- Email: [email protected]
Genetics of Familial FSGS
Genetics of Familial FSGS
Dr. Demetri: In answer to your first question, there is a ton of work that has been performed in cell-cycle inhibitors of various types and it has been remarkably fruitless. However, that also may be because it was not performed in the right patient population. For some reason, the minute they go into hormone-receptor–positive breast cancer, CDK4 inhibitors have breakthrough designations, which is extraordinary. Part of the problem again is matching the drug with the right patient, validating the targets, and figuring out how to hit the target effectively. We have to be careful not to jump to conclusions. We often say the drug failed, but, instead, it may just be the wrong patient. To your second question about how many drugs may be needed for personalized cancer treatment: who would have guessed that the same drug would treat leukemia and solid sarcoma? This again suggests that we must bin tumors by mechanism and we still do not know how to do that well. I suspect you will be binning things by mechanism of glomerulonephritis, perhaps by mechanisms that also apply to interstitial pulmonary fibrosis (IPF). The world has just had a couple of really great targeted drugs approved for IPF. IPF is worse than most cancers and IPF could have mechanisms that really affect what you are working on in nephrology.
467
12. Blanke CD, et al. Long-term results from a randomized phase II trial of standard- versus higher-dose imatinib mesylate for patients with unresectable or metastatic gastrointestinal stromal tumors expressing KIT. J Clin Oncol. 2008;26:620-5. 13. Demetri GD, et al. Molecular target modulation, imaging, and clinical evaluation of gastrointestinal stromal tumor patients treated with sunitinib malate after imatinib failure. Clin Cancer Res. 2009;15:5902-9. 14. Yap TA, et al. First-in-human phase I trial of two schedules of OSI-930, a novel multikinase inhibitor, incorporating translational proof-of-mechanism studies. Clin Cancer Res. 2013; 19:909-19. 15. Patrikidou A, et al. Influence of imatinib interruption and rechallenge on the residual disease in patients with advanced GIST: results of the BFR14 prospective French Sarcoma Group randomised, phase III trial. Ann Oncol. 2013;24:1087-93. 16. Demetri GD. Identification and treatment of chemoresistant inoperable or metastatic GIST: experience with the selective tyrosine kinase inhibitor imatinib mesylate (STI571). Eur J Cancer. 2002;38(Suppl 5):S52-9.
Genetics of Familial FSGS Martin Pollak, MD Keywords: Autosomal dominant, TRPC6, actinin-4, APOL1, HIV nephropathy, hypertensive nephropathy
INF2,
REFERENCES 1. O’Bryan RM, et al. Phase II evaluation of Adriamycin in human neoplasia. Cancer. 1973;32:1-8. 2. Grier HE, et al. Addition of ifosfamide and etoposide to standard chemotherapy for Ewing’s sarcoma and primitive neuroectodermal tumor of bone. N Engl J Med. 2003; 348:694-701. 3. Rous P. A transmissible avian neoplasm. (Sarcoma of the common fowl) by Peyton Rous, M.D., Experimental Medicine for Sept. 1, 1910, vol. 12, pp. 696-705. J Exp Med. 1979;150: 738-53. 4. Weisberg E, et al. Effects of PKC412, nilotinib, and imatinib against GIST-associated PDGFRA mutants with differential imatinib sensitivity. Gastroenterology. 2006;131:1734-42. 5. Bauer S, et al. KIT oncogenic signaling mechanisms in imatinib-resistant gastrointestinal stromal tumor: PI3-kinase/ AKT is a crucial survival pathway. Oncogene. 2007;26:7560-8. 6. Zhu MJ, et al. KIT oncoprotein interactions in gastrointestinal stromal tumors: therapeutic relevance. Oncogene. 2007;26: 6386-95. 7. Duensing A, et al. Mechanisms of oncogenic KIT signal transduction in primary gastrointestinal stromal tumors (GISTs). Oncogene. 2004;23:3999-4006. 8. Singer S, et al. Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol. 2002;20:3898-905. 9. Rubin BP, et al. KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res. 2001;61:8118-21. 10. Tuveson DA, et al. STI571 inactivation of the gastrointestinal stromal tumor c-KIT oncoprotein: biological and clinical implications. Oncogene. 2001;20:5054-8. 11. Lux ML, et al. KIT extracellular and kinase domain mutations in gastrointestinal stromal tumors. Am J Pathol. 2000; 156:791-5.
I am going to talk about familial focal segmental glomerulosclerosis (FSGS) today. As everyone realizes, FSGS is a histopathologic entity, a pattern of injury that is seen on a kidney biopsy. Nephrotic syndrome is just that, a syndrome. Hypertensive kidney disease also shares features with FSGS. There is considerable overlap between all of these entities. Even though these are just names, I think the fact that we do not have really good ways for the clinicians and the pathologists and patients to talk to each other about these entities does cause some problems. Nevertheless, today we will focus on familial FSGS. As a historical note, while I was still in training, I came across a family with inherited kidney disease in Oklahoma, in a report published in 1998.1 I flew out to a little town outside of Tulsa, Oklahoma, and spent 2 days rounding up members of this family and taking their blood pressures, talking to them, and obtaining urine and blood samples.
Financial disclosure and conflict of interest statements: none. Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA. Address reprint requests to: Martin Pollak, MD, Department of Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Farr 8, Boston, MA 02215. E-mail: mpollak@bidmc. harvard.edu
468
They had a dominant inheritance of a phenotype that was characterized by proteinuria. A few individuals had biopsy-proven FSGS. We started to perform genetic studies to try to identify the underlying gene and also scour the world for other similar families. The advantage of this kind of approach is that one does not need to worry about whether the amount of proteinuria or the histology looks exactly the same from person to person. It alleviates the concern about having a good way to classify patients diagnostically. We therefore performed what at the time were fairly laborious genetic studies, and ultimately this led us to the identification of mutations in the α-actinin-4 gene as the cause of the disease in two families with FSGS, as well as several others.2-4 All of the mutations that segregated with disease in these families were in the actin-binding domain of α-actinin-4, matching well with the best described function of α-actinin-4, which is to cross-link actin filaments.2-4 Indeed, the actin-binding domains associated with FSGS-causing mutations have a much greater affinity to actin filaments than wild type α-actinin-4. I therefore suggest that these mutations are causing a dominant effect by a gain-of-function change in their behavior. This led us to start thinking about what may be going on structurally and we ultimately found that the actinbinding domain could exist as both a closed and an open confirmation.5,6 We therefore concluded that mutations that were associated with kidney disease and led to increased actin binding interfered with the transition between the closed and open states.5,6 We hypothesized further that the presence of diseaseassociated mutations may alter the behavior of the podocyte cytoskeleton. We have been very interested in what this means for the biophysical properties of the podocyte and how we can connect in vitro cytoskeletal behavior with how the podocyte actually behaves. We are making slow progress toward this. If we create reconstituted α-actinin-4/actin networks and perform rheologic studies, in work performed in collaboration with colleagues from the Physics Department at Harvard, we note that when we apply increasing stress to wild-type α-actinin-4/actin networks, we find increasing stiffness of the network until finally it breaks at a certain level of stress. With these disease-associated mutations, this occurs at a much lower level of stress, resulting in excess stiffening and breaking of the actin cytoskeleton. When the actin-binding domain is removed from αactinin-4, the cytoskeleton networks do not stiffen first, they just break all of a sudden.7 The big question, and what we actively are trying to understand now using a variety of biophysical methods, is to see if this behavior translates into altered biophysical behavior of cells in culture, and we hope eventually also in cells in vivo.
