Slide 1

Thank you. Thank you very much for the kind introduction and this is really my great honour to be here and I’d like to thank for giving me a chance to show my recent studies on klotho. My laboratory has been working on the klotho gene for many years and we recently found that a klotho gene is deeply involved in the phosphate/calcium metabolism. So, I’m going to show some of our recent studies today.

Slide 2

By the way, the gene was named after a Greek goddess Cloto who spun the thread of life.

Slide 3

The klotho gene was originally identified as a gene mutated in the klotho mouse. The klotho mouse is actually generated serendipitously in the transgenic mouse experiment for other purposes. The klotho mouse is one of the transgenic mouse lines that do not express the transgene so therefore, is supposed to be useless. However, we realised that homozygous for the transgene developed multiple aging-like phenotypes around 4 weeks of age due to insertional mutation of the transgene.

Slide 4

So the transgene was integrated into a certain area of the mouse genome and accidentally disrupted the gene located there. The aging-like phenotypes include short lifespan, the klotho mutant mice, the average lifespan of klotho mutant mice is about 2 months.

Slide 5

And they don’t grow very well, they don’t grow weight after weaning until they die.

Slide 6

And they are infertile because of hypogonadism and they have premature thymic involution, skin atrophy, muscle atrophy and vascular calcification.

Slide 7

So, this is the HE staining of the aorta and you can appreciate severe medial calcification and in the kidney small arteries showed a severe calcium calcification. So they also develop osteoporosis. This is a bone radiograph of the femur and tibia of wild type mice and klotho mutant mice. You can see a generalised decrease of bone radiodensity in klotho mutant mice. Micro CT analysis showed that the thickness of cortical bone is significantly reduced in klotho mutant mice.

Slide 8

They have pulmonary emphysema. So these are sections of lung taken at the same magnification of the wild type mice and the klotho mutant mice and you can see significant enlargement of air spaces in klotho mutant mice which is very similar to pulmonary emphysema in humans.

Slide 9

They have ectopic calcification that you can see very well here but this is a section of the stomach stained with von Kossa staining and this staining detects calcium deposits as a black staining like this so you can see a lot of ectopic calcification in klotho mutant mice. So they also have motor neuron degeneration similar to ARS patients in humans and they have some hearing disorders.

Slide 10

So all these phenotypes are here every single homozygous with 100% penetrance, so we concluded that the klotho mouse is the first documented mammalian model for human aging in general that manifests multiple aging-like phenotypes in a single individual.

Slide 11

So, the next question is what gene was disrupted by insertional mutation of the transgene? We analysed the transgene integration site or klotho locus and found that the transgene was integrated in tandem at a single locus generating a simple deletion of 8 kb. We determined a nuclear sequence of this region and found that the lesion occurred just upstream of the unknown gene at the time. Hydrophobicity analysis of the predicted aminoacid sequence of the protein of this gene detected 2 hydrophobic stretches at the amino and carboxy terminus indicating that this gene encodes a single pass transmembrane protein with a big extracellular domain and short intracellular domain. The extracellular domain --- to beta glucosidases of bacterial plants however we were unable to detect enzymatic activity in recombinant klotho protein and later on we and other groups found that the klotho protein has weak beta-glucoronidase activity or sialidase activity. This raises a possibility that the klotho protein may be able to modify sugar chains of glycoproteins and/or glycolipids. I’ll get back to this point later.

Slide 12

The klotho gene was expressed predominantly in the kidney. This is the in situ hybridization and the immunohistochemistry of klotho in the kidney and we can detect weak expression of the klotho gene in proximal tubules and much stronger expression in distal convoluted tubules. Northern blot analysis of klotho mutant mice fails to detect this transcript, so our klotho mutant mice turned out to be null mutant for this transcript.

Slide 13

The conclusion at this point is that a defect in klotho gene expression leads to a syndrome resembling aging. The klotho gene encodes a single-pass transmembrane protein and is expressed in limited tissues.

Slide 14

So disruption of the klotho gene looks like --- aging-like phenotypes so that the next question is whether klotho functions as an aging suppressor gene in other words klotho overexpression may extend lifespan in mice. We’ve got two independent transgenic mouse lines that overexpress klotho and there is a control of human elongation factor 1 alpha promoter which is ubiquitously active, so these transgenic mice as shown in the red and blue lines, they live longer than wild type controls. The average lifespan was 20-30% longer than the wild type mice indicating that the klotho gene is an aging suppressor gene that accelerates aging when disrupted and extends lifespan when overexpressed in mice.

