Cloning of Syrian Hamster (Mesocricetus auratus) Cytokine cDNAs and Analysis of Cytokine mRNA Expression in Experimental Visceral Leishmaniasis

Abstract

The Syrian golden hamster (Mesocricetus auratus) is uniquely susceptible to a variety of intracellular pathogens and is an excellent model for a number of human infectious diseases. The molecular basis for this high level of susceptibility is unknown, and immunological studies related to this model have been limited by the lack of available reagents. In this report we describe the cloning and sequence analysis of portions of the Syrian hamster interleukin 2 (IL-2), IL-4, gamma interferon (IFN-γ), tumor necrosis factor alpha, IL-10, IL-12p40, and transforming growth factor β cDNAs. In addition, we examined the cytokine response to infection with the intracellular protozoan Leishmania donovani in this animal model. Sequence analysis of the hamster cytokines revealed 69 to 93% homology with the corresponding mouse, rat, and human nucleotide sequences and 48 to 100% homology with the deduced amino acid sequences. The hamster IFN-γ, compared with the mouse and rat homologs, had an additional 17 amino acids at the C terminus that could decrease the biological activity of this molecule and thus contribute to the extreme susceptibility of this animal to intracellular pathogens. The splenic expression of these genes in response to infection with L. donovani, the cause of visceral leishmaniasis (VL), was determined by Northern blotting. VL in the hamster is a progressive, lethal disease which very closely mimics active human disease. In this model there was pronounced expression of the Th1 cytokine mRNAs, with transcripts being detected as early as 1 week postinfection. Basal expression of IL-4 in uninfected hamsters was prominent but did not increase in response to infection with L. donovani. IL-12 transcript expression was detected at low levels in infected animals and paralleled the expression of IFN-γ. Expression of IL-10, a potent macrophage deactivator, increased throughout the course of infection and could contribute to the progressive nature of this infection. These initial studies are the first to examine the molecular immunopathogenesis of a hamster model of VL infection and indicate that progressive disease in this model of VL is not associated with early polarization of the splenic cellular immune response toward a Th2 phenotype and away from a Th1 phenotype.

The Syrian golden hamster has proven to be an excellent experimental model for a number of human infectious diseases, including syphilis, leishmaniasis, and mycobacterial, fungal, and arboviral infections (for a review, see reference 13). For a number of these pathogens (e.g., Treponema pallidum and Leishmania [Viannia] spp.), other suitable animal models are not available. The Syrian hamster is highly susceptible to many intracellular organisms and as such has been used as an experimental host for the isolation of a number of human pathogens. The reason for this extreme susceptibility is unknown, and because of a lack of reagents, substantive molecular immunological studies of these models of infectious diseases have not been undertaken.

To better understand this animal’s immune response to important pathogens, we determined the nucleotide sequence of hamster cytokine genes and characterized their expression in a model of infection. We isolated and cloned the Syrian hamster interleukin 2 (IL-2), IL-4, gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), IL-10, IL-12p40, and transforming growth factor β (TGF-β) cDNAs, as well as the constitutively expressed hypoxanthine phosphoribosyltransferase (HPRT) gene. We used cross-species DNA amplification (reverse transcription [RT]-PCR) techniques (31) using primers which targeted regions of similarity found among the published gene sequences in other species. After nucleotide sequence analysis of the cloned cDNAs we characterized cytokine gene expression in the hamster model of visceral leishmaniasis (VL). Systemic infection of the hamster with Leishmania donovani results in a relentlessly increasing visceral parasite burden, progressive cachexia, hepatosplenomegaly, pancytopenia, hypergammaglobulinemia, and, ultimately, death (13, 15). These clinicopathologic features closely mimic active human VL. This is in contrast to the murine model of VL, in which infection is controlled without the development of overt clinical disease. Our results, which comprise the first description of the molecular immunopathogenesis of disease in the hamster, do not support the paradigm that progressive, uncontrolled leishmanial infection is determined by the induction of a strong Th2 cytokine response and the absence of a Th1 response (20). These studies provide the foundation to further dissect the immunological mechanisms of progressive disease in this unique model and may offer important insights into human disease.

