HTLV-I-Specific Cytotoxic T Lymphocytes in the Cerebrostinal Fluid of Patients with HTLV-I-associated Neurological Disease Steven Jacobson, PhD," Dale E. McFarlin, MD," Susan Robinson, BS," Rhonda Voskuhl, MD," Roland Martin, MD," Alfred Brewah, MS,t Amy J. Newell, BS,t and Scott Koenig, PhD, MDf

Recently, it has been shown that in patients with human T-cell lymphotropic virus type I (HTLV-&associated neurological disease, high levels of HTLV-I-specific cytotoxic T lymphocytes (CTLs) could be detected in the peripheral blood. These CTLs predominantly recognized products of the pX region of HTLV-I, had a CD8+ phenotype, and were human leukocyte class I restricted. Moreover, these responses were not detected in asymptomatic, HTLV-Iseropositive individuals. This implied a role for these CTLs in the pathogenesis of the neurological disorder associated with HTLV-I. We have extended these observations by demonstrating HTLV-I-specific CTLs directly from lymphocytes obtained from the cerebrospinal fluid of patients with HTLV-I-associated myelopathyltropical spastic paraparesis. Uncultured cerebrospinal fluid lymphocytes were used directly as effectors on a variety of targets expressing HTLV-I. These cells were lysed in a virus-specific and HLA class I-restricted manner. Moreover, the cerebrospinal fluid lymphocytes were sorted into purified CDS+ populations, cloned by limiting dilution, and assayed for CTL activity. An exceedingly high proportion of these resultant lines were shown to be cytolytic and precursor frequency analysis indicated that as many as 1 in 500 cells were HTLV-I-specific CTLs. The majority of these CTL lines recognized HTLV-I gene products encoded within the pX region of HTLV-I. The significance of these HTLV-Ispecific CTLs in the central nervous system of patients with HTLV-I-associated neurological disease is discussed with regard to the potential role of CTLs in the pathogenesis of this disease. Jacobson S, McFarlin DE, Robinson S, Voskuhl R, Martin R, Brewah A, Newell AJ, Koenig S. HTLV-I-specific cytotoxic T lymphocytes in the cerebrospinal fluid of patients with HTLV-I-associated neurological disease. Ann Neurol 1992;32:65 1-657

The human T-cell lymphotropic virus type I (HTLV-I) is endemic in many parts of the world, particularly in the Caribbean, the southern regions of Japan, South America, and recently has been shown to be present in the southeastern United States [l}. In all these areas, a slowly progressive chronic neurological disease termed HTLV-I-associated myelopathyltropical spastic paraparesis (HAM/TSP) has been observed in individuals seropositive for the virus 12-43. Not all HTLVI-seropositive individuals manifest this neurological disease. The vast majority, in fact, remain asymptomatic carriers of the virus and only 1% develop either HAM/TSP or the neoplastic disorder, adult T-cell leukemia 151. Why one group of HTLV-I-seropositive individuals develops neurological symptoms, another develops a neoplastic condition, and a third group remains as asymptomatic carriers is not known. A number of possibilities including genetic factors, virus mu-

tations, or differences in host-immune responses could explain this spectrum of HTLV-I-associated conditions. Recently, high levels of circulating HTLV-I-specific cytotoxic T lymphocytes (CTLs) have been shown to be present in the peripheral blood of patients with HTLV-I-associated neurological disease {GI. These CTLs were CD8+, human leukocyte (histocompatibility leukocyte antigen [HLA)) class I restricted, and, although not exclusively, predominantly recognized the HTLV-I gene products encoded in the HTLV-I regulatory region, pX. Importantly, these CTLs were not demonstrable directly from the peripheral blood of HTL.V-I-seropositive, asymptomatic individuals [GI. This suggested that HTLV-I-specific CTLS may contribute to the pathogenesis of HTLV-I-associated neurological disease. We postulated that if these CTLs do play a role in the pathogenesis of this disorder, then

-

From the *NeuroimmunologyBranch, National Institutes of Health, Bethesda, and tMedImmune Inc, Gaithersburg, MD.

Address correspondence to Dr Jacobson, NIH/NINDS, Building 10, Room 5B-16, Bethesda, M D 20892.

Received Jan 17, 1992, and in revised form Mar 6 and Apr 28. Accepted for publication May 3, 1992.

