Example of detection of human insulinlike growth factor 1 gene (hIGF1) transcripts and polymerase chain reaction detection of hIGF1 plasmid vector. A, Reverse transcriptase polymerase chain reaction shows a positive band (458 base pairs [bp]) from hIGF1 transcripts in thyroarytenoid muscle after multiple injections of hIGF1 formulation (left lane) and a negative control (right lane). B, Polymerase chain reaction shows a positive band (490 bp) from hIGF1 plasmid DNA in thyroarytenoid muscle after multiple injections of hIGF1 formulation (left lane) and a negative control (right lane).
Representative transverse sections of thyroarytenoid muscle specimens stained for adenosine triphosphatase at pH 10.4 (original magnification ×400) and taken from (A) denervated control animal, (B) denervated animal given a single injection of the hIGF1 gene, (C) denervated animal given multiple injections of the hIGF1 gene, and (D) normal animal.
Comparison of the mean lesser diameter of thyroarytenoid muscle fiber among the denervated control group, animals given a single injection of the hIGF1 gene (SIIGF), animals given multiple injections of hIGF1 (MIIGF), and normal animals. Values are given as mean±SD. Asterisk indicates the difference is significant (P <.05) compared with control group.
Representative axial sections of thyroarytenoid muscle specimens stained with acetylcholinesterase-silver or gold (original magnification ×400) and taken from (A) denervated control animal, (B) denervated animal given a single injection of the hIGF1 gene, (C) denervated animal given multiple injections of hIGF1, and (D) normal animal.
Comparison of the motor endplate length in the thyroarytenoid muscles of the denervated control group, the group given a single injection of the hIGF1 gene (SIIGF), the group given multiple injections of hIGF1 (MIIGF), and normal animals. Values are given as mean±SD. Asterisk indicates a significant increase (P <.05) compared with control group.
Comparison of percentage of endplates with nerve contact in the thyroarytenoid muscle of animals receiving single injections of the hIGF1 gene (SIIGF), multiple injections of hIGF1 (MIIGF), and control groups. Values are given as mean±SD. Asterisk indicates a significant increase (P <.05) compared with control group.
Shiotani A, O'Malley BW, Coleman ME, Flint PW. Human Insulinlike Growth Factor 1 Gene Transfer Into Paralyzed Rat LarynxSingle vs Multiple Injection. Arch Otolaryngol Head Neck Surg. 1999;125(5):555–560. doi:10.1001/archotol.125.5.555
To compare the biological effects of single vs multiple treatment of rat denervated laryngeal muscle with human insulinlike growth factor 1 (hIGF1) gene therapy.
Experimental Methods or Design
A muscle-specific nonviral vector containing the α-actin promoter and hIGF1 gene formulated with polyvinyl polymers was injected into denervated adult rat thyroarytenoid muscle. The effects on animals given a single injection (n=16) vs those given multiple injections (n=14) vs control groups (n=18) were evaluated. Twenty-eight days after the first injection, gene expression, muscle fiber size, motor endplate length, and nerve-to-motor endplate contact were evaluated.
Gene expression, detected by reverse transcriptase polymerase chain reaction for hIGF1 messenger RNA, occurred in 13 (81%) of 16 animals receiving single injections and 14 (100%) of 14 animals receiving multiple injections. Compared with controls, hIGF1– transfected animals in both single- and multiple-injection groups had a significant increase in the lesser diameter of muscle fiber, a significant decrease in motor endplate length, and a significant increase in the percentage of endplates with nerve contact (P <.05 for all). There was no statistical difference between single- and multiple-injection groups.
Applied to laryngeal paralysis, hIGF1 gene therapy provides an opportunity to augment surgical treatment modalities by the prevention or reversal of muscle atrophy, and enhancement of nerve sprouting and muscle reinnervation. Although the percentage of denervated muscles demonstrating hIGF1 expression was increased following multiple injections, no difference was observed in the biological response compared with that in the single-injection treatment groups. Further investigation will be conducted to assess long-term benefits and physiological responses and to define the limitations of this potentially valuable therapeutic strategy.
