Inflammatory Response to Escherichia coli Urinary Tract Infection in the Neurogenic Bladder of the Spinal Cord Injured Host Rajeev Chaudhry,* Ramiro J. Madden-Fuentes,* Tara K. Ortiz, Zarine Balsara, Yuping Tang, Unwanaobong Nseyo, John S. Wiener,† Sherry S. Ross and Patrick C. Seed‡ From the Division of Urologic Surgery, Department of Surgery (RC, RJM-F, TKO, ZB, UN, JSW, SSR, PCS), Department of Pediatrics (YT, JSW, SSR, PCS), Department of Molecular Genetics and Microbiology and Center for Microbial Pathogenesis (PCS), Duke University Medical Center (PCS), Durham, North Carolina
Abbreviations and Acronyms hpi ¼ hours after infection IL ¼ interleukin NGB ¼ neurogenic bladder PCR ¼ polymerase chain reaction qPCR ¼ quantitative PCR SCI ¼ spinal cord injury UTI ¼ urinary tract infection Accepted for publication December 9, 2013. Study received Duke University institutional animal care and use committee approval. Supported by the Paralyzed Veterans Association. * Equal study contribution. † Financial interest and/or other relationship with GlaxoSmithKline. ‡ Correspondence: Duke University School of Medicine, Durham, North Carolina 27710 (telephone: 919-684-9590; FAX: 919-681-2089).
Purpose: Urinary tract infections cause significant morbidity in patients with spinal cord injury. An in vivo spinal cord injured rat model of experimental Escherichia coli urinary tract infection mimics human disease with enhanced susceptibility to urinary tract infection compared to controls. We hypothesized that a dysregulated inflammatory response contributes to enhanced susceptibility to urinary tract infection. Materials and Methods: Spinal cord injured and sham injured rats were inoculated transurethrally with E. coli. Transcript levels of 84 inflammatory pathway genes were measured in bladder tissue of each group before infection, 24 hours after infection and after 5 days of antibiotic therapy. Results: Before infection quantitative polymerase chain reaction array revealed greater than twofold up-regulation in the proinflammatory factor transcripts slc11a1, ccl4 and il1b, and down-regulation of the antimicrobial peptides lcn2 and mpo in spinal cord injured vs control bladders. At 24 hours after infection spinal cord injured bladders showed an attenuated innate immune response with decreased expression of il6, slc11a1, il1b and lcn2, and decreased il10 and slpi expression compared to controls. Despite clearance of bacteriuria with antibiotics spinal cord injured rats had delayed induction of il6 transcription and a delayed anti-inflammatory response with decreased il10 and slpi transcript levels relative to controls. Conclusions: Spinal cord injured bladders fail to mount a characteristic inflammatory response to E. coli infection and cannot suppress inflammation after infection is eliminated. This may lead to increased susceptibility to urinary tract infection and persistent chronic inflammation through neural mediated pathways, which to our knowledge remain to be defined. Key Words: urinary bladder, neurogenic; urinary tract infections; spinal cord injuries; cytokines; Escherichia coli
NEUROGENIC bladder results from abnormal neural communication between bladder and nervous system, resulting in abnormal bladder storage or emptying. More than an estimated 400,000 children and adults in the United States have NGB secondary
1454
j
www.jurology.com
to SCI, spina bifida, cerebral palsy, multiple sclerosis or Parkinson disease.1e5 NGB is associated with increased morbidity and mortality due in part to a high risk of recurrent UTI with an annual incidence as high of as 20% to 25%.6e8 Recurrent UTIs
0022-5347/14/1915-1454/0 THE JOURNAL OF UROLOGY® © 2014 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2013.12.013 Vol. 191, 1454-1461, May 2014 Printed in U.S.A.
INFLAMMATORY RESPONSE TO URINARY TRACT INFECTION IN SPINAL CORD INJURED HOST
are a leading cause of rehospitalization in patients with NGB and they are associated with increased morbidity and mortality as the result of end stage renal disease and urosepsis.9 However, the underlying mechanisms accountable for an enhanced susceptibility to infection remain unclear. Increased post-void residual volume and urinary stasis were postulated to predispose to bacterial overgrowth and infection. However, human studies do not universally support this proposition. Correlations among post-void residual urine volume, urinary retention and UTI risk are controversial.10e12 A major challenge of defining the mechanisms underlying susceptibility to UTI in the SCI host has been the lack of an appropriate in vivo model. In a study of UTI in a SCI rat model using a clinical isolate of uropathogenic Escherichia coli we could not correlate the degree of post-void residual urine volume with susceptibility to UTI.13 While these data do not completely eliminate voiding dysfunction in SCI cases as a contributor to UTI susceptibility, they suggest that other factors may have a substantial role in modifying the infection risk. SCI produces systemic and organ specific changes, including the induction of bladder remodeling and activation of bladder inflammation. Studies by several groups demonstrated remodeling of SCI bladders, as evidenced by increased levels of tropoelastase and lysyl oxidase along with enhanced production of transforming growth factor-b1 and insulin-like growth factor-1.14,15 Also, various proinflammatory transcripts are up-regulated in bladders after SCI, including CD74, S100A9 and THY-1.15 We hypothesized that a dysregulated inflammatory response to bladder infection may be a mechanism contributing to enhanced susceptibility to UTI in SCI bladders. Using our rat SCI model we found altered inflammatory responses before, during and in the resolution of E. coli UTI.