65 1

these cells would be expected to be present within affected tissues of the central nervous system (CNS). Support for this idea came from previous observations that in 1 patient with TSP there was an infiltration of CD8' cells within the spinal cord [73. In this study, additional evidence is presented for an immunopathogenic role of HTLV-I-specific CTLS in HTLV-I-associated neurological disease. HTLV-Ispecific CTLs were demonstrated directly from lymphocytes present in the cerebrospinal fluid (CSF) of affected individuals. CSF lymphocytes from patients with HAMITSP were able to lyse HTLV-I-infected targets in a virus-specific and HL.A class I-restricted manner. Moreover, CTL. precursor frequency analysis of a subset of patients indicated a high frequency of cytotoxic CD8+ CSF lymphocytes specific for HTLV-I.

Materials and Methods Patients Four patients with HTLV-I-associated neurological disease were admitted to the Neuroimmunology Clinical Service at the National Institutes of Health. Patient 1 was a 34-year-old black man who developed HAM/TSP and had a history of intravenous drug abuse and sexual promiscuity. Laboratory studies included high titers of HTLV-I antibodies in the serum and CSF. His CSF contained 86 white blood cells/mm3, a protein of 83 mg/dl, albumin quotient of 11.2 (normal, 3.2-9.0), IgG index of 1.1 (normal, 0.26-0.62), and oligoclonal bands. Patient 2 was an HTLV-I-seropositive white man with no history of transfusion or drug abuse, but who frequently traveled to HTLV-I-endemic areas. He had weakness and spasticity in both lower extremities. The CSF showed 10 lymphocytes/mm3, total protein of 76 mgidl, glucose of 62 mg/dl, albumin quotient of 5.0, IgG of 7.2 mg/ dl, IgG index of 0.58, and oligoclonal bands were present. Patient 3 was an HTLV-I-seropositive 45-year-old woman who has been described previously {6}. Patient 4 was a 68year-old man who received a transfusion of HTLV-Icontaminated blood and subsequently developed HAM/ TSP. HTLV-I antibodies were found in serum and CSF. His CSF contained 3 lymphocytes/mm3, total protein of 43 mg/ dl, IgG was increased to 8.4 mg/dl with an index of 1.20, and oligoclonal bands were present. All patients were seronegative for hunian immunodeficiency virus (HIV).

Cytotoxicity Asiays PERIPHERAL BLOOD LYMPHOCYTES. Unfracttonated peripheral blood lymphocytes (PBLs) were separated on Ficollsodium metritoate gradients (LSM, Organon Technika, Durham, NC) and directly assayed on the day of separation. Fresh PBLs were used as effectors at the indicated effector-totarget (E:T) ratio in a 4-hour 51Crrelease assay. Supernatants were harvested in a Skatron harvesting system (Skatron, Inc, Lier, Norway). Percent specific lysis was calculated as described [b}and is presented as the mean of triplicate assays. Standard deviations were less than 5%. Spontaneous release ranged between 15% and 25%. As targets, 1 x lo6 autologous Epstein-Bm virus-transformed lymphoblastoid cell

