Low back pain is currently one of the most common health problems, affecting >60% of adults and 30% of adolescents1,2. Back pain is associated with specific degenerative changes in the intervertebral discs3,4, and recent work has shown a particularly close association with Modic changes in or near the vertebral end plates5.
Strongly associated with lumbar disc herniation (LDH) and degeneration, Modic changes represent signal changes in the vertebral end plate and subchondral bone, and are visualized on T1 and T2-weighted magnetic resonance imaging (MRI) scans. First described in 1988, Modic changes are classified into 3 types6. Type-I changes show increased signal intensity on T2-weighted images and decreased signal intensity on T1-weighted images, reflecting fissures and microfractures of the trabeculae. Type-II changes demonstrate increased signal intensity on both T2 and T1-weighted images, representing damage and fatty degeneration. Type-III changes display decreased signal intensity on both T2 and T1-weighted images, which is considered to be sclerotic bone.
Type-II Modic changes predominate6-8, and the presence and size of Modic changes increase with age8. However, despite >2 decades of study, understanding of the pathogenesis of Modic changes remains limited. The prevalence of Modic changes has varied greatly among different studies, ranging from 0.5% to 62%, depending on how they were defined and identified, and on the age of the population9,10. Because of limited biopsy specimens, the pathological nature of Modic changes remains controversial.
It has been reported that 53% of patients with LDH have anaerobic microorganisms (P. acnes and Corynebacterium propinquum) within their intervertebral discs11. Albert et al. showed that Modic changes could be related to the presence of bacteria12, and that antibiotics (amoxicillin and clavulanate) are more effective at treating low back pain if Modic changes were present13. However, an animal model is required to demonstrate conclusively that anaerobic bacteria can cause Modic changes.
The objectives of the present study were (1) to demonstrate that infection by P. acnes can lead to the development of Modic changes in New Zealand White rabbits and (2) to compare the properties of infected and noninfected (control) subchondral bone tissues.
Materials and Methods
Twenty clean-grade (noninfected) male New Zealand White rabbits, with an age of approximately 3 months and a mean body weight of 2 kg (range, 1.8 to 2.4 kg), were used in this study. The animals were provided by the Shanghai Laboratory Animal Company (SLAC). Spinal deformity was excluded by radiographic examination. The study protocol was reviewed and approved by the institutional review board and ethics committee of the authors’ institution.
Twenty rabbits were randomly assigned to a P. acnes injection group or to a control group, each with 10 animals (Fig. 1). In the control animals, the subchondral bone superior to the L4-L5 and L5-L6 discs was injected percutaneously with either nothing (sham) or 1 mL of normal saline solution (vehicle), respectively. The subchondral bone superior to the L3-L4 discs was undisturbed (blank). In the P. acnes-injected animals, 1 mL of P. acnes (ATCC [American Type Culture Collection]-6919 provided by Guangzhou Type Culture Collection, 1.6 × 107 colony forming units [CFU]/mL in normal saline solution) was injected into the subchondral bone superior to the L4-L5 and L5-L6 discs. Although it is possible that the injection needle caused some damage to the end plate, the use of sham and vehicle controls enabled the effects of any such damage to be separated from the specific influences of the bacteria.
The rabbits were anesthetized with an intravenous injection of 15 mg/kg of pentobarbital sodium and were positioned prone on a platform. All processes were performed under sterile surgical conditions. Under the guidance of fluoroscopy (Luminos dRF; Siemens), percutaneous puncture was performed, from the animal’s right side, into subchondral bone superior to the L4-L5 and L5-L6 discs, using a 22-gauge needle with a so-called screw-in method to penetrate cortical bone (Fig. 2). The needle tip entered the subchondral bone at a distance of 1 to 2 mm from the adjacent end plate, at an angle of 45° from the coronal and cross-sectional planes (Fig. 2). When the needle tip arrived at the correct location, the needle stylet was withdrawn and P. acnes or normal saline solution was injected slowly into the marrow space under the cartilage end plate. After the procedure, the rabbits were housed individually with free access to food and water.
