activity of human trypanosome lytic factor in mice

Molecular & Biochemical Parasitology 117 (2001) 129–136

Activity of human trypanosome lytic factor in mice

Chad Barker a, Karen W. Barbour b, Franklin G. Berger b, Stephen L. Hajduk a,*a Department of Biochemistry and Molecular Genetics, Schools of Medicine and Dentistry, Uni�ersity of Alabama at Birmingham, Birmingham,

AL 35394, USAb Department of Biological Sciences, Uni�ersity of South Carolina, Columbia, SC 29036, USA

Received 9 May 2001; accepted in revised form 13 July 2001

Abstract

The inability of the cattle pathogen Trypanosoma brucei brucei to infect humans is due to an innate factor in human serumtermed Trypanosome Lytic Factor (TLF). Human haptoglobin-related protein is the proposed toxin in TLF and can exist eitheras a component of a minor subclass of high-density lipoprotein (TLF-1) or as a lipid free, high molecular weight protein complex(TLF-2). The trypanolytic activity of both TLF-1 and TLF-2 has been studied in vitro but their relative contributions toprotection against T. b. brucei infection in vivo has not been established. In the present studies we show that treatment of T. b.brucei infected mice with TLF-1 resulted in a dose dependent decrease in parasite numbers but did not affect parasite numbersin mice infected with Trypanosoma brucei rhodesiense, the causative agent of the human sleeping sickness. Similarly, pretreatmentof mice with TLF-1 resulted in protection against a challenge by T. b. brucei but had no effect on T. b. rhodesiense challenge.Induction of the acute phase protein haptoglobin, a natural antagonist of TLF-1, diminished but did not abolish the protectionagainst trypanosome challenge. In addition, haptoglobin knockout mice showed higher levels of TLF-1 mediated protectionagainst a T. b. brucei challenge. These results suggest that while TLF-1 is active in vivo, even in the presence of elevated levelsof haptoglobin, its activity is modulated in a dose dependent fashion by haptoglobin in the circulation. © 2001 Elsevier ScienceB.V. All rights reserved.

Keywords: Trypanosoma brucei brucei ; Haptoglobin related protein; Haptoglobin; Acute phase; Knockout mice; High-density lipoprotein;Trypanosome lytic factor

www.parasitology-online.com.

1. Introduction

African trypanosomes infect a wide range of mam-mals including humans, domestic animals and wildgame. Trypanosoma brucei brucei causes a wasting dis-ease, called Nagana, in cattle and though morphologi-cally indistinguishable from the sleeping sicknessparasite Trypanosoma brucei rhodesiense, is unable toinfect humans [1,2]. Innate protection of humansagainst T. b. brucei infection is a consequence of acirculating non-immune trypanosome lytic factor (TLF)[3,4]. Two structurally distinct complexes in humanserum have trypanolytic activity. TLF-1 is a minorsubclass of human high-density lipoprotein (HDL) with

a molecular mass of 500 kDa and a density of 1.21–1.24 g ml−1 [5]. This trypanolytic subspecies of humanHDL contains apolipoprotein AI (apo A-I), and hap-toglobin-related protein (HPR) [6]. TLF-2 is a lipid-freeprotein complex greater than 1000 kDa containing apoA-I, HPR and immunoglobulin M (IgM) [7,8].

Several lines of evidence support the notion thatHPR is the molecule responsible for trypanosome lysis.The HPR gene arose during primate evolution as aresult of the duplication and mutation of a copy of thegene for the acute phase protein haptoglobin [9,10].Analysis of sera from a range of mammals showed thatTLF activity is restricted to primates with the HPRgene. Interestingly, a single point mutation in the HPRgene in chimpanzees correlates with a loss of TLFactivity in chimpanzee sera [11]. Recent studies haveshown that highly purified HPR has low but detectablelytic activity (J. Bishop, University of Alabama atBirmingham, personal communication) and reconstitu-

* Corresponding author. Tel.: +1-205-934-6033; fax: +1-205-975-2547.

E-mail address: [email protected] (S.L. Hajduk).

0166-6851/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.PII: S0166-6851(01)00339-5

C. Barker et al. / Molecular & Biochemical Parasitology 117 (2001) 129–136130

tion of purified HPR, apo A-I and lipid into a lipo-protein complex increased the trypanolytic activity[12]. Finally, HPR is associated with both TLF-1 andTLF-2, which are the only two trypanolytic complexesidentified in human serum [6,13,14].

