Pipeline compounds in Phase 2/3 Studies

Simon Collins, HIV i-Base

Early data was presented on several new agents at the first of the opening oral sessions on new antiretrovirals.

Many of these new molecules included new targets and mechanisms of action, but were also still only in pre-clinical stages of development. These included pyranodiprymidine integrase inhibitors, AK605 CCR5 inhibitor and TAK-220 and UK-427,857 CCR5 agonists and PA-457, a new budding inhibitor (see abstract 9, 10, 11, 12 and14).

There were also important new results from compounds already in Phase 2/3 studies in HIV-positive people that are likely to be available to treatment-experienced patients in the UK in clinical trials, expanded access or as licensed drugs over the next year.


TMC-114 is a protease inhibitor developed by Tibotec that has strong antiviral activity against a broad range of protease-resistant virus. In a Phase II, open label dose finding study, 50 multiple treatment experienced patients were randomised to one of four arms – adding BID doses of 300, 600 or 900mg TMC-114, with 100mg ritonavir BID for 14-days to their regimen (arms A, B, C), or to continue their existing combination (arm D). [1]

Baseline median viral load and CD4 counts were 4.3 log and 297 cells/mm3 median number of three primary PI mutations (0-5) and six additional PI mutations (2-11) between the four arms. Almost half the patients were resistant to all PIs at baseline, with approximately 25% sensitive to only one PI and 25% sensitive to ≥2 PIs.

Median viral load change from baseline to day 14 was -1.24 (-0.51 to –2.06), -1.50 (-0.47 to –2.49), -1.13 (0.49 to –2.13) and +0.02 (+0.37 to –0.41) in the 300, 600, 900 and placebo arms respectively. 69%, 92%, 69% and 17% had at least a >1log reduction and in the TMC-114/r groups 97% of patients had at least a 0.5 log reduction.

One patient in the 900mg arm discontinued treatment due to GI discomfort and one patient in the 600mg arm due to hepatitis. Grade 3/4 ALT, AST or GGT enzymes occurred in two, two, and one of arms A, B, and C respectively with one Grade 3 ALT in the control group. One patient had a G4 rash. Grade ≥2 triglycerides and cholesterol increases occurred in eight and 12 of the TMC-114 group compared to three and one of the control group respectively.

T-20 and T-1249

The next antiretroviral agent to be licensed will be T-20 (long expected in Europe to be in June or July, but possibly now delayed to December). The TORO 1 and TORO 2 studies showed clear benefit of using T-20 over optimised background therapy alone when previously presented individually at the Barcelona, ICAAC and Glasgow conferences, so it was no great surprise that the pooled data presented at CROI showed similar results. [2]

One poster focused on the injection site reactions associated with T-20 administration and reported pathological changes following biopsy from seven patients. T-20 was detected in all samples and concentration correlated positively with areas of greatest inflammatory response. [3]

Other findings reported that this response, consistent with a localised hypersensitivity reaction, was seen in all patients, including patients with erythema without nodules, palpable nodules or no clinical reaction; that the pattern of eosinophil and histiocyte proliferation and collagen degeneration resembled that of granuloma annulare and the recently described interstitial granulomatous drug reaction, and that there was no relation between the clinical reaction and the degree or localisation of inflammation, or the depth of injection.

Patients on early studies have reported that nodules are very slow to disappear, and although Roche suggests that they have been seen less frequently since the introduction of a more recent formulation that requires less preparation time (nodules may be associated with a drug that was not entirely constituted), the poster concludes: “Due to the high incidence of this adverse reaction, clinical dermatologists should become familiar with the findings in this study prior to caring for HIV-infected patients undergoing this new mode of therapy.”

It is also becoming clear that like all other drugs, T-20 has limited benefits unless supported by activity of additional drugs in the combination. People receiving T-20 in early studies or the expanded access programme who are not able to achieve maximal viral suppression accumulate resistance to T-20, and derive limited benefit.

