The antimicrobial peptide cathelicidin protects the urinary tract - Nature

© 2006 Nature Publishing Group http://www.nature.com/naturemedicine

ARTICLES

The antimicrobial peptide cathelicidin protects the urinary tract against invasive bacterial infection Milan Chromek1, Zuzana Slamova´1, Peter Bergman2, La´szlo´ Kova´cs3, L’udmila Podracka´4, Ingrid Ehre´n5, Tomas Ho¨kfelt6, Gudmundur H Gudmundsson7, Richard L Gallo8, Birgitta Agerberth2 & Annelie Brauner1 The urinary tract functions in close proximity to the outside environment, yet must remain free of microbial colonization to avoid disease. The mechanisms for establishing an antimicrobial barrier in this area are not completely understood. Here, we describe the production and function of the cathelicidin antimicrobial peptides LL-37, its precursor hCAP-18 and its ortholog CRAMP in epithelial cells of human and mouse urinary tract, respectively. Bacterial contact with epithelial cells resulted in rapid production and secretion of the respective peptides, and in humans LL-37/hCAP-18 was released into urine. Epithelium-derived cathelicidin substantially contributed to the protection of the urinary tract against infection, as shown using CRAMP-deficient and neutrophildepleted mice. In addition, clinical E. coli strains that were more resistant to LL-37 caused more severe urinary tract infections than did susceptible strains. Thus, cathelicidin seems to be a key factor in mucosal immunity of the urinary tract.

Under healthy conditions, the urinary tract, except for the external part of the urethra, is free from microbial colonization. Mechanisms involved in the clearance of bacteria include mechanical factors, such as urine flow and regular bladder emptying, chemical-defense components of epithelia and, upon bacterial stimulation, epithelial shedding and influx of effector immune cells. Components of epithelial defense include antimicrobial substances, chemokines and cytokines. Recently, antimicrobial peptides have been shown to have an important role in the first line of mucosal immunity1. Two main families of antimicrobial peptides in mammals, the defensins and the cathelicidins, are expressed in immune cells and at epithelial surfaces2,3. Gene-deficient models have confirmed the importance of cathelicidin in protection of skin4 and b-defensin-1 in respiratory tract5. Furthermore, b-defensin-1 has been suggested to be a major antimicrobial peptide in urogenital tissues6–8. Cathelicidins are expressed in circulating neutrophils and myeloid bone marrow cells, in epithelial cells of the skin and gastrointestinal tract, as well as in the epididymis and lungs9–14. In contrast to the multiple defensins, only one cathelicidin gene, CAMP, has been found in humans12. The gene product is synthesized as a propeptide, designated human cationic antimicrobial peptide-18 (hCAP-18) or pro-LL-37. The neutrophil propeptide is cleaved extracellularly into cathelin and the C-terminal peptide LL-37 (ref. 15). Both cleavage products have a broad antimicrobial effect with complementary action16. Likewise, there is only one mouse gene encoding cathelicidin, Camp (also known as CnLp), and it is very similar to the human gene.

The mouse cathelicidin proform is processed to the mature bioactive peptide CRAMP (cathelin-related antimicrobial peptide)14,17. LL-37 and CRAMP are amphipathic, a-helical peptides that preferentially bind to negatively charged groups of the outer leaflet of the bacterial membrane. This binding results in damage to the membrane, and different models explaining the precise mechanism of the peptides’ action have been proposed18. In addition to antimicrobial activity, antimicrobial peptides were suggested to be multifunctional modulators of different immune reactions19–21. Notably, cathelicidins in vitro inhibited the growth of pathogens that have a major role in the urinary tract, whereas they were rather ineffective against urogenital commensal bacteria22. Therefore, we hypothesized that production of cathelicidin in vivo could be a natural mechanism protecting the urinary tract. Here, we show that epithelial cells of the urinary tract in humans and mice produce cathelicidins. We also provide evidence of rapid production and secretion of cathelicidin upon contact with bacteria. Moreover, we show the relevance of epithelium-derived as well as neutrophil-derived cathelicidin using CRAMP-deficient mice, neutrophil depletion and in vitro experiments. Furthermore, our data indicate that severity of bacterial invasion is linked to bacterial resistance to cathelicidin. RESULTS Cathelicidin is present in urine To analyze the presence of the human cathelicidin LL-37/hCAP-18, we examined urine from both healthy children and children with urinary

1Department

of Clinical Microbiology, Microbiology and Tumorbiology Center, Karolinska University Hospital and Karolinska Institutet, SE-171 76 Stockholm, Sweden. of Medical Biochemistry and Biophysics, Karolinska Institutet, SE-171 77 Stockholm, Sweden. 3Department of Pediatrics, Comenius University School of Medicine, 833 40 Bratislava, Slovakia. 4Department of Pediatrics, Pavol Jozef Sˇafa´rik University School of Medicine, 040 66 Kosˇice, Slovakia. 5Department of Urology, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden. 6Department of Neuroscience, Karolinska Institutet, SE-171 77 Stockholm, Sweden. 7Biology Institute, University of Iceland, 101 Reykjavik, Iceland. 8Division of Dermatology, University of California, San Diego, California 92161, USA. Correspondence should be addressed to A.B. ([email protected]). 2Department