SURGICAL treatments of unilateral laryngeal paralysis—vocal-fold injection, thyroplasty, and arytenoid adduction—achieve vocal-fold medialization1 by static change to the vocal-fold tissue or laryngeal framework. The surgical outcome with these procedures may be limited by muscle atrophy associated with denervation. Laryngeal reinnervation procedures have had little effect on the return of dynamic laryngeal function and are not widely accepted as treatment options. The failure of reinnervation procedures to achieve functional results reflects multiple factors, including a decrease in motor fiber density, existing muscle atrophy, central motoneuron loss, and inappropriate or misdirected innervation by antagonistic motoneurons.2,3
A multitude of neurotrophic factors have been shown to prevent central motoneuron loss, enhance nerve sprouting, preserve motor endplate structure, and promote reinnervation.4- 6 Of these, insulinlike growth factor 1 (IGF-1) is both neurotrophic and myotrophic and plays an important role in the maintenance and regeneration of muscle and peripheral nerves.6- 8 The use of IGF-1 and other trophic factors may prove beneficial for the treatment of peripheral nerve injuries. The clinical application of trophic factor therapy, however, is limited by bioavailability, the inability to achieve steady-state levels, and toxic effects associated with parenteral delivery.9,10 Gene therapy provides a mechanism whereby infrequent administration achieves long-term steady-state levels of the gene product using common routes of delivery.11,12
In previous investigations,13,14 a rat model of laryngeal paralysis was established to evaluate the potential of muscle-specific nonviral gene therapy based on human IGF-1 (hIGF1) plasmid formulated with povidone (polyvinylpyrrolidone). A single injection of the formulation produced comparable effects on muscle fiber diameter and motor endplate structure to those achieved using a cumbersome model of daily recombinant IGF-1 protein injections for 3 to 6 weeks.15,16
In the present study, we compare the effects of hIGF1 gene transfer after single and multiple injections. Local expression of hIGF1 is confirmed using reverse transcriptase–polymerase chain reaction (RT-PCR) to detect hIGF1 messenger RNA (mRNA). The biological effect is assessed by measuring muscle fiber diameter, motor endplate length, and nerve-to-motor endplate contact.
Details and construction of the muscle-specific α-actin– hIGF1 hybrid gene have been previously described.17,18 The construct consists of a chicken skeletal α-actin promoter, the hIGF1 coding sequence, and human growth hormone 3‘-UTR region. The promoter region is skeletal muscle specific and up-regulated by the gene product hIGF-1, and the human growth hormone 3‘-UTR region enhances the secretion of the hIGF-1 protein product and, thus, increases the paracrine effects of hIGF1. The secretion of the gene product will, theoretically, produce a more homogeneous response in muscle tissue at effectively lower doses.
The construct was cloned into the kanamycin-resistant plasmid backbone pBluescript II KS (Stratagene, La Jolla, Calif), which contains the plasmid origin of replication and kanamycin-resistance gene to avoid the use of lactams in production. The construct is used in formulation with povidone19 to enhance delivery into muscle cells.
Institutional guidelines, in accordance with the Animal Welfare Act of 1966 and all subsequent revisions, including those made in 1985, and National Institutes of Health guidelines were followed for the handling and care of laboratory animals.
Sprague-Dawley rats (150-200 g) were randomly assigned to a group receiving hIGF1 in a single injection (8 animals for the fiber diameter study and 8 animals for the motor endplate study), a group receiving hIGF1 in multiple injections (7 animals for the fiber diameter study and 7 animals for the motor endplate study), and a control group (10 animals for the fiber diameter study and 8 animals for the motor endplate study). Rats in the control group received either a β-galactosidase formulation or isotonic sodium chloride solution. Animals were anesthetized by the intramuscular administration of ketamine hydrochloride (35 mg/kg) and xylazine hydrochloride (5 mg/kg). Using an operating microscope, the larynx was approached through a midline incision. After the left recurrent laryngeal nerve and the left superior laryngeal nerve were identified, a 1-cm segment was removed from the left recurrent laryngeal nerve, and the proximal and distal ends were ligated with 10-0 nylon sutures. The left superior laryngeal nerve was also divided. A midline thyrotomy was performed, and the laryngeal cavity was visualized to confirm vocal-fold motion impairment. Vector solution (hIGF1 or β-galactosidase, 90 µg of DNA) (30 mL) or sodium chloride solution (30 mL) was injected into the left thyroarytenoid muscle using a Hamilton syringe with a 30-gauge needle. The larynx was closed using 10-0 nylon suture, and the skin was also sutured. Rats in the multiple-injection group received 2 more injections 7 and 14 days after the initial injection (total, 3 injections).
