Kidney Growth During Catabolic Illness: What It Does Not Destroy Makes It Grow Stronger

Proteolysis Regulating Growth 

It is easy to understand how synthesis of new proteins could produce proteins necessary for cell growth. The transcriptional and translational machinery of the cell allow exquisite control of synthesis of any required protein. It has been more difficult to understand how kidney cells could selectively reduce breakdown of only the proteins needed for growth. The classic pathway of lysosomal macroautophagy is not considered to be a highly selective system of intracellular proteolysis, because bulk, unselected cytoplasmic proteins are delivered to the harsh, acidic environment of the lysosome where proteases are activated by the low pH.8 During the past 15 years, selective pathways of protein degradation have been described. The ubiquitin-proteasome system, the most important system for destroying intracellular protein, is controlled by a large number, perhaps hundreds, of ubiquitin ligases that recognize specific proteins to be destroyed.9 By tagging targeted proteins with another small protein, ubiquitin, these ligases provide a signal allowing specific proteins to be destroyed by a second protein complex, the proteasome. In muscle, the ubiquitin proteasome system is specifically activated during many catabolic conditions.9

A specific lysosomal pathway, chaperone-mediated autophagy (CMA) (Fig. 1), is of particular interest because of its relationship to growth, nutrition, and kidney disease. In CMA, a protein, a heat shock cognate protein, hsc73, binds to a five amino acid sequence similar to Lys-Phe-Glu-Arg-Gln (KFERQ) on the target protein.10 Once this binding takes place, hsc73 binds to the lysosome-associated membrane protein 2a (LAMP2a) on the lysosomal membrane, which facilitates delivery of the protein into the lysosome.10 In response to stimuli that increase CMA in liver cells, LAMP2a abundance increases, especially in the lysosomal membrane.10 Overexpression of LAMP2a stimulates activity of CMA, suggesting that LAMP2a abundance is a rate-limiting step for the activity of the pathway.10

View Large Image Download to PowerPoint

Figure 1. Chaperone-mediated autophagy (CMA). In CMA, cytoplasmic proteins with the KFERQ target sequence bind hsc73 and are unfolded. The complex binds to LAMP2a, and then the protein is transported through the lysosomal membrane. KFERQ, Lys-Phe-Glu-Arg-Gln; HSC 73, heat shock cognate protein of 73 kilodaltons; LAMP2a, lysosome-associated membrane protein 2a. Modified from Franch HA: Pathways of proteolysis affecting renal cell growth. Curr Opin Nephrol Hypertens 11:445–450, 2002.

One clue that CMA may be regulated in renal hypertrophy comes from studies of metabolic acidosis.11, 12 During acidosis, the kidney increases ammonia production as a way of excreting acid. By itself, acidosis does not promote hypertrophy in cultured renal tubular cells, but ammonia does at the concentrations seen in the kidney during acidosis.11, 13 As ammonia enters lysosomes, it traps hydrogen ions, increases the pH in lysosomes, and decreases proteolytic activity of both unselective macroautophagy and selective CMA. Any inhibitor of lysosomal function, whether it works through alkalization or as a protease inhibitor, also leads to hypertrophy.11, 12 Because KFERQ-containing proteins (those destroyed by CMA) account for up to one third of total protein in kidney, suppressing lysosomal degradation would lead to the accumulation of this class of proteins.14 Many enzymes abundant in renal hypertrophy contain a KFERQ sequence. These include phosphofructokinase, glyceraldehyde 6 phosphate dehydrogenase (GAPDH), pyruvate kinase, choline kinase, and phosphatidylcholine transferase. These glycolytic and synthetic enzymes along with transcription factors with KFERQ motifs, including c-fos and the paired box transcription factor, pax2, are abundant during growth in the kidney.6, 10, 15

Furthermore, KFERQ proteins in the kidney are selectively degraded during calorie starvation.10 Because starvation increases the activity of CMA, the observation that calorie starvation prevents the renal hypertrophy that occurs after uninephrectomy suggests that these proteins are required for renal hypertrophy.4 Although there is no direct evidence that reducing CMA is enough to make kidneys grow, studies in the cathepsin A knockout mouse suggest that too much CMA activity will prevent renal growth.16, 17 Cathepsin A knockout mice have constitutively active CMA from decreased ability to downregulate lysosomal LAMP2a.17 These mice have the same phenotype as humans with cathepsin A deletions: death from renal failure at an early age.17, 18