Preliminary data on efficacy of the second generation fusion inhibitor T-1249, which binds to a slightly different sequence of gp41 than T-20 in T-20 resistant patients, were presented as a late-breaker oral session on the second generation fusion-inhibitor from Trimeris/Roche. [4]

Fifty patients who had already developed resistance to T-20 were enrolled in the first T-1249 Phase II study, receiving 192mg QD for 10 days. Planned interim analysis was presented for the first 25 patients. Mean viral load and duration of previous T-20 therapy at study entry was 5.1 logs and 70 weeks, with mean time on T-20 with a viral load >5,000 copies was 56 weeks (range 28-136 weeks).

Median viral load change at day 11 was -1.12 (1.50-0-83, 95% CI). Fifteen patients had >-1 log drop and 19 had a drop >0.05 log. When analysed by time on previous (failing) T-20 regimen, patients with least exposure to a failing regimen (24-48 weeks) whilst only 8/17 people who had continued >48 weeks achieved >1 log drop.

One patient who had pre-existing severe chronic obstructive pulmonary disease at baseline developed an upper respiratory tract infection and died of pneumonia before the end of the trial. The two other significant adverse events were one case of elevated liver enzymes and one case of acute bronchitis.

This study demonstrates antiviral activity of a once-daily fusion inhibitor prior to the approval of T-20 and access to similar roll-over studies is essential for patients unable to derive sustained benefit from the first programme, especially as response to the new agent is limited by previous time on a failing T-20 regimen.


Longer-term efficacy and safety data were presented for atazanavir, particularly important as the profile for this once-daily azapeptide protease inhibitor has promised a kinder metabolic profile and consequent hope for a reduced risk of lipodystrophy than has yet to be seen in clinical studies.

This prospective open-label roll-over study (BMS-044) assessed the longer-term efficacy and safety of ATV beyond 72 weeks and the efficacy and safety in patients switched from nelfinavir (NFV) to atazanavir (ATV) in the earlier Phase II study (BMS-008). [5]

A total of 346 patients (63% male, 37% female) were enrolled and treated in AI424-044; the median cumulative time on therapy was approximately 108 weeks (72 weeks in 008 and 36 weeks in 044).

Viral load, CD4 and lipid effects are summarised in the table below:

ATV 400mg
044 Entry
(n = 139) Week 24 NFV to ATV 044 Entry (n = 63) Week 24
Response Criteria† Observed/Evaluable (%)
< 400 c/mL ITT 111/139 (80) 54/63 (86)
< 50 c/mL As treated 63/129 (49) 80/133 (60) 30/60 (50) 37/62 (60)
< 50 c/mL ITT 80/139 (58) 37/63 (59)
Median CD4 472 556 543 584
Total Chol * 180 (129) 176 (128) 202 (56) 169 (56)‡
Fasting LDL-C * 110 (60) 105 (86) 132 (33) 99 (40)‡
Fasting TG * 105 (103) 104 (110) 127 (47) 102 (52)‡
* Median mg/dL (n)
†As treated analysis, results maintained through 108 weeks from start of AI424-008. ITT analysis for 008/044 selected cohort
‡p < 0.0001, NFV to ATV, mean % change, week 24 vs entry

Discontinuations due to adverse events were infrequent and comparable across cohorts (ATV 400 mg, 1%; ATV 600 mg, 2%; NFV to ATV, 3%), and no new safety issues were identified after approximately 108 weeks of cumulative ATV treatment. Asymptomatic elevation in indirect bilirubin (without hepatic transaminase elevation) was the most frequent laboratory abnormality.

ATV treatment did not produce clinically relevant increases in TC, fasting LDL-C, or fasting TG. After 24 weeks of ATV, patients switching from NFV, experienced significant decreases in TC, fasting LDL-C, and fasting TG towards pre-antiretroviral treatment levels.


The Phase II study used to select the dose for development programme for tipranavir (see also January/February issue of HTB) was presented at the conference.

Gathe and colleagues presented results from an international, blinded, multicentred trial of three investigational doses of tipranavir/ritonavir dose regimens in 216 patients experienced to all three classes, and at least one primary PI mutation. [6]

The three doses of tipranavir/ritonavir (TPV/RTV) used were 500mg/100mg, 500mg/200mg and 750mg/200mg added to existing therapy for the first two weeks to assess change in viral load. Background therapy could then be optimised.

