AACR 2026 – a CCR8 reality check

20 April 2026

Amgen and Gilead deliver lacklustre early clinical results with AMG 355 and denikitug respectively.

Since companies started the first clinical trials with anti-CCR8 molecules in 2021, data have come out slowly and in small doses. But at this year’s AACR meeting two companies, Amgen and Gilead, are presenting the first results with their respective molecules, AMG 355 and denikitug, giving the field a bit more clarity. The picture isn’t pretty. For Amgen, the story was stark: AMG 355 failed to demonstrate meaningful activity even when paired with Keytruda, a shortcoming that prompted the company to halt its phase 1 trial at just 77 patients, a fraction of the 535 originally planned. Gilead’s results were not dramatically better, though the company did manage to clock a few more responses, and combination data with the anti-PD-1 MAb zimberelimab are still to come.Limited clinical activityAMG 355’s results, presented at AACR on Saturday, showed just two responses among 77 patients in the phase 1 solid tumour stud, in a melanoma patient in the monotherapy arm and in a gastric cancer patient in the combination cohort. Doses ranged from 0.28mg to 700mg for monotherapy, and 70-700mg plus 200mg of Keytruda for the combination arm. About 40% of patients had prior exposure to PD-(L)1 inhibitors. The presenter, Dr Marwan Fakih of City of Hope, admitted that clinical activity was “limited”.Treatment-related adverse events (TRAEs) occurred in 75% and 80% of patients in the monotherapy and combination arms respectively, including 4% and 17% respective rates of grade ≥3 TRAEs; these included signs of drug-induced liver injury.DenikitugThings looked somewhat more favourable for Gilead’s denikitug, though the available data cover only the monotherapy arm so far. Results are set to be reported at AACR on Tuesday, but the the abstract shows four responses among 52 evaluable patients treated at 1-100mg.This is still only an 8% response rate, but arguably a cleaner signal than with AMG 355. Responses came in patients with NSCLC, triple-negative breast cancer, vaginal squamous cell carcinoma, gastroesophageal junction adenocarcinoma. Results from the combination cohort with zimberelimab are still pending. On the safety side, the abstract does not report the overall proportion of patients who experienced adverse events; however, the most common TRAEs were pruritus and maculopapular rash (37% each), diarrhoea (25%), fatigue (23%) and decreased appetite (13%). There were no grade 4 or 5 treatment-related adverse events.Gilead expanded the trial from 62 to around 412 planned participants, a signal that it sees enough here to keep pushing, at least until it has the combination data in hand.Elsewhere in CCR8, data have been mixed, with perhaps the strongest single-agent signal so far coming from Qilu's QLP2117, which showed a 33% response rate among nine patients treated at the highest dose in a phase 1 basket trial. However, other assets have fallen by the wayside, including BeOne's BGB-A3055, Zai Lab's ZL-1218 and Bayer's lanerkitug.

The picture isn’t pretty. For Amgen, the story was stark: AMG 355 failed to demonstrate meaningful activity even when paired with Keytruda, a shortcoming that prompted the company to halt its phase 1 trial at just 77 patients, a fraction of the 535 originally planned. Gilead’s results were not dramatically better, though the company did manage to clock a few more responses, and combination data with the anti-PD-1 MAb zimberelimab are still to come.

Limited clinical activity

AMG 355’s results, presented at AACR on Saturday, showed just two responses among 77 patients in the phase 1 solid tumour stud, in a melanoma patient in the monotherapy arm and in a gastric cancer patient in the combination cohort.

Doses ranged from 0.28mg to 700mg for monotherapy, and 70-700mg plus 200mg of Keytruda for the combination arm. About 40% of patients had prior exposure to PD-(L)1 inhibitors. The presenter, Dr Marwan Fakih of City of Hope, admitted that clinical activity was “limited”.

Treatment-related adverse events (TRAEs) occurred in 75% and 80% of patients in the monotherapy and combination arms respectively, including 4% and 17% respective rates of grade ≥3 TRAEs; these included signs of drug-induced liver injury.

Denikitug

Things looked somewhat more favourable for Gilead’s denikitug, though the available data cover only the monotherapy arm so far. Results are set to be reported at AACR on Tuesday, but the the abstract shows four responses among 52 evaluable patients treated at 1-100mg.

This is still only an 8% response rate, but arguably a cleaner signal than with AMG 355. Responses came in patients with NSCLC, triple-negative breast cancer, vaginal squamous cell carcinoma, gastroesophageal junction adenocarcinoma. Results from the combination cohort with zimberelimab are still pending.

On the safety side, the abstract does not report the overall proportion of patients who experienced adverse events; however, the most common TRAEs were pruritus and maculopapular rash (37% each), diarrhoea (25%), fatigue (23%) and decreased appetite (13%). There were no grade 4 or 5 treatment-related adverse events.

Gilead expanded the trial from 62 to around 412 planned participants, a signal that it sees enough here to keep pushing, at least until it has the combination data in hand.

Elsewhere in CCR8, data have been mixed, with perhaps the strongest single-agent signal so far coming from Qilu's QLP2117, which showed a 33% response rate among nine patients treated at the highest dose in a phase 1 basket trial. However, other assets have fallen by the wayside, including BeOne's BGB-A3055, Zai Lab's ZL-1218 and Bayer's lanerkitug.