A Non-Nephrotoxic Gentamicin Congener That Retains Antimicrobial Efficacy
Ruben M. Sandoval*,,
James P. Reilly,
William Running,
Silvia B. Campos*,,
Joseph R. Santos*,
Carrie L. Phillips*,, and
Bruce A. Molitoris*,,¶
* Indiana University School of Medicine, Division of Nephrology, Indiana Center for Biological Microscopy, Indiana University Department of Pathology & Laboratory Medicine, and ¶ Roudebush Veterans Administration Medical Center, Indianapolis; and Indiana University Department of Chemistry, Bloomington, Indiana
Address correspondence to: Dr. Bruce A. Molitoris, Department of Medicine, Nephrology Division, Indiana University School of Medicine, 950 W. Walnut Street, R2-202C, Indianapolis, IN 46202. Phone: 317-274-5287; Fax: 317-274-8575; E-mail: bmolitor{at}iupui.edu
Received for publication October 27, 2005.
Accepted for publication August 2, 2006.
Aminoglycoside antibiotics, although of major clinical importancein the treatment of serious Gram- negative infections and apotential therapeutic agent in the amelioration of diseasesthat are characterized by premature stop mutations, are associatedwith a high incidence of acute renal failure. With the use ofHPLC techniques, the four components (congeners) of gentamicin,the most commonly used aminoglycoside, were isolated and characterized.Described here is a congener with minimal cytotoxicity in cellculture and animal studies that retained normal bactericidalproperties in both Bacillus subtilis and a multidrug-resistantform of Klebsiella pneumoniae. Furthermore, in animal studies,this congener failed to induce the functional and pathologicchanges that are characteristic of gentamicin nephrotoxicitythat is seen with the native compound. Finally, internalizationof this non-nephrotoxic component was unaltered, but the subcellulardistribution was different from native gentamicin or the otherthree cytotoxic congeners. These studies have identified a componentof the native gentamicin congener mixture that retains its bactericidalproperties with minimal or no apparent nephrotoxicity.
Aminoglycoside antibiotics have remained a mainstay for thetreatment of Gram-negative infections for >40 yr despitetheir associated nephrotoxicity. With the dramatic increasein sepsis (1), sepsis-induced acute renal failure (ARF) (2),the lack of new antibiotics, and the emergence of multidrug-resistantbacteria (3,4), the continued and even increased use of aminoglycosidesis occurring. Unfortunately, a high percentage of patients whoare treated with aminoglycosides develop ARF. In addition, patientswith existing kidney, liver, or cardiac diseases have a rateof aminoglycoside-induced ARF as high as 50% (5,6). This increasedrate of ARF adds morbidity, mortality, and cost to the adverseeffects that are associated with this clinically important butpotentially dangerous class of antibiotics. Therefore, understandingthe intracellular mechanism of aminoglycoside toxicity may leadto approaches that minimize this barrier to its use. Recentdata from several genetic diseases secondary to premature stopmutations indicate that gentamicin suppresses reading of thestop codon, resulting in expression of the full-length protein(7,8).
Data from our laboratory have elucidated a trafficking pathwayfor gentamicin that is responsible in large part for the associatedintracellular pathologies (9). In proximal tubule cells, aminoglycosidesare internalized by the multiligand receptor megalin (7,10,11)and sequestered by the proximal tubule epithelia primarily intolysosomes. However, a small percentage of the absorbed gentamicintraffics retrograde via the endocytic pathway through the Golgicomplex and into the endoplasmic reticulum. It subsequentlyis released into the cytosol, freeing it for toxic associationwith subcellular organelles such as the nucleus, mitochondria,intracellular vesicles, and the inner leaflet of surface membranes(9).
Commercial gentamicin is a mixture of various congeners C1 andC1a and the enantiomers C2 and C2a. Structural differences betweenthem are minor, differing only by a methyl or hydrogen substitutionin two R groups on the purpurosamine residue. In a publicationby Kohlhepp et al. (12), the nephrotoxicity of the various congenersoriginally was investigated. Unfortunately, this work was hamperedby technical limitations in the ability to separate the enantiomersC2-C2a as well as cross-contamination of the various purifiedcongeners with other congeners in ranges of 2 to 12%. In addition,the C2a content of these purified components was unknown. Inthis study, we isolated the individual congeners of gentamicin,including the enantiomers C2 and C2a, using current HPLC techniques.The enantiomer C2 exhibited little cellular toxicity and nonephrotoxicity while maintaining bactericidal efficacy. Resultsfrom this study could lead to more widespread use of gentamicinin the form of the isolated C2 congener, especially in high-riskpatients. This purified congener also could find long-term dailyuse in treating genetic diseases that are induced by prematurestop codon defects in which the presence of gentamicin inducesa read-through of the stop codon in the defective message (7,8).