MATERIALS AND METHODS

Hamsters.

Six- to eight-week-old outbred Syrian golden hamsters (Mesocricetus auratus) were obtained from Charles River Laboratories and maintained in a specific-pathogen-free facility. Animals were handled according to local and federal regulations, and research protocols were approved by our Institutional Animal Care and Use Committee.

Spleen cell culture.

Hamster spleen cells were isolated by passage of the organs through a wire screen and then nylon mesh. The erythrocytes were lysed in 0.83% ammonium chloride in 0.01 M Tris HCl, and the remaining cells were washed in RPMI medium. The spleen cells were cultured in RPMI medium containing 10% heat-inactivated calf serum (Hyclone), 50 μg of gentamicin per ml, 1 mM glutamine, and 25 mM HEPES at 10⁶ cells/ml in a 5% CO₂ atmosphere at 37°C in the presence or absence of phorbol myristate acetate (10 ng/ml) and ionomycin (500 ng/ml) for 6 to 8 h or 5 mM concanavalin A for 24 h prior to isolation of the RNA.

Oligonucleotide primers.

Oligonucleotide primers specific for the HPRT gene were designed from the published Chinese hamster HPRT cDNA sequence (28) with the assumption that there would be a high level of sequence homology. The primers for amplification of cytokine genes were designed from regions of homology found among the corresponding published human, mouse, rat, and gerbil cDNA sequences. In most instances, the use of multiple primer combinations was required to successfully amplify a specific product of the appropriate size. The oligonucleotides were synthesized on a DNA synthesizer (Applied Biosystems, Foster City, Calif.) and purified by reverse-phase high-performance liquid chromatography. Degenerate primers were used when there was incomplete homology among the published sequences and when nondegenerate primers failed to yield an amplification product. The sequences of the primers used to successfully amplify cytokine or HPRT cDNAs are as follows: IL-2: forward, ATGTACAGCAKGCAGCTCGC; reverse, TGTTGAGATGRYRCTTTGAC; IL-4: forward, CATTGCATYGTTAGCRTCTC; reverse, TTCCAGGAAGTCTTTCAGTG; IL-10: forward, ACAATAACTGCACCCACTTC; reverse, AGGCTTCTATGCAGTTGATG; IL-12: forward, GTACACCTGYCACAAAGGAG; reverse, GATGTCCCTGATGAAGAAGC; IFN-γ: forward, GGATATCTGGAGGAACTGGC; reverse, CGACTCCTTTTCCGCTTCCT; 5′ untranslated region (UTR), TAGARRAGAMASATCAGYYRA; 3′ UTR, GYCTKGSYKAATTAGTCAGAAA; TNF-α: forward, GACCACAGAAAGCATGATCC; reverse, TGACTCCAAAGTAGACCTGC; TGF-β: forward, CCCTGGAYACCAACTATTGC; reverse, ATGTTGGACARCTGCTCCAC; HPRT: forward, GACAGGACTGAAAGACTTC; reverse, ATCCAACACTTCGAGAGGTC. Degenerate bases are indicated by the appropriate International Union of Pure and Applied Chemistry (IUPAC) single-letter designation (K = G or T, Y = C or T, R = A or G, S = G or C, M = A or C).

Cloning of cytokine and HPRT cDNAs.

