Background

Intervertebral disc (IVD) disorders are a major cause of low back pain (LBP) [1], often associated with ageing and degeneration [2, 3]. During IVD degeneration, the annulus fibrosus (AF) integrity is compromised and its extracellular matrix (ECM) becomes disorganized, impairing the nucleus pulposus (NP) integrity [1]. IVD bulging may occur, without rupture of the AF, resulting in the so-called contained disc herniation, when NP extrusion is prevented by either the AF or the posterior longitudinal ligament (PLL) (hereby designated by protused and contained hernias, respectively, following the Lumbar Disc Nomenclature: version 2.0) [4]. When AF and PLL rupture occurs, the NP may be extruded to the epidural space [5, 6] (hereby designated by uncontained hernias). IVD degeneration is currently best evaluated through magnetic resonance imaging (MRI) that allows the analysis of both disc morphology and hydration [7], which is clinically assessed by one of the five Pfirrmann grades that classify disc degeneration regarding disc height, structure and signal intensity [7, 8]. Nevertheless, the Pfirrmann grade does not relate to pain symptoms and hernia containment.

IVD degeneration is characterized by biochemical and mechanical alterations, with disorganization and degradation of ECM, and chronic inflammation, with increased expression of pro-inflammatory cytokines [9]. Recently, the dysregulation of the IVD ECM during degeneration has been associated to tissue fibrosis [10,11,12]. Fibrosis is a pathological event occurring in several organs, characterized by a dysregulation of conventional tissue repair, in response to chronic inflammation, resulting in ECM remodelling and excessive accumulation of specific ECM components, which may alter the biomechanical properties of the tissue [13]. Particularly, ECM alterations are characterized by an increase in collagen I (Col I) content and fibronectin (FN) deposition, accompanied by matrix disorganization, leading to scar tissue formation and eventually resulting in organ dysfunction [10, 14]. Alterations in the concentration and morphology of Col I, collagen II (Col II) and FN have been associated with the progression of human IVD degeneration [11, 15]. In human AF (hAF) tissue, a reduction in Col I has been associated with the degeneration level and ageing [1, 11, 16, 17], suggesting a dual effect of both factors in Col I reduction. Col II has also been shown to be reduced in degenerated IVD, compared to healthy IVD [11, 36]. Figure 2B shows representative images of hAF samples stained with PSR and imaged under polarized light, while Fig. 2C shows the proportion of the collagen fibres (in red, yellow and green) for the different conditions analysed. Results clearly show a lower frequency (< 7%) of the red birefringence (thicker fibres) for all the samples, in comparison with the more mature fibres (yellow/green birefringence). Moreover, fibres with red birefringence were reduced in the herniated samples, compared with AIS. On the other hand, the thinner fibres (green) were present in higher proportion (38–86%) in all the conditions analysed, while the frequency of collagen fibres with intermediate thickness (yellow) range from 12 to 60%. Together with thicker fibres, the intermediate size collagen fibres are also reduced in all herniated samples compared with AIS control, more specifically in protused hernias (12%, **p < 0.01). These fibres slightly increase in contained hernias (30%, p = 0.0718) compared with protused hernias.

In what concerns thinner collagen fibres, an opposite trend is observed. In this case, the AF from all herniated discs present a higher proportion of green fibres, compared with AIS discs, being this significantly higher in the case of protused hernias (87.85%, **p < 0.01). With herniation progression, the presence of these fibres is slightly reduced (69.35%, p = 0.0656 contained hernias). These results suggest an increased synthesis of new collagen fibres in the AF from herniated discs, compared with AIS controls, particularly in protused hernias.

hAF matrix biochemical analysis with herniation progression

A biochemical analysis of the ECM of hAF was performed by histological/IHC analysis of Col I, Col II, FN and sGAG, generally lost during fibrosis (Fig. 3A) (low magnification images are presented in Supplementary Figure S2). From the images, the percentage of staining area was quantified and the results are presented as the median percentage of area or positive cells and respective interquartile range for each marker (Fig. 3B). Figure 3A (a–d) shows representative images for sGAG staining (blue) in the hAF of different samples. By AB/PSR staining, it can be observed a clear higher stained area of sGAG (blue) in hAF from most AIS samples [77.1% (43.8–90.6)] compared to all herniated samples, particularly in contained hernias [7.1% (2.1–21.3), p = 0.0515] and uncontained hernias [1.78% (0.2–81.9), p = 0.0555].

