Microdialysis update: optimizing the advantages

Microdialysis was introduced in the early 1970s as a method to measure dynamic release of substances in the brain (see Tossman & Ungerstedt, 1986). The technique has been refined over the past three decades due to the development of new materials for dialysis membranes and commercial availability of smaller, more consistently fabricated probes. A typical microdialysis probe consists of rigid metal concentric tubing with a semipermeable region at the tip (Fig. 1). Molecules of restricted size passively diffuse from the brain through the dialysis membrane into an infusion solution which is then directed out of the brain and collected in tubes for serial analysis of substance content. Probes are inserted into the brain region of interest, typically making lesions during their travel through the brain and at the sampling site. Once the trauma of insertion subsides, usually after an hour or so, probes collect substances released from axons projecting to dendrites and cell bodies of the targeted area. Substances surrounding the semipermeable region of the probe passively diffuse down a concentration gradient into the solution infused through the probe. Substance recovery from the brain decreases exponentially with faster infusion rates. A high precision infusion pump is critical for maintaining constant flow through the probes to ensure that altered substance content in the dialysates reflects changes in release by the brain and not variable diffusion gradients resulting from sporadic changes in flow rates through the probes. High performance liquid chromatography (HPLC) is commonly used to measure target substances in the dialysates, but other methods such as radioimmunoassay may be employed. The development of microbore columns for HPLC (Durkin et al. 1985) and their commercial availability by the mid 1990s has made it possible to accurately measure smaller amounts of substances in the dialysates.

Figure 1. Schematic diagram illustrating flow through a microdialysis probe and probe size relative to brainstem regions where substance release was measured by Richter et al. (1999).

After leaving the probe, the 3-4 μl samples were collected in tubes and substance content measured with HPLC. Peak location is used to identify the target substance. Either peak height or area is typically measured to determine the amount present in the sample.

In a paper in this issue of The Journal of Physiology, Richter et al. (1999) have optimized the advantages of microdialysis to measure the release of glutamate, GABA, adenosine and serotonin during hypoxic episodes in a key brainstem region involved in the control of breathing in cats. Since this region, the ventral respiratory group (VRG), is a cell group which extends rostrocaudally for several millimetres, insertion of a microdialysis probe into the VRG allows the effective measurement of neurotransmitter release onto cell bodies rostral and caudal to the probe, as well as distal dendrites of the (injured) cell bodies at the probe tip. Compression trauma and tissue destruction from a 0.5 mm (diameter) microdialysis probe can extend 100-300 μm beyond the probe circumference. By documenting that phrenic discharge was not adversely affected by probe insertion, Richter et al. (1999) demonstrated that critical brainstem pathways for rhythm generation and transmission from the VRG to phrenic motoneurones remained intact, despite predictably large lesions from the probes.

A concern with microdialysis is the effect of variable membrane permeability on substance recovery. Since the percentage recovery in microdialysis studies varies, it is ordinarily determined for each individual probe and substance of interest. Factors affecting dialysis through semi-permeable membranes include: (i) length and diameter of the membrane, (ii) flow rate, (iii) temperature, (iv) pH, (v) molecular weight, charge and shape of the substance, (vi) binding of the substance to the dialysis membrane or outlet tubing, and (vii) stability of the substance at the temperature and pH of the perfusion medium. Probe recovery characteristics vary with repeated or long-term use since protein accumulates on dialysis membranes inserted in brain tissue and can gradually block the pores. Richter et al. (1999) were careful to report in vitro recovery rates for each neurotransmitter and only reported results where recovery characteristics of the probes were similar before and after each experiment.

Increased sensitivity of the methods for analysing substance content in the dialysates has made it feasible to measure substance release over shorter time intervals. This improved time resolution is due primarily to the development of microbore columns and electrochemical detectors with less ‘dead space’ for the HPLC analysis (Durkin et al. 1985). These improvements make it possible to detect smaller amounts in each dialysate sample and allowed Richter et al. (1999) to collect measurable (3-4 μl) samples every 1-1.5 min using high perfusion rates (2-4 μl min⁻¹). Although these rates yielded only 4-9 % recovery of the actual amount of substance in the dialysed brain region, detectable changes in release were clearly demonstrated over the short time intervals and fully characterized during hypoxic episodes.

Sample collection over a minute is an inordinately long time. Although substances released during four to five breaths in a cat are measured during this time frame, the alternative is to collect brain tissue for biochemical assay of the substance content at one time point per animal. Despite the time resolution of minutes rather than milliseconds, brain lesions from the probes and the potential for variable membrane permeability, microdialysis is still a powerful technique for directly measuring substance release by the brain. Using this technique, Richter et al. (1999) now provide direct evidence for temporal release of four key neurotransmitters and clarify the role for each in shaping a complex respiratory pattern during hypoxic episodes.