Journal of Neuroimmunology 223 (2010) 138–140

Contents lists available at ScienceDirect

Journal of Neuroimmunology

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

Short communication

The serum and cerebrospinal fluid pharmacokinetics of anakinra after intravenousadministration to non-human primates

Elizabeth Fox a,⁎, Nalini Jayaprakash b, Tuyet-Hang Pham c, Ayana Rowley c, Cynthia L. McCully b,Frank Pucino c,d, Raphaela Goldbach-Mansky c

a The Children's Hospital of Philadelphia, Philadelphia, PA, United Statesb Pediatric Oncology Branch, National Cancer Institute, Bethesda, MD, United Statesc National Institute of Arthritis and Musculoskeletal and Skin Diseases, NIH, Bethesda, MD, United Statesd Center for Drug Evaluation and Research, FDA, DHHS, Silver Spring, MD, United States

⁎ Corresponding author. The Children's HospitalOncology-CTRB 4016, 3501 Civic Center Blvd, PhiladelTel.: +1 267 425 3010; fax: +1 267 425 0113.

E-mail address: Foxe@email.chop.edu (E. Fox).

0165-5728/$ – see front matter © 2010 Elsevier B.V. Adoi:10.1016/j.jneuroim.2010.03.022

a b s t r a c t

a r t i c l e i n f o

Article history:Received 28 December 2009Received in revised form 26 March 2010Accepted 29 March 2010

Keywords:IL-1RaIL-1AnakinraPharmacokineticsNOMID

Anakinra improves the central nervous system manifestations of neonatal-onset multisystem inflammatorydisease, which is mediated by IL-1ß oversecretion. The cerebrospinal fluid (CSF) penetration of the IL-1receptor antagonist anakinra was studied in rhesus monkeys after intravenous doses of 3 and 10 mg/kg. Drugexposure (area under concentration–time curve) in CSF was 0.28% of that in serum. The average CSFconcentration at 3 mg/kg was 1.8 ng/mL, which is 30-fold higher than endogenous CSF levels of IL-1Ra. TheCSF penetration was not dose-dependent, indicating that the CSF penetration was not saturated in the 3 to10 mg/kg dose range.

of Philadelphia, Division ofphia, PA 19104, United States.

ll rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Interleukin-1 (IL-1), implicated in the pathogenesis of neonatal-onset multisystem inflammatory disease (NOMID), mediates pro-inflammatory effects through binding to the IL-1 receptor which ispresent on many nucleated cells (Dinarello, 2004). The recombinantIL-1 receptor antagonist, anakinra (IL-1Ra, Kineret™), improves theclinical and laboratory manifestations of NOMID in patients treatedwith 1–2 mg/kg daily by subcutaneous injection (Goldbach-Manskyet al., 2006; Lovell et al., 2005). In addition to the disappearance of thecutaneous manifestations, an improvement in the central nervoussystem (CNS) manifestations was observed, including alleviation ofheadache, reduction in increased intracranial pressure, and decreasedleptomeningeal and cochlear enhancement on MRI. The median(interquartile range) serum IL-1Ra concentration in patients withNOMIDwas 0.36 ng/mL (0.23–1.26 ng/mL) pretreatment and 43 ng/mL(8.8–200 ng/mL) after 3 months of anakinra. In CSF, IL-1Ra concentra-tion was 0.21 ng/mL (0.077–0.35 ng/mL) and increased 5-fold onanakinra to 1.14 ng/mL (0.50–1.69 ng/mL) (Goldbach-Mansky et al.,2006).

The clinical improvement in the CNSmanifestations of NOMID andthe increase in CSF IL-1Ra concentrations measured at a single timepoint post-treatment in children treated with anakinra prompted usto more fully characterize the CSF pharmacokinetics of anakinra in anon-human primate model, which is highly predictive of CNSpharmacology in humans (Bacher et al., 1994; McCully et al., 1990;Jacobs et al., 2005; Meany et al., 2007; Muscal et al., 2010). Animalshave chronically indwelling Ommaya reservoirs that permit serialatraumatic sampling of ventricular CSF to assess the time course ofCSF drug concentrations after a systemic dose of drug. Drug exposurein CSF is compared to serum using the area under the concentration–time curve (AUC) rather than single time point measurements. CSFdrug penetration in this model is a surrogate for blood-brain barrierpenetration (Fox et al., 2002).

