[¹⁷⁷Lu]Pertuzumab: Experimental Therapy of HER-2–Expressing Xenografts Free

The basic characteristics of [¹⁷⁷Lu]pertuzumab, a radioactive labeled monoclonal antibody binding to HER-2, have recently been described (1). The high SKOV-3 xenograft radionuclide uptake compared with other organs encouraged us to do experimental therapy using the same animal model.

HER-2, or ErbB2, is a transmembrane receptor-like protein, which is part of the HER family of tyrosine kinase receptors (2, 3). HER-2 is unique in this family because it has no known ligand but seems to be constitutively open to dimerization (whereas other members of the receptor family have to be stimulated by a ligand; ref. 4). HER-2 has been found to be the preferred partner for heterodimerizations with the other HER family members, and this may be the main reason for the oncogenicity of HER-2 (5). Pertuzumab (rhuMab 2C4) is designed to bind to the dimerization domain (domain II) of HER-2 and sterically hinder heterodimerization or homodimerization (6, 7). The antibody has been found safe and clinically active in a recently published phase I clinical trial (8).

HER-2 is most commonly known to be overexpressed in cancers of the breast, especially in cancers with bone metastases (9, 10), but has also been reported to be overexpressed in bladder cancer (11) and, to some extent, in cancers of the ovaries (12) and gastric region (13).

Different radionuclides can be used for targeted radionuclide therapy. We have selected the low-energy β emitter ¹⁷⁷Lu, which has a spectrum of emitted β energies suitable for targeting of micrometastases (14, 15). The CHX-pertuzumab complex gives a chelate-mediated binding of ¹⁷⁷Lu. Micrometastases, not detectable at the time of primary tumor treatment, often cause tumor relapse. We hope that the application of [¹⁷⁷Lu]pertuzumab in an adjuvant setting can help alleviate this problem.

Cell line, chemicals, equipment, and abbreviations. All experiments were done using the SKOV-3 cell line (ATCC HTB 77). The cells were cultivated in 25-, 75-, or 175-cm² Δflasks (Nunc, Wiesbaden, Germany) using HAM's F-10 medium complemented with 10% (v/v) fetal bovine serum (Sigma, St. Louis, MO), 2 mmol/L l-glutamine, and 100 IU/mL penicillin/streptavidin (Biochrom, Berlin, Germany).

Omnitarg was a kind gift from Genentech (South San Francisco, CA). The antibody pertuzumab was purified by size-exclusion chromatography of the Omnitarg preparation using MilliQ-water as eluent on a PD-10 column (GE Healthcare, Uppsala, Sweden). The purified antibody was freeze-dried in a Heto FD 3 (Heto, Holten, Denmark) overnight. Isothiocyanate-benzyl-CHX-A″-diethylenetriaminepentaacetic acid (DTPA) was purchased from Macrocyclics (Dallas, TX). Gamma counting was done using an automated well crystal scintillator (Wallac Wizard 1480 3″, Perkin-Elmer, Wellesley, MA) within an energy window of 30 to 340 keV. Instant TLC SG strips were purchased from Gelman Sciences (Ann Arbor, MI) and were analyzed using a cyclone phosphor-imaging system (Perkin-Elmer).

The chimeric monoclonal antibody cMAb U36 against CD44v6 was used as negative control because SKOV-3 does not express CD44v6 in detectable amounts. The selection, production, and purification of the antibody have been described previously (16–18). The antibody was a kind gift from Prof. van Dongen (Department of Otolaryngology/Head and Neck Surgery, Vrije University Medical Center, Amsterdam, the Netherlands).

Samples for histology were dehydrated in a Ventana vacuum-infiltrating processing (Dalab, Sweden) using an 11-step program with increasing concentrations of ethanol, followed by xylene, and paraffin treatment. The samples were embedded in paraffin using a Tissue-Tec TEC embedding machine (Sakura Finetek, Tokyo, Japan) and cut in a Leica RM2165 Microtome (Leica, Leipzig, Germany).

¹⁷⁷Lu was purchased from NRG (Petten, The Netherlands).

Animals. Xenografts were formed by s.c. inoculation in the right posterior leg of BALB/c (nu/nu) mice (M&B, Ry, Denmark) injecting (5-7) × 10⁶ cells in 100 μL medium. The experiments comply with current Swedish law and were done with permission from the Uppsala committee of animal research ethics.

