Breast cancer brain metastases are a deadly sequela of primary breast tumors that overexpress human epidermal growth factor receptor 2 (HER2); median survival for patients with these tumors is 10 to 13 months from the time of diagnosis. Current treatments for HER2-positive breast cancer brain metastases are invasive, toxic, and largely ineffective. Here, we have developed an adeno-associated virus serotype 9 (AAV9) vector to express the anti-HER2 monoclonal antibody trastuzumab (Herceptin) in vivo. A single prophylactic intrathecal administration of AAV9.trastuzumab vector in a novel orthotopic Rag1−/− murine xenograft model of HER2-positive breast cancer brain metastases significantly increased median survival, attenuated brain tumor growth, and preserved both the HER2 antigen specificity and the natural killer cell–associated mechanism of action of trastuzumab. When administered as a tumor treatment, AAV9.trastuzumab increased median survival. Dose-escalation studies revealed that higher doses of AAV9.trastuzumab resulted in smaller tumor volumes. Our results indicate that intrathecal AAV9.trastuzumab may provide significant antitumor activity in patients with HER2-positive breast cancer brain metastases.

Significance: Intrathecal delivery of trastuzumab via adeno-associated virus has the potential to become a novel, integral part of adjuvant therapy for patients with HER2-positive breast cancer brain metastases. Cancer Res; 78(21); 6171–82. ©2018 AACR.

Breast cancer is the most commonly diagnosed malignancy in women in the United States with an estimated 268,670 new cases in 2018 (1). Approximately 20% of breast cancers overexpress human epidermal growth factor receptor 2 (HER2) and are considered to be more aggressive and more likely to metastasize to the brain than other breast cancer subtypes (2–4). About 30% of patients with metastatic HER2-positive (HER2+) breast cancer will develop breast cancer brain metastases (BCBM; refs. 5–8). In the registHER prospective study of patients with newly diagnosed HER2+ metastatic breast cancer, 37.3% of the 1,012 patients studied developed brain metastases within 10.8 months of the initial diagnosis of metastatic disease (9). Brain metastases can significantly lower patient quality of life by inducing nausea, sensory loss, aphasia, motor deficits, ataxia, seizures, stroke, and paralysis (3, 10–12). The median age at diagnosis of HER2+ BCBM is 48 years (13), and the incidence of HER2+ BCBM is rising (4). Evidence suggests that this increase is due to many factors. Targeted, effective therapies for HER2+ tumors in the periphery have increased the duration of patient survival, thereby allowing adequate time for outgrowth of brain metastases that would have otherwise remained subclinical before death (10, 14). Importantly, many of these therapies, including biological therapeutics, do not reach adequate concentrations in cerebrospinal fluid (CSF) after systemic administration (15, 16).

The current standard-of-care treatments for HER2+ BCBM are invasive, can cause cognitive impairment, and provide suboptimal survival benefits. Patients often undergo a combination of neurosurgical tumor resection, stereotaxic radiosurgery, whole-brain radiotherapy, systemic and/or intrathecal (i.t.) chemotherapy, steroids, or anti-HER2 agents (3, 10, 14). Even with these treatments, survival from the time of diagnosis of HER2+ BCBM ranges from 3 to 25 months with a median of 10 to 13 months (5, 9, 17, 18). Clearly, there is an unmet need for more effective, targeted treatments for patients with HER2+ BCBM.

Trastuzumab (Herceptin, Roche) is a humanized monoclonal antibody (mAb) directed against HER2 that extends survival of patients when used with chemotherapy to treat primary and systemically metastatic HER2+ disease (5, 19, 20). However, trastuzumab does not cross the intact blood–brain barrier to treat central nervous system (CNS) tumors (2, 3). Additionally, the CSF concentration of trastuzumab after intravenous (i.v.) administration is 300- to 400-fold lower than that in serum (15, 16). As such, patients with concurrent HER2+ BCBM and systemic HER2+ disease who receive i.v. trastuzumab often experience regression or stabilization of systemic tumor burden while brain metastases progress (5).

Trastuzumab administered i.t. has been reported to increase survival and delay progression of HER2+ brain metastases. Bousquet and colleagues administered i.t. trastuzumab to a patient with HER2+ cerebellar and epidural breast cancer metastases, resulting in a 6-month halt in disease progression (21). Colozza and colleagues described a patient with HER2+ intraparenchymal cortical metastases who was treated for 19 months with i.t. trastuzumab (22). In both case reports, patients were still alive at the time of publication. A 2013 meta-analysis by Zagouri and colleagues demonstrated that the median survival of patients with HER2+ leptomeningeal carcinomatosis, a particularly deadly form of BCBM, increases from 5.9 months in historical controls to 13.5 months with i.t. trastuzumab treatment (23). A recent phase I clinical trial in Europe demonstrated that administering doses of up to 150 μg of trastuzumab i.t. to patients with leptomeningeal HER2+ resulted in no serious adverse events (24).

Despite these promising reports, i.t. administration of trastuzumab has disadvantages. For instance, multiple i.t. administrations are required. Perhaps more importantly, the normal, rapid turnover of CSF results in a widely fluctuating pharmacokinetic profile of trastuzumab in CSF, resulting in a CSF half-life of just 12 hours (21, 25). It is therefore unlikely that tumor cells in the CNS receive optimal exposure to trastuzumab after i.t. administration.

Gene therapy offers the potential for a one-shot solution to the problem of mAb delivery across the blood–brain barrier. Adeno-associated viral (AAV) vectors, particularly serotype 9, can safely and efficiently deliver exogenous genes, such as the gene for trastuzumab, to neurons and astrocytes throughout the brain and spinal cord after a single i.t. administration. This results in long-term, stable expression of the transgene product in the brain parenchyma and CSF (26, 27).

In this study, we aimed to use AAV9.trastuzumab delivered i.t. to bypass the blood–brain barrier for localized expression of trastuzumab in situ. We developed a novel Rag1−/− murine orthotopic xenograft model of HER2+ BCBM, and then delivered AAV9.trastuzumab i.t. by intracranioventricular (ICV) injection either as tumor prophylaxis or treatment. In both cases, a single dose of i.t. AAV9.trastuzumab significantly extended median survival of mice compared with no treatment or control AAV vector treatment. Higher doses of vector led to smaller tumor volumes when measured 35 days after tumor implantation. We also showed that trastuzumab expressed by CNS cells still binds to HER2 on tumors and maintains the clinical product's principal mechanism of action against tumors: facilitating antibody-dependent cell-mediated cytotoxicity through natural killer (NK) cells. Looking ahead, we predict that i.t. AAV9.trastuzumab could prolong survival of patients with existing HER2+ CNS metastases in addition to patients with primary or metastatic HER2+ breast cancer at risk for developing BCBM.

Experimental design

Six- to 9-week-old B6.129S7-Rag1tm1Mom/J (Rag1−/−) and NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice were obtained from The Jackson Laboratory and housed at the barrier facility of the Translational Research Laboratories Vivarium at the University of Pennsylvania. A female cynomolgus macaque was obtained from Covance Inc. and housed at the animal facility of the Biomedical Research Building II/III at the University of Pennsylvania. All animals were maintained according to NIH guidelines for the care and use of animals in research.

All procedures and protocols were approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania. The 201Ig IA (201IA) control mAb is a rhesus antisimian immunodeficiency virus IgG immunoadhesin. The 2.10A control mAb is a rhesus antisimian/human immunodeficiency virus IgG. Vectors expressing these two antibodies served as negative controls for vector-mediated antibody production. AAV9.null, which serves as a negative control for the presence of vector transduction in the CNS, has an intact genome and expression cassette but has no transgene. Phosphate-buffered saline (PBS) was used as a control for the volume of vector administration. AAV9var is a closely related AAV9 variant that performs similarly in vivo.