The relative contribution of TLF-1 and TLF-2 toprotection against T. b. brucei infection in vivo isunclear. Estimates of the relative amount of TLF-1and TLF-2, based on recovery of units of activity,indicates that normal human serum contains 10 �g ofTLF-1 (1000 U) ml−1 and 12 �g of TLF-2 (20 U)ml−1 [8]. This suggests that TLF-1 plays a major rolein human resistance to T. b. brucei infection. However,Raper and co-workers [15] have proposed that TLF-2is the sole trypanolytic molecule in normal humanserum and that TLF-1 is activity is inhibited by thehuman blood protein haptoglobin. The trypanolyticactivities of TLF-1 and TLF-2 are not equally inhib-ited by the acute phase protein haptoglobin [15].When TLF-1 is incubated, in vitro, with human hap-toglobin T. b. brucei lysis is inhibited in a concentra-tion dependent fashion [16,17]. Most importantly, nearcomplete inhibition of TLF-1 was obtained at 0.1 mgml−1 haptoglobin, well below the concentration ofhaptoglobin found in normal human serum. Despitesharing common proteins, HPR and apo A-I it ap-pears that TLF-2 activity in not inhibited by hap-toglobin, bringing into question the relevance ofTLF-1 in vivo[15].

In this paper, we have examined the effects ofTLF-1 on T. b. brucei infection in mice. Treatment ofmice with TLF-1 resulted in a dose dependent resis-tance to parasite challenge. Similarly, treatment of T.b. brucei infected mice with TLF-1 produced a dosedependent inhibition of parasite growth and decline inparasiteamia. The role of serum haptoglobin as apotential modulator of TLF-1 activity in vivo was alsoevaluated. In vitro both mouse and human hap-toglobin had identical TLF-1 inhibitory activities. Weevaluated the effects of low haptoglobin levels onTLF-1 activity using haptoglobin knockout mice andthe effects of high haptoglobin levels in mice undergo-ing an acute phase response. The haptoglobin geneknockout increased the effectiveness of TLF-1 treat-ment of T. b. brucei infected mice. Induction of anacute phase response in mice diminished the effective-ness of TLF-1 treatment, however, significant TLF-1activity remained even at concentrations of hap-toglobin exceeding 2 mg ml−1. Based on the resultsobtained with this mouse model we propose thatTLF-1 is active in vivo and that ahaptoglobinemia,either due to genetic defect or as a result of parasiteinfection, may enhance the activity of TLF-1 in hu-mans.

2. Materials and methods

2.1. Parasites and mouse lines

The bloodstream developmental stage of either T. b.brucei ILTat 1.3 (ILRAD Trypanozoon antigenic type1.3) and a clonal line of T. b. rhodesiense were used inall studies [18]. Parasites were maintained at −196 °Cin 7.5% DMSO and infections in mice were initiatedby intraperitoneal injection of approximately 3×107

parasites. Swiss CD-1 and C57BL/6J female mice,approximately 4–6 weeks old, were used in thesestudies.

Haptoglobin knockout mice were produced as de-scribed earlier [19]. Briefly, a construct containing adominant selectable marker (PGK-neo) in place ofexons 2 and 3 of the mouse haptoglobin gene wasgenerated, and introduced into ES cell line E14. Malechimeric mice from one appropriately targeted ES cellclone were used to produce heterozygous knockoutmice, which were intercrossed to generate ho-mozygotes. The line was verified as nullizygous for thehaptoglobin gene by the presence of the targetedchromosomal gene in DNA, the absence of the wild-type gene, a lack of detectable haptoglobin mRNA inliver and other organs, and a complete absence ofhaptoglobin in the plasma. The knockout mice de-velop normally, appear healthy, exhibit a normallifespan, and are fertile [19].

2.2. In �itro TLF reactions

TLF-1 activity was assayed in vitro as previouslydescribed [5]. Briefly, blood was collected by cardiacpuncture from trypanosome infected mice at 1–5×108

parasites ml−1 of blood. Trypanosomes were sepa-rated from blood constituents by DE52 ion exchangechromatography in bicine buffered saline buffer (BBS)(50 mM bicine, 150 mM NaCl, 1% w/v glucose, 10 Uml−1 sodium heparin). Unbound trypanosomes wereconcentrated by centrifugation (1500×g, 10 min,4 °C), washed with BBS, re-centrifuged and resus-pended at a final concentration of 3×107 try-panosomes ml−1 in lysis assay buffer (F12 media, 10%heat inactivated fetal bovine serum, 1% w/v glucose,10 mM Hepes (pH 7.4)). To each assay, 100 �l oftrypanosome suspension (1×107 trypanosomes) and aspecified amount of purified TLF-1 was added thenbrought to a final volume of 300 �l with phosphatebuffered saline (137 mM NaCl, 2.6 mM KCl, 10 mMNa2HPO4, 1.8 mM KH2PO4) (PBS) containing 0.3mM EDTA (PBSE). Samples were incubated for 2 hat 37 °C and the percentage of lysed trypanosomesdetermined by phase contrast microscopy.

