Molecular Genetics and Metabolism 109 (2013) 339–344

Contents lists available at SciVerse ScienceDirect

Molecular Genetics and Metabolism

j ourna l homepage: www.e lsev ie r .com/ locate /ymgme

Erythrocytes encapsulated with phenylalanine hydroxylase exhibit improvedpharmacokinetics and lowered plasma phenylalanine levels in normal mice

Nelson S. Yew a,⁎, Emmanuelle Dufour b, Malgorzata Przybylska a, Julie Putelat b, Cristin Crawley a,Meta Foster a, Sarah Gentry a, David Reczek a, Alla Kloss a, Aurélien Meyzaud b, Françoise Horand b,Seng H. Cheng a, Yann Godfrin b

a Genzyme, a Sanofi Company, 49 New York Avenue, Framingham, MA 01701-9322, USAb ERYtech Pharma, Batiment Adenine, 60, Avenue Rockefeller, 69008 Lyon, France

⁎ Corresponding author.E-mail address: nelson.yew@genzyme.com (N.S. Yew

1096-7192/$ – see front matter © 2013 Elsevier Inc. Allhttp://dx.doi.org/10.1016/j.ymgme.2013.05.011

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 March 2013Received in revised form 15 May 2013Accepted 15 May 2013Available online 16 July 2013

Keywords:Enzyme therapyErythrocytesDrug deliveryPharmacokineticsPhenylalaninePhenylketonuria

Enzyme replacement therapy is often hampered by the rapid clearance and degradation of the administeredenzyme, limiting its efficacy and requiring frequent dosing. Encapsulation of therapeutic molecules into redblood cells (RBCs) is a clinically proven approach to improve the pharmacokinetics and efficacy of biologicsand small molecule drugs. Here we evaluated the ability of RBCs encapsulated with phenylalanine hydroxy-lase (PAH) to metabolize phenylalanine (Phe) from the blood and confer sustained enzymatic activity in thecirculation. Significant quantities of PAH were successfully encapsulated within murine RBCs (PAH-RBCs)with minimal loss of endogenous hemoglobin. While intravenously administered free PAH enzyme was rap-idly eliminated from the blood within a few hours, PAH-RBCs persisted in the circulation for at least 10 days.A single injection of PAH-RBCs was able to decrease Phe levels by nearly 80% in normal mice. These resultsdemonstrate the ability of enzyme-loaded RBCs tometabolize circulating amino acids and highlight the potentialto treat disorders of amino acid metabolism.

© 2013 Elsevier Inc. All rights reserved.

1. Introduction

Red blood cells (RBCs) have long been of interest as drug carriersand delivery vehicles, possessing several desirable features for this pur-pose, including complete biocompatibility, long lifespan (120 days inhumans), and natural degradation and elimination by the body [1–4].Purified populations of RBCs can be readily obtained in large quantities,and the methods for encapsulating small molecule drugs or proteinsare well developed [5]. As a delivery vehicle, encapsulated RBCs canbemodified to be efficiently taken up by cells of the reticuloendothelialsystem, delivering anti-inflammatory drugs, anti-cancer agents or anti-gens [1]. Alternatively, enzyme-loaded RBCs can serve as circulatingbioreactors that can continuously detoxify toxins or excess metabolitespresent in the blood [4]. In this regard, channels and transporterspresent in the erythrocyte membrane allow the transport of ions,amino acids, and various small molecules into and out of the cell[6]. Importantly, encapsulated proteins considered foreign by theimmune system are shielded from any neutralizing antibodies thatmay be present in the circulation.

).

rights reserved.

Elevated or altered levels of amino acids in blood are diagnostic forsome inborn metabolic disorders. One example is phenylketonuria(PKU), an autosomal recessive disorder caused by a deficiency in phe-nylalanine hydroxylase (PAH), which metabolizes phenylalanine totyrosine in the presence of molecular oxygen, iron and the essentialcofactor tetrahydrobiopterin (BH4) [7,8]. If untreated, Phe levels inthe blood rise to extremely high levels, leading to neurocognitive de-cline and severe mental retardation. Genetic screening now identifiesnewborns with PKU who are then immediately placed on a severePhe-restricted diet [9]. However, diet compliance is poor and thereremains a significant unmet medical need for this disorder [10].

To begin to develop therapies for PKU, we have evaluated the use ofrecombinant PAH encapsulated in erythrocytes in mice. The encapsula-tion process involves reversible hypotonic swelling of the RBCs thattransiently opens pores in the membrane, allowing drugs to enter thecells [5,11]. Upon return to isotonic medium, the pores are closed andthe drug is entrapped. While the procedure is highly reproducible, theinherent properties of the encapsulated protein are critical, with effica-cy dependent on the stability and specific activity of the enzyme. Mam-malian PAH is a large multimeric enzyme (200 kD as a tetramer) withrelatively low specific activity [12]. We therefore chose to evaluatePAH from the cyanobacterium Chromobacterium violaceum, which is a33 kDmonomer that is more thermally stable and reportedly more ac-tive than the mammalian enzyme [13].

