ARTHRITIS & RHEUMATISM Vol. 39, No. 6, June 1996, pp 1071-1076 0 1996. American College of Rheumatology 1071 LETTERS More on vasodilatation, joint swelling, and nitric oxide To the Editor: Drs. Evans and Stefanovic-Racic (1) kindly comment on our views (2) that (i) high synovial fluid (SF) hydrostatic pressures (Psf) require locally raised mean capillary pressure (P,.,) that (ii) may depend on arteriolar vasodilatation and (iii) contribute to swelling. Vasodilatation was suspected (iv) to be driven by cartilage metabolic demands, and possibly (v) mediated by chondrocytic nitric oxide (NO). Their comments call for clarification. (i) Capillary pressure. Fluid flow across endothelia depends on Starling pressures (3-5). Normal escape of fluid into S F (direction +, resorption -) depends on a positive net pressure (CP), the difference between a positive hydrostatic pressure difference (Pcnp - PSp)and a net back-sucking (-) colloid osmotic pressure difference (COP,,,, - COP,,) of varying efficiency (100-0% = a reflection coefficient (r of 1-0). The normally negative P,, “sucks“ [ -( -) = t] into SF. Thus, Equation 1: z p = (Pmp ~ Pr,) - (COPl,,“,- COP,,) In the near-normal example, simplified for didactic reasons, T P is + 7 (unit cm H,O). XP = [20 - 7 = 25 ( - 5)] - 0.9 (35 - 15) - I8 If the effective COP difference drops to 0 (COP,,,, = COP,,, or (r = 0 at m permeability), both the CP (by 18) and the flow into SF increase. Due to limited joint compliance, the effusion gradually raises the P,,, decreasing the CP. If the CP returns to 7, this decrease in the COP difference may raise the P,, to 13 cm H,O (-5 + 18). A negative COP difference (COP,,,, < COP,,), a lower CP, and a higher PC:,,, permit a higher Psp A CP of 0 does not induce swelling (equation 2). At negative CP, S F is resorbed. In 1940 Bauer et al (6) proposed that rheumatoid S F “colloids’’ raise COP,,, as these “colloids” do (7), sucking fluid into SF. They never considered hydrostatic P (equation I not yet formulated). The failure to grasp that high P,, requires elevated PcaP explains how the proposal (6,7), in otherwise excellent studies, could be found misleading (8,9). This interlude does not weaken the conclusion that in a steady state (normal or swelling), Pcap n u s t exceed (or equal) P,, + (r(COP,,,, - COP,,), a sum of negative terms (equation I), which opposes flow into S F (see Knox et al as referenced in 5). In >50% of the traumatic joints in our study (2). the “opposing” sum exceeded Pcap averages, and the effusions would already have been resorbed unless the Pcapwas elevated to give a positive CP. (ii) Arteriolar dilatation. PJnduced venous tamponade may decrease blood flow and raise Pcap.The former can be compensated for by arteriolar dilatation, raising the Pcap to levels “required” (item i) or higher (2). This (and item i) agrees with Levick’s statement (3), that “A fall in precapillary resistance and a possible rise in postcapillary resistance readily explain . . . that intra-articular pressures frequently exceed normal capillary pressure in acute rheumatoid effusions.” (iii) SF volume. Transendothelial flow is a product of CP (+ or -1 equation 1) and of endothelial hydraulic conductance (L,) (high in fenestrated capillaries and increased by “pore” enlargement) and surface area (A) (ref. 3, cf. ref. 4). Excess S F is removed by the lymphatics (Jv,,ymph). Thus, an increase or decrease in S F volume (V,,) per unit of time (dV,,/dt) is reflected (cf. ref. 3) by Equation 2: A n increase (items i and ii) in [CP] increases the formation of SF (3). The effect on SF accumulation (dVJdt+) is stronger if permeability (Ll, t , (r J, ) and A are increased or if Jv~,ym,,,lis decreased. In heavy exercise, muscle blood flow may increase >20-fold, which is mainly due to an increase in arteriolar conductance (correlates with the fourth power of the lumen radius) (10). Fenestral flow “jets,“ increasing COP differences ( 5 ) . and increase of