Received 27 February; accepted 7 April; published online 4 June 2006; doi:10.1038/nm1407

636

VOLUME 12

[

NUMBER 6

[

JUNE 2006

NATURE MEDICINE


ARTICLES

a

b

c

d

e

f

g

h

Figure 1 Immunohistochemical staining of sections from healthy human renal cortical tissue (a–d), a piece of human renal cortex incubated in cell culture medium for 24 h (e,f), and a piece infected with uropathogenic E. coli for the same time (g,h). The sections are stained using polyclonal antibody to LL-37 (a,c–e,g) and normal rabbit immunoglobulin (b,f,h). In the healthy renal tissue, we found LL-37/hCAP-18 immunoreactivity in the hyaline substance in the lumen of renal tubuli (a,c, brown) and in neutrophils (d) indicated by arrows. Parallel sections of the kidney (a,b) are shown. We did not observe any staining in the hyaline substance when we used normal rabbit immunoglobulin instead of primary antibody (b). After 24 h of in vitro incubation, tubular epithelial cells displayed morphological signs of cell death (e, arrow). Moreover, in the presence of bacteria, epithelial cells stained positive for LL-37/hCAP-18 (g, brown, arrow). Parallel sections of the kidney (e–h) are shown. We did not detect any staining when we used normal rabbit immunoglobulin instead of primary antibody (f,h). Scale bars in a, b and d, 50 mm; in c, e, f, g and h, 20 mm.

histochemistry. Although all the homogenized samples of renal tissue were positive for LL-37/hCAP-18 by ELISA, resident renal cells did not stain for cathelicidin; however, we found positive staining in the hyaline substance in the lumen of renal tubuli (Fig. 1a–c) and in neutrophils (Fig. 1d). Accordingly, normal human proximal tubular cells (hPTC) from three individuals, a human renal epithelial cell line (A498) and human uroepithelial cell lines (J82 and UROtsa) were positive for CAMP mRNA transcript but negative for LL-37/hCAP-18 (Supplementary Fig. 2).

tract infection. Cathelicidin was detected in all samples of 28 healthy children (median, 0.3 ng/ml; range, 0.2–5.9 ng/ml). In 29 children with pyelonephritis or cystitis, cathelicidin levels were significantly increased (median, 2.4 ng/ml; range, 0–312.5 ng/ml; P o 0.001; Supplementary Fig. 1 online). All children with urinary tract infections had leukocyturia, as measured microscopically in urinary sediment (mean, 32 leukocytes per high-power field, s.d., 19.6). We found a small positive correlation between cathelicidin and leukocytes (R ¼ 0.60, P o 0.01) and myeloperoxidase (MPO), a typical neutrophil enzyme23 (R ¼ 0.62, P o 0.01), suggesting additional nonleukocyte origins of cathelicidin in urine. Epithelial cells of the urinary tract express cathelicidin We examined homogenized pieces of noninfected kidneys from three nephrectomized individuals for the presence of human cathelicidin LL-37/hCAP-18 at the mRNA and protein level with real-time PCR and ELISA, respectively. We focused on pelvis renalis, the central part of the kidney with uroepithelium, which is the first epithelial surface to come in contact with bacteria. We also analyzed renal cortex, a highly vascularized peripheral part of the kidney, a common bacterial target and the site of massive inflammation during pyelonephritis. All tissue samples were positive for CAMP mRNA and cathelicidin peptide (Supplementary Fig. 2 online). Levels of MPO in kidney tissues correlated with levels of cathelicidin (R ¼ 0.91, P o 0.01). To investigate the production of human cathelicidin at the cellular level, we analyzed renal samples from 12 individuals by immuno-

NATURE MEDICINE

VOLUME 12

[

NUMBER 6

[

JUNE 2006

Bacteria induce rapid secretion of cathelicidin To investigate cathelicidin expression during bacterial infection, we incubated healthy samples of renal cortex with uropathogenic E. coli for 24 h, and performed immunohistochemistry to determine the presence of human cathelicidin LL-37/hCAP-18. After infection, cathelicidin was clearly visible in the cytoplasm of tubular epithelial cells (Fig. 1g). To confirm the epithelial source of cathelicidin, as well as to exclude the possibility of peptide uptake from degranulated leukocytes, we infected renal epithelial (both A498 and hPTC) and uroepithelial (both J82 and UROtsa) cells with uropathogenic E. coli. Upon contact with bacteria, we observed a rapid increase in CAMP mRNA in all cell types, with maximum levels detected at 5 min after stimulation. The initial increase was followed by a rapid decrease of the transcript levels, and, after infection for 135 min, the level of CAMP mRNA was the same as in control cells (Fig. 2a). At the protein level, both pro-LL-37 (18 kDa) and LL-37 (4.5 kDa) were released into the cell medium after 5 min of bacterial exposure. In contrast to the observed decrease in mRNA, the levels of secreted cathelicidin increased at later time points analyzed (Fig. 2b). After 24 h of infection, mRNA decreased between 42% and 64% compared to uninfected cells, indicating inhibition of transcription or increased decay of mRNA (Fig. 2c). Yet in the same cells, we showed release of the peptide after bacterial infection (Fig. 2d). Bacterial infection did not influence the viability of cells (data not shown). Mice secrete cathelicidin during urinary tract infection To investigate the in vivo significance of our cell-culture results, we used a mouse model of ascendant pyelonephritis and determined the presence of the mouse cathelicidin CRAMP in kidneys by immunostaining. In noninfected mice, resident renal cells did not stain for CRAMP, whereas leukocytes in blood vessels stained positive (data not shown). During urinary tract infection, we observed areas of the kidney with visible bacteria, but without massive leukocyte infiltration, corresponding to the early stages of urinary tract infection (Fig. 3a–d). In these areas, tubular epithelial cells stained positive

