RAPID COMMUNICATION 488J o u r n a l o fJ o u r n a l o f

CellularPhysiologyCellularPhysiology

Vertebral Body Innervation:

Implications for Pain

MICHELANGELO BUONOCORE,1,2* ANNA MARIA ALOISI,2,3 MASSIMO BARBIERI,2

ANNA MARIA GATTI,1 AND CESARE BONEZZI2

1Unit of Clinical Neurophysiology & Neurodiagnostic Skin Biopsy, Pavia, Italy2Unit of Pain Medicine IRCCS Fondazione Salvatore Maugeri Pavia, Pavia, Italy3Department of Physiology, University of Siena, Siena, Italy

Vertebral fractures often cause intractable pain. To define the involvement of vertebral body innervation in pain, we collected specimensfrom male and female patients during percutaneous kyphoplasty, a procedure used for reconstruction of the vertebral body. Specimenswere taken from 31 patients (9 men and 22 women) suffering high-intensity pain before surgery. In total, 1,876 histological preparationswere obtained and analysed. Immunohistochemical techniques were used to locate the nerves in the specimens. The nerve fibres werelabelled by indirect immunofluorescence with the primary antibody directed against Protein Gene Product 9.5 (PGP 9.5), a pan-neuronalmarker; another primary antibody directed against type IV collagen (Col IV) was used to identify vessels and to determine theirrelationship with vertebral nerve fibres. The mean percentage of samples in which it was possible to identify nerve fibres was 35% in menand 29% in women. The percentages varied depending on the spinal level considered and the sex of the subject, nerve fibres being mostlypresent around vessels (95%). In conclusion, there is scarce innervation of the vertebral bodies, with a clear prevalence of fibres locatedaround vessels. It seems unlikely that this pattern of vertebral body innervation is involved in vertebral pain or in pain relief followingkyphoplasty.

J. Cell. Physiol. 222: 488–491, 2010. � 2009 Wiley-Liss, Inc.

The authors declare the absence of conflicts of interest.

*Correspondence to: Michelangelo Buonocore, Unit of ClinicalNeurophysiology, Fondazione Maugeri; Via Maugeri 10, 27100Pavia, Italy. E-mail: michelangelo.buonocore@fsm.it

Received 11 August 2009; Accepted 26 October 2009

Published online in Wiley InterScience(www.interscience.wiley.com.), 17 December 2009.DOI: 10.1002/jcp.21996

The vertebral body is a complex laboratory in which a dynamiccondition (continuous bone deposition and resorption) allowsa static function (postural maintenance). The bone remodellinginvolves a series of metabolic processes that change with ageing,becoming clinically relevant in osteoporotic subjects. Thedecline of gonadal hormones in both sexes is the major factorresponsible for bone weakening (Imai et al., 2009). Thiscondition is often accompanied by vertebral fractures andoften, but not always, pain. Unfortunately this painful conditionis resistant to most available drugs and its nature must still beexplained. If pain arises from the vertebral body, as recentlysuggested (Niv et al., 2003), there must be the local presence ofnerve endings. Literature reports indicate the presence ofneural fibres innervating the vertebral body in both rodents andhumans (Sherman, 1963; Antonacci et al., 1998; Ohtori et al.,2007) and that some of these fibres contain substance P,suggesting a possible involvement in nociception (Brown et al.,1997; Fras et al., 2003).

Percutaneous kyphoplasty (PK) has recently beenintroduced into clinical practice for the treatment of vertebralpain deriving from fracture or osteolytic lesion of the vertebralbodies. The technique combines vertebroplasty with aninflatable bone tamp (Lieberman et al., 2001; Pateder et al.,2007). The effectiveness of PK in pain relief has beendemonstrated in several clinical conditions associated withvertebral involvement such as osteoporosis (Jensen and Dion,1997; Deramond and Mathis, 2002) and multiple myeloma(Masala et al., 2008). The technique also provides the possibilityto collect bone specimens directly from the vertebral bodyduring the procedure (Huang et al., 2005). This possibility wasexploited by our laboratory to quantify the vertebral bodyinnervation in male and female patients undergoing kyphoplastyfor compression fracture. Clinical outcomes were alsodetermined.