We administered vector to mice at least 21 days before tumor implantation in the prophylaxis studies and 3 days after tumor implantation in the treatment studies. Prophylactic AAV9var.trastuzumab was administered 2 weeks before tumor implantation. Ten to 14 days are required for vector expression to peak and reach steady state; we wanted to ensure expression was at a steady state when tumor was injected.

We used a minimum of 8 mice per experimental group to allow for robust statistical analysis in case mice were euthanized due to complications during or after tumor implantation. The exception is the pilot study represented in Supplementary Fig. S1A and S1B. The large prophylaxis study began with 20 mice per group, the treatment study with 8 mice per group, and the NSG and NK cell depletion studies with 12 mice per group. Macrophage depletion studies began with 9 mice per group, with an additional 2 mice added to AAV9.trastuzumab–treated arms due to clodronate-related toxicity. Mice found dead rather than euthanized were included only in the survival analyses and not in downstream transgene expression, histologic, or biodistribution analysis. Studies were conducted without blinding.

Statistical analysis

Survival study and tumor volume P values were calculated using the log-rank (Mantel–Cox) test in GraphPad Prism software for Windows 7. Hazard ratios were calculated with a Cox regression model using the sts package (StataCorp).

Vector construction

Sequences matching the World Health Organization–published amino acid sequences of the heavy and light chains of trastuzumab were back translated, codon optimized, and synthesized by GeneArt (Life Technologies). Heavy and light chain sequences were preceded by a human IL2 secretion signal. The heavy and light chain sequences were cloned into an AAV expression construct containing an upstream hybrid cytomegalovirus (CMV) immediate early enhancer/chicken β-actin promoter, a chimeric intron (Promega), and a downstream simian virus 40 (SV40) polyadenylation signal. The heavy and light chain sequences were separated from each other by a foot-and-mouth disease virus-derived self-cleaving peptide to ensure 1:1 production of heavy and light chain protein. The construct was flanked by AAV2 inverted terminal repeats. The resulting pAAV.CMV.PI.trastuzumab.SV40 expression construct was packaged in an AAV9 or AAV9var capsid by triple transfection of 293 cells and purified as previously described (28). The regulatory elements in the AAV9.201IA expression cassette are a chicken β-actin (CB7) promoter, a chimeric intron, and a downstream rabbit beta-globin polyadenylation signal (rGB). The regulatory elements in the AAV9.null expression cassette are a thyroid hormone-binding globulin promoter, a microglobin/bikunin enhancer (TBG), and a bovine growth hormone (bGH) polyadenylation signal. Vector was titrated using standard quantitative polymerase chain reaction (qPCR) or digital droplet PCR (AAV9var.trastuzumab only). AAV9.CMV.PI.2.10AmAb.SV40, AAV9.CB7.PI.201IA.rBG, and AAV9.TBG.PI.null.bGH were obtained from the Penn Vector Core.

Vector administration

For tumor prophylaxis, mice were injected ICV with 1 × 1011 genome copies (GC) of AAV9.CMV.PI.trastuzumab.SV40, AAV9.CB7.PI.201IA.rBG, AAV9.TBG.PI.null.bGH, or AAV9.CMV.PI.2.10AmAb.SV40 diluted in sterile PBS at least 21 days prior to BT474.M1.ffluc tumor implantation. For tumor treatment, mice received BT474.M1.ffluc tumors first, and vector was administered by ICV injection 3 days after tumor implantation. AAV9var.CMV.PI.trastuzumab.SV40 was administered in half-logs from 1e10 to 3e11 GC/mouse for tumor volume studies.

Orthotopic xenograft model of HER2+ breast cancer brain tumors in Rag1−/− and NSG mice

The HER2+ BT474.M1 human ductal carcinoma cell line was a generous gift from Lewis Chodosh and Jason Ruth (University of Pennsylvania). IMPACT I testing (IDEXX BioResearch) indicated no Mycoplasma spp. contamination. Cell authentication was not conducted. Cells were transduced with the lentiviral vector VSVG.HIV.SIN.cPPT.CMV.ffluciferase.WPRE (Penn Vector Core) and cryopreserved in liquid nitrogen. One week before tumor implantation, BT474-M1.ffluc cells at passage 56 were thawed, expanded in DMEM/F12 (Corning) with 10% fetal bovine serum and 1% penicillin/streptomycin at 37°C and 5% CO2, and passaged 3 days before tumor cell implantation. Normal cell morphology was confirmed at time of passage and on the day of harvest. On the day of tumor injection, cells were counted with a Countess II cytometer (Life Technologies) and suspended at 1 × 105 cells per 5 μL in 50%/50% (v/v) PBS/MatriGel (Corning).

For the injection procedure, mice were anesthetized with ketamine/xylazine. Fur on the scalp and neck was sheared. A 1.7-mg, 90-day time-release 17-β estradiol pellet (Innovative Research of America) was implanted subcutaneously in the dorsum of the neck and readministered every 90 days during the study. Mice were fixed in a stereotaxic apparatus. Exposed skin was cleansed with povidone iodine and 70% ethanol. A 1-cm anterior–posterior incision was made over the top of the skull, and the bregma was identified. A pneumatic drill was used to drill a burr hole in the skull that was positioned 0.8 mm posterior and 2.2 mm left of the bregma.

A 25-μL syringe (Hamilton Company) was loaded with 5 μL of cell suspension and positioned in a mechanical injector on the stereotaxic frame. After bringing the needle to 0.8 mm posterior and 2.2 mm left of the bregma, the needle was moved 4.0 mm deep into the brain parenchyma and then lifted 1.0 mm to create a pocket into which to inject cells. The needle was left in place for 5 minutes before the cell suspension was mechanically injected over 10 minutes; the needle was left in place for another 5 minutes before being removed slowly. After suturing incisions, mice recovered on a 37°C heating pad and were given 100 μL of 2 mg/kg enrofloxacin in PBS and 0.3 mg/kg buprenorphine in PBS subcutaneously.

Mice were monitored daily. When moribund, mice were euthanized by overexposure to CO2 followed by cervical dislocation. At necropsy, brains and tumor were either snap frozen on dry ice and stored at −80°C for transgene expression and GC analysis, cryopreserved in optimal cutting temperature medium using liquid nitrogen, or preserved in formalin.

Tumor volume

To determine the dose effect of our treatment, we used an AAV9var vector, which behaves similarly to AAV9 in vivo. Vector was administered 2 weeks before tumor implantation. Thirty-five days after tumors were implanted, brains were harvested. Blunt dissection at the tumor injection needle track was used to isolate tumors from surrounding brain tissue. Measurement of tumor diameter was performed with digital Vernier calipers (Thermo Fisher). The tumor diameter was then measured in 3 dimensions (x, y, and z), and the tumor volume was calculated as the volume of an ellipsoid, 4/3 * π * x/2 * y/2 * z/2.

ICV Herceptin and AAV9.trastuzumab administration for quantification of trastuzumab in brain tissue

AAV9.trastuzumab was diluted in PBS. A total of 1e11 GC per Rag1−/− mouse was delivered ICV in a final volume of 10 μL. Two weeks later, brains were harvested.