M. Pollak
This observation raises another question, which I think has therapeutic implications. Even though these mutations are an exceptionally rare cause of kidney disease, I think the hope is that this will help us understand more common forms of disease, not just diseases caused by point mutations in α-actinin-4. From an evolutionary point of view, although the actin-binding domain of αactinin-4 normally is hidden, it is highly conserved in evolution. Even Drosophila α-actinins have a hidden actin-binding domain. Why would this exist if all that it does is cause kidney disease when it is exposed to actin through gain-of-function mutations? The answer is that it must have other functions. The notion is that perhaps there are physiologic signals in the cell that are important in controlling whether αactinin-4 has a closed or open confirmation and essentially control the strength of its interactions with actin. This is something we have been exploring using a variety of methods. For example, we are trying to understand whether actin binding is controlled through posttranslational modifications of α-actinin-4. This is a work in progress, but I hope it shows that understanding rare forms of disease can lead us to make interesting hypotheses in understanding more generalized mechanisms of disease. Mutations in INF2 cause familial FSGS.8 This is based on work performed by Elizabeth Brown in my laboratory and now replicated by many groups. There are a large number of independent mutations in a protein called INF2 that progressive chronic kidney disease with proteinuria, typically FSGS on kidney biopsy, but generally characterized by subnephrotic ranges of proteinuria.8 A variety of mutations now have been identified in INF2 and they all cluster in the same domain of the protein, a region called the diaphanous inhibitory domain (DID). What is INF2? It is a member of the formin subfamily of proteins. These proteins regulate the actin cytoskeleton, but in a very different way than α-actinin-4 does. INF2 and other members of the formin family sit on the so-called barbed end of an actin filament and they stimulate oligomerization of actin filaments. INF2 exists as a dimer, as does α-actinin-4. The business end of the molecule that stimulates actin polymerization is near the C terminus of the molecule. An interaction between the domain DID and the diaphanous autoregulatory domain inhibits the actin regulatory function of the INF2 protein. All the mutations that cause human disease are in the DID, where they result in the loss of the normal inhibitory effect of INF2 on actin. However, we do not think that is the sole reason for disease. The end terminus, the DID region of INF2, also interacts with other formin family members. Formin family members are downstream effectors of the Rho family of small guanosine triphosphatases
Genetics of Familial FSGS
469
(GTPases), whose implication in nephrotic syndrome comes from the convergence of evidence from a large number of research groups in this field. Many genetic studies and other cell biological and physiological studies point to the Rho GTPases as critically important for podocyte function. The formins are activated by Rho. Formins, in turn, can stimulate actin polymerization. INF2, the DID of INF2 specifically, interacts directly with the diaphanous autoregulatory domain of a variety of other formins that are downstream of Rho signaling. We believe that this interaction is important in inhibiting the downstream effects of Rho activation, therefore modulating its function. We believe that when this interaction is interrupted by FSGS-causing mutations, Rho signaling essentially progresses unchecked (Fig. 1). Will we knock out INF2 in a zebrafish model, we see severe disease, very disordered podocytes, and decreased survival of zebra fish as well as a phenotype characterized by increasing proteinuria. In conclusion, INF2 basically functions as a modulator of Rho signaling mechanisms and human mutations lead to essentially unopposed Rho signaling and ultimately kidney disease (Fig. 1). It is intriguing to note that, if one looks up all the variants in these two proteins, INF2 and α-actinin-4, in public databases, one sees a large number of variants that are present in what is presumably a normal population. This indicates that the notion that we can just take a patient with FSGS and sequence INF2 and α-actinin-4 and find a variant that has never been identified before and call that the cause of the disease is an oversimplification. I think this is actually a problem in both clinical care and in research studies into relatively rare diseases, understanding what is signal and what is noise within the huge variation in the human genome. Taken together, α-actinin-4, INF2, and TRPC6 mutations account for approximately 50% of the families with late-onset FSGS in a dominant pattern of inheritance. There is also a long and growing list of genes that when mutated cause recessive forms of FSGS and nephrotic syndrome, which we will hear about from Friedhelm Hildebrandt in the next talk. INF2 mutaon Leads to unopposed Rho signaling Rho/mDia Rho/Rock Stress fibers
Figure 1. INF2 modulates actin dynamics by coordinating Rho/ Rac/CDC42 signaling balance. Arp, actin related proteins; mDia, mouse Diaphanous-related formin 1; NWasp, neuronal Wiskott– Aldrich Syndrome protein.
Another point I would like to make, to return to the cancer analogy, is that making a molecular diagnosis is the correct way to determine therapy, and more important than the histopathology. For example, I think everyone would agree that you would not want to treat someone whose biopsy looks like FSGS but has a mutation in a nephronophthisis gene with immunosuppressive therapy. Therefore, I think the notion that molecular diagnoses may be an important component of determining therapy is one that we should really consider seriously. The last gene I would like to talk about is APOL1. As everyone here knows, kidney disease is much more common in people of recent African ancestry than essentially all other ethnicities. The rates of kidney disease have been growing fast in this population.9 There is indeed a four-fold increased risk of nondiabetic hypertension associated with end-stage kidney disease in blacks versus whites. There is also a vastly increased risk of human immunodeficiency virus (HIV)-associated nephropathy in the same population.10 How much of what we call hypertension-associated end-stage renal disease is really FSGS? Is our diagnosis just a function of how and when we are making the diagnosis as opposed to what the actual disease is? In 2008, Kopp et al11 published a really remarkable article showing that a variation of one locus in the human genome seemed to explain most of the high rate of nondiabetic kidney disease in African Americans. A single locus on chromosome 22 was identified using a tool called admixture mapping. This was, I think, a remarkable article and it was very surprising that one locus seemed to drive the majority of this association.11 Building on this work, together with Kopp et al, we were able to identify variants in the APOL1 gene in the region of the genome that Kopp et al had identified as what we believe are the causal drivers of this association.9 We have come to call these two variants G1 and G2. G1 refers to a pair of amino acid changes that almost always are inherited together.9 G2 encodes a sequence near the end terminus of the APOL1 protein. Finally, because these are so common, we no longer call the European form of APOL1 wild-type, we call it G0, for simplicity.9 People who have two copies of G0 are not at increased risk of kidney disease. This is the typical white or non–kidney risk allele. People who inherit two copies of the risk alleles, either G1 or G2, are the individuals at very high risk for kidney disease. These mutant variants, however, are quite common. The allele frequencies of G1 and G2 in this country are on the order of 40%, so these are not rare variants. It is quite unusual for common variants to follow a recessive pattern of inheritance, which makes this story quite unique.9 These associations have been replicated widely, not just in African Americans but in native Africans. Kopp
470
