gagCEST effect strongly depends on GAG molecular composition
Emma Olsson1, Pernilla Peterson1, André Struglics2, Michael Gottschalk3, Patrik Önnerfjord4, and Jonas Svensson1,5

1Medical Radiation Physics, Department of Translational Medicine, Lund University, Malmö, Sweden, 2Orthopaedics, Department of Clinical Sciences Lund, Lund University, Lund, Sweden, 3Lund University Bioimaging Center, Lund University, Lund, Sweden, 4Rheumatology and Molecular Skeletal Biology, Department of Clinical Sciences Lund, Lund University, Lund, Sweden, 5Medical Imaging and Physiology, Skåne University Hospital, Lund, Sweden


gagCEST has been suggested as a method for in vivo evaluation of cartilage GAG content. The main type of GAG in cartilage is chondroitin sulfate (CS), most commonly CS-A and CS-C. Validation of gagCEST have mostly been performed using CS-A but the main type in mature human articular cartilage is CS-C. In this study we evaluate the gagCEST effect from GAG in different forms. Our results indicate that mainly CS-A is contributing to gagCEST effect in cartilage, while no or little effect is seen from CS-C. gagCEST may therefore not correctly reflect the GAG content of human articular cartilage.


Proteoglycans, mainly aggrecan, are major components of articular cartilage and the negatively charged glycosaminoglycan (GAG) side chains are key elements for the tissue function1. gagCEST has been suggested as a possible method for in vivo evaluation of osteoarthritis progression through assessment of cartilage GAG content.

The main GAG type in aggrecan, chondroitin sulfate (CS), exist in two common variations, sulfated at the 4th (CS-A) and 6th (CS-C) carbon of the amino sugar, respectively2. The proportions of the two types changes with age and in mature human articular cartilage the 6-sulfated CS-C is dominating3. Attempts to validate the gagCEST method however, commonly includes phantoms with CS-A.

In this study we aim to investigate how well gagCEST reflects the actual GAG content of articular cartilage through gagCEST experiments with different types and forms of GAG.


Two series of phantoms were made using CS-A (85 % 4- and 15 % 6-sulfated, bovine trachea, Sigma-Aldrich) and CS-C/A (50 % 4- and 50 % 6-sulfated, shark cartilage, Sigma-Aldrich) dissolved in phosphate buffered saline (PBS) to seven different concentrations (0-30 mg/ml). Additionally, two phantoms were made with intact aggrecan (protein bound CS) from calf articular cartilage, extracted using guanidinium hydrochloride. The purified aggrecan was freeze dried and dissolved in PBS to concentrations of 10 and 20 mg/ml. pH of all phantom solutions were set to 7.3-7.4.

18 samples of ex vivo cartilage were collected from eight human femoral heads (harvested during hip replacement and stored at -20 °C). With the femoral heads immersed in PBS, gagCEST was measured in the cartilage before the samples were cut out. GAG content was subsequently measured in the cartilage samples using Alcian blue precipitation, a standard method for determination of GAG concentration4.

gagCEST measurements were conducted at 7 T (Philips Achieva) using a 32 channel head coil (NOVA). Saturation was achieved using a pulse train of 5 99 ms long RF pulses with 800° flip angle and 1 ms interpulse delay. 52 frequency offsets were imaged in an interval of ±2.5 ppm around the water frequency, using FFE readout with TE/TR=2.4/10 ms, FOV=120x120 mm2, voxel size=1x1x3/0.7x0.7x3 (phantoms/ex vivo) mm3 and bandwidth=338 Hz/pixel. Correction for B0 inhomogeneity was made using WASSR5.

Magnetization transfer ratio asymmetry (MTRasym) was calculated for a frequency interval of 0.7-1 ppm from water, as a measure of gagCEST effect. Mean values of MTRasym were calculated within regions of interest (ROIs) drawn within each phantom and cartilage sample.


For both of the phantom series, MTRasym resulted in a clear linear dependence on CS concentration. However, the gagCEST effect was markedly lower (on average 52 % lower in mean MTRasym per mg/ml CS) for the CS-C/A phantoms than the CS-A phantoms (figure 1). When adjusting the scale to reflect 4-sulfated CS only, the gagCEST effect of the two phantom series are very similar (figure 2).

For the two phantoms containing intact aggrecan, the gagCEST effect was similar to the effect of the CS-A phantoms (figure 1).

For the human ex vivo cartilage samples, presumably consisting of mainly 6-sulfated CS, only a weak correlation (Pearsons rho = 0.55, 95% CI: 0.094, 0.81) was found with GAG concentration as assessed by Alcian blue (figure 3).


Although there is a clear linear dependence of MTRasym on CS concentration, there is also a clear difference in gagCEST effect between the CS-A and CS-C/A phantoms, suggesting that not all CS types contributes equally to the gagCEST effect. It seems that gagCEST works very well for 4-sulfated CS while there is no or little effect from 6-sulfated CS.

The immature calf articular cartilage contains a larger proportion of 4-sulfated CS which is also reflected in the results of the aggrecan phantoms. The similar result of CS bound in aggrecan and free CS (CS-A) suggests that the aggrecan structure does not have a large impact on the gagCEST effect.

GAG in mature human articular cartilage consists of mainly 6-sulfated CS. Our phantom experiments hence indicate that the gagCEST method may not correctly reflect the total GAG content in vivo. The fairly poor correlation between gagCEST effect and GAG content observed in our ex vivo experiments also supports this conclusion.


We conclude that it is mainly 4-sulfated CS that contributes to the gagCEST effect. Since the main GAG type in mature human articular cartilage is 6-sulfated CS, gagCEST may not correctly reflect the total GAG content in this tissue.


Prof. Leif Dahlberg is acknowledged for providing the ex vivo tissue.


1. A. J. S. Fox, A. Bedi and S. A. Rodeo. The basic science of articular cartilage: structure, composition, and function. Sports Health 2009; 1(6): 461-8.
2. V. H. Pomin. NMR chemical shifts in structural biology of glycosaminoglycans. Anal Chem 2014; 86(1): 65-94.
3. M. T. Bayliss, D. Osborne, S. Woodhouse, et al. Sulfation of chondroitin sulfate in human articular cartilage. The effect of age, topographical position, and zone of cartilage on tissue composition. J Biol Chem 1999; 274(22): 15892-900.
4. S. Bjornsson. Simultaneous preparation and quantitation of proteoglycans by precipitation with alcian blue. Anal Biochem 1993; 210(2): 282-91.
5. M. Kim, J. Gillen, B. A. Landman, et al. Water saturation shift referencing (WASSR) for chemical exchange saturation transfer (CEST) experiments. Magn Reson Med 2009; 61(6): 1441-50.


Figure 1: Mean MTRasym in the three phantom series. For both CS-A and CS-C/A MTRasym depends linearly on the CS-concentration although the gagCEST effect is lower for CS-C/A. The gagCEST effect of the intact calf aggrecan agree well with that of the CS-A phantoms.

Figure 2: Mean MTRasym in the phantoms with CS-A (85 % 4-sulfated, 15 % 6-sulfated) and CS-C/A (50 % 4- sulfated, 50 % 6-sulfated), same as presented in figure 1, although instead as a function of only 4-sulfated CS concentration.

Figure 3: Mean MTRasym in the human cartilage samples as a function of GAG concentration (determined by Alcian blue precipitation and converted to percent of wet weight). One sample was excluded due to problems in the chemical preparation. Pearsons rho was 0.55 with 95 % CI from 0.094 to 0.81.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)