Introduction
As a topoisomerase I (Top1) inhibitor, the alkaloid SN38 (7-ethyl-10-hydroxy camptothecin) is one of the most potent cytotoxic camptothecins (CPTs) against cancer cells [
1,
2]. Although SN38 has great potential to treat many malignancies, such as colorectal, lung, gastric, and ovarian cancers, it cannot be used directly in clinical applications due to its poor water solubility and spontaneous hydrolytic instability of the lactone form (active form) to the carboxylate form (inactive form) [
3,
4]. Various prodrug approaches have been developed to solve the poor solubility problem, leading to the successful development of 3 FDA-approved drugs, irinotecan (CPT-11, 10ʹ-OH group on ring A is conjugated to a water-soluble moiety) [
5,
6], ONIVYDE (nanoliposome irinotecan) [
7], and Trodelvy (antibody drug conjugate, 20ʹ-OH group on ring E is conjugated to a sacituzumap via an acid sensitive linker, targeting the Trop-2 receptor in cancer cells) [
8,
9]. The CPT derivative irinotecan has been widely used since 1996 in advanced colorectal cancer (CRC) as a standard treatment agent in both monotherapy and combination therapy. However, only 2–8% of irinotecan [
10‐
13] can be transformed in vivo into the active metabolite, SN38, and 55% of the drug is excreted as intact irinotecan in humans [
14]. SN38 is 100–1000 times more cytotoxic than irinotecan [
13,
15]. Therefore, irinotecan itself without SN38 transformation is inactive and has practically no therapeutic value. Irinotecan conversion into active SN38 in vivo is achieved by carboxylesterases in the liver [
12,
16‐
20]. However, human liver carboxylesterase activity can vary widely among individual patients [
21,
22], which can lead to patient-specific irinotecan PK [
22] and antitumor efficacy. These intrinsic limitations of irinotecan significantly reduce its clinical potential [
6].
To overcome these problems of using SN38 as an anticancer drug, numerous drug delivery systems, such as prodrugs, polymeric micelles, and liposome-based formulations, have been studied extensively [
23]. These approaches can alter the properties of SN38, such as water solubility. The formulated SN38 has shown good efficacy against various tumors in preclinical research but showed disappointing results in the human clinical setting. SN38 liposome particles [
24,
25], PEG-SN38 [
26], and SN38 polymer micelle [
27] did not present better antitumor efficacy relative to irinotecan in human phase II trials. Problems associated with these drug delivery systems include low drug loading, poor tumor penetration, non-targeting effects, and unfavorable drug release. Therefore, new approaches are urgently needed to formulate SN38 for higher anti-cancer efficacy and lower toxicity.
It is well-documented that human serum albumin (HSA) is a desired drug delivery carrier [
28‐
30] due to its unique properties, such as being endocytosed and transcytosed into and across the cell via receptors [
29], long half-life of 19 days [
31‐
34], able to accumulate at the tumor tissue due to the enhanced permeability & retention (EPR) effect; and being preferentially taken up and metabolized by cancer cells to serve as nutrients [
35‐
39]. We previously developed the single protein encapsulation (SPE) technology to carry a predefined number of DOX (doxorubicin) molecules to form uniform HSA-DOX complexes (SPEDOXs) by an unmodified monomeric HSA molecule [
34], thereby avoiding the issues associated with synthetic polymers, conjugated HSA, and HSA nanoparticles (NPs). In vivo studies with mice demonstrated better PK, lower toxicity, and superior tumor inhibitory activity of SPEDOXs compared with unformulated DOX [
34]. Furthermore, our recent study demonstrated robust SPEDOX-6 uptake and efficacy in killing human cancer cells, while displaying low cytotoxicity to hiPSC-CMs (human induced pluripotent stem cell-derived cardiomyocytes) and hiPSC-CSs (multi-lineage cardiac spheroids) [
40], indicating that the SPE technology may provide an excellent platform for cancer drug formation. The FDA has granted “Orphan Drug Designation” to SPEDOX-6 for treatment of soft tissue sarcoma (STS) patients. Phase Ib/IIa human clinical trials of SPEDOX-6 are being planned (IND# 152154).
In this study, by adopting the SPE technology, we have successfully encapsulated SN38 to create two SPESN38 complexes, SPESN38-5 (5 SN38 molecules per HSA) and SPESN38-8 (8 SN38 molecules per HSA as the maximum capacity). Preclinical evaluations using CRC and STS mouse models show that SPESN38 complexes have better PK values than those of irinotecan, resulting in 1.8-fold higher SN38 AUC0-∞. SPESN38 also has a higher antitumor efficacy than that of irinotecan without increased toxicity. These results demonstrate SPESN38 complexes as novel effective anticancer agents with great potential for clinical applications, thereby warranting further studies to develop them into cancer therapeutics.