Separation and Characterization of Congeners
Separation of native gentamicin was carried out using a C18column (250 x 4.60 mm), and the chromatogram was developed usinga Waters 600e HPLC pump (Milford, MA). A gradient of 100% ofa 0.1% (vol/vol) mixture of tri-fluoro-acetic acid in waterto 100% of a 0.1% (vol/vol) mixture of tri-fluoro-acetic acidin acetonitrile was used to develop the chromatogram. Retentiontimes of 12.87, 22.07, 23.84, and 25.85 min were noted for congenersC1, C2, C2a, and C1a, respectively. There was no overlap betweenadjacent peaks. For all studies, the same lot of gentamicinwas used. This ensured that the results obtained were independentof potential alterations in the ratio of gentamicin components.
Lyophilized gentamicin congeners were resuspended in ddH20 toa concentration of 20 mg/ml. Actual concentrations then weredetermined at the clinical chemistry laboratory of the RoudebushVA Hospital. Discrepancies between perceived and actual concentrationslikely are due to the hygroscopic nature of gentamicin; actualmeasured concentrations were used for all dosing calculations.
Bactericidal assays were conducted using a multidrug-resistantform of Klebsiella pneumoniae that was grown in #11 medium (Difco,Sparks, MD) and Bacillus subtilis that was grown on regulartryptic soy agar medium in large 245 x 245-mm agar plates forat least 24 h. Wells were loaded with 100 µl of nativegentamicin or the separated congeners at concentrations of 5or 10 µg/ml (control wells contained PBS). Plates wereexamined the next day to determine areas of clearance. Areasthat were completely clear were quantified for kill zone areas;area that were slightly more opaque than surrounding areas weredeemed inhibition areas. Generally, the B. subtilis presentedonly kill zones.
Cell Toxicity Assays
LLC-PK1 porcine proximal tubule cells were plated on 18-mm circularcoverslips and grown in K-P medium supplemented with 10% FBS(Hyclone, Logan, UT) and penicillin streptomycin (Sigma-Aldrich,St. Louis, MO). Upon reaching approximately 40% confluence,cells were incubated in media that contained the various congenersor native gentamicin at a final concentration of 1 mg/ml. Afterapproximately 24 h, the cells were washed once with medium andplaced in medium that contained 2.5 µg/ml Hoechst 333442(Molecular Probes, Eugene, OR) and 0.667 µg/ml propidiumiodide (Molecular Probes) for 10 min. The cells then were washedin normal medium and visualized live using confocal microscopy.
In Vivo Nephrotoxicity Assay
Male Sprague-Dawley rats (200 to 250 g; n = 3; Harlan, Indianapolis,IN) received a single daily dose of either native gentamicinor the congener C2 at 100 mg/kg or a normal saline solutionintraperitoneally for 6 d. Blood samples were taken initiallyto establish a baseline and daily up until 24 h after the finalinjection. Serum creatinine then was measured on a Beckman CreatinineAnalyzer 2 (Beckman Instruments, Brea, CA) and reported in mg/dl.All protocols were approved by the Institutional Animal Careand Use Committee.
Histology and Indirect Immunofluorescence Localization in Tissue
After completion of the in vivo nephrotoxicity study, rat kidneyswere perfusion-fixed with fresh 4% paraformaldehyde in PBS (pH7.4). Some of the tissue was embedded in paraffin, and 3-µmsections were cut for histology, stained with hematoxylin andeosin (H&E) or periodic acid-Schiff (PAS), and scored bya pathologist (C.L.P.) who was blinded to the interventions.At least two representative sections of each kidney were examined.Morphologic evaluation of injury was assessed by grading theextent of necrosis of the proximal convoluted tubules as outlinedby Jablonski et al. (13). Injury scores were assigned on thebasis of the predominant pattern present. For indirect immunofluorescence,100-µm-thick sections were cut on a vibratome and stainedwith the monoclonal anti-gentamicin antibody (Biodesign International,Saco, ME) and a fluorescein-conjugated secondary antibody (JacksonImmunoresearch, West Grove, PA) and lightly counterstained withAlexa 647–phalloidin (Molecular Probes) to localize filamentousactin for visualization using confocal microscopy.
Indirect Immunofluorescence Localization in LLC-PK1 Cells
Because of the high degree of cell death that was seen in theprevious cytotoxicity assay with the congeners, LLC-PK1 cellswere allowed to reach approximately 70% confluence before beingincubated in medium that contained 1 mg/ml of either nativegentamicin or the congeners. During this period of exposureto native gentamicin or the congeners, lysosomes were labeledby adding a 10,000 molecular weight rhodamine dextran to theculture medium (0.5 mg/ml; Molecular Probes). The cells thenwere fixed and stained for indirect immunofluorescence and visualizedvia confocal microscopy. Briefly, the Golgi complex was localizedwith a fluorescein-conjugated lectin from Helix pomatia at 5µg/ml (EY Laboratories, San Mateo, CA) diluted in solutionthat contained the CY-5–conjugated secondary antibodythat was used in the indirect immunolocalization of gentamicin.Visualization of nonlysosomal, cytosolic gentamicin was carriedout as described previously (9), using a Tyramide Signal Amplification(Molecular Probes).