C. Barker et al. / Molecular & Biochemical Parasitology 117 (2001) 129–136 131

2.3. In �i�o acti�ity of TLF in mice

TLF-1 was purified as described previously and thespecific activity of each preparation was established bytitration of purified TLF-1 to establish the concentra-tion necessary to lyse 50% of the trypanosomes in astandard in vitro lysis assay (1 U TLF) [5]. For injec-tion into mice 3, 10, 30, 300 or 600 U of TLF-1 werebrought to a final volume of 200 �l with PBSE. Pre-treated mice were injected intraperitoneally withpurified TLF-1 48 h prior to challenge with 1×107 T.b. brucei. Trypanosome infected mice were treated withTLF-1 after parasite numbers had reached 1×107

ml−1 of blood in the mouse. The effect of a mouseacute phase response on the activity of TLF-1 wasexamined in Swiss CD-1 female mice following irradia-tion (600 rads). Mice were infected 24 h post irradia-tion, parasiteamia allowed to reach approximately 1/107

parasites ml−1 and mice treated with 3, 30 or 300 U ofTLF-1.

2.4. Determination of the stability of TLF in mice

Initially the stability of TLF-1 activity in mice wasestablished by intraperitoneal injection of 300 U ofTLF-1 then sequentially challenging mice at 24 h inter-vals with 1×107 parasites by intraperitoneal injection.Parasite numbers in the circulation were monitoreddaily by microscopic examination of tail bleeds.

To better evaluate the turnover rate of TLF-1 in micethe clearance of 125I-TLF-1 and non-lytic human HDLwas examined. Purified TLF-1 (60 �g) or non-lytichuman HDL (60 �g) was iodinated by the chloramine-T method to a specific activity of 1×105 cpm �g−1 andseparated from unincorporated 125I as previously de-scribed [20]. Iodinated samples (1 �g=2 U of TLF-1)were adjusted to final volume of 200 �l with PBSE andwere injected intraperitoneally into Swiss CD-1 mice.The appearance of iodinated TLF-1 and HDL in theblood was monitored in 5 �l samples of blood takenfrom the tail and counted with a gamma counter. Thedistribution of 125I-TLF-1 and 125I-HDL in HDL andnon-HDL fractions of mouse serum was established inblood samples taken 72 h post injection. Mouse serumwas adjusted to 1.063 g ml−1 with concentrated NaBrsolution and centrifuged at 50000 rpm in a Ti70 rotor(Beckman) for 24 h at 14 °C. The float containingLDLs was removed and counted. The bottom third ofthe tube was then adjusted to 1.26 g ml−1 with thesolution and centrifugation at 50000 rpm continued for42 h at 14 °C. The float containing total serum HDLs(top third of tube) and bottom lipoprotein deficientserum (bottom third of tube) were removed and sam-ples were counted.

2.5. Mouse haptoglobin purification

Swiss CD-1 female mice were irradiated (600 rads),to induce an acute phase response, 36 h prior tocollecting blood. Irradiation results in a 50–500 foldincrease in serum haptoglobin concentrations. Serumwas isolated, a protease inhibitor mixture containingleupeptin (1.5 �g ml−1), PMSF (70 �g ml−1), pepstatin(1 �g ml−1), EDTA (1.5 mM) was added, and thepreparation was fractionated on an anti-human hap-toglobin affinity column in PBSE. Mouse haptoglobinwas eluted with 4 M guanidine and dialyzed againstPBSE. Mouse haptoglobin was further purified byMono-Q (Pharmacia) anion exchange chromatographyin 20 mM bis Tris buffer using a linear NaCl gradientof 0–0.5 M. The mouse haptoglobin eluted between290 and 350 mM NaCl. The purity of the mousehaptoglobin from the Mono-Q column was evaluatedby SDS-PAGE and western blot analysis with anti-hu-man haptoglobin (Sigma). This affinity purificationprocedure yielded approximately 200 �g of mouse hap-toglobin ml−1 of acute phase sera.

2.6. In �itro lysis inhibition assay

Purified mouse haptoglobin was assayed for TLF-1inhibition by addition to a standard TLF assay. Indi-cated amounts of mouse haptoglobin were added to theTLF lysis assays incubated for 2 h at 37 °C and thepercentage of ILTat 1.3 trypanosomes lysed was deter-mined by microscopic examination.

3. Results

3.1. Dose dependent inhibition of trypanosome infectionby TLF-1

Based on in vitro lysis assays, it was recently pro-posed that TLF-1 is inactive in vivo and that the majortrypanolytic activity in humans is associated with TLF-2 [15]. In order to investigate the in vivo activity ofTLF-1, mice were treated with highly purified TLF-1before and after T. b. brucei infection. Similar studieswith purified TLF-2 were not possible due to the insta-bility of this trypanolytic complex.

Swiss CD1 and C57BL/6J mice were treated withdifferent amounts of TLF-1 by intraperitoneal injection48 h prior to challenge with T. b. brucei (Fig. 1A,B).Pretreatment of mice with increasing concentrations ofTLF-1 resulted in longer delays in patency of para-siteamia and death. Both Swiss and C57BL/6J miceresponded similarly to trypanosome challenge followingtreatment with TLF-1 except that at higher concentra-tions of TLF-1 (30 and 300 U) the pre-patent periodwas lengthened in the C57BL/6J mice and at the higher


dosage (300 U) mice were protected from T. b. bruceichallenge (Fig. 1B).