Jv,,ymp,,, etc., may well prevent short exercises that result in a 7-fold blood flow increase (1) (considerably less for Pcap in equation 2) from inducing clinical swelling. (iv) Mechanism of arteriolar dilatation (item ii). We suspected metabolic vasodilatation. Finding joint cartilage avascular, Hunter (1 1) described synovial vessels: “From these (large branches) again there arises a Crop of small short Twigs, that shoot towards the outer Surface; and whether they serve for nourishing oizfv [our italics] or if they pour out a dewy Fluid, I shall not pretend to determine. However that be, I cannot help observing, that the Distribution of the Blood-vessels to the articulating Cartilages . . . seems calculated for obviating great Inconveniences.” This fits our “capillaries of cartilage in the synovium” (see ref. 6 in ref. 2). Cartilage gets energy from glycolysis. It is true that Bywaters found O? uqtake to be very small (“if there is any at all”) in slices (ret. 8 in ref. 1) soon considered to represent “senile equine cartilage” (12): it was higher in younger cattle. In 1940, this Philadelphia group (12) found that low glucose further enhances 0, uptake, a Crabtree effect that was confirmed later (ref. 9 in ref. 2). If lactate release (-3.54 pmoles/l2 hours + 295 nmoles/hour) and 0, uptake (-3 nnioles/minute + 180 nmoles/hour) rates in 10’ control chondrocytes (ref. 9 in ref. 1) are converted (+) to the same unit, ATP stoichiometry does not support the belief that oxidation is unimportant (1). If metabolic (-low glucose) “acidotic hypoxia” (Crabtree) explains the links between SF lactate and 0, better than “hypoxic acidosis’’ (Pasteur), the lack of effect of cyanide on glycolysis (ref. 9 in ref. 1) cannot support Evans and Stefanovic-Racic’s belief. They note (ref. 9 in ref. 1) that a low energy supply decreases cartilage synthesis of macromolecules, the breakdown of which is possibly enhanced in acidosis. What they fail to appreciate is that both are opposed by vasodilatation (item ii) and the Crabtree switch. Correlations between SF acidotic hypoxia and elevated P,, (-Vsf) which may lead to reactive hyperemia still fit (2) metabolic vasodilatation (ref. 10 and Renkin in ref. 4). LETTERS 1072 (v) NO. Chondrocytic NO could (1) b e vasodilatatory. NO seemed to fit the role better (2) than “old” mediator metabolites (ref. 10 and Renkin in ref. 4) “diluted” en route to synovial arterioles. Views i-iii are false if I have simplified physiologists’ (3-5) messages incorrectly. It has not been disproved (1) that the vasodilatation (item ii) could be metabolic (item iv), but any other type of arteriolar dilatation may be responsible. Proposal v is false if there was no increase in NO synthesis in those traumatic joints (1) in which P,, exceeded 10-20 cm H,O (item i). Johan Ahlqvist, M D Sibbvik Vastarzfiard, Finland 1. Evans CH, Stefanovic-Racic M: Reply (letter). Arthritis Rheum 38:1530-1531, 1995 2. Ahlqvist J, Osterlund K: Possible role of inducible nitric oxide synthase in articular chondrocytes in the pathogenesis of arthritic swelling (letter). Arthritis Rheum 38:1529-1530, 1995 3. Levick JR: Synovial fluid dynamics: the regulation of volume and pressure. In, Studies in Joint Disease. Edited by A Maroudas, EJ Holborow. London, Pitman Books, 1983 4. Renkin EM, Michel CC, editors: Microcirculation. In, Handbook of Physiology. Vol. IV. Cardiovascular System. Bethesda, MD, American Physiological Society, 1984 5. Levick JR: An analysis of the interaction between interstitial plasma protein, interstitial flow. and fenestral filtration and its application to synovium. Microvasc Res 47:90-125, 1994 6. Bauer W, Ropes MW, Waine H: The physiology of articular structures. Physiol Rev 20:272-312, 1940 7. Makinen (Miikisara) P: Synovial fluid in rheumatoid arthritis. Ann Med Exp Biol Fenn 36 (suppl 7):l-70, 1958 8. Lipson RL, Baldes El, Anderson JA, Polley HF: Osmotic pressure gradients and joint effusions. Arthritis Rheum 8:29-37, 