counts decreased also more noticebly in the responders (34 cells/␮L) than in the non-responders (11 cells/␮L). After a median of 16 weeks under the new therapy, the VL3 decreased to a median of 1143 copies/mL and CD43

280

Table 1 Patients’ characteristics # 1

3 4 5 6 7 8 9

T2

ABC, 3TC, NVP, LPV/r ddI, 3TC, ABC, NVP ABC, AZT, 3TC, LPV/r ddI, d4T, ABC, LPV/r ddI, ABC, NVP

ABC, 3TC, NVP, LPV/r 3TC, TDF, APV, RTV, LPV/r TDF,3TC, IDV, LPV/r AZT, 3TC, ABC, TDF, LPV/r d4T, ABC, LPV/r, TDF NVP, SQV, LPV/r

d4T, 3TC, NVP, APV, TDF EFV, ddI, LPV/r, APV 3TC, APV, LPV/r

12

d4T, RTV, IDV, SQV AZT, ABC, EFV AZT, 3TC, ABC, TDF ABC, 3TC, EFV

13

3TC, d4T, EFV

14

d4T, ABC, APV

10 11

Median Minimum Maximum

3TC, TDF, IDV, LPV/r, APV 3TC, TDF, IDV, LPV/r 3TC, TDF, APV, RTV TDF, 3TC, LPV/r 3TC, LPV/r TDF, 3TC, ABC, LPV/r TDF, 3TC, ABC, LPV/r 3TC, ddI, TDF, EFV, LPV/r

Years since 1st , HIV diagnosis

Years on therapy before TI

Age (years)

No. of PIs before TI

No. of NRTIs before TI

No. of NNRTIs before TI

Duration of TI (days)

Reason for TI

12

5

40

5

4

1

175

Tuberculosis

14

14

60

7

6

1

66

Resistance

6

5

40

3

4

0

79

Resistance

7

5

43

5

6

1

42

Toxicity

12

12

61

1

6

1

159

Resistance

10

10

41

4

6

2

112

Resistance

10

6

48

6

6

2

70

Patient’s choice

5

4

44

4

5

1

99

Unknown

11

6

38

6

5

2

74

Leucoencephalopathy

>12 8

12 5

44 39

2 3

6 6

1 1

23 95

Toxicity Toxicity

17

11

41

2

6

1

76

Toxicity

15

11

55

1

4

1

91

Resistance

14

10

61

4

6

1

82

Toxicity

12 5 17

8 4 14

43.5 38 61

4 1 7

6 4 6

1 0 2

80.5 23 175

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

2

To

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

Fig. 2. Viral susceptibility to antiretroviral drugs (PI: protease inhibitors, NRTI: nucleoside reverse transcriptase inhibitors, NNRTI: non-nucleoside reverse transcriptase inhibitors) before, during and after therapy interruption. The class and median number of predicted fully susceptible drugs in the eight non-responder and the six responder patients are represented in bars. The estimations were made with the bioinformatic tools g2p and Stanford according to the genotype before interruption (G0 ), genotype during interruption (G1 ) and the genotype under the new treatment (G2 ).

increased to a median of 230 cells/␮L. After a median of 36 weeks under the new therapy the VL4 was in median 3369 copies/mL and all responders reached and maintained CD4+ -T-cell counts around 300–400 cells/␮L. In the nonresponders, post-TI CD4+ -T-cell counts largely differed from one patient to another. No statistically significant differences between responders and non-responders concerning duration of TI, VL and CD4 levels and modifications prior, during or after TI were detected. 3.3. Resistance profile, drug susceptibility and efficacy of the regimens as estimated from g2p and Stanford Genotyping of the viral pol gene before, during and after TI was performed. The resistance-associated mutations were analysed and the susceptibility was estimated using the interpretation systems g2p and Stanford (Fig. 2). In our analysis, drugs were considered either fully active or not-fully active. At baseline, 12 patients carried a virus with at least one primary and one secondary PI-resistance-associated mutation (median of all patients = 7.5 PI-mutations). In all patients the virus had at least 3 NRTI-resistance mutations (median of all patients = 5.5 NRTI-mutations), and in 10 patients NNRTIresistance mutations were present (median of all patients = 2 NNRTI-mutations) (Table 2). No correlation between the initial number of NRTI- and/or NNRTI-resistance mutations and the response to T2 was detected, but a significant correlation between a number of ≤8 PI-resistant mutations in G0 and response to the new therapy was detected (p = 0.028). However, three patients with ≤8 PI-resistance mutations at baseline did not respond to T2 (pats. #1, #6 and #12). At baseline, all eight non-responders and four responders had been not-fully susceptible to the PI included in their future