et al10 have shown that the relative odds ratio of FSGS associated with this high-risk genotype is quite high at 17-fold, and is almost 30-fold for patients with HIVassociated nephropathy. Let us also think of this in a different way. If an African American patient walked into your clinic with FSGS, they have almost a 75% chance of having this high-risk APOL1 genotype. In contrast, an African American without FSGS has an approximately 13% chance of having the high-risk genotype. Thus, most African Americans with FSGS have this high-risk APOL1 genotype. This high frequency can lead to what I am calling here a pseudodominant FSGS. One puzzling question is the trend toward disease in diabetic versus nondiabetic African American patients. Recently, Parsa et al12 showed that APOL1 seems to be a risk factor for progression of disease whether or not patients have diabetic or nondiabetic nephropathy. However, the fact that early disease seems to be less common in diabetic patients is still very puzzling, to me at least, because from the point of view of a clinical nephrologist, if a patient presents to the clinic with chronic kidney disease, diabetes, and hypertension, it is hard to identify if the patient has diabetic nephropathy versus hypertensive nephropathy, or the other way around, which is a confounding factor in many studies. One issue that has come up a lot is whether these variants in APOL1 are really the causal driver, the biological drivers of this association with FSGS. We are pretty convinced that they are for a variety of reasons, including the fact that they are two independent mutations that alter the coding sequence of APOL1.9 There are combination hot spots surrounding APOL1, which make it hard to believe that there are other things outside of this region that drive the disease.9 In answer to the question, what is APOL1, these are relatively new genes. They are evolving rapidly. They were introduced into the genome only approximately 60 million years ago, which is not very long ago in evolutionary terms. APOL1 has a role in human resistance to trypanosome infections.9,13-15 This is something that has been known well before we started studying these mutations. APOL1 has three major domains: the APOL1 protein pore-forming domain, the membrane-addressing domain, and the SRA-interacting domain.9 The C terminus of APOL1, the region of APOL1 that harbors these diseaseassociated variants, interacts with a protein called SRA, which is produced by certain subspecies of African trypanosomes.9,13,14 The parental species of African trypanosomes, Trypanosome brucei, enters into the human bloodstream where it ingests high-density lipoprotein particles.15 APOL1 circulates in the blood as part of a highdensity lipoprotein-3 complex. The thinking is that the
M. Pollak
change in the pH level liberates APOL1, which in turn, by unclear mechanisms, leads to swelling and parasitic death. The subspecies Trypanosome rhodesiense can infect human beings, as compared with T brucei, which does not cause disease in human beings because we have APOL1, a form of innate immunity that protects us against T brucei infection.15 In contrast, T rhodesiense makes a protein called SRA, which interacts with APOL1 and inactivates its antitrypanosomal activity.15 In human beings with the G1 or G2 forms of APOL1, the presence of these variants allows APOL1 to evade inactivation by SRA, and this is something we verified by laboratory studies.9 Thus, people with one or two copies of G1 or G2 are protected against T rhodesiense infection.9 If you look in the human genome, we see signals that evolution has been acting on these G1 and G2 variants to bring them to high frequency quickly. If we look at Africa, which is a very heterogeneous continent from a genetic standpoint, these kidney disease risk alleles are more common in western sub-Saharan Africa than in eastern sub-Saharan Africa.9 One of the puzzles that we are trying to understand better is that T rhodesiense is more common in eastern Africa than in western Africa, therefore the distribution of these alleles and the distribution of T brucei versus T rhodesiense do not overlap and it is not clear why. It could be because other infectious agents besides T rhodesiense also are important in the evolutionary pressures involved. It may be that the frequencies of these alleles have changed, and perhaps T rhodesiense could not survive in this part of Africa because of the high frequency of these alleles. There are a lot of possible explanations, but it is going to be challenging to figure this out. In short, these alleles arose very recently in western sub-Saharan Africa, approximately 5,000 years ago, which explains why we see them in high frequency among African Americans, but essentially at 0% frequency in people who do not have recent African ancestry. We think there are probably approximately 4,000 individuals in the United States with this high-risk genotype, which is a substantial number of individuals at high risk for kidney disease. This is a difficult gene to study for a variety of reasons, including the fact that mice do not have an APOL1. This is a primate-only gene, so, for example, we cannot knock it out or mutate it in mice. APOL1 is a very toxic protein. Even the wild-type form seems to be quite toxic when expressed in vitro. Therefore, many challenges remain in understanding APOL1. Finally, one big question is that even though the relative risk of kidney disease in people with these genotypes is high, it is not sufficient to explain disease. Most people walking around with two APOL1 high-risk alleles do not develop kidney disease. Therefore, the question is, what is different about the people who do and
Genetics of Familial FSGS
do not develop kidney disease? One idea is that it is probably important is certain inflammatory pathways.16,17 Friedman described a series of patients who had received interferon treatment for hepatitis C, multiple sclerosis, and so forth, who had developed FSGS or collapsing nephropathy after interferon treatment. We genotyped the DNA from these biopsy specimens and found that everyone had the high-risk APOL1 genotype.16,17 This was pretty striking and suggested that at least one of the ways a patient with a high-risk genotype may be tipped over into overt disease is through stimulation of certain immune pathways.16,17 Let me stop here and thank the people who have contributed to all this work over the years.
FEATURED QUESTION AND DISCUSSION Dr. David Salant: Thank you Martin. Just to return to the α-actinin-4 story, I wonder if there is not a common theme emerging here with the TRP channels and intracellular calcium increase? Several years ago it was shown that in experimental membranous nephropathy with complement-mediated injury that there is a massive influx of calcium in podocytes. I wonder if you are not identifying for us a more common mechanism that might contribute to this collapse of the actin cytoskeleton that is common to all of these diseases that we see? Dr. Martin Pollak: I think that is a great thought. Aberrant calcium leading to cytoskeletal dysregulation through an imbalance in Rho-GTPase signaling is likely a pathway we have converged on through physiology, cell biology, and genetics. Dr. Peter Mundel: Martin, thank you. What I do not understand is, do you think that APOL1 is diseasecausing or disease-modifying? If we treat the underlying disease, such as HIV nephropathy, for example, do we need to care or worry about APOL1? Also, related to that: let us say that we have a patient who carries the high-risk mutation, develops end-stage kidney disease, and is lucky to receive a transplant. If you place a G0 kidney in this patient, is this patient safe? Finally, the third question: do you think that APOL1 should be targeted and depleted from serum as a protective measure? Dr. Martin Pollak: With regard to your last question, I do not know about depletion, but the answer to the second question may inform the answer to that question. The transplant issue, I did not show this but there are studies from Barry Freedman’s group and studies we have performed in collaboration with Anil Chandraker that suggest that in the transplant setting it is the kidney that is at high risk of disease, not the recipient. Therefore, if you place a G0 kidney into a G1 person, that kidney is not at high risk. If you do the reverse, risk travels with the kidney. This suggests that it probably is not the
471
circulating APOL1, but it is the APOL1 that is expressed in cells that is important. That is why I do not think depleting it from the serum is likely to be an effective therapy. Finally, let me answer your question regarding the disease modifier. One answer might be that maybe this is largely semantics. Does smoking cigarettes cause lung cancer? Not everyone who smokes cigarettes develops lung cancer. Or, does obesity cause diabetes? Does diabetes cause nephropathy? This question can be asked about any risk factor. You could argue that in many cases, what matters is the degree of penetrance. For example, we think of α-actinin-4–associated FSGS as a dominant Mendelian disease, but there are occasional people with INF2 or ACTN4 mutations who do not even have proteinuria even when they are relatively old. Therefore, we cannot even say that these genes cause disease, but we can say that they probably set up your kidneys to be susceptible to disease.
