Gene Delivery by Viral Vectors in Primary Cultures of Lacrimal Gland Tissue

Abstract

purpose. To test the feasibility of gene transfer into lacrimal gland tissue in primary culture, using different viral vectors.

methods. Lacrimal glands were dissected from adult Sabra rats and divided by pincers to 0.3–0.4 mm fragments. Tissue was maintained under primary organ culture conditions using the “raft” technique. The ability of three different viral vectors to conduct β-galactosidase (β-gal) gene delivery was examined: adenovirus (Ad5CMVLacZ), vaccinia (VSC9), and herpesvirus (tkLTRZ₁). Tissue fragments were incubated for 60 minutes with one of the viral vectors and transferred to fresh medium. After 3 and 7 days, β-gal expression was examined by X-gal staining in gross preparations and in histologic sections.

results. At 3 days, β-gal expression was observed in 33% of tissue fragments exposed to the vaccinia vector and in 18% and 14% of fragments exposed to the adenoviral and herpes vectors, respectively. After 7 days in culture, successful gene delivery occurred in 77% of vaccinia, 41% of adenovirus, and only 13% of herpesvirus applications. Vector-specific reporter gene expression patterns were observed: With the vaccinia vector, lacrimal duct cells were predominantly stained; in contrast, the adenoviral vector tended to transduce the interacinar areas, with β-gal expression mainly occurring within the myoepithelial cells.

conclusions. Vaccinia and adenovirus are efficient vectors for gene transfer into lacrimal gland tissue in primary culture. The specific expression pattern obtained by the vaccinia vector probably reflects its characteristic tissue tropism to lacrimal duct cells. The results presented in this ex vivo system may be a first step toward expressing genes with products that could be continuously delivered to the eye through the tears. Such proteins could include anti-inflammatory, anti-angiogenic, anti-herpetic, anti-bacterial, or anti-glaucomatous agents, among others.

Many diseases of the eye such as glaucoma, dry eye, keratitis, uveitis and others, require chronic, long-term local administration of different therapeutic preparations. Among these medications are lubricants, IOP-lowering drugs, anti-inflammatory drugs, antimicrobials, and steroids. The traditional and basic way of delivery of these preparations is by topical application of eye drops and eye ointments. Although there are great advantages to such delivery, especially in minimizing systemic side effects and increasing efficacy, there are also many drawbacks to this type of treatment. Among them are large fluctuations in concentration of the active ingredient, short residence time of the medication on the eye, frequent applications that are cumbersome for the patient and often lead to low compliance, allergies and toxicity that may arise from preservatives or vehicle components. ¹  

A potential alternative approach that could allow continuous local delivery of different substances is to modify the lacrimal glands genetically to express and chronically secrete the necessary products. This approach could be a means to modulate tear composition and also secretion rate. Gene delivery and expression have been extensively studied as a therapeutic approach in different human diseases, mainly genetic in origin. ² In the eye, gene therapy of inherited retinal degenerations has shown promise in a number of animal models, ^{3 4} and our group has shown the feasibility of gene transfer by viral vectors to ocular blood vessels in an animal model of retinopathy of prematurity. ⁵ Androgen regulation in primary cultures of lacrimal gland was shown by use of recombinant adenoviral vectors, ⁶ but gene delivery and expression to modulate lacrimal gland function for therapeutic purposes has not been extensively explored. 

The purpose of the present study was to test the feasibility of gene transfer into lacrimal gland tissue in primary culture, using different viral vectors. For this, we adopted a modification of the methodology introduced by Hunt et al. ⁷ for maintaining lacrimal gland tissue fragments in culture for extended periods. Using electron microscopy, they showed that structural integrity and polarity of the tissue is preserved in organ culture for up to 22 days, as is secretory activity. Our results indicate that the three viral vectors examined—vaccinia, adenovirus, and herpes simplex—were capable of delivering a reporter gene to lacrimal gland tissue under these ex vivo conditions, albeit at different efficiency and tropism. 

Methods

Lacrimal Gland Organ Cultures

A modification of the lacrimal gland fragment organ culture technique introduced by Hunt et al. ⁷ was used. 

