The association between retraction of the torn rotator cuff and increasing expression of hypoxia inducible factor 1α and vascular endothelial growth factor expression: an immunohistological study

Rotator cuff injury affects both the tendons and the associated muscles. Full thickness tears may result in muscle and tendon retraction, leading to a change in the angulation of muscle fibers and subsequent fatty infiltration (FI) [1]. FI was first observed by computed tomography (CT), which led to the five- stage1 classification system that is based upon fat-to-muscle ratio, one of the most important prognostic indicators after rotator cuff repair [1]. Furthermore, rotator cuff muscle atrophy (MA) is important in assessing treatment of rotator cuff tears [2‐5]. Although these radiological parameters are good indicators of tendon degeneration, the histopathology of the degenerating rotator cuff are not as clear. Benson et al. proposed that hypoxic changes in the cuff contributes to loss of cells by apoptosis, which may be responsible for the development of tears [6]. These authors found different levels of hypoxia inducible factor 1α (HIF) in the damaged rotator cuff [7]. HIF is known to be an important upregulator of vascular endothelial growth factor (VEGF), which stimulates endothelial cells and vessels to invade hypovascularized tissue [8]. Pufe et al. and Peterson et al. found higher concentrations of VEGF in degenerative Achilles tendon as compared with healthy tissue [9, 10]. Although the rotator cuff injury has been attributed to hypovascularity, there is some evidence that neovascularization also occurs [11]. Lewis et al. described increased neovascularity in patients with rotator cuff tendinopathy; increased HIF and VEGF expression in the diseased rotator cuff, but not in tendon retraction [12]. Furthermore, neovascularity has not been definitively proven histologically until this point.

The purpose of this study is to examine HIF and VEGF expression, and neovascularity in torn rotator cuff tissue from patients with medium size (3-5 cm, formally Bateman stage 3) full-thickness tears and tendon retraction, compared with healthy control samples. A second goal is to find an association between FI, MA and the amount of VEGF and HIF expression and neovascularity within both torn rotator cuffs.

We aimed to clarity the relationship between VEGF/HIF expression and neovascularization and tendon retraction in full-thickness rotator cuff injuries. We showed that HIF and VEGF are upregulated in full-thickness rotator cuff tears (highest in group IV, Patte stage1 III), leading to higher vessel density while angiogenetic peptides are not detected in healthy tendon. In the torn rotator cuff, vessel density was lower than in our control group but increased with worsening tendon retraction. Our results confirm the findings of Matthews at al., who reported smaller numbers of vessels in massive tears compared to healthy controls [15]. However, vessel density was not explored in that work; for the first time, we have shown increasing vessel density correlating with retraction. This could be due to high VEGF levels, although we did not find this in our patient population. This may be due to the small number of patients in groups I and II.

While VEGF mediated angiogenesis contributes to the repair and remodeling of degenerated tendons, invasion of endothelial cells may also weaken the mechanical stability of the tendon [8]. We assume that increasing vessel density weakens the tendons and leads to the accretive nature of tendon retraction. This has been supported by the recently published theoretical model of Peers et al, showing that chronic tendon loading causes mechanical trauma with multiple microruptures of tendon microvasculature [16]. These microruptures initiate a VEGF-mediated cascade of vascular remodeling that becomes chronically pathologic. Hypervasularity has been demonstrated in degenerated human and rabbit Achilles tendons, in diseased patellar and long head biceps tendons [17‐20]. Chronic tendon pathology appears to be a highly active process of ongoing neovascularisation [21]. Why this process does not lead to tissue repair, but to pain and chronic disease, is unknown.

Primary reconstruction of massive rotator cuff tears with minimal retraction is currently recommended in clinical practice; more extensive reconstruction with muscle transfers and prostheses are used in cases of moderate to severe retraction [22‐24]. Our findings present histological backing for these guidelines and imply a potential benefit to treating patients with partial thickness tears more aggressively, in order to prevent complications of injury progression.

