Advertisement

International Orthopaedics

, Volume 42, Issue 7, pp 1557–1563 | Cite as

Raman spectroscopy reveals differences in molecular structure between human femoral heads affected by steroid-associated and alcohol-associated osteonecrosis

  • Ema Nakahara
  • Wenliang Zhu
  • Giuseppe Pezzotti
  • Hidetoshi Hamada
  • Masaki Takao
  • Takashi Sakai
  • Nobuhiko Sugano
Original Paper

Abstract

Purpose

The purposes of this study were to document novel Raman spectroscopic findings in femoral heads affected by osteonecrosis and to identify molecular structure differences based on aetiology.

Methods

We obtained 13 femoral heads with osteonecrosis from 13 different patients who underwent total hip arthroplasty. Comparisons were made between the viable zones of each femoral head examined. The samples were scanned with X-ray micro-CT for structural mapping and a central coronal section slab was prepared for Raman spectroscopy and histological analyses. Raman spectra were collected at different locations, including the viable and necrotic zones of the femoral head, using a highly spectrally resolved Raman microprobe.

Results

Significant alterations in the spectral morphology in the high wavenumber region were found, with a pronounced inhibition of peculiar lipid signals in the frequency interval 2851 ~ 2890 cm−1 and at ~ 1750 cm−1. The necrotic zone in steroid-associated osteonecrosis showed an increase in the ratio of lipid-related bands to protein-related bands, while alcohol-associated osteonecrosis exhibited a decrease in this ratio.

Conclusions

We systematically found a decrease in Raman intensity for sphingomyelin and phenylalanine fingerprint bands in the necrotic zones, and these differences may be related to the etiology of osteonecrosis.

Keywords

Osteonecrosis of the femoral head Raman Spectroscopy Aetiologic factor Histology 

Introduction

Osteonecrosis of the femoral head (ONFH) is the most common hip disease in many Asian countries and a major cause of total hip arthroplasty (THA) [1]. Although various aetiologic factors have been reported, the pathological mechanisms of femoral head ischemia have not yet been clarified [1, 2]. Histological evaluation can only partially verify femoral head osteonecrosis and reparative tissue reactions even in biopsy specimens in the early stage of osteonecrosis, primarily because histological diagnosis relies on cell and tissue morphology. Therefore, it is difficult to identify histological differences between the femoral heads of patients with osteonecrosis due to different aetiologies, such as corticosteroids, alcohol, or fracture-induced mechanical vascular disruption.

Recently, Raman spectroscopy was introduced as a suitable tool to examine the biochemical composition and the changes in molecular structure of the extracellular matrix in bone and cartilage [3, 4, 5]. Accordingly, we hypothesized that Raman analyses of unknown pathologies of ONFH could also be developed to help create suitable vibrational algorithms to distinguish between the different aetiologies of ONFH. We systematically analyzed a number of femoral heads obtained directly from the operating room to look for differences in the bone marrow at the molecular level, which may relate to aetiologic factors. The goals of this study were to document new Raman spectroscopic findings in femoral heads with osteonecrosis and to identify differences due to the patients’ aetiologies.

Materials and methods

Our institutional review board approved this study, and informed consent was obtained from each patient. We obtained 13 necrotic femoral heads from 13 different patients who underwent THA. Eleven of them were non-traumatic ONFH, and the remaining two were secondary ONFH. The demographic information of the patients is presented in Table 1. The stage of osteonecrosis was evaluated using the Japanese Investigation Committee (JIC) classification [6]. The femoral head samples were scanned with X-ray micro-CT for structural mapping, and a central coronal section slab was prepared for Raman spectroscopy and histology. Examples of optical and CT scan results are shown in Fig. 1a, b, respectively, for patient number 1 in Table 1.
Table 1

Patient demographics

Patient number

Age (years)