MRI and Evaluation
Lumbar MRI scans were performed before the operation and at 2 weeks and 1, 2, 3, and 6 months postoperatively, with a GE Signa CV/i scanner (1.5 T) using a knee-joint surface coil. Sagittal T2-weighted images were acquired using a repetition time/echo time of 2,500/120 ms, bandwidth of 31.25 Hz, and echo train length of 17, whereas T1-weighted images used 550/10 ms, 31.25 Hz, and echo train length of 2. A central sagittal image of the lumbar spine was selected as a locating image for the next T2-weighted FSE (fast spin echo) cross-sectional scans at 3,000/102 ms, 31.25 Hz, and echo train length of 17. The matrix size for each image was the same (matrix size, 384 × 256; field of view, 20 × 20 mm; slice thickness, 2 mm; and interslice gap, 0.2 mm).
All MRI scans were evaluated separately by an experienced radiologist and a spinal surgeon, who were blinded to the animal grouping. A consensus was reached by discussion when they disagreed in particular cases. Marrow signal changes adjacent to a target segment were recorded, together with disc degeneration or any other disorders revealed by the MRI scan. Modic changes were divided into 3 types6, and disc degeneration was classified initially from grade I to V, according to the system described by Pfirrmann et al.14.
Besides visual evaluation, we also performed quantitative measurements of the end-plate region signal intensity. ImageJ software (version 1.46; National Institutes of Health), was used to measure signal intensity of the superior subchondral bone region adjacent to L3-L4, L4-L5, and L5-L6 on T2 and T1-weighted central images, as previously reported15. The signal intensity of cerebrospinal fluid was used as contrast.
Blood samples from each rabbit were collected before and after the first injection and before the final MRI. After the final MRI examination, the rabbits were immediately killed with an overdose of pentobarbital sodium (40 mg/kg). The spine from L2 to L7 was excised, trimmed of soft tissue, and dissected into several vertebral body-disc-vertebral body specimens, each of which was bisected into 2 halves in the sagittal plane. One half was fixed for microcomputed tomography (micro-CT) and histological analysis. The other half was used to provide a tissue sample from the end plate for culture and was then frozen at −80°C for real-time polymerase chain reaction (PCR).
The fixed specimens were analyzed using a high-resolution (20-μm resolution, 0.48-mm field of view, and 1024 × 1024-pixel) micro-CT scanner (μCT 100; Scanco Medical). The scanning parameters were peak voltage of 70 kVp, current of 200 μA, and integration time of 300 ms. Trabecular bone parameters (bone volume per total volume [BV/TV], trabecular bone-specific surface [BS/BV], trabecular plate thickness [TbTh], trabecular number [TbN], and intertrabecular separation [TbSp]) and porosity were measured with the manufacturer’s program.
Histology and Blood Analysis
Specimens for histology were fixed with 4% paraformaldehyde at 4°C for 24 hours and then with 70% ethanol and decalcifying agent for 10 days. They were sequentially dehydrated, embedded in paraffin, and sectioned at 5 μm. Five sections of each specimen, chosen at random, were stained with hematoxylin and eosin (HE) for structural analysis and with safranin O-fast green for assessment of proteoglycans.
End plates were considered as degenerated if they had any of the following signs: end-plate structural defects, including cartilage erosions or fractures8, or granulation tissue, fibrous tissue, or fat cells replacing normal marrow tissue6.
The total white blood-cell count (WBC), C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) in blood were tested to investigate if general infection was present at 1 day and 6 months after injection.
Tissue Culture and Detection of Bacteria
Immediately after extraction, end-plate tissue samples were spread onto and embedded into Columbia blood agar plates and were incubated under anaerobic conditions for 14 days at 37°C. Chromosomal DNA was extracted from single presumptive bacterial colonies with ZymoBead Genomic DNA Kit (Zymo Research) amplified to a 600-bp region with P. acnes 16S rDNA primer (Table I), and tested with DNA agarose gel electrophoresis. A positive control (DNA from ATCC 6919) and negative control (sterile double distilled water) were included.
