Dehydration Accelerates Cytogenesis and Cyst Growth in Pkd1−/− Mice by Regulating Macrophage M2 Polarization

Adult autosomal dominant polycystic kidney disease (ADPKD) has been shown to be related as a “third hit” to the occurrence of acute or chronic kidney injury. Here, we examined whether dehydration, as a common kidney risk factor, could cause cystogenesis in chronic-onset Pkd1−/− mice by regulating macrophage activation. First, we confirmed that dehydration accelerated cytogenesis in Pkd1−/− mice and that macrophages infiltrated the kidney tissues even earlier than macroscopic cyst formation. Then, microarray analysis suggested that glycolysis pathway may be involved in macrophage activation in Pkd1−/− kidneys under conditions of dehydration. Further, we confirmed glycolysis pathway was activated and lactic acid (L-LA) was overproduced in the Pkd1−/− kidney under conditions of dehydration. We have already proved that L-LA strongly stimulated M2 macrophage polarization and overproduction of polyamine in macrophage in vitro, and in the present study, we further discovered that M2 polarization-induced polyamine production shortened the primary cilia length by disrupting the PC1/PC2 complex. Finally, the activation of L-LA–arginase 1–polyamine pathway contributed to cystogenesis and progressive cyst growth in Pkd1−/− mice recurrently exposed to dehydration.


Autosomal dominant polycystic kidney disease (ADPKD)
is an inherited renal disease that is the fourth leading cause of renal replacement therapy (RRT) in the USA [1] and approximately 10% of end-stage renal disease (ESRD) in Europe [2]. Recently, there is only one Food and Drug Administration (FDA) approved treatment, tolvaptan, for ADPKD [3]. However, the long-term administration of the agent will cause some side effects, including thirst [4], polyuria [5], and liver injury [6]. Therefore, there is an urgent need to develop more effective and safe treatments with the better understanding of molecular mechanisms of ADPKD. The roles of inflammatory cells, especially macrophages in the pathogenesis of ADPKD, has become a hot focus in the past 10 years, [7][8][9] and it is much necessary to understand the roles and underlying mechanisms that drive macrophage plasticity in ADPKD, which may facilitate the development of novel therapeutic strategy for ADPKD treatment. Our previous study confirmed that M2-type macrophages promote cyst enlargement in the later stages of ADPKD in a rapid-onset polycystic kidney disease (PKD) mouse model, and showed that l-lactic acid (l-LA) acted as a key regulator influencing macrophage polarization in the polycystic kidney microenvironment [10]. Although we confirmed that macrophages play a vital role in the later phase of cyst enlargement (postnatal day [PD] [22][23][24][25][26][27][28][29][30][31][32], it is necessary to explore whether macrophages also affect initial cystogenesis (PD16- 21) in the PKD mouse model. As the lifespan of rapid-onset PKD mice can be as short as PD35 and there is an unavoidable overlap between kidney development phase (before PD21) and the cystogenesis phase (PD16-21), it is understandable that the trigger of cystogenesis in the rapid-onset mouse model was kidney development. The limitations of the model were problematic for further exploration of the definite mechanism by which macrophages caused initial cystogenesis. To resolve this issue, we inactivated the Pkd1 gene in mice at PD30 to generate a chronic-onset mouse model. However, isolated microcysts were not detected until PD180 in this model. It has been reported that ischemia-reperfusion injury (IRI) could notably accelerate the development of PKD in a chronic-onset mouse model [11] and renal macrophages undergo a classical transition from M1 to M2 type [12]. However, IRI is not common in adult ADPKD, suggesting that other factors are involved in cystogenesis in adult ADPKD. Dehydration is a common cause of heat stress nephropathy [13], and severe dehydration may lead to acute kidney injury (AKI) [14]. Non-severe or infrequent dehydration is often asymptomatic, but recurrent and severe dehydration may promote irreversible renal cell death, fibrosis, and dysfunction [15]. Therefore, dehydration-induced kidney injury (DKI) can contribute to the AKI-to-chronic kidney disease (CKD) transition. Roncal Jimenez et al. reported that DKI in wild-type mice resulted in the development of proximal tubular injury, macrophage infiltration, and kidney fibrosis, indicating that DKI is also likely to induce cystogenesis in mice with Pkd1 deficiency [16]. In the present study, we showed that DKI accelerated cystogenesis in a chronic-onset mouse model, thus providing an opportunity to comprehensively investigate the roles of macrophages in the development of ADPKD.