Material and methods
Material and instruments
HSA (25% solution) and SN38 were purchased from Octapharma USA and GLPbio Technology, respectively. Methanol, ethanol, other chemicals and suppliers were purchased from VWR. UV spectrum measurement and quantitation were conducted on a UV-1600 PC spectrometer (VWR). Both SPESN38-5 and SPESN38-5 were prepared following similar protocols for making SPEDOX-6 [
34].
In vivo studies
All in vivo studies were performed at Roswell Park Comprehensive Cancer Center Animal Facility following the animal protocol approved by the Institutional Animal Care and Use Committee (IACUC). Male and female CD-1(ICR) mice (haired) (5 to 7 weeks old) were purchased from Charles River Lab. Severe combined immunodeficiency (SCID) mice (CB17SC, strain C.B-Igh-1b/IcrTac-Prkdcscid, 5 to 7 weeks old were from Roswell internal breeding.
MTD study
For SPESN38-5, the lyophilized yellowish powder was dissolved in DI water to form a clear SPESN38-5 solution with light yellowish color. In PO route, the SPESN38-5 solution was fed to mice (2 female mice/group) at doses of 300, 250 and 200 mg/kg on Day 1. For IV route, the SPESN38-5 solution was intravenously injected to mice at doses of 80 (2 female mice/group), 55 (2 female mice/group), 45 (2 female and 2 male mice/group) and 35 mg/kg (2 female and 2 male mice/group). The percent body weight change of all mice was recorded vs days. For SPESN38-8, the same procedures were used for preparation and the resulting SPESN38-8 solution was intravenously injected to mice at doses of 45 and 35 mg/kg (2 female and 2 male mice/group).
PK study
After PO administration of SPESN38-5 at the dose of 250 mg/kg to six groups (3 female CD-1 mice/group), blood samples were collected into 1.5 mL Li-Heparin LH/1.3 tubes after anesthetizing mouse with CO2 are at the timepoints of 1, 2, 4, 8, 12, and 24 h (triplicate blood samples at each time point). Serum from each blood sample was obtained by 2,500 rpm centrifugation for 3 min, and the serum on the top layer was collected using pipet and transferred into 1.5 mL Eppendorf tubes and then frozen immediately in liquid nitrogen until PK analysis. After IV administration of SPESN38-5 (tail vein injection) was performed at the dose of 55 mg/kg to six groups (3 female CD-1 mice/group), the same procedure was used for sample preparation.
Mouse plasma samples were analyzed for SN-38 and SN-38G by LC–MS/MS using a previously described method [
41] over the calibration range of 0.200 to 200 ng/mL for each analyte. Briefly, an aliquot of plasma (100 uL) was mixed with acidified methanol containing the internal standards [irinotecan-d
10 (Toronto Research Chemicals, Toronto, Canada) and camptothecin (Sigma-Aldrich, St. Louis, Missouri), respectively] for a protein precipitation extraction, followed by centrifugation and injection of the supernatant for analysis. Chromatographic separation was achieved using a Waters CORTECS C18+ LC column (100 mm × 2.1 mm, 2.7 um) maintained at 50 ℃ and sample elution carried out at flow rate of 300 µL/min with a biphasic gradient (water with 0.1% acetic acid and acetonitrile with 0.1% acetic acid). SN-38 and SN-38G were detected by multiple reaction monitoring (MRM) using an AB SCIEX 5500 mass spectrometer with an electrospray ionization source in positive ion mode controlled by AB SCIEX Analyst
® software, version 1.6.2. All sample results were obtained within one analytical run. Samples above the calibration range were diluted to be below the point of saturation of the detector.
Non-compartmental analysis (NCA) was performed utilizing mouse plasma concentrations of SN38 and SN38-G that were obtained by LC–MS/MS. Plasma samples that were included in the NCA were collected at t = 0, 1, 2, 4, 8, 12, and 24 h post-dose for PO route and at t = 0, 0.083, 1, 2, 4, 8, and 24 h post-dose for IV route. The PK parameters were calculated using Phoenix WinNonlin software: maximum plasma concentration (Cmax), Area Under the plasma Concentration–time curve (AUC), elimination half-life (t1/2), apparent clearance (CL/F), and clearance (CL). AUC values were calculated using the linear-up log-down method.