Labeling. Conjugation and labeling was done according to previously described and optimized protocols (1, 19). Briefly, the antibodies (60 nmol/L) were incubated overnight with isothiocyanate-benzyl-CHX-A″-DTPA (130 nmol/L) in 0.07 mol/L borate buffer (pH 9.2). The conjugated antibodies were purified on a NAP-5 column (GE Healthcare), and the buffer was changed to metal-free 1 mol/L sodium acetate buffer with 5 mg/mL ascorbic acid (pH 5.6). Chelation to ¹⁷⁷Lu was done for 30 min to 1 h, and the chelated products were purified from free ¹⁷⁷Lu on NAP-5 columns with PBS (pH 7.2) as running buffer. Radiochemical purity after purification was >99% according to instant TLC (SG plates) with 0.2 mol/L citric acid buffer. The specific activity after labeling was 22 MBq/mg for the biodistribution study with [¹⁷⁷Lu]pertuzumab. For the tumor treatment study, the specific activities were 200 MBq/mg for [¹⁷⁷Lu]pertuzumab and 300 MBq/mg for [¹⁷⁷Lu]U36.

Biodistribution. Xenografts were allowed to grow to a size of ∼5 mm diameter before [¹⁷⁷Lu]pertuzumab was injected i.v. into the tail vein of the mice. Four animals were sacrificed at each time point, and a panel of organs (Table 1) was collected for gamma counting. Euthanasia was done by an overdose of anesthesia (ketamin and xylazin) followed by exsanguination. The mice tails and empty injection syringes were also collected and measured to assess the amount of radioactivity released in the bloodstream after injection. Calibration was made against a set of three standards. The results are given as % injected dose/gram in Fig. 1. In Table 1, the results are given as standard uptake values (SUV) calculated by the formula: SUV = (A_o × m_m) / (A_m × m_o), where A_o is the number of radioactive counts per minute (cpm) in the studied organ, m_m is the weight (g) of the mouse, A_m is the total cpm value for the mouse, and m_o is the weight (g) of the studied organ. SUV >1 denotes accumulation in that organ.

Gamma camera imaging. Static images with a gamma camera were collected as described in Persson et. al (1) and will only be described briefly here. Two mice were injected i.v. with 2 MBq [¹⁷⁷Lu]pertuzumab (0.3 pmol). After 3 days, the mice were killed and imaged for 600 s with a gamma camera [Millennium VG with a 5/8″ NaI(Tl) crystal, General Electric Medical Systems, Haifa, Israel] equipped with a pinhole collimator [MEGP (VPC5), zoom factor 2.0]. Data from the two set energy windows (113 keV ± 10% and 208 keV ± 10%) were added together and presented as a 256 × 256 matrix.

Dosimetry. The organ uptake values, non-corrected for physical half-life, were time integrated to obtain the residence time per gram tissue for dosimetry calculations. Integration between time 0 and 14 days was made by the trapezoid method. The two last time points were fitted to a single exponential function, which was used to estimate the residence time from 14 days to infinity. The extrapolated area was in all organs <18% of the total calculated area.

The mean uptake values and their SDs given in Table 1 were used to randomly generate new statistically normally distributed uptake values. These new sets of uptake values were time integrated in the same way as described above. This produced a set of 30 randomly distributed residence time values. The relative error (1 relative SD) in this data set was used as the relative error of the calculated absorbed dose in the organ.

The radioactive decay of ¹⁷⁷Lu produces mainly low energy β particles. S values for ¹⁷⁷Lu were obtained from RADAR phantoms (Unit Density Spheres) published on the Internet.⁴

Tumor treatment. The agents for tumor treatment were given i.v. in tail vein 7 days after xenograft inoculation. The study was a semi-randomized blind setup. Groups of mice were randomly selected for one of five treatments: 7 MBq [¹⁷⁷Lu]pertuzumab (n = 8 mice), 5 MBq [¹⁷⁷Lu]pertuzumab (n = 8 mice), 7 MBq [¹⁷⁷Lu]U36 (unspecific control, n = 7 mice), native pertuzumab (non-labeled antibody control, n = 9 mice), and PBS (vehicle, n = 8 mice). After injection, the key was sealed in an envelope, keeping the identity of the groups secret until after the data analysis.

The mice were weighed, and tumor sizes were measured two to three times per week. The tumor size measurement was done using an electronic slide caliper. The tumor volume was calculated using an ellipsoid estimation of tumor form. The mice were followed until their tumors reached a volume of ∼1 cm³ before they were euthanized and dissected. Dissection consisted of removal of tumor for final volume measurement (length, breadth, and depth) and removal of kidneys for analysis of possible renal damage. When mice had to be euthanized due to tumor size, their tumor size was at all later times considered to be 1 cm³.

Survival curve plots and Kaplan-Meier analysis were done using Prism 4 (GraphPad Software, San Diego, CA). The end point was a tumor volume reaching 0.5 cm³. Mice removed from the study due to other reasons than having reached the end point were censored. The study was, for the remaining animals, finished after 102 days at which time they were censored. Euthanasia was done with an overdose of anesthesia (ketamin and xylazin) followed by cervical dislocation.