RNA isolation and RT-PCR were performed as previously described (35). Total RNA was isolated from stimulated and unstimulated spleen cells with lysis buffer containing guanidinium isothiocyanate (Ultraspec; Biotecx, Friendswood, Tex.) according to the manufacturer’s instructions. One microgram of total RNA was reverse transcribed into cDNA with Moloney murine leukemia virus reverse transcriptase (Gibco BRL) and random hexamers. Amplification of specific cDNA was attempted with multiple different primer pair combinations. Amplification was performed initially at a low annealing temperature (42°C), with extension at 70°C for 2 min and denaturation at 95°C for 1 min. Taq polymerase was used, except in the case of the full-length IFN-γ cDNA, for which a mixture of Taq and Pfu polymerases (15:1) was used. Reactions that produced amplification products were repeated at incrementally (4°C) higher annealing temperatures (up to 55°C) until a single product was obtained. The amplified DNA was cloned if it had a size similar to that predicted from the published homologous sequences, if it was the predominant amplification product of the reaction, and, in the case of the cytokines, if its expression was increased in cDNA derived from the activated cells. In most instances it was necessary to use multiple primer combinations to generate an amplification product meeting these criteria. The amplification product was cloned by ligation into the pCRII plasmid (Stratagene) and transformed into competent Escherichia coli DH5α cells according to the manufacturer’s instructions. The recombinant plasmids were isolated by alkaline lysis and purified by cesium chloride gradient centrifugation. The DNA insert was sequenced with vector-specific primers and an automated, fluorescent DNA sequencer (model 373A; Applied Biosystems). The resulting sequences were identified by a search of the NCBI databases for homologous sequences that used BLAST (1). Sequence comparisons were conducted with the Genetics Computer Group (GCG) package, which makes optimal alignments by inserting gaps to maximize the number of matches according to the local homology algorithm of Smith and Waterman. For amino acid alignment, the gap creation penalty was set at 3.0 and the gap extension penalty was set at 0.10. For nucleotide alignment, these values were set at 5.0 and 0.30, respectively.

Parasites and infection.

Hamsters were infected with L. donovani (1S strain) amastigotes, which were maintained by serial passage in hamsters. Hamsters were infected by intracardial inoculation with 5 × 10⁶ purified amastigotes (37). At 7, 14, and 28 days postinfection the animals were euthanized, the hepatic parasite burden was determined with an impression smear (51), and the spleens were harvested and snap frozen in liquid nitrogen.

Analysis of cytokine gene expression.

The in situ splenic cytokine expression in uninfected and L. donovani-infected hamsters was analyzed by Northern blotting. Total RNA was extracted from the frozen spleens with acid-guanidinium isothiocyanate-phenol-chloroform (5), and Northern blotting was performed as previously described (4). RNA (30 μg) was separated on formaldehyde-agarose gels, electroblotted onto a nitrocellulose membrane (Schleicher and Schuell, Inc., Keene, N.H.), and cross-linked by UV light (Stratalinker 2400; Stratagene, La Jolla, Calif.). The blot was prehybridized in standard prehybridization buffer, and the blots were then hybridized at 42°C for 16 h with a [α-³²P]dCTP-labeled cDNA probe (6 × 10⁵ cpm/ml). The blots were washed and then exposed at −80°C to Kodak XAR-5 film with Kodak intensifying screens. The probes used for the Northern blotting were obtained by digestion of the recombinant vector with EcoRI, and the cloned insert was separated from the vector fragment by agarose gel electrophoresis. The cloned hamster gene fragments were extracted from the gel and used as probes as described above. Hybridization with the HPRT probe was used to assess loading equivalency and RNA integrity.

To quantify the intensity of the autoradiographic signals of the Northern blots, we scanned the autoradiograms with standard video imaging equipment connected to a Macintosh Power personal computer and analyzed the image with an NIH 1.59 image analysis software package with an integrated density program (32). The area analyzed for each band was kept constant, and the background density of the autoradiogram was subtracted from the densitometric data for each band. The results were expressed as a ratio of cytokine expression to HPRT expression to control for equivalent loading and transfer of RNA.

Nucleotide sequence accession numbers.

The GenBank accession numbers for the hamster cytokine and HPRT cDNAs are as follows: IFN-γ, AF034482; IL-2, AF046212; IL-4, AF046213; IL-10, AF046210; IL-12p40, AF046211; TNF-α, AF046215; TGF-β, AF046214; and HPRT, AF047041.

