MSC-based intradiscal treatment has recently attracted attention for its potential to revolutionize the treatment of chronic discogenic LBP by repopulating the IVD and restoring functional tissue through matrix synthesis by implanted cells and possible beneficial effects on native cells [

5

]. At least three published clinical studies and one unpublished clinical trial (ClinicalTrials.gov NCT01290367) have used MSC-based therapies to treat chronic discogenic LBP [

12

,

13

,

33

] (Table 

5

). Yoshikawa et al. [

13

] first reported significant LBP reduction and rehydration of the treated disc following percutaneous injection of autologous BM-MSCs within a collagen sponge in two older patients with markedly degenerated discs. Orozco et al. [

12

] reported a pilot study of 10 patients with chronic discogenic LBP treated with percutaneous intradiscal injection of autologous BM-MSCs, in which 90% of the subjects reported clinical benefit, and both LBP and disability were greatly reduced at 3 months after transplantation, followed by modest additional improvement at 6 and 12 months. MRI also revealed marginally improved hydration of the treated disc at 12 months [

9

,

12

]. Similarly, Pettine et al. [

33

] evaluated the use of autologous, nonexpanded concentrated BM cells considered to contain multiple stem and progenitor cells, including MSCs. In total, 26 patients with chronic LBP underwent intradiscal injection of the concentrated BM cells, and significant improvement in the VAS, ODI, and modified Pfirrmann score was found at 3, 6, and 12 months post transplantation. Although the exact mechanism in Pettine et al.’s study remains unclear, patients who received > 2000 colony-forming unit-fibroblasts (CFU-F) per milliliter of bone marrow aspirate showed significantly greater improvement of discogenic LBP compared to patients receiving 

9

,

34

]. The 36-month results from a 100-patient, four-arm (high dose of 1.2 × 10

7

mesenchymal precursor cells (MPCs) with HA, low dose of 0.6 × 10

7

MPCs with HA, saline injection, and HA injection), randomized, placebo-controlled phase 2 trial were recently announced. In that trial, 41% of the low-dose group and 35% of the high-dose group achieved 50% pain reduction at 24 months, whereas 82% of the low-dose group who achieved 50% pain reduction over 24 months maintained treatment success at 36 months (Mesoblast website: file:///C:/Users/user/Downloads/Durable%20Three%20Year%20Outcomes%20in%20Disc%20Disease.pdf) (Table 

5

).

Table 5

Clinical trials using mesenchymal stem cell-based therapies for degenerative disc disease

2010, Yoshikawa et al. [13]

Autologous BM-MSCs

105 cells/ml

2

Intradiscal, single

24

Improvement in pain score and rehydration of the disc in both patients

2011, Orozco et al. [12]

Autologous BM-MSCs

23 ± 5 × 106

10

Intradiscal, single

12

Improvement in pain, disability, and disc hydration

2015, Pettine et al. [33]

Autologous BM concentrated cells

121 (±11) × 106

26

Intradiscal, single or double

12

Improvement in pain scores in patients with higher CFU-F concentrations; improvement on MRI (n = 8)

2017, Mesoblast Ltd (unpublished)

Allogeneic MPCs in HA carrier

6 × 106 1.8 × 107

100

Intradiscal, single

36

Improvement in VAS and ODI in 6 million MPCs injected group

Present study

AT-MSCs in HA carrier

2 × 107 (n = 5) 4 × 107 (n = 5)

10

Intradiscal, single

12

Improvement in VAS, ODI, SF-36 (n = 6); improvement of water content on diffusion MRI (n = 3)

Despite these positive results showing that MSCs are promising for degenerative disc repair, low cell survival resulting from the hypoxic and inflammatory host environment, cell leakage leading to osteophyte formation, and identification of the optimal cell source and optimal delivery method still present major challenges to MSC therapy [2, 8, 32, 35]. Thus far, it seems that both autologous and allogeneic MSCs can be implanted safely, but immune rejection remains a significant issue [36]. It has been reported that 90% of the transplanted cells leaked out of the degenerated disc following injection in aqueous solution, but this leakage was reduced to 50% with fibrin glue coadministration [2, 35]. Similarly, our previous preclinical study showed that coadministration with Tissuefill® (HA derivative) could enhance the efficacy of intradiscally injected MSCs [32]. The present study differs from previous clinical studies [12, 13, 33] and the four-arm clinical trial (Mesoblast’s clinical trial) described earlier because our patients were coadministered a HA derivative and autologous AT-MSCs rather than bone marrow-derived cells. AT-MSCs can be obtained with a procedure that is less invasive and presents lower risk than harvesting of BM-MSCs. Furthermore, the autologous AT-MSCs in this study showed increased expression of TβRIII, a potential predictor of chondrogenesis. AT-MSCs expressing TβRIII at higher levels are thought to have higher chondrogenic potential through increased susceptibility to TGF-β3-induced chondrogenesis [34]. Although intradiscal injection can cause further injury, leading to more degeneration and cell leakage, current delivery methods are limited to direct injection into the affected discs [37]. To the best of our knowledge, this is the first clinical trial showing the safety and tolerability of combined implantation of autologous AT-MSCs and a HA derivative in patients with chronic discogenic LBP.

