Exposure to a sudden increase in inflammatory cytokines TNF-α and TGF-β changes barrier function at interfaces between the vascular system and the respective tissue compartments making up the osteoarthritic knee joint. Based on the results of this multimodal, multiscale and correlative imaging study, TGF-β exposure diminishes barrier function significantly more compared to TNF-α exposure, which may reflect differences in dosage or respective differences in the effects of pro- and anti-inflammatory cytokines on barrier function. These data implicate the capacity of acute events, such as trauma and cytokine storms induced by viral disease, to significantly impair barrier function between tissues and thereby molecular trafficking, even in osteoarthritic synovial joints which would be exposed to chronic inflammation. Further, the study implicated cytokine induced changes to pericellular and extracellular transport of the large molecular tracer (70 kDa) that mimics transport of albumin (67 kDa), the main carrier protein in the blood plasma. Exposure to both cytokines appeared to increase transport of the large molecular weight tracer to the pericellular region of chondrocytes (the chondron), which may offer a means to open transport pathways for concurrent delivery of pharmaceutical agents that improve chondrocyte viability as well as matrix regeneration by chondrocytes to enable cartilage repair in OA patients.24 As a whole, the study provides new insights into pathways of transport between the cardiovascular system and the cellular inhabitants of the musculoskeletal system. From a methodological and mechanistic perspective, the study underscores the power of correlative imaging to enable an integrated understanding of complex biosystems and their cellular inhabitants.
The multimodal, correlative imaging across length scales enabled seamless linking of loss in structure and function in OA, from subcellular to organ system length scales. This was particularly apparent with respect to transport of the fluorescent tracer that tracked large molecule traffic from the heart to the cells populating the respective tissues of the synovial joint (Fig. 3–5); akin to tracking truck traffic in road transport networks, future imaging modalities may enable real time assessment of traffic dynamics which would be invaluable for understanding mechanisms of pathophysiology, novel drug delivery strategies, and new preventative health measures.16,25−28 The current study highlights the near instantaneous nature of the circulation to the joint, with the dextran perfusing all musculoskeletal tissues within the five minute circulation period of the study. The integrated, correlative and multimodal imaging approach also remarkably enabled connection of articular surface degeneration at smallest length scales to mesoscopic biomechanical considerations of joint structure and function, where compositional changes of the surface observed with electron microscopy could be tied to the metaphysis of the joint, where stress concentrations would occur during physiological loading (Fig. 4,5A). All knee joints studied showed cartilage surface fibrillation, marrow space openings into the articular cartilage, and vascular invasion into the articular cartilage, with no tidemark duplication at the osteochondral interface; these all represent morphological features characteristic of early to mild osteoarthritis.27,29
A limitation and strength of the study is that the Dunkin-Hartley guinea pigs naturally develop osteoarthritis with age, which precludes study of age-matched controls without OA yet provides a relevant model for the human condition, where onset of OA correlates with age.15 Use of younger animals prior to onset of OA would not provide an appropriate, independent control, since our previous studies using the same animal model, including a younger cohort, demonstrate an age as well as a disease state dependence on distribution and intensity of fluorescence tracers.15
Another recent study from our group, implementing episcopic cryoimaging, showed that exposure to an acute spike of inflammatory cytokine TGF-β significantly reduced vascular barrier function in bone and muscle compared to age-matched controls.1 This observation provided the impetus for the current study, where plastic embedding and multimodal, higher resolution imaging modalities allowed for quantitative assessment of barrier function in different tissue compartments of the knee joint. For that reason the study design compared tracer transport in two age matched cohorts of the same species with similar stages of OA, where one group was administered a pro-inflammatory cytokine TNF-α and the other was administered an anti-inflammatory cytokine TGF-β.
The study design emulates those carried out to probe the blood-brain barrier and other tissue barrier functions in human patients and animal models.19–22,30−32 For example, an increase in vascular albumin permeability was reported with exogenously administered TNF-α in in vivo mouse model studies of lung, liver, kidney, brain, sciatic nerve, retina, duodenum, jejunum, cecum, thoracic aorta, skin and diaphragm.33 In contrast to the current acute study observing changes 5 minutes after TNF-α delivery, significant increases were not observed in the mouse study until 12–18 hours after delivery of TNF-α. Furthermore, the published mouse study administered TNF-α at 20 ng/g,33 a greater than thirty-fold increase in dose compared to the current study. These differences may also reflect intrinsic differences between the healthy murine model and the natural model of OA in the guinea pig. A limitation of the current study is that it was not possible to carry out a dose-response analysis of each cytokine studied due to resource contraints, both with regard to the animal model as well as access to chemical agents (cytokines). The current study, however, provides new insights to guide future investigations and published reports of serum cytokine levels in healthy and diseased, as well as young and aged human cohorts,1,3,12−14 are also increasing; these will also provide new foundations for future studies.