RESULTS AND DISCUSSION

The molecular basis for the extreme susceptibility of the Syrian hamster to intracellular pathogens is unknown. We reasoned that the cytokines which promote the expansion of Th1 cells (IL-2, IFN-γ, and IL-12) and/or induce macrophage activation (IFN-γ and TNF-α) could have unique structural features which could cause a diminished biological effect or could be expressed in vivo at a low level to account for this susceptibility. Alternatively, the prominent expression of cytokines which promote Th2 cell expansion (IL-4 and IL-10) or suppress macrophage activation (IL-4, IL-10, and TGF-β) could lead to a permissive host response. To address this issue, we cloned and sequenced the hamster IL-2, IL-4, IFN-γ, TNF-α, IL-10, IL-12p40, TGF-β, and HPRT genes. The genes were cloned after amplification by RT-PCR with primers designed from regions of homology in published sequences from other species. A high level of sequence conservation of the cloned cDNAs confirmed their identity. In most cases most of the coding region of the cDNA was obtained. The sequences were compared to the published mouse, rat, and human sequences to identify common and/or unique features which could relate to the immunobiology of this animal. The deduced amino acid sequences of the cloned cytokine genes, and their alignment with the mouse, rat, and human sequences, are shown in Fig. 1. In general, the cloned hamster gene sequences showed the greatest homology to the published rat sequences and the homology to rat and mouse sequences was significantly greater than the homology to the human sequences (Table 1).

FIG. 1.

Alignment of the deduced amino acid sequences of the cloned hamster (Ha) cytokine cDNAs with those of the mouse (M), rat (R), and human (Hu) homologs. (a) IFN-γ. (b) IL-2. (c) IL-4. (d) TNF-α. (e) IL-10. (f) IL-12p40. (g) TGF-β. The schematic at the top of each sequence shows the full coding region based on the mouse sequence (approximate scale). The solid vertical lines indicate the exons of the mouse mRNA, and the broken vertical lines indicate the cleavage sites for the putative signal peptides or precursor protein (TGF-β). The sequences are numbered, with 1 being the first amino acid of the mature protein. The portion of the hamster cDNA that we cloned relative to the mouse sequence is represented by the stippled area. The alignment was accomplished with the GCG BestFit program. Murine, rat, and human amino acids which are identical to the hamster amino acid are designated with a dot. Inserted spaces (to improve or preserve the best alignment) are denoted by a dash. Conserved cysteine residues are indicated by a double dagger (‡). Other amino acid residues which have been shown to be critical for biological activity of the mouse or human cytokines are designated by a diamond (◊). Regions known to be involved in receptor binding are overlined. The GenBank accession numbers used in the sequence comparison are listed in Table 1, footnote a.

TABLE 1.

Sequence identities between hamster, mouse, rat, and human cytokines

Cytokine	% Nucleotide (% amino acid) identitya
	Hamster/mouse	Hamster/rat	Hamster/human	Mouse/rat	Mouse/human	Rat/human
IFN-γ	78 (55)	78 (59)	74 (51)	88 (84)	69 (41)	67 (39)
IL-2	84 (76)	88 (83)	80 (74)	90 (81)	73 (62)	76 (65)
IL-4	76 (55)	80 (64)	64 (43)	77 (59)	65 (42)	68 (42)
TNF-α	88 (87)	86 (84)	82 (77)	93 (93)	82 (79)	82 (78)
IL-10	91 (88)	92 (92)	82 (78)	93 (90)	83 (77)	84 (78)
IL-12	84 (78)	85 (80)	78 (72)	93 (92)	71 (64)	72 (64)
TGF-β	93 (100)	93 (100)	92 (99)	98 (100)	91 (99)	90 (99)

^aThe sequence identities shown are for the portion of the sequence corresponding to the hamster cDNA reported here. The amino acid sequence is deduced from the nucleotide sequence. Nucleotide and deduced amino acid sequences corresponding to the primer sequences were excluded from the analyses of sequence identities. The GenBank accession numbers used in the sequence comparison were as follows: IL-2, K02797, M22899, and A14844; IL-4, M25892, X03532, M13238, X16058, and M13982; IFN-γ, K00083, M29317, and A07051; TGF-β, M13177, X52498, and X02812; IL-10, M37897, L02926, and M57627; IL-12p40, M86671, U16674, and M65272; TNF-α, M13049, X66539, and X01394. 