METHODSSubjects

In the period May 2007 to October 2008, male and female patientsundergoing percutaneous kyphoplasty in the Salvatore Maugeri

� 2 0 0 9 W I L E Y - L I S S , I N C .

Pain Medicine Unit (Pavia, Italy) for the treatment of severe backpain deriving from a vertebral compression fracture were asked toparticipate in the present study (see Table 1). Patients wereincluded in the study based only on clinical indications and werealways independent of the present study. Patients were informedabout the experimental procedures for bone specimen collectionand treatment (as approved by the local ethics committee) andwere asked to sign the informed consent form.

Clinical global impression of change (CGIC)

The CGIC, a 7-point Likert scale (1¼ very much improved to7¼ very much worse), was completed by the physician before thepatient’s discharge in order to evaluate the patient’s condition, withparticular attention to pain relief after kyphoplasty (Guy, 1976).

Vertebral bone biopsy

Balloon kyphoplasty involves the introduction of a cannula into thevertebral body, followed by insertion of an inflatable bone tampdesigned to return the vertebral body to its original height whilecreating a cavity to be filled with polymethylmethacrylate cement.After an extrapedicular or transpedicular entry point into thevertebral body is identified using a guide pin and fluoroscopy, acannulated obturator is placed over the guide wire. The obturatoris then tapped into the bone over the guide wire and a workingcannula is placed over the obturator and advanced until the tip ofthe cannula is seated in the posterior portion of the vertebral body.To remove the bone specimen, a cannula is then inserted into thevertebral body, rotated and removed while suctioning with asyringe.

TABLE 1. Detailed description of the male and female patients

Patients Age Pathol. Spinal level Sections Sections with fibres Percentage Fibres around vessels Sections with free terminals CGIC

MalesDM 41 MYE T11 43 21 49 20 1 2PB 79 OST T12 31 11 35 11 0 2RT 47 OST T12 64 19 30 19 0 3SE 75 OST L1 45 27 60 24 3 3CC 88 OST L1 70 17 24 17 0 3CM 67 OST L1 79 19 24 18 1 2DA 79 MYE L1 172 53 31 53 0 2CC 88 OST L2 33 9 27 9 0 3CVE 58 OST L2 29 6 20 6 0 2

M¼ 69.5 M¼ 35.7%Females

ZM 55 OST T7 72 0 0 0 0 2DL 74 OST T9 51 14 27 14 0 2FG 67 MET T9 40 0 0 0 0 2LR 76 OST T10 43 4 9 4 0 3NA 69 OST T12 44 8 18 8 0 3MP 73 OST T12 36 1 3 1 0 2PR 86 OST L1 23 3 13 3 0 2OL 69 OST L1 58 9 15 9 0 2CE 58 OST L1 26 6 23 6 0 3CG 77 OST L1 36 11 30 11 0 3LR 76 OST L1 106 39 37 39 0 3BF 67 MYE L1 76 23 30 23 0 3OA 82 OST L2 34 27 79 23 4 3ZM 55 OST L2 62 16 26 16 0 2VM 71 OST L3 27 25 92 22 3 3MG 72 OST L3 77 10 13 10 0 3MA 66 OST L3 57 17 30 17 0 3MA 74 OST L3 75 31 41 31 0 2BL 73 MYE L3 18 17 94 17 1 2CD 80 OST L4 66 7 11 7 0 3OM 56 OST L4 84 22 26 22 0 2BG 79 OST L4 41 17 41 15 2 3LL 70 OST L4 34 10 29 10 0 3FC 68 OST L4 63 5 8 5 0 2

M¼ 70.5 M¼ 29%

OST, osteoporosis; MYE, myeloma; MET, metastasis.Spinal levels are grouped and the numbers of fibres not related to vessels are reported. The clinical global impression of change (CGIC) is reported in the last column.