1965 9. Palmer DG, Myers DB: Some observations of joint effusions. Arthritis Rheum 11:745-755, 1968 10. Guyton AC: Textbook of Medical Physiology. Philadelphia, WB Saunders, 1981 11. Hunter W: On the structure and diseases of articular cartilages. Philos Trans R SOC42514-521, 1743 12. Rosenthal 0. Bowie MA, Wagoner G: Studies in the metabolism of articular cartilage. J Cell Comp Physiol 17:221-233, 1941 Fibroblast adhesion to articular cartilage To the Editor: We read with interest the recent article by McCurdy et al (l),in which they show enhancement of rheumatoid synovial fibroblast adhesion to articular cartilage exposed to neutrophil proteases. In our recent work (2), exposure of cartilage to neutrophil proteases did not result in increased adhesion of fibroblasts unless: 1) the fibroblasts were previously treated with collagenase to free the membrane-bound collagen receptors from collagen synthesized by these cells, thus allowing adhesion to the cartilage surface collagen; or 2) the damaged cartilage surface was replenished with fibronectin, which in the case of McCurdy et al, it was provided by the fetal calf serum used for incubation with the cartilage prior to the adhesion assay. It is also unclear in their work if the results reported represent cell adhesion to the artifactually produced cut surface of the cartilage pieces or to the true articular surface. We would also like to point out that our previous work on the effects of neutrophil proteases unmasking collagen epitopes at the cartilage surface (3) was misquoted by McCurdy et al. We used EDTA, and not EGTA, in an unsuccessful attempt to inhibit the attack by neutrophil granules. The failure to show a role for metalloproteinases in damage to the articular surface was supported by the negative results obtained with cartilage explants stimulated in vitro with cytokines. In those experiments, in spite of evidence of matrix depletion presumably due to metalloproteinase secretion by activated chondrocytes, there was no increase in collagen antibody binding to the surface. Hugo E. Jasin, M D University of Arkansas for Medical Sciences Little Rock, A R 1. McCurdy L, Chatham WW, Blackburn WD Jr: Rheumatoid synovial fibroblast adhesion to human articular cartilage: enhancement by neutrophil proteases. Arthritis Rheum 38:1694-1700, 1995 2. Noyori K, Jasin HE: Inhibition of human fibroblast adhesion by cartilage surface proteoglycans. Arthritis Rheum 37:1656-1663, 1994 3. Jasin HE, Taurog JD: Mechanisms of disruption of the articular cartilage surface in inflammation: neutrophil elastase increases availability of collagen I1 epitopes for binding with antibodies on the surface of articular cartilage. J Clin Invest 87:1531-1536, 1991 To the Editor: We appreciate Dr. Jasin’s interest in our recent article and recognize his previous work in the area of fibroblast adhesion. As Dr. Jasin has demonstrated, using human skin fibroblasts and bovine cartilage, pretreatment with collagenase and addition of fibronectin is necessary for enhanced fibroblast adhesion to cartilage (1). In our system, using human synovial fibroblasts and human cartilage, we did not need to pretreat the cells with collagenase (2). As Dr. Jasin suggested, fibronectin may have played a role it our results, but was not explored in our study. Regarding adhesion to “native” versus cut articular surface, we were very careful in our studies to use thin discs (1-2 mm thickness) of cartilage, which were studied with the cut surface down. Therefore, we do not think that adhesion to the cut surface was responsible for our results. Dr. Jasin’s observation that the addition of cytokines to cartilage explants did not result in articular surface damage would seem to be more relevant to chondrocyte-mediated tissue resorption. Our studies involve the effects of neutrophils, which as we have previously demonstrated, do degrade surface cartilage by a mechanism that, in part, is mediated by collagenase (3).