Abstract 4607: A translational study of Tumor Treating Fields in non small cell lung cancer

Background Lung cancer is the second most common cancer and the leading cause of cancer related death in the US. Non-small-cell-lung cancer (NSCLC) accounts for approximately 80% of all cases of lung cancer. Prognosis of NSCLC remains poor despite extensive research. Tumor Treating Fields (TTF, TTFields) are low intensity, intermediate frequency alternating electric fields that have been demonstrated to disrupt cancer cell replication. A phase III clinical trial has previously demonstrated the efficacy of TTF in glioblastoma patients. Objectives This study was designed to test the effects of TTF in NSCLC, both as a single therapy and in combination with chemotherapies, from bench to the bedside. Methods H1299 (Human NSCLC) and LLC1 (murine Lewis lung carcinoma) cells were cultured while being subjected to different frequencies of low intensity (1-2 V/cm) TTF alone or in combination with different chemotherapies (cisplatin, pemetrexed, erlotinib and paclitaxel). An autologous model of C57BL/5 mice bearing LLC1 tumors was used to test the in vivo application of continuous TTF. A phase II clinical trial of TTF in combination with pemetrexed (standard regimen) was performed in 42 stage IIIB and IV NSCLC patients following progression after at least one line of chemotherapy. Treatment was made using 2 pairs of insulated, non-invasive transducer arrays applied to the skin of the thorax. Results In-vitro: The treatment of H1299 and LLC1 cultures with TTF for 72 hours had a significant inhibitory effect on proliferation in both cell lines, compared to control cultures. The optimal frequency for culture growth inhibition was 150 kHz. The combination of optimal frequency-TTF with different concentrations of chemotherapeutic agents led to further decrease in culture growth rate, compared to TTF-naïve cultures. In-vivo: TTF resulted in a significant reduction of approximately 30% in tumor volume and weight in mice bearing NSCLC tumors, as a monotherapy. This therapeutic effect almost doubled in mice that were treated with pemetrexed concomitant to the TTF. No pathological changes could be demonstrated in healthy tissues excised from TTF-treated animals. Clinical: NSCLC human patients with advanced-stage disease were treated with TTF (median of 11.2 hours/day) using the NovoTTF-100L System (NovoCure Ltd.) in combination with pemetrexed. Median progression-free survival was 23 weeks, median overall survival was 13 months and 1-year-survival was 54% - an improvement over pemetrexed-treated historical controls. The only device-related adverse event was mild to moderate contact dermatitis. Conclusions The inhibitory effect of TTF in different NSCLC models was demonstrated both as a single treatment and in combination with chemotherapies. The promising efficacy and low toxicity of TTF therapy combined with pemetrexed in NSCLC patients should lead to further clinical investigation of TTF as a potential treatment for NSCLC patients.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4607. doi:1538-7445.AM2012-4607

Incubator weaning in preterm infants and associated practice variation

OBJECTIVE

To evaluate the relationship of weight of preterm infants when first placed into an open crib with days to full oral feedings, growth velocity and length of stay (LOS), and to identify unwarranted variation in incubator weaning after adjusting for severity indices.

STUDY DESIGN

A retrospective study using the ParadigmHealth neonatal database from 2003 to 2006 reviewed incubator weaning to an open crib in appropriate-for-gestational-age (AGA) infants from 22 to weeks gestation. Primary outcome measurements included days to full oral (PO) feeding, weight gain from open crib to discharge and length of stay. Models were severity adjusted. To understand hospital practice variation, we also used a regression model to estimate the weight at open crib for the top 10 volume hospitals.

RESULT

In all 2908 infants met the inclusion criteria for the study. Their mean weight at open crib was 1850 g. On average every additional 100 g an infant weighed at the open crib was associated with increased time to full PO feeding by 0.8 days, decreased weight gained per day by 1 gram and increased LOS by 0.9 days. For the top 10 volume hospitals, severity variables alone accounted for 9% of the variation in weight at open crib, whereas the hospital in which the baby was treated accounted for an additional 19% of the variation.

CONCLUSION

Even after controlling for severity, significant practice variation exists in weaning to an open crib, leading to potential delays in achieving full-volume oral feeds, decreased growth velocity and prolonged LOS.