Microscopy
Histologic sections were viewed with a Widefield Nikon microscopewith an attached color CCD camera (Nikon, Melville, NY). Fluorescenceimages were acquired using a Bio-Rad MRC 1024 confocal microscope(Bio-Rad, Hercules, CA) on a Nikon platform (The Fryer Co.,Huntley, IL). Co-localization studies with Texas Red dextran,FITC-Helix pomatia, and gentamicin were conducted on a Zeiss510 confocal microscope with images illuminated and acquiredsequentially to eliminate bleed-over emissions from the differentfluorescence emission spectra.
Statistical Analyses T tests were conducted on serum creatinine values between congenerC2 and the native compound at the various injection days andon Jablonski score values between congener C2 and the nativecompound using Excel 2003 (Microsoft Corp., Redmond, WA). Differencesin necrosis, apoptosis, and cell density among the various congenerson the human and porcine cell lines were determined using anANOVA function with Systat 11 software (Systat Software, PointRichmond, CA). Statistical significance was achieved at P 0.05.All reported values are mean ± SEM.
Individual Congeners of Gentamicin Separated by HPLC Retain Their Bactericidal Properties
To determine whether the various congeners of gentamicin retainedtheir bactericidal properties, we first separated the congenersusing HPLC techniques. The congeners eluted off the HPLC columnin the order C1, C2, C2a, and C1a with retention times of 12.87,22.07, 23.84, and 25.85 min, respectively (Figure 1A).
Figure 1. Individual congeners of gentamicin retain their bactericidal properties. Using HPLC techniques, individual congeners of gentamicin, including the enantiomers C2 and C2a, were separated without cross-contamination (A). Bactericidal assays using Bacillus subtilis and a multidrug-resistant form of Klebsiella pneumoniae were conducted on agar plates. The kill zones and growth inhibition zones were measured and graphed, as shown in B. Overall, little variation occurred between the individual congeners and the native gentamicin compound. Assays on B. subtilis always resulted in a clear kill zone around the 50-µl well and no zone of inhibition. With K. pneumoniae, a smaller kill zone was seen than with B. subtilis, surrounded by a larger, more opaque region that represents a zone of growth inhibition (n = 3; data are means ± SEM). Native refers to commercial gentamicin.
Bacterial assays using B. subtilis demonstrated that all individualcongeners retained the bactericidal efficacy of the native compound(Figure 1B). Kill zone measurements in the agar plates indicatedsimilar clearance areas, between 12 and 15 mm at a concentrationof 10 µg/ml. Use of a multidrug-resistant form of K. pneumoniaeyielded similar results to the B. subtilis. With K. pneumoniae,bactericidal efficacy decreased, producing smaller kill zones(between 9 and 11 mm at 10 µg/ml) and the appearance ofzones of growth inhibition between 13 and 14 mm that were notseen with B. subtilis (histogram, Figure 1B). These kill andinhibition zones were equivalent for the four congeners, aswell as for the native commercial mixture.
Enantiomer C2 Exhibited Reduced Toxicity in LLC-PK1 Cells
Because all the congeners were found to retain bactericidalactivity, we next conducted cytotoxicity assays in culture usingLLC-PK1 proximal tubule cells to determine the toxicity of eachindividual congener. The nuclear dyes propidium iodide and Hoechst33342 were used as markers for necrosis and apoptosis, respectively.In this assay, the cell permeant dye Hoechst 33342 labeled nucleiof all cells with an even distribution (Figure 2A, Blank). Apoptoticcells were discerned from normal cells by condensation of nuclearmaterial and increased localized fluorescence (Figure 2A, Native,arrow). Necrotic cells incorporated the cell-impermeant dyepropidium iodide and displayed a pink color when combined withthe cyan color of the Hoechst 33342 (Figure 2A, C1, C1a, andC2a, arrowheads). Here, a marked difference in cytotoxicityamong the various congeners was observed. All congeners exceptC2 had 100% toxicity at 24 h of exposure (Figure 2). Conversely,the percentage of cell death with the C2 congener was 3.1 ±1.3 and 8.9 ± 3.4% for the native mixture. This combinationof cell death and inhibition of cell growth resulted in celldensity values that were significantly higher for congener C2when compared with the native compound, with values of 80.3± 5.7 and 58.5 ± 7.7% (P 0.05) of untreated values,respectively. Both of these values were elevated when comparedwith those that were obtained for congeners C1, C1a, and C2a,which all were <20% of untreated values (Figure 2B). Thedata suggest that the C2 component of gentamicin had minimalcytotoxicity and actually reduced the toxicity of the othergentamicin components when present in the native mixture.