lines (LCLs) were infected with various HTLV-I vaccinia recombinants I61 at a multiplicity of infection of 10.0 for 10 to 14 hours and then incubated with 0.2 rnCi of Na451Crfor 1.5 hours. Target cells were used at a final concentration of 5 x lo3cells/well. Target cells infected with the vaccinia-p27x vaccinia recombinant expressed the HTLV-I pX gene products, p27x (rex), p40x (tax), and p21x; vaccinia-p40x expressed p40x (tax) and p2 Ix; vaccinia-p2 l x expressed p2 Ix; vaccinia-gag expressed p19, p24, and p l 5 internal structural proteins; vaccinia-e&vexpressed gp62, gp46, and gp21 surface glycoprotein; control vaccinia-Ha - expressed no HTLV-I proteins. Long-term T-cell lines derived from patients with HAM/TSP in whom HLA phenotypes were known were also used as targets. These lines expressed HTLV-I and have been described previously {8} and all lines expressed comparable levels of HTL.V-I on cell surface membranes. Lymphocytes were obtained from CSF by atraumatic lumbar punctures in which less than 1 red blood cell/mm3 was present. When sufficient lymphocytes were obtained (no less than 3 x lo' CSF cells in a total volume of 20 to 25 ml of CSF), these were used directly in a 4-hour 51Crrelease assay at the indicated effector-to-target ratio as described above. Target cells were used at a final concentration of 1 x lo3 cells/well. Targets consisted of autologous LCLs infected with the various HTLV-I vaccinia recombinants or HLA-typed allogeneic T-cell lines expressing HTLV-I (see above). CSF lymphocytes (or PBLs) were also sorted into purified CD8+ populations on an Epics flow cytometer (Coulter, Hialeah, FL) and cloned by limiting dilution as described previously {9, 101. The percentage of CD8+ cells in the total CSF or PBL population was also determined by flow cytometry. Purified CD8+ lymphocytes were plated in 96-well Costar plates (Cambridge, MA catalog no. 3799) at final concentrations of 100, 50, 25, 12.5, 6, and 3 cells per well in 0.2 ml of media containing RPMI plus 10% fetal calf serum (FCS). Allogeneic feeders (SOOOR) were added to each well at a concentration of 1.0 x lo5cells/well. Feeders were prestimulated with phytohemagglutinin (GIBCO, Grand Island, NY) at a final concentration of 1 pg/well for 30 minutes at 37°C. Twenty-four hours later, 100 pl of supernatant from each well was removed and replaced with 100 p1 of fresh media containing RPMI plus 10% FCS, 10% purified interleukin-2 (IL-2) (Cellular Products, Buffalo, NY) and 100 U/ml of recombinant IL-2 (Cellular Products, Buffalo, NY). All wells were refed twice per week with IL-2-containing media. After 2 to 3 weeks, cell buttons became prominent and lines derived from this cloning were screened for cytotoxicity against 51Cr-labeled autologous LCLs infected with the HTLV-I vaccinia-p40x, gag, e m , or vaccinia-Ha - constructs. Precursor frequency analyses were determined by cloning cells in limiting dilution and testing individual wells for HTLV-I-specific Iysis of autologous LCLs infected with the HTLV-I o r HA-control vaccinia recombinants. Individual wells demonstrating >20% HTLV-I-specific lysis were screened as positive. Results were plotted graphically (number of cells versus percentage of negative-responding wells) and the precursor frequency was extrapolated from a point

CEREBROSPINAL FLUID LYMPHOCYTES.

652 Annals of Neurology Vol 32 No 5 November 1992

-

Effectors: Patient \ Fresh PBL Targets: LCL infected with HTLV-I Vacclnia Recombinants

-

-

1

\

100:1

a's

25:l Effec1or:Target

so:1

100:1

50:l

25:l

C

-

HLA Class

I match (644)

-

I

HLA Class I match (644)

4

-

Effectors: Patient II Fresh PBL Targets: HTLV-I Expressing T cell Llnes

Effectors: Patient I Fresh PBL Targets: HTLV-I expressing T cell lines

-

12.5:l

12.5:l

Effect0r:Target

A

-

Effectors: Patlent II Fresh PBL Targets: LCL Infected wlth HTLV-I Vacclnla Recomblnants

\

HLA Class II metch (DR2,3) HLA mlrmitch

0

0

100:1

B

50:l

25:l

1oo:i

12.5:l

Effec1Or:Target

50:l 25:l Effector:Target

12.5:l

D

Fig 1. (A, C). Peripheral blood leukocytes (PBLJ) isolatedfrom Patients 1 and 2, respectively, were incubated with autologous B cells infected with the various human T-cell &mphotropic virus type I (HTLV-I) vaccinia recombinants as described in Materials and Methods. The vaccinia constructs are s h w n in each jigure. (B) PBL from Patient 1 were incubated with HTLV-Iexpressing T-cell lines. Open symbols represent a T-cell line histocompatibility leukoqte antigen (HLA) mismatched and a T-cell line HLA class I1 matched (DR2,3) with Patient 1. Closed symbols represent 2 T-cell lines H L A class I matched (B44) with Patient 1. (0)PBL from Patient 2 were incubated with HTLV-I-expressing T-cell lines. Open symbols represent 2 T-cell lines HLA mismatched and 1 T-cell line HLA class I1 matched (DIG'). Closed circles represent 1 T-cell line HLA class I matched (A23,B7) with Patient 2. LCL = lymphoblastoid cell line.

where 37% of cultures were unreactive for HTLV-I. Wells displaying >20% specific lysis remained HTLV-I specific when expanded and retested. Precursor frequency for the total CSF or PBL lymphocyte population was calculated as the percentage of CD8+ cells times the precursor frequency of the CD8+ population.