The hamster IFN-γ cDNA that we initially cloned was 309 nucleotides (nt) in length; it started at position 225 of the published mouse sequence and terminated at the end of the coding region (17). This partial-length cDNA was used in the Northern blot studies of splenic mRNA expression (see below). We subsequently amplified and cloned a cDNA encoding the full-length IFN-γ protein (Fig. 1a) with degenerate primers which targeted the 5′ and 3′ UTRs. Homology of the hamster IFN-γ gene sequence to mouse, rat, and human sequences was considerably greater at the nucleotide level than at the amino acid level (Table 1) (10, 17, 18). The divergence of the hamster sequence is in contrast to the previously reported high level of homology between the corresponding mouse and rat IFN-γ protein sequences (10). Of most significance was the low level of homology to the C-terminal and N-terminal α-helices of the mouse and rat proteins (52 to 58% sequence identity), regions which are critical for binding to the IFN-γ receptor (19, 30). In contrast, in these regions there is 80% and 83% homology between the mouse and rat sequences. This divergence is likely to influence the species specificity of this cytokine. The unique RKRKR polycationic tail at the 3′ end of the C-terminal α-helix, which is required for biological activity (52), is conserved in the hamster sequence. In the hamster IFN-γ, however, there is a substitution of asparagine for the histidine at position 111, which could result in diminished biological activity (30). The hamster and human sequences have an additional 17 and 9 amino acids (aa) at the C terminus, respectively, compared with the mouse and rat sequences. Removal of the C-terminal 9 aa from the human IFN-γ protein was found to significantly enhance its antiviral activity (29), and so it is thought that these additional residues sterically block the proximal residues from a strong interaction with the IFN-γ receptor. The additional residues on the C terminus of the hamster IFN-γ thus have the potential to diminish the biological activity of this molecule and therefore could contribute to the hamster’s susceptibility to intracellular pathogens. Studies are under way to test this hypothesis.

The hamster IL-2 cDNA that we cloned was 456 nt in length and extended from the start of the coding region to a site 4 nt short of the end of the coding region in the rat, mouse, and human homologs (Fig. 1b) (26, 33, 53). It had greater nucleotide sequence homology to rat IL-2 than to mouse IL-2 (Table 1). A 42-nt segment containing an unusual CAG triplet repeat coding for 10 glutamine residues uniquely found in exon 1 of the mouse gene is absent in the hamster gene, as it is in the rat, gerbil, and human homologs. This repeat sequence is not thought to have functional significance in the mouse (26). The two cysteine residues (Cys-58 and Cys-105 in the human sequence) which form a disulfide bond required for biological activity of mouse and human IL-2 (47) are conserved in the hamster sequence. The aspartic acid residue (at position 20 in the human sequence) which is required for biological activity (6) is also conserved in the hamster sequence. There was no striking divergence in the hamster sequence corresponding to the mouse and rat A and B α-helices, which are involved in IL-2 receptor binding (24).

