Evaluation of Nonpeptidic Ligand Conjugates for the Treatment of Hypoxic and Carbonic Anhydrase IX–Expressing Cancers

Because of the defective formation of vasculature, the deposition of a dense extracellular matrix, and the insufficient assembly of lymphatic vessels, oxygen delivery to tumor tissues is often inadequate (1–3). While the extent of hypoxia can vary both temporally and spatially within a tumor (1–3), its clinical impact is almost always detrimental, with increases in chemoresistance, angiogenesis, radioresistance, metastasis, resistance to cell death, genomic instability, and the generation of cancer stem cells being the more common adverse consequences (4–7). Given the central contributions of these hypoxia-induced changes to tumor progression and survival, tumor hypoxia might well be considered one of the most important targets yet to be exploited in oncology.

Recent estimates suggest that 1% to 1.5% of all genes are hypoxia regulated (4). Thus, cells in hypoxic regions of tumors are commonly phenotypically different from normoxic cells, with some proteins increasing and others decreasing in abundance. In an effort to identify a hypoxia-induced protein that might be upregulated sufficiently in anaerobic cancers to be exploited for selective drug targeting, we searched for cell surface proteins that were dramatically induced by hypoxia yet absent or nearly absent in normoxic cells. The most prominent candidate to emerge from this search was carbonic anhydrase IX (CA IX), a membrane-spanning isoform of the intracellular carbonic anhydrases (8) that is significantly overexpressed in cancers of the lung (9), colorectum (10), stomach (11), pancreas (12), breast (13), cervix (14), bladder (15), ovaries (16), brain (17), head & neck (18), and kidneys (19). CA IX also appears to be upregulated on cancer cells on the edge of invasive tumors (20). In addition, CA IX is constitutively expressed in certain cancers, such as clear cell renal carcinomas where mutations in the VHL gene lead to continuous HIF-1α activation (19, 21, 22). Importantly, CA IX is absent from many healthy tissues (23), with high expression found only in the gastrointestinal tract and liver/gallbladder where most of the CA IX appears to be in a catalytically inactive form (23, 24). Low levels of CA IX can be found in skin (primarily hair follicles/sebaceous glands), testis, and salivary glands (23, 25).

Because of CA IX's cancer-enriched expression pattern, considerable effort has been expended to identify CA IX–specific ligands for use in tumor targeting. In fact, multiple CA IX ligands have been identified and conjugated to fluorophores (26–35), SPECT probes (34–48), and PET (49–59) agents with the goal of imaging hypoxic tumors. While therapeutic applications of some of these ligands have been explored in the literature, on the whole, less attention has been devoted to the important goal of eliminating malignant masses. Although an anti-CA IX antibody linked to a therapeutic radionuclide achieved only suppression of murine tumor xenograft growth (45), several small monovalent and bivalent CA IX ligands have shown much more promising results when coupled to cytotoxic drugs such as DM1 and duocarmycin (60, 61). Encouraged by these initial successes, we have undertaken to further explore the use of low molecular weight CA IX–targeting ligands to treat hypoxic tumors; only in this study, we have examined a different CA IX–specific ligand conjugated to an unrelated cytotoxic drug, tubulysin B.

Protected amino acids were purchased from Chem-Impex Intl. H-Cys (Trt)-2-Cl-Trt resin was obtained from Novabiochem. Tubulysin B and its activated derivatives were a kind gift from Endocyte Inc. 2-(1H-7-Azabenzotriazole-1-yl)-1,1,3,3-tetramethyl uronium hexafluorophosphate methanaminium (HATU) was obtained from Genscript Inc. Sulfuric acid, methanol, DMSO, DMF, TFA, isopropyl alcohol, NH₂-PEG₁₂-COOH-tBu, diisopropylethylamine (DIPEA), piperidine, CF₃COOH, CH₂Cl₂, K₂CO₃, tyramine, and all other chemical reagents were purchased from Sigma Aldrich. Purecoat Amine 24-well microtiter plates were purchased from BD Biosciences. All other cell culture reagents, syringes, and disposable items were purchased from VWR.

The CA IX ligand was prepared as described previously (62). For fluorescence imaging, the ligand was coupled with the rhodamine derivative (5-((5-aminopentyl)carbamoyl)-2-(6-(dimethylamino)-3-(dimethyliminio)-3H-xanthen-9-yl)benzoate hydrochloride) in the presence of EDC.HCl and HOBt in DMSO for 12 hours to yield the target conjugate. LRMS-LC/MS (m/z): [M + H]⁺ calculated for C₅₂H₆₀N₁₀O₁₃S₂, 1,097.39; found, 1,097.

