B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te /b ra in res

Research Report

Effect of caffeine and morphine on the developingpre-mature brain

Amy M. Blacka, Shawna Pandyaa, Darren Clarkb,Edward A. Armstronga, Jerome Y. Yagera,⁎aDepartment of Pediatrics, Section of Pediatric Neurosciences, Stollery Children's Hospital, University of Alberta, Edmonton, Alberta, CanadabDepartment of Behavioral Neurosciences, University of Alberta, Edmonton, Alberta, Canada

A R T I C L E I N F O

⁎ Corresponding author. Section of PediatricAberhart Centre One, 114 University Avenue

E-mail address: jyager@ualberta.ca (J.Y. Y

0006-8993/$ – see front matter © 2008 Elsevidoi:10.1016/j.brainres.2008.04.066

A B S T R A C T

Article history:Accepted 17 April 2008Available online 1 May 2008

Apnea of pre-maturity is common, occurring in 85% of infants born less than 34 weekgestation. Oral caffeine is the most frequent form of therapy, often in conjunction with theuse of intubation and intermittent ventilation. Morphine is used to reduce the pain believedto be associated with the latter. Little information is available on the effects of caffeine,morphine or their combination, on the developing brain. We determined the effect ofcaffeine andmorphine alone and in combination of cell death on the developing brain of therat. Cell death, measured by Fluoro-jade B and activated caspase-3, was significantlyincreased at 12 and 24 hour post-caffeine injection (P < 0.05) in the cortex, caudate, nucleusaccumbens, hypothalamus, hippocampus and superior colliculus. No alterations were seenfollowing morphine injection alone. However, in the thalamus, the combination of caffeineand morphine did increase cell death to a significantly greater extent than caffeine alone.Further research is required to determine the long-term pathologic and functional effects ofcaffeine and the combination of caffeine and morphine on the developing immature brain.

© 2008 Elsevier B.V. All rights reserved.

Keywords:ApneaPre-maturityCaffeineMorphineCell death

1. Introduction

Pre-term birth, defined as any infant born between 20 and37week gestation, constitutes approximately 6–12% of all birthsin industrialized countries and accounts for 70% of neonatalmortality and 75% of neonatal morbidity (Joseph et al., 2007;Wen et al., 2004). The rate of pre-term birth is expected toincrease asmedical interventions continue to bemore success-ful, making the care of these infants a major public healthconcern (Wen et al., 2004). Depending on gestational age andbirth weight, pre-term infants present with a wide range ofabnormal physiologic responses due to their immature organsystems. One of the most common problems is apnea of pre-

Neurosciences, Division, Edmonton, Alberta, Canager).

er B.V. All rights reserved

maturity, reported to afflict 85% of infants born at <34 weekgestation (Schmidt et al., 2006).

Apnea of pre-maturity is defined as cessation of breathinglasting longer than 15s and accompanied by bradycardia orhypoxia (Schmidt et al., 2006). The most common method oftreatment is continuous positive airway pressure and admin-istration of amethylxanthine, generally in the form of caffeine(1,3,7-trimethylxanthine). Caffeine is thought to stimulaterespiratory function through its excitatory effects on thecentral nervous system (Bauer et al., 2001; Janvier et al., 2004;Schmidt et al., 2006). It mainly acts as an antagonist atadenosine receptors, but at high doses can also inhibitphosphodiesterases, block GABAA receptors or cause a release

of Pediatric Neurology, Stollery Children's Hospital, Room 7315,ada T6G 2J3. Fax: +1 780 407 8283.

.

Table 1 – Fluoro-jade B cell count statistics at 12 h

Caffeine/morphine

Caffeine Morphine Saline

Cortex 23.55±8.522b,d

19.89±3.817

8.792±2.746 10.09±1.708

Caudate 54.56±18.71

94.2±24.11b,d

17.43±3.842 11.38±2.609

Nucleusaccumbens

354.2±99.54

201±40.44

236.5±63.55 106±28.33

Thalamus 636.9±88.07a,c

501.6±85.32a,d

159±24.27 93.2±30.47

Hypothalamus 455.3±102.6b

343.1±46.70

260.6±48.46 141.3±31.3

CA1 290.0±80.00b,d

109.9±27.59

47.89±18.90 29.61±15.41

Superiorcolliculus

366.9±27.01a,c

294.4±26.77b,d

163.7±13.48 160.9±27.52

aSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.001.bSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.05.cSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.001.dSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.05.

