CD71-Directed Probody-Drug Conjugate Shows Safety, Signals of Activity in Advanced Solid Tumors

Melissa Johnson, MD

The CD71-directed probody-drug conjugate (PDC) of monomethyl auristatin E (MMAE) CX-2029 showed tolerability and antitumor activity in patients with advanced cancer, according to findings from the first-in-human phase 1/2 PROCLAIM-CX-2029 trial (NCT03543813) that were presented during the 2020 ASCO Virtual Scientific Program.

Among 45 evaluable patients, at a dose of 2 mg/kg or higher, 2 patients with squamous non—small cell lung cancer (NSCLC) and 1 patient with head and neck squamous cell carcinoma (HNSCC) experienced a confirmed partial response (PR). Additionally, 18 patients experienced stable disease, including 1 patient with ocular melanoma.

“The results of this first-in-human trial validate CD71 or the transferrin receptor as a viable therapeutic cancer target,” lead study author Melissa L. Johnson, MD, associate director of Lung Cancer Research at Sarah Cannon Research Institute, said in a virtual presentation during the meeting. “Probody technology enables targeting CD71, a previously undruggable antibody-drug conjugate (ADC) target, at tolerable doses associated with clinical activity.”

CD71 is a transmembrane glycoprotein that internalizes iron-bound transferrin. The target is highly expressed on tumor cells and normal cells, resulting in lethal on-target toxicity with traditional targeted strategies such as ADCs. PDCs are recombinant antibody prodrugs that are activated by tumor-associated proteases, thereby sparing normal cells from on-target effects.

CX-2029 is a masked humanized monoclonal antibody that binds to CD71 in the presence of tumor-associated proteases. The agent is conjugated to MMAE, a cytotoxic microtubule payload in a 2:1 drug to antibody ratio.

In preclinical studies, CX-2029 demonstrated potent activity in mouse models and tolerability in cynomolgus monkeys. In comparison, the unmasked ADC was tolerated at 0.6 mg/kg but became lethal at a dose of 2.0 mg/kg, whereas the masked PDC was tolerated up to a dose of 6.0 mg/kg.

In PROCLAIM-CX-2029, patients with metastatic or locally advanced unresectable solid tumors and an ECOG performance status (PS) of 0 or 1 who had received at least 1 prior line of systemic therapy were enrolled into 1 of 8 dose-escalation cohorts. Patients with stable brain metastases were allowed to enroll on trial. Patients with transfusion-dependent anemia or iron metabolism and grade 2 or higher neuropathy were excluded from the trial in anticipation of expected MMAE toxicities.

Given the unknown potential for toxicity in humans, the starting dose was 1/20 rather than the typical 1/6 of the highest non-severely toxic dose from preclinical models. The PDC was given intravenously in 3-week cycles, across 8 cohorts: 0.1 mg/kg (n = 3), 0.25 mg/kg (n = 3), 0.5 mg/kg (n = 6), 1.0 mg/kg (n = 3), 2.0 mg/kg (n = 8), 3.0 mg/kg (n = 12), 4.0 mg/kg (n = 6), and 5.0 mg/kg (n = 4).

The nonclinical pharmacokinetic and toxicology profile of the agent predicted for a tolerable dose range of 2.0 mg/kg to 4.0 mg/kg in patients with hematopoietic suppression and neuropathy, both of which are expected of an MMAE payload.

The maximum-tolerated dose (MTD), safety, antitumor activity, and potential biomarkers of response in the plasma and tissue served as key end points of the trial.

At the time of data cutoff in April 2020, 45 patients were enrolled, the majority of whom were male (62%) and had an ECOG PS of 1 (71%). Patients had a median age of 60 years and had received a median of 3 prior lines of therapy (range, 1-16). CD71 staining by immunohistochemistry showed comparable levels of high (at least 50% staining of 2+/3+), low, and unknown expression of 33%, 36%, and 31%, respectively.