2. Materials and methods

2.1. Drug

Anakinra (100 mg/0.67 mL; Biovitrum, AB, Stockholm, Sweden)waspurchased from commercial sources. Anakinra (17.3 kDa) differs fromnative human IL-1Ra (23–25 kDa) by the addition of a methionineat the amino terminus and the absence of glycosylation. Two dose levels(3 and 10 mg/kg) were studied. Each dose was diluted in normal salineto a final concentration of 25–50 mg/mL in 1–3 mL and administered asan intravenous (IV) bolus.

139E. Fox et al. / Journal of Neuroimmunology 223 (2010) 138–140

2.2. Animals

Three adult male rhesus monkeys (Macaca mulatta), weighing10.2–14.4 kg, were used for this study. Animals were group housed inaccordance with the Guide for the Care and Use of Laboratory Animals(http://oacu.od.nih.gov/regs/guide/guide.pdf). The experimental pro-tocol was approved by the National Cancer Institute's Animal Care andUse Committee. IV bolus anakinra was delivered through a jugularvenous port, and blood samples were drawn from a saphenous venouscatheter on the side contralateral to drug administration. CSF sampleswere obtained from an indwelling Pudenz silicon catheter attached toa subcutaneously implanted Ommaya reservoir (Bacher et al., 1994).

2.3. Experiments

Blood samples (3 mL)were collected before, and 0.25, 0.5, 1, 2, 3, 4,6, 8, 10, 24, and 48 h after anakinra administration. The blood wasallowed to clot for 30 min and serum was separated by centrifugationand stored at −70 °C. CSF samples (0.3 mL) were collected at thesame time points as blood. The Ommaya reservoir was pumped 3times prior to collection of each sample to insure adequate mixing ofCSF. CSF samples were immediately frozen and stored at −70 °C.Animals RQ3633 and RQ3588 were studied after 3 and 10 mg/kg ofanakinra with N2 months break between experiments. Animal 15398was studied after 10 mg/kg.

2.4. Sample analysis

Serum and CSF anakinra concentrations were measured with ahuman IL-1Ra ELISA (Invitrogen, BioSource™). Spectrophotometricendpoint detection (OD 450 nM) of the biotin–streptavidin horse-radish peroxidase product was quantified using a SpectraMax M5(Molecular Devices Sunnyvale, CA). Standard curves were generatedusing anakinra rather than the supplied native IL-1Ra standard. CSFstandards were prepared by diluting anakinra in phosphate bufferedsaline with 1% bovine serum albumin. All standards, controls andsamples were analyzed in duplicate and the assay validated accordingto FDA guidelines (US DHHS, 2001). Standard curves using a fourparameter logistic fit for serum and CSF anakinra concentrationsranged from 0.062 to 4.0 ng/mL. The median (range) R2 for standardcurves generated on 9 days of sample analysis was 0.999 (0.998–1.000). The lower limit of detection was 0.032 ng/mL and the lowerlimit of quantification was 0.062 ng/mL. For low (0.062 ng/mL), mid(0.50 ng/mL) and high (2.0 ng/mL) serum controls the day to daycoefficient of variationwas 8%, 2%, and 3%, respectively. The day to daycoefficient of variation for CSF controls was 3%, 1% and 1% at low, midand high concentrations, respectively.

2.5. Pharmacokinetic analysis

The serum and CSF pharmacokinetics of anakinra were analyzedby non-compartmental methods (Gibaldi and Pierrier, 1982). The areaunder the concentration–time curve (AUC0–last) was calculated withthe linear trapezoidal method and extrapolated to infinity (AUC0–∞)

Table 1Non-compartmental pharmacokinetic parameters for anakinra.