Table 2 – Fluoro-jade B cell count statistics at 24 h

137B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

of intracellular calcium (Daly and Fredholm, 1998; Kang et al.,2002). In this regard, animal studies have shown caffeinetreatment to alter adenosine receptor expression and distribu-tion (Montandon et al., 2006), cause transient motor impair-ments (Tchekalarova et al., 2005) and be neurotoxic to thenewborn rat (Kang et al., 2002). Despite limited research onsafety and long-term consequences, caffeine has been thestandard pharmacologic treatment of apnea since the early1970's and is currently among the most commonly adminis-tered drugs in neonatal intensive care (Bauer et al., 2001; Duet al., 2006; Janvier et al., 2004).

Due to the difficulties of pre-term birth, infants oftenrequire multiple pharmacologic treatments. It has beenreported that at least 15% of neonates in the NICU receivemore than ten different drugs, putting them in a high-riskgroup for adverse drug reactions (Du et al., 2006). Morphine isused in conjunction with caffeine when pain relief of intuba-tion and mechanical ventilation is required for infants withimmature respiratory systems (Anand et al., 2004; Du et al.,2006). Morphine is variably considered a safe and effectiverecommendation when administered for severe or continuingpain in neonates. Indeed several studies have shown theopioids, includingmorphine, to be neuroprotective both in cellculture and in the human pre-term newborn (Angeles et al.,2005; Kim et al., 2001). However, controversy remains, in thatothers have found morphine to be associated with adverseneurologic effects (Anand et al., 2004).

Pre-mature infants are at high risk for significant cognitiveand motor impairment. In this regard, studies have revealedlower cognitive scores in pre-term infants compared to termcontrols. Moreover, these infants have higher rates of atten-tion deficit disorder, and internalizing and externalizingbehaviors (Bhutta et al., 2002). Though outcome of pre-maturebirth is multi-factorial in nature, few studies have determinedthe effects of caffeine andmorphine, alone or in combination,on the immature developing brain. Research involving the

Fig. 1 – Graph depicting the weight of rat pups at 12, 24, and72 h following the injection of caffeine, morphine, or caffeineand morphine, or saline. #P<0.05 compared to all othergroups and, *P<0.05 compared to saline group.

pre-mature infant is limited and although combinations ofdrugs are routinely administered, safety and efficacy of thispolypharmacy have not been thoroughly evaluated. In thisstudy, we examined the acute neurotoxic effects of caffeine,morphine and their use in combination, on the neonatal ratbrain.

Caffeine/Morphine

Caffeine Morphine Saline

Cortex 14.13±2.427b

13.5±1.755

9.42±1.198 7.457±1.209

Caudate 60.45±13.00b,d

36.74±8.384

19.43±3.164 18.84±2.268

Nucleusaccumbens

251.7±53.28b,d

248.7±30.06b,d

101.7±17.17 98.49±12.39

Thalamus 445.6±75.15a,c,e

270.4±28.95b,d

92.44±11.79 101.9±11.81

Hypothalamus 442.8±60.40a,c

430.2±65.82a,c

102.8±14.95 144.3±23.03

CA1 410.5±136.5b,d

222.3±55.36

128±25.01 91.46±15.04

Superiorcolliculus

313.8±68.42b,c

373.5±44.64a,c

79.88±14.82 109.1±7.358

aSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.001.bSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.05.cSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.001.dSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.05.eSignificantly different from caffeine by ANOVA with Tukey's posthoc test P<0.05.

Table 3 – Cleaved caspase-3 cell count statistics at 12 h

Caffeine/Morphine

Caffeine Morphine Saline

Cortex 6763±392.3

6613±687.5

6550±170.2 6265±460.1

Thalamus 8425±315.2d,e

10,635±774.8a,c

6279±361.2 6710±677.8

Hypothalamus 8513±310.4b,d

8755±475b,d

6446±481.6 6230±122.8

CA1 10,663±409.5

11,140±672.2b

9600±511.9 8775±605.3

Superiorcolliculus

9440±860.1

10,081±796.7b

7135±766.6 6569±504.8

aSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.001.bSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.05.c

138 B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

2. Results

Prior to injection (P3), there were no significant differences inmean body weight between groups. At the 12 hour time pointfollowing injection, the caffeine + morphine group mean bodyweight was significantly lower than all other groups (C +M=7.13±0.15, C=8.21±0.14, M=7.75±0.10, and S=8.47±0.16;P<0.05). At the 24 hour time point, the caffeine + morphinegroup mean weight differed significantly from the salinegroup while no other group differences were found (C +M=7.43±0.39, and S=8.91±0.30; P<0.05). By 72h, the differencein mean weight for the caffeine + morphine group had disap-peared (Fig. 1). Statistical analysis of the data indicates that ratpups receiving the combination treatment of caffeine +morphine experienced an attenuated growth rate for thefirst 24 h. There was no mortality in any of the groups.