V E R T E B R A L B O D Y I N N E R V A T I O N 489

Histological preparations

The nerve fibres were labelled by indirect immunofluorescence,using a pan-neuronal marker (Protein Gene Product 9.5, PGP 9.5)for nerve fibres and a marker directed against type IV collagen(Col IV) for vertebral blood vessels. In particular, specimens wereimmediately fixed in cold Zamboni’s solution at 48C overnightand then decalcified in ethylenediamine tetraacetic acid (EDTA)solution for 12 h at 48C. After decalcification the samples werewashed three times in phosphate-buffered saline (PBS 0.1 M pH7.4), cryoprotected with 20% sucrose in PBS and stored at 48C untilused. Once frozen with a synthetic resin, specimens were cut into50mm sections with a freezing sliding microtome (HM 450, MicromInternational, Walldorf, Germany), washed in PBS and left as freefloating slices in the block solution: PBSþ 0.3% Triton X-100þ 5%Normal Donkey Serum (NDS, Jackson Immuno Research, WestGrove, PA) for 3 h. The floating sections were then exposed to theprimary antibodies r-PGP 9.5 (Biogenesis, Poole England, 1:1,000)and m-Col IV (Chemicon International, Temecula CA, 1:800) for2–5 h at room temperature and then at 48C overnight. After threewashes in wash solution (PBSþ 0.3% Triton X-100 and 1% NDS),the sections were incubated with the secondary antibodies,Donkey anti-rabbit IgG labelled with Cy3 (1:400) and Donkeyanti-mouse IgG labelled with Cy2 (1:200) (Jackson ImmunoResearch), for 2–4 h at room temperature and then at 48Covernight. After three washes in wash solution and two in PBS, thesections were adhered to coverslips with agar and dehydrated viaan alcohol series. Finally, sections were cleared with methylsalicylate and mounted in dibutyl phthalate xylene (DPX, DDK,Milan, Italy).

JOURNAL OF CELLULAR PHYSIOLOGY

Nerve fibre visualisation and quantification

All slices were visualised with a fluorescence microscope system(Axioskop 40 FL, Zeiss, Gottingen, Germany) equipped with digitalphotocamera (AxioCam MRc 5) and supported by Axio Visionsoftware for imaging collection. Slices were analysed by twoindependent operators.

RESULTS

The results are summarised in Table 1.

Subjects

Following the inclusion criteria, 31 patients were enrolledin the study: 9 males and 22 females with a mean age of70 years for both sexes (range 41–88). As shown in Table 1,the diagnoses were as follows: osteoporosis (n¼ 26),multiple myeloma (n¼ 4), and metastatic lesion(n¼ 1). Three patients had percutaneous kyphoplastic (PK)twice.

Clinical global impression of change (CGIC)

The kyphoplastic procedure, carried out in all patients withstandardised protocol, had a positive outcome as evidenced bythe physicians’ CGIC data collected from all patients: 16reported score 2 (‘much improved’, 42%) and 18 score 3(‘minimally improved’, 58%). The mean CGIC score was 2.58.

Fig. 2. Schematic representation of nerve fibre percentages in maleand female patients. Data are means W SEM

490 B U O N O C O R E E T A L .

PGP 9.5 and Col IV determinations

On average, 55.1 slices per bone specimen were prepared,visualised and analysed. Since concordance between resultsobtained from the two operators was over 90%, the meanof the two measures was considered (Figs. 1A,B and 2).

Nerve fibres visualised with the pan-neuronal marker PGP9.5 were easily identified by fluorescence microscopy(Fig. 1A,B). As reported in detail in Table 1, nerve fibres wereidentified in 539 of the 1,876 histological preparations. All themale specimens showed fibres, while two female samples hadno fibres and another three had a percentage lower than10% notwithstanding the high number of slices per sample. Theconcomitant staining with Col IV antibody revealed that the vastmajority of these fibres accompanied vessels (95.3%; Fig. 1B).

As seen in Figure 2, most of the male specimens derived fromthe lower thoracic (T11–T12) or first lumbar (L1–L2) levels,while in females they were more widely distributed amongthoracic (T7–T12) and lumbar vertebrae (L1–L4). In males, thepercentage of specimens with fibres was higher at T11 andlargely similar at the other levels (30%). In women, the mean

Fig. 1. Nerve fibres revealed by fluorescence microscopy: (A) freenerve fibres (arrows) in the vertebral body of a patient withosteoporosis; (B) nerve fibres (arrows) related to blood vessels(arrowheads) in a patient with myeloma (in green, PGP 9.5 staining ofnerve fibres; in red, type IV collagen staining of blood vessels).Bar U 100mm.