Real-Time PCR Analysis
The total RNA of subchondral end-plate bone and of cartilage were extracted separately from the specimens and purified with an RNeasy Mini Kit (Qiagen). These samples did not contain the anulus or nucleus. Reverse transcription was performed at 45°C for 50 minutes and 82°C for 5 minutes using 5× PrimeScript RT Master Mix (TaKaRa Bio). In end-plate bone specimens, the following genes were quantified as markers of end-plate inflammation: tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IL-4, interferon-γ (IFN-γ), and platelet-derived growth factor-β (PDGF-β). The housekeeping gene β-actin was used as a control. Our pretest showed that RNA from P. acnes does not lead to nonspecific amplification of the primers mentioned above (Fig. 3). Reactions were set up in triplicate in 96-well plates (20 μL per well) using 2 μL cDNA with 10 μL SsoFast EvaGreen Supermix (Bio-Rad Laboratories), 7 μL double-distilled water, and 10 μM gene-specific forward and reverse PCR primers (synthesized by Sangon Biotech) (Table I). PCR reactions were performed at 95°C for 10 minutes (activation), followed by 40 cycles of 95°C for 10 seconds, 60°C for 20 seconds, and 72°C for 20 seconds (amplification), and a final extension at 72°C for 1 minute, in an ABI Prism 7500 system (Applied Biosystems).
Data were expressed as the mean and standard deviation. Analysis of variance (ANOVA) was used to compare mean data among groups, with differences between each group being analyzed with the least significant difference (LSD). Proportions were compared using the chi-square test. Statistical analyses were performed using SPSS 16.0 (SPSS). A p value of <0.05 was considered significant.
At the third and sixth months, 7 of the 20 segments injected with P. acnes were identified as having type-II Modic changes, while no segments from blank, sham, or vehicle spinal levels showed such changes (p < 0.05). The T1-weighted MRI signal of all 4 groups decreased at 2 weeks postoperatively, which might reflect the influence of the operation. At 3 months, the signal intensity presented an increasing trend, but there were no significant differences between groups. At 6 months, signal intensity of T1-weighted images increased significantly in P. acnes animals (mean and standard deviation, 3.43 ± 0.41 [range, 2.42 to 4.44] compared with 2.43 ± 0.66 [range, 1.98 to 2.87] before the injection; p = 0.026). The T2-weighted signal also showed a higher intensity in P. acnes animals, at all times between 1 and 6 months, but differences were not significant (p = 0.115). At 6 months, discs adjacent to end plates injected with P. acnes showed no significant changes (Table II). No discs showed signs of degeneration (Fig. 4).
Typical micro-CT images (Fig. 5) showed normal trabecular bone. Trabecular parameters from blank, sham, and vehicle groups showed no significant differences (Table III). End-plate regions injected with P. acnes also showed no significant differences in micro-CT parameters compared with other groups.
Fluid from 11 of 20 end-plate regions injected with P. acnes showed bacterial colonies. Among them, 9 were identified as P. acnes with a specific primer (Fig. 6), and the other 2 regions grew staphylococci. No bacterial colonies grew in the same culture environment from fluid of the other groups; the differences between P. acnes and control groups were significant (p = 0.013). A detailed comparison of MRI and PCR results is shown in Table IV.
Three of the 20 specimens in the P. acnes group presented an end-plate fissure (Fig. 7). All 3 were visually confirmed to be type-II Modic changes. Compared with intervertebral discs in the blank group, discs in the sham and vehicle groups seemed normal. The cell density of the anulus, nucleus, cartilage, and osseous end plates also presented no significant changes in specimens injected with normal saline solution or P. acnes, and histological analysis showed no indication of inflammatory cells (Fig. 7).
The WBC, CRP, and ESR showed no significant differences among the groups or within the groups before or after injection or at 6 months (results not presented).
Real-Time PCR Analysis
The results of the real-time PCR are presented in Figure 3. Compared with the vehicle groups, the expression of IL-4 in the end-plate region injected with P. acnes was slightly but not significantly increased. On the other hand, the expressions of TNF-α, IL-1β, and IFN-γ were significantly upregulated following the injection of P. acnes (p = 0.001, 0.013, and p < 0.001, respectively). The expression of PDGF-β was similar between the vehicle and P. acnes groups (p = 0.526), with both increased significantly compared with the blank and sham groups (p = 0.003 and 0.001, respectively).