637

ARTICLES

Uninfected Infected

*

60

b

50

*

40 30

c +

u

inf 5′ 15′ 45′ 135′

19 17 11 6

20 10

3

Uninfected

Infected

1.2

d

+

1.0 0.8

A498 hPTC u i u i

J82 u i

UROtsa u i

19 17 11 6

0.6 0.4 0.2

3

0.0

0 0


1.4 CAMP/GAPDH mRNA (normalized ratio)

70 CAMP/GAPDH mRNA (ratio × 1,000)

a

15 30 45 60 75 90 105 120 135 Time (min)

A498

hPTC J82 Cell type

UROtsa

Figure 2 Cathelicidin LL-37/hCAP-18 in vitro. (a) mRNA encoding LL-37/hCAP-18 in uroepithelial cells (UROtsa) without bacteria (continuous line) and after infection with uropathogenic E. coli (dotted line). All cell lines studied, both uroepithelial and renal epithelial, reacted similarly after bacterial exposure. Data are presented as median and range of CAMP/GAPDH 1,000 mRNA ratios from two experiments in which each sample was analyzed in triplicate. *P o 0.05. (b) Western blot of supernatants from uroepithelial cells (UROtsa) without bacteria (uninfected, u) and after 5, 15, 45 and 135 min infection with uropathogenic E. coli (infected, inf). Similar results were obtained in all cell types studied, both uroepithelial and renal epithelial. Synthetic LL-37 served as a positive control (+). Molecular weights are in kDa. Discrepancy in the migration of the detected peptide and the synthetic LL-37 is due to salt content in cell culture medium (data not shown). (c) mRNA encoding LL-37/hCAP-18 in renal epithelial cells (A498 and hPTC) and uroepithelial cells (J82 and UROtsa) without bacteria (open bars) and after 24-h infection with uropathogenic E. coli (gray bars). mRNA levels from noninfected cells are considered 1 and relative mRNA levels are calculated. P o 0.01, compared with uninfected cells at 24 h. Data are presented as median and range of normalized CAMP/GAPDH mRNA ratios from three independent experiments. (d) Western blot of supernatants from renal epithelial cells (A498 and hPTC) and uroepithelial cells (J82 and UROtsa) without bacteria (uninfected, u) and after 24-h infection with uropathogenic E. coli (infected, i). Synthetic LL-37 served as a positive control (+). Molecular weights are in kDa. Discrepancy in the migration of the detected peptide and the synthetic LL-37 is due to salt content in cell culture medium (data not shown).

for CRAMP with apparent peptide release into the lumen of the renal tubuli (Fig. 3c,d). In areas of massive infection, on the other hand, invading polymorphonuclear leukocytes destroyed the normal kidney structure (Fig. 3e–h). Leukocytes stained strongly for CRAMP, as they were a major source of cathelicidin during the late stages of infection (Fig. 3g,h). These findings confirmed two sources of mouse cathelicidin in renal tissue: epithelial cells and neutrophils. Cathelicidin peptides kill uropathogenic E. coli To investigate the significance of cathelicidin production within the urinary tract, we measured the sensitivity of the most common uropathogen, E. coli, to the synthetic peptides LL-37 and CRAMP. We tested the clinical isolate E. coli CFTO73, which we used for both mouse studies and cell-culture experiments. The minimal inhibitory concentration (MIC) was 8 mM for both peptides, which corresponds to 36 mg/ml LL-37 and 31 mg/ml CRAMP. The peptide concentration killing 50% of the bacteria, EC5024, was 4 mM for LL-37 and 8 mM for CRAMP.

Epithelial cathelicidin protects the urinary tract To address the relevance of cathelicidin derived from epithelial cells, we studied the antimicrobial activity of urinary bladder mucosa. We infected CRAMP-producing (Camp+/+) and CRAMP-deficient (Camp–/–) mice with uropathogenic E. coli. We analyzed the number of bacteria that attached to urinary bladders 1 h after infection, when no influx of leukocytes into the urinary tract occurs25,26. We found significantly more bacteria attached to bladders of Camp–/– mice (P o 0.05, Fig. 4a). The infection rate, as measured by the presence of bacteria in the urinary tract of mice 48 h after challenge, was also significantly higher in Camp–/– mice (P o 0.05) and did not seem to be influenced by neutrophil-derived cathelicidin. The infection rate remained the same irrespective of whether the mice had neutrophils or whether we induced transient neutropenia by treatment with neutrophil-specific monoclonal antibody (Fig. 4b). Cathelicidin significantly contributed to the antimicrobial activity of epithelial cells in vitro, as the number of bacteria surviving on epithelial cells from Camp–/– mice after 30 min was significantly higher than the number

