Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Aminoglycoside antibiotics are indispensable for treatment of serious bacterial infections, and despite careful attention to dosage regimens, nephrotoxicity and ototoxicity still cause concerns (Lortholary et al., 1995). In the present study, we tested whether side effects of aminoglycoside therapy could be limited by expression of genes of antibiotic resistance in vivo. As a model system, we used aminoglycoside antibiotic hygromycin B and a transgenic mouse strain bearing the antibiotic resistance gene hyg^R under the control of the constitutive promoter Pgk1 (Johnson et al., 1995).

Hygromycin B is an aminoglycoside antibiotic produced by Streptomyces hygroscopicus (Pettinger et al., 1953) that is active against both prokaryotic and eukaryotic cells. It has been shown that in eukaryotic cells, hygromycin B acts by interfering with protein synthesis, especially by inhibition of translocation, which is thought to be the result of its interaction with the eukaryotic ribosome displacing EF-2 from the ribosome or interfering with the activity of EF-2 and the stabilization of peptidyl-t-RNA bound to the ribosomal acceptor site (Gonzalez et al., 1978). Toxic effects of hygromycin B in tissue culture can be prevented by expression of hyg^R (Blochlinger and Diggelmann, 1984). This enzyme adds phosphate to position 7 of the destomic acid ring of hygromycin B, which results in complete loss of biological activity both in vitro and in vivo (Pardo et al., 1985). Therefore, hyg^R has been widely used as a positive selective marker in the construction of transgenic animals via ESC. The transgenic construct containing hyg^R is introduced into ESC. Then clones of ESC bearing recombinant DNA with hyg^R can be selected in medium containing hygromycin B. Although toxic effects of hygromycin B have been studied on the cellular level in tissue culture (see, for instance, Chen et al., 1995; Gonzalez et al., 1978; Pardo et al., 1985), we know of no previous reports on acute or tissue-specific hygromycin B toxicity in transgenic mice bearing hyg^R. For our experiments, we have used transgenic mice carrying hyg^R driven by the constitutive Pgk1 promoter (Johnson et al., 1995). These mice develop normally and do not exhibit any apparent abnormal phenotype.

The aim of our study was to investigate the influence of constitutive expression of hyg^R on the toxic effects of hygromycin B in vivo. We have examined expression of hyg^R mRNA in different tissues, and we have determined the approximate lethal dose of hygromycin B for hyg^R and wild-type mice. Tissue-specific changes were then evaluated by histopathological examination and serum biochemical analysis. Our approach of using transgenic mice bearing hyg^R might contribute to better understanding of the mechanism of action and toxicity of aminoglycoside antibiotics.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and drug treatment. Homozygous transgenic mice C57BL/6J-TgN(pPWL512hyg)1Ems carrying hyg^R (Johnson et al., 1995) and wild-type C57BL/6J were obtained from the Jackson Laboratory (Bar Harbor, ME). The animals were kept in specific pathogen free conditions and were supplied with standard diet and water ad libitum. The room was held at 22°C with humidity 70% and a 12-hr dark-light cycle. Animal care and experimental procedures were carried out in accordance with institutional guidelines. Hygromycin B was purchased from Sigma Chemical Corp (St. Louis, MO). The compound was dissolved in sterile water before application, and the solution was injected i.p.

Acute toxicity, approximate lethal dose. The approximate lethal dose of hygromycin B after single i.p. injections was determined according to Deichmann and LeBlanc (1943). This protocol significantly reduces the number of experimental animals, limits unnecessary suffering and complies with current guidelines on animal studies (IARC, 1993). Two sets of 6-month-old male mice (22-28 g) were used for all acute toxicity experiments. The animals were treated with a single dose of hygromycin B i.p. at doses that started at 2.7 mg/kg and increased by 50% for each consecutive dose. Control mice were treated with the same volume of sterile saline. Total volume injected was 0.5 ml. The health status and body weights of animals were monitored daily for 10 consecutive days. The lowest dose in the dose response at which one mouse died was considered the approximate lethal dose. Dead or sacrificed moribund animals were necropsied, and the organs were stored in 10% buffered neutral formalin until further analysis. The animals were checked twice a day. The maximum theoretical time the animals could be dead before necropsy was performed was 12 hr. Because animals did not seem moribund when checked, we assume that the actual time was much less.