Histology. Tumors and kidneys were fixed in 4% phosphate-buffered formalin, processed, and embedded in paraffin. From selected blocks, 4-μm-thick sections were cut, put on coated slides (Superfrost plus, Menzel-Gläser, Germany), and dried for 12 h at 37°C. The sections were deparaffinized in xylene and rehydrated through graded concentrations of ethanol to distilled water.

The kidney slices were then stained with H&E.

The tumor slices were treated with hydrogen peroxide doped PBS for 10 min to quench endogenous peroxidase. Retrieval was preformed by pressure boiling the samples in citrate buffer for 7 min. The samples were stained for HER-2 with the A0485 antibody (DAKO, Glostrup, Denmark). Antibody dilutions 1:300 and 1:900 were applied, and the samples were incubated in 4°C overnight. After rinsing, the samples are treated with one drop EnVision antibody (DAKO) per sample and incubated for 45 min in room temperature. The samples were rinsed again and then developed using Fexin plus (DAKO) for 15 min.

The uptake of [¹⁷⁷Lu]pertuzumab was low in most organs as can be seen in Table 1 and Fig. 1A  to C. The radioactivity concentration, expressed as percent injected activity per gram organ, was always below 8% for all normal organs, excluding blood. The tumor uptake values were at all times higher than the values in normal organs, including blood. The SUV were below 1.8 for all normal organs, excluding blood. This concentration decreased with time in all organs, excluding the tumor xenografts, which continued to accumulate radioactivity until 5 days after injection. The accumulation of radioactivity in the tumor combined with decreasing activity concentration in other organs leads to a nuclide half-life corrected tumor-to-blood ratio starting at 1.4 at 8 hours after injection and continually increasing to a maximum of 19 after 2 weeks.

The gamma camera images collected 3 days after injection are shown in Fig. 2. The tumors accumulated large amounts of ¹⁷⁷Lu. Small amounts of radioactivity could also be found in the liver area. A region of interest analysis of the tumor and corresponding lateral region gave a tumor to background ratio of 29 for the left mouse and 24 for the right mouse.

The dosimetry calculations indicated that the absorbed dose to the tumor would, in our case, be at least five times the dose given to blood. For other normal organs, the situation was even better. Calculated absorbed doses to selected organs are presented in Table 2.

No groups of mice showed any significant weight loss after injection. The only exception was one individual mouse in the native pertuzumab group (see below). The weight gain that could be discerned did not vary between the groups.

Tumor growth. The tumor growth did not vary between groups until 25 days after xenograft inoculation (18 days after treatment start), but as can be seen in Fig. 3, the groups that received PBS and native antibody then experienced a rapid tumor growth during an additional period of ∼25 days. The mice with tumors that reached the size of about 1 cm³ were sacrificed. The growth of tumors treated with radioactivity (targeted with pertuzumab antibody or non-targeted with U36 antibody) was slower, but only the tumors that received radioactivity targeted via pertuzumab delivery showed a clear delay in tumor growth (Fig. 3).

Kaplan-Meier analysis. The result from the Kaplan-Meier analysis is shown in Fig. 4. The end point criterion was tumors growing ≥0.5 cm³. This end point was selected because it was noticed that after reaching 0.5 cm³, tumors grew exponentially. Most mice (31 of 40) reached the end point, but nine mice did not: six of these mice had still not reached the end point at the end of the experiment, three in the 7 MBq [¹⁷⁷Lu]pertuzumab group, two in the 5 MBq [¹⁷⁷Lu]pertuzumab group, and one in the 7 MBq [¹⁷⁷Lu]U36 group. The [¹⁷⁷Lu]U36-treated mouse had no obvious xenograft take. One mouse from the 7 MBq [¹⁷⁷Lu]pertuzumab group was killed at day 100 due to large apparent tumor size, but at dissection, it was discovered that the tumor was rather small. Instead, most of the apparent volume was formed by a large reservoir of fluid, possibly an inflammatory process. This tumor was considered to not have reached the end point. All these mice were censored.

One mouse in the native pertuzumab group had to be removed from the study after 28 days due to weight loss. This mouse did not receive radioactivity, and the weight loss was probably due to stress or illness. One mouse from the 5 MBq [¹⁷⁷Lu]pertuzumab group was killed at day 72 due to tumor necrosis that spread to the dermis, causing open wounds. Both of these mice were censored. Open wounds were also observed in a few other mice but only after they reached the end point tumor volume of 0.5 cm³.

The obtained tumor take values corresponded well to previous experiences with this model system where the tumor takes have been ∼90% to 95% (data not shown).

Analysis of survival with Kaplan-Meier analysis showed that the group that received 7 MBq [¹⁷⁷Lu]pertuzumab had a significantly improved time to tumor progression versus mice that received PBS, native pertuzumab, or 7 MBq [¹⁷⁷Lu]U36 (P < 0.0001, P < 0.0001, and P < 0.01, respectively). Mice receiving 5 MBq [¹⁷⁷Lu]pertuzumab had a significantly longer time to tumor progression than mice that received PBS or native pertuzumab (P < 0.0001 and P < 0.0005, respectively).