Herceptin was diluted in PBS and administered ICV in a 10 μL volume at a dose of 15 μg/mouse brain, which is roughly equivalent to a 50-mg i.t. dose in humans by brain mass. Brains were harvested 24, 48, and 168 hours after ICV administration of Herceptin and homogenized as described above in 1-mL tissue lysis buffer. Protein A enzyme-linked immunosorbent assay (ELISA), performed as described below, was used to determine ng/mL of trastuzumab in brain homogenate. Multiplying this trastuzumab concentration by the homogenate volume of 1 mL yielded the amount of trastuzumab per brain in ng.

The half-life of Herceptin in brain tissue was calculated as follows:

t1/2 = t ln(2)/(ln (xi) − ln(xt)), where t = time point, xi is the initial quantity at t = i, and xt is the quantity at time point t.

NK cell depletion

PK136 (anti-mouse NK1.1) antibody was purified from PK136 hybridoma supernatant (ATCC HB-191) using the protein A purification kit (Sigma-Aldrich) and supplemented as needed with PK136 available commercially (Leinco Technologies). Mice were given 100 μg PK136 or PBS intraperitoneally (i.p.) on days 5 and 1 before tumor implantation and then weekly for the duration of the experiment.

Systemic macrophage depletion

Chlodronate liposomes or PBS liposomes (Nico van Rooijen, Vrije University Medical Center, Amsterdam, the Netherlands) were injected i.p. at 10 μL/g body weight on days 5 and 1 before tumor implantation, and then weekly for the duration of the experiment.

Nonhuman primate study

An adult female cynomolgus macaque weighing 4.38 kg was anesthetized with ketamine/dexmedetomidine, and the hair overlying the occiput and dorsal neck was shaved. A spinal needle was inserted suboccipitally directly into the cisterna magna, and 1 mL of CSF was collected. Vector was then infused into the cisterna magna in a volume of 1 mL over 1 minute. The needle was removed slowly followed by application of pressure to the puncture site. CSF was collected by the same procedure at indicated time points over the course of 6 months.

Preparation of serum

Blood was collected by retro-orbital or submandibular bleed into Z-Gel microtube serum separators (Sarstedt) and incubated at room temperature for 20 minutes. After centrifuging for 5 minutes at 5,000 RPM in a tabletop microcentrifuge (accuSpin micro 17, Thermo Fisher Scientific), serum was stored at −80°C.

Preparation of brain homogenate

Brain homogenates were prepared by chipping frozen brain into tissue lysis buffer (25 mmol/L Tris–HCl, 5 mmol/L EDTA, 1% Triton-X, 150 mmol/L NaCl, pH 7.6) containing 3 times the normal concentration of cOmplete Protease Inhibitor Cocktail Tablets (Roche). Samples were homogenized with stainless-steel beads on a TissueLyzer II (Qiagen) at 30 Hz for 2 minutes, frozen at −80°C, thawed, and then centrifuged at 17,000 × g for 60 minutes at 4°C to remove myelin debris. A bicinchoninic acid assay (Thermo Fisher Scientific) was used to determine the protein concentration of brain homogenates.

Protein A and mac251 enzyme-linked immunosorbent assay

Protein A ELISA was used to quantify trastuzumab and 2.10A mAb expression in serum and brain homogenates. 201IA was quantified using a specific SIV mac251 gp120 antigen ELISA. All steps were performed at room temperature unless otherwise stated. Plates were washed with a 405TS microplate washer (BioTek) with PBS + 0.05% Tween-20. Protein A (Sigma-Aldrich) was suspended in PBS and stored at −20°C. Costar 96-well EasyWash ELISA assay plates (Corning) were coated with 5 μg/mL protein A or 2 μg/mL mac251 gp120 in PBS overnight at 4°C and then blocked with PBS + 0.5% bovine serum albumin (protein A ELISA) or fetal bovine serum (201IA ELISA). Samples were diluted in PBS and plated. Herceptin (Roche) or purified 201IA was used as a quantitative standard. For brain homogenate ELISAs, brain homogenate from untreated mice was spiked into the standard curve wells at a dilution equal to that of samples on the plate. After washing, plates were incubated with AffiniPure polyclonal goat anti-human IgG-biotin (Jackson ImmunoResearch Laboratories) followed by streptavidin–horseradish peroxidase (Abcam). Plates were developed with TMB substrate, stopped with 2N H2SO4, and read using a SpectraMax M3 plate reader (Molecular Devices) at 450 nm.

Biodistribution by qPCR

AAV9 vector GCs in brain were quantified using TaqMan qPCR (Thermo Fisher Scientific). Briefly, frozen brain tissue was chipped into ALT Buffer (Qiagen) and homogenized with steel beads using a TissueLyzerII (Qiagen). Phenol:chloroform:isoamyl alcohol (25:24:1; Sigma-Aldrich) extraction and isopropanol precipitation was used to isolate tissue DNA. TaqMan qPCR primers and probe were designed against the SV40 polyadenylation signal of the vector.

Histology

HER2 staining was performed on formalin-fixed, paraffin-embedded tissue samples. Sections were deparaffinized through a xylene and ethanol series, boiled in a microwave for 6 minutes in 10 mmol/L citrate buffer (pH 6.0) for antigen retrieval, treated sequentially with 2% H2O2 (15 minutes; Sigma-Aldrich), avidin/biotin blocking reagents (15 minutes each; Vector Laboratories), and blocking buffer (1% donkey serum in PBS + 0.2% Triton for 10 minutes) followed by incubation with rabbit anti-HER2 primary antibody (1 hour; Abcam ab2428) and biotinylated donkey anti-rabbit secondary antibody (45 minutes; Jackson ImmunoResearch Laboratories) diluted in blocking buffer. Bound antibodies were visualized with a Vectastain Elite ABC kit (Vector Laboratories) using DAB as the substrate. Sections were counterstained with hematoxylin to show nuclei.

Immunofluorescence staining for human IgG was performed on cryosections. Sections were fixed in 4% paraformaldehyde in PBS for 10 minutes, permeabilized and blocked in 0.2% Triton in PBS containing 1% donkey serum for 30 minutes, and incubated for 1 hour with a goat antibody against the crystallizable fragment (Fc) of human IgG (Jackson ImmunoResearch Laboratories #109-005-098) diluted in 1% donkey serum in PBS. After washing sections in PBS, bound primary antibodies were detected with fluorescein isothiocyanate-labeled secondary donkey anti-goat antibodies (Jackson ImmunoResearch Laboratories) diluted in 1% donkey serum in PBS. After washing in PBS, sections were mounted with Vectashield (Vector Laboratories) containing DAPI as a nuclear counterstain.

IHC to detect NK cells and HER2 was also performed on cryosections. Sections for NK cell staining were fixed in acetone at −20°C for 7 minutes, air dried, sequentially treated with 0.3% H2O2 in PBS for 10 minutes, avidin/biotin blocking reagents (15 minutes each; Vector Laboratories), and blocking buffer (1% donkey serum in PBS, 20 minutes). Sections for HER2 staining were fixed in 4% paraformaldehyde in PBS for 10 minutes, permeabilized in 0.2% Triton in PBS for 30 minutes, and sequentially treated with 0.3% H2O2 in PBS (10 minutes) and avidin/biotin blocking reagents (15 minutes each; Vector Laboratories). Sections were then blocked with 1% donkey serum in PBS for 20 minutes, treated with primary antibody (1 hour), and corresponding biotinylated secondary antibody (45 minutes; Jackson ImmunoResearch Laboratories) diluted in 1% donkey serum in PBS. Primary antibodies were as follows: rat anti-mouse Ly-49G2 (clone 4D11, BD Pharmingen, BD Biosciences) for NK cells or rabbit anti-mouse Erb2 (Abcam ab2428) for HER2. A Vectastain Elite ABC kit (Vector Laboratories) was used according to the manufacturer's instructions with DAB as the substrate.

