Which factors matter the most? Revisiting and dissecting antibody therapeutic doses

Abstract

Factors such as antibody clearance and target affinity can influence antibodies' effective doses for specific indications. However, these factors vary considerably across antibody classes, precluding direct and quantitative comparisons. Here, we apply a dimensionless metric, the therapeutic exposure affinity ratio (TEAR), which normalizes the therapeutic doses by antibody bioavailability, systemic clearance and target-binding property to enable direct and quantitative comparisons of therapeutic doses. Using TEAR, we revisited and dissected the doses of up to 60 approved antibodies. We failed to detect a significant influence of target baselines, turnovers or anatomical locations on antibody therapeutic doses, challenging the traditional perceptions. We highlight the importance of antibodies' modes of action for therapeutic doses and dose selections; antibodies that work through neutralizing soluble targets show higher TEARs than those working through other mechanisms. Overall, our analysis provides insights into the factors that influence antibody doses, and the factors that are crucial for antibodies' pharmacological effects.

Figure 1.

Schematic of the TEAR. Factors such as the PK in the peripheral blood, antibody distribution, target turnover and baseline, antibody affinity, and MoAs are routinely considered in antibody development. TEAR takes into account the components with higher quantification certainty, including bioavailability (F), systemic clearance (CL), and binding affinity (K_D), and serves as a metric reflecting the pRO. The factors beyond pRO that are not considered in TEAR, such as the target distribution, target properties and MoAs, will lead to variance in the TEARs. The impact of these factors can be investigated on the basis of the TEARs.

Figure 2.

Most approved antibodies have therapeutic doses oversaturating the targets in plasma, as shown by the TEAR metric. Up to 60 licensed antibodies were included in the analysis. Most of the surveyed antibodies had TEARs > 2, indicating that such antibodies could saturate their targets in the peripheral blood (pRO τ 99%) at the therapeutic doses. As shown in the bar chart, two antibodies had TEARs > 5, 15 had TEARs > 4, 23 had TEARs = 3 and 4, and 17 had TEARs = 2 and 3, indicating that factors other than CL, F and K_D affect therapeutic dose selection. Each dot represents the TEAR of an antibody in mean ± standard deviation [70]. The antibodies are numbered in alphabetical order and are shown on the x-axis. F, bioavailability; CL, systemic clearance; τ, dosing interval; C_target, antibody concentration at the site of action.

Figure 3.

Effects of target anatomical locations, forms and turnovers on therapeutic doses. (a) Target anatomical location does not have a significant impact on antibody therapeutic dose, whereas target cellular location might be associated with therapeutic dose selection. Although there was no significant difference between the TEARs of the circulation and tissue groups (p = 0.99, unpaired Student’s t-test), the TEARs of the soluble and membranous groups were significantly different (p = 0.0022, unpaired Student’s t-test). Each dot represents the mean TEAR of an antibody. The bars represent the median and quartiles. (b) The antibody TEARs are different between certain disease-target scenarios. As shown above, the antibodies were categorized into four disease-target scenarios. The major diseases in each scenario are autoimmune diseases with targets in the circulation (circulation-soluble), hematologic malignancies (circulation-membranous), tissues (tissue-soluble) and solid tumors (tissue-membranous). A significant difference was observed only between the hematologic-malignancies and tissue-target groups (p = 0.0065, unpaired Student’s t-test). No difference in TEAR was observed between the other groups. (c) TEARs are different between tumor anatomical sites. The anticancer antibodies targeting the blood and lymph system had significantly lower TEARs than those targeting the GI system (p = 0.006, unpaired Student’s t-test) but have no substantial difference from the ones targeting skin (p = 0.88, unpaired Student’s t-test). (d) The pathologic locations of autoimmune diseases have no significant impact on antibody TEARs. The TEARs were not significantly different between the target anatomical locations in autoimmune diseases (p = 0.39, ordinary one-way ANOVA). In (b)–(d), each dot represents the TEAR of an antibody. All the data are represented as mean ± SD. (e) Target turnover is not relevant to antibody therapeutic dose. Target turnover was found not to be a significant factor affecting the therapeutic doses in the circulation, tissue, membranous and soluble groups (p = 0.22, 0.98, 0.48 and 0.47, respectively; Pearson’s correlation). The horizontal bars represent the SD in the target turnover rates, the vertical bars represent the SD in the TEARs, and the shadows represent the 90% prediction intervals.

Figure 4.

The MoA is a pivotal factor in discerning therapeutic doses. (a) Antibodies elicit therapeutic efficacy via three distinct mechanisms: soluble target neutralizing, membranous signaling suppression and immunomodulatory function. (b) The TEARs are significantly different between antibodies with varying mechanisms of action. The antibodies in the immunomodulatory-function group had the lowest TEARs (mean TEAR = 2.7), whereas the antibodies that act through soluble-target-neutralization seem to have the highest therapeutic doses (mean TEAR = 3.7). The antibodies in the signaling-suppression group had significantly lower TEARs compared with the ones in the soluble-target-neutralization group (3.2 versus 3.7, p = 0.04, unpaired Student’s t-test). The TEARs in the signaling-suppression and soluble-target-neutralization groups were significantly higher than those in the immunomodulatory-function group (p = 0.02, P = 0.0001, unpaired Student’s t-test). Each dot represents the TEAR of an antibody. All the data are represented as mean ± SD. (c) The target turnover rates were correlated with the TEARs of the antibodies with an immunomodulatory function (p = 0.03, respectively; Pearson’s correlation), but not with the TEARs of the antibodies that neutralize soluble targets or suppress membrane signaling (p = 0.47 and 0.46, respectively; Pearson’s correlation). The horizontal bars represent the SD in the target turnover rates, the vertical bars represent the SD in the TEARs, and the gray shadows represent the 90% prediction intervals.

Figure 5.

MoA has an impact on antibodies’ FIHD escalations. (a) The dose-escalating ranges in the FIH trials are denoted by gray bars, whose lower and upper edges represent the FIHDs and MADs in the FIH trials, respectively. The TEARs of the P2Ds (TEAR^P2D) are given in mean ± SD and are represented by red bars and shadows. The solid black circles pertain to the TEARs of the therapeutic doses. The blue and yellow triangle symbols denote the antibodies that used MABEL approaches or treatment efficacy as the P2D selection rationale. Forty antibodies were included in the analysis and were grouped by their MoAs. The names of the tested antibodies are indicated on the x-axis. (b) The TEAR^FIH values were significantly different between antibodies with varying MoAs. The TEAR^FIH values in the signaling-suppression and soluble-target-neutralization groups were significantly higher than those in the immunomodulatory-function group (p = 0.01, P < 0.0001, unpaired Student’s t-test). Each dot represents the mean TEAR^FIH value of an antibody. All the data are represented as mean ± SD.