JOURNAL OF CELLULAR PHYSIOLOGY

percentage of specimens with fibres was 29%, but as shown inFigure 2 the percentage was quite different according to thevertebra. The percentage was quite low at thoracic levels(about 11%), while at L2–L3 it increased to 50%.

DISCUSSION

In the present study, we quantified the vertebral bodyinnervation in male and female patients undergoing kyphoplastyfor compression fracture. The main result was that onlyone-third (539/1,876) of the histological preparationscontained nerve fibres.

The presence of nerve endings in normal human vertebraltrabecular bone (also known as cancellous or spongy bone) hasbeen the subject of a few reports, which like the present oneshowed scarce innervation. An early study showed thepresence of few nerve endings in the bone marrow of twopatients, one with neurofibromatosis and another withreticulosarcomatosis (Sherman, 1963); later, Kumar and Davis(1973) reported a lack of innervation in vertebral bodies andintervertebral discs. More recently, nerve fibres weredescribed in the bone marrow and basivertebral vessels ofintact vertebral bodies obtained at autopsy (Antonacci et al.,1998).

The presence of substance P in the basivertebral nerve (Fraset al., 2003) was considered strong evidence that its fibres havethe potential for transmitting pain signals. However, the case ofsarcoma tumour contrasts with the observations supportingthis hypothesis: it arises centrally in a major bone and is oftensilent until it grows sufficiently to irritate the periosteum (Rittset al., 1987).

Percutaneous kyphoplasty (PK) is a well-known procedureto induce immediate pain relief in patients suffering back painbecause of vertebral fracture of different origin. On the basis ofthe present data, it is difficult to consider the pain relief as due tonerve block, as recently suggested by Niv et al. (see later). Thereis evidence that restoration of morphological relationshipsamong spinal elements (bone, muscle, ligaments) has ananalgesic effect (Dreyfuss et al., 2008). Ligaments, together withother structures such as muscles, may participate in painexperienced by patients with vertebral body fractures. Indeed,nerve endings have been described in joint capsules, anteriorand posterior longitudinal ligaments, and annulus fibrosus, andanaesthesia of the posterior spinal elements is able to relievepain (Nath et al., 2008).

V E R T E B R A L B O D Y I N N E R V A T I O N 491

On the other hand, it has been suggested that the analgesicaction of PK is due to a direct effect of the biomaterial injectedinto the vertebral body: it would provide immediate pain reliefvia mechanical and thermal destruction of the vertebral nervefibres (Niv et al., 2003). The present data do not fully supportthis hypothesis since nerve fibres were very scarce in thevertebral bodies and some subjects had no fibresnotwithstanding the high number of slices observed (i.e. 0/72, 0/40). In addition, most of the detected nerve fibres were aroundblood vessels; although a strict sensory function cannot beexcluded, this pattern of distribution suggests that these fibresare mostly sympathetic efferents, that is dedicated to vascularcontrol (Groen et al., 1990).

It is difficult to determine if the small number of fibrespresent in the specimens was due to loss of nerve fibres inconsequence of the compression fracture or if the observedfibres were the result of inflammatory processes in which newvessels were accompanied by nerves, which are absent innormal vertebral bodies. The latter hypothesis implies a paucityof innervation which could result in a low blood supply to thetrabecular bone. Indeed, if the nerve fibres in the vertebraewere involved in modulation of the blood supply, their scarcitysuggests absence of vascularisation, which would be responsiblefor weakening of the bone with consequent fractures. It can behypothesised that the hormonal/metabolic alterations presentin most aged subjects cause a decrease of innervation (withconsequent vascular disregulation) or that vertebral bodyfractures result in deformity, causing abnormal biochemicalstresses which by their nature and location produceinflammation and pain.

Interestingly, it must be underlined that this condition ismuch more evident in women than in men. Indeed, there is aclear sex difference in the vertebral innervation at differentspinal levels. In males, the percentage of vertebral innervationwas ‘relatively’ high in the spinal levels taken into considerationfor treatment (T12, L1 and L2, about 35%), whereas in femalesthe percentage of innervation markedly differed when thoracicor lumbar vertebrae were considered (ranging from 10% to50%). If innervation was a normal constituent of vertebrae, thepercentage would likely be the same at all levels. The differencesrecorded in our study suggest functional differences, meaningthat the innervation can be attributed to the fracture and/or tometabolic disregulation leading to an inflammatory condition.