Fig. 2 – Graphs depicting an increase in Fluoro-jade Bstaining in the thalamus at 12 (a), and 24 (b) hourpost-caffeine and/or morphine injection. a — Significantlydifferent from saline by ANOVA with Tukey's post hoc testP<0.001. b — Significantly different from saline by ANOVAwith Tukey's post hoc test P<0.05. c — Significantly differentfrom morphine by ANOVA with Tukey's post hoc testP<0.001. d — Significantly different from morphine byANOVA with Tukey's post hoc test P<0.05. e — Significantlydifferent from caffeine by ANOVA with Tukey's post hoc testP<0.05.

Significantly different from morphine by ANOVA with Tukey'spost hoc test P<0.001.dSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.05.eSignificantly different from caffeine by ANOVA with Tukey's posthoc test P<0.05.

Areas from Fluoro-jade B stained caffeine treated sections,and caffeine plus morphine treated sections showed signifi-cant differences (P<0.05; P < 0.001, respectively) in positive cellcounts when compared to saline and morphine groups at 12and 24 hour time points (see Tables 1 and 2). At 12 h, thoseareas with significantly increased evidence of cell death in thecaffeine and morphine group, compared to saline controlsincluded the cortex, thalamus, hypothalamus, hippocampus,and the superior colliculus. Differences between the caffeinegroup alone and saline controls at this time point included thecaudate, thalamus, and the superior colliculus. By 24 hourpost-injection, increases were also seen in the caudate andnucleus accumbens for the caffeine andmorphine groups, and

Table 4 – Cleaved caspase-3 cell count statistics at 24 h

Caffeine/Morphine

Caffeine Morphine Saline

Cortex 6843±191.8

7058±215.8

6308±91.06 6436±313.3

Thalamus 7588±284.4a,c

7515±157.6a,c

5506±305.5 5673±217.8

Hypothalamus 7863±224.8b,c

7845±238.7b,c

6044±257 6473±258.7

CA1 8938±175.3b,d

8315±250.7

7428±353.2 7321±322.6

Superiorcolliculus

7257±206.4b,d

7230±222.5a,d

5957±233 5831±256.7

aSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.001.bSignificantly different from saline by ANOVA with Tukey's posthoc test P<0.05.cSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.001.dSignificantly different from morphine by ANOVA with Tukey'spost hoc test P<0.05.

Fig. 3 – Graphs depicting an increase in caspase-3 staining,also at 12 (a), and 24 (b) hour post-caffeine and/or morphineinjection. a — Significantly different from saline by ANOVAwith Tukey's post hoc test P<0.001. b — Significantlydifferent from saline by ANOVA with Tukey's post hoc testP<0.05. c — Significantly different from morphine by ANOVAwith Tukey's post hoc test P<0.001. d — Significantlydifferent from morphine by ANOVA with Tukey's post hoctest P<0.05. e — Significantly different from caffeine byANOVA with Tukey's post hoc test P<0.05.

139B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

in the nucleus accumbens and hypothalamus for the caffeinealone group (Fig. 2). No differenceswere exhibited by 72 h (datanot shown).

Similarly, areas from cleaved caspase-3 stained, caffeinetreated and caffeine plus morphine treated sections at 12 and24 h showed significant differences when compared to salineandmorphine groups in cleaved caspase-3 positive cell counts(see Tables 3 and 4). By 12 hour post-injection the combinedcaffeine/morphine group showed significantly increasedstaining compared to controls in the thalamus and hypotha-lamus and by 24 h, the CA1 and superior colliculus alsodisplayed enhanced staining in the combined caffeine/mor-phine group. At 12 h in the caffeine alone treated group, thethalamus, hypothalamus, hippocampus and superior collicu-lus exhibited a significant increase in caspase-3 stainingwhichpersisted for the first 24 h, with the exception of thehippocampus (Fig. 3).

The effect becomes more pronounced at 24 h, particularlywith Fluoro-jade B staining where every region counted shows

a significant difference. Rat pups treated with morphine didnot differ significantly from the saline group at any time point.Interestingly, with Fluoro-jade staining at 24 h, a significantdifference was identified in the thalamus, in the caffeine plusmorphine group (P < 0.05) compared to caffeine alone, withsimilar trends in the CA1 region of the hippocampus, and thecaudate. This suggests that although morphine alone did notresult in significantly increased cell death, the combination ofcaffeine and morphine had an additive effect.

3. Discussion

The purpose of this study was to determine the effects ofcaffeine and morphine alone, and in combination, on theimmature rat brain. Our most important findings relate to thealteration in cell death, seen in the rat pups very shortly afterexposure to caffeine. The lack of weight gain in the first 24 h ofthe caffeine and morphine groups, is of interest, thoughdifficult to explain. Neither the caffeine nor morphine groupsshowed a similarly significant reduction in weight gaincompared to the combination treatment group. However,notably the caffeine group tended to be more active than theother groups, and the morphine group was clearly morelethargic, and fed less well than did the other groups. It ispossible that the combination of these effects was enough tocause a decrease inweight gain for the 24 h following injection.Interestingly, the findings are similar to those observed in thehuman newborn (Schmidt et al., 2006, 2007).