281

T2 regimen, while two of the responders (pats. #10 and #11) had been fully susceptible to LPV/r which was included in their T2 after TI. During TI, G1 was made after a median of 67 days in the absence of drugs (range 34–159). Loss of all drug-resistanceassociated mutations was observed in three patients, a reduction in the number of resistance mutations was observed in additional seven patients, and no modification was seen in four patients. Classified in responders and non-responders, the loss of primary and secondary PI-mutations was 87.5% versus 36.7% (overall 49.2%). For NRTIs 52.6% mutations were lost in responders and 29.7% in non-responders (overall 41.%). NNRTI mutations were lost in responders with a frequency of 42.9% versus 40% in non-responders (overall 42.9%). After the cessation of therapy, the mutations RT-K70R (pat. #12) and RT-G190A (pat. #1) were newly detected. The number of PI-resistance mutations in G1 was a predictive parameter for the success of the subsequent therapy: a significant correlation between a number of ≤1 secondary PI-resistance mutations (p = 0.0093) or absence of primary mutations during TI (p = 0.028) and success of T2 was detected. There was also a statistically significant correlation between absence of primary PI-resistance mutations with a sampling date of at least 48 days in the absence of drugs (p = 0.045). Both interpretation systems predicted an increase of the number of fully active drugs during the TI. There was a statistically significant correlation between the increase number of ≥1 active PI and response to T2 (p = 0.027). No correlation between the number of NRTI- and/or NNRTIresistance mutations or susceptibility during TI and response to T2 was detected. After a median of 39 days under the new treatment, genotyping of the circulating virus was made. In nine patients less resistance mutations compared to baseline were detected, in four patients the baseline strain was found and the virus from patient #1 even accumulated resistance mutations. In our cohort, PI resistance mutations were reduced in 29%, NRTI mutations in 36% and NNRTI mutations in 41% compared to the number of baseline mutations. In the responders, 88% fewer PI mutations, 53% fewer NRTI mutations and 86% fewer NNRTI mutations were detected. On the other hand, in the non-responders the number of PI mutations was only reduced by 10%, for NRTI by 19% and for NNRTI by 27%. The mutations RT-K219Q and RT-D67G (pat. #1), RT-V108I (pat. #6), and PR-K20R and RT-K219Q (pat. #12) were newly detected under the new treatment T2 . Under the new treatment, the susceptibility profiles remained similar to those during the TI. In order to find a correlation between drug-susceptibility in G2 and the success of the therapies, the number of working drugs was calculated. We defined working drugs as the fully active substances included in the concomitant regimen (Table 3). In the baseline treatments, (all of them failing regimens) the median number of working drugs was 0 [accordingly to g2p, (range 0–1) and Stanford, (range 0–1)]. In the new therapy (T2 ) of the non-responders, the median number of working drugs was

282

Table 2 Resistance-associated mutations found in G0 , G1 and G2 in the plasma samples of the patients PI K 20