The radioimaging conjugate was synthesized by the following solid-phase methodology: H-Cys(Trt)-2-chlorotrityl resin (100 mg, 0.56 mmol/L) was swollen with 3 mL of dichloromethane (DCM) followed by 3 mL of dimethylformamide (DMF). For three times, a 3-mL solution of 20% piperidine in DMF was added to the resin with argon bubbled through for 5 minutes. The resin was washed three times with 3 mL of DMF and 3 times with 3 mL isopropyl alcohol (i-PrOH). After swelling the resin in DMF, a solution of Fmoc-Asp(tBu)-OH (2.5 eq), PyBOP (2.5 eq), and DIPEA (4.0 eq) in DMF was added. Argon was bubbled for 2 hours, and resin was washed three times with 3 mL of DMF and 3 times with 3 mL i-PrOH.

The above sequence was repeated for three more coupling steps for addition of Boc-DAP(Fmoc)-OH, Fmoc-NH-PEG₁₂-COOH, and CA IX ligand. The final compound was cleaved from the resin using a trifluoroacetic acid (TFA): H2O: tri-isopropylsilane: cocktail (95:2.5:2.5) and concentrated under vacuum. The concentrated product was precipitated in diethyl ether and dried under vacuum. Crude conjugate was purified by preparative RP-HPLC [A = 2 mmol/L ammonium acetate buffer (pH 7.0), B = CH₃CN, solvent gradient: 0% B to 100% B in 30 minutes] to yield the requisite product named CA IX-PEG₁₂-EC20. LRMS-LC/MS (m/z): [M + H]⁺ calculated for C₅₉H₉₇N₁₁O₂₈S₃, 1503.57; found, 752.8 (half mass). For the L2-PEG₃₆-EC20 conjugate, LRMS-LC/MS (m/z): [M + H]⁺ calculated for C₁₀₈H₁₉₅N₁₁O₅₂S₃, 2,574.21; found, 1,281.6 (half mass).

CA IX–tubulysin B was synthesized from the CA IX–PEG₁₂–EC20 conjugate above. A solution of saturated sodium bicarbonate (2 mL) in HPLC grade water was bubbled with argon continuously for 10 min. CA IX-PEG₁₂-EC20 (50 mg, 0.033 mmol) was dissolved in argon-purged HPLC grade water (2.0 mL) and the pH of the reaction mixture was increased to 7 using argon purged sodium bicarbonate. A solution of disulfide activated-tubulysin B (37.47 mg, 0.034 mmol) in THF (2.0 mL) was then added to the reaction mixture. The progress of the reaction was monitored using analytic LC-MS, and after stirring for 30 minutes, the reaction was found to reach completion. Crude CA IX-tubulysin B was purified by preparative RP-HPLC [A = 2 mmol/L ammonium acetate buffer (pH 7.0), B = CH₃CN, solvent gradient: 0% B to 100% B in 30 minutes] to yield the requisite product. LRMS-LC/MS (m/z): [M + H]⁺ calculated for C₁₀₄H₁₆₄N₁₈O₃₉S₅, 2,450.84; found, 1,225 (half mass).

HCT-116 (transfected with a CA IX expression vector), HT-29, RCC4, and SK-RC-38 human cancer cell lines were obtained from the ATCC in February 2015. No authentication was done by the authors prior to experimentation. All cells were cultured in RPMI1640 medium supplemented with 10% FBS and 1% penicillin–streptomycin at 37°C in a humidified 95% air–5% CO₂ atmosphere. Cells were split in an approximately 1:8 ratio when flasks reached confluence. All testing was performed by the 10th passage.

HCT-116, HT-29, RCC4, and SK-RC-38 human cancer cell lines (10⁵) were seeded into chambered coverglass plates and allowed to grow to confluence over 48–72 hours. Spent medium was replaced with 0.5 mL of fresh medium containing 0.5% BSA, and various concentrations of the dye conjugate alone or the dye conjugate plus 100-fold excess CA IX ligand. After incubation for 1 hour at 37°C, cells were rinsed with incubation solution (2 × 1.0 mL) to remove unbound conjugate then washed with PBS (1 × 1.0 mL). Images were acquired using confocal microscopy (FV 1000, Olympus).