Figure 3 Immunofluorescent staining of two sections (one section is shown in a–d and the other is shown in e–h) from the renal cortex of a * * * * NMRI mouse at 24 h of pyelonephritis. Sections are stained with propidium iodide (a,e, blue), + + + + neutrophil-specific antibody (b,f, red), antibody to CRAMP (c,g, green), and images showing these three sets of panels merged (d,h). The upper panel represents marginal areas of infection corresponding to the early stage of inflammation. Bacteria are stained in the lumen of renal tubulus by propidium iodide (a, arrowheads) and stimulate the production and release of CRAMP (c, green, arrowheads). A neutrophil with typical segmented nucleus (b, starred arrowhead) and positive for a neutrophil marker (c, starred arrowhead) is also positive for CRAMP (c,d, starred arrowhead). Cross indicates the lumen of the renal tubulus. The lower panel shows an area of massive infection with abundance of bacteria, destruction of normal parenchyma and accumulation of neutrophils (f) that exhibit CRAMP immunoreactivity (g,h). After blocking of antibody with synthetic CRAMP (1–39) peptide, we did not observe any staining, and, likewise, there was no staining in CRAMP-deficient mice (data not shown). Scale bars in a–d, 20 mm; in e–h, 50 mm.

638

a

b

c

d

e

f

g

h

VOLUME 12

[

NUMBER 6

[

JUNE 2006

NATURE MEDICINE

ARTICLES Figure 4 Relevance of epithelium-derived Neu (+) 100,000 100 cathelicidin for the protection of the urinary 100 Neu (–) tract against infection. (a) Number of bacteria 75 10,000 80 attached to urinary bladders of CRAMP-producing 60 50 1,000 +/+ –/– (Camp ) and CRAMP-deficient (Camp ) mice 40 25 100 1 h after bacterial inoculation. CFU, colony20 forming units. Data are presented as median and 0 10 0 +/+ –/– +/+ –/– +/+ –/– range from one of two experiments with three to Camp Camp Camp Camp Camp Camp Mice Mice Mouse PTC four mice per group. P o 0.05. (b) Infection rate of Camp+/+ and Camp–/– mice with (Neu (+)) or without neutrophils (Neu (–)). The infection rate was higher in Camp–/– mice, irrespective of the presence of neutrophils (P o 0.05). Data are presented as percentage of infected mice versus mice challenged with bacteria from three independent experiments using a total of 36 mice. (c) Survival of bacteria attached to Camp+/+ and Camp–/– mouse proximal tubular cells (Mouse PTC). Percentage of viable bacteria attached to the cells at 30 min of incubation in relation to the counts at 10 min is depicted. Data are presented as mean + s.d. from two independent experiments with nine samples per group. P o 0.05.

b

DISCUSSION Here we show that cathelicidin is constitutively expressed in the urinary tract. Immediately after contact with bacteria, epithelial cells substantially increased the synthesis and secretion of cathelicidin, protecting the urinary tract from bacterial adherence. In later stages of inflammation, invading leukocytes became the main source of the peptide. Clinical E. coli strains resistant to human cathelicidin were more prone to invade the upper urinary tract than susceptible strains. In vivo experiments and mathematical simulations have suggested that urine flow with regular emptying of the bladder is only one of the defense mechanisms of the urinary tract. Other antibacterial factors are essential to keep this niche free of microbes27,28. Bacterial attachment results in exfoliation of host bladder epithelial cells29. Moreover, epithelial cells of the urinary tract have been shown to possess antibacterial properties25, but the active antimicrobial components have not been systematically characterized. Some factors of innate immunity, however, have been shown to interfere with adherence30,31, inhibit growth32 or directly kill uropathogens6–8. In addition to these factors, our study shows the crucial importance of the antimicrobial peptide cathelicidin in defense of the urinary tract.

NATURE MEDICINE

VOLUME 12

[

NUMBER 6

[

JUNE 2006

a

b Percent of body weight

Resistance to LL-37 correlates with invasive infection To investigate the importance of cathelicidin production in relation to bacterial invasion, we tested 35 E. coli strains for sensitivity to synthetic LL-37. E. coli strains from individuals with more invasive infection of the upper urinary tract, pyelonephritis, were significantly more resistant to LL-37 than strains from children with infection of the lower urinary tract, cystitis (P o 0.05; Fig. 6).