In the first 24h followingadministration, caffeine increasedneurological cell death compared to controls, while morphinealonedidnot significantly affect cell death. The combination ofcaffeine and morphine resulted in an effect that is mostlycomparable to the effect of caffeine alone. In the thalamus,caffeine plus morphine had a combined effect on cell deaththat was additive, in that there was significantly greater celldeath in the combined treatment group compared to caffeinealone.

By 24 h, regional differences with respect to the vulner-ability of the effects of caffeine and caffeine and morphinewere observed. In particular, a significant increase in Fluoro-jade staining was seen in the cortex, caudate nucleus, and CA1

region only when caffeine and morphine were given incombination, and not with either drug alone. Similarly,Fluoro-jade staining of the thalamus at 24 h was significantlygreater following the combination of caffeine/morphine thanwith caffeine alone. These findings do indicate that in certainregions of the brain, an additive effect is seen when the twomedications are used in combination. Nonetheless, morphineadministration alone was insufficient to produce a significantneurotoxic effect in the brain regions examined in our study.

Further to Fluoro-jade B staining, immunohistologicalapplication of cleaved caspase-3 antibody to detect apoptosisalso showed increased positivity in the caffeine and thecaffeine and morphine combination experimental groups.These findings suggest that cell death due to caffeine, orcaffeine andmorphine is at least partially apoptotic in nature.Our findings confirm the previously reported widespreadneurotoxic effects of caffeine administration on apoptoticcell death in the neonatal rat brain, and suggest the addition of

140 B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

morphine may exacerbate these effects in some brain regions(Bakuridze et al., 2005; Kang et al., 2002).

The neonatal brain is well known to be particularly vulner-able to potentially neurotoxic agents and in addition toenhancing cell death, previous research indicates that caffeinehas adverse effects on neonatal morbidity. Although findingsremain controversial, human studies have linked high caffeineintake to low birth weight and an increased risk for fetal loss(Bech et al., 2005; Dlugosz and Bracken, 1992; Martin andBracken, 1987). Caffeine has also been shown, at high doses, tohave amild teratogenic effect in fetal rats (Dlugosz andBracken,1992). Schmidt et al. found human infants receiving caffeine inthe NICU had significantly lower growth rates over the first3 weeks of caffeine administration when compared to controls(Schmidt et al., 2006). Neurobehavioral side effects of caffeineexposure in neonatal rats have also been studied. Chroniccaffeine treatment results in transient motor impairments andalterations in adenosine receptor regulated functions likehyperalgesia, low-anxiety and learning and/or memory relatedtasks (Fisher and Guillet, 1997; Pan and Chen, 2007; Tchekalar-ova et al., 2005). Morphological changes in adenosine receptordensity and dendritic arborization caused by caffeine adminis-tration have been suggested as a basis for differences incognitive function and motor behavior (Etzel and Guillet, 1994;Fisher andGuillet, 1997; Guillet, 1990; Juarez-Mendez et al., 2006;Tchekalarova et al., 2005).

In the brain, caffeine acts predominantly as a competitiveantagonist at A1 and A2A adenosine receptors (Daly andFredholm, 1998). Adenosine receptors (A1) are among the mostwidespreadG-protein coupled receptors in the brain enriched inareas including the hippocampus and thalamic nuclei (Wardas,2002). Adenosine acts to inhibit classical neurotransmitters, butin particular focuses on the excitatory amino acid, glutamate(Dunwiddie and Masino, 2001). A1 receptors are also coupled totheactivationofK+ channels and the inhibition ofCa2+ channels(Dunwiddie and Masino, 2001). A2A adenosine receptors areabundant in the striatum, hippocampus and nucleus accum-bens, functioning mainly in the activation of adenyl cyclase,which synthesizes c-AMP. In addition, they play an importantrole in modulating excitatory glutamate input to the dopami-nergic reward system of the striatum (Dunwiddie and Masino,2001; Rebola et al., 2005; Wardas, 2002). The net effect ofadenosine receptor activation is to reduce excitability andarousal in the brain. Hence, antagonism by caffeine at thesereceptors leads toapositive stimulatory effect. Cleavedcaspase-3 positivity indicates a role for apoptosis in the cell death weobserved. However, the caffeine blockade of adenosine recep-tors leading to an increase of glutamate in the cell implicatesexcitotoxic processes in the neurotoxic effects of caffeineadministration on the neonatal rat brain. In the future, electronmicroscopic examination of the dying cellswould be required todiscern the exact nature of cell death.