Establishment of Chronic-Onset PKD Mice
We crossed C57/BL6 Pkd1 cond/cond mice with C57/ BL6 tamoxifen-Cre (B6. Cg-Tg [Cre/Esr1]) 5Amc/J mice (stock 004,682; Jackson Laboratories) to produce the mice used in this study. To generate the chronic-onset PKD mouse model, we induced Cre recombinase activity in mice at PD30 by intraperitoneal injection with tamoxifen (total dose: 300 mg/kg and 100 mg/kg on three sequential days) in coin oil [10,17]. All experiments were performed using protocols approved by the Second Military Medical University Animal Care and Use Committee (SMMU-ACUC-20150430).

Animal Sample Collection
Mice were killed at the end of the dehydration period (that is, after the 7-h dehydration protocol at 8 weeks) by anesthesia and cardiac exsanguination, and kidneys were processed for DNA and RNA extraction and histological examination (fixed in 4% paraformaldehyde buffered solution, pH 7.3-7.4). A sample of the kidney specimens was flash frozen in liquid nitrogen and immediately stored at − 80 °C for further investigation.

Animal Experimental Protocol
The dehydration protocol consists of placing mice in a heated environment at 40 °C for 30 min each hour for a total of 7 h, 5 days a week, for a total duration of 8 weeks. During the dehydration treatment, the mice in one group received 25% glucose at a dose of 10 µL/g body weight daily via oral gavage using a plastic feeding tube (Instech Laboratories, Plymouth Meeting, PA, USA) for 8 weeks. We gavaged the mice in another group with (R)-GNE-140, which is an effective LDH inhibitor, at a dose of 5 mg/kg/day, for 8 weeks. Meanwhile, the mice in the control group received an equal volume of saline daily via oral gavage for 8 weeks.

Microarray
The Agilent Sure Print G3 Mouse GE Microarray (8*60 K, ID: 028005) was used. Total RNA was quantified with a Nano Drop ND-2000 (Thermo Scientific), and RNA integrity was assessed with an Agilent Bioanalyzer 2100 (Agilent Technologies). Sample labeling, microarray hybridization, and washing were performed according to the manufacturer's instructions. Briefly, total RNA was transcribed to doublestrand complementary DNA (cDNA), synthesized into cRNA, and labeled with cyanine-3-CTP; the labeled cRNA was hybridized onto the microarray. After three washes, the arrays were scanned with the Agilent Scanner G2505C. Feature Extraction (version 10.7.1.1, Agilent Technologies) was used to analyze array images to obtain raw data. Gene Spring (version 13.1, Agilent Technologies) was used to complete basic analyses of the raw data. The raw data were first normalized with the quantile algorithm. Probes with at least 100% of the values in any one of the conditions with flags in "detected" were chosen for further data analyses. Then, we identified DEGs through fold changes. The threshold set for up-and downregulated genes was a fold change ≥ 2.0. Afterward, gene ontology and KEGG analyses were used to determine the roles of the differentially expressed mRNA.

Renal Function Evaluation
Plasma levels of creatinine were measured using a mouse creatinine assay kit (#80350, Crystal Chem, IL, USA). Plasma levels of blood urea nitrogen (BUN) were measured using a BUN colorimetric detection kit (#K024-H5, Arbor Assays, Ann Arbor, MI, USA).

Measurements of Urinary Parameters
Urinary albumin value was measured by an immunoassay (Bayer, Elkhart, IN, USA). Urinary albumin-to-creatinine ratio (UACR) was calculated as urine albumin/urine creatinine (mg/g).

