Electrolyte Channels and Aldosteronism

Over the past few years, it has become apparent that hyperaldosteronism is far commoner than was once suspected and screening of unselected patients with hypertension reveals that about 5-10% of patients have primary hyperaldosteronism. In patients with resistant hypertension, that percentage increases to 15-20%. About 30% of hyperaldosteronism is caused by aldosterone producing adrenal adenomas (APA). Most of the rest is related to bilateral adrenal hyperplasia with less than 5% of cases being familial. The secretion of aldosterone in adrenal cells is dependent on the intracellular calcium concentration and increases in response to higher plasma calcium. Entry of calcium into the cells is in turn dependent on voltage-gated membrane calcium channels (which allow calcium influx when the cells are depolarized) and a calcium ATPase which removes calcium from the cells. Under normal circumstances, adrenal cells are hyperpolarized thus keeping these calcium channels closed. Cell polarization is maintained by a combination of the action of the Na-K-ATPase (which exchanges 3 intracellular Na for 2 extracellular K) and membrane K channels lead to K loss from the cells.

Angiotensin II inhibits the Na-K-ATPase leading to cell depolarization, calcium influx into cells and aldosterone secretion. Similar effects are seen when cells are treated with oubain, a specific Na-K-ATPase inhibitor that also leads to hyperaldosteronism.

In 2011 in a seminal paper in Science, Choi et al reported finding somatic mutations in KCNJ5, a membrane potassium channel in patients with APA. These were identified by sequencing tissue from the tumors and comparing with the surrounding tissue. Subsequently, it has been found that about 30-40% of patients with APA have somatic mutations in KCNJ5. These mutations are believed to reduce the ion selectivity of the channels, allowing Na to move into the cell and reduce the resting membrane potential. A number of families have been identified with KCNJ5 mutations resulting in bilateral hyperplasia - now called Familial Hyperaldosteronism type III.

Recently, a paper was published in Nature Genetics which attempted to determine if there were other somatic mutations in patients with APA. In this study, they took KCNJ5-normal patients and sequenced the exons of the tumors and the surrounding tissue. There were very few mutations identified but 5/9 patients had mutations in ATP1A1, a component of the Na-K-ATPase or ATP2B3, a component of the calcium ATPase that removes calcium from adrenal cells. Follow-up targeted sequencing of 300 patients with APA revealed that about 7% had mutations in one of these two genes. Patients with these mutations had higher aldosterone levels, lower minimum potassium levels and higher systolic BP, all indicators of more severe disease. Notably, no families have been identified with these mutations. In vitro studies revealed that cells with these mutations have very low membrane potentials and it is speculated that if this was a germline mutation, it would likely not be compatible with life. This is a fascinating insight into how very small changes in electrolyte channels can have far-reaching consequences and shows a great progression from exome sequencing to the bench and to clinical investigation.

The images in this post are taken from the recent paper in Nature Genetics. One would wonder if somatic mutations explain some of the missing heritability that were are seeing in genetic studies of common diseases. See this previous post by Lisa on the genetic causes of hypertension.

New potential drug targets in ADPKD

Currently there is no good treatment for the most common inherited cause of ESRD, adult polycystic kidney disease. There have been a number of high profile trials in ADPKD in recent years. These trials have endeavored to show a reduction in cyst growth and GFR decline with everolimus, sirolimus) and most recently Tolvaptan (TEMPO). The longer (2years) and larger (433 patients) of the two mTOR inhibitor trials (everolimus) did show a significant reduction in cyst growth at one year but not in GFR reduction. The shorter sirolimus trial failed to show a reduction in cyst growth or GFR decline. The TEMPO trial was over 3 years, had 1445 patients and did show that the V2 antagonist Tolvaptan slowed GFR decline (reciprocal of the serum creatinine level, −2.61 [mg per milliliter]−1 per year vs. −3.81 [mg per milliliter]−1 per year; p=0.001) and cyst growth, 2.8% per year (95% confidence interval [CI], 2.5 to 3.1), versus 5.5% per year in the placebo group (95% CI, 5.1 to 6.0). The jury is still out about the clinical applicability of these drugs and there have been criticisms. For example, tolvaptan is very expensive and would need to be used long term. In the mTOR inhibitor trials some argue doses could have been higher and the lack of hard end points speaks for itself.

However, all is not lost. A potential new drug target in ADPKD was reported by Rowe et al. in Nature Medicine last month. A good overview of the topic can also be found in the same issue.