HCT-116 model efficacy study
HCT-1116 cell line (CCL-247) was purchased from ATCC. After growing in Eagle's Minimum Essential, HCT-116 cells were harvested by trypsinization and washed twice with PBS. HCT-116 cells (2 × 106 per injection) were suspended in 200 µL of a 1:1 solution of ice-cold PBS and Matrigel (Corning Incorporated, Corning, NY) solution. HCT-116 cancer xenograft tumors were first generated by injecting 2 × 106 cancer cells into the flank area of SCID mice. After the tumors grew to 800–1200 mm3, they were isolated, and approximately 50 mg of non-necrotic tumor masses were subcutaneously implanted into the flank area of individual female SCID mice. When the implanted xenograft tumors grew to 250 to 350 mm3 on Day 7 after tumor transplantation, mice were randomly divided into 4 groups for intravenous injection: (1) vehicle (saline, 8 females), (2) SPESN38-5 (PO route at 200 mg/kg, 8 females), (3) irinotecan (50 mg/kg, 8 females), (4) SPESN38-5 (IV route at 55 mg/kg, 8 females). The intended dose for irinotecan (pharmaceutical grade for human application) at 100 mg/kg was attempted on 2 SCID mice bearing HCT-116. Surprisingly, both died immediately. Other doses at 75 and 50 mg/kg were tried on healthy CD-1 mice. Both mice from 75 mg/kg IV died instantly, but 2 mice from 50 mg/kg IV were safe, which is consistent with the literature report. Therefore, irinotecan treatment group had only 6 female mice for this study. Mice in group 1 and group 2 on Day 10 were sacrificed due to the large tumor size with diameter ≥ 20 mm. One mouse from group 4 showed health issues early on and was euphanized. Tumor volume (TV) and BW were measured three times per week or daily depending on the condition of the mouse. TV was calculated using the formula: v = 0.5 (L x W2). Progression at the endpoint was a tumor size with diameter ≥ 20 mm or a moribund condition.
SK-LMS-1 model efficacy study
SK-LMS-1 cell line (HTB-88) was purchased from ATCC. After growing in Eagle's Minimum Essential, SK-LMS-1 cells were harvested by trypsinization and washed twice with PBS. SK-LMS-1 cells (1 × 106 per injection) were suspended in 200 µL of a 1:1 solution of ice-cold PBS and Matrigel (Corning Incorporated, Corning, NY) solution. SK-LMS-1 cancer xenograft tumors were first generated by injecting 1 × 106 cancer cells into the flank area of SCID mice. After the tumors grew to 800–1200 mm3, they were isolated, and approximately 50 mg of non-necrotic tumor masses were subcutaneously implanted into the flank area of individual mice. Equal number (12) of female and male mice were used in this experiment. When the implanted xenograft tumors grew to 250 to 350 mm3 on Day 7 after tumor transplantation, mice were randomly divided into 6 groups for intravenous injection: (1) vehicle (saline, 4 females), (2) DOX (5 mg/kg, 4 females), (3) SPESN38-8 (IV at 33 mg/kg, 4 females), (4) vehicle (saline, 4 males), (5) DOX (5 mg/kg, 4 males), (6) SPESN38-8 (IV at 33 mg/kg, 4 males). The intended schedule for drug or vehicle treatment was weekly for 3 doses. However, mice in groups 2 and 5 with DOX at 5 mg/kg after 2 doses lost > 20% BW, indicating severe toxicity. Mice in group 1, 2, 4 and 5 on Day 9 were sacrificed due to the large tumor size with diameter ≥ 20 mm and severe BW loss (> 20%). One male mouse from group 6 had some health issues early on and was sacrificed on Day 14. Tumor volume (TV) and BW were measured three times per week or daily depending on the condition of the mouse. TV was calculated using the formula: v = 0.5 (L x W2). Progression at the endpoint was a tumor size with diameter ≥ 20 mm or a moribund condition.
Tumor tissue preparations and staining study
Tumor tissues from Sk-LMS-1 mouse study were fixed in 10% neutral buffered formalin for 24, and then transferred into 70% ethanol for up to 4 days. The fixed tissues were embedded in paraffin and sectioned at 5 microns at any time when tissues were moved into 70% ethanol. All the specimens were formalin-fixed and paraffin-embedded.
H & E staining
Dako CoverStainer was utilized for H & E staining analyses on the paraffin-embedded SK-LMS-1 tumor tissues with a DAKO H&E kit.