What about diabetes and uremia? There is no increase in ammonia production in these conditions to downregulate CMA. It is known that growth factors, including epidermal growth factor (EGF), are increased in the kidney in uremia and diabetes and that growth factors can suppress protein breakdown.19, 20 With inhibitors of lysosomal function, we examined the pathway of proteolysis regulated by EGF in primary cultures of proximal tubule cells and found that a lysosomal system was being regulated.21 To investigate whether CMA is regulated in renal hypertrophy we used a cultured renal tubular cell model (NRK-52E cells). EGF, prolonged the half-life of GAPDH protein (a classic substrate for CMA22) as much as ammonia did. EGF caused a similar increase in the half-life of the pax2 transcription factor. This increase in half-lives was accompanied by an increased accumulation of proteins with a KFERQ motif, including GAPDH, the M2 isoform of pyruvate kinase and pax2. Notably, ammonia also increased the level of these proteins, including pax2. Proteins with no KFERQ sequence (e.g., hexokinase22) did not accumulate. Lysosomal import of KFERQ proteins depends on how much LAMP2a and the hsc73 are associated with the lysosome.10 We found that EGF decreased the amount of LAMP2a and hsc73 associated with lysosomes. We concluded that in cultured renal cells, EGF slows the breakdown of proteins targeted for destruction by CMA, just as ammonia does.

To show that diabetes also regulates CMA, we used a pair-fed rat model of acute diabetes to be sure that there was no influence of differences in caloric or protein intake that could affect CMA.6 Lysosomal protein breakdown in the renal cortex of the diabetic animals was lower. Just as we had seen with EGF in the cell culture model, we found increases in KFERQ containing proteins, especially GAPDH and pax2.6 Moreover, lysosomes isolated from the renal cortex of the diabetic animals had less LAMP2a and hsc73 associated with them, consistent with reduced CMA.6 We concluded that diabetes reduces CMA in the kidney.

Signal-Regulating Proteolysis 

Growth factors activate well-defined intracellular signaling. Pathways activated by insulin and insulin-like growth factor-1 (IGF-1) receptors have formed a paradigm for understanding regulation of cellular protein metabolism in skeletal muscle, especially the phosphatidyl-3 inositol kinase (PI 3 kinase) pathway (Fig. 2).23 In cultured hepatic carcinoma cells, addition of the phosphoinositide products of type 1 PI 3-kinase suppresses proteolysis,24 and inhibiting pentaerythritol tetranitrate, a phosphatase that inactivates type 1 PI-3 kinase, seems to specifically suppress macroautophagy in intestinal cancer cells.25 Similarly in muscle cells, we found that metabolic acidosis, a consequence of chronic kidney disease, inhibited insulin receptor substrate 1–associated PI 3-kinase and attenuated the suppression of protein degradation by insulin.26 PI 3-kinase is increased in the renal cortex during diabetic renal hypertrophy.27

View Large Image Download to PowerPoint

Figure 2. Phosphatidyl-3 inositol kinase (PI 3 kinase) signaling in kidney cells. This cartoon shows growth factors (e.g., epidermal growth factor [EGF]) activating the p85 regulatory and p110 effector subunits of phosphatidyl 3 kinase (PI 3 kinase). The PI 3 kinase pathway stimulates protein synthesis through phosphoinositol dependent kinase (PDK1) and the mammalian target of rapamycin (mTOR) activating p70 S6 kinase (p70S6K). Akt blocks protein degradation in kidney cells by preventing the forkhead proteins (FoxO) from activating CMA. Note that in muscle cells, FoxO activates the proteasome by this same pathway. The PI 3 kinase pathway also has a weak effect to reduce proteasomal proteolysis through mTOR. Nutrient intake supports activation of mTOR while inhibiting FoxO activity by SIRT1. CMA, chaperone-mediated autophagy; Sirt1, Sirtulin-1; p110, phospinositide 3 kinase p110; p85, p85 subunits.

We examined how EGF suppresses proteolysis in NRK-52E renal tubular cells. We first showed that growth factor tyrosine kinase receptors and the associated Ras signaling upstream of PI 3 kinase regulated protein breakdown.28 Ras activates the regulatory PI 3 kinase p85 subunit of which recruits and activates the PI 3 kinase p110 kinase subunit. We created separate adenoviruses expressing either a dominant negative p85 or constitutively active p110 subunits.28 Expressing dominant negative p85 accelerated the basal rate of proteolysis and completely eliminated the ability of EGF to suppress proteolysis. This shows that class 1 PI-3 kinase activity is required for EGF to regulate proteolysis. Transfection with this constitutively active p110 construct increased PI 3-kinase activity and Akt phosphorylation, and increased cell protein half-life by approximately 33%, suggesting that p110 activation is sufficient to regulate proteolysis. Expression of this dominant negative PI 3-kinase p85 regulatory subunit in L6 muscle cells also blunted the effect of insulin on proteolysis.26 PI 3 kinase regulated different proteolytic systems in kidney and muscle. In skeletal muscle, PI 3 kinase regulated proteasomal proteolysis, but in the kidney cells, it regulated a lysosomal system.26, 28, 29 Thus, PI 3 kinase regulated protein breakdown in both cell types but affected a different system of proteolysis in each.