Fig. 3
figure 3

Human AF matrix biochemical characterization with herniation progression. A Histological/IHC staining for a–d: Alcian Blue/Picro-Sirius Red, scale bar: 100 μm; e–h: Collagen I, scale bar: 100 μm; i–l: Collagen II, scale bar: 100 μm; m–p: Fibronectin, scale bar: 100 μm. B Quantification of each staining per herniation type. Data presented using dot plots, with median and interquartile range. Kruskal-Wallis test followed by corrected Dunn’s were performed. *p < 0.05; **p < 0.01. C Multivariate analysis of interaction of hernia containment level with age for each staining

Full size image

Regarding the expression of Col I (Fig. 3A (e–h) and 3B), it is highly present in AF tissue from AIS samples (Fig. 3A, e) [84.0% (65.4–98.1)] and significantly decreased in the AF of all herniated samples [8.0% (3.1–16.0), **p < 0.01, in protused hernias; 11.9% (2.2–49.7), **p < 0.01, in contained hernias; and 13.5% (1.4–47.4) *p < 0.05, in uncontained hernias] (Fig. 3A, f–h). Col II stained area was also assessed (Fig. 3A (i–l) and 3B), showing an opposite trend of Col I. The stained area of Col II is lower in hAF from AIS samples [20.5% (3.5–35.8)] (Fig. 3A, i) increasing in hAF from herniated samples (Fig. 3A, j–l). This increase is close to significant in hAF of contained hernias [65.4% (32.8–82.5), p = 0.0717] and uncontained hernias [68.9% (26.8–88.8), p = 0.0976].

Relatively to FN expression (Fig. 3A (m–p) and 3B), the results show low and heterogeneous FN expression in protused hernias [0.5% (0.04–15.8)], with 2 (out of 9) samples showing higher FN stained area (about 25.8 and 30.3%). Nevertheless, FN expression tends to increase in hAF from herniated IVD, being significantly higher in contained hernias samples [23.5% (3.2–59.0), *p < 0.05], and in uncontained samples [14.8% (6.6–42.8), p = 0.0516].

In addition, to discard that the differences observed might be due to age differences, the interaction between the variables “age” and “hernia containment” for all the markers analysed was addressed using a multivariate analysis. The p values obtained (Fig. 3C) demonstrate an absence of interaction between the variables “age” and “herniation progression” (p > 0.05) for sGAG, Col I, Col II and FN presence in hAF, reinforcing that the differences in the biochemical composition of hAF ECM with herniation progression are not due to age differences between the different donors.

Moreover, for each herniation stage (protused, contained and uncontained), a linear regression between hAF ECM stained area and donor age was performed (Fig. 4). With this analysis, it is possible to verify a positive correlation for sGAG and Col I with ageing only in the group of hernias contained by PLL (r2 = 0.2346, p = 0.0488 for sGAGs and r2 = 0.259, p = 0.037 for Col I, respectively), suggesting that, within this herniation containment level, Col I and sGAG increase with ageing.

Fig. 4
figure 4

Correlation of expression levels of hAF matrix components with age, within each herniation type. Data presented using dot plots, with 95% confidence intervals and indication of r2 and p value for each linear regression

Full size image

hAF fibrotic analysis with herniation progression at the cell level

Additionally, relevant cellular markers to tissue/IVD fibrosis (α-SMA+ cells, MMP12+ cells and CD68) were also evaluated in hAF samples by IHC analysis.

The α-SMA+ cells were absent from hAF of AIS samples (Fig. 5A (a–d) and 5B), as well as in a high percentage of herniated samples (in 33.3% of the protused hernias, 47% of the contained hernias and 38.5% of the uncontained hernias), suggesting a high heterogeneity of the α-SMA expression in hAF. Nevertheless, in the samples presenting α-SMA+ cells, an increased expression of α-SMA in the AF from herniated tissues was observed, ranging from 2.27 to 94.71%, when compared to AIS samples.

Fig. 5
figure 5

Human AF fibrotic analysis at the cell level with herniation progression. A IHC staining for a–d: α-SMA, scale bar: 50 μm; e–h: MMP12, scale bar: 50 μm; i–l: macrophages (CD68) scale bar: 50 μm. B Quantification of each staining per herniation type. Data presented using dot plots, with median and interquartile range. Kruskal-Wallis test followed by corrected Dunn’s were performed. *p < 0.05; **p < 0.01. C Multivariate analysis of interaction of hernia containment level with age for α-SMA. D Correlation of α-SMA staining quantification with age for each herniation type. Data presented using dot plots with 95% confidence intervals and indication of r2 and p value for each linear regression

Full size image

In what concerns MMP12 (Fig. 5A (e–h) and 5B), another marker associated with IVD fibrosis, heterogeneity was also observed, since MMP12+ cells were absent in 33.3% of AIS samples, in 11.1% protused hernias, in 5.9% of contained hernias and in 9.1% of uncontained samples. No differences were observed regarding the presence of MMP12 in the different herniated conditions, with most samples exhibiting MMP12+ cells below 50%.

