In a previous investigation by Wiig & Tenstad (2001) it was determined that, the exclusion volume of the negatively charged IgG4 (pI = 6.6) is almost twice that of the positively charged IgG1 (pI = 8.7). Furthermore, in a recent in vitro study, Wiig et al. (2003) were able to titrate the pI of albumin and could thereby study a graded effect of interstitial charge on exclusion. This latter study documented that, in fully swollen dermis, almost 40 % of the total albumin exclusion can be accounted for by the presence of negatively charged GAGs. These two studies are precursors of the present work, where we hypothesized that fixed interstitial negative charges have a significant contribution to excluding albumin from tissues of otherwise normally hydrated animals. In this study we used a charged modified albumin probe, cHSA with no net negative charge, and as such, not involved in electrostatic interaction with the polyanionic interstitial matrix constituents. By comparing directly its distribution volumes in selected tissues to the corresponding values of a normally charged HSA we were able to approximate the steric and electrostatic components of the total volume exclusion, and found that the excluded volume of cHSA in rat skeletal muscle and skin is about 40 % and 25 %, respectively, lower than the corresponding values for HSA.
In a study involving skin and muscle tissues in rabbits, Gandhi & Bell (1992) investigated in detail the transvascular transport of charge-modified albumin. On the other hand, to our knowledge, the effect of modifying the net electrical charge of albumin on the extravascular distribution of this species in vivo has been addressed only briefly before. Specifically, in an abstract, Bell (1985) presented that the interstitial exclusion volume of charge-modified albumin is lower than the corresponding one for native albumin. The present study is therefore the first thorough documentation of the electrostatic exclusion of albumin in vivo.
The experimental method employed here has been presented and evaluated in several series of experiments that established the volume distributions of rat serum albumin (RSA) and IgG (Wiig et al. 1992, 1994; Wiig & Tenstad, 2001). These publications discuss in detail the method and its important steps along with the potential sources of errors. In what follows, therefore, only some key points that are relevant to the present study will be mentioned.
Our accurate estimation of albumin distribution volume depends on a series of requirements that have to be satisfied and can be summarized as follows: the labelled albumin has to reach and maintain a steady-state distribution within tissues; the estimates of exclusion volumes have to account for the probe present only in the fluid-phase of the interstitium; the labelled macromolecule has to retain its label for the entire duration of infusion; the method of sampling interstitial fluid has to be reliable.
Achievement of steady state prior to tissue and fluid analysis is an important requirement in this study. As presented in Fig. 1, a 72 h infusion time, proved to be a duration sufficient for the establishment of a steady-state plasma tracer concentration for both probes. Nonetheless, to ensure that both tracers reached a steady state in the interstitium at the time of tissue sampling, we allowed for a longer tracer equilibration period, i.e. 5-7 days. As showed by Fig. 2, cHSA and HSA distribution volumes at t= 120 h and t= 168 h are not statistically different, suggesting no further uptake occurred during this point and that a steady state was achieved.
One concern in these experiments regarded the potential of non-specific tissue binding of the two macromolecular probes. As presented in Table 1, about 93-95 % of HSA (pI = 5.0) was eluted from the rat muscle and skin tissues. These results are similar to the ones reported for RSA by Wiig et al. (1992), a proof that the heterologous HSA protein does not have a higher tendency to bind to the tissue. cHSA (pI = 7.6), however, presented a slightly higher degree of non-specific binding. About 90 % of this probe was eluted from medial and lateral hindlimb rat muscles and slightly less, i.e. about 81 % and 84 % from hindlimb and back skin, respectively. In these experiments we experienced a lower degree of binding than reported in previous in vivo studies involving IgG (Wiig & Tenstad, 2001) or, the more recent in vitro study involving HSA (Wiig et al., 2003). Nevertheless, in agreement with these two previous reports we also observed that cHSA exhibits a higher degree of binding within tissues as opposed to the unmodified HSA, and furthermore that the degree of binding appears to be directly correlated with the increase in the pI of the probe.
Another problem that may arise when using long equilibration times is related to probe alteration and degradation during the infusion period. The low amount of free iodine (< 1 %) and the lack of significant probe degradation products (see Fig. 4) suggest that these problems were negligible. The stock solution of HSA was dimer-free, whereas cationization of HSA led to less than 5 % dimer formation in the stock solution. After circulation in the animal for 7 days, the amount of dimer increased up to a maximum of 10 % of total albumin amount for both tracers (see Fig. 4). Theoretically, the formation of dimers would lead to a slight underestimation of the absolute extravascular distribution volume. However, since the percentage of dimer is similar for both species, the effect of dimerization on the relative distribution of the probes will be negligible.
In this study we used heterologous HSA and not homologous rat serum albumin. Our choice was based on our experience that HSA is more stable and has a lower tendency to form dimers; these properties are important in experiments involving chemical modification of albumin. One potential problem with infusing this probe could have been related to the possibility of antibody development in response to heterologous HSA infusion. The immunological response is dependent on the amount of antigen infused and its persistence in the animal body (Modabber & Sercarz, 1970). Our experiments involved infusion of trace amounts of HSA and the duration of our experiments was relatively short, therefore the immunological responses might not play a dominant role in these experiments. As far as the HSA probe is concerned, it should be noted that this probe was eluted from tissues in the same proportion as RSA and furthermore, as will be discussed shortly, its distribution volume in rat tissue is very similar to that of RSA. Therefore, since there were negligible probe degradation products and no free iodine accumulated we conclude that use of heterologous albumin allowed for an accurate determination of the interstitial distribution volumes.
Interstitial fluid was sampled by post-mortem implantation of nylon wicks within selected tissues of rats. The wick method was evaluated and described in detail in experimental studies involving tissues such as skin (Aukland & Fadnes, 1973; Wiig et al. 1988) and muscle (Wiig et al. 1991). As shown in these earlier reports, samples obtained by implantable wicks are representative of the interstitial fluid. Furthermore, studies in rabbits have shown that the colloid osmotic pressure of wick fluid and lymph is similar under steady-state conditions (Fadnes, 1981), the latter being a more commonly used reference for interstitial fluid. Our protocol of sampling interstitial volumes was identical with the one documented in these studies, including the period of 20 min wick implantation; the duration was determined to be sufficient for interstitial fluid equilibration. Evaporation of wick fluid may be a potential problem that was controlled for by wick handling at 100 % relative humidity. Furthermore, it should be noted that available and excluded volumes of albumin are calculated as ratios of Vi, thereby cancelling out the volume of wick fluid from the equation (eqn (5)). Our evaluation of tracer distribution volumes is thus based on relative counts only and with gamma-emitting tracers of the entire tissue samples and wicks. Potential errors involved in evaporative losses and measurement of very small weights or volumes are thereby eliminated (Wiig et al. 1992).