P. acnes survived in 9 of 20 end plates at 6 months, creating a mild inflammatory-like response based on the detection of inflammatory cytokines and increased signal intensity in T1-weighted images in a manner that resembled Modic changes.
There is no consensus concerning the pathology of Modic changes and their relationship with disc degeneration and disc infection16,17. Reported pathological findings in Modic changes include inflammation, fatty changes in bone marrow6, defects (fractures or erosions) in the adjacent end plate, and neoinnervation18. Adams et al. suggested that microfractures and fissures within the end plate can arise from biomechanical stress16 and can initiate Modic changes16,17 by allowing various substances to pass between the vertebral bone and intervertebral disc15,19,20. A clinical study by Albert et al. involving 61 patients with LDH showed that extruded nucleus material from 46% of the patients contained bacteria, with P. acnes being the main microorganism (86%)12. Their further research, a double-blind randomized clinical trial with 1 year of follow-up, showed that amoxicillin-clavulanate treatment reduced chronic low back pain in patients with Modic type-I changes13. Agarwal et al. successfully cultured P. acnes from materials harvested by microdiscectomy for LDH (7 of 52 patients)21. Zhou et al. found that P. acnes is more likely to be present in herniated discs with anular tears, suggesting that P. acnes can enter the disc nucleus via defects in adjacent tissues22.
Our culture results confirm that P. acnes can survive and multiply in tissues adjacent to the vertebral end plate. On the other hand, we did not find P. acnes in the blank, sham, or vehicle groups, which confirmed that P. acnes was not present in normal discs. The MRI results show increased signal intensity in T1 and T2-weighted images of the end-plate region, suggesting an inflammatory-like reaction characteristic of Modic type-II changes. It is conceivable that the injection caused focal tissue damage, which allowed increased swelling and tissue hydration, but this was not observed in the histological analysis. Three samples from the P. acnes group showed irregular end-plate lesions, which could be the result of P. acnes infection. Their rough edges suggest that they were not caused by needle puncture, and the process was performed under fluoroscopy so that the needle would not penetrate the disc. Histological analysis also showed no increase in disc degeneration during the 6-month study, suggesting that the inflammation-like changes were initiated directly by the bacteria. Although Modic changes are considered to be associated with disc degeneration, not all Modic changes are found together with degeneration in the clinic. The real-time PCR results indicated that this focal inflammation is mediated by cytokines such as IL-1β, IFN-γ, and TNF-α, which presumably arise from host cells in the vertebral bone. The fact that these cytokines were not upregulated in the sham and/or vehicle groups suggests that they were induced by toxins from the bacteria, rather than induced by the cytokines upregulated by the damaged end plate (PDGF-β and IL-4). Similar cytokines are activated in herniated discs23-25. Because disc degeneration and herniation always coexist with Modic changes6,26, the inflammatory-like responses associated with P. acnes, Modic changes, and LDH may share a common mechanism.
It is possible that the early effects of P. acnes are so weak that they are not fully reflected by MRI, because a previous study has shown that Modic change-like pathologies can be missed by MRI18. It should be noted that direct inoculation does not fully reflect clinical conditions of infection, and there was no subsequent invasion of inflammatory cells. It is also possible that the inflammation sites were missed in our histological analysis. Needle puncture and infection together led to Modic changes in the present experiment, but it cannot be directly proven that P. acnes alone can cause Modic changes in humans. It is possible that the infection only accelerates the changes initiated by focal damage arising from injection. Thus, P. acnes alone may lead to only a weak inflammatory-like reaction, whereas severe Modic changes may depend also on the altered biomechanical and biochemical environment associated with end-plate defects, inflammatory cell invasion, and accompanying disc degeneration.
As far as we know, this study is the first attempt to establish an in vivo animal model of Modic changes. The use of multiple techniques, including percutaneous puncture, histology, PCR, MRI, and micro-CT, to evaluate the effects of P. acnes is a major strength. One limitation of this study is that we used only the typical strain of P. acnes (ATCC 6919). Second, we could not completely exclude the possibility that the P. acnes that we confirmed with PCR was derived from exogenous contamination, although the operation was performed in strictly sterile conditions. Third, a 6-month quadrupedal rabbit model may not completely represent biomechanical influences found in humans. The etiology of different types of Modic changes may be sequential7, whereas our results concern only Modic type-II changes. There is evidence of significant osseous changes in Modic type-III changes, for example, in the research by Perilli et al.27. Additional studies with primates over a long time period and with more intervening factors would enable us to increase our understanding of the occurrence and development of Modic changes.
