The meniscus is a fibrocartilagenous disc in the knee that protects the joint from damage. Meniscal injuries are common, however repair efforts are largely unsuccessful and are not able to prevent the degenerative changes that result in development of osteoarthritis. Tissue regeneration in adults often recapitulates events of embryonic development, suggesting the regulatory pathways controlling morphogenesis are candidate repair signals. Here we use laser capture microdissection to collect mouse embryonic day 16 (E16) meniscus, articular cartilage, and cruciate ligaments. RNA isolated from these tissues was then used to perform genome-wide microarray analysis. We found 38 genes were differentially expressed between E16 meniscus and articular cartilage and 43 genes were differentially expressed between E16 meniscus and cruciate ligaments. Included in our data set were extracellular matrix proteins, transcription factors, and growth factors, including TGF-β modulators (Lox, Dpt) and IGF-1 pathway members (Igf-1, Igfbp2, Igfbp3, Igfbp5). Ingenuity Pathway Analysis revealed that IGF-1 signaling was enriched in the meniscus compared to the other joint structures, while qPCR showed that Igf-1, Igfbp2, and Igfbp3 expression declined with age. We also found that several meniscus-enriched genes were expressed either in the inner or outer meniscus, establishing that regionalization of the meniscus occurs early in development. © 2013 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:46–53, 2014.
The meniscus has a pivotal role in protecting the articular cartilages (ACs) of the femur and tibia during routine knee movement. Meniscal injuries are common, accounting for more than half of the knee arthroscopies performed each year. Although treatment of acute meniscal injuries has evolved dramatically, based largely on the findings that removal of damaged meniscus resulted in greater knee destabilization and accelerated cartilage degeneration,[1, 2] current surgical repair procedures cannot reliably prevent the deleterious changes and clinical symptoms that presage the development of knee osteoarthritis.
Increasing evidence suggests that the regulatory factors directing formation of musculoskeletal tissues during embryogenesis also play an instructional role in adult tissue regeneration.[4, 5] Despite the importance of the meniscus in knee joint homeostasis, relatively little is known about the cellular and molecular pathways that regulate meniscus development, allowing it to change from the uniform connective tissue structure seen within the joint interzone during embryogenesis to the highly complex and distinctly regionalized structure present in the adult knee.[5-8] Based on the contention that a better understanding of meniscal morphogenesis would provide a more informed basis for investigating and improving meniscal repair, we used laser capture microdissection (LCM) to isolate tissue from the meniscus, AC, and cruciate ligaments (CL) of embryonic day 16 (E16) mouse knees. RNA from these tissues was then subjected to microarray analyses and gene expression profiling. By comparing gene expression in meniscus, CL, and AC, we demonstrate that the developing meniscus expresses a unique gene signature that distinguishes it from other intra-articular connective tissues. Localization studies confirmed these results and revealed that by E16, the meniscus is already a heterogeneous tissue with regionalized patterns of gene expression. In addition, our data confirm a role for TGF-β signaling in knee joint morphogenesis and suggest that IGF-1 signaling may function in the specification and maintenance of meniscal progenitors. The individual genes and specific signaling pathways uncovered by our study can now be examined when developing new strategies aimed at enhancing meniscal regeneration and preventing osteoarthritis.
MATERIALS AND METHODS
Timed pregnant CD-1 and C57BL/6 mice were obtained from Charles River labs. Mice were maintained in a virus and parasite-free barrier facility and exposed to a 12 h light/dark cycle. All studies were approved by the Harvard Medical School Institutional Animal Care and Use Committee.
Laser Capture Microdissection, RNA Extraction, and Amplification
Freshly dissected E16 CD-1 mouse knee joints were embedded in optimal cutting temperature (O.C.T.) and snap frozen using liquid nitrogen. Cryosections were cut at a thickness of 14 µm at −20°C, placed onto polyethylene naphthalate (PEN) membrane glass slides (MDS Analytical Technologies, Sunnyvale, CA), stored briefly on dry ice, and then used for LCM. Slides were processed using the HistoGene LCM frozen section staining kit (Arcturus, Foster City, CA). Laser capture was performed using a Leica laser capture microdissection LMD 6000 microscope. Tissue was collected in extraction buffer supplied in the Picopure RNA isolation kit (Arcturus). Approximately 30 sections were collected per RNA sample. Each RNA sample represents knee joint tissue from three to four embryos. RNA was isolated using the Picopure RNA isolation kit (Arcturus) and RNA quality assessed by Agilent Bioanalyzer analysis. RNA integrity number (RIN) values were >5 which is within the accepted range for LCM expression analysis (Partners HealthCare Center for Personalized Genetic Medicine). RNA amplification was performed using 29 ng of total RNA per sample as an input for the Ovation Pico WTA System V2 (Nugen, San Carlos, CA).