Median baseline CD4 count and viral load were 153 copies/mm3 and 4.5 log copies/mL respectively and previous PI use ranged from 37.5% for lopinavir to 79.6% for indinavir.

Median decreases in viral load by intent-to-treat analysis were -0.9 log in the 500/100mg arm, -1.0 log in the 500/200mg arm, and -1.2 log in the 750/200mg arm.

The most common adverse events associated with tipranavir were diarrhoea, nausea, fatigue, headache, vomiting and elevated liver transaminases.

A second poster analysed baseline PI susceptibility with median baseline fold WT IC50 in 157 isolates to tipranavir of 1.1 (range 0.3–100.2). This was 76.5 (0.5–165.3) for lopinavir, 8.7 (0.3–150.7) for amprenavir, 7.0 (0.1–108.9) for saquinavir, 12.2 (0.4–197.4) for indinavir, 36.8 (0.3–96.4) for nelfinavir, and 94.2 (0.2–808.5) for ritonavir. [7]

Viral isolates from 200/216 (93%) patients had >10 protease gene mutations at baseline and 41/216 (19%) patients had genotypes indicative of resistance to all currently marketed PIs. Twenty-two per cent of patients had virus with three universal protease inhibitor-associated mutations (UPAMs) – L33I/V/F, V82A/F/L/T, I84V, L90M.

The distribution of TPV susceptibility in these isolates was 42% ≤ 1-fold WT, 27% >1 to 2-fold WT, 18% >2 to 4-fold WT, and 12% >4-fold WT. The overall median viral load changes after initial two weeks were -1.23 log, -1.24 log, -0.21 log, and -0.19 log, respectively. This indicates a strong antiviral response for TPV IC50 ≤ two-fold WT, and a lesser response for TPV IC50 >two-fold WT.

As indicated by the clinical results, baseline phenotypic susceptibility to TPV was maintained in the majority of isolates that were highly resistant to available PIs. The reduced response to TPV susceptibility at an IC50 approximately two-fold WT required the accumulation of 16-20 protease gene mutations including three UPAMs.

The 500/200mg dose has been selected to take forward to Phase III RESIST 1 and 2 studies, which were planned to start enrolment from February 2003, but which have still not been approved at UK clinics.


FTC (emtricitabine, Coviracil), is a nucleoside (cytosine) analogue similar to 3TC, that was recently acquired by Gilead following a protracted development programme with Triangle Pharmaceuticals. Several posters were presented showing comparable or similar efficacy to d4T or 3TC in treatment naïve and experienced adults and children, perhaps with advantages of an improved resistance profile and once-daily dosing.

In FTC-301, 571 patients were randomised to receive FTC or d4T (285 d4T, 286 FTC) with background ddI and efavirenz. Median BL VL was 4.9 log10 and the mean BL CD4+ was 318 cells/mm3. [8]

The proportion of patients having virological failure through week 48 was 5.3% in the FTC arm and 12.7% in the d4T arm (p < 0.05). The mean increase from baseline to week 48 in CD4+ was significantly greater in the FTC arm (153 cells/mm3) than the d4T arm (120 cells/mm3) (p < 0.05).

Genotypic analysis performed on 46 of the 49 confirmed virological failures showed 32 (97%) patients had mutations in the d4T arm compared to 69% (9/13) from the FTC subgroup (p < 0.05). The M184V mutation was observed in 46% (6/13) of the FTC group while TAMs were observed in 8% (1/13) of the FTC subset and 21% (7/33) of the d4T subset. Both the ddI-associated mutations and the NNRTI mutations were less prevalent in the FTC subset, 0% and 69% (9/13) as compared to the d4T subset, 12% (4/33) and 88% (29/33), although these differences did not reach statistical significance.