1

G0 G1 G2

L 10 I I I

L 24 I I I

NRTI

3

G0 G1 G2

F – F

R – R

I – I

4

G0 G1 G2

I I I

6

G0 G1 G2

F F –

7

G0 G1 G2

I I I

8

G0 G1 G2

I – I

9

G0 G1 G2

I I I

12

G0 G1 G2

I – I

2

G0 G1 G2

F – I

5

G0 G1 G2

I – –

10

G0 G1 G2

– – I

11

G0 G1 G2

13

G0 G1 G2

I I –

14

G0 G1 G2

I – –

Sum all

G0 G1 G2

12 6 9

V 32

I I I

L 33

M 36

M 46 L L L

I I I I I I

I 54

L 63 P P P

A 71

I – I

V V V

P P P

V V V

I I I

V V V

P P P

V V V

L L L

V V V

P P P

V V V

P P P

V V V

I I I

V – V

P P P

V – V

I I I

V V V

P P P

V V V

V – –

P P P

F F F

I I I

F – F

I – I I I I

– – R

I 47

G 48

I 50

F 53

V V V V – V

L – L

I I I

I – – I – –

G 73

V 77

V 82

L – –

P P P

L 90

A – A A A A

I 93

A A – V V V A – A

M M M M M M

L – –

M M M

L L L

M – M

2 1 2

2 1 2

1 1 1

2 1 2

5 3 3

V V V S – – V – –

V – –

M M M

L L L

M – –

L L L

M – – M – –

1 1 1

1 0 0

1 0 1

2 0 1

L L –

D 67

T 69

K 70 R R R

N N N

N N N

R R R

D D –

N N –

N – N

V – –

S – –

I – –

A – –

7 4 5

13 12 11

8 5 6

3 1 1

4 3 3

4 1 2

5 4 4

9 4 5

V 118

M 184 V – –

L L L D D D

V – –

6 5 5

7 3 4

V V V

3 3 2

0 0 0

I I I

D – D

I – –

R R R

8 4 4

7 5 6

7 7 7

W – –

V – –

V – –

Y – –

– – Q

F – –

E – –

W W W

Y Y Y

W – –

Y – –

I I I 1 1 1

5 3 3

10 4 4

F 116

Q 151

A 98

L 100

K 101

I – –

K 103

V 106

V 108

A A –

– – I

V 179

Y 181

Y 188

G 190 – A A

N – –

E – E I I I

L L L

Y Y Y

5 3 2

Q Q Q

F – –

Q – –

10 5 6

8 5 7

I I I C – C

M M M

D D D N – –

A – A

C C C C – – C – –

N N – G G G

F F F

A A – D D D

Q Q –

V V –

4 1 2

F 77

I I I

E E E

V – –

D D D

D N N

Y Y Y

Q – Q

R R R

N – –

W W –

F – F

I – –

NNRTI V 75

SSA SSA SSA

V – V

N – –

R R R

Y Y Y

V – V

V – –

N N N

W W W

E E E

V – –

T 69

Q Q Q

Y Y Y

R – –

N N N

V V –

K 219 – – Q

– – V

N – –

N N N

T 215

V V V

– R R

L – –

L 210

V V V

I I I

N – N

L – –

M – –

P – –

V 115

F F F

L – L

T – –

L 74

D D – R R R

L L L

P P –

8 5 7

MDR K 65

L – L

I I I

V – –

E 44 D D D

L L L

P P P

I – –

M 41

L L L V V V

P P P

M M –

I 84 V V V

A – –

C – –

N N N

L L –

L L –

1 1 1

1 1 1

1 1 1

1 1 1

1 1 1

1 1 1

2 1 1

1 0 1

4 2 1

1 1 0

0 0 1

2 2 2

6 2 3

2 2 0

3 2 2

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

Pat.

5 4 4 6 4 5 4 4 4 3 1 2 2 2 2 2 1 1 6 5 6 8 8 8 6 4 5 1 0 1 1 0 1 0 0 0 1 1 1 7 5 7 4 3 3 2 1 2 1 1 1 2 1 2 1 0 2 8 5 7 G0 G1 G2 Sum non-resp

PI: protease inhibitor associated mutations; NRTI: nucleoside reverse transcriptase inhibitor associated mutations; MDR: multi NRTI resistance associated mutations; NNRTI: non-nucleoside reverse transcriptase inhibitor associated mutations; G0 : genotype before therapy interruption; G1 : genotype within therapy interruption; G2 : genotype under a new therapy; responders: patients with VL under level of detection under the new treatment (patient number underlined); non-responders: patients with VL >2000 under the new treatment; The underlined amino acid positions indicate primary mutations for PI resistance. The PR-amino acid positions 63, 71, 77, 93 are changes which per se slightly affect drug susceptibility but reinforce the effect of other mutations or increase the viral fitness of resistant strains (Oette et al., 2003).

2 2 2 0 0 0 4 2 3 2 2 2 0 0 1 1 1 0 2 0 0 1 0 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 3 6 5 3 4 6 3 4 2 2 1

0 0 0

3 2 2

3 2 2

3 4 4

2 1 2

1 1 1

1 1 1

3 2 1

0 0 0 5 2 2 4 1 0 1 1 0 G0 G1 G2 Sum resp

4 1 2

0 0 0

0 0 0

0 0 0

1 0 0

1 0 0

0 0 0

1 0 0

0 0 0

1 0 0

1 0 0

5 4 3

2 0 0

1 0 0

2 1 1

1 0 0

1 0 0

3 0 0

1 1 1

3 1 2

1 1 1

0 0 0

5 2 2

4 3 4

4 3 3

2 0 0

0 0 0

4 2 2

2 1 1

4 2 1

0 0 0

0 0 0

0 0 0

0 0 0

0 0 0

0 0 0

0 0 0

2 2 1

0 0 0

0 0 0

0 0 0

2 0 0

2 2 0

1 0 0

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

283

0.5 [corresponding to g2p, (range 0–2)], and 0 [according to Stanford, (range 0–2)]. In the T2 of the responders, the median number of working drugs was 2 [accordingly to g2p, (range 2–4)] and 3 [as predicted by Stanford, (range 1–4)]. There was a significant correlation between presence of ≥1 working boosted-PI and successful response to T2 (p = 0.0003 accordingly to g2p, and p = 0.0023 accordingly to Stanford). This prediction was accurate for the 6 responding patients but failed for the non-responders, patients #8 and #12, who according to G1 were estimated to include active LPV/r in their T2 . But at failure these two patients displayed a strain with drug resistance mutations not visible in the prior analysis. No correlation between the number of working NRTIs and/or NNRTIs and the success of the regimen was detected.