340 N.S. Yew et al. / Molecular Genetics and Metabolism 109 (2013) 339–344

2. Methods

2.1. Expression and purification of recombinant C. violaceum PAH

The sequence encoding C. violaceum PAH (GenBank accession#AF6711) (EC 1.14.16.1) was codon-optimized for expression inEscherichia coli and synthesized (DNA 2.0, Menlo Park, CA USA). Thesequence was cloned into pET29a (EMDMillipore, Darmstadt, Germany)so that the 6XHis tag sequence was added in-frame to the 3′ end of thePAH gene. The pET29a-PAH construct was used to transform BL21(DE3)star cells (Life Technologies, Carlsbad, CA USA), which were inducedwith 0.5 mM IPTG at 20 °C for 18 h. Cell pellets were resuspendedin lysis buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mMβ-mercaptoethanol, protease inhibitors). The cell suspension wassonicated then centrifuged at 17,000 ×g for 30 min. The supernatantwas decanted and FeSO4 was added to a final concentration of0.6 mM. The supernatant was stirred for 30 min at 4 °C then filteredthrough a 0.2 μm filter.

A 5 mL HisTrap FF column (GE Healthcare Life Sciences, Pittsburgh,PA USA) was equilibrated with 10 column volumes (CV) of equili-bration buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mMβ-mercaptoethanol, 5 mM imidazole). After loading the supernatant,the column was washed with 50 CV of Triton X-114 buffer (50 mMTris–HCl, pH 7.4, 150 mM NaCl, 5 mM β-mercaptoethanol, 0.1% TritonX-114) followed by 20 CV of equilibration buffer. The protein was theneluted with elution buffer (50 mM Tris–HCl, pH 7.4, 150 mM NaCl,5 mM β-mercaptoethanol, 300 mM imidazole). The protein containingfractions were pooled, buffer-exchanged into 50 mM sodium acetate,pH 6.2, concentrated by ultrafiltration (Amicon PD-10 columns) (EMDMillipore) and stored at −20 °C.

2.2. PAH enzyme assay

The activity of purified C. violaceum PAH was measured accordingto the procedure described by Nakata et al. [13] with some modifica-tions. The assay mixture contained 100 mM Tris–HCl, pH 7.5, 4 mMDTT, 4 mM L-phenylalanine, 100 μg bovine catalase, and 10–50 μg ofPAH enzyme in a volume of 250 μL. The reaction was initiated bythe addition of DMPH4 to a final concentration of 0.4 mM and wascarried out at 23 °C for 0.5–1 h with shaking. For PAH activity mea-surement of blood samples, 10–20 μL of washed RBCs was added tothe mixture and the reaction time increased to 2–6 h. The reactionwas stopped by the addition of 250 μL of 5% trichloroacetic acid. Tyro-sine was detected colorimetrically by adding 250 μL of 0.1% 1-nitroso-2-napthol in 100 mMNaOH and 250 μL of 20% (v/v) HNO3 with 0.05%(w/v) NaNO2. The mixture was incubated at 55 °C for 30 min, cooledto 25 °C, then centrifuged 3 min × 5000 rpm. The supernatant wastransferred to a microplate and read at 450 nm. Values were com-pared to tyrosine standards (0–280 μM).

2.3. Encapsulation of PAH in RBCs

The PAH enzyme was loaded into mouse RBCs by the method ofreversible hypotonic dialysis using an AN69 hollow fiber dialyzer(Gambro, Lyon, France) as described by Dufour et al. [14] with somemodifications. Whole blood from OF1 mice was centrifuged and theplasma was removed. The RBCs were washed three times with 0.9%(v/v) NaCl. Purified PAH was then added to the washed RBCs to afinal concentration of 1 mg/mL, resulting in a cell suspension with ahematocrit of approximately 70%. The RBCs plus PAH were dialyzedagainst a hypotonic buffer of 40 mOsmol/kg and then resealed withthe addition of a hypertonic solution (10% v/v) as described previously[15]. The suspension was then incubated at 37 °C for 30 min. After finalwashes, the final product was resuspended in SAG-Mannitol plus 6%bovine serum albumin to a hematocrit of 50%. The product was storedat 2–8 °C and was injected no more than 16 h after preparation.

As a control, processed control RBCs (CON-RBCs) were preparedby dialyzing RBCs as described above but without added PAH.