FI and MA of the torn rotator cuff correlate with the duration and size of the tear [3, 4, 25]. MA and FI levels could be used as outcome predictors for surgically treated injuries [26]. There is accumulating evidence that these parameters correlate closely with tendon degeneration. While age has been previously shown to correlate with FI and MA, we did not confirm this in our study. This may be due to the consistency of age among our patients groups. We demonstrated a correlation between FI and MA and VEGF/HIF expression. Therefore, VEGF and HIF expression as well as neovascularity may be used as a monitoring tool in the clinical setting, for the level of tendon degeneration that occurs during tendon degeneration. Neovascularization has been previously described in massive rotator cuff tears as compared to smaller tears of smaller size [27]. Neovascularization and lipid accumulation may be an indication of enhanced aerobic metabolism in the degenerated muscle. Further histological studies are required to investigate the mechanism of this finding; HIF and VEGF upregulation may help to explain these observations.

HIF, which is induced by cell hypoxia, is one of the most important regulators of VEGF gene expression [28]. The original function of VEGF as a strong inducer of embryonic vasculogenesis was shown in knockout mice [29]. VEGF plays a pivotal role in tumor angiogenesis (e.g. glioblastomas), as well as other disease states (diabetic retinopathy, age-related macular degeneration or rheumatoid arthritis) [28]. While the role of VEGF in tendon degeneration is largely unknown, the process is thought to occur in areas of poor perfusion. Rahburn and Macnab identified a hypovascularized zone in the rotator cuff that coincides with the most common location for tears [30]. In contrast, we found hypervascularisation within our samples. Our findings are supported by studies showing increased Doppler flow in diseased rotator cuffs, likely reflecting a neovascularization process [12, 31].

The etiology of rotator cuff tears is still a matter of debate. Some authors ascribe mechanical damage to subacromial impingement, while others suspect a primary process of tendon degeneration [11]. On the basis of ultrasound findings, Gohlke found that patients with degenerative full-thickness rotator cuff tears and significant retraction were older [32]. In a recent literature review, Nho et al. conclude that hypovascularity is a minimal contributor to cuff tears, which supports our finding that hypervascularisation occurs late in the course of cuff tendinopathy [33]. Since HIF induction was initially described in conditions of decreased oxygen partial pressure, the majority of studies on HIF regulation have addressed hypoxic conditions [34]. However, there is evidence that HIF can also be induced under normal conditions by growth factors, hormones or nitric oxide [35].

Cyclic loading is as important inducer of HIF expression. Benson et al. showed HIF levels to be high in mild impingement, small, medium, and large tears, but reduced in massive tears. We assume that complete tendon rupture reduces mechanical cyclic loading, resulting in reduced HIF expression. If hypoxia underlies HIF expression, this contradicts Matthews's results describing the poorest blood supply in massive tears [15]. Biomechanical studies of Andarawis-Puri and Perry found out that cyclic tendon loading attenuates as retraction increases [36, 37]. In our findings, HIF expression significantly increases with rising Patte grade. Therefore, it remains unclear whether hypoxia or cyclic tendon loading is the underlying reason for HIF and VEGF expression.

Our study has several limitations. Our control group is comparatively small; tumor patients were excluded from the study due to their requirement for prosthesis implantation or upper limb amputation. Therefore, we elected to include trauma patients of matchable age as controls (difficult due to the age discrepancy in this population as compared with the other groups and the reason for our low numbers).

Full-thickness rotator cuff tears may develop into cuff arthropathy and be associated with osteoarthritis of the shoulder [38]. Osteoarthritis was not measured, but may act as a confounder for VEGF and HIF expression. This bias was addressed by limiting our subjects to patients with medium sized tears (3-5 cm; unclassified due to lack of international consensus). Extent of tendon retraction was the only fluctuating parameter.

We demonstrated a correlation between vessel density, VEGF, and HIF expression as well as FI and MA in full-thickness medium-sized rotator cuff tears and tendon retraction. Neovascularization in cuff and tendon tissues is mediated by VEGF. Increasing vessel density accelerates tendon degeneration by weakening its underlying structure. VEGF expression and neovascularity could be used in the clinical setting to monitor tendon degeneration in patients with rotator cuff injuries.