Slide 15

So we didn’t know what the biological function of klotho protein was until quite recently but recently we realised that a klotho protein is involved in the regulation of fibroblast growth factor signalling pathway.

Slide 16

We came up with this idea when we found a report on FGF 23 knockout mice several years ago. Klotho deficient mice as I said, developed multiple aging-like phenotypes as listed here and quite surprisingly FGF23 knockout mice also developed identical abnormalities addressed here. So, these two mutant mice not only shared pathological changes, aging-like pathological changes but also very similar metabolic abnormalities represented by high blood phosphate levels, high blood vitamin D levels, high blood calcium levels and low blood glucose levels.

Slide 17

So, what is FGF23? So many of you in the audience know about FGF 23, just briefly FGF 23 is one of the newest members of the FGF ligand super family, identified as a bone derived phosphaturic hormone that suppresses renal phosphate resorption and induces phosphate excretion. The gain-of-function of FGF23 causes phosphate wasting phenotype. So actually this gene was characterised very well in patients with autosomal dominant hypophosphatemic rickets or ADHR. These patients carry a mutation in the FGF23 gene that confers resistance to the proteolytic degeneration of FGF23 protein and therefore, these patients have very high blood FGF23 levels and that causes, induces phosphate excretion from the kidney and defect in bone demineralisation.
In contrast, loss of function mutations like I said develops phosphate retention phenotype as observed in the FGF23 knockout mice. Very interestingly, klotho deficient mice have more than 1000 times higher serum blood FGF23 levels than wild type mice. So these observations have led us to the hypothesis that klotho may be involved in the FGF23 signalling pathway.

Slide 18

Because klotho protein encodes a single-pass membrane protein and therefore if klotho has something to do with FGF23 signalling pathway it may directly interact with FGF receptors on the same surface. So to test this idea we did a very simple co-immunoprecipitation experiment. We transfected 293 cells with klotho expression vector or FGF receptor expression vector and asked if these two proteins are co-immunoprecipitated. FGF receptors had multiple isoforms, so at least there are 4 distinct genes that encode FGF receptors, FGF receptor 1, 2, 3, 4 and each gene has multiple or alternative stress variants and therefore, the total number of FGF receptor isoforms would be I don’t know exactly but maybe more than 30 or 40. So I just picked up the representative receptor isoforms and transfected each FGF receptor isoform expression vector and asked which isoform associated the best with klotho protein. The answer was klotho protein was co-immunoprecipitated very well with FGF receptor 1c, 3c and 4c.

Slide 19

Next we asked whether the association of klotho protein to FGF receptors may change the affinity to FGF23. To address this question we generated –bees conjugated with FGF receptor alone or a klotho protein alone or a klotho/FGF receptor complex and applied FGF23 protein to these –bees and asked which bees can pull down FGF23 most efficiently in vitro. The answer was very clear.

Slide 20

As you can see here, only FGF receptor/ klotho complexes can successfully pull down FGF23. So that means that FGF23 can bind only to klotho/FGF receptor complex. So this is also consistent with its ability to support FGF signalling pathway. So 293 cells don’t express klotho endogenously but they express FGF receptors, so that while basic FGF can stimulate FGF signalling pathway in this cell line, so by the way we monitor FGF signalling activity by looking at phosphorylation of FGF receptor substrate or FRS2 protein and its downstream target ERK.
But when you transfect 293 cells with klotho expression vector and then stimulate them with FGF23, you’ll get a nice dose response of the FGF23. So our working hypothesis is klotho forms complexes with FGF receptors on the cell surface to provide a high affinity binding site to FGF23 and facilitates FGF23 signalling activation.

Slide 21

Conclusions at this point are klotho forms complexes with FGF receptors. FGF23 binds to klotho/FGF receptor complexes. So in conclusion FGF23 requires klotho to activate FGF signalling.

Slide 22

What happens after FGF23 binds to FGF receptor7klotho complex? The physiological output is negative phosphate balance. So the klotho/FGF system achieves this job through promoting renal phosphate excretion in proximal tubules and suppressing renal vitamin D production in proximal tubules.