Furthermore, macrophage presence in hAF samples was also assessed by the expression of CD68+ cells (a common marker used for macrophages) (Fig. 5A, i–l, and Fig. 5B). CD68+ cells were observed in 33% of AIS samples, in 24% of contained hernias and in 11.7% of uncontained samples, but not in protused hernias (Fig. 5A,i–l). Macrophages were present in AF tissue both, dispersed (Fig. 5A, k) and agglomerated (Fig. 5A, l), the last predominantly in hAF tissue borders (Fig. 5A, l). Due to this observation, macrophage infiltration area was quantified (Fig. 5B), instead of percentage of positive cells. Nevertheless, the quantification of CD68+ area revealed no significant differences between hernia conditions.

As previously described for ECM markers, the interaction between the variables “hernia containment” and “age” was addressed for α-SMA expression using a multivariate analysis (Fig. 5C), but no interaction was found between the two variables (p value = 0.291). Moreover, no linear correlations were found between the increase of α-SMA expression with ageing for each hernia containment group.

Discussion

The AF plays a functional and crucial role in IVD homeostasis since it is responsible for the containment of the NP, preventing its extrusion. A pro-inflammatory environment and extensive mechanical stresses can weaken AF mechanical properties [37], ultimately leading to IVD herniation. But the mechanisms behind IVD herniation and AF disruption remain unclear. The IVD, and particularly the AF ECM, undergoes structural, biochemical and biomechanical alterations that have been studied in different models, as canine, rabbit, mice or bovine [38,39,40,41]. IVD degeneration has been associated with tissue fibrosis [10,11,12], but this has been poorly addressed. It is crucial to understand tissue fibrosis in the IVD, so we could disclose novel cellular and molecular cell targets to stop/revert IVD herniation. Taking this into account, this study aimed to evaluate the fibrotic alterations of AF with hernia progression (here referred to the medical evaluation of hernia containment levels).

Currently used in IVD clinical evaluation, Pfirrmann’s MRI-based classification lacks suitability to assess herniation progression, among other problems [7, 8]. Although considering the structure, brightness and height of the disc [8], Pfirrmann grading system does not distinguish disc herniation progression. For example, in this study, only 5 samples were described as Pfirrmann grade V, characterized by the collapse of IVD to the epidural space [8], but 13 samples were identified as extruded to the epidural space by the surgery team. It is important to combine different classification systems that embrace both image analysis and clinical evaluation, to a more accurate definition of each herniation condition. In the future, systemic markers would be important to improve the accurate diagnosis of IVD herniation progression stages.

The ECM of hAF was first analysed regarding its ultra-structure and biochemical composition. Matrix disorganization and impairment in the AF tissue are characteristics of IVD degeneration and herniation. As expected, in the most severe herniation stage, a more disorganized matrix was found. A low percentage of thicker (more mature) collagen fibres (red) was revealed in all samples. A higher amount of thinner collagen fibres (green birefringence) present within the herniated tissue was observed, suggesting recent ECM remodelling.

Regarding ECM biochemical alterations of hAF, this work used a panel of fibrotic ECM-related markers (Col I, Col II, FN, sGAG) and compared AF from herniated IVD with AF from AIS patients, as control group, due to the impossibility of obtaining human adult healthy IVD. A tendency for lower sGAG stained area in AF tissue from contained hernias and uncontained compared to AIS samples was observed. These changes are in line with the loss of sGAGs previously reported for IVD degeneration in human, as well as in canine and rabbit models [16, 41,42,43]. Regarding Col I, a decreased expression in herniated IVD was observed compared with AIS samples but increasing with herniation progression. The reduction of Col I stained area in AF from herniated tissue, together with increased disorganization of the fibres may contribute to an increase in AF susceptibility to rupture, impairing NP confinement and allowing herniation [1]. A reduction in Col I in AF tissue has been associated with ageing [1, 11, 16], and with the degeneration level in both goat [44] and human (in vitro study) [17].

Col II stained area in the AF showed a trend to increase in the most herniated tissues, in opposite to Col I, suggesting ECM turnover regarding the collagen composition of the AF. It is possible that this ECM turnover is also accompanied by alterations in the mechanical properties of the tissue and, therefore, it may become more susceptible to herniation, since Col II has a smaller elastic modulus than Col I [45]. Further biomechanical tests should be conducted to evaluate this possibility. Nevertheless, the literature describes a reduction of Col II from non-degenerated to degenerated or herniated IVD, both in AF tissue [11] or whole IVD [18, 19]. In fact, there is a lack of knowledge of the herniation containment level in most of the literature. Additionally, FN stained area was shown to increase from hernias contained by AF to more severe herniation stages. This is in accordance with the literature, in which FN showed to increase from low to high degeneration grades (evaluated by Thompson’s scale) [15], as well as between protrusion and extrusion [18], both in human IVD.