Renal Tissue Analysis
An aliquot of each individual kidney sample or collected cells was precisely weighed and transferred to a microfuge tube. After adding 1 mL extraction solution (acetonitrile:methanol:water = 2:2:1), the samples were vortexed for 30 s, homogenized at 45 Hz for 4 min, and sonicated for 5 min in an ice-water bath. The homogenization and sonication steps were repeated three times, followed by incubation at − 20 °C for 1 h and centrifugation at 12,000 rpm and 4 °C for 15 min. An 80 µL aliquot of the clear supernatant was transferred to an auto-sampler vial for liquid chromatography tandem mass spectrometry analyses. Stock solutions were prepared individually by dissolving or diluting each standard substance to a final concentration of 10 mmol/L. All standards were purchased from Sigma-Aldrich: pyruvic acid (107360), putrescine (51799), spermine (S4264), spermidine (S0266), mixed amino acid standards (A9906), and 4-aminobutyraldehyde (A44150). A series of calibration standard solutions were prepared by stepwise dilution of this mixed standard solution with extraction solution. UHPLC separation was performed using an Agilent 1290 Infinity I Series UHPLC system (Agilent Technologies) equipped with a Water Acquity UPLC BEH Amide column (2.1 × 100 mm, 1.7 μm). An Agilent 6460 triple quadrupole mass spectrometer equipped with an AJS-ESI interface was used for assay development. The MRM parameters were optimized for each of the targeted analytes using flow injection analyses by injecting the standard solutions of the individual analytes into the API source of the mass spectrometer. At least two MRM transitions (i.e., the Q1/Q3 pairs) per analyte were obtained, and the two most sensitive transitions were used in the MRM scan mode to optimize the collision energy for each Q1/Q3 pair. Between the two MRM transitions per analyte, the Q1/Q3 pairs that showed the highest sensitivity and selectivity were used as the MRM transitions for quantitative monitoring. The additional transitions acted as qualifiers for verifying the identity of the target analytes. Agilent Mass Hunter Work Station (B.08.00) was used for MRM data acquisition and processing. The working standard solution was diluted in a stepwise manner with the extraction solution, with double dilution factors. Standard solutions were subjected to UHPLC-MRM-MS analyses. Signal-to-noise ratios were used to determine LLODs and LLOQs. LLODs and LLOQs were defined as the analyte concentrations that led to peaks with signal-to-noise ratios of 5 and 20, respectively. Kidney tissue (50 mg) was homogenized in 1 mL ice-cold 10% trichloroacetic acid (TCA) on ice. The same kidney tissue: TCA ratio was used for all samples. Samples were centrifuged for 5 min at maximum speed. A 0.5 mL aliquot of the supernatant was transferred to a new microtube. Saturated ether (0.5 mL) was added to the supernatant, and the samples were vortexed for 20 s. The last two steps were repeated three times. Samples were left open in the fume hood for 30 min, and 50 µL each sample was mixed with 50 μL polyethylene glycol solution. The solution was mixed vigorously, incubated on ice for 30 min, and centrifuged at 13,000 rpm for 5 min at 4 °C. The supernatant was transferred to a fresh tube and diluted five-fold with ice-cold dH 2 O prior to the assay. Then, an L-lactate (L-LA) assay kit (Cat#: A-108S and A-108L, Department of Biochemistry, University at Buffalo, Buffalo, NY, USA) was used to measure the levels of lactic acid in the kidney samples. The activity of pyruvate kinase and lactate dehydrogenase (LDH) in kidney samples were measured with colorimetric pyruvate kinase assay kit (#Ab83432) and colorimetric lactate dehydrogenase assay kit (#Ab102526).

Immunohistochemical Staining for Macrophage
Macrophage infiltration was assessed by immunostaining with a polyclonal anti-mouse monocytemacrophage marker F4/80 diluted 1:50 (#ab111101). The number of positive cells for F4/80 was counted using an Image Pro Plus software (version 6.0). The software allows color recognition and positive cells were identified as % positive color saturation at 40 × magnification in a blinded manner using at least 20 fields for each biopsy section. Macrophages highly expressing arginase 1 (ARG-1) were stained with monoclonal antibody against mouse ARG-1 diluted 1:50 (ab124917) and digital images were quantified at 40 × magnification using Image Pro Plus software (version 6.0).

Cystic Index Calculation
Kidney Sects. (2 µm thick) were stained with hematoxylin and eosin (H&E) using routine procedures. Cyst formation was quantified from sagittal sections of whole kidneys. All specimens were scanned using Aperio XT, Leica Aperio (Leica Biosystems), and the raw data were read using Imagescope (version 12.3). The CI of whole kidney sections was calculated. Briefly, high-resolution whole kidney images were divided (along the longitudinal axis) into 20-30 small regions. Cystic areas in each split image and the whole kidney area were measured using Image Pro Plus (version 6.0) and the cystic areas were summed using Microsoft Excel 2010 as follows: CI of kidney specimen A = (Area of cysts in spit image 1 + 2 + 3 + 4… + n)/(Total area of kidney). Cystic indices (CIs) were calculated from 5 to 6 kidney sections at each group and the mean CI was determined.