Cloning of the hamster IL-4 cDNA required the use of a greater number of primer combinations, including degenerate primers, to obtain an amplification product. The cloned hamster IL-4 cDNA included 46 nt of the 5′ UTR and extended to 33 nt short of the end of the coding region of the mouse homolog (42). Analysis of the deduced amino acid sequence (Fig. 1c) revealed that it had substantially less homology to the mouse and rat proteins than the IFN-γ and IL-2 gene sequences had to the corresponding mouse and rat sequences (34, 42) (Table 1). This lower degree of sequence homology for IL-4 compared with other cytokines has previously been reported for the mouse, rat, and human IL-4 (34, 42, 63). Additionally, the much lower level of homology of the amino acid sequence compared with the nucleotide sequence is not peculiar to the hamster but is also the case in the corresponding mouse and rat sequences. The hamster cDNA, like the mouse and rat homologs, lacks several stretches of nucleotides (the largest contains 24 nt) that are present in the human coding region. The biological function related to these sequence differences is unknown. The six cysteine residues which form three disulfide bonds in the mature human and mouse IL-4 proteins which are required for biological activity (60) are conserved in the hamster sequence.

The hamster TNF-α cDNA that we cloned included 688 nt, which corresponded to most of the coding regions of the mouse, rat, and human cDNAs (12, 44, 45). Only the first 2 and last 13 nt were lacking. The deduced amino acid sequence (Fig. 1d) showed a high level of homology to the mouse, rat, and human sequences (Table 1). A number of features in the mouse, rat, and human homologs are conserved in the hamster TNF-α: (i) the unusually long signal sequence (76 aa); (ii) the 29-aa hydrophobic region, present in the signal sequence, which is thought to be the transmembrane domain; (iii) the two-cysteine residues (numbers 69 and 101 in the human sequence), which form a disulfide bond (41); and (iv) the histidine residue (at position 15 of the mature human TNF-α), which is required for full cytotoxic activity (62).

The IL-10 cDNA fragment that we cloned was 432 nt in length and contained no nucleotide deletions or insertions compared with the mouse, rat, and human IL-10 sequences (16, 39, 56). The deduced amino acid sequence (Fig. 1e) showed a high degree of sequence identity with the mouse, rat, and human homologs. The portion of the hamster cDNA that we cloned lacked the first 27 aa (18 of which are the putative signal peptide in the homologous proteins) and the last 8 aa of the coding region. There are four cysteine residues conserved between the mouse and human sequences which are required for biological activity (59). The amino-terminal cysteine was outside of the portion of the hamster IL-10 that we cloned, but the remaining three residues were conserved.

We cloned a cDNA coding for a 149-aa portion of the hamster IL-12p40 polypeptide (Fig. 1f). The deduced amino acid sequence showed considerable divergence from the highly similar corresponding mouse and rat sequences (49) (Table 1). Surprisingly, it had substantially greater homology to the human amino acid sequence than did either the mouse or rat sequence (49, 61). Within this partial hamster IL-12p40 sequence there are 5 cysteine residues and 3 potential N-glycosylation sites that are conserved from the corresponding mouse and human sequences. The hamster sequence contains a 3-aa stretch that is absent in the mouse and rat sequences but present in the human gene. At another site the human polypeptide contains a single amino acid deletion compared with the mouse, rat, and hamster sequences. The biological significance of these differences is unknown.

The mammalian TGF-β1 mRNA encodes a 390-aa precursor, the C-terminal 112 aa of which encode the mature TGF-β1 protein. We cloned the cDNA that includes most of the region which encodes the mature protein (Fig. 1g). This cDNA showed 92 to 93% homology to the mouse, rat, and human cDNAs (8, 9, 46). The deduced amino acid sequence was identical to the corresponding mouse and rat sequences and differed from the human sequence at only a single amino acid residue (Table 1). This high level of homology has been recognized among all mammalian TGF-β1 polypeptide sequences.

A 391-bp portion of the hamster HPRT gene coding region (sequence not shown) was isolated for use in characterization of expression of this constitutive housekeeping gene relative to the expression of the inducible cytokine genes. This cDNA showed a high degree of homology (96%) to the previously published Chinese hamster HPRT gene nucleotide sequence (28) but less homology to the mouse, rat, and human sequences (22, 23, 28). In general the level of homology between the HPRT gene sequences was greater than that between the cytokine gene sequences, consistent with a previous observation that gene sequence divergence of the cytokines and other host defense molecules is greater than that of the intracellular enzymes (40). As was the case for all of the hamster cytokines except TGF-β, the HPRT cDNA showed substantially greater homology to the mouse, rat, and human nucleotide sequences than to the corresponding deduced amino acid sequences.