Lacrimal glands were dissected from adult Sabra rats and divided by pincers to 0.3–0.4 mm fragments. Animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The “raft” culture technique ⁸ was applied: Between three and six tissue fragments were suspended on filter paper (Hogla-Kimberly Ltd., Ramle, Israel) within 4-cm center-well organ culture plates (model 353037; Falcon-BD Biosciences, Bedford, MA). DMEM medium (Beit Haemek Industries, Kibbutz Beit-Haemek, Israel) was in contact with the filter paper from below, with 10% fetal calf serum, 100 U/mL penicillin, 100 μg/mL streptomycin, 1% glutamine, and 10 mM HEPES buffer to maintain pH at 7.0. The tissue fragments were thus at the interface between liquid media from below and air above. Plates were maintained in an incubator at 37°C and under 5% CO₂. Viability of the tissue was assessed by the tetrazolium dye 3-(4,5-dfimethylthiazol-2-yl)-2,5-diphyenyltetrazolium bromide mitochondrial assay (MTT) adapted for organ culture as described by Connelly et al. ⁹ The MTT dye, reduced by active mitochondria and other cellular enzymes, forms a dark purple precipitate in viable areas within the tissue fragments. Fragments assayed two hours after excision served as a positive control, showing that living cells are present throughout the specimen at this time (Fig. 1A) . Freshly excised fragments maintained for 2 hours in distilled water to cause cellular death served as a negative control, and, indeed, no MTT staining was seen (Fig. 1B) . After 7 to 10 days in culture, viability of the tissue was largely maintained. Figures 1C and 1D exemplify the range of MTT staining observed at this time. Whereas viability was reduced to a certain degree in some fragments (Fig 1C) , large viable regions could be identified in all cases and staining was often present throughout the tissue (Fig. 1D) . 

Viral Vectors

The ability of three different viral vectors to transfer the β-galactosidase (β-gal) reporter gene to lacrimal gland tissue was examined: adenovirus (Ad5CMVLacZ, 10⁸ plaque-forming units [pfu]/mL), vaccinia (VSC9, 1–2 × 10⁸ pfu/mL), and herpesvirus (tkLTRZ₁, 8 × 10⁸ pfu/mL). The adenoviral vector containing the β-gal reporter gene under the control of the cytomegalovirus (CMV) promoter was E1A-defective and propagated in 293 cells. The vaccinia expression vector, constructed with the β-gal gene, was under control of a vaccinia early promoter. The herpes simplex virus type I vector contained β-gal under control of the murine leukemia virus (MuLV) LTR promoter inserted at the thymidine kinase gene locus. ⁵  

Study Design and Analyses

One of the viral vectors (or mock medium without virus as control), was added to the growth medium and three to six lacrimal gland tissue fragments were incubated with the vector at 37°C for 60 minutes in small wells (96-well culture plates, Falcon-BD Biosciences). Final concentrations in the incubating medium of the adenovirus, vaccinia, and herpes virus were 10⁷ pfu/mL. After infection, tissue fragments were transferred to fresh virus-free medium using the raft technique described earlier. Expression of β-gal was examined by X-gal staining in gross preparations and in histologic sections, as previously described. ⁵ Expression of β-gal could be identified as early as 24 hours after exposure to the adenoviral and vaccinia vectors, but only to a limited extent. Therefore, levels of expression were examined at 3 and 7 days after exposure. Two independent observers performed counting of expression-positive tissue fragments. A tissue fragment was considered positive (i.e., successful reporter gene transfer and expression occurred) only if histologic (microscopic) examination showed clear expression of β-gal within specific cells or structures. Background carmine staining assisted in identification of tissue structure. In a number of cases, immunostaining with CD34 (Signet QBEnd/10) was used to identify endothelial cells within blood vessels according to methods previously published. ¹⁰  

All in all, more than 260 lacrimal gland tissue fragments in 74 different plates were examined in 10 separate experimental cycles. Twenty-six plates were exposed to vaccinia, 20 to adenovirus, and 15 to herpes virus. Thirteen plates served as three separate controls to verify that background staining or artifacts that may be mistaken as positivity for X-gal did not occur: In four plates, exposure to vaccinia virus for 3 and 7 days without X-gal staining (two plates at each time point); in five plates, no exposure to virus and then X-gal staining at 3 days (two plates) and 7 days (three plates); and in an additional four plates, no virus or X-gal staining. 