Histopathological examination and serum biochemical analysis. The cadavers were necropsied, and the organs were stored in 10% buffered neutral formalin. The organs were prepared as paraffin-embedded glass slides stained with hematoxylin and eosin and were evaluated in terms of the NTP (National Toxicology Program of NIEHS) standards. When possible, a complete cross section of each organ was evaluated (liver, spleen, gastrointestinal tract, femoral bone marrow, mandibular and mesenteric lymph nodes, kidney, brain, testes, lungs and heart). For liver, two cross sections, one of each of the two largest liver lobes, were examined. For kidneys, an entire cross section (left longitudinal, right transverse) was evaluated. Lungs had two cross sections (one of each of the two largest lobes). The entire sections on the slides (all fields) were evaluated under blind conditions for lesions and were scored (graded) on a subjective basis compared with control animals. The grades were as follows: 1 = minimal, 2 = mild, 3 = moderate, 4 = marked and 0 = no pathological changes. The preparation and evaluation of slides used the NTP criteria and terminology (Chhabra et al., 1990).

Northern blot analysis. Total RNA samples were isolated from spleen, heart, thigh muscle, lung, kidney, liver, brain and testis by guanidinium thiocyanate-phenol extraction (Chomczynski and Sacchi, 1987), separated by electrophoresis in a formaldehyde/agarose gel and transferred to a nylon HybondN+ membrane (Amersham, Arlington Heights, IL) by capillary blotting. To compare loading of RNA samples we photographed the ethidium bromide-stained gels. The blots were hybridized to a hyg^R cDNA probe (Johnson et al., 1995) that was labeled with ³²P dCTP (DuPont NEN, Boston, MA) using random oligonucleotide primers (^T7QuickPrime, Pharmacia, Piscataway, NJ). The autoradiograms were exposed for 48 to 72 hr. The bands of hyg^R mRNA on autoradiograms and the ethidium bromide-stained 18S rRNA bands of corresponding samples were analyzed by scanning densitometry with a BioImage (Millipore, Bedford, MA) system. To compare expression of hyg^R mRNA, we calculated the relative IOD as total IOD of the autoradiographed band normalized to the intensity of the corresponding 18S ribosomal band visualized by ethidium bromide.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue-specific expression of hyg^R. The steady-state levels of hyg^R mRNA in tissues of transgenic mice were examined by Northern blot analysis. Total RNA was isolated from spleen, heart, thigh muscle, lung, kidney, liver, brain and testes of male hyg^R-bearing transgenic and wild-type mice using guanidinium thiocyanate-phenol extraction. The RNA samples were separated by electrophoresis in a formaldehyde/agarose gel, transferred to a nylon membrane by capillary blotting and hybridized with ³²P-labeled hyg^R cDNA probe. The autoradiograms were evaluated by densitometry, and the results represent the average of three transgenic mice (fig. 1). The hyg^R mRNA reached different levels in different tissues of the hyg^R-bearing transgenic mice. No signal was detectable in wild-type mice. The highest level of hyg^R mRNA expression was detected in the brain. Mid-level expression was detected in skeletal muscle, testis, kidney spleen and liver, and somewhat lower levels were detected in heart and lungs.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Expression of hygR mRNA in mouse tissues. Total RNA was isolated and separated by electrophoresis in a formaldehyde/agarose gel, transferred to a nylon HybondN+ membrane and hybridized with labeled hygR cDNA. Autoradiograms were exposed 48 to 72 hr. A) Relative levels of expression normalized to amount of RNA loaded. Bars represents averages ± S.E.M. B) Autoradiogram. C) Amounts of RNA loaded in different lanes were compared on ethidium bromide-stained gels.

Acute toxicity approximate lethal dose. The approximate lethal dose (Deichmann and LeBlanc, 1943) of hygromycin B in transgenic hyg^R-bearing as well as wild-type mice was determined as a measure of the acute toxicity. The mice were treated with single doses of hygromycin B i.p. (2 mice per dose) that increased by 50% for each consecutive dose. The first dose in the increasing sequence of doses at which the mice died was considered the approximate lethal dose. The health status of animals was monitored for consecutive 10 days. Hyg^R transgenic and wild-type mice tolerated a single i.p. injection of hygromycin B without immediate toxic symptoms or distress. The approximate lethal dose of hygromycin B for wild-type mice was 6 mg/kg (table 1). Expression of hyg^R in transgenic mice caused a substantial increase in resistance to hygromycin B. The approximate lethal dose for hyg^R animals was 535 mg/kg, which represents an 89-fold increase over the wild-type strain. Lethal doses in both hyg^R-expressing transgenic and wild-type strains caused the same signs of decreased activity, which progressed to lethargy and death. The body weight of animals after lethal doses decreased, an effect that was less severe at shorter survival duration after hygromycin treatment (table 1).

Serum biochemical analysis. Hyg^R mice were treated with the lethal dose of 803 mg/kg hygromycin B, and wild-type mice were treated with the lethal dose of 9 mg/kg hygromycin B. Control hyg^R mice were also treated with the nontoxic dose of 9 mg/kg. Serum samples taken 48 hr after treatment were evaluated for the levels of BUN, AST and ALT. Elevated BUN is associated with dehydration and/or renal insufficiency. Elevated activities of ALT and AST are characteristic of liver damage, particularly necrosis, cirrhosis and hepatitis. Increased AST levels are also characteristic for muscle trauma, myocardial infarction and myositis.