2.4. Estimation of PAH levels in PAH-RBCs and in blood

A 5 μL aliquot of the PAH-RBCs was serially diluted with phosphate-buffered saline (PBS) such that an equivalent of 5 nL of the PAH-RBCswas loaded onto an acrylamide gel. After electrophoresis and transferto nitrocellulose, the blot was probed with a rabbit polyclonal antibodyto PAH (1:10,000 dilution of a 1 mg/mL stock). The antibodywas gener-ated by immunizing rabbits with the purified C. violaceum PAH protein.A goat anti-rabbit secondary antibody (Santa Cruz Biotechnology,Dallas, TXUSA) followed by a chemiluminescent substrate (SuperSignalWest Pico chemiluminescent substrate, Thermo Scientific, Rockford, ILUSA) was used to detect the bands. For detecting levels of PAH in thecirculation, whole blood was centrifuged to pellet the RBCs, whichwere washed 2× with PBS. An equivalent of 0.5 μL of washed RBCswas loaded per lane.

2.5. Mice

OF1 mice, 6–8 weeks old, were purchased from Charles RiverLaboratories (Lyon, France). BTBR-PAHenu2/J mice were purchasedfrom Jackson Laboratories (Bar Harbor, ME, USA) and a colony wasmaintained in-house. The PAHenu2 animals were cared for in an AAALACaccredited facility in accordance with the guidelines established by theNational Research Council. Animals had access to food and water adlibitum.

2.6. In vivo studies

The PAH-RBCs or negative control processed RBCs were injectedintravenously (10–12 mL/kg) via the tail vein. Tetrahydrobiopterinpowder (Schircks Laboratories, Jona, Switzerland) was dissolved in aphosphate buffer, pH 7.4 immediately prior to intraperitoneal injec-tion (200 mg/kg) 30 min prior to bleeding the animals. Blood wascollected under anesthesia (3–5% isoflurane) via the orbital venousplexus. An aliquot of whole blood was frozen at −80 °C for enzymeactivity analysis. The remaining blood was centrifuged, plasma wascollected and frozen at −80 °C. The pelleted RBCs were washedtwice with phosphate-buffered saline (PBS), then resuspended to50% hematocrit in PBS and frozen at −80 °C.

2.7. Measurement of phenylalanine and BH4 levels in plasma

The levels of Phe and BH4 in plasma were analyzed by UPLC–MS/MS, using an Acquity UPLC (Waters Corporation, Milford, MA USA)hyphenated to an API 5000 triple quadrupole mass spectrometer(AB SCIEX, Framingham, MA USA). L-phenylalanine (Sigma-Aldrich,St. Louis, MO, USA) was used to prepare standard solutions, andlabeled L-phenylalanine-13C9, 15N (Sigma-Aldrich) was used as theinternal standard. Phe analysis was performed using an Acquity BEHC18 column (1.7 μm, 2.1 mm × 50 mm) with gradient separation,which included a 0.5 min hold at 100% mobile phase A (0.5%trifluoroacetic acid, 0.3% heptafluorobutyric acid in water) followed bya 0–30%mobile phase B (acetonitrile) gradient over 2.5 min, an increaseto 90% B over 3 min, a 1 minwash at 90% B, and re-equilibration at 100%A for 1.9 min at a flow rate was 0.5 mL/min. The MS/MS detection wascarried out in a positive ion mode with declustering potential (DP) andcollision energy (CE) manually optimized to 20 and 25 respectively,while the ion-spray potential and temperature were set to 1500 V and500 °C. MS/MS transitions were: 166.1/120.1 for phenylalanine and176.1/129.1 for labeled phenylalanine.

BH4 is highly unstable, therefore 100 ng/mL cysteine, 0.1%dithioerythritol was used as the sample diluent buffer to minimizeoxidation. Analysis of BH4 was performed using a Phenomenex Luna

341N.S. Yew et al. / Molecular Genetics and Metabolism 109 (2013) 339–344

3 μm C18(2) Mercury 2.0 mm × 2.0 mm column cartridge kept at30 °C. The chromatographic run was performed at 1 mL/min flowunder isocratic conditionswith amobile phase consisting of 1% acetoni-trile with 99% of a 0.5% trifluoroacetic acid and 0.3% heptafluorobutyricacid solution in water. Mass spectrometric detection was carried outusing electrospray ionization in the positive ion mode. Theprecursor-product ion transition was 242.0/166.1. The DP and CEwere manually optimized to 50 and 25, respectively, while theion-spray potential and temperature were set to 1500 V and 500 °C.

2.8. Statistical analysis

Data were analyzed by one-way analysis of variance (ANOVA)followed by Bonferroni's post-test. Statistical significance was assignedas follows: * indicates P b 0.01 and ** indicates P b 0.001.

3. Results

3.1. Purification of recombinant C. violaceum PAH from E. coli and stabilityof purified PAH in vitro

The sequence encoding PAH from C. violaceum was codon-optimized for expression in E. coli, synthesized, and the syntheticgene then cloned into the inducible plasmid expression vector pET29asuch that a 6× His tag was incorporated into the C-terminus of the pro-tein. Expression of PAHwas induced by IPTG, and the protein was puri-fied from the soluble E. coli lysate by affinity chromatography over anickel-agarose resin. The column was washed extensively with 0.1%Triton-X114 buffer to remove endotoxin. Approximately 130 mg ofprotein of>90% puritywas obtained from6 L of culture (Fig. 1A). Endo-toxin levels were typically less than 0.5 EU/mg and the specific activityof the purified recombinant enzyme was ~0.3 U/mg.