METHODS Methods and T10 Spinal Cord Transection All animal use and procedures were approved by the Duke University institutional animal care and use committee. Female Sprague DawleyÒ rats underwent complete T10 spinal cord transection or sham surgery consisting of anesthesia, surgical incision over T10 and closure. During a 14-day postoperative recovery period SCI rats underwent a manual Crede maneuver twice daily. SCI rats and controls were injected twice daily with 10 mg/kg enrofloxacin for prophylaxis against UTI during the recovery period. Each group had free access to food.
UTI Induction and Antibiotic Treatment Antibiotic prophylaxis was stopped 72 hours before infection. SCI and control rats were infected with approximately 5 103 and 5 106 cfu, respectively, of the
1455
human cystitis E. coli isolate UTI89. This was estimated to be about 1.5 log greater than the 50% infectious dose, as previously established at our laboratory (Balsara and Seed, unpublished data). A subset of SCI rats was also infected with approximately 5 106 cfu to ensure that any detectable difference in inflammatory response and cfu were not attributable to the disparity in inocula. With the rat under inhaled isoflurane anesthesia the perineum was prepared with iodine solution and 70% ethanol. A lubricated 22 gauge No. 381123 BD AngiocathÔ angiocatheter (catalogue No. 381123) was introduced in the urethra. The bladder was fully emptied of urine and the bacterial inoculum was instilled directly in the bladder. Urine was collected before infection and each 24 hours thereafter by a manual Cred e maneuver. To determine cfu urine samples were plated on Luria-Bertani agar with kanamycin (50 mg/ml) and on MacConkey agar plates. A group of rats was treated with enrofloxacin 48 hpi for 5 days (10 mg/kg intramuscular injection twice daily). Urine was collected daily during therapy and plated to enumerate cfu, as described. Unless otherwise indicated experiments were performed in biological replicates on separate days.
Organ Harvest and Homogenization Rats were sacrificed by CO2 inhalation. Bladders were procured in sterile fashion. Half of each tissue was preserved in 10% formalin for histopathology, a quarter was preserved in RNAlaterÒ for RNA extraction and a quarter was placed in 4 ml phosphate buffered saline with 0.02% Triton X-100Ô for homogenization. Tissues were homogenized with a handheld BioGen PRO200Ò tissue homogenizer. To enumerate cfu homogenates were plated in serial dilutions on Luria-Bertani agar plates containing 50 mg/ml kanamycin or on MacConkey agar plates.
Polymerase Chain Reaction Array. Transcript levels of 84 innate immunity targets were measured using RT2 Profiler PCR Antibacterial Response Arrays (No. PARN-148A, QiagenÒ). Bladder tissue RNA was extracted using the RNeasyÒ Mini Kit. RNA quality, concentration and purity were analyzed using ultraviolet spectrophotometry and gel electrophoresis. First strand cDNA synthesis was performed using random hexamers and reverse transcriptase (Qiagen and InvitrogenÔ). Pooled cDNA from each group of 5 rats was used for input into each array run. Arrays were run on a 7900HT Fast Real-Time PCR System (Applied BiosystemsÒ) and analyzed by the 2eDDCT method. The supplementary table (http://jurology.com/) lists genes in the array. Quantitative. The supplementary Appendix (http:// jurology.com/) lists primers (IDTÒ). qPCR reactions were performed under 95C for 1 minute and then 40 cycles at 95C for 10 seconds, at 60C for 30 seconds and at 72C for 15 seconds. Fluorescence was measured at a 79C step for 2 seconds after the extension step.16 Melting curve analysis was performed immediately after the run from 58C to 90C. Data were analyzed by relative quantification using the 2eDDCT method.
Neutrophil Enumeration Formalin fixed bladder tissue was embedded in paraffin, sectioned at 5 mm and mounted on glass slides for
1456
INFLAMMATORY RESPONSE TO URINARY TRACT INFECTION IN SPINAL CORD INJURED HOST
hematoxylin and eosin staining. Each slide contained 9 to 12 bladder sections from a single rat. Tissue sections were examined under light microscopy at 400 power and 20 high power fields of each slide were photographed. Each image included epithelium and submucosa. Images were randomly numbered and separated from the key coding each image to the section of origin. Neutrophils were counted by a blinded recorder. After 2 independent counts per image replicate counts were averaged and compiled to obtain the mean SD count per section.