4. Discussion 4.1. Resistance mutation pattern evolution The evolution of resistance during and after TI followed different patterns in 14 heavily pretreated patients. It has been previously described that in viruses with multiple resistance-associated mutations, a complete stop of therapy often leads to a replacement of the resistant virus by a fully susceptible or a less resistant variant (Devereux et al., 1999; Verhofstede et al., 1999; Miller et al., 2000; Izopet et al., 2000; Deeks et al., 2001; Delaugerre et al., 2001; Hance et al., 2001; Falkensammer et al., 2002; Kijak et al., 2002; Halfon et al., 2003). However, viral rebound of previous resistant variants is generally observed when a therapy is restarted (Chun et al., 2000; Delaugerre et al., 2001; Imamichi et al., 2001; Kijak et al., 2002; Deeks et al., 2003). In our study, cessation of treatment lead to a significant reduction in the number of resistance-associated mutations in ten patients, while in the other four only minor or no changes at all were detected. After implementation of the new therapy, eight patients harboured viruses identical or with more resistanceassociated mutations than baseline, but in six, a susceptible virus different from the resistant baseline remained. The different resistance evolution patterns during and after TI could not be predicted by any of the analysed baseline parameters. The modification of the resistance patterns during and after TI had consequences for the success of the subsequent therapy. In the eight patients where the virus remained identical or similar to the baseline, the following therapy failed and no significant reductions in the VL were observed (nonresponder patients). In contrast, in the six patients who had a susceptible virus during TI and maintained it under the new treatment, sustained undetectable viral loads for longer than one year were achieved (responder patients). One problem of TI is the emergence of resistance viruses (Bonhoeffer et al., 2000; Dorman et al., 2000; Girard et al., 2001; Metzner et al., 2002). This risk is specially high for NNRTIs due to the long half-life of these drugs which leads to

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

284

Table 3 Working drugs in treatment before interruption (T0 ) and under the new treatment (T2 ) #

1 3 4 6 7 8 9 12 2 5 10 11 13 14

g2p

Stanford

Working drugs in T0 as estimated from G0

Working drugs in T2 as estimated from G2

PI

NRTI

NNRTI

PI

NRTI

– – – – – – – – – – – – – –

– – ddI – – – – – – – – – – –

NVP – – – – – – – – – – – – –

– – – – – – – – LPV/r LPV/r LPV/r LPV/r LPV/r LPV/r

3TC – 3TC – – – – 3TC + ABC 3TC + TDF d4T* 3TC 3TC 3TC 3TC + ddI

Working drugs in T0 as estimated from G0

Working drugs in T2 as estimated from G2

NNRTI

PI

NRTI

NNRTI

PI

NRTI

NNRTI

– – – NVP – – – – – – – – – EFV

– – – – – – – – – – – – – –

– – – – – – – – – – – – – –

NVP – – – – – – – – – – – – –

– – – – – – – LPV/r APV+LPV/r LPV/r LPV/r LPV/r LPV/r LPV

3TC – – – – – – 3TC 3TC+TDF d4T* + 3TC + TDF – – 3TC 3TC + TDF + ABC

– – – – – – – – – – – – – EFV

Working drugs were included in the regimen and estimated as fully active from genotype before interruption (G0 ) and genotype under the new treatment (G2 ) by g2p and Stanford. The drugs marked in italics indicate a successful treatment.; *d4T is after 195 days under therapy taken out of the regimen. The patient continued with LPV/r, ABC and TDF, but only LPV/r as fully active substance. APV: amprenavir; LPV/r: boosted lopinavir; ddI: didanosine; d4T: stavudine; 3TC: lamivudine; ABC: abacavir; TDF: tenofovir; NVP: nevirapine.