The patient population was representative of a number of solid tumors, including NSCLC (n = 9; 20%), HNSCC (n = 8; 18%), colorectal cancer (n = 7; 16%), soft tissue sarcoma (n = 4; 9%), prostate cancer (n = 3; 7%), and other (n = 14; 31%).

Patients received a median of 3 (range, 1-12) doses of CX-2029. The median duration of exposure to CX-2029 was 9 weeks (range, 3-26).

There were no dose-limiting toxicities (DLTs) in the expansion cohorts leading up to the 3 mg/kg cohort. Though, 1 patient in the 0.1 mg/kg cohort discontinued treatment due to an infusion-related hypoxic event. The first assessment of patients who received 4 mg/kg of CX-2029 showed no DLTs. However, the 4 mg/kg and 5 mg/kg doses were declared intolerable.

The most common all-grade treatment-related adverse effects (TRAEs) occurring in at least 10% of patients consisted of infusion-related reactions [(IRR) 90%; primarily low grade and at first infusion], followed by anemia (80%), fatigue (31.6%), nausea (15.8%), neutropenia (31.6%), and leukopenia (28.2%).

In general, IRRs occurred within the first 2 to 3 hours after infusion and resolved within the same time frame after initiation of supportive care measures, such as antihistamines, meperidine, corticosteroids, and slowed infusion.

Grade 3 or higher TRAEs occurring in at least 2 patients included anemia (67.4%), neutropenia (31.6%), leukopenia (18.2%), and IRRs (6%). The majority of patients required at least 1 red blood cell (RBC) transfusion (74.8%), with a median number of 2 transfusions. The median time to first RBC transfusion was approximately 5 weeks, apart from the 5.0 mg/kg cohort, in which the median time to transfusion was approximately 2 weeks. No treatment-related deaths were reported with CX-2029.

“CX-2029 demonstrated dose-dependent hematopoietic toxicities consistent with MMAE payload,” said Johnson.

Regarding pharmacokinetics, preliminary plasma concentrations over time showed similar downward slopes across dose levels for masked in-tact CX-2029, providing no evidence of dose proportionality, said Johnson. The clearance, volume of distribution, and terminal half-life of CX-2029 was 0.55-2.7 L/day, 3.2-10.6 L, and 2.3-9.8 days, respectively.

“Importantly, the median ratio of free MMAE to total CX-2029 was very low, at less than 4%, indicating that most of the drug in the peripheral circulation was the potent microtubule inhibitor bound to the probody. This was consistent across all dose levels examined,” said Johnson.

The waterfall plot summarizing patients’ target lesion shrinkage in the 2.0 mg/kg to 5.0 mg/kg cohorts demonstrated striking activity in patients with HNSCC and squamous NSCLC who received at least 3.0 mg/kg of CX-2029, defined by at least 30% tumor shrinkage.

No responses were observed in patients who received less than 2.0 mg/kg. Though, the 0.5 mg/kg dose led to stable disease lasting 8 months in 1 patient with ocular melanoma.

To capture the activity of CX-2029 in squamous histologies, Johnson shared a case of a 75-year-old woman who was diagnosed with stage III NSCLC who received carboplatin/paclitaxel with radiation, followed by durvalumab (Imfinzi), gemcitabine, and docetaxel/ramucirumab (Cyramza). After 8 weeks of CX-2029, the women achieved a PR with near resolution of a right-sided perifissural lesion, as well as improvement in the left and right lower lobe nodular regions.

CX-2029 will be evaluated at a dose of 3.0 mg/kg in expansion cohorts in squamous NSCLC, HNSCC, esophageal and gastroesophageal junction cancers, and diffuse large B-cell lymphoma, concluded Johnson.

Johnson ML, El-Khoueiry AB, Hafez N, et al. CX-2029, a PROBODY drug conjugate targeting CD71 (transferrin receptor): results from a first-in-human study (PROCLAIM-CX-2029) in patients (pts) with advanced cancer. J Clin Oncol. 2020;38(suppl 15; abstr 3502). doi:10.1200/JCO.2020.38.15_suppl.3502

<<< 2020 ASCO Virtual Scientific Program