In my postdoctoral studies at Stanford University, I worked in a cystic fibrosis research laboratory under the mentorship of Dr. Jeffrey J. Wine. In cystic fibrosis, the mutation of a single gene leads to the loss of regulated chloride transport across epithelia in the lungs, pancreas, gut and other wet epithelial surfaces in the human body. The mutated protein, the cystic fibrosis transmembrane conductance regulator (CFTR), is a cAMP-stimulated chloride channel from a class of proteins referred to as ATP binding cassette (ABC) transporters. Mutations of genes in this family had been linked to various clinical maladies including: ALD gene -Adrenoleukodystrophy, SUR gene -Diabetes, CFTR gene -Cystic fibrosis, MDR gene -Multidrug resistance in cancer. For many years the function of proteins in the ABC gene family were assumed to be equally homologous to their structure. The genetically related ABC transporters appeared uniformly to split ATP and use that energy to "pump" substrates upstream across cell membranes. After the cloning of CFTR in 1989 by Riordan, Collins and Tsui, perspectives changed. The finding that CFTR was not a pump, but, in fact, a Cl- channel led investigators to re-examine the activity of other ABC transporters.

In my first year of my postdoctoral research I was interested in designing novel approaches to study the structure and function of CFTR, an 'ATPase-like' chloride channel. Since I had good success with creating chimeric cDNAs between P-type ATPases in my dissertation research, I wanted to see how feasible this would be with the CFTR and another homologous ABC transporter, P-glycoprotein. During the sequence analysis I used alignment and phylogenetic software to run comparisons at the nucleotide and amino acid levels for CFTR homology with P-glycoprotein. After much effort, the two genes weren't as related as I would have liked and this project seemed unlikely to succeed. Then serendipity struck and a new finding caught our lab's attention.

In late 1992 the function of the multidrug resistance protein (MDR gene product P-glycoprotein) was dramatically revised. The well-characterized P-glycoprotein (P-gp) transporter, thought to split ATP and pump chemotherapy drugs out of cancer cells, was found to be tightly linked to volume-activated Cl- channel activity by Miguel Valverde. This indicated it was "bifunctional." In fact, in the paper, Valverde used the high homology between P-glycoprotein and CFTR to support their claim. Of course I had just decided the two weren't very homologous and became interested in testing their finding. Soon the CF field became interesting in whether CFTR might be bifunctional and maybe responsible for other functions. As a result, investigating the hypothesized Cl- channel activity of P-gp became an active effort in Jeff Wine's lab and soon a popular subfield in ABC transporter research.

Hence, in preliminary studies of the possible multifunctionality of CFTR, I wanted to first confirm the bifunctional nature reported for its homolog, the P-glycoprotein. The P-glycoprotein was reported to be a swelling-activated chloride channel and I studied whether there was direct evidence linking P-glycoprotein expression levels with up-regulation of swelling-induced chloride currents. Patch clamp and isotopic efflux studies (via I-125) of the chloride currents of cells lines that expressed low/no P-glycoprotein (mouse fibroblasts, NIH 3T3 cell line) and cell lines that were selected for high expression of P-glycoprotein (COL-1000 cell line) indicated the level of expression correlated with an enhanced swelling response, yet the data suggested P-glycoprotein was not functioning as a channel but perhaps is just a regulator in the swelling response pathway.
 

Postdoctoral Research Summary
(excerpts from my NIH progress report of August 1995)

At Stanford University the aim of my research was to pursue two lines of investigation on the potential multifunctionality of two members of the ATP binding cassette (ABC) transporter family; the cystic fibrosis transmembrane conductance regulator, and the multi-drug resistance (MDR) gene product, the P-glycoprotein (Pgp). My greatest progress occurred in the first area: to test the hypothesis that P-glycoprotein, which is known to be a chemotherapy drug pump, directly regulated and enhanced the sensitivity of swelling-activated channels.

To review, P-glycoprotein is the product of the multi-drug-resistance (MDR) gene. The laboratories of Valverde and Higgins originally reported that the transfection of the MDR1 gene confered swelling-activated Cl- currents in cells, and thus P-gp was likely to be both a drug pump and a Cl- channel. While my initial research (with colchicine-selected NIH/3T3 cell lines) ruled out the possibility that P-gp had intrinsic channel activity, I did find an enhancement in the threshold and magnitude of the swelling-activated Cl- and K+ conductance in cells overexpressing P-gp. In order to establish if the human MDR1 gene product (P-glycoprotein) was a facilitator of the regulatory volume decrease (RVD) mechanism of osmotically stressed cells, I tested another set of epithelial cell lines, MES-SA (P-gp-) and DX5 (P-gp+). The MES-SA cell line was a human uterine epithelial line that showed no detectable mRNA for the P-glycoprotein even after 40 cycles of RT-PCR, and the DX5 cells were doxorubicin-selected multi-drug-resistant MES-SA's that overexpressed the P-glycoprotein.