To detect NK cells on formalin-fixed, paraffin-embedded sections, in situ hybridization (ISH) was performed using the ViewRNA ISH Tissue Assay Kit (Thermo Fisher Scientific) according to the manufacturer's protocol. Z-shaped probe pairs specific for mouse Ncr1 (natural cytotoxicity triggering receptor 1, NKp46) RNA were synthesized by the kit manufacturer. The deposition of Fast Red precipitates indicating positive signals was imaged by fluorescence microscopy using a rhodamine filter set. Sections were counterstained with DAPI to show nuclei.

IT AAV9.trastuzumab can prevent or treat tumors in a novel Rag1−/− orthotopic xenograft model of HER2+ BCBM

I.t.-administered trastuzumab can increase survival and delay tumor progression in patients with HER2+ BCBM. However, trastuzumab has been reported to have a half-life of 12 hours in human CSF (21, 25). We aimed to administer AAV9.trastuzumab i.t., enabling constitutive secretion of trastuzumab by neurons and astrocytes in situ, better tumor exposure to the immunotherapy, and thus longer median survival.

We began by developing a novel xenograft mouse model to study the effect of i.t. AAV9.trastuzumab against HER2+ BCBM. In order to ensure our model had no endogenous IgG and was not “leaky” with respect to immune cell development, we chose to use Rag1−/− mice. As trastuzumab does not bind to murine HER2, using a xenografted cell line that expresses human HER2 was necessary for these experiments. Therefore, we selected the HER2+ human ductal carcinoma cell line BT474.M1.ffluc for use in our experiments. Importantly, BT474.M1.ffluc cells are relatively sensitive to trastuzumab treatment, making detection of small changes in efficacy after administration of immunomodulatory elements possible. We used these cells across all of our experiments in order to allow continuity between efficacy studies and studies conducted to determine the mechanism of our treatment.

To establish the model, we administered BT474.M1.ffluc cells (100,000) stereotactically into the caudate/putamen region of the mouse brain. We originally monitored tumor bioluminescence twice weekly following i.p. administration of luciferin. However, we found that the administration of luciferin concurrent with the presence of the subcutaneous 17-beta estradiol time-release pellets, used routinely to enhance engraftment, led to significant urinary and kidney pathology, requiring euthanasia of animals. We therefore ceased luciferase imaging for the remainder of our experiments and instead used survival and tumor volume as our primary endpoints.

We first tested AAV9.trastuzumab as HER2+ CNS tumor prophylaxis in our Rag1−/− xenograft model. In a pilot experiment, we administered AAV9.trastuzumab, AAV9.201IA (a negative control vector expressing a rhesus antisimian immunodeficiency virus IgG immunoadhesin), AAV9.null (AAV9 vector with an intact genome and expression cassette but no transgene), or PBS i.t. by ICV injection into Rag1−/− mice at least 3 weeks before stereotactic BT474.M1.ffluc tumor implantation (100,000 cells). Mice that received AAV9.trastuzumab tumor prophylaxis survived significantly longer than mice that received control treatments [Supplementary Fig. S1A and S1B; 111 days vs. 48.5 days (PBS control); P = 0.0012]. Importantly, the median survival among groups of mice that received control treatments did not significantly differ.

We next repeated this experiment with additional mice in each treatment group and using a full-length IgG-negative control vector (AAV9.2.10AmAb). We administered AAV9.trastuzumab or AAV9.2.10AmAb i.t. by ICV injection at least 21 days before tumor implantation. Untreated mice served as an additional control. We implanted HER2+ BT474-M1.ffluc cells into the brain parenchyma stereotaxically and monitored mice daily until moribund. Kaplan–Meier survival curves indicated that the median survival of mice that received i.t. AAV9.trastuzumab tumor prophylaxis was significantly greater than mice that received AAV9.2.10AmAb (Fig. 1A and B; 124 days vs. 46.5 days; P < 0.0001, hazard ratio = 0.0306). The median survival of mice that received AAV9.2.10AmAb was not significantly different from mice without treatment (Fig. 1A and B; 46.5 days vs. 50 days; P = 0.4306, hazard ratio = 1.1868). Protein A ELISA quantification of IgG transgene expression in serum (Fig. 1C and D) and brain tissue homogenate, normalized to total protein in brain homogenate (Fig. 1E), indicated that transgene expression was similar between groups that received AAV9.trastuzumab and AAV9.2.10AmAb. Biodistribution analysis of AAV9 vector GCs in brain tissue demonstrated equivalent genome deposition between groups that received vector (Supplementary Fig. S2A). Tumors remained HER2+ at the time of necropsy (Supplementary Fig. S2B).

Figure 1.

Intrathecal AAV9.trastuzumab tumor prophylaxis. A, Kaplan–Meier survival curves for mice that received i.t. AAV9.trastuzumab tumor prophylaxis, i.t. AAV9.2.10AmAb-negative control treatment, or no treatment at least 21 days before tumor challenge. B, Tabulated survival and statistical data for A. C and D, Serum expression of trastuzumab and 2.10A mAb measured by protein A ELISA. E, IgG transgene expression in brain tissue measured by protein A ELISA of brain homogenate and graphed as a percentage of total protein in brain homogenate.

Figure 1.

Intrathecal AAV9.trastuzumab tumor prophylaxis. A, Kaplan–Meier survival curves for mice that received i.t. AAV9.trastuzumab tumor prophylaxis, i.t. AAV9.2.10AmAb-negative control treatment, or no treatment at least 21 days before tumor challenge. B, Tabulated survival and statistical data for A. C and D, Serum expression of trastuzumab and 2.10A mAb measured by protein A ELISA. E, IgG transgene expression in brain tissue measured by protein A ELISA of brain homogenate and graphed as a percentage of total protein in brain homogenate.

Close modal

To determine if i.t. AAV9.trastuzumab can serve as a treatment for existing HER2+ BCBM, we implanted HER2+ BT474-M1.ffluc tumors into Rag1−/− mice and administered i.t. AAV9.trastuzumab or no treatment 3 days after tumor implantation. Kaplan–Meier survival curves indicated that the median survival of mice that received AAV9.trastuzumab was significantly greater than mice that received no treatment (Supplementary Fig. S3A and S3B; 82 days vs. 61 days; P = 0.002).

Dose escalation of AAV9var.trastuzumab leads to smaller day 35 tumor volume

We next performed a dose-escalation tumor prophylaxis study using a closely related variant of AAV9 that performs similarly in vivo. We administered 4 doses of AAV9var.trastuzumab 2 weeks before tumor implantation. A control group received no treatment. At day 35 after tumor implantation, we harvested brains and carefully dissected tumors from surrounding tissue. We measured the tumors in 3 dimensions with digital Vernier calipers and calculated the tumor volume. The mean volume of tumors decreased as the dose of AAV9var.trastuzumab increased, and mice that received no treatment had the largest mean tumor volume (Fig. 2A and B).

Figure 2.

Day 35 tumor volume following AAV9var.trastuzumab tumor prophylaxis. A, Tumor volumes from mice that received the indicated doses of AAV9var.trastuzumab tumor prophylaxis or no treatment. Tumor volume was measured 35 days after tumor implantation. B, Tabulated results and statistical data for A.

Figure 2.

Day 35 tumor volume following AAV9var.trastuzumab tumor prophylaxis. A, Tumor volumes from mice that received the indicated doses of AAV9var.trastuzumab tumor prophylaxis or no treatment. Tumor volume was measured 35 days after tumor implantation. B, Tabulated results and statistical data for A.