We localized the production of human cathelicidin to resident uroepithelial and tubular epithelial renal cells. This was shown by realtime PCR analysis of mRNA levels and, after bacterial infection, by western blot and immunohistochemical analyses of protein levels. In noninfected tissue, we detected cathelicidin only in the hyaline substance inside renal tubuli. This finding is in line with earlier data on b-defensin-1 (ref. 6), which also could be found only in the lumen of renal tubuli without obvious localization inside cells. Our results indicate low constitutive expression of cathelicidin in renal epithelial and uroepithelial cells. Upon contact with bacteria, levels of CAMP mRNA increased, indicating either rapid transcriptional activation or sudden inhibition of constitutive mRNA degradation. Notably, within minutes, we detected release of cathelicidin. This prompt release did not seem to depend on the increase of mRNA. Regulation at the level of translation, for example, by unblocking of peptide synthesis, is likely. Regulation of translation could also explain why we could detect CAMP mRNA but not cathelicidin peptide in uninfected cells. Thus, when sensing the presence of bacteria, innate antimicrobial defenses seem to be mobilized by affecting transcription, translation and release. This fast and multilevel regulation emphasizes the importance of cathelicidin as a first line of mucosal defense. After an initial rapid increase, we observed a decrease in the ratio of CAMP to GAPDH mRNA to levels below those seen in noninfected cells. Throughout infection, the housekeeping gene GAPDH remained unaffected (data not shown). The reduction in CAMP mRNA levels could therefore be attributed either to inhibition of cathelicidin transcription or to mRNA decay exceeding synthesis. Transcriptional downregulation of CAMP mRNA has been documented during

CFU/kidney

Neutrophil cathelicidin influences severity of infection In Camp–/– mice with urinary tract infection, the number of bacteria in kidneys and the size of kidneys were significantly higher than those of Camp+/+ mice (P o 0.05; Fig. 5). Correspondingly, Camp–/– mice had also more severe systemic signs of infection, for example, weight loss and increased mortality resulting from septicemia. Camp–/– mice lost 11.2% (s.d., 7.3) of their body weight during 2 d of infection, whereas wild-type mice lost 3.3% (s.d., 5.1) of their body weight during the same time (P o 0.05). In addition, 3 out of 18 CRAMPdeficient mice (17%) died as a result of septicemia, as confirmed by bacterial cultures from blood. In contrast, none of the wild-type mice died during the experiments. All of these parameters seemed to be influenced by cathelicidin produced by neutrophils, as we did not observe corresponding differences in neutropenic mice (data not shown).

Percent bacteria 30 min/10 min

CFU/bladder


on Camp+/+ cells (P o 0.05; Fig. 4c). The number of bacteria attached at 10 min did not differ (data not shown).

c

Percent of inoculated mice

a

100,000,000 10,000,000 1,000,000 100,000 10,000 1,000 100 10 Camp +/+

Camp –/–

Mice

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Camp +/+

Camp –/–

Mice

Figure 5 The course of experimental urinary tract infection in CRAMPproducing Camp+/+ and CRAMP-deficient Camp–/– mice. The number of bacteria in kidneys (a) and the size of kidneys (b) at 48 h of infection were higher in Camp–/– mice. P o 0.05. Accordingly, systemic response to infection was stronger in Camp–/– mice (data not shown). The differences were observed only in the presence of neutrophils and not after neutrophil depletion (data not shown). Three independent experiments were performed with five to six mice per group. Data are presented as medians with dot plots (a) and mean + s.d. (b) from representative experiments.

639

ARTICLES respectively. During infection, epithelial cells rapidly increase production of cathelicidin peptide to protect the urinary tract from bacterial invasion, and the ‘second wave’ of cathelicidin comes with invading leukocytes. In summary, our data suggest that antimicrobial peptides of the cathelicidin family are important in bacterial clearance of the urinary tract.

120

MIC (µg/ml)

100 80 60 40 20

METHODS

0


Cystitis

APN

Figure 6 Sensitivity of uropathogenic E. coli to synthetic LL-37 peptide expressed as a minimal inhibitory concentration (MIC) of LL-37. Strains isolated from more invasive infection, pyelonephritis (APN, n ¼ 21), were more resistant to LL-37 (P o 0.05) than strains isolated from infection of the lower urinary tract, cystitis (n ¼ 14). Individual levels and means are shown.

bacterial infection in the gut and in a cervical epithelial cell line33,34. This was suggested as a novel immune escape mechanism promoting persistence of bacteria33. The observed decrease in mRNA could be a result of negative feedback after the increased secretion of the cathelicidin peptide. Indeed, we have observed that LL-37 was able to downregulate its own transcription (P.B., unpublished data). Such a negative-feedback mechanism could be an important way of controlling the levels of antimicrobial peptides, as they are cytotoxic to eukaryotic cells at high concentrations35. Neutrophils are known to produce cathelicidin36 and may occasionally be present in the healthy tissue or urine. To distinguish the source of cathelicidin in our experiments, we analyzed the urine samples for the presence of visible neutrophils and MPO. Although we found a positive correlation between cathelicidin, MPO and leukocyturia, the small correlation coefficients suggested other sources of cathelicidin, likely epithelial cells. Once reaching the urinary tract, the majority of bacteria are washed away by urine flow. To initiate infection, bacteria need to attach and to resist mucosal defense mechanisms25. Our experiments showed that the antimicrobial properties of the urinary tract epithelium of CRAMP-deficient mice were substantially compromised. As a consequence, Camp–/– mice were easier to infect. This primary epithelial defense was not influenced by neutrophils, as the infection rate was the same in mice with neutrophils and during transient neutropenia. Epithelium-derived cathelicidin, therefore, seems to have an important role in the first line of defense against attaching bacteria, despite its low concentration in vivo. Such a discrepancy has been observed before and may be attributed to higher activity in the specific chemical environment37, the synergy between various antimicrobial peptides38 or the very close proximity to epithelial cells. Similar to our finding, in another organ system, the skin, epithelial production of cathelicidin was shown to be of major importance39. Once mice were infected, however, the severity of infection was influenced by cathelicidin derived from neutrophils. Therefore, we suggest that cathelicidin participates in different immune processes during urinary tract infection; direct antimicrobial defense of epithelium, recruitment of immune system cells and killing of bacteria by neutrophils. Notably, we observed substantial variability in the sensitivity of clinical E. coli isolates to LL-37. Pyelonephritic strains were more resistant to cathelicidin than cystitic strains. Resistance to LL-37 could therefore be linked to the invasiveness of uropathogenic E. coli. We conclude that the cathelicidins LL-37/hCAP-18 and CRAMP are expressed in epithelial cells in the human and mouse urinary tract,