A dose dependent effect was also seen when T. b.brucei infected mice were treated with TLF-1 (Fig.1C,D). Mice were infected by intraperitoneal injectionand parasite numbers were monitored until the para-siteamia reached 1×107 trypanosomes ml−1 in theblood. Mice were then treated with 3–600 U of TLF-1by intraperitoneal injection. The number of parasites inthe circulation decreased dramatically after 2 h and themorphological appearance of the parasites was consis-tent with the pathway of TLF-1 lysis proposed basedon in vitro studies [20]. Although the elapsed time forlysis is longer in vivo, the changing morphology of thetrypanosomes is identical. Treatment of C57BL/6J micewith high concentrations of TLF-1 (600 U) cured T. b.brucei infections (Fig. 1D). Based on these results weconclude that TLF-1 is active in normal mice both indelaying the development of parasiteamia and in treat-ing existing T. b. brucei infections.

3.2. TLF-1 clearance in mice

Despite its in vivo activity, only animals treated withthe highest levels of TLF-1 (600 U) were cured oftrypanosome infections or were completely protectedfrom T. b. brucei challenge. Assuming a rapid absorp-tion of TLF-1 into the mouse circulation a concentra-tion of 600 U ml−1 approaches the level of TLF-1 inthe circulation of humans. The effectiveness of TLF-1

Fig. 2. Stability of TLF-1 following injection into mice. (A) 125I-TLF(closed circles) or non-lytic human HDL (open circles) was injectedinto the peritoneum of mice and blood samples counted to determinethe rate of appearance in the circulation and clearance rates. Low butdetectable levels of iodinated apolipoproteins were detected up to 72h post injection. (B) The trypanolytic activity of the circulatingTLF-1 was evaluated by challenging mice with T. b. brucei 24 h(closed diamonds), 48 h (open circles), 72 h (open triangles), 96 h(closed squares) and 120 h (closed circles) after injection with TLF-1.Growth of T. b. brucei in an untreated mouse (open squares).

when injected into mice may be influenced by the rateof TLF-1 clearance from the circulation following treat-ment. In order to address this possibility we examinedthe clearance rate of 125I-TLF-1 and 125I-non-lytic hu-man HDL in mice (Fig. 2A). When injected into themouse peritoneal cavity both TLF-1 and non-lytic hu-man HDLs quickly equilibrated in the circulationreaching maximum levels approximately 4 h post injec-tion (Fig. 2A). The levels of TLF-1 and non-lytic HDLgradually declined over a 72 h period to approximatelyone-tenth maximal levels.

We next determined whether the apolipoproteins ofTLF-1 remained associated with HDL or exchangedwith other protein complexes within the circulation. Toexamine this possibility mice were sacrificed and bloodcollected 72 h following intraperitoneal injection withTLF-1. When serum HDL was purified by ultracen-trifugation, greater than 70% of the radioactivity frac-tionated with the HDL particles indicating that theTLF-1 particle remained intact prior to clearance bythe mouse (data not shown).

The clearance of TLF-1 from the circulation corre-lated with decreased protection against T. b. bruceichallenge. Mice were treated with 300 U of TLF-1 andchallenged with T. b. brucei 24, 48, 72, 96 or 120 h later(Fig. 2B). The level of protection against trypanosomechallenge decreased with increasing time post treatmentwith TLF-1. These results suggest that while TLF-1 is

Fig. 1. Inhibition of T. b. brucei growth by TLF-1 in mice. Pretreat-ment of (A) CD-1 mice and (B) C57BL/6J mice 48 h prior tochallenge with 3×107 trypanosomes. (C) CD-1 mice or (D) C57BL/6J mice were infected with T. b. brucei (ILTat 1.3) and treated withTLF-1 after parasiteamia reached 1×107 per ml. Dilutions of TLF-1were prepared in PBSE adjusted to a final volume of 200 �l. Controlinjections of 200 �l PBSE (open squares), 3 U TLF-1 (open dia-monds), 10 U TLF-1 (closed circles), 30 U TLF-1 (open circles), 300U TLF-1 (open triangles), 600 U TLF-1 (closed squares).


active in mice its rapid clearance from the circulationdiminishes its curative and protective activities.

