Follistatin in chondrocytes: the link between TRPV4 channelopathies and skeletal malformations

Humanized porcine chondrocyte model

Chondrocytes were isolated from the femoral condyles of female adult pigs as described previously (27). Briefly, cartilage was minced, removed from the bone, and sequentially digested in Pronase (Calbiochem, San Diego, CA, USA) and type 2 collagenase (Worthington, Lakewood, NJ, USA). Chondrocytes were maintained in feed medium [Dulbecco's modified Eagle's medium–high glucose (DMEM-HG), 15 mM HEPES, 0.1 mM nonessential amino acids, and 10% FBS; Life Technologies, Grand Island, NY, USA] at 380 mOsm (osmolarity was adjusted by adding sucrose) and pH 7.4.

Cells were plated at 1.9 × 10⁵ cells/cm² for 3 d and then transfected with huTRPV4_WT or mutant huTRPV4 using Amaxa's Nucleofector device (Lonza, Basel, Switzerland) and the primary human chondrocyte kit, following the manufacturer's protocol. Cells were transfected with TRPV4 mutations that cause skeletal dysplasias of varying severity: V620I = mild (11), A716S = moderate (10), T89I = severe (14), or an arthropathy-causing mutation (F273L; ref. 16). WT TRPV4 and all mutant constructs were C-terminally tagged with red fluorescent protein (RFP) in pcDNA3.1, and mutagenesis was performed using the Quick-Change site-directed mutagenesis kit (Stratagene, Santa Clara, CA, USA). All constructs were verified by sequencing. TRPV4-RFP expressing cells were selected 48 h after the transfection using FACS DiVa Cell Sorter (BD Biosciences, Franklin Lakes, NJ, USA), thus ensuring similar expression levels of all transgenes. Given the dominant gain-of-function nature of the mutations, the endogenous porcine TRPV4 was unaltered in these cells to reflect the presence of the healthy TRPV4 allele in human patients with a TRPV4 channelopathy.

Ca²⁺ imaging

Cells were loaded with 2 μM Fura-2-AM (Life Technologies), and imaged using a protocol for ratiometric Ca²⁺ imaging using 340/380-nm blue light for dual excitation, recording emissions with specific filter sets, and generating ratios. Basal Ca²⁺ levels (averaged over 10 images prior to stimulus) and peak Ca²⁺ levels (maximum measurement after stimulus) in response to −100 mOsm stimulation were determined.

Gene array

RNA was extracted from porcine chondrocytes that were transfected with huTRPV4_WT or huTRPV4_V620I and exposed to isoosmotic or hypoosmotic (−100 mOsm) medium for 15 min, using TRIzol reagent (Life Technologies) according to the manufacturer's protocol. This RNA was run on Agilent 4x44k porcine microarrays (representing all known protein-coding mRNAs from Sus scrofa). Microarrays were conducted in triplicate. Hybridization was background subtracted, normalized for a panel of housekeeping genes, and analyzed as described previously (28). Differences between huTRPV4_WT and huTRPV4_V620I were calculated for genes with significant up-regulation.

Gene expression

FST gene expression was measured in transfected chondrocytes using quantitative reverse transcription polymerase chain reaction (qRT-PCR). RNA was isolated from transfected chondrocytes using TRIzol reagent (Life Technologies) according to the manufacturer's protocol. After purification, 2 μg of total RNA was treated with RNase-free DNase (Qiagen, Valencia, CA, USA). RNA (0.5 μg) was then reverse transcribed using oligo dT primers and the SuperScript first strand synthesis kit (Life Technologies) according to the manufacturer's protocol. Gene expression was assessed by qRT-PCR, using the ABI 7900HT Fast Real Time PCR System and Power SYBR Green PCR Master Mix (Applied Biosystems, Grand Island, NY, USA). Specificity of primers was verified by dissociation/melting curve for the amplicons when using SYBR Green as a detector. mRNA abundance was quantified using the 2^−ΔΔCt method with all data normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and the huTRPV4_WT sample as the calibrator.