Concurrently, hAF fibrotic markers were also evaluated at the cellular level. In the current work, α-SMA+ cells were absent in all AIS samples, suggesting MF activity in AF tissue only during IVD herniation. α-SMA+ cells have been previously reported to increase from non-degenerated to degenerated IVD, both in human and rat [12]. Hastreiter et al. reported an increase percentage of α-SMA+ cells in extruded tissues, when compared to scoliosis samples [46]. In our study, most of the samples present low levels of α-SMA+ cells (< 20%), in agreement with previous reports (4–15% of α-SMA expression) [46]. However, some heterogeneity between the samples was observed, suggesting that, although a reliable marker for fibrosis in several tissues, α-SMA might not be optimal for IVD herniation.

We have also investigated the expression of MMP12, a pro-fibrotic marker upregulated under inflammatory conditions [12]. MMP12 has been associated with IVD degeneration in a rat model [29] and in human AF [12]. However, although Lv et al. suggested MMP12 as a degeneration marker [12], our study did not show any correlation between MMP12 and the progression of IVD herniation, since MMP12+ cells were observed in most of control samples and absent in some of the herniated samples. Moreover, although MMP12 has been associated with α-SMA, herein our results indicate higher expression of MMP12 than α-SMA [12]. Lastly, we also evaluated the presence of macrophages, important players in tissue fibrosis. We observed macrophage infiltration in a small number of samples and in a low percentage (< 2.0%), distributed within the outer layers of the AF. Moreover, although macrophages were found in some AIS samples, these cells were not present in protused hernias. Previous studies detected CD68+ macrophages in the AF of pig, mice and human [27, 28, 47]. Gronblad et al. highlighted their abundant presence in AF of human herniated tissue, but without analysing differences between hernia types [48]. Our findings are in agreement with previous reports where macrophages were detected in herniated IVD and are believed to be key players in hernia resorption through phagocytosis [3, 49]. Additionally, an evaluation of a possible interaction of age within the comparisons regarding hernia containment was conducted, revealing no influence of this factor in the observed differences. However, no differences were observed between different herniated conditions suggesting that the ECM fibrotic alterations do not correlate directly with the frequency of macrophages present on hAF, indicating that other players might be involved.

This study presents some limitations. The use of thoracic/lumbar discs collected during anterior release surgeries for AIS may not be the ideal control for lumbar discs collected from patients operated for herniated lumbar discs. As in practice IVD samples from healthy donors are almost inaccessible, AIS samples were the most suitable option for control group. However, it has to be considered that although MRI analysis suggests that the NP from scoliosis patients retains in normal conditions of water, the IVD of these conditions might be under abnormal spine load and effects on AF tissue are still not well defined [12, 50]. AIS discs have been described to present sparse elastic fibres with some disorganization in collagen and elastic fibre networks, loss of lamellar structure in the AF and presence of cell clusters [51]. In addition, AIS AF presents a decrease in elastin and Col I and decrease of Col III and Collagen IV [52]. Therefore, caution should be taken when analysing the results since some differences can either be covered or less attenuated [12]. Moreover, a reduced sample size of different hernia types might influence the statistical power of comparisons, but still it was possible to observe significant differences between parameters analysed in the hAF tissues. Although our conclusions have shown to be independent of the age of the patients, we cannot exclude the influence that other external factors (gender, body mass index, smoking habits or other co-morbidities) might have in the analysis. In the future, it would be relevant to address other fibrosis markers for a better understanding of AF failure in IVD herniation.

Conclusions

To the best of our knowledge, this is the first study that characterizes fibrotic events in the AF tissue of herniated IVD in a systematic way, at distinct stages of hernia containment. This study suggests important alterations in the hAF tissue during herniation that have been associated with tissue fibrosis. Structural changes were observed in the distribution of collagen fibres in herniated AF, in particular their lack of organization in uncontained samples. The stained area of sGAG and Col I were decreased in all the containment levels of herniated AF when compared with AIS AF, while FN and Col II were increased in AF of more advanced stages of IVD herniation. Moreover, α-SMA tends to increase in the AF of uncontained hernias. Hence, our work indicates herniation progression stage as an important parameter to understand the process of tissue fibrosis in hAF and consequently, of AF failure and IVD herniation.