We used these cloned cDNAs to analyze cytokine mRNA expression in the Syrian hamster model of VL, which is caused by the intracellular protozoan L. donovani. A number of investigators have previously shown that systemic infection of hamsters with L. donovani results in a progressively increasing hepatic and splenic parasite burden, hepatosplenomegaly, cachexia, pancytopenia, and ultimately, death (13, 15). These features closely mimic active human disease. Also as in humans, hamsters with active VL fail to mount an antigen-specific cellular immune response (15). In contrast, infection of susceptible mice with L. donovani does not cause overt clinical disease or death and is associated with a parasite-specific cellular immune response that is ultimately able to control parasite replication (38, 50). We were therefore interested in the immunopathogenesis of VL in the hamster model because of its unique similarities to active human disease.

In these studies, administration of 5 × 10⁶ parasites by intracardial inoculation resulted in a 65-fold increase in parasite burden over the first month of infection (data not shown). In our experience with this model, cachexia is evident at 2 to 3 months postinfection and death occurs 3 to 4 months postinfection (data not shown). These data are similar to a number of previous descriptions of the course of infection in this model (15, 43). We analyzed the expression of cytokine mRNA in the spleens of control (uninfected) hamsters and in hamsters at 7, 14, and 28 days postinfection (Fig. 2 and 3). We reasoned that the immunopathogenic mechanisms related to progressive parasite replication would be evident during the first month of infection.

FIG. 2.

Cytokine mRNA expression in spleens of control and L. donovani-infected hamsters. The spleens of uninfected controls and hamsters infected for 7, 14, and 28 days were analyzed. Thirty micrograms of hamster spleen total RNA was separated by agarose gel electrophoresis and probed with the labeled HPRT, IL-2, IL-4, IFN-γ, TNF-α, IL-10, IL-12p40, and TGF-β cDNA probes. Hybridization with the HPRT cDNA was used to assess loading equivalency and RNA integrity. The data presented are from two blots. The upper HPRT panel corresponds to the IL-2, IFN-γ, and IL-4; the lower HPRT panel corresponds to the IL-10, IL-12, TNF-α, and TGF-β.

FIG. 3.

Semiquantitative analysis of mRNA expression by densitometry. The autoradiographic bands shown in Fig. 2 were quantified by video image analysis. Shown are the means ± standard errors of the mean of the ratio of the cytokine to HPRT band from the same sample.

Significant baseline expression of IL-4 and TGF-β, but not of the other cytokines, was detected in uninfected animals. The constitutive expression of TGF-β was not surprising since studies in other mammalian cells showed that TGF-β was not strictly transcriptionally regulated (2). The finding of strong basal expression of IL-4 in healthy hamster tissue, however, contrasts with what has been found in mice and rats, in which transcripts in resting cells are difficult to detect even with DNA amplification techniques. This finding raises the possibility that the regulatory sequences in the hamster gene are different from those of the other species. Whether or not this basal expression of IL-4 could inhibit the development of a protective immune response to leishmanial infection cannot be determined until reagents are available for neutralization of this cytokine.

The Th1 cytokine IFN-γ plays a key role in the control of infection with many intracellular pathogens, including Leishmania spp., and is the cytokine primarily responsible for macrophage activation and killing of the intracellular parasite (20, 50). We found a considerable increase in the mRNAs for the Th-1 cytokines, IL-2 and IFN-γ, as early as 1 week postinfection. This suggests that the expression of IFN-γ, which is associated with control of infection in the murine L. major model (20) and the nonlethal murine L. donovani model (50), is not sufficient to control the parasite replication which ultimately leads to death in this animal. Our results concur with another study which demonstrated that splenic T cells with a functional Th1 phenotype (capable of transferring delayed-type hypersensitivity to naive animals) were present in hamsters with active VL (15). Similarly, prominent expression of IFN-γ mRNA was observed in bone marrow specimens obtained from patients with active VL (25). Together these data suggest that an insufficient effector cell response to IFN-γ, rather than absence of the cytokine, contributes to the uncontrolled visceral infection.