We observed compartmentalization of molecular transport within the joint similar to that reported in previous studies, and the significant changes in barrier function with exposure to TGF-β confirm those reported in previous cryo-imaging studies,1,15 irrespective of intrinsic differences in the technologies used for fluorescent imaging of episcopic specimens and laser scanning confocal microscopy (LSCM) imaging. LSCM in general allows for more precise and powerful excitation of fluorophores and capture of emitted light in targeted wavelengths. Further, differences between cryo- and chemically fixed tissues are expected to be minimal, since the cryofixed specimens are fixed and frozen using liquid nitrogen and OCT compared to the chemical fixation of lysine-fixable dextrans for multimodal imaging. Cryofixation occurs at such low temperatures that any potential diffusive processes are expected to be slowed, and chemical fixation of the “fixable dextrans” containing a primary amine chemically fixes the dextrans in place.34–36
Previous experiments have demonstrated the molecular sieving properties of bone where diffusivity decreases with increasing molecular size; at 70 kDa, or circa 7 nm Stokes diameter, molecules are no longer able to pass through the bone lacunocanalicular system (LCS).2,17,37,38 Molecules of molecular weight in the range of TNF-α (17 kDa) and TGF-β (25 kDa) have been reported to permeate the lacunocanalicular system and may modulate barrier properties within the LCS. Here we observed the 70 kDa molecule in the LCS which contradicts previously reports that molecules of this size are excluded from the LCS, even with the aid of mechanical stimuli and associated convective transport.2,17,39 The circulation time of the tracer in the current study was chosen to match that of previous studies tracing transport from the vascular system to the osteocytes in healthy and acute fracture rodent models; sample sizes also matched those of previous studies.2,17,18,39,40 Taken together, the current study suggests that the chronic OA disease state, combined with acute exposure to inflammatory cytokines TNF-α and TGF-β, increases LCS permeability and associated transport to osteocytes.
Similarly, the permeability of large molecules in cartilage has been shown to be largely dependent on cartilage composition, which itself depends on health or disease state.41 It has been demonstrated that changes brought on by OA, i.e. cartilage degradation, also alter tissue structure and composition.42 In the current study we observed a small amount of tracer within the articular cartilage, with elevated levels of 70 kDa tracer observed inwards toward the articular surface. This suggests that a greater permeability to macromolecules in the cartilage in the middle and deep zones is accompanied by decreased retention in the tissue. This is consistent with findings that different sized antibodies, e.g. 25 kDa, 50 kDa, 150 kDa and 200 kDa in size, diffuse into the cartilage with greatest concentrations observed at the articular surface.43 The diffusion characteristics of 70 kDa dextrans in cartilage PCM have been shown to change with disease state in a porcine model, where healthy animals exhibit significant differences in PCM and ECM diffusion coefficients, and OA animals do not.44 This is also consistent with our findings that the differences in tracer levels in ECM and PCM were not significant in the guinea pigs which spontaneously develop OA with age. However, the accumulations in the chondron in TGF-β animals but not TNF-α animals suggest that TGF-β may differentially change permeability characteristics of articular cartilage, providing impetus for further studies.
A comparison of tissue compartments between and across the TNF-α and TGF-β groups, did not demonstrate significant differences in bulk permeability of respective tissue compartments. However, the data did show that there was a significant modulation of transport between the bone microvasculature and other tissue compartments in the femur, in the respective groups. Although these cytokines may not necessarily modify the tight junctions at tissue compartment boundaries, they do alter the function of the microvasculature within the joint. TGF-β has been shown to be vital to endothelial barrier function and thus maintenance of microvasculature, with an increase in TGF-β being implicated with the breakdown of the functional tissue barrier at the blood-retinal boundary.30,31 Similarly, TNF-α has been shown induce increased vascular permeability through endothelial dysfunction in the brain and lung.31,32 There is mounting evidence of the association of vascular pathology and OA pathogenesis.45,46 TNF-α and TGF-β were shown to have an immediate effect on vasculature, and it is hypothesized that cycles of intermittent increase in pro-inflammatory cytokines associated with age-related chronic inflammation may cause further deterioration of the functional barrier at vascular interfaces and may be a contributing factor to OA pathogenesis.
The investigation into transport of macromolecules to and between the compartments of the joint, via the circulatory system through to interstitial transport requires imaging modalities and specimen preparation which allow visualization across multiple length scales, from molecular, cellular, tissue and to organ, e.g. such as the methodologies described here. The current study was not without limitation, as imaging in a single plane (comprising thousands of images) and at a single timepoint captures "a snapshot" during a dynamic process. An ideal imaging modality would also allow for temporal assessment in real time. In vivo imaging, albeit currently available at much lower resolution, may provide an additional degree of temporal fidelity; imaging modalities and respective computational capacity are constantly increasing in resolution and speed.
Furthermore, while the molecular weight of the dextran is comparable to that of albumin, dextrans differ both in shape and molecular charge to globular proteins; these factors can affect the mobility of large molecules within tissues. Also, although we observed changes in barrier function between the two groups, in this present study it is not possible to separate the immediate effects of inflammatory cytokines and the chronic effects of exposure to inflammatory cytokines in osteoarthritic, age matched subjects.