Figure 2. Cytotoxicity assays using cultured cells revealed that the first enantiomer separated, which we call congener C2, had markedly reduced toxicity after 24 h of continual exposure. Results in the porcine proximal tubule cell line LLC-PK1 were striking. (A) Apoptosis was assessed using the nuclear dye Hoechst 33342, which labeled nuclei of normal cells uniformly with a lower intensity and the nuclei of apoptotic cells more intensely and demonstrated the presence of condensed apoptotic bodies. The nuclear dye propidium iodide is cell impermeant and labels only necrotic cells, characterized by permeable cell membranes; co-localization of the two dyes gives a reddish-pink color. For all congeners except C2, evidence of widespread necrosis existed (A). For the congener C2, minimal evidence of toxicity was present. A histogram that was generated form the microscopic data revealed a toxicity index for C2 lower than any other individual component and lower than the native compound, 3.1 ± 1.3 and 8.9 ± 3.4%, respectively (B). Cell density for the toxic C1, C1a, and C2a congeners dropped below 20% of control values. The values for C2 were significantly higher than any other congener and also higher than the native compound (*P 0.05) with reported values of 80.3 ± 5.7 and 58.5 ± 7.7%, respectively. Values are means ± SEM. Bar = approximately 10 µm.
Six Days of Daily Exposure to Enantiomer C2 Is Not Nephrotoxic in Sprague-Dawley Rats
Having identified a gentamicin congener with reduced toxicityin cultured cells, we next set out to determine the toxicityof this congener in vivo. Male Sprague-Dawley rats that weregiven either congener C2 at 100 mg/kg per d or normal salineintraperitoneally for six daily doses showed no elevation inserum creatinine (Figure 3A). In these groups, serum creatininevalues remained 0.2 mg/dl up to 24 h after the last injection.However, rats that were given the same dose of the native gentamicincompound exhibited an elevation in serum creatinine 24 h afterthe first injection, with a near tripling of serum creatininevalues to a peak value of 0.55 mg/dl 24 h after the fourth injection.We specifically used six daily doses for these studies to allowadditional time for the C2 component to induce injury, becauseit is widely known that serum creatinine rises early in thismodel when native gentamicin is used. Therefore, even afterprolonged exposure to C2, there was no change in serum creatininevalues, which remained significantly lower than those of nativegentamicin–treated rats (P 0.05).
Figure 3. Congener C2 did not induce nephrotoxicity in Sprague-Dawley rats. Congener C2 or the native compound was administered daily for 5 d. Serum creatinine levels were determined and graphed showing levels 24 h after injection (A). Rats exposed to C2 or normal saline exhibited no change in serum creatinine levels, which remained significantly lower than values of native gentamicin–treated rats (*P 0.05), a finding that is indicative of no alteration in renal function. Hematoxylin and eosin (H&E)-stained sections from rats given native gentamicin exhibited cell sloughing and debris in the tubular lumens (B, arrows). In addition, large perinuclear vacuoles (B, arrowheads) could be seen in these tissues. Sections from C2-treated rats contained few epithelial cells in the tubular lumens, with intracellular vacuoles when present, smaller and more evenly dispersed throughout the cytosol (C, arrowheads). Sections also were stained with periodic-acid Schiff (PAS) to label the glycocalyx in the brush border of proximal tubule cells. Exposure to the native compound resulted in a reduced apical brush border staining (D). Thin arrows clearly show a reduction in the purple color along the brush border of these heavily vacuolized cells (arrowheads). Debris within the tubular lumen again was seen in rats given the native compound (thick arrow). Rats exposed to congener C2 exhibited brighter PAS staining (E, thin arrows), indicative of a thicker apical brush border. Again, fewer large, intracellular vacuoles and less debris were seen in the tubular lumens. H&E-stained sections were scored using a modified Jablonski score with results shown in F. The reduced cellular injury that was seen in C2-treated rats was significantly lower than that seen in native gentamicin–treated rats. Localization of C2 or the native compound via indirect immunofluorescence revealed different localization patterns. Native gentamicin formed large aggregates around the nuclei of proximal tubule cells (G, arrowheads), with a few large punctate aggregates. Congener C2 localized in smaller, discrete punctate structures dispersed throughout the cytosol, forming fewer perinuclear aggregates (H). Bar = approximately 20 µm.
Six Days of Daily Exposure to the C2 Congener Induces Few Alterations in Kidney Morphology
A general histologic survey of the rat kidney tissues was conductednext to determine whether gross morphologic changes accompanieddifferences that were observed in renal function. Again, thesestudies were conducted after 7 d of exposure, or six doses ofgentamicin. This was to ensure that we would see any delayednephrotoxicity from the C2 component. These studies occurredat a time when the native gentamicin–treated animals wererecovering, as evidenced by a reduction in serum creatinine.Therefore, our studies underestimate the actual histologic damagethat was seen early for native gentamicin. H&E-stained kidneysections of rats that were exposed to the native gentamicincompound showed extensive proximal tubule cell damage (Figure 3B).Here, proximal tubules from native gentamicin–treatedrats contained cast material and shed cells in the lumen (thickarrow). In addition, many of the proximal tubule cells containedlarge vacuoles that localized around the nucleus (Figure 3B,arrowheads). In contrast, kidney sections from rats that weretreated with congener C2 exhibited normal morphology (Figure 3C).Tubular epithelial cells were cuboidal, and the lumens lackedcellular debris. Occasionally, vacuoles appeared within thetubular epithelia in C2-treated rats, but they generally weremuch smaller in size, fewer in number, and not localized tothe perinuclear area (Figure 3C, arrowheads). Rats that wereexposed to normal saline exhibited morphology similar to thatseen with the congener C2 (data not shown).
