As already noted in Comment 1 posted by Dr. Frauke Stanke, the study by Sousa and colleagues published recently in this Journal [1] strongly supports our early notion [2-7; not referenced in the Sousa paper] that bioelectric measurements on human rectal biopsies performed ex vivo in Ussing chambers (“ICM”) can be used as a highly sensitive technique for fine-diagnosis of cystic fibrosis (CF) and potentially also for ex vivo testing of CFTR correctors and potentiators in a native epithelium, avoiding the possible occurrence of cell culturing artifacts. Whereas our diagnostic data were obtained using the socalled Rotterdam protocol (suction biopsies; short-circuit current/Isc measurements; recirculating bath fluid by airlifts; HCO3-/CO2 pH buffering; adding Ca2+-linked secretagogues e.g. carbachol and histamine prior to cAMP agonists), the Sousa paper uses the socalled Freiburg protocol which differs substantially from the Rotterdam ICM technique (forceps biopsies; open circuit/equivalent Isc measurements; continuous perfusion with bicarbonate free Ringer; adding cAMP agonists prior to carbachol). Recently, a modified Rotterdam protocol has been developed introducing recurrent washing steps aimed to lower basal cAMP levels, and adding cAMP-based secretagogues (forskolin/IBMX) prior to carbachol, resulting in a strong (~5-fold) accentuation of the CFTR-mediated Isc response to cAMP elevation [5]. This modified protocol has become the basis of a new European ICM SOP which is presently validated in multiple European CF Centers and has been endorsed by the ECFS Diagnostic Network Working Group and the ECFS Clinical Trials Network. A very similar ICM SOP has been set-up and tested in several CFF-TDN Centers in the USA, and a fully validated, harmonized ICM protocol using commercially available recirculating Ussing chambers, Isc measurements and HCO3-/CO2 pH buffering will be soon at the disposal of CF Centers worldwide.

Major advantages, limitations and pitfalls of the historic Rotterdam and Freiburg technique have been summarized and discussed in multiple publications [6,8,9]. Regretfully, Sousa and colleagues do not restrict themselves to a discussion of their own valuable results, but also choose to downplay the diagnostic value of the Rotterdam technique (“leading to conflicting results; precluding good correlations with clinical symptoms”), despite overwhelming evidence for the ability of both techniques to discriminate between non-CF, classic CF, and non-classic CF. The clear cut-off value between PS-CF and non-CF presented in the first comprehensive diagnostic ICM validation study using the historic Rotterdam protocol [10] already provided strong evidence that ICM can be sensitively used for fine-diagnosis of CF. Correlation of CFTR function to clinical phenotype in general was seen also in that study for patients with known CF, although it should be noted that the mild clinical CF-like features in questionable CF patients entering the extended diagnostic evaluation typically are not distinguishable from symptoms resulting from other differential diagnoses. This is not surprising for “questionable” CF patients with equivocal results in the standard CF diagnostic tests, and provides the main reason for a more precise evaluation of CFTR dysfunction i.e. by nasal potential difference or ICM.