Identification of Cell Proliferation in Kidney Sections
Ki-67 IHC staining was performed on paraffin Sects. (2 µm thick) of formalin-fixed tissue. All kidney sections were first incubated in 3% H 2 O 2 followed by 4% serum from the host animal in which the relevant secondary antibody was generated. Then, sections were incubated overnight at 4 °C with primary antibodies against Ki-67 (#sc7844, 1:100, Santa Cruz Biotechnology; #ab15580, 1:100, Abcam). Signals were detected using Dako EnVision Detection Systems Peroxidase/ DAB kit, Rabbit/Mouse (#K5007), and VECTASTAIN ® Elite ® ABC-HRP kits (peroxidase, goat IgG, #PK-6105, CA, USA). Micrographs were randomly generated with a light microscope at 40 × magnification. The numbers of positive and total nuclei were counted using Image Pro Plus (version 6.0). The ratio of Ki-67-positive nuclei to the total number of nuclei was calculated as a percentage using Microsoft Excel (version 2010). Images were captured from 5 to 6 kidney specimens at each postnatal age and 12-15 images were taken from each specimen (covering the cortex, C-M junction, and medulla).

Renal Macrophage Isolation and Sorting
Mouse kidneys were perfused through the left ventricle with ice-cold phosphate-buffered saline (PBS), excised, minced into small pieces (0.3-0.5 mm 3 ), and subjected to enzymatic digestion with a mixture of 500 U/mL collagenase I, 125 U/mL collagen XI, 60 U/mL DNase I, and 60 U/mL hyaluronidase for 1 h at 37 °C. The kidney tissues were triturated and filtered through a 40-μm nylon mesh. After erythrocyte lysis according to the manufacturer's instructions (

Real-Time PCR
Total RNA was extracted using an RNase minikit (Invitrogen, Carlsbad, CA, USA) and reversetranscribed. The primer sequences were detailed in the Supplementary materials. Real-time PCR was performed using SYBR Green PCR Master Mix (Toyobo, Osaka, Japan) and the Rotor-Gene 3000A real-time PCR system (Corbett, Sydney, Australia) according to the manufacturers' instructions. In brief, the PCR amplification reaction mixture (20 μL) contained 2 μL cDNA, 0.4 μL sense (F) primer, 0.4 μL anti-sense (R) primer, and 10 μL SYBR Green I. After initial denaturation at 95 °C for 1 min, the reaction was cycled 45 times. Each cycle consisted of denaturation at 95 °C for 15 s and primer annealing and extension at 60 °C for 31 s. Results are shown as the relative expression of the targeted genes normalized to the expression of Gapdh. Real-time PCR was performed in triplicate for each experiment, and the average values were measured. Each experiment was repeated three times. Using the gene specific efficiencies, mRNA relative expression folds were calculated as 2 −ΔΔ circle thresholdCT .

Statistical Analysis
All results are presented as mean ± SD. Variables were compared using Student's t-tests and Mann-Whitney U tests as appropriate. Data analyses were performed using SPSS (version 16.0), and P < 0.05 was considered to indicate significance.

Dehydration Promoted Cyst Growth in Mice with Pkd1 Gene Deficiency
In the present study, the Pkd1 gene was inactivated by tamoxifen at PD30 to generate a chronic-onset conditional knockout mouse model and without any intervention, microcysts could be detected in the cortex and outer medulla in Pkd1 −/− mice at PD180 [19]. The cystic phenotype was severe at PD300 (Fig. 1A). Pkd1 encoding protein, PC1, dominatingly expressed in RTECs (Fig. 1B), and the mice that present stable and heritable Pkd1 knockdown in RTECs (~ 80%, Fig. 1C and D). However, cyst growth was more accelerated and progressive under conditions of dehydration. The dehydration protocol, which is described in Fig. 1E, resulted in varying degrees of sweating and weight loss in Pkd1 −/− mice. Although the mice deprived of water during dehydration (water access at night [D-WAN], Fig. 1F) would drink more water at night (Fig. 1G), they still showed significantly more weight loss compared with their Pkd1 −/− littermates that had free access to water during dehydration (water at all times [D-WAT], Fig. 1H).
Interestingly, although D-WAN and D-WAT mice drank approximately the same amounts of water during 24 h (Fig. 1I), after 2 months of the intervention, renal dysfunction ( Fig. 1J and K), polycystic characteristics ( Fig. 1L-N), and urine protein excretion (Fig. 1O) showed much greater progression in the D-WAN Pkd1 −/− mice and their survival rates were significantly decreased compared with their D-WAT littermates (Fig. 1P). Renal cell proliferation was compared among different groups and the proliferative indices (PI%) were calculated: the D-WAN Pkd1 −/− mice > the D-WAT Pkd1 −/− and control mice, P all < 0.05 ( Fig. 1Q and R).   Fig. 2A). A total of 520 differentially expressed genes (DEGs) were common to pools 1-3 (Fig. 2B). Microarray data have been deposited in the Array Express database at EMBL-EBI (www. ebi. ac. uk/ array expre ss) under the accession number E-MTAB-10449. By comprehensive bioinformatics analysis based on the overlapping 520 DGEs, we designated glycolysis pathway as the primary target to elaborate how dehydration facilitated cyst formation ( Fig. 2C and D).