The ontogeny of adenosine receptors begins in utero,occurring in a phylogenetic time-dependent fashion fromearliest formation in the brainstem to later development intheouter cortical layersof theneonatalbrain (Gaytanet al., 2006;Herlenius et al., 2002). Distribution and formation of adenosinereceptors in the neonatal rat are not complete in the cerebralcortex until around P30 (Herlenius et al., 2002). Our findingsindicate that themost robust appearance of cell deathdue to the

application of the caffeine treatment is evident in the deep graymatter areas of the neonatal brain, namely the thalamus,hypothalamusandsuperior colliculus. If our findings are relatedto the blockade of adenosine receptors, it is reasonable that atpost-natal day 3, these deep gray areas would have a greaterabundanceandnormal distributionof adenosine receptors thanmore recently evolved brain areas like the cortex. For thisreason, the thalamus, hypothalamus and superior colliculusmay be more vulnerable to the effects of caffeine, and blockadeof adenosine at these receptors.

Morphine is an agonist at opioid receptors, which are mostabundant in the brainstem, thalamus, hypothalamus, hippocam-pus, striatum and regions of the CNS commonly known to beinvolved in the inhibition of pain transmission (Liu et al., 2004).There is limited research on interactions between adenosine andopioid receptors, but it appears that connections between the twoare limited. In the immature brain, adenosine and endogenousopioids work in concert in the brainstem to exert inhibitorycontrol over respiratoryneurons (SimakajornboonandKuptanon,2005). Methylxanthines, caffeine in particular, are used tocounteract this normal inhibition of respiratory neurons andartificially accelerate respiratory maturation until the infant isable to breathe continuously on its own. In theory, morphineshould act preferentially against the effects of caffeine in thebrainstem as endogenous opioids would. In the brain regionsunder investigation in our studyhowever,morphinedidnot havean attenuating effect on the neurotoxicity of caffeine.

Research on morphine in the neonatal brain is scant incomparison to caffeine and indicates that morphine hasextremely variable effects depending on the route of adminis-tration, theareaof thebrain involved, gestational age andwhichreceptor you choose to investigate. While some studies showthat morphine contributes to adverse outcome in pre-matureinfants, others have found neuroprotection (Anand and Hall,2006; Anand et al., 2004; Angeles et al., 2005; Hall et al., 2005; Kimet al., 2001; Nandi and Fitzgerald, 2005). The high level ofinconsistency and controversy in the data suggests multipleactions of morphine on the neonatal brain. Our finding of anincrease in cell death within the thalamus with both caffeineand morphine may simply be the result of a concentration ofopioid receptors in this region (Narita et al., 2008). More focusedresearch studying the effects of morphine alone is requiredbefore conclusionsmay be drawnwith respect to the immatureinfant's developing brain.

The effects of caffeine on the developing brain remaincontroversial. The most recent studies by Schmidt (2005) andSchmidt et al. (2006, 2007) certainly suggest a beneficial effect onthe pre-mature infant with respect to overall death anddisability, andmore specifically on cerebral palsy and cognitivedelay at 18 to 21months of follow up. Differences between thehuman and animal data are not overly surprising. In theaforementioned studies, thepre-maturehuman infant is clearlyexposed to a number of insults which may influence outcome.Importantly, bronchopulmonary dysplasia is one of thosefactors which was also very significantly improved with theuse of caffeine, the results of which likely probably affected theoverall improvement in death and disability. Moreover, onewould not anticipate ‘gross’ brain injury as a result of caffeineuse, the type of whichwould bemanifested as cerebral palsy. Inthis regard, caffeine ismore likely to be subtle in its effects, and

141B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

as such, later results determining mild to moderate learningdifficulties, school readiness, and attentional difficulties, willalso be of interest as they relate to the effects of caffeine use inthe pre-mature infant.

The results of our study suggest that the administration ofcaffeine has a neurotoxic effect on the immature brain. Thefindings are in keepingwith those of Desfrere et al. (2007), whoshowed an alteration of astrogenesis. They did not, however,see an increase in programmed cell death or caspase-3staining. In this study, the authors injected caffeine at lowerdoses and for a longer duration than did our study. It ispossible that these differences are the result of the differencesin dosage and duration. Morphine administration did not, onits own cause an increase in cell death. Though the resultsfrom the human and animal data differ, the complexity ofteasing out the effects of caffeine is more easily done in theanimal model. In this regard, caution should continue to bemaintained when utilizing caffeine in the pre-mature infant.Clearly, further, long-term studies are required to determinethe effects of thesemedications, and on balance, their value inthe NICU setting of the pre-mature baby.