Close modal

Secreted trastuzumab binds to HER2+ brain tumors

We next used anti-IgG immunofluorescence of day 47 tumor cryosections to determine if secreted trastuzumab could bind to HER2+ tumor cells in vivo. Micrographs showed trastuzumab decorating tumors from mice that received i.t. AAV9.trastuzumab but not tumors from mice that received i.t. AAV9.2.10AmAb or no treatment (Fig. 3; component channels and merge are shown in Supplementary Fig. S4). Additionally, neurons transduced by AAV9.trastuzumab or AAV9.2.10AmAb stained positive for IgG. Secreted trastuzumab also decorated day 20 tumors from mice that received AAV9.trastuzumab (Supplementary Fig. S5).

Figure 3.

IgG immunofluorescence of day 47 tumor cryosections. Green channel (human/rhesus IgG) overexposures of day 47 tumors from mice that prophylactically received i.t. AAV9.trastuzumab, i.t. AAV9.2.10AmAb, or no treatment. Staining of tumors from mice that received AAV9.trastuzumab and AAV9.2.10AmAb shows neurons expressing IgG (arrows) and tumors decorated with trastuzumab. Tumors from mice that received AAV9.2.10AmAb were not decorated by IgG. DAPI (blue), tumor (T), and brain parenchyma (B).

Figure 3.

IgG immunofluorescence of day 47 tumor cryosections. Green channel (human/rhesus IgG) overexposures of day 47 tumors from mice that prophylactically received i.t. AAV9.trastuzumab, i.t. AAV9.2.10AmAb, or no treatment. Staining of tumors from mice that received AAV9.trastuzumab and AAV9.2.10AmAb shows neurons expressing IgG (arrows) and tumors decorated with trastuzumab. Tumors from mice that received AAV9.2.10AmAb were not decorated by IgG. DAPI (blue), tumor (T), and brain parenchyma (B).

Close modal

NK cells mediate i.t. AAV9.trastuzumab tumor prophylaxis

Trastuzumab exerts its effect against peripheral tumors mainly by facilitating antibody-dependent cell-mediated cytotoxicity (ADCC) of HER2+ tumor cells (28–30). We hypothesized that the same mechanism would govern AAV9.trastuzumab tumor prophylaxis in the CNS. We first evaluated AAV9.trastuzumab tumor prophylaxis in NSG mice, which lack NK cells and functional macrophages, both of which can perform antibody-dependent tumor cytotoxicity. As expected, AAV9.trastuzumab tumor prophylaxis failed in NSG mice (Supplementary Fig. S6A–S6C). Survival was comparable between groups that received no treatment (37 days) or AAV9.trastuzumab (40 days, P = 0.862). Transgene expression in brain tissue was similar among mice that received vector.

We next sought to determine if NK cells or macrophages played a larger role in i.t. AAV9.trastuzumab tumor prophylaxis in the CNS. To do so, we first used continuous NK cell depletion by i.p. administration of the anti-NK1.1 antibody PK136 in our Rag1−/− xenograft tumor model (Fig. 4A and B). Mice that received AAV9.2.10AmAb with or without NK cell depletion survived a median of 53 and 50 days, respectively. Mice that received AAV9.trastuzumab tumor prophylaxis and no NK cell depletion lived a median of 156 days. The median survival of mice that received AAV9.trastuzumab tumor prophylaxis with continuous NK cell depletion was significantly shorter (73 days; P < 0.0001, compared with AAV9.trastuzumab without NK cell depletion).

Figure 4.

Intrathecal AAV9.trastuzumab tumor prophylaxis in the setting of continuous NK cell depletion. A, Kaplan–Meier survival curves for mice that received i.t. AAV9.trastuzumab tumor prophylaxis or i.t. AAV9.2.10AmAb control treatment with or without continuous NK cell depletion using PK136 (anti-NK1.1 antibody). B, Tabulated survival and statistical data for Fig. 5A. C,In situ hybridization for Ncr1 (NKp46) RNA (red) in formalin-fixed paraffin-embedded tumors harvested at necropsy from the experiment in Fig. 5A. Larger images of the AAV9.2.10AmAb-treated tumors are shown in Supplementary Fig. S7. DAPI (blue), tumor (T), and brain parenchyma (B).

Figure 4.

Intrathecal AAV9.trastuzumab tumor prophylaxis in the setting of continuous NK cell depletion. A, Kaplan–Meier survival curves for mice that received i.t. AAV9.trastuzumab tumor prophylaxis or i.t. AAV9.2.10AmAb control treatment with or without continuous NK cell depletion using PK136 (anti-NK1.1 antibody). B, Tabulated survival and statistical data for Fig. 5A. C,In situ hybridization for Ncr1 (NKp46) RNA (red) in formalin-fixed paraffin-embedded tumors harvested at necropsy from the experiment in Fig. 5A. Larger images of the AAV9.2.10AmAb-treated tumors are shown in Supplementary Fig. S7. DAPI (blue), tumor (T), and brain parenchyma (B).

Close modal

We subjected the formalin-fixed tumors from mice in the NK cell depletion experiment to ISH for NK cell–specific Ncr1 (NKp46) RNA. Micrographs indicate that NK cells infiltrated only tumors from mice given AAV9.trastuzumab without NK cell depletion (Fig. 4C; Supplementary Fig. S7). IgG transgene expression in brain tissue homogenate was comparable between groups that received vector (Supplementary Fig. S8A). Ncr1 ISH of spleens confirmed successful depletion of NK cells (Supplementary Fig. S8B). We also cryosectioned day 20 tumors from mice that received prophylactic AAV9.trastuzumab, AAV9.2.10AmAb, or no treatment; this experiment was conducted without NK cell depletion. IHC staining for Ly-49G2 confirmed that NK cells infiltrated day 20 tumors only from mice that received AAV9.trastuzumab tumor prophylaxis (Fig. 5).

Figure 5.

NK cell IHC staining of day 20 tumors. Mice received AAV9.trastuzumab, AAV9.2.10AmAb, or no treatment 21 days before tumor implantation. Tumors were harvested 21 days after implantation, and IHC staining for NK cells (Ly-49G2) was performed on sample cryosections. Tumor (T), brain parenchyma (B), and necrotic tumor (N).

Figure 5.

NK cell IHC staining of day 20 tumors. Mice received AAV9.trastuzumab, AAV9.2.10AmAb, or no treatment 21 days before tumor implantation. Tumors were harvested 21 days after implantation, and IHC staining for NK cells (Ly-49G2) was performed on sample cryosections. Tumor (T), brain parenchyma (B), and necrotic tumor (N).

Close modal

To determine if systemically circulating macrophages play a role in AAV9.trastuzumab tumor prophylaxis in our model, we used continuous macrophage depletion by i.p. administration of clodronate liposomes. Mice that received no treatment with or without macrophage depletion survived a median of 53 and 50 days, respectively (Supplementary Fig. S9A and S9B). Mice that received AAV9.trastuzumab without macrophage depletion survived a median of 92 days (P ≤ 0.0001 compared with mice that received no treatment without macrophage depletion). Mice that received AAV9.trastuzumab with macrophage depletion survived a median of 70.5 days (P = 0.3268, compared with AAV9.trastuzumab without macrophage depletion). IgG transgene expression in brain tissue was comparable among groups that received vector (Supplementary Fig. S9C), and macrophage depletion was confirmed by IHC staining of spleens for CD68 (Supplementary Fig. S9D).