Animal Dose level[mg/kg]

Serum

AUC0–24 h

[ng h/mL]AUC0–∞[ng h/mL]

Cmax

[ng/mL]Half-life[h]

Clearanc[mL/min

RQ3588 3 14,900 14,900 25,000 2.3 3.3RQ3633 3 17,400 17,400 17,700 1.7 2.8RQ3588 10 59,900 59,900 114,000 2.5 2.9RQ3633 10 44,700 44,800 70,300 3.3 3.815398 10 68,300 69,800 9810 5.0 2.4

using the terminal rate constant derived from the slope of the naturallog-transformed concentrations and times on the terminal elimina-tion phase of the decay curve. Average concentration (Cave) at steadystate was derived from the AUC0–∞/24 h. Terminal half-life wascalculated by dividing 0.693 by the terminal rate constant, which wasderived from the slope of the natural log-transformed concentrationsand times on the terminal elimination phase of the plasma and CSFcurves. Anakinra clearance was calculated by dividing the dose by theAUC0–∞. Volume of distribution at steady state was determined fromthe anakinra area under the moment curve and AUC0–∞ withcorrection for the short duration of drug infusion. CSF penetration ofanakinra was calculated from the AUCCSF:AUCserum ratio.

3. Results

Pretreatment serum and CSF native IL-1Ra concentrations weredetectable (N0.032 ng/mL) but not quantifiable (b0.062 ng/mL). Serumand CSF pharmacokinetic parameters for anakinra are provided inTable 1, and the mean serum and CSF concentration–time profiles forthe 3 and 10 mg/kg doses are shown in Fig. 1. After the bolus dose,serum anakinra concentrations declined by 5 logs over 24 to 48 hwith amean (±SD) terminal half-life in serum of 2.9±1.3 h. The AUC0–∞ was16,200±1800 ng h/mL at 3 mg/kg and 58,200±12,600 ng h/mL at10 mg/kg, yielding Cave of 670 and 2420 ng/mL at 3 and 10 mg/kg.Mean (±SD) clearance was 3.0±0.5 mL/min/kg and Vdss was 0.45±0.36 L/kg.

CSF anakinra concentration peaked 1–2 h post-dose. The CSFAUC0–∞was 43±24 ng h/mL at 3 mg/kg and 164±11 ng h/mL at 10 mg/kg,and the Cave in CSF at 3 and 10 mg/kg were 1.8 and 6.8 ng/mL,respectively. Anakinra exposure in the CSFwas substantially lower thanserum (AUCCSF:AUCserum was 0.28%±0.10%). However, the decline inCSF anakinra concentrationswas slower (T1/2, 4.7 h) than in serum, suchthat by the last measurable time point the CSF anakinra concentrationwas 56% of the serum concentration.

4. Discussion

The overall drug exposure in CSF as measured by the AUC0–∞ wasb1% of the AUC0–∞ in serum. Bolus IV dosing of anakinra permittedaccurate measurement of the systemic and CSF exposure andcalculation of the CSF penetration as the ratio of the AUCCSF:AUCserum.The half-life was similar to the 1.8 h half-life reported for healthyhumans after IV administration of anakinra (Granowitz et al., 1992).Unlike comparing single time point serum and CSF measurements,the AUCCSF:AUCserum method of assessing penetration is independentof the shape of the concentration–time curves in serum and CSF.The slower decline in CSF anakinra concentrations results in higherCSF:serum ratios at later times when single time points are used toassess penetration. Despite the low penetration into CSF relative toserum, the CSF Cave at the 3 mg/kg dose was 30-fold higher thanpretreatment native IL-1Ra CSF concentrations and is comparable tothe CSF concentration measured in patients with NOMID (Goldbach-Mansky et al., 2006).

Cerebrospinal fluid Ratio

e/kg]

Vdss

[L/kg]AUC0–24 h

[ng h/mL]AUC0–∞[ng h/mL]

Cmax

[ng/mL]Half-life[h]

AUCCSF:AUCSerum [%]

0.25 59 60 16 4.6 0.400.54 26 27 6 1.0 0.150.18 158 166 17 6.0 0.260.25 148 152 24 5.0 0.341.05 164 175 21 6.8 0.24

Fig. 1. Mean serum and CSF anakinra concentration–time curves for 3 and 10 mg/kganakinra administered as an IV bolus dose.