Longer-term efficacy and safety results were presented by Wakeford and colleagues for 215 patients who continued to receive FTC after 48-week successful treatment in FTC-303. Of these, 152 of 294 (51%) maintained suppression of HIV-1 RNA ≤ 400 copies/mL and 139 (47%) ≤ 50 copies/mL through week 120 (2.3 years). Median time on FTC was 140 weeks and the Kaplan-Meier probability of virological failure at four years was 11%. [9]

Most adverse events were mild or moderate; the annual incidences of drug-related severe or potentially life threatening adverse events were 3% and < 1%, respectively. The annual incidences of Grade 3 and Grade 4 laboratory abnormalities were 11% and 10%, respectively. Of these, asymptomatic and transient elevations in CPK accounted for more than two-thirds of the overall Grade 4 events.

Although the cumulative virological failure in patients switching to FTC appears high (49% failure between weeks 48 and 120) understanding this is frustrated by the focus on results based on suppression to only <400 copies/ml – as partial suppression to only 50-400 copies/ml is already well understood to lead to virological rebound. The long-term safety data from this long-awaited drug is nevertheless encouraging.

Molena and colleagues reported benefits from randomising 355 patients already on treatment (with viral load <400 copies) to a once daily regimen of FTC, ddI and efavirenz or continuing existing therapy over 48 weeks. [10]

The proportion of patients with plasma HIV RNA < 50 copies/mL at week 48 was significantly higher in the once-daily arm than in the continuation arm (95% versus 87%, p = 0.01). Median CD4 count increase was similar between arms (+21 and +13/mm3) as were rates of treatment discontinuations (10.1% and 12.4% respectively).

FTC is also formulated for paediatric dosing as a palatable flavoured syrup and results from two paediatric studies were also presented.

Fifty-one naive and 31 experienced children (median four years prior ART), reported results from using FTC at a FTC dose of 6mg/kg QD in children of all age groups (range three months-17 years old) produced similar plasma levels to adults receiving 200 mg QD. [11]

In the NP group, at week 20, the median changes in HIV-1 RNA and CD4 were -2.6 log10 and +213 cells/mm3 respectively. Naïve patients received FTC with lopinavir/r and d4T and treatment experienced children substituted FTC for 3TC similar to the protocol for the adult 303 study described above.

Across both groups, by Non-Completer=Failure analysis, at week 24, 89% (73/82) of evaluable patients achieved a VL <400 copies/ml.

Virological response at week 24:

Efficacy endpoint ARTnaïve (n=51) ARTexp (n=31) Total (n=82)
%< 400 copies/mL(NC=F) 47 (92.2%) 26 (83.9%) 73(89.0%)
%< 50 copies/mL (NC=F) 32 (62.7%) 22 (71.0%) 54(66.0%)
Median change from B/line RNA -3.09 (-4.03, -0.07) -0.00 (-0.81, 0.96) -2.52 (-4.03, 0.96)

Grade 3/4 lab abnormalities were reported in 17% (14/82) of patients. Three patients discontinued the study, one for SAE (anemia), one for virologic failure, and one withdrew consent.

The incidence of serious adverse events was 13.7% (7/51) in ART-naïve patients and 16.1% (5/31) in ART-experienced patients. There were five events of at least moderate severity that were possibly or probably related to study drug (one case each of pancreatitis, vomiting, leucopenia, anaemia and pleural effusion).

Grade 3/4 lab abnormalities were reported in 6% (5/82) of patients; 2% (1/51) ART-naïve and 13% (4/31) ART-experienced patients.

PACTG 1021 is an ongoing Phase I/II study of once-daily FTC in combination with ddI and efavirenz in minimally treated or treatment naïve children. [12]

Thirty children were enrolled at 15 sites into two cohorts (three-12 years, 13-21 years), and were stratified as either therapy naïve or < six weeks of perinatal prophylaxis. At baseline, median age: 10.5 years (17 <12 years old; 13 13-21 years). Median CD4 count and percentage was 302 cells/mm3 and 16.5%. Median viral load was 50,000 copies/ml. FTC was dosed were FTC 6 mg/kg.

Viral results (ITT, discontinuation = failure): at week 16 (n = 23), 87% < 400 copies/ml, 74% < 50 copies/ml. For children >12 yrs old (n = 13), 92% < 400 copies/ml.

Three patients permanently discontinued therapy (one viral failure, one voluntary withdrawal, one Grade 3 rash). At week 16 (n=19), median CD4 increase was 220 cells (9%) with all children showing a CD4 increase from baseline.