Slide 23

I’ll talk about vitamin D metabolism first.

Slide 24

As I said FGF23 knockout mice and klotho knockout mice have significantly high blood active vitamin D blood levels. When you inject FGF23 into normal mice, you can observe a decrease of blood vitamin D levels in a time and dose dependent manner. Active form of vitamin D is 1, 25-dihydroxyvitamin D3 and this is synthesised from its precursor in the kidney by the enzyme 1 alpha hydroxylase.

Slide 25

Active form of vitamin D is inactivated by 24-hydroxylase which is expressed not only in the kidney but also in peripheral tissues. So actually FGF23 can affect expression levels of these activating and inactivating enzymes, so FGF23 injection into mice causes reduction of 1 alpha-hydroxylase expression and increase of 24-hydroxyalse expression and therefore, its net effect is a significant decrease of active form of vitamin D. So that FGF23 klotho system suppresses vitamin D.

Slide 26

Conversely vitamin D actually enhances the FGF23/ klotho system. We know that administration of vitamin D in normal mice or rats increases expression of FGF23 in the bone and klotho in the kidney. If you inject vitamin D, you can see a significant increase of serum FGF23 levels in a time and dose dependent manner.

Slide 27

So these observations strongly suggest that a novel bone-kidney endocrine axis exists mediated by FGF23, klotho and vitamins D. So vitamin D is synthesised in the kidney by 1-alpha hydroxylase and vitamin D reaches to the bone and binds to its nuclear receptor or vitamin D receptor and lying under bound vitamin D receptor forms a heterodimer with another nuclear receptor RXR and directly binds to the promoter region of the FGF23 gene and upregulates expression. FGF23 secreted from bone in turn acts on FGF receptor/klotho complex expressed on the kidney and which results in suppression of 1-alpha hydroxylase and upregulation of 24-hydroxylase and suppresses vitamin D levels. So this negative feedback group is physiologically very important because upregulation of any single component of this group disturbs and messes up the vitamin D metabolism.

Slide 28

Another important factor that regulates vitamin D levels is of course PTH. PTH acts on the kidney to upregulate 1-alpha hydroxylase and increases vitamin D levels. But quite interestingly, parathyroid is one of the few organs that expresses klotho endogenously.

Slide 29

So, this means that FGF23 can act on this FGF receptor/klotho complex expressed on parathyroid cells. Actually FGF23 can suppress PTH transcription and PTH secretion which results in the suppression of vitamin D.

Slide 30

Therefore we can conclude that the klotho/FGF23 system counteracts vitamin D through down regulating 1-alpha hydroxylase in the kidney, upregulating 24-hydroxylase in the kidney and downregulating PTH in parathyroid. So reduction of serum vitamin D levels is obtained by the activation of this system.

Slide 31

Ok so, so far so good and we can explain why klotho deficient and FGF23 deficient mice develop identical phenotypes because klotho FGF23 requires klotho for its activity but there are many other questions that remain to be addressed. So one question is the phosphate resorption and vitamin D biosynthesis take place primarily in the proximal tubules in the kidney whereas klotho is much more abundant in distal convoluted tubules, DCT than in PT. So the question is what is klotho doing in DCT?

Slide 32

Another question to be addressed is the fact that as I said, the klotho protein is a single pass transmembrane protein but we and other groups identify that the klotho can be shed and cleaved on the cell surface by a membrane associated protease ADAM10 and released in the extracellular space.
So what is the secreted form of klotho protein doing? This is another interesting question to be addressed. So to address this question we generated a soluble form of recombinant klotho protein and injected them into rats and monitored their phosphate metabolism. So when we injected recombinant klotho protein into normal rats, we observed a significant decrease of blood phosphate levels and concomitantly urinary phosphate excretion is increased which means that klotho induced phosphaturia. So the secreted form of klotho protein may function as a phosphaturic protein.

Slide 33

So then what is the mechanism? How does klotho do this job? So phosphate metabolism is, as I said, primarily resolved and regulated in the proximal tubules. 70-80% of phosphate is reabsorbed from the proximal tubule and reused. In the proximal tubule sodium phosphate co-transporter type 2a is expressed in the brush border membrane or apical membrane and primarily responsible for phosphate uptake from luminal fluid.