Gene Expression Microarrays and Analysis
Microarray analysis was performed at the Partners HealthCare Center for Personalized Genetic Medicine (Cambridge, MA). Illumina Mouse WG-6 microarrays were hybridized with 1.5 μg of labeled cDNAs. Three independent biological replicates were run to account for variability. Microarray data were subjected to quantile normalization and a t-test was performed on the replicates followed by multiple hypothesis correction using the Benjami Hochberg test. Only fold changes with a p-value of less than 0.05 and a false discovery rate of less than 0.05 are reported. Analysis of differentially regulated genes was performed using Ingenuity Pathway Analysis (analysis.ingenuity.com). Microarray data was deposited in gene expression omnibus (GEO accession number: GSE47799 pending).
Histology, In Situ Hybridization, and Immunofluorescence
For histological and localization studies, E16 mouse knees were cut in the sagittal plane and three separate hindlimbs were analyzed. For in situ hybridization, in vitro transcription to generate riboprobes was performed using standard protocols and reagents (Promega, Madison, WI). Probes for Col1a1, Col2a1, Lox, and Igfbp3 were generated as previously described.[9-11] Probe for Acan was generated using the following primers 5′-AGCTTCAGTCCCAGAAGCCAGC-3′ and 5′-AGCTGGTAATTGCAGGGGACGT-3′; for Igfbp2 5′-ATGTTGGGAGGTGGTAGCAG-3′ and 5′-GAGGTTCAGCTTAACAGCCG-3′. Plasmid used to generate Igf1 probe was purchase from a commercial source (Open Biosystems, Huntsville, AL; clone 4194295). Histology and in situ hybridization were carried out as previously described.[12, 13]
For IGF1Rβ immunofluorescence, hindlimbs from E16 mice were fixed overnight in 4% paraformaldehyde and embedded in paraffin. Sections were rehydrated in ethanol and antigen detection was performed using a rabbit anti-human IGF1Rβ antibody (1:50; Santa Cruz Biotechnology, Dallas, TX, c-20). Subsequently, sections were incubated with AF555 tagged, goat anti-rabbit IgG antibody (1:400; Jackson ImmunoResearch, West Grove, PA), and mounted in Vectashield with DAPI (Vector Labs, Burlingame, CA) prior to viewing.
Medial and lateral menisci isolated from 8-week and 9-month old C57/Bl6 mice were homogenized and RNA isolated using Trizol reagent according to the manufacturer's instructions (Invitrogen, Grass Island, NY). cDNA was synthesized using the Transcriptor First Strand cDNA Synthesis Kit reagent per the manufacturer's instructions (Roche Applied Science, Indianapolis, IN). Quantitative RT-PCR (qPCR) was performed using the Universal Probe Library System (Roche Applied Science) and the primers listed in Table 1. PCR reactions were performed using the Roche Lightcycler 480 Real-Time PCR System. Data are based on triplicate reactions of at least two biological samples. Values were normalized to β-actin using the 2-ΔΔCt method.
|Gene||Primer Pair (5′ → 3′)|
The E16 mouse knee was chosen as the starting material for these studies because it is the earliest time point when all of the structures present in the adult knee joint are clearly distinguishable by histology (Fig. 1). At E16, the meniscus contains proliferating cells that will be responsible for the rapid growth in size that occurs before birth, as well as a population of cells that have stopped proliferating and are now actively producing extracellular matrix proteins (ECM). As seen in Figure 1A–C, the embryonic meniscus can be easily distinguished within the mouse knee as a triangular-shaped structure located between the articular surfaces of the femur and tibia. Along with collecting meniscus-specific tissue, we also sampled the AC and CL, two tissues that have a shared developmental origin with the meniscus.[7, 15, 16] These structures are also easily identified based on their location and histological staining within the forming joint (see Fig. 1).