Gender

Side

Aetiology of osteonecrosis

JIC stage6

1

66

M

L

Alcohol

3A

2

40

M

L

Steroid

3A

3

33

M

L

Alcohol

3A

4

53

F

R

Steroid

3B

5

68

F

L

Post-femoral neck fracture

3A

6

63

M

L

Alcohol

3B

7

68

M

R

Steroid

1

8

54

M

L

Alcohol

3A

9

80

F

L

Rapidly destructive coxarthrosis

Not applicable

10

63

F

L

Steroid

4

11

27

F

L

Steroid

4

12

64

M

L

Steroid

3B

13

60

F

R

Steroid

3B

Fig. 1

Optical picture (a) and X-ray micro-CT scan image (b) of the sectioned centre slab of a femoral head with stage 3A alcohol-associated osteonecrosis in a 66-year-old man (patient number 1 in Table 1)

Raman spectra were collected at different locations, including the viable and necrotic zones of the femoral head, using a Raman microprobe spectrometer. Raman experiments were conducted within one week of surgery, and the samples were kept in an evacuated environment at around 0 °C. The Raman excitation source in the present experiments used a 532-nm Nd:YVO4 diode-pumped solid-state laser operating with a power of 200 mW (SOC JUNO, Showa Optronics Co., Ltd., Tokyo, Japan). An objective lens with a numerical aperture of 0.5 was used both to focus the laser beam on the sample surface and to collect the scattered Raman light. A pinhole aperture of 100 μm was adopted while employing an objective lens with a magnification of ×100. For each studied sample, several tens of spectra were collected, from which an average spectrum could be extracted with statistical validity. Average spectra were deconvoluted upon fitting them to mixed Gaussian-Lorentzian sub-bands. In the case of patients 2 and 10 (Table 1), a strong fluorescence background was found, resulting in difficulty in collecting a reliable signal.

To limit data scatter related to the intrinsic characteristics of individual patients, comparisons were always carried out between healthy and necrotic regions of the femoral head in the same patient. Statistical histograms were then prepared using spectral mapping of the whole head, and a comparison was carried out between different samples with respect to the healthy and necrotic zones (HZ and ONZ, respectively) of the osteonecrotic bone.

Results

As can be observed in the optical and X-ray micro-CT images taken of patient number 1 (Fig. 1a, b, respectively), the presence of a necrotic zone could be clearly identified in the femoral head, while the other zone was found to be viable. Typical Raman spectra recorded in osteonecrotic and viable zones of the sample (cf. locations marked in Fig. 1a) are shown in Fig. 2. The Raman spectra were collected at three different spectral regions: the hydroxyapatite (HAP) region (Fig. 2a), the amide bond region (Fig. 2b), and the C–H vibration-related high wavenumber region (Fig. 2c). The spectra demonstrated significant variations between the viable and necrotic zones, but a homogeneous morphology within the distinct zones. In the high wavenumber region, the bands located in the range from 2800 to 2900 cm−1 were conspicuously suppressed in the necrotic zone, which had a Raman spectral line shape that was highly similar to that of pure type I collagen (reference spectrum; cf. Fig. 2d). Figures 3 and 4 show representative examples of bones affected by osteonecrosis associated with alcohol (patient number 1) and steroids (patient number 7), respectively. These figures compared deconvoluted spectra collected at the necrotic and viable zones. Assignments of all of the observed bands are shown in Table 2. In femoral heads with osteonecrosis, vibrational bands of the inorganic HAP component dominated the Raman spectrum of the femoral head in the low wavenumber region, but emissions from organic tissues also appeared, including structural proteins, amino acids, and lipids (Figs. 3 and 4, Table 2). Regarding emissions from the inorganic components of the spectrum, emissions from apatite at different locations of the viable zones were relatively uniform within the same femoral head (Fig. 2a), but showed clear variations between different samples (Figs. 3 and 4). The presence of different calcium phosphate phases in the bone (Ca3(PO4)2 (TCP) and CaHPO4) were observed at the higher wavenumber shoulder of the main HAP band, through their peculiar bands. The spectra showed slight variations in the contents of such phases in the necrotic zones. A decrease in the intensity of the P–O symmetric stretching band of HAP were observed in the necrotic zone, as well as broadening of its deconvoluted lower wavenumber shoulder, owing to lattice distortions induced by the generation of vacancies and PO4 vacant sites. For all samples, the value of the band area ratio between the shoulder at ~ 950 cm−1 and the main band at 960 cm−1 showed an increase in the necrotic zones compared with the respective viable zones (cf. Fig. 5a).
Fig. 2