4. Experimental procedures

FemaleLong-Evansrats (CharlesRiverLaboratories)werebred inthe Health Sciences Laboratory Animal facility at the Universityof Alberta. Rat pups were born spontaneously, and reared withtheir dams inconventionalhousing.Ten littersof Long-Evansratpups were used in this study (n=94). All animals weremaintained on a 12-hour light/dark schedule and received foodand water ad libitum throughout the study. All procedures wereapproved by the Health Sciences Animal Policy and WelfareCommittee at the University of Alberta and were in accordancewith the Canadian Council on Animal Care guidelines.

4.1. Experimental paradigm

On post-natal day 3 (P3), rat pups were sexed and randomlyassigned to one of the four treatment groups: Group I —Caffeine + Morphine (C + M); Group II — Caffeine (C) (100mg/kg); Group III — Morphine (M) (10mg/kg); and Group IV —Saline Control (S). Rat pups were utilized at P3 because at thisage they are presumed to be pre-mature at a level of 26–30 week gestation in comparison to the human infant (Backet al., 2007; Back and Rivkees, 2004). Each pup subsequentlyreceived one of the treatments by subcutaneous injection,with all volumes being adjusted to 50μl. The pups wererecovered for 30 min following each injection in an incubatormaintained at 34.5°C (nesting temperature) prior to beingreturned to their dams.

4.2. Histopathology

Pups were euthanized with 5% halothane and decapitated at12 h (n=30), 24 h (n=44) and 72 h (n=20) post-injection. Brainswere immediately extracted, frozen in iso-pentane and storedat − 70°C. Twenty micron coronal sections were taken on acryostat. Anterior sections were taken at the level of theanterior commissure, posterior sections through the mamil-

lary bodies, and through the midbrain. All sections were thenstained with Fluoro-jade B (Chemicon) to detect degeneratingneurons, and cleaved caspase-3 (Cell Signaling Technology) toquantify apoptotic cell death. Previous authors have showncleaved caspase-3 staining to be the most prominent before24 hour post-insult, with diminished appearance subse-quently (Olney et al., 2002). Hence cleaved caspase-3 stainingwas done at only the 12 and 24 hour post-injection timepoints.

Immunostaining was performed using the apoptotic effec-tor antibody, cleaved caspase-3. Fresh frozen sections werepost-fixed in formalin and then cleared and dehydrated in agraded series of ethanol washes. Sections were then washedwith 1% hydrogen peroxide to quench endogenous peroxi-dases and subsequently blocked with normal horse serummixed with Triton X-100 (0.2%). Sections were then allowed toincubate with the primary cleaved caspase-3 antibody over-night. After rinsing and 30 min of incubation with thesecondary anti-rabbit antibody, the sections were rinsedagain and incubated with an avidin–biotin complex (ABC,Vector Laboratories). The immunoreactivity is visualized withdiaminobenzedine tetrahydrochloride (DAB, Sigma).

4.3. Cell counting

Fluoro-jade B stained cells were counted in areas of the cortex,striatum, globus pallidus, thalamus, hypothalamus, nucleusaccumbens, hippocampus and superior colliculus. Cleavedcaspase-3 stained cells were counted in the areas of the cortex,CA1 region of the hippocampus, thalamus, hypothalamus andsuperior colliculus. Regions were chosen based on previousresearch, to assess neural damage over a broad range of brainareas, and in areas most likely to be affected by caffeine ormorphine based on their effect on brain receptors.

Fluoro-jade B stained cells were counted under a fluor-escent microscope. Three different locations from an anteriorsection of the cerebral cortex were counted and an averagewas taken while two locations from each of the thalamus andsuperior colliculus were averaged. One location was countedfor the remaining areas (n≥5 for 12 h and n≥7 for 24 h groups).Neuroanatomical landmarks were established to align thefield of view of themicroscope consistently across sections foreach area counted. Fluoro-jade B positive cells from theappropriate areas were counted at a magnification of 400×,in a 0.2mm2 field of view.

Cleaved capase-3 positive cells were counted from picturesof the sections taken using a Spot Flex 64MB camera and Spotsoftware (Diagnostic Instruments Inc. MI). Only one locationper brain area was counted for this stain due to the extremelyhigh number of positively staining cells (n≥5 for 12 h and n≥7for 24 h groups). Cleaved caspase-3 positive cells were countedfrom a 0.04mm2 field at 400× magnification.

4.4. Statistical analysis

Histological evaluation and cell counting were accomplishedin a blinded fashion, such that there was no previousknowledge by the individual counting, of the experimentalgroup. Between group comparisons were analyzed by analysisof variance (ANOVA) followed by Tukey's post hoc test.