3.3. Inhibition of TLF-1 acti�ity by mouse haptoglobin

Results presented here indicate that purified TLF-1retains its trypanolytic activity when injected into nor-mal mice. In these animals the levels of the acute phaseprotein, haptoglobin, ranges from 10 to 200 �g ml−1.Induction of an acute phase response in these animalsresults in a 10–50-fold increase in haptoglobin in the

blood (Fig. 3C). To test whether TLF-1 was active, invivo, in the presence of elevated levels of haptoglobin,CD-1 mice were irradiated (600 rads) 24 h prior totreatment with TLF-1 (30 or 300 U) and subsequentlychallenged with T. b. brucei 24 h post TLF-1 treatment(Fig. 3A). The effects of TLF-1 on T. b. brucei infectiondecreased in acute phase mice consistent with the in-hibitory activity of human haptoglobin in vitro [16].However, treatment of acute phase mice with 300 U ofTLF-1 significantly delayed parasite growth relative tountreated acute phase mice challenged with T. b. brucei.This suggests that TLF-1 is active even in the presenceof the inhibitor haptoglobin and plays a major role inprotection against T. b. brucei infection. Mouse hap-toglobin is 80% homologous to human haptoglobin, isrecognized by polyclonal antibodies against the humanprotein and has similar characteristics as an acute-phase reactant protein [21,22]. However, it remainedpossible that the mouse model did not accurately mimicthe situation in humans. Of particular concern waswhether mouse haptoglobin was an effective inhibitorof TLF-1 mediated lysis.

To test this possibility, the inhibitory activity ofmouse haptoglobin for TLF-1 was examined by addingmouse sera containing different amounts of hap-toglobin to an in vitro TLF-1 lysis assay (Fig. 3B).Induction of acute phase response in mice, by irradia-tion 24 h prior to collection of serum results in a10–50-fold increase in the levels of serum haptoglobin(Fig. 3C, lanes 3 and 4). Sera from the irradiated miceinhibited TLF-1 lysis to a greater extent than sera fromnon-irradiated animals. While these results are consis-tent with mouse haptoglobin inhibiting TLF-1 in vitroa number of serum proteins are increased during anacute phase response [23].

In order to directly determine its inhibitory activity,mouse haptoglobin was purified and tested for TLF-1inhibition. Normal Swiss CD-1 mice were irradiated toinduce haptoglobin expression and blood was collected36 h later. Mouse haptoglobin was purified by affinitychromatography with anti-human haptoglobin andMono-Q chromatography (Fig. 4A). Purified mousehaptoglobin inhibits TLF-1 activity in vitro with anIC50 of approximately 9 �g ml−1 (Fig. 4B). This isapproximately equivalent to the IC50 for human hap-toglobin (6 �g ml−1). Inhibition of lysis with mousehaptoglobin is dose dependent suggesting that bothhuman and mouse haptoglobin reduce TLF-1 lysis by asimilar mechanism.

These results demonstrate that mouse haptoglobinhas similar inhibitory activity to human haptoglobin invitro and suggests that the in vivo lytic activity ofTLF-1 in mice mimics that expected in humans. Whilehaptoglobin is an effective inhibitor of TLF-1 activityin vitro, significant trypanolytic activity remains inmice, even in acute phase animals expressing elevated

Fig. 3. Acute phase mice have elevated haptoglobin levels and are lessresponsive to TLF-1 treatment. Irradiation of mice induces an acutephase response and increases the inhibitory activity of mouse sera forTLF-1. (A) Normal CD-1 mice (closed symbols) and irradiated (600rads) mice (open symbols), were treated 24 h later after irradiationwith either 30 U (circles) or 300 U (squares) of TLF-1 and subse-quently challenged 24 h later with T. b. brucei. Untreated, non-irradi-ated mice infected with T. b. brucei were monitored as a control(broken line). (B) Effect of acute phase serum on TLF-1 mediatedlysis in vitro. Parasite lysis was determined following addition dilu-tions of serum from irradiated (closed circles and squares) or non-ir-radiated mice (open circles and squares) to in vitro assays with T. b.brucei (1×107) and TLF-1 (2 U). (C) Serum samples were run onSDS-PAGE and analyzed by western blot with an antibody againsthuman haptoglobin that cross-reacts with mouse haptoglobin. Serafrom non-irradiated mice (lanes 1 and 2) contained low levels ofhaptoglobin. Sera taken 24 h post-irradiation (lanes 3 and 4) showedan increase in haptoglobin. Serum from mouse c1 (open squares)had slightly elevated haptoglobin levels consistent with the intermedi-ate level of inhibition seen.


Fig. 4. Purification of mouse haptoglobin from acute phase mice. (A)Mouse haptoglobin was purified by affinity and anion exchangechromatography from the serum of irradiated mice. The samples ofmouse and human haptoglobin were run under reducing conditionson SDS-PAGE and western blotted with anti-human haptoglobin.The major proteins in the mouse and human haptoglobin prepara-tions reacted with the anti-haptoglobin. The positions of the � and �sub-units of human and mouse haptoglobin are indicated. The �subunit of human HPR is also indicated (�-HPR). (B) Inhibition ofTLF-1 by mouse and human haptoglobin. Increasing amounts ofhuman and mouse haptoglobin were added to TLF-1 (1 �g=2 U). T.b. brucei (1×107) were then added, incubated for 2 h at 37 °C andthe percentage of lysed trypanosomes determined.

induction of the acute phase response while no hap-toglobin was detected in the serum from haptoglobingene knockout animals (Fig. 5A). Serum from normaland haptoglobin null mice was tested for inhibition ofTLF-1 activity in vitro. Irradiated and non-irradiatedC57BL/6J mouse sera inhibited TLF-1 in a dose depen-dent fashion while serum from haptoglobin knockoutmice failed to inhibit lysis (Fig. 5B). These results showthat haptoglobin and not other acute phase proteins areresponsible for inhibition of TLF-1 activity in vitro.