HT-29 cells (150,000 cells/well in 500 μL) were seeded into 24-well Falcon plates and allowed to form monolayers over 48 hours. Spent medium in each well was replaced with 0.5 mL fresh medium containing increasing concentrations CA IX-PEG₁₂-EC20 bound with ^99mTc or CA IX-rhodamine in the presence or absence of 100-fold excess CA IX ligand where appropriate. After incubating for 1 hour at 37°C, cells were rinsed twice with 1 mL of medium and 1 mL of Tris buffer. After dissolving in 0.5 mL of 0.25 mol/L NaOH (aq), for ^99mTc conjugates, lysate was transferred into individual γ-counter tubes and radioactivity was counted using a γ-counter (Packard, Packard Instrument Company). For CA IX-rhodamine–treated cells, lysate was added to a quartz cuvette and fluorescence was measured in a fluorimeter. Apparent K_d was calculated by plotting bound radioactivity versus the concentration of radiotracer using GraphPad Prism 4.

Athymic nu/nu mice were purchased from Harlan Laboratories, housed in a sterile environment on a standard 12-hour light–dark cycle and maintained on normal rodent chow. All animal procedures were approved by the Purdue Animal Care and Use Committee in accordance with NIH guidelines.

Five-week-old male nu/nu mice were inoculated subcutaneously with HT-29 cells (5.0 × 10⁶/mouse) on their shoulders. Growth of the tumors was measured in two perpendicular directions every 2 days using a caliper (body weights were monitored on the same schedule), and the volumes of the tumors were calculated as 0.5 × L × W² (L = longest axis and W = axis perpendicular to L in millimeters). Once tumors reached between 400 and 500 mm³ in volume, animals were administered ^99mTc-bound conjugates (10 nmol, 150 μCi) in saline (100 μL) via tail vein injection. At various times, animals were sacrificed by CO₂ asphyxiation. Images were acquired after shielding the kidneys using a Kodak Imaging Station (In-Vivo FX, Eastman Kodak Company) in combination with CCD camera and Kodak molecular imaging software (version 4.0). For radioimages, illumination source = radio isotope, acquisition time = 3 minutes, f-stop = 4, focal plane = 5, FOV = 160, binning = 4. For white light images, illumination source = white light transillumination, acquisition time = 0.05 s, f-stop = 16, focal plane = 5, FOV = 160 with no binning. Following imaging, animals were dissected and selected tissues were collected to preweighed γ-counter tubes. Radioactivity of preweighed tissues and ^99mTc-bound conjugates (10 nmol, 150 μCi) in saline (100 μL) was counted in a γ-counter. CPM values were decay corrected, and results were calculated as % injected dose (ID)/gram of wet tissue.

HT-29 cells were seeded on amine-coated 24-well plates and allowed to form monolayers. The spent medium in each well was replaced with fresh medium (0.5 mL) containing various concentrations of tubulysin B or CA IX-tubulysin B in the presence or absence of 100-fold excess CA IX ligand. After incubating for 1 hour at 37°C, cells were rinsed 3 times with fresh medium and then incubated an additional 24 hours at 37°C in fresh medium. Spent medium in each well was again replaced with fresh medium (0.5 mL) containing ³H-thymidine (1 μCi/mL) and the cells were incubated for an additional 4 hours. After washing the cells three times with medium, they were dissolved in 0.5 mL of 0.25 mol/L NaOH. Thymidine incorporation was then determined by counting cell-associated radioactivity using a scintillation counter (Packard, Packard Instrument Company). The IC₅₀ value was derived from a plot of the percent of ³H-thymidine incorporation versus log concentration using Graph Pad Prism 4 and TableCurve 2D software.

Various concentrations of CA IX-tubulysin B were administered to male nu/nu mice thrice weekly via tail vein injection. Visibly sick mice were recorded and if necessary euthanized before the end of the experiment.

HT29 cells (5.0 × 10⁶) were injected into the shoulders of 5- to 6-week-old female nu/nu mice. Tumors were measured in two perpendicular directions thrice weekly with vernier calipers and their volumes were calculated as 0.5 × L × W², where L is the longest axis (in millimeters), and W is the axis perpendicular to L (in millimeters). Dosing of CA IX-tubulysin B in the presence or absence of 100-fold excess CA IX ligand was initiated when the subcutaneous tumors reached approximately 100 mm³ in volume. Dosing solutions were prepared in saline and filtered through a 0.22-μm filter. Solutions were administered via tail vein injection or intraperitoneally. Each mouse received 2 μmol/kg CA IX-tubulysin B per injection. Injections were given thrice weekly for 3 weeks and the mice were weighed concurrently.