A prerequisite for encapsulating an enzymewithin RBCs is that theenzyme should be stable, comparable with the median half-life of theRBCs, and not prone to inactivation. The stability of solutions of puri-fied PAH were evaluated in vitro by incubation at 4, 23, or 37 °C for21 days. The enzyme exhibited little loss of activity over the first14 days at all the temperatures tested, with a moderate decrease inactivity at 37 °C after three weeks of incubation (Fig. 1B). Thus, thepurified PAH was relatively stable over time in vitro.

Fig. 1. Purification and in vitro stability of recombinant C. violaceum PAH. A) The PAH was eCoomassie Blue stained gel of the purified enzyme (5 μg load). B) Stability of recombinant C4 °C, 23 °C, or 37 °C for three weeks. Enzyme activity was measured at various time points

3.2. Encapsulation of C. violaceum PAH into mouse RBCs

We next encapsulated the purified PAH protein into normal mouseRBCs by reversible hypotonic dialysis and resealing. The procedure uti-lizes a hollow fiber dialyzer to transiently disrupt the erythrocytemem-brane under controlled conditions such that hemolysis is minimized.Currently this is the only procedure that can process a sufficient volumeof blood for clinical use. The hematologic characteristics of a representa-tive batch of PAH-encapsulated RBCs are shown in Table 1. The corpus-cular hemoglobin concentration decreased only slightly compared tothe concentration before dialysis, indicating that hemoglobin waslargely retained within the PAH-loaded RBCs. The low extracellularhemoglobin concentration indicated that there was little hemolysisof the PAH-RBCs upon storage overnight at 4 °C.

To determine the efficiency of encapsulation and the amountencapsulated, the mixture of RBCs and purified PAH before encapsu-lation as well as the final product were analyzed by immunoblotting.The RBC plus PAH mixture before encapsulation was diluted 2- and4-fold prior to loading the sample on a gel to prevent oversaturationof the signal. By comparison to purified PAH standards, approxi-mately one-quarter of the initial PAH was successfully encapsulatedwithin the RBCs, with the final product containing approximately0.4 mg of PAH per mL (Fig. 2A). Analysis of the PAH-RBCs after oneday of storage at 4 °C showed that the cells remained intact withno detectable leakage of PAH into the supernatant. Enzymatic analy-sis of the PAH-RBCs showed that the PAH enzyme remained activeafter encapsulation, with no loss of activity after one day of storageat 4 °C (Fig. 2B).

3.3. Pharmacokinetics of PAH-RBCs in normal mice

To determine the pharmacokinetic profile of the encapsulatedPAH in vivo, mice were injected intravenously with PAH-RBCs at adose of 10 mL/kg (250 μL for a 25 g mouse). An equivalent amountof free PAH (4 mg/kg, i.e. 0.1 mg for a 25 g mouse) was injectedinto a separate group of mice for comparison, and blood was collectedat 15′, 6 h, 1, 3, 6, and 10 days post-injection. Although abundantlevels of PAH were present at the 15′ timepoint in the group injectedwith free PAH, circulating enzyme was undetectable by 6 h post-injection (Fig. 3A). In contrast, PAH levels in the blood of mice injectedwith the PAH-RBCs declined over the first few days but persisted

xpressed in E. coli and purified by affinity chromatography (see Methods). Shown is a. violaceum PAH enzyme activity in vitro. Aliquots of purified enzyme were incubated at.

Table 1Hematologic characterization of PAH-RBCs.

Day 0 Day 1

RBC + PAH PAH-RBC PAH-RBC

Hematocrit (%) 73.7 49.4 47.3Globular volume (μm3) 54.4 47.8 46.6Extracellular hemoglobin (g/dL) Nd 0.2 1.0Corpuscular hemoglobin (g/dL) 27.7 23.4 23.5RBC/mL (×109) 15.8 12.9 13.2

Nd — not determined.

342 N.S. Yew et al. / Molecular Genetics and Metabolism 109 (2013) 339–344

through day 10 (Figs. 3B and C). Enzyme activity of washed RBCs sam-pled at each time point showed that PAH activity was present for theduration of the study (Fig. 3D). These results demonstrate that encapsu-lation of PAHwithin RBCs greatly improved the pharmacokinetics of theenzyme in the circulation.