Rather than stirring up the pro-and con discussion in detail again, we here like to focus on two key differences between the two historic techniques, i.e. the use of bicarbonate (HCO3-) containing (Rotterdam) versus bicarbonate-free (Freiburg) perfusion solutions, and recirculation (Rotterdam) versus continuous perfusion (Freiburg). The bicarbonate-free condition, though clearly non-physiologically, is advocated in the Sousa and previous papers “to exclude a possible contribution of CFTR-independent electrogenic HCO3- secretion, mimicking residual Cl- channel function in CF colonic epithelia”. However, with the sole exception of heat-stable E.coli enterotoxin (STa)-provoked, cyclic GMP-mediated HCO3- secretion in human duodenum [11], evidence for the occurrence of CFTR-independent, cyclic AMP or carbachol-induced electrogenic HCO3- secretion in mouse or human intestine is completely lacking [11,12]. Consequently, we do not see any signs of cAMP-and carbachol-induced electrogenic anion secretion in rectal biopsies or rectal organoids from CFTR-null patients and in colon from Cftr-/- mice bathed in HCO3- -rich medium [12,13]. The HCO3- -free condition however may explain why in continuously perfused Ussing chambers, extensive depletion of endogenous prostaglandins in the presence of the cyclooxygenase inhibitor indomethacin, might eventually lead to a false-positive, CF-like current response to carbachol in non-CF rectal biopsies (cf. Fig. 1A, Sousa paper). This condition apparently results in a severe depletion of endogenous cAMP, the intracellular messenger required for protein kinase A activation and phosphorylation/activation of the CFTR channel. Addition of agonists increasing cAMP restore the normal, non-CF response which is why the Freiburg protocol contains a third application of carbachol in the presence of forskolin/IBMX.
However, as shown recently by Dr. Martin Hug (University Medical Centre Freiburg; personal communication), a switch from HCO3- -free to HCO3- -rich medium has a similar effect, i.e. it completely normalizes the current response to carbachol even in the absence of cAMP agonists, mimicking our findings using the Rotterdam protocol. Our preliminary studies suggest that this recovery of the anion secretory response to carbachol in non-CF biopsies results from the stimulation of cAMP production by a HCO3- -dependent soluble adenylyl cyclase (cf. ref. 14). The relevance of this observation for CF diagnosis became immediately apparent when we compared ICM tracings obtained with the original Rotterdam protocol (adding carbachol first, and cAMP agonists later) with the results of a modified protocol (adding cAMP agonists first, followed by carbachol, i.e. similar to the Sousa protocol) in compound heterozygotes carrying F508del and a “mild” other mutation (R117H; A455E): whereas the original protocol was able to detect a strongly reduced Isc response to carbachol in most patients, the effect of carbachol on a background of saturating cAMP (modified/Sousa protocol) was much less reduced (De Jonge/Derichs, study in progress). Apparently the subsaturating cAMP levels prevailing in the old protocol cause submaximal phosphorylation and activation of mutant-CFTR whereas full activation by excess cAMP (i.e. forskolin/IBMX) is needed to reach “wild type” Isc responses. In contrast, non-CF biopsies generate maximal Isc responses already upon phosphorylation and activation of ~20% of the total pool of apical CFTR [8]. Aside its effect on cellular cAMP levels, bicarbonate may also contribute to CFTR-dependent electrogenic anion secretion through the colonic SLC26A3 (“DRA”) chloride-bicarbonate exchanger which is physically and functionally coupled to CFTR [15]. The putative lack of this component in HCO3- -free bath fluid may possibly explain why almost all classical CF patients in the Sousa study showed zero values for CFTR-mediated Isc responses, whereas in the Rotterdam ICM protocol, as well as in fluid secretion assays carried out with intestinal organoids generated from rectal biopsies, a considerable proportion of classical CF patients (in particular F508del homozygotes), but not CFTR null patients, showed a strongly reduced but measurable CFTR activity [4, 13]. Furthermore, a lack of bicarbonate in the bath may also lead to an underestimation of CF severity in CFTR mutants that retain near-normal Cl- channel activity but show a strongly reduced or absent HCO3- transport activity [16]. Finally, the failure to generate HCO3- is known to result in CF-like salt wasting by the sweat gland, emphasizing the importance of bicarbonate for the proper functioning of CFTR-dependent salt transport in vivo [17].

In summary, both ICM protocols discussed above can be used as valuable biomarkers for CF diagnosis and prognosis. In fact, the new ECFS ICM SOP implements major advantages of both historic protocols and combines them into a state-of-the art protocol which takes previous experiences and improvements into account. The lack of bicarbonate in the Sousa protocol and the use of saturating cAMP levels (probably far beyond the level prevailing under physiological conditions) may have important consequences, including the misclassification of some non-classic CF patients as non-CF, and the masking of (low levels of) residual CFTR activity in patients with classic CF. Furthermore the use of large volumes of bath fluid in the Sousa protocol (requiring continuous perfusion) limits its application for ex vivo validation of novel CFTR modulators and correctors that are usually available only in small quantities. In our opinion, investigators should be aware of these pros and cons that may help them in selecting their own method of choice.

Hugo de Jonge
Professor/Group Leader
Department of Gastroenterology & Hepatology
Erasmus University Medical Center
Rotterdam, The Netherlands
h.dejonge@erasmusmc.nl

Nico Derichs
Director CFTR Biomarker Center & Translational CF Research Group
CF Center, Pediatric Pulmonology
Charité University Berlin, Germany
nico.derichs@charite.de