Thin histologic sections also were stained with PAS to localizethe glycocalyx that was associated with the brush border membraneof proximal tubules. Here, a violet color that is associatedwith the glycocalyx-rich microvilli helps delineate changesto the brush border. In rats that were exposed to the nativegentamicin compound, the proximal tubular brush border seemedto be reduced in size (Figure 3D, thin arrows). As with theH&E sections, these tubules contained intralumenal debris(Figure 3D, arrows) and intracellular vacuoles (arrowheads).The proximal tubular brush borders of rats that were exposedto congener C2 exhibited more abundant and enriched PAS staining,defining the taller apical membranes (Figure 3E, thin arrows),similar to that observed in saline-treated rats (data not shown).Results from a modified Jablonski score (Figure 3F) to assessdamage in the cortex showed that injury was significantly reducedin rats that were treated with C2 as compared with those thatwere treated with native gentamicin (P 0.05).
When tissues from rats that were exposed to the native commercialgentamicin compound or the C2 congener were processed for indirectimmunofluorescence localization of gentamicin, a differencein the intracellular accumulation and distribution emerged.The formation of myeloid bodies or cytosegresomes within proximaltubule cells has been a hallmark alteration associated withprolonged exposure to aminoglycosides (14,15). Here, formationof these structures in rats that were exposed to the nativegentamicin compound was seen (Figure 3G, arrowheads). In thesetissues, the lysosomes appeared swollen and greatly reducedin number, and they often were localized around the nucleus.In contrast, rats that were exposed to congener C2 exhibitedwhat would be deemed normal lysosomal morphology (Figure 3H).In these tissue sections, lysosomes seemed much more numerous,smaller in diameter, and diffusely distributed throughout thecytosol. They were not segregated around the nucleus as notedfor the native compound. Tissue sections from rats that weretreated with normal saline alone and processed identically forimmunofluorescence localization produced no fluorescence associatedwith gentamicin localization (data not shown).
Short-Term Exposure to the Individual Congeners Reveals No Difference in Early Nonlysosomal Trafficking
Previous studies from two laboratories have documented the importanceof cytosolic release of aminoglycosides in subsequent LLC-PK1cell injury (9,16). Therefore, we hypothesized the non-nephrotoxicC2 component of gentamicin might not be released as rapidlyor to the same extent as the other toxic gentamicin components.To test this hypothesis directly, we used our previously reportedendosomal/lysosomal quenching and Tyramide amplification techniquesto evaluate the early phase of cytosolic gentamicin trafficking(9). The results shown in Figure 4 indicate that we could discernno difference in the extent of cytosolic release 1 h after exposureto any gentamicin component. In all cells, for all congeners,there was rapid appearance of a homogeneous cytosolic distributionof gentamicin. Therefore, all components reached the cytosolrapidly, and these data could not explain the differences thatwere noted in cell injury among the various components.
Figure 4. Short-term exposure to native gentamicin. The individual cytotoxic congeners or C2 revealed no difference in early, nonlysosomal trafficking. Cells were exposed to the various congeners of gentamicin (1 mg/ml) for 1 h and processed for lysosomal quenching, and Tyramide Signal Amplification of gentamicin (see Materials and Methods). Trafficking to the endoplasmic reticulum (ER) was observed uniformly for all congeners. Cytosolic release was not yet detected at this early time point as is evident by the lack of localization within the nucleus. Bar = approximately 10 µm. A, C1; B, C1a; C, C2; D, C2a.
Large Perinuclear Accumulations of Cytotoxic Gentamicin Congeners Contain Elements of Both Lysosomes and the Golgi Complex
To study long-term differences in the intracellular distributionof gentamicin, we undertook additional studies in LLC-PK1 cells.After 24 h of continuous exposure to the native gentamicin mixture,congener C2, or the cytotoxic congeners and a 10,000 molecularweight Texas Red dextran to localize the lysosomes, cells werefixed and processed for the localization of the Golgi complexwith a FITC-conjugated lectin from Helix pomatia (17) and indirectimmunofluorescence localization of gentamicin with a Cy-5–conjugatedsecondary antibody. The staining patterns that were observedfor the native compound, congener C2 and congener C1a (as arepresentative of the altered morphology seen with all cytotoxiccongeners) are shown in Figure 5. Exposure to gentamicin, aseither the native mixture or the individual congeners, showedco-localization between the dextrans and gentamicin in lysosomes,as expected. In untreated cells or cells that received congenerC2, lysosomes seemed numerous, small, and evenly distributedthroughout the cytosol, as previously noted in the rat studies.In contrast, the lysosomes in cells that were exposed to cytotoxiccongeners or the native gentamicin mixture were located in aperinuclear position and swollen, as seen here with C1a. TheGolgi complex in cells that were treated with C2 but not thosethat were exposed to the cytotoxic congeners was faint. Somestaining in these cells also occurred in the periphery of thecell, likely as a result of recognition of carbohydrate moietiesat the cells’ surface. Cells that were exposed to thecytotoxic congeners produced a Golgi staining pattern that wasidentical to that of lysosomes and gentamicin as seen with C1a.Here, the intensity of the lectin staining was markedly increasedwith little or no staining within the remaining cytosol or thecell surface. In staining that was reminiscent of immunolocalizednative gentamicin in rats, cells that were treated with C1aexhibited large grape-like clusters around the nucleus and containedgentamicin and both lysosomal and Golgi complex markers. Cellsthat were exposed to C2 lacked these structures, and there wasno overlap between the lysosomal and Golgi complex markers.The staining pattern for the native compound was more heterogeneous,with accumulation patterns in between those seen for eitherC2 or C1a. These data suggest that the cytotoxic gentamicincongeners induce a trafficking abnormality after prolonged exposure.They further indicate that inclusion of the C2 component reducesthe intracellular trafficking abnormality that is induced bythe nephrotoxic congeners, because treatment with the nativegentamicin mixture showed results in between the two extremes.