140 E. Fox et al. / Journal of Neuroimmunology 223 (2010) 138–140

In humans with subarachnoid hemorrhage (SAH) treated withintravenous anakinra, the CSF penetration was 2–4% (Gueorguievaet al., 2008; Clark et al., 2008). Higher concentrations of anakinra inCSF of these patients may be expected after subarachnoid hemorrhagewith leakage of blood into CSF, use of an extravascular drain to collectCSF samples, and compromise of the blood-brain barrier after SAH.

Proteins are generally restricted from crossing the blood-CSF barrier,as evidenced by the low concentration of proteins in normal CSF. Inmurinemodels, cytokines including IL1α and IL-1β have been shown tocross the blood-brain barrier via saturable transport mechanisms(Banks et al., 1991, 1995; Gutierrez et al., 1994). Gutierrez et al.demonstrated that the peak percent of an IL-1Ra IV dose per gram ofbrain was 0.33% to 0.65% and occurred 30 to 40 min after injection.However at 10 min after injection, the percent of the dose enteringthe brain was 4–8 times lower. This demonstrates the differences inTmax and concentration–time profiles of ILRa in serum and CSF afterIV injection and demonstrates the inconsistencies inherent in usingindividual time points to determine CNS penetration. By comparison ofexposure (AUC0–∞) in CSF and serum, we demonstrate that the CSFpenetration of anakinra in non-human primates is 0.2–0.3%. Comparingthe 3 mg/kg and 10mg/kg dose, the exposure in both serum and CSFincreased by 3.7 fold at the higher dose and the CSF penetration wassimilar. This indicates that therewasno apparent saturation of transportof anakinra into the CSF in this dose range.

Inflammation of the meninges can increase blood-CSF barrierpenetration (Nag, 2003). After systemic administration of anakinra,childrenwith NOMID have an improvement of the CNSmanifestations(decreased CSF leukocytosis and protein) and a decrease in leptomen-ingeal enhancement on MRI (Goldbach-Mansky et al., 2006, 2007).Our non-human primate model has an intact blood-CSF barrier anddemonstrates that increasing the systemic dose of anakinra propor-tionally increased the systemic (serum) and CSF drug exposure. In

addition, the CSF penetration did not change indicating that the pen-etration into CSF was not dose-dependent over this dose range of 3 to10 mg/kg. These data provide a rationale to increase anakinra doses inpatients to improve control of CNS inflammation. Doses of 1–5 mg/kgare currently under investigation in a clinical trial with NOMIDpatients to obtain better control of the CNS inflammation that seemsto persist even if complete systemic inflammatory control is obtained(http://clinicaltrials.gov/ct2/show/NCT00069329).

References

Bacher, J.D., Balis, F.M., McCully, C.L., Godwin, K.S., 1994. Cerebral subarachnoidsampling of cerebrospinal fluid in the rhesus monkey. Lab. Anim. Sci. 44, 148–152.

Banks, W.A., Ortiz, L., Plotkin, S.R., Kastin, A.J., 1991. Human interleukin (IL) 1 alpha,murine IL-1 alpha and murine IL-1 beta are transported from blood to brain in themouse by a shared saturable mechanism. J. Pharmacol. Exp. Ther. 259, 988–996.

Banks, W.A., Kastin, A.J., Broadwell, R.D., 1995. Passage of cytokines across the blood-brain barrier. Neuroimmunomodulation 2, 241–248.

Clark, S.R., McMahon, C.J., Gueorguieva, I., Rowland, M., Scarth, S., Georgiou, R., Tyrrell,P.J., Hopkins, S.J., Rothwell, N.J., 2008. Interleukin-1 receptor antagonist penetrateshuman brain at experimentally therapeutic concentration. J. Cereb. Blood FlowMetab. 28 (2), 387–394.

Dinarello, C.A., 2004. Therapeutic strategies to reduce IL-1 activity in treating local andsystemic inflammation. Curr. Opin. Pharmacol. 4, 378–385.

Fox, E., Bungay, P.M., Bacher, J., McCully, C.L., Dedrick, R.L., Balis, F.M., 2002. Zidovudineconcentration in brain extracellular fluid measured by microdialysis: steady-stateand transient results in rhesus monkey. J. Pharmacol. Exp. Ther. 301, 1003–1011.