Raman spectra in (a) low, (b) middle, and (c) high wavenumber spectral regions, collected at different locations in the viable zone (H1, H2, H3) and the necrotic zone (O1, O2, O3) of a sectioned femoral head with stage 3A osteonecrosis in a 66-year-old man with alcohol-associated osteonecrosis (patient number 1 in Table 1). (d) A comparison of the average spectra from the healthy and necrotic zones with those of a type I collagen sample

Fig. 3

Deconvolution of the averaged Raman spectra in different spectral regions collected from the viable zone and necrotic zone of a sectioned femoral head with stage 3A alcohol-associated osteonecrosis in a 66-year-old man (patient number 1 in Table 1)

Fig. 4

Deconvolution of the averaged Raman spectra in different spectral regions collected in the healthy zone and the necrotic zone of a sectioned femoral head with stage 1 steroid-associated osteonecrosis in a 68-year-old woman (patient number 7 in Table 1)

Table 2

Assignment of the Raman bands

Band label

Band position (cm−l)

Principal assignment

Band 1

960

v1(PO4)

Band 2

1005

v(C–C) in phenylalanine

Band 3

1020~1070

v3(PO4)

Band 4

1081

v(CO32−), v(PO43−) in lipids

Band 5

1156

v(C–C) in sphingomyelin

Band 6

1173

C–O–C in proteins

Band 7

1208

Stretching mode of C–C6H5 in tyrosine and phenylalanine

1242

Amide III, C–N stretching, and CH2 wagging

1258

Amide III, adenine, and cytosine β-sheet structure

Band 8

1315

Amide III, CH2 bending mode in ɑ-helix

1338

Amide III, ɑ-helix N–H bending, C–N stretching

Band 9

1370

Ring and C–N stretching in cytosine and guanine

Band 10

1444

δ(CH, CH2) in proteins + lipids

Band 11

1526

v(C=C) in sphingomyelin

Band 12

1555

Amide II, N–H bending, and C–N stretching

Band 13

1605

C=C stretching in phenylalanine and tyrosine

Band 14

1638

Amide I, ɑ-helix + β-sheet (C=O stretching vibrations)

Band 15

1660

Amide I, C=O stretching vibrations in ɑ-helix

Band 16

1681

Amide I, v(C=O) in disordered structure

Band 17

1750

C=O stretching mode in lipids and phospholipids

Band 18

2851

vs(CH3) in lipids (liquid)

Band 19

2877

vs(CH3) in lipids (hexagonal) and symmetric stretching of CH3 units in collagen

Band 20

2888

vs(CH3) in lipids (orthorhombic) and symmetric stretching of CH3 units in collagen

Band 21

2910

v(CH2) and v(CH3)

Band 22

2930

vs(CH3) in proteins

Band 23

2956

vas(CH3) in proteins

Band 24

2990

CHɑ,ɑ′ stretching

Band 25

3010

v(HC=CH)

Fig. 5

Intensity ratios of (a) I954/I960 in hydroxyapatite and (b) I1680/I1656 for the amide I bands obtained in the viable zone (VZ) and necrotic zone (ONZ) of the investigated bone samples

Moreover, a significant decrease in the intensity of the Raman bands of sphingomyelin (band 5 at 1156 cm−1 and band 11 at 1526 cm−1) and phenylalanine (band 2 at 1005 cm−1) was observed in the necrotic zones of all of the investigated samples. Amide I components were quite sensitive to changes in collagen secondary structure. A pronounced increase in the area ratio between band 16 (amide I, C=O stretching in the disordered structure) and 15 (amide I, C=O stretching vibration in the α-helix), as well as a broadening on the higher wavenumber side were observed in the osteonecrotic zones (Figs. 3 and 4). An increase in the band ratio of amide I in the necrotic zone was observed in all of the investigated samples, irrespective of the etiology of necrosis (cf. Fig. 5b). Despite the reproducibility of this result, no clear spectral difference was found between the effects of alcohol and steroids in this spectral zone.

In contrast, a significant alteration in the spectral morphology in the high wavenumber region was found in the necrotic zone of patient number 1, exhibiting a pronounced inhibition of lipid signals at 2851 ~ 2890 cm−1 and of band 17 (C=O stretching mode in lipids and phospholipids).