Alternating electric fields arrest cell proliferation in animal tumor models and human brain tumors

We have recently shown that low intensity, intermediate frequency, electric fields inhibit by an anti-microtubule mechanism of action, cancerous cell growth in vitro. Using implanted electrodes, these fields were also shown to inhibit the growth of dermal tumors in mice. The present study extends these findings to additional cell lines [human breast carcinoma; MDA-MB-231, and human non-small-cell lung carcinoma (H1299)] and to animal tumor models (intradermal B16F1 melanoma and intracranial F-98 glioma) using external insulated electrodes. These findings led to the initiation of a pilot clinical trial of the effects of TTFields in 10 patients with recurrent glioblastoma (GBM). Median time to disease progression in these patients was 26.1 weeks and median overall survival was 62.2 weeks. These time to disease progression and OS values are more than double the reported medians of historical control patients. No device-related serious adverse events were seen after >70 months of cumulative treatment in all of the patients. The only device-related side effect seen was a mild to moderate contact dermatitis beneath the field delivering electrodes. We conclude that TTFields are a safe and effective new treatment modality which effectively slows down tumor growth in vitro, in vivo and, as demonstrated here, in human cancer patients.

Disruption of cancer cell replication by alternating electric fields

Low-intensity, intermediate-frequency (100-300 kHz), alternating electric fields, delivered by means of insulated electrodes, were found to have a profound inhibitory effect on the growth rate of a variety of human and rodent tumor cell lines (Patricia C, U-118, U-87, H-1299, MDA231, PC3, B16F1, F-98, C-6, RG2, and CT-26) and malignant tumors in animals. This effect, shown to be nonthermal, selectively affects dividing cells while quiescent cells are left intact. These fields act in two modes: arrest of cell proliferation and destruction of cells while undergoing division. Both effects are demonstrated when such fields are applied for 24 h to cells undergoing mitosis that is oriented roughly along the field direction. The first mode of action is manifested by interference with the proper formation of the mitotic spindle, whereas the second results in rapid disintegration of the dividing cells. Both effects, which are frequency dependent, are consistent with the computed directional forces exerted by these specific fields on charges and dipoles within the dividing cells. In vivo treatment of tumors in C57BL/6 and BALB/c mice (B16F1 and CT-26 syngeneic tumor models, respectively), resulted in significant slowing of tumor growth and extensive destruction of tumor cells within 3-6 days. These findings demonstrate the potential applicability of the described electric fields as a novel therapeutic modality for malignant tumors.

Diffusion and partition of solutes in cartilage under static load

We describe experimental apparatus, methodology and mathematical algorithms to measure diffusion and partition for typical small ionic solutes and inulin (a medium size solute) in statically loaded cartilage. The partition coefficient based on tissue water (K(H(2)O)) of Na(+) increased from 1.8 to 4.5 and for SO(4)(-2) decreased from 0.5 to 0.1, when the applied pressure was raised from zero to 22 atm K(H(2)O) of inulin decreased from 0.3 to 0.05, for an increase in pressure from zero to 11 atm. Our theoretical interpretation of the results is that the partition coefficient can be expressed as a function of fixed charge density (FCD) for both loaded and unloaded cartilage. The partition coefficient shows good agreement with the ideal Gibbs-Donnan equilibrium, particularly when FCD is based on extrafibrillar water (EFW). The diffusion coefficients, D also decreased with an increase in applied pressure; raising the pressure from 0 to 22 atm resulted in the following changes in the values of D: for Na(+) from 2.86 x 10(-6) to 1.51 x 10(-6) cm(2)/s, for SO(4)(-2) from 1.58 x 10(-6) to 7.5 x 10(-7) cm(2)/s, for leucine from 1.69 x 10(-6) to 8.30 x 10(-7) cm(2)/s and for inulin from 1.80 x 10(-7) to 3.30 x 10(-8) cm(2)/s. For the three small solutes (two charged and one neutral) the diffusion coefficient D is highly correlated with the fraction of fluid volume in the tissue. These experimental results show good agreement with the simple model of Mackie and Meares: hence solute charge does not affect the diffusion of small solutes under load. For inulin D & K show some agreement with a modified Ogston model based on two major components, viz., glycosaminoglycans (GAG) and core protein. We conclude that the changes in the partition and diffusion coefficients of small and medium size solutes in statically loaded cartilage can be interpreted as being due to the reduction in hydration and increase in FCD. The change in the latter affects the partition of small ionic solutes and the partition and diffusion of larger molecules. Our results throw light on the ionic environment of chondrocytes in loaded cartilage as well as on the transport of solutes through the matrix.