suboptimal plasmatic concentration for a long time after their withdrawal. In our study, NNRTI-resistance mutations arose during TI (G190A in patient #1) or shortly after resumption of treatment (V108I in patients #2 and #6) in three patients who had included NVP in their T0 . These new mutations led to a reduction in the efficacy of NVP reused in the T2 of patients #1 and #6. Therefore, when a TI is planned, it is advisable to withdraw the NNRTIs before the other drugs. In addition, also PI- and NRTI-resistance associated mutations were newly detected during or after TI. In patient #12, the PIresistance mutation K20R was observed under T2 . Although this mutation could not be detected in G0 and G1 , it had been earlier seen in a genotype determined one year before G0 , but the whole resistance pattern of this older sample was divergent from G2 (patient’s data prior to this study, not shown). The presence of this K20R mutation could therefore be a consequence of reverse mutation or a selection by LPV/r of an old strain followed by a new evolutionary process. One new NRTI-resistance mutation was also detected during the TI at codon 70 and three more after reinitiation of treatment. The appearance of the NRTI-resistance mutations in patients #1 and #12 could be explained by the resistance development pathway under AZT treatment previously described (Beerenwinkel et al., 2004; Boucher et al., 1992; Larder and Kemp, 1989). This pathway starts with the appearance of K70R and may continue in two different branches: development of K219Q and subsequently D67N, or development of M41L and T215Y together with a non-systematic fading of K70R. In patient #12, the mutations M41L and T215Y but not K70R were visible at baseline. During the TI, M41L and T215Y waned and K70R was detectable. It seems that either through reverse mutations or by reemergence of an old strain containing only the K70R, the virus

followed the resistance development pathway in the reverse direction and returned to the origin. Under the new selective pressure of T2, it proceeded along the alternative branch (K219Q, D67N). The presence of the mutations M41L and T215Y but not K70R, and the reappearance of K70R during a TI has also been observed in two other patients not included in this study (Balduin et al., 2004). Other studies (Miller et al., 2000; Deeks et al., 2001; Deeks et al., 2003)] described a massive decrease in CD4+ T-cell counts as another problem of TI. In our analysis, although CD4 counts decreased moderately during TI, postinterruption CD4 counts were in most patients similar to baseline values. 4.2. Predictors of therapy success after TI The length of the TI, baseline VL and/or CD4, or VL and/or CD4 changes during the TI were not significant predictors for the success of the new therapy in our study, in contrast to other authors (Miller et al., 2000; Izopet et al., 2000; Deeks et al., 2001; Ru´ız et al., 2003; Gianotti et al., 2004). In our study, a statistically significant difference between responders and non-responders in the reduction of the number of PI-resistance mutations during TI was found. Moreover, a number of ≤1 secondary PI-resistance mutation or absence of primary mutations during TI statistically correlated with success of T2 . Therefore, G1 was a good predictor of therapy success. Our results, as well as previous studies, suggest that genotyping should not be made too close to TI start, because resistance mutations loss may be underestimated (Ru´ız et al., 2003). A two months period off-therapy seems to be an

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

appropiate timespan for a genotyping with sufficient predictive value. As a consequence of the reduction of PI-resistance mutations during and after TI, a significant increase of the PIsusceptibility in the responders and in two non-responders could be predicted using the g2p and Stanford prediction tools. No phenotypic assays to verify these predictions were available although some studies have demonstrated the validity of different prediction systems usage in therapy optimization (Durant et al., 1999; Baxter et al., 2000; Shafer et al., 2001; Cingolani et al., 2002). After restarting the therapy, mutation- and drug susceptibility-profiles remained similar to those during TI, with the exception of patient #12 who developed new resistance-mutations, and patient #8 where a re-emergence of a resistant virus was observed. In order to find a relation between drug-susceptibility and the success of the corresponding therapies, the number of working drugs was calculated (working PI defined as fully active PI concomitantly included in the regimen). In our study, only regimens which included ≥1 working boosted PI achieved a reduction and maintenance of viral loads ≤ 50 copies/mL. The importance of PIs for the success of a new therapy in heavily pretreated patients has also been previously reported (Deeks et al., 2001; Katlama et al., 2003). In contrast, no correlations between the number of NRTI- and/or NNRTI-resistance mutations, NRTI- and/or NNRTI-susceptibility profile, or number of NRTI- and/or NNRTI-working drugs with success of the new therapy could be detected, respectively. Some authors have claimed the importance of the baseline susceptibility profile a predictive value for response to the new treatment (Deeks et al., 2003; Ru´ız et al., 2003). These studies indicated that only treatments for which the viruses were susceptible before the TI will result in success. In our study, although a correlation between the baseline number of PI-resistance mutations and response to T2 was found, the baseline susceptibility profile was not predictive for the success of the subsequent therapy. Two of the responders (pats. #10 and #11) had at baseline viruses fully susceptible to the PI they received afterwards in T2 , while for the other 4 responders (#2, #5, #13 and #14) a resensitisation occurred during TI. Our study illustrates the complexity of the consequences of treatment interruptions. The obsevation of different viral evolution and furthermore the detection of resistance pathways in a reverse direction could help to understand and prevent therapy failure. This may lead to selection of a new effective therapy regimen after therapy interruption.