Specific Gene Alteration in Polycystic Kidney During Dehydration
Cell location based on the overlapping DEGs suggested that macrophages are key cells involved in the development of ADPKD in Pkd1 −/− mice (Fig. 3A). Immunohistochemical staining confirmed that there were clusters of macrophages infiltrating around the interstitial tissues even before microcyst formation in . Upregulated (red) and downregulated (green) differentially expressed genes (DEGs) in pools 1, 2, and 3. The microarray data was achieved from independent three mice in one group. Data are mean ± SD; one-way ANOVA. B In total, 520 DEGs were common to pools 1-3. C, D Significantly activated signaling pathways, including glycolysis, complement cascade, and regulation of platelet activation, in Pkd1 −/− kidneys under dehydration situation. C P value pattern. D Signaling pathway model. the D-WAN Pkd1 −/− mice ( Supplementary Fig. 1). Gene Ontology (GO) enrichment further indicated that macrophage plasticity, primary including polarization and activation, were both associated with cyst growth after dehydration in Pkd1 −/− mice (Fig. 3B). By KEGG pathway, we found that D-WAN Pkd1 −/− kidney tissues exhibited strong activation of glycolysis, pyruvate metabolism, HIF-1 signaling pathway, and arginine-polyamine metabolism, compared with D-WAT and control littermates (Fig. 3C). Kidney cyst growth in Pkd1 −/− mice is associated with regional hypoxia, which leads to activation of hypoxia-inducible transcription factors, HIF-1 and HIF-2 expressed in RTECs [20]. Supplementary Fig. 2 shows that HIF-1α not only express in RTECs but also express in macrophages in polycystic kidneys in D-WAN mice at PD180. Wiki pathway analysis indicated that the pathways representing glycolysis and macrophage activation were both significantly activated in D-WAN Pkd1 −/− mice (Fig. 3D). By protein-protein interaction network in 520 overlapping gene lists, we designated that almost all the genes (green arrow) that encoded enzymes regulating the glycolysis pathway, including Eno1, Pkm, Ldhal6b, Pklr, and Ldhb had presented significantly different expressions (Fig. 3E).

Glycolysis Was Activated in Pkd1 −/− Mice When Exposed to Dehydration
Pyruvic acid and l-lactic acid (l-LA) were considered to be key metabolites during glycolysis (Supplementary Fig. 3), and we showed that pyruvic acid and l-LA had already been overproduced in kidney tissues even before macroscopic cysts had formed in Pkd1 −/− mice ( Supplementary Fig. 4), confirming that glycolysis was activated preceding organic renal hypoxia, which was due to cyst compression and vascular bed destruction (the Warburg effect). To investigate whether increased glucose intake promoted cyst development when exposed to recurrent dehydration, we gavaged Pkd1 −/− mice with 25% glucose solution at 10 µL/g body weight (D-GLU group) via a plastic feeding tube (Instech Laboratories, Plymouth Meeting, PA, USA) under conditions of dehydration. Increased aerobic glycolysis is a prominent feature of the Pkd1 −/− kidney and among the enzymes involved in glycolysis, lactate dehydrogenase (LDH) is emerging as an attractive target for possible pharmacological approaches in PKD therapy. We gavaged D-WAN mice with (R)-GNE-140 (D-GNE group), which is an effective LDH inhibitor, at a dose of 5 mg/kg/day. All groups had free access to water during the night (Fig. 4A). Glucose intake significantly upregulated LDH protein expression ( Fig. 4B and C), enhanced the activity of LDH (Fig. 4D), and produced more l-LA in polycystic kidneys (Fig. 4E). As expected, a marked decrease in survival rate accompanied progressive cyst enlargement and poorer renal function in the D-GLU group (Fig. 4F and G), paralleling with marked promotion of cyst growth (Fig. 4H-J) and stimulation of CLEC proliferation in mice ( Fig. 4K and L). The LDH activity and production of l-LA in polycystic kidneys were significantly inhibited after (R)-GNE-140 treatment, and the polycystic characteristics, renal function, and survival rates were notably improved in the D-GNE group despite identical glucose load to the D-GLU group. These observations strongly suggested that glycolysis and its final product, l-LA, play a vital role in dehydration-induced cyst enlargement in Pkd1 −/− mice.