Figure 5. Long-term exposure to the cytotoxic congeners induced coalescence of lysosomes and the Golgi complex in LLC-PK1 cells. Texas Red dextran (0.5 mg/ml) was given in the media during gentamicin exposure to label the lysosomes. In untreated cells and congener C2-treated cells, the lysosomes appear as small, discrete vesicles found throughout the cellular cytosol. Cells exposed to native gentamicin exhibited some perinuclear accumulation of gentamicin and Texas Red dextran, although both still largely were confined within smaller vesicles. Cells exposed to the cytotoxic congeners, as shown here with C1a, produced much larger swollen vesicles that encompassed the nucleus. When gentamicin was introduced to the cells and visualized, co-localization with the lysosomes occurred. Visualization of the ER-Golgi-intermediate-compartment/cis-Golgi complex with a lectin from Helix pomatia was faint in untreated cells, congener C2–treated cells, and the native gentamicin–treated cells. Movement of the lectin epitope during normal cellular trafficking may account for the faint staining that was seen at this concentration. Cells exposed to congener C1a exhibited intense staining of Helix pomatia, producing a pattern that was identical to that seen with the Texas Red dextran and gentamicin. Bar = approximately 10 µm.
Clinical Implications
Numerous strategies have been developed to minimize the nephrotoxicitythat is associated with aminoglycoside antibiotics. Althoughcertain aminoglycoside antibiotics are less nephrotoxic, allstill result in an unacceptably elevated level of nephrotoxicity,especially in high-risk patients. Approaches to limit this associatedtoxicity have included introduction of an agent to alter uptakeor intracellular trafficking of aminoglycosides. In this regard,high levels of dietary calcium (18) or the use of polyasparticacid has shown beneficial effects in animal models (19–22).Another approach has been purification of a nontoxic nativecomponent, or modification of the parent compounds to produceless toxic substances. This approach has yielded two commerciallyavailable aminoglycosides that have somewhat less nephrotoxicityand commonly are used today. For example, modification of KanamycinA yielded Amikacin, and Tobramycin was purified from Nebramycin(23).
Early clinical observations indicated the nephrotoxicity ofdifferent gentamicin manufacturing lots differed (12). Thisled Kohlhepp et al. (12) to initiate studies to determine therelative toxicity of the various gentamicin components in animalstudies. Unfortunately, at the time of their studies, it wasnot possible to separate gentamicin cleanly into its variouscomponents. Therefore, their studies were conducted using mixturesthat were enriched for a given component C1, C1a, or C2. Inaddition, the C2a component, a component of the native mixturethat we showed to have a high level of cellular toxicity, couldnot be identified in these early studies. Therefore, the actuallevel of cross-contamination reported in the study may representan underestimation, because no data as to where and to whatextent C2a was present in each of the enriched component mixturesexist.
Using improved technology, we were able to separate gentamicincomponents into pure fractions. Our data clearly identifiedthe C2 component as nontoxic in cell culture; the consistencybetween cell lines and the in vivo rat data confirm the nontoxicnature of the C2 component.
In our studies, the non-nephrotoxic C2 congener was found tobe just as potent against Gram-positive and multidrug-resistantGram-negative bacteria as native gentamicin or any of the individualcomponents. Therefore, the C2 congener could have a profoundclinical impact in the treatment of infectious diseases. Moreaggressive and broader therapeutic regimens that now could includepatients who previously were deemed unsuitable because of preexistingrisk factors may be developed. Moreover, understanding the mechanismsthat are inherent to toxicity may provide a means by which effectiveco-therapies can be developed. Finally, the congener C2 holdsgreat promise as a therapeutic agent for the treatment of geneticdiseases that arise from premature stop mutation defects, inwhich continuous daily doses to increase transcription of completeproteins may be possible (7,8). A series of studies have alreadyshown gentamicin’s potential benefit against diseasessuch as cystic fibrosis, Hurler’s syndrome, and Duchenneand Becker muscular dystrophies (7,8). Additional data indicatethat premature stop codon mutations in the IDUA and p53 genescan be suppressed by gentamicin or Amikacin but not by Tobramycin(24).