5. Conclusions Therapeutical benefit of treatment interruptions is discussed controversially, but nevertheless therapy interruptions do occur. They are a consequence of resistance development, severe toxicity, pharmacokinetic interactions or patients’ per-

285

sonal decision. In our study, six patients responded to the newly initiated treatment after TI, while eight patients did not. We focussed on the evolution of the genotypic patterns before, during and after TI to gain new insights into different resistance pathways and resistance development in general. After cessation of the medication, virus-strains can evolve according to different mutation patterns. Depending on the evolution of virus-strains during TI the interpretation of susceptibility to drugs was possible and had a high likelihood. A prediction of drug susceptibility by any of the clinical baseline parameters was not possible. The resistance mutation- and drug susceptibility-profiles achieved during the TI have predictable value for the success of the subsequent therapy. A close clinical and virological follow-up of the patient throughout the period off-therapy should be carried out in order to observe evolution of resistance mutations. The use of genotype interpretation tools to analyse the genotype during TI can provide indications to establish a subsequent effective treatment. Future interpretation systems that combine resistance-prediction tools with other data like toxicity, plasma levels, dosage or boosting of the drugs, or other patient characteristics would be an invaluable help in the implementation of an individual patient-optimised antiretroviral regimen.

Acknowledgements We gratefully acknowledge the help of A. Kroidl, L. Kajala, R. Leidel, A. Woehrmann, N. Qurishi and N. Schmeisser in delivering the clinical data, and D. Hammerschmidt for unvaluable help in sample processing. This work was supported by grants from DFG (H01582, KA1569) and BMBF (0312708).

References Balduin M, Sierra S, D¨aumer M, Rockstroh JK, Oette M, F¨atkenheuer G, et al. Evolution of NRTI resistance patterns in heavily pre-treated HIV-1-infected patients undergoing treatment interruption. In: Proceedings of the XIII International HIV Drug Resistance Workshop: Basic Principles and Clinical Implications, 8–12 June 2004; 2004, Abstract 67. Bartlett JA, DeMasi R, Quinn J, Moxham C, Rousseau F. Overview of the effectiveness of triple combination therapy in antiretroviral-naive HIV-1 infected adults. AIDS 2001;15:1369–77. Baxter JD, Mayers DL, Wentworth DN, Neaton JD, Hoover ML, Winters MA, et al. A randomized study of antiretroviral management based on plasma genotypic antiretroviral resistance testing in patients failing therapy CPCRA 046 Study Team for the Terry Beirn Community Programs for Clinical Research on AIDS. AIDS 2000;14:F83–93. Beerenwinkel N, Daumer M, Oette M, Korn K, Hoffmann D, Kaiser R, et al. Geno2pheno: estimating phenotypic drug resistance from HIV-1 genotypes. Nucleic Acids Res 2003;31:3850–5. Beerenwinkel N, Rahnenfuhrer J, Daumer M, Hoffmann D, Kaiser R, Selbig J, et al. Learning multiple evolutionary pathways from crosssectional data. In: Proceedings of the 8th Annual International Con-