142 B R A I N R E S E A R C H 1 2 1 9 ( 2 0 0 8 ) 1 3 6 – 1 4 2

R E F E R E N C E S

Anand, K.J., Hall, R.W., 2006. Morphine, hypotension, andintraventricular hemorrhage. Pediatrics 117, 250–252 authorreply 252–253.

Anand, K.J., Hall, R.W., Desai, N., Shephard, B., Bergqvist, L.L.,Young, T.E., Boyle, E.M., Carbajal, R., Bhutani, V.K., Moore, M.B.,Kronsberg, S.S., Barton, B.A., 2004. Effects of morphineanalgesia in ventilated preterm neonates: primary outcomesfrom the NEOPAIN randomised trial. Lancet 363, 1673–1682.

Angeles, D.M., Wycliffe, N., Michelson, D., Holshouser, B.A.,Deming, D.D., Pearce,W.J., Sowers, L.C., Ashwal, S., 2005. Use ofopioids in asphyxiated termneonates: effects on neuroimagingand clinical outcome. Pediatr. Res. 57, 873–878.

Back, S.A., Rivkees, S.A., 2004. Emerging concepts inperiventricular white matter injury. Semin. Perinatol. 28, 405–414.

Back, S.A., Riddle, A., McClure, M.M., 2007. Maturation-dependentvulnerability of perinatal white matter in premature birth.Stroke 38, 724–730.

Bakuridze,K.,Duzenli, S.,Gepdiremen,A., 2005.The roleofRutheniumred as a partial agonist in caffeine-induced neurotoxicity incerebellar granular cell culture of rats. Int. J. Neurosci. 115, 13–21.

Bauer, J., Maier, K., Linderkamp, O., Hentschel, R., 2001. Effect ofcaffeine on oxygen consumption and metabolic rate in very lowbirthweight infantswith idiopathic apnea. Pediatrics 107, 660–663.

Bech, B.H., Nohr, E.A., Vaeth, M., Henriksen, T.B., Olsen, J., 2005.Coffee and fetal death: a cohort study with prospective data.Am. J. Epidemiol. 162, 983–990.

Bhutta, A.T., Cleves, M.A., Casey, P.H., Cradock, M.M., Anand, K.J.,2002. Cognitiveandbehavioral outcomesof school-agedchildrenwho were born preterm: a meta-analysis. JAMA 288, 728–737.

Daly, J.W., Fredholm, B.B., 1998. Caffeine — an atypical drug ofdependence. Drug Alcohol Depend. 51, 199–206.

Desfrere, L., Olivier, P., Schwendimann, L., Verney, C., Gressens, P.,2007. Transient inhibition of astrocytogenesis in developingmouse brain following postnatal caffeine exposure. Pediatr.Res. 62, 604–609.

Dlugosz, L., Bracken, M.B., 1992. Reproductive effects of caffeine: areview and theoretical analysis. Epidemiol. Rev. 14, 83–100.

Du,W., Warrier, I., Tutag Lehr, V., Salari, V., Ostrea, E., Aranda, J.V.,2006. Changing patterns of drug utilization in a neonatalintensive care population. Am. J. Perinatol. 23, 279–285.

Dunwiddie, T.V., Masino, S.A., 2001. The role and regulation ofadenosine in the central nervous system. Annu. Rev. Neurosci.24, 31–55.

Etzel, B.A., Guillet, R., 1994. Effects of neonatal exposure to caffeineon adenosine A1 receptor ontogeny using autoradiography.Brain Res. Dev. Brain Res. 82, 223–230.

Fisher, S., Guillet, R., 1997. Neonatal caffeine alters passiveavoidance retention in rats in an age- and gender-relatedmanner. Brain Res. Dev. Brain Res. 98, 145–149.

Gaytan, S.P., Saadani-Makki, F., Bodineau, L., Frugiere, A., Larnicol, N.,Pasaro, R., 2006. Effect of postnatal exposure to caffeine on thepatternof adenosineA1 receptordistribution in respiration-relatednuclei of the rat brainstem. Auton. Neurosci. 126–127, 339–346.

Guillet, R., 1990. Neonatal caffeine exposure alters adenosinereceptor control of locomotor activity in the developing rat.Dev. Pharmacol. Ther. 15, 94–100.

Hall, R.W., Kronsberg, S.S., Barton, B.A., Kaiser, J.R., Anand, K.J.,2005. Morphine, hypotension, and adverse outcomes amongpreterm neonates: who's to blame? Secondary results from theNEOPAIN trial. Pediatrics 115, 1351–1359.

Herlenius, E., Aden, U., Tang, L.Q., Lagercrantz, H., 2002. Perinatalrespiratory control and its modulation by adenosine andcaffeine in the rat. Pediatr. Res. 51, 4–12.