In conclusion, the presence of very few nerve fibres invertebral bodies (mostly related to vessels) suggests that

JOURNAL OF CELLULAR PHYSIOLOGY

vertebral body innervation is not involved in the pain reliefobtained with kyphoplasty.

Acknowledgments

We thank Dr. Georgia Amato for the invaluable help in theprocessing of specimens.

Literature Cited

Antonacci MD, Mody DR, Heggeness MH. 1998. Innervation of the human vertebral body: Ahistologic study. J Spinal Disord 11:526–531.

Brown MF, Hukkanen MVJ, McCarthy ID, Redfern DRM, Batten JJ, Crock HV, Hughes SPF,Polak JM. 1997. Sensory and sympathetic innervation of the vertebral endplate in patientswith degenerative disc disease. J Bone Joint Surg (Br) 79:147–153.

Deramond H, Mathis JM. 2002. Vertebroplasty in osteoporosis. Semin Musculoskelet Radiol6:263–268.

Dreyfuss P, Snyder BD, Park K, Willard F, Carreiro J, Bogduk N. 2008. The ability of single site,single depth sacral lateral branch blocks to anesthetize the sacroiliac joint complex. PainMed 9:844–850.

Fras C, Kravetz P, Mody DR, Heggeness MH. 2003. Substance P-containing nerves within thehuman vertebral body: An immunohistochemical study of the basivertebral nerve. Spine J3:63–67.

Groen GJ, Baliet B, Drukker J. 1990. Nerves and nerve plexuses of the human vertebralcolumn. Am J Anat 188:282–296.

Guy W. 1976. Clinical global impressions. In: ECDEU Assessment Manual forPsychopharmacology, 1st edition. Rockville, USA: National Institute of Mental Health.pp. 221–227.

Huang KY, Yan JJ, Lin RM. 2005. Histopathologic findings of retrieved specimens ofvertebroplasty with polymethylmethacrylate cement: Case control study. Spine 30:E585–E588.

Imai Y, Nakamura T, Matsumoto T, Takaoka K, Kato S. 2009. Molecular mechanismsunderlying the effects of sex steroids on bone and mineral metabolism. J Bone Miner Metab27:127–130.

Jensen ME, Dion JE. 1997. Vertebroplasty relieves osteoporosis pain. Diagn Imaging (SanFranc) 19:68–72.

Kumar S, Davis PR. 1973. Lumbar vertebral innervation and intra-abdominal pressure. J Anat114:47–53.

Lieberman IH, Dudeney S, Reinhardt MK, Bell G. 2001. Initial outcome and efficacy of‘‘kyphoplasty’’ in the treatment of painful osteoporotic vertebral compression fractures.Spine 26:1631–1638.

Masala S, Anselmetti GC, Marcia S, Massari F, Manca A, Simonetti G. 2008. Percutaneousvertebroplasty in multiple myeloma vertebral involvement. J Spinal Disord Tech 21:344–348.

Nath S, Nath CA, Pettersson K. 2008. Percutaneous lumbar zygapophysial (Facet) jointneurotomy using radiofrequency current, in the management of chronic low back pain: Arandomized double-blind trial. Spine 33:1291–1297.

Niv D, Gofeld M, Devor M. 2003. Causes of pain in degenerative bone and joint disease: Alesson from vertebroplasty. Pain 105:387–392.

Ohtori S, Inoue G, Koshi T, Ito T, Watanabe T, Yamashita M, Yamauchi K, Suzuki M, Doya H,Moriya H, Takahashi Y, Takahashi K. 2007. Sensory innervation of lumbar vertebral bodiesin rats. Spine 32:1498–1502.

Pateder DB, Khanna AJ, Lieberman IH. 2007. Vertebroplasty and kyphoplasty for themanagement of osteoporotic vertebral compression fractures. Orthop Clin North Am38:409–418.

Ritts GD, Pritchard DJ, Unni KK, Beabout JW, Eckardt JJ. 1987. Periosteal osteosarcoma. ClinOrthop Relat Res 219:299–307.

Sherman MS. 1963. The nerves of bone. J Bone Joint Surg (Am) 45:522–528.