REFERENCES 1. Mathis BJ, et al. A locus for inherited focal segmental glomerulosclerosis maps to chromosome 19q13. Kidney Int. 1998;53:282-6. 2. Dandapani SV, et al. Alpha-actinin-4 is required for normal podocyte adhesion. J Biol Chem. 2007;282:467-77. 3. Henderson JM, et al. Mice with altered alpha-actinin-4 expression have distinct morphologic patterns of glomerular disease. Kidney Int. 2008;73:741-50. 4. Kaplan JM, et al. Mutations in ACTN4, encoding alpha-actinin4, cause familial focal segmental glomerulosclerosis. Nat Genet. 2000;24:251-6. 5. Weins A, et al. Mutational and biological analysis of alphaactinin-4 in focal segmental glomerulosclerosis. J Am Soc Nephrol. 2005;16:3694-701. 6. Weins A, et al. Disease-associated mutant alpha-actinin-4 reveals a mechanism for regulating its F-actin-binding affinity. Proc Natl Acad Sci U S A. 2007;104:16080-5. 7. Ward SM, et al. Dynamic viscoelasticity of actin cross-linked with wild-type and disease-causing mutant alpha-actinin-4. Biophys J. 2008;95:4915-23. 8. Brown EJ, et al. Mutations in the formin gene INF2 cause focal segmental glomerulosclerosis. Nat Genet. 2010;42:72-6. 9. Genovese G, et al. Association of trypanolytic ApoL1 variants with kidney disease in African Americans. Science. 2010;329: 841-5. 10. Kopp JB, et al. APOL1 genetic variants in focal segmental glomerulosclerosis and HIV-associated nephropathy. J Am Soc Nephrol. 2011;22:2129-37. 11. Kopp JB, et al. MYH9 is a major-effect risk gene for focal segmental glomerulosclerosis. Nat Genet. 2008;40:1175-84. 12. Parsa A, et al. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med. 2013;369:2183-96. 13. Vanwalleghem G, et al. Coupling of lysosomal and mitochondrial membrane permeabilization in trypanolysis by APOL1. Nat Commun. 2015;6:8078. 14. Lecordier L, et al. Adaptation of Trypanosoma rhodesiense to hypohaptoglobinaemic serum requires transcription of the APOL1 resistance gene in a RNA polymerase I locus. Mol Microbiol. 2015;97:397-407. 15. Vanhamme L, et al. Apolipoprotein L-I is the trypanosome lytic factor of human serum. Nature. 2003;422:83-7.
472
16. Friedman DJ, et al. Population-based risk assessment of APOL1 on renal disease. J Am Soc Nephrol. 2011;22:2098-105. 17. Friedman DJ, Pollak MR. Genetics of kidney failure and the evolving story of APOL1. J Clin Invest. 2011;121:3367-74.
Genetics of Kidney Diseases Friedhelm Hildebrandt, MD Keywords: Recessive mutations, ADCK4, COQ6, ARHGDIA, KANK, MYO1E
I would like to start out by telling the story of a patient, a 3-year-old boy who developed nephrotic syndrome with generalized edema. His nephrotic syndrome was steroidresistant. Renal histology showed focal segmental glomerulosclerosis (FSGS), and, as is typical for FSGS, he had to enter into the chronic dialysis and transplantation program within a few years after presentation. If we ask ourselves, why does this boy have to go through this trouble? The molecular answer is that he has a mutation in the podocin gene. This, in my mind, is one of the strongest cause-and-effect relationships that we can see in clinical medicine. A single mutation in the 3.3 billion bases we have in our genome is, in and of itself, sufficient to cause FSGS. This is in fact typical for recessive monogenic diseases, otherwise known as single-gene disorders or Mendelian disorders. They are defined by the fact that in a given patient, the disease is caused by a mutation in only 1 of approximately 22,000 genes that we have in our genome. The fact that we call this monogenic does not preclude the fact that in different patients different genes may cause a similar disease. For instance, we know that FSGS can be caused by podocin or nephrin mutations. There are approximately 23 known recessive causes of nephrotic syndrome right now, and approximately 7 dominant genes. If we look at the causes of chronic kidney disease in the first 20 years of life in the North American Pediatric Renal Transplant Cooperative Study, in which approximately 9,000 patients who had chronic kidney disease were studied, then we see that the cause of chronic kidney disease in approximately 50% of these individuals was congenital anomalies of the kidneys, but 15% was steroid-resistant nephrotic syndrome (SRNS).1 Financial disclosure and conflict of interest statements: none. Department of Medicine, Division of Nephrology, Children’s Hospital and Harvard Medical School, Boston, MA. Address reprint requests to: Friedhelm Hildebrandt, MD, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail: [email protected]
F. Hildebrandt
Only in the past few years did it become apparent that there actually are multiple monogenic causes of these diseases in nephrotic syndrome. There are now 30 known causes and the number is increasing rapidly. This is similar to the development of the field of cystic kidney disease, in which approximately 20 years ago we wanted to find the gene for nephronophthisis, and we now have found 95 genes. In addition, in tubulopathies this number is similarly very high.2,3 As you know, the identification of single-gene causes of nephrotic syndrome such as podocin have put the podocyte at the center of the pathogenesis of FSGS. Figure 1 includes all the mutations that my laboratory has found as recessive genes leading to FSGS. What is fascinating is that, in the past few years, we have found that many of these proteins coalesce not only to tight protein complexes but also to pathogenic pathways, and, increasingly, as more proteins become known, many of them map back to the same pathway. In other words, each of these proteins is, in and of itself, necessary to avoid the disease phenotype of SRNS. To show this, we asked a very simple question. Given that there are approximately 30 monogenic causes of steroid-resistant nephrotic syndrome, what is the percentage of the nephrotic syndrome cases that have a singlegene cause? I use the term cause here because, in recessive diseases, if there is a homozygous truncating mutation or two compound truncating mutations, or if you have functional evidence that a mutation abrogates function, then you can infer cause. Because I am a pediatric nephrologist, my cohort is younger than 25 years of age. Over about 10 years we accrued samples from 1,718 different families with SRNS all over the world, and we sequenced these for all the known genes. Specifically, we looked into 27 genes, 20 of which were recessive and 7 of which were dominant. We have 1,200 families from Europe, approximately 360 families from the United States, 125 families from India, and 30 and 25 families from Asia and Australia, respectively. When we finished this analysis, we found, stunningly, that in 30% of these families we actually could find a causative mutation. Otherwise put, in these patients who manifest with SRNS before 25 years of age, we found mutations in 21 of 27 genes tested. The most frequent mutations were found in podocin, which was 10% of this cohort. The second most frequent was nephrin at 7%, Wilms Tumor (WT1) gene at 5%, Phospholipase C Epsilon (PLC1) gene at 2%, and then 17 other genes had to share into another 6%. One can reasonably expect that this number will increase rapidly as we and others nowadays can find a novel diseasecausing gene at a rate of every 2 to 3 months. If you look at the age distribution, you can see in infants with nephrotic syndrome that there are mutations in approximately 60% of patients. It is inversely age-
Dr. Demetri: In answer to your first question, there is a ton of work that has been performed in cell-cycle inhibitors of various types and it has been remarkably fruitless. However, that also may be because it was not performed in the right patient population. For some reason, the minute they go into hormone-receptor–positive breast cancer, CDK4 inhibitors have breakthrough designations, which is extraordinary. Part of the problem again is matching the drug with the right patient, validating the targets, and figuring out how to hit the target effectively. We have to be careful not to jump to conclusions. We often say the drug failed, but, instead, it may just be the wrong patient. To your second question about how many drugs may be needed for personalized cancer treatment: who would have guessed that the same drug would treat leukemia and solid sarcoma? This again suggests that we must bin tumors by mechanism and we still do not know how to do that well. I suspect you will be binning things by mechanism of glomerulonephritis, perhaps by mechanisms that also apply to interstitial pulmonary fibrosis (IPF). The world has just had a couple of really great targeted drugs approved for IPF. IPF is worse than most cancers and IPF could have mechanisms that really affect what you are working on in nephrology.
