The Sixteenth Conference on Retroviruses and Opportunistic Infections heard today that two large, seven-year-long, international trials of the immune-stimulating drug interleukin-2 (IL-2, Proleukin) in people with HIV had produced a negative result. Both trials found that there was no clinical benefit in using the injectable therapy and that it produced significant serious side-effects. Exactly why these trials, which involved 5800 participants and cost $85 million, made no difference in terms of deaths or AIDS diagnoses, despite producing CD4 cell increases in participants, is still unknown and may lead to other discoveries about the immune system.
Background to both studies
IL-2 is a cytokine, a natural substance produced by T-cells, which can also be made artificially. It stimulates the production and maturation of CD4 T-cells. Investigators have been interested for 15 years in whether it is beneficial to treat people with HIV with IL-2 to boost immune functioning and delay disease progression.
ESPRIT (Evaluation of Subcutaneous Proleukin in a Randomized International Trial) and SILCAAT (Subcutaneous IL-2 in patients with Low CD4 Counts under Active Antiretroviral Therapy) were international, open-label phase III studies examining the effects of IL-2 injections on the progression of HIV disease and death.
The main difference between the two studies was the CD4 count of study participants at entry: ESPRIT participants had to have a CD4 cell count of at least 300 when they entered the study, whereas SILCAAT participants had a CD4 cell count between 50 and 300.
In both studies, half the participants received injections of IL-2 under the skin alongside their antiretroviral therapy (ART), while the rest received ART alone. In ESPRIT, participants received three five-day courses (‘cycles’) of IL-2 injections at eight-week intervals, followed by additional cycles, as required, to keep CD4 counts either at least twice the baseline level or over 1000. In SILCAAT, participants received six five-day cycles of a lower dose of IL-2 at eight-week intervals, followed by additional cycles as required to maintain CD4 cell counts an average of 150 above baseline.
The primary endpoint – the main outcome measured – for both studies was death from any cause or AIDS-related illness; serious non-AIDS events and serious clinical reactions to treatment were also measured.
ESPRIT included 4111 participants from 252 clinical sites in 25 countries, and SILCAAT 1695 patients from 114 sites in eleven countries. The studies needed a minimum number of primary endpoints in order to be able to demonstrate a significant effect. In ESPRIT, 320 primary endpoints were required in order to demonstrate whether or not it worked: in fact, 323, affecting 7.8% of participants, were seen. In SILCAAT, 300 were required but only 227 had been seen by November last year, affecting 13.4% of participants. At that point the trial’s Data and Safety Monitoring Board decided SILCAAT was unlikely to reach the number of events needed, and it was decided to close the trial on 18 November.
The participants in both trials were similar in some respects. Their average age was around 40 and they had been on ART for about four years, with 80% having an undetectable viral load at trial entry. Twenty-six per cent of ESPRIT patients and 32% of SILCAAT patients had had an AIDS diagnosis. By the time the studies had closed the average time spent in the study was 6.9 years in ESPRIT and 7.6 years in SILCAAT. The ESPRIT participants’ average CD4 count at trial entry was 457 and their mean lowest-ever CD4 count (the CD4 nadir) was 197. In SILCAAT, participants’ average CD4 count at entry was 202 and their CD4 nadir was 60.
After seven years, the bottom-line finding was that in both trials and for all endpoints (with one crucial exception) there was no difference in the clinical outcomes for participants on IL-2 and those on placebo, despite higher CD4 gains in the IL-2 patients.
In ESPRIT, participants receiving IL-2, averaged over the whole trial period, had 153 more CD4 cells than those not on IL-2, and after six years their count was 128 cells higher. Both differences were statistically significant (p
In SILCAAT a similar pattern was seen, except that IL-2 had less of an effect. Averaged over the trial period IL-2 recipients had 67 CD4 cells more than those on ART alone, and at six years had 37 cells more. Although this difference was small, it was still statistically significant. CD4 counts in IL-2 recipients rose from 197 to around 320 in the first year but then stayed the same, never exceeding 340. In contrast, the CD4 count in ART-only subjects rose slowly but steadily from 197 to around 320 over the six years.
