Induced resistance to ofatumumab-mediated cell clearance mechanisms, including complement-dependent cytotoxicity, in chronic lymphocytic leukemia

Abstract

Ofatumumab (OFA), a human CD20-targeting mAb, kills B lymphocytes using the innate immune system including complement-dependent cytotoxicity (CDC). The efficacy of OFA in patients with chronic lymphocytic leukemia (CLL) is limited by drug resistance, which is not well characterized. To better understand mechanisms of resistance, we prospectively studied CLL cells isolated from blood samples collected before and after in vivo exposure to the initial dose of OFA therapy in 25 patients undergoing their first treatment for progressive CLL. As previously reported, OFA therapy rapidly decreased the absolute lymphocyte count, CD20 expression by CLL cells, and serum complement levels. We now show that after administration of the first dose of OFA, there was a modest rebound in the absolute lymphocyte count and serum complement levels, but substantial ongoing loss of CD20 expression by CLL cells. These post-OFA treatment CLL cells were highly resistant to OFA-mediated CDC but retained sensitivity to alemtuzumab-mediated CDC in vitro. Posttherapy serum OFA levels correlated inversely with both the amount of pretreatment circulating cell-bound CD20 and with the decrease in this value following treatment. In vitro OFA-mediated CDC did not predict clinical responses, and the patients with first-dose reactions to OFA did not have markers of increased complement activation in vivo. We propose that optimal efficacy of CD20- targeted therapy for CLL requires determining an mAb dose size and frequency that optimizes CLL killing without exceeding the capacity of the cytotoxic mechanisms and thus minimizes loss of CD20 expression in the surviving CLL cells.

Trial registration: ClinicalTrials.gov NCT01024010.

Figure 1. Changes in Absolute Lymphocyte Counts (ALC), Serum Complement Levels (CH50), and Serum Ofatumumab Concentration (OFA) after Administration of the First Dose of OFA

A: The median ALC and CH50 values 8 hours (8hr) and 24 hours (24hr) after the initiation of the infusion of the first dose of OFA were calculated as a percentage of the baseline pre-treatment levels for each patient and the error bars represent the standard deviation. The median ALC decreased 45% (range 28–91) at 8hr (p<0.0001) and then increased 27% between 8hr and 24hr (p=0.0002) despite interim administration of pentostatin and cyclophosphamide. The median CH50 decreased to 14% of baseline at 8hr (p<0.0001) and then rebounded slightly to 20% of baselineat 24hr (p=0.05). Median OFA levels were 33.7 μg/ml at 8hr and 28.8 μg/ml at 24hr. B–D: These scatterplots show the ALC (B), CH50 (C) and OFA (D) values for each patient. Solid circles represent patients receiving a 300 mg first dose of OFA on day 1 and open circles represent those patients who received less than 300 mg of OFA. The bars represent the median of the values.

Figure 2. The Serum Ofatumumab (OFA) Level at 8 Hours After Starting Therapy is Inversely Proportional to the Total Pre-treatment Cell Associated Circulating CD20 and the Decrease in this Value after the Initial Dose of 300 mg of OFA

A: The total cell-associated CD20 in the circulation (Total CD20 binding sites) was calculated as the product of the CD20 expression of the CLL cells (dMESF) and the absolute lymphocyte count. B: The change in the total cell-associated CD20 in the circulation (ΔCD20 binding sites) 8 hours after starting the administration of the first dose of OFA was calculated by subtracting the calculated number of binding sites in the 8hr specimen from the pre-treatment value. Serum OFA levels in the patients receiving the full 300 mg initial dose (n = 17) had a significant inverse correlation coefficient (r) with total CD20 binding sites (A) and ΔCD20 binding sites (B) using Spearman Correlation statistics.