After survival for 28 days from denervation and the first injection, the animals were killed using a lethal dose of sodium pentobarbital. The whole larynx was excised andimmediately snap-frozen with OCT (Miles, Elkhart, Ill) inliquid nitrogen, and the larynges were stored at − 80°C until frozen sectioning.
Sequential axial sections cut 35-µm thick and coronal sections cut 12-µm thick were placed on lysine-coated slides sequentially and air dried. Every fifth axial section was processed for acetylcholinesterase nerve staining, and every seventh coronal section was processed for adenosine triphosphatase staining. The remaining sections were stored at − 80°C.
Coronal sections were stained with adenosine triphosphatase using the method of Guth and Samaha.20 When stained at pH 10.4, all fibers in the thyroarytenoid muscle are fast type,21 darkly staining, and have adequate contrast for computer-assisted measurement.
Measurements were obtained from the midbody of the thyroarytenoid muscle relative to the anterior aspect of the vocal process. Five randomly selected high-power (×400) views were digitally recorded. All complete fibers within the field of view were measured. The lesser diameter was obtained for comparison between treated and control specimens. The lesser-fiber diameter is the maximum diameter across the lesser aspect of the muscle fiber. This measurement overcomes the distortion that occurs when a muscle fiber is cut obliquely and is considered more accurate than the measured fiber area.22 A minimum of 146 (mean, 213.9) fibers per animal were counted in hIGF1 -treated and control groups. The investigator (A.S.) was blinded to the treatment of the animals.
Axial sections were stained to identify motor endplates and peripheral nerves in thyroarytenoid muscle using the acetylcholinesterase staining method and the modified silver-gold impregnation method of Pestronk and Drachman.23,24 This staining produces a transparent blue cholinesterase stain in contrast to the black silver–stained axons and nerve terminals.
The effect of hIGF1 gene transfection on neuromuscular junctions was evaluated by measuring the length of motor endplates, as outlined by the cholinesterase stain, and the percentage of endplates with nerve contact. The motor endplate length is a quantitative index of denervation effect.23,24 A minimum of 51 endplates (mean, 81.2) were assessed per animal for motor endplate length, and a minimum of 66 (mean, 159.4) were assessed per animal for the percentage of endplates with nerve contact in hIGF1-treated and control animals. The investigator (A.S.) was blinded to the treatment of the animals.
For morphometric analysis, microscopic images were digitally captured using a CCD camera (Hanamatsu Photonics, Hanamatsu City, Japan) and computer system (Apple Macintosh Quadra 840 AV; Apple Computer Inc, Cupertino, Calif). The lesser diameter of the muscle fiber and motor endplate length were calculated from digitized images using National Institutes of Health software (image 1.58f; Scientific Computing Resource Center, Bethesda, Md).
To evaluate the expression of hIGF1 transcript and the presence of the hIGF1 plasmid vector, RT-PCR and PCR were performed using RNA and DNA obtained from sequential fresh frozen sections on slides stored at −80 C. Sections of hIGF1 gene-injected thyroarytenoid muscle were microdissected from the slide and put into a microcentrifuge tube containing phenol and guanidine isothiocyanate (Trizol) solution (Gibco BRL, Gaithersburg, Md) for RNA and DNA extraction. Total RNA and DNA were isolated according to the manufacturer's instructions.