Results

Three days after exposure to the different viral vectors under primary organ culture conditions, β-gal expression was observed in 33% of lacrimal gland tissue fragments exposed to the vaccinia vector, in 18% of tissue fragments exposed to the adenoviral vector, and in 14% of fragments exposed to the herpes viral vector (Table 1 , left). After 7 days in culture, β-gal expression was observed in 77% of cases of vaccinia, 41% of cases of adenovirus, and only in 13% of herpesvirus applications (Table 1 , right). Identification of β-gal-positive tissue fragments was highly reliable. The two independent observers differed on only two tissue specimens in which extent of expression was relatively small. 

The different viral vectors exhibited different reporter gene expression patterns. After exposure to the vaccinia vector, tubular β-gal-positive structures were clearly seen under low magnification in gross lacrimal gland fragments (Fig. 2A) . After adenovirus gene delivery, the gross pattern of lacZ expression was different, showing up as irregular spots within the tissue specimens (Fig. 2B) . 

On histologic examination, β-gal expression was clearly seen in lacrimal duct cells at 3 and 7 days after exposure to the vaccinia vector (Figs. 3A 3B) . Immunostaining using the CD34 antibody to identify endothelial cells showed blood vessels adjacent to the β-gal-expressing tubular structures, supporting their identification as lacrimal ducts (Fig. 3C) . At the 7-day time point (but not at 3 days), β-gal expression was also detected in some acinar cells (Fig. 3D) . 

In contrast, with the adenovirus vector, staining was especially prominent in the interacinar areas, mainly in the myoepithelial cells surrounding the acini (Fig. 4A) . Some expression was also noted in acinar cells, but this was often accompanied by degradation of these cells (Fig. 4B) . 

Three different controls were examined: exposure to vaccinia virus for 3 and 7 days without X-gal staining, no exposure to virus (mock) followed by X-gal staining at 3 and 7 days, and a no virus-no staining group. In all cases (13 different plates, 54 tissue fragments) no discrete X-gal staining was seen in gross preparations or histologic sections (data not shown). 

Discussion

In the present study, vaccinia and adenovirus were shown to be efficient vectors for gene transfer to lacrimal gland tissue in an ex vivo system. Vector-specific expression patterns of reporter genes were observed. The vaccinia vector preferentially delivered the β-gal gene to lacrimal duct cells and acini. The adenoviral vector tended to yield expression mainly within the myoepithelial cells surrounding the acini. Because ubiquitous, non-cell-specific promoters were used in all viral vectors, this apparent selectivity most probably reflects differential tropism of the viruses to specific cell types within the glands. 

One of the potential problems that should be addressed when viral-vector-mediated gene delivery is considered is possible toxicity of the vector to the transfected cells. The degradation observed in acinar cells expressing β-gal after exposure to the adenovirus vector in the present study may well be the result of such toxicity. Similar cellular toxicity is a known potential effect of E1-deleted first-generation adenoviral vectors, such as the one we used in the present study. ¹¹ New high-capacity “gutless” adenoviral vectors have recently been developed that may circumvent this problem (in addition to allowing delivery of larger constructs). ¹² Use of improved vectors such as these can enhance the efficacy and safety of viral-vector-mediated gene delivery. Other challenges in applying this type of gene delivery technique in vivo include control of the inflammatory response associated with viral vector use and attaining long-term, stable expression of the delivered gene. These problems may also be better addressed with more advanced viral vectors than the ones used in this study. As an alternative, new gene transfer methods that are not dependent on viral vectors, such as complexing the DNA expression cassette to cationic liposomes or cationic polymers, could be used to circumvent these viral vector-related difficulties. ¹³  

Extrapolation of the results from this ex vivo study to in vivo conditions would also require modification and development of techniques to introduce the vector directly into the intact lacrimal gland, as opposed to simple, direct exposure in the culture dish. One option would be to inject the vector directly into the gland, but a better possibility could be to cannulate the glandular excretory tear duct and perform retrograde injection. Such cannulation would also allow collection of tears to examine whether the delivered gene product was indeed expressed. ¹⁴ A similar approach has been taken in gene transfer to salivary glands. ¹⁵  