3.4. Efficacy of PAH-RBCs in normal and PKU mice

We next asked if the injected PAH-RBCs were capable of metaboliz-ing Phe in the blood. Normal OF1 mice were injected with PAH-RBCs(10 mL/kg) and Phe levels were measured 6 h post-injection. SincePAH requires the cofactor BH4 for enzymatic activity, BH4 was injectedeither intravenously or intraperitoneally 30 min prior to bleeding theanimals. BH4 has been shown to readily enter erythrocytes by passivetransport [16]. The time of injection was chosen based on the reportedshort half-life and rapid clearance of BH4 in vivo [17,18]. Injection of

Fig. 2. Encapsulation of C. violaceum PAH in RBCs. A) Western blot analysis to estimatethe amount of PAH entrapped within the RBCs. Legend: (RBC + PAH), the mixture ofRBCs and PAH prior to encapsulation; (PAH-RBC), encapsulated RBCs on day 0 (d0)and day 1 (d1) after encapsulation; (S), the supernatant of the PAH-RBC suspension,or RBCs alone (RBC). An equivalent of 5 nL of each sample was loaded per lane, withthe exception that the RBC + PAH sample was diluted 0.5× and 0.25× before loading(i.e. an equivalent of 2.5 and 1.25 nL of sample loaded). The blot was probed with ananti-PAH antibody. B) Enzyme activity of the PAH-RBCs. 10 μL of the PAH-RBCs wasassayed on day 0 and day 1. Data shown is representative of several independentencapsulations.

Fig. 3. Pharmacokinetics of free PAH and PAH-RBCs after IV administration into normalmice. A) Western blot of whole blood 15 min and 6 h after injection of free PAH en-zyme (4 mg/kg). B)Western blot of whole blood 6 h, 1, 3, 6, and 10 days after injectionof PAH-RBCs. For both A) and B) an equivalent of 0.5 μL of blood was loaded per lane.N = 3 mice per time point. C) Quantitation of PAH levels in blood over time after injec-tion of PAH-RBCs. The PAH signal from B)was quantitated by densitometry and comparedto purified PAH standards. D) PAH enzyme activity in blood over time after injection ofPAH-RBCs. PAH activity was measured from whole blood samples taken at 6 h, 1, 3, 6,and 10 days post-injection. N = 3 mice per time point. The data are expressed as themean ± SEM.

BH4 alone resulted in a small decrease in Phe levels, possibly due to ac-tivation of the endogenous liver PAH present in these normal animals(Fig. 4A). High levels of BH4 were present in the blood at the 6 h timepoint after either IV or IP injection of BH4 at 5.5 h after injection of thePAH-RBCs (Fig. 4B). Administration of the PAH-RBCs along with eitherIV or IP injection of BH4 resulted in a substantial decrease in Phe levelsof 59 and 78% respectively (P b 0.001) (Fig. 4A). Consistent with thisobservation was a corresponding increase in plasma tyrosine levels inthe groups injected with the PAH-RBCs (data not shown). Hence, theresults demonstrated that the PAH-RBCs could effectively metabolizePhe to tyrosine in the blood and lower Phe levels in normal mice.

The PAH-RBCs were then evaluated in the PAHenu2 mouse, whichcontains a mutation in the active site of PAH and therefore lacksPAH activity. As a result this mouse exhibits highly elevated levelsof Phe in plasma, reaching 10–20× higher than the levels found innormal animals. To ensure the RBCs from OF1 mice were compatiblewith the BTBR strain of the PAHenu2 mice, hemagglutination assayswere performed and were negative (data not shown). Male PAHenu2

mice were injected IV with PAH-RBCS (12 mL/kg) and Phe levelswere measured 6 h post-injection. As a control, one group of micewas injected with processed RBCs dialyzed without PAH (CON-RBC).To ensure that sufficient levels of BH4 cofactor were continuouslypresent in the blood over the 6 h period, BH4 was injected immedi-ately after dosing the PAH-RBCs and then hourly thereafter. Unlike

Fig. 4. Effect of IV administration of PAH-RBCs on plasma Phe levels. A) Plasma Phe levels 6 h post-injection of PAH-RBCs into normal mice. BH4 was injected IP or IV (200 mg/kg)5.5 h after injection of PAH-RBCs. N = 3 mice per group. B) Plasma BH4 levels 6 h post-injection of PAH-RBCs. N = 1 for IV alone group. C) Plasma Phe levels 6 h post-injection ofPAH-RBCs into PAHenu2 mice. BH4 was injected IP (150 mg/kg) at 0, 1.5, 2.5, 3.5, 4.5, and 5.5 h after injection of PAH-RBCs. D) Plasma BH4 levels 6 h post-injection of PAH-RBCs intoPAHenu2 mice. Het, heterozygous mice. N = 8 mice per group (N = 3 mice Naïve het). **P b 0.001, *P b 0.01. The data are expressed as the mean ± SEM.

343N.S. Yew et al. / Molecular Genetics and Metabolism 109 (2013) 339–344

what was observed in normal animals, administration of the PAH-RBCs did not reduce Phe levels in the PAHenu2 mice (Fig. 4C). As wasobserved previously in normal mice, high levels of BH4 were presentin the blood at the 6 h time point (Fig. 4D), indicating that sufficientlevels of cofactor were available for the enzyme. We estimated thatapproximately 0.4 mg of PAH (0.12 U) was delivered into the circula-tion and the enzyme was confirmed to be active at the 6 h timepoint(data not shown). Nevertheless, the results indicate that this amountof PAH was insufficient to significantly decrease the extremely highlevels of Phe present in the circulation of PAHenu2 mice.