286

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287

ference on Research in Computational Molecular Biology; 2004. p. 36–44. Benson C, Downey GV, Havlir DV, Vaida F, Lederman M, Gulick R, et al., Team tAAS. A 16-week treatment interruption does not improve the virologic response to multidrug salvage therapy in treatmentexperienced patients: 48-week results from ACTG A5086. In: Proceedings of the 11th Conference on Retroviruses and Opportunistic Infections, 8–11 February 2004; 2004, Abstract 58. Benson CA. Structured treatment interruption in HIV infection. AIDS Read 2001;11:99–102. Bonhoeffer S, Rembiszewski M, Ortiz GM, Nixon DF. Risks and benefits of structured antiretroviral drug therapy interruptions in HIV-1 infection. AIDS 2000;14:2313–22. Boucher CA, O’Sullivan E, Mulder JW, Ramautarsing C, Kellam P, Darby G, et al. Ordered appearance of zidovudine resistance mutations during treatment of 18 human immunodeficiency virus-positive subjects. J Infect Dis 1992;165:105–10. Carrieri P, Spire B, Duran S, Katlama C, Peyramond D, Francois C, et al. Health-related quality of life after 1 year of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 2003;32:38–47. Chun TW, Davey Jr RT, Engel D, Lane HC, Fauci AS. Re-emergence of HIV after stopping therapy. Nature 1999;401:874–5. Chun TW, Davey Jr RT, Ostrowski M, Shawn Justement J, Engel D, Mullins JI, et al. Relationship between pre-existing viral reservoirs and the re-emergence of plasma viremia after discontinuation of highly active anti-retroviral therapy. Nat Med 2000;6:757–61. Cingolani A, Antinori A, Rizzo MG, Murri R, Ammassari A, Baldini F, et al. Usefulness of monitoring HIV drug resistance and adherence in individuals failing highly active antiretroviral therapy: a randomized study (ARGENTA). AIDS 2002;16:369–79. d’Arminio-Monforte A, Lepri AC, Rezza G, Pezzotti P, Antinori A, Phillips AN, et al. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients I. CO. N. A. Study Group. Italian Cohort of Antiretroviral-Naive Patients. AIDS 2000;14:499–507. Deeks SG, Wrin T, Liegler T, Hoh R, Hayden M, Barbour JD, et al. Virologic and immunologic consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viremia. N Engl J Med 2001;344:472–80. Deeks SG, Grant RM, Wrin T, Paxinos EE, Liegler T, Hoh R, et al. Persistence of drug-resistant HIV-1 after a structured treatment interruption and its impact on treatment response. AIDS 2003;17:361–70. Delaugerre C, Valantin MA, Mouroux M, Bonmarchand M, Carcelain G, Duvivier C, et al. Re-occurrence of HIV-1 drug mutations after treatment re-initiation following interruption in patients with multiple treatment failure. AIDS 2001;15:2189–91. Devereux HL, Youle M, Johnson MA, Loveday C. Rapid decline in detectability of HIV-1 drug resistance mutations after stopping therapy. AIDS 1999;13:F123–7. Dieterich DT. Long-term complications of nucleoside reverse transcriptase inhibitor therapy. AIDS Read 2003;13:176–84, 187. Dorman KS, Kaplan AH, Lange K, Sinsheimer JS. Mutation takes no vacation: can structured treatment interruptions increase the risk of drug-resistant HIV-1? J Acquir Immune Defic Syndr 2000;25:398–402. Duran S, Saves M, Spire B, Cailleton V, Sobel A, Carrieri P, et al. Failure to maintain long-term adherence to highly active antiretroviral therapy: the role of lipodystrophy. AIDS 2001;15:2441–4. Durant J, Clevenbergh P, Halfon P, Delgiudice P, Porsin S, Simonet P, et al. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet 1999;353:2195–9. Falkensammer B, El Attal R, Mullauer B, Sarcletti M, Parson W, Dierich MP, et al. Loss of zidovudine related mutations in the reverse transcriptase gene of HIV after switching therapy. Eur J Clin Microbiol Infect Dis 2002;7:379–86. Gianotti N, Soria A, Galli L, Castagna A, Cernuschi M, Hasson H, et al. Predictors of sustained response to therapy resumption after treatment