640

Peptides and antibodies. LL-37, CRAMP (34 amino acids) and an extended form, CRAMP (1–39), were synthesized by Innovagen. We prepared polyclonal rabbit antibodies to LL-37 and CRAMP (1–39) as previously described12 and by Innovagen, respectively. We prepared the monoclonal LL-37–specific antibody as previously described40. Subjects. The ethics committees of the Karolinska Hospital and Karolinska Institute approved the study and informed consent was obtained. The clinical study included 35 children with acute urinary tract infection and 28 healthy counterparts without symptoms or signs of any renal disease. The diagnostic criterion of acute urinary tract infection in the study was the presence of Z105 E. coli/ml freshly voided urine. Except for bacteriuria, the diagnostic criteria of acute pyelonephritis were: a body temperature of Z38 1C and laboratory signs of systemic inflammation, either of C-reactive protein Z20 mg/L or erythrocyte sedimentation rate of Z20 mm/h. We obtained healthy tissue from different parts of the kidney of 12 individuals nephrectomized because of a localized malignant tumor affecting the kidney. Mice. We used Bki NMRI, C57Bl/6 and 129/SvJ mice (Charles River). We generated CRAMP-deficient (Camp–/–) mice as previously described4. Briefly, after targeted disruption of Camp by homologous recombination, we injected 129/SvJ embryonic stem cells into C57Bl/6 blastocysts and transferred them into foster mothers. We crossed chimeric offspring with C57Bl/6 females. Afterward, we performed backcrosses into 129/SvJ for seven generations. We used 129/SvJ mice (Camp+/+) as a control. We did not observe any difference in the course of urinary tract infection when we compared the parental strains 129/SvJ and C57Bl/6 (data not shown). Cells. We isolated normal human and mouse proximal tubular cells (PTC) according to a modified protocol based on a previously described protocol41 (described in Supplementary Methods online). Cancer-derived human renal epithelial cells (A498), uroepithelial cells (J82) (both from American Type Culture Collection) and virus-immortalized UROtsa cells (provided by S. Garrett, University of North Dakota) were used as previously described42 (Supplementary Methods). Bacteria. For both mouse and in vitro experiments, we used E. coli CFT O73. This strain was isolated from an individual with pyelonephritis; it expresses type 1, P and S fimbriae and hemolysin, and induces pyelonephritis in experimentally challenged normal mice43. We grew bacteria overnight on cystine lactose electrolyte-deficient agar at 37 1C and in Luria-Bertani broth for 5 h, to reach logarithmic phase of growth, then washed them twice with PBS. The bacterial concentration was measured by spectrophotometry and confirmed by viable count on blood agar plates. Mouse model of ascendant pyelonephritis. Mouse experiments were approved by the Northern Stockholm Animal Ethics Committee. We performed ascendant urinary tract infections of mice as previously described44 (Supplementary Methods). We measured bacteria attached to urinary bladders after 1 h of infection. After 48 h of infection, we analyzed blood, kidneys and urinary bladders of mice. We carried out three independent mouse experiments with five to six mice in each group. For neutrophil depletion, we administered 100 ml RB6-C85 monoclonal antibody (R&D systems) intravenously 24 h before infection. Such treatment has been shown to induce neutropenia lasting 5 d45. Bacterial infection of renal cortical pieces and infection of cells. We infected 1 mm–thick kidney pieces (B100 mg), and we used epithelial cell lines when they formed a confluent layer. We carried out experiments in serum-free medium supplemented with gentamycin (40 mg/ml) in 24-well plates (Costar).

VOLUME 12

[

NUMBER 6

[

JUNE 2006

NATURE MEDICINE

ARTICLES We added 108 bacteria into each well and incubated plates in 37 1C, 5% CO2, 95% O2 and 80% humidity (Supplementary Methods).