Updated results from two Phase III studies of fosamprenavir (908) were presented at the meeting.

Nadler and colleagues reported 48-week results from an open-label randomised (2:1) study comparing 1400mg 908 BID to 1250mg nelfinavir (NFV) BID, conducted in the USA, Panama, Puerto Rico and South Africa. [13]

251 treatment naïve individuals in closely matched arms had baseline values (all approximated) including: median CD4 210 (range 2-1000) with 45% <200 cells/mm3 and 20% <50 cells/mm3; median viral load 4.8 log with 45% > 100,000 copies/ml; median age 38; approximately 30% women; 50% heterosexual; 24% White, 32% African-American, 44% Hispanic.

At week 48, by intent-to-treat analysis, 58% and 42% patients had achieved and maintained viral load reductions to <50 copies/ml in the 908 and NFV arms respectively. Analysis by stratified baseline viral load showed 56% vs 57% <100,000 and 60% vs 24% when >100,000 copies/ml in the 908 and NFV arms respectively.

The overall incidence of drug related Grade 2–4 adverse events was comparable; with only greater incidence of diarrhoea in the nelfinavir arm showing significantly different AE being higher on NFV (18% vs 5%, p = 0.002).

In practice, and in previous studies (ie the SOLO study), fosamprenavir is co-administered with low-dose ritonavir. DeJesus and colleagues randomised 315 PI-experienced patients to either 908/RTV 1400mg/200mg once daily, 908/RTV 700mg/100mg BID or to lopinavir/ritonavir BID (1:1:1) together with two sensitive RTIs, with the aim of showing non-inferiority of the 908 arms. [14]

With low baseline median baseline viral load of only 4.14 log (range not provided) efficacy was assessed by measuring time-averaged change in vRNA from baseline (AAUCMB) at both 24 and 48 weeks. At 24 weeks this was, by a confusing ‘intent-to-treat observed’ analysis, -1.46, -1.48 and -1.63 log in the 908 QD, 908 BID and LPV/r arms respectively. The upper limits of both CIs were below the pre-specified 0.5 non-inferiority margin, although these are both very short-term and limited data.

Finally, MacManus and colleagues presented data on resistance for people experiencing treatment failure on the SOLO and NEAT studies. The SOLO study (48-week results were presented as a late breaker in Glasgow) compared boosted 908 QD to NFV BID with an abacavir + 3TC background in 660 treatment naïve individuals. All failures showed evidence of ongoing replication, but a statistically significant difference in patient incidence of both PI and NRTI-associated mutations in favour of 908 (0% vs 56% and 9% vs 57% respectively). [15]

In contrast, patients failing on non-ritonavir boosted, fosamprenavir regimens in the NEAT study showed mutations characteristic of development of APV resistance in virus from 5/29 (17%) 908 treated subjects analyzed (3% subjects exposed) and included I54L/M, V32I + I47V and M46I. NFV-selected mutations (D30N, N88D/S, or L90M) were detected in 6/26 (23%) NFV-treated subjects analysed (7% subjects exposed).

Anti-CD4 mAb (TNX-355)

Finally, interesting results were shown from using a novel entry inhibitor, TNX-355, a humanised IgG4 anti-CD4 domain 2 mAb, that showed potent anti-HIV-1 activity following a single dose of TNX-355 in HIV-positive subjects. [16]

Five sequential cohorts of six HIV-1-infected subjects received single iv doses of TNX-355 in an open-label dose-escalation study (0.3, 1, 3, 10, and 25 mg/kg). Mean baseline CD4 count was 354 cells/mm3 and viral load was 5.1 log10 copies/mL. All were HAART-experienced and 19/30 were on failing HAART at entry.

Mean peak decreases in VL of 0.2, 0.68, 1.48, and 1.09 log10, occurred on days two, four, 14, and 21 for the 1, 3, 10, and 25 mg/kg dose cohorts, respectively. The number of patients achieving >1.0-log10 decrease was nought out of six, two out of six, five out of six and five out of six for the 1, 3, 10, and 25 mg/kg cohorts, respectively. Duration of complete CD4 cell coating with TNX-355 ranged from one to two days at 1 mg/kg to 15–27 days at 25 mg/kg, and correlated with day of VL nadir. TNX-355 was well tolerated. No serious adverse events were reported. CD4 cell depletion was not observed in any cohort.