Histopathological examination. Morphological manifestations of hygromycin B tissue-specific toxicity in animals treated with lethal doses was assessed by microscopic analysis. Lethal doses of hygromycin B (9 mg/kg) in wild-type animals caused nephrotoxicity (table 3). These mice had tubule eosinophilia, degeneration and necrosis characterized by cytoplasmic eosinophilia associated with necrosis, loss of tubule cellular and nuclear detail or degeneration and pyknotic nuclei and fragmentation of cells (fig. 2). These lesions are typical of acute tubular nephrosis. Remaining tissues (liver, spleen, GI tract, femoral bone marrow, mandibular and mesenteric lymph nodes, brain, testes, lungs and heart) were within normal limits (only kidney data are shown for comparison in fig. 2). The hyg^R mice treated with lethal doses of hygromycin B (803 mg/kg) had liver damage characterized as hepatocellular fatty change, acute inflammation and hepatocellular necrosis (table 3; fig. 3). Liver lesions in the hyg^R transgenic mice were typified by hepatocellular necrosis with nuclear pyknosis and loss of cellular detail, acute infiltration with small foci of neutrophils and fatty change with hepatocellular intracytoplasmic large, distinct, clear vacuoles that displaced nuclei. These liver lesions are characteristic of acute liver damage. The remaining tissues (spleen, GI tract, femoral bone marrow, mandibular and mesenteric lymph nodes, kidney, brain, testes, lungs and heart) were within normal limits (only liver data are shown for comparison in fig. 3).

View larger version (146K): [in this window] [in a new window]   Fig. 2.   Kidney of wild-type and hygR+/+ transgenic mice after treatment with lethal doses of hygromycine B. A) Section of kidney from a wild-type mouse administered 9 mg/kg of hygromycin B has scattered renal tubules degeneration and necrosis in the cortex. Affected tubules are characterized by loss of tubule cell details (arrow) and pyknosis or small nuclei (arrowheads). B) The renal cortical tubules from a transgenic hygR+/+ mouse treated with 803 mg/kg of hygromycin B are within normal limits (arrows). (Magnification 40×; stained with hematoxylin and eosin).

View larger version (151K): [in this window] [in a new window]   Fig. 3.   Liver of wild-type and hygR+/+ transgenic mice after treatment with lethal doses of hygromycin B. A) Liver from a wild-type mouse administered 9 mg/kg of hygromycin B had normal size and shape of hepatocytes (arrows) within normal limits. B) The hygR+/+ transgenic mice given 803 mg/kg of hygromycin B had scattered hepatocellular necrosis characterized by pyknosis or loss of hepatocellular detail (arrows) and hepatocellular fatty change typified by clear, eccentric round vacuoles in hepatocellular cytoplasm (arrowheads). (Magnification 40×; stained with hematoxylin and eosin).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, we characterized the acute and tissue-specific toxicity of hygromycin B in transgenic mice bearing the hyg^R gene under control of a constitutive promoter. We characterized the tissue-specific expression of hyg^R mRNA, and we investigated the acute toxicity of hygromycin B in hyg^R and wild-type mice. The hyg^R expression in transgenic animals caused an 89-fold increase in the approximate lethal dose of hygromycin B compared with wild-type mice. Serum biochemical analysis of hyg^R and wild-type mice treated with lethal doses of hygromycin B indicated liver and kidney damage as measured in terms of ALT, AST and BUN. On the morphological level, these changes were manifested as acute tubular nephrosis in wild-type mice and acute liver damage in hyg^R mice. Our results show that the constitutive expression of the bacterial hyg^R gene in transgenic mice in vivo confers resistance to hygromycin B.

To evaluate hyg^R expression in vivo, we measured steady-state levels of hyg^R mRNA in different mouse tissues by Northern blot analysis. Although hyg^R mRNA was detected in all tissues we examined, the levels of expression varied. The highest level was detected in brain, and mid-levels were measured in skeletal muscle, testis, kidney, spleen and liver. Lower levels were measured in heart and lungs. It has been reported that Pgk1-driven transcription of lacZ in transgenic mice is not uniform among the tissues and that the highest levels were reached in testes and brain, that were measured mid-levels in kidney and heart followed by liver and skeletal muscle and that the lowest levels were detected in spleen (McBurney et al., 1994). The slightly different pattern of transgene expression in our study may be caused by the length of the Pgk1 promoter sequence used for hyg^R regulation, by the lack of introns and/or by differences between integration sites in the genome. In case of the hyg^R mouse, the Pgk1 promoter fragment contained a promoter-specific enhancer placed directly upstream of the hyg^R gene (Johnson et al., 1995). In contrast, the lacZ construct contained the first three exons and two introns in addition to the Pgk1 promoter-specific enhancer (McBurney et al., 1994), which may have influenced expression of the transgene (Palmiter et al., 1991).