Mechanism
Cytosolic release, a pathway for gentamicin trafficking thatwas identified previously in our laboratory (9) and confirmedrecently by Servais et al. (16), was a characteristic sharedby all congeners in this study, including the non-nephrotoxicC2 congener. Although this potential explanation for the lackof toxicity that was seen with C2 was eliminated by our studies,a deleterious alteration to the Golgi complex and lysosomes,characterized by coalescing of the two into a large amorphouscompartment, was observed. These structures, presumed to be"myeloid bodies" that classically have been associated withgentamicin toxicity (14,15), were produced by the cytotoxiccongeners but not with the non-nephrotoxic C2 congener. Thisalteration was apparent in cultured cells after 24 h of continuousexposure and in proximal tubule epithelial cells in rats thatwere subjected to daily exposure for 6 d. This effect is similarto the trafficking effects that were induced by Brefeldin A,another fungal metabolite, in certain cell lines (25).
In cells that were exposed to the congener C2, staining withHelix pomatia was faint. This lectin is known to label the cis-Golgiand endoplasmic reticulum–Golgi intermediate compartments.Hence, shuttling of the lectin epitope between the endoplasmicreticulum and Golgi complex could account for the reduced intensityat the concentration used. In contrast, the intensity that wasseen at the Golgi complex when the cytotoxic congeners wereused (represented by C1a in Figure 5) was very intense and localizedin grape-like clusters around the nucleus. We postulate thatthis arises from an accumulation of Golgi elements in a stagnantorganelle. Of note, these alterations occurred within the entirelysosomal pool and not just within newer endosomal bodies thatformed after gentamicin exposure, because preincubation withdextrans to label lysosomes, followed by a chase and subsequentgentamicin exposure for 24 h, produced identical results (datanot shown). Similar structures also were seen in our previousstudies and allowed for easy identification of the Golgi complex(9,26). The affects on this pathway were much more severe whencultured cells were exposed to the individual cytotoxic congenersas compared with the native compound. This observation againstrongly suggests that the inclusion of the C2 congener, inthe native mixture, affords protection against the effects ofthe cytotoxic congeners. Therefore, on the basis of the datapresented here, the naturally occurring variation in the percentageof C2 congener composition of native gentamicin from commerciallot to lot could account for the previous clinical observationsregarding variable nephrotoxicity (13).
The gentamicin congener C2 was isolated from native gentamicinand shown to induce no cellular injury and no nephrotoxicityin a rat model of gentamicin toxicity while retaining normalbactericidal properties. The C2 congener failed to induce theintracellular trafficking abnormalities that were seen withother congeners and native gentamicin. Taken together, thesedata indicate the C2 congener of gentamicin may be of clinicalvalue for the treatment of serious infections, especially inpatients who are at high risk for developing ARF.
Acknowledgments
This work was supported by a VA Merit Review Grant awarded toB.A.M. Additional support came from the National Institute ofDiabetes and Digestive and Kidney Diseases via a George M. O'Briencenter of excellence award (P50-DK61594 and P01-DK53465 to B.A.M.).
We acknowledge support from the Indiana Genomics Initiative(INGEN), presented to Indiana University Medical Center by theEli Lilly Foundation.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR: Epidemiology of severe sepsis in the United States: Analysis of incidence, outcome, and associated costs of care.
Crit Care Med 29
: 1303
–1310, 2001[CrossRef][Medline]
Thakar CV, Yared JP, Worley S, Cotman K, Paganini EP: Renal dysfunction and serious infections after open-heart surgery.
Kidney Int 64
: 239
–246, 2003[CrossRef][Medline]
Ridley RG: Research on infectious diseases requires better coordination.
Nat Med 10[Suppl]
: S137
–S140, 2004[CrossRef][Medline]
Levy SB, Marshall B: Antibacterial resistance worldwide: Causes, challenges and responses.
Nat Med 10[Suppl]
: S122
–S129, 2004[CrossRef][Medline]
Wilschanski M, Yahav Y, Yaacov Y, Blau H, Bentur L, Rivlin J, Aviram M, Bdolah-Abram T, Bebok Z, Shushi L, Kerem B, Kerem E: Gentamicin-induced correction of CFTR function in patients with cystic fibrosis and CFTR stop mutations.
N Engl J Med 349
: 1433
–1441, 2003[Abstract/Free Full Text]
Sandoval RM, Molitoris BA: Gentamicin traffics retrograde through the secretory pathway and is released in the cytosol via the endoplasmic reticulum.
Am J Physiol Renal Physiol 286
: F617
–F624, 2004[Abstract/Free Full Text]
Moestrup SK, Cui S, Vorum H, Bregengard C, Bjorn SE, Norris K, Gliemann J, Christensen EI: Evidence that epithelial glycoprotein 330/megalin mediates uptake of polybasic drugs.
J Clin Invest 96
: 1404
–1413, 1995[Medline]
Saito A, Pietromonaco S, Loo AK, Farquhar MG: Complete cloning and sequencing of rat gp330/"megalin," a distinctive member of the low density lipoprotein receptor gene family.