What are the downstream effectors of type 1 PI 3-kinase on proteolysis? Akt and regulate the mammalian target of rapamycin (mTOR) are two major pathways downstream of PI 3 kinase. Akt phosphorylates and inactivates glycogen synthetase kinase and subclass O of the F Box (Forkhead) family of transcription factors, the FoxOs to regulate glucose metabolism.30 mTOR integrates nutritional signals with PI 3 kinase signaling to regulate protein synthesis by p70S6 kinase and similar pathways.31 We found that the mTOR inhibitor, rapamycin, seemed to partly reverse EGF-mediated suppression of phosphate dehydrogenase.29 However, neither proteasome inhibitor MG132 nor lactocystin altered the effect of EGF on proteolysis, but both prevented rapamycin from having any effect on proteolysis. This result suggests mTOR does not regulate CMA, but that mTOR mediates a separate weak effect of EGF to prevent activation of the proteasome.

Because mTOR did not appear to mediate the signal-regulating CMA, we examined whether Akt was required.29 We used a dominant negative Akt1 virus in the NRK-52E cells and blocked lysosomal proteolysis with methylamine. In the presence of the lysosome inhibitor, 10 mM methylamine, dominant negative Akt1 no longer stimulated proteolysis reflecting activity of the dominant negative Akt1 on lysosomal function. By using a constitutively active Akt1, we found a decrease in phosphate dehydrogenase that was two thirds of that induced by EGF. If PI 3 kinase signaling is important for inhibiting CMA, then inhibitors of the pathway should cause depletion of KFERQ proteins, such as pax2. Confirming this expectation, expression of a dominant negative Akt1 but not the mTOR inhibitor rapamycin decreases pax2 protein. Similarly, treatment with the constitutively active Akt1 increases pax2 expression. These results suggest direct regulation of CMA by Akt1.

Taken together, our results suggest that EGF regulates CMA in renal tubular cells by a pathway that includes Ras, type 1 PI 3-kinase, and Akt. What lies downstream of Akt? The forkhead family of transcription factors (the FoxOs) move to the nucleus to regulate the transcription of many metabolic genes important for surviving the unfed state.32 Akt phosphorylation of FoxOs leads to their inactivation by excluding them from the nucleus. Catabolic hormones (e.g., glucocorticoids) or conditions that interfere with IGF-1/insulin signaling in muscle decrease the phosphorylation of FoxO1 and FoxO3.33, 34 Decreased FoxO phosphorylation led to increased transcription of muscle-specific ubiquitin C3 ligases. Preliminary experiments with a constitutively active FoxO suggest a role for the FoxOs in regulation of CMA in kidney cells (W. Shen, MD, PhD, and H. Franch, MD, unpublished data, 2006). This is important because oxidative stress, which acts counter to Akt on FoxO activity and is increased in protein and calorie starvation, accelerates CMA.35 It increases FoxO activity by increasing nuclear import (by the JUN terminal kinase) and enhances FoxO transcriptional activity (by the SIRT1 protein deacetylase).36 The SIRT1 pathway is of particular interest because of its role in mediating nutritional signals. Further experiments will be needed to examine whether nutritional signals regulate CMA through FoxO signaling and whether it provides an integration point between dietary and growth factor effects on renal growth and the progression of renal disease.

Conclusion 

Kidneys grow and muscle atrophies in catabolic conditions such as diabetes, because of opposite activation of conserved intracellular signaling pathways. Insulin and IGF-1 resistance accelerates muscle protein breakdown through the proteasome, whereas activation of renal growth factors, such as EGF, allows accumulation of specific substrates needed for growth of the kidney. It is of interest that the PI 3 kinase signaling pathway regulates different proteolytic systems in different cell types. We hypothesize that there is a common growth factor-activated signaling pathway regulating proteolysis in epithelial and muscle cells, but divergence in downstream effectors. In evolution, such divergence occurred so that specialized cell-type specific systems of proteolysis are regulated by common signaling intermediates. Our hypothesis is that the PI 3 kinase/Akt pathway will interact with nutritional signaling pathways to regulate proteolytic systems in many clinically important conditions.