Western blotting

For the detection of FST and actin proteins, huTRPV4-RFP transfected and sorted chondrocytes were lysed in 100 μl of CHAPS buffer containing 30 mM Tris-Cl (pH 7.5), 150 mM NaCl, 1% CHAPS, 1 mM phenylmethylsulfonyl fluoride, 1× protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA), incubated on ice for 30 min and centrifuged at 10,000 rpm for 3 min. Proteins from the cleared supernatant were then quantified using Bradford reagent (Sigma-Aldrich, St. Louis, MO, USA). Equal amounts of total proteins were analyzed by SDS-polyacrylamide gel electrophoresis and transferred onto Immobilon-P membranes (Millipore, Billerica, MA, USA) using a semidry transfer apparatus. Membranes were blocked in phosphate-buffered saline (PBS), 0.1% Tween 20 containing 5% nonfat powdered milk. Proteins were detected by incubating with antibodies at 0.2 μg/ml in blocking solution. Anti-FST antibodies used in this study were from AbCam (Cambridge, MA, USA), and antiactin antibodies were from Santa Cruz Biotechnology (Dallas, TX, USA). Primary antibodies were detected with goat anti-rabbit or anti-mouse secondary antibody conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA, USA). Immunocomplexes were detected using ECL Plus Western blotting Detection Reagents (GE Healthcare Life Sciences, Pittsburgh, PA, USA).

Transfection of FST-promoter-driven luciferase reporter into ATDC5 cells

Transient transfections of ATDC5 cells were performed in 12-well plates at 70–85% confluence with Lipofectamine 2000 transfection reagent (Life Technologies) according to the manufacturer's instructions. A 2.4-kb fragment of the pig follistatin promoter was cloned into the pGL3 basic luciferase vector (Promega, Madison, WI, USA), as either a WT isoform or a cAMP-response-element (CRE) mutant that comprises the transcription start site (Supplemental Fig. S1). These two promoter constructs were cotransfected with either huTRPV4_WT or huTRPV4_V620I expression constructs (pcDNA3.1 backbone). At 6 h after transfection, cells were treated with insulin-transferrin-sodium selenite (ITS) liquid medium supplement (Sigma-Aldrich) for chondrogenic differentiation and harvested on d 3. Lysates were assayed for luciferase activity following manufacturer's instructions (Promega). Results were normalized to Renilla luciferase activity resulting from cotransfection of the pRL Renilla luciferase reporter vector (Promega).

Humanized TRPV4 mouse generation

All animal protocols were approved by the Duke University Institutional Animal Care and Use Committee, in full compliance with U.S. National Institutes of Health guidelines. Transgenic humanized TRPV4 mice were generated using a human bacterial artificial chromosome (BAC; RP11-126K19; Children's Hospital Oakland Research Unit, Oakland, CA, USA). The human BAC was modified by recombineering (29–31) so that the IRES-yellow fluorescent protein (YFP) Venus cassette was inserted immediately following the translational stop signal, into the 3′ untranslated region of the human gene (huTRPV4_WT); for the mutant version, a point mutation was introduced in exon 12, resulting in the V620I mutant protein (huTRPV4_V620I). All modifications were sequence verified. Modified BAC DNA was injected into pronuclei of B6SJLF1 mice following standard procedures (32). Several transgenic lines were obtained per construct, which transmitted the transgene to the next generation. We detected human TRPV4 cDNA, using specific primers, in similar abundance from the two lines (one each WT, V620I) that were ultimately selected for further experimentation. These lines were backcrossed to C57BL/6N over >5 generations. We also tested for expression of YFP-Venus (30, 33–36). Its expression in cartilage was confirmed by confocal microscopy (LSM 510; Carl Zeiss, Inc., Thornwood, NY, USA; Supplemental Fig. S2). Mouse femurs were dissected from 3 animals/group: huTRPV4_WT, huTRPV4_V620I, and nontransgenic control. Femurs were soaked in a 1.5 μM solution of Mitotracker Deep Red (Life Technologies) to stain all living cells. YFP expression and Mitotracker were imaged in intact femoral condyles on a confocal microscope. Cells were visible in all mice using the Mitotracker, but YFP was detectable only in chondrocytes from the transgenic mice, with expression found in virtually all articular chondrocytes, similar to native TRPV4.