For RT-PCR, RNA (1 µg) was incubated with DNase I (Gibco BRL) and reverse transcriptase (SuperScript II; Gibco BRL). The hIGF1 transcripts were amplified by 40 cycles of RT-PCR (30 seconds at 94°C, 30 seconds at 67°C, and 30 seconds at 72°C) using a 5‘ primer homologous to the α-actin promoter region (5‘-CGACGCGCAGTCAGCACAG-3‘) and a 3‘ primer homologous to the hIGF1 sequence (5‘-CTGCGGTGGCATGTCACTCTTCA-3‘). There is an intron sequence between 2 exons where these primers lie; therefore, RT-PCR products from the hIGF1 transcript (458 base pairs) can be distinguished from the plasmid PCR product (656 base pairs) on the basis of size. Quantitative RT-PCR was performed for RT-PCR–positive specimens to quantify the copies of transcripts.
For PCR, DNA (0.1 µg) was amplified by 35 cycles of PCR (30 seconds at 94°C, 30 seconds at 62°C, and 30 seconds at 72°C) using a 5‘ primer (5‘-CATGTCCTCCTCGCATCTCT-3‘) and a 3‘ primer (5‘-GGCACTGGAGTGGCAACTTC-3‘) homologous to the hIGF1 gene vector sequence.
The PCR fragments were separated by 2% agarose gel electrophoresis and visualized by gold staining (SYBR Gold; Molecular Probes, Inc, Eugene, Ore).
Mean values of the lesser diameter of the muscle fiber, motor endplate length, and percentage of endplates with nerve contact in each animal were obtained. One-way analysis of variance, followed by the Fisher protected least significant difference test, was performed using commercial software (Statview SE version 1.04; Abacus Concepts, Inc, Berkeley, Calif) to compare differences among the group given hIGF1 as a single injection, the multiple-injection group, and control groups (β-galactosidase and sodium chloride group). Statistical significance is set at P <.05. Values are given as mean±SD. For comparison of the number of transcript copies obtained by quantitative RT-PCR, the Mann-Whitney U test was used, with statistical significance set at P <.05.
To confirm successful gene transfer and transcription, PCR and RT-PCR were used to detect the presence of plasmid construct and mRNA for hIGF1 in injected laryngeal muscles (Figure 1). Four weeks after laryngeal denervation and the first injection of the hIGF1 formulation, PCR analysis identified hIGF1 plasmid DNA in 16 (100%) of 16 animals in the single-injection group and 14 (100%) of 14 animals in the multiple-injection group. Using RT-PCR analysis, hIGF1 mRNA was detected in 13 (81%) of 16 animals in the single-injection group and 14 (100%) of 14 animals in the multiple-injection group. hIGF1 plasmid DNA and mRNA were not detected in control animals. Quantitative RT-PCR for RT-PCR–positive animals showed no significant difference in transcript copies when comparing the 2 groups (P = .37).
Representative cross-sections of the thyroarytenoid muscle specimens stained for adenosine triphosphatase are shown in Figure 2. The mean lesser diameters of muscle fibers in single- and multiple-injection and control groups are presented in Figure 3. The mean lesser diameters of muscle fibers from the single-injection (17.56±0.97 µm) and multiple-injection (16.36±1.60 µm) groups were significantly larger than those of the denervated control group (14.70±1.43 µm) (P <.001 and P =.04, respectively). There was no significant difference between single- and multiple-injection groups in lesser-fiber diameter (P =.10).
Microscopic analysis of sequential sections stained with hematoxylin and fast red revealed no evidence of inflammatory cell infiltrate.
Representative axial sections of thyroarytenoid muscle specimens stained for acetylcholinesterase and silver-gold impregnation are shown in Figure 4. In the denervated condition, cholinesterase-stained areas are elongated and more diffuse. The percentage of endplates with nerve contact is also diminished in the denervated condition.
