The pp38 Gene of Marek's Disease Virus (MDV) Is Necessary for Cytolytic Infection of B Cells and Maintenance of the Transformed State but Not for Cytolytic Infection of the Feather Follicle Epithelium and Horizontal Spread of MDV

Abstract

Marek's disease virus has a unique phosphoprotein, pp38, which is suspected to play an important role in Marek's disease pathogenesis. The objective of the present study was to utilize a mutant virus lacking the pp38 gene (rMd5Δpp38) to better characterize the biological function of pp38. This work shows that the pp38 gene is necessary to establish cytolytic infection in B cells but not in feather follicle epithelium, to produce an adequate level of latently infected T cells, and to maintain the transformed status in vivo.

The biological function of the Marek's disease (MD) virus (MDV) unique pp38 gene in the pathogenesis of MD is poorly understood. pp38 has been associated with lymphoid tropism (18), reactivation from latency (19, 21), maintenance of tumors (20), and immunosuppression (8). Three recombinant viruses (rMd5, rMd5Δpp38, and rMd5/pp38CVI) were used in this work to expand the previous knowledge of the role of pp38 in viral replication and transmission, transformation and viral regulation of apoptosis, latency, reactivation, and cell cycle regulation. Construction of the recombinant viruses rMd5 and rMd5Δpp38 has been described previously (15). The construction of rMd5/pp38CVI is described elsewhere (L. F. Lee, X. Cai, I. M. Gimeno, and S. M. Reddy, submitted for publication). Briefly, the RecA-assisted restriction endonuclease cleavage procedure was used to insert the pp38 gene from CVI988/Rispens at the same site that previously contained the Md5 pp38 gene, and insertion of the pp38 gene at the correct site was later confirmed by PCR and DNA sequencing. In vitro and in vivo expression of pp38 in rMd5/pp38CVI was confirmed by immunostaining with the monoclonal antibody T65 (L. F. Lee, unpublished data) that specifically reacts with CVI988/Rispens pp38.

To study replication and transmission, 1-day-old 15x7 chickens (4) were inoculated with either rMd5, rMd5Δpp38, or rMd5/pp38CVI. Horizontal virus transmission was monitored by using contact chickens. Samples of bursa, thymus, and spleen were collected at 3, 6, and 9 days postinoculation (dpi). At 26 dpi, samples of feather follicle epithelium (FFE) and spleens of contact chickens were collected. Monoclonal antibodies 1AN86.17 and T81 (17) specific to MDV glycoprotein B (gB) and ribonucleotide reductase (RR), respectively, were used to study MDV replication. At 6 dpi, late-antigen gB and early-antigen RR expression could not be detected in bursa of Fabricius, thymus, or spleen after inoculation with rMd5Δpp38 (Fig. 1A), but both viral antigens were highly expressed in all tested chickens following rMd5 or rMd5/pp38CVI inoculation (Fig. 1C and E). The failure of rMd5Δpp38 to induce cytolytic infection was cell type specific because B lymphocytes but not FFE were affected (Fig. 1B, D, and F). Horizontal virus transmission was confirmed by isolating MDV from every contact chicken (data not shown).

FIG. 1.

Expression of RR and gB in various tissues. (A) Expression of RR in the bursa of Fabricius 6 days after rMd5Δpp38 inoculation. (B) Expression of gB in the FFE 26 days after rMd5Δpp38 inoculation. (C) Expression of RR in the bursa of Fabricius 6 days after rMd5 inoculation. (D) Expression of gB in the FFE 26 days after rMd5 inoculation. (E) Expression of RR in the bursa of Fabricius 6 days after rMd5/pp38CVI inoculation. (F) Expression of gB in the FFE 26 days after rMd5/pp38CVI inoculation. The scale of each image is represented in micrometers by the bar in the lower right corner of each panel.