Glycolysis-Induced M2 Polarization Caused Cytogenesis in Pkd1 −/− Mice
We have reported that l-LA could enhance the activity of arginase-1 (ARG1) and stimulate polyamine production in macrophages in vitro and confirmed that treatment with N-hydroxy-nor-l-arginine (nor-NOHA) markedly inhibited M2-macrophage polarization and postponed cyst development in chronic-onset Pkd1 −/− mice [10]. In the present study, we showed that glucose intake under conditions of dehydration further stimulated the macrophage infiltrating polycystic kidneys (Fig. 5A). More than that, more macrophages were driven toward the M2 type ( Fig. 5B-E), and the expression of ARG1 were enhanced in the macrophages (Fig. 5F-I). The mRNA expressions of the genes encoding regulator enzymes of polyamine metabolism in renal macrophages were described as follows: for Arg1, D-GLU > D-WAN > D-GEN, P all < 0.05; for Odc1, D-GLU > D-WAN and D-GEN, P both < 0.05; for Smox, there were not significantly differences among three groups (Fig. 5J). As a result, greater amounts of intermediate and final products of polyamine metabolism, including L-arginine (Arg), ornithine (Orn), putrescine (Put), 4-aminobutyraldehyde (4-Amino), and spermidine (Sperm), were produced in the isolated macrophages in D-GLU PKD kidneys, compared with D-WAN counterparts (Fig. 5K). By contrast, the renal characteristics described above in the D-GLU group were significantly reversed in the D-GEN group.

Polyamine Shortened the Cilia Lengths on OX-161 Cells by Disrupting PC2-PC1 Combination
Our previous work demonstrated that putrescine could stimulate the proliferation of fibroblasts and upregulate the expression of TGF-β both in vivo and in vitro. [21] It would be helpful to determine the marked progression of renal fibrosis at the end stage of ADPKD. In the present study, we showed that putrescine significantly shortened the primary cilia length of OX-161 cells, a human cyst-lining epithelial cell line, according to the dosage-dependent manner. The effect of spermidine to the primary cilia length on OX-161 H Two kidney weights/body weight (2KW/BW) ratios among groups. I Gross appearance of polycystic kidneys were compared between D-GLU and D-GNE groups. Black arrows show cysts on the surface of kidneys. J Cystic characteristics were notably deteriorated in D-GLU groups but remarkably improved after D-GNE treatment. Scale bars: 1 mm. K, L Renal cell proliferation was evaluated using Ki-67 IHC staining. Arrowheads show the Ki-67 stained nuclei of proliferative cells. Scale bars: 50 μm. The average value of one group was achieved from six independent mice. Data are mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001, analyzed by one-way ANOVA. , and spermidine (Sperm) in renal macrophages were measured using the NMR-based metabonomic method. The average value of one group was achieved from six independent mice. Data are mean ± SD; *, P < 0.05; **, P < 0.01; ***, P < 0.001, analyzed by one-way ANOVA. cells was not observed (Fig. 6A and B). As shown in Fig. 6C, many proteins are associated with the expression of PC2 or PC1. In vivo, we compared the mRNA expression levels of these proteins between D-WAN and D-WAT polycystic kidney tissues (Supplementary Fig. 5) and found that CD2AP showed strongest upregulation in D-WAN Pkd1 −/− kidney tissues (black arrow, Fig. 6D). The expression of CD2AP was significantly upregulated in D-WAN Pkd1 −/− kidneys, compared with D-WAT counterparts (Fig. 6E). IHC staining shows that CD2AP primary expresses in the glomeruli rather than the RTEC in Pkd1 −/− kidneys of D-WAT mice; however, it both highly expressed in glomeruli and RTEC in Pkd1 −/− kidneys of D-WAN mice (Fig. 6F). Further, we found that the expressions of CD2AP in primary RTEC isolated from D-WAN Pkd1 −/− kidneys were significantly upregulated compared with those from D-WAT counterparts. Immunofluorescence analysis revealed that while endogenous PC2 localized to the plasma membrane ( Fig. 6G-I) and > 85% of the primary cilia in untreated OX-161 cells, those treated with putrescine at 1 mM showed striking loss of PC2 staining from these locations (~ 45%, Fig. 6J). Besides, the lengths of primary cilia on the PC2 − OX-161 cells were significantly shorter than those on the PC2 + OX-161 cells when they were both treated with 1 mM putrescine for 16 h (Fig. 6K). Putrescine at 1 mM significantly upregulated the expression of CD2AP and downregulated the expression of PC2 in OX-161 cells in vitro (Fig. 6L). For CD2AP dominantly expressed in the cytoplasm of OX-161 cells, the effect likely interrupted formation of the PC1/PC2 complex and disrupted the function of primary cilia (Fig. 7).