467
12. Blanke CD, et al. Long-term results from a randomized phase II trial of standard- versus higher-dose imatinib mesylate for patients with unresectable or metastatic gastrointestinal stromal tumors expressing KIT. J Clin Oncol. 2008;26:620-5. 13. Demetri GD, et al. Molecular target modulation, imaging, and clinical evaluation of gastrointestinal stromal tumor patients treated with sunitinib malate after imatinib failure. Clin Cancer Res. 2009;15:5902-9. 14. Yap TA, et al. First-in-human phase I trial of two schedules of OSI-930, a novel multikinase inhibitor, incorporating translational proof-of-mechanism studies. Clin Cancer Res. 2013; 19:909-19. 15. Patrikidou A, et al. Influence of imatinib interruption and rechallenge on the residual disease in patients with advanced GIST: results of the BFR14 prospective French Sarcoma Group randomised, phase III trial. Ann Oncol. 2013;24:1087-93. 16. Demetri GD. Identification and treatment of chemoresistant inoperable or metastatic GIST: experience with the selective tyrosine kinase inhibitor imatinib mesylate (STI571). Eur J Cancer. 2002;38(Suppl 5):S52-9.
Genetics of Familial FSGS Martin Pollak, MD Keywords: Autosomal dominant, TRPC6, actinin-4, APOL1, HIV nephropathy, hypertensive nephropathy
INF2,
REFERENCES 1. O’Bryan RM, et al. Phase II evaluation of Adriamycin in human neoplasia. Cancer. 1973;32:1-8. 2. Grier HE, et al. Addition of ifosfamide and etoposide to standard chemotherapy for Ewing’s sarcoma and primitive neuroectodermal tumor of bone. N Engl J Med. 2003; 348:694-701. 3. Rous P. A transmissible avian neoplasm. (Sarcoma of the common fowl) by Peyton Rous, M.D., Experimental Medicine for Sept. 1, 1910, vol. 12, pp. 696-705. J Exp Med. 1979;150: 738-53. 4. Weisberg E, et al. Effects of PKC412, nilotinib, and imatinib against GIST-associated PDGFRA mutants with differential imatinib sensitivity. Gastroenterology. 2006;131:1734-42. 5. Bauer S, et al. KIT oncogenic signaling mechanisms in imatinib-resistant gastrointestinal stromal tumor: PI3-kinase/ AKT is a crucial survival pathway. Oncogene. 2007;26:7560-8. 6. Zhu MJ, et al. KIT oncoprotein interactions in gastrointestinal stromal tumors: therapeutic relevance. Oncogene. 2007;26: 6386-95. 7. Duensing A, et al. Mechanisms of oncogenic KIT signal transduction in primary gastrointestinal stromal tumors (GISTs). Oncogene. 2004;23:3999-4006. 8. Singer S, et al. Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol. 2002;20:3898-905. 9. Rubin BP, et al. KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res. 2001;61:8118-21. 10. Tuveson DA, et al. STI571 inactivation of the gastrointestinal stromal tumor c-KIT oncoprotein: biological and clinical implications. Oncogene. 2001;20:5054-8. 11. Lux ML, et al. KIT extracellular and kinase domain mutations in gastrointestinal stromal tumors. Am J Pathol. 2000; 156:791-5.
I am going to talk about familial focal segmental glomerulosclerosis (FSGS) today. As everyone realizes, FSGS is a histopathologic entity, a pattern of injury that is seen on a kidney biopsy. Nephrotic syndrome is just that, a syndrome. Hypertensive kidney disease also shares features with FSGS. There is considerable overlap between all of these entities. Even though these are just names, I think the fact that we do not have really good ways for the clinicians and the pathologists and patients to talk to each other about these entities does cause some problems. Nevertheless, today we will focus on familial FSGS. As a historical note, while I was still in training, I came across a family with inherited kidney disease in Oklahoma, in a report published in 1998.1 I flew out to a little town outside of Tulsa, Oklahoma, and spent 2 days rounding up members of this family and taking their blood pressures, talking to them, and obtaining urine and blood samples.
Financial disclosure and conflict of interest statements: none. Department of Medicine, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA. Address reprint requests to: Martin Pollak, MD, Department of Medicine, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Farr 8, Boston, MA 02215. E-mail: mpollak@bidmc. harvard.edu
468
They had a dominant inheritance of a phenotype that was characterized by proteinuria. A few individuals had biopsy-proven FSGS. We started to perform genetic studies to try to identify the underlying gene and also scour the world for other similar families. The advantage of this kind of approach is that one does not need to worry about whether the amount of proteinuria or the histology looks exactly the same from person to person. It alleviates the concern about having a good way to classify patients diagnostically. We therefore performed what at the time were fairly laborious genetic studies, and ultimately this led us to the identification of mutations in the α-actinin-4 gene as the cause of the disease in two families with FSGS, as well as several others.2-4 All of the mutations that segregated with disease in these families were in the actin-binding domain of α-actinin-4, matching well with the best described function of α-actinin-4, which is to cross-link actin filaments.2-4 Indeed, the actin-binding domains associated with FSGS-causing mutations have a much greater affinity to actin filaments than wild type α-actinin-4. I therefore suggest that these mutations are causing a dominant effect by a gain-of-function change in their behavior. This led us to start thinking about what may be going on structurally and we ultimately found that the actinbinding domain could exist as both a closed and an open confirmation.5,6 We therefore concluded that mutations that were associated with kidney disease and led to increased actin binding interfered with the transition between the closed and open states.5,6 We hypothesized further that the presence of diseaseassociated mutations may alter the behavior of the podocyte cytoskeleton. We have been very interested in what this means for the biophysical properties of the podocyte and how we can connect in vitro cytoskeletal behavior with how the podocyte actually behaves. We are making slow progress toward this. If we create reconstituted α-actinin-4/actin networks and perform rheologic studies, in work performed in collaboration with colleagues from the Physics Department at Harvard, we note that when we apply increasing stress to wild-type α-actinin-4/actin networks, we find increasing stiffness of the network until finally it breaks at a certain level of stress. With these disease-associated mutations, this occurs at a much lower level of stress, resulting in excess stiffening and breaking of the actin cytoskeleton. When the actin-binding domain is removed from αactinin-4, the cytoskeleton networks do not stiffen first, they just break all of a sudden.7 The big question, and what we actively are trying to understand now using a variety of biophysical methods, is to see if this behavior translates into altered biophysical behavior of cells in culture, and we hope eventually also in cells in vivo.