P-glycoprotein does not directly regulate swelling-activated Cl- channels.

Drug resistance tests, P-gp mRNA expression analysis (RT-PCR), and P-gp protein expression analysis (Western Blots) were performed to establish that the parental MES-SA cell line expressed no P-gp while the doxorubicin-selected MES-SA cells (hence called DX5's) expressed high levels of P-gp. Next, whole-cell patch clamp studies and I-125 efflux studies were performed on the cells to characterize and quantitate their swelling-activated Cl- currents. Whole cell patch clamp identified an intermediate conductance swelling-channel that has been described recently as VSOAC (Volume Sensitive Osmolyte/Anion Channel [3]). VSOAC is the swelling-induced, depolarization-inactivating, outwardly rectifying intermediate conductance Cl- channel (40-50pS@+120mV, 15pS@0mV) that is common to various cell lines, as well as the channel identified by Valverde and Higgins as being P-gp. Although we found previously in NIH/3T3 cells that P-gp expression appeared to facilitate swelling current activation, in P-gp-upregulated MES-SA cells the opposite was found.

In both whole cell patch clamp and I-125 efflux studies, swelling activated Cl- currents in DX5 cells (high P-gp) were lower than those in MES-SA cells (no P-gp). These findings made the hypothesis that P-glycoprotein is a Cl- channel facilitator highly unlikely. In addition, further studies with cation (Rb-86) efflux demonstrated that K+ currents were also lowered in the DX5 cells. This suggested PGP could not be a Cl- channel-specific regulator since alterations effected both Cl- and K+ swelling currents.

To test whether these effects are linked directly to PGP expression, P-glycoprotein inhibitors that had been previously reported to inhibit swelling currents, were tested. P-glycoprotein inhibitors (100uM Verapamil, 100uM Dideoxyforskolin) either had no effect or had equal inhibition of both the DX5 (high PGP) and MES-SA (no PGP) cell lines' swelling currents. We believed these results suggested the drug selection procedure might have played a larger role in the swelling current alterations of these and other cell lines. In an attempt to learn about the mechanism involved in swelling channel activation, we tested whether manipulation of Ca levels could elicit or inhibit the swelling-activated Cl- currents of MES-SA cell lines as found previously in other cell lines. Elevation of intracellular Ca (via 1.3uM ionomycin application) had neither a stimulatory or inhibitory effect on swelling currents in either cell line.

In summary, our findings did not support the hypothesis that P-gp was a facilitator or regulator of swelling-activated Cl- channels (VSOAC). Rather, we believed drug selection might have been the cause of RVD activation threshold changes found by Higgins and Valverde. They transfected MDR1 cDNA and then selected the cells in colchicine. By what mechanism selection elicited such a change had yet to be identified. Given the pathway between osmotic stress and channel activation had not yet been elucidated, it would have been difficult to determine which step in that path is altered by cell selection.
 

References

D.B. Luckie, M.E. Krouse, T.C. Law, B.I. Sikic, and J.J. Wine. (1996) Doxorubicin selection for MDR1/P-glycoprotein reduces swelling-activated K+ and Cl- currents in MES-SA cells. Am. J. Physiol. (Cell. Physiol) C1029-C1036.

(Abstract) D.B. Luckie, K.L. Harper, M.E. Krouse, T.C. Law, B. Sikic, and J.J. Wine (1995) MDR/P-glycoprotein expression is associated with reduced swelling-activated K+ and Cl- efflux in Messa and DX5 cells. Biophys. J., 68 (2):A273.

D.B. Luckie, M.E. Krouse, K.L. Harper, T.C. Law, and J.J. Wine (1994) Selection for MDR/P-glycoprotein enhances swelling-activated K+ and Cl- currents in NIH/3T3 cells. Am. J. Physiol. 267 (Cell Physiol 36): C650-C658.

(Abstract) M.E. Krouse, D.B. Luckie, K.L. Harper, T.C. Law, B.I. Sikic, and J.J. Wine (1993) MDR/P-glycoprotein expression facilitates swelling Cl- current activation but is probably not the channel. Pediatric Pulmonology, January Appendix. 1: 5A.