Ligand-targeted therapeutic agents are designed to selectively deliver imaging or therapeutic agents directly to the pathologic cells. The direct targeting of therapeutics typically reduces the efficacy as compared with the free drug; however, normal tissues are spared exposure leading to greatly reduced levels of toxicity. So, in an effort to identify ligands for use in targeting hypoxic tumors, a literature search for potent ligands of the hypoxia-induced cell surface enzyme, carbonic anhydrase IX (CA IX), was conducted. A suitable polyamino-polycarboxylamido aromatic sulfonamide ligand was chosen for conjugation (Fig. 1) due to its K_i of 7.8 nmol/L (62). In addition, the divalent nature of this molecule may aid in initial binding and internalization. Initially, this ligand was conjugated to a rhodamine via a five carbon alkyl chain to both explore the impact of conjugation on cellular binding and to ensure that the conjugate would bind to a variety of CA IX–expressing cancers. For this purpose, the CA IX–rhodamine conjugate was incubated with the CA IX–expressing cell lines HCT-116, HT-29, RCC4, and SK-RC-38 and examined by confocal microscopy. As shown in Fig. 2, binding is observed in every cell line tested, although HT-29 cells appeared to internalize the conjugate at a slower rate. To further assess binding, various concentrations of CA IX-rhodamine were incubated with HT-29 cells in the presence or absence of unbound ligand. The apparent K_d was found to be 105 nmol/L. Taken together, these data suggest that conjugation to one of the carboxylic acids does not markedly compromise binding. Importantly, when the fluorescent conjugate was incubated in the presence of 100-fold competing unconjugated CA IX ligand, essentially no binding was observed. This result shows that the binding is indeed receptor-mediated and that the conjugate was able to bind CA IX expressed on multiple different cell types (e.g., colon cancer and renal cell carcinomas). In addition, the punctate appearance of the fluorescence is indicative of internalization of the bound conjugate into endosomes during the 1-hour incubation period. This rapid uptake may be beneficial for the delivery of therapeutic agents.

Once a suitable conjugation point was identified on the CA IX ligand, a PEG₁₂ linker and ^99mTc-binding moiety (EC20) were coupled to the ligand on the same carboxylic acid via the same amide reaction used for synthesis of the rhodamine conjugate (Fig. 1). The binding of this conjugate was saturable with an apparent dissociation constant of 54 nmol/L, and binding was successfully competed by addition of 100-fold excess free ligand. As the original ligand is reported to have a K_i of 7.8 nmol/L (62), the addition of the PEG₁₂ linker and the ^99mTc-binding moiety resulted in an approximately 7-fold loss of binding affinity. This loss of binding affinity is similar to other ligand-targeted conjugates we have synthesized and the overall affinity was deemed acceptable to move into in vivo studies.

Next, to determine whether the CA IX conjugate would specifically accumulate in CA IX–expressing tumors, a biodistribution was determined in mice bearing HT-29 xenografts. To accomplish this, 8 nmol of ^99mTc-chelated conjugate was injected via the tail vein and mice were euthanized 4 or 22 hours later. The major organs/tissues were removed and the amount of radioactivity was determined. As shown in Fig. 3, the tumor exhibited the greatest percentage of injected dose (ID/g) at the 4-hour time point. Most of the uptake by the tumor was competed when 100-fold excess of unconjugated ligand was co-administered, showing that binding in vivo is also receptor mediated. In addition, the tumor was the only tissue that showed a statistically significant difference (P = 0.038) between the competed and non-competed groups at the 4-hour timepoint (Supplementary Fig. S1). The kidneys appear to be the primary route of excretion; however, signal in the intestine in the competition group may be due to some hepatic clearance as well. After 22 hours, the conjugate was nearly completely removed from all tissues including the tumor. Overall, these data support the fact that the CA IX–technetium conjugate specifically accumulates in the CA IX–expressing tumor.

When compared against several other CA IX–imaging agents, the accumulation within the tumor of this ^99mTc-chelated conjugate is somewhat lower. For example, CA IX conjugates with the best BioD had uptake of 10%–20% ID/g in the tumor (34, 53, 54) at 4 hours. Other conjugates exhibited approximately 1%–5% ID/g (51) and a few poorly localized to the tumor with <1% ID/g (38, 39). Interestingly, the linker appears to play a large role in determining the biodistribution of polyamino-polycarboxylamido aromatic sulfonamide CA IX ligands. For example, when this ligand was coupled to longer PEG₃₆ and semi-rigid Proline₃-PEG₁₂ linkers, tumor uptake was approximately 5% ID/g (47). Why the linker has such an impact on this class of ligands is currently not known, but shows that exploring different linkers can be very important when optimizing the biodistribution of drug conjugates.