Results Demonstration of HTLV-I-specific Cytotoxic T Lymphocytes Directly from Lympboqtes in the Cerebrospinal Fluid

The presence of circulating HTLV-I-specific CTLs in freshly isolated peripheral blood lymphocytes was demonstrated in Patients 1 and 2 as described previously in other patients with HTL.V-I-associated neurological disease 161. Figure 1 shows that HTLV-I-

Jacobson et al: HTLV-I CTL in CSF of HAM/TSP Patients 653

~~~

--

Effectors -Patient I Fresh CSF Lymphocytes Targets: HTLV-1 expressing T cell Lines & LCL Infected with Vaccinia recombinants

Effectors: Patient I CSF Lymphocytes Targets: HTLV-I expressing T cell lines 30,

-

I

I

I

H U Cbs1 I N t e h (BU)

--

20

; i ap

\

1

HLA Class I match (814)

I \

I

7 HLA Class 11 match (DR 2,3)

n 50:l

12.5:l

v.se-pag

3.1.1

1 2 5:l Etfector:Tsrget

50:l

E1fector:TargeI

3.1:l

C

A

-

-

30

I

-

Effectors: Patient IV Fresh PEL Targets: LCL Infected with HTLV-I Vaccinia Recombinants

-

Effectors: Patient II CSF Lymphocytes Taregts: HTLV-I expressing T cell IlneS

3\

I \ 40:l

R

10:1

100:1

2.5:l

Effector:Targel

Figure 2. (A) Lynphocytes were obtained from cerebrospinal fluid (CSF) of Patient 1 and incubated with human T-cell !ymphotropic virus type I (HTLV-+expressing T-cell lines. Open circles represent a T-cell line human leukocyte (HLA) class I1 matched (DR2,3)with Patient 1. Closed circles represent a T-cell line HLA class I matched (844) with Patient 1. (B) CSF lymphocytesfrom Patient 2 were incubated with HTLV-I-expressing T-cell lines HLA mismatched (open circle) or HLA clajs I matched (A23,B7) (closed circles) with Patient 2. (C) CSF lymphocytes were obtained from Patient 1 , 1 month after the results shown in Figure 2A. These cells were incubated with 2 HTLV-I-expressing T-cell lines HLA class I matched (B44) (dosed triangles) and I T-cell line HLA class I1 matched (DR2,3) (open triangle) with Patient 1 . Autologous B cells infected with the HTLV-I vaccinia recombinants, p40x, env, gag, and H a - , were also used as targets. (D) Peripheral blood leukocytes (PBLs) isolated from Patient 4 were incubated with autologous B cells infected with the various HTLV-I vaccinia recombinants (shown in Figure 2 0 ) .

50:l 25:l Effector:Target

1251

D

specific CTLs could be demonstrated in the peripheral blood that predominantly recognized HTLV-I products encoded within the pX region (lysis of autologous B cells infected with the vaccinia-p27x and p40x constructs; Fig lA, C) as well as env (lysis of the vacciniaenv construct). To assess the HLA restriction of these circulating HTLV-I-specific CTLs, a panel of HLAtyped, long-term T-cell lines expressing HTLV-I was used as targets. As shown in Figure IB and D, only those lines matched at an HLA class I allele were lysed. Targets that were HLA mismatched or only matched at HLA class I1 with the circulating CTL effectors were not killed (see Fig 1B). HLA-matched T-cell lines not expressing HTLV-I were not lysed (data not shown). Because of a high pleocytosis in Patient 1 (86 cells/ mm3) and a milder pleocytosis in Patient 2 (10 cells/ mm3), it was possible to assess the CTL activity directly from circulating lymphocytes present in the CSF. HTLV-I-expressing long-term T-cell lines of a known HLA type were used as targets (these were also used in Figure 1B and D). As shown in Figure 2A and B,

654 Annals of Neurology Vol 32 No 5 November 1992

lymphocytes from the CSF lysed only the HLA class I-matched target, but not the HTLV-I-expressing target either matched at HLA class 11 or HLA mismatched. The presence of CTLs directly in the CSF of Patient 1 was reevaluated approximately 1 month later. The CSF still showed a pleocytosis (16 cells/mm3) and these CSF lymphocytes lysed two HTLV-I-expressing targets matched at HLA class I, but did not lyse the HLA class 11-matched target (Fig 2C). Moreover, autologous B cells infected with either the vaccinia-p40x Or vaccinia-env were 'ysed (see Fig but not B cells infected with the vaccinia-gag or vacciniainfluenza Ha - construct. These data indicate that HTLV-I-specific, HLA class I-restricted CTLs could be demonstrated directly from lymphocytes present in the CSF of patients with HTLV-I-associated neurological disease. 2c)9