Proc Natl Acad Sci U S A 91
: 9725
–9729, 1994[Abstract/Free Full Text]
Kohlhepp SJ, Loveless MO, Kohnen PW, Houghton DC, Bennett WM, Gilbert DN: Nephrotoxicity of the constituents of the gentamicin complex.
J Infect Dis 149
: 605
–614, 1984[Medline]
Jablonski P, Howden BO, Rae DA, Birrell CS, Marshall VC, Tange J: An experimental model for assessment of renal recovery from warm ischemia.
Transplantation 35
: 198
–204, 1983[Medline]
Houghton DC, Hartnett M, Campbell-Boswell M, Porter G, Bennett W: A light and electron microscopic analysis of gentamicin nephrotoxicity in rats.
Am J Pathol 82
: 589
–612, 1976[Abstract]
Beauchamp D, Gourde P, Bergeron MG: Subcellular distribution of gentamicin in proximal tubular cells, determined by immunogold labeling.
Antimicrob Agents Chemother 35
: 2173
–2179, 1991[Abstract/Free Full Text]
Servais H, Van Der Smissen P, Thirion G, Van der Essen G, Van Bambeke F, Tulkens PM, Mingeot-Leclercq MP: Gentamicin-induced apoptosis in LLC-PK1 cells: Involvement of lysosomes and mitochondria.
Toxicol Appl Pharmacol 206
: 321
–333, 2005[CrossRef][Medline]
Pavelka M, Ellinger A: Localization of binding sites for concanavalin A,
Ricinus communis I and Helix pomatia lectin in the Golgi apparatus of rat small intestinal absorptive cells. J Histochem Cytochem 33
: 905
–914, 1985
Bennett WM, Elliott WC, Houghton DC, Gilbert DN, DeFehr J, McCarron DA: Reduction of experimental gentamicin nephrotoxicity in rats by dietary calcium loading.
Antimicrob Agents Chemother 22
: 508
–512, 1982[Abstract/Free Full Text]
Swan SK, Kohlhepp SJ, Kohnen PW, Gilbert DN, Bennett WM: Long-term protection of polyaspartic acid in experimental gentamicin nephrotoxicity.
Antimicrob Agents Chemother 35
: 2591
–2595, 1991[Abstract/Free Full Text]
Whittem T, Schnellmann RG, Ferguson DC: Poly-L-aspartic acid does but triiodothyronine does not protect against gentamicin-induced cytotoxicity in the porcine kidney cell line LLC-PK1.
J Pharmacol Exp Ther 262
: 834
–840, 1992[Abstract/Free Full Text]
Kishore BK, Lambricht P, Laurent G, Maldague P, Wagner R, Tulkens PM: Mechanism of protection afforded by polyaspartic acid against gentamicin-induced phospholipidosis. II. Comparative in vitro and in vivo studies with poly-L-aspartic, poly-L-glutamic and poly-D-glutamic acids.
J Pharmacol Exp Ther 255
: 875
–885, 1990[Abstract/Free Full Text]
Beauchamp D, Laurent G, Maldague P, Abid S, Kishore BK, Tulkens PM: Protection against gentamicin-induced early renal alterations (phospholipidosis and increased DNA synthesis) by coadministration of poly-L-aspartic acid.
J Pharmacol Exp Ther 255
: 858
–866, 1990[Abstract/Free Full Text]
Budavari S, O’Neil M, Smith A, Heckelman P:
The Merck Index, 11th Ed., Merck & Co., Inc., 1989
Keeling KM, Bedwell DM: Clinically relevant aminoglycosides can suppress disease-associated premature stop mutations in the IDUA and P53 cDNAs in a mammalian translation system.
J Mol Med 80
: 367
–376, 2002[CrossRef][Medline]
Lippincott-Schwartz J, Yuan L, Tipper C, Amherdt M, Orci L, Klausner RD: Brefeldin A’s effects on endosomes, lysosomes, and the TGN suggest a general mechanism for regulating organelle structure and membrane traffic.
Cell 67
: 601
–616, 1991[CrossRef][Medline]
Sandoval RM, Dunn KW, Molitoris BA: Gentamicin traffics rapidly and directly to the Golgi complex in LLC-PK(1) cells.
Am J Physiol Renal Physiol 279
: F884
–F890, 2000[Abstract/Free Full Text]
This article has been cited by other articles:
R. Azimov, N. Abuladze, P. Sassani, D. Newman, L. Kao, W. Liu, N. Orozco, P. Ruchala, A. Pushkin, and I. Kurtz G418-mediated ribosomal read-through of a nonsense mutation causing autosomal recessive proximal renal tubular acidosis
Am J Physiol Renal Physiol,
September 1, 2008;
295(3):
F633 - F641.
[Abstract][Full Text][PDF]
S. Bolisetty and A. Agarwal Subclinical kidney injury incites endotoxin hyperresponsiveness
Am J Physiol Renal Physiol,
July 1, 2007;
293(1):
F41 - F42.
[Full Text][PDF]
Top
Abstract
Introduction
0.05. All reported values are mean ± SEM. 