IL-12 is a strong inducer of IFN-γ production and plays a major role in the development of the Th1 cell response and control of experimental L. major infection in mice (48). In the hamster, splenic IL-12 mRNA was detected at a modest level starting 7 days postinfection and its expression paralleled that of IFN-γ. Thus, progressive disease in this model was not associated with the absence of IL-12 expression early in the course of infection. Whether or not the expression of mRNA for IL-12 and the Th-1 cytokines is downregulated later in the course of disease is currently being studied.

TNF-α is a proinflammatory cytokine that can have both protective and pathologic consequences for the host. TNF-α acts synergistically with IFN-γ to activate macrophages to kill Leishmania spp. (54) and to promote resolution of murine L. donovani infection (55). At the same time, TNF-α induces cachexia, a prominent feature of progressive VL. We found that mRNA for TNF-α was increased within 1 week of infection but that the level did not increase further during the first month of infection. Pearson and colleagues demonstrated that macrophages from L. donovani-infected hamsters produced high levels of TNF-α activity (measured in the L929 bioassay) 4 to 8 weeks postinfection and suggested that it had a role in the development of cachexia (43). Further studies to characterize its expression later in the course of disease are under way.

In spite of the prominent baseline IL-4 expression, we did not observe an increase in IL-4 transcripts during the course of infection. This is in striking contrast to observations in L. major-infected BALB/c mice, in which parasite-induced IL-4 has been shown to play a critical role in the observed parasite dissemination and disease progression in the murine model (20). Although a previous study of L. donovani infection in BALB/c mice demonstrated in situ hepatic expression of IL-4 by RT-PCR 10 days to 8 weeks after infection (38), this cytokine does not appear to play a major role in the mouse’s modest susceptibility to L. donovani (27).

IL-10 has been demonstrated to antagonize Th1 cytokine synthesis and to inhibit macrophage activation and killing of intracellular parasites (7, 57). We found increasing levels of IL-10 expression in splenic tissue over the first 4 weeks after infection, suggesting that IL-10 may play a role in the progressive disease seen in the hamster. Studies by Wilson et al. (58) and our laboratory (36) of the murine model of VL demonstrated a high level of splenic IL-10 expression which contributed to the suppression of splenic T-cell function associated with visceral parasite replication. Similarly, IL-10 has been shown to play a role in the suppression of T-cell responsiveness and IFN-γ secretion in active human VL (14, 21). Another potent inhibitor of macrophage function, TGF-β (11), has been shown to exacerbate L. major infection in mice (3). We found no increase in the expression of TGF-β mRNA following infection, but since its production is posttranscriptionally regulated (2), it could still play a role in the progressive visceral disease in the hamster. The role of TGF-β in this model needs to be confirmed with analysis of protein expression.

In summary, sequence analysis of seven Syrian hamster cytokine genes revealed that IFN-γ had unique features that could contribute to the extreme susceptibility of this animal to Leishmania spp. and other intracellular pathogens. The other hamster cytokines showed no striking sequence divergence from the published sequences of homologs that would indicate altered biological function, but these cDNAs were incomplete. Use of these cDNAs enabled the first characterization of the molecular immunopathogenesis of a hamster model of infection. In this model of active VL, early parasite replication, which ultimately leads to death, was not associated with polarization of the immune response toward a Th2 cytokine profile and away from a Th1 cytokine profile. This is similar to what has been observed in human VL but contrasts with the dominant Th2 responses observed in progressive L. major infection in BALB/c mice. Further studies to characterize the mechanisms of susceptibility in this model should provide valuable insight into the immunopathogenesis of human VL.