Comparison of trastuzumab levels in brain after administration of i.t. Herceptin or i.t. AAV9.trastuzumab

To compare the amount of trastuzumab in brain tissue after ICV administration of AAV9.trastuzumab versus Herceptin, we administered 1e11 GC AAV9.trastuzumab or 15 μg Herceptin ICV to Rag1−/− mice. We harvested brains from mice that received AAV9.trastuzumab 2 weeks after vector administration, and brains that received Herceptin 24, 48, and 168 hours after administration. Figure 6A indicates that the median amount of trastuzumab expressed in brain tissue in mice that received AAV9.trastuzumab is equivalent to the amount of Herceptin in brain tissue 24 hours after administration of 15 μg antibody. Over time, the amount of Herceptin in brain tissue decreased, as expected. We calculated the half-life of Herceptin in mouse brain tissue to be 42 hours between the 24- and 48-hour intervals and 46 hours between the 48- and 168-hour intervals.

Figure 6.

Trastuzumab quantitation in the CNS of mice and a cynomolgus macaque. A, Trastuzumab quantified in brain tissue at the time points indicated following either a 15 μg ICV bolus of Herceptin or 2 weeks after ICV AAV9.trastuzumab vector administration (1e11 GC/mouse). B, Expression of 201IA in the CSF of a cynomolgus macaque that received AAV9.201IA by intracisterna vector administration (1e12 GC/kg).

Figure 6.

Trastuzumab quantitation in the CNS of mice and a cynomolgus macaque. A, Trastuzumab quantified in brain tissue at the time points indicated following either a 15 μg ICV bolus of Herceptin or 2 weeks after ICV AAV9.trastuzumab vector administration (1e11 GC/mouse). B, Expression of 201IA in the CSF of a cynomolgus macaque that received AAV9.201IA by intracisterna vector administration (1e12 GC/kg).

Close modal

Vector-mediated antibody expression in macaque CSF

To examine the safety of using an AAV9 vector to express an antibody in the CNS of a nonhuman primate, we delivered 1 × 1012 GC/kg of AAV9.201IA i.t. into the cisterna magna of a female cynomolgus macaque. We collected CSF at regular intervals over the next 6 months and quantified 201IA expression in CSF by ELISA. Results in Fig. 6B show a peak concentration of 201IA in CSF of 0.350 μg/mL at day 48, which plateaued to 0.165 μg/mL by day 77 and 0.153 μg/mL by the end of the study (day 168). Importantly, no serious adverse events occurred after administration of vector.

Metastasis of HER2+ breast cancer to the brain is a devastating diagnosis with a poor prognosis due to a lack of targeted and effective treatments. Although i.t. administration of trastuzumab has been reported to slow disease progression and prolong survival, the benefit is modest, likely due to the rapid turnover of CSF reported by others, resulting in poor tumor exposure to antibody. An unmet need for effective treatments exists for patients with this disease.

The advent of AAV vectors for gene transfer has revolutionized the field of gene therapy. The discovery of CNS-tropic vectors, such as AAV9, represents a boon for clinicians seeking to localize biological treatments behind the blood–brain barrier. A single i.t. administration of AAV9 vector into the CSF by way of the lateral ventricles or cisterna magna leads to widespread transduction of neurons and astrocytes throughout the cortex, cerebellum, and spinal cord, resulting in long-lived production of transgene product in the CNS compartment (26, 27, 31).

We successfully demonstrated that AAV9.trastuzumab administered i.t. at a moderate vector dose, both as tumor prophylaxis and as tumor treatment, significantly extends median survival in a xenograft mouse model in which the human HER2+ cancer cell line BT747-M1.ffluc is implanted into the brains of Rag1−/− mice. In the prophylactic setting, the median survival was 2.48 times greater for mice that received AAV9.trastuzumab tumor prophylaxis than for mice that received control AAV vector treatment or no treatment. Additionally, dose-escalation studies using AAV9var.trastuzumab tumor prophylaxis indicated that higher doses of vector led to smaller tumors 35 days after implantation. In the treatment setting, the median survival benefit from AAV9.trastuzumab administration was significant but less, perhaps in part due to the shorter total duration of exposure of tumor cells to expressed trastuzumab.

Our studies also indicate that trastuzumab secreted by neurons and astrocytes maintains its antigen specificity and effector function despite the fact that these cells do not normally produce IgG. Using immunofluorescence microscopy, we showed that HER2+ tumors from mice treated with AAV9.trastuzumab, but not AAV9.2.10AmAb, are decorated with expressed IgG.

The effector functions of trastuzumab have been widely studied (28, 29). Trastuzumab bound to HER2 on tumor cells (i) mediates ADCC when its Fc binds to activating Fcγ receptors (e.g., FcγRIIIa) on immune cells such as NK cells and macrophages; (ii) prevents initiation of signal transduction when HER2 homodimerizes or heterodimerizes with other members of the HER family, thus slowing tumor growth; (iii) blocks proteolytic cleavage of the HER2 ectodomain, preventing the remaining membrane-bound p96 protein from activating progrowth signaling pathways; (iv) disrupts proangiogenic pathways; and (v) disrupts DNA repair pathways.

The predominant mechanism by which trastuzumab acts against peripheral HER2+ tumors in patients is ADCC by immune cells carrying Fcγ receptors, such as NK cells and macrophages. We hypothesized that the same would be the case in our xenograft model. AAV9.trastuzumab tumor prophylaxis failed when administered to NSG mice, which lack NK cells and functional macrophages, compared with Rag1−/− mice; this result suggests an NK cell- or macrophage-mediated mechanism of tumor prophylaxis. When we depleted NK cells in the Rag1−/− model, the median survival of mice decreased substantially, indicating that NK cell–driven ADCC of tumor cells is responsible for a majority of the survival benefit. Additionally, immunofluorescence and IHC staining of NK cells in tumors indicated that NK cells infiltrated only into tumors of mice that received AAV9.trastuzumab tumor prophylaxis, not in tumors from untreated or control-treated mice. These NK cells were present in tumors as early as 20 days after tumor implantation and as late as the time of necropsy. To our knowledge, this is the first report revealing the predominant mechanism by which trastuzumab works against HER2+ tumors in the CNS.

Our macrophage depletion studies indicated that these cells do not contribute significantly to the antitumor effect of AAV9.trastuzumab tumor prophylaxis in this model. However, as clodronate liposomes do not cross the blood–brain barrier, these experiments can only address whether systemically circulating macrophages play a role in AAV9.trastuzumab tumor prophylaxis. Tissue-resident macrophages/microglia may still contribute a small part to the antitumor effect of AAV9.trastuzumab in this model.

We were unable to administer i.t. Herceptin as a positive control in our experiments due to Institutional Animal Care and Use Committee restrictions, which prohibit us from performing more than a single ICV injection during the lifetime of a mouse. In lieu of this, we report that the half-life of Herceptin in the brain tissue of Rag1−/− mice is around 42 to 46 hours. We also showed that the median amount of trastuzumab in brains of mice 2 weeks after AAV9 vector administration was equivalent to the amount in brains 24 hours after receiving 15 μg by ICV injection. Finally, we were able to demonstrate safe, constant expression of an antibody in the CNS of a nonhuman primate for 6 months after ICM administration of AAV9 vector.

The BT474.M1 cell line highly overexpresses HER2, meaning our results likely overestimate the response to our therapy in humans given the heterogeneity of HER2 expression both within tumors and among different HER2+ BCBM in patients. Although using other cell lines in this model may shed light on the expected efficacy of our treatment in humans, our results illustrate proof of concept that AAV9.trastuzumab can reduce tumor burden and increase survival when administered to mice with a HER2+ breast cancer brain tumor. It is also important to note that patients whose tumors become resistant to trastuzumab treatment would likely not benefit from our therapy. This will be an important consideration when choosing a target population for clinical studies and treatment.