To study the effect of the pp38 gene on transformation and apoptosis, two experiments were conducted. In experiment 1, 1-day-old 15x7 chickens (4) were inoculated with either rMd5 or rMd5Δpp38. Half of the animals in each treatment group were euthanized at 6 weeks postinoculation (wpi), and the other half were euthanized at 15 wpi. Samples of vagus, brachial and sciatic nerves, gonad, lung, liver, and spleen were collected for histopathology and immunohistochemistry and for the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) assay. Experiment 2 was conducted to confirm that differences between rMd5Δpp38 and rMd5 were due solely to the deletion of pp38. The experimental design was the same as that for replicate 1 but included chickens inoculated with rMd5/pp38CVI.

The frequency of gross lesions induced by rMd5Δpp38 and rMd5 was evaluated at 6 and 15 wpi in replicate 1 (Table 1). rMd5 induced a high frequency of lymphoproliferative lesions at 6 wpi (60 to 80%) that tended to increase at 15 wpi (up to 100%). rMd5Δpp38 induced microscopic lymphoproliferative lesions in only 25% of the chickens at 6 wpi. The frequency of lesions induced by rMd5Δpp38 also tended to increase at 15 wpi (up to 44%) but was significantly lower (P < 0.05) than that in the group inoculated with rMd5. Most of the lymphoproliferative lesions induced by rMd5Δpp38 did not progress to gross tumors. This contrasted with the lymphoproliferative lesions induced by rMd5, which at 6 and 15 wpi were evident gross and histological lesions. No differences in the frequency of lymphoproliferative lesions between rMd5/pp38CVI and rMd5 were found at either 6 or 15 wpi in replicate 2 (Table 1). The phenotype of transformed cells in tumors induced by rMd5, rMd5/pp38CVI, and rMd5Δpp38 was in all cases CD4⁺ CD8⁻ Meq⁺ MATSA⁺ (6, 11, 12; data not shown). In the lymphoproliferative lesions induced by rMd5Δpp38 that could be detected only by microscopic examination, many cells had homogeneously condensed chromatin, and numerous apoptotic bodies could be detected (Fig. 2B). Fragmentation of DNA was confirmed by the TUNEL technique (Fig. 2E). Lymphoproliferative lesions induced by either rMd5 or rMd5/pp38CVI did not show morphological evidence of apoptosis (Fig. 2A and C), and fragmentation of DNA could not be found by the TUNEL assay (Fig. 2D and F). Many viruses have developed different mechanisms to block apoptosis (1-3, 5, 7, 10, 14). This study shows that pp38 is involved in blocking apoptosis, and this might at least partially explain the inability of lymphoproliferative lesions to progress to gross tumors. However, it is unclear whether pp38 is directly blocking apoptosis or whether it is interfering with the cytotoxic T lymphocyte response against MD tumors.

TABLE 1.

Frequency of lesions at 6 and 15 wpi^a

Replicate	Inoculum	Chickens (%) with indicated type of lesions at:
		6 wpi				15 wpi
		Visceral		Nerve		Visceral		Nerve
		Gross	Histological	Gross	Histological	Gross	Histological	Gross	Histological
1	rMd5Δpp38	0a	25ab	0a	0a	6a	44b	25ab	37a
	rMd5	60b	60b	73b	80b	79b	100a	100b	100b
	Control	0a	0a	0a	0a	0a	0c	0a	0c
2	rMd5Δpp38	0a	20ab	0a	0a	0a	60ab	8a	40ab
	rMd5	43b	80b	86b	100b	60b	100b	100b	100b
	rMd5/pp38CVI	35b	80b	76b	80b	41b	100b	100b	100b
	Control	0a	0a	0a	0a	0a	0a	0a	0a

^aData indicate the percentage of chickens that developed lesions in each group at 6 and 15 wpi. In replicate 1, there were 34 chickens per treatment and for each time point (6 and 15 wpi). In replicate 2, there were 17 chickens per treatment and for each time point (6 and 15 wpi). Differences between treatment results were statistically studied (analysis of variance test, P < 0.05). The same letter placed next to numbers within a replicate means that the differences were not significant.

FIG. 2.