Conversely, in the case of steroid-associated ONFH, the lipid bands in the high wavenumber range were still strong with respect to the collagen band in the necrotic regions despite a decrease in the overall intensity of the spectrum (Figs. 3d and 4d). Figure 6 shows a comparison of the intensity ratios of lipid-related bands to protein-related bands, I18+19+20/I21+22+23, in bones with and without osteonecrosis for all of the investigated samples. The necrotic zone in steroid-associated ONFH showed an increase in the ratio of lipid-related bands versus protein-related bands, while alcohol-associated ONFH experienced a decrease in this ratio.
Fig. 6

Intensity ratio of lipid-related bands vs. protein-related bands for the investigated samples. The necrotic zone (ONZ) in steroid-associated osteonecrosis showed an increase in this ratio, while that in alcohol-associated osteonecrosis revealed a decrease in this ratio compared with the respective viable zone (VZ)

Finally, it should be noted that two cases of ONFH in Table 1 were neither related to steroids nor to alcohol use. Patient number 5 developed ONFH due to a femoral neck fracture and subsequent vascular disruption, while the necrotic segment in patient number 9 occurred secondary to fragmentation due to rapidly destructive coxarthrosis. The patterns of these two cases were somewhat different from the other patterns observed in this study. Although there were similarities in the Raman spectra collected in the intermediate wavenumber region between these samples and the ONFH classified here (i.e., degradation of phenylalanine owing to a lack of nutritional supplementation and of sphingomyelin caused by membrane disruption and cell death), we did not have sufficient statistical power (only one case for each type of sample) to make these spectroscopic results generalizable.

Discussion

The Raman spectrum can provide information peculiar to the molecular structure of a chemical species, and it is possible to identify an unknown material by comparing it with a previously built database. Since the morphology of a characteristic Raman peak for a given functional group strongly depends on stoichiometric and crystallographic alterations in its chemical structure, Raman spectroscopy can be used to examine the biochemical alterations of molecular structures arising from different pathologies. Previous studies have applied Raman spectroscopy to the evaluation of the extracellular matrix in degenerated cartilage, osteoporotic bone, and osteonecrotic bone [3, 4, 5]. The intensity ratios of the bands of HAP, the amide I ratio, and the carbonate/HAP ratio, are parameters that indicate alterations in composition and structure of the necrotic bone tissue of patients with osteonecrosis.

In the cases included in this study, the observed increase in the ratio of I950/I960 revealed a structural distortion of the HAP lattice, due to an unbalanced process of bone generation and resorption in the necrotic zone. The alterations in structure and composition of inorganic mineral components in the bone can in turn affect the collagen fibril matrix through molecular unfolding of the α-helix, as revealed by the pronounced increase in the band area ratio between band 16 and 15, as well as a broadening on the higher wavenumber side of the necrotic zones. The morphological feature of the amide I stretching band has been shown to be related to bone quality [7], so the fracture resistance and tensile strength of the bone in the necrotic zone may be greatly reduced.

The observed significant inhibition of the Raman bands of sphingomyelin and phenylalanine in the necrotic zones indicated degradation of sphingomyelin, possibly caused by membrane disruption due to cell death, and of phenylalanine, potentially owing to a lack of nutritional supplementation. The disruption of the membrane may also result in a reduction in amphipathic lipids, and this can be revealed by a decrease in the intensity of the lipid bands observed in alcohol-associated ONFH. This phenomenon was indeed represented by a decrease in the ratio of lipid-related bands to protein-related bands as observed in alcohol-associated ONFH. The decomposition of lipid in the necrotic bones results in the presence of type I collagen fibres in the matrix. In cases of steroid-associated ONFH, it is assumed that the administration of steroids may suppress the degradation of membrane lipids, and thus increase the intensity ratio of lipid-related bands to collagen bands with respect to viable zones (Fig. 6).

A limitation of this study was that because of the overlapping of the Raman bands of proteins and lipids in the high wavenumber region (2870–2920 cm−1), uncertainty may eventually arise with respect to the exact spectral locations of the sub-band components. Accordingly, spectral deconvolution may contain errors and data scatter. Despite this shortcoming, this study confirmed that intensity variations of the Raman bands in the necrotic zones were systematically observed in relation to the etiology of osteonecrosis. Future studies should involve a larger number of patients to obtain more spectral diversifications for specific pathologies and to obtain greater statistical power.