As a quantitative parameter of hygromycin B toxicity, we determined the approximate lethal dose that corresponds to the LD₅₀ ± 30% for most chemicals (Deichmann and LeBlanc, 1943). This protocol minimizes the number of animals needed for the assessment of acute toxicity without unduly compromising the accuracy (Ecobichon, 1992). The approximate lethal dose in wild-type mice for hygromycin B after i.p. application in our experiments was 6 mg/kg. This value is identical to the LD₅₀ estimate after i.v. application using conventional methods (Berdy, 1980), which indicates that the approximate lethal dose is an appropriate parameter for evaluating the acute toxicity of hygromycin B in vivo. Transgenic hyg^R positive mice showed a dramatic 89-fold increase in the approximate lethal dose over the wild-type strain. This suggests that the level of Hyg^R enzyme present in the cells of the transgenic mice effectively decreased hygromycin B toxicity in vivo.

To characterize the organ targets of hygromycin B toxicity in wild-type and hyg^R transgenic mouse, we performed biochemical analysis and histopathological evaluation of mice treated with lethal doses of hygromycin B. The kidney damage was characterized as an alteration in BUN levels, and liver damage was evaluated as a change of ALT and AST activities. Serum biochemical analysis of both wild-type and hyg^R mice treated with lethal doses of hygromycin B (9 mg/kg and 803 mg/kg, respectively) revealed renal injury characterized as increased levels of BUN and liver damage characterized as an increase in ALT and AST enzymatic activities. Morphological correlates to the serum biochemical results were characterized using histopathological analysis. The wild-type mice treated with lethal doses of hygromycin B showed histological signs of nephrotoxicity identified as tubule eosinophilia, degeneration and necrosis. In contrast, the transgenic hyg^R transgenic mice treated with toxic doses of hygromycin B showed histopathological signs of liver damage characterized as hepatocellular fatty changes, acute inflammation and hepatocellular necrosis. Although the serum biochemical analysis and histopathological findings in wild-type and hyg^R mice differ, the liver and kidney seem to be the targets of hygromycin B toxicity in both mouse strains. The weak correlation between biochemical markers and histopathological findings has been reported in other instances, so organ-specific damage must be characterized by using more than one parameter (Baum et al., 1975; Herrera, 1993). It seems that in hyg^R mice, the lethal doses of hygromycin B overloaded the metabolic capacities of Hyg^R enzyme to inactivate the drug. The absence of hepatic or renal histopathological manifestation of hygromycin B toxicity in wild-type animals or hyg^R mice, respectively, may be explained by different availability of active hygromycin B molecules in liver and kidney.

Pathological manifestations of renal toxicity associated with aminoglycoside therapy consist largely of acute tubular lesions and necrosis resulting in kidney failure (Laurent et al., 1990; Tulkens, 1989). This pathological pattern of tissue injury correlates with the pharmacokinetics and disposition of aminoglycosides. After application, the aminoglycosides are excreted unchanged through the kidneys (Sande and Mandell, 1985), where a portion of the dose is reabsorbed and cumulated in the tubular cells of the renal cortex (Giuliano et al., 1986; Kuhar et al., 1979). The association of aminoglycosides with negatively charged phospholipids and their accumulation in the lysosomes of tubular cells leads to phospholipidosis by inhibition of lysosomal phospholipases (Laurent et al., 1982), which may trigger necrosis (Laurent et al., 1990). On the other hand, it has also been shown on the cellular level that hygromycin B interferes with protein synthesis, especially by inhibition of translocation (Gonzalez et al., 1978). Our results indicate that compared with aminoglycosides used in therapy, hygromycin B caused not only clinically similar kidney damage but also liver toxicity. The exact molecular mechanism of hygromycin B toxicity and the degree of its possible interaction with phospholipids and/or protein synthesis in vivo remain to be elucidated.

Aminoglycoside antibiotics are indispensable for treatment of serious infectious diseases. Despite careful attention to dosage regimens designed to achieve targeted levels in the serum, their toxicity still causes concern (Lortholary et al., 1995). Although hygromycin B is not used for treatment of human patients, our results clearly show that its toxicity can be prevented by expression of hyg^R in vivo. The general approach of using transgenic animals with specific changes in drug metabolism pathways can contribute to our understanding of the mode of action and toxicity of drugs. This could ultimately lead to improvement of drug design and treatment regimens.