M. Pollak
This observation raises another question, which I think has therapeutic implications. Even though these mutations are an exceptionally rare cause of kidney disease, I think the hope is that this will help us understand more common forms of disease, not just diseases caused by point mutations in α-actinin-4. From an evolutionary point of view, although the actin-binding domain of αactinin-4 normally is hidden, it is highly conserved in evolution. Even Drosophila α-actinins have a hidden actin-binding domain. Why would this exist if all that it does is cause kidney disease when it is exposed to actin through gain-of-function mutations? The answer is that it must have other functions. The notion is that perhaps there are physiologic signals in the cell that are important in controlling whether αactinin-4 has a closed or open confirmation and essentially control the strength of its interactions with actin. This is something we have been exploring using a variety of methods. For example, we are trying to understand whether actin binding is controlled through posttranslational modifications of α-actinin-4. This is a work in progress, but I hope it shows that understanding rare forms of disease can lead us to make interesting hypotheses in understanding more generalized mechanisms of disease. Mutations in INF2 cause familial FSGS.8 This is based on work performed by Elizabeth Brown in my laboratory and now replicated by many groups. There are a large number of independent mutations in a protein called INF2 that progressive chronic kidney disease with proteinuria, typically FSGS on kidney biopsy, but generally characterized by subnephrotic ranges of proteinuria.8 A variety of mutations now have been identified in INF2 and they all cluster in the same domain of the protein, a region called the diaphanous inhibitory domain (DID). What is INF2? It is a member of the formin subfamily of proteins. These proteins regulate the actin cytoskeleton, but in a very different way than α-actinin-4 does. INF2 and other members of the formin family sit on the so-called barbed end of an actin filament and they stimulate oligomerization of actin filaments. INF2 exists as a dimer, as does α-actinin-4. The business end of the molecule that stimulates actin polymerization is near the C terminus of the molecule. An interaction between the domain DID and the diaphanous autoregulatory domain inhibits the actin regulatory function of the INF2 protein. All the mutations that cause human disease are in the DID, where they result in the loss of the normal inhibitory effect of INF2 on actin. However, we do not think that is the sole reason for disease. The end terminus, the DID region of INF2, also interacts with other formin family members. Formin family members are downstream effectors of the Rho family of small guanosine triphosphatases
Genetics of Familial FSGS
469
(GTPases), whose implication in nephrotic syndrome comes from the convergence of evidence from a large number of research groups in this field. Many genetic studies and other cell biological and physiological studies point to the Rho GTPases as critically important for podocyte function. The formins are activated by Rho. Formins, in turn, can stimulate actin polymerization. INF2, the DID of INF2 specifically, interacts directly with the diaphanous autoregulatory domain of a variety of other formins that are downstream of Rho signaling. We believe that this interaction is important in inhibiting the downstream effects of Rho activation, therefore modulating its function. We believe that when this interaction is interrupted by FSGS-causing mutations, Rho signaling essentially progresses unchecked (Fig. 1). Will we knock out INF2 in a zebrafish model, we see severe disease, very disordered podocytes, and decreased survival of zebra fish as well as a phenotype characterized by increasing proteinuria. In conclusion, INF2 basically functions as a modulator of Rho signaling mechanisms and human mutations lead to essentially unopposed Rho signaling and ultimately kidney disease (Fig. 1). It is intriguing to note that, if one looks up all the variants in these two proteins, INF2 and α-actinin-4, in public databases, one sees a large number of variants that are present in what is presumably a normal population. This indicates that the notion that we can just take a patient with FSGS and sequence INF2 and α-actinin-4 and find a variant that has never been identified before and call that the cause of the disease is an oversimplification. I think this is actually a problem in both clinical care and in research studies into relatively rare diseases, understanding what is signal and what is noise within the huge variation in the human genome. Taken together, α-actinin-4, INF2, and TRPC6 mutations account for approximately 50% of the families with late-onset FSGS in a dominant pattern of inheritance. There is also a long and growing list of genes that when mutated cause recessive forms of FSGS and nephrotic syndrome, which we will hear about from Friedhelm Hildebrandt in the next talk. INF2 mutaon Leads to unopposed Rho signaling Rho/mDia Rho/Rock Stress fibers
Figure 1. INF2 modulates actin dynamics by coordinating Rho/ Rac/CDC42 signaling balance. Arp, actin related proteins; mDia, mouse Diaphanous-related formin 1; NWasp, neuronal Wiskott– Aldrich Syndrome protein.
Another point I would like to make, to return to the cancer analogy, is that making a molecular diagnosis is the correct way to determine therapy, and more important than the histopathology. For example, I think everyone would agree that you would not want to treat someone whose biopsy looks like FSGS but has a mutation in a nephronophthisis gene with immunosuppressive therapy. Therefore, I think the notion that molecular diagnoses may be an important component of determining therapy is one that we should really consider seriously. The last gene I would like to talk about is APOL1. As everyone here knows, kidney disease is much more common in people of recent African ancestry than essentially all other ethnicities. The rates of kidney disease have been growing fast in this population.9 There is indeed a four-fold increased risk of nondiabetic hypertension associated with end-stage kidney disease in blacks versus whites. There is also a vastly increased risk of human immunodeficiency virus (HIV)-associated nephropathy in the same population.10 How much of what we call hypertension-associated end-stage renal disease is really FSGS? Is our diagnosis just a function of how and when we are making the diagnosis as opposed to what the actual disease is? In 2008, Kopp et al11 published a really remarkable article showing that a variation of one locus in the human genome seemed to explain most of the high rate of nondiabetic kidney disease in African Americans. A single locus on chromosome 22 was identified using a tool called admixture mapping. This was, I think, a remarkable article and it was very surprising that one locus seemed to drive the majority of this association.11 Building on this work, together with Kopp et al, we were able to identify variants in the APOL1 gene in the region of the genome that Kopp et al had identified as what we believe are the causal drivers of this association.9 We have come to call these two variants G1 and G2. G1 refers to a pair of amino acid changes that almost always are inherited together.9 G2 encodes a sequence near the end terminus of the APOL1 protein. Finally, because these are so common, we no longer call the European form of APOL1 wild-type, we call it G0, for simplicity.9 People who have two copies of G0 are not at increased risk of kidney disease. This is the typical white or non–kidney risk allele. People who inherit two copies of the risk alleles, either G1 or G2, are the individuals at very high risk for kidney disease. These mutant variants, however, are quite common. The allele frequencies of G1 and G2 in this country are on the order of 40%, so these are not rare variants. It is quite unusual for common variants to follow a recessive pattern of inheritance, which makes this story quite unique.9 These associations have been replicated widely, not just in African Americans but in native Africans. Kopp
470