By performing comparative microarray analyses on the LCM tissues, we were able to create tissue-specific profiles for the meniscus (M), AC, and CL. For our studies, genes with a fold change less than 2 were considered common between the three intra-articular connective tissues. These included Gdf5, Wnt9a, Erg, Noggin, and Chordin, all previously validated as markers of synovial joint development. Altogether, 38 genes were differentially expressed between E16 meniscus and E16 AC, while 43 genes were differentially expressed between E16 meniscus and E16 CL. The top 20 differentially expressed genes from each comparison (Meniscus vs. AC, Meniscus vs. CL) are shown in Tables 2 and 3. Included in the lists are transcription factors, growth factors, and extracellular matrix molecules, consistent with the developmental events occurring in the E16 knee. Quantitative reverse transcriptase-polymerase chain reaction (qPCR) analysis confirmed the differential expression of the subset of genes we refer to as meniscus signature genes (Supplemental Fig. S1).
|Gene Symbol||Gene Name||Fold Change||p-Value|
|Sepp1||Selenoprotein P, plasma, 1||3.0||0.04|
|Adamts4||A disintegrin-like and metallopeptidase (reprolysin type) with thrombospondin type 1 motif, 4||2.6||0.02|
|Col4a1||Collagen, type IV, alpha 1||2.4||0.0008|
|Growth factors and regulators|
|Rspo2||R-spondin 2 homolog||3.8||0.05|
|Igf1||Insulin-like growth factor 1||3.5||0.002|
|Igfbp3||Insulin-like growth factor binding protein 3||3.1||0.001|
|Rgs5||Regulator of G-protein signaling 5||2.9||0.007|
|Socs2||Suppressor of cytokine signaling 2||2.8||0.03|
|Cell membrane protein|
|Laptm5||Lysosomal protein transmembrane 5||3.3||0.1|
|P2ry6||Pyrimidinergic receptor P2Y, G-protein coupled, 6||3.0||0.03|
|Mrc1||Mannose receptor, C type 1||3.0||0.03|
|Arap3||ArfGAP with RhoGAP domain, ankyrin repeat, and PH domain 3||2.9||0.01|
|Tyrobp||TYRO protein tyrosine kinase binding protein||2.4||0.02|
|Vat1l||Vesicle amine transport protein 1 homolog-like||2.5||0.0004|
|Ralyl||RALY RNA binding protein-like||2.3||0.004|
|Gene Symbol||Gene Name||Fold Change||p-Value|
|Egr1||Early growth response 1||3.4||0.002|
|Snai3||Snail homolog 3||2.8||0.0004|
|Col8a1||Collagen, type VIII, alpha 1||3.3||0.03|
|1500015O10Rik/Ecrg4||Riken cDNA 1500015O10||3.2||0.002|
|Col15a1||Collagen, type XV, alpha 1||3.0||0.005|
|Growth factors and regulators|
|Hbegf||Heparin-binding EGF-like growth factor||3.1||0.003|
|Igfbp2||Insulin-like growth factor binding protein 2||2.9||0.001|
|Igfbp5||Insulin-like growth factor binding protein 5||2.5||0.001|
|Csrp2||Cysteine and glycine-rich protein 2||3.0||0.006|
|Prkd2||Protein kinase D2||2.5||0.009|
|Errfi1||ERBB receptor feedback inhibitor 1||2.5||0.02|
|Magee1||Melanoma antigen, family E, 1||2.5||0.002|
|Cell membrane protein|
|Slc9a3r1||Solute carrier family 9 (sodium/hydrogen exchanger), member 3 regulator 1||2.8||0.00006|
|Ralyl||RALY RNA binding protein-like||2.6||0.004|
|Ckb||Creatine kinase, brain||2.4||0.004|
|Idi1||Isopentenyl-diphosphate delta isomerase||2.3||0.001|
|N6amt2||N-6 adenine-specific DNA methyltransferase 2||2.3||0.02|
|Higd1a||HIG1 domain family, member 1A||2.3||0.001|
When we compared gene expression patterns between E16 meniscus and AC, a number of interesting trends emerged. We found that lysyl oxidase (Lox), a gene that functions to covalently crosslink collagen and elastin, is highly enriched in the developing mouse meniscus, consistent with previous reports of its expression in adult human meniscus. Dermatopontin (Dpt), a protein involved in collagen fibrillogenesis, is also highly expressed in the E16 meniscus. As Lox and Dpt are modulators of TGF-β activity,[19, 20] their abundance in the meniscus is in keeping with the recently established requirement for TGF-β signaling in knee joint morphogenesis.[13, 21] R-Spondin 2 (Rspo2), an activator of Wnt signaling is also enriched in E16 meniscus, consistent with its reported localization to the inner mouse meniscus. Intriguingly, Igf-1 and the Igf-1 binding protein 3 (Igfbp3) are more than threefold higher in meniscus tissue versus AC, suggesting that the meniscus may be a source of IGF-1 during joint development.