No differences were seen in the number and rates of opportunistic disease or death in either study. In ESPRIT the number of patients developing AIDS or dying (of any cause) was 158 on IL-2 (107 deaths) and 165 on ART (114 deaths). In SILCAAT the figures were 109 on IL-2 (80 deaths) and 118 on ART (75 deaths). The difference in these figures was not statistically significant. The figures for non-AIDS related illnesses were also the same in IL-2 recipients and those on IL-2 alone.
The only difference seen between the two arms was in the number and rates of grade-4 clinical adverse events (AEs) – serious side-effects, in other words. Grade 4 means life-threatening. In ESPRIT 466 IL-2 recipients developed grade-4 AEs versus 383 on ART alone. This 23% difference was statistically significant (p = 0.03). In SILCAAT there was no difference in grade-4 AEs over the whole study: there were 199 in patients on IL-2 and 184 on ART alone. However, during the first year, which was when most of the IL-2 was administered, there was a difference, with twice as many patients on IL-2 getting AEs (38 versus 19). This was significant (p = 0.01). There were also significantly more IL-2 patients over the whole trial developing gastrointestinal and psychiatric symptoms.
The main grade-4 AE seen in IL-2 patients that was not seen in ART patients was deep vein thrombosis (DVT). This comprised 80% of the excess AEs in both trials, with 40 cases of DVT in ESPRIT participants on IL-2 versus 14 on ART alone (p = 0.001). DVT was twice as common in SILCAAT and nearly three times as common in ESPRIT among IL-2 recipients. Serious psychiatric symptoms were also 50% more common in IL-2 recipients in ESPRIT and over twice as common in SILCAAT.
Verdict on the trials
The verdict of the trial investigators was definitive. ESPRIT investigator Marcelo Losso told the conference that, “Any potential benefit of IL-2, however moderate, can be ruled out.” SILCAAT investigator Yves Lévy was if anything even more emphatic. When asked at a press conference if this was “the death knell for IL-2”, he replied: “I think so. I don’t see any possible benefit in continuing with IL-2 studies.”
Both investigators were keen, however, to emphasise that they did not regard the trials as failures. There had been strong laboratory data gathered in the early 1990s that suggested the CD4 gains produced by IL-2 might have an effect on clinical progression, but large phase III trials were the only way to find out.
The fact that the CD4 gains did not translate into clinical benefit was itself a puzzling finding, which needed to be investigated further. It was calculated that the kind of cell increases seen should have resulted in 27% fewer deaths and AIDS diagnoses, but this was not seen. Losso and Lévy suggested two explanations. Firstly, any clinical benefit of IL-2 might simply be cancelled out by its adverse effects. Alternatively, the kind of CD4 T-cells produced by IL-2 were not as effective in fighting disease as the ones arising from the suppression of HIV. A third explanation might be that the body has a finite reservoir of embryo T-cells and all that IL-2 therapy was doing was depleting it faster.
DVTs and another common side-effect, high blood pressure, might be caused by IL-2 inflaming the lining of blood vessels. Since the studies started it has been discovered that both HIV infection and some drugs (such as abacavir) do this, and it might be possible to analyse stored samples to find out if there are higher levels of inflammatory marker chemicals such as D-dimer in IL-2 recipients.
The investigators also emphasised that it was not the end of the cytokine story, even if it was the end of IL-2. Yves Lévy is investigating another one called IL-7 which appears to have fewer side effects. Marcelo Losso commented: “The lesson is that when you perturb the immune system in some way you may be surprised by the clinical outcomes you get. This may apply to IL-7 too.”
Losso M et al. Effect of Interleukin-2 on clinical outcomes in patients with a CD4+ cell count of 300/mm3: primary results of the ESPRIT study. Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, abstract 90aLB, 2009.
Lévy Y et al. Effect of Interleukin-2 on clinical outcomes in patients with CD4+ cell count 50 to 299/mm3: primary results of the SILCAAT study. Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, abstract 90bLB, 2009.