No morphologic damage was detectable in sections of either irradiated or nonirradiated kidneys, neither in a blind or open examination. The tumor cells, analyzed in tissue sections, were found to express HER-2 whether treated or not. Treated tumors seemed, morphologically, somewhat more chaotic than untreated, showing large heterogeneity with partial degradation, fibrotic regions, and lymphocyte infiltration.

The biodistribution study was designed to give information on both early and late kinetics (uptake, retention, and clearance) to allow for dosimetry calculations. The results from the biodistribution study showed that the tumor uptake of [¹⁷⁷Lu] was higher than the uptake in all analyzed normal organs and in blood. Furthermore, the low radioactivity concentration in bone indicated that the lutetium was not lost from the chelator in vivo because free lutetium is a known bone “seeker” (20). The residence times in the tumors were high, as could be expected from our previously published retention studies (1). The tumor-to-blood ratio increased from about 1.4, 8 h after injection, up to about 19 after 2 weeks, which is a promising result.

The simplified dosimetry calculation used in this article, which do not include β crossfire and photon contribution, is motivated by the local absorption of the low-energy β particles and the low abundance of photons and other penetrating radiations in the ¹⁷⁷Lu decay. Thus, the main uncertainty in the dose calculations is not due this simplified model but rather due to the variation in kinetic data. The data show that the absorbed dose to the tumor is dominant to all normal organs.

For many radiolabeled antibodies used in targeted therapy, the critical organ usually is the red bone marrow due to crossfire from radioactivity in the blood. Because blood in our study is the organ that is obtaining the highest absorbed dose besides the tumor, one may suspect that bone marrow, also in targeting therapy with [¹⁷⁷Lu]pertuzumab, is the critical organ. However, the generally accepted coupling factor (red marrow-to-blood activity concentration ratio) is 0.36 (21). Because the absorbed dose tumor-to-blood ratio is ∼5, the expected tumor-to-bone marrow ratio would be 14, which is still an advantageous factor and indicate a possibility to obtain good therapy results.

The goal of clinical treatment with [¹⁷⁷Lu]pertuzumab is primarily to eliminate micrometastases not visible at the time of surgery. With this in mind, [¹⁷⁷Lu]pertuzumab was, in the experimental therapy experiments, given 7 days after xenograft inoculation. The xenografts had time to establish but not to grow into a detectable solid tumor.

The tumor growth curves were most retarded in the groups that received radioactivity targeted with pertuzumab, which of course is promising for the planning for clinical therapy trials on HER-2–overexpressing tumors. Furthermore, no groups of mice showed any significant weight loss after injection, indicating no major acute toxicity from the antibodies and/or the radiation.

To facilitate statistical analysis of the treatments, a survival curve analysis of Kaplan-Meier type was done. This analysis showed that the group that received 7 MBq [¹⁷⁷Lu]pertuzumab had a significantly increased time to tumor progression in relation to the controls. Mice receiving 5 MBq [¹⁷⁷Lu]pertuzumab also had a significantly longer time to tumor progression than mice that received PBS or native pertuzumab. Thus, the HER-2–specific radionuclide therapy proved more efficient than not HER-2–specific treatments.

However, although the time to tumor progression was significantly improved using HER-2–targeted radiotherapy, only two of all HER-2–targeted mice did not resume tumor progression. Immunohistology of sections from these two tumors showed presence of HER-2 in the remaining tissue, but because no tumor growth was observed in these cases, the HER-2–positive tumor cells seemed unable to divide.

We believe that the therapy could be improved further by using higher amounts of [¹⁷⁷Lu]pertuzumab. It is probable that the maximum tolerated doses have not been reached in this study. This is also indicated because no adverse effects could be found by the treatment. The obtained therapy results are encouraging, especially considering that the transplanted SKOV-3 cells have been reported to be radioresistant (22).

The results showed that [¹⁷⁷Lu]pertuzumab can be used as an agent for HER-2–targeted radionuclide therapy in vivo. We believe that the optimal use for [¹⁷⁷Lu]pertuzumab would be in an adjuvant setting when there are strong indications of metastatic spread of tumor cells. We are currently planning studies of radionuclide uptake, when applying [¹⁷⁷Lu]pertuzumab in clinical phase I/II studies on breast and urinary bladder cancers overexpressing HER-2.

Grant support: The Swedish Cancer Society.

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.

We thank Veronica Asplund Eriksson for technical support, Genentech for the generous gift of Omnitarg, and Marika Nestor and Prof. Gus van Dongen for their kind donation of the control antibody U36.