Depth-dependent compressive properties of normal aged human femoral head articular cartilage: relationship to fixed charge density

OBJECTIVES

Determine the depth-varying confined and osmotic compression moduli of normal human articular cartilage from the femoral head, and test whether these moduli are dependent on fixed charge density.

METHODS AND RESULTS

Using an automated instrument to allow epifluorescence microscopy analysis during confined compression testing on cartilage samples, the equilibrium confined compression modulus (H(A 0)) was found to vary markedly with depth (z=0-1500 microm) from the articular surface. H(A 0) increased from 1.16+/-0.20 MPa in the superficial (0-125 microm) layer to 7.75+/-1.45 MPa in the deepest (1250-1500 microm) layer tested, and was fit by the expression, H(A 0)(z) [MPa]=1.44 exp(0.0012.z [microm]). Also, in successive slices of cartilage extending from the articular surface to the middle-deep regions, the bulk modulus (K(0)) and fixed charge density (FCD) increased, consistent with previous findings. While H(A 0), K(0), and FCD each varied with depth from the articular surface, the dependence of H(A 0) and K(0) on depth did not appear to be completely related to variations in FCD.

CONCLUSIONS

The confined compression modulus of normal aged human femoral head articular cartilage increases markedly with depth from the articular surface, a trend similar to that observed for articular cartilage from other joints in animals but with an absolute amplitude that is several-fold higher. The compressive properties were not simply related to FCD at different depths from the articular surface, suggesting that other as yet undefined factors also contribute to compressive properties.

Partition and diffusion of sodium and chloride ions in soft charged foam: the effect of external salt concentration and mechanical deformation

The partition and diffusion characteristics of an acrylic acid/acrylamide hydrogel, copolymerized in the pores of a polyurethane foam with sodium and chloride ions, were studied by radiochemical methodologies. The hydrogel foam swells by 51%, 80%% and 260% relative to its raw state under bath salt concentrations of 2.0, 1.0, and 0.15 M, respectively. The corresponding partition coefficients are 1.13, 1.29, and 1.99 for sodium (Na+) and 0.89, 0. 85, and 0.65 for chloride (Cl-). The diffusion coefficients are independent of bath concentration and increase linearly with hydration towards their values in water. Deformation affects partition and diffusion solely by dilatation, which determines the swelling and hydration. Comparison of the hydrogel foam with cartilage and intervertebral disc shows considerable similarities and suggests that the same mechanisms control their function.

The osmotic pressure of chondroitin sulphate solutions: experimental measurements and theoretical analysis

We used equilibrium dialysis to measure the osmotic pressure of chondroitin sulphate (CS) solutions as a function of their concentration and fixed charge density (FCD) and the ionic strength and composition of the solution. Osmotic pressure varied nonlinearly with the concentration of chondroitin sulphate and in 0.15 M NaCl at FCDs typical of uncompressed cartilage (approximately 0.4 mmol/g extrafibrillar H2O) was approximately 3 atmospheres. Osmotic pressure fell by 60% as solution ionic strength increased up to about 1 M, but remained relatively constant at higher ionic strengths. The ratio of Ca2+ to Na+ in the medium was a minor determinant of osmotic pressure. The data are compared with a theoretical model of the electrostatic contribution to osmotic pressure calculated from the Poisson-Boltzmann equation using a rod-in-cell model for CS. The effective radius of the polyelectrolyte rod is taken as a free parameter. The model qualitatively reproduces the non-linear concentration dependence, but underestimates the osmotic pressure by an amount that is independent of ionic strength. This difference, presumably arising from oncotic and entropic effects, is approximately 1/3 of the total osmotic pressure at physiological polymer concentrations and ionic strength.

Effect of allopurinol on NMDA receptor modification following recurrent asphyxia in newborn piglets