DISCUSSION
This study was performed to elaborate the mechanism by which the microenvironment influences macrophage M2 polarization in polycystic kidney tissue under conditions of recurrent dehydration. Based on the results of bioinformatics analysis and our experience, we focused on the glycolysis-arginine-polyamine pathway, which caused macrophages to manifest M2 phenotype.
Previous studies have demonstrated that increased surface area of macroscopic cysts in ADPKD attributed to the high proliferative rate of RTECs [22][23][24]. In rapidonset PKD model (Pkd1 inactivation before PD10), we discovered that the proliferation rate in PKD kidneys (~ 24%) was approximately three-fold higher than that in wildtype littermates (~ 8%) at PD18 ~ 20 [10]. The finding indicated that high proliferative rate in the rapid-onset PKD kidneys is probably the overlapping effects of the organ developmental status and the enhanced susceptibility to Pkd1 inactivation. Furtherly, we discovered that proliferation rates in specimens of chronic-onset PKD kidneys were not appreciably higher than in agematched control kidney samples at PD180. Alternatively, in chronic-onset ADPKD, the proliferation may act as an accelerant in the appropriate context [11]. For instance, in mice with Pkd1 deficiency after the development, during an episode of the subsequent tubule repairment after kidney injury, RTECs undergo several processes having some degree of similarity between processes that are at play during kidney development, including proliferation, dedifferentiation, and redifferentiation. Therefore, it is conceivable that an episode of renal injury/repair in an adult kidney harboring Pkd1-deficient RTECs will also accelerate PKD [25,26].
The tubular segment with Pkd1 deficiency is considered to be an important factor influencing the susceptibility of cyst formation. For instance, Piontek et al. ever reported that Pkd1 inactivation at PD2 would result in grossly enlarged, homogenously cystic kidney within 2 weeks, and cysts originated from all tubular segments [17]. Leonhard et al. proposed that the activation of Cre and Pkd1 gene inactivation by tamoxifen at PD10 would led to many rapidly progressing cysts in the outer medulla. They also raised that Pkd1 is inactivated around PD10, the outer medulla is more susceptible to cyst formation than other regions of the kidney, while the inner medulla seems highly resistant to cyst formation [25]. Piontek et al. ever intraperitoneally injected mice with 250 mg/kg tamoxifen at 6-week-old and generated chronic-onset PKD mice (> 50% Pkd1 inactivation); they observed that the grossly cystic kidneys derived from all nephron segments at 6-month-old [17]. Takakura et al. reported that ischemia reperfusion injury-induced cyst formation was extensively observed in both the inner cortex and the outer stripe of the outer medulla in the kidneys from chronic-onset PKD mouse model [11]. Legué and Liem also have observed that cysts initially generated in proximal tubules in the adult PKD mice with Tulp 3K407I mutants and they assessed that initial cystogenesis has been linked to accelerated cell proliferation in proximal tubules and collecting ducts in PKD kidneys [27]. The causes for the susceptibility of tubular segments to Pkd1 deficiency differences are not Dehydration Accelerates Cytogenesis and Cyst Growth... well known. Some researchers speculate that, especially during neonatal kidney development in the outer medulla, considerable tubule lengthening occurs, which requires the full potential of PKD protein function to coordinate correct directional cell division [25].
Recent studies have suggested that alternations in metabolic pathways may be critical in ADPKD [28] and that defective glucose metabolism may be a hallmark offering new insight for investigating the roles of metabolic alterations in cystogenesis [29,30]. Glucose is the main source of energy for the cell, and in the presence of oxygen, pyruvate is normally transported into mitochondria where it is converted into acetyl-CoA, which enters the tricarboxylic acid and is completely oxidized [31]. However, under hyperproliferative conditions, such as in early ADPKD, CLECs tend to show this inefficient process, even when oxygen is available. Rowe et al. suggested that cells lacking the Pkd1 gene tend to rely heavily on aerobic glycolysis as an energy source even when oxygen is available [29]. We also found that pyruvic acid and lactic acid were overproduced before massive cyst formation, confirming activation of glycolysis preceded hypoxia in Pkd1 −/− kidney tissues. Many animal and clinical studies have confirmed that interfering with glycolysis could have a beneficial effect on the progression of ADPKD. We confirmed that intake of fluid supplemented with glucose further enhanced the glycolysis burden in Pkd1 −/− mice to produce more l-LA in kidney tissues. These observations indicate that sugared beverages could not replace water to supplement body fluid, especially for individuals with Pkd1 deficiency.
Hopp et al. have identified several biologically relevant metabolic pathways that were altered early in chronic-onset ADPKD, the most highly represented being arginine biosynthesis and metabolism. They also found that levels of kynurenines and polyamines were also augmented in chronic-onset PKD kidneys compared with wild-type counterparts [32]. Trott et al. reported that arginine depletion would result in a dose-dependent compensatory increase in argininosuccinate synthase-1 levels as well as decreased cystogenesis in vitro and ex vivo. [33] In the present study, we have proved that heat-stress kidney injury induced by recurrent dehydration markedly accelerates the cyst enlargement via glycolysisactivated L-LA-ARG1-polyamine pathway in the chronic-onset PKD model. As a cytokine, ARG1 could directly induce CLEC proliferation by the ERK-MRK pathway in vitro [10]. As a key enzyme in regulation of arginine metabolism, activated ARG1 will overproduce greater polyamines, including putrescine, spermidine, and spermine. Spermidine and spermine are intracellular molecules, while putrescine can diffuse into the cell membrane, and therefore putrescine is more likely to participate in intercellular interactions [34]. Putrescine has been shown to significantly stimulate TGF-β expression in fibroblasts [21]. It is reasonable to speculate that it plays an important role in renal fibrosis in the end stage of ADPKD in chronic-onset Pkd1 −/− mice. With regard to the observation that putrescine shortened the lengths of primary cilia in CLECs, we focused more on the effects of polyamine on cilia function. In the adult kidney, primary cilia function as mechanosensors for luminal flow and play important roles in the control of cell proliferation [35]. There is accumulating evidence that primary cilia play key roles in normal physiological functions of RTECs, and that defects in ciliary function cause cytogenesis [36,37]. PC2 functions in a complex with PC1, and the present study included a number of experiments to examine the effects of polyamine on PC1-PC2 complex formation in renal epithelial cells. Among all proteins that may combine with PC1 or PC2, CD2AP showed the most significant upregulation in Pkd1 −/− mice under conditions of dehydration. We found that putrescine elevated the expression of CD2AP by approximately 1.5fold in CLECs, but not PC2, in vitro. CD2AP was named for its ability to bind CD2 and promote CD2 aggregation to stabilize the interaction between T cells and antigenpresenting cells [38]. It is ubiquitously expressed at higher levels in immune cells, epithelial cells, and neurons. Previous studies have shown that CD2AP is necessary for signaling at the slit diaphragm of the kidney [39]. It can combine with PC2 and disrupt PC1-PC2 combination to cause cystogenesis [40]. Data are mean ± SD; *, P < 0.05; ***, P < 0.001, analyzed by t-test.