et al10 have shown that the relative odds ratio of FSGS associated with this high-risk genotype is quite high at 17-fold, and is almost 30-fold for patients with HIVassociated nephropathy. Let us also think of this in a different way. If an African American patient walked into your clinic with FSGS, they have almost a 75% chance of having this high-risk APOL1 genotype. In contrast, an African American without FSGS has an approximately 13% chance of having the high-risk genotype. Thus, most African Americans with FSGS have this high-risk APOL1 genotype. This high frequency can lead to what I am calling here a pseudodominant FSGS. One puzzling question is the trend toward disease in diabetic versus nondiabetic African American patients. Recently, Parsa et al12 showed that APOL1 seems to be a risk factor for progression of disease whether or not patients have diabetic or nondiabetic nephropathy. However, the fact that early disease seems to be less common in diabetic patients is still very puzzling, to me at least, because from the point of view of a clinical nephrologist, if a patient presents to the clinic with chronic kidney disease, diabetes, and hypertension, it is hard to identify if the patient has diabetic nephropathy versus hypertensive nephropathy, or the other way around, which is a confounding factor in many studies. One issue that has come up a lot is whether these variants in APOL1 are really the causal driver, the biological drivers of this association with FSGS. We are pretty convinced that they are for a variety of reasons, including the fact that they are two independent mutations that alter the coding sequence of APOL1.9 There are combination hot spots surrounding APOL1, which make it hard to believe that there are other things outside of this region that drive the disease.9 In answer to the question, what is APOL1, these are relatively new genes. They are evolving rapidly. They were introduced into the genome only approximately 60 million years ago, which is not very long ago in evolutionary terms. APOL1 has a role in human resistance to trypanosome infections.9,13-15 This is something that has been known well before we started studying these mutations. APOL1 has three major domains: the APOL1 protein pore-forming domain, the membrane-addressing domain, and the SRA-interacting domain.9 The C terminus of APOL1, the region of APOL1 that harbors these diseaseassociated variants, interacts with a protein called SRA, which is produced by certain subspecies of African trypanosomes.9,13,14 The parental species of African trypanosomes, Trypanosome brucei, enters into the human bloodstream where it ingests high-density lipoprotein particles.15 APOL1 circulates in the blood as part of a highdensity lipoprotein-3 complex. The thinking is that the
M. Pollak
change in the pH level liberates APOL1, which in turn, by unclear mechanisms, leads to swelling and parasitic death. The subspecies Trypanosome rhodesiense can infect human beings, as compared with T brucei, which does not cause disease in human beings because we have APOL1, a form of innate immunity that protects us against T brucei infection.15 In contrast, T rhodesiense makes a protein called SRA, which interacts with APOL1 and inactivates its antitrypanosomal activity.15 In human beings with the G1 or G2 forms of APOL1, the presence of these variants allows APOL1 to evade inactivation by SRA, and this is something we verified by laboratory studies.9 Thus, people with one or two copies of G1 or G2 are protected against T rhodesiense infection.9 If you look in the human genome, we see signals that evolution has been acting on these G1 and G2 variants to bring them to high frequency quickly. If we look at Africa, which is a very heterogeneous continent from a genetic standpoint, these kidney disease risk alleles are more common in western sub-Saharan Africa than in eastern sub-Saharan Africa.9 One of the puzzles that we are trying to understand better is that T rhodesiense is more common in eastern Africa than in western Africa, therefore the distribution of these alleles and the distribution of T brucei versus T rhodesiense do not overlap and it is not clear why. It could be because other infectious agents besides T rhodesiense also are important in the evolutionary pressures involved. It may be that the frequencies of these alleles have changed, and perhaps T rhodesiense could not survive in this part of Africa because of the high frequency of these alleles. There are a lot of possible explanations, but it is going to be challenging to figure this out. In short, these alleles arose very recently in western sub-Saharan Africa, approximately 5,000 years ago, which explains why we see them in high frequency among African Americans, but essentially at 0% frequency in people who do not have recent African ancestry. We think there are probably approximately 4,000 individuals in the United States with this high-risk genotype, which is a substantial number of individuals at high risk for kidney disease. This is a difficult gene to study for a variety of reasons, including the fact that mice do not have an APOL1. This is a primate-only gene, so, for example, we cannot knock it out or mutate it in mice. APOL1 is a very toxic protein. Even the wild-type form seems to be quite toxic when expressed in vitro. Therefore, many challenges remain in understanding APOL1. Finally, one big question is that even though the relative risk of kidney disease in people with these genotypes is high, it is not sufficient to explain disease. Most people walking around with two APOL1 high-risk alleles do not develop kidney disease. Therefore, the question is, what is different about the people who do and
Genetics of Familial FSGS
do not develop kidney disease? One idea is that it is probably important is certain inflammatory pathways.16,17 Friedman described a series of patients who had received interferon treatment for hepatitis C, multiple sclerosis, and so forth, who had developed FSGS or collapsing nephropathy after interferon treatment. We genotyped the DNA from these biopsy specimens and found that everyone had the high-risk APOL1 genotype.16,17 This was pretty striking and suggested that at least one of the ways a patient with a high-risk genotype may be tipped over into overt disease is through stimulation of certain immune pathways.16,17 Let me stop here and thank the people who have contributed to all this work over the years.
FEATURED QUESTION AND DISCUSSION Dr. David Salant: Thank you Martin. Just to return to the α-actinin-4 story, I wonder if there is not a common theme emerging here with the TRP channels and intracellular calcium increase? Several years ago it was shown that in experimental membranous nephropathy with complement-mediated injury that there is a massive influx of calcium in podocytes. I wonder if you are not identifying for us a more common mechanism that might contribute to this collapse of the actin cytoskeleton that is common to all of these diseases that we see? Dr. Martin Pollak: I think that is a great thought. Aberrant calcium leading to cytoskeletal dysregulation through an imbalance in Rho-GTPase signaling is likely a pathway we have converged on through physiology, cell biology, and genetics. Dr. Peter Mundel: Martin, thank you. What I do not understand is, do you think that APOL1 is diseasecausing or disease-modifying? If we treat the underlying disease, such as HIV nephropathy, for example, do we need to care or worry about APOL1? Also, related to that: let us say that we have a patient who carries the high-risk mutation, develops end-stage kidney disease, and is lucky to receive a transplant. If you place a G0 kidney in this patient, is this patient safe? Finally, the third question: do you think that APOL1 should be targeted and depleted from serum as a protective measure? Dr. Martin Pollak: With regard to your last question, I do not know about depletion, but the answer to the second question may inform the answer to that question. The transplant issue, I did not show this but there are studies from Barry Freedman’s group and studies we have performed in collaboration with Anil Chandraker that suggest that in the transplant setting it is the kidney that is at high risk of disease, not the recipient. Therefore, if you place a G0 kidney into a G1 person, that kidney is not at high risk. If you do the reverse, risk travels with the kidney. This suggests that it probably is not the
471
circulating APOL1, but it is the APOL1 that is expressed in cells that is important. That is why I do not think depleting it from the serum is likely to be an effective therapy. Finally, let me answer your question regarding the disease modifier. One answer might be that maybe this is largely semantics. Does smoking cigarettes cause lung cancer? Not everyone who smokes cigarettes develops lung cancer. Or, does obesity cause diabetes? Does diabetes cause nephropathy? This question can be asked about any risk factor. You could argue that in many cases, what matters is the degree of penetrance. For example, we think of α-actinin-4–associated FSGS as a dominant Mendelian disease, but there are occasional people with INF2 or ACTN4 mutations who do not even have proteinuria even when they are relatively old. Therefore, we cannot even say that these genes cause disease, but we can say that they probably set up your kidneys to be susceptible to disease.