The present study tests the hypothesis that repeated episodes of asphyxia will lead to alterations in the characteristics of the N-methyl-d-aspartate (NMDA) receptor in the brain cell membrane of newborn piglets and that pre-treatment with allopurinol, a xanthine oxidase inhibitor, will prevent these modifications. Eighteen newborn piglets were studied. Six untreated and six allopurinol treated animals were subjected to eight asphyxial episodes and compared to six normoxic, normocapneic controls. Brain cell membrane Na+,K+-ATPase activity was determined to assess membrane function. Na+,K+-ATPase activity was decreased from control following asphyxia in both the untreated and treated animals (47.7+/-3.2 vs. 43.0+/-2.2 and 41.0+/-5.3 micromol Pi/mg protein/h, p<0.05, respectively). 3H-MK-801 binding studies were performed to measure NMDA receptor binding characteristics. The receptor density (Bmax) in the untreated asphyxia group was decreased compared to control animals (0.80+/-0.11 vs. 1.13+/-0.33, p<0.05); furthermore, the dissociation constant (Kd) was also decreased (3.8+/-0.7 vs. 9.2+/-2.2, p<0.05), indicating an increase in receptor affinity. In contrast, Bmax in the allopurinol treated asphyxia group was similar to control (1. 06+/-0.37); and Kd was higher (lower affinity) than in the untreated group (6.5+/-1.4, p<0.05). The data indicate that recurrent asphyxial episodes lead to alterations in NMDA receptor characteristics; and that despite cell membrane dysfunction as seen by a decrease in Na+,K+-ATPase activity, allopurinol prevents modification of NMDA receptor-ion channel binding characteristics induced by repeated episodes of asphyxia.

Mechanical properties of the collagen network in human articular cartilage as measured by osmotic stress technique

We have used an isotropic osmotic stress technique to assess the swelling pressures of human articular cartilage over a wide range of hydrations in order to determine from these measurements, for the first time, the tensile stress in the collagen network, Pc, as a function of hydration. Osmotic stress was applied by means of calibrated solutions of polyethylene glycol. Calculations of osmotic stress were based on the balance, at equilibrium, between the applied stress, the collagen stress, and the proteoglycan osmotic pressure, piPG, acting within the extrafibrillar matrix compartment. Pc vs hydration was determined for several normal human samples, both native and trypsin-treated, and for cartilage from one osteoarthritic (OA) joint. We found that for normal cartilage the collagen network does not become "limp" until the volume of cartilage has decreased by 20-25% of its initial value and that its contribution to the balance of forces in cartilage therefore must be taken into account over a much wider range of hydrations than was previously thought. For normal cartilage, the Pc vs hydration curves exhibit a steep increase with increasing hydration; trypsin treatment does not change their slope, showing that PG concentration does not influence the inherent stiffness of the collagen network. By contrast, the curves for OA specimens are considerably shallower and displaced to higher hydrations. Our findings thus highlight the role of the stiffness of the collagen network in limiting hydration in normal cartilage and ensuring a high PG concentration in the matrix, which is essential for effective load-bearing and is lost in OA.

Aggrecan turnover in human articular cartilage: use of aspartic acid racemization as a marker of molecular age

Aggrecan is a key component of the cartilage matrix. During aging, many changes occur in its composition and structure; in particular, there is an increase in the proportion of lower molecular weight monomers and of the "free" binding region. An important question has been whether these changes represent alterations in biosynthesis or whether they are due to the accumulation with age of the partially degraded fragments of the originally synthesized large monomer. In the present work we have used an independent tool, viz., the extent of racemization of aspartic acid to study the molecular "age" of different buoyant density fractions of the aggrecan of human articular cartilage, as well as of isolated free binding region and link protein. By measuring the D/LAsp ratio of the different aggrecan species, we were able to establish directly the relative residence times of these molecules in the cartilage matrix and, in combination with compositional and structural analyses, to define their "history" and calculate some of the kinetics constants characterizing their turnover. The value of the turnover constant for the large monomer in fraction A1D1 is 0.206 per year, which corresponds to a half-life of 3.4 years, while the turnover constant for the free binding region is 0.027 per year, which corresponds to a half-life of 25 years. It is thus clear that the rate of formation and turnover of the large monomer is much more rapid than the final degradation of the free binding region fragments, which explains the accumulation of the latter in cartilage during aging.

Insulin-like growth factor-I and its complexes in normal human articular cartilage: studies of partition and diffusion

Insulin-like growth factor-I (IGF-I) plays a major role in cartilage homeostasis. Our objective was to study the penetration of IGF-I, both alone and bound to serum proteins, into the different zones of normal human cartilage using radioactively labeled IGF-I. The uptake of free IGF-I was higher than that predicted on the basis of excluded volume calculations and showed concentration dependence: we attributed this to reversible binding of the hormone to the tissue. Since the extent of binding was much higher than that calculated for binding to cell receptors, we concluded that IGF-I binds to matrix components. The kinetics of desorption of IGF-I from cartilage confirmed our conclusions regarding binding. The degree of uptake of IGF-I protein complexes prepared by labeling human serum with [125I]IGF-I showed that such complexes are largely excluded from normal cartilage and that the amounts present in the tissue are too low to affect proteoglycan metabolism.