Gibaldi, M., Pierrier, D., 1982. Pharmacokinetics. Marcel Dekker, New York. 445–457.Goldbach-Mansky, R., Dailey, N.J., Canna, S.W., Gelabert, A., Jones, J., Rubin, B.I., Kim, H.J.,

Brewer, C., Zalewski, C., Wiggs, E., Hill, S., Turner, M.L., Karp, B.I., Aksentijevich, I.,Pucino, F., Penzak, S.R., Haverkamp, M.H., Stein, L., Adams, B.S., Moore, T.L.,Fuhlbrigge, R.C., Shaham, B., Jarvis, J.N., O'Neil, K., Vehe, R.K., Beitz, L.O., Gardner, G.,Hannan, W.P., Warren, R.W., Horn, W., Cole, J.L., Paul, S.M., Hawkins, P.N., Pham, T.H.,Snyder, C.,Wesley, R.A., Hoffmann, S.C., Holland, S.M., Butman, J.A., Kastner, D.L., 2006.Neonatal-onset multisystem inflammatory disease responsive to interleukin-1betainhibition. N. Engl. J. Med. 355, 581–592.

Goldbach-Mansky, R., Pucino, F., Kastner, D.L., 2007. Treatment of patients withneonatal-onset multisystem inflammatory disease/chronic infantile neurologic,cutaneous, articular syndrome: comment on the article byMatsubara et al. ArthritisRheum. 56, 2099–2101 author reply 2101-2.

Granowitz, E.V., Porat, R., Mier, J.W., Pribble, J.P., Stiles, D.M., Bloedow, D.C., Catalano, M.A.,Wolff, S.M., Dinarello, C.A., 1992. Pharmacokinetics safety and immunomodulatoryeffects of human recombinant interleukin-1 receptor antagonist in healthy humans.Cytokine 4, 353–360.

Gueorguieva, I., Clark, S.R., McMahon, C.J., Scarth, S., Rothwell, N.J., Tyrell, P.J., Hopkins,S.J., Rowland, M., 2008. Pharmacokinetic modeling of interleukin-1 receptorantagonist in plasma and cerebrospinal fluid of patients following subarachnoidhemorrhage. Br. J. Clin. Pharmacol. 65 (3), 317–325.

Gutierrez, E.G., Banks, W.A., Kastin, A.J., 1994. Blood-borne interleukin-1 receptorantagonist crosses the blood-brain barrier. J. Neuroimmunol. 55, 153–160.

Jacobs, S.S., Fox, E., Dennie, C., Morgan, L.B., McCully, C.L., Balis, F.M., 2005. Plasma andcerebrospinal fluid pharmacokinetics of intravenous oxalplatin, cisplatin andcarboplatin in nonhuman primates. Clin. Cancer Res. 11, 1669–1674.

Lovell, D.J., Bowyer, S.L., Solinger, A.M., 2005. Interleukin-1 blockade by anakinraimproves clinical symptoms in patients with neonatal-onset multisystem inflam-matory disease. Arthritis Rheum. 52, 1283–1286.

McCully, C.L., Balis, F.M., Bacher, J., Phillips, J., Poplack, D.G., 1990. A rhesus monkeymodel for continuous infusion of drugs into cerebrospinal fluid. Lab. Anim. Sci. 40,520–525.

Meany, H.J., Fox, E., McCully, C., Tucker, C., Balis, F.M., 2007. The plasma andcerebrospinal fluid pharmacokinetics of erlotinib and its active metabolite (OSI-420) after intravenous administration of erlotinib in non-human primates. CancerChemother. Pharmacol. 62, 387–392.

Muscal, J.A., Thompson, P.A., Giranda, V.L., Dayton, B.D., Bauch, J., Horton, T., McGuffy, L.,Nuchtern, J.G., Dauser, R.C., Gibson, B.W., Su, J.M., 2010. Plasma and cerebrospinalfluid pharmacokinetics of ABT-888 after oral administration in non-humanprimates. Cancer Chemother. Pharmacol. 65 (3), 419–425.

Nag, S., 2003. Pathophysiology of blood-brain barrier breakdown. Methods Mol. Med.89, 97–119.

US Department of Health and Human Services (DHHS), FDA: Guidance for industry,bioanalytical method validation, 2001.