While revising this manuscript, we learned of a simultaneous experiment on rats with similar objectives and whose results support the conclusions of the present paper28.
In summary, P. acnes infection can be a potential cause of Modic changes, for it can reside within the end-plate region and lead to a mild inflammatory-like response.
NOTE: The authors thank Professor Michael Adams and Professor Trish Dolan for their assistance with the manuscript.
Investigation performed at the Department of Orthopaedics, Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, People’s Republic of China
Disclosure: This study was partially supported by the National Natural Science Foundation of China (No. 31270997, No. 81171739, and No. 81672208). The Disclosure of Potential Conflicts of Interest forms are provided with the online version of this article (http://links.lww.com/JBJS/A143).
1. Waddell G, Burton AK. Occupational health guidelines for the management of low back pain at work: evidence review. Occup Med (Lond). 2001 ;51(2):124–35.
2. Shan Z, Deng G, Li J, Li Y, Zhang Y, Zhao Q. Correlational analysis of neck/shoulder pain and low back pain with the use of digital products, physical activity and psychological status among adolescents in Shanghai. PLoS One. 2013;8(10):e78109. Epub 2013 Oct 11.
3. Cheung KM, Karppinen J, Chan D, Ho DW, Song YQ, Sham P, Cheah KS, Leong JC, Luk KD. Prevalence and pattern of lumbar magnetic resonance imaging changes in a population study of one thousand forty-three individuals. Spine (Phila Pa 1976). 2009 ;34(9):934–40.
4. de Schepper EI, Damen J, van Meurs JB, Ginai AZ, Popham M, Hofman A, Koes BW, Bierma-Zeinstra SM. The association between lumbar disc degeneration and low back pain: the influence of age, gender, and individual radiographic features. Spine (Phila Pa 1976). 2010 ;35(5):531–6.
5. Jensen TS, Karppinen J, Sorensen JS, Niinimäki J, Leboeuf-Yde C. Vertebral endplate signal changes (Modic change): a systematic literature review of prevalence and association with non-specific low back pain. Eur Spine J. 2008 ;17(11):1407–22. Epub 2008 Sep 12.
6. Modic MT, Steinberg PM, Ross JS, Masaryk TJ, Carter JR. Degenerative disk disease: assessment of changes in vertebral body marrow with MR imaging. Radiology. 1988 ;166(1 Pt 1):193–9.
7. Zhang YH, Zhao CQ, Jiang LS, Chen XD, Dai LY. Modic changes: a systematic review of the literature. Eur Spine J. 2008 ;17(10):1289–99. Epub 2008 Aug 27.
8. Wang Y, Videman T, Battié MC. Modic changes: prevalence, distribution patterns, and association with age in white men. Spine J. 2012 ;12(5):411–6. Epub 2012 Apr 18.
9. Kjaer P, Leboeuf-Yde C, Sorensen JS, Bendix T. An epidemiologic study of MRI and low back pain in 13-year-old children. Spine (Phila Pa 1976). 2005 ;30(7):798–806.
10. Kleinstück F, Dvorak J, Mannion AF. Are “structural abnormalities” on magnetic resonance imaging a contraindication to the successful conservative treatment of chronic nonspecific low back pain? Spine (Phila Pa 1976). 2006 ;31(19):2250–7.
11. Stirling A, Worthington T, Rafiq M, Lambert PA, Elliott TS. Association between sciatica and Propionibacterium acnes. Lancet. 2001 ;357(9273):2024–5.
12. Albert HB, Lambert P, Rollason J, Sorensen JS, Worthington T, Pedersen MB, Nørgaard HS, Vernallis A, Busch F, Manniche C, Elliott T. Does nuclear tissue infected with bacteria following disc herniations lead to Modic changes in the adjacent vertebrae? Eur Spine J. 2013 ;22(4):690–6. Epub 2013 Feb 10.