[1] Sousa et al PLOS One 2012
[2] Veeze HJ, Sinaasappel M, Bijman J, Bouquet J, De Jonge HR (1991) Ion transport abnormalities in rectal suction biopsies from children with cystic fibrosis. Gastroenterology 101: 398-401
[3] Veeze HJ, Halley DJJ, Bijman J, De Jongste JC, De Jonge HR, Sinaasappel M (1994) Determinants of mild clinical symptoms in cystic fibrosis patients. J Clin Invest 93: 461-466
[4] Bronsveld I, Mekus F, Bijman J, Ballman M, Greipel J, Hundrieser J, Halley DJ, Laabs U, Busche R, De Jonge HR, Tümmler B, Veeze HJ (2000) Residual chloride secretion in intestinal tissue of deltaF508 homozygous twins and siblings with cystic fibrosis. The European CF Twin and Sibling Study Consortium. Gastronterology 119: 32-40
[5] Bronsveld I, Mekus F, Bijman J, Ballman M, De Jonge HR, Laabs U, Halley DJ, Ellemunter H, Mastella G, Thomas S, Veeze HJ, Tümmler B (2001) Chloride conductance and genetic background modulate the cystic fibrosis phenotype of Delta F508 homozygous twins and siblings. J Clin Invest 108: 1705-1715
[6] De Jonge HR, Ballmann M, Veeze H, Bronsveld I, Stanke F, Tümmler B, Sinaasappel M (2004) Ex vivo CF diagnosis by intestinal current measurements (ICM) in small aperture, circulating Ussing chambers. J Cyst Fibros 3 (Suppl 2): 159-163
[7] Derichs N, Mekus F, Bronsveld I, Bijman J, Veeze HJ, Tümmler B, von der Hardt H, Ballmann M (2004) CFTR-mediated residual chloride secretion does not protect against early chronic Pseudomonas aeruginosa infection in F508del homozygous cystic fibrosis patients. Pediatr Res 55:69-75.
[8] De Boeck K, Derichs N, Fajac I, De Jonge HR, Bronsveld I, Sermet I, Vermeulen F, Sheppard DN, Cuppens H, Hug M, Melottii P, Middleton PG, Wilschanski M, ECFS Diagnostic Network Working Group. (2011) New clinical diagnostic procedures for cystic fibrosis in Europe. J Cyst Fibros 10 Suppl 2:S53-66.
[9] Mall M, Hirtz S, Gonska T, Kunzelmann K (2004) Assesment of CFTR function in rectal biopsies for the diagnosis of cystic fibrosis. J Cyst Fibros 3 (Suppl 2): 165-169
[10] Derichs N, Sanz J, von Kaenel T, Stolpe C, Zapf A, Tümmler B, Gallati S, Ballmann M (2010) Intestinal current measurement for diagnostic classification of patients with questionable cystic fibrosis: validation and reference data. Thorax 65(7):594-99.
[11] Pratha VS, Hogan DL, Matensson BA, Bernard J, Zhou R, Isenberg JL (2000) Identification of transport abnormalities in duodenal mucosa and duodenal enterocytes from patients with cystic fibrosis. Gastroenterology 118: 1051-1060
[12] Xiao F, Li J, Singh AK, Riederer B, Wang J, Sultan A, Park H, Lee MG, Lamprecht G, Scholte BJ, De Jonge HR, Seidler U (2012) Rescue of epithelial HCO3- secretion in murine intestine by apical membrane expression of the CFTR mutant F508del. J Physiol 590.21: 5317-5334
[13] Dekkers JF, Wiegerinck CL, De Jonge HR, Bronsveld I, Janssens HM, De Winter-De Groot KM,
Bijvelds MJC, Nieuwenhuis EES, Van den Brink S, Clevers H, Van der Ent CK, Middendorp S,
Beekman JM (2013) A functional CFTR assay using primary cystic fibrosis intestinal organoids. Nature
Med in press
[14] Halm ST, Zhang J, Halm DR (2010) beta-Adrenergic activation of electrogenic K+ and Cl- secretion in guinea pig distal colonic epithelium proceeds via separate cAMP signaling pathways. Am J Physiol 299: G81-G95
[15] Ko SB, Shcheynikov N, Choi JY, Luo X, Ishibashi K, Thomas PJ, Kim JY, Kim KH, Lee MG, Naruse S, Muallem S (2002) A molecular mechanism for aberrant CFTR-dependent HCO3- transport in cystic fibrosis. EMBO J 21: 5662-5672
[16] Choi JY, Muallem D, Kiselyov K, Lee MG, Thomas PJ, Muallem S (2001) Aberrant CFTR-dependent HCO3- transport in mutations associated with cystic fibrosis. Nature 410: 94-97
[17] Feldshtein M, Elkrinawi S, Yerushakmi B, Marcus B, Vullo D, Romi H, Ofir R, Landau D, Sivan S, Supuran CT, Birk OS (2010) Hyperchlorhidrosis caused by homozygous mutation in CA12, encoding carbonic anhydrase XII. Am J Hum Genet 87: 713-720