4. Discussion

In this study, a prokaryotic, monomeric form of PAH was entrappedinto mouse RBCs and demonstrated to function as phenylalanine me-tabolizing bioreactors when injected into the bloodstream. Althoughnot directly shown, the subsequent decrease in Phe levels upon injec-tion of PAH-RBCs and BH4 strongly suggest that the PAH-RBCs wereable to take up both Phe and the necessary cofactor BH4 from the circu-lation and convert phenylalanine to tyrosine. Compared to free PAH,which survived in the blood for less than 1 h, the PAH-RBCs persistedin the circulation and remained active for at least 10 days, confirmingthe ability of RBC encapsulation to greatly improve the pharmacokinet-ics of the therapeutic enzyme.

Given that the concept of encapsulating enzymes within RBCs hasbeen in existence for a while, the examples and applications usingthis approach over the past several years are relatively few [19]. En-zymes have been entrapped to remove toxic compounds from theblood, e.g. thiosulfate-cyanide sulfurtransferase (to treat cyanidepoisoning), δ-aminolevulinate dehydrogenase (lead intoxication),and alcohol dehydrogenase/oxidase (alcohol poisoning) [20–24].Encapsulated enzymes have been evaluated for various genetic

diseases, for example thymidine phosphorylase (mitochondrialneurogastrointestinal encephalomyopathy), arginase (familialhyperargininemia) and adenosine deaminase (ADA deficiency) [25–30].

One of the most promising and advanced applications is the en-capsulation of asparaginase for the treatment of acute lymphoblasticleukemia (ALL) [31]. While asparaginase is widely used to treat ALL,the enzyme has a short half-life in plasma (from 15 to 30 hdepending on the bacterial source of the protein), thus requiring fre-quent readministrations and inducing significant adverse immuneresponses. Clinical trials using asparaginase-loaded RBCs haveshown greatly improved pharmacokinetics and pharmacodynamicscompared to the free enzyme and significantly reduced adverse im-mune effects [4,32,33].

Although the PAH-RBCs were effective in lowering Phe levels innormal mice, the lack of efficacy in PKU mice may be more due tothe low specific activity of the enzyme and the large amount of sub-strate involved rather than a limitation of the technology. The specificactivity of our enzyme preparation was ~0.3 U/mg protein (one unitbeing defined as 1 μmol of tyrosine formed per minute at 23 °C). Al-though not directly comparable, the specific activity of asparaginaseis greater than 200 U/mg protein (one unit being defined as 1 μmolof ammonia per min at 37 °C).

Secondly, very high plasma Phe levels (1600–1800 μM) are presentin the PKUmouse, 10–20 times that of normalmice. However, based onthe amount of PAH delivered and the specific activity of the enzyme, weestimate that ~5 μmol of Phe could have been consumed over a six hourperiod. Possible explanations are that there is a continuous influx of Pheinto the circulation, and that the in vivo activity of PAH was lower thanthe activity measured in vitro.

Nevertheless, a pegylated form of phenylalanine ammonia lyase(PAL) has been shown to decrease Phe levels in the PKU mouse [34].PAL is a 240 kD tetrameric enzyme that converts Phe to trans-cinnamic

344 N.S. Yew et al. / Molecular Genetics and Metabolism 109 (2013) 339–344

acid plus ammonia. We have observed that PAL-loaded RBCs lowerPhe levels in normal mice but the pharmacokinetics of PAL-RBCswere uncharacteristically short, with PAL being lost from the circula-tion within 24 h (data not shown). Further studies are needed to im-prove the viability of PAL-RBCs, to increase the activity of our PAHenzyme preparation, and to investigate the use of PAH and PAL fromother organisms. Use of a low Phe diet may also aid in demonstratingefficacy in the PKU mouse.

Although PKU represents a relatively high hurdle, the use ofenzyme-loaded RBCs can potentially be applied to a number of othergenetic disorders of amino acid metabolism, such as methylmalonicacidemia and related branched-chain amino acid defects [35]. Providedthat a sufficiently robust enzyme can be found for the given indication,the testing of the RBC-encapsulated enzyme in an animal diseasemodelwould be relatively straightforward, which may lead to further clinicalapplications of this highly promising and novel form of enzyme replace-ment therapy.

Conflict of interest statement

ED, JP, AM, FH, and YG are employees of ERYtech Pharma. NSY, MP,CC, MF, SG, DR, AK, and SHC are employees of Genzyme.

Acknowledgments

We thank Estelle Houde, Zhengyu Luo, Robert Jones, and the Com-parative Medicine staff for the technical assistance, and Andrew Legerfor critically reading the manuscript.