Slide 34

So to make a long story short, I’m just showing here the scheme of the potential mechanisms or our working hypothesis. So klotho is expressed primarily in the distal tubular cells. We know that klotho appears to be expressed in both the basal and apical side of the distal tubular cells. So NaPi -2a is expressed in the apical membrane or proximal tubular cells and absolutely involved in the phosphate reabsorption.

Slide 35

So when klotho is shed and released into the extracellular space, we still don’t know the mechanisms but it appears that the klotho protein is transported into the apical side of proximal tubular cells probably through transcytosis and apical klotho protein can directly interact with sodium phosphate co-transporter and through its glycosidase-like enzymatic activity klotho removes – sugar chains on the NaPi-2a protein which results in the reduction of the channel transport activity and suppresses phosphate absorption. So this is one piece of data which supports this idea. So we prepared brush border membrane vesicles, this is a cell from the kidney so this is a cell-free system consisting just of an apical membrane, brush border membrane and ion transporters embedded there. So we can measure phosphate uptake by adding radioactive phosphate in the medium and we can actually measure phosphate uptake in test tubes. We treated this brush border membrane vesicles with – or recombinant klotho protein and did western blot analysis using NaPi-2a antibody and found that klotho treatment significantly reduced the glycosylated form of NaPi-2a and concomitantly increased unglycosylated form of NaPi-2a. So this is associated with reduction of phosphate uptake activity of this unglycosylated NaPi-2a. The same effect was observed when we used bacterial beta glucoronidase purchased from Sigma and actually but we need to use 4000 times higher concentration of beta-glucoronidase than klotho, so probably klotho is much more specific to this reaction. Anyway the ability of klotho and beta-glucoronidase to suppress phosphate uptake was rescued by a chemical inhibitor for beta-glucoronidase DSAL, so probably this activity depends on the enzymatic activity as glycocytes.

Slide 36

So klotho protein can modify sugar chains of NaPi-2a which leads to inactivation of this transporter. Actually klotho can also act on sugar chains on TRPV5. This work was reported by Doctor Binder’s group a couple of years ago and they found that klotho modifies –chains from TRPV5 which results in activation of TRPV5 and TRPV5 is a major calcium entry gate in the distal convoluted tubules.

Slide 37

So activation of TRPV5 potentially results in increased renal calcium resorption. So there are many functions of secreted klotho protein. So one of the major functions is as I said here klotho protein can regulate ion channels or transport activity through modifying their sugar chains on the cell surface. The target of klotho action is NaPi-2a and TRPV5 and this list of ion channels is growing.

Slide 38

We previously reported that the secreted form of klotho protein can suppress insulin IgF signalling. We and others reported that klotho can reduce oxidative stress. Other groups reported that klotho can suppress --- signalling so there are so many functions reported recently.
We still don’t know what is the common mechanism behind these multiple functions and we still don’t know what is the physiological importance for these activities and we still don’t know how klotho shedding or how klotho secretion is regulated.
So there are still many open questions but one thing we know for sure is why klotho FGF23 deficiency looks like aging. We have some different answers for this. So probably aging-like phenotypes in klotho deficient mice and FGF23 deficient mice are caused by phosphate toxicity because restriction of dietary phosphate alleviates many aging-like phenotypes in these mutants.
In addition restriction of dietary vitamin D lowers blood phosphate levels and alleviates many if not all aging-like phenotypes in these mutant mice. So our working hypothesis is phosphate accelerates aging. So naturally if the opposite is true, phosphate restriction may slow down aging.

Slide 39

So actually this might be true, partly true in patients with chronic kidney disease. I’m not a nephrologist so I won’t go into this topic in great detail but it has been reported that klotho is decreased in the kidney of CKD patients. Mice lacking klotho and ESRD patients share several common pathologies including hyperphosphatemia, vascular calcification and high blood FGF23 levels.
So my very crude hypothesis is CKD maybe viewed as a state of klotho deficiency of course I don’t believe that klotho deficiency can explain everything about the CKD phenotype but klotho deficiency may contribute to the pathophysiology of CKD patients. So there are so many open questions and there still remains a lot to do so maybe in the next few years we’ll get more interesting findings hopefully.

Slide 40

So I’ll just spend the last 5 minutes or so expanding our findings on klotho and FGF23 to other endocrine systems.