REFERENCES 1. Mathis BJ, et al. A locus for inherited focal segmental glomerulosclerosis maps to chromosome 19q13. Kidney Int. 1998;53:282-6. 2. Dandapani SV, et al. Alpha-actinin-4 is required for normal podocyte adhesion. J Biol Chem. 2007;282:467-77. 3. Henderson JM, et al. Mice with altered alpha-actinin-4 expression have distinct morphologic patterns of glomerular disease. Kidney Int. 2008;73:741-50. 4. Kaplan JM, et al. Mutations in ACTN4, encoding alpha-actinin4, cause familial focal segmental glomerulosclerosis. Nat Genet. 2000;24:251-6. 5. Weins A, et al. Mutational and biological analysis of alphaactinin-4 in focal segmental glomerulosclerosis. J Am Soc Nephrol. 2005;16:3694-701. 6. Weins A, et al. Disease-associated mutant alpha-actinin-4 reveals a mechanism for regulating its F-actin-binding affinity. Proc Natl Acad Sci U S A. 2007;104:16080-5. 7. Ward SM, et al. Dynamic viscoelasticity of actin cross-linked with wild-type and disease-causing mutant alpha-actinin-4. Biophys J. 2008;95:4915-23. 8. Brown EJ, et al. Mutations in the formin gene INF2 cause focal segmental glomerulosclerosis. Nat Genet. 2010;42:72-6. 9. Genovese G, et al. Association of trypanolytic ApoL1 variants with kidney disease in African Americans. Science. 2010;329: 841-5. 10. Kopp JB, et al. APOL1 genetic variants in focal segmental glomerulosclerosis and HIV-associated nephropathy. J Am Soc Nephrol. 2011;22:2129-37. 11. Kopp JB, et al. MYH9 is a major-effect risk gene for focal segmental glomerulosclerosis. Nat Genet. 2008;40:1175-84. 12. Parsa A, et al. APOL1 risk variants, race, and progression of chronic kidney disease. N Engl J Med. 2013;369:2183-96. 13. Vanwalleghem G, et al. Coupling of lysosomal and mitochondrial membrane permeabilization in trypanolysis by APOL1. Nat Commun. 2015;6:8078. 14. Lecordier L, et al. Adaptation of Trypanosoma rhodesiense to hypohaptoglobinaemic serum requires transcription of the APOL1 resistance gene in a RNA polymerase I locus. Mol Microbiol. 2015;97:397-407. 15. Vanhamme L, et al. Apolipoprotein L-I is the trypanosome lytic factor of human serum. Nature. 2003;422:83-7.
472
16. Friedman DJ, et al. Population-based risk assessment of APOL1 on renal disease. J Am Soc Nephrol. 2011;22:2098-105. 17. Friedman DJ, Pollak MR. Genetics of kidney failure and the evolving story of APOL1. J Clin Invest. 2011;121:3367-74.
Genetics of Kidney Diseases Friedhelm Hildebrandt, MD Keywords: Recessive mutations, ADCK4, COQ6, ARHGDIA, KANK, MYO1E
I would like to start out by telling the story of a patient, a 3-year-old boy who developed nephrotic syndrome with generalized edema. His nephrotic syndrome was steroidresistant. Renal histology showed focal segmental glomerulosclerosis (FSGS), and, as is typical for FSGS, he had to enter into the chronic dialysis and transplantation program within a few years after presentation. If we ask ourselves, why does this boy have to go through this trouble? The molecular answer is that he has a mutation in the podocin gene. This, in my mind, is one of the strongest cause-and-effect relationships that we can see in clinical medicine. A single mutation in the 3.3 billion bases we have in our genome is, in and of itself, sufficient to cause FSGS. This is in fact typical for recessive monogenic diseases, otherwise known as single-gene disorders or Mendelian disorders. They are defined by the fact that in a given patient, the disease is caused by a mutation in only 1 of approximately 22,000 genes that we have in our genome. The fact that we call this monogenic does not preclude the fact that in different patients different genes may cause a similar disease. For instance, we know that FSGS can be caused by podocin or nephrin mutations. There are approximately 23 known recessive causes of nephrotic syndrome right now, and approximately 7 dominant genes. If we look at the causes of chronic kidney disease in the first 20 years of life in the North American Pediatric Renal Transplant Cooperative Study, in which approximately 9,000 patients who had chronic kidney disease were studied, then we see that the cause of chronic kidney disease in approximately 50% of these individuals was congenital anomalies of the kidneys, but 15% was steroid-resistant nephrotic syndrome (SRNS).1 Financial disclosure and conflict of interest statements: none. Department of Medicine, Division of Nephrology, Children’s Hospital and Harvard Medical School, Boston, MA. Address reprint requests to: Friedhelm Hildebrandt, MD, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail: [email protected]
F. Hildebrandt
Only in the past few years did it become apparent that there actually are multiple monogenic causes of these diseases in nephrotic syndrome. There are now 30 known causes and the number is increasing rapidly. This is similar to the development of the field of cystic kidney disease, in which approximately 20 years ago we wanted to find the gene for nephronophthisis, and we now have found 95 genes. In addition, in tubulopathies this number is similarly very high.2,3 As you know, the identification of single-gene causes of nephrotic syndrome such as podocin have put the podocyte at the center of the pathogenesis of FSGS. Figure 1 includes all the mutations that my laboratory has found as recessive genes leading to FSGS. What is fascinating is that, in the past few years, we have found that many of these proteins coalesce not only to tight protein complexes but also to pathogenic pathways, and, increasingly, as more proteins become known, many of them map back to the same pathway. In other words, each of these proteins is, in and of itself, necessary to avoid the disease phenotype of SRNS. To show this, we asked a very simple question. Given that there are approximately 30 monogenic causes of steroid-resistant nephrotic syndrome, what is the percentage of the nephrotic syndrome cases that have a singlegene cause? I use the term cause here because, in recessive diseases, if there is a homozygous truncating mutation or two compound truncating mutations, or if you have functional evidence that a mutation abrogates function, then you can infer cause. Because I am a pediatric nephrologist, my cohort is younger than 25 years of age. Over about 10 years we accrued samples from 1,718 different families with SRNS all over the world, and we sequenced these for all the known genes. Specifically, we looked into 27 genes, 20 of which were recessive and 7 of which were dominant. We have 1,200 families from Europe, approximately 360 families from the United States, 125 families from India, and 30 and 25 families from Asia and Australia, respectively. When we finished this analysis, we found, stunningly, that in 30% of these families we actually could find a causative mutation. Otherwise put, in these patients who manifest with SRNS before 25 years of age, we found mutations in 21 of 27 genes tested. The most frequent mutations were found in podocin, which was 10% of this cohort. The second most frequent was nephrin at 7%, Wilms Tumor (WT1) gene at 5%, Phospholipase C Epsilon (PLC1) gene at 2%, and then 17 other genes had to share into another 6%. One can reasonably expect that this number will increase rapidly as we and others nowadays can find a novel diseasecausing gene at a rate of every 2 to 3 months. If you look at the age distribution, you can see in infants with nephrotic syndrome that there are mutations in approximately 60% of patients. It is inversely age-