To the best of our knowledge, this is the first report to examine the feasibility of gene transfer to lacrimal gland tissue with the purpose of ultimately using this technique for therapeutic purposes. Genes encoding secreted proteins could be delivered to lacrimal glands with the goal of maintaining stable therapeutic levels of the delivered gene product in the tears. Thus, as an example, one can envision continuous delivery by the tears of products such as interferon in patients with recurrent herpes infections. ¹⁶ Alternatively, genes may be delivered to modify lacrimal gland function, to affect volume of tear secretion or tear composition. By applying the recently developed transcription control switches such as the “tet-on” systems, temporal control can be gained over these effects. ¹⁷ Clearly, much remains to be done to achieve these goals. The feasibility of in vivo gene delivery to the lacrimal gland remains to be shown, toxicity and side effects must be assessed, and the efficacy in the treatment of ocular diseases has to be determined. 

In conclusion, evidence for the feasibility of gene delivery to lacrimal gland tissue is provided in the present ex vivo study. This may be the first step toward expressing genes whose products could be continuously delivered to the eye through the tears. 

 Supported in part by the Yedidut and Yael Research Grants.

 Submitted for publication June 3, 2002; revised October 19, 2002; accepted November 7, 2002.

 Disclosure: E. Banin, None; A. Obolensky, None; E. Piontek, None; H. Falk, None; E. Pikarsky, None; J. Pe’er, None; A. Panet, None; I. Chowers, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Viability of lacrimal gland tissue fragments in organ culture using the raft technique. Cell viability was assessed by mitochondrial tetrazolium dye reduction (MTT). (A) MTT staining performed in a tissue fragment 2 hours after excision showed widespread reduction of the dye, with formation of dark-purple precipitate, indicating living cells throughout the specimen. (B) In contrast, a freshly excised fragment that was maintained for 2 hours in distilled water to cause cellular death, shows no MTT staining. (C, D) After 7 to 10 days in culture, viable regions were identified in all fragments. In some, viability was reduced to a certain degree (C), but in others viability was very well maintained throughout (D).

Table 1.

 

Expression of β-Gal 3 and 7 Days after Exposure to the Different Viral Vectors

Viral Vector	Number of Exposed Specimens 3 days	β-Gal-Positive Specimens* (Yield in %), † 3 days	Number of Exposed Specimens 7 days	β-Gal-Positive Specimens* (Yield in %), † 7 days
Vaccinia	45	15 (33)	53	41 (77)
Adenovirus	33	6 (18)	22	9 (41)
Herpes	28	4 (14)	32	4 (13)

Figure 2.

 

Gross preparations of lacrimal gland tissue three days after exposure to two different viral vectors. (A) Note β-gal expression (blue) in tubular structures (arrows) after exposure to the vaccinia vector. (B) In contrast, after transduction with the adenovirus vector, an irregular spotted pattern of β-gal expression (arrows) was seen within the tissue.

Figure 3.

 

Gene expression pattern of the vaccinia vector at the cellular level. (A, B) Histologic sections of lacrimal gland tissue 3 days after exposure to the vaccinia vector, stained with carmine. β-Gal expression was seen within tubular structures (black arrows), the lacrimal ducts, sectioned crosswise (A) and lengthwise (B). Adjacent to the stained ductal structures, blood vessels were visible (red arrows). (C) Immunostaining with CD34 to identify endothelial cells (red arrow). Note that β-gal expressing cells were those of the adjacent lacrimal ducts (black arrows) and not the blood vessels. (D) Seven days after exposure to the vaccinia virus vector, β-gal expression was also occasionally present within acinar cells.

Figure 4.

 

Gene expression pattern of the adenovirus vector. Histologic sections stained with carmine, 7 days after exposure. (A) β-Gal expression was especially prominent in myoepithelial cells surrounding the acini (arrows). (B) Some cells within the acini were also occasionally transfected. Such cells often showed signs of degradation.