To directly examine the in vivo effects of haptoglobinon the activity of TLF-1 normal (Fig. 5C) and hap-toglobin null mice (Fig. 5D) were infected with T. b.brucei, and treated with TLF-1 when parasiteamiareached 1×107 per ml. Treatment of normal mice withTLF-1 significantly reduced parasite numbers in theblood with the highest concentrations clearing parasitesfrom the blood for more than 5 days. TLF-1 treatmentof T. b. brucei infected haptoglobin null mice extendedthe pre-patent period following TLF-1 treatment rela-tive to normal mice at each concentration (Fig. 5D).Haptoglobin null mice treated with 100 U of TLF-1failed to develop parasiteamia up to two weeks post-

Fig. 5. Haptoglobin gene knockout mice have enhanced responsive-ness to TLF-1 treatment of T. b. brucei infection. (A) Western blotanalysis of serum from wild-type C57BL/6J mice and haptoglobingene knockout mice. Sera from non-irradiated (lanes 1, 2, 5, 6) andirradiated (lanes 3, 4, 7, 8) mice was analyzed by western blot withanti-human haptoglobin. Low levels of haptoglobin were present inwild-type, non-irradiated mice and increased following irradiation.No haptoglobin was detected in the sera of haptoglobin gene knock-out animals. (B) Parasite lysis was determined following additiondilutions of serum from irradiated wild-type (closed circles andsquares) or irradiated haptoglobin gene knockout mice (open circlesand squares) to in vitro assays with T. b. brucei (1×107) and TLF-1(2 U). (C) Post-treatment of C57BL/6J mice (open symbols) and (D)haptoglobin gene knockout mice (closed symbols). T. b. brucei in-fected mice were treated with TLF-1 when parasite numbers reached1×107 ml−1. Dilutions of TLF-1 were prepared in PBSE adjusted toa final volume of 200 �l. Control injections 200 �l PBSE into wildtype (broken line) and haptoglobin knockout mice (diamonds), 3 UTLF-1 (circles), 10 U TLF-1 (inverted triangles), 30 U TLF-1 (trian-gles), 100 U TLF-1 (squares).

levels of haptoglobin, supporting the idea that TLF-1plays a significant role in protection against T. b. bruceiinfection.

3.4. TLF-1 acti�ity is enhanced in haptoglobin nullmice

Haptoglobin deficiencies are widespread in Africadue to malaria and other infectious diseases [24]. Adecrease in haptoglobin levels may provide enhancedprotection against challenge by T. b. brucei. The de-creased activity of TLF-1 in acute phase mouse serasuggests that modulation of haptoglobin levels in micemimics the predicted inverse correlation of trypanolyticactivity in human serum and haptoglobin levels [16]. Inorder to determine whether haptoglobin deficiency infl-uences the activity of TLF-1, we examined the activityof TLF-1 both in vivo and in vitro with irradiated andnon-irradiated C57BL/6J mice and haptoglobin nullmice (Fig. 5A and B). Serum haptoglobin levels wereevaluated by western blot analysis with antibodiesagainst human haptoglobin. Haptoglobin levels in-creased in the normal mice following irradiation due to


treatment (closed squares). These studies show thathaptoglobin inhibits TLF-1 activity in vivo and invitro and that haptoglobinanemia may influence theprotection TLF-1 provides against challenge with T. b.brucei.

4. Discussion

We have examined the trypanolytic activity of TLF-1, a subclass of human HDL containing apolipo-proteins AI and HPR, in mice. These studies showthat TLF-1 is active against T. b. brucei infections inmice and that TLF-1 treatment of mice extends thepre-patent period following challenge by T. b. brucei(Figs. 1 and 2). These results support the possibilitythat TLF-1 is an important factor in protectionagainst trypanosome infection in humans despite hap-toglobin inhibition in vitro [16]. We have also used amouse model to examine the effects of haptoglobinlevels on TLF-1 activity in animals. Purified mouseand human haptoglobin inhibits TLF-1 with nearlyidentical dose response in vitro (Fig. 4). Elevating hap-toglobin levels in mouse serum by induction of acutephase response decreased the both the in vivo and invitro activity of TLF-1 but failed to completely abro-gate its trypanolytic activity (Fig. 3). Previous predic-tions that ahaptoglobinemia would increase thepotency of TLF-1 in humans were supported by re-sults obtained with haptoglobin gene knockout mice inwhich the activity of TLF-1 was significantly enhancedrelative to normal mice (Fig. 5).