Demonstration of HTLV-I-specific Cytotoxic T Lymphocytes fmm Lymphocytes in the Cerebrospinal Fluid as Defined by Limiting Dilution Cultures Typically, insufficient numbers of lymphocytes are present in the CSF to assay for direct CTL activity on all possible effectodtarget combinations. To determine frequency of CTLs in the CSF that are directed at HTLV-I, techniques were adopted in which CSF lymphocytes were sorted by flow cytometry into purified CD8' populations and subsequently cloned by limiting dilution. The resultant lines were screened for HTLV-I-specific CTL activity by their capacity to lyse autologous B cells infected with either the vacciniap27x, p40x, env, or influenza Ha - constructs. Initially, this approach was developed with CSF containing only 5 cells/mm3 from an HTLV-I-seropositive individual with neurological disease (Patient ST {6], here described as Patient 3) previously shown to have strong HTLV-I-specific CTL activity from lymphocytes in her peripheral blood [6]. The results in the Table show that a high proportion of purified CD8+ CSF cloned lymphocytes were CTLs that recognized the HTLV-I p40x construct in the context of HLA A2 (see legend to the Table). The purified CD8+ PBLs from Patient 3 plated in limiting dilution also contained a high proportion of HTLV-I p40xspecific CTLs (see Table). These results were confirmed from CSF lymphocytes obtained from HAM/TSP Patient 4. Initially, circulating HTLV-I-specific CTLs were detected from fresh PBL of this patient (Fig 2D). The number of lymphocytes in the CSF was insufficient for the assessment of direct CTL activity. However, the CSF lymphocytes that were obtained ( 5 cells/mm3) were able to be sorted into purified CD8' cells and cloned by limiting dilution following the approaches used for Patient 3. HTLV-I-specific CTLs were also demonstrated in these CSF lines (see Table) again, with a high

Precursor Freqizency of HTLV-I-spec& C T L in the Peripheral Blood and CerebrospinalFlizid in Patients with HTLV-I-associated Neurological Disease Precursor Frequencya

CD8+ CSFb Patient 3 4

Total CSF Lymphocytes

1/60 11195

11125 11488

CD8+ PBLb Total PBL 1/86 11100

11390 11333

aLymphocytes from the peripheral blood and cerebrospinal fluid (CSF) were into pur&ed CD8+ cells, cloned by limiting &Iution, and precursor frequencies were calculated a~ described in Materials and Methods. bCD8t cells represent 48% and 22% of the total CSF lymphocytes and peripheral blood lymphocytes (PBLs), respectively, for Patient 3, and 40% and 3096, respectively, for Patient 4. The majority of Patients 3 and 4's cytotoxic T lymphocytes (CrrS) were specific for the pX region of HTLV-I. These responses were shown to be predominantly HLA A2 and A3 restricted, for Patients 3 and 4, respectively (data not shown). HTLV-I = human T-cell lymphotropic virus type I; HLA compatibility leukocyte antigen.

=

histo-

CTL precursor frequency, comparable with PBLs (see Table). Analysis of the HTLV-I gene products recognized by these CTLs indicated that the majority were specific for HTLV-I env although some lines also reacted with the HTLV-I p40x construct. Sorting of CD8' cells and subsequent cloning by limiting dilution of CSF lymphocytes from Patient 2 was also attempted, but because of poor cloning efficiency it was not possible to determine an HTLV-I CTL precursor frequency. In this experiment, however, T-cell lines did grow from wells plated at 100 cells per well and 16 of 40 lines tested were cytotoxic for the HTLV-I p40x region. This demonstrated a high proportion of HTLV-I-specific CTLs in the CSF of patients with HTLV-I-associated neurological disease.