Additionally, our mice bore one brain tumor, but patients often develop multiple metastases. Testing AAV9.trastuzumab tumor prophylaxis in a model where tumor cells are administered via carotid artery injection would indicate whether our treatment is effective against multiple brain tumors. We also used AAV9.trastuzumab in these studies as a monotherapy. Unsurprisingly, all mice eventually succumbed to outgrowth of tumors, which remain HER2+ at necropsy. Combining AAV.trastuzumab with chemotherapy and/or radiotherapy will likely be synergistic against HER2+ BCBM, both in this xenograft model and in patients.

Trastuzumab infusion is standard of care for patients with HER2-overexpressing breast cancer, and i.t. delivery of trastuzumab has been shown to be beneficial. However, trastuzumab expression cannot be turned off after AAV9-mediated gene transfer, although there are efforts in the gene therapy field to develop an inducible expression system. Related to this, cardiotoxicity is a known side effect that occurs in a small subset of patients who receive systemic trastuzumab treatment. Risk factors for cardiotoxicity have been identified, including previous anthracycline-containing chemotherapeutic regimens, age, and existing cardiac dysfunction (32–34).

After i.t. administration of i.t. AAV9.trastuzumab, we anticipate a lower concentration of trastuzumab in blood than after i.v. administration of Herceptin. This would be less likely to lead to cardiotoxicity. Regardless, women with CNS metastasis from HER2+ breast cancer almost always have systemic disease, necessitating treatment with systemic administration of trastuzumab where safety monitoring of cardiac function is standard. We will incorporate this aspect into the design of our first-in-human phase I safety trial.

Given this fact and the results of our experiments, we intend to move AAV9.trastuzumab toward the clinic after successful completion of necessary preclinical studies, including safety and toxicology experiments in large animal models. We plan to first assess the safety, efficacy, and pharmacokinetics of AAV9.trastuzumab in women with documented CNS lesions from HER2+ breast cancer. Subject to safety and efficacy in this setting, we will subsequently assess AAV9.trastuzumab in a similar patient population with systemically metastatic HER2+ disease prior to clinical or radiologic diagnosis of CNS lesions. Given that AAV transgene expression has been documented to persist for years in nonhuman primates and humans (35, 36), this approach has the potential to be an integral part of the adjuvant therapy for patients with early diagnosis of HER2+ breast cancer after curative resection of the breast lesion. Indeed, our AAV platform has the potential to be expanded to express other therapeutic immunotherapies behind the blood–brain barrier to treat CNS diseases and address other unmet therapeutic needs.

No potential conflicts of interest were disclosed.

Conception and design: W.T. Rothwell, P. Bell, M.P. Limberis, M. Li, J.M. Wilson

Development of methodology: W.T. Rothwell, P. Bell, L.K. Richman, M.P. Limberis, M. Li

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

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): W.T. Rothwell, P. Bell, J.M. Wilson

Writing, review, and/or revision of the manuscript: W.T. Rothwell, P. Bell, M.P. Limberis, J.M. Wilson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.K. Richman, A.P. Tretiakova, M. Li

Study supervision: W.T. Rothwell, J.M. Wilson

Other (oversaw histology work): P. Bell

We would like to thank Lewis Chodosh and Jason Ruth for their generous gift of BT474.M1 cells; Lewis Chodosh, José Conejo-Garcia, Chi Van Dang, and Laura A. Johnson for their guidance and advice; Tamara Goode for procedural instruction and advice with the xenograft model; Deirdre McMenamin, Christine Draper, and the Penn Gene Therapy Program (GTP) Program for Comparative Medicine for procedural assistance; Jamunabai Prakash and the Penn GTP Cell Morphology Core for assistance with histopathology; the Penn Vector Core; Christian Hinderer for advice and direction; Jenny Greig for her generous gift of AAV9.2.10AmAb and for advice and direction; and Tarek Sahmoud for his assistance with the clinical development plan.

This work was funded by the University of Pennsylvania Gene Therapy Program.

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.