Statistical Analysis Statistical analysis was done with PrismÒ with significance considered at p
Abbreviations and Acronyms hpi ¼ hours after infection IL ¼ interleukin NGB ¼ neurogenic bladder PCR ¼ polymerase chain reaction qPCR ¼ quantitative PCR SCI ¼ spinal cord injury UTI ¼ urinary tract infection Accepted for publication December 9, 2013. Study received Duke University institutional animal care and use committee approval. Supported by the Paralyzed Veterans Association. * Equal study contribution. † Financial interest and/or other relationship with GlaxoSmithKline. ‡ Correspondence: Duke University School of Medicine, Durham, North Carolina 27710 (telephone: 919-684-9590; FAX: 919-681-2089).
Purpose: Urinary tract infections cause significant morbidity in patients with spinal cord injury. An in vivo spinal cord injured rat model of experimental Escherichia coli urinary tract infection mimics human disease with enhanced susceptibility to urinary tract infection compared to controls. We hypothesized that a dysregulated inflammatory response contributes to enhanced susceptibility to urinary tract infection. Materials and Methods: Spinal cord injured and sham injured rats were inoculated transurethrally with E. coli. Transcript levels of 84 inflammatory pathway genes were measured in bladder tissue of each group before infection, 24 hours after infection and after 5 days of antibiotic therapy. Results: Before infection quantitative polymerase chain reaction array revealed greater than twofold up-regulation in the proinflammatory factor transcripts slc11a1, ccl4 and il1b, and down-regulation of the antimicrobial peptides lcn2 and mpo in spinal cord injured vs control bladders. At 24 hours after infection spinal cord injured bladders showed an attenuated innate immune response with decreased expression of il6, slc11a1, il1b and lcn2, and decreased il10 and slpi expression compared to controls. Despite clearance of bacteriuria with antibiotics spinal cord injured rats had delayed induction of il6 transcription and a delayed anti-inflammatory response with decreased il10 and slpi transcript levels relative to controls. Conclusions: Spinal cord injured bladders fail to mount a characteristic inflammatory response to E. coli infection and cannot suppress inflammation after infection is eliminated. This may lead to increased susceptibility to urinary tract infection and persistent chronic inflammation through neural mediated pathways, which to our knowledge remain to be defined. Key Words: urinary bladder, neurogenic; urinary tract infections; spinal cord injuries; cytokines; Escherichia coli
NEUROGENIC bladder results from abnormal neural communication between bladder and nervous system, resulting in abnormal bladder storage or emptying. More than an estimated 400,000 children and adults in the United States have NGB secondary
1454
j
www.jurology.com
to SCI, spina bifida, cerebral palsy, multiple sclerosis or Parkinson disease.1e5 NGB is associated with increased morbidity and mortality due in part to a high risk of recurrent UTI with an annual incidence as high of as 20% to 25%.6e8 Recurrent UTIs
0022-5347/14/1915-1454/0 THE JOURNAL OF UROLOGY® © 2014 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RESEARCH, INC.
http://dx.doi.org/10.1016/j.juro.2013.12.013 Vol. 191, 1454-1461, May 2014 Printed in U.S.A.
INFLAMMATORY RESPONSE TO URINARY TRACT INFECTION IN SPINAL CORD INJURED HOST
are a leading cause of rehospitalization in patients with NGB and they are associated with increased morbidity and mortality as the result of end stage renal disease and urosepsis.9 However, the underlying mechanisms accountable for an enhanced susceptibility to infection remain unclear. Increased post-void residual volume and urinary stasis were postulated to predispose to bacterial overgrowth and infection. However, human studies do not universally support this proposition. Correlations among post-void residual urine volume, urinary retention and UTI risk are controversial.10e12 A major challenge of defining the mechanisms underlying susceptibility to UTI in the SCI host has been the lack of an appropriate in vivo model. In a study of UTI in a SCI rat model using a clinical isolate of uropathogenic Escherichia coli we could not correlate the degree of post-void residual urine volume with susceptibility to UTI.13 While these data do not completely eliminate voiding dysfunction in SCI cases as a contributor to UTI susceptibility, they suggest that other factors may have a substantial role in modifying the infection risk. SCI produces systemic and organ specific changes, including the induction of bladder remodeling and activation of bladder inflammation. Studies by several groups demonstrated remodeling of SCI bladders, as evidenced by increased levels of tropoelastase and lysyl oxidase along with enhanced production of transforming growth factor-b1 and insulin-like growth factor-1.14,15 Also, various proinflammatory transcripts are up-regulated in bladders after SCI, including CD74, S100A9 and THY-1.15 We hypothesized that a dysregulated inflammatory response to bladder infection may be a mechanism contributing to enhanced susceptibility to UTI in SCI bladders. Using our rat SCI model we found altered inflammatory responses before, during and in the resolution of E. coli UTI.