Figure 3. Circulating CLL Cells Sampled after the First Dose of Ofatumumab (OFA) have Low Levels of CD20 Expression, Decreased OFA Binding Capacity, and Complement Fragment Deposition after in vitro Incubation with OFA and C5-deficient Serum

A: Median OFA binding to CLL cells (dMESF OFA, based on probing with mAb HB43 specific for the Fc region of OFA) sampled from patients before initiation of treatment was measured after incubation of cells with 10 μg/ml of OFA and was found to be 9918 (range 1786 – 39694) (max binding 0hr). Median OFA binding in cells from patients sampled 8 hours after the initiation of OFA treatment was 2077 (range 96 – 5769) (bound in vivo 8hr) and decreased to 258 (range 0 – 2542) at 24hr (bound in vivo 24hr). CLL cells sampled at 8hr and 24hr were then incubated in vitro with 10 μg/ml of OFA. Compared to the 0hr sample (max binding 0hr), the OFA binding to the CLL cells from the 8hr sample was significantly lower at 8hr (max binding 8hr, median dMESF 3113, range 276 – 7951, p<0.0001) and at 24hr (max binding 24hr, median dMESF 1124, range 0 – 6870, p<0.0001). There were no significant differences in median dMESF OFA in patients that received an initial OFA dose <300 mg (open circles, n=8) compared to those that received a dose of 300 mg of OFA (closed circles, n=17) B: Median C3 fragment deposition (dMESF C3, based on probing with FITCmAb 1H8 specific for C3 fragments) in CLL cells collected before the initiation of therapy and incubated in vitro with 10 μg/ml OFA and 10% C5-depleted serum was 40667 (range 2914–89828) (max binding 0hr). dMESF C3 in CLL cells collected at 8hr after starting OFA therapy (bound in vivo 8hr) was lower at 7150 (range 757–35809) and decreased further at 24hr (bound in vivo 24hr) to 2476 (range 0–24139). Compared to the 0hr sample, CLL cells collected at 8hr after initiation of OFA therapy and then incubated with 10 μg/ml of OFA and 10% C5-depleted serum in vitro (max binding 8hr) had a significantly lower median dMESFC3 of 8027 (range 1225–44856)(p<0.0001) which decreased further in CLL cells sampled at 24hr (max binding 24hr) to 5325 (range 927–25940)(p<0.0001). dMESF C3 was not significantly different in CLL cells from patients who received lower (<300 mg) (open circles) compared to standard 300 mg (solid circles) initial doses of OFA.

Figure 4. Levels of Ofatumumab(OFA) Binding to CLL Cells Equivalent to Those Measured in Circulating CLL Cells After Initiation of OFA Therapy Have Low in Vitro Complement Activating Activity

The OFA concentration resulting in binding to CLL cells equivalent to that measured in CLL cells sampled 8 hours after initial therapy with 300 mg of OFA was 0.025 μg/ml. Pre-treatment CLL cells from a randomly selected subset of 5 patients were incubated in vitro with OFA (0.025 μg/ml, 0.05 μg/ml and 0.1 μg/ml) and 10% normal human serum (NHS) as a source of complement and assayed for OFA binding, C3 fragment deposition, and complement dependent cytotoxicity (CDC). The results of experiments were expressed as a percentage the control experiments which measured the OFA binding, C3 fragment deposition, and CDC achieved with a saturating dose of OFA (10 μg/ml) and 10% NHS in the same specimens. Compared to control experiments, CLL cells treated with 0.025 μg/ml had significantly lower median OFA binding (25%, range 17–28), C3 fragment deposition (0%, range 0–3%), and CDC (0%, range 0–3%). Sequential doubling of the OFA concentration results in an appreciable increase in the median OFA binding to 38% (range 31 –45%) at 0.05 μg/ml and 51% (range 51 – 71%) at 0.1 μg/ml but this was associated with minimal C3 fragment deposition (median 0% at 0.05μg/ml, and 1% at 0.1 μg/ml) or CDC (median 0% at 0.05 μg/ml, and 3% at 0.1 μg/ml).

Figure 5. In Vitro Complement Dependent Cytotoxicity (CDC)

CLL cells obtained from patients before treatment (0hr), and then at 8hr and 24hr after the start of the administration of the first dose of ofatumumab (OFA) were tested for in vitro CDC using saturating doses (10 μg/ml) of rituximab (RTX), OFA, or alemtuzumab (ALM)and 10% normal human serum (NHS) as a source of complement. RTX induced low levels of CDC. Compared to OFA CDC in pre-treatment specimens (median 37% cytotoxicity, range 0 – 83) there was a significant decrease in OFA CDC at 8hr (median 0%, range 0 – 5)(p<0.0001) and 24hr (median 0%, range 0 – 32) (p<0.0001). In contrast, ALM CDC was not decreased by in vivo exposure of CLL cells to OFA.