A comparison between meniscus and CL also highlighted the differences between the ECM of these two connective tissues. Two members of the non-fibrillar collagen gene family that are components of basement membrane, collagen, type VIII, alpha 1 (Col8a1) and collagen, type XV, alpha 1 (Col15a1) were more abundant in meniscus than CL.[22, 23] The TGF-β and IGF signaling pathways were also prominent; early growth response 1 (Egr1), a mediator of TGF-β induced fibrosis, showed the greatest differential expression between meniscus and CL, while Igfbp2 and Igfbp5, proteins that sequester IGF-1 in the ECM thereby regulating its bioavailability, were both highly enriched in meniscus. Esophageal cancer-related gene 4 (Ecrg4), previously reported as expressed in human intra-articular tissues but decreased in osteoarthritic cartilage is another gene highly enriched in the E16 mouse meniscus.
We also employed Ingenuity Pathway Analysis on our data sets to determine which canonical signaling pathways are active in the E16 knee. The most striking discovery was that IGF-1 signaling pathway molecules were significantly enriched in both comparisons (meniscus vs. AC p < 0.0020; meniscus vs. CL p < 0.026). This result was not surprising as the data in Tables 2 and 3 show that Igf-1, Igfbp2, Igfbp3, and Igfbp5, were all highly expressed in the meniscus compared to the other joint tissues. While IGF-1 signaling has been previously shown to play a role in the formation of the temporomandibular joint, this is the first time to our knowledge that the IGF system has been implicated in knee joint development and meniscal morphogenesis.
To verify the data obtained from our expression analysis, we next examined the spatial localization patterns of a select group of meniscus signature genes. For comparison sake, we also determined the expression pattern of ECM genes normally found in the adult meniscus. Although these ECM molecules are not part of the meniscus signature, including their expression allows us to create a developmental map of knee at E16. We found that collagen, type I, alpha 1 (Col1a1), the primary ECM component of the adult meniscus, was expressed in a broad domain of the E16 meniscus and was also found in the patellar tendon (Fig. 2B). Aggrecan (Acan), the major large proteoglycan in the adult meniscus, displayed a similar expression pattern in the meniscus to Col1a1, but was also strongly detected in the cartilage of the developing femur and tibia (Fig. 2C). Collagen, type II, alpha 1 (Col2a1), a major collagen component of the adult inner meniscus, was highly expressed in the cartilage similar to Acan, but was not detected in the meniscus at this stage (Fig. 2D). This result is consistent with previous reports showing that expression of Col2a1 in the mouse meniscus is not detected until 1 week of age. Overall our data provide evidence that many of the critical ECM components found in the adult meniscus are first expressed during meniscus development.
Having established the spatial expression pattern of meniscal ECM components, we next focused on localizing the meniscus-enriched genes identified in our array. We found that Lox was specifically expressed in the inner meniscus region of the E16 mouse knee (Fig. 2E) while Dpt localized to the outer meniscus as well as the developing gastrocnemius tendon (Fig. 2F). Igf-1 was detected throughout the meniscus as well as the cartilage of the femur and tibia (Fig. 2G). In contrast, Igfbp2 appeared to be expressed at higher levels in the inner region of the meniscus, while Igfbp3 localized to the outer meniscus (Fig. 2H,I). When we examined the expression of IGF-1R, the major receptor for IGF-1, we found that it was present in both AC and the developing cartilage of the femur and tibia but was absent from the E16 meniscus (Fig. 2J), suggesting that the knee cartilages may be the targets of meniscal produced IGF activity. To determine if genes present in the developing meniscus have a role in the adult knee, we isolated RNA from menisci of 8-week and 9-month-old mice and analyzed gene expression using qPCR (Fig. 3). We found that the levels of Col2a1, Col1a1, and Acan present in meniscus increased with age (Fig. 3A). In contrast, expression of Lox and Dpt remained relatively low and constant (Fig. 3B), while IGF-1 signaling pathway members, Igf-1, Igfbp2, and Igfbp3 decreased with age (Fig. 3C). Most surprisingly, expression of the Igf-1r could be measured in adult meniscus (Fig. 3C) suggesting a change in tissue responsiveness to IGF-1 in the adult meniscus compared to the developing meniscus.