References

[1] V.R. Muzykantov, Drug delivery by red blood cells: vascular carriers designed bymother nature, Expert. Opin. Drug Deliv. 7 (2010) 403–427.

[2] M.Magnani, F. Pierige, L. Rossi, Erythrocytes as a novel delivery vehicle for biologics:from enzymes to nucleic acid-based therapeutics, Ther. Deliv. 3 (2012) 405–414.

[3] Y. Godfrin, F. Horand, R. Franco, E. Dufour, E. Kosenko, B.E. Bax, A. Banz, O.A. Skorokhod,J.M. Lanao, V. Vitvitsky, E. Sinauridze, V. Bourgeaux, K.C. Gunter, International seminaron the red blood cells as vehicles for drugs, Expert. Opin. Biol. Ther. 12 (2012) 127–133.

[4] Y. Godfrin, B.E. Bax, Enzyme bioreactors as drugs, Drugs Future 37 (2012) 263–272.[5] M. Hamidi, H. Tajerzadeh, Carrier erythrocytes: an overview, Deliv. Rev. 10 (2003)

9–20.[6] G. Tunnicliff, Amino acid transport by human erythrocyte membranes, Comp.

Biochem. Physiol. Comp. Physiol. 108 (1994) 471–478.[7] C.R. Scriver, S. Kaufman, Hyperphenylalanemia: phenylalanine hydroxylase defi-

ciency, in: C.R. Scriver, A.L. Beaudet, W.S. Sly, D. Valle (Eds.), The Metabolic andMolecular Bases of Inherited Disease Phenylketonuria, McGraw-Hill, New York,2001, pp. 1667–1724.

[8] N. Blau, F.J. van Spronsen, H.L. Levy, Lancet 376 (2010) 1417–1427.[9] J.J. Mitchell, Y.J. Trakadis, C.R. Scriver, Phenylalanine hydroxylase deficiency,

Genet. Med. 13 (2011) 697–707.[10] F.J. van Spronsen, Phenylketonuria: a 21st century perspective, Nat. Rev. Endocrinol.

6 (2010) 509–514.[11] R. Kravtzoff, C. Ropars, M. Laguerre, J.P. Muh, M. Chassaigne, Erythrocytes as carriers

for L-asparaginase. Methodological and mouse in-vivo studies, J. Pharm. Pharmacol.42 (1990) 473–476.

[12] S. Kaufman, Phenylalanine 4-monooxygenase from rat liver, Methods Enzymol.142 (1987) 3–17.

[13] H.Nakata, T. Yamauchi, H. Fujisawa, Phenylalaninehydroxylase fromChromobacteriumviolaceum. Purification and characterization, J. Biol. Chem. 254 (1979) 1829–1833.

[14] E. Dufour, F. Gay, K. Aguera, J.Y. Scoazec, F. Horand, P.L. Lorenzi, Y. Godfrin, Pancreatictumor sensitivity to plasma L-asparagine starvation, Pancreas 41 (2012) 940–948.

[15] V. Bourgeaux, O. Hequet, Y. Campion, G. Delcambre, A.M. Chevrier, D. Rigal, Y.Godfrin, Inositol hexaphosphate-loaded red blood cells prevent in vitro sickling,Transfusion 50 (2010) 2176–2184.

[16] K. Azumagawa, S. Suzuki, M. Mino, Developmental changes of tetrahydrobiopterinin rat and human erythrocytes, J. Nutr. Sci. Vitaminol. (Tokyo) 41 (1995) 607–617.

[17] C.O. Harding, M. Neff, K. Wild, K. Jones, L. Elzaouk, B. Thony, S. Milstien, The fate ofintravenously administered tetrahydrobiopterin and its implications for heterol-ogous gene therapy of phenylketonuria, Mol. Genet. Metab. 81 (2004) 52–57.

[18] A. Ohashi, Y. Suetake, Y. Saeki, T. Harada, S. Aizawa, H. Hasegawa, Rapid clearanceof supplemented tetrahydrobiopterin is driven by high-capacity transporters inthe kidney, Mol. Genet. Metab. 105 (2012) 575–581.

[19] C.G. Millan, M.L. Marinero, A.Z. Castaneda, J.M. Lanao, Drug, enzyme and peptidedelivery using erythrocytes as carriers, J. Control. Release 95 (2004) 27–49.

[20] P. Leung, L.E. Ray, C. Sander, J.L. Way, D.M. Sylvester, Encapsulation of thiosulfate:cyanide sulfurtransferase by mouse erythrocytes, Toxicol. Appl. Pharmacol. 83(1986) 101–107.

[21] L. Pei Petrikovics, W.D. McGuinn, E.P. Cannon, J.L. Way, Encapsulation of rhodaneseand organic thiosulfonates by mouse erythrocytes, Fundam. Appl. Toxicol. 23 (1994)70–75.