The present study was designed as a phase I clinical trial, for which the calculation of adequate sample size is often difficult. Here, we expected that 10 patients would be sufficient to achieve the trial endpoints because the active sample size in phase I clinical trials has been generally reported to be between six and 10 active subjects [38]. Using this sample size and study design, we were able to obtain the following primary findings: combined implantation of AT-MSCs and the HA derivative is safe and tolerable for treatment of chronic discogenic LBP; additionally, at 6 months following AT-MSC transplantation, seven patients (cases 2, 4, 5, 6, 7, 8, and 9) showed significant reduction of VAS, ODI, and SF-36, whereas at 12 months only six patients (case 2, 4, 5, 7, 8, and 9) showed significant improvement of pain and disability. The treatment success rate was not different between the low-dose (2 × 107 cells/disc) and high-dose (4 × 107 cells/disc) groups, and among the six subjects (cases 2, 4, 5, 7, 8, and 9) who achieved treatment success, three (cases 4, 8, and 9) were determined to have increased water content based on the increased ADC value determined from diffusion MRI. We used the ADC values to assess the water content in the degenerated disc because ADC values have been reported to be a more reliable method of assessing subtle changes in water content compared with T2-weighted imaging and T2 mapping [39] (Table 4).

The exact mechanism by which the combined implantation of AT-MSCs and the HA derivative led to improvement of chronic discogenic LBP in the present study remains unclear. Based on our previous preclinical study, however, we assume that injection of MSCs into the degenerated disc improves ECM production by the degenerated host NP cells, increases NP-like gene expression, and modulates the immunological response of NP cells to inflammatory cytokines. The immunomodulatory effects of MSCs on NP cells within the degenerated disc could potentially inhibit the inflammatory cascade, thereby preventing ingrowth of pain-inducing vasculature and nerve fibers [2, 8, 32]. Additionally, suspension of the cells in HA derivative for coadministration may prevent cell leakage, reduce the risk of uncontrolled differentiation of MSCs into osteoblasts, and enhance cell attachment and cell survival [32].

We also evaluated the potential causes of treatment failure (pain reduction 

3

). Initially, case 3 complained of LBP and pain in both buttocks. Initial lumbar X-ray imaging and MRI showed degenerative spondylolisthesis (forward displacement of L4 on L5), facet joint arthritis, and spinal stenosis at the L4/L5 level; flexion and extension radiography showed no lumbar instability (Fig. 

4

). The patient had no neurogenic intermittent claudication or radiculopathy, which was supported by electromyography and a nerve conduction study. The pain did not improve even after medial branch nerve blocks, and the patient was included in this study based on discographic findings. Case 6 was obese (38 kg/m

2

) and had a disc protruding to the left side at the L5/S1 level, with a disc height reduction of approximately 21.4%. Case 10 had depressive symptoms (Beck Depression Inventory initial score 14), which might have resulted in treatment failure. Although all patients were enrolled after lumbar medial branch block and discography to rule out other causes of LBP, the LBP was still not successfully eliminated in cases 6 and 10, possibly because of other potential confounding factors for chronic LBP such as obesity (case 6) and depression (case 10). In addition, other structural etiologies for chronic LBP that could have prevented treatment success here may include spondylolithesis (case 3), spinal stenosis (case 3), facet joint arthritis (case 3), decreased disc height (case 6), and disc herniation (case 6) (Fig. 

4

).

Fig. 4

Possible causes of treatment failures in cases 3 and 6. Lumbar lateral X-ray imaging (a), T2 sagittal MRI (b), and T2 axial MRI (c) of case 3 showing degenerative spondylolisthesis of L4 on L5 (slippage or displacement of L4 vertebra compared to L5 vertebra) and spinal stenosis. T2 sagittal MRI (d) and T2 axial MRI (e) of case 9 revealed left-sided L4–5 herniated NP with decreased disc height

Thus, careful patient selection is essential for achieving therapeutic success in stem cell therapy for chronic discogenic LBP because MSC-based therapies should not be thought of as a cure-all for spinal pain [40]. Given the multifactorial nature of chronic LBP, it is difficult to isolate patients in which IVD degeneration is the only contributor to LBP without a larger-scale study, but the current study supports the safety and tolerability of such a clinical trial. Moderate IVD degeneration has been considered an ideal target for stem cell therapy. Patients with nondiscogenic LBP and those with advanced disc degeneration or severe annular compromise may not be ideal candidates for this therapy [40]. Furthermore, it is necessary to consider the ideal cell dose to administer. The present study showed no significant difference between the low cell dose and high cell dose, suggesting that an even lower cell dose may still be efficacious. A low cell dose might actually be beneficial due to the poor nutritional supply of the NP, which could result an inability to support higher numbers of transplanted cells. In such a case, a high cell dose that exceeds the limit for donor cell viability in the NP environment could lead to deleterious accumulation of dead cells and waste products [2, 8]. Other important considerations could be the duration of cell survival after implantation and the degradability of the scaffold.

Our study is a single-arm, open-label, phase I pilot study, and thus caution should be applied when drawing any conclusions regarding long-term safety and efficacy. Although there was no placebo group in this pilot study, a placebo effect is unlikely because coadministration of AT-MSCs and HA derivative provided significant improvement of patient outcomes (VAS/ODI and MRI improvement), and both the patients and attending physician were blinded to the treatments that each patient received. Large-scale clinical trials assessing the optimal cell source, cell dose, scaffold, and relevant clinical endpoints are needed to define the true pathology that will benefit from stem cell therapy and the appropriate therapeutic regimen. However, we propose that coadministration of AT-MSCs and the HA derivative may provide a safe and tolerable treatment for chronic discogenic LBP resulting from moderate IVD degeneration.