METHODS Methods and T10 Spinal Cord Transection All animal use and procedures were approved by the Duke University institutional animal care and use committee. Female Sprague DawleyÒ rats underwent complete T10 spinal cord transection or sham surgery consisting of anesthesia, surgical incision over T10 and closure. During a 14-day postoperative recovery period SCI rats underwent a manual Crede maneuver twice daily. SCI rats and controls were injected twice daily with 10 mg/kg enrofloxacin for prophylaxis against UTI during the recovery period. Each group had free access to food.
UTI Induction and Antibiotic Treatment Antibiotic prophylaxis was stopped 72 hours before infection. SCI and control rats were infected with approximately 5 103 and 5 106 cfu, respectively, of the
1455
human cystitis E. coli isolate UTI89. This was estimated to be about 1.5 log greater than the 50% infectious dose, as previously established at our laboratory (Balsara and Seed, unpublished data). A subset of SCI rats was also infected with approximately 5 106 cfu to ensure that any detectable difference in inflammatory response and cfu were not attributable to the disparity in inocula. With the rat under inhaled isoflurane anesthesia the perineum was prepared with iodine solution and 70% ethanol. A lubricated 22 gauge No. 381123 BD AngiocathÔ angiocatheter (catalogue No. 381123) was introduced in the urethra. The bladder was fully emptied of urine and the bacterial inoculum was instilled directly in the bladder. Urine was collected before infection and each 24 hours thereafter by a manual Cred e maneuver. To determine cfu urine samples were plated on Luria-Bertani agar with kanamycin (50 mg/ml) and on MacConkey agar plates. A group of rats was treated with enrofloxacin 48 hpi for 5 days (10 mg/kg intramuscular injection twice daily). Urine was collected daily during therapy and plated to enumerate cfu, as described. Unless otherwise indicated experiments were performed in biological replicates on separate days.
Organ Harvest and Homogenization Rats were sacrificed by CO2 inhalation. Bladders were procured in sterile fashion. Half of each tissue was preserved in 10% formalin for histopathology, a quarter was preserved in RNAlaterÒ for RNA extraction and a quarter was placed in 4 ml phosphate buffered saline with 0.02% Triton X-100Ô for homogenization. Tissues were homogenized with a handheld BioGen PRO200Ò tissue homogenizer. To enumerate cfu homogenates were plated in serial dilutions on Luria-Bertani agar plates containing 50 mg/ml kanamycin or on MacConkey agar plates.
Polymerase Chain Reaction Array. Transcript levels of 84 innate immunity targets were measured using RT2 Profiler PCR Antibacterial Response Arrays (No. PARN-148A, QiagenÒ). Bladder tissue RNA was extracted using the RNeasyÒ Mini Kit. RNA quality, concentration and purity were analyzed using ultraviolet spectrophotometry and gel electrophoresis. First strand cDNA synthesis was performed using random hexamers and reverse transcriptase (Qiagen and InvitrogenÔ). Pooled cDNA from each group of 5 rats was used for input into each array run. Arrays were run on a 7900HT Fast Real-Time PCR System (Applied BiosystemsÒ) and analyzed by the 2eDDCT method. The supplementary table (http://jurology.com/) lists genes in the array. Quantitative. The supplementary Appendix (http:// jurology.com/) lists primers (IDTÒ). qPCR reactions were performed under 95C for 1 minute and then 40 cycles at 95C for 10 seconds, at 60C for 30 seconds and at 72C for 15 seconds. Fluorescence was measured at a 79C step for 2 seconds after the extension step.16 Melting curve analysis was performed immediately after the run from 58C to 90C. Data were analyzed by relative quantification using the 2eDDCT method.
Neutrophil Enumeration Formalin fixed bladder tissue was embedded in paraffin, sectioned at 5 mm and mounted on glass slides for
1456
INFLAMMATORY RESPONSE TO URINARY TRACT INFECTION IN SPINAL CORD INJURED HOST
hematoxylin and eosin staining. Each slide contained 9 to 12 bladder sections from a single rat. Tissue sections were examined under light microscopy at 400 power and 20 high power fields of each slide were photographed. Each image included epithelium and submucosa. Images were randomly numbered and separated from the key coding each image to the section of origin. Neutrophils were counted by a blinded recorder. After 2 independent counts per image replicate counts were averaged and compiled to obtain the mean SD count per section.
Statistical Analysis Statistical analysis was done with PrismÒ with significance considered at p