13. Albert HB, Sorensen JS, Christensen BS, Manniche C. Antibiotic treatment in patients with chronic low back pain and vertebral bone edema (Modic type 1 changes): a double-blind randomized clinical controlled trial of efficacy. Eur Spine J. 2013 ;22(4):697–707. Epub 2013 Feb 13.
14. Pfirrmann CW, Metzdorf A, Zanetti M, Hodler J, Boos N. Magnetic resonance classification of lumbar intervertebral disc degeneration. Spine (Phila Pa 1976). 2001 ;26(17):1873–8.
15. Shan Z, Fan S, Xie Q, Suyou L, Liu J, Wang C, Zhao F. Spontaneous resorption of lumbar disc herniation is less likely when Modic changes are present. Spine (Phila Pa 1976). 2014 ;39(9):736–44.
16. Adams MA, Freeman BJ, Morrison HP, Nelson IW, Dolan P. Mechanical initiation of intervertebral disc degeneration. Spine (Phila Pa 1976). 2000 ;25(13):1625–36.
17. Adams MA, McNally DS, Dolan P. ‘Stress’ distributions inside intervertebral discs. The effects of age and degeneration. J Bone Joint Surg Br. 1996 ;78(6):965–72.
18. Fields AJ, Liebenberg EC, Lotz JC. Innervation of pathologies in the lumbar vertebral end plate and intervertebral disc. Spine J. 2014 ;14(3):513–21. Epub 2013 Oct 18.
19. Braithwaite I, White J, Saifuddin A, Renton P, Taylor BA. Vertebral end-plate (Modic) changes on lumbar spine MRI: correlation with pain reproduction at lumbar discography. Eur Spine J. 1998;7(5):363–8.
20. Lama P, Zehra U, Balkovec C, Claireaux HA, Flower L, Harding IJ, Dolan P, Adams MA. Significance of cartilage endplate within herniated disc tissue. Eur Spine J. 2014 ;23(9):1869–77. Epub 2014 Jun 20.
21. Agarwal V, Golish SR, Alamin TF. Bacteriologic culture of excised intervertebral disc from immunocompetent patients undergoing single level primary lumbar microdiscectomy. J Spinal Disord Tech. 2011 ;24(6):397–400.
22. Zhou Z, Chen Z, Zheng Y, Cao P, Liang Y, Zhang X, Wu W, Xiao J, Qiu S. Relationship between annular tear and presence of Propionibacterium acnes in lumbar intervertebral disc. Eur Spine J. 2015 ;24(11):2496–502. Epub 2015 Aug 19.
23. Ahn SH, Cho YW, Ahn MW, Jang SH, Sohn YK, Kim HS. mRNA expression of cytokines and chemokines in herniated lumbar intervertebral discs. Spine (Phila Pa 1976). 2002 ;27(9):911–7.
24. Burke JG, Watson RW, McCormack D, Dowling FE, Walsh MG, Fitzpatrick JM. Intervertebral discs which cause low back pain secrete high levels of proinflammatory mediators. J Bone Joint Surg Br. 2002 ;84(2):196–201.
25. Lee JM, Song JY, Baek M, Jung HY, Kang H, Han IB, Kwon YD, Shin DE. Interleukin-1β induces angiogenesis and innervation in human intervertebral disc degeneration. J Orthop Res. 2011 ;29(2):265–9.
26. de Roos A, Kressel H, Spritzer C, Dalinka M. MR imaging of marrow changes adjacent to end plates in degenerative lumbar disk disease. AJR Am J Roentgenol. 1987 ;149(3):531–4.
27. Perilli E, Parkinson IH, Truong LH, Chong KC, Fazzalari NL, Osti OL. Modic (endplate) changes in the lumbar spine: bone micro-architecture and remodelling. Eur Spine J. 2015 ;24(9):1926–34. Epub 2014 Jul 26.
28. Dudli S, Liebenberg E, Magnitsky S, Miller S, Demir-Deviren S, Lotz JC. Propionibacterium acnes infected intervertebral discs cause vertebral bone marrow lesions consistent with Modic changes. J Orthop Res. 2016 ;34(8):1447–55. Epub 2016 Aug 3.
Supplemental Digital Content
Copyright 2017 by The Journal of Bone and Joint Surgery, Incorporated