1.
Sakai
N
,
Vaisman
BL
,
Koch
CA
,
Hoyt
RF
 Jr
,
Meyn
SM
,
Talley
GD
, et al
Targeted disruption of the mouse lecithin:cholesterol acyltransferase (LCAT) gene. Generation of a new animal model for human LCAT deficiency
.
J Biol Chem
1997
;
272
:
7506
10
.
2.
Koo
T
,
Kim
IA
. 
Brain metastasis in human epidermal growth factor receptor 2-positive breast cancer: from biology to treatment
.
Radiat Oncol J
2016
;
34
:
1
9
.
3.
Lin
NU
,
Bellon
JR
,
Winer
EP
. 
CNS metastases in breast cancer
.
J Clin Oncol
2004
;
22
:
3608
17
.
4.
Kennecke
H
,
Yerushalmi
R
,
Woods
R
,
Cheang
MC
,
Voduc
D
,
Speers
CH
, et al
Metastatic behavior of breast cancer subtypes
.
J Clin Oncol
2010
;
28
:
3271
7
.
5.
Bendell
JC
,
Domchek
SM
,
Burstein
HJ
,
Harris
L
,
Younger
J
,
Kuter
I
, et al
Central nervous system metastases in women who receive trastuzumab-based therapy for metastatic breast carcinoma
.
Cancer
2003
;
97
:
2972
7
.
6.
Clayton
AJ
,
Danson
S
,
Jolly
S
,
Ryder
WD
,
Burt
PA
,
Stewart
AL
, et al
Incidence of cerebral metastases in patients treated with trastuzumab for metastatic breast cancer
.
Br J Cancer
2004
;
91
:
639
43
.
7.
Gori
S
,
Rimondini
S
,
De Angelis
V
,
Colozza
M
,
Bisagni
G
,
Moretti
G
, et al
Central nervous system metastases in HER-2 positive metastatic breast cancer patients treated with trastuzumab: incidence, survival, and risk factors
.
Oncologist
2007
;
12
:
766
73
.
8.
Stemmler
HJ
,
Kahlert
S
,
Siekiera
W
,
Untch
M
,
Heinrich
B
,
Heinemann
V
. 
Characteristics of patients with brain metastases receiving trastuzumab for HER2 overexpressing metastatic breast cancer
.
Breast
2006
;
15
:
219
25
.
9.
Brufsky
AM
,
Mayer
M
,
Rugo
HS
,
Kaufman
PA
,
Tan-Chiu
E
,
Tripathy
D
, et al
Central nervous system metastases in patients with HER2-positive metastatic breast cancer: incidence, treatment, and survival in patients from registHER
.
Clin Cancer Res
2011
;
17
:
4834
43
.
10.
Mehta
AI
,
Brufsky
AM
,
Sampson
JH
. 
Therapeutic approaches for HER2-positive brain metastases: circumventing the blood-brain barrier
.
Cancer Treat Rev
2013
;
39
:
261
9
.
11.
Forsyth
PA
,
Weaver
S
,
Fulton
D
,
Brasher
PM
,
Sutherland
G
,
Stewart
D
, et al
Prophylactic anticonvulsants in patients with brain tumour
.
Can J Neurol Sci
2003
;
30
:
106
12
.
12.
Glantz
MJ
,
Cole
BF
,
Friedberg
MH
,
Lathi
E
,
Choy
H
,
Furie
K
, et al
A randomized, blinded, placebo-controlled trial of divalproex sodium prophylaxis in adults with newly diagnosed brain tumors
.
Neurology
1996
;
46
:
985
91
.
13.
Park
IH
,
Ro
J
,
Lee
KS
,
Nam
BH
,
Kwon
Y
,
Shin
KH
. 
Trastuzumab treatment beyond brain progression in HER2-positive metastatic breast cancer
.
Ann Oncol
2009
;
20
:
56
62
.
14.
Ramakrishna
N
,
Temin
S
,
Chandarlapaty
S
,
Crews
JR
,
Davidson
NE
,
Esteva
FJ
, et al
Recommendations on disease management for patients with advanced human epidermal growth factor receptor 2-positive breast cancer and brain metastases: American Society of Clinical Oncology clinical practice guideline
.
J Clin Oncol
2014
;
32
:
2100
8
.
15.
Pestalozzi
BC
,
Brignoli
S
. 
Trastuzumab in CSF
.
J Clin Oncol
2000
;
18
:
2349
51
.
16.
Stemmler
HJ
,
Schmitt
M
,
Willems
A
,
Bernhard
H
,
Harbeck
N
,
Heinemann
V
. 
Ratio of trastuzumab levels in serum and cerebrospinal fluid is altered in HER2-positive breast cancer patients with brain metastases and impairment of blood-brain barrier
.
Anticancer Drugs
2007
;
18
:
23
8
.
17.
Sperduto
PW
,
Kased
N
,
Roberge
D
,
Xu
Z
,
Shanley
R
,
Luo
X
, et al
Summary report on the graded prognostic assessment: an accurate and facile diagnosis-specific tool to estimate survival for patients with brain metastases
.
J Clin Oncol
2012
;
30
:
419
25
.
18.
Leone
JP
,
Leone
J
,
Zwenger
AO
,
Iturbe
J
,
Leone
BA
,
Vallejo
CT
. 
Prognostic factors and survival according to tumour subtype in women presenting with breast cancer brain metastases at initial diagnosis
.
Eur J Cancer
2017
;
74
:
17
25
.
19.
Slamon
DJ
,
Leyland-Jones
B
,
Shak
S
,
Fuchs
H
,
Paton
V
,
Bajamonde
A
, et al
Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2
.
N Engl J Med
2001
;
344
:
783
92
.
20.
Cobleigh
MA
,
Vogel
CL
,
Tripathy
D
,
Robert
NJ
,
Scholl
S
,
Fehrenbacher
L
, et al
Multinational study of the efficacy and safety of humanized anti-HER2 monoclonal antibody in women who have HER2-overexpressing metastatic breast cancer that has progressed after chemotherapy for metastatic disease
.
J Clin Oncol
1999
;
17
:
2639
48
.
21.
Bousquet
G
,
Darrouzain
F
,
de Bazelaire
C
,
Ternant
D
,
Barranger
E
,
Winterman
S
, et al
Intrathecal trastuzumab halts progression of CNS metastases in breast cancer
.
J Clin Oncol
2016
;
34
:
e151
5
.
22.
Colozza
M
,
Minenza
E
,
Gori
S
,
Fenocchio
D
,
Paolucci
C
,
Aristei
C
, et al
Extended survival of a HER-2-positive metastatic breast cancer patient with brain metastases also treated with intrathecal trastuzumab
.
Cancer Chemother Pharmacol
2009
;
63
:
1157
9
.
23.
Zagouri
F
,
Sergentanis
TN
,
Bartsch
R
,
Berghoff
AS
,
Chrysikos
D
,
de Azambuja
E
, et al
Intrathecal administration of trastuzumab for the treatment of meningeal carcinomatosis in HER2-positive metastatic breast cancer: a systematic review and pooled analysis
.
Breast Cancer Res Treat
2013
;
139
:
13
22
.
24.
Bonneau
C
,
Paintaud
G
,
Tredan
O
,
Dubot
C
,
Desvignes
C
,
Dieras
V
, et al
Phase I feasibility study for intrathecal administration of trastuzumab in patients with HER2 positive breast carcinomatous meningitis
.
Eur J Cancer
2018
;
95
:
75
84
.
25.
Bergman
I
,
Burckart
GJ
,
Pohl
CR
,
Venkataramanan
R
,
Barmada
MA
,
Griffin
JA
, et al
Pharmacokinetics of IgG and IgM anti-ganglioside antibodies in rats and monkeys after intrathecal administration
.
J Pharmacol Exp Ther
1998
;
284
:
111
5
.
26.
Cearley
CN
,
Wolfe
JH
. 
Transduction characteristics of adeno-associated virus vectors expressing cap serotypes 7, 8, 9, and Rh10 in the mouse brain
.
Mol Ther
2006
;
13
:
528
37
.
27.
Hinderer
C
,
Bell
P
,
Vite
CH
,
Louboutin
JP
,
Grant
R
,
Bote
E
, et al
Widespread gene transfer in the central nervous system of cynomolgus macaques following delivery of AAV9 into the cisterna magna
.
Mol Ther Methods Clin Dev
2014
;
1
:
14051
.
28.
Spector
NL
,
Blackwell
KL
. 
Understanding the mechanisms behind trastuzumab therapy for human epidermal growth factor receptor 2-positive breast cancer
.
J Clin Oncol
2009
;
27
:
5838
47
.
29.
Shi
Y
,
Fan
X
,
Deng
H
,
Brezski
RJ
,
Rycyzyn
M
,
Jordan
RE
, et al
Trastuzumab triggers phagocytic killing of high HER2 cancer cells in vitro and in vivo by interaction with Fcgamma receptors on macrophages
.
J Immunol
2015
;
194
:
4379
86
.
30.
Clynes
RA
,
Towers
TL
,
Presta
LG
,
Ravetch
JV
. 
Inhibitory Fc receptors modulate in vivo cytotoxicity against tumor targets
.
Nat Med
2000
;
6
:
443
6
.
31.
Bell
CL
,
Vandenberghe
LH
,
Bell
P
,
Limberis
MP
,
Gao
GP
,
Van Vliet
K
, et al
The AAV9 receptor and its modification to improve in vivo lung gene transfer in mice
.
J Clin Invest
2011
;
121
:
2427
35
.
32.
Mayer
EL
,
Lin
NU
. 
Long Term Follow-Up of National Surgical Adjuvant Breast and Bowel Project Trial B-31: How Well Can We Predict Cardiac Toxicity With Trastuzumab?
J Clin Oncol
2012
;
30
:
3
.
33.
de Azambuja
E
,
Procter
MJ
,
van Veldhuisen
DJ
,
Agbor-Tarh
D
,
Metzger-Filho
O
,
Steinseifer
J
, et al
Trastuzumab-associated cardiac events at 8 years of median follow-up in the Herceptin Adjuvant trial (BIG 1-01)
.
J Clin Oncol
2014
;
32
:
2159
65
.
34.
Mantarro
S
,
Rossi
M
,
Bonifazi
M
,
D'Amico
R
,
Blandizzi
C
,
La Vecchia
C
, et al
Risk of severe cardiotoxicity following treatment with trastuzumab: a meta-analysis of randomized and cohort studies of 29,000 women with breast cancer
.
Intern Emerg Med
2016
;
11
:
123
40
.
35.
Rivera
VM
,
Gao
GP
,
Grant
RL
,
Schnell
MA
,
Zoltick
PW
,
Rozamus
LW
, et al
Long-term pharmacologically regulated expression of erythropoietin in primates following AAV-mediated gene transfer
.
Blood
2005
;
105
:
1424
30
.
36.
Nathwani
AC
,
Reiss
UM
,
Tuddenham
EG
,
Rosales
C
,
Chowdary
P
,
McIntosh
J
, et al
Long-term safety and efficacy of factor IX gene therapy in hemophilia B
.
N Engl J Med
2014
;
371
:
1994
2004
.