Immunohistochemistry (IHC) analysis on Ki67 and cleaved caspase-3
Deparaffinized tissue sections were rehydrated and incubated in 1 × pH6 citrate buffer (Invitrogen Cat #00–5000) for 20 min using a DAKO PT Link. With an Autostainer, the following steps and reagents were used for IHC analysis:
(1) Incubation in 3% H2O2 for 15 min; (2) Incubation with 10% normal goat serum 10 min (Thermo Fisher #50062Z) 10 min; (3) Incubation with Avidin/Biotin block (Vector Labs Cat#SP-2001); (4) Incubation with primary KI67 antibody (Abcam #ab15580 or Cleaved Caspase-3 (Asp175) antibody (Cell Signaling Cat #9661) diluted in 1% BSA for 30 min; (5) Incubation with secondary Goat anti Rabbit (Vector labs #BA-1000) for 15 min; (6) Incubation with ABC reagent (Vector Labs Cat #PK 6100) for 30 min; (7) Incubation with DAB substrate (Dako Cat #K3467) for 5 min; (8) Counterstained with DAKO Hematoxylin for 20 s; (9) Coversliped slides.
Statistic analysis
Statistic analyses of tumor volume and tumor weight change are described in Additional file.
Discussion
CPTs belong to the class of TOP1 inhibitors [
2,
47‐
53]. Among CPTs [
54], SN38 stands out as one of the most potent cancer therapeutics. It is well known that the lactone form (L form) of CPTs undergoes pH-dependent and reversible ring opening through hydrolysis [
3] to produce the inactive carboxylate form (C form) [
4]. Different CPTs have similar t
1/2 for the ring opening reaction and reach the equilibrium state with ~ 20/80% L/C forms in PBS buffer at 37 ℃. However, the presence of HSA significantly changes the kinetics and thermodynamics of the ring opening reaction, due to differential HSA binding to L/C forms of different CPTs. While HSA binds the L/C forms of CPT with a 157-fold higher affinity for the C form, SN38 binding to HSA is reversed, with 4.3-fold stronger binding for the L form. As a result, HSA increases the ring opening of CPT with decreasing t
1/2 and < 0.5% L form at the equilibrium. On the contrary, SN38’s t
1/2 and the L form at the equilibrium both increase substantially in the presence of HSA, from 20 to 35 min and 13 to 38%, respectively [
4]. For the past decades, the unique properties of SN38 attracted many research attempts, but its poor aqueous solubility has hindered its development as an unmodified drug. Consequently, the prodrug approach led to an irinotecan approved by the FDA, generating active metabolite SN38 by a biotransformation [
12,
16‐
20]. The complex PK and low fraction conversion (2–8%) of irinotecan in human setting [
10‐
13] have resulted in inconsistent PK behaviors and efficacy among different patients [
21,
22].
The current study represents the first example of developing unmodified SN38 in soluble and stable forms for in vivo antitumor evaluation. Based on the similar preparation procedure except for different SN38/HSA ratios, SPESN38-5 and SPESN38-8 are expected to display similar pharmacological properties such as MTD, PK, and antitumor efficacy. Toxicity study in mice indicated SPESN38-5 and SPESN38-8 with respective 55 and 45 mg/kg MTD, confirming their similar but not identical properties. We thus proceeded to conduct PK and different tumor model studies with either SPESN38-5 or SPESN38-8.
SN38 AUC
0-∞ of SPESN38-5 at a single IV dose of 55 mg/kg (Table
1) is similar to that of irinotecan at a dose of 200 mg/kg (Additional file
1: Table S3) [
55]. Based on their molecular weights, the SN38 AUC
0-∞ value for SPESN38-5 is estimated to be 1.8 times higher than that for equivalent irinotecan. Since the carboxylesterase activity is much lower in humans than in mice [
56], much smaller % irinotecan biotransformation to SN38 in humans relative to mice is expected, likely resulting in much lower SN38 AUC
0-∞ value of irinotecan compared to SPESN38-5 in human plasma. Unlike the prodrug irinotecan, SPESN38-5 does not need biotransformation to SN38 by carboxylesterases, minimizing insistency among different patients. As demonstrated for SPEDOX-6 [
34], delivery of SPESN38-5 to cancer cells via endocytosis, followed by SN38 dissociation and/or HSA hydrolysis by proteases, releases unmodified SN38 into the cytosol of cancer cells. Furthermore, while HSA has a long half-life of 3 weeks in human serum, due to its effective rescue and recycling through strong HSA-hFcRn (human FcRn) binding, MSA-mFcRn (mouse serum albumin-mouse FcRn) binding is weak [
57], leading to a short HSA half-life in mice. As a result, SPESN38-5 is expected to have even better SN38 AUC
0-∞ value relative to irinotecan in humans than in mice.