Concentration and size distribution of insulin-like growth factor-I in human normal and osteoarthritic synovial fluid and cartilage

The concentration of free insulin-like growth Factor-I (IGF-I) and its complexes was determined in human normal and osteoarthritic synovial fluids, using ultrafiltration through 20- and 100-kDa membranes, followed by a radioimmunoassay of each fraction. In addition, freshly obtained samples of normal and osteoarthritic cartilage were incubated for several days, at both 4 and 37 degrees C. The incubation media (desorbates) were analyzed the same way as the synovial fluid samples to yield the concentration of IGF-I in cartilage in situ. Our findings are (i) Free IGF-I content is extremely low in both human serum and synovial fluid and there is no significant difference between the two; (ii) The concentration of total IGF-I in normal human synovial fluid is an order of magnitude lower than that in serum due mainly to the decrease in the concentration of the large complex; (iii) Preliminary results show that the total IGF-I in osteoarthritic synovial fluids is twice as high as in normal fluids; (iv) In normal human cartilage the levels of IGF-I in all its forms are very low and are consistent with the expected exclusion of large molecules by the extracellular matrix; (v) By contrast, in osteoarthritic cartilage, the concentrations of all forms of IGF-I are high, probably due to increased permeability of the matrix and binding; (vi) The levels of IGF-I found in normal human cartilage are more than an order of magnitude lower than those which stimulate proteoglycan synthesis in human cartilage in culture, while the IGF-I levels in osteoarthritic cartilage lie in the range in which stimulation does occur.

Age-related changes in collagen packing of human articular cartilage

We have used X-ray scattering techniques to determine if the lateral packing of collagen molecules in the fibrils of human articular cartilage changes with age. Such changes would affect the available intrafibrillar volume and consequently the amount of intrafibrillar water. Measurements were made both in the presence and absence of compression on samples from donors aged 20 to 90 years. We find a weak though statistically significant tendency towards less dense collagen packing in native tissue as a function of age. However, the increase in packing density in response to pressure does not change with age, and the packing density in articular cartilage from which the proteoglycan molecules have been removed is similarly not age-dependent. The small increase in intrafibrillar water indicated by our data is insufficient to explain the reported increase in fibril diameter in samples from aged donors.

Brain cell membrane modification following hypercapnia and recovery in newborn piglets

The effect of hypercapnia on brain cell membrane structure and function was studied in anesthetized newborn piglets. Lipid peroxidation products (conjugated dienes and fluorescent compounds), Na+,K(+)-ATPase activity and enzyme affinity to ATP (substrate), K+ and Na+ ions (activators), and strophanthidin (inhibitor) were measured in three groups of animals: controls, those exposed to 90 minutes of PaCO2 > 80 mmHg (hypercapnia) and those exposed the same way, following restoration of normal PaCO2 (recovery). Enzyme activity was unchanged by hypercapnia, but enzyme affinity was altered as indicated by an increase in ATP affinity. Affinities to Na+, K+, and strophanthidin were unchanged. Restoration of normal PaCO2 resulted in an increase in conjugated dienes. The data demonstrate that hypercapnia followed by restoration of normal PaCO2 in healthy term newborn piglets is associated with mild modification of brain cell membrane Na+,K(+)-ATPase, possibly due to lipid peroxidation.

Effects of a continuous infusion of tris(hydroxymethyl)aminomethane on acidosis, oxygen affinity, and serum osmolality

The effects of a continuous infusion of tris(hydroxymethyl)aminomethane (THAM) on pH, base excess, p50, serum osmolality, and plasma drug concentration during respiratory acidosis were studied in newborn piglets. Measurements were made during three experimental periods: (1) control period with normal blood gases; (2) hypercapnia period, and (3) hypercapnia plus THAM period (THAM infusion: 1.65 mmol/kg/h). pH decreased and paCO2 increased between control period (7.40 +/- 0.05 and 45 +/- 3 mm Hg) and hypercapnia period (7.24 +/- 0.06 and 59 +/- 2 mm Hg; p < 0.001; mean +/- SD). pH returned to baseline (7.37 +/- 0.04) during the hypercapnia plus THAM period, while paCO2 remained elevated (63 +/- 4 mm Hg). p50 increased from 30.7 +/- 5.9 to 38.3 +/- 4.7 (p < 0.05) during hypercapnia and decreased with hypercapnia plus THAM. THAM concentration and base excess increased with time and were linearly related. Serum osmolality was unchanged during the THAM infusion. We conclude that continuous infusion of THAM is effective in normalizing pH during respiratory acidosis in the piglet.