Slide 41

I talked about FGF23, FGF23 actually belongs to a subfamily of FGF ligand super family called endocrine FGF. As this name suggests, these 3 FGFs, FGF19 and FGF21 and FGF23 are very different from the other classical FGF ligands because these factors can circulate in the blood and regulate metabolism in adulthood. On the other hand, classical FGFs usually function in an autocrine paracrine fashion and regulate development and organ genesis. So the endocrine FGFs are very different from the classical FGFs. FGF19 is secreted from intestine upon feeding and acts on liver to suppress bile acid synthesis.
FGF21 is secreted from liver upon fasting and acts on adipose tissues to increase lipolysis.
FGF23 as I said, is secreted from bone and acts on kidney to suppress phosphate resorption and this factor requires klotho.

Slide 42

There’s another protein that has similarities with klotho it’s named beta-klotho, so beta-klotho was identified based on cDNA sequence homology with klotho and also encodes a single pass transmembrane protein but its tissue distribution is quite different from that of klotho. Klotho was expressed in the kidney but beta-klotho is expressed in liver and white adipose tissue. Beta-klotho knockout mice were reported and they developed increased bile acid synthesis. Actually, it has been known that FGF receptor 4 knockout mice FGF15 this is a mouse autolog of human FGF19 knockout mice also exceeded increased bile acid synthesis.

Slide 43

So we tested this possibility by doing a very similar experiment to what we did previously just by replacing klotho expression vector with beta-klotho expression vector. Co-immunoprecipitation in this experiment detected the strong induction between beta-klotho and FGF receptor 1c and 4. Again we looked at the similarities. Again we used 293 cells, as I said 293 cells don’t express klotho and don’t express beta-klotho and therefore, the HEK293 cells don’t respond to any of these endocrine FGFs. If you transfect the klotho, then the 293 cells become responsive to FGF23.

Slide 44

When you transfect beta-klotho expression vector and then stimulate the cells with endocrine FGFs we get good response not only to FGF19 but also to FGF21. So from the genetic evidence we expected that beta-klotho may support FGF19 signalling but its ability to support FGF21 is actually quite unexpected. But anyway I believe that we can successfully identify multiple novel endocrine axis mediated by klotho gene family members and endocrine FGF in family members. So beta-klotho is expressed in the liver and white adipose tissue.

Slide 45

Klotho is expressed in the kidney and parathyroid gland. FGF19 is secreted from the intestine and acts on liver where beta-klotho is expressed to suppress bile acid synthesis and actually beta-klotho is expressed in the gallbladder and induces gallbladder filling forming a negative feedback loop for post mesangial bile acid release and synthesis

Slide 46

So the take home message here is FGF19, 21, 23 these endocrine FGFs circulate in the blood but FGF receptors are expressed virtually in every tissue. So tissue specific expression of klotho and beta-klotho determines where these endocrine FGFs should work. So probably the klotho gene family has evolved in the regulation of bioactivity, tissue specific bioactivity of the endocrine FGFs.

Slide 47

So one point I should mention here is that this endocrine systems are regulated in very similar signalling molecules or molecular designs. As I said vitamin D upregulates FGF23 levels transcription and FGF23 in turn suppresses enzymes involved in vitamin D synthesis. So in FGF 19 beta-klotho system bile acid actually binds through its nuclear receptor effect are indirectly transcriptionally activate FGF19 gene expression and FGF19 then binds to beta-klotho/FGF receptor for complex expressed in the liver which suppresses CYP7A1 gene. This encodes late limiting enzyme of bile acid synthesis from cholesterol. So this again forms a negative feedback loop and disruption of any single component of this group like FGF19 knockout or beta-klotho knockout or FGF receptor knockout results in the identical phenotype which is increased bile acid synthesis.
So we still don’t know very well about FGF21 feedback group but fatty acids can actually bind to PPAR alpha and we know that PPAR alpha is the most important transcriptional upregulator of FGF21. FGF21 secreted from liver acts on beta-klotho FGF receptor 1 complex on the white adipose tissue and probably some CYP enzymes might be involved in this negative feedback loop. We still don’t know this identity.

Slide 48

Ok this is the end of my talk and these works have been done by very talented post docs and graduate students in my lab and I would like to thank Doctor Orson Moe and Doctor Chou-long Huang who are my best collaborators.

Slide 49

And I won’t go into detail in this list but I would like to thank all my collaborators. Thank you very much.