Bacterial sensitivity assays and epithelial cell antimicrobial assay. Bacterial sensitivity assays and epithelial cell antimicrobial assay are described in Supplementary Methods. Analysis of cathelicidin protein and CAMP mRNA. We used immunohistochemistry, western blot and ELISA for analysis of cathelicidin peptide and real-time PCR to measure CAMP mRNA transcript levels (Supplementary Methods). Data analysis. Normally distributed data are presented as mean and standard deviation (whiskers), otherwise as median and range excluding outliers (whiskers) from at least three independent experiments. Alternatively, dot plots are presented, with mean for normally distributed data, or median for data, which did not meet the criteria for normal distribution. Outliers were not excluded from statistical analysis. Each set of experiments using normal human cells or tissue samples was repeated with cells or samples from at least three subjects. For the data with normal distribution, t-test for independent samples or analysis of variance (ANOVA) were used. Mann-Whitney U test, KruskalWallis ANOVA and Spearman rank-correlation test were used for the rest of the data. Differences with P o 0.05 were considered statistically significant. Note: Supplementary information is available on the Nature Medicine website. ACKNOWLEDGMENTS We thank Z. Fehervı´zyova´ and T. Baltesova´ for help collecting urine samples, M. Lindh for technical assistance and G. Kronvall for help with single-strain regression analysis. This work was supported by ALF Project Funding, The Swedish Society of Medicine, The Swedish Association of Kidney Patients, funds from the Karolinska Institute, Magn. Bergvalls Foundation, Capio Foundation, The Swedish Research Council (04X-2887, 06X-11217), The Swedish Foundation for International Cooperation in Research and Higher Education (STINT), and the Knut and Alice Wallenberg Foundation. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Published online at http://www.nature.com/naturemedicine/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/

1. Zasloff, M. Antimicrobial peptides of multicellular organisms. Nature 415, 389–395 (2002). 2. Ganz, T. Defensins: antimicrobial peptides of innate immunity. Nat. Rev. Immunol. 3, 710–720 (2003). 3. Zanetti, M. Cathelicidins, multifunctional peptides of the innate immunity. J. Leukoc. Biol. 75, 39–48 (2004). 4. Nizet, V. et al. Innate antimicrobial peptide protects the skin from invasive bacterial infection. Nature 414, 454–457 (2001). 5. Moser, C. et al. b-Defensin 1 contributes to pulmonary innate immunity in mice. Infect. Immun. 70, 3068–3072 (2002). 6. Valore, E.V. et al. Human b-defensin-1: an antimicrobial peptide of urogenital tissues. J. Clin. Invest. 101, 1633–1642 (1998). 7. Morrison, G., Kilanowski, F., Davidson, D. & Dorin, J. Characterization of the mouse beta defensin 1, Defb1, mutant mouse model. Infect. Immun. 70, 3053–3060 (2002). 8. Zucht, H.D. et al. Human b-defensin-1: A urinary peptide present in variant molecular forms and its putative functional implication. Eur. J. Med. Res. 3, 315–323 (1998). 9. Zaiou, M. & Gallo, R.L. Cathelicidins, essential gene-encoded mammalian antibiotics. J. Mol. Med. 80, 549–561 (2002). 10. Tollin, M. et al. Antimicrobial peptides in the first line defence of human colon mucosa. Peptides 24, 523–530 (2003). 11. Agerberth, B. et al. Antibacterial components in bronchoalveolar lavage fluid from healthy individuals and sarcoidosis patients. Am. J. Respir. Crit. Care Med. 160, 283–290 (1999). 12. Gudmundsson, G.H. et al. The human gene FALL39 and processing of the cathelin precursor to the antibacterial peptide LL-37 in granulocytes. Eur. J. Biochem. 238, 325–332 (1996). 13. Malm, J. et al. The human cationic antimicrobial protein (hCAP-18) is expressed in the epithelium of human epididymis, is present in seminal plasma at high concentrations, and is attached to spermatozoa. Infect. Immun. 68, 4297–4302 (2000).