As expected from the PK values, SPESN38-5 in the PO route did not provide a viable option for treating cancer because of low oral bioavailability. However, both SPESN38-5 and SPESN38-8 in the IV route demonstrated excellent antitumor efficacy in 2 mouse models. In the HCT-116 CRC model, SPESN38-5 at 55 mg/kg showed superior antitumor activity compared to irinotecan (Additional file
1: Table S1). Separately, in the SK-LMS-1 STS model, excellent antitumor activity was achieved by SPESN38-8 at 33 mg/kg, resulting in 6 of 7 mice tumor-free. In stark contrast, conventional DOX at 5 mg/kg (MTD) was ineffective. Due to the fact that each HSA molecule in SPESN38-8 carries 60% more of SN38 than SPESN38-5 without lowering anticancer efficacy and higher toxicity, SPESN38-8 is the preferred drug candidate for further investigation.
It is known that tumor cells aggressively take up HSA as nutrients to support fast growing tumor cells [
35‐
37,
39]. As such, SPESN38 complexes may achieve targeted SN38 delivery to cancer cells due to: (1) HSA (in SPESN38) is taken up by tumor cells, and dissociation and/or enzymatic degradation of HSA release SN38; (2) the acidic tumor microenvironment destabilizes SPESN38 as demonstrated for SPEDOX-6 [
34], resulting in HSA’s conformation change and liberation of SN38; (3) Secreted Protein Acidic and Rich in Cysteine (SPARC) with binding affinity to HSA, may play an important role in promoting tumor uptake of HSA and ABRAXANE [
58,
59], although recent clinical trials did not reveal a significant correlation between SPARC expression and the treatment outcome of ABRAXANE [
60].
The long half-life of HSA can be attributed to FcRn-mediated rescue and recycling mechanism [
31‐
33,
61]. If cancers express less FcRn, they are expected to have less HSA (SPESN38) recycling capacity, leading to increased endocytosis, SN38 dissociation, and lysosomal degradation of HSA (SPESN38). Consequently, cancer cells would get higher concentrations of SN38 relative to normal cells in cancer patients. Published reports [
35,
37‐
39,
62] and a database [
63] convincingly show that many types of cancer, including breast cancer, lung cancer, cervical cancer, ovarian cancer, pancreatic cancer [
64], CRC [
39], have significantly lower levels of FcRn, which promotes tumor growth by increasing HSA endocytosis and consumption. Therefore, FcRn expression levels might offer a promising cancer-targeting strategy for development of HSA-encapsulated drugs for attacking various cancers [
65].
Conventional drug-containing NPs are usually assembled from lipids, synthetic and natural polymers, and inorganic materials. These NPs can be made in different size ranges and are heterogeneous in size distribution (polydisperse). Furthermore, drug molecules are often linked to the carrier through covalent conjugation. In contrast, the SPE technology has the following unique properties: (1) The formulation process involves no chemical steps. HSA encapsulation of drug molecules are achieved through multiple non-covalent interactions between HSA and drug molecules under a specific set of conditions; (2) The binary system contains a single native HSA molecule that encapsulates a predefined number of a specific drug molecule in its unmodified form; (3) The resulting SPEDRUG complex is uniform in size (monodisperse) and has the same size of a native HSA molecule; (4) The number of drug molecules per HSA molecule may be adjusted according to specific application; (5) The HSA-drug binding strength is tunable by adjusting formulation conditions to effect PK and antitumor efficacy. The successful development of SPEDOX-6 [
34], SPESN38-5, and SPESN38-8 has demonstrated the utility and versatility of the SPE platform. We are actively developing other SPEDRUG complexes, and different SPEDRUG complexes are expected in the future.
Conclusion
Using the newly developed SPE technology, we prepared SPESN38-5 and SPESN38-8, demonstrating the first examples of unmodified SN38 in clear, stable, and injectable solution. Compared with irinotecan and DOX in animal models, SPESN38-5 and SPESN38-8 showed favorable pharmacokinetic values, superior antitumor efficacy against CRC and STS, and lower systemic toxicity. The successful development of SPEDOX-6, SPESN38-5, and SPESN38-8 has validated the SPE platform in drug formulation. These SPEDRUG complexes represent a new uniform macromolecular nanodrug that may be used to target low FcRn expressing cancer cells, further improving their antitumor efficacy while reducing side effect toxicities. These promising preclinical results have prompted these SPEDRUG complexes to be aggressively pursued for their clinical applications.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.