Using MEF cells from pkd-/- and pkd+/+ mouse littermates they found that growth medium from the pkd-/- cells was more acidic and that the pkd-/- cells had a higher ATP content. To investigate which metabolic pathways might be causing this difference they used NMR spectroscopy and found lower glucose and higher lactate levels in the knock out cells. They then used a mitochondrial ATPase inhibitor to determine the source of higher ATP and found only wt cell had a reduction in ATP with this treatment.  Then the investigators did a real-time PCR analysis on the pkd-/- cells and found an upregulated glycolysis signature. They thus concluded that the pkd-/- cells rely on aerobic glycolysis for their energy demands. This is known as the Warburg effect described in cancer cells (Otto Warburg, a physician-scientist, received the Nobel Prize in Physiology or Medicine in 1931). To see if these in vitro findings translated into in vivo they used Ksp-Cre; Pkd1flox/− mice, which develop early and severe PKD and measured 13C-glucose or 13C-lactate using 13C-NMR. The findings were the same. The authors then used 2-deoxyglucose (2-DG) an analogue of glucose that is unmetabolised. They treated wt and pkd deficient mice this compound and found that the pkd deficient mice had a lower cyst index and lower 13C-glucose consumption as measured using 13C-NMR.

This interesting study proposes that the use of drugs targeting this pathway in combination with other drugs may reduce cystogenesis and progression of CKD in ADPKD. The authors do stress that their summary with regard to human treatments is speculative. In all I think the future is not so gloomy for ADPKD.

Posted by Andrew Malone

Bad Odor

Cystinosis is an autosomal recessive disease caused by a mutation in CTNS, which encodes the lysosomal transporter of cystine. This leads to intracellular cystine accumulation which leads to renal, neurological and cardiac damage. The treatment for this condition is life-long cysteamine. Back in 2011, we reported on a new formulation of cysteamine that is given just twice daily (rather than q6 hours) and is associated with a lower incidence of side effects including halitosis and body odor. There was a higher incidence of GI side effects in patients treated with the new drug. The halitosis occurs because a proportion of the drug is converted to dimethylsulfide and this can appear in expired air. The reason why the new formulation is associated with less halitosis is because, although serum drug levels are the same as with the older drug, the total dose is reduced so there is less overall conversion to the offending metabolites. This is welcome news for patients given that the drug should be started before age 5 and the halitosis and body odor cause major social problems for patients which can result in non-compliance.

Yesterday, the FDA approved cysteamine bitartrate ER (Procysbi) for the treatment of cystinosis. Essentially, this is an enteric-coated delayed-release version of the drug that is largely absorbed in the small intestine. The major issue, of course, is cost. The traditional form of the drug costs about $8,000/year while the new formulation will cost approximately $250,000. That is an enormous difference for a drug which does no more than reduce side effects. However, when you read this New York Times article on procysbi, you can better understand why the parents of these children think that this is entirely worth it.

According to the same article, it is estimated that by 2018, spending on orphan drugs will account for about 16% of overall spending on prescription drugs. This is bound to lead to conflict between insurers, patients and the manufacturers when trying to decide who to treat and who will pay. Is it worth $242,000 every year for a drug to be given twice, rather than four times daily? The problem, of course, is that this is the only way that the company can recoup the cost of development because there are so few patients with the disease - approximately 300 in the US. It should be remembered, of course, that the ulitmate cost of non-compliance in this case is very high - dialysis, transplantation and loss of future earnings so the cost calculation is not as simple as I suggested above. This is a debate that is certainly going to continue over the next few years.

Does nephrology need personalized medicine?

Systems biology is one of science’s growth areas. Sequencing technologies and software tools developed on the back of the human genome project have reduced the cost of, and therefore increased access to, large and complex datasets (ending in -ome) of genome sequences (genomics), gene expression (transcriptomics) and proteins and metabolites (proteomics and metabolomics). Systems biological techniques integrate these datasets and provide insights into how phenotypes may emerge from interacting biological processes rather than isolated genes or proteins.

A recent editorial in the journal Nephrology Dialysis Transplantation examined this field in general and its relevance to nephrology. The authors mention that –omic datasets have been useful in modeling “self-organized highly interconnected networks”, and that such networks have implicated unexpected candidates in disease pathogenesis (see for example, this paper on cardiac hypertrophy). 

The review goes on to suggest that using the tools of systems biology to finely phenotype individuals will usher in an era of truly personalized medicine. However, it is not clear to me that a definite sequel to this type of analysis will be the personalization of treatment or even that the concept of personalized medicine is particularly suited to our current view of what constitutes clinical evidence.

Diseases such as the ANCA-associated vasculitides (AAV) are now known to exhibit genomic variability. Randomised controlled trials (RCTs) in AAV (such as here and here) have been hampered by: 

1. Short follow-up times 
2. Inter-group heterogeneity which may have affected outcomes. These factors have contributed to ongoing debate about the applicability of the results of these trials (see correspondence here). 
3. Additionally a recent trial in membranous nephropathy, likely to represent another disease with distinct –omic subsets, was marked by slow recruitment. 

 

All these points together suggest that it may be difficult to conduct meaningful clinical studies of distinct –omic subtypes in nephrological diseases. Currently, primacy is given to RCTs when evaluating the efficacy of new treatments; and in nephrology the community is finally beginning to produce the RCTs which have been absent historically. 

If the focus is to switch away from RCTs with their large, well-matched study groups and towards splitting groups up by some -omic fingerprint I am able to envisage a time when one has to choose between giving more credence to the results of larger, “non-personalised” trials or smaller studies featuring –omic data but lacking the controlled element of RCTs.  Would this represent progress?