Gene expression

Knee and hip cartilage from 2-wk-old huTRPV4_WT and huTRPV4_V620I mice (n=3 samples/genotype; each sample contained pooled tissue from 6 mice) were collected and stored in RNA later (Life Technologies) until RNA extraction. Cartilage was then transferred to Trizol (Life Technologies) and homogenized using a Mini-Beadbeater (Biospec, Bartlesville, OK, USA). The RNA was extracted according the Trizol protocol through the precipitation step and then was processed according to the manufacturer's protocol for the RNeasy Plus Kit (Qiagen). RNA (0.5 μg) was then reverse transcribed using Superscript VILO (Life Technologies), according to the manufacturer's protocol. Gene expression was assessed by qRT-PCR as described above. Using the 2^−ΔΔCt method, data were normalized for Gapdh, and a huTRPV4_WT sample was set as the calibrator. Fst mRNA abundance was normalized for TRPV4 mRNA.

Chick limb ossification

Fertilized Ross Hubert chick eggs (Gallus gallus domesticus, Pilgrim's Pride Hatchery, Siler City, NC, USA) were incubated for a total of 9 d at 37°C and 70% humidity. Embryos were staged according to Hamburger and Hamilton (HH; ref. 38). Eggs were windowed on d 2 (HH14), sealed with tape, and incubated until d 7. Affi-gel blue gel 150- to 300-μm-diameter beads (Bio-Rad, Hercules, CA, USA) were incubated overnight with either PBS (Mediatech, Manassas, VA, USA) or 100 μg/ml carrier free recombinant human FST 288 (R&D Systems, Minneapolis, MN, USA). On d 7 (HH30), a single PBS- or FST-soaked bead was placed in the right leg of each chick near the tibial growth plate. The window chambers were resealed and incubated for an additional 2 d. On d 9 (HH34), embryos were harvested (n=14 for PBS treated and n=18 for FST treated), both legs were removed and stained with alcian blue (Acros Organics, Morris Plains, NJ, USA) and alizarin red (Electron Microscopy Sciences, Hatfield, PA, USA) to identify cartilage and bone, respectively, using a previously published protocol (39). The stained legs were imaged, and the bone length and ossification regions were measured for the tibia and femur using the LSM Image 5 browser (Zeiss). The percentage ossification of the tibia and femur were calculated as follows: (length of ossification region/total bone length) × 100.

Human chondrocytes

Primary human chondrocytes were isolated from the femoral cartilage of control (n=7) or International Skeletal Dysplasia Registry case R94-316, who has the neonatally lethal TRPV4_T89I mutation (14). Human TRPV4, human FST, and human GAPDH gene expression in the chondrocytes was determined by isolating RNA from the cells using the Trizol (Life Technologies) protocol and reverse transcription using Superscript VILO (Life Technologies), followed by qRT-PCR as described above. Using the 2^−ΔΔCt method, data were normalized for GAPDH, and a control sample was set as the calibrator. FST mRNA abundance was normalized for TRPV4 mRNA. The levels of FST secreted into the culture medium were determined using a human FST ELISA kit (R&D Systems) according to the manufacturer's instructions. Total FST secreted into the medium was normalized for total micrograms of RNA in the cell lysate.

Statistical analyses

All data are presented as means ± se. Data were tested for normality with Kolmogorov-Smirnov test. Non-normal data were log transformed to achieve normality, and statistics were performed on log transformed data. Student's t test or ANOVA was used to compare means (α=0.05). The percentage of murine chondrocytes responding with a Ca²⁺ signal was compared between conditions and genotypes by χ² test of proportions. Paired 2-tailed t tests were used to determine significant differences between the right and left legs in the chick experiments.