Janvier, A., Khairy, M., Kokkotis, A., Cormier, C., Messmer, D.,Barrington, K.J., 2004. Apnea is associated with

neurodevelopmental impairment in very low birthweight infants.J. Perinatol. 24, 763–768.

Joseph, K.S., Huang, L., Liu, S., Ananth, C.V., Allen, A.C., Sauve, R.,Kramer, M.S., 2007. Reconciling the high rates of preterm andpostterm birth in the United States. Obstet. Gynecol. 109,813–822.

Juarez-Mendez, S., Carretero, R., Martinez-Tellez, R., Silva-Gomez,A.B., Flores, G., 2006. Neonatal caffeine administration causes apermanent increase in the dendritic length of prefrontalcortical neurons of rats. Synapse 60, 450–455.

Kang, S.H., Lee, Y.A., Won, S.J., Rhee, K.H., Gwag, B.J., 2002.Caffeine-induced neuronal death in neonatal rat brain andcortical cell cultures. NeuroReport 13, 1945–1950.

Kim, M.S., Cheong, Y.P., So, H.S., Lee, K.M., Kim, T.Y., Oh, J., Chung,Y.T., Son, Y., Kim, B.R., Park, R., 2001. Protective effects ofmorphine in peroxynitrite-induced apoptosis of primary ratneonatal astrocytes: potential involvement of G protein andphosphatidylinositol 3-kinase (PI3 kinase). Biochem. Pharma-col. 61, 779–786.

Liu, J.G., Rovnaghi, C.R., Garg, S., Anand, K.J., 2004. Opioid receptordesensitization contributes to thermal hyperalgesia in infantrats. Eur. J. Pharmacol. 491, 127–136.

Martin, T.R., Bracken, M.B., 1987. The association between lowbirth weight and caffeine consumption during pregnancy. Am.J. Epidemiol. 126, 813–821.

Montandon, G., Bairam, A., Kinkead, R., 2006. Long-termconsequences of neonatal caffeine on ventilation, occurrence ofapneas, and hypercapnic chemoreflex in male and female rats.Pediatr. Res. 59, 519–524.

Nandi, R., Fitzgerald, M., 2005. Opioid analgesia in the newborn.Eur. J. Pain 9, 105–108.

Narita, M., Hashimoto, K., Amano, T., Narita, M., Niikura, K.,Nakamura, A., Suzuki, T., 2008. Post-synaptic action ofmorphine on glutamatergic neuronal transmission related tothe descending antinociceptive pathway in the rat thalamus.J. Neurochem. 104, 469–478.

Olney, J.W., Tenkova, T., Dikranian, K., Muglia, L.J., Jermakowicz, W.J.,D'Sa, C., Roth, K.A., 2002. Ethanol-induced caspase-3 activation inthe in vivo developing mouse brain. Neurobiol. Dis. 9, 205–219.

Pan, H.Z., Chen, H.H., 2007. Hyperalgesia, low-anxiety, andimpairment of avoidance learning in neonatal caffeine-treatedrats. Psychopharmacology (Berl) 191, 119–125.

Rebola, N., Canas, P.M., Oliveira, C.R., Cunha, R.A., 2005. Differentsynaptic and subsynaptic localization of adenosine A2Areceptors in the hippocampus and striatum of the rat.Neuroscience 132, 893–903.

Schmidt, B., 2005. Methylxanthine therapy for apnea ofprematurity: evaluation of treatment benefits and risks at age5 years in the international caffeine for apnea of prematurity(CAP) trial. Biol. Neonate 88, 208–213.

Schmidt, B., Roberts, R.S., Davis, P., Doyle, L.W., Barrington, K.J.,Ohlsson, A., Solimano, A., Tin, W., 2006. Caffeine therapy forapnea of prematurity. N. Engl. J. Med. 354, 2112–2121.

Schmidt, B., Roberts, R.S., Davis, P., Doyle, L.W., Barrington, K.J.,Ohlsson, A., Solimano, A., Tin, W., 2007. Long-term effects ofcaffeine therapy for apnea of prematurity. N. Engl. J. Med. 357,1893–1902.

Simakajornboon, N., Kuptanon, T., 2005. Maturational changes inneuromodulation of central pathways underlying hypoxicventilatory response. Respir. Physiol. Neurobiol. 149, 273–286.

Tchekalarova, J., Kubova, H., Mares, P., 2005. Postnatal caffeineexposure: effects on motor skills and locomotor activity duringontogenesis. Behav. Brain Res. 160, 99–106.

Wardas, J., 2002. Neuroprotective role of adenosine in the CNS. Pol.J. Pharmacol. 54, 313–326.

Wen, S.W., Smith, G., Yang, Q., Walker, M., 2004. Epidemiology ofpreterm birth and neonatal outcome. Semin. Fetal NeonatalMed. 9, 429–435.