Raper and co-workers have proposed that the lipid-free apo A-I/HPR protein complex, TLF-2, is theprincipal trypanolytic factor in normal human serum[15]. Their conclusion was based on the findings that,(1) TLF-1 was inhibited by human haptoglobin invitro whereas TLF-2 was not; (2) addition of largeamounts of haptoglobin to normal human serumfailed to inhibit trypanolytic activity. These studiesand our own work have established that human hap-toglobin inhibits TLF-1 activity in vitro in a dosedependent fashion [15,16]. However, these in vitrostudies are based on a relatively short exposure toTLF-1 (2–3 h) and may not accurately mimic thesituation in the circulation. Treatment of mice withTLF-1 prior or post-infection with T. b. brucei signifi-cantly reduced parasite numbers. Concentrations ofTLF-1 similar to those normally found in humanserum (10–1000 U/ml) increased the pre-patent periodfor T. b. brucei challenged mice and reduced parasitenumbers in infected mice. Treatment of the humansleeping parasite T. b. rhodesiense infected mice withTLF-1 had no affect on parasiteamia, supporting thespecificity of TLF-1 in vivo (data not shown). Ourstudies do not address the activity of TLF-2 in the

mouse model since purification of TLF-2 to homo-geneity has not been possible due to the instability ofthe complex.

Our results indicate that the concentration of hap-toglobin in the circulation of mice influences the activ-ity of TLF-1. We have previously shown that thetrypanolytic activity of normal human serum is in-versely proportional to the concentration of hap-toglobin [16]. The identification of haptoglobin as anatural antagonist for TLF-1 activity in vitro is notdisputed. However, the influence of haptoglobin inhi-bition on trypanosome infections was unclear. Raperand co-workers argue that haptoglobin levels do notinfluence the trypanolytic activity of intact serum be-cause TLF-2 is not inhibited by haptoglobin [15]. Ourfindings show that at normal concentrations of hap-toglobin TLF-1 is active against trypanosomes in thecirculation of mice.

We have recently reported that T. b. brucei bindshaptoglobin [25]. These studies further suggest thatTLF-1 binds to a haptoglobin receptor in the flagellarpocket and that binding of either TLF-1 or hap-toglobin to the haptoglobin receptor results in endocy-tosis and lysosomal targeting in T. b. brucei. This ledus to propose that inhibition of TLF-1 killing of try-panosomes by haptoglobin is a consequence of hap-toglobin competition for TLF-1 binding. In thesestudies we also found that purified HPR competes forTLF-1 binding indicating that HPR is the ligand forthe TLF-1 receptor in T. b. brucei. The reported lackof haptoglobin inhibition for TLF-2 killing in vitromay be a consequence of lower affinity of TLF-2associated HPR for the haptoglobin receptor.

In haptoglobin null mice the trypanolytic activity ofTLF-1 is enhanced (Fig. 5). These results may berelevant to conditions in human populations sinceahaptoglobinemia is common in Africans and islargely a consequence of chronic malaria infection [26].The relative contribution of TLF-1 in human resis-tance to trypanosome infection may be further en-hanced since patients with paroxysmal nocturnalhemoglobinuria have very low levels of circulatinghaptoglobin and little or no TLF-2 [15]. Serum fromthese individuals is highly trypanolytic presumably dueto TLF-1 activity.

The importance of TLF-1 in innate protection fromT. b. brucei infection has been questioned [15]. Ourstudies demonstrate that TLF-1 is active in mice evenin the presence of high concentrations of haptoglobin.We further show that treatment of T. b. brucei in-fected haptoglobin gene knockout mice with TLF-1results in enhanced clearance of the parasites and de-lays or prevents infection. These results support a rolefor TLF-1 in protection against T. b. brucei infectionin humans and suggests that the relative levels of TLFand haptoglobin might influence an individuals suscep-tibility to trypanosome infection.


Acknowledgements

We would like to thank members of the Hajduk lab,especially, Rusty Bishop, Jerome Drain, Masako Shi-mamura, Monika Oli and Sara Faulkner for helpfulcomments on this work. We also thank members of theBiology of Parasitism course at the Marine BiologicalLaboratories (1995 and 1996) for initial studies onTLF-1 in animals. These studies were support by grantsfrom the National Institutes of Health (AI39033) toSLH, and the UNDP/World Bank/WHO special pro-gramme for research and training in tropical diseases toSLH.

References

[1] Hajduk SL, Hager KM, Esko JD. Human high-denstiy lipo-protein killing of African trypanosomes. Ann Rev Microbiol1994;48:139–62.

[2] Tomlinson S, Raper J. Natural human-immunity to try-panosomes. Parasitol Today 1998;14:354–9.