Although the developmental anatomy of the meniscus has been well described, little information exists on the molecular mechanisms that direct meniscal morphogenesis. Using LCM combined with genome-wide microarray analysis to compare the gene expression profiles of E16 mouse meniscus, AC and CL, we uncovered a gene expression pattern that distinguishes the meniscus from the other intraarticular tissues present in the embryonic knee. We found 38 genes differentially expressed between the E16 meniscus and AC, and 43 genes differentially expressed between the E16 meniscus and CL. In addition to these quantitative differences, visualization of gene expression on E16 mouse knees revealed that several of the genes displayed distinct patterns, localizing to either the inner or outer area of the meniscus. As the mature meniscus is known to be a highly specialized structure with a unique shape and anatomy, our results suggest that even at early stages of joint formation, the meniscus has begun the critical regionalization necessary for its proper function within the adult knee.
The comparative microarray analysis we performed revealed that several modulators of the TGF-β family, including Lox and Dpt, were highly upregulated in the E16 meniscus. The TGF-β pathway is known to be critical for joint morphogenesis[29, 30] and recent evidence has shown that TGF-β signaling is necessary for knee joint development as hindlimbs from Tgfbr2Prx1cre knockout mice lack proper formation of meniscus, tendons, and ligaments.[13, 21] Our data provide additional evidence for the fundamental role of TGF-β signaling in meniscus formation. As enhancing TGF-β activity in the meniscus has shown success in animal models of meniscal injury, and incorporation of TGF-β into biomaterials used as meniscal substitutes has also shown efficacy,[31, 32] we believe that our developmental biology based approach will identify other signaling pathways that can be manipulated to aid in meniscal regeneration. To this end, bioinformatics analysis of our microarray data revealed a significant enrichment of the IGF-1 signaling axis in the E16 meniscus. This is a novel finding as IGF-1 is known to play critical roles in bone and cartilage development and homeostasis, but has not been previously implicated in the formation of the meniscus. Data from our study identifies the E16 meniscus as a potential source of IGF signals in the developing knee joint and show that prior to birth, meniscus cells are not likely targets for these signals because they have few IGF receptors. Levels of IGF1R mRNA increase in the meniscus after birth, suggesting that IGF signaling may have a regulatory role in the mature meniscus. IGF-1 increases the proliferation and migration of adult meniscal cells but has had limited success in defect models of meniscal repair.[33, 34] Since our data show that developing meniscus is enriched in IGF-1 pathway members, it may be necessary to treat with agents that turn on endogenous IGF-1 production rather than providing exogenous growth factor in order to stimulate healing. This will require a better understanding of the genes and signaling pathways that regulate IGF-1 expression during knee joint development.
It is likely that successful therapies for meniscus repair will require a more complete understanding of the regulatory molecules that control meniscal morphogenesis. At present few, if any, genes are known to be directly involved in meniscus formation. Our study provides a gene signature for the embryonic meniscus that can be used to gain new insights into how the meniscus forms and obtains its unique properties. As tissue regeneration often recapitulates development, the genes identified in our study can be used as markers to help define a reparative cell population resident in adult meniscus as well as to track the development process of engineered meniscal constructs to ensure the proper formation of this type of fibrocartilage. As strides continue to be made in generating functional meniscal tissue, information gained from studies like ours on the regulatory factors and pathways used in meniscal development will greatly enhance these efforts.
We thank Dr. Richard Maas (Harvard Medical School) for the use of the laser capture microdissection instrument. We also thank the Partners HealthCare Center for Personalized Genetic Medicine (Cambridge, MA) for assistance with microarray analysis.