[22] N.L. Bustos, A.M. Batlle, Enzyme replacement therapy in porphyrias—V. In vivocorrection of delta-aminolaevulinate dehydratase defective in erythrocytes inlead intoxicated animals by enzyme-loaded red blood cell ghosts, Drug Des.Deliv. 5 (1989) 125–131.

[23] M. Magnani, U. Mancini, M. Bianchi, A. Fazi, Comparison of uricase-bound anduricase-loaded erythrocytes as bioreactors for uric acid degradation, Adv. Exp.Med. Biol. 326 (1992) 189–194.

[24] C. Lizano, M.T. Perez, M. Pinilla, Mouse erythrocytes as carriers for coencapsulatedalcohol and aldehyde dehydrogenase obtained by electroporation in vivo survivalrate in circulation, organ distribution and ethanol degradation, Life Sci. 68 (2001)2001–2016.

[25] B.E. Bax, M.D. Bain, C.P. Ward, A.H. Fensom, R.A. Chalmers, The entrapment ofmannose-terminated glucocerebrosidase (Alglucerase) in human carrier erythro-cytes, Biochem. Soc. Trans. 24 (1996) 441S.

[26] M. Levene, D.G. Coleman, H.C. Kilpatrick, L.D. Fairbanks, B. Gangadharan, C.Gasson, B.E. Bax, Preclinical toxicity evaluation of erythrocyte-encapsulated thy-midine phosphorylase in BALB/c mice and Beagle dogs: an enzyme replacementtherapy for mitochondrial neurogastrointestinal encephalomyopathy, Toxicol.Sci. 131 (2012) 311–324.

[27] K. Adriaenssens, D. Karcher, B. Marescau, C. Van Broeckhoven, A. Lowenthal, H.C.Terheggen, Hyperargininemia: the rat as a model for the human disease and thecomparative response to enzyme replacement therapy with free arginase andarginase-loaded erythrocytes in vivo, Int. J. Biochem. 16 (1984) 779–786.

[28] Y.G. Kaminsky, E.A. Kosenko, Argocytes containing enzyme nanoparticles reducetoxic concentrations of arginine in the blood, Bull. Exp. Biol. Med. 153 (2012)406–408.

[29] B.E. Bax, M.D. Bain, L.D. Fairbanks, A.D. Webster, R.A. Chalmers, In vitro and invivo studies with human carrier erythrocytes loaded with polyethylene glycol-conjugated and native adenosine deaminase, Br. J. Haematol. 109 (2000) 549–554.

[30] B.E. Bax, M.D. Bain, L.D. Fairbanks, A.D. Webster, P.W. Ind, M.S. Hershfield, R.A.Chalmers, A 9-yr evaluation of carrier erythrocyte encapsulated adenosine deam-inase (ADA) therapy in a patient with adult-type ADA deficiency, Eur. J. Haematol.79 (2007) 338–348.

[31] Y.M. Kwon, H.S. Chung, C. Moon, J. Yockman, Y.J. Park, S.D. Gitlin, A.E. David, V.C.Yang, L-Asparaginase encapsulated intact erythrocytes for treatment of acutelymphoblastic leukemia (ALL), J. Control Release 139 (2009) 182–189.

[32] R. Kravtzoff, I. Desbois, J.P. Lamagnere, J.P. Muh, C. Valat, M. Chassaigne, P.Colombat, C. Ropars, Improved pharmacodynamics of L-asparaginase-loaded inhuman red blood cells, Eur. J. Clin. Pharmacol. 49 (1996) 465–470.

[33] C. Domenech, X. Thomas, S. Chabaud, A. Baruchel, F. Gueyffier, F. Mazingue, A.Auvrignon, S. Corm, H. Dombret, P. Chevallier, C. Galambrun, F. Huguet, F.Legrand, F. Mechinaud, N. Vey, I. Philip, D. Liens, Y. Godfrin, D. Rigal, Y. Bertrand,L-asparaginase loaded red blood cells in refractory or relapsing acute lymphoblasticleukaemia in children and adults: results of the GRASPALL 2005-01 randomized trial,Br. J. Haematol. 153 (2011) 58–65.

[34] C.N. Sarkissian, A. Gamez, L. Wang, M. Charbonneau, P. Fitzpatrick, J.F. Lemontt, B.Zhao, M. Vellard, S.M. Bell, C. Henschell, A. Lambert, L. Tsuruda, R.C. Stevens, C.R.Scriver, Preclinical evaluation of multiple species of PEGylated recombinant phe-nylalanine ammonia lyase for the treatment of phenylketonuria, Proc. Natl. Acad.Sci. U.S.A. 105 (2008) 20894–20899.

[35] I. Knerr, N. Weinhold, J. Vockley, K.M. Gibson, Advances and challenges in thetreatment of branched-chain amino/keto acid metabolic defects, J. Inherit. Metab.Dis. 35 (2012) 29–40.