◂
This study had four limitations. First, the regulation of macrophage polarization is not limited to l-LA-arginine-polyamine pathway. Second, the composition of macrophage phenotypes in polycystic kidneys was not homogeneous and would change at different stages of ADPKD. Third, the elements regulating macrophage M2 polarization were not isolated but were closely interconnected with one another. In addition, macrophage polarization was likely to change according to alterations in protein expression. Finally, alteration of the microenvironment did not unbalance macrophage polarization, but rather caused macrophages to achieve a new equilibrium to better adapt to the altered homeostasis. It may be necessary to maintain the stability of homeostasis, but at the whole-body level it likely further drove progression of the disease.

FUNDING
The present work was supported by the following grants: (1) The National Natural Science Foundation of China (81700579, Fig. 7 L-LA-putrescine-CD2AP axis promoted cytogenesis in Pkd1 −/− mice when they were exposed to recurrent dehydration. (1) RTECs lacking the Pkd1 gene tend to convert glucose into lactate and preferentially use anaerobic glycolysis. (2) Lactate was accumulated when exposed to recurrent dehydration. (3) Lactic acid stimulated the production of polyamine in macrophage. (4) Putrescine directly interacted with RTECs, upregulate CD2AP expression, promote its combination with PC2 (5), and disrupt the stability of the PC1-PC2 complex on primary cilia of RTECs. (6) The dysfunction of primary cilia would finally cause initial cystogenesis.