Further assessment of therapeutic potential awaits data from longer duration trials; a phase 1b multiple- dose study is planned.


Unless stated otherwise, all references are to the Programme and Abstracts of the 10th Conference on Retroviruses and Opportunisitc Infections (CROI), 10–14 February 2003, Boston.

  1. Arasteh K, Clumeck N, Pozniak A et al. First clinical results on antiretroviral activity, pharmacokinetics, and safety of TMC114, an HIV-1 protease inhibitor, in multiple PI-experienced patients. Abstract 8
  2. Greenberg ML, Melby T, Sista P et al. Baseline and on-treatment susceptibility to enfuvirtide seen in TORO 1 and TORO 2 to 24 weeks. Abstract 141.
  3. Ball RA, Kinchelow T. Pathology of injection site reactions with enfuvirtide. Abstract 714.
  4. Miralles GD, Lalezari JP, Bellos N et al. T-1249 demonstrates potent antiviral activity over 10 day dosing in most patients who have failed a regimen containing enfuvirtide (ENF): planned interim analysis of T1249-102, a Phase I/II study. Abstract 14lb.
  5. Murphy R, Pokrovsky V, Rozenbaum W et al. Long-term efficacy and safety of atazanavir with stavudine and lamivudine in patients previously treated with nelfinavir or ATV: 108-week results of BMS Study 008/044. Abstract 555
  6. Gathe J, Kohlbrenner VM, Mayers D et al. Tipranavir/ritonavir demonstrates potent efficacy in multiple protease inhibitor experienced patients: BI 1182.52. Abstract 179.
  7. Cooper D, Hall D, Jayaweera D et al. Baseline phenotypic suceptibility to tipranavir/ritonavir is retained in Isolates from patients with multiple protease inhibitor experience (BI 1182.52). Abstract 596.
  8. Cahn P, Raffi F, Saag M et al. Virologic efficacy and patterns of resistance mutations in ART-naïve patients receiving combination therapy with once-daily emtricitabine compared to twice-daily stavudine in a randomized, double-blind, multi-center clinical trial. Abstract 606.
  9. Wakeford C, Shen G, Rousseau F et al. Long-term efficacy and safety of emtricitabine in HIV+ adults switching from a lamivudine containing HAART egimen. Abstract 550
  10. Molina JM, Ferchal F, Rancinan C et al. Once-daily combination of emtricitabine, didanosine, and efavirenz vs continued PI-based HAART in HIV-infected adults with undetectable plasma HIV-RNA: 48-week results of a prospective randomized multicenter trial (ALIZE-ANRS 99).
  11. Saez-Llorens X, Violari A, Ndiweni D et al .Once-daily emtricitabine in HIV-infected pediatric patients with other antiretroviral agents. Abstract 872.
  12. PACTG 1021: An ongoing Phase I/II Study of once-daily emtricabine, didanosine, and efavirenz in therapy-naive or minimally treated pediatric patients. Abstract 873.
  13. Nadler J, Rodriguez-French A, Millard J et al. The NEAT study: GW433908 efficacy and safety in ART-naïve subjects, final 48-week analysis. Abstract 177.
  14. DeJesus E, LaMarca A, Sension M et al. The context study: efficacy and safety of GW433908/RTV in PI-experienced subjects with virological failure (24 week results). Abstract 178.
  15. Macmanus S, Yates P, White S et al. GW433908 in ART-naive subjects: absence of resistance at 48 weeks with boosted regimen and APV-like resistance profile with unboosted regimen. Abstract 598.
  16. Kuritzkes DR, Jacobson JM, Powderly W et al. Safety and preliminary anti-HIV activity of an anti-CD4 mAb (TNX-355; formerly Hu5A8) in HIV-infected patients. Abstract 13.

Links to other websites are current at date of posting but not maintained.