Structure‒tissue exposure/selectivity relationship (STR) correlates with clinical efficacy/safety

Abstract

Drug optimization, which improves drug potency/specificity by structure‒activity relationship (SAR) and drug-like properties, is rigorously performed to select drug candidates for clinical trials. However, the current drug optimization may overlook the structure‒tissue exposure/selectivity-relationship (STR) in disease-targeted tissues vs. normal tissues, which may mislead the drug candidate selection and impact the balance of clinical efficacy/toxicity. In this study, we investigated the STR in correlation with observed clinical efficacy/toxicity using seven selective estrogen receptor modulators (SERMs) that have similar structures, same molecular target, and similar/different pharmacokinetics. The results showed that drug's plasma exposure was not correlated with drug's exposures in the target tissues (tumor, fat pad, bone, uterus), while tissue exposure/selectivity of SERMs was correlated with clinical efficacy/safety. Slight structure modifications of four SERMs did not change drug's plasma exposure but altered drug's tissue exposure/selectivity. Seven SERMs with high protein binding showed higher accumulation in tumors compared to surrounding normal tissues, which is likely due to tumor EPR effect of protein-bound drugs. These suggest that STR alters drug's tissue exposure/selectivity in disease-targeted tissues vs. normal tissues impacting clinical efficacy/toxicity. Drug optimization needs to balance the SAR and STR in selecting drug candidate for clinical trial to improve success of clinical drug development.

Keywords: Clinical efficacy/toxicity; Drug development; Drug optimization; Structure-tissue exposure/selectivity relationship (STR); Structure‒activity-relationship (SAR).

Graphical abstract

Figure 1

Drug exposure in the plasma was not correlated with drug exposure in the disease targeted tissue among seven SERMs with different or similar chemical structures. (A) Structures of seven SERMs. (B‒D) Plasma AUC vs. tissue AUC of tumor (B), fatpad (C) and bone (D). (E) Concentration‒time curve of nafoxidine vs. raloxifene, in which nafoxidine had higher drug concentration in the plasma and tumor. (F) Concentration‒time curve of tamoxifen vs. raloxifene, in which two drugs had similar plasma concentration, but tamoxifen had much higher tumor concentration. (G) Concentration‒time curve of toremifene vs raloxifene, in which toremifene had lower plasma concentration, but higher drug concentration in the tumor with raloxifene. MMTV-PyMT transgenic mice with spontaneous breast cancer were orally administered with tamoxifen, toremifene, afimoxifene, droloxifene, lasofoxifene, nafoxidine and raloxifene (5.0 mg/kg). Three mice were sacrificed at each time point to collect plasma and other tissues. The drug concentration in all samples and their calculated AUC by non-compartment model were compared among different compounds (Supporting Information Figs. S1 and S2). Data in E, F and G were represented as mean ± SEM (n = 3 per time point, two-tailed t-test). Asterisks indicate the following P-value: ∗P < 0.05, ∗∗P < 0.01.

Figure 2

Drug exposure in the tissue, not in the plasma, was correlated with drug clinical efficacy/safety. (A) Concentration‒time curve of tamoxifen vs. raloxifene, in which two drugs had similar plasma concentration, but tamoxifen had much higher concentration in tumor, fatpad, bone, uterus, skin, stomach, lung and brain. (B) Clinical efficacy of tamoxifen (20 mg/d) vs. raloxifene (60 mg/d). (C) Comparison of clinical adverse effects associated with different organs. Data of (C) and (D) were summarized from FDA labels of both drugs^,. Data were represented as mean ± SEM (n = 3 per time point, two-tailed t-test). Asterisks indicate the following P-value: ∗P < 0.05, ∗∗P < 0.01.

Figure 3

Drug tissue selectivity is a critical parameter that tips the balance of efficacy/toxicity. (A and B) Comparison of drug exposure (A) and selectivity (B) in tissues like tumor, fatpad and bone between tamoxifen and raloxifene after oral administration (5 mg/kg). (C and D) Comparison of drug exposure (C) and selectivity (D) in tissues like lung, tumor and fatpad between tamoxifen and toremifene after oral administration (5 mg/kg). (E and F) Comparison of drug exposure (E) and selectivity (F) in tissues like fatpad, tumor and skin between tamoxifen and nafoxidine after oral administration (5 mg/kg) in MMTV-PyMT transgenic mice with spontaneous breast cancer. Data were represented as mean ± SEM (n = 3, two-tailed t-test). Asterisks indicate the following P-value: ∗P < 0.05, ∗∗P < 0.01.

Figure 4

Slight structure modification altered drug exposure and selectivity in tissues despite similar exposure in the plasma. (A) The chemical structure of afimoxifene, droloxifene, tamoxifen and toremifene. (B) Concentration‒time curve after i.v. administration of afimoxifene, droloxifene, tamoxifen and toremifene (i.v. 2.5 mg/kg) on MMTV-PyMT transgenic mice with spontaneous breast cancer (n = 3 at each time point). Data were represented as mean ± SEM. (C) AUC ratio of tamoxifen vs. toremifene, tamoxifen vs. afimoxifene and afimoxifene vs. droloxifene. Average AUC in each tissue was used for the ratio calculation. (D) Drug tissue selectivity calculated by AUC_tissue/AUC _total using data collected in (B).

Figure 5

Slight structure modification altered drug exposure and selectivity in both plasma and tissues. (A) The chemical structure of lasofoxifene and nafoxidine. (B) AUC ratio of nafoxidine vs lasofoxifene. Average AUC in each tissue was used for the ratio calculation. (C) Concentration‒time curve after i.v. administration of lasofoxifene and nafoxidine (IV 2.5 mg/kg) on MMTV-PyMT transgenic mice with spontaneous breast cancer (n = 3 at each time point). Data were represented as mean ± SEM. (D) Drug tissue selectivity calculated by AUC_tissue/AUC_total using data collected in (B).

Figure 6

An enhanced accumulation of seven SERMs in tumors compared to normal fat pad tissue. (A‒G) Concentration‒time curve of tumor and fatpad (normal tissue surrounding tumors) after i.v. administration of seven SERMs (2.5 mg/kg) on MMTV-PyMT transgenic mice with spontaneous breast cancer (n = 3 at each time point); (H) An illustration of mechanism of enhanced tumor accumulation. Tumor vascular are abnormal and leaky compared to vascular in normal tissue, which allows more protein/drug complex enters to tumors resulting an enhanced drug accumulation. Data were represented as mean ± SEM (n = 3 per time point, two-tailed t-test). Asterisks indicate the following P-value: ∗P < 0.05, ∗∗P < 0.01.

Figure 7

Molecular structure descriptors influence drug exposure in tissues. (A) PCA analysis to decompose molecule structure descriptors to 3 components. (B) Percentage of molecule property variance explained by components. (C) Ordinary least squares analysis of drug's partition coefficient in tissue (represented by K_p = AUC_tissue/AUC_plasma) against molecule structure (represented by components) for both i.v. and oral data. (D) Univariate feature analysis of collected descriptors (clustered) in target tissues including fatpad, tumor and bone. (E) Representative descriptors selected from (D, oral) to explain the difference of properties correlated to different tissues.