interruption in HIV-infected patients failing antiretroviral therapy. J Med Virol 2004;72:181–6. Girard PM, Schneider V, Dehee A, Mariot P, Jacomet C, Delphin N, et al. Treatment interruption after one year of triple nucleoside analogue therapy for primary HIV infection. AIDS 2001;15:275–7. Halfon P, Durant J, Clevenbergh P, Carsenti H, Celis L, Khiri H, et al. Kinetics of disappearance of resistance mutations and reappearance of wild-type during structured treatment interruptions. AIDS 2003;17:1351–61. Hance AJ, Lemiale V, Izopet J, Lecossier D, Joly V, Massip P, et al. Changes in human immunodeficiency virus type 1 populations after treatment interruption in patients failing antiretroviral therapy. J Virol 2001;75:6410–7. Hogg RS, O’Shaughnessy MV, Gataric N, Yip B, Craib K, Schechter MT, et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997;349:1294. Imamichi H, Crandall KA, Natarajan V, Jiang MK, Dewar RL, Berg S, et al. Human immunodeficiency virus type 1 quasi species that rebound after discontinuation of highly active antiretroviral therapy are similar to the viral quasi species present before initiation of therapy. J Infect Dis 2001;183:36–50. Izopet J, Massip P, Souyris C, Sandres K, Puissant B, Obadia M, et al. Shift in HIV resistance genotype after treatment interruption and short-term antiviral effect following a new salvage regimen. AIDS 2000;14:2247–55. Izopet J, Souyris C, Hance A, Sandres-Saune K, Alvarez M, Pasquier C, et al. Evolution of human immunodeficiency virus type 1 populations after resumption of therapy following treatment interruption and shift in resistance genotype. J Infect Dis 2002;185:1506–10. Katlama C, Dom´ınguez S, Duvivier C, Delaugerre C, Peytavin G, Legrand M, et al. Long-term benefit of treatment interruption in salvage therapy (GIGHAART ANRS 097). In: Proceedings of the 10th Conference on Retroviruses and Opportunistic Infections, 11–14 February, 2003; 2003, Abstract 68. Kijak GH, Simon V, Balfe P, Vanderhoeven J, Pampuro SE, Zala C, et al. Origin of human immunodeficiency virus type 1 quasispecies emerging after antiretroviral treatment interruption in patients with therapeutic failure. J Virol 2002;76:7000–9. Larder BA, Kemp SD. Multiple mutations in HIV-1 reverse transcriptase confer high-level resistance to zidovudine (AZT). Science 1989;246:1155–8. Lawrence J, Mayers DL, Hullsiek KH, Collins G, Abrams DI, Reisler RB, et al. Structured treatment interruption in patients with multidrug-resistant human immunodeficiency virus. N Engl J Med 2003;349:837–46. Metzner KJ, Bonhoeffer S, Fischer M, Karanicolas R, Weber R, Hirschel B, et al. Detection of minor populations of drug-resistant-viruses in patients undergoing structured treatment interruptions. Antivir Ther 2002;7:S73. Miller V, Sabin C, Hertogs K, Bloor S, Martinez-Picado J, D’Aquila R, et al. Virological and immunological effects of treatment interruptions in HIV-1 infected patients with treatment failure. AIDS 2000;14:2857–67. Mocroft A, Youle M, Moore A, Sabin CA, Madge S, Lepri AC, et al. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment center. AIDS 2001;15:185–94. Oette M, Kaiser R, H¨aussinger D. Resistenz in der HIV-Therapie: Diagnostik und Management. Bremen: UNI-MED; 2003. Ort´ız GM, Wellons M, Brancato J, Vo HT, Zinn RL, Clarkson DE, et al. Structured antiretroviral treatment interruptions in chronically HIV-1infected subjects. Proc Natl Acad Sci USA 2001;98:13288–93. Oxenius A, Price DA, Dawson SJ, G¨unthard HF, Fischer M, Perrin L, et al. Residual HIV-specific CD4 and CD8 T cell frequencies after prolonged antiretroviral therapy reflect pretreatment plasma virus load. AIDS 2002;16:2317–22. Palella Jr FJ, Delaney KM, Moorman AC, Loveless MO, Fuhrer J, Satten GA, et al. Declining morbidity and mortality among patients with

M. Balduin et al. / Journal of Clinical Virology 34 (2005) 277–287 advanced human immunodeficiency virus infection HIV Outpatient Study Investigators. N Engl J Med 1998;338:853–60. Richman D, Bozzette S, Morton S, Chien S, Wrin T, Dawson K, et al. The prevalence of antiretroviral drug resistance in the US. In: Proceedings of the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC 2001); 2001, Abstract LB-17. Richman DD, Morton SC, Wrin T, Hellmann N, Berry S, Shapiro MF, et al. The prevalence of antiretroviral drug resistance in the United States. AIDS 2004;18:1393–401. Ru´ız L, Mart´ınez-Picado J, Romeu J, Paredes R, Zayat MK, Marfil S, et al. Structured treatment interruption in chronically HIV-1 infected patients after long-term viral suppression. AIDS 2000;14:397–403. Ru´ız L, Ribera E, Bonjoch A, Romeu J, Mart´Inez-Picado J, Paredes R, et al. Role of structured treatment interruption before a 5-drug salvage antiretroviral regimen: the Retrogene Study. J Infect Dis 2003;188:977–85.

287

Shafer RW, Gonzales MJ, Brun-Vezinet F. Online comparison of HIV-1 drug resistance algorithms identifies rates and causes of discordant interpretations. Antivir Ther 2001;6:101. Tuldr´a A, Fumaz CR, Ferrer MJ, Paredes R, Romeu J, Ru´ız L, et al. Psychological impact of structured treatment interruptions in patients with prolonged undetectable HIV-1 viral loads. AIDS 2001;15: 1904–6. Verhofstede C, Wanzeele FV, Van Der Gucht B, De Cabooter N, Plum J. Interruption of reverse transcriptase inhibitors or a switch from reverse transcriptase to protease inhibitors resulted in a fast reappearance of virus strains with a reverse transcriptase inhibitor-sensitive genotype. AIDS 1999;13:2541–6. Yozviak JL, Doerfler RE, Woodward WC. Resuppression of virus load after interruption in treatment with nevirapine and 2 nucleoside reverse-transcriptase inhibitors. Clin Infect Dis 2002;34:547– 50.