A significant decrease in motor endplate length was observed in groups given hIGF1 as a single injection (20.88±1.42 µm) and as multiple injections (21.88±2.19 µm) compared with the denervated control group (25.41±3.19 µm) (P =.002 and P =.03, respectively). There was no significant difference between single- and multiple-injection groups in motor endplate length (P =.31) (Figure 5). A significantly increased percentage of endplates with nerve contact was observed for groups given a single injection (20.3%±13.9%) or multiple injections (24.8%±13.8%) compared with the control group (4.4%±4.2%) (P =.008 and P =.002, respectively). There was no significant difference between single- and multiple-injection groups in the percentage of endplates with nerve contact (P =.54) (Figure 6).
The muscle-specific gene delivery system used in this study contains an avian skeletal α-actin promoter and has been shown to increase specific gene expression more than 10-fold in muscle.17,18 The plasmid is formulated with povidone, which has been shown to increase the level of gene transfer expression 10- to 100-fold more than DNA alone.19 This formulation has been modified to produce hIGF1 efficiently and to enhance secretion, resulting in less intracellular accumulation.17 Although this modification may benefit the therapeutic response, it limits our ability to interpret immunohistochemical staining for hIGF-1 protein expression. Furthermore, the size of the rat thyroarytenoid muscle (about 1×3 mm) prohibits quantitative protein measurement, such as Western blot or enzyme-linked immunosorbent assay, when tissue is processed simultaneously for histological examination. Local expression was, therefore, confirmed at the mRNA level (RT-PCR) in thyroarytenoid muscle microdissected from sequential microscopic slides.
Although gene transfer at the DNA level confirmed by PCR was 100% in both groups, in the multiple-injection group, the detection of the hIGF1 transcript was increased to 100% from 81% in the single-injection group. The increase in percentage of animals expressing hIGF1, however, was not associated with an increase in mRNA expression as determined by quantitative RT-PCR.
Compared with controls, hIGF1–transfected animals in both single- and multiple-injection groups had a significant increase in the lesser diameter of muscle fibers, a significant decrease in motor endplate length, and a significant increase in percentage of endplates with nerve contact. These changes are demonstrative of the myotrophic and neurotrophic effects of hIGF1. No significant difference in muscle fiber size and motor endplate structure was observed when comparing single-injection with multiple-injection groups. This observation suggests that a single injection of the hIGF1 formulation is sufficient to achieve therapeutic effect and underscores the potential limitation of multiple-injection strategies.
Equivalent biological effect following single and multiple injections may be attributed to an hIGF1 dose-response plateau reached after the first injection. The relationship between biological effect and the hIGF1 dose has been demonstrated to be linear within a therapeutic window for in vitro study.7 Increased levels do not necessarily increase the biological response in vitro, and excess hIGF1 may even decrease the in vitro activity.7 Alternatively, the absence of dose response following multiple injections may reflect inefficient hIGF1 gene transfer, or a rate-limiting step at the transcription level thus limiting the number of available mRNA copies. Immune or inflammatory responses to the plasmid formulation or hIGF-1 protein product may also limit the biological effect. Although we have not performed specific immune studies to rule out this possibility, microscopic evaluation of histological sections revealed no cellular infiltration or evidence of inflammation in either single- or multiple-injection groups. Similarly, no inflammatory response was observed in rat hind limb muscle after injection of this plasmid formulation.17
Future experiments will focus on extending the number of treatment injections and increasing dosing intervals in an attempt to improve the biological response. Investigations will also be conducted to determine whether a dose-response plateau to hIGF1 exists in vivo. Finally, we are developing an experimental system to assess physiological parameters to evaluate the therapeutic benefit of thyroarytenoid muscle hypertrophy and reinnervation after hIGF1 gene therapy.
Accepted for publication October 27, 1998.
This research is supported in part by grant 5K08 DC0081-05 from the National Institute on Deafness and Other Communication Disorders, Bethesda, Md, and The Cal Ripken/Lou Gehrig Fund for Neuromuscular Research, Baltimore, Md.
We thank Norman Hardman, PhD, and Eric Tomlinson, PhD, for contributing the plasmid formulation used in this study.
Reprints: Paul W. Flint, MD, Department of Otolaryngology–Head and Neck Surgery, Johns Hopkins Medicine, 601 N Caroline St, Baltimore, MD 21287-0910 (e-mail: email@example.com).