Discussion This study demonstrates the presence of CD8+, HTLV-I-specific CTLs in the CSF of patients with HTLV-I-associated neurological disease. Cells migrating into the CSF may more closely reflect events in the CNS than PBLs. Therefore, experimental observations made with these CSF cells should provide insight about the immunological events in the CNS of affected patients. Previously, HTLV-I-specific CTLs were shown to be present in the peripheral blood of HTLV-Iseropositive individuals with neurological disease. These responses could not be found in asymptomatic, HTLV-I-seropositive spouses or carriers [GI. On the basis of these findings, it was postulated that HTLV-Ispecific cytotoxic responses are related to the clinical and pathological disorders associated with HAMITSP. If HTLV-I-specific CTLs contribute to the pathogenesis of HAM/TSP, several conditions must be ful-

Jacobson et al: HTLV-I CTL in CSF of HAMITSP Patients

655

filled. Because all T-cell-mediated immune responses depend on recognition of antigen in the context of HLA molecules through interaction with antigenspecific receptors 111, 121, the three components of T-cell recognition (viral antigen, HLA molecules, and the T cell) should be demonstrated at lesion sites. Some, but not all of these conditions have been met by work reported to date. Both the HTLV-I p19 core protein and HLA class I molecules have been demonstrated in spinal cord lesions of 1 patient with HAM/ TSP 181. CD8' cells have also been observed in these lesions {8}. These previous reports, coupled with the results from the present study demonstrating functional CD8+, HTLV-I-specific CTLs in the CSF of patients with HAM/TSP, support the view that a component of these CD8' cells in HAM/TSP spinal cord lesions may be HTLV-I-specific CTLs. If so, these lymphocytes may cause direct lysis of infected cells in the spinal cord, which could explain the pathological and clinical changes observed in this disease 113, 141. Moreover, production of high levels of cytokines such as tumor necrosis factor-&,which can affect the replication of retroviruses as well as the function of lymphocytes, macrophages, and glia 115, 167, might also contribute to the impairment of neurological function. The hypothesis of an immunopathological role for HTLV-I-specific CTLs in the pathogenesis of HTLVI-associated neurological disease is supported further by the high precursor frequency of the CD8+, HTLVI-specific CTLs in the CSF. Precursor frequencies of CTLs for common viruses (i.e., measles, influenza, and mumps) in the peripheral blood are typically in the range of 1/10,000 to 1/100,000 of total lymphocytes 1171. By contrast, HTLV-I-specific CTL frequencies of total lymphocytes of 1/500 or greater (see Table) were observed in the blood and CSF of patients with HAMITSP. Although the presence of CTLs in the CNS of these patients supports a role for these cells in the pathogenesis of this disease, it still remains uncertain as to when and in which compartment, i.e., periphery or CNS, the initial activation of these CTLs occur. The high precursor frequency of CD8+, HTLV-I-specific CTLs in the CSF could be due to localized expansion of these CTLs within the CNS or increased migration from the peripheral blood. Alternatively, increased HTLV-Ispecific CTLs may be a response to a high viral burden present in patients with HAM/TSP [b}. The possibility that HTLV-I-specific CTLs contribute to the pathogenesis of HTLV-I-associated neurological disease implies that immunotherapeutic strategies could be beneficial in the treatment of this disorder. The elimination or reduction of retroviral antigens with antiretroviral agents might be possible, particularly as new agents are identified for the treatment of HIV infections. Alternatively, the use of drugs with

antilymphocytic effects that would reduce the number of HTLV-I-specific CTLs or interfere with their function could be beneficial. In this regard, HAM/TSP Patient 1 in the present study has been treated with high-dose steroids 1181, and remarkable neurological improvement has been observed with a concomitant reduction (22% specific lysis before steroids, 3% specific lysis after) in HTLV-I-specific CTLs as detected in PBLs. Other more tailored immune-mediated therapies might include elimination of particular T-cell. subsets, mimicking some strategies that have been successful in the treatment and prevention of T-cell-mediated experimental allergic encephalomyelitis [19-23}. Hopefully, by defining the immunopathogenic effector T cell or the antigenic nature of the retroviral target, or both, effective therapies can be designed to intervene in HTLV-I-associated neurological disease.