Effect of hyperventilation on oxygenation of the brain cortex of neonates

A new phosphorescence imaging method (Rumsey et al, Science (1988) 1649) has been used to continuously monitor the oxygen pressure in the blood of the cerebral cortex of newborn pigs. The animals' blood pressure was continuously measured and PaCO2, PaO2 and arterial blood pH were measured periodically. The oxygen pressure in the blood was quantitatively determined for regions of about 100 um square within the image (from a total field of about 3 mm diameter). It was observed that during hyperventilation, which lowered PaCO2 and increased pH of the blood, oxygen pressure decreased in proportion to the decrease in PaCO2. For example, hyperventilation which decreased PaCO2 from its normal value of 40 Torr to 10 Torr caused a rapid (within 5 minutes) decrease in oxygen pressure in the blood of capillaries and veins to approximately 1/4 of normal.

Effect of hyperventilation on oxygenation of the brain cortex of newborn piglets

A new phosphorescence imaging method (Rumsey et al. Science Wash. DC 241: 1649-1651, 1988) has been used to continuously monitor the PO2 in the blood of the cerebral cortex of newborn pigs. A window was prepared in the skull and the brain superfused with artificial cerebrospinal fluid. The phosphorescent probe for PO2, Pd-meso-tetra(4-carboxyphenyl)porphine, was injected directly into the systemic blood. The phosphorescence of the probe was imaged, and the lifetimes were measured using flash illumination and a gated video camera. The PO2 in the blood of the veins and capillary beds of the cortex was calculated from the lifetimes. Systemic blood pressure was continuously monitored while the systemic arterial PCO2, PO2, and blood pH were measured periodically. The PO2 in the blood was quantitated for 60- to 200 microns2 regions within the image (from a total field of approximately 3 mm diam). The PO2 in the microvasculature was not uniform across the viewing field but increased or decreased in each region independently of the other regions. Thus at any point in time the PO2 in a region could be substantially above or below the average value. During hyperventilation, which lowered arterial PCO2 and increased pH of the blood, the average PO2 decreased in proportion to the decrease in arterial PCO2. For example, hyperventilation, which decreased arterial PCO2 from its normal value of 40 Torr to 10 Torr, caused a rapid (within 5 min) decrease in PO2 in the blood of capillaries and veins to approximately one-third of normal.

Some biochemical and biophysical parameters for the study of the pathogenesis of osteoarthritis: a comparison between the processes of ageing and degeneration in human hip cartilage

We have investigated the changes in some of the biochemical and biophysical properties of human femoral head cartilage on the one hand during ageing and on the other hand in osteoarthritis. Topographical variations were also investigated. The parameters studied were those relevant to cartilage function, viz., proteoglycan concentration (as expressed by the concentration of negatively charged groups), the rate of glycosaminoglycan synthesis, water content, osmotic pressure and fluid loss during compression. During ageing the fixed charge density was found to increase at all sites of the femoral head provided fibrillation was absent: osmotic pressure increased accordingly whilst loss of fluid under the effect of externally applied compression diminished. In cartilage from osteoarthritic joints the opposite changes were found. The rate of GAG synthesis varied considerably with site on the femoral head. It decreased somewhat with age on the superior surface, but increased on the inferior surface. When the same sites were compared, the rate of GAG synthesis in cartilage from osteoarthritic heads was either the same as or lower than in cartilage form normal heads in the same group.

"Free" and "exchangeable" or "trapped" and "non-exchangeable" water in cartilage

We repeated some of our own previous experiments, as well as some of Torzilli's recent experiments (11) on which he bases his conclusions relating to a nonexchangeable "trapped water" in cartilage. We are unable to confirm Torzilli's findings. We observed partition coefficients for 3H.HO very close to unity. That both the extrafibrillar and most of the intrafibrillar water is freely exchangeable and behaves as available water towards small solutes has been independently shown (3) for other collagenous tissues. All the different permutations of partition experiments have yielded results that are fully consistent with our original picture of the very major fraction of cartilage water being free.