Humanized TRPV4 porcine chondrocyte model: assessing huTRPV4_V620I mutant channels

ATDC5 chondrocyte cell line model: assessment of transcriptional regulation of FST by huTRPV4_V620I

To better understand how excess Ca²⁺ entering the cell through huTRPV4_V620I up-regulates FST transcription, we conducted reporter gene experiments in a chondrogenic cell line. We transfected ATDC5 cells with huTRPV4_V620I, huTRPV4_WT, and a FST promoter construct that drives luciferase expression. Two versions of the FST promoter were used, 2.4 kb of the WT porcine FST proximal promoter (FST_CRE-WT) and the same DNA sequence with an engineered mutation in the CRE that is part of the FST transcriptional start site (FST_CRE-mut; ref. 42 and Supplemental Fig. S1). Undifferentiated ATDC5 cells expressing huTRPV4_V620I did not exhibit altered luciferase activity. Notably, ATDC5 cells that were differentiated toward a chondrogenic phenotype showed a significant increase in FST reporter activity in the presence of the V620I mutation, which was abolished when the FST_CRE-mut promoter was present (Fig. 2C). These findings demonstrate that huTRPV4_V620I channels, compared with huTRPV4_WT, increase transcriptional drive on a 2.4-kb proximal promoter fragment of porcine FST, but only once cells are driven into the chondrogenic lineage. These data are likely relevant for all mammals because of the highly conserved nature of the CRE sequence and relative position in the FST promoter in different mammalian species.

huTRPV4_V620I transgenic mice: humanized mouse model of brachyolmia caused by TRPV4_V620I

With evidence for Ca²⁺ influx through huTRPV4_V620I activating FST transcription and leading to its overexpression in primary chondrocytes, we examined the in vivo function of huTRPV4_V620I in a mouse model. huTRPV4_V620I transgenic mice were generated to recapitulate the skeletal features of brachyolmia in these mice, and to probe Ca²⁺ signaling and Fst expression abnormalities in chondrocytes of these mice. For controls, huTRPV4_WT mice were generated.

Humanized TRPV4 porcine chondrocyte model: assessing a wider range of TRPV4 mutations

To address whether other skeletal dysplasia-inducing TRPV4 mutations would also up-regulate FST, we expanded our humanized porcine chondrocyte model to include examples of a clinical severity spectrum of skeletal dysplasia mutations [V620I=mild (11), A716S=moderate (10), T89I=severe (14)], and used an arthropathy-causing mutation, F273L (16), which is not associated with skeletal dysplasia to test the specificity of our proposed FST channelopathy mechanism.

FST decreases bone ossification in a chick embryo model system

To show that FST up-regulation could cause alterations in the developing skeleton, we measured the effects of excess FST on bone ossification in chick legs (Fig. 5A). The treatment of chick legs with PBS soaked beads did not alter the percent ossification of either the tibia or the femur (Fig. 5B, C), as compared to the contralateral untreated limb. However, FST bead implantation significantly decreased the percentage ossification of both the tibia (Fig. 5B) and the femur (Fig. 5C) when compared to the contralateral untreated leg. Thus, excess FST significantly decreases ossification of long bones in the developing skeleton of a vertebrate embryo.

Native TRPV4_T89I in human chondrocytes increases FST

In a human case study, we measured TRPV4 and FST mRNA and secreted FST protein levels in human chondrocytes from a neonatally deceased TRPV4_T89I individual, compared to a cohort of control human chondrocytes. Human TRPV4 was expressed by both control human chondrocytes (0.92±0.21; fold change) and chondrocytes from the TRPV4_T89I patient (0.5). The TRPV4 expression levels were 1.8-fold higher in the control chondrocytes, perhaps indicative of an adaptive response by the T89I chondrocytes to the cellular injury that a channel that gates excess Ca²⁺ can inflict. On normalization to TRPV4 expression levels, FST gene expression was elevated nearly 4-fold in T89I chondrocytes (Fig. 6A). Notably, FST protein secretion from TRPV4_T89I chondrocytes was 3.8-fold higher than from control chondrocytes (Fig. 6B). These findings indicate that FST is overexpressed in chondrocytes in a case of a rare but lethal human skeletal dysplasia caused by a TRPV4 mutation, TRPV4_T89I.