With receptor-mediated accumulation established in the tumor, the cytotoxic drug, tubulysin B, was conjugated to the sulfur of the CA IX-PEG₁₂-EC20 via a disulfide bond (Fig. 1). Tubulysin B, a natural product that inhibits tubulin polymerization was chosen due to its potent toxicity toward numerous cell lines (63). The long hydrophilic spacer was used to successfully block passive diffusion of the tubulysin B conjugate into nontargeted cells. To determine the cytotoxicity of this conjugate in a CA IX–dependent and independent fashion, first, CA IX-tubulysin B was incubated with HT-29 cells for 1 hour in the presence or absence of 100-fold unconjugated CA IX ligand. Then, cells were washed followed by a 24-hour incubation period. The number of viable cells was determined via ³H-thymidine uptake. As shown in Fig. 4A, the IC₅₀ value for the CA IX–targeted tubulysin B was 4.4 nmol/L. In contrast, when the targeted conjugate was competed with 100-fold excess of the free ligand, essentially no cytotoxicity was observed up to 1 μmol/L. The fact that free tubulysin B exhibited an IC₅₀ value of 0.08 nmol/L suggests that little free drug was released prior to receptor-mediated uptake by the target cells. Taken together, these results demonstrate the requirement of receptor binding for cytotoxicity and the inability of the conjugate to passively diffuse through the cell membrane.

Prior to in vivo efficacy studies, a brief MTD study was performed. Three treatment groups consisting of 3 mice per group were administered increasing concentrations of CA IX-tubulysin B. As shown in Fig. 4B, no adverse events were observed when 1 μmol/kg CA IX-tubulysin B was administered. However, at the highest dose, 5 μmol/kg, all mice showed signs of toxicity. Therefore, a dose of 2 μmol/kg was used for all further in vivo efficacy studies. For evaluation of in vivo efficacy, 2 μmol/kg of drug was administered intraperitoneally or via tail vein injection every other day for 9 doses. As shown in Fig. 5, both administration routes essentially halted growth of the HT-29 tumor xenografts while CA IX-tubulysin B is administered. As the tumor did not completely regress, cancer regrowth would likely occur once administration is halted in this mouse model. When the cytotoxic conjugate was administered in conjunction with 100-fold excess CA IX ligand, growth of the tumors was similar to that of the control group (statistical summary can be found in Supplementary Figs. S2 and S3). A free tubulysin B treatment group was not included in these experiments due to the known high toxicity of systemic administration of the free drug. These data demonstrate that the therapeutic efficacy in vivo is also mediated by CA IX–specific targeting. As no obvious toxicity was observed and no weight loss was detected, we conclude that this dosing level was both safe and effective in this murine tumor model (Fig. 5).

Because hypoxic regions of tumors are reported to harbor cells that are unusually difficult to eradicate (5–7), new strategies are required to assure that chemotherapeutic agents can reach these oxygen-depleted environments. Recognizing the strong upregulation of CA IX by hypoxia, the constitutive expression in certain cancer types, and presence on the leading edge of invasive tumors, we undertook to develop a CA IX–targeted drug that could concentrate in these more CA IX–enriched areas of a tumor. Data presented above reveal that the CA IX ligand identified here can indeed promote suppression of CA IX–expressing tumor growth when conjugated to the highly potent microtubule inhibitor, tubulysin B. It will now be important to find chemotherapeutic agents known to eradicate normoxic tumors that will most effectively synergize with this and other CA IX–targeted drugs.

P.S. Low is a chief scientific officer at and reports receiving a commercial research grant from Endocyte, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: P.-C. Lv, J. Roy, P.S. Low

Development of methodology: P.-C. Lv, J. Roy

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.-C. Lv

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.-C. Lv, K.S. Putt

Writing, review, and/or revision of the manuscript: P.-C. Lv, J. Roy, K.S. Putt, P.S. Low

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.-C. Lv

Study supervision: P.S. Low

The authors gratefully acknowledge the campus-wide mass spectroscopy facility and support from the Purdue University Center for Cancer Research (P30CA023168).

Funding for this project was made possible by a grant from Endocyte, Inc (to P.S. Low).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.