[3] Rifkin MR. Identification of the trypanocidal factor in normalhuman serum: high-density lipoprotein. Proc Natl Acad Sci USA1978;75:3450–4.

[4] Hawking F. The differentiation of Trypanosoma rhodesiensefrom T. brucei by means of human serum. Trans R Soc TropMed Hyg 1973;67:517–27.

[5] Hajduk SL, Moore DR, Vasudevaharya J, Siqueira H, Torri AF,Tytler EM, Esko JD. Lysis of Trypanosoma brucei by a toxicsubspecies of human high density lipoprotein. J Biol Chem1989;264:5210–7.

[6] Smith AB, Esko JD, Hajduk SL. Killing of trypanosomes byhaptoglobin related protein. Science 1995;268:284–6.

[7] Tomlinson S, Jansen A-M, Koudinov A, Ghiso JA, Choi-MiuraN-H, Rifkin MR, Ohtaki S, Nusenzweig V. High-density lipo-protein independent killing of Trypanosoma brucei by humanserum. Mol Biochem Parasitol 1995;70:131–8.

[8] Raper J, Fung R, Ghiso J, Nussenzweig V, Tomlinson S. Char-acterization of a novel trypanosome lytic factor from humanserum. Infec Immun 1999;67:1910–6.

[9] Maeda N. Nucleotide sequence of the haptoglobin and hap-toglobin-related gene pair. J Biol Chem 1985;260:6698–709.

[10] McEvoy SM, Maeda N. Complex events in the evolution of thehaptoglobin gene cluster in primates. J Biol Chem1988;263:15740–7.

[11] Seed JR, Sechelski JB, Loomis MR. A survey for a trypanocidalfactor in primate sera. J Protozool 1990;37:393–400.

[12] Tytler EM, Moore DR, Pierce MA, Hager KM, Esko JD,Hajduk SL. Reconstitution of the trypanosome lytic factor. Therole of lipids and apolipoproteins in the cytotoxicity of a subspe-cies of human high density lipoproteins. Mol Biochem Parasitol1995;69:9–17.

[13] Tomlinson S, Muranjan M, Nussenzweig V, Raper J. Hap-toglobin related protein and apolipoprotein AI are componentsof the two trypanolytic factors in human serum. Mol BiochemParasitol 1997;86:117–20.

[14] Muranjan M, Nussenzweig V, Tomlinson S. Characterization ofthe human serum trypanosome toxin, haptoglobin-relatedprotein. J Biol Chem 1998;273:3884–7.

[15] Raper J, Nussenzweig V, Tomlinson S. The main lytic factor ofTrypanosoma brucei brucei in normal human serum is not high-density lipoprotein. J Exp Med 1996;183:1023–9.

[16] Smith AB, Hajduk SL. Identification of haptoglobin as a naturalinhibitor of trypanocidal activity in human serum. Proc NatlAcad Sci USA 1995;92:10262–6.

[17] Raper J, Nussenzweig V, Tomlinson S. Lack of correlationbetween haptoglobin concentration and trypanolytic activity ofnormal human serum. Mol Biochem Parasitol 1996;76:337–8.

[18] Hager KM, Hajduk SL. Mechanism of resistance of Africantrypanosomes to cytotoxic human HDL. Nature 1997;385:823–6.

[19] Lim S-K, Kim H, Lim SK, Ali A, Lim YK, Wang Y, ChongSM, Constantini F, Baumann H. Increased susceptibility in Hpknockout mice during acute hemolysis. Blood 1998;92:1870–7.

[20] Hager KM, Pierce MA, Moore DR, Tytler EM, Esko JD,Hajduk SL. Endocytosis of a cytotoxic human high-densitylipoprotein results in disruption of acidic intracellular vesiclesand subsequent killing of African trypanosomes. J Cell Biol1994;126:155–67.

[21] Yang F, Lineham LA, Friedrichs WE, Lalley PA, SakaguchiAY, Bowman BH. Characterization of the mouse haptoglobingene. Genomics 1993;18:374–80.

[22] Shapiro SZ, Black SJ. Identification of an acute-phase reactantin murine infections with Trypanosoma brucei. Infect Immun1992;60:3921–4.

[23] Gabay C, Kushner I. Mechanisms of disease: acute-phaseproteins and other systemic responses to inflamation. New EnglJ Med 1999;340:448–54.

[24] Giblett ER. Haptoglobin types in American negroes. Nature1959;183:192–3.

[25] Drain J, Bishop JR, Hajduk SL. Haptoglobin related proteinmediates trypanosome lytic factor binding to trypanosomes. JBiol Chem 2001;30254–30260.

[26] Trape JF, Fribourg-Blanc A. Ahaptoglobinemia in African pop-ulations and its relation to malaria endemicity. Am J Epidemol1988;127:1282–8.

activity of human trypanosome lytic factor in mice

Documents