NATURE MEDICINE

VOLUME 12

[

NUMBER 6

[

JUNE 2006

14. Gallo, R.L. et al. Identification of CRAMP, a cathelin-related antimicrobial peptide expressed in the embryonic and adult mouse. J. Biol. Chem. 272, 13088–13093 (1997). 15. Sorensen, O.E. et al. Human cathelicidin, hCAP-18, is processed to the antimicrobial peptide LL-37 by extracellular cleavage with proteinase 3. Blood 97, 3951–3959 (2001). 16. Zaiou, M., Nizet, V. & Gallo, R.L. Antimicrobial and protease inhibitory functions of the human cathelicidin (hCAP18/LL-37) prosequence. J. Invest. Dermatol. 120, 810–816 (2003). 17. Pestonjamasp, V.K., Huttner, K.H. & Gallo, R.L. Processing site and gene structure for the murine antimicrobial peptide CRAMP. Peptides 22, 1643–1650 (2001). 18. Brogden, K.A. Antimicrobial peptides: pore formers or metabolic inhibitors in bacteria? Nat. Rev. Microbiol. 3, 238–250 (2005). 19. Scott, M.G., Davidson, D.J., Gold, M.R., Bowdish, D. & Hancock, R.E. The human antimicrobial peptide LL-37 is a multifunctional modulator of innate immune responses. J. Immunol. 169, 3883–3891 (2002). 20. Yang, D., Biragyn, A., Kwak, L.W. & Oppenheim, J.J. Mammalian defensins in immunity: more than just microbicidal. Trends Immunol. 23, 291–296 (2002). 21. Bowdish, D.M., Davidson, D.J. & Hancock, R.E. A re-evaluation of the role of host defence peptides in mammalian immunity. Curr. Protein Pept. Sci. 6, 35–51 (2005). 22. Turner, J., Cho, Y., Dinh, N.N., Waring, A.J. & Lehrer, R.I. Activities of LL-37, a cathelin-associated antimicrobial peptide of human neutrophils. Antimicrob. Agents Chemother. 42, 2206–2214 (1998). 23. Faurschou, M. & Borregaard, N. Neutrophil granules and secretory vesicles in inflammation. Microbes Infect. 5, 1317–1327 (2003). 24. Hase, K., Eckmann, L., Leopard, J.D., Varki, N. & Kagnoff, M.F. Cell differentiation is a key determinant of cathelicidin LL-37/human cationic antimicrobial protein 18 expression by human colon epithelium. Infect. Immun. 70, 953–963 (2002). 25. Norden, C.W., Green, G.M. & Kass, E.H. Antibacterial mechanisms of the urinary bladder. J. Clin. Invest. 47, 2689–2700 (1968). 26. Justice, S.S. et al. Differentiation and developmental pathways of uropathogenic Escherichia coli in urinary tract pathogenesis. Proc. Natl. Acad. Sci. USA 101, 1333–1338 (2004). 27. Gordon, D.M. & Riley, M.A. A theoretical and experimental analysis of bacterial growth in the bladder. Mol. Microbiol. 6, 555–562 (1992). 28. Cox, C.E. & Hinman, F., Jr. Experiments with induced bacteriuria, vesical emptying and bacterial growth on the mechanism of bladder defense to infection. J. Urol. 86, 739–748 (1961). 29. Mulvey, M.A. et al. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282, 1494–1497 (1998). 30. Bates, J.M. et al. Tamm-Horsfall protein knockout mice are more prone to urinary tract infection: rapid communication. Kidney Int. 65, 791–797 (2004). 31. Fliedner, M., Mehls, O., Rauterberg, E.W. & Ritz, E. Urinary sIgA in children with urinary tract infection. J. Pediatr. 109, 416–421 (1986). 32. Abrink, M., Larsson, E., Gobl, A. & Hellman, L. Expression of lactoferrin in the kidney: implications for innate immunity and iron metabolism. Kidney Int. 57, 2004–2010 (2000). 33. Islam, D. et al. Downregulation of bactericidal peptides in enteric infections: a novel immune escape mechanism with bacterial DNA as a potential regulator. Nat. Med. 7, 180–185 (2001). 34. Bergman, P. et al. Neisseria gonorrhoeae downregulates expression of the human antimicrobial peptide LL-37. Cell. Microbiol. 7, 1009–1017 (2005). 35. Johansson, J., Gudmundsson, G.H., Rottenberg, M.E., Berndt, K.D. & Agerberth, B. Conformation-dependent antibacterial activity of the naturally occurring human peptide LL-37. J. Biol. Chem. 273, 3718–3724 (1998). 36. Sorensen, O., Arnljots, K., Cowland, J.B., Bainton, D.F. & Borregaard, N. The human antibacterial cathelicidin, hCAP-18, is synthesized in myelocytes and metamyelocytes and localized to specific granules in neutrophils. Blood 90, 2796–2803 (1997). 37. Dorschner, R.A. et al. The mammalian ionic environment dictates microbial susceptibility to antimicrobial defense peptides. FASEB J. 20, 35–42 (2006). 38. Yan, H. & Hancock, R.E. Synergistic interactions between mammalian antimicrobial defense peptides. Antimicrob. Agents Chemother. 45, 1558–1560 (2001). 39. Braff, M.H., Zaiou, M., Fierer, J., Nizet, V. & Gallo, R.L. Keratinocyte production of cathelicidin provides direct activity against bacterial skin pathogens. Infect. Immun. 73, 6771–6781 (2005). 40. Yoshio, H. et al. Antimicrobial polypeptides of human vernix caseosa and amniotic fluid: implications for newborn innate defense. Pediatr. Res. 53, 211–216 (2003). 41. Todd, J.H., McMartin, K.E. & Sens, D.A. Enzymatic isolation and serum-free culture of human renal cells. in Human cell culture protocols (ed. Jones, G.E.) 431–436 (Humana Press, Totowa, 1996). 42. Rossi, M.R. et al. The immortalized UROtsa cell line as a potential cell culture model of human urothelium. Environ. Health Perspect. 109, 801–808 (2001). 43. Khalil, A. et al. Cytokine gene expression during experimental Escherichia coli pyelonephritis in mice. J. Urol. 158, 1576–1580 (1997). 44. Chromek, M. et al. Matrix metalloproteinase-9 and tissue inhibitor of metalloproteinases-1 in acute pyelonephritis and renal scarring. Pediatr. Res. 53, 698–705 (2003). 45. Black, C.A. et al. Acute neutropenia decreases inflammation associated with murine vaginal candidiasis but has no effect on the course of infection. Infect. Immun. 66, 1273–1275 (1998).

641