Conclusion

The Raman bands arising from both inorganic (i.e., HAP) and organic (i.e., lipids and proteins) components of the bone tissue showed clear changes in spectral line-shape for necrotic zones compared with those of the viable zones. We also observed significant degradation of sphingomyelin and phenylalanine in the necrotic zones, and there were important differences influenced by the aetiology of osteonecrosis in the Raman spectra. Finally, the intensity ratio of lipid-related bands to protein-related bands, I18+19+20/I21+22+23, exhibited different trends in the necrotic zone compared with the viable zone in osteonecrosis due to different aetiologies; the necrotic zone in steroid-associated ONFH showed an increase in this ratio, while in alcohol-associated ONFH, there was a decrease in the lipid/protein band ratio in the viable zones.

Notes

Funding information

The principal investigator (NS) of the Japanese Investigation Committee under the auspices of the Ministry of Health Labour and Welfare received a research grant from the Ministry of Health and Welfare of Japan (Research on Measures for Intractable Diseases).

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. 1.
    Wakaba F, Mikihiro F, Toshikazu K et al (2010) Nationwide epidemiologic survey of idiopathic osteonecrosis of the femoral head. Clin Orthop Relat Res 468:2715–2724CrossRefGoogle Scholar
  2. 2.
    Choi H-R, Steinberg ME, Cheng EY (2015) Osteonecrosis of the femoral head: diagnosis and classification systems. Curr Rev Musculoskelet Med 8:210–220CrossRefPubMedPubMedCentralGoogle Scholar
  3. 3.
    T B, Niciejewski K, Kozielski M, Szybowicz M, Siatkowski M, Krauss H (2012) Identifying compositional and structural changes in spongy and subchondral bone from the hip joints of patients with osteoarthritis using Raman spectroscopy. J Biomed Opt 17(1):017007CrossRefGoogle Scholar
  4. 4.
    Takahashi Y, Sugano N, Takao M, Sakai T, Nishii T, Pezzotti G (2014) Raman spectroscopy investigation of load-assisted microstructural alterations in human knee cartilage: preliminary study into diagnostic potential for osteoarthritis. J Mech Behav Biomed Mater 31:77–85CrossRefPubMedGoogle Scholar
  5. 5.
    Pezzotti G, Rondinella A, Marin E, Zhu W, Aldini NN, Ulian G, Valdrè G (2017) Raman spectroscopic investigation on the molecular structure of apatite and collagen in osteoporotic cortical bone. J Mech Behav Biomed Mater 65:264–273CrossRefPubMedGoogle Scholar
  6. 6.
    Sugano N, Atsumi T, Ohzono K, Kubo T, Hotokebuchi T, Takaoka K (2002) The 2001 revised criteria for diagnosis, classification, and staging of idiopathic osteonecrosis of the femoral head. J Orthop Sci 7:601–605CrossRefPubMedGoogle Scholar
  7. 7.
    Oshima Y, Iimura T, Saitou T, Imamura T (2015) Changes in chemical composition of bone matrix in ovariectomized (OVX) rats detected by Raman spectroscopy and multivariate analysis, Proc. SPIE 9303, Photonic Therapeutics and Diagnostics XI, 93033S:1–4Google Scholar

Copyright information

© SICOT aisbl 2018
corrected publication April/2018

Authors and Affiliations

  • Ema Nakahara
    • 1
  • Wenliang Zhu
    • 1
  • Giuseppe Pezzotti
    • 2
  • Hidetoshi Hamada
    • 3
  • Masaki Takao
    • 1
  • Takashi Sakai
    • 3
  • Nobuhiko Sugano
    • 1
  1. 1.Department of Orthopaedic Medical EngineeringOsaka University Graduate School of MedicineOsakaJapan
  2. 2.Ceramic Physics LaboratoryKyoto Institute of TechnologyKyotoJapan
  3. 3.Department of Orthopaedic SurgeryOsaka University Graduate School of MedicineOsakaJapan

Personalised recommendations