Lymphoproliferative lesions in viscera and nerves. (A) Hematoxylin and eosin (H&E)-stained section of kidney of rMd5-inoculated chicken. (B) H&E-stained section of ovary of rMd5Δpp38-inoculated chicken. (C) H&E-stained section of ovary of rMd5/pp38CVI-inoculated chicken. (D) TUNEL assay in a section of lung of rMd5-inoculated chicken. (E) TUNEL assay in a section of lung of rMd5Δpp38-inoculated chicken. (F) TUNEL assay in a section of lung of rMd5/pp38CVI-inoculated chicken. (G) H&E-stained section of a nerve of rMd5-inoculated chicken. (H) H&E-stained section of a nerve of rMd5Δpp38-inoculated chicken. (I) H&E-stained section of a nerve of rMd5/pp38CVI-inoculated chicken. The scale of each image is represented in micrometers by the bar on the lower right corner of each panel.

Nerve lesions induced by rMd5Δpp38 differed from those induced by rMd5. While rMd5 and rMd5/pp38CVI induced typically neoplastic lesions (type A) (Fig. 2G and I), rMd5Δpp38 induced more inflammatory lesions (type B) characterized by abundant edema and extensive infiltration of plasma cells (Fig. 2H). It is generally considered that type B lesions have an important autoimmune component (13). Our results provide additional evidence for the relevance of pp38 in regulating the immune response and suggest the need for further study.

To study the role of pp38 on latency, reactivation, and cell cycle, 1-day-old 15x7 chickens (4) were inoculated with either rMd5 or rMd5Δpp38. Spleens were collected at 3, 6, 14, and 26 dpi to study the distribution of the splenocytes in the different phases of the cell cycle by propidium iodide staining and subsequent flow cytometry analysis. We have found that infection with either rMd5 or rMd5Δpp38 increased the percentage of splenocytes in the S phase as early as 3 dpi, but no difference between the two viruses was found (Table 2). It is not possible to determine from our results if MDV is regulating the cell cycle or if there is a change in the spleen cell populations after MDV infection. Seven additional spleens were collected 14 dpi for isolation by plaque assays and real-time PCR (9). The standard plaque assay was a good indicator of the ability of latent MDV to reactivate in cell culture, while real-time PCR is a relative measure of the number of viral genome copies in the sample. Our results show that regardless of the lower number of latently infected T cells in chickens inoculated with rMd5Δpp38 (10-fold lower), the ability of the virus to reactivate from latency in cell culture was not affected (i.e., 10-fold lower reactivation, but also relatively 10-fold fewer copies of the viral genome). Therefore, our results do not support a role of pp38 in reactivation (16), since no differences between rMd5 and rMd5Δpp38 in the ability to reactivate from latency were found.

TABLE 2.

Percentages of splenocytes at different stages of the cell cycle after infection with rMd5 and rMd5Δpp38^a

No. of dpi	Group	Splenocytes (%) at indicated stageb
		G0/G1	S	G2/M
3	rMd5Δpp38	61.30a	19.84b	17.95a
	rMd5	72.49a	13.21ab	13.42b
	Control	74.72a	4.33a	20.10a
6	rMd5Δpp38	69.97a	5.06a	24.20a
	rMd5	71.45a	7.21b	20.14b
	Control	71.17a	3.94a	24.37a
14	rMd5Δpp38	80.58a	3.11a	15.37a
	rMd5	78.34a	5.00b	14.81a
	Control	77.42a	3.15a	16.65a
26	rMd5Δpp38	76.09b	5.22b	14.98ab
	rMd5	73.88b	7.08c	13.71b
	Control	79.41a	2.88a	16.71a

^aThe values of seven chickens for each treatment, killed at different time points, were averaged.

^bDifferences between treatment results were statistically studied (analysis of variance test, P < 0.05): The same letter placed next to numbers means that the differences were not significant.

In conclusion, our results show that pp38 is required to establish cytolytic infection in B lymphocytes but not in FFE, to produce an adequate level of latently infected T cells, and to maintain the transformed status of lymphocytes by preventing apoptosis. In this work, the deletion of pp38 results in an early onset of B-type lesions in the nerves and indicates that pp38 might be involved in immunomodulating the immune response against MD. Taken together, these findings indicate that rMd5Δpp38 might have the potential of being used as a model to study regression of tumors and the immune response against MDV and MDV-induced tumors.