References 1. McFarlin DE, Blattner WA. Non-AIDS retroviral infections in humans. Annu Rev Med 1991;42:97-105 2. Gessain A, Vernant JC, Maurs L, et al. Antibodies to human T-lymphotropic virus I in patients with tropical spastic paraparesis. Lancet 1985;2:407-410 3. Rodgers-Johnson P, Gajdusek DC, Morgan OS, et al. HTLV-I and HTLV-111 antibodies and tropical spastic paraparesis. Lancet 1985;2:1247-1248 4. Osame M, Usuku K, Itumo S, et al. HTLV-I associated myelopathy: a new clinical entity. Lancet 1986;1:1031-1032 5. Murphy EL, Hanchard B, Figueroa JP, et al. Modelling the risk of adult T cell leukemidlymphoma in persons infected with human T-lymphotropic virus type I. Int J Cancer 1989;43: 250-25 3 6. Jacobson S, Shida H, McFarlin DE, et al. Circulating CD8' cytotoxic T lymphocytes specific for HTLV-I pX in patients with HTLV-I associated neurologic disease. Nature 1990;348: 245-248 7. Wayne Moore GR, Traugott U, Schienberg LC, Raine CS. Tropical spastic paraparesis: a model of virus induced, cytotoxic T-cell-mediated demyelination? Ann Neurol 1989;26:523-5 30 8. Jacobson S, Raine CS, Mingioli ES, McFarlin DE. Isolation of an HTLV-I like retrovirus from patients with tropical spastic paraparesis. Nature 1988;331:540-543 9. Koenig S, Earl P, Powell D, et al. Group specific, major histocompatibility complex class I restricted cytotoxic responses to human immunodeficiency virus 1 (HIV-1) envelope proteins by cloned peripheral blood T cells from an HIV-1 infected individual. Proc Natl Acad Sci USA 1988;85:8638-8642 0. Moretta A. Frequency and surface phenotype of human T lymphocytes producing interleukin 2. Analysis by limiting dilution and cell cloning. Eur J Immunol 1988;15:148-155 1. Davis MM, Bjorkman PJ. T-cell antigen receptor genes and T-cell recognition. Nature 1988;334:395-402 2. Rothbard JB, Gefter ML. Interactions between immunogenic peptides and MHC proteins. Annu Rev Immunol 1991;9:527565 3. Mattson DH, McFarlin DE, Mora C , et al. Central nervous systems lesions detected by magnetic resonance imaging (MRI) of tropical spastic paraplegia. Neurology 1987;2:49 14. Cruickshank JK, Rudge P, Dalgleish AG, et al. Tropical spastic paraparesis and human T cell lymphotropic virus type I in the United Kingdom. Brain 1990;112:1057-1090

656 Annals of Neurology Vol 32 No 5 November 1992

15. Poli G, Kinter A, Justement JS, et al. Tumor necrosis factor a functions in an autocrine manner in the induction of human immunodeficiency virus expression. Proc Natl Acad Sci USA 1990;87:782-785 16. Vyakarnam A, McKeating J, Meager A, Beverly PC. Tumour necrosis factor (a,b) induced by HIV-1 in peripheral blood mononuclear cells potentiate virus replication. AIDS 1991;4: 21-27 17. McFarland HF, Goodman A, Jacobson S. Virus specific cytotoxic T cells in multiple sclerosis. Ann N Y Acad Sci 1988;532: 273-279 18. Osame M, Matsumoto M, Usuku K, et al. Chronic progressive myelopathy associated with elevated antibodies to human T lymphotropic type I and adult T-cell leukemia-like cells. Ann Neurol 1987;21:117-122 19. Wraith DC, McDevitt HO, Steinman L, Acha-Orbea H. T cell

20.

2 1.

22.

23.

recognition as the target for immune intervention in autoimmune disease. Cell 1989;57:709-715 Wraith DC, Smilek DE, Mitchell JD, et al. Antigen recognition in autoimmune encephalomyelitis and the potential for peptide mediated immunotherapy. Cell 1989;59:247-25 5 Sakai K, Zamvil SS, Mitchell DJ, et al. Prevention of experimental encephalomyelitis with peptides that block interaction of T cells with major histocompatibility complex proteins. Proc Natl Acad Sci USA 1982869470-9474 Howell MD, Winters ST, Olee T, et al. Vaccination against experimental allergic encephalomyelitis with T cell receptor peptides. Science 1989;246:668-670 Vandenbark AA, Hashim G, Offner H. Immunization with a synthetic T-cell receptor V-region peptide protects against experimental allergic encephalomyelitis. Nature 1989;341